Performance Measurement and Management Control : Measuring and Rewarding Performance 9781849505710, 9780762314799

Citation preview

PERFORMANCE MEASUREMENT AND MANAGEMENT CONTROL: MEASURING AND REWARDING PERFORMANCE Marc J. Epstein Jean-Franc- ois Manzoni Editors

STUDIES IN MANAGERIAL AND FINANCIAL ACCOUNTING Series Editor: Marc J. Epstein Recent Volumes: Volume 1:

Setting the Standard for the New Auditors Report: An Analysis of Attempts to Influence the Auditing Standards Board

Volume 2:

The Shareholders Use of Corporate Annual Reports

Volume 3:

Applications of Fuzzy Logic and the Theory of Evidence to Accounting

Volume 4:

The Usefulness of Corporate Annual Reports to Shareholders in Australia, New Zealand, and the United States: An International Comparison

Volume 5:

A Power Control Exchange Framework of Accounting Applications to Management Control Systems

Volume 6:

Throughout Modeling: Financial Information Used by Decision Makers

Volume 7:

Applications of Fuzzy Sets and the Theory of Evidence to Accounting II

Volume 8:

Corporate Governance, Accountability, and Pressures to Perform: An International Study

Volume 9:

The January Effect and Other Seasonal Anomalies: A Common Theoretical Framework

Volume 10: Organizational Change and Development in Management Control Systems: Process Innovation for Internal Auditing and Management Accounting Volume 11: US Individual Federal Income Taxation: Historical, Contemporary and Prospective Policy Issues Volume 12: Performance Measurement and Management Control: A Compendium of Research Volume 13: Information Asymmetry: A Unifying Concept for Financial and Managerial Accounting Theories Volume 14: Performance Measurement and Management Control: Superior Organization Performance Volume 15: A Comparative Study of Professional Accountants’ Judgements Volume 16: Performance Measurement and Management Control: Improving Organizations and Society Volume 17: Non-financial Performance Measurement and Management Practices in Manufacturing Firms: A Comparative International Analysis

STUDIES IN MANAGERIAL AND FINANCIAL ACCOUNTING VOLUME 18

PERFORMANCE MEASUREMENT AND MANAGEMENT CONTROL: MEASURING AND REWARDING PERFORMANCE EDITED BY

MARC J. EPSTEIN Jesse H. Jones Graduate School of Management Rice University, Texas, USA

JEAN-FRANC - OIS MANZONI IMD, Lausanne, Switzerland

United Kingdom – North America – Japan India – Malaysia – China

JAI Press is an imprint of Emerald Group Publishing Limited Howard House, Wagon Lane, Bingley BD16 1WA, UK First edition 2008 Copyright r 2008 Emerald Group Publishing Limited Reprints and permission service Contact: [email protected] No part of this book may be reproduced, stored in a retrieval system, transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without either the prior written permission of the publisher or a licence permitting restricted copying issued in the UK by The Copyright Licensing Agency and in the USA by The Copyright Clearance Center. No responsibility is accepted for the accuracy of information contained in the text, illustrations or advertisements. The opinions expressed in these chapters are not necessarily those of the Editor or the publisher. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-7623-1479-9 ISSN: 1479-3512 (Series)

Awarded in recognition of Emerald’s production department’s adherence to quality systems and processes when preparing scholarly journals for print

Marc J. Epstein

CONTENTS LIST OF CONTRIBUTORS

ix

PREFACE

xiii

PART I: MEASURING AND REWARDING PERFORMANCE ALIGNING, MEASURING, AND REWARDING PERFORMANCE IN COMPLEX ORGANIZATIONS Marc J. Epstein

3

ON THE FOLLY OF HOPING FOR A, SIMPLY BECAUSE YOU ARE TRYING TO PAY FOR A Jean-Franc- ois Manzoni

19

PERFORMANCE MEASUREMENT AND MANAGEMENT CONTROL SYSTEMS: CURRENT RESEARCH AND IDEAS GOING FORWARD Antonio Davila

43

PART II: EVALUATING AND MOTIVATING EMPLOYEE PERFORMANCE EARLY PERFORMANCE RATINGS AND PERSONALITY MEASURES IN CAREER ADVANCEMENT Mark A. Mishken and Krisztina Juhasz v

73

TYING VS. NOT-TYING POST-COMPLETION REVIEWS TO EXTRINSIC REWARDS Stefan Linder OBJECTIVE REWARDING, MANAGERIAL MOTIVATION AND ORGANIZATIONAL COMMITMENT: THE INTERVENING ROLE OF JUSTICE Frank G.H. Hartmann and Sergeja Slapnicˇar

95

127

PART III: EVALUATING PERFORMANCE OF ORGANIZATIONAL UNITS EMPIRICAL EVIDENCE ON THE (PERCEIVED) VALUE OF INFORMATION FOR PERFORMANCE MEASUREMENT PURPOSES IN AN ERPS ENVIRONMENT Xavier Gabrie¨ls and Ann Jorissen THE MISSING LINK BETWEEN INFORMATION AND ACTION: HASTENINGS AND DELAYS AS UNIVERSAL REACTIONS TO PERFORMANCE FEEDBACK Tom De Schryver, Rob Eisinga, Christine Teelken and Erik Poutsma THE LIFETIME VALUE SCORECARD: FROM E-METRICS TO INTERNET CUSTOMER VALUE M. Bonacchi, M. Ferrari and M. Pellegrini

vi

147

177

193

EXAMINING THE CONSTRUCT VALIDITY OF THE BALANCED SCORECARD USING THE MULTITRAIT–MULTIMETHOD MATRIX Emilio Boulianne

227

ASSESSING PRIVATIZED AIRPORT PERFORMANCE FROM STAKEHOLDER VIEWPOINTS: A STUDY OF SYDNEY AIRPORT Dorothea Zakrzewski

253

CAUSALITY IN A PERFORMANCE MEASUREMENT MODEL: A CASE STUDY IN A BRAZILIAN POWER DISTRIBUTION COMPANY Andre´ Carlos Busanelli de Aquino, Ricardo Lopes Cardoso, Marcelo Sanches Pagliarussi and Vale´ria Lobo Archete Boya

273

PART IV: REWARDING PERFORMANCE PERFORMANCE MEASUREMENT AND EXECUTIVE COMPENSATION: PRACTICES OF HIGH-PERFORMANCE COMPANIES Belverd E. Needles, Marian Powers and Mark L. Frigo PRIVATE PERFORMANCE INFORMATION IN CEO INCENTIVE COMPENSATION Eduardo Schiehll

vii

303

323

This page intentionally left blank

LIST OF CONTRIBUTORS M. Bonacchi

Department of Business Administration, University of Napoli ‘‘Parthenope’’, Napoli, Italy

Emilio Boulianne

John Molson School of Business, Concordia University, Montreal, Quebec, Canada

Vale´ria Lobo Archete Boya

FUCAPE Business School, Vitoria, ES, Brazil

Andre´ Carlos Busanelli de Aquino

FEARP University of Sa˜o Paulo, Ribeirao Preto, SP, Brazil

Ricardo Lopes Cardoso

EBAPE Getulio Vargas Foundation, Praia de Botafogo, Rio de Janeiro, RJ, Brazil

Antonio Davila

IESE Business School, University of Navarra, Spain

Tom De Schryver

Radboud University Nijmegen, Nijmegen, The Netherlands

Rob Eisinga

Radboud University Nijmegen, Nijmegen, The Netherlands

Marc J. Epstein

Rice University, Houston, TX, USA

M. Ferrari

Department of Social, Cognitive and Quantitative Sciences, Faculty of Communication and Economics Science, University of Modena and Reggio Emilia, Reggio Emilia, Italy

Mark L. Frigo

DePaul University, Chicago, IL, USA ix

x

LIST OF CONTRIBUTORS

Xavier Gabrie¨ls

Business Economics, European University College Brussels, Brussels, Belgium; University of Antwerp, Antwerp, Belgium

Frank G.H. Hartmann

RSM Erasmus University, Rotterdam, The Netherlands

Ann Jorissen

University of Antwerp, Antwerp, Belgium

Krisztina Juhasz

New York State Office of Court Administration, New York, NY, USA

Stefan Linder

CTcon GmbH, Bonn, Germany

Jean-Franc- ois Manzoni

International Institute for Management Development (IMD), Lausanne, Switzerland

Mark A. Mishken

New York State Office of Court Administration, New York, NY, USA; Pace University, New York, NY, USA

Belverd E. Needles

DePaul University, Chicago, IL, USA

Marcelo Sanches Pagliarussi

FUCAPE Business School, Vitoria, ES, Brazil

M. Pellegrini

DADA mobile (DADA Group), Firenze, Italy

Erik Poutsma

Radboud University Nijmegen, Nijmegen, The Netherlands

Marian Powers

Northwestern University, Evanston, IL, USA

Eduardo Schiehll

HEC Montreal, Montreal, Quebec, Canada

Sergeja Slapnicˇar

Faculty of Economics, University of Ljubljana, Ljubljana, Slovenia

xi

List of Contributors

Christine Teelken

Radboud University Nijmegen, Nijmegen, The Netherlands

Dorothea Zakrzewski

College of Business, School of Accounting, University of Western Sydney (UWS), Centre for Innovation and Industry Studies, Australia

This page intentionally left blank

PREFACE In 2001, we gathered a group of researchers in Nice, France to focus discussion on performance measurement and management control. Following the success of that conference, we held subsequent conferences in 2003, 2005, and 2007. This volume contains some of the exemplary papers that were presented at the most recent conference. The conference has grown in number of participants, quality of presentations, and reputation and this year attracted leading researchers in the field from North America, South America, Europe, Asia, Australia, and Africa. Though the conference has been generally focused on performance measurement and management control and has included presentations on many facets of the topic, each year we have also focused on a particular theme of current interest. This year’s theme was directed at measuring and rewarding performance. This includes evaluating and rewarding the performance of organizations, units, teams, and individuals. It includes empirical, analytical, and experimental research. The last few years have seen an explosion of academic research and managerial interest in performance evaluation and rewards. Some of this is related to the dramatic increase in compensation among senior corporate executives and associated discussions related to the drivers and measures of success. Some of it relates to whether and how measurement and rewards are to be linked for improved performance and how critical the performance evaluation system is in driving superior performance. There were three plenary sessions at this conference and the papers are included here. Marc Epstein presented three current research projects and examined the relationship between alignment, performance measurement, and rewards discussing both formal and informal systems. Jean-Francois Manzoni looked more carefully at the necessity to focus more extensively on the informal systems that have often been neglected in the management control discussions. Tony Davila presented a summary of the 120 papers presented in the concurrent sessions and used this data to discuss the current state of research in the field. In addition to the three plenary sessions, this volume also includes some of the other excellent papers presented at the conference. The call for papers xiii

xiv

PREFACE

drew a wonderful response of 250 submissions, so the competition to make a presentation at the conference was quite high. Further, given the space limitations in this book, another competitive selection was required. The contents of this book represent a collection of leading research in management control and performance measurement and provide a significant contribution to the growing literature in the area. This collection of papers also covers a representative set of topics, research settings, and research methods. From the first year, the conference has relied heavily on EIASM and Graziella Michelante for organization and management and their enthusiastic participation and excellent work has been critical to the conference success. This year we also welcomed a new Conference Chairman, Antonio Davila, who provided excellent intellectual and administrative leadership to the conference. We thank them and all of the speakers and participants at the conference. Their attendance and enthusiastic participation made the conference an enjoyable learning experience. We are hopeful that this book will continue the search for additional understanding and development in performance measurement and management control, and provide guidance for both academic researchers and managers as they work toward improving organizations. Marc J. Epstein Jean-Franc- ois Manzoni Editors

PART I: MEASURING AND REWARDING PERFORMANCE

This page intentionally left blank

ALIGNING, MEASURING, AND REWARDING PERFORMANCE IN COMPLEX ORGANIZATIONS Marc J. Epstein ABSTRACT The management control and performance measurement literature reflects a long history of discussion related to organizational, team, and individual rewards. Yet, much of the research and guidance in the academic and managerial literature has been inadequate. Reflecting work on three current research studies, this chapter examines the gaps in our current understanding of the relationship of performance measurement, rewards, and performance and suggests some research questions that are of significant interest.

Much of our academic research in performance measurement and management control has focused on the premise that when organizations more carefully align strategy, structure, systems, performance measures, and rewards they become more successful. Research reported in the previous conferences here in Nice (Epstein & Manzoni, 2002, 2004, 2006) along with leading authors (for example, Simons, 1995, 2000, 2005; Kaplan & Norton, 2004, 2006) provide support for this premise. In previous contributions to this conference along with other articles (Epstein & Performance Measurement and Management Control: Measuring and Rewarding Performance Studies in Managerial and Financial Accounting, Volume 18, 3–17 Copyright r 2008 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-3512/doi:10.1016/S1479-3512(08)18001-7

3

4

MARC J. EPSTEIN

Westbrook, 2001; Wisner, Epstein, & Bagozzi, 2006; Davila, Epstein, & Shelton, 2006) I have also both proposed and tested models of the importance of alignment of these management control, performance measurement, and rewards systems. But, managers continuously report that they find these models very difficult to implement. If this alignment is so well accepted and important and the processes well defined why does it remain so challenging? In three of my current research projects, I have been struck by the commonalities of the challenges and the general lack of recent academic research and managerial best practices that are available to resolve these challenges. After decades of focus in both the research and managerial press around the design and implementation of management control and performance measurement systems, there is still much work to be done to successfully address these issues. It remains challenging in both for-profit and non-profit organizations, in various organizational functions, and in business and geographical units. And, it is challenging in well-managed organizations with smart, dedicated, and diligent managers. Three very different types of organizations and organizational functions – and surprisingly similar research projects and challenges – provide evidence of the pervasiveness of these challenges. In examining the gaming industry, the microfinance industry, and the corporate social responsibility (CSR) function in organizations, the differences in the typical organizational designs and the significantly different missions and goals are readily apparent.

Primary Organizational Missions

financial performance

dual or blended

social performance

The gaming industry along with most other for-profit institutions is primarily oriented toward excellence in financial performance. The microfinance industry and most non-profit or socially oriented organizations are primarily focused on achieving social goals. The CSR (or sustainability) function in most organizations, in contrast, is typically faced with achieving multiple goals through a blended or dual mission. Though on

Aligning, Measuring, and Rewarding Performance in Complex Organizations

5

the surface these organizations seem quite different, some similarities are striking and the challenges are critical and unresolved. And, the examples that are provided here are from the leading organizations in the world with smart and dedicated managers. Yet, these managers are often measuring and rewarding items that appear to be inconsistent with their mission. They are challenged to design management control, performance measurement, and reward systems that will drive high performance and superior success in achieving their mission. So, one challenge is to identify those principles of management control, performance evaluation, and reward systems that transcend for-profit and non-profit organizational distinctions at various organizational and individual levels.

GAMING INDUSTRY Las Vegas, Nevada in the United States has been dominated by the gaming (casino) industry since its early days. For most of the one hundred year history of the city, the burgeoning hotel industry focused on inexpensive accommodations, food, and alcohol to attract customers while earning its profits from gambling. The activities, processes, and revenues were all oriented toward gambling. In the last decade, dramatic changes have occurred. The industry is now dominated by two large companies with many large hotels (MGM Mirage and Harrah’s Entertainment). And, the orientation is that these are entertainment businesses with a wide variety of offerings rather than only casino operations. Until recently, the vast majority of both revenue and profit was earned from the casino floor. But, current industry estimates are that only about 35% of total revenues come from gaming, while the other 65% are derived from non-gaming attractions such as hotel rooms, entertainment venues, pools and spas, and restaurants (Campbell, Epstein, & Martinez-Jerez, 2006). Even though the habits and demands of Las Vegas casino and resort customers have changed dramatically over time, the way in which most of these companies measure and manage customer profitability has not adapted to these changes. Customer profitability is still calculated and managed primarily based only on gaming related performance measures. Thus, casino hosts, the employees who manage the customers’ experience and reward profitable patrons, focus their efforts on enhancing gaming related performance measures even though the strategy has changed to more explicitly focus on the business as entertainment rather than gaming and the revenue sources have changed dramatically. Thus, though these employees

6

MARC J. EPSTEIN

could be important in generating significantly increased entertainment revenues and profits, their rewards are typically based exclusively on a small portion of the total customer experience and spending. So, since entertainment and hotel revenue is not measured, it is often not included in the reward systems that are developed for the customer facing employees (hosts). Since only a portion of customer profitability is measured and rewarded, many casinos and resorts in Las Vegas are not fully aligning their strategy of customer orientation with the way customer and organizational performance is measured and rewarded – and may not be achieving these objectives. Employee and customer turnover often exceeds 50% and neither customer nor corporate profitability is being effectively managed. Just as with employees, rewards for customers are also often not aligned with the strategy and mission of the organization. (Complimentaries (comps) of free or discounted rooms, food, and entertainment are typically rewarded exclusively on gaming activity.) These issues deserve attention. Further, there are many unanswered questions as to what decision rights should be provided to the casino hosts, whether individually based or team based compensation would provide better results, whether performance measures and rewards should be primarily based on individual unit performance (individual hotels) or group performance, and whether these should be short-term or long-term evaluations. Some of the casinos have begun to use decision analytics, utilizing improved data-collection technology and analysis to derive the maximum amount of value from its business processes (Davenport & Harris, 2007). Through the use of decision analytics, Harrah’s Entertainment, for example, has been able to more fully discern what their customers want, how much each customer is willing to pay for various services, and what keeps each customer loyal. It has also begun linking employee incentives with key metrics related to customer loyalty and satisfaction (Loveman, 2003). Field experiments can be more extensively used to provide answers to some of the questions above and both managers and researchers can learn much from these experiments. Research and practice in management control, marketing, and human resources suggest that an organization’s profitability is often driven by employee satisfaction and loyalty, which then affects customer loyalty, retention, and satisfaction, resulting in increased profitability. By empowering employees by increasing decision rights, and implementing incentive programs that more fully align themselves with performance measures articulated by the organizations’ strategies, companies

Aligning, Measuring, and Rewarding Performance in Complex Organizations

7

like those in Las Vegas can improve employee retention, customer retention, and customer profitability articulated in their corporate strategy. Then, why have not these very well-managed organizations done this? Why is this so difficult? It may be that this is more complex than many of our management control and performance measurement models describe. It may be that the alignment literature does not adequately address all of the important variables and that reward systems and formal processes alone are often inadequate. It may be that informal systems of culture and people and more effective utilization of intrinsic rewards are also critical. For-profit corporations are likely to be mostly on the left side of the continuum above. Thus, it would appear that these should be among the easiest missions to implement since the mission, strategy, and goals are clear. And, alignment should be easier to attain. Though there are fewer challenges in for-profit organizations, research and practice in the gaming industry demonstrates in part the deficiencies in our current models and that more complex and nuanced approaches are necessary for success.

MICROFINANCE INDUSTRY In our last conference here in Nice, I presented a model that describes the drivers and measures of microfinance success (Epstein, 2006). I had recently visited microfinance clients in Ghana to better understand the drivers and measures of microfinance success. Subsequently, along with coauthors, I began a larger research project with a new model of the mission, strategy, implementation, and measures of success. The study has included surveys, interviews, archival data, and field experiments. Whereas most microfinance research over the past 30 years has focused on the success of the microfinance institution, our current research is focused on identifying the drivers and measures of client success (Datar, Epstein, & Yuthas, 2008). When for-profit corporations have a financial mission and are on the far left side of the continuum above, it is not surprising that they are focused primarily on their own success (even though they may be dependent on satisfying client needs). But, the mission of non-profit organizations should be primarily socially oriented as the ultimate outcome and they should be on the right side of the continuum. Thus, their focus should be primarily oriented toward the success of the client or beneficiary rather than the financial success of the institution. We find that in many cases this is contrary to common current industry practice.

8

MARC J. EPSTEIN

Microfinance, which has been used as a tool for alleviating global poverty for about 40 years, usually refers to loans of approximately $100–500 to groups of 5–40 people who guarantee each others’ loans. Though microfinance often also includes a wider set of financial services such as savings and insurance, the loans are typically business loans, provided to individuals (mostly women) to start or expand their businesses. Microfinance institutions ultimately have a social mission, aiming to provide financial and social transformative opportunities for impoverished peoples in the areas in which they operate. Though initially established almost exclusively as non-profit institutions, they are increasingly converting to forprofit regulated banks primarily to access the capital markets for greater growth and liquidity and to provide a full range of financial services. Though this can provide greater access to capital and new needed services to the poor, it can also provide new challenges in staying focused, managing, and achieving the primary social mission. To ensure the success of microfinance institutions in achieving their mission of alleviating poverty, it is imperative that the organization’s mission be clearly articulated and aligned with performance measurement and incentives at the various organizational levels. After an extensive review of microfinance related literature and interviews with microfinance clients, Epstein and Crane (2007) found that most studies on microfinance focus on measures such as organizational growth, client satisfaction and retention, and default rates as the primary measures of organizational success. These are measures that, while important for the financial sustainability of the organization, do not effectively address the social mission of microfinance. Organizations primarily with social missions such as those within the microfinance industry should be implementing performance measures that are aligned with their social mission and employee incentives should be based upon measures that motivate social performance as well as financial sustainability. Aligning staff incentives such as promotions to social goals has been found to contribute positively to loan officer motivation, retention, and productivity (de Aghion & Morduch, 2005). But, that is not enough. The design of a management control system would certainly suggest that the rewards be consistent with the mission. In this case, they are not only inconsistent and do not support the mission, they may actually be in direct conflict with the mission. And, providing the incentives for loan officers to be committed to a social mission must rely on culture, people, intrinsic motivation, and other forms of informal processes in addition to the formal systems in place.

Aligning, Measuring, and Rewarding Performance in Complex Organizations

9

Many microfinance institutions, however, continue to motivate their loan officers through rewards based upon financial indicators and compensation such as default rates and portfolio growth, ignoring compensation based on social measures such as success of the micro-enterprise, increased household income of the client, or improved school attendance, nutrition, housing, or self-esteem. This leads to an increased risk of a preoccupation with profitability at the expense of poverty reduction and other social goals such as women’s empowerment, resulting in mission drift (Copestake, 2006). And, it is not only the loan officers that are measured and rewarded based on financial rather than social performance. The institutions and their senior executives are similarly evaluated. The focus has been almost entirely on short-term, easy to measure goals even though they are not the primary mission of the organizations. Though we do know that it is common for organizations and individuals to gravitate toward the easier to measure financial measures rather than the often more difficult to measure and relevant non-financial performance measures, it is particularly critical in non-profit organizations where this not only leads to ineffective management practices, but goals that are contrary to the organizational mission. And, like the gaming industry, employee and customer turnover often exceeds 50%. In addition to an increased risk of mission drift, linking loan officer compensation to default rates and portfolio growth also results in strategic misalignment because repayment and default rates are measured at the group level. Because group clients guarantee each others’ loans, it is certain that individual default rates are higher than what is reflected by group default rates. Like the hosts that manage customer experience within the Las Vegas’ casino and resort industry, the measurement of performance and incentives which drive the actions of employees within the microfinance industry are misaligned with the company’s strategy of improving customer experience and retention. So, what can be done to increase loan officer and client retention and drive improved performance that is consistent with the organizational mission? What can improved management control and performance measurement systems do to increase success? This requires not only focus on the mission but the design and implementation of systems to provide consistency throughout the organization. Thus, mechanisms are required – along with organizational culture and people that are supportive. That is, both formal and informal systems are required. Too often, our management control literature has emphasized the formal systems and the alignment of strategy, structure, systems, performance measures, and rewards as the

10

MARC J. EPSTEIN

solution to effective management control and strategy execution. Increasingly, we must explicitly acknowledge the importance of the informal systems that are also necessary for success. Loan officers in this socially oriented industry have personal goals that are also important for their satisfaction. Whereas they may have joined the industry to achieve social goals, they spend much of their efforts collecting loan payments from delinquent clients. They would like to provide more support to the clients but need additional training to be effective. They would like to focus their activities on client success, but their performance measures and rewards are based on institutional success. Their value is measured on default rate and loan growth rather than achieving the social mission. Realignment can be accomplished through the use of a mix of both formal and informal systems. For the formal systems, performance measures and rewards can be realigned. In addition, additional investment in people and culture can also change the informal systems that exist. And, particularly important in this context, the organization can be reoriented to provide for significantly more engagement between the loan officer and the client. So first, this issue is partly about alignment around customer focus and increasing client and loan officer retention. Second, it is about focusing more clearly on the organizational mission and avoiding mission drift. Third, it is about developing performance evaluation and reward systems throughout the organizations that are consistent and supportive rather than in direct conflict with the missions. (This is the case for both the microfinance industry and the gaming industry!) Our new research in microfinance is intended to determine and develop specific products, services, and actions that are likely to lead to increased success and new approaches to measuring success though field experiments with randomized control groups and new training, guidance, and support approaches for the loan officers and clients.

CORPORATE SOCIAL RESPONSIBILITY FUNCTION The managers in the CSR (or sustainability) function in organizations are faced with a different challenge. On the continuum above, they are challenged by a difficult paradox of trying to achieve excellence in both social and financial performance simultaneously. In recent years, CEOs have increasingly acknowledged the importance of sustainability for (a) fulfilling responsibilities to communities, (b) increasing shareholder value, and

Aligning, Measuring, and Rewarding Performance in Complex Organizations

11

(c) improving both social and financial performance. So, then, why is sustainability implementation so challenging? It is because implementing sustainability is fundamentally different. For operating goals, the direct link to profit is usually clear. For innovation and new product development, though long term and difficult to predict and measure, the intermediate goal is developing new products and the ultimate goal is increased profit. In these cases, companies set missions and strategies and develop aligned systems, structures, culture, performance measures, and rewards. The standard management control and performance measurement models for implementation can be applied. For sustainability, the goal is to simultaneously achieve excellence in both social and financial performance. Measuring and managing this paradox creates significantly more challenges. First, it is often unclear how to make the tradeoffs. Second, it is often unclear how the stakeholders will respond to managerial actions. Third, both corporate and societal priorities often change. And fourth, the costs of implementing sustainability constantly changes. So, the result is that the standard implementation approaches often fail. Excellence in strategy and leadership are minimum enablers of success in sustainability. Companies need a clear sustainability strategy and committed senior leadership for effective sustainability implementation. Companies can use formal and informal processes – hard and soft implementation tools – for effective execution. So, what factors determine the role of each of these tools in the implementation of strategy? How critical are the performance measurement and reward systems? What we have found is that relevant and formal performance evaluation and reward systems are often needed at various organizational and individual levels. But, informal systems may be just as critical. So, to achieve implementation success in sustainability it may be that the following is necessary: (1) Sustainability must be an integral component of strategy. (2) Leadership must be committed to sustainability and building additional organizational capacity. Actions are more difficult to specify so distributed leadership is more critical. (3) This must be supported with management control and performance measurement and reward systems as appropriate. It should also be supported with mission, culture, and people as appropriate. The choices and balance of the formal and informal processes will depend on the strategy, organizational design, systems in place, culture, people, and many other factors.

12

MARC J. EPSTEIN

(4) This must all be used to implement learning throughout the organization to provide guidance and support to other managers as to how to make the tradeoffs and make the challenging managerial decisions. Managers must integrate sustainability into all strategic decisions. Then, additional systems and rewards can be introduced to formalize and support. Despite the intended benefits that CSR actions hope to achieve through improved social and environmental performance, implementation can be surprisingly difficult. Because the goal of CSR is to simultaneously achieve excellence in both social and financial performance, it is often unclear as to how to make tradeoffs between the social and financial performance of the company, especially when employees have substantial incentive pressure and are compensated on performance that does not include social (or environmental) measures. Because of the complexity of the implementation of sustainability and the paradox and challenges described above, implementation efforts have often failed. Even companies and their senior managers that have committed themselves to sustainability have been often unable to do it successfully. The traditional formal systems of management control, performance measurement, and rewards has not been successful and a new model that relies more on a combination of formal and informal systems may be required. A broader understanding of these challenges and the need for new approaches may drive the actions that can achieve greater success simultaneously in financial, social, and environmental performance. (For more on the implementation of sustainability in organizations, see Epstein, 2008.)

A MORE BALANCED MODEL OF IMPLEMENTATION Management control and performance measurement research has often relied too heavily on the formal systems of organizational design, organizational systems, performance evaluation, and rewards to both drive and explain organizational performance. It has often focused on the important role of compensation and rewards to drive improved performance. It has relied on these tools to motivate employees and develop strategic management systems to implement strategy. But, a more balanced approach to implementation may be needed. Sure, rewards can be important. But, so can culture and people and providing employees with the intrinsic motivation to achieve individual and organizational goals. This

Aligning, Measuring, and Rewarding Performance in Complex Organizations

13

is particularly true in non-profit organizations where employees often accept lower salaries to participate in a social mission. So, both intrinsic and extrinsic rewards are important elements of the drive for superior organizational performance. An organization’s incentive system is one way that it often allocates extrinsic rewards and punishments in order to affect the purposeful behavior of its members. Thus, while the purpose of an organization may be to create value, the purpose of an organization’s incentive system is to motivate value-creating behavior (Hall, 2002). So, though tying employee compensation to tangible measures of performance and output can help motivate employee effort (Baron & Kreps, 1999), it is not the only motivator. Much more focus on the use of informal systems and the role of culture and people is critical. Management control researchers must examine informal systems effects in addition to the formal systems when evaluating drivers of success. Though compensation and explicit incentives are important, the implicit motivators can be just as critical. And, though we often rely on the notion of the alleged importance of incentives linked to performance measures articulated by organizational strategy, little research has been done that empirically addresses the effectiveness of linking incentives to performance measures. But, that which has been completed shows positive results regarding incentives as subsequently affecting employee actions (Indjejikian, 1999). This stresses the importance of aligning company strategy and performance measures with incentives that motivate employee behavior. Successful implementation of incentives that align with the performance measures established by an organization’s strategy and mission can motivate individuals to act in the best interest of the company, resulting in increased organizational performance. But, identifying those incentives is not clear and the balance between formal and informal systems and financial and non-financial motivators is also unclear. Further, they do require that the strategy being aligned must be properly defined. In some cases (as outlined above) the stated mission and strategy and the one being tacitly implemented may be quite different. The main goal of any reward system is to affect motivation. The motivation through an effective reward system is only one form of motivation that drives individuals to create value within their organization. Empowering employees through the implementation of increased decision rights, in addition to strategy-aligned incentives, is another important component of potentially increasing and maintaining employee satisfaction as well as increasing organizational value (Hall, 2002).

14

MARC J. EPSTEIN

Despite the importance of alignment between incentives and rewards with company strategy, many note that this alignment is not often discussed or implemented within organizations (Kerr, 2004). As a result, rewards structures and the performance measures on which these structures are based often fail to align with the organization’s overall mission and strategy. Many managers have witnessed incentives which drive behavior that is not in the best interest of the organization’s strategy. This comes as a result of rewards systems based upon performance measures that do not fully represent the organization’s strategy and objectives. This is in part because most jobs require multiple tasks and require decisions based on a variety of different factors. Performance in these tasks are sometimes easy to measure, and sometimes not. An organization that implements its reward systems based on narrow, specific performance measures that are easy to measure will motivate behavior that is geared toward excelling in those particular measures. As a consequence, the implementation of incentives based upon narrow, easy-tomeasure performance metrics often motivates performance that is incomplete or in conflict with the organization’s overall strategy (Kohn, 1993). This is a major challenge for all organizations, whether they are for-profit, non-profit, or a hybrid.

WHAT CAN RESEARCHERS DO? So, significantly increased research is needed in management control, performance measurement, and rewards to better understand how to drive improved performance in complex organizations. The needed studies are examples of some of the serious gaps in our knowledge and impacts both current research and managerial success. Some of the needed research would include: (1) Examine the factors that drive success – conceptually and empirically. (2) Better understand mission drift in for-profits and non-profits and how performance evaluation and rewards can help retain organizational focus. (3) Examine the role of performance evaluation and rewards in driving organizational success. What is the role of formal and informal (hard and soft) processes? (4) Develop better measures for short-term and long-term success – especially the challenging non-financial metrics.

Aligning, Measuring, and Rewarding Performance in Complex Organizations

15

(5) Identify and test the causal relationships and implementing this in organizations. Surveys, archival data, interviews, observation, and field experiments can all be used. (6) Empirically test the relationships between performance, performance measures, and incentives. When this is most effective is not clearly established in the research. (7) Examine the individual and organizational barriers to alignment and organizational change. As mentioned above, the examples cited here have been from industry leaders, rather than laggards. Though we have made progress in better understanding and explaining management control and performance measurement, the challenges remain both critical and prevalent.

SUMMARY Many management control and performance measurement systems have been developed in both research and practice. And, some have been extensively tested. Yet, organizations of all types still have significant difficulty designing and implementing performance evaluation and reward systems at various organizational and individual levels. And, the success of these systems is often unclear. Similarities and challenges among the three very different types of organizations discussed here are striking. Challenges exist around performance evaluation, customer and employee retention, reward systems, and organizational alignment around mission. And, we find that mission drift can occur in both for-profit and non-profit organizations alike. Management control researchers have often proposed that careful alignment will resolve these issues. But though the reward systems are important, it is the combination of formal and informal systems that is critical. Managers have significantly more choices than are typically described in the management control and performance evaluation and rewards literature. And, leading managers use these choices to create flexibility and achieve excellence in performance – sometimes relying more on informal processes than the formal processes. This is too often ignored in our literature and research. Though, we often describe the importance of compensation, we also see that organizations can succeed without heavy reliance on the reward systems. But, they typically fail if they are in direct conflict with organizational goals.

16

MARC J. EPSTEIN

These issues are very broad and generic. The issues exist with many types of organizations and at various units of analysis and organizational levels. They exist with boards of directors and top managers and with business unit, geographical, and functional managers. From an organizational perspective, both for-profit and non-profit organizations face the same challenge of implementing incentive and rewards systems that are based upon measures aligned with the organization’s strategy. The examples discussed above illustrate that these challenges are pervasive throughout different types of organizations, emphasizing the need for managers to address the alignment of their employee incentive structures with performance measures and organizational strategy. Individuals can be motivated both intrinsically and extrinsically through leadership that inspires empowerment, as well as the implementation of incentives that are aligned with strategic measures of performance. The effective linkage of a firm’s performance measures with employee performance goals and incentive programs can help drive performance that is aligned with the organization’s overall mission and strategy. But, it is often not enough. Broader understanding, use, and integration of a combination of both formal and informal systems must be included in both organizational research and practice to drive improved performance and achieve organizational goals.

REFERENCES Baron, J. N., & Kreps, D. M. (1999). Strategic human resources: Frameworks for general managers. New York, NY: Wiley. Campbell, D., Epstein, M. J., & Martinez-Jerez, F. A. (2006). Slots, tables, and all that Jazz: Managing customer profitability at the MGM grand hotel. Harvard Business School Case Study, Case N-9-106-029, September, pp. 1–25. Copestake, J. (2006). Mainstreaming microfinance: Social performance management or mission drift? Draft. Department of Economics and International Development, University of Bath, Bath, January 13. Datar, S., Epstein, M. J., & Yuthas, K. (2008). In microfinance, clients must come first. Stanford Social Innovation Review (Winter). Davenport, T. H., & Harris, J. G. (2007). Competing on analytics. Boston, MA: Harvard Business School Press. Davila, T., Epstein, M. J., & Shelton, R. (2006). Making innovation work how to manage it, measure it, and profit from it. NJ: Wharton School Publishing. de Aghion, B. A., & Morduch, J. (2005). The economics of microfinance. Cambridge, MA: MIT Press. Epstein, M. J. (2006). Improving organizations and society: The role of performance measurement and management control. In: Epstein & Manzoni (Eds), Performance

Aligning, Measuring, and Rewarding Performance in Complex Organizations

17

measurement and management control: Improving organizations and society. London: Elsevier Science. Epstein, M. J. (2008). Making sustainability work best practices in managing and measuring corporate social, environmental and economic impacts. UK/San Francisco, CA: Greenleaf Publishing/Berrett-Koehler Publishers, Inc. Epstein, M. J., & Crane, C. A. (2007). Alleviating global poverty through microfinance: Factors and measures of financial, economic, and social performance. In: Rangan, Quelch, Herrero & Barton (Eds), Business solutions for the global poor. San Francisco, California: Jossey-Bass Publishers. Epstein, M. J., & Manzoni, J. F. (Eds) (2002). Performance measurement and management control: A compendium of research. London: Elsevier. Epstein, M. J., & Manzoni, J. F. (Eds) (2004). Performance measurement and management control: Superior organizational performance. London: Elsevier. Epstein, M. J., & Manzoni, J. F. (Eds) (2006). Performance measurement and management control: Improving organizations and society. London: Elsevier. Epstein, M. J., & Westbrook, R. A. (2001). Linking actions to profits in strategic decision making. MIT Sloan Management Review (Spring), 39–49. Hall, B. J. (2002). Incentive strategy within organizations. Harvard Business School Case, 9-902131, March 13. Indjejikian, R. J. (1999). Performance evaluation and compensation research: An agency perspective. Accounting Horizons, 13(2), 147–157. Kaplan, R. S., & Norton, D. P. (2004). Strategy maps converting intangible assets into tangible outcomes. Boston, MA: Harvard Business School Press. Kaplan, R. S., & Norton, D. P. (2006). Alignment using the balanced scorecard to create corporate synergies. Boston, MA: Harvard Business School Press. Kerr, S. (2004). Executives ask: How and why should firms and their employees set goals? Academy of Management Executive, 18(4), 122–142. Kohn, A. (1993). Why incentive plans cannot work (September). Harvard Business Review. Loveman, G. (2003). Diamonds in the data mine (May). Harvard Business Review. Simons, R. (1995). Levers of control how managers use innovative control systems to drive strategic renewal. Boston, MA: Harvard Business School Press. Simons, R. (2000). Performance measurement & control systems for implementing strategy. NJ: Prentice Hall. Simons, R. (2005). Levers of organization design how managers use accountability systems for greater performance and commitment. Boston, MA: Harvard Business School Press. Wisner, P. S., Epstein, M. J., & Bagozzi, R. P. (2006). Organizational antecedents and consequences of environmental performance. Advances in Environmental Accounting and Management, 3, 143–167.

This page intentionally left blank

ON THE FOLLY OF HOPING FOR A, SIMPLY BECAUSE YOU ARE TRYING TO PAY FOR A Jean-Franc- ois Manzoni ABSTRACT Kerr’s (1975) examination of the ‘‘folly of rewarding A while hoping for B’’ led him to encourage organizations to align reward system and desired employee behavior. Since then, much of the accounting and control literature has increasingly reduced the reward system to one of its components – incentive compensation plans – and has increasingly ceased to examine other behavioral levers used by corporations, thus implicitly or explicitly treating measurement and reward as a sufficient condition to obtain desired employee behavior. This chapter considers the complexity of the reward system (including its inevitable subjective dimension) and discusses its role, in connection with other important managerial levers, in corporations’ broader efforts to shape employee behavior. The chapter concludes with a review of literature streams in economics and psychology, suggesting that an intense incentive alignment approach may be self-fulfilling and hence counter-productive.

When his article ‘‘On the folly of rewarding A while hoping for B’’ got reprinted twenty years later as an Academy of Management Classic, Performance Measurement and Management Control: Measuring and Rewarding Performance Studies in Managerial and Financial Accounting, Volume 18, 19–41 Copyright r 2008 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-3512/doi:10.1016/S1479-3512(08)18002-9

19

20

JEAN-FRANC - OIS MANZONI

Steve Kerr recalled that the original article almost did not get published. The reviewers disagreed on the article’s merits and the editor had to weigh in. He did so in favor of Kerr, and the article went on to become one of the most cited articles of its time. Kerr’s (1975, 1995) basic point was the following:  Human beings tend to be ‘‘rational’’. They tend to engage in behaviors that get rewarded, ‘‘often to the virtual exclusion of activities not rewarded’’.  Yet, Kerr added, ‘‘many reward systems y are fouled up in that the types of behavior rewarded are those which the rewarder is trying to discourage, while the behavior desired is not being rewarded at all’’. For example, we hope for stretch targets but we reward ‘‘making the numbers’’. In the article, Kerr reviewed examples from a variety of sectors, including medicine, universities, politics and business. A few of these examples were indeed cases of ‘‘an organization rewards A while hoping for B’’, where A is negatively correlated with B. That is obviously not terribly smart. More frequently, however, the example presented a slightly more subtle situation: The organization was rewarding something it wanted (say, A1), while hoping people would also produce other behaviors (A2 and A3), which were not perfectly positively correlated – and could even be negatively correlated – with A1. In most cases, the organization was in fact trying to reward the behaviors it was hoping to see, but was doing so imperfectly – typically using incomplete measure(s) of the relevant performance dimension(s). For example, an insurance company wants claim resolution that is speedy and accurate, uses measures for speed and accuracy but gets the accuracy part wrong and as a result, gets speed and excessive payouts to clients. This frequent case could be summarized as ‘‘on the folly of hoping for B, while rewarding a bad proxy for B’’. Or, ‘‘on the folly of hoping for a certain mix of A, B and C, while rewarding a different mix’’. But of course Kerr’s title was simpler and has stuck in people’s minds. The basic point is quite fair: If you want A, you should not reward B. The reward system should not ‘‘get in the way’’. The tag line could be: ‘‘Avoid conflict between what you desire and what you reward’’. There are of course several ways of avoiding conflict. One is to avoid tying rewards too directly with specific outcomes. Kerr concedes that employee behavior is not 100% determined by ‘‘formal rewards and punishments’’ and that ‘‘certainly it is true that in the absence of formal reinforcement

On the Folly of Hoping for A, Simply because You are Trying to Pay for A

21

some soldiers will be patriotic, some players will be team-oriented, and some employees will care about doing their job well’’. But, he adds, ‘‘in such cases the rewarder is not causing the behavior desired but is only a fortunate bystander. For an organization to act upon its members, the formal reward system should positively reinforce desired behavior, not constitute on obstacle to be overcome’’ (Kerr, 1995, italics added). So, in Kerr’s terms, ‘‘if you want it done, you should measure and reward it’’. The oft-heard aphorism ‘‘if it doesn’t get measured, it doesn’t get done’’ makes the same point in negative: Measurement and reward are necessary conditions to obtain desired behaviors from members of an organization. Over the past 20 years, the performance measurement and reward rhetoric has continued to evolve. First, the notion of reward has become progressively narrower. In many articles I read and in conversation with accounting and control colleagues, the question seems to have shifted from ‘‘Rewarding A’’ to ‘‘Paying for A’’, to ‘‘Designing an annual incentive plan based on A’’. Second, ‘‘if it doesn’t get measured, it doesn’t get done’’ has often become ‘‘what gets measured gets done’’, and ‘‘you get what you measure’’. The latter two sentences go much further than the first, as they position measurement and reward as sufficient conditions, when previously they were presented as necessary ones. A sufficient condition means you can get the desired result by simply measuring and rewarding for it. ‘‘What gets measured gets done’’y . This is a gross over-simplification. Studying and working with executives a significant proportion of my time, I see many things that are being rewarded but do not get done, and I see things that are not specifically rewarded still getting done. I also see organizations use a great variety of influencing mechanisms beyond the reward system. These two gradual shifts seem to underpin what I perceive to be a pervasive implicit acceptance of ‘‘incentive alignment’’ as the dominant paradigm of the accounting and control community. In fact, it sometimes feel as if ‘‘let’s get the incentives right’’ has become our only response. Illustrating this one-track orientation, a colleague was recently telling me that his university seemed to be unable to obtain enough collegiality and commitment from senior faculty. His question was: ‘‘How do you at IMD incentivize (sic) your senior faculty to stay engaged and contribute?’’. Well, we do offer incentives, but we also do many other things beyond rewarding people to get them to do things. Let me take these two issues in turn, starting with the breadth of the reward system.

22

JEAN-FRANC - OIS MANZONI

REWARDING PERFORMANCE BEYOND ‘‘ANNUAL VARIABLE COMPENSATION PLANS’’ One of Many Rewards There is no doubt that in a number of cases, variable compensation plans purely driven by quantitative measures of performance overwhelm any and all other considerations. The financial services industry, in particular, features every year a number of traders and investment fund managers who collect vast sums of money, based on the measurable returns their decisions produced. Clearly, for these individuals, bonus-related considerations must be extremely salient in their minds. I do not meet many millionaire traders. I do meet large numbers of executives, many of them very senior. Almost all are part of a variable compensation program of some sort, often several programs. Do they think about these plans? Sure! Do these plans influence (some of) their decisions? Very likely. But it is also very clear that these incentive compensation programs are but one of many rewards that managers are hoping – and working hard – to receive. Consider a few other formal and informal rewards:  salary increase (which has a number of advantages over variable pay, starting with recurrence but also often including various kinds of fringe benefits associated with fixed pay but not necessarily variable pay);  more autonomy/latitude on the job (less need to check with the boss before making decisions, and less second-guessing from the boss after making decisions);  resources under one’s responsibility (or attributed to one’s unit);  mobility (upward, i.e., promotion; but also functional or geographical, to broaden one’s skills and experience);  other development opportunities, both internal (e.g., training, task forces) and external to the organization (e.g., providing the individual with external visibility);  reputation within the organization – expression of professional respect, public support and protection; and  boss’s appreciation/satisfaction/recognition. These rewards are clearly not independent. The boss’s performance appraisal and, more generally, the quality of one’s relationship with one’s boss, influence to a non-trivial extent the distribution of the other rewards.

On the Folly of Hoping for A, Simply because You are Trying to Pay for A

23

An extensive body of research known as leader-member exchange theory confirms that employees enjoying higher quality relationships with their bosses are more likely to receive challenging task assignments, training opportunities, resources, information and support (Liden, Wayne, & Sparrowe, 2000; Scandura, Graen, & Novak, 1986). Reputation within the organization (itself influenced by the boss’s views/sponsorship) certainly also intervenes in the allocation of mobility, resources and autonomy. It is of course impossible to make a general statement regarding the relative importance for ‘‘managers’’ of variable compensation versus other rewards listed above. The relative importance probably depends on the relative size of the rewards and, in particular, of the incentive compensation potential, as well as personal preferences. There may also be hierarchical effects and differences across national cultures. At the risk of sounding naı¨ ve, I think I have rarely met managers prepared to commit reprehensible acts to secure their typical annual bonus. I have, however, met quite a few managers that would, or did commit reprehensible acts in order to stay in their boss’s good graces.

The Importance of ‘‘Expert Judgment’’ (aka Subjectivity) The accounting and control community has invested heavily in studying which quantitative performance measures should be linked with pay, and how. This is certainly an interesting subject. But it is not clear that variable compensation plans always function entirely ‘‘objectively’’ or that payments are 100% based on the achievement of non-renegotiated and unchanged quantitative targets. Again, the range of situations across hierarchical levels, organizations, industries and countries is simply too wide to allow a clear answer, but I have certainly encountered many companies where the incentive compensation routinely featured ‘‘individual qualitative goals’’ that provided the boss(es) with leeway for subjective assessment. I have also come across numerous cases where the supposedly quantitatively driven payouts were adjusted to reflect a state of nature perceived to be too different from the expectations the targets were based on, or the fact that the quantitative outcomes incorporated more exogenous influences than were deemed acceptable. Incentive compensation is hence not always 100% objective, as exemplified in the audit firm that Chapman (2007) studied and presented at the Conference, where he found a rather sophisticated process leading to individual awards for partners. This process was informed, but not

24

JEAN-FRANC - OIS MANZONI

mechanically driven by, a series of quantitative performance indicators, and Chapman likened it to an expert system. The importance of ‘‘expert judgment’’ is likely to be even greater for the other rewards listed above, most of which are determined or influenced by the boss. This is indeed what I found in a previous study (Manzoni, 1993), where I asked 50 (manufacturing) managers to assess to what extent each of six rewards depended on ‘‘my unit’s performance against quantitative targets’’, ‘‘my immediate superior’s overall judgment’’ and ‘‘other factors’’. Five of the six rewards were clearly perceived to depend more on boss judgment than on target achievement, with about three quarters of the sample attributing more weight to boss judgment (including about 40% who attributed twice as much weight to that factor). The one exception was ‘‘Reputation within the company’’, which was still far from ‘‘objectivity’’ with 46% of the respondents attributing more weight to boss judgment (including 31% attributing to it more than twice as much weight). The importance of subjectivity in performance assessment is also supported by a growing amount of evidence on the role of organizational citizenship behaviors (OCBs) in subordinate performance evaluation and reward. (OCBs are discretionary behaviors that are not explicitly recognized by the reward system and which ‘‘promote the effective functioning of the organization’’ (Organ, 1988). They include dimensions such as altruism, civic virtue and conscientiousness, which are difficult to assess quantitatively.) Several OCB studies were conducted in sales settings, where ‘‘sales’’ (a good measure of quantitative performance) would be expected to have a large impact on subordinate performance evaluations. In fact, (intrinsically subjective) OCBs generally account for a larger proportion of the variance in subordinate performance evaluations than objective sales productivity measures (MacKenzie, Podsakoff, & Fetter, 1991, 1993). That is not very difficult, though, as in various studies reviewed by MacKenzie et al. (1993), ‘‘objective sales productivity usually accounts for only a fraction of the variance in managerial evaluations of salesperson performance (in the range of 5% to 8%)’’. MacKenzie, Podsakoff, and Paine’s (1999) study of 987 multi-line insurance agents and 161 agency managers reports similar findings and adds that the relative explanatory role of OCBs was even higher for managers’ performance assessments than for agents’ assessments. After reviewing these and many more studies in this area, Organ, Podsakoff, and MacKenzie (2006) conclude that the impact of OCBs on

On the Folly of Hoping for A, Simply because You are Trying to Pay for A

25

managerial evaluations is strong and positive, and at least as great as the impact of in-role behavior. Podsakoff, MacKenzie, Paine, and Bachrach (2000) worded it even more strongly: ‘‘This suggests that OCBs accounted for substantially more variance in performance evaluations than objective performance’’. Such subjectivity is, simply, unavoidable. Even when bosses have at their disposal several quantitative measures of performance, they will have to aggregate subjectively the results across the various dimensions. Even if they committed ex ante to an objective aggregation process, there can be no guarantee that the target setting process was ‘‘objective’’. For example, we know that the target setting process allows managers to attribute more slack (and hence easier targets) to profit center managers whom they trust to make good use of it (Merchant & Manzoni, 1989). A final source of subjectivity, in the evaluation and reward process, are the centrally imposed quotas. Many organizations issue some form of guidance on the distribution of performance evaluation ratings (and other rewards associated with such ratings). The relative nature of such systems implies that a group of bosses must sit down and coordinate. In other words, they engage in a ‘‘horse trading process’’. Some would argue that such discussion processes can reduce the impact of any individual’s subjectivity and hence make the overall process more ‘‘objective’’. This may be true in environments where the process receives considerable time and attention from many parties, which seems to be the case in some investment banks (Eccles & Crane, 1988). But in many cases I have observed and discussed with executives, the process also introduces distorting factors, such as the (more or less empirically grounded) perception of the boss’s peers and the relative bargaining power of the various players, which are not necessarily strongly related to the subordinates’ performance. To summarize this first point: In practice, reward systems incorporate more subjectivity than much accounting and control research tends to consider. Annual incentive plans matter, but promotion, autonomy and development opportunities also constitute powerful rewards. Human perception (particularly that of the boss) determines or at least influences several of these other rewards. Incentive compensation plans themselves also often include subjective assessments and/or adjustments. As a result, reward systems remain largely subjective affairs. Performance measures can and do influence the outcome, but the ‘‘expert system’’ considers a much wider set of variables.

JEAN-FRANC - OIS MANZONI

26

INFLUENCING BEHAVIOR THROUGH ALL AVAILABLE LEVERS (AND HENCE BEYOND REWARD SYSTEMS) Having established that the reward system is much richer and more subjective than generally recognized, I now want to examine the other behavioral levers available to managers to influence employee behavior – most of which seem to me to be under-studied by the accounting and control community. To start with, we must agree that employee behavior is not random. It is influenced by individual factors, of course, but also by a series of ‘‘levers’’ that shape behavior by influencing employee willingness and ability to produce the behavior desired by management. There are several ways to describe such levers. Pascale and Athos (1981) provided a model emphasizing Seven Ss (Strategy, Structure, Systems, Skills, Staff, Style and Shared values). Galbraith (1995) proposed the ‘‘Star Model’’, featuring strategy, structure, process, rewards and people. I tend to use a slightly different framework, presented in Fig. 1. This framework posits that employee behavior is influenced by:  Key Performance Indicators (KPIs) and incentives, of course. We discussed this aspect in the previous section.

Top management behavior

Technology

Processes

Enough people Culture of the organization

Employee behavior

Long enough

Structure

KPIs & incentives

Fig. 1.

Skills

Forces Influencing Employee Behavior.

On the Folly of Hoping for A, Simply because You are Trying to Pay for A

27

 The organizational structure, which defines each individual’s boss, peers, clients and suppliers, and influences individuals’ organizational identity (who is ‘‘we’’?).  Processes, that is, the way work is being conducted and flows across individuals and units. Processes can be explicit or not, they can be managed to a greater or a lesser extent, but they can certainly enable or disable individual actions.  Top management behavior is of course part of the incentive-shaping system. A CEO who cares about customers will usually reward managers who do too. I list ‘‘top management behavior’’ separately from ‘‘KPIs and incentives’’, though, for two reasons: First, because top management behavior is not always aligned with the organization’s ‘‘official’’ KPIs and incentives. Second, because beyond its incentive value, top management behavior also has a modeling value; it legitimizes other managers behaving that way, and it can also demonstrate complex behaviors that subordinates may not yet master.  Technology is a broad term that includes here the information that is available and can (or cannot) be exchanged by individuals and units. An organization I was working with wanted its sub-units to behave less independently and instead cooperate to sell ‘‘solutions’’ to the customer. Unfortunately, the group was incapable of telling the divisions how much business it was doing, across all five divisions, with each client.  Individuals need to want to do things, they also need to be able to. That is the Skills dimension. Skills are typically affected by three types of flow: organizations hire new skills, eliminate redundant ones and train everybody else. Employee behavior can be changed reasonably rapidly by applying a strong and consistent set of forces on the employees. All large-scale changes I have studied, read about or worked on featured simultaneous and aligned changes along most or all of these levers. How rapidly employee behavior changes depends on several endogenous and exogenous factors, including how large the required change is, how deeply anchored are the previous behaviors, how powerful and aligned are the applied forces, how many individuals need to be affected, how geographically dispersed these individuals are and how easy it is to monitor behavior and enforce the changes. But if sufficient determination and force guide the effort, the behavior of many members of the organization can be changed rather quickly. And then, of course, there is the organization’s culture. ‘‘Culture’’ is a word that is very often mis-used. Managers often use it to refer to the

28

JEAN-FRANC - OIS MANZONI

organization’s ‘‘atmosphere’’, the ‘‘feel of the place’’. That is incorrect and unfortunate, because it misleads managers to think that the organization’s culture can be changed quickly. One can indeed change the atmosphere of an organization in a matter of days, but an organization’s culture is much more, and is much more robust, than today’s or tomorrow’s atmosphere. The culture of an organization is ‘‘the pattern of shared basic assumptions that the group learned as it solved its problems (y), that has worked well enough to be considered valid and, therefore, to be taught to new members as the correct way to perceive, think and feel in relation to those problems’’ (Schein, 1992). At any point in time, the culture of an organization influences the behavior of its members. In Fig. 1, this relationship is noted by the arrow flowing from right to left from culture to employee behavior. In light with this definition, it is clear that people who have worked for General Electric over the past 20 years are likely to behave differently than individuals who have been working for, say, a municipal government over the same period. Yet, culture does not develop out of thin air. It develops over time, based on patterns of behavior that people repeat again and again, in part because they are functional for the actors. To change the culture of the organization hence requires a lot more than modifying temporarily the behavior of some of its members. It requires re-shaping the behavior of enough individuals , for long enough to allow them to internalize the new behavior. This internalization process takes time. How much time again depends on several factors, but for non-trivial changes we typically talk in years rather than months. The IMD Example Let me now go back to the question I was asked regarding ‘‘how do you at IMD incentivize the faculty?’’. IMD is a good example for this discussion because (a) it is an academic work environment most readers will be familiar with and will hence be able to compare and contrast with their own, and (b) it represents an excellent illustration of the way behavior can be shaped by multiple behavioral levers. The following description and analysis is based in part on my personal experience over the past four years, but also on IMD President’s insightful accounts on the way the institution has been managed over the past 15 years (Lorange, 2002, 2008). First comes the identification of ‘‘desired behavior’’. What do we want faculty and staff to do, and how do we want them to behave?

On the Folly of Hoping for A, Simply because You are Trying to Pay for A

29

To be sustainable, this desired behavior must be connected to the organization’s ‘‘business model’’. IMD is a separate institution, without financial connection to a university or any other wealthy parent. We live off the revenues we generate, that is, mostly from the tuition fees we charge participants to attend our programs. In colloquial terms, ‘‘we eat what we kill’’. Desired behavior should also be connected with the organization’s mission, that is, to the way the perimeter of activities is defined. At IMD, we work with executives to help them face more effectively their challenges of today and tomorrow. This means helping these executives to become more willing and better able to keep developing themselves over time, but it can also include other types of activities closer to supporting strategy definition and/or deployment, to building organizational capabilities or improving the effectiveness of a top management team. To help executives face more effectively their current and future challenges, we must of course conduct research. Research is important at IMD, as reflected by the very significant financial resources we allocate to it. But it is a means to an end (where the end is helping executives), as opposed to an end in itself. We also need to devote significant resources to developing pedagogical material that helps us help executives (and flows as quickly as possible from our research activities). And of course we need to be good at working directly with executives in activities ranging from teaching in openenrollment programs to facilitating top management meetings. Last but not least, our small size and independence requires faculty and staff to be ‘‘low maintenance and collegial’’. We simply cannot afford layers of administration, politics or any other form of bureaucracy, nor the kind of adversarial relationship between faculty and staff that is present in many business schools. To obtain this behavior from faculty and staff, IMD does use the formal reward system. Very consistent with Kerr’s (1975) encouragement, IMD rewards what it hopes to observe.  Our salary band is fairly narrow (most faculty are paid within a 25–30% range) and positively correlated with age. Salaries are not adjusted every year (which is clearly facilitated by Switzerland’s relatively low inflation).  Faculty and senior staff share a ‘‘profit sharing’’ bonus pool, distributed as a percentage of salary.  Faculty is also eligible for an individual bonus (the total envelope of which is equal to the faculty profit sharing pool). This individual bonus is

30

JEAN-FRANC - OIS MANZONI

determined by the President and our informal Faculty Dean (i.e., he does not have that title but plays that role), based on three criteria. Research (50%), teaching (30%) and institutional contribution (20%). – Research performance is graded on a relative basis (four or five groups) based on a complex point system, where the number of points received for each publication is determined by the President and faculty dean, who read everyone’s outputs, based on the publication’s outlet and estimated ‘‘impact’’. This grading process is hence partly subjective. – Teaching performance is assessed by the same individuals, taking into account quality of delivery but also quantity and quality of innovation. Delivery is assessed on the basis of participant ratings, after the Faculty Dean removes each individual’s 10% or so weakest ratings received during the year, to allow for some unevaluated innovation. Innovation can pertain to pedagogical material or pedagogical process, and is hence assessed somewhat subjectively. Here, again, individual absolute performances are ranked in four or five ‘‘buckets’’. – Institutional contribution is assessed by the Faculty Dean based on an extremely complex accounting system, fed by faculty self-reports as well as data provided by individuals in charge of various processes (such as recruitment, faculty meetings, alumni department, etc.). We all know that the ultimate result is somewhat subjective, but we also know that the Faculty Dean’s ‘‘expert system’’ is informed by a considerable mass of information, in part provided by each of us.

I certainly do not want to under-estimate the importance of our compensation practices. The amounts involved are not trivial (in a good collective and individual year, the two bonuses could amount to 50% of annual salary), and the determination of the individual bonus has clearly been carefully thought through over the years. Consistent with a considerable amount of evidence that the perceived fairness of a process helps individuals accept the process’s outcomes (see Cropanzano, Bowen, & Gilliland, 2007 for a recent review), the faculty’s high degree of respect for, and trust in the Faculty Dean is also very helpful. Are we conscious of the existence of these bonuses? Absolutely, we are. Many of my faculty colleagues are surprisingly unaware of the individual bonus determination mechanisms (when asked, they often refer to ‘‘a complicated process and calculation run by Jim’’), but we basically all know the three performance dimensions that are being considered. As for the profit sharing system, its motivational impact is of course not due to any

On the Folly of Hoping for A, Simply because You are Trying to Pay for A

• • • •

“We, we, we, not me, me me” “Great!!” “Good can always be done better” President’s personal example

• Staffing=> interdependence • OWP program => innovation • Two meetings/yr with President & “Faculty Dean” • No areas • No rank

Fig. 2.

Faculty & Staff behavior

KPIs & Incentives

31

• Weekly memo on program enrollment • Transparency on individual teaching loads & research output Culture-nurturing mechanisms • X-mas party • OWP Party • Two IMD outings/year • Faculty & staff dinner • Old timers transmit history and explain “the way things are done around here” • Very careful recruiting • OWP as knowledge management mechanism • R&D funding

Shaping Faculty and Staff Behavior at IMD.

direct return on additional effort calculation, but it reminds all of us that we can only succeed if faculty cooperates with each other and staff. But I doubt this compensation framework would be enough to get us to do all the things we do. This framework is but one of many powerful behavioral shaping mechanisms the IMD management has set up over the years. Let me review a few of them, using the model presented above as an organizing framework (Fig. 2 for a summary):  Our structure: IMD has no departments, no areas. This does not eliminate ‘‘turf battles’’, but it drastically reduces their potential. We also have no formal ranks; everyone’s called Professor.  Among key processes, I want to select four that I believe are particularly influential. The first one is the program staffing process. We do not get assigned to programs, we get asked to contribute by each Program Director. We call it a ‘‘market system’’, although strictly speaking it is not really a market, as we do not allow for differential rates. It is really a voluntary system of mutual seduction and reciprocal obligations. Since most of us direct programs, we need colleagues to ‘‘teach for us’’. To induce colleagues to teach in programs we direct, we often find it necessary to teach in programs they direct. We are hence interdependent

32

JEAN-FRANC - OIS MANZONI

and all have to manage our goodwill accounts with colleagues, or at least with those colleagues whose cooperation we may need at some point. Another key process is a large executive development program we offer once a year, and where we try very hard only to use material developed during the year. This program acts as a powerful innovation engine. The process is led by the program director asking us in September to commit to certain sessions for next June. June is sufficiently far away, we all feel we can accomplish great things over the coming nine months and so it is not very difficult to get us to commit. (By the way, individual sessions are not rated in this program, and the overall atmosphere is very much one of mutual stimulation and experimentation. So we all know that this program is one where experimentation is not only less risky, but is in fact encouraged.) Once made, the commitment acts as a powerful behavioral driver. The program director also tends to ask pairs or trios of faculty to take charge of groups of sessions, thus further enhancing cooperation and exchange across faculty. A third very important process: The President and Faculty Dean meet with each faculty member twice a year for an hour. We review past, present and future, we discuss plans and resources; we get a chance to express concerns and grievances, and management gets an opportunity to share their appreciation and/or encourage us to increase or re-direct our efforts. These meetings are prepared by both parties, in particular by faculty who are required to review their activities in writing. These three processes contribute to a considerable exchange of information and favors among faculty members and between faculty members and management. The fourth key process pertains to our performance evaluation process relative to contract renewals. IMD does not have a tenure concept. (That is one of the very first things the President told me at our initial meeting: ‘‘Are you aware we don’t have tenure here?’’.) We have three or five year contracts, which can be renewed once before getting to what the Swiss law calls an ‘‘open contract’’, where formal periodic renewal is no longer necessary but employment can be discontinued if the employee no longer performs satisfactorily. Renewals and conversion to open contract involve some of the external information search that goes on in every business school, but it involves less of it – we tend to contact fewer people – and this external information is complemented by internal information that receives substantial weight in the decision-making process.  The behavior of our top management and, in particular, of our President, is also an important behavioral driver. He has a few mantras he repeats over and over again, to the point that they have become part of our

On the Folly of Hoping for A, Simply because You are Trying to Pay for A

33

collective vocabulary. He insists that we need to behave and talk about ‘‘we, we, we, not me, me, me’’. (John Kennedy had made this point more elegantly when he encouraged Americans not to ask what their country can do for them but rather what they can do for their country, but ‘‘we, we, we’’ works surprisingly well too!) He is reasonably systematically encouraging and warm toward faculty achievements, but also consistently repeats to the group that as successful as we have been, ‘‘good can always be done better’’ (and how can we do better next time? y). He also works extremely hard on behalf of the institution.  The information we report and discuss is also important. As probably most schools, we have various reports listing faculty research activities and outputs. We also have a weekly report showing registration and forecasts for all programs over the next 12 months. We obviously do not all read this report line by line every week, but I certainly look at it occasionally. Its arrival also acts as a weekly reminder of where revenues come from. Also on the information front, there is complete transparency (and we get regular reports) on the whole faculty’s teaching activities. We do not get program ratings (except for program in which we teach, of course), but we do get a list of who is doing how much in each program and over the year.  We also take very seriously the composition and skills of our faculty. I have already mentioned several of the mechanisms that encourage information exchange and stimulate cooperation. Actually, this process starts at the recruiting stage. We recruit very carefully, predominantly at the ‘‘mid-career’’ level. Professional qualifications and individual skills matter a lot, but so does the individual’s cultural fit. To assess this cultural fit, we work on several fronts, starting with a solid information search before discussing any invitation (a process that can run over years), and continuing with extensive involvement of existing faculty in individual meetings and in the candidate’s presentation. One of us acts as host for the visit and will have to summarize the reports submitted by everyone. It is not unusual for a candidate to meet 15 of us (that is 30% of the faculty) for an hour each. So we invite carefully, and then we interview and debate even more carefully.  Last, but definitely not least, IMD features a large number of ‘‘culturenurturing mechanisms’’: Several parties, where we know the staff will count how many faculty members attend and hence we make sure we participate occasionally; an annual faculty and senior staff dinner; a staff induction program that includes a presentation on the institute’s business model by

JEAN-FRANC - OIS MANZONI

34

the President – and which new faculty are also encouraged to attend; and of course long-serving colleagues, who have made IMD what it is, are still very active and involved in the Institute’s activities (and sit at the top of the faculty pay range). They make sure that through various conversations, newcomers understand the organization’s history and values.

Other Examples My objective above is not to portray IMD as an ideal environment. First, because our system is not perfect. It is very effective on many fronts, slightly less on other fronts (that we have identified and are currently discussing). Second, because the model summarized above and discussed more extensively in Lorange (2002, 2008) is appropriate for an organization like ours (a small, independent organization, with a distinct positioning, mission and strategy), but clearly not for other organizations pursuing a different objective function. What I hope I illustrated is the way this particular organization is using a multitude of levers coherently and rather forcefully to shape employee behavior. These levers include short-term compensation, but they also include many other facets. Readers working in business schools will be able to assess to what extent their own organization is deploying such a convergent set of forces to shape faculty and staff behavior. I could have used other examples as well. During my intervention at the conference, for example, I also referred to Haier, a Chinese group that has developed a very driven, performanceoriented culture (and has exceptional achievements to show for it). This performance-oriented culture rests on an intense use of information (particularly accounting information), which also strongly influences financial rewards and one’s public standing in the organization. (e.g., yesterday’s lowest performing worker is encouraged to stand in footsteps painted on the floor and explain publicly how s/he will do better today. Similarly, managers’ monthly rankings are listed on a billboard posted at the cafeteria’s entrance. See Lin (2005a, 2005b, 2006) for a description of some of the company’s performance measurement, evaluation and reward practices.) But employee behavior is also shaped by many other levers such as those illustrated above. (See Yi & Ye, 2003; Pucik & Xin, 2004 for excellent descriptions.) The Haier system would certainly not work with all organizations or employees. Haier hence consciously selects employees that are more likely to thrive in their system. The key point is that

On the Folly of Hoping for A, Simply because You are Trying to Pay for A

35

Haier consciously and relentlessly shapes its employees’ behavior and, by doing so over a number years, has developed a strong culture that now also contributes to nurturing the system. I could also have talked about Egon Zehnder International, one of the leading firms in the executive search industry. In this traditionally very individualistic sector, Egon Zehnder stands apart with a compensation system driven by seniority and company-wide performance. (See Zehnder, 2001, or Lowe, 2004 for an excellent description of the firm’s approach.) The firm’s founder believes that far from being a handicap, this quasi profit sharing compensation system is an important part of the firm’s success, as a major component of a complex, integrated system of behavioral levers directed at a workforce recruited in no small measure based on its cultural compatibility. With a different vocabulary and highlighting slightly different behavioral levers, the point above is reminiscent of Nadler and Tushman’s (1980, 1998) congruence theory. It is certainly not a revolutionary thought in the management literature, but somehow it seems to have been receiving a decreasing amount of time and attention in much of the accounting and control literature.

CONCLUDING THOUGHTS The accounting and control research community spends much time analyzing the use of quantitative measures of performance and the way they should be linked to incentive compensation systems to drive the desired behavior. Clearly, this is an important research area, but I wonder whether we are not devoting too high a proportion of our time to what is only a small component of the reward system. There is no doubt that linking significant financial rewards to the achievement of specific quantitative yardsticks works! If you offer enough money, you will end up getting more of the activity/output you are paying for. The problem is that only in very rare circumstances are we able to pay people for exactly what we want them to do. Measurement is an imperfect process. Sometimes we use measures that are incomplete measures of the individual’s impact on the organization. Paying based on an incomplete measure will typically yield more of the desired outcomes, but also more of other outcomes that may not be desired. More complete measures of performance will typically be partly uncontrollable by individuals, which

36

JEAN-FRANC - OIS MANZONI

leads these individuals to bear risk and decreases the tightness of the link between efforts and rewards. (See Epstein & Manzoni, 1998 for a discussion of the completeness–controllability issue). One alternative is of course to use a Balanced Scorecard type of approach, linking rewards to a panel of indicators, but this process has drawbacks of its own: It dilutes the motivational power of each dimension and it requires setting weights that may not correspond to the allocation of effort and resources needed by the organization. In practice, it seems to me that (most) managers and organizations have been more careful and, dare I say, wiser. Beyond short-term incentive plans, they manage a complex set of rewards. They tend to do so in a way that is informed by quantitative performance measures, but also contains much human judgment. This judgment can be called ‘‘subjective’’ when we do not like it, or ‘‘expert system’’ if we think it can capture complex phenomena and relationships. Smart organizations also use all the levers at their disposal to shape employee behavior and, ultimately, the culture of the organization. So a first concluding thought would be to encourage accounting and control researchers to broaden their examination of reward systems beyond short-term incentive plans, and to broaden their examination of behavioral levers beyond reward systems. Merchant, Van der Stede, and Zheng (2003) made a similar plea for a broader and more cross-disciplinary approach to the study of organizational incentives which, they asserted, is a broader subject that many studies narrowly focused on incentive contracting. I support their plea and extend it to a broader focus on shaping employee behavior. Organizational incentive systems (OIS) are an important behavioral lever but they tend to be part of a broader influence system, which includes several other powerful levers. It seems to me that to understand the functioning and impact of one component in a complex system, one needs to have a broad understanding of the system and its various components, or else to approach the problem jointly with people who do (have such an understanding). Going one step further, however, I have started actively to wonder to what extent the intense focus on incentive alignment may possess perverse self-fulfilling properties. That is, are we designing increasingly complex systems to prevent people from cheating the systems y that are leading them to cheat in the first place? This self-fulfilling process might be caused by two phenomena: The imperfection of the measurement process, and motivational considerations. Let me examine these two dimensions in turn. Going back to Kerr’s (1975) examples, in most cases the problem that Kerr was encouraging the organization to solve through better measurement

On the Folly of Hoping for A, Simply because You are Trying to Pay for A

37

and reward was caused, in the first place, by the organization’s attempts to encourage specific outcomes by tying them to various rewards. Kerr’s subsequent encouragement actively to reward the desired behaviors assumes that it is possible to define better measures to capture performance. But in reality, we can rarely find the ‘‘right measures’’! Organizations then keep changing metrics to correct for side effects of the previous metrics; they try to shift behavior a bit one way, only to find it needs to be pulled back in another direction y Could it be that in many cases, there will not be a perfect measurement solution and we will be condemned to keep changing metrics to correct side effects that we created in the first place? The second reason why an intense incentive alignment focus may be selffulfilling is motivational. Ghoshal (2005, p. 77) presented the argument as follows: ‘‘A particular ideology (essentially grounded in a set of pessimistic assumptions about both individuals and institutions) has increasingly penetrated most of the disciplines in which management theories are rooted’’. (y) A theory that assumes that people can behave opportunistically and draws its conclusions for managing people based on that assumption can induce managerial actions that are likely to enhance opportunistic behavior among people’’. Ghoshal did not really explain how this process would develop. Others have proposed a number of related paths. Adult human beings have all learned how to be selfish and narrowly (short-term) self-interested. Most of us can produce this kind of behavior if we want to. Most of us have also learned to broaden our focus and to behave collaboratively and less selfishly, especially for a group we care about. An intense focus on individual short-term rewards may actually ‘‘activate/reinforce’’ employees’ selfish side, thus increasing the need to guide and constrain employee behavior through carefully designed incentives that people will take into account in order to increase the rewards they draw from them. Ferraro, Pfeffer, and Sutton (2005) support this view: ‘‘A growing body of evidence suggests that self-interested behavior is learned behavior’’ (p. 14). ‘‘There are feedback processes that cause an emphasis on pay and extrinsic incentives to create attitudes and behavior that make emphasizing pay essential for motivating and directing behavior. That is because emphasizing pay actually makes pay more important to employees (y) creating a cycle of behavior that makes the use of incentives, once begun, more and more necessary to continue to motivate and direct behavior’’ (p. 20). Underlying this argument is the hypothesis that individuals who choose their behavior in order to obtain behavior- or outcome-dependent rewards

38

JEAN-FRANC - OIS MANZONI

will continue to do so, and in fact will increasingly do so. This hypothesis has been the object of considerable research over the last few years. In psychology, the propensity of extrinsic rewards to displace intrinsic motivation, often called the ‘‘over-justification effect’’, is a core prediction of Deci and Ryan’s Cognitive Evaluation Theory, a component of their broader ‘‘Self-Determination Theory’’ (see, e.g., Deci & Ryan, 1985 or Ryan & Deci, 2000). Unfortunately, the evidence is not perfectly conclusive. Several meta-analyses and literature reviews report support for the hypothesis (e.g., Tang & Hall, 1995 or Deci, Koestner, & Ryan, 1999), but others disagree (e.g., Eisenberger & Cameron, 1996). In economics, this effect has also been researched extensively, typically under the name ‘‘motivation crowding theory’’ (Frey, 1997). Literature reviews by Frey and Jegen (2001) and Lindenberg (2001), for example, report frequent support for the hypothesis. More recently, Weibel, Rost, and Osterloh’s (2007) meta-analysis and vignette study finds that while performance-contingent pay strengthens extrinsic motivation, it also tends to decrease intrinsic motivation. More generally, they argue, pay-forperformance ‘‘always produces hidden costs of rewards’’. Interestingly, and in contrast with the lack of cross fertilization and mono-discipline focus observed in the accounting literature by Merchant et al. (2003), these two lines of research cross-reference one another extensively. Some individuals even publish in both literatures, as for example, Mulder, van Dijk, Wilke, and De Cremer (2005), Mulder, van Dijk, and De Cremer (2006a), and Mulder, van Dijk, De Cremer, and Wilke (2006b) fascinating studies on sanctioning systems (showing that the presence or introduction of sanctioning systems can lead to a decrease of cooperation, in part because they decrease individuals’ confidence that others will behave cooperatively). Intuitively, I think we will find over time that the over-justification effect tends to occur more frequently and more intensely under some circumstances. Some evidence already points in that direction (see Gagne´ & Deci’s, 2005, review). I am struck, in particular, by Forehand’s (2000) fascinating study examining the reaction of customers to marketers’ promotion efforts. The motivational impact of promotional rewards turned out to depend on the consumer’s attribution of the marketer’s objective: When the marketer was perceived to be ‘‘promotion-focused’’ (i.e., to care mostly about maximizing sales), customers reported lower purchasing intention than when they perceived the marketer to be ‘‘reward-focused’’. (This is exactly what Tesco’s Vice-President Tim Mason said, when he argued that the first retailer to introduce a consumer-friendly innovation

On the Folly of Hoping for A, Simply because You are Trying to Pay for A

39

(such as Tesco’s ‘‘loyalty card’’) has a much greater likelihood of being perceived by consumers as ‘‘doing something for them’’, while followers/ imitators’ efforts are more likely to be attributed to a desire to reduce the first mover’s advantage.) James (2005) examined the issue using analytical models, and he too proposed a contingent outcome: Motivation crowding out will occur when rewards are perceived as controlling, which is more likely to occur when the object of the agent’s intrinsic motivation is the source of the agent’s extrinsic compensation and when the incentives offered to the agent are too large. It is important for the accounting and control community to become more cognizant of these exciting developments in other fields, including those cited in this chapter, i.e., psychology, economics and marketing. That would be a first step to taking a more active and effective part in the development of the world’s knowledge on these questions.

REFERENCES Chapman, C. (2007). Performance measures, pay and accountability: A field study in an audit firm. Paper presented at the 4th Conference on Performance Measurement and Management Control, Nice (France), September 26–28. Cropanzano, R., Bowen, D. E., & Gilliland, S. W. (2007). The management of organizational justice. Academy of Management Perspectives, 21, 34–48. Deci, E. L., Koestner, R., & Ryan, R. M. (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125(3), 627–668. Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human behavior. New York: Plenum. Eccles, R. G., & Crane, D. B. (1988). Doing deals: Investment banks at work. Boston, MA: Harvard Business School Press. Eisenberger, R., & Cameron, J. (1996). Detrimental effects of reward: Reality of myth? American Psychologist, 51, 1153–1166. Epstein, M., & Manzoni, J.-F. (1998). Implementing corporate strategy: From tableaux de bord to balanced scorecards. European Management Journal, 16(2), 190–204. Ferraro, F., Pfeffer, J., & Sutton, R. I. (2005). Economics language and assumptions: How theories can become self-fulfilling. Academy of Management Review, 30(1), 8–24. Forehand, M. R. (2000). Extending overjustification: The effect of perceived reward-giver intention on response to rewards. Journal of Applied Psychology, 85(6), 919–931. Frey, B. S. (1997). Not just for the money: An economic theory of personal motivation. Broookfield: Edward Elgar Publishing. Frey, B. S., & Jegen, R. (2001). Motivation crowding theory. Journal of Economic Surveys, 15(5), 589–611. Gagne´, M., & Deci, E. L. (2005). Self-determination theory and work motivation. Journal of Organizational Behavior, 26, 331–362.

40

JEAN-FRANC - OIS MANZONI

Galbraith, J. R. (1995). Designing organizations. San Francisco, CA: Jossey-Bass. Ghoshal, S. (2005). Bad management theories are destroying good management practices. Academy of Management Learning and Education, 4(1), 75–91. James, H. S., Jr. (2005). Why did you do that? An economic examination of the effect of extrinsic compensation on intrinsic motivation and performance. Journal of Economic Psychology, 26, 549–566. Kerr, S. (1975). On the folly of rewarding A, while hoping for B. Academy of Management Journal, 18(4), 769–783. Kerr, S. (1995). On the folly of rewarding A, while hoping for B. Academy of Management Executive, 9(1), 7–14. Liden, R. C., Wayne, S. J., & Sparrowe, R. T. (2000). An examination of the mediating role of psychological empowerment on the relations between the job, interpersonal relationships, and work outcomes. Journal of Applied Psychology, 85(3), 407–416. Lin, T. W. (2005a). Effective OEC management control at China Haier group. Strategic Finance, 86(11), 39–45. Lin, T. W. (2005b). OEC management control system helps China Haier group achieve competitive advantage. Management Accounting Quarterly, 6(3), 6–11. Lin, T. W. (2006). Lessons from China. Strategic Finance, 88(4), 48–55. Lindenberg, S. (2001). Intrinsic motivation in a new light. KYKLOS, 54, 317–342. Lorange, P. (2002). New vision for management education: Leadership challenges. Oxford: Pergamon Press. Lorange, P. (2008). Thought leadership meets business: How business schools can become more successful. Cambridge: Cambridge University Press. Lowe, S. C. (2004). Marketplace masters: How professional service firms compete to win. Westport, CT: Praeger Publishers. MacKenzie, S. B., Podsakoff, P. M., & Fetter, R. (1991). Organizational citizenship behavior and objective productivity as determinants of managerial evaluations of salespersons’ performance. Organizational Behavior and Human Decision Processes, 50(1), 123–150. MacKenzie, S. B., Podsakoff, P. M., & Fetter, R. (1993). The impact of organizational citizenship behavior on evaluations of salesperson performance. Journal of Marketing, 57(1), 70–80. MacKenzie, S. B., Podsakoff, P. M., & Paine, J. B. (1999). Do citizens behaviors matter more for managers than for salespeople? Journal of the Academy of Marketing Science, 27(4), 396–410. Manzoni, J.-F. (1993). Use of quantitative feedback by superiors: Causes and consequences. Unpublished Doctoral Dissertation. Harvard University Graduate School of Business Administration, Boston. Merchant, K. A., & Manzoni, J.-F. (1989). The achievability of budget targets in profit centers: A field study. The Accounting Review, 44(3), 539–558. Merchant, K. A., Van der Stede, W., & Zheng, L. (2003). Disciplinary constraints on the advancement of knowledge: The case of organizational incentive systems. Accounting, Organizations and Society, 28, 251–286. Mulder, L. B., van Dijk, E., & De Cremer, D. (2006a). Fighting noncooperative behavior in organizations: The dark side of sanctions. Research on Managing Groups and Teams, 8, 59–81. Mulder, L. B., van Dijk, E., De Cremer, D., & Wilke, H. A. M. (2006b). Undermining trust and cooperation: The paradox of sanctioning systems in social dilemmas. Journal of Experimental Social Psychology, 42(2), 147–162.

On the Folly of Hoping for A, Simply because You are Trying to Pay for A

41

Mulder, L. B., van Dijk, E., Wilke, H. A. M., & De Cremer, D. (2005). The effect of feedback on support for a sanctioning system in social dilemmas: The difference between installing and maintaining the sanction. Journal of Economic Psychology, 26(3), 443–458. Nadler, D., & Tushman, M. (1980). A model for diagnosing organizational behavior. Organizational Dynamics, 9(2), 35–51. Nadler, D., & Tushman, M. (1998). Competing by design: The power of organizational architectures. NY: Oxford University Press. Organ, D. W. (1988). Organizational citizenship behavior: The good soldier syndrome. Lexington, MA: Lexington Books. Organ, D. W., Podsakoff, P. M., & MacKenzie, S. B. (2006). Organizational citizenship behavior: Its nature, antecedents, and consequences. Thousand Oaks, CA: Sage. Pascale, R., & Athos, A. (1981). The art of Japanese management. London: Penguin Books. Podsakoff, P. M., MacKenzie, S. B., Paine, J. B., & Bachrach, D. G. (2000). Organizational citizenship behaviors: A critical review of the theoretical and empirical literature and suggestions for future research. Journal of Management, 26(3), 513–563. Pucik, V., & Xin, K. (2004). Managing performance at Haier (A). Case study, IMD, Lausanne, IMD-3-1332. Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55, 68–78. Scandura, T. A., Graen, G. B., & Novak, M. A. (1986). When managers decide not to decide autocratically: An investigation of leader-member exchange and decision influence. Journal of Applied Psychology, 71, 579–584. Schein, E. (1992). Organizational culture and leadership. San Francisco, CA: Jossey-Bass. Tang, S-H., & Hall, V. C. (1995). The overjustification effect: A meta-analysis. Applied Cognitive Psychology, 9, 365–404. Weibel, A., Rost, K., & Osterloh, M. (2007). Crowding-out of intrinsic motivation – Opening the black-box. Working paper available at SSRN – http://ssrn/com, ID 957770. Yi, J. J., & Ye, S. X. (2003). The Haier way: The making of a Chinese business leader and a global brand. Dumont, NJ: Homa & Sekey Books. Zehnder, E. (2001). A simpler way to pay. Harvard Business Review, 79(4), 53–61.

This page intentionally left blank

PERFORMANCE MEASUREMENT AND MANAGEMENT CONTROL SYSTEMS: CURRENT RESEARCH AND IDEAS GOING FORWARD Antonio Davila ABSTRACT Based on the more than 120 papers presented at the fourth conference on Performance Measurement and Control, this paper examines the current state of research in this field. It examines the diversity in research settings, theoretical backgrounds, research designs, and topics covered. The picture that emerges is that of a dynamic field where different ideas and perspectives converge to create a rich and interesting environment. The papers show the progress that this field has made both in terms of the quality of the research as well as the attractiveness of the research questions being addressed. The paper concludes with some thoughts about how to improve even more the quality going forward and an optimistic assessment of the future of the field.

A biannual research conference such as the fourth Conference on Performance Measurement and Management Control is a unique opportunity to see the trends and state of the art in a particular research field. Performance Measurement and Management Control: Measuring and Rewarding Performance Studies in Managerial and Financial Accounting, Volume 18, 43–69 Copyright r 2008 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-3512/doi:10.1016/S1479-3512(08)18003-0

43

44

ANTONIO DAVILA

It gives a clear picture of what researchers are interested in, what topics are being explored, what research methods are being used, and what background theories are sustaining the empirical work. It also provides an exceptional set up to compare more than a hundred papers and to draw some conclusions about why some of them are more convincing than others. These conclusions are not about research methodology – numerous papers and books address the issue of how to do design a reliable and valid research project for any methodological approach that you choose; they are about basic ideas that can be easily addressed to improve a paper significantly. This piece is about these two topics. The first part describes the state of the art in performance measurement and management control. It provides descriptive statistics on the research settings that are attracting attention, the topics that are being researched, the methods that are being used, and the theories that support empirical studies. Its aim is to give map of the current research field. In contrast to a review paper, this paper does not focus on a particular topic. It is not biased by publication – which reflects the preferences of the editorial team towards certain topics and certain research designs. Rather, it lets the field speak up through a conference that is open to diversity in terms of topics and methods but most importantly in terms of researchers. The second part of the paper offers basic suggestions on how to make research papers better. After reading the papers in the conference, certain issues emerge as to why certain papers are more convincing than others. This second part of the paper is about these issues – simple changes that can drastically change the perception about a paper.

1. DESCRIPTIVE STATISTICS ON STATE-OF-THE-ART RESEARCH IN PERFORMANCE MEASUREMENT AND MANAGEMENT CONTROL 1.1. Where is Research Happening? An overwhelming majority of the papers are empirical, reflecting the applied nature of performance measurement and management control. Our aim as researchers is to understand practice, identify the challenges that managers are facing to focus our efforts on addressing them, uncover best practices that we can analyze and disseminate through our teaching, and advance knowledge and its applications.

Performance Measurement and Management Control Systems

45

Large companies typically face more complex environments. Coordination problems associated with managing thousands of people often require novel approaches. Global operations quickly bring to the forefront crosscultural challenges especially in a field such as ours. Cross-functional coordination is more demanding in large companies. Learning processes, where management control systems have a key role, are also more complex. Moreover, managers in these companies face these challenges on a daily basis and have thought about how to solve them; these managers can be interesting research partners. It is not surprising that 62 papers in the conference (out of 120) focus on these organizations (Fig. 1) (Lucianetti, 2007; Matejka & Maas, 2007; Anderson, Davis, & Widener, 2007; AbdelMaksoud & Pollanen, 2007; Silvi, Bartolini, Raffoni, & Visani, 2007). These companies have more sophisticated performance measurement and control systems and thus offer very rich research settings. Moreover, the unit of analysis can move from the company level to business units and functions. These different alternatives to execute a project also expand the opportunities for research. Public data on these companies enhances information available to have richer studies and triangulate data. These characteristics can be both a blessing and a threat. First, the advancement of the field requires studying these sophisticated systems; but existing concepts might be too crude to capture and communicate the complexity of these new systems. Second, going down the hierarchy of a company increases the 70 60

Number of papers

50 40 30 20 10 0

Large companies

Fig. 1.

Public sector

Health care

Where is Research Happening?

Small and medium companies

46

ANTONIO DAVILA

degrees of freedom, reduce noise in the research design and thus lead to sharper hypothesis testing; however if the researcher is not careful enough, correlated error terms and correlated omitted variables can question the validity of the findings. The second largest set of papers – 22 papers – focuses on the public sector (Chiaravalloti & Meer-Kooistra, 2007; Al Sharif, 2007; Amenta, 2007; Verbeeten, 2007; Mucciarone & Neilson, 2007). While this sector has traditionally been of interest, the current amount of work is definitely much higher than in the past. This significant effort to understand performance measurement and management control systems in government organizations probably reflects the social pressure to have more effective and efficient public sector. New public sector theory is built on the assumption that management techniques that are used in the private sector can be replicated in the public sector to enhance its performance. Performance measurement and control systems are among the most prevalent techniques in companies and understanding how they are adapted and how they are used in the public sector is of outmost importance to document the rebirth of government organizations. Six out of these twenty-two papers study universities probably reflecting that access to our own institutions may be easier, but also indicating the need to improve our own coordination and motivation systems (Broad & Jadsri, 2007; Mpabanga, 2007; Agasisti & Bianco, 2007; Ja¨a¨skela¨inen, Lo¨nnqvist, Laaksonen, & Kulmala, 2007; Macdonald & Kam, 2007). From the perspective of our field, these organizations present several unique challenges. First, their objectives are much fuzzier than those of forprofit companies. While we may argue about the objectives of companies (even more today with the increasing relevance of corporate social responsibility), economic value creation will always play a dominant role. In contrast, government organizations have a much broader role around social service, with political forces acting on top of it and power spread around numerous actors with frames of mind very different from what we are used to in companies. In these settings, the traditional goal divergence problem emerges in awkward ways. Second, the complexity of the pubic sector suggests that simply translating management tools that work in companies will not address these challenges. These tools need to be modified and adapted; trial and error will be at the core of this transformation process but also development of sound theory. Researchers need to be there to document what works and what does not; but most importantly to ground these observations in theory to understand the why behind them. Nine studies chose the healthcare sector for their research (Franchino, Laura, & Paolo, 2007; Baraldi, 2007; Kastberg, 2007; Turolla, Vola,

Performance Measurement and Management Control Systems

47

Carenzo, & Franchino, 2007). While often part of the public sector, healthcare presents its own challenges. It is one of the most important sectors in the more advanced economies, accounting for more than 10% of the GDP. Its sheer weight in the economy deserves special attention. The tension between different objectives and different actors in the industry requires its own performance measurement and control systems. The nature of the decisions that are taken on a daily basis provide a different dimension to the concept (and the measurement) of efficiency. Biotechnology is a piece of the healthcare sector that is increasing in its relevance. This industry works with extremely long development cycles where performance is hard to measure and learning, coordination, and control cannot be achieved with the traditional tools (Ishak & Lee, 2007). Going forward, biotechnology should get more research attention because of its growing importance and the challenges that it presents to our knowledge. Eight papers focus on the issues of small and medium companies (Reheul & Jorissen, 2007; Simpson, Padmore, & Frecknall-Hughes, 2007; Misiaszek, Oriot, & Otley, 2007). Granted that larger companies present more complex measurement and control problems, still small companies are a large part of our economies and our audience in teaching programs. Moreover, the more aggressive segments of these companies are a fascinating population. For instance, back in 1984, SMEs spent about 6% of R&D expenses in the US; this number was up to about 14% in 1994 and again up to 21% by 2003. These companies are taking on more complex tasks with innovative management solutions that we should be looking into. To conclude this section, I would like to point out to what I felt was missing. There were no studies centered on not-for-profit organizations. These organizations such as micro-credit banks, foundations, or social entrepreneurial startups are increasing in their importance – not only in terms of size but most importantly in terms of their relevance to society. Another part of the economy that was underrepresented is the service sector such as audit firms (because of our affiliation with the accounting profession I was expecting more interaction with them), the tourist sector, or banking. The weight of these sectors in the economy is increasing but our heritage in manufacturing still drives a lot of our research.

1.2. What Topics are Being Researched? The central role of performance measurement to organizations and to our research community is reflected in the attention it receives (Fig. 2)

48

ANTONIO DAVILA 35 30

Number of papers

25 20 15 10 5

Fig. 2.

Tr pr ans ic fe in r g

P po ubl lic ic ie s

nc er ov G

st Co

na

ste sy

ste sy IT

e

s m

m s

s et dg Bu

B Sc ala or nc ec ed a Pe rd r ev fo a rm re luat an w io ce a n Co rds / m pe ns at io n

m

Pe

r ea form su a re nc m e en t M an co age nt me ro n l t

0

What Topics are Being Researched?

(Kruis & Spekle´, 2007; Szanto & Wimmer, 2007; Chiesa, Frattini, Lazzarotti, & Manzini, 2007; Claes, 2007; Pesalj, 2007). Current research in this area is moving towards tackling harder performance measurement issues such as measuring intangible assets, perceptions, taste, human and intellectual capital (Garelli & Longobardi, 2007; Chiucchi, 2007), ethical behavior, corporate social responsibility (Fiori, Donato, & Izzo, 2007), environmental impact (Delai & Takahashi, 2007; Ja¨rvenpa¨a¨ & La¨nsiluoto, 2007), entrepreneurship (Bratnicki & Dyduch, 2007), and innovation (Yamane & Asada, 2007). Managers need guidance in answering these questions and our research tradition gives us the knowledge to add to the discussion. Another trend is to go deeper into organizations and examine measurement challenges in functions such as marketing, new product development, or human resources. New industries such as biotechnology or web-based business models are also attracting attention. Traditional issues in performance measurement such as the shortcomings of financial measures, measuring efficiency, understanding how measures are being used, or moving measures throughout the organizational structure are still in the research agenda. Another large topic is management control systems which is the typical dependent variable in contingency studies (Ding, Dekker, & Groot, 2007;

Performance Measurement and Management Control Systems

49

Campbell, Datar, & Sandino, 2007). Research is moving towards the top management team to better understand the role that these systems have in formulating and implementing strategy – what some researchers identify as strategic performance management systems. Moving up the hierarchy leads to more challenging research questions because various forces come into play interacting with one another. Disentangling these forces requires smart and innovative research designs. This increased difficulty reflects the importance of the research questions addressed. This trend also enhances the interaction with the strategy research field. A few years ago the intersection between management control and strategy had few studies; this intersection now includes numerous papers. The dynamism of the strategy field is also a good source of stimuli to bring new ideas and enrich our own field. Research is also moving into the interplay between formal control systems and their impact on organizational behavior (Da Silva & Rojas Lezana, 2007; Hartmann & Slapnicar, 2007; Kalagnanam, Kobussen, & Vaidyanathan, 2007). Rather than explaining cross-sectional differences in design, current research is going a step forward in understanding how particular designs change the social dynamics of organizations. The research designs and concepts in the papers that address these issues are also innovative. Again, the interaction with organizational theory – also a quickly evolving field – promises to bring new and stimulating ideas in how we understand our own research questions. New research topics that are emerging include inter-organizational control systems and the reasons-foradoption of management systems. The Balanced Scorecard also had a prominent role at the conference (Frezatti, Relvas, & Junqueira, 2007; Cugini & Michelon, 2007; Yongvanich & Guthrie, 2007; Vola, Broccardo, & Truant, 2007; Shulver, Lawrie, Barney, & Kalff, 2007; Boulianne, 2007). It was developed 15 years ago and it is now reaching the population of organizations at large. Research on this topic focuses mostly around implementation cases where researchers describe the process of adopting the system. The number of welldocumented examples is large enough that a meta-analysis of the existing evidence is needed to summarize this accumulated knowledge. Performance measurement frameworks such as the Balanced Scorecard have not seen a significant conceptual innovation for a few years. However, leading companies are moving beyond a focus on internal processes and their impact on the market – the view that characterizes existing frameworks – to use performance measurement systems to look around their ecosystem. Other topics that are part of the management control concept but attract enough attention to treat them separately is performance evaluation,

50

ANTONIO DAVILA

rewards, and compensation (Pekkola, Ukko, & Rantanen, 2007; De Waal & Roobol, 2007; Johansson & Siverbo, 2007; Barontini & Bozzi, 2007; Likierman, 2007). Performance measures play a significant role in evaluation and rewards and research in this area is expanding to include the interplay of these systems with sociological forces within organizations such as gamesmanship, autonomy, and psychological aspects such as perception of justice, role conflict, and role ambiguity (Ukko, Karhu, & Pekkola, 2007). These are important issues that these papers address in a very interesting way. These new concepts bring new perspectives to one of the most important aspects of management. They also show that those researchers working in performance measurement and control need to interact and understand very well other fields such as human resources in this particular topic. Compensation studies focus on the design of economic incentives and often address the question from an economic perspective. The progress in this area is also clear from reading the conference papers. Information systems (Enterprise Resource Planning – ERP – or Customer Relationship Management – CRM) are becoming central to performance measurement and control systems. Several papers examine this intersection and offer interesting insights (Folgueras, Crespo, Berbis, & Mezcua, 2007; Wall, 2007; Gabrie¨ls & Jorissen, 2007; Argyropoulou, Ioannou, & Koufopoulos, 2007). However, the centrality of information systems to our research field suggests that more resources should be devoted to the topic. Again the interaction with other fields, information technology in this case, is critical. Researchers in information technology have centered part of their efforts on ERP, CRM, and similar business software, but they lack the perspective that management control systems brings to better understand this important phenomenon. Similarly, governance – also at the core of the control function – is beginning to receive attention (Speckbacher & Wentges, 2007). Other fields such as finance have looked into it in detail already, but they lack the perspective that the control literature offers. I hope that over time, this relevant topic will increase in importance. Traditional topics that are still in the research agenda but are not as pervasive as they use to be are: internal control (Chen & Sandino, 2007), cost management (Suomola, Varila, & Jokioinen, 2007), budgets (Karbhari & Mohamad, 2007; Arena, Arnaboldi, & Azzone, 2007), and transfer pricing (Fischbacher, Stefani, & Pfaff, 2007). These topics have been central to our field for several decades and while all the open issues have not been closed, the diversification into other topics is a healthy move.

51

Performance Measurement and Management Control Systems

1.3. What Methods are Being Used? Fig. 3 provides descriptive statistics on the methods used. The most common research method is case studies; which as been the most popular method for some time (Parisi & Sansalvadore, 2007; Fujino, 2007; Micheli & Kennerley, 2007; Agyemang & Ryan, 2007; O¨zer & Kocakoc- , 2007; Fazzini & Terzani, 2007). Case studies are an attractive approach to get into the detail of performance measurement and control systems. Advances in the practice of management often start in certain companies; it is our role as researchers to identify these innovators to understand what they are doing, how they are going about it, and why do their approaches work or fail. Therefore, an important part of case studies is to carefully identify organization(s) unique in some distinct and relevant aspect; case research is time consuming so purposefully choosing outliers often provides much more insight than picking up an average organization. The significant advantage of case research is the possibility of collecting very rich data to fully understand the phenomenon. It means that case research is very time consuming; relevant data comes from analyzing external and internal documents and quantitative data, from a large enough number of interviews

45 40

30 25 20 15 10 5

Fig. 3.

What Methods are Being Used?

Si m ul at io n

A na ly tic al m od el s

Ex pe rim en ts

Pr es cr ip tiv e

da ta Fi el d

op (In in fo io rm n/ e Th d) eo ry

ey rv Su

se

stu

di

es

s

0

Ca

Number of papers

35

52

ANTONIO DAVILA

with different informants, from observation such as meetings, and sometimes even from questionnaires. Rich data allow triangulation through multiple informants and qualitative and quantitative evidence to enhance the validity of the case findings. Longitudinal case studies where the researcher interacts with the organization over a long period of time often add an important evolutionary dimension. These cases document not only a snapshot at a point in time, but changes over time that are critical to fully understand these systems. Similarly, going from one case to three or four cases allows for interesting comparisons. Another way of increasing the credibility of case research is to contrast the evidence with different theories to show how the data supports or contradicts theoretical predictions. The link to theory and the literature is critical to a good case study. A difficult balance in case research is between the need for detailed descriptions and crispy learning points. Without convincing descriptions the reader may remind skeptical about the validity of the conclusions, but too much description may obscure critical findings. Some papers at the conference were based on action research where the researchers were engaged as actors in the process (Rejc Buhovac, Savicˇ, & Kecˇanovic´, 2007; Bonacchi, Ferrari, & Pellegrini, 2007). These papers offer an interesting perspective – as long as they are written as objectively as possible without emphasizing the excellent job done by the researcher in her or his role as participant in the process. They provide a point of view that is often lacking in other approaches to case research. However, action research has traditionally faced significant challenges in terms of being published in research journals (these papers often appear in edited books or practitioners’ magazines) because validity concerns are difficult to overcome. Surveys are the second most popular research design (Bodnar, Danko, Kiss, & Vas, 2007; Kudo, 2007; Forslund & Jonsson, 2007; Van Der Merwe & Visser, 2007; Constanti, Krambia-Kapardis, & Constanti, 2007). They allow testing specific hypotheses using statistical tools. This research is not about outliers doing something unique but rather about how the average population behaves. However, survey research commonly includes performance as a dependent variable, a tradition that addresses managers’ perennial concern which is whether whatever is being tested works. Because of validity concerns, researchers often rely on existing variables for which there are valid measurement instruments. Therefore contributions to the literature are incremental and related to testing new relationships or examining a new variable with its associated instrument. The statistical approach to hypothesis testing raises issues such as the validity of the model. For instance, the large number of contingency variables that have been

Performance Measurement and Management Control Systems

53

identified in the budgeting literature requires new research to consider these variables if they want to avoid omitted correlated variables or mis-specified models. Other relevant issues in statistical analysis are endogeneity and independence of observations. Another set of papers aimed at developing new ideas, frameworks, and theories without empirical data (Megali & Becatti; Zraly, 2007; Decramer, Christiaens, & Vanderstraeten, 2007; Mueller, 2007; Soulerot & Farjaudon, 2007). There is no easy research project and theory work is no exception. However, excellent theory papers are a pleasure to read. Good theory papers have to be well grounded in the literature – both theoretical and empirical. They also need to be well-argued. It is not enough to provide a line of reasoning but also to explain why alternative explanations do not work. Theory papers are scarcer that they should be. Given the significant amount of empirical evidence that the field generates, it is important to capture all this dispersed learning not only in periodic literature review papers but in frameworks that organize this learning in an effective way. Archival field data is yet another research design common in performance measurement and management control (Schiehll, 2007; Bhattacharya, Black, Black, & Christensen, 2007; Needles, Powers, Shigaev, & Frigo, 2007b; Schryver, Eisinga, Teelken, & Poutsma, 2007; Ulucan, 2007). For instance, compensation studies often rely on public sources of data such as company filings. Public data sources are available to various research teams and have often been examined with significant detail. Progress using these data requires creative hypotheses that are well-argued and grounded on the literature. Companies are also rich sources of archival data – for instance, a company may have an interesting data set on how targets for salespeople evolve over time. Access to good quality private databases provides a competitive advantage to researchers; their challenge becomes how to link this advantage to theory in order to identify significant contributions. Prescriptive studies build on existing theory to tell organizations how to manage themselves (Garengo, Biazzo, & Bernardi, 2007; De Souza Lenz, Hansen, Fedrizzi, & Roldan, 2007; Michel & Michel, 2007). The audience for these studies is often managers who need to go back to their work with new ideas. Nevertheless, these studies have room in research conferences because our aim is precisely to be of relevance to managers. Indeed, research should be communicated to managers at some point in time if we want to fulfill our mission. However, the outlet for this type of work is seldom research journals; but rather books, chapters in edited books, or practitioners’ magazines.

54

ANTONIO DAVILA

Three other research methods were used by different papers. They are methods addressing research questions that are hard to answer using more common approaches but yet they are relevant to our field. These research methods include experiments (Booker & Heitger, 2007; Roberts, Cauvin, & Neumann, 2007), analytical models (Schultze & Weiler, 2007), and simulation (Kunz, 2007). Experiments have a high degree of control over the research setting and the researcher can manipulate particular variables to examine human behavior. The challenge of designing experiments is to make sure that there are no confounding variables that might be varying with the manipulated one. Analytical models have a long tradition grounded in microeconomics. Their significant advantage is that they force researchers to be sharp and unequivocal in concepts and arguments. Because they are translated into mathematical language, there is no room for misunderstanding what a particular concept means or for sloppy thinking. Their challenge is to balance the need to abstract reality enough to have clean models and yet be close enough to reality to explain and predict behavior. Simulations are also growing in importance; they are difficult to set up in a way that they are close enough to reality but they are powerful in exploring complex behavior.

1.4. What Theories are Being Used? Fig. 4 provides an overview of the various theories used in the papers. The diversity of theories shows the applied nature of performance measurement and control systems. Because of this nature, all the perspectives are needed and each brings a different view on the phenomena that we research. Twenty papers use psychology as their theoretical background (Baraldi, 2007; Linder, 2007; Arizzio, 2007; Madini, 2007; Jansen, 2007; Mishken & Juhasz, 2007). Given the impact that control systems have on people, it is reasonable that psychology drives a significant amount of research. The duality between extrinsic and intrinsic motivation has been debated for a long time and the fact that it is still at the core of current papers shows that the debate is not yet solved. A topic that has been gaining importance over the last few years is the concept of justice. Reward systems are just a way to allocate rents between shareholders and employees and justice is at the core of this process; perceptions of justice in this distributive process can have important consequences in terms of motivation and overall performance of the organization. Other topics in psychology that are also addressed are role theory, socio-cognitive models, participation (a classic variable in budget

55

Performance Measurement and Management Control Systems 30

25

20

15

10

5

Fig. 4.

Th eo ry -le ss

th eo ry G ro un de d

St ra te gy

So ci ol og y

l M an ge m th ent eo c ry on tro l Ec on om ic s/r at io na l

rg an th iza eo tio ry na

O

Ps

yc

ho

lo

gy

0

What is the Theory Background?

studies), variable risk preference model, expectancy-valence motivation (Emmanuel, Kominis, & Slapnicar, 2007), and leadership styles. This diversity in topics reflects the richness of perspectives that can be brought to study performance measurement and control systems but also the richness of our research community. Nineteen papers used organizational theory (which is often seen as part of sociology but that we treat separately here) as their background (Dekker & Van Den Abbeele, 2007; Pavlov & Bourne, 2007). Contingency theory was the most popular perspective in the conference within organizational theory papers (Nevries, Hoffjan, & Winn, 2007; Groot, Dekker, & Schoute, 2007). Its popularity reflects its long tradition in the management control literature and also its versatility to adapt to multiple research settings. Its arguments are closely related to the concept of efficiency in economics. The concept of ‘‘fit’’ argues that the best (often interpreted in some form of economic efficiency) control systems’ design varies across settings depending on contingency variables. Contingency theory still faces significant challenges going forward. For instance, the number of contingency variables identified in the literature has grown so much that it is hard to control for all the

56

ANTONIO DAVILA

potential contingencies and to know what variables have the highest impact. It appears that deeper theories are needed to explain why certain variables affect the design of control systems, rather than invoking contingency theory as the end point. As an example, the variable country is an important contingency variable in explaining differences in control system design across countries; however, country itself does not explain why these differences exist. We need to go deeper into understanding what is behind the country variable. Other organizational theories used as background in various papers include social systems theory, resource dependence, new public sector theory, power theory, network theory, stakeholder theory, information processing, human capital, and stakeholder theory. Evolutionary theory has been part of organizational theory for quite some time; however, control systems’ research has often been cross-sectional in nature. The lack of an evolutionary perspective on performance measurement and control systems is starting to be addressed as several papers in the conference prove. Traditional sociology is the background for five papers (Chapman, 2007; Frow, Ogden, & Marginson, 2007). Institutionalism is an important theory to explain the adoption of practices and therefore central to understanding how systems are adopted across companies. Adoption of performance measurement and control systems is commonly based on cost-benefit tradeoff arguments. However, these systems are embedded in social forces. Institutionalism provides the tools to examine alternative arguments. For instance adoption may be linked to functional pressure where weak performance forces the organization to look for solutions; it might be also associated to political forces where a shift in the power structure demands changes in the design of the organization; adoption may also be due to social pressure where changes in values at the social level demand changes in organizational practices – the corporate social responsibility move is a clear illustration of this point. Foucauldian analysis and constructivism are also theories that provide an interesting perspective to our understanding of performance measurement and control systems. Economics is also the background for a significant number of papers (14) (Needles, Powers, & Frigo, 2007a; Horva´th, Lattwein, & Moeller, 2007; Kandel, 2007; Busanelli De Quino, Cardoso, Boya, & Pagliarussi, 2007). The concept of efficiency is at the core of economics and also present in the minds of managers. In the particular case of performance measurement and control systems, the question is whether and how these systems help improving the efficiency of the organization without sacrificing long-term effectiveness. The difficulty in answering the cost-benefit trade-off question

Performance Measurement and Management Control Systems

57

comes from estimating costs – especially the costs of running these systems – and, even harder, estimating the benefits. Often, organizational performance is used as a proxy for these benefits. The challenge in case studies is to separate the impact of these systems from the other changes that happen at the same; in survey research the noise term captures the random variability of these other variables, but this assumption only holds if they are uncorrelated with performance measurement and control systems. While the concept of efficiency goes back to the early days of economics; more recent aspects of efficiency are also present in papers. Risk aversion and risk premium associated with agency relationships and efficient contracting are core to control systems and in particular to compensation systems where an explicit contract happens between the organization and the employee. Other aspects of economics include transaction cost economics, information economics, and causality models. Fifteen papers used what I have labeled management control theory (Andersson-Fele, 2007; Naranjo-Gil, Maas, & Hartmann, 2007; Wagner & Soljakova; Zakrzewski & Juchau, 2007). These papers are grounded on theories and frameworks that have been developed within our field. These frameworks include the traditional Thompson/Ouchi/Merchant framework on personnel, action, and results control; Simons’ levers of control; the duality of formal and informal controls; and various performance measurement and management frameworks, value based management, and measurement theory (Zurwehme, 2007). Most of these papers are empirical using these frameworks as background for the empirical work. Theory work appears to move much slower than empirical findings. The most popular frameworks the personnel–action–results and the levers of control are two (at least) and one decade old; even if there has been significant findings and changes in the world since they were developed. In contrast, fields such as strategy or organizational theory (just to pick up two close to control theory) have been fertile in terms of new concepts and ideas. One paper relied on grounded theory; a somewhat surprising observation given the need to understand new practices in the field. Given the need to frame observations of new practices and the fact that theory that serves as background to empirical studies was developed a few years ago; the field going forward need to rely on heavier use of grounded theory. The strategy literature was the background for three papers (Chenhall, Kallunki, & Silvola, 2007). Network theory was the background for one of them. This theory has proved to be an innovative and effective way to look at new organizational forms that have emerged over the last few years. Performance measurement and control systems in these networks have some

58

ANTONIO DAVILA

unique features; yet we don’t know much about how centrality or structural holes, just to cite to key aspects of network theory, interact with these systems. Another paper was grounded in entrepreneurial theory. Again, entrepreneurship is a very dynamic field both in research and practice. Just to bring a piece of evidence, governments around the world are experimenting with different solutions to create the dynamics of Silicon Valley – they see entrepreneurship as a fundamental aspect for economic growth. However, the energies devoted to this segment of the economy are lower than its importance (and newness) would require. Finally, 29 papers lacked a clear theoretical background. This fact reflects the need to work hard on linking field observations with theories. Research is a two-way street with empirical work and theory development evolving together. The absence of theory on a significant number of papers identifies the need to keep on working on building this relationship between the two faces of research.

2. SOME THOUGHTS ON HOW TO MAKE OUR RESEARCH EVEN BETTER Table 1 outlines 10 ideas that are not costly to implement in a research project but have a significant impact on the quality of a paper. These ideas came up as I read the papers in the conference and doing the mental exercise of teasing the reasons why some sounded more convincing while others left the impression of being weaker. The first idea is the need to be specific. An advice that we often give PhD students is to narrow the research question. Sometimes we tackle research questions that are too broad. For example a research question like ‘‘why do companies use performance management systems?’’ is not a question for a research project but for a research field! Papers with broad research questions provide answers that do not meet expectations – at best they provide a partial answer to the questions and at worst the answer is hard to interpret. A similar problem happens with conclusions; these are either too broad (overestimating the actual finding) or trivial. For instance ‘‘the study shows that performance measurement systems affect operations’’ is too broad and obvious – we know that they affect operations; this is why we study them. Or, ‘‘the problem are related to the organization being in the public sector;’’ assigning the particular problem identified in the paper to the fact that the research was done in a public sector organization is too rough.

Performance Measurement and Management Control Systems

Table 1.

59

Random Thoughts on Performance Measurement and Control Systems’ Research.

1. Be specific: research questions and conclusions need to be sharp. A research question that is too broad is unlikely to be answered in the paper and conclusions that are too broad fail to convey the point of the research project. 2. Be convincing: your arguments need to be grounded in accepted (and acceptable) assumptions and contrast different explanations. Also don’t over promise and under deliver; it reduces the credibility of the work. 3. Be concise: You do not need to show in one paper all that you know about a topic. 4. Grab the attention of the reader: Academic literature is literature nevertheless and it needs to have an interesting hook to grab the reader. 5. Be smart about motivating the paper: ‘‘Nobody has done it’’ is usually a bad way to ‘‘sell’’ a paper. 6. Be sensitive to the power of language: The same research design can come across as dull and weak or as innovative and powerful. 7. Going beyond descriptions: Research is about describing what you do but also about analysis and learning points. 8. Be careful about your research design: Credibility in a research paper comes from a welldesigned and executed research design. Spend time describing it and the decisions you took to make it robust. 9. Use sharp concepts and arguments: Research quality increases when concepts arguments and hypothesis testing are sharp rather than vague. 10. Be creative: At the end of the day, every research paper has to have a spark of creativity that moves knowledge forward.

Probably, there are other organizations in the public sector that do not face this problem (which makes falsifying the conclusion an easy task); thus the more likely explanation is the presence of an omitted correlated variable but we don’t know which one is it. Or ‘‘specialization in the design of control systems is done only under specific conditions’’ (of course, what else would we expect! This conclusion has been proved again and again in contingency studies for the past 30 years). The second idea is to be convincing in the research design and writing. A paper needs to have assumptions that are well-argued or that are accepted in the literature. Weak assumptions limit the credibility of a paper. Certain papers just need better arguments (presenting at conferences and workshops and circulating the paper among colleagues is an effective way of identifying these weaker arguments); but sometimes authors present weak assumptions that are not even relevant to the paper! Examples of assumptions that hurt the credibility of the paper are: ‘‘I will assume that low performing companies need more external financing than high performing ones’’ (why is this? Growth companies need lots of funding and they grow because they

60

ANTONIO DAVILA

are successful), ‘‘The most frequent means of motivation is bonuses’’ (what about intrinsic motivation? There have been lots of papers on precisely this). Careful thinking (again presentations and discussions are very effective in being careful) is important to avoid these weak assumptions that, in most cases, can be easily corrected. Another way in which credibility decreases is by over promising and under delivering. For instance, starting the paper saying that the strength of the paper is using multiple paradigms but then using just one does not help. Criticizing financial measures but proposing no solution to it also weakens credibility. A third way to decrease credibility is to offer one view and one interpretation of the facts. This is especially relevant in case studies where arguing that the evidence is consistent with one theory is not enough. To be powerful, case studies also need to provide evidence inconsistent with alternative explanations. For instance: ‘‘using our framework led to improved company performance’’ (was it your framework or something else that managers used?), ‘‘In our opinion, the problem is y’’ (research needs facts not opinions; opinions need arguments in their favor and arguments that invalidate alternative opinions), ‘‘the relationship depends on various factors, one of them is learning’’ (how relevant it is compared to these other factors). A third idea is being concise. Most research projects can be explained in less than 35 pages (double space) (and often with 5 þ/2 hypotheses) and readers appreciate it. Before people read a paper, they look at its length and longer papers decrease the likelihood of a careful reading. Most often papers are long because of long literature reviews. A paper does not need to summarize all the literature that might be related to the question. Rather, the literature review for an empirical paper needs to be short and to the point, highlighting the contribution of the findings. Papers are not intended to show how knowledgeable the author is about a topic; they communicate findings and explain how these findings were achieved. Another reason why papers can become too lengthy is when they attempt to answer too many questions; papers address one or two questions. If there are more, then the author needs to identify the key findings to focus on having them come across rather than hide them amidst too many messages. A fourth idea is that a paper is a piece of literature. It needs to grab the reader from line one as a good novel does. Because papers are about research we tend to forget that they need to be read; we have to think about how we say it, not just what we say. The abstract is not about summarizing the sections of the paper, but about convincing readers that they will learn from reading it. Findings reported in the abstract that do not encourage

Performance Measurement and Management Control Systems

61

reading the paper include broad conclusions such as: ‘‘changing systems can lead to behavioral issues’’ (no need to read the paper to know it), or ‘‘public sector using non-financial rewards and private sector financial rewards’’ (we kind of know this). Also using key words such as ‘‘strategy’’ repeatedly does not make the paper more interesting: ‘‘the analysis of the identified strategic problem with strategic workshops to develop a strategic management tool and a strategic plan of action.’’ The fifth idea is to carefully work on motivating the paper. A well-known researcher said that he spends most of his time writing and re-writing the abstract and the first five pages. This is where the motivation of the paper is argued. Motivating the paper through ‘‘this is the first research paper that examines this question’’ or ‘‘there are no studies that address the question that we address’’ seldom work. First, somebody may have done it (and then the research becomes redundant) and second, doing a research project because nobody has done it before is not a good reason to do it. Motivating the paper based on the fact that this study happens in country X also fails often. Why do we expect country X to be different? If we expect it to be different, why don’t we study directly the variables that drive these differences? The sixth idea is to be sensitive to the power of language. The first aspect of this idea is to write as clearly as possible. An example of a sentence that does not make it easy for the reader is: ‘‘the performance measurement system to be practically applied given the context of the company in order to verify its proposals of new performance measurement bases.’’ A second aspect is to use language to the advantage of the authors instead of against them. For instance, if a research team follows a company for five years, interview four people several times, go to meetings, and go over internal documents; then the paper should not start the description of the research saying ‘‘I interviewed four people.’’ A third aspect is to be careful with tables. They are a key part of any paper and thus they need to be easy to read and self-explanatory. We often pay little attention to tables even if they communicate a key part of the message (especially when using research designs based on statistical tools). Also, the writing should communicate passion about the research rather than doubts: ‘‘we did field research, kind of a survey’’ (so it was not a survey?), ‘‘we interviewed some people’’ (how many?), ‘‘our first hypothesis provides secondary research questions’’ (why is it? Your first hypothesis is not interesting enough to address the primary research question?) The seventh idea is to go beyond descriptions. Descriptions of a company or descriptive statistics are not a full research paper. They are the first stage

62

ANTONIO DAVILA

of a research project – tables 1 and 2 out of several tables. Research is about advancing knowledge and learning points; thick descriptions might be interesting but they have to be carefully analyzed and lead to findings. Otherwise, descriptions just document observations, something that is very hard to publish. To go beyond this first stage, descriptions need to be related to the literature and identify learning points. Large sample studies need to go beyond descriptive statistics and t-tests into multivariate analysis to control for potential omitted variables. The eighth idea is to be careful about the research design: including the research question, hypotheses, variable measurement, sample selection, and analysis and the presentation of the paper. For instance, using one questionnaire item to measure one variable is too noisy and you cannot establish its reliability. This problem is compounded if what you measure is performance which we know is multi-dimensional. A careful description of where the data comes from, what is the sample, how big is the sample, how data was collected is essential. Some papers do not explain where the data comes from or try to avoid it: ‘‘in this section we briefly mention our research method’’ (no need to be brief in the research section – brevity is for the literature review section). In the statistical analysis the typical threats are endogeneity and omitted correlated variables. The ninth idea is to use sharp concepts and arguments. Hypotheses need to be well-argued; a hypothesis is not the same as an opinion. Moreover, they cannot be simplistic such as ‘‘low employee turnover is associated with higher stock price’’ (the relationship between these two variables, if it exists, is a lot more complex than this). The same idea holds for concepts, SMEs is not the same as private companies and country is not a good proxy for country characteristics (such as education or culture) because even if country turns out to be significant in a statistical analysis, it is unclear what is driving the result because there are multiple variables correlated with country. Thus, the fact that country is significant is not much of a finding if we don’t know why it matters. Finally be creative. Research is about creativity and discipline. Do not repeat what has been already done. If your data allows to test an existing result do not write it as a hypothesis. Summarizing what has been already done is helpful for PhD students but has no contribution. Bringing ideas from other fields is often a good idea, but it still needs creativity. For instance, bringing a concept such as customer equity – that is becoming very popular in the marketing literature – into accounting is not enough. In other words, describing what customer equity is and how it is applied in a company might become good teaching material but it needs a bit more to be

Performance Measurement and Management Control Systems

63

researched. All this information is already out there, albeit in a different literature. Bringing concepts from other fields is important – it encourages cross-disciplinary work which is a critical aspect of performance measurement and control systems; but they need to be studied from the perspective of our field so we can add to knowledge – either refining the concept or applying the concept into new settings.

3. CONCLUSIONS The quality and quantity of the papers at the conference speak to the dynamism of research in performance measurement and control systems. The field is tackling questions that are important for theory development but also for managers. There is diversity in terms of research approaches, theoretical perspectives, and research settings. This diversity indicates a rich environment that helps advancing knowledge pushed in several directions at the same time. Most papers at the conference show a depth and quality of research that promise a bright future to the field. A vast majority of the shortcomings identified relate to the work in progress nature of the papers; these shortcomings are details that significantly enhance the perception of quality. In 2009, the fifth conference in performance measurement and control systems will take place. The evolution of the field that we have witnessed over the last four editions promises lots of new ideas and insights that will prove to enrich our understanding of this important aspect of management.

REFERENCES Abdel-Maksoud, A., & Pollanen, R. (2007). Relationships among advanced managerial practices, management accounting techniques, competition and performance measurement: C. Working Paper. University of Sharjah. Agasisti, T., & Bianco, A. D. (2007). Efficiency measurement in Italian university system: Non-parametric and parametric approaches. Working Paper. Politecnico di Milano. Agyemang, G., & Ryan, B. (2007). Accountability relationships in the public and private sectors: Fuzzy connections and missing links. Working Paper. University of London. Al Sharif, N. (2007). New public management and performance measurement in france: A national cultural analysis. Working Paper. Orleans University. Amenta, C. (2007). A performance indicator to reward museum marketing effort. Working Paper. Universita` di Palermo.

64

ANTONIO DAVILA

Anderson, S. W., Davis, G., & Widener, S. (2007). Strategy, cost structure and performance in the U.S. domestic airline industry: An investigation of causality. Working Paper. Rice University. Andersson-Fele, L. (2007). Time to revive Luther Gulick-on span of control and organisation quality. Working Paper. Gothenburg University. Arena, M., Arnaboldi, M., & Azzone, G. (2007). Risk management and budgeting: The quest for integration. Working Paper. Politecnico di Milano. Argyropoulou, M., Ioannou, G., & Koufopoulos, D. (2007). Performance measures for erp system implementation. Working Paper. Athens University of Economics and Business. Arizzio, M. E. (2007). Human resources competence and performance evaluation empirical model in a consultant company. Working Paper. Milano Polytechnic University. Baraldi, S. (2007). Balancing formal rewarding and intrinsic motivation: The role of the balanced scorecard in professional organizations. Working Paper. Catholic University Milan. Barontini, R., & Bozzi. S. (2007). Executive compensation and ownership structure: Empirical evidence for Italian listed companies. Working Paper. Scuola Superiore Sant’Anna. Bhattacharya, N., Black, E., Black, D., & Christensen, T. E. (2007). The effects of regulation on pro forma reporting trends. Working Paper. Southern Methodist University. Bodnar, V., Danko, D., Kiss, N., & Vas, G. (2007). Changes in the management control toolkit of Hungarian organizations (1996–2006). Working Paper. Corvinus University of Budapest. Bonacchi, M., Ferrari, M., & Pellegrini, M. (2007). The life time value scorecard: From E-metrics to internet customer value. Working Paper. University of Naples. Booker, D., & Heitger, D. (2007). The effect of causal performance measure knowledge on reducing individuals’ discounting of performance measures in profit pred. Working Paper. Miami University. Boulianne, E. (2007). Examination of the balanced scorecard construct validity with the multitrait-multimethod matrix. Working Paper. Concordia University. Bratnicki, M., & Dyduch, W. (2007). Entrepreneurship-based performance measurement system. Working Paper. The Karol Adamiecki University of Economics in Katowice. Broad, M., & Jadsri, P. (2007). Accounting and performance control systems in public universities. Working Paper. Jyvaskyla University. Busanelli De Quino, A. C., Cardoso, R. L., Boya, V. L. A., & Pagliarussi, M. S. (2007). Causality in a performance measurement model: A field study in a Brazilian power distribution company. Working Paper. University of Sao Paulo. Campbell, D., Datar, S., & Sandino, T. (2007). Franchising and control across multiple markets: The case of the convenience store industry. Working Paper. Harvard Business School. Chapman, C. (2007). Performance measures, pay and accountability: A field study in an audit firm. Working Paper. Oxford University. Chen, X., & Sandino, T. (2007). Do internal management controls mitigate employee theft in chain organizations? Working Paper. University of Illinois at Urbana-Champaign. Chenhall, R., Kallunki, J.-P., and Silvola, H. (2007). Control within the cultural context of blat: A study of management control systems, strategy, innovation and performance within. Working Paper. Monash University. Chiaravalloti, F., & Meer-Kooistra, J. V. D. (2007). Performance measurement and management for performing arts. Including artistic quality indicators in the PMS’s of publicly fund. Working Paper. University of Groningen.

Performance Measurement and Management Control Systems

65

Chiesa, V., Frattini, F., Lazzarotti, V., & Manzini, R. (2007). Are there any differences in performance measurement between research and development? Evidence from an empirical study. Working Paper. Politecnico di Milano. Chiucchi, M. S. (2007). The organizational and strategic benefits of measuring intellectual capital performance: Evidence from some Italian companies. Working Paper. University of the Marche. Claes, P. (2007). Implementation of performance management in a Dutch diversified multinational: The resurrection of value-based management. Working Paper. Vrije Universiteit Amsterdam. Constanti, P., Krambia-Kapardis, M., & Constanti, A. (2007). Performance management: Cyprus, an exploratory study. Working Paper. Intercollege Limassol. Cugini, A., & Michelon, G. (2007). Performance evaluation in research departments: From the balanced scorecard to the strategy map. Working Paper. University of Padova. Da Silva, S. L., & Rojas Lezana, A. G. (2007). A social performance management model under stakeholders’ orientation. Working Paper. Santa Catarina Federal University. De Souza Lenz, G., Hansen, P. B., Fedrizzi, L. d. B., & Roldan L. B. (2007). Towards a multitheoretical approach to performance measurement in cooperative horizontal networks: A system proposal based upon. Working Paper. Pontificia Universidade Catolica do RS. De Waal, A. A., & Roobol, M. (2007). Performance-related pay in the home furnishing industry: Any chance of success? Working Paper. Maastricht School of Management. Decramer, A., Christiaens, J., & Vanderstraeten, A. (2007). Individual performance management in higher education institutions: Strategic responses of faculties towards institutional pre. Working Paper. University College Ghent. Dekker, H. C., & Van Den Abbeele, A. (2007). Partner selection, knowledge acquisition and interfirm governance design. Working Paper. Vrije University, Amsterdam. Delai, I., & Takahashi, S. (2007). Including sustainability in the organization performance measurement system. Working Paper. Universidade de Sao Paulo. Ding, R., Dekker, H., & Groot, T. (2007). An exploration of use of inter-firm cooperation by Dutch firms and involvement of financial managers. Working Paper. Vrije University, Amsterdam. Emmanuel, C. R., Kominis, G., & Slapnicar, S. (2007). The impact of target setting on managerial motivation: The case of an Eastern European bank. Working Paper. University of Glasgow. Fazzini, M., & Terzani, S. (2007). The use of balanced scorecard in Italian fashion companies. Developing multiple performance measures. Working Paper. Universita degli studi di Napoli ‘‘Parthenope’’. Fiori, G., Donato, F. D., & Izzo, M. F. (2007). Corporate social responsibility and firm performance. An analysis on Italian listed companies. Working Paper. Luiss Guido Carli University. Fischbacher, U., Stefani, U., & Pfaff, D. (2007). Specific investments, synergy effects and transfer pricing – experimental results. Working Paper. Zurich University. Folgueras, A., Crespo, A. G., Berbis, J. M. G., & Mezcua, B. R. (2007). Information economics: A value added system driven model. Working Paper. Carlos III University of Madrid. Forslund, H., & Jonsson, P. (2007). The performance management process concerning on-time delivery – state-of-the-art description and perceived performance. Working Paper. Vaxjo University.

66

ANTONIO DAVILA

Franchino, J., Laura, B., & Paolo, C. (2007). The usefulness of making a selection among balanced performance indicators. The multidimensional reports structured through key. Working Paper. University of Eastern Piedmont. Frezatti, F., Relvas, T. R. S., & Junqueira, E. (2007). Balanced scorecard and management accounting attributes structure: An analysis in the Brazilian environment. Working Paper. University of Sao Paulo. Frow, N., Ogden, P. S., & Marginson, P. D. (2007). Legitimising budget flexibility whilst maintaining budgetary control: Evidence from a multi-national organisation. Working Paper. Warwick Business School. Fujino, M. (2007). The role of performance measurement systems in public management reforms. Working Paper. Nihon University. Gabrie¨ls, X., & Jorissen, A. (2007). The (perceived) value of information for performance measurement purposes after an ERP adoption. Working Paper. University of Antwerp. Garelli, R., & Longobardi, A. (2007). A journey through the intellectual capital. Working Paper. Universita di Genova. Garengo, P., Biazzo, S., & Bernardi, G. (2007). A circular approach to performance measurement in Smes. Working Paper. University of Padua. Groot, T., Dekker, H., & Schoute, M. (2007). Determining performance targets: The role of historical performance, future expectations and benchmark information. Working Paper. Vrije University, Amsterdam. Hartmann, F., & Slapnicar, S. (2007). How formality of managerial control systems affects managerial justice perceptions, motivation and organisational commitment. Working Paper. RSM Erasmus University. Horva´th, P., Lattwein, J., & Moeller, K. (2007). Strategy, control and performance in the automotive supplier industry. Working Paper. International Performance Research Institute. Ishak, A., & Lee, A. (2007). Strategy and performance measures for organisational performance: A study of the biotechnology industry. Working Paper. Curtin University of Technology. Ja¨a¨skela¨inen, A., Lo¨nnqvist, A., Laaksonen, T., & Kulmala, H. I. (2007). Empirical analysis on the factors affecting the design of performance measurement systems in public sector. Working Paper. Tampere University of Technology. Jansen, E. P. (2007). Leadership style and changes in management accounting & control: A case study of a multi-outlet car dealership. Working Paper. University of Groningen. Ja¨rvenpa¨a¨, M., & La¨nsiluoto, A. (2007). Exploring environmental management accounting change: Mapping the journey to the (promised) green land and making good corporate. Working Paper. University of Jyvaskyla. Johansson, T., & Siverbo, S. (2007). Rational, political and institutional/cultural explanations of the utilization of relative performance evaluation in Swedish Lo. Working Paper. Goteborg University. Kalagnanam, S., Kobussen, G., & Vaidyanathan, G. (2007). The introduction, development and use of performance management systems in government owned (crown) corporations: The role of I. Working Paper. University of Saskatchewan. Kandel, A. (2007). A fallacy of the outcome effect hypothesis. Working Paper. Caldwell College. Karbhari, Y., & Mohamad, M. H. S. (2007). The impact of the modified budgeting systems (Mbs) on performance measurement in Malaysian central government organisations. Working Paper. Cardiff University. Kastberg, G. (2007). Performance measurement in health care: Balanced scorecard in a professional context. Working Paper. Gothenburg University.

Performance Measurement and Management Control Systems

67

Kruis, A.-M., & Spekle´, R. F. (2007). Management control system design and effectiveness: An empirical analysis of 258 Dutch organizational units. Working Paper. Nyenrode Business University. Kudo, H. (2007). Performance measurement and policy evaluation in Japanese local government: Between enthusiasm, expectation, and skepticism. Working Paper. Chuo University. Kunz, J. (2007). Influence of performance measures on organizational long-term performance. Working Paper. Johann Wolfgang Goethe University. Likierman, A. (2007). Measuring the success of performance related pay for executives. Working Paper. London Business School. Linder, S. (2007). Tying vs. not-tying post-completion reviews to extrinsic rewards and/or punishments. Working Paper. CTcon GmbH. Lucianetti, L. (2007). An empirical analysis of performance measures and management accounting practices in large Italian manufacturing firms. Working Paper. University of Chieti. Macdonald, S., & Kam, J. (2007). Never mind the quality: Rewarding gamesmanship in UK universities. Working Paper. University of Sheffield. Madini, P. (2007). Performance management: An exploratory case study of a negotiated budgetary process. Working Paper. 2GC Active Management. Matejka, M., & Maas, V. (2007). Balancing the dual responsibilities of business unit controllers: Field and survey evidence. Working Paper. University of Michigan. Michel, J.-E. G. E., & Michel, D. E. (2007). Managers acceptance criterias for performance measurement. Working Paper. HEC School of Management. Micheli, P., & Kennerley, M. (2007). The roles of performance measurement in the public sector in England. Working Paper. Cranfield School of Management. Mishken, M. A., & Juhasz, K. (2007). Early performance ratings and personality measures in career advancement. Working Paper. Pace University. Misiaszek, F., Oriot, F., & Otley, D. (2007). Strategic scorecarding systems in action in French Smes. Working Paper. Ecole Superieure de Commerce de Toulouse. Mpabanga, D. (2007). Performance management and appraisal at the university of Botswana: Is it management control or a quest for superior performance. Working Paper. University of Botswana. Mucciarone, M., & Neilson, J. (2007). Performance reporting in an Australian government context: A comparative study. Working Paper. Curtin University of Technology. Mueller, F. (2007). Sales management control strategies in banking – strategic fit and performance impact. Working Paper. Europan Business School. Naranjo-Gil, D., Maas, V., & Hartmann, F. (2007). Explaining management accounting innovation: The effects of strategy, historical performance and Cfo characteristics. Working Paper. Pablo de Olavide University. Needles, B. E., Powers, M., & Frigo, M. L. (2007a). Performance measurement and executive compensation: Practices of high performance companies. Working Paper. DePaul University. Needles, B. E., Powers, M., Shigaev, A., & Frigo, M. L. (2007b). High performance companies in developing and developed countries: The case of India and the United States. Working Paper. DePaul University. Nevries, P., Hoffjan, A., & Winn, M. (2007). Comparative management accounting – An analysis of French and German management accounting practices. Working Paper. Otto Beisheim Graduate School of Management.

68

ANTONIO DAVILA

O¨zer, P., & Kocakoc- , I. D. (2007). Assessing the relative performance weights of employee performance criteria with AHP: A case study from a Turkish automotive co. Working Paper. Dokuz Eylul University. Parisi, C., & Sansalvadore, F. (2007). An explorative study on the implementation of the balanced scorecard on value creation. The case of car distribution sector. Working Paper. University of Florence. Pavlov, A., & Bourne, M. (2007). Understanding the foundations of ‘managing through measures’: The impact of performance measurement on organizational process. Working Paper. Cranfield School of Management. Pekkola, S., Ukko, J., & Rantanen, H. (2007). Linking rewards to performance measurement: Challenges in the private and public sector. Working Paper. Lappeenranta University of Technology. Pesalj, B. (2007). The application of network theory on a conceptualization of performance measurement system of MNEs. Working Paper. Faculty of Economics Belgrade. Reheul, A.-M., & Jorissen, A. (2007). The role of performance on the choice of Mcs in Smes: The integration of resource dependence theory in the conventional conti. Working Paper. University of Antwerp. Rejc Buhovac, A., Savicˇ, N., & Kecˇanovic´, B. (2007). Strategic management control over the use of police force. Working Paper. University of Ljubljana. Roberts, M., Cauvin, E., & Neumann, B. (2007). Effects of financial versus nonfinancial performance, financial measures’ importance, and presentation order on corporate manag. Working Paper. University of Colorado at Denver. Schiehll, E. (2007). Private performance information in CEO incentive compensation. Working Paper. HEC Montreal. Schryver, T. D., Eisinga, R., Teelken, C., & Poutsma, E. (2007). The missing link between information and action: Hastenings and delays as universal reactions to performance feedback. Working Paper. Radboud Universiteit Nijmegen. Schultze, W., & Weiler, A. (2007). Performance measurement, value creation, and managerial compensation: The missing link. Working Paper. University of Jena. Shulver, M., Lawrie, G., Barney, W., & Kalff, D. (2007). Aligning individual performance management with business strategy: A case study in the UK insurance industry. Working Paper. 2GC Active Management. Silvi, R., Bartolini, M., Raffoni, A., & Visani, F. (2007). Strategic control and management accounting systems in Italian companies: Evidences from the field. Working Paper. University of Bologna. Simpson, M., Padmore, J., & Frecknall-Hughes, J. (2007). The difficulties of measuring performance in small and medium sized enterprises. Working Paper. Sheffield University. Soulerot, M., & Farjaudon, A.-L. (2007). Management control’s implications of the exploitation/ exploration dilemma: Empirical evidence from a fast-moving consumer goods. Working Paper. DRM-CREFIGE. Speckbacher, G., & Wentges, P. (2007). The impact of firm size and family ownership on management control systems in small and medium sized enterprises. Working Paper. Wirtschaftsuniversitat Wien. Suomola, P., Varila, M., & Jokioinen, I. (2007). Validity problem in performance measurement: A conceptual analysis with an illustrative case on quality cost measurement. Working Paper. Tampere University of Technology.

Performance Measurement and Management Control Systems

69

Szanto, R., & Wimmer, A. (2007). Performance management and value creation – a stakeholder approach. Working Paper. Corvinus University of Budapest. Turolla, A., Vola, P., Carenzo, P., & Franchino, J. (2007). Balanced scorecard in the pubblic sector: The case study of the health sector in the north of Italy. Working Paper. University of Eastern Piedmont. Ukko, J., Karhu, J., & Pekkola, S. (2007). Employees satisfied with performance measurement and rewards: Is it even possible? Working Paper. Lappeenranta University of Technology. Ulucan, A. (2007). Efficiency evaluation with multi-dea analysis: An application in a world bank supported project. Working Paper. Hacettepe University. Van Der Merwe, N., & Visser, S. (2007). Demonstrating improved performance management in the automotive industry of a developing country. Working Paper. North-West University. Verbeeten, F. (2007). The use of control systems in public sector organizations: An empirical investigation in Dutch municipalities. Working Paper. Rotterdam School of Management. Vola, P., Broccardo, L., & Truant, E. (2007). Performance measurement under balanced scorecard: The case study of a co-operative credit bank in piedmont. Working Paper. University of Turin. Wall, F. (2007). Performance measurement and behavioural control: The moderating role of Erp-systems. Working Paper. University of Witten/Herdecke. Yamane, S., & Asada, T. (2007). An empirical study on the function of strategic management control in environmental new product development. Working Paper. Osaka University. Yongvanich, K., & Guthrie, J. (2007). The balanced scorecard of Thailand listed companies and performance implications. Working Paper. Ramkhamhaeng University. Zakrzewski, D., & Juchau, R. (2007). Airport privatization: How to measure airport performance from stakeholder viewpoints. Working Paper. University of Western Sydney. Zraly, M. (2007). Integration concept of management control and its contribution to performance management. Working Paper. Czech Technical University in Prague. Zurwehme, A. (2007). Does management accounting influence organizational performance of service units? The case of training institutions. Working Paper. Dresden University of Technology.

This page intentionally left blank

PART II: EVALUATING AND MOTIVATING EMPLOYEE PERFORMANCE

EARLY PERFORMANCE RATINGS AND PERSONALITY MEASURES IN CAREER ADVANCEMENT Mark A. Mishken and Krisztina Juhasz ABSTRACT Although performance assessment systems are widely used in government agencies, they are underutilized in promotional decisions. Most promotions are based on the results of knowledge-based exams to the exclusion of other data such as job performance ratings, personality constructs, and motivation. Two empirical studies examined this and made recommendations for change. The first study examined the relationship between early career performance appraisal scores and future organizational advancement. The second study examined how personality variables correlated with desire to advance in the organization to attain supervisory positions. Findings provide support for incorporating job performance measures and personality constructs in promotional decision-making.

INTRODUCTION Most companies are interested in determining which employees are likely to progress within their organization; succession planning has become

Performance Measurement and Management Control: Measuring and Rewarding Performance Studies in Managerial and Financial Accounting, Volume 18, 73–94 Copyright r 2008 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-3512/doi:10.1016/S1479-3512(08)18004-2

73

74

MARK A. MISHKEN AND KRISZTINA JUHASZ

common practice in many firms (Garman & Glawe, 2004). How do we identify these employees early on in their careers? Can early performance appraisal ratings determine which employees are likely to advance in the organization? Also, are there certain traits that differentiate between employees who desire to advance versus those who do not? And finally, can we screen a large applicant pool quickly and efficiently to identify employees likely to progress in the organization. Several methods may be employed by an organization to identify individuals who are likely to advance, including performance appraisal ratings, job-knowledge tests, or personality measures. Tools such as the performance appraisal are often used in the private sector. However, in the public sector organization, often decisions for promotion are based solely on some knowledge-based tests. Although knowledge-based tests may predict job performance, they do not tap into the motivational components of job performance. Vroom’s (1964) model of performance concludes that job performance is determined at least partly by motivation. For example, Ferris, Witt, and Hochwarter (2001) found that employees who scored high on general mental ability (GMA) tests and social skills had the highest levels of performance and salary. Individuals with low social skills and high GMA (or vice versa) did not perform as well. Also, GMA has been shown to be a valid predictor of job performance, but more so for complex job categories (Hunter & Hunter, 1984; Salgado et al., 2003; Salgado, Anderson, Moscoso, Bertua, & Fruyt, 2003a). The findings that job complexity moderates the validity of cognitive ability tests was demonstrated in American (Hunter & Hunter, 1984) as well as European samples (Salgado et al., 2003a, 2003b). Recently, Bartram (2005) demonstrated that ability tests correlate with four of the eight main job competencies, while ability and personality tests correlates with all eight of the main job competencies. As indicated by the above results, different factors (such as GMA and social skills) contribute to high job performance, job-knowledge tests may not be sufficient to evaluate who should and should not be promoted and may not be a strong predictor of which employees desire to advance in the organization. In-house developed job-knowledge tests (as opposed to standardized GMA tests and test batteries) are usually the most common tools used in the public sector organizations to make promotional decisions (Salgado et al., 2003a, 2003b). The current chapter examines two additional methods not traditionally used for these purposes in the public sector – performance appraisal ratings and personality measures.

Early Performance Ratings and Personality Measures

75

Performance appraisals tap into the knowledge and other competencies of performance. As a result, performance appraisals have become increasingly important to organizations as a tool for major decisions such as promotions and layoffs (Harris, Gilbreath, & Sunday, 1998; Martin, Bartol, & Kehoe, 2000), especially in the private sector. Likewise, personality measures have also been shown to tap into motivation and may also be a useful means to identify candidates or employees who aspire to advance in an organization. Several studies provide evidence that there are trait differences between those who seek managerial positions versus those who do not. The trait of ambition has been one of the most useful predictors of managerial advancement (Howard & Bray, 1988; Tharenou, 2001). More, recently, research found that traits such as managerial aspiration (the extent to which employees desire advancement into management positions), masculinity and adaptability predicted managerial advancement (Tharenou, 2001; Metz, 2004). The chapter explores how these two tools – performance appraisals and personality testing – can be used to identify individuals likely to advance in a government agency.

EARLY PERFORMANCE APPRAISAL RATINGS AND CAREER PROGRESSION Bartram (2004) argued that organizations are focused more on determining the validity of selection procedures than on the validity of their post-hire instruments such as performance assessments. This is especially true for government agencies. The overall consensus is that they lag behind the private sector when it comes to performance appraisals (Grote, 2000; Tziner, Kopelman, & Joanis, 1997). Many agencies still rely on traditional performance appraisal and promotion methods. For various reasons (such as union involvement, lack of merit pay incentives, and competency exams), most government agencies have not been at the forefront of new performance management developments. Organizations tend to rely on performance appraisal ratings to make promotion decisions. However, in government organizations, the method in which promotion decisions are made can be very different for various positions within the organization. In many governmental positions, scores on a job-knowledge examination rather than performance appraisal ratings may be the major (or exclusive) determinant of promotion. These positions where knowledge-based test are used for promotion are usually termed

76

MARK A. MISHKEN AND KRISZTINA JUHASZ

‘‘competitive’’ positions. While performance appraisal ratings are likely related to the acquisition of the job knowledge required to do well on a promotional job-knowledge test, these ratings may not enter directly into the promotion process. Performance appraisal ratings may also tap into such aspects as interpersonal components or contextual factors (Fletcher, 2001; Viswesvaran, Schmidt, & Ones, 2005) of job performance whereas jobknowledge tests may not. Southworth (2000) makes the argument that jobknowledge exams are insufficient at determining the best candidates for promotion. Often, poor performers on the job are able to receive promotions based on their ranking on the knowledge-based exam, whereas some high performers are hindered from progressing in the organization due to their low ranking on these exams (Southworth, 2000). Positions in which a job-knowledge test is not used may allow managers a larger degree of latitude in the types of information that can be considered in promotion decisions. These positions, where knowledge-based tests are not used for promotion are termed ‘‘non-competitive’’ positions. Because performance appraisal ratings are readily available in such situations, it is likely that they may play a role in promotion decisions. The correlation between performance ratings and merit pay has been found to range between .17 and .27 (Harris et al., 1998). In their review, Gerhart and Milkovich (1990) argue that the relationship between performance ratings and pay raises is not very strong, especially in some organizations. Ferris et al. (2001) suggest that the weak relationship may be due to the managers’ discomfort with creating large differentials between the salaries of the employees. This relationship between performance ratings and salary/promotional progression may be even weaker in government organizations that employ job knowledge-based exams as a basis for promotion to most job titles.

STUDY ONE Research on the extent to which performance appraisal ratings are related to promotional progression may prove useful. A commonly cited aphorism in psychological research is that past behavior predicts future behavior. Similarly, it could be expected that past employee performance should predict future performance. Although individuals may vary in their performance over time and in different contexts, and thus may receive varied performance appraisal ratings over the years, some consistency in behavior can be expected. In recently published research, the case for

Early Performance Ratings and Personality Measures

77

performance dynamism, that is the claim that performance is not static but changes over time, has been supported. However in the same study (Sturman, Cheramie, & Cashen, 2005) the researcher found that over time there was still a significant amount of variance accounted for by previous measures of job performance. In another study, Viswesvaran et al. (2005) found that although halo error affected job performance ratings, there was a substantial ‘‘general factor’’ comprised of cognitive ability and contextual or citizenship elements that were accounting for a significant amount of variance. If this were so, then it would make sense for earlier performance ratings to predict future organizational performance. Specifically we hypothesized that: (1) In a longitudinal study, overall job performance ratings will significantly correlate with organization progression. (2) The overall job performance ratings will significantly correlate with organizational progression higher for the ‘‘non-competitive’’ employee group than for the ‘‘competitive’’ employee group.

METHOD Sample The sample consisted of 299 (58% male and 42% female) employees of a large state government agency in the Northeastern United States who were in their first year of a job within the competitive title series of Court Clerk during the period 1990–2007. Individuals in this series typically perform the tasks of lower to mid-level managers, and this job group is considered a key ‘jump off’ point from which court employees can progress through the organizational hierarchy. Typical duties of the Court Clerk series may include: swearing of witnesses, polling jurors, maintaining custody of exhibits, keeping minutes of court proceedings, supervision of court bailiffs and other court personnel, reviewing calendaring decisions, court motions and forms, interacting with the public and attorneys, answering inquires, and interpreting court orders. The self-reported ethnic background of participants was 74% Caucasian, 19% African American, and 6% Hispanic. See Table 1 for descriptive statistics. Of the 299 employees, 87% of the clerks progressed into competitive positions (based on knowledge tests) and 13% of the employees were promoted within non-competitive positions (not based on knowledge tests).

78

MARK A. MISHKEN AND KRISZTINA JUHASZ

Table 1. Study One – Descriptive Statistics of the Overall Sample.

Year of appraisal Overall job performance Grade level during appraisal Grade level change Year of hire

N

Mean

Standard Deviation

299 295 298 251 298

1992.46 4.471 19.7148 2.5618 1985.69

1.574 .6395 2.95577 3.17981 5.899

The average tenure in the organization did not differ for the group of individuals in competitive (14.53 years) versus non-competitive (14.74 years) positions. However, the non-competitive sample of employees was hired into the organization at a significantly higher mean grade level (21.18) than the competitive sample (19.50), and had a significantly higher mean grade change (6.11) than the competitive sample (1.99). In addition, the noncompetitive sample received significantly higher overall job performance ratings (4.7) than the competitive sample (4.4). We controlled for these differences in our analyses. The two samples also differed in gender composition. The competitive sample consisted of 60.5% female and 39.1% male. The non-competitive sample consisted of 39.5% female and 60.5% male. If individuals obtained multiple Court Clerk jobs within the 1990–2007 period, the earlier performance appraisal was used and the later one was deleted from analyses in order to ensure independence of observations. The mean number of years between their performance appraisal ratings and progression data were gathered was 15.54 years. There was no significant difference in the latter for the competitive and noncompetitive groups (14.54 vs. 14.71).

The Performance Appraisal Form The evaluation form used for probationary ratings utilized a Behavioral Observation Scale (BOS) format (Latham & Wexley, 1977, 1981). This instrument is behavioral as opposed to one that is graphical or global in nature (Spector, 2003). One of the main goals of performance appraisal is the giving of feedback to employees to aid in their development. (Fletcher, 1986, 2001). Though there is some disagreement with respect to what type of appraisal suits this purpose (Fletcher, 2001; Kane & Bernardine, 1982), the

Early Performance Ratings and Personality Measures

79

BOS has received support in this area and its general usefulness from numerous sources (Latham & Wexley, 1977, 1981; Wiersma & Latham, 1986; Jelley & Goffin, 2001; Latham & Mann, 2006). This specific form required supervisors to rate subordinate behaviors related to four jobrelevant work dimensions. First, subordinates were rated on their performance of court clerical activities. This dimension included behaviors related to correctly completing court forms, reviewing documents for defects, and accurately calculating fees and fines. Second, subordinates were rated on their planning and organizing of work. This dimension contained behaviors related to completing high priority work first, making prompt decisions, avoiding duplication of work, and basing plans on unit objectives. Third, managers rated subordinates on providing effective supervision. This dimensions contained behaviors related to effectively delegating work, ensuring that subordinates’ work follows organizational procedures, training subordinates, and evaluating subordinates’ work. Finally subordinates were assessed on their interaction with others in the organization. This dimension contained behaviors such as working cooperatively and tactfully with co-workers, sharing knowledge with colleagues, and presenting ideas and opinions in an effective way. These four dimensions will be referred to as ‘Court Clerk Activities,’ ‘Planning and Organizing,’ ‘Effective Supervision,’ and ‘Interaction with Others.’ Supervisors rated the importance of each of these dimension to the specific job of the subordinate. This rating was made on a four-point scale anchored by ‘not important/not relevant’ and ‘extremely important.’ In addition, under each work activity dimension a number of critical observable behaviors (as derived from a job analysis of court clerk positions) required to accomplish the major functions of the job were listed. Supervisors were asked to rate the extent to which the employee exhibited each behavior. These ratings were made on a four-point scale anchored by ‘very rarely’ and ‘almost always.’ Supervisors had the additional options of marking a behavior as not relevant to the successful performance of the employee’s job and adding up to three additional job-related behaviors that they felt would provide better representation of the employee’s job. After completing the performance appraisal form, supervisors rated subordinates’ overall job performance on a five point scale ranging from 1 ¼ unsatisfactory to 5 ¼ excellent. Supervisors were given the instructions to base their ratings of overall job performance on a consideration of the behaviors included in the performance appraisal form and any other jobrelated factors they felt were appropriate.

80

MARK A. MISHKEN AND KRISZTINA JUHASZ

Promotional Progression Because changes in salary are often tied to tenure in government organizations, salary was not used to measure progression through the organization. Instead, grade changes were used as a measure of progression. Thus, a greater change in grade indicates greater progression upward in the organizational hierarchy. Therefore, in order to obtain a grade change score, the salary grade at the time of the original performance appraisal was subtracted from the individual’s salary grade for the year 2007.

RESULTS – STUDY ONE To explore the relationship between overall performance ratings and grade change, we conducted a partial correlation analysis controlling for year of evaluation, grade at time of evaluation, and year of hire. The partial correlations were necessary because individuals with chronologically earlier review periods would have had a greater chance of obtaining promotions, individuals in higher grade levels may have had less opportunity to move upwards, and individuals who have worked in the organization longer have had more years to advance. The results indicated a significant positive relationship between overall ratings and grade change r ¼ .18, p ¼ .002, n ¼ 241. We corrected for unreliability in performance ratings using the formula noted in Bobko (2001). As suggested by Viswesvaran, Ones, and Schmidt (1996), we used .52 in the formula as a measure of interrater reliability in performance appraisals. The correction for unreliability yielded a correlation of .25. A hierarchical regression analysis was computed. The three control variables (year of evaluation, grade at time of evaluation, and year of hire) were added into the regression equation in step 1 and then overall performance ratings were entered in step 2 to predict organizational progression. Overall job performance ratings added significant variance in predicting organizational progression (see Table 2 for a summary of the results). Thus Hypothesis 1 was supported. In this longitudinal study, overall job performance ratings significantly correlated with organizational progression.

Organizational Progression: Competitive versus Non-Competitive Jobs For individuals whose current (year 2007) position was obtained through taking a job-knowledge examination (competitive), the partial correlation

81

Early Performance Ratings and Personality Measures

Table 2.

Study One – Hierarchical Regression Analysis: Prediction of Organizational Progression.

Variable

Unstandardized Coefficients

Step 1 (Constant) Year of hire Year of appraisal Grade level in the organization at time of appraisal Step 2 (Constant) Year of hire Year of appraisal Grade level in the organization at time of appraisal Overall job performance

Standardized Coefficients

B

Standard error

Beta

t

Sig.

222.473 .091 .201 .081

217.330 .037 .104 .074

– .188 .110 .084

1.024 2.467 1.926 1.100

.307 .014 .055 .272

273.626 .076 .213 .149

213.123 .037 .102 .074

– .156 .116 .154

1.284 2.075 2.080 2.003

.200 .039 .038 .046

.953

.257

.213

3.713

.000

Note: Dependent variable: Organizational progression. R2 ¼ .07, R ¼ .26 for Step 1; R2 ¼ .11, R ¼ .33 for Step 2.

coefficient (controlling for year of probationary evaluation, grade at time of evaluation, and year of hire into the organization) between performance ratings and progression was r ¼ .14 (p ¼ .024, n ¼ 207). Correcting for unreliability resulted in a correlation of .19. However, for positions which did not require individuals to take a job-knowledge exam (non-competitive), the partial correlation coefficient (controlling for year of evaluation, grade at time of evaluation, year of hire, and number of years in the noncompetitive title) between job performance and progression was r ¼ .32 (p ¼ .038 (1-tail), n ¼ 29). Correcting for unreliability resulted in a correlation of .44. The difference between the two correlations (uncorrected) was not significant. Thus, Hypothesis 2 was not supported likely due to the small non-competitive sample size of 29. A previous study by the authors found a statistically significant difference between the two groups of employees (Mishken, Ruminson, & Juhasz, 2005).

DISCUSSION Overall, the results were supportive of a relationship between supervisors’ performance appraisal ratings and future promotional progression. As

82

MARK A. MISHKEN AND KRISZTINA JUHASZ

mentioned, the mean years between the ratings and the snapshot of how the employees progressed as measured by grade change was 15.54 years. This is a significant amount of time. This seems to support the current research (Sturman et al., 2005; Viswesvaran et al., 2005) that although behavior dynamism indeed exists, there is still some stability in job performance. This research also finds that performance appraisals measure a general factor comprised of cognitive and contextual or citizenship skills. Combined, these measures should have value in predicting future performance. Although not statistically significantly different (see Mishken, Ruminson, & Juhasz, 2006), the finding that overall performance appraisal ratings correlate higher for ‘‘non-competitive’’ employees’ progression than for ‘‘competitive’’ employees may be accounted for by the way these employees are promoted. The richness of the overall performance rating in the performance appraisal which is based on cognitive and contextual elements (Viswesvaran et al., 2005) may be a better predictor than knowledge tests which may have a cognitive component but little if any citizenship or contextual factors.

PERSONALITY MEASURES AND ORGANIZATIONAL PROGRESSION Most companies are interested in determining which employees are likely to progress within their organization to assume supervisory roles. As Kirkpatrick and Locke (1991) note, it is difficult to judge which employees have a desire to supervise – especially when the workers are at entry-level and have little or no work experience. They suggest that organizations should utilize assessment centers to determine both the motivation and ability to lead. However, assessment centers are an expensive and timeconsuming process and it would be difficult to screen a large number of new candidates. Bartram (2004) noted that there is ample research indicating that personality attributes have important work implications (i.e., Barrick, Mount, & Judge, 2001; Borman, Penner, Allen, & Motowildo, 2001; Hermelin & Robertson, 2001). Personality testing may be used to prescreen entry-level candidates to select individuals with a strong desire to supervise. Subsequently, the assessment center could be used to measure the candidates’ ability to perform in supervisory roles.

Early Performance Ratings and Personality Measures

83

STUDY TWO The second study investigated whether specific personality characteristics (as measured by the California Psychological Inventory (CPI)) can be used to select candidates who desire to advance to management roles within their organization. Such tools as the CPI may be especially useful for screening large applicant pools. Organizations may be able to use these personality measures to quickly pre-screen candidates for motivation (not ability) to lead. Public sector organizations usually have to screen large applicant pools for entry-level positions. It would be especially useful to these organizations to be able to differentiate between candidates who will or will not fill future leadership roles within their organizations. These organizations can then base their selection decisions on the need for future leaders within their organization. Can current assessment tools, such as the CPI be used to screen large applicant pools for these desired qualities? Can we identify employees who are more proactive in seeking advancement? For our second study (see Juhasz & Mishken, 2005), we were interested in examining if the CPI can be used to identify entry-level law enforcement employees who choose to actively pursue a supervisory role.

The Use of the CPI in Law Enforcement Organizations The CPI has been extensively used to screen law enforcement officers. Selection of candidates for positions within a law enforcement type organization (e.g., law enforcement, court security, etc.) traditionally includes a system for eliminating candidates who possess personal characteristics that indicate the likelihood of serious job problems (Hargrave & Hiatt, 1989). Such psychological screening systems generally focus on eliminating individuals who are likely to cause problems, rather than predicting which individuals are likely to succeed on the job. CPI psychological screening is commonly used in law enforcement (Hargrave & Hiatt, 1987). It is an inventory designed to measure the personality of individuals who do not suffer from psychiatric disorders (Gough & Bradley, 1996). There have been a number of published investigations of the CPI within a law enforcement context. Such studies have examined the relationship between CPI scales and job performance (e.g., Hogan & Kurtines, 1975; Mills & Bohannon, 1980; Pugh, 1985; Hiatt & Hargrave, 1988; Hargrave & Hiatt, 1989), performance during training (e.g., Hargrave & Hiatt, 1987), or

84

MARK A. MISHKEN AND KRISZTINA JUHASZ

ratings of suitability for work as a law enforcement officer (e.g., Hogan & Kurtines, 1975; James, Campbell, & Lovegrove, 1984; Hargrave, Hiatt, & Gaffney, 1986; Hargrave & Hiatt, 1989). A few CPI studies have looked at the characteristics of effective managers (as noted by Hoffman & Davis, 1995) but have mainly concentrated on managerial performance and not on prediction of leadership desire. To date, the authors are only aware of one study examining why law enforcement officers decline to participate in the sergeants promotional process (Whetstone, 2001). However, the study did not examine personality characteristics. This study was designed to examine the extent to which CPI personality scales could be used to predict a desire for a leadership role in an organization. More specifically, we were interested in six CPI scales: Managerial Potential, Work Orientation, Leadership, Law Enforcement Orientation, Achievement via Independence, and Achievement via Conformance. It was hypothesized that the six CPI scales would account for a statistically significant portion of the variance in predicting which candidates applied for a newly created leadership position within the organization. The following scales were examined in more detail and specific hypotheses were formulated:

Managerial Potential Managerial Potential (Mp) is the part of the CPI special scales developed by Gough (1984) to identify individuals who seek managerial roles and who may be good in these leadership roles. Entrepreneurs, banking and business executives, and sales managers scored higher on the Managerial Potential scale, whereas art students and prison inmates scored lower (Gough & Bradley, 1996). Jacobs’ (1992) study found differences on the Managerial Potential scale between individuals who advanced to midmanagement versus those who did not (as noted in Gough & Bradley, 1996). Others found that employees who scored high on the Managerial Potential scale were described two-years later as having more managerial potential (Hoffman & Davis, 1995). In addition, Young, Arthur, and Finch (2000) found that Managerial Potential predicted ratings on three criteria of managerial performance better than other traits such as anxiousness and intelligence. We were interested in determining whether the Managerial Potential scale would be able to predict differences between candidates who applied to this newly created supervisory position versus those who did not.

Early Performance Ratings and Personality Measures

85

Hypothesis 1. Employees who apply to the supervisory position will score higher on the CPI Managerial Potential scale than those who do not apply to the position. Work Orientation Work Orientation (Wo) is part of Gough’s (1985) special CPI scales. It was constructed to identify individuals with strong work ethics and willingness to work. Gough & Bradley, 1996 notes that teachers and police officers scored higher on the work orientation scale and juvenile delinquents and psychiatric patients scored lower. Individuals who score high on the scale are described as ‘‘dependable, moderate, organized, and reasonable’’ (Gough & Bradley, 1996) Low scorers are described as ‘‘careless, changeable, distractible, reckless, and temperamental’’ (Gough & Bradley, 1996). Hoffman and Davis (1995) found that candidates’ Work Orientation scores correlated with personnel directors’ ratings two-years later. Therefore, we expected that individuals with stronger work ethics, as measured by the Work Orientation scale, would desire to take on more responsibility and to advance in the organization. Hypothesis 2. Employees who apply for the supervisory position will score higher on the Work Orientation scale. Leadership Potential Leadership (Lp) is part of the specialized scales developed by Gough. There has not been much research on the leadership scale. However, as Gough and Bradley (1996) notes, those who scored high on the scale include MBA candidates, sales managers, and military officers. Those scoring lower include high school disciplinary problems, psychiatric patients, and art students. High scorers are described as ‘‘alert, ambitious, energetic, poised, and resourceful’’ and low scorers are described as ‘‘awkward, dissatisfied, immature, quitting, and timid’’ (Gough & Bradley, 1996). As Steers, Porter, and Bigley (1996) noted, one of the key characteristics of leaders is ambition. Therefore, we expected the leadership scale to distinguish between candidates who pursued the supervisory positions versus those who did not. Hypothesis 3. Employees who pursue the supervisory position will score higher on Leadership Potential.

86

MARK A. MISHKEN AND KRISZTINA JUHASZ

Law Enforcement Orientation Law Enforcement Orientation (Leo) is part of Gough’s special CPI scales and is used to distinguish between people who want to pursue a career in law enforcement from other populations. The Law Enforcement Orientation scale has been tested on several groups. The highest scorers were the probation officers and the lowest scorers included psychology graduate students, social graduate students, and mathematicians. Description of highscoring individuals includes ‘‘confident, conventional, healthy, organized, and practical’’ and low-scoring individuals are described as ‘‘absent-minded, artistic, lazy, pessimistic, and worrying’’ (Gough & Bradley, 1996, p. 175). For our study, we expected that the Law Enforcement Orientation scale would distinguish between those who desire advancement in the law enforcement field versus those who are less proactive about advancement. Hypothesis 4. Employees who apply for the supervisory position will score higher on the Law Enforcement Orientation scale than those who do not apply. Achievement via Independence versus Achievement via Conformance As noted earlier, one of the key characteristics of leaders is their high need for achievement (Bartram, 2004). The CPI folk scales distinguish between two types of achievement needs – Achievement via Independence and Achievement via Conformance. Achievement via Independence and Achievement via Conformance are part of the 20 main folk scales of the CPI. Because of the distinction between the two concepts, we expected these scales to differentiate between individuals who applied for the supervisory position versus those who did not. High scorers on the Achievement via Conformance scale are described as having a ‘‘strong drive to do well; like to work in settings where tasks and expectations are clearly defined; efficient and well organized’’ (Gough & Bradley, 1996, p. 13). Low scorers were described as having ‘‘difficulty in doing best work in settings that have strict rules and regulations; easily distracted; tends to stop working when things do not go well’’ (Gough & Bradley, 1996, p. 13). However, high scorers on the Achievement via Independence scale were described as having a ‘‘strong drive to do well; likes to work in settings that encourage freedom and individual initiative, clear thinking and intelligent’’ (Gough & Bradley, 1996, p. 13). Low scorers were described as having ‘‘difficulty in doing best work in settings that are vague,

Early Performance Ratings and Personality Measures

87

poorly defined and lacking in intellectual or cognitive endeavors’’ (Gough & Bradley, 1996, p. 13). Hypothesis 5. Employees who apply to the leadership position will score higher on Achievement via Independence and higher on Achievement via Conformance than those who do not apply.

METHOD Sample The sample consisted of 321 applicants who applied and were hired for the entry-level position of Bailiff (court security) in a state court system in 1996. These employees were followed for a period of eight years. Of the 321 court security personnel, 40 (12.5%) applied for a newly created leadership position as a lieutenant. Of the 40 applicants, 6 (15%) were women, 30 (75%) were men, and 4 (10%) were unknown. There was no significant difference between the groups with respect to length of time in the organization. Procedures As part of the application process for the entry position of Bailiff, all applicants completed Form 434 of the CPI to determine their psychological suitability to perform the job. In addition to the psychological evaluation, all candidates passed a full background investigation, a medical and a physical ability test. The newly hired employees were followed for approximately eight years after applying for the entry-level position. Their pre-hire CPI psychological profiles were examined to determine if there were any differences in those who later applied for the newly created supervisory position versus those who did not apply. The employees who applied to the newly created supervisory position of lieutenant had to go through an extensive application process that included obtaining letters of recommendation, essay writing, evaluation by a committee, and an interview. Measures California Psychological Inventory This version of the CPI consists of 434 items, which includes 20 folk scales and 13 special purpose scales (Gough & Bradley, 1996). This study used a

88

MARK A. MISHKEN AND KRISZTINA JUHASZ

database of CPI scores for applicants, which consisted of standardized scores (t-scores) based on total norms (not separated by gender) for the following scales: Managerial Potential, Work Orientation, Leadership, Law Enforcement Orientation, Achievement via Conformance, and Achievement via Independence.

RESULTS – STUDY TWO Tables 3 and 4 summarize the means, SD, and intercorrelations of the variables studied. All the variables were positively correlated with each other except Achievement via Independence and Law Enforcement Orientation (zero-correlation). t-Tests (one-tail) were conducted to determine whether subjects differed on the special purpose CPI scales (Managerial Potential, Work Orientation, Leadership, and Law Enforcement Orientation) and the two folk scales (Achievement via Independence and Achievement via Conformance). Hypothesis 1 was supported. Individuals who applied for the newly created supervisory position scored significantly higher on Managerial Potential than those who did not (69.6 vs. 64.8, po.01). Hypothesis 2 was also supported. Individuals who applied for the newly created supervisory position scored significantly higher on Work Orientation than those who did not (65.2 vs. 63.7, po.05). Hypothesis 3 was not supported. There was no statistically significant difference on the Leadership scale between individuals who applied to the supervisory position versus those who did not. Hypothesis 4 was also not supported. There was no statistically significant difference on the Law Enforcement Orientation scale between individuals who applied to the supervisory position versus those who did not. Hypothesis 5 was partially supported. Individuals who applied for the Table 3.

Study Two – Intercorrelations among the CPI Scales.

Variables 1. 2. 3. 4. 5. 6.

Managerial Potential Work Orientation Leadership Law Enforcement Orientation Achievement via Conformance Achievement via Independence

Note: po.01.

1

2

3

4

5

6

– .59 .72 .26 .52 .57

– – .49 .20 .48 .48

– – – .34 .58 .29

– – – – .28 .08

– – – – – .15

– – – – – –

89

Early Performance Ratings and Personality Measures

Table 4. Study Two – Means and SD of the CPI Scales. Variable Managerial Potential Work Orientation Leadership Law Enforcement Orientation Achievement via Conformance Achievement via Independence

Group

N

Mean

SD

A B A B A B A B A B A B

87 15 281 40 281 40 281 40 281 40 281 40

64.75 69.60 63.74 65.15 62.77 63.60 66.77 65.70 61.25 61.38 63.25 66.28

6.8 3.7 5.0 4.5 5.7 4.0 7.4 6.8 5.4 5.8 6.3 3.8

A, did not apply to the leadership position; B, applied to the leadership position.

newly created supervisory position scored significantly higher on Achievement via Independence than those who did not (66.3 vs. 63.3, po.01). There was no significant difference in the two groups on the Achievement via Conformance scale. A multiple regression analysis was conducted using the step-wise technique for entering variables to examine which of the six CPI scales predict whether an employee would apply for the supervisory position. The criterion for entry was set at .10 and the criteria for removal was .15. All of the above six CPI scales were entered into the regression equation. Only Work Orientation was a significant predictor of candidates’ decision to apply for the supervisory position (R ¼ .33, po.001).

DISCUSSION Overall, the results indicate there are significant personality differences between individuals who pursue supervisory roles versus those who do not. We found support for our hypothesis that those individuals who score higher on the Managerial Potential scale, the Work Orientation scale, and the Achievement via Independence scales are more likely to apply for the supervisory position than those who scored lower. That is, those individuals who applied to the supervisory position have a strong desire to lead, are conscientious, dependable and reasonable, and prefer to work in settings where they have autonomy. However, we did not find significant differences

90

MARK A. MISHKEN AND KRISZTINA JUHASZ

between the groups on Leadership, Law Enforcement Orientation, and the Achievement via Conformance scales. Achievement via Conformance and Achievement via Independence were significantly correlated. However, the correlation was small (r ¼ .15) (see Table 3). Need for Achievement via Independence may be a better measure of Need for Achievement commonly used to predict managerial behaviors (Ghiselli, 1971; Steers et al., 1996). Although the two groups differed on Achievement via Independence, Managerial Potential and Work Orientation, only the Work Orientation scale significantly predicted which employees pursued the supervisory role. It should be noted that the Work Orientation scale was significantly correlated with Managerial Potential (.59) and Achievement via Independence (.48), which may explain why only the Work Orientation scale differentiated between the two groups in the regression equation. The Work Orientation scale may measure a different aspect of leadership motivation than the other two scales. With the exception of the Law Enforcement Orientation scale, all five of the scales were highly intercorrelated (see Table 3). Overall, the results indicate that personality screening may be a useful tool to distinguish between applicants who are interested in taking on leadership roles versus those who are not. However, there were some limitations to the study. There was likely a restriction in range due to the candidates being pre-selected based partly on their CPI scores and the psychological screening process. One of the other limitations of the study was that we did not have the CPI profiles of all the candidates who applied for the supervisory position. We may have obtained significant relationships for the other variable and may have explained a greater amount of variance in the decision to apply to a supervisory position. Future research may want to explore these relationships on a larger sample. It would also be interesting to survey the entry-level candidates at the time of application about their desire for advancement and compare it to the results obtained with the CPI. In addition, future studies should explore the personality differences between those who applied for a supervisory position and were rejected from those who were accepted.

OVERALL CONCLUSION The first study found that performance appraisals administered at an earlier time period could predict future progression in an organization. It also supported the notion that there was differential prediction of progression depending upon the class or type of position.

Early Performance Ratings and Personality Measures

91

The results of the second study provide further evidence that motivation to perform a supervisory role can be measured by personality measures such as the CPI. Organizations may benefit by pre-screening applicants in order to select candidates with a strong motivation to lead. More expensive and time-consuming selection tools can then be used later on in the process to select candidates with the ability to lead. Overall, the findings from these two studies demonstrate that organizations can identify individuals likely to advance in their organizations. Such tools as early performance appraisal ratings and personality measures have the potential to be valuable in making promotional decisions.

REFERENCES Barrick, M. R., Mount, M. K., & Judge, T. A. (2001). Personality and performance at the beginning of the new millennium: What do we know and where do we go next? International Journal of Selection and Assessment, 9(1–2), 9–30. Bartram, D. (2004). Assessment in organizations. Applied Psychology: An International Review, 53(2), 237–259. Bartram, D. (2005). The great eight competencies: A criterion-centric approach to validation. Journal of Applied Psychology, 90(6), 1185–1203. Bobko, P. (2001). Correlation and Regression: Applications for industrial and organizational psychology and management. Thousand Oaks, CA: Sage Publications, Inc. Borman, W. C., Penner, L. A., Allen, T. D., & Motowildo, S. J. (2001). Personality predictors of citizenship performance. International Journal of Selection and Assessment, 9, 52–69. Ferris, G. L., Witt, L. A., & Hochwarter, W. A. (2001). Interaction of social skill and general mental ability on job performance and salary. Journal of Applied Psychology, 86(6), 1075–1082. Fletcher, C. (1986). The effects of performance review in appraisal: Evidence and implications. The Journal of Management Development, 5(4), 3–12. Fletcher, C. (2001). Performance appraisal and management: The developing research agenda. Journal of Occupational and Organizational Psychology, 74(4), 473–487. Garman, A. N., & Glawe, J. (2004). Succession planning. Consulting Psychology Journal: Practice and Research, 56(2), 119–128. Gerhart, B., & Milkovich, G. T. (1990). Organizational differences in managerial compensation and financial performance. Academy of Management Journal, 33(4), 663–691. Ghiselli, E. E. (1971). Explorations in managerial talent. Pacific Palisades, CA: Goodyear Publishing Company. Gough, H. G. (1984). A managerial potential scale for the California psychological inventory. Journal of Applied Psychology, 69, 233–240. Gough, H. G. (1985). A Work Orientation scale for the California psychological inventory. Journal of Applied Psychology, 70, 505–513. Gough, H. G., & Bradley, P. (1996). CPI Manual (3rd ed). Palo Alto, CA: Consulting Psychologists Press.

92

MARK A. MISHKEN AND KRISZTINA JUHASZ

Grote, D. (2000). Public sector organizations: Today’s innovative leaders in performance management. Public Personnel Management, 29(1), 1–19. Hargrave, G. E., & Hiatt, D. (1987). Law enforcement selection with the interview, MMPI, and CPI: A study of reliability and validity. Journal of Police Science and Administration, 15(2), 110–117. Hargrave, G. E., & Hiatt, D. (1989). Use of the California psychological inventory in law enforcement officer selection. Journal of Personality Assessment, 53(2), 267–277. Hargrave, G. E., Hiatt, D., & Gaffney, T. W. (1986). A comparison of MMPI and CPI test profiles for traffic officers and deputy sheriffs. Journal of Police Science and Administration, 14(3), 250–258. Harris, M. M., Gilbreath, B., & Sunday, J. A. (1998). A longitudinal examination of a merit pay system: Relationships among performance ratings, merit increases, and total pay increases. Journal of Applied Psychology, 83(5), 825–831. Hermelin, E., & Robertson, I. T. (2001). A critique and standardization of meta-analytic validity coefficients in personnel selection. Journal of Occupational an Organizational Psychology, 74, 253–277. Hiatt, D., & Hargrave, G. E. (1988). Predicting job performance problems with psychological screening. Journal of Police Science and Administration, 16(20), 122–125. Hoffman, R. G., & Davis, G. L. (1995). Prospective validity study: CPI Work Orientation and Managerial Potential scales. Educational and Psychological Measurement, 55, 881–890. Hogan, R., & Kurtines, W. (1975). Personological correlates of police effectiveness. The Journal of Psychology, 91, 289–295. Howard, A., & Bray, D. W. (1988). Managerial lives in transition. New York: Guilford Press. Hunter, J. E., & Hunter, R. F. (1984). Validity and utility of alternative predictors of job performance. Psychological Bulletin, 96, 72–98. Jacobs R. L. (1992). Moving up the corporate ladder: A longitudinal study of motivation, personality, and managerial success in women and men. Unpublished doctoral dissertation, Boston University; James, S. P., Campbell, I. M., & Lovegrove, S. A. (1992). Personality differentiation in a police-selection interview. Journal of Applied Psychology, 69(1), 129–134 James, S. P., Campbell, I. M., & Lovegrove, S. A. (1984). Personality differentiation in a policeselection interview. Journal of Applied Psychology, 69(1), 129–134. Jelley, R. B., & Goffin, R. D. (2001). Can performance-feedback accuracy be improved? Effects of rater priming and rating-scale format on rating accuracy. Journal of Applied Psychology, 86(1), 134–144. Juhasz, K. & Mishken, M.A. (2005). Using the CPI to predict employees’ management-seeking behaviors. Paper presented at the 113th annual meeting of the American Psychological Association, August. Washington, DC. Kane, J. S., & Bernardine, H. J. (1982). Behavioral observation scales and the evaluation of performance appraisal effectiveness. Personnel Psychology, 35(3), 635–641. Kirkpatrick, S. A., & Locke, E. A. (1991). Leadership: Do traits matter? Academy of Management Executive, 5(2), 48–60. Latham, G. P., & Mann, S. (2006). Advances in the science of performance appraisal: Implications for practice. International Review of Industrial and Organizational Psychology, 21, 295–337. Latham, G. P., & Wexley, K. N. (1977). Behavioral observation scales for performance appraisal purposes. Personnel Psychology, 30, 255–268.

Early Performance Ratings and Personality Measures

93

Latham, G. P., & Wexley, K. N. (1981). Increasing productivity through performance appraisal. Reading, MA: Addison-Wesley. Martin, D. C., Bartol, K. M., & Kehoe, P. E. (2000). The legal ramifications of performance appraisal: The growing significance. Public Personnel Management, 29(3), 381–409. Metz, I. (2004). Do personality traits indirectly affect women’s advancement? Journal of Managerial Psychology, 19(7), 695–707. Mills, C. J., & Bohannon, W. E. (1980). Personality characteristics of effective state police officers. Journal of Applied Psychology, 65(6), 680–684. Mishken, M. A., Ruminson, K. C., & Juhasz, K. (2005). Performance appraisal field study: The use of behavioral observation scales. Paper presented at the European Academy of Management Conference, May. Munich, Germany. Mishken, M. A., Ruminson, K. C., & Juhasz, K. (2006). Findings from the public sector: Performance ratings and promotional progression. Paper presented at the SIOP national conference, April. Dallas, Texas. Pugh, G. (1985). The California psychological inventory and police selection. Journal of Police Science and Administration, 13(2), 172–177. Salgado, J. F., Anderson, N., Moscoso, S., Bertua, C., & Fruyt, F. D. (2003a). International validity generalization of GMA and cognitive abilities: A European community metaanalysis. Personnel Psychology, 56(3), 573–605. Salgado, J. F., Anderson, N., Moscoso, S., Bertua, C., Fruyt, F. D., & Rolland (2003b). A meta-analytic study of general mental ability validity for different occupations in the European community. Journal of Applied Psychology, 88(6), 1068–1081. Southworth, D. (2000). Using job performance as a component of civil service examinations. Public Personnel Management, 29(3), 410–418. Spector, P. (2003). Industrial organizational psychology: Research and practice. New York, NY: Wiley Publishing. Steers, R. M., Porter, L. W., & Bigley, G. A. (1996). Motivation and leadership at work. Boston, MA: McGraw Hill. Sturman, M. C., Cheramie, R. A., & Cashen, L. H. (2005). The impact of job complexity and performance measurement on the temporal consistency, stability, and test–retest reliability of employee job performance ratings. Journal of Applied Psychology, 90, 269–283. Tharenou, P. (2001). Academy of Management Journal, 44(5), 1005–1017. Tziner, A., Kopelman, R., & Joanis, C. (1997). Investigation of raters’ and ratees’ reactions to three methods of performance appraisal: BOS, BARS, and GRS. Canadian Journal of Administrative Sciences, 14(4), 396–404. Viswesvaran, C., Ones, D. S., & Schmidt, F. L. (1996). Comparative analysis of the reliability of job performance ratings. Journal of Applied Psychology, 81, 557–574. Viswesvaran, C., Schmidt, F. L., & Ones, D. S. (2005). Is there a general factor in ratings of job performance? A meta-analytic framework for disentangling substantive and error influences. Journal of Applied Psychology, 90, 108–131. Vroom, V. H. (1964). Work and motivation. New York, NY: John Wiley & Sons. Whetstone, T. S. (2001). Copping out: Why police officers decline to participate in the sergeant’s promotional process. American Journal of Criminal Justice, 25(2), 147–159. Wiersma, U., & Latham, G. P. (1986). The practicality of behavioral observation scales, behavioral expectation scales, and trait scales. Personnel Psychology, 39(3), 619–628. Young, B. S., Arthur, W., & Finch, J. (2000). Predictors of managerial performance: More than cognitive ability. Journal of Business and Psychology, 15(1), 53–72.

This page intentionally left blank

TYING VS. NOT-TYING POST-COMPLETION REVIEWS TO EXTRINSIC REWARDS Stefan Linder ABSTRACT The literature on post-completion reviews (PCRs) either does not deal with the tying of PCRs to extrinsic rewards or provides scant theoretical reasoning or empirical analysis to back up its recommendations. Based on research from psychology and empirical studies, the present chapter proposes that several effects of a PCR, which must be deemed rather dysfunctional, will increase when extrinsic rewards are linked to such a review. At the same time some possibly functional effects, however, are likely to remain constant. The propositions, therefore, call the usefulness of tying PCRs to rewards into question and call for further investigation.

1. INTRODUCTION In contrast to the very profound research interest that seeks methods to evaluate proposed capital expenditure projects, post-completion reviews (PCRs) or post-completion audits of capital expenditures have received little

Performance Measurement and Management Control: Measuring and Rewarding Performance Studies in Managerial and Financial Accounting, Volume 18, 95–125 Copyright r 2008 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-3512/doi:10.1016/S1479-3512(08)18005-4

95

96

STEFAN LINDER

attention (Charreaux, 2001). This holds particularly true for the problem of linking vs. not linking such reviews to extrinsic rewards and/or to punishments for the person who has been subjected to a PCR. Even though Helfert (1960), Istvan (1961), and Nicholson (1962) have already raised this issue, Anglo-American literature, just like the French and German literature, rarely deals with tying PCRs to rewards. Furthermore, except for Ha¨gg (1977), the existing literature does not provide a sound theoretical rationale for its recommendations with respect to tying vs. nottying. Not surprisingly then, empirical studies about whether PCRs are linked to rewards and/or punishments in practice and the respective reasons for doing so show a significant variance and confusion. This chapter wants to shed more light on the question of tying vs. nottying PCRs to extrinsic rewards and/or punishments by deriving propositions about the impact of such a tying on the effects of PCRs on the people involved. For the sake of simplicity, the analysis will be limited to a twoperson setting: a person whose capital expenditure project has been subjected to a PCR and who will receive the rewards and/or punishments (an agent or subordinate) and an individual (a controller, a superior, or a principal) in charge of conducting the PCR and of distributing the rewards and/or punishments. In order to arrive at some propositions that have the potential to mirror reality very closely and that lend themselves to empirical testing, concepts, models, and empirical evidence from social-cognitive psychology will be the bases of the analysis. In contrast to many other chapters, the present one will focus on the development of propositions, that is, hypotheses; it does not include an empirical study that tests propositions. The chapter thereby follows Dubin’s (1978) and Popper’s (1980) clear distinction between developing and testing hypotheses. It devotes considerably more space to hypothesis development than would a chapter that combines development and testing. Given the state of research on tying or not-tying PCRs to extrinsic incentives, the analysis should therefore profit from such a separation. The chapter is structured as follows: First, the existing literature on tying of PCRs to rewards and/or punishments will be critically reviewed. Then the methodology of this chapter will be described in more detail. In addition, the social-cognitive foundations for an analysis of the impact of tying PCRs to rewards will be laid. The effects PCRs have on the people involved will then be analyzed. Based on the structuring and presentation of these effects, the impact of tying rewards and/or punishments has on these effects shall then be investigated in detail and propositions will be established. Next, the

Tying vs. Not-Tying Post-Completion Reviews to Extrinsic Rewards

97

propositions derived will be discussed. Finally, a conclusion with respect to future research will be drawn.

2. STATE OF PCR RESEARCH ON TYING VS. NOT-TYING Most of the research on capital expenditure management and capital budgeting has dealt with developing and discussing methods to evaluate proposed capital expenditure projects (like Net Present Value, Internal Rate of Return, or the real options approach), and measuring the practical usage of these methods. In fact, ‘‘probably more surveys have been undertaken on the use of capital budgeting techniques than on any other accounting and finance topic’’ (Drury & Tayles, 1996, p. 371). In comparison to this significant research interest targeted at methods for evaluating capital expenditure projects, other parts of the capital expenditure process – and PCRs in particular – have received little attention (Charreaux, 2001). PCRs are those actions in the capital expenditure process in which actual values of a certain capital expenditure project are compared to their planned values, and in which the reasons for potential discrepancies between the two sets of values are analyzed. They have been discussed occasionally in the literature since the late 1950s under several labels – including post-completion audits, post-audits, postmortems, and postimplementation reviews (Mills & Kennedy, 1990; Neale, 1989). The main reason for conducting such reviews, according to conceptual works and in the overwhelming majority of empirical studies from Canada, Germany, Italy, the U.K. and the U.S., is to foster knowledge creation for future capital expenditure projects (to avoid the repetition of mistakes) and to solve problems with current projects (Epstein & Rejc, 2005; Mills & Kennedy, 1990; Neale, 1989; Riggs, Bedworth, & Randhawa, 1998; Segelod, 1995; Soares, Coutinho, & Martins, 2007). Apart from this benefit with respect to knowledge, researchers and practitioners alike recommend PCRs as a means of getting employees responsible for the capital expenditure projects to work harder and to make more valid estimates (see Linder, 2006 for a review). In the limited stream of literature on PCRs, Helfert (1960), Istvan (1961), and Nicholson (1962) have already pointed to the question of linking or not linking PCRs to extrinsic rewards (like bonuses or promotions) and punishments (like public criticism). Istvan’s (1961) study of large North

98

STEFAN LINDER

American corporations, for example, shows, that in some of them ‘‘employees were reprimanded for repeated errors’’ (Istvan, 1961, p. 43) undisclosed by PCRs. Similarly, Nicholson (1962) shows that this linking of rewards and punishments to PCRs can span the continuum from insignificant rewards and punishments to the loss of a job. Surprisingly, however, later works on PCRs mainly do not deal with tying PCRs to incentives. This holds as true for empirical research as for theoretical works and even more practical ‘‘how-to’’ publications. Furthermore, the few empirical works do not look at the influence such a tying or not-tying has. They either only point to a considerable confusion and disagreement among practitioners about whether to link PCRs to rewards/punishments, or to similarly diverging degrees of companies actually tying their PCRs to rewards and punishments (see, e.g. De Bodt & Bouquin, 2001; Posey, Roth, & Dittrich, 1985; Saatmann, 1970). The notable exception is Ha¨gg (1977). He tries to answer the question of tying vs. not tying of rewards by analyzing the impact of such a linking on the effects of PCRs on the abilities and learning of the people involved. In contrast to the claims of PCR literature, his study does not find a statistically significant influence of tying (Ha¨gg, 1977). Since his study is based on a very small number of firms, one could, however, attribute this result to a sampling problem. Hence, the question of whether or not tying PCRs to rewards (and/or punishments) impacts learning from PCRs still awaits clarification. But just as empirical studies on PCRs do not answer the question of tying vs. not-tying, neither do theoretically- or conceptually-oriented works. While the majority of this literature stresses that tying does have a positive impact on the magnitude of learning (i.e. knowledge/ability development) of the people subjected to a PCR (e.g. Dillon & Caldwell, 1981), the theoretical foundations for these postulations are not truly satisfying. The same holds true for a dysfunctional increase of a PCR’s disciplining effect due to linking PCRs to rewards and punishments that has been postulated by (among others) Azzone and Maccarrone (2001), Borer (1978), and Nicholson (1962) and that should lead to an ‘‘over-disciplination’’ in the form of an ‘‘adherence by the letter’’ phenomenon. Once again, Ha¨gg (1977) makes the first serious attempt to give such ideas a sound theoretical basis. His work is on the one hand inspired by the behavioral research of Dalton and Lawrence (1971), Bruns and DeCoster (1969), and Vroom (1964a). On the other hand, he uses economic research about budgets and budgetary slacks. Based on these foundations, he proposes that the magnitude of change in behavior caused by PCRs depends on the perceived rewards (Ha¨gg, 1977). However, his focus is limited to the impacts already postulated in literature

Tying vs. Not-Tying Post-Completion Reviews to Extrinsic Rewards

99

and in particular on whether the results of a PCR are used by managers to improve future capital expenditure plans or simply as a sign of progressive management. Thus, his analysis is primarily behaviorist. Secondarily, it does not differentiate the incentives linked to a particular PCR from the rewards or punishments tied to future actions. Overall, he makes a very important first step at exploring PCRs in a causal way but does not provide a truly satisfactory analysis of all likely effects of PCRs and their interplay with a tying of PCRs to rewards and/or punishments from a social-cognitive perspective. Such an analysis would first require an examination of the general effects of PCRs. This would allow an understanding of the entire field when linking PCRs to rewards. It would therefore be useful to circumvent the danger of overlooking potentially important effects of PCRs (and their interaction with a tying of rewards) that the literature has not yet discussed. However, such a compilation of the conceivable effects is still missing from the PCR research. Overall, the research on the impact of tying vs. not-tying PCRs to rewards and punishments, hence, is still very limited.

3. METHODOLOGY Following Dubin (1978) and Popper (1972, 1980), philosophy of science typically distinguishes two research objectives: (1) proposition or hypothesis development and (2) hypothesis testing. The purpose of hypothesis development is exploration and tentative knowledge generation. Such creative processes of exploring the real world are neither capable of following a strict scientific process of discovery nor do they require such a process from a philosophical point of view (Popper, 1980). However, philosophers of science point out that any such hypotheses need to be critically tested before they may be used for practical recommendations (ibid.). In contrast to hypothesis development, testing is intended to secure our knowledge of real world phenomena by critically checking the falsity of the hypotheses (Albert, 1984; Popper, 1980). It is a clearly structured process and leads to a ranking of hypotheses about a certain phenomenon based on their degree of non-falsification (Popper, 1972). The hypotheses about a certain phenomenon that fully resisted the test of falsification (or at least did so to a higher degree than potential other hypotheses about that phenomenon) are those that can become part of more elaborate theories and guidelines for practitioners.

100

STEFAN LINDER

Both research objectives are interlinked and equally important (Popper, 1972): Hypothesis development strives to provide a pool of potentially true explanations that lend themselves to hypothesis testing. The latter needs a pool of hypotheses to be tested as a prerequisite, and at the same time, ensures through falsification of some hypothesis that the pool of potential explanations does not become unwieldy. In managerial research, however, hypothesis testing is much more common. Both Dubin (1978) and Popper (1980) call for a clear separation of these two objectives. This allows avoiding ‘‘ad hoc’’ hypotheses that can be found in some parts of economic research just as much as circumventing the ever-looming danger of quickly adapting one’s hypotheses to the results of one’s empirical investigation without an in-depth analysis of whether this is really necessary (i.e. whether the error lies with the hypotheses or with the empirical study). Furthermore, such a separation frees up considerably more space for hypothesis development than a chapter combining development and testing would allow. This can be very important when there is relatively little research about a phenomenon. In contrast to the overwhelming number of publications on the topic, the present one will concentrate on the development of hypotheses and propositions. It, consequently, does not include an empirical study testing the propositions that are presented in the following paragraphs. Given the lack of research on tying PCRs to extrinsic incentives and, hence, the resulting space required to arrive at some first propositions, the analysis should profit from such a separation. In order to arrive at hypotheses, most behavioral managerial research falls into one of two streams (Ewert & Wagenhofer, 2003): It either makes profound use of economic models like principal-agent theory or game theoretic approaches (e.g. Holmstro¨m & Milgrom, 1991; Jensen & Meckling, 1976), or relies on theories and empirical studies from the behavioral sciences – particularly from psychology (e.g. Fedor, 1991; Ilgen & Feldman, 1983; Simon, 1957). Each has its advantages and disadvantages. Principal-agent or game theoretic models stand out by their precision, resulting from their mathematical formulation. Theories and studies from the behavioral sciences – and especially from psychology – offer the advantage of coming a bit closer to human behavior, since they do not require as tightly set premises as do principal-agent or game theoretic models. This should facilitate empirical testing (and subsequent derivation of practical guidelines). Furthermore, analyses based on psychological research allow for the easy inclusion of aspects of human decision processes

Tying vs. Not-Tying Post-Completion Reviews to Extrinsic Rewards

101

and behaviors that do not necessarily fit the more rational principal-agent and game theoretic approaches. Since the present chapter concentrates on developing hypotheses, the closeness to reality and the ease of future empirical testing of the hypotheses generated must be seen as very important. For this reason the chapter will rely on psychological research. In particular, it will use concepts from the dominant social-cognitive perspective in psychology. It will, however, occasionally draw links to results from principal-agent models in order to complement the perspective and ensure proper inclusion of available knowledge on PCRs and reviews in general.

4. SOCIAL-COGNITIVE FOUNDATIONS OF ANALYSIS The development of the initial capital expenditure plan, its implementation and the following PCR are all carried out by human beings. Hence, an analysis of PCRs has to consider these individuals and their characteristics. These characteristics will determine how the PCR takes place and will, at the same time, be influenced by a PCR (e.g. through learning or the satisfaction (or not) of needs). Since the inception of modern psychology one hundred years ago, psychologists have strived to identify and describe the characteristics or factors that influence a person’s cognitive and physical actions. While there are still competing definitions and theories, most psychologists would agree, that these characteristics can be grouped under four headings: cognitive and physical abilities, expectations, needs, and attitudes. A person’s behavior clearly is influenced by his or her abilities (i.e. potential) (Ha¨gg, 1977; Porter & Lawler, 1968 ). The literature differentiates cognitive abilities or knowledge from physical abilities (George & Jones, 2005). Cognitive abilities or knowledge comprise what an individual knows (know that) and can do (know how) (Ryle, 1949). Physical abilities, in contrast, refer to aspects like physical strength, endurance, flexibility, and balance (Daft & Noe, 2001). Some authors differentiate abilities into categories depending on whether the individual acquired them by training or by experience (e.g. Ha¨gg, 1977). The majority, however, do not draw this distinction. Individuals differ in the magnitude of their knowledge, but cognitive abilities are generally limited in

102

STEFAN LINDER

comparison to what the environment would ask for (e.g. Porter & Lawler, 1968). This bounded knowledge creates uncertainty in the individual about the environment, its elements, and their interaction (March & Olsen, 1975; March & Simon, 1958). This in turn, can cause the individual to make erroneous decisions about what to do in a situation. Fortunately, however, a person’s knowledge is not an entirely given or a fixed factor, but can be developed through interaction with the environment (George & Jones, 2005; McDermott, 1999). Such a change in knowledge (which can but does not have to result in a change in behavior) is typically referred to as learning (George & Jones, 2005). Expectations as the second determinant of human behavior can be characterized as an individual’s assumptions about future events (e.g. the behavior of another person) and their consequences for that individual’s needs (George & Jones, 2005; Porter & Lawler, 1968). They are probability estimates about, firstly, the impact of a person’s actions and efforts on the results obtained by these actions and, secondly, the consequences (i.e. rewards and punishments) that are linked to these results. The first are ‘‘selfefficacy’’ perceptions or ‘‘effort-performance’’ (E-P) expectations, while the second can be termed ‘‘instrumentalities’’ or ‘‘performance/results– reward’’ (P-R) expectancies (Bandura, 1997; Porter & Lawler, 1968; Rotter, 1954; Vroom, 1964a). Needs are a person’s latent preparedness to react in a particular way to specific incentives: rewards or punishments (George & Jones, 2005; Krech, Crutchfield, & Ballachey, 1962). Fulfilling one’s needs is associated with positive feelings and ‘‘satisfaction’’; frustration of those needs is associated with negative feelings and dissatisfaction. Typically, an individual will behave in a utilitarian or hedonistic way (i.e. the person will strive to optimize the balance of good over bad feelings and choose her actions accordingly) (e.g. Galbraith & Cummings, 1967). Researchers still disagree over the exact list of needs and their interaction. The theories advocated by Alderfer (1972), Maslow (1954), or Yukl (1990), for example, include physiological needs (e.g. for food), safety needs (including a need for remuneration), needs for affiliation and friendship, needs for esteem and power, and needs for achievement and self-fulfillment. Except for the physiological needs and the safety needs, this list is the same as McClelland, Atkinson, Clark, and Lowell’s (1953) well-known motivation theory. Thus, relying on this list of needs for further analysis is reasonable given the current state of psychological research. While not a proponent of the social-cognitive perspective, Skinner (1965) developed a definition of rewards and punishments that lends itself to

Tying vs. Not-Tying Post-Completion Reviews to Extrinsic Rewards

103

cognitive analyses as much as to behaviorist ones. Following Skinner (1965), rewards can be defined as all those tangible or intangible liked objects that, when given or available to a person, are able to satisfy his or her currently active needs (Type I reward), or objects disliked that, when taken from the person or lost by it, are able to increase the current level of need satisfaction (Type II reward). Type I rewards can take the form of monetary compensation, increase in job security, promotions, autonomy, and social recognition (e.g. Merchant & Van der Stede, 2003). Similarly, any disliked tangible or intangible object that is allocated to a person shall be called a punishment (Type I punishment) as well as and any liked object which is taken away from the person (Type II punishment) (Skinner, 1965). Rewards and punishments need to fit the currently active needs, that is, be in line with the level of satisfaction of the respective needs in order to be ‘‘relevant’’ to the individual (see, e.g. Lawler, 1976; Melton, 1973). Following the respective underlying needs, a reward or punishment can be classified as intrinsic or extrinsic (Ukko, Karhu, & Pekkola, 2007; Vroom, 1964a). While literature considers physiological needs, safety needs, needs for affiliation and friendship, and needs for esteem/power to be needs that can be extrinsically satisfied or frustrated (i.e. from outside the individual), the needs for achievement/self-fulfillment can only be satisfied (or frustrated) from within the individual (by satisfaction with one’s performance on a specific job) (e.g. Alderfer, 1969; Alderfer, 1972). Therefore, they can be called intrinsic needs and are sources of the so-called intrinsic motivation, which has attracted growing research attention since the seminal works by Deci (1971, 1972). Much of this research has been targeted at the interaction of extrinsic and intrinsic needs (and their corresponding rewards). Some authors have proposed that there is no such interaction, others argue in favor of a positive relationship, and still others for a negative one – some form of ‘‘crowding-out’’ (see, e.g. Alderfer, 1972; Amabile, Hill, Hennessey, & Tighe, 1994; Deci, 1971; Deci, 1972; Frey, 1997; Lao, 1981; Vroom, 1964a). Even though researchers have not found a unanimously accepted answer to this question yet, there is a tendency away from extreme positions – that is, the view is gaining ground that such an interaction may take place, but probably only in some specific and rather rare cases (see, e.g. Cameron & Pierce, 2002; Kunz & Pfaff, 2002; Reinholt, 2006). Following this view and in line with Alderfer’s (1969, 1972) motivation theory, it can be assumed that the intrinsic needs for achievement will always play a role

104

STEFAN LINDER

in an individual’s decisions about how to act in a certain situation and that these needs, in contrast to all extrinsic needs, can never be fully satisfied. Hence, except for a few specific cases, intrinsic needs play an important role, while extrinsic motivation depends on the relevance of the extrinsic rewards and punishments provided, that is their fit with the current level of satisfaction of an individual’s extrinsic needs (see, e.g. Lawler, 1976; Melton, 1973). Extrinsic motivation, therefore, may not be important when there are no extrinsic rewards or punishments that are capable of influencing a person’s current degree of need satisfaction, either because these extrinsic needs are satisfied (i.e. making rewards irrelevant) or because they are so frustrated that they cannot be frustrated any further (i.e. making punishments irrelevant). Since more than one need requires external satisfaction, extrinsic rewards and punishments in practice can take a wide range of forms, including non-material awards (Frey, 2007). Additionally, a single extrinsic reward may satisfy several of the extrinsic needs just described. Some awards, for example, may imply social recognition and social affiliation to a group of people or social recognition and security by the accompanying monetary or career implications (Foss, 2007). The influence of rewards and punishments on human motivation will, however, depend on their fit with the currently active needs. For example, non-material awards or prizes indicate social recognition, and hence, may be a strong incentive for someone whose needs for esteem/power are less satisfied than her other needs but not be an incentive for someone whose other needs more strongly crave for satisfaction. Attitudes, which in literature are sometimes mistaken for needs and expectations, are evaluations of certain objects (e.g. actions, types or outcomes of behavior) with respect to the satisfaction or frustration of a person’s needs (Ajzen & Fishbein, 1980; Ajzen, 1988; Crawford, Luka, & Cacioppo, 2002; Krech et al., 1962). If an object is experienced or considered useful in satisfying the individual’s needs, a positive attitude towards the object will result; if the person experiences the object as useless or even counterproductive to satisfying his needs, a negative attitude will be learnt (Crawford et al., 2002). Based on this description of the factors determining human behavior and the assumption that a satisfied person will be more productive and, therefore, more desirable for a company than a dissatisfied one, it is possible to identify and describe the possible effects of a PCR on the (presumably two) people involved in it: the agent and the superior.

Tying vs. Not-Tying Post-Completion Reviews to Extrinsic Rewards

105

5. EFFECTS OF PCRS In order to understand the impact of linking PCRs to extrinsic rewards and punishments, an analysis of the potential effects of a PCR on the individuals involved (and their characteristics) can be a good basis. It will allow discussion of the changes in these effects when PCRs are linked to extrinsic incentives. Although PCRs take place ‘‘after the fact’’ (i.e. after the respective action of capital expenditure), they can have ex-ante and ex-post effects. The exante effects are caused by the existence of a system for conducting PCRs: by employees believing that the results of their actions will, or at least could be checked later. The ex-post effects, however, are caused by the knowledge of these results gained in the course of a PCR.

5.1. Ex-Ante Effects Ex-ante distrust effect. Announcing a PCR or introducing a system for conducting PCRs can give the individual occupied with capital expenditure planning or project implementation the impression that the firm or the superior considers him unwilling to do his job as expected (Barkema, 1995; Frey, 1993; Manzoni & Barsoux, 2005 for control systems). If the agent interprets the principal’s announcement or decision to implement PCRs as a signal that she considers the agent unwilling to do a good job, then the agent’s need for affiliation will be hurt. The results of a laboratory experiment by Falk and Kosfeld (2006) for control systems point to the creation of such dissatisfaction and bad feeling. Following Alderfer’s (1972) theory of motivation, and compatible with the results Cohen (1965) obtained in his study on control systems, this frustration of the need for affiliation will lead to a regression of the agent to lower-order safety and monetary needs. At the same time, the reduced affiliation between the agent and the superior, stemming from the principal’s distrust of the agent’s willingness to perform well, increases the marginal net returns of opportunistic behavior for the agent (Frey, 1993 for controls). Before the PCR was announced, marginal returns of opportunistic behavior were low because of high emotional ‘‘costs’’ (i.e. negative consequences on his needs for affiliation) associated with behaving opportunistically. However, introducing PCRs that are linked to a breakdown of the former relationship and affiliation between principal and agent reduces these emotional costs.

106

STEFAN LINDER

Consequently, marginal net returns of opportunistic behavior increase and may induce the agent to pursue his other needs more opportunistically. This crowding-out of trust by control can, in fact, create a self-enforcing spiral of distrust between the principal and the agent (Frey, 1993). The magnitude of the effect will depend, among other things, on the personal affiliation between the agent and the principal that existed prior to the principal’s announcement of PCRs: the more intense this affiliation was, the stronger the dysfunctional effect is likely to be (see Barkema, 1995 for control systems). This effect, therefore, can be called a ‘‘distrust effect of PCRs.’’ Ex-ante alienation effect. If the pure existence or announcement of PCRs can hurt an agent’s needs for affiliation, then we can expect the agent to become alienated from the principal. However, if the superior values affiliation with the agent, then the principal’s own needs for affiliation and friendship will be frustrated. While it has not yet received attention in PCR literature, such a problem is well known in human resource management and employee appraisal literature (Wing, 2000). A superior in charge of checking on someone else’s capital investment project may, therefore, dislike PCRs. This effect of PCRs could be termed as ‘‘alienation’’ or ‘‘frustration of principal’s needs for affiliation effect of PCRs.’’ Ex-ante pinpointing effect. Measuring the results of a person’s actions draws that person’s attention to the measured tasks and the way in which they should be performed (Cammann, 1974; Hopwood, 1972; Locke & Latham, 1990; Ridgeway, 1956). This is synonymous with an increase in an agent’s knowledge and a reduction in the probability of misunderstandings between the employee and her superiors (i.e. the firm) (Locke & Latham, 1984). Clear, explicit targets, that is, knowledge about where to go are, as shown by many psychological experiments on ‘‘goal setting,’’ a prerequisite for productivity and high performance (Locke, 1968; Locke & Latham, 1984; Locke & Latham, 1990). The individual who does not know what is important to the superior is likely to face difficulties in making decisions about how to behave in the face of multiple options. This influence of a review on the employee’s knowledge ex-ante has, in fact, been recognized in general management control literature for some time. Its correspondence when conducting PCRs can be called a ‘‘pinpointing effect of a PCR.’’ Ex-ante expectancy effect. While this increase in knowledge about where the company wants the individual to go raises the ability to do ‘‘a good job,’’ a PCR also has an ex-ante impact on the employee’s expectations. It affects the willingness to do a good job since it lowers the results–rewards

Tying vs. Not-Tying Post-Completion Reviews to Extrinsic Rewards

107

probabilities to obtain rewards/avoid punishments by actions that lead to results unflavored by the firm and increases the results–rewards expectancies of alternative courses of action that lead to results measured and presumably favored by the organization. This change in expectations (and its behavioral implications) has been postulated in PCR literature since its earliest days (e.g. Grimes, 1954; Nicholson, 1962). The corresponding effect in performance measurement theory has received much attention not only in principal-agent and game theoretical models (e.g. Holmstro¨m & Milgrom, 1991; Wagenhofer, 1992) but also in psychological experiments (e.g. Churchill & Cooper, 1964), and empirical studies on control systems and performance measurement (e.g. Cammann, 1974). Such a change in expectations is likely to encourage an employee to work harder in solving a problem in a way that maximizes the measured results compared to the situation in which the individual knows that the results of the actions chosen will probably never be assessed (Holmstro¨m & Milgrom, 1991). This modification of behavior, hence, depends on the person’s subjective interpretation of the likelihood of a review, of what parts of her work might be subjected to it, and of the (negative) consequences that she will have to bear if the principal managed to disclose her non-conforming behavior in such a PCR. In short, it depends on the subjective estimation of a certain performance outcome’s instrumentality (P-R) for obtaining desired consequences or avoiding undesired ones. However, this effect has a flip side. Following analyses and empirical studies on performance measurement and control systems (e.g. Cammann, 1974; Holmstro¨m & Milgrom, 1991), it is likely that the agent will not work harder in all areas of her work (as is assumed in many articles and books that mention the positive effect of PCRs on employee motivation), but rather invest all her effort on the parts of the work that will be subjected to a PCR with a sufficiently high probability. For all other outputs (i.e. parts of her work that can either not be measured at all or that are likely not to be checked in the course of a PCR because of the costs associated with measuring them), the agent will leave the level of effort unchanged or even reduce the level of effort in order to release capacities for the measured (i.e. checked) parts of her work. Consequently, the agent will not necessarily act to achieve the company-wide optimum, but will rather strive for a local or even individual optimization of the parts of his work that will influence his evaluation by the principal (see Cammann, 1974 for a similar effect of key performance indicators). This can at the extreme lead to an ‘‘adherence by the letter’’ to the company procedures for outcomes subjected to PCRs and to opportunistic behavior.

108

STEFAN LINDER

This effect of PCRs on an agent’s expectations ex-ante shall be called ‘‘exante expectancy effect of a PCR.’’

5.2. Ex-Post Effects Besides these ex-ante effects, several ‘‘after the fact’’ effects of PCRs on the abilities, expectations, needs, and attitudes of the individuals involved in a PCR can be identified. Ex-post knowledge generation/learning effect on the employee. The first ex-post effect is on the knowledge and abilities of the individual subjected to a PCR. It results from learning via feedback of results measured in the course of a PCR. This effect has been discussed and empirically analyzed both in management control literature and in the psychological feedback and employee evaluation literature (e.g. Annett, 1969; Bolger & O¨nkalAtay, 2004; Bonner & Walker, 1994; Hirst & Luckett, 1992; Luckett & Eggleton, 1991; Vroom, 1964b). Similarly, researchers on PCRs have studdied this effect for many years conceputally and empirically (Chenhall & Morris, 1993; Ha¨gg, 1977; Kuszla & Pezet, 2006; Lu¨der, 1969). It is probably the best-researched effect of a PCR. Since this effect is about knowledge generation by means of learning from feedback, it shall be called an ‘‘ex-post learning effect of PCRs with respect to the subordinate’s knowledge.’’ Ex-post knowledge generation/learning effect on the superior. While the superior in charge of conducting a PCR on someone else’s work does not receive direct feedback on his own work by the means of this PCR (he checks the subordinate’s work, not his own), he, nevertheless, may learn something about capital expenditures. As long as he is not omniscient about capital expenditure planning and implementation, the likelihood is high that he will learn about factors in the success or failure of a capital expenditure project when conducting a variance analysis in the course of a PCR. This will allow him either to support the agent in developing ideas for corrections and carrying them out (if necessary) or to conduct them himself. This effect has occasionally been mentioned in PCR literature (e.g. Lu¨der, 1969) and can be called a ‘‘knowledge generation effect ex-post’’ that takes place with the superior. Ex-post expectations development effect. The third effect deals with changes in the employee’s expectations. It has been considered in management control literature, empirical studies, and game theoretic models for quite some time (e.g. Earley, 1986). This effect is twofold. On the one hand,

Tying vs. Not-Tying Post-Completion Reviews to Extrinsic Rewards

109

control systems induce learning of the ‘‘locus of control’’ (i.e. the results– rewards expectations) (Rotter, 1954). On the other hand, control systems also have an impact on the development of self-efficacy beliefs (i.e. action– results expectations). Literature from psychology has noted that self-efficacy increases when a certain course of action has led to the wished-for results but will decline if it has not (e.g. Bandura, 1997). Therefore, the change in self-efficacy is likely to induce revised decisions in future decision processes. This learning with respect to the subordinate’s expectancies shall be called ‘‘ex-post expectancy development effect of a PCR.’’ Ex-post effect on the employee’s needs. As stated earlier, some of a person’s needs require extrinsic satisfaction while the status of satisfaction of the need for achievement is influenced intrinsically. Thus, even when PCRs are (presumably) not linked to extrinsic rewards or punishments, they nevertheless have an ‘‘after the fact’’ influence on the individual’s level of need satisfaction or frustration. If the review shows that the results obtained by the employee’s actions are favorable, he should feel intrinsic satisfaction of his needs for achievement, while he will experience frustration of these needs if the PCR points to (unfavorable) deviations from the anticipated targets. Even in the case of a mere measuring of achieved results without linking these results obtained to any form of extrinsic reward or punishment the intrinsic needs for achievement can be satisfied (or frustrated) by the PCR. As, according to Alderfer (1972), a person’s intrinsic needs of achievement cannot be fully satisfied, but will always be relevant, a partial satisfaction of them due to a positive outcome of a review process, will not compromise the employee’s willingness to ‘‘do a better job next time.’’ Therefore, when there are only intrinsic incentives linked to PCRs, no significant changes will occur from PCRs with respect to future decision processes. Any such influence of a PCR on the subordinate’s degree of satisfaction of her (extrinsic and/or intrinsic) needs shall be called a ‘‘need satisfaction (or frustration) effect of a PCR.’’ Ex-post attitude development effect. Even though need satisfaction or frustration is limited to the achievement needs when there is no tying of extrinsic rewards, changes in the individual’s attitudes are likely: if the results measured in the PCR are accompanied by the wished-for consequences of achievement need satisfaction, a positive attitude is likely to be created (or sustained); if the consequences are a frustration of the needs for achievement due to an unfavorable result of the agent’s actions uncovered in the PCR, a negative attitude towards the respective task (and potentially the PCR and the person conducting it) will be formed (or sustained).

110

STEFAN LINDER

This effect can be called an ‘‘ex-post attitude development effect.’’ Ex-post effect on the supervisor’s needs for power: Transferring the right to a person to control another person can be viewed as a form of French and Raven’s (1959) ‘‘legitimate power’’ (p. 159). Provided that his work is not the object of a control by the individual whose work he reviews, appointing someone to check someone else’s work implies that the former has a higher role than the latter. Furthermore, subjecting a person’s results to a control is commonly considered (and often felt) to placing that person, her abilities, and her motivation under scrutiny. Hence, it can be assumed that a PCR can be an instrument (partially) satisfying the first person’s needs for power. This effect shall be called ‘‘power effect of a PCR.’’

6. EFFECTS OF PCRS WHEN TIED TO EXTRINSIC REWARDS After having described a grid of potential effects of a PCR on the (presumably two) individuals involved, the influence of the tying of rewards and punishments on the effects of a PCR can follow. Provided that the extrinsic rewards or punishments for the subordinate linked to PCRs are relevant to his needs, such a tying of the results measured in a PCR to extrinsic incentives will have several influences on the effects just described. Impact on ex-ante distrust effect. Tying extrinsic rewards or punishments that fit the agent’s needs for safety or remuneration, social belonging, social reputation, and/or power to the results obtained in a PCR raises the importance of a PCR’s results for the agent’s need satisfaction above the level of mere intrinsic interest stemming from his needs for achievement. This is likely to increase the agent’s belief that the firm distrusts him. By linking rewards to PCR results the firm/the superior seems to believe that it is necessary to punish or to reward the agent extrinsically; in other words she seems not to believe in the agent’s (intrinsic) willingness to do a good job (Blau, 1955; Osterloh, 2006). This kind of increase of a ‘‘distrust effect’’ is likely to raise (or create) tensions between the person conducting the review and the person subjected to it. In his study of accounting data and performance evaluations, Hopwood (1972) found such a link between tying of incentives to accounting data and the tensions between the people involved. Falk and Kosfeld (2006) found

Tying vs. Not-Tying Post-Completion Reviews to Extrinsic Rewards

111

a similar link in a laboratory study of control systems. Since a growth in the tensions between the person responsible for carrying out a review and a person subjected to it will probably lead to a reduction of satisfaction of the first person’s needs for friendship, we, therefore, can formulate our first proposition: Proposition 1. Tying extrinsic rewards or punishments for the subordinate to PCRs will increase the distrust effect of PCRs. Influence on the ex-ante alienation effect. The increased distrust instilled by tying extrinsic incentives to PCRs will further alienate the agent from the superior above the level already implied by merely conducting a PCR (without any link to extrinsic incentives) (Reber & Van Gilder, 1982 about supervision). As a consequence, the principal will be less able to keep the current level of satisfaction of her own needs for affiliation (i.e. she will suffer a more intense frustration of these needs). Therefore, we propose that tying PCRs to extrinsic rewards and punishments will exacerbate the frustration of the principal’s needs for affiliation implied by a PCR. Consequently, we suggest Proposition 2. Tying extrinsic rewards or punishments for the subordinate to PCRs will increase the alienation effect of PCRs. Impact on the ex-ante pinpointing effect. Parts of the performance measurement literature postulate that linking incentives to performance will inform an employee of the duties that are particularly important (see, e.g. Dalton, 1971; Merchant & Van der Stede, 2003). However, the postulated impact actually seems doubtful. Goal-setting theory proposes that clear and precise targets provide the basis for high performance since targets set this way reduce the probability of misunderstandings (Locke & Latham, 1984; Locke & Latham, 1990). However, merely linking rewards and punishments to the achievement of certain results measured in a review does not make these targets clearer than the pure announcement of the measuring already done on its own. It does not provide the employee with additional information on what results to obtain then in the case of the pure measuring of these results. A tying only increases the valence of reaching these results for the employee but does not help to clarify the desired results. Consequently, PCRs linked to rewards are likely not to have a different impact on the knowledge processes via a change in abilities/knowledge of the individual than PCRs have on their own.

112

STEFAN LINDER

Proposition 3, therefore, is Proposition 3. Tying extrinsic rewards or punishments for the subordinate to PCRs does not have an impact on the knowledge of the individual ex-ante – that is, on the pinpointing effect of PCRs. Influence on the ex-ante expectancy effect. Linking extrinsic rewards or punishments to the achievement of certain results will affect the utility the individual attributes to attaining these results, since ‘‘more is at stake.’’ Consequently, she will be induced to behave in a way that maximizes the measured results determining her (most valued) rewards. This impact has been highlighted and analyzed in analytical models and conceptual works for performance management systems (see, e.g. Holmstro¨m & Milgrom, 1991). Similarly, Azzone and Maccarrone (2001), Borer (1978), and Nicholson (1962) mention such an impact for PCRs tied to incentives. However, no conceptual or empirical investigation of this effect in a PCR setting is available. Empirical and experimental studies investigating the impact of monetary incentives (tied to results) on performance in general performance measurement settings have, however, come to mixed results (Bonner & Sprinkle, 2002; Kunz & Pfaff, 2002). Since most of the studies either have some methodological problems or were not targeted to testing this proposition in a multi-task/multi-parts setting (ibid.), it still seems reasonable to stick to the proposition developed in the conceptual works. This even more so as experimental research has concentrated on financial incentives when analyzing the impact of tying performance measurement to rewards and punishments. Non-monetary extrinsic rewards and punishments, however, are significant in most ‘‘natural’’ settings (e.g. avoiding ‘‘loss of face,’’ a career-advancement, public recognition of success), perhaps making financial incentives just one aspect that employees consider when optimizing their work direction and effort. Furthermore, as stated earlier, the relevance of extrinsic incentives depends on the person’s current needs. Some of the empirical investigations might not have been sufficiently controlled for this factor. Thus, in this chapter we suggest – remaining in line with conceptual works about performance measurement and with the postulates in PCR literature – that (as long as the incentives are relevant to the person) an increase in the expectancy effect will take place when PCRs are tied to extrinsic rewards and/or punishments. Proposition 4. Tying extrinsic rewards or punishments for the subordinate to PCRs increases the ex-ante expectancy effect.

Tying vs. Not-Tying Post-Completion Reviews to Extrinsic Rewards

113

Influence on the subordinate’s ex-post learning effect. Numerous publications on the administration of extrinsic rewards and punishments postulate a greater increase in learning when feedback systems are linked to such incentives than in situations without such a link (e.g. Dalton, 1971; Dillon & Caldwell, 1981). This hypothesis, however, appears doubtful from the perspective of cognitive psychology. It follows the classical behavioral works by Skinner (1965) and Thorndike (1911). Their animal experiments showed that an increase in incentives leads to a more significant shift in behavior. However, as cognitive psychology starting from a famous experiment by Tolman and Honzik (1930) points out, it is important to differentiate learning (i.e. knowledge acquisition) from the use of knowledge (i.e. what can actually be seen in behavior). Classical, behaviorist studies and theories do not make this distinction. They, therefore, are likely to confound the two effects – and, consequently, their root causes. As already shown by Tolman and Honzik (1930), learning itself seems not to be influenced by external incentives – only the degree by which the acquired knowledge is applied in actual behavior is, provided that the same incentives are still expected and considered valuable in the future. Based on this and other experiments from social-cognitive psychology, linking extrinsic rewards and punishments to the achievement or non-achievement of certain results may impact an individual’s decisions of how to behave in the future, but not the amount of knowledge generation or remembrance (see, e.g. Mazur, 2002). Consequently, we can assume that no direct influence of tying PCRs to extrinsic rewards or punishments exists on the knowledge a person learns from seeing the results of her actions. Such incentives may only impact the way that person might behave in the future (provided that similar rewards and punishments are to be administered at that point in time and that they remain relevant). Therefore, we suggest Proposition 5. Tying extrinsic rewards or punishments for the subordinate to PCRs does not increase the magnitude of the knowledge the subordinate generates from feedback. Influence on the principal’s ex-post learning effect. Since we assumed that the incentives tied to a PCR are rewards and/or punishments for the employee subjected to a PCR and not for the superior, no influence of these rewards should be expected on the superior’s learning. This even more so given the irrelevance of incentives for learning just described following

114

STEFAN LINDER

social-cognitive psychology. Proposition 6. Tying extrinsic rewards or punishments for the subordinate to PCRs does not increase the knowledge the principal generates by learning about the success or failure of the agent’s actions. Implications for the ex-post expectancies development effect. The employee learns which rewards or punishments to expect based on her present experiences (Porter & Lawler, 1968). This will influence her decisions about how to behave. However, this development of an individual’s expectations does not differ from the learning effect on her abilities apart from the fact that it involves the person’s expectations, not her abilities. Consequently, the same should hold true for the expectations development effect as for the ability/knowledge development effect ex-post. Hence, we can assume that linking PCRs to incentives will not have an impact on how well the developed expectations will be remembered. Proposition 7. Tying extrinsic rewards or punishments for the subordinate to PCRs does not increase memorization of expectancy changes caused by feedback about the results obtained. Interplay with the ex-post effects on the employee’s needs Only if extrinsic rewards and/or punishments are tied to a PCR, such a control will be linked with satisfaction (positive outcome of a control) or frustration (negative outcome of a review) of the extrinsic needs. As no crowding out of intrinsic by extrinsic motivation is likely to take place except in rare cases (see, e.g. Cameron & Pierce, 2002; Lao, 1981), linking PCRs to rewards implies that an individual subjected to such a review will experience satisfaction or frustration of both intrinsic and extrinsic needs. Since extrinsic needs can (in contrast to the intrinsic needs for achievement) be fully satisfied (cf. Alderfer, 1972), extrinsic rewards may become irrelevant. In this case, tying PCRs to extrinsic rewards will not differ from a situation with purely intrinsic incentives (i.e. a sole ‘‘need for achievement’’) even though the organization provides extrinsic rewards. However, whenever the extrinsic incentives are (or remain) relevant to the individual, the following should hold true: Proposition 8. Provided that incentives offered are relevant to the current need state of the individual, tying extrinsic rewards or punishments for the subordinate to PCRs increases the satisfaction or frustration over the level of a situation with pure intrinsic motivation.

Tying vs. Not-Tying Post-Completion Reviews to Extrinsic Rewards

115

Influence on the ex-post attitude development effect. Since a tying of PCRs to rewards and punishments implies that an individual’s extrinsic needs will either be satisfied or frustrated, and since attitude development depends on the usefulness of an object experienced to satisfy a person’s needs (Crawford et al., 2002; Krech et al., 1962), linking incentives to PCRs will increase the ex-post attitude development effect compared to a situation of a ‘‘pure’’ PCR. Provided that the incentives are relevant to the person’s needs, we would assume that the larger the rewards and punishments are, the more significant the shift in attitudes should be. While not directed at PCRs, Kay, Meyer, and French (1965) actually support the positive impact of linking rewards to appraisals on the magnitude of the attitude (further-) development of employees. Therefore, we can expect the following proposition to hold for linking PCRs to extrinsic rewards and punishments: Proposition 9. Tying extrinsic rewards or punishments for the subordinate to PCRs increases the ex-post attitude development effect experienced by the agent. Impact on the ex-post effect on the superior’s needs for power. The power of the superior over the person subjected to a PCR depends (besides ‘‘legitimate power’’) on the ability to cause good or negative feelings. By linking PCRs to extrinsic rewards and punishments, the power of the principal grows: he gains the opportunity to distribute rewards to the subordinate based on the conclusions he draws (voluntarily) about the results achieved. In contrast, a PCR without any linking of extrinsic rewards and/or punishments lacks ‘‘coercive power’’ (French & Raven, 1959, p. 156). For this reason, we can assume that PCRs that are tied to extrinsic rewards and/or punishments for the subordinate will be more useful from the superior’s point of view to satisfy her own needs for power. The magnitude of the power effect should, therefore, grow as PCRs are linked to extrinsic incentives for the agent. Proposition 10. Tying extrinsic rewards or punishments for the subordinate to PCRs increases the ex-post effect on the principal’s needs for power.

7. DISCUSSION The preceding analysis and the propositions lead to the conclusion that PCRs have a multitude of direct effects on the abilities, attitudes,

116

STEFAN LINDER

expectations, and needs of the individuals involved in them. Some of these effects seem to increase when PCRs are tied to extrinsic rewards and/or punishments while others remain unchanged. It is important to note that the propositions suggest that a tying of PCRs to extrinsic rewards and punishments will not change the learning and knowledge generation effects ex-ante and ex-post of a PCR. However, it increases the distrust-, alienation-, and power-effects just like it impacts the effects of need satisfaction/frustration and attitude formation. Furthermore, tying PCRs to extrinsic rewards and/or punishments is not likely to change the probabilities underlying the expectations about the PCR held by a person who is subjected to such a review. However, it influences the importance of the results of certain actions for the agent’s need satisfaction or frustration. Thereby, it is likely to draw the employee’s effort towards the parts measured – and rewarded or punished. This can have dysfunctional consequences from the superior’s point of view. Even though these propositions require empirical non-falsification before offering a basis for practical recommendations on how to design PCRs, they nevertheless call for a careful selection and design of PCRs. This includes taking into account the objectives that a firm wants to attain by using PCRs when deciding whether or not to tie. Following the propositions, tying may be expected to support goal achievement from the firm’s/principal’s perspective just as much as it may hamper it – depending on the firm’s objectives:  For fostering knowledge acquisition by individuals ex-ante and/or expost, tying PCRs to rewards seems unnecessary. Neither the pinpointing nor the ex-post learning effects (at the agent and the principal) increase when PCRs are tied to extrinsic incentives. Considering the potential dysfunctional consequences like an increase in distrust or alienation, such a tying does not seem advisable. The present discussion, therefore, suggests that the lack of a positive effect on managerial learning as found in Ha¨gg’s (1977) study is likely not to be due to a mere sampling problem, but resulting from the way human learning works (i.e. to be due to the ineffectiveness of incentives to foster such a learning based on feedback by means of a PCR).  For increasing the effort that employees exert on the parts subjected to a PCR, linking such reviews to extrinsic incentives, however, might be advisable. Nevertheless, it should not be overlooked that this increased concentration of work effort can also be dysfunctional – if the success of the capital expenditure projects subjected to the PCRs can only be partly

Tying vs. Not-Tying Post-Completion Reviews to Extrinsic Rewards

117

measured and if the employee can promote the success of the measured parts at the expense of the parts that do not lend themselves to being checked in a PCR. In many practical cases cash inflows of a firm are likely to be much harder to ascribe to certain capital expenditure projects than are the cash outflows for the projects. Thus, in this case, tying PCRs to extrinsic incentives is prone to lead employees to invest a great deal of effort on staying within the capital expenditure budget (i.e. the cost budget of a project), while putting less emphasis on the promised cash inflows or perhaps voluntarily overstating the future cash inflows of a proposed capital expenditure project in order to ensure that it will pass the firm’s hurdle rates (Soares et al., 2007). It would be necessary to track the accuracy of both cash outflows and inflows to reduce this dysfunctional implication of tying rewards to PCRs.  For raising the superior’s work motivation, tying of PCRs to extrinsic incentives for the employee can work in two ways. It can increase satisfaction of his needs for power. At the same time, however, it may hamper the satisfaction of the needs for affiliation. While these aspects have not been discussed in the literature on PCRs, the propositions call for taking them into account in future analyses. The size of the two counteracting effects on the superior’s need satisfaction and motivation should depend on the importance of the two needs to the principal: a person in charge of conducting PCRs and whose needs for affiliation are satisfied through interaction with his peers and/or family or friends, is less likely to suffer a significant reduction of his level of need satisfaction than someone who depends on satisfying his needs for affiliation with the person whose work he is checking. Thus, one could even conceive a case in which a firm decides to link PCRs to rewards for the sole reason of satisfying the superior’s needs for power. This might be a viable strategy if the superior could satisfy his needs for affiliation either with other employees than the one subjected to the PCRs, or even outside the firm (i.e. in their private lives of members of a sports club, family, or friends). According to the overwhelming majority of empirical studies, the primary goal of firms in Canada, Germany, Italy, the U.K. and the U.S. for conducting PCRs is increased knowledge for future capital expenditure projects (to avoid repeating mistakes) and for solving problems with current projects (Epstein & Rejc, 2005; Riggs et al., 1998; Segelod, 1995; Soares et al., 2007). Disciplining the employees responsible for the capital expenditure projects to work harder and more carefully, is, according to

118

STEFAN LINDER

these studies, at best only a secondary objective of PCRs (Linder, 2006). An increase in the satisfaction of the superior’s needs for power, in fact, is not mentioned at all in these empirical studies (ibid.). Following the propositions, a tying of PCRs to extrinsic rewards and/or punishments does not increase the knowledge generation (i.e. change in abilities) of a PCR. Therefore, such a tying does not seem to be necessary in order to achieve the objectives for which most firms strive for by conducting PCRs. Considering the potential negative effects of such a tying in the form of increased distrust- and alienation-effects, it does not seem to be advisable to tie PCRs to extrinsic rewards (or punishments). Only for the few firms who want both knowledge generation and a concentration of employees’ work effort, such a tying might seem to be an option. However, since there are considerable undesirable side effects of a tying of PCRs to rewards, it seems preferable to rely on the disciplining function of PCRs that are not tied to extrinsic incentives.

8. TENTATIVE CONCLUSIONS AND IMPLICATIONS FOR RESEARCH The present chapter focuses on an often overlooked, yet quite important aspect of the capital expenditure process: tying vs. not tying of PCRs to extrinsic rewards and/or punishments. Based on a model of human characteristics, the chapter discussed potential effects of PCRs on the individuals involved. It then looked at what happens to the magnitude of these effects of a PCR when the review is tied to extrinsic rewards and/or punishments for the person subjected to a PCR. The results of this analysis were formulated as a set of propositions, which were then discussed. The propositions suggest that tying PCRs to extrinsic rewards and punishments will leave the learning and knowledge generation effects of a PCR unchanged, while increasing the distrust-, alienation-, and powereffects. Furthermore, while it does not change the probabilities underlying the agent’s expectations about the PCR, it influences the valence of certain actions’ results. Thereby, it leads to an increase in the effort exerted on the parts measured in the PCR. This, however, will take place at the expense of other, not measured or less/not rewarded or punished parts. Careful selection and design of PCRs and their linking to extrinsic incentives, hence, seems to be very important. Depending on the objectives that the firm wants

Tying vs. Not-Tying Post-Completion Reviews to Extrinsic Rewards

119

to attain by using PCRs, we would expect that tying can, in fact, either support or interfere with the achievement of the firm’s goals. Since many empirical studies have found that learning from feedback is the predominant objective that Western companies try to attain by using PCRs, no tying of PCRs to rewards and punishments seems necessary for full goal achievement. In addition, considering the potential increase in the dysfunctional effects of PCRs from a tying of PCRs to extrinsic rewards, such a tying does not seem to be advisable. Therefore, in general the propositions and the goals stated by firms in empirical studies call for not tying PCRs to extrinsic rewards and/or punishments. There are, however, a couple of limitations to the present analysis and its conclusions: First, the analysis is based upon a psychological model of human beings. Since psychology is still a young academic field, further psychological research might imply changes in the number, structure, and definition of the underlying factors that influence human behavior. Hence, while we assume that the model covers all important variables, some minor modifications of the propositions may become necessary as psychological research progresses. Closely linked to the first limitation is a second one: the analysis does not distinguish the magnitude of impact extrinsic rewards have from that of extrinsic punishments on human need satisfaction. This is because psychological researchers are still not entirely sure about whether humans are risk averse, risk neutral, or risk seeking (Mazur, 2002). Some older studies point to risk seeking, since they report a more powerful impact of rewards than punishments (cf. e.g. Annett, 1969; Skinner, 1965). In contrast, more recent studies point to a greater impact of punishments (e.g. Kahneman & Lovallo, 1993; Tversky & Kahneman, 1991). However, many of the available studies have methodological problems (Mazur, 2002). Therefore, the present analysis assumes risk neutrality. Future research should – as psychological studies progress – abandon this assumption and adapt the propositions accordingly. Furthermore, it is assumed that it is possible to measure the success or failure of a capital expenditure project, to identify the reasons for potential deviations, and to reward or punish. Such a measuring of results may, of course, not always be practicable (Pfeffer, 1998 for performance measurement in general). However, this assumption reduces the complexity of the analysis to a manageable level. Future research might want to relax this assumption and the propositions derived therefrom.

120

STEFAN LINDER

The present chapter is limited to what Dubin (1978) and Popper (1980) call hypothesis/proposition development. It does not try to provide an empirical examination of the propositions presented. Instead, it concentrates on deriving propositions for future empirical testing. Consequently, these propositions need to be empirically scrutinized (and perhaps refined) before they can be used by practitioners. Moreover, this chapter is limited to the analysis of a two-person situation. In practice, however, PCRs are often carried out by a group of people. Similarly, it is typically also a group of people that is subjected to a PCR since capital expenditures are quite often developed and implemented by project teams. In these cases, improvements in knowledge can (if communication takes place) extend to the respective groups. This should be considered when talking about the full value that PCRs may offer to a firm with respect to knowledge management (Kuszla & Pezet, 2006). Similar aspects can be expected with respect to the development of shared attitudes (i.e. culture) in a firm. Furthermore, since studies on the impact of control systems on trust and cooperation in multi-person collaborative environments show differences from studies in single-individual settings (e.g. Coletti, Sedatole, & Towry, 2005), the propositions derived in this chapter for linking PCRs to rewards on attitudes might need refinement in order to apply to capital expenditure projects carried out by a group of people. Incorporating these aspects into the analysis would, however, have made it much more complicated, while it would probably not have – given the current dearth of research on the tying of PCRs to rewards – yielded significantly better insights than the two-person analysis chosen. Future research should, however, extend the examination to multi-person settings. Thus, while this chapter sheds some light on the question of tying/not tying, much more research is needed from management control researchers on the social-cognitive implications of PCRs and of tying such reviews to rewards and/or punishments.

REFERENCES Ajzen, I. (1988). Attitudes, personality, and behavior. Stony Stratford: Open University Press. Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs: Prentice Hall. Albert, H. (1984). Wissenschaftstheorie. In: E. Grochla, & W. Wittman, (Eds), Handwo¨rterbuch der Betriebswirtschaft (4th ed., Vol. 1, cols. 4674–4692). Stuttgart: C. E. Poeschel Verlag. Alderfer, C. P. (1969). An empirical test of a new theory of human needs. Organizational Behavior and Human Performance, 4, 142–175.

Tying vs. Not-Tying Post-Completion Reviews to Extrinsic Rewards

121

Alderfer, C. P. (1972). Existence, relatedness and growth: Human needs in organizational settings. New York: The Free Press. Amabile, T., Hill, K., Hennessey, B., & Tighe, E. (1994). The work preference inventory: Assessing intrinsic and extrinsic motivational orientations. Journal of Personality and Social Psychology, 66, 950–967. Annett, J. (1969). Feedback and human behavior: The effects of knowledge of results, incentives and reinforcement on learning and performance. Baltimore: Penguin Books. Azzone, G., & Maccarrone, P. (2001). The design of the investment post-audit process in large organizations: Evidence from a survey. European Journal of Innovation Management, 4, 73–87. Bandura, A. (1997). Self-efficacy: The exercise of control. New York: W. H. Freeman and Company. Barkema, H. G. (1995). Do managers work harder when they are monitored? Kyklos, 48, 19–42. Blau, P. M. (1955). The dynamics of bureaucracy: A study of interpersonal relations in two government agencies. Chicago: The University of Chicago Press. Bolger, F., & O¨nkal-Atay, D. (2004). The effects of feedback on judgemental interval predictions. International Journal of Forecasting, 20, 29–39. Bonner, S. E., & Sprinkle, G. B. (2002). The effects of monetary incentives on effort and task performance: Theories, evidence, and a framework for research. Accounting, Organizations and Society, 27, 303–345. Bonner, S. E., & Walker, P. L. (1994). The effects of instruction and experience on the acquisition of auditing knowledge. The Accounting Review, 69, 157–178. Borer, D. (1978). Innerbetriebliche Investitionskontrolle in Theorie und Praxis. Bern: Verlag Paul Haupt. Bruns, W. J., Jr., & DeCoster, D. T. (Eds). (1969). Accounting and its behavioral implications. New York: McGraw-Hill. Cameron, J., & Pierce, W. D. (2002). Rewards and intrinsic motivation: Resolving the controversy. London: Bergin and Garvey. Cammann, C. (1974). The impact of a feedback system on managerial attitudes and performance. Dissertation. Yale University. Charreaux, G. (2001). Introduction. In: G. Charreaux (Ed.), Images de l’investissement: Au-dela` de l’e´valuation financie`re: une lecture organisationnelle et strate´gique (pp. 1–10). Paris: Librairie Vuibert. Chenhall, R. H., & Morris, D. (1993). The role of post completion audits, managerial learning, environmental uncertainty and performance. Behavioral Research in Accounting, 5, 170–186. Churchill, N. C., & Cooper, W. W. (1964). Effects of auditing records: Individual task accomplishment and organization objectives. In: W. W. Cooper, H. Leavitt & M. W. Shelly, II. (Eds), New perspectives in organization research (pp. 270–275). New York: Wiley. Cohen, H. (1965). The demonics of bureaucracy: Problems of change in a government agency. Ames: Iowa State University. Coletti, A. L., Sedatole, K. L., & Towry, K. L. (2005). The effect of control systems on trust and cooperation in collaborative environments. The Accounting Review, 80, 477–500. Crawford, E. L., Luka, B., & Cacioppo, J. T. (2002). Social behavior. In: H. Pashler & R. Gallistel (Eds), Stevens’ handbook of experimental psychology. Learning, motivation, and emotion (3rd ed., vol. 3, 737–799). New York: Wiley.

122

STEFAN LINDER

Daft, R. L., & Noe, R. A. (2001). Organizational behavior. Fort Worth: Harcourt College Publishers. Dalton, G. W. (1971). Motivation and control in organizations. In: G. W. Dalton & P. R. Lawrence (Eds), Motivation and control in organizations (pp. 1–35). Homewood and Georgetown: Richard D. Irwin and The Dorsey Press. Dalton, G. W., & Lawrence, P. R. (Eds). (1971). Motivation and control in organizations. Homewood and Georgetown: Richard D. Irwin and The Dorsey Press. De Bodt, E., & Bouquin, H. (2001). Le controˆle de l’investissement. In: G. Charreaux (Ed.), Images de l’investissement: Au-dela` de l’e´valuation financie`re: une lecture organisationnelle et strate´gique (pp. 115–166). Paris: Librairie Vuibert. Deci, E. L. (1971). Effects of externally mediated rewards on intrinsic motivation. Journal of Personality and Social Psychology, 18, 105–115. Deci, E. L. (1972). The effects of contingent and noncontingent rewards and controls on intrinsic motivation. Organizational Behavior and Human Performance, 8, 217–229. Dillon, R., & Caldwell, J. (1981). A system for postauditing capital projects. Managerial Planning (Planning Review), 9, 18–2230. Drury, C., & Tayles, M. (1996). UK capital budgeting practices: Some additional survey evidence. The European Journal of Finance, 2, 371–388. Dubin, R. (1978). Theory building. New York: The Free Press. Earley, P. C. (1986). An examination of the mechanisms underlying the relation of feedback to performance. Academy of Management Proceedings, 214–218. Epstein, M. J., & Rejc, A. (2005). Measuring the payoffs of IT investments. CMA Management (December/January), 20–24. Ewert, R., & Wagenhofer, A. (2003). Interne Unternehmensrechnung (5th ed.). Berlin: Springer. Falk, A., & Kosfeld, M. (2006). The hidden cost of control. American Economic Review, 96, 1611–1630. Fedor, D. B. (1991). Recipient responses to performance feedback: A proposed model and its implications. Research in Personnel and Human Resources Management, 9, 73–120. Foss, N. J. (2007). Comment – Awards as compensation. European Management Review, 4, 21–23. French, J. R. P., Jr., & Raven, B. (1959). The bases of social power. In: D. Cartwright (Ed.), Studies in social power (pp. 150–167). Ann Arbor: The University of Michigan. Frey, B. S. (1993). Does monitoring increase work effort? The rivalry with trust and loyalty. Economic Inquiry, 31, 663–670. Frey, B. S. (1997). Not just for the money: An economic theory of personal motivation. Cheltenham and Brookfield: Edward Elgar. Frey, B. S. (2007). Awards as compensation. European Management Review, 4, 6–14. Galbraith, J., & Cummings, L. L. (1967). An empirical investigation of the motivational determinants of task performance: Interactive effects between instrumentality-valence and motivation-ability. Organizational Behavior and Human Performance, 2, 237–257. George, J. M., & Jones, R. J. (2005). Understanding and managing organizational behavior (4th ed.). Upper Saddle River: Prentice Hall. Grimes, L. E. (1954). A comprehensive system of preformance reports. In: American Management Association (Eds), Tested Approaches to Capital Equipment Replacement, AMA Special Report No. 1 (pp. 86–94), New York: AMA. Helfert, E. A. (1960). Checkpoints for administering capital expenditures. California Management Review, 2, 84–98.

Tying vs. Not-Tying Post-Completion Reviews to Extrinsic Rewards

123

Hirst, M. K., & Luckett, P. F. (1992). The relative effectiveness of different types of feedback in performance evaluation. Behavioral Research in Accounting, 4, 1–22. Holmstro¨m, B., & Milgrom, P. (1991). Multitask principal-agent analyses: Incentive contracts, asset ownership, and job design. Journal of Law, Economics and Organization, 7, 24–52. Hopwood, A. G. (1972). An empirical study of the role of accounting data in performance evaluation. Journal of Accounting Research, 10(Supplement), 156–182. Ha¨gg, I. (1977). Reviews of capital investments – Empirical studies. Research Report no. 1977/3. Department of Business Administration, University of Uppsala, Uppsala. Ilgen, D. R., & Feldman, J. M. (1983). Performance appraisal: A process focus. Research on Organizational Behavior, 5, 141–197. Istvan, D. F. (1961). Capital-expenditure decisions: How they are made in large corporations. Indiana Business Report no. 33. Bureau of Research – Graduate School of Business, Indiana University, Indianapolis. Jensen, M., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3, 305–360. Kahneman, D., & Lovallo, D. (1993). Timid choices and bold forecasts: A cognitive perspective on risk taking. Management Science, 39, 17–31. Kay, E., Meyer, H. H., & French, J. R. P. (1965). Effects of threat in a performance appraisal interview. Journal of Applied Psychology, 49, 311–317. Krech, D., Crutchfield, R. S., & Ballachey, E. L. (1962). Individual in society: A textbook of social psychology. New York: McGraw-Hill. Kunz, A. H., & Pfaff, D. (2002). Agency theory, performance evaluation, and the hypothetical construct of intrinsic motivation. Accounting, Organizations and Society, 27, 275–295. Kuszla, C., & Pezet, A. (2006). La post-e´valuation des investissements: Un dispositif d’apprentissage et de gouvernance. Paper presented at the Congre`s annuel de l’AFC, Lille. Lao, R. C. (1981). Effect of monetary reward on subsequent motivation. The Journal of Psychology, 107, 41–43. Lawler, E. E., III. (1976). Control systems in organizations. In: M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 1247–1291). Chicago: Rand McNally. Linder, S. (2006). Investitionskontrolle: Grundzu¨ge einer verhaltensorientierten Theorie. Wiesbaden: Deutscher Universita¨ts-Verlag. Locke, E. A. (1968). Towards a theory of task motivation and incentives. Organizational Behavior and Human Performance, 3, 157–189. Locke, E. A., & Latham, G. P. (1984). Goal setting: A motivational technique that works! Englewood Cliffs: Prentice Hall. Locke, E. A., & Latham, G. P. (1990). A theory of goal setting & task performance. Englewood Cliffs: Prentice Hall. Luckett, P. F., & Eggleton, I. R. C. (1991). Feedback and management accounting: A review of the research into behavioral consequences. Accounting, Organizations and Society, 16, 371–394. Lu¨der, K. (1969). Investitionskontrolle: Die Kontrolle des wirtschaftlichen Ergebnisses von Investitionen. Wiesbaden: Gabler Verlag. Manzoni, J.-F., & Barsoux, J. L. (2005). How bosses reveal their attitudes towards employees. European Business Forum, 20, 86–88.

124

STEFAN LINDER

March, J. G., & Olsen, J. P. (1975). The uncertainty of the past: Organizational learning under ambiguity. European Journal of Political Research, 3, 147–171. March, J. G., & Simon, H. A. (1958). Organizations. New York: Wiley. Maslow, A. H. (1954). Motivation and personality. New York: Harper & Row. Mazur, J. E. (2002). Learning and behavior (5th ed.). London: Pearson. McClelland, D. C., Atkinson, J. W., Clark, R. A., & Lowell, E. L. (1953). The achievement motive. New York: Appleton-Century-Crofts. McDermott, R. (1999). Why information technology inspired but cannot deliver knowledge management. California Management Review, 41, 103–117. Melton, A. W. (1973). Motivation and learning. In: D. C. McClelland & R. S. Steele (Eds), Human motivation: A book of readings (pp. 31–34). Morristown: General Learning Press. Merchant, K. E., & Van der Stede, W. (2003). Management control systems. Harlow: Prentice Hall. Mills, R. W., & Kennedy, J. A. (1990). Post-Completion audit of capital expenditure projects. London: The Chartered Institute of Management Accountants. Neale, C. W. (1989). Post auditing practices by UK firms: Aims, benefits and shortcomings. The British Accounting Review, 21, 309–328. Nicholson, J. T. (1962). Post completion review of capital expenditures: An examination of the procedures, the administration and the effects of post completion reviews in six large industrial companies. Unpublished dissertation, Harvard Business School. Osterloh, M. (2006). Human resources management and knowledge creation. In: I. Nonaka & K. Ichijo (Eds), Knowledge creation and management: New challenges for managers (pp. 225–248). Oxford: Oxford University Press. Pfeffer, J. (1998). Six dangerous myths about pay. Harvard Business Review, 76(June), 109–119. Popper, K. R. (1972). Objective knowledge. Oxford: The Clarendon Press. Popper, K. R. (1980). The logic of scientific discovery (10th ed.). London: Hutchinson. Porter, L. W., & Lawler, E. E., III. (1968). Managerial attitudes and performance. Homewood: Richard D. Irwin. Posey, I. A., Roth, H. P., & Dittrich, N. E. (1985). Capital investment post auditing practices. Cost & Management, 59(June), 14–18. Reber, R. W., & Van Gilder, G. (1982). Behavioral insights for supervision (2nd ed.). Englewood-Cliffs: Hall. Reinholt, M. (2006). No more polarization, please! Towards a more nuanced perspective on motivation in organizations. Working Paper. Center for Strategic Management and Globalization, Copenhagen. Ridgeway, V. F. (1956). Dysfunctional consequences of performance measurements. Administrative Science Quarterly, 1, 240–247. Riggs, J. L., Bedworth, D. D., & Randhawa, S. U. (1998). Engineering economics (4th ed.). Singapore: McGraw-Hill. Rotter, J. B. (1954). Social learning and clinical psychology. New York: Hall. Ryle, G. (1949). The concept of mind. London: Hutchinson. Saatmann, A. (1970). Die Erfolgskontrolle von Investitionen. Bochum: Universita¨tsverlag. Segelod, E. (1995). Resource allocation in divisionalized groups. Aldershot: Avebury. Simon, H. (1957). Models of man: Social and rational. New York: Wiley. Skinner, B. F. (1965). Science and human behavior. New York: The Free Press. Soares, J. O., Coutinho, M. C., & Martins, C. V. (2007). Forecasting errors in capital budgeting, a multi-firm post-audit study. The Engineering Economist, 52, 21–39.

Tying vs. Not-Tying Post-Completion Reviews to Extrinsic Rewards

125

Thorndike, E. L. (1911). Animal intelligence: Experimental studies. New York: Macmillan. Tolman, E. C., & Honzik, C. H. (1930). Introduction and removal of reward, and maze performance in rats. University of California Publications in Psychology, 4, 257–275. Tversky, A., & Kahneman, D. (1991). Loss aversion in riskless choice: A reference dependent model. The Quarterly Journal of Economics, 106, 1039–1061. Ukko, J., Karhu, J., & Pekkola, S. (2007). Employees satisfied with performance measurement and rewards: Is it even possible? Paper presented at the 4th EIASM Conference on Performance Measurement and Management, Nice. Vroom, V. H. (1964a). Work and motivation. New York: Wiley. Vroom, V. H. (1964b). Some psychological aspects of organizational control. In: W. W. Cooper, H. J. Leavitt & M. W. Shelly, II. (Eds), New perspectives in organization research (pp. 72–86). New York: Wiley. Wagenhofer, A. (1992). Abweichungsanalysen bei Erfolgskontrolle aus agencytheoretischer Sicht. Betriebswirtschaftliche Forschung und Praxis, 44, 319–338. Wing, K. T. (2000). Painless performance appraisals. Strategic Finance, 82(August), 62–65. Yukl, G. A. (1990). Skills for managers and leaders: Text, cases, and exercises. Englewood Cliffs: Hall.

This page intentionally left blank

OBJECTIVE REWARDING, MANAGERIAL MOTIVATION AND ORGANIZATIONAL COMMITMENT: THE INTERVENING ROLE OF JUSTICE Frank G. H. Hartmann and Sergeja Slapnicˇar ABSTRACT The behavioural accounting literature suggests that managerial motivation and commitment are affected by the way in which managerial performance is evaluated, but little is known about what aspects of rewarding system are crucial in evoking desired outcomes. In this chapter we explore whether managerial motivation and commitment depend on the level of objectivity of the rewarding system, and what variables mediate this relationship. We develop a causal model in which we set up hypotheses about the intervening role of managerial perceptions of justice. We test these hypotheses using survey data from a sample of 161 managers from 11 commercial banks. The chapter contributes to the literature by providing empirical evidence on the antecedents and consequences of justice in an applied setting, suggesting the important mediating role of justice in the relationship between rewarding systems and managerial motivation. Performance Measurement and Management Control: Measuring and Rewarding Performance Studies in Managerial and Financial Accounting, Volume 18, 127–145 Copyright r 2008 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-3512/doi:10.1016/S1479-3512(08)18006-6

127

128

FRANK G. H. HARTMANN AND SERGEJA SLAPNICˇAR

1. INTRODUCTION The organizational literature broadly evidences that managers’ perceptions of the justice of organizational arrangements is crucial for their motivation and subsequent performance (Colquitt, Conlon, Wesson, Porter, & Ng, 2001). Although the performance evaluation process is amongst the most critical of such organizational arrangements, there is little systematic evidence on the set of attributes of this process that invoke justice perceptions, nor on the relative importance of these attributes (Greenberg, 2001). Such knowledge is important, however, to inform current debates about the desirable characteristics of systems, metrics and methods for performance evaluation (Ittner, Larcker, & Meyer, 2003). In this chapter we investigate how the objectivity of rewarding system affects managerial justice perceptions and their subsequent motivation. We define objectivity in line with extant performance evaluation research as the extent to which the rewarding is formulaic and traceable, rather than subjective and implicit. Based on the organizational justice literature we argue that objectivity of rewarding and justice may be related in at least two ways. On the one hand, objective rewarding systems may have the benefit of providing managers’ with procedurally fair rewards (Leventhal, Karuza, & Fry, 1980), which will enhance subsequent managerial motivation (Cohen-Charash & Spector, 2001). On the other hand, however, objective performance evaluation procedures may disregard the subtleties of managers’ specific job characteristics, which would result in lower managerial perceptions of justice (Diekman, Barsness, & Sodnak, 2004), and subsequent motivation. Understanding the role of objectivity of rewarding may thus involve understanding the trade-off between its potential positive and negative consequences on justice. In this chapter, we aim to contribute to the organizational justice and performance evaluation literatures by exploring causal paths between objectivity of rewarding, managerial perceptions of procedural and distributive justice, and subsequent motivation. Regarding the performance evaluation literature, we extend current analyses on the (dys)functional consequences of performance evaluation systems that have addressed justice only tangentially (Hartmann, 2005). Regarding the organizational justice literature, we explore objective rewarding system as an antecedent of managerial justice perceptions, thus extending existing framework in which the causes of justice perceptions have received only negligible attention (Liao & Rupp, 2005). The remainder of this chapter is structured as follows. Section 2 below provides an overview of the literature that has addressed the relationship

Objective Rewarding, Managerial Motivation and Organizational Commitment

129

between managerial perceptions of justice of performance evaluation and rewarding processes. Section 3 then develops hypotheses about the relationships between the objectivity of rewarding system, managerial justice perceptions and behavioural outcomes. Section 4 describes the design of the empirical analysis that was conducted to test the hypotheses. Section 5 reports on the outcomes of the analyses. Section 6 concludes this chapter, with a discussion of the results, the strengths, the weaknesses, and the implications of the current study.

2. LITERATURE REVIEW Contemporary research in organizational behaviour demonstrates the importance of employees’ justice perceptions about a variety of organizational arrangements (Liao & Rupp, 2005) and for a variety of organizational outcomes (Cohen-Charash & Spector, 2001). Studies performed across industries and organizational levels consistently demonstrate that employees who feel that they are treated fairly by the organization, their superior and their colleagues are more motivated and perform better than employees perceiving their working environment as unfair (Colquitt et al., 2001; Cropenzano, Byrne, Bobocel, & Rupp, 2001). The literature, however, is less determined about the dimensionality of justice perceptions, the sources of various justice perceptions, and about the predictive power of these various types of justice on subsequent attitudes and behaviours (Lind, 2001). Regarding the dimensionality of justice, the justice literature appears undecided about the exact types of justice that can be meaningfully distinguished. The initial concept of justice, which is distributive justice, reflects the equity of the division of resources (e.g., pay) over employees. Equity thinking predicts that employees form and compare rewards–effort ratios, and will feel injustice if their ratio is less favourable than the benchmark (Koys, 1990). These employees are then expected to feel less motivated and adapt their efforts downwards to restore the ratio to an equitable level (Lawler, 1968). Typically, survey studies explore how comparative pay levels affect distributive justice judgments (Lawler, Koplin, Young, & Fadem, 1968). They demonstrate the validity of equity thinking, but over time established that justice perceptions also depend on the reward allocation process (Koys, 1990). Employees will judge a reward level as more fair if they perceive the procedures that have led to this reward level as fair, which is labelled procedural justice (Colquitt, 2001).

130

FRANK G. H. HARTMANN AND SERGEJA SLAPNICˇAR

So far, procedural justice has been operationalized as a normative set of attributes that make up a fair process (Leventhal et al., 1980). Typical survey studies assess the level of procedural justice by scores attached to attributes as openness, communication, participation, and the availability of a grievance system and demonstrate that procedural justice affects such factors as organizational commitment, job satisfaction, evaluation of authority, and performance (Colquitt et al., 2001; Cropenzano, Byrne, Bobocel, & Rupp, 2001). It is not clear yet whether the two concepts of justice are exhaustive, as several studies have argued for the further decomposition of procedural justice. Factors such as interactional justice and informational justice purport to reflect the fairness of procedural aspects, such as related to superior–subordinate relationships and information exchange (Colquitt, 2001). Evidence on the viability of these concepts is mixed (Colquitt et al., 2001; Cropenzano et al., 2001), but supports the need for a more analytical understanding of procedural justice than the normative model allows. Three directions seem worth exploring. First, by assuming a fixed set of procedural fairness dimensions, the normative model hinders an understanding of what elements of any given process are important. This may be a reason why studies have paid little attention to developing and testing theory about the causes of justice perceptions, and predominantly address the effects of justice perceptions. Second, regarding the sources of justice, the justice literature therefore provides ample opportunity for extension towards understanding the antecedents of justice (Liao & Rupp, 2005). This is also true for the rewarding process as a potentially important antecedent of justice (Folger & Greenberg, 1985; Koys, 1990). Third, regarding the type of justice perceptions, there is little systematic evidence on the relative predictive power of distributive justice and procedural justice neither on work-related motivation, nor on the question how justice perceptions are affected by the rewarding process (Colquitt et al., 2001; Cropenzano et al., 2001). Understanding the functioning of rewarding process, however, requires such more detailed knowledge, as the performance evaluation literature currently evidences (Ittner et al., 2003; Lipe & Salterio, 2002).

3. RESEARCH MODEL AND HYPOTHESES In the management accounting literature, a crucial question concerns the positive and negative effects of using various sorts of performance

Objective Rewarding, Managerial Motivation and Organizational Commitment

131

measurement systems on which rewarding is based. As these systems come in various forms, research is undecided about the most useful and predictive system element classification. Contemporary research has investigated, for example, the use of financial versus non-financial indicators, and the use of multiple versus single indicators (Lipe & Salterio, 2002; Moers, 2005; Ittner et al., 2003). Although each of these dimensions and many other (Hartmann, 2000) do explain some of the variance in the behavioural outcomes associated with the systems, we propose here that the objectivity versus subjectivity of the performance measurement and rewarding system may be a more complete variable that describes the relevant characteristics of rewarding. We define objectivity as the extent to which systems follow a procedural logic in the measurement and rewarding of managerial performance that is part of the organizational set of rules and regulations. Objectivity is the opposite of informal system, in which the superior has a large amount of discretion in establishing performance levels and allocating rewards. Objectivity refers to the use of formula based rewards. We believe that objective evaluations are likely to be a good source of clear and specific performance objectives. Based on the literature on motivation that stresses the importance of clear and specific targets, we expect that such can be the positive results of using objective evaluations (Locke, Latham, & Erez, 1988; Locke & Latham, 1990). Thus, we expect that there is a positive relationship between objectivity of rewarding, motivation, and subsequent organizational commitment. We present a set of related hypothesis as follows. H1. Objectivity of reward determination has a positive effect on motivation. Our aim is to explore whether this relationship is mediated by justice perceptions. In the next step we propose that organizational justice perception is a mediator between objectivity of rewarding and motivation. Based on the extant literature we distinguish between procedural justice and distributive justice (Colquitt, 2001). Procedural justice reflects the procedures used in establishing performance evaluations and rewards. Distributive justice, which relates to feelings of fairness associated with reward to effort ratio, is not directly related to the objectivity of rewarding. It is rather related to the procedural justice, which helps determine the level of effort for a given outcome. Subordinates, who perceive the procedures of a certain organizational process as fair, are also

FRANK G. H. HARTMANN AND SERGEJA SLAPNICˇAR

132

inclined to perceive the outcomes of these processes as fair. Therefore we expect that: H2. Objectivity of reward determination has a positive effect on procedural justice. H3. Procedural justice has a positive effect on distributive justice. The literature shows that these two types of justice have different effects on the feelings of subordinates towards work and organization. Prior research (Folger & Konovsky, 1989; Colquitt, 2001) shows that motivation is more directly affected by distributive justice. Distributive justice accounts for more variance in satisfaction with pay and consequently motivation. Organizational commitment, in contrast, is better explained directly by procedural justice as procedural justice is a characteristic of an institution or organization and hence, accounts for more variance in the attitudes about organization. In line with this evidence we propose the following three hypotheses. H4. Procedural justice has a positive effect on organizational commitment. H5. Distributive justice has a positive effect on motivation. H6. Organizational commitment has a positive effect on motivation. In Fig. 1 we give a graphical overview of the relationships that we intend to test. H4+

H2 +

Objectivity of rewarding

Procedural justice

Organisational commitment H6+

H3+ H5+ Distributive justice H1 +

Fig. 1.

Conceptual Model.

Motivation

Objective Rewarding, Managerial Motivation and Organizational Commitment

133

4. METHOD 4.1. Sample and Data We tested the model on a sample of 161 managers from Slovenian banks. Constructs have been measured via a web-survey. The survey method was chosen in line with the nature of investigation. The sensitivity in disclosing private information about motivation, commitment, and justice of certain organizational procedures requires full confidentiality between researcher and participant. Participants required an explicit confirmation that the data collection method was sufficiently protected and that the findings would not be disclosed on individual basis to the management board of the bank. Surveys are used in the majority of studies on which the present study builds, that investigate behavioural impacts of various organizational processes due to the fact that no archival data exist or are relevant. Furthermore, the choice of the method adds to the continuity of the literature, and the comparability of findings. The sample was selected in the following process: we approached management boards of all 18 Slovenian banks. Banks were considered appropriate, as we aimed for a sample of managers from a single industry with a homogenous and transparent business process. In 11 banks management boards agreed to participate in the research. They appointed the HR manager or HR executive director as a contact person. The contact persons selected those upper-middle managers of the bank that are in their position for at least one year and have an employment contract in which at least part of their yearly compensation would be contingent on performance. A total of 260 respondents were selected for participation. After obtaining e-mail addresses of all respondents, an individualized e-mail message was sent to each respondent with the request to participate, a reference to the endorsement from the bank, a unique and anonymous code, and a link to a webpage of one of the sponsoring universities. Respondents could login to the webpage with the questionnaire using the code. To increase the overall quality of the instrument, as well as the response rate we followed the measures proposed by Dillman (2000): an attractive layout on the webpage was designed with an easy filling in the form. Where we could, we used previously established items from the extant literature to measure the constructs. The exceptions are procedural justice and objectivity of reward determination, the latter being carefully constructed as a formative variable as explained below. The full questionnaire was drafted in English, after which it was translated into Slovenian using a translate–retranslate

FRANK G. H. HARTMANN AND SERGEJA SLAPNICˇAR

134

procedure. We tested the Slovenian version on five bank managers from three different banks to assess its length and possible difficulties respondents might have in understanding. To increase the response rate the nonresponding managers were asked two more times to participate if their code has not yet appeared in our database. The final response rate was 61.9%. Tables 1 and 2 contain descriptive information on the sample of respondents. The managers in our sample are in charge of distinct areas of responsibility, have a separate budget, and have experienced at least one performance evaluation cycle. The number of observations per organization ranged from 4 to 35, reflecting the organization’s size and responsibility centre structure (see Table 2).

Table 1.

Valid Missing

N Mean Median Standard deviation Minimum Maximum Percentiles

25 50 75

Table 2.

Sample Descriptives. Age

Years in Bank

Tenure in Function

153 8 43.45 44.00 8.176 29 58 36.00 44.00 50.00

157 4 9.82 8.00 8.580 1 35 3.00 8.00 13.00

158 3 4.76 3.25 4.730 1 30 1.88 3.25 6.00

Responsibility Areas of Respondents.

Job Trading Back office Investment banking and asset management Treasury Administration Internal operations Other Missing Total

Frequency

Percent

Cumulative

59 43 8 4 35 4 2 6

36.6 26.7 5.0 2.5 21.7 2.5 1.2 3.7

38.1 65.8 71.0 73.5 96.1 98.7 100.0

161

Objective Rewarding, Managerial Motivation and Organizational Commitment

135

4.2. Variable Measurement We specify the Objectivity of reward determination (OBJEC) as an index (Rossiter, 2002). The choice of the measurement is grounded on the fact that objectivity is a result of different (observable) attributes, rather than being caused by them. It is derived as a mean of four indicators – objectivity of measurement (indicating the predominant source of information for evaluation of a manager which on one extreme comes from the information system and on the other extreme from subjective superior’s judgment, but it can be combined) and quantification of targets (indicating whether performance targets have been on one extreme quantitatively defined and on the other extreme qualitatively agreed). Both indicators are measured on a scale from 1 to 5 where 1 indicates one extreme position (superior’s judgment and qualitative targets, respectively), and 5 the other extreme position (information system and quantitative targets, respectively). As we measure both aspects of objectivity for fixed and variable pay separately, we use four indicators. The index is subsequently calculated as a mean on all four indicators as objectivity of fixed pay does not necessarily correlate with the objectivity of variable pay so depicting them as reflective indicators would be a misspecification. All other constructs in the model are measured by reflective indicators. For each construct, three indicators measured managerial perceptions on five-point Likert scales. To measure the perception of Procedural justice (PROCJ) we developed our own instrument. In our instrument we asked respondents fairness perceptions about three elements of the performance evaluation system, without a priori delineating the normative dimensions of procedural justice. Hence, respondents were asked to indicate to what extent they believe that the system of target setting, performance measurement, and rewarding is going to fairly determine their pay. This was required since this study investigates antecedents of procedural justice perception, in contrast with studies on procedural justice that only study its effects (Cropenzano et al., 2001). This also allowed a better distinction between procedural and distributive justice, which is about the perception of justice that arises from comparisons with some sort of benchmark, such as the effort-to-pay of other periods or effortto-pay of their colleagues. Distributive justice (DISTRJ) was thus measured with instrument developed previously by Leventhal (1976). To measure Motivation (MOTIV) we used the intrinsic motivation inventory (Deci, Eghrari, Patrick, & Leone, 1994). Measurement of Affective organizational commitment (COM) was derived from Allan and Meyer (1990). When more items per latent construct were available, we used exploratory factor

136

FRANK G. H. HARTMANN AND SERGEJA SLAPNICˇAR

analysis using SPSS with Varimax rotation to identify the best indicators for the measurement model. After eliminating 12 items on the basis of their low loadings on the selected 4 factors (Joseph, Rolph, Ronald, & William, 1995), we retained 12 indicators (see the Appendix for an overview of constructs and indicators). Since the critical sample size depends on the number of variables in the model (Jo¨reskog & So¨rbom, 2001a, 2001b), LISREL 8.72 provides the estimation of the critical number for the given model. For our models critical size numbers is 150 indicating that our actual sample size of 161 is appropriate.

4.3. Analysis We test our hypotheses simultaneously by using structural equation modelling (SEM). SEM is composed of a measurement and a structural equation model. The measurement model provides an assessment of convergent and discriminant validity of constructs. The relationship between the observed variables and the latent variables provide us with information about the extent to which a given observed variable is able to measure the latent variable. A factor loading serves as a validity coefficient, whereas measurement error is a measure of reliability (Schumacker & Lomax, 1996). An advantage of such assessment of reliability over the Cronbach’s alpha is that in the latter case the measurement error is not incorporated explicitly into the statistical analysis to adjust directly the latent variables (Smith & Langfield-Smith, 2004). The exclusion of the measurement error in the estimation of the relationships may result in misstated regression coefficients. The second step in the SEM is the estimation of the specified structural equation model that provides an assessment of nomological validity. It facilitates testing of direct and indirect effects on dependent variables. There may be important theoretical differences between direct and indirect effects (Bollen, 1989) which are taken into account in SEM. Our model consists of one exogenous variable – Objectivity of rewarding and two determination system, two intervening variables Procedural and Distributive justice. Motivation and Organizational commitment are the outcomes (dependent variables) but as we predict that commitment also affects motivation the former is technically an intervening variable. The structure coefficients indicate the strength and direction of the relationships among the latent variables, whereas the disturbance term indicates the portion of a latent variable that is not explained by other specified latent variable(s) in the

Objective Rewarding, Managerial Motivation and Organizational Commitment

137

model. The parameters are estimated using LISREL 8.72 (Jo¨reskog & So¨rbom, 2001a) with the maximum likelihood estimation method.

5. RESULTS 5.1. Measurement Model The results of the analysis of the measurement model are contained in Table 3 and Fig. 2. For an easier interpretation of the outcomes, the estimated values are standardized. All factor loadings were greater than 0.70, which provided evidence of a satisfactory level of convergent validity, whereas relatively low measurement errors may be interpreted as a sign of appropriate measurement reliability. No further modifications of the model in terms of specifying indicators’ error covariances are suggested. All loadings are statistically significant at po0.01. Objectivity of reward determination is measured in the model by one indicator, as explained previously. In the case of one observed variable the literature (Jo¨reskog & So¨rbom, 2001a; Schumacker & Lomax, 1996, p. 80) argues that an assumption of no measurement error of a single indicator is as arbitrary as specifying an arbitrary measurement error. It has been suggested that a conservative value for standard error of 0.10 should be chosen or still arbitrary, but based on the given research context, to set standard error to the smallest value found for the other indicators (Anderson & Gerbing, 1988). We have chosen the first approach, but specified an even more conservative error variance of 0.20 (unstandardized) for that indicator (Cadez, 2005).

Table 3.

OBJEC PROCJ DISTRJ MOTIV COM

Correlations Matrix of the Latent Variables in the Measurement Model.

OBJEC

PROCJ

DISTRJ

MOTIV

COM

1.0 0.26 0.11 0.37 0.20

– 1.0 0.55 0.22 0.33

– – 1.0 0.14 0.12

– – – 1.0 0.40

– – – – 1.0

Note: po0.01, po0.05.

FRANK G. H. HARTMANN AND SERGEJA SLAPNICˇAR

138

0.16

0.22

objec1

0.92

OBJEC

0.84 MOTIV

distrj1

0.07

distrj2

0.32

distrj3

0.73 motiv2

0.29 0.46

0.76

0.88 0.96

motiv1

motiv3

DISTRJ

0.43

0.82

0.82 0.49 0.19 0.33

procj1

0.72 0.90

procj2

COM

com1

0.81 com2

0.34 0.34

0.86

PROCJ

com3

0.82

0.26

procj3 Goodness of Fit Statistics

χ 2 = 59.86, df = 56, p = 0.34, RMSEA = 0.021, CFI = 1.00, NFI = 0.96, NNFI = 0.99, GFI = 0.95,

AGFI = 0.91

Fig. 2. Estimated Measurement Model.

The goodness of fit measures assess the fit on three dimensions: (i) the overall fit of the model (a non-significant w2, RMSEAr0.05 and GFIZ0.9), (ii) comparative fit to a baseline model (the incremental fit: NFIZ0.9, NNFI or Trucker and Lewis’ indexZ0.9 and CFIZ0.95) and model parsimony (adjusts the fit for the number of parameters in the model: AGFIZ0.9) (see, e.g., Smith & Langfield-Smith, 2004; Cadez, 2005). All reported statistics are above acceptable levels recommended by the literature.

5.2. Structural Models Fig. 3 presents the statistics of the estimated structural model. We tested six hypotheses. All, but one have the expected sign and reasonably large structural coefficients. Objectivity of reward determination has a strong direct effect on Motivation (0.30) and an indirect affect on Motivation through intervening variables Procedural justice and

Objective Rewarding, Managerial Motivation and Organizational Commitment

R2=0.11

0.89 Objectivity of rewarding

0.26 ***

Procedural justice

0.33***

139

Organisational commitment

0.93 R2=0.07 R2=0.30

0.70

0.54***

0.34*** 0.07

Distributive justice

Motivation 0.76

0.30***

R2=0.24 Goodness of Fit Statistics χ = 63.06, df = 60, p = 0.37, RMSEA = 0.018, CFI = 1.00, NFI = 0.95, NNFI = 0.99, GFI = 0.94, AGFI = 0.91 2

Fig. 3.

Estimated Structural Model. Note: po0.01, po0.05.

Organizational commitment which together explain 24% of variance of Motivation. 11% of variance of Organizational commitment is explained by variation of Procedural justice. Both results suggest that Procedural justice perceptions of rewarding system is an important intervening variable explaining Motivation. In line with prediction Procedural justice strongly affects Distributive justice perception, however its perception is not significantly linked to Motivation. This calls for a further investigation of indirect links. Overall, the fit statistics prove a good fit between the theoretical model and underlying variance–covariance matrix.

6. CONCLUSIONS Our analysis shows that objectivity of reward determination matters for managerial motivation and organizational commitment. We explored two possible paths through which objectivity would infer one’s motivation and commitments. The first one is the direct relationship and the second one the indirect relationship through perception of justice. The results support our hypotheses of the importance of justice for motivation and commitment. We find that perception of procedural justice is crucial for the link between objectivity of rewarding and motivation. It has power to invoke organizational commitment and motivation. Moreover, it also affects the perception that the rewards are fair. Overall, the analysis provided great support for the

140

FRANK G. H. HARTMANN AND SERGEJA SLAPNICˇAR

relevance of the variables that we explored. This suggests that the objectivity aspect of performance measurement and rewarding systems may be a dimension that explains much of the variance in subsequent managerial behaviour. This has implications for the wider management literature, but also for the discussion on performance evaluation that we find in the accounting branch of the management literature. Of course, this study has limitations that are attributable to the method of data gathering and data analysis. The typical limitations of a survey study, which depends on perceptual measurement and self-reports, potentially limit the validity of our findings. However, we believe that the care by which we organized the survey and the high response rates guarantee that the results in the study cannot only be explained by random or systematic measurement error. Future studies may build on the present study by exploring other dimensions of performance evaluation systems, and the level of abstraction that we applied in this chapter as well. Therefore, we recommend that future studies complement their analysis of details of such systems with broader design characteristics.

REFERENCES Allan, N. J., & Meyer, J. P. (1990). The measurement and antecedents of affective, continuance, and normative commitment to the organization. Journal of Occupational Psychology, 63, 1–18. Anderson, J., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103(3), 411–423. Bollen, K. (1989). Structural equations with latent variables. New York, NY: John Wiley and Sons. Cadez, S. (2005). A contingency approach in designing a strategic management accounting system: A SEM model tested in the Slovene companies. University of Ljubljana. Unpublished doctoral dissertation. Cohen-Charash, Y., & Spector, P. E. (2001). The role of justice in organizations: A metaanalysis. Organizational Behavior and Human Decision Processes, 86, 278–321. Colquitt, J. A. (2001). On the dimensionality of organizational justice: A construct validation of a measure. Journal of Applied Psychology, 86, 386–400. Colquitt, J. A., Conlon, D. E., Wesson, M. J., Porter, C. O. L. H., & Ng, K. T. (2001). Justice at the millennium: A meta-analytic review of 25 years of organizational justice research. Journal of Applied Psychology, 86, 425–455. Cropenzano, R., Byrne, Z. S., Bobocel, D. R., & Rupp, D. E. (2001). Moral virtues, fairness heuristics, social entities and other denizens of organizational justice. Journal of Vocational Behavior, 58, 164–209.

Objective Rewarding, Managerial Motivation and Organizational Commitment

141

Deci, E. L., Eghrari, H., Patrick, B. C., & Leone, D. (1994). Facilitating internalization: The self-determination theory perspective. Journal of Personality, 62, 119–142. Diekman, K. A., Barsness, Z. I., & Sodnak, H. (2004). Uncertainty, fairness perceptions and job satisfaction: A field study. Social Justice Research, 17, 237–255. Dillman, D. A. (2000). Mail and Internet surveys: The tailored design method (2nd ed.). New York, NY: John Wiley & Sons. Folger, R., & Greenberg, J. (1985). Procedural justice: An interpretative analysis of personnel systems. Research in Personnel and Human Resource Management, 3, 141–183. Folger, R., & Konovsky, M. A. (1989). Effects of procedural and distributive justice on reactions to pay raise decisions. Academy of Management Journal, 32, 115–130. Greenberg, J. (2001). Setting the justice agenda, seven unanswered questions about ‘‘what, why, how’’. Journal of Vocational Behavior, 58, 210–219. Hartmann, F. G. H. (2000). The appropriateness of RAPM: Toward the further development of theory. Accounting, Organizations and Society, 25, 451–482. Hartmann, F. G. H. (2005). The effects of tolerance for ambiguity and uncertainty on the appropriateness of accounting performance measures. ABACUS, 41, 241–264. Ittner, C. D., Larcker, D. F., & Meyer, M. (2003). Subjectivity and the weighting of performance measures: Evidence from a balanced scorecard. The Accounting Review, 78(3), 725–758. Jo¨reskog, K., & So¨rbom, D. (2001a). LISREL 8 user’s reference guide. Chicago: Scientific Software International. Jo¨reskog, K., & So¨rbom, D. (2001b). Structural equation modeling with the SIMPLIS command language. Chicago: Scientific Software International. Joseph, F. H., Jr., Rolph, E. A., Ronald, L. T., & William, C. B. (1995). Multivariate data analysis (4th ed.). Englewood Cliffs, NJ: Prentice-Hall. Koys, D. J. (1990). Process equity in compensation administration. Labor Law Journal, 41(8), 586–591. Lawler, E. E. (1968). Effects of hourly overpayment on productivity and work quality. Journal of Personality and Social Psychology, 10, 306–313. Lawler, E. E., Koplin, C. A., Young, T. F., & Fadem, J. A. (1968). Inequity reduction over time in an induced overpayment situation. Organizational Behavior and Human Performance, 3, 253–268. Leventhal, G. S. (1976). Fairness in social relationships. In: J. W. Thibaut, J. T. Spence & R. C. Carson (Eds), Contemporary topics in social psychology (pp. 211–239). Morristown, NJ: General Learning Press. Leventhal, G. S., Karuza, J., & Fry, W. R. (1980). Beyond fairness: A theory of allocation preferences. In: G. Mikula (Ed.), Justice and social interaction (pp. 167–218). NY: Springer-Verlag. Liao, H., & Rupp, D. E. (2005). The impact of justice climate and justice orientation on work outcomes: A cross-level multifoci framework. Journal of Applied Psychology, 90, 242–256. Lind, E. A. (2001). Thinking critically about justice judgments. Journal of Vocational Behavior, 58, 220–226. Lipe, M., & Salterio, S. (2002). A note on the judgmental effects of the balanced scorecard’s information organization. Accounting, Organizations and Society, 27(6), 531–540. Locke, E. A., & Latham, G. P. (1990). A theory of goal setting and task performance. NJ: Prentice-Hall.

142

FRANK G. H. HARTMANN AND SERGEJA SLAPNICˇAR

Locke, E. A., Latham, G. P., & Erez, M. (1988). The determinants of goal commitment. Academy of Management Review, 13, 23–39. Moers, F. (2005). Discretion and bias in performance evaluation: The impact of diversity and subjectivity. Accounting, Organizations and Society, 30, 67–80. Rossiter, J. H. (2002). The C-OAR-SE procedure for scale development in marketing. International Journal of Research in Marketing, 19, 305–335. Schumacker, R. E., & Lomax, R. G. (1996). A beginner’s guide to structural equation modeling. Mahwah: Lawrence Erlbaum Associates. Smith, D., & Langfield-Smith, K. (2004). Structural equation modeling in management accounting research. Journal of Accounting Literature, 23, 49–84.

APPENDIX Questionnaire Items

1

1

1

1

2

2

2

2

3

3

3

3

4

4

4

4

5

5

5

5

objective information from the information system my performance in quantitative terms

on objective information from the information system my performance in quantitative terms

Objectivity of Reward Determination (OBJEC) Respondents were asked to indicate whether they agreed more with the statement on the right or on the left. My fixed pay is based on my supervisorus personal judgement of my performance My fixed pay is based on my performance in qualitative terms My bonus is based on my supervisorus personal judgement of my performance My bonus is based on my performance in qualitative terms

For each of the following questions, respondents were asked to indicate their level of agreement with the statements on a five-point Likert scale (1 ¼ I completely disagree, 2 ¼ I disagree, 3 ¼ neutral, 4 ¼ I agree 5 ¼ I completely agree).

Objective Rewarding, Managerial Motivation and Organizational Commitment

143

1 1 1

1 1 1

1 1 1

1 1 1

2 2 2

2 2 2

2 2 2

2 2 2

3 3 3

3 3 3

3 3 3

3 3 3

4 4 4

4 4 4

4 4 4

4 4 4

5 5 5

5 5 5

5 5 5

5 5 5

144

I have full confidence in the system’s fairness in determining targets. I have full confidence in the system’s fairness in evaluating performance. I have full confidence in the system’s fairness in determining pay.

Procedural Justice (PROCJ) procj1 procj2 procj3

My pay level reflects the effort I have put into my work. My pay is appropriate for the work I have completed. My pay reflects what I have contributed to the organization.

Distributive Justice (DISTRJ) distrj1 distrj2 distrj1

I enjoy my tasks as manager in this bank very much. The activities in my job are challenging and exciting to do. While working, I often think about how much I enjoy my work.

Motivation (MOTIV) motiv1 motiv2 motiv3

I do not feel a strong sense of belonging to my organization (reverse). I do not feel ‘emotionally attached’ to this organization (reverse). I do not feel like ‘part of the family’ at this organization (reverse).

Affective Organizational Commitment (COM) com1 com2 com3

FRANK G. H. HARTMANN AND SERGEJA SLAPNICˇAR

PART III: EVALUATING PERFORMANCE OF ORGANIZATIONAL UNITS

This page intentionally left blank

EMPIRICAL EVIDENCE ON THE (PERCEIVED) VALUE OF INFORMATION FOR PERFORMANCE MEASUREMENT PURPOSES IN AN ERPS ENVIRONMENT Xavier Gabrie¨ls and Ann Jorissen ABSTRACT This chapter investigates if and how the introduction of an Enterprise Resource Planning System (ERPS) influences the information characteristics quality, timeliness and complexity. Subsequently we analyze whether the influence of an ERPS adoption on these information characteristics has an impact on the perceived value of information available for performance measurement (PM) purposes. On the basis of the extant literature a structural model is developed which tries to capture the direct and indirect effects of the degree of ERPS adoption on the perceived value of the information available for PM.

Performance Measurement and Management Control: Measuring and Rewarding Performance Studies in Managerial and Financial Accounting, Volume 18, 147–175 Copyright r 2008 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-3512/doi:10.1016/S1479-3512(08)18007-8

147

148

XAVIER GABRIE¨LS AND ANN JORISSEN

INTRODUCTION Organizations constantly seek different ways and mechanisms to reduce uncertainty. Galbraith (1973) defined uncertainty as the gap between the amount of information that an organization possesses, and the amount of information required for a given level of performance. Reducing this gap leads to better decisions and subsequently to higher company performance. A mechanism suggested by Galbraith (1973) to close this information gap is an investment in a vertical information system. These vertical information systems will, according to Galbraith, increase the timing of the information flows, the scope of the data and the capacity to process information. Being developed in the late eighties and implemented in the nineties especially in large multinational companies, Enterprise Resource Planning System (ERPS) posses many characteristics of these vertical information systems, described by Galbraith in the early seventies. ERPS can be defined as ‘‘integrated software packages that control all personnel, material, monetary and information module-based flows of a company’’ (Granlund & Malmi, 2002) or as ‘‘enterprise wide packages that tightly integrate business functions into a single system with a shared database’’ (Quattrone & Hopper, 2005; Newell, Huang, Galliers, & Pan, 2003). These enterprise-wide database systems allow the integration of information between the various management areas of the organization, such as marketing, human resources management (HRM), accounting, inventory management, logistics and production. They are supposed to provide more real-time information for external parties through financial reporting, gather information that is useful for taking decisions more quickly and/or efficiently and facilitate (strategic) planning and control. Impact of these integrated information systems on the information available for performance measurement (PM) purposes is the focus of this chapter. The role and influence of ERPS on organizations has been studied from different perspectives in the extant literature. Several authors have studied the change processes which occurred in organizations implementing ERPS, others have focused on the impact of ERPS on organization design, management control and company performance, while still others have concentrated on the changing role of the accountant (Bradford & Florin, 2003; Caglio, 2003; Chapman, 2005; Dechow & Mouritsen, 2005; Granlund & Malmi, 2002; Hyvo¨nen, 2003; Lodh & Gaffikin, 2003; Scapens & Jazayeri, 2003; Nicolaou, 2004; Quattrone & Hopper, 2005; Dechow & Mouritsen, 2005).

Empirical Evidence on the (Perceived) Value of Information

149

An ERPS comprises a set of integration applications modules, which span most business functions. Given this feature of ERPS, we want to explicitly measure the extent of adoption of ERPS in an organization by taking into account the number of modules which have been implemented. This is done by defining the degree of ERPS adoption as the number of modules adopted. In this manner we integrate the extent to which a change has been integrated into the company. Building on the results of the extant IS and management control literature, the influence of (user) implementation satisfaction and centralization on the perceived value of information for performance evaluation is taken into consideration. We also take into account the influence of the degree of ERPS adoption on the characteristics of the information available in the company. Therefore, this chapter examines the relationships between the degree of ERPS adoption, information characteristics, centralization, (user) implementation satisfaction and the perceived value of the available information for PM. The majority of prior literature investigating the impact of ERPS on an organization is mainly based on case studies. We will however use the survey method, in order to gain more widespread evidence. The research data were collected through a survey which was sent to the 150 largest companies in Belgium. Based on the extant management accounting and control as well as the IS literature, hypotheses were developed and a structural model was derived in order to measure the impact of the degree of ERPS adoption on the perceived value of information for PM. The structural model is estimated with the use of a Partial Least Squares (PLS) path model. The remainder of this chapter is organized as follows. First, a theoretical framework which supports the empirical analysis will be formulated on the basis of the existing literature. Second, the data collection method will be presented together with the measurement of the variables and the method of analysis used. In the third part, the results of the path analysis will be presented and discussed. Finally, the chapter ends with a conclusion, the limitations and avenues for future research.

LITERATURE REVIEW AND DEVELOPMENT OF RESEARCH HYPOTHESES Extant research in different disciplines has illustrated the importance of Accounting Information Systems (AIS) to facilitate decision management as

XAVIER GABRIE¨LS AND ANN JORISSEN

150

well as decision control (Sprinkle, 2003; Libby & Waterhouse, 1996; Zimmerman, 2000; Abernethy & Vagnoni, 2004). Although ERPS can also be considered as AIS, they relate to all areas of management and their possible effects are supposed to have greater impact than their ‘‘legacy’’ predecessors (Bradford & Florin, 2003). Fisher (1998) found that management control systems (MCS) are, among others, dependent on technology. The research model, presented in Fig. 1, includes both direct, as well as indirect, effects of technology on the value of information for PM. MCS can be defined as the formalized procedures and systems that use information to maintain or alter patterns in organizational activity (Simons, 1990). This chapter will focus on one aspect of MCS, namely PM. The goal of PM is to implement strategy (Anthony & Govindarajan, 2005). PM affects employees’ incentives to increase shareholder value (Indjejikian, 1999). As the choice of performance measurement system (PMS) is consequently one of the most critical challenges an organization faces, much research has focused on the design, effects and use of PMS (Jensen & Meckling, 1976; Indjejikian, 1999; Ittner & Larcker, 1998; Simons, 1990; Vandenbosch, 1999; Bisbe & Otley, 2004; Henri, 2006). However, in order to use a PMS, one needs information, which is perceived as valuable by organizational participants. Therefore, this chapter focuses on the perceived value of the information provided by ERPS for PM.

-Quality - Timeliness - Complexity Information Characteristics

- Training - Setup satisfaction (User ) implementation satisfaction

H4

H3 H5

Degree of ERPS adoption

H2

H1

Information value for PM C3

C1

C2 Centralization

Fig. 1.

Research Model to Guide the Framework of Analysis.

Empirical Evidence on the (Perceived) Value of Information

151

Demski and Feltham (1976) already illustrated that employee behavior and motivation can be affected by managerial accounting information since managerial accounting information plays a role in monitoring, measuring, evaluating and rewarding actions. A common problem is finding and having access to accurate data to monitor, measure, evaluate and reward actions. This gap between the information necessary and the information available to control employee behavior results in uncertainty. An increase in the capacity to handle information is one way to reduce this uncertainty1 (Galbraith, 1973). ERPS are supposed to exponentially increase the organizations information processing capabilities and therefore reduce the uncertainty associated with performance evaluation (Scapens, 1994; Demski & Feltham, 1976; Galbraith, 1973; Bouwens & Abernethy, 2000). According to agency theory, people do not act in the organizations best interests, but rather in their own (Abernethy & Vagnoni, 2004; Mas-Colell, Whinston, & Green, 1995). ERPS may also serve as an ex-post control mechanism which provides information to ensure that the agent’s action choices are consistent with organizational goals (Mauldin & Ruchala, 1999). By including the number of modules adopted in the organizations as an independent variable, we take explicitly into account the extent to which ERPS has been adopted by the organization. The extent of module adoption should lead to more seamless integration, enhancing consistency of and access to information, and consequently reduce uncertainty. This results in a first hypothesis, in which we investigate the direct impact of the ERPS introduction on value of information available for PM purposes. H1. As the degree of ERPS adoption increases, the perceived value of information for PM enhances. The MIS literature and the literature on the adoption of ERPS and ABC systems informs us that the attitude of the individual towards an innovation is an important intervening variable which determines whether or not the individual will use the innovation. In this case, the attitude of the individual might influence whether or not the individual will act upon the information provided by the ERPS (Foster & Ward, 1994; Lawrence, 1954; Scapens, 1994; Burns & Scapens, 2000; Granlund, 2001). Literature shows that resistance can negatively impact the use of an innovation. Research results provide evidence that resistance can be overcome by training and involvement of the users in the design and the implementation of the actual new system2, so as to ensure that users can identify and commit themselves with/to the new implementation (Abernethy & Bouwens, 2005; Jermias, 2001; McGowan & Klammer, 1997; Krumwiede, 1998; Hartwick & Barki,

152

XAVIER GABRIE¨LS AND ANN JORISSEN

1994; Cooper, Kaplan, Maisel, Morrissey, & Oehm, 1992; Watson, Rainer, & Houdeshel, 1997; DeLone & McLean, 2003). An adequate implementation process is therefore absolutely necessary (Clemons, Thatcher, & Row, 1995). Otherwise users might stay too much in their old way of thinking, which would destroy the intended advantages of the new system (Bagranoff & Brewer, 2003). Furthermore, user (implementation) satisfaction is driven by the satisfaction with implementation resources. Resources can be defined as money, people and time that are required to successfully complete the project (Ein-Dor & Segev, 1978). Wixom and Watson (2001) found support for the relation between a high level of resources and implementation success in their study on data warehousing success. These observations from the literature lead to the following hypothesis. H2. User (implementation) satisfaction has a positive effect on the perceived value of information for PM. In the academic as well as in the professional literature it is stated that ERPS change the characteristics of the data available in a company. The basic idea is that ERPS makes real-time availability of information possible. More timely information has the potential to reduce uncertainty (Bouwens & Abernethy, 2000). However, better (informed) management control is also dependent upon the quality of the information3 and not only the amount of data. Information quality is defined as the degree to which the information meets the needs of his users (Seddon & Kiew, 1994), and can consequently be considered an important aspect in closing the ‘‘information gap’’ and reducing uncertainty (Galbraith, 1973). The centralized coordination and high integration of information under ERPS are supposed to have a positive influence on the quality of the information. The importance of information quality and timeliness is also substantially documented in the MIS literature4 (Raghunathan, 1999; DeLone & McLean, 1992; Wixom & Watson, 2001; Khalil & Elkordy, 2005; Teo & Wong, 1998; Seddon, 1997). ERPS are typically considered complex (Bingi, Sharma, & Godla, 1999). Complexity is defined by Rogers (1983) as the degree to which a certain innovation is difficult to understand and use. Due to the information increase, it may become impossible for individuals to process all the information available (Tiessen & Waterhouse, 1983). Furthermore, it is also important to limit the (perceived) complexity (of information retrieval), as complexity in general can destroy the advantages of the reduced uncertainty and ultimately lead to resistance due to the lack of skills and knowledge (Rogers, 1983). Argyris (1977, p. 113) already identified the complexity associated with MIS-implementations as one plausible explanation for the

Empirical Evidence on the (Perceived) Value of Information

153

often-unmet a priori expectations. Granlund and Malmi (2002) observed that system complexity could be one of the factors explaining their general observation that ERPS had limited impact on managerial accounting. The above discussion leads to the following hypothesis with regard to the characteristics of the information: H3. The degree of ERPS adoption improves the quality of information (H3a), positively influences the timeliness of information (H3b), but also increases the complexity of information retrieval (H3c). The importance of user (implementation) satisfaction for the perception about the value of the information for PM purposes has already been highlighted, when H2 was developed. However, evidence is available in the literature that user (implementation) satisfaction may alter the perceived characteristics of the information available. The recent study of Khalil and Elkordy (2005) illustrates the importance of continuous training and orientation programs in order to improve the information quality and timeliness. Given the importance of the attitude of the individual, we will also test the following hypothesis: H4. User (implementation) satisfaction positively influences the information characteristics quality, timeliness and complexity. Although the information may be more quickly available and of higher quality, people have to perceive this information as valuable for performance evaluation purposes (Argyris & Kaplan, 1994, p. 84). Information quality is believed to be one of the most important characteristics that determine the degree to which information is perceived as valuable (O’Reilly, 1982). Therefore, we assume that the perceived value of PM is affected by the quality of information. If information is of high-quality, timely and not utterly complex, managers are more likely to rely on it for decisions making or PM purposes (Low & Mohr, 2001). Gelderman (1998) found high correlations between ease of use5, timeliness and frequency of usage. A number of IS studies have investigated the influence of perceived easy of use on the actual use and usefulness of information for decisionmaking and performance evaluation and obtained mixed results (Davis, 1989; Davis, Bagozzi, & Warshaw, 1989; Adams, Nelson, & Todd, 1992; Igbaria, Zinatelli, Cragg, & Cavaye Angele, 1997). It seems logical to hypothesize that systems where information retrieval is less complex, of higher quality and quickly available, can satisfy more information needs. Consequently these systems’ uncertainty and the information is perceived to be of higher value.

154

XAVIER GABRIE¨LS AND ANN JORISSEN

Therefore, we will test the direct influence of the information characteristics on the perceived value of information for PM. So we include the following hypothesis. H5. The information characteristics quality of information and timeliness have a positive effect on the perceived value of ERPS-provided information for PM (H5a & H5b). The increased complexity of information retrieval negatively affects the perceived value of ERPS-provided information available for PM purposes (H5c). As ERPS are characterized by a centralized structure and it is advisable to adapt the organization to the system, instead of the other way around (Hyvo¨nen, 2003; Abernethy & Bouwens, 2005), a decentralized organization is expected to experience much more changes after an ERPS adoption. However, following Huber (1990, p. 57), we are not pretending that a decentralized organization will become more centralized (and vice versa). It may very well be that the increased information will enable a (de)centralized organization to operate in an even more (de)centralized manner. Therefore, we will control for centralization by including the following relationship in our model: Control 1 (C1). The degree of ERPS adoption influences the degree of centralization6. However, no expectation is put forward regarding the sign of the relationship. Centralization relocates responsibility and access to information to management at headquarters (Chenhall & Morris, 1986). In centralized organizations, local management has minimal decision authority in the design of reward systems. The manager’s preference for control over the factors he is being assessed on, entails his inclination for decentralized reward structures (Merchant & Manzoni, 1989, p. 554). Therefore, one could expect that managers in more decentralized organizations embrace easier possible changes to the PMS. As opposed to their colleagues in more centralized management structures, they can participate into the actual decisions regarding the design of the PMS. It seems logical that participation in developing their own assessment measurement leads to higher perceived value of information. This leads to a second control relationship: Control 2 (C2). Centralization has a negative effect on the perceived value of information for PM.

155

Empirical Evidence on the (Perceived) Value of Information C3a Centralization

Quality H3a

Degree of ERP adoption

C3c H5b

Timeliness H3b

H5a

H4b1 H1 H 2b4

H3c H4a1

C2

C3b

C1

H5c

Information value for PM

Complexity H4a2

H4a3 H4b3

H2b

H2a Setup satisfaction

Training

(User) implementation satisfaction

Fig. 2.

Structural Model.

Given the importance attached in previous literature to the influence of centralization on the information characteristics (Chenhall & Morris, 1986; Bouwens & Abernethy, 2000; Gosselin, 1997), we will also test a last control relationship: Control 3 (C3). Centralization has a positive influence on the quality of information (C3a), timeliness (C3b) and the (perceived) complexity of the system (C3c). Taking into account the findings of prior literature and the hypotheses developed, Fig. 2 presents in a graphical way the hypothesized relationships between the different variables studied.

THE RESEARCH METHOD Data Collection and Sample Description Given the high cost level involved with the implementation of an ERPS, only large companies have enough resources to implement a full ERPS planning system. Therefore we have chosen as the research population for the survey the 150 largest companies of Belgium. These companies all have a

156

XAVIER GABRIE¨LS AND ANN JORISSEN

personnel staff above 350 fulltime equivalents and a sales volume greater than 200 million euro. The choice of the population was made on the basis of the Belfirst database, which includes all companies which have to publish their annual accounts in Belgium. A total of 150 companies was felt adequate as augmenting the sample selection with 50, revealed that those companies were almost all non-ERPS adopters. After correcting for double listings7, we were left with 137 usable entries. All 137 companies were primarily telephoned to ask whether or not they had an ERPS installed such as SAP, Oracle or JD Edwards. Based on this question, 95 companies were identified as ERPS adopters. Possible respondents were identified from the Trends database, which lists the accounting and finance manager in each of the companies selected. The initial questionnaire was pilot-tested by ten accounting and IT-professionals8, as well as three academics. Necessary revisions were made in order to maximize understanding of the questions by the target respondents. To maximize our response rate, respondents were offered several modes of response (Pinsonneault & Kraemer, 1993; Sudman & Blair, 1999). Respondents, with publicly known e-mail address, were contacted by e-mail. In addition to providing an online link to the survey, a portable data format (PDF) version of the survey – which could be mailed (postal) or faxed back – was added in attachment. Respondents of whom the e-mail address was not publicly available9 were contacted by postal mail. This mail included the possibility to fill in the questionnaire online. A cover letter explaining the focus of the research and the need for cooperation was included in each case (Dillman, 2000). Furthermore, the target respondent, accounting and finance manager10, was given the opportunity to pass on the survey in case he felt he was not the appropriate person to answer the questions. Forced response was turned on in the online questionnaire, in order to avoid missing values. A month after the initial mail, a follow-up mail, which elaborated on the importance of the respondent’s contribution to this research project, was sent out. A total of 47 responses was returned. Unfortunately, due to missing answers, five companies had to be deleted, leaving us with 42 usable responses. This process yielded a usable response rate of 44.21%. As people were given the opportunity to pass on the questionnaire, we received in total 13 answers from IT-managers. A Mann– Whitney U-test was run to verify if this difference in response profile yielded any bias. No statistically significant differences were found at the 0.05 level. Regarding education, only 6 respondents did not have an academic degree, from the 35 respondents who did have a university degree, 14 had

Empirical Evidence on the (Perceived) Value of Information

157

a post-graduate degree (Masters, MBA or PhD) also. A Mann–Whitney U-test11 did not reveal any significant differences concerning education. Out of the 42 respondents, 37 used the SAP system, 3 used Oracle, 1 Baan and 1 JD Edwards. Out of the 37 SAP users, 3 also indicated to use Oracle as well. No significant differences in answers concerning the package used could be detected. Finally, to test for non-respondent bias, again a Mann–Whitney U-test12 was run to identify significant differences between early and late respondents (Brownell, 1995; Moore & Tarnai, 2002). No significant differences could be found. Furthermore, follow-up telephone calls were made to a selection of non-respondents to verify their reasons for non-response. Most reasons mentioned concerned current mergers going on, the absence of a clear IT-structure at present, the outsourcing of accounting and/or sometimes the ERPS-database to a coordination center.

The Survey Contents and the Measurement of the Research Variables The survey started with questions asking for the respondent’s function, education and hierarchical position. In the group of ERPS adopters some specific information about the ERPS in use was asked for. Questions included the ERPS they used, the reasons for adoption, the implementation strategy and the introduction of other management accounting techniques such as balanced scorecards or activity-based costing. The remainder of the survey contained questions with regard to the characteristics of the information available in the company as well as the perceived value of this information for PM purposes. Based on the survey responses, the following research variables were constructed.

The Degree of ERPS Adoption The implementation of an ERPS is typically organized around the modules of the system which will be adopted (Granlund & Malmi, 2002). To capture these differences among firms in the survey we measured the construct ‘‘degree of ERPS adoption as the number of relevant modules installed in each company’’. Respondents were asked to indicate which of the following 12 modules they had installed in their local subsidiary: budgeting and planning, performance evaluation, procurement, sales and marketing, inventory management, supply chain management (SCM), manufacturing

158

XAVIER GABRIE¨LS AND ANN JORISSEN

and product management, human resources management (HRM), customer relationship management (CRM), service and maintenance, e-commerce applications, and business intelligence and analytics. The division into 12 modules was based on a comparison between the module-based structure of SAP and Oracle, to accommodate users of both packages equally. The respondent scores for this item were then compared to the maximum number of modules each company could install given the business it is in. For example, a service organization does not always need a manufacturing and production module13. The score resulting from the division of the number of modules installed and the modules considered relevant for that industry was used to measure the construct ‘‘ERPS introduction’’. With regard to the measurement of the other dependent and independent variables, we used instruments published in the literature as often as possible in order to enhance the reliability and validity of our measures. Unless mentioned otherwise, each of the variables below was measured using a fivepoint Likert scale, ranging from 1 (strongly disagree) to 5 (strongly agree). After checking the factorability of items14, the design of all measurement instruments was based on the results of principal component analysis with Varimax rotations (Table 1). Summated scales for the factors identified15 were constructed and Cronbach (1951) alpha statistics were used to verify the internal reliability of the new scale (Nunnally, 1978).

(User) Implementation Satisfaction A new instrument of six items, based on previous literature, was initially used to verify the level of (user) implementation satisfaction. The first three items asked respondents to indicate the adequacy of implementation budgets and time as well as knowledge and expertise on a five-point Likert scale ranging from 1 (not at all sufficient) to 5 (abundant). These items were added based on the results of a case study conducted in two subsidiaries of large multinational organizations (Gabrie¨ls, 2006). The following item was derived from the resistance to MAS change measure from Abernethy and Bouwens (2005) and reflects the satisfaction with change management. The satisfaction with training was taken from McGowan and Klammer (1997). Factor analysis revealed that the six items loaded on two components, explaining in total 66% of the total variance. The first component is interpreted as satisfaction with the setup (setup satisfaction) and explains about 33% of the total variance. The Cronbach’s alpha is 72.6%.

159

Empirical Evidence on the (Perceived) Value of Information

Table 1.

Survey Items and Results from the Rotated (Varimax) Factor Component Analysis. I

A. (User) implementation satisfaction Budgets during implementation Time to complete the implementation Knowledge and expertise during implementation I believe that the implementation team has done a great job in bringing about the change in IS. People in this organization have received proper training and orientation about the ERPS after the implementation. Post-implementation training is also satisfactory Setup satisfaction Training B. Information characteristics Information is provided real-time It is easy to get in-between months data Information is quickly available I am satisfied with the overall accuracy of the information I receive from the ERPS The information provided by the ERPS usually meets my needs The information provided by the ERPS supports this division’s operations well Given the goals this division tries to meet, the information provided by the ERPS is relevant The overall quality of the information provided is high It is easy to get the information I need out of the information system Learning to use the ERPS has been easy for me Overall, the ERPS is easy to use Interpretation of the information provided by the ERPS is easy for me Interpretation of the information provided by the ERPS is easy for my superior Quality Complexity Timeliness

II

.770 .761 .400

.010 .331 .516

.826

.089

.052

.862

.098

.921

III

a ¼ 72.6% a ¼ 82.5% .323 .379 .494 .729

.009 .056 .218 .235

.791 .837 .702 .254

.810

.241

.296

.811

.007

.178

.798

.293

.000

.750

.152

.387

.247

.774

.227

.113 .214 .440

.814 .629 .678

.017 .396 .108

.395

.638

.033

a ¼ 87.9% a ¼ 72.5% a ¼ 90.4%

160

XAVIER GABRIE¨LS AND ANN JORISSEN

Table 1.

(Continued ). I

C. Centralization Strategic decisions Human resource decisions Investment decisions Budget allocations Marketing decisions Centralization D. Perceived value of information for performance measurement (PM) The information we receive is suited to support the performance measures The measurement of performance is objective and verifiable The performance measures provide a good indication of my effort The performance of this division is similar to divisions from competing firms Information value for PM

.803 .626 .816 .835 .719 a ¼ 82% .787 .900 .713 .508 a ¼ 76%

The second component also explains also about 33% of the variance and loads on the two items representing training, and is consequently named training. The Cronbach alpha for this construct was 82.5%.

Information Characteristics The characteristics of the information available were measured by including instruments that tried to capture the quality, timeliness and complexity of this information. Five questions were identified from the quality of information measure used by Abernethy and Bouwens (2005), which was based on Ives, Olson, and Baroudi (1983). These questions related to the quality of the information available and asked for the accuracy as well as the relevancy of the information and whether the information provided support for the respondent’s needs as well as for their division’s operations. Based on the results of a case study conducted in two subsidiaries of large multinational organizations, three questions were added to assess the timeliness of the information (Gabrie¨ls, 2006). Five questions regarding the perceived complexity were adapted from the Bradford and Florin (2003) study. Respondents were asked to identify the easiness of information retrieval as well as the easiness of becoming familiar with the new system and its use. Furthermore, they had to indicate to what extent it was difficult to interpret the information available.

Empirical Evidence on the (Perceived) Value of Information

161

Three components were identified explaining 68% of the total variance. Four items representing quality of information loaded significantly on a first component, explaining 28.5% of the total variance. A summated scale was labeled accordingly and internal reliability analysis revealed a Cronbach alpha of 87.9%. A second component represented the easiness of information retrieval (two items loaded high on this component) and explained 20% of the total variance. After reverse coding these two items, this component was labeled (perceived) complexity (of information retrieval). The Cronbach alpha is 72.5%. A third component could be interpreted as timeliness of the information as two items representing this aspect loaded high (W.79) on this component. This component explained 19.5% of the total variance and it’s internal reliability is 90.4% (Cronbach alpha).

Degree of Centralization The degree of centralization was measured using the instrument developed by Gordon and Narayanan (1984). Respondents were asked, on a fivepoint Likert scale, ranging from 1 (my subsidiary has all the influence) to 5 (headquarters has all the influence), to indicate the level of centralization with regard to the following decisions: strategic, human resource, investment, budget and marketing. All items loaded on one component, explaining 58.3% of the total variance. The factor loading for item 2 (human resource decisions) was below .70, so it was dismissed from further analysis. A summated scale was created for the remaining four items and labeled ‘‘centralization’’. The Cronbach’s alpha is 81.6%.

Performance Measurement Four questions regarding the usefulness and the adequacy of the information for PM purposes were identified from previous research (Shields, 1995; Moers, 2001) and adapted for the research study at hand. All four items loaded on one component, explaining approximately 55% of the total variance. The factor loading for the fourth item was rather low, so it was dismissed from further analysis. As all remaining three items represented the perceived value of the information available for PM, the summated scale resulting from these three items was named information value for PM. The Cronbach alpha is 75.8%.

XAVIER GABRIE¨LS AND ANN JORISSEN

162

Method of Analysis Given our sample size, our hypotheses will be tested with path analysis using the PLS technique. As the factor analysis identified new constructs and prior knowledge hereabout is rather scarce, this technique seems appropriate. Furthermore, PLS also requires fewer assumptions to be met regarding measurement scales and residual distributions (Chin, 1998; Barclay, Higgins, & Thompson, 1995). Apart from the assessment of the structural model that identifies relations among constructs, PLS also allows confirmation of the above described factor and reliability analysis,16 as it also measures relations between manifest items (observed values for each survey question) and the latent constructs they represent (the components identified by factor analysis).

RESEARCH RESULTS From the descriptive statistics in Table 2 we learn that on average 62.6% (standard deviation: 18.02%) of the total available ERPS modules are installed in the companies under study. Respondents are very satisfied with the quality (15.3/20) and timeliness (7.5/10) of the information and are rather indifferent about the complexity of information retrieval (5.2/10). Satisfaction with training (7.2/10) may explain this indifference. As setup satisfaction scores higher than average (9.0/15), it can be assumed that change management and resources during setup where also satisfactory. Decision authority is almost equally Table 2.

Descriptive Statistics of the Variables (N ¼ 42).

Variable

Degree of ERP adoption Quality Timeliness Complexity Setup satisfaction Training Centralization Information value for PM

Mean

62.64% 15.29 7.50 5.21 9.05 7.14 13.45 10.60

Standard Deviation

18.02% 2.71 1.88 1.70 1.94 1.55 3.67 1.91

Actual Range

Theoretical Range

Min

Max

Min

Max

33% 8 2 2 4 4 4 5

100% 20 10 8 13 10 20 14

0% 4 2 2 3 2 4 3

100% 20 10 10 15 10 20 15

163

Empirical Evidence on the (Perceived) Value of Information

Pearson Correlation Coefficients (N ¼ 42).

Table 3. Variable

1

2

3

4

5

6

7

8

1. Degree of ERP n.a. adoption 2. Quality .107 0.863 3. Timeliness .195 .542 0.955 .202 0.882 4. Complexity .234 .299 5. Setup satisfaction .031 .634 .530 .247 0.804 6. Training .119 .268 .134 .214 .281 0.923 7. Centralization .156 .036 .186 .015 .186 .003 0.808 8. Information value .148 .414 .241 .294 .387 .405 .242 0.822 for PM ,, refers to po0.10, o0.05, o0.01, respectively, and square root of variance extracted

(diagonally).

distributed between the local subsidiary and headquarters, with a slight prevalence for headquarters as the score for centralization is 13.45/20. The information value for PM is perceived above average (10.6/15). The results do not indicate high multicollinearity (Table 3)17 (Tabachnick & Fidell, 2001, p. 84). Next, we focus on the estimation of the structural model that we developed on the basis of the extant literature. As Table 4 shows, composite reliability (CR) of latent variables (cfr. Cronbach alpha) is for each latent variable estimated always above 0.84 (Chin, 1998), confirming sufficient construct reliability. Furthermore, average variance extracted (AVE) is for all latent variables in each model significantly above .50, meaning that more than 50% of the variance of the indicators is accounted for and confirming adequate convergent validity (Fornell & Larcker, 1981). Barclay et al. (1995) describe that adequate discriminate validity is attained when the square root of the AVE (diagonal entries in Table 3) is greater than the off-diagonal entries. As illustrated in Table 3, this condition is satisfied. Loadings of manifest variables on latent variables are ranging from .73 to .96 and are usually above .80, so may be considered high. This confirms the adequacy of the factor and reliability analysis described above. Fig. 2 illustrates the structural model together with the hypothesized relations between the different latent constructs that will be tested. Fig. 3 presents the significant paths in this structural model. Path coefficients, t-statistics and significance levels, as well as R2 are displayed only if the path is significant. Significance of the path coefficients was

XAVIER GABRIE¨LS AND ANN JORISSEN

164

Table 4.

Factor Loadings of Manifest Items on Latent Variables in PLS-Model.

Manifest Items

Loadings on Latent Variables (LV)

ERP adoption (%)

1.000

CR; (AVE): Degree of ERP adoption The information provided by the ERPS usually meets my needs The information provided by the ERPS supports this division’s operations well Given the goals this division tries to meet, the information provided by the ERPS is relevant The overall quality of the information provided is high

1.00; (1.00) 0.859 0.878

0.897

CR; (AVE): Quality Information is provided real-time It is easy to get in-between months data

.921; (.745) 0.951 0.960

CR; (AVE): Timeliness It is easy to get the information I need out of the information system Learning to use the ERPS has been easy for me

.954; (.913) 0.928 0.833

CR; (AVE): Complexity Sufficiency of budgets during implementation Sufficiency of time to complete the implementation I believe that the implementation team has done a great job in bringing about the change in IS

.875; (.778) 0.784 0.764 0.862

CR; (AVE): Setup satisfaction People in this organization have received proper training and orientation about the ERPS after the implementation Post-implementation training is also satisfactory

.846; (.647) 0.909 0.937

CR; (AVE): Training Strategic decisions Investment decisions Budget allocations Marketing decisions

.920; (.852) 0.821 0.846 0.827 0.735

CR; (AVE): Centralization The information we receive is suited to support the performance measures The measurement of performance is objective and verifiable The performance measures provide a good indication of my effort

.883; (.653) 0.833

CR; (AVE): Information value for PM

0.817

0.891 0.736 .862; (.676)

165

Empirical Evidence on the (Perceived) Value of Information -0,087** (2,087) Centralization Quality

0,075*** (4,328)

R²= 0.472

R²= 0.008 -0,317*** (7,266)

0,162* (1,848)

Timeliness Degree of ERP adoption 0,074*** (2,911)

0,090*** (3,792)

0,088* (1,930)

Training

-0,095*** (3,315)

R²= 0.324

-0,109* (2,020)

Complexity 0,667*** (11,883)

-0,145* (1,810) 0,292*** (5,935)

Information value for PM

R²= 0.104 0,547*** (9,643)

-0,247*** (3,566)

R²= 0.394

0,195** (2,429)

Setup satisfaction

Fig. 3. PLS Results for Structural Model: Path Coefficients and (t-Values); ,, refers to po0.10, o0.05, o0.01, Respectively (Only Statistically Significant Paths and Their Path Coefficients are Shown).

assessed by a bootstrapping procedure using 200 samples with 200 cases per sample. As Fig. 3 and Table 5 illustrate, the expected direct influence of the degree of ERPS adoption on the perceived value of information for PM is indeed significant (0.095, po0.01; H1). However, contrary to expectations, the direction is negative, meaning that the installation of more modules leads to less valuable information for PM purposes. H2 is supported, as both training (0.292, po0.01; H2a) and setup satisfaction (0.195, po0.05; H2b) load significantly positive on the perceived information value for PM. H3 is also confirmed, as the degree of ERPS adoption has a significantly positive effect on the information characteristics quality (0.075, po0.01; H3a), timeliness (0.090, po0.01; H3b) and complexity (0.074, po0.01; H3c). Although we do not find a significant relationship between the information characteristic timeliness and the value of information for PM (H5b), both other information characteristics load significantly on the information value for PM. In line with the hypotheses, quality loads significantly positive on the information value for PM (0.162, po0.10; H5a) whereas complexity loads significantly negative on the information value for PM (0.109, po0.10; H5c). This confirms the expected indirect relationship between the

Path to

Table 5.

R2 ¼ 0.008

Centralization

R2 ¼ 0.472

Quality

0.090 (3.792) 0.020 (0.3945) 0.547 (9.643) 0.091 (1.468)

R2 ¼ 0.324

Timeliness

0.074 (2.911) 0.145 (1.810) 0.247 (3.566) 0.052 (0.0578)

R2 ¼ 0.104

Complexity

R2 ¼ 0.394

Information Value for PM

Results from PLS Analysis ‘‘Information Value for PM’’ (N ¼ 42).

0.087 (2.087)

0.075 (4.328) 0.088 (1.930) 0.667 (11.883) 0.066 (1.323)

0.095 (3.315) 0.292 (5.935) 0.195 (2.429) 0.317 (7.266) 0.162 (1.848) 0.062 (1.048) 0.109 (2.020)

166

Path from Degree of ERP adoption Training Setup satisfaction Centralization Quality Timeliness Complexity ,, refers to po0.10, o0.05, o0.01, respectively.

XAVIER GABRIE¨LS AND ANN JORISSEN

Empirical Evidence on the (Perceived) Value of Information

167

degree of ERPS adoption and the perceived value of information for PM. The data supports H4 which hypothesized a relationship between user satisfaction and the information characteristics. Setup satisfaction loads significantly on all three information characteristics (0.667, po0.01: H4b1 0.547, po0.01: H4b2 and 0.247, po0.01: H4b3) and training on quality and complexity (0.088, po0.10: H4a1 and 0.145, po0.10: H4a3), but not on timeliness (H4a2). The significant relationships are, as expected, positive for quality and timeliness and negative for complexity. As both quality and complexity (vide supra) significantly influence the information value for PM, an indirect path between training and setup satisfaction via the information characteristics on the perceived value of information is established. Fig. 3 illustrates the significant and negative relation between the degree of ERPS adoption and centralization (0.087, po0.01; C1). Furthermore, we do observe a significant negative relationship between centralization and the perceived information value for PM (0.317, po0.01; C2), providing support for a second indirect relationship (through centralization) between the degree of ERPS adoption and the perceived value of information for PM purposes. No significant relationship could be found between centralization and any of the three information characteristics. Consequently, control relationship 3 is not supported in this data model. Although the direct influence of the degree of ERPS adoption leads to information that is less valuable for PM purposes, both indirect relationships of the degree of ERPS adoption report a positive effect on the perceived information value forPM. The results of the direct relationships show that the degree of ERPS adoption leads to better information quality, more timely information and more complexity. Although, better information quality results in more valuable information for PM, this effect is tempered by the increased complexity associated with the new system. Further, our model finds a reduction in centralization as more ERPS modules are installed. The data indicate that when more decisions are decentralized, information under an ERPS system is perceived of higher value for PM purposes. This confirms our theoretical argument that the perceived value of information for PM increases when local management can participate in the choice of the performance indicators and system (C2). The importance of adequate change management, through sufficient resources and enough training on the perceived information value for PM is confirmed by the results. Apart from the absence of a relationship between training and timeliness, which seems rather logical, as the timeliness of information can be considered more system related, setup satisfaction and training are considered very important to improve quality, reduce

168

XAVIER GABRIE¨LS AND ANN JORISSEN

complexity and directly and indirectly improve the information value for PM. Since all our hypotheses are confirmed (except for H4a2 & C3) and an overall R2 of nearly 40% is reached, we conclude that our model fits the theory rather well. Of course, additional refinements can be made to enhance our understanding of the phenomenon under study.

CONCLUSION The general aim of this chapter was to assess the impact of the degree of ERPS adoption on the perceived value of information available for PM. Extending prior research, we developed a structural model which aims to explain the direct and indirect influence of the degree of ERPS adoption on the perceived value of information available for PM. We measured the degree of ERPS adoption by taking into account the number of modules installed in a company. The path model was estimated with the use of the PLS method and the data were obtained through a survey sent to the 150 largest companies in Belgium. The results of the path analyses indicate that the direct impact of degree of ERPS adoption has a negative influence on the perceived value of information for PM. So we observe, contrary to expectations, that more modules lead to less valuable information for PM. This negative influence could be explained by agency theory. Increased information availability brings about improved transparency. In case reward systems are tied to this more transparent information, the agent may oppose the new system, as he may be opposed to the availability of this transparent information. Although there is a direct negative impact of the degree of ERPS adoption on the perceived value of information for PM purposes, the data reveals that the indirect influence through the information characteristics is positive. The results indicate that the degree of ERPS adoption leads on the one hand to more timely information and to information of better quality, but on the other hand it creates more complexity. The results of the model show that this enhanced quality of information leads to information that is perceived more valuable for PM. This positive effect is somewhat tempered by a negative impact of complexity. The results show that this complexity can be diminished by the availability of sufficient resources which accompany the setup of the system. As a result, the study confirms that behavioral aspects, such as training and setup satisfaction, should not be neglected. These behavioral aspects have a direct positive influence on the perceived value of information for PM. Furthermore, setup satisfaction (sufficient resources

Empirical Evidence on the (Perceived) Value of Information

169

and adequate change management) has a positive influence on the timeliness and quality of information and training enhances the quality of information. The results indicate an influence of the degree of ERPS adoption on centralization. As the relationship is negative, introducing ERPS and installing more modules leads to less centralization. Apparently, as more ERPS modules lead to more timely and better quality information, organizational transparency is increased. This could allow for more decentralization to cope with local uncertainties. Finally, the results also reveal the expected negative influence between centralization and the perceived value of information for PM, as less centralization brings about more local influence on performance measures and consequently entails acceptance and (perceived) value of information for PM purposes. In line with the theory of Galbraith (1973), the study confirms that investing in ERPS indirectly improves the value of the information for PM purposes, as perceived by the employees in the organizations under study. ERPS reduce uncertainty and thus investing in this information processing capacity increases at least indirectly the perceived value of the use of this information in the context of management control. Behavioral aspects, such as setup satisfaction and training, should however not be neglected. This study is subject to several limitations. In spite of a careful administration of the survey, it is impossible to rule out all possible alternative plausible explanations for the results, since one only possesses a limited amount of data on the various variables, that may co-vary with the variables in question (Abernethy, Chua, Luckett, & Selto, 1999, pp. 18–19). This is an inherent limitation of survey research. Therefore, threats to internal validity always remain and they impede the drawing of causal relationships, so only associations can be reported in this chapter. However, our PLS models indicate a good fit and variance explained of the dependent variables by the independent variables. As we sometimes ask self-assessment of the respondents, this may also bias our results. However, Venkatraman and Ramanujan (1987) found support for the proposition that managers’ self-ratings of performance are less biased than researchers typically give them credit for. Nevertheless, our findings could benefit from an extension of the respondents per organization, just as from an extension of our research population. This extension could include more subsidiaries of the same organization, or more organizations in different countries. Although we investigate the perceived value of information for PM, this information not necessarily leads to more efficient PM. Therefore, further research could investigate to what extent this information actually leads to changes to the design and efficiency of PM.

170

XAVIER GABRIE¨LS AND ANN JORISSEN

NOTES 1. The other option is to reduce the need for information by organizing for that. 2. Of course, there are numerous other factors during the design and implementation phases that determine the resistance to the ERPS afterwards. Jermias (2001, p. 157) warns in this regard against the significant role some purely technical and/or economical factors might play in prohibiting the adoption of the ‘‘new’’ system. 3. ‘‘The quality of IS information refers to the reliability, relevance, accuracy, precision and completeness of IS information (Bailey & Pearson, 1983; Teng, Cheon, & Grover, 1995; Wang & Strong, 1996)’’ (Dunk, 2000, p. 5). 4. ‘‘Although recognized as a major determinant of IS success, information quality received little attention as a construct in past empirical research on information systems, yand was often treated as a component of a general measure of user satisfaction’’ (Khalil & Elkordy, 2005). 5. ‘‘Complexity parallels perceived ease of use quite closely’’ (Davis, 1989). 6. And, as already remarked before, the relationship is clearly not the other way around, as early evidence from Granlund and Malmi (2002) suggests. Although ERPS are driving changes in organizational structures, such as the degree of centralization, these factors are not responsible for the degree of ERPS adoption. 7. Some companies recently merged from two or more organizations into one entity, others were listed both as a production facility and coordination centre. Each time, only one instance was retained for further analysis. 8. These were data-users as well as possible respondents. So the three groups, suggested by Dillman (2000) were all included in the pre-test. 9. In total nine companies had to be contacted through postal mail. 10. To avoid common method bias, the IT-manager in 10 (þ20%) of our respondent organizations was also contacted. Six responses were received and no significant differences were found. 11. One respondent did not reveal his level of education and was treated as a missing value in this test. 12. To test the Mann–Whitney U null-hypothesis, people who only answered after the reminder (4 weeks) were considered late respondents. Nevertheless, as the time frame of response ranged from 0 to 13 weeks, a Kruskal–Wallis test, based on the number of weeks after the initial mail, was also run to verify any differences between respondent groups. Again, no statistically significant differences at the 0.05 level were found. 13. To assess the relevant modules for each industry, the listing of Vluggen (2005) was used. As our research study included all modules, the relevance of those modules not included in the Vluggen study was determined by the authors in cooperation with an external ERPS specialist. 14. The Bartlett test of sphericity showed that non-zero correlations existed at the significant level of 0.000 for all the variables. The Kaizer–Meyer–Olkin (KMO) measure of sampling adequacy was met in all cases with an MS of W0.639 (Hair, Anderson, Tatham, & Black, 1998). 15. Unless mentioned otherwise, given our sample size a factor loading of .75 and preferably above was considered significant to be retained for further analysis (Hair et al., 1998).

Empirical Evidence on the (Perceived) Value of Information

171

16. Individual item reliability is considered adequate when an item has a factor loading that is greater than 0.707 on its respective construct (Carmines & Zeller, 1979). 17. The statistical problems created by singularity and multicollinearity occur at bivariate correlations of .90 and above (Tabachnick & Fidell, 2001).

ACKNOWLEDGEMENTS We gratefully acknowledge the helpful comments of Jan Bouwens (Tilburg University), Ann Vanstraelen (University of Antwerp), the cooperation of Marcel Weverbergh and assistant (University of Antwerp) for the online version of the survey, Wim Van Grembergen (University of Antwerp), Wynne W. Chin (University of Houston) for providing us with PLS-Graph and the respondents of our (pilot) survey for their time spent.

REFERENCES Abernethy, M. A., & Bouwens, J. (2005). Determinants of accounting innovation. Abacus, 41(3), 217–240. Abernethy, M. A., Chua, W. F., Luckett, P. F., & Selto, F. H. (1999). Research in managerial accounting: Learning from others’ experiences. Accounting and Finance, 39, 1–27. Abernethy, M. A., & Vagnoni, E. (2004). Power, organization design and managerial behavior. Accounting, Organizations and Society, 29(3–4), 207–225. Adams, D. A., Nelson, R. R., & Todd, P. A. (1992). Perceived usefulness, ease of use, and usage of information technology: A replication. MIS Quarterly (June), 227–247. Anthony, R. N., & Govindarajan, V. (2005). Management control systems (11th ed., 851pp.). McGraw Hill: Irwin, NY. Argyris, C. (1977). Organizational learning and management information systems. Accounting, Organizations and Society, 2(2), 113–123. Argyris, C., & Kaplan, R. S. (1994). Implementing new knowledge: The case of activity-based costing. Accounting Horizon, 8(3), 83–105. Bagranoff, N. A., & Brewer, P. C. (2003). PMB investments: An enterprise system implementation. Journal of Information Systems, 17(1), 85–106. Bailey, J. E., & Pearson, S. W. (1983). Development of a tool for measuring and analyzing computer user satisfaction. Management Science, 29(5), 530–545. Barclay, D., Higgins, C., & Thompson, R. (1995). The partial least squares approach to causal modeling, personal computer adoption and use as an illustration. Technology Studies, 2. Bingi, P., Sharma, M., & Godla, J. (1999). Critical issues affecting an ERP implementation. Information Systems Management, 16(3), 7–14. Bisbe, J., & Otley, D. (2004). The effects of interactive use of management control systems on product innovation. Accounting, Organizations and Society, 29, 709–737.

172

XAVIER GABRIE¨LS AND ANN JORISSEN

Bouwens, J., & Abernethy, M. A. (2000). The consequences of customization on management accounting system design. Accounting, Organizations and Society, 25, 221–241. Bradford, M., & Florin, J. (2003). Examining the role of innovation diffusion factors on the implementation success of enterprise resource planning systems. International Journal of Accounting Information Systems, 4, 205–225. Brownell, P. (1995). Research methods in management accounting. Melbourne: Coopers and Lybrand. Burns, J., & Scapens, R. W. (2000). Conceptualizing management accounting change: An institutional framework. Management Accounting Research, 11, 3–25. Caglio, A. (2003). Enterprise resource planning systems and accountants: Towards hybridization? European Accounting Review, 12(1), 123–153. Carmines, E. G., & Zeller, R. A. (1979). Reliability and validity assessment, sage university paper series on quantitative applications in the social sciences. Beverly Hills, CA: Sage. Chapman, C. (2005). Not because they are new: Developing the contribution of enterprise resource planning systems to management control research. Accounting, Organizations and Society, 30(7–8), 685–689. Chenhall, R. H., & Morris, D. (1986). The impact of structure, environment, and interdependence on the perceived usefulness of management accounting systems. The Accounting Review, LXI(1, January 1986), 16–35. Chin, W. (1998). Issues and opinion on structural equation modelling. MIS Quarterly, 22(1), 7–16. Clemons, E. K., Thatcher, M. E., & Row, M. C. (1995). Identifying sources of reengineering failures a study of the behavioral factors contributing to reengineering risks. Journal of Management Information Systems, 12(2), 9–36. Cooper, R., Kaplan, R. S., Maisel, L. S., Morrissey, E., & Oehm, R. M. (1992). Implementing activity-based cost management: Moving from analysis to action. Montvale, NJ: Institute of Management Accountants. Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16(3), 297–334. Davis, F. D. (1989). Perceived usefulness, perceived ease of use and user acceptance of information technology. MIS Quarterly (September), 319–340. Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of computer technology: A comparison of two theoretical models. Management Science, 35(8), 982–1003. Dechow, N., & Mouritsen, J. (2005). Enterprise resource planning systems, management control and the quest for integration. Accounting, Organizations and Society, 30(7/8), 691–733. DeLone, W. H., & McLean, E. R. (1992). Information systems success: The quest for the dependent variable. Information Systems Research, 3(1), 60–95. DeLone, W. H., & McLean, E. R. (2003). The DeLone and McLean model of information systems success: A ten-year update. Journal of Management Information Systems, 19(4), 9–30. Demski, J. S., & Feltham, G. A. (1976). Cost determination: A conceptual approach. Ames, IA: Iowa State University Press. Dillman, D. A. (2000). Mail and Internet surveys: The tailored design method (2nd ed). Wiley. Dunk, A. S. (2000). Quality of IS information, innovation budget pressure, and departemental performance. Working paper, March, University of Canberra, Canberra, Australia.

Empirical Evidence on the (Perceived) Value of Information

173

Ein-Dor, P., & Segev, E. (1978). Organizational context and the success of management information systems. Management Science, 24(10), 1064–1077. Fisher, J. (1998). Contingency theory, management control systems, and firm outcomes. Behavioral Research in Accounting, 10, 45–64. Fornell, C., & Larcker, D. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39–50. Foster, G., & Ward, T. J. (1994). Theory of perpetual management accounting innovation lag in hierarchical organizations. Accounting, Organizations and Society, 19(4/5), 401–411. Gabrie¨ls, X. (2006). The impact of enterprise resource planning systems (ERPS) on management accounting. Unpublished doctoral dissertation. Faculty of Applied Economics, University of Antwerp, Antwerp, Belgium, 155p. Galbraith, J. (1973). Designing complex organizations. Reading, MA: Addison-Wesley. Gelderman, M. (1998). The relation between user satisfaction, usage of information systems and performance. Information and Management, 34, 11–18. Gordon, L. A., & Narayanan, V. K. (1984). Management accounting systems, perceived environmental uncertainty and organization structure: An empirical investigation. Accounting, Organizations and Society, 9(1), 33–47. Gosselin, M. (1997). The effect of strategy and organizational structure on the adoption and implementation of activity-based costing. Accounting, Organizations and Society, 22(2), 105–122. Granlund, M. (2001). Towards explaining stability in and around management accounting systems. Management Accounting Research, 12, 141–166. Granlund, M., & Malmi, T. (2002). Moderate impact of ERPS on management accounting: A lag or permanent outcome? Management Accounting Research, 13, 299–321. Hair, J. F., Jr., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis (5th ed). Upper Saddle River, NJ: Prentice Hall. Hartwick, J., & Barki, H. (1994). Explaining the role of user participation in information system use. Management Science, 40(4), 440–465. Henri, J.-F. (2006). Organizational culture and performance measurement systems. Accounting, Organizations and Society, 31(1), 77–103. Huber, G. P. (1990). A theory of the effects of advanced information technologies on organizational design, intelligence and decision making. Academy of Management Review, 5(1), 47–71. Hyvo¨nen, T. (2003). Management accounting and information systems: ERP versus BoB. European Accounting Review, 12(1), 155–173. Igbaria, M., Zinatelli, N., Cragg, P., & Cavaye Angele, L. M. (1997). Personal computing acceptance factors in small firms: A structural equation model. MIS Quarterly, 21(3), 279–305. Indjejikian, R. J. (1999). Performance evaluation and compensation research: An agency perspective. Accounting Horizons, 13(2), 147–157. Ittner, C. D., & Larcker, D. F. (1998). Innovations in performance measurement: Trends and research implications. Journal of Management Accounting Research, 10, 205–238. Ives, B., Olson, M. H., & Baroudi, J. J. (1983). The measurement of user information satisfaction. Communications of the ACM, 26(10), 785–793. Jensen, M., & Meckling, W. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3, 305–360. .

174

XAVIER GABRIE¨LS AND ANN JORISSEN

Jermias, J. (2001). Cognitive dissonance and resistance to change: The influence of commitment confirmation and feedback on judgement usefulness of accounting systems. Accounting, Organizations and Society, 26, 141–160. Khalil, E. M., & Elkordy, M. M. (2005). EIS information: Use and quality determinants. Information Resources Management Journal, 18(2), 68–93. Krumwiede, K. R. (1998). The implementation stages of activity-based costing and the impact of contextual and organizational factors. Journal of Management Accounting Research, 10, 239–277. Lawrence, P. R. (1954). How to deal with resistance to change. Harvard Business Review, 32, 49–66. Libby, T., & Waterhouse, J. H. (1996). Predicting change in management accounting systems. Journal of Management Accounting Research (Fall), 137–150. Lodh, S. C., & Gaffikin, M. J. R. (2003). Implementation of an integrated accounting and cost management system using the SAP system: A field study. European Accounting Review, 12(1), 85–121. Low, G. S., & Mohr, J. (2001). Factors affecting the use of information in the evaluation of marketing communications productivity. Academy of Marketing Science Journal, 29(1), 70–88. Mas-Colell, A., Whinston, M. D., & Green, J. R. (1995). Micro-economic theory (998pp). New York: Oxford University Press. Mauldin, E. G., & Ruchala, L. V. (1999). Towards a meta-theory of accounting information systems. Accounting, Organizations and Society, 24, 317–331. McGowan, A. S., & Klammer, P. (1997). Satisfaction with activity-based cost management implementation. Journal of Management Accounting Research (9), 216–237. Merchant, K. A., & Manzoni, J. F. (1989). The achievability of budget targets in profit centers: A field study. The Accounting Review, 64, 539–558. Moers, F. (2001). Performance measures and performance targets in incentive systems: An empirical study of use and effects. Maastricht University. Moore, D. L., & Tarnai, J. (2002). Evaluating nonresponse error in mail surveys. In: R. M. Groves, D. A. Dillman, L. Eltinge & R. J. A. Little (Eds), Survey nonresponse (pp. 197–211). New York: Wiley. Newell, S., Huang, J., Galliers, R., & Pan, S. (2003). Implementing enterprise resource planning and knowledge systems in tandem: Fostering efficiency and innovation complementarity. Information and Organization, 13, 25–52. Nicolaou, A. I. (2004). Firm performance effectgs in relation to the implementation and use of enterprise resource planning systems. Journal of Information Systems, 18(2), 79–105. Nunnally, J. C. (1978). Psychometric theory (2nd ed). New York: McGraw-Hill. O’Reilly, C. A. (1982). Variations in decision makers’ use of information sources: The impact of quality and accessibility of information. Academy of Management Review, 24(4), 756–771. Pinsonneault, A., & Kraemer, K. L. (1993). Survey research methodology in management information systems: An assessment. Journal of Management Information Systems, 10(2), 75–85. Quattrone, P., & Hopper, T. (2005). A time-space odyssey: Management control systems in two multinationals. Accounting, Organizations and Society, 30(7/8), 735–764. Raghunathan, S. (1999). Impact of information quality and decision-maker quality on decision quality: A theoretical model and simulation analysis. Decision Support Systems, 26, 275–286.

Empirical Evidence on the (Perceived) Value of Information

175

Rogers, E. M. (1983). Diffusion of innovations (3rd ed.). New York: Free Press. Scapens, R. W. (1994). Never mind the gap: Towards an institutional perspective on management accounting practice. Management Accounting Research, 5, 301–321. Scapens, R. W., & Jazayeri, M. (2003). ERP systems and management accounting change: Opportunities or impacts? A research note, European Accounting Review, 12(1), 201–233. Seddon, P. (1997). A respecification and extension of the DeLone and McLean Model of IS success. Information Systems Research, 8(3), 240–253. Seddon, P., & Kiew, M. (1994). A partial test and development of the DeLone and McLean model of IS success. Proceedings of the international conference on information systems (ICIS 94), Vancouver, Canada (pp. 99–109). Shields, M. D. (1995). An empirical analysis of firms’ implementation experiences with activitybased costing. Journal of Management Accounting Research (Fall), 148–166. Simons, R. (1990). The role of management control systems in creating competitive advantage: New perspectives. Accounting, Organizations and Society, 15(1–2), 127–143. Sprinkle, G. B. (2003). Perspectives on experimental research in managerial accounting. Accounting, Organizations and Society, 28(2/3), 287–319. Sudman, S., & Blair, E. (1999). Sampling in the twenty-first century. Academy of Marketing Science, 27(2), 269–277. Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th ed.). Allyn & Bacon: Needham Heights, MA. Teng, J. T. C., Cheon, M. J., & Grover, V. (1995). Decisions to outsource information system functions: Testing a strategy-theoretic discrepancy model. Decision Sciences, 26(1), 75–103. Teo, T. S., & Wong, P. K. (1998). An empirical study of the performance impact of computerization in the retail industry. Omega (The International Journal of Management Science), 26(5), 611–621. Tiessen, P., & Waterhouse, J. H. (1983). Towards a descriptive theory of management accounting. Accounting, Organizations and Society, 8(4/5), 251–267. Vandenbosch, B. (1999). An empirical analysis of the association between the use of executive support systems and perceived organizational competitiveness. Accounting, Organizations and Society, 24, 77–92. Venkatraman, N., & Ramanujan, V. (1987). Measurement of business economic performance: An examination of method convergence. Journal of Management (Spring), 109–122. Vluggen, M. (2005). The determinants of enterprise resource planning (ERP) systems usage levels: An empirical study. Paper presented at the EAA Conference, Go¨teborg, Sweden. Wang, R. Y., & Strong, D. M. (1996). Beyond accuracy: What data quality means to data consumers. Journal of Management Information Systems, 12(4), 5–34. Watson, H. J., Rainer, K., & Houdeshel, G. (1997). Building executive information systems and other decision support applications. New York: Wiley. Wixom, B. H., & Watson, H. J. (2001). An empirical investigation of the factors affecting data warehousing success. MIS Quarterly, 25(1), 17–41. Zimmerman, J. L. (2000). Accounting for decision making and control (3rd ed). New York: R.D. Irwin.

This page intentionally left blank

THE MISSING LINK BETWEEN INFORMATION AND ACTION: HASTENINGS AND DELAYS AS UNIVERSAL REACTIONS TO PERFORMANCE FEEDBACK Tom De Schryver, Rob Eisinga, Christine Teelken and Erik Poutsma ABSTRACT This chapter focuses on what the key decision makers in organizations decide after having received information on the current state of the organizational performance. Because of strong attributions to success and failure, it is impossible to predict in advance which concrete actions will occur. We can however find out what kinds of actions are decided upon by means of an organizational learning model that focuses on the hastenings and delays after performance feedback. As an illustration, the responses to performance signals by trainers and club owners in Dutch soccer clubs are analyzed.

Performance Measurement and Management Control: Measuring and Rewarding Performance Studies in Managerial and Financial Accounting, Volume 18, 177–192 Copyright r 2008 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-3512/doi:10.1016/S1479-3512(08)18008-X

177

178

TOM DE SCHRYVER ET AL.

Investing in information systems makes sense if it offers interesting pieces of information that reduces the environmental complexity into manageable pieces so that the decision situation becomes clear (Lant & Montgomery, 1987). Yet, decision making can remain challenging when the information technology to produce good information is present (Lampel & Shapira, 2001). These problems happen especially with complex information systems, reporting on multiple overlapping indicators like scorecards with leading, lagging, financial and non-financial indicators because they create inconsistent feedback; some information suggests that the organization is doing well, while other information indicates – at the same time – that the organization is not doing so well. As a result, detailed charts of the organizational reality do not necessarily reduce the environmental complexity. To handle this problem, decision makers need to filter information themselves. Filtering of information by humans is natural as decision makers can only scrutinize limited pieces of information at a time. Yet, it remains important to assess which information is picked up for decision making without having to assume that they will focus on problems first. It is equally possible that decision makers will chose information they already know, or information that best serves their goals. This chapter explores these different possibilities in a systematic way. The first section explains why there is variation in individual responses to organizational performance feedback. The second section proposes a framework based on piecewise linear models. The third section relates the framework to the reactions that can be expected from various organizational learning theories. The fourth section applies the model to decision making after performance feedback in Dutch soccer clubs.

DECISION MAKERS VERSUS STAKEHOLDERS Even when there is a fixed organizational goal and when there are clear decision allocation rights, organizational decision makers disagree upon the actions that need to be taken after performance feedback. It is known that performance appraisal can lead to actions to improve the implementation of strategies or to rethink the strategic orientation (Simons, 1991). Moreover, actions to implement strategies can mean a lot of different things. Actions to implement strategies happen because it is believed that strategies are poorly understood, because performance feedback highlights principal-agent problems, or because of single-loop learning. Similarly, when new performance information is used to discuss future strategic

The Missing Link between Information and Action

179

orientations, disagreement may arise. Some decision makers may champion alternatives and make actions that are deliberately different from the common organizational orientation. Since clear signs can appeal to different decision makers in the organization, it becomes difficult to discern what the response will be of organizations after performance feedback. Put differently, decision makers cannot be reduced to a simple means to an end. When decision makers are seen as stakeholders with different but overlapping goals, with different organizational roles and capabilities, then it becomes important to find out who will respond to what kind of information for what goal. Goal conflict allows us to contrast self-serving behavior in organizations with self serving by organizations (Johns, 1999). On the one hand, goal conflict can lead to more information processing in an attempt to reconcile different points of views. The information then serves the organization. Alternatively, information processing can be used to serve the individuals because inconsistent feedback increases the possibilities for framing and filtering information. As such, inconsistent feedback is convenient to explain away a particular action. Other decision makers may react themselves with increasing their self-serving behavior because they feel that other problems or opportunities are left unattended. A framework and organizational learning theory is needed to test the individual reactions of different stakeholders by means of a model that relates information to action.

HASTENINGS AND DELAYS IN PIECEWISE LINEAR MODELS Since we want to find out how individual stakeholders respond to performance feedback, it is not really an option to rely on very broad measures of action that lump together all different responses of all different decision makers. There however exists a promising methodological approach that can explain ex post what has happened after performance signals. The methodology is based on piecewise linear models. It has been used first by Greve (2003) in organizational decision sciences and has since then been adopted by many scholars (e.g. Audia & Greve, 2006; Baum, Rowley, Shipilov, & Chuang, 2005; Park, 2007). The reason why we make use of it, is that the piecewise linear model can be applied to any commitment to action, as long as it is observable

180

TOM DE SCHRYVER ET AL.

shortly after performance appraisal. There does not have to be a prior indication that the commitment to action actually has occurred, nor that performance information is effectively scrutinized. The minimum requirement is that actions can be correlated with performance signals in timeframes nearby. Piecewise linear models then split performance information into one or more zones of success and failure by multiplying the performance signals with indicator variables. It results in Eq. (1).

hðmÞ ¼

X

b1ij ðSijk DðPijkLijkÞ Þt þ

X

b2ij ðS ijk DðPijk4LijkÞ Þt

(1)

Legend i ¼ indicator for the performance indicator j ¼ indicator for the aspiration level or benchmark t ¼ repeated measures indicator k ¼ indicator for the decision maker S ¼ performance signal ¼ P–L P ¼ observed value of the performance indicator L ¼ aspired value of the performance indicator mij ¼ linear predictor h ¼ link function DPipLij ¼ indicator variable for failure: 0 when PiWLij, 1 when PipLij DPiWLij ¼ indicator variable for success: 0 when PipLij, 1 when PiWLij

Key for piecewise models is that the response after success and failure can differ due the separate effect of low performance (b1) and high performance (b2). Feelings of success typically occur when actual performance (Pi) outnumbers an aspiration level (Lij) and feelings of failure occur when there is negative performance feedback. Success can further lead to complacence and delays in some occasions and lead to confidence and hastenings in other occasions. Similarly, signs of failure can trigger hastenings because solutions to the problems need to be found and they can lead to delays because there is some threat rigidity. The strong attributions will be translated into shifts in concrete commitments to action. Since these commitments are observable, they can be analyzed by Eq. (1).

The Missing Link between Information and Action

181

HASTENINGS AND DELAYS IN ORGANIZATIONAL LEARNING THEORY Piecewise linear models are especially powerful when they are used in combination with organizational learning theories because they give detailed explanations as to why different decision makers would reveal hastenings and delays after success or failure. We will show how Threat Rigidity Theory (TRT) introduced by Staw, Sandelands, and Dutton (1981), A Behavioral Theory of the Firm introduced by Cyert and March (1992) and the Variable Risk Preference Model (VRPM) introduced by March and Shapira (1992) can be combined with piecewise linear models and can result in propositions on hastenings by different decision makers in the organization. Hastenings and Delays in Threat Rigidity Theory (TRT) According to TRT, threats will increasingly lead to a restriction in information processing, a constriction in control and a focus toward efficiency (Staw et al., 1981). Therefore the organization becomes rigid in its response. The restriction in information processing implies that initial signs of negative performance signals spark decision making, but that later signals will be left unattended. Therefore, a restriction in information processing corresponds with b1ijkW0. This is the first manifestation of hastenings starting in the top of the organization. It also leads to a constriction of control, which implies an increased centralization, a strengthening of tightly coupled links and dissolution of weak links. It means that top managers will restrict the managerial discretion of their middle managers and even will replace middle managers. Moreover, TRT assumes that the best solution to put a halt to the threat is to better implement the strategy and to focus on efficiency. For example, various sorts of cost cutting can be a remedy to level off repeated signs of negative performance feedback. When there is no longer a threat and performance restores to aspiration levels, TRT assumes that the top relaxes the information restrictions, loosens control mechanisms, and no longer single out efficiency as main strategic orientation. TRT further assumes that successful organizations will tolerate experimentation more and allow strategy changes. Commitments to efficiency, constriction in control, and centralization of information can lower. The extent to which strongly positive performance feedback extends these behavioral responses is not explored in TRT. Therefore, we argue that

182

TOM DE SCHRYVER ET AL.

(1) b1ijk has to be positive and (2) b2ijk has to be less positive than b1ijk in TRT for strategy changes and that (1) b1ijk has to be negative and (2) b2ijk has to be less negative than b1ijk in TRT for strategy-implementation issues. Hastenings and Delays in A Behavioral Theory of the Firms (ABTF) In ABTF, quasi-resolution of conflict, uncertainty avoidance, problemistic search and organizational learning determine decision making (Cyert & March, 1992). A lack of confidence in the current strategy makes strategy changes much more likely after negative performance feedback. Contrary to TRT, the top of the organization is unwilling or unable to impose its own solution. As a result it is possible that all decision makers take action. Negative performance feedback creates dissatisfaction, resulting in increased odds of conflict between stakeholders and experimentation to form new coalitions or to find new solutions. Different solutions by different stakeholders will be put on the agenda. Only some of the solutions will be implemented. Yet, the search for solutions does not need to end. Instead it will stop once performance restores to levels that are acceptable for all remaining stakeholders. Contrary to TRT, ABTF gives clear indications when to expect hastenings or delays when performance is much better than aspired. Positive performance feedback gives a feeling of satisfaction to all stakeholders. Consequently, the stakeholders will be willing to stay in the organizational coalition and are unlikely to suggest decisions that might endanger the coalition. Therefore, the more performance feedback is positive, the more risky decisions are delayed. When performance is beyond aspirations, organizations are complacent and ignore signals from performance feedback. Because problemistic search can be fruitless, the effects of positive performance feedback on actions are assumed to be stronger than the effects of negative performance feedback on actions in ABTF. Therefore, we argue that (1) b2ijk has to be negative and (2) b1ijk has to be less negative than b2ijk in ABTF for all actions that put pressure on the dominant coalition.

Hastenings and Delays in the Variable Risk Preference Model (VRPM) The VRPM has shed interesting new lights on the theory of decision making. Counter to ABTF, the VRPM assumes moderate risk taking as

The Missing Link between Information and Action

183

positive performance feedback gets stronger. The VRPM argues that managers in highly successful organizations resort to risk taking because they feel that they can deliver a positive return on organizational resources. Some successful managers are thus assumed to be confident and to take risk or to institute changes that they feel they can implement (Shapira, 1995). There is some empirical evidence for this line of thought. With a broad measure of risk taking, Miller and Chen (2004) found no support that high success leads to additional risk taking. Their result is in support of ABTF: satisfaction gets stronger as performance gets better. Yet, it does not rule out that highly successful firms undertake specific opportunistic actions. For example, Baum et al. (2005) have showed that extremely successful banks are more willing to underwrite risky strategic alliances because the success buffers the downside risk of failure. VRPM can also contribute to modeling information ambiguity. In TRT and ABTF stakeholders may hold different views on the way to turnaround low performance, but everybody agrees upon the state of the performance. This can happen when there is only one signal or when all signals of a complex performance system generate the same information. If all elements of a complex performance system always align then, on grounds of efficiency, there is no need to keep complex information systems operational. Rather, complex information systems only make sense when they provide inconsistent feedback. However, inconsistent feedback makes the relationship between information and actions even more elusive because it gives the decision maker the opportunity to consider organizational performance a success or a failure. The organizational learning theories have primarily focused on the effects of negative performance feedback. Yet, self-serving theory (e.g. Johns, 1999) has focused on the extreme bias to focus on favorable evaluations even when there are clear indications of failure. It for instance leads to impression management of financial statements (Lewellen, Park, & Ro, 1996). More extreme is Audia and Brion (2007) who found that decision makers rather appeal to less important performance indicators when they generate feelings of success to counter feelings of failure from primary performance indicators to stall new product launches. Inconsistent feedback thus may lead to satisfaction when performance appraisal does not uniformly indicate toward success. As such, complex performance systems may well overshoot their target. While they aim at reducing the environmental complexity by increasing the number of performance indicators, they may support opportunistic or failure-induced decisions without knowing why.

184

TOM DE SCHRYVER ET AL.

In this line of thought, we place the VRPM that suggests that the overall organizational propensity toward risk taking will be low when there is information ambiguity. The aggregated organizational response will take into account that some decision makers advocate no action because they are satisfied, while other decision makers would rather like to repair the gap with aspiration and institute changes. As a result, a move toward consensus makes the organizational response lie somewhere in between two extremes of risk taking and inaction. The question then becomes who in the organization will focus on signs of failure and who will focus on success. In that respect, there may be important differences between top and middle managers. Because top managers are representatives of the chosen strategic orientation, top managers will reveal stronger commitments to the past than middle managers. Consequently, they will use positive performance indicators as long as possible and ignore negative information as long as possible. It could be that top managers postpone strategy changes as long as possible, salve in the case of extreme success or failure. On the contrary, it may very well be so that middle managers will institute plenty of changes even if there is partial information that they are successful, because actions show their mastery in many domains and actions are taken to ensure that their targets are being met. When success is overwhelming, middle managers have no time to challenge the existing way of doing. Therefore, we argue that (1) b1ijk has to be positive for middle managers and (2) b2ijk has to be negative. Similarly, (1) b1ijk has to be negative for top managers and (2) b2ijk has to be positive. This way, the strength of the VRPM is optimally used: tendencies to institute strategy changes need to be balanced all the time against tendencies to remain the status quo.

HASTENINGS AND DELAYS IN THE DUTCH SOCCER LEAGUE The Dutch soccer premier league is an interesting setting because of at least two reasons. First, there are multiple performance indicators. It also seems that short-term performance matters more than long-term performance signals for club owners to change trainers within the season (e.g. Audas, Dobson, & Goddard, 1999; Dobson & Goddard, 2001). Second, the response of the trainers to performance appraisal is underexposed in academic literature.

The Missing Link between Information and Action

185

We will thus take into account two relevant decision makers in soccer clubs: club owners (which correspond to the top managers) and trainers (which correspond to the middle management). Together they have four strategic choices: club owners replacing trainers, trainers shifting to more offensive strategies, to more defensive or to more efficient strategies. These are the only decisions that can be taken during the soccer season. Club owners can only replace trainers without legal restrictions during the season. Trainers have more options for action during the season. They can motivate, train, select players from the talent pool differently. This can be achieved in numerous ways. ‘Catenaccio’ is an ideal type of defensive strategy requiring a highly organized and effective backline defense, which aims to prevent goals. ‘Brazilian Samba’ soccer is the ideal type of a prospector strategy where the team tries to provoke confusion in their adversaries, fascination in the public and goals for the team. Finally, there is the ‘German’ soccer style that tries to attain a clinical efficiency, which corresponds to an analyzers strategy. In this chapter, we will assume that clubs are able to combine the strategies into hybrid forms. A highly successful hybrid strategy is ‘Total Soccer’, first adopted in the Netherlands, which combines elements of German efficiency with the Latin elements of surprise (ExpertFootball.com, 2004). Hypotheses In this section we tailor the propositions of the different organizational learning theories into concrete hypotheses for testing. TRT expects first middle managers and then top to institute hastenings after first signs of negative performance feedback. Initiatives of middle managers that are not related toward efficiency will be curtailed by the top as negative performance feedback gets worse. The organization starts restricting the processing of information. Only initiatives aimed at improving efficiency may last in this period. Moreover, because of the constriction in control, club owners will hasten dissolving weak links. We therefore expect higher odds that trainers will be replaced after first signs of negative performance feedback. Finally, if trainers did not already focus on efficiency, club owners will increasingly enforce them to do so. Consequently, the TRT hypotheses for soccer clubs are Hypothesis 1a. Club owners hasten to replace trainers just after performance has dropped below aspiration (i.e. b1ijkW0 and b1ijkWb2ijk). Hypothesis 1b. Trainers hasten the adoption of defensive or offensive strategies just after performance has dropped below aspiration (i.e. b1ijkW0 and b1ijkWb2ijk).

186

TOM DE SCHRYVER ET AL.

Hypothesis 1c. Trainers hasten the adoption of efficiency strategies as long as there is negative performance feedback (i.e. b1ijkob2ijko0). In ABTF, dissatisfaction prevails over confidence in the current strategy. Negative performance feedback sparks latent conflicts and a pessimistic outlook. Consequently, both trainers and club owners independently look for responses to problems. Thus, the odds increase that trainers will try out other strategies and that club owners will try out other trainers. Consequently, the ABTF-hypotheses for soccer clubs are Hypothesis 2a. Club owners hasten to replace trainers as long as there is negative performance feedback (i.e. b2ijkob1ijko0). Hypothesis 2b. Trainers hasten to adopt other strategies as long as there is negative performance feedback (i.e. b2ijkob1ijko0). According to VRPM hastenings will not only occur when performance feedback is negative but also when it is exceptionally positive. In the soccer setting, the only plausible opportunistic action is that trainers shift strategy because they think that the existing talent pool of players can deliver even better results. Also according to VRPM inconsistent performance feedback may lead to diversity of actions. Hypothesis 3a. Club owners hasten to replace trainers as long as there is negative performance feedback (i.e. b2ijkob1ijko0). Hypothesis 3b. Trainers increasingly adopt new strategies when signals of performance feedback get stronger (i.e. b1ijko0 and b2ijk W0). Data, Measures and Models Various websites were consulted to collect the data. The information consulted at http://home.wanadoo.nl/ronald.zwiers/ was found to be valid. The final sample consists of 16 leagues between 1990/1991 and 2005/2006. Since a league lets 18 clubs play twice against each other, there are 34 repeated measures for 288 ( ¼ 18  16) club years, giving 9792 observations in total. We rely on Generalized Estimation Equations to replicate the data in the sample. These statistical models estimate the effects of multiple regressors on discrete and interval responses in a repeated measurement setting (Diggle, Liang, & Zeger, 2002). The models that we fitted contain some

The Missing Link between Information and Action

187

confounding variables. The first set came from visual inspection of the data. It appeared that club owners stall trainer turnover decisions either at the start or at the end of the season. Therefore, we include starting up and cooling down effects. We also added a linear effect as we observed that the number of trainer turnover differs from year to year. The second set of baseline variables is the result of soccer-specific knowledge. We took into account that there was a shift in the incentive structure, where wins were rewarded with 3 points against 2 points from 1994/1995 onwards. We also took into account that a large number of consecutive losses must lead to trainer turnover (Colin & Muller, 1996, 2002). Finally, reputation effects and effects of current strategic orientation may influence the decision making in clubs. In particular for actions by the club owner, we add experience in the premier league, the date of incorporation, the tenure of the trainer in the club, and the turnover experience of the club. For actions undertaken by the trainer, the experience in the premier league, the date of incorporation and strategic orientation variables are included. The measures for the action variables are Trainer turnover is a dichotomous variable measure, with value 1 if the club owner replaces a trainer after a match, 0 if the club owner decides to keep the trainer. Offensive shift. Trainers can shift their selection, training and motivation to move toward a prospector strategy. An offensive shift toward a prospector strategy is measured by the realized increase after four weeks in the number of goals scored per match. Defensive shift. Trainers can shift their selection, training and motivation to move toward a defender strategy. A successful shift toward a defensive strategy is measured by the realized decrease after four weeks in the number of goals taken per match. Efficiency shift. Trainers can shift their selection, training and motivation to move toward an efficiency strategy. A successful shift toward efficiency is measured by the increase of goal difference after four weeks. The measures for the performance variables are: Short-term performance met monitors the performance of a single match. The match result is compared with the score that could be expected from the current ranking. In particular, we expect clubs to win games from opponents having currently less than 4 points in the ranking. Otherwise, clubs should expect to obtain a draw. This can lead to positive performance

188

TOM DE SCHRYVER ET AL.

signals – ‘short-term performance met’ – when clubs beat higher-ranked teams and negative performance signals – ‘short-term performance not met’ – when clubs loose or when clubs obtain a draw against lower–ranked teams. Long-term performance met monitors the ranking. It signals whether the club is going to attain a ranking similar as the average ranking of the last three years before. This performance signal was constructed by subtracting the current ranking by a weighted moving average of the ranking of the last three years. This signal is substantially different from the short-term performance because even when teams have lost a game, it is possible to get positive performance feedback if the match result does not affect the ranking. Minimum performance met monitors if the club will not be able to play next year in the premier league. The signals therefore compare the current ranking of the club with the minimum ranking that is necessary to attain next year’s premier league.

Results Before discussing the main results, we skim over the baseline variables in Table 1. The first column of Table 1 reports strong cooling down and starting up effects, indicating that in the beginning and at the end of the season, club owners do not decide to change trainers. We also find that the number of consecutive losses is an important predictor for trainer turnover. Replacing trainers within the season seems to have changed since the incentive system rewards more offensive strategies. The changed attitude toward trainer turnover has however not been gradual or linear over the years. Finally, the tenure of the current trainer and recent experience with trainer turnover have small but significant effects on the decision of club owners to replace trainers. The second, third and fourth columns describe the decision making context for trainers. The results clearly show that shifts toward more offensive, defensive or efficient strategies are largely determined by the current strategic orientation. Table 1 indicates that trainers cannot make a more offensive shift when they already have followed a prospector strategy. Nor are defenders likely to make offensive shift. Offensive shifts are to be considered by analyzers, although the effect is small. It seems that a defensive shift will be increasingly considered by defenders, but it may also be considered by prospectors. Efficiency shifts are unlikely to be realized by trainers following a defender strategy, but might be considered by a trainer

189

The Missing Link between Information and Action

Table 1. Effect of Performance Feedback on Trainer and Club Owner.

Cooling down Starting up Season (linear) Before 1994/1995 Number of consecutive losses Experience in premier league Date of incorporation Turnover experience Tenure of trainer in club (weeks) Analyzer strategy Defender strategy Prospector strategy Trainer turnover Contrast short-term aspiration Short-term aspiration not met Short-term aspiration met Contrast long-term aspiration Long-term aspiration not met Long-term aspiration met Contrast minimum aspiration Minimum aspiration not met Minimum aspiration met Quasi-likelihood under the Independence model Criterion (QIC) (Pan, 2001, p. 122)

Trainer Turnover

Offensive Shift

Defensive Shift

Efficiency Shift

1.37 (.00) 1.65 (.00) 0.04 (.35) 0.84 (.01) 0.22 (.00) 0.02 (.08) 0.00 (.35) 0.16 (.00) 0.01 (.00) – – – – – – – – 0.01 (.98) 0.51 (.04) 0.52 (.08) 0.11 (.39) 0.21 (.06) 0.10 (.01) 0.01 (.96) 0.02 (.86) 0.01 (.80) 1006

0.39 (.00) 0.14 (.35) 0.00 (.00) 0.13 (.45) 0.07 (.20) 0.01 (.02) 0.00 (.08) 0.01 (.67) 0.00 (.98) 0.02 (.02) 0.56 (.00) 2.37 (.00) 0.77 (.01) 0.39 (.01) 0.22 (.02) 0.17 (.01) 0.03 (.49) 0.16 (.00) 0.14 (.00) 0.16 (.01) 0.10 (.09) 0.06 (.02) 8407

0.37 (.01) 0.22 (.12) 0.00 (.00) 0.12 (.47) 0.05 (.33) 0.01 (.16) 0.00 (.07) 0.00 (.93) 0.00 (.42) 0.01 (.07) 2.57 (.00) 0.44 (.00) 0.28 (.45) 0.05 (.74) 0.00 (.96) 0.05 (.53) 0.01 (.85) 0.12 (.00) 0.13 (.00) 0.01 (.94) 0.12 (.12) 0.12 (.00) 8412

0.00 (.98) 0.03 (.87) 0.00 (.00) 0.01 (.96) 0.10 (.17) 0.03 (.00) 0.00 (.95) 0.01 (.86) 0.00 (.72) 0.02 (.04) 0.63 (.00) 0.80 (.00) 0.46 (.27) 0.56 (.00) 0.68 (.00) 0.11 (.25) 0.05 (.40) 0.33 (.00) 0.28 (.00) 0.09 (.43) 0.25 (.03) 0.34 (.00) 8439

with a prospector strategy. Overall, defenders seem to be increasingly good at defending but loose out their attacking abilities. And prospectors, on average, seem to be unable to improve their offensive skills. The baseline variables suggest that the models describe a plausible context in which the results of performance feedback can be embedded. Before focusing on the analysis of the individual slopes for the performance effects, it is important to inspect whether the slopes for good and bad performance are statistically different from each other. Three contrasts indicate that the actions of the trainers differ after signs of success versus failure. It is apparent from Table 1 that only trainers tend to shift strategies after performance feedback. While the contrasts rule out that club owners hasten to change trainers after information, they do not rule out that club owners are sensitive to performance information. The first column

190

TOM DE SCHRYVER ET AL.

reveals that club owners tend to stall trainer turnover the more the short-term performance is met. The same model also suggests that on average the higher the ranking of a club, the more common it becomes to consider trainer turnover. Still the reaction of the club owner is seasoned because none of the contrasts is significant. In terms of TRT, there is thus no support that club owners shift to TRT responses. Not only the insignificant contrasts, but also the fact that new trainers are unable to impose an efficiency strategy in the four weeks after their arrival is an indication that the club owners are not imposing responses to the trainer. Thus there is no support for Hypotheses 1a–1c, ruling out that the organization as a whole responds in a TRT way. Still trainers show signs of threat rigidity in some occasions. The significant contrasts for short-term performance indicate that the trainer shifts strategies after a disappointing game. In particular, trainers respond to losses against lower-ranked teams, by increasing the efficiency of the team (b1ij ¼ 0.68) or by increasing the offensive power of the team (b1ij ¼ 0.22). The contrast also indicates satisfaction when clubs just have beaten a stronger opponent. The results suggest that success may backfire as clubs lose offensive power as indicated by a b2ij of 0.17. It seems that trainers are then unable to impose a feeling of urgency, which makes it hard for clubs to keep the strong offensive results in the next four matches. One contrast is however not enough for full support of ABTF. While good short-term performance somehow lowers the odds of club owners to shift to trainer turnover, there is no guarantee that club owners will not replace trainers then. There is thus only partial support for Hypotheses 2, notably for Hypothesis 2b. Another indication against ABTF is that trainers show little satisfaction when the historical performance feedback is good. Rather, they increasingly keep on experimenting different approaches as their ranking is better than expected. Yet these experiments do not reflect truly opportunistic behavior after performance appraisal, as the contrasts indicate that shifts in strategy happen gradually. Finally, there is partial support for the VRPM. The third significant contrast is for offensive shifts after signs that the minimum performance is not met. The contrast reveals a pattern that is in support of Hypothesis 3b. A negative sign for b1ij of 0.1 indicates that trainers look for quick wins when they are close to relegation. The positive sign for b2ij further indicates that trainers face offensive problems when the minimum ranking is just met. They are only able to institute offensive changes the further they are away from the relegation zone. Again there is no support for Hypothesis 3a that club owners shift attention.

The Missing Link between Information and Action

191

Overall, the results of this study correspond with conclusions from earlier empirical studies on performance appraisal (e.g. Ketchen & Palmer, 1999; Lant & Hurley, 1999) that organizational behavior after performance feedback reveals streaks of TRT, ABTF and VRPM, instead of one uniform pattern. We add that the different streaks might be divided among groups and individual members. Our study indicates that during soccer leagues, club owners use performance information in a seasoned way. On the contrary, trainers show signs of TRT, ABTF and VRPM. Overall, they are much more responsive to performance feedback than club owners.

CONCLUSION We have put the decision maker to the fore and the information systems to the back because information only becomes valuable when it is used by human beings for decision making. Information does not always lead to action, because of processes of selection and choice. This chapter investigated selection processes that are fueled by feelings of success and failure. They influence confidence, satisfaction or both. This makes it too challenging to predict where these feelings of success or failure will lead to. Yet it is realistic to assume that success or failure affect the decision speed of actions. Since decision speed is observable for most commitments to action, it puts research on the elusive link between information and action a step further. Rich links with many different organizational learning theories can be made. An illustration in this chapter with soccer clubs has shown how a focus on decision speed explains different reactions to performance feedback by trainer and club owner.

ACKNOWLEDGMENTS We would like to thank the EIASM institute, Professor Marc J. Epstein and Professor Jean-Franc- ois Manzoni for being able to present our chapter at the 4th EIASM Conference on Performance Measurement and Management control in Nice. The chapter has benefited from the remarks at the conference, notably from Professor Davila. We would also like to thank the FWO-Vlaanderen for the travel grant and the colleagues at the Department of Economics, University Antwerp for useful comments during the lunch seminar.

192

TOM DE SCHRYVER ET AL.

REFERENCES Audas, R., Dobson, S., & Goddard, J. (1999). Organizational performance and managerial turnover. Managerial and Decision Economics, 20(6), 305–318. Audia, P. G., & Brion, S. (2007). Reluctant to change: Self-enhancing responses to diverging performance measures. Organizational Behavior and Human Decision Processes, 102(2), 255–269. Audia, P. G., & Greve, H. R. (2006). Less likely to fail: Low performance, firm size, and factory expansion in the shipbuilding industry. Management Science, 52(1), 83–94. Baum, J. A. C., Rowley, T. J., Shipilov, A. V., & Chuang, Y. T. (2005). Dancing with strangers: Aspiration performance and the search for underwriting syndicate partners. Administrative Science Quarterly, 50(4), 536–575. Colin, F., & Muller, L. (1996). De Gouden Voetbalgids. Antwerpen: Standaard. Colin, F., & Muller, L. (2002). De Grote Voetbalencyclopedie. Antwerpen: Houtekiet. Cyert, R. M., & March, J. G. (1992). A behavioral theory of the firm (2nd ed.). Cambridge, MA: Blackwell Business. Diggle, P. J., Liang, K.-Y., & Zeger, S. L. (2002). Analysis of longitudinal data (2nd ed.). Oxford: Oxford University Press. Dobson, S., & Goddard, J. A. (2001). The economics of football. New York: Cambridge University Press. ExpertFootball.com. (2004). Soccer Styles of Play. Retrieved on March 14, 2007, from http:// expertfootball.com/coaching/styles.php Greve, H. R. (2003). Organizational learning from performance feedback: A behavioral perspective on innovation and change. Cambridge: Cambridge University Press. Johns, G. (1999). A multi-level theory of self-serving behavior in and by organizations. Research in Organizational Behavior, 21, 1–38. Ketchen, D. J., & Palmer, T. B. (1999). Strategic responses to poor organizational performance: A test of competing perspectives. Journal of Management, 25(5), 683–706. Lampel, J., & Shapira, Z. (2001). Judgmental errors, interactive norms, and the difficulty of detecting strategic surprises. Organization Science, 12(5), 599–611. Lant, T. K., & Hurley, A. E. (1999). A contingency model of response to performance feedbackescalation of commitment and incremental adaptation in resource investment decisions. Group & Organization Management, 24(4), 421–437. Lant, T. K., & Montgomery, D. B. (1987). Learning from strategic success and failure. Journal of Business Research, 15(6), 503–517. Lewellen, W. G., Park, T., & Ro, B. T. (1996). Self-serving behavior in managers’ discretionary information disclosure decisions. Journal of Accounting & Economics, 21(2), 227–251. March, J. G., & Shapira, Z. (1992). Variable risk preferences and the focus of attention. Psychological Review, 99(1), 172–183. Miller, K. D., & Chen, W. R. (2004). Variable organizational risk preferences: Tests of the march-shapira model. Academy of Management Journal, 47(1), 105–115. Park, K. M. (2007). Antecedents of convergence and divergence in strategic positioning: The effects of performance and aspiration on the direction of strategic change. Organization Science, 18(3), 386–402. Shapira, Z. (1995). Risk taking: A managerial perspective. New York: Russell Sage Foundation. Simons, R. (1991). Strategic orientation and top management attention to control-systems. Strategic Management Journal, 12(1), 49–62. Staw, B. M., Sandelands, L. E., & Dutton, J. E. (1981). Threat-rigidity effects in organizational behavior – A multilevel analysis. Administrative Science Quarterly, 26(4), 501–524.

THE LIFETIME VALUE SCORECARD: FROM E-METRICS TO INTERNET CUSTOMER VALUE M. Bonacchi, M. Ferrari and M. Pellegrini ABSTRACT The aim of this chapter is to develop a performance measurement framework for understanding the relationships among drivers of customer profitability in internet companies. We recognize an opportunity to improve management control systems for internet companies, where performance measurement systems currently focus on measuring web data, such as number of customers, cost of service, cost of acquisition (CoA), and churn rate. However these indicators, taken separately, do not provide useful information to make decisions. To fill this gap we developed a framework, which we designate as the Lifetime Value Scorecard, to investigate the relationships between customer data and financial data, providing an early indication as to whether or not the marketing strategies being implemented are successful. We then offer an application of the Lifetime Value Scorecard to the mobile value-added services industry, where content and services are provided to consumer cell phones, mainly using wireless networks.

Performance Measurement and Management Control: Measuring and Rewarding Performance Studies in Managerial and Financial Accounting, Volume 18, 193–228 Copyright r 2008 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-3512/doi:10.1016/S1479-3512(08)18009-1

193

194

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

1. INTRODUCTION Management accounting in internet companies is moving from e-metrics to drivers of profitability. Early evaluations of internet companies were based mainly on observing the numbers of visitors to the site. The assumption was that the number of visitors translated directly into revenues, so that the higher the number of visitors, the greater the revenues would be. E-metrics related to the customer base were used to justify the price of internet pure player stocks until the first half of 2000. But since the internet shakeout customer base has no longer been considered an asset, unless it is profitable (Lev, 2001). Internet pure players, more so than other type of company, realized at that time the value of fundamental economics. To plan any kind of investment or marketing campaign, internet companies need to verify both their attraction, conversion and retention capacities, and consequently the profitability of their customer portfolio as well as the cost of customer acquisition (Agrawal, Arjona, & Lemmens, 2001). In other words, companies can only afford to acquire new customers for amounts less than the customer lifetime value (CLV) (Berger & Nasr, 1998). In fact, maximizing the customer base does not mean increasing the value of the company, as there is no relationship between customer acquisition and financial performance, unless it is possible to demonstrate that the stream of profit realized from new customers (CLV) will be higher than the cost of their acquisition (CoA). In the last couple of years, advertising models have evolved as well, from a focus on ‘‘brand awareness’’ to a focus on ‘‘direct and measurable’’ customer acquisitions (Economist, 2006a, 2006b, 2007b; Epstein, 2007; Epstein & Yuthas, 2007; French, 2007). Unlike television advertising, internet advertisers pay only when a user clicks through to their website, gaining a reliable measurement of customer acquisition costs (Court, 2005; Laffey, 2007). In addition, a large amount of computer readable data on marketing performance is now becoming available via search engines and other information technology tools (Economist, 2007a; Epstein, 2007; Varian, 2006). This allows firms to use data on revealed preferences rather than intentions, and sampling is no longer necessary when you have the whole population available (Gupta, Hanssens, Hardie, & Kahn, 2006). Given the above, we recognize an opportunity to improve management control systems for internet companies. The main issues to deal with are: the determination of the expected value of customers, and the CoA. These concepts are not new (Gupta et al., 2006; Gupta, Lehmann, & Stuart, 2004; Gupta & Zeithaml, 2006; Neslin, Gupta,

The Lifetime Value Scorecard

195

Kamakura, Lu, & Mason, 2006), yet not well addressed in management accounting literature. In essence we extend the concept of CLV and the works of several researchers to the arena of management accounting. The challenge for internet company performance measurement systems is moving from the current focus on measuring web data (such as clickthrough rate, cost of acquisition (CoA), and churn rate) to managing the relation between CLV and CoA. This will permit the investigation of the relationships between customer behaviour (through web metrics) and financial data (Epstein, 2007). The aim of this chapter is to develop a framework for understanding the causal relationships among drivers of customer profitability in internet pure player companies. Our approach, which we designate as the Lifetime Value Scorecard, provides an integrated framework for evaluating internet pure player strategies intended to maximize customer base value. To develop the model, we examine the case study of a hypothetical mobile content & service provider in the sector of mobile value-added services (VAS), where content and services are provided to consumer cell phones, mainly using wireless networks. In fact, the worldwide mobile value-added services market has recorded extremely high growth rates in the last three years (Bertele`, Rangone, & Renga, 2007). In addition, the Italian example is an exportable model at an international level because Italy is a leader in this industry, thanks to the proliferation of mobile users. The chapter is structured as follows: Section 2 provides a synthesis of the literature, Section 3 formalizes the research question and discusses the research method, while Section 4 traces the evolution of internet company strategic choice for the maximization of the customer base. Section 5 presents the Lifetime Value Scorecard. Section 6 provides a description of the mobile VAS industry used to test the framework we suggest. In the final section we will offer some concluding remarks.

2. LITERATURE REVIEW The literature on internet companies coming from international academics and professionals started to flourish during the euphoric period of the bubble as well as after its collapse. A wide range of topics was touched on, including business model aspects, the financial evaluation of internet companies, planning and control, and the value of customers.

196

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

Focusing on the analysis of the literature dealing with the variables significant to our purposes, we can group it into four main areas that cover: a. b. c. d.

business model; market value; planning and control; CLV.

a. Business model Literature on the business model developed greatly just around the period of the internet euphoria. This is not incidental; in fact, there was an attempt to label and define many new businesses that were set up in those years. In this wealth of literature, some authors tried to define the business model concept and to identify its purpose and key elements by analysing their relation to strategies and business processes. (Amit & Zott, 2001; Magretta, 2002; Osterwalder & Pigneur, 2002; Tapscott, Lowy, & Ticoll, 1998; Tapscott, Lowy, & Ticoll, 2000; Timmers, 1998). Other authors explored the field of the business model in terms of components or sub-components, with an evolution and an overlapping area of analysis with respect to previous ones (Afuah & Tucci, 2000; Alt & Zimmermann, 2001; Hamel, 2000; Magretta, 2002; Osterwalder & Pigneur, 2002; Weill & Vitale, 2001). There are further studies aimed at identifying criteria for assessing the feasibility and profitability of business models or evaluating them in comparison with other alternatives (Afuah & Tucci, 2000; Hamel, 2000; Weill & Vitale, 2001). They tend to use financial indicators even if they are often difficult to measure in this context. The business model is an ongoing field of research which aims to explain how these companies operate. b. Market value With regard to literature on market value, studies about traffic data on sites owned by publicly listed internet companies show that the more popular sites provide significantly better stock returns than the less popular ones, thanks to their superior ability to attract advertising revenues and extract greater compensation from affiliated sites (Lazer, Lev, & Livnat, 2001). Literature on customer acquisition cost and market value highlights that investments in CoA, in order to be qualified as assets, require a solid basis for an expectation of a stable, multiyear customer base. For management and investment purposes, it is useful to include output measures in customer

The Lifetime Value Scorecard

197

indicators, capturing the value of the intangible, so as to associate them with a firm’s market value, even if they are not standardized and publicly reported, and consequently not comparable across firms. The positive correlation between customer acquisition cost and market value, clearly visible before the bubble collapsed, confirmed the assumption immediately after that a reasonable expectation of a multiyear stream of benefits from customers, demonstrated as historically stable, is needed (Demers & Lev, 2001; Lev, 2001). Internet traffic measures can also be associated with a firm’s value as long as they reflect the actual purchase behaviour of customers rather than the number of visits and time spent on the site. Thus, as more and more technology firms got listed on financial markets – often at very early stages in their life cycles – traditional valuation methods and metrics often seemed ill suited to them. In reality, while the estimation challenges are different for these firms, the fundamentals of valuation do not and should not change in evaluating technology firms (Copeland, Koller, & Murrin, 2000; Damodaran, 2001). c. Planning and control With reference to the planning and control systems of dot-coms and click and mortars, we can find few contributions, many of them with a strict focus on the customer. An attempt to link customer acquisition cost and financial performance was made by identifying a tableau de board with the main categories of indicators (Keen, 1999):  Customer relationship to evaluate its duration (repeat business customer retention, revenue growth rate per established customer, net cash flow value for customer transaction).  Operational efficiencies to measure the impact of the internet on the cost structure (customer service cost effectiveness, total relationship cost per customer).  Balance sheet strength: measures linked to Economic Value Added and shareholder value (working capital per unit of revenue). Other indicators such as hits per page and impressions were commonly used as web metrics in the 2000s, but ‘‘hits are a poor measure of business performance because they reflect site design, thus failing to adequately reflect customer behaviour. Hence, hits should not be used as a core metric for measuring success’’ (Cutler & Sterne, 2000). The data that must be managed are only those of customers considering certain ratios such as: reach,

198

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

acquisition, conversion, retention, loyalty, abandonment, attrition, and churn that represent different stages of the customer life cycle (Table 1). The contributions of Keen and Sterne and Cutler point out the central role of the customer, but both suffer from a lack of a systemic framework. Also the e-performance scorecard is customer-centric: the three categories of indicators – attraction, conversion and retention – express the efficiency of e-business cost, and effectiveness of a site’s operation. Some of these indicators are static, while others are dynamic, but this scorecard has nothing to do with the balanced scorecard; in fact, the only perspective considered is that of the customer (efficiency is implicitly measured while innovation and financial indicators are totally missing). It has an external and ex-post focus aimed at finding a relationship between e-indicators and financial performance of internet companies according to an investor point of view. What lacks is a Table 1.

Different Stages of Customer Life Cycle.

 Reach refers to the potential to gain the attention of your target audience. A different way to calculate reach is as a percentage of the total potential market. In the reach stage, the goal is primarily that of awareness  Acquisition, in which the goal is customer participation. Once a potential prospect takes an overt action expressing interest (e.g., signing up for a newsletter, taking a survey, joining a discussion list, filling out a form in return for a white paper, or downloading a demo), a new prospect has been acquired. But, here actually making the sale is still very much in question  Conversion is the point at which a sale has been made and the prospect has been converted into a customer. However, in the online realm, conversion has a myriad of meanings depending on the goals of the particular site, confirming the difficulty in having a unique definition of customer ratios  Retention implies the marketing assumption that it is far less costly to sell additional products or services to an existing, retained customer than it is to find and secure a new one. Consistent retention is what makes the cost of customer ownership so much lower than customer acquisition. It is what gives one customer a higher lifetime value than another one  Loyalty is the act of binding yourself intellectually or emotionally to a course of action. The customers with the greatest lifetime value are generally those who are not only loyal to the products, but also loyal to the company. On the Web, loyalty also refers to site visits over time  Abandonment rate is the number of people who commence but do not complete the buying process  Attrition rate is the percentage of existing, converted customers who have ceased buying and have gone elsewhere during a specific period of time  Churn measures how much of the customer base rolls over during a given period of time. To calculate churn, divide the number of customers who attrite during the given time period by the total number of customers at the end of the time period Source: Cutler and Sterne (2000).

The Lifetime Value Scorecard

199

systemic framework that links them in a cause and effect relationship in an inside and forward-looking perspective (Agrawal et al., 2001). Focusing on B2C, literature on e-indicators based on customers seems to be concerned about justifying poor economic results. This explains the substitution of short-term measures with customer-oriented measures and periodical measures with cross-periodical ones. On the basis of this background, Maccarone (2002) suggests that according to the value-based approach, the firm can be considered a result of a portfolio of long-term projects, each of which contributes to creating the value of the firm. This sacrifices short-term profitability in order to maximize long-term performance. Therefore, the value of the firm is the sum of the value of the firm’s customers along its life cycle. The limit of this financial approach is the missing link in the firm’s strategy: indicators such as retention rate or customer conversion rate represent the success of the strategy, but they are qualitative indicators and require further measures of key success factors in the performance system. Maccarone points out the necessity to shift towards a long-term perspective and to adapt the performance measurement system with indicators coming from the web site that can change according to the evolution of strategic priorities during the different phases of the life cycle. However, his proposal is not so clear in terms of a management control system for internet companies. As for CRM, literature suggests an implementation of a strategic CRM performance measure, to involve the entire organization in contributing to CRM strategy success. The approach reflects the balanced scorecard philosophy even if it is focused only on the CRM system (Brewton & Schiemann, 2003). There is an effort to put the indicators focused on customer perspective into a systemic framework with the aim of evaluating the achievement of the CRM goals at each level of the organization. The integration of the customer perspective of the balanced scorecard with few measures such as CLV or customer retention is not sufficient to translate the CRM strategy and link it to customer profitability objectives. Brewton and Schiemann (2003) offer a systemic conceptualization, but it is strictly focused on the CRM system and many internet companies often lack CRM systems. The idea that many studies have focused on a particular perspective and consequently on a particular stakeholder group comes from Neely, Marr, Adams, and Kapashi (2002) who, in fact, propose a performance prism to draw a picture about what internet companies do and should measure, analysing the status of measurement exigencies with respect to many

200

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

stakeholders and prescribing different indicators for each of them. However, they do not develop a measurement system for this type of company. Further literature has developed customer profitability metrics in order to measure the success of customer-based initiative, but always in a CRM context (Peppers & Rogers, 2004). d. Customer lifetime value The literature on CLV is mainly the result of the transition from productcentric to a customer-centric context, but it is narrowly focused on specific aspects with a marketing connotation in most cases. Some contributions have tried using a mathematical approach to determine the lifetime value (LTV) in evaluating the customer assets (customer equity) by establishing the determinant variables in the LTV calculation (De Bonis, Balinski, & Allen, 2002; Rust, Zeithaml, & Lemon, 2000). This kind of literature is strictly concentrated on the calculation of the LTV as a single value, overlooking other wider strategic implications. Also customer behaviour has been examined in depth, establishing a particular set of measures coming from a value-based measurement priorities matrix, in order to exactly understand customer exigencies and to improve customer value (Doyle, 2002). Although the structure is characterized by a sequence of steps, a systemic framework is lacking. The lifetime model also finds applications in evaluating marketing efforts in terms of acquiring or retaining customers and selling them more products, indicating how much you should spend on each function. ‘‘The real value of lifetime value is not in the numbers, but in the use that is made of the numbers. Without imaginative strategies, the LTV numbers are useless’’ (Hughes, 2003). This approach tries to use lifetime value tables to evaluate the effectiveness of various relationship-building strategies. Notwithstanding the step forward towards a cause and effect relation between LTV and strategy, there is not a holistic view of the e-enterprise, but a focus on marketing function. In a marketing background, LTV can be used to evaluate how a cost of a campaign or of a product R&D can be allocated to the target customers (Peppers & Rogers, 2005). However, the LTV equations used for these kinds of assessments are based, in our opinion, on variables with a high degree of subjectivity. Further literature underlines the difficulties in estimating LTV especially if it is based on historic data (Faulkner, 2003) or future revenue projection. In addition, there is the need to know what it costs to serve that customer and this requires a sophisticated information system (Barnes, 2006).

The Lifetime Value Scorecard

201

The difficulties stemming from the estimation of CLV on the basis of historic data are highlighted also by Malthouse and Blattberg (2004), who propose a model over a long period rather than an entire lifetime, in order to obtain a higher accuracy of prediction of the LTV. A systemic framework based on LTV has been drawn to assess an entire marketing mix (Yang, 2005). The historic marketing mix reveals how retention efforts can eventually turn a costly acquisition into a profitable outcome. However, long-term profitability may not be sufficient to ensure an optimal investment. In fact, an investor also needs to investigate a time horizon for achieving a breakeven. Thus, even though this marketing mix is ultimately profitable, investors must simultaneously consider if they can financially bear the loss within the first three years. The model develops a new benchmark, LTV BE% (break even rate) for decision making, aiming at long-term benefits. In substance, there is in this field of studies robust research analysing the effects of marketing actions on customer retention and CLV (Bolton, Lemon, & Verhoef, 2004; Reinartz & Kumar, 2003). Recently there was an exploration of the link between CLV and financial performance (Gupta & Lehmann, 2003; Gupta et al., 2004). Further developments in literature have highlighted the importance of managing customer value with a ‘‘systematic approach for addressing customer value issues that includes: customer segmentation, measuring profitability, estimating CLV, identifying additional sources of customer value, and managing to enhance customer profitability. It demonstrates how organizations can create more value for and derive increased value from the customer’’ (Epstein & Yuthas, 2007, p. 5).

3. RESEARCH STRATEGY The ‘‘fil rouge’’ linking the literature on planning and control and CLV mentioned above is an attempt to give a specific tool for new exigencies coming from the digital processes using in part the Balanced Scorecard logic but not the instrument, or in other cases further indicators to add to the existing ones to understand web trends. What lacks instead, in our opinion, is a systematic framework in which the e-metrics are integrated with a cause and effect relationship and strategically considered also from a financial perspective. Performance measurement systems in internet companies focus mainly on measuring web data, such as click-through rate, CoA, and churn rate. While this approach does offer information about customer behaviour when visiting

202

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

a company’s site or when interested in making a transaction, the indicators, taken separately, do not provide useful translation into financial data. For this reason, a systemic framework that can examine in depth a cause and effect relationship between web metrics and financial information is necessary. In this context, the research question is: what kind of evolution does the performance measurement system have to deal with in order to be able to manage data coming from the web for supporting strategic decisions in a B2C internet company? In particular, it is necessary to understand how CLV at the individual level, and consequently the customer equity at the firm level, are affected by changes in marketing strategy in a forward looking perspective. After having drawn the theoretical model, the research question is tested using the mobile VAS industry as a case study. What is interesting about this industry is that the environment and consequently the strategies evolve in a timely fashion. This industry is relatively new, so there is not a significant number of cases to deal with. However, not only can case study be used as an empirical base, but also as experiments in relation to theory (Grandori, 1991). In particular, our illustrative case study (Scapens, 2004) aims to point out the results achieved in a real context and how the performance measurement system should evolve to keep up with the pure player’s environment and in particular with the information needs. As mentioned above, the choice of this particular business model as a research field is legitimated by the great expansion of this kind of market, where Italy has a leadership position, as shown in Table 2, making the Italian model exportable at the international level. Table 2. Italy Mobile Mobile Mobile Mobile

a

telephone users – ’000 telephone calls – mn minutesb SMS messages sent – mnc telecommunication revenuesd

Trends in Mobile Industry in Italy. 2001

2002

2003

2004

2005

51.246.0 72.000.0 11.600.0 45.9

54.200.0 74.468.6 18.067.47 47.7

56.770.0 75.839.5 22.187.49 48.7

62.750.0 76.660.6 25.941.40 39.8

71.500.0 77.278.5 34.197.39 39.8

2006

2007

76.083.5 78.438.3 77.820.9 78.319.9 45.080.89 NA 40.3 41.5

Sources: Euromonitor International (2008). a Mobile telephone users: International Telecommunications Union/World Bank/Trade Sources/Euromonitor International. b Mobile telephone calls: Euromonitor International from International Telecommunications Union. c Mobile SMS messages sent: Euromonitor International from International Telecommunications Union. d Mobile telecommunication revenues: % of telecom revenue – Euromonitor International from International Telecommunications Union.

203

The Lifetime Value Scorecard

4. INTERNET COMPANY STRATEGIES: CUSTOMER BASE MAXIMIZATION AND ONLINE MARKETING It is widely known that attracting and retaining customers is critical to corporate success. However, for internet pure player success, it is even more crucial than for other industries. In fact, internet companies are predominantly service-based and derive revenue from the creation and sustenance of long-term relationships with their customers. The strategic choices of internet pure players are focused on maximizing the value of their customer base, which is a business asset whose potential value needs to be measured in order to understand the effectiveness of their strategies (Gupta & Lehmann, 2005; Kumar & Petersen, 2005; Peppers & Rogers, 2005). Thus to manage customer-based strategies these companies have to pay attention to: acquisition, retention, and growth (Fig. 1): a. Acquistion: characterized by aggressive marketing techniques of customer acquisition aimed at building up a new user base. Sometimes these direct marketing techniques go along with promotional campaigns that try to instil ‘‘brand awareness’’ into the target segments.

GROWTH (user base growth, tackle with churn)

ACQUISITION (customer acquisition techniques)

RETENTION (user loyalty, CRM and lifetime increase)

Fig. 1.

The Main Strategies for Managing Customer Base.

204

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

b. Retention: focused on measuring the lifetime of users and their value. In this phase, techniques that typically stimulate user retention and minimize churn rate are adopted. c. Growth: once the user base is built and the churn rate is under control, the focus shifts towards planning for the organic growth of the user base, and defining the new client target acquisition number for each period (e.g. month or quarter) in order to balance churn rate and reach the target growth rate. To manage the customer base we have to define it. In doing so we will consider the life cycle of customers, arriving from the target market and progressing towards an established loyal customer base. Of course, along the way, many individual customer life cycles are cut short by abandonment and attrition. Customer base is generally categorized in a. Browsers: users that browse the web site. b. Profiled users: users that provide information about themselves and their behaviour in exchange for commercial benefits such as an information newsletter. c. Clients: users that make purchases. d. Subscribers: users that subscribe to recurring services. A conversion relationship exits between each one of the ‘‘states’’ mentioned above (browsers; profiled users; clients; subscribers). The value of users to a company increases as they go from one state to the next. What is extremely important is that the company should be able to assign a value to each one of these states and carefully monitor it over time. For these reasons dot-coms have to create a long-term relationship with customers, and pay attention to their lifetime value. Internet marketing initiatives are the critical factors for internet pure players in order to acquire, retain, and increase the customer base. By the new millennium, however, disillusionment with online advertising (and its dominant technique, the so-called banner Ad) set in, as it appeared to be an ineffective method of reaching customers. But this was not to be the end of the story: a new method of online advertising, termed paid search (Laffey, 2007), was to emerge whereby the use of search engines would trigger a display of advertisements based on the topic of the search. Paid search would become the dominant form of online advertising, powering revenues past peak levels of the dot-com boom as organizations were attracted to its benefits (Bughin, Erbenich, & Shenkan, 2007; Grosso, Shenkan, & Sichel, 2006).

The Lifetime Value Scorecard

205

The collapse of banner advertising left search engines/portals desperate for new sources of revenue. In 1998, a California-based start up, Goto.com (later renamed Overture), emerged on the scene, offering a solution via a new method of advertising known as paid search. This revolutionary approach would transform online advertising. Advertising comes in many forms, but they all have one thing in common: a desire to replace the old approach, where advertisers paid for the privilege of ‘‘exposing’’ a theoretical audience to their message, with a new one where advertisers only pay for real and measurable actions by consumers, such as clicking on a web link, sharing a video, placing a call, and printing a coupon or buying something (Economist, 2006b). In the paid search model, listing position is determined by how much an advertiser is prepared to pay for a keyword or phrase. When a user searches for a specific keyword, the order of the results they see is determined by current bids in the auction. Payment is made by advertisers each time a user searches for a term and then clicks on their link. Monitoring the behaviour of click-throughs via paid search is essential, as it provides a precise measurement of how successful the advertising method is in terms of achieving the objectives set forth. Data collected from such tracking should then be fed back into the process to make performance more effective. It may be, for example, that poor quality leads are being attracted (perhaps through the use of the wrong keywords) or that clicks from certain sources work better than others. Furthermore, monitoring can be used to ensure that a firm is not paying too much for clicks; for instance, being in position b or c may generate as much business as being in position a. Ensuring that customers are taken to an appropriate landing page is also seen as crucial (Epstein, 2007). From a management accounting point of view the innovation that effectively addresses these challenges is the possibility to calculate the CoA of each single customer. This is possible, thanks to the link between the advertising (and its costs) and the number of customers acquired when Ads are displayed. This is very useful since it helps to assess the efficiency of customer acquisition methods and evaluate sales and distribution strategies. In fact, advertising is evolving from ‘‘cost of contact’’ (or cost per impression) to effective ‘‘CoA’’. The evolution can be traced looking at the different price models, in particular (Sterne, 2002): a. Cost per thousand impressions (CPM): The CPM model refers to advertising bought on the basis of impressions. This is in contrast to the

206

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

types of pay-for-performance advertising, whereby payment is only triggered by a mutually agreed upon activity (i.e. click-through, registration, sale). b. Cost per click (CPC): Introduced by search engines; thanks to keyword advertising, it is becoming one of the dominant models. Cost is determined by effective clicks on the advertisement, regardless of the number of impressions required to generate it. The other key feature is the process that determines the price to be paid that comes from a bid process among different advertisers on a specific keyword. c. Cost per acquisition (CPA): CoA of either a profiled customer or a onetime sale or subscriber. In a nutshell, we are moving forward in the value chain, where the advertising network shares the risk of the campaign by trusting in an effective conversion rate of the retailer. A relationship exists among the different models, and can be measured with user behaviour metrics such us click-through rate and conversion rate, in particular: CPM ! click-through rate ðCTRÞ ! CPC ! conversion rate ðCRÞ ! CPA

Media buyers pay attention to this relationship in order to evaluate the costs of each Ad (Economist, 2007a). For example if you have to allocate a marketing budget of $20,000, under the hypothesis exemplified in Table 3, you can choose between the following options: a. CPM of $0.02 in order to get 1,000,000 potential customers who see your ad; b. CPC of $2.00 in order to reach 10,000 potential customers who click on your ad; c. CPA of $25.00 in order to acquire a total of 800 customers. This evolution in adverting also implies an evolution in the marketing strategies that since the earlier stages of the internet era have shown new exigencies in measuring their effectiveness.

5. CUSTOMER LIFETIME VALUE FRAMEWORK Measuring e-business transactions was one of the consequences ‘‘forced’’ by the new economy in order to understand customers’ behaviour. However,

207

The Lifetime Value Scorecard

Table 3.

The Relations among Ads Price Model.

Marketing budget (a) Cost per thousand impressions Estimated impressions Cost per impression (b) Cost per click Estimated impressions Estimated click through rate Estimated clicks Cost per click (c) Cost per acquisition Estimated clicks Estimated conversion rate Number of new costumers Cost per acquisition

20,000 1,000,000 0.02 1,000,000 1% 10,000 2.00 10,000 8% 800 25.00

new metrics coming from the web indicating different paths that customers take when visiting a site cannot be considered stand-alone as simple indicators. They need to be integrated into a framework of measures, capable of explaining the development of the business (Epstein, 2007). The question is raised herein (Berger & Nasr, 1998): how much is it worth spending to acquire a new customer today, the cost of which will be recovered through future business dealings with that customer? As such, three major issues should be settled: first, CLV has to be defined from a mathematical point of view; second, it is necessary to organize the useful data, and third, and most important, a systematic framework has to be constructed in order to assess the relationship among CLV and its drivers.

5.1. The Algorithm CLV is the present value of the future cash flows a customer will contribute to the business as long as the customer remains such (Gupta & Lehmann, 2005; Novo, 2004; Pfeifer, Haskins, & Conroy, 2005). In general a CLV model has two components: a. contribution margin (customer value); b. customer length of service (lifetime).

208

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

Most standard introductions to the notion of CLV centre around a formula similar to CLV ¼

T X ðpt  ct Þ t¼0

rt ð1 þ iÞt

(1)

where pt ¼ price paid by a consumer at time t ct ¼ cash cost of servicing the customer at time t i ¼ cost of capital for the firm rt ¼ the probability of the customer repeat buying or being ‘‘alive’’ at time t T ¼ time horizon for estimating CLV At a purely conceptual level, the calculation of CLV is similar to the discounted cash flow approach used in finance. However, the reality is more complex. The main differences are a. Customer segments need to be identified, in fact there is no average customer. b. Cash costs need to be traced to the customers. c. Lifetime estimation plays a crucial role in the formula, in fact there is the need to explicitly incorporate the possibility that a customer may defect in the future. While the first two points will be discussed in details in the rest of the chapter, let us concentrate our attention on how to calculate lifetime, i.e. the amount of time a customer is estimated to remain such and create value before defecting. Lifetime is usually described as a survival function which explains the probability that the customer will still be active at a certain time (Gupta et al., 2006; Fader & Hardie, 2007; Rosset, Neumann, Eick, & Vatnik, 2003). In Eq. (1) lifetime is a function of retention rate. The retention rate for period t(rt) is defined as the proportion of customers active at the end of period t1 who are still active at the end of period t. Usually, for the mobile company, we speak about churn rate which is the opposite of the retention. In fact the churn rate for a given period is defined as the proportion of customers active at the end of period t1 who dropped out in period t. If we take for granted that past and current customer behaviour is the best predictor of future customer behaviour, it is possible to take a series of past retention numbers and project them into the future in order to make a prediction about expected churn rate.

209

The Lifetime Value Scorecard

As analyses of the various parametric and non-parametric models are not the subject of this work, we will refer to the model of Gupta and Lehman (2003) that considers an infinite time horizon. In fact, retention rate accounts for the fact that over time, the chances of a customer staying with the companies decrease significantly. They also demonstrate that the typical method for the conversion of retention rate to expected lifetime (1/churn rate) and then the calculation of present value over that finite time period overestimates lifetime value. In addition they show that, using infinite time horizon, if margin and retention rate are constant over time, Eq. (1) can be simplified as follows CLV ¼

1 X

ðp  cÞ

t¼0

rt r ¼m 1þir ð1 þ iÞt

(2)

In other words, CLV simply becomes margin (m) times a margin multiple that depends on customer retention probability and the company’s cost of capital. As one of the main reasons for calculating CLV is to determine what you can pay for customer acquisition, it is important not to include what you pay for Ads in the CLV, since they are discretional costs (Berger & Nasr, 1998; Hogan et al., 2002). It is more useful to explicitly compare the two figures: CLV versus CoA. The difference between CLV and CoA is what we indicate as net customer lifetime value. NCLV ¼ m

r  CoA 1þir

(3)

By using Eq. (3) managers can decide how much to spend on prospective customer acquisition. They are also able to check the difference in profitability among customer segment.

5.2. The Data Mart To adopt the CLV model all the required data have to be prepared, and data tracking and mining could be the biggest obstacle. A useful tool could be a data mart, i.e. a database, or collection of databases, designed to help managers make strategic decisions about their business. Whereas a data warehouse combines databases across an entire enterprise, data marts are usually smaller and focused on a particular

210

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

subject or department and always tailored to support the specific pattern of reporting for which they are designed. With a restricted scope of content for analytical processes, they serve a single department, part of an organization, and/or a particular data analysis problem domain. Thus, a data mart is defined by its function and/or audience; in our specific case it is focused on customers. In fact, CLV analysis is based on a customer’s historical transaction record, which should be consistent, robust, and clean. The dataset keys are the following (Yang, 2005): a. Customer: Uniquely identified by customer ID, the customer is the core within an entire relational CLV data mart. Each individual customer can be tracked, evaluated, and scored in terms of historic transactions from invoices. b. Invoice: One invoice number has multiple relations to customer IDs and promotion codes since we need to know at least: i. source of marketing promotion, ii. first purchase date, iii. billing event, iv. last purchase date. c. Acquisition: After each marketing campaign there is the necessity to track the cost of customer acquisition (cost of the campaign/no. of acquired customers). This is feasible because the promotions are typically customer driven, linking the cost of the campaign (airtime, keyword bid, etc.) with the purchase of a product through a specific code to type on the web or dial on the mobile phone. It is important to understand that there is no average customer: a segmentation is necessary in order to obtain different customer groups, each of them with their own market behaviour. For instance, the main segmentation criteria (Kotler, 1999) in the Mobile VAS industry, could be:  Socio-demographic characteristics such as age, sex, marital status, children, education, occupation, and personal income. Together, these features help to profile the internet user and to provide evidence for the emergence of a new customer segment with specific exigencies.  Customer tenure. It is demonstrated that the retention rate among different tenure cohorts may be different. For example, it is commonly observed that customers with long tenure such as a loyal customers are much more likely to be retained as loyal customers than are new customers.

211

The Lifetime Value Scorecard

Table 4.

Linking Customer Profile to Economics. Profile A

Profile A

Profile A

Total

Revenue Cost of services Contribution margin CoA Customer margin Other fixed cost EBIT

 Type of media used to attract the customer. The type of media or offer used to attract customers can have a dramatic effect on the long-term behaviour, and customers who come into business on the same media will tend to behave in similar ways over time.  Billing methods. Operators usually develop a pricing structure that supports groups of users that behave differently from one another. For example, an operator may target a segment of users of a particular service with a subscription type offer (e.g. flat fee), while other segments pay on a per item basis. Another example is to offer ad-supported MMS (with insertion of advertisements in messages) for lower prices to the teen and pre-paid segment, while charging full price to other segments. The subdivision into segments can be carried out according to the different acquisition campaigns, the distribution channel or the time that the subscribers have spent with the company. For example, with the shifted-beta-geometric model, Fader and Hardie (2007) have demonstrated that the retention rate (defined as the opposite of the churn rate) is an increasing function of time. Therefore, the longer the time that subscribers are with the company, the lower the probability that they are going to churn. The criteria must be defined on the basis of their relation to customer behaviour that influence the main drivers of CLV. Once customer profiles are identified, a coherent reporting has to follow. An example of such a reporting is presented in Table 4. 5.3. The Scorecard From a management accounting perspective, the challenge is not only measuring customer value, but managing it as well. In other words, it is necessary to link customer behaviour, and the actions that could influence it, with future financial results.

212

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

This goal cannot be reached solely through financial data that fails to provide vital management input. In fact, they neither tell of a future decline in business due to a faulty change in marketing strategy, nor do they predict a future rise in profitability due to a beneficial return on the original marketing strategy. Only a performance management accounting system based on measuring and managing the future value of a customer can do this. In fact, the critical difference between CLV and periodic accounting is the way time is handled. Firms using customer accounting understand and believe in the concept of lifetime value; they know that as long as they continue to acquire customers with higher and higher lifetime values, the periodic financial statements will reflect their positive impact. Firms that are managed by relying only on periodic financial statements may even be surprised when profitability rises or falls due to operational or advertising changes that affect the customer behaviour. A performance measurement system has thus to integrate specific lag and lead indicators (Epstein & Manzoni, 1998) capable of providing an early sign about whether a strategy is being implemented successfully. This is the core idea behind the net lifetime value scorecard: tracking the relationships between customer behaviour and customer value (Fig. 2).

Measures

ARPU

Sales per each customer segment

Cost of product/service

Contribution margin per customer

Retention

Churn rate

Usage

# Transactions

Attraction

Reach rate

Conversion

Conversion rate

Marketing policies

Cost of contact

CONTRIBUTION CONTRIBUTION MARGIN

NET NET LIFETIME LIFE TIMEVALUE VALUE

LIFETIME

COST OF COSTOF ACQUISITION ACQUISITION

feedback

Fig. 2.

The CLV Scorecard.

The Lifetime Value Scorecard

213

To make the scorecard operational, it is necessary to first track the net customer lifetime value (NCLV), for instance each month, week, day, depending on the nature of the company, and then to investigate the drivers of each component contributing to change the NCLV. In fact, it is not as important to know the absolute or exact value of a customer as it is to know whether this value is rising or falling over time. In particular, the main drivers are: contribution margin, cost of customer acquisition, and lifetime. In this context, particular attention has to be paid to the relation between marketing campaign, lifetime, and CoA. This relation could be investigated though the analysis of lead indicators that are able to track the relation between marketing campaigns and changes in customer behaviour that typically precede change in customer value. This means that if you track these changes in behaviour, you can forecast a change in value, and consequently you can direct your marketing campaign accordingly. For instance, churn rate is a useful metric to track changes in customer retention over time and to help evaluate how changes in business and service offerings affect customer retention. To complete the analysis, once data on indicators has been collected, statistical tools, such as multiple regression, should be used to analyse and test the validity of the customized model as hypothesized. The implementation of the NCLV scorecard could be useful not only in determining what you can pay for customer acquisition cost, but also for other managerial decisions, such as 1. Estimating the future revenue given a certain amount of money you intend to spend in marketing. As already mentioned, this would be possible looking at past data using statistical analysis such as multiple regression; 2. Evaluating the company through the embedded value of subscriber base, i.e. customer equity (Gupta et al., 2004); 3. Improving business reporting (Lev, 2001; IASB, 2005; Wiesel, Skiera, & Villanueva, 2008). Regarding the third point, it is worthwhile to mention that some companies are already supplementing and complementing financial information with unaudited non-financial data such as customer base, churn rate, and CoA. This is the case of the initial public offering of Virgin Mobile USA, Inc., where the S-1 form contains the data reported in Table 5.

Other Data (Unaudited)

Table 5.

2,666,194 4.3 993,625 3,844,777 $ (48,275) $ 22.54 $ 14.94 $ 118.62

Virgin Mobile USA, Inc. Disclosure of Customer Data.

2,328,830 3.9 1,422,855 2,851,152 $ (133,567) $ 24.24 $ 16.85 $ 131.58

3,013,781 4.8 729,313 4,574,090 $ 47,884 $ 21.48 $ 13.15 $ 120.55

214

Gross additionsa Churnb (%) Net customer additonsc End-of-period customersd Adjusted EBITDA (in thousands)e,f Average revenue per usere,g CCPUe,h CPGAe,i

Source: Form S1 pages 14–16. Available on Edgar data base: http://www.sec.gov/Archives/edgar/data/1396546/000119312507097779/ds1.htm a Gross additions represents the number of new customers that activated our handsets during a period, unadjusted for churn during the same period. In measuring gross additions, we begin with handset activations and exclude any customer that has replaced one of our handsets with another one, retailer returns, customers who have reactivated within seven months of deactivation and fraudulent activations. b Churn is used to measure customer turnover. Churn is calculated as the ratio of the net number of customers that disconnect from our service to the weighted average number of customers, divided by the number of months during the period being measured. The net number of customers that disconnect from our service is calculated as the total number of customers that disconnect less the adjustments noted under gross additions above. These adjustments are applied in order to arrive at a more meaningful measure of churn. Churn includes those customers who we automatically disconnect from our service when they have not replenished their account for 150 days as well as those customers who voluntarily disconnect from our service. We believe churn is a useful metric to track changes in customer retention over time and to help evaluate how changes in our business and services offerings affect customer retention. In addition, churn is also useful for comparing our customer turnover to that of other wireless communications providers. c Net customer additions represents the number of new customers that activated our handsets during a period, adjusted for churn during the same period. d End-of-period customers are the total number of customers at the end of the period being measured. e We use several financial performance metrics, including Adjusted EBITDA, ARPU, CCPU and CPGA, which are not calculated in accordance with generally accepted accounting principles (GAAP). A non-GAAP financial metric is defined as a numerical measure of a

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

company’s financial performance that (i) excludes amounts, or is subject to adjustments that have the effect of excluding amounts, that are included in the comparable measure calculated and presented in accordance with GAAP in the statement of operations or statement of cash flows or (ii) includes amounts, or is subject to adjustments that have the effect of including amounts, that are excluded from the comparable measure so calculated and presented. We believe that the non-GAAP financial metrics that we use are helpful in understanding our operating performance from period to period, and although not every company in the wireless communication industry defines these metrics in precisely the same way that we do, we believe that these metrics as we use them, facilitate comparisons with other wireless communication companies. These metrics should not be considered substitutes for any performance metric determined in accordance with GAAP. f Adjusted EBITDA is calculated as net income (loss) plus depreciation and amortization, interest expense, non-cash compensation expense, equity issued to a member and debt extinguishment costs. We find Adjusted EBITDA to be useful as a measure for understanding the performance of our operations from period to period. g Average revenue per user, or ARPU, is used to measure and track the average revenue generated by our customers on a monthly basis. ARPU is calculated as net service revenue for the period divided by the weighted average number of customers for that period being measured, further divided by the number of months in the period being measured. The weighted average number of customers is the sum of the average customers for each day during the period measured divided by the number of days in that period. ARPU helps us to evaluate customer performance based on customer revenue and forecast our future service revenues. h Cash cost per user, or CCPU, is used to measure and track our costs to provide support for our services to our existing customers. The costs included in this calculation are our cost of service (exclusive of depreciation and amortization), excluding cost of service associated with initial customer acquisition, general and administrative expenses, excluding any marketing, selling, and distribution expenses associated with initial customer acquisition, non-cash compensation expense, net loss on equipment sold to existing customers, cooperative advertising expenses in support of existing customers and other (income)/expense, net of debt extinguishment costs. These costs are then divided by our weighted average number of customers for the period being measured, further divided by the number of months in the period being measured. CCPU helps us to assess our ongoing business operations on a per customer basis, and evaluate how changes in our business operations affect the support costs per customer. Given its use throughout the industry, CCPU also serves as a standard by which we compare our performance against that of other wireless communication companies. i Cost per gross addition, or CPGA, is used to measure the cost of acquiring a new customer. The costs included in this calculation are our selling expenses, our net loss on equipment sales (cost of equipment less net equipment revenue), excluding the net loss on equipment sold to existing customers, equity issued to a member, cooperative advertising in support of existing customers and cost of service associated with initial customer acquisition. CPGA helps us to assess the efficiency of our customer acquisition methods and evaluate our sales and distribution strategies. CPGA also allows us to compare our average acquisition costs to those of other wireless communication companies.

The Lifetime Value Scorecard

215

216

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

6. THE MOBILE VALUE-ADDED SERVICE CASE STUDY There is not only one right way of calculating CLV across all businesses. First of all, it depends on the industry to which you apply the formula, and also on the cost structure and the revenue stream of the single company. For this reason, we present a case study, in order to show what has been achieved in practice and to highlight the evolution needed in the performance measurement system in dealing with internet environment. We apply the advocated framework to the mobile VAS industry that is populated by typical internet pure players. In 2005, the worldwide mobile VAS market recorded extremely high growth rates that were difficult for other segments to match. The latest market analyses – including Ovum Research – estimate the global VAS market at approximately 18 billion euro for 2005, and predict that this figure could more than double by 2010, with a compound average growth rate of about 22% for the 2005–2010 period. The main catalysts for growth include: the increase in the number of mobile telephone users around the world, which is expected to grow from 2 billion currently to 3 billion in 2010, and the widening of the age range of VAS consumers, with the number of VAS mobile users expected to increase from 15% currently to 30%. Furthermore, the increasing penetration of mobile phone use in emerging markets, such as Latin America, Russia, and Asia Pacific, will lead to a partial reduction in average revenue per user (ARPU) from 50 to 40 euro. The Italian market is one of the most developed worldwide (Fig. 3). According to a study (Bertele` et al., 2007), the Italian market in 2006 accounted for about 1.031 million euros with a 14% growth from the previous year (903 million euros in 2005). The current scenario is very promising in maintaining growth over the next couple of years. The advent of new multimedia services (such as video download) and of mobile TV (under the Digital Video Broadcasting Handheld standard) seems to guarantee further development of the market, where Telco, traditional, and new media companies fight to gain market shares and play a relevant role. In this industry a major role is played by mobile content & service providers (MCSP) that provide content and services to consumer cell phones mainly, but not exclusively, using, wireless networks. Some examples of the biggest players in this industry worldwide are Buongiorno Vitaminic, DADA, Jamba (Verisign), LaNetro Zed.

217

The Lifetime Value Scorecard 1200 1031 903

million EURO

900

816

600 408 253

300

0 2002

Fig. 3.

Content origination

2005

2006

Italian Mobile VAS Market. Source: Bertele` et al. (2007).

Mobile Content & Service providers

Service management

Aggregation & publishing of content and services in different formats (sms, MMS, WAP..)

Fig. 4.

Media companies

Telcos Billing

Mktg & Display

Marketing and promotion of VAS through MEDIA: TV, print or online networks

Network delivery

Network used to distribute VAS to end users (Mobile operators)

User

CRM

Service demand

R&D activity Design and project of the product/service

2004

Payment

Content Owners

2003

Mobile Vas Value Chain.

The mobile VAS value chain is characterized by the following activities (Fig. 4): a. b. c. d. e. f.

content origination; service management; marketing and display; network delivery; customer relationship management; billing.

218

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

To measure business efficiency and effectiveness, it is necessary to define each activity in terms of revenue generation, resource absorption, and value drivers. a. Content origination The content origination relates to the creation of content that is suitable to be transferred and used in mobile cell phones. The content plays one of the key roles in this industry, representing one of the reasons to buy a mobile data service. Content can be classified according to different categories.  News: information services (breaking news, sports, trading news);  Phone personalization: ring tones, real tones, wallpaper, and everything that is related to self-expression through personalization of the cell phone;  Entertainment: java games, time killers, and other self-generated content;  User-generated content: content that is generated into a community and that is made accessible to other community members through wap navigation. The players of this piece of the value chain are basically publishers (music or games publishers) and distributors (the major record labels in the music industry) who distribute, in digital form, content that is traditionally produced for other markets, such as traditional CDs. The same thing happens in the news market where traditional newspaper or TV publishers implement a digital extension of their platform to deliver their content through the wireless channel. b. Service management This activity is performed by players that aggregate content from different originators and make it available to wireless carriers for distribution to their user base. These players, who are usually defined as MCSP, manage and run a platform that is able to    

Ingest the content from many originators and aggregate it. Maintain a physical connection to the various wireless carriers. Deliver the content to the wireless carriers upon request. Adapt the content according to the different format requests by the user agent.  Request the billing to wireless carriers.  Manage subscriber lists on subscription-based services.

The Lifetime Value Scorecard

219

The business model of MCSP is very scalable since it relies on a technology platform that has a negligible marginal incremental cost. For example, doubling the customer base does not require a change in the equipment plant. c. Marketing and display Marketing can be done either off-line (print, TV, and other below-the-line tools) or online. In the last couple of years we have seen a shift from off-line advertising to online advertising. As we have already mentioned the penetration of the internet in the marketing mix allows for the model to emphasize the measurement of CoA in the various stages (browser, profiled user, buyer, subscriber – see above). d. Network delivery The content delivery happens through wireless networks. So, wireless carriers have a key role. The delivery can be made through:  Standard SMS: Extremely popular application for the growth of this industry, the SMS is still very effective for news and alerts.  MMS: Maintains the characteristics of the SMS, but it adds some features of multimedia which are very useful for enhancing the user experience.  Wap-push: Consists of an alert via WAP message (i.e. a message containing a link to a WAP page); by clicking on the message users can open a WAP page where they can download the requested content. This is very commonly used for downloading content.  Wap 2.0: This is a real html page (or xhtml to be more precise) that makes the user experience very similar to standard web navigation. When the delivery happens, the owner of the relationship with the end-user has to pay for:  the clearance of the rights to content originators;  delivery and transportation costs to the wireless network. e. Customer relationship management Since this industry is pretty new and the value chain is so convoluted, many problems can happen while performing a download. Therefore, the customer relationship management becomes another very important piece in the value chain. Customers may experience problems in downloading or interacting with the platform. The CRM is usually performed by wireless carriers and the content and service providers that establish a main point of contact via IVR,1 mail, or live agents to solve those problems. An analysis of

220

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

the main problems encountered by users is crucial to understanding churn and to improving the lifetime. f. Billing In the mobile VAS industry the billing is performed by wireless carriers through premium SMS. The content and service providers that have a relationship with carriers will receive an out-payment of what carriers are able to provide to the end-users. The out-payment is used to pay the delivery of content and service, to make the clearance on rights, and to pay the subscribers CoA. One of the key activities in any internet transaction is the ability to effectively make users pay. According to the payment means requested (credit card, debit card, paypal, PayPal mobile account) a different successful billing rate (i.e. the effective yield of the billing process) may be achieved. Failure in billing can vary from very low percentages, for example when credit cards are used, to pretty large percentages when the payment is performed using the telephone account (SIM card). In the latter situation leakage is mainly linked to some factors, the first and most common is the lack of credit on the pre-paid card, or because of overspending in the account. To render the framework operational we provide a simple comprehensive example, using a hypothetical company that will be called ‘‘i-Tones’’, that operates as a MCSP in several countries in all continents. A brief description of i-Tones is as follows: ‘‘i-Tones’’ works with all major record labels who supply contents and has agreements with all major carriers who provide the network with content delivery and bill customers on their mobile account. Customers are reached predominantly through paid search advertisement, using word such as ‘‘free-ringtones’’. Subscribers have to pay one fixed fee monthly subscription plan that provides the right to receive a specified number of downloads of content.

Given the above, the business model for ‘‘i-Tones’’ would consist of  Investing in customer acquisition through direct marketing and advertising.  Raising revenue through direct billing to end-users made on its behalf by carriers and getting its out-payments from them.  Managing the platform for providing services to end-users and incurring the related fixed costs, plus paying the delivery fee for content distribution.  Clearing rights for distributed content through the proper holders. Each activity has an impact on the profit and loss of the business (Fig. 5). According to the business model defined above, it is possible to draw a Lifetime Value Scorecard that can describe the main value drivers for

221

The Lifetime Value Scorecard

Content origination

Service management

Mktg & Display

Network delivery

User

Revenues Cost of delivery Contribution margin Cost of acquisition Customer Margin Fixed cost EBIT

Fig. 5. Mobile and Content Provider Profit and Loss.

‘‘i-Tones’’ (Fig. 6). These drivers can be considered as lead indicators and must be monitored on a daily basis to verify that a daily value creation process is in place. In theory, the issue of the time value of the money also has to be considered, usually discounted in the CLV formula, in order to get an idea of future profits today. Discounting may be very appropriate in some businesses, particularly very long-cycle, high ticket retail and B2B. However, we believe that this practice may confuse the issues in the B2C where the environment is very fast and it is better to be simple. On one hand the lifetime is forecasted according to the churn rate of the current period, taking into consideration a forecast of the churn rate in the future. On the other hand, CoA is calculated dividing the expenses relating to advertising on all various channels (including, web, TV, press, etc.) to acquire new subscribers and the customers acquired in the considered unit of time. In synthesis, we can say that this comparison between CoA and CLV is vital for players in this market. Therefore, a real time tool showing the trend of those data is highly recommended. The tool should show CLV and CoA:  for specific campaigns in place (if different marketing campaigns are in place);  in different countries in case of international business;  for each hour, to monitor different consumer behaviour;  in the target audience (if the campaign is targeting different segments).

222

M. BONACCHI, M. FERRARI AND M. PELLEGRINI Measures

ARPU

successful billing rate

Cost of product/service

Cost of delivery Cost per download

Retention

Instant/Historic churn

Usage

# downloads

Attraction

CTR, # visits

Conversion

Conversion rate

Marketing policies

CPM, CPC, CPA

CONTRIBUTION CONTRIBUTION MARGIN MARGIN

NET NET LIFETIME VALUE

LIFETIME LIFETIME

COST OF ACQUISITION ACQUISITION

feedback

Fig. 6.

The CLV Scorecard of ‘‘i-Tones’’.

For example, the analysis can be conducted by looking at the retention behaviour of newly acquired customers to understand which lifetime those users are projecting. Therefore, a daily or hourly fine-tuning of the campaign should be performed: for example, by changing the bid price on the keyword advertising. Needless to say, this never ending process is driven by many factors, some of which are not under the control of the company, such as competitor actions, regulation changes, or technological discontinuities, any of which may affect the consumer behaviour in both conversions and retention.

7. CONCLUSIONS The speed of business that is completely based over the internet is a big challenge for existing performance measurement systems; they need to be adapted to support decisions in a timely fashion. The evolution of business models, in fact, has involved a development of the planning and control system, which was forced, to a certain extent, to evaluate ‘‘new market exigencies’’ with the consequence of taking a step

223

The Lifetime Value Scorecard

forward in measuring a particular intangible aspect – the value of customers – through more objective variables. The Lifetime Value Scorecard model we propose is aimed not only at evaluating the effectiveness of marketing strategies, but more importantly it is a way to measure and manage the intangible asset represented by customers of internet pure players that is crucial for an accurate evaluation process and improving financial reporting. The model is able to translate web metrics of internet pure player companies into financial data, developing a framework for understanding the causal relationships among drivers of customer profitability and to assess the implementation degree of strategies intended for maximizing customer base value. The Lifetime Value Scorecard can be considered stand-alone if the e-enterprise is of small to medium size, otherwise it could be integrated into a more complex planning and control system such as the balanced scorecard, in order to deeply investigate customer perspective. Another alternative could be, in absence of a BSC, an integration into a CRM system. In fact, a flexible and integrated measurement system capable of capturing the peculiarity of any kind of business could represent a guide for strategic decisions and also a possible competitive advantage. This framework could be replicated with some adaptations in other industries that have on line activities (telecommunications, banks, etc.) both with an examination of case studies and with statistical analysis, in order to investigate how customer features, such as socio-demographic characteristics, customer tenure, and type of media used, affect subscription period and their lifetime value.

NOTE 1. Short for interactive voice response, a telephony technology in which someone uses a touch-tone telephone to interact with a database to acquire information from or enter data into the database. IVR technology does not require human interaction over the telephone as the user’s interaction with the database is predetermined by what the IVR system will allow the user access to.

ACKNOWLEDGMENTS The authors would like to thank participants at the fourth Workshop on Performance Measurement and Management Control for their comments and

224

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

suggestions. Although this chapter is the result of teamwork, Massimiliano Bonacchi can be considered the author of Sections 1, 4 and 5; Mascia Ferrari of Sections 2, 3 and 7; Massimiliano Pellegrini of Section 6.

REFERENCES Afuah, A., & Tucci, C. L. (2000). Internet business models and strategies: Text and cases. New York: McGraw-Hill. Agrawal, V., Arjona, L. D., & Lemmens, R. (2001). E-performance: The path of rational exuberance. McKinsey Quarterly (1), 30–43. Alt, R., & Zimmermann, H.-D. (2001). Preface: Introduction to special section – Business models. Electronic Markets, 11(1), 3–9. Amit, R., & Zott, C. (2001). Value creation in e-business. Strategic Management Journal, 22(6/7), 493–520. Barnes, J. G. (2006). Build your customer strategy. New York: Wiley. Berger, P. D., & Nasr, N. I. (1998). Customer lifetime value: Marketing models and applications. Journal of Interactive Marketing, 12(1), 17–30. Bertele`, U., Rangone, A., & Renga, F. (2007). Mobile content: Aspettando mobile internet e Mobile TV. From http://www.osservatori.net Bolton, R. N., Lemon, K. N., & Verhoef, P. C. (2004). The theoretical underpinnings of customer asset management: A framework and propositions for future research. Academy of Marketing Science Journal, 32(3), 271–292. Brewton, J., & Schiemann, W. A. (2003). Measurement: The missing ingredients in today’s customer relationship management strategies. Journal of Cost Management, 17(1), 5–14. Bughin, J., Erbenich, C., & Shenkan, A. (2007). How companies are marketing online: A McKinsey global survey [Electronic Version]. McKinsey Quarterly. Retrieved on November 27, 2007 from http://www.mckinseyquarterly.com/links/27538 Copeland, T., Koller, T., & Murrin, J. (2000). Valuation: Measuring and managing the value of companies. New York: Wiley. Court, D. C. (2005). Boosting returns on marketing investment. McKinsey Quarterly (2), 36–47. Cutler, M., & Sterne, J. (2000). E-metrics business metrics for the new economy. NetGenesis Corporation. http://www.targeting.com Damodaran, A. (2001). The dark side of valuation. New York: Prentice Hall. De Bonis, J. N., Balinski, E., & Allen, P. (2002). Value-based marketing for bottom-line success. New York: McGraw-Hill. Demers, E., & Lev, B. (2001). A rude awakening: Internet shakeout in 2000. Review of Accounting Studies, 6, 331–359. Doyle, P. (2002). Value based marketing for bottom line success. New York: McGraw-Hill. Economist. (2006a). E-commerce [Electronic Version]. The Economist from http://www. economist.com/background/displayBackground.cfm?story_id=7138905 Economist. (2006b). The ultimate marketing machine. The Economist, 61–64. Economist. (2007a). Many ways to skin a cat. The Economist, 80–82. Economist. (2007b). Word of mouse. The Economist, 77–78.

The Lifetime Value Scorecard

225

Epstein, M. J. (2007). Evaluating the effectiveness of internet marketing initiatives. The Society of Management Accountants of Canada, the American Institute of Certified Public Accountants and The Chartered Institute of Management Accountants. Epstein, M. J., & Manzoni, J. F. (1998). Implementing corporate strategy: From Tableaux de Bord to balanced scorecard. European Management Journal, 16(2), 190–203. Epstein, M. J., & Yuthas, K. (2007). Managing customer value. The Society of Management Accountants of Canada, the American Institute of Certified Public Accountants and The Chartered Institute of Management Accountants. Fader, P. S., & Hardie, B. G. S. (2007). How to project customer retention. Journal of Interactive Marketing, 21(1), 76–90. Faulkner, M. (2003). Customer management excellence. Chichester: Wiley. French, T. D. (2007). Confronting proliferation y in online media: An interview with Yahoo!’s senior marketer [Electronic Version]. McKinsey Quarterly, 18–27. Retrieved on November 27, 2007 from http://www.mckinseyquarterly.com/Marketing/Digital_ Marketing/Confronting_proliferation__in_online_media_An_interview_with_Yahoo!s_ senior_marketer_2018 Grandori, A. (1991). Epistemologia ed economia aziendale: Note per un dibattito. Finanza Marketing e Produzione, 9(2). Grosso, C., Shenkan, A. G., & Sichel, H. P. (2006). A reality check for online advertising. McKinsey Quarterly (3), 10–11. Gupta, S., Hanssens, D., Hardie, B., & Kahn, W. (2006). Modeling customer lifetime value. Journal of Service Research, 9(2), 139. Gupta, S., & Lehmann, D. R. (2003). Customers as asset. Journal of Interactive Marketing, 17(1), 9–24. Gupta, S., & Lehmann, D. R. (2005). Managing customers as investment. The strategic value of customer in the long run. Upper Saddle River, NJ: Wharton School Publishing. Gupta, S., Lehmann, D. R., & Stuart, J. A. (2004). Valuing customers. Journal of Marketing Research, 41(1), 7–18. Gupta, S., & Zeithaml, V. (2006). Customer metrics and their impact on financial performance. Marketing Science, 25(6), 718. Hamel, G. (2000). Leading the revolution. Boston: Harvard Business School Press. Hogan, J. E., Lehmann, D. R., Merino, M., Srivastava, R. K., Thomas, J. S., & Verhoef, P. C. (2002). Linking customer assets to financial performance. Journal of Service Research, 5(1), 26–38. Hughes, A. M. (2003). The customer loyalty solution. New York: McGraw-Hill. IASB. (2005). Management commentary. London: International Accounting Standards Board. Keen, P. (1999). New sources and measures of value for eCommerce. Accenture Internal Report. Kotler, P. (1999). Kotler on marketing: How to create, win, and dominate markets. New York: Free Press. Kumar, V., & Petersen, J. A. (2005). Using a customer-level marketing strategy to enhance firm performance: A review of theoretical and empirical evidence. Academy of Marketing Science Journal, 33(4), 504. Laffey, D. (2007). Paid search: The innovation that changed the web. Business Horizons, 50(3), 211–218. Lazer, R., Lev, B., & Livnat, J. (2001). Internet traffic and portfolio returns. Financial Analysts Journal, 57(3), 30.

226

M. BONACCHI, M. FERRARI AND M. PELLEGRINI

Lev, B. (2001). Intangibles: Management, measurement, and reporting. Washington, DC: The Brookings Institution Press. Maccarone, P. (2002). The design of performance measurement system for internet ‘‘pure plays’’: Is a new paradigm really needed? In: M. J. Epstein & J. F. Manzoni (Eds), Performance measurement and management control: A compendium of research. Elsevier. Magretta, J. (2002). Why business models matter. Harvard Business Review, 80(5), 86–92. Malthouse, E. C., & Blattberg, R. C. (2004). Can we predict customer lifetime value? Journal of Interactive Marketing, 19(1), 2. Neely, A., Marr, B., Adams, C., & Kapashi, N. (2002). Measuring ebusiness performance. In: A. Neely (Ed.), Business performance measurement. Cambridge: Cambridge University Press. Neslin, S. A., Gupta, S., Kamakura, W., Lu, J., & Mason, C. H. (2006). Defection detection: Measuring and understanding the predictive accuracy of customer churn models. Journal of Marketing Research, 43(2), 204–211. Novo, J. (2004). Drilling down turning customer data into profits with a spreadsheet (3rd ed.). Saint Petersburg: Jim Novo. Osterwalder, A., & Pigneur, J. (2002). An ebusiness model ontology for modeling ebusiness. Paper presented at the 15th Bled Electronic Commerce Conference-eReality: Constructing the eEconomy, Bled, Slovenia, June 17–19. Peppers, D., & Rogers, M. (2004). Managing customer relationships. Hoboken, NJ: Wiley. Peppers, D., & Rogers, M. (2005). Return on customer. New York: Currency Doubleday. Pfeifer, P. E., Haskins, M. E., & Conroy, R. M. (2005). Customer lifetime value, customer profitability, and the treatment of acquisition spending. Journal of Managerial Issues, 17(1), 11–25. Reinartz, W. J., & Kumar, V. (2003). The impact of customer relationship characteristics on profitable lifetime duration. Journal of Marketing, 67(1), 77–99. Rosset, S., Neumann, E., Eick, U., & Vatnik, N. (2003). Customer lifetime value models for decision support. Data Mining and Knowledge Discovery, 7, 321–339. Rust, R., Zeithaml, V., & Lemon, K. (2000). Driving customer equity: How customer lifetime value is reshaping corporate strategy. New York: The Free Press. Scapens, R. W. (2004). Doing case study research. In: C. Humphrey & B. Lee (Eds), The real life guide to accounting research. Oxford: Elsevier. Sterne, J. (2002). Web metrics-proven methods for measuring web site success. New York: Wiley. Tapscott, D., Lowy, A., & Ticoll, D. (1998). Blueprint to the digital economy: Creating wealth in the era of the E-business. New York: McGraw-Hill. Tapscott, D., Lowy, A., & Ticoll, D. (2000). Digital capital. Boston: Harvard Business School Press. Timmers, P. (1998). Business models for electronic markets. Journal on Electronic Markets, 8(2). Varian, H. R. (2006). The economics of internet search. Rivista di Politica Economica (IX–X), 177–191. Weill, P., & Vitale, M. R. (2001). Place to space. Boston: Harvard Business School Press. Wiesel, T., Skiera, B., & Villanueva, J. (2008). Customer equity: An Integral part of financial reporting. Journal of Marketing, 72(2), 1–14. Yang, A. X. (2005). Using lifetime value to gain long-term profitability. Journal of Database Marketing & Customer Strategy Management, 12(2), 142–152.

EXAMINING THE CONSTRUCT VALIDITY OF THE BALANCED SCORECARD USING THE MULTITRAIT–MULTIMETHOD MATRIX Emilio Boulianne ABSTRACT The Balanced Scorecard (BSC) proposes four dimensions to represent business performance: Financial, Customer, Innovation and Internal Business, and Learning and Growth. Surprisingly, little attention has been paid to the BSC as a valid construct representing performance. Despite surveys reporting that a growing number of firms use the BSC, little is known about the reliability and validity of the measures and dimensions it proposes. Validity problems impact on the importance and credibility allocated by management to certain BSC measures. This study’s objective is to empirically examine the construct validity of the BSC. Through a literature review and field study, a set of measures associated with all four BSC dimensions is selected. Next, survey research is conducted to examine the reliability of selected measures and the structure of BSC dimensions. Lastly, we examine the convergent and

Performance Measurement and Management Control: Measuring and Rewarding Performance Studies in Managerial and Financial Accounting, Volume 18, 227–251 Copyright r 2008 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-3512/doi:10.1016/S1479-3512(08)18010-8

227

228

EMILIO BOULIANNE

discriminate validity of the BSC’s measures using the multitrait– multimethod (MTMM) matrix. Results show that the BSC – with its four dimensions and related measures – represents a valid construct. This study responds to research calls on the importance of validating the BSC framework – and its associated measures – in order to enhance consistency on BSC research.

1. INTRODUCTION To obtain better assessments of business-unit performance, a consensus was reached among academics and practitioners regarding the design of performance-measurement systems to consider non-financial measures (Chenhall & Langfield-Smith, 1998; Ullrich & Tuttle, 2004). The Balanced Scorecard (BSC), introduced by Kaplan and Norton (1992), received a great deal of interest in this respect and was one of the major topics examined within management accounting research throughout the past decade. The BSC proposes the following four dimensions to represent performance: Financial, Customer, Innovation and Internal Business, and Learning and Growth. Surprisingly, little attention has been paid to the BSC as a valid performance measurement model, the form originally proposed by Kaplan and Norton.1 Despite surveys reporting that a growing number of firms use the BSC, little is known about the reliability and validity of the measures and dimensions it proposes (Ittner, Larcker, & Meyer, 2003; Chenhall, 2003). Surveys report executives are concerned about the quality and reliability of the non-financial performance measures utilized in BSCs (Lingle & Schiemann, 1996; Ittner & Larcker, 2001; Reck, 2001). Kaplan and Norton (2001) report that the reason for delay in BSC implementation may be that firms have not developed reliable and valid measures for the scorecards. Problems with reliability and validity of measures have also impacted on the credibility and importance allocated by managers to the BSC’s dimensions (Lipe & Salterio, 2002). Changes in importance are functions of the credibility of BSC measures. One year, for example, a firm weighted at 20% the BSC’s Learning and Growth dimension for performance evaluation, to turn around the next year and weight it at a mere 4% because management considered the measures associated with that dimension not reliable (Malina & Selto, 2001). This study’s primary objective is to empirically examine the construct validity of the BSC. Using a literature review and field study, we begin by

Examining the Construct Validity of the Balanced Scorecard

229

selecting a set of common BSC measures associated with all four BSC dimensions. As indicated by Lipe and Salterio (2000) and Dilla and Steinbart (2005), only common measures qualify as evaluations of performance across business units. Our survey examines the reliability of these measures, as well as the structure of the four BSC dimensions. Furthermore, we use the multitrait–multimethod (MTMM) matrix to examine the convergent and discriminant validity of the measures selected. According to Brownell, ‘‘the multitrait-multimethod matrix is an extremely powerful approach, much underutilized in management accounting research’’ (1995, p. 51). Results show that the BSC, with its four claimed dimensions and related measures, represents a valid performance model. Chenhall (2003) points out the importance of developing robust and valid BSC measures to enhance the consistency between various studies on the BSC. Others find it is important to first establish the validity of measures used and the associated framework before suggesting new business models (Ittner, Larcker, & Meyer, 1997). In response to these concerns, the present study provides empirical evidence using measures that have passed rigorous validity steps. To the best of our knowledge, no other study has empirically investigated the BSC from a construct validity perspective. This text is organized as follows: Section 2 presents a literature review; Section 3 describes the research methodology; Section 4 reports on the reliability of the BSC measures and the structure of the BSC dimensions; and Section 5 examines the convergent and discriminant validity of measures. The last section discusses limitations and presents a conclusion.

2. LITERATURE REVIEW The importance of reliable and valid measures for performance measurement has already been pointed out in organizational psychology literature (Blum & Naylor, 1968). For instance, subjective measures for performance assessments are often considered to be less accurate and reliable than objective measures because they may be influenced by the rater’s biases (Heneman, 1986; Campbell, 1990). Reliability is also regarded as an important factor in the choice of performance measures (Ittner et al., 2003). According to Malina and Selto (2001) ‘‘to be effective as a management control device, the BSC should result in evaluations that are accurate’’ (p. 75). We begin our analysis with the content validity step.

230

EMILIO BOULIANNE

2.1. Selection of BSC Dimensions and Measures Content validity refers to the use of appropriate measures and dimensions to represent a construct based on existing literature (Kwok & Sharp, 1998). The literature on the BSC clearly proposes four dimensions: Financial, Customers, Innovation and Internal Business, and Learning and Growth. The literature suggests that these dimensions may be essential to all organizations (Malina & Selto, 2001) and puts forward a list of measures associated with each dimension. The measures can be unique to particular units or common to units. Kaplan and Norton (1996) state that all BSCs have certain common measures that are relevant across units, a core measurement set regardless of business objectives, and that units at the same organizational level have common measures. This belief has been supported by Lipe and Salterio (2000) and Dilla and Steinbart (2005) who report that participants in experiments evaluated their division’s performance based solely on common measures, unique performance measures having no effect on managers’ evaluation judgments. These observations support the viewpoint that the BSC should include critical performance measures, which are mainly reflected through measures common to numerous units. It may therefore be considered appropriate to use a set of common performance measures in examining several business units. In addition to content validity, we conducted a field study with managers to determine which common measures to retain. This initiative is used to obtain face validity – the use of appropriate measures and dimensions from the point of view of managers (Kwok & Sharp, 1998).2 A survey research was then performed among several business units using the measures retained from the field study. The next section describes the methods used to collect the information.

3. RESEARCH METHODOLOGY 3.1. Field Study To obtain validation and feedback from business managers we employed a field-study research method.3 In order to maintain equilibrium between external and internal validity, we relied on the deliberate sampling for heterogeneity approach from Cook and Campbell (1979) for the selection of business units. The selection of units was based on the following two criteria: (1) size, based on number of employees and (2) industry, based on

Examining the Construct Validity of the Balanced Scorecard

231

manufacturing or service sector. This approach intentionally creates variance by looking for different sizes and industries and restricts the sample to only those units that meet these criteria. Considering external and internal validity, we aimed to have managers from 4 to 6 units participate in interviews. Participation of more units would provide a better understanding of results and add to their generalizability; however, interviewees’ availability vs. time required for the commitment limited the number of units we could interview. With the two selection criteria in mind, we worked with a professional accounting organization, the Certified General Accountants of Canada, Quebec region, to obtain participation of interested managers.4 Four managers from four different business units showed an interest in participating. We discussed with managers of different balanced scorecards (different dimensions and measures) based on Kaplan and Norton (1996, 2001, 2004) and Kaplan and Atkinson (1998). The managers were asked to indicate which measures they considered to be: (1) relevant for performance evaluation and (2) common to different business units. Managers were told that there are no right or wrong scorecards, dimensions, or measures. These managers were knowledgeable about the BSC approach. The four business units included in the field study are located in Montreal, Canada. The interviews were 45–65 min in duration. The profiles of the units are as follows: Unit A is large in size, part of a multi-unit company, and operates in the manufacturing industry; Unit B is large in size and operates in the service industry; Unit C is medium in size and operates in the manufacturing industry; and Unit D is small in size and operates in the service industry. The number of employees in units ranges from 50 to 190. The field study helped us to elaborate on the information to obtain for the research survey. Table 1 shows the measures considered by managers relevant for performance evaluation and common to different business units, together with the respective dimensions. The selection of measures takes account of the managers’ judgment and experience; this is a limitation of the present research. For the Financial dimension, return on assets (ROAs) and net profit margin intended to reflect financial performance, and working capital ratio asset utilization. For the Customer dimension, marketing expenses to revenues ratio reflects marketing efforts to solicit new customers, and revenue growth is a proxy for market share. Two measures were selected to represent the Customer aspect while previous research has often used a single measure (see, e.g., Banker, Potter, & Srinivasan, 2000; Ittner et al., 1997). We have observed in some BSC studies that revenue

232

EMILIO BOULIANNE

Table 1. Measures, by Dimension, Considered by Managers Relevant for Performance Evaluation and Common to Different Business Units. Dimensions Financial

Customer Innovation and Internal

Learning and Growth

Measures Return on asset Net profit margin Working capital ratio Marketing expenses to revenues Revenue growth Number of new products Number of products offers R&D expenses to revenues Employee absenteeism rate Employee turnover rate Training expenses to revenues Revenue per employee

growth appears either in the Financial dimension or in the Customer dimension, depending on the nature attributed to this measure. Revenue growth may be seen as an indicator of financial performance, or as an indicator of competitiveness with a customer focus, reflecting the relative market share and position. (For example, growth in sales volume appears in the Customer dimension of Nova Scotia Power’s scorecard; see Kaplan & Norton, 2001, p. 122). In examining the literature, other performance measurement models have a similar classification. In Lynch and Cross’s (1991) Performance Pyramid Model (1991), revenue growth is associated with the Market dimension, instead of the Financial dimension, and in Fitzgerald, Johnston, Brignall, Silvestro, and Voss’s (1991) Determinants and Results Matrix, revenue growth is associated with the Competitiveness dimension instead of the Financial performance dimension. The managers interviewed found revenue growth to be an appropriate proxy for market share as the latter is very difficult to obtain at the businessunit level. For the Innovation and Internal Business dimension, number of new products introduced over the last three years, number of product offers, and R&D expenses to revenues reflect innovation initiatives. Finally, for the Learning and Growth dimension, employee absenteeism rate and employee turnover rate reflect employee satisfaction, training expenses to revenue reflects commitment to employee development and training efforts, and revenue per employee reflects employee productivity.

Examining the Construct Validity of the Balanced Scorecard

233

3.2. Survey Research A research survey and secondary data analysis were employed to collect from business units the measures in Table 1. Since we intended to perform a MTMM matrix to examine convergent and discriminant validity, we collected data from different sources and modes of assessment. As managers are reluctant to permit disclosure of objective data on their units (see for instance Chenhall, 1997),5 we once more worked with CGA-Quebec to support the study and used their members’ directory to select business units. We contacted members by telephone and first asked whether the firm was organized as a business unit. For those units that fulfilled these criteria, we explained the nature of the study, which required two respondents per unit, and elaborated upon the information they would be asked to provide. To encourage participation, respondents were promised summarized outcomes of the study. Two questionnaires were forwarded to the firms that agreed to participate. The first questionnaire asked for objective financial and non-financial data/measures, to be answered by the person who occupied the highest position in accounting/finance in the unit. The second questionnaire asked for subjective assessments, to be answered by the person who occupied the highest management position in the unit. Questionnaires were reviewed for clarity.6 Fig. 1 summarizes data sources and modes of assessment. The 380 units that agreed to participate received questionnaires.7 We conducted three telephone reminders at intervals of two weeks, four weeks, and six weeks. We received return questionnaires from 128 firms, although responses from 38 firms were eliminated either because the questionnaires were incomplete, or because they were received from only one of the two respondents. In the present study, to count for one observation, we required both questionnaires fully completed from the same business unit. The sample consequently consists of 90 pairs of usable questionnaires, for a response rate of 24% – a rate similar to other studies that have used multiple respondents (e.g., Chan, Huff, Barcaly, & Copeland, 1997; Sabherwal & Chan, 2001 used multiple respondents and obtained response rates of 19% and 7%, respectively). The main reasons mentioned for non-participation in the study were confidentiality concerns regarding objective data and the involvement of two respondents per business unit. Concerning the highest management position respondents, previous research indicates time constraints, rather than the nature of the questionnaire, as a potential reason for non-participation (Assael & Keon, 1982). The profile of the person holding the highest position in accounting/finance is a comptroller who holds a bachelor’s degree in commerce with an

234

EMILIO BOULIANNE

Objective (Based on records)

Mode of Assessment Subjective (Based on judgment)

Persons with the highest position in accounting/ finance were asked to provide financial and non-financial data/measures.

Persons with the highest management position were asked to provide subjective assessments on their business units. Primary (Data obtained from the business unit)

Fig. 1.

Relative position of indicators, by quartile, obtained through a database

(Not examined)

Secondary (Data obtained from outside the business unit) Data source

Data Sources and Modes of Assessment Utilized (adapted from Venkatraman et al., 1987).

accounting designation and has an average age of 42 years. The profile of the person holding the highest management position is a general manager who holds a bachelor’s degree in commerce, and has an average age of 45 years. They have held their current positions for an average of 8 years, and have been employed by the same firm for an average of 14 years. At the business-unit level, the average number of employees is 156, with average revenues of 22 million Canadian dollars.8 To estimate the non-response bias, we compared late respondents vs. early respondents and results indicate that we do not have the presence of non-respondent bias. Table 2 provides descriptive statistics of measures collected, while Table 3 provides a correlation matrix. The instruments used to collect the data are presented in the appendix and explained in the following sections.

4. RELIABILITY OF MEASURES AND FACTORIAL VALIDITY (STRUCTURE) OF DIMENSIONS We used Cronbach’s alpha coefficient to estimate the internal consistency of measures and to analyze the quality of the BSC as a performancemeasurement model. As in other fields, Cronbach’s alpha is the most

Maximum

1–4 1–4 1–4 1–4 1–7 1–7 1–3

Theoretical Range

Descriptive Statistics of Measures Collected. Minimum

4 4 4 4 7 7 3

Table 2. SD

1 1 1 1 2 1 1

Mean

Measures

0.99 0.94 0.79 1.03 1.25 1.63 0.76

4 4 4 4

1–4 1–4 1–4 1–4

Examining the Construct Validity of the Balanced Scorecard

1 1 1 1

Theoretical range does not apply for Panel B

Panel A: Subjective assessment obtained from managers Return on assets (ROA) 3.06 Net profit margin (NPM) 3.02 Working capital ratio (WC) 3.10 Revenue growth (REVGR) 2.90 Employee absenteeism rate (ABS) 5.41 Employee turnover rate (TURN) 5.24 Innovation (INO) 1.97

0.80 0.79 0.91 0.56

Panel B: Financial and non-financial data obtained from position in accounting/finance respondents Return on assets (ROA) 9.49 6.52 4.70 31.40 Net profit margin (NPM) 5.05 5.36 3.00 35.00 Working capital ratio (WC) 1.73 1.25 0.19 10.90 Marketing expenses to revenues (MRK) 0.02 0.04 0 0.32 Revenue growth (REVGR) 6.81 18.04 38.5 71.2 Number of new products (NEWP) 15.29 19.64 0 50 Number of product offers (POFF) 20.28 35.36 1 90 R&D expenses to revenues (R&D) 0.0095 0.0158 0 0.12 Training expenses to revenues (TRAI) 0.0026 0.0056 0 0.02 Revenue per employee (RPE) 254,300 465,233 25,000 4,277,992 Panel C: Relative positioning found with the financial database Return on assets (ROA) 2.78 Net profit margin (NPM) 3.06 Working capital ratio (WC) 2.44 Revenue growth (REVGR) 3.43 Note: n ¼ 90.

235

Panel A ROA NPM WC REVGR ABS TURN INO Panel B ROA NPM WC MRK REVGR NEWP POFF R&D TRAI RPE Panel C ROA NPM WC REVGR

Note: n ¼ 90.

INO

0.00 0.05 0.05 0.02

ROA

Table 3.

TURN

Subjective Assessments from Managers ABS

REVGR

NPM

1.00

WC

ROA

1.00 0.09

0.03 0.06 0.01 0.01

0.04 0.13 0.12 0.08

0.81 0.58 0.58 0.17

Correlation Matrix.

WC

0.15 0.07 0.09 0.62

1.00 0.39 0.11 0.07 0.06 0.16 0.08 0.07 0.08

1.00 0.13 1.00 0.01 0.03

0.27 0.14 0.13 0.04

1.00

RPE

1.00 0.03 0.17 0.02

0.20 0.09 0.02 0.21 0.03 0.2. 0.09 0.16

TRAI

0.06 0.01 0.01 0.02

MRK REVGR NEWP POFF R&D

Financial and Non-Financial Data from Accounting/Finance Respondents NPM

0.15 0.05 0.05 0.03

1.00 0.45 1.00 0.04 0.06 1.00 0.12 0.08 0.08 0.15 0.02 0.11 0.17 0.22 0.12 0.17 0.05 0.24 0.02 0.14 0.04 0.51 0.07 0.71 0.21 0.69 0.21 0.08 0.12

0.05 1.00 0.17 0.71 1.00 0.08 0.07 0.26 0.12 0.16 0.04 0.14 0.11 0.04 0.01 0.08 0.37 0.30 0.06 0.01 0.07 0.10 0.00 0.09 0.01 0.04 0.11 0.27 0.14

1.00 0.43 0.03

0.29 0.25 0.27 0.39

0.04 0.13 0.27 0.01 0.06 0.07 0.01 0.04 0.15 0.02

1.00 0.43 1.00 0.56 0.42 1.00 0.12 0.07 0.07 0.05 0.04 0.01 0.16 0.06 0.11

0.23 0.24 0.24 0.04

0.27 0.06 0.14 0.12 0.01 0.04 0.07 0.02 0.48 0.10 0.06 0.11 0.04 0.03 0.18 0.02 0.143 0.13 0.08 0.21

1.00 0.90 0.51 0.54 0.04 0.08 0.21

0.47 0.38 0.39 0.09

0.39 0.20 0.46 0.23 0.00 0.26 0.08 0.01 0.00 0.09 0.29 0.17 0.33 0.14 0.30 0.11 0.02 0.13 0.12 0.02 0.11 0.18 0.16 0.10 0.02 0.03 0.14 0.07 0.04 0.04 0.50 0.35 0.36 0.14

Pearson correlation is significant at the 0.01 level, significant at the 0.05 level, (2-tailed).

ROA

WC

1.00

REVGR

Relative Positioning

NPM

1.00 0.57 1.00 0.56 0.99 1.00 0.12 0.01 0.020

237

Examining the Construct Validity of the Balanced Scorecard

Table 4. Measures with Cronbach Alpha Coefficient Per Dimension. Dimension

Financial

Customer

Innovation and Internal

Learning and Growth

Measure

Cronbach Alpha Coefficient

Alpha if Item Deleted After 1st Iteration

Final Cronbach Alpha

Return on asset Net profit margin Working capital ratio Marketing expenses to revenues Revenue growth Number of new products Number of product offers R&D expenses to revenues

0.64

0.21 0.05 0.82 0.23

0.64

Employee absenteeism rate Employee turnover rate Training expenses to revenues Revenue per employee

0.51

0.03 0.00

0.42

0.55

0.01 0.55

0.00

1st iteration

2nd iteration

0.00

0.00

0.00

0.00

0.00

0.58

0.43

deleted

0.58

Note: n ¼ 90. Descriptive statistics for the measures above are available in Table 2, Panel B, and for the employee absenteeism rate and the employee turnover rate in Panel A. Due to the high kurtosis index, Innovation and Internal Business measures have been transformed using the square foot for use in reliability analyses. The nine measures in bold will be examined in further analyses.

recognized estimation of internal consistency reliability in accounting research (Brownell, 1995; Kwok & Sharp, 1998). The BSC is a construct with four identifiable dimensions designed to represent business unit performance; coefficient alpha must therefore be calculated for each dimension (Churchill, 1979). Table 4 presents the BSC measures along with Cronbach’s alpha coefficients for each dimension. As we can see in Table 4, we obtained a Cronbach’s alpha coefficient of 0.64 for the three measures of the Financial dimension. This coefficient would be higher if we delete the working capital ratio but we are keeping it because of its sound content validity and because at an early stage, a coefficient of around 0.60 is considered reasonable (Nunnally, 1967).

238

EMILIO BOULIANNE

For the Customer dimension the alpha coefficient is 0.51, which shows that the two measures are compatible enough for purposes of reliability. For the Innovation and Internal business dimension, we must delete the R&D expenses to revenues measure to obtain an alpha coefficient of 0.55. For the Learning and Growth dimension, two iterations are necessary; first, we must delete the revenue per employee measure to obtain an alpha of 0.43, and then delete the training expenses to revenue measure to obtain an alpha of 0.58. Theoretical arguments support this iterative process – Cronbach’s alpha coefficient computation, deletion of items, and re-computation until an acceptable alpha coefficient is achieved for each dimension (see, Churchill, 1979, p. 69). Factor analysis can then be used to validate whether the structure of four BSC dimensions can be observed empirically, which would allow us to examine the factorial validity of the BSC. Factorial validity refers to the degree to which an empirical factor analysis is coherent with a priori theoretical expectations (Kerlinger, 1986). We therefore performed a principal components analysis – Varimax rotation – with the reliable measures (measures in bold in Table 4). Table 5 presents results that confirm the four BSC dimensions, as proposed by Kaplan and Norton, results are also consistent with Hoque and James’ (2000) study. Only the working capital ratio measure does not clearly fit the four BSC dimensions, with a loading of 0.268 for the Financial Table 5. Factor Analysis of the Selected Measures. Measure

Factor Loadings F1 financial

Return on asset Net profit margin Working capital ratio Marketing expenses to revenues Revenue growth Number of new products Number of product offers Employee absenteeism rate Employee turnover rate Eigenvalues

0.914 0.917 0.268

F2 learning and growth

F3 innovation and internal

F4 customer

0.275

0.349

0.487 0.505 0.850

0.127

1.948

0.144 0.777 0.842 0.836 0.823 1.456

1.341

0.229 0.240 1.088

Note: Extraction Method, principal component analysis; Rotation Method, Varimax with Kaiser Normalization; Variance explained with the four factors, 72.9%; Absolute values less than 0.10 have been suppressed.

Examining the Construct Validity of the Balanced Scorecard

239

dimension and a loading of 0.275 for the Learning and Growth dimension. The previous reliability analysis indicated that Cronbach’s alpha coefficient for the Financial dimension could be improved from 0.64 to 0.82 by deleting the working capital ratio measure; this measure is therefore deleted from further analyses. All in all, the above results combined (Tables 4 and 5) demonstrate reliability of eight measures, and these measures are associated with their related dimensions. Results also support the specific BSC structure of four dimensions in showing factorial validity (that is, coherence between theoretical expectations and empirical results).

5. EXAMINING CONVERGENT AND DISCRIMINANT VALIDITY OF MEASURES USING A MULTITRAIT– MULTIMETHOD (MTMM)MATRIX One principle of scientific research is that a construct should be measurable by at least two different methods; otherwise, the researcher has no way of knowing whether the measure obtained is anything more than an artifact of the assessment process (Churchill, 1979). We therefore refer to the concepts of convergent and discriminant validity. Convergent validity represents the extent to which measures of a construct agree with one another, that is, it is an assessment of the capacity of different measurement methods to produce highly correlated measures for a given trait (Campbell & Fiske, 1959; Brownell, 1995). Discriminant validity represents the extent to which measures of different constructs are distinct, that is, it is an assessment of two different traits measured in ways that produce low correlations, whatever the methods used (Kwok & Sharp, 1998). The MTMM matrix allows to assess both convergent and discriminant validity. For the measures examined through content validity, face validity, reliability, and factorial validity, data were collected from different sources and different modes of assessment. Detailed explanations on data collection are provided below (see Table 6).

5.1. Subjective Assessments For the subjective measures of the Financial and Customer dimensions (Method 1), management respondents were asked to classify their business

240

EMILIO BOULIANNE

Table 6. Dimension

Data Sources and Modes of Assessment for the MTMM Matrix. Measures

Financial

Return on asset Net profit margin

Customer

Marketing expense to revenue Revenue growth

Innovation and Number of new Internal products Number of product offers

Learning and Growth

Employee absenteeism rate Employee turnover rate

Method 1: Subjective Respondent: Manager

Method 2: Objective Respondent: Comptroller

Questionnaire Worksheet (appendix, (appendix, Panel A): Panel C): Objective data Subjective assessment obtained by comptroller, obtained by manager and computation of relative position, by quartile, with a financial database Questionnaire Same as above, but relative (appendix, Panel A positioning only for for revenue growth) revenue growth (appendix, Panel C): Questionnaire Worksheet: Objective data/ (appendix, Panel B): measure obtained by Subjective assessment comptroller obtained by manager on the degree of firm innovation Questionnaire: Worksheet: Objective data Subjective assessment obtained by comptroller obtained by manager on employee productivity (revenue per employee)

unit’s performance compared with its peers, using a 4-point scale, for the ROA, net profit margin, and revenue growth measures. We thereby obtained a perceptual assessment ranging from 4 (best) to 1 (worst) for each measure. (See appendix, Panel A.) For the Innovation and Internal business dimension, we used the strategy typology of Miles and Snow (1978, 1994) to capture the degree of a firm’s innovation. Management respondents were presented with a short description of three different firms and were asked to choose which description most closely fit their business unit compared to other firms in the industry. The innovation component was underlined to emphasize the degree of innovation. The questionnaire did not include the terms defender (less innovative, score 1), analyzer (mid position, score 2), or prospector (highly innovative, score 3), and the instructions indicated that none of the strategic-types was necessarily good or bad. Respondents classified the 90 business units as follows: 27 defenders (30%), 39 analyzers (43%), and 24

Examining the Construct Validity of the Balanced Scorecard

241

prospectors (27%). This higher percentage of analyzers is coherent with previous research (see Collins, Holzmann, & Mendoza, 1997; Sabherwal & Chan, 2001). Abernethy and Lillis (2001) used a similar approach to assess the innovation construct. (See appendix, Panel B.) For the Learning and Growth dimension, management respondents were asked to classify their business unit’s performance compared with peers using a 7-point scale (1 meaning a high rate, 7 a low rate) for the employee absenteeism rate and employee turnover rate measures.

5.2. Objective Assessments For the objective measures of the Financial and Customer dimensions (Method 2), the position in accounting/finance respondents were asked to provide financial and non-financial data/measures. In addition to computing the ROA, net profit margin, revenue growth and marketing expenses to revenues measures, respondents of each unit were asked to provide annual revenues for size classification and the Standard Industrial Classification (SIC) code for industry classification. Measures were then estimated relative to size and industry. The calculated measures were compared to a database in which the data are organized by quartile.9 We classified the relative positions of the first three indicators above for each business unit and obtained an assessment ranging from 4 (best) to 1 (worst). For example, a business unit that reported $25 million of revenue, an SIC code of 1450, and a ROA of 5% would be assigned a score of 3 if the database showed that, for this specific size and industry, a ROA ranging from 0 to 10% corresponded to quartile 3. (See appendix, Panel C.) For the Innovation and Internal Business dimension, the position in accounting/finance respondent of each business unit provided the number of new products and the number of product offers. Finally, for the Learning and Growth dimension, which reflects employee productivity, the position in accounting/finance respondent provided the objective data to calculate the revenue per employee measure. Subjective and objective assessments are different approaches to evaluation. Measurement based on secondary data allows replication, although it is not always available in the format desired or it may suffer from inaccuracy (Kern & Morris, 1994; Courtenay & Keller, 1994). Measurement based on primary data allows collecting data in the required format; however, bias may be introduced. Accordingly, both strengths and weaknesses of methods and data sources are therefore mitigated when using

242

EMILIO BOULIANNE

a multimethod approach. In addition, the use of two different respondents per business unit, as in the present study, reduces the threat of functional or response bias (Huber & Power, 1985). Our intention with the MTMM matrix was to respect Campbell and Fiske’s recommendations to use independent methods, respondents, and sources of data.

5.3. Multitrait–Multimethod Matrix Results Table 7 presents the MTMM matrix results. Convergent validity is the first of the four criteria to examine in interpreting the matrix. When the diagonal coefficients in the lower left quadrant of the matrix are positive and significant (values in bold), the convergent validity is considered to be satisfied. In other words, convergent validity of a measure is present by the extent to which it correlates with other methods designed to measure the same trait of a construct (Churchill, 1979). In Table 7, six of the eight diagonal coefficients are positive and significant (0.50, po 0.01; 0.38, po 0.01; 0.39, po 0.01; 0.37, po 0.01; 0.30, po 0.01; and 0.21, po 0.05); thus convergent validity is, in general, supported. The exceptions are the marketing expenses to revenues/revenue growth relationship and employee turnover rate/revenue per employee relationship, which are both positive but not significant. The three other criteria examine discriminant validity. The second criterion requires that each diagonal coefficient be higher than the different trait/different method correlations – those entries that are in the same column and row as the diagonal coefficient (entries in triangles with dotted lines). In other words, the correlation between two different measures of the same variable should be higher than the correlations between that variable and other variables that have neither trait nor method in common (Campbell & Fiske, 1959). This requirement is fulfilled in 79/88 cases; the second criterion is thus satisfied at the level of 90%. The third criterion requires that each diagonal coefficient be higher than the different trait/same method correlations (entries in triangles with full lines). That is, the correlation within a trait measured by different methods must be higher than the correlations between traits that have the method in common (Churchill, 1979). The results show that this requirement is fulfilled in 71/88 cases; the third criterion is thus satisfied at the level of 80%. These are interesting results since this criterion is not always satisfied once the ‘‘method effect’’ occurs (Althauser & Heberlin, 1970; Becker & Vance, 1993). The last criterion requires that the pattern of correlations of the four triangles in the matrix be identical. A Friedman test was conducted to

Table 7.

Examination of the Convergent and Discriminant Validity of BSC Measures with the MTMM Matrix.

Examining the Construct Validity of the Balanced Scorecard

245

investigate the triangles’ similitude. w2 statistic is 14.8 (df ¼ 3), significant at po 0.01; we can thereby infer that the pattern of correlations is not identical (that is, there are significant differences between the triangles structures, so criterion 4 is not satisfied). As reported in several previous studies, criterion 4 is seldom satisfied. For example, Venkatraman and Ramanujam (1987) found similar results in which the first three criteria were satisfied but not the last regarding the pattern of correlations of triangles. To rule on convergent and discriminant validity with the MTMM matrix, the four criteria need to be satisfied overall (Peter, 1981) and results show that the criteria are in general satisfied. Our objective with the multimethod approach was to examine the convergent and discriminant validity of measures that had earlier passed through validity steps.

6. DISCUSSION, LIMITATIONS, AND CONCLUSION The chapter’s objective was to empirically examine the construct validity of selected measures and the four dimensions to determine whether the BSC is indeed a valid performance-measurement model. Recall that Chenhall (2003) points out the importance of developing reliable and valid measures to enhance consistency between studies on the BSC. With reference to content validity, face validity, reliability, factorial validity, and convergent and discriminant validity, results indicate that the four BSC dimensions and related measures represent a valid performance model. In using a multimethod–multiple respondent approach to reduce the threat of response bias, the present study also offers a methodological contribution. Churchill (1979) states that when using validated measures, if a construct is more than a measurement artifact, results should be reproducible with a new sample. Accordingly, future research projects should examine the validity of the BSC measures analyzed here with other firms and in different settings. This would enhance the robustness and reliability of studies on the BSC, and offer a stronger base for theory development. Accounting research should be conducted with reference to relevant theories but theory is not yet developed for performance evaluation; the present study is therefore one initiative toward a theory-building perspective. The present study has limitations. First, the measures examined are those considered by managers of the field study to be relevant for performance evaluation and common to different business units and were selected on that basis. Second, a larger sample for the survey research would increase confidence in the results but the difficulties of obtaining information at the

246

EMILIO BOULIANNE

business-unit level limited the response rate. Third, although we referred to and applied rigorous validity concepts, these concepts themselves have limits (e.g., the reliability level is an estimation (Peter, 1979) and to rule on convergent and discriminant validity, the four criteria often are not all met empirically but assessed overall (Kwok & Sharp, 1998; Peter, 1981)). Finally, measures developed for planning, compensation, or performance evaluation, are most likely not appropriate for these three contexts; the present results therefore apply to performance evaluation only. With the advent of integrated information systems such as the BSC, the function of the accountant as an information producer has become more challenging. To be effective in the design of BSC for performance evaluation, accountants need pertinent and reliable measures within a valid framework. Rigorous research on the BSC is only beginning to emerge and the present study intends to be among them.

NOTES 1. ‘‘Several years ago, we introduced the Balanced Scorecard. At that time, the Balanced Scorecard was about performance measurement, not about strategy’’ (Kaplan & Norton, 2001, p. 3). 2. With face validity, we obtained the point of view of managers in practice. Managers have knowledge and experience with performance measurement systems, and are therefore competent to judge whether the measures suggested reflect reality as they know it (Kwok & Sharp, 1998, p. 142). 3. Field study research is defined here as the examination of a real-world phenomenon through direct contact with managers (Yin, 2002; Ahrens & Dent, 1998). The objective is to obtain a rich and real understanding of a relevant business world phenomenon (Merchant & Van der Stede, 2006). Interview, questionnaire, and examination of documents are used to obtain evidence in field study research (McKinnon, 1988). 4. The Certified General Accountants of Canada (CGA-Canada) is a professional accounting association representing 55,000 members and students. We worked with CGA-Quebec, an affiliate of CGA-Canada. We had developed a relationship with this organization in previous research projects. 5. For example, Chenhall states: ‘‘In the first instance it was planned to collect objective data on divisional profitability such as growth in sales and return on assets. However, this was unsuccessful as many of the chief managers were reluctant to permit the disclosure of such data. As an alternative, perceived measures of performance were used’’ (1997, p. 196). 6. Two academics and an adviser in linguistics reviewed the questionnaires. 7. The managers and units that participated in the field study have not been contacted for the survey research.

Examining the Construct Validity of the Balanced Scorecard

247

8. The business units were in pulp and paper, textile, transformation, construction, industrial products, food products, retail, wholesale, leasing, and dealers. The percentage per industry is similar to the 500 firms contacted. 9. The database is Performance Indicators for Canadian Business from Statistics Canada. In this database, indicators are designed to serve as performance benchmarks against which firms can be compared.

ACKNOWLEDGMENTS I acknowledge the comments of participants at the 4th Conference on Performance Measurement and Management Control (EIASM) held in September 26–28, 2007 in Nice, France. Thanks also to the Fonds Que´be´cois de Recherche sur la Socie´te´ et la Culture (FQRSC), CGACanada, and CGA-Que´bec for their important financial support.

REFERENCES Abernethy, M. A., & Lillis, A. M. (2001). Interdependencies in organization design: A test in hospitals. Journal of Management Accounting Research, 13, 107–129. Ahrens, R., & Dent, J. F. (1998). Accounting and organizations: Realizing the richness of field research. Journal of Management Accounting Research, 10, 1–39. Althauser, R. P., & Heberlin, T. A. (1970). Validity and the multitrait-multimethod matrix. In: E. F. Borgatta & G. W. Bohrnstedt (Eds), Sociological methodology (pp. 151–169). San Francisco: Jossey-Bass Inc. Assael, I., & Keon, J. (1982). Non sampling vs. sampling errors in survey research. Journal of Marketing, 46(2), 114–123. Banker, R. D., Potter, G., & Srinivasan, D. (2000). An empirical investigation of an incentive plan that includes nonfinancial performance measures. The Accounting Review, 75, 65–92. Becker, T. E., & Vance, R. J. (1993). Construct validity of three types of organizational citizenship behavior: An illustration of the direct product model with refinements. Journal of Management, 19(3), 663–682. Blum, M. L., & Naylor, J. C. (1968). Industrial psychology. New York, NY: Harper and Row. Brownell, P. (1995). Research Methods in Management Accounting, Coopers & Lybrand Accounting Research Methodology, Monograph no. 2. Australia and New Zealand: Coopers & Lybrand. Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait multimethod matrix. Psychological Bulletin, 56, 81–105. Campbell, J. P. (1990). Modeling the performance prediction problem in industrial and organizational psychology. In: M. D. Dunnette & L. M. Hough (Eds), Handbook of industrial and organizational psychology (2nd ed., Vol. 1, pp. 687–732). Palo Alto, CA: Consulting Psychologists Press.

248

EMILIO BOULIANNE

Chan, Y. E., Huff, S. L., Barcaly, D. W., & Copeland, D. G. (1997). Business strategic orientation, information systems strategic orientation, and strategic alignment. Information Systems Research, 8(2), 125–150. Chenhall, R. H. (1997). Reliance on manufacturing performance measures, total quality management and organizational performance. Management Accounting Research, 8, 187–206. Chenhall, R. H. (2003). Management control systems design within its organizational context: Findings from contingency-based research and directions for the future. Accounting, Organization, and Society, 28, 127–168. Chenhall, R. H., & Langfield-Smith, K. (1998). The relationship between strategic priorities, management techniques and management accounting: An empirical investigation using a system approach. Accounting, Organization and Society, 23(3), 243–264. Churchill, G. A. (1979). A paradigm for developing better measures of marketing constructs. Journal of Marketing Research, 16(1), 64–73. Collins, F., Holzmann, O., & Mendoza, R. (1997). Strategy, budgeting, and crisis in Latin America. Accounting, Organization and Society, 22(7), 669–689. Cook, T., & Campbell, D. (1979). Quasi-experimentation: Design and analysis issues for field settings. Boston, MA: Houghton Mifflin Company. Courtenay, S. M., & Keller, S. B. (1994). Errors in database revisited: An examination of the CRSP shares-outstanding data. The Accounting Review, 69(1), 285–291. Dilla, W., & Steinbart, P. (2005). Relative weighting of common and unique balanced scorecard measures by knowledge decision makers. Behavioral Research in Accounting, 17, 43–53. Financial Performance Indicators for Canadian Business. (2007). Statistics Canada publisher, industrial organization and finance division. Ottawa: Canada. Fitzgerald, L., Johnston, R., Brignall, S., Silvestro, R., & Voss, C. (1991). Performance measurement in service businesses. London, UK: CIMA. Heneman, R. L. (1986). The relationship between supervisory ratings and results-oriented measures of performance: A meta-analysis. Personnel Psychology, 39(4), 811–826. Hoque, Z., & James, W. (2000). Linking balanced scorecard measures to size and market dactors: Impact on organizational performance. Journal of Management Accounting Research, 12, 1–17. Huber, G., & Power, D. (1985). Retrospective reports of strategic-level managers. Strategic Management Journal, 6, 171–180. Ittner, C. D., & Larcker, D. F. (2001). Assessing empirical research in managerial accounting: A value based management perspective. Journal of Accounting and Economics, 32, 349–410. Ittner, C. D., Larcker, D. F., & Meyer, M. W. (1997). Performance, compensation, and the balanced scorecard. Working Paper. The Wharton School, University of Pennsylvania. Ittner, C. D., Larcker, D. F., & Meyer, M. W. (2003). Subjectivity and the weighting of performance measures: Evidence from a balanced scorecard. The Accounting Review, 78(3), 725–758. Kaplan, R. S., & Atkinson, A. A. (1998). Advanced management accounting (3rd ed.). New Jersey: Prentice Hall. Kaplan, R. S., & Norton, D. P. (1992). The balanced scorecard – Measures that drive performance. Harvard Business Review (Jan–Feb), 71–79. Kaplan, R. S., & Norton, D. P. (1996). The balanced scorecard: Translating strategy into action. Boston, MA: Harvard Business School Press.

Examining the Construct Validity of the Balanced Scorecard

249

Kaplan, R. S., & Norton, D. P. (2001). The strategy-focused organization: How balanced scorecard companies thrive in the new business environment. Boston, MA: Harvard Business School Press. Kaplan, R. S., & Norton, D. P. (2004). Strategy maps: Converting intangible assets into tangible outcomes. Boston, MA: Harvard Business School Press. Kerlinger, F. N. (1986). Foundations of behavioral research. New York, NY: Holt, Rinehart and Wilson. Kern, B. B., & Morris, M. H. (1994). Differences in the compustat and expanded value line databases and the potential impact on empirical research. The Accounting Review, 69(1), 274–284. Kwok, W. C., & Sharp, D. J. (1998). A review of construct measurement issues in behavioral accounting research. Journal of Accounting Literature, 17, 137–174. Lingle, J. H., & Schiemann, W. A. (1996). From balanced scorecard to strategic gauges: Is measurement worth it? Management Review, 85(3), 56–61. Lipe, M. G., & Salterio, S. E. (2000). The balanced scorecard: Judgmental effects of common and unique performance measures. The Accounting Review, 75(3), 283–298. Lipe, M. G., & Salterio, S. E. (2002). A note on the judgmental effects of the balanced scorecard’s information organization. Accounting, Organization and Society, 27, 531–540. Lynch, R. L., & Cross, K. F. (1991). Measure up! Yardsticks for continuous improvements. Cambridge, MA: Blackwell. Malina, M. A., & Selto, F. H. (2001). Communicating and controlling strategy: An empirical study of the effectiveness of the balanced scorecard. Journal of Management Accounting Research, 13, 47–90. McKinnon, J. (1988). Reliability and validity in field research: Some strategies and tactics. Accounting, Auditing, and Accountability Journal, 1(1), 34–54. Merchant, K. A., & Van der Stede, W. A. (2006). Field-based research in accounting: Accomplishment and prospects. Behavioral Research in Accounting, 18, 117–134. Miles, R. E., & Snow, C. C. (1978). Organizational strategy, structure and process. New York: McGraw Hill. Miles, R. E., & Snow, C. C. (1994). Fit, failure, and the hall of fame. New York: The Free Press. Nunnally, J. (1967). Psychometric methods. New York: McGraw-Hill Book. Peter, J. P. (1979). Reliability: A review of psychometric basics and recent marketing practices. Journal of Marketing Research, 16(1), 6–17. Peter, P. J. (1981). Construct validity: A review of basics issues and marketing practices. Journal of Marketing Research, 18(2), 133–145. Reck, J. L. (2001). The usefulness of financial and nonfinancial performance information in resource allocation decisions. Journal of Accounting and Public Policy, 20, 45–71. Sabherwal, R., & Chan, Y. E. (2001). Alignment between business and IS strategies: A study of prospectors, analyzers, and defenders. Information Systems Research, 12(1), 11–33. Ullrich, M. J., & Tuttle, B. M. (2004). The effects of comprehensive information reporting systems and economic incentives on managers’ time-planning decisions. Behavioral Research in Accounting, 16, 89–105. Venkatraman, N., & Ramanujam, V. (1987). Measurement of business economic performance: An examination of method convergence. Journal of Management, 13(1), 109–122. Yin, R. K. (2002). Case study research, design, and methods (3rd ed.). Newbury Park, CA: Sage.

250

EMILIO BOULIANNE

APPENDIX Panel A Unit Performance; Subjective assessment; Respondent: Manager. When you compare your business unit with peers (e.g., units of the same size, same industry), indicate in which quartile it types:  Q4 corresponds to the first best at the 25th percentile,  Q3 corresponds at the 50th percentile, e.g., there are 50% of firms above,  Q2 corresponds at the 75th percentile, e.g., there are 75% of firms above, and  Q1 corresponds to the worst set of firms. Q4

Q3

Q2

Q1

Return on assets Net profit margin Revenue growth Working capital

Panel B Degree of firm’s innovation; Subjective assessment; Respondent: Manager. By referring to the following three descriptions, identify which unit more closely corresponds to your business unit presently, when compared with others in your industry. (Note that none of these descriptions is either good or bad.) Business 1 Business 1 maintains a secure niche in a relatively stable product or service area. In general, Business 1 is not at the forefront of developments in regard to new products and services, or market developments. Business 1 tends to ignore changes that have no direct influence on current areas of operation, and concentrates on doing the best job possible in its area. Business 2 Business 2 makes constant changes in its group of products/services. Business 2 acts as a pioneer and innovator in new markets even if some of these

251

Examining the Construct Validity of the Balanced Scorecard

efforts prove not to be profitable. Business 2 responds very rapidly to early signals of new market needs or business opportunities. Business 3 Business 3 brings changes in its group of products/services. Business 3 attempts to be leader for some products/services based on concepts introduced by innovative firms. Business 3 responds quite rapidly to early signals of new market needs or business opportunities. My business unit is best described by:

1

2

3

Panel C Unit Performance; Objective assessment; Respondent: Comptroller. Based on the definitions provided, calculate the following financial indicators: Return on asset:

For your unit, the return on asset is:

net profit þ interest expense total assets

___________________

Net profit margin:

For your unit, the net profit margin is:

net profit total revenue

___________________

Revenue growth:

For your unit, the revenue growth is:

sales current year ðlessÞ sales previous year sales previous year

____________________

Working capital:

For your unit, the working capital is:

Current Assets Current Liabilities

____________________

Marketing expenses to revenues:

For your unit, marketing expenses to revenues is:

marketing expenses total revenue

___________________

This page intentionally left blank

ASSESSING PRIVATIZED AIRPORT PERFORMANCE FROM STAKEHOLDER VIEWPOINTS: A STUDY OF SYDNEY AIRPORT Dorothea Zakrzewski ABSTRACT Airport privatization has led to the emergence of new industry players enabling them, it is claimed, to raise additional capital, improve efficiency, reduce costs, generate new revenue streams and engage in new commercial airport investments in the market economy. A question remains about the impact and benefit of airport privatization reforms on stakeholder groups. It is also debatable how overall privatized airport performance can be formally assessed. This chapter reports the perceptions of key stakeholder groups on the privatization of Sydney Airport. Preliminary attributes and indicators of the airport performance stakeholder model, with an emphasis on assessing privatized airport performance from stakeholders’ perspectives are considered. A qualitative paradigm was applied to this field research. Leximancer data mining software was used for the thematic analysis of the interviews conducted with key Sydney Airport stakeholders.

Performance Measurement and Management Control: Measuring and Rewarding Performance Studies in Managerial and Financial Accounting, Volume 18, 253–272 Copyright r 2008 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-3512/doi:10.1016/S1479-3512(08)18011-X

253

254

DOROTHEA ZAKRZEWSKI

1. INTRODUCTION In accordance with Beesley (1992) privatization, the imperative for economic and institutional reform of public enterprises is considered as the optimal transformation to deal with the constraint of governments’ budgets. These reforms have been embraced by all nations, developed as well as the developing countries, with the underpinning assumption that the whole economy benefits from the efficiency changes generated by privatized organizations. Parker (2003a, 2003b) and others (Parker, 2003a; Vickers & Yarrow, 1995; Wiltshire, 1987) claim that efficiency gains through the transfer of ownership and management of state-owned enterprises into private hands is to be the ultimate objective of privatization reforms. Private enterprises driven by the profitability imperative are expected to operate more efficiently. Under the privatization philosophy Australian airports have been gradually sold off to conform to the pressures of deregulation reforms. The globalization and the deregulation of the transport industry have prompted a commercial approach to airport operations. Commercialoriented approaches to airports have also been a driver to seek private sector involvement in airport operations. Australia’s first airport privatization occurred in 1996. The question remains about the success and consequences of these changes. This chapter is part of the still ongoing PhD study examining the impacts of airport privatization at Sydney Airport from stakeholder perspectives. This study accepts the existence of multiple parties having a legitimate interest or stake in their business. Emphasized by Donaldson (2002) and Mitchell, Agle, and Wood (1997) stakeholders interact and give meaning to the corporation. It is argued that in the air transport industry most stakeholder interests are interdependent. This chapter proposes an integrated performance measurement model to link stakeholder theory with performance measurement and is founded on Kaplan and Norton’s (1996, 2001) Balanced Scorecard Framework. The focal point of this chapter is the Sydney Airport privatization. It is proposed that the outcome of airport privatization reforms should be assessed by taking stakeholder interests and views into account. The research uses a qualitative paradigm combining interviews and archival content analysis to analyze the impact of Sydney Airport privatization on its stakeholders. The chapter is organized as follows: the background to the Australian airport privatization is provided, followed by a discussion of airport stakeholders and importance for performance measurement. The research design and framework is stated. In conclusion, the findings of what

Assessing Privatized Airport Performance

255

stakeholder view is on airport privatization is discussed, followed by the preliminary findings of the attributes of the stakeholder-based airport performance assessment tool.

2. BACKGROUND ON AIRPORT PRIVATIZATION AND STAKEHOLDER SIGNIFICANCE 2.1. The Australian Experience Up to the 1980s the Australian government has funded most airport operations. Airport management was controlled by and was held accountable to government, the major stakeholder. Airport management, with its capital-intensive operations, was dependent upon the public sector to fund airport growth. As argued by Sharp (1996), Baird (1996) and McGhee (1996), the sum of investment needed for Australian airports to achieve international standards and to maintain their competitiveness challenged government and their capacity to run airports efficiently. More than 80 countries have introduced some form of privatization or commercialization of airports these days and it was argued (Enright & Ng, 2001; Humphreys & Francis, 2002a) that by the year 2000 more than 100 airports in more than 30 countries were privatized. Key arguments supporting airport privatization in the literature are as follows: (a) the governments’ disposal of assets, lowers the budget deficit and promotes investment into other sectors to foster economic growth and development; (b) the airlines and the general public profit from a wide range of commercial services; (c) the airport operators generate more return on investment linked to long-term plans and expansion strategies and (d) the overall airport business is to run as a profitable business enticing airport operators to rethink the concept of self funding and accountability. As indicated by Murphy (1996) by 1997 the Australian government had privatized all major city and large regional airports of the Federal Airport Corporation (FAC).1 Smaller regional and local airports owned by the Australian government were transferred to the ownership and responsibility of local authorities, usually local government councils, in the mid 1980s. Although the current political philosophy in Australia espouses the superiority of free markets and privatization, the acclaimed efficiency gains of Australian airport deregulation and privatization on all airport users remain debatable.

256

DOROTHEA ZAKRZEWSKI

Little evidence however exists about whether the underpinning reasoning and existing claims of privatization such as increased efficiency and profitability are correct. Recent research in the area of airport performance measurement and strategic performance management (Graham, 2003; Humphreys & Francis, 2000, 2002b; Walsh, Lok, & Jones, 2005) imply that corporate performance measurement needs to take into account external, internal as well as operational factors. Privatization reforms enticed state-owned entities to become more efficient, profit oriented and accountable to various stakeholders. Therefore, in line with Brenner (1995) it is stakeholders who are impacted by the organizations (performance). It is suggested here that the impact of privatization should be validated by analyzing airport performance through the eyes of stakeholders. It is being proposed that through interpretation of Jensen (2002) ideology of the Balanced Scorecard as a managerial equivalent to stakeholder theory, the underpinning assumptions of stakeholder theory and performance measurement could be interlinked to test the efficiency claims of privatization. This chapter considers the preliminary stages of such model development.

2.2. Stakeholder Identification and Performance Measurement Stakeholders are vital to overall success of business operations. The modern corporations play an important role in today’s society. Corporations support the nations economic growth and in a sense are clusters of individuals working together in the transformation process of goods and services (Andrews, 1971; Ansoff, 1965; Bantel & Osborn, 1995; Porter, 1990; Watson, 1993). This is particularly true for complex entities such as airports where different stakeholders rely on one another for operating success. The concept of ‘stakeholders’ is embedded in the management and social science literature (Donaldson, 2002). However, there exist various identification rationales based on normative, descriptive, instrumental or prescriptive theories as to ‘who or what’ are stakeholders (Donaldson & Preston, 1995; Friedman & Miles, 2002; Mitchell et al., 1997; Savage, Nix, Whitehead, & Blair, 1991). This chapter does not aim to elaborate on a definition of stakeholder theory nor to offer a rationale of stakeholder identification. It accepts the notion of stakeholders as defined by Freeman (1984, p. 46) ‘being any group or individual who can affect or is affected by the achievements of the organizations objectives’. Furthermore, it is contended that the underpinning assumptions of stakeholder theory and performance measurement can be linked in the

Assessing Privatized Airport Performance

257

development of the model to assess the performance outcomes of privatized airports. Mitchell et al. (1997) has identified key stakeholder groups relevant to privatized airport operations. Mitchell also classified stakeholders based on the idea that stakeholders become salient to managers to the extent that managers perceive them as possessing key constructs such as (a) power to influence a firm (Dahl, 1957; Pfeffer, 1981; Weber, 1947), (b) legitimacy with the firm (Weber, 1947) and (c) urgency towards claims on the firm (Hill & Jones, 1992; Wartick & Mahon, 1994). For the purpose of the undertaken study, the above-stated attributes have assisted in distinguishing and categorizing airport stakeholders (Table 1) used in the analysis. Privatized airports are more efficient and profitable in their operations, it is argued, because of the change in ownership rights and intensified competition. Current research indicates (Beesley, 1992; Emmons, 2000; Parker, 2003b) that the impact of privatization on stakeholders and overall privatized entity performance is, however, difficult to assess, as the political and economic effects of ownership, competition, regulation and technological changes need to be separated out from the entities’ present position. Writers have noted that organizational performance measurement requires a focus on appropriate performance indices of inputs, processes, outputs, outcomes as well as the environment and institutional strategy (Bantel & Osborn, 1995; Kaplan & Norton, 2001; Kotter & Heskett, 1996; Manzoni, 2004). With respect to performance measurement this chapter accepts that  performance measurement is critical for management decision-making of firms operating in a competitive environment,  performance measurement does have pitfalls and challenges,  to ensure competitive advantage and the objective of wealth maximization, performance measurement needs to relate to stakeholders,  performance measurement is not easy as the factors that need to be considered for measurement purposes differ amongst entities within the same industry and  there is no ‘one-size fits all measurement concept’ that can be applied to organizations. Privatization, in particular airport privatization, deals with commercial, political and social issues. Privatized airports have become more commercially oriented; a question remains of how to measure privatized airport performance taking the needs of stakeholders into account. This chapter proposes the stakeholder-based airport performance assessment model that includes key performance attributes identified by stakeholders during the undertaken field research.

258

DOROTHEA ZAKRZEWSKI

Table 1.

Airport Stakeholder Identification.

Stakeholder Type

Attributes and Definition (Mitchell et al., 1997)

Dormant stakeholder

Possesses power to impose their will on a firm – does not have a legitimate relationship or urgent claim. These stakeholders have little or no interaction with the firm; however, they can become significant if they acquire legitimacy and urgency Possesses legitimacy but have no power to influence the firm and no urgent claims. Important stakeholders for corporate social responsibility and performance. Managers have no need to engage in active relationship with such stakeholders Possesses urgency as key attribute but has neither power nor legitimacy. These are stakeholders that appear bothersome for management until they acquire some power and legitimacy to make a claim Possesses power and legitimacy, has influence in firm. They expect and receive a lot of managers attention and are those stakeholders that are claimed by scholars to be the ‘only’ stakeholders Possesses urgency and power, but lacks legitimacy. This group is suggested to use coercion to influence stakeholder/ management relationship and individuals. Society refuses to acknowledge the existence of this stakeholder group Possesses urgent and legitimate claims; however, there is a lack of power. These stakeholders depend on other dominant stakeholders who have common interest Possesses all three key attributes most important stakeholder for any organization to satisfy

Discretionary stakeholder

Demanding stakeholder

Dominant stakeholder

Dangerous stakeholder

Dependent stakeholder

Definitive stakeholder

Airport Stakeholders

Regulators, government, employees

Community, council

Environmental groups, unions

Investors/shareholders, consortia members

N.A.

Others: infrastructure asset providers, concessionaries Airlines (domestic and international), passengers, airport operator

Source: Modified from Mitchell et al. (1997) and used to identify airport stakeholders by the author.

259

Assessing Privatized Airport Performance

3. RESEARCH DESIGN Interviews were conducted to assess the impact of privatization on Sydney Airport performance and its stakeholders. Open-ended questions were used to get an understanding of the interviewee’s viewpoint on the impact of privatization reforms. These questions asked for the participants’ interpretation of (a) airport privatization reforms and viewpoints on how privatized airport performance could be assessed when Sydney Airport is portrayed as a subset of key stakeholder operations and (b) whether an integrated performance measurement framework would be a useful tool for privatized airport performance assessment by industry standards. The participant sample was selected based on their role, corporate and stakeholder interest and involvement in the Sydney Airport privatization as well as being a key stakeholder of the airport. Avoiding data overload, quality of interviews rather than quantity mattered. The tabular representation (Table 2) indicates the sampling method chosen for each key stakeholder group and the position held by most interviewees. Leximancer data mining software assisted with the thematic analysis. The analysis that follows was done in two parts:  Firstly, a conceptual analysis was undertaken with the aim to discover the presence, frequency, strength and definition of key concepts from all the interviews.  Secondly, relational analysis assessed the research propositions behind the objectives of the privatization reforms by ‘drilling down’ into the data to find the meaning in not only the relatedness of the concepts but also in the context of airport privatization. Table 2. Sampling Method Used for Each Stakeholder Group. Airport Stakeholders

Sampling Method

Interviewees

Regulators and government

Convenience

Community and council Environmental groups Investors/shareholders, consortia members Other airport users: infrastructure asset providers, concessionaries Airlines (domestic and international), passengers, airport operator

Snow-ball Snow-ball Convenience Convenience

Executive managers, deputy secretaries Mayors, councilors Leader of environmental groups Senior management Senior management

Convenience

Senior management

260

DOROTHEA ZAKRZEWSKI

Leximancer was used to display (a) the words comprising the vocabulary of the most frequent concepts used for subsequent analyses, (b) the connectivity of concepts in similar contexts and (c) the relation between individual concepts in form of a visual display of a diagram referred to as ‘concept map’. This software provides a means of both quantifying and displaying the conceptual structure of the interviews and enables the researcher to explore interesting conceptual features. Leximancer extracted the most important concepts discussed in the textual documents. It identifies concepts based on ‘seed’ words as a starting point. The software then learns a concept by identifying the words that occur around these seed words and builds the concept based on seed and associated words (Leximancer, 2007). This was measured by examining how often two concepts are discussed within the same passage of text, which in this analysis was identified as two sentences within the same paragraph that was proportionate to the length of the responses. All analysis conducted by Leximancer was on solely the responses from the interviewees. The interviewer text prompting the question has been eliminated from the analysis by using the ‘kill class’ function that is available on Leximancer. Such approach allows for a more focused examination and exploration of only the interviewees responses related to questions on airport privatization and airport performance assessment. The limitations of this study are in line with general limitations as applied to case research and using qualitative methods such as interviews.

4. SYDNEY AIRPORT: AT A GLANCE Since its privatization, Sydney Airport has received multiple awards. In terms of capital expenditure $ 1.2 billion has been poured into Sydney Airport between 1994 and 2002 including significant investment made for the Sydney Olympic Games; and a further $ 2.5 billion is planned between 2002 and 2022 for further expansion (Dodson, 2005; Meacham, 2005). It is asserted that Sydney Airport is a major engine of growth (Mather, 2004). Sydney (Kingsford Smith) Airport is argued to be the largest and most important domestic as well as the busiest international airport in Australia. Sydney Airport has been successful in its operations providing double-digit growth earnings since its acquisition. It is said that its increased profitability and efficiency of operation contributes on average $ 7 billion a year to NSW economy and houses over 500 businesses and government bodies (Mather, 2004). The consortium acquiring the airport consists of Macquarie Airports

261

Assessing Privatized Airport Performance

and Managed Funds (62.79%), Ferrovial Aeropuertos (20.68%), Hochtief Airport (11.57%) and Ontarios Teachers Trust (4.6%). The privatization mode was a trade sale to encourage the airport and the airlines to develop commercial agreements without the need of government involvement. The government retained control on key areas such as noise, security and price control through the implementation of the light-handed regulatory regime.

5. STAKEHOLDERS’ VIEWS ON PRIVATIZATION This section identifies the preliminary themes from the data analysis. The thematic conceptual analysis that was undertaken involved the detection and quantification of the presence of concepts within the interviews. Table 3 illustrates the most common concepts as identified by Leximancer from the 20 semi-structured interviews used in this study, examining the research proposition related to stakeholders’ views on the objectives behind airport privatization. The key concepts of ‘airport’, ‘government’ and ‘privatization’ are the three most frequent-occurring ones. The concept ‘airport’ was mentioned 127 times by the respondents. An interesting aspect of this study is that although Sydney Airport has been privatized and has become commercially oriented, multiple references are made to the post privatization role of the government. Interestingly, most of the other key concepts related to the research proposition on exploring the objectives behind privatization reforms are of financial and economic nature. Table 3. Concept Airport Government Privatization Infrastructure Money Private Capital Community Assets Investment Million

Key Concepts Related to Airport Privatization. Absolute Count

Relative Count (%)

127 67 30 27 19 18 16 16 14 11 11

100 52.7 23.6 21.2 14.9 14.1 12.5 12.5 11 8.6 8.6

262

DOROTHEA ZAKRZEWSKI

This is furthermore illustrated by Regarding the privatization, there are many reasons for privatizations. There is one fundamental reason why governments privatize and that is cash/returns! – Interviewee 1.

Considering the key concepts of ‘money’ and ‘private’ related to ‘airport’ and ‘privatization’ interviewees raised the argument that the private sector at the time of privatization was prepared, or was more likely prepared, to make capital investments to improve the experience for customers or passengers to plan for the future better than the Commonwealth was likely to do. This is illustrated by The qualitative aspect of privatization is to create an environment which promotes A) future investment in the airport to support expected capacity growth and B) also service levels. Obviously on the economic side of things, privatization is a revenue generating event generally speaking. So if you have a government which has a deficit or has you know future funding commitments then the ability to generate, not just the lump sum of revenue that comes from the privatization (from the sale of the asset) but also the potential to maintain an ongoing income stream, potentially form license fees or rates things like that, is quite important. – Interviewee 2.

A detailed examination of the meaning and connectedness of the secondary concepts ‘infrastructure’, ‘money’, ‘capital’, ‘investment’ and ‘assets’ in the context as to why Sydney Airport was privatized, emphasized that the primary aim was to (a) maximize the sale proceeds and (b) secure a purchaser with the financial capability to fulfill the development and expansion plans needed to align the airport operations to global standard levels. It was also pointed out in the interviews that it was all about the passenger experience, as airports are more commercially driven now than in the past. One can argue that the improvement at Sydney Airport has taken place as a result of private ownership – an example of that would be, more retail shops which gives the travelling public a greater range of services; there are more eating facilities which gives the travelling public a better experience, even though it complicates the operations of airlines because everyone’s queued up. – Interviewee 3.

Fig. 1 enhances the above-discussed themes and connectivity of the concepts related to the privatization objectives through stakeholder eyes through a visual display of the conceptual map. As noted on the Leximancer map above there is a strong link between the concepts of ‘privatization’, ‘airport’ and ‘government’ as shown through the thick line and thematic circles. Also the concepts ‘infrastructure’, ‘capital’,

Assessing Privatized Airport Performance

Fig. 1.

263

Conceptual Mapping Pointing Out ‘Privatization Objectives’.

are quite central to the overall theme of airport privatization. This is supportive of existing literature on airport privatization indicating that the government should do and invest in the things that the private sector cannot do, but which governments can do reasonably well, which are social services, defense, immigration and all the things that the private sector really does not have a great deal of interest in. The perceived government philosophy in Australia by the respondents was that the private sector should do what the private sector is best at; and the government should do, what the government is best at – and it was identified that the government was not best at running commercial entities, the private sector is best at running commercial entities. Interestingly, in Australia the airports were moved into the FAC, which was a full government business enterprise with dividend targets and returns

264

DOROTHEA ZAKRZEWSKI

on capital targets prior to privatization. Governments’ working on annual budgets found it very difficult to invest in infrastructure, with a long-term focus. Particularly, during the time of increased competition and globalization governments increasingly around the world in the mid-1980s faced budget difficulties, thus funding infrastructure and being more customer focus was not the priority. Other underpinning reasons for privatizing Australian airports and Sydney Airport in particular was to have the new owners invest into infrastructure a great deal more than the government had available to invest. Sydney Airport was perceived as the most critical single piece of infrastructure for Sydney (NSW), and so obviously Australia’s major airport. It was pointed out that at the time of the privatization, while the aero infrastructure was in good shape, the terminals were not at the most critical piece of national infrastructure sitting in the Sydney basin. Thus, from an airline perspective the objectives were to actually stimulate competition in the market, and also to stimulate investment in the airports: When the government announced privatisation, they did have a plan for phasing in the efforts. Thus, two other objectives behind the privatization reforms were to have the environmental and planning insurance operated under the Commonwealth legislation. The key part of the regulatory framework was to ensure that private owners invest in the infrastructure. – Interviewee 4.

Other privatization objectives that were mentioned by the respondents included land and property development. Particularity with Sydney Airport, the community stakeholders pointed out that the airport land bank was within a very short distance from the central business district, allowing the new owners to develop the land that was undeveloped. Also, it was indicated that the government made it very clear at the time of privatization, that there should be commercial negotiations and transparency between the various stakeholders who are involved in capital planning and pricing decisions. Sydney Airport however has been named in multiple circumstances as a ‘frustrating experience trying to negotiate any contract and new price with’, particularly from the airlines perspectives. A further objective was aviation affordability. By reducing aeronautical charges, the airport operators could make aviation more affordable, which will bring more people to the site, which in turn would help create better opportunities for rental income for retail and other commercial services on the airport sites. However, it was mentioned that it has not happened in the way some stakeholders hoped for due to the commercial and economic orientation of the airport business.

265

Assessing Privatized Airport Performance

6. STAKEHOLDER INTERDEPENDENCE OF INTERESTS The analysis further suggests the existence of strong interdependence amongst particularly three stakeholders: the airport operator, the government and airlines (see Fig. 2). In line with the existing literature there exists stakeholder interdependence at airports. Airports cannot co-exist without the airlines and although Sydney Airport is privatized the concept government has a strong presence. The airport operator relies on the revenue streams coming from the passenger revenue and the charges imposed on the airlines. At the same time however the airport operator needs to fulfill its obligation to its investors whilst implementing and complying with regulatory changes imposed by government. The community is seen as a partner from the airport operators’ perspective, and is involved in debates on airport development in accordance with state government planning policies. The degree of involvement and say is however debatable as indicated in the interviews. The airport operator and the government identified the principal impact of airport privatization to be related to the growth in commercial activities at airports. Catering for the end consumer and satisfying stakeholders through improved company performance led to the acclaimed shareholder wealth maximization at Sydney Airport. It was further identified that shareholder wealth maximization is the key driver behind airport operators’ predominant focus on commercial strategies and expansion. It was further acknowledged by the definitive and dependent stakeholders that revenue streams from the commercial-oriented approach of privatized business operations have outperformed the traditional aeronautical revenue streams at Sydney Airport. Due to this strong interdependence and symbiotic relationship between stakeholders one would expect to see cooperation amongst stakeholder

Government

Airport Operator

Investors

Fig. 2.

Stakeholder Interdependence.

Airlines

266

DOROTHEA ZAKRZEWSKI

demands. Supportive of the Productivity Commissions’ Inquiry on the pricing regime at airports, some airport users have expressed dissatisfaction with increases in airport charges, the ongoing asset valuation debate and overall commercial negotiation procedures. A certain lack of power to negotiate the terms for supply of services in an efficient and effective manner with the airport operator was identified by airport users, despite the acknowledgement of the existing interdependence of operations. It is particularly the smaller airport users who are questioning the airport operators’ dominance. It seems that bargaining power from the dominant and definitive stakeholders (airlines, passengers, airport operator and investors) has increased post privatization, whilst the dormant stakeholder (the government) maintains control of monitoring noise, pollution and price charges (introduction of light-handed regime). The dormant stakeholder also exhibits indirect influence on imposing or reinforcing stricter operating processes in terms of security, safety, but also border control such as quarantine and customs,2 ensuring world standards. In these instances the airport users, investors and operators have neither negotiation power nor say. As indicated at the beginning of this chapter, airports are complex entities. The focus on stakeholders has increased as an aftermath of privatization. It is therefore proposed in the next section of this chapter that due to the interdependence of airport stakeholder operations and the public accountability function of airports, privatized airport performance outcomes need to be assessed in dimensions beyond standard financial measures.

7. AIRPORT PERFORMANCE ASSESSMENT TOOL Kaplan and Norton’s (1996, 2001) Balanced Scorecard provides a framework of relational and integrated performance measures. Due to the stated interdependence amongst stakeholders, the three perspectives examined (financial, operational and community) that consist of different stakeholders are proposed to be equally important. The underpinning rationale of the model is as follows: 1. Performance measurement of airports is considered to be (a) challenging as various factors need to be considered, especially privatization consequences (b) complex as these entities are accountable to and influenced by multiple stakeholders and (c) the overall concept of measurement is subjective: organizations measure what they want to

Assessing Privatized Airport Performance

267

measure. Therefore it is important to consider the information needs of stakeholders. 2. Stakeholders rely on airport performance assessment due to their interdependence of interests given the complexity of airport operations. Therefore the sole focus of airport performance is not limited to economic performance (financial measures). Internal operational as well as external factors, derived from key stakeholder perspectives are proposed when assessing privatized airport operations. 3. Key performance measures are defined as performing three core functions according to Neely (1998): a. assisting the businesses to comply with regulatory obligation (financial measures/data in annual reports), b. assisting management to check financial and non-financial health of the business and c. supporting the strategic direction and vision of the business. It is argued that when assessing privatized airports the performance measures reflecting multi stakeholder interests have to be used instead of indicators reporting solely the financial health of the business. These measures should reflect the needs of the key stakeholder groups and items over which the airport operator has control. Therefore all dimensions have equal weighting.

7.1. The Model to Date The model to date shown in Fig. 3 relates to Sydney Airport relevant airport performance measures from stakeholders perspectives. At this stage it is not a comprehensive performance tool given the progress of the study. The overall aim is to derive three to five measures most relevant for each stakeholder category. The above model is still in its embryonic stage and the field research to date has identified the following key performance attributes as most relevant for particular stakeholders:  Airlines have identified baggage reliability, accessibility to airport facilities important measure to assess airport performance.  Desired participation in the airport development planning process by ensuring convenient layout and design of airport terminals was indicated by airlines.

Operational Perspective Operational Perspective

Revenue Traffic income per passenger Traffic income per WLU Traffic income per turnover% Commercial income per passenger Concession income per passenger Duty & Tax free income per international departing passenger Property income per passenger Property income per WLU

• • • •

Other:

•

• •

• • • •

Profitability Activity Solvency Dividend Yield

Staff cost/ employee Passenger/employee WLU per employee Staff cost per passenger Staff cost per WLU Other direct cost per passenger Other direct cost per WLU

Cost

Financial Measures/ Stakeholder Power (FLAP indicators)

Financial Perspective

“Privatized” AIRPORT “Privatized” AIRPORTPERFORMANCE PERFORMANCE

• • • • • •

• •

Security •Reported Breaches

Safety •Reported Incidence

Government Perspective

Airport Operations •Traffic Growth • Airport Development

The Airport Performance Model to Date.

• •

•

•

•

•

• • • •

Community Perspective Community Perspective

Environment

Key Environment indicator

Air quality Community relations: Response time to complaints Energy consumption (kWh/m squared floor area) Ground transport (public transport usage) Traffic Growth/ Throughput through local areas Noise (% of aircraft operations per annum) Area affected by noise Waste recycling (%) Waste disposal (weight) Water conservation (water consumption per passenger)

268

Customers/ Service / Process

Passengers: ••Cleanliness of Airport** ••Service time: check in, baggage claim, waiting times, variability of wait ••Connecting time, flight alternatives* ••Costs: food and drinks, departure fees, connection fees, airport tax ••Comfort: crowding, sound level, noise, temperature, choice of things to do Airlines (Domestic & International): ••Baggage transfer reliability** ••Accessibility to Airport Facilities* ••Design and Layout of Terminal to ensure on time departures* • •Negotiated Airport Charges Airport Operator ••Passenger served per unit time ••Baggage handled per unit time ••Flight ground delays ••People accommodated per unit time ••Gate utilization ••Power/ fuel consumption ••Security effectiveness/ crime/ theft

Fig. 3.

DOROTHEA ZAKRZEWSKI

Assessing Privatized Airport Performance

269

 Passengers identified the cleanliness of the airport and the variety of retail facilities as most important.  From the community perspective the response time to complaints, traffic growth in the surrounding areas and employment opportunities have been recognized as important indicators.  Regulators view the traffic growth as the most critical indicator of successful privatized airport operations, followed by satisfactory cooperative relations between operator and airport users measured by no intervention and less public dilemmas.  Environment groups mentioned corporate social responsibility reporting on environmental issues such as wastage reduction and air pollution and noise. Note: the measures relevant to the airport operator and financial shareholders are yet to be examined. A thorough discussion of the model attributes, its usefulness and validation is currently being undertaken and will be reported in a subsequent chapter.

8. CONCLUDING COMMENT Australia and other nations are still subject to ongoing privatization reforms and it is evident that such reforms are accepted by the public. It is asserted that the deregulation of the air transport industry generated more freedom, raised commercial attitude and created a more balanced public–private sector relationship. Stakeholders were identified using Mitchell et al. (1997) model. The preliminary findings of this ongoing study on Sydney Airport identified strong interdependence amongst the stakeholders. Sydney Airport has undergone expansion in their commercial activities. The objectives of privatizing Sydney Airport as regarded by the key stakeholders were discussed. It was identified that the shareholders and the airport operator have benefited the most from the privatization process; other airport users on the other hand claim that airports are natural monopolies. The first findings of the integrated airport performance measurement model with the aim to assess airport performance based on attributes relevant to stakeholder perspectives, was introduced. As airport privatization, particularly in Australia is quite a recent phenomenon there exists an extensive field and demand for future research of these privatized entities.

270

DOROTHEA ZAKRZEWSKI

NOTES 1. The first phase was completed in 1997 (Melbourne, Brisbane and Perth) followed by the second phase in 1998 (Darwin, Adelaide, Canberra and Hobart) and Sydney airport in 2001. 2. The stated additional security, safety and border control issues are not linked to the investigation of the privatization phenomena and thus will not be discussed in any further details in this chapter, nor form part of the performance model.

ACKNOWLEDGMENTS The author would like to acknowledge the assistance for her ongoing PhD study that is provided by the IMA FAR (Institute of Management Accountants Foundation for Applied Research) Doctoral Grant scheme. The author would like to acknowledge the assistance of her supervisors Prof. Roger Juchau and Prof. Ross Chapman in this study. This chapter was presented at the EIASM conference in Nice, France September 2007 and the author would like to thank for the feedback received from the audience.

REFERENCES Andrews, K. (1971). The concept of corporate strategy. Homewood, IL: Dow Jones-Irwin. Ansoff, H. (1965). Corporate strategy: An analytical approach to business policy for growth and expansion. New York: McGraw Hill. Baird, B. (1996). Airport privatisation: Australia’s potential as a tourist destination. Paper presented at the Airport Privatisation Conference, Hilton Hotel, Sydney, September 26. Bantel, K., & Osborn, R. (1995). The influence of performance, environment and size on the identifiability of firm strategy. British Journal of Management, 6, 235–248. Beesley, M. E. (1992). Privatization, regulation and deregulation (2nd ed.). New York: Routledge. Brenner, S. N. (1995). Stakeholder theory of the firm: Its consistency with current management techniques. In: J. Nasi (Ed.), Understanding stakeholder thinking (pp. 75–96). Helsinki: LSR-Julkaisut Oy. Dahl, R. A. (1957). The concept of power. Behavioral Science, 2, 201–215. Dodson, L. (2005). Higher fares to pay for airport police. Sydney Morning Herald, September 21. Donaldson, T. (2002). The stakeholder revolution and the clarkson principles. Business Ethics Quarterly, 12(2), 107–111. Donaldson, T., & Preston, L. E. (1995). The stakeholder theory of the corporation: Concepts, evidence and implications. Academy of Management review, 20, 65–91. Emmons, W. (2000). The evolving bargain – Strategic implications of deregulation and privatization. Boston: Harvard Business School Press. Enright, M. J., & Ng, F. (2001). Airport privatisation. HBR field case study. Centre of Asian Business Cases, University of Hong Kong.

Assessing Privatized Airport Performance

271

Friedman, A., & Miles, S. (2002). Developing stakeholder theory. Journal of Management Studies, 39(1), 1–21. Graham, A. (2003). Managing airports – An international perspective. London: ButterworthHeineman. Hill, C. W., & Jones, T. M. (1992). Stakeholder – Agency theory. Journal of Management Studies, 29(2), 131–154. Humphreys, I., & Francis, G. (2000). Traditional airport performance indicators: A critical perspective. Journal of the Transportation Research Board, 1703(Paper No. 00-0575), 24–30. Humphreys, I., & Francis, G. (2002a). Airport privatisation. British Economy Survey, 31(2), 51–56. Humphreys, I., & Francis, G. (2002b). Performance measurement: A review of airports. International Journal of Transport Management, 1, 79–85. Jensen, M. (2002). Value maximization, stakeholder theory and the corporate objective function. Business Ethics Quarterly, 12(2), 235–256. Kaplan, R., & Norton, S. (2001). The strategy focused organization: How balanced scorecard companies thrive in the new business environment. Boston: Harvard Business School Press. Kotter, J. P., & Heskett, J. L. (1996). Corporate culture and performance. New York: Free Press. Leximancer. (2007). Leximancer Manual v. 2.23. Workshop Manual. Leximancer Software Copyrighted by Andrew E. Smith – www.leximancer.com Manzoni, J.-F. (2004). From high performance organizations to an organizational excellence framework. Studies in Managerial and Financial Accounting, 14, 19–44. Mather, K. (2004). ‘‘Macquarie Airports – Airport privatisation Buyer Perspective on Sydney Airport’’ South China Morning Post (SCMP) Conference, Business and Economic Policy Seminar (Plenary Speaker), Sept. 15, 2004 Hong Kong. McGhee, M. (1996). Airports: Valuation issues. Paper presented at the Airport Privatization Conference, Sydney. Meacham, S. (2005). JetCity. The Sydney Magazine insert in the Sydney Morning Herald, September, pp. 70–78. Mitchell, R., Agle, B., & Wood, D. (1997). Towards a theory of stakeholder identification and salience: Defining the principle of who and what really counts. The Academy of Management Review, 22(4). Murphy, B. (1996). Restructuring of the Federal Airports Corporation (FAC) in a privatised environment. Paper presented at the Airport Privatisation, Sydney. Neely, A. (1998). Measuring business performance. London: The Economist Books. Parker, D. (2003a). Performance, risk and strategy in privatised, regulated industries – The UK’s experience. The International Journal of Public Sector Management, 16(1), 75–100. Parker, D. (2003b). The UK’s privatisation experiment: The passage of time permits a sober assessment. Paper presented at the Privatisation Experience in the EU, Villa La Collina, Cadenabbia (Como), November 1–2. Pfeffer, J. (1981). Power in organizations. Marshfield, MA: Pitman. Porter, M. (1990). The competitive advantage of nations. New York: The Free Press. Savage, G. T., Nix, T. H., Whitehead, C. J., & Blair, J. D. (1991). Strategies for assessing and managing organizational stakeholders. Academy of Management Executive, 5, 61–75. Sharp, J. (1996). The Government’s airport reforms. Paper presented at the The Airport Privatisation Conference Sydney, Sydney, September 26.

272

DOROTHEA ZAKRZEWSKI

Vickers, J., & Yarrow, G. (1995). Privatization – An economic analysis. Massachusetts: The MIT Press. Walsh, P., Lok, P., & Jones, M. (2005). The measurement and management of strategic change. Frenchs Forest, NSW: Pearson Prentice Hall. Wartick, S. L., & Mahon, J. M. (1994). Towards a substantive definition of the corporate issue construct: A review and synthesis of the literature. Business and Society, 33, 293–311. Watson, G. (1993). Strategic benchmarking – How to rate your company’s performace against the world’s best. New York: Wiley. Weber, M. (1947). The theory of social and economic organization. New York: Free Press. Wiltshire, K. (1987). Privatisation – The British experience: An Australian perspective. Melbourne: Longman Cheshire.

CAUSALITY IN A PERFORMANCE MEASUREMENT MODEL: A CASE STUDY IN A BRAZILIAN POWER DISTRIBUTION COMPANY Andre´ Carlos Busanelli de Aquino, Ricardo Lopes Cardoso, Marcelo Sanches Pagliarussi and Vale´ria Lobo Archete Boya ABSTRACT This study extends prior balanced scorecard (BSC) research by incorporating the effects of uncertainty, payment schemes and the strength of causal relations proposed in the performance measurement model (PMM) on the budgetary dynamics. Our analysis was restricted to two strategic business units (SBU), engineering projects and electricity distribution service, from a Brazilian electric power concessionaire. We postulate a mediated moderation association between uncertainty (treatment), bonus scheme (mediator), dispersion of payment scheme and the strength of causal relations proposed in the PMM (moderators) on budgetary slack (outcome). Additionally, we postulate that the use of accounting-based measures (ABM) also mediates the effect of uncertainty on budgetary slack. We gathered monthly observations from 102 indicators containing the target and achievement values throughout Performance Measurement and Management Control: Measuring and Rewarding Performance Studies in Managerial and Financial Accounting, Volume 18, 273–299 Copyright r 2008 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-3512/doi:10.1016/S1479-3512(08)18012-1

273

274

ANDRE´ CARLOS BUSANELLI DE AQUINO ET AL.

2002–2006. Managers were later asked to answer questionnaires about the possible cause–effect relations between these indicators, then 215 causal maps of the department and corporate indicators were drawn up. Econometric analysis provided evidence that the budgetary slack observed is directly impacted by uncertainty, and this impact is moderated by the dispersion of payment scheme. However, we did not find any evidence that supported the mediation process proposed between uncertainty, ABM and budgetary slack. Incomplete implementation of BSC and the level of analysis adopted are possible explanations for that.

1. INTRODUCTION In this study we tested the impact of associations between indicators within a Performance Measurement Model (PMM), proposed by executive officers and managers, on budgetary slack. The purpose of this study is to assess how the combination of indicators in a PMM impacts on the budgetary slack previously arranged in the goal-setting process, according to the typology proposed by Malina and Selto (2004b). PMM are used as means to promote individual behavior alignment to organization’s strategy (Epstein & Manzoni, 1997). They help to describe causal association between decisions and their intended outcomes, hence giving support to the development, communication and execution of strategic actions (Malina & Selto, 2004a). PMMs are usually associated with budgetary processes in organizations. Goals are set in the course of budgeting negotiations, and the evaluation of individual and team performance is tied both to goal attainment and to variance analysis. Accordingly, budgetary dynamics are employed to provide incentives to agents. From a communication point of view, the use of PMM can further the dialogue between members of distinct organizational levels about performance variations, and hence corrective actions may be proposed, with the help and – at the same time – enhancing the organizational learning (Abernethy & Brownell, 1999). As a result, organizations frequently use the same budget dynamics for planning the resources allocation and evaluating the performance of operations. Sprinkle (2003) suggests that this simultaneous use, for both decision-influencing and decision-facilitating purposes, should be considered in investigations aimed at understanding the role of budgets in organizations. Sprinkle’s suggestion is aligned with a perspective that considers performance-evaluation and reward systems as having both a motivational and an informational role (Merchant, 1998).

Causality in a Performance Measurement Model

275

It has also been suggested in the literature that the nature of indicators present in a PMM interacts with the contingencies of its use, with subsequent effects on the budgetary process. These effects can encompass goal acceptance by personnel, individual motivation, less information asymmetry, occurrence of budgetary slack and manipulation of accounting measures (Hartmann, 2000). Thus, it is expected that the cause–effect relations proposed in a PMM impact on the perception of how the work should be done in order to achieve goals. As a result, the perceived cause– effect relations exert influence on goal negotiation, individual motivation and task commitment (Webb, 2004). Once the PMM indicators are adopted or chosen, a bargaining process begins to set goals to be attained. The perception of a valid cause-and-effect relation between action, performance and reward should motivate the individual to pursue the aspired goals, but given the prospect of both reward and enforcement, the individual will strive to set easily attainable goals. Thus, indicators included in the PMM that reflect valid causal connection between action and outcome are considered more effective than lists of indicators without this characteristic. Nevertheless, causal relations can be mistaken both for logical relations and for finality relations. The former are established using a financial calculation and cannot be empirically tested, and the latter are based on human desires, in which an individual outlines routes or means which are believed to lead to the planned objectives. Finality relations, therefore, can be part of a credible tale constructed by the organization (Malina & Selto, 2004b), which, if institutionalized, might dictate behavior. PMMs that posit the existence of valid causal relations have been criticized with regard to authenticity of these relations. Critics have centered attention over the rhetorical arguments used by balanced scorecards (BSCs) advocates, a prevalent PMM in organizations (Norreklit, 2000, 2003). Despite the existence of numerous studies that tested BSC effects in diverse organizational settings,1 empirical tests on causality in PMMs have not resulted in statistical significance of the proposed causality associations (Malina & Selto, 2004b). Malina and Selto (2004b) tested 18 causal relationships among 9 indicators using Granger’s causality test and found only 3 significant relations. This study extends prior research by incorporating the effect of uncertainty, payment schemes and the strength of causal relations proposed in the PMM on budgetary dynamics, a motivation raised by Sprinkle (2003) and Hartmann (2000), respectively. Sprinkle (2003) questions about the impact of incentive plans on budgetary slack in PMMs that contain both

276

ANDRE´ CARLOS BUSANELLI DE AQUINO ET AL.

financial and non-financial indicators, such as BSC. Also, Sprinkle (2003) asserts the need to understand how evaluators weigh and integrate the various performance measures to form an overall appraisal of performance, and how this occurs in group settings. Thus, we included team incentives in our analysis. Hartmann (2000) suggests that uncertainty impacts on the appropriateness of choosing accounting-based measures (ABM) in PMM. Then, managers should balance the use of indicators in PMMs according to their consequence on functional and dysfunctional behaviors (Hartmann, 2000). We then tested the impact of accounting-based indicators as mediators of budgetary slack. Besides that, we incorporate the effect of payment schemes on budgetary dynamics. As a result, we postulate a mediated moderation association between uncertainty (treatment), bonus scheme (mediator), dispersion of payment scheme and the strength of causal relations proposed in the PMM (moderators) on budgetary slack (outcome). The effect of uncertainty on budgetary slack is moderated both by the dispersion of payment scheme and by the strength of causal relations proposed in the PMM. Also, this moderation is mediated by the use of bonus schemes. Additionally, we postulate that the use of ABM also mediates the effect of uncertainty on budgetary slack. The causal association assumes that the effect of uncertainty on budgetary slack is mediated by the use of ABM. Our analysis was performed in an electric power concessionaire, listed on the Sa˜o Paulo Stock Exchange (Bovespa). The company has around 1,000 employees and provides services not only in power distribution and transmission, but also in hydroelectric power plant construction and maintenance. The company is accountable to the Brazilian regulatory electric power agency (ANEEL). This study covers two strategic business units (SBU) from the holding corporation, which are engineering projects and electricity distribution service. Each of them pursue distinct targets and amounts of bonus, and the holding corporation is responsible for various targets related with both SBU. In 2001, the company started the implementation of the BSC. This process was completed in late 2006, when the consulting firm transferred BSC coordination to internal support team. Department teams in each SBU have payment schemes associated to financial and non-financial indicators, even so bonuses are granted only if the main target is attained. That applies to both executive officers and business unit managers. Additional incentives, such as promotions and prerogatives of dismissal exist, but are contingent to discretionary evaluations by superior officers.

Causality in a Performance Measurement Model

277

2. THEORETICAL STRUCTURE AND HYPOTHESIS DEVELOPMENT Budgets, as part of control systems, incorporate indicators that will be used for decision and control purposes. Comparison of individual and team performance in the organization and an association with a more objective or subjective reward distribution are examples of the use of budgetary information. Comparisons between threshold indicators and actual data indicators are periodic. Agents and team workers whose performance is measured in comparison to goals have their stress, motivation, satisfaction and effort affected, in order to keep it closer to target. The intensity of these effects depends on budget dynamics, for instance, on agents’ participation on goals establishment, on how difficult it is to reach the threshold, and or on rewards (punishment) association to thresholds’ attainment (Luft & Shields, 2003). Incentives and uncertainties related to goal’s attainment increase the relative importance of each target and their attractiveness (Webb, 2004). That may increase commitment (Hollenbeck & Klein, 1987; Klein, Wesson, Hollenbeck, & Alge, 1999), and dysfunctions such as data manipulation and budgetary slack (Hopwood, 1972; Onsi, 1973; Hirst & Yetton, 1984; Merchant, 1985a; Hughes & Kwon, 1990; Lal, Dunk, & Smith, 1996). While data manipulation is related to ex post data of actual performance, budgetary slack2 is related to ex ante data of potential performance. In both cases, information asymmetry between those that monitor performance and those who have their performance monitored increases the probability of occurrence of dysfunctions (Merchant, 1985a), and drives some firms to adopt participative budget (Shields & Young, 1993). This study emphasizes on budgetary slack as a dysfunction because it affects information relevance for decision-making process. 2.1. Budgetary Slack, Incentives and Uncertainty Agents are able to use private information to protect themselves from uncertainties related to changes on organization’s environment and on their own activities technology (Hartmann, 2000) – that affect negatively future performance forecast and judgment. Since agents’ reward (punishment) is related to budget’s thresholds attainment, they may seek to keep their performance close to those thresholds, even if some dysfunction becomes ‘‘necessary’’, such as budget slack (ex ante) or actual data manipulation

278

ANDRE´ CARLOS BUSANELLI DE AQUINO ET AL.

(ex post). Consequently, the higher the uncertainty level, more information of actual performance is needed, both financial and non-financial in nature. Also the higher the uncertainty about agent’s ability to reach thresholds related to their bonus, the lower will be the incentives’ enforcement, increasing budget slack probability. Agency Theory and prior studies (Van der Stede, 2000; Merchant & Manzoni, 1989; Merchant, 1985b; Simons, 1988; Young, 1985; Chow, Cooper, & Waller, 1988; Holmstrom & Milgrom, 1991) help in understanding that rationale. A potential enforcement is related to the possibility of ending an employment relationship or loss of potential promotion, or even a cut in annual bonus. Under such circumstances, the agent looks for means to protect himself from the risk of failing to achieve the goals and consequently receiving an unfavorable evaluation, and one of the possible protection modes is to manipulate goals, making them easily attainable (Van der Stede, 2000). In addition to goal-achievement rewards, intermediary managers are motivated to bargain looser goals to maintain autonomy and credibility, and are motivated also by the feeling of being winners when they achieve the goals which may induce them to negotiate them downward (Merchant & Manzoni, 1989). Senior managers are also encouraged to keep the subordinates’ goals highly attainable in order to increase the predictability of corporate gains, and because easily achievable goals reduce the risk of lack of commitment, and the risk of engaging in accounting information manipulation (Sunder, 1997; Mulford & Comiskey, 2002). Easily attainable budgetary goals help to keep low negative variations that reduce the need for senior managers’ analysis and interference in control. Additionally, goal flexibility permits the firm to reward intermediary managers’ performance and assure a competitive compensation package in order to prevent intermediary managers’ remuneration to be incompatible with the market.

2.2. Dispersion of Pay Structure Financial bonuses associated with teamwork tend to provoke contest that may be either positive or negative to the organization goals. So, the effect of financial bonuses is moderated by bonus-sharing criteria. Individual commitment and individualized bonus-sharing criteria may be jointly used, in order to avoid free-riding (Alchian & Demsetz, 1972; Besanko, Dranove, Shanley, & Schaefer, 2007; Leibowitz & Tollison, 1980), even if subjective evaluation indicators are used (Besanko et al., 2007). This effect increases in

Causality in a Performance Measurement Model

279

multi-tasks scenarios, in which the higher the number and complexity of activities under responsibility of each group, the higher is the difficulty in measuring individual contribution to output. Hence, bonus sharing moderates the strength of bonus incentives (Widener, 2006). This moderation is more apparent in egalitarian payment structures. Even tough egalitarian structures promote cooperation, they also increase the incidence of free-riding and could provoke dissatisfaction on most qualified employees (Bloom, 1999; Fiessbach et al., 2007; Pfeffer & Langton, 1993).

2.3. Causality and Indicators Attractiveness PMM models with valid cause–effect relations between indicators are deemed to be important for organizations on account of more reliable predictions of the future effects of present actions. They also provide insights about how actions affect results, and improve motivation and incentives (Malina & Selto, 2003, 2004a), since they differ from random lists of indicators that do not harmonize organizational objectives. Individual’s commitment to goals depends on the belief that goals are attainable, and on self-efficacy beliefs (Webb, 2004). Also, the individual’s commitment is affected by the belief that the efforts can affect performance measurements in which rewards are based (Malina & Selto, 2004a; Klein et al., 1999). Firms usually associate non-financial indicators to bonus scheme, often related to goals based on customers and internal process (Malina & Selto, 2001; Ittner & Larcker, 1998), even if these indicators have low weight (Ittner, Larcker, & Meyer, 2003). Webb (2004) suggests that firms tend to associate non-financial indicators as cause of financial indicators related to bonus.3 He found that managers are more committed when there are stronger associations between non-financial indicators and financial indicators, in comparison to when associations are weak. Thus, the perception of a valid cause–effect relation between action, performance and reward should motivate the individual to pursue the goals. The precedent discussion suggests the following hypothesis: H1. The relation between uncertainty, bonus-related indicators and propensity to budgetary slack occurs by means of a mediated moderation process. Moderation occurs through both the diffusion of payment scheme and the strength of causal relations.

ANDRE´ CARLOS BUSANELLI DE AQUINO ET AL.

280

2.4. Uncertainty, Emphasis on Accounting Numbers and Budgetary Slack Uncertainty is usually assumed as being positively related to budget variances (Lukka, 1988). Also, uncertainty is mediated by risks assumed by agents – based on how performance is evaluated (Hartmann, 2000). In the literature about reliance on accounting performance measures (RAPM) one can find that PMM may have more or less emphasis on ABM. ABM are more objective than other measures, and they alleviate ambiguity of interpretation. So, when managers’ performance measures are based on ABM, their probability to create budgetary slack increases (Onsi, 1973; Hirst & Yetton, 1984; Merchant, 1985a; Hughes & Kwon, 1990; Lal et al., 1996). In addition, uncertainty increases the risks related to noncontrollable events (Onsi, 1973; Hartmann, 2000). However, the effects of non-controllable events are mitigated by the use of subjectivity, that leads to the usage of non-accounting-based measures (NABM) (Holmstrom, 1979). Gibbs, Merchant, Van der Stede, and Vargus (2004) found that NABM included in bonus schemes are used to mitigate distortions and risks associated to non-controllable events, particularly when difficulty targets and bonuses criteria that could result in reputation and/or payment losses are at stake. The precedent discussion suggests the following hypothesis: H2. Uncertainty, ABM and propensity to budgetary slack are associated in a mediating process. Fig. 1 presents how variables are associated to each other.

[DPS] [CES] [BNS] H1 [UNC]

[SLK] [ABM]

H2

[SLK] - Budgetary slack [UNC] - Uncertanty [BNS] - Bonus [ABM] - Accounting-based measures [DPS] - Diffusion pay scheme [CES] - Cause-Effect strenght Positive, linear, additive relations Negative, linear, additive relations Positive moderator effect

Fig. 1.

Proposed Relations between the Key Variables.

Causality in a Performance Measurement Model

281

3. COMPANY BACKGROUND The company discussed in this case study is a holding corporation with shares traded on the Sa˜o Paulo Stock Exchange (Bovespa), although it is not certified by Bovespa in either of its enhanced levels of corporate governance. The holding company is controlled by members of a Brazilian family. It has nine SBUs, but this study focuses only on two of them (departmental level) besides the holding company itself (corporate level). The holding corporation controls both the SBU with 100% of their total contributed capital. Their operational activities are related to (1) hydroelectric plant engineering, maintenance and operation; and (2) electric power distribution. The entire company has a decentralized structure. The corporate level has six executive boards: financial; administrative; technical; regulatory and strategic affairs; commercial and distribution; and energy market. The SBUs’ officers’ report directly to the holding corporation’s CEO. Currently (2007–2010), three of the six executive officers were selected among employees. The CEO, CFO and the head of investor relations are members of the family that controls the holding company, as are the chairman and vice-chairman of the board of directors. That family holds 62% of both SBUs common shares (with voting rights). The holding company does not publicly disclose any information related to executive compensation, besides minimum and maximum annual amounts allowed. Fig. 2 presents this corporate structure. Implementation of the current PMM was started by the holding company in 2001. Before that, management indicators were focused on strategic planning indices belonging to a 15-year forecast and were not linked to rewards. In 2001, the company began the process to implement the BSC, which was done by external consultants with the help of internal support staff. The process was started in the holding corporation and then spread to every SBU. This process was completed in late 2006, when the consulting firm transferred BSC coordination to the internal support team, but kept a helpdesk person available for any inquiries. The PMM incorporates financial and operational performance indicators, distributed in five dimensions (financial, customers, internal processes, employees and continuous improvement) applicable to all levels of the organization, both firm level and business unit level. The process to implant the PMM was spread throughout the company, but the scope of implementation differs among the executive boards. Although it began in 2001, the process gained impetus only since 2003, when the past series of these indicators can be consulted. Even so, some indicators that already existed before implementing the PMM have continued to be

282

ANDRE´ CARLOS BUSANELLI DE AQUINO ET AL.

Majority shareholder [1]

other shareholders

Level of analysis TC = 34% VC = 62%

Corporate indicators

SBU indicators

Holding [1] TC = 100%

TC = 100%

SBU Electric distribution

SBU Engineering

other SBUs

SBU: Strategic Business Unit TC: Total Capital VC: Voting Capital [1]: Family members that participate of board of directors, such as chief executive officer.

Fig. 2.

Case Analysis Scheme.

used, and therefore have a longer history. Besides those indicators, each SBU suggests, adopts and measures specific indicator sets in order to get attained to its strategic goals. The SBU indicator thresholds are established within each SBU. Corporate targets are established through a negotiation process, and ongoing performance is measured against these targets. Variance reports, disseminated throughout the business units and areas of the company, through charts and the corporate intranet, are used to highlight deviations from targets. The evaluation process, based on BSC measures, is systemized and ongoing, at monthly meetings attended by officers, managers, advisors and key employees. Results are compared against targets and preventive and corrective actions are proposed and/or submitted for discussion. Explanations are required if targets are not met along with proposals for corrections, through total quality control (TCQ) routines. Justifications of

Causality in a Performance Measurement Model

283

events outside the control of the area responsible for the indicator (i.e., possible influences of other areas or unforeseen contingencies) were mentioned as a common practice in interviews. The evaluation process is based on BSC measures. The distribution of bonuses is tied to meeting the targets under the responsibility of each area. However, greater weight is given to financial metrics. This weighting can be an example of compensation as cited in Ittner et al. (2003), in which executives limit the impact of the subjectivity present in BSC measures. The company is now developing a positions and salaries plan, with performance measurements by competencies. For now, however the performance incentives of managers and officers consist of sporadic promotions and compensation variations, without clear assessment criteria. The bonus system has existed since 2002. As other indicators have been implemented, the basis for the bonuses has been altered. The personnel are divided into three categories: A, B and C. Level A personnel consist of operational staff, at a lower level in the hierarchy. Level B employees also consist of operational staff, at both low and intermediate positions, but distinguished by their basic salaries. Level C personnel consist of operational and executive staff at the intermediate and top positions in the hierarchy. Annual remuneration uniformly rewards A and B levels staff. In other words, the employees in the A and B groups receive the same fixed amount as profit sharing, independently of a particular individual’s salary or position, while C-level staff members receive a share that varies according to each employee’s responsibilities. The distribution of bonuses is tied to meeting the targets under the responsibility of each area. The profit sharing depends on the each SBU operational income and on some financial and non-financial indicators; however, greater weight is given to financial metrics. Under this scheme, if the indicators are achieved but the profits do not reach the target, no profit sharing is distributed. If it is exceeded, the amount to be distributed is increased proportionally to how much the target was surpassed. This target is established by upper management and approved by the board. Each area, within each SBU, receives half of its profit sharing bonus based on the attainment of its respective goals. The other half of bonus depends on the SBU performance. Individual bonus and associated incentives vary according to its level in firm’s hierarchy, as described above (categories A, B and C). People at the A and B levels have a reduced incentive, because the fixed amount does not vary with their individual salary, and thus equates various levels of intellectual and operational contributions. Neuropsychological studies have

284

ANDRE´ CARLOS BUSANELLI DE AQUINO ET AL.

shown that satisfaction with compensation is a result of comparison of an individual’s own pay package with that of his or her peers (Fiessbach et al., 2007). It is expected that providing the same bonuses to employees at different pay levels and positions in the company should reduce the satisfaction of those with higher salaries and positions within classes A and B, reducing even more the impact of the bonus incentive on individual behavior in a team. Hence, the greater the salary spread among employees at levels A and B in a particular area, the greater the loss of bonus incentive power and the stronger the propensity for free-rider behavior. In contrast, for executives and personnel at level C the individual bonus is proportional to the salary and varies according to position, seniority and duties. The bonus received by an individual at level C is always higher than that of a subordinate at the same level in the same area. Finally, a change from level B to level C within the company, even if within the same area, is seen as a promotion because of the increased compensation due to the better profit-sharing bonus (proportional to salary) and status. Regarding vertical mobility within the company, employees at levels A and B can expect to be promoted to the next higher classification, while those at level C can expect to be promoted to executive positions. From the current department managers, only two of them did not start their careers within the company and all have worked for it for more than 10 years. Starting in 2000, the company began an expansion process and gained new markets. This resulted in the promotion and transfer of staff, including managers and executives. Therefore, the SBUs are related to teams, composed of employees at levels A and B, who receive less varied salaries and equal bonuses. Because of this more egalitarian situation, they are more susceptible to incentives related to promotion to level C, or the disincentive of being let go. The corporate level indicators are related to upper level managers and executives, who receive more differentiated salaries, besides bonuses tied to their salaries and thus are less egalitarian.

4. DATA COLLECTION AND METHOD Data collection involved several stages. The first stage involved visits to all departments making up the executive boards, where we interviewed each of the 10 department managers to get an idea of the activities undertaken by them. Also, we wanted to know how the departments were structured, the number of employees, hierarchy, rules, procedures and control systems.

Causality in a Performance Measurement Model

285

We checked the existence of formalized activities, rules and regulations, as well as reward systems. This was useful to find the organization’s degree of decentralization. We also gathered information about each department’s performance indicators and the implementation status of the PMM. After the interviews, the head of the supporting team assigned to initiate the organization’s PMM provided more comprehensive information. After the first stage, we gathered the past series of indicators containing the target and achievement values of each department’s staff. In addition, the executives also provided corporate indicators. We grouped the indicators according to their relation to the reward system, whether financial or not, and with regard to other issues such as the calculation formula, first year of use and the direction of the result desired by the company for the indicator (the bigger the better or the smaller the better), as well as indicators’ frequency. Data were provided in soft copies and the authenticity of the series could be seen by comparing indicators that were controlled at both corporate and SBU levels at the same time. Some financial indicators were not provided for each SBU. Overall, past series of 102 indicators were collected for the 2002–2006 period, with monthly observations. However, the areas differ with regard to the PMM operational status, and consequently this reflected on the creation and inclusion of indicators, resulting in very short historic series and incomplete data for some indicators. Thus, the number of valid indicators for analysis was reduced to 55 between corporate and SBU indicators. Managers were later asked to indicate in questionnaires, according to their opinions, the possible cause–effect relations between indicators. They indicated the relations existing between the department’s own indicators and the relation between these and the corporate indicators. Then, 215 causal maps of the department and corporate indicators were drawn up. Table 1 presents a summary of descriptive statistics comparing the sample means and the company’s average. In the table it can be seen that the sample contains more indicators in the customer than in the BSC dimension, fewer indicators not related to bonuses and more indicators composed of formulas. The sample also contains more accounting-based indicators. All the indicators are measured quantitatively, in objective terms. Eighty percent of the indicators are financial and customer based. From a financial perspective, the majority are accounting based, unlike what occurs in the customer category, as shown on Table 2. From both perspectives, indicators are equally associated with bonuses.

ANDRE´ CARLOS BUSANELLI DE AQUINO ET AL.

286

Table 1.

Characteristics of BSC Measures for Sample and Total Company Indicators.

Number of indicators BSC dimension: financial BSC dimension: customers BSC dimension: internal process BSC dimension: employees BSC dimension: improvement Non-bonus related Bonus related Non-formula based Formula based Corporate level SBU level Non-accounting-based measure Accounting-based measure Objective/quantitative measure Subjective measure

Table 2.

Total

Sample

88 48% 27% 20% 3% 1% 63% 38% 56% 44% 49% 51% 58% 42% 100% –

55 47% 33% 18% 2% – 56% 44% 60% 40% 47% 53% 51% 49% 100% –

Comparison of BSC Measure Characteristics: ABM, Associated Bonus and Formulae Based. ABM

BSC Dimension Financial (%) Customers (%) Internal process (%) Employees (%) Improvement (%) Total (%)

NABM ABM 13 32 4 2 – 51

35 – 14 – – 49

Bonus Associated?

Formula Based?

Total

No

Yes

Total

No

Yes

Total

48 32% 18 2 – 100

25 16 13 2 – 56

22 17 5 – – 44

47 33 18 2 – 100

27 27 4 2 – 60

20 5 15 – – 40

47 33 18 2 – 100

Notes: n ¼ 55 indicators. ABM, accounting-based measure; NABM, non-accounting-based measure.

4.1. Dependent Variable: Budgetary Slack To approximate budgetary slack (SLK), we used a time series of target’s attainment. The index r ¼ target value/attained value, in a given period, represents the degree of achievement of a given target. The series was

Causality in a Performance Measurement Model

287

treated in order to result only in positive values and to obtain periodic values. For ‘‘the higher the better’’ indicators, rW1 indicates a not fully achieved target, and rr1 indicates a fully achieved target. Recurrent results of rr1 are associated with loose indicators, which have a lower incentive power. Recurrent results of rW1 suggest that the target is being ignored by management or is impossible to be achieved. Slack is associated with either rW1 or ro1, depending on whether the indicator is the higher the better or the lower the better. The proxy for SLK was the sum of modulus of all occurrences of rr1 (for ‘‘the higher the better’’ indicators) or r ¼ c1 (for ‘‘the lower the better’’ indicators) in the series, divided by the series size.4

4.2. Independent Variables: Bonus, ABM and Uncertainty The presence of an indicator in the bonus-calculation formula, as well as the link of attainment of this indicator’s target to the area’s right to share the bonus, was captured by the dummy BNS, as follows: 1 if indicator is associated with the bonus and 0 if not associated. Uncertainty hinders probability distribution of indicator’s future values’ forecast, and hinders ex ante adjustments, resulting in more experience and effort needed to consider contingences. On the other hand, it encourages setting more modest goals, creating a propensity to more slack. Uncertainty (UNC) in this case was approximated by the variability in the attainment value, measured by a standard/average deviation of that value on the series. From the various constructs presented for RAPM (see Hartmann, 2000), we adopted the nature of the accounting-financial metric as the proxy, as in Langhfield-Smith (1997), rather than just the quantitative nature of the indicator, which could exist even for non-financial indicators. This accounting-financial nature is associated with rigidity, formality and objectivity (Langhfield-Smith, 1997). So, the accounting-based nature of an indicator is captured by the dummy variable ABM, where: 1 for ABM and 0 for non-ABM.

4.3. Moderate Variables: Dispersion of Pay Scheme and Cause–Effect Pairs The dispersion in the bonus payment scheme (DPS) was approximated by the hierarchical level at which the indicator was generated and managed, as

288

ANDRE´ CARLOS BUSANELLI DE AQUINO ET AL.

well as to which team that bonus was associated. Thus, the dummy takes on 0 for indicators at the SBU level, meaning low diffusion, thus neutralizing the effect of the bonus on generating slack, and 1 for indicators at the corporate level, which means high diffusion, heightening the effect of the bonus on the generation of slack. In general, for other incentives, such as promotions, delimitation of responsibilities and risk of firing, the expected behavior is similar to the distribution of the bonus. Egalitarian structures generally are associated with larger teams, with less capacity to hold individuals responsible, smaller efficiency salaries, less specific job positions with more similar opportunities in job market, and more chance for horizontal promotion. Hierarchical structures are generally associated with smaller teams, greater ease of assessing individual responsibility, upper management positions, high vertical promotion incentives (inside or outside the firm). Hence, we believe this proxy controls for other present incentives. Regarding the strength of the causal pairs involved, we used the typology of Malina and Selto (2004b) to group the constructs that characterize the subjectivity involved in these relations, and thus the strength of the relation to make the indicator attractive. According to Malina and Selto (2004a), relation type affects indicators’ attractiveness. They suggest that performance indicators moderate the effect of incentive schemes, because in the absence of trustful cause–effect relations employees believe that reward is random-walk or based on exogenous variables – both lead to dispersion. They divided the measures relations into three types: (i) cause–effect relations, (ii) finality relations and (iii) logical relations. In the cause–effect relations, the occurrence of event X naturally implies another event Y. These relations need to be empirically observed to be proven and rejected. Finality relations are articulations of one of the possible means to achieve a desired objective. When the individual wishes to achieve an objective, he or she follows the routes presumably leading to the intended objectives, and thereby finality relations are created artificially through individuals’ wishes. Lastly, logical relations are not empirically testable, because they are part of a mechanical concept and are established by an accounting or financial calculation. Considering this logic, there is no cause–effect relation between turnover of assets and return on investment (ROI), since the ROI calculation contains in its formula turnover of assets and is therefore a logical relation. This relation cannot be empirically verified but it needs to be assessed based on financial logic or a mathematical concept. We used the cause–effect strength as proxy of its relation nature, as suggested by Webb (2004). Hence, relations based on logic and on

289

Causality in a Performance Measurement Model

mathematical concepts are strong-deductive nature (type (iii), from Malina & Selto, 2004b). Therefore, indicators’ attractiveness based on type (iii) relations are higher than the attractiveness of indicators based on type (i) and (ii) relations. Second, based on others’ studies, Sprinkle (2003) point that the number of performance measures may be inversely related to an evaluator’s ability to form accurate assessments of performance, since the agent’s bounded rationality. Furthermore, the optimal amount of performance data that should be supplied to evaluators is unclear, and may be related to the combinations and types of financial and non-financial measures employed. The higher the number of existing causal relations for one indicator the more scattered the dependence on their performance, and therefore more effort is needed to be in control of it. Similarly, it is easier for the teams to find excuses, especially if one of the causal relations is related to other responsibility centers in the company, without being able to clearly deduce the responsibility for the poor performance. We identified 215 cause–effect relations, that we clustered according to the presence of ‘‘cause’’ measure on ‘‘effect’’ measurement formula – herein denoted as deductive relation (type (iii) from Malina & Selto, 2004b), and according to ABM nature. There were identified 6 cases of type (iii) relations, according to Table 3; none of them was ABM–ABM structure. A significant portion of cause–effect relations analyzed in our sample of 55 indicators are from type (i) or (ii). Empirical evidences suggest that ABM–ABM relations are not necessarily classified as type (iii); many were classified as type (i) or (ii), since their cause indicators are not related to the respective effect indicators’ formula. Finally, strength relation of nondeductive relations (type (i) or (ii)) that are not associated to ABM–ABM Table 3. Frequency of BSC Relations. Type

ABM ’ ABM ABM ’ NABM NABM ’ ABM NABM ’ NABM Total

(iii)

(i) or (ii)

Total

0 5 0 1 6

45 88 15 61 209

45 93 15 62 215

Notes: 215 relations; 55 indicators. ABM, accounting-based measure; NABM, non-accountingbased measure.

290

ANDRE´ CARLOS BUSANELLI DE AQUINO ET AL.

nature depends on how managers identify their causality, even if that causality does not actually exist. Experience in indicators’ usage and manager’s rhetoric affect causality identification. For each effect indicator, one or more relations were listed by the respondents. The variable strength of the relation (CES) was constructed by dividing the sum of the number of type (iii) relations and the number of ABM–ABM relations by the total number of relations associated with each indicator. This methodology aims to identify relations’ strength dispersion on each of the 55 indicators analyzed. A variables’ summary is presented in Table 4.

5. RESULTS To test H1 we estimate coefficients from four regressions (Baron & Kenny, 1986; Muller, Judd, & Yzerbyt, 2005). Moderation of the effect of UNC on SLK was estimated through Eq. (1), and the effects of UNC on BNS and ABM were tested through Eqs. (2) and (3). SLK ¼ b11 þ b12 UNC þ b13 BNS þ b14 DPS þ b15 CES þ b16 UNCnDPS þ b17 UNCnCES þ 1

ð1Þ

BNS ¼ b21 þ b22 UNC þ 2

(2)

ABM ¼ b31 þ b32 UNC þ 3

(3)

Finally, the moderated effects of both BNS and ABM in SLK were estimated through Eq. (4), which also shows that both the residual direct effect of UNC on SLK (b42) and the mediator (BNS) partial effect on SLK (b43) are moderated, as seen in b47, b48, b49 and b410, respectively. SLK ¼ b41 þ b42 UNC þ b43 BNS þ b44 ABM þ b45 DPS þ b46 CES þ b47 UNCnDPS þ b48 UNCnCES þ b49 BNSnDPS þ b410 BNSnCES þ 4

ð4Þ

Table 5 provides interpretations of the slope parameters of equations 1 through 4.

1 2 3 4 5 6 7 8 9 10

SLK UNC BNS ABM DPS CES UNCDPS UNCCES BNSDPS BNSCES

9.53 0.72 0.43 0.49 0.47 0.28 0.48 0.24 0.18 0.12

Mean 26.96 0.73 0.50 0.50 0.50 0.33 0.76 0.58 0.38 0.26

SD 0 0 0 0 0 0 0 0 0 0

Minimum

Table 4.

136.62 3.53 1.00 1.00 1.00 1.00 3.53 3.53 1.00 1.00

Maximum

1 0.381 0.372 0.313 0.131 0.047 0.032 0.064 0.078 0.142

1

1 0.005 0.168 0.382 0.167 0.826 0.687 0.041 0.009

2

1 0.016 0.098 0.014 0.177 0.075 0.535 0.541

3

1 0.017 0.186 0.071 0.217 0.085 0.139

4

1 0.173 0.670 0.296 0.497 0.079

5

Descriptive Statistics and Correlation Matrix.

1 0.255 0.617 0.125 0.537

6

1 0.723 0.190 0.005

7

1 0.067 0.217

8

1 0.463

9

1

10

Causality in a Performance Measurement Model

291

ANDRE´ CARLOS BUSANELLI DE AQUINO ET AL.

292

Table 5.

Interpretation of the Slope Parameters in Eqs. (1)–(4).

Slope Parameters b12 b14 b15 b16 b17 b22 b32 b42 b43 b45 b47 b48 b49 b410

Interpretation of Slope Parameters Overall effect of UNC on SLK, controlled by average level of DPS and CES Moderator effect of DPS on SLK Moderator effect of CES on SLK Change in UNC effect on SLK as DPS increases Change in UNC effect on SLK as CES increases Overall effect of UNC on BNS Overall effect of UNC on ABM Residual direct effect of UNC on SLK at the average level of DPS and CES Mediator effect of BNS on SLK at the average level of DPS and CES Mediator effect of ABM on SLK Change in residual effect of UNC on SLK as DPS increases Change in residual effect of UNC on SLK as CES increases Change in mediator effect of BNS on SLK as DPS increases Change in mediator effect of BNS on SLK as CES increases

To corroborate H1 we need H1A. b166¼0, b226¼0, b496¼0. If b47 ¼ 0, we have full mediated moderation. H1B. b176¼0, b226¼0, b4106¼0. If b48 ¼ 0, we have full mediated moderation. With mediated moderation, there is overall moderation of the treatment effect that is b166¼0. There must be mediation and one or both of the indirect paths from the treatment (UNC) to the outcome (SLK) must be moderated (Muller et al., 2005). That is b166¼0, b226¼0 and b496¼0, and/or b176¼0, b226¼0 and b4106¼0. Table 6 presents expected values for slope parameters and their interpretations. To test H2, we should use Eqs. (5), (6) and (7) as follows (Muller et al., 2005): SLK ¼ b51 þ b52 UNC þ 5

(5)

ABM ¼ b61 þ b62 UNC þ 6

(6)

SLK ¼ b71 þ b72 UNC þ b73 ABM þ 7

(7)

293

Causality in a Performance Measurement Model

Table 6. Expected Values and Interpretation for Slope Parameters for H1 and H2. Slope Parameters’ Expected Values

Interpretation

b16o0 b22o0 b47 ¼ 0

The effect of UNC on SLK decreases as DPS increases UNC is negatively associated with BNS The effect of UNC in SLK is independent of DPS

b49o0 b17o0 b22o0 b48 ¼ 0 b410o0

The effect of BNS in SLK decreases as DPS increases The effect of UNC on SLK decreases as CES increases UNC is negatively associated with BNS The effect of UNC in SLK is independent of CES The effect of BNS in SLK decreases as CES increases

b52 W0 b62o0 b44 W0 |b72|o|b52|

UNC is negatively associated with SLK UNC is negatively associated with ABM ABM is positively associated with SLK Mediate effect of ABM on SLK

To corroborate H2 we need: b526¼0, b626¼0 and b446¼0. In addition, |b72|o|b52| However, to avoid problems with omitted variables, we use b12, b32, b42 and b44 as surrogates for b52, b62, b72 and b73, respectively. Table 7 presents the regression models that estimate Eqs. (1)–(4) with these variables. Table 8 presents a comparison of expected and estimated values for slope parameters. Overall, the results are inconsistent with both expectations for H1. However, the high degree of multicolinearity between the multiplicative terms and the individual constructs precluded a meaningful interpretation of the coefficients. Uncertainty directly affects propensity to budgetary slack in a positive way (b12 W0; b42 W0 ), though not mediated by the presence of indicator on bonus scheme (b22 ¼ 0). The effect of uncertainty on budgetary slack is moderated by bonus dilution (b42W0). Accordingly, H1A was not corroborated. Uncertainty is not related to the presence of indicator on bonus scheme (b22 ¼ 0), and bonus dilution does not moderate the effect of bonus on the propensity to budgetary slack. Also, association of indicators to bonus is not related to propensity to budgetary slack. These results may indicate that there are other (and stronger) incentives present at the organization. In recent years, one of those SBUs received annual bonus that roughly amounted for one month of managerial level salary. At the operational

ANDRE´ CARLOS BUSANELLI DE AQUINO ET AL.

294

Table 7. Least Squares Regression Results for Meditated Moderation.

UNC ABM BNS DPS CES UNCDPS UNCCES BNSDPS BNSCES Constant F Adj. R2 VIF N

SLK

BNS

ABM

SLK

1.69 (2.74) – 0.37 (1.51) 0.36 (1.56) 0.03 (0.07) 1.45 (2.34) 0.14 (0.28) – – 0.63 (3.41) 1.91 0.61 3.59 55

0.004 (0.04) – – – – – – – – 0.43 (4.53) 0.02 0.00

0.12 (1.24) – – – – – – – – 0.57 (6.03) 1.54 0.03

1.77 (2.77) 0.36 (1.72) 0.15 (0.35) 0.25 (0.86) 0.10 (0.20) 1.51 (2.29) 0.12 (0.23) 0.26 (0.51) 0.27 (0.24) 0.58 (3.03) 1.91 0.61 4.07 55

55

55

Notes: Robust standard errors (White test). t-statistics are given in parenthesis. p-valueo.1, p-valueo.05, p-valueo.01.

Table 8. Expected vs. Estimated Values for Slope Parameters. Expected Values for Slope Parameters

Estimated Values for Slope Parameters

b16o0 b22o0 b47 ¼ 0 b49o0 b17o0 b32o0 b48 ¼ 0 b410o0 b62 ¼ b32o0 b52 ¼ b12W0 b72 ¼ b42W0 b44W0 |b72|o|b52|

b166¼0 b22 ¼ 0 b476¼0 b49 ¼ 0 b17 ¼ 0 b32 ¼ 0 b48 ¼ 0 b410 ¼ 0 b32 ¼ 0 b12W0 b42W0 b44o0 |b42|W|b12|

level, in which the salaries are lower, the bonus was greater than one month of salary. On the other hand, middle level workers got bonus on the same amount as of operational level workers, who are their subordinates. In 2005 there was no bonus payment at the other SBUs. This situation had changed as a new executive officer has been assigned. The entrance of a new executive

295

Causality in a Performance Measurement Model

officer may have affected workers’ stress related to the threatening of being dismissed. Another explanation is that the effect of uncertainty on budgetary slack does not depend on its association to bonus because there must be other indicators, associated to promotion (dismissal) beliefs, with stronger effects on budgetary slack. Results point to no moderation of the direct effect of uncertainty on slack. It seems that, in spite of the informational content of cause–effect relations, managers keep protecting themselves from uncertainty through budgetary slack. Our finds do not confirm those presented by Webb (2004). With respect to H2, although b12 and b42 are both significant, regression of ABM on UNC is not, meaning that uncertainty impacts directly on the propensity to budgetary slack, irrespective of the nature of the indicator. Thus, results are inconsistent with H2 also. In order to retreat multicolinearity we estimate regression parameters with a stepwise procedure, aiming at understanding how the most significant factors affect the dependent variable slack. Table 9 presents the results from the stepwise regression between SLK and all other independent variables. As shown in Table 9, the effect of uncertainty on slack is again positive and moderated by bonus dispersion. The effect of bonus on slack increased in significance. However, the effect of accounting-based indicators on slack is also negative, contrary to the theoretical propositions. Table 9. Coefficients Estimated with Stepwise Procedure. SLK UNC ABM BNS DPS CES UNCDPS UNCCES BNSDPS BNSCES Constant F Adj. R2 VIF N

1.51 (6.57) 0.39 (2.12) 0.40 (2.12) – – 1.22 (5.47) – – – 0.490 (5.47) 18.20 0.56 2.33 55

Notes: Robust standard errors (White test). t statistics are given in parenthesis. p-valueo.1, p-valueo.05, p-valueo.01.

296

ANDRE´ CARLOS BUSANELLI DE AQUINO ET AL.

The effect of uncertainty on the indicators might be attenuated by the presence of subjective assessment, goal commitment or another dysfunctional behavior such as ex post information management (Hartmann, 2000). Results indicate that uncertainty impacts on slack irrespective of the nature of performance indicators.

6. FINAL COMMENTS Firms are increasingly using financial indicators jointly with non-financial indicators to enhance decision making and controllability. Currently, effort is still needed to understand the impacts that resulted from different mixes of indicators. Dysfunctional behaviors have been suggested to result from the choice of indicators. These behaviors can include budgetary slack and manipulation of accounting information. From an economic perspective, managers want to set motivating goals. On the other hand, they want to alleviate uncertainty and other risks associated with measurement, which could cause a reduction of the agents’ efforts. In this work we have emphasized the effect of uncertainty on the choice of performance indicators and its subsequent effect on budgetary slack. Results were inconsistent both with the hypothesis of mediated moderation process between uncertainty, bonus, dispersion of payment scheme and strength of causality with budgetary slack, and with the mediation process proposed between uncertainty, ABM and budgetary slack. The budgetary slack observed in the indicators series are directly impacted by uncertainty, and this impact is moderated by the dispersion of payment scheme. The strength of causal relations does not moderate the impact of uncertainty on slack, as proposed here and empirically observed by Webb (2004). The findings are subject to a number of limitations. The company is accountable to the Brazilian regulatory electric power agency (ANEEL) and has recently concluded the start up of BSC, with rough definitions of both positions and salaries plan and incentive plans – all these factors can affect budgetary slack incentives and were not controlled on our study. Also, the analysis was conducted at the business unit level, and some variables associated with budgetary slack operate at the individual level, such as motivation, satisfaction and performance. The BSC is not fully operational, and this might have consequences on the managers’ behaviors and on shorter indicators series. Finally, sample size and the use of multiplicative terms impacted on the robustness of econometric analysis.

Causality in a Performance Measurement Model

297

NOTES 1. See Wong-On-Wing, Guo, Li, and Yang (2007) for a review of studies that tested BSC effects in diverse organizational settings. 2. Simons (1988, p.268) appraises that ‘‘budget slack is the outcome of setting easily attainable budget goals so that individuals receive organizational rewards for performance that is below the level that would be expected if goals were tightly set’’. Dunk and Perera (1997) present other definitions for slack, but they are all related to managers’ intentional action in the sense to make goals easily attainable. 3. Webb (2004, p.931) defines causal relation strength as manager’s expectations about how much of a change on outcome is an effect of a change on the cause. 4. Indicator’s sample size was controlled because time series differ in range, from one indicator to another. Walker and Johnson (1999) use the proxy frequency, or the sum of all r occurrences in the series, to capture the estimation bias in sales targets in the negotiation between sales representatives and their managers.

REFERENCES Abernethy, M. A., & Brownell, P. (1999). The role of budget in organizations facing strategic change: An exploratory study. Accounting, Organizations and Society, 24(3), 189–204. Alchian, A. A., & Demsetz, H. (1972). Production, information costs and economic organization. The American Economic Review, 62(5), 777–795. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Besanko, D., Dranove, D., Shanley, M., & Schaefer, S. (2007). Economics of strategy (4th ed.). New York: Wiley. Bloom, M. (1999). The performance effects of pay dispersion on individuals and organizations. Academy of Management Journal, 42(1), 25–40. Chow, C. W., Cooper, J. C., & Waller, W. S. (1988). Participative budgeting: Effects of a truth inducing pay scheme and information asymmetry on slack and performance. The Accounting Review, 63, 111–122. Dunk, A. S., & Perera, H. (1997). The incidence of budgetary slack: A field study exploration. Accounting, Auditing & Accountability Journal, 10(5), 649–664. Epstein, M. J., & Manzoni, J. C. (1997). The balanced scorecard and tableau de board: Translating strategy into action. Management Accounting, 79, 28–36. Fiessbach, K., Weber, B., Trautner, P., Dohmen, T., Sunde, U., Elger, C. E., & Falk, A. (2007). Social comparison affects reward-related brain activity in the human ventral striatum. Science, 318, 1305–1308. Gibbs, M., Merchant, K. A., Van Der Stede, W., & Vargus, M. E. (2004). Determinants and effects of subjectivity in incentives. The Accounting Review, 79(2), 409–436. Hartmann, F. G. H. (2000). The appropriateness of RAPM: Toward the further development of theory. Accounting, Organizations and Society, 25, 451–482.

298

ANDRE´ CARLOS BUSANELLI DE AQUINO ET AL.

Hirst, M. K., & Yetton, P. (1984). Influence of reliance on accounting performance measures and job structure on role ambiguity for production and non-production jobs. Australian Journal of Management, 9, 53–62. Hollenbeck, J. R., & Klein, H. J. (1987). Goal commitment and the goal-setting process: Problems, prospects, and proposals for future research. Journal of Applied Psychology, 72(2), 212–220. Holmstrom, B. (1979). Moral hazard and observability. Bell Journal of Economics, 10(1), 74–91. Holmstrom, B., & Milgrom, P. (1991). Multitask principal-agent analyses: Incentive contracts, asset ownership, and job design. Journal of Law, Economics and Organization, 7, 24–52. Hopwood, A. G. (1972). An empirical study of the role of accounting data in performance evaluation. Journal of Accounting Research, 10, 156–182. Hughes, M. A., & Kwon, S. Y. (1990). An integrative framework for theory construction and testing. Accounting, Organizations and Society, 15, 179–191. Ittner, C. D., & Larcker, D. F. (1998). Innovations in performance measurement: Trends and research implications. Journal of Management Accounting Research, 10, 205–238. Ittner, C. D., Larcker, D. F., & Meyer, M. (2003). Subjectivity and the weighting of performance measures: Evidence from a balanced scorecard. The Accounting Review, 78(3), 725–758. Klein, H. J., Wesson, M. J., Hollenbeck, J. R., & Alge, B. J. (1999). Goal commitment and the goal-setting process: Conceptual clarification and empirical synthesis. The Journal of Applied Psychology, 84(6), 885–896. Lal, M., Dunk, A. S., & Smith, G. D. (1996). The propensity of managers to create budgetary slack: A cross-national reexamination using random sampling. The International Journal of Accounting, 31, 483–496. Langhfield-Smith, K. (1997). Management control systems and strategy: A critical review. Accounting, Organizations and Society, 22, 207–232. Leibowitz, A., & Tollison, R. (1980). Free riding, shirking, and team production in legal partnerships. Economic Inquiry, 18(3), 380–394. Luft, J., & Shields, M. D. (2003). Mapping management accounting: Graphics and guidelines for theory-consistent empirical research. Accounting, Organizations and Society, 28(2–3), 169–249. Lukka, K. (1988). Budgetary biasing in organizations: Theoretical framework and empirical evidence. Accounting Organizations and Society, 13, 281–301. Malina, M. A., & Selto, F. H. (2001). Communicating and controlling strategy: An empirical study of the effectiveness of the balanced scorecard. Journal of Management Accounting Research, 13, 441–469. Malina, M. A., & Selto, F. H. (2004a). Choice and change of measures in performance measurement models. Management Accounting Research, 15(4), 441–469. Malina, M. A. & Selto, F. H. (2004b). Causality in a performance measurement model. Retrieved on April 12, 2007, from http://papers.ssrn.com/sol3/papers.cfm?abstract_id ¼ 488144. Merchant, K. A. (1985a). Organizational controls and discretionary program decision-making: A field study. Accounting, Organizations and Society, 10, 67–85. Merchant, K. A. (1985b). Budgeting and the propensity to create budget slack. Accounting, Organizations and Society, 10, 201–210. Merchant, K. A. (1998). Modern management control systems: Text and cases. Upper Saddle River, NJ: Prentice Hall.

Causality in a Performance Measurement Model

299

Merchant, K. A., & Manzoni, J.-F. (1989). The achievability of budget targets in profit centers: A field study. The Accounting Review, 64(3), 539–559. Mulford, C. W., & Comiskey, E. E. (2002). The financial numbers game: Detecting creative accounting practices. New York: Wiley. Muller, D., Judd, C. M., & Yzerbyt, V. Y. (2005). When moderation is mediated and mediation is moderated. Journal of Personality and Social Psychology, 89, 852–863. Norreklit, H. (2000). The balance on the balanced scorecard: A critical analysis of some of its assumptions. Management Accounting Research, 11, 65–88. Norreklit, H. (2003). The balance on the balanced scorecard: What is the core? A rhetorical analysis of the balanced scorecard. Accounting, Organizations and Society, 28, 591–619. Onsi, M. (1973). Factor analysis of behavioral variables affecting budgetary slack. The Accounting Review, 48(3), 535–548. Pfeffer, J., & Langton, N. (1993). The effect of wage dispersion on satisfaction, productivity, and working collaboratively: Evidence from college and university faculty. Administrative Science Quarterly, 38(3), 382–407. Shields, M. D., & Young, S. M. (1993). Antecedents and consequences of participative budgeting: Evidence on the effects of asymmetrical information. Journal of Management Accounting Research, 5, 265–280. Simons, R. (1988). Analysis of the organizational characteristics related to tight budget goals. Contemporary Accounting Research, 5(1), 267–283. Sprinkle, G. B. (2003). Perspectives on experimental research in managerial accounting. Accounting, Organizations and Society, 28, 287–318. Sunder, S. (1997). Theory of accounting and control. Cincinnati, OH: South-Western Publishing. Van Der Stede, W. A. (2000). The relationship between two consequences of budgetary controls: Budget slack creation and managerial short-term orientation. Accounting, Organizations and Society, 25, 609–622. Webb, A. (2004). Managers’ commitment to the goals contained in a strategic performance measurement system. Contemporary Accounting Research, 21(4), 925–958. Walker, K. B., & Johnson, E. N. (1999). The effects of a budget-based incentive compensation scheme on the budgeting behaviour of managers and subordinates. Journal of Management Accounting Research, 11, 1–29. Widener, S. (2006). Human capital, pay structure, and the use of performance measures in bonus compensation. Management Accounting Research, 17, 198–221. Wong-On-Wing, B., Guo, L., Li, W., & Yang, D. (2007). Reducing conflict in balanced scorecard evaluations. Accounting, Organizations and Society, 32, 363–377. Young, S. M. (1985). Participative budgeting: The effects of risk aversion and asymmetric information on budgetary slack. Journal of Accounting Research, 23, 829–842.

This page intentionally left blank

PART IV: REWARDING PERFORMANCE

This page intentionally left blank

PERFORMANCE MEASUREMENT AND EXECUTIVE COMPENSATION: PRACTICES OF HIGH-PERFORMANCE COMPANIES Belverd E. Needles, Marian Powers and Mark L. Frigo ABSTRACT This study examines the links between financial performance and executive compensation for high-performance companies (HPC). HPC display sustained and superior cash flow returns, asset growth, and total shareholder returns. In previous empirical analysis, HPC companies displayed specific identifiable financial performance drivers and measures when compared to companies in the S&P 500 (Needles et al., 2004). Most recently, HPC sustained their high performance when compared to the S&P 500 over varied economic periods. Further, the research identified operating asset management characteristics of these companies, especially as they relate to the cash cycle (Needles et al., 2004). Continuing this stream of research, this study first identifies the financial and non-financial performance measures related to compensation of top management of HPC as reported in the companies’ public disclosures. Then, these findings for HPC are matched to a set of comparable Performance Measurement and Management Control: Measuring and Rewarding Performance Studies in Managerial and Financial Accounting, Volume 18, 303–323 Copyright r 2008 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-3512/doi:10.1016/S1479-3512(08)18013-3

303

304

BELVERD E. NEEDLES ET AL.

non-HPC. Finally, we evaluate the stated performance measures for executive compensation in light of the performance drivers and measures identified by previous research to be distinguishing characteristics of HPC. We hypothesize that HPC will more closely align stated performance measures for executive compensation with performance characteristics that have been shown to be characteristics of HPC. We find that HPC are more focused and unambiguous in their use of both financial and non-financial performance measures in executive compensation. This study continues our exploration of the links between strategy, execution, and financial performance by examining the links between financial performance and executive compensation for high-performance companies (HPC). HPC display sustained and superior cash flow returns, asset growth, and total shareholder returns. In previous empirical analysis, HPC companies displayed specific identifiable financial performance drivers and measures when compared to companies in the S&P 500 (Needles, Frigo, & Powers, 2006). Most recently, HPC sustained their high performance when compared to the S&P 500 over varied economic periods. Further, the research identified operating asset management characteristics of these companies, especially as they relate to the cash cycle (Needles, Frigo, & Powers, 2004). In the current study, the financial and non-financial performance measures related to compensation of top management of HPC as reported in the companies’ public disclosures are identified. Then, these findings for HPC are matched to a set of comparable non-HPC. Finally, we evaluate the stated performance measures for executive compensation in light of the performance drivers and measures identified by previous research to be distinguishing characteristics of HPC. We hypothesize that HPC will more closely align stated performance measures for executive compensation with performance characteristics that have been shown to be characteristics of HPC. Indeed, HPC are more focused and unambiguous in their use of both financial and non-financial performance measures in executive compensation and HPC outperform comparable companies on the financial measures.

PRIOR RESEARCH RELATED TO EXECUTIVE COMPENSATION Typically, compensation programs are comprised of a mixture of base salary and short-term and long-term incentives; the incentive elements rely on a combination of performance measures (Epstein & Roy, 2005). However, the

Performance Measurement and Executive Compensation

305

real concern and the focus of the current study lies with determining the performance measures that serve as a basis for the annual bonuses, or shortterm incentives, of top management. These bonuses are most commonly based on past or ex post financial or other performance incentives. Longterm incentives, such as stock options, are more difficult to evaluate due their objective of promoting future performance or ex ante measures. We focus in this chapter on the annual bonus contract because historical accounting literature is based on investigations that scrutinize the selection and behavioral consequences of annual bonus contracts (Ittner, Larcker, & Rajan, 1997). Core, Guay, and Verrecchia (2003) use agency theory to investigate the relative weights of price and non-price performance measures in total compensation and find that cash pay plays little role in CEO’s total incentives, conflicting with standard agency predictions. Based on their results, they suggest further research on performance measure use in CEO compensation. A topic of much heated debate contains the question of whether top executives, especially CEOs, actually earn their pay. In an article from the Chicago Tribune titled ‘‘CEO Pay Runs Way Ahead of Performance,’’ a study found that chief executives from 11 companies from the Standard & Poor’s 500 received $865 million over five years while operating at a loss of $640 million in shareholder value. According to the article, among one of the companies was AT&T, in which the CEO received $17.2 million last year, while AT&T shares declined five percent (Bloomberg News, 2006). Although the public eye seems to surround issues such as the fairness of these immense CEO compensation arrangements, the scholarly press focuses on conclusions based on comprehensive analysis and research. It is presumed that public companies’ boards of directors bargain at arm’s length with CEOs to negotiate pay arrangements designed to serve shareholders’ interests in an effort to legitimize compensation arrangements through an underlying corporate law-based approach (Bebchuk & Fried, 2004). This fundamental conjecture of executive compensation leads to the assumption that the board bargains at arm’s length with executives about compensation, exclusively considering the best interest of the entity and its stakeholders. The decision to provide the bonus portion of the compensation arrangement depends on the judgment of the board or its compensation committee (Bebchuk & Fried, 2004). If management teams are not driven through compensation measures, it may result in a failure to create value for a firm. Katz, Gomez-Mejia, Tosi, and Werner (2000) evaluated relationships between firm size, performance, and CEO pay. The foundation for the theory was formulated based on the agency theory. Agency theory concerns the

306

BELVERD E. NEEDLES ET AL.

relationship between a principal, the shareholder, and an agent of the principal, the company’s managers (i.e. CEO). In essence, it entails the costs of resolving disagreements between the principals and agents and aligning interests of the two groups. The principal can align these interests through monitoring of the agent to guarantee that the principal’s interests are being met. This is frequently impractical, and therefore the principal will align the interests through executive compensation. Executive compensation consists of base salary, bonus, and equity compensation such as stock options. The goal of equity compensation is for the agent to have similar interests as the shareholders and therefore be motivated to take on riskier projects that will produce higher returns. The research assessed throughout the study provided evidence that supports the theory in which organizational size is a significant determinant of total CEO pay. Combined indicators of firm size explain approximately nine times the amount of variance in total CEO pay as compared to the most highly associated performance measure. Fascinating enough, further exploration concludes that firm size accounts for more than 40 percent of the fluctuations in total CEO pay, while a firm’s operational performance accounts for less than 5 percent of the variance (Katz et al., 2000). As indicated above, many annual bonus awards rely on financial results and in prior years these measures have been criticized for encouraging an exaggerated misrepresentation on short-term accounting profits and hindering the emphasis on long-term investments (Ittner et al., 1997). Performance measures such as earnings and return on investment (ROI) have limited value and can be easily manipulated, such as through the timing of transaction recognition, when it comes to compensation of top executives. In addition, changes in share price are not a good indicator of a manager’s own performance based on the fact that a company’s stock price can increase for reasons unrelated to a manager’s own efforts and decision making (Bebchuk & Fried, 2004). ROI is one of the most common performance measures, and has been criticized for not taking into consideration the cost of capital and for being unduly influenced by external reporting rules (Ittner & Larcker, 1998). In the case of Fannie Mae, for example, the chief executive Franklin Raines received nearly $52 million from 1999 through 2003 based on performance measures such as a 15 percent annual earnings growth. Then, in September of 2004, Fannie collapsed due to the discovery of accounting improprieties, which caused investors to question whether Raines had manipulated the numbers in order to take home more money in his pocket (MacDonald & Ozanian, 2005). In other words, accounting earnings are a key factor in measuring performance for the rationale of executive compensation. Furthermore, another recent

Performance Measurement and Executive Compensation

307

study examined the outcome of earnings persistence on the style and nature of executive compensation. The study determined that accounting earnings obtain more weight in executive compensation contracts for firms with high earnings persistence than those with low earnings persistence (Ashley & Yang, 2004) Further, relying primarily on accounting earnings becomes problematic when the accounting data are noisy. Yermack (1995) found that the noisier the accounting data, the more likely it was that a board of directors would provide incentives from stock options to monitor the performance of the CEO. Hayes and Schaefer (2000) investigated observable and unobservable (to outsiders of the entity) measures of executive performance. In essence, unobservable measures are those that are only visible to those inside the firm. The research observed the premise that the unexplained variation in executive compensation contracts should predict future variation in firm performance if the unobservable measures are positively correlated with future firm performance. In other words, the hypothesis of the study is that executive compensation is a circuitous indicator of future firm performance. After testing the hypothesis through the use of executive compensation data from the Forbes Executive Compensation Surveys, the study concluded that strong evidence supports the unexplained variation in current executive compensation to be related to future performance. Implications further confirmed that as the variance of observable (to outsiders) measures of performance is higher, the relationship between unexplained variation in current compensation and future performance is stronger. To rephrase the concluding analysis of the study, the unknown fluctuations in compensation to top executives are connected to the future operations of a company. Furthermore, when the performance measures that are observable only to those inside the firm are lower, the correlation between the unknown fluctuations in current executive compensation amounts and the future operations of a company is improved. Hayes and Schaefer (2000) determined that this inference is consistent with the fact that firms substitute away from performance measures visible to the public toward measures that are unobservable to outsiders as the public measures become more strident.

EMPIRICAL OBJECTIVES In order to measure the compensation as listed in the annual bonus contract, boards of directors bestow a number of benchmarks such as strategic initiatives, fundamental performance drivers, and a widespread set of both

308

BELVERD E. NEEDLES ET AL.

financial and non-financial performance measures (Epstein & Roy, 2005). Over the past decade, more emphasis has been placed on incorporating non-financial metrics into the performance measurement process. More specifically, the use of non-financial objectives such as product innovation, customer satisfaction, and employee satisfaction has taken a significant jump in recent years (Ittner et al., 1997). In other words, both objective and subjective criteria can be used for quantitatively determining an executive’s bonus qualifications. Objective measures are goals whose attainment can readily be determined, as with financial performance measures. Subjective or discretionary measures often lead to disagreements regarding whether the executive has in fact achieved the goals, as with non-financial performance measures (Bebchuk & Fried, 2004). In accord with this background, we conduct tests of significant differences among the top 10 financial performance measures and the leading 4 nonfinancial performance measures. Further, we examine the performance of HPC versus comparables on the identified value-added financial measures over the period 2001–2005. The data for this study comes from the DEF14A, or the definitive proxy statement, the primary source of information about management’s strategies for the firm as well as management compensation. Included in the proxy statement is a summary of how members of management are paid, how much they are paid, and their incentives for payment. We expect the HPC, in contrast to their comparable companies, will more closely align stated performance measures for execution compensation with measurement characteristics that have been shown to be the attributes of HPC. We have divided the measurement results into the following three criteria: strategic goals and initiatives, key financial measures, and various non-financial measures. Any performance metrics enumerated in the proxy statement that did not meet those categories were classified separately according to the balanced scorecard.

Strategy As previously mentioned, the first performance metric that we analyzed was strategic goals and initiatives. To connect corporate operations with corporate strategic goals, the performance judgments of management must consist of key factors that provide insight into the organization’s capabilities to cultivate its future competitive position and allow for the forecast of future performance (Epstein & Roy, 2005). Strategic goals and initiatives go hand in hand with developing a comprehensive strategy to maximize the

Performance Measurement and Executive Compensation

309

potential of a variety of business opportunities and attaining selected strategic goals along with a set of individually defined strategic initiatives. A true business strategy expert must focus on emphasizing one firm goal that should drive all his or her analysis and decision making: helping the business maximize the creation of financial value (Frigo & Litman, 2004). In order for management to achieve the goal of supporting strategic objectives, he or she must have demonstrated the development and execution of strategic plans. In addition, the term strategy indirectly imposes the standard of strategically positioning the entity’s assets and strategic alliances.

Financial Performance Measures Traditional performance has been measured according to financial results. Therefore, compensation has a history of being defined in terms of financial metrics. Many companies today conventionally still use financial measures as the sole basis of measuring executive compensation. The following financial performance measures were evaluated in our study of executive compensation:          

Stock return Net income Earnings per share Earnings Before Interest and Taxes (EBIT), Earnings Before Interest, Taxes, Depreciation and Amortization (EBITDA) or earnings before taxes Operating profit/operating profit margin Cash flows Return on assets Return on equity Return on investment Earnings goals and sales growth

Non-Financial Performance Measures Traditional financial performance measures often represent lagging indicators, quantifying past or present results but demonstrating failure to forecast future performance or anticipate behavior that will result in executing and obtaining future performance objectives (Epstein & Roy, 2005). As previously mentioned, annual bonus awards calculated in conjunction with financial performance measures have been linked to management decisions

310

BELVERD E. NEEDLES ET AL.

that avoid focus on long-term investments and actually create strategies that centralize on short-term results. However, firms that have traditionally relied almost entirely on financial performance measures such as earnings, accounting profits, and stock returns, are now beginning to realize that heavy emphasis placed on financial measures is inconsistent with their relative significance (Ittner & Larcker, 1998). In other words, it is suggested that non-financial measures essentially lead to greater financial performance. The primary reasons suggested for the use of non-financial performance measures in incentive contracts for executive management are that these measures are more superior indicators of projected financial performance than conventional accounting, or financial measures and they are functional in assessing and motivating managerial performance (Banker, Potter, & Srinivasan, 2000). Instead of pertaining to short-term performance as financial performance measures, non-financial measures are positively correlated to the long-term benefits and economical well being of the entity. Non-financial measures such as meeting customer needs, internal process improvements, and an organization’s innovation of product and brand offerings reflect current managerial decisions that do not expose such efforts until subsequent years pass (Banker et al., 2000). For example, current research and development expenditures of a pharmaceutical or technology company are not likely to generate economic benefits until future years due to the extensive investigation and testing procedures of the product offerings. By incorporating non-financial indicators into the measurement systems pertaining to award contracts, many firms seek to create a wider set of measures that capture not only firm value, but also the factors leading to the creation of value in the business (Ittner & Larcker, 1998). Our study has concluded that the primary non-financial performance measures fall into four categories: human resource management, production and operations, marketing and customer service, and management performance and company-related objectives. Firstly, human resource management is comprised of employee survey results and employee retention. How well the human resource department of a company is managed reflects on the employee turnover calculation. An effective human resource department is reflected in the achievement of departmental work plans. In addition, efficiency is organized into the development of management and employees and the exercise of leadership within the industry and in the communities. Secondly, allocating production and operations to the non-financial performance sector of executive compensation encompasses the commitment to the quality of products and/or services and manufacturing

Performance Measurement and Executive Compensation

311

productivity. When companies utilize production and operations as a nonfinancial measurement in determining executive compensation, it can also include any acquisitions of products, patents, and product registrations. Objectives pertaining to this category tend to include any product cost reduction targets and innovation of certain products and/or services that promote development, growth, and expansion. In addition, ensuring an ample product supply and an effective launch of new offerings supports the leadership in advancing growth through new product development and the licensing of new products. Thirdly, marketing and customer-related non-financial objectives could be defined in terms of customer survey results and customer retention. Valuing customer-oriented goals supports the promotion of customer satisfaction and the improvement of community satisfaction. Furthermore, a large portion of marketing performance measures target factors such as market penetration and marketing expansion efforts. Lastly, management performance and company related objectives are a non-financial performance measure that directs successful leadership, guidance, and ethics. Achievement of company-related objectives could involve the implementation or completion of critical projects. Personal and individual goals of executives as approved by a company’s compensation committee along with annual bonus awards are based on the attainment of specific business and management objectives. In relation to the establishment of policies, directives, and organizational goals to position the company for growth, leadership qualities are measured by reviews from the executive’s subordinates, peers, and superiors. Individual performance goals pertain to the level of responsibility and commitment, level of performance, and past and present contribution to the achievement of organizational goals and contributions to the business unit. Management performance is demonstrated through progress toward or achievement of milestones in such executive’s area of responsibility with respect to the company’s financial performance. In addition, individual objectives of executives entail the delivery of strong financial performance along with driving the company’s growth through organizational leadership and the development of enhancing globalization in relation to the company’s business.

EMPIRICAL SAMPLE As previously mentioned, our investigation focused on two groups of companies: ‘‘High-performance companies (HPC)’’ and three publicly

312

BELVERD E. NEEDLES ET AL.

traded comparable companies (see Appendix). The comparable companies were chosen from within the same industry code and similar size and operations but without regard to financial performance. International companies were not included in the empirical sample since they do not issue statements comparable to the proxy statement. The data for the publicly traded comparable companies were found using the Standard & Poor’s Net Advantage database. As noted in previous research, the first group consisted of 38 HPC that have met the following strict criteria (Frigo, 2002a, 2002b):  10þ Years of Cash Flow Return on Investment (CFROI) about double or more the cost of capital,  10þ Years Asset Growth rates exceeding GDP, and  10þ Years Total Shareholder Return (TSR) consistent with ROIs and Growth. The Return Driven Strategy Initiative is an ongoing research study spearheaded at the Center for Strategy, Execution, and Valuation in the Kellstadt Graduate School of Business at DePaul University. The research involves the screening of more than 15,000 public companies and the identifying, documenting, and benchmarking of the strategic activities that separate the best performers from the worst (Frigo & Litman, 2004). The Return Driven Strategy Initiative influenced the development of a framework for strategic analysis designed to focus on the prioritization of business activities that lead to the highest levels of financial performance (Litman, 2003). This research was conducted in correlation with the CSFB HOLT’s Value Search database of cash flow performance and valuations of tens of thousands of companies (Frigo & Litman, 2004). Intense investigation through the use of this database led to the discovery of the 38 companies that have exhibited extraordinary financial performance, closely paralleling the Return Driven Strategy framework consisting of a set of strict requirements, or tenets, that compel the success of a firm. In doing the analyses, the HPC were grouped alphabetically according to their ticker symbols along with the ticker symbols of each of the three comparable companies listed in accordance. The data for executive compensation with regard to foreign comparable companies was excluded from the study.

Performance Measurement and Executive Compensation

313

DISCUSSION OF RESULTS Table 1 shows the study results organized for HPC and their comparable companies into the three categories:  Strategic goals and initiatives (1 measure)  Financial measures (10 measures)  Non-financial measures (4 measures).

Findings The data findings were determined through close and careful examination of proxy statements for the most recent year (usually 2005) for each individual company studied. The totals for each category were then calculated separately for the HPC and their comparable companies. Percentages were computed according to the amount of companies that illustrated positive results for the category in terms of the total number of companies and again, separately for the HPC and their comparable companies. Any non-US companies were excluded from all calculations with regard to totals due to lack of comparable reporting. Neither HPC nor comparables tend to emphasize overall strategic goals and incentives. Only about one in five (22 percent) of HPC and comparable companies mention these areas as executive compensation criteria. However, HPC are clearly more focused in execution compensation policies. For example, HPC use significantly fewer measures – both financial (2.45 per HPC versus 3.17 on average for comparables) and non-financial (.97 per HPC versus 1.31 on average for comparables). These differences are statistically significant. Further, HPC emphasize unadjusted value-creating measures, especially earnings per share (69 percent versus 32 percent) and earnings goals and sales growth (61 percent versus 43 percent). HPC are also more frequent users of net income (33 percent versus 15 percent), cash flows (17 percent versus 13 percent), return on assets (19 percent versus 5 percent), and return on equity (19 percent versus 7 percent). All these differences are statistically significant except cash flows. In contrast, comparable companies tend to use more adjusted financial measures such as EBITDA and EBIT (17 percent for comparables versus 14 percent for HPC) and ROI (12 percent versus

Table 1.

Financial and Non-Financial Measures Used in Executive Compensation.

1 12 11 1,100.00

ROI

0.033638

22 45 23 104.55

Earnings goals and sales growth

261.11

0.001061

94 174 80 85.11

Total

314

0.019543

61.11

167.31

Financial Measures OP profit/ Cash flow OP profit margin

7 5 2 28.57

2.78

43.27

ROE

0.327636 0.021496 0.040680

6 14 8 133.33

19.44

11.54

ROA

EPS

EBITDA, EBIT, or pretax earnings targets

3 17 14 466.67

19.44

6.73

Net income

5 18 13 260.00

0.090010

16.67

4.81

Stock return

0.311787

8.33

13.46

Strategic goals and initiatives

13.89

16.35

93.80 35.92

7 7 0 0.00

25 33 8 32.00

17.31

17.84 29.20

12 16 4 33.33

69.44

0.069725 0.022475 0.000048

6 7 1 16.67

33.33

31.73

8.76 315.38

8 23 15 187.50

16.67

15.38

12.71 65.38

0.494780 22.22

6.73

14.64 75.27

HPC totals Comparables totals Difference Percentage difference t-test

22.12

3.21 19.23

0.001061

8.01 96.15

0.033638

3.42 24.62

0.327636 0.021496 0.040680

0.019543

37.71 54.31

0.090010

17.95 53.85

0.069725 0.022475 0.000048

0.311787

9.94 59.62 0.494780

0.11 0.48

HPC percentages (of 36 HPC) Comparables percentages (of 104 comparables) Difference (%) Percentage difference t-test

BELVERD E. NEEDLES ET AL.

HPC totals Comparables totals Difference Percentage difference t-test

17.09 55.94 0.024697

30.56 13.46

11 14 3 27.27 0.024697

Human resources management

14.64 75.27 0.021496

19.44 4.81

7 5 2 28.57 0.021496

Production/ operations

8.01 48.08 0.124943

16.67 8.65

6 9 3 50.00 0.124943

Marketing/ customer related

9.08 25.15 0.171023

36.11 45.19

13 47 34 261.54 0.171023

Management performance and individual/company-related objectives

30.66 29.83 0.057698

102.78 72.12

37 75 38 102.70 0.057698

Total

Non-Financial Objectives

HPC percentages (of 36 HPC) Comparables percentages (of 104 comparables) Difference (%) Percentage difference t-test

Note: EPS, Earnings per share; OP, Operating; ROA, Return on assets; ROE, Return on equity; ROI, Return on investments.

Performance Measurement and Executive Compensation

315

316

BELVERD E. NEEDLES ET AL.

3 percent). These latter measures are areas in which judgment can play a role by excluding negatives from the measurements. When examining the results of the non-financial objectives in relation to executive compensation, the HPC used statistically significantly higher percentages for metrics in specific strategic areas such as human resource management (31 percent versus 13 percent), production and operations (19 percent versus 5 percent), and marketing/customer related (17 percent versus 9 percent). These results sustain the Return Driven Strategy structure in that engaging employees is one of the eleven principles of applying the framework (Frigo, 2002a, 2002b). Obtaining the right workforce and engaging it in activities that challenge and develop its ability to innovate, operate, and build on a company’s brand is a primary competitive advantage. Management and employees must have the proper incentives to be motivated and aligned toward the company’s objectives, especially through the exercise of the quality of leadership, values, and culture. In order to allow growth and prosperity of a firm, constant re-invention and integrating strategies that focus on creating new products and services is necessary. Innovative offerings is a second principle within the Return Driven Strategy framework and further supports evidence that HPC prioritize product innovation, understanding that differentiating the offering leads to value execution. Comparable companies tend to use more general statements about management performance and company-related objectives than HPC (45 percent versus 36 percent) as opposed to the specific areas discussed in the previous paragraph. As a result, in a similar manner to the financial measures, there is less focus and opportunity for the use of ‘‘judgment’’ in evaluating performance of executives in the comparable companies.

Performance Measurement In the previous section, it was observed that HPC tended to emphasize value-creating financial measures in its executive compensation practices. The performance of HPC companies was compared to the comparables for the period 12/31/2001–12/31/2005 to test whether HPC indeed performed better in these areas. Earnings per share were excluded because of the difficulty of comparing this measure among companies. Sales growth, return on assets, return on equity, cash flows returns on assets, cash flows returns

Performance Measurement and Executive Compensation

317

on stockholders’ equity, and cash flows returns on sales were included. The following hypothesis was tested for each of these measures: H. There is no significant difference between the HPC and the comparable companies. The hypothesis was rejected in every case, indicating that HPC performed significantly better on those value-creating measures that were identified as the basis of executive compensation.

FUTURE RESEARCH As noted in the discussion above, the issue of equity-based compensation is a complex one, especially as regards it being an ex ante or ex post incentive. Companies’ proxy statements enable the determination of stock options for company CEO’s and the company’s top management (including the CEO). Data is also available for the amount of exercisable and unexercisable options outstanding along with the dollar value amount. An extension of the current study to be done in the future will look at the role equity-based compensation for HPC versus comparables.

CONCLUSION In this study, financial and non-financial performances related to executive compensation were examined for HPC and a set of comparable companies. This is a continuation of our work involving the identification of characteristics of HPC. The measurement results were divided into the following three criteria: strategic goals and initiatives, key financial measures, and various non-financial measures. Tests of significant differences were conducted among the top 10 financial performance measures and the leading 4 non-financial performance measures. HPC are more likely to use unadjusted value-creating measures, especially earnings per share and earnings goals and sales growth. HPC are also more frequent users of net income, cash flows, return on assets, and return on equity. In contrast, comparable companies tend to use more adjusted financial measures such as EBITDA and EBIT and ROI. These latter measures are areas in which judgment can play a role by excluding and overcoming negatives from the measurements. When examining the results of the non-financial objectives in

318

BELVERD E. NEEDLES ET AL.

relation to executive compensation, the HPC used statistically significantly higher percentages for metrics in specific strategic areas such as human resource management production and operations, and marketing/customer related. Comparable companies tended to use more general non-financial measures than did HPC. Finally, we examined the performance of HPC versus comparables on the identified value-added financial measures over the period 2001–2005. HPC performed significantly better on those value-creating measures that were identified as the basis of executive compensation. In summary, HPC are more focused and unambiguous in their use of both financial and non-financial performance measures in executive compensation and HPC outperform comparable companies on the financial measures.

ACKNOWLEDGMENTS We wish to thank Jennifer Maes, Karen Hansen, and Robert Bellinski, who served as Research Assistants on this study.

REFERENCES Ashley, A. S., & Yang, S. S. (2004). Executive compensation and earnings persistence. Journal of Business Ethics, 50(4), 369–382. Banker, R. D., Potter, G., & Srinivasan, D. (2000). An empirical investigation of an incentive plan that includes non-financial performance measures. The Accounting Review, 75(1), 65–92. Bebchuk, L., & Fried, J. (2004). Pay without performance: The unfulfilled promise of executive compensation. Cambridge, MA: Harvard University Press. Bloomberg News. (2006). CEO pay runs way ahead of performance, study finds. Chicago Tribune, April 1. Core, J. E., Guay, W. R., & Verrecchia, R. E. (2003). Price versus non-price performance measures in optimal CEO compensation contracts. The Accounting Review, 78(4), 957–981. Epstein, M. J., & Roy, M.-J. (2005). Evaluating and monitoring CEO performance: Evidence from US compensation committee reports. Corporate Governance, 13, 75–87. Frigo, M. L. (2002a). Strategic competencies of return driven strategy. Strategic Finance, 42(12), 6–9. Frigo, M. L. (2002b). Supporting tenets of return driven strategy. Strategic Finance, 43(1), 10–12. Frigo, M. L., & Litman, J. (2004). When strategy and valuation meet. Strategic Finance, 45(2), 31–39.

319

Performance Measurement and Executive Compensation

Hayes, R. M., & Schaefer, S. (2000). Implicit contracts and the explanatory power of top executive compensation for future performance. RAND Journal of Economics, 31(2), 273–293. Ittner, C. D., & Larcker, D. F. (1998). Innovations and performance measurement: Trends and research implications. Journal of Management Accounting Research, 10, 205–238. Ittner, C. D., Larcker, D. F., & Rajan, M. V. (1997). The choice of performance measures in annual bonus contracts. The Accounting Review, 72(2), 231–256. Katz, J. P., Gomez-Mejia, L. R., Tosi, H. L., & Werner, S. (2000). How much does performance matter? A meta-analysis of CEO pay studies. Journal of Management, 26(2), 301–339. Litman, J. (2003). Drive thy value, Research and application of the most advanced strategy and valuation frameworks, Chicago and New York: Credit Suisse and CSFB HOLT. MacDonald, E., & Ozanian, M. (2005). Paychecks on Steroids. Forbes, 134–138. Needles, B. E., Frigo, M., & Powers, M. (2004). Strategy and integrated financial ratio performance measures: Empirical evidence of the financial performance scorecard and high performance companies. Studies in Managerial and Financial Accounting, 14, 113–151. Needles, B. E., Frigo, M., & Powers, M. (2006). Strategy and integrated financial ratio performance measures: Further evidence of the financial performance scorecard and high performance companies. Studies in Managerial and Financial Accounting, 16, 241–267. Yermack, D. (1995). Do corporations award CEO stock options effectively? Journal of Financial Economics, 39, 237–269.

APPENDIX High-Performing Companies Symbol

Company name

ABT

Abbot Laboratories

ADP

Automatic Data Processing, Inc.

AMGN

Amgen Inc.

AXP

American Express Company

AZN

AstraZeneca plc

Comparables Symbol AGN MYL PARS ASF CEN FDC CRL IVGN AFFX COF FMD ACF GSK SEPR BRL

Company name Allergan Inc. Mylan Labs Inc. Pharmos Corp. Administaff Inc. Ceridian Corporation First Data Corp. Charles River Laboratories International Inc. Invitrogen Corp. Affymetrix Inc. Capital One Financial Corp. First Marblehead Corp. AmeriCredit Corp. Glaxosmithkline plc Sepracor, Inc. Barr Pharmaceuticals Inc.

320

BELVERD E. NEEDLES ET AL.

APPENDIX (Continued ) High-Performing Companies Symbol

Company name

BBBY

Bed Bath & Beyond Inc.

BVF

Biovail Corp.

CTAS

Cintas Corp.

DELL

Dell Inc.

DHR

Danaher Corp.

ESRX

Express Scripts Inc.

FNM

Fannie Mae

FRX

Forest Laboratories Inc.

GE

General Electric Co.

GPS

Gap Inc.

HD

The Home Depot, Inc.

HOG

Harley-Davidson Inc.

Comparables Symbol PIR LIN WSM ALKS ADRX IPXL.PK RMK GKSR KAR HPQ IBM GTW DOV TKR CR CMX OCR APR FRE SOV CFC SNY ENDP WPI MMM TYC TXT ANF AEOS ARO LOW SHW KGFHY.PK DMH VPWS.OB UMCC.PK

Company name Pier 1 Imports Inc. Litcomp PLC Williams-Sonoma Inc. Alkermes, Inc. Andrx Group Corp. Impax Laboratories Inc. Ready Mix Inc. G&K Services Inc. Kardex Hewlett-Packard Co. International Business Machines Corp. Gateway Inc. Dover Corp. Timken Co. Crane Co. Catalyst Media Group PLC Omnicare Inc. April Group Freddie Mac Sovereign Bancorp Inc. Countrywide Financial Corp. Sanofi-Aventis Endo Pharmaceuticals Holdings Inc. Watson Pharmaceuticals Inc. 3M Co. Tyco International Ltd. Textron Inc. Abercrombie & Fitch Co. American Eagle Outfitters Inc. Aeropostale Inc. Lowe’s Companies Inc. Sherwin-Williams Co. Kingfisher New ADR Ducati Motor Holding SpA Viper Powersports Inc. Ultra Motorcycle Co.

321

Performance Measurement and Executive Compensation

APPENDIX (Continued ) High-Performing Companies Symbol

Company name

INTC

Intel Corp.

ITW

Illinois Tool Works Inc.

JNJ

Johnson & Johnson

JNY

Jones Apparel Group Inc.

KO

Coca-Cola Co.

LLY

Eli Lilly & Co.

MDT

Medtronic Inc.

MRK

Merck & Co. Inc.

MSFT

Microsoft Corp.

MXIM

Maxim Integrated Products Inc.

OMC

Omnicom Group Inc.

ORCL

Oracle Corp.

Comparables Symbol AMD TXN LLTC PNR HSC DCI PG KV-B MRX RL FOSL LIZ JSDA PEP FIZ AL PRX BNT BSX STJ BDX BMY PTIE SUPG SYMC CA RHT MU MCHP ALTR IPG LAMR RHD BMC MFE NOVL

Company name Advanced Micro Devices Inc. Texas Instruments Inc. Linear Technology Corp. Pentair Inc. Harsco Corp. Donaldson Company Inc. Procter & Gamble Co. K V Pharma CL B Medicis Pharmaceutical Corp. Polo Ralph Lauren Corp. Fossil Inc. Liz Clairborne Inc. Jones Soda Co. Pepsico, Inc. National Beverage Corp. Alcan, Inc. Par Pharmaceutical Companies Inc. Bentley Pharmaceuticals Inc. Boston Scientific Corp. St. Jude Medical Inc. Becton Dickinson & Co. Bristol-Myers Squibb Co. Pain Therapeutics Inc. SuperGen Inc. Symantec Corp. CA, Inc. Red Hat Inc. Micron Technology Inc. Microchip Technology Inc. Altera Corp. Interpublic Group of Companies Inc. Lamar Advertising Co. RH Donnelley Corp. BMC Software Inc. McAfee Inc. Novell Inc.

322 PAYX

BELVERD E. NEEDLES ET AL. Paychex Inc.

ACS TSS MGI

PFE

Pfizer Inc.

NVS DDD VRX

PII

Polaris Industries Inc.

RHI

Robert Half International Inc. Schering-Plough Corp.

ACAT HMC MPX MAN KELYA KFY KG QGLY QSC

SGP

SYK

Stryker Corp.

SYY

Sysco Corp.

WMT

Wal-Mart Stores Inc.

WYE

Wyeth

WAT ZMH HSP PFGC UNFI NAFC COST PSMT BJ TEVA IVX NXXI

Affiliated Computer Services, Inc. Total System Services, Inc. Moneygram International Inc. Novartis AG SCOLR Pharma Inc. Valeant Pharmaceuticals International Arctic Cat Inc. Honda Motor Co. Ltd. Marine Products Corp. Manpower Inc. Kelly Services Inc. Korn/Ferry International King Pharmaceuticals Inc. Quigley Corp. Questcor Pharmaceuticals Inc. Waters Corp. Zimmer Holdings Inc. Hospira Inc. Performance Food Group Co. United Natural Foods Inc. Nash Finch Co. Costco Wholesale Corp. PriceSmart Inc. BJ’s Wholesale Club Inc. Teva Pharmaceutical Industries Ltd. Ivax Corp. Nutrition 21 Inc.

PRIVATE PERFORMANCE INFORMATION IN CEO INCENTIVE COMPENSATION Eduardo Schiehll ABSTRACT Following the optimal contracting hypothesis, this study investigates the issue of whether the board of director’s ex ante choice to incorporate individual performance evaluation (IPE) measures into the CEO bonus plan rewards managerial decisions not reflected in measures of the firm’s current financial performance. Empirical results provide evidence that the use of IPE in the CEO bonus plan is an increasing function of the proportion of outsider directors on the board and a decreasing function of the informativeness of financial performance measures. This study also demonstrates how the use of IPE in incentive contracting can explain CEO cash compensation that is not explained by the firm’s current performance and governance variables. Finally, the CEO incentive cash compensation not explained by observable performance measures or governance structure is positively associated with firm future performance one year after its award. Overall, results support the optimal contracting hypothesis. IPE appears to be used to increase the informativeness of CEO actions and determine the level of current CEO cash incentive compensation.

Performance Measurement and Management Control: Measuring and Rewarding Performance Studies in Managerial and Financial Accounting, Volume 18, 323–356 Copyright r 2008 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1479-3512/doi:10.1016/S1479-3512(08)18014-5

323

324

EDUARDO SCHIEHLL

1. INTRODUCTION This study investigates the association between the use of individual performance evaluation (IPE) measures in incentive contracting and Chief Executive Officer (CEO) compensation. Specifically, it looks at the issue of whether the board of director’s ex ante choice to incorporate more subjective and private performance assessments, such as IPE, into the CEO cash bonus plan rewards managerial decisions that are not reflected in traditional financial measures of firm performance. The premise that firms use private, less aggregate performance measures to correct for imperfections in traditional financial performance measures is known as the optimal contracting hypothesis (Prendergast, 1999; Murphy & Oyer, 2001). Similar to Bushman, Indjejikian, and Smith (1996) and Ittner, Larcker, and Rajan (1997), the present research focuses strictly on the CEO annual cash bonus, and takes the perspective that an empirical analysis of the bonus component is of particular interest for two reasons. First, risk-incentive tradeoffs have been documented in many multiple-performance-measure agency models (e.g., Holmstrom, 1979; Banker & Datar, 1989; Feltham & Xie, 1994; Prendergast, 1999; Murphy & Oyer, 2001). These models suggest that the ideal performance measure should reflect a manager’s true contribution to firm value, purged of factors beyond managerial control, but including the effect of current actions on future profitability. This implies that most traditional firm financial performance measures fall short of meeting the incentive aspect. Second, from a governance standpoint, the private observable (or non-publicly observable) nature of certain types of performance measures poses additional constraints on ex ante choices. While traditional public accounting and market-based performance measures are considered either too broad or too narrow to capture the multi-dimensional tasks that CEOs must perform, an assessment of individual contribution to firm value, such as IPE, implies the use of private information that is non-verifiable by external parties such as minority shareholders. To empirically investigate this issue, three related research questions are addressed. First, whether the presence of IPE in the CEO bonus plan is associated with proxies for the informativeness of the firm’s financial performance measures and the independence of the board of directors. Second, adopting an approach similar to Core, Holthausen, and Larcker (1999), whether the presence of IPE in the CEO bonus plan explains CEO incentive compensation that is not explained by the observable current financial performance and governance structure. Third, to further

Private Performance Information in CEO Incentive Compensation

325

investigate the use of IPE for optimal incentive contracting, whether the presence of IPE in the CEO bonus plan is associated with firm future performance. If the optimal contracting hypothesis holds true, IPE is expected to be associated with unexplained variations in CEO cash bonus compensation and firm future performance. Overall, results support the optimal contracting hypothesis. The empirical evidence suggests that the presence of IPE in the CEO bonus plan is an increasing function of the firm’s investment opportunities, lack of informativeness of the accounting performance measures, and independence of the board of directors. Results also show that the presence of IPE in the CEO bonus plan can explain the amount of CEO cash bonus compensation that is not explained by the firm’s current financial performance and board independence. Finally, results show a significant association between the proportion of CEO cash bonus compensation related to the use of IPE and firm performance in the first year after the award. In sum, IPE appears to be incorporated in the CEO cash bonus plan to correct or adjust for the failure of traditional financial performance measures to capture CEO performance, and to correctly determine the level of CEO performance-based compensation. Furthermore, the link between the use of IPE and firm future performance suggests that IPE is not generally used to award excessive compensation (Core et al., 1999). The following section provides the theoretical background and research motivations. The third section describes the sample and research methods used to empirically investigate the proposed hypotheses. The fourth section presents and discusses the main empirical results. The final section summarizes the research findings and their implications.

2. THEORETICAL BACKGROUND AND RESEARCH HYPOTHESES CEO compensation typically consists of three components: a base salary, an annual cash bonus plan (short-term incentive), and a stock-based plan (long-term incentive). While salary is based on an annual fixed dollar amount and long-term incentive typically links CEO compensation to the firm’s share price at some future date, short-term incentive payoffs usually stem from more immediate, operational performance drivers.1 The CEO cash bonus plan therefore depends on the board’s ex ante choices among the many performance measures available to assess CEO performance. Moreover, performance measures for the cash bonus plan should take into

326

EDUARDO SCHIEHLL

account risk-incentive tradeoffs. That is, they should motivate without either rewarding inadequate performance or discouraging reasonable risktaking.2 According to Prendergast (1999), the primary constraint on performance-related pay is that it imposes additional risk on the agent. Thus, optimal incentive depends on the incremental compensation that awards the agent’s additional efforts, the precision with which the desired efforts are assessed the agent’s risk tolerance, and the agent’s responsiveness to incentives. Such tradeoffs make the task of designing effective incentive contracts one of the most critical components of the governance process (Murphy & Oyer, 2001). Multiple-performance-measures agency models developed by Holmstrom (1979), Banker and Datar (1989), and Feltham and Xie (1994), for example, propose that a particular performance measure should be included in a performance measure portfolio if, and only if, it provides incremental information about managerial actions over and above other, less costly measures. This theory has motivated much of the empirical research on performance measure choices in incentive contracting. Thus, it has been widely suggested that performance measures are chosen for their precision as well as their ability to provide incremental information on managerial efforts (the informativeness principle). The more precise or sensitive to managerial effort the measure, the more it reduces information asymmetry as well as risk to the agent. Since the optimal contract trades off risks and incentives at the margin, performance measures that provide lower-risk incentives are favored. Following this line of reasoning, several taxonomies have been used to empirically investigate the choice of performance measures for contracting purposes (e.g., Bushman et al., 1996; Keating, 1997; Ittner et al., 1997; Banker, Potter, & Srinivasan, 2000; Ittner & Larcker, 2002).3 Overall, these studies suggest that firms avoid the use of aggregate, firm wide financial performance measures when they are less indicative of managerial performance. Similar to Bushman et al. (1996), the present research focuses on the board of director’s use of IPE in the performance measurement portfolio to define the CEO annual cash bonus. While it is difficult to characterize the specific performance criteria or judgments upon which IPE payoffs are based, IPE measures clearly differ from traditional aggregate accounting earnings and stock price measures of corporate performance. Moreover, it is assumed that IPE contrasts with the more objective, observable, and verifiable accounting and stock-price-based performance measures. Consequently, IPE measures may involve discretionary and subjective performance assessments (Murphy & Oyer, 2001), as well as

Private Performance Information in CEO Incentive Compensation

327

non-financial and financial performance criteria that are usually privy to the contracting parties alone. Three related research questions are addressed. First, whether the presence of IPE in the CEO bonus plan is explained by proxies of the informativeness of the firm’s financial performance measures and board independence. Second, whether the use of IPE is associated with the level of CEO cash bonus compensation. Third, whether the CEO cash bonus compensation linked to the use of IPE is associated with firm future performance. Research hypotheses related to the three above research questions are developed in the following paragraphs.

2.1. Determinants of the Use of IPE in the CEO Bonus Plan Consistent with multiple-performance-measures agency models (Holmstrom, 1979; Baker, Gibbons, & Murphy, 1994; Feltham & Xie, 1994; Murphy & Oyer, 2001), performance measures for incentive contracting tend to be chosen for their ability to provide precise, timely information on managerial efforts. In theory, the relative weight placed on a performance measure should reflect its ability to convey incremental information about the performance impact of managerial actions. For example, if the firm’s stock price returns fluctuate wildly, they would provide a less precise measure of executive actions. Similarly, if a firm has lengthy product development or life cycles, then current accounting performance measures would be less sensitive to current executive actions. In addition, given the aggregate and historic aspects of accounting information, the use of financial performance measures alone might not be the most efficient way to motivate agents to act in the best interests of the firm (Keating, 1997; Bushman & Smith, 2001). Incorporating IPE into the CEO cash bonus plan could therefore correct for certain imperfections in traditional financial performance measures. This argument parallels those of Bushman et al. (1996) and Ittner et al. (1997). These authors showed that firms tend to substitute away from traditional firmwide financial performance measures when they fail to fully reflect managerial performance. Firms are therefore expected to incorporate IPE in the CEO bonus plan when traditional financial measures – normally accounting- or stock-based – were particularly poor at measuring the consequences of contemporaneous managerial actions. Thus, important factors affecting the informativeness (Holmstrom, 1979) of traditional financial performance measures are firm strategy, environment uncertainty, R&D expenditures, growth opportunities, and industry regulations

328

EDUARDO SCHIEHLL

(Banker & Datar, 1989; Bushman et al., 1996; Ittner et al., 1997). In light of this evidence, it was assumed that the use of IPE in the CEO bonus contract provides incremental information about CEO performance, and that the presence of IPE would be positively associated with factors that decrease the informativeness of traditional financial performance measures. This leads to the following research hypothesis: H1a. The presence of IPE in the CEO bonus plan tends to be positively associated with factors decreasing the level of informativeness of the firm’s financial performance measures. The balance of power among corporate managers, shareholders, and directors is established when shareholders nominate a board of directors to represent and protect their interests and ensure that senior managers fulfill their responsibilities to the firm. Board composition therefore refers to the distribution of members according to their primary allegiance, which may be either to shareholders (outside) or management (inside) (Bellavance & Schiehll, 2006). The governance literature also suggests that vigilant outside boards tend to demand complex governance mechanisms. The aim is to reduce information asymmetry and more closely monitor managerial decision-making. In contrast, a board with a higher proportion of inside directors may demand less private information, such as IPE, either because the directors already have good knowledge of the firm’s performance or because they are reluctant to monitor fellow CEOs (Core et al., 1999; Bushman & Smith, 2001; Sloan, 2001). Assuming that the presence of IPE in the CEO bonus plan provides incremental information on managerial actions, thereby improving optimal contracting (Bushman et al., 1996; Murphy & Oyer, 2001), the presence of IPE in the CEO bonus plan is expected to be associated with board independence. Specifically, the presence of IPE in the CEO bonus plan should be positively associated with the proportion of independent directors on the board and negatively associated with a combined board chair/CEO position (CEO duality). The following hypotheses were tested: H1b. The presence of IPE in the CEO bonus plan tends to be positively associated with the proportion of independent directors on the board. H1c. The presence of IPE in the CEO bonus plan tends to be negatively associated with CEO duality.

Private Performance Information in CEO Incentive Compensation

329

2.2. IPE and CEO Compensation This study aims to investigate the argument that boards of directors incorporate IPE into the CEO bonus plan to mitigate incentive distortions caused by the lack of informativeness of traditional accounting and marketbased performance measures. Thus, the association between the presence of IPE in the CEO bonus plan and CEO incentive compensation is examined. Following the results of Aggarwal and Samwick (2003), it can be argued that this association is a necessary condition to support the optimal contracting hypothesis. The premise is that the CEO bonus plan incorporates private, less aggregate, and non-verifiable performance information, such as IPE, in order to better monitor, evaluate, and reward actions not captured by publicly observable measures of firm performance. Therefore, the presence of IPE should be associated with compensation levels that are unexplained by the firm’s contemporaneous financial performance or governance structure. For instance, Baker et al. (1994) show that when a firm cannot benefit from contracting based on verifiable measures alone, publicly and privately observable performance measures are brought in to bolster the optimal contracting process. Accordingly, this study aims to investigate the ability of the presence of IPE to explain CEO cash bonus compensation levels. However, because IPE can be used to adjust incentive compensation upwards or downwards, the direction of this association is not predicted. Similar to the studies by Core et al. (1999), and Hayes and Schaefer (2000), the association between the presence of IPE in the CEO bonus plan and CEO incentive compensation is investigated, while controlling for observable measures of contemporaneous performance, board independence, and other governance attributes. This leads to the following hypothesis: H2. The presence of IPE in the CEO bonus plan tends to be significantly associated with the proportion of CEO incentive cash compensation that is unexplained by the firm’s observable measures of financial performance and board independence.

2.3. CEO Cash Bonus and Firm Future Performance As discussed earlier, the optimal contracting hypothesis predicts that the use of less aggregate, privately observable performance information for incentive purposes improves the incentive contracting process by

330

EDUARDO SCHIEHLL

incorporating performance dimensions of multi-task managerial efforts that are not captured by traditional financial measures. More precisely, the presence of IPE in the CEO cash bonus plan is assumed to improve the incentive contracting process by allowing compensation committees to reward the CEO based on qualitative, strategic, and more forward-looking aspects of firm performance (Baker et al., 1994; Murphy, 1999; Fisher, Maines, Peffer, & Sprinkle, 2005). Hence, and consistent with Core et al. (1999) and Hayes and Schaefer (2000), boards of directors are expected to incorporate non-verifiable performance assessments, such as IPE, to reward executive decisions that are not reflected by the firm’s contemporaneous financial performance, but which aim to benefit firm future performance. The level of executive incentive compensation that is unexplained by the observable measures of contemporaneous performance and board independence should therefore be positively associated with firm future performance. The premise is that the CEO cash bonus award incorporates both public and private (unobservable to outsiders) performance information, and that private information, when used optimally, should be associated with firm future financial performance (Hayes & Schaefer, 2000). This leads to the following research hypothesis: H3. The CEO cash bonus compensation that is unexplained by the firm’s financial measures of contemporaneous performance and board independence tends to be positively associated with the firm’s future financial performance.

3. METHODS 3.1. Sample and Data Collection Empirical testing was performed on data from a sample of 178 Canadian public companies listed on the Toronto Stock Exchange (TSE) in 1999 in the following industries: biotechnology, forest products, industrial products, high technology, consumer products, merchandising, and media. Data was collected from two separate sources. Information on CEO compensation and firm governance structure was retrieved from the SEDAR (System for Electronic Document Analysis and Retrieval) database, as disclosed in the firms’ Management Proxy Circulars. Financial information such as firm’s growth opportunities, noise in accounting information, performance, and other control variables were obtained from the Compustat database. Firms

331

Private Performance Information in CEO Incentive Compensation

Table 1. Industry

Sample Distribution by Industry. N (%)

IPE ¼ 1

Growth

R&D Intensity

Biotechnology High Technology Knowledge-based Forest products Industrial products Consumer products Merchandising Media Capital-based

17 (12.6%) 15 (11.1%) 32 (23.7%) 15 (11.1%) 42 (31.1%) 20 (14.8%) 14 (10.4)% 12 (8.9%) 103 (76.3%)

58.3% 58.8% 58.6% 64.7% 57.1% 63.2% 25.0% 58.3% 54.7%

4.760 2.392 3.576 1.029 1.409 2.094 2.238 2.437 1.839

3.272 0.219 1.746 0.031 0.056 0.024 0.105 0.039 0.051

Total

135 (100%)

55.6%

2.241

0.475

Note: IPE ¼ 1 if firm uses IPE in bonus contracts; Growth is measured by the market-to-book ratio; R&D intensity is measured by the ratio of R&D expenditures to the book value of total assets.

must have had available data from a minimum of five years before and three years after the compensation award so that certain variables could be measured appropriately. Eliminating observations for which data was missing left 135 usable observations. Table 1 presents sample composition by industry. According to the data, the biotechnology and high technology industries present quite different levels of growth opportunities and R&D intensity. Hence, these knowledge-based industries are distinguished from the more capital-based industries in subsequent analyses.

3.2. Variables 3.2.1. Dependent Variables Three dependent variables were used for the empirical investigation: presence of IPE in the CEO cash bonus plan, CEO cash bonus compensation, and firm future performance. Presence of IPE in the CEO bonus plan was measured by a dummy variable, coded one if the compensation committee report contained in the firm’s Management Proxy Circular included expressions such as ‘‘individual performance evaluation’’ or ‘‘individual performance objectives’’ in its description of the CEO cash bonus plan, and coded zero otherwise. The approach is similar to those used by Ittner et al. (1997) and Ittner and Larcker (2002) to capture the relative importance placed on financial and non-financial performance measures in

332

EDUARDO SCHIEHLL

incentive plans. The appendix documents examples of sampled firms that scored one for this variable.4 For purposes of this investigation, cash incentive compensation (cash bonus alone or cash bonus plus salary) was taken as a more appropriate measure than total CEO pay-related wealth. Thus, two measures of CEO incentive cash compensation were used: cash bonus alone and total cash compensation (cash bonus plus salary). Panel B, Table 2, summarizes the data, including salary, cash bonuses, total cash compensation, stock options,5 and total compensation. Similar to the study by Core et al. (1999), two variables were used to capture firm future performance: return on equity (ROE)6 (an accountingbased performance measure), and stock price return (MKT) (a marketbased performance measure). Firm future performance was measured at one (tþ1) and three (tþ3) years after the incentive cash compensation award. ROE was computed as: ROEtþ1 ¼ Earnings before taxestþ1 =Book value of equityt ROEtþ3 ¼ Earnings before taxestþ3 =Book value of equitytþ2 MKT as: MKTtþ1 ¼ ðStock pricetþ1 þ dividends per sharetþ1 Þ=Stock pricet MKTtþ3 ¼ ðStock pricetþ3 þ dividends per sharetþ3 Þ=Stock pricetþ2

3.2.2. Independent Variables 3.2.2.1. Informativeness of Financial Performance Measures. Consistent with the prior literature, three variables were used to capture the factors expected to influence the precision and sensitivity (informativeness) of traditional financial performance measures. Similar to Bushman et al. (1996), and Ittner and Larcker (2002), the firms’ investment opportunities were measured by the market-to-book ratio and R&D intensity was measured by the R&D expenditures to total assets ratio.7 In a performance evaluation context, the intuition underlying the market-to-book ratio is that it captures the consequences of managerial actions and decisions that are reflected in the market price but not in the accounting information. Thus, the higher the market-to-book ratio, the lower the precision with which the accounting information captures the impact of executive decisions on firm performance. Whereas the intuition underlying the explanatory power of R&D intensity is that it captures the extent to which the firm follows an innovation-oriented strategy. The literature assumes that, in an innovation-oriented context, the

Private Performance Information in CEO Incentive Compensation

333

consequences of managerial actions are not immediately reflected in traditional financial information (sensitivity), leading to the use of alternative performance measures. Hence, the presence of IPE in the CEO cash bonus plan is expected to be positively related to the firm’s growth opportunities and R&D intensity. Following a similar proxy used by Ittner et al. (1997), exogenous noise in the firm’s financial information was measured using Fisher z-scores8 for the correlations between return on assets and stock market returns over five years prior to the proxy data. Noise in financial information is assumed as inversely related to the correlation between short-term accounting and market returns. As such, the presence of IPE in the CEO bonus plan is expected to be negatively associated with the noise variable. 3.2.2.2. Board Independence and Other Governance Variables. Corporate governance data were collected directly from Management Proxy Circulars contained in the SEDAR9 database for the 1999 and 2000 fiscal years. The database was designed to represent the corporate governance mechanisms present at the beginning of and throughout the performance year when the cash incentive compensation was awarded. Thus, governance information was collected from the firm’s 1999 Management Proxy Circular and the amount of the 1999 cash bonus was collected from the firm’s 2000 Management Proxy Circular. The proportion of outside directors on the board represents board independence and its efforts to oversee senior management, particularly the CEO.10 An outside director is independent of management and free of any interest, business or other, that could, or could reasonably be perceived to, materially interfere with the director’s ability to monitor management.11 A number of empirical studies suggest that agency problems are higher when the CEO is also the board chair (e.g., Barkema & Gomez-Mejia 1998; Core et al. 1999). Therefore, CEO duality was used as an indicator variable, set at one if the CEO was also the board chair and zero otherwise. Three other corporate governance variables were measured to control for governance mechanisms that may affect board independence and performance measure choices in the CEO bonus plan. CEO ownership was measured by the ratio of number of shares directly or indirectly controlled by the CEO to total outstanding shares in the same period. This variable was intended to measure ownership by professional CEOs, and excluded CEOs who were also large shareholders. As such, CEO ownership greater than (or closer to) 10% and CEOs with family connections to a large shareholder was considered large shareholdership. Board shareholdings were also considered, measured by the ratio of the number of shares directly or indirectly controlled

334

EDUARDO SCHIEHLL

by the board of directors as a group to total outstanding shares, excluding CEO ownership if the CEO was also a director. Finally, the monitoring effect of large shareholdings in the ownership structure was measured, and whether the largest shareholder was an individual (including a family) or institution. 3.2.2.3. Economic Determinants of CEO Cash Bonus Compensation. The standard agency research on the determinants of CEO incentive compensation (e.g., Tosi, Werner, Katz, & Gomez-Mejia, 1998; Core et al., 1999) predicts that CEO compensation is an increasing function of firm performance and environmental uncertainty. Accordingly, return on assets and standard deviation of the MKT were considered for the years when the compensation was awarded. Executive incentive compensation also tends to increase with CEO tenure, which is measured by the number of years as CEO in the present firm (Barkema & Gomez-Mejia, 1998). It is also expected that larger firms with greater growth opportunities and more complex operations would award greater cash incentive compensation to their CEOs. Finally, an industry indicator variable is included to control for industry differences in the demand for CEO incentive compensation. 3.2.2.4. Economic Determinants of Firm Future Performance. Following similar model specifications used by Core et al. (1999), five variables are used to explain the firm’s future performance. Sales and market-to-book ratio are measured for the year prior to which firm future performance is measured to control for past performance and growth opportunities, respectively. The industry variable is used to control for macroeconomic factors that may have affected the firm’s performance. Standard deviation of the ROE and MKT over the five years prior to the proxy data are used to control for firm uncertainty. Finally, natural log of the firm’s market capitalization is used to control for firm size.

3.3. Descriptive Statistics In Table 2, Panels A, B, C, and D report the descriptive statistics for the dependent and independent variables used in the empirical investigation. Panel A, Table 2, presents the IPE and corporate governance measures. Fifty-five percent (75 out of 135) of the CEO bonus plans in the sample incorporated IPE criteria. Cluster analysis reveals that no firm in the sample used IPE exclusively in the CEO bonus plan. In addition, return on assets and ROE were the most frequently used financial performance measures,

Private Performance Information in CEO Incentive Compensation

335

along with IPE, in the CEO cash bonus plan descriptions. Table 1 reports the distribution of IPE by industry. Results show a relatively similar pattern across industries, except for lower use of IPE in the merchandising industry. Consistent with prior studies, the distribution of CEO ownership is skewed, with a mean of 8.4% of the firm’s outstanding equity, while the variable measuring large shareholder ownership shows that this sample of Canadian public companies presents a relatively significant level of ownership concentration, at a mean of 35%. Ownership concentration is also reflected in board of directors’ ownership, at a mean of 22% of the firm’s outstanding equity. For instance, Panel A of Table 2 shows that 44% of firms are familyowned (59 out of 135), with an average 17.1% block of voting shares, and 48% have institutional ownership (65 out of 135), with an average 18.6% block of voting shares. Eleven firms in the sample have diffuse ownership.12 Average board size is nine members (3–16 members), with two-thirds of members outsiders and 42% of board chairs acting simultaneously as CEO. Table 3 shows the correlations between variables. As expected, family ownership is negatively related to institutional ownership and proportion of outsiders on the board, but positively related to CEO ownership and CEO duality.13 Panel B of Table 2 presents the compensation data. It reports descriptive statistics on CEO salaries and cash bonuses as well as stock option values. Although the determinants of stock options were not considered in this study, one of the objectives was to document the relevance of the cash bonus compensation to total CEO incentive compensation. Average (median) salary and cash bonus in the sample are $378,000 ($340,000) and 289,000 (110,000), respectively, for an average (median) total cash compensation of $667,000 ($501,250). CEOs were also granted options, at an average (median) value of $348,000 ($28,150), and an average (median) total compensation of $1,103,000 ($593,000). Table 3 presents the Pearson’ correlation matrix. It shows that cash bonus compensation is positively correlated with firm size and family shareholdings, but negatively correlated with presence of IPE in the CEO bonus plan and the noise variable, which measures the informativeness of financial performance measures. Results are consistent with Bushman et al. (1996) and Ittner et al. (1997). Panel C of Table 2 presents some financial data on the sampled firms. Average (median) sales and market value are $1.3 billion (349 million) and $1.7 billion (250 million), respectively. Average (median) total assets and R&D intensity are $1.5 billion (288 million) and $0.5 (0.0), respectively. The growth opportunities variable is right-skewed, with a mean (median) of 2.24 (1.64), while the noise of financial measures and standard deviation of ROE variables are left-skewed, with a mean (median) of 0.05 (0.23) and 14%

Minimum

1.000 0.877 44.000 1.000 16.000 1.000

Maximum

0.221 0.357 0.437 0.171 0.481 0.186 0.081

0.555 0.0838 5.793 0.422 8.867 0.660

Mean

0.114 0.258 0.000 0.000 0.000 0.000 0.000

1.000 0.006 3.000 0.000 8.000 0.714

Median

0.251 0.254 0.498 0.247 0.501 0.260 0.274

0.499 0.175 7.829 0.496 2.667 0.181

Standard Deviation

Salary

135 289,425 470,476 0.0000 3,065,000

Bonus

135 667,353 601,155 37,100 3,565,000

Total Cash Compensation

135 347,872 931,837 0.0000 5,727,596

Options

135 1,102,569 1,461,828 37,100 8,177,596

Total Compensation

336

N

0.000 0.000 0.000 0.000 3.000 0.250

0.937 0.937 1.000 0.929 1.000 0.732 1.000

Descriptive Statistics.

135 135 135 135 135 135 0.001 0.000 0.000 0.000 0.000 0.000 0.000

Table 2.

Variable Name

135 135 135 135 135 135 135

Panel A: IPE and corporate governance variables

Use of IPE in the CEO bonus plan CEO ownership CEO tenure CEO duality Board size Proportion of outsider directors on the board Board shareholdings Large shareholdings Family shareholder dummy Family shareholdings Institutional shareholder dummy Institutional shareholdings Diffuse ownership dummy

135 377,927 229,728 27,400 1,450,000

Panel B: Compensation data (in Canadian dollars)

N Mean Standard deviation Minimum Maximum

EDUARDO SCHIEHLL

Percentiles

Panel C: Financial data

N Mean Standard deviation Minimum Maximum Percentiles 25 50 75

Sales

135 1356.4 3286.3 .0000 27080.2 79.9 349.0 1070.0

25 50 75

Assets

135 1484.6 3770.3 7.835 30338.2 80.193 288.3 1075.7

210,000 340,000 500,000

Growth

271,941 501,250 859,460

Standard Deviation of Return on Equity

0.0000 110,000 362,000

R&D Intensity

Noise in Financial Measures

135 0.029 0.130 0.638 0.376 0.354 0.044 0.080

337,477 593,482 1,071,790

135 0.495 0.323 0.159 2.059 0.257 0.488 0.640

Return on Standard Assets t1 Deviation of Market Return

0.0000 28,125 217,500

Market Capital.

135 1699.7 5561.0 8.0675 54405.3 89.1 249.9 1190.9

Market Returntþ3

135 0.140 0.246 0.008 2.133 0.027 0.066 0.145

Market Returntþ1

105 0.056 0.418 0.873 1.217 0.237 0.016 0.313

135 0.050 1.247 4.144 3.541 0.791 0.228 0.714

Return on Equitytþ3

135 0.247 0.626 0.651 2.339 0.094 0.065 135

135 0.475 2.276 0.000 16.940 0.000 0.000 0.069

Return on Equitytþ1

105 0.020 0.661 5.523 2.503 0.030 0.083 0.132

135 2.241 2.333 0.307 18.320 1.010 1.640 2.527

135 0.008 0.516 4.452 0.521 0.025 0.094 0.157

Panel D: Variables measuring future performance

N Mean Standard deviation Minimum Maximum Percentiles

25 50 75

Private Performance Information in CEO Incentive Compensation

337

338

Table 3.

EDUARDO SCHIEHLL

Correlation matrix: Dependent and independent variables. IPE

1 2 3 4

5

6

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Growth R&D intensity Noise in financial measures Standard deviation of return on equity Standard deviation of market return Percentage of outside directors CEO duality Board shareholdings Large shareholdings Family large shareholdings Institutional large shareholdings CEO ownership CEO tenure Market return tþ1 Market return tþ3 Return on equity tþ1 Return on equity tþ3 Cash compensation Bonus Ln sales t LogN assets t (size) Ln market capitalization Return on assets t

1

2

3

4

5

6

7

8

9

0.180 0.052 0.192 –0.082 –0.221 –0.022 0.124

0.109

0.143

–0.115

0.062

0.174

0.061

–0.177

0.111

0.016

0.152

–0.111 –0.120

–0.138 –0.042

–0.044 –0.117

–0.036

0.040

–0.158

–0.035

–0.007

–0.116

–0.002

0.046

–0.044

0.273

0.008

0.068

–0.131

–0.048 0.053

–0.154 –0.098

–0.044 –0.125

–0.211 –0.319

0.215

–0.158

–0.169

–0.309

0.065

0.365

–0.021

–0.045

–0.277

0.202

0.479

0.224 –0.133

–0.123

–0.040

–0.050 –0.016 –0.276 –0.142 –0790

–0.051 –0.010 0.614 –0.102 –0.017

–0.223 0.391 0.551 0.231 –0.240 –0.417 0.256 0.035 –0.054 –0.047 –0.148 –0.141 –0.084 –0.037 0.034 0.078 –0.145 0.153 0.066 0.125

–0.066

–0.163

–0.032

0.055

0.124

0.029

–0.173

–0.044

–0.047

0.048

–0.026

0.140

–0.146 0.080 –0.083 –0.226 –0.172 0.001 –0.129 –0.349 –0.207 –0.087 –0.017 –0.240 –0.373 0.085 0.161 –0.095 –0.172 0.236 –0.263 –0.223 0.001

0.117 0.039 0.032

–0.046 0.065 0.024

0.117 0.180 0.058

–0183

–0.222

0.063

0.162 –0.069

–0.156 –0.031 0.164 –0.126 –0.059

–0.171 –0.147 0.136 0.041 –0.060

–0.054 –0.058 –0.030 –0.191 –0.309

–0.099

–0.006

–0.088

0.067

–0.100

0.059

–0.113

–0.128

0.094 –0.076

0.325 –0.061 0.018

–0.042 0.052 –0.056 –0.035 0.101

–0.009

–0.387 –0.053

–0207

–0.025

–0.346 –0.074

0.205 –0.159

–0.128

0.096

significance at 1%; significance at 5%; significance at 10% level (two-tailed).

–0.098

0.144

0.459 0.540

0.218

339

Private Performance Information in CEO Incentive Compensation

10

11

12

13

14

15

–0.176 0.029

0.241

–0.196

0.201

16

17

18

19

20

21

22

–0.499 0.502 –0.250 –0.168 0.396 0.214 –0.026 –0.113 –0.071 0.084 0.035 –0.129 0.205 –0.097 0.006 0.117 0.002 0.077 0.083

–0.054

0.100

0.123

0.199

–0.052

–0.097

–0.055

0.189 0.069 0.036

–0.065 0.110 0.022

–0.067 –0.059 –0.041 0.047 –0.179 –0.006

0.063 0.122

0.123 0.097

–0.089 –0.107 –0.084 –0.153

–0.339 –0.279 –0.036 0.083

0.018

–0.124

0.031 0.149

–0.112

0.009 0.176 0.042

0.125 0.442 0.249

–0.122 0.013 0.003

0.937 0.487 0.3586 0.386 0.249 0.584

0.093

0.202

–0.179

0.596 0.503

0.505 0.447

0.164

0.634 0.165

–0.007

0.191

0.431

0.145

0.153

0.232

340

EDUARDO SCHIEHLL

(6.6%), respectively. These statistics are comparable to those obtained by Craighead, Magnan, and Thorne (2004) on a sample of 139 of the 150 largest firms listed on the TSE.

4. DATA ANALYSIS 4.1. Determinants of IPE and CEO Cash Incentive Compensation As discussed earlier, this empirical investigation begins by examining whether the presence of IPE in the CEO bonus plan is associated with variables that measure the informativeness of the firm’s financial performance measures and the independence of the board of directors. As such, the following regression model is proposed: IPEit ¼ f ðInformativeness of financial performance measures þ outside directors on the board þ CEO duality þ control variablesÞ IPEit ¼ a0 þ b1 GROWTHit þ b2 R&Dit þ b3 NOISEit þ b4 OUTit þ b5 DUALit þ b6 BSHit þ b7 LSHit þ b8 INDi þ b9 SIZEit þ 1

ðModel 1Þ

where IPEit ¼ dummy variable, which takes the value 1 if the CEO bonus plan incorporates IPE and otherwise equals 0 Variables measuring the informativeness of firms’ financial performance measures GROWTHit ¼ market-to-book ratio R&Dit ¼ R&D intensity (R&D/Assets) NOISEit ¼ standard deviation of annual ROE over past five years Variables measuring independence of the board of directors OUTit ¼ percentage of outside directors on the board DUALit ¼ dummy variable, which takes the value 1 if CEO is board chair and otherwise equals 0

Private Performance Information in CEO Incentive Compensation

341

Control variables BSHit ¼ board of directors ownership LSHit ¼ ownership percentage by large shareholders SIZEit ¼ natural log of total sales INDit ¼ dummy variable, which takes the value 1 if firm is a capital-based industry and otherwise equals 0 To examine the association between the presence of IPE in the CEO bonus contract and CEO incentive compensation, the following regression model is proposed: BONUSit ¼ f ðIPE þ economic determinants of CEO incentive compensation þ CEO entrenchment þ control variablesÞ

BONUSit ¼ a0 þ b1 ðIPEÞ þ b2 GROWTHit þ b3 ROAit þ b4 RISKi þ b5 CEOSHit þ b6 DUALit þ b7 CEOTENUREit þ b8 LSHit þ b9 SIZEit þ b10 INDi þ 1 ðModel 2Þ where BONUSit ¼ CEO incentive cash compensation award for performance year t IPEit ¼ dummy variable, which takes the value 1 if the CEO bonus contract incorporates IPE and otherwise equals 0 Economic determinants of CEO incentive compensation GROWTHit ¼ market-to-book ratio ROAit ¼ ROE RISKit ¼ standard deviation of annual market return over the past five years Variables measuring CEO entrenchment CEOSHit ¼ level of CEO ownership DUALit ¼ dummy variable, which takes the value 1 if CEO is board chair and otherwise equals 0 CEOTENUREit ¼ Number of years as CEO in the present organization LSHit ¼ large shareholder ownership

342

EDUARDO SCHIEHLL

Control variables SIZEit ¼ natural log of total sales INDit ¼ dummy variable, which takes the value 1 if firm is a capital-based industry and otherwise equals 0

Table 4 reports the two-stage least square (2SLS) coefficients for the regression models of firm’s use of IPE in the CEO bonus plan (Model 1) and CEO incentive cash compensation (Model 2), estimated simultaneously. Note that two variables are used to measure CEO cash incentive compensation: amount of cash bonus alone – columns (1) and (3), and total cash compensation, i.e., cash bonus plus salary – columns (2) and (4).14 Columns (1) and (2) of Table 4 report the coefficients for Model 1, which regresses the presence of IPE in the CEO bonus plan on variables measuring the informativeness of the firm’s financial performance measures, board independence, and CEO duality (H1a, H1b, and H1c). As expected, estimates indicate a significant positive coefficient on firm’s growth opportunities. The coefficient on firm’s R&D intensity is positive, but statistically insignificant. This is due to potential multi-collinearity between the variables measuring R&D intensity and growth opportunities. The coefficient estimate and other robustness tests not reported in the table indicate that R&D intensity is positive and significant at conventional levels when the model is estimated without the variable measuring firm’s growth opportunities. Results for these variables offer reasonable evidence that IPE is more important in determining CEO short-term incentive payouts when investments are highly firm-specific and when the consequences of managerial actions, such as R&D expenses, are likely to have a long-term impact. As expected, the variable measuring noise in the firms’ financial measures is negative and shows a statistically significant association with the presence of IPE in the CEO bonus plan. Taken together, these results support H1a, and are similar to those documented by Bushman et al. (1996). We may therefore conclude that the presence of IPE in the CEO bonus plan is used to improve short-term contracting by reducing information asymmetry and driving managerial attention towards operational aspects of the firm’s value creation process. Columns (1) and (2) of Table 4 also report a positive and statistically significant association between the proportion of outside directors and the presence of IPE in the CEO bonus plan. Consistent with the optimal contracting hypothesis, this result supports hypothesis H1b: more independent boards tend to improve incentive contracting with CEOs by incorporating IPE to determine bonus payouts. However, the CEO duality

Table 4. 2SLS Regression of Firm’s Use of IPE in the CEO Bonus Plan and Level of CEO Incentive Compensation on the Informativeness of Financial Performance Measures (FPM), Board Independence, Ownership Structure, Economic Determinants of CEO Incentive Compensation, Firm Size, and Industry Control variables.

632456.3

50540.14 271132.3 113488.5

444022.4

Cash compensation 2SLS with IPE (4)

?

52301.81 169273.1 105171.7

127674.5 160065.6

Bonus 2SLS with IPE (3)

IPE

þ þ þ

188609.7 181558.8

Sign

Economic determinants Growth Return on assets Risk (sandard deviation stock return)

þ þ

Variables

0.035 0.219 0.031

CEO entrenchment CEO ownership CEO duality (CEO ¼ COB)

Sign

0.034 0.212 0.028

0.348 0.041

Variables

þ þ 

0.462 0.041

IPE 2SLS with Cash Compensation (2)

Informativeness of FPM Growth opportunities R&D intensity Noise in financial measures

þ 

IPE SLS with Bonus (1)

Board independence Outsiders on board CEO duality (CEO ¼ COB)

163412.2 31062.5

255058.5 201849.4

Private Performance Information in CEO Incentive Compensation

0.142 22.22 0.000 135

2653497

0.174 59.29 0.000 135

1731343

737.56 168709.6 0.145 0.034 0.024 0.244

Size Capital-based companies Constant

2795.72 100655.8 0.059 0.032 0.025 0.238

0.043

Adjusted R2 w2 p-value N

þ 

Board shareholdings Large shareholdings Size Capital-based companies

0.147

0.142 22.22 0.023 135

CEO tenure Large shareholdings

Constant

0.137 22.43 0.021 135

þ þ

Adjusted R2 w2 p-value N

significant at 0.01; significant at 0.05; significant at 0.10.

343

344

EDUARDO SCHIEHLL

coefficient is positive but not statistically significant. Although hypothesis H1c is not supported, the results in Columns (1) and (2) of Table 4 support the argument that the firm’s choice to incorporate IPE in the CEO bonus plan is a function of the informativeness of the financial performance measures (H1a) and the monitoring effort of an independent board of directors (H1b). Columns (3) and (4) of Table 4 present coefficients of CEO incentive cash compensation regressed on presence of IPE in the CEO bonus plan, economic determinants of incentive compensation, and control variables measuring CEO entrenchment. As expected, column (3) of Table 4 shows a significant negative association between CEO bonus amount and the variable measuring presence of IPE in the CEO cash bonus plan. This result supports hypothesis H2. The IPE coefficient suggests that CEOs with bonus plans incorporating IPE received about $632,000 less in cash compensation on average. This significant association corroborates the argument that nonpublicly observable measures of performance, such as IPE, help explain the level of CEO incentive compensation. A possible explanation for the negative sign is that short-term incentives may be less relevant in firms that base their compensation packages on more private, non-observable performance measures, such as IPE. In other words, when publicly available financial performance measures are less informative of managerial performance, short-term incentives may also be relatively lower. For example, a firm with greater growth opportunities and use of intangible assets might focus on long-term performance, and consequently place more emphasis on long-term incentives such as stock options (Murphy, 1999; Core et al., 1999). An alternative explanation, which is also consistent with the optimal contracting hypothesis, is that the use of private performance information such as IPE gives boards of directors more discretion in the award process. In other words, they could adjust the incentive compensation downwards to compensate for imperfections in objective performance measures. Prendergast (1999), for example, suggests that the advantage of incorporating subjective performance assessments is that it allows a more holistic picture of managerial performance, which is only partially captured by objective measures alone. Column (3) of Table 4 also shows that CEO cash compensation tends to be positively associated with firm’s growth opportunities (þ$ 52,000), and is greater when the CEO is also board chair (þ$181,558) and working in a larger firm. This is consistent with previous research on the determinants of CEO short-term incentive compensation suggesting that larger firms demand higher quality CEOs and are willing to pay for them. Results also

Private Performance Information in CEO Incentive Compensation

345

support the assertion that incentive compensation is used to mitigate monitoring difficulties associated with the executive use of private information to guide investment decisions and seize opportunities for task reallocation in order to increase rewards (Prendergast, 1999). Standard agency models suggest that the level of CEO incentive compensation is an increasing function of firm performance. Surprisingly, the coefficient in Table 4 on contemporaneous return on assets shows no statistically significant association with either measure of CEO incentive compensation. A possible explanation is that return on assets was measured over a single period, therefore failing to take into account average past performance. Although firm risk is a potentially important determinant of CEO compensation level, Table 4 shows that the coefficient of standard deviation of stock return is not statistically significant. This is however consistent with other empirical research on compensation, such as Core et al. (1999), who document no significant association between level of CEO compensation and two measures of firm risk. Furthermore, theoretical models developed by, e.g., Banker and Datar (1989), suggest that incentive compensation may either increase or decrease with firm risk, which may justify the lack of clear association in the empirical investigation. These coefficients also strongly support previous empirical research provided by Murphy (1999) and Magnan, St-Onge, and Calloc’h (2000) suggesting that Canadian CEO compensation contracts are sub-optimally designed, since they inadequately reflect firm performance and risk. In fact, ownership concentration and CEO-dominated boards are potential explanations for sub-optimally designed incentive contracts (Barkema & Gomez-Mejia, 1998), which was a motivation for the present investigation using a sample of Canadian firms. The next section investigates whether the presence of IPE in the CEO bonus plan is associated with firm future performance.

4.2. IPE in the CEO Bonus Plan and Firm Future Performance From the earlier evidence, the question arises as to whether CEO incentive cash compensation related to the presence of IPE in the CEO bonus plan is associated with increased shareholder wealth. Accordingly, the following question is examined: whether the CEO cash incentive compensation level that is unexplained by the firm’s observable measures of performance and board independence, which I label, predicted excess compensation, is associated with firm future performance (H3).

346

EDUARDO SCHIEHLL

Following a similar approach to Core et al. (1999) and Hayes and Schaefer (2000), excess compensation is computed using the regression coefficients of CEO incentive cash compensation on its economic and governance determinants, as in Model 2. However, as this investigation is interested in the portion of the CEO incentive compensation that might be explained by the use of IPE, I re-estimated Model 2 without the independent IPE variable. The estimated coefficients of Model 2 were then used in conjunction with each firm’s economic determinants and governance attributes to predict excess CEO incentive compensation. As such, predicted excess incentive compensation is computed by the difference between actual and predicted CEO incentive compensation. I interpret predicted excess compensation as the portion of the CEO cash incentive compensation due to the use of IPE in the CEO incentive award process. To distinguish between the firm’s future accounting and market performance, both return on book value of equity and MKT were measured one (tþ1) and three years (tþ3) after the cash bonus was awarded. Hence, the regression equations used to investigate the association between predicted excess compensation and firm future performance were specified as follows: ROEtþ1 ¼ f ðPredicted excess compensation þ control variablesÞ (Model 3) ROEtþ1 ¼ a0 þ b3 ðPredicted excess compensationÞ þ b2 STDROE þ b3 SALE þ b4 IND þ  where ROEitþ1

¼ return on equity for the fiscal year subsequent to the cash bonus award STDROEit1 ¼ standard deviation of annual ROE over the past five years SALEit1 ¼ natural log of sales INDit ¼ industry indicator variable and MKTitþ1 ¼ f ðPredicted excess compensation þ control variablesÞ (Model 4) MKTitþ1 ¼ a0 þ b1 ðexcess compensationÞ þ b2 STDMKTit1 þ b3 MTBit1 þ b4 LNMKCAP þ b5 IND þ 

Private Performance Information in CEO Incentive Compensation

347

where MKTitþ1

¼ market return for the fiscal year subsequent to the cash bonus award STDMKTit1 ¼ standard deviation of annual market return over the past five years MTBit1 ¼ market-to-book ratio LNMKCAPit1 ¼ natural log of market value of equity in the preceding year INDit ¼ industry indicator variable

Table 5 presents the regression coefficients for future performance on predicted excess compensation and control variables. Similar to Core et al. (1999), the variables in the above models relate to time periods as follows: firm performance one year (2001) and three years (2003) after the compensation award is regressed on predicted excess compensation, estimated based on 2000 incentive compensation data. All other variables are measured as at the end of 1999, the year prior to that used to estimate predicted excess compensation. Panel A of Table 5 presents the results with return on equity (ROE) as the dependent variable, while Panel B presents the results with MKT as the dependent variable. Two variables were used to measure predicted excess compensation, one based on the amount of cash bonus alone, or excess bonus (columns (1) and (2)), and the other based on cash bonus plus salary, or excess cash compensation (columns (3) and (4)). Both measures of predicted excess compensation were deflated by firm’s total cash compensation to correct for scale difference across observations. Column (1) of Table 5 shows a positive and statistically significant association between predicted excess bonus and return on equity at one year after the compensation award. Column (3) shows that excess cash compensation is positively associated with both measures of firm future performance in the first year after the compensation award (ROEtþ1; MKTtþ1). Columns (2) and (3) of Table 5, however, report that neither measure of predicted excess compensation is significantly associated with firm future performance measured over a lengthier time span, more specifically, three years after the compensation award (ROEtþ3; MKTtþ3). Although not reported here, neither measure of excess compensation is associated with firm performance measured two years after the compensation award. The lack of association with firm future performance measured later than one year after the compensation award may be explained by the

348

EDUARDO SCHIEHLL

Table 5.

OLS Regression of Firm Future Performance on Predicted CEO Excess Compensation and Control Variables.

Panel A: Firm’s return on equity (ROE) at one and three years after CEO incentive cash compensation award Variables

Sign

Excess bonus Excess cash compensation Standard deviation of ROE Sales

þ þ  þ

2

Adjusted R F N

ROE

tþ1

(1)

ROE

0.082 1.540 0.000

tþ3

(2)

0.097 0.236 0.001

0.619 34.38 135

0.229 6.10 105

ROE

tþ1

(3)

0.056 1.636 0.000 0.669 41.10 135

ROE

tþ3

(4)

0.053 0.165 0.002 0.233 6.12 105

Panel B: Firm’s stock price return (MKT) at one and three years after CEO incentive cash compensation award Variables Excess bonus Excess cash compensation Standard deviation of market return Growth opportunities Market capitalization Adjusted R2 F N

Sign MKTtþ1 (1) MKTtþ3 (2) MKTtþ1 (3) MKT ? ? þ þ þ

0.055 1.149

0.090

0.000 0.012

0.126 0.006 0.013

0.375 11.54 135

0.003 0.96 105

0.101 1.139 0.002 0.016 0.385 11.63 135

tþ3

(4)

0.053 0.134 0.007 0.015 0.011 0.84 105

Note: White heteroskedasticity – consistent standard errors. Significant at 0.01; significant at 0.05; significant at 0.10.

fact that cash bonus plans are by definition short-term performance incentives. To sum up, the estimated coefficients reported in Table 5 for the predictable component of the CEO incentive compensation due to nonobservable aspects of firm performance, such as IPE, show a significant positive association with firm future performance one year after the compensation award, but not with long-term future performance. These results support hypothesis H3. Although the association is limited to one year after the bonus award, results support the argument that IPE is used in incentive compensation to reward CEOs for actions that improve the firm’s future performance, but which are not reflected in the firm’s contemporaneous, observable financial performance. The link between the

Private Performance Information in CEO Incentive Compensation

349

predicted excess component of the CEO incentive compensation and the firm’s future performance implies that the use of IPE in the CEO cash bonus plans follows the optimal contracting hypothesis. IPE appears to be used to increase the informativeness on CEO actions (H1) and to correctly determine the level of current CEO cash incentive compensation (H2). This finding bears directly on the current governance debate as to whether boards use all available mechanisms to monitor executive actions. It is also consistent with Hayes and Schaefer (2000), who show that executive compensation tends to be a better predictor of firm future performance when unobservable measures of performance are given greater weight in the incentive contract.

5. SUMMARY AND CONCLUSION Theoretical research based on the optimal contracting hypothesis suggests that incentive contracts benefit from the inclusion of performance information that is observable only to the contracting parties, such as IPE. The premise is that privately observable performance measures may be used to decrease information asymmetry and mitigate incentive distortions caused by the lack of informativeness of traditional accounting and marketbased performance measures of executive contribution to firm performance. From a governance perspective, however, the use of information that is unobservable to outsiders, and by definition unverifiable, creates a constraint. The board of directors can use such private information for ex post rationalizations instead of optimal contracting. This paper therefore examined whether the presence of IPE in the CEO cash bonus plan follows the optimal contracting hypothesis. Three issues were addressed. First, whether the presence of IPE in the CEO bonus plan was associated with proxies of the informativeness of the firm’s financial performance measures and board independence. Consistent with prior research on the determinants of performance measure choices in incentive contracting, results show that the use of IPE in the CEO bonus plans is an increasing function of the firm’s investment opportunities and the informative precision of traditional financial information on managerial actions. With respect to board independence, a significant positive association is found between the use of IPE and the proportion of outside directors on the board. This corroborates the argument for the monitoring effectiveness of independent boards and implies that outside directors tend to implicitly monitor managerial decisions by improving governance and incentive

350

EDUARDO SCHIEHLL

mechanisms. Second, to further explore the argument that boards of directors incorporate IPE in the CEO bonus plan to account for decisions that are not reflected by financial performance measures, the association between the presence of IPE in the bonus plan and the CEO incentive compensation was investigated. The empirical evidence suggests that the use of IPE is significantly and negatively associated with the CEO cash incentive compensation. Results support the ability of private (non-verifiable to outsiders) performance measures to explain CEO incentive compensation levels. This negative association may be explained by the fact that the presence of IPE in the CEO bonus plan allows boards of directors to better align performance-contingent payouts with managerial performance. An alternative explanation is that, in firms where publicly observable financial measures of performance are less informative of managerial actions, and therefore less useful as contracting instruments, cash bonus payoffs may be relatively lower. For example, higher growth opportunities and R&D expenses are attributes of knowledge-based firms and short life-cycle products, with consequently less emphasis on short-term incentives. Finally, this study demonstrates that the predictable incentive component of the CEO incentive compensation due to the presence of IPE in the CEO bonus plan is positively associated with the firm’s future performance one year after the compensation was awarded. This suggests that IPE is incorporated in the CEO bonus plan to reward CEOs for actions that are not reflected in the firm’s contemporaneous financial performance, but which nevertheless improve future performance. The findings of this study reasonably support the economically efficient use of private performance measures such as IPE (optimal contracting hypothesis). IPE measures appear to be used by boards of directors to award CEOs for actions not reflected in traditional contemporaneous financial measures of firm performance. These results have implications for the understanding of incentive mechanisms, in particular the nature of the information that boards of directors use to reward top managers. Moreover, unlike prior research on the determinants of performance measure choices, this study contributes by documenting the ability of non-publicly observable performance measures to explain executive incentive compensation levels and firm future performance. Contrary evidence is uncovered to fuel the heated discussion on whether boards of directors have delegated the tasks of monitoring and rewarding executives to capital markets as a by-product of the extensive use of equity-based compensation. According to the study’s empirical evidence, boards of directors instead appear to oversee executive actions by collecting and evaluating performance-related information that is not available to outsiders.

Private Performance Information in CEO Incentive Compensation

351

As in any empirical investigation, certain limitations must be acknowledged. The primary caveat is the lack of information on the specific weights applied to the IPE criteria when the firms computed their CEO bonus payouts. As mentioned earlier, public and private performance information is used to complement the incentive contracting process. Knowledge of the weights given to the performance measure mix in the CEO bonus plan would certainly improve the capacity to distinguish the proportion of cash compensation that is based on privately observable performance measures. Along the same line, reliable information about the composition of the firms’ compensation committees would certainly enhance the analysis of the association between the presence of IPE in the CEO bonus plan and board independence. In addition, a certain ambiguity in the compensation disclosure requirements may well have introduced noise into the compensation data and the variable measuring the presence of IPE in the CEO bonus plan. Future research initiatives in this field could attempt to overcome these limitations by using surveys or private consulting data on executive compensation practices. Furthermore, given the fact that compensation disclosure in Canada is a relatively recent phenomenon, this study investigated the presence of IPE as one component among other traditional performance measures in incentive contracting, not the determinants of its adoption. Future research could aim to overcome this limitation and pursue the investigation by examining what motivates firms to include or change the relative weight of privately observable performance measures in the CEO bonus plan.

NOTES 1. The balanced scorecard approach of performance evaluation, as proposed by Kaplan and Norton (2001), is an example of a scorecard-based reward system intended to drive managerial attentions to multiple dimensions of the firm value creation process. 2. Business history is littered with examples of firms that have rewarded the wrong performance. At the H. J. Heinz Company, division managers received bonuses only if earnings increased over the prior year. Managers delivered consistent earnings growth by manipulating the timing of shipments to customers and preparing for services not yet received. In each of these cases, executives took actions to increase their compensation, but these actions were seemingly at the expense of long-run firm value (Baker et al., 1994). 3. Bushman et al. (1996), for example, examine the use of IPE versus traditional accounting earnings and stock price measures of corporate performance. Keating (1997) addresses division versus firm accounting performance measures, while Ittner et al. (1997), Banker et al. (2000), and Ittner and Larcker (2002) distinguish between financial (accounting) measures and non-financial (operational) performance measures.

352

EDUARDO SCHIEHLL

4. Although executive stock and stock option holdings may provide a strong incentive, this study focused strictly on the CEO cash bonus plan. Cash bonus represents the incentive compensation awarded to the CEO, based on the board’s ex ante choices of firm and/or individual performance criteria. As such, the cash bonus plan represents a relevant governance mechanism to collect and evaluate incremental, performance-related information that is not available to outsiders. Moreover, information on performance criteria linked to the executive’s target cash bonus is included in the compensation committee’s ‘‘Report on Executive Compensation’’ (see examples in the Appendix), while information on performance criteria linked to shares and stock option grants are very rarely disclosed. Hence, neither this issue nor the issue of the optimal choice of compensation mix was addressed, both interesting questions that are beyond the scope of this chapter. 5. Stock options were valued at 25% of the exercise price at the time of grant. Murphy (1999) raises some issues related to the evaluation of stock options granted to executives (distinguishing between the cost of options to the firm and the value to executives) and the fact that there is no recognized valuation methodology. Lambert, Larcker, and Weigelt (1993) state that evaluating options at 25% of their exercise price generates values similar to those obtained with more sophisticated evaluation models. Here, the stock option valuation method used by Core et al. (1999) was adopted. 6. The data collected on variable IPE and other performance measure categories used in incentive contracts shows that ROE and return on assets were the most frequently used financial measures by firms to compute the CEO cash bonus payout. 7. The Canadian accounting rules for R&D are contained in the CICA Handbook, Section 3450, and essentially mirror those of IAS 38. Research is defined as ‘‘original and planned investigation undertaken with the prospect of gaining new scientific or technical knowledge and understanding,’’ while development is the ‘‘application of research findings or other knowledge to a plan or design for the production of new or substantially improved materials, devices, products, processes, systems or services prior to the commencement of commercial production or use.’’ All costs engaged in a research phase must be expensed immediately. In contrast, development cost should be capitalized if, and only if, a company can demonstrate that a specific development cost satisfies six specific conditions. Meeting all six conditions at once is very difficult, which results in a relative low level of capitalization of R&D expenses in Canada (see, e.g., the study by Ding et al., 2004). 8. Fisher’s z-score is a transformation of Pearson’s r correlation coefficient. It is obtained by dividing the correlation plus 1 by the same correlation minus 1, then taking the natural log of the result and dividing by 2. Fisher’s transformation reduces skew and makes the sampling scores distribution more normal as sample size increases (Gujarati, 2003). Transforming Pearson’s correlation coefficients into a z-score is primarily used for hypothesis testing, but is also recommended when correlation coefficients are used as independent variables, as in this study. 9. The SEDAR is a Website that has been operated by The Canadian Securities Administration since 1997. It provides filings of publicly available documents for all Canadian public companies in order to facilitate the electronic filing of securities information, as required by Canadian securities regulatory agencies. 10. Bill 198, introduced in 2002 by the Ontario Securities Commission (OSC) in response to the reforms taking place in the U.S. under the Sarbanes–Oxley Act,

Private Performance Information in CEO Incentive Compensation

353

granted boards of directors additional responsibilities and legally strengthened their monitoring role. 11. Corporate Governance Guidelines: definition used by the Toronto Stock Exchange (TSE, 1994). 12. Canadian securities regulations require disclosure of significant shareholders, both individual and institutional, which beneficially own or exercise control over at least 10% of the firm’s outstanding shares. Therefore, only the identities of large shareholders are disclosed in the firms’ Management Proxy Circulars. If no large shareholder is indicated, the ownership structure is considered diffuse, according to Canadian standards. 13. The correlation matrix allows a first scrutiny for potential multi-collinearity in the regression analysis. Although several governance variables are significantly correlated, coefficients are low. The exception is the correlation between CEO ownership, board shareholding, and the dummies for the large shareholders identity (family and institutional), which was taken into account in the model specifications. Furthermore, results of multi-collinearity and robustness tests do not appear to upset results reported in Tables 4 and 5. 14. Note that, following most prior empirical research in this area, board and ownership structures are treated as exogenous, whereas economic theory argues for endogeny. In addition, economic theory suggests that compensation is simultaneously determined by board and ownership structures. To the extent that this holds true in practice, I acknowledge that some inferences based on Table 4 may be affected by a simultaneous equations bias. However, the focus of this study is the presence of IPE and its ability to explain CEO incentive compensation. This simultaneity is taken into account in Table 4 by the simultaneous estimation (2SLS) of the two models.

ACKNOWLEDGMENTS Financial support from the Canadian Accounting Academic Association (CAAA) and the Society of Management Accountants of Canada (CMA) is gratefully acknowledged. Also acknowledged are the helpful comments of Heather Wier (University of Alberta), Amin Mawani (York University), Howard Thomas (Warwick Business School), Gilbert Laporte, Pascal Franc- ois (HEC Montreal), and participants of 4th EIASM Conference on Performance Measurement and Management Control.

REFERENCES Aggarwal, R., & Samwick, A. (2003). Performance incentives within firms: The effect of managerial responsibility. The Journal of Finance, LVIII(4), 1613–1649. Baker, G., Gibbons, R., & Murphy, K. J. (1994). Subjective performance measures in optimal incentive contracts. Quarterly Journal of Economics, 109(4), 1124–1156.

354

EDUARDO SCHIEHLL

Banker, R. D., & Datar, S. M. (1989). Sensitivity, precision, and linear aggregation of signals for performance evaluation. Journal of Accounting Research (27), 21–39. Banker, R. D., Potter, G., & Srinivasan, D. (2000). An empirical investigation of an incentive plan that includes nonfinancial performance measures. The Accounting Review, 75(1), 65–92. Barkema, H. G., & Gomez-Mejia, L. R. (1998). Managerial compensation and firm performance: A general research framework. Academy of Management Journal, 41(2), 135–145. Bellavance, F., & Schiehll, E. (2006). The complementarities between Governance Structure and Incentive Contracting. Working paper, p. 36. HEC Montreal, Montreal, Quebec. Bushman, R. M., Indjejikian, R. J., & Smith, A. (1996). CEO compensation: The role of individual performance evaluation. Journal of Accounting and Economics, 21, 161–193. Bushman, R. M., & Smith, A. J. (2001). Financial accounting information and corporate governance. Journal of Accounting and Economics, 32, 237–333. Core, J. E., Holthausen, R. W., & Larcker, D. (1999). Corporate governance, chief executive officer compensation, and firm performance. Journal of Financial Economics, 51(3), 371–406. Craighead, J. A., Magnan, M. L., & Thorne, L. (2004). The impact of mandated disclosure on performance-based CEO compensation. Contemporary Accounting Research, 21(2), 369–398. Ding, Y., Entwistle, G., & Stolowy, H. (2004). International differences in R&D disclosure practices: Evidence in a French and Canadian context. Advances in International Accounting, 17, 55–72. Feltham, G. A., & Xie, J. Z. (1994). Voluntary financial disclosure in and entry game with continua of types. Contemporary Accounting Research, 9(3), 46–80. Fisher, J. G., Maines, L. A., Peffer, S. A., & Sprinkle, G. B. (2005). An experimental investigation of employer discretion in employee performance evaluation and compensation. The Accounting Review, 80(2), 563–583. Gujarati, D. N. (2003). Basic econometrics. New York: McGraw-Hill/Irwin. Hayes, R. M., & Schaefer, S. (2000). Implicit contracts and the explanatory power of top executive compensation for future performance. RAND Journal of Economics, 31(2), 273–293. Holmstrom, B. R. (1979). Moral hazard and observability. Bell Journal of Economics, 10, 74–91. Ittner, C. D., & Larcker, D. F. (2002). Determinants of performance measure choices in worker incentive plans. Journal of Labor Economics, 20(2), 58–90. Ittner, C. D., Larcker, D. F., & Rajan, M. V. (1997). The choice of performance measures in annual bonus contracts. The Accounting Review, 72(2), 231–255. Kaplan, R. S., & Norton, D. P. (2001). The strategy-focused organization: How balanced scorecard companies thrive in the new business environment. Boston, MA: Harvard School Business Press. Keating, S. A. (1997). Determinants of divisional performance evaluation practices. Journal of Accounting and Economics, 24, 243–273. Lambert, R. A., Larcker, D. F., & Weigelt, K. (1993). The structure of organizational incentives. Administrative Science Quarterly, 38(3), 438–461. Magnan, M., St-Onge, S., & Calloc’h, Y. (2000). The impact of directors’ power on the use of performance-contingent compensation for CEOs. Working paper, p. 34. HEC Montreal, Montreal, Quebec. Murphy, K. J. (1999). Executive compensation. In: O. Ashenfelter & D. Card (Eds), Handbook of Labor Economics (3rd ed.). Amsterdam.

Private Performance Information in CEO Incentive Compensation

355

Murphy, K. J., & Oyer, P. (2001). Discretion in executive incentive contracts: Theory and evidence. Working paper, p. 48. Marshall School of Business, University of Southern California, California. Prendergast, C. (1999). The provision of incentives in firms. Journal of Economic Literature, 37(1), 7–63. Sloan, R. G. (2001). Financial accounting and corporate governance: A discussion. Journal of Accounting and Economics, 32, 335–347. Tosi, H. L., Werner, S., Katz, J., & Gomez-Mejia, L. R. (1998). How much does performance matter? A meta-analysis of CEO pay studies. Journal of Management, 26(2), 301–324. TSE – Toronto Stock Exchange Committee on Corporate Governance in Canada. (1994). Where were the Directors? The Dey Report. Toronto, Canada.

APPENDIX Biomira Inc. – Management Proxy Circular, 1999 (p. 3) Annual Incentive Compensation: The committee has sought to provide annual incentive compensation for certain executive officers through bonus arrangements. Awards are contingent upon the achievement of corporate and individual objectives. Consumer Packaging Inc. – Report on Compensation (p. 5) Senior managers and executive officers are eligible for annual incentive awards under the Management Incentive Plan. Each participant’s award is expressed as a percentage of base salary which increases with level of responsibility. Actual awards are determined on the basis of a combination of factors including the achievement of the profit plan (specifically, operating income, and cash flow) as well as meeting budgets and individual performance of personal objectives. Dupont Canada – Report on Executive Compensation (p. 8) Variable compensation: The committee believes that the compensation of all senior managerial and professional employees, including the CEO and NEOs, should be related to the attainment of the Corporation’s financial objectives. Annual incentives are available to NEOs subject to the achievement of pre-established threshold and target annual objectives. The criteria used to determine incentive awards include combinations of the Corporation’s ROE, Business Earnings and Cash Generated. A portion of incentives awards is based on judgmental assessment of personal objectives successfully achieved. Gesco Industries Inc. – Report on Executive Compensation The Corporation’s current compensation policy for its executive officers, including the Named Executives, emphasizes base salary and bonus incentives over other forms of compensation. A portion of the bonus incentives is tied

356

EDUARDO SCHIEHLL

into personal performance goals. Other bonus payments and other forms of compensation, including the granting of stock purchase options, are tied to the general performance of the Corporation. Compensation packages are built to reward executives for personal and Corporate achievements while ensuring these executives are treated equitably based on prevailing employment market conditions (Management Proxy Circular, 1998).