Work Sharing during the Great Recession : New Developments and Beyond [1 ed.] 9789221245643, 9789221245636

Citation preview

Work Sharing during the Great Recession

M3125 - MESSENGER TEXT.indd i

14/05/2013 13:33

M3125 - MESSENGER TEXT.indd ii

14/05/2013 13:33

Work Sharing during the Great Recession New Developments and Beyond

Edited by

Jon C. Messenger Senior Research Officer, International Labour Office, Geneva, Switzerland

Naj Ghosheh Research Officer, International Labour Office, Geneva, Switzerland

Edward Elgar Cheltenham, UK • Northampton, MA, USA

International Labour Office Geneva, Switzerland

M3125 - MESSENGER TEXT.indd iii

16/05/2013 10:05

© International Labour Organization 2013 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or otherwise without the prior permission of the publisher. Publications of the International Labour Office enjoy copyright under Protocol 2 of the Universal Copyright Convention. Nevertheless, short excerpts from them may be reproduced without authorization, on condition that the source is indicated. For rights of reproduction or translation, application should be made to ILO Publishing (Rights and Licensing), International Labour Office, CH-1211 Geneva 22, Switzerland, or by email: [email protected]. The International Labour Office welcomes such applications. The designations employed in ILO publications, which are in conformity with United Nations practice, and the presentation of material therein do not imply the expression of any opinion whatsoever on the part of the International Labour Office concerning the legal status of any country, area or territory or of its authorities, or concerning the delimitation of its frontiers. The responsibility for opinions expressed in studies and other contributions rests solely with their authors, and publication does not constitute an endorsement by the International Labour Office of the opinions expressed in them. Reference to names of firms and commercial products and processes does not imply their endorsement by the International Labour Office, and any failure to mention a particular firm, commercial product or process is not a sign of disapproval. Published by Edward Elgar Publishing Limited The Lypiatts 15 Lansdown Road Cheltenham Glos GL50 2JA UK

Edward Elgar Publishing, Inc. William Pratt House 9 Dewey Court Northampton, MA 01060 USA

In association with International Labour Office 4 route des Morillons 1211 Geneva 22 Switzerland ISBN 978-92-2-124563-6 A catalogue record for this book is available from the British Library Library of Congress Control Number: 2012952237 This book is available electronically in the ElgarOnline.com Economics Subject Collection, E-ISBN 978 1 78254 088 5

ISBN 978 1 78254 087 8 (cased) Typeset by Servis Filmsetting Ltd, Stockport, Cheshire Printed and bound in Great Britain by T.J. International Ltd, Padstow

M3125 - MESSENGER TEXT.indd iv

24/05/2013 09:02

Contents List of contributors

vii

1. An introduction to work sharing: A strategy for preserving jobs, creating new employment and improving individual well-being Jon C. Messenger and Naj Ghosheh 2. Work sharing as an alternative to layoffs: Lessons from the German experience during the crisis Lutz Bellmann, Andreas Crimmann, Hans-Dieter Gerner and Frank Wießner 3. European diversity of work sharing as a crisis measure: The experiences of Austria, Belgium, France and the Netherlands Jörg Flecker and Annika Schönauer 4. Work sharing in Japan Kazuya Ogura

1

24

72 99

5. The Turkish experience with work-sharing policy during the global economic crisis, 2008–2010 Erinç Yeldan

119

6. Results of the implementation of the suspension and partial unemployment insurance programmes in Uruguay, 2009–2010 María José González Fernández

151

7. Work sharing as a potential policy tool for creating more and better employment: A review of the evidence Lonnie Golden and Stuart Glosser

203

8. Conclusion: Lessons learned from the Great Recession and implications for policy Jon C. Messenger and Naj Ghosheh

259

Index

299

v

M3125 - MESSENGER TEXT.indd v

14/05/2013 13:33

M3125 - MESSENGER TEXT.indd vi

14/05/2013 13:33

Contributors Lutz Bellmann Chair of Labour Economics, Friedrich-AlexanderUniversity Erlangen-Nuremberg and Head of the Research Department Establishments and Employment of the Institute for Employment Research – Research Institute of the Federal Employment Agency, Nuremberg, Germany. Andreas Crimmann Senior Researcher at the Research Department Establishments and Employment of the Institute for Employment Research – Research Institute of the Federal Employment Agency, Nuremberg, Germany. Jörg Flecker Director, Forschungs- und Beratungsstelle Arbeitswelt (FORBA) Institute, Vienna, Austria. Hans-Dieter Gerner Senior Researcher at the Research Department Establishments and Employment of the Institute for Employment Research – Research Institute of the Federal Employment Agency, Nuremberg, Germany. Naj Ghosheh Co-Editor, Research Officer in the Conditions of Work and Employment Programme, International Labour Office, Geneva, Switzerland. His most recent publication is Offshoring and Working Conditions in Remote Work, with Jon C. Messenger (Palgrave Macmillan and ILO, 2010). Stuart Glosser Professor of Economics, Department of Economics, University of Wisconsin at Whitewater, Whitewater, WI, USA. Lonnie Golden Professor of Economics and Labor Studies, Department of Economics and Labor Studies, Pennsylvania State University, Abington, PA, USA. María José González Fernández Professor Adjunto, Universidad Católica Dámaso Antonio Larrañaga, Montevideo, Uruguay. Jon C. Messenger Co-Editor, Senior Research Officer with the Conditions of Work and Employment Programme, International Labour Office, Geneva, Switzerland. His most recent publications  include

vii

M3125 - MESSENGER TEXT.indd vii

14/05/2013 13:33

viii

Contributors

Offshoring and Working Conditions in Remote Work (edited with Naj Ghosheh, Palgrave Macmillan and ILO, 2010) and Working Time Around the World (with Sangheon Lee and Deirdre McCann, Routledge and ILO, 2007). Kazuya Ogura Associate Professor, School of Commerce, Waseda University, Tokyo, Japan. Annika Schönauer Researcher and PhD candidate, Forschungs- und Beratungsstelle Arbeitswelt (FORBA) Institute, Vienna, Austria. Frank Wießner Senior Researcher at the Research Department Establishments and Employment of the Institute for Employment Research – Research Institute of the Federal Employment Agency, Nuremberg, Germany. Erinç Yeldan Dean of the School of Economics, Department of Economics, Yasar University, Izmir, Turkey.

M3125 - MESSENGER TEXT.indd viii

14/05/2013 13:33

1.

An introduction to work sharing: A strategy for preserving jobs, creating new employment and improving individual well-being Jon C. Messenger and Naj Ghosheh

‘[S]hort-time work’ (Kurzarbeit) is fast becoming an international word . . . In 2009, our economy contracted by almost 5 per cent because we are so heavily dependent upon exports. And as our country basically does not have any natural resources, we had to pause and consider what it is that makes Germany strong . . . And it was clear that, on the one hand, our strength lies in our employers – not only our large companies, but also our small and medium-sized enterprises – and on the other hand, in our workers, who have often been with a company for a long period of time. We understood that it was vital to keep this bond between employers and employees during difficult times [and] we therefore introduced short-time work as an instrument proposed by the State . . . The good thing is that it all turned out well, simply because the companies believed that they had a future, and because the State was ready and willing to actually invest quite a lot of money into the programme. (Address by Her Excellency Ms Angela Merkel, Chancellor, Federal Republic of Germany, to the 2011 International Labour Conference in Geneva, Switzerland)

1.

THE GREAT RECESSION OF 2008–09: AN ECONOMIC ‘SHOCK WAVE’

The global financial crisis that first hit the United States (which was already in a recession) in the autumn of 2008, and then the rest of the world shortly thereafter, can be likened to a ‘shock wave’. A massive wave of reduced liquidity in the financial sector unleashed by the dramatic collapse of Lehman Brothers generated a negative macroeconomic shock which drove down aggregate demand in countries around the world to a degree unseen since the Great Depression: credit markets froze, bringing lending to a virtual standstill, halting business investment, and driving consumer spending into a downward spiral. Thus, the economic ‘shock 1

M3125 - MESSENGER TEXT.indd 1

14/05/2013 13:33

2

Work sharing during the Great Recession

wave’ emanating from the financial sector had a profound effect on the ‘real’ global economy as well: for example, 28 of the 30 OECD countries were hit by a recession, and that recession was greater than the historical average in the depth of its impact on GDP in all but five OECD countries, the Czech Republic, Greece, the Republic of Korea, New Zealand and Switzerland (OECD 2010: 33). And the impact on GDP was not limited to the developed world, but stretched around the globe from China to Latin America – turning the global financial crisis into the Great Recession of 2008–09. National governments responded to this economic shock wave with a variety of economic policies designed to avoid a potential cataclysm. First and foremost, many national governments deployed Keynesian macroeconomic fiscal ‘stimulus packages’ (and often monetary ‘stimulus’ as well, through both interest rate cuts and so-called ‘quantitative easing’) to bolster aggregate demand. Governments also developed or expanded a variety of labour market policies, ranging from unemployment insurance to skill training and a variety of other types of employment assistance. These last policies were typically based on the assumption that unemployment rates would follow historical patterns and move more or less in tandem with the decline in GDP. Yet, curiously, the response of unemployment in many countries did not mirror the change in GDP, as would normally be expected. Indeed, some countries experienced much greater increases in unemployment than would normally have occurred given the percentage drop in GDP – notably Spain and the United States, but also Canada, Greece, Ireland and New Zealand – while in European countries such as Belgium, Finland, Germany, Italy and the Netherlands, and in some Asian countries, notably the Republic of Korea and Japan, changes in unemployment rates were relatively small compared with the decline in GDP and also compared with previous recessions. In Spain, for example, a GDP decline of approximately 4 per cent in 2008–09 resulted in nearly a 10 per cent increase in the unemployment rate. In stark contrast, in Germany GDP dropped more than 6 per cent in 2008–09, but the unemployment rate over the period actually declined (ibid.: 33–4). Why did this happen? A number of recent studies have concluded that these seemingly incongruous results regarding unemployment during the Great Recession stem in large part from the extent to which companies in different countries chose to respond to output declines mainly by reducing employment levels, as in Spain, or alternatively, to respond primarily by reducing per capita working hours, as in Germany (for example, Arpaia et al. 2010; Crimmann et al., 2010; OECD 2010; Hijzen and Venn 2011). Also, the decision of many companies to reduce hours of work across the board rather than

M3125 - MESSENGER TEXT.indd 2

14/05/2013 13:33

An introduction to work sharing

3

resorting to layoffs was the result of another, very different type of labour market policy developed and implemented by governments to respond to the economic shock wave. This type of policy is called Kurzarbeit (‘short-time work’) in Germany – and it is indeed, as Chancellor Merkel stated, ‘fast becoming an international word’. Germany’s Kurzarbeit was in fact merely the largest and the best known of many similar measures developed and implemented by national (or subnational) governments in a broad range of countries to promote reduced working hours in lieu of cutting jobs. These policy measures, which are known by a variety of different names in different countries, can collectively be referred to as ‘work sharing’. This chapter will provide a general background and overview of what is meant by work sharing that will inform the rest of this volume. Section 2 provides the definitions and objectives of work sharing as an employment policy tool. Section 3 identifies and defines the theoretical underpinnings of work sharing, including the principle that work sharing can be used not only to preserve or increase employment, but also to improve individual and societal well-being. Section 4 provides a brief history of work sharing in different countries during different periods of the twentieth century. Section 5 then explains why work sharing emerged during the Great Recession in both developed countries and, for the first time, some developing countries as well. Section 6 concludes the chapter by setting out what the rest of the volume will cover, namely what work-sharing policies have been used in different countries during the Great Recession and what the future holds for work-sharing policies more generally.

2.

WHAT IS WORK SHARING? DEFINITIONS AND OBJECTIVES

Work-sharing measures can take two distinct forms. The first form, of which Germany’s Kurzarbeit is perhaps the quintessential example, can be defined as a labour market instrument based on the reduction of working time intended to spread a reduced volume of work over the same (or a similar) number of workers in order to avoid layoffs.1 In the framework of national work-sharing programmes, enterprises can receive benefits when they refrain from laying off workers, and instead ‘share’ the lower volume of available work by reducing the working hours of either all their employees or all members of a specific work unit. In times of economic crisis, the use of such temporary, ‘crisis work-sharing measures’ as a labour market policy tool aimed at preserving existing jobs during a cyclical downturn not only helps to avoid mass layoffs and increased unemployment, but

M3125 - MESSENGER TEXT.indd 3

14/05/2013 13:33

4

Work sharing during the Great Recession

also allows businesses to retain their skilled workforces, thus minimizing firing and (re)hiring costs, preserving functioning plants, and bolstering staff morale during difficult times. If crisis work-sharing measures are properly designed and implemented, the result can be a ‘win–win–win’ solution: enabling workers to keep their jobs and even to prepare for the future; assisting companies not only to survive an economic crisis, but to be well-positioned to prosper when growth returns; and minimizing the costs of unemployment and, ultimately, social exclusion for governments and society as a whole. This temporary, crisis form of work sharing is the primary focus of this volume and also of most of its individual contributions (see the last section of this chapter for details). However, there is a second distinct form of work sharing as well. This can be defined as follows: a government policy that is designed to induce permanent downward adjustments in average working hours for the purpose of encouraging additional hiring and thus increasing the level of employment. As will be discussed later in this chapter, such permanent working-time reductions can be achieved via different mixes of policies, including – but certainly not limited to – legally mandated reductions in the normal or standard workweek in a country. This second form of work sharing remains a subject of often contentious debate among economists and policymakers, as will be discussed in Section 3. None the less, the dramatic expansion of crisis work-sharing measures during the Great Recession, and the growing evidence of their success in preserving jobs in many countries by reducing per capita working hours, has sparked renewed interest in exploring the potential of permanent reductions in working hours – the other form of work sharing – as a policy tool designed to create new employment. This latter form of ‘permanent’ work sharing, if properly implemented, not only holds the potential to increase the job content of any particular level of output, but as we shall see later, also offers the possibility of improving individual well-being, thus contributing to more sustainable economies and societies. While this permanent form of work sharing is a secondary focus of this volume, it is the primary focus of Chapter 7, ‘Work sharing as a potential policy tool for creating more and better employment: A review of the evidence’ (Golden and Glosser). Finally, it is worth noting that work sharing should not be confused with job sharing (which is a very common mistake). Job sharing refers to an individual arrangement whereby two persons – each of whom is working part-time – decide to share a single, full-time job. Job sharing and other similar forms of ‘time sharing’ among individual workers are (unlike work sharing) individual rather than collective working-time arrangements, and such arrangements are not discussed in this volume.

M3125 - MESSENGER TEXT.indd 4

14/05/2013 13:33

An introduction to work sharing

3.

THE THEORETICAL UNDERPINNINGS OF WORK SHARING

3.1

Work Sharing as a Job Preservation Tool (Crisis Work-sharing Measures)

5

Crisis work-sharing measures designed to preserve existing jobs encourage companies to respond to a reduction in demand for their products or services by spreading the reduced volume of work over all workers in the enterprise or all workers in a highly impacted work unit, rather than laying off a portion of those workers. This type of ‘sharing’ of the available work is achieved by reducing the number of hours worked by each affected employee.2 For example, with a 20 per cent reduction in demand for its products or services, instead of laying off 20 per cent of its workforce, a company could instead reduce the average hours worked by all of its employees by approximately 20 per cent across the board. Under this scenario, if workers were previously working a normal 40-hour workweek, with work sharing they would be working 32 hours per week. This reduction in working hours is typically accompanied by a corresponding (pro rata) reduction in employees’ wages or salaries. In this way, the company would be able to cut its labour costs in line with the reduction in demand for its products or services – but without the need to resort to layoffs. Of course, the cuts in wages which typically accompany crisis work-sharing measures involve some hardship for the firms’ employees – although not as serious a hardship as losing a job for those who would otherwise be impacted by layoffs. For this reason, most of those countries that have developed crisis work-sharing programmes provide affected workers with some type of a wage supplement giving partial replacement of the employees’ lost earnings, generally around one-half or slightly more of the amount by which their salaries have been reduced. These wage supplements are public subsidies which often, but not always, take the form of partial unemployment insurance (UI) payments during the period of work sharing; in some countries, these partial UI payments are called ‘shorttime compensation’. Most crisis work-sharing measures establish specific time limits on the work-sharing period. Such limits help to ensure that the work-sharing measure is a temporary measure in response to an economic crisis or another short-term reduction in demand, and not a long-term reduction in hours and pay. Making such measures temporary also helps to minimize ‘deadweight’ – the amount of public work-sharing subsidies received by firms who would not have engaged in layoffs in any case – as well as limiting any potential displacement effects – that is, a crowding out of emerging

M3125 - MESSENGER TEXT.indd 5

14/05/2013 13:33

6

Work sharing during the Great Recession

businesses and industries by existing, inefficient ones – which could potentially result from the public subsidies provided to work-sharing firms. Finally, while not an integral element of work sharing, many worksharing measures include links to training or retraining activities for participating workers. The idea of such linkages is to help affected workers to upgrade their skills during the period of work reduction, so that either they can be more productive when demand for the company’s products or services rebounds, or should that fail to occur, those workers would at least be better prepared to move on to new jobs during an economic recovery. Unfortunately, as we shall see, the training and retraining components of work-sharing measures have proven to be some of the most problematic and least effective elements of these programmes. The main focus of this volume is on the use of such crisis work-sharing measures as a job preservation tool and their effects during the Great Recession of 2008–09 and its aftermath. Chapters 2–6 focus primarily (although not always exclusively) on such crisis work-sharing measures. 3.2

Work-sharing Policies to Increase Employment (‘Permanent’ Work Sharing)

Work-sharing policies designed to create new employment also involve a downward adjustment of hours of work, and the basic logic underlying such policies is the same: spreading any given volume of work in a company over a larger number of workers. In this case, however, the reduction of working hours is not a temporary policy in response to a decline in the demand for a firm’s products or services, but rather is a permanent reduction in working hours designed to encourage hiring and thus increase the overall level of employment. Such permanent reductions in average hours of work can be achieved by different methods, ranging from legally mandated or collectively bargained reductions in the normal or standard working week (the ‘classic’ and most studied approach) to the use of tax incentives (for example, reduced payroll taxes, tax credits) provided to employers who reduce the normal workweek of employees in their companies on a weekly, monthly, or even an annual basis (see, for example, the US tax credit proposal in Baker 2009), to the promotion of reduced actual working hours via part-time or part-year work (as in the Netherlands)¸ or some combination of policies. Regardless of the specific method used, the objective of this form of work sharing is the same: to reduce average hours of work per worker in order to promote hiring and thus increase total employment. For example, reducing average hours of work by 10 per cent could theoretically result in a 10 per cent increase in new job vacancies.

M3125 - MESSENGER TEXT.indd 6

14/05/2013 13:33

An introduction to work sharing

7

In reality, however, it is highly unlikely that such a one-for-one conversion of permanent hours reductions into new job creation would occur, although the expected employment effects of such reductions depend considerably on the theoretical framework. For example, the neoclassical economic view is that output must inevitably decline as a result of such a reduction in working hours (especially the standard legal workweek), thus ensuring that permanent working-time reductions cannot create new jobs and might actually decrease employment levels. In contrast, the Keynesian perspective assumes that total spending and output would not decrease with permanent working-time reductions, although the composition of the output (and hence of employment) might shift across sectors. And a range of empirical studies of the actual employment effects of work-sharing policies show a wide range of results – ranging from substantially positive to neutral to moderately negative effects on employment (see Golden and Glosser, ch. 7 in this volume for a review of this literature). More concretely, the size of the employment effects of a particular work-sharing policy depends upon both the direct effects of the policy, plus a number of indirect effects. These indirect effects are factors which are likely to reduce the employment-creation effects of permanent worksharing policies, and which, taken together, mean that these effects will likely be substantially less than the maximum potential in theory. First, some employers might find it in their interest to schedule a workweek beyond the new, reduced normal workweek for at least some workers; this phenomenon has been referred to as the ‘overtime leakage’ effect. Second, a ‘productivity effect’ may undermine the effort to expand job creation – that is, a shortened workweek might help reverse the ‘vicious circle’ of longer hours to achieve greater production or productivity per worker but at the expense of lower per hour productivity, and turn it into a virtuous cycle of shortened workweeks and improved productivity per hour – and thus, on balance, lower unit labour costs (see below). Third, workers are not perfect substitutes for each other – referred to in academic literature as ‘indivisibility of labour’ – and, therefore, the existence of skill shortages in industries or occupations where hours are reduced can act as a serious bottleneck to filling any new job openings created by this form of work sharing (see, for example, Freeman 1998). This issue should not be confused with the so-called ‘lump-of-labour fallacy’: the concept that there is a fixed amount of work in an economy,3 and the further assumption that work sharing somehow involves parcelling out this ‘fixed amount of work’ (see Walker 2007, for a historical review of the lump-of-labour fallacy and a refutation of its theoretical arguments4). Finally and perhaps most importantly, is the issue of real labour costs per hour – that is ‘unit labour costs’. Most microeconomic studies of

M3125 - MESSENGER TEXT.indd 7

14/05/2013 13:33

8

Work sharing during the Great Recession

permanent work-sharing measures emphasize the importance of labour costs. The neoclassical economic view is that with permanently reduced hours, labour costs are likely to increase, particularly if hourly wages are not reduced in line with the reduction in working hours; this will, in turn, encourage firms to substitute capital for labour, thus undermining any potential employment creation (see, for example, Hunt 1998, 1999; Kapteyn et al. 2000). In contrast, in their seminal article on working-time reduction and employment, Bosch and Lehndorff (2001: 211) emphasize that ‘the decisive factor for firms is unit costs, not absolute wage costs’. If hours of work are reduced but total earnings are held constant, then ceteris paribus, a firm’s unit labour costs will increase; if that occurs, firms may, in turn, be forced to raise the price of their goods or services, which may reduce the demand for them. The problem for employers would be the most severe if there were no reduction in earnings accompanying the permanently reduced hours of work. If, on the other hand, earnings (more correctly, total payroll expenses) were reduced proportionally to the reduction in hours (as is typically the case with work sharing when it is used as a job-preservation tool), or if wage increases were phased in over a period of time simultaneously with reductions in working hours, then much of the concern regarding higher unit labour costs due to this form of work sharing would be unwarranted. Alternatively, even if total earnings remained constant, the potential increases in labour cost per unit can be countered or overcome with public subsidies in the short term, such as by reducing payroll tax rates. Work-sharing policies designed to increase employment via permanent reductions in working hours, including both the theoretical and empirical literature regarding the potential employment-creating effects of this form of work sharing, will be reviewed and analysed in depth in Chapter 7 (Golden and Glosser). 3.3

Work Sharing as Contributor to Improved Individual and Societal Well-being

Interestingly, working time is also considered to be an important contributor to individual well-being in the burgeoning new literature regarding the economics of ‘happiness’. The happiness literature uses an individual’s ‘reported subjective well-being as a proxy measure for individual welfare’ (Frey and Stutzer 2010: 3). In this literature, unemployment is universally found to be detrimental to happiness, not only due to reduced income, but also because paid work provides people with a sense of identity and opportunities for socialization, while ceteris paribus more work hours are positively associated with reported levels of happiness (ibid.; see also Stiglitz et

M3125 - MESSENGER TEXT.indd 8

14/05/2013 13:33

An introduction to work sharing

9

al. 2009). However, this generally positive association between paid work and happiness tends to turn negative when hours of work become long; for example, when individuals perceive that they have insufficient time to spend with their families or to handle domestic chores, they report being less happy with their lives in general (Böhnke 2005; Pouwels et al. 2008; Golden and Glosser, ch. 7). This finding regarding the detrimental effects of long hours of work on happiness suggests that permanent work-sharing policies have the potential to benefit employed workers as well. Specifically, two groups of employed workers would potentially benefit from work sharing. First, those workers who are most at risk of overwork – those working long hours and suffering from symptoms of work-related fatigue and stress – could clearly benefit from reduced hours. Second, those workers who are ‘overemployed’ – that is, those workers who, even if they are not working long hours, would prefer shorter hours of paid work even if it results in reduced earnings – could benefit from reduced hours as well (Golden and Glosser, ch. 7). Thus, part of the rationale for permanent work-sharing policies lies in the potential improvements in well-being for the overworked and the overemployed, as well as for those workers for whom unemployment (or underemployment) could be prevented (or lessened) via the jobs and hours gained from work sharing. The possibility for combined gains benefiting both of these groups of workers suggests that work sharing has the potential to improve the well-being not only of those workers who do not become unemployed due to work sharing and unemployed workers who become re-employed because of it, but also the happiness of employed individuals – and thus, the well-being of society as a whole. Although this potential effect of permanently reduced working hours on improved individual and societal well-being will be touched upon only briefly in this volume (particularly in Chapter 7), it is worth mentioning here simply to emphasize that the potential benefits of permanent working-time reductions go far beyond their employment effects alone.

4.

A BRIEF HISTORY OF WORK SHARING

4.1

Early Origins: The Reduction of the Working Week

Given the prominent role which Kurzarbeit (‘short-time work’) played as a job-preservation tool during the Great Recession, it is not especially surprising to learn that work sharing originated in Germany. In fact, the first agreements in Germany regarding the reduction of working time with

M3125 - MESSENGER TEXT.indd 9

14/05/2013 13:33

10

Work sharing during the Great Recession

monetary compensation for the reductions (in some cases) were made all the way back in 1891, and unemployment compensation, including short-time work benefits, was introduced at the beginning of the Weimar Republic in 1918 (Holzmayer 1989: 6). With the establishment of a government employment agency in Germany in 1927, short-time work became one of its labour market policy tools, where (despite some changes in the institutional framework in the post-Second World War period) it has remained to this day (ibid.) Although work sharing in its ‘short-time work’ form first appeared in Germany, work sharing as a measure to create new employment was part of the broader international push during the nineteenth and early twentieth centuries for the reduction of working hours and the establishment of an eight-hour day – a worker-led movement which ultimately resulted in the creation of the first international labour standard, the ILO Hours of Work (Industry) Convention, 1919 (No. 1). While the primarily objectives of reducing hours of work were the provision of social protection and more ‘free time’ for workers, the idea that shorter hours of work for those already employed would mean that more jobs would be created for those without work was part and parcel of this movement. This linkage can perhaps be encapsulated in the famous declaration by the President of the American Federation of Labor (AFL), Samuel Gompers, in 1887: ‘As long as we have one person seeking work who cannot find it, the hours of work are too long’ (Samuel Gompers, as quoted in Best 1981: 2–3). 4.2

The Great Depression: Mass Unemployment and the 40-hour Week

Despite these earlier antecedents, the concept of work sharing as we know it today appears to have taken form during the Great Depression of the 1930s. Even though, as described above, the primary objective of Convention No. 1 – and later, the Hours of Work (Commerce and Offices) Convention, 1930 (No. 30) – was the protection of workers, the economic and social crisis of the Great Depression resulted in the reduction of working hours being seen as a tool for combating unemployment as well. The concept of work sharing as a reduction in hours of work for the purpose of increasing employment is clearly reflected in the spirit of the ILO Forty-Hour Week Convention, 1935 (No. 47), adopted at the height of the Depression, which established the principle of the 40-hour week. Convention No. 47 noted that ‘unemployment has become so widespread and long continued that there are at the present time millions of workers around the world suffering hardship and privation from which they are justly entitled to be relieved’, and from this perspective, it advocated ‘that

M3125 - MESSENGER TEXT.indd 10

14/05/2013 13:33

An introduction to work sharing

11

a continuous effort should be made to reduce hours of work in all forms of employment to such extent as is possible’ (Preamble). These words still ring true today. During the Great Depression, a wide range of work-sharing initiatives, both industry and government led, emerged in both Europe and North America. According to a review of articles published in the International Labour Review between 1931 and 1939, ‘One of the earlier responses to the crisis . . . was the management of working hours’ (Mamudi 2009: 6). Legislation reducing the standard workweek was enacted in several European countries. For example, the German government enacted an Emergency Decree to reduce hours of work to 48 per week in 1931 and then moved to establish an even lower legal limit of 40 hours per week early in 1933, and France followed suit with legislation establishing a 40-hour workweek by 1936 (ibid.: 7). Italy also adopted the 40-hour workweek, first via a national agreement, implemented at the industry level, between the trade unions and the Federation of Italian Industry; this agreement was later codified in law in 1937 (Mattesini and Quintieri 2006: 422). Perhaps the most widespread use of work sharing to combat the Depression – both to preserve existing jobs and also to create new ones – was in the United States. Work sharing in the US began with voluntary, business-led efforts to shorten working hours as an alternative to layoffs during the soaring unemployment of the early 1930s, with strong support from the Hoover Administration (Best 1981). This early Depresssion form of work sharing, while appearing to save jobs, was often unpopular with workers both because it was unilaterally imposed by employers and also because it was uncompensated, and thus entailed substantial cuts in earnings (Nemirow 1984). With the arrival of the Roosevelt administration, a modified version of Depresssion-era work sharing became one of a number of major initiatives (along with public job creation and unemployment insurance) to combat unemployment in the US (Best 1981). A new programme called the President’s Reemployment Agreement (PRA), announced in July 1933, provided strong, patriotic encouragement for individual companies to sign an agreement between their firm and the President (Taylor 2011). The PRA ‘blanket’ agreement consisted of three parts: (i) a shorter workweek of 35 to 40 hours; (ii) an increase in hourly wage rates; and (iii) a clause recognizing workers’ rights to collective bargaining (ibid.: 135). This was a massive effort, which, depending on the specific estimate one chooses to believe, put between 1.34 and 4 million Americans back to work (ibid.: 133–4). These blanket agreements were later supplanted by industryspecific regulatory Codes under the National Industrial Recovery Act

M3125 - MESSENGER TEXT.indd 11

14/05/2013 13:33

12

Work sharing during the Great Recession

(NIRA), enacted in the same year, which established standard working hours and wage rates in each industry (ibid.; see also Mamudi 2009). As in the European countries discussed above, the shorter, 40-hour workweek in the United States was later enshrined in US Federal legislation, the Fair Labor Standards Act (FLSA) of 1938 (Best 1981). 4.3

Work Sharing in the Post-Second World War Era

During the period of prosperity that followed the Second World War, the concept of work sharing faded into the background as the economies of Western countries boomed and generated a quantity of jobs that was sufficient to keep most of the developed world at or near full employment for nearly three decades. It was only with the ‘stagflation’5 that followed in the wake of the oil-price shocks of the 1970s that the reduction of working hours as an employment policy re-emerged as a major issue in these countries. And even then, the policy experimentation with working-time reductions intended to create or preserve employment which occurred during that period was primarily in a handful of Western European countries, notably Belgium, Denmark, France, Germany and the Netherlands. These more recent European experiences with working-time reductions designed to create or preserve employment began during the late 1970s and continued through the 1980s and 1990s at varying paces in the different countries. The first initiatives during this period were undertaken by Belgium, beginning with a government proposal in 1979 for a subsidized reduction in the normal workweek from 40 to 36 hours with ‘wage moderation’ that was rejected by employers (and some unions), and continuing with the ‘3–5–3’ plan to increase employment (3 per cent lower increase in wages, combined with a 5 per cent reduction in working hours and a 3 per cent increase in employment) from 1983 to 1986, but appears to have generated only 23,000 new jobs (Kapteyn et al. 2000: 31). France, too, was an early leader with a Socialist government proposal in 1981 for a five year phased-in reduction of the legal workweek from 40 to 35 hours, but which was subsequently abandoned after the initial reduction from 40 to 39 hours (ibid.) – only to be revived by a later Socialist government as the Aubry I and Aubry II Laws (discussed below). The Netherlands, too, saw a number of working-time reduction initiatives during this period, beginning with a centralized agreement in 1982 – in the context of a high level of unemployment – between the social partners to gradually reduce the collectively agreed workweek from 40 to 38 hours, combined with a freeze on annual wage growth (Kapteyn et al. 2000). A 1985 survey found that 80 per cent of enterprises in the Netherlands reported reductions

M3125 - MESSENGER TEXT.indd 12

14/05/2013 13:33

An introduction to work sharing

13

in working hours, 17 per cent of firms reported that new jobs had been created, and another 26 per cent anticipated new job creation in the future (De Neubourg 1991). Perhaps the most substantial collective reductions in working hours during this period occurred in Germany, with the German trade union IG Metall successfully negotiating, in stages, a reduction in working hours in the metalworking industry from 40 hours per week in 1984 to 35 hours a week by 1995, with substantial reported positive employment effects estimated at between 21 and 70 per cent of the hours reduction (that is, of the theoretical maximum potential effect with a one-for-one conversion of hours to jobs) depending on the specific survey (Kapteyn et al. 2000: 33; Bosch and Lehndorff 2001: 222). More broadly, it has been estimated that the overall effects of collectively agreed working-time reductions in Germany during this period accounted for approximately 20 per cent of all new employment – approximately 420,000 out of 2.12 million new jobs (Seifert 1991). During this same period, work-sharing programmes designed to preserve existing jobs were introduced in Canada and in some states in the US as well. In Canada, in addition to a number of experiments with work-sharing schemes at the enterprise level, in 1982 the Canadian Federal government introduced a national work-sharing programme with time-limited ‘short-time compensation’ (STC) benefits payable out of unemployment compensation funds (Koeltz 2009; Golden and Glosser, ch. 7). Canada’s Work-Sharing Program provides businesses with an alternative to layoffs in situations where there is a reduction in demand for a company’s products or services; an evaluation of the programme’s effects during the 1990–91 recession in North America concluded that it successfully reduced layoffs, saving an annual average of 43,200 jobs per year during this two-year period (Graves and Dugas 1993, cited in Gray 1996: 805) In the United States, work sharing began in 1979 with a statelevel programme in California (Best 1981). Work-sharing programmes in the US – while authorized under a provision of Federal law enacted in 1992 – exist on a state-by-state basis (there is no national programme). A total of 17 US states6 had work-sharing programmes with STC benefits at the onset of the Great Recession, but most of them were extremely small (Balducchi and Wandner 2008). Without doubt, however, the most significant work-sharing initiative prior to the Great Recession was the recent French 35-hour workweek, perhaps the most large-scale work-sharing initiative since the Great Depression, and certainly the best known – and the most hotly debated. The ‘35 hours’ in France took the form of two laws, Aubry I and Aubry II,7 enacted in 1998 and 2000, respectively (see Askenazy 2008, and

M3125 - MESSENGER TEXT.indd 13

14/05/2013 13:33

14

Work sharing during the Great Recession

LaJeunesse 2009: 213–21, for comprehensive reviews of the French experience with the 35-hour week). Aubry I did not immediately impose a 35-hour normal workweek, but instead attempted to stimulate collective bargaining by providing firms with financial incentives to reduce working hours by 10 per cent and increase employment by 6 per cent prior to 1 January 2000 – at which time the legal limit on normal hours of work (before overtime) would be set at 35 hours per week for firms with more than 20 employees (1 January 2002 for enterprises with 20 employees or fewer). Despite some initial resistance from French employers, Aubry I did indeed stimulate collective negotiation, such that by the end of 1999, some 30,000 enterprises with two million employees had reached agreements to establish a 35-hour workweek and create new jobs (Askenazy 2008: 6). Aubry II codified the new 35-hour workweek, but also permitted the ‘35 hours’ to be annualized – that is, to be calculated on the basis of 1,600 hours per year, provided that working hours did not exceed 48 per week or an average of 44 hours over a 12-week period (LaJeunesse 2009). Aubry II also offered employers a number of concessions, especially in terms of how working hours are counted (for example, firms could exclude certain types of breaks, days off and training periods from this calculation), as well as removing the requirement that firms had to increase employment by 6 per cent in order to receive financial assistance from the government (ibid.) Perhaps not surprisingly the employment effects of the Aubry I law appear to have been far greater than those of Aubry II (according to Askenazy 2008, there was a net increase in employment of between 6 and 9 per cent under Aubry I versus 3 per cent for Aubry II); none the less, the overall employment effects of the two Aubry laws during the 1998–2002 period appear to have been quite substantial – with the most widely cited figure showing a net increase of between 300,000 and 350,000 jobs (Askenazy 2008: 17; LaJeunesse 2009: 217; see also Golden and Glosser, ch. 7). With the election of a new conservative government in France in 2002, a number of ‘adjustments’ (referred to as assouplissements in French, literally ‘softenings’) were made to the original Aubry laws. These adjustments raised statutory maximum limits in overtime hours and decoupled reductions in firms’ social insurance contributions from working-time reductions; later, under President Nicolas Sarkozy’s slogan ‘work more to earn more’, employers’ social insurance contributions were further reduced, and employees’ taxes on overtime pay were eliminated entirely (ibid.) While these various revisions did not actually repeal the original laws, they have none the less clearly ended the use of the ‘35 hours’ as a mechanism for promoting work sharing.

M3125 - MESSENGER TEXT.indd 14

14/05/2013 13:33

An introduction to work sharing

5.

15

THE RE-EMERGENCE OF WORK SHARING DURING THE GREAT RECESSION

The onset of the Great Recession and the global jobs crisis that it spawned led to a dramatic re-emergence of work sharing as a labour market policy tool aimed at preserving existing jobs. In the framework of national (and subnational) work-sharing measures, enterprises received benefits when they refrained from the use of layoffs, and instead ‘shared’ the lower amount of available work by reducing the working hours of all employees or all members of a work unit. The reduction in working hours under this form of work sharing is often (although not always) coupled with reductions in wages, which are typically proportional to the reduction in workers’ working hours (although this may not always be the case). This important constraint was often alleviated by the provision of wage supplements to affected workers, most typically by means of some form of partial unemployment compensation, although in some cases general government revenues were instead used to fund these payments. 5.1

The Expansion of Existing Work-sharing Measures in Developed Countries

Work-sharing programmes had already been implemented in a number of countries in the industrialized world prior to the onset of the global economic crisis, including: Austria, Belgium, Canada, France, Germany, Switzerland and the Netherlands, and small programmes in some individual states in the United States. Many of these existing work-sharing programmes were revised and expanded during the Great Recession. For example, the German Federal work-sharing programme, Kurzarbeit, was dramatically expanded. Specifically, the German government made the following change in the programme: the application procedure for the scheme was simplified; employers’ contributions for social insurance of short-time workers were reduced when their employees participated in training or qualification measures while working reduced hours; and the maximum duration of the scheme was increased first to 12 months, then extended to 18 months, and finally to 24 months. As a result, Kurzarbeit became the largest work-sharing programme in the world during the crisis, reaching a maximum participation of approximately 64,000 establishments and 1.5 million employees at the height of the crisis in mid-20098 (see Crimmann et al. 2010). Likewise, the French chômage partiel (‘partial unemployment’) programme extended the upper limit of non-worked hours covered by the partial unemployment contractual allowance from 600 to 800 hours per year, and up to 1,000 hours for firms in particularly

M3125 - MESSENGER TEXT.indd 15

14/05/2013 13:33

16

Work sharing during the Great Recession

vulnerable industries, such as textiles, garments and automobiles. All in all, the following developed countries had work-sharing measures in some form during the Great Recession: Austria, Belgium, Canada, Denmark, Finland, France, Germany, Italy, Japan, Luxembourg, the Republic of Korea, the Netherlands, Sweden and Switzerland, and small programmes in a number of individual US states.9 A number of different approaches were used for implementing worksharing measures during the Great Recession. First, national laws (and also subnational laws in Federal systems) often promoted – but did not mandate – the use of work sharing, and provided companies with financial incentives for adopting them. For example, Germany’s Kurzarbeit was (and remains) available to establishments facing a temporary, unavoidable loss of employment due to economic factors (see Crimmann et al., 2010). Work sharing was also widely used in the Republic of Korea, based on experiments with different forms of work sharing during the financial crisis of 1997–98. The Korean government actively promoted work sharing in individual companies during the crisis by introducing a wide range of financial incentives (see Korea International Labour Foundation 2009). Second, national-level agreements in some countries provided an overall framework for action at the enterprise level. The example of Japan demonstrates how a tripartite agreement at the national level can promote the adoption of work-sharing schemes in specific companies, providing both strong normative encouragement as well as financial incentives. A tripartite agreement to attain employment security and employment creation was concluded on 23 March 2009 by the Prime Minister, the Urban Business Federation (Nippon Keidanren), the Japan Chamber of Commerce and Industry, the National Federation of Small Business Associations, and the Japanese Trade Union Confederation (RENGO). This tripartite agreement included four components, one of which was the ‘Maintenance of employment through promoting the (so-called) “Japanese-model” of work sharing’ (see Messenger 2009: 3). Finally, work sharing was also adopted and implemented via collective bargaining at both industry and establishment levels, and not surprisingly, it was one of the top collective bargaining issues in Europe during 2009 (see Eurofound 2010). In Germany, most industry- and establishmentlevel collective agreements already contained provisions allowing enterprises to reduce working hours by as much as 20 per cent in order to avoid the loss of employment (see Bosch 2009). However, even in the absence of national work-sharing programmes or national tripartite framework agreements, work-sharing measures were none the less adopted and implemented via collective agreements at sectoral and enterprise levels.

M3125 - MESSENGER TEXT.indd 16

14/05/2013 13:33

An introduction to work sharing

17

For example, while there was no statutory work-sharing programme in Sweden, a collective agreement in the manufacturing sector in March 2009 allowed for the introduction of reduced working hours and temporary layoffs with compensation to ensure that workers received at least 80 per cent of their regular wages (Eurofound 2010: 18). Likewise, in Denmark, a large number of manufacturing firms concluded enterprise-level worksharing agreements during the crisis based on an earlier sectoral collective agreement, reached in 2007, which provided for a period of work sharing from 13 up to a maximum of 26 weeks (Glassner and Keune 2010: 15). 5.2

New Developments in Work Sharing in Middle-income Countries

Preserving jobs during the global crisis was also a top priority in many middle-income countries, which were particularly hard hit by job losses in their formal economies, often in export-oriented or consumer goods industries. As a result, during the Great Recession a number of these countries acted to discuss, negotiate and implement some basic forms of work sharing or similar working-time adjustments – often (but not always) with links to training. These countries include: Argentina, Chile, Mexico and Uruguay in Latin America; Bulgaria, the Czech Republic, Hungary, Romania, Slovakia and Slovenia in Eastern Europe; and also South Africa and Turkey. Among those work-sharing measures that were implemented in middleincome countries, the first revealing aspect is the different terminology used for the various schemes. Some countries referred to ‘reduced working time’ (Turkey), ‘flexiconto’ (Slovakia), or names based on the days worked plus the days not worked and those spent in training, such as the ‘411’ programme (Hungary); there was also the ‘labour training permission’ (Chile) and ‘Paros técnicos’, (‘technical unemployment’ in Mexico). These variations in terminology highlight differences in the design of the individual national measures. In some countries, only support for either reduced (weekly) working hours or temporary work stoppages (of a few weeks or months) were possible, while others offered provisions for both types of reduced working hours, with the same or varying conditions, and sometimes linked with training as well. In addition, some countries specifically adapted their work-sharing measures to the needs of small and medium-sized enterprises, such as in Argentina, Hungary and Romania. The timeframe for the use of these measures was nearly always temporary, ranging from three to 12 months in duration. Exceptionally in a few of these countries, the measures are a permanent feature of the labour law, which can be activated when necessary, for example Turkey (Messenger and Rodríguez 2010). Despite the differences among the crisis work-sharing measures across

M3125 - MESSENGER TEXT.indd 17

14/05/2013 13:33

18

Work sharing during the Great Recession

these countries, some common principles prevailed. One important similarity is that the level of development and implementation of work-sharing programmes in the two regions has been mainly at national level. Another similarity is that many countries made an important effort to extend unemployment benefit schemes and/or expand their coverage to workers with reduced hours; for example, some countries expanded the application, eligibility and coverage of partial unemployment benefits. In most cases, these work-sharing programmes also included a requirement for companies to demonstrate clear economic reasons in order to be eligible for any reduced working time or partial unemployment scheme. Also, there was often a requirement that employers maintain their previous level of employment while receiving a work-sharing subsidy or participating in a programme. Moreover, employers were required to continue paying their employees’ wages and social security contributions, although the latter were sometimes also at reduced levels (ibid.). Work-sharing measures to preserve jobs have been endorsed at international level by both the International Labour Organization (ILO) and the European Union (EU). In this context work sharing with partial unemployment benefits are policy responses suggested by the ILO Global Jobs Pact, adopted by the ILO’s tripartite constituents in June 2009, to limit or avoid job losses and to support enterprises in retaining their workforces (ILO 2009, Section III, Point 11.4). Likewise, various EU bodies highlighted the use of temporary ‘short-time work’ arrangements as one of the key measures to mitigate the impact of the global economic crisis and maintain employment, especially when accompanied by financial support to reduce workers’ income losses and training measures (see Council of the European Union 2009; European Commission 2009).

6.

MOVING BEYOND THE GREAT RECESSION: THE FOCUS OF THIS BOOK

This volume will examine the resurgence in the interest in and use of work sharing as a job-preservation strategy during the Great Recession, with the objective of learning lessons for the future from its successful use as a crisis-response measure in a number of countries. In addition, the volume will also consider the implications of the crisis experience for the application of work sharing as an employment-creation measure, thus contributing to the ongoing debate on the efficacy of this form of work sharing. This chapter has introduced the concept and history of work sharing and related working-time adjustment measures (for example, workingtime accounts), and has explained how work sharing can be used, first, as

M3125 - MESSENGER TEXT.indd 18

14/05/2013 13:33

An introduction to work sharing

19

a strategy for preserving existing jobs, and also as a tool for promoting increased employment and, simultaneously, improving individual and societal well-being. As discussed at the beginning of this chapter, the main focus of this volume is on crisis work-sharing measures, and Chapters 2–6 focus primarily (although not always exclusively) on such measures and their use and effects during the Great Recession of 2008–09 and its aftermath. Chapters 2 and 3 begin the review of these crisis work-sharing measures by investigating the use of work sharing and some allied measures (for example, working-time accounts, establishment-level agreements) as a job-preservation strategy in Europe during the crisis. Chapter 2 focuses on Germany, whose work-sharing programme, Kurzarbeit, was the largest in the world during the Great Recession, while Chapter 3 reviews four other European countries that substantially revised and expanded their worksharing programmes during the crisis – Austria, Belgium, France and the Netherlands. Chapter 4 rounds out our review of crisis measures in the developed world by analysing the case of a large Asian country, Japan, which made extensive use of several different mechanisms for the adjustment of working hours to avoid layoffs and thus ‘share’ the lower level of available work. Chapters 5 and 6 turn to the experiences of the new work-sharing measures which emerged in a number of middle-income developing countries for the first time during the Great Recession. Chapter 5 analyses the Reduced Working Time Programme in Turkey, which was by far the largest work-sharing measure in any developing country during the recession. Chapter 6 turns the focus to Latin America, and specifically to Uruguay, where two different instruments were implemented – one combining work sharing with training and the other sharing work by linking rotating, temporary layoffs with unemployment benefits. This chapter will compare and contrast those two measures and their effects. Following the analyses of these various crisis-related measures, Chapter 7 changes the focus by exploring the potential of the second form of work sharing – a government policy designed to induce permanent downward adjustments in average working hours (via a variety of possible methods) – to encourage additional hiring and thus increase the overall level of employment. This permanent form of work sharing is the primary focus of Chapter 7. The chapter will also consider the potential benefits of such permanent work-sharing policies beyond their possible effects on employment and unemployment, including promoting improvements in individual well-being. Finally, the concluding chapter will synthesize the main findings of the preceding chapters and consider their implications for the future of work

M3125 - MESSENGER TEXT.indd 19

14/05/2013 13:33

20

Work sharing during the Great Recession

sharing in both of its forms. First, it will present the lessons learned from the Great Recession and their implications for crisis work-sharing policies. Then it will move beyond the crisis experience to consider how to design permanent work-sharing policies in a manner that would provide appropriate incentives for generating positive employment effects, and thus help to promote a sustainable and job-rich economic recovery.

NOTES 1. This type of work sharing is also referred to as ‘short-time work’ and ‘partial’ or ‘ temporary’ unemployment (among other terms). 2. It should be noted that, with or without work-sharing measures, the available work is going to be ‘shared’ (distributed) in one form or another. In the absence of such measures, a market-induced form of ‘work sharing’ may involve increasing hours of paid work for a proportion of the workforce, while other workers receive no work at all, that is, they are unemployed. 3. Many critical analyses of permanent work sharing assume that the existence of a fixed amount of work in an economy is a key assumption underlying work sharing, but actually this is not the case. 4. In essence, Walker argues that there is a crucial distinction between shorter hours of work and the restriction of output. He cites Sydney J. Chapman’s (1909) theory of hours of labour, which emphasizes that the effects of technological progress will increase production and raise the relative value of leisure, and therefore, that ‘the optimal length of the working day must progressively decline’ (Walker 2007: 286). According to Walker, Chapman’s theory suggests that shorter hours of work lead to increased efficiency (a socalled ‘efficiency week’), resulting in lower prices and, ultimately, increased (rather than fixed) demand for products and labour. 5. The term ‘stagflation’, which was first coined in the 1970s, refers to the combination of economc stagnation, including higher unemployment levels, with relatively high levels of inflation – a new phenomenon at that time. 6. Arizona, Arkansas, California, Connecticut, Florida, Iowa, Kansas, Maryland, Massachusetts, Minnesota, Missouri, New York, Oregon, Rhode Island, Texas, Vermont and Washington (Ridley 2009). Six additional states plus the District of Columbia have recently established work-sharing programmes in the wake of the Great Recession: Colorado, Maine, New Hampshire, New Jersey, Oklahoma, and Pennsylvania. 7. The laws are named after the French Minister of Labour at that time, Martine Aubry. 8. Several other working-time adjustment measures were also widely used in Germany during the crisis, for example working-time accounts and establishment-level agreements. 9. Twenty-three states and the District of Columbia have established work-sharing programmes in the wake of the Great Recession (see n. 6 for exact states). In addition, new US Federal legislation was enacted in 2012 that provides incentive grants and general revenue funds to support the extension of work-sharing programmes to additional states, as well as the expansion of existing state programmes.

REFERENCES Address by Her Excellency Ms Angela Merkel, Chancellor, Federal Republic of Germany, to the International Labour Conference, Tuesday, 14 June 2011, 12

M3125 - MESSENGER TEXT.indd 20

14/05/2013 13:33

An introduction to work sharing

21

p.m., Provisional Record (PR No.22), International Labour Conference, 100e Session, Geneva, June 2011. Arpaia, A., N. Curci, E. Meyermans, J. Peschner and F. Pierini (2010), ‘Short-time working arrangements as response to cyclical fluctuations’, European Economy Occasional Papers No. 64, European Commission, Directorate-General for Economic and Financial Affairs and Directorate-General for Employment, Social Affairs and Equal Opportunities, Brussels. Askenazy, P. (2008), ‘A primer on the 35-hour in France, 1997–2007’, Discussion Paper No. 3402, Institute for the Study of Labour (Forschunginstitut zur Zunkunft der Arbeit, IZA), Bonn. Baker, D. (2009), ‘Job sharing: tax credits to prevent layoffs and stimulate employment’, Center for Economic and Policy Research, Washington, DC, October. Balducchi, D.E. and S.A. Wandner (2008), ‘Work sharing policy: power sharing and stalemate in American federalism’, Publius. The Journal of Federalism, 38 (1) (Winter): 111–36. Best, F. (1981), Work Sharing: Issues, Policy Options, and Prospects, Kalamazoo, MI: W.E. Upjohn Institute for Employment Research. Böhnke, P. (2005), First European Quality of Life Survey: Life Satisfaction, Happiness and a Sense of Belonging, Luxembourg: Office for Official Publications of the European Communities. Bosch, G. (2009), The European Debate on Work Sharing in the Economic Crisis, Seoul: Korean Labour Institute. Bosch, G. and S. Lehndorff (2001), ‘Working time reduction and employment: experiences in Europe and economic policy recommendations’, Cambridge Journal of Economics, 25: 209–43. Council of the European Union (2009), ‘Employment Committee’s contribution to the informal Employment Summit – Analysis carried out by the EMCO on short-time working arrangements’, Brussels, 26 May. Crimmann, A., F. Wiebner and L. Bellmann (2010), The German Work-Sharing Scheme: An Instrument for the Crisis, Conditions of Work and Employment Series No. 25, Geneva: ILO. De Neubourg, C. (1991), ‘Where have all the hours gone? Working time reduction policies in the Netherlands’, in K. Hinrichs, W.K. Roche and C. Sirianni (eds), Working Time in Transition: The Political Economy of Working Hours  in  Industrial Nations, Philadelphia, PA: Temple University Press. pp. 129–47. European Commission (2009), Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of Regions: A Shared Commitment for Employment, COM (2009) 257 final, European Commission, Brussels, 3 June. European Foundation For the Improvement of Living and Working Conditions (Eurofound) (2010), ‘Extending Flexi-security – the potential of short time working schemes’, ERM Report, European Foundation, Dublin. Freeman, R. (1998), ‘Work sharing to full employment: serious option or populist fallacy?’, in R. Freeman and P. Gottschalk (eds), Generating Jobs, New York: Russell Sage, pp. 195–222. Frey, B.S. and A. Stutzer (2010), ‘Happiness: a new approach in economics’, Happiness and Economic Policy, CESinfo DICE Report, 8 (4): 3–7. Glassner, V. and M. Keune (2010), ‘Negotiating the crisis? Collective bargaining in

M3125 - MESSENGER TEXT.indd 21

14/05/2013 13:33

22

Work sharing during the Great Recession

Europe during the economic downturn’, Industrial and Employment Relations Department, Working Paper, No. 10, ILO, Geneva. Gray, D. (1996), ‘Work-sharing benefits in Canada: an effective employment stabilization policy measure?’, Industrial Relations Quebec, 51 (4) (Fall): 802–22. Hijzen, A. and D. Venn (2011), ‘The role of short-time work schemes during the 2008–09 recession’, OECD Social, Employment and Migration Working Papers, No. 115, OECD, Paris. Holzmayer, W.T. (1989), Kurzarbeitergeld und Schlechtwettergeld: im entwicklungsgeschichtlicher Vergleich, Rheinfelden, Freiburg and Berlin: Schäuble. Hunt, J. (1998), ‘Hours reductions as work-sharing’, Brookings Papers on Economic Activity, 1: 339–81. Hunt, J. (1999), ‘Has work-sharing worked in Germany?’, Quarterly Journal of Economics, 114 (1): 117–48. International Labour Office (ILO) (2009), ‘Recovering from the crisis: a global jobs pact’, ILO, Geneva, 19 June. Kapteyn, A., A. Kalwij and A. Zaidi (2000), ‘Myth of Worksharing’, Department of Economics Discussion Paper Series No. 32, University of Oxford, Oxford. Koeltz, D. (2009), ‘Canada. Work sharing: an alternative to layoffs’, unpublished paper, ILO, Geneva. Korea International Labour Foundation (2009), ‘Twenty-three per cent of workplaces with over 100 employees join work sharing’, Labour Today, No. 615, 15 April. LaJeunesse, R. (2009), Work Time Regulation as Sustainable Full Employment Strategy: The Social Effort Bargain, London and New York: Routledge. Mamudi, R. (2009), ‘A survey of the Great Depression as recorded in the International Labour Review, 1931–1939’, Employment Working Paper No. 42, ILO, Geneva. Mattesini, F. and B. Quintieri (2006), ‘Does a reduction in the length of the working week reduce unemployment? Some evidence from the Italian economy during the Great Depression’, Explorations in Economic History, 43 (3): 413–37. Messenger, J.C. (2009), ‘Work sharing: a strategy to preserve jobs during the global jobs crisis’, Policy Brief No. 1, Conditions of Work and Employment Programme, ILO, Geneva, June. Messenger, J.C. and S. Rodríguez (2010), ‘New developments in work sharing in middle income countries’, Policy Brief No. 2, Conditions of Work and Employment Programme, ILO, Geneva, February. Nemirow, M. (1984), ‘Work-sharing approaches: past and present’, Monthly Labor Review, 107 (9): 34–9. Organization for Economic Cooperation and Development (OECD) (2010), OECD Employment Outlook: Moving beyond the Jobs Crisis, Paris: OECD. Pouwels, B., J. Stiegers and J.D. Vlasblom (2008), ‘Income, working hours, and happiness’, Economics Letters, 99: 72–4. Ridley, N. (2009), ‘Work sharing – an alternative to layoffs for tough times’, Center for Law and Social Policy (CLASP), Washington, DC. Seifert, H. (1991), ‘Employment effects of working time reductions in the former  Federal Republic of Germany’, International Labour Review, 130: 495–510. Stiglitz, J.E. A. Sen and J.-P. Fitoussi (2009), ‘Report by the Commission on the

M3125 - MESSENGER TEXT.indd 22

14/05/2013 13:33

An introduction to work sharing

23

Measurement of Economic Performance and Social Progress’, Commission on the Measurement of Economic Performance and Social Progress, Paris. Taylor, J.E. (2011), ‘Work-sharing during the Great Depression: did the “President’s Reemployment Agreement” promote reemployment?’, Economica, 78, 133–58. Walker, T. (2007), ‘Why economists dislike a lump of labor’, Review of Social Economy, 65 (3) (September): 279–91.

M3125 - MESSENGER TEXT.indd 23

14/05/2013 13:33

2.

Work sharing as an alternative to layoffs: Lessons from the German experience during the crisis Lutz Bellmann, Andreas Crimmann, Hans-Dieter Gerner and Frank Wießner*

1.

INTRODUCTION

In the second half of 2008, the global economy was hit by a massive shock originating from a deep financial crisis. Like many other countries around the globe, Germany searched for remedies and interventions, especially to ease the strain on the labour market. Unlike many other countries, however, Germany was reasonably considered a success. This chapter reviews the roots and causes of this positive result, and it mainly sheds light on Kurzarbeit, the German work-sharing scheme, which was a central policy in saving jobs. Current economic developments are still somewhat uncertain, difficult to predict and even harder to direct. There are even indications of a global debt crisis, bringing with it the danger of a renewed and perhaps long-lasting worldwide recession. For the time being it is difficult to judge whether the crisis came to an end in 2009, whether it will remain ongoing or whether we are facing a new crisis. Or maybe the financial crisis has temporarily submerged, and is just transforming into an economic one again. In any case, the situation remains uncertain. Although Germany itself was not hit as hard by the financial crisis as some other economies, the crisis revealed economic and structural shortcomings in many countries. The financial problems of some EU member states, such as Greece, Italy, Spain, Portugal and Ireland, as well as the level of household debt in the United States, are affecting Germany indirectly and with a time lag. Given the huge financial support and guarantees required to stabilize the euro system, these problems are touching the country more and more directly today in areas where the German economy has its own structural deficits. When jobs are endangered, individuals mainly focus on their own 24

M3125 - MESSENGER TEXT.indd 24

14/05/2013 13:33

Work sharing as an alternative to layoffs

25

situation. Labour market policy, on the other hand, has to offer measures to avoid layoffs and job losses more broadly based on reasoned understanding of the crisis in order to provide appropriate remedies. Programmes for employment protection alone cannot fully overcome such developments if they occur, but they can ease the individual consequences. Therefore a fact-finding process with regard to work sharing is still worthwhile. Work sharing may safeguard jobs for a certain time and under certain circumstances. Furthermore, it is one option to buy time for the development and implementation of concerted measures. However, work sharing is not an inexpensive option (although perhaps less expensive than macroeconomic stimulus packages) and thus reduces the financial resources available for structural economic adjustments. This chapter is structured as follows: first we take a look at the German economy with a special focus on the labour market (Section 2). In Section 3, we explain Kurzarbeit, the German work-sharing scheme, as one important instrument used to cushion the economic and financial crisis of the Great Recession of 2008–09. In Section 4 we present some descriptive information on the use of work sharing at establishment level and the use of training during periods of work sharing. In Section 5 we provide an empirical analysis of the determinants of the use of work sharing and its intensity, and in Section 6 we estimate the related employment effects. The final section draws some conclusions about the ‘safeguarding effect’ of work sharing.

2.

GERMANY’S LABOUR MARKET RESPONSE TO THE 2008–09 CRISIS

2.1

The State of the German Economy

Looking back at the turbulence of the last three years in late 2011, it seems that Germany apparently found a suitable policy mix to overcome the crisis with only slight employment losses. Despite the fact that the full consequences – and hence any repercussions from the crisis – have yet to be fully determined, the German economy is presently benefiting from a significant upturn, which began in the first quarter of 2010. Figure 2.1 shows the overall developments regarding gross domestic product (GDP) between 2006 and the first quarter of 2011. The decline becomes visible from the second quarter of 2008. As the line indicates, the first quarter of 2009 already marks a turning point of sorts. Between the third quarter of 2008 and the third quarter of 2009 we can see a clearly V-shaped course of these developments. The downturn in 2009 was

M3125 - MESSENGER TEXT.indd 25

14/05/2013 13:33

26

Work sharing during the Great Recession

6 4 2 0 −2 −4 −6 −8

I

II

III IV

2006

I

II

III IV

2007

I

II

III IV

2008

Changes in relation to previous year (%)

I

II

III IV

2009

I

II

III IV

2010

I 2011

Changes in relation to previous quarter (%)

Note: Adjusted in prices; index: year 2000 5 100. Source: Statistisches Bundesamt (2011), Fachserie 18, Reihe 1.2; own figures.

Figure 2.1

Development of German GDP 2006–2010 (in real terms)

the strongest ever observed in Germany since the Second World War. Interestingly, from 2010 onwards an unprecedented and lasting growth has followed, which still held at the end of 2011. As a result of this growth the GDP losses experienced in 2008–09 had been compensated almost entirely by the beginning of 2011. In fact, the shock waves from the financial and economic crisis reached the German labour market with a certain time lag. After the outbreak of the crisis, firms first managed to survive for a time on a comfortable backlog of orders. According to the German ‘Sachverständigenrat’ (the German Council of Economic Experts – GCEE), Germany currently finds itself in an uncertain situation. In the course of the recovery process, GDP regained its pre-crisis level in mid-2011. The GCEE forecast a GDP growth rate of 3.0 per cent for 2011 and of 0.9 per cent for 2012, which is a deceleration of the economic growth (Sachverständigenrat 2011: 10ff.). Furthermore the GCEE warned: The uncertainty surrounding the resolution of the sovereign debt crisis may have a major impact on world trade. There is a real danger that the already fraught funding conditions for sovereigns may tighten further. If the escalation of the crisis is confined to the euro area, the scenario projects a decrease in the volume of world trade from 4.9 per cent to 3.5 per cent. In that case German

M3125 - MESSENGER TEXT.indd 26

14/05/2013 13:33

Work sharing as an alternative to layoffs

27

9.7%

9.6%

2008 2010

8.6% 7.1% 6.1%

7.6% 7.1%

5.8%

OECD

United States

EU

Germany

Source: OECD Statistics, own figures.

Figure 2.2

Harmonized unemployment rates, 2008–10

GDP would expand by only 0.4 per cent. If the many risks and uncertainties lead to a stagnation of world trade in 2012, the rate of change of German GDP would drop to –0.5 per cent and thus to a recession. (p. 12)

2.2

The German Labour Market during the Crisis

A comparison of annual unemployment rates for the years 2008 and 2010 (Figure 2.2) shows that unemployment rose by 2.5 percentage points across all OECD countries, and even by 3.8 percentage points in the United States and by 2.6 percentage points within the European Union. With a declining unemployment rate (–0.5 percentage points), Germany was the big exception. Despite the fact that a large number of establishments were hit by the Great Recession of 2008–09, Krugman (2009) none the less noted of ‘Germany’s jobs miracle’. As we see from Table 2.1, the crisis hit the German economy in a rather selective way. The industries with the largest proportion of affected establishments were the automotive, mechanical engineering and chemical industries. Also affected were the temporary employment agencies, which had an ambiguous role, as temporary agency work in Germany basically functions as a flexibility buffer. First, it buffers demand peaks and second, it protects internal labour markets when demand is low. Besides the direct impact of the crisis on the demand for temporary agency work, they additionally functioned as a buffer for other industries which first let temporary agency work contracts expire before dismissing their own staff.

M3125 - MESSENGER TEXT.indd 27

14/05/2013 13:33

28

Work sharing during the Great Recession

Table 2.1

Proportion of establishments in Germany affected by the global crisis, and development of employment; selected industries (%)

Industries

Proportion of establishments affected by the crisis

Development of employment between 2008 and 2009

45 52 45 19 61 24 22 12 24 20 66 27

–3 –6 –3 0 –3 –1 2 0 0 –1 –34 –1

Manufacturing industry Automotive Chemicals Food Mechanical engineering Construction Hotels and restaurants Banking and insurance Wholesale and retail Additional service activities Temporary employment agencies Total

Sources: IAB Establishment Panel Survey 2008 and 2009 (see Bellmann and Gerner 2011b).

Unsurprisingly, in Germany the crisis primarily hit sectors which serve global markets. For instance, in September 2009, 220,000 employees in the engineering sector were participating in work-sharing schemes for economic reasons. This is equivalent to 22.7 per cent of all employees in this sector. Other sectors were affected only indirectly by the decline in orders from the exporting establishments. Consequently, the global players among German companies suffered more than small or medium-sized suppliers on local and regional markets. Beyond a rather small decline of 3 per cent in the entire manufacturing industry, temporary agency work was the big loser of the crisis with an employment decline of 34 per cent. The right-hand column of Table 2.1 shows that employment indeed remained remarkably stable in most industries, even if they were strongly affected by the crisis. 2.3

Reasons for the ‘German Miracle’

In the German case, both employment and unemployment remained stable although the economy suffered from a much stronger decline in real GDP than many other OECD countries, as Figure 2.3 shows. For instance, in the United States the unemployment rate rose inversely to the development of the economy. Spain, as another example, experienced a below-average decline in real GDP, but a tremendous increase in the

M3125 - MESSENGER TEXT.indd 28

14/05/2013 13:33

Work sharing as an alternative to layoffs 8

Countries with Short-Time Work Compensation (STC) programmes Countries with no STC programmes

E

Delta unemployment rate

IRL

6

IS

TR *

4

USA **

DK

2

FIN

MEX

H LUX

0 –10

Notes:

29

UK EU S J I D

OECD CDN P

CZ

SK

A

F

B

NL

–5 Real GDP growth

NZ GR

AUS PL

ROK N

0

*Value for Turkey is estimated; **work-sharing programmes in 17 states.

Sources: Own calculations based on OECD (2010a, b); see also Bell and Blanchflower (2009) and Möller (2010: 327).

Figure 2.3

Changes in real GDP and unemployment rates for 30 OECD countries (2009Q2 versus 2008Q2)

unemployment rate. Arpaia et al. (2010) provide a detailed international review of work-sharing arrangements in EU countries. A number of countries have also been using similar work-sharing approaches for several decades and some other countries implemented new schemes in the course of the recent crisis. Messenger (2009: 3) reports that work sharing is mostly found in the industrialized world: Austria, Belgium, Canada, Finland, France, Germany, Japan, the Republic of Korea, the Netherlands, Switzerland and a number of individual states in the United States. . . . Nonetheless, in the context of the global economic and jobs crises, a few developing and transition countries have begun to discuss and experiment with some basic forms of work sharing, including Chile, Mexico and Uruguay in Latin America, and the Czech Republic, Hungary, Slovakia and Slovenia in Eastern Europe.

Figure 2.3 also shows that many of the OECD economies which were hit by a strong decline in real GDP used work-sharing schemes (white circles in the figure), thereby managing to avoid stronger increases in unemployment. Cahuc and Carcillo (2011: 30) come to the conclusion ‘that shorttime work programs used in the recent downturn had significant beneficial

M3125 - MESSENGER TEXT.indd 29

14/05/2013 13:33

30

Work sharing during the Great Recession

effects. This suggests that countries which do not have short-time compensation programs could benefit from their introduction’. Although the correlation between the use of work sharing and the stability of the labour market during the crisis is striking, the German ‘jobs miracle’ is not based on that alone. Even before the crisis, labour market flexibility by hours instead of flexibility by numbers had a long tradition in most German establishments, which is currently fostered by company-level ‘pacts for employment and competitiveness’ that allow for flexibility in human resource management.1 In the first stage they generated a reduction of overtime and of the positive balances on working-time accounts and a hiring freeze that contributed to stabilizing employment. After that, temporary workers and freelancers were laid off and fixed-term contracts expired without being extended. Then, firms did not lay off their core employees but reduced their working hours temporarily. This was affordable due to previous productivity gains in connection with moderate collective bargaining. In addition to this, previous deregulations of the labour market had increased its general flexibility. Flexi-time arrangements, working-time accounts and company-level pacts for employment enabled establishments to adjust their production quickly to market changes without reducing their staff. Finally, after utilizing these flexibility buffers, the establishments started an extensive use of work sharing, which provides an attractive financial support for both employers and employees. All in all, the combination of these reforms and instruments increased both the external flexibility of the labour market and the internal flexibility of the establishments. As a consequence, ‘labour hoarding’ became an attractive alternative to ‘hiring and firing’. Another motivation for labour hoarding is the expected – and in some cases already noticeable – shortage of skilled labour, which will become even more serious in the coming years due to demographic changes. Meanwhile a lot of empirical research has been conducted to highlight the role of work sharing during the crisis. Bellmann and Gerner (2011b) assessed the impact of work-sharing arrangements on employment and earnings by comparing the change in the wage per employee and the number of employees per establishment between the first half of 2008 and the first half of 2009. Using data from the IAB Establishment Panel Survey (Fischer et al. 2009) and applying a multivariate framework, they distinguish between establishments which suffered from a decline in demand for their products or services due to the 2008–09 crisis and those that did not. Within the group of establishments affected by the crisis, the employment reduction concentrated on those that did not emphasize work sharing. While they reported an employment loss of approximately

M3125 - MESSENGER TEXT.indd 30

14/05/2013 13:33

Work sharing as an alternative to layoffs

31

6 per cent during the crisis, establishments which did use work sharing were able to stabilize their employment level. Furthermore, Bellmann and Gerner (2011b) show that work sharing did not cut earnings substantially, as many companies topped up the work-sharing compensation provided by the Federal Employment Agency. Based on the same data from the IAB Establishment Panel Survey, Boeri and Brücker (2011) estimate the employment effects of work sharing and identify small deadweight effects. Brenke et al. (2011: 12) also conclude that work sharing preserved jobs during the Great Recession: In periods in which the economy is struggling, the social blow of lost working hours can be cushioned; moreover when the situation improves, the necessary personnel are available immediately. Therefore, it was correct to make the regulations governing short-time work more attractive to those affected by the crisis.

Hijzen and Venn (2011) study the impact of work sharing during the 2008–09 crisis on the basis of data from 19 OECD countries. Of these countries, 11 utilized a work-sharing scheme which had already existed before the recent crisis, five introduced work sharing in the years 2008–09, and three had never had such schemes. Hijzen and Venn found the largest impact of work sharing on employment in Germany and Japan. According to their investigations, the positive impact of work sharing was mainly limited to workers with permanent contracts. This indicates an increased labour market segmentation between workers in regular jobs and workers in temporary and part-time jobs. Last but not least, according to their estimates only small deadweight occurred, which is in line with the results obtained by Boeri and Brücker (2011).

3.

KURZARBEIT – THE GERMAN WORK-SHARING SCHEME

3.1

Legal Conditions of Work Sharing

The basic idea of work sharing is not new at all. In fact, Kurzarbeit has been known in Germany since the 1920s, as discussed in Chapter  1. Clearly, Germany has a lot of experience in the use of work sharing (Box 2.1). After German reunification, work sharing was used on a massive scale in eastern Germany in the early 1990s. Nowadays there are three types of work-sharing measures in Germany:

M3125 - MESSENGER TEXT.indd 31

14/05/2013 13:33

32

Work sharing during the Great Recession

BOX 2.1 THE HISTORY OF WORK SHARING AND COMPENSATION IN GERMANY ●

●

● ● ● ● ●

First agreements on the temporary reduction of work as early as 1891, occasionally compensated by a ‘waiting allowance’ (Wartegeld) (Holzmayer 1989: 6). 1909: Tobacco Tax Amendment Act (Gesetz zur Änderung des Tabaksteuergesetzes) – first publicly financed compensation for a shortage of work due to higher taxes and customs fees. 1910: Kali-Law – temporary compensation for workers in the potash and fertilizer industry in the case of overproduction. Use of Kurzarbeit during and after the First World War at municipality level. Massive use of Kurzarbeit during the inflation years of 1922–23. Use of Kurzarbeit interrupted during the depression in 1924–26 due to lack of funds. Since 1927, Kurzarbeit has been used as a standard instrument of labour market policy (§§116ff. AVAVG – Gesetz über Arbeitsvermittlung und Arbeitslosenversicherung (Law on Job Placement and Unemployment Insurance)).

Source:

●

●

Own compilation.

Transfer-Kurzarbeit is used for permanent loss of employment2 due to re-structuring measures at establishment level (§216b Social Code Book III). This instrument was used extensively in former East Germany after reunification, when large parts of the former East German industry collapsed during the course of the economic transition. It was mainly intended to avoid widespread layoffs. According to Brenke et al. (2011: 4) it is not to overcome a temporary decline in production, but rather as ‘first aid’ to cushion the effects of the economic changes socially. The amount of short-time work compensation is similar to unemployment benefits. Seasonal work sharing (§175 Social Code Book III) is mainly used in the construction sector and for other ‘outdoor occupations’, such as landscape gardeners, to compensate for non-productive times due to weather conditions in the winter season. Otherwise workers in these occupations would be laid off at the beginning of the winter and

M3125 - MESSENGER TEXT.indd 32

14/05/2013 13:33

Work sharing as an alternative to layoffs

●

33

rehired in springtime. Seasonal work sharing compensation can be granted from 1 December to 31 March. Work sharing for economic reasons is used if there is a temporary, unavoidable reduction in the volume of work due to economic factors or an unavoidable event – the most important type of Kurzarbeit (§170 Social Code Book III). It provides a respite for establishments during an economic crisis, thus helping to protect jobs and maintain employment (this is the core instrument dealt with in this chapter).

In Germany either the management or the works council can request work sharing for economic reasons for the establishment. Beforehand, the establishment has to prove that other flexibility buffers such as the reduction of overtime, the reduction of positive balances in working-time accounts and paid leave have been utilized as far as possible. The establishment then has to submit a work-sharing plan to the local employment agency3 which indicates the estimated amount of work reduction in terms of the number of employees and the amount of working time. Short-time work compensation (Kurzerbeitergeld: KuG) is available for all employees covered by the social security system with a loss of at least 10 per cent of their gross monthly earnings. The calculation of the individual compensation is similar to that for unemployment benefits. For the lost working hours, employees with at least one dependent child receive compensation amounting to 67 per cent of the net difference to the regular wage, whereas those without dependants receive 60 per cent. For a 100 per cent loss of work the compensation is the same amount as the unemployment benefit (see, for example, Crimmann and Wießner 2009 or Crimmann et al. 2010). Figure 2.4 shows the calculation of short-time work compensation. Unlike unemployment benefits, compensation for work sharing in Germany (KuG) is granted to the establishment and not to the individual. First the establishment has to calculate these payments and add them to the individual paychecks according to the reduced working hours. By the end of the month, the employer has to report exact monthly figures to the local employment agency for an exact final accounting. On this basis, the establishment is paid the work-sharing compensation as a reimbursement. The maximum duration of work sharing used to be 12 months. In January 2009 it was extended to a maximum of 18 months as an initial response to the crisis, and on 5 June 2009, the maximum duration was extended to as long as 24 months following a decision made by the Federal Ministry of Labour and Social Affairs (BMAS). Since 1 January 2010 the duration was limited to 18 months again. Furthermore, in 2009 the government simplified the application procedure and reduced employers’

M3125 - MESSENGER TEXT.indd 33

14/05/2013 13:33

34

Work sharing during the Great Recession

50% loss of work KuG 60% or 67% +

100% loss of work KuG 60% or 67%

Work

Net income

Work

Income

Example 1: 50% loss of work

Work

Income

Example 2: 100% loss of work

Note: Kurzerbeitergeld (KuG) is 60 per cent of the net wage loss due to work sharing and 67 per cent with at least one dependent child living in the household.

Figure 2.4

Calculation of KuG (short-time work compensation) for partial and total loss of work

contributions to social security for employees participating in work sharing. Since these regulations were only temporary most of them have since been revoked. Basically, access to short-time compensation (STC) is more difficult now and conditions are more restrictive again, reflecting a limited duration designed to reduce deadweight effects. Table 2.2 shows the most important requirements and conditions of work sharing in Germany. 3.2

Advantages and Disadvantages of Work Sharing

3.2.1 Labour market policy perspective Of the instruments of labour market policy, work sharing is exceptional as it serves a hybrid function: in terms of active labour market policy it subsidizes the establishment to enable it to overcome a temporary shortage of work and to avoid layoffs. At individual level the compensation functions as an instrument of passive labour market policy similar to unemployment benefits. From the perspective of unemployment insurance, work sharing is rational if the shortage of work is temporary and not due to structural deficits in the affected industries. Work sharing pays off for the unemployment insurance system as long as the costs for the work-sharing compensation are lower than the

M3125 - MESSENGER TEXT.indd 34

14/05/2013 13:33

Work sharing as an alternative to layoffs

Table 2.2

35

Requirements for KuG (short-time work compensation) in Germany (according to §§169ff. Social Code Book III; from 1 April 2012 to §§95ff. Social Code Book III)

Requirements

Comments

Significant unavoidable shortage of work

The shortage of work must be for economic reasons (not structural ones) and unavoidable. It must be temporary (not permanent, i.e., there should be a good chance of returning to a normal workload in the future). At least one-third of the entire staff (excluding apprentices or trainees) must be affected with an estimated loss of income of at least 10%. At individual level, working hours can be reduced by up to 100% From 1 February 2009 until the end of 2010 a shortage of work also counted as significant if less than one-third of the entire staff is affected by a loss of 10%. The compensation can even be granted if the establishment is not able to provide full employment to at least one employee Since 1 January 2012, again at least one-third of the entire staff (excluding apprentices or trainees) must be affected with an estimated loss of income of at least 10%. At individual level, working hours can be reduced by up to 100% There must be at least one employee in regular employment (§171 Social Code Book III). There should be a good chance that the establishment will resume regular working hours with a reasonable workload in the future The employment contract must be maintained, i.e., the employee must not be laid off, nor may there be any agreements to dissolve the contract. The individual is also eligible for KuG in the case of incapacity to work due to illness. While receiving the compensation the individual is obliged to accept other job offers from the employment agency Either the employer or the works council is obliged to report the estimated shortage of work to the local employment agency and to prove the conditions of eligibility for receiving KuG For the number of working hours lost due to work sharing, the social security contributions are reduced to 80%. The employer has to cover the full amount*

Temporary modifications

Latest reform

Requirements of the establishment

Individual requirements

Notification

Contributions to social security

M3125 - MESSENGER TEXT.indd 35

14/05/2013 13:33

36

Table 2.2

Work sharing during the Great Recession

(continued)

Requirements

Comments

Temporary modifications

From 1 February 2009 the Federal Employment Agency covers 50% of the social security contributions for the shortage of work. In the case of training on the job during the shortage of work the Federal Employment Agency covers as much as 100% of the contributions. Under certain conditions up to 100% of the costs of the training measures can be reimbursed ● From 1 July 2009 the Federal Employment Agency covers the social security contributions for the shortage of work completely from the seventh month of work sharing – even without any training measures and no matter how many working hours are lost Since 1 January 2012, for the number of working hours lost due to work sharing, the social security contributions are reduced to 80%. The employer has to cover the full amount Recipients of unemployment benefit or similar benefits who are participating in publicly financed training measures do not qualify for KuG The duration of work sharing is institutionalized for all the employees of an establishment. It starts with the first month of the payment of KuG. If work sharing is interrupted for at least one month, the duration is extended correspondingly. After a break of at least three months a new period of work sharing may be started The maximum duration of KuG used to be 12 months. In January 2009 it was extended to a maximum of 18 months. On 5 June 2009, the maximum duration was extended to as long as 24 months following a decision made by BMAS. From 1 January 2010 the duration was limited to 18 months again Since 1 January 2012, the maximum duration of work sharing is now reduced to six months From 1 July 2009 employers or works councils do not necessarily have to reapply for KuG if the interruption is less than two months Since 1 January 2012, employers or works councils have to reapply for KuG. There must be at least three months between two consecutive periods of STC receipt

Latest reform

Disqualification

Duration

Latest reform Temporary modification Latest reform

M3125 - MESSENGER TEXT.indd 36

●

14/05/2013 13:33

Work sharing as an alternative to layoffs

Table 2.2

37

(continued)

Requirements

Comments

Latest reform

Since 1 January 2012, temporary agency work is excluded from receipt of KuG

Note: * Normally the social security contributions (retirement fund, unemployment insurance, health insurance, nursing-care insurance) are paid equally by employer and employee. The contributions to social security amount to approx. 40 per cent of the gross wage (that is, both employer and employee have to pay a share of 20 per cent). The contributions to statutory accident insurance must be paid entirely by the employer. Source: Crimmann et al. (2010), authors’ compilation.

unemployment benefits that would have been paid if the employee had been laid off. None the less, with the work-sharing option additional costs for unemployment benefits can occur if it is not possible to save the job in the long run.4 Hence, the decisive factor is the total duration of work sharing and/or unemployment. Work sharing becomes less attractive over time. A high probability of the individuals concerned becoming unemployed despite work sharing also reduces its cost-effectiveness. One reason why Germany decided to opt for large-scale work sharing during the Great Recession was the intention to avoid an increased level and persistence of unemployment. Especially in a recession, the average duration of unemployment tends to be longer. This would have devalued the latest labour market reforms and caused higher costs of unemployment (Blanchard 2006). However, it is clear that work sharing leads to some deadweight effects as well (Boeri and Brücker 2011: 26). In Germany, the fear of such potential negative effects was obviously outweighed by the fear of mass unemployment during the course of the crisis. 3.2.2 The employees’ perspective From the employees’ perspective, work sharing can have two different sides, as it may prevent unemployment or may only delay it. It clearly indicates that the job is in danger. On the other hand it is also a signal that the company would prefer to keep an employee rather than fire him/her. Income reductions and social costs are the price the employee has to pay for remaining employed. However, it is unclear whether this is sufficient to save the job and whether the future with the company is really worth it. However, leaving the establishment instead and looking for a new job can also involve substantial risks. In principle, one has to weigh up the probability of returning to regular work after work sharing against the counter probability of becoming

M3125 - MESSENGER TEXT.indd 37

14/05/2013 13:33

38

Work sharing during the Great Recession

unemployed after work sharing and finding a new job again. If an employee makes a decision to leave the company, he/she has to take into account a certain period of unemployment until he/she finds an adequate position again. A rational employee will prefer work sharing if the anticipated income reduction and social losses are lower than the expected costs of changing the job. 3.2.3 The employers’ perspective According to neoclassical economic theory, a declining demand for goods and products also reduces the demand for labour. In a theoretical optimum the marginal productivity of labour (MPL) is equal to the wage rate w. If this balance is disturbed, the wage rate exceeds the MPL, which leads to a declining demand for labour L. In the short term, labour can hardly be substituted by capital. Therefore production can only be adjusted by varying L. This can be brought about either by reducing staff or by reducing working hours. Human capital theory (Becker 1964) explains which employees should be retained and which should be laid off. In a simple model we differentiate between two groups of employees: ● ●

Group A: employees with general, unspecified human capital. These receive a wage of wA according to their marginal productivity. Group B: employees with specific, establishment-related human capital. These receive a wage of wB below their marginal productivity (Backes-Gellner et al. 2001).

With declining MPL due to underutilization and fixed wages due to collective agreements the rational choice of an employer is: ● ●

●

In the short term, to lay off A and keep B because wA . MPLA and wB , MPLB. In the medium term, to still keep B despite wB . MPLB because recruiting and training costs are still higher than the costs of firing if the company returns to the previous production level after the recession. In the long term, to lay off all employees with w . MPL.

This explains why work sharing is not equally suitable for all employees in an establishment. From the employer’s point of view there are two strategic options. On the one hand, the employer could go for a ‘fire and re-hire’ strategy, which seems to be quite suitable for group A. In that case the establishment suffers a loss of skilled labour and has to bear direct monetary costs due to the layoffs and later on for new recruitments and

M3125 - MESSENGER TEXT.indd 38

14/05/2013 13:33

Work sharing as an alternative to layoffs

39

training on the job. So-called ‘labour hoarding’ (that is, the retention of skilled labour despite temporarily slack demand) as the preferred option for group B implies costs for work sharing and other residual costs. The latter costs arise from a disproportionately small adjustment of operating costs to the reduction of work. For example, residual costs are incurred by the maintenance of machinery which is temporarily not being used due to a shortage of work. Work sharing is an option to reduce labour costs but the shortage of work itself is also costly: during the shortage of work contributions to health insurance, nursing-care insurance and retirement pension still have to be paid. As the compensation is provided by the Federal Employment Agency, the employer does not have to pay contributions to unemployment insurance while there is a shortage of work. The social security contributions paid during the shortage of work are based on 80 per cent of the difference between the regular gross wage and the gross wage while participating in work sharing. Whereas the social security contributions are normally shared equally between employer and employee, in the case of work sharing, the employer has to cover the full amount during the period of work sharing. From February 2009 until December 2011 the Federal Employment Agency exceptionally covered 50 per cent of these social security contributions.5 Work sharing becomes an option for the employer if the expected costs of work sharing and the residual costs are lower than the expected costs of hiring and firing. The latter costs are higher the more specific, establishment-related human capital an employee possesses. The costs of work sharing can be calculated precisely for every single employee, whereas the residual costs can only be estimated as an average per employee. With the ‘fire and re-hire’ strategy costs rise for severance payments and possibly for legal costs if an employee sues the employer regarding the dismissal. Additionally, the expected costs for hiring new staff have to be taken into consideration. These include expenditures for recruitment and training on the job. Furthermore, the new employee can be expected to be less productive during the starting phase of his/her employment. Work sharing reduces labour costs for the establishments since they only have to pay for the work actually performed. Losses of companyspecific skills or the loss of skilled labour to competing companies are avoided, as well as internal turbulence due to mass layoffs. Such ‘labour hoarding’ saves time and money if the workforce is going to be increased again when the economy recovers from the recession. As these considerations explain, work sharing has a twofold impact on labour costs. First, it reduces current labour costs immediately, thus

M3125 - MESSENGER TEXT.indd 39

14/05/2013 13:33

40

Work sharing during the Great Recession

helping to stabilize the establishment. Second, there is a sound motivation to discontinue work sharing soon, as a single working hour is more expensive than an hour under regular conditions. However, the model also shows that crisis work sharing is not suitable for a long-lasting or permanent shortage of work.

4.

THE USE OF WORK SHARING

4.1

Work Sharing for Economic Reasons

70

1,600 1,500 1,400 1,300 1,200 1,100 1,000 900 800 700 600 500 400 300 200 100 0

50 40 30 20

Establishments in thousands

60

10 0 Jan Feb March April May June July Aug Sept Oct Nov Dec Jan Feb March April May June July Aug Sept Oct Nov Dec Jan Feb March April May June July Aug Sept Oct Nov Dec Jan Feb March April May June July Aug Sept Oct Nov Dec Jan Feb March April May June July Aug Sept Oct Nov Dec

Individuals in thousands

In the years before the economic and financial crisis, the role of work sharing was rather marginal. Only seasonal work sharing was used to any degree, and it was known more as a temporary benefit for employees mainly in the construction sector and other outdoor occupations. From the third quarter of 2008 onwards we observe a tremendous increase in work sharing for economic reasons (Figure 2.5). In May 2009 the renewed work-sharing scheme reached its peak, with more than 1.4 million participants in approximately 60,000 establishments. The decline of work sharing in 2010 goes hand in hand with the recovery of the economy. Even

2007

2008

2009

2010

2011

Persons in work sharing for economic reasons Establishments with work sharing for economic reasons

Source: Own calculation based on data of the Federal Employment Agency.

Figure 2.5

M3125 - MESSENGER TEXT.indd 40

Work sharing in Germany (2007–2011)

14/05/2013 13:33

Work sharing as an alternative to layoffs

41

in 2011 we find quite a large number of establishments still using work sharing. Maybe some of these establishments are suffering more from structural problems than from a lack of orders, and are therefore unable to benefit from the economic upturn. 4.2

Flexibility at Establishment Level

4.2.1 Database: the IAB Establishment Panel For our empirical analysis of flexibility measures during the crisis – and especially of work sharing at establishment level – we use data from the IAB Establishment Panel. This annual survey is unique in Germany, as it represents all industries and establishment sizes nationwide. The survey on the basis of personal interviews started in western Germany in 1993 and has been conducted in eastern Germany too since 1996, making it a nationwide survey. The IAB Establishment Panel is conceived as a longitudinal survey, that is, a large majority of the same establishments are interviewed every year. Consequently, it permits both an analysis of developments across time by comparing cross-sectional data at different points in time, and also longitudinal studies of individual establishments. Currently, approximately 16,000 establishments are surveyed on a large number of subjects associated with employment policy, including, for example, working hours and continuing training. The survey also covers varying focal topics every year. The population of the IAB Establishment Panel consists of all establishments in Germany with at least one employee subject to social security as of 30 June of the previous year. The basis for sampling is the establishment file of the Federal Employment Agency, which contains around two million establishments that notify the social security agencies of their employees as required. The sample for the IAB Establishment Panel is drawn according to the ‘principle of optimum stratification’. As a result, large establishments, small federal states and small industries and the manufacturing industry in eastern Germany are overrepresented (Fischer et al. 2009). For the first time we use new data from the survey wave of 2009, the year of the crisis, and compare it to panel data from the year 2010, which was characterized by recovery from the recession. 4.2.2 The use of ‘flexibility’ measures In Germany, establishments can use a rich variety of ‘flexibility’ instruments. These can first be classified as internal and external measures. Second, they can be categorized according to the mechanism as numerical, monetary and functional instruments (Table 2.3). On the one hand, work sharing belongs to the internal and numerical instruments. On the other,

M3125 - MESSENGER TEXT.indd 41

14/05/2013 13:33

42

Table 2.3

Work sharing during the Great Recession

Classification of ‘flexibility’ measures in Germany Internal

Numerical

● ● ● ● ●

Monetary

● ● ●

Functional

● ●

overtime flexi-time, working-time accounts, part-time work, early retirement reduction of working time company holidays work sharing performance-related payment establishment-level agreements ‘Mini Jobs’* job rotation internal training

External ● ● ● ● ● ● ● ● ●

hire & fire, recalls fixed-term contracts temporary agency work internships voluntary work wage subsidies work sharing external training (occupational) retraining

Note: * In Germany so-called ‘Mini Jobs’ are ‘marginal’ part-time jobs. The jobholders are either paid a maximum wage of €400 per month or employed only short term (i.e., for a maximum of two months or 50 days per year). Source: Own compilation (see Will 2010: 2).

the financial compensation granted by the unemployment insurance can also be regarded as an external monetary measure. It is characteristic of industrial relations in Germany that there is a broad consensus between employers and unions regarding these instruments within the so-called ‘social partnership’. Indeed, many of them are even codified in collective agreements, which facilitates their implementation at establishment level. In practice, the use of such measures is agreed between the establishment and its works council. Figure 2.6 illustrates how establishments in Germany made use of different flexibility instruments during the economic and financial crisis. In the 2010 survey wave of the IAB Establishment Panel Survey, the establishments were asked if they had been affected by the financial and economic crisis during the previous two years, that is, 2008 and 2009. We distinguish between establishments which regard themselves as being ‘predominantly negatively affected by the crisis’ and establishments which see themselves as not particularly negatively affected. Around 36 per cent of all establishments regarded themselves as being affected by the crisis in both years. Figure 2.6 shows first that establishments affected by the crisis made much more use of all of the flexibility instruments than those which were not affected. Regardless of whether an establishment was in difficulty or not, the most common measures for capacity adjustment were a reduction

M3125 - MESSENGER TEXT.indd 42

14/05/2013 13:33

Work sharing as an alternative to layoffs Reduction of overtime/neg. balances on working-time accounts No refilling of vacant positions

24 9 18 5 18

Company holiday

5 17

Reduced increase of staff

4 16

Work sharing

3 16

Operational layoffs

3 12

Reduction of working hours

3 7

Reduction of temporary employment

2 6

Enhancement for further training

4 6

Other measures

2 5

Reduction in number of trainees taken on Reduction of temporary agency work 0%

43

2 4 1 5%

10%

Establishments in crisis

15%

20%

25%

Establishments not in crisis

Source: IAB Establishment Panel Survey 2010; own calculations.

Figure 2.6

The use of ‘flexibility’ measures during the crisis

of overtime and even allowing negative balances on working-time accounts. Second, we see that work sharing is neither the first choice nor the most frequently used instrument. This is in line with the legal requirements that establishments first have to utilize their so-called ‘flexibility buffers’ (for example, reducing working-time account balances to zero) prior to applying for work sharing. By programme design, work sharing is complementary: an option to be used after other flexibility measures have already been utilized. Only when these measures had reached their limits did work sharing take on a crucial role in avoiding layoffs. These characteristics of work sharing as a special option emerge even more clearly if we focus on those establishments affected by the crisis that were using work sharing. From Figure 2.7 we can see that one in two of these establishments had reduced overtime and even allowed negative balances on working-time accounts. During the crisis, company holidays were also used in many cases to bridge periods with a lack of orders. Then, in third place we find operational layoffs as a response in order to adjust the size of the workforce to a reduced demand for labour. Work sharing was not used equally by all establishments. Table 2.4 illustrates that it was much more common among medium-sized and large establishments. This may be due to the fact that in Germany larger

M3125 - MESSENGER TEXT.indd 43

14/05/2013 13:33

44

Work sharing during the Great Recession

Reduction of overtime/neg. balances on working-time accounts

50%

No refilling of vacant positions

26% 44%

Company holiday Reduced increase of staff

23%

Operational layoffs

30% 21%

Reduction of working hours 15%

Reduction of temporary employment Enhancement for further training

13% 8%

Other measures Reduction in number of trainees taken on

10% 13%

Reduction of temporary agency work 0%

10%

20%

30%

40%

50%

Source: IAB Establishment Panel Survey 2010; own calculations.

Figure 2.7

Table 2.4

‘Flexibility’ measures of establishments affected by the crisis and using work sharing Share of establishments using work sharing by size (%)

Establishment size (no. of employees) Small (1–49) Medium (50–249) Large (250 and more) All establishments

2009

2010

4 15 20 5

5 13 17 6

Source: IAB Establishment Panel Survey 2009–2010; own calculations.

establishments were affected by the crisis to a greater extent than smaller ones. Another possible explanation is that larger establishments have more flexibility options as well as professional human resource departments to put them into practice. Figure 2.8 shows in detail that the manufacturing industry and construction mainly opted for a reduction of working hours by work sharing, while this hardly played a role in trade, the service sector or the food industry. Even more interestingly, we can already observe a recovery in the industrial goods sector in 2010, whereas the use of work sharing still

M3125 - MESSENGER TEXT.indd 44

14/05/2013 13:33

Work sharing as an alternative to layoffs 30%

45

28%

25%

22%

22% 19%

20%

20%

16%

16%

15%

12% 10%

6% 5%

3% 1%

0%

5%

4% 2%

1%

3%

2009 2010 2009 2010 2009 2010 2009 2010 2009 2010 2009 2010 2009 2010 2009 2010 Food

Consumer Industrial Investment Constructgoods goods goods ion

Trade

Service

All sectors

Source: IAB Establishment Panel Survey 2009–2010; own calculations.

Figure 2.8

Share of establishments using work sharing by industry

increased in all other sectors in the same year. Once again, this indicates that the economic and financial crisis reached the labour market, though after a time lag. 4.2.3 Further training during periods of work sharing One of the innovations in Germany in response to the economic and financial crisis was the subsidization of training on the job during work sharing (see Table 2.2). For a descriptive overview, we differentiate between establishments by size. Small establishments have fewer than 50 employees, medium-sized ones have between 50 and 249, and establishments with 250 or more employees are referred to as large establishments (Figure 2.9). Of the establishments which used work sharing in 2009 (that is, 5 per cent of all establishments), 13 per cent organized continuing training during the period of work sharing. In 2010 this proportion increased to 14 per cent of the establishments using work sharing. In both years, the proportion of all establishments which used the working-time reduction due to a shortage of work for continuing training clearly remained under 1 per cent of all establishments. Figure 2.9 illustrates that such measures were much more popular among large establishments than among small or medium-sized ones.

M3125 - MESSENGER TEXT.indd 45

14/05/2013 13:33

46 35%

Work sharing during the Great Recession Establishments with training during work sharing Establishments with work sharing

30% 25% 20% 15% 10% 5% 10%

11%

25%

29%

45%

55%

13%

14%

2009

2010

2009

2010

2009

2010

2009

2010

0%

1−49

50−249

250+

All

Number of employees

Source: IAB Establishment Panel 2009–2010; authors’ calculations.

Figure 2.9

Proportion of establishments with work sharing and establishments providing training during work sharing in 2009 and 2010

Some 45 per cent of the establishments with 250 or more employees implemented training measures during work sharing in 2009 and 55 per cent in 2010. In general, larger establishments are more likely to invest in human capital than small or medium-sized ones: the higher the staff’s qualification level, the higher is an establishment’s propensity to support continuing training. This human resource strategy is obviously also applied to participants in work sharing. Compared to small establishments, larger establishments have more demand for continuing training, but also have other options to find skilled personnel. Similar to the use of work sharing, it was again the manufacturing industry which made most use of training during periods of work sharing. One in five establishments using work sharing in this sector offered continuing training to their staff (industrial goods: 14 per cent, construction sector: 8 per cent). Despite the fact that the combination of work sharing with continuing training was highly prominent in political terms during the years of the economic and financial crisis, we find indications that it was not of great importance for establishments. In 2009, establishments using work sharing requested only 17 per cent of the financial budget that the Federal

M3125 - MESSENGER TEXT.indd 46

14/05/2013 13:33

Work sharing as an alternative to layoffs

47

Employment Agency provided especially for continuing training during periods of work sharing. In relation to the total annual budget for work sharing, this was less than 1 per cent. From the establishments’ perspective, the expected returns on investments in further training are uncertain. Participants might be reluctant or simply not capable (Allaart et al. 2009). Furthermore, employees who have participated in company-financed training courses might leave the establishment before the investment has been fully ‘amortized’. These uncertainties appear to be even larger in a recessionary period. By subsidizing continuing training during work sharing, the government intended to create incentives to increase the incidence and intensity of continuing training in Germany, which is still below international levels (Behringer et al. 2008). Because of the demographic challenges with which Germany is confronted, additional efforts in the field of continuing training seem to be justified (Fuchs and Söhnlein 2009). Furthermore, analyses of further training during the 2008–09 crisis show that the incidence of continuing training declined, but the intensity did not (Bellmann et al. 2011). Thus the lower opportunity costs of training measures due to the shortage of work were not sufficient to raise the establishments’ commitment in this respect. However, generous subsidies and grants for continuing training increase the possibility of deadweight effects. Under these conditions, establishments might even be encouraged to invest in further training regardless of the level of expected returns; this was the case if the establishments were faced with the uncertainty caused by the global crisis. On the other hand, the subsidy would be efficient as long as the training costs were lower than the costs saved due to the reduction of social security contributions while using work sharing. This misguided incentive loses its attractiveness from the seventh month of work sharing, as the Federal Employment Agency covers the social security contributions for the period of work sharing then anyway. Additional problems and obstacles to the provision of continuing training in conjunction with work-sharing schemes were the lack of training programmes, goals and personnel needed, because to fulfil those prerequisites more time would be necessary than was available during the 2008–09 crisis. Even if these problems did not exist, the establishments still feared that the participants might be forced to interrupt or even cancel the training course if they were needed again at the establishment. Last but not least, the employees’ participation in training courses (in conjunction with work sharing) requires that these employees must have the opportunity to seek new jobs elsewhere, not only in the establishment with the worksharing scheme.

M3125 - MESSENGER TEXT.indd 47

14/05/2013 13:33

48

Work sharing during the Great Recession 5

13

14 26

Completely public

19

Mainly public 16

Equally shared

9

Mainly establishment Completely establishment

12 53 33

2009

2010

Source: IAB Establishment Panel 2009–2010.

Figure 2.10

Financing of training during periods of work sharing (in %)

Comparing the financial sources of the training measures during work sharing (Figure 2.10) we observe a shift from a larger share of public financing in 2009 to a larger contribution from the establishment side in 2010. This indicates that the subsidy alone was not the crucial factor for the establishments’ decision to opt for continuing training while implementing work sharing. If they considered training to be useful they did not mind financing it themselves, even without public support. Moreover, there is also some indication from interviews with human resource managers that the conditions for eligibility and the distribution of the subsidies for continuing training were rather bureaucratic and not well adapted to the establishments’ needs. This might have given the establishments another incentive to run training measures on their own budget without public support. Again, it must be mentioned that the financial support for training during work sharing did not play a major role among the efforts to combat the effects of the economic and financial crisis. There are clear indications that the establishments that opted for training during work sharing were mainly those that had already invested in the human capital of their staff as well. Thus these were complementary measures rather than an impulse for additional continuing training.

M3125 - MESSENGER TEXT.indd 48

14/05/2013 13:33

Work sharing as an alternative to layoffs

5.

49

ECONOMETRIC ANALYSES: DETERMINANTS OF THE USE OF WORK SHARING AND ITS INTENSITY

Cahuc and Carcillo (2011: 31) propose the use of larger sets of observations collected at firm level in addition to existing macroeconomic evaluations. A micro-econometric analysis seems to be most suitable for understanding how establishments reacted during the crisis. With a series of probit regression models, we estimate the probability of work sharing being used in establishments6 and the determinants of its use (for details see Appendix 2A (B), Table 2A.1). With the same set of explanatory variables we also estimate the intensity of the use of work sharing with truncated regression models (see Appendix 2A (B), Table 2A.2). The intensity is measured as the proportion of employees in work-sharing schemes in the first half of the year as a percentage of all employees in the same establishment,7 that is, the dependent variable ranges from 0 to 100 per cent. Our models do not include agriculture or the public sector, as work sharing does not play a role in these sectors. All in all, with our regression models we test the following hypotheses: Hypothesis 1: Before implementing work sharing, firms first try to reduce labour capacities by using both internal and external flexibility buffers. Hypothesis 2: Firms try to avoid loss of human capital or a shortage of skilled employees; work sharing is more likely for firms with higher investments in continuing training. Hypothesis 3: Firms are more likely to use work sharing (and to do so more intensively) the more they are affected by the global crisis. In Subsections 3.2.2 and 3.2.3 we showed that work sharing is not a cheap option for the establishment and that employees also have to cope with income cuts. Consequently, the establishments would be expected to use other flexibility buffers first (see Figure 2.6), before opting for work sharing (hypothesis 1). In our analyses we control for both internal and external flexibility measures. Labour capacities can be reduced externally by downsizing the number of freelancers or temporary agency workers. Internal flexibility is controlled by means of employees with fixed-term contracts, the share of employees with working-time accounts, and the existence of part-time work. Legally, both the works council and the management of an establishment can apply for work sharing. Hence, we expect that the existence of a works council will have a positive influence on the

M3125 - MESSENGER TEXT.indd 49

14/05/2013 13:33

50

Work sharing during the Great Recession

probability of work sharing.8 Collective agreements can also increase the probability of work sharing as they may contain special dismissal protection arrangements. For the manufacturing industry, it is also known that some collective agreements guaranteed additional monetary compensation in the case of work sharing. In Subsection 3.2.3 we explained that establishments have a strong motivation to avoid loss of human capital (hypothesis 2). The motivation should be even larger if the establishment has invested in training measures for the staff in the previous year.9 Additionally, we take vacancies for skilled employees in the previous year as an indicator of a shortage of ‘know-how’. As a standard we control for the (logarithmized) size of the establishment and the sector of industry. The descriptive statistics presented in Table 2.1 indicated that the crisis did not hit all sectors equally. With regard to the establishment size, scale effects would be expected, that is, there would be a higher probability of work sharing with increasing size. Establishments affected by the crisis (see Subsection 4.2.2) were also asked to rate the extent to which they were affected on a scale ranging from 1 for ‘low’ to 5 for ‘very strong’.10 In addition we control for the export ratio of the establishment’s revenues (hypothesis 3). The flexibility strategies which can be taken into account in our models include fixed-term contracting, the use of freelancers, temporary agency work and part-time work.11 Additionally, we have information about the share of employees with working-time accounts. In the case of declining demand for labour it is likely that the establishment will preserve its core staff to the disadvantage of the more flexible and less protected groups of staff. In addition, one important condition of eligibility for work sharing is that flexibility buffers must already have been utilized as far as possible. In general, all observed types of flexibility buffers show a negative tendency regarding the probability and the intensity of work sharing in both years. More employees with working-time accounts increase the probability but have no significant influence on the intensity of work-sharing use. Working-time accounts are complementary to work sharing but can be utilized only to a certain degree to respond to shortages of work. Hence working-time accounts lead to a higher probability of work sharing, but they do not increase its intensity. Co-determination is crucial for German industrial relations. However, it does not play a role for the probability of using work sharing. In a crisis the works council would probably first try to avoid layoffs, but second it would also attempt to prevent higher wage losses. As the truncated regression models show, the existence of a works council has a clearly positive impact on the intensity of work sharing in both years. A possible

M3125 - MESSENGER TEXT.indd 50

14/05/2013 13:33

Work sharing as an alternative to layoffs

51

explanation is that works councils pursue a strategy of distributing the shortage of work among more employees. Establishments may also be subject to collective agreements. In Germany, collective agreements usually go hand in hand with the size of the establishment. In our models such agreements do not have a significant influence on the probability or the intensity of work sharing. In several branches of industry, collective agreements guarantee additional payments for employees participating in work sharing. This again increases the costs of work sharing and results in a type of disadvantage for such companies. Furthermore, establishments have a vital interest in retaining highly productive and capable employees in the company also in times of temporary labour slack. Apart from layoff costs such as severance payments or other compensation payments, the loss of human capital is also of primary importance. Training measures in the previous year have only a slight influence on the probability of work sharing in the first half of 2010. If the use of work sharing in 2009 is controlled for, this significance disappears. For the intensity of work sharing, we find negative effects of training measures in the first half of 2010 (M5, M6). A possible explanation for this result is that establishments making use of continuing training anticipated a forthcoming increase in demand earlier and therefore already reduced the intensity of work sharing in the first half of 2010. Job vacancies in the previous year do not have a significant effect on the probability of work sharing; however, they reduce the intensity of work sharing significantly in all of our models. As a strategy the establishments might intend to avoid further losses of skilled labour; but this is relevant only for certain jobs and employees. With vacant jobs the establishment is also affected less by declining demand for work due to the crisis. The use of work sharing in 2009 strongly increased both the probability and the intensity of work sharing in 2010 (M3, M6). Evidently work sharing was not used as a short-term measure. Regardless of the general recovery of the economy beginning in 2010 (see Figure 2.1), from the establishments’ perception the crisis was not yet fully overcome by mid-2010. Establishment size is included in the models as the logarithmized total number of employees. A larger workforce increases the probability of work sharing in both years but at the same time reduces its intensity. On the one hand, larger establishments are more likely to have at least some of their employees participating in work sharing. On the other, larger establishments have more and different flexibility reserves than their smaller counterparts. They also often have greater capacities to organize alternative and flexible working processes. This might explain their

M3125 - MESSENGER TEXT.indd 51

14/05/2013 13:33

52

Work sharing during the Great Recession

lower intensity of work sharing. However, some new legislative amendments were aimed at promoting easier access to work sharing for smaller establishments. A higher export ratio of revenues increases the probability of work sharing only in 2009. This indicates that primarily export-oriented establishments were hit by the global economic and financial crisis due to declining foreign demand. The fact that export orientation has no significant impact on the probability or intensity of work sharing in 2010 might be explained by the broader range of the crisis at that time. After the ‘global players’ of German industry were first hit directly, their suppliers were also impacted by the crisis but with a time lag. As a consequence, differences between export-oriented establishments and those serving domestic markets diminished. In line with our expectations, the probability and the intensity of work sharing increase with the intensity of the crisis. For the year 2009, establishments with a low score for the intensity of the crisis show a negative intensity of work sharing. Presumably these establishments first tried to apply other flexibility measures before opting for work sharing on a larger scale. In order to analyse the influence of the industrial sector, we use the service sector (without trade) as a reference category. Only in 2009 did establishments in the food sector have a lower probability of work sharing than in the service sector.

6.

EMPLOYMENT EFFECTS OF WORK SHARING

In Figure 2.1, we saw that Germany suffered a massive decline in GDP due the Great Recession, mainly because of the strong export orientation of the economy. As Möller (2010: 330) explains, a cyclical decline in employment of –4.2 per cent in 2009 and –3.3 per cent in 2010 would have been expected as a worst-case scenario, which is based on the assumption of a close relationship between the cyclical components of production and employment. As employment subject to social security contributions stood at roughly 27.5 million in 2008, these decreases would imply a reduction of 0.88 million workers in 2009 and 0.61 million in 2010, amounting to as many as 1.5 million workers in total. Typically, more than 80 per cent (or 1.2 million workers) of the affected employees would appear on the unemployment registers sooner or later. On the contrary, however, as noted earlier, overall employment actually increased during the crisis, whereas the unemployment rate declined from 7.6 to 7.1 per cent (Figure  2.2). From Table 2.1 we know that the industries which were hardest hit by

M3125 - MESSENGER TEXT.indd 52

14/05/2013 13:33

Work sharing as an alternative to layoffs

Table 2.5

53

Change in employment in 2008–09 due to the crisis (dependent variable: logarithmized number of employees) All establishments

Crisis Year 2009 Crisis * Year 2009

0.175*** –0.004 –0.060***

Establishments Establishments without work sharing with work sharing 0.117*** 0.004 –0.057***

0.109 –0.089*** –0.011

Note: */**/*** Indicates significance at the 10%/5%/1% levels, respectively. Source: Bellmann and Gerner (2011b); IAB Establishment Panel 2008–2009.

the crisis were automotive, mechanical engineering and chemicals. In these industries, we find a large proportion of skilled workers who are even more expensive to replace than to retain in the case of a reduction in orders. This fact is likely to lead to ‘labour hoarding’ and may also be one of the important reasons for ‘Germany’s jobs miracle’ (Krugman 2009). In their empirical analysis using data from the IAB Establishment Panel survey Bellmann and Gerner (2011b) compare the change in the number of employees in establishments which reported being affected by the 2008–09 crisis with those which were not, between the first half of 2008 and the first half of 2009.12 The change in employment for establishments suffering from the crisis and those that did not and those using work sharing and those that did not is presented in Table 2.5. Furthermore, in order to assess the moderating effect of work sharing, Bellmann and Gerner (ibid.) conducted the same analysis separately for establishments that adopted work sharing and those that made no use of it. As we see from Table 2.5, establishments affected by the crisis (column 1) were 17.5 per cent larger than those that were unaffected. Between 2008 and 2009 there were no significant changes in employment (–0.004) in establishments not affected by the economic and financial crisis. In contrast, establishments affected by the crisis showed a significant decline in employment of around 6 per cent over time (interaction term consisting of crisis and time dummy, column 1, last row). Looking at the effect of the time dummy, we find a significant decline in employment (–8.9 per cent) for establishments that were using work sharing (column 3, row 2) whereas this has no significance for establishments without work sharing (column 2, row 2). Finally, for establishments not using work sharing the estimates exhibit a significant difference in the time trend from 2008 to 2009 and between crisis-stricken establishments and those establishments not affected (interaction term) of 5.7 per cent (column 2, row 3). For

M3125 - MESSENGER TEXT.indd 53

14/05/2013 13:33

54

Work sharing during the Great Recession

establishments using work sharing, we find no significant difference over time between establishments affected by the crisis and those that were not (column 3, row 3). Bellmann and Gerner (2011b) therefore summarize that despite the German jobs miracle there were substantial employment losses in establishments which were affected by the crisis. We only find weak evidence of work sharing playing a moderating role in the employment adjustment during the economic and financial crisis. Establishments which used work sharing reduced their employment between the first half of 2008 and the first half of 2009, irrespective of whether they were affected by the crisis or not. This result makes sense, because firms using work sharing are required to demonstrate that they are in a poor economic situation in order to be subsidized by KuG, so they would have likely suffered substantial employment losses in any case. In contrast to those findings, with the IAB Establishment Panel data, Boeri and Brücker (2011) estimate a positive employment effect of work sharing. They find that the number of jobs saved was approximately 400,000 in 2009 with only minor deadweight effects. Dietz et al. (2010), Möller (2010) and Ohanian (2010) also provide some empirical evidence based on aggregate data for Germany, and they show a strong decline in labour productivity from 2008 to 2009. As discussed above, estimations based on micro data indicate that those establishments which were affected by the economic and financial crisis decreased their employment significantly (Bellmann and Gerner 2011b). These estimates showed no significant moderating effects on employment adjustment in establishments affected by the crisis which used work sharing or in those which did not. At first glance, these results seem to contradict the expectations about the safeguarding effect of work sharing. However, they do in fact correspond if the reduction in employment at establishment level is disproportionately lower on average than the reduction in output. In order to identify employment retention due to labour hoarding at the establishment level, we therefore investigate the development of output per employee over time, especially from 2008 to 2009. Evidence of such labour hoarding is identified in a multivariate setting.13 The estimates are reported in Table 2.6. Since the base year for the estimation is 2009, our results indicate a strong decline in labour productivity for establishments which used work sharing only. For other establishments the decline in labour productivity is still significant in statistical terms, but not relevant in economic terms. For the year 2008, we find a 1 per cent higher productivity compared to the reference year of 2009. Taking into account the interaction between the use of work sharing and the time dummy, productivity drops by 16 per cent from

M3125 - MESSENGER TEXT.indd 54

14/05/2013 13:33

Work sharing as an alternative to layoffs

Table 2.6

55

Fixed-effects estimates of logarithmized output per employee

Year dummies (base category 2009) Year 2004 Year 2005 Year 2006 Year 2007 Year 2008 Interactions work sharing in 2009 and year dummies Year 2004 Year 2005 Year 2006 Year 2007 Year 2008 Number of employees (logarithm) Sum of investments (logarithm) Number of agency workers (logarithm) Technical state of the plant (1, very good . . . 5, poor) % part-time employees % women % qualified % trainees Works council Sectoral collective bargaining Firm-level collective bargaining Exporting firm Number of observations

–0.04*** –0.04*** 0.00 0.00 0.01*** 0.11*** 0.14*** 0.17*** 0.20*** 0.16*** –0.50*** 0.01*** 0.04*** –0.01*** –0.13*** –0.05* 0.01 –0.14*** 0.03 0.02** 0.01 0.02** 38,672

Note: ***/**/* Indicate significance at the 1%/5%/10% levels, respectively. Source: Own calculations based on the IAB Establishment Panel Survey 2004–2010.

2008 to 2009. For establishments which used work sharing, this amounts to a total reduction in productivity of minus 17 per cent. Putting the decline in productivity into relation with the decline in employment presented in Table 2.5, we find a disproportionately small adjustment of employment to productivity. This can be regarded as first evidence of employment retention due to labour hoarding arising from work sharing. However, the strong decline in labour productivity in establishments which adopted work sharing is mainly due to a strong decline in production which was still maintained with the same number of employees. We also observe a strong decline in labour productivity from 2008 to 2009 of around 4 per cent across all establishments, which means that

M3125 - MESSENGER TEXT.indd 55

14/05/2013 13:33

56

Work sharing during the Great Recession

they did not reduce their workforces in proportion with the reduction of output. Therefore, in the light of the overall fall in GDP of around 5 per cent from 2008 to 2009, labour hoarding seems to be an important explanatory factor for the German jobs miracle. Evidently, work sharing plays an important role in this context. Again, this makes sense because establishments which adopted work sharing were much more strongly affected by the economic and financial crisis in the first place.

7.

CONCLUSIONS: THE SAFEGUARDING EFFECT OF WORK SHARING

As was subsequently revealed, the perception of an ‘unprecedented deep short crisis’ was true. All in all, the recession was overcome quite quickly and the economy soon recovered, in the German case without many job losses. Moreover, during the upturn that followed the crisis, unemployment in Germany declined to a pre-crisis level. Given these facts, the set of interventions was obviously an appropriate response to the German type of crisis, which had a V-shaped course and in some cases differed from the shape of the crisis with which other countries had to cope. Considering the sovereign debts in the eurozone as well as the US budget deficit, it is still too early for a general all-clear. Basically it is possible that all the above-mentioned flexibility tools will have to be utilized again sooner than expected. However, if the recent economic upturn lost momentum or if another shock occurred, work sharing for economic reasons might lose its curativeness and might potentially cause unintended and even negative effects. The future situation in Germany would be completely different if the crisis returned and followed a W-, U- or L-shaped course. In the past the extensive use of work sharing has been an expensive but still affordable option, as German industry was hit only very selectively by the crisis. In accordance with the German model of social partnership employers, employees and the unemployment insurance contributed their share, thereby making the measures socially and politically acceptable (Crimmann et al. 2009). If another shock wave were to hit the economy on an even larger scale it might be difficult to maintain such interventions over the long term. As work sharing induces labour hoarding, it might postpone or even hamper the implementation of inevitable changes in an establishment’s organizational environment. Supporting jobs in sectors that are in structural decline could slow down job reallocation (Cahuc and Carcillo 2011: 30). If this leads to a slowdown of structural change or even to ‘structural conservatism’, it will even be an obstacle to future recovery

M3125 - MESSENGER TEXT.indd 56

14/05/2013 13:33

Work sharing as an alternative to layoffs

57

and growth. For instance, even in a recession there is still considerable hiring. As a matter of fact, work sharing, in interrelation with other flexibility instruments such as working-time accounts or ‘company-level pacts for employment and competitiveness’ turned out to be an important employment stabilizer during the crisis. Naturally, it would be interesting to know at least approximately how many jobs were saved by means of work sharing. A reliable estimation or even a serious guess is extremely difficult. Basically, we cannot control for relevant macroeconomic effects such as the stabilization of domestic demand with the given microeconomic data. Additionally, the true counterfactual situation of a labour market without work sharing cannot be modelled. None the less, some approaches have been published: Möller (2010: 330) develops a worst-case scenario in which as many as 1.5 million jobs could have been lost due to the crisis. However, one has to understand that this does not simply correspond to the number of jobs saved. In this context, Boeri and Brücker (2011) estimate that work sharing safeguarded approximately 400,000 jobs in 2009. On the basis of the number of participants in work sharing and the average shortage of work, the Federal Employment Agency reports around 300,000 full-time equivalents for 2009 with a peak of around 393,000 in April 2009. In addition, we have shown that in establishments which were using work sharing (that is, around 60,000 establishments with an average of 1.1 million participants in 2009) labour productivity declined by 17 per cent, and by 1 per cent in establishments without work sharing. With careful and conservative consideration we would interpret our results as supporting the estimations of Boeri and Brücker for 2009. Once again, it must be stressed that this does not hold for the counterfactual situation of a ‘world without work sharing’. In that case, establishments would have made use of other flexibility options, with layoffs being only one of several options. As mentioned earlier, it has to be stressed again that in the German case the crisis hit the economy rather selectively. Affected establishments were mostly found in the automotive industry, in mechanical and chemical engineering, and among temporary employment agencies (which have the function of a flexibility buffer in any case). Furthermore, most of the affected establishments were export oriented. However, many of these global players had – and still have – a very strong position in their specific markets. In other words, these establishments were not as vulnerable as they could have been and work sharing did not have to cushion the impact of the crisis alone, but rather extended and enhanced their resistance. Beyond questions of the necessity of any measures or interventions the government showed a strong commitment to stabilizing the economy as a

M3125 - MESSENGER TEXT.indd 57

14/05/2013 13:33

58

Work sharing during the Great Recession

whole and the easy access to KuG gave a strong signal not only to affected establishments, but also to indirectly-related ones. Reinhart and Rogoff (2009) examined the aftermath of severe financial crises and found deep and lasting effects on asset prices, output and employment. They found that increases in unemployment and decreases in housing prices last for five and six years, respectively. If these findings also hold for Germany, then work sharing might not have avoided layoffs, but simply postponed them. However, the 2008–09 crisis was exceptional in many respects and did not fit into the normal pattern. The programme can therefore be regarded as quite successful. For the time being, the induced labour hoarding gave the German economy a jump-start after the crisis, leading to a quick recovery and an economic upturn, which continued through to the end of 2011. Regarding this as a contribution to future productivity, the indirect effect of work sharing during and even after the crisis might be an even stronger stimulus than that of most of the other economic programmes. Whereas work sharing itself can be regarded as a success, the outcomes of training during work sharing were rather sobering. The value added due to continuing training during reduced working hours remained clearly below expectations. For a number of reasons the scale of the scheme was much too small for one to be able to speak honestly of strengthened competitiveness or strong commitment and motivation of employees and employers. In addition to the microeconomic analyses presented in this chapter, macroeconomic research is needed, too, not only to assess possible deadweight effects, but also to compare work sharing with other instruments (see Balleer et al. 2011). Furthermore, in conjunction with these measures which seek to preserve jobs directly, it would be interesting to investigate the effect of fiscal stimulus programmes as well. Since such stimulus programmes were also used not only in Germany, but also in many other countries during the 2008–09 crisis, a comparative analysis is feasible. Intuitively, work sharing has clear advantages in relation to macroeconomic stabilization policies, because of its not only direct but also immediate response to the Great Recession which unfolded after the fall of Lehman Brothers.

NOTES *

The authors would like to thank three anonymous peer reviewers for their helpful comments and Mr Harald Maus from the German Federal Employment Agency who commented on the legislative framework for work sharing.

M3125 - MESSENGER TEXT.indd 58

14/05/2013 13:33

Work sharing as an alternative to layoffs 1.

2. 3. 4. 5.

6. 7. 8. 9. 10.

11. 12. 13.

59

‘Pacts for employment and competitiveness’ consist on the one hand of employees’ concessions concerning working time, monetary, organizational and further training measures and on the other of employers‘ pledges concerning employment and investment guarantees (see Bellmann and Gerner 2011a). Note: The legal status of the participants of Transfer-Kurzarbeit is still ‘employed’. The German Federal Employment Agency has 180 local employment agencies nationwide with a total of around 640 local offices. For a technical discussion of this issue, see Appendix 2A (A). If the establishment provides training on the job while implementing work sharing, the Federal Employment Agency covers the entire cost of social security contributions for the shortage of work. From the seventh month of work sharing, the Federal Employment Agency also covers the entire social security contributions for the shortage of work. The exact wording of the question in the IAB Establishment Panel Survey is ‘Did you have to make use of short-time work in the first half of 2009?’ (or 2010) with ‘Yes’ and ‘No’ as possible answers. The denominator of this fraction is the number of employees as of the reference date of 30 June. A legal precondition for the election of a works council in an establishment is that the establishment has at least five employees in permanent employment. To identify these establishments we control for such investments in the 2008 survey wave. Establishments in the estimation for 2009 must have participated in both waves. The crisis indicator and related questions about the extent to which they were affected by the crisis were only asked of establishments which participated in the 2010 survey wave. Retrospective estimations for 2009 are possible only for establishments which participated in both waves. Defined in our model as less than 20 working hours per week. For technical details, see Appendix 2A (C). For technical details, see Appendix 2A (D).

REFERENCES Allaart, P., L. Bellmann and U. Leber (2009), ‘Company-provided further training in Germany and the Netherlands’, Empirical Research in Vocational Education and Training, 1 (2): 103–22. Arpaia, A., N. Curci, E. Meyermans, J. Peschner and F. Pierini (2010), ‘Work sharing arrangements as response to cyclical fluctuations’, European Economy Occasional Paper No. 64, European Commission, Brussels. Backes-Gellner, U., E. Lazear and B. Wolff (2001), Personalökonomik: Fortgeschrittene Anwendungen für das Management, Stuttgart: Schäffer-Poeschel. Balleer, A., B. Gehrke, W. Lechthaler and Chr. Merkl (2011), ‘Work sharing and the macro-economy’, University of Erlangen-Nuremberg (manuscript). Becker, G. (1964), Human Capital: A Theoretical and Empirical Analysis, with Special Reference to Education, New York: Columbia University Press. Behringer, F., D. Moraal and G. Schönfeld (2008), ‘Betriebliche Weiterbildung in Europa: Deutschland weiterhin nur im Mittelfeld’, Berufsbildung in Wissenschaft und Praxis, 37 (1): 9–14. Bell, D. and D. Blanchflower (2009), ‘What should be done about rising unemployment in the OECD?’, IZA Discussion Paper No. 4455, Bonn. Bellmann, L. and H.-D. Gerner (2011a), ‘Further training and company-level pacts for employment in Germany’, Jahrbücher für Nationalökonomie und Statistik, 32 (2): 98–115.

M3125 - MESSENGER TEXT.indd 59

14/05/2013 13:33

60

Work sharing during the Great Recession

Bellmann, L. and H.-D. Gerner (2011b), ‘Reversed roles? Wage and employment effects of the current crisis’, in H. Immervoll, A. Peichi and K. Tatsiramos (eds), Who Loses in the Downturn? Economic Crisis, Employment and Income Distribution, Research in Labor Economics, Vol. 32, Bingley, UK: Emerald, pp. 181–206. Bellmann, L., H.-D. Gerner and U. Leber (2011), ‘Further training during the crisis’, Paper prepared for the meeting of the International Working Party on Labor Market Segmentation (IWPLMS), Bamberg, 11–13 July. Blanchard, O. (2006), ‘European unemployment. The evolution of facts and ideas’, Economic Policy. A European Forum, No. 45: 5–59. Boeri, T. and H. Brücker (2011), ‘Work sharing benefits revisited: some lessons from the Great Recession’, IZA Discussion Paper No. 5635, Bonn. Brenke, K., U. Rinne and K.F. Zimmermann (2011), ‘Work sharing: the German answer to the Great Recession’, IZA Discussion Paper No. 5780, Bonn. Cahuc, P. and S. Carcillo (2011), ‘Is work sharing a good method to keep unemployment down?’, IZA Discussion Paper No. 5430, Bonn. Crimmann, A. and F. Wießner (2009), ‘Wirtschafts- und Finanzkrise: Verschnaufpause dank Kurzarbeit’, IAB-Kurzbericht 14/2009, Nürnberg. Crimmann, A., F. Wießner and L. Bellmann (2010), ‘The German work-sharing scheme: an instrument for the crisis’, Conditions of Work and Employment Series, No. 25, ILO, Geneva. Dietz, M., M. Stops and U. Walwei (2010), ‘Safeguarding jobs through labor hoarding in Germany’, in K.F. Zimmermann and C. Wey (eds), The Economy, Crises, and the Labor Market: Can Institutions Serve as a Protective Shield for Employment?, Berlin: Duncker & Humblot, pp. 125–49. Fischer, G., F. Janik, D. Müller and A. Schmucker (2009), ‘The IAB establishment panel – things users should know’, Schmollers Jahrbuch. Journal of Applied Social Science Studies, 129: 133–48. Fuchs, J. and D. Söhnlein (2009), ‘Der Einfluss der Bevölkerungsentwicklung auf das  künftige Erwerbspersonenpotenzial in Ost- und Westdeutschland’, in I. Cassens, M. Luya and R. Scholz (eds), Die Bevölkerung in Ostund  Westdeutschland. Demografische, gesellschaftliche und wirtschaftliche Entwicklungen seit der Wende, Wiesbaden: Vs Verlag für Sozialwissenschaften, pp. 200–222. Hijzen, A. and D. Venn (2011), ‘The role of short-time-work schemes during the 2008–09 Recession’, OECD Social, Employment and Migration Working Paper No. 115, OECD, Paris. Holzmayer, W.T. (1989), Kurzarbeitergeld und Schlechtwettergeld: ein entwicklungsgeschichtlicher Vergleich, Rheinfelden, Freiburg, Berlin: Schäuble. Krugman, Paul (2009), ‘Free to lose’, New York Times, 12 November 2009, available at http://www.nytimes.com/2009/11/13/opinion/13krugman.html. Messenger, J. (2009), ‘Work sharing: a strategy to preserve jobs during the global jobs crisis’, Policy Brief No. 1, Conditions of Work and Employment Programme, ILO, Geneva, June. Möller, J. (2010), ‘The German labor market response in the world recession – de-mystifying a miracle’, Zeitschrift für Arbeitsmarktforschung, 42 (4): 325–36. OECD (2010a), ‘Quarterly National Accounts: Quarterly Growth Rates of GDP’, available at: http://stats.oecd.org/index.aspx?queryid5350 (accessed 4 March 2010).

M3125 - MESSENGER TEXT.indd 60

14/05/2013 13:33

Work sharing as an alternative to layoffs

61

OECD (2010b), ‘Key Short-Term Economic Indicators: Harmonised Unemployment Rate’, available at: http://stats.oecd.org/Index.aspx?DataSetCode5CSP2009 (accessed 4 March 2010). Ohanian, L.E. (2010), ‘The economic crisis from a neoclassical perspective’. Journal of Economic Perspectives, 24: 45–66. Reinhart, C. and K. Rogoff (2009), ‘The aftermath of financial crises’, National Bureau of Economic Research, NBER Working Paper 14656, Cambridge, MA. Sachverständigenrat zur Begutachtung der gesamtwirtschaftlichen Entwicklung (2011), Verantwortung für Europa wahrnehmen. Jahresgutachten 2011/12, Wiesbaden: Statistisches Bundesamt. Statistisches Bundesamt (2011), ‘Fachserie 18, Reihe 1.2’ (Volkswirtschaftliche Gesamtrechnungen, Inlandsproduktsberechnung, Vierteljahresergebnisse, 3. Vierteljahr 2011), Wiesbaden. Will, H. (2010), ‘Kurzarbeit als Flexibilisierungsinstrument. Hemmnis strukturellen Wandels oder konjunkturelle Brücke für Beschäftigung?’, IMK Study 5/2010, Düsseldorf.

M3125 - MESSENGER TEXT.indd 61

14/05/2013 13:33

62

Work sharing during the Great Recession

APPENDIX 2A Appendix A: Formal Discussion of the Advantages and Disadvantages of Work Sharing 1. Labour market policy perspective From the perspective of the unemployment insurance system, work sharing is rational if the shortage of work is temporary and not due to structural deficits of the affected industries. Formally we have: Ec a

H G Cws Cue Cue ( ) 1 1 2 p , E d c d. a a t t t t51 (1 1 r) t5T (1 1 r) t51 (1 1 r) T

Work sharing pays off for the unemployment insurance system as long as the costs for the work-sharing compensation Cws are lower than the unemployment benefits Cue to be paid if the employee were laid off. None the less, for the work-sharing option, additional costs for unemployment benefits Cue can occur if it is not possible to save the job in the long run. As we see from the equation, the decisive factor is the duration of work sharing or of unemployment. Work sharing becomes less attractive over time. A high probability of individuals becoming unemployed despite work sharing also reduces its cost-effectiveness. 2. The employees’ perspective Like the employer, an employee will assess the opportunity costs of staying with the company or leaving it. Ee a

F H F Uws Ul Uue Ul 1 pa 1 (1 2 p) c a 1 a df t t t t t51 (1 1 r) t5T (1 1 r) t5T (1 1 r) t5H (1 1 r) T

G F Uue Ul . Ec a 1 a (1 1 r) t d . t ( ) 1 1 r t51 t5G

Formally we have a utility function U, which consists of income and social utility, for example, social contacts and participation, acknowledgement or a satisfying occupation. U has different values for work (Ul), work sharing (Uws) and unemployment (Uue). For work sharing and for unemployment we assume a lower U than for regular work. Furthermore, the probability of returning to regular work after work sharing and the counter probability of becoming unemployed after work sharing and finding a new job again also have to be considered.

M3125 - MESSENGER TEXT.indd 62

14/05/2013 13:33

Work sharing as an alternative to layoffs

63

3. The employers’ perspective Basically labour costs consist of two parts:1 ● ●

Current costs of labour change with the amount of work being performed. This is true of gross wages and social security contributions. Quasi-fixed labour costs do not depend on working hours but on the number of employees, for example, costs for hiring and firing or training on the job or the maintenance of a canteen.

In our considerations we analyse the current costs of work sharing for an establishment. We make the simplifying assumption that the current costs of labour CW are a function of the wage rate w and the volume of work L in terms of working hours performed plus the employer’s contributions to social security CS (which is a function of the contribution rate s, the wage rate w and the volume of work L) plus the quasi-fixed costs of labour CQF: CW 5 wL 1 CS 1 CQF. In the case of work sharing, the employer has to pay for the work being performed (wL) plus contributions to social security. The latter consist of the employer’s share for the work being performed Cs(work) and for both the employer’s and the employee’s share for the period of the shortage of work Cs(loss). Taking the contracted working time as 1, in the case of work sharing an employee performs a number of working hours L with a shortage of work of (1 – L). The current costs of labour total up as follows: CW 5 wL (1 1 CS(work)) 1 w (1 2 L) CS(loss) 1 CQF. The amount of social security contributions is linked to the wage rate for regular work w – according to the employment conditions before work sharing. Consequently the burden of social security contributions for work sharing depends on the base and the duration of work sharing. With an increasing shortage of work the costs per hour rise disproportionately. So there is no motivation for an excessive use of work sharing for the establishments. Work sharing becomes an option for the employer if the expected costs of work sharing and the residual costs are lower than the expected costs of hiring and firing. The latter costs are higher, the more specific, establishment-related human capital an employee has. Formally we can express this relation as:

M3125 - MESSENGER TEXT.indd 63

14/05/2013 13:33

64

Work sharing during the Great Recession T T E (C ) E (CWS 1 CR) E (CF) H , a 1 a . t t t (1 1 r) t51 t5i (1 1 r) t5j (1 1 r) T

a

The costs of work sharing CWS for every single employee can be calculated exactly. The residual costs CR are quasi-fixed costs of employment and can only be estimated. They are apportioned among all the participants of work sharing. On the right-hand side of the equation we have the expected costs for firing CF which arise from severance payments and costs for trials if an employee sues the establishment for the dismissal. Additionally, the expected costs for hiring new staff have to be taken into consideration. These include costs for recruitment and training on the job. Furthermore, the new employee can be expected to be less productive during the starting phase. Appendix B:

Details on Regression Models

With a series of probit regression models (M1–M3) we estimate the probability of establishments using work sharing2 and its determinants (Table 2A.1). With the same set of explanatory variables we also estimate the intensity of the use of work sharing with truncated regression models (Table 2A.2, M4–M6). The intensity is measured as the proportion of employees involved in work sharing in the first half of the year as a percentage of all employees in the same establishment,3 that is, the dependent variable ranges between 0 and 100 per cent. Our models do not include agriculture or the public sector as work sharing does not play a role in these sectors. The results of the probit estimation for the first half of 2009 are presented under M1 and for the first half of 2010 under M2. We also include another model M3 where we control for whether the establishment had already used work sharing in 2009.4 This information is not available for the previous year. The results of the truncated regressions are presented following the same structure: M4 and M5 present the results for 2009 and 2010, respectively, and M6 additionally controls for the use of work sharing in 2009. Appendix C:

Detailed Regression Results of the Change in Employment during the 2008–09 Crisis

In order to assess the impact of work sharing on employment Bellmann and Gerner (2011b) estimate a difference-in-differences regression with the following form (Meyer 1995): logNit 5b0 1 b1Ci 1b2T2009 1 b3CiT2009 1 xitg 1 eit, with t 5 2008, 2009

M3125 - MESSENGER TEXT.indd 64

14/05/2013 13:33

Work sharing as an alternative to layoffs

Table 2A.1

65

Probit estimations for the probability of work sharing in 2009 and 2010

Dependent variable: Establishment using work sharing in the 1st half of the year (15yes); Establishments with five or more employees in the private sector excluding agriculture 2009 (M1) Employees with fixed-term contracts (yes51) Freelancers (yes51) Temporary agency workers (yes51) Part-time employees (yes51) Working-time accounts (share of employees) Works council (yes51) Collective agreements (yes51) Training measures in previous year (yes51) Vacancies for skilled employees in previous year (yes51) Work sharing in 2009 (yes51) Size of establishment (log. number of employees) Export orientation (export ratio of revenues) Intensity of the crisis 15 low (yes51) 2 (yes51) 3 (yes51) 4 (yes51) 5 5 very strong (yes51) Sectors Food (yes51) Consumer goods (yes51) Industrial goods (yes51)

M3125 - MESSENGER TEXT.indd 65

–0.0238*** (0.009) –0.0246** (0.011) –0.0676*** (0.007) –0.0167 (0.011) 0.0179** (0.009) –0.0040 (0.011) –0.0114 (0.009) –0.0073 (0.010) 0.0146 (0.011)

2010 (M2) –0.0232** (0.009) –0.0233** (0.011) –0.0507*** (0.008) –0.0368*** (0.012) 0.0287*** (0.009) 0.0024 (0.011) –0.0033 (0.008) –0.0157* (0.009) –0.0187 (0.012)

2010 (M3) –0.0288*** (0.009) –0.0018 (0.013) –0.0453*** (0.009) –0.0341*** (0.012) 0.0277*** (0.009) –0.0012 (0.011) –0.0010 (0.008) 0.0012 (0.009) –0.0078 (0.013) 0.3640*** (0.019) 0.0181*** (0.004) –0.0231 (0.020)

0.0473*** 0.0342*** (0.004) (0.004) 0.0463** 0.0021 (0.019) (0.019) (Ref: 0 5 not affected) –0.0323 0.0667* 0.0869** (0.027) (0.039) (0.042) 0.0356* 0.0720*** 0.0681*** (0.020) (0.021) (0.021) 0.1472*** 0.2026*** 0.1495*** (0.015) (0.015) (0.015) 0.2994*** 0.3548*** 0.2336*** (0.022) (0.021) (0.021) 0.4020*** 0.4491*** 0.2870*** (0.030) (0.027) (0.030) (Ref: Services without trade) –0.0628*** –0.0234 0.0052 (0.017) (0.024) (0.027) 0.2737*** 0.1954*** 0.1083*** (0.037) (0.032) (0.029) 0.3344*** 0.2365*** 0.1185*** (0.030) (0.027) (0.024)

14/05/2013 13:33

66

Work sharing during the Great Recession

Table 2A.1

(continued)

Dependent variable: Establishment using work sharing in the 1st half of the year (15yes); Establishments with five or more employees in the private sector excluding agriculture

Investment goods (yes51) Construction (yes51) Trade (yes51) c2 Pseudo-R2 (McFadden) Number of cases

2009 (M1)

2010 (M2)

2010 (M3)

0.3008*** (0.023) 0.3039*** (0.029) 0.0197 (0.014) 1,364.95*** 0.33 6,422

0.2748*** (0.020) 0.3474*** (0.027) 0.0050 (0.013) 1,505.39*** 0.28 7,389

0.1768*** (0.019) 0.2550*** (0.026) 0.0097 (0.013) 2,091.32*** 0.39 7,382

Note: Standard errors in parentheses; */**/*** Indicate significance at the 10%/5%/1% levels, respectively. Marginal effects are presented. Source:

IAB Establishment Panel 2008–2010.

Table 2A.2

Truncated regressions for estimating the intensity of work sharing in 2009 and 2010

Dependent variable: Intensity of work sharing in the 1st half of the year; Establishments using work sharing with five or more employees in the private sector excluding agriculture

Employees with fixed-term contracts (yes51) Freelancers (yes51) Temporary agency workers (yes51) Part-time employees (yes51) Working-time accounts (share of employees) Works council (yes51) Collective agreements (yes51) Training measures in previous year (yes51) Vacancies for skilled employees in previous year (yes51) Work sharing in 2009 (yes51)

M3125 - MESSENGER TEXT.indd 66

2009 (M4)

2010 (M5)

2010 (M6)

0.0025 (0.023) –0.0590* (0.035) –0.1121*** (0.029) –0.0720*** (0.025) –0.0361 (0.023) 0.1502*** (0.029) –0.0355 (0.023) 0.0313 (0.027) –0.0529** (0.025)

–0.0217 (0.028) –0.0907* (0.048) –0.1676*** (0.036) –0.0808*** (0.028) 0.0092 (0.027) 0.1752*** (0.034) –0.0133 (0.027) –0.0511** (0.025) –0.1041* (0.056)

–0.0229 (0.028) –0.0832* (0.047) –0.1665*** (0.035) –0.0842*** (0.027) 0.0139 (0.026) 0.1729*** (0.034) –0.0131 (0.027) –0.0511** (0.025) –0.1099* (0.056) 0.0825*** (0.025)

14/05/2013 13:33

Work sharing as an alternative to layoffs

Table 2A.2

67

(continued)

Dependent variable: Intensity of work sharing in the 1st half of the year; Establishments using work sharing with five or more employees in the private sector excluding agriculture 2009 (M4) Size of establishment (log. number of employees) Export orientation (export ratio of revenues) Intensity of the crisis 15 low (yes51) 2 (yes51) 3 (yes51) 4 (yes51) 5 5 very strong (yes51) Sectors Food (yes51) Consumer goods (yes51) Industrial goods (yes51) Investment goods (yes51) Construction (yes51) Trade (yes51) Constant Sigma Constant c2 Number of cases

2010 (M5)

2010 (M6)

–0.0326*** –0.0837*** –0.0870*** (0.011) (0.014) (0.014) 0.0171 0.0612 0.0505 (0.044) (0.056) (0.056) (Ref: 0 5 not affected) –0.3601** –0.0580 –0.0372 (0.176) (0.130) (0.131) –0.0469 0.0095 0.0153 (0.054) (0.066) (0.067) 0.0119 0.0249 0.0221 (0.031) (0.037) (0.036) 0.0776** 0.1282*** 0.1160*** (0.030) (0.037) (0.037) 0.0920*** 0.1125*** 0.0984** (0.033) (0.041) (0.041) (Ref: Services without trade) 0.2508* 0.3884** 0.3992** (0.148) (0.186) (0.184) 0.3083*** 0.2223*** 0.2035*** (0.063) (0.065) (0.065) 0.3727*** 0.3146*** 0.2930*** (0.056) (0.059) (0.058) 0.3704*** 0.3461*** 0.3294*** (0.054) (0.053) (0.053) 0.2945*** 0.2643*** 0.2489*** (0.058) (0.057) (0.056) 0.2220*** 0.2139*** 0.2071*** (0.062) (0.064) (0.063) 0.3495*** 0.4663*** 0.4511*** (0.067) (0.071) (0.071) 0.2716*** 0.3271*** 0.3248*** (0.006) (0.008) (0.008) 221.36*** 255.95*** 271.57*** 1,002 1,261 1,260

Note: Standard errors in parentheses; */**/*** Indicate significance at the 10%/5%/1% levels, respectively. Source: IAB Establishment Panel 2008–2010.

M3125 - MESSENGER TEXT.indd 67

14/05/2013 13:33

68

M3125 - MESSENGER TEXT.indd 68

14/05/2013 13:33

Independent variables Export ratio of revenues Employees with fixed-term contracts Freelancers Temporary agency workers Part-time employees Share of employees with working-time accounts Works council Collective agreements Training measures Vacancies for skilled employees in previous year Work sharing (previous year) Size of establishment (log) Intensity of the crisis (0 5 not affected) 0.088 0.448 0.116 0.150 0.849 0.494 0.335 0.508 0.746 0.122 3.682

10,744

0.160 0.789

Mean

9,440 10,717 10,629 10,621 10,732 10,653 10,708 10,734 8,796 10,744

10,732 10,548

Number of observations

2009

1.512

0.197 0.497 0.320 0.357 0.358 0.461 0.472 0.500 0.435 0.327

0.367 0.219

Standard deviation

9,273 10,482 10,392 10,398 10,494 10,386 10,473 10,299 8,855 8,831 8,853 10,508

10,494 10,414

Number of observations

0.085 0.449 0.116 0.172 0.861 0.438 0.316 0.464 0.691 0.075 0.161 3.568

0.154 0.068

Mean

2010

0.195 0.497 0.320 0.377 0.346 0.462 0.465 0.499 0.462 0.264 0.368 1.487

0.361 0.199

Standard deviation

Number of observations, means and standard deviations of dependent and independent variables presented in Tables 2A.1 and 2A.2

Dependent variables Use of work sharing in the 1st half of the year Intensity of work sharing in the 1st half of the year

Variable

Table 2A.3

69

M3125 - MESSENGER TEXT.indd 69

14/05/2013 13:33

Source: IAB Establishment Panel 2008–2010.

15 low 2 3 4 5 5 very strong Sectors (reference 5 services) Food Consumer goods Industrial goods Investment goods Construction Trade

0.021 0.070 0.174 0.109 0.058 0.036 0.036 0.074 0.164 0.081 0.200

8,831 8,831 8,831 8,831 8,831 10,744 10,744 10,744 10,744 10,744 10,744

0.185 0.187 0.261 0.371 0.273 0.400

0.143 0.256 0.379 0.311 0.233 10,508 10,508 10,508 10,508 10,508 10,508

10,481 10,481 10,481 10,481 10,481

0.035 0.038 0.076 0.163 0.079 0.199

0.020 0.069 0.172 0.107 0.056

0.183 0.191 0.266 0.369 0.270 0.399

0.141 0.254 0.377 0.309 0.230

70

Table 2A.4

Work sharing during the Great Recession

Estimated effects on the number of employees (Table 2.5, column 1) ln(number of employees)

Crisis Year 2009 Crisis*year 2009 % qualified % part-time Weekly working time Exporting firm Eastern Germany Sectoral collective bargaining Firm-level collective bargaining Works council 9 establishment size dummies 39 sector dummies Number of observations R2

0.175*** –0.004 –0.060*** 0.698*** 0.045 –0.007 0.664*** –0.249*** 0.382*** 0.452*** 1.880*** – *** 13,190 0.523

Note: ***/**/*, Significant at the 1%/5%/10% levels, respectively. Source: IAB Establishment Panel Surveys 2008 and 2009.

where logNit is the logarithmized number of employees in establishment i in year t. Ci is a dummy variable indicating whether the establishment was affected by the crisis or not. T2009 is a time dummy for the year 2009. CiT2009 is an interaction term between the time dummy and the crisis indicator. Therefore, b3 gives the difference in the development of the outcome variable between establishments affected by the crisis and those not affected between 2008 and 2009. Finally, xit is a vector of confounding control variables and [it is an error term. Appendix D:

Technical Details on Labour Hoarding due to Work Sharing

Assuming that the use of work sharing is a suitable predictor for being affected by the economic and financial crisis, our estimations are based on the following linear relationship: Yit loga b 5ai 1g1log (Nit) 1g2 log (Kit) 1g3WS1trt t1(WSt)trtWS 1xrit b1eit. Nit

M3125 - MESSENGER TEXT.indd 70

14/05/2013 13:33

Work sharing as an alternative to layoffs

71

Y The dependent variable log (Nitit) gives the output per employee observed for establishment i in year t. The coefficient ai is a firm-specific fixed effect. Nit is the number of employees and Kit the capital stock (approximated by the investments) for plant i in year t. We therefore assume a Cobb–Douglas production technology for our analysis. WS is a dummy which assumes the value of one if an establishment used work sharing. t9t is a vector of time dummies and t the corresponding coefficient vector. The vector (WSt)9 includes interaction terms between the time dummies and the WS dummy, tWS therefore identifies the difference in the time trends between establishments which use work sharing and those which do not. Moreover, the estimating equation includes a vector x9it of further time-varying control variables measured at establishment level and finally, eit an idiosyncratic error term. After a within transformation g3 is therefore not identified, the equation is estimated by OLS. The standard errors are adjusted post hoc to account for correlated outcomes within establishments by applying a modified sandwich estimator (Hardin and Hilbe 2007); that is, we cluster our standard errors at plant level.

Notes 1. In our model we focus only on labour costs; fixed costs for the establishment are not taken into account. 2. The exact wording of the question in the IAB Establishment Panel Survey is ‘Did you have to make use of short-time work in the first half of 2009?’ (or 2010) with ‘Yes’ and ‘No’ as possible answers. The dependent variable is dichotomous with a value of 0 for ‘no use of work sharing’ and 1 for ‘work sharing was used in the establishment’. 3. The denominator of this fraction is the number of employees as of the reference date of 30 June. 4. Such information is not available for the year 2008.

References Bellmann, L. and H.-D. Gerner (2011b), ‘Reversed roles? Wage and employment effects of the current crisis’, in H. Immervoll, A. Peichi and K. Tatsiramos (eds), Who Loses in the Downturn? Economic Crisis, Employment and Income Distribution, Research in Labor Economics, Vol. 32, Bingley, UK: Emerald, pp. 181–206. Hardin, J.W. and J.M. Hilbe (2007), Generalized Linear Models and Extensions, 2nd edn, College Station, TX: Stata Press. Meyer, B.D. (1995), ‘Natural and quasi-experiments in economics’, Journal of Business and Economic Statistics, 13, 151–61.

M3125 - MESSENGER TEXT.indd 71

14/05/2013 13:33

3.

European diversity of work sharing as a crisis measure: The experiences of Austria, Belgium, France and the Netherlands Jörg Flecker and Annika Schönauer*

1.

INTRODUCTION: WORK SHARING AS A CRISIS MEASURE

The member states of the European Union introduced a range of measures directly or indirectly aimed at supporting the labour market during the Great Recession of 2008 and 2009. These instruments were intended to help avoiding mass dismissals and minimizing the firing and (re)hiring costs for businesses. One type of measure is intended to maintain employment: it has a preventive character that aims to keep people in employment by supporting companies or providing income support for workers who have accepted pay cuts to safeguard their jobs (Messenger 2009). Work sharing as a job-preservation tool falls into this category of measures, and it has to be distinguished from measures to create employment and measures which support individuals in case of redundancy. Work sharing as a policy to combat the consequences of the economic crisis, such as rising unemployment, has become very widespread in Europe and other industrialized countries in recent years and can be described as the most comprehensive and important measure for reducing working hours in Europe in the last decade. The widespread approval for work sharing is related to the fact that it reduces the companies’ labour costs, while at the same time employees with firm-specific knowledge and experience remain available for an economic upswing. For the workers affected, work sharing offers at least the opportunity to temporarily maintain employment and income, even if the reduction in working hours leads to some pay reductions. Apart from this, work sharing temporarily protects against the stigma and social isolation associated with unemployment. In many

72

M3125 - MESSENGER TEXT.indd 72

14/05/2013 13:33

European diversity of work sharing as a crisis measure

73

countries, government financial support has been introduced for employees in work-sharing arrangements or for companies with work-sharing measures. Some of the existing instruments were adapted and expanded in order to deal with the needs of companies and employees in the crisis. This also applies to Austria, Belgium and France, three of the countries researched for this book. Other countries, such as the Netherlands, introduced a new work-sharing measure in the course of the economic crisis (European Foundation for the Improvement of Living and Working Conditions 2009; Arpaia et al. 2010). At first glance the work-sharing measures appear similar, but the countries in reality applied them differently. A glance at the country-specific descriptions of the models makes it clear that there are differences. Austria and the Netherlands, for example, speak of short-time working (Kurzarbeit and Werktijdverkorting) and Belgium and France call it temporary or partial unemployment (chômage temporaire and chômage partiel), respectively. Also, these countries differ in regard to the role the respective measures play in their ensemble of labour market policy instruments. While in Austria and the Netherlands short-time work had always been a marginal phenomenon until the recent crisis, Belgium and France have a long tradition regarding their models of temporary and partial unemployment. But while the French version had fallen into disuse since the beginning of the 2000s and is aimed at overcoming unexpected difficulties in the short term, the Belgian version has continuously been used and is intended to act as a buffer, easing the impact of the business cycle on employment (European Foundation for the Improvement of Living and Working Conditions 2009; Cour des comptes 2011). There are essential differences between the models regarding in certain points: ● ● ●

● ●

the specification of the groups entitled to participate in the programme (which employees, which companies); the preconditions at company level (turnover trends, measures required during short-time working); the type and extent of public financial support (wage subsidy payments are paid directly to the employees in the form of an unemployment payment which then supplements employees’ wages for the reduced working hours or is paid as a subsidy to the employer); the length and period that support can be provided; and whether or not the measure is used in combination with training.

Table 3.1 shows the essential elements of the measures in Austria, Belgium, France and the Netherlands (Council of the European Union 2009; Deeke

M3125 - MESSENGER TEXT.indd 73

14/05/2013 13:33

74

M3125 - MESSENGER TEXT.indd 74

14/05/2013 13:33

Temporary unemployment (chômage temporaire)

Partial unemployment (chomage partiel)

BE

F

All employees, including parttime workers, temporary

All employees, including temporary agency workers, except apprentices, CEOs and board members; other possibilities have to be exhausted first, such as use of overtime and holiday entitlements, or workingtime accounts Employees, including temporary agency workers and workers on fixed-term employment contracts; in some cases only blue-collar workers The employer pays 75% of the hourly gross wage for non-

Complete unemployment for a maximum of 1 month, partial unemployment for a maximum of 12 months Six consecutive weeks, in total 1,000 hours

6 months, possibility of extensions by additional max. 6 months, maximum 24 months, if application is accepted before 12/2010

Short-time working allowance equivalent to unemployment benefit (about 55% of the net wage) plus sickness and retirement insurance and family allowances if applicable. Support is increased by 15% if the employer offers training Maximum 75% of the previous income, depending on family status, up to a gross monthly wage of €2,206.46

Short-time work (Kurzarbeit)

AT

Duration

Type and extent of support

Name

Country

Eligibility

Measures of short-time work in Austria, Belgium, France and the Netherlands

Table 3.1

M3125 - MESSENGER TEXT.indd 75

Short-time work (Werktijdverkorting) until the end of March 2009, partial unemployment thereafter

worked hours at a minimum of €6.84 an hour; the state pays SMEs €3.84 and large companies €3.33 per hour and worker. In cases of a company closure of more than 3 months, employees are entitled to unemployment benefits 70% of the wage for the nonworked hours

Maximum of 15 months

per employee a year (can be prolonged by an additional 6 months)

Sources: Council of the European Union (2009); Deeke (2009); European Foundation for the Improvement of Living and Working Conditions (2009); Jørgensen (2009).

NL

agency workers and those on fixed-term employment contracts, if their employer has a decrease in sales of at least 50% in the case of short-time work All employees in case of temporary layoff No temporary agency workers; employers have to undertake training of the staff on short-time work

.

75

14/05/2013 13:33

76

Work sharing during the Great Recession

2009; European Foundation for the Improvement of Living and Working Conditions 2009; Jørgensen 2009). In all the measures outlined in the table, amendments of existing laws were made during the course of the crisis. The chapter proceeds as follows. The experiences of Austria, Belgium, France and the Netherlands are dealt with in Sections 2–5. Section 6 concludes.

2.

KURZARBEIT IN AUSTRIA

Short-time work in Austria is based on a social partner agreement and means a temporary reduction of normal working time (Normalarbeitszeit). It was first introduced in 1934, and since 1949 it has existed as an instrument of labour market policy in Austrian law, but has played only a marginal role until recently. It was initially designed as an instrument to avoid layoffs in cases of unforeseen and temporary declines in production due to incidents such as interruptions of transport lines, accidents, strikes in supplying plants, or natural disasters. Its purpose therefore is solely the preservation, and not the creation, of jobs (Fuchs 2009). All employers except the state (Bund) and its federal states (Bundesländer), municipalities (Gemeinden), associations of municipalities (Gemeindeverbände) and other corporate bodies under public law, are entitled to the measure of short-time work as long as their employees are affected by a reduction of working time which causes a loss of earnings. While initially the eligibility for Kurzarbeit was restricted to senior bluecollar workers, reforms between 2000 and 2008 successively expanded its applicability to all employees, including temporary agency workers. Apprentices, CEOs and board members are exempted from short-time work (AMS Österreich 2009a; BMASK 2009). Short-time work is approved for six months at first, but can be prolonged for six months at a time, up to a maximum of 18 months. As a measure to combat the recent crisis, enterprises commencing short-time work until the end of 2010 may provisionally even extend it up to 24 months (AMS Österreich 2009a; BMASK 2009). There are several conditions a company has to fulfil to be entitled to introduce short-time work. The organization has to suffer from temporary economic difficulties due to factors external to the enterprise. This does not include seasonal effects or ‘normal’ variations of natural factors, for example, bad weather in the construction sector. The Austrian Employment Service (AMS) has to be informed about economic difficulties six weeks before the implementation of short-time work and four

M3125 - MESSENGER TEXT.indd 76

14/05/2013 13:33

European diversity of work sharing as a crisis measure

77

weeks before its prolongation (this is not necessary in the case of a natural disaster). A formal application must be submitted at least three weeks before the start or prolonging of short-time work. All other options for support have to be considered before and discussed among the employer, works council and the ‘social partners’ (that is, trade unions and the employers’ federation). The reduction of working time during the period of short-time work has to be at least on average 10 per cent and not more than 90 per cent of the normal working time, or, in the case of part-time work, of the defined working time. Additionally there has to be an agreement between the social partners on the specifications of short-time work, in particular regarding the scope, the timeframe, and the preservation of the employment relationship during short-time work and afterwards (Behaltefrist). The social partners may also agree on further training and qualification during short-time work. For every reduced hour the employer pays Kurzarbeiterunterstützung to the employee. If some time is used for training he/she pays Qualifizierungsunterstützung instead. The AMS replaces the employers’ costs for Kurzarbeiterunterstützung and Qualifizierungsunterstützung with an hourly flat rate. The hourly flat rate for Kurzarbeiterunterstützung is based on the unemployment insurance expenses for unemployment benefits and social insurance contributions that would arise instead. The normal working time (Normalarbeitszeit), the monthly gross wage plus social insurance contributions before short-time work and the number of children are decisive factors (Table 3.2). The hourly flat rate for Qualifizierungsunterstützung (Table 3.3) contains a supplement of 15 per cent for additional expenses due to training measures. The employer has to pay increased social insurance contributions; these contributions are paid by the AMS after six months of short-time work. Before 2008, the main reasons for using Kurzarbeit were flooding in Table 3.2

Example of flat rate of Kurzarbeiterunterstützung for a fulltime employee (40 hours per week), paid by the AMS to the employer

Monthly gross wage incl.special payment From 2,071.48

To 2,109.99

Flat rate for each cancelled working hour: Kurzarbeiterunterstützung 0 1 children child 7.55

7.80

2 3 4 5 or more children children children children 8.04

8.28

8.52

8.76

Source: AMS Österreich (2009b).

M3125 - MESSENGER TEXT.indd 77

14/05/2013 13:33

78

Work sharing during the Great Recession

Table 3.3

Example of flat rate of Qualifizierungsunterstützung for a fulltime employee (40 hours per week), paid by the AMS to the employer

Monthly gross wage incl.special payment From

To

2,071.48

2,109.99

Flat rate for each cancelled working hour: Qualifizierungsunterstützung 0 1 children child 8.77

9.05

2 3 4 5 or more children children children children 9.33

9.61

9.90

10.18

Source: AMS Österreich (2009b).

Table 3.4

Persons receiving short-time work subsidies, cumulated, by year in Austria

Male Female Total Costs (€, total) Sources:

2007*

2008

2009

n.a.n.a.782* 388,000*

7.088 1.067 8.155 1,008,000*

49.374 10.066 59.440 250,000,000**

AMS Österreich (2010); *BMASK (2009);**federal budget estimates.

1997, 2002 and 2006, as well as the negative effects of the 9/11 attacks on the tourism industry in 2002. The number of benefit recipients varied between 480 (2004) and 5,746 (2002) per year between 1997 and 2007. The onset of the crisis in 2008 led to the reforms mentioned above, which transformed the rarely used instrument of Kurzarbeit into one of the main pillars of the government’s strategy to combat unemployment. Subsequently, the number of employees benefiting from the programme rose to about 8,000 in 2008 to be followed by an explosion to about 60,000 in 2009 (BMASK 2009: 51; BMASK 2010b; AMS Österreich 2010). Due to the structure of the economy, the financial crisis affected (besides the financial branch) mainly the industrial, export-oriented sector. Since industrial labour in Austria is still heavily dominated by male workers, around 80 per cent of the workers in that sector registered for short-time work in 2009 were male (see Tables 3.4–6). Figure 3.1 gives an overview of the number of workers who were preregistered for short-time work and those who actually were in short-time work between October 2008 and February 2010. The figure shows the

M3125 - MESSENGER TEXT.indd 78

14/05/2013 13:33

European diversity of work sharing as a crisis measure

Table 3.5

79

Persons receiving training while on short-time work subsidies, cumulated, by year in Austria

Male Female Total

2007*

2008

2009

n.a. n.a. n.a.

3 4 7

2.271 266 2.537

Source: AMS Österreich (2010).

Table 3.6

Persons receiving short-time work subsidies in 2009, by age in Austria

,25

26–50

.50

6,210

43,163

11,088

Source: BMASK (2010a).

56,728

60,000

53,911

56,626

53,181

51,671

50,000

47,158 38,937

40,000

35,904

35,135 33,481

37,368

29,292

30,000

36,079 33,326

22,411

29,384

35,013

26,425

33,049 26,169

20,000

21,467

22,257

15,573 8,957

10,000

11,338

O ct . N 08 ov .0 D 8 ec .0 8 Ja n. 0 Fe 9 b. 0 M 9 ar .0 9 A pr .0 M 9 ay 0 Ju 9 n. 09 Ju l. A 09 ug .0 9 Se p. 09 O ct . N 09 ov .0 D 9 ec .0 9 Ja n. 10 Fe b. 10

0

739 5,705 578 4,296 7,287

Pre-registered workers

De facto subsidized workers

Source: BMASK (2010b).

Figure 3.1

Pre-registered and de facto subsidized short-time workers in Austria

M3125 - MESSENGER TEXT.indd 79

14/05/2013 13:33

80

Work sharing during the Great Recession

number of workers in short-time work in a given month. The monthly values do not add up to the total number of workers affected because short-time working often takes longer than one month. Conversely, the cumulative annual figure in Table 3.4 is not reached in any single month. From November 2009 on, no data on de facto subsidized workers in shorttime work were available. The duration and extent of short-time work varied strongly during the course of 2009. The peak was reached in April with 56,728 employees pre-registered for short-time work. By December, 35,135 employees from 253 enterprises remained listed. The average proportion of de facto subsidized short-time workers of all pre-registered workers was 64 per cent. It was highest in December 2008 (81 per cent) and lowest in March 2009 (46 per cent). On average a company using this measure had 118 workers in short-time work, which reduced working time on average by 25 per cent. Some 20 per cent of all employees were involved in training measures during short-time work within the illustrated period (BMASK 2010a). Based on these assumptions, the Austrian Institute of Economic Research estimated a total of 8,400 jobs preserved by short-time working. Including qualification measures such as Bildungskarenz, the estimated reduction of unemployment would amount to 12,700 jobs saved (Mahringer 2009). An estimation by the Federal Ministry of Labour, Social Affairs and Consumer Protection (BMASK) shows that a third of all jobs saved through anti-crisis labour market measure are due to short-time working, which is assumed to have resulted in approximately 30,000 jobs preserved (BMASK 2010b: 304). However, some other evaluations show much lower employment impacts: Stiglbauer (2010) assumes that 26 per cent of all workers on short-time working would have lost their job otherwise, and therefore calculates an employment effect of nearly 6,800 jobs. The expenditures for short-time work were negligible before 2008. For 2009, BMASK expects costs of roughly €250 million. These costs were part of the stimulus packages to combat the financial crisis (Schratzenstaller 2009: 589). However, the estimated budget for short-time work was not fully used. To be on the safe side, companies tended to pre-register more people for short-time work than those who actually used the scheme. This concerns some large companies, especially those that pre-registered several thousand workers and actually used short-time work only modestly or not at all. It can be assumed that the negative job impact of the crisis was expected to be much stronger and its duration much longer. In particular, the rather quick recovery in Austria meant that the work-sharing measure was used to a lesser extent than expected, and that it was highly effective where it was adopted.

M3125 - MESSENGER TEXT.indd 80

14/05/2013 13:33

European diversity of work sharing as a crisis measure

3.

81

CHÔMAGE TEMPORAIRE IN BELGIUM

Like most European countries, Belgium had already established a model of short-time working a long time before the current global economic crisis: temporary unemployment for blue-collar workers. Compared to the other countries analysed in this chapter, the Belgian chômage temporaire was the most widely used programme. Temporary unemployment means that the work duty is reduced or temporarily suspended while the worker remains employed by the company and is entitled to a wage replacement payment, the temporary unemployment benefit. The replacement rate is 70 or 75 per cent of the previous gross wage up to a maximum level, depending on the family situation. Temporary unemployment may be justified by technical accidents, force majeure or economic reasons. The Belgian model of short-time work has for a long time provided companies with an inexpensive form of numerical flexibility, making it possible to adapt employment to fluctuations of the company’s turnover or orders. The preconditions for temporary unemployment are not specified in any detail by law. This is why the measure was well established and heavily used in Belgium even before the crisis. The system is also called ‘partial unemployment’ because workers alternate between days or weeks of work and days or weeks of unemployment. The unemployment spell is limited to four continuous weeks which, however, can be repeated after one week of work. The measure is very flexible because the employer has to notify the public employment agency only seven days in advance – no approval is needed. Up to 2009, temporary unemployment for economic reasons was limited to blue-collar workers, students doing manual work and temporary agency workers (Conseil Supérieur de l’Emploi 2010). In reaction to the crisis, the model was adapted. First, the wellestablished temporary unemployment scheme was modified by increasing the maximum amount of income-related transfer payments in cases of temporary unemployment for blue-collar workers. It was raised from €1,250 to €1,650 per month. In addition, the worker receives a premium of €200 a month from the employer. Second, new forms of short-time working were introduced which include white-collar workers who had not previously been entitled to participate in temporary unemployment schemes. These measures were agreed on in April 2009 and applied at first for the period between 1 July 2009 and 31 December 2009. An extension until 30 June 2010 was passed by parliament in December 2009. The measures were preceded by negotiations between the social partners (trade unions and the employers’ federation) concerning the extension of the temporary unemployment scheme to white-collar workers. However, the

M3125 - MESSENGER TEXT.indd 81

14/05/2013 13:33

82

Work sharing during the Great Recession

negotiations collapsed and, as a reaction, new and separate models for blue- and white-collar workers were introduced directly by the government (European Foundation for the Improvement of Living and Working Conditions 2009; Van Gyes 2009a). Even before the introduction of the new models, companies took recourse to other schemes to be able to reduce working hours in particular of white-collar workers in order to safeguard employment. The ‘time-credit scheme’ – which is intended to improve work–life balance by supporting individual working-hour reductions – was widely used as an anti-crisis measure at company level (Van Gyes 2009b). The government initiative was based on both the time-credit policy and the temporary unemployment scheme, and further introduced three new forms of shorttime working in 2009 (European Foundation for the Improvement of Living and Working Conditions 2009; Service publique fédéral Emploi 2009; Conseil Supérieur de l’Emploi 2010): ●

●

●

Temporary collective reduction of working hours This measure is open to all companies. The company is supported by a reduction in the social security contributions. This reduction is increased if short-time working is accompanied by the introduction of a fourday week. Temporary individual reduction in working hours (‘crédit-temps de crise’) If the company is in difficulty, that is, suffers a reduction of turnover of 15 per cent (which was the rate from 1 January 2010, but was previously set at 20 per cent), and is covered by a collective agreement, the workers are entitled to a larger allowance as compared to the ‘normal’ time-credit scheme. Temporary or economically conditioned unemployment for whitecollar workers Companies in difficulty may also use the temporary unemployment scheme for white-collar workers if they reduce the working hours by at least two days a week. The duration is limited to 16 weeks in the case of a full suspension of work duties or to 26 weeks in the case of a partial suspension.

The white-collar unions viewed the introduction of the new models as a success, as it made it possible to preserve their white-collar status. In addition it was possible to agree a guarantee that the measures were extended to 2010 at the latest. However, it was expected that significant advances in the harmonization of the employment status of white- and blue-collar workers would be made (Van Gyes 2009a). See Table 3.7 for an overview of the new anti-crisis employment measures. In the course of the crisis, the number of workers in temporary

M3125 - MESSENGER TEXT.indd 82

14/05/2013 13:33

83

M3125 - MESSENGER TEXT.indd 83

14/05/2013 13:33

Company collective agreement

Monthly compensation paid by the employer for reduced

Employee compensation

For all employees or a specific category of employees

Conditions

Procedure

Reduction in working time by one-fifth or a quarter

Temporary collective reduction of working time Based on existing time-credit scheme – parties decide to reduce working time by mutual agreement Company must be in financial difficulties; working-time reduction by one-fifth or a half for a minimum period of 1 month and up to a maximum period of 6 months; normal conditions of the timecredit scheme do not apply – temporary reduction of working time is not considered an employee’s right in this case; applicable to full-time employees only Consent required by employee in the form of an agreement between the worker and the employer; collective agreement at sectoral level before 1 June 2009; if none exists, collective agreement at company level or a company plan approved by an ad hoc commission (the latter must contain measures for maintaining a specified level of employment); several further technical conditions Allowance paid by the National Employment Office (Rijksdienst voor

Temporary individual reduction of working time

Overview of the new anti-crisis employment measures in Belgium

Measure

Table 3.7

Compensation paid by the Auxiliary Unemployment Benefits Fund

Company in financial difficulties; white-collar workers (measure already exists for blue-collar workers); applicable to a certain number of employees and only after exhausting their recuperation days; total (during all days of the week) or partial (at least keeping two working days a week) suspension of the employment contract for a period of at least 1 or 2 weeks and up to a maximum of 16 or 26 weeks a year Collective agreement at sectoral level before 1 June 2009; if none exists, company collective agreement or company plan approved by an ad hoc commission – both must provide information on the employment level maintained, the duration of suspension and the fee to be paid by the employer; unemployment office and employee must be informed seven days prior to implementation

Collective, total or partial suspension of employment contract

Temporary unemployment

84

M3125 - MESSENGER TEXT.indd 84

14/05/2013 13:33

During working-time reduction, the employer benefits from a reduction in social security contributions: by €600 a quarter if working time is reduced by one-fifth, by €750 a quarter if working time is reduced by a quarter; these amounts are increased by €400 if a 4-day working week is implemented

No provision

Arbeidsvoorziening/Office National de l’Emploi); working-time reduction: by one-fifth €188 (€248 when the worker is aged 50 years or over), by half €442; the employer can pay an additional compensation; however, this is not compulsory; limit: the wage and allowance combined must remain below the employee’s previous full-time wage

wages in the case of working-time reduction: by one-fifth – a minimum of €150; by a quarter – minimum of €187.50; limit: the wage and compensation combined must remain below the employee’s last full-time wage; these amounts are increased by €100 if a 4-day working week is implemented

Source: Van Gyes (2009a).

Employer benefit

Temporary individual reduction of working time

Temporary collective reduction of working time

Table 3.7 (continued)

(Caisse auxiliaire de paiement des allocations de chômage/Hulpkas Voor Werkloosheidsuitkeringen, CAPAC/HVW) for each day of suspension amounting to: 70% if cohabiting, to 75% (if single or main family breadwinner) of the worker’s gross wage which is limited to a maximum of €2,206 a month; the employer can pay an additional compensation that is equal to the compensation granted by an employer to blue-collar workers receiving unemployment benefits in the case of a total suspension of an employment contract due to economic difficulties No provision

Temporary unemployment

European diversity of work sharing as a crisis measure

Table 3.8

85

Number of temporary unemployment subsidies, yearly averages in Belgium 2007

Male Female Total Days spent on partial unemployment, total Costs, total €

2008

2009

97,738 22,211 119,949 9,365,303

108,362 26,375 134,737 10,132,569

172,578 38,287 210,865 18,905,846

381,410,369

430,637,359

1,052,999,323

Source: ONEM (2010).

Table 3.9

Number of temporary unemployment subsidies, yearly averages, by age in Belgium

, 25 a 25–49 a $ 50 a Total

2007

2008

2009

14,464 86,074 19,411 119,949

16,040 96,128 22,568 134,736

20,854 150,217 39,794 210,865

Source: ONEM (2010).

unemployment nearly doubled: from 120,000 in 2006–07 to 210,000 in 2009 (yearly average) (Table 3.8). At its peak in March 2009, 313,000 workers received subsidies for partial unemployment (Table 3.9). The number of white-collar workers receiving temporary unemployment benefits remained limited, reaching its peak of 9,000 in November 2009 (Conseil Supérieur de l’Emploi 2010: 30). The most important increase of temporary unemployment was registered in the manufacturing industry, which in August 2009 accounted for half of all workers in temporary unemployment. The vast majority of temporary unemployment benefit recipients were male: in 2009, on average, nearly 172,600 men and only 38,300 women were covered by the scheme. During 2009, the average duration of partial unemployment was about 89 days. The costs of the programme during the year surpassed €1 billion, more than doubling compared to 2008. In regional terms, the take-up of temporary unemployment corresponds with the size of the total population of Flanders (about 6.1 million) and Wallonia (about 3.5 million), although the general unemployment rate in

M3125 - MESSENGER TEXT.indd 85

14/05/2013 13:33

86

Table 3.10

Work sharing during the Great Recession

Number of temporary unemployment subsidies, yearly averages, in Belgium by region

Région flamande Région wallonne Région Bruxelles-Capitale Total

2007

2008

2009

74,956 40,155 4,838 119,949

85,991 43,685 5,060 134,736

138,939 64,802 7,124 210,864

Source: ONEM (2010).

Table 3.11

Daily rates of temporary unemployment subsidies, per person, in € in Belgium

Family breadwinner Single Single, living in multi-person household

Min.

Max.

38.75 32.65 24.40

63.65 63.65 59.40

Source: ONEM (2010).

Table 3.12

Individual working-hour reduction (Crédit-temps – reduction of work duties)

Year

Number (average)

Days (total)

Costs (€, total)

2007 2008 2009

100,261 108,526 118,740

31,152,265 33,733,488 36,901,209

276,555,120 312,510,372 349,213,488

Source: ONEM (2010).

Wallonia is more than twice as high as in Flanders. Only the region of the capital, Brussels, is underrepresented. (Table 3.10). As described in Table 3.11, the remuneration during partial unemployment varies in line with salary, household size and the number of children. The table shows the differences by family situation in the daily rates of temporary unemployment benefits. The take-up of the temporary individual working-hour reduction via the time-credit scheme also increased during the crisis (Table 3.12).

M3125 - MESSENGER TEXT.indd 86

14/05/2013 13:33

European diversity of work sharing as a crisis measure

87

However, this was the case only for the reduction of work duties, while the numbers of sabbaticals decreased between 2007 and 2009. In December 2009 the new measure crédit-temps de crise was used by 2,574 workers (Conseil Supérieur de l’Emploi 2010: 31). Thus, the Belgian experience exhibits a widespread take-up of both well-established and adapted and newly introduced anti-crisis measures. In particular, the temporary unemployment scheme (chômage temporaire) was used by up to more than 300,000 workers. In 2009, the costs for this type of short-time working exceeded €1 billion. Negotiations regarding the extension of temporary unemployment to also cover white-collar workers had failed in 2009. As a consequence, the government introduced anticrisis measures for white-collar workers based on the time-credit scheme on the one hand and temporary unemployment on the other. Compared to other countries, there is now a wider variety of options in Belgium for individual and collective reductions of working time aiming at the protection of employment. As far as the employment effect of short-time working in Belgium is concerned, Hijzen and Venn (2011: 34) calculated that ‘the decline in permanent employment from the start of the crisis to the end of 2009 Q3 was about 55,000 jobs less (1.6 per cent of total employees) in Belgium than what it would have been in the absence of the STW scheme’. However, taking into account the intensity of use of short-time working before the crisis, the net impact during the crisis was only 14,000 jobs according to their estimates (ibid.). Yet, the average jobs impact they show is 42,600, and thus by far the highest number of jobs preserved among the countries covered in this chapter.

4.

CHÔMAGE PARTIEL IN FRANCE

In France, the measure to prevent loss of jobs in a severe downturn of the business cycle is called chômage partiel (partial unemployment). It can also be used under other exceptional circumstances such as interruptions in the provision of raw materials or energy, disasters or the restructuring of a company. The reduction or suspension of activities has to be both temporary and exceptional. If companies, for such temporary reasons, reduce the level of activity or suspend their activities altogether, the partial unemployment scheme partly compensates the loss of the employees’ salary. The measure thus makes it possible for the company to maintain employment in the case of temporary closures of the company (or parts of it) or by means of reducing the working hours of the workforce or parts of it.

M3125 - MESSENGER TEXT.indd 87

14/05/2013 13:33

88

Work sharing during the Great Recession

On the basis of an agreement by the social partners (trade unions and the employers federation) in November 2008, the conditions for partial unemployment were changed considerably at the beginning of 2009. From then on, workers received 60 per cent of their normal wages for the hours not worked (in contrast to 50 per cent previously). The state pays the company an hourly rate (called an allocation spécifique de chômage partiel) for the hours not worked, which was increased from €244 to €384 for small and medium-sized enterprises (SMEs) and from €213 to €333 for larger companies. Since April 2009, employees’ compensation for hours not worked increased again to a rate of 75 per cent of normal wages, or a minimum of €6.84 per hour, paid by the employer. The costs for the company can also be reduced by further government support depending on the gravity of the economic situation, the number of layoffs avoided and efforts undertaken regarding working-hour reduction or flexibilization. The government support can amount to 50, 80 or, in exceptional cases, 100 per cent of the payments covered by the employer. Furthermore, the worker receives an allowance paid by the company to bring his/her income above the monthly minimum wage. This allowance could also be partly reimbursed by the state. After six weeks of total suspension of business activities, the workers on partial unemployment receive the normal unemployment benefits even if they are not laid off. However, this situation is limited to six months, after which workers have to be made redundant. Workers may increase their income by taking on additional employment and still receive the partial unemployment benefit. The total number of hours for which payments can be made was also increased from 600 to 800 hours per employee during the crisis; for the textile, clothing, leather and automotive industries including the subcontractors, the limit was set to 1,000 hours per employee in January 2009. In September 2009, the regulations for the latter were expanded to all private economic sectors. Within this quota, the support for partial unemployment can be used over a period of six months, which can be renewed once resulting in one year in total. In the case of a closure of the company, the support is paid for a maximum of six weeks. If a company is not in a position to fully take up its activities after the period of partial unemployment and, as a consequence, layoffs are necessary, severance pay is calculated on the basis of the months of full activity. Partial unemployment can be linked with training measures in different ways. One option for the workers is to use training leave (congé individuel de formation), which provides them with a remuneration of 80 per cent of their net salary, a level higher than the partial unemployment allowance. A temporary work agency may also use the measure of partial unemployment, provided that

M3125 - MESSENGER TEXT.indd 88

14/05/2013 13:33

European diversity of work sharing as a crisis measure

Table 3.13

Persons receiving partial unemployment benefits, by year 2007

Persons (31000) Rate in % of employed persons

89

2008

2009

QI

QII QIII QIV

QI

QII QIII QIV

85

92

33

98

0.3

0.4

85 0.3

51 0.2

0.1

44

0.4

0.2

132 0.5

QI 182

QII QIII 258

0.7

1.0

143 0.6

Source: INSEE (2010).

the business relationship with the employer company of its worker is at risk. Partial unemployment was used widely in France before 1996, when it peaked at 1.8 million employees spending a total of 12 million workdays in the programme. With the introduction of the 35-hour week in the late 1990s, the numbers declined significantly to around 60,000 employees per year in the early 2000s. With the financial crisis, the numbers rose again from 44,000 employees in the third quarter of 2008 to 258,000 in the second quarter of 2009, declining again to 143,000 in the third quarter of 2009 (Table 3.13). At its peak in 2009, partial unemployment affected 1 per cent of the total workforce. In 2009, 77.6 million hours were spent on partial unemployment, with estimated costs of €319 million (Chef d’entreprise 2010; Cour des comptes 2011). Interestingly, the hours of partial unemployment demanded by the companies and accepted by the government were much higher than the actual take-up in particular in the year 2009. Subsidies were actually used for only 30 per cent of the ‘authorized’ hours (Table 3.14). The debate on the job-preservation effects of partial unemployment Table 3.14

Hours of partial unemployment ‘authorized’ and paid (in million hours)

‘Authorized’ hours Paid hours Share of actual use (%) (hours paid / hours authorized)

2007

2008

2009

2010 (Jan–May)

8 4 50

35 16 46

258 78 30

87 30 34

Source: Cours des comptes (2011: 164).

M3125 - MESSENGER TEXT.indd 89

14/05/2013 13:33

90

Work sharing during the Great Recession

in France is controversial. Analyses using data from previous recessions paint a rather negative picture. For example, Calavrezo et al. (2009) investigated the relationship between the short-time compensation programme and the redundancy behaviour of companies between 1996 and 2004. The authors concluded that participation in the programme did not provide protection from redundancies. This conclusion is based on findings showing that companies with short-time work were more likely to make workers redundant than those that did not participate in such a programme, controlling for profitability. It could mean that both shorttime work and redundancies are used to face economic difficulties. Thus, it is possible that short-time work is a policy for establishments in structural decline or that companies use short-time work to calm down social tensions before a planned redundancy scheme. In our view, however, no final conclusions can be drawn from this comparative analysis regarding the job-preservation effects within the group of establishments that took part in the programme. Regarding the crisis from 2008 and 2009, observers seem to agree that the use of the partial unemployment scheme in France was rather modest in particular if compared with countries such as Germany and Belgium. According to a peer review paper in the Mutual Learning Programme 2010 ‘the number of non-worked hours effectively subsidised corresponded to 50,000 full-time equivalent jobs in 2009’. The reason why French employers used partial unemployment very little compared to employers in other countries was that they mainly chose other measures to adapt their organizational needs: the non-renewal of fixed-term contracts and a sharp decrease in the use of temporary agency work (Wilthagen 2010: 4). This view is shared by the Audit Office of the French government, which argued in a report that the partial unemployment scheme was insufficiently used during the crisis. With 275,000 workers, only 0.8 per cent of the active population benefited from the measure, while in Germany the share was 3.2 per cent and in Belgium 5.6 per cent. As a consequence, the employment impact was only 18,000 jobs saved in contrast to 221,500 jobs saved in Germany (Cour des comptes 2011). The report gives various reasons for the rather low take-up: the delays caused by the time needed for the reactivation and modernization of the measure; regulations that are less attractive to companies as compared to other countries; and the evolution of the labour market. The last point is of particular interest to us and gives another explanation why the French model of partial unemployment was not as successful as in other countries. Since the end of the 1990s, France has strongly increased the flexibility of working hours in the context of the 35-hour week. The annualization of working hours and the potential use of external flexibility offered by fixed-term contracts

M3125 - MESSENGER TEXT.indd 90

14/05/2013 13:33

European diversity of work sharing as a crisis measure

91

and temporary agency work provided a substitute for the work-sharing measure of partial unemployment (ibid.: 169f.).

5.

WERKTIJDVERKORTING AND DEELTIJDWERKLOOSHEIDSWET (WW) IN THE NETHERLANDS

The policy rules on short-time working (Werktijdverkorting) in the Netherlands were originally not intended for economic crises. However, the financial crisis triggered such a rapid and unexpected decline in the real economy that some companies found themselves facing dramatic decreases in demand from one day to the next. In order to spare such companies from having to make hasty decisions about downsizing their workforces amidst rapidly changing circumstances, on 30 November 2008 the Dutch government decided to subsidize short-term working for companies whose turnover dropped by 30 per cent or more in a period of two months (Extended Special Policy Rules regarding the Prohibition on Short-Time Working 2008; Council of the European Union 2009). This provision was seen as a way of preventing companies from overreacting to the crisis by dismissing personnel who would be needed once demand rebounded. Employers could apply for a six-week period of Werktijdverkorting, which could be extended three times up to a total of 24 weeks. Other characteristics of the special policy rules were, for example, several obligations with respect to training, continued payment of wages, and the obligation to maintain the contract until four weeks after the end of the short-time working period. The employees received compensation from unemployment insurance for the reduced working hours (Werkloosheidswet). Employees were guaranteed four weeks’ protection against dismissal following the short-time working period (ibid.). However, these measures were available at the cost of eligibility for unemployment benefits in the case of unemployment, which has always been heavily criticized by the trade unions. In December 2008, the three trade union confederations and the employers’ associations asked the Minister for Social Affairs and Employment to soften the rules for short-time working. The deadline for applications was extended, first to 15 January and then to 1 March 2009. In February, the Minister for Social Affairs and Employment announced a further extension and an increase of the budget earmarked for short-time working (Collective Bargaining Newsletter 10/2008, 1/2009 and 2/2009). This short-time working arrangement expired on 21 March 2009. At the height of the economic crisis in 2009, based on the view that a new phase in the economic crisis demanded new measures, a new scheme

M3125 - MESSENGER TEXT.indd 91

14/05/2013 13:33

92

Work sharing during the Great Recession

called part-time unemployment (Deeltijd-Werkloosheidswet) came into force. Werktijdverkorting was used to prevent companies from overreacting to the economic crisis, but the objective of partial unemployment is different. Part-time unemployment is meant to preserve the viability of companies, by enabling them to keep employing the workers that are essential for the company when demand starts growing again. In partial unemployment the regulations regarding the drop in turnover (eligibility criteria for short-time working) were omitted, and the eligibility was instead tied to consensus between the company and employees/labour unions about the necessity of short-time working for that particular company. Also, part-time unemployment permitted companies to reduce the working hours of employees, and affected employees could apply for partial unemployment benefits. Agency workers or workers whose contracts would be terminated during partial unemployment were excluded from participating in the scheme. The working hours of the employees could be reduced by a minimum of 20 per cent and a maximum of 50 per cent, for at least 26 weeks. The maximum duration of part-time unemployment is dependent on the number of employees for which it is requested. If the share of employees for which it is requested exceeds 60 per cent of the employers’ total staff, the maximum duration is 39 weeks; if the share is between 30 and 60 per cent the maximum duration is 52 weeks; and a share of less than 30 per cent of employees relates to a maximum duration of 65 weeks. This condition implies that firms have to consider carefully whether they really need part-time unemployment, because the inactivity of their employees cannot be reversed immediately. As compensation, the employees receive 70 per cent of the normal wage for the reduced part of the working hours. The employees thereby lose the remaining 30 per cent, that is, 15 per cent of their total earnings. Companies can decide to supplement the compensation up to 100 per cent of earnings. Apart from this, hours spent in Deeltijd-Werkloosheidswet are subtracted from the accumulated entitlement to unemployment benefit (Grünell 2009). The part-time unemployment scheme met some resistance from unions. Some unions demanded that employers should supplement workers’ unemployment benefits to 100 per cent of the former wage. In many company agreements, trade unions obtained additional payments from employers (Collective Bargaining Newsletter 4/2009 and 5/2009). The scheme was initially introduced as a temporary measure and later on was extended several times. Financial constraints played a role. The Minister for Social Affairs and Employment announced a halt of the measure because the total appropriation of €375 million had been spent. Following protests, an extension of the scheme was announced, however with tighter conditions

M3125 - MESSENGER TEXT.indd 92

14/05/2013 13:33

European diversity of work sharing as a crisis measure

93

(Collective Bargaining Newsletter 6/2009). Nevertheless, between 23 June and 20 July 2009 employers could not apply for partial unemployment. The trade unions demanded additional compensation from employers in order for the workers to achieve a 100 per cent wage compensation, which received a frosty reception from employers. In addition they demanded that the companies guarantee employees at least six months’ protection against redundancy following the short-time working. This demand went beyond the government’s requirement of three to five months’ protection against redundancy. Unions also demanded that the periods of unemployment be credited towards workers’ pension rights (Grünell 2009). Owing to this the Ministry for Social Affairs and Employment set up a telephone hotline to help inform and mediate between the parties at company level. Disputes were referred to the Labour Foundation (Stichting van de Arbeid, STAR); if no solution could be found there, then the minister decided. STAR has also developed guidelines for the new scheme. These said that works councils cannot refuse to reduce working hours solely on the grounds that the company does not supplement the compensation up to 100 per cent of the wage. In response, companies were not allowed to rule out additional payments in the run-up to negotiations. The question as to whether or not additional payments were to be made may thus not be a precondition for negotiations (ibid.). In Werktijdverkorting and Deeltijd-Werkloosheidswet it is obligatory that participating workers should be involved in training measures. Employers can claim special financial assistance with training if they offer their employees further training (European Foundation for the Improvement of Living and Working Conditions 2009). In 2010, the Ministry prolonged part-time unemployment again, until 1 July 2011, or alternatively, until the amount of €70 million that was earmarked for the measure was exhausted. This extension was seen as necessary to support those sectors that have been confronted with the effects of the crisis in a later stage, such as the construction and metal industries. As a result, firms that have not received any support according to the Deeltijd-Werkloosheidswet or Werktijdverkorting in the past can continue to apply for the scheme beyond 1 April. At the same time the final date of closing Deeltijd-Werkloosheidswet was set at 1 July 2011 (65 weeks after 1 April 2010). So even firms that applied for support in November 2010 (for 30 per cent of their staff) could only use it until 1 July 2011. In addition, firms were obligated to arrange (external) training of employees on parttime unemployment. A representative study among employers in 2009 demonstrated that they were most familiar with Werktijdverkorting (68 per cent) and DeeltijdWerkloosheidswet (78 per cent), among all the available policy measures

M3125 - MESSENGER TEXT.indd 93

14/05/2013 13:33

94

Work sharing during the Great Recession

aimed at tackling the crisis (Intomart 2010). Both measures were used quite modestly (3 and 5 per cent of enterprises, respectively), but firms that used either of the measures were quite satisfied with them. Some 78 per cent of the firms that use Werktijdverkorting valued this instrument as being adequate or even good. This percentage was even higher, 91 per cent, among users of Deeltijd-Werkloosheidswet. According to Statistics Netherlands, between April 2009 and April 2010, 69,170 workers were covered by the part-time unemployment scheme. The majority were male workers in the metal industry aged between 31 and 46 years (Chkalova 2010). While the situation in the Netherlands is very similar to other countries, for example Austria, with regard to the number of people in short-time work, the Dutch government spent more money on this measure. Yet, regarding the effectiveness of the scheme the available assessments have been rather positive. Only a small minority of workers included in the scheme were laid off afterwards, meaning that most of those jobs saved were then preserved. In July 2009 it was estimated that four out of 10 jobs in danger could be saved due to the part-time unemployment scheme (De Klaver and De Ruig quoted in Wilthagen 2010).

6.

CONCLUSIONS

In this chapter the work-sharing measures in Austria, Belgium, France and the Netherlands were presented. In these countries previously existing measures were adapted or new measures were introduced in response to the economic crisis. Although ‘short-time working’ was used in all these countries to protect jobs and to prevent unemployment the actual measures differed considerably, as did the adaptation measures taken in response to the crisis. The duration of the short-time work allowance was extended particularly significantly in Austria, where it was prolonged in several steps from three to 24 months. Another major change of the schemes was the extension of the coverage and the inclusion of atypical workers. Before the crisis, in Austria, Belgium and France, short-time work had been limited to workers in standard employment and was extended to temporary agency workers and workers with fixed-term contracts during the crisis. In Belgium, where only blue-collar workers had been covered, an additional work-sharing scheme was established for white-collar workers. While incentives for training workers were introduced in all countries, taking part in training was made obligatory only in the Netherlands. While the significance of work sharing as an anti-crisis policy is underlined by the generally high increase in the take-up of these measures, the numerical importance differs markedly between the countries. Belgium

M3125 - MESSENGER TEXT.indd 94

14/05/2013 13:33

European diversity of work sharing as a crisis measure

95

stands out with a very widespread use of chômage temporaire (temporary unemployment), a well-established measure in labour market policy. Reaching 5.6 per cent of dependent employment, the share of workers participating in short-time work was well above the take-up rate in all other European countries (OECD 2010). The four countries described in this contribution represent the diversity that exists in Europe relating to work-sharing policies and take-up patterns. According to Mandl et al. (2010), Austria and Belgium belong to the ‘traditional cluster’ of countries with a long tradition of short-time working schemes. In this group of countries, short-time workers typically have a permanent contract and work in medium-sized or large companies in a subordinate position. The cluster of ‘unconventional countries’ comprises the more economically liberal member states with established flexicurity regimes, such as the Netherlands. In these countries, workers with a fixed-term contract and working in a small company are more likely to be covered by short-time working schemes. Finally, France belongs to a group of countries with more diverse characteristics and where workers with a temporary or parttime contract were more likely to be eligible for short-time working (ibid.: 42ff.). In Belgium and the Netherlands the work-sharing measures became a bone of contention in the debate between the social partners and the government. In Belgium, the issue was the harmonization of employment status of blue- and white-collar workers. In the Netherlands, the social partners concordantly and repeatedly asked the government for adaptations and extensions of short-time working support; however, the available budget limited the number of workers benefiting from the measures. In contrast, in Austria the budget that was earmarked for short-time working in 2009 was not exhausted because the actual take-up reached only some 65 per cent of the pre-registered company demands. In France, an even lower proportion of 30 per cent (2009) of all short-time working hours that companies applied for were actually implemented in the end. The reason may be that companies acted particularly cautiously in view of the biggest crisis in the world economy since the 1930s and demanded much higher support than they actually needed. But instead of only short-time working they also used other measures such as the reduction of working hours in the framework of annualized working hours or laying off temporary workers and workers with fixed-term contracts. At least for the Austrian case with rather high levels of unemployed temporary agency workers at the end of 2008 and in 2009, it could be argued that the recently increased contractual flexibility of the labour market stood in the way of a higher take-up of the measure of work sharing. There were also differences among the countries with regard to

M3125 - MESSENGER TEXT.indd 95

14/05/2013 13:33

96

Work sharing during the Great Recession

social security contributions and unemployment insurance: in Austria, Belgium and France, the social security contributions during short-time working were calculated on the basis of the full-time income, while in the Netherlands this was not the case. What is more, the Dutch scheme stipulated that the benefits workers receive during short-time work were deducted from their unemployment benefit entitlements. The Belgian and also the Austrian examples show that not only collective measures such as short-time working are used to share work more equally in a crisis situation. Individual measures, which were designed to support workers taking leave for training or other purposes, were also frequently taken up by companies to avoid layoffs. In Belgium, individual working-time reduction schemes were used as a basis for the extension of the temporary unemployment measure to white-collar workers. In Austria, a rarely used training leave scheme became more popular because of the crisis. Overall, the findings show that work-sharing measures were crucial to the countries’ responses to the crisis, in spite of considerable variation in both the design and take-up of the measures. While diverging estimates exist regarding the precise job-preservation effects, there is general agreement that short-time work contributed strongly to keeping people in employment. In addition, the data presented in this contribution on the gap between companies’ planned and realized take-up in some countries indicate that there was still a buffer: if the downturn had been stronger and longer, companies could and probably would have used work sharing more intensively.

NOTE *

The authors are indebted to the experts in several countries who contributed to this chapter by sharing their knowledge: Maarten van Klaveren (University of Amsterdam), Hugo Erken (Ministry for Social Affairs and Employment, The Netherlands), André Simon (Ministry for Employment and Labour, Belgium), Guy van Gyes (HIVA, KU Leuven) and Sigrid Röhrich (Ministry for Labour, Social Affairs and Consumer Protection). Lukas Hofstätter supported us with research. We would also like to thank Jon Messenger from the International Labour Office for his comments and support.

REFERENCES AMS Österreich (2009a), ‘Infoblatt Kurzarbeit’, available at: http://www.ams. at/_docs/001_KUA_Infoblatt.pdf (accessed 10 December 2011). AMS Österreich (2009b), ‘Pauschalsatztabellen Kurzarbeit’, available at: http:// www.ams.at/sfu/14087.html (accessed 10 December 2011).

M3125 - MESSENGER TEXT.indd 96

14/05/2013 13:33

European diversity of work sharing as a crisis measure

97

AMS Österreich (2010), ‘AMS Arbeitsmarktdaten’, Webabfrage, available at: http://iambweb.ams.or.at/ (accessed 10 December, 2011). Arpaia, A., N. Curis, E. Meyermans, J. Peschner and F. Pierini (2010), ‘Short time working arrangements as response to cyclical fluctuations’, European Economy, Occasional Papers 64, European Commission, Brussels. BMASK (2009), ‘Aktive Arbeitsmarktpolitik in Österrreich. 1994–Mitte 2009’, Vienna, available at: http://www.bmsk.gv.at/cms/site/attachments/2/7/9/CH0690 /CMS1249975678352/dokumentation_aktive_amp_in_oesterreich_1994-2009_ final_juli_2009.pdf. BMASK (2010a), ‘ELISweb Wirtschafts- und Arbeitmarkinformationssystem’, Onlineabfrage, available at: http://www.dnet.at/elis/Default.aspx (accessed 10 December 2011). BMASK (2010b), ‘Aktive Arbeitsmarktpolitik in Österreich 1994–2010’, Vienna. Calavrezo, O., R. Duhautois and E.Walkowiak (2009), ‘The short-time compenstion program in France: an efficient measure against redundancies?’, Centre for Employment Studies (Centre d’Études de l’Emploi, CEE), Working Paper 114, Noisy le Grand, available at: http://www.cee-recherche.fr/fr/doctrav/114chomage_partiel_licenciements_economiques.pdf (accessed 16 November 2012). Chef d’entreprise (2010), ‘Chômage partiel – près de 80 millions d’heures consommées en 2009’, in Chef d’enterprise.com, Le site des dirigeants de TPE et PME, 5 February, available at: http://www.chefdentreprise.com/Breves/Chomagepartiel-pres-de-80-millions-d-heures-consommees-en-2009-32257.htm (accessed 16 November 2012). Chkalova, K. (2010), ‘Deeltijd-WW in beeld’, in Sociaaleconomische trends – statistisch kwartaalblad over arbeidsmarkt, sociale zekerheid en inkomen, 3rd quarter, Centraal Bureau voor de Statistiek, Den Haag/Heerlen. Conseil Supérieur de l’Emploi (2010), ‘Évolutions récentes sur le marché du travail’, Bruxelles. Council of the European Union (2009), Addendum 2 to the note from Presidency to Delegations. Subject: Employment Committee’s contribution to the informal Employment Summit – Analyses carried out by the EMCO on short-time working arrangements. Ref: EMCO/04/160209/EN. Cour des comptes (2011), ‘Le système français d’indemnisation du chômage partiel: un outil insuffisamment utilisé’, Rapport public annuel 2011, Paris, February. Deeke, A. (2009), ‘Konjunkturelle Kurzarbeit – Was kann bei vorübergehendem Arbeitsausfall bewirkt werden?’, WSI Mitteilungen 8/2009. European Foundation for the Improvement of Living and Working Conditions (2009), ‘Tackling the recession: employment-related public initiatives in the EU member states and Norway’, Dublin, available at: http://www.eurofound. europa.eu/docs/erm/tn0907020s/tn0907020s.pdf (accessed 16 November 2012). Fuchs, J. (2009), ‘Kurzarbeit – Die aktuelle Rechtslage’, Diploma Thesis, University of Graz. Grünell, M. (2009), ‘Trade unions place new demands on part-time unemployment scheme’, available at: http://www.eurofound.europa.eu/eiro/2009/05/articles/ nl0905029i.htm (accessed 8 September 2011). Hijzen, A. and D. Venn (2011), ‘The role of short-time work schemes during the 2008–09 recession’, OECD Social, Employment and Migration Working Papers 115, OECD, Paris.

M3125 - MESSENGER TEXT.indd 97

14/05/2013 13:33

98

Work sharing during the Great Recession

INSEE (Institut national de la statistique et des études économiques) (2010), Online Abfrage, available at: http://www.insee.fr/fr/default.asp (accessed 10 December 2011). Intomart (2010), ‘Werkgevers over crisis, crisismaatregelen en arbeidsmarkt, Kwantitatief onderzoek uitgevoerd in opdracht van het Ministerie Van Sociale Zaken en Werkgelegenheid’, Employers talking about the crisis: quantitative research commissioned by the Dutch Ministry of Employment and social Affairs. Jørgensen, C. (2009), ‘Economic crisis leads to extensive use of work-sharing’, available at: http://www.eurofound.europa.eu/eiro/2009/03/articles/dk0903021i. htm (accessed 2 September 2011). Mahringer, H. (2009), ‘Der Arbeitsmarkt in der Finanzmarkt- und Wirtschaftskrise’, WIFO Monatsberichte, 12/2009: 967–78. Mandl, I., J. Hurley, M. Mascherini and D. Storrie (2010), ‘Extending flexicurity – The potential of short-time working schemes’, ERM Report 2010, Dublin. Messenger, J.C. (2009), ‘Work sharing: a strategy to preserve jobs during the global job crisis’, Policy Brief No. 1, Conditions of Work and Employment Programme, ILO, Geneva, June. OECD (2010), Employment Outlook 2010: Moving Beyond the Jobs Crisis, Paris: OECD. ONEM (Office National de l’Èmploi) (2010), Online Abfrage, available at: http:// www.rva.be/home/menufr.htm (accessed 10 January 2012). Schratzenstaller, M. (2009), ‘Bundesvoranschlag 2009/10 und Finanzrahmen 2009 bis 2013’, WIFO Monatsberichte, 8/2009: 583–602. Service publique fédéral Emploi, Travail e. C. s. (2009), ‘Mesures anti-crise: nouveautés’, 24 December, available at: http//www.emploi.belgique.de/default News.aspx?id526028 (accessed 18 March 2010). Stiglbauer, A. (2010), ‘The Austrian labour market and the Great Recession: developments and measures taken’, Monetary Policy and the Economy, Q3/2010: 25–44. Van Gyes, G. (2009a), ‘Reducing working time as anti-crisis measure’, available at: http://www.eurofound.europa.eu/eiro/2009/06/articles/be0906029i.htm (accessed 7 September 2011). Van Gyes, G. (2009b), ‘Temporary unemployment is buffer to economic crisis’, available at: http://www.eurofound.europa.eu/eiro/2009/04/articles/be0904019i. htm (accessed 7 September 2011). Wilthagen, T. (2010), ‘Going Dutch? A comparison of the more restrictive Dutch approach to short-time work with the French partial activity scheme, Mutual Learning Programme’, Peer Review on ‘Employment measures to tackle the economic downturn: short time working arrangements/partial unemployment schemes’, France, 27–28 September.

M3125 - MESSENGER TEXT.indd 98

14/05/2013 13:33

4.

Work sharing in Japan Kazuya Ogura

1.

WORKING HOURS AND EMPLOYMENT VOLUME: A HISTORICAL PERSPECTIVE

In recent years, the debate on work sharing in Japan has become increasingly active, particularly since around 2002, when the annual average unemployment rate reached a historic high level of 5.4 per cent. In Japan, particularly from the time of the 1970s’ oil crises, it became customary for many companies to adjust the volume of employment through increasing or decreasing overtime. By maintaining a certain amount of overtime on a constant basis, dismissals were not suddenly required when business decreased and thus employment levels were maintained. At the same time, rather than recruiting large numbers of new workers when business increased, the existing workforce handled the situation through overtime to a certain extent. Over the following decades, overtime hours became the most typical method by which Japanese companies’ adjusted to changes in market demands. Adjusting the workforce to market fluctuations via overtime can be seen as a type of work sharing that is already in practice. The scale of this adjustment is quite large, and will be estimated later in this chapter. First, in order to understand the ‘Japanese style’ of work sharing, we look at employment practices in Japanese companies and their relationship to the typical method of employment adjustment via overtime hours. Section 2 examines the recent situation. Section 3 analyses the employment effect of the Employment Adjustment Subsidy. Section 4 discusses the complexity of working hours and Section 5 examines diverse employment work sharing. Section 6 concludes. Some experts have pointed out that post-war employment practices in Japan established up until the late 1950s had three primary characteristics: (i) long-term employment, (ii) seniority wages and (iii) labour unions organized at the company level (see, among others, Dore 1973; Gordon 1985). The long-term employment and the seniority wage practices are mutually dependent. Japanese companies train their employees through

99

M3125 - MESSENGER TEXT.indd 99

14/05/2013 13:33

100

Work sharing during the Great Recession

Table 4.1

Year

1974

1975 1976 1977

The rate by which companies maintained the volume of employees by working-hour reductions (manufacturing industry, %) Month

1–6 7–9 10–12 1–3 4–6 7–9 7–9

Working-hour reductions (M.A.) Overtime reduction

Increase of holidays

Layoffs

12 25 46 55 16 19 18

4 6 7 10 7 2 2

3 6 17 24 7 2 1

Source: Survey on Labour Economy Trends (MHLW).

‘on-the-job training’ (OJT) from the beginning of the employees’ job career for many years, and in this way, employees gradually acquire higher job skills in accordance with their length of service with the company. Since companies pay the higher wages according to their employees’ length of service (seniority wages), their employees are motivated to remain with the company, often for their entire career. In addition, Japanese labour unions regard the reciprocal relationship between the long-term employment and seniority wage practices as positive. Most labour unions consist solely of the regular employees at a company, and they are largely unaffected by union federations outside of the company; therefore, the unions are often very cooperative with the employers. Also, the employers regard cooperative relations with the unions as important to maintain the stability of the company, so they have made considerable efforts to preserve their employees’ jobs. The features mentioned above have played an important role – in particular at the time of the first oil crisis during 1973–74. As a result of that crisis, the unemployment rate in Japan increased from 1.3 per cent in 1973 to 1.9 per cent in 1975.1 Also, overtime hours decreased from an average of 184.8 hours per employee in 1973 to 127.2 hours per employee in 1975.2 Table 4.1 shows the rate by which companies utilized different methods of employment adjustment to reduced demand – overtime reductions, increased holidays and layoffs. As can be seen from the table, the rate of overtime reduction is relatively high from the latter half of 1974, in comparison with the other methods. This shows that many companies made an effort to preserve their employees’ jobs by reducing overtime, and labour

M3125 - MESSENGER TEXT.indd 100

14/05/2013 13:33

Work sharing in Japan

101

unions cooperated with this approach. The Japanese method of employment adjustment (so-called ‘Japanese-style’ work sharing) was established at that time.

2.

THE PROTECTION OF EMPLOYMENT BY REDUCTIONS IN OVERTIME: RECENT USE

Next, we look at the recent situation. Figure 4.1 shows the annual hours of overtime per regular employee, the growth rate of real GDP, and the unemployment rate from 1993 to 2010. As GDP increases from 1993 to 1996, so does overtime. These figures are particularly remarkable for manufacturing. However, with the slump of 1997–98, overtime decreases. Faced with an unprecedented recession in 2008–09, overtime was drastically reduced, and the unemployment rate has surged upward simultaneously. However, even with negative GDP, overtime is drastically reduced but still does not reach zero. Table 4.2 shows the rate by which companies maintained the same volume of employment by means of a reduction in employees’ working 250

10

200

6

Hours (overtime)

4 150

2 0

100

−2 −4

50

−6

% (GDP, unemployment rate)

8

−8 0

19

93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10

−10

Overtime (industry total) Unemployment rate

Overtime (manufacturing) Growth rate of real GDP

Sources: Based on Monthly Labour Survey (MHLW), Labour Force Survey (MIC), National Accounts (Cabinet Office).

Figure 4.1

Growth rate of real GDP, unemployment rate, annual overtime per regular employee, 1993–2010

M3125 - MESSENGER TEXT.indd 101

14/05/2013 13:33

102

M3125 - MESSENGER TEXT.indd 102

14/05/2013 13:33

2009

2008

2010

2009

2008

Year

1–3 4–6 7–9 10–12 1–3 4–6 7–9 10–12 1–3 4–6 7–9 10–12 1–3 4–6 7–9 10–12 1–3 4–6 7–9 10–12

Month

100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100

Total

13 14 16 35 47 49 45 43 44 40 36 34 16 16 20 50 69 71 61 55

Maintained

6 6 8 21 30 29 26 26 25 22 19 17 8 7 11 34 52 48 38 37

Overtime reduction 2 2 3 7 8 9 8 9 8 9 9 8 2 2 4 11 14 14 10 11

Increase of holidays 0 0 0 2 13 14 11 10 8 6 5 4 0 0 1 5 31 31 22 20

Lay off 1 1 2 7 12 9 7 7 6 5 3 3 1 1 3 14 25 19 13 11

Output reduction

Ways of working hours reduction (M.A.)

87 86 84 65 53 51 55 57 56 60 64 66 84 84 80 50 31 29 39 45

Not done

The rate by which companies maintained employment by working-hour reductions, 2008–2010 (in %)

Manufacturing

Total

Table 4.2

103

M3125 - MESSENGER TEXT.indd 103

14/05/2013 13:33

2010

2009

2008

1–3 4–6 7–9 10–12 1–3 4–6 7–9 10–12 1–3 4–6 7–9 10–12 1–3 4–6 7–9 10–12

100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100

Source: Survey on Labour Economy Trends (MHLW).

Wholesale, retail

2010

54 47 42 40 13 14 15 27 35 40 41 41 44 39 36 32

32 28 23 22 6 7 7 16 21 24 26 27 29 24 22 19

10 9 8 8 0 2 3 3 4 6 6 7 6 8 8 6

15 11 9 7 0 0 0 0 2 2 2 2 1 1 1 1

8 7 5 4 0 1 1 2 3 4 4 4 4 4 2 2

46 53 58 60 87 86 85 73 65 60 59 59 56 61 64 68

104

Work sharing during the Great Recession

hours, primarily (although not exclusively) overtime. In comparison with the third quarter of 2008, from the fourth quarter there is a great increase in the rate of companies taking some sort of measures. By the second quarter of 2009, 49 per cent of companies are doing so. As can be seen in the table, there are several methods of maintaining employment, but the majority of companies are making use of overtime reductions. The percentage of companies implementing overtime reductions in the manufacturing industry is particularly high. It is worth noting that from the 1990s onward, the number of companies adjusting the number of employees rather than working hours in response to changes in demand began to increase. The secular increase in non-regular employees who are not guaranteed long-term employment began during this period. This means that many companies now have an alternative mechanism to adjust employment volume other than overtime. With the crisis situation (which began in autumn 2008), although the adjustment through working hours was widely utilized, large numbers of non-regular employees were also dismissed. Further information from the MIC Labour Force Survey indicates that the number of workers in specific industries decreased during the Great Recession of 2008–09. For example, in 2008 there were 10.77 million workers in the manufacturing industry, but this number declined to 10.15 million in 2009. A total of 620,000 workers were laid off, and the rate of employment reduction was 5.8 per cent, which was the highest among all industries. The second-highest rate of employment reduction was in the construction industry (3.4 per cent), where 4.37 million workers in 2008 declined to 4.22 million in 2009. In contrast, the medical and welfare service industry increased its workforce, rising from 5.65 million workers in 2008 to 5.88 million in 2009 (the rate of increase was 4.1 per cent). On balance, therefore, the Great Recession resulted in a reduction of the workforce, especially in manufacturing and construction. Thus, many workers lost their jobs during the Great Recession of 2008–09, but it also may be true that many more people could have lost their jobs without the reduction of working hours. For this reason, we shall now estimate how much employment has been protected through overtime reduction. This estimate is shown in Table 4.33 (due to limitations of the available statistical data, only firms with five or more employees were used4). First, we present the annual working hours per employee for 2008 and 2009 (A and B). Regular employees worked 2,031.6 hours in 2008 and 1,976.4 hours in 2009, while part-time employees worked 1,111.2 and 1,082.4 hours, respectively. And the differences between 2008 and 2009 are 55.2 and 28.8 hours (C).

M3125 - MESSENGER TEXT.indd 104

14/05/2013 13:33

Work sharing in Japan

Table 4.3

105

Estimated number of employees whose employment status was maintained by reduction of overtime from 2008 to 2009 Actual working hours per employee

Regular employee Part-time employee

2008

2009

Difference

A

B

C5A−B

2,031.6

1,976.4

55.2

1,111.2

1,082.4

28.8

% of companies that reduced overtime (2009) D Regular employee Part-time employee

Regular employee Part-time employee

Reduced w. hours at companies that reduced overtime per employee E51/0.2775*C

Total no. of employees (2009)

F

27.75%

198.92

31,986,000

27.75%

103.78

12,004,000

No. of employees at firms that reduced overtime

Total reduced w. hours

No. of employees whose employment status was maintained

G5F*D

H5E*G

I5H/A

8,876,115

1,765,636,796

869,087

3,331,110

345,702,596

311,107

I (total) 1,180,194

Source: Based on and calculated using Monthly Labour Survey (MHLW).

Next, the rate of companies that reduced overtime in 2009 is calculated. The figures (from Table 4.2) are for the yearly quarters, yielding an annual average of 27.75 per cent (D). Here, we assume that the shortening of working hours from 2008 to 2009 has only been achieved by those companies that reduced overtime. Thus, we estimate that the reduced working hours per employee in 27.75 per cent of firms5 was 198.92 hours for regular employees and 103.78 hours for part-time employees (E). The total number of employees in 2009 stood at 31,986,000 regular employees and 12,004,000 part-time employees (F). Calculating the

M3125 - MESSENGER TEXT.indd 105

14/05/2013 13:33

106

Work sharing during the Great Recession

number of employees in those firms that reduced overtime yields a total of 8,876,115 regular employees and 3,331,110 part-time employees (G). Multiplying this number of employees by the reduced working hours calculated in (E) yields a total of 1,765,636,796 hours for regular employees and 345,702,596 hours for part-time employees (H). This is the total amount of reduced working hours in all firms that reduced overtime in 2009. If we divide the hours in (H) by the number of working hours in 2008 (A), the result is 869,087 regular employees and 311,107 part-time employees, giving a total of 1,180,194. We estimate that this is the total number of employees whose employment was protected by overtime reduction in firms that undertook such measures in 2009. This data covers only firms with five or more employees, but such firms make up 87 per cent of all employees. The remaining 13 per cent of employees work in firms with less than four employees. There is a high possibility that firms employing less than four employees also utilized overtime reduction, but the rate of regular and part-time employees is unknown, and hence such firms are not included in the estimate. In 2009 there were a total of 3,360,000 unemployed persons. This figure includes part-time employees, and adding 1,180,000 to this figure results in 4,540,000 people. The 1,180,000 persons represent 1.78 per cent of the 2009 labour force population of 66,170,000. Accordingly, if there were no maintenance of employment status by overtime reduction in 2009 and the decreased production were met entirely with dismissals, we can estimate that in 2009, around 4,540,000 or more workers would have been unemployed. The unemployment rate under this scenario would have been approximately 6.9 per cent, instead of 5.1 per cent (5.1 per cent 1 1.78 per cent).6 Thus, overtime reduction, the main means of employment adjustment by companies, can be estimated to have maintained the employment of more than a million workers from 2008 to 2009. In the previous recession of 2001–02, a government–labour– management agreement on work sharing was made (in March 2002), but its effect was minimal. During that recession, few companies implemented employment maintenance measures (as crisis measures) to shorten regular employees’ contracted hours (normally scheduled hours, not overtime) and reduce pay accordingly. The main reasons are thought to be because of the structure of Japanese working hours (the function of overtime), and the weakness of employment security for non-regular employees but the earlier than expected recovery is also likely to have affected the situation. However, the effects of this financial crisis and the resulting Great Recession have been quite severe; it was uncertain when an economic recovery would occur, which led to large expectations for work sharing

M3125 - MESSENGER TEXT.indd 106

14/05/2013 13:33

Work sharing in Japan

107

from all parties. Manufacturing companies implemented crisis work sharing in the form of shortened scheduled hours and accordingly reduced wages on quite a wide scale. Factories of such representative major companies as Toyota, Nissan, Hitachi and Fujitsu all reduced their production and capacity. This also affected their subcontractors, and work sharing with increased holidays and lowered wages also spread widely throughout the manufacturing industries. Work sharing in the manufacturing industry is considered possible because unpaid overtime work is relatively small in scale compared to other industrial sectors. However, service sector or ‘white-collar’ work could not be reduced as easily as in the case of reduced operations in a factory. Basically, ‘blue-collar’ workers have to be physically present at a factory (or other workplace) because means of production and raw materials are also at the factory. So, work sharing with reduced working hours and wages for them reduces those blue-collar workers’ volume of work. However, many white-collar workers often work late at night or in their own home without additional compensation (that is, unpaid overtime). Such a work attitude leads to a situation in which although the volume of work remains unchanged wages are none the less reduced; and even if these workers have more days off, work sharing may be a formality that exists only on paper rather than in practice.

3.

THE EMPLOYMENT EFFECT OF THE EMPLOYMENT ADJUSTMENT SUBSIDY

The government expanded the supporting measures for work sharing during the crisis, and these measures have been seen to have some effect on maintaining employment during the Great Recession. The most important measure of this kind is the Employment Adjustment Subsidy (EAS). This is a public subsidy paid to employers whose production or sales decreased due to an economic downturn in the most recent three months and who are making efforts to save their employees’ positions for a maximum of one year, and which may also include vocational training. The subsidy payments are covered 100 per cent by the Employment Insurance Programme, and the subsidy is paid for a maximum of 300 days over a period of three years, based on the actual number of days of reduced work. The expansion of the EAS resulted from the tripartite agreement to attain employment security and employment creation, which was concluded on 23 March 2009 by the Prime Minister, the Urban Business Federation (Nippon Keidanren), the Japan Chamber of Commerce and Industry, the National Federation of Small Business Associations, and the

M3125 - MESSENGER TEXT.indd 107

14/05/2013 13:33

108

Work sharing during the Great Recession

Japan Trade Union Confederation (RENGO). This tripartite agreement included five components, one of which was the ‘Maintenance of employment through promoting the “Japanese model” of work sharing’. The Japanese model of work sharing in this tripartite agreement contains several features for employment protection, such as maintaining the volume of employment through the reduction of working hours and wages; protecting the employment status of workers by transferring them from the parent company to subsidiaries; and relaxing of requirements for and also increasing the amount of the EAS. In line with the tripartite agreement, in June 2009 the government expanded the EAS by relaxing the eligibility criteria for the subsidy and increasing the amount of the payment. Specifically, these criteria were relaxed as follows: 1.

2.

3.

Previously, workers had to have an enrolment period for unemployment insurance of at least six months, but now there is no longer a required minimum enrolment period. Previously, eligibility requirements for support were based upon a reduction of the production volume, but now a reduction in sales is also applicable. Thus, a reduction of business activities in monetary terms has become an alternative eligibility criterion.7 The amount of financial support provided to workers requiring occupational training accompanying the reduction of production has risen from the previous 1,200 to 4,000 yen per day for large firms and 6,000 yen per day for small and medium-sized firms.

Since autumn 2008, the number of firms applying for an EAS increased rapidly. Table 4.4 shows applications for and monetary value of EAS subsidies provided by the Japanese Ministry of Health, Labour and Welfare (MHLW) to support work sharing. Less than a hundred firms applied for an EAS in the first half of 2008, but this number has risen considerably from the beginning of 2009. The largest figures were 84,481 participating firms in October 2009 and 2,530,659 participating employees in April 2009, with the amount of EAS in September 2009 reaching a historic maximum of 77.7 billion yen. However, employees out of work due to being temporarily laid off for only one day in the month are still counted as employees in Table 4.4. The number of cases of EAS assistance for the costs of non-company vocational training is included, as well as the subsidy for layoff allowances. Thus, employees only receiving vocational training but not laid off are also counted as one employee in the table. However, a single employee receiving both the layoff and vocational training subsidies is counted twice. In

M3125 - MESSENGER TEXT.indd 108

14/05/2013 13:33

Work sharing in Japan

Table 4.4

109

Number of applications for and total amount of EAS

2008 Apr May Jun Jul Aug Sep Oct Nov Dec 2009 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 2010 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Firms

Employees

Amount of EAS (yen)

63 79 92 96 123 107 140 198 1,707

1,343 2,601 1,774 2,429 3,060 2,970 3,632 8,598 138,549

34,691,000 30,466,000 37,208,000 44,586,000 43,819,000 41,214,000 53,935,000 36,834,000 49,683,000

12,209 29,137 46,558 61,360 67,216 75,273 82,982 79,904 80,908 84,481 81,231 81,756

879,614 1,865,792 2,379,069 2,530,659 2,334,312 2,376,995 2,439,417 2,114,021 1,994,071 1,974,517 1,853,495 1,866,717

81,122,000 499,907,000 5,825,942,000 13,804,158,000 32,224,658,000 54,081,890,000 76,588,301,000 76,059,714,000 77,650,955,000 73,557,536,000 57,455,223,000 59,659,160,000

83,079 79,648 82,962 80,890 76,936 76,050 72,499 69,022 67,259 64,420 60,635 57,764

1,727,778 1,608,713 1,597,700 1,497,346 1,327,747 1,282,450 1,214,566 1,124,706 1,101,037 1,056,667 1,004,667 996,869

47,617,259,000 40,813,469,000 43,959,530,000 33,824,369,000 31,523,562,000 37,725,420,000 34,236,089,000 33,795,850,000 29,226,661,000 25,828,462,000 22,966,416,000 21,549,350,000

Source: Public report on the Employment Adjustment Subsidy (MHLW).

consideration of the limitations of these data, we have estimated the effect of the EAS in maintaining employment, based on the assumption of one full month being taken off work. Table 4.5 is an estimate of potential unemployed persons. This applies

M3125 - MESSENGER TEXT.indd 109

14/05/2013 13:33

110

M3125 - MESSENGER TEXT.indd 110

14/05/2013 13:33

I/J

F1G B/H

E*0.66 E*0.8

C/30 D*0.6

Persons based on the applications for EAS at June 2009 Each rate of the total (637,087 1 1,745,844) Total expenditures for EAS at September 2009 (yen) 77,650,955,000*0.267, 0.733 Average monthly salary per employee at June 2009 (yen) Average salary per employee per day (yen) Average amount of layoff allowance per employee per day (yen) Average amount of subsidy for layoff allowance per employee per day (yen) Subsidy for training per employee per day (yen) Total amount of subsidy per employee per day (yen) Total days of layoff Average workdays at June 2009 Number of potential unemployed persons

Large firm

Small & medium-sized firm

4,000 8,701 2,382,798 19.7 120,954

4,701

Total 370,630

6,000 10,762 5,288,599 21.2 249,676

4,762

637,087 1,745,844 0.267 0.733 (Breakdown) 20,732,804,985 56,918,150,015 356,139 297,652 11,871 9,922 7,123 5,953

Estimate of potential unemployed persons based on the EAS

Source: Based on and calculated using Monthly Labour Survey, Public Report on the Employment Adjustment Subsidy (MHLW).

G H I J K

F

C D E

A1 A2 B

Table 4.5

Work sharing in Japan

111

to regular employees only. The calculation is based on September 2009, the month when the support payments of the EAS were the highest. The amounts paid via the EAS in support for large firms and small and medium-sized firms differ, but no data were available on the different scales of the amounts. The amount for September 2009 is the total expenditure for that month, but it usually takes three months from application until a decision is made to actually supply subsidies. Thus, the amount for September 2009 can be assumed to be based upon applications in June 2009. The number of applications in June 2009 was A1 according to a report by the MHLW;8 the rate of this is then calculated (A2), as is the breakdown of the total amount of EAS in row B. Row C is the June 2009 average monthly salary per employee. Under Japanese law, the daily layoff allowance is 60 per cent or more of the daily salary. However, the daily salary is calculated based upon the calendar days (30 days in June), not the working days, and thus we can estimate the daily salary (D). Companies paying over 60 per cent of the layoff allowance are assumed to be few, and the daily layoff allowance is then calculated (E). The amount of the subsidy for layoffs from the EAS is 66 per cent of the daily layoff allowance for large firms and 80 per cent for small and medium-sized firms (F). The subsidy for vocational training is 6,000 yen in large firms and 4,000 yen for small and medium-sized firms (G). The amount assumed for both the subsidy for the layoff allowance and the subsidy for vocational training for one person is (H). Row (I) is (B) divided by (H), and it can be defined as the total days of layoff. Row J is the average number of workdays in June 2009; (I) divided by (J) yields the number of potential unemployed persons (K). This is 120,954 for large firms, 249,676 for small and medium-sized firms, and in total 370,630 potentially unemployed persons. Thus the EAS can be assumed to have protected the employment of about 370,000 regular employees per month. This figure is 0.56 per cent of the labour force population for September 2009 of 66,200,000 persons. The above estimate assumes that those who received a subsidy were off work for an entire month, but in fact it includes those who might have been off work for only one or two weeks, and hence the actual number of people whose employment status was protected by the subsidy may have been greater. In addition, the effects of overtime reduction in maintaining employment were estimated in Table 4.3, but the figures for Table 4.5 may be entirely included in those of Table 4.3. At least there is a high possibility that a portion of Table 4.5 is included in Table 4.3, and hence the number of potentially unemployed persons in Table 4.5 cannot simply be added to the number from Table 4.3 to estimate the total effect in maintaining

M3125 - MESSENGER TEXT.indd 111

14/05/2013 13:33

112

Work sharing during the Great Recession

employment. Unfortunately, it is very difficult to separate out these two estimates due to the statistical limitations mentioned earlier.

4.

THE COMPLEXITY OF WORKING HOURS IN JAPAN: COULD PERMANENT OVERTIME REDUCTIONS HELP TO CREATE JOBS?

Basically, work sharing is the concept of maintaining or creating employment by adjusting working hours and wages. However, the complexity of problems with working hours in Japan makes it difficult to make a simple comparison with other developed countries. First, the overtime premium pay rate is typically lower in Japan than in most European countries. In Japan the overtime rate is only 25 per cent, with a 35 per cent premium for work on holidays. This means that the premium rates are not set at a rate to ensure cost losses for business owners.9 Moreover, regular workers work approximately 30 hours of overtime per month, including around 16 hours of unpaid overtime work on average and 10 days of unused paid annual leave per year – all of which benefits business owners and allows companies to achieve the same level of production with fewer employees. (This system also makes it hard to calculate working hours correctly.) Table 4.6 shows the estimated number of potential jobs that could be created if unpaid overtime were abolished. This applies to regular employees only. A survey of 8,000 regular employees conducted by the author in 2008 found 16.2 hours of unpaid monthly overtime per person. If this is multiplied by 12, the annual unpaid overtime is 194.4 hours (A). The total number of regular employees is 34,324,000 (B), and multiplying A and B gives 6,672,585,600 hours. This is the total unpaid overtime of regular Table 4.6

Potential volume of job creation by removal of unpaid overtime (2008, regular employees)

Annual unpaid overtime per reg. employee A 194.4

Annual working hours per reg. employee (excl. unpaid overtime) D

Potential volume of job creation

B

Total unpaid overtime of reg. employees C5A*B

34,324,000

6,672,585,600

2,031

3,284,723

No. of reg. employees (2007)

E5C/D

Sources: Based on and calculated using Monthly Labour Survey (MHLW), Employment Status Survey (MIC) and the author’s survey.

M3125 - MESSENGER TEXT.indd 112

14/05/2013 13:33

Work sharing in Japan

Table 4.7

113

Potential volume of job creation by removal of untaken paid leave (2009, regular employee)

Entitled days of paid leave per reg. employee A 17.9 Total untaken days of paid leave E5C*D 317,720,000

Days taken per reg. employee

Days untaken per reg. employee

No. of reg. employees (2007)

B

C5A−B

D

8.5

9.4

33,800,000

Annual holidays per reg. employee (excl. paid leave) F

Total holidays 1 paid leave per reg. employee G5B1F

Annual workdays per reg. employee H5365−G

Potential volume of job creation

113.4

121.9

243.1

1,306,952

I5E/H

Sources: Based on and calculated using General Survey on Working Conditions (MHLW), Employment Status Survey (MIC).

employees (C). The annual working hours of regular employees was 2,031 hours (D). Dividing (C) by (D) yields a total 3,284,723 full-time-equivalent positions (E). Thus, if the unpaid overtime of regular employees were entirely abolished and additional employment of new regular employees is assumed to be created on a proportionate basis, the theoretical estimate would be that a maximum of 3,285,000 full-time positions could be created. However, unpaid overtime has a skewed distribution for companies and workers. Hence, even under a scenario in which unpaid overtime was completely eliminated and new employment increased, the actual results would differ. New employment as a substitute for unpaid overtime would also increase wage costs, and as such could not be undertaken lightly by companies. Paid annual leave presents a similar situation. In Japan, the legally entitled amount of paid annual leave is never fully used. Table 4.7 calculates estimates of potential number of jobs which could be created by the elimination of unused days of paid annual leave. The number of days of paid annual leave entitlement per regular employee in 2008 was 17.9 days (A), but the number of leave days actually taken in 2008 was only 8.5 per employee (B). Subtracting (A) minus (B) means that the annual unused days of paid leave per regular employee was 9.4 days in 2009 (C). Multiplying the total number of regular employees, 33,800,000 (D), by

M3125 - MESSENGER TEXT.indd 113

14/05/2013 13:33

114

Work sharing during the Great Recession

(C) results in a total of 317,720,000 days (E). This is the total amount of unused days of paid annual leave. Next, we calculate the total annual workdays per regular employee per year. The annual number of non-working days other than paid annual leave (weekends, public holidays) of regular employees equals 113.4 (F). Adding the unused days of paid annual leave (B) and (F) yields 121.9 days (G); this is the total number of days that are not worked annually by regular employees. Subtracting (G) from 365 days in a year yields a total of 243.1 days (H); this is the annual number of workdays per regular employee. (E) divided by (H) yields an equivalent of 1,306,952 full-time positions (I). Thus, if all entitled paid annual leave were taken, the additional employment for about 1,307,000 workers might become available. However, just as with unpaid overtime, companies will make every effort to avoid new wage costs, and therefore, the possibility that such job creation would actually occur is extremely low.

5.

DIVERSE EMPLOYMENT WORK SHARING

Many companies have made efforts to protect their regular employees’ status through the reduction of overtime and also by means of the EAS. However, crisis work sharing in Japan is only a short-term stopgap. When we consider the future changes in the economic environment and the structure of the labour market, what is more important in the long run is to create a society where diverse forms of work are recognized. What this means is that the spread of equal working conditions with the same wages for the same type of work and level of difficulty, regardless of factors such as gender, age or type of employment, is needed in Japanese society from now on. When we consider equal working conditions, we see various differences between regular and non-regular employees. There are several major issues, including employment security for non-regular employees, working hours (regular employees work long hours), wages and other benefits (including allowances, bonuses and severance pay), career development and training (poor opportunities for non-regular employees) and social security. It would be extremely difficult to try to resolve all these issues simultaneously. The Employment Insurance Programme is an important system of income security for workers losing their job, but it currently has enrolment requirements stipulating weekly working hours of 20 or more and an employment period of six months or more. Accordingly, those with low working hours or a short period of employment are unable to access this income security when they lose their job.

M3125 - MESSENGER TEXT.indd 114

14/05/2013 13:33

Work sharing in Japan

115

With regard to the public basic pension, employees aged 20 or more have to pay the pension contribution, and if they work for a private company or public organization, they also have to pay the earnings-related pension contribution. However, a dependent family whose annual income is less than 1.3 million yen can gain the basic pension even if they do not contribute to the basic pension. Because of this income ‘cap’, many housewives are not willing to work full-time. The lower threshold for income tax is an income of 1.03 million yen, which can also be considered as a disincentive for part-time housewives to work full-time. If even only one point is changed, there would be a chorus of protest from those impacted by that change. Imagine, for example, if the income tax minimum of 1.03 million yen annually were abolished. This would mean a major change from the household to the individual person as a basic unit of the tax system. Currently, the conventional wisdom is that the income of working housewives with a low-paid part-time job (in the majority of cases) should not be taxed, as they must also do housework. Furthermore, most companies pay the husbands of such working housewives a dependant’s allowance of around 10,000 yen monthly if the wife’s annual income is less than 1.03 million yen. If this 1.03 million yen figure were abolished, one can expect opposition not only from part-time employees, but also from employees receiving the dependant’s allowance. As another example, if all the monetary working conditions enjoyed by regular employees were to be abolished, their incentive to stay with their current company would be substantially reduced. In the typical case of large Japanese companies where a large number of young employees are trained to become employees with firm-specific skills through on-the-job training, the so-called seniority-based pay system and the security of longterm employment creates incentives for them to stay in the company. The problem in Japan, where the labour market does not have an occupational basis and where there are relatively well-paid regular employees with high skills and low-paid and low-skilled non-regular employees, is the existence of a sort of dual social status based on the form of employment. Eliminating the differences between regular and non-regular employees means that, relatively, the existing rights of regular employees will be somewhat reduced. To avoid this situation, the current working conditions of regular employees must be maintained, while those of non-regular employees must be improved both absolutely and relatively during periods of economic growth. Yet, despite the healthy economy of the years prior to the Great Recession, no such bold moves were made. Some laws were revised, but with the advent of a once-in-a-century crisis during the process of these reforms, the favourable conditions for change have disappeared.

M3125 - MESSENGER TEXT.indd 115

14/05/2013 13:33

116

6.

Work sharing during the Great Recession

CONCLUSION

In Japan, since the 1970s’ oil crises, many companies have introduced the employment maintenance type of work sharing. By means of a reduction of overtime hours during a recession, employment volume is maintained to a large extent with a relatively small number of dismissals. Over the following decades, this approach became the most typical work-sharing method in Japanese companies. This employment maintenance type of work sharing has also been fully utilized during the Great Recession of 2008–09. According to the author’s estimates, the employment of more than one million workers was protected by this type of work sharing. In addition, the government expanded the EAS in 2009 by relaxing the eligibility criteria and increasing the amount of support provided to companies. According to the author’s estimates, these changes led to the preservation of employment for at least 370,000 workers per month. However, there are at least two major issues regarding the future of work sharing in Japan. One is the difficulty of introducing the employmentcreation type of work sharing by which unemployed people can become employed without economic growth. This strategy is negatively affected by the major differences between regular and non-regular employees in terms of employment periods, wages, types of job and so on. If such differences among workers cannot be reduced, then the introduction of the employment-creation type of work sharing would be difficult. In addition, a big problem to consider is the wage costs. That is, if the wages of nonregular employees increase and the wages of regular employees remain at the same level, then the wage gap between non-regular and regular employees would be reduced. However, under this scenario, total wage costs would increase, and therefore, this option is unlikely to be accepted by employers. On the other hand, if the wages of regular employees decrease while those of non-regular employees remain at the same level or increase, the wage gap also becomes smaller. However, it would be very difficult for regular employees to accept such wage cuts, since most of them are the main ‘breadwinners’ in their household and their wages are a larger share of household income than the wages of non-regular employees who support the household. Therefore, it is very difficult to solve the problem of such differences between regular and non-regular employees without an economic recovery. Another important issue is the existence of unpaid overtime and/or unused days of paid annual leave. Many workers, particularly white-collar workers, often work late at night and/or while on holiday without additional wages and also never fully take their legally entitled paid annual leave; therefore, the employment-maintenance type of work sharing does

M3125 - MESSENGER TEXT.indd 116

14/05/2013 13:33

Work sharing in Japan

117

not always function as effectively as it could. However, addressing this point would require many major changes to the Japanese way of working, and it is simply not realistic to expect that the situation will change dramatically in the near future.

NOTES 1. MIC Labour Force Survey, Ministry of Internal Affairs and Communications, Tokyo. 2. MHLW Monthly Labour Survey, Ministry of Health, Labour and Welfare, Tokyo. 3. In both cases, working hours were longer in 2008 than in 2009. In order to estimate the effect of employment protection, we use the reduction of working hours from 2008 to 2009. 4. It should be noted that these are firms, not companies; one company may contain several firms. 5. Companies are used for overtime reduction percentages, as no studies concentrating on firms are available. 6. The labour force population includes workers who are self-employed. Although selfemployed workers can also be considered to protect their own employment through reducing working hours to avoid losing their work, this is not the subject of the current study. 7. The condition is a reduction of 5 per cent or more in the production or sales in the last three months compared with the previous three months or the same three months during the last year. 8. Those numbers are a quick estimation by the Ministry, so the sum of the two is different from the definite number of June 2009 in Table 4.5. Unfortunately we could not obtain a precise figure with a breakdown by the size of the firm. 9. Furthermore, some allowances and bonus payments are not included in the calculation of the base wage for the overtime premium rate. This means that those rates would become lower than the expected rate in which the allowances and the bonus payments are included in the calculation. For ordinary regular employees at large firms, the bonus payments account for 20 to 30 per cent of their annual income.

BIBLIOGRAPHY Dore, R. (1973), British Factory – Japanese Factory – The Origin of National  Diversity in Industrial Relations, Berkeley, CA: University of California Press. Gordon, A. (1985), The Evolution of Labor Relations in Japan – Heavy Industry 1853–1955, Harvard East Asian Monographs 117, Harvard University, Cambridge, MA. Higuchi, Y. (ed.) (2002), Nihongata Worksharing no Jissen (Japanese Worksharing in Practice), Japan Productivity Center, Tokyo. Japan Institute for Labour Policy and Training (JILPT) (2009), ‘Research on Diversification of Workplace and Working Hours’ JILPT Research Report No. 106. Ogura, K. (2001), Oshu no Worksharing (Worksharing in Europe), Japan Institute for Labour Policy and Training, Tokyo. Ogura, K. (2008), ‘Worksharing ha Koyo Sokushin ni Yuko datta ka’ (Was

M3125 - MESSENGER TEXT.indd 117

14/05/2013 13:33

118

Work sharing during the Great Recession

worksharing effective for promoting employment?), Nihon Rodo Kenkyu Zasshi (The Monthly Journal of Labour Studies), No. 573, 84–7. Wakisaka, A. (2002), Nihongata Worksharing (Japanese Worksharing), Tokyo: PHP Kenkyujo. Yamato, K. (2009), ‘Koyo Chousei Joseikin no Shitsugyo Yokusei Kouka’ (Employment Adjustment Subsidy’s Inhibiting Effect of Unemployment), in Mizuho Keizai Insaito (Mizuho Economic Insight), 5 August, 1–3.

M3125 - MESSENGER TEXT.indd 118

14/05/2013 13:33

5.

The Turkish experience with work-sharing policy during the global economic crisis, 2008–2010 Erinç Yeldan*

1.

INTRODUCTION

In terms of employment vulnerability, Turkey was hard hit by the global financial crisis that began in 2007. The repercussions of the crisis began to be felt by the third quarter of 2008 when the gross domestic product (GDP) decelerated to 0.9 per cent in real terms (annualized). With a further contraction of 7 per cent in the last quarter of 2008, and a historical collapse by 14.9 per cent in the first quarter of 2009, the Turkish labour market faced a severe contraction in demand. The open unemployment ratio increased by almost six percentage points from an average of 9.5 per cent during the summer months of 2008 to a historical high of 16.1 per cent in February 2009. Industrial production felt the burden of adjustment severely as it dropped by 24 per cent in January 2009, and may have reached the pre-crisis levels only by July 2010. In response to the darkening economic conditions, the Turkish government enacted a series of stimulus packages starting in the last quarter of 2008. The first employment package was announced in October 2008, with mostly provisions for reductions in social security premiums and other cost items for the employers. This package was complemented in February 2009 with the work-sharing programme (Reduced Working-Time (RWT), which is the English translation of the official Turkish title) that was then instituted. A further package was announced in May 2009 that included a wider set of fiscal stimulus measures directed towards employment and aggregate demand. Even though it had a relatively low fiscal component, the policy of RWT funding was the most widely implemented and most visible aspect of employment-related measures. The programme was initiated to compensate those workers employed in those enterprises where, due to either the crisis or other adverse economic conditions, production and employment hours 119

M3125 - MESSENGER TEXT.indd 119

14/05/2013 13:33

120

Work sharing during the Great Recession

were reduced, or even completely stopped. With approval of the Ministry of Labour, those workers were granted payments from the RWT Fund. According to the reports of the Turkish Employment Agency (ISKUR), the total number of participants engaged in the RWT programme reached 268,081 over a period of two and a half years since November 2008. Total outlays of the fund reached to TL 162.5 million1 (approx. US$110 million) in 2009, and dropped to TL 39.1 million (approx. US$26 million) in 2010. The aim of this chapter is to investigate the employment effects and fiscal implications of the RWT programme over the course of the global crisis in Turkey. The chapter will be structured as follows: a brief historical outline will be presented of the Turkish system of industrial relations, with an emphasis on the institutional and legal infrastructure (Section 2). An overview of the macroeconomic conditions pertaining to the Turkish labour markets over the 2000s will then be considered, with an examination of the effects of the global crisis in 2008 and beyond (Section 3). In Section 4, the RWT programme is introduced and evaluated in terms of its overall coverage, aims, benefits, and costs. Section 5 engages in qualitative assessments and direct interviews with the social partners of the programme. The chapter concludes with a summary assessment.

2.

INSTITUTIONS AND THE LEGAL INFRASTRUCTURE

To better understand work sharing in Turkey, it will be useful to conduct a brief review of the institutional and legal framework regarding the labour markets and industrial relations. Modern legislation governing these dates back to the early days of the Republic in 1924.2 The 1924 Law instituted a 6-day workweek with Fridays as the official holiday. The Ministry of Labour was founded in 1945, and with the switch to official development planning in early 1960s, a more liberal and tolerant environment was generated to labour’s organizational rights (Boratav 1991). Union activities such as collective bargaining and the right to strike were recognized in 1963 with the establishment of the Labour Code, No. 1475. With the oil price shocks of 1974 and 1979, resulting in stagnation of the global economy, Turkey also suffered a severe economic and political crisis. A high inflation–high unemployment spiral took place, and the labour movement became increasingly active by the end of the 1970s. The 1980 coup d’état severely restricted labour’s rights with a new constitution and amendments to the Labour Law. Trade unions were banned and their leaders were arrested, facing treason charges. Under the repressive conditions of the 1980s, the labour movement experienced a considerable

M3125 - MESSENGER TEXT.indd 120

14/05/2013 13:33

The Turkish experience with work-sharing policy, 2008–2010

121

setback. These adverse conditions had a direct effect on labour remuneration and social rights. Real wages went on a downward spiral resulting in cumulative decline of 25 per cent in the period from 1980 to 1987 (Voyvoda and Yeldan 2001; Yeldan, 2001; Boratav et al. 2002). The 1980 coup was followed by a comprehensive privatization programme over the rest of the decade and into the 1990s. Trade union density fell to about 8 per cent of the aggregate labour force (with rural employment accounting for roughly half of total employment at the time) (Tunali 2003; Dayıoglu and Ercan 2010). The legislation on industrial relations and collective bargaining, in particular, was developed in the post-coup era, in 1983, and has been amended several times since then. By 1995, most of the restrictions over trade union activities from the military era were removed. Turkish accession negotiations with the EU set a positive tone towards labour’s right to organize (Tunali 2003). Public sector civilian workers’ right to organize in trade unions with proper recognition of collective bargaining and strikes was initiated in 2001 with the advent of the Law on Public Employee Membership in Trade Unions. Social securitization and public pension rights were redesigned via reforms in 2002 and 2008. Turkish labour’s unionization is overseen via Trade Union Law No. 2821 and the Law on Strikes, Lock-outs No. 2822. Law No. 2821 established 28 work activities as the organizational domain for both the employees and employer unions. According to this law, in order to be able to engage in collective bargaining, a union ought to represent at least 50 per cent of the employees at the enterprise level, and at least 10 per cent of the aggregate labour force of the identified work activities. These restrictions had a negative effect on trade unions’ power to act as a representative at the enterprise level.3 Currently, there are about 100 workers’ unions in the private sector and 50 unions in the public sector (Dayıoglu and Ercan 2010). Employees’ unions are represented by three major worker confederations: Turk-Is (Confederation of Turkish Labour Unions) is the largest organization. It is structured mostly around the American unionization model with a pre-commitment to non-partisan style of politics. Turk-Is makes it plain that it has no ideological relationship with any of the political parties, their ideologies, or their programmes; and that its mandate rests solely on enhancing labour’s working conditions and wage incomes through collective bargaining. DISK (Confederation of Progressive Trade Unions of Turkey) has a more radical stance, especially given the historical struggle experiences of its leaders during the 1970s. The Hak-Is Trade Union Confederation is generally regarded as a pro-Islamist and pro-(current) government organization. There are some anecdotal exceptions in the

M3125 - MESSENGER TEXT.indd 121

14/05/2013 13:33

122

Work sharing during the Great Recession

relationship between Hak-Is and the position of the government but they generally are in agreement.4 By contrast, employers’ unions number around 50 and are represented by a single confederation, TISK (Turkish Confederation of Employers’ Unions). Social dialogue (among employers, employees and the government bodies) is further engaged via certain national institutions, among which the Economic and Social Council is the most prominent. The Council is chaired by the prime minister, and includes the ministers of finance, labour and social security, and the undersecretariat of the Treasury, State Planning Organization5 and Foreign Trade. It also comprises Chambers of Commerce and Bourses (TOBB), of Agriculture (TZOB), and of Tradesmen and Artisans (TESK), the employers’ confederation (TISK), and all three workers’ confederations. The Council was established in 1995 and has an advisory role to address economic and social issues at the nationwide level. Tunali observes: [T]he legislative structure that governs industrial relations includes a stipulation for social dialogue at the enterprise level. Union representatives have a say on various matters such as annual leave, health, safety, and disciplinary regulations at the plant level. (Tunali, 2003: ch 7)

The legal minimum wage rate is set by the Minimum Wage Determination Commission. This rate not only sets a floor for remunerations of the formal employees but also provides a benchmark for a host of cases, such as public sector civilian workers’ salary grades and also retirees’ pensions. The Commission comprises representatives from the government, workers’ confederations and employee associations. The Unemployment Insurance Fund and its supporting legislation were enacted in 1999 and began functioning in March 2002. Its operations are run by ISKUR (the Employment Agency). The unemployment insurance scheme had been under severe criticism from both the unions and the employers’ associations. Workers’ unions claim that the fund’s coverage is not adequate and its conditions are too strict (Bahçe et al. 2011; ISSA 2011). Employers, on the other hand, object to the additional wage costs, and claim that the fund should have replaced severance pay obligations (TISK 2010). Conditions on layoffs and overall legislation on employment protection are arguably strict in Turkey. An OECD study conducted in 2008 claimed that among 40 OECD countries, Turkey ranks the highest in terms of the employment protection legislation (EPL) index (OECD 2008 see also TISK 2010). An in-depth investigation by Taymaz and Ozler (2005) has found that the main reason for Turkey’s rather high EPL scores stem from

M3125 - MESSENGER TEXT.indd 122

14/05/2013 13:33

The Turkish experience with work-sharing policy, 2008–2010

123

the restrictions on the use of temporary forms of employment contracts and the lack of a legal framework for temporary work (see Dayιoglu and Ercan 2010: 48). Other reasons include high severance and notice obligations (kıdem ve ihbar tazminatı) and generous dismissal compensation. The main counter-argument set forth by the workers’ unions is that the labour markets are severely fragmented and informalized. At the same time the real wage rate is, in fact, quite flexible, though notably in a downward direction. A study by Voyvoda and Yeldan (2001) found that manufacturing wages were indeed quite flexible, and furthermore, that under many crisis episodes such as 1994 and 2001, real wages had taken the burden of adjustment (see Figure 5.1). Figure 5.1 displays the dynamics of the real wages in private manufacturing, denominated both in Turkish lira and also in US dollars. Data in Figure 5.1 reveal that after a brief surge over 1990-93, real wages had plummeted during the 1994 financial crisis, and thus have borne the brunt of adjustment to the crisis. During 1996–98, real wages generally kept their momentum. However, after the 1999 wave of crises, real wages in private manufacturing faced a second cycle of contraction. This contraction was especially pronounced in US dollar terms. Turkish manufacturing labour entered the 2000s under such conditions. With this background, the chapter will now focus on the employment effects of the Great Recession starting in 2008.

3.

POST-2000 TURKEY: RAPID ECONOMIC GROWTH, FALTERING EMPLOYMENT PERFORMANCE

During the 2000s, despite rapid growth and a significant surge in exports, the economy could not generate jobs at the desired rate. The unemployment rate, which stood at 6.5 per cent in 2000, had jumped to 10.3 per cent in 2002 in the aftermath of the February 2001 financial crisis. The annual rate of growth averaged 5.8 per cent over 2003 to 2008. Yet, employment generation capacity of this rapid growth had not met expectations, and the unemployment rate could not be brought down below 9 per cent by the end of 2007, just prior to the eruption of the current global economic crisis. Despite rapid expansion of production in many sectors, civilian employment increased sluggishly, and labour participation remained below its levels as observed during the 1990s. Table 5.1 offers general statistics regarding the labour markets. Over the first decade of the 2000, the civilian labour force is observed to have increased by around 2.6 million. Comparable sectoral data reveals a secular

M3125 - MESSENGER TEXT.indd 123

14/05/2013 13:33

124

1988-I

2003.I

2002.I

2001.I

1999.I

1998-I

1997-I

1996-I

1995-I

1994-I

1993-I

1992-I

1991-I

1990-I

1989-I

Productivity Index

Real wages (TL)

Unit wage costs (US$)

Productivity and real wages in Turkish private manufacturing (Index 1988.15100)

Voyvoda and Yeldan (2001), calculations based on Turkstat data from manufacturing industry annual surveys.

Figure 5.1

Source:

40

70

100

130

160

190

220

2001 crisis

2004.I

250

1994 crisis

2005.I

280

1990−93 expansion via hot money inflows

2006.I

310

2000.I

Post-2001 adjustments

2007.I

340

2008.I

M3125 - MESSENGER TEXT.indd 124

14/05/2013 13:33

125

M3125 - MESSENGER TEXT.indd 125

14/05/2013 13:33

6,088 4,407 1,267 10,569

7,769 3,810 1,364 8,637

5,713 4,136 1,189 9,918

48,485 23,250 20,954 2,295 9.9 1,959 16.9

2006

4,867 4,314 1,231 10,327

49,994 23,114 20,738 2,376 10.3 1,805 16.8

2007

5,016 4,441 1,241 10,495

50,772 23,805 21,194 2,611 11.0 1,850 17.4

2008

New Series

5,240 4,079 1,306 10,650

51,686 24,748 21,277 3,471 14.0 2,061 20.6

2009

52,541 25,641 22,594 3,046 11.9 2,013 18.3   5,683 4,496 1,431 10,986

2010

Source:

Turkstat, Household Labour Force (HLF) Surveys.

Notes: a. Persons not looking for a job yet ready to work if offered a job: (i) Seeking employment and ready to work within 15 days, and yet did not use any of the job search channels in the last 3 months; plus (ii) discouraged workers. b. Total (open 1 disguised) unemployment accounting for the persons ‘not in the labour force’.

51,668 24,776 22,330 2,446 9.9 2,087 16.9

2006

46,209 23,078 21,581 1,497 6.5 1,139 10.9

2000

Developments in the Turkish labour market (1,000 persons)

151 Age Population Civilian Labour Force Civilian Employment Unemployed (Open) Open Unemployment Ratio (%) Disguised Unemploymenta Total Unemployment Ratiob (%) Civilian Employment by Sectors Agriculture Industry Construction Services

Table 5.1

126

Work sharing during the Great Recession

decline in the rural workforce from 7.8 million in 2000 to 4.9 million in 2007 just before the outbreak of the global crisis. Gains in industrial employment were limited, with approximately 500,000 during this period. With construction employment remaining roughly constant at around 1.2 million, most of the rural labour surplus seems to have been absorbed by services sectors. Many researchers trace the roots of fragmentation and informalization of the urban labour market to this trend, where small-scale services based on self-employment and family labour was mostly the observed norm (see, for example, Taymaz and Ozler 2005; Ercan and Tansel 2006; Agénor et al. 2007; Bahçe et al. 2011; Demιrhan et al. 2011; ISSA 2011). From the table one can further trace the impact of the crisis in the labour markets. The number of unemployed increased by 860,000 between 2008 and 2009. Consequently, the unemployment rate jumped from 11 to 14 per cent from 2008 to 2009. The fact that the rural economy could have absorbed much of the excess labour surplus with additional agricultural ‘jobs’ totalling 230,000 in 2009 was conducive in mitigating the adverse effects of the crisis. In particular, industrial sectors had shed, on average, 320,000 jobs in 2009. Figure 5.2 extends this picture over a broader time horizon utilizing quarterly averages from 1998 to 2010. From 1998 to the end of 2000, the average unemployment rate stood at 6.9 per cent. With the onset of the 2001 financial crisis it jumped to 10.1 per cent. The global crisis pushed the unemployment rate to 12.9 per cent. As of January 2011, the unemployment rate is estimated at 11.2 per cent.6 An important group of people not covered in those numbers is the group of ‘discouraged’ workers. The Turkstat surveys identifies this group as: ‘Persons not looking for a job yet ready to work if offered a job: (i) Seeking employment and ready to work within 15 days, and yet did not use any of the job search channels in the last 3 months; plus (ii) discouraged workers’. This group of people is not counted as part of the civilian labour force and is not included among the unemployed. This number had been consistently rising over the course of the 2000s and, according to Turkstat’s Household Labour Force Survey (HLFS) results, had reached to 2,070,000 by the end of 2010. If we add the Turkstat data on the ‘disguised unemployment’ defined as the excess labour supply (unemployed and disguised), it is observed to reach 18.3 per cent of the labour force. 3.1

Impact of the of the Global Crisis on the Turkish Labour Markets

The effects of the global crisis on the economy were increasingly felt at the beginning of the third quarter of 2008. As the growth rate in GDP decelerated to 0.9 per cent as an average for the whole of 2008, it registered a

M3125 - MESSENGER TEXT.indd 126

14/05/2013 13:33

M3125 - MESSENGER TEXT.indd 127

14/05/2013 13:33

127

Figure 5.2 Open unemployment by quarters, Turkey (%)

Source: Calculations based on Turkstat, HLF Surveys.

4

6

8

10

12

14

98

16

19

-I

19 9

8.1 6.0 6.0 7.5

9.

8.9

I

6.8 6.8

20

00 .

8.2 8.3

I 6.1 5.5

20 0

6.2

1.

8.5

I

6.7 7.8

20 0

10.4

2.

I

11.5 9.3 9.6

20

11.0

03 .

12.3

I

10.0 9.4

20 0

10.3

4.

12.4

I

20 0

9.3 9.5 10.0

5.

11.4

I

20 0

9.4 9.4 10.6

6.

11.5

I

20 0

9.2 9.0 9.8

7.

I

10.9 9.2 9.1

20 0

10.1

8.

11.5

I

9.5 10.3

20 0

12.6

9.

15.8

I 13.8 13.2

20 1

13.2

0.

I

14.2 11.2 11.1 11.2

128

Work sharing during the Great Recession

20.0 16.1

15.5

16.0

15.8 14.9 13.6

14.5

14.4 13.7 11.9 11.0

11.6

12.0

9.9

12.0 9.2

13.0

11.0

13.4

12.8

9.4 10.5

13.4

14.0 13.5

13.1 12.6

13.0

10.7 11.2 9.9 10.2 11.4 11.3 11.2 10.6

11.0

11.4

8.0 2008 4.0

2009 2010

be r

be r

ec em D

be r

em ov

N

O ct o

be r

t us

em Se pt

y Ju l

A ug

e

ay

Ju n

M

ch A pr il

ar

ry M

ua Fe br

Ja n

ua

ry

0.0

Source: Turkstat, HLF Surveys.

Figure 5.3

Turkey: open unemployment ratio under the global crisis

further decline of 6.8 per cent over the first half of 2009. The burden of adjustment increasingly fell on the real economy, in particular the industrial sectors and the non-agricultural labour market. Industrial output fell by 24 per cent by January 2009 and could have reached the pre-crisis levels only as late as July 2010. The unemployment rate rose steadily in the second half of 2008 and jumped to a new plateau in 2009, before finally receding to its pre-crisis levels. This also included significant wage losses and extended informalization of the workplace (Figure 5.3). The effect of the global crisis varied by sector. Industrial output fell in the first six months of the crisis by a cumulative 40 per cent from October 2008 to March 2009. Industry had shed about 400,000 jobs over that period; the nonagricultural unemployment rate jumped to 17 per cent in February 2009. Among the industrial sectors, the hardest hit were the export-oriented manufacturing sectors such as automobiles, auto parts and consumer durables. The trends of employment in the agricultural versus manufacturing industries reveal that the burden of adjustment in the labour markets was mostly absorbed by the rural economy. Agricultural employment accounted for about 4.2 million people in early 2008. As the effects of the Great Recession took hold throughout 2008 and 2009, agricultural employment increased by 2 million, bringing the number of workers employed in agricultural activities to 6.2 million. Employment in the manufacturing industries did not return to pre-crisis levels, even as late as June 2010 (see Figure 5.4).

M3125 - MESSENGER TEXT.indd 128

14/05/2013 13:33

The Turkish experience with work-sharing policy, 2008–2010

129

6,500 6,000 5,500 5,000 4,500 4,000 3,500 3,000 Agriculture Manufacturing industry

2,500

20

08

Ja Fe nua br ry u M ary ar A ch pr M il a Ju y ne J u Se Au ly pt gu e s Om t N ctober ov b e e 20 Dec mb r 09 em er Ja be Fe nua r br ry u M ary ar A ch pr M il a Ju y ne J Se Au uly pt gu e s Om t N ctober ov b e e 20 Dec mb r 10 em er Ja b Fe nua er br ry u M ary ar A ch pr M il a Ju y ne J Se Au uly g pt u e s Om t N ctober ov b D em er ec b em er be r

2,000

Source: Turkstat, HLF Surveys.

Figure 5.4

Employment in agriculture and in manufacturing (1,000 persons)

The rather unexpected rise in agricultural employment has cast doubts over the quality of employment statistics. Sonmez (2011), for instance, argues that the rise in rural employment had been a reflection of the defective questionnaire set in the Turkstat, Household Labour Force (HLF) Surveys, and that such a rise in agricultural employment is not consistent with the reported national income statistics that reveal a contraction of the agricultural sectors followed by only a modest expansion in 2010. BETAM (Bahcesehir University Economic Research Centre) researchers claim, on the other hand, that this sort of expansion was mostly centred on cash crops and was in line with the overall rise in food prices worldwide.7 A significant characteristic of the unemployment problem over this period was the rapid rise of long-term unemployment (that is, those who had been unemployed for six months and more). In 2008, the annual average of long-term unemployment for six months or longer was 1.112 million people, or 42 per cent of the total openly unemployed. In 2009 the share of long-term unemployed to total increased to 45 per cent, or 1.56 million people. By the end of 2010 the share of long-term unemployed stood at 49 per cent (Turkstat, HLF surveys).

M3125 - MESSENGER TEXT.indd 129

14/05/2013 13:33

130

4.

Work sharing during the Great Recession

ELEMENTS OF THE REDUCED WORKING-TIME PROGRAMME

In response to the unfolding recession, the government enacted a series of stimulus packages to combat aggravated unemployment and output losses. These measures were introduced under three broad policy measures and were spread over the last quarter of 2008 and the first half of 2009. They mainly relied on tax reductions and subsidies to promote investment and employment. It is estimated that as a ratio to GDP, the fiscal costs of the stimulus measures that were directed exclusively at employment preservation amounted to 0.01 per cent in 2008, 0.45 per cent in 2009 and 0.49 per cent in 2010. These numbers can be contrasted with the costs of the overall stimulus package (including financial measures), which were estimated to be on the order of 0.99 per cent of GDP in 2008, 3.41 per cent in 2009 and 2.23 per cent in 2010 (see Yeldan 2010; Ercan 2010). As stated above, the policy of RWT funding was the most widely implemented and most visible aspect of employment-related measures. Initial steps towards an officially recognized programme on work sharing were taken as early as May 2003, when the concept was first referred to in the Labour Law 4857, Article 65. The programme was openly described and its principles were laid out in the Official Gazette dated 31 March 2004. Further amendments were introduced with two stimulus packages introduced in January and May 2009 under conditions of the global crisis. It was put into its final form as part of the Omnibus Law8 which came into effect in February 2011. The RWT concept was initially identified in the Labour Law (4857, art. 65) as a ‘reduction of weekly working time by at least one-third over a period of at least four weeks in a continuous fashion, and at most three months’. It is noteworthy that the law clearly distinguishes ‘reduced working time’ from ‘short time temporary work’. The concept of ‘temporary employment’ allows for both parties to agree on the temporary nature and the overall timeframe of the work contract. Under ‘reduced working time’, by contrast, the nature of the work is permanent but is subject to necessary reduction due to external factors outside the enterprise. The law initially set the conditions for RWT under two categories: (i)  the presence of a general economic crisis, and (ii) unfavourable conditions. Note that in either of these cases, the fact that they ought to be generated from external factors to the firm is important. Thus, given the official definition, RWT is understood to be present if due to either a general economic crisis or unfavourable conditions external to the firm, production activities have been stopped or were reduced significantly (meaning by at least one-third) over a period of at least four continuous

M3125 - MESSENGER TEXT.indd 130

14/05/2013 13:33

The Turkish experience with work-sharing policy, 2008–2010

131

weeks. Conditions of the general economic crisis were regarded by law as those domestic and international developments that had a profound effect on the national economy, and thus, on the specific production unit. If these conditions were prevalent, then the employer had to make an application to ISKUR, with a detailed statement concerning the conditions faced by the firm. The employer also needs to inform employees of the status and application procedures. If the enterprise was also subject to collective bargaining, the employer had to inform the workers’ union as well. In February 2011, a series of articles and amendments pertaining to labour relations and social securitization issues were lumped together in what was known by the public as the ‘Omnibus Law’. Numbered 6111, the Omnibus Law also contained new decrees concerning the RWT structure. The most important new addition was the extension of the coverage of the RWT cases to include sectoral or regional crisis conditions. Accordingly, now firms can apply for RWT coverage for a period of three months if, due to the general economic, sectoral, or regional crisis, the firm experiences a reduction in its weekly activity in all or some of its departments. Under the newly stated conditions, a ‘sectoral crisis’ is recognized in the law as being ‘due to developments experienced in the national or international scale, certain productive sectors of the domestic economy are thought to be adversely affected, downsizing the enterprises within that sector’. In contrast, a ‘regional crisis’ in the law recognizes such adverse conditions within the context of municipalities or regions. Recognition of the conditions of a general economic crisis was initially determined by the Ministry of Labour, which could declare it independently. The Omnibus Law amended this authorization, and transferred the power to declare a general economic crisis from the ministry to the Council of Ministers. Currently, the Council of Ministers is charged with the authority of extending RWT coverage up to six months. Given these final amendments, the RWT payments are to be paid from the Unemployment Insurance Fund at a rate of 60 per cent of the gross monthly wages. This sum cannot exceed 150 per cent of the gross legal minimum wage rate. In principle, if unfavourable conditions faced by the firm entail elements specific to managerial decisions (such as insufficiency of cash funds or undesired inventory accumulation) then these petitions are rejected by ISKUR. Consequently, unfavourable conditions faced by the firm need to be of an external nature to be recognized as a basis for RWT. These typically involve enterprise-specific adverse developments that had not occurred due to management decisions and were unfavourable in nature such as natural disasters, widespread epidemics, terrorist acts/wars or similar conditions.

M3125 - MESSENGER TEXT.indd 131

14/05/2013 13:33

132

Work sharing during the Great Recession

Notably, the law clearly indicates that seasonal factors are not recognized among the general, sectoral or regional crisis conditions, nor are the unfavourable developments pertaining to demand. Thus, adverse effects of this sort are not eligible for RWT funding. During the application process, the contents of the application package are included in a CD that is then prepared by the employer. The materials include: ● ● ● ●

●

identification of the impact of the crisis or unfavourable external conditions affecting the enterprise and its activities; identification of the time period in which the enterprise can resort to RWT; declaration of whether the RWT procedures would apply to the whole or specific departments of the enterprise; declaration of the employees by name, social security identification, average employment hours weekly, and the proposed programme for RWT; and if the enterprise activity has been stopped altogether, declaration of the time period of inactivity along with the names and identification of the workers affected.

The application package of the enterprise is investigated by inspectors from the Ministry of Labour. If the application is approved, then the ministry sends a report to ISKUR granting the beginning and ending dates of the RWT programme. ISKUR informs the employer who, in turn, has to inform its employees and their union about the general terms of the RWT. Once the enterprise is accepted under the RWT programme, the employer does not pay any wages to its employees, or their social security or unemployment insurance premiums. Irrespective of the level of the monthly wages, those workers under the RWT programme receive their upcoming payments from ISKUR under conditions set by law. Also during the RWT coverage, two-thirds of the employees’ health and maternity premiums are paid by the Social Security Institute. These premiums are calculated by the Institute based on the lowest income threshold. In order to gain access to the RWT programme, workers have to satisfy the general conditions of eligibility for receiving unemployment insurance benefits, including: ●

the employee should be employed continuously for at least 120 consecutive days up until the time of RWT coverage; and

M3125 - MESSENGER TEXT.indd 132

14/05/2013 13:33

The Turkish experience with work-sharing policy, 2008–2010 ●

133

the employee should have worked and paid his/her unemployment insurance premiums for at least 20 months (600 days) over the last three years.

Given the RWT coverage opportunity at the enterprise, the employee is faced with two options. First, provided that he/she meets the overall requirements of eligibility, the employee can apply to ISKUR within 30 days and seek assistance under the RWT programme. Second, within six days of the recognition and announcement of the enterprise for eligibility to the RWT programme, the employee can ask for his/her contract to be terminated under conditions set forth in the Labour Law (4857, Article 24). In this situation, the employee recognizes that he/she has terminated his/her contract as a right and is therefore eligible for severance pay and other rights that had been stipulated in the contract. Notification Pay (ihbar tazminatı), however, is not granted to such employees who had voluntarily declared their intention to leave their job. On the other hand, workers who wish to continue with their job, now protected under the RWT programme, reserve the right to return to that job when conditions become favourable at the enterprise. This right is not recognized for workers who had chosen to leave the enterprise and had asked for the severance payments from the employer. As conditions worsened as a result of the global crisis, the RWT coverage was extended first from three to six months in February of 2009, and then up to one year in May 2009 as part of the first two stimulus packages. In its initial format, before 2009, if employees were to lose their job at the end of the RWT programme, they would start receiving unemployment insurance compensation after the first three months’ payments of RWT were deducted. After May 2009, these conditions were further  relaxed in favour of the workers. Accordingly, the most recent procedure provides that if workers had already benefited from the RWT for up to the maximum of one year, and were laid off after this, then the first six months’ RWT payments would not be deducted from the unemployment insurance compensation. Those RWT payments that belong to the following months would be subtracted from the compensation payment. The programme had a very weak start in 2005 when it was first initiated. As demonstrated in greater detail in Table 5.2, in 2005 only 21 employees were granted a total of TL 10,567 (US$8,000). In 2007, 40 workers were eligible for the programme and received a total of TL 22,051 (US$18,000). The number of people on the programme increased starting in March 2009 and accelerated to a peak of 82,439 people in June. In June alone a total sum of TL 23.925 million (approximately US$16 million) was disbursed.

M3125 - MESSENGER TEXT.indd 133

14/05/2013 13:33

134

Work sharing during the Great Recession

Table 5.2

RWT Fund: expenditures and employment gains Number of participating workers

Expenditures (thousands TL)

21 217 40 650 190,223 27,147 3,025

10.6 64.4 22.1 70.6 162,473.1 39,144.4 1,998.3

2005 2006 2007 2008 2009 2010 2011*

Note: * As of April 2011. Source: ISKUR monthly Bulletin, April 2011.

82,439

90,000.0

Number of participating workers

80,000.0

634

414

2,455

32,760

40,378

Jan.09 Feb.09 Mar.09 Apr.09 May.09 Jun.09 Jul.09 Aug.09 Sep.09 Oct.09 Nov.09 Dec.09 Jan.10 Feb.10 Mar.10 Apr.10 May.10 Jun.10 Jul.10 Aug.10 Sep.10 Oct.10 Nov.10 Dec.10 Jan.11 Feb.11 Mar.11 Apr.11 May.11

0.0

810

2,242

1,213

1,269

3,851

3,152

6,153

7,365

5,095

651

10,000.0

3,797

6,935

12,954

20,000.0

23,514

30,000.0

28,500

27,491

40,000.0

38,911

41,753

50,000.0

52,301

60,000.0

45,105

53,734

70,000.0

47,176

66,405

Expenditures (thousands TL)

Source: ISKUR Monthly Bulletins.

Figure 5.5

RWT Fund: participants and expenditures, 2009–2010

Over November 2008 to April 2011, a total of 268,081 persons from 3,632 enterprises benefited from the programme, with a total disbursement of TL 203.7 million (approximately US$125 million). The pattern of those who had been engaged in the RWT Fund together with the corresponding disbursements on a monthly basis is presented in Figure 5.5.

M3125 - MESSENGER TEXT.indd 134

14/05/2013 13:33

The Turkish experience with work-sharing policy, 2008–2010

Table 5.3

135

RWT Fund: selected indicators No. of applications

No. of approvals

Rejections and withdrawals

Enterprises Persons Enterprises Persons Enterprises Persons engaged engaged engaged November 2008 December 2008 January 2009 February 2009 March 2009 April 2009 May 2009 June 2009 July 2009 August 2009 September 2009 October 2009 November 2009 December 2009 January 2010 February 2010 March 2010 April 2010 May 2010 June 2010 July 2010 August 2010 September 2010 October 2010 November 2010 December 2010 January 2011 February 2011 March 2011 April 2011 Total

33 271 384 678 1,345 666 331 244 292 277 358 176 160 215 0 44 116 50 39 21 23 11 22 12 10 48 4 0 19 16 5,865

8,627 31,633 37,767 49,388 68,227 23,177 13,725 10,809 28,162 21,575 21,283 14,252 9,357 16,239 0 3,293 6,103 2,783 3,074 1,540 2,101 742 2,580 2,634 1,062 4,009 579 0 2,506 1,832 389,059

19 161 250 379 740 334 200 137 220 224 292 140 118 147 0 27 77 31 27 16 17 7 12 8 9 35 0 0 5 0 3,632

6,451 19,912 26,145 32,953 43,803 13,034 8,383 7,053 25,256 19,831 15,650 12,346 7,157 9,762 0 2,381 3,875 2,018 2,290 1,050 1,744 630 888 482 859 2,965 0 0 1,163 0 268,081

14 110 134 299 605 332 131 107 72 53 66 36 42 68 0 17 39 19 12 5 6 4 10 4 1 13 4 0 5 5 2,213

2,176 11,709 11,494 16,406 24,448 10,208 5,342 3,740 3,010 1,914 5,541 2,297 2,166 6,474 0 912 2,231 808 784 490 357 112 1,692 2,152 203 1,044 579 0 299 295 118,883

Sources: ISKUR, monthly Bulletins; ISKUR Vth General Assembly Report, November 2009, Ankara.

Table 5.3 further documents the situation of the applicants and approvals over the post-October 2008 crisis era. The number of applicants to the programme was 8,627 in November 2008, that is at the onset of the crisis. Applications grew in February and March 2009. From October

M3125 - MESSENGER TEXT.indd 135

14/05/2013 13:33

136

Work sharing during the Great Recession

2008 to April 2011, the total number of applications was 389,545, with total approvals of 268,081, which yields an acceptance ratio of 69 per cent. Across sectors, the RWT programme was mostly concentrated in three sectors: metal products, textiles, and the manufacture of motor vehicles and equipment. These three sectors had preserved a total of 110,000 jobs in 2009. This accounted for 58 per cent of those engaged in the RWT programme that year. In 2010, in line with the overall trend, the number of workers engaged in these three sectors declined to 13,721, which was about half of the total. A particular sector of interest is construction, one of the most labour-intensive sectors, which had suffered severely from the crisis. The RWT programme reached a total of 2,375 workers in 2009, but only 321 in 2010. Mining and quarrying also had a relatively low share, with 885 persons in 2009 and 88 in 2010. The overall sectoral distribution of the RWT programme over 2009 and 2010 is summarized in Table 5.4. The age and sex composition of those workers who had benefited from the RWT programme is tabulated in Table 5.5. In 2009, out of the total 190,000 participants, 160,000 were men, with a majority being middle-aged. Interestingly, in 2009 the number of young women in the RWT programme (aged 20–24 years) exceeded the number of men of the same age group. This is the only exception to the observation that the RWT participants across all age groups were predominantly male. This follows the overall trend of job applications to ISKUR. According to data by Dayıoglu and Ercan (2010), on average, in 2008 and 2009 only about 30 per cent of the total applicants to ISKUR were women. Job placement rates were 15.6 per cent for males (156,000 in one million male applicants), and 12 per cent for females (50,000 in 415,000 applicants). Comprehensive as it is, a major deficiency of the overall employmentpreservation programmes can be found in their short-term structure in general, and the rather low participation rates in comparison to their fiscal costs. A major shortcoming of the package was that these jobs were actually created for a maximum of six months only, and lacked formal channels for further follow-up towards formal employment. Karadeniz (2010), for instance, argues that a major mishap had been the fact that many training courses that were designed as part of the RWT programme were provided without due marketing research about the job prospects. He is one of those who argue that ISKUR overall is understaffed and lacks the infrastructure and means to provide such services for its clientele. There are also widespread concerns over employers’ lack of interest, in general, towards ISKUR’s education campaigns. In interviews with the social partners, a major concern was that a shift of focus

M3125 - MESSENGER TEXT.indd 136

14/05/2013 13:33

The Turkish experience with work-sharing policy, 2008–2010

Table 5.4

137

Sectoral distribution of persons engaged in RWT programmes, 2009–2010

Sector

2009

2010

Fabricated metal products, excluding basic metals and machinery and components Textiles, textile and leather products Manufacture of transport vehicles and equipment Machinery and equipment manufacturing, not classified elsewhere Rubber and plastics Food products, beverages and tobacco Design and repair of machinery and other machine products Wholesale and retail trade Manufacturing of electrical machinery Manufacturing of computers and optical equipment Manufacturing of wood products and paper and publishing Construction Other services Transport and depots Law and accounting services; technical engineering services; managerial services Manufacturing of chemical products Television broadcasts and arts services Administrative support services Mining and quarrying Hotels and lodging and restaurants Sewage and water services Electricity and gas Agriculture, fisheries and animal husbandry Other professional services Manufacturing of pharmaceutical products Education services Finance and insurance Cultural activities Human health services IT and other communications Coal mining and crude petroleum Real estate services Telecommunications Total

42,232

4,925

41,341 26,620 19,267

4,818 3,978 3,928

13,637 11,276 7,094

2,663 1,706 1,238

5,135 3,323 2,897 2,556

662 502 428 350

2,375 1,906 1,850 1,354

321 315 238 207

1,298 1,143 1,050 885 877 493 416 276 255 180 159 153 68 52 13 8 6 3 190,198

196 152 126 88 83 77 59 43 35 19 5 0 0 0 0 0 0 0 27,162

Source:

ISKUR, Monthly Bulletins, 2011.

M3125 - MESSENGER TEXT.indd 137

14/05/2013 13:33

138

Table 5.5

Work sharing during the Great Recession

RWT participants by age and sex composition, 2009 and 2010

Age

15–19 20–24 25–29 30–34 35–39 40–44 45–64 651 Total

2009

2010

Men

Women

Men

Women

75 2,404 34,719 46,567 36,604 24,768 14,767 189 160,093

63 3,572 7,294 7,461 5,466 3,700 2,488 61 30,105

39 646 4,446 6,197 5,776 3,791 1,942 24 22,861

21 505 964 1,012 835 567 391 6 4,301

Source: ISKUR, Monthly Bulletins, 2011.

is severely needed if the labour force is to switch from its traditional base towards maintaining a supply of technical expertise based upon modern job skills. Finally, it is generally agreed that despite its clear focus otherwise, the RWT programme could not deliver a strong remedy to the extent of the informalization of the labour markets, especially in the rural sectors. Indeed, informal, unregistered employment without any social securitization coverage has been an indispensable trait of Turkish labour markets over the 2000s. The ratio of the unregistered to total employment had fluctuated between 45 and 50 per cent over the decade, and apparently constituted one of the major (arguably the main) mechanisms of adjustment. As informal labour bore the burden of adjustment, wage remunerations fell and quality of jobs unavoidably worsened. Figure 5.6 reports on the available data from the Turkstat database regarding the extent of informalization and duality in the labour markets. Data reveal that in 2010, unregistered workers in agriculture numbered 4,915,000 persons. With an addition of 4,857,000 workers employed in the non-agricultural sectors under unregistered conditions, this brings the total unregistered employment to 9.8 million people, which is 43.25 per cent of total employment. In fact, unregistered informal employment gains mostly explain gains in total employment in the first half of 2010. As Table 5.6 shows, there was a rapid expansion in total employment from the onset of the crisis in 2008 to 2010. Yet a close inspection of the status of these data reveals that of the 1,403,000 newly employed workers, 552,000 were ‘employed’ without

M3125 - MESSENGER TEXT.indd 138

14/05/2013 13:33

The Turkish experience with work-sharing policy, 2008–2010 12,000

139

Agriculture Non-agriculture

10,000

8,000 4,708

5,119

5,285

5,132

4,814

4,825

4,915

5,136

4,547

4,307

4,290

4,406

4,503

4,857

2004

2005

2006

2007

2008

2009

2010

6,000

4,000

2,000

Source: Turkstat, HLF Surveys.

Figure 5.6

Table 5.6

Unregistered (informal) employment (1,000 persons)

Employment and unregistered informal employment, 2008–2010 Total

Employed 2008 21,193 2010 22,596 2010–2008 difference 1,403 Unregistered informal employed 2008 9,220 2010 9,772 2010–2008 difference 552

Agriculture

5,016 5,683 667

Non-agriculture Industry Construction 4,441 4,496

1,241 1,431 736

Services

10,495 10,986  

4,406 4,857 451

4814 4915 101

 

Source: Turkstat, HLF Surveys.

any social security coverage, as ‘unregistered’ employees. This phenomenon was particularly acute in agriculture, where of the total 667,000 new employment, 451,000 employees were unregistered as informal workers, with a ratio of 68 per cent.

M3125 - MESSENGER TEXT.indd 139

14/05/2013 13:33

140

4.1

Work sharing during the Great Recession

Two Case Studies

More nuanced information about these policies can be found in certain applications that were presented to the legal authorities for decisions. The examples presented are based on law reports from the Sakarya Labour Court in Turkey. Firm A was a producer of agricultural motor vehicles and tractors. It employed 165 workers in production, together with an additional 60 managerial workers, totalling 225 employees in January 2007. Starting in mid-2008, the firm experienced serious difficulties in sales revenues. The number of tractor vehicle orders fell from 1,065 in 2007 to 887 in 2008, and to 341 in 2009. The sales orders recovered in the first five months of 2010 to 1,169, but the downswing cycle was quite severe in 2009 as the sales revenues declined by 61.5 per cent relative to 2008. In response, the firm management applied for RWT coverage on 30 March 2009 and the application was granted by ISKUR. RWT coverage was set for a total of six months from 30 March to 26 July 2009. However, the ending date was extended for an additional 12 months to 26 March 2010. Employment of production workers had fallen initially to the level of 125–130 workers in the period from September 2007 to September 2008. In January 2009 production workers’ employment fell to 100, and then to 85 in March of 2009, when the firm decided to apply for RWT. Employment stabilized at that level for the rest of 2009, but another cut took place to reduce the number of workers (in production) to 70 in March 2010. Employment of managerial workers was also reduced, but at a very modest scale. The firm had not resorted to overtime work in a significant fashion and generally asked workers to take vacations without pay. Annual production of tractors fell by 30.1 per cent in 2008 and by an additional 64.3 per cent in 2009. The production level was virtually zero during the first quarter of 2009. It can be argued that the firm was in a quite dire situation in the spring of 2009 and the RWT programme succeeded in keeping employment and production levels afloat. By March 2010, the firm’s production and employment levels were almost half of their pre-crisis levels but were at least on a positive trajectory. Firm B was a producer of furniture and forestry products. The firm was adversely affected by the crisis in its export sales, where export revenues of its wood tiles and wood planks fell by more than 60 per cent in 2008. In response, the firm’s management successfully switched to the domestic market and the share of domestic sales increased from 53.5 to 97.2 per cent

M3125 - MESSENGER TEXT.indd 140

14/05/2013 13:33

The Turkish experience with work-sharing policy, 2008–2010

141

in 2009. Total monthly sales revenues of the company were on the order of TL 10 million in summer 2007 but were reduced to TL 4.2 million in January 2008. With a decisive switch to the domestic market, monthly sales revenues accelerated to TL 13 million in July 2008, but receded back to the TL 4 million during the winter months of 2009. This swing was mostly due to the seasonal effects of construction activities in the domestic economy. The firm applied for the RWT programme three times over the course of March 2009 to June 2010. The first application was dated 12 March 2009 and was granted RWT coverage by ISKUR covering six months from 13 April to 10 October 2009. However, the firm’s management changed its view and asked for a refusal of the RWT on 16 March 2009. The likely argument here was that the firm’s management was concerned with the ‘negative’ image of being involved with the privileged treatment of the RWT, especially given fierce domestic competition. This, however, was not the end of the story. Again, the seasonality of the domestic demand forced the firm to reapply for RWT on 13 October 2009 and then again on 31 December 2009. Both of these additional applications were rejected by ISKUR on the grounds that the firm’s economic activity was not affected significantly by the global crisis conditions from 2009. Enterprise-level data on employment and production suggest that ISKUR’s analysis of the firm activity after 2009 was sound. Employment of production workers fell by 6.8 per cent in 2008 and increased by 7.6 per cent in 2009 in annual terms. Production of wood tiles, on the other hand, is reported to have increased by 32.7 per cent in 2008, and by a further 32.3 per cent in 2009. The RWT application, then, seems to have been based on seasonality of total sales (see Figure 5.7), a factor that had not been interpreted favourably by ISKUR.

5.

VOICES OF THE CONSTITUENTS

Almost everyone concerned with industrial relations and overall conditions of the work environment in Turkey agrees that high unemployment rates along with low participation in the workforce, concerns over macroeconomic stability, widespread informalization and fragmentation are among the key problems of the labour markets. These concerns have been high on the agenda of both the employers’ associations and the labour unions and their confederations. However, the solutions offered reflect very different perspectives. Employers’ associations and their affiliates called for a more flexible labour market, with deregulation both at the hiring end and for termination of contracts, more flexibility in the initial

M3125 - MESSENGER TEXT.indd 141

14/05/2013 13:33

M3125 - MESSENGER TEXT.indd 142

14/05/2013 13:33

142

Figure 5.7

Thousand TL

Total sales revenues, Firm B

0

2,000

4,000

6,000

8,000

10,000

12,000

14,000

16,000

Jan.07 Mar.07 May.07 Jul.07 Sep.07 Nov.07 Jan.08 Mar.08 May.08 Jul.08 Sep.08 Nov.08 Jan.09 Mar.09 May.09 Jul.09 Sep.09 Nov.09 Jan.10 Mar.10

The Turkish experience with work-sharing policy, 2008–2010

143

design of the contracts, and a more stable macroeconomic environment with better incentives for entrepreneurship. Trade unions, on the other hand, emphasize more social protection, job security and wage increases that reflect gains in productivity and welfare. The RWT initiative seems to have been one of the few policy interventions welcomed by both parties. Although objections and complaints about the ongoing practices remain, key elements of the programme had general support from all parties engaged. The Confederation of Turkish Labour Unions (Turk-Is) is the largest confederation of workers’ unions in Turkey. Turk-Is clearly indicates that one of its main principles is to promote necessary policy measures for protecting and enhancing employment opportunities against the adverse conditions of the global crisis. Thus, in principle Turk-Is supported the RWT and similar policy interventions. It also highlights the importance of maintaining high-quality jobs, while at the same time monitoring the application of the policy against fraud and mis-use of the unemployment insurance fund for purposes outside its main rationale. Turk-Is’s main objection to the employment packages thus far has been that they were mostly employer-focused. According to Turk-Is, reductions in social security premiums did not reach their desired outcome to create sufficient employment, and this was recognized by government officials. The reductions in premiums should be devised such that they take into account the rights of the employees, and they should be applicable only to employers whose labour force is mostly (at least 50 per cent) composed of organized/formal labour. Specifically regarding the day-to-day implementation of the RWT programme, Turk-Is favoured the recent Omnibus Law amendments to include sectoral and regional crisis conditions as part of the RWT coverage. However, in order to introduce more objectivity and reduce misunderstandings, Turk-Is suggested referral to the State Planning Organization (now the Ministry of Development) and Turkstat data. Furthermore, there was also an objection to the omission of the phrase ‘conditions for continuous reduction or complete stopping of economic activity’ from the original law document (No. 4447). Turk-Is claims that omitting the word ‘continuous’, could lead to mis-use and fraud, and seems to contradict the main purpose of the programme, which is to protect employee rights.9 Turk-Is has also suggested that if the workers at the enterprise seeking RWT assistance were unionized, the management could start RWT procedures directly with the workers’ union. According to Turk-Is, this would promote formalization and increased incentives for unionization. It would also reduce bureaucracy, and help to eliminate corruption and fraud. The Confederation of Progressive Trade Unions of Turkey (DISK) is

M3125 - MESSENGER TEXT.indd 143

14/05/2013 13:33

144

Work sharing during the Great Recession

also favourably disposed towards the package. DISK, as clearly stated in its reports and in our interviews, considers that the measures taken in the face of the crisis were all in favour of the employers, disregarding the demands of the workers. In DISK’s view, flexible working schemes hardly affected employment. On the contrary, the firms which had benefited from the RWT allowances started to lay off their workers by the end of the payment period. During such periods, the demand of the employers’ associations for flexible working schemes was based in their view on nothing but opportunism. Demands for flexibility, ‘flexicure’ labour markets, and private employment offices were always recommended by employers to increase employment within this work scheme. Nevertheless, DISK’s view was that these policies have had little effect in decreasing unemployment and bore little relationship to work sharing. DISK is of the opinion that the Unemployment Insurance Fund is one of the mechanisms that was put into force based on social concerns. But in practice, this fund has been used in other economic areas which were beyond the scope of the fund’s essential functions. The use of the fund outside of its intended purpose and the strict conditions had kept the fund from fulfilling its obligations, as well as preventing authorities from being able to correct injustices and substitute lost wage income. DISK was favourably disposed towards the RWT package as long as it was not an instrument for introducing deregulation, flexibility and informalization of the employment conditions.10 The Hak-Is Trade Union Confederation generally claimed that the measures taken as a reaction to the crisis should have been complemented by employment-focused policies, and furthermore, economic and social policies should have been implemented simultaneously. Otherwise, the consequences on the labour market, which is severely distorted already, would be devastating. Therefore, a social dialogue mechanism devised among the social partners on egalitarian grounds and sustaining the flow of information was viewed as important in reaching common ground for policies to work. In Hak-Is’s view, in addition to the rearrangements of the Unemployment Insurance Fund, RWT schemes and severance payments served as passive policy measures that could eliminate the adverse effect of the crisis on the most vulnerable groups. These groups mainly consist of low-income groups or those who are living in poverty. For this reason, especially during the era of the crisis, a ‘positive discrimination’ principle should be adopted in social protection policies. In order to minimize the damage of the crisis to the most vulnerable, measures such as direct income support and family support should be planned and activated.11 On the employers’ side the major organization is TISK, the employers’

M3125 - MESSENGER TEXT.indd 144

14/05/2013 13:33

The Turkish experience with work-sharing policy, 2008–2010

145

confederation, whose overall stance in interviews was that, although the measures taken by the government are appreciated, they were perceived to be inadequate. Only a couple of these measures were seen to be effective, such as the RWT allowance and sales tax reductions. But, the effect of another important measure that TISK proposed, the reductions in the employers’ share of social security premium payments, has been offset by other regulations. TISK also favoured the recent amendments to the RWT brought by the Omnibus Law. The additional recognition of ‘sectorial crisis conditions’ and the application of the RWT to include ‘individual departmental units as well as the whole of the enterprise’ and omission of the ‘requirement of continuity for at least four weeks’ were found to be quite acceptable by the TISK members. TISK, in particular, was very keen to extend the coverage of the RWT to bring it on par with the applications in certain EU countries such as Belgium, Switzerland, France and Germany, where short-time work eligibility includes technical breakdowns, bad weather conditions, granting simultaneous annual vacations, or difficulties in obtaining raw materials and intermediate inputs. Another set of issues that TISK was keen on addressing were bureaucratic delays in setting the beginning and ending periods, and the lack of an overall grievance mechanism to possibly challenge the decisions set forth by the Ministry of Labour inspectors. As it currently stands, the verdict by the ministry is final and no petition is allowed against the reports prepared by the inspectors.12 The Istanbul Chamber of Industry (ISO) also took a positive view of the RWT, noting that reduced-time allowance has had a positive effect on both the enterprise and the workers. Of the 200,000 people who have benefited from this arrangement, most were employed by big companies, indicating that due to information and administration deficiencies small and medium-sized enterprises were outside the scope of this system. Shorttime allowances and other flexible working schemes would also be beneficial for both the enterprises and the workers if they were made permanent, rather than just temporary arrangements that were specific responses to economic crisis situations. ISKUR (the Employment Agency) has a direct mandate over the RWT framework and is quite enthusiastic for its wider application. ISKUR regards the RWT framework as one of the key programmes achieving social cohesion among the social partners at low fiscal cost. ISKUR’s management regards the training component as very important, and is now developing a programme where the prospective employers will participate in employment-guaranteed training programmes. The costs of these programmes will be met from the Unemployment Insurance Fund

M3125 - MESSENGER TEXT.indd 145

14/05/2013 13:33

146

Work sharing during the Great Recession

as administered by ISKUR. Yet, in this regard it should also be noted that ISKUR is a relatively small public agency with around 2,827 personnel (509 in the Ankara headquarters) in 2009 (ISKUR, Monthly Bulletins, 2010). The understaffing remains an acute problem in the management of diverse functions such as administering unemployment insurance compensation, RWT eligibility and job accession opportunities.

6.

CONCLUSION: ASSESSMENTS AND POLICY CHALLENGES FOR THE FUTURE

Turkey was hit by the global crisis in its real production sectors, notably in industry. Over the first three months of the crisis that began in November 2008, industrial output fell by 40 per cent and industrial employment fell by 400,000 jobs. Against this adverse development, the government was criticized for its delayed response, wasting valuable time. The main policy intervention came only after May 2009 and, by then, Turkey faced a severe reduction in fiscal resources due to the fall in tax revenues. Ercan et al. (2010) concluded that the policy response was too late and was relatively less efficient in comparison with its OECD counterparts. Nevertheless, RWT was an important component and a major experience. It was also a welcome policy intervention appreciated by the employers’ and the employees’ associations, and the government bureaucracy. The programme that started in March 2009 involved a total of 82,439 people in June 2009. From November 2008 to April 2011, 268,081 people from 3,632 enterprises had benefited from the programme with a total disbursement of TL 203.7 million (approximately US$125 million). Given these figures, the estimated cost of the programme amounts to US$727 per job preserved. This is a significantly lower cost for Turkey compared with many other OECD countries following similar programmes, and can be mostly explained by the low minimum wages rates at which the RWT payments are scaled. It is hard to quantify the net employment gains from the programme, as data from ISKUR are incomplete at the time of writing. An independent study conducted by the ISKUR researchers, however, attempted to examine this question by keeping track of the citizenship ID numbers of those engaged in the RWT programme. By continuously keeping track of the net additions to the programme every month through meticulous tracking of the ID numbers, ISKUR researchers revealed that as of 31 May 2011 a total of 152,261 people were followed over a year’s involvement in the programme. Of these, 51,726 had left their job by the end of the year. Thus, 100,535 jobs could have been protected through the RWT

M3125 - MESSENGER TEXT.indd 146

14/05/2013 13:33

The Turkish experience with work-sharing policy, 2008–2010

147

programme from May 2010 to 2011. A straightforward application of this finding to the overall labour market would suggest that if the RWT programme had not been utilized in 2010, the total number of persons unemployed would have been 3,147,000 rather than 3,046,000. Thus the unemployment rate would have increased to 12.3 per cent, rather than the realized rate of 11.9 per cent in 2010, which suggests a gain of 0.4 percentage points off the unemployment ratio for that year. Thus, in summary it can be argued that the RWT programme was relatively successful in securing jobs during the most severe months of the global recession. Furthermore, it was operated at a relatively low fiscal cost. The key question at the current juncture, however, is whether the RWT scheme could be operational beyond crisis conditions, as an employment generation programme. How can it be redesigned to serve as a tool aiding enterprises that face the costs of restructuring and adjustment? How can it be restructured to accommodate various individually specific needs of the workers, as well as the business considerations of the enterprise management? Here a distinction, highlighted by Tucker and Folkard, will prove useful. Tucker and Folkard (2012: 29) caution that it is important to distinguish between flexible working-hour arrangements controlled by the employer (‘employer-led variability’) and those that permit employees to have discretion regarding their work schedules (‘employee-led flexibility’). They report that research studies regarding flexible working-hour arrangements based on employer-led variability tend to show negative impacts on workers’ health and well-being, while those based on employee-led flexibility typically show positive effects on a range of measures, related to not only occupational safety and health, but also work–life balance and organizational performance (draft paper cited in ILO 2011: 58). One of the significant challenges of the Turkish experience was that the RWT programme revealed an undesirable feature for employers that restricted them to a specific pattern and duration of time reduction. Commanding the enterprises via a specific timetable for a reduction in hours proved constraining for the management, especially under the global crisis conditions of considerable uncertainty with regard to demand and overall business conditions. A further challenge was the concern to protect the company’s ‘image’ against any likely reputation loss due to its appearance as ‘in need’ of public protection in the eyes of the potential customers and creditors. Our interviews revealed that in many instances, the managers had weighed their decisions to request RWT funds against such reputation concerns in a competitive and unsympathetic market. Within a more general context, Baker (2011) suggested that such problems can be avoided if employers were provided with more discretion

M3125 - MESSENGER TEXT.indd 147

14/05/2013 13:33

148

Work sharing during the Great Recession

over ‘which workers were to be covered’, and ‘how much time is actually reduced’. According to Baker, this would likely mean that the payments are made based on scheduled reductions in hours and then adjusted in accordance with the actual reduction in hours on a quarterly or annual basis. The two very clear positive aspects of the programme were, first and foremost, the ability of the RWT scheme to maintain continuation of the employment contract, thus keeping the worker attached to his/her job. Second, the programme introduced clear incentives for increased formalization of the workforce. Clearly, in the end the main objective ought to be: to structure a short time compensation programme in ways that make it more desirable than laying off workers from the standpoint of employers, even at a somewhat higher cost to the government than the current unemployment insurance system. (Ibid.: 5)

Accordingly, since there is a clear welfare gain publicly to maintain employment status and to guarantee formalization within labour markets, it would be desirable to grant a premium to the participants in excess of what the workers would be receiving in unemployment insurance benefits. This could take the form of an employer credit, and be subject to control both by the authorities and also by the workers’ union. Such a scheme would give the employers the flexibility to redesign work hours as well as tasks. Such a remuneration scheme could further be extended to cover basic health and benefit pensions for a broader timeframe. All these would also likely encourage employee willingness to consider participation in short-time work and work-sharing programmes.

NOTES *

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

Prepared as part of the ILO project, ‘Work Sharing Policies and Programmes in the Context of the Global Recession and the Resulting Global Crisis’, Ankara. I am indebted to Jon Messenger, Naj Ghosheh, Mustafa Biçerli, Fatma Demir, Enis Bağdadioğlu, Murat Özveri, Umit Efendioglu, and to colleagues at Bilkent and the ILO office Ankara for their invaluable comments and guidance. Needless to say, all views and possible errors are solely mine. TL is the Turkish lira, that is, the national currency of Turkey. Turkish membership in the ILO dates to 1932. See Demιrhan et al. (2011) for a more detailed assessment. A good source in English is Tunali (2003: ch. 7). ‘Hak’ translates as ‘right’ as in workers’ rights. It is one of the names for God in Arabic. The SPO was dissolved and became the Ministry of Development in June 2011. Turkstat report Household Labour Force Survey data as three-month moving averages. Thus, the ‘January 2011’ data refer to the arithmetic average of December 2010 to February 2011.

M3125 - MESSENGER TEXT.indd 148

14/05/2013 13:33

The Turkish experience with work-sharing policy, 2008–2010 7. 8.

9. 10. 11. 12.

149

See BETAM research notes available at www.betam.bahceshir.edu.tr. This rather unconventional term describes the sweeping amnesty and many other seemingly related policy measures and litigation matters lumped together under Law Code 6661. Given its broad ambit, it was referred to as the ‘bag law’ in the popular Turkish media. Interview with Enis Bagdadioglu, chief researcher at Turk-Is headquarters, Ankara, 2011. Interview with research department personnel at the DISK Ankara Office, 2011. Interview with research department personnel at the Hak-Is Ankara Office, 2011. Interview with the research department personnel at the TISK Ankara Office.

REFERENCES Agénor, P.R., H.T. Jensen, M. Verghis and E. Yeldan (2007), ‘Disinflation, fiscal sustainability, and labor market adjustment in Turkey’, Chapter 7 in A. Richard, A. Izquierdo and H. T. Jensen (eds), Adjustment Policies, Poverty and Unemployment: The IMMPA Framework, Oxford: Blackwell. Bahçe, S., G.Y. Ozdemir, E. Voyvoda, M. Ozdemir and M. Candan (2011), Labour Policies: What is Happening and What Should Be Done? (in Turkish), Ankara: Belediye-Is. Baker, D. (2011), ‘Work sharing: the quick route back to full employment’, CEPR Research Paper, Washington, DC, June. Boratav, K. (1991), Social Classes and Distribution of Income in Turkey (in Turkish), Istanbul: Yeni Dizi: 7. Boratav, K., E. Yeldan and A. Köse (2002), ‘Globalization, distribution and social policy: Turkey, 1980–1998’, in L. Taylor (ed.), External Liberalization  and  Social Policy, Oxford and New York: Oxford University Press, pp. 317–62. Dayιoglu, M. and H. Ercan (2010), ‘Building more effective labour market policies and institutions in emerging market economies: Turkey’, Ankara, ILO Office for Turkey, December, mimeo. Demιrhan, A.E., C. Katz, J. Petras, M. Özveri and Y. Akkaya (2011), Struggle for Rights with Its Historical and Theoretical Aspects (in Turkish), Istanbul: Nota Bene Pub. Ercan, H. (2010), ‘The impact of the global financial crisis on employment in Turkey’, in Ercan et al. (eds), pp. 73–97. Ercan, H. and A. Tansel (2006), ‘How to approach the challenge of reconciling labor flexibility with job security and social cohesion in Turkey’, paper prepared by the Turkish expert group for European Council’s FORUM 2005 (Strasbourg). Ercan, H., E. Taymaz and Erinc Yeldan (eds) (2010), Crisis and Turkey: Impact Analysis of Crisis Response Measures, Ankara: ILO Office for Turkey. ILO (International Labour Organization) (2011), ‘Working Time in the Twenty First Century’, Report for discussion at the tripartite meeting of experts on working-time arrangements, Geneva, 17–21 October. ISKUR (Turkish Employment Agency), Monthly Bulletins, Ankara. ISKUR (Turkish Employment Agency) (2009), Report to the Fifth General Assembly, Ankara. ISSA (Independent Social Scientists’ Association) (2011), Annual Report: Wage

M3125 - MESSENGER TEXT.indd 149

14/05/2013 13:33

150

Work sharing during the Great Recession

Labour and Capital under the Great Recession (in Turkish), Istanbul: Yordam Pub. Karadeniz, O. (2010) ‘A review of the measures on reduction of the effects of the global crisis on employment and their efficiency’ (in Turkish), Seminar presented at Pamukkale University, mimeo. OECD (2008), OECD Economic Surveys: Turkey, vol. 2008/13, Paris, September. Sonmez, M. (2011), Crisis and Poverty, Istanbul: Iletisim Pub. Taymaz, E. and S. Ozler (2005), ‘Labour market policies and EU accession: problems and prospects for Turkey’, METU/ERC Working Paper no. 04/05, Ankara. TISK (Turkish Confederation of Employers’ Unions) (2010), ‘Report to the 24th Congress’, Ankara, December. Tucker, P. and S. Folkard (2012), ‘Working time, health, and safety: a research synthesis paper’, Conditions of Work and Employment Series No. 31, ILO, Geneva. Tunali, I. (2003), ‘Background Study on the Labour Market and Employment in Turkey’, prepared for the European Training Foundation, mimeo. Turkstat (various years), Household Labour Force Participation Surveys, Ankara, available at: www.tuik.gov.tr. Voyvoda, E. and E. Yeldan (2001), ‘Patterns of productivity growth and the wage cycle in Turkish manufacturing’, International Review of Applied Economics, 15 (4): 375–96. Yeldan, E. (2001), Turkish Economy under Globalization (in Turkish), Istanbul: Iletisim Pub. Yeldan, E. (2010), ‘Crisis impact on the Turkish economy’, in Ercan et al. (eds), pp. 9–40.

M3125 - MESSENGER TEXT.indd 150

14/05/2013 13:33

6.

Results of the implementation of the suspension and partial unemployment insurance programmes in Uruguay, 2009–2010 María José González Fernández

1.

INTRODUCTION

Two worker retention and work-sharing programmes were used in Uruguay to address the contraction in labour demand by companies affected by the global crisis that began in 2008. These programmes are components of the general unemployment insurance subsidy scheme. The mechanism known as ‘partial unemployment insurance’ (PUI), on the one hand, is a subsidy granted in cases where there is a partial suspension of the worker’s activities (a reduction in the number of hours worked in the week and devoted to training). Created in 2009, it targeted specific sectors and companies affected by the crisis. On the other hand, since the introduction of unemployment insurance in 1981 there has been a subsidy available for cases of full suspension of a worker’s activities during the month. All companies and economic sectors were eligible for the subsidy. Considering all the forms of subsidy for full or partial suspension of activities that were used as work-sharing subsidies, a monthly average of 0.7 per cent of private sector salaried workers participated in these programmes in the period studied (June 2009–December 2010), which is equal in turn to 30 per cent of the beneficiaries of unemployment insurance for dismissal over that same period. This represents a considerable impact in relative terms as a response to the downturn in economic activity faced by some sectors during that period. This chapter will examine these two worker-retention and work-sharing mechanisms that were implemented in Uruguay under the general unemployment insurance scheme as a response to the contraction in labour demand by companies. Section 2 provides an overview of the country’s economic and labour market conditions. The leading characteristics of

151

M3125 - MESSENGER TEXT.indd 151

14/05/2013 13:33

152

Work sharing during the Great Recession

the general unemployment insurance scheme are outlined in Section 3. Section 4 describes the work-sharing subsidy mechanism known as Partial Unemployment Insurance introduced in 2009, and the main results of its implementation. Section 5 presents the results of another worker-retention subsidy instrument for temporary reductions in company activity: the Unemployment Subsidy granted for cases of suspension of work activities (or unemployment insurance on the grounds of suspension), which was created in 1981 as one of the available unemployment insurance options. These two subsidy programmes are then compared, their impact in terms of job preservation is assessed, and an estimate of their costs is given (Sections 6–8). Finally, Section 9 sets out recommendations based on the assessment of the operation and results of both worker-retention subsidy programmes as a response to the negative effects of the global economic crisis on the labour market.

2.

ECONOMIC AND LABOUR MARKET CONDITIONS

Uruguay is generally viewed as a small economy and with a matching population. In 2010 the country had a (projected) total population of 3,356,584 inhabitants, with a working-age population (aged 15 and over) of 2,654,602. Gross domestic product (GDP) for 2010 stood at US$40.27 billion, with a per capita GDP of US$12,000. With regard to the distribution of employment by sector of activity, the services sectors have the largest share of labour demand. The agricultural sector and the manufacturing industry account for 12.7 and 14 per cent of employment, respectively, and the remaining sectors (various services) represent almost three-quarters of all employment (Figure 6.1). In the 2005–10 period, important public policy amendments were introduced, many of which played a key role in mitigating the national repercussions of the recent global crisis. In 2009, Uruguay was recognized by the ILO as one of the only five countries in the world that continued to grow and reduce unemployment. While the GDP growth rate slowed significantly, the economy still grew by 2.9 per cent, and the unemployment rate dropped 0.4 percentage points to 7.3 per cent. There was also a considerable increase in real wages (5.6 per cent). This positive performance was driven by several factors. These include the implementation of economic policies focused on reducing economic and social vulnerabilities and generating sustainable macroeconomic equilibria; greater diversification of exports; a number of structural reforms implemented as of 2005, which strengthened the country’s ability

M3125 - MESSENGER TEXT.indd 152

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010 Household services 8.6%

Primary activities 11.5%

Other community services 4.7% Soc. serv. & health 7.0% Education 5.6% Public Adm. & Defence 5.9%

153

Manufacturing & electricity, gas & water 14.0% Construction 7.3% Trade, Rest. & Hotels 21.5%

Banking, finances, bus. services 8.4%

Transport and communications 5.4%

Source: Prepared by author with INE (Instituto Nacional de Estadística: National Statistics Institute) data.

Figure 6.1

Distribution of employment by activity sector (2010)

to withstand negative shocks; institutional changes in labour market operation; and increased levels of employment protection. In the context of industrial relations in Uruguay, sectoral or industry trade unions are grouped in a single confederation of workers called PITCNT. From 1985 to 2001, the country experienced a steady drop in union membership, with the unionization rate (the ratio of labour confederation members to employed salaried workers) falling from 38 per cent in 1985 to 16 per cent in 2000. The unionization rate of private sector salaried workers fell by 60 per cent, while that of public sector employees decreased by 20 per cent. Various explanations have been given for the drop in union membership during the 1990s. The government’s withdrawal from collective wage bargaining is considered a decisive factor, as is the shift from sector- to company-level bargaining, which ‘modified one of the historical structuring features of the Uruguayan labour movement’.1 As of 2005, a new institutional framework for labour relations was put in place and a new scenario for union activity emerged. With the new government came a change in the wage-setting policy that had been applied until then. These changes included the reinstatement of wage councils by industry in the private sector, the adoption of protective measures for union activities, and the elimination of a decree that authorized the evacuation of workplaces in the event of worker occupation. As a result of these institutional changes, trade unions gained greater power and the number of unionized workers increased significantly, reversing the decreasing trend experienced in the 1990s. An increase of around 16 per cent in

M3125 - MESSENGER TEXT.indd 153

14/05/2013 13:33

154

Work sharing during the Great Recession 65 17

61

15

59 13

57 55

11

53 9

51 49

Unemployment rate

Employment & activity rates

63

7

45

1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 sem 2011

47

Activity

Employment

5

Unemployment (right)

Source: INE.

Figure 6.2

Activity, employment and unemployment rates (country total, urban until 2005)

union membership was estimated from 2003 to 2005, a growing trend that appears to have continued in subsequent years. From 2005 to 2010, the economy was marked by strong growth, a considerable expansion in public social policies aimed at reducing poverty, destitution and inequality,2 and a substantial improvement in labour market indicators, achieving historic lows in unemployment rates (Figure 6.2) and better-quality employment, with a significant expansion of social security coverage. However, unemployment rates remained high among certain segments of the population, such as rates among women and young people (Appendix Tables 6A2.1 and 6A2.2). With respect to the evolution of economic activity, as of 2004 Uruguay’s economy entered a phase of sustained high growth, after the marked economic contraction of the previous years.3 This expansion was driven by growth in foreign and domestic demand, and it extended to every sector of activity. This steady growth was only affected in 2009, as a result of the severe crisis that hit the global economy starting in the second semester of 2008 (BCU 2011). Uruguay’s economy felt the repercussions of the crisis in the first quarter of the year 2009, when GDP fell by 1.9 per cent in seasonally adjusted terms. The impact on the economy (and the labour market) was more moderate than was initially feared. Technically speaking, the

M3125 - MESSENGER TEXT.indd 154

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

155

economy never entered a recession, as in the second quarter of 2009 it experienced a slight growth (1.3 per cent), and in the following quarters it recovered its dynamic level of activity. In 2009, the economy grew moderately on average (ibid. 2011). In addition, the unemployment rate, which had climbed to 8 per cent during the second quarter of 2009, dropped again in the following quarters, bringing down the annual average (Table 6A2.1). On another level, collective bargaining between unionized workers and employers may also have provided solutions that mitigated the impact of the tightening phases of the economic cycle on employment. In particular, in Uruguay’s current sector-level bargaining (wage councils by sector of activity), the introduction of safeguard clauses in bargaining agreements provides an instrument for tripartite reconsideration of wage agreements in the event of significant changes in economic conditions. An analysis of the sectoral wage agreements that emerged from the 2010–11 wage council negotiations reveals a certain maturity in Uruguay’s labour relations. In a context of uncertainty concerning the evolution of the world economy, 56 per cent of all sectors negotiated safeguard clauses. Also, 42 per cent of the agreements merely provide for the possibility of convening a new wage council meeting, upon request of any of the parties, if a substantial variation in economic conditions occurs, while virtually all agreements lack clauses to mitigate the impact of negative shocks on employment. Another instrument for mitigating the negative effects of these shocks on employment consisted of the incorporation of a sectoral performance adjustment component into wage adjustment agreements, with the aim of moderating real wage rigidity. Most of the wage adjustments negotiated included a fixed percentage for real increases. However, of the 230 wage agreements that were in force in 2009 (based on the 2008 wage round), only 10 took sectoral performance into account. Lastly, the measures to encourage working-time reduction were introduced with the aim of tempering the negative effects of the economic downturn on employment.

3.

URUGUAY’S UNEMPLOYMENT INSURANCE SCHEME: THE FOUNDATION OF WORKSHARING MEASURES

Two subsidy measures for worker retention and work sharing were implemented as a response to the cyclical contraction in labour demand by companies from sectors affected by the global crisis. The logic underlying the design and operation of these programmes is that of a trade-off in the short term between work hours and job preservation. The ultimate aim

M3125 - MESSENGER TEXT.indd 155

14/05/2013 13:33

156

Work sharing during the Great Recession

BOX 6.1 GENERAL UNEMPLOYMENT INSURANCE SCHEME: AMENDMENTS The general scheme established in 1981 was amended in October 2008 by Law No. 18,399. The leading changes were as follows: ●

●

● ●

A flexibilization of the eligibility conditions in force. Workers are now allowed to hold another job while receiving the unemployment benefit for the lost or suspended job. A decreasing benefit is established with the aim of encouraging beneficiaries to look for a new job before the insurance coverage period expires. On average, there is no change in the previous replacement rate, which was 50 per cent of the nominal wage earned during the six months before the occurrence of the event that gave rise to the need for unemployment insurance (or 60 per cent for workers with dependants).4 A cap amount for the benefit is established.5 The maximum coverage period is cut from six to four months and 48 workdays in the case of the subsidy granted for suspension of activities, and in the case of the subsidy granted for dismissal the maximum coverage period is maintained at six months for monthly workers and 73 workdays for day labourers.

was not a permanent increase in employment in these sectors through the subsidies for partial or total suspension of activities, especially as these subsidies were granted for only a short number of months. It was under the general unemployment insurance scheme that worker-retention and work-sharing mechanisms were implemented as a response to the cyclical contraction of labour demand by companies. The characteristics of the general scheme are outlined below. 3.1

Legal and Institutional Aspects

The general unemployment insurance scheme currently in force was created in 1981 (Degree-Law No. 15,180) and subsequently amended to expand its coverage (Box 6.1).6 It covers all private sector salaried workers, except those employed in the financial system. The subsidy is managed by the Social Security Agency (Banco de Previsión Social – BPS).

M3125 - MESSENGER TEXT.indd 156

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

157

In order to be eligible for this benefit, workers must have been employed for a set minimum period by the company applying for the subsidy; they must have made contributions for at least six months or 150 workdays in the 12 months preceding the application for unemployment insurance; they cannot have been fired or suspended for misconduct or quit voluntarily; and they must not have used the benefit recently (if they received unemployment benefits before, at least 12 months must have elapsed in order to be eligible for the benefit again). The grounds stipulated for receiving the unemployment benefit under the general scheme are dismissal, (full) suspension of activities, and a reduction in the number of workdays or work hours. In the case of reduction of work hours or workdays, the subsidy is granted when the days worked in a month or the hours worked in a day are reduced by 25 per cent or more of the legal or usual days/hours worked in normal seasons, except in those cases in which the work reduction is the result of an express agreement; or if it is characteristic of the profession or job; or in the case of monthly employees. Thus, this unemployment subsidy for work-time reduction is designed exclusively for non-monthly workers. In addition, the Executive is authorized to establish a total or partial unemployment insurance scheme for highly specialized workers for a term no greater than one year, in certain work categories or economic activities; and to extend the duration of the benefit for identical reasons and term, provided that the temporary nature of the lack or reduction of work is documented and that the company’s commitment to preserve the jobs is secured. 3.2

Financing

Contributors to the social security system are entitled to receive unemployment benefits provided that they meet certain eligibility criteria. Unemployment insurance is financed together with all the other benefits of the social security system (benefits paid to workers, retired persons and pensioners). Access to unemployment benefits is not associated with a specific contribution. It is financed with the social security contributions of employers and workers, and any social security system deficit is covered by tax and general revenue allocations.7 An implicit contribution rate can be calculated for the insurance (which would cover the benefits effectively paid). Velásquez (2005) estimated it at 3.48 per cent of the payroll in 2002. However, besides the necessary assumptions involved in calculating the contribution rate, this calculation is relative, as only part of social security spending is financed with contributions from beneficiaries. Currently only half of the BPS’s social

M3125 - MESSENGER TEXT.indd 157

14/05/2013 13:33

Beneficiaries (average)

Source:

2008

2006

2004

2002

2000

1998

1996

1994

1992

180 160 140 120 100 80 60 40

1990

Work sharing during the Great Recession

1995 Index = 100

158

Expenditure

Prepared by author with data from BPS Statistical Yearbook.

Figure 6.3

UIBs and expenditure

security spending is financed by beneficiary contributions; the other half is financed by taxes and general revenue allocations. Even considering the weight of non-contributory benefit expenditure (family allowances, old-age pensions), a substantial part of contributory benefit expenditure, including unemployment insurance, is financed through taxes and general revenue (Table 6A2.3). 3.3

Expenditure

The weight of unemployment insurance expenditures is relatively small, when considered with respect both to GDP and to all social security expenditure (Table 6A2.3). From 1990 to 2009, unemployment insurance spending ranged from 0.2 to 0.3 per cent of GDP to a maximum of 0.49 per cent when the unemployment rate peaked at 20 per cent during the 2002 crisis. The weight of unemployment insurance expenditures in overall BPS spending is also not significant. From 1993 to 2009, it only exceeded 3 per cent of BPS expenditures during the periods of rising unemployment: the 2000–02 recession (when it reached 4 per cent) and 2009. As is to be expected, both the number of unemployment insurance beneficiaries (UIBs) and the corresponding expenditures present a counter-cyclical behaviour: they rise during recessions, when the number of dismissed and temporarily laid-off workers in the economy increases (Table 6A2.4). Expenditure is determined primarily by the number of beneficiaries, as evidenced by the synchronized evolution of both variables in the 1990–2009 period (Figure 6.3). However, real-wage alterations have had no effect on expenditures, mainly because unemployment benefits have a ceiling, as indicated above in Subsection 3.1. This cap, which is

M3125 - MESSENGER TEXT.indd 158

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010 UIBs/unemployed (%)

159

UIBs/laid-off workers (%)

20%

0%

Source:

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

10%

Prepared by author with data from BPS Statistical Yearbook.

Figure 6.4

Percentage of UIBs in total unemployed (%)

determined administratively (associated with the BPC),8 has been adjusted to control social security spending, when even considering the lower incidence of unemployment insurance. 3.4

Coverage and Evolution of the Number of Unemployment Insurance Beneficiaries (UIBs)

The scope of unemployment insurance coverage is what indicates the extent to which the unemployment and work-sharing subsidy programmes created under the general unemployment insurance scheme can been used. The unemployment insurance scheme provides coverage for formal salaried workers in the private sector (with the exception of financial system employees), who in 2008 accounted for a third of total employment, before the incorporation of domestic workers. Workers not covered by unemployment insurance include private sector salaried workers who do not contribute to social security, independent workers, and public sector employees. The last two categories combined accounted for 43.9 per cent of the employed population in 2009 (BPS 2011). Overall, one-third of all employed persons do not contribute to social security. According to Amarante and Buchelli (2008) these three elements explain why 57 per cent of all employed workers were not covered against the risk of unemployment in 2005 (also see Table 6A2.5). The number of UIBs with respect to the number of total unemployed workers represented an average of 13.6 per cent in the 1990–2010 period, peaking at 22.5 per cent in 2010 (Figure 6.4; also see Table 6A2.6). In that year, UIBs accounted for 26.3 per cent of all laid-off workers, almost doubling the average for 1990–2010 (14.6 per cent) (Table 6A2.4). It should

M3125 - MESSENGER TEXT.indd 159

14/05/2013 13:33

160

Work sharing during the Great Recession 40,000 35,000 30,000 25,000 20,000 15,000 10,000 5,000 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

0

Source: BPS Statistical Yearbook.

Figure 6.5

Unemployment insurance beneficiaries

be noted that the information on beneficiaries provided by the BPS refers to benefits. In line with Amarante and Buchelli (ibid.), the main reasons for the subsidy’s relatively low coverage of the unemployed population are found to be: (i) the weight of unemployed workers who are not private sector salaried workers (primarily independent workers); (ii) the number of workers entering the labour market for the first time or after a long period of unemployment (in which they were inactive); (iii) long-term unemployment; and (iv) the significant proportion of private sector salaried workers who do not contribute to social security. The second reason accounted for 48 per cent of all unemployed persons in 2005 (ibid.), and this factor explains the lack of coverage of almost half of the unemployed population in the recent past. The fourth reason explains the lack of coverage of another 24.8 per cent of the unemployed population in 2005. With regard to the economic cycle’s effect on unemployment insurance coverage, it might be expected that during a downturn in the economic cycle (for example, during the 1999–2002 recession), when unemployment is swelled by laid-off workers, the proportion of total unemployed workers who are UI benefits will tend to rise. However, that was not the case, as can be observed in Figure 6.4. Instead, a steady growth in that percentage can be observed starting in 2004, coinciding with a period of strong economic expansion. During recessions, the number of unemployed workers on unemployment insurance increased considerably (Figure 6.5). However, the degree of coverage (the percentage of unemployed workers covered by the subsidy) was boosted by the strong economic expansion experienced from 2004 to 2009: the growth in employment was accompanied by an upward trend in the number of contributors to social security.

M3125 - MESSENGER TEXT.indd 160

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

161

1,200,000 1,000,000 800,000 600,000

2008

2006

2004

2002

2000

1998

1996

1994

1992

400,000

Source: BPS Statistical Yearbook.

Figure 6.6

Social security contributors (dependent workers)

Moreover, the increase in economic activity seems to have pushed up social security coverage percentages among private sector salaried workers over the 2004–09 period. The number of dependent workers who contribute to social security grew at an average rate of 8 per cent, with a cumulative growth of 48 per cent for that period (Figure 6.6). Another factor associated with the cycle’s growth phase, which has enabled the increase in the proportion of beneficiaries covered by the subsidy in relation to the total unemployed population, has been a strong reduction in the duration of unemployment, as a result of the strong job creation trend in the 2004–10 period (Table 6A2.7). Thus, in short, the incidence of two of the factors that determine the degree of coverage of unemployed workers (the existence of workers not registered in social security, and long-term unemployment) has been significantly reduced. This has increased the percentage of workers who are covered by social security benefits, including unemployment insurance. 3.5

Characteristics of Workers on Unemployment Insurance

The main characteristics of workers on unemployment insurance are outlined below, based on Continuous Household Survey (CHS) data.9 Unemployment insurance coverage is highest among male (two-thirds of the total), middle-aged (only one-third is 29 years or younger) and highly educated workers. Just 20 per cent of the total has completed primary education only, while almost the same percentage (16 per cent) has higher education (Tables 6A2.8 and 6A2.9). With regard to the areas of activities to which subsidy beneficiaries are connected, in 2010 almost half of the beneficiaries were concentrated in

M3125 - MESSENGER TEXT.indd 161

14/05/2013 13:33

162

Work sharing during the Great Recession

two sectors: manufacturing and construction, with each contributing close to a quarter. The services sector accounts for 17.8 per cent of workers on unemployment insurance, with a contribution that is significantly lower than its weight in total employment. This fact may be linked to the high incidence of workers without social security coverage and of independent workers in the sector (Table 6A2.8). 3.6

Grounds for Receiving Benefits

The grounds stipulated for receiving unemployment benefits under the general scheme are dismissal, suspension of work activities, and a reduction in the number of workdays or work hours. Suspension has traditionally been used as a mechanism to address drops in business activity and avoid firing workers. It operates as a worker retention subsidy. It is often implemented in combination with the practice of ‘rotating workers on unemployment insurance’ and with the consent of company and sector unions, and therefore in those cases it also operates as a work-sharing scheme.10 This strategy basically consists in the company establishing (more or less explicitly) a timetable for putting workers who are on the payroll on unemployment insurance for a predefined number of months, after which they are reincorporated into the company’s activities, as other workers go on unemployment insurance for the same number of months. The reduction option is a subsidy for work-time reduction, but it is designed exclusively for non-monthly workers. As it is associated with the low participation of non-monthly workers in total employment, the reduction option represents less than one in every 10 workers covered by unemployment insurance, with 2010 being the year in which it had its greatest incidence (Figure 6.7). From the early 1990s to 2010, the proportion of unemployment insurance granted for suspension of activities varied sharply, from a high of 63 per cent to a low of 23 per cent, with the lowest percentage registered in 2010 (Figure 6.7 and Table 6A2.10). At the same time, from 2004 there was an increasing incidence of unemployment insurance granted for dismissal. The steady increase in the number of UIBs under the dismissal option since 2004, in a context of strong economic growth and falling unemployment rates (from 16.9 per cent in 2003 to 7.1 per cent in 2010) is most likely explained by: (i) the higher number of workers contributing to social security during this period; (ii) the dynamism of the economy and the labour market, which entailed a strong growth in job creation and elimination of worker mobility between jobs; and (iii) the extension of the benefit for additional periods of up to one year by the Executive and for longer terms by parliament. Even as the

M3125 - MESSENGER TEXT.indd 162

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

163

8,000 7,000 6,000 5,000 4,000 3,000 2,000

Reduction

Source:

Suspension

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

0

1993

1,000

Dismissal

Prepared by author from BPS Statistical Yearbook.

Figure 6.7

Total number of new UIBs, by grounds (annual monthly average)

average duration of unemployment drops, this discretional component presents a certain downward rigidity. 3.7

Evolution of the UIBs and the Work-suspension Subsidy

The relative weight of the different grounds for unemployment insurance tends to vary from one phase to another of the economic cycle. In particular, companies tended to use the programme that operated as a workerretention subsidy (suspension option) more intensely during economic downturns. The inclusion of work suspension as grounds for accessing unemployment benefits was based on the need to address the risk of unemployment that emerges from an unexpected and temporary lull in a company’s activities (ECLAC, 2010). It is used more intensively in certain sectors, in particular in the manufacturing industry. In 2009 the manufacturing industry had 36.2 per cent of its workers on unemployment insurance under the suspension option, and in 2010 that percentage had gone up to 43.1 per cent (Table 6A2.10). In sum, in the case of subsidies granted for dismissals the predominant logic is to use unemployment insurance as a means of providing coverage against the risk of unemployment faced by workers. However, in the case of subsidies granted for suspension the predominant logic is to subsidize worker retention by preserving the employment relationship between the worker and the company, as noted above, without incurring any labour

M3125 - MESSENGER TEXT.indd 163

14/05/2013 13:33

164

Work sharing during the Great Recession

costs for the company (for a period of up to four months in one year). Also, use of the subsidy does not entail any costs, as the contribution to social security is fixed and independent of how intensively unemployment insurance is used. This logic of subsidizing worker retention coincides with a strong increase in unemployment subsidies under the suspension option during the periods of economic contraction: in 1995 and 1999–2002 these workers represented 57.9 and 62.8 per cent, respectively, of all workers on unemployment insurance. In 2009, while aggregate employment did not suffer a contraction, some sectors were affected and the percentage of unemployment beneficiaries under the suspension option totalled 38.7 per cent of all UIBs (Table 6A2.10). In addition to the positive relationship between subsidies granted under the suspension option, on the one hand, and the unemployment rate, on the other, the evolution of these suspended workers tends to somewhat precede the evolution of the unemployment rate. This phenomenon would appear to indicate that, at the start of a downward economic phase, companies tend first to put workers on unemployment insurance under the suspension option, and only later resort to dismissal. It should be noted that the evolution of the geographical distribution of applications for unemployment insurance demonstrates a substantial increase as of 2008 in the number and proportion of subsidies outside the capital city of Montevideo (in the rest of the country). Applications for subsidies in the rest of the country vastly exceed those from the capital (in 2009 they accounted for 56.2 per cent of all unemployed workers receiving unemployment insurance) (Table 6A2.4). This evolution of the unemployment insurance subsidy in the country’s interior would indicate that the impact of the global crisis was greater outside of Montevideo, as employment is more closely associated with the prospects of exportoriented agricultural production. Additionally, the increase in the number of workers contributing to social security in recent years was very strong in the interior of the country.

4.

URUGUAY’S NEW WORK-SHARING SCHEME: PARTIAL UNEMPLOYMENT INSURANCE

4.1

Programme Design: Policies on Work Sharing

In response to the impact of the global economic crisis that began in 2008, even though throughout 2009 its effects in Uruguay appeared to be moderate, the government adopted measures to support the economic sectors

M3125 - MESSENGER TEXT.indd 164

14/05/2013 13:33

Physical volume index 2006 = 100

Partial unemployment insurance programmes in Uruguay, 2009–2010

165

120 100 80 60

Tanneries

Jan-11

Jan-10

Jan-09

Jan-08

Jan-07

Jan-06

Jan-05

Jan-04

Jan-03

Jan-02

40

Textiles

Source: INE.

Figure 6.8

Evolution of textile and tanning industries (physical volume of production)

most affected by the crisis and to mitigate the negative impact on employment. The manufacturing industry was the sector worst hit by the world crisis; industrial output fell by 3.7 per cent in 2009, and some segments registered very steep drops. By Decree No. 2009/00510 of 7 July 2009, the Executive created a special new work-sharing scheme for a term of one year using the powers granted under the general unemployment insurance scheme for creating a special partial unemployment subsidy programme.11 It covered workers from the leather, textile and garments, wood and wood products, and metalworking industries. As noted above in the analysis of the country’s economic conditions in 2009, even though the economy as a whole grew on average during that year, as of mid-2008 and throughout 2009 several segments of the manufacturing industry saw their levels of activity seriously affected (Figure  6.8). For that reason it was decided that the worker-retention subsidy would target those sectors. As the information on the level of activity of those sectors was provided by national statistics, there was a relative consensus that their situation was critical. The legal framework on special unemployment (for highly specialized workers, in certain economic activities or categories) included in the general unemployment scheme was one of the aspects that changed in 2008 (Law No. 18,399). The degree of Executive discretion in the implementation and extension of special unemployment subsidy schemes was reduced with the addition of new conditions, namely that the benefit could only be granted ‘for a term no greater than one year’ and by requiring that

M3125 - MESSENGER TEXT.indd 165

14/05/2013 13:33

166

Work sharing during the Great Recession

in the event of extensions ‘the temporary nature of the lack or reduction of work must be documented and the commitment to preserve the jobs must be secured’. These changes determined (and restricted) the conditions designed for the Partial Unemployment Insurance (PUI) scheme, which was created under this special regime: a limited term and the temporary nature of the situation and the programme. Thus, this instrument was conceived as a short-term job-retention policy. This design was, moreover, consistent with the moderate effect that the global crisis was having on the economy as a whole in the year 2009. However, Uruguay has a longstanding tradition in the use of so-called ‘extensions’ in unemployment insurance, managed discretionally by the executive branch of government for terms of up to one year (and by parliament for longer terms), under the discretional powers granted by the general unemployment insurance scheme. These discretionary powers enable the extension of measures originally designed as transitory. It should be noted that these requirements apply to the PUI scheme but not to any of the other unemployment insurance options (work suspension and reduction for non-monthly workers) that operate as work-sharing or worker-retention subsidies. The programme’s design sought to simultaneously achieve two main goals within a general job-preservation strategy. On the one hand, through the granting of benefits under the PUI mechanism, it aimed to minimize the loss of income suffered by workers due to time not worked. This goal is similar to that of the ‘traditional’ unemployment insurance scheme. On the other hand, it aimed to provide the opportunity for workers to spend the time not worked (due to decreased economic activity) in training. This second factor addressed several worker and business concerns. From the perspective of workers, it sought to develop their skills towards enhancing their employability and productivity. From the perspective of businesses, by enabling workers to remain in the company, it increased the probability of maintaining the employment relationship and reduced the costs associated with labour turnover (which is greater when workers are placed on unemployment insurance, as some of those workers look for and find new jobs). According to Casanova (2009), the programme further offered additional advantages for job retention including: it did not put greater pressure on national finances; as social security is covered by general revenue and taxes, the advantage of this system is that the sector coverage was limited and the fact that the training components were financed from resources from the Job Retraining Fund (Fondo de Reconversión Laboral – FRL); the programme has a counter-cyclical effect as it stimulated worker retention in a situation of economic contraction to prevent a rise

M3125 - MESSENGER TEXT.indd 166

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

167

in unemployment; it was also an example of an innovative combination of passive policies (unemployment insurance) and active policies (training) working in concert; it contributed to social dialogue, as collective bargaining and employer–worker agreement over the presentation of applications is essential to accessing benefits as well as distributing working hours among all workers in a company. In addition to these aspects, the training component came with a grant or economic stipend for the trainee. This stipend was financed by the FRL, which had significant resources at the time of the programme’s development. Thus, an additional source of financing was secured for the programme through the incorporation of the training component, and this entailed a considerable reduction in the loss of income suffered by workers as a result of the reduction in workdays. The National Employment and Vocational Training Institute (Instituto Nacional de Empleo y Formación Profesional – INEFOP), which manages the FRL, was entrusted with implementing training courses, provided that such courses were stipulated under the collective bargaining agreement indicated under item (f) in Box 6.2. 4.2

Main Results of the Implementation of the Partial Unemployment Insurance Programme

The most notable result of the 2009 implementation of the PUI worksharing scheme was its minimal usage by companies. In the year that the PUI was in force, it was used for only three months (November 2009 and February and March 2010). Following its adoption in July 2009, only 730 workers were placed on PUI (under the reduction option established by decree) from July 2009 to June 2010, representing 0.3 per cent of all workers on unemployment insurance during that period, and 1.3 per cent of workers on unemployment insurance for suspension and reduction (excluding dismissals) (Table 6A2.11). In November 2009, the month with the largest number of new beneficiaries, there were four companies with workers on unemployment insurance under the PUI programme: a tannery, two textile companies, and one metalworking company. In the following months, another company from the textile sector applied for coverage under this programme, bringing the total number of participating companies to five. As noted above, the low incidence of the PUI scheme can be explained. In the first place, most companies opted for the suspension subsidy, a competitive work-sharing mechanism that was already available under the general unemployment scheme (detailed below). Other determining factors may have been the low contribution to employment of the sectors included in the design of the PUI scheme; the high proportion of small

M3125 - MESSENGER TEXT.indd 167

14/05/2013 13:33

168

Work sharing during the Great Recession

BOX 6.2 PARTIAL UNEMPLOYMENT SCHEME IN URUGUAY The conditions that had to be met to qualify for this benefit, as established in the legal decree, were: a. A 15% or greater drop in company activities in the quarter prior to application for coverage under the special regime, as compared to the average level of activity for the same quarter in the two previous years.12 b. A reduction of at least one and no more than two workdays per workweek for each worker involved. c. The company cannot have dismissed more than 5% of its personnel in the quarter prior to the application for coverage under the special regime (not including dismissals for gross misconduct). d. If the company has workers covered by unemployment insurance under the suspension option, it must reinstate them before applying for the subsidy. e. While the benefit is in effect, the company cannot suspend or dismiss workers, except in cases of disciplinary suspensions or in the event of gross misconduct. f. Adoption of a collective bargaining agreement, covering all workers and for a term extending through the duration of the special regime subsidy, including the provisions set forth in items (b) and (e) above, as work-sharing mechanisms, and, where applicable, the workers’ obligation to attend training courses or return to formal education. companies in the economy; the conditions companies were required to meet to participate; relatively weak incentives for company participation in work-sharing programmes; a certain time lag in its implementation; and the country’s swift economic recovery. 4.2.1

Sector focus: Low participation in total employment of the sectors eligible for PUI The PUI scheme was designed to target certain industrial sectors, and thus only companies from those sectors were eligible for coverage under this type of unemployment insurance. Salaried workers from these sectors (leather, textile and garments, wood and wood products, and

M3125 - MESSENGER TEXT.indd 168

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

169

metalworking) had a limited participation in total employment (accounting for 4.9 per cent of salaried workers in 2009, or an estimated 42,000 workers) (Table 6A2.12). Consequently, by focusing coverage on these sectors only, the potential impact of the PUI scheme on the economy as a whole was quite limited. At the time of the programme’s introduction there was significant concern over the fiscal cost of a subsidy to companies affected by the global crisis. The focus on the sectors that suffered significant downturns could probably be explained by the fear that the global crisis would have a severe and persistent impact on the economy and that considerable revenue would have to be spent in mitigating that impact. However, the sectoral focus would appear not to have been an important factor as in the sectors eligible for PUI the percentage of suspended workers on PUI is very small (Table 6A2.13). Most notably, for selected months in which new PUI beneficiaries were incorporated (November 2009 and March 2010), workers receiving partial unemployment benefits in the sectors eligible for this type of unemployment insurance represented 29 to 32 per cent of those sectors’ workers on unemployment insurance under the suspension option. This means that the PUI programme had a ceiling of one-third of all suspended workers on unemployment insurance. However, in spite of this evidence, this was not one of the factors given as an explanation for the programme’s results in interviews with qualified company representatives and union leaders. 4.2.2 Incidence of company size in the use of the PUI programme The companies with a greater number of workers on PUI were also those that used this resource more intensively. The proportion of UIBs who received partial unemployment benefits increased as the number of workers on unemployment insurance in general went up.13 This phenomenon may be due to the fact that the partial subsidy mechanism was used more intensively by the larger companies, which had a greater number of workers on the payroll. Moreover, job creation during this period was concentrated in mediumsized and large enterprises. While jobs in general grew by 26 per cent, large companies saw an increase of 37 per cent, while for micro and small enterprises it was 15 per cent. As for formal employment (those with social security coverage), close to 90 per cent of Uruguay’s companies (micro and small enterprises) generate a mere 15 per cent of all jobs, even though between 2005 and 2008, micro and small enterprises (employing up to 19 workers) accounted for around 95 per cent of all companies in Uruguay.14 The costs of managing a new system, the numerous eligibility requirements for participating in the PUI scheme, and the lack of mechanisms

M3125 - MESSENGER TEXT.indd 169

14/05/2013 13:33

170

Work sharing during the Great Recession

providing flexibility for work-time distribution combined to make the work-sharing system implemented through the PUI scheme unattractive for 95 per cent of Uruguay’s companies. Although in terms of the employment represented by micro and small enterprises, the percentage of workers who could not access the benefit dropped to 15 per cent of the workers who contributed to social security, these companies did not use it. This would indicate that company size was another limiting factor in the implementation of this work-sharing system, under these particular conditions. 4.2.3 Eligibility conditions for the work-sharing programme The set of conditions that must be met in order to be eligible to receive the subsidy for reduced workdays under the programme was probably a determining factor in the scarce use of this mechanism by companies. First, in addition to the two-day maximum and one-day minimum, a condition for companies to receive the subsidy was that the company could not have dismissed more than 5 per cent of its workers in the quarter prior to application and could not suspend or dismiss workers during the term of the benefit. Consequently, for reductions in labour demand that exceeded the maximum (5 per cent) plus the compensation for reduced hours through work sharing (2.0 per cent), the work-sharing programme was not attractive for reductions above 7.1 per cent, as opting for the subsidy entailed waiving the possibility of reducing labour demand by more than 7.1 per cent. Second, a source of transaction costs involved in participating in the programme may arise from the need to enter into a collective bargaining agreement with workers to establish the work-sharing mechanism and possibly the participation of workers in training courses. However, this factor was not mentioned as an obstacle by sources interviewed in companies or union leaders from the sectors involved. Third, another condition originally included in the work-sharing scheme was the reinstatement of workers on unemployment under the suspension option. This is a source of significant rigidity, as suspension was the option that had normally been used to address a reduction in economic activities and avoid having to dismiss workers. This new work-sharing scheme, then, competed directly with a mechanism that had been traditionally in use and was accepted by both workers and employers. This restriction, however, was lifted one month after it was created.15 4.2.4

Weak incentives for companies to participate in work-sharing programmes Work-hour reduction has been proposed primarily by governments and/or unions as a response to growing unemployment. Incentives were

M3125 - MESSENGER TEXT.indd 170

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

171

necessary to persuade companies to adopt work-sharing schemes when such schemes are not mandatory, as is the case, for example, with the legal reduction in the normal number of work hours, above which payment of compensation is required. The weight of the wage bill16 in gross value added (GVA) by industry (Amarante and Brum 2010) was high only in the leather sector (92 per cent of GVA). For other industries included under the work-sharing scheme the weight of the wage bill on GVA ranged from 26 to 33 per cent, which is approximately the same as the weight of the wage bill in overall GVA (33 per cent). Consequently, for most companies the economic incentive for participating in the PUI scheme was relatively weak. Generally, there were no flexibility mechanisms in the economy for distributing work time, either over the workweek or throughout the year. The reduced work-hour scheme implemented in the country did not include any kind of workday flexibility mechanisms for employers that would bring down the costs of participating in the work-sharing programme. Instead, a rigid minimum of one and maximum of two workdays was established for the reduction of workdays per week. For example, employers were not given the option of changing the number of daily work hours depending on the day, or of having weeks with a different number of workdays. Furthermore, the scheme did not give employers the flexibility to implement the reduction of work during the week in ways other than by reducing the number of workdays per week, or of reducing the number of work hours evenly over every day of the week. Employers were also not given the option of choosing whether they preferred to reduce the number of workdays or the number of work hours in a workday, a preference that could vary depending on how intensively labour was used (the more intensively it is used, the more employers prefer to reduce the number of hours) and how the production process was organized. 4.2.5 A certain lag in implementation coupled with swift economic recovery As evidenced by the examination of sector activity data and according to the opinion of government stakeholders who participated in the programme’s implementation, the negative economic conditions that affected companies at the time this work-sharing subsidy option was effectively made available (in the middle of 2009) have for the most part been overcome. Connected in part with this improvement in economic conditions, the scarce use given by beneficiary sectors to the PUI programme may be explained by a delay in its implementation with respect to the moment in which the sectors’ companies were worst hit by the global crisis and would have needed the subsidy most (Figure 6.9).

M3125 - MESSENGER TEXT.indd 171

14/05/2013 13:33

Reduction

Suspension

Nov-09

Sep-09

Jul-09

May-09

Mar-09

Jan-09

Nov-08

Sep-08

Jul-08

Jan-08

9,000 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0

May-08

Work sharing during the Great Recession

Mar-08

172

Dismissal

Source: BPS Statistical Yearbook.

Figure 6.9

Evolution of the number of new subsidies for dismissal, suspension and reduction

Physical volume index 2006 = 100

200 150 100 50

Wood

Jan-11

Jan-10

Jan-09

Jan-08

Jan-07

Jan-06

Jan-05

Jan-04

Jan-03

Jan-02

0

Automobile

Source: INE.

Figure 6.10

Evolution of automobile and wood industries (physical volume of production)

The evolution of the physical volume of production index of the targeted sectors reveals that as the effects of the global crisis in Uruguay neither persisted nor were intensified, when the work-sharing scheme was created (July 2009) most sectors were already on the path to recovery (or were consolidating their recovery). This evolution is evident in the wood and automobile sectors, but the trend can also be observed in all targeted sectors and in the manufacturing industry as a whole (Figure 6.10 and Table 6A2.14).

M3125 - MESSENGER TEXT.indd 172

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

173

4.2.6 Views from social partners17 Interviews were conducted with social partners to assess the results of the PUI programme and the use of the suspension grounds or option, as a response to the impact of the recent global economic crisis on Uruguay. Programme managers, entrepreneurs and labour leaders from the PUI beneficiary sectors were interviewed. With respect to the causes that would explain why companies did not make use of this programme more intensively, the following reasons were given consistently by respondents (Box 6.3). First, the negative economic conditions that affected companies when this unemployment insurance option was introduced (mid-2009) were swiftly overcome. Second, there was a delay in the creation of the PUI with respect to the moment when companies suffered the worst effects of the global crisis and thus required support. Third, the work-sharing mechanism stipulated in the programme required that companies reorganize their productive processes and the costs entailed were ultimately greater than the benefits obtained. Fourth, companies opted for unemployment insurance on suspension grounds,

BOX 6.3 SOCIAL ACTOR VIEWS OF THE PARTIAL UNEMPLOYMENT PROGRAMME* The social partners interviewed for this chapter in general considered the following factors as not being very decisive in the PUI’s limited success: 1. the fact that applying for PUI entailed meeting very demanding eligibility requirements in terms of demonstrating the drop in activity level or presentation of the agreement between workers and employers; 2. problems in programme management (with the exception of its insufficient dissemination); and 3. resistance from workers or unions to accepting this new form of unemployment insurance. In this sense there was one exception: the union of the Paycueros company (leather industry) refused to participate in the PUI scheme, as it entailed a certain wage reduction for a group of the company’s workers. Note: * Based on views of interviewed individuals in Appendix 6A1.

M3125 - MESSENGER TEXT.indd 173

14/05/2013 13:33

174

Work sharing during the Great Recession

as this mechanism was economically beneficial for the company and was widely known and used in the past. Finally, there was a certain unawareness of the benefits of the new scheme; the communication campaigns conducted by the programme’s managing committee to disseminate this unemployment insurance option were not massive in scope. All those interviewed considered that some form of employment subsidy was desirable, in order to preserve jobs when companies face a downturn, even if it meant that some companies or sectors would resort to it repeatedly. The fact that this unemployment insurance entails a transfer from sectors with greater job stability to sectors with less job stability was not viewed negatively, and employment stability measures were generally viewed as positive actions. From the perspective of businesses, it was stressed that the larger companies employing more workers were the companies that tended to use unemployment subsidies most, and particularly the partial subsidy mechanism. Another factor noted by company representatives was how much they valued the freedom that the unemployment insurance on grounds of suspension allowed them in terms of managing their human resources. This partially explained their lack of interest in participating in the new work-sharing programme, in view of options available.

5.

UNEMPLOYMENT INSURANCE ON GROUNDS OF SUSPENSION AS A SUBSIDY FOR WORKER RETENTION

The intensity with which companies resorted to unemployment insurance on grounds of suspension is evidence that it was used by them as a subsidy for worker retention (accounting for 18.6 per cent of workers on unemployment insurance in the July 2009–June 2010 period) (Table 6A2.11). Moreover, as indicated in the previous section, it was often implemented in combination with the practice of rotating workers on unemployment insurance, and with the consent of company and sector unions, to address downturns and avoid having to dismiss workers. Consequently, it operated as a work-sharing mechanism in which the distribution of work was not done by reducing the number of workdays but rather by reducing the working months of different workers by a certain number of months. Therefore, by incorporating the suspension option, unemployment insurance played a double role: protecting workers against the risk of unemployment, and subsidizing worker retention, as it enabled employers to reduce the number of dismissals when the company faced a drop in its level of activity.

M3125 - MESSENGER TEXT.indd 174

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

175

6,000 5,000 4,000 3,000 2,000 1,000 Jun‐09 Jul/Aug‐09 Sep‐09 Oct‐09 Nov‐09 Dec‐09 Jan‐10 Feb‐10 Mar‐10 Apr‐10 May‐10 Jun‐10 Jul‐10 Aug‐10 Sep‐10 Oct‐10 Nov‐10 Dec‐10 Jan‐11

0

Men Source:

MTSS based on BPS data.

Figure 6.11 5.1

Women

UIBs on suspension grounds, by gender

Characteristics of UIBs on Suspension Grounds

The beneficiaries of unemployment insurance on grounds of suspension in the period studied for which information is available (June 2009–January 2011) were for the most part men (65.8 per cent on average). However, for women the proportion of insurance benefits that corresponded to unemployment insurance on suspension grounds was greater than for men (Figure 6.11). When considering age, most beneficiaries were in the 25–34 age group, which was followed in importance by the over 45 age group (accounting for almost a third each) (Figure 6.12). 5.2

Areas of Activity of the Companies that Use Unemployment Insurance on Suspension Grounds

With regard to the areas of activity of the companies that used the unemployment insurance on suspension grounds, the most prominent was by far the manufacturing industry, which in 2010 accounted for 43 per cent of all suspended workers (annual monthly average), and 36 per cent in the year 2009 (Figure 6.13 and Table 6A2.13 and 6A2.15). The relative intensity of use of the suspension option in the manufacturing industry was higher than in other sectors (half of its workers on unemployment insurance correspond to the suspension option – Figure 6.14 and Table 6A2.15). In order to examine the influence of seasonality in the different sectors

M3125 - MESSENGER TEXT.indd 175

14/05/2013 13:33

176

Work sharing during the Great Recession 8,000 7,000 6,000 5,000 4,000 3,000 2,000

45 and older

35 to 44

Jan-11

Nov-10

Sep-10

Jul-10

May-10

Mar-10

Jan-10

Nov-09

Sep-09

0

Jun-09

1,000

25 to 34

18 to 24

Source: MTSS based on BPS data.

Figure 6.12

UIBs on suspension grounds, by age

Finances, business services 7%

Computers, communications 1%

Community, personal & social services 7%

Mining & quarrying 0% Agriculture, forestry, fishery 5%

Lodgings, food services 3%

Transport, storage 4%

Trade 12%

Construction 16%

Manufacturing industry 43%

Electricity, gas, water 2%

Source: MTSS based on BPS data.

Figure 6.13

M3125 - MESSENGER TEXT.indd 176

UIBs on suspension grounds, by activity (2010)

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010 Community, social, personal services 11%

Mining & quarrying 0%

177

Agriculture, forestry & fisheries 11%

Finance, business services 9% Manufacturing 20%

Computers, communications 1%

Lodgings, food services 3%

Electricity, gas, water 1%

Transport, storage 5%

Source:

Construction 24%

Trade 15%

MTSS based on BPS data.

Figure 6.14

UIBs, by activity (2010)

6,000 5,000 4,000 3,000 2,000

Trade

Source:

Construction

Dec-10

Oct-10

Aug-10

Jun-10

Apr-10

Feb-10

Dec-09

Oct-09

Aug-09

0

Jun-09

1,000

Manufacturing industry

MTSS based on BPS data.

Figure 6.15

UIBs on suspension grounds, by activity

as a factor to explain the intensity of use of unemployment insurance on suspension grounds, the evolution of suspended workers on unemployment insurance in the three sectors that most resorted to it (commerce and construction combined account for 28 per cent) is presented in Figure  6.15. The three sectors displayed a similar evolution, which

M3125 - MESSENGER TEXT.indd 177

14/05/2013 13:33

178

Work sharing during the Great Recession

indicates that movements were basically linked to the influence of the labour market and economic factors that affected all sectors, and not to sector-specific seasonal patterns. If seasonality had an effect it would reveal dissimilar behaviours in different sectors, as seasonality is specific to each sector. The companies with a greater number of workers on unemployment insurance are also those with a more intensive use of the suspension option. The contribution of suspended workers to the total number of workers on unemployment is greater when the company puts more workers on unemployment insurance in general.18 This phenomenon may be due to the fact that the suspension subsidy option is used more intensively by the larger companies, which have a greater number of workers on the payroll.

6.

COMPARATIVE ANALYSIS OF THE PUI PROGRAMME AND UNEMPLOYMENT INSURANCE ON SUSPENSION GROUNDS

6.1

Full Suspension as an Alternative to PUI

One of the main reasons for the scarce use of the PUI programme is that companies resorted to an alternative through which they obtained an economic subsidy that allowed them to temporarily interrupt a worker’s activities (and the associated labour costs) without severance of the employment relationship. Companies clearly used the reduction or suspension options of the general unemployment subsidy scheme as an alternative to firing workers, as they had been doing in the past. During the year in which the PUI programme was in force (July 2009–June 2010), workers on unemployment insurance under the suspension and reduction options combined accounted for 21 per cent of all workers on unemployment (a monthly average of over 5,000 people). Their incidence is markedly higher than that of workers on PUI (who account for 0.3 per cent of the total) (Table 6A2.11). Outlined below are the main differences between the work-sharing subsidy programmes traditionally in force (suspension and reduction options) and the PUI work-sharing scheme created in 2009 for sectors affected by the global crisis: 1.

When a company suffered a drop in sales it could apply for coverage under one of the work-sharing subsidy programmes that had been historically available (suspension and reduction options), at the company’s sole discretion. In contrast to the new PUI scheme, no

M3125 - MESSENGER TEXT.indd 178

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

2.

3.

4.

5.

6.

6.2

179

minimum drop in sales is required under these options for companies to qualify for the subsidy, nor are there any restrictions regarding the period in which such a drop in sales occurs – in certain sectors, such as the textile industry, there is a time lag between production and sales – or any obligation to document it. Under the PUI programme, the benefits for the companies were relatively limited (with regard to duration, and restricted to certain sectors and companies). This was meant to prevent an excessively intensive use of this mechanism by companies. In contrast, with the unemployment subsidy under the suspension and reduction options, companies have rather generous benefits. Companies could access the unemployment subsidy on grounds of suspension and reduction without having to undertake any commitment to preserve jobs, as was required by the PUI programme. Access to the subsidy under the suspension and reduction options does not require co-financing from the companies that use it, in contrast to the PUI programme. One of the conditions that operated to limit company interest in the PUI work-sharing programme and led companies to choose the suspension option instead – that they could not be covered by both subsidies at once – was later eliminated by decree. As the PUI programme limits dismissals, in order to apply for this work-sharing scheme companies had to abandon the possibility of adjusting their labour demand downwards. Accessing the subsidy on grounds of suspension and reduction does not limit the simultaneous use of the subsidy on grounds of dismissal. Comparison of the Benefits of the Different Worker-retention or Work-sharing Subsidy Programmes

The low participation of companies in this programme raises the question of whether companies are interested in training. The empirical evidence available indicates that most companies that invested were very interested in training.19 Moreover, in 90 per cent of the courses, the companies themselves bore the costs of the training, thus clearly demonstrating that they considered training as having a positive impact on worker productivity. However, this training may not have been as appealing during the downward phase of the economic cycle. Business interviewees did not mention training as an important incentive for applying for the work-sharing scheme. For their part, union leaders who supported training noted that the time required to arrange it was an obstacle to its implementation. By the time the training was ready to be launched, the company had regained

M3125 - MESSENGER TEXT.indd 179

14/05/2013 13:33

180

Work sharing during the Great Recession

its level of activity, which meant that it was not possible to conduct the training courses. 6.3

Comparison of the Marginal Costs of the Benefits of the Different Worker-retention or Work-sharing Subsidy Programmes

With respect to the subsidy amount, unemployment beneficiaries under the full suspension option, with fixed or variable wages, received a benefit that was equal to 50 per cent of the monthly average of the nominal wages earned by them in the six months immediately prior to the event that caused the suspension,20 and 60 per cent in the case of workers who were married or in a common-law relationship or have disabled relatives or minors (under 21 years of age) in their care. For workers with partially suspended activities or reduced work time, the amount of the subsidy would be equal to the difference between the subsidy as calculated (above) and the amount effectively earned during the period they receive the benefit. Thus, as the replacement rate was the same for total and partial suspension, the benefit had a lower marginal cost in the second case, as it was calculated as the difference between the subsidy amount and what the beneficiary was still earning for continuing on the job. The PUI programme stipulated a maximum reduction in work hours of 33 per cent, and the 50 to 60 per cent subsidy is paid for that difference. The loss of income for the worker was no more than 16.5 per cent.

7.

ESTIMATED IMPACT OF THE WORK-SHARING MEASURES (PARTIAL, REDUCTION AND SUSPENSION)

The results of the PUI work-sharing scheme implemented in Uruguay, which were obtained by gathering data at the company level, are consistent with the results yielded by the few empirical studies conducted with micro data reviewed in Section 3. The effect of the time-sharing measures on employment is close to zero. The maximum reduction in workdays established under the scheme is two days per week. Two days in a six-day week represents 33.33 per cent of the standard workweek for the industries targeted by the PUI. An elasticity of employment to reduced hours of –0.5 as yielded by the time-series data, would generate an increase of 16.7 per cent in the number of workers demanded; an elasticity of the magnitude yielded by company micro-data (for example, an elasticity of –0.6, as estimated by Hunt, 1996) would generate an increase in labour demand of 2 per cent.

M3125 - MESSENGER TEXT.indd 180

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

181

Although the work-sharing subsidy innovation was barely used, the mechanisms that were already in force were used relatively intensively to address the (moderate) effects of the crisis. The number of suspended workers (monthly average) had a cumulative increase of 24 and 57 per cent in 2008 and 2009, respectively. The average number of workers who received the subsidy under the suspension option in the June 2009– December 2010 period (more than 5,000 workers) represented 0.6 per cent of private sector salaried workers (Table 6A2.11). In sum, if the various modes of subsidy for full or partial suspension of activities that were used as work-sharing subsidies are considered together, in the period studied (June 2009–December 2010) an average of 5,800 workers participated every month in these subsidies (accounting for 0.7 per cent of all private sector salaried workers of 2009). This includes all UIBs with the exception of workers on unemployment on dismissal grounds (that is, it included workers on unemployment under the reduction, full suspension and partial options). This represents a considerable impact in relative terms as a response to the downward phase of the economic cycle during that period; they represented 30 per cent of the workers placed on unemployment insurance for dismissal during that same period (Table 6A2.11). This can be considered the minimum estimate of the number of salaried workers whose jobs were protected through the subsidy in 2009, as the estimate is based on a monthly average. The number of UIBs for full or partial suspension totalled more than 30,000 in 2009.

8.

ESTIMATED COSTS OF UNEMPLOYMENT INSURANCE FOR PARTIAL UNEMPLOYMENT AND SUSPENSION

In 2009, unemployment insurance expenditures totalled US$100 million, accounting for 0.28 per cent of GDP (BPS 2011). In 2009 there was an average of 25,872 UIBs, which translates into an estimated annual cost of US$3,868 per transfer subsidy and an estimated monthly cost of US$320 per subsidy.21 The cost of each of the grounds for unemployment insurance is basically determined by the number of benefits paid (the rationale for this has been explained). Consequently, an adequate estimate of the costs of each programme was its contribution in terms of number of beneficiaries. The cost of PUI was minimal. First, due to its low level of implementation it amounted to only 0.3 per cent of average benefits in one year (the 12month period from July 2009 to June 2010) (Table 6A2.11). Second, the

M3125 - MESSENGER TEXT.indd 181

14/05/2013 13:33

182

Work sharing during the Great Recession

subsidy is granted for a reduction in work hours (of up to 33 per cent), so that on average benefits are a third of the average monthly benefit. The total amount of PUI benefits can be estimated at around US$100,000. The cost of unemployment insurance for suspensions was greater, as it represented an average of 21 per cent of the benefits in the period for which data are available (June 2009–January 2011) and 18.6 per cent in the year the PUI programme was in force. In one year, over US$20 million was spent in unemployment subsidies for full suspensions (an equivalent of more than US$300 per monthly benefit, in line with the average benefit). In sum, unemployment insurance programmes have a relatively moderate cost in Uruguay, in terms of average monthly benefits paid, as these have a ceiling and the replacement rate is 50 per cent (or 60 per cent for workers with dependants). Moreover, the maximum number of monthly benefits per worker is four in the case of full suspension and six in the case of partial suspension.22

9.

OBSERVATIONS AND RECOMMENDATIONS BASED ON THE URUGUAY SITUATION

Some observations can now be made about the experience with worksharing policies in Uruguay. The operation of the different work-sharing or employment subsidy programmes should be integrated. In this sense unemployment insurance on grounds of suspension should be integrated with PUI. They clearly have different requirements for participation and only the PUI programme clearly establishes a work-sharing mechanism and provides for training courses. By the same token, unemployment insurance on grounds of reduction should be combined with the decree option (PUI), as their similarities entail a reduction in the number of hours worked, but the first applies to non-monthly workers whereas the second applies to workers paid by the month. Alternatively, the worksharing partial unemployment programme should be redesigned, since it competes with other forms of work-sharing subsidies. A possibility is that it operate when a sector registers two quarters with a drop in activity (compared to the average of the two previous years), or when a company suffers a significant drop in sales, or it could be applied in a similar way as the work-sharing subsidy programmes currently in force (suspension and reduction), at the company’s sole discretion. Work-sharing programmes should also be implemented to support sectors and companies affected by downturns, including mechanisms to automatically place workers on unemployment and/or providing for permanent operation of the programme. This should be done in order

M3125 - MESSENGER TEXT.indd 182

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

183

to shorten the response times of policies aimed at addressing deteriorating economic situations, as this is a key factor in job preservation, and improve access to information on the programme’s existence. The consistency of the goals of the various schemes should be considered: relatively restrictive objectives that seek to prevent an excessively intensive use by companies (such as in the case of the PUI programme) coexist with others that are extraordinarily generous (such as the goals of the other schemes currently in force), which require no commitment to preserve jobs and no co-financing by the companies that use them. Requirements of work-sharing policies should be reconsidered. The aims pursued by the work-sharing partial unemployment programme of limiting the Executive’s degree of discretion in the implementation of special unemployment subsidy schemes and the requirement that beneficiary companies undertake certain job-stability commitments in return for the subsidy are considered advisable. However, by applying the more restrictive conditions only to the new schemes that subsidize employment and not to the schemes currently in force (suspension and reduction), the participation of companies in the new schemes is drastically limited compared to the competing schemes, as is indicated by its limited use in addressing the crisis in 2009. Working time in the context of work sharing needs be reconsidered. The requirements that operated to limit entrepreneur interest in using the work-sharing PUI scheme should be revised. One of the most restrictive requirements – incompatibility with the unemployment insurance on grounds of suspension – has already been lifted. A subsequent decree allowed companies to have a percentage of workers on unemployment on grounds of suspension at the same time as it participated in the new worksharing programme. Also, the combination of ‘percentage by which dismissals are limited’ and ‘maximum reduction of work hours’ needs to be reconsidered, because in practice this entails that companies have to virtually give up the possibility of adjusting their labour demand downwards in order to participate in this work-sharing regime. Companies should be given greater flexibility in the use of work hours. Some forms of flexibilization in the distribution of work hours were, in fact, included as part of agreements reached between companies and workers to address the crisis. But including such flexibility as part of the work-sharing subsidy would encourage its use among companies. The benefits of preventing job loss through the reduction of work hours are obvious in the case of workers and the government, but not always in the case of companies. Consequently, incentives need to be included to encourage companies to adopt work-sharing mechanisms through a reduction in work hours. Flexibility can be derived from distributing hours differently

M3125 - MESSENGER TEXT.indd 183

14/05/2013 13:33

184

Work sharing during the Great Recession

for each day of the week or by changing the distribution from one week to the next, with a set number of hours for a given number of months. The way work sharing is implemented can also be made more flexible, through a reduction in the number of hours in a workday or a reduction in the number of workdays and in the way the workdays are distributed. Finally, policymakers should reconsider the of cost of work-sharing policies. Given that there is a large company bias in the intensity of use of the subsidies, mechanisms need to be introduced to reduce the information and transaction costs involved in participating in PUI-type programmes; these costs are now relatively high for medium-sized companies and they operate as a barrier to their participation. Temporary work policies in other countries may offer an example of how to represent small and medium-sized enterprises in this situation by acting as programme promoters and application reception offices.

NOTES 1.

2.

3.

4.

5.

6. 7. 8.

Unionization report prepared by Instituto Cuesta Duarte, the PIT-CNT’s research centre. Unionization data are estimated based on the number of members reported by unions prior to each conference of the confederation to determine the number of conference delegates per union. Through a major increase in public social spending (from 2004 to 2009 it grew by more than 3 percentage points of GDP) focused on households below the poverty line, the incidence of destitution dropped from 4.6 per cent in 2004 to 1.2 per cent in 2010, while the percentage of people living below the poverty line went from 40 to 18 per cent during that same period. In the 2005–10 period, Uruguay had a cumulative GDP growth of 35 per cent, equivalent to an average annual rate of 7 per cent, having grown above the average for Latin America and exceeding the country’s historical highs. In 2009, the economy grew by 2.9 per cent, in contrast to the performance of the global economy, which suffered a 0.6 per cent contraction, and to the performance of Latin American economies, which experienced a 1.8 per cent contraction, as a result of the marked downturn in the most advanced economies. The amount received by dismissed workers as a percentage of the previous compensations is as follows: 66 per cent for the first month of the benefit; 57 per cent for the second; 50 per cent for the third; 45 per cent for the fourth; 42 per cent for the fifth; and 40 per cent for the sixth. The amount of the benefit cannot exceed a maximum of eight base units for benefits and contributions (Bases de Prestaciones y Contribuciones – BPCs) per month for workers in a situation of full suspension of activities or reduced work, and in the event of dismissal, the cap amount for the first month is 11 BPCs, dropping gradually to six BPCs in the sixth month. The BPC was created in 2004 as a unit of reference for calculating social benefits. In 2001, rural workers were incorporated by Decree No. 211/01, amended in 2005. Domestic workers were included in 2007, by Law No. 18,065 of December 2006, regulated in June of the following year. Health benefits are the only services associated with a specific contribution. On the one hand, to control spending, the 8 BPC cap has not been increased. The BPC

M3125 - MESSENGER TEXT.indd 184

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

9. 10.

11. 12. 13.

14. 15.

16. 17.

18.

19.

185

can also be adjusted on a discretional basis within a range of up to 20 per cent above or below the consumer price index (CPI) or the average wage index. Since its establishment, the BPC has been adjusted in line with the CPI, so that it has not received the positive effect of the increase in real wages. It should be noted that the number of cases of unemployed workers on unemployment insurance gathered by the CHS is relatively low. The information must be examined and compared to administrative records in light of this fact. In February 2009, faced with the impact of the global crisis on the metal industry, which meant placing a considerable number of workers on unemployment insurance, the sector’s union proposed the ‘rotation of workers on unemployment insurance’ as a strategy (according to statements in the newspaper El País, Uruguay, 25 February 2009). Leather industry union leaders from companies that were unable to participate in the PUI work-sharing programme declared, for their part, that the union organized the rotation of workers on unemployment insurance (on grounds of work suspension) as a way of avoiding dismissals through work sharing. As stipulated under Article 10 of Decree-Law No. 15,180 of 20 August 1981, in the wording given by Article 1 of Law No. 18,399 of 24 October 2008. This must be backed by evidence proving that sales for the quarter prior to application were at least 15 per cent lower, in constant prices, than the simple average sales for the same quarter of the two previous years. In companies with more than 10 workers on unemployment, the percentage of workers on partial unemployment is as high as 4 per cent of all workers on unemployment, and that percentage drops to 0.02 per cent in companies with 10 or fewer workers on unemployment (data as of November 2009, when new partial unemployment beneficiaries were incorporated). Observatory of Business Dynamics (Observatorio de Dinámica Empresarial – ODE), OPP-BPS, 2009. This condition is relaxed by the Decree of 10 August 2009, most likely in response to arguments put forward by workers and employers regarding the widespread use of the suspension option as a way of avoiding the severance of the employment relationship. This decree replaces the condition originally set with the possibility of ‘allowing for up to a maximum of 20 per cent of the workers on the payroll to be suspended’. Calculated based on all types of wages, year 2008, Continuous Household Survey, INE. The people interviewed were: a union leader from a leather industry company that participated in the PUI scheme and a union leader from a non-participating company also from the leather industry; a union leader from the workers’ confederation; three public sector officials who participated in the programme’s design and management (one from the Ministry of Labour and Social Protection and two from the Planning and Budget Office); and an official from the Ministry of Industry. Several company representatives were also consulted. Most reported that they were either unaware of the existence of the PUI programme or knew little about it and how it worked, and thus contributed no elements to the analysis except the fact that the programme is largely unknown. The opinions gathered from the business sector thus basically come from only two qualified business informants. The names of the people interviewed are listed in Appendix 6A1. In companies with more than 10 workers on unemployment, the percentage of suspended workers can be as high as 25.5 per cent of all workers on unemployment, and that drops to 19.3 per cent in companies with 10 or fewer workers on unemployment (data as of November 2009, when new partial unemployment beneficiaries were incorporated). According to data from MTSS 2010. This report also gives the following elements that indicate the importance that training has for companies: (i) only 15 per cent of companies do not consider it necessary to train their workforce; (ii) of the 85 per cent that report they plan to conduct training next year they will do so intensively, with an average of 2.8 courses held per company, involving a very large number of workers; (iii) companies are willing to finance their training needs. With regard to the training

M3125 - MESSENGER TEXT.indd 185

14/05/2013 13:33

186

20.

21. 22.

Work sharing during the Great Recession modalities chosen, training courses by company staff account for 32 per cent of all training, courses conducted in the company by outside instructors account for 22 per cent, courses held in private centres account for 22 per cent, and training by equipment (or technology) suppliers accounts for 11 per cent. A small percentage (6 per cent) of training is conducted with funding from state programmes; and (iv) the areas covered by the courses are General Management (7 per cent), Administration and Information Technology (16 per cent), Marketing and Sales (6 per cent), Production (21 per cent), Accounting and Finances (19 per cent), Human Resources (21 per cent), and Unspecified (11 per cent). For workers paid by the day or by the hour, the amount will be equal to 12 daily wages per month, with the amount of each wage obtained by dividing by 150 the total nominal wages received in the six months immediately preceding the event that caused the suspension. Author calculation based on BPS (2011). That is in the general scheme. However, the Executive can grant extensions of up to 12 months, for a total of 18 months of benefits. Parliament can also pass laws authorizing the extension of benefits past the term of 18 months. In 2010, 212 resolutions were adopted by law to extend unemployment insurance benefits, covering a total of 3,931 beneficiaries. In January 2011, the number of workers on unemployment insurance on any grounds totalled 25,000, which means that these extensions had a significant incidence.

REFERENCES Amarante, V. and M. Brum (2010), ‘Empleo y valor agregado de los sectores de actividad en Uruguay’, First progress report, Agreement between OPP and the Economy Institute, Facultad de Ciencias Económicas y de Administración. Amarante, V. and M. Buchelli (2008), ‘El seguro de desempleo en Uruguay’, Cuadernos del CLAEH 96–97, Series 2, Year 31: 175–207. Banco Central del Uruguay (BCU) (Central Bank of Uruguay), Statistics 2011. Banco de Previsión Social (BPS) (Social Security Agency), Statistical Yearbooks. Casanova, F. (2009), ‘Uruguay Programme for Job Preservation by Reducing Working Hours Combined with Training’, Notes on the Crisis, ILO, Geneva. ECLAC (2010), ‘La reacción de los gobiernos de las Américas frente a la crisis internacional: una presentación sintética de las medidas de política anunciadas hasta el 31 de diciembre de 2009’, Notes, January. Hunt, J. (1996), ‘Has work sharing worked in Germany?’, NBER Working Paper Series No. 5724, National Bureau of Economic Research, Cambridge, MA, August. MTSS (2010), ‘Relevamiento de Necesidades de Calificación y Personal de las empresas’, Observatorio de Mercado de Trabajo de la DINAE, MTSS, Ministerio de Trabajo y Seguridad Social (Ministry of Labour and Social Security), Uruguay. Velásquez, M. (2005), ‘Seguro de desempleo: evaluación y lineamientos para su modernización’, in Uruguay. Empleo y protección social. De la crisis al crecimiento, Santiago: ILO Office for Chile, Paraguay and Uruguay, pp. 373–95.

M3125 - MESSENGER TEXT.indd 186

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

APPENDIX 6A1

187

INTERVIEWEES

Alvaro Eq Brunini Manager of the Office of Planning and Budget (worked on the design process and implementation of the programme). Yanina Corsini (Esq.) Ministry of Industry, Energy and Environment representative on the Technical Advisory Committee. Pablo Garcia Daniel Pagano

Official of the Tannery Workers’ Union (UOC). Union leader, Leather Industry Union.

Ivonne Pandiani Ministry of Labour and Social Security representative on the Technical Advisory Committee. Fernando Pereira (Esq.) Union leader, member of the governing committee of the workers’ confederation, PIT-CNT. Ricardo Perez Official of the tannery sector (the company did not meet the requirements to participate in the policy). Juan Manuel Rodríguez (Esq.) Office of Planning and Budget representative on the Technical Advisory Committee, created to facilitate service. Technical advisers for various businesses Anonymous, by request.

M3125 - MESSENGER TEXT.indd 187

14/05/2013 13:33

188

Work sharing during the Great Recession

APPENDIX 6A2 Table 6A2.1

Year

1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Activity, employment, and unemployment rates, total in country (%) Activity rate

Employment rate

Unemployment rate

Total

Men Women Total

Men Women Total

Men Women

55.7 57.9 57.4 57.9 57.3 57.6 57.4 56.9 58.2 59.3 58.4 57.9 60.4 59.3 59.6 60.6 59.1 58.1 58.5 58.5 60.8 62.5 62.4 63.1 62.7

73.7 75.3 74.4 74.5 73.5 73.5 72.5 72.0 73.3 74.0 72.0 71.5 73.5 72.0 71.9 72.2 70.7 69.0 70.0 69.3 72.3 73.9 73.1 73.8 72.9

68.0 70.2 69.7 69.9 68.5 68.3 67.6 67.3 68.1 68.1 64.9 65.0 67.8 65.7 64.1 63.9 61.2 59.7 62.9 62.7 66.3 69.1 69.2 70.0 69.3

7.8 6.8 6.4 6.2 6.9 7.0 6.8 6.4 7.0 8.0 9.8 9.1 7.8 8.7 10.9 11.5 13.5 13.5 10.2 9.6 8.2 6.6 5.4 5.2 5.0

40.6 43.2 43.0 43.9 43.8 44.1 44.8 44.4 45.6 46.9 47.1 46.3 49.3 48.5 49.1 50.9 49.4 48.9 48.7 49.5 50.9 52.7 53.4 53.9 53.7

50.2 52.6 52.4 53.2 52.4 52.5 52.3 52.1 52.9 53.2 51.4 51.2 54.3 52.6 51.5 51.4 49.1 48.3 50.8 51.4 54.1 56.7 57.7 58.5 58.4

35.4 37.7 37.9 39.2 39.0 39.1 39.4 39.5 40.1 40.6 40.2 39.5 42.8 41.4 40.8 40.9 38.9 38.8 40.6 41.9 43.7 46.1 47.9 48.7 48.9

9.8 9.2 8.7 8.1 8.6 8.8 9.0 8.4 9.2 10.3 12.0 11.5 10.1 11.3 13.6 15.3 17.0 16.9 13.1 12.2 10.9 9.2 7.6 7.3 6.8

12.8 12.7 11.9 10.7 11.0 11.4 11.9 11.0 12.1 13.4 14.7 14.7 13.0 14.6 17.0 19.7 21.2 20.8 16.5 15.3 14.2 12.4 10.1 9.8 9.0

Note: Up to 2005 is urban total. Source: INE.

M3125 - MESSENGER TEXT.indd 188

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

Table 6A2.2

Specific unemployment rate by gender and age group, Montevideo (%)

Year

1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Source:

189

24 years & younger

25 years & older

Total

Total

Men

Women

Total

Men

Women

9.1 8.6 9.2 8.9 9.0 8.4 9.1 10.8 12.4 11.6 10.2 11.8 13.9 15.5 17.1 16.7 12.9 11.8 10.6 8.6 7.3 7.6 6.9

24.9 23.7 26.2 25.0 24.4 23.2 25.5 25.5 28.0 27.2 25.5 28.0 31.7 36.2 40.0 39.1 32.4 30.0 29.0 24.1 21.0 21.3 20.4

20.6 19.7 23.3 21.9 20.3 19.8 21.4 22.7 25.5 22.9 22.5 23.9 27.4 31.0 34.3 35.3 29.2 26.9 24.1 20.5 17.9 18.6 17.4

30.3 29.2 29.9 29.7 29.4 27.4 30.9 29.1 31.0 33.0 29.1 33.0 36.5 42.0 46.7 43.8 36.5 34.0 34.9 28.5 24.8 24.5 23.8

5.2 4.7 5.1 4.9 5.2 4.6 4.8 6.8 8.0 7.6 6.5 7.9 9.6 10.6 12.2 12.3 9.0 8.3 7.1 5.7 4.7 5.0 4.4

3.3 3.2 3.3 3.4 3.3 2.8 2.9 4.4 6.2 5.4 4.4 5.6 6.9 6.7 9.6 9.5 6.8 6.1 5.3 3.8 3.0 3.2 2.9

7.7 6.7 7.3 6.8 7.6 6.8 7.1 9.5 10.2 10.2 8.9 10.5 12.7 14.6 15.0 15.1 11.4 10.6 8.9 7.6 6.4 6.9 5.9

INE and MTSS, based on INE.

M3125 - MESSENGER TEXT.indd 189

14/05/2013 13:33

190

M3125 - MESSENGER TEXT.indd 190

14/05/2013 13:33

5,216,072,006 7,648,595,997 10,452,985,873 13,413,011,673 16,305,671,983 18,724,617,509 20,298,680,875 20,324,880,019 19,307,597,864 18,571,410,193 18,953,784,626 21,503,153,896 26,315,576,776 32,528,236,154 36,583,516,982 33,271,229,071 38,499,904,714

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 20081 2009

1,182,722,795 2,266,673,204 3,443,160,451 5,481,095,999 7,057,193,144 7,924,772,074 9,851,453,783 11,010,285,275 11,018,768,995 12,523,626,392 11,782,588,681 10,115,945,587 8,698,130,145 7,387,196,253 4,846,386,430 7,048,367,523 6,297,525,177

General revenue contributions*

1,200,173,041 1,824,732,529 2,592,495,459 3,463,966,443 4,565,109,338 5,386,163,666 5,394,159,578 5,170,372,585 6,404,169,139 7,073,557,427 8,981,121,488 11,017,727,840 12,475,479,432 14,629,867,196 18,679,469,280 21,195,520,926 23,314,672,742

Allocated taxes**

7,598,967,842 11,740,001,730 16,488,641,783 22,358,074,115 27,927,974,465 32,035,553,249 35,544,294,236 36,505,537,879 36,730,535,998 38,168,594,012 39,717,494,795 42,636,827,323 47,489,186,353 54,545,299,602 60,109,372,692 61,515,117,520 68,112,102,633

Total resources

7,729,597,199 11,736,315,237 16,873,306,820 22,512,388,244 28,169,019,540 32,416,599,321 35,595,408,482 36,450,937,774 37,196,584,747 38,471,872,215 40,041,762,959 42,636,827,324 46,655,497,702 52,491,219,953 58,361,477,414 58,758,727,901 69,576,163,333

Total expenditure

167,845,787 262,299,703 445,596,584 526,067,722 582,438,982 651,327,106 1,005,502,470 1,105,836,415 1,341,372,151 1,643,175,742 1,063,869,508 699,982,584 712,270,403 917,009,926 1,172,986,802 1,388,435,056 2,258,520,226

Unemployment insurance – total

2.2 2.2 2.6 2.3 2.1 2.0 2.8 3.0 3.6 4.3 2.7 1.6 1.5 1.7 2.0 2.4 3.2

Unemployment insurance –% expenditure 2,382,895,836 4,091,405,733 6,035,655,910 8,945,062,442 11,622,302,482 13,310,935,740 15,245,613,361 16,180,657,860 17,422,938,134 19,597,183,819 20,763,710,169 21,133,673,427 21,173,609,577 22,017,063,449 23,525,855,710 28,243,888,449 29,612,197,919

General revenue & taxes

30.8 34.9 35.8 39.7 41.3 41.1 42.8 44.4 46.8 50.9 51.9 49.6 45.4 41.9 40.3 48.1 42.6

General revenue – % expenditure

Source:

BPS.

Notes: 1. In 2008, the National Health Fund was created by Law 18,131, so as of 01/01/08 sickness insurance payments are no longer part of own resources. * Includes COFIS (Social Security Financing Contribution Tax, Law 18,083 Art. 109), Family Allowances (Laws 18,227 and 17,139) and Net BPS Transfers. ** Includes allocated VAT, Sweepstakes Tax, COFIS (Law 17,345, Art. 22 until July 2007) and IASS (Social Security Assistance Tax, Law 18,314 as of July 2008).

Resources – misc. income allocations

Overall BPS resources and expenditure (in current UYU)

Year

Table 6A2.3

Partial unemployment insurance programmes in Uruguay, 2009–2010

Table 6A2.4

Year

Montevideo

191

Unemployment insurance beneficiaries (UIBs) (annual average) Interior

1990 n.a. n.a. 1991 n.a. n.a. 1992 n.a. n.a. 1993 n.a. n.a. 1994 n.a. n.a. 1995 n.a. n.a. 1996 n.a. n.a. 1997 n.a. n.a. 1998 n.a. n.a. 1999 n.a. n.a. 2000 n.a. n.a. 2001 n.a. n.a. 2002 n.a. n.a. 2003 8,929 7,141 2004 7,840 6,300 2005 7,679 6,716 2006 8,707 8,279 2007 8,794 9,794 2008 8,884 11,003 2009 11,067 14,805 2010 11,126 14,390 1990–2010 Average

UIB Total

Unemp. rate (urban country)

12,667 15,253 15,288 16,070 18,107 21,686 19,258 17,100 17,652 23,384 26,200 31,340 37,302 22,372 14,140 14,394 16,985 18,588 19,886 25,872 25,515 20,431

8.5 8.9 9.0 8.3 9.2 10.3 11.9 11.4 10.1 11.3 13.6 15.3 17.0 16.9 13.1 12.2 11.4 9.6 7.9 7.7 7.1 11.0

UnemUIB/ Laid-off UIB/ ployed Unemp. unemLaid(urban (%) ployed off (%) country) (urban country) 74,339 109,621 113,137 103,500 120,315 135,171 156,133 150,835 140,729 155,940 188,823 218,373 237,026 233,452 181,908 170,229 167,600 147,050 121,012 120,031 113,236 150,403

17.0 13.9 13.5 15.5 15.0 16.0 12.3 11.3 12.5 15.0 13.9 14.4 15.7 9.6 7.8 8.5 10.1 12.6 16.4 21.6 22.5 13.6

n.a. n.a. n.a. n.a. n.a. 105,133 132,569 116,305 123,368 135,262 162,910 191,345 209,244 206,581 158,236 140,543 131,775 119,204 100,959 102,252 96,935 139,539

n.a. n.a. n.a. n.a. n.a. 20.6 14.5 14.7 14.3 17.3 16.1 16.4 17.8 10.8 8.9 10.2 12.9 15.6 19.7 25.3 26.3 14.6

Source: BPS, without reassessments and calculated by author based on CHS data.

Table 6A2.5

Composition of employment (%)

Private salaried workers Public salaried workers Employers Independent workers w/o place of business Independent workers with place of business Others

2001

2004

2005

2009

54.5 16.6 3.9 8.8 14.6 1.6

52.6 17.7 3.5 9.2 15.2 1.8

54.5 16.6 3.9 8.3 15.2 1.8

56.1 14.3 4.8 3.6 19.1 2.1

Source: Based on CHS data.

M3125 - MESSENGER TEXT.indd 191

14/05/2013 13:33

192

Work sharing during the Great Recession

Table 6A2.6

BPS contributors (annual monthly average, in number of jobs)

Year

Total

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

860,934 918,128 929,885 926,809 929,995 946,880 966,533 956,622 927,742 901,046 852,878 865,943 916,147 1,005,143 1,083,841 1,166,716 1,248,623 1,283,197

Var. Dependents Var. (%) (%)

Private

683,476 732,849 745,034 740,829 744,010 756,686 775,003 773,292 751,282 734,087 691,767 705,809 749,457 831,566 904,969 983,166 1,059,457 1,090,743

492,721 543,849 555,005 546,060 554,709 572,755 592,107 584,344 562,317 544,939 503,718 522,365 564,978 644,380 716,050 787,699 861,803 884,041

6.6 1.3 –0.3 0.3 1.8 2.1 –1.0 –3.0 –2.9 –5.3 1.5 5.8 9.7 7.8 7.6 7.0 2.8

7.2 1.7 –0.6 0.4 1.7 2.4 –0.2 –2.8 –2.3 –5.8 2.0 6.2 11.0 8.8 8.6 7.8 3.0

Var. Domestic Industry (%) & Trade 10.4 2.1 –1.6 1.6 3.3 3.4 –1.3 –3.8 –3.1 –7.6 3.7 8.2 14.1 11.1 10.0 9.4 2.6

23,885 23,749 22,850 23,586 25,053 27,803 29,484 30,502 30,785 31,249 34,964 36,905 38,564 40,713 43,272 47,393 51,087 54,618

394,491 418,435 424,100 418,159 421,305 430,728 439,418 429,821 423,181 412,404 374,398 379,571 411,075 473,317 521,478 581,322 641,017 664,143

Source: BPS.

Table 6A2.7

Average duration of unemployment (weeks) and unregistered in social security (%)

Year

Duration of Unemployment*

Unregistered in Social Sec.

2004 2005 2006 2007 2008 2009 2010

16.0 14.0 12.2 10.2 9.1 9.0 8.3

– – – – 33.3 32.0 31.3

Note: *To calculate the duration of unemployment, the INE considers the spells of time in which the unemployed person had no work whatsoever, including odd jobs. Source: INE.

M3125 - MESSENGER TEXT.indd 192

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

Table 6A2.8

193

Characteristics of unemployed persons on unemployment insurance

% of unemployed on unemployment insurance Gender Men Women Educational level Up to primary Secondary – incomplete Secondary – complete Technical secondary school Technical training Teachers Higher education

2004

2009

4.2

4.2

52 48

66 34

28 31 14 14 n.a. 2 11

20 33 7 4 19 1 16

Note: 193 cases in 2009. Source: Based on CHS data.

M3125 - MESSENGER TEXT.indd 193

14/05/2013 13:33

194

Table 6A2.9

Work sharing during the Great Recession

Characteristics of unemployed workers on unemployment insurance

Gender Men Women Region Montevideo Rest of the country Age Under 20 years 20 to 29 years 30 to 39 years 40 to 49 years 50 to 59 years 60 and older Area of activity No data Trade, restaurants, hotels Agriculture, forestry and fisheries Electricity, gas and water Finances Mining and quarrying Construction Manufacturing industry Services Transport, storage and community

1994

2004

2009

2010

69.1 30.9

63.0 37.0

70.0 30.0

68.3 31.7

57.4 42.6

54.5 45.5

43.8 56.2

42.4 57.6

3.7 33.6 27.9 19.5 12.6 2.8

1.5 31.7 29.8 21.1 12 2.8

2.1 32.6 29.6 19.6 13.0 3.1

n.a. n.a. n.a. n.a. n.a. n.a.

n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.

0.2 21 7.8 0.1 9.7 0.1 15.4 19.8 18.7 7.2

0.4 14.6 11.7 0.7 0.8 0.2 23.5 22.1 15.8 6.0

0.3 13.9 9.4 0.9 0.5 0.1 22.7 24.0 17.8 6.0

Note: Annual average of new beneficiaries; annual average of beneficiaries by industry. Source: BPS.

M3125 - MESSENGER TEXT.indd 194

14/05/2013 13:33

195

M3125 - MESSENGER TEXT.indd 195

14/05/2013 13:33

2,601 2,194 3,012 2,130 1,713 1,999 3,152 3,374 3,835 4,511 1,897 1,054 1,096 1,197 1,047 1,296 2,029 1,605

* Some rounded to 100%.

1,496 1,859 2,161 2,065 2,218 2,283 2,518 2,552 2,270 2,541 1,845 1,713 2,100 2,505 2,975 3,317 3,787 4,629

56 51 26 32 23 23 18 5 2 662 344 221 307 324 365 453 278 648

4,153 4,104 5,199 4,227 3,954 4,305 5,688 5,931 6,107 7,714 4,086 2,988 3,503 4,026 4,387 5,066 5,244 6,882

Dismissal Suspen- Reduc- Total sion tion

Variation (%)

36.0 45.3 41.6 48.8 56.1 53.0 44.3 43.0 37.2 32.9 45.2 57.3 60.0 62.2 67.8 65.5 72.2 67.3

62.6 53.5 57.9 50.4 43.3 46.4 55.4 56.9 62.8 58.5 46.4 35.3 31.3 29.7 23.9 25.6 38.7 23.3

1.4 1.2 0.5 0.8 0.6 0.5 0.3 0.1 0.0 8.6 8.4 7.4 8.8 8.0 8.3 8.9 5.3 9.4

100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 24 16 –4 7 3 10 1 –11 12 –27 –7 23 19 19 11 14 22

–16 37 –29 –20 17 58 7 14 18 –58 –44 4 9 –13 24 57 –21

–9 –49 25 –29 1 –25 –69 –57 28,289 –48 –36 39 5 13 24 –39 133

–1 27 –19 –6 9 32 4 3 26 –47 –27 17 15 9 15 4 31

Dismissal Suspension Reduction Total* Dismissal Suspension Reduction Total

Structure (%)

New UIBs, by grounds or option (annual monthly average)

Source: Prepared by author based on BPS data.

Note:

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Year

Table 6A2.10

196

M3125 - MESSENGER TEXT.indd 196

14/05/2013 13:33

Jun-09 Jul/Aug-09 Sep-09 Oct-09 Nov-09 Dec-09 Jan-10 Feb-10 Mar-10 Apr-10 May-10 Jun-10 Jul-10 Aug-10 Sep-10

Table 6A2.11

13,390 15,248 14,844 14,362 15,150 14,814 16,267 15,344 15,391 14,469 14,719 14,572 14,014 14,439 13,739

2,439 3,059 3,429 3,737 4,179 4,550 5,165 5,234 5,185 4,922 4,911 4,938 4,782 4,783 4,671

32 310 37 2 4 0 0 0 0 0 0 0 0 1 0

215 187 192 195 182 172 186 186 184 162 189 177 165 175 177

46 46 54 65 73 74 80 77 84 80 83 80 74 83 89

0 0 0 0 349 0 1 214 166 0 0 0 0 0 0

175 152 114 103 107 75 106 115 86 86 97 105 115 139 131

7,390 7,066 5,543 4,853 5,453 3,692 4,257 4,373 3,793 3,424 3,951 4,292 5,937 7,382 6,849

0 0 0 0 0 0 694 694 694 6 4 3 3 1 0

23,687 26,068 24,213 23,317 25,497 23,377 26,756 26,237 25,583 23,149 23,954 24,167 25,090 27,003 25,656

Dismissal Dismissal Work Reduction Reduction Reduction Reduction Suspen- Suspension Grand by decree due to sion esp. total – over 50 reduction due to due to suspension Metzen dismissal dismissal – over 50

Number of UIBs, by unemployment grounds or option

197

M3125 - MESSENGER TEXT.indd 197

14/05/2013 13:33

Source:

18.1

4,483

15,016

60.7

4,658 4,865 5,070 5,277 4,519 18.0

13,627 14,419 15,235 16,238 14,752 58.8

0.1

32

0 0 0 0 20 0.1

Prepared by author, and MTSS with BPS data.

Oct-10 Nov-10 Dec-10 Jan-11 Average Total weight (%) Total Jul-09/ Jun-10 Total weight (%) 0.7

183

176 192 166 185 182 0.7

0.3

72

82 77 83 84 74 0.3

0.3

66

0 0 0 0 38 0.2

0.4

104

117 104 95 138 114 0.5

18.6

4,609

6,892 5,783 4,318 4,739 5,262 21.0

0.8

190

0 0 1 1 111 0.4

100.0

24,756

25,552 25,440 24,968 26,662 25,072 100.0

198

Table 6A2.12

Work sharing during the Great Recession

Employed persons covered by PUI

Industry Textiles Garments Tanneries and leather products Wood and wood products Basic metal industries Metal product manufacturing Machinery & equipment manufacturing Machinery & equipment n.e.c. Radio, TV, communication machinery Electrical machinery and optical apparatus Transport equipment Transport equipment Total salaried workers on partial unemployment Total salaried workers Total employed

Percentage

Salaried workers

0.64 1.06 0.6 0.5 0.04 1.16 0.32 0.14 0.01 0.16 0.14 0.12 4.9 100

5,537 9,171 5,191 4,326 346 10,036 2,769 1,211 87 1,384 1,211 1,038 42,307 865,164 1,543,329

Source: Prepared by author with CHS data, 2009.

M3125 - MESSENGER TEXT.indd 198

14/05/2013 13:33

Partial unemployment insurance programmes in Uruguay, 2009–2010

Table 6A2.13

Suspension and PUI beneficiaries, by manufacturing industry (selected months with new PUI beneficiaries) November 2009 Suspension

Sectors with PUI Textiles Garments Tanneries and leather products Wood and wood products Basic metal industries Metal product manufacturing Computers, electrical product manufacturing Electrical equip. manufacturing Machinery & equip. n.e.c. Automobile manufacturing Manufacture of other transport equip. Industries w/o PUI Food Beverages Tobacco Paper/paper product manufacturing Printing, publishing Chemicals manufacturing Pharmaceuticals manufacturing Rubber/plastics products Non-metallic mineral products Furniture manufacturing Other industries n.e.c. Industry suspension total PUI sectors Suspension total Industry/suspension total (%) PUI sectors/industry (%) PUI sectors/suspension total (%) PUI/PUI sect. suspension (%) Source:

199

March 2010

%

PUI

219 343 202

10.02 15.69 9.24

286 0 32

86 36 102 17

3.93 1.65 4.67 0.78

13

%

PUI

52 86 172

4.46 7.37 14.74

166 0 0

0 0 0 0

71 49 81 9

6.08 4.20 6.94 0.77

0 0 0 0

0.59

0

2

0.17

0

23 30 9

1.05 1.37 0.41

31 0 0

18 19 4

1.54 1.63 0.34

0 0 0

707 17 1 6

32.34 0.78 0.05 0.27

– – – –

269 6 0 12

23.05 0.51 0 1.03

– – – –

39 36 6

1.78 1.65 0.27

– – –

50 19 7

4.28 1.63 0.60

– – –

194 45 48 7 2,186 1,080 5,453 40 49 20 32

8.87 2.06 2.2 0.32 100

108 96 17 20 1,167 563 3,793 31 48 15 29

9.25 8.23 1.46 1.71 100

– – – – – 349

Suspension

– – – – – 166

Prepared by author based on BPS data.

M3125 - MESSENGER TEXT.indd 199

14/05/2013 13:33

200

M3125 - MESSENGER TEXT.indd 200

14/05/2013 13:33

Jan-07 Feb-07 Mar-07 Apr-07 May-07 Jun-07 Jul-07 Aug-07 Sep-07 Oct-07 Nov-07 Dec-07 Jan-08 Feb-08 Mar-08 Apr-08 May-08 Jun-08 Jul-08 Aug-08 Sep-08 Oct-08 Nov-08

 

103.4 103.4 106.2 103.7 105.4 104.8 106.2 106.5 106.6 106.6 107.6 107.6 108.8 108.9 108.4 108.8 108.4 107.8 108.5 108.2 108.5 109.0 106.4

92.9 90.6 88.9 89.9 99.1 93.9 96.1 101.2 105.3 99.6 100.0 100.2 100.7 99.5 97.5 95.6 95.2 96.1 99.0 93.7 91.4 89.7 87.5

94.9 97.5 100.4 95.5 96.7 99.7 98.4 92.5 93.4 91.1 88.6 85.0 84.9 83.0 86.9 94.3 87.5 95.4 93.3 97.8 96.2 95.0 86.1

Garments 102.2 101.8 106.6 108.9 111.1 106.2 108.6 108.0 105.3 107.6 107.5 111.6 107.6 113.7 111.1 111.8 113.5 108.0 102.9 98.2 99.5 98.5 95.4

107.4 113.9 121.5 121.2 112.2 126.9 117.1 124.8 117.7 131.1 133.0 126.9 148.3 157.0 135.4 128.9 136.7 133.1 135.6 122.3 129.0 125.5 127.8

Tanneries Wood

Employment Index (1996 5 100)

D manu- Textile fact. ind.

Table 6A2.14

101.1 104.1 103.4 103.5 106.8 105.7 104.0 107.7 107.7 107.1 110.7 111.4 114.7 114.9 115.7 115.7 117.8 119.8 119.5 120.6 118.6 119.1 118.8

Paper

105.5 105.2 105.5 107.4 108.3 112.7 111.7 112.7 112.7 114.3 115.5 102.2 112.8 110.9 112.1 113.3 115.1 112.2 115.0 114.2 107.0 107.4 106.9

Basic metals 111.3 102.5 113.4 104.3 113.7 108.4 114.1 109.9 110.0 107.7 109.2 107.0 107.6 109.6 109.6 106.7 103.8 108.1 108.3 98.5 101.9 105.9 102.4

Metal prod. manuf. 99.8 98.2 100.7 98.2 100.6 104.3 104.2 111.0 110.5 112.1 123.0 112.3 111.5 104.1 108.0 108.4 108.5 110.8 111.6 113.9 119.8 119.9 120.6

Machinery man. 106.4 103.2 102.2 100.2 109.0 110.8 118.4 118.3 110.9 110.3 112.7 113.5 116.0 120.2 120.1 127.6 120.9 124.0 123.5 122.4 121.0 113.2 114.4

103.5 103.9 103.3 103.4 104.4 107.6 103.5 107.5 115.9 115.0 109.6 105.7 111.6 93.9 107.7 117.8 115.4 115.3 118.6 114.6 117.9 122.1 110.2

Electric. Autoapp. mobile manuf. man.

115.4 110.4 118.9 123.3 124.9 119.7 138.2 140.4 127.1 107.3 119.7 112.1 105.4 107.7 106.9 117.1 112.8 107.9 106.6 107.9 98.6 87.5 102.9

Man. other transp.

201

M3125 - MESSENGER TEXT.indd 201

14/05/2013 13:33

107.4 105.4 104.6 105.8 105.6 104.2 103.6 102.0 103.1 105.3 105.1 107.2 107.4 105.8 107.0 108.8 108.2 108.1 108.1 106.7 105.8 106.0 105.1 107.1 108.1

Source: INE.

Dec-08 Jan-09 Feb-09 Mar-09 Apr-09 May-09 Jun-09 Jul-09 Aug-09 Sep-09 Oct-09 Nov-09 Dec-09 Jan-10 Feb-10 Mar-10 Apr-10 May-10 Jun-10 Jul-10 Aug-10 Sep-10 Oct-10 Nov-10 Dec-10

83.4 79.6 75.7 78.9 75.4 77.3 75.6 74.9 72.6 75.8 76.8 78.2 77.9 78.9 78.6 81.2 81.0 83.2 80.9 79.6 77.6 80.2 76.8 76.0 79.1

82.8 83.6 81.4 78.5 80.0 78.8 79.4 80.6 79.3 80.0 77.9 75.6 75.2 74.3 77.0 78.7 76.2 78.0 80.2 79.7 74.7 73.5 75.1 75.5 76.0

97.7 96.2 87.6 82.7 79.0 79.0 81.1 79.0 82.9 88.4 91.5 93.4 94.5 96.8 99.2 102.0 103.3 107.6 106.7 95.6 87.3 86.9 87.5 85.3 82.2

119.7 124.3 119.5 120.1 125.4 117.7 121.4 123.7 108.2 117.2 114.7 117.3 121.8 121.4 127.4 129.2 121.5 130.6 132.7 142.7 129.3 124.1 125.2 118.4 123.9

118.7 116.5 116.6 119.3 119.5 119.1 120.6 122.4 121.0 125.8 123.5 123.9 126.1 125.7 126.2 131.9 134.2 133.5 137.3 136.8 135.3 133.3 131.2 131.7 134.8

102.3 99.2 101.4 95.3 99.2 96.9 94.0 98.5 93.6 95.5 98.0 95.9 94.8 91.9 89.7 97.1 97.4 96.4 96.0 92.6 92.6 93.1 96.7 95.5 92.6

106.8 103.4 108.0 118.7 112.7 108.8 111.0 111.3 113.6 112.9 115.9 113.0 105.9 108.7 106.2 105.0 98.1 94.7 96.0 100.4 102.2 98.5 99.9 100.4 100.7

121.2 116.3 116.7 111.7 109.8 111.3 110.4 110.2 111.7 115.3 116.8 115.7 117.2 117.5 121.8 119.8 120.7 117.5 117.4 117.6 118.9 118.8 117.9 117.2 119.8

110.6 110.1 106.2 107.0 107.6 106.5 107.1 107.4 105.2 105.9 103.7 102.2 104.2 105.6 103.6 105.0 102.3 104.3 104.3 105.8 104.9 106.3 105.9 105.8 106.4

106.8 93.7 82.5 73.7 80.1 84.5 85.7 77.9 77.7 81.4 93.0 94.4 111.4 101.6 102.0 102.4 105.7 115.0 115.8 121.1 121.6 122.4 121.8 131.6 142.7

109.3 87.0 84.1 74.3 118.6 81.6 81.3 82.0 94.2 96.1 98.9 97.6 96.9 89.6 91.1 92.6 91.4 90.8 93.4 89.0 87.9 82.4 82.8 81.0 84.3

202

Table 6A2.15

Work sharing during the Great Recession

Suspension UIBs, by area of activity (annual monthly average and %)

Mining and quarrying Agriculture, forestry and fisheries Manufacturing industry Electricity, gas, water Construction Trade Transport, storage Lodgings, food services Computers, communications Finances, business activities Community, social & personal services Total

2009

%

2010

%

21 654 2,050 71 820 786 308 191 69 368 329 5,667

0.4 11.5 36.2 1.2 14.5 13.9 5.4 3.4 1.2 6.5 5.8 100.0

9 277 2,278 86 823 618 229 166 55 344 395 5,280

0.2 5.2 43.1 1.6 15.6 11.7 4.3 3.1 1.0 6.5 7.5 100.0*

Note: * Rounded. Source: Prepared by author, and MTSS based on BPS data.

M3125 - MESSENGER TEXT.indd 202

14/05/2013 13:33

7.

Work sharing as a potential policy tool for creating more and better employment: A review of the evidence Lonnie Golden and Stuart Glosser

1.

WORK SHARING AS AN INSTITUTIONALIZED PRACTICE TO PRESERVE OR CREATE EMPLOYMENT: AN OVERVIEW

‘Work sharing’, generally, is considered to be any type of policy-induced, downward adjustment of working time. Work sharing falls into two types. One is when it is designed to induce a permanent reduction in the length of work hours among all, or large subsets, of workers, in the form of shorter weekly or annual hours. These reductions may take the form of shortened standard or legal workweeks (hours after which is considered ‘overtime’), hard limits on the duration of workers’ annual, weekly or daily overtime work hours, or annual leave periods or various other forms of paid time off. The second type of work sharing is designed to induce reductions in work hours that may be temporary, including those triggered by economic crises, such as government programmes designed for the purpose of preventing or postponing planned layoffs by employers, to preserve employment or curb increases in unemployment. Such temporary work-sharing measures are usually adopted, or if already in place promoted, during a cyclical downturn, such as the recent global financial crisis and subsequent worldwide recession. The goal of work sharing is typically to add to the employment level. In the case of temporary work sharing, the aim is to promote downward adjustments of work hours to cushion the impact of a recession on as many workers as possible, while also preserving job and employer attachments that will benefit employers in the longer run by saving them often burdensome training and hiring expenses when a downturn eventually ends. The main goal of permanent work sharing is to foster more hiring instead

203

M3125 - MESSENGER TEXT.indd 203

14/05/2013 13:33

204

Work sharing during the Great Recession

of lengthening work hours. Both types of work sharing are intended to curb the rate of unemployment, the first by reducing unemployment rates in the longer term, and the second by preventing or postponing layoffs that would likely lead to higher unemployment in the short term. Both, however, may act to reduce the incidence and/or duration of unemployment, by counteracting not just cyclical but also the structural forces that may restrain job creation or encourage job destruction. Thus, even a crisis work-sharing system, if institutionalized, could become one of the built-in, automatic stabilizers for the macroeconomy, as a more proactive tool for the next recession. Its effectiveness may be facilitated if and when it is supported concurrently by other economic incentives for employers to shorten the duration of work hours, hold on to, retrain or hire workers. Moreover, both types of work sharing may also provide more job opportunities in recovery or expansionary periods. Often overlooked is the fact that work sharing may not only spread work among a larger number of workers, but also spread well-being more generally, to the extent that the layoffs would have been concentrated on certain worker subgroups who tend to suffer most in terms of income loss and re-employment fortunes, and also if the hours that are reduced would have otherwise led to symptoms of overwork, such as the adverse mental and physical health effects of longer work hours (which are well established in the literature). Support for the well-being effects of work sharing is derived from a theoretical foundation that it would counteract some of the longer-term forces that have created an increasingly inherent bias towards layoffs rather than hours reduction during recessions, and lengthening work hours during expansionary periods, such as higher fixed costs of employment (particularly in countries such as the United States). Thus, work sharing may have broader goals, of which maximizing available employment opportunities is an important subset. However, these goals also include creating more non-work time for more workers; greater potential well-being on the job for those employed, such as reduced job insecurity, reduced work stress and fatigue; and improved work–life balance. Such benefits are generally the reverse side of the costs to workers of working long hours (that is, over 48 hours per week). The structure of this chapter is as follows. Section 2 examines the theoretical case regarding work sharing and social welfare. Section 3 discusses the direction of the potential employment response to reduced work hours. Section 4 reviews previous studies that sought to estimate the number of jobs created and/or saved by reducing hours among various European and other countries. Section 5 reviews evidence regarding the employment effects of work-time reductions in North America, as well as the current state of work-sharing policies and practices there. Section 6 examines how

M3125 - MESSENGER TEXT.indd 204

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

205

the recent ‘Great Recession’ in the United States affected the behaviour of hours, employment, unemployment and underemployment. Section 7 examines empirically with time-series analysis the US manufacturing sector, in order to assess whether a case can be made for inducing hours reductions with work sharing to produce or save jobs in this sector. It also considers the nuances as to the effects of such hours reductions if implemented during periods of economic contractions vis-à-vis economic expansion periods. Section 8 provides a review on the potential worksharing effects on worker well-being. Section 9 concludes.

2.

THE THEORETICAL CASE FOR FACILITATING WORK SHARING: WELFARE, EFFICIENCY, EQUITY EFFECTS

The outcomes of any market or policy can be evaluated using five criteria – the effects on individual and social or welfare, productive efficiency, dynamic efficiency (economic growth) and social equity. In economic theory, market forces are presumed to adjust so that no one can be made better off without making someone else equally worse off, that is, in theory, there are no ‘losers’ at all. But in practice, real-world impediments such as limited information, negative externalities and relative immobility of labour and its relatively lower bargaining power vis-à-vis employers, and so on, may make it possible to improve societal well-being, on balance, with a carefully designed labour market policy intervention, such as temporary work sharing with wage supplements – referred to as ‘shorttime compensation’ (STC). Historically, both temporary and permanent work-sharing measures have been promoted mainly in the interest of promoting overall social welfare and equity, while not impinging on economic growth. By spreading the distribution of the impact of recessions more thinly, but more widely, a small number of workers would be unemployed, although at the possible risk of more workers being temporarily underemployed. The economist’s criteria may judge a policy-induced reallocation of hours as desirable as long as the ‘winners’ who benefit from the change gain more than the ‘losers’ made worse off by the change lose, on balance. Workers with reduced hours – for example, to part-time status (less than 35 hours per week) – even if voluntary, may not necessarily experience improved welfare if this leads to reduced income and/or inferior working conditions (Gash et al. 2012). Thus, in the case of temporary work sharing, the provision of partial wage compensation such as STC is crucial for improving social, and perhaps even individual, welfare. The individual welfare case derives not only from the potential cushioning of income loss

M3125 - MESSENGER TEXT.indd 205

14/05/2013 13:33

206

Work sharing during the Great Recession

for individuals or reduced risk of being laid off, but in its provision of non-monetary benefits such as time off, a more flexible employment relationship, and potentially lower stress and fatigue caused by longer hours of work. Work sharing is also a labour market policy that promotes a sort of ‘risk sharing’, by pooling the risk of the probability of unemployment. The social welfare case for a reallocation of work time generated by worksharing policies would be, in the best case, ‘Pareto improving’ in economic terms – that is, making at least one individual better off without making any other individual worse off. In the more realistic case, a net social welfare gain would occur if the marginal reductions in welfare among those who do not get the extra work hours they prefer are exceeded by the size of the gains in welfare experienced by those who would otherwise have been rendered unemployed or more underemployed. This net social welfare gain is particularly likely had some workers been working more hours than they might have preferred, that is, if they were overemployed or working involuntary but required overtime hours – earning incomes they might have been prepared to sacrifice to attain shorter hours, but not given the choice to do so. The net social welfare outcome largely depends on the extent to which actual hours and changes in them align with (initial or eventual) preferences (Lee et al. 2007). With work-sharing policies, some workers may have a perceived aversion to loss of income, which may be greater than their aversion to the loss of time. If their perceived probability of being laid off is low, then their individual well-being may be lowered. However, this situation can be altered with the use of wage supplements such as STC. Indeed, the theoretical economic case for regulatory limits on working time generally are to prevent increases in: (i) unemployment and overemployment mismatches that have proven, unfortunately, not to resolve themselves automatically (for example, Sousa-Posa and Henneberger 2002; Golden and Gebreselassie 2007; Davoine and Méda 2009; Otterbach 2009), and (ii) the ‘crowding out’ by paid work of other uses of time that have a beneficial purpose for future human and social capital development – for example, parenting, civic activity, time for educational activities and so on (see, for example, Public Agenda 2009). The latter argument also supports the long-term growth case for either form of work sharing. The efficiency case for adopting more work sharing is that it would utilize more of the available labour resources to the extent that it minimizes layoffs that lead workers to experience extended spells of unemployment. This is pertinent particularly in the case of the US in the aftermath of the Great Recession, where as high as 42 per cent of the unemployed are considered long-term unemployed – that is, they have been unemployed for over half a year (US BLS 2010). In addition, work sharing may actually contribute to efficiency, to the extent that shortened hours help firms to minimize

M3125 - MESSENGER TEXT.indd 206

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

207

their unit labour costs in the longer run through cost savings on hiring and training. Similarly, to the extent that work sharing shortens the duration and depth of an economic contraction, it contributes to the longer-term rate of economic growth. Thus, there are both a macroeconomic and a microeconomic efficiency case for work-sharing policies.

3.

EMPLOYMENT EFFECTS OF WORK SHARING: RESTRAINTS, OFFSETS AND LEAKAGES LIMITING EMPLOYMENT CREATION

What would be the realistic expectations of the employment creating effects  of a work-sharing policy? Most analyses of the potential employment-creating effects of any type of work hours reduction, especially permanent but also temporary reductions, start from a theoretical labour demand perspective. That is, they begin with a model that presumes that at any given rate of wage (W) cost, employers will choose an optimal bundle of employees (E) and work hours (H). In the ‘ideal’ worksharing policy model, the response of employment would be a one-for-one substitution following a reduction in hours per worker (Freeman 1998). That is, reducing hours by 10 per cent would theoretically open up 10 per cent more job vacancies. For example, shortening the workweek from 40 to 36 hours at a firm with 50 employees would create five new jobs at that firm. This, however, is only the best-case scenario. It is more realistic to assume that a fully proportional employment response is prevented by many constraints and complications. This is particularly likely in the short-run span of time for some effects, and in the long-run span of time for others. Thus, the predicted employment effects of work-sharing policies will depend on the timeframe, microeconomic versus macroeconomic framing, and whether the specific policy in question is a temporary versus a permanent type of hours reduction. 3.1

Potential Indivisibilities of Labour

One possible constraint on employment creation with work-sharing policies is that workers are sometimes not perfect substitutes for each other. Thus, the reduced hours of one worker may not be fully transformed into a job or more work hours for another worker (see, for example, Freeman 1998). This represents the so-called ‘indivisibility’ of labour issue. However, this argument may have merit only for skilled labour positions and knowledge workers, particularly in the short run and within the same firm. In those industries in which workers may be somewhat interchangeable and

M3125 - MESSENGER TEXT.indd 207

14/05/2013 13:33

208

Work sharing during the Great Recession

jobs or tasks are often routine, such as in manufacturing and the retail trade, the indivisibility of labour concern may be overstated. Indeed, in their quest to dismiss potential positive employment effects, work-sharing opponents have often taken this potential critique beyond the point of credibility – by claiming a belief on the part of work-sharing advocates in a so-called ‘lump of labour fallacy’. The latter term refers to the notion that at any one point in time, there is a fixed amount of work to be done in an economy, such that any increase in the amount each worker can produce would reduce the number of available jobs. Historically, this argument has been invoked to refute the truly erroneous fear that automation leads to mass unemployment. Indeed, if these structural changes are accompanied by sustained aggregate demand, there would be a displacement of job opportunities, but not necessarily unemployment. Similarly, as long as aggregate demand is healthy enough to sustain the overall demand for output, it is not fallacious to predict that either more jobs or more hours in sectors elsewhere in the economy might appear following an induced reduction in the length of the workweek, even if there is no such thing as a ‘fixed amount of labour to be done’. The work-sharing critics’ argument can be countered by any evidence that demonstrates that overall output does not fall outright following an induced decline in work hours. Thus, what is the theoretical step-by-step sequence of how policyinduced reductions in work hours may, or may not, be translated into gains in employment – that is the ‘employment effects’ and their potential ‘offsets’? Both the microeconomic and macroeconomic perspectives offer some possible complications in the transformation of hours reductions for some workers into employment opportunities for others. Many possible ‘filters’ exist in converting a reduction in hours into eventual employment-creation effects, and many of them will mitigate the extent to which employment responds to a given reduction in hours. These ‘indirect effects’ most often work against job creation rather than for it, but sometimes they occur only in the longer run. In recessionary times, holding down the rate of layoffs and preventing unemployment in the short run may take precedence. In brief, the net effect of a reduction in work hours on employment levels will ultimately depend upon the following: the extent to which employers choose to use more overtime hours above a new, lower standard workweek (if the hours reduction is made by reducing the legal workweek); the extent to which weekly wages fall in proportion to the hours reduction and the relative importance of fixed costs of labour which cannot be reduced when hours are decreased; the extent to which labour productivity per hour changes, perhaps due to reduced fatigue, reduced downtime and labour hoarding, or a greater work pace and work effort; and the lags in the adjustment process of hours and employment

M3125 - MESSENGER TEXT.indd 208

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

209

changes over time. We shall explore first the direct effect on output, and then each indirect effect separately. While there is a cost to following any policy, there is also a cost to not adopting a policy, and this includes the cost of not having temporary work-sharing measures already in place at the beginning of a recession. In addition, policies that promote reductions in hours create a potential range of lasting benefits – even beyond the potential job creation or layoff prevention – many of which may accrue in the longer run and are not apparent in the short run. Briefly, work sharing could improve overall worker welfare, particularly if it is well targeted to the sectors where hours of work tend to be longer than is healthy and/or longer than is preferred by employees. 3.2

Output Effects

There are direct effects when hours are reduced via a public policy of work sharing, particularly when the intention is to permanently reduce the standard (or ‘normal’) hours of work across all sectors.1 Employers have to either reduce output proportionately, or create more jobs to recoup the potential drop in output. While few believe that the increase will be proportional, some maintain that the fraction will be significantly greater than zero while others hold that the fraction may be closer to zero and perhaps even negative. Theoretically speaking, the simple arithmetic of reducing hours can be represented as: DH S

DQ . D(H*E)

If standard hours (H-bar) are reduced, via institutional (for example, legislative) processes, this triggers both a direct and indirect effect on employment (E). First, it produces a direct effect on actual hours worked (H). To produce the same level of output (Q) as previously, a direct effect of reduced hours would be to increase employment. However, a positive employment effect occurs only to the extent that hours and employment are potentially substitutable in the total amount of labour input in a production process. To the extent that there is indivisibility of labour (a point which was discussed earlier), it makes it difficult to replace the lost output with the hiring of a new employee, leaving output lower. Output also drops if there are complementarities between employees and work hours in an organization’s or production process’ use of labour input; for example, fewer support staff positions might be needed if factory, office or store operating/opening hours per week are shortened as a result of the shorter hours for individual workers.

M3125 - MESSENGER TEXT.indd 209

14/05/2013 13:33

210

Work sharing during the Great Recession

The ultimate effect of a reduction in work hours on employment turns on whether a macroeconomic view, rather than a strictly microeconomic one, is adopted. The microeconomic perspective is particularly useful in analysing the effect of hours reductions on employment under contrasting assumptions. For example, suppose that there are diminishing marginal returns in employment but constant returns to hours – that is, what if hiring additional employees contributes ever smaller increases to a firm’s overall output? If this is the case, then if hours are reduced by, say, 10 per cent, then the corresponding increase in employment needed to keep constant the firm’s level of output needed would be greater than 10 per cent. Thus, under such conditions, the employment effect of reduced hours would likely be positive, all else equal. In contrast, the macroeconomic (Keynesian) view emphasizes the crucial role of aggregate demand. The Keynesian perspective assumes that the total spending throughout the economy would not be altered with work sharing, as long as the total amount of disposal income does not decrease, although the composition of output produced might be altered. Thus, employment opportunities might shift across sectors, depending on the type of consumption spending that increased or decreased after work sharing, but employment would not decline in total. Indeed, in the case of work sharing, preventing layoffs and consequent unemployment might curb the spiralling downward of aggregate demand, even if underemployment might be more widespread. In brief, positive employment effects may occur at the macroeconomic level, but only if aggregate demand (that is, total spending) is not reduced, but rather is shifted in composition. This result depends upon adding more employees who are subject to ‘diminishing returns’ and on daily ‘set-up costs’ for employers being relatively low. If this is the case, it would mitigate the potentially negative effects on employment if overall output were reduced. To minimize any direct negative effects on output due to restricting work hours, evidence suggests that shortening of working hours each workday is superior to reducing the number of working days per week, given the fixed costs of employment for employers. 3.3

Real Labour Costs per Hour

The creation of employment via a reduction in hours of work is largely contingent upon the extent to which hourly wage rates effectively rise. There is little chance that permanent work-sharing policies can expand employment much if employees’ weekly earnings remain the same after the decrease in their hours of work (with the exception of a prior or consequent rise in labour’s productivity). Unit labour cost (ULC) per

M3125 - MESSENGER TEXT.indd 210

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

211

hour is likely to rise, if weekly pay is not reduced proportionately, unless the work-sharing policy includes some kind of wage subsidy to offset a rise in labour costs for employers. Otherwise, shortening hours of work raises the employer’s per hour labour cost and perhaps increases the prices of their goods or services. Such a rise in hourly labour costs may reduce demand for a firm’s output (an adverse ‘scale effect’), and thus cut the overall quantity of labour demanded. It might also in the longer run encourage firms to replace labour with capital in their production processes (which is known as the ‘substitution effect’). In addition, if work sharing results in effectively higher real wage rates per hour, there is a potential labour supply response in the longer run, in that higher real-wage rates might ultimately induce greater labour force participation. This is not a bad thing; however, in a recessionary period this might actually boost up the aggregate (industry or national) unemployment rate, particularly if many of those workers whose hours are lowered take secondary jobs outside their firm that might otherwise have gone to an unemployed job seeker. 3.4

The Role of Fixed Labour Costs and Underutilized Capital Capacity

A further complicating factor in predicting the employment effects of reductions in working hours is that there are often substantial fixed costs of hiring additional labour. These include three types of fixed costs associated with adding employees: the costs of recruiting and hiring new employees; the cost of training them; and expenses that are fixed per employee, but do not vary with hours worked, such as insurance and paid annual leave. The presence of any fixed costs associated with adjusting the employment level may result in a labour market that creates longer hours and fewer short-hours jobs than the workforce might prefer (Rebitzer and Taylor 1995). The presence and growth of such fixed costs incentivizes employers to extend hours of work per employee rather than increase new hiring during expansionary periods. Moreover, and more pertinently, fixed costs of employment incentivize employers to use layoffs rather than hours reductions during downturns. Anything that raises the ratio of fixed costs of labour thus would lead to less ‘labour hoarding’ (that is, employers holding on to employees longer than necessary) and more ‘hours hoarding’ (that is, longer hours than the minimal number needed), rather than job creation (Glosser and Golden 2004). There is a legitimate concern that firms’ capital capacity becomes more underutilized when work sharing shortens typical shift lengths. A rise in ULC will be reinforced all the more if decreased hours per worker result in reduced rates of capital capacity utilization (that is, due to shorter operating hours).

M3125 - MESSENGER TEXT.indd 211

14/05/2013 13:33

212

3.5

Work sharing during the Great Recession

Overtime Leakage

If a permanent reduction in hours of work is adopted by means of cutting the length of the standard (or ‘normal’) workweek and this policy inhibits firms’ output, then employers may choose to schedule some employees’ hours longer than the new standard hours – that is, they will increase the total number of overtime hours. This phenomenon reduces the potential amount of new job creation and is referred to as ‘overtime leakage’ (Rubin and Richardson 1997). A rise in ULC following shorter standard hours (either because of an effectively higher hourly wage rate and/or higher ratio of fixed to variable labour costs) is all the more likely when there are premium wages paid for the additional overtime work performed. Thus, employers may find it rational to keep actual hours of work nearly as long as they were before the reduced standard hours. Thus, actual hours would not be reduced commensurately with the new legal standard (again, assuming that the work-sharing policy involved a reduction in standard hours). The theoretical prediction here is that, in firms currently using no overtime hours, there will be net positive employment effects, but in firms using overtime hours ‘in equilibrium’ there may well be net negative employment effects (Hart 1987; Andrews and Simmons 2001). The ‘meta-estimate’ average elasticity is 0.80, or slightly less than unity (Hart and Wilson 1988; Owen 1989; Andrews and Simmons 2001). This means that a five-hour reduction in the standard workweek length would reduce employers’ average hours scheduled by four hours (the actual range, based on several countries’ manufacturing sectors during the 1980s was from 0.67 to 1.21). Thus, a consensus estimate of the overtime leakage is estimated at about 20 per cent, and perhaps as low as 0.10 of the initial reduction in hours when looking at the national level and not just aggregate industry level (Bosch and Lehndorff 2001). Firms facing shortages of skilled labour will raise the level of overtime hours to keep output from falling. This response may, in turn, raise hourly wage costs, increasing ULC and suppressing subsequent gains in employment. However, an overtime leakage effect tends to be higher in the short run, but dissipates in the longer run, when management can implement changes in work organization that preclude the need for regular overtime. 3.6

Skill Shortages and Mismatches

The most stubborn limit, especially regarding achieving the ultimate goal of shrinking unemployment, is the extent to which there is a mismatch or difference in the skills between the unemployed and the employed. In some cases, reducing the hours of the employed does not necessarily induce

M3125 - MESSENGER TEXT.indd 212

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

213

hiring of those who are unemployed (Freeman 1998). One potentially serious obstacle to job creation following a reduction in hours is present if there are already skill shortages in the occupation or industry where hours are reduced. Even if employers increase the number of positions, they may not be able to fill the resulting vacancies if there are skill shortages (ibid.). This divisibility of jobs problem can be serious (White 1987). Some economists predict that imposing shorter working hours under the assumption of indivisibility of labour (which was discussed earlier) would lead to more general skill shortages and consequently, undesired increases in (costly) overtime hours because of the inability to fill job vacancies. This economic view suggests that worktime redistribution will not be able to resolve unemployment or underemployment between skill groups, only within skill groups. Reducing hours can conceivably even ‘backfire’ to the extent that skilled and unskilled jobs are complementary and physical and human capital gets used less (Freeman 1998). Unskilled, skilled and high-skilled workers have different labour demand elasticities with respect to real wage rates. If most of the unemployed are unskilled (as may be the case during a period of strong economic growth), and skilled and unskilled labour are complementary, then a reduction in the worktime of skilled labour may even decrease the demand for the unskilled unemployed (ibid.; Bauer and Zimmermann 1999). 3.7

Productivity Offset

Working-time reduction in the form of a shorter standard workweek is likely to reduce productivity per worker, because shorter hours are clearly likely to reduce output (Q) levels, at least initially. However, one of the sources of confusion in the debate over the eventual employment effects of hours reductions is that there are two, separate definitions of ‘productivity’. Not only is there productivity per worker, but there is also the productivity per hour of work. Even while total output is reduced, work-sharing policies may yield an increase in average productivity per hour. Most likely, the decline in output will be less than proportionate to the decline in actual hours. This is the case if the last hour of work, on the margin, were less productive than the average hour of work. If the average productivity of labour rises as hours are reduced, then average productivity per hour (Q/H) may increase to keep output (Q) constant. The resulting ‘productivity offset’ precludes the need for employers to add employees, at least less than a number proportionate to the decline in hours per employee. A productivity offset may develop if worker fatigue is reduced, the pace of work per hour is intensified, or so-called ‘hoarding’ of labour is reduced (Whitely and Wilson 1986; Owen 1989). For example, if hours are

M3125 - MESSENGER TEXT.indd 213

14/05/2013 13:33

214

Work sharing during the Great Recession

reduced by 10 per cent but output drops by only 5 per cent, the productivity offset amounts to 50 per cent. Employers will thus be able to recoup the lost output while adding only proportionately half of the drop in hours. Following an hours reduction, management has a window of opportunity to reorganize work and production so as to improve productivity (White 1987). In addition, firms may respond to a reduced standard workweek by rearranging (rationalizing) work hours. This may occur across the day, week or even, perhaps, across the year (‘annualization’) in ways that promote higher productivity per hour. Thus, reductions in working hours may generate ‘shock effects’ that increase hourly productivity. If firms lose a proportionately smaller amount of output by shortening standard hours of work, they need not hire as many new employees to make up the lost output. The productivity effect of reducing hours may not be symmetric at every level of hours. The size of the productivity offset tends to be positively correlated with the level of average hours. That is, if hours of work are already lengthy, a reduction in hours policy may engender a large productivity offset. However, the productivity offset may be much smaller when average hours are already in the range of 35 to 40 hours per week. That is, a 10 per cent reduction from 60 to 54 hours may yield productivity per hour gains in excess of 50 per cent, but the same 10 per cent reduction from 35 to 31.5 hours per week might yield only a tiny gain in productivity per hour. The net effect on job creation may turn on the existing extent of unproductive time in the course of the working day. Reducing the workweek will generally increase the employment level, but this policy may serve to aggravate the employment situation if daily productivity were already very high and the resulting productivity offset is low, insufficient to cover a substantial part of the costs of employment adjustment (Domingueza et al. 2011). Based on available studies, the estimated size of the productivity offset varies between 10 and 50 per cent (Roche et al. 1996; Bosch and Lehndorff 2001). While the gain in productivity may obviate the need to hire more labour, thus undermining the goal of an immediate gain in employment, such a result does have a longer-term advantage. Thus, from the micro-level perspective, greater productivity of labour per hour induced by hours reduction surely restrains subsequent growth in jobs (Owen 1989). However, when expanding the scope to the aggregate (industry or macroeconomic) level, labour demand theory suggests higher productivity increases the relative value of an hour of labour input vis-à-vis other resources. If hours reduction leads to higher per hour labour productivity, then labour eventually becomes relatively more attractive. For example, labour becoming a more productive resource might lead firms to add

M3125 - MESSENGER TEXT.indd 214

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

215

a job as opposed to making an investment in a financial instrument. In the long run, the benefit of productivity gains per hour can also show up in the form of greater labour retention, as well as increased market competitiveness (White 1987). Firms’ products and services become more price competitive both domestically and abroad, thus enhancing rather than stifling job-creation effects. Thus, higher labour productivity per hour may stimulate more job creation in the aggregate over the longer run, even while suppressing employment at any particular firm in the short run (Owen 1989).2 3.8

Offsetting the Offsets? Conditions under which Employment would Expand

In sum, the size of the subsequent employment effects of reduced work hours may be suppressed indirectly by any subsequent increase in average overtime hours used; a drop in output supplied or demanded due to higher unit labour costs for employers per hour; the extent of fixed relative to variable labour costs; the rise in average labour productivity per hour; a reduced rate of capital capacity utilization; a greater extent to which technologies used in production allow work to be divisible among positions and hours; and the adequacy of qualified labour available to fill the new positions created. The overall net effect on employment depends on the extent to which the potential direct effect of increasing employment is offset by potential indirect effects that obviate a need for new hiring. If work sharing leads inevitably to higher non-wage labour costs, upward pressure on wages (ironically by reducing the extent of unemployment), and even reduces the pressure on government to pursue alternative solutions to unemployment (Snower 1995), then the prospects would be bleak for it as an employment-enhancing policy tool. However, the conventional microeconomics-based critique of the effects of hours reduction on employment may well be overstated. For example, shortened shifts conceivably can be accompanied by adding more shifts. Indeed, firms that add a new shift may actually boost employment if they seek new employees, provided that aggregate demand is sufficient to purchase the output being produced by the shorter but greater number of work shifts. In addition, the pessimistic view may assume too much immobility between skill levels and occupations. Mobility or upgrading may occur in actuality at the workplace level, particularly during periods of macroeconomic expansion. For example, such mobility can occur if employers can change job content, for example, eliminating clerical tasks from professional jobs (perhaps reversing direction of the tendency which has been occurring up until now because of computerization). Moreover,

M3125 - MESSENGER TEXT.indd 215

14/05/2013 13:33

216

Work sharing during the Great Recession

the potential problem of an inadequate supply of skilled labour and a resort to more overtime hours is likely to occur only in countries known to be experiencing skill shortages rather than countries with extensive systems of vocational training (Bosch and Lehndoff 2001). All of the assumed adjustments are subject to lags, making it difficult to predict and disentangle short-, medium- and long-run effects (Owen 1989). For example, productivity offsets may exist only in the short run and evaporate over the longer term. Projections of the potential employment effects span the spectrum, from clearly positive effects (Jacobson and Ohlsson 2000; Lanoie et al. 2000) to perhaps even a negative employment impact (Kapteyn et al. 2004). For example, employment effects might be positive with small reductions in working hours, while larger reductions may reduce employment (Marimon and Zilibotti 2000). In addition, examining micro-level cases suffers from a potential bias that would understate how many viable jobs crisis work-sharing schemes can save. For example, studying only firms that apply for such programmes is subject to a selection bias – participating firms tend to be either less competitive than other firms or performing more poorly (Calavrezo et al. 2009). Thus, such firms are less likely to have avoided layoffs or to have created employment in the absence of the programme. With the existence of so many alternative scenarios, work-sharing advocates need to show how reductions in working hours of various types can potentially lead to increased employment, rather than just engage in counter-critiques of the claims made by opponents of work hours reduction (White 1987). The wider lens suggests that the employment effects of working-time reductions might be positive if the initial level of hours is already low; the reduced hours are accompanied by restrained per hour wages and/or reinforcing government subsidies for hiring, such as reduced payroll taxes (Bosch et al. 1994); gains in productivity make labour a relatively more attractive resource to retain or hire, or firms become more competitive in international trade (Cette and Taddei 1993); new work shifts are added; the unemployed are trained in the skills needed to get hired by firms whose hours per worker have been shortened; and the ratio of fixed to variable labour costs is reduced, for example, by pro-rating the costs of employee benefits. Depending on how hours reductions are implemented, it is quite possible to do so without driving up unit labour costs, increasing overtime work, restraining capital capacity, having unfilled skilled job vacancies, or hindering growth in output. Perhaps most importantly, a significant (perhaps near proportional) reduction in wages would enhance the employment effects of working-time reductions (see for example, van Ginnekan 1984). To generate more jobs, the use of overtime could be discouraged by changing the underlying economics of the

M3125 - MESSENGER TEXT.indd 216

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

217

hours decision, for example, reducing the ratio of fixed to variable costs of labour. This objective could be accomplished by transforming into variable any costs that are currently fixed per employee. Pro-rating holiday and vacation (annual leave) time and employer contributions to insurance and to payroll taxes – based on hours of work rather than the number of employees – would help counter much of the bias towards layoffs and longer hours incentivized by fixed labour costs. What will be needed is a range of policies adopted to ‘offset the offsets’. These policies should be designed to reduce the fixed costs of employment (disincentives to hiring); encourage the reorganization of working time along with increased capital capacity utilization; and reduce the long-run adverse output effects, for example, by encouraging productive leisure and inducing new labour force participation.

4.

ESTIMATED EMPLOYMENT EFFECTS: EVIDENCE FROM EUROPE AND OTHER COUNTRIES

There is no single, definitive estimate of the number of jobs created or saved with work-sharing policies. However, there is a broad set of estimates from a range of case studies, industry studies and macroeconomic simulations. Much empirical analysis has developed since the firstgeneration literature considered the potential employment effects, from the micro to the macro level (for example, Conference Board in Europe 1981; Ehrenberg and Schumann 1982; Cuvillier 1984; Nemirow 1984; Williams 1984; Drèze et al. 1986; Zachmann 1986; Booth and Schiantarelli 1987; Hill 1987; Nyland 1989; Niefer-Dichmann 1991; Taddei 1991; Fromont 1993; International Labor Review 1993, 1995; Roche et al. 1996; Blyton and Trinczek 1997; EIRR 1997). Different estimates of the employment effects of working-time reductions may be caused by different conditions in the implementation of the reduction, by diverse methodological approaches, and by divergent theoretical views (Bosch et al. 1994). There are simply too many idiosyncrasies involved to reduce to a single number the size of the employment effect of a given reduction in working hours. First, the hours reduction can take on a variety of forms – temporary or permanent, general or targeted by industry, occupation, age group or across the board, or, via shorter standard weekly hours, by limitations on overtime hours, or by encouraging part-time jobs, earlier retirement, or leaves of absence. It also matters whether the measures are voluntary or compulsory, statutory or incentive based, aimed at weekly, annual or lifetime hours and so on. Such measures may or may not include

M3125 - MESSENGER TEXT.indd 217

14/05/2013 13:33

218

Work sharing during the Great Recession

government-subsidized crisis work-sharing schemes during cyclical downturns, targeted public subsidies for hiring or training and unemployment insurance (Messenger and Rodríguez 2010). The ultimate employment effect may depend in part on whether an hours reduction is pushed by the government, workers or employers, and the degree to which it is voluntarily accepted on the part of workers and complied with (or resisted) on the part of employers. Second, the employment effects of working hours reductions may differ depending upon the magnitude, location (country), timing (year or business-cycle stage) and the pace of the hours reduction. It also may vary according to occupation (white collar or blue collar) and industry (goods or services production) and the concurrent changes or matches in labour supply and demand in the industries or occupations experiencing hours reduction. Each enterprise, industry and country may have experienced or introduced working-time reduction along with varying degrees of work reorganization, hours flexibilization, extended operating hours or capital capacity utilization, wage rate increase or restraint, public sector subsidization and so on. Third, a particular measure adopted might anticipate the potential offsets and explicitly attempt to prevent them as part of the hours reduction scheme. For example, firms and unions have negotiated increases in the planned utilization rate in the framework of working-time reductions. Fourth, the actual effects of work sharing on employment are difficult to ascertain empirically also because some of the effect occurs through job retention rather than job creation (Blyton and Trinczek 1997). Indeed, job preservation may be the predominant and most valuable contribution of work-sharing policies. It is rational from an economic theoretical point of view that temporary work-sharing measures that promote job preservation should be stronger than the job-creation effects of reduced hours. In the latter, firms must incur the costs of recruitment, hiring and training, whereas in the former, firms do not incur these costs because the employees are already with the company (Bosch and Lehndorff 2001). Fifth, estimation procedures each employ a particular model whose scope may or may not be broad enough to capture all of the possible offsetting and indirect effects of hours reductions on employment in the short, medium or long run. The timeframe matters, in order for all the offsets and eventual employment effects to work themselves out (Owen 1989). The short-term effect of hours reductions on employment is considered to be over a single phase of the business cycle. In contrast, the longterm effects take into consideration eventual adjustments to the capital stock or education and training of labour. The medium term is a period longer than one business cycle, for changes in the (physical and human)

M3125 - MESSENGER TEXT.indd 218

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

219

capital stock to begin adjusting to a change in work hours, but not yet be fully adjusted. Thus, for example, even when there are measurable effects on overtime hours, they may tend to dissipate over the longer run, when the next cyclical downturn arrives. Thus the job-preserving effects of hours reductions may be delayed and not be felt until a future recession (Bosch and Lehndorff 2001). Finally, deriving an estimate of the employment effect of a working-time reduction involves disentangling the effects of working time, economic growth and productivity. Each of these effects will offset or reinforce the direct employment effect of whatever form of hours reduction is pursued. While most empirical studies presume that the analysis starts from a ‘shock’ to working time that leads to increased productivity and employment, we know that productivity growth in the long run may lead back to hours reduction. Thus, not only are these variables simultaneous, but they are mutually causal or endogenous. Virtually all empirical studies of actual European cases show some positive employment effects of working-time reductions. Nevertheless, these studies also find increases in overtime and labour costs that prove to be restraining factors that inhibit the employment creation effects (ibid.). Most countries in Europe have initiated both work-sharing policies designed to spread employment opportunities, as well as temporary crisis work-sharing measures (often referred to as ‘short-time work’ schemes). Indeed work sharing is among the most common forms of European responses to downturns in employment (Cazes et al. 2009). Like many of the conventional, microeconomic-based studies, projecting the effects of restricting work hours, such as a reduction to no more than 35 hours per week, depends largely on the set of assumptions adopted regarding the productivity, cost and wage effects of the reduction. For example, if output per worker remained at at least 90 per cent of its previous level, employment would grow by 6.3 per cent, rather than decline as it would if output shrank substantially. Moreover, if total wages paid were to decline by 1.4 per cent, then employment would actually grow by 13.3 per cent (Fitzgerald 1998). When hours reductions are accompanied by an employment subsidy to reduce quasi-fixed costs of employment, while attempting to maintain workers’ take-home pay by preventing downward adjustments in their earnings, this policy may be likely to increase employment (Erbas et al. 2001). Empirical studies from the 1980s indicate, on average, a positive employment effect of at least 40 per cent and perhaps as high as over 70 per cent of the initial change in hours (Bosch and Lehndorff 2001). Larger-sized reductions in hours appear to have greater employment effects than only marginal changes in hours. The net employment effect of an imposed reduction in hours is estimated to be

M3125 - MESSENGER TEXT.indd 219

14/05/2013 13:33

220

Work sharing during the Great Recession

positive in eight of the 12 industries examined by a UK study (Booth and Ravallion 1993). In the UK, five of six studies found negative elasticities, ranging from –0.39 to –1.47 between actual hours reduction and subsequent employment increases; the computed average among all six studies is an elasticity of about –0.60 – meaning that a 10 per cent decline in hours is associated with a 6 per cent increase in employment in these samples (Andrews and Simmons 2001). In a panel of 52 engineering industries over 1978–82, robustly negative estimates of the employment elasticity with respect to standard hours were found (Hart and Wilson 1988); for the majority of firms in the sample who use overtime hours, the elasticity of employment averaged –0.50. However, for the nine firms offering no overtime, the elasticity was actually positive. Thus, an important precondition for reduced standard hours to translate into increased employment appears to be that firms start out having overtime hours, which presumably cushions them against unit labour cost increases when they convert hours into jobs. In Germany, labour unions began to reduce standard hours on an industry-by-industry basis during the 1980s in an attempt to raise employment. One analysis of industry-level data suggests that such ‘work sharing’ may have reduced employment in the 1984–94 period, but this was because the hourly wages rose enough to offset the decline in actual hours worked (Hunt 1999). Another German study simulates the reverse case of people working more hours for the same pay, and projecting whether employment will increase or decrease; it finds that no positive effect on the number of persons employed can be expected (Conrad et al. 2008). The most studied policy regarding permanent working-time reductions is reductions in the standard (or normal) legal workweek (International Labor Review 1993; Boissonat 1996). Most prominent among recent reductions in the standard workweek was the 2000 reduction from 39 to 35 hours per week in France (which was based on the two laws, Aubry I and Aubry II). This working-time reduction was supported by concurrent wage subsidies that effectively lowered non-wage labour costs for employers, as well as providing employers with opportunities to vary hours per week above the standard, as long as they averaged 35 hours per week over a year and did not exceed other prescribed limits (for example, a maximum of 48 hours per week). Its success in raising employment is at least partly attributable to the moderation of wage costs (Schreiber 2008). The employment level grew faster than forecast and the unemployment rate fell more than would have been predicted. Wages and labour demand afterwards slowly recovered from the initial shortening of working hours, as output appeared to be relatively unaffected (Logeay and Schreiber 2006). The most widely cited figure regarding the employment effect of the

M3125 - MESSENGER TEXT.indd 220

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

221

French ‘35 hours’ is a net creation of 350,000 jobs, and perhaps as many as 500,000, over and above the job creation that would have occurred even without the working-time reduction. However, this result fell short of the government’s original promise and forecast of 700,000 jobs over three years, in large part because work time fell less than originally expected, as many firms avoided a full 10 per cent cut in hours by changing their way of calculating work time, and also because under Aubry II firms were no longer required to hire at least 6 per cent more workers to receive the subsidies, as they were in the first round of workweek cuts – allowing companies to absorb much more of the impact than expected through hourly productivity gains. The most valid estimates of net employment gains from the French policy were 6.0 to 7.5 per cent in firms that benefited from the first round (Aubry I and 1996 Robien law incentives), and an initial estimate of about 3 per cent in firms that reduced hours without these additional incentives. Not only did the policy package avoid labour cost increases for businesses, but most of the 35-hour employees cited quality-of-life improvements despite the wage moderation, greater variability in their schedules, and intensification of work (Hayden 2006). In Portugal, in 1996, the maximum standard workweek was reduced from 44 to 40 hours. For workers involved, the reduction in Portugal appears to have resulted in a reduced separation rate (the sum of layoffs and quits). The fact that monthly earnings remained constant for workers also meant that some of them actually faced a higher risk of job loss – those who were working fewer than 40 hours, however, not those whose hours were reduced as a consequence of the law. It also increased hourly wage rates, which kept workers’ monthly earnings approximately constant. However, those whose hours were reduced as a consequence of the law were not adversely affected (Raposo and van Ours 2010). The Brazilian government is now considering a similar reduction of 44 down to 40 hours per work as the standard, along with an increase in the overtime wage premium. The latter may help accomplish these ends, perhaps even by itself, since a simulated tax rate of 12 per cent of overtime wages was found to reduce the workweek from 40 to 35 hours per week. Indeed, this tax change was projected to boost employment by 7 per cent; however, it was also projected to reduce productivity per worker by about 4 per cent (Osuna and Ríos-Rul 2003). In Venezuela in 1991, the standard length of the workweek was reduced, but only for blue-collar workers (Bujanda and Fairris 2011). The impact of this legislation on blue-collar type employment was estimated (with panel data and a difference-in-differences identification strategy), to calculate transitions into employment for affected workers. Work-sharing policies were found to be successful in raising blue-collar employment in Venezuela, by a moderately sized

M3125 - MESSENGER TEXT.indd 221

14/05/2013 13:33

222

Work sharing during the Great Recession

but statistically significant amount; results suggest a 6 percentage point increase in employment in the second period following the reform. Recent OECD analysis found that work-sharing measures in Japan, preserved jobs that otherwise would have been lost in the recession (OECD 2010b; see also Ogura, ch. 4 in this volume). The OECD identified a key source of the effect as the encouragement of employers to adjust to lower demand by reducing workers’ hours rather than resorting to layoffs. The OECD concluded that in up to 24 countries where some type of crisis work-sharing measure (‘short-time work’) was operated during the recession, such measures were effective in preserving jobs, at least those of permanent employees (OECD 2010a). In most countries, the largest part of the reduction in average working time occurred outside of such short-time work schemes, however; these other working-time reduction measures included reduced overtime hours, employer-initiated reductions in standard working hours for full-time workers, and increased numbers of (often) involuntary part-time workers (Houseman 2011). Estimation results of the impact of crisis work-sharing (short-time work) measures during the Great Recession found clear evidence in 11 of 19 countries that permanent (but not temporary) jobs were preserved during the economic downturn. These measures also increased the incidence of average hours reductions among permanent employees. There was a smaller negative effect on permanent employment of a given 10 per cent reduction in output during the crisis period in those countries with crisis work-sharing (short-time work) measures, relative to countries without such schemes (OECD 2010a: Table 1.4).

5.

ESTIMATED EMPLOYMENT EFFECTS: EVIDENCE FROM NORTH AMERICA

Samuel Gompers, the founding leader of the American Federation of Labor (AFL) in the 1880s once famously declared, ‘if one worker is unemployed then hours of work are too long’. While the eight-hour standard working day and 40-hour standard working week were established in the US (and a number of other developed countries) in the 1930s, working hours have not declined markedly in the US since that time as compared to most EU countries (Bluestone and Rose 2000). Despite this, US policymakers have not seriously considered policies of working-time reduction, in stark contrast to most continental European countries, with the exception of periodic, limited discussion of temporary work-sharing programmes during recessions. Perhaps this in large part explains the paucity of empirical and case studies in the US regarding the employment effects of working-time reductions. Almost all studies of the effects of permanent

M3125 - MESSENGER TEXT.indd 222

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

223

work-sharing measures are of countries that have implemented such laws in Europe, or European industries that have collectively bargained such hours reductions. Some of these studies include a pooled sample of OECD countries, which may or not include North America. Freeman’s (1998) survey of studies of UK time series found that for every 1.0 per cent (unit) decline in hours, induced by market forces, employment responded upward by 0.1 to 0.7 per cent. In many EU countries, the employment effects of shortening work hours are on the relatively small side, which may be because working hours are already relatively short in those countries. Yet, the potential for positive employment effects from work-time reductions may actually be greater in countries such as the US, where fulltime workweeks are longer than in most European countries. However, the presence of strong employment protection laws in the EU limits employers’ recourse to layoffs (Gray 1996). This also might serve to limit the amount of employment that can be created with work-sharing policies. Some estimates of the relationship between hours and employment in the US found that (relative to Germany), US employment adjusts considerably more than average hours to a given reduction in output demand (Abraham and Houseman 1995). Similarly, in contrast to other OECD countries, most of the adjustment of labour input to a negative ‘shock’ to output occurs through employment rather than working hours in the US (see, for example, van Audenrode 1994). However, since 1979, average workweeks have been absorbing more of a given shock to output demand than in previous cyclical swings (Glosser and Golden 1997). While the response of work hours has become even more concentrated immediately following downward shocks to output, the lag time of the employment response to increased output demand has lengthened, particularly in expansion phases (Glosser and Golden 2004). This phenomenon has produced more extended periods of ‘jobless recoveries’ following recessions (Glosser and Golden 2005). This may reflect in part the absence of a policy of work hours reduction or restraint in the US, which might help offset the factors that are stoking demand for longer hours and delaying hiring, such as skill-upgrading of jobs and ‘lean staffing’ strategies during upturns, and also reducing employment more than hours during downturns. A simulation of various working-time reduction policies computed the impact of a policy of reducing the number of days in the workweek to four days (Hamermesh 1996), and found that average hours would decline by 18 to 20 per cent (the largest reductions in hours would occur in certain private industries – coal mining, oil and gas, water transport, and the textile manufacturing industries). If instead, working time were reduced via a restriction to a 7.5 hour maximum workday, average hours would be reduced by only 6 to 9 per cent (the largest impact in manufacturing

M3125 - MESSENGER TEXT.indd 223

14/05/2013 13:33

224

Work sharing during the Great Recession

would be in the cement, dairy products and beverage industries). The change in employment necessary to keep output constant, following an imposed reduction in work hours, through reducing either the number of days or the number of daily hours, was estimated. Generally, output could be made up through relatively smaller increases in employment than increases in daily hours. The simulated loss of output found was large, the smallest being as high as 4 per cent. Thus, a four-day reduced workweek might well create jobs, but at a steep cost in terms of reduced output. On the other hand, a mandatory reduction in the length of daily hours has a less severe potential impact on output. To make up the lost output following a mandated shorter working day, the number of days per week need only be raised by 1.5 to 5 per cent. Indeed, it would be relatively easy to restore the lost output by increasing employment, but more difficult to do so by raising daily hours. Hamermesh (ibid.) concludes that employment might be expanded significantly following a mandated reduction in daily hours of work. In Canada, it was estimated that if only half the 6.4 million hours of overtime worked by Canadian workers could be converted into full-time jobs, employment would be boosted by up to 80,000 individuals. A simulated 10 per cent reduction in standard hours phased in over five years would reduce the number of unemployed by 4.1 per cent, with only a negligible impact on output growth and inflation (Donner Commission 1995). However, cutting in half the existing overemployment rate, from 10 to 5 per cent of workers (that is, redistributing hours from the overworked to the underworked) would have a limited ability to create new jobs filled by the currently unemployed, albeit it may still be a worthwhile public policy (Drolet and Morissette 1997). In addition, a natural experiment of sorts occurred between 1997 and 2000 in the Canadian province of Quebec. It reduced its standard workweek from 44 down to 40 hours, with the primary aim of stimulating employment growth. The Quebec policy involved no mandated wage increases to compensate workers for the reduced earnings due to fewer hours. The evidence suggests that weekly hours did experience a 20 per cent reduction (among full-time workers) in the weekly hours worked beyond 40. However, the policy did not appear to raise employment at either the provincial level or within industries where hours of work were most affected (Skuterud 2007). Interest in work-sharing policies in North America tends to revive with every recession, and descriptive analyses of the programmes follow a few years afterward, usually touting their favourable features or results (for example, McCarthy and Rosenberg 1981; Meltz et al. 1981; Nemirow 1984; Kerachsky et al. 1985; Best 1988; Owen 1989; Golden 1990; Hunnicutt 1992). Estimates of their employment impact date back to the

M3125 - MESSENGER TEXT.indd 224

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

225

President’s Reemployment Agreement (PRA) of 1933 – it appears to have created nearly 2.5 million new employment opportunities, in around four months. However, the programme also required firms to raise hourly wage rates, offsetting close to half of these gains (Taylor 2011). Although cutbacks in work hours and work-sharing policy was encouraged in the US and especially in Canada in the late 1970s and early 1980s, it turned out to be ephemeral (Jacoby 1985; Huberman 1997), whereas in the Canadian unemployment insurance (UI) system, temporary work sharing was integrated only in the 1990s (Huberman and Minns 2007). In terms of the empirical findings of studies of crisis work-sharing programmes in North America, the first such programme in the US was established in the State of California in 1979, and an 1986 assessment of that California programme found that the unemployment rate of employees in firms was 11 to 12 per cent using work sharing with STC, versus 14 per cent in the control group of non-users (Vroman and Brusentsev 2009). In the US, a study of three state UI systems using STC displayed varying results (Kerachsky et al. 1986): Oregon was able to achieve a one-for-one ‘layoff conversion rate’, where each hour of STC substitutes precisely for an hour of layoff, and Arizona’s programme averted some layoffs (the results from the third state, California, were invalid due to a flawed sample). Using an approach that relied on surveying employers, the New York State Department of Labor computed that work sharing with STC averted about 4,000 layoffs per year during the period from 1988 to 1994 (Needles and Nicholson 1988, claim that this figure is implausibly high). In the California programme during the 2001 recession, 67 per cent of firms who used the programme were in the manufacturing sector; even within manufacturing, just over 5 per cent of firms in groups with UI used work sharing with STC (MaCurdy et al. 2004: Table 1). Since some firms do not belong to a UI group, this amounted to only 3.2 per cent of all firms in the state. Nevertheless, a few of the industries stand out in terms of the intensity of their use of STC, in particular, electronics and primary metals manufacturing, and to a somewhat lesser extent, instruments, fabricated metals and furniture. Across the US, while use of work sharing with STC has generally been slight, it spiked during and immediately following the last two recessions, with the most widespread use in the manufacturing and wholesale trade (Gray 1996; MaCurdy et al. 2004; Balducchi and Wandner 2008; Messenger 2009; Vroman and Brusentsev 2009; Shelton 2011). Through the end of 2008, short-time compensation benefits rarely reached even 1 per cent of all the unemployment benefits paid annually. However, this ratio rose to 2 per cent in 2009 and 1.2 per cent in 2010. In the Great Recession years of 2008–09, about 220,000 employees were involved in the 23 states’ work-sharing/STC programmes (a 24th state,

M3125 - MESSENGER TEXT.indd 225

14/05/2013 13:33

226

Work sharing during the Great Recession

New Jersey, adopted a programme in early 2012), amounting to 0.73 per cent of the total full-time equivalent change in private sector employment during that period. While this is small as a proportion of production worker employment, in manufacturing it constituted a less trivial-looking proportion of almost 5 per cent. In Canada, use of the Work Sharing Program between 1990–91 and 2001–02, measured as the proportion of all job separations, was almost as meagre as in the US. The number of participants ranged from a low of 7,683 in 1997–98 to a high of 112,357 in 1990–91. The corresponding average number of layoffs attributed to a shortage of work was about 3 million a year, and one half of these layoffs resulted in a UI claim. This translates into work-sharing claims being only 1.2 per cent of all layoffs and only 2.4 per cent of all claims. None the less, work-sharing arrangements involved almost 300,000 workers and almost 11,000 employers since 2008, averting thousands of layoffs (‘American Income Life & National Income Life’ 2011). In Canada, the majority of participants (67 per cent) in the Work Sharing Program were employed in the manufacturing sector. The incidence of work sharing in manufacturing was almost four times as high as in the entire workforce (8.4 versus 2.3 per cent), and almost the same rates in the wholesale and retail trade and business services, although the percentages were trivial elsewhere (Human Resources and Skills Development Canada 2004). Evaluations found an estimated 43 to 67 thousand jobs were saved by Canada’s Work Sharing Program during the years 1989 through 1991, mainly by averting layoffs (Needles and Nicholson 1988; Gray 1996). The greater success in Canada speaks to their far more widespread use in Canada versus the US (Meltz et al. 1981). The scale to which temporary work sharing is used in Canada is still quite small, however, because the current design of the programme limits the demand for work sharing on the part of firms (Siedule et al. 1996). Assessing the efficacy of the programme is complicated by the fact that work sharing with STC may just postpone rather than prevent layoffs. Nevertheless, more individuals likely would have been unemployed, which would have raised unemployment by 0.70 percentage points in 1987, reaching 1.5 per cent higher by 1990. In addition, because of the positive health and job satisfaction outcomes of work-sharing participants versus those who were laid off, the programme’s benefit to cost ratio was estimated to be as high as 2.6 to 1 in the long run, over twice as high as in the shorter term (Freeman 1998). More recently, in two US states that promoted work-sharing programmes with STC benefits more extensively among employers, Rhode Island and Connecticut, the proportions of private sector employment that appeared to have been preserved by the programmes were 4 and

M3125 - MESSENGER TEXT.indd 226

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

227

2 per cent, respectively, and much higher in the manufacturing sector – 16 and 23 per cent of manufacturing employment (Houseman 2011). The Rhode Island State Labor Department calculates that the state processed more than 12,000 initial work-sharing claims in 2010, which is estimated to have  helped hold down its unemployment rate. Launched in 1992 as WorkShare, the Rhode Island programme is credited with averting more than 14,000 layoffs in Rhode Island since 2007. The State Labor Department said work sharing helped avoid an estimated 9,550 layoffs in 2009 and 2010 (Wall St. Journal, 21 November 2011). In 2009, Rhode Island’s employers were five times as likely to turn to work sharing as employers in the other US states that had programmes, on average. Nationally, according to the US Department of Labor, work sharing programmes saved approximately 165,000 jobs in 2009 – nearly triple the number of jobs saved in 2008 – plus another 100,000 jobs in 2010. Of the five million jobs lost across the US in 2009, 220,000 of them might have been saved if the whole country were on a work-sharing programme with STC (Houseman 2011). If as many employees had participated in work sharing/STC in 2009 as the rates observed for Kurzarbeit in Germany, the number of jobs preserved would have been almost one million – avoiding as much as one in eight job losses during the recession. Moreover, one prominent macroeconomic study estimated that if work-sharing programmes with STC benefits were expanded to all 50 US states, for every $1 in Federal government funds devoted to finance it, the programmes would return $1.69 in real GDP one year later (Zandi 2009).

6.

THE GREAT RECESSION IN THE US: EMPLOYMENT REDUCTION STILL DOMINATES HOURS ADJUSTMENT

Employment adjustments bear the brunt of recessions in the US, particularly during the most recent crisis. The rate of layoff pattern is quite similar in the durables and non-durables production industries, which contrasts with average weekly overtime hours in manufacturing, Figure 7.1 shows rates of layoffs (from JOLTS data that began only in 2000, and therefore are not used further). Over the course of the Great Recession in the US, aggregate weekly hours of work (total employee-hours) plummeted by over 9 per cent (Kroll 2011). Employment reductions comprised 6.6 per cent of that change, while hours reductions constituted only 2.6 per cent. The average workweek fell during the period, but by less than one hour in the private sector as a whole. The domination of employment reduction as opposed to hours reduction was apparent in the manufacturing sector,

M3125 - MESSENGER TEXT.indd 227

14/05/2013 13:33

228

Work sharing during the Great Recession 2.3 2.2 2.1

Rate

2.0 1.9 1.8 1.7 1.6 1.5 1.4 2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Note: Shaded areas indicate US recessions.

Figure 7.1

Rate of layoffs and discharges, total private employment, 2001–2011 (seasonally adjusted data)

as well as in the construction industry. This result occurred despite rather sizeable downward adjustments in the workweek in the goods-producing sectors. However, hours adjustments occurred more in sectors where employment was not declining during the recession – utilities, education and health services. Thus, while hours of work declined during the recession in the US, this reduction makes up for no more than roughly onequarter of the overall adjustment in labour input which occurred over the period. On the other hand, weekly overtime hours have not been eliminated– despite the economic downturn. They did drop by an hour over the recessionary period, at least in the sector for which these data are collected, manufacturing (Figures 7.2–7.5). For production and non-supervisory workers, the drop and subsequent increase appear to be somewhat starker in contrast to all employees, but otherwise similar in pattern. It is also worth noting that in the US, there has been a labour marketinduced, ‘work sharing’ of sorts, although a haphazard one reflecting cyclical market forces. This situation likely reflects neither a shift in workers’ preferences nor a public policy intention. Indeed, underemployment rose markedly during the Great Recession – which began earlier in the US – and persisted in the subsequent ‘jobless recovery’. By the end of 2009, the share of the workforce working part-time hours ‘for economic reasons’ peaked at over 6 per cent. Specifically, a greater proportion of workers are employed ‘part-time’ (defined by the US Bureau of Labor Statistics (BLS) as 1–34 hours per week) for ‘Economic Reasons’, specifically due to ‘Slack Work or Business Conditions’. Figure 7.6 displays the recent,

M3125 - MESSENGER TEXT.indd 228

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

229

3.50 3.25

Hours

3.00 2.75 2.50 2.25 2.00 2006

2007

2008

2009

2010

2011

Note: Shaded area indicates US recession.

Figure 7.2

Average weekly overtime hours: manufacturing, 2006–2011 (seasonally adjusted data)

3.50 3.25

Hours

3.00 2.75 2.50 2.25 2.00 1.75 2006

2007

2008

2009

2010

2011

Note: Shaded area indicates US recession.

Figure 7.3

Average weekly overtime hours of all employees: durable goods, 2006–2011 (seasonally adjusted data)

unprecedented spike in the number of such workers. That number remains about double the number that it was prior to the recession. In 2009, over seven million workers reported that their less than full-time hours are due to ‘slack work or unfavourable business conditions’ (including seasonal declines in demand), plus another just over two million workers who reported that they ‘could only find part-time work’. The part-time workforce that is willing to work for longer hours for the additional income is surely just a subset of the ‘underemployed’, those

M3125 - MESSENGER TEXT.indd 229

14/05/2013 13:33

230

Work sharing during the Great Recession

6.0 5.5 5.0 Hours

4.5 4.0 3.5 3.0 2.5

99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11

98

19

97

19

96

19

95

19

94

19

93

19

92

19

91

19

19

19

90

2.0

Note: Shaded areas indicate US recessions.

Figure 7.4

Average weekly overtime hours of production and nonsupervisory employees: durable goods (seasonally adjusted data)

5.00 4.75 4.50 Hours

4.25 4.00 3.75 3.50 3.25

10 20

08 20

06 20

04 20

02 20

00 20

98 19

96 19

94 19

92 19

19 90

3.00

Note: Shaded areas indicate US recessions.

Figure 7.5

Average weekly overtime hours of production and nonsupervisory employees: non-durable goods (seasonally adjusted data)

workers who would prefer more work hours in order to earn more income, regardless of whether they are employed at full-time or part-time hours (see Golden 2006). During the latest recession, this hours reduction has occurred largely as an unguided adjustment by employers, some of which has been through reduced hours in companies’ full-time workforces, and

M3125 - MESSENGER TEXT.indd 230

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

231

7,000

Thousands of persons

6,000 5,000 4,000 3,000 2,000 1,000

1990

1992

1994

1996

1998

2000

2002

2004

2006

2008

2010

Note: Shaded areas indicate US recessions.

Figure 7.6

Employment: part-time due to slack work or business conditions (seasonally adjusted data)

undoubtedly, also in their non-regular workforces – that is, part-time and temporary workers. In 2009, the average number of weekly hours declined to 33 hours per week (based on establishment survey data), the lowest level on record since data have been collected. Average hours of those employed in the cyclical ‘bellwether’ industry of temporary help services showed a dramatic dip from over 34 hours per week in the autumn of 2008 down to 31.9 hours in September 2009. In contrast, the average workweek in manufacturing, and in particular overtime hours, dipped by only about half an hour over the course of the recession. Thus, the cyclical downturn was absorbed by non-supervisory and production workers’ hours in the more cyclically sensitive, part-time and temporary type of jobs. As a consequence, one of the challenges of promoting crisis work-sharing measures is that while they have the potential to curb layoffs and reduce unemployment, employers already do so, for some workers, in recessions. However, this reduction is done in a way that exposes more workers to the risk of becoming ‘underemployed’, working shorter hours or fewer days or weeks than they prefer. The ‘underemployment rates’ by industry illustrate the degree to which shorter hours have become more prevalent in the US for cyclical reasons, roughly doubling across most industries. It also appears that the climb in ‘involuntary part-time’ work by 2010 was steepest in public administration, construction, trade and information services. With the exception of leisure and hospitality and other service industries, these also happen to be the sectors with the highest rate of underemployment (Konczal 2010). Underemployment rates largely tend to mirror, although not precisely, the industry pattern of unemployment. Unemployment in this recession has

M3125 - MESSENGER TEXT.indd 231

14/05/2013 13:33

232

Work sharing during the Great Recession

been considerably more concentrated in construction, and is also above the national average in the manufacturing, agricultural and mining, professional and business services, and the recreation and leisure industries (US BLS, 2010). These patterns of hours suggest that even though some formal worksharing plans were adopted by some companies in the 24 US states that now have work-sharing/STC programmes, employers have adjusted hours downward just as workers seem to prefer more income. Perhaps average hours might have dipped even further, if not for some companies’ formal implementation of policies providing for some guaranteed minimum average number of hours per employee per week.

7.

TIME-SERIES RELATIONSHIPS BETWEEN OUTPUT, HOURS, EMPLOYMENT AND LAYOFFS: CASE ANALYSIS WITH US DATA

Over the business cycle, hours of work are known to complement employment, decreasing when the economy is slowing down and rising when the macroeconomy is recovering or expanding. However, working time is also a potential substitute for adjustment in employment during recessionary periods, with reduced hours acting as a brake on reductions in jobs. Thus, work-sharing policies have to spread employment opportunities among a larger group of workers, whether such opportunities are increasing or decreasing. There are many reasons why jobs may not be directly substitutable for reduced hours, at least proportionately, which are explored in detail above. Indeed, even without these complications, since both workweek lengths and employment levels are moved by the same, business-cycle forces, it may be all but impossible to disentangle hours, employment and layoffs. Figure 7.7, illustrates that the inherent challenge – output variation, as captured by the industrial production index – is closely tied to the length of the workweek, in a bi-directional way. The figure shows the percentage changes from the previous year in output, employment and the workweek. Workweek adjustments immediately follow output changes (for example, Glosser and Golden 1997), absorbed mainly in the first month following an output increase or decrease. The complicating factor is that longer hours not only eventually signal increases in new hiring by employers, but they often result in additional income, particularly among non-supervisory and production workers, which in turn will be spent and fed back into greater output. The converse case would be replicated in the downward direction when average hours per worker contracts for a given level of employment.

M3125 - MESSENGER TEXT.indd 232

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

233

15 10 5 0 −5 −10 −15

Output

Hours

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

−20

Employment

Note: Shaded areas indicate US recessions. Source: FRED-Federal Research Economic Data.

Figure 7.7

Average weekly hours, employment and output in manufacturing, 1990–2011 (percentage change from year ago)

Note that in the shaded areas of Figure 7.7, average weekly hours (in manufacturing) dropped during the recession, but the extent of that drop was dwarfed by the precipitous drop in output and employment. Indeed, the employment-level change appears to closely track the change in output, with a lag of about a couple of months. In the previous two recessions, a workweek drop preceded the drop in output and employment. However in the recent recession, the workweek drop was more marked, but led an eventual recovery first in output and then in employment. Employment clearly has bounced back in the period from mid-2009 to 2011, but the magnitude of that increase has not been enough to avoid being characterized as another ‘jobless recovery’. While the workweek appears to have been more responsive to the drop in output, it still pales in comparison to the dramatic drop in employment in the US during the recession. We first observe simple correlation associations. We expect that these will be positive, because of the inherent simultaneity between output, hours and employment. What we may find, however, is that some sectors may exhibit less positive, zero, or perhaps even an inverse correlation. The more negative, or smaller positive, suggests a greater potential for substitutability. Table 7.1 shows correlations between changes in both hours and employment, lagged for six months each. It observes four periods: the whole decade from 2000 to 2011; just the 2007–09 recession; the overall

M3125 - MESSENGER TEXT.indd 233

14/05/2013 13:33

234

Table 7.1

Work sharing during the Great Recession

Correlations between weekly hours and employment levels, 6 months of lags

Decade Durables manufacturing Non-durables manufacturing Wholesale trade Retail trade Transp and warehousing Utilities Financial activities Education and health services Recession Durables manufacturing Non-durables manufacturing Wholesale trade Retail trade Transp and warehousing Utilities Financial activities Education and health services Slump Durables manufacturing Non-durables manufacturing Wholesale trade Retail trade Transp and warehousing Utilities Financial activities Education and health services

M3125 - MESSENGER TEXT.indd 234

0

1

2

3

4

5

6

0.176

0.206

0.233

0.220

0.273

0.262

0.239

0.050

0.107

0.123

0.135

0.089

0.074

0.166

0.151 –0.031 0.082

0.161 0.007 0.116

0.184 0.169 0.082 –0.018 0.226 0.069

0.161 0.103 0.138

0.174 0.035 0.155

0.125 0.029 0.131

0.109 0.061 –0.054 –0.080

0.021 –0.087 0.018 –0.122

0.013 0.055

0.078 –0.127 0.044 –0.113

0.045 –0.049 –0.018

0.130

0.013

0.018

0.051

0.330

0.463

0.375

0.374

0.290

0.302

0.326

0.254

0.232

0.320

0.266

0.143

0.313

0.229

0.415 0.028 0.220

0.304 0.274 0.135

0.268 0.111 0.219

0.161 –0.003 –0.050 0.173 0.239 0.150 0.018 0.045 –0.026

0.019 0.172 0.218

0.292 0.227 –0.203 –0.200 0.140 –0.179 0.153 0.114 0.009 –0.086 0.002 –0.060 0.042 –0.033 –0.041 –0.111 –0.068 –0.111 0.028 0.041 0.101

0.221

0.296

0.314

0.254

0.262

0.256

0.327

0.120

0.140

0.168

0.188

0.111

0.189

0.159

0.305 0.022 0.227

0.283 0.147 0.106

0.278 0.113 0.217

0.266 0.082 0.017

0.185 0.194 0.109

0.146 0.131 0.004

0.174 0.075 0.179

0.152 –0.016 –0.034 –0.093 0.040 –0.016 0.008 –0.043 –0.081 –0.068 –0.027 –0.015 –0.021 –0.051 –0.063 –0.086 –0.083 –0.148 –0.052 –0.151 –0.100

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

Table 7.1

235

(continued) 0

Recovery Durables manufacturing Non-durables manufacturing Wholesale trade Retail trade Transp and warehousing Utilities Financial activities Education and health services

1

–0.285 –0.248

2

3

0.090 –0.164

4

5

0.035 –0.020

6 0.190

–0.376 –0.148 –0.263 –0.074 –0.161 –0.111 –0.102 0.119 0.226 0.060 –0.210 0.262 0.059

0.303 0.307 0.197 –0.216 0.297 –0.115

0.287 0.255 0.086

0.239 0.293 0.023 0.071 0.019 –0.027

–0.073 –0.372 0.211 –0.022 –0.339 0.213 –0.009 –0.190 –0.291 –0.093 –0.190 0.031 –0.213 –0.101 –0.142 –0.050 –0.064 –0.051 0.043 –0.154 –0.030

‘slump’ from 2007–11; and just the recovery period following the recession from mid-2009 to 2011. Such potential substitutability appears to be slightly greater in the non-durables type of manufacturing than in durables. Substitution between hours and employment is most apparent, when the scope includes the entire decade, in the construction and information industries. It appears to be greater also in certain industries during the recession (and the continuing slump since then) – information and professional and business services. Negative correlations are more apparent during the mid-2009 to 2011 recovery period. That result suggests that employment has been restrained by the increasing length of the workweek during the recovery. Thus, work-sharing policies may be just as powerful a tool to promote employment, and counter a jobless recovery, during the recovery period following a recession as it might be during the recession itself. Next, we model the relationship between hours and employment in a more sophisticated fashion, using time-series analysis techniques.3 In order to identify the impulse responses, we apply a Choleski decomposition to a priori impose the following causal ordering {Q, H, P, W}: Q →H→P→W. This technique will permit us to observe the existing pattern of both hours and employment following a ‘shock’ (one-time adjustment) of a given (‘unit’) size in output (Q, industrial production index), and, where applicable, shocks to the average workweek (H), and real weekly wages (W). A shock to the average workweek would represent a work-sharing simulation. However, the difficulty of observing this result will be compounded by the likelihood that a reduction in work hours, if not accompanied

M3125 - MESSENGER TEXT.indd 235

14/05/2013 13:33

236

Work sharing during the Great Recession

0.4 0.3 0.2 0.1 0.0 −0.1 −0.2 0

Figure 7.8a

5

10

15

20

25

30

35

40

45

Impulse response of hours (H) to a shock in output (Q) in manufacturing sector

by wage compensation, then eventually feeds back negatively on output demand – reflecting the simultaneity problem that makes it so difficult for us to isolate the substitutability of hours for employment. We first run this model on the aggregates for which monthly data and longer time-series data are available, notably the manufacturing sector, then the durables and non-durable goods production subsectors of manufacturing. We explore whether sectors that experienced a greater downward adjustment in hours also underwent a subsequently smaller decline in employment, all else constant. The estimates include both the size of the hours and employment responses and, importantly, the lengths of lag times in the response (represented by the bold black line, in the figures). Finally, we attempt to infer the extent to which a forced shock downward in hours – for example, a policy of work sharing – might be expected to impact positively on employment, and how that might vary by sector, including those outside of manufacturing. The impulse response (IR) estimate will be considered the ‘realistic case scenario’, whereas the upper and lower bounds of the confidence intervals will reflect the ‘bestcase scenario’ and ‘worst-case scenario’, respectively. Figure 7.8a shows that a one unit increase in the level of output would raise the workweek by about 0.20 to 0.30 of a unit.4 The effect of that output increase declines after about a year, and dissipates to zero after about the 15th month. However, the size and length of the effect of an output shock on hours pales in comparison to the effect of output on employment adjustments, shown in Figure 7.8b. The effect on employment, however, builds only gradually, peaking at about the same month as when an output shock’s effect on working hours begins to decline. Moreover, the effect on employment, although lagged, is in scale about three times the size of the effect on hours. This suggests that, (i) at least

M3125 - MESSENGER TEXT.indd 236

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

237

1.25 1.00 0.75 0.50 0.25 0.00 −0.25 0

Figure 7.8b

5

10

15

20

25

30

35

40

45

Impulse response of employment (P) to a shock in output (Q) in manufacturing sector

0.30 0.25 0.20 0.15 0.10 0.05 0.00 −0.05 −0.10 0

Figure 7.9a

5

10

15

20

25

30

35

40

45

Impulse response of hours (H) to a shock in output (Q) in manufacturing sector during expansionary periods

after some lag time, employment adjustments dominate hours adjustments among employers in the manufacturing sector, and (ii) after about a year, the substitution within labour inputs, from hours to employment, starts to settle in. Both of these results subtly point to the potential utility of worksharing measures – which would both raise the responsiveness of working hours, vis-à-vis employment, so that more of the adjustment to the change in output is absorbed through hours, and perhaps also delay the time between employers’ adjustment of hours and their adjustment of employment during downturns, as well as quicken the lag time during upturns. Figures 7.9a and 7.9b show the impulse responses during expansion periods only, to distinguish between the entire period and the expansion

M3125 - MESSENGER TEXT.indd 237

14/05/2013 13:33

238

Work sharing during the Great Recession

0.5

0.4

0.3

0.2

0.1 0.0

0

Figure 7.9b

5

10

15

20

25

30

35

40

45

Impulse response of employment (P) to a shock in output (Q) in manufacturing sector during expansionary periods

periods alone. By default, we can observe whether the response to a shock would behave differently during the downturn phases of the cycle (data back to the year 1990 do not provide sufficient months of downturns to yield reliable degrees of freedom for generating tests of statistical significance, unfortunately). The behaviour is somewhat different. Figure 7.9a shows that the response of hours (H) to a one-unit shock in output (Q) in a given month is slightly less than in the overall period, about 0.20 in the initial month following that shock. In the next 12 months, the effect dwindles, reaching zero at about the 12th month out (after which the confidence interval suggests that the impulse response is no longer statistically significant). By default, this suggests that hours adjustment is a bit stronger during downturns than during upturns. In Figure 7.9b, production worker employment (P) adjustment appears to be far smaller, only about half in scale (only 0.40), and also more persistent (significant for over four years after the shock as opposed to only three), during upturns. This result suggests there is perhaps some room for work-sharing measures to both curb and delay downward employment adjustments during the downturn stages of the economic cycle. It also accounts for the recently observed phenomenon of ‘jobless recoveries’ during upturns, with longer lagged responses in employment to a given increase in output demand. Figures 7.10 and 7.11 apply data from 1990 to 2011 to distinguish the behaviour of the two main subsectors of manufacturing, durables and non-durables goods production. They portray the adjustment of employment and hours, to both a unit-sized shock in output and a shock in hours. The left-hand column of Figure 7.10 shows that a shock to output in durables manufacturing appears to initially impact on hours, at 0.30 the size of the shock, which then decreases in size until it is wiped out after about

M3125 - MESSENGER TEXT.indd 238

14/05/2013 13:33

239

Output

Hours

Response of:

1.75 1.50 1.25 1.00 0.75 0.50 0.25 0.00 −0.25

0.5 0.4 0.3 0.2 0.1 0.0 −0.1 −0.2

2.25 2.00 1.75 1.50 1.25 1.00 0.75 0.50 0.25 0.00

Figure 7.10

Employment

M3125 - MESSENGER TEXT.indd 239

14/05/2013 13:33

10

10

10

15

15

15

20

20

20

25

25

25

30

30

30

35

35

35

40

40

40

45

45

45

1.75 1.50 1.25 1.00 0.75 0.50 0.25 0.00 −0.25

0.5 0.4 0.3 0.2 0.1 0.0 −0.1 −0.2

10

15

20

25

30

35

40

45

0.50

0.75

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 −0.1 −0.2 5

5

10

10

15

15

20

20

25

25

30

30

35

35

40

40

45

45

−0.50

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 −0.1 −0.2

0.5 0.4 0.3 0.2 0.1 0.0 −0.1

−0.50

−0.25

0.5 0.4 0.3 0.2 0.1 0.0 −0.1

0.00 −0.25

0.00

0.25

5

Hours

0.25

0.50

0.75

Shocks to: 2.25 2.00 1.75 1.50 1.25 1.00 0.75 0.50 0.25 0.00

Impulse responses to shocks in output and hours: durables manufacturing, 1990 to 2011

5

5

5

Output

240

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2

Figure 7.11

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

0.30 0.25 0.20 0.15 0.10 0.05 0.00 −0.05 −0.10

Output

Response of: Hours

Employment

M3125 - MESSENGER TEXT.indd 240

14/05/2013 13:33

10

10

10

15

15

15

20

20

20

25

25

25

30

30

30

35

35

35

40

40

40

45

45

45

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

0.30 0.25 0.20 0.15 0.10 0.05 0.00 −0.05 −0.10

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 −0.1

0.5 0.4 0.3 0.2 0.1 0.0 −0.1

Shocks to: 0.9 0.7 0.6 0.8 0.5 0.7 0.4 0.6 0.3 0.2 0.5 0.1 0.4 0.0 0.3 −0.1 0.2 −0.2

5

5

5

10

10

10

15

15

15

20

20

20

25

25

25

Hours

30

30

30

35

35

35

40

40

40

45

45

45

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 −0.1

0.5 0.4 0.3 0.2 0.1 0.0 −0.1

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 −0.1 −0.2

Impulse responses to shocks in output and hours: durables manufacturing, expansionary periods between 1990 and 2011

5

5

5

Output

Work sharing as a potential policy tool for creating employment

241

15 months (when the lower bound of the confidence interval crosses zero, suggesting that the impulse response cannot be concluded to be above zero). Conversely, the initial impact on employment is negligible in the immediate months following a shock, but the effect builds gradually over time. It peaks at about the same time the effect on hours has run out. This suggests that there is a substitution of employment for hours to a one-time shock in output, but only after a several months’ lagged response. What about the effects of a given shock in average weekly hours? This is shown in the right-hand column of Figure 7.10. Using this scenario to capture the potential effects of work-sharing policies and practices is complicated, however. The reason is that adjustments in work hours, particularly among the non-supervisory and production workers, is directly and positively associated with income, which when spent, may well push up output demand, and then indirectly, also the level of hiring and thus employment. Consequently, it is likely, because of this complementarity, that a positive (negative) shock to hours might yield an overall positive (negative) response in employment. However, if we can somehow screen out the size of the effect of hours on output demand (top right figure), then the net effect is more likely to reflect the potential effect of altered hours on employment. In the case of durable goods manufacturing, Figure 7.11 shows that this effect of hours on output (shown in the top panel of the right-hand column of the Figure) is, however, not statistically significant. Thus, a shock in hours in this sector has not been associated with inverse adjustments in employment (bottom right panel of the figure); indeed there is some indication to the contrary. Nevertheless, during expansionary periods, an increase in hours does in fact raise output, at least after a time lag of a few months. Thus, it might appear that the potential employment-creation effect of limiting hours (via work-sharing policies) is more present in the recovery and expansion phases in the cycle than during the downturn phase. This result may seem curious, but it does suggest that introducing more work sharing, if sustained through a recovery, could promote a more rapid switchover among employers to hiring than might have occurred otherwise. In the non-durables industry, the magnitudes of the responses of both hours and employment appear to be more muted (see Figures 7.12 and 7.13), but the pattern is similar to durables. The employment response peaks at about a year following an output shock (third row, first column of Figure 7.12), about the same time that its effect on hours dissipates fully. The effect on both hours and employment degenerates more quickly, however, than it does during expansionary phases in the durables industry. A superior way of replicating the effect of work-sharing policies, using VAR estimates, would be to shock hours in the downward direction,

M3125 - MESSENGER TEXT.indd 241

14/05/2013 13:33

242

1.0 0.8 0.6 0.4 0.2 0.0 −0.2

Response of:

Figure 7.12

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 −0.1 −0.2

0.30 0.25 0.20 0.15 0.10 0.05 0.00 −0.05 −0.10

Output

Hours

Employment

M3125 - MESSENGER TEXT.indd 242

14/05/2013 13:33

10

10

10

15

15

15

20

20

20

25

25

25

30

30

30

35

35

35

40

40

40

45

45

45

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 −0.1 −0.2

0.30 0.25 0.20 0.15 0.10 0.05 0.00 −0.05 −0.10

1.0 0.8 0.6 0.4 0.2 0.0 −0.2

0.2 0.1 0.0 −0.1 −0.2 −0.3 −0.4 −0.5

0.6 0.5 0.4 0.3 0.2 0.1 0.0 −0.1 −0.2

0.3 0.2 0.1 0.0 −0.1 −0.2 −0.3 −0.4 −0.5

Shocks to:

5

5

5

10

10

10

15

15

15

20

20

20

25

25

25

Hours

30

30

30

35

35

35

Impulse responses to shocks in output and hours: non-durables manufacturing, 1990 to 2011

5

5

5

Output

40

40

40

45

45

45

0.2 0.1 0.0 −0.1 −0.2 −0.3 −0.4 −0.5

0.6 0.5 0.4 0.3 0.2 0.1 0.0 −0.1 −0.2

0.3 0.2 0.1 0.0 −0.1 −0.2 −0.3 −0.4 −0.5

243

Output

Hours

Response of:

0.2 0.1 0.0 −0.1 −0.2 −0.3 −0.4

0.25 0.20 0.15 0.10 0.05 0.00 −0.05 −0.10

0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15

Figure 7.13

Employment

M3125 - MESSENGER TEXT.indd 243

14/05/2013 13:33

10

10

10

15

15

15

20

20

20

25

25

25

30

30

30

35

35

35

40

40

40

45

45

45

0.2 0.1 0.0 −0.1 −0.2 −0.3 −0.4

0.25 0.20 0.15 0.10 0.05 0.00 −0.05 −0.10 30

30

35

35

40

40

45

45 0.6 0.5 0.4 0.3 0.2 0.1 0.0 −0.1 −0.2

25

30

35

40

45

−0.2 20

−0.2 15

−0.1 10

0.0 −0.1

0.0

5

0.1

25

25

0.2

20

20

0.1

15

15

0.3

10

10

0.2 0.1 0.0 −0.1 −0.2 −0.3 −0.4

0.2

5

5

Hours

0.3

0.6 0.5 0.4 0.3 0.2 0.1 0.0 −0.1 −0.2

0.2 0.1 0.0 −0.1 −0.2 −0.3 −0.4

Shocks to: 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15

Impulse responses to shocks in output and hours: non-durables manufacturing, expansionary periods between 1990 and 2011

5

5

5

Output

244

Work sharing during the Great Recession

1

0.5

0 1

2

3

4

5

6

7

8

9

10

11

12

−0.5

−1

−1.5

O OH OHW OW

−2

Figure 7.14

Impulse responses of employment to simultaneous shocks in output (O), hours (H) and/or wages (W): durables manufacturing, contractionary periods between 1978 and 2011

simultaneously with a shock downward in output, to simulate a recession, using only data from months in the downturn (contraction) periods. To yield more observations, we added data back to 1978 and considered ‘downturn’ periods to be months of ‘negative growth gaps’, where output’s growth rate was less than the average growth rate between 1978 and 2011. The resulting downturn periods correspond closely with the traditional US National Bureau of Employment Research (NBER) recession dates. Figures 7.14 and 7.15 display the results of this exercise for the durables and non-durables industries, respectively. Figure 7.14 displays the most crucial result of this research: in the durables sector, when only an output (O) shock occurs, employment decreases substantially. The O impulse response line suggests that employment in the overall durables manufacturing sector, if there were no corresponding reduction in the workweek, would drop by almost 1 per cent following a 1 per cent shock downward in output. However, when a downward hours shock is introduced at the same time as the output shock, employment responds positively, for at least 10 months following the shock. This represents the ‘realistic case’ scenario (the small sample size, however, precludes the possibility of generating reliable confidence intervals). Shocking hours downward by one unit, proportionally equal to the unit decline in output,

M3125 - MESSENGER TEXT.indd 244

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

245

0.4 0.2 0 1

2

3

4

5

6

7

8

9

10

11

12

−0.2 −0.4 −0.6 −0.8

O OH OHW OW

−1

Figure 7.15

Impulse responses of employment to simultaneous shocks in output (O), hours (H) and/or wages (W): non-durables manufacturing, contractionary periods between 1978 and 2011

is associated with a moderately sized employment gain which builds up to about 0.80 gain in the sixth month following the shock. The difference between the O and OH is quite consistent, indicating an employmentcreating effect. Given that the level of employment in durables manufacturing is 8.8 million before the start of the recession, this translates into an estimated 80,000 jobs lost. However, if there had been a 1 per cent reduction in work hours accompanying the downward output shock, the OH implies that, within six months, employment would have been higher than otherwise. This result translates roughly into 26,000 jobs potentially saved in durables manufacturing, if workweeks would have declined proportionally with a given output drop. This preliminary estimate is our key finding that would support the employment case for work-sharing policies. In addition, if workers’ earnings were reduced proportionately with output, this would have somewhat exacerbated the employment decline and limited the prevention of employment loss. This implies that protecting wages from a proportional decline would prevent some employment loss as well. In the non-durables goods industry, however (Figure 7.15), if a reduction in hours accompanied a negative output shock, the subsequent drop in employment does not appear to be mitigated substantially. There is, however, a slight moderation of the employment decline that would

M3125 - MESSENGER TEXT.indd 245

14/05/2013 13:33

246

Work sharing during the Great Recession

0.8

O OH OHW OW

0.6 0.4 0.2 0 1

2

3

4

5

6

7

8

9

10

11

12

−0.2 −0.4 −0.6

Figure 7.16

Impulse responses of employment to simultaneous shocks in output (O), hours (H) and/or wages (W): durables manufacturing, expansionary periods between 1978 and 2011

have occurred without hours being shocked downward, provided that the output and hours reductions are accompanied by reduced wages. This neutralizing effect occurs particularly within the first month or two following the negative shock to output. Figures 7.16 and 7.17 show the impulse responses in the durables and non-durables manufacturing industries, respectively, to a downward shock to hours when there is an upward shock to output. In the durables manufacturing industry, reducing hours during an expansion may have little employment creation effect, and might even have counterproductive effects unless accompanied by corresponding reductions in wages. In the non-durables industry, in contrast, reducing hours during an expansionary period actually does slightly increase employment, eventually, beginning in about the 9th month following the output gain. Reducing wages proportionately (OHW lines) would work against employment creation during expansions in the non-durables industry. In sum, there appears to be some promise for preventing severe employment declines during downturns with some form of work-sharing policy – a reduction in workweek lengths in conjunction with downturns in output – in preserving employment during downturns. This would occur most profoundly in the durable goods production industries in manufacturing (although not in the non-durables industries). Aside from the downturn periods, employment appears to take a long time to

M3125 - MESSENGER TEXT.indd 246

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

247

0.2 0.1 0 1

2

3

4

5

6

7

8

9

10

11

12

−0.1 −0.2 −0.3 −0.4 −0.5 −0.6

O OH OHW OW

−0.7

Figure 7.17

Impulse responses of employment to simultaneous shocks in output (O), hours (H) and/or wages (W): non-durables manufacturing, expansionary periods between 1978 and 2011

substitute for increases in hours, particularly during expansions. This inertia in employment clearly also reflects forces other than the degree of variability in average work hours. However, in addition, there are some nuances in the patterns by sector. Examining the effect of a potential unanticipated alteration in average work hours is difficult because it cannot be completely disentangled from the effect of such a shock on output demand. Nevertheless, there appears to be no net negative effects on employment, once consideration is taken that output moves with hours. Thus, realizing the potential positive effects of work-sharing policies on employment is going to be a more targeted and nuanced rather than a one-size-fits-all affair. The effects of work-sharing policies would vary by industry type and stage of the business cycle. The existing patterns of adjustment in the US suggest that reductions in working hours may be a more powerful tool in terms of protecting or increasing employment levels in durables manufacturing when the economy is heading into a recession, and in non-durables manufacturing when it is already recovering from a recession. Future empirical research along these lines ought to broaden the sample to earlier years, so that disaggregated industries could be identified within the durables manufacturing sector where the impact of hours reduction during recessions would be most fruitful for preserving

M3125 - MESSENGER TEXT.indd 247

14/05/2013 13:33

248

Work sharing during the Great Recession

or creating greater levels of employment with subsidized work-sharing efforts. Further research should also incorporate into the VAR model some additional variables, such as overtime weekly hours and multifactor productivity indexes (available quarterly), to account for some further nuances in the relationship between hours and employment. Finally, consideration should also be given as to why the durable goods sector behaves so differently than non-durable goods during recessions and also during expansions.

8.

WORK SHARING AND SOCIAL WELFARE: FROM SPREADING WORK TO SPREADING WELL-BEING

The potential benefits of work sharing ought to be considered on a wider scope than its effect just on employment and unemployment (important as that is), but also on the potential increase in overall well-being of those who are or remain employed. First, to the ears of most macroeconomists, it must sound puzzling that work sharing and other forms of hours reduction might be judged to be of limited value or even detrimental to the economy because they trigger a boost in average labour (or organizational) productivity. Even if all of an hours reduction is absorbed through the productivity offset rather than employment gains, this may still be a worthwhile economic policy. If unit labour cost increases can be avoided, for example, via intensification of work or increased capital capacity utilization, hours reduction can be a ‘win–win’ public policy. It may promote greater welfare, efficiency and equity, all without harming economic growth. While permanent work-sharing policies may not profoundly reduce unemployment, the social and economic benefits of hours reduction occur through a variety of channels to both individuals and the macroeconomy. Second, the effects of reductions in working hours extend not only to the unemployed, but also to those who are employed as well. The increased non-work time might be welcomed by two groups of workers: first, those who are most at risk of overwork symptoms, such as work stress, fatigue, pain and injury risk, and second, the overemployed, who have an unrealized preference for fewer hours of work despite the reduction in income that it entails. Regarding the first group of workers, the potential benefits to the employed of shortened work hours can be ascertained by simply reversing the evidence of the adverse effects, risks and hazards of long working hours on workers’ mental and physical health, including an improved sense of work–life balance (Dembe et al. 2008). For example, work sharing with STC in Canada resulted in better morale, attitudes towards work and

M3125 - MESSENGER TEXT.indd 248

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

249

management, and overall health for the work-sharing participants’ group in contrast to other workers (Siedule et al. 1996). Thus, many workers whose hours are reduced will likely experience reduced work stress, fatigue from work and work–family conflict, even if it were not their voluntary choice to reduce their working hours (and income). Reduced stress is all the more likely if hours were involuntarily long, because, when employers set hours without employee input, interference of one’s job with family is relatively stronger. Interference is also stronger among full-time workers who prefer part-time hours than among full-timers who prefer full-time jobs. In contrast, being part-time but preferring full-time hours (one form of underemployment) has no bearing either way on work–family conflict. Thus, full-time workers who have their hours reduced to part-time will lessen their work–family conflict, especially if the reduction is preferred, but this appears to occur even if the reduced hours were not preferred. When employees feel overemployed, this exacerbates work–family conflict and work stress. Being underemployed, working fewer than desired hours, does not reinforce the adverse effects of long hours on these indicators of welfare (Golden et al. 2011). Indeed, workers’ reported overall level of ‘happiness’, when controlling for workers’ level of annual income, health and all demographic factors, is negatively associated with the number of days per month that they work ‘extra hours’ beyond their usual hours among (hourly paid but not salaried) workers. Happiness is also greater for workers who have flexibility in their work schedules, the ability to take time off during the day, and those whose shifts are regular and predictable. When employees feel overemployed, it also exacerbates unhappiness, all else constant. This supports evidence from one work-sharing programme that reducing hours (and earnings) by 10 per cent, produced a generally favourable employee reaction, particularly for those who used the extra unpaid non-work time for leisure activities. A four-day workweek (via compressed workweeks, furloughs or extended unpaid vacation time) also produced unexpected gains in job and work–life satisfaction, while also averting some layoffs, in the public sector (Facer and Wadsworth 2010). Thus, shortened work hours via work sharing, despite decreased income, is likely to be associated with improvements in welfare among workers even if the shorter hours are not strictly voluntary, although especially when the reduced hours would be preferred, since overemployment is a more potentially damaging ‘hours mismatch’ than underemployment (see, for example, Wooden et al. 2009). The net social welfare gain would be the sum of the increased time off for the employed (welcomed especially by the overemployed), plus the gain for those who are unemployed or underemployed if some of those reduced hours of work can be transformed into jobs and earnings for them.

M3125 - MESSENGER TEXT.indd 249

14/05/2013 13:33

250

9.

Work sharing during the Great Recession

CONCLUSION: TOWARD TARGETED, PREFERENCE-BASED AND INDUSTRYSPECIFIC HOURS REDUCTIONS

The analysis to this point suggests that the biggest gains in overall wellbeing could be achieved if work-sharing measures, whether temporary or permanent, were targeted most assertively to particular types of workers or to particular industries – those in which unemployment, underemployment and overemployment rates are highest or most detrimental. Even in the much milder recession of 2001, overemployment was significant in a few sectors (see Golden 2006). That is, the proportion of workers who report a desire for working fewer hours, even if that means less income, varies by the industry in which workers are employed. Thus, work-sharing policies might be most fruitful, from the point of view of minimizing workers’ well-being loss, when adopted for workers whose hours exceed 40 hours per week (in the US, there appears to be a strong preference for work hours at precisely the legal standard of 40 hours per week). This is in contrast to countries such as Germany or the Netherlands, where a job with 36 hours appears to be more acceptable to most workers who harbour preferences for either a part-time job or a new full-time job (see, for example, Bloemen 2008). There are many types of employer actions that may lead to reduced working hours per worker, not only involuntary but voluntary arrangements. This includes obvious permanent work-sharing measures, such as a shortening of the standard workweek after which overtime pay is owed, for all or a given classification or industry of workers. However, it also includes options such as limiting mandatory overtime work; requiring that employers provide employees opportunities to reduce to part-time hours at their current job with pro-rated compensation; and increases in paid and unpaid time off (including vacations, seasonal, sabbatical or other forms of extended temporary leave). The difference between these forms of reduced hours is likely to be minimal in terms of their potential employment impact, although shorter workdays are likely to be less costly for employers than fewer working days per week (see, for example, Hamermesh 1996). In the US, legislation has just been enacted that funds the 24 US states (plus the District of Columbia) that already include STC paid to worksharing participants as part of their state’s UI programmes to expand these programmes, and also incentivizes other states to adopt work-sharing/ STC programmes. The motivation behind this new law is that additional, federal funding would spread the use of work sharing with STC, and thus save even more jobs. It could also save states’ UI trust funds from being drained or having to borrow from the Federal government. In addition, a

M3125 - MESSENGER TEXT.indd 250

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

251

proposed new tax credit would allow employers to reduce employee work time, while still maintaining their pay. One attractive feature of expanding such programmes is that they would dovetail with some of the pending policy proposals in the US that also involve work hours reductions. This includes proposals for paid leave, in the form of paid sick leave and parental leave time (for Federal employees). The most comprehensive and relevant legislative proposal regarding weekly hours of work is the ‘Working Families Flexibility Act’ (H.R. 1274). The latter proposal would authorize an employee to request from his/her employer a change in the terms or conditions of the employee’s employment if the request relates to: (i) the number of hours the employee is required to work; (ii) the times when the employee is required to work; or (iii) the location where the employee is required to work. Policies that promote a more flexible workplace by better matching actual working hours with those preferred by employees can not only reduce overemployment, but also achieve job preservation, particularly in firms that are contracting. Moreover, reduced hours are bound to also serve the goal of environmental sustainability (see, for example, Rosnick and Weisbrot 2006; LaJeunesse 2009), and are likely to promote the longer-term sustainability of working humans as well. Thus, work-sharing programmes with STC could be a key subset of a potentially broader effort to more formally and permanently institutionalize reductions in working hours, in whatever form. Such reductions are currently not on the menu of options for employers, employees and policymakers in many countries, but they have the unrealized potential to improve both the quantity and quality of employment opportunities. Policy innovations such as the recent Middle Class Tax Relief and Job Creation Act of 2012 in the US, which allocated funds for existing and new states to adopt, implement, and expand work-sharing STC in their UI systems, are important first steps towards this ultimate goal. In summary, this chapter has demonstrated the potential power of institutionalizing work sharing, whether temporary or permanent, to protect employment levels in recessions or perhaps feed employment growth during recoveries. From the empirical analyses and simulations generated herein, using a four-equation VAR model, variance decompositions and IR functions, we may conclude from the empirical analysis of the US case (a relatively difficult environment given that relatively high fixed costs of employment in the US mitigate against employment creation) that during economic downturns and slumps, a ‘realistic case’ scenario is that some additional employment can be gained if a reduction in the length of weekly working hours can be induced in conjunction with declines in output in the durables goods manufacturing sector, but not the non-durables sector. Given the evidence that there is some potential substitutability

M3125 - MESSENGER TEXT.indd 251

14/05/2013 13:33

252

Work sharing during the Great Recession

of employment for hours, caution is warranted because the complementarity between hours and employment is strong. Thus, a crucial component for ensuring that crisis work-sharing measures preserve jobs is the provision of some type of partial income replacement, such as STC, to prevent (or lessen) the macroeconomic repercussions of underemployment. Institutionalizing built-in stabilizers to absorb a downturn should be put formally in place before a ‘crisis’ starts, rather than being installed during or after it. The results suggest that such measures would appear to be a powerful tool to lessen the severity of employment reductions during recessions and thus ought to appear on the menu of crisis-response measures, particularly in countries currently lacking formal programmes that promote adjustment through hours reductions. It cannot be discerned from the data analysed in this chapter what form of working-time reduction – a shorter workweek, more opportunities for quality part-time work, limits on overtime or types of reduced annual work hours – would be most likely to deliver employment gains. None the less, it may be generalized from the results that it matters little what form such weekly hours reductions take, only that they should be accompanied by partial, rather than either no or full, wage replacement. It is especially important that any reduction in working hours be accompanied by a partial restoration of earnings – so as to prevent reinforcing a macroeconomic contraction, while also not raising employers’ unit labour costs. Based on the case analysis of the US situation in the manufacturing sector, slight gains in employment seem possible with reduced hours during contractionary periods in the durable goods manufacturing industry and somewhat during expansionary periods in non-durables manufacturing (this chapter did not have sufficient data to simulate the potential effects of work sharing in the service-producing sectors). The greatest potential gain in well-being for workers would occur if work-sharing measures were targeted towards the sectors where employment gains are most realizable, and also where hours reductions are mainly preference based, in those industries where involuntarily long work hours are the highest.

NOTES 1. As discussed earlier, there are a range of possible methods to induce reductions in actual hours of work, many of which do not involve a reduction in standard hours and/or do not apply across all sectors (for example, tax incentives to reduce hours of work in specific industries). 2. This abstracts from the bi-directional causality issue that productivity gains sustained over time may lead to reduced hours via the ‘net income effect’ on labour supply, for example, workers’ preference for shorter hours (see Rubin and Richardson 1997).

M3125 - MESSENGER TEXT.indd 252

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

253

3. Vector autoregression (VAR) analysis establishes the dynamic interrelationship through time in a system using monthly time series of related variables. Each equation in an ‘ordered VAR system’ is then estimated recursively. The VAR system is composed of output (Q), average workweek (H), employment of production workers (P) and real weekly wages (W). The data are ordered in that exact sequence, to reflect the actual nature of such adjustments in labour input, after an output change. Then, the parameters of each of the estimated four VAR systems are used to construct dynamic, ‘orthogonalized impulse response’ (IR) functions. 4. Error bands representing 2 standard errors for the responses are also depicted in Figures  7.8–7.13. The error bands were constructed using Monte Carlo integration. These error bands are used to help us judge whether the IR is statistically significant or not. We observe the simulated impact of the shock to output, and hours, for the 12 months following such a one-time, one-month shock.

BIBLIOGRAPHY Abraham, K. and S. Houseman (1995), ‘Labor adjustment under different institutional structures’, in F. Buttler, W. Franz, R. Schettkat and D. Soskice (eds), Institutional Frameworks and Labor Market Performance: Comparative Views on the U.S. and German Economies, London: Routledge, pp. 285–315. Abraham, K. and S. Houseman (2009), ‘Short-time compensation is a missing social safety net for US economy in recessions’, Employment Research, 16 (3): 3. ‘American Income Life & National Income Life’ (2011), Labour Letter Canada, Labour Advisory Board, 12 (7), December: 1. Andrews, M.J. and R. Simmons (2001), ‘Friday may never be the same again: some results on work sharing from union-firm bargaining models’, Scottish Journal of Political Economy, 48 (5): 488–516. Balducchi, D.E. and S.A. Wandner (2008), ‘Work sharing policy: power sharing and stalemate in American federalism’, Journal of Federalism, 38 (1):111–36. Bauer, T. and K.F. Zimmermann (1999), ‘Overtime work and overtime compensation in Germany’, Scottish Journal of Political Economy, 46: 419–36. Best, F. (1988), Reducing Workweeks to Prevent Layoffs: The Economic and Social Impacts of Unemployment Insurance-Supported Work Sharing, Philadelphia, PA: Temple University Press. Bloemen, H.G. (2008), ‘Job search, hours restrictions, and desired hours of work’, Journal of Labor Economics, 26 (1): 137–79. Bluestone, B. and S. Rose (2000), ‘The enigma of working time trends’, in L. Golden & D.M. Figart (eds), Working Time: International Trends, Theory and Policy Perspectives, London: Routledge Press, pp. 21–37. Blyton, P. and R. Trinczek (1997), ‘Renewed interest in work sharing? Assessing recent developments in Germany’, Industrial Relations Journal, 28 (1): 3–13. Boissonat, J. (1996), ‘Combating unemployment, restructuring work: reflections on a French study’, International Labour Review, 135 (1): 5–15. Booth, A. and M. Ravallion (1993), ‘Employment and length of the working week in a unionized economy in which hours of work influence productivity’, The Economic Record, 69 (207) (December): 428–36. Booth, A. and F. Schiantarelli (1987), ‘The employment effects of a shorter working week’, Economica, 54 (214): 237–48.

M3125 - MESSENGER TEXT.indd 253

14/05/2013 13:33

254

Work sharing during the Great Recession

Bosch, G., P. Dawkins and F. Michon (1994), Times Are Changing: Working Time in 14 Industrialised Countries, Geneva: International Institute for Labour Studies. Bosch, G. and S. Lehndorff (2001), ‘Working-time reduction and employment: experiences in Europe and economic policy recommendations’, Cambridge Journal of Economics, 25: 209–43. Bujanda, L.F. and D. Fairris (2011), ‘Mandatory work hours reduction and work sharing in Venezuela’, Draft, Eastern Economic Association Meetings, New York, February. Burda, M.C. and J. Hunt (2011), ‘What explains the German labor market miracle in the Great Recession?’, SFB 649 Discussion Paper 2011-031, Berlin. Calavrezo, O., R. Duhautois and E. Walkoviak (2009), ‘Chômage partiel et licenciements économiques’, Connaissance de l’emploi, Centre d’Études de l’Emploi, No. 63, March. Cazes, S., S. Verick and C. Heuer (2009), ‘Employment: labour market policies in times of crisis’, Employment Working Paper No. 35, ILO, Geneva. Cette, G. and D. Taddei (1993), ‘The economic effects of reducing and reorganizing working time’, Futures, June: 561–77. Conference Board in Europe (1981), ‘Working Time Reductions and Unemployment: A Socio-Economic Assessment’, Brussels. Conrad, K., H. Koschel and A. Löschel (2008), ‘Extending working hours: why not work 42 hours rather than 38? A CGE analysis for Germany’, Empirica, 35 (3) (July): 255–66. Cuvillier, R. (1984), The Reduction of Working Time, Geneva: ILO. Davoine, L. and D. Méda (2009), ‘Work more to earn more? The mixed feelings of Europeans’, International Labour Review, 148 (1–2): 15–46. Dembe, A.E., R. Delbos and J.B. Erickson (2008), ‘The effect of occupation and industry on the injury risks from demanding work schedules’, Journal of Occupational and Environmental Medicine, 50 (10) (October): 1185–94. Domingueza, E., M. Ullíbarria and I. Zabaletaa (2011), ‘Reduction of working hours as a policy of work sharing in the face of an economic crisis’, Applied Economics Letters, 18 (7): 683–86. Donner Commission (1995), ‘Shorter Hours, More Jobs?’, Worklife Report, Ottawa. Drèze, J., T. Persson and M. Miller (1986), ‘Work-sharing: some theory and recent European experience’, Economic Policy, 1 (3) (October): 562–619. Drolet, M. and R. Morissette (1997), ‘Working more? Less? What do workers prefer?’, Perspectives on Labour and Income, 9 (4): 32–8. Ehrenberg, R. and P. Schumann (1982), Longer Hours or More Jobs?, Ithaca, NY: ILR Cornell University Press. Erbas¸ S., N. Chera and L. Sayers (2001), ‘Can a shorter workweek induce higher employment? Mandatory reductions in the workweek and employment subsidies’, International Tax and Public Finance, 8: 485–509. European Industrial Relations Review (EIRR) (1997), ‘Emphasis on working time across the EU’, no. 278 (March): 1–13. Facer, R.L. and L.L. Wadsworth (2010), ‘Four-day work weeks: current research and practice’, Connecticut Law Review, Symposium: Redefining Work: Implications of the Four-Day Work Week, 42 (May): 1031–46. Facer, R.L. and L.L. Wadsworth and C.A. Arbon (2010), ‘Alternative work schedules in local government: cui bono?’, Review of Public Personnel Administration, 30 (3) (September): 322–40.

M3125 - MESSENGER TEXT.indd 254

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

255

Fitzgerald, T. (1998), ‘Work schedules, wages and employment in a general equilibrium model with team production’, Review of Economic Dynamics, 1 (4) (October): 809–34. Freeman, R. (1998), ‘Work sharing to full employment: serious option or populist fallacy?’, in Freeman and P. Gottschalk (eds), Generating Jobs, New York: Russell Sage, pp. 195–222. Fromont, M. (1993), ‘Negotiating flexibility to save jobs’, World of Work, No. 5, (September): 9–11. Fuchs, V. and J. Jacobsen (1991), ‘Employee response to compulsory short-time work’, Industrial Relations, 30 (3) (Fall): 501–13. Gash, V., A. Mertens and L.R. Gordo (2012), ‘The influence of changing hours of work on women’s life satisfaction’, The Manchester School, 80 (1) (January): 51–74. Glosser, S. and L. Golden (1997), ‘Average work hours as a leading economic variable in U.S. manufacturing industries’, International Journal of Forecasting, June: 725–46. Glosser, S. and L. Golden (2004), ‘The changing nature of hours and employment adjustment in US manufacturing: a contributing cause of the jobless recovery?’, International Journal of Manpower, 25 (7/8): 618–42. Glosser, S. and L. Golden (2005), ‘Is labour becoming more or less flexible? Changing dynamic behavior and asymmetries of labour input in US manufacturing’, Cambridge Journal of Economics, 29, 535–57. Golden, L. (1990), ‘The insensitive workweek: trends and determinants of adjustment in average hours’, Journal of Post Keynesian Economics, 13: 79–110. Golden, L. (2006), ‘Overemployment in the US: which workers are willing to reduce their work hours and income?’, in J.-Y. Boulin, M. Lallement, J.C. Messenger and F. Michon (eds), Decent Working Time: New Trends, New Issues, Geneva: ILO, pp. 209–61. Golden, L. and T. Gebreselassie (2007), ‘Overemployment mismatches in the US’, Monthly Labor Review, April: 18–37. Golden, L., J. Henly S. Lambert and B. Wiens-Tuers (2011), ‘Working time in the employment relationship: perceived control and work–life balance’, Chapter 10 in K. Townsend and A. Wilkinson (eds), The Edward Elgar Research Handbook on Work and Employment Relations, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 188–211. Gray, D. (1996), ‘Work sharing benefits in Canada: an effective employment stabilization policy measure?’, Industrial Relations–Quebec, 51 (4) (Fall): 802–22. Hamermesh, D. (1996), Workdays, Work Hours and Work Schedules: Evidence for the United States and Germany, Kalamazoo, MI: W.E. Upjohn Institute for Employment Research. Hart, R. (1987), Working Time and Employment, Boston, MA: Allen & Unwin. Hart, R. and N. Wilson (1988), ‘The demand for workers and hours: micro evidence  from the U.K. metal industry’, in Hart (Ed.), Employment, Unemployment, and Labor Utilization, Boston, MA: Unwin Hyman. Hayden, A. (2006), France’s 35-hour week: attack on business? Win–win reform? Or betrayal of disadvantaged workers? Politics & Society, 34: 503–42. Hill, S. (1987), ‘Working time changes and employment growth’, Lloyd’s Bank Review, no. 163 (January): 31–46. Houseman, S. (2011), ‘Labor market flexibility: a view from the United States’,

M3125 - MESSENGER TEXT.indd 255

14/05/2013 13:33

256

Work sharing during the Great Recession

W.E. Upjohn Institute for Employment Research Working Paper, Presented at conference, ‘Increasing Labor Market Flexibility: Boon or Bane?’, Institute for Employment Research, Nuremberg, 18–19 March. Huberman, M. (1997), ‘An economic and business history of worksharing: the Bell Canada and Volkswagen experiences’, Business and Economic History, 26 (Winter): 404–12. Huberman, M. and C. Minns (2007), ‘The times they are not changin’: days and hours of work in old and new worlds, 1870–2000’, Explorations in Economic History, 44 (4), 538–67. Human Resources and Skills Development Canada (2004), ‘Evaluation of the Work Sharing Program’, Ottawa, February. Hunnicutt, B.K. (1992), ‘Kellogg’s six-hour day: a capitalist vision of liberation through managed work reduction’, The Business History Review, 66 (3) (Autumn): 475–522. Hunt, J. (1999), ‘Has work sharing worked in Germany?’, Quarterly Journal of Economics, 114 (1): 117–48. International Labour Review (1993), ‘France: company work sharing agreements’, International Labour Review, 132 (3): 293–8. International Labour Review (1995), ‘Perspectives: working time and employment–new arrangements’, International Labour Review, 134 (2): 259–72. Jacobson, T. and H. Ohlsson (2000), ‘Working time, employment, and work sharing: evidence from Sweden’, Empirical Economics, 25 (1): 169–87 Jacoby, S. (1985), Employing Bureaucracy, New York: Columbia University Press. Kapteyn, A., A. Kalwij and A. Zaidi (2004), ‘The myth of worksharing’, Labour Economics, 11 (3) (June): 293–313. Kerachsky, S. and others (1985), An Evaluation of Short-Time Compensation Programs: Summary Report, Princeton, NJ: Mathematica Policy Research. Kerachsky, S. et al. (1986), ‘Work sharing programs: an analysis of their use’, Monthly Labor Review, May: 31–3. Konczal, M. (2010), ‘The stagnating labor market’, Roosevelt Institute Working Paper, New York, 22 October. Kroll, S. (2011), ‘The decline in work hours during the 2007–09 recession’, Monthly Labor Review, 134 (4) (April): 53–9. LaJeunesse, R. (2009), Working Time Regulation as Sustainable Full Employment Policy: The Social Effort Bargain, Series: Routledge Frontiers of Political Economy, London: Routledge. Lanoie, P., A. Béjaoui Bouiet and F. Raymond (2000), ‘Qui veut réduire ses heures de travail? Le profil des travailleurs adhérent à un programme de partage de l’emploi’, Relations industrielles / Industrial Relations, 55 (3): 505–23. Lee, S., D. McCann and J.C. Messenger (2007), Working Time around the World: Trends in Working Hours, Laws and Policies in a Global Comparative Perspective, London and New York: Routledge; Geneva: ILO. Logeay C. and S. Schreiber (2006), ‘Testing the effectiveness of the French work sharing reform: a forecasting approach’, Applied Economics, 38 (17): 2053–68. MaCurdy T., J. Pearce and R. Kihlthau (2004), ‘An alternative to layoffs: work sharing unemployment insurance’, California Policy Review, August: 1–11. Marimon, R. and F. Zilibotti (2000), ‘Employment and distributional effects of restricting working time’, European Economic Review, 44: 1291–326. McCarthy, M. and G. Rosenberg (1981), Work sharing Case Studies, Kalamazoo, MI: W.E. Upjohn Institute for Employment Research.

M3125 - MESSENGER TEXT.indd 256

14/05/2013 13:33

Work sharing as a potential policy tool for creating employment

257

Meltz, N., F. Reid and G. Swartz (1981), Sharing the Work: An Analysis of the Issues in Worksharing and Job Sharing, Toronto: University of Toronto Press. Messenger, J.C. (2009), ‘Work sharing: a strategy to preserve jobs during the global jobs crisis’, Policy Brief No. 1, Conditions of Work and Employment Programme, ILO, Geneva, June. Messenger, J.C. and S. Rodríguez (2010), ‘New developments in work-sharing in middle-income countries’, Policy Brief, No. 2, Conditions of Work and Employment Programme, ILO, Geneva, February. Needles, K. and W. Nicholson (1988), ‘Measuring the effects of STC on workforce dynamics’, in S. Houseman and G. Wong (eds), Changes in Working Time in Canada and the US, Kalanazoo, MI: W.E. Upjohn Institute for Employment Research. Nemirow, M. (1984), ‘Work-sharing approaches: past and present’, Monthly Labor Review, 107 (9) (September): 34–40. Niefer-Dichmann, E. (1991), ‘Working time reductions in the former federal republic of Germany: a dead-end for employment policy’, International Labor Review, 130 (4): 511–22. Nyland, C. (1989), Reduced Worktime and the Management of Production, Cambridge: Cambridge University Press. OECD (2010a), Employment Outlook 2010 – How does JAPAN compare?, Paris: Organization for Economic Cooperation and Development. OECD (2010b), Employment Outlook 2010: Moving Beyond the Job Crisis, Paris: Organization for Economic Cooperation and Development. Osuna, V. and J.V. Ríos-Rul (2003), ‘Implementing the 35 hour workweek by means of overtime taxation’, Review of Economic Dynamics, 6 (1) (January): 179–206. Otterbach, S. (2009), ‘Mismatches between actual and preferred work time: empirical evidence of hours constraints in 21 countries’, Journal of Consumer Policy, 33 (2): 143–61. Owen, J.D. (1989), Reduced Working Hours: Cure for Unemployment or Economic Burden?, Baltimore, MD: Johns Hopkins University Press. Public Agenda (2009), ‘With Their Whole Lives Ahead of Them: Myths and Realities About Why So Many Students Fail to Finish College’, Report to the Bill and Melinda Gates Foundation. Raposo, P.S. and J.C. van Ours (2010), ‘How working time reduction affects jobs and wages’, Economics Letters, 106 (1) (January): 61–63. Rebitzer, J. and L. Taylor (1995), ‘Do labor markets provide enough short hour jobs? An economic analysis of work hours and work incentives’, Economic Inquiry, 33 (2) (April): 257–74. Roche, W., B. Fynes and T. Morrisey (1996), ‘Working time and employment: a  review of international evidence’, International Labour Review, 135 (2): 129–57. Rosnick, D. and M. Weisbrot (2006), ‘Are shorter work hours good for the environment? A comparison of U.S. and European energy consumption’, CEPR report, December. Rubin, M. and R. Richardson (1997), The Microeconomics of the Shorter Workweek, Aldershot, UK: Avebury, Ashgate. Schreiber, S. (2008), ‘Did work-sharing work in France? Evidence from a structural co-integrated VAR model’, European Journal of Political Economy, 24 (2) (June): 478–90.

M3125 - MESSENGER TEXT.indd 257

14/05/2013 13:33

258

Work sharing during the Great Recession

Shelton, A.M. (2011), ‘Compensated work sharing arrangements (short-time compensation) as an alternative to layoffs’, Congressional Research Service, 7-5700, 15 February. Siedule, T., C. Guest and G. Wong (1996), ‘Economic activities and the demand for work sharing in Canada’, in Wong and G. Picot (eds), Working Time in Comparative Perspective, Vol. I, Kalamazoo, MI: W.E. Upjohn Institute. Skuterud, M. (2007), ‘Identifying the potential of work sharing as a job-creation strategy’, Journal of Labor Economics, 25 (2): 265–87. Snower, D. (1995), ‘Evaluating unemployment policies: what the underlying theories tell us?’, Oxford Review of Economic Policy, 11 (1): 110–35. Sousa-Poza, A. and F. Henneberger (2002), ‘An empirical analysis of workinghours constraints in twenty-one countries’, Review of Social Economy, 60 (2) (June): 209–42. Taddei, D. (1991), ‘Social Europe: Working Time, Employment and Production Capacity’, April Supplement, Directorate-General, Commission of the European Communities, Luxembourg. Taylor, J.E. (2011), ‘Work sharing during the great depression: did the “President’s reemployment agreement” promote reemployment?’, Economica, 78: 133–58. US Bureau of Labor Statistics (US BLS) (2010), The Employment Situation – March 2010, BLS, Washington, DC, USDL-10-0394. van Audenrode, S. (1994), ‘Short-time compensation, job security and employment contracts: evidence from selected OECD countries’, Journal of Political Economy, 102 (1): 76–102. van Ginnekan, W. (1984), ‘Employment and the reduction of the work week: a comparison of seven European macroeconomic models’, International Labour Review, 123: 35–52. Vroman, W. and V. Brusentsev (2009), ‘Short-Time Compensation as a Policy to Stabilize Employment’, The Urban Institute, Washington, DC, November. White, M. (1987), ‘Working Hours: Assessing the Potential for Reduction’, ILO, Geneva. Whitely, J. and R. Wilson, (1986), ‘The impact on employment of a reduction in the length of the working week’, Cambridge Journal of Economics, 10: 43–59. Williams, B. (1984), ‘Shorter hours – increased jobs?’, Three Banks Review, September: 3–16. Wooden, M., D. Warren and R.W. Drago (2009), ‘Working time mismatch and subjective well-being’, British Journal of Industrial Relations, 47 (1) (March): 147–79. Zachmann, R. (1986), ‘Reduction of working time as a means to reduce unemployment: a microeconomic perspective’, International Labour Review, 125 (2) (March/April): 163–75. Zandi, M. (2009), Written Testimony before the Joint Economic Committee, The Impact of the Recovery Act on Economic Growth, 29 October, Washington, DC.

M3125 - MESSENGER TEXT.indd 258

14/05/2013 13:33

8.

Conclusion: Lessons learned from the Great Recession and implications for policy Jon C. Messenger and Naj Ghosheh

1.

INTRODUCTION

As we have seen throughout this volume, work-sharing programmes and measures played a critical role in the crisis-response strategies of many countries around the world. Although the programme designs, target populations and eligibility criteria varied considerably from country to country, the overall objective of all of these measures was the same: to preserve jobs and maintain firms which were suffering from the severe, financial-crisis-induced economic downturn that has come to be known as the ‘Great Recession’. Indeed, those countries which had work-sharing programmes that pre-dated the crisis almost uniformly expanded those measures by relaxing eligibility requirements; increasing the wage supplements for affected workers and/or subsidies for eligible firms; and extending the duration of the measures – often for an entire year or even longer. In addition, a number of countries that had never before used work sharing developed and implemented new programmes, particularly in Eastern Europe and Latin America. Thus, the Great Recession of 2008–09 and its immediate aftermath provide a wealth of experiences from which we can draw some lessons for future economic crises regarding both the optimal design of crisis work-sharing programmes as job-preservation measures and their potential effects on maintaining employment levels. In this chapter, rather than simply summarizing the individual contributions in this volume, we shall instead synthesize the work-sharing experiences that were presented and attempt to answer the question: what is the ideal programme design for a crisis work-sharing programme to preserve jobs during an economic downturn? The experience of the Great Recession also raises a larger question: given the substantial overall success of working-time reductions in maintaining employment during an economic downturn, does the available 259

M3125 - MESSENGER TEXT.indd 259

14/05/2013 13:33

260

Work sharing during the Great Recession

evidence suggest that work sharing can be something more than a measure which is solely intended for crisis response? Can work sharing possibly also serve as an effective tool for creating more, and perhaps even better jobs, in a more positive economic environment? In other words, to what extent can permanent reductions in work hours (for example, reduced standard workweeks, increased periods of paid leave, targeted hours reductions in specific industries, voluntary, preference-based reductions in working hours and so on) help to create employment while improving individual well-being, and thus contribute to more sustainable economies and societies? This chapter will consider those questions as well. The remainder of the chapter will seek to examine and explain three different issues. Using the information from the previous chapters, Section 2 will summarize what crisis work-sharing policies have been used in different countries during the Great Recession. To do this it will examine the experiences of the countries in key elements of crisis work-sharing policies, namely: the targeting and eligibility requirements for companies and workers; adjustments in working time; provisions regarding wage supplements for work-sharing participants; time limitations on the period of work sharing; links between work-sharing policies and training; and the programme participation and employment effects of work-sharing policies. Section 3 will examine the implications of this experience for work-sharing policies, including how they can be ideally designed for employment preservation during economic crises and potentially also be used to promote job creation as well. Section 4 concludes with recommendations and observations regarding the prospects of work sharing as a tool in the employment policy arsenal in the future.

2.

THE USE OF WORK SHARING DURING THE GREAT RECESSION: SUMMARY AND LESSONS LEARNED

2.1

The Design of Crisis Work-sharing Policies and Programmes to Preserve Jobs

Crisis work-sharing policies and programmes are designed so that the burdens of a difficult economic situation are shared – not only among workers, but between workers, employers and governments as well. If work-sharing policies are properly designed and implemented, the result can be a ‘win–win–win’ solution: enabling workers to keep their job and even to prepare for the future; assisting companies not only to survive a crisis, but to be well positioned to prosper when growth returns; and mini-

M3125 - MESSENGER TEXT.indd 260

14/05/2013 13:33

Conclusion

261

mizing the costs of social transfer payments and, ultimately, social exclusion for governments and society as a whole. Crisis work-sharing measures not only help to avoid layoffs, buy they also allow businesses to retain their workforces, thus minimizing firing and (re)hiring costs, preserving functioning plants, and bolstering staff morale during difficult times. The question is: how can crisis work-sharing policies and programmes be designed to make them more likely to deliver such a ‘win–win–win’ solution? There are essentially five key elements that may be included in crisis work-sharing programmes and measures designed to avoid layoffs, not all of which are present in every work-sharing policy measure or programme. These key elements are as follows: ● ●

● ●

●

programme targeting and eligibility criteria for both companies and workers; the reduction of working hours for all eligible workers in an eligible company or a specific work unit within an eligible company, in lieu of layoffs, and a corresponding (typically but not always pro rata) reduction in total earnings; the provision of wage supplements to affected workers to ‘cushion’ the effects of temporary reductions in earnings; the establishment of specific time limits on the period of crisis work sharing to ensure that the programme is a temporary measure in response to an economic crisis; and the creation of links between work-sharing programmes and training activities.

In addition to these elements, engaging employers’ associations and trade unions in the design and implementation of government crisis worksharing programmes and measures is common and can increase their likelihood of success. We shall now review each of these elements in turn, considering how they were applied in the countries analysed in this volume, and the lessons that we can learn from these experiences. 2.2

Programme Targeting and Eligibility Criteria for Companies and Workers

The criteria used for eligibility to crisis work-sharing measures are not universal. Countries have established different criteria for eligibility to participate in such work-sharing programmes, as shown in Appendix Table 8A.1. Each crisis work-sharing measure has a set of eligibility requirements that may be more or less restrictive depending on how the programme is

M3125 - MESSENGER TEXT.indd 261

14/05/2013 13:33

262

Work sharing during the Great Recession

designed. The main common thread across countries regarding their eligibility criteria for crisis work-sharing programmes is that companies need to be able to demonstrate that they are suffering from some form of serious business difficulty due to economic conditions beyond their control that might lead them to reduce their workforces via layoffs. Different countries have established different criteria as to what constitutes economic difficulties (see Table 8A.1). A few crisis work-sharing programmes and measures address specific economic problems associated with the industry during a period of the year or a specific national circumstance, but most of them are designed to address major declines in business activity due to economic factors on a national or global scale. In this context the most developed work-sharing policies are found in Germany, where (as discussed in Chapters 1 and 2) work sharing has been used off and on for nearly a century to address various economic challenges. Three different types of work sharing exist. The first type of German worksharing measure is seasonal work sharing, where in industries such as gardening and construction this policy is used during unproductive times of the year due to adverse weather, for example, winter. The second form of work sharing in Germany is ‘Transfer-Kurzarbeit’, which is focused on structural economic problems, and was developed mainly for German reunification, in order to avoid widespread mass layoffs in the formerly state-owned East German industry. Finally, and most relevant for this volume, there is work sharing for economic reasons, which was developed to respond to temporary, unavoidable loss of work due to cyclical economic conditions or an unavoidable event, such as the Great Recession of 2008–09 and its aftermath. Most countries with national crisis work-sharing programmes or measures have developed them solely to respond to national or global economic conditions that threaten companies within the country or in a particular industry or occupational group (see Table 8A.1). Thus, in Belgium, for example, the only eligibility criterion is that the company must be in financial difficulties. In Austria, as in Germany, the national work-sharing measure (which is also called ‘Kurzarbeit’) requires that the establishment must be suffering from temporary economic difficulties due to factors external to the enterprise (but unlike in Germany, there is no scheme designed to respond to normal seasonal effects on businesses). The crisis work-sharing measure in France (chômage partiel) also refers to exceptional circumstances, but in addition to economic conditions these also include the provision of raw materials or energy, natural disasters, or the restructuring of a company – which in all cases must be temporary and exceptional. Crisis work-sharing measures in countries such as Japan, the Netherlands, and Uruguay establish specific baselines of economic

M3125 - MESSENGER TEXT.indd 262

14/05/2013 13:33

Conclusion

263

contraction over time periods among their eligibility criteria for firms. For example, the Employment Adjustment Subsidy (EAS) in Japan can only be applied if there is a reduction of 5 per cent or more in production or sales in the last three months compared to the previous three months or the same three-month period in the previous year. In the Netherlands, a decline of 30 per cent or more during a period of two months is required to participate in the Part-Time Unemployment work-sharing measure, and in Uruguay there must be a 15 per cent drop in company activities in the quarter prior to application for coverage (as compared with the average level or productivity in the same quarter during the previous two years) to participate in the Partial Unemployment Insurance (UI) scheme (one form of work sharing in that country). Eligibility criteria for the Reduced Working Time (RWT) programme in Turkey include a time element, in that firms must suffer from a general economic crisis caused by domestic or international developments that reduced firm activity over a period of at least four consecutive weeks. In Japan, the amount of corporate earnings lost as opposed to the percentage of business lost is important, as a reduction in sales or business activity in monetary terms is part of the firm’s eligibility requirements for the EAS. Thus, while the eligibility criteria for firms seeking to participate in crisis work-sharing measures vary among countries, most of them establish at least the following requirements: exceptional economic circumstances; a significant drop in sales or production or in financial terms; and some condition that these losses must generally be short-term economic losses in comparison with some other, recent period. Finally, some measures are limited to specific sectors or industries, such as the Partial UI scheme, which targeted the leather, textiles and garments, wood and wood products, and metalworking industries. While the focus of crisis work-sharing measures is primarily on the economic problems faced by companies, they may also include requirements regarding the groups of employees that are covered. For example, as noted before, in Germany some seasonal workers can participate in one form of work sharing available to workers in industries such as construction and gardening. In Belgium, as a result of the global financial crisis in 2008, the temporary unemployment measure (chômage temporaire) was extended from covering only blue-collar workers to include white-collar workers as well, including temporary agency workers. In Uruguay two forms of work sharing developed, both of which included provisions that used employed workers as the basis for programme eligibility. The Partial UI work-sharing scheme in Uruguay requires that the company cannot have dismissed more than 5 per cent of its personnel in the quarter prior to the application for coverage under the measure (this does not include

M3125 - MESSENGER TEXT.indd 263

14/05/2013 13:33

264

Work sharing during the Great Recession

dismissals based on misconduct), whereas the suspension provision of the UI programme relies on a temporary suspension of work activities for different groups of workers – with different groups of workers rotating rapidly into and out of such temporary suspensions every few weeks – or alternatively, on a reduction in the number of work days or working hours. A notable variation on this theme can be found in the Netherlands, which explicitly takes into account the number of employees that will be taking part in work sharing. The newer Dutch work-sharing measure, the Part-Time Unemployment scheme (Deeltijd-Werkloosheidswet), reduces the hours of employees by a minimum of 20 per cent and a maximum of 50 per cent for a period of at least 26 weeks. The maximum duration of part-time unemployment is dependent upon the number of employees for which it was requested. If the share of employees exceeds 60 per cent, the maximum duration is 39 weeks; if the share is between 30 and 60 per cent, the maximum duration is 52 weeks; and if the share of employees is less than 30 per cent, then the maximum duration is 65 weeks. Perhaps the most important criterion affecting firms’ participation in crisis work-sharing measures is the length of the period required to qualify for work-sharing assistance in a particular country. Some countries have rapid approval processes, while others have established quite timeconsuming reviews. For example, countries such as Austria and Japan have quite lengthy approval mechanisms, with the Austrian Employment Service requiring six weeks’ notification to initiate the programme, while a period of approximately three months is needed in Japan before the subsidies from the EAS are paid. By contrast, in Belgium the review period is very quick, as the employer is only required to notify the public employment agency seven days in advance of the use of temporary unemployment (chômage temporaire) and no formal approval is needed. Such a feature may have an impact on firms’ decision to participate, and in this regard it is worth noting that Belgium’s various work-sharing measures, taken together, reach a larger proportion of the labour force – 5.6 per cent – than those in any other country (Flecker and Schönauer, ch. 3 in this volume; see also OECD 2010). What becomes clear from a review of these various crisis work-sharing measures is that there are a number of ways to structure eligibility for crisis work-sharing support. Declines in a firm’s revenues over a period of time, the types of industries and workers covered, and the levels of employee participation in work sharing within firms are all mechanisms that have been used as eligibility criteria for these programmes. While these criteria are all important, perhaps the most critical one is how quickly a company can obtain approval to participate in a work-sharing programme. Also, during crisis periods the number of firms looking to participate will likely

M3125 - MESSENGER TEXT.indd 264

14/05/2013 13:33

Conclusion

265

be much higher and demand to access the programme will grow – often dramatically. While lengthier programme eligibility mechanisms can act as something of a filter, if properly established, they can also permit rapid decision-making by government officials for firms seeking access to worksharing programmes. 2.3

Adjustments in Working Time: The Extent of Reduced Hours (and Earnings)

Adjustments to working hours and earnings are at the heart of crisis work-sharing policies, since the whole point of these programmes is to allow companies to reduce labour costs in response to economic crises without resorting to layoffs. The Great Recession hit many establishments quite hard, and as we have seen, the development and implementation of work-sharing measures have been important in preserving employment in many countries. However, as noted earlier, the manner in which these programmes have developed has been different across countries, and this is also true with regard to reductions in working hours and corresponding (typically, but not always, pro rata) reductions in total earnings, as well as the wage subsidies offered by governments to affected workers (which will be discussed in the next subsection). With regard to working hours, the work-sharing policy environment in Germany provides some context for the policies found not only in that country, but in many others as well. In Germany not only the policy of Kurzarbeit, but also an establishment’s ability to reduce overtime hours of work, was instrumental in the adjustment of actual working hours to reduced demand. In addition, workers’ working-time accounts (which include hours accumulated for overtime work) needed to be reduced to zero in order to qualify for Kurzarbeit. Reductions in the amount of overtime were also important to sharing work in Japan, and such reductions actually saved more jobs than the ‘official’ crisis work-sharing measure in that country, the EAS (Ogura, ch. 4 in this volume). Thus, in both of these countries, reductions in overtime also served as a type of crisis worksharing mechanism. Another important feature regarding the reduction of working hours for the application of work-sharing policies is the establishment of thresholds for the minimum and maximum permissible reductions in working hours. Many countries apply such thresholds, and in some cases they have linked the amount of compensation that they will pay to the percentage reduction in working time. Thus, in Austria companies in the Kurzarbeit programme required a reduction of working hours by at least 10 per cent and not more than 90 per cent of working time (or in the case of part-time work, of

M3125 - MESSENGER TEXT.indd 265

14/05/2013 13:33

266

Work sharing during the Great Recession

defined working time). In Belgium, the reduction in working time must be between one-fifth and one-half for a minimum period of one month and a maximum of six months to qualify for one of the work-sharing measures in the country (temporary collective reduction of working time), but in another scheme (temporary unemployment), workers alternate between days or weeks of work and days or weeks of unemployment, limited to four continuous weeks, but which can be repeated after one week of work. Similarly, the Part-Time Unemployment scheme in the Netherlands reduces the hours of employees by a minimum of 20 per cent and a maximum of 50 per cent for at least 26 weeks for firms in the Netherlands; in the Turkish crisis work-sharing programme (RWT) firms must reduce weekly working hours by at least one-third over a period of at least four weeks continuously and for three months at the most. The extent of flexibility in working hours is a big difference between the two work-sharing measures in Uruguay. One of these measures (Seguro de desempleo por causal suspension, the temporary suspension provision of the UI) allows employers and workers some flexibility by requiring a reduction of either working hours or work days, but the other measure (Seguro de desempleo parcial, the Partial Unemployment scheme) requires a reduction of at least one full and no more than two full working days per workweek for each worker involved – with zero flexibility in the number of working hours. In fact, this difference appears to be one important reason why the former work-sharing measure was used so much more extensively than the latter (González Fernández, ch. 6 in this volume). In short, crisis work-sharing programmes and measures are meant to help ‘bridge’ companies and workers to better economic times. As such, it is imperative that such programmes provide companies with a maximum degree of flexibility to adjust working hours in response to changing economic conditions – albeit for a limited period. 2.4

The Provision of Wage Supplements to Affected Workers: Who, How Much, and from What Sources?

With the reduction in working hours there is likely to be a corresponding (pro rata) reduction in workers’ earnings, and because of this situation the work-sharing programmes in many countries have established financial compensation to partially offset this reduction in earnings and thus raise workers’ total income under work sharing. Indeed, unlike many work-sharing measures during earlier periods of history (see the brief history of  work sharing in Chapter 1), wage supplements are central to today’s crisis work-sharing programmes and measures. As wage supplements in crisis work-sharing programmes are tied to the prevention of

M3125 - MESSENGER TEXT.indd 266

14/05/2013 13:33

Conclusion

267

unemployment, in many cases it is UI that pays these wage supplements. It is also the reason why the calculation of benefit amounts for many of these wage supplements goes beyond just the worker, and also takes into account whether the worker is the family breadwinner and the number of dependants that person may be supporting. In addition, some measures also take into account whether a worker is participating in some form of training in determining the amount of this payment. As noted above, crisis work sharing is a policy that stems from trying to protect against unemployment, so it is hardly surprising that most countries use UI calculations to determine any wage-related compensation. For example, this situation is the case in the Kurzarbeit system in Germany. The crisis work-sharing programmes in Austria, France and Turkey provide wage supplements based on: the monthly gross wage plus social contributions prior to work sharing in the Austrian case; nominal wages for hours not worked in the French case; and gross monthly wages in the Turkish case. The UI Fund in Turkey pays 60 per cent of the gross monthly wages and cannot exceed 150 per cent of the gross legal minimum wage rate. The French system of ‘partial unemployment’ pays 75 per cent of normal wages for hours not worked. Both the Canadian Work Sharing Program and the various state-level work-sharing programmes operated by states in the US also use unemployment insurance criteria for determining eligibility for and the amount of ‘short-time compensation’ that is payable to participating workers. However, in Japan the EAS sets the daily layoff allowance at 60 per cent or more of the daily salary, which is calculated on the basis of calendar (not work) days and paid at a higher rate for small- and medium-sized enterprises (SMEs) than larger ones. While most countries include wage supplements as part of their worksharing programmes, in some countries whether or not the worker is the principal breadwinner for a family is (as noted above) an important consideration as well. The German Kurzarbeit system maintains a ShortTime Compensation Fund, which compensates job holders who lose at least 10 per cent of their gross monthly earnings due to the reduced hours. A German worker with no children receives 60 per cent of the net difference in the regular wage, but 67 per cent if he/she has a dependent child. Comparably, in Austria a flat hourly rate used to calculate UI is also used to calculate the work-sharing payment, which factors in the monthly gross wage plus social contributions prior to work sharing and the number of children in the worker’s family. In Belgium, the amount of payment for temporary unemployment has been set at 70 per cent of the reduced wage if the worker is cohabiting, or 75 per cent of the wage if the worker is the main breadwinner (limited to a maximum of €2,206 per month). In Uruguay the UI scheme, which forms the basis of both

M3125 - MESSENGER TEXT.indd 267

14/05/2013 13:33

268

Work sharing during the Great Recession

of the work-sharing programmes in that country, pays 50 per cent of the nominal wage to a worker for six months, but 60 per cent to workers with dependants. The crisis work-sharing measures in Uruguay are notable in that in one of them, workers progressively received less compensation over time, and in the other, training was incorporated as an integral part of the programme. The temporary suspension provision of UI gradually reduces the amount of benefits available to a worker during the six-month period from 66 per cent of the nominal wage earned in the first month to 50 per cent of that wage by the end of the period, and the total amount of compensation is also capped, in order to encourage a speedy return to work. By contrast, the Partial UI scheme extends the period of payment from UI for up to one year, but also requires training provided by the National Employment and Vocational Training Institute for workers who are participating in this scheme. The relationship between crisis work-sharing measures and training programmes also influences the amount of wage supplements available to the worker. The Austrian work-sharing programme begins with a flat hourly rate, but adds 15 per cent to cover additional expenses due to training measures. Partial unemployment (chômage partiel) in France that is linked with training leave is remunerated at 80 per cent of the net salary, which is a bit higher than the normal partial unemployment allowance without training. For workers in Japan requiring occupational training during the reduction of production operations, payments went up from 1,200 yen per day to 4,000 yen per day for large firms and 6,000 yen per day for SMEs. In the Netherlands, Part-Time Unemployment requires Dutch workers to participate in training, for which employers can claim special financial assistance. Wage supplement payments for work-sharing participants are usually made from social insurance funds relating to unemployment, but they may be administered by different government agencies. For example, the German state (Länder) governments deal with UI and requests to participate in work-sharing programmes, both regarding work sharing for economic reasons and other forms of Kurzarbeit as well. National governments, through their social insurance systems, pay the costs of worksharing compensation in Austria, France and the Netherlands. However, in the Netherlands these social insurance benefits can be deducted from workers’ unemployment benefit entitlements – which is a point of contention with the Dutch trade unions. In Turkey, the Turkish Employment Agency (ISKUR) administers the crisis work-sharing programme, RWT, which is paid through social insurance. Social insurance also pays for UI in Uruguay, which is based on social security contributions as well as tax

M3125 - MESSENGER TEXT.indd 268

14/05/2013 13:33

Conclusion

269

and general revenue allocations. Japan is a notable exception to the rule for funding crisis work-sharing payments, as the EAS is paid solely from general government revenues. Government-provided wage supplements are thus an important feature of work sharing today. Ensuring that social insurance is properly funded is therefore a critical consideration in the development and implementation of work-sharing policies. Also, as many crisis work-sharing programmes and measures are linked with UI, it is notable that dependants become an element in the calculations of the amounts available to workers whose firms participate in work-sharing programmes. Owing to the fact that such work-sharing policies have been implemented explicitly to address economic crises, more streamlined procedures to link work sharing with social insurance, and thus help speed payments to workers and their dependants, merit consideration by policymakers. 2.5

Time Limits and Other Programme Features to Minimize Deadweight and Displacement Effects

Crisis work-sharing measures are often developed with some features designed to minimize deadweight and displacement effects. In this context (as discussed in Chapter 1), the term ‘deadweight’ refers to the amount of public work-sharing subsidies received by firms which would not have engaged in layoffs in any case, and the term ‘displacement effects’ refers to the crowding out of emerging businesses by existing inefficient ones which could potentially result from public subsidies for work sharing (see Chapter 1 for further information regarding these issues). This subsection will highlight some of mechanisms found in crisis work-sharing policies to limit these undesirable results. One of most common features to limit the potential deadweight and displacement effects of crisis work-sharing measures is a simple time limit regarding how long these measures can remain in effect. As a result of the global economic crisis which emerged in 2008, most countries which either had an existing crisis work-sharing programme or introduced a new one extended their originally established length from six months to one year or more, in an attempt to help companies ride out the economic storm. For example, prior to the onset of the Great Recession, Austria and France had crisis work-sharing programmes lasting six months, but in 2009 the duration of the Austrian programme was extended to one year and the French programme was extended to a maximum of 18 months. The German Kurzarbeit programme has had one of the longest durations, and its length has varied throughout the crisis years. The maximum duration for the German work-sharing programme used to be 12 months; it

M3125 - MESSENGER TEXT.indd 269

14/05/2013 13:33

270

Work sharing during the Great Recession

was then extended to 18 months and then to 24 months in 2009; and has since been reduced, first to 18 months and then back down to 12 months. The Dutch Part-Time Unemployment measure is perhaps unique in this context, as the duration of the scheme is dependent upon the proportion of workers in a firm to which the scheme will apply. As noted earlier, this duration can range from 39 weeks if 60 per cent of staff in a company are included to 65 weeks if less than 30 per cent of staff are involved. In order to ensure that companies refrain from abusing the subsidies offered by crisis work-sharing programmes, many of the programmes include provisions that are explicit in only applying to demonstrable consequences of severe economic strain, such as the global economic crisis which began in 2008. For example, the Austrian Kurzarbeit programme requires that the organization must be suffering from temporary economic difficulties due to factors external to the enterprise, but not seasonal effects or normal variations such as changes in weather that routinely affect certain sectors of the economy (for example, construction). The Belgian chômage temporaire (individual reduction of working time) measure requires a company to be in demonstrable financial difficulties, and the Netherlands, Japan, Turkey and Uruguay have similar requirements. As the main objective of work sharing in crisis situations is to save jobs, many of these work-sharing programmes also have features that could be referred to as ‘employment-related’ restrictions. Such restrictions attempt to prevent firms from taking the financial benefits and subsidies offered by these programmes and then laying off the affected workers anyway. For example, the EAS in Japan protects the employment status of workers by prohibiting layoffs and restricting the transfer of workers from a parent company to a subsidiary during the work-sharing subsidy period. Partial UI in Uruguay requires that the company cannot have dismissed more than 5 per cent of its personnel in the quarter prior to the application of the coverage of this form of insurance. And the old Dutch Short-Time Work programme (Werktijdverkorting) also included a provision in which employees were guaranteed four weeks’ protection against dismissal following the short-time work period. What is important to remember is that when economic crises take place, many ‘losses’ will result. Firms lose profitability, and may even go bankrupt; workers lose hours and wages, and perhaps even their jobs; and governments lose tax revenues while they need to increase unemployment expenditures, and perhaps even lose their political support as well. What crisis work-sharing programmes attempt to do is to limit the amount of damage that any one group or individual will suffer during the economic ‘tsunami’ that may come from international or national sources beyond the control of the firm. These programmes are not meant to be infinite in

M3125 - MESSENGER TEXT.indd 270

14/05/2013 13:33

Conclusion

271

the amount of time that they apply, nor are most of them meant to sustain companies which had problems before any economic crisis. The main aim of these subsidies is to ensure company survival for economically viable firms and employment for their workers during economic crises. Thus, crisis work-sharing subsidies are not meant to go on in perpetuity, but merely to create a bridge for firms and workers from times of economic crisis until a recovery; hence time limits are essential for such programmes to function properly. 2.6

Links with Training: The Use (and Non-use) of Training and Retraining for Work-sharing Participants

Much as with many UI programmes, efforts were made in some countries to link their work-sharing programmes with training and retraining opportunities for participating workers during their time off work. The idea underlying this approach was that such training would be an opportunity to use the time off work to improve workers’ skill sets and thus make them more productive and hence more employable. In most countries, crisis work-sharing programme links with training typically included extra payments to workers (for example, for trainingrelated expenses) and sometimes special funds that companies could tap to pay the costs of training, although in a few countries training was an integral part of the crisis work-sharing programme. In Austria, the training supplement was 15 per cent of the flat hourly rate for additional expenses. However, in the Netherlands, the Short-Time Work and PartTime Unemployment schemes required workers to be involved in training, and Dutch employers could then claim special financial assistance to pay for this training. Germany similarly provided special funds for further training while a company was engaging in work sharing. The Japanese EAS can include subsidies for vocational training, although in practice these funds were little used for this purpose during the Great Recession. In Uruguay the Partial UI scheme allows worker training financed from the Job Retraining Fund, on the condition that there is a collective agreement covering these workers, with the training provided by the National Employment and Vocational Training Institute. Unfortunately, the linkage between work-sharing programmes and training/retraining has not always been a successful match, as demonstrated by the very low rates of training participation in nearly all of the countries covered in this volume, other than those schemes in which training participation was mandatory (most notably, Part-Time Unemployment in the Netherlands). Even in Germany, only 13 to 14 per cent of establishments participating in work sharing for economic reasons

M3125 - MESSENGER TEXT.indd 271

14/05/2013 13:33

272

Work sharing during the Great Recession

were organizing training courses for work-sharing participants. Thus, the share of establishments using the reduced hours for continuing training remained under 1 per cent of all German establishments (Bellmann et al., ch. 2 in this volume). A number of different factors appear to have contributed to this rather disappointing result. These included cumbersome administrative requirements to access training funds; lack of appropriate training programmes (and qualified instructors); workers’ concerns about their ability to successfully complete training courses; unpredictability regarding the amount of time available to workers for training, when they may be required to interrupt their training to return to work if demand picked up; and also the fact that providing training courses for work-sharing participants could potentially result in workers seeking new employment outside of the company providing the training. In Austria and the Netherlands, workers’ participation in training also raised costs for companies that participated in work-sharing programmes, although they might have been able to recover some of these financial outlays (as in the Dutch case). Perhaps the core issue with regard to the training components and work-sharing policies is that they were not made compatible. National training policies are meant to train people for new jobs or how to use new technology, but these training and retraining programmes must by definition have a beginning and an ending date. Crisis work sharing, by contrast, is a policy meant to maintain employment during uncertain times, and as such it is difficult to predict how long such measures will continue and thus difficult to plan for training and retraining courses for work-sharing participants in a systematic way. This is not to suggest that training/retraining and work sharing are incompatible, but owing to the need for quick action on the part of policymakers in many countries, the compatibility issues between them (such as those discussed above) were not properly reconciled. Better integration between work sharing and training might occur if certain factors were considered. First, simpler administrative procedures for companies and workers might encourage more companies to participate in such programmes. Also, short-term or modular courses could be developed that allow workers to rotate out of work for short periods (from a few days to two weeks) and then come back to work after completing the courses. In addition, to address any concerns about workers leaving firms following training, a requirement might be placed that the worker must stay with the establishment for a limited specified period of time. Such measures may provide better short- and medium-term solutions to promote employment retention – which is, after all, the primary aim of crisis work-sharing programmes.

M3125 - MESSENGER TEXT.indd 272

14/05/2013 13:33

Conclusion

2.7

273

Programme Participation and Employment Effects of Crisis Work-sharing Programmes

The participation rates of firms and employees in crisis work-sharing programmes varied substantially among the countries that implemented such measures, which is reflected in the programmes reviewed in this book. Germany, where crisis work-sharing measures were the most highly developed, had the highest participation levels: approximately 60,000 establishments and 1.4 million workers participated in Kurzarbeit for economic reasons at the height of the crisis in May 2009 – which represents approximately 5 per cent of private sector wage employment (Bellmann et al., ch. 2 in this volume). Likewise, in Belgium, during 2009 the various temporary unemployment measures were used by more than 300,000 workers – 5.6 per cent of private wage employment (Flecker and Schönauer, ch. 3 in this volume). In France, where partial unemployment (chômage partiel) was used as a work-sharing programme, 258,000 workers were using this programme at its peak in 2009, while the Part-Time Unemployment programme in the Netherlands covered 69,170 workers during the period between April 2009 and April 2010. In contrast to Belgium and Germany, in France and the Netherlands, as well as Austria, the rate of participation in their various work-sharing measures represented less than 1 per cent of private wage employment (ibid.) Outside of Europe, countries with crisis work-sharing policies may have been limited, but none the less had substantial levels of participation. At its peak in April 2009, the EAS in Japan had 84,481 firms and 2.5 million employees participating in 2009, which is 3.8 per cent of private wage employment in that country (Ogura, ch. 4 in this volume). In Turkey, 268,000 workers were approved for participation in the RWT programme between 2008 and 2011 (or just over 1 per cent of formal employment) – making RWT the largest crisis work-sharing programme in any developing country (Yeldan, ch. 5 in this volume). In Uruguay, the number of beneficiaries under the temporary suspension provision of UI (either full or partial suspension) totalled over 30,000 in 2009; although this number may sound small, it actually represents approximately 3.5 per cent of private sector salaried employment in that small country; in contrast, participation in the newly created Partial UI programme was almost nil (González Fernández, ch. 6 in this volume). It is also useful to bear in mind that many of these countries developed new programmes, or at least adapted their policies, very quickly to respond to the onset of the Great Recession. Thus, although the participation rates were generally substantial, they might have been higher had the work-sharing programmes been better institutionalized for implementation in these exceptional economic circumstances.

M3125 - MESSENGER TEXT.indd 273

14/05/2013 13:33

274

Work sharing during the Great Recession

It is important to note that, from a gender perspective, the participants in the work-sharing programmes and measures are predominantly male. This is the case not only in Germany, but in all of those countries reviewed in this volume. The most likely explanation for this phenomenon is that the crisis in these countries mainly hit male-dominated sectors such as manufacturing and construction, whereas female-dominated sectors such as health services and education were less affected (at least initially). None the less, there is also an issue regarding the structure of some crisis worksharing measures, in that these measures are not designed to cover parttime and temporary workers – who statistically are predominantly female in most countries. Thus, in those countries in which workers on part-time or temporary contracts are eligible for work-sharing measures, such as France and the Netherlands, women were more likely to be able to take advantage of these measures. Regarding the employment effects of work-sharing programmes and measures, a number of recent studies specifically investigating the effects of these measures on job preservation during the Great Recession have concluded that these programmes do in fact reduce layoffs by increasing per capita reductions in working hours (see, for example, Arpaia et al. 2010: 40; Crimmann et al. 2010: 35; OECD 2010: 15; Hijzen and Venn 2011: 4). In addition, other types of working-time arrangements – most notably working-time accounts – clearly also contributed to the positive employment effects in countries such as Germany, since these accounts must be drawn down to zero before an enterprise is eligible to apply for Kurzarbeit (see, for example, Crimmann et al. 2010). Nearly all of the crisis work-sharing programmes and measures in the countries reviewed in this volume had positive effects on employment, as shown in Table 8.1. Although the estimates presented in this table are not strictly comparable, it is clear that the effect of work sharing on employment was much larger in some countries than in others. While the largest effect in absolute numbers was clearly in Japan and Germany, as a proportion of the labour force the employment effect was actually the largest in Belgium (Flecker and Schönauer, ch. 3 in this volume). Also, in some cases the impact of one work-sharing policy was much higher within a country than another work-sharing policy, most notably in Japan but also in Uruguay as well. A number of different factors which were considered to have influenced these results are highlighted in the remainder of this section. In spite of the fact that Germany has been considered the originator – and even the epicentre – of work-sharing policies, the employment effects of work sharing were actually the most pronounced in Japan, and as a percentage of workers (private sector wage employment only), these effects

M3125 - MESSENGER TEXT.indd 274

14/05/2013 13:33

Conclusion

Table 8.1

275

Summary of estimated employment effects of work-sharing programmes and measures reviewed in this volume

Country

Work-sharing scheme

Estimated employment effects

Germany

Short-time work (Kurzarbeit)

Austria

Short-time work (Kurzarbeit)

Belgium

Temporary unemployment (chômage temporaire) Partial unemployment (chômage partiel)

400,000 jobs preserved estimated for 2009a Estimates ranging from 6,800 to over 30,000 jobs preserved in 2009b 42,600 jobs preserved in 2009c Estimates ranging up to 18,000 jobs preserved in 2009d 27,000–28,000 jobs preserved through parttime unemployment in 2009e 1,180,194 jobs preserved in 2009f 370,630 jobs preserved per month in 2009g 100,535 jobs preserved from May 2010 to 2011h Approximately 165,000 jobs preserved in 2009 and 100,000 jobs in 2010i Combined effect of all types of suspensions estimated at 5,800 jobs preserved per monthj in 2009–10 No effect

France

Netherlands

Japan

Short-time work (Werktijdverkorting) until the end of March 2009, part-time unemployment (Werkloosheidswet) thereafter Reduction of overtime

Turkey

Employment Adjustment Subsidy (EAS) Reduced Working Time (RWT)

United States

State work-sharing programmes with Short-Time Compensation (STC)

Uruguay

Unemployment insurance on suspension grounds (Seguro de desempleo por causal suspension) Partial Unemployment Insurance programme (Seguro de desempleo parcial)

Notes: a. Bellmann et al., ch. 2 in this volume. b. Flecker and Schönauer, ch. 3 in this volume. c. Flecker and Schönauer, ch. 3 in this volume. d. Flecker and Schönauer, ch. 3 in this volume. e. Flecker and Schönauer, ch. 3 in this volume. f. Ogura, ch. 4 in this volume. g. Ogura, ch. 4 in this volume. h. Yeldan, ch. 5 in this volume. i. Golden and Glosser, ch. 7 in this volume. j. González Fernández, ch. 6 in this volume.

M3125 - MESSENGER TEXT.indd 275

14/05/2013 13:33

276

Work sharing during the Great Recession

were actually the largest in Belgium. However, one has to be careful with such an interpretation because, while the rather conservative employment estimates presented in the chapter on Germany focus only on the effects of Kurzarbeit (and only for 2009), it is well established that reductions in overtime and drawing down the time credit balances in workers’ workingtime accounts (which are widely used in German industry) to zero or even to negative levels, also contributed substantially to the positive employment results known as the ‘German jobs miracle’ (Bellmann et al., ch. 2 in this volume). Moreover, while the EAS in Japan accounted for a substantial number of jobs preserved, the reduction of overtime hours actually provided the most powerful results – and in fact, many of the positions counted as being preserved due to the EAS scheme may also have benefited from overtime reductions. A few key factors may account for the results in Japan. With regard to overtime, as pointed out in the chapter on Japan (Ogura, ch. 4 in this volume), the financial cost of overtime is not as high in Japan as in most other developed countries (the overtime premium is only 25 per cent), and in fact, for many ‘white-collar’ workers, overtime hours are not compensated at all; partly as a result, overtime is widely used by most Japanese companies. In a country with famously long working hours for regular employment, the substantial reductions in the vast reserve of overtime appear to have been an extremely important measure in preserving jobs. The flexibility of a work-sharing programme in terms of variations in working hours and days is an important consideration in encouraging employers to seek cost reductions by cutting working hours instead of jobs. As shown in Table 8.1, the estimated employment effects of worksharing measures in Uruguay were nearly all attributable to the temporary suspension provisions of the UI programme in that country. Uruguay’s new Partial UI programme had very few participants, and thus little if any effect, owing both to its limited sectoral coverage and also to the fact that the reduced hours scheme in this programme required a minimum of one and maximum of two workdays, thus providing little flexibility regarding which days or, more importantly, which hours could be reduced. This inflexibility in the programme design meant that employers were reluctant to participate in this measure because it would have restricted their decision-making at a critical juncture. When a measure is meant to address a crisis, imposing such restrictions on firms’ decision-making can severely undermine programme participation – thus virtually ensuring that the measure will not have any effect. Another factor that posed challenges to the crisis work-sharing measures in these countries is employers’ perceptions. The Turkish experience highlights a concern that may not get enough attention as a work-sharing

M3125 - MESSENGER TEXT.indd 276

14/05/2013 13:33

Conclusion

277

programme design consideration. Specifically, larger Turkish employers were generally hesitant to engage in the RWT programme because they did not want the business community, and especially markets and their competitors, to think that their companies had substantial problems. Whether this perception actually manifested itself in reality is not in itself material; what is most important is that these larger Turkish employers believed it, and as a result, it meant that they were less likely to participate in the RWT programme. Likewise, some German employers were concerned about work sharing and training, believing that workers who got more training might subsequently consider leaving the company. Once again, the extent to which this situation actually occurs is not the main issue, but rather the concern is that it could happen. Paying closer attention to employers’ sensitivities, such as those described above, is essential if crisis work-sharing programmes are to have their desired effect. Workers’ perceptions of crisis work-sharing programmes and measures are very important as well. Key worker concerns regarding these measures include: the level of wage supplements paid to participating workers, and also their interest (or lack of it) in combining work sharing with training. For example, the German experience demonstrates that workers may be reluctant to engage in training because they might be concerned about their ability to handle the training materials and/or to successfully complete the training programme. Another notable factor affecting companies’ and workers’ participation in crisis work-sharing programmes is how easy or difficult they are to access. For example, the work-sharing measures in Austria, the Netherlands and Turkey can require weeks or even months to obtain government approval for participation. The problem with such long timeframes is that when a firm is in economic turmoil, it may not be able to survive during such a long waiting period. As noted above, on a proportional basis the Belgian temporary unemployment programme (chômage temporaire) may actually have been the most effective work-sharing programme, and that programme only requires that companies provide seven days’ notice to government officials and these officials do not need to issue any type of a formal approval. In short, it must be remembered that economic crises are stressful periods for both employers and workers, and if crisis work sharing is to serve its role as an economic ‘life preserver’, it should not require firms to ‘drown’ while waiting for permission to use the life preserver! Linked with the need to make crisis work-sharing programmes and measures more accessible is the need to make them preferable to other available alternatives – including mass layoffs. The relatively weak employment effects of the partial unemployment programme (chômage

M3125 - MESSENGER TEXT.indd 277

14/05/2013 13:33

278

Work sharing during the Great Recession

partiel) in France are illustrative of this ‘real-world’ test. Delays caused by programme reactivation and modernization, less attractive provisions compared with those in other countries, and the evolution of the labour market, including the increased use of temporary agency workers, all undermined the attractiveness of the French programme to employers (Flecker and Schönauer, ch. 3 in this volume). This experience highlights the need to enhance crisis work-sharing programmes to make them more attractive to employers, with simpler forms and administrative procedures; speedier approvals backed by the use of advanced IT systems; and greater flexibility in their application. Crisis work-sharing programmes and measures also need to be more effectively promoted, in order to stress to employers that by adopting work-sharing policies, the workforce and its firm-specific skill set are preserved, so that when economic conditions improve regular operations can be rapidly ‘ramped up’ and at minimal additional cost.

3.

IMPLICATIONS FOR POLICY

3.1

What is the Ideal Programme Design for a Crisis Work-sharing Programme to Preserve Jobs?

Based on the review of the lessons learned regarding the operation of crisis work-sharing programmes during the Great Recession and its aftermath in the previous section, there are a number of programme design features that appear to be of particular importance for making such programmes more effective. First, balanced eligibility criteria are essential. Such criteria must be stringent enough to screen out those companies that are clearly suffering from structural problems, rather than a cyclical downturn in sales and/or production. However, they must not be so restrictive as to discourage companies that are actually suffering from cyclical problems from applying for the programme. Obviously, this can be a difficult balance to achieve in practice, but typically a sharp drop in sales and/or production due to factors external to the enterprise are fairly reliable indicators of cyclical problems, and thus can serve as reasonable eligibility criteria for crisis work-sharing programmes. Some such programmes even require companies to demonstrate (via a business plan or sales projections) that they will be able to recover to their previous level of sales and/or production within a specified period (for example, the Canadian Work Sharing Program), but care must be taken not to make such requirements too burdensome. In fact, burdensome administrative requirements for companies,

M3125 - MESSENGER TEXT.indd 278

14/05/2013 13:33

Conclusion

279

extensive paperwork, and delays in programme activation (start-up) and the processing of company applications must be avoided at all costs. While some such requirements and paperwork are unavoidable – and indeed essential for reducing the potential deadweight and displacement effects of work-sharing programmes and measures – care must be taken to ensure that these are not viewed as unnecessarily burdensome by companies; otherwise, they will not even bother to submit an application for the programme. For example, the application form for the German Kurzarbeit programme is only a single page (both sides) in length, and this is most certainly a feature of the programme which companies find attractive. Likewise, delays in programme activation or re-activation at the onset of a crisis (which occurred to some extent in France – see Flecker and Schönauer, ch. 3 in this volume) or in processing applications can reduce firms’ participation. The need for flexibility in the volume and patterns of hours reductions which are permitted and supported is another important feature of a well-designed crisis work-sharing programme. In this case, the German Kurzarbeit programme can perhaps be regarded as an extreme example, because the working hours of individual workers can be reduced by up to 100 per cent (see Bellmann et al., ch. 2 in this volume). The main point here is that some companies may require greater reductions in working hours (and hence labour costs) than others in order to remain viable, and welldesigned programmes should provide sufficient flexibility to respond to a variety of circumstances. In some situations, the pattern of hours reduction may also be important; for example, some firms may find it more convenient to reduce work hours on a daily basis, while others may prefer to maintain the same daily hours but reduce the length of the workweek – for example, to move from a five- to a four-day workweek. Well-designed programmes should be able to accommodate these types of variations in work hours as well. In addition, wage supplements for affected workers are essential for partially offsetting the reduction in their earnings due to the temporary reduction in their hours of work. These wage subsidies, which often take the form of partial unemployment benefits (called ‘short-time compensation’ in many countries) should be sufficiently generous that they are able to replace a substantial proportion (50 per cent or more) of participating workers’ lost income. This wage replacement is important, for helping participating workers to maintain not only their standard of living, but also their purchasing power – thus supporting aggregate demand. Next, reasonable but fixed time limits on work-sharing subsidies are crucial for minimizing the potential deadweight and displacement effects of such measures. Most crisis work-sharing measures establish specific

M3125 - MESSENGER TEXT.indd 279

14/05/2013 13:33

280

Work sharing during the Great Recession

time limits on the period of work sharing. Such limits help to ensure that work sharing is a temporary measure in response to an economic crisis or another short-term reduction in demand, and not a permanent reduction in hours (and pay). As we saw in the previous section, the precise duration of the work-sharing subsidies for individual companies and their workers varies substantially from one programme to another. However, based upon the experience of the Great Recession, a duration of between six months and one year appears to be a reasonable time limit, and according to the contributions in this volume, this does not appear to generate any significant deadweight effects. And, while displacement effects are notoriously difficult to measure, the mere existence of such time limits should help to minimize any potential displacement effects associated with these types of programmes. Crisis work-sharing programmes and measures are important in their own right, but if they are to be combined with training and retraining programmes, then careful consideration should be given as to how this accompanying training is structured and conducted. Better integration between crisis work-sharing programmes and training might best occur if certain factors were considered. First, simpler administrative procedures for organizations and workers might encourage more companies to participate in such programmes. Also, short-term or modular courses could be developed that allow workers to rotate out of work for short periods (from a few days to two weeks) and come back to work after completing the course. In addition, to address any concerns about workers leaving firms immediately following the completion of training, a requirement might be placed that the worker must stay with the establishment for a specified period of time after taking a training course. Such measures may provide better short- and medium-term solutions to promote employment retention, which is, after all, the primary aim of crisis work-sharing programmes. Last, but most certainly not least, it is essential for governments to take an active role in supporting and promoting crisis work-sharing programmes and measures. It is certainly possible for individual companies to implement work-sharing measures acting on their own – and this in fact happened in firms in several countries (for example, Serbia, the United Kingdom) where no national crisis work-sharing programmes exist. However, the lack of public support makes this a difficult and expensive proposition for companies. In addition, with individual company work-sharing arrangements, wage supplements for affected workers are normally not available, since most companies (with a few exceptions in countries such as Germany and the Netherlands) do not establish funds for this purpose. In any case, providing company-funded wage supplements to affected workers defeats one of the key objectives of crisis work-

M3125 - MESSENGER TEXT.indd 280

14/05/2013 13:33

Conclusion

281

sharing measures: an immediate reduction of companies’ labour costs without layoffs. Nevertheless, simply creating a government-supported crisis worksharing programme is not enough. Many such programmes which have been created with the best of intentions never attract a sufficient level of participation to have any real economic impact. For example, many of the (now) 24 state-level work-sharing programmes (plus the District of Columbia) in the US are so poorly publicized that they are small and remain unknown to the vast majority of potentially eligible companies and workers. If such state work-sharing programmes with ‘short-time compensation’ (that is, partial unemployment benefits) had been implemented in all US states and taken to scale, these programmes, which preserved a mere 165,000 jobs in 2009, could have potentially saved far more jobs (see, for example, Houseman 2011; Golden and Glosser, ch. 7 in this volume). Therefore, it is absolutely essential for crisis work-sharing programmes to engage in aggressive informational campaigns and outreach to potentially eligible companies and workers. Moreover, the experience of crisis work-sharing programmes and measures during the Great Recession of 2008–09 and its aftermath suggests that it is crucial to have such measures in place prior to the onset of a crisis in order for them to function effectively (see, for example, OECD 2010; Yeldan, ch. 5 in this volume). Once such measures are in place, they can then be readily expanded and/or adapted as circumstances warrant (Bellmann et al.; Flecker and Schönauer; Ogura; and Yeldan, chs 2–5 in this volume). 3.2

Can Work Sharing also be an Effective Job-creation Tool?

The preponderance of the evidence presented in this volume (and many other recent studies as well, including those summarized earlier in this chapter) demonstrates that one of the key lessons learned from the experience of crisis work-sharing programmes and measures during the Great Recession and its aftermath is that in general cutting hours of work has positive effects on employment levels during economic crises. Although available estimates of the employment effects of such programmes and measures vary considerably across countries – and even within countries as well – nearly all of them concur that crisis work-sharing measures preserved jobs, typically with minimal deadweight effects. Obviously, this suggests that crisis work-sharing measures are important economic stabilizers, but it also raises an intriguing question: can reductions in work hours also have positive effects on employment during periods of economic recovery and growth as well? As was discussed in Chapter 1, permanent work-sharing measures are

M3125 - MESSENGER TEXT.indd 281

14/05/2013 13:33

282

Work sharing during the Great Recession

different from crisis work-sharing programmes and measures – which are fundamentally temporary in nature. Work sharing as a measure designed to spur hiring and thus increase employment also involves a downward adjustment of hours of work. However, in this case the reduction of working hours is not being made in response to a decline in the demand for a firm’s products or services, but rather is the result of an explicit government policy. Such reductions in hours of work can be achieved by different methods, ranging from legally mandated reductions in the normal or standard legal workweek in a country to collective bargaining in specific industries to the use of tax incentives (for example, reduced payroll taxes or tax credits) provided to companies that reduce the average workweek of employees in their enterprises on a weekly, monthly, or even an annual basis, or some combination of these and policies (including the promotion of part-time work, that is, shorter than full-time hours of work). Regardless of the specific method used, the objective of this form of work sharing is the same: to reduce actual average hours of work per worker in order to increase the aggregate employment level. However, as was discussed briefly in Chapter 1 and in much greater detail in Chapter 7, there are a number of different factors (for example, overtime ‘leakages’; productivity ‘offsets’; the fact that workers are not perfect substitutes for each other (so-called ‘indivisibility of labour’), which could potentially lead to skills shortages in industries or occupations where hours are reduced; the potential for increases in unit labour costs, and so on) that might substantially reduce the extent to which permanent reductions in hours of work are translated into new employment. So what would actually happen if the hours of work in a country or in a particular industry were reduced? In Chapter 7, Golden and Glosser attempt to answer that question. Specifically, they review the theoretical and empirical evidence regarding the use of permanent reductions in hours of work to generate employment from a range of countries. They find that, based on evidence from a range of mainly European country and industry studies, permanent reductions in working hours have generally (although certainly not always) shown net positive employment effects – although the actual effects can vary substantially depending upon a range of factors, particularly the mix of policies used. Golden and Glosser emphasize that ‘[v]irtually all empirical studies of actual European cases show some positive employment effects of working-time reductions’ (Golden and Glosser, ch. 7 in this volume, p. 219). In general, they find that larger reductions in hours of work appear to have greater employment effects than more marginal reductions, and that reducing the length of the workday is likely to be less costly for employers than cutting the number of working days in a week. Nevertheless, they note that many of these same studies also find

M3125 - MESSENGER TEXT.indd 282

14/05/2013 13:33

Conclusion

283

increases in overtime and labour costs associated with reductions in work hours, and that these factors appear to restrain the job creation effects of the shorter hours (see, for example, Bosch and Lehndorff 2001; see also Chapter 7 for additional references to this literature). After reviewing the available evidence from cases of actual, permanent reductions in working hours, Golden and Glosser then use time-series analysis techniques1 to simulate the effects of a downward shock – that is, a one-time adjustment of a given size – in weekly hours of work in one country: the United States. Their findings suggest not only that work sharing – even without short-time compensation – could have had a ‘considerable neutralizing effect’ on job loss, but also that employment growth in the US during the period of economic recovery following the Great Recession (mid-2009 through 2011) was restrained in part by an increase in hours of work during the recovery period – resulting in what is commonly known as a ‘jobless recovery’. Consequently, it appears that a reduction in average weekly hours via work sharing could potentially have induced employers to move more quickly to expand hiring during the subsequent economic recovery than would otherwise have been the case. These results regarding the potential employment effects of reductions in weekly hours in the US under different sets of assumptions – combined with studies estimating the empirical effects of actual hours reductions in Europe and elsewhere – suggest that working-time reductions also have the potential to be an effective tool for modestly increasing employment levels. Whether this potential can be realized depends on the national environment (economic, social and political) in the particular country, and crucially, on the specific mix of policies that are used. In terms of the national environment, the potential for job creation via reductions in working hours is generally the greatest where companies use substantial amounts of overtime (at premium rates), as this appears to help offset potential increases in unit labour costs. In addition, the higher the fixed costs of employment per employee in a country (as opposed to variable costs), the greater will be the incentive for companies to extend actual hours of work for their employees, and therefore, the less the potential for increased employment resulting from reductions in working hours. In terms of the specific mix of policies used to reduce hours of work, as stated at the end of Chapter 7, it is not clear from the available evidence what form of working-time reduction – a shorter standard workweek (that is, normal weekly hours of work), limits on overtime, more opportunities for quality part-time work, or reduced annual hours of work (for example, increased periods of paid annual leave) – would be most likely to deliver employment gains. However, it does appear that shortening

M3125 - MESSENGER TEXT.indd 283

14/05/2013 13:33

284

Work sharing during the Great Recession

hours of work each day is superior to reducing the number of workdays per week. Moreover, it seems crucial that any reductions in weekly hours of work should be accompanied by partial – rather than no or full – wage compensation, in order avoid a negative macroeconomic impact, while not raising employers’ unit labour costs. Finally, and perhaps most importantly, Golden and Glosser emphasize that ‘[t]he greatest potential gain in well-being for workers would occur if work-sharing measures were targeted towards the sectors where employment gains are most realizable, and also where hours reductions are mainly preference based, in those industries where involuntarily long work hours are the highest’ (p. 252). In other words, by carefully targeting permanent reductions in hours of work to specific industry sectors and groups of workers based on industry circumstances and workers’ preferences, the potential employment gains associated with such hours reductions can be maximized. At the least, the findings from Chapter 7 would appear to justify experimentation with various approaches for using working-time reductions as a job-creation tool, such as for example tax and/or other incentives (for example, converting fixed labour costs to variable ones) for companies that act to reduce weekly hours of work, combined with careful evaluation of the results of such experiments.

4.

PROSPECTS FOR THE FUTURE

The proven job-preservation effects and the potential job-creation effects of work-sharing programmes and measures are important reasons for considering the use of such measures. Their benefits to those who would be unemployed without work sharing are obvious. However, there are other potential benefits of such measures as well. For one thing, the existence of substantial portions of workers who are overworked – and thus more likely to experience health problems and work–life conflict (see Tucker and Folkard 2012 and Fagan et al., 2012 for a review of the recent evidence) – side by side with substantial proportions of workers who are unemployed or underemployed, suggests that the current distribution of hours of work is less than optimal in many countries. As was discussed in Chapter 1, those workers who are most at risk of overwork – those suffering from symptoms of work-related fatigue and stress – could benefit from reduced hours. In addition, workers who are overemployed – that is, those who would prefer shorter hours of paid work even if this results in reduced earnings – could benefit from reduced hours as well (Golden and Glosser, ch. 7 in this volume). Both of these groups of workers could also potentially benefit from improved work–life balance due to shorter hours.

M3125 - MESSENGER TEXT.indd 284

14/05/2013 13:33

Conclusion

285

On a more macro level, there is also the question of the environmental impacts of reductions in working hours. As the recent, high-profile debates over greenhouse gas emissions and global climate change indicate, environmental issues are a growing area of concern for the entire world. As a result, the longstanding emphasis on economic growth as the primary measure of both economic and social progress is increasingly being called into question (see, for example, Stiglitz et al. 2009; Schor 2012). However, creating more jobs with limited growth seems virtually impossible unless more employment can be realized from an existing (or at least a similar) level of output. There are very few public policies that have even the potential to contribute to achieving this objective, but work sharing is one of them. While the potential environmental benefits of more permanent work-sharing policies may be even more difficult to quantify than their potential employment effects, they are certainly no less important. All of this is not to say that work sharing is some sort of a magical ‘silver bullet’. However, as countries – particularly those with advanced economies – are discovering, regardless of their economic and fiscal policies, global economic growth is increasingly moving to the developing world. Perhaps this is simply inevitable, and even desirable, as countries which have lagged behind the West economically for various reasons are now starting to close the gap. None the less, given the growth rate of the labour force in many developing countries, it is not clear that – even with high and sustained levels of economic growth – they will be able to fully absorb their rapidly expanding workforces. And if such growth is not forthcoming, the result could be mass unemployment in a large number of developing countries. For developed countries, where substantial, sustained economic growth appears unlikely and perhaps even worrying from an environmental perspective, the question is even starker: how can these countries maintain their employment levels and living standards in the face of this new global economic reality? Can work sharing support the creation of more sustainable economies and societies? While it is clear that work sharing cannot meet such an enormous challenge alone, it can be one of a number of measures which help to promote increased employment, improved work– life balance, more sustainable economies, and ultimately, more equitable societies.

NOTE 1. Specifically, they use a time-series analysis technique known as vector autoregression (VAR) analysis. Methodological details can be found in Chapter 7.

M3125 - MESSENGER TEXT.indd 285

14/05/2013 13:33

286

Work sharing during the Great Recession

BIBLIOGRAPHY Arpaia, A., N. Curci, E. Meyermans, J. Peschner and F. Pierini (2010), ‘Short-time working arrangements as response to cyclical fluctuations’, European Economy Occasional Papers No. 64, European Commission, Directorate-General for Economic and Financial Affairs and Directorate-General for Employment, Social Affairs and Equal Opportunities, Brussels. Bosch, G., and S. Lehndorff (2001), ‘Working time reduction and employment: experiences in Europe and economic policy recommendations’, Cambridge Journal of Economics, 25: 209–43. Crimmann, A., F. Wiebner and L. Bellmann (2010), The German Work-Sharing Scheme: An Instrument for the Crisis, Conditions of Work and Employment Series No. 25, Geneva: ILO. Fagan, C., C. Lyonette, M. Smith and A. Saldana (2012), ‘The influence of working time arrangements on work-life integration or “balance”: a review of the international evidence’, Conditions of Work and Employment Series No. 32, ILO, Geneva. Hijzen, A. and D. Venn (2011), ‘The role of short-time work schemes during the 2008–09 recession’, OECD Social, Employment and Migration Working Papers, No. 115, OECD, Paris. Houseman, S. (2011), ‘Labor market flexibility: a view from the United States’, presentation prepared for the conference on ‘Increasing Labour Market Flexibility: Boon or Bane?’, Institute for Employment Research, Nuremberg, Germany, 18–19 March. Messenger, J.C. (2009), ‘Work sharing: a strategy to preserve jobs during the global jobs crisis’, Policy Brief No. 1, Conditions of Work and Employment Programme, ILO, Geneva, June. Organisation for Economic Co-operation and Development (OECD) (2010), OECD Employment Outlook: Moving Beyond the Jobs Crisis, Paris: OECD. Schor, J. (2012), ‘Working hours in the debate about growth and sustainability’, nef/CASE About Time Colloquium, 12 January. Stiglitz, J.E., A. Sen and J.-P. Fitoussi (2009), ‘Report by the Commission on the Measurement of Economic Performance and Social Progress’, Commission on the Measurement of Economic Performance and Social Progress, Paris. Tucker, P. and S. Folkard (2012), ‘Working time, health, and safety: a research synthesis paper’, Conditions of Work and Employment Series No. 31, ILO, Geneva.

M3125 - MESSENGER TEXT.indd 286

14/05/2013 13:33

287

M3125 - MESSENGER TEXT.indd 287

14/05/2013 13:33

Austria

Table 8A.1

Short-time work (Kurzarbeit)

Work-sharing programme name All employees, including temporary agency workers, except apprentices, CEOs & board members; other possibilities have to be exhausted first, such as use of overtime & holiday entitlements, or workingtime accounts

Eligibility criteria (workers & employers) Short-time work allowance (Kurzarbeiterunterstützung) equivalent to unemployment benefit (about 55% of the net wage) plus sickness & retirement insurance & family allowances if applicable. Support is increased by 15% if the employer offers training

Type & amount of support (to workers & employers) Social partners may also agree on further training for short-time work (Behaltefrist); if work sharing includes training, then the flat hourly rate (Qualifizierungsunterstützung) contains a supplement of 15% for additional expenses related to training; these contributions are paid by Employment Service after 6 months of shorttime work

Training provisions in work-sharing programmes

National work crisis-sharing measures/programmes (reviewed in this volume)

APPENDIX 8A

Short-time work approved for 6 months (1st), but can be prolonged up to a maximum of 18 months; if application was accepted before 12/2010, then short-time work could be extended up to 24 months

Duration (time limits of programme)

288

M3125 - MESSENGER TEXT.indd 288

14/05/2013 13:33

Belgium

Table 8A.1

Temporary collective reduction of working time

Temporary unemployment (chômage temporaire)

Work-sharing programme name

(continued)

Reduction in working time by 1/5 or a quarter; company collective agreement required; company is supported by a reduction in the social security contributions

Employees, including temporary agency workers & fixed-term employment contracts; in some cases only bluecollar workers

Eligibility criteria (workers & employers)

Company collective agreement monthly compensation paid by employer for reduced wages in the case of working-time

Maximum 75% of the previous income, depending on family status, up to a gross monthly wage of €2,206.40

Type & amount of support (to workers & employers)

Training provisions in work-sharing programmes

Can alternate between days and weeks of unemployment, but overall unemployment can be for a maximum of four continuous weeks; partial unemployment for a maximum of 12 months

Duration (time limits of programme)

289

M3125 - MESSENGER TEXT.indd 289

14/05/2013 13:33

Temporary individual reduction of working time

Based on existing time credit schemes. Parties decide to reduce working hours by mutual agreement. Company must be in financial difficulties; working-time reduction by 1/5 or half the minimum period of 6 months; normal

reductions – by 1/5 – a minimum of €150 – by a quarter – minimum €187.50; limit: the wage & compensation combined must remain below the employee’s last full-time wage: these amounts are increased by €100 per day if a 4-day week is implemented; employer benefits from a reduction in social security contributions Allowance is paid by the national employment office, workingtime reduction by 1/5 is €188 (€248 if worker is aged 50 or over), by half €442; the employer can pay an Duration is limited to 16 weeks in the case of full suspension of work duties or 26 weeks in the case of partial suspension

290

M3125 - MESSENGER TEXT.indd 290

14/05/2013 13:33

France

Table 8A.1

Partial unemployment (chômage partiel)

Work-sharing programme name

(continued)

additional compensation, but not compulsory

conditions of time credit do not apply; applicable to full time employees only. Consent required of employee in the form of an agreement between the worker & the employer; collective agreement at company level or a company plan must be approved by an ad hoc commission. Wage must be below employee’s previous full-time wage All employees, including part-time workers, temporary agency workers and those on fixed-term employment contracts, if their employer has a decrease in sales at least 50%. All employees in the case of temporary layoff The employer pays 75% of the hourly gross wage nonworked hours at a minimum of €6.84 an hour; the state pays SMEs €3.84 & large companies €3.33 per hour per worker. In the case

Type & amount of support (to workers & employers)

Eligibility criteria (workers & employers)

Training leave (congé individual de formation) provides remuneration of 80% of net salary (a level higher than partial unemployment insurance)

Training provisions in work-sharing programmes

800 hours per employee during crisis for a period of 6 months with one renewal making the maximum total 1 year. For textile,

Duration (time limits of programme)

291

M3125 - MESSENGER TEXT.indd 291

14/05/2013 13:33

Germany

Work sharing for economic reasons

Seasonal work sharing

Mainly used in the construction industry & other outdoor professions (gardeners, etc.) during nonproductive times (winter season) (§175 Social Code III) Temporary unavoidable loss of work due to economic factors or an unavoidable event (§170 Social Code III); provides support to establishments in crisis to maintain employment

of a company closure of more than 3 months, employees are entitled to unemployment benefits

clothing, leather, and automotive industry the maximum hours per worker was set at 1,000 hours over the same period (6 months, plus 6 more months maximum) Winter season; can be granted from 1 December to 31 March

292

M3125 - MESSENGER TEXT.indd 292

14/05/2013 13:33

Table 8A.1 Eligibility criteria (workers & employers) Loss of work must be due to economic reasons (not structural), must be temporary, and at least 1/3 of staff (including apprentices & trainees) must be affected with an estimated loss of income of at least 10%; jobholders who have lost at least 10% of gross monthly earnings individual-level working time can be reduced up to 100%; loss of work also counts if less than 1/3 of entire staff is affected by income loss of at least 10%. Sound chance establishment will return to reasonable workload

Work-sharing programme name

Short-time work compensation (Kurzarbeitforgeld) (§169ff. of Social Code)a

(continued)

Worker with 1 dependant paid 67% of net difference to regular wage & those without dependants get 60%; for loss of work of 100% compensation is at the level of unemployment benefits. Employment contract must be maintained. Social insurance is reduced to 80% & employer must cover full amount, but since 2009 covers 50% of social insurance contributions.

Type & amount of support (to workers & employers) Training-related work sharing. In the case of continuing training Federal Employment Agency covers 100% of the contributions & under conditions the costs of training can be reimbursed up to 100%

Training provisions in work-sharing programmes

Duration of work is for all employees of firm and starts with first payment of STC. If there is an interruption of at least 1 month there is an extension to an equivalent period; if a break of 3 months then new period of work sharing may start. Maximum duration overall was 12 months,

Duration (time limits of programme)

293

M3125 - MESSENGER TEXT.indd 293

14/05/2013 13:33

Japan

Employment Adjustment Subsidy (EAS)

Public subsidy paid to employers whose production or sales decreased by the economic slump in the last 3 months and who make efforts to save employee jobs for a maximum of 1 year. Enrolment period for workers is unquestioned, reduction of production volume, sales or business activity in monetary terms (added in 2009); Funds cover 100% by employment insurance; amount of support to workers requiring vocational training raised from 1,200 yen per day to 4,000 yen for a large firm & 6,000 yen for SMEs

Compensation paid by establishment which is reimbursed from the Ministry of Labour Unemployment Insurance Funds

Subsidy for vocational training is 6,000 yen for large firms and 4,000 yen for SMEs

extended in January 2009 to 18 months, and extended further in June 2009 to 24 months (since STC duration reduced to 18 months again, and further reduced to 12 months in 2011 EAS is paid for a maximum of 300 days for up to 3 years

294

M3125 - MESSENGER TEXT.indd 294

14/05/2013 13:33

Netherlands

Table 8A.1

Part-Time Unemployment (DeeltijdWerkloossheidswet)

Short-Time Work (Werktijdverkorting)

Work-sharing programme name

(continued)

No drop in turnover; agreement between workers & employer about necessity of shorttime work; employees could reduce hours & apply for part-time unemployment benefits; agency workers not

5% or greater reduction in production or sales in last 3 months compared with previous 3 months or same 3 months in the previous year No temporary agency workers; employers have to undertake training of the staff on short-time work; company must have a drop in turnover of 30% or more in 2 months

Eligibility criteria (workers & employers)

Employees receive 70% of the normal wage for reduced part of working hours (employees lose remaining 30%, i.e., 15% of their total earnings).

70% of the wage for non-worked hours

Type & amount of support (to workers & employers)

Training is obligatory with employers able to claim special financial assistance if they offer further training to employees Training is obligatory with employers able to claim special financial assistance if they offer further training to employees

Training provisions in work-sharing programmes

26 weeks normally; if share of workers exceeds 60% of total staff in the company then duration is 39 weeks;

Maximum of 15 months

Duration (time limits of programme)

295

M3125 - MESSENGER TEXT.indd 295

14/05/2013 13:33

Turkey

Reduced Working Time (RWT) (Labour Law 4867, art. 65)b

Companies can supplement the wage up to 100% (based on collective agreement) If accepted into the programme the employer does not pay any wages to their employees, or social security or UI. Unemployment Insurance Fund (ISKUR) pays wages to employees under conditions set by law. 2/3 of employee health & maternity benefits paid by Social Security Institute. Employees should be covered for

included; working hours reduced by minimum 20% & maximum 20%

Presence of economic crisis and unfavourable conditions trigger the policy (needs to be an economic crisis beyond the decision making of the firm, i.e., a sectoral, national, or international economic crisis). Reduction of weekly working time by at least 1/3 over a period of at least 4 weeks in a continuous fashion and at most for 3 months. Employer requests application form which is posted to them. Labour inspectors investigate claims made by firm after application

share of workers between 30 and 60% then 52 weeks; less than 30% is 65 weeks RWT initially allowed firms to apply for reduced working-time coverage for a period of 3 months, which was extended to 6 months in February 2009 and then to 1 year in May 2009

296

M3125 - MESSENGER TEXT.indd 296

14/05/2013 13:33

Uruguay

Unemployment Insurance on temporary suspension grounds

Work-sharing programme name

Table 8A.1 (continued)

All industries are eligible. Unemployment benefit can be given for dismissal, suspension of work activities, and reduction in the number of workdays or work hours. Suspension used for drops in business activities to avoid firing workers (a worker retention system). Rotation of workers on UI with consent of company and trade union

Eligibility criteria (workers & employers) at least 120 days uninterrupted up until time of coverage and the employee should have worked and paid UI premiums for at least 20 months (600 days) over the last 3 years Subsidy is financed through social security contributions from workers and employers (any deficits covered by tax and general revenue allocations)

Type & amount of support (to workers & employers)

Training provisions in work-sharing programmes

Up to 1 year, but extension based on discretionary decision by the Executive (Cabinet) & parliament allows for further extension

Duration (time limits of programme)

297

M3125 - MESSENGER TEXT.indd 297

14/05/2013 13:33

Partial Unemployment Insurance

Industries covered include only leather, textiles and garments, wood and wood products, & metalworking. Paid by unemployment scheme, but also meant to make sure workers not working are in training; 15% drop in company activity over quarter (compared with previous 2 years); reduction of at least 1 and no more than 2 days per workweek; company cannot have dismissed more than 5% of workers in quarter under coverage; if company has workers on UI (suspension option, see above) then they must be reinstated before this subsidy applies; while benefit is paid, no suspension or dismissal of workers except for disciplinary reasons; Subsidy is financed through social security contributions from workers and employers (any deficits covered by tax and general revenue allocations) Workers obligated to attend training courses or return to formal education

No longer than 1 year and in event of extensions the temporary nature of the lack or reduction of work must be documented and the commitment to preserve jobs must be secured.c

298

M3125 - MESSENGER TEXT.indd 298

14/05/2013 13:33

Work-sharing programme name

(continued)

adoption of Collective Bargaining Agreement to cover all workers for a term through the duration of the subsidy and workers are obligated to attend training courses or return to formal education

Eligibility criteria (workers & employers)

Type & amount of support (to workers & employers)

Training provisions in work-sharing programmes

Duration (time limits of programme)

Notes: a. For reasons of completeness, ‘Transfer Kurzarbeit’ was a type of work sharing used during reunification of East and West Germany. b. Turkish law clearly distinguishes reduced working time and short-time temporary work. Temporary employment allows for both parties to agree on the temporary nature and overall timeframe of the work contract. In reduced working time, by contrast, the nature of the work is permanent, but is subject to necessary reduction due to external factors outside the enterprise. c. Other than for disciplinary reasons or gross misconduct this programme does not allow for dismissal of workers for the duration of the programme (and benefit paid).

Table 8A.1

Index accession negotiations with EU Turkish trade union recognition 121 activity, employment and unemployment rates total, Uruguay 188 additional hiring, substantial costs 4, 211 age and sex combination of workers in RWT programme, Turkey 136 aggregate demand, crucial role 210 agricultural employment in Turkey 129 agricultural motor vehicle construction Turkey workers and managers 140 agricultural versus manufacturing industry, Turkey 128 agriculture, unregistered workers, Turkey 138 American Federation of Labor (AFL) founding leader, Samuel Gompers 222 quotation of, on working hours 10, 222 applications for and total amount of EAS 109 application to ISKUR for assistance from RWT programme 133 approval mechanisms, lengthy, Austria, Japan 264 Arizona layoff aversions 225 asset prices 58 Austrian Employment Service (AMS) 76–7 Austrian experience of work sharing 72 Austrian Institute of Economic Research estimation of jobs 80 automation and mass unemployment fear 208 automobile and wood industries physical volume of production 172

automotive industries, Germany 27–8 hardest hit by crisis 52–3 average duration of unemployment (weeks) 192 average weekly hours, employment and output in manufacturing, 1990–2011 233 average weekly overtime hours in manufacturing, 2006–2011, US 229 average weekly overtime hours of employees 2006–2011, US 229 durable goods and non-durable goods, US 230 Banco de Provisión Social contributors (annual monthly average, in number of jobs), 192 resources and expenditure, Uruguay 190 Belgium 40 to 36 hours reduction, 1979 12 anti-crisis measures 87 eligibility criterion, financial difficulties of company 262 experience of work sharing 72 blue-collar workers and white-collar workers arrangements, Belgium 81–2 BPS, see Banco de Provisión Social California, 1979 first US state work-sharing programme 13, 225 study of crisis work sharing 225 Canada estimate of overtime working 224 out of unemployment compensation funds 13

299

M3125 - MESSENGER TEXT.indd 299

14/05/2013 13:33

300

Index

‘short-time compensation’ (STC) benefits capital capacity, under-use 211 career development and training poor opportunities for non-regular employees, Japan 114 cash crops, expansion, Turkey 129 change in employment 2008–2009, Germany 53 changes in real GDP for OECD countries 29 characteristics of unemployed persons on unemployment insurance 193–4 characteristics of Unemployment Insurance Beneficiaries on suspension grounds ages 175 by gender 175 chemical industries, Germany 27–8 chômage partiel in France 87–91 chômage temporaire in Belgium 81–7 civic activity, ‘crowded out’ by work 206 clerical tasks, elimination from professional jobs 215 co-determination in German industrial relations 50 collective agreements in Germany 50, 51 collective bargaining 14, 16, 170 Turkey 121 Companies’ labour costs, reduction 72 company activity areas using unemployment insurance 175, 177–8 company-level pacts for employment 30 company products, reduction in demand 13 compensation for employees, Netherlands 92 compensation to establishment 33, 39 complexity of working hours, Japan 112–14 composition of employment percentage 191 computerization 215 Confederation of progressive Trade Unions of Turkey (DISK) 143–4

M3125 - MESSENGER TEXT.indd 300

radical stance 121 Confederation of Turkish Labour Unions (Turk-Is) largest in Turkey 143 Connecticut, USA 226–7 construction and metal industries, Netherlands 93 Continuous Household Survey (CHS) Uruguay 161–2 correlations between weekly hours and employment levels, US 234–5 costs of hiring and firing 39 country-specific description 73 crisis work sharing, Japan shortened scheduled hours, wage reduction 106 short-term stopgap 114 crisis work-sharing measures 19, 222 design of 259–60 crisis work-sharing programmes, North America 225 criteria for eligibility for companies and workers 259–65 daily rates of temporary unemployment subsidies per person, in Belgium 86 debt crisis in Germany impact on world trade 26–7 decline in labour productivity work sharing as establishments’ only use 54 Deeltijd-werkloosheidswet (WW), Netherlands 91–4, 264 Denmark manufacturing firms work-sharing agreements 17 dependent and independent variables 68–9 dismissal compensation, generous, Turkey 123 dismissal, suspension, reduction, subsidies Uruguay 172 diversification of exports, Uruguay 152 downturns in employment work-sharing responses in Europe 219 downward adjustment of working time 203

14/05/2013 13:33

Index dual social status based on employment form, Japan 115 duration of short-time work, variance in 2009 80 duration of work sharing, Germany 33 Dutch Part-Time Unemployment measure 270 earnings-related pension contribution, Japan 115 EAS, see Employment Adjustment Subsidy Eastern Europe, introduction of work sharing 259 East German reunification 32 econometric analysis 49–71 economic activity increase, Uruguay 161 economic and labour market conditions Uruguay 151–5 economic crises 203 economic difficulties, national circumstances 262 economic policies, response 2 educational activities, ‘crowded out’ by work 206 education services correlation between weekly hours and employment levels 234 elasticity of employment for majority of firms 220 eligibility conditions for work-sharing programme, Uruguay conditions for companies to receive subsidy 170 eligibility criteria for companies and workers 261–5 financial difficulties of company 262 eligibility criteria for short-time working, Netherlands 92 Emergency Decree, German government, 1931 reduction of working hours 11 employability enhancement 166 employed persons covered by PUI 198 employee numbers, estimated effects on 70 employee unions in Turkey 121 employee’s perspective of work sharing 37

M3125 - MESSENGER TEXT.indd 301

301

employees with fixed-terms contracts 50 employer credit 148 employers’ lack of interest in ISKUR’s education campaign 136–8 employers’ perspective of work sharing 38–40 employment adjustment in Japan 100 Employment Adjustment Subsidy (EAS), Japan 108–11, 263 effect on employment 107–12 Employment Agency, Turkey (ISKUR) 122 direct mandate over RWT 145 employment and unemployment rates, urban, Uruguay, 2003–2005 154 employment-creation effects 1, 207–8 employment decline establishments affected by crisis 53 employment decrease 7 employment distribution by activity sector, Uruguay, 2010 153 employment effects, estimated evidence from Europe 217–22 employment effects of working-time reductions 216 employment effects of work sharing country by country 275 Germany 52–6 programmes 273–8 employment expansion conditions 215–17 employment in agriculture and in manufacturing, Turkey 129 employment index 200–201 Employment Insurance Programme, Japan 114 employment level increase 4 employment losses in establishments, Germany 54 employment maintenance work sharing, Japan 116 employment measures, anti-crisis, Belgium temporary measures 83–4 employment opportunities, maximum 204 employment packages, employer-focus 143

14/05/2013 13:33

302

Index

employment part-time, slack work or business conditions 231 employment-preservation programmes deficiency of short-term structure 130, 136 employment promotion 19 employment protection in Japan, EAS 111–12 employment protection laws in EU 223 employment protection legislation, Turkey 122–3 ‘employment-related’ restrictions 270 employment security for non-regular employees, Japan 114 employment stability in Germany decline in GDP 28 employment subsidy 174 employment, unregistered informal, 2008–2010 139 Turkey 138 engineering industries negative estimates of employment elasticity in standard hours 220 entrepreneur interest in work-sharing PUI schemes 183 establishments, larger share of public financing 48 establishments using work sharing by industry 45 estimates, fixed-effect, output per employee, Germany 55 European diversity of work sharing 72 European Union (EU) 18 euro system stabilization 24 evolution of Unemployment Insurance Beneficiaries 163 expenditure for short-time work, Austria 80 expenditure on unemployment insurance 158–9 export orientation of German economy 52 Fair Labor Standards Act (FLSA), 1938 12 Federal Employment Agency, Germany

M3125 - MESSENGER TEXT.indd 302

social security contributions 35 training activities 46–7 Federation of Italian Industry legislation on 40-hour working week 11 financial activities correlation between weekly hours and employment levels 234 financial assistance with training, employers, Netherlands 93 ‘fire and re-hire’ costs 4 strategy 38 fixed-term contracts 30, 50 non-renewal, France 90 flat rate for full-time employees 77–8 flexibility measures 50 during crisis, Germany 43 measures of establishments using work sharing, Germany 44 ‘flexibility’ measures, use of, Germany classification of 41–5 flexibility of working hours in France 90–91 flexible working schemes, Turkey 144 flexi-time arrangements 30 FLSA, see Fair Labour Standards Act, 1938 Forty-Hour Week Convention, 1935 (No. 47) 10–11 France ‘35 hours’ workweek 12 Aubry I and Aubry II, 1998, 2000 13–14 crisis work-sharing measure experience of work sharing 72 legislation on 40-hour working week 11 raw materials provision, natural disasters, company restructuring 262 Socialist government proposal, 1981 French chômage partiel (partial unemployment) 15–16 full suspension as alternative to PUI, Uruguay 178 furniture and forestry products firm, Turkey crisis in export sales 140–41 future prospects 284–5

14/05/2013 13:33

Index GDP, decline in, Spain versus Germany vis-à-vis employment effect 2, 28–9 general unemployment insurance scheme, 1981 Uruguay’s legal scheme for UI 155 German Council of Economic Experts (GCEE) 26 German decline in GDP in Great Recession 52 German economic state 25–7 German GDP, development, 2006–2010 26 German ‘jobs miracle’ 30 labour hoarding, explanation 56 German labour market during crisis 27–8 German labour unions reduction of standard hours, industry-by-industry 220 ‘German Miracle’, reasons for 28–31 German sectors serving global markets, crisis 28 German trade union IG Metall reduction in working hours in metalworking industry 13 German work-sharing scheme, Kurzarbeit 1, 3, 9, 15, 16, 19, 24, 25, 31–3, 73, 74, 76–8, 227, 262, 265, 268, 269, 273–6, 279, 287 Germany, Kurzarbeit, expansion procedures 15 global financial crisis, 2008–2010 119, 203–59 United States, 2008, ‘shock wave’ 1 Gompers, Samuel, founding leader of American Federation of Labor (AFL) 222 government approval need in Austria, Netherlands, Turkey 277 government expansion of EAS, Japan 116 government labour-management agreement on work sharing, Japan, 2001–2002 106 government subsidies for hiring 216 Great Depression, 1930s 1, 10 Great Recession, 2008–2009 1, 259 in Japan 104

M3125 - MESSENGER TEXT.indd 303

303

in the US, effect of 204–5, 227–32 work-sharing measures 16 gross domestic product, Germany, 2006–2011 25–6 gross value added (GVA), wage bill weight 171 grounds of suspension, unemployment insurance, Uruguay 174 Hak-Is Trade Union Confederation pro-Islamist 121–2, 144 health services correlations between weekly hours and employment levels 234 high unemployment rates, Turkey 141 higher non-wage labour costs 215 ‘hoarding’ of labour 213 Hoover Administration, USA 11 hourly productivity 214 hours and employment relationship 235 hours, flexibilization of 218 hours of partial unemployment ‘authorized’ and paid, France 89 Hours of Work (Commerce and Offices) Convention, 1930 (No. 30) primary objectives 10 Hours of Work (Industry) Convention, 1919 (No. 1) 10 hours reduction, employment effects dependence on industry (goods or services production) 218 dependence on magnitude or location (country) 218 dependence on occupation (white collar or blue collar) 218 hours reduction, forms of 217–18 household debt in United States, effects on Germany 24 Household Labour Force (HLF) Surveys, Turkey 129 housing price decreases, Germany 58 human capital theory 38, 46, 51 human resource management flexibility 30 IAB Establishment Panel Survey data 30–31 IAB Establishment Panel, unique to Germany 41, 43, 45, 46

14/05/2013 13:33

304

Index

ideal programme design for crisis work-sharing programme 278–81 ILO, see International Labour Organization implementation lag 171–2 impulse responses durables manufacturing, contractionary periods between 1978 and 2011 244 durables manufacturing, expansionary periods between 1978 and 2011 246 of employment (P) 237, 238 of hours (H) 236 non-durables manufacturing, contractionary periods between 1978 and 2011 245 non-durables manufacturing, expansionary periods between 1978 and 2011 247 impulse responses to shocks in output and hours durables manufacturing, 1990–1911 239 durables manufacturing, expansionary periods between 1990 and 2011 240 non-durables manufacturing, 1990–2011 242 non-durables manufacturing, expansionary periods between 1990 and 2011 243 impulse response of hours (H) 237 to shocks in output and hours 239 incentives, weakness for company participation, Uruguay 170–71 increase in non-regular employees, Japan 104 increased periods of paid leave 260 independent workers no unemployment insurance 159 individual and societal well-being work-sharing contribution 8–9 individual reduction in working hours, temporary 82 individual well-being 1, 4, 8, 19, 206, 260 ‘indivisibility of labour’ issue 7

M3125 - MESSENGER TEXT.indd 304

industrial relations in Turkey 141 industrial relations in Uruguay trade union PIT-CNT 153 industries, vulnerable 16 Instituto Nacional de Empleo y Formación Profesional – INEFOP 167 insurance, for unemployment 162 internal flexibility control 49 International Labour Organization (ILO) Forty-Hour Week Convention, 1935 (No. 47) 10–11 Global Jobs Pact 18 Hours of Work (Industry) Convention, 1919 (No. 1) 10 wage supplements 10 International Labour Review articles, 1931–1939 11 Istanbul Chamber of Industry (ISO) positive view of RWT 145 Japan growth rate of real GDP, unemployment rate 101 job preservation 222 no compensation for overtime hours 276 work-sharing measures 262 Japan Chamber of Commerce and Industry Japanese Trade Union Confederation (RENGO) tripartite agreement 16, 107, 108 Job Creation Act of 2012, US 251 job creation estimation 204 job creation in medium and large companies, Uruguay 169 job crisis, global 15 job insecurity, reduction in 204 ‘jobless recoveries’ in US 223, 228 job-preservation effects 1, 93 job retention advantages, Uruguay 166 Job Retraining Fund (Fondo de Reconversión Laboral – FRL), Uruguay 166–7

14/05/2013 13:33

Index job rotation 42 job sharing, individual arrangement 4 Kali-Law, potash and fertilizer industry, 1910 32 Keynesian ‘stimulus packages’ 2 knowledge workers 207–8 Korea International Labour Foundation 16 Kurzarbeit, German work-sharing scheme 31–40 Kurzarbeit in Austria 76–80, 270 labour costs 39–40, 211, 215 labour demand contraction, Uruguay 151, 170 Labour Foundation (Stichting van de Arbeid, STAR) 93 ‘labour hoarding’ 30 due to work sharing 70–71 establishment level, employment retention 54 retention of skilled labour despite slack demand 39 Labour Law, Turkey, on reduced working time 120, 130 labour market flexibility by hours, Germany 30 fragmentation, Turkey 123 policies 2, 206 problems in Turkey 141 response, Germany 25 work sharing, exceptional 34 labour movement in Turkey 120–21 labour productivity, labour retention 214–15 labour relations in Uruguay 155 labour unions, Japan 99 Latin America, introduction of work sharing 259 Law on Public Employee Membership in Trade Unions, Turkey 121 law reports from Turkey, case studies 140–41 layoff use 211 Japan 104 refraining from 15 legal conditions of work sharing 31–40 legal framework for temporary work none in Turkey 123

M3125 - MESSENGER TEXT.indd 305

305

legal minimum wage rate, Turkey Minimum Wage Determination Commission 122 legislation on collective bargaining, Turkey 121 legislation on industrial relations, Turkey 121 Lehman Brothers, collapse 1 long-term employment, Japan 99 long-term unemployment rise in Turkey 129 long working hours, negative 9 losses in economic crises 270 low coverage of unemployed in Uruguay reasons 160 low participation in workforce, Turkey 141 ‘lump of labour fallacy’ argument 7, 208 macroeconomic shock 1 macroeconomic stability concerns, Turkey 141 manufacturing and construction industries, Germany work-sharing options 44 manufacturing industry in Uruguay worst hit by world crisis 165 marginal costs of benefits different worker-retention or worksharing programmes 180 marginal productivity of labour (MPL) equal to wage rate 38 mass unemployment and 40-hour week Great Depression, 1930s 10–12 measures of short-time work Austria, Belgium, France, Netherlands 74–5 mechanical engineering and chemical industries, Germany hardest hit by crisis 53 mechanical industries, Germany 27–8 metal products, textiles, motor vehicles and equipment RWT concentration 136 MIC Labour Force Survey, Japan 104 micro-econometric analysis 49 Middle Class Tax Relief and Job Creation Act, United States 251

14/05/2013 13:33

306

Index

Minister for Social Affairs and Employment, Netherlands 92–3 monthly gross wage, Austria 77 National Employment and Vocational Training Institute, Uruguay 167 National Federation of Small Business Associations, Japan tripartite agreement 16, 107 National Industrial Recovery Act (NIRA), United States 11–12 national institutions in Turkey 122 national laws on work sharing 16 national work crisis-sharing measures, summary 287–98 natural disasters 77 Netherlands 91–4 experience of work sharing 72 high unemployment 12–13 requirement to participate 263 Short-Time Work and Part-Time Unemployment scheme 271 work sharing, decline of 30% working-time reduction initiatives new employment creation, by work sharing 10 New York State Department of Labor work sharing with STC, layoff aversions 225 non-work time creation 204 normal working time (Normalarbeitszeit) Austria, 1934 76, 77 North America, estimated employment effects 222–7 North America, reduction of layoffs, 1990s 13 Notification Pay, Turkey 133 oil crises, 1970s 116 oil price shocks, 1974, 1979 Turkey crisis 12, 120 Omnibus Law, 2011, Turkey 131, 143 on-the-job training (OJT), Japan 100, 115 open unemployment by quarters, Turkey 127–8 Oregon, one-for-one layoff conversion rate 225

M3125 - MESSENGER TEXT.indd 306

output decline less need for labour input 209 variation 7, 209, 232 overall stimulus package costs, Turkey 130 overtime 42 hours reduction 116, 222 Japan, estimated numbers of employees, 2008–2009 105–6 market fluctuations 99 premium pay rate 112 unpaid 113 overtime leakage (overtime hours increase) protection of employment 101 reduction of job creation 212 paid annual leave, Japan 113–14 paid leave untaken and potential job creation, Japan 113 paid work ‘crowding out’ of parenting, civic activity, educational activities 206 parental leave time in US 251 parenting, ‘crowded out’ by work 206 ‘Pareto improving’ 206 partial income replacement (STC) 252 partial unemployment Belgium 81 benefits 18 circumstances for use, temporary and exceptional, France 87–91 Partial Unemployment Insurance (PIU), Uruguay 263 company size, in use of 169–70 comparative analysis 178–80 estimated costs, Uruguay 181–2 Job Retraining Fund 271–2 legal conditions to qualify for 168 low participation in sectors eligible 168–9, 178, 276 main results of 167–9 unemployment insurance on suspension grounds 178–80 work-sharing subsidies 151, 165 partial wage compensation for temporary work sharing 205 part-time employees, Japan 106 part-time unemployment, Netherlands 92

14/05/2013 13:33

Index Part-Time Unemployment scheme, Netherlands 264 part-time work 6 increased numbers 222 in United States 228 part-year work in Netherlands 6 payroll expenses, reduction 8 payroll taxes, reduced 216 pension contributions, Japan 115 performance-related payment 42 permanent overtime reductions, Japan 112–14 permanent work hour reduction 4, 6, 203 persons receiving partial unemployment benefits by year, France 89 persons receiving training while on short-time work subsidies, Austria 79 policies on work sharing, Uruguay 164–7 policy of work hours reduction or restraint absence of, in US 223 Portugal, reduced separation rate 221 Post-Second World War era full employment 12–14 post-war employment practices, Japan three primary characteristics 99 potential employment response 204 preference-based reductions in working hours 260 preservation of jobs, Austria 76, 80 President’s Reemployment Agreement (PRA), USA, 1933 shorter workweek of 35 to 40 hours 11 private sector salaried workers no unemployment insurance 159 privatization programme, 1980s, Turkey 121 probit estimations for work sharing, 2009 and 2010 65–6 productive leisure 217 ‘productivity effect’ 7, 166 productivity growth, leading to hours reduction 218 productivity offset 213–15 programme designs 259 programme participation

M3125 - MESSENGER TEXT.indd 307

307

Europe, Japan, Turkey, Uruguay 273–8 programme targeting 261–5 programmes to preserve jobs 259–60 protection against dismissal 91 public financial support 48, 73 public sector employees no unemployment insurance 159 public sector subsidization 218 PUI, see Partial Unemployment Insurance Quebec, reduction of standard workweek to stimulate employment growth 224 rate of layoffs and discharges, US recession 228 rates of employment reduction, Japan construction industry, medical and welfare service industry 104 rates of underemployment by industry, US shorter hours prevalent 231 real labour costs per hour 210–11 reasons for companies’ negative interest, Uruguay 173–4 recession, 2008, 2009 72 impact 205 in OECD countries 2 reduced hours and earnings, extent of Germany, Austria, Uruguay, Netherlands 265–6 reduced standard workweeks 260 Reduced Working-Time (RWT) programme, Turkey 119–20, 130–41, 263 among social partners 145 application process 132 concentration in metal products, textiles, motor vehicles 136 coverage extension, Turkey 133 employment gains 134 expenditures of participating workers, 2009–2010 134 final form, Omnibus Law, 2011 130 importance and success of 146–7 key programme for social cohesion 145

14/05/2013 13:33

308

Index

participants, by age and sex composition, 2009–2010 138 programmes, sectoral distribution of persons, Turkey, 2009–2010 137 selected indicators, Turkey 135 reduction of working hours 2, 3–5 net employment effect, estimate 219 reductions in standard workweek, France based on laws, Aubry I and II 220–21 reductions in working hours, benefits 248–9 redundancies, protection from 90 redundancy behaviour of companies, 1996–2004, France 90 regular and non-regular employees, Japan 116 Republic of Korea, work sharing 16 retail trade correlations between weekly hours and employment levels 234 Rhode Island State Labor Department work-sharing claims, high level 226–7 rise in unit labour costs, reinforcement 211 ‘risk sharing’ 206 routine tasks in manufacturing 208 retail trade 208 rural economy in Turkey 128–9 RWT, see Reduced Working Time safeguarding effect of work sharing, Germany 54 seasonal factors, no recognition 132 seasonal work sharing in construction sector 32–3 in Germany 262 Second World War 12 sector-level bargaining, Uruguay 155 seniority wage practices, Japan 99 length of service wages 100 sense of identity from paid work 8 service sector and food industry work-sharing options 44 services sector labour demand in Uruguay 151

M3125 - MESSENGER TEXT.indd 308

severance payments 51 Turkey 133 share of establishments using work sharing by size, Germany 44 shocks on employment Uruguay 155 short-term job retention policy 165 short-time allowances, beneficial if permanent 145 short-time compensation (STC) 5, 30, 148 temporary work sharing with wage supplements 205 short-time work 222 Austria, Netherlands 73 compensation in Germany (Kurzarbeitergeld: KuG) 33–6 Kurzarbeit, in Germany 3 Netherlands, obligations for training 91–4 requirements for 35–7 six months approval 76 subsidies, persons receiving 78 skill mismatches difference between employed and unemployed 212–13 skill shortages 7, 212–13 skill training 2 skilled labour positions 207–8 skilled labour shortage 30 skilled work forces, retention of 4 skills upgrading 6 ‘Slack Work or Business Conditions’ US definition of US Bureau of Labor Statistics 228 social dialogue in Turkey 122 social insurance contributions, Austria 77 social insurance of short-time workers, Germany 15 social isolation of unemployment 72 social partners (trade unions and employer federation), Japan 88 social partners, views from interviews to assess results of PUI programme 173–4 ‘social partnerships’ in Germany 42 agreement necessary 77

14/05/2013 13:33

Index social security contributions 18, 39, 52 Europe 93 premium payments, Turkey 145 premiums, reductions 143 Social Security Agency (Banco de Previsión Social – BPS), Uruguay eligibility grounds and financing 156–8 social security, unregistered, Uruguay 192 social welfare 204, 206 socialization opportunities of work 8 staff morale bolstering 4 STC, see short-time compensation stigma of unemployment 72 subsidization of training, Germany 47 subsidy for layoffs from EAS, Japan 111 to promote investment 130 worker retention 174 suspension and PUI beneficiaries by manufacturing industry, Uruguay 199 suspension grounds, use of, Uruguay 173–4 suspension UIBs by area of activity 202 Sweden collective agreement in manufacturing sector 16–17 target populations 259 tax incentives reduced payroll taxes, tax credits 6 tax reductions 130 taxes on overtime pay, elimination 14 temporary agency work, decrease in France 90 temporary employment agencies, Germany 27–8 temporary labour slack retention of productive employees 51 temporary unemployment France, Belgium 73 justification for 81 subsidies Belgium yearly averages, by region 85–6 white-collar workers 82 temporary work reduction planned layoffs by employers 203

M3125 - MESSENGER TEXT.indd 309

309

termination of contracts under Labour Law, Turkey 133 terminology variations in countries 17 textile and tanning industries, evolution 165 textile, leather, clothing and automotive industries, France limit of 1,000 hours’ payment per employee, January 2009 88 ‘time-credit scheme’ Belgium 82 time limits, minimization of deadweight and displacement effects 269–71 time-series analysis 205, 235 time-series relationships, US case analysis between output, hours, employment and layoffs 232–48 Tobacco Tax Amendment Act, 1909, Germany 32 total sales revenues of a firm, Turkey 142 Toyota, Nissan, Hitachi, Fujitsu shortened scheduled hours, wage reductions 107 tractor vehicles, fall in sales orders 140 trade unions density fall in Turkey 121 laws in Turkey 121 Netherlands 91 social protection, job security, wage increases 143 trades distribution in RWT programmes, Turkey 137 training activities 6, 42, 179–80 training and retraining for work sharing participants 271–2 training components and work-sharing policies incompatibility 272 training for work, Uruguay 166 training leave (congé individuel de formation) France remuneration of 80% of net salary 88 training, obstacles to, Germany lack of training programmes 47 training on the job, subsidization of, Germany 45

14/05/2013 13:33

310

Index

transaction costs 170 Transfer-Kurzarbeit for permanent loss of employment 32 Germany for German reunification 262 transport and warehousing correlations between weekly hours and employment levels 234 tripartite agreement components 108 Turkey 6-day working week, Fridays as holiday 120 ‘discouraged’ workers 126 post-2000 123–9 ‘reduced working time’ Sakarya Labour Court cases 140 temporary time-frame, three to twelve months 17 Turk-Is (Confederation of Turkish Labour Unions) 143 American unionization model 121 Turkish Confederation of Employers’ Unions (TISK), 122 Turkish employers, large, hesitance to engage 277 Turkish Employment Agency (ISKUR) 120 in favour of recent amendments to RWT 145 on short-time work eligibility inclusions 145 Turkish experience in work sharing during global economic crisis, 2008–2010 119 Turkish labour market development 125–6 impact of global crisis 126–9 informal unregistered employment 138 Turkish private manufacturing, productivity and real wages 124 UIBs, see Unemployment Insurance Beneficiaries unemployed persons, estimated potential based on EAS, Japan 110 unemployment decline in Germany 56

M3125 - MESSENGER TEXT.indd 310

increase 2 Japan 99 unemployment after work sharing 37–8 unemployment benefit schemes extension of 18 unemployment benefits 33 dismissal, work days or hours reduction work activities suspension 162–3 Netherlands 91 Unemployment Insurance Beneficiaries (UIBs) percentage in total unemployed 159 Turkey, eligibility for 132–3 unemployment insurance beneficiaries (UIBs) annual average, Uruguay 191 by activity (2010) 177 new, by grounds or option 195 number, by unemployment 196–7 on suspension grounds by activity 177 on suspension grounds by age 176 on suspension grounds by gender 175 total number of new (annual monthly average) 163 unemployment insurance beneficiaries, Uruguay 160 unemployment insurance compensation understaffing, Turkey 146 Unemployment Insurance Fund, Turkey, 1999 122, 144 unemployment insurance scheme, Uruguay work-sharing measures 155–64 unemployment insurance system 34 (UI) payments 5 unemployment rate by gender and age group, Uruguay 189 harmonized, 2008–2010 27 OECD countries 29 rise, 2008–2010, European Union and United States 27 union activity from 2005 in Uruguay 153 union resistance, in Netherlands 92–3 unionization rates in Uruguay 153

14/05/2013 13:33

Index unit costs and wage costs 8 unit labour costs (ULC) 7–8, 211, 212, 216 United States work sharing shortening of working hours 11 unpaid overtime in Japan 116 Urban Business Federation (Nippon Keidanren) tripartite agreement 16, 107 Uruguay, 2009–2010 151 Uruguay, drop in company activities 263 Uruguay situation, work-sharing policies 182–4 Uruguay’s partial unemployment insurance work-sharing scheme 164–75 utilities correlations between weekly hours and employment levels 234 Venezuela, 1991 blue-collar workers, workweek reduction 221 vocational training, extensive systems 216 vocational training subsidy, Japan 111 wage agreements, Uruguay 155 compensation 236 or salary reduction 5 replacement payment 81 subsidy 211 supplements to affected workers 266−9 wage supplements 5 affected workers, provision of, Europe, Turkey, Uruguay 266–9 wage-setting policy, Uruguay 153 wages in Japan 116 Weimar Republic, Germany, 1918, unemployment compensation 10 welfare gain, to maintain employment status, Turkey 148 well-being spread through work sharing 204 well-being spreading, from work spreading 248–9

M3125 - MESSENGER TEXT.indd 311

311

Werktijdverkorting, Netherlands 91, 92 protection against dismissal 270 Western European countries working-time reductions, 1970s 12 white-collar unions, Belgium 82 white-collar workers, blue-collar workers, Japan 107 wholesale trade, correlations between weekly hours and employment levels 234 women and young people’s employment rates, Uruguay 154 work environment in Turkey 141 workday flexibility mechanisms, none for employers 171 worker fatigue, reduction of 213 worker retention programme, Uruguay 151 worker substitution, potential problem 207–8 workers’ hours, reduction in Japan 222 workers on unemployed insurance, characteristics, Uruguay 161–2 workforces, large 51 ‘Working Families Flexibility Act’ (H.R. 1274), US 251 working-hour reductions individual, Belgium 86 in Japan, 2008–2010 rate of employment maintenance 100, 102–3 working hours reduction 5, 6, 18 eight-hour day 10 working hours, temporary collective reduction 82 working housewives in Japan 115 working-time account balances 43 employees with 50 reduction policies 223 work–life balance, improving 204 work reorganization 218 work sharing advantages and disadvantages 34, 62 employer’s perspective of work sharing 62–4

14/05/2013 13:33

312

Index

alternative to layoffs, German experience 24 compensation 31 countries 29 crisis measure, European Union 72 definitions and objectives 3 during Great Recession 15–18, 48 economic reasons, Germany 33, 40–41 effective job-creation tool 281–4 government policy, permanent working time reductions 4 history, origin in Germany (Kurzarbeit) (2007–2011) 9–10, 40 impact of, Germany and Japan 31 intensity estimation 66–7 Japan 99 job preservation tool 5, 72 labour hoarding 56 middle-income countries 17–18 ‘permanent’, for job creation 6 role in employment adjustment 54 safeguarding effect 56–9 social welfare boost in average labour 248–9 subsidy programmes differences with PUI work-sharing scheme 178–9

M3125 - MESSENGER TEXT.indd 312

tool for creating employment 203 Transfer-Kurzarbeit, Germany 262 Turkey 120 United States 228 work-sharing measures Canada and US 13 indirect effects 7 partial, reduction and suspension estimated impact, Uruguay 180–81 range of results 7 targeting worker types and industries 250 Uruguay 151 Work-Sharing Program in Canada, 1990–1991, 2001–2002 13 majority in manufacturing sector 226 work stress and fatigue, reduction in 204 work-suspension subsidy, Uruguay 163–4 work-time reduction in North America 204 work training during work sharing, 2009–2010, Germany 46 workweeks, full-time, longer in US than in Europe 223 worldwide recession 203

14/05/2013 13:33