India’s Contemporary Macroeconomic Themes: Looking Beyond 2020 (India Studies in Business and Economics) 9819957273, 9789819957279

This book extensively examines various contemporary macroeconomic themes of India, namely growth and macro policies, tax

126 2 12MB

English Pages 624 [602] Year 2023

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Foreword
Celebrating 90 Years of a Visionary Economist, Dr. C. Rangarajan
Speech of Dr. Manmohan Singh, Former Prime Minister at the Conference of Madras School of Economics to Honor Dr. Rangarajan on his 90th Birth Anniversary on 21–22 April, 2023
Contents
Editors and Contributors
1 Introducing the Volume on India’s Contemporary Macroeocnomic Themes-Looking Beyond 2020
1.1 Introduction
1.2 Overviews of Themes and Chapters
Part I Growth and Macro Policies
2 The Indian Economy in the Post-pandemic World: Opportunities and Challenges
2.1 Introduction
2.2 Pre-pandemic Economic Conditions
2.2.1 Aggregate Macro Conditions
2.2.2 Agriculture, Informal Sector and MSMEs
2.2.3 Formal Sector
2.3 Post-pandemic Economic Conditions
2.3.1 Impact of Covid-19 Pandemic
2.3.2 Aggregate Macroeconomic Conditions
2.3.3 Formal Sector Conditions
2.4 Opportunities and Challenges Going Forward
2.4.1 Short and Medium Term Challenges
2.4.2 Engines of Growth
2.5 State Finances
2.6 Freebies
2.6.1 Structural Transformation
2.6.2 Climate Change
References
3 India’s Economy in the Twenty-First Century: Role of State-Differentiated Demographic Dividend
3.1 Introduction
3.2 Size of the Indian Economy in the Twenty-First Century: Review of Recent Studies
3.3 India’s US$26 Trillion Economy by 2047–48: Assumptions and Findings (EY Study)
3.4 Interdependence of Saving, Investment, and the Demographic Dividend
3.5 State-Differentiated Demographic Dividend
3.6 Messages for Policy Formulation
3.7 Conclusions
Appendix
References
4 Post-covid Fiscal Recovery in India: Uncertainty, Growth, and Fiscal Prudence
4.1 Introduction
4.2 Growth Outlook: Various Estimates
4.2.1 Post Covid Global Growth Recovery: A Comparative Picture
4.2.2 Why Growing at More Than 6% is Critical?
4.3 Fiscal Imbalance
4.3.1 The General Government Debt
4.4 Fiscal Stance for Higher Capital Expenditure
4.5 Framework of Fiscal Responsibility
4.6 Conclusions
References
Part II State Finances and Intergovernmental Fiscal Transfers
5 The Relationship Between Government Revenue, Government Expenditure and Economic Growth in India: An Empirical Investigation at the Sub-national Level
5.1 Introduction
5.2 Literature Review
5.2.1 Nexus Between Government Revenue and Government Expenditure
5.2.2 Relationship Between Government Revenue and Economic Growth
5.2.3 Association Between Government Expenditure and Economic Growth
5.2.4 Factors Driving Public Expenditure
5.3 Stylised Facts
5.4 Data and Methodology
5.4.1 Data
5.4.2 Methodology
5.5 Empirical Estimation and Results
5.6 Concluding Observations
Statistical Annex
References
6 Revenue Implications of GST on Indian State Finances
6.1 Introduction
6.2 State of the Economy
6.3 Fiscal Health of Indian States
6.4 Revenue Performance of GST
6.5 Revenue Implications of GST on Indian State Finances
6.5.1 Comparison of Revenue Under Protection with State GST
6.5.2 Analysis State Tax Revenue: Pre- Versus Post-GST
6.5.3 Fiscal Health of States: Pre- Versus Post-GST
6.6 Summary and Conclusions
Appendix
References
7 Equalization Transfers Policy Based on Expenditure Needs and Own Revenue Capacity of Indian State Governments
7.1 Introduction
7.2 Literature Review
7.3 Empirical Model, Data, and Estimation
7.4 Empirical Results
7.4.1 Estimation Results of State Governments’ Revenue Expenditures
7.4.2 Estimation Results of State Governments’ Own Revenues
7.4.3 State-Wise Efficiency Scores
7.4.4 Determining Fiscal Equalization Transfers
7.5 Summary and Policy Conclusion
References
Part III Fiscal Reforms
8 Goods and Services Tax in India: A Stocktaking
8.1 Introduction
8.2 Salient Features of Indian GST
8.3 Impact of GST on Minimizing Distortions and Revenue Productivity
8.4 Multiplicity of Rates: Complications, Distortions, and Inverted Duty Structure
8.5 The Reform Issues and Strategy for Implementation
8.6 In Conclusion
References
9 Recent Reforms in India’s Corporate Income Tax Regime: Rationale, Impacts, and Improvements
9.1 Introduction
9.2 Literature Review
9.3 Tax Rate Evolution in India: Stylized Facts and Analysis of Effective Tax Rates
9.3.1 Basic Tax Rates
9.3.2 Minimum Alternate Tax
9.3.3 Effective Corporate Tax Rates
9.4 Recent Optional Lower Tax Rate Regime
9.5 User Cost of Capital Method and Analysis
9.5.1 Methodology
9.5.2 Industry-Level User Costs of Capital for Basic Tax Rates
9.5.3 Impact of Optional Tax Rates
9.5.4 Loss-Making Companies
9.6 Conclusions, Caveats, and Policy Recommendations
9.7 Annexure 1
9.8 Annexure 2
References
10 The Anatomy of Public Debt Reductions: Case of India
References
11 Measuring Tax Impact on Corporate Dividend Behavior in India
11.1 Introduction
11.1.1 Recent Developments Pertaining to the Dividend Taxation
11.2 Literature
11.2.1 The Lintner Study
11.2.2 Studies Testing Tax Impact on Dividend Behavior
11.2.3 Other Recent Studies
11.3 The Model and the Methodology
11.3.1 Inter-Temporal Adjustment
11.3.2 Interpretation of the Model in the Indian Context
11.3.3 Influence of Other Factors
11.4 Empirical Analysis
11.4.1 Regression Analysis (Cobb-Douglas Assumptions)
11.4.2 Regression Analysis (CES Assumptions)
11.5 Summary
References
Part IV Banking and Monetary Policy
12 Non-performing Assets of Indian Banking: An Evolutionary Journey
12.1 Introduction
12.2 Broad Trends of NPAs of Indian Banks
12.2.1 Some Definitional Issues
12.2.2 The Timeline and Identification of Twists and Turns
12.3 Phase I (1996–2008): Falling NPAs
12.4 Phase II (2008–09 Through 2017–18): Rising NPAs
12.4.1 Broad Trends
12.4.2 Regulatory Forbearance
12.4.3 Bad Fundamentals and Steep Fall of Commodity Prices
12.4.4 Corporate Sector Debt Problems
12.4.5 Corporate Governance and Corruption Related Issues
12.4.6 To Sum Up …
12.5 Improvements in NPAs Since 2018
12.6 Concluding Observations and the Way Ahead
References
13 Conduct of Monetary Policy in India: The Journey so Far and Contemporary Challenges
13.1 Introduction
13.2 Evolution of Thinking on the Role of Monetary Policy
13.2.1 Theoretical Foundation: Quantity Theory of Money
13.2.2 Keynesian Revolution and Role of Monetary Policy
13.2.3 Monetarist View
13.2.4 AD-AS Model
13.3 Conduct of Monetary Policy in India: The Evolving Process
13.3.1 Credit Planning and Social Control of Banking Business
13.3.2 Monetary Targeting Framework
13.3.3 Multiple Indicators Approach
13.3.4 Inflation Targeting
13.4 Contemporary Issues and Challenges
13.4.1 Conflict in Objectives
13.4.2 Monetary Policy Response to Negative Supply Shocks
13.4.3 External Sector Dominance on the RBI’s Monetary Policy
13.4.4 Monetary Policy Without Money
13.5 Concluding Remarks
References
14 Macroeconomics of Digitalization—Evolving Issues and Perspectives
14.1 Introduction
14.2 Measurement of Digitalization:
14.3 Digitalization and Economic Growth
14.3.1 Supply Channel: Productivity Enhancements
14.3.2 Demand Channel—Boost in Consumption
14.4 Digitalization and Inflation
14.5 Digitalization and Financial Markets
14.6 Digitalization and Cash Adoption
14.7 Issues in Digitalization
14.7.1 Cyber Security, Data Protection, and Privacy
14.7.2 Socio-economic Challenges—Digital and Financial Exclusion
14.7.3 Labor Market Implications of Digitalization
14.8 Conclusion
References
15 Does Financial Frictions Matter for Monetary Policy Transmission in India?
15.1 Introduction
15.2 Related Literature
15.3 Analytical Framework
15.4 Data and Methodology
15.4.1 Data Description
15.4.2 Methodology
15.5 Empirical Analysis
15.5.1 Unit Root Tests & the BDS Test
15.5.2 Preliminary Analysis Using the SVAR Model
15.5.3 The MS-VAR Results
15.6 Conclusion and Policy Implications
References
16 Cash and Debt Management in India
16.1 The Indian Economy in the Eighties and Early Nineties
16.2 Monetary Policy and the Monetary-Fiscal Co-ordination
16.3 Stoppage of Automatic Monetization
16.4 Transparency in the Budgetary Operations
16.5 Administered Interest Rates and the Financial Sector Reforms
References
Part V Environment and Social Sector Policies
17 Role of Fiscal Policy in Climate Change Mitigation in India
17.1 Introduction
17.2 Economic Modeling of Climate Change
17.3 India’s Policies Toward Mitigation of Climate Change
17.4 Issues in Framing Mitigation Policies in a Macro Framework
17.5 Concluding Remarks
References
18 Ecological Fiscal Transfers and State-Level Budgetary Spending in India: Analyzing the Flypaper Effects
18.1 Introduction
18.2 Review of Literature
18.3 Interpreting Data
18.4 The Econometric Models and Results
18.4.1 Medium Dense Forest Cover Models
18.4.2 Very Dense Cover Models
18.5 Conclusion
Appendix 1: Dense Cover and Tax Devolution—Outlier Elimination
References
19 Measurement of Multidimensional Inequality of Opportunity in India
19.1 Introduction
19.2 Relevant Literature
19.3 Methodology
19.3.1 Econometric Specification
19.3.2 Identification
19.3.3 Estimation
19.3.4 Discussion of Identification
19.4 Data
19.5 Results
19.5.1 Presence of Heteroskedasticity
19.5.2 Estimation Results
19.5.3 Contributions of Circumstances and Efforts to Inequality in Work Wellbeing
19.6 Conclusion
19.7 Appendix
19.8 Tables
19.8.1 Estimation Results for the Wellbeing Equation of the SEM Using Klein and Vella (2010) Procedure of Identification
19.8.2 Estimation Results for an Income—Monthly Per Capita Expenditure—Equation
19.8.3 Check for Heteroskedasticity
19.8.4 Shares
References
20 Youth Labor Market Challenges in India: Education, Employment, and Sustainable Development Goals
20.1 Introduction
20.2 Labor Market Scenario in India and the Youth
20.2.1 Structural Transformation and Its Reversal
20.2.2 Rising Youth Population, But Inadequate Growth of Workforce
20.2.3 Sectoral Decomposition of Youth Employment
20.2.4 Youth Work Participation: Role of Education and Training
20.2.5 Youth Active Job Seekers and Passive Unemployment
20.3 Effectiveness of the Past Skill Development and Employment Generation Measures
20.4 Policies to Achieve Sustainable Development Goals Through Creation of Jobs
20.5 Conclusion
References
Part VI Emerging Economic and Policy Challenges
21 Forecasting State-Level Fiscal Imbalances in India
21.1 Introduction
21.2 Review of Literature
21.3 Data Characteristics and Methodology
21.3.1 Data Characteristics and Sample Period
21.3.2 Model Specification
21.4 Results: Estimated Model
21.4.1 State-Level Real GSDP
21.4.2 State-Level Implicit Price Deflators
21.4.3 State-Level Total Expenditure
21.4.4 State-Level Effective Interest Rate
21.5 Model Validation and Sample Period Estimation
21.6 Forecasting State-Wise Fiscal Imbalances
21.7 Conclusion
Appendix
References
22 Machine Learning in Macroeconomics: Application to DSGE Models
22.1 Introduction
22.2 What is Machine Learning and What Are Its Main Components?
22.3 Machine Learning and Macroeconomics: A Brief Overview
22.4 Techniques of Machine Learning
22.4.1 Relevant Machine Learning Models
22.4.2 Random Forests
22.4.3 Support Vector Machine
22.4.4 Relevance of Random Forests and SVM in Economic Modeling
22.5 Evaluation of Machine Learning Techniques
22.6 DSGE Models and Machine Learning—rationale and Findings
22.6.1 How Are Machine Learning Methods Applied to DSGE Classification?
22.7 Conclusion
References
23 Friends with Benefits: The Role of Internal Capital Markets During Financial Stress
23.1 Introduction
23.2 Literature Review
23.3 Data and Methodology
23.3.1 Data Description
23.3.2 Methodology
23.4 Results and Discussions
23.4.1 Results of Base Models
23.4.2 Result of Stand-Alone Versus Group Member Model
23.4.3 Result of Net Transfer Model
23.5 Conclusion
Appendix
References
24 Reforming the Indian Census
24.1 Introduction
24.2 Background
24.3 Census Schedules
24.4 Concluding Remarks
References
25 The Performance of India’s Merchandise Exports: An Analytical Perspective
25.1 Introduction
25.2 India’s Export Performance
25.2.1 Growth and Composition of India’s Exports: A Disaggregated Sectoral View
25.2.2 India’s Exports in the Global Context
25.2.3 Direction of India’s Exports
25.2.4 Import Intensity of India’s Exports
25.3 Determinants of Exports
25.3.1 Determinants of India’s Export: Review of Recent Literature
25.3.2 Determinants of India’s Merchandise Export: An Empirical Investigation
25.4 Summary and Conclusions
Annexure 1: Vector Error Correction Estimates
References
Recommend Papers

India’s Contemporary Macroeconomic Themes: Looking Beyond 2020 (India Studies in Business and Economics)
 9819957273, 9789819957279

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

India Studies in Business and Economics

D. K. Srivastava K. R. Shanmugam   Editors

India’s Contemporary Macroeconomic Themes Looking Beyond 2020 Foreword by Dr. Manmohan Singh

India Studies in Business and Economics

The Indian economy is one of the fastest growing economies of the world with India being an important G-20 member. Ever since the Indian economy made its presence felt on the global platform, the research community is now even more interested in studying and analyzing what India has to offer. This series aims to bring forth the latest studies and research about India from the areas of economics, business, and management science, with strong social science linkages. The titles featured in this series present rigorous empirical research, often accompanied by policy recommendations, evoke and evaluate various aspects of the economy and the business and management landscape in India, with a special focus on India’s relationship with the world in terms of business and trade. The series also tracks research on India’s position on social issues, on health, on politics, on agriculture, on rights, and many such topics which directly or indirectly affect sustainable growth of the country. Review Process The proposal for each volume undergoes at least two double blind peer review where a detailed concept note along with extended chapter abstracts and a sample chapter is peer reviewed by experienced academics. The reviews can be more detailed if recommended by reviewers. Ethical Compliance The series follows the Ethics Statement found in the Springer standard guidelines here. https://www.springer.com/us/authors-editors/journal-author/journal-aut hor-helpdesk/before-you-start/before-you-start/1330#c14214

D. K. Srivastava · K. R. Shanmugam Editors

India’s Contemporary Macroeconomic Themes Looking Beyond 2020

Foreword by Dr. Manmohan Singh

Editors D. K. Srivastava Ernst & Young India New Delhi, Delhi, India

K. R. Shanmugam Madras School of Economics Chennai, Tamil Nadu, India

ISSN 2198-0012 ISSN 2198-0020 (electronic) India Studies in Business and Economics ISBN 978-981-99-5727-9 ISBN 978-981-99-5728-6 (eBook) https://doi.org/10.1007/978-981-99-5728-6 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd. The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721, Singapore Paper in this product is recyclable.

Dedicated to Dr. C. Rangarajan, Celebrating 90 years of a visionary economist

Foreword

Dr. C. Rangarajan’s 90th Birth Anniversary is an occasion for us to reflect on his immense contribution towards the Indian economy and the monetary policy of our country. He is one of the rare economists who are capable of straddling the two worlds of academia and government. He has completed 90 fulfilling years. This journey has taken him around and at the peaks of the financial and economic world globally. I have not known anyone more dependable, more impartial, and more valuable in the advice he gave as the Chairman of the Prime Minister’s Economic Advisory Council. As the Governor of the Reserve Bank of India, he exerted profound influence in the formulation of monetary policy and exchange rate policies. Dr. Rangarajan taught at the Wharton School of Finance and Commerce, Graduate School of Business Administration, New York University, and IIM, Ahmedabad. While chairing various committees and commissions, he applied academic principles to practical problems and found undisputed solutions. Growth and inclusion are two sides of the same coin. Dr. Rangarajan understood this and wove these together to support a cohesive pattern of growth that generated additional resources to be utilized to meet the goals of socio-economic equity and greater inclusion. Computerization of bank branch operations was a far-reaching reform that was led by Dr. Rangarajan. This book examines various contemporary macroeconomic themes of India: growth and macro policies, fiscal reforms, government finances and intergovernmental fiscal transfers, banking and monetary policy, and environmental and social sector policies. Forty-four eminent economists who have had the privilege of working with and learning from Dr. C. Rangarajan have contributed 25 chapters.

vii

viii

Foreword

I am sure that this book will serve as an excellent reference for professionals such as policymakers and financial analysts, students, and other stakeholders interested in the field of economics and finance.

Dr. Manmohan Singh Former Prime Minister of India Eminent Economist and Architect of the 1991 Economic Reforms of India New Delhi, India

Celebrating 90 Years of a Visionary Economist, Dr. C. Rangarajan

India is currently poised to lead growth in the world economy in the twenty-first century. It can arguably be said that the country has arrived at this advantageous position due to the economic reforms that were undertaken in the early 1990s. A major architect of those reforms was Dr. C. Rangarajan, who occupied critical positions for about 25 years following the launch of the reforms—first as Governor of the Reserve Bank of India (RBI), covering the period from 1992 to 1997; then as Chairman, Twelfth Finance Commission (FC12) from 2002 to 2004; and subsequently as Chairman of the Economic Advisory Council to the Prime Minister from 2005 to 2014.1 In the sphere of economic reforms, Dr. Rangarajan made significant contributions in several dimensions of workings of the economy, which continue to serve well in contemporary India. First, during his tenure at the RBI, he worked toward gradually making India move to a regime of flexible exchange rates, the beginnings of which were made by two successive devaluations in the early 1990s. This was to serve as a major tool to make India more competitive in the world economy while India started to de-license and de-regulate its industrial sectors. Second, in the mid-1990s, Dr. Rangarajan successfully brought the Ministry of Finance and the RBI together to agree to a discontinuation of the practice of issuing ad-hoc treasury bills to fill up shortfalls in government revenues to finance its expenditures. This major move facilitated, in due course, the entire framework of Fiscal Responsibility Legislations (FRLs) for the central and state governments from 2003 onwards. It was at about this time that Dr. Rangarajan became Chairman of FC12. It was this Commission, which complemented the Centre’s Fiscal Responsibility and Budget Management Act (FRBMA) of 2003 by preparing the framework for extending the provisions of Centre’s FRL to the state governments and for the combined account of central and state governments. The entire framework of the combined debt and deficit relative 1

According to Dr. Y. V. Reddy (in his message to the Conference, honoring and celebrating the 90th birth anniversary of Dr. C. Rangarajan, April 21–22, 2023), “Dr. Rangarajan is among the foremost economists of the country, who has had a profound, lasting and decisive impact in charting the economic direction of the country over the last quarter century.” ix

x

Celebrating 90 Years of a Visionary Economist, Dr. C. Rangarajan

to GDP was linked to the saving and investment profiles of the broader economy, thereby weaving together the government and the private sectors in one logically consistent underlying framework. The FC12 also provided a major incentive scheme of debt rescheduling, which persuaded the state governments to agree to the conditions specified in the debt sustainability framework in enacting their individual FRLs. In the years following the FC12 recommendations, eventually all states enacted their FRLs. The FC12 also laid down a framework for guiding fiscal transfers from the center to the states under an equalization principle with an emphasis on equalizing health and education services. States showed revenue surplus from 2005 to 2016 onwards and the debt-GDP ratio started declining for both the center and the states. As Chairman of the Economic Advisory Council to the Prime Minister of India (2005–2014), Dr. Rangarajan established the practice of bringing out two publications each year, providing an overview of the developments in the Indian economy. This was an innovation, which gave insight into the workings of the Advisory Council. During these years, several critical recommendations that Dr. Rangarajan was able to formulate, while heading various committees and commissions, made substantive contributions to the ongoing economic reforms in India. These initiatives facilitated the spread of financial inclusion to lower income households with a view to making a dent in the incidence of poverty in India. He also made salient methodological contributions for the measurement of the incidence of poverty in India. Dr. Rangarajan also became instrumental in providing a solid foundation to India’s statistical system in his capacity as the Chairman of the National Statistical Commission in 2000–2001.2 He also headed the High-Level Expert Committee on the management of Public Expenditure, which recommended the concept of effective revenue deficit with a view to encouraging the formation of capital assets at the subnational level. Some of the other prominent Committees and Commissions that he headed include Task force on Jammu and Kashmir, constituted by the Prime Minister (2005–2006), Committee on Financial Inclusion (2006–2008), High level Committee on Estimation of Saving and Investment (2007–2009), Expert Group to Review the Methodology for Measuring Poverty (2012), and Committee on Production Sharing Contracts in Hydrocarbon Exploration (2012). He also chaired the Committee to review the functioning of the sugar industry and made important recommendations for the decontrol of the sector. In May 2020, the Tamil Nadu government formed a High-Level Committee, chaired by Dr. Rangarajan, to advise the government on the medium-term response after COVID-19 lockdown period.3 2

He also chaired the Committee on Balance of Payment (1991) which laid a firm basis for our external sector management. He also chaired the Committee on Disinvestment (1992) which defined the rationale and framework for government disengagement from public enterprises. 3 Dr. Y. V. Reddy (in his message in 2023) said that Dr. Rangarajan’s presidential address at the Economic Association on “Issues in Monetary Management” (1988) is considered a landmark statement. His Kutty Memorial Lecture on “Autonomy of Central Banks” (1993) was decisive in putting an end to the 40-year flawed system of monetization of fiscal deficit. His Anantha Ramakrishnan Memorial Lecture on “Dimensions of Monetary Policy” (1997) set out the rationale for a strong and unswerving monetary policy and the need for identifying the central objectives of monetary policy which has to be the control of inflation. His monograph on the “Inter-relationship

Celebrating 90 Years of a Visionary Economist, Dr. C. Rangarajan

xi

Earlier, Dr. Rangarajan was entrusted with several apex positions of responsibility, including the governorship of Andhra Pradesh (1997), Odisha (1998–1999), and Tamil Nadu (2001–2002). In these years, Dr. Rangarajan, along with his wife, Dr. Haripriya Rangarajan who was a well-known iconographer and historian, made salient contributions to the cultural heritage and the economic landscape of these states. For a limited period, Dr. Rangarajan was a member of the Rajya Sabha during 2008–2009. He made well-appreciated contributions to the discussions in the Rajya Sabha on economic issues pertaining to growth, inflation, and fiscal imbalances, especially in the context of global economic and financial crisis. Dr. Rangarajan’s contribution as an academician and a policymaker was substantially enhanced by his playing an active role in strengthening eminent institutions of research and teaching, such as the National Institute of Public Finance and Policy, New Delhi, and the Madras School of Economics, Chennai, which gave an opportunity to a number of economists and research scholars to contribute to India’s ongoing policy debates. These institutions became so well-recognized in the economic research and policy-making landscape of the country that several of their faculty members were entrusted, from time to time, with positions of responsibility as Members of the Finance Commissions, Planning Commission, and the Economic Advisory Council to the Prime Minister. Dr. Rangarajan’s early academic life was characterized by salient accomplishments and distinctions.4 He was associated with institutions of repute such as the Wharton School of Finance and Commerce, University of Pennsylvania; Graduate School of Business Administration, New York University; Indian Institute of Management, Ahmedabad; Indian Statistical Institute, New Delhi; and University of Rajasthan, Jaipur. He was also a Fellow at the International Food Policy Research Institute, Washington for some time. Some of the other institutions of excellence have benefited from his guidance and supervision at the apex level, including C. R. Rao Institute of Mathematics, Statistics and Computer Science, Hyderabad, University of Hyderabad (2015 to date), and ICFAI Foundation for Higher Education, Hyderabad (2015 to date). Dr. Rangarajan was President of the Indian Economic Association in 1988 and of the Indian Econometric Society in 1994. He was the Conference President of the Indian Economic Association in the centenary year of 2017. Besides this, he was also President, Indian Statistical Institute, Kolkata; Chairman, GENOME Foundation; and President, CUTS Institute for Regulation and Competition. Dr. Rangarajan has recently reflected upon his long and exemplary journey of academic achievements and contributions to India’s economic reforms by putting together his experiences in the form of his memoirs titled “Forks in the Road: My between Agricultural Growth and Industrial Development” (1982) is cited even today as standard reference which testifies to its seminal value. His works “Short term Investment Forecasting (1974) and Innovations in Banking (1982) are highly quoted. 4 Dr. Rangarajan graduated with an Honors degree from University of Madras and he obtained his Ph.D. degree in Economics from University of Pennsylvania. He was conferred D.Sc. Honoris Causa Degree by University of Hyderabad in 2010. He also received the Honorary doctorate from several institutions.

xii

Celebrating 90 Years of a Visionary Economist, Dr. C. Rangarajan

Days at the RBI and Beyond.” In his own assessment, the opportunities that came his way during these years were partly an outcome of his choices and partly that of chance. It is only natural that through his long and distinguished career, Dr. Rangarajan has been honored by numerous awards. In recognition of his contribution to India’s economic policy-making, he was conferred Padma Vibhushan in 2002. In 2020, Dr. Rangarajan was conferred the P. C. Mahalanobis Lifetime Achievement Award in recognition of his contribution to official statistics. Among others, he received the Lifetime Achievement Awards from Financial Express and Business World both in 2010. In 2007, he was given the SKOCH Lifetime Achievement Award as also the Ravi J. Mathai National Fellow Award. He was also the recipient of the Lifetime Achievement Award by Corporate Affair Ministry in 2009. He was awarded the Hall of Fame Award 2008 by Outlook Magazine. Earlier in 2002, he received the Alumni award for outstanding leadership from the Wharton India Economic Forum, University of Pennsylvania. In spite of handling various responsibilities of an administrative nature from time to time, Dr. Rangarajan never deviated from the path of being a devoted academician. This is reflected in the long and consistent list of his academic contributions in the form of books, articles, monographs, working papers, and newspaper columns. Some of his well-known books include the most recent one, namely, “Forks in the Road” (2022), “Counting the Poor in India: Where Do We Stand” (2016), “Federalism and Fiscal Transfers in India” (2011), “Principles of Macroeconomics” (1979), and “Indian Economy: Essays on Money and Finance” (1998). It is only befitting that forty-four eminent economists, most of whom had the privilege of working with, and learning from Dr. Rangarajan, have contributed to this volume titled, India’s Contemporary Macroeconomic Themes: Looking Beyond 2020 published by Springer Nature. This volume is the outcome of papers presented by these authors in a Conference organized at Madras School of Economics for “Honoring and Celebrating the 90th Birth Anniversary of Dr. C. Rangarajan” on 21 and 22 April 2023.5 The articles contained in this volume reflect not only on India’s current economic and policy challenges but also on its future possibilities and potential. New Delhi, India Chennai, India

5

D. K. Srivastava K. R. Shanmugam

The conference was graced by messages and tributes by eminent public figures and economists including Dr. Manmohan Singh, former PM of India, Shri N. R. Singh, chairman of FC15, Dr. Y. V. Redday, chairman of FC14, and Dr. D. Subbarao, former Governor of RBI. It was sponsored by Sundaram Finance, Dr. A. C. Muthiah, Indian Bank, Trigen TEchnologies Ltd, HDFC Ltd., etc.

Speech of Dr. Manmohan Singh, Former Prime Minister at the Conference of Madras School of Economics to Honor Dr. Rangarajan on his 90th Birth Anniversary on 21–22 April, 2023

Dr. Rangarajan’s Birth centenary is an occasion for us to reflect on his immense contribution toward Indian economy and the monetary policy of our country. He is one of those rare economists who are capable of straddling the two worlds of academia and the government. I am very happy that Madras School of Economics is organizing a Conference to honor and celebrate Dr. C. Rangarajan’s 90th Birth Anniversary on 21–22 April 2023. I join numerous admirers and friends of Dr. Rangarajan in conveying my warmest greetings for having completed 90 years of a very creative life in the service of the nation. Dr. C. Rangarajan has been a very close friend and a distinguished economist. I have the privilege of knowing him for over 50 years. He is completing 90 fulfilling years that started in a small town called Ariyalur, now Perambalur. This journey has taken him around and at the peaks of financial and economic world globally. I have not known anymore more dependable, more impartial and more valuable in the advice he gave as the Chairman of the Prime Minister’s Economic Advisory Council during my tenure as the Prime Minister. I remain deeply indebted for the advice that he gave to the government at that time. As Governor of Reserve Bank of India he exerted profound influence in the formulation of monetary policy and exchange rate policies. Indeed, Dr. Rangarajan has excelled himself in many of the public positions he has held by tempering his economic convictions with a very sound discrimination between what is practically feasible and what is not. Right from the days when he taught at the Wharton School of Finance and Commerce, Graduate School of Business Administration, New York University and IIM-Ahmedabad, his love of academics has been overwhelming. He applied academic principles at practical problems and found undisputed solutions. Upon completion of his term as Chairman to the Economic Advisory Council, he returned to academics and is serving and actively contributing as Chairman, Madras School of Economics. I have never seen any controversies in what he speaks or writes and rises above all biases. As Governor of the Reserve Bank of India, he led the reforms from the front and was part of the team of those men and women who shared with him the dream of an equal, just and inclusive India. He has been a great votary of financial inclusion. It xiii

xiv

Speech of Dr. Manmohan Singh, Former Prime Minister …

was he who sowed the seeds of financial inclusion with women at the center. Having allowed SHGs to be recognized as formal entities and open bank accounts was a game changer that economically empowered millions of women in the country and spurred financial inclusion. Growth and inclusion are two sides of the same coin. Dr. Rangarajan acknowledged it well and weaved these together to produce a cohesive pattern of growth that generated additional resources to be utilized to meet the socio-economic goals and facilitate greater inclusion. Computerization of the bank branch operations was a far-reaching reform that was led by Dr. Rangarajan. He had the vision of taking banking at the doorstep of people and servicing customers at a quicker place. He has been a pillar of India’s economic community and anchored policies as they evolved. India’s management of triple crisis—fiscal, balance of payments and financial— experienced in the late-80s and the mid-90s, much of it under Dr. Rangarajan’s stewardship set the stage for New Delhi’s response to the Asian and then the transAtlantic financial crisis. In fact, after the external payments crisis in 1991, India undertook transformative changes in the country’s trade, tariffs, and foreign investment policies. These helped improve India’s external economic profile, including a reduction in external debt-to-GDP ratios and an increase in trade-GDP and foreign investment-GDP ratios. Most importantly, the changes have enabled India to accumulate foreign currency reserves. The strategy of gradual economic liberalization combined with risk-averse prudential regulation in the banking and financial sector helped limit India’s exposure to the recent financial crisis and the subsequent global slowdown. I am sure, for many more years, he will continue to give us the benefit of his wisdom and guidance through his writings. I wish my friend, a distinguished scholar and an outstanding public servant, Dr. C. Rangarajan all the very best and a long time in the service of our motherland.

Contents

1

Introducing the Volume on India’s Contemporary Macroeocnomic Themes-Looking Beyond 2020 . . . . . . . . . . . . . . . . . . D. K. Srivastava and K. R. Shanmugam

Part I 2

3

4

5

Growth and Macro Policies

The Indian Economy in the Post-pandemic World: Opportunities and Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S. Mahendra Dev and Rajeswari Sengupta India’s Economy in the Twenty-First Century: Role of State-Differentiated Demographic Dividend . . . . . . . . . . . . . . . . . . . D. K. Srivastava, Muralikrishna Bharadwaj, Tarrung Kapur, and Ragini Trehan Post-covid Fiscal Recovery in India: Uncertainty, Growth, and Fiscal Prudence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pinaki Chakraborty

Part II

1

9

53

73

State Finances and Intergovernmental Fiscal Transfers

The Relationship Between Government Revenue, Government Expenditure and Economic Growth in India: An Empirical Investigation at the Sub-national Level . . . . . . . . . . . . . . . . . . . . . . . . . . Deba Prasad Rath, Samir Ranjan Behera, Bichitrananda Seth, Anoop K. Suresh, and Rachit Solanki

89

6

Revenue Implications of GST on Indian State Finances . . . . . . . . . . . 123 Sacchidananda Mukherjee

7

Equalization Transfers Policy Based on Expenditure Needs and Own Revenue Capacity of Indian State Governments . . . . . . . . . 153 K. R. Shanmugam and K. Shanmugam

xv

xvi

Contents

Part III Fiscal Reforms 8

Goods and Services Tax in India: A Stocktaking . . . . . . . . . . . . . . . . . 183 Govinda Rao

9

Recent Reforms in India’s Corporate Income Tax Regime: Rationale, Impacts, and Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . 207 Supriyo De

10 The Anatomy of Public Debt Reductions: Case of India . . . . . . . . . . . 237 Prachi Mishra and Nikhil Patel 11 Measuring Tax Impact on Corporate Dividend Behavior in India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 J. V. M. Sarma Part IV Banking and Monetary Policy 12 Non-performing Assets of Indian Banking: An Evolutionary Journey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267 Rakesh Mohan and Partha Ray 13 Conduct of Monetary Policy in India: The Journey so Far and Contemporary Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315 Amaresh Samantaraya 14 Macroeconomics of Digitalization—Evolving Issues and Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343 Sakshi Awasthy and Sarat Dhal 15 Does Financial Frictions Matter for Monetary Policy Transmission in India? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357 Ranjan Kumar Mohanty and N. R. Bhanumurthy 16 Cash and Debt Management in India . . . . . . . . . . . . . . . . . . . . . . . . . . . 385 Charan Singh Part V

Environment and Social Sector Policies

17 Role of Fiscal Policy in Climate Change Mitigation in India . . . . . . . 401 U. Sankar 18 Ecological Fiscal Transfers and State-Level Budgetary Spending in India: Analyzing the Flypaper Effects . . . . . . . . . . . . . . . 421 Amandeep Kaur, Ranjan Kumar Mohanty, Lekha Chakraborty, and Divy Rangan 19 Measurement of Multidimensional Inequality of Opportunity in India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437 Sayli Javadekar and Jaya Krishnakumar

Contents

xvii

20 Youth Labor Market Challenges in India: Education, Employment, and Sustainable Development Goals . . . . . . . . . . . . . . . . 479 Jajati Keshari Parida and S. Madheswaran Part VI

Emerging Economic and Policy Challenges

21 Forecasting State-Level Fiscal Imbalances in India . . . . . . . . . . . . . . . 499 D. K. Srivastava, Muralikrishna Bharadwaj, Tarrung Kapur, and Ragini Trehan 22 Machine Learning in Macroeconomics: Application to DSGE Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 521 D. M. Nachane and Aditi Chaubal 23 Friends with Benefits: The Role of Internal Capital Markets During Financial Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 543 Suparna Biswas and Saumitra N. Bhaduri 24 Reforming the Indian Census . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 571 P. G. Babu and Vikas Kumar 25 The Performance of India’s Merchandise Exports: An Analytical Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 583 Pushpa Trivedi

Editors and Contributors

About the Editors D. K. Srivastava is a noted economist and currently the chief policy advisor, Ernst & Young India. He is also an honorary professor at Madras School of Economics, Chennai, India, and was a member of the Advisory Council to the 15th Finance Commission of India. He was also a professor of National Institute for Public Finance and Policy (NIPFP), Delhi, India, and has published many books including Development and Public Finance: Essays in Honour of Raja J. Chelliah (co-edited, 2005), Federalism and Fiscal Transfers in India (co-edited, 2011), and Environment and Fiscal Reforms in India (co-edited, 2014). K. R. Shanmugam is Professor at, and Director of Madras School of Economics, Chennai, India. He has six books (the latest one in 2019 with Springer, co-authored) and numerous journal articles to his credit on topics of his research interest such as public finance, corporate finance, macroeconomic models, applied economics, economics of human resources and environmental economics.

Contributors Sakshi Awasthy Department of Economic and Policy Research, Reserve Bank of India, Mumbai, India P. G. Babu Indira Gandhi Institute of Development Research, Mumbai, India Samir Ranjan Behera Department of Economic and Policy Research, Reserve Bank of India, Mumbai, India Saumitra N. Bhaduri Madras School of Economics, Chennai, Tamilnadu, India

xix

xx

Editors and Contributors

N. R. Bhanumurthy Dr. B.R. Ambedkar School of Economics (BASE) University, Bengaluru, India Muralikrishna Bharadwaj Senior Manager, Macro-Fiscal Unit, Tax and Economic Policy Group, Ernst & Young, Gurgaon, India; Senior Manager, Tax and Economic Policy Group, EY India, Gurugram, India Suparna Biswas Madras scholar of Economics, Chennai, Tamilnadu, India Lekha Chakraborty National Institute of Public Finance and Policy, New Delhi, India Pinaki Chakraborty Vice-Chairman, Institute of Development Studies, Jaipur, India Aditi Chaubal Indian Institute of Technology Bombay, Mumbai, India Supriyo De National Institute of Public Finance and Policy, New Delhi, India S. Mahendra Dev Faculty of Social Sciences, ICFAI, Hyderabad, India; Former Director, IGIDR, Mumbai, India Sarat Dhal Department of Economic and Policy Research, Reserve Bank of India, Mumbai, India Sayli Javadekar Department of Economics, University of Bath, Bath, UK Tarrung Kapur Senior Manager, Macro-Fiscal Unit, Tax and Economic Policy Group, Ernst & Young, Gurgaon, India; Senior Manager, Tax and Economic Policy Group, EY India, Gurugram, India Amandeep Kaur National Institute of Public Finance and Policy, New Delhi, India Jaya Krishnakumar Institute of Economics and Econometrics, Geneva School of Economics and Management, University of Geneva, Geneva, Switzerland Vikas Kumar Azim Premji University, Bengaluru, India S. Madheswaran Centre for Economic Studies and Policy, Institute for Social and Economic Change, Nagarbhavi, Bangalore, India Prachi Mishra International Monetary Fund, Washington, USA Rakesh Mohan Centre for Social and Economic Progress (CSEP), New Delhi, India Ranjan Kumar Mohanty Xavier Institute of Management, XIM University, Bhubaneswar, Odisha, India Sacchidananda Mukherjee National Institute of Public Finance and Policy (NIPFP), New Delhi, India D. M. Nachane Centre for Economic and Social Studies, Hyderabad, India

Editors and Contributors

xxi

Jajati Keshari Parida School of Economics, University of Hyderabad, Gachibowli, Hyderabad, India Nikhil Patel International Monetary Fund, Washington, USA Divy Rangan The Convergence Foundation, New Delhi-, India Govinda Rao Takshashila Institution, Jakkur, Bengaluru, Karnataka, India Deba Prasad Rath Department of Economic and Policy Research, Reserve Bank of India, Mumbai, India Partha Ray National Institute of Bank Management (NIBM), Pune, India Amaresh Samantaraya Department of Economics, Pondicherry University, Pondicherry, India U. Sankar Madras School of Economics, Chennai, India J. V. M. Sarma Centre for Public Finance, Madras School of Economics, Chennai, India Rajeswari Sengupta Indira Gandhi Institute of Development Research (IGIDR), Mumbai, India Bichitrananda Seth Department of Economic and Policy Research, Reserve Bank of India, Mumbai, India K. Shanmugam Government of Tamil Nadu, Chennai, India K. R. Shanmugam Madras School of Economics, Chennai, India Charan Singh CEO, Foundation for Economic Growth and Welfare, Noida, Uttar Pradesh, India Rachit Solanki Department of Economic and Policy Research, Reserve Bank of India, Mumbai, India D. K. Srivastava Chief Policy Advisor, EY India and Formerly Director, Madras School of Economics, Gurugram, India Anoop K. Suresh Department of Economic and Policy Research, Reserve Bank of India, Mumbai, India Ragini Trehan Senior Manager, Macro-Fiscal Unit, Tax and Economic Policy Group, Ernst & Young, Gurgaon, India; Senior Manager, Tax and Economic Policy Group, EY India, Gurugram, India Pushpa Trivedi Department of Economics, School of Science and Humanities, Shiv Nadar University Chennai, Chennai, India

Chapter 1

Introducing the Volume on India’s Contemporary Macroeocnomic Themes-Looking Beyond 2020 D. K. Srivastava and K. R. Shanmugam

Abstract This chapter initially highlights the current status of the Indian economy and the challenges faced by it. Then, it introduces the six broad themes of the volume: Growth and Macro Policies, State Finances and Intergovernmental Fiscal Transfers, Fiscal Reforms, Banking and Monetary Policy, Environment and Social Policies and Emerging Economic and Policy Challenges. It also introduces the contributors of this volume who joined hands to pay tribute to Dr. C. Rangarajan. Keywords Indian economy · Macroeconomic themes · Growth · Monetary · Environment and social policies · Emerging challenges

1.1 Introduction India is the largest democracy in the world. Since 1991, its economy has pursued free market liberalization, greater openness in trade and increased investment in infrastructure and has achieved considerable progress. It has emerged as the third largest economy in terms of purchasing power parity. It is among the fastest growing nations in the world. The Indian economy is projected to be the world’s second largest before 2050. It has made great advances in many fields, particularly Information Technology (IT). On December 21, 2022, India assumed the presidency of the G-20 forum; it is a member of the IMF and the World Bank. It is a destination for larger FDI. It has launched its own satellites and sent a spacecraft to the moon and Mars. However, it faces a number of challenges. One of the biggest challenges is the population. It has surpassed China and become the world’s most populous country. While it is the third largest economy and the fastest growing one, its per capita D. K. Srivastava · K. R. Shanmugam (B) Madras School of Economics, Chennai, India e-mail: [email protected] D. K. Srivastava e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 D. K. Srivastava and K. R. Shanmugam (eds.), India’s Contemporary Macroeconomic Themes, India Studies in Business and Economics, https://doi.org/10.1007/978-981-99-5728-6_1

1

2

D. K. Srivastava and K. R. Shanmugam

income was only US$2256.6 in 2021 as against the USA’s per capita income of US$70,248.6 and China’s per capita income of US$12,556.3. Another challenge is its huge income inequality. The new Oxfam report highlights that more than 40% of the wealth created in India from 2012 to 2021 had gone to just 1% of the population while only 3% had trickled down to the bottom 50%. High investment growth is critical to meet India’s rising demand and ensure noninflationary growth in the long run. Capital investment, particularly in the private sector, has not yet picked up sustainably. However, evidence indicates that the nonperforming assets have come down after 2017 and banks’ willingness to lend has gone up substantially. Additionally, the savings-investment gap has widened and the capacity utilization of manufacturing firms has increased, indicating that India may be at the cusp of a private investment revival. According to the World Bank, India has improved its ease of doing business significantly, from its 142nd place out of 190 countries in 2014 to 63rd rank in 2022. However, it suffers from high inflation, an unstable rupee, and a large current account deficit. As a net importer of crude oil, India’s economy is sensitive to increases in the price of oil and other commodities, such as gas, steel and precious metals. The high price of oil in 2021–22 has worsened India’s current account deficit and put upward pressure on consumer prices. The ongoing Russia-Ukraine war has also aggravated the problem. While the Indian agriculture sector has managed to live up to the demands of India’s large population, about 45% of the working population is engaged in agriculture. However, it contributes about 15% of the national income, indicating a low productivity per person in the sector. Further, India has a poor quality of human capital. According to the UNDP’s HDI, which is based on life expectancy, education and per capita income India ranked 132 out of 191 countries in 2021. Other challenges faced by the Indian economy are poverty and unemployment. More than 10% of the population (i.e., about 15 crore) lives below the World Bank poverty line. This means that these people are not able to participate in the economy and this leads to a vicious cycle of poverty. Between 1991 to 2019, India’s unemployment rate ranged between 2.37% and 5.76%, but it increased to 8% in 2020. Additionally, India has one of the largest budget deficits in the developing world. Excluding subsidies, it amounts to nearly 8% of GDP. The current level of its public debt is unsustainable. It still allows little scope for increasing investment in public services like health and education. India also only manages to collect tax revenues averaging about 17%1 of GDP on the combined account of central and state governments, which is relatively low in comparison to other countries at similar per capita income levels. India’s Constitution provides for the transfer of resources from the Center to the subnational governments. These provisions, however, do not specifically mention the need for making equalization transfers. There are considerable disparities in the level of publicly provided services across states.

1

Average over the period 2017–18 to 2021–22.

1 Introducing the Volume on India’s Contemporary Macroeocnomic …

3

1.2 Overviews of Themes and Chapters Against this backdrop, this book is devoted to analyzing salient ongoing economic and policy themes with a forward-looking outlook. It is divided into six broad sections covering six critical themes. These themes are not mutually exclusive and often cover common ground. This is due to the nature of interdependence of the economic issues. The first theme relates to Growth and Macro Polices. Growth is the central challenge for the Indian economy in the twenty-first century, especially in the period up to 2047–48 known as the ‘Amrit Kaal’. In this period, India is slated to become one of the largest economies in the world, leading global growth. In this section, three contributions have been included. The first article by Mahendra Dev and Sengupta looks at India’s medium-term outlook in the post-pandemic world in the context of an uncertain global environment and India’s structural problems. The article provides an assessment of the factors that need to be critically addressed in order for India to not only achieve and sustain a high rate of growth but also to make the leap to a high-income country and create adequate jobs. The second article by Srivastava, Bharadwaj, Kapur and Trehan has a longer-term focus, highlighting the role of the demographic dividend in the evolution of the Indian economy. It especially draws attention to the role of the differential pace at which this dividend would unfold for different states in India and how the opportunities that are opening up can be utilized to obtain optimal growth results. The third article in this section by Chakraborty focuses on the short to medium-term, emphasizing the need to restore fiscal balances after the COVID shock so as to put the Indian economy back on the high growth path. The second theme is on State finances and Intergovernmental fiscal transfers. Management of intergovernmental fiscal transfers through the Finance Commissions has been a subject of intense research and debate. It continues to remain critical for India’s future since much of the investment in human resources has been entrusted to the state governments as per our constitutional scheme. Robust state finances through fiscal transfers as well as through their own revenue effort are therefore critical for India’s development and welfare. In this section, the article by Rath, Behera, Seth, Suresh and Solanki explores the relationship between government revenues, government expenditures, and economic growth. They advocate that states should strive to boost their revenue growth while ensuring containment of their expenditure growth relative to their revenue counterpart. The next article in this section by Mukherjee on GST’s revenue implications on Indian state finances makes a case for undertaking a comprehensive assessment of the revenue performance of both GST and non-GST taxes for examining the impact of GST on state finances. The third paper in this section by Shanmugam and Shanmugam proposes a comprehensive methodology for determining equalization transfers in India by jointly considering expenditure needs and revenue capacities of the states. The third theme relates to fiscal reforms in India. Fiscal reforms have been an ongoing theme in India over a long period of time, encompassing taxation reforms, Fiscal Responsibility Legislations (FRLs), and public expenditure reforms. Many

4

D. K. Srivastava and K. R. Shanmugam

fiscal reform initiatives have been embarked upon in recent years, such as the GST and the reforms in the Corporate Income Tax (CIT) system. In this section, Govinda Rao provides a comprehensive stocktaking of India’s GST, arguing that these reforms are by no means complete and much more needs to be done to enhance its revenue productivity while lowering its collection, compliance, and distortion costs. Rao provides a detailed reform agenda for fully realizing the potential gains from the tax in terms of efficiency and revenues. The next paper by Supriyo De in the context of fiscal reforms relates to the recent CIT reforms. It utilizes the cost of capital approach to examine the impact of the CIT reforms across various sectors and ownership types. The results indicate that in terms of user cost, the various lower tax options are not attractive, and under certain situations may be worse for younger and smaller companies. Public debt sustainability has been subjected to a major challenge due to COVID. In the next article, Mishra and Patel outline the costs and risks associated with high public debt in India, drawing on experiences across countries as well as India’s own past. They conclude with a discussion of the benefits of reducing debt and highlight possible scenarios for achieving a sizable reduction in debt levels over the next decade. Continuing with the theme of fiscal reforms, Sarma analyzes the impact of taxation of income on corporate dividend behavior in India using a time series analysis. The fourth theme pertains to the subject of Banking and Monetary Policy. The monetary and financial dimension of India’s macroeconomy has a critical role to play both in price stabilization and facilitation of economic growth. In this section, the first paper by Mohan and Ray takes up the important issue of non-performing assets (NPAs) of the Indian banks. The authors narrate the journey of the emergence of NPAs in India and identify factors that explain the upward and downward phases of NPAs. The next paper by Samantharaya is on the conduct of monetary policy in India. He highlights the issues relating to single and multiple objective frameworks for determining monetary policy with a view to strengthening policy effectiveness. In the next paper, Awasthy and Dhal take up another forward-looking critical aspect of the Indian economy pertaining to digitalization and its macroeconomic implications. They provide a generalized perspective on the impact of ongoing digitalization on key economic indicators relating to productivity and economic growth, inflation dynamics, financial markets, and currency demand. Continuing the discussion on monetary policy, Mohanty and Bhanumurthy take up the issue of monetary policy transmission in India and analyze the role of financial frictions in the context of flexible inflation targeting. They contend that financial frictions influence the extent of policy transmission in India and argue for adding output growth as a monetary policy target in addition to inflation. Charan Singh discusses automatic monetization, transparency of budget operations and the dismantling of the administered rates regime in the country. The fifth theme pertains to the critical area of environment and social sector policies. India has a critical global responsibility in ensuring that it plays its due role in attaining global climate targets. The Government of India has also made a number of commitments that would affect future policies and the working of the economy. In this section, Sankar analyzes, in detail, the role of fiscal policy in climate change

1 Introducing the Volume on India’s Contemporary Macroeocnomic …

5

mitigation in India and recommends that India may prepare to introduce carbon tax systems or emission trading systems. Within this broad theme, Kaur, Mohanty, Chakraborty and Rangan take up the subject of ecological fiscal transfers and assess whether state-level budgetary spending shows a flypaper effect linked to these transfers. Javadekar and Krishnakumar examine the issue of measuring multidimensional inequality of opportunity in India. They consider well-being in the work domain as a multifaceted concept and model it as a latent variable measured by multiple indicators and apply this approach to Indian data using a regression-based approach. Parida and Madheswaran take up the subject of youth labor market challenges in India in the context of education, employment, and Sustainable Development Goals (SDGs). They argue that policies aimed at the development of infrastructure along with promotion of industrial growth are necessary to create adequate jobs for the growing educated labor force, to revive the structural transformation, and to achieve the SDGs. The sixth theme, namely, Emerging Economic and Policy Challenges, has a forward-looking orientation. In this section, we endeavor to cover certain emerging dimensions and challenges that are critical for India’s future. Papers in this section take up subjects that are often missed out in more conventional discussions of economic growth and policy. The first paper by Srivastava, Bharadwaj, Kapur and Trehan develops a forecasting framework for projecting state-level fiscal imbalances in the medium-term by using a panel model approach in a multi-equation framework. Nachane and Chaubal pick-up another forward-looking and under-analyzed subject of the role of machine learning in macroeconomics, especially in the context of Dynamic Stochastic General Equilibrium (DSGE) models. They critically analyze the implications of introducing machine learning techniques on DSGE models and related policy implications. Biswas and Bhaduri analyze the subject of financial stress and its impact. They show that group-affiliated firms have lesser chances of facing financial distress as opposed to stand-alone firms in challenging economic conditions. Babu and Kumar provide a well-thought-out contribution on the subject of reforming the Indian Statistical System in a rapidly changing world so that availability of adequate statistics can contribute to India’s future economic performance. The last paper in this section by Trivedi provides a critical perspective on India’s external sector which is bound to play a key role in India’s future growth performance. The study covers both trade and capital flows as aspects of the external sector. This book provides, on the whole, an assessment of critical and contemporary macroeconomic challenges of the Indian economy while highlighting the opportunities and policy choices for India’s bright economic future in the forthcoming decades. The contributions have come from established scholars in the field of macroeconomics who joined hands to pay tribute to an outstanding economist and policymaker, Dr. C. Rangarajan, who was a key contributor to India’s economic reforms. That journey of reforms and opportunities continues, making India’s position in the global economy progressively stronger.

Part I

Growth and Macro Policies

Chapter 2

The Indian Economy in the Post-pandemic World: Opportunities and Challenges S. Mahendra Dev and Rajeswari Sengupta

Abstract In this descriptive study, we first analyze what the medium term future looks like for the Indian economy in a post-pandemic world, both in the context of an uncertain global environment and also in the context of India’s own structural problems. We also highlight the structural challenges and opportunities for the Indian economy—what are the obstacles in India’s growth trajectory going forward, and what can be done to overcome them as India maneuvers through a complex global environment where geopolitics dominate trade, climate change concerns become critical and authoritarian regimes may increasingly lose favor with foreign investors and corporations. The objective is to provide a broad assessment of the factors that need to be critically addressed in order for India to not only achieve and sustain a high rate of growth but also to make the leap to a high income country and create adequate jobs. Keywords Pandemic · K-shaped recovery · Job creation · Structural challenges · Global uncertainty · Economic reforms JEL Codes E2 · E5 · E6 · G2

S. M. Dev (B) Faculty of Social Sciences, ICFAI, Hyderabad, India e-mail: [email protected] Former Director, IGIDR, Mumbai, India R. Sengupta Indira Gandhi Institute of Development Research (IGIDR), Mumbai, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 D. K. Srivastava and K. R. Shanmugam (eds.), India’s Contemporary Macroeconomic Themes, India Studies in Business and Economics, https://doi.org/10.1007/978-981-99-5728-6_2

9

10

S. M. Dev and R. Sengupta

2.1 Introduction It has been three years since the Covid-19 pandemic began spreading in India. By now the pandemic seems behind us. Even though the health shock seems over, the economic recovery from its impact has progressed at an uneven pace and in a haphazard manner. While the formal economy may have recovered from the shock the same perhaps cannot be said about the informal sector, which is also not adequately captured in the official data. Over and above the pandemic there have been other shocks to the economic recovery in the form of severe supply chain bottlenecks triggered by the year-long land war in Europe between Russia and Ukraine, relapse of the pandemic in China and associated draconian restrictions imposed by their government on the movements of people as well as goods, and persistently high inflation in the Western economies of the US, UK and EU and consequent aggressive monetary tightening by the respective central banks. The most recent shock has manifested itself in the form of a banking sector turmoil in the US which started with the collapse of the Silicon Valley Bank in California. These global shocks have further aggravated the growth challenges facing the Indian economy even as it has been struggling to recover from the pandemic. By some measures, large parts of the economy are back to the pre-pandemic levels of activity. However, it is worthwhile to remember in this context that even before the pandemic the Indian economy was in a precarious state. There may have been structural challenges, cyclical factors as well as policy issues that may have jointly contributed to the slowing down of the economy in the run-up to the pandemic. This implies that even if the economy returns to the pre-pandemic levels, the longstanding challenges may make it difficult to achieve a high and sustainable growth rate which is essential if India wants to lift millions out of poverty, achieve a high income status and create jobs for the millions entering the workforce every year. To add to the challenges, India is now facing a slowing global economy, a fragmented geopolitical environment, growing threat of climate change and the related uncertainties, increasing automation and its impact on the labor market, and potentially a protectionist world that is withdrawing from the uninterrupted globalization that had created vast economic gains in the post-world war period. In this study, we first analyze what the medium term future looks like for the Indian economy in a post-pandemic world, both in the context of an uncertain global environment and also in the context of India’s own structural problems. We then outline some of the structural challenges and opportunities for the Indian economy as it struggles to restore high growth in the post-pandemic period and highlight some of the factors that need urgent attention if India has to make a leap from middle to a high income country in the next few decades. The high growth will also help in increasing employment and reducing poverty. We start by providing a brief description of how the economy was faring in the pre-pandemic period (Sect. 2.2) followed by a comprehensive analysis of how the pandemic affected various sectors of the economy and delineating the patterns of

2 The Indian Economy in the Post-pandemic World: Opportunities …

11

economy recovery from the shock (Sect. 2.3). We then move on to a broader discussion of the challenges and problems that might pose as obstacles in India’s future growth trajectory and also throw some light on some of the important opportunities that policymakers need to be cognisant of (Sect. 2.4).

2.2 Pre-pandemic Economic Conditions In case of India, the pandemic was even more of an acute and longer lasting shock compared to other affected countries, owing to the state the economy was in, in the pre-Covid-19 period. By the time the first Covid-19 case was reported in India, the economy had deteriorated significantly after years of feeble performance. The pandemic compounded the existing problems of unemployment, low incomes, rural distress, malnutrition and widespread inequality.

2.2.1 Aggregate Macro Conditions As discussed in detail in Dev and Sengupta (2022a), real GDP (gross domestic product) growth rate had been on a downward trajectory since 2015–16 and by 2019–20, it had slowed down to 3.9%, the lowest level since 2002–03 (see Fig. 2.1 for example). Unemployment reached a 45-year high increasing from 2.2% in 2011–12 to 6.1% in 2017–18. A major driver of growth in any economy is investment by the private corporate sector. In the pre-Covid-19 period, private sector investment had been declining. According to the official data, gross fixed capital formation (GFCF) as percent of GDP declined from 34.3% in 2011–12 to 28.6% in 2019–20. In 2019–20, it grew (in real terms) only by 1.1% down from 11.2% in the previous year. Capex data from CMIE shows a similar picture of declining investment (see Fig. 2.13a, b). Aggregate consumption expenditure (private and government) had also been falling, for the first time in several decades. Exports of goods and services contracted by 3.4% in 2019– 20 after growing by 11.9% in the previous year; exports of goods alone contracted sharply by 6.1%.1 In general, non-oil, non-gold exports were pretty much stagnant in the run-up to the pandemic (see Fig. 2.4).

1

Data obtained from the Second Revised Estimates for 2020–21 released on February 28, 2023 by the MOSPI (Ministry of Statistics and Program Implementation). It may be noted that national accounts statistics in India especially GDP data are potentially fraught with measurement errors as detailed in multiple studies (see for example, Nagaraj et al. (2020), Nagaraj and Srinivasan (2017), among others) and hence should be interpreted with caution. Wherever possible we have therefore supplemented this information with data from the private sector databases such as those published by the Centre for Monitoring Indian Economy (CMIE). All growth numbers are in constant prices (2011–12).

12

S. M. Dev and R. Sengupta

Fig. 2.1 GVA growth at constant prices. Source CMIE Economic Outlook. This graph shows non-agricultural, non-government GVA (gross value added) growth at constant prices

Some of the key macroeconomic parameters on the other hand were relatively stable with high reserves of foreign exchange and low inflation.

2.2.2 Agriculture, Informal Sector and MSMEs 2.2.2.1

Agricultural Sector

This sector is critical as a large number of workers and the entire country’s population are dependent on it. The performance of agriculture is also key to the state of rural demand. In the pre-Covid-19 period, agricultural GDP experienced an average growth rate of 3.5% per year in the eight-year period from 2012–13 to 2019–20 with intermittent fluctuations (Table 2.1). However, the terms of trade moved against agriculture from 2016–17 to 2018– 19 due to bumper crop and horticultural production which caused a decline in food prices. Terms of trade for agriculture improved in 2019–20 as the nominal agricultural GDP growth was 12.5% as compared to real growth of 4.3%. Growth in rural wages was subdued in the pre-Covid-19 period, particularly for agricultural labor in both nominal and real terms, partly due to the slowdown in the construction sector (see Fig. 2.2).

2 The Indian Economy in the Post-pandemic World: Opportunities …

13

Fig. 2.2 Growth in rural wages. Source RBI (2020)

Table 2.1 Growth rates of GVA in agriculture and allied activities Growth rates (%)

Year 2012–13

1.5

2013–14

5.6

2014–15

−0.2

2015–16

0.6

2016–17

6.8

2017–18

6.6

2018–19

2.6

2019–20

4.3

Average growth rate during 2012–13 and 2019–20

3.48

Average growth rate during 2014–15 and 2019–20

3.45

Source National Accounts Statistics

2.2.2.2

Informal Sector

India has a vast informal sector, arguably the largest in the world, employing close to 90% of its working population and contributing more than 45% to its overall GDP. The share of informal employment in total employment is very high. The share, which includes agricultural workers, has declined marginally from 94% in 2004–05 to 91% in 2017–18 (see Table 2.2). Out of a total of 465 million workers, 422 million were informal workers in 2017–18. Even in non-farm sector (manufacturing and services), the share of informal workers was around 84% in the same year. There are significant inequalities between informal and formal sector workers. The informal/unorganized workers do not have access to any social security benefits

14

S. M. Dev and R. Sengupta

Table 2.2 Informal Employment: Number and Shares Total employment (in Informal employment % Share of informal workers in total millions) (in millions) employment 2004–05 459.4

430.9

93.8

2011–12 474.2

436.6

92.5

2017–18 465.1

421.9

90.7

Source Mehrotra and Parida (2019)

and also face the uncertainty of work. Out of the total workers, the shares of selfemployed, casual and regular workers respectively were 51.3%, 23.3% and 23.4%. Most of the self-employed and casual employees are informal workers. In the pre-Covid period, the informal sector was hit by two consecutive shocks in a short span of time, from 2016 to 2019. The first shock was Demonetisation in November 2016 when 86% of the cash in circulation in the economy became unusable overnight owing to a government decree, followed by the haphazard introduction of the Goods and Services tax in 2017 (Sengupta, 2016; Shah, 2016).With the Covid19 outbreak, the already struggling informal sector was disproportionately affected (Bhattacharya et al., 2017).

2.2.2.3

MSMEs

The micro, small and medium enterprises as a whole form a major chunk of manufacturing in India and play an important role in providing large-scale employment. Recent annual reports on MSMEs indicate that the sector contributes around 30% of India’s GDP, and based on conservative estimates, employs around 50% of industrial workers and contributes half of the overall exports. Over 98% of MSMEs can be classified as micro firms, and 94% remain unregistered with the government. Many of the micro-enterprises are small, household-run businesses. However, many aspects of government policy are at best scale neutral and do not explicitly take into consideration these enterprises. This sector also does not have access to adequate, timely and affordable institutional credit. More than 81% MSMEs are self-financed with only around 7% borrowing from formal institutions and government sources (Economic Census, 2013). The MSMEs are present in manufacturing, trade and service sectors. Table 2.3 provides growth rates of industry-wise deployment of bank credit. It shows that growth of credit was either low or negative for the MSMEs. Demonetisation and GST also contributed to the poor performance of MSMEs. The problems with the NBFC sector from 2018 onwards have further hampered credit allocation to this sector.

2 The Indian Economy in the Post-pandemic World: Opportunities …

15

Table 2.3 Growth in industry-wise deployment of bank credit by major sectors (YoY, %) Item Non-food credit

March-15 8.6

March-16 9.1

March-17 8.4

March-18 8.4

March-19 12.3

Nov-19# 7.2

Industry

5.6

2.7

−1.9

0.7

6.9

2.4

Micro and small

9.1

−2.3

−0.5

0.9

0.7

−0.1

Medium

0.4

−7.8

−8.7

−1.1

2.6

−2.4

5.3

4.2

−1.7

0.8

8.2

3.0

Textiles

−0.1

1.9

−4.6

6.9

−3.0

−6.1

Infrastructure

10.5

4.4

−6.1

−1.7

18.5

7.0

Large

Source Economic Survey 2019–20; # as on November 22, 2019

2.2.3 Formal Sector In the pre-pandemic period, one of the major problems that the Indian economy was grappling with was the Twin Balance Sheet (TBS) crisis. This manifested in the form of high levels of non-performing assets (NPAs) in the banking sector, especially in the inadequately capitalized public sector banks, and high levels of debt in large, financially stressed companies, particularly in the infrastructure sector (Sengupta and Vardhan, 2020a). The problems started around the time of the Global Financial Crisis in 2008–09 when the world economy began slowing down and the spate of infrastructure projects that had been supported by a massive credit boom in the mid-2000s began to fail (Subramanian & Felman, 2019). The balance sheet stress in both the banking sector and the private corporate sector peaked in 2018 when gross NPAs reached a level of almost 14% of total loans. On the private corporate side, Credit Suisse reported that by early 2017, around 40% of the corporate debt monitored by it was owed by companies that had an interest coverage ratio of less than 1, meaning they did not earn enough to pay the interest obligations on their loans. This TBS crisis triggered the introduction of the asset quality review (AQR) by RBI in 2016, which forced the banks to recognize stressed assets on their books. The banking sector’s response to this growing bad-loans crisis and to the series of steps taken by the government and the RBI to address the crisis was to avoid risks (Sengupta and Vardhan, 2020a) and curtail lending. And the private corporate sector’s response to the TBS crisis was to cut back investment spending and start deleveraging (Vardhan, 2021). The combination of an impaired banking sector and a cautious private corporate sector led to a drastic decline in the share of industrial credit (large firms and MSMEs) in total bank credit which fell from 44% in 2011 to 31% in 2020. In particular, credit off-take during 2019–20 was muted with non-food credit growth at 6.1% being less than half the growth of 14.4% of the previous year. This was the lowest growth rate of bank credit in nearly six decades (see Fig. 2.3).

16

S. M. Dev and R. Sengupta

Fig. 2.3 Credit and deposit growth of the banking sector (YoY). Source ICRA report

As the banking sector withdrew from lending, this gap in the commercial credit landscape to a large extent was filled up by non-banking finance companies (NBFCs). Between 2015–16 and 2019–20, the share of NBFCs and HFCs (housing finance companies) in institutional credit (i.e., credit from banks and non-bank financial institutions) increased from 20 to 27%, net of bank credit. This implies that some part of the shortfall in credit from the banking sector was compensated by flows of credit from NBFCs (Sengupta and Vardhan, 2022). Almost at the peak of the NBFC credit boom, the Indian financial system received another major blow when a big NBFC named IL&FS (Infrastructure Leasing & Financial Services) defaulted on its debts in September 2018. This sent shockwaves through the banking system as well as the debt markets—the two biggest funding sources for the NBFC sector. This was followed by several other low-impact shocks in the financial sector. Bottom line is that by 2019–20, the credit landscape was in serious turmoil and credit being the engine of growth it is not surprising that all engines of growth were sputtering, particularly private investment. Gross fixed capital formation (GFCF) growth, a measure of investment, turned negative in Q2 and Q3, 2019–20. Two key indicators of investment demand, production and imports of capital goods remained in contraction mode in January and February 2020 (RBI, 2020) before the pandemic hit India. Capacity utilization in the manufacturing sector also declined below the long-term average in 2019–20. As shown in Fig. 2.4, the profit margins of the firms in the private corporate sector were also stagnant in the pre-pandemic period.

2 The Indian Economy in the Post-pandemic World: Opportunities …

17

Fig. 2.4 Non-oil, non-gold exports. Source CMIE database

2.3 Post-pandemic Economic Conditions 2.3.1 Impact of Covid-19 Pandemic The Covid-19 pandemic was a huge shock that hit an arguably fragile Indian economy that was struggling in multiple aspects, as explained in the previous section, and impacted various segments of the economy. We discuss the impact and recovery from the pandemic in great detail in Dev and Sengupta (2022b) from March 2020 to February 2022. When the pandemic began spreading in India in March 2020, the central government announced one of the largest and most stringent nationwide lockdowns in the world at the time (based on data from the Oxford COVID-19 Government Response Tracker) in order to contain the rapid spread of the contagious disease. The first wave peaked in mid-September 2020, and cases declined thereafter till the end of the year. As the severity of the first wave of the pandemic began subsiding, many of the nationwide mobility restrictions were gradually relaxed starting in June 2020. In the summer of 2021 India was hit by a second wave of the pandemic, which peaked in May 2021, was more widespread, more severe, its geographic coverage was much greater and larger percentage of population was affected. This time around there was no nationwide lockdown; instead the lockdowns were regional and hence more scattered. Throughout the pandemic period, economic activity continued in stops and starts, depending on the heterogeneity in the geographical distribution of the Covid-19 cases as well as the extent and intensity of mobility restrictions imposed by the governments—both central and states.

18

S. M. Dev and R. Sengupta

Eventually the population gained herd immunity through the disease and also through mass vaccinations, but in the initial phases, at least till the second wave, uncertainty loomed large because of substantial delays in universal vaccination and threats from potential mutations of the virus. The severity of the pandemic began subsiding in September 2021 and despite a brief third wave in the Dec 2021-Jan 2022 period, by early 2022, the pandemic had mostly been brought under control in India, it had become endemic in nature and there were no further mobility restrictions. By April 2022, more than 60% of the population was fully vaccinated. The pandemic arguably further compounded the pre-existing problems of the Indian economy especially, unemployment, depressed consumption demand, stagnating incomes, rural distress, stress in the informal sector and widespread inequality. According to official data, real GDP shrank by 5.8% in 2020–21, with the biggest contraction reported by contact-intensive sectors such as trade, transportation, hotels and restaurants, i.e., primarily in the services sector. Among the components of aggregate demand, private sector consumption (in real terms) shrank on a year-onyear basis by 5.2%, gross fixed capital formation contracted similarly by 7.3% while exports of goods and services contracted by 9.1%. Clearly the pandemic was a severe, unprecedented shock for a slowing economy.2 From the middle of 2021, the economy, especially the formal sector, began recovering from this shock, primarily because of an unexpected boom in exports and also aided by the herd immunity of the population which allowed the services sector to resume activities. As the developed world came out of the pandemic and began opening up their respective economies, there was a tremendous surge in exports, especially merchandise exports, from India (see Fig. 2.4). This contributed majorly to the recovery of the economy from the pandemic. The exports boom continued almost till the middle of 2022. In general the formal sector seems to have fared much better than the vast informal sector, during the pandemic. While the informal sector and the MSMEs were already reeling under the previous shocks of demonetization in 2016, the GST implementation in 2017 and the NBFC crisis in 2018, the formal sector firms who had experienced balance sheet stress during the TBS crisis, had begun cleaning up their balance sheets by the time the pandemic hit India. They had been deleveraging, new investment projects had not been taken up in a big way and also the Insolvency and Bankruptcy Code (IBC, 2016) aided the process of resolution of bad debts. As a result, they were in a relatively better shape when the pandemic began spreading. So while the informal sector was disproportionately impacted by the pandemic not only because they were already struggling but also because they did not have the wherewithal to deal with such a big shock, it appears that the formal sector where the firms could also cut costs, gained market share at the expense of the vast majority of MSMEs when Covid-19 struck. This is demonstrated for example by the net profit margin of the listed companies which increased during the pandemic (see Fig. 2.5). 2

Data obtained from the Second Revised Estimates for 2020–21 released on February 28, 2023 by the MOSPI.

2 The Indian Economy in the Post-pandemic World: Opportunities …

19

Fig. 2.5 Net profit margin of non-finance, non-oil listed companies. Source CMIE database

Consequently, the manner in which economic recovery happened was heterogeneous and uneven led mostly by these large companies. This creates the impression that the recovery from the pandemic may have been K-shaped and the informal sector is yet to recover fully. Real GDP grew by 9.1% in 2021–22 on a year-on-year basis fuelled by resumption of the services sector activities. Compared to 2020–21, construction grew by close to 15%, trade, hotels and transportation grew by nearly 14%, while manufacturing expanded by 11% (all in real terms).3 Since March 2022 the economy began facing serious headwinds from multiple quarters – an extremely volatile global environment causing unprecedented uncertainty. The land war in Europe and associated sanctions on Russia imposed by the western countries, China’s zero-Covid restrictions amidst a resurgence of cases, developed countries such as the US, UK and EU nations experiencing the worst streak of inflation in four decades, forcing their respective central banks to aggressively raise interest rates—these simultaneous global shocks created widespread risk aversion, and imposed a drag on global economic growth. The IMF projected that global growth will fall from 3.4% in 2022 to 2.9% in 2023, much lower than the historical average of 3.8%. A fragmented geopolitical landscape and, persistent global macroeconomic uncertainty act as serious headwinds for the Indian economy’s future growth trajectory. By early 2023, it became clear that while large parts of the economy had gone back to pre-Covid levels of activity (interpreted as a sign of “recovery”) in effect the 3

Data obtained from the First Revised Estimates for 2021–22 released on February 28, 2023 by the MOSPI.

20

S. M. Dev and R. Sengupta

economy had grown at barely 3–3.5% during the three years of the pandemic and recovery period, i.e., from 2020–21 to 2022–23. The biggest challenge facing the Indian economy at the current juncture, therefore, is achieving a high, sustainable growth rate, and creating a sufficient number of jobs to absorb the millions of unemployed, amidst a highly volatile and uncertain global economic environment that is dealing with the repercussions of multiple adverse shocks. And even as the world was coming to terms with certain levels of uncertainty triggered by the shocks of 2022, financial market turmoil unfolding in the US with the collapse of two mid-sized banks (Silicon Valley Bank and Signature Bank) and in the EU with the forced take over by UBS of Credit Suisse-one of the most systemically important banks at a global level, has sent renewed shockwaves across countries and, deepened fears of a financial contagion. India is unlikely to remain decoupled from the impact of this evolving crisis. In the event of more such bank failures, the crisis is likely to further aggravate the headwinds for India’s already weak medium term growth prospects.

2.3.2 Aggregate Macroeconomic Conditions 2.3.2.1

Growth and Unemployment

Three years since the pandemic, the growth momentum of the Indian economy seems to be slowing down. The reopening of the economy in 2021, the revival of service sector activity as well as resurgence of merchandise exports helped fuel the recovery in 2021–22. However in 2022–23, additional global headwinds and, pre-existing weaknesses of the economy along with some potential scarring created by the pandemic have begun to act as a drag on growth. Table 2.4 provides sector-wise recovery of gross value added over the three years from FY20 to FY23. It shows that GVA increased from Rs.132 lakh crore in FY20 to Rs.147 lakh crore in FY23—a growth of roughly 3.7% per year. Agriculture has shown a growth rate of nearly 4% per year whereas mining recorded very low growth of less than 0.5%. Similarly, trade, hotels, transport and communications also registered a low growth rate of 1.4% per year. On the demand side, GDP increased from Rs.145.4 lakh crore in FY 20 to Rs.159.7 lakh crore—increase of 9.9% (see Table 2.5). In other words, growth rate in GDP was 3.3% per year during the 3-year period 2019–20 to 2022–23. Gross fixed capital formation grew at 6% while private final consumption expenditure recorded a growth rate of 4.4% during the same period. Exports and imports showed growth rates of 10% and 8% respectively. Government final consumption expenditure grew at the rate of 2.3% only. According to official statistics, in 2022–23 the Indian economy is scheduled to grow at 7% in real terms relative to 2021–22. While service sector and agriculture recovery/growth are expected to continue, manufacturing, however, is scheduled

2 The Indian Economy in the Post-pandemic World: Opportunities …

21

Table 2.4 Sector-wise recovery in Gross Value Added: 2022–23 over 2019–20 2019–20 (Rs. Lakh crore)*

2022–23 (Rs. In Lakh crore)**

2022–23 over 2019–20 (Change in percent)

Average annual growth rate over three years 2019–20 to 2022–23

Agriculture, forestry and fishing

19.94

22.21

11.38

3.8

Mining and quarrying

3.17

3.21

1.26

0.4

Manufacturing

22.60

25.97

14.91

5.0

Electricity, gas, water supply and other utility services

3.01

3.45

14.62

4.9

Construction

10.44

12.32

18.01

6.0

Trade, hotels, transport, communication and services relating to broadcasting

26.9

28.05

4.28

1.4

Financial, real estate and professional services

28.99

33.11

14.21

4.7

Public administration, defense and other services

17.33

18.80

8.48

2.8

Aggregate gross 132.36 value Added

147.13

11.16

3.7

*Third revised estimates **Second advance estimates Source National Accounts Statistics, MOSPI

to grow by 0.6% as opposed to the 11.1% growth of last year.4 In fact lackluster performance of the manufacturing sector continues to be a serious concern for India. For the April-December period of 2022–23, the economy grew by 7.7%. Among various sectors, manufacturing grew only by 0.4% (on year-on-year basis) compared to 15.6% over the same period in 2021–22. Growth in the construction sector and mining also slowed down, while service sector, especially financial services exhibited relatively better performance. 4

Data obtained from the Second Advance Estimates for 2022–23 released on February 28, 2023 by the MOSPI.

22

S. M. Dev and R. Sengupta

Table 2.5 Components of Aggregate Demand: 2022–23 over 2019–20 2019–20 (Rs. Lakh crore)*

2022–23 (Rs. Lakh crore)**

2022–23 over 2019–20 (change in percent)

Average annual growth rate over three years 2019–20 to 2022–23

Private final consumer expenditure

82.56

93.42

13.15

4.4

Government final consumption Expenditure

14.92

15.95

6.9

2.3

Gross fixed capital formation

45.93

54.26

18.1

6.0 10.3

Exports

28.14

36.86

30.99

Imports

33.22

41.51

25.0

8.3

145.35

159.71

9.9

3.3

Aggregate GDP

*Third revised estimates **Second advance estimates Source National Accounts Statistics, MOSPI

Another crucial challenge facing the Indian economy now is related to job creation. Job creation has always been a major problem in India, and the pandemic has in all probability left permanent scars on the Indian labor market. A deeply worrisome piece of data in this context is the total number of persons working in the economy

Fig. 2.6 Number of persons working. Source CMIE CPHS database

2 The Indian Economy in the Post-pandemic World: Opportunities …

23

(across both formal and informal sectors). As estimated by the CMIE (using the Consumer Pyramids Households Survey or CPHS) this number has been roughly stagnant at 400 million since 2016. There has been no secular growth in the total employment. Once the shock of the pandemic subsided and the economy reopened, this number came back to around 400 million by early 2023 (see Fig. 2.6). The labor force participation rate (LFPR) was nearly 43% in the pre-pandemic quarter of October-December, 2019 and averaged 44% between March 2016 and March 2020. It recorded a sharp decline to 38.3% in the April–June quarter of 2020 during the first wave of the pandemic, and by January 2023 it had risen to merely 39.8% (see Fig. 2.7), much below the pre-pandemic levels. In other words, the CMIE data, in general, shows that employment is yet to recover to the pre-pandemic level. The unemployment rate (see Table 2.6) looks almost similar across the two periods but given that many people left the labor force in the pandemic and post-pandemic periods, the true unemployment post-pandemic has been higher. This is clear evidence of deeper problems persisting in the economy.

Fig. 2.7 Labor force participation rate. Source CMIE CPHS data

Table 2.6 Labor force participation rate and unemployment rate Indicator

May–August 2019

September-December 2019

May–August 2022

September-December 2022

Labor force participate rate(%)

42.85

42.71

39.17

39.51

7.46

7.52

7.43

7.47

Unemployment Rate (%) Source CMIE

24

S. M. Dev and R. Sengupta

Table 2.7 No. of households employed and person days generated: FY20 to FY23 Months

No. of households employed (in millions)

No. of person days generated (in millions)

2019–20 2020–21 2021–22 2022–23 2019–20 2020–21 2021–22 2022–23 April

16.9

11.1

21.3

18.7

271.0

141.7

340.7

285.9

May

21.0

33.1

22.3

26.2

365.4

569.5

371.6

435.3

June

21.5

38.9

29.4

27.6

319.0

640.6

451.8

421.9

July

15.0

27.6

26.8

17.6

193.3

391.1

379.5

235.5

August

12.3

20.1

21.1

13.8

152.6

260.3

278.2

167.1

September 12.0

20.0

20.8

14.4

146.8

263.6

278.4

179.3

10.9

19.9

17.4

13.4

137.9

262.4

221.7

162.3

November 12.5

18.4

17.5

15.9

169.2

235.8

228.6

203.6

December 14.1

20.8

21.4

18.5

204.0

284.4

297.9

244.8

January

15.7

20.9

20.0

16.7

230.8

278.2

269.8

206.9

February

18.7

22.8

20.2

16.7

267.6

308.0

270.0

202.8

March

16.0

20.1

20.0



182.9

255.6

245.2



October

Source MGNREGA, Ministry of Rural Development

This shows that there may have been some permanent damage from the pandemic. Moreover, given that the majority (85%) of the workers are in the informal sector and assuming the formal sector has been less affected by the overall numbers we can conclude that informal sector employment in particular is yet to recover from the pre-pandemic period. The demand for MNREGA also provides some idea about the slow progress in employment generation in the overall economy. Table 2.7 shows that the average number of households employed under the MNREGA scheme in FY20 was 15.6 million; it has gone up to 18.1 million in FY23 so far, down from an average of 21.5 million in FY22. This implies that even after the pandemic has receded, enrollment under this scheme continues to be higher than the pre-pandemic level. Likewise the average number of person days generated in FY23 is higher than the pre-pandemic period. While Table 2.7 shows the actual uptake of the employment under this scheme, Table 2.8 shows the demand for work. We find that the CAGR of average number of households as well as of persons demanding work under MNREGA is higher in FY23 compared to FY20. All these point toward higher uptake of this scheme in the post-pandemic period, implying that there still aren’t enough jobs in the informal sector to absorb these workers who are instead applying for the employment guarantee scheme. This in turn highlights the weak recovery of overall employment in the post-pandemic period. Furthermore, trends in real wages reveal that rural areas are yet to recover from the stagnant/decline in real wages during the pandemic and pre-pandemic periods. Real wage growth declined sharply across all agricultural and non-agricultural operations in FY16-FY20 as compared to those of FY12-FY15 (see Table 2.9). Covid-19

2 The Indian Economy in the Post-pandemic World: Opportunities …

25

Table 2.8 Demand for work by households persons under MGNREGA: FY20 to FY23 Months

Demand for work, No. of Households (in millions)

Demand for work, No. of Persons (in millions)

2019–20 2020–21 2021–22 2022–23 2019–20 2020–21 2021–22 2022–23 April

21.05

13.41

26.19

23.27

30.38

20.01

37.85

31.29

May

24.76

37.35

26.58

30.75

35.69

54.26

39.12

41.34

June

25.43

44.79

33.97

31.77

35.38

63.50

48.15

40.91

July

18.35

31.99

31.35

20.41

24.08

42.90

41.62

23.80

August

14.60

24.32

24.66

15.98

18.28

31.59

31.74

18.15

September 14.26

24.39

24.02

16.76

17.73

31.29

30.26

18.96

12.92

24.37

20.46

15.54

16.03

30.95

25.60

17.36

November 15.21

22.76

20.63

18.55

19.27

28.92

25.50

20.96

December 17.04

26.54

24.04

21.18

22.26

34.87

30.09

24.24

January

18.88

26.35

23.37

20.69

24.95

34.37

29.82

24.50

February

22.25

28.68

23.78

21.15

29.47

38.39

30.72

25.10

March

20.74

26.24

24.06



27.64

35.91

31.58



October

Source MGNREGA, Ministry of Rural Development

worsened the trend of poor growth in rural farm and non-farm wages seen during FY16-20. Real wage growth was slightly above 1% in FY21 but it turned negative in FY22. In agricultural operations, real wages exhibited negative growth of 1% to 4% for different operations. Table 2.9 Growth in real wages for farm and non-farm activities in rural areas Activities

FY12-FY15

FY16-FY20

FY2020-21

FY2021-22

Real on-farm wage growth Sowing

10.6

1.3

0.1

−1.0

Plouwing/tilling

9.0

0.6

−0.7

−4.1

Harvesting/winnowing/threshing

11.2

0.4

1.0

−1.3

Picking worker

6.0

1.7

1.3

−1.3

Animal husbandry worker

16.8

1.8

1.6

−2.0

Real non-farm wage growth Carpenter

7.4

0.9

1.0

−3.1

Blacksmith

9.8

0.9

−0.3

−2.2

Mason

7.6

0.9

1.0

−3.2

LMV& tractor driver

10.1

0.6

0.2

−0.8

Non-farm labor

10.5

0.1

1.3

−2.6

Sweeper

14.8

0.8

1.8

−0.3

Source Sinha and Pant (2022), India ratings and Research

26

S. M. Dev and R. Sengupta

Fig. 2.8 Index of consumer sentiment. Source CMIE database

In the case of non-agricultural operations, it ranged from −0.3 to −3.2%. The Economic Survey (2023) also says nominal rural wages have increased at a steady positive rate during FY23 (till November 2022). However, the report also indicates that growth in real rural wages has been negative due to elevated inflation. It remains a matter of concern that real wages in rural areas have not recovered from low/negative growth in the pandemic and pre-pandemic period. Data on consumer sentiment index compiled by the CMIE further underscore the possibility that recovery from the pandemic has been feeble at best (see Fig. 2.8). Prior to the pandemic, the all-India index was roughly at 100; it dropped drastically during the lockdown of 2020 and since then, while it has been steadily going up, it still remains much below the pre-pandemic level.

2.3.2.2

Poverty Trends

Poverty in India has been declining in the pre-pandemic period. The estimates of poverty based on the Tendulkar committee methodology show that the number of poor came down by 137 million between 2004–05 and 2011–12 (GOI, 2009). The Rangarajan Committee also showed poverty declined between 2009–10 and 2011–12 (GOI, 2014). UNDP and Oxford multi-dimensional poverty (OPHI & University of Oxford, 2018) index show poverty was almost halved between 2005–06 and 2015– 16, climbing down to 27.5% (decline of poor by 271 million). It declined from 27.5% in 2015–16 to 16.4% in 2019–20. Is there evidence of pro-poor growth in consumption? As Table 2.10 shows, the growth of per capita consumption has been higher for top deciles compared to the

2 The Indian Economy in the Post-pandemic World: Opportunities …

27

Table 2.10 Decile-wise Growth in Per Capita Consumption (% per year, compound) 1993–94 to 2004–05

2004–05 to 2011–12

Rural

Urban

Rural

Urban

First Decile

0.70

0.66

2.91

2.96

Second Decile

0.49

0.54

3.00

3.28

Third Decile

0.56

0.66

3.15

3.39

Fourth Decile

0.55

0.91

3.17

3.42

Fifth Decile

0.54

1.00

3.17

3.41

Sixth Decile

0.55

1.24

3.30

3.35

Seventh Decile

0.52

1.36

3.40

3.30

Eighth Decile

0.61

1.35

3.45

3.40

Ninth Decile

0.71

1.47

3.48

3.45

Tenth Decile

1.61

2.30

3.71

4.52

Bottom Five Deciles

0.57

0.75

3.08

3.29

Top Five Deciles

0.80

1.54

3.47

3.60

Decile

Note The growth rates are in real terms and derived from URP (universal reference period) consumption data. Source Twelfth Five Year Plan, Planning Commission, Government of India

bottom deciles. But, the ratio between the two periods’ growth is higher for bottom deciles. We do not have official data on consumer expenditure after 2011–12. In the last one year there have been several studies on poverty using indirect methods and using CMIE, NSS and PLFS data sources. There have been extreme views on poverty trends in the post-2011–12 period. Using the “leaked” data for 2017–18 Subramaniam (2019) shows poverty increased during 2011–12 and 2017–18. Bhalla et al. (2022) present estimates of poverty and inequality for the period 2004–05 to 2020–21. According to them, the extreme poverty (pp $ 1.9) was as low as 0.8% in the prepandemic year 2019.5 Panagariya and More (2023) estimated poverty and inequality before and after COVID-19. They show that rural poverty continued to fall except in the strict lockdown quarter of April-June quarter of 2020. The rural poverty increased to 36.4% in April-June quarter 2020 from 33.50% in January-March quarter 2020. However, rural poverty continuously declined and it was 26.10% in April-June 2020– 21. Urban poverty rose from 16.3% in January-March 2020 quarter to 20.20% in April-June 2020. It was around 21.50% in January-March quarter of 2020–21 but declined to 19.70% in April-June quarter of 2020–21. Another indicator for levels of living is real wages which we have discussed earlier. The growth rate of real wages between 2014–15 and 2021–22 was below 1 percent per year across the board. The rate of growth was 0.9 percent, 0.2 percent and 0.3 percent for agricultural labor, construction workers and non-agricultural labor respectively (Dreze, 2023). Growth

5

Also see Roy et al. (2022), Dreze and Somanchi (2023) on poverty estimates.

28

S. M. Dev and R. Sengupta

of real rural wages was 4% to 5% per annum when poverty declined faster during 2004–05 to 2011–12. Our view is that in the post-2011–12 period, as the growth rate of GDP declined, the rate of decline in poverty must have slowed down. Policy-makers must continue to follow the two-fold strategy of letting the economy grow rapidly and attacking poverty directly through poverty alleviation programs.

2.3.2.3

Inflation

One problem that the Indian economy has been struggling to deal with throughout the pandemic and especially during the recovery period is inflation. The RBI has an inflation targeting mandate wherein it is supposed to maintain CPI inflation at 4% with a flexible band of ±2% around it. During the pandemic period from March 2020 to December 2021, headline CPI inflation averaged at roughly 5.8%. However, between January 2022 and February 2023, the average inflation went up to 6.7%, well above the upper threshold of the RBI’s tolerance band (see Fig. 2.9). This was primarily due to supply chain bottlenecks triggered by the war in Europe and China’s pandemic related lockdowns of 2022. WPI inflation, which is a proxy for inflation in the pipeline, also reached as high as 13%. Rising inflation in India also coincided with high and stubborn inflation in the developed economies of the US, UK and the EU. Global food inflation also went up dramatically for many commodities due to supply chain constraints on one hand and a significant increase in fiscal stimulus and injection of abundant global liquidity especially by the developed countries on the

Fig. 2.9 Headline CPI inflation. Source CMIE database

2 The Indian Economy in the Post-pandemic World: Opportunities …

29

other hand which boosted aggregate demand. The FAO global food index increased significantly by 43% due to the Russia-Ukraine war in 2022. Similarly the prices of cereals and vegetal oils rose respectively by 54% and 88% in 2022. Although prices moderated in the last six months, the levels of food prices are still high at global level. The RBI initially fell behind the curve and got delayed in responding to the rising inflation because it was arguably focusing on reviving growth as the economy was recovering from the shock of the pandemic. Subsequently it started raising the policy repo rate and tightened liquidity conditions (as reflected in the increase in the 91-day T-bill rate in Fig. 2.10). By March 2023, it has raised the repo rate by 2.5% starting from a pandemic low of 4%. However inflation continues to hover around 6.5% and especially core inflation (non-food, non-fuel) has been highly persistent at around 6% for nearly 18 months now. This implies that inflationary pressures are deeply embedded within the system possibly because firms have been passing on the high input prices gradually given the weak demand conditions in the economy. Ironically CPI inflation is still high even though WPI inflation has significantly softened (to around 3–4%) because of the lowering of commodity prices as global supply constraints have eased up to a large extent by now. Given that the RBI’s target is 4% CPI inflation, it may need to maintain a contractionary monetary policy stance for rest of this year in order to bring inflation down to the target. But this also means that monetary policy will act against growth over the medium term.

Fig. 2.10 RBI’s monetary policy. Source CMIE database

30

2.3.2.4

S. M. Dev and R. Sengupta

Fiscal Situation

On the fiscal front, as discussed in Sengupta (2023), in the pre-pandemic financial year of 2019–20, the fiscal deficit of the central government alone was more than 4.5% of GDP, much higher than the 3% medium term target set by the Fiscal Responsibility and Budget Management (FRBM) Act. For several years the government has been struggling to spend within its means (see Fig. 2.11) During the pandemic period, the deficit shot up, first to 9.2% of GDP in 2020–21, and then to 6.9% in 2021–22. The consolidated central and state government deficits were as high as 13.3% and 10.2% of GDP in 2020–21 and 2021–22, respectively. Amidst slowing investment and exports, in response to the pressure to boost growth and create jobs, the government has increased its capex spending significantly in the last 3 years. Capital expenditure increased from Rs 5.93 lakh crore in 2021–22 to Rs 7.28 lakh crore (RE) in 2022–23 and to Rs. 10 lakh crore (BE) in 2023–24. As a result of persistently high deficits, the total central and state government debt amounted to 84% of GDP in 2022–23: an uncomfortably high level compared to an average of 74% in the period from 2010–11 to 2019–20. When the government presented the Union Budget for 2023–24, they adhered to the fiscal deficit target of 6.4% for 2022–23 and projected a fiscal deficit of 5.9% for 2023–24. While this has been widely appreciated as an attempt toward much needed fiscal consolidation, there has not been any clear and well-defined strategy that delineates how the government plans to bring fiscal deficit down to the announced target of less than 4.5% by 2025–26, which itself is on the higher side, given the FRBM target of 3%.

Fig. 2.11 Fiscal deficit as a percentage of GDP, from 2013–14 to 2024–25 (projected). Source Kathuria (2023)

2 The Indian Economy in the Post-pandemic World: Opportunities …

31

This implies that both on the monetary and fiscal aspects, the room left to maneuver policy in order to revive growth remains significantly limited.

2.3.3 Formal Sector Conditions During the pandemic the large firms in the private corporate sector had benefitted by gaining market share, as explained in Sect. 2.2. In 2022, the multiple global shocks such as the Russia-Ukraine war and China’s strict lockdowns led to a renewal of supply chain bottlenecks which in turn pushed up prices of several important commodities such as crude oil, natural gas, fertilizers and wheat. The resultant increase in cost pressures squeezed the profit margins of the companies which began declining (see Fig. 2.12). Even as input prices began increasing, exports which had played a crucial role in facilitating the recovery of the economy from the pandemic stopped growing toward the later part of 2022, owing to global headwinds (see Fig. 2.13). This too adversely impacted the performance of the corporate sector that had benefitted from the exports boom of 2021–22. The slowdown of exports is also problematic because the other important driver of growth, i.e., private sector investment has not picked up in a big way either and, the proverbial capex (capital expenditure) cycle continues to remain elusive. Figure 2.14a shows the total, real value of all investment projects under implementation over

Fig. 2.12 Net profit margin of listed, non-oil, non-financial firms. Source CMIE database

32

S. M. Dev and R. Sengupta

Fig. 2.13 Non-oil, non-gold exports. Source CMIE database

the period from March 2017 to December 2022.6 After the sharp decline in the lockdown period of 2020, projects under implementation picked up as the economy gradually reopened but started declining again roughly from March 2021 onwards. Projects under implementation act as a proxy for the general investment activity in the economy and as the figure shows, this remains sluggish. Total “under implementation” projects (real terms) in the private sector (see Fig. 2.15) alone show a similar declining trend in recent times, after a brief uptick in 2021–22. Financial sector: As mentioned in the previous section, the pandemic hit the Indian economy at a time when the financial sector, and in particular the banking sector was dealing with secularly declining credit growth due to heightened risk aversion in banks as well as in large (commercial) borrowers after years of a series of balance sheet crises. The pandemic which was an unprecedented shock to the economy in general dealt a further blow to credit growth and arguably worsened the risk aversion of the financial sector. During the first year of the pandemic, credit growth from all sources slowed down. Bank credit growth almost halved from a CAGR of 11.3% in the previous decade to 5.6%—the lowest in almost six decades (Sengupta & Vardhan, 2022). It recovered to 8.6% in the second year of the pandemic. Also drastic was the decline in credit growth from NBFCs, from 25.5% to 15.1% in the first year followed by a contraction in the second year. The NBFCs were already struggling in the pre-pandemic period, as mentioned in the previous section, and the pandemic was yet another massive 6

Nominal values have been converted into real by using consumer price index (CPI) of the corresponding quarter.

2 The Indian Economy in the Post-pandemic World: Opportunities …

33

0.92 0.90 0.88 0.86 0.84 0.82

Nov-26

Jul-26

Mar-26

Nov-25

Jul-25

Mar-25

Nov-24

Jul-24

Mar-24

Nov-23

Jul-23

Mar-23

Nov-22

Jul-22

Mar-22

Nov-21

Jul-21

0.78

Mar-21

0.80

Fig. 2.14 Total “under implementation” projects (real values, Rupees billion). Source CMIE database

0.34 0.33

0.32 0.31 0.3 0.29 0.28 0.27

Nov-26

Jul-26

Mar-26

Nov-25

Jul-25

Mar-25

Nov-24

Jul-24

Mar-24

Nov-23

Jul-23

Mar-23

Nov-22

Jul-22

Mar-22

Nov-21

Jul-21

0.25

Mar-21

0.26

Fig. 2.15 Total “under implementation” projects in the private sector (real values, Rs bn). Source CMIE database

34

S. M. Dev and R. Sengupta

blow to their balance sheets. The bond market credit however grew at a steady rate of 11% during the pandemic (see Table 2.11). By the second year of the pandemic, the health of the banking sector had improved substantially. This improvement in banks’ financials primarily came about due to multiple rounds of capital infusion in public sector banks by the government, resolution of bad assets by the Insolvency and Bankruptcy Code (IBC), and also due to the decline in credit growth rate discussed earlier. Two key indicators demonstrate the banking system’s progress. Successive waves of recapitalization have given banks enough resources to write off most of their bad loans. As a result, they have been able to bring down their gross NPAs from 11% of total advances in 2017–18 to 5.9% in 2021–22. NPAs for industrial credit have been reduced even more dramatically, from 23% to 8.4%. Even after these large write-offs, most banks retain comfortable levels of capital. This financial turnaround gave the banks the space to resume their business of extending credit. Since the pandemic period, credit growth has nearly doubled. In 2022 on average, bank credit growth was about 18%, and bond market issuances were also strong. Deposit growth, on the other hand, remained muted at slightly below 10% (see Fig. 2.16). Table 2.11 Growth (CAGR) of credit across sources, 2011–2022 Source

2011–2020 (%)

2020–2021 (%)

2021–2022 (%)

2011–2022 (%)

Bonds

15.5

11.0

11.2

14.7

Banks

11.3

5.6

8.6

10.6

NBFCs

25.5

15.1

−9.7

20.9

CPs

17.6

5.8

−3.3

14.4

ECB

15.6

−3.6

7.6

13.0

Total

12.9

6.0

8.2

11.8

Source Sengupta and Vardhan (2022)

Fig. 2.16 Year-on-year (YoY) credit and deposit growth of the banking sector

2 The Indian Economy in the Post-pandemic World: Opportunities …

35

The strong credit growth seems primarily driven by growth in unsecured consumer credit as well as home loans. Growth of credit to MSMEs remains strong on the back of the credit guarantee scheme which initiated by the government during the pandemic and extended several times since then (Dev & Sengupta, 2022b). There was also some uptick in credit demand due to capital expenditure in sectors such as renewable energy and logistics. Government expenditure on infrastructure such as roads has also been creating demand for credit from EPC contractors and construction companies.

2.4 Opportunities and Challenges Going Forward It is clear from the discussion so far that the macroeconomic fundamentals continue to be weak and now there are additional headwinds from the global economy given the renewal of uncertainty triggered by the financial market turmoil in the US. A report of the Confederation of Indian Industry (CII) says that India’s GDP can grow from the current $3 trillion to $5 trillion by 2026–27, to $9 trillion by 2030 and to $40 trillion by 2047 if its population is productively employed. According to RBI (2022a): “the pre-Covid trend growth rate works out to 6.6% (CAGR for FY13 to FY20) and excluding the slowdown years, it works out to 7.1% (CAGR for FY13 to FY17). Taking the actual growth rate of (−) 6.6% for FY21, 8.9% for FY22 and assuming growth rate of 7.2% for FY23, and 7.5% beyond that, India is expected to overcome Covid-19 losses by FY35. The output losses for individual

Fig. 2.17 Medium term growth path. Source RBI (2022a)

36

S. M. Dev and R. Sengupta

years have been worked out to be Rs 19.1 lakh crore, Rs 17.1 lakh crore and Rs 16.4 lakh crore for FY21, FY22 and FY23, respectively”. In other words, it will take as many as 12 more years to overcome the loss of income due to the pandemic. Therefore, significant efforts are needed to improve growth in the medium to long term. Figure 2.17 depicts a medium term GDP path as projected by the RBI.

2.4.1 Short and Medium Term Challenges The growth rate of GDP for FY 2024 is expected to be around 7%. The RBI estimates 6.5% for FY24 while the IMF predicts 5.9% in the same year. India has to first take care of the short and medium term challenges such as high and entrenched inflation, twin deficit (current account, fiscal) and debt problems. Monetary and fiscal policies would need to be managed effectively in order to tackle these challenges in the context of growing global uncertainties. Rangarajan and Srivastava (2023) discuss the immediate growth prospects for the Indian economy. According to them, if fiscal stimulus in the form of capex continues, growth rate can come closer to that of RBI’s prediction in FY24. If however revenue expenditure increases due to impending national elections, then growth is likely to be closer to 6%. A steady growth of 6% to 7% can be ensured in the medium term if fixed capital formation is raised by another 2 percentage points. In addition to the scars left by the pandemic, what are the structural challenges that make India’s growth story complicated? Against the background of two years of the pandemic, a fragmental geopolitical landscape and heightened global economic uncertainty, in this section we discuss the challenges and opportunities that the Indian economy is likely to face in the near future as well as longer term as it strives to achieve a sustained high growth rate and create new jobs. We also throw light on some of the important reforms and policy initiatives that must be implemented in order to achieve this objective.

2.4.2 Engines of Growth As discussed in the previous section, nearly three years after the pandemic the two main engines of growth—exports as well as private sector investment, are stuttering, similar to the immediate pre-pandemic period of 2019–20.

2.4.2.1

Investment

The government has also given a much needed capex push in its Union Budgets of 2022–23 and 2023–24, yet the translation of this into private investment is still not

2 The Indian Economy in the Post-pandemic World: Opportunities …

37

visible. Investment continues to be sluggish despite a sharp cut in corporate taxes introduced by the government in 2019. The persistent monetary policy tightening by the RBI in order to tame inflationary pressures might act as an additional drag on investment revival. The gross fixed capital formation (GFCF) as percent of GDP increased from 27.3% in FY21 to 28.9% in FY22 and to 29.2% in FY23. It increased only marginally in FY23 as compared to 28.5% of FY20, the pre-pandemic year. For growth to increase it is imperative that the GFCF/GDP ratio goes up. What is required perhaps to unleash the animal spirits of the private sector is policy certainty and creation of a level-playing field through government actions. The Indian government has undertaken several structural reforms in recent times such as announcement of privatization and asset monetization, tax reforms (GST introduction and corporate tax rationalization), the production linked incentive (PLI) scheme; insolvency and bankruptcy code (IBC) to improve the credit culture and resource allocation mechanism, labor reforms, and a fiscal policy focused on capex and infrastructure (RBI, 2022a). A lot more needs to be done on a sustained basis to create a conducive environment for private sector investment which in turn can create much needed jobs. Also attracting foreign companies to produce in India must be given a high priority now given that many of these companies are looking for alternatives to China and Russia. India needs to take full advantage of this opportunity.

2.4.2.2

Exports

The tepid economic recovery in FY22 was largely dependent on exports which grew exceptionally rapidly. However, as mentioned earlier, the growth rate of exports declined in FY23. Rising global interest rates, winding down of economic stimulus packages and consequent slowdown in global growth will likely have a negative impact on India’s trade activity going forward. Volatility in commodity prices and continued geopolitical tensions will make trade developments uncertain. If the US and Indian business cycles continue to diverge, meaning that the US goes into a slow growth while Indian economy grows relatively faster, India’s CAD will widen even more because exports will continue to slow down but the import bill may keep rising. The World Trade Organization (WTO) has pegged the global trade growth at 1% for FY23, down from 3.4% amidst rising apprehensions about a global slowdown. In recent times, services exports have picked up substantially as economies around the world have opened up fully after 2 years of the pandemic and this has been working in India’s favor given that services exports are our strength. However, this is likely to be a temporary development and will hit a plateau as global economy continues to slow down. Merchandise exports from India have already been contracting. Economic theory suggests that in the face of adverse terms of trade shock, a weaker currency helps in expenditure switching toward higher exports and lower imports and hence improves the trade balance. Weaker exchange rate boosts non-oil exports and helps reduce non-oil imports by increasing the price of imports. Since the US Fed began tightening monetary policy in 2022, the US dollar has been appreciating and

38

S. M. Dev and R. Sengupta

hence the Indian rupee has faced significant depreciation pressures. However, the RBI has been intervening actively in the forex market to prevent the rupee from depreciating. It lost roughly $100 billion in 2022 in this endeavor. If the rupee fails to follow when other EM currencies are depreciating, then India’s exports will lose competitiveness. Already, the rupee has appreciated significantly against other Asian currencies such as the South Korean won, the Thai baht and the Taiwanese dollar. If competitiveness is further eroded at a time when the global economic environment is turning difficult, export growth could really suffer. In this context, therefore, the RBI’s attempts to prevent the rupee from depreciating, might not be the most suitable policy reaction. What is required instead is a real depreciation of the rupee; instead, real effective exchange rate (REER) has been entirely flat in recent times. From a more general and broader perspective, notwithstanding the ongoing slowdown in exports, the way international trade stands now might present a historic opportunity for India to join the club of great exporting nations. In 2022 China, the main export engine of the world had begun locking down its factories owing to a rapid resurgence of Covid-19 cases. This resulted in international firms scouting for new production locations. In fact several large corporations started operations in other Asian countries such as Vietnam, in order to de-risk from China. Even as China opens up now, this trend may continue as MNCs look to diversify their operations across multiple geographical locations in order to avoid over exposure to China. This is the so-called “China plus one” strategy, under which MNCs plan to build more of their new factories outside of China. Indeed geopolitical considerations will increasingly drive global commerce in the future. On the other hand, Russia is still under tight economic sanctions imposed by the Western countries in the aftermath of its invasion of Ukraine. This implies that two large Asian countries are reducing their presence on the international trade landscape, thereby creating an unprecedented scope for India to attract international firms to produce and export from here. Moreover, geopolitical tensions between the US and China can also put India in a favorable position. India also has a fast-growing young workforce, compared to China’s shrinking and aging labor force. The government’s capex push has triggered a fair amount of infrastructure creation in the public sector. India also has world-class public digital infrastructure which is facilitating innovation, productivity improvements and access to services. In a nutshell the country has tremendous potential to catch up on China whose economy is five times as big. In order to take advantage of these opportunities, India needs a liberal, stable and consistent trade policy regime. Unfortunately government policies with respect to international trade have turned increasingly protectionist. Import tariffs have been going up since 2015. India’s import tariff rates (MFN-based average) increased from the lowest level of about 12% in 2008 to 15% in 2019. For the year 2018, China’s import tariff rate was 9.6% compared to India’s 13.5%. While 11.9% of the tariff lines had rates above 15% in 2010–11, that proportion has gone up to 25.4% in 2020–21. As mentioned in Mitra (2023), in 2017 and 2018, India doubled its import duties on beauty aids, watches, toys, furniture, footwear, kites and candles all of which

2 The Indian Economy in the Post-pandemic World: Opportunities …

39

are produced in labor intensive industries taking advantage of India’s vast pool of low-skilled labor but this was a clear acknowledgement of lack of competitiveness in these sectors. Import duties were also raised for electronics and communication devices and related inputs and parts which make it costlier to manufacture these items in India. There were further tariff increases in 2020 and 2021 on imports of electronics and automobile components, fabrics, agricultural products, etc. International trade today is entirely dependent on global value chains. Import duties especially on inputs hamper this process because they increase the cost of importing, thereby disrupting the production chain. Higher import duties convey to foreign firms that doing business in India is going to be costly and difficult for them. The government has also been imposing bans and taxes on exports in order to deal with surging inflation. In 2022 they banned wheat exports, imposed an export duty on steel products at the rate of 15%, increased the export duty on iron ore from 30 to 50%, and imposed a 20% export duty on rice varieties commanding a 28–30% share of annual exports. While the government’s bans and market interventions will do little to dent inflation, they are likely to damage growth by undermining exports. India is also not part of the trade pillar of America’s Indo-Pacific Economic Framework for Prosperity or the Comprehensive and Progressive Agreement for Trans-Pacific Partnership. It also refused to join the RCEP and so far does not have any bilateral trade agreements with its principal trading partners such as the US and EU. As a result, the economy does not enjoy the full benefits of vigorous competition and the prices of some components are higher than on the global markets. In order to compete with other peers such as Vietnam who are benefitting from the China + 1 strategy perhaps more than India at the current juncture, Indian policymakers will need to make an explicit export-oriented push and signal to the world that the country is open for business. So far that has not happened but this is one of the biggest opportunities.

2.4.2.3

Other Factors

Over and above harnessing the potential to export and reviving private sector investment, there are several other opportunities for India in the medium term to improve economic growth. A recent issue of the Economist magazine (Economist, 2022) says that as the pandemic recedes, four pillars are visible that might support growth in the next decade; (1) forging of a single national market through the GST; (2) an expansion of industry owing to the shift to renewable energy, and a move in supply chains away from China (3) improvements in technology, IT services and outsourcing industry; and (4) a high-tech, welfare safety-net for the hundreds of millions left behind by all this. Moreover, India is on its way to becoming Asia’s top financial technology (Fin Tech) hub with a staggering 87% Fin Tech adoption rate against the global average of 64%. The growth rate of Indian Payment systems like UPI (United Payments Interface) and Aadhar Enabled payment services (AePS) has been phenomenal. According to RBI (2022a), the long strides taken in the digital finance arena need to be leveraged

40

S. M. Dev and R. Sengupta

to promote growth. There are growing opportunities for new investment in areas like e-commerce, start-ups, renewables and supply chain logistics.7 India’s fast-growing young workforce is another advantage which few countries enjoy in today’s world. This alone can help India leapfrog to a high growth trajectory provided of course there are adequate and good quality education and skilling opportunities as well as meaningful employment options to absorb the youth.

2.4.2.4

Financial Sector

Finance is the backbone of any economy. India being a bank dominated economy the strong credit growth revival in the post-pandemic period (as discussed in the last section) is indeed welcome news after a period of lackluster credit growth. However, very little of this credit is going to large-scale industry or for financing investment. Lending to large industries has been stagnant in nominal terms during the last two years, implying that it has declined sharply in real terms, and there has been little lending for private sector investment. Over the last 2 years, bank lending to infrastructure has grown but this was fuelled mainly by public sector capex. Meanwhile, much of the lending to private industry has been in the form of working capital loans, necessitated by the increase in commodity prices, which has led to a sharp rise in the cost of holding inventories. A big reason behind the lackluster growth of industrial credit is because private sector investment has been sluggish for nearly a decade and continues to be so now. Firms seem to have finally used up much of their spare capacity. But the fundamental problems that led to the difficulties of the past decade still have not been resolved. There is still no framework that will reduce the risk of private sector investment in infrastructure. Nor is there any reassurance for the banks that if problems do develop, they can be resolved expeditiously, since the Insolvency and Bankruptcy Code (IBC, 2016) has been plagued by delays and other problems. Now, heightened global macroeconomic uncertainty, growing geopolitical tensions and uncertain recovery prospects of the domestic economy are likely to make matters worse. In fact in 2023, credit growth seems to be tapering off to some extent as aggregate demand in the economy has been slowing down. Additionally, the monetary contraction that is being implemented by the RBI to tackle inflation might also eventually dampen credit growth. To what extent the credit growth would decline and what impact the decline would have on economic output remains to be seen.

2.4.2.5

Policy Levers

One of the important policy levers for boosting growth is to use the fiscal space to encourage demand. However, as already mentioned in the previous section, fiscal space in India is likely to remain limited in the foreseeable future. And if growth 7

See Srivastava (2023a) on how digital transformation will help accelerate growth.

2 The Indian Economy in the Post-pandemic World: Opportunities …

41

starts to slow down, this will make the task of fiscal consolidation even more challenging. Lowering the deficit and debt to more sustainable levels is imperative for ensuring macroeconomic stability which in turn is an important precondition for growth especially amidst growing global uncertainty. As Table 2.12 shows the general government outstanding liabilities were less than 70% during the period from FY11 to FY18. But it accelerated to 89.4% in FY21. This is significantly higher than FRBM target of 60% and it is a risk for medium term macroeconomic stability. At least for the next few years, fiscal policy has to follow the path of consolidation which implies that there is not much room to use this level for stimulating growth, even if aggregate demand starts to slow down. This will be a serious short to medium term challenge for Indian economy, especially given that monetary policy, the other important policy level, will need to remain focused on bringing the CPI inflation down to the RBI’s target level of 4%. The government has rightly been focusing on capital expenditure in the last three budgets. In August 2020 they also outlined an infrastructure project pipeline to be implemented over the next five years, which will serve as one of the key drivers of faster economic growth. Using the data on annual nominal growth in tax revenue, government expenditure and GDP for the period 1981–82 to 2019–20, RBI (2022a) estimates general government (center + states) fiscal multipliers for total expenditure and its components (Table 2.13). The multiplier is more than one only for capital expenditure. It indicates that only capital expenditure leads to proportionately higher rise in GDP. Table 2.12 Fiscal Deficit and outstanding liabilities (% of GDP): center and states Year

Gross fiscal deficit Center

Outstanding liabilities States

Center

States

2011–12

5.9

2.0

51.7

23.2

2012–13

4.9

2.0

51.0

22.6

2013–14

4.5

2.2

50.5

22.3

2014–15

4.1

2.6

50.1

22.0

2015–16

3.9

3.0

50.1

23.7

2016–17

3.5

3.5

48.4

25.1

2017–18

3.5

2.4

48.3

25.1

2018–19

3.4

2.4

48.5

25.3

2019–20

4.7

2.6

51.6

26.7

2020–21

9.2

4.2 (PA)

61.7

31.1 (PA)

2021–22

6.7 (RE)

3.5 (BE)

58.1 (RE)

29.4 (BE)

2022–23

6.4 (BE)



59.5 (BE)



PA: Provisional Accounts; RE: Revised Estimates; BE (Budget estimates) Source RBI (2022a), Annual Report 2021–22

42

S. M. Dev and R. Sengupta

Table 2.13 Overall Fiscal Multipliers

Impact multiplier Total expenditure

0.72

Revenue Expenditure

0.79

Revenue Expenditure net of Interest Payments and subsidies

0.84

Capital expenditure

1.32

Source RBI (2022a)

At the same time there should be some balance between revenue and capital expenditure. Most of the expenditures on health and education are in revenue account. These expenditures on human capital should not be compromised. Fiscal consolidation must focus on raising tax revenue and as well as expenditure control. Tax/ GDP ratio has to be improved by measures such as widening the tax base, removing exemptions and unproductive subsidies, further reforms in GST, etc.

2.5 State Finances Consolidation in state finances is equally important as they spend more than the center. The state governments allocate significant amount of funds to agriculture in their budgets. They spend 60% of the total government expenditure, 70% of education and health spending, and a larger share in public capital expenditure. Capital expenditure by States/UTs in India is more than two-thirds of the total capital expenditure incurred by the general government. RBI released a report on “State Finances: A Study of Budgets.” Some of the findings of the report are the following. 1. In 2020–21, States’ consolidated gross fiscal deficit (GFD) rose to 4.1 percent of gross domestic product (GDP), from 2.6 percent in 2019–20 (Table 2.14). The rise, however, was short-lived and a reversion to consolidation was done in 2021–22, as shown by the provisional accounts (PA) taking the GFD down to 2.8 percent of GDP, as against the BE of 3.5 percent and RE of 3.7 percent for that year. This correction was brought about by higher than expected growth in both tax and non-tax revenues. The revenue deficit also declined from 1.9% in FY 21 to 0.4% in FY22. Table 2.14 Key deficit indicators: 31 states and Union Territories Gross fiscal deficit

2019–20

2020–21

2021–22 (PA)

2022–23 (BE)

2.6

4.1

2.8

3.4

Revenue deficit

0.6

1.9

0.4

0.3

Primary deficit

0.9

2.1

1.1

1.6

PA: Provisional Accounts; BE: Budget Estimates Source RBI (2023)

2 The Indian Economy in the Post-pandemic World: Opportunities …

43

Table 2.15 State-wise Gross Fiscal deficit and Revenue deficit States

2020–21

2021–22 (RE)

2022–23 (BE)

Revenue deficit

Gross Fiscal deficit

Revenue deficit

Gross Fiscal deficit

Revenue deficit

Gross Fiscal deficit

Andhra Pradesh

3.5

5.4

1.6

3.2

1.3

3.6

Assam

–0.4

3.3

1.0

9.5

–0.7

3.5

Bihar

1.9

5.1

5.5

11.4

–0.6

3.5

Chhattisgarh

2.0

4.5

0.3

3.8

–0.2

3.3

Gujarat

1.4

2.5

0.0

1.5

0.0

1.7

Haryana

3.0

3.8

1.4

3.0

1.0

3.0

Himachal Pradesh

0.1

3.6

–0.2

4.0

2.0

5.0

Jharkhand

1.0

5.0

–0.1

3.2

–1.8

3.0

Karnataka

1.1

3.9

0.3

2.4

0.6

2.7

Kerala

3.2

5.1

3.5

5.1

2.3

3.9

Madhya Pradesh 1.9

5.1

0.5

3.7

0.3

4.0

Maharashtra

1.5

2.6

1.0

2.8

0.7

2.5

Odisha

–1.7

1.8

–3.3

0.4

–2.5

3.0

Punjab

3.2

4.2

3.6

5.6

2.0

3.7

Rajasthan

4.3

5.9

3.0

5.2

1.8

4.4

Tamil Nadu

3.4

5.2

2.7

4.4

2.2

4.1

Telangana

2.3

5.1

–0.4

3.9

–0.3

4.0

Uttar Pradesh

0.1

3.3

–1.2

4.0

–2.0

3.7

West Bengal

2.3

3.4

–2.1

3.5

1.7

3.6

All States/UTs

1.9

4.1

0.9

3.7

0.3

3.4

Source RBI (2023)

2. Capital outlay of States showed a robust growth of 31.7 percent in 2021–22 (PA). Strong growth in tax and non-tax revenues, coupled with the advancement of payment by the Center of tax devolution and GST compensation, provided the required fiscal space to accelerate capital expenditure. The consolidated capital outlay of the States is budgeted to grow by 38.4 percent in 2022–23 (over 2021– 22 PA). The capital outlay to GDP ratio is expected to improve from 2.3 percent in 2021–22 (PA) to 2.9 percent in 2022–23 (BE). 3. In 2021–22 (RE), the gross fiscal deficit was more than 4% of GSDP for 6 states: Bihar (11.4%), Assam (9.5%), Punjab (5.6%), Rajasthan (5.2%), Kerala (5.1%) and Tamil Nadu (4.4%) (Table 2.15). Revenue deficit is also high in some of these states. 4. States’ debt to GDP ratio increased sharply at end-March 2021 to meet pandemic related expenditure. This ratio is estimated to decline slightly by end-March 2022

44

5.

6.

7.

8.

S. M. Dev and R. Sengupta

but is budgeted higher at 29.5 percent by end-March 2023. At a disaggregated level, the ratio is expected to be higher than 25 percent 17 for 26 States and UTs at end-March 2023. The fiscal health of States has rebounded from pandemic induced stress due to buoyant revenue collections and prudent expenditure management. Improvement in key deficit indicators has enabled States to reduce their outstanding liabilities. The report cautions that a major risk for states is the likely reversion to the old pension scheme by some States. By postponing the current expenses to the future, States risk the accumulation of unfunded pension liabilities in the coming years. It also indicates that all the States to continue with the current capex push, to sustain the quality of expenditure and maintain capital assets so that their longevity improves. In addition, States should also step-up capex in areas like research and development and green energy. Climate change is another area that deserves special attention in the coming years. The report also suggests that state governments should set up Finance Commissions (SFC) in a regular and timely manner to decide on the assignment of taxes, fees and other revenues to local governments.

2.6 Freebies Recently, there has been a lot of discussion on freebies given by the states.8 To derive an estimate of freebies, RBI (2022f) collated data on major financial assistance/ cash transfers, utility subsidies, loan or fee waivers and interest free loans announced by the states in their latest budget speeches (i.e., for FY23). These estimates show that the expenditure on freebies ranges from 0.1 to 2.7% of the GSDP for different states (Table 2.16). The freebies as percent of GSDP were more than 2% for some of the highly indebted states such as Punjab and Andhra Pradesh (Table 2.16). The budgets may not give the entire picture of freebies as some of them happen off budget, beyond the pale of FRBM tracking (Subbarao, 2022).9 The amount of freebies could be even higher if we take into account these extra-budgetary subsidies. Some kind of social protection measures for the poor and vulnerable groups, and informal workers are needed in any country. However, it should not be financed by increasing debt. Rangarajan (2022) suggests that overall fiscal support to such schemes should be limited to less than 10% of the total expenditure of the central government and state governments until their revenue to GDP or GSDP ratios is increased in a sustainable manner. 8

Singh, N.K. (2022), “Freebies are a passport to fiscal disasters”, Indian Express, April 22, 2022 https://indianexpress.com/article/opinion/columns/freebies-are-a-passport-to-fiscal-disaster7879244/; 9 Subbarao (2022), “States, Freebies and the costs of fiscal profligacy”, The Hindu, June 27, 2022, https://www.thehindu.com/opinion/lead/states-freebies-and-the-costs-of-fiscal-pro fligacy/article65573164.ece; Rangarajan (2022), “Good and Bad Freebies”, Indian Express, June 16, 2022.

2 The Indian Economy in the Post-pandemic World: Opportunities …

45

Table 2.16 Freebies Announced by the States in 2022–23 (As a percent of GSDP)

(As a percent of Revenue Receipts)

(As a percent of Own Tax Revenue)

Andhra Pradesh

2.1

14.1

30.3

Bihar

0.1

0.6

2.7

Haryana

0.1

0.6

0.9

Jharkhand

1.7

8.0

26.7

Kerala

0

0

0.1

Madhya Pradesh

1.6

10.8

28.8

Punjab*

2.7

17.8

45.4

Rajasthan

0.6

3.9

8.6

West Bengal

1.1

9.5

23.8

Note Dhasmana (2022). “Not all states are so financially weak that they can’t announce freebies.” Business Standard. April 2022 Source RBI (2022d) based on budget documents of the state governments

2.6.1 Structural Transformation One of the long standing challenges facing the Indian economy is structural transformation in agriculture, industry and services.

2.6.1.1

Reforms in Agriculture Sector

The Economic Survey (GOI, 2023) calls for the reorientation of agriculture due to challenges such as climate change, rising input costs, fragmented landholdings, suboptimal farm mechanization, low productivity and disguised unemployment. The Reserve Bank of India (RBI 2022b) report on currency and finance says that “the agriculture sector suffers from low capital formation, declining R&D, low crop yields, inadequate crop diversity and intensity, with excessive dependence on subsidies and price support schemes.” There has been significant progress in the country’s agricultural development since independence with a remarkable transformation from food scarcity to food self-sufficiency. However, the Green Revolution approach led to water logging, soil erosion, ground water depletion and unsustainability of agriculture. The policies today are still based on “deficit” mind set of the 1960s. Also, the procurement, subsidies and water policies are biased toward rice and wheat. A change of narrative is required in Indian agriculture toward more diversified, high value production, better remunerative prices and farm incomes, marketing and trade reforms, high productivity with less inputs, less chemical and pesticide-based production approaches, inclusive in terms of women and young farmers, small farmers, nutrition sensitive and sustainable models (Dev, 2023).

46

S. M. Dev and R. Sengupta

2.6.1.2

Industry and Services

In a larger context, structural transformation of the economy from agriculture toward manufacturing and services sectors can be of critical importance when it comes to generating employment opportunities. India’s development trajectory so far stands out among other countries because the economy has transformed from agriculture to services bypassing the industrial route. However, there is a deep disconnect between the shares of GDP and shares of employment across sectors. In terms of GDP, there has been structural change from agriculture to services but in terms of employment, agriculture is still the largest employer at 46% (Table 2.17). Moreover employment may have shifted over the years from agriculture to services to some extent but not adequately to manufacturing. Indeed, of particular concern is the inability of the Indian manufacturing sector to absorb labor. The share of manufacturing in employment was only 11% in 2019–20 (Table 2.17). This is problematic because getting absorbed in the services sector often requires specialized skills which the vast majority of the workforce may not possess. Therefore a widespread manufacturing push is much needed to generate millions of jobs.10 There are two sources of productivity. One is productivity increase within sectors, and the other is shifting workers from low productivity sectors to high productivity sectors. India must focus on both sources to raise growth and employment. For the manufacturing sector, production Linked Incentive (PLI) schemes can improve performance. However, more efforts are required to improve the manufacturing sector. Table 2.17 Share in Gross Value Added (GVA)and Employment 2019–20 (%) 2019–20 share in GVA

2019–20 (%) share in employment

Agriculture and Allied Activities

15.0

45.6

Manufacturing

17.1

11.2

7.9

11.6

Industry (Secondary Sector)

29.7

23.7

Trade, Hotels and restaurants, transport, storage and communications

20.3

18.7

Financing, real estate, business services

21.9

3.1

Community, social personal services

13.1

8.9

Services (tertiary)

55.3

30.7

Construction

Non-agriculture Total

85.0

54.4

100.0

100.0

Source PLFS 2019–20 and National accounts Statistics 10

See Mitra (2023) for arguments in favour of manufacturing-fed and export-led growth.

2 The Indian Economy in the Post-pandemic World: Opportunities …

47

Another view is that we should invest more in services sector as scope in manufacturing sector is limited for employment creation (Rajan & Lamba, 2023). There are indeed a lot of opportunities for India in the service sector. Brand and customer centricity are also important here (Dev, 2022). Growing start-ups including unicorns in manufacturing and services can also be part of this effort. At the same time, both manufacturing and services have to be developed together. A study by Chanda (2017) deals with the interdependence between services and manufacturing sectors and argues that a vibrant service sector should be seen as an enabler for the manufacturing sector and not as a competitor to manufacturing. Its 3year action plan (Aayog, 2017) also indicates that India has the advantage of walking on two legs: manufacturing and services. It offers specific proposals for jumpstarting some of the key manufacturing and services sectors, including apparel, electronics, gems and jewellery, financial services, tourism and cultural industries and real estate. Among other things, it recommends the creation of a handful of Coastal Employment Zones, which may attract multinational firms in labor-intensive sectors away from China to India. India has a major advantage of demographic dividend. However, it might soon become a liability if enough productive jobs are not created. And meaningful structural transformation is key to employment generation. According to the UN Population Statistics database, India will add another 183 million people to the working age group of 15–64 years between 2020 and 2050. Thus, a whopping 22% of the incremental global workforce over the next three decades will come from India. This further underscores the importance of creating productive employment opportunities which might however prove challenging especially now given the scarring of the pandemic. A crucial element of this structural transformation is the role played by the MSMEs who form a major chunk of manufacturing and services in India and hence can play an important role in providing large-scale employment and also reducing income, social and reduce regional disparities. Yet, many aspects of government policy are at best scale neutral and do not really favor the MSMEs. This sector does not get adequate, timely and affordable availability of institutional credit. The policies have to give a positive bias toward MSMEs so that they can be a driver for employment generation. Short and long-term initiatives are required specifically for the development of MSMEs.

2.6.1.3

Health and Education

Access to good health and education is essential for improving human capital. Yet India’s progress on both these aspects leaves much to be desired. We also have a great quality dichotomy in both these sectors. There are islands of excellence that can compete internationally in education while vast majority of them churn masses of children with poor learning achievements and unemployable graduates. One has to fix this dichotomy in health and education.

48

S. M. Dev and R. Sengupta

Few years back, the Deputy Prime Minister of Singapore cautioned about school education in India. He said, “schools are the biggest crisis in India today and have been for a long time. Schools are the biggest gap between India and East Asia. And it is a crisis that cannot be justified.”11 Skill deficiency of workers is well known. While promotion of technology and knowledge economy will add to growth, one cannot have “demographic dividend” for growth with low human capital. Apart from enhancing productivity and boosting private investment, education and skill development will be the biggest enablers for achieving this dividend. In order to have structural change from agriculture to non-agriculture and from unorganized to organized, education and skill development are needed. Moreover, women’s labor participation rates have been low. Raising women’s human capital and participation rates need to be prioritized in order to improve economic growth.

2.6.2 Climate Change Climate change is now a serious challenge for India’s long-term growth. Reducing carbon emissions and accelerating energy transition is a challenge as well as an opportunity. In the recent COP20 meeting at Glasgow, Prime Minister Narendra Modi announced that India will aim to attain net zero emissions by 2070. Net zero, or becoming carbon neutral, means not adding to the amount of greenhouse gasses in the atmosphere. China has announced plans for carbon neutrality by 2060, while the US and EU aim to hit net zero by 2050. PM also announced that India will draw 50% of its consumed energy from renewable sources by 2030, and cut its carbon emissions by a billion tons by the same year. India wants commitments of developed countries on providing finance, transfer of technology and emission reductions due to historically high consumption patterns. Climate justice is another issue. Developed countries are historically responsible but rich in developing countries also have to pay for their consumption patterns. Acknowledgements We thank Dr. Govinda Rao and other participants for their comments on an earlier version of the paper presented at the conference on “India’s Contemporary Macroeconomic Themes”, honouring and celebrating the 90th Birth Anniversary of Dr. C. Rangarajan, held at the Madras School of Economics, April 21 and 22, 2023. The chapter also draws from the authors’ IGIDR working paper http://www.igidr.ac.in/pdf/publication/WP-2023-006.pdf

11

First Lecture of Niti Aayog’s ‘Transforming India” initiative, August 26, 2016.

2 The Indian Economy in the Post-pandemic World: Opportunities …

49

References Bhalla, S., Bhasin, K., & Virmani, M. A. (2022). Pandemic, poverty and inequality. IMF Working Paper no. 2022/069, Washington DC. Bhattacharya, K., Mitra, S., Pal, S., & Saha, B. (2017). Reviving the informal sector from the throes of demonetisation. Ideas for India, February 13, International Growth Centre, London School of Economics and Political Science, UK. https://www.ideasforindia.in/topics/macroeconomics/ reviving-the-informal-sector-from-the-throes-of-demonetisation.html Dreze, J. (2023). Wages are the worry, not just unemployment. Indian Express, April 13, 2023–05– 05. Dreze, J. & Somanchi, A. (2023). Weighty evidence? Poverty Estimation with missing data. Ideas for India, April, 2023. Chanda, R. (2017). Services for Indian Manufacturing. In Dev, S.Mahendra (ed., 2017), India Development Report 2017, Oxford University Press, New Delhi. Dev, S. M. (2023). Agriculture and Rural Areas in Budget 2023–24: A need for Comprehensive Approach for Transformation. Economic and Political Weekly, March 25, 2023. Dev, S. M. (2022). How Budgets can generate higher growth and jobs. Indian Express, January 29, 2022. Dev, S. M. R., & Sengupta, R. (2022a). Covid-19: Impact on the Indian economy,” India Development Report, Mumbai, Oxford University Press and Indira Gandhi Institute of Development Research (IGIDR). Dev, S. M. R., & Sengupta, R. (2022b). Covid-19 Pandemic: Impact, Recovery, and the Road Ahead for the Indian Economy,” Indira Gandhi Institute of Development Research, Mumbai Working Papers 2022b-016, Indira Gandhi Institute of Development Research, Mumbai, India. Felman, J., & Subramanian, A. (2019). India’s Great Slowdown: What Happened? What’s the Way Out?” CID Faculty Working Paper No. 370. https://www.hks.harvard.edu/centers/cid/publicati ons/faculty-working-papers/india-great-slowdown GOI. (2023). Economic Survey, 2023. Government of India. GOI. (2009). Report of the Expert Group to Review the Methodology for Estimation of Poverty, Chaired by S.D. Tendulkar, Planning Commission, Government of India. GOI. (2014). Report of the Expert Group to Review the Methodology for Measurement of Poverty. Chaired by C. Rangarajan, Planning Commission, Government of India. Himanshu. (2022). Our employment data should be interpreted cautiously. Mint, 16 June 2022. Mehrotra, S., & J.K. Parida (2019). India’s Employment Crisis: Rising Education Levels and Falling non-agricultural job growth” https://cse.azimpremjiuniversity.edu.in/wp-content/uploads/2019/ 10/Mehrotra_Parida_India_Employment_Crisis.pdf Mitra, D. (2023). Manufacturing-fed and export-led growth for gainful employment and skill creation. paper presented at Columbia Summit on Indian Economy, New York, March 24–25, 2023. Mukherji, S. (2022). Tepid rural wage growth poses stiff challenges for economic recovery. Business Standard, April 20, 2022. Niti, A. (2017). India Three Year Action Agenda 2017–18 to 2019–20. Government of India Kathuria, R. (2023). Pre-election Budget 2023–24: Short on extravagance, long on substance. Ideas for India, February 22, International Growth Centre, London School of Economics and Political Science, UK. https://www.ideasforindia.in/topics/macroeconomics/pre-election-budget-202324-short-on-extravagance-long-on-substance.html Nagaraj, R., Sapre, A., & Sengupta, R. (2020). Four Years After the Base-Year Revision: Taking Stock of the Debate Surrounding India’s National Accounts Estimates. India Policy Forum, National Council of Applied Economic Research, 16(1), 55–107. Nagaraj, R., & Srinivasan, T. N. (2017). Measuring India’s GDP growth: Unpacking the Analytics and Data Issues behind a Controversy that Refuses to Go Away. India Policy Forum (NCAER), Vol. 13, July 2017.

50

S. M. Dev and R. Sengupta

OPHI and University of Oxford (2018). Global Multi-dimensional Poverty Index 218. published by Oxford Poverty and Human Development Initiative (OPHI) and University of Oxford, UK. Panagariya, A. with Vishal More (2023). Poverty and Inequality Before and after Covid. paper presented at Columbia Summit on Indian Economy, New York, March 24–25, 2023 Rajan, R., & Lamba, R. (2023). How services can create next wave of India’s Globalisation. Times of India, April 11, 2023. Rangarajan, C., & Mahendra Dev, S. (2022). Poverty in India: Measurement, Trends and Other Issues. in Hashim, S., Mukherjee, R., & B. Mishra (ed.2022). Perspectives on Inclusive Policies for Development. Springer. Rangarajan, C., & Srivastava, D. K. (2023). What are India’s Immediate Growth Prospects. The Hindu, March 15, 2023. RBI. (2021). RBI Bulletin May, 2021, Reserve Bank of India, Mumbai. RBI. (2022a). Report on Currency and Finance, Reserve bank of India, April, 2022. RBI. (2022b). Financial Stability Report. December, Reserve bank of India. RBI. (2022c). RBI Bulletin July, 2022a, Reserve Bank of India, Mumbai. RBI. (2022d). Scars of the pandemic. Report on Currency and Finance, 2022. RBI. (2022e), Annual Report 2021–22, Reserve Bank of India, Mumbai. RBI. (2022f). State finances: A Risk Analysis. RBI Bulletin, June, 2022. RBI. (2023). State Finances: A Study of Budgets. RBI, January 2023. Sinha Roy, S., & Van Der Weide, R. Poverty in India Has Declined Over the Last Decade But Not as Much as Previously Thought’, World Bank Policy Research Working Paper 9994. Sengupta, R. (2023). Budget 2023–24: Fiscally conservative but lacking economic strategy. Ideas for India, February 27, International Growth Centre, London School of Economics and Political Science, UK. https://www.ideasforindia.in/topics/macroeconomics/budget-2023-24fiscally-conservative-but-lacking-economic-strategy.html Sengupta, R., & Harsh, V. (2022). India’s credit landscape in a post-pandemic world. Indira Gandhi Institute of Development Research, Mumbai Working Papers 2022–019, Indira Gandhi Institute of Development Research, Mumbai, India. Sengupta, R., Son, L.L., & Harsh, V. (2022). A Study of the Non-Banking Finance Companies in India. Indira Gandhi Institute of Development Research, Mumbai Working Papers 2022–009, Indira Gandhi Institute of Development Research, Mumbai, India. Sengupta, R., & Harsh, V. (2021). Consumerisation of banking in India: Cyclical or structural? Ideas for India, July 23, International Growth Centre, London School of Economics and Political Science, UK. https://www.ideasforindia.in/topics/money-finance/consumerisation-of-bankingin-india-cyclical-or-structural.html Sengupta, R., & Harsh, V. (2020a). Policymaking at a time of high risk-aversion. Ideas for India, April 6, International Growth Centre, London School of Economics and Political Science, UK. https://www.ideasforindia.in/topics/money-finance/policymaking-at-a-timeof-high-risk-aversion.html Sengupta, R., & Harsh, V. (2020b). The Indian corporate bond market: From the IL&FS default to the pandemic. The Leap Blog, August 7. Sengupta, R., & Harsh, V. (2020c). The pandemic and the package. Ideas for India, June 4, International Growth Centre, London School of Economics and Political Science, UK. https://www. ideasforindia.in/topics/macroeconomics/the-pandemic-and-the-package.htm Sengupta, R., & Harsh, V. (2019). How banking crisis is impeding India’s economy, East Asia Forum, October 3. https://www.eastasiaforum.org/2019/10/03/banking-crisis-impedes-indiaseconomy/ Sengupta, R., & Harsh, V. (2017). This time it is different: Non-performing assets in Indian banks. Economic & Political Weekly, 52(12), 25 March 2017. Sengupta, R. (2016). A macro view of India’s currency ban, Ideas for India, November 16, International Growth Centre, London School of Economics and Political Science, UK. https://www. ideasforindia.in/topics/macroeconomics/a-macro-view-of-india-s-currency-ban.html

2 The Indian Economy in the Post-pandemic World: Opportunities …

51

Shah, A. (2016). A monetary economics view of the demonetization, Ideas for India, International Growth Centre, London School of Economics and Political Science, UK. https://www.ideasfori ndia.in/topics/money-finance/a-monetary-economics-view-of-the-demonetisation.html Srivastava, D. K. (2023a). Balancing Growth and Fiscal Consolidation, Economic and Political Weekly, Vol.58, No.12, March 25, 2023 Srivastava, D. K. (2023b). How Digital Transformation will help India accelerate its growth in the coming year, EY Economy Watch, March, 25 April. Subramanian, S. (2019). What is happening to rural welfare, poverty and inequality in India, India Forum, 27th November 2019. Vardhan, H. (2021). A decade of credit collapse in India. Ideas for India, June 25, 2021. https:// www.ideasforindia.in/topics/money-finance/a-decade-of-credit-collapse-in-india.html

Chapter 3

India’s Economy in the Twenty-First Century: Role of State-Differentiated Demographic Dividend D. K. Srivastava, Muralikrishna Bharadwaj, Tarrung Kapur, and Ragini Trehan

Abstract A number of recent studies have highlighted that in the twenty-first century, India is slated to become the global growth leader, maintaining relatively high growth rates over several decades. This is estimated to result in India becoming the second largest economy in terms of its size of GDP as measured in PPP terms by 2039–40 and the largest economy by 2051–52 (EY, 2023). Even in market exchange rate terms, India is estimated to become the second largest economy by 2075 (Goldman Sachs, 2022). This robust growth performance is largely dependent on the interface of India’s unfolding demographic dividend with its saving and investment rates. Saving rate tends to be high when dependency rate is low and the available population in the working age group is productively employed. In this paper, we have highlighted that the demographic profile in the next few decades is expected to be noticeably different across Indian states. States that are at present considered relatively more developed in economic terms are also the states that would be aging ahead of the states that are presently less developed. To some extent, the growing discrepancy in the age profiles of different states would be moderated by inter-state migration. Using the differential features of the evolution of age structure across states, policymakers both at the central and state levels can orient their policies to maximize overall and state-specific growth by finetuning over time, their expenditure

D. K. Srivastava Chief Policy Advisor, EY India and Formerly Director, Madras School of Economics, Gurugram, India e-mail: [email protected]; [email protected] M. Bharadwaj · T. Kapur (B) · R. Trehan Senior Manager, Tax and Economic Policy Group, EY India, Gurugram, India e-mail: [email protected]; [email protected] M. Bharadwaj e-mail: [email protected] R. Trehan e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 D. K. Srivastava and K. R. Shanmugam (eds.), India’s Contemporary Macroeconomic Themes, India Studies in Business and Economics, https://doi.org/10.1007/978-981-99-5728-6_3

53

54

D. K. Srivastava et al.

prioritization with respect to education, health, and skill building. In terms of urbanization and spread of infrastructure facilities also, suitable messages can be derived for optimal allocation of infrastructure investment across space and over time. Keywords Indian economy · Demographic dividend · Urbanization · Saving-investment rate · GDP growth · Growth-leading states

3.1 Introduction Recently, there has been considerable interest in projecting the size of the Indian economy as it evolves during the twenty-first century. Policymakers in India have drawn attention to India’s performance during the Amrit Kaal1 covering the 25-year period from 2022–23 to 2047–48, marking the completion of 100 years of independence. Two important international studies brought out by the OECD (Guillemette & Turner, 2021) and Goldman Sachs (2022) provide comparative performance of the Indian economy vis-à-vis. other leading global economies. The OECD study provides projections up to 2060 in purchasing power parity (PPP) terms. The Goldman Sachs study provides projections up to 2075 in market exchange rate (MX) terms. Three India-specific studies in this context have been released by EY India (2023), FICCI-McKinsey (2022), and Morgan Stanley (2022). In the August 2022 issue of the EY publication titled “Economy Watch,” it was highlighted that India would overtake the US in terms of relative size by 2039–40 and the Chinese economy by 2057–58 in PPP terms. An updated version of these projections was utilized in the document released by EY India titled “India@100: realizing the potential of a US$26 trillion economy”2 released at the World Economic Forum on January 19, 2023. The central driving force for India’s comparatively robust economic performance is its demographic dividend as reflected by its bulging share of working age population during the next several decades. This is closely linked to the availability of investible resources in the form of higher savings. In the present paper, we bring out the contribution of states in the evolving growth story of India highlighting that the demographic dividend experience for major Indian states is going to be quite different. Some of the present-day developed states may well be past the peak of their demographic dividend while some of the less developed states may experience the peaking of the share of working age population in the next few decades. This feature will have significant policy implications in terms of the focus of policy attention toward expanding health, education, and skilling facilities for the populations of these states.

1 2

‘Amrit Kaal’ is a term drawn from Vaidic astrology, indicating a uniquely auspicious period. https://www.ey.com/en_in/india-at-100.

3 India’s Economy in the Twenty-First Century: Role …

55

This paper is divided into seven sections. Apart from the introduction, Sect. 3.2 provides a brief review of recent studies undertaken by various institutions for projecting the size of the Indian economy in the twenty-first century. Section 3.3 discusses in detail, the assumptions and findings of the EY India (2023) study. Section 3.4 contains a discussion on two key growth drivers linked to saving and investment rates, and emerging demographic trends. Section 3.5 provides a statewise differentiation of the demographic trends and highlights their implications. Based on these state-wise population patterns, Sect. 3.6 provides messages for policy formulation at the central and the state levels. Section 3.7 provides concluding observations.

3.2 Size of the Indian Economy in the Twenty-First Century: Review of Recent Studies In this section, we provide a brief outline of the findings of five major recent studies projecting India’s GDP as indicated above. The OECD Study (Guillemette & Turner, 2021) The OECD Study (Guillemette & Turner, 2021) focused on projecting the potential or trend output growth based on a long-term model in which the main determinants of growth include growth of capital stock and its productivity, growth of labor force and its productivity, and the pace of technological progress (Guillemette & Turner, 2018; OECD, 2014; Johansson et al., 2013). Table 3.1 gives the estimated growth rates for a selected group of countries in the OECD baseline scenario. In the case of the world, China, India, and the US, average growth rates progressively fall, decade after decade, although the rate of this decline eases after the 2030s. The main reason for the falling growth rates is the declining marginal productivity of capital, the increasing share of consumption of fixed capital (CFC), and the falling contribution of technological progress. Table 3.1 Period average growth rates (%) for OECD baseline scenario: 2022–2060 Period 2022 to 2025 2026 to 2030 2031 to 2035 2036 to 2040 2041 to 2045 2046 to 2050 2051 to 2055 2056 to 2060

China 4.9 3.7 2.9 2.1 1.6 1.3 1.3 1.2

Germany 2.0 0.7 0.7 0.8 0.9 0.9 0.9 1.0

Japan 1.1 0.6 0.5 0.4 0.3 0.2 0.4 0.5

US 2.2 1.6 1.5 1.4 1.4 1.3 1.3 1.3

Source (basic data) Guillemette and Turner (2021) Note For India, data is on fiscal year basis. 2022 implies 2022–23 and so on

India 8.2 5.9 4.7 3.8 3.1 2.7 2.4 2.3

World 3.7 2.7 2.3 1.9 1.7 1.5 1.5 1.5

56

D. K. Srivastava et al.

Table 3.2 gives the estimated size of the respective GDPs for these selected economies as per the OECD baseline scenario. The size of the Indian economy in PPP terms is projected to reach $30.9 trillion by 2047–48, from $10.8 trillion in 2022–23, making it the second largest economy after China. The Goldman Sachs Study (2022) Goldman Sachs (2022) in their study titled “The Path to 2075—Slower Global Growth, But Convergence Remains Intact” has provided growth rates of major economies up to 2079 based on GDP measured in market exchange rate terms. As per their estimates, India is projected to have a GDP size of US$52.5 trillion by 2075, overtaking the US (Table 3.3). The study recognizes the impact of slowing down of population growth rates and aging of the population across countries. On the capital side, they utilize the estimates of initial capital stock, growth in capital stock, and depreciation rate. Productivity growth is modeled with reference to US productivity growth and adding to it, two components that are derived with reference to the relativity of a country’s per-capita GDP vis-à-vis. US per-capita GDP, and relative growth momentum in the past decade.

Table 3.2 Size of the economy as per the OECD baseline: 2022–2060 (Size in terms of GDP measured in PPP terms; PPP$ trillion) Period 2023 2025 2030 2035 2040 2045 2050 2055 2060

China 28.3 30.8 37.0 42.6 47.3 51.2 54.8 58.4 62.1

Germany 4.3 4.4 4.6 4.7 4.9 5.1 5.4 5.6 5.9

Japan 5.4 5.5 5.6 5.8 5.9 6.0 6.1 6.2 6.3

US 21.7 22.4 24.3 26.2 28.1 30.0 32.1 34.3 36.5

India 10.8 12.5 16.6 20.9 25.1 29.2 33.4 37.6 42.2

World 116.6 124.2 142.0 159.1 175.2 190.4 205.4 221.1 237.8

Source (basic data) Guillemette and Turner (2021) Notes (1) For India, data is on fiscal year basis. 2022 implies 2022–23 and so (2) The magnitudes are given in trillions of PPP$

Table 3.3 Size of the economy as per the Goldman Sachs estimates: GDP in US$ trillion 2020

Decade

World 86.6

2030

121.4

2040

171.6

2050

227.9

2060

US 21.8

China 15.5

6.6

27.0

24.5

13.2

32.0

34.1

22.2

37.2

41.9

291.4

33.2

42.8

48.6

2070

363.9

45.8

48.6

54.8

2075

402.5

52.5

51.5

57

Source (basic data) Goldman Sachs (2022)

India 2.8

3 India’s Economy in the Twenty-First Century: Role …

57

The Morgan Stanley Study (2022) Morgan Stanley, in its study titled “Why this is India’s decade?” released on October 31, 2022 has also projected the size of the Indian economy up to 2031–32 under three scenarios namely, base, bear, and bull (Table 3.4). In their base case scenario, continued policy reforms which focus on improving the investment climate are expected to lay the foundations for India’s real GDP growth to move higher to an average of 6.5% over the next 10 years. The Indian economy is expected to expand by about US$400–500 billion annually over the next 10 years, reaching a size of US$7.9 trillion—a feat that has only been achieved by the US and China thus far. This would make India the third-largest economy (up from fifth currently) in the world by 2031–32. The FICCI-McKinsey Study (2022) The FICCI-McKinsey (2022) study titled “India’s century—achieving sustainable inclusive growth” was released in December 2022. This study establishes that if the Indian economy shows a real GDP growth of 6.1% (scenario 2), it would result in a nominal GDP magnitude of US$28 trillion by 2047–48. However, realizing the aspiration of reaching 60 crore jobs would likely require in their estimation, a growth of 7.7% annually. Achieving this level of growth would likely make India a US$40 trillion economy (Table 3.5). This would mean that India could be close to becoming a high-income nation, with a per-capita income of above US$12,000. Table 3.4 Size of the Indian economy (Nominal GDP in US$ trillion, and nominal GDP per-capita in US$) Variable GDP (US$ trillion) Per-capita GDP (US$)

2021-22 3.2 2278

2031-32 Bull 9.5 6277

Base 7.9 5242

Bear 6.2 4123

Source Morgan Stanley (2022)

Table 3.5 India’s 2047 GDP estimates for various growth scenarios Variable Real GDP growth (% ann) Nominal GDP (US$ trillion)

Source FICCI-McKinsey (2022)

Scenario 1 5.2-5.4 23.0

Scenario 2 6.1 28.0

Scenario 3 7.7 40.0

58

D. K. Srivastava et al.

3.3 India’s US$26 Trillion Economy by 2047–48: Assumptions and Findings (EY Study) In this study, projections of India’s GDP growth are undertaken for the period 2022– 23 to 2060–61. In these projections, for the period from 2022–23 to 2027–28, real growth rates are taken from the October 2022 issue of the IMF World Economic Outlook. After that, growth rates are projected under three alternative scenarios, namely, S1, S2, and S3. Projection Methodology The key determinant of growth in India is investment which is largely financed by domestic savings. Labor supply is going to be abundant in India due to the increasing number of working age persons in line with the unfolding demographic trends. In the projections, nominal savings rate serves as the starting point. This is converted into the nominal investible resources to GDP ratio by adding the current account deficit and change in foreign exchange reserves relative to GDP. The nominal investible resources to GDP ratio is then converted into its real counterpart by using the differential in the deflators pertaining to gross capital formation (GCF) vis-à-vis. GDP. Then, the investible resources are converted from gross terms to net fixed capital formation which is relevant for determining growth. Net fixed capital formation is then related to the GDP growth rate by a multiple as measured by the ratio of net fixed capital formation to GDP relative to GDP growth. This factor captures the impact of technological progress or an improvement in multi-factor productivity. From real GDP projections, nominal GDP is derived by using projections regarding IPD-based deflator. Nominal GDP is then converted into PPP terms and MX terms using alternative depreciation profiles.3 Two changes pertaining to (1) nominal savings rate and (2) impact of technological progress factor are critical differentiators between alternative simulations. Between S1 and S2, assumptions regarding nominal savings rate are changed, and between S2 and S3, assumptions regarding both nominal savings rate and technological progress have been changed. Chart 3.1 shows the time profile of real GDP growth rates at constant 2011–12 prices under alternative simulations. In the base scenario (S1), it increases to a peak of over 6% in the early 2030s and then falls to a little less than 5%. This base run growth profile is progressively uplifted in simulations S2 and S3. To smoothen the annual fluctuations, it is the trend values4 of the projected growth rates that are plotted in Chart 3.1. In Table 3.6, real GDP growth rates in PPP terms for the alternative scenarios are given. Chart 3.2 shows that in market exchange rate terms, the critical thresholds of US$5, 10 and 30 trillion are crossed, respectively, in FY2028, FY2036, and FY2051. In PPP 3 4

For details on methodology, see EY (2023). Estimated using HP-Filter in E-Views.

3 India’s Economy in the Twenty-First Century: Role …

59

7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 2000-01 2002-03 2004-05 2006-07 2008-09 2010-11 2012-13 2014-15 2016-17 2018-19 2020-21 2022-23 2024-25 2026-27 2028-29 2030-31 2032-33 2034-35 2036-37 2038-39 2040-41 2042-43 2044-45 2046-47 2048-49 2050-51 2052-53 2054-55 2056-57 2058-59 2060-61

3.0

S1

S2

S3

Chart 3.1 India’s trend growth rate (%) using real GDP projections under alternative simulations. Source (basic data) EY (2023)

Table 3.6 India’s projected growth rates (five yearly averages) in PPP terms: S1, S2, and S3 Period 2022-23 to 2025-26 2026-27 to 2030-31 2031-32 to 2035-36 2036-37 to 2040-41 2041-42 to 2045-46 2046-47 to 2050-51 2051-52 to 2055-56 2056-57 to 2060-61

S1 7.46 5.98 5.94 5.80 5.32 5.12 4.96 4.67

S2 7.46 6.00 6.11 6.08 5.68 5.56 5.49 5.29

S3 7.46 6.01 6.19 6.25 5.93 5.89 5.90 5.77

Source (basic data) EY (2023) Note For India, data is on fiscal year bases. 2022 implies 2022–23 and so on

terms, India has already crossed the PPP$5 trillion and PP$10 trillion thresholds in FY2011 and FY2023, respectively. It is projected to cross PPP$30 trillion by FY2041. Chart 3.3 shows that in PPP terms, India has a share which is already higher than those of Germany and Japan. It is projected to overtake the US by FY2040 and China by FY2052.5 India’s per-capita GDP is projected to cross $13,000 threshold6 in PPP terms in FY2033 and in market exchange rate terms in FY2045. By FY2048, which is the last year of the Amrit Kaal, India’s per-capita GDP in market exchange terms is expected to exceed US$15,000.

5

For cross country comparisons, OECD baseline scenario were used for countries other than India. For India, simulation S3 is being used as drawn from ongoing studies at EY. 6 A per capita income threshold of US$13,205 is considered synonymous with reaching a developed country status (World Bank, 2023); https://datahelpdesk.worldbank.org/knowledgebase/art icles/906519-world-bank-country-and-lending-groups.

60

D. K. Srivastava et al.

FY01 FY03 FY05 FY07 FY09 FY11 FY13 FY15 FY17 FY19 FY21 FY23 FY25 FY27 FY29 FY31 FY33 FY35 FY37 FY39 FY41 FY43 FY45 FY47 FY49 FY51 FY53 FY55 FY57 FY59 FY61

100 90 80 70 60 50 40 30 20 10 0

Nominal GDP (US$ trillion)

GDP (PPP international $ trillion)

Chart 3.2 Projected size of the Indian economy in PPP and MX terms—(S3). Source (basic data) EY (2023) 35.0 30.0 25.0 20.0 15.0 10.0 5.0 0.0

China

Germany

India

Japan

US

Chart 3.3 Share in global GDP: selected major countries (PPP terms)—S3. Source (basic data) Authors’ calculation

India is projected to regain7 by the early 2050s, its pre-eminent position in the global economy which prevailed in 1700 A.D. when its share in global GDP was at 24.5% (Table 3.7) (Maddison, 2007). This share had fallen to just 4.2% by 1950. India’s current share in world GDP is 6.95% (FY2022) in PPP terms.8 In contrast, India’s share in world population is 17.8% implying that its per-capita GDP in PPP$ is only 39.1% of world per-capita GDP. Thus, India has a considerable gap to cover before it can provide for its average citizen, a level of prosperity comparable to world average. By the end of the Amrit Kaal (FY2048), India’s share in world population 7

Srivastava et al. (2022). GDP based on purchasing power parity (PPP) share of world total, sourced from IMF World Economic Outlook (October 2022).

8

3 India’s Economy in the Twenty-First Century: Role …

61

Table 3.7 India’s share in world GDP (%) Year (A.D.) 1 1000 1500 1600 1700 1820 1870 1900 1913 1940 1950 1960 1970 1980 1990 2000 2008

India 32.0 27.8 24.4 22.4 24.5 16.1 12.2 8.6 7.5 5.9 4.2 3.9 3.4 3.2 4.0 5.2 6.7

2048 2053

21.8 25.7

China Japan 25.4 1.1 22.7 2.6 24.9 3.1 29.0 2.9 22.3 4.1 33.0 3.0 17.1 2.3 11.1 2.6 8.8 2.6 5.4 4.7 4.6 3.0 5.2 4.4 4.6 7.4 5.2 7.8 7.8 8.6 11.8 7.2 17.5 5.7 Simulation 3 (S3) 24.9 2.8 23.6 2.6

Germany 1.2 1.2 3.3 3.8 3.7 3.9 6.5 8.2 8.7 8.4 5.0 6.6 6.1 5.5 4.7 4.2 3.4

US 0.3 0.4 0.3 0.2 0.1 1.8 8.9 15.8 18.9 20.6 27.3 24.3 22.4 21.1 21.4 21.9 18.6

2.5 2.3

14.6 13.9

Source Maddison Project Database (2020) and EY (2023) Note The original magnitudes are given in 1990 Geary Khamis international dollars

is estimated to be 17.3% while its share in global PPP GDP is estimated to be 21.8% under S3 indicating that India’s per-capita GDP would be 1.26 times the average global per-capita GDP.

3.4 Interdependence of Saving, Investment, and the Demographic Dividend The key enabler facilitating India’s comparatively high growth is its unfolding demographic dividend. The share of its working age population (WAP) in the total population is expected to reach 68.9% by 2030. Several features distinguish India’s demographic trends in comparison to China as seen in Chart 3.4. We note that (1) India’s overall population becomes the largest, overtaking that of China by 2023; (2) Share of WAP to total population peaks at 68.9% by 2030 (104.3 crores working age persons), overtaking China; (3) India’s overall dependency ratio becomes lowest in its history by 2030 at 31.2%; (4) Excess of India’s WAP share over that of China would be at a maximum by 2056 (10.9% points) and (5) India’s old dependency ratio overtakes the young dependency ratio by 2056.

62

D. K. Srivastava et al.

75

72.9

22.5

Highest share of WAP in 2030: India 68.9

70

15.0

65

7.5

60 0.0

55

WA Pop: India minus China-% points (RHS)

WA Pop - China: %

2100

2090

2080

2070

2060

2050

2040

2030

2020

2010

2000

1990

1980

-15.0 1970

45 1960

-7.5

1950

50

WA Pop - India: %

Chart 3.4 Share of working age population in total population (%): India and China. Source UN (2022)

There is a critical interrelationship between saving and investment rates and the share of WAP. As the share of WAP increases, the share of the dependent population falls. This would lead to a higher savings rate which can then be converted to a higher investment rate and a higher growth rate by uplifting the rate of growth of the capital stock. This can be combined with the higher labor supply linked to the higher WAP provided the available persons in the working age group are adequately skilled, trained, and educated in order to become part of the productive labor force. From the viewpoint of policy intervention, another aspect of the demographic dividend is critical. The overall dependency ratio (D) can be decomposed into young dependency (D1) and old dependency (D2). The D1 curve remains higher than D2 up to 2055 (Chart 3.5). This is the period when the relatively higher emphasis should be provided to educating and skilling this young population which would potentially be joining the labor force. It is only after 2055 that the old dependency ratio becomes higher than the young dependency ratio, which would call for higher expenditures on health in the latter part of the twenty-first century. Given the importance of the demographic dividend, it is critical to recognize that there are significant differences in the state-wise profile of the demographic trends which require to be acknowledged and taken into account in policy making at least at the level of the central government.

3.5 State-Differentiated Demographic Dividend Projections of state-wise populations up to 2036 based on the population census of 2011 have been prepared by Ministry of Health and Family Welfare, Government of India. Table 3.8 shows that while over time, the population growth rate falls for all states, a good number of states by 2036 would have population growth rates that are quite low. Relatively high per-capita income states like Tamil Nadu, Andhra Pradesh,

3 India’s Economy in the Twenty-First Century: Role …

63

50

40

31.2

30 20 10 0 -10 -20 -30 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060 2065 2070 2075 2080 2085 2090 2095 2100

-40

Old age pop - Young age pop Old age (65+) - D2

Young age (25%) dum

(0.77) −0.34*

Positive output gap (real)

(0.18) −0.57***

Negative output gap (real)

(0.15) −0.01

DM25

(0.04) DL25

0.01 (0.06)

Observations

135

135

135

135

135

135

135

AR(2)

0.25

0.39

0.34

0.26

0.26

0.26

0.27

Hansen

0.71

0.72

0.83

0.71

0.63

0.72

0.72

Notes 1. Standard errors in parentheses 2. The asterisks stand for the p-value significance levels (***pp < 0.01, **p < 0.05, *p < 0.1) Source Authors’ estimates

Thirdly, we introduced a positive output gap variable in column (IV) that takes the value of the variable if the output gap is positive, and 0 otherwise. Similarly, a negative output gap variable was specified in Column (V) that takes the value of the variable if the output gap is negative, and 0 otherwise. The results indicate that States pursue a countercyclical expenditure policy in both positive and negative output gap scenarios. Additionally, we found that their countercyclicality is more pronounced during periods of negative output gap than during positive ones. This trend has been particularly noticeable in recent years, as States have increased their expenditures to respond to the COVID-19 pandemic. or effectively zero), the resulting slope coefficient of debt when the output gap is positive is also -0.04.

114

D. P. Rath et al.

Fourthly, we explore the impact of output gap on government primary expenditure across varying levels of debt (i.e., more than 25% and equal to or less than 25%) by introducing an interaction term between a dummy variable for debt more than 25% (with a value of 1 when the debt exceeds 25% and 0 otherwise) and the output gap variable. Our findings suggest that the States’ reactions to output gap do not differ significantly based on prior debt levels, whether it exceeds 25% or is less than or equal to 25%.

5.6 Concluding Observations The paper seeks to investigate the long-term association between the revenue, expenditure, and economic growth at the sub-national level in India, while also identifying the drivers of States’ primary expenditure. An empirical investigation was conducted on 18 Indian States, covering the period from 2005–2006 to 2019–2020. The analysis confirms the bi-directional relationship between government revenue and expenditure, validating the fiscal synchronization hypothesis at the subnational level in India. Additionally, there exists a relationship between government expenditure and economic growth, supporting the Wagner’s hypothesis in the context of the Indian States. The study also confirms the existence of a virtuous cycle between government revenue, expenditure, and economic growth in India. Higher government revenue leads to greater spending and a multiplier effect, catalysing economic growth. Higher economic growth leads to higher revenue, and so on. A one% increase in GSDP in the long run would lead to an increase in government’s revenue and expenditure by 1.08–1.14 and 1.19–1.20%, respectively. Similarly, an increase in government revenue by 1% would result in an increase in government expenditure and GSDP by 1.10 and 0.88–0.93%, respectively. Regarding the factors driving States’ primary expenditure, our analysis for the period 2011–2012 to 2019–2020 for 15 States suggest that the primary expenditure has a persistent effect, with past decisions on primary expenditure influencing current-year decisions. Factors such as elections and natural calamities turn out to be insignificant in influencing States’ decision on primary expenditure. Interestingly, the government employees’ pay revisions turn out to be a driver of State’s primary expenditure but in combination with other indicators they lose their significance. States pursue a countercyclical expenditure policy in positive and negative output gap scenarios, with more pronounced countercyclicality during negative output gap periods. Regarding the States’ sensitivity to their debt levels, it was found that they respond to the level of debt countercyclically in case of positive output gap. The findings of this study carry several important policy implications that warrant consideration. Firstly, policymakers should prioritize efforts to enhance government revenue collection to stimulate economic growth. Specifically, higher revenue collections may induce greater government spending, generating a multiplier effect on the economy. Secondly, it is desirable for States to increase their capital expenditure during positive output gaps while reducing unproductive revenue expenditure. This

5 The Relationship Between Government Revenue, Government …

115

strategic reallocation of resources may result in improved quality of expenditure and enhanced medium-term growth prospects while keeping the fiscal deficit under control. Lastly, it is essential for States with higher debt levels to factor in debt reduction considerations while making decisions about primary expenditure, in the interest of long-term fiscal sustainability and financial stability considerations.

Statistical Annex Pedroni’s Co-integration Test The Pedroni’s panel co-integration test which is based on the Engle Granger procedure is an improvement over the conventional co-integration tests since it allows for heterogeneity among individual members of the panel. This co-integration test evaluates the regression residual with I(1) variables in order to detect the presence of unit root. If the residuals of the regression turns out to be I(0), it implies that the variables under consideration are co-integrated. In its generalized form, Pedroni’s panel co-integration regression could be mathematically expressed as: yit = σi + δi t + βi xi,t +

i,t

(5.1)

where t = 1,….,T; and i = 1,….,N. In Eq. (5.1), y and x are presumed to be I(1) and σi andδi represents the individual and trend effects. In this test, the null hypothesis presupposes that the residual is I(1). Following this, the residual is tested for the presence of unit root, and if the null hypothesis is rejected, co-integration between the variables is assumed to exist. This co-integration test in fact provides eleven statistics which has varying degrees of properties (size and power for different N and T). Kao’s Cointegration Test For ensuring the robustness of existence of co-integration detected by the Pedroni’s test, Kao’s co-integration test is also undertaken. This test is also based on Engle Granger procedure and is also a two-stage procedure for detecting co-integration similar to Pedroni’s co-integration test (Kao’s regression equation has an algebraic expression that is comparable to Pedroni’s). However, this test differs from the Pedroni’s since it specifies cross-section specific intercepts and homogenous coefficients on the first stage regressors. Thus, in the first stage, Kao test assumes homogenous coefficient and different intercept of regression equation for every cross section, but in the second stage, it examines the stationary test of residual error series of regression equation at first stage. As the residual is tested for unit root and the null hypothesis is rejected, it implies that there exist co-integration amongst the variables.

116

D. P. Rath et al.

Panel Fully Modified Ordinary Least Squares Fully modified OLS (FMOLS) for elasticity estimation in panel co-integration analysis has been recommended by Pedroni (Pedroni, 1996). While estimating dynamic co-integrated panels, heterogeneity is a major issue encountered on account of the differences in means among the individuals as well as differences in individuals’ responses to short-run disturbances from co-integrating equilibrium. This is overcome by FMOLS estimator by incorporating the individual specific intercepts into the regression and by allowing serial correlation properties of the error processes to vary across individual members of the panel. Following the Mitic et al. (2017) approach, the standard form of the OLS panel estimator is given in Eq. (5.2): N T N T ∑t=1 β N T = ∑i=1 ∑t=1 xi,t − x i yi,t − y i (xi,t − x i )−1 ∑i=1

The covariance matrix represented by

i

(5.2)

of the vector error term is given by:

11i

12i

21i

22i

where 11i is the long-run variance of error term εi,t , 22i is the long-run covariance matrix of i,t and 21i = 12i shows the long- run covariance between independent variable and its residual vector. The modified OLS estimator provides FMOLS estimator as Eq. (5.3): −1

N T β F M O L S = ∑i=1 xi,t − x i L 22i ∑t=1

2 −1

−1

−1

N T ∗ xi,t − x i yi,t ∑i=1 L 11i L 22i (∑t=1 − T δi )

(5.3) where, L 11i =

11i



21i

∗ yi,t = yi,t − y i −

and δ i =

21i

+

0 21i



−1 21i

21i

L 21i L 22i

L 21i ( L 22i

22i

, L 12i = 0, L 21i = xi,t +

+

L 21i − L 22i L 22i

21i

−1/2 21i ,

L 22i =

1/2 21i

β xi,t − x i

0 22i )

Panel Dynamic Ordinary Least Squares Panel dynamic ordinary least squares (DOLS) proposed by Kao and Chiang (2000) is an extension of the time series DOLS put forth by Stock and Watson (1993). DOLS offers certain built-in advantages over OLS and FMOLS estimation. Firstly, the issue of asymptotic bias prevalent in OLS estimation is well addressed in DOLS estimation by including lags and leads of the difference series of variables. Secondly, the adoption of DOLS is useful in coping with the problem of serial correlation irrespective

5 The Relationship Between Government Revenue, Government …

117

of the order of integration and the existence or absence of co-integration. Thirdly, DOLS is more computationally convenient than OLS or FMOLS. Fourthly, the usage of DOLS is justified even when the dependent variable is endogenous since DOLS estimator are asymptotically unbiased and normally distributed. Fifthly, the DOLS estimator accounts for the heteroscedasticity between the groups by computing the mean group estimator. Finally, the ‘t’ statistic obtained by the DOLS tends to follow the standard normal distribution as compared to the ‘t’ statistic computed using OLS or FMOLS. The Dynamic OLS estimator can be computed using the following regression Eq. (5.4): q

yi,t = βi xi,t + ∑ j=−q ςi j xi,t+ j + γli Dli + εi,t

(5.4)

where q denotes the number of lags or leads required. Two-Way System Generalized Method of Moments (GMM) The two-way system generalized method of moments (GMM) is an econometric technique used to estimate a system of equations where endogenous variables are jointly determined. This method involves the use of instrumental variables (IVs) to account for the endogeneity of explanatory variables. The GMM estimator uses a set of moment conditions based on the orthogonality conditions between the errors and IV. It is a preferred choice over other econometric techniques due to its ability to handle various data structures, address endogeneity issues, and provide consistent and efficient estimates. However, it can be computationally intensive, and the validity of the IV used in the GMM estimator must be carefully evaluated to ensure they meet the necessary assumptions. Moreover, the two-way system GMM requires the correct specification of moment conditions and weighting matrix, which can be challenging in certain cases.

References Afonso, A., & Jalles, J. T. (2014). Causality for the government budget and economic growth. Working Papers. Department of Economics. Lisbon School of Economics and Management. Al-Faris, A. F. (2002). Public expenditure and economic growth in the gulf cooperation council countries. Applied Economics, 34(9), 1187–1193. Alkasasbeh, O. M. A., & Haron F. N. (2018). Fiscal policy and its relationship with economic growth: A review study. Saudi Journal of Business and Management Studies. Scholars Middle Publishers, Dubai, United Arab Emirates. ISSN: 2415-6671. Aschauer, D. A. (1990). Is government spending stimulative? Contemporary Economic Policy, 8, 30–46. Atems, B. (2015). Another look at tax policy and state economic growth: The long-run and short-run of it. Economics Letters. Athanasios, A., Constantinos, K., & Emmanouil, T. (2014). Government spending and revenues in the Greek economy: Evidence from nonlinear cointegration. Empirica. Springer. Austrian Institute for Economic Research. 41(2), 365–376.

118

D. P. Rath et al.

Bacchiocchi, E., Borghi, E., & Missale, A. (2011). Public investment under fiscal constraints. Fiscal Studies, 32(1), 11–42. Baffes, J., & Shah, A. (1994). Causality and co-movement between taxes and expenditures: Historical evidence from Argentina, Brazil and Mexico. Journal of Development Economics, 44, 311–331. Bergh, A., & Henrekson, M. (2011). Government size and growth: A survey and interpretation of the evidence. Research Institute of Industrial Economics. Working Paper Series. No. 858. Bhat, K. S., Nirmala, V., & Kamiah, B. (1993). Causality between tax revenue and expenditure of Indian states. Indian Economic Journal, 40, 108–109. Bishnoi, N. K., & Juneja, T. (2016). An examination of interdependence between revenue and expenditure of government of Haryana. The Indian Economic Journal, 64(1–4), 176–185. Blanchard, O., & Perotti, R. (2002). An empirical characterization of the dynamic effects of changes in government spending and taxes on output. The Quarterly Journal of Economics. Burton, A. (1999). The effect of government size on the unemployment rate. Public Choice, 99, 3–4. Carboni, O., & Medda, G. (2010). A neoclassical growth model with public spending. Centre for North South Economic Research. University of Cagliari and Sassari. Working Paper CRE. No. S201033. Chalil, M. (2018). Political cycles, government spending, and efficiency of Indonesia’ Local Governments. MPRA Paper 88082. University Library of Munich, Germany. Chang, T., & Ho, Y. H. (2002). A note on testing tax-and-spend, spend-and-tax or fiscal synchronization: The case of China. Journal of Economic Development, 27, 151–160. Choi, I. (2001). Unit root tests for panel data. Journal of International Money and Finance, 20(2), 249–272. Colombier, C. (2009). Growth effects of fiscal policies: An application of robust modified Mestimator. Applied Economics, 41, 899–912. Dahlberg, M., & Johansson, E. (2000). An examination of the dynamic behaviour of local governments using GMM bootstrapping methods. Journal of Applied Econometrics, John Wiley & Sons Limited, 15(4), 401–416. Dasalegn, M. J. (2014). Taxation for economic growth: Evidence from Ethiopia. International Journal of Economics and Empirical Research. The Economics and Social Development Organization, 2(7), 294–300. Dash, B. B., & Raja, A. V. (2012). Political determinants of the allocation of public expenditures: A study of the Indian states. Working Papers 12/101, National Institute of Public Finance and Policy. Dhanasekaran, K. (2001). Government tax revenue, expenditure and causality: The experience of India. Indian Economic Review, 36(2). Dladla, K., & Khobai, H. (2018). The impact of taxation on economic growth in South Africa. MPRA Paper 86219. University Library of Munich. Germany. Ebeke, C., & Olçer, D. (2013). Fiscal Policy over the election cycle in low-income countries. IMF Working Paper, No. 13/153, June. Edame, G. E., & Okoi, W. W. (2014). The impact of taxation on investment and economic development in Nigeria. Academic Journal of Interdisciplinary Studies, 3(4), 209–218. Eita, J. H., & Mbazima, D. (2008). The causal relationship between government revenue and expenditure in Namibia. MPRA Paper, 9154. University Library of Munich, Germany. Engen, M. E., & Skinner, J. (1992). Fiscal policy and economic growth. National Bureau of Economic Research. Working Paper No. 4223. Federici, D., Parisi, V., & Elliott, C. (2015). Do corporate taxes reduce investments? Evidence from Italian Firm-level Panel Data. Cogent Economics & Finance. Folster, S., & Henrekson, M. (2001). Growth effects of government expenditure and taxation in rich countries, European Economic Review, 45(8), 1501–1520.

5 The Relationship Between Government Revenue, Government …

119

Funashima, Y., & Hiraga, K. (2017). Wagner’s Law, fiscal discipline and inter-governmental transfer: empirical evidence at the US and German state levels. International Tax and Public Finance, 24, 652. Gaspar, V., & Gopinath, G. (2020). Fiscal policies for a transformed world. IMF Blog. Gemmell, N., & Au, J. (2012). Government size, fiscal policy and the level and growth of output: A review of recent evidence. Journal of the Asia Pacific Economy. Gurdal, T., Aydin, M., & Inal, V. (2021). The relationship between tax revenue, government expenditure and economic growth in G7 countries: A new evidence from time and frequency domain approaches. Economic Change and Restructuring. Springer Science and Business Media. Helms, J. L. (1985). The effect of state and local taxes on economic growth: A time series—cross section approach. The Review of Economics and Statistics, 67(4). MIT Press. Hondroyiannis, G., & Papapetrou, E. (1995). An explanation of Wagner’s law for Greece: A cointegration analysis. Public Finance., 50, 67–79. Horton, M., & Ganainy A. E. (2020). Fiscal policy: Taking and giving away. Finance and Development Magazine, International Monetary Fund. Im, K. S., Pesaran, M., & Shin, Y. (2003). Testing for unit roots in heterogeneous panels. Journal of Econometrics, 115(1), 53–74. International Monetary Fund. (2015). Fiscal policy and long-term growth. Policy Paper. Kelly, T. (1997). Public expenditures and growth. Journal of Development Studies. Taylor & Francis Journals, 34(1), 60–84. Khairul, A., Nazimuddin, Raja, M., & Hasdi, A. (2019). Is there a causality relationship between local tax revenue and regional economic growth? A panel data evidence from Indonesia. Regional Science Inquiry, 11(1), 73–84. Hellenic Association of Regional Scientists, June. Khundrakpam, J. K. (2003). Public Sector Spending and Economic Growth in India. RBI Occasional Papers, 22(1–3), 1–17. Kollias, C., & Paleologou, S. M. (2013). Guns, highways and economic growth in the United States. Economic Modelling, 30(C), 449–455. Elsevier. Konukcu-Onal, D., & Tosun, A. N. (2008). Government revenue-expenditure nexus: Evidence from several transitional economies. Economic Annals, 53, 145–156. Korinek, A., & Stiglitz, J. E. (2022). Revisiting the fiscal-monetary policy mix and correcting macro economic externalities. Hutchins Centre Working Paper. No.78. August. Lee, Y., & Gordon, R.H. (2005). Tax structure and economic growth. Journal of Public Economics, 89(5–6), 1027–1043. Levin, A., Lin, C. F., & Chu, C. S. J. (2002). Unit root tests in panel data: Asymptotic and finitesample properties. Journal of Econometrics, 108, 1–24. Li, X. (2001). Government revenue, government expenditure, and temporal causality: Evidence from China. Applied Economics, 33, 485–497. Manage, N., & Marlow, M. L. (1986). The Causal relation between federal expenditures and receipts. Southern Economic Journal, 52, 617–629. Magazzino, C. (2014). The relationship between revenue and expenditure in the ASEAN Countries. East Asia. Mdanat, M. F., Shotar, M., Samawi, G., Mulot, J., Arabiyat, T. S., & Alzyadat, M. A. (2018). Tax structure and economic growth in Jordan. 1980–2015. EuroMed Journal of Business, 13(1), 102–127. Misra, S., Behera, R. S., Seth, B., & Sood, S. (2021). fiscal framework and quality of expenditure in India. Reserve Bank of India Bulletin, June, Reserve Bank of India. Mitic, P., Ivanovic, O., & Zdravkovic, A. (2017). A cointegration analysis of real GDP and CO2 emissions in transitional countries. Sustainability. April. Mofidi, A., & Stone, A. J. (1990). Do state and local taxes affect economic growth? The Review of Economics and Statistics, 72(4), 686–691. Mohanty, A. R., & Mishra, B. R. (2017). Cointegration between revenue and expenditure: Evidence from India. Advances in Economics and Business, 5(1), 33–40.

120

D. P. Rath et al.

Naidu, C. R., Mohsin, M., & Naidu, V. J. (1995). Government expenditure in Andhra Pradesh: An analysis of growth and determinants. Prajnan, 23(3), 332–352. Narayan, K. P. (2005). The government revenue and government expenditure nexus: Empirical evidence from nine Asian countries. Journal of Asian Economics, 1203–1216. Elsevier. Narayan, S., Rath, N. B., & Narayan, K. P. (2012). Evidence of Wagner’s law from Indian states. Economic Modelling (pp. 1548–1557). Elsevier. Nayak, K. D., & Hazarika, B. (2022). Linkage between income and government expenditure at Indian Sub-nationals: A second-generation panel cointegration techniques. National Institute of Public Finance and Policy Working Paper. No. 374. Neog, Y., & Gaur, K. A. (2020). Tax structure and economic growth: A study of selected Indian states. Journal of Economic Structures. Nirola, N., & Sahu, S. (2020). Revisiting the Wagner’s law for Indian states using second generation panel cointegration. Economic Change and Restructuring, 53, 241–263. Nyamongo, M. E., Siehei, M. M., & Schoeman, N. J. (2007). Government revenue and expenditure nexus in South Africa. South African Journal of Economic and Management Sciences, 10, 256–268. Payne, J. (1997). The tax-spend debate: The case of Canada. Applied Economics Letters. Taylor & Francis Journals, 4(6), 381–386. Pedroni, P. (1996). Fully modified OLS for heterogeneous cointegrated panels and the case of purchasing power parity. Indiana University. Pereira, A., & Sagale, R. O. (2011). Long-term effects of fiscal policies in Portugal. Journal of Economic Studies, 38, 114–127. O¸legs, T., & Karlis, V. (2019). The impact of government borrowing costs on fiscal discipline, Kyklos, 72(3), 446–471. Wiley Blackwell, August. Rahn, R., & Fox. H. (1996). What is the optimum size of government. Vernon K. Krieble Foundation. Reserve Bank of India. (2021). State finances: A study of budget of 2021–22. November Reserve Bank of India. (2023). State finances: A study of budgets of 2022–23. January. Richter, C., & Paparas, D. (2013). The validity of Wagner’s Law in Greece during the Last 2 centuries. Applied Economics Quarterly. Ro¸soiu, I. (2015). The impact of the government revenues and expenditures on the economic growth. Procedia Economics and Finance., 32, 526–533. Sagdic, N. E., Sasmaz, U. M., & Tuncer, G. (2020). Wagner versus Keynes: Empirical evidence from Turkey’s Provinces. Panoeconomicus, 67, 657–674. Scully, G. W. (2003). Optimal taxation, economic growth and income inequality. Public Choice, 115, 299–312. Shand, R., & Bhide, S. (2000). Sources of economic growth: Regional dimensions of reforms. Economic and Political Weekly, 3747–3757, October. Srithongrung, A., & Juarez, S. I. (2015). Fiscal policies and subnational economic growth in Mexico. International Journal of Economics and Financial Issues, 5(1), 11–22. Srithongrung, A., & Kriz, A. K. (2014). The impact of subnational fiscal policies on economic growth: A dynamic analysis approach. Journal of Policy Analysis and Management, 33(4), 912–928. Wiley. Stoilova, D. (2017). Tax structure and economic growth: Evidence from the European Union. Contaduria y Administracion, 62(3), 1041–1057. Stokey, N., & Rebelo, S. (1995). Growth effects of flat-rate taxes. Journal of Political Economy, 103(3), 519–550. Szarowska, I. (2011). Changes in taxation and their impact on economic growth in the European Union. Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis, 59, 325–332. Tomljanovich, M. (2008). The role of state fiscal policy in state economic growth. Contemporary Economic Policy. July. Trivedi, P., & Rajmal. (2011). Growth effects of fiscal policy of Indian States. Millennial Asia, 2(2), 141–162.

5 The Relationship Between Government Revenue, Government …

121

Ugwunta, D., & Ugwuanyi, U. (2015). Effect of distortionary and non-distortionary taxes on economic growth: Evidence from sub-Saharan African Countries. Journal of Accounting and Taxation. Wang, T., Gao, K., Wen, C., Xiao, Y., & Bingzheng, Y. (2022). Assessing the nexus between fiscal policy, COVID-19 and economic growth. Environmental Science and Pollution Research International.

Chapter 6

Revenue Implications of GST on Indian State Finances Sacchidananda Mukherjee

Abstract Assessing the revenue implications of GST on Indian state finances cannot be contained to compare the revenue stream which is subsumed into GST with State GST collection alone. Since GST subsumes many taxes from state tax bases, comparing the revenue performance of taxes which are outside the GST framework would be equally important. Moreover, in the federal system revenue implication of shortfall in tax collection of the federal government is also likely to spill-over to subnational finances in terms of lower tax devolution. Sustaining revenue streams of state governments is important for sustainable Public Finance Management (PFM) and therefore a comprehensive assessment of state finances before and after GST would be important. This paper attempts to fill the gap in exiting literature by assessing the revenue of 18 major states during pre- and post-GST periods. Keywords Revenue assessment · Goods and services tax (GST) · State finances · Revenue protection · GST compensation · India JEL Codes H20 · E62 · H26

6.1 Introduction Indian GST completes five years on 30 June 2022. During last five years, many changes are made in the rate structure as well as rules and regulations of GST (Mehta & Mukherjee, 2021). Revenue implications of those changes cannot be This chapter is a re-publication of the author’s working paper and is being re-used here with permission Mukherjee, S. (2023). Revenue Implications of GST on Indian State Finances, NIPFP Working Paper Series, Working Paper No. 388. National Institute of Public Finance and Policy, New Delhi: India. S. Mukherjee (B) National Institute of Public Finance and Policy (NIPFP), 18/2, Satsang Vihar Marg, Special Institutional Area, New Delhi 110 067, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 D. K. Srivastava and K. R. Shanmugam (eds.), India’s Contemporary Macroeconomic Themes, India Studies in Business and Economics, https://doi.org/10.1007/978-981-99-5728-6_6

123

124

S. Mukherjee

ignored and therefore assessing the impact of GST on Indian state finances will be important. For example, in a recent paper Mukherjee (2023) estimates that effective tax rate of GST was 12.88% in 2017–2018, and it falls in three consecutive years (12.71% in 2018–2019, 11.20% in 2019–2020 and 10.91% in 2020–2021), thereafter it rises to 12.21% in 2021–2022 and 12.56% in 2022–2023 (upto Q3 of 2022–2023). Moreover like World economy, Indian economy has also gone through a major economic shock due to the COVID-19 pandemic. Being a broad measure of tax base, slowing down of economic growth during 2019–2021 and increasing demands for public expenditures during the pandemic have impacts on fiscal situation of the economy (Mukherjee, 2022). Sustaining the revenue stream which is subsumed into GST is important for sustainable Public Finance Management (PFM) for states. Since GST compensation period ends on 30 June 2022, it will be important for states to assess the revenue performance of both GST as well as other indirect taxes which are outside the GST framework. The present study attempts to fill the gap in exiting literature by assessing the revenue of 18 major states during pre- and post-GST periods. So far actual (or audited) statement of accounts of state finances is available up to 2020–2021 either from state Finance Accounts or Budget Documents. We have used both the data sources in this paper and present our analysis. We present the trends of economic growth of Indian economy as well as aggregate growth rate of 18 major states during 2006–2007 to 2021–2022 in the next section. Since, revenue mobilization is dependent on economic growth, in Sect. 6.3 we present aggregate fiscal health of 18 major states during 2005–2006 to 2022– 23BE. In Sect. 6.4 we present revenue performance of GST and this is followed by discussion on revenue implications of GST on Indian state finances in Sect. 6.5. We draw our conclusions in Sect. 6.6.

6.2 State of the Economy We observe a falling trend in the annual economic growth rate since 2006–2007. It is to be highlighted that like World economy, Indian economy has also faced two major shocks during 2006–2007 to 2021–2022—Global Financial Crisis (GFC, 2008–2009 to 2009–2010) and COVID-19 Pandemic (2020–2021). The impact of COVID-19 pandemic on Indian economic growth was much larger than the impact of GFC (Fig. 6.1). Being a broad measure of tax base of the economy, slowing down of economic growth is expected to have impact on the fiscal health of governments.

6.3 Fiscal Health of Indian States Like economic growth, Indian states also faced fiscal stresses during 2008–2010, 2015–2017 and 2020–2021 onwards (Fig. 6.2). In addition to revenue stress due to GFC, implementations of 6th Pay Commission recommendations by many state

6 Revenue Implications of GST on Indian State Finances

125

22 18 14 10 6 2 -2

All India

18 Major States

Fig. 6.1 Annual growth rate of GDP/GSDP at market prices (at current prices, 2011–2012 series) (%). Source Computed by author based on EPWRF India time series database and budget documents of state governments

governments increased revenue expenditures and resulted in rises of revenue as well as fiscal deficits during 2008–2010 (Mukherjee, 2019). States faced a relatively stronger fiscal stress during 2015–2016 to 2016–2017. To improve the financial health and operational efficiency of debt-ridden power distribution companies, the Union government introduced the Ujwal DISCOM Assurance Yojana (UDAY) scheme in November 2015 to provide debt relief to public Power Distribution Companies (DISCOM). The basic objective of the scheme was to clean up the balance sheet of the DISCOM by taking over 75% of outstanding debt (as on 30 September 2015) by the participating State government and free the credit blocked by creditors (mostly Public Sector Banks). Participating states took over 75% of outstanding debt of public DISCOM in two tranches—50% in 2015–2016 and 25% in 2016–2017. This resulted in fiscal stress for states without any impact on revenue deficit (Mukherjee, 2019). States again faced fiscal shock during 2020–2021 due to the COVID-19 pandemic. The economic contraction during the pandemic created pressures on Public Finance Management (PFM) in terms of lower revenue mobilization and higher public expenditures. Both the Union and state governments faced dual problems of arresting economic contraction and managing public finance with limited public resources available for disposal (Mukherjee & Badola, 2022).

126

S. Mukherjee

5 4 3

2008-09

2009-10

2015-16

2016-17

2020-21

2 1

0 -1 -2

Revenue Deficit

Fiscal Deficit

Fig. 6.2 Aggregate fiscal health of 18 major states (% of GSDP) during 2005–2006 to 2022–2023 (BE). Source Computed by author based on finance accounts and budget documents of states

6.4 Revenue Performance of GST India introduced Goods and Services Tax (GST) on 1 July 2017 by subsuming taxes from the Union and state tax bases. GST is a comprehensive multi-stage Value Added Tax (VAT) encompassing both goods and services with concurrent taxation power of the Union and state governments. During Q3 of 2017–2018 to Q4 of 2021–2022, GST collection was hovering between 6 and 6.5% of nominal GDP, except during Q1 and Q2 of 2020–2021 (on account of COVID-19 Pandemic). During Q1 and Q2 of 2022–2023, GST collection has crossed 6.5% of GDP. Tax Buoyancy in GST (i.e., Growth Rate of GST Collection/Growth Rate of GDP) was lower during 2019–2020, otherwise it lies above 1 (Fig. 6.3). Before COVID-19 pandemic growth rate in quarterly (year-on-year) GST collection shows a falling trend and the growth rate was hovering between 2 and 16% during Q3 of 2018–2019 to Q4 of 2019–2020. During Q2 of 2020–2021 to Q1 of 2021– 2022, the growth rate improves and it is mostly attributed to lower base effect due to COVID-19 pandemic. In Q2 of 2021–2022, the growth rate again falls and thereafter it is hovering between 16 and 34%. The fall in growth rate in GST collection in Q2 of 2021–2022 could be due to the second wave of the COVID-19 pandemic and associated economic restrictions. It is also to be highlighted that the pre-pandemic quarterly growth rate of GST from imports (IGST and GST Compensation cess from imports) was lower than growth rate of GST from domestic components. However, growth rate in GST from imports surpasses the growth rate in GST from domestic components since Q3 of 2020–2021. On average GST from imports contributes onefourth share in total GST collection in India. The higher growth rate in GST collection

6 Revenue Implications of GST on Indian State Finances

127

8

4 3.5

7

3 2.5

6.5

2

6

1.5

5.5

Tax Buoyancy

GST Collection (as % of GDP)

7.5

1

5

0.5

4.5

0

4

Tax Buoyancy (RHS)

GST (as % of GDP)

Fig. 6.3 Quarterly GST collection and tax buoyancy. Source Computed by author based on monthly press releases of the Department of Revenue, Government of India

from imports may be attributed to the post pandemic rise in global prices of goods and services. Improvement in GST collection during 2021–2022 and 2022–2023 is largely driven by rise in prices of goods and services and improvements in the GST compliance (Fig. 6.4). 90 70 50 30 10 -10 -30 -50

GST-Total

GST-Domestic

GST-Import

Fig. 6.4 Quarterly (year-on-year) growth rate in GST collection. Source Computed by author based on monthly press releases of the Department of Revenue, Government of India

128

S. Mukherjee

Table 6.1 GST revenue performance of 18 major states in India Description

2018–2019

2019–2020

2020–2021

Revenue under protection (Rs. crore) (A)

530,545

604,821

689,496

SGST collection (including IGST settlement) (Rs. crore) (B)

468,722

457,743

420,025

B as % of A

88.3

75.7

60.9

SGST collection (including IGST settlement and GST compensation) (Rs. crore) (C)

526,993

562,158

542,887

C as % of A

99.3

92.9

78.7

SGST collection (including IGST settlement and GST compensation from all sources*) (Rs. crore) (D)

631,390

D as % of A

91.6

Source Computed by author based on data obtained from state finance accounts/budget documents

Analysis of GST collections of 18 major states is presented in Table 6.1. This shows that on average states collected 88.3% of Revenue Under Protection (or aggregate projected revenue of states in GST) in 2018–2019.1 However, there exists variation across states ranging from 63% in Punjab to 103% in Andhra Pradesh in the achievement (Appendix Table A1). In 2019–2020, on average SGST collection (including IGST settlement) could meet only 76% of the RUP of 2019–2020 and it further falls to 61% in 2020–2021. As compared to 2018–2019, in 2019–2020 the largest slippage in GST collection with reference to Revenue Under Protection is observed for Rajasthan, followed by Andhra Pradesh. As compared to 2019–2020, in 2020–2021 GST collection further falls for all states and the larger falls are observed for Maharashtra and Madhya Pradesh. GST compensation helped states to protect revenue and with GST compensation on average states achieved 99.3% of RUP in 2018–2019. In 2019–2020, with GST compensation states achieved on average 93% of RUP. Due to shortfall in GST compensation cess collection in 2020–2021, states could achieve 78.7% of RUP with GST compensation from GST compensation fund. However, in 2020–2021 states also receive GST compensation in terms of back-toback loans from the Union government in lieu of shortfall in GST compensation cess collection and this helped states to achieve on average 91.6% of RUP in 2020–2021. This analysis shows that state GST collection (as percentage of RUP) is falling over the years during 2018–2021 and dependence on GST compensation has grown up.

1

During GST transition period (1 July 2017 to 30 June 2022) states receive GST compensation based on the difference between projected state GST revenue and actual state GST collection (including IGST settlement). The projection of state GST revenue is based on the revenue that states collected in 2015–2016 (base year) from taxes subsumed into GST (also known as revenue under protection) and 14% annual growth (Year-on-Year) in revenue under protection from the base year 2015–2016.

6 Revenue Implications of GST on Indian State Finances

129

6.5 Revenue Implications of GST on Indian State Finances In this section we assess the revenue impact of GST on Indian state finances. We adopt different measures to assess the revenue stream associated with GST and compare the revenue performance of major states over the periods (pre-GST versus post-GST). We present state-wise average annual growth rate in GSDP (at market prices, current prices, 2011–2012 series) during pre-GST (2012–2013 to 2016–2017) and post-GST (2017–2018 to 2021–22RE) periods in Fig. 6.5. We see that except in West Bengal, all other major states show fall in the average annual growth rate of GSDP during the post-GST period (2017–2018 to 2021–22RE) as compared to the pre-GST period (2012–2013 to 2016–2017). It is expected that fall in the growth rate in GSDP may have some impact on GST collection for states. Slowdown in economic growth in 2019–2020 and COVID-19 pandemic in 2020–2021 resulted in revenue as well as fiscal stresses for states (Mukherjee & Badola, 2022). In this paper we assess the revenue implications of GST for States only. In this exercise we carry out the following comparisons: (a) Comparison of Revenue Under Protection (RUP) for Pre-GST period with State GST collection (including IGST Settlement) for Post-GST period (b) Comparison of State Revenue Basket (either partially or fully subsumed into GST) for Pre-GST period with Post-GST period (as % of GSDP) (c) Comparison of State’s Share in the Union Taxes (either partially or fully subsumed into GST) for Pre-GST with Post-GST period (as % of GSDP) (d) Comparison of State’s Own Tax Revenue (OTR) for Pre-GST period with PostGST period (as % of GSDP) 20 15 10 5 0 -5

Pre-GST (2012-13 to 2016-17)

Post-GST (2017-18 to 2021-22RE)

Post-GST - Pre-GST

Fig. 6.5 Average annual growth rate of GSDP (at current prices, market prices, 2011–2012 series) (in %). Source Computed by author based on data obtained from EPWRRF India time series database and budget documents of state governments

130

S. Mukherjee

(e) Comparison of State Tax Revenue (OTR & State’s Share in the Union Taxes) for Pre-GST period with Post-GST period (as % of GSDP) It is to be mentioned that by Revenue Under Protection (RUP) we mean the revenue from taxes (from state tax base) which are subsumed into GST. Being the transition year we avoid taking into account GST collection in 2017–2018 in our empirical analysis. Since both the Union and state governments settled transitional credits of pre-GST taxes with GST liability, GST collection is not likely to reflect the actual GST potential of states in 2017–2018. We estimate the growth rate in GST collection in 2019–2020 with reference to GST collection in 2018–2019 and so on. Since data on RUP for Gujarat and Haryana is not available for pre-GST period, except for 2015–2016, it has restricted our analysis.

6.5.1 Comparison of Revenue Under Protection with State GST We first compare average annual growth rate of RUP for the Pre-GST period with State GST (including IGST settlement) for the post-GST period. In addition to State GST (including IGST settlement), states receive GST compensation. To make comparable series of revenue for pre- and post-GST, we add GST compensation receipts of states with SGST revenue. We construct three alternative streams of revenue for the post-GST period by adding GST compensation receipts from the GST compensation fund and back-to-back loans that states receive in lieu of shortfall in GST compensation cess collection vis-à-vis the amount required for full GST compensation payments to states. The four series of data that we have created are presented below: . Pre-GST: Average Annual Growth Rate in Revenue Under Protection during 2014–2015 to 2016–2017 . Post-GST1: Average Annual Growth Rate in State GST Collection (including IGST Settlement) during 2019–2020 to 2020–2021 . Post-GST2: Average Annual Growth Rate in State GST Collection (including IGST Settlement and GST Compensation) during 2019–2020 to 2020–2021 . Post-GST3: Average Annual Growth Rate in State GST Collection (including IGST Settlement, GST Compensation, and Back-to-Back Loans) during 2019– 2020 to 2020–2021 Figure 6.6a shows that in Goa the pre-GST average annual growth rate in revenue from subsumed taxes in GST (hereafter RUP) was 9.9% and average annual growth rate in SGST (including IGST settlement) was −11.1% during 2019–2020 to 2020– 2021 (hereafter Post-GST1). Average annual growth rate in SGST with GST compensation (from GST compensation fund) was 10.5% during 2019–2021 (hereafter PostGST2). This implies that GST compensation payments helped Goa to achieve the average annual growth rate in SGST which is higher than what was prevailing during

6 Revenue Implications of GST on Indian State Finances

131

pre-GST period. States receive back-to-back loans from the Union government in 2020–2021 and 2021–2022 in lieu of shortfall in GST compensation cess collection in addition to GST compensation from the GST compensation fund. Since the obligations of servicing the debt (both interest and principal payments) of these loans rest on the Union government and for this GST Compensation cess collection has been extended till 31 March 2026, we take into account GST compensation (from all sources) as revenue receipts of states and estimate the growth rate in SGST collection. We find that with all GST compensation, Goa achieved an average annual growth rate of 23.4% during 2019–2021 (hereafter Post-GST3). Except Bihar, other three states (viz., Andhra Pradesh, Chhattisgarh, Goa) achieved growth rate in the PostGST3 higher than the growth rate that was prevailing prior to introduction of GST (Fig. 6.6a). It is to be highlighted that Telangana is created from undivided Andhra Pradesh in June 2014 and therefore pre-GST growth rate of Andhra Pradesh is lower. Figure 6.6b shows that all four states attain a lower average annual growth rate in GST collection during Post-GST1 than the growth rate prevailing during the pre-GST period. Even with GST compensation from all sources (under Post-GST3), except Kerala none of the other three states could attain an average annual growth rate that was prevailing during the pre-GST period. Figure 6.6c shows that except Odisha none of the other three states could achieve average annual growth rate in the Post-GST3 higher than the average annual growth

9.90 Goa

-11.09 10.51 23.40

7.44 -4.26

Chhattisgarh

0.96 15.12 15.76 -2.01

Bihar

2.85 12.95

-12.24 -5.77

Andhra Pradesh

2.66 7.89 -14

-10

-6 Pre-GST

-2

2

Post-GST1

6 Post-GST2

10

14

18

22

26

Post-GST3

Fig. 6.6a Average annual growth rate of RUP and state GST* (%). Note *—including IGST settlement. Source Computed by author based on Data obtained from state finance accounts/budget documents

132

S. Mukherjee 10.18 -6.04

Madhya Pradesh

0.36 9.45 8.67 -3.23

Kerala

5.00 16.08 11.00 -5.87

Karnataka

-1.52 9.43 13.32 -1.57

Jharkhand

3.59 12.08 -10

-6

-2

Pre-GST

Post-GST1

2

6 Post-GST2

10

14

18

Post-GST3

Fig. 6.6b Average annual growth rate of RUP and state GST* (%). Note *—including IGST settlement. Source Computed by author based on Data obtained from state finance accounts/budget documents

rate in taxes subsumed into GST prior to the introduction of GST (Pre-GST). Among major states, Odisha is the only state which has achieved a positive average annual growth rate in GST collection without any GST compensation in the Post-GST1. GST compensation in terms of back-to-back loans helped states to moderate the impact of shortfall in GST collection in 2020–2021. Except a few states (viz., Chhattisgarh, Goa, Kerala, Odisha, Tamil Nadu), with GST compensations (from all sources) the annual average growth rate in GST collection falls short of 14%. Barring Telangana, all other states in Fig. 6.6d achieved the average annual growth rate in the Post-GST3 higher than the average annual growth rate prevalent in the PreGST period. None of the states achieved the average annual growth rate in the postGST1 that was prevailing in the pre-GST period. This shows that GST compensation was necessary for states for sustaining the revenue stream which has subsumed into GST. Telangana is created from undivided Andhra Pradesh on June 2014 and we have taken the growth rate for 2016–2017 only for Telangana for the pre-GST period. Except Bihar, the average annual Pre-GST growth rate in subsumed taxes was lower than 14% for other states. Many states could not attain 14% average annual growth rate in GST collection during 2019–2020 to 2020–2021 even with GST Compensation. Though, GST compensation helped states to moderate the revenue impact of shortfall in GST collection, it could not wipe out the entire shortfall in GST collection for some states.

6 Revenue Implications of GST on Indian State Finances

133 10.88

-6.54

Rajasthan

0.87 9.59 5.47 -6.46

Punjab

2.12 4.21

8.22 1.63

Odisha

4.24 15.39 6.07

Maharashtra -8.01

-1.91 4.22

-10

-6

-2

2

Pre-GST

Post-GST1

6

10

Post-GST2

14

18

Post-GST3

Fig. 6.6c Average annual growth rate of RUP and state GST* (%). Note *—including IGST settlement. Source Computed by author based on Data obtained from state finance accounts/budget documents 7.99 -3.89

West Bengal

2.72 9.72 8.85 -6.24

Uttar Pradesh

3.14 8.87 20.06 -4.24

Telangana*

2.31 6.92 6.56 -1.54

Tamil Nadu

7.24 13.84 -10

-6 Pre-GST

-2

2

Post-GST1

6 Post-GST2

10

14

18

22

Post-GST3

Fig. 6.6d Average annual growth rate of RUP and state GST** (%). Note *—including IGST settlement. Source Computed by author based on Data obtained from state finance accounts/budget documents

134

6.5.1.1

S. Mukherjee

Analysis of GST Revenue: Pre- Versus Post-GST

Protection of revenue from taxes subsumed into GST could also be seen in terms of sustaining the revenue stream as percentage of nominal GSDP. In this sub-section, we compare the revenue from taxes subsumed into GST for the pre-GST period with State GST (including IGST settlement) for the post-GST period. We have also taken into account GST compensation receipts of states and created two additional series of revenue stream, viz., Post-GST2 and Post-GST3. . Pre-GST: Average Annual Growth Rate in Revenue Under Protection during 2014–2015 to 2016–2017 . Post-GST1: Average Annual Growth Rate in State GST Collection (including IGST Settlement) during 2019–2020 to 2020–2021 . Post-GST2: Average Annual Growth Rate in State GST Collection (including IGST Settlement and GST Compensation) during 2019–2020 to 2020–2021 . Post-GST3: Average Annual Growth Rate in State GST Collection (including IGST Settlement, GST Compensation, and Back-to-Back Loans) during 2019– 2020 to 2020–2021 Figure 6.7a shows that none of the states could sustain the revenue stream corresponding to taxes subsumed into GST (in terms of percentage share in GSDP) in the post-GST period without GST compensation. With GST compensation, except Andhra Pradesh, other states presented in Fig. 6.7a have managed to attain average revenue which is marginally higher than the revenue that is subsumed into GST (as % of GSDP). In other words, GST compensation helped states to sustain the revenue stream which is subsumed into GST. Telangana is created from undivided Andhra Pradesh on June 2014 and it has an impact on the revenue stream of Andhra Pradesh. Figure 6.7b shows that GST compensation helped states to maintain the revenue corresponding to taxes subsumed into GST and there was no windfall benefit to any state from the GST compensation in terms of substantial higher share of GST in GSDP. It is to be highlighted that without GST compensation it would have been difficult for states to sustain the revenue that is subsumed into GST. Figure 6.7c shows that even after GST compensation (from all sources), Punjab and Telangana are not able to sustain the revenue corresponding to taxes subsumed into GST post introduction of GST. Other states with GST compensation could manage the revenue sustainability. This shows that both GST compensation from GST compensation fund and back-to-back loans in lieu of shortfall in GST compensation cess collection were necessary for sustaining the revenue streams of state governments. We present summary results of 18 major states in terms of share of either RUP or SGST in GSDP over the periods in Table 6.2. This shows that distribution of states across value range of share of RUP/SGST in GSDP (excluding GST compensation) has moved towards lower value ranges during 2018–2021 as compared to 2015– 2016. In other words, post-GST State GST (including IGST settlement) collection has suffered for states.

Andhra Pradesh*

Bihar Pre-GST

Post-GST1

Goa Post-GST2

3.1 2.7

2.8

Chhattisgarh

3.1

3.9 3.3

3.1

3.4

2.3

3.2

3.3

5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0

135

4.4

6 Revenue Implications of GST on Indian State Finances

Gujarat**

Haryana**

Post-GST3

Jharkhand

Karnataka Pre-GST

Kerala Post-GST1

3.2

3.3

3.5

3.2 2.6

2.8

3.2

3.0

3.4

3.4

3.1

5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0

3.1

Fig. 6.7a a Average annual share of RUP and state GST# in GSDP (%). Note *—Telangana is created in June 2014 from undivided Andhra Pradesh and it results fall share of RUP in GSDP. **— for Pre-GST the revenue corresponds to 2015–2016 only. #—including IGST settlement. Source Computed by author based on data obtained from state finance accounts/budget documents

Madhya Pradesh Post-GST2

Maharashtra

Odisha

Post-GST3

Fig. 6.7b Average annual share of RUP and state GST* in GSDP (%). Note *—Telangana is created in June 2014 from undivided Andhra Pradesh and it results fall share of RUP in GSDP. **—for PreGST the revenue corresponds to 2015–2016 only. #—including IGST settlement. Source Computed by author based on data obtained from state finance accounts/budget documents

6.5.2 Analysis State Tax Revenue: Pre- Versus Post-GST Various taxes from tax bases of the Union and state governments have subsumed into GST. Separation of revenue stream of individual taxes, those are subsumed into GST, is difficult. The problem of separation of the revenue stream becomes difficult when GST subsumes only one or two components (or leave out a few commodities from the

Punjab

Rajasthan Pre-GST

Tamil Nadu Post-GST1

2.6

2.6

3.1

2.9

2.7

2.9

2.6

2.5

2.5

2.7

4.0

5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0

S. Mukherjee

4.3

136

Telangana Uttar Pradesh West Bengal Post-GST2

Post-GST3

Fig. 6.7c Average annual share of RUP and state GST* in GSDP (%). Note *—Telangana is created in June 2014 from undivided Andhra Pradesh and it results fall share of RUP in GSDP. **—for PreGST the revenue corresponds to 2015–2016 only. #—including IGST settlement. Source Computed by author based on data obtained from state finance accounts/budget documents

Table 6.2 Distribution of 18 major states by share of RUP or state GST* in GSDP Share of RUP or SGST in GSDP

2015–2016

2018–2019

2019–2020

2020–2021

3.5%

3

(17)

1

(6)

0

(0)

0

(0)

Total no. states 18

18

18

18

Notes *—including IGST settlement but excluding GST compensation Figures in the parenthesis show the percentage share in Total Number of States Source Computed by author

GST system) of a major tax head (Table 6.1). In other words, if subsummation of taxes into GST is partial for a tax head, it becomes difficult to separate revenue stream into GST and non-GST components, given the disaggregated level of data available either from State Budget Documents or Finance Accounts. Since GST subsumes various tax components from multiple tax heads, GST collection may improve, depending on size of the tax base and buoyancy of the concerned tax components, whereas tax collection of the original tax head (i.e., tax head from where tax components are taken into the GST base) may suffer revenue loss (or the revenue stream may dry out) if the entire tax head is not subsumed into GST. Comparison of RUP for the pre-GST period with SGST for the post-GST period may not necessarily reveal all the revenue effects of GST. To overcome this data related problem, in this section we take up

6 Revenue Implications of GST on Indian State Finances

137

revenue analysis of all tax heads from where tax components that are subsumed into GST. To make the series comparable with the pre-GST period we also consider State GST (including IGST settlement) for the post-GST period. It is to be mentioned that in 2017–2018 and 2018–2019 a part of IGST settlements was based on tax devolution formula of the 14th Finance Commission. So, we have separately taken into account the state’s share in IGST (under the major budget head of 0008) in our analysis. Like State Basket of Revenue where taxes are either partially or fully subsumed into GST, we have separately prepared revenue stream corresponding to the Union taxes which are either partially or fully subsumed into GST and from where states receive their share through tax devolution (Tables 6.2, 6.3 and 6.4). Table 6.5 shows that except Maharashtra, average share of the State Revenue Basket in GSDP has fallen during 2018–2019 to 2020–2021 (hereafter Post-GST period) for all other major states as compared to the pre-GST period (2014–2015 to 2016–2017). Except Goa, average state’s share in the Union Taxes (in selected indirect taxes, as % of GSDP) has fallen during 2018–2019 to 2020–2021. Even with the GST compensation, five states (viz., Andhra Pradesh, Gujarat, Madhya Pradesh, Tamil Nadu, West Bengal) could not meet the average share of State Revenue Basket in GSDP as it was prevalent during 2014–2015 to 2016–2017 (Table 6.4). The end of GST compensation regime from 1 July 2022 may have serious revenue implication for some states. For some states (viz., Bihar, Jharkhand, Karnataka, Telangana, Uttar Pradesh) fall in the share of Union taxes (related to indirect taxes having bearing with the Table 6.3 State revenue basket (either partially or fully subsumed into GST) Major head

Description

0040

Taxes on sales, trade etc Partial • Except petrol, diesel, ATF, natural gas, crude petroleum and alcoholic beverages for human consumption

0042

Tax on goods and passengers

Partial • Tax on entry of goods into Local Areas (0042-106)

0045

Other taxes and duties on commodities and services

Partial e.g., • Entertainment tax (0045-101) • Betting tax (0045-102) • Luxury tax (0045-105) • Taxes on advertisement exhibited in cinema theatres (0045-111)

0023

Hotel receipts tax

Fully

0006

State goods and services Fully tax (SGST)

0008

Integrated goods and services tax (IGST)

Source Author

Taxes subsumed into GST

In 2017–2018 and 2018–2019, a part of IGST Settlement was based on tax devolution formula of the 14th finance commission

138

S. Mukherjee

Table 6.4 The union government revenue basket (either partially or fully subsumed into GST)* Major head

Description

Taxes subsumed into GST

0005

Central goods and services tax (CGST)

Central GST (including IGST settlement)

0037

Customs

Partial • Additional Customs Duty commonly known as Countervailing Duty (or CVD) • Special Additional Duty of Customs (or SAD) • Cesses and surcharges

0038

Union excise duties (UED)

Partial • Union excise duty (except petrol, diesel, ATF, crude petroleum, natural gas, tobacco) • Additional excise duty • Excise duty levied under the medicinal and toilet preparations (Excise duties) Act, 1955 • Cesses and surcharges

0044

Services tax

Full

0045

Other taxes and duties on commodities and services

Partial • State’s share in the tax

Note *—Pertaining to the union of taxes from where states receive their share through tax devolution Source Author

GST) during the post-GST period exceeds the gain (or positive difference between the post-GST and pre-GST average share of the State Revenue Basket in GSDP) from the GST regime (post-GST compensation) (Table 6.6). This results in fall in the share of Combined Revenue Basket of the concerned states during the post-GST period. This shows that states not only faced the revenue shortage due to State GST collection (including IGST settlement) but also fall in revenue of selected Union taxes (related to GST) resulted in lower tax devolution from the Union government. For some states, the spill-over effect of revenue shortage of the Union taxes in terms lower tax devolution is stronger than fall in own GST collection (including GST compensation receipts). Therefore, in a federal system, sub-national governments may face the impact of revenue shortfall differently than the federal government and also the impact will vary across sub-national governments, depending on tax devolution from the federal government. The revenue impact on the sub-national government will depend not only on their own revenue performance but also on the revenue performance of the federal government, as the latter may spill-over to subnational finances through tax devolution. In other words, revenue implications of any tax reform may be felt differently by different levels of governments and across sub-national governments. Revenue compensation to sub-national governments to mitigate revenue uncertainty associated with any tax reform may not completely take care of revenue uncertainties which are related to revenue shortfall of the federal government.

6 Revenue Implications of GST on Indian State Finances

139

Table 6.5 State-wise average share of state revenue as well as state’s share in the selected union taxes—without GST compensation (% of GSDP)*

Seven State

State Revenue Basket (1)

State's Share in the Selected Union Taxes (2)

Combined Revenue Basket of State (1+2)

PreGST (A)

Pre-GST (C )

Post-GST (D)

Pre-GST (E )

PostGST (B)

B-A

D-A

PostGST (F)

F-E

Andhra Pradesh

5.15

4.29

-0.86

1.54

1.16

-0.38

6.69

5.45

-1.24

Bihar

4.22

3.92

-0.30

5.62

4.39

-1.23

9.84

8.31

-1.53

Chhattisgarh

4.33

3.66

-0.68

2.68

2.44

-0.24

7.01

6.10

-0.91

Goa

5.43

4.56

-0.87

1.34

1.37

0.03

6.77

5.94

-0.84

Gujarat**

4.42

3.42

-1.00

0.63

0.52

-0.12

5.05

3.94

-1.11

Haryana**

4.39

3.70

-0.69

0.46

0.38

-0.08

4.85

4.08

-0.77

Jharkhand

4.19

3.85

-0.34

3.01

2.62

-0.39

7.20

6.47

-0.73

Karnataka

4.38

3.53

-0.85

0.93

0.71

-0.22

5.31

4.24

-1.07

Kerala

5.43

4.92

-0.52

0.92

0.74

-0.18

6.36

5.66

-0.70

Madhya Pradesh

4.24

3.43

-0.81

2.85

2.17

-0.68

7.09

5.60

-1.49

Maharashtra

3.87

4.34

0.47

0.59

0.55

-0.04

4.46

4.89

0.43

Odisha

4.23

3.90

-0.33

2.90

2.27

-0.63

7.13

6.17

-0.96

Punjab

4.20

3.47

-0.73

0.84

0.80

-0.04

5.04

4.27

-0.76

Rajasthan

4.00

3.93

-0.07

1.74

1.50

-0.24

5.74

5.43

-0.31

Tamil Nadu

5.24

4.64

-0.61

0.77

0.60

-0.17

6.01

5.23

-0.78

Telangana#

4.96

4.78

-0.19

0.89

0.66

-0.23

5.86

5.43

-0.42

Uttar Pradesh

4.21

4.11

-0.10

3.41

2.78

-0.63

7.62

6.89

-0.73

West Bengal

3.46

3.01

-0.45

1.96

1.60

-0.36

5.42

4.61

-0.81

Note *—Pre-GST implies average share during 2014–2015 to 2016–2017 and post-GST implies average Share during 2018–2019 to 2020–2021. **—Pre-GST figure corresponds to 2015–2016 only. #—Pre-GST figure corresponds to average of 2015–2016 and 2016–2017 Source Computed by author based on data obtained from state finance accounts/budget documents

6.5.2.1

Comparison of Tax Revenue of States: Pre-GST Versus Post-GST

We compare state’s Own Tax Revenue (OTR) and state’s Tax Revenue (STR = OTR + state’s share in the Union taxes) between the pre- and post-GST periods to see if there is any change in the corresponding revenue stream (as % of GSDP) over the periods (Table 6.7). Except Maharashtra and Telangana, there is fall in the average share of OTR in GSDP for all other major states during the post-GST period. Except Chhattisgarh, Goa, Jharkhand, Maharashtra and Punjab, average state’s share in the Union taxes (as % of GSDP) has also fallen during the post-GST period for all other major states. As a result, the average share of STR (in GSDP) has fallen for states, except Maharashtra and Tamil Nadu during the post-GST period. Introduction of GST may not be the only reason behind this fall in the shares of OTR and STR in GSDP, there may be several other factors like slowing down of economic growth and COVID-19 pandemic. However, in this analysis we do not take into account GST compensations receipts of states.

140

S. Mukherjee

Table 6.6 State-wise average share of state revenue as well as state’s share in the selected union taxes—with GST compensation (% of GSDP)* State Revenue Basket (1) State

PreGST (A)

PostGST (B)

B-A

State's Share in the Selected Union Taxes (2) PreGST (C )

Post-GST (D)

D-C

Combined Revenue Basket of State (1+2) PreGST ( E)

Post-GST (F)

F-E

Andhra Pradesh

5.15

4.55

-0.60

1.54

1.16

-0.38

6.69

5.71

-0.98

Bihar

4.22

4.73

0.50

5.62

4.39

-1.23

9.84

9.11

-0.73

Chhattisgarh

4.33

4.79

0.46

2.68

2.44

-0.24

7.01

7.23

0.22

Goa

5.43

6.22

0.79

1.34

1.37

0.03

6.77

7.59

0.82

Gujarat**

4.42

4.19

-0.23

0.63

0.52

-0.12

5.05

4.71

-0.34

Haryana**

4.39

4.49

0.10

0.46

0.38

-0.08

4.85

4.87

0.02

Jharkhand

4.19

4.50

0.32

3.01

2.62

-0.39

7.20

7.12

-0.07

Karnataka

4.38

4.57

0.19

0.93

0.71

-0.22

5.31

5.28

-0.03

Kerala

5.43

5.78

0.35

0.92

0.74

-0.18

6.36

6.53

0.17

Madhya Pradesh

4.24

4.04

-0.20

2.85

2.17

-0.68

7.09

6.21

-0.88

Maharashtra

3.87

4.99

1.12

0.59

0.55

-0.04

4.46

5.55

1.08

Odisha

4.23

4.87

0.64

2.90

2.27

-0.63

7.13

7.14

0.01

Punjab

4.20

5.13

0.93

0.84

0.80

-0.04

5.04

5.93

0.89

Rajasthan

4.00

4.50

0.50

1.74

1.50

-0.24

5.74

6.00

0.26

Tamil Nadu

5.24

5.16

-0.08

0.77

0.60

-0.17

6.01

5.76

-0.25

Telangana#

4.96

5.04

0.08

0.89

0.66

-0.23

5.86

5.70

-0.15

Uttar Pradesh

4.21

4.52

0.31

3.41

2.78

-0.63

7.62

7.30

-0.33

West Bengal

3.46

3.45

-0.01

1.96

1.60

-0.36

5.42

5.05

-0.37

Note Pre-GST implies average share during 2014–2015 to 2016–2017 and post-GST implies average share during 2018–2019 to 2020–2021. GST compensation includes compensation from the GST compensation fund as well as back-to-back loan from the union government. **—Pre-GST figure corresponds to 2015–2016 only. #-Pre-GST figure corresponds to average of 2015–2016 and 2016– 2017 Source Computed by author based on data obtained from state finance accounts/budget documents

Except Andhra Pradesh, Bihar, Gujarat, Karnataka, Madhya Pradesh and Tamil Nadu, difference in the average share of state’s OTR (with GST compensation) in GSDP between the post- and pre-GST periods becomes positive for other major states. Commensurate with GST compensation receipt of a state, the average share of the state’s STR in GSDP improves during the post-GST period. However, as mentioned earlier average state’s share in the Union taxes has fallen for many states during the post-GST period. As a result, the improvement in average share of STR in GSDP is restricted to only some states during the post-GST period. Like state governments, the Union government has also faced revenue shortfall in GST collection and in the absence of any mechanism for revenue compensation from the GST compensation fund, the Union government has raised the “Non-Shareable Duties” and “Cesses on Commodities” under Union Excise Duties (Mukherjee, 2022). Though this helped the Union government to mitigate the revenue shortfall in GST collection to some

6 Revenue Implications of GST on Indian State Finances

141

Table 6.7 State-wise average share of own tax revenue and state’s tax revenue (% of GSDP)— without GST compensation

State's Own Tax Revenue State's Tax Revenue Pre-GST (A) Post-GST (B) B-A Pre-GST (C) Post-GST (D) Andhra Pradesh 7.06 6.09 -0.97 10.52 9.12 Bihar 6.18 5.18 -1.00 18.82 16.61 Chhattisgarh 7.30 6.56 -0.74 13.27 12.91 Goa 7.38 6.17 -1.21 10.39 9.62 Gujarat 6.09 4.82 -1.27 7.51 6.17 Haryana 6.21 5.75 -0.46 7.24 6.74 Jharkhand 5.31 5.12 -0.18 12.03 11.94 Karnataka 7.26 6.17 -1.09 9.35 8.03 Kerala 6.82 6.16 -0.65 8.89 8.11 Madhya Pradesh 7.28 5.89 -1.40 13.69 11.55 Maharashtra 6.37 6.76 0.38 7.69 8.19 Odisha 6.33 6.10 -0.23 12.84 12.01 Punjab 6.85 5.75 -1.09 8.72 7.83 Rajasthan 6.13 6.06 -0.07 10.04 9.96 Tamil Nadu 6.92 6.01 -0.91 8.65 7.56 Telangana 6.69 7.18 0.50 8.69 8.90 Uttar Pradesh 7.04 7.25 0.21 14.73 14.50 West Bengal 5.34 5.06 -0.29 9.74 9.21 State

D-C -1.40 -2.21 -0.36 -0.76 -1.34 -0.50 -0.09 -1.32 -0.77 -2.13 0.50 -0.83 -0.89 -0.08 -1.09 0.21 -0.24 -0.53

Note Pre-GST implies average share during 2014–2015 to 2016–2017 and post-GST implies average share during 2018–2019 to 2020–2021 Source Computed by author based on data obtained from state finance accounts/budget documents

extent, this has resulted in lower tax devolution to states, as the revenue from these sources are not shared with the states (Table 6.8). GST compensation has helped states to reduce the difference between post- and pre-GST average share of OTR in GSDP. However, even after GST compensation the difference remains negative for Andhra Pradesh, Bihar, Gujarat, Madhya Pradesh and Tamil Nadu. This shows that though GST compensation helped states to moderate the revenue shortfall but it could not wipe out the entire difference between the average shares of OTR in GSDP between pre- and post-GST periods for all states. Some states could not the maintain average share of OTR in GSDP during 2018–2019 to 2020–2021 even after receiving GST compensation as it was prevalent during pre-GST period (Fig. 6.8). Like state’s OTR, the difference in state’s Tax Revenue (STR = OTR + State’s share in the Union taxes) in GSDP the between post- and pre-GST periods varies across states. Even with GST compensation the difference cannot be wiped out for some states, as a result those states faced revenue stress during the post-GST period (Fig. 6.9).

142

S. Mukherjee

Table 6.8 State-wise average share of own tax revenue and state’s tax revenue (% of GSDP)—with GST compensation

State's Own Tax Revenue State's Tax Revenue Pre-GST (A) Post-GST (B) B-A Pre-GST (C) Post-GST (D) Andhra Pradesh 7.06 6.34 -0.72 10.52 9.37 Bihar 6.18 5.99 -0.19 18.82 17.42 Chhattisgarh 7.30 7.70 0.40 13.27 14.04 Goa 7.38 7.83 0.45 10.39 11.28 Gujarat 6.09 5.59 -0.50 7.51 6.93 Haryana 6.21 6.54 0.33 7.24 7.52 Jharkhand 5.31 5.78 0.47 12.03 12.59 Karnataka 7.26 7.21 -0.05 9.35 9.08 Kerala 6.82 7.03 0.21 8.89 8.98 Madhya Pradesh 7.28 6.50 -0.78 13.69 12.17 Maharashtra 6.37 7.41 1.04 7.69 8.84 Odisha 6.33 7.07 0.74 12.84 12.98 Punjab 6.85 7.41 0.57 8.72 9.49 Rajasthan 6.13 6.62 0.49 10.04 10.53 Tamil Nadu 6.92 6.54 -0.39 8.65 8.09 Telangana 6.69 7.45 0.76 8.69 9.17 Uttar Pradesh 7.04 7.66 0.61 14.73 14.90 West Bengal 5.34 5.50 0.15 9.74 9.66 State

D-C -1.15 -1.40 0.77 0.90 -0.57 0.28 0.56 -0.27 0.10 -1.52 1.15 0.14 0.76 0.49 -0.56 0.48 0.17 -0.09

Note Pre-GST implies average share during 2014–2015 to 2016–2017 and post-GST implies average share during 2018–2019 to 2020–2021. GST compensation includes compensation from the GST compensation fund as well as back-to-back loan from the union government Source Computed by author based on data obtained from state finance accounts/budget documents

6.5.3 Fiscal Health of States: Pre- Versus Post-GST Except Maharashtra, Uttar Pradesh and West Bengal, rise in average revenue deficit is observed for all major states during the post-GST period (2017–2018 to 2020– 2021) as compared to the pre-GST period (2013–2014 to 2016–2017). Since GST compensation receipts are accounted under “Grants-in-Aids from the Centre” in the State Accounts, total revenue receipts account for GST compensation receipts also. The rise in revenue deficits during the post-GST period may not be solely attributed to revenue shortfall on account of GST collection; there are various other drivers for rising revenue deficits, e.g., COVID-19 pandemic, economic slowdown, rise in public expenditure due to the pandemic etc,. Except 8 states (viz., Bihar, Chhattisgarh, Goa, Karnataka, Kerala, MP, Odisha, Tamil Nadu) the average fiscal deficit has fallen during the post-GST period as compared to the pre-GST period. Since backto-back loan received by states from the Union government in lieu of shortfall in GST compensation cess collection is accounted under “Loans and Advances from the Central Government” in State Accounts, fiscal deficit figures for 2020–2021 are

6 Revenue Implications of GST on Indian State Finances

143

1.00 0.50 0.00 -0.50 -1.00 -1.50

Difference* in OTR (as % of GSDP) - Without GST Compensation Difference* in OTR (as % of GSDP) - With GST Compensation

Fig. 6.8 Difference in state’s own tax revenue (OTR) with and without GST compensation. Note *—Difference: average post-GST own tax revenue (OTR, as % of GSDP)—average pre-GST own tax revenue (OTR, as % of GSDP). Source Computed by author based on data obtained from state finance accounts/budget documents 1.50 1.00 0.50 0.00 -0.50 -1.00 -1.50 -2.00 -2.50

Difference* in STR (as % of GSDP) - Without GST Compensation Difference* in STR (as % of GSDP) - With GST Compensation

Fig. 6.9 Difference in state’s tax revenue (STR) with and without GST compensation. Notes *— Difference: average post-GST state tax revenue (STR, as % of GSDP)—average pre-GST state tax revenue (STR, as % of GSDP). Source Computed by author based on data obtained from state finance accounts/budget documents

144

S. Mukherjee

Table 6.9 State-wise average revenue and fiscal deficits (as % of GSDP) State

Average annual revenue deficit (% of GSDP)

Average annual fiscal deficit (% of GSDP)

2013–2014 to 2017–2018 to B-A 2013–2014 to 2017–2018 to D-C 2016–2017 (A) 2020–2021 (B) 2016–2017 (C) 2020–2021 (D) Andhra Pradesh

2.07

2.47

Rise 4.52

4.43

Fall

−2.42

−0.69

Rise 3.14

3.14

Rise

Chhattisgarh −0.51

0.83

Rise 2.51

3.69

Rise

Goa

−0.24

−0.02

Rise 2.49

2.71

Rise

Gujarat

−0.46

0.16

Rise 1.98

1.83

Fall

Bihar

Haryana

2.02

2.11

Rise 4.00

3.48

Fall

Jharkhand

−1.03

−0.56

Rise 3.47

3.15

Fall

Karnataka

−0.09

0.16

Rise 2.11

2.79

Rise

2.32

2.40

Rise 3.66

3.82

Rise

−1.07

0.12

Rise 2.87

3.59

Rise

Fall

Kerala Madhya Pradesh Maharashtra

0.41

0.40

1.64

1.63

Fall

Odisha

−2.11

−2.00

Rise 1.95

2.35

Rise

Punjab

2.00

2.60

Rise 5.63

3.27

Fall

Rajasthan

0.99

3.34

Rise 5.25

4.11

Fall

Tamil Nadu

0.70

2.05

Rise 2.94

3.48

Rise

Telangana*

−0.13

0.50

Rise 4.28

3.77

Fall

Uttar Pradesh

−1.53

−1.62

Fall

3.80

1.67

Fall

2.04

1.46

Fall

3.29

3.14

Fall

West Bengal

Note *—Pre-GST figure is average of 2015–2016 and 2016–2017 Source Computed by author based on data obtained from state finance accounts/budget documents

not corrected for back-to-back loans received by states. This shows that despite the rising average revenue deficit during the post-GST period, many states have contained average fiscal deficit during the post-GST period (Table 6.9).

6.6 Summary and Conclusions We observe a falling trend in the annual economic growth rate of India since 2006– 2007. Like World economy, Indian economy has also faced two major shocks during 2006–2007 to 2021–2022—Global Financial Crisis (GFC, 2008–2009 to 2009–2010) and COVID-19 Pandemic (2020–2021). The impact of COVID-19 pandemic on Indian economic growth was much stronger than the impact of GFC.

6 Revenue Implications of GST on Indian State Finances

145

We observe that except in West Bengal, all other major states show fall in the average annual growth rate of GSDP during post-GST period (2017–2018 to 2021–22RE) as compared to the pre-GST period (2012–2013 to 2016–2017). It is expected that fall in the growth rate in GSDP may have some impact on GST collection. Slowdown in economic growth in 2019–2020 and COVID-19 pandemic in 2020–2021 resulted in revenue as well as fiscal stresses for states. Like economic growth, Indian states also faced fiscal stresses during 2008–2010, 2015–2017 and 2020–2021 onwards. In addition to revenue stress due to GFC, implementations of 6th Pay Commission recommendations by many states increased revenue expenditures and resulted in rises of revenue as well as fiscal deficits during 2008–2010. States also faced a relatively stronger fiscal stress during 2015–2016 to 2016–2017 due to adoption of Ujwal DISCOM Assurance Yojana (UDAY) scheme to provide debt relief to public Power Distribution Companies (DISCOM). States again faced fiscal shock during 2020–2021 due to the COVID-19 pandemic. The economic contraction during the pandemic created pressures on Public Finance Management (PFM) in terms of lower revenue mobilization and higher expenditure demands. Both the Union and state governments faced dual problems of arresting economic contraction and managing public finance with limited public resources available for disposal. Analysis of GST collections of 18 major states shows that on average states collected 88.3% of Revenue Under Protection (or aggregate projected revenue of states in GST) in 2018–2019. In 2019–2020, on average SGST collection (including IGST settlement) could meet only 76% of the RUP and it further falls to 61% in 2020– 2021. GST compensation helped states to protect revenue and with GST compensation on average states achieved 99.3% of RUP in 2018–2019. In 2019–2020, with GST compensation states achieved on average 93% of RUP. Due to shortfall in GST compensation cess collection in 2020–2021, states could achieve 78.7% of RUP with GST compensation from the GST compensation fund. However, in 2020– 2021 states also received GST compensation in terms of back-to-back loans from the Union government in lieu of shortfall in GST compensation cess collection and this helped states to achieve on average 91.6% of RUP in 2020–2021. The analysis shows that state GST collection (as percentage of RUP) is falling over the years during 2018–2021 and the dependence on GST compensation has grown up. To assess the revenue impact of GST on Indian state finances, we adopt different measures of revenue streams associated with GST and compare the revenue performance of major states over the periods (pre-GST versus post-GST). We compare the average annual growth rate of revenue from taxes subsumed into GST for the Pre-GST period with State GST (including IGST settlement) for the post-GST period. In addition to State GST (including IGST settlement), states receive GST compensation. To make comparable series of revenue for pre- and postGST periods, we add GST compensation receipts of states with the SGST revenue. We construct three alternative revenue streams for the post-GST period by adding GST compensation receipts from the GST compensation fund and back-to-back loans that states receive in lieu of shortfall in GST compensation cess collection. Our observations are as follows:

146

S. Mukherjee

. Except Andhra Pradesh, the average annual growth rate in revenue from taxes subsumed into GST during the pre-GST period (2014–2015 to 2016–2017) is higher than average annual growth rate in SGST (including IGST settlement but excluding GST compensation) during the post-GST period (2019–2020 to 2020– 2021) for other major Indian states. It is to be highlighted that Telangana is created from undivided Andhra Pradesh in June 2014 and therefore pre-GST growth rate of Andhra Pradesh was lower. . Except Andhra Pradesh, Goa and Tamil Nadu, the pre-GST average annual growth rate in revenue from taxes subsumed into GST is higher than the post-GST average annual growth rate in SGST (including IGST settlement and GST compensation from GST compensation fund) for other major states. . Except, Bihar, Jharkhand, Karnataka, Madhya Pradesh, Maharashtra, Punjab, Rajasthan and Telangana, the pre-GST average annual growth rate in revenue from taxes subsumed into GST is lower than the post-GST average annual growth rate in SGST (including IGST settlement and GST compensation from GST compensation fund as well as back-to-back loans from the Union government) for other major states. . Among major states, Odisha is the only state which has achieved a positive average annual growth rate in GST collection without any GST compensation in the postGST period. . Except for a few states (viz., Chhattisgarh, Goa, Kerala, Odisha), post-GST average annual growth rate in GST collection (with GST compensations from all sources) is lower than 14%. . Many States could not attain 14% average annual growth rate in GST collection during 2019–2020 to 2020–2021 even with GST Compensation (from all sources). This implies that though GST compensation helped states to moderate the revenue shortfall in GST collection during 2019–2020 to 2020–2021, it could not wipe out the entire revenue shortfall in the GST collection for some states. Therefore, GST compensation was necessary for states to sustain the revenue stream which has subsumed into GST. Protection of revenue from taxes which are subsumed into GST could also be seen in terms of sustaining the revenue stream as percentage of GSDP. Our analysis shows that none of the states could sustain the revenue stream corresponding to taxes subsumed into GST (in terms of percentage share in GSDP) in the postGST period without GST compensation. Even after GST compensation (from all sources), Andhra Pradesh, Madhya Pradesh and Telangana are not able to sustain the revenue corresponding to taxes subsumed into GST post introduction of GST. This shows that GST compensation helped states to maintain the revenue corresponding to taxes subsumed into GST and there was no windfall benefit to any state from the GST compensation in terms of substantial higher share of GST in GSDP. It is to be highlighted that without GST compensation it would have been difficult for states to sustain the revenue that is subsumed into GST. Comparison of RUP for the pre-GST period with SGST for the post-GST period may not necessarily reveal all revenue effects of GST. To overcome this data related

6 Revenue Implications of GST on Indian State Finances

147

problem, we take up revenue analysis of all tax heads from where tax components are subsumed into GST. To make the series comparable with the pre-GST period we also consider State GST (including IGST settlement) for the post-GST period. Like State Basket of Revenue where taxes are either partially or fully subsumed into GST, we have separately prepared revenue stream corresponding to the selected Union taxes which are either partially or fully subsumed into GST and from where states’ receive their share through tax devolution. We notice that except Maharashtra, the average share of State Revenue Basket in GSDP has fallen during 2018–2019 to 2020–2021 for all other major states as compared to the pre-GST period (2014–2015 to 2016–2017). Except Goa, the average state’s share in the Union Taxes (in indirect taxes, as % of GSDP) has fallen during 2018–2019 to 2020–2021 for other states. Even with GST compensation, five states (viz., Andhra Pradesh, Gujarat, Madhya Pradesh, Tamil Nadu, West Bengal) could not meet the average share of State Revenue Basket in GSDP as it was prevalent during 2014–2015 to 2016–2017. The end of GST compensation regime from 1 July 2022 may have serious fiscal implication for some states. For some states (viz., Bihar, Jharkhand, Karnataka, Telangana, Uttar Pradesh) fall in the share of Union taxes (related to indirect taxes having bearing with the GST, as % of GSDP) during the post-GST period exceeds the gain (or positive difference between the post-GST and pre-GST average share of the State Revenue Basket in GSDP) from the GST regime (post GST compensation). This results in fall in the share of Combined Revenue Basket (State’s Revenue+Share in the Selected Union taxes) of concerned states during the post-GST period. This shows that states not only faced the revenue shortage due to State GST collection (including IGST settlement) but also a fall in the selected Union taxes collections (related to GST) which resulted in lower tax devolution from the Union government. For some states, the spill-over effect of revenue shortage of the Union taxes in terms lower tax devolution is stronger than the fall in own GST collection (including GST compensation receipts). Therefore, in a federal system, sub-national governments may face the impact of revenue shortfall differently than the federal government and also the impact will vary across sub-national governments. The revenue impact on the sub-national government will depend not only on their own revenue performance but also on revenue performance of the federal government, as the latter may spill-over to sub-national finances through tax devolution. In other words, revenue implications of any tax reform may be felt differently by different levels of governments and across sub-national governments. Revenue compensation to sub-national governments to mitigate revenue uncertainty associated with any tax reform may not completely take care of revenue uncertainties which are related to revenue shortfall of the federal government. We compare the state’s Own Tax Revenue (OTR) and state’s Tax Revenue (STR = OTR + state’s share in the Union taxes) between the pre- and post-GST period to see if there is any change in the share of GSDP over the periods. Except Maharashtra and Telangana, there is fall in the average share of OTR in GSDP for all other major states during the post-GST period. Except Chhattisgarh, Goa, Jharkhand, Maharashtra and Punjab, the average state’s share in the Union taxes (as % of GSDP) also falls during the post-GST period for all other major states. As a result, the average share of STR

148

S. Mukherjee

(in GSDP) falls for states, except Maharashtra and Telangana during the post-GST period. Introduction of GST may not be the only reason behind this fall in the shares of OTR and STR in GSDP. There may be several other factors like slowing down of economic growth and COVID-19 pandemic. However, in this analysis we do not take into account GST compensations receipts of states. Except Andhra Pradesh, Bihar, Gujarat, Karnataka, Madhya Pradesh and Tamil Nadu, difference in the average share of state’s OTR (with GST compensation) in GSDP between post- and pre-GST period becomes positive for other major states. Commensurate with GST compensation receipt of a state, average share of state’s STR in GSDP improves. However, as mentioned earlier average state’s share in the Union taxes has fallen for many states during the post-GST period. As a result, the improvement in average share of STR in GSDP is restricted to some states during the post-GST period. Like state governments, the Union government also faced revenue shortfall in GST collection and in the absence of any mechanism for revenue compensation from the GST compensation fund, the Union government raised the “Non-Shareable Duties” and “Cesses on Commodities” under Union Excise Duties (Mukherjee 2022). Though this helped the Union government to mitigate the revenue shortfall in GST collection to some extent, this resulted in lower tax devolution to states. Except Maharashtra, Uttar Pradesh and West Bengal, rise in average revenue deficit is observed for all major states during post-GST period (2017–2018 to 2020– 2021) as compared to pre-GST period (2013–2014 to 2016–2017). Since GST compensation receipts are accounted under “Grants-in-Aids from the Centre” in the State Accounts, total revenue receipts accounts for GST compensation receipts also. The rise in revenue deficits during post-GST period may not be solely attributed to revenue shortfall on account of GST collection alone; there are various other drivers for rising revenue deficits, e.g., COVID-19 pandemic, economic slowdown, rise in public expenditure due to the pandemic. Except 8 states (viz., Bihar, Chhattisgarh, Goa, Karnataka, Kerala, MP, Odisha, Tamil Nadu) average fiscal deficit has fallen during the post-GST period as compared to the pre-GST period. Since back-to-back loans received by states from the Union government in lieu of shortfall in GST compensation cess collection is accounted under “Loans and Advances from the Central Government” in State Accounts, fiscal deficit figures for 2020–2021 are not corrected for back-to-back loans received by states. This shows that despite rising average revenue deficit during post-GST period, many states have contained average fiscal deficit during the post-GST period.

Appendix See Table A1.

18,698

11,956

3,232

42,752

22,565

9,497

53,549

24,922

22,711

89,640

16,370

21,441

25,421

44,130

23,866

Chhattisgarh

Goa

Gujarat

Haryana

Jharkhand

Karnataka

Kerala

Madhya Pradesh

Maharashtra

Odisha

Punjab

Rajasthan

Tamil Nadu

Telangana

(103)

24,206 (101)

39,137 (89)

23,763 (93)

13,510 (63)

12,639 (77)

83,181 (93)

19,751 (87)

21,390 (86)

42,663 (80)

8,201 (86)

18,775 (83)

35,351 (83)

2,529 (78)

8,665 (72)

16,738 (90)

21,257

24,206

42,288

25,939

20,639

16,029

91,511

22,617

24,274

53,417

9,230

21,595

41,500

3,005

10,926

19,309

21,257

(101)

(96)

(102)

(96)

(98)

(102)

(100)

(97)

(100)

(97)

(96)

(97)

(93)

(91)

(103)

(103)

27,207

50,308

28,980

24,442

18,662

102,190

25,890

28,411

61,046

10,827

25,724

48,737

3,684

13,630

21,316

23,431

(86)

23,517 (86)

38,376 (76)

21,954 (76)

12,751 (52)

13,204 (71)

82,602 (81)

20,448 (79)

20,447 (72)

42,147 (69)

8,418 (78)

18,873 (73)

34,107 (70)

2,438 (66)

7,895 (58)

15,801 (74)

20,227

Revenue SGST under (including IGST protection settlement)

20,554

2019–2020

Revenue SGST under (including IGST protection settlement)

SGST (including IGST settlement and GST compensation)

2018–2019

Bihar

Andhra Pradesh

State

Table A1 State-wise revenue performance in GST (Rs. crore)

25,780

47,298

26,394

21,556

17,132

97,620

24,979

26,022

56,644

9,950

24,326

44,753

3,257

10,976

19,325

22,068

(95)

(94)

(91)

(88)

(92)

(96)

(96)

(92)

(93)

(92)

(95)

(92)

(88)

(81)

(91)

(94)

SGST (including IGST settlement and GST compensation)

31,016

57,351

33,037

27,864

21,275

116,496

29,515

32,388

69,592

12,343

29,325

55,561

4,200

15,538

24,300

26,712

(71)

22,190 (72)

37,942 (66)

20,755 (63)

11,819 (42)

13,043 (61)

69,949 (60)

17,257 (58)

20,028 (62)

37,711 (54)

7,931 (64)

18,236 (62)

29,459 (53)

1,985 (47)

7,925 (51)

16,050 (66)

18,871

Revenue SGST under (including IGST protection settlement)

2020–21

25,293

48,545

26,388

21,513

17,405

87,372

22,551

26,750

51,500

9,889

23,302

40,793

3,668

11,137

20,410

22,399

(82)

(85)

(80)

(77)

(82)

(75)

(76)

(83)

(74)

(80)

(79)

(73)

(87)

(72)

(84)

(84)

SGST (including IGST settlement and GST compensation)

(93)

(continued)

27,673 (89)

54,786 (96)

30,992 (94)

22,411 (80)

21,227 (100)

99,349 (85)

27,093 (92)

32,516 (100)

63,907 (92)

11,578 (94)

27,654 (94)

50,015 (90)

4,508 (107)

14,246 (92)

24,315 (100)

24,710

SGST (including IGST settlement and GST compensation from all sources*)

6 Revenue Implications of GST on Indian State Finances 149

29,776

468,722 (88)

28,166 (95)

48,802 (99)

526,993

30,143

49,110

(99)

(101)

(99)

604,821

33,944

56,391

457,743 (76)

27,308 (80)

47,232 (84)

562,158

31,666

52,412

(93)

(93)

(93)

SGST (including IGST settlement and GST compensation)

2020–21

689,496

38,696

64,285

420,025 (61)

26,013 (67)

42,860 (67)

Revenue SGST under (including IGST protection settlement)

542,887

31,790

52,184

(79)

(82)

(81)

SGST (including IGST settlement and GST compensation)

631,390 (92)

36,221 (94)

58,191 (91)

SGST (including IGST settlement and GST compensation from all sources*)

Note *—includes GST compensations from GST compensation fund as well as back-to-back loans from the union government in lieu of shortfall in GST compensation cess collection. Figures in the parenthesis show the percentage share in revenue under protection Source Computed by author based on data obtained from state finance accounts/budget documents

530,545

West Bengal

Major States

49,466

2019–2020 Revenue SGST under (including IGST protection settlement)

SGST (including IGST settlement and GST compensation)

2018–2019

Revenue SGST under (including IGST protection settlement)

Uttar Pradesh

State

Table A1 (continued)

150 S. Mukherjee

6 Revenue Implications of GST on Indian State Finances

151

References Mehta, D., & Mukherjee, S. (2021). Emerging issues in GST law and procedures: An assessment. NIPFP Working Paper No. 347, National Institute of Public Finance and Policy, New Delhi. Mukherjee, S. (2019). State of public finance and fiscal management in India during 2001–16. Working Papers 265, National Institute of Public Finance and Policy, New Delhi. Mukherjee, S. (2022). Revenue assessment of goods and services tax (GST) in India. NIPFP Working Paper No. 385, National Institute of Public Finance and Policy, New Delhi. Mukherjee, S. (2023). Revenue performance assessment of Indian GST. NIPFP Working Paper No. 392, National Institute of Public Finance and Policy, New Delhi. Mukherjee, S., & Badola, S. (2022). Public finance management in India in the time of COVID-19 pandemic. Indian Economic Journal, 70(3), 452–547.

Chapter 7

Equalization Transfers Policy Based on Expenditure Needs and Own Revenue Capacity of Indian State Governments K. R. Shanmugam and K. Shanmugam

Abstract This study addresses an important policy issue pertaining to the determination of equalization transfers to Indian States. It measures the revenue expenditure needs and own revenue fiscal capacity of States normatively and determines the equalization transfers in four alternative scenarios. The estimated amounts of equalization transfers for all 29 States range from | 3,73,956 crore to | 14,77,282 crore. These findings will be useful to policymakers and other researchers to design more effectively the transfers policy besides general-purpose transfers to address the resource gap such that the Indian States can provide a standard level of public services. Keywords Equalization transfers · Expenditure gaps · Revenue capacity · Indian states · Panel methodology JEL Classification H77 · H73 · H72 · C23

7.1 Introduction India has a federal system of Governments: Center or national Government and State or sub-national Governments. The Indian Constitution has assigned separate revenue sources and spending responsibilities to the Center and States based on the Principle of Separation. Most buoyant and mobile taxes like income tax, corporate income tax, central excise, etc. are assigned to the Centre e, while less buoyant and immobile taxes including sales tax, motor vehicle tax, stamps, and registration are allocated to the States. At the same time, relatively more expenditure functions are given to K. R. Shanmugam (B) Madras School of Economics, Chennai, India e-mail: [email protected] K. Shanmugam Government of Tamil Nadu, Chennai, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 D. K. Srivastava and K. R. Shanmugam (eds.), India’s Contemporary Macroeconomic Themes, India Studies in Business and Economics, https://doi.org/10.1007/978-981-99-5728-6_7

153

154

K. R. Shanmugam and K. Shanmugam

the States. Thus, by design, asymmetry arises in allocating revenue resources and spending responsibilities between the sub-national and the national Governments, causing vertical imbalances. As State Governments differ in their fiscal capacity to raise resources due to various factors like economic structure and institutional factors, rich States are better able to finance public service provisions compared to other poor States. The States may also differ in their expenditure priorities or needs, depending on their socioeconomic and demographic profiles. Further, in States with equal fiscal capacity, the costs incurred on the same level of public goods may differ because of variations in price levels, climatic conditions, etc. This leads to horizontal fiscal imbalances across jurisdictions/States. Both vertical and horizontal fiscal imbalances pose a challenge to States in providing a standard bundle of public services across jurisdictions. As a remedy, Buchanan (1950) suggested the provision of equalization transfers. Other supporters like Boadway (1980), Boadway and Flatters (1991) also viewed that the equalization transfers could bring equity, because it could ensure equal fiscal treatment of identical jurisdictions in a federation. It could also discourage fiscally induced migration and help the jurisdictions to provide a minimum standard level of public services, thereby enhancing economic “efficiency” (Shah, 1994).1 Thus, they could ensure the normative aspects of both equity and efficiency (Munoz et al., 2016). However, the median voter theorem argues that the transfers would crowd out local revenues as sub-national Governments would pass them to local citizens in the form of reductions in local taxes/fees (Broadford & Oates, 1971). Thus, they would likely exert disincentive effects on State’s tax/revenue efforts. Further, distributing the transfers as lower taxes would crowd out local spending (Scott, 1952). Countries like Australia, Canada, Germany, and Switzerland have developed their own models of equalization with different implications for equity, incentives, and distribution (Bahl et al., 1992; Blair, 1992; Boadway, 2004; Ladd & Yinger, 1994; and Ridge, 1992). Among them, the Canadian and Australian systems are two wellestablished models of equalization: the former focuses on fiscal capacity equalization while the latter focuses on both fiscal capacity and expenditure equalization. However, in both models, there is no reference to the efficiency consideration (Rangarajan & Srivastava, 2004). In India, to mitigate both vertical as well as horizontal imbalances, the Constitution provided for an intergovernmental transfer mechanism. The Finance Commission (FC) has been constituted once every five years to determine transfers (tax devolution and grants) from the Center to the States. This was done mainly using the “gap-filling approach”, which determines the assessment of tax revenues of the States based on past performance. This approach does not consider the efficiency in raising revenues. The Ninth FC attempted to use the “Representative Tax System” (RTS) to some 1

Boadway et al. (1993) argue that the equalization transfers that reduce net fiscal benefit differentials create one of those rare instances in economics when equity and efficiency considerations coincide. Other considerations used for equalization transfers include the prevention of secessionist tendencies in countries with relatively high regional tension (Martinez-Vazquez and Searle, 2007; Spahn, 2007).

7 Equalization Transfers Policy Based on Expenditure Needs and Own …

155

extent. The problem with the RTS is that it derives revenue capacity which is closer to the average and not the maximum/potential level. Since 1950, the Planning Commission Grants and grants through Centrally sponsored Schemes have also supplemented the Finance Commission’s transfers. After the abolition of the Planning Commission in 2014, the finance commission’s transfers have been dominant. In the State Government’s budget, the Central transfers have a considerable share. From 2011–12 to 2018–19, the Central transfers constituted around 38.3% (2013–14) to 47.1% (2016–17) of the total revenue receipts of (29) Indian States.2 Most of these transfers were general (i.e., unconditional), and only a small fraction of them were specific (conditional).3 Further, each FC uses a different formula/criterion which leads to uncertainty in the shares of States. As a result, the central transfers have varied widely across States and years. For instance, in 2018–19, the per capita transfer was the lowest in Haryana (|5,434) and the highest in Arunachal Pradesh (|90,124). There are also wide variations in other fiscal parameters. In 2018–19, the per capita own revenue was about |2,800 (lowest) in Bihar and |40,500 (highest) in Goa. Similarly, the per capita revenue expenditure was about |10,500 (lowest) in Bihar and |79,200 (highest) in Sikkim. These variations indicate that some States are not taking full effort to generate their revenues, i.e., they do not realize their potential level of revenues. A solution to these fiscal disparities is “Equalization Transfers”, with normatively determined own revenue capacity and expenditures of State Governments in India. This study is an attempt to design the “Equalization Transfers” based on normatively determined revenue capacity and expenditure needs of Indian States from 2005–06 to 2018–19. Specifically, it empirically determines the additional central transfers needed to provide a standard level of public services considering the full or average revenue potential of State Governments. The main contributions of this study are as follows: (i) while a large number of studies have emerged in the literature to examine the advantages and standards of equalization for different nations, only limited literature has examined how to practically equalize the fiscal capacity and spending needs (Maarten & Lewis, 2011). This study is a new addition to this sparsely researched literature; (ii) studies analyzing equalization transfers in developing nations (Munaz et al., 2016; Saraf & Srivastava, 2009) are very limited. The present study attempts to design the equalization transfers with normatively determined revenue capacity and revenue expenditure needs of General Category States (GCSs) and Special Category States (SCSs) in India, a developing country; (iii) the present study empirically examines whether transfers have an incentive or disincentive or no effect on own revenues and revenue expenses in the Indian States. The State-specific results of the study will help policymakers to take appropriate steps to obtain a horizontal balance. 2

Source: 15th Finance Commission Report, Vol. IV The States, 2020. While the approach pursued by the Finance Commission has an equalizing content, none of the Commission so far has formulated an explicit methodology on normative basis to derive the equalization transfers. The partial gap filling approach also creates an adverse incentive among States. Further, this approach is partially equalizing the cost conditions. Thus, the aim of achieving the horizontal equalization is not yet fulfilled in India.

3

156

K. R. Shanmugam and K. Shanmugam

This study is organized as follows. Section 7.2 briefly reviews the literature on the study topic. Section 7.3 discusses the empirical model, the data sources, and the estimation methods employed. The empirical results are presented and discussed in Sect. 7.4, while the policy conclusions are given in Sect. 7.5.

7.2 Literature Review Two contrary views emerged on the transfers to the sub-national Governments: (i) the first one is “First Generation Theory” (FGT) developed by Tiebout (1956), Musgrave (1959), and Oates (1972). This theory views equalization transfers as an important instrument in preventing rich regions from attracting more investments at the expense of poor regions; (ii) the other one is the recently emerged “Second Generation Theory” (SGT). It strongly favors a strong own revenue capacity of State Governments. This theory stresses the significance of a competitive environment for State/local Governments to achieve economic efficiency and it also argues that the Federal Government should not intervene in sub-national taxing and spending activities. Further, it argues that the Center’s fiscal interference is distortionary and creates incentive compatibility problems, which happen because the Center’s fiscal intervention may induce the State/local Governments to spend more, borrow heavily, and be dependent on the Center’s transfers. Although these two theoretical views contradict each other, large volumes of literature have emerged to analyze intergovernmental transfers and horizontal fiscal inequalities. They show that it is not possible for all jurisdictions to offer a standard bundle of public services at a comparable level of taxation. The sub-national Governments’ needs may differ, and their costs of providing public services may also differ. Some of them with a strong tax base may be richer than others and are in a better position to finance their public service provisions. These forces are the main reasons for net fiscal benefits which are the sources of inefficiency and inequality (Shanmugam & Shanmugam, 2022a).4 As a remedy, Buchanan (1950), Boadway (1980), etc. proposed equalization transfers. These transfers from the Central Government can discourage fiscally induced migration and ensure that all State Governments would be able to provide a standard bundle of public services at a comparable level of the tax rate. But Courchene (1978) and Scott (1952) strongly oppose this view and argue that the equalization transfers can increase inefficiency in the regional allocation of resources due to the fact that such transfers can discourage out-migration of labor to high-income States, where it will be more productive (Shah, 1994). 4

The literature argues that people will choose jurisdictions where they can optimize their net fiscal benefits. This will lead to inefficient allocation of labor across jurisdictions. If people with the same income level will not migrate despite differences in net fiscal benefits, there will be horizontal inequality. In such a case, the transfers will ensure efficiency in labor allocation and horizontal equity (Shanmugam & Shanmugam, 2022a).

7 Equalization Transfers Policy Based on Expenditure Needs and Own …

157

Shah (1988) and other opponents argue that with full capitalization, there is hardly any efficiency and equity basis for equalization transfers. The reason for this is that citizens in richer States pay more for private services than public services. On the other hand, citizens in poor States pay more for public services than private services. Another reason is that net benefits can be capitalized into property values. However, only at the time of property sale, a capital gain or loss on account of the local public sector can be realized (Shanmugam & Shanmugam, 2022a, 2022b). Therefore, there is no guarantee for Tiebout’s prescription that a system of local Governments would ensure optimal levels of local public services (Shah, 1994). The Median voter theorem argues that transfers will affect the local revenues of sub-national Governments negatively (i.e., crowing out effect) because these Governments would pass on the benefits of transfers to local citizens in the form of reduced local taxes and fees. Thus, they can lead to disincentive effects on the State’s tax/ revenue efforts. Further, this would crowd out local spending. Therefore, the effects of transfers on public spending and revenues should be determined empirically. Thus, these arguments on the equalization transfers form the basis for estimating the expenditure needs and fiscal capacity. As a result, many methodologies have emerged in the literature to measure expenditure needs and fiscal capacity (Munaz et al., 2016). Measuring Expenditure Needs The simplest approach to measuring expenditure needs uses historical expenditure patterns (Boex and Martinez-Vazquez, 2004). While this approach is simple, past expenditures may not accurately reflect spending needs. In addition, the expenditure norms and priorities may change over time. In such a case, the previous year’s expenditure may fail to reflect the current year’s policy objectives (Vaillancourt & Bird, 2007). The next approach posits that the expenditure needs are identical in all jurisdictions and, therefore, each jurisdiction can get an equal amount. Although it is also simple, it may widen the gap in per capita resource availability. The third approach uses the regression method in which the actual expenditures are regressed on need indicators and other determinants of State/local expenditure. Then, an allocation formula is determined using the estimated parameters associated with the need indicators, keeping constant the impact of non-need expenditure factors (Ladd, 1994). The limitation of this procedure is that it needs relevant data on regional characteristics, influencing regional expenditures. In addition, the validity of this approach depends on how well actual expenditures represent the expenditure needs. The fourth approach is the Representative Expenditure System (RES) method. In this approach, a sub-national government’s per capita expenditure need is estimated as the sum of its workload for each category of service weighted by the average spending on each unit of service, divided by the population (Shanmugam & Shanmugam, 2022b). A major limitation of this procedure is that it requires data on different categories of spending, workload, etc. (Maarten & Lewis, 2011). Measuring Fiscal Capacity Fiscal capacity is the ability of national and sub-national Governments to raise revenues from their own resource bases. Different methodologies are available

158

K. R. Shanmugam and K. Shanmugam

to measure the fiscal capacity of sub-national/State Governments. The simplest approach utilizes the data on current or past year’s revenue collections. However, this procedure has many limitations: (i) although the tax rates may not directly affect the potential or maximum possible ability to raise revenue, the fiscal effort and tax payer compliance affect the actual revenue; (ii) if the current revenue is used as the fiscal capacity, it may provide an incentive for the regional Governments to levy a low tax or to less effort in order to obtain more transfers; (iii) the use of past revenue collection may have time inconsistency because sub-national Governments may think that if they generate more current revenues by increased tax rates/efforts, they may get less transfers in future (Vaillancourt & Bird, 2005). The second approach considers a macroeconomic indicator (like GDP) at the regional level to measure fiscal capacity. Since it indicates the maximum capacity and not the actual fiscal capacity, it may not be a good indicator. In general, Governments’ revenues are significantly lower than their GDP. The third approach, called the Representative Revenue System (RRS), measures the fiscal capacity using the standard tax bases and rates. It basically employs the regression procedure to estimate the average revenue effort. The accuracy of RRS depends on the availability of data. For instance, Canada views the fiscal capacity of a province as its ability to raise revenues from personal income tax, corporate income tax, sales taxes, property tax, etc. It uses the data on the size of respective tax bases and the average effective tax rates to compute the revenue potential of each province. The Canadian province gets transfers if its revenue potential is below a threshold or “standard” (Boex and Martinez-Vazquez, 2004). Since the RRS uses the regression method, it provides the estimates of the average revenue capacity and not the maximum/potential capacity. Since the regression approach measures the average effort, the frontier approach is useful for measuring the potential revenues of regions. Broadly, two frontier approaches are available in the literature: the data envelopment approach (DEA) and the stochastic frontier approach (SFA). The DEA method, developed by Farrell (1957), considers that the actual revenue, R, is less than or equal to the potential or maximum possible revenue, R*(=f(X)), i.e., R ≤ R*, where R* is a deterministic quantity and X is a vector of determinants of revenue, including the revenue base. The revenue gap is given by u = R*-R and, due to a non-linear relationship, this relation can be written as R = f(.)e−u . The SFA approach for cross-section data, developed independently by Aigner et al. (1977) and Meeusen and van den Broeck (1977), considers that the potential revenue is not deterministic, but stochastic due to random factors and so the actual revenue can be written due to random factors as R = f (.) e−u ev = f(.) eε , where v is the regular stochastic error term and ε is the composite error term (=v-u). The Maximum Likelihood Estimation (MLE) method can be used to estimate this frontier revenue equation, assuming that the one-sided technical inefficiency term u follows either half normal or truncated normal or exponential or gamma distribution. Jondrow et al. (1982) procedure suggests how to compute the individual specific efficiency of sample units as the conditional expectation of e−u given ε (i.e., Revenue Efficiency = E (e−u | ε)).

7 Equalization Transfers Policy Based on Expenditure Needs and Own …

159

Schmidt and Sickles (1984) introduced the panel data version of SFA. This approach assumes that efficiency is time-invariant. Utilizing the Cobb–Douglas functional form (and lower cases indicate the logarithmic values), the fixed effects panel data model of the revenue equation is written as rit = α + xit ’ β + vit − ui , where ui is independently and identically distributed (i.i.d.) with mean μ and variance σ2 u , and x is a vector of determinants of revenues. If αi = α − ui , this equation becomes ' rit = αi + xit β + vit . The αi is an individual (region) specific effect (Ravirajan & Shanmugam, 2021). This equation can be estimated using either the fixed effects (FE) or the random effects (RE) method. The choice of the estimation method is based on Hausman Statistics. In the FE method, the individual specific αi values can be estimated, and the highest value of αi is α* [=max (αi )], and it is treated as the performance of the Most Efficient Region (MER). The relative efficiency of the ith region can be measured as (−ui ). In ui = α* − αi , and the own revenue efficiency is computed as ORE i = exp∑ the RE method, the individual effect of a region is calculated as αi = (1/T) ε it ; i = 1, 2,…, N. The highest value of αi (i.e., α*) is the performance of the MER, and the relative efficiency is calculated as ui = α* − αi and ORE as in the FE method. The MLE method can also be used alternatively to estimate the equation rit = α + xit ’ β + vit − ui , assuming that u follows either half normal or truncated normal distribution. Later, this time-invariant approach is extended to time-varying efficiency models by Battese and Coelli (1992), Cornwell et al. (1990), Kumbhakar (1990). Greene (1993) and Kalirajan and Shand (1994) provide a comprehensive review of these frontier models. Equalization Transfers Across Nations Different nations have framed different equalization mechanisms. Australia considers both revenue and expenditures to determine the transfers. It basically computes a “standard budget” for each service using an all-State average of expenditures and revenues (Shanmugam & Shanmugam, 2022b). Thus, it uses the average efficiency. Germany uses the average nation-wise per capita tax revenue as the proxy for the expenditure of each sub-national Government, while Switzerland uses population density, mountain zones, productive area, etc., to compute the expenditure needs of cantons. Both of these nations consider the spending needs in fiscal equalization. Canada uses an elaborate (tax-by-tax) RTS approach. For each tax, the representative tax effort is calculated by subtracting the actual revenue base from the standard revenue base.5 Each region’s equalization transfer is computed as its deficiency in the measured tax base relative to the national average, multiplied by a target tax rate (national average tax rate). If all regions chose the target rate, then capacity differences are fully utilized and all regions have the same (per capita) fiscal resources. 5

The U.S. Advisory Commission on intergovernmental relations developed this approach. See Akin (1973) for a review. This approach involves two steps. The first step quantifies the tax base for each tax levied by the sub national Government. The second step applies the average effective tax rates of all States (the ratio between total tax revenue and total value of tax base) on the tax bases of individual States and derives the taxable capacity. Finally, the taxable capacities of individual taxes add up to get the aggregate taxable capacity of a State.

160

K. R. Shanmugam and K. Shanmugam

However, if a region sets its rate above the target rate, its base will decline and its transfers rise in consequence (Bird & Smart, 2002). Denmark and Sweden explicitly design transfers with the assumption that an average national-local tax rate is applied. As per this procedure, regions that levy above-average taxes are not penalized, and regions that levy below-average taxes are not rewarded. Thus, it is an incentive for regions to levy at least average taxes. Refer to Hansjörg & Claire (2008), Ma (1997), and Vaillancourt and Bird (2007) for the important characteristics of fiscal equalization schemes in select countries. Empirical Studies on Measuring Expenditure Needs Both income and transfers are considered as the most significant determinants of public expenditure/spending in the most empirical literature. Many hypotheses are put forward to analyze the effects of them. These are Wagner’s Hypothesis, Veil Hypothesis, Disincentive Effect Hypothesis, the Flypaper Effect Hypothesis, etc. Wagner’s Hypothesis discusses a long-run positive effect of income on public expenditure and income. The Veil hypothesis argues that unconditional transfers (e.g., lump sum transfers) can be utilized by the local Governments either to provide any combination of public goods/services or tax concessions. As a result, they don’t affect relative prices (i.e., no substitution effect). Specifically, the effect of them is the same as the effect of distributing the lump sum transfers directly to local citizens. This means that the effect of one dollar increase in local citizens’ income on local spending is exactly equal to the effect of receiving one dollar of transfers on spending (Bradford and Oates, 1971). Scott (1952) argues that if sub-national governments distribute the transfers in the form of lower taxes, this will crowd out local spending. This is called the Disincentive effect hypothesis. The Flypaper effect hypothesis argues that the unconditional transfers given to a jurisdiction can have a greater stimulatory effect on its spending than the equivalent increase in the median voter’ income. That is, “money sticks where it hits” (Shanmugam & Shanmugam, 2022b). A handful of empirical studies emerged in the literature to study the more general redistributive effects of transfer systems in many countries. A major objective of the majority of these studies is to examine the equalization capacity of existing systems and provide alternatives to reduce the horizontal imbalance. Please refer to ACIR (1986 and 1988) for the USA; Martínez-Vázquez and Boex (1999) for the Russian Federation; Ruggeri and Yu (2000) for Canada; Hierrio et al. (2007) for five federal countries—Germany, Australia, Canada, Spain, and Switzerland; Hofman and Guerra (2005) for East Asian countries—China, Indonesia, the Philippines, Thailand, and Vietnam; and OECD (2014) for OECD countries (Shanmugam & Shanmugam, 2022a). In the Indian context, a few studies emerged on the topic. The emprical studies by Lalvani (2002) and Panda (2015) confirm the presence of teh fly-paper effect hypothesis, Panda and Velan (2013) confirm the presence of the incentive effects of fiscal transfers (i.e., crowd-in effect) on the spending of 22 States from 1980–81 to 2004–05. Panda’s (2017) study also shows that the transfers had a crowd-in effect on States’ revenue expenditures, capital disbursements, and aggregate expenditures

7 Equalization Transfers Policy Based on Expenditure Needs and Own …

161

from 1980–81 to 2007–08. However, these studies have not attempted to measure the spending needs of State Governments to determine the equalization transfers. A study by Saraf and Srivastava (2009) calculates the expenditure needs-based equalization transfers only for the health and education sectors using the Australian approach (Shanmugam & Shanmugam, 2022b). Empirical Studies Estimating Revenue Function and Fiscal Capacity: A few studies use the SFA to measure revenue or tax or fiscal efficiency. Pessino and Fenochietto (2010, 2013) employed the SFA to estimate the tax capacity of 96 countries. Cyan, Martinez-Varquez, and Vulovic (2013) compute the tax capacity of 94 countries using the SFA. Alm and Duncan (2014) estimate the tax efficiency of OECD and select non-OECD nations using both DEA and SFA. Jha and Sahni (1997) used Cornwell et al. (1990) time-varying stochastic frontier approach and computed the tax efficiency of Canadian Provincial Governments from 1971 to 1993. The study by Alfirman (2003) measured the tax efficiency of Indonesia’s provincial Governments from 1996 to 1999, using the frontier model for cross-section data. Interestingly, both income and transfers have emerged as significant economic factors determining local revenues in most of the past literature. These studies use income as a proxy for revenue base. They include transfers to test the crowd-out effect of the transfers/grants from upper tier Governments on revenues from local taxes, considering the conceptual foundation of the median-voter model. For instance, Zhuravskaya (2000) shows a crowd-out effect in Russia, Mogues and Benin (2012) in Ghana, and Baretti et al. (2002) in Germany. Some studies do not find a crowding-in effect. For example, Dahlberg et al. (2007) show a crowding-in effect of transfers on local tax revenues after econometrically addressing the potential endogeneity of transfers. Studies such as Skidmore (1999) also show a positive (crowding-in) effect of transfers on local revenues.6 In the Indian context, only a limited number of studies have emerged. For instance, Dash and Raja (2013), Naganathan and Sivagnanam (2000), Panda (2009), and Sarma (1991) show that the transfer effect on States’ own revenues/tax revenues is negative. But these studies analyze only the average effect of transfers on own/tax revenues and not State-specific effects. Further, they bypass the issue of tax effort. While Piancastelli (2001) and Purohit (2006) estimate the tax capacity/effort using an income or RTS or aggregate regression approach, they do not use tax effort to design transfers. Jha (1999) uses the Battese and Coelli (1992) panel frontier approach to calculate the tax efficiency of 15 major Indian States from 1980–81 to 1992–93. Sandhya et al. (2016) use the SFA for panel data and compute the tax capacity of 14 major Indian States from 1991–92 to 2010–11. But they also do not use the fiscal capacity to derive the equalization transfers. 6

Studies such as Allers et al. (2001) and Solé-Ollé (2006) considered political economy factors. Some other studies included natural, social, and demographic factors as determinants of local revenues. For instance, regions enriched with a large volume of natural resources may collect larger revenues in the form of royalties from mining, etc. Regions with greater non-farm economic activities may be able to collect more fees and tax revenues.

162

K. R. Shanmugam and K. Shanmugam

7.3 Empirical Model, Data, and Estimation This study utilizes the Australian Transfer Mechanism (ATM). The methodology involves five steps: I. Estimate the per capita (real) revenue expenditure (EXit ) equation7 : per capita (actual) expenditure cannot be considered because it depends on various known and unobserved factors and, therefore, we consider the estimated expenditure (E X it ) which is conditional on various determinants and this is only the explained part of the regression. The estimated values from the regression model represent how much the sub-national Government i spends, conditional on its spending needs, cost profiles, and revenues at period t. II. Estimate the expenditure needs (ENit ) for each State utilizing the predicted values estimated from the model and standard benchmark. The difference between the benchmark and actual estimated expenditure is the expenditure need; we consider two benchmarks—top three States’ average and all States’ average, (E X it∗ ), i.e., ENit = E X it∗ − (E X it ). Multiplying ENit for each State by its population and GSDP deflator will provide an estimate of the expenditure need in nominal term, E N ∗i t . For States spending more than benchmark, we assign only zero expenditure need value. III. Estimate the per capita own revenue equation using a stochastic frontier method and measure own revenue potential (fiscal capacity) for each State, OR* it . IV. Calculate the excess fiscal capacity EFit , which is the difference between potential revenue and actual revenue of the State; we consider two benchmarks: one based on the top three States’ average own revenue efficiency and the other based on all States’ average efficiency; For States having more than benchmark, we assign only zero excess capacity. V. Fiscal equalization transfer for each State is determined as Tit = EN* it − EF* it . If the difference is negative for any State, the equalization transfer is zero. Δ

Δ

Δ

Δ

Revenue Expenditure Equation Since States spend money on different sectors, it is a difficult task to specify and estimate sector-specific expenditure equations. Following past studies on the topic, this study specifies the linear panel data expenditure model as8 EXit = β0 + β1 · TRit + β2 · ORit + β3 · GSDPit + β4 · NPSit + β5 · PDit + β6 · URit + β7 · ROADit + β8 · PCit·t + λi + μt + μi,t (7.1)

7

Capital expenditure is investment nature; as per FRBM act, States are meeting capital expenses using public debt and so we consider only revenue expenditure. 8 Initially, this study considered various functional forms like log-linear, log-log, etc. However, the linear form was chosen due to its suitability for the data. Past studies by Lalvani (2002) and Panda (2017) also employed the linear model.

7 Equalization Transfers Policy Based on Expenditure Needs and Own …

163

In (1), EXit is the revenue expenditure of the ith State in year t; TRit represents the central transfers, ORit is the own revenues, GSDPit is income, and NPSit is the share of the non-primary sector in total GSDP. These are economic variables and all are annual figures in real and per capita terms. PDit is population density, URit is the degree of urbanization, ROADit is the length of road length, and PCit is per capita power consumption. These are basically infrastructure variables. λi captures the State-specific unobserved heterogeneity effect, and μt is the time effect; uit is the regular disturbance term.9 ,10 The NPS is expected to have a negative effect on revenue expenditure. The ROAD can have a similar effect. However, population density and urbanization may have a negative or positive effect on State expenditures. Power consumption may represent prosperity and it may have a positive effect. But if power is constrained, then the State needs to spend more to procure from others for uninterrupted supply. In that case, power consumption may have a negative effect. Equation (7.1) can be estimated using either the fixed effects or the random effects model estimation method. The choice of the appropriate model is based on Hausman Statistics. Separate models for GCS and SCS are estimated. Further, transfers interaction terms with State dummies are used to check its effect in each State in an alternative specification of the equation. If the estimated β1 is negative (positive) and significant, it will indicate the disincentive (incentive) effect of transfers. The flypaper is present if this coefficient is larger than the income coefficient β3 . If β1 = β3 (i.e., the transfer coefficient is equal to the income coefficient), then it confirms/ supports the Veil hypothesis. The Wagner law can be held good if β3 is positive and significant (Shanmugam & Shanmugam, 2022a). Own Revenue Equation: This study utilizes the following stochastic frontier own revenue equation for panel data to analyze the effect of transfers on own revenues and estimate the fiscal or revenue capacity for 29 Indian States: Ln O Rit = αi + γ1 + Ln T Rit + γ2 Ln G S D Pit + γ3 Ln N P Sit + γ4 Ln U Rit + γ5 Ln L I Tit + γ6 Ln PCit + γ7 T R E N D + vi,t (7.2) In (7.2), OR, TR, GSDP, NPS, UR, and PC are as explained above. LITit is literacy rate. All are in log form. TREND is a year’s trend. vit is the stochastic error term. The αi is a State-specific effect, and Eq. (7.2) can be estimated using the standard 9

It is noted that the Central transfers variable is included as an independent variable in (1). This may create a potential endogeneity problem. In the preliminary investigation, it is found that the results do not change even after excluding this variable. This means that there is no endogeneity issue and even if there exists endogeneity, the bias may be negligible. 10 One may argue that many political economy variables may play a role in determining State expenditures; one can capture their effects by including dummies for election year, ruling party, etc. Since the term λi in (1) can also capture these individual-specific factors, this study does not include the political factors directly in order to avoid perfect multi-collinearity problem.

164

K. R. Shanmugam and K. Shanmugam

panel FE method or RE method based on the Hausman statistics. From the estimated results of the model, the time-invariant own revenue efficiency for each State can be computed. As GCS and SCS differ in their characteristics, we have done a separate (split sample) analysis for the GCS and SCS. In alternative specifications of the model, the fiscal transfer terms interacting with State dummies are included in order to examine the effect of fiscal transfers on own revenues of each individual State. Data The data for the study come from various secondary sources. The per capita GSDP (in 2011–12 prices) and the share of the non-primary sector in total GSDP for 29 States from 2005–06 to 2018–19 are compiled by the National Account Statistics, MOSPI, and EPW Research Foundation. All fiscal variables—revenue expenditures, transfers to each State and own revenues are collected from the Comptroller and Auditor General (CAG) of India Audit Reports and the Finance Accounts of the State Governments. Using the GSDP deflator and population of the respective States, these fiscal variables are converted into real and per capita terms. The population density and literacy data from Census 2001 and 2011 are used to extrapolate their values for the remaining periods of the study. The projected urban ratios are compiled from the Office of the Registrar General and Census Commissioner (2006) till 2010–11 and from the National Commission on Population (2019) after 2010–11. The data source for the road length and per capita power consumption is the RBI’s Handbook of Statistics on the State Economy. The data used is a balanced panel data with (29 × 14 = ) 406 observations. The list of General Category States (GCSs) and the Special Category States (SCSs) covered in the sample is shown in Table 7.3.11 Table 7.1 presents the means and standard deviations of the study variables. It is observed that mean values/standard deviations of almost all variables are different for GCS and SCSs, indicating that GCSs and SCSs have different characteristics and they require separate treatment in determining equalization transfers, i.e., the common benchmark for GCSs and SCSs cannot serve the purpose.

7.4 Empirical Results 7.4.1 Estimation Results of State Governments’ Revenue Expenditures Table 7.2 presents the two-way random effects (RE) estimation results of real per capita revenue expenditure Eq. (7.1) for General Category States and one-way fixed effects (FE) results for Special Category States. The transfers variable has a positive and significant effect at a 1% level in both GCS and SCS, indicating that revenue 11

It is noted that while all other SCSs are smaller and hilly States, Goa is small but not hilly. The GCSs are major Indian States.

7 Equalization Transfers Policy Based on Expenditure Needs and Own …

165

Table 7.1 Means and standard deviations of the study variables Variables

General category states

Special category states

Mean

S.D

Mean

6,375.82

2,985.82

7,026.92

7,746.26

EX—Real Per Capita Revenue Expenditure (|) 10,541.29 4,215.56

23,835.4

11,717.81

19,810.7

12,213.79

OR—Real Per Capita Own Revenues (|)

S.D

TR—Real Per Capita Transfers (|)

3,957.67

1,887.79

GSDP—Real Per Capita Income in |

82,258

39,207.13 93,324.94 69,174.83

NPS—Non-Primary Sector Share in GSDP (%) 77.05

7.53

75.67

9.10

LIT—Literacy Rate (%)

74.96

8.05

80.53

9.15

UR—Urban Ratio (%)

32.56

11.3

29.53

15.68

ROAD—Road Length (kms.)

1,98,705.0 1,25,496.2 48,803.2

77,548.1

PD—Population Density

488.11

294.05

177.02

134.81

PC—Per Capita Power Consumption (kwh)

1,069.31

514.11

726.88

577.42

Sample Size (N)

252

154

Sources Authors’ own calculation using the basic data explained above

expenditure increases with higher transfers. That is, this result confirms a strong incentive (or crowd-in) effect of transfers on revenue expenditures of Indian States. The per capita income also has a positive and significant impact on both GCS and SCS. This result strongly supports the Wagner hypothesis. Interestingly, the parameter associated with transfers is larger than the GSDP parameter, thereby confirming the existence of the flypaper effect in both GCSs and SCSs. Thus, the impacts of these two economic variables are as expected. States may also distribute the transfers as lower taxes. In such a case, the transfers affect negatively the own revenue effort and this in turn will affect negatively the expenditure level, i.e., transfers will crowd out local spending (Scott, 1952). But in Table 7.2, the per capita own revenue has a positive and significant effect on the per capita revenue expenditure in GCSs and SCSs, implying that States with more fiscal capacity incur more revenue expenditures. Population density and urban ratio are positively and significantly associated with real per capita revenue expenditures of both GCSs and SCSs, while NPS and ROAD are negatively and significantly related to the per capita revenue expenditure of GCSs. However, the share of the non-primary sector has a positive and significant effect in the case of SCSs. The per capita power consumption also has a positive and significant effect on the per capita revenue expenditures of SCSs. Table 7.3 shows the results of the expenditure equations of GCSs and SCSs in which transfers terms are allowed to interact with State dummies. These interaction terms are positive and statistically significant in all GCSs, except Bihar, Gujarat, Maharashtra, Uttar Pradesh, and West Bengal. They are also positive and significant in all SCSs, except Goa, Meghalaya, Sikkim, and Tripura. The effects of other variables are more or less the same for GCSs and SCSs as in Table 7.2, except that road length

166

K. R. Shanmugam and K. Shanmugam

Table 7.2 Panel model estimation results (Dependent variable: real per capita revenue expenditures of Indian States from 2005–06 to 2018–19) Variables

Constant

General category states

Special category states

Two-way RE

One way FE

Coefficient

t-value

Coefficient

2459.01

(2.22)**



t-value

Per capita transfers (TR)

0.88

(11.79)*

0.629

(17.7)*

Per capita own revenues (OR)

0.512

(7.76)*

0.558

(5.69)*

Per capita income (GSDP)

0.047

(9.61)*

0.019

(2.19)**

Non-primary sector share (NPS)

−59.265

(−3.69)*

82.609

(2.27)**

Population density (PD)

1.645

(2.99)*

39.956

(2.68)*

Urban ratio (UR)

52.481

(4.77)*

136.75

(3.44)*

(−2.17)**

Road length (ROAD)

−0.002

Per capita power consumption (PC)



2.889

R square

0.961

0.9963

Hausman statistics

23.83

20.73

State dummies

Included

Included

Time effect

Included

Not Included

– (2.15)**

Source Authors’ own estimation

is not significant in the revenue expenditure equation of GCSs, and urban ratio and power consumption become insignificant in SCSs.

7.4.2 Estimation Results of State Governments’ Own Revenues Table 7.4 reports the one-way fixed effects (FE) estimation results of (log) real per capita own revenue Eq. (7.2) for both GCSs and SCSs. The effect of transfers is expected to be either neutral or encourage the own revenue efforts of States. But the coefficient of per capita transfers (γ1 ) is negative and statistically significant at a 5% level for GCSs, indicating a strong disincentive (or crowd-out) effect. This result is consistent with the median voter model hypothesis. However, for SCSs, the coefficient of transfers is positive and significant, indicating that higher per capita transfers from the Center head to higher per capita own revenue of SCSs. That is, there is no adverse effect of transfers on the own revenues of SCSs. It seems that instead of substituting for own revenue, the fiscal transfers are complements to own revenue efforts of SCSs. The parameter of real per capita income is positive and significant at a 5% level in both GCSs and SCSs. The magnitude of this coefficient indicates that a 1% increase in per capita income leads to a 0.28% increase in per capita own revenue of GCSs

7 Equalization Transfers Policy Based on Expenditure Needs and Own …

167

Table 7.3 Panel model estimation results of revenue expenditures for GCSs and SCSs states with transfers interaction Variables

Two-way RE

Variables

Coefficient t ratio

One-way FE Coefficient t-ratio

Constant

2292.936

0.82

Constant

Per capita own revenues

0.456

8.74*

Per capita own revenues

Per capita income

0.026

3.05*

Per capita income

Non-primary sector share

−30.992

Population density



– 0.341

3.40*

0.059

4.72*

−2.22** Non-primary sector share

151.136

3.36*

3.215

6.61*

Population density

139.8

3.94*

Urban ratio

48.552

4.71*

Urban ratio

71.501

1.01

Road length

0.001

1.3

Road length





Per capita power consumption





Per capita power consumption

0.212

0.15

R square

0.988

R square

0.997

General category states

Special category states

Andhra Pradesh *Per capita transfers

0.753

8.41*

Arunachal Pradesh *Per 0.599 capita transfers

13.83*

Bihar *Per capita transfers

−0.349

−1.58

Assam *Per capita transfers

−1.68

−2.71*

Chhattisgarh *Per capita transfers

0.633

7.41*

Goa *Per capita transfers

0.09

0.56

Gujarat *Per capita transfers

0.086

0.56

Himachal Pradesh * Per 0.358 capita transfers

2.73*

Haryana *Per capita transfers

0.896

4.95*

Jammu & Kashmir *Per 0.404 capita transfers

2.01**

Jharkhand *Per capita 0.252 transfers

1.77***

Manipur *Per capita transfers

0.513

2.78*

Karnataka *Per capita 0.401 transfers

3.44*

Meghalaya *Per capita transfers

0.216

1.12

Kerala *Per capita transfers

4.80*

Mizoram *Per capita transfers

0.579

10.44*

Madhya Pradesh *Per 0.297 capita transfers

2.26**

Nagaland *Per capita transfers

0.869

6.23*

Maharashtra *Per capita transfers

0.064

0.33

Sikkim *Per capita transfers

0.174

1.51

Orissa *Per capita transfers

0.475

5.67*

Tripura *Per capita transfers

0.135

0.49

0.646

(continued)

168

K. R. Shanmugam and K. Shanmugam

Table 7.3 (continued) Variables

Two-way RE

Variables

One-way FE

Coefficient t ratio Punjab *Per capita transfers

0.739

5.20*

Rajasthan *Per capita 0.701 transfers

5.90*

Tamil Nadu *Per capita transfers

0.349

2.58*

Telangana *Per capita 0.472 transfers

4.06*

Coefficient t-ratio

Uttar Pradesh *Per capita transfers

−0.243

−1.22

Uttarakhand *Per capita transfers

0.788

11.83*

West Bengal *Per capita transfers

−0.149

−0.8

State effect

Included

State effect

Included

Year effect

Included

Year effect

Not included

Source Authors’ own estimation Table 7.4 One-way FE model estimation results of stochastic frontier own revenue functions for GCSs and SCSs (2005–06 to 2018–19) Variables

General category states

Special category states

Coefficient

Coefficient

t-value

t-value

Constant

−1582

(−0.800)

−1.914

(−1.75)

Ln per capita transfers (TR)

−0.109

(−2.37)

0.241

(2.55)

Ln per capita income (GSDP)

0.279

(2.49)

0.345

(3.14)

Ln non-primary sector share (NPS)



Ln urban ratio (UR)

0.293

(4.20)



Ln literacy rate (LIT)

1.459

(4.96)

0.982



Ln per capita power consumption (PC)

0.125

(1.96)



Trend

0.02

(2.10)



R square

0.9786

0.9563

Hausman statistics

19.26

127.19

State (n−1) Dummies

Included

Included

Source Authors’ own estimation

(2.41)

7 Equalization Transfers Policy Based on Expenditure Needs and Own …

169

and a 0.35% increase in per capita own revenue of SCSs. As expected, the controlling variables, urban ratio, literacy rate, and per capita power consumption are significant and positively associated with the own revenue of GCSs. Interestingly, the trend coefficient is positive and statistically significant at a 5% level, indicating that the real per capita own revenue of GCSs on average grew at about 2% per annum. In the case of SCSs, the literacy rate has a positive and significant coefficient. Table 7.5 presents the estimation results of the alternative specifications of own revenue equations for GCSs and SCSs. These equations allow the transfers terms to interact with State dummies. In 7 States—Chhattisgarh, Kerala, Maharashtra, Odisha, Rajasthan, Telangana, and Uttarakhand, the effect of transfers on own revenue is positive, but significant at 10% level only in Kerala, Maharashtra, and Telangana. In Andhra Pradesh, Bihar, Gujarat, Haryana, Karnataka, Punjab, and Uttar Pradesh, the effect is negative and significant at 5% or 10% level. In Jharkhand, Madhya Pradesh, Tamil Nadu, and West Bengal, the transfers’ effect is negative, but not significant even at the 10% level. The effects of other variables are almost the same as in Table 7.4 except that the power consumption and trend variables become insignificant. The transfers had a crowd-in (positive and significant) effect on own revenue efforts in Jammu & Kashmir, Manipur, Meghalaya, Mizoram, Nagaland, and Tripura. In the remaining 5 SCSs, it had no impact.

7.4.3 State-Wise Efficiency Scores Table 7.6 shows the State-wise efficiency scores using the coefficients of State dummies (effects) from Table 7.4 (not shown). The overall mean own revenue efficiency score for GCSs is 74.67%, indicating that on average the GCSs approximately utilized only about 75% of their own revenue potential during the study period. Therefore, it could be possible for them to raise their existing (own) revenues by 25% more with existing resource bases. The efficiency scores varied widely from 42.18% (in West Bengal) to 100% (in Andhra Pradesh). In the case of SCSs, the average own revenue efficiency score is only 29.19%, indicating that there is a greater possibility for them to improve their own revenues. Goa is the most efficient among the SCSs.

7.4.4 Determining Fiscal Equalization Transfers In order to determine the revenue expenditure needs of each of the GCSs and SCSs, the predicted values of real per capita revenue expenditure of these States from the estimated results given in Table 7.2 are used. We use two alternate benchmarks: the average of the top 3 States (benchmark 1) and the average of all States (benchmark 2). We use separate benchmarks for GCSs and SCSs. As explained in Sect. 7.3, the expenditure need is calculated as the difference between the benchmark expenditure

170

K. R. Shanmugam and K. Shanmugam

Table 7.5 Panel model estimation results for own revenue effort equations for GCSs and SCSs with transfer interaction Variables

Coefficient

t ratio

Variables

Coefficient

t ratio

Constant

−9.595

−2.78

Constant

−2.106

−1.46

Ln per capita income (GSDP)

0.383

2.72

Ln per capita income (GSDP)

0.284

1.64

Ln non-primary sector share (NPS)





Ln non-primary sector share





Ln urban ratio (UR)

0.27

3.09

Ln urban ratio (UR)





Ln literacy rate (LT)

3.144

4.57

Ln literacy rate (LT)

0.745

1.08

Ln per capita power consumption (PC)

−0.003

−0.05

Ln per capita power consumption (PC)





Trend

−0.001

−0.10

Trend





R square

0.9999

R square

0.9997

Hausman statistics

136.76

Hausman statistics

62.75

Special category states

General category states Andhra Pradesh *Ln per capita transfers

−0.27

-4.90

Arunachal Pradesh *Ln per capita transfers

−0.224

−1.10

Bihar *Ln per capita transfers

−0.493

−2.49

Assam *Ln per capita transfers

0.328

1.59

Chhattisgarh *Ln per capita transfers

0.064

0.93

Goa *Ln per capita transfers

0.173

1.63

Gujarat *Ln per capita transfers

−0.144

−1.94

Himachal Pradesh * Ln per capita transfers

0.255

1.33

Haryana * Ln per capita transfers

−0.382

-4.78

Jammu and Kashmir * Ln per capita transfers

0.712

2.19

Jharkhand *Ln per capita transfers

−0.137

−1.39

Manipur *Ln per capita transfers

0.612

2.06

Karnataka *Ln per capita transfers

−0.199

−2.77

Meghalaya *Ln per capita 0.577 transfers

2.75

1.91

Mizoram *Ln per capita transfers

0.784

3.16

Kerala *LN per capita 0.185 transfers Madhya Pradesh *Ln per capita transfers

−0.001

−0.02

Nagaland *Ln per capita transfers

0.629

2.54

Maharashtra *Ln per capita transfers

0.15

1.83

Sikkim *Ln per capita transfers

−0.098

−0.30

Orissa *Ln per capita transfers

0.062

0.83

Tripura *Ln per capita transfers

0.97

3.45

Punjab *Ln per capita −0.241 transfers

−3.50

Rajasthan *Ln per capita transfers

0.38

0.03

(continued)

7 Equalization Transfers Policy Based on Expenditure Needs and Own …

171

Table 7.5 (continued) Variables

Coefficient

t ratio

Tamil Nadu *Ln per capita transfers

−0.055

−0.67

Telangana *Ln per capita transfers

0.313

3.24

Uttar Pradesh *Ln per −0.205 capita transfers

−1.95

Uttarakhand *Ln per capita transfers

0.159

1.40

West Bengal *Ln per capita transfers

−0.013

−0.17

Variables

Coefficient

t ratio

Source Authors’ own estimation Table 7.6 State-wise own revenue efficiency scores GCSs

Own rev. efficiency

SCSs

Own rev. efficiency

Andhra Pradesh

100.00 (1)

Arnica Pradesh

26.63(5)

Bihar

49.32 (17)

Assam

25.62(6)

Chhattisgarh

87.38 (4)

Goa

100.00(1)

Gujarat

68.99 (13)

Himachal Pradesh

36.33(2)

Haryana

92.74 (3)

Jammu & Kashmir

31.81(4)

Jharkhand

55.92 (16)

Manipur

10.89(10)

Karnataka

86.51 (5)

Meghalaya

19.42(7)

Kerala

66.81 (14)

Mizoram

12.81(8)

Madhya Pradesh

69.86 (12)

Nagaland

10.66(11)

Maharashtra

70.02 (11)

Sikkim

34.19(3)

Orissa

79.32 (8)

Tripura

12.69(9)

Punjab

80.27 (7)

Rajasthan

85.74 (6)

Tamil Nadu

72.89 (10)

Telangana

97.71 (2)

Uttar Pradesh

64.39 (15)

Uttarakhand

73.29 (9)

West Bengal

42.89 (18)

Mean TE%

74.67

Mean TE%

29.19

Source Authors’ own estimation

172

K. R. Shanmugam and K. Shanmugam

and the predicted value of expenditure of the State i. The estimated expenditure needs for GCSs and SCSs in 2018–19 are shown in Table 7.7.12 The average of the top 3 States (i.e., benchmark 1) for GCSs is computed using the estimated revenue expenditures of Uttarakhand, Kerala, and Haryana. A State is considered as a zero-need State if its expenditure exceeds the benchmark (i.e., if it gets negative expenditure needs value). Kerala and Uttarakhand have negative expenditure needs; these States are assigned zero when we use benchmark 1. Table 7.7 shows that Haryana’s required revenue expenditure need is the minimum. Uttar Pradesh has the highest revenue expenditure need |.3996.5 billion, followed by Bihar and West Bengal. With the average benchmark of 2, only 8 GCSs are identified to have revenue expenditure needs. The total revenue expenditure need of GCSs is calculated at |.13,146.6 billion based on benchmark 1 and |.5,295 billion based on benchmark 2. The estimated total revenue expenditure needs for all SCSs is |.2,995.9 billion using benchmark 1 and |.1,249.7 billion using benchmark 2. The overall expenditure need for all 29 States is estimated at |.16,142.6 billion (with benchmark 1) and |.6,544.7 billion (with benchmark 2). Then we calculate the excess fiscal capacity (EFit ), which is the difference between potential revenue (calculated using the efficiency scores given in Table 7.6) and actual revenue of the State; we consider two benchmarks: one based on the top three States’ average own revenue efficiency and other based on all States’ average efficiency; for States having more than benchmark, we assign only zero excess capacity. With benchmark 1 (top 3 average), West Bengal’s excess fiscal capacity, |.809.6 billion is the highest. West Bengal needs to generate this amount additionally. Maharashtra and Uttar Pradesh rank second and third in terms of excess fiscal capacity to be exploited. Assam and Jammu & Kashmir have larger excess fiscal capacity among SCSs. With benchmark 2, West Bengal, Uttar Pradesh, Bihar, and Maharashtra are the top 4 States with larger fiscal capacity to be exploited. Assam, Tripura, and Manipur have a larger fiscal capacity to be exploited. The total estimated excess fiscal capacity of all GCSs is about |.4,632 billion based on benchmark 1 and |.1,327 billion based on benchmark 2. Similarly, the total estimated excess fiscal capacity of all SCSs is about |.750 billion based on benchmark 1 and |.129 billion based on benchmark 2. The overall excess fiscal capacity for all States is estimated at |.5,383 billion (benchmark 1) and |.1,456 billion (benchmark 2). Fiscal equalization transfer for a State i is Tit = EN* it − EF* it . If the difference is negative for any State, the equalization transfer is zero. We consider four alternative scenarios by considering (i) revenue expenditure needs and excess fiscal capacity, both based on the top 3 States’ average benchmarks (Scenario 1); (ii) revenue expenditure needs based on average benchmark and excess revenue capacity calculated using top 3 States’ average benchmark (Scenario 2); (iii) average revenue expenditure needs and excess fiscal capacity estimated using the average benchmark (Scenario

12

Results for only the latest year 2018–19 are reported due to space constraint. We find that results using Tables 7.2 and 7.3 are also almost the same as these results.

0

0

21,373

35,690

Chhattisgarh

Gujarat

215,697

0

399,649

0

142,488

Uttar Pradesh

Uttarakhand

West Bengal

61,340

0

0

30,585

15,584

25,677

Tamil Nadu

85,159

Rajasthan

0

1243

0

56,515

Telangana

32,650

15,312

Odisha

Punjab

59,877

Maharashtra

0

0

127,444

Kerala

Madhya Pradesh

0

25,940

Karnataka

26,010

6517

53,484

Haryana

Jharkhand

153

0

142,871

26,227

236,680

Andhra Pradesh

Bihar

0

80,959

4975

75,652

0

39,304

9816

8072

9837

77,796

24,225

28,038

12,343

16,828

2222

37,721

3146

32,299

47,709

292

23,981

0

2922

0

0

0

13,497

4322

7343

0

7715

0

7698

0

17,238

0

Average

Top 3 Avg

Top 3 Avg

Average

Excess revenue (EF*)

Rev. exp. need (EN*)

General category states

States

61,529

0

323,997

15,584

0

75,342

7240

22,813

0

103,219

0

13,597

36,656

4294

0

18,227

204,381

26,227

0

0

140,045

0

0

15,861

0

0

0

32,289

0

0

9181

0

0

0

110,572

0

Top 3 EN*, Top 3 EF* Avg. EN*, Top 3 EF*

Equalization transfers

13,631

0

191,716

0

0

25,677

0

1243

0

52,193

0

0

18,295

0

0

153

125,633

0

Avg. EN*, Avg. EF*

94,779

0 (continued)

375,668

15,584

27,663

85,159

15,312

32,650

46,380

123,122

0

25,940

45,769

6517

27,992

21,373

219,442

26,227

Top 3 EN*, Avg. EF*

Table 7.7 Equalization Transfers for Indian States with Normatively Determined Revenue Expenditure Needs and Own Revenue Fiscal Capacity in 2018–19 (|. Crore)

7 Equalization Transfers Policy Based on Expenditure Needs and Own … 173

124,965

299,598

1,614,257

All SCS Total

All states

Source Authors’ own estimation

6562

16,552

Tripura

654,470

558

0

6609

129

Nagaland

0

6281

Sikkim

685

Mizoram

6889

15,951

15,101

Manipur

Meghalaya

15,449

51,720

Jammu and Kashmir

0

1866

0

19,586

Goa

Himachal Pradesh

87,359

0

173,265

Arunachal Pradesh

0

529,505

1,314,659

538,262

75,029

7438

1027

4772

4044

4280

5114

11,154

5873

0

29,424

1902

463,233

Top 3 Avg

Average

Top 3 Avg

145,590

12,873

2779

0

1914

1504

1117

2036

0

0

0

3361

161

132,717

Average

Excess revenue (EF*)

Rev. exp. need (EN*)

Assam

Special category states

All GCS Total

States

Table 7.7 (continued) Equalization transfers

1,143,836

230,729

9114

0

1837

0

10,821

10,837

40,566

13,713

0

143,841

0

913,107

373,956

66,007

0

0

0

0

2002

1775

4295

0

0

57,935

0

307,949

Top 3 EN*, Top 3 EF* Avg. EN*, Top 3 EF*

543,654

115,114

3783

0

0

0

5164

4852

15,449

1866

0

83,999

0

428,540

Avg. EN*, Avg. EF*

1,477,282

287,705

13,773

129

4695

0

13,984

13,915

51,720

19,586

0

169,905

0

1,189,577

Top 3 EN*, Avg. EF*

174 K. R. Shanmugam and K. Shanmugam

7 Equalization Transfers Policy Based on Expenditure Needs and Own …

175

3); and (iv) revenue expenditure needs using the top 3 average benchmark and excess fiscal capacity estimated using the average benchmark (Scenario 4). In Scenario 1, the total additional equalization transfers required for all 29 States is |.11,438 billion (6.23% of the aggregate GSDP of 29 States). In Scenario 2, it is estimated at |.3,740 billion (2.03% of GSDP), and in Scenario 3, it is computed at |.5,436 billion (2.96% of GSDP). In Scenario 4, the total additional transfers required for all 29 States is calculated at |.14,773 billion (8.04% of GSDP). It is noted that since Scenario 3 calculates the total equalization transfers considering excess fiscal capacity and revenue expenditure needs both based on average benchmarks, the result under this scenario is consistent with the Australian approach which also equalizes with respect to average benchmarks (Shanmugam & Shanmugam, 2022b). To start with, the Center can consider Scenario 3. Under this scenario, 14 out of 29 States would get additional transfers. Over the years, the Center may target to reach the best Scenario 1.

7.5 Summary and Policy Conclusion In this study, an attempt has been made to determine the equalization transfers from the Center to State Governments in India with normatively determined expenditure needs and fiscal own revenue capacities of States. This procedure is consistent with both efficiency and equity. Specifying the per capita revenue expenditure equation, this study has estimated the revenue expenditure needs of 29 Indian State Governments from 2005–06 to 2018–19 empirically. Specifying the stochastic own revenue function, this study has also estimated the own revenue potentials and excess fiscal capacity of the State Governments. It has considered two benchmarks (top 3 average and all States’ average) to compute excess fiscal capacity and revenue expenditure needs and derive equalization fiscal transfers so that each State can provide the same level of services to citizens in the country. As fiscal attributes have varied among small and hilly states (SCSs) and larger or general category states (GCSs), this study has considered a separate benchmark for them. The empirical results strongly support the incentive or crowding-in effect of fiscal transfers on revenue expenditures of the Indian States. They also confirm the validity of the Wagner hypothesis and the flypaper effect hypothesis for both GCSs and SCSs. The effect of transfers is positive and significant in all GCSs, except Bihar, Gujarat, Maharashtra, Uttar Pradesh, and West Bengal. It is also positive and significant in all SCSs except Goa, Meghalaya, Sikkim, and Tripura. The results of the study also indicate a strong disincentive or the crowding-out effect of transfers on the own revenue effort of GCSs and a strong incentive or crowding-in effect on the own revenue effort of SCSs. In Kerala, Maharashtra, Telangana, Jammu & Kashmir, Manipur, Meghalaya, Mizoram, Nagaland, and Tripura, the fiscal transfers significantly and positively contribute to the own revenue effort. In Andhra Pradesh, Bihar, Gujarat, Haryana, Karnataka, Punjab, and Uttar Pradesh,

176

K. R. Shanmugam and K. Shanmugam

the transfers effect is negative. Own revenue efficiency scores of 10 GCSs and 7 SCSs are below the mean efficiency levels of the respective groups. The analyses show that with the normatively determined revenue expenditure needs and fiscal capacity (own revenue effort), the aggregate equalization transfers amount for all 29 Indian States was about |.11,438 billion (6.23% of GDP) in Scenario 1, |.3740 billion (2.03%) in Scenario 2, |.5,436 billion (2.96%) in Scenario 3, and |.1,4773 billion (8.01%) in Scenario 4. Given the fact that in 2018–19, the actual gross revenue receipt (GRR) of the Central Government was |.25,679 billion, it seems to be possible for the Center to fully or mostly equalize the transfers.13 To begin with, the Center can consider Scenario 3 (which is consistent with the Australian approach which equalizes with respect to average benchmarks) and target to achieve Scenario 1 over the years. Thus, the results of the study indicate the relevance of the First-Generation Theorem, which strongly favous the equalization transfers from the Center to States in the federal system. The advantage of this approach is that it provides a single measure of transfers instead of the existing complicated approach (or criteria) used by FCs. It is simple to implement. The benchmark levels can be altered based on fund availability. It takes into account the current (actual) population, the same level of public services to all citizens. It also provides an incentive for the States to raise their own revenues efficiently. As States will get additional transfers, the Center can make it conditional that States should exploit their full fiscal capacity. Nevertheless, the study is not free from limitations. First, it considers only revenue equalization and does not consider capital expenditure. In the Australian model, the capital expenditure needs are supplemented by an elaborate framework of distribution of loans for the States (Shanmugam & Shanmugam, 2022a). This study ignores this and also the deficit financing. Second, the endogeneity effect of transfers is ignored. However, as stated in footnote 9, the endogeneity bias could be minimum. Last, the results of this study are sensitive to the benchmarks used. Despite these limitations, we hope that policy makers and other stakeholders may find useful the results of this study to make appropriate strategies to determine effectively the equalization transfers policy to Indian States so that all citizens living in different regions can avail a comparable (standard) level of public services.

References ACIR. (1986). Measuring state fiscal capacity: Alternative methods and their uses. Advisory Commission on Intergovernmental Relations. Information Reports, n.M-150. ACIR. (1988). State fiscal capacity and effort. Advisory Commission on Intergovernmental Relations, Information Reports, n. M-170. 13

Evidences indicate that India’s tax GDP ratio is low. As India is moving toward achieving the status of developed nations, there is a need to raise our tax-GDP ratio such that the Centre can equalize the transfers.

7 Equalization Transfers Policy Based on Expenditure Needs and Own …

177

Aigner, D. J., Lovell, C. A. K., & Schmidt, P. (1977). Formulation and estimation of stochastic frontier production function models. Journal of Econometrics, 6, 21–37. Akin, J. S. (1973). Fiscal capacity and the estimation method of advisory commission on intergovernmental relations. National Tax Journal, 26(2). Allers, M., de Haan, J., & Sterks, C. (2001). Partisan influence on the local tax burden in the Netherlands. Public Choice, 106(3–4), 351–363. Alfirman, L. (2003). Estimating stochastic frontier tax potential: Can Indonesian local governments increase tax revenues under decentralization. Working Paper No. 03–19, CFEA, University of Colorado at Boulder. Alm, J., & Duncan, D. (2014). Estimating tax agency efficiency. Public Budgeting and Finance, 34, 92–110. Bahl, R., Martinez-Vazquez, J., & Sjoquist, D. L. (1992). Central city-suburban fiscal disparities. Public Finance Quarterly, 20, 420–432. Baretti, C., Huber, B., & Lichtblau, K. (2002). A tax-on-tax revenue: The incentive effects of equalizing transfers: Evidence from Germany. International Tax and Public Finance, 9, 631–649. Battese, E. C., & Coelli, T. J. (1992). Frontier production functions, technical efficiency and panel data: With application to paddy farmers in India. Journal of Productivity, 3, 153–169. Blair, P. (1992). Financial equalization between local and regional authorities in European countries. Local Government Studies, 18, 7–27. Bird, R. M., & Smart, M. (2002). Intergovernmental fiscal transfers: International lessons for developing countries. World Development, 30(6), 899–912. Boadway, R. (1980). Intergovernmental fiscal transfers in Canada, Canadian Tax Foundation, Toronto, Canada. Boadway, R. (2004). The theory and practice of equalization. Cesifo Economic Studies, 50, 211–254. Boadway, R., & Flatters, F. (1991). Federal-provincial relations revisited: Consequences of recent constitutional and policy developments. In McMillan (Ed.), Provincial public finances: Plaudits, problems, prospects (pp. 87–121). Toronto: Canadian Tax Foundation. Boadway, R., Sandra, R., & Shah, A. (1993). The reform of fiscal systems in developing countries: A federalism perspective. Paper presented at the conference on fiscal reform and structural change, New Delhi. Boex, J., & Martinez-Vazquez, J. (2004). Designing intergovernmental equalization transfers with imperfect data: concepts, practices, and lessons. Working Paper No. 04–21, Georgia State University. Bradford, D. F., & Oates, W. E. (1971). The analysis of revenue sharing in a new approach to collective fiscal decisions. Quarterly Journal of Economics, 85, 434–448. Buchanan, J. M. (1950). Federalism and fiscal equity. American Economic Review, 40, 583–599. Buchanan, J. M. (1952). Federal grants and resource allocation. Journal of Political Economy, 60, 208–217. Cornwell, C., Schmidt, P., & Sickles, R. (1990). Production frontiers with cross sectional and time series variation in efficiency levels. Journal of Econometrics, 46, 185–200. Courchene, T. J. (1978). Avenues of adjustment: The transfer system and regional disparities. In M. Walker (Ed.), Canadian confederation at the crossroads: The search for federal-provincial balance (pp.145–86). Vancouver: The Fraser Institute. Cyan, M. R., Martinez-Vazquez, J., & Vulovic, V. (2013). Measuring tax effort: Does the estimation approach matter and should effort be linked to expenditure goals. Working Paper, International Centre for Public Policy, Andrew Young School of Policy Studies, Georgia State University, Atlanta. Dahlberg, M., Mork, E., Rattese, J., & Agren, H. (2007). Using a discontinuous grant rule to identify the effect of grants on local taxes and spending. Journal of Public Economics, 92(12), 2320–2335. Dash, B. B., & Raja, A. V. (2013). Do political determinants affect the size and composition of public expenditure? A study of the Indian states. International Review of Economics, 60(3), 293–317.

178

K. R. Shanmugam and K. Shanmugam

Greene, W. H. (1993). The econometric approach to efficiency analysis. In H. O. Fried & S. S. Schmidt (Eds.), The measurement of productive efficiency: Techniques and applications (pp. 68– 1190). Oxford University Press. Hansjörg, B., & Claire, C. (2008). Fiscal equalization. OECD Economic Studies, 44, 1–22. Hierro, L., Atienza, P., & Patiño, D. (2007). Inequality and progressiveness in the distribution of revenues of the states in federal countries: A comparative study. Working Paper No. 07–03, Georgia State University, Andrew Young School of Policy Studies. Hofman, B. and Guerra, C.S. (2005). Fiscal disparities in east Asia: How large and do they matter?. East Asia decentralizes: Making local government work, Washington, DC: The World Bank. Jha, R. (1999). Tax efficiency in selected Indian states. Empirical Economics, 24, 641–654. Jha, R., & Sahni, B. S. (1997). Tax efficiency and productivity analysis: The case of Canadian fiscal federalism. Public Finance, 52(2), 186–197. Jondrow, J., Lovell, C. A. K., Materov, I. S., & Schmidt, P. (1982). On the estimation of technical inefficiency in the stochastic frontier production model. Journal of Econometrics, 19, 233–238. Kalirajan, K. P., & Shand, R. T. (1994). Economics in disequilibrium: An approach from the frontier. Macmillan India Ltd. Kumbhakar, S. (1990). Production frontiers and panel data and time varying technical inefficiency. Journal of Econometrics, 46, 201–211. Ladd, H. F. (1994). Measuring disparities in the fiscal condition of local governments. In J. E. Anderson (Ed.), Fiscal equalization for state and local government finance (pp. 21–54). Praeger. Ladd, H., & Yinger, J. (1994). The case of equalization aid. National Tax Journal, XLVII, 211–224. Lalvani, M. (2002). Flypaper effect: Evidence from India. Public Budgeting & Finance, 22(3), 67–88. Ma, Y. (1997). Intergovernmental fiscal transfers in nine countries: Lessons for developing countries. Working Paper No. WP 1822, World Bank Policy Research. Maarten, A., & Lewis, I. J. (2011). Equalizing spending needs of sub-national governments in a developing country: The case of Tanzania. Environment and Planning: Government and Policy, 29, 487–501. Martínez-Vázquez, J., & Boex, J. (1999). Fiscal decentralization in the Russian federation during the transition. Working Paper No. 3, Andrew Young School of Policy Studies, Georgia State University. Martínez-Vázquez, J., & Searle, B. (2007). Fiscal equalization: Challenges in the design of intergovernmental transfers. Springer Science & Business Media. Meeusen, W., & Broeck, J. V. D. (1977). Efficiency estimation from Cobb Douglas production function with composed Error. International Economic Review, 18, 435–444. Mogues, T. J., & Benin, S. (2012). Do external grants to district governments discourage own revenue generation? A look at local public finance dynamics in Ghana. World Development, 40(5), 1054–1067. Munoz, A. F., Gustavo, A. R., & Claudia, B. (2016). Sub-national fiscal disparities and intergovernmental transfers in LAC. Review of Public Economics, 219(4), 35–66. Musgrave, R. (1959). Theory of public finance: A study in public economy. McGraw-Hill. Naganathan, M., & Sivagnanam, K. J. (2000). Federal transfers and tax efforts of states in India. The Indian Economic Journal, 47, 101–110. Oates, W. E. (1972). Fiscal federalism. Harcourt Brace Jovanovich Inc. Panda, P. K. (2009). Central fiscal transfers and states own revenue efforts in India: Panel data models. Margin: The Journal of Applied Economic Research, 3(3), 223–242. Panda, P. K., & Velan, N. (2013). Central fiscal transfers and states’ spending in India: An analysis of incentive effect. Economics Bulletin, 33(2), 1229–1246. Panda, P. K. (2015). Fly- paper effect of fiscal transfers in India: A dynamic panel analysis. The Empirical Econometrics and Quantitative Economics Letters, 4(2), 69–79. Panda, P. K. (2017). Budgetary impacts of central fiscal transfers in India: Evidence from state level data, VISION: Journal of Indian Taxation, 4(2), 20–38.

7 Equalization Transfers Policy Based on Expenditure Needs and Own …

179

Piancastelli, M. (2001). Measuring the tax effort of developed and developing countries: Cross country panel data analysis-1985/95. IPEA Working Paper No. 818. Purohit, M. C. (2006). Tax efforts and taxable capacity of central and state governments. Economic and Political Weekly, 41, 747–751. Rangarajan, C., & Srivastava, D. K. (2004). Fiscal transfers in Australia: Review and relevance to India. Economic and Political Weekly, 3709–3722. Ravirajan, K., & Shanmugam, K. R. (2021). Efficiency of commercial banks in India after global financial crises. Theoretical and Applied Economics, 3(628), 65–82. Ridge, M. (1992). Local government finance and equalization: The case of Ireland. Fiscal Studies, 13, 54–73. Ruggeri, G., & Yu, W. (2000). Federal fiscal balances and redistribution in Canada: 1992–1997. Canadian Tax Journal, 48(3), 626–655. Sandhya, G., Goyal, A., & Pal, R. (2016). Why tax effort falls short of tax capacity in Indian states: A stochastic frontier approach. Public Finance Review 1–28. Saraf, R., & Srivastava, D. K. (2009). Determining general and specific purpose transfers: An integrated approach. Working Paper No. 40, Madras School of Economics. Sarma, J. V. M. (1991). Panel data models and measurement of states’ tax effort in India. Journal of Quantitative Economics, 7(1), 145–164. Schmidt, P., & Sickles, R. C. (1984). Production frontiers and panel data. Journal of Business and Economic Statistics, 2, 367–374. Scott, A. D. (1952). Federal grants and resource allocation. Journal of Political Economy, 60, 534–538. Shah, A. (1988). Capitalization and the theory of local public finance: An interpretive essay. Journal of Economic Surveys, 2(3), 209–243. Shah, A. (1994). A fiscal needs approach to equalization transfers in a decentralized federation. Working Paper No.1289, World Bank Policy Research. Shanmugam, K., & Shanmugam, K. R. (2022a). Equalization transfers based on spending needs and fiscal capacity of state governments in India. Margin-the Journal of Applied Economic Research, 16(2), 139–165. Shanmugam, K., & Shanmugam, K. R. (2022b). Equalization transfers based on spending needs and fiscal capacity of state governments in India. MSE Working Paper No. 238. Skidmore, M. (1999). Tax and expenditure limitations and the fiscal relationship between state and local governments. Public Choice, 99(1–2), 77–102. Sole´-Olle´, A. (2006). The effects of party competition on budget outcomes: Empirical evidence from local governments in Spain. Public Choice, 126(1), 145–176 Spahn, P. (2007). Equity and efficiency aspects of interagency transfers in a multi-government framework. In R. Boadway, & A. Shah (Eds.), Intergovernmental fiscal transfers-principles and practice. The World Bank. Tiebout, C. M. (1956). A pure theory of local expenditures. Journal of Political Economy, 64(5), 416–424. Vaillancourt, F., & Bird, R. (2005). Expenditure-based equalization transfers. ITP Paper No.0512. Vaillancourt, F., & Bird, R. (2007). The interregional incidence of central budgets in federation: Some evidence from Canada. Public Budgeting and Finance, 27(1), 1–19. Zhuravskaya, E. V. (2000). Incentives to provide local public goods: Fiscal federalism, Russian style. Journal of Public Economics, 76(3), 337–368.

Part III

Fiscal Reforms

Chapter 8

Goods and Services Tax in India: A Stocktaking Govinda Rao

Abstract The Goods and Services Tax (GST) that was implemented in July 2017 to simplify and harmonize the domestic consumption tax system has taken a firm root and is a reality. However, the potential gains from the reform are yet to be fully realized, and the reform is still a work in progress. In the last five years, the tax has settled, and there are signs of improvement in revenue collections. There have been many gains from the tax, and these include the unification of many domestic trade taxes and harmonized structure, removal of impediments to trade and reduced transaction cost of inter-state transactions, greater formalization of the economy through digitization, getting rid of inter-state tax exportation by converting the tax from origin-based to destination-based, and more efficient supply chain management. However, much more remains to be done to make it “a money machine”, and to lower the collection, compliance, and distortion costs. The paper details the reform agenda for fully realizing the potential gains from the tax in terms of efficiency and revenues. Keywords Tax reform · Goods and services tax · Value added tax JEL Classification H 20 · H 25

The chapter revised version of the author’s working paper and sections are re-used with permission. Govinda Rao, M. (2022). Evolving Issues and Future Directions in GST Reform in India, Working Paper 221/2022. Madras School of Economics, Chennai: India. The author is grateful to Prof. C. Rangarajan for his incisive comments on the earlier draft of the paper. G. Rao (B) Takshashila Institution, Jakkur, Bengaluru, Karnataka 560064, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 D. K. Srivastava and K. R. Shanmugam (eds.), India’s Contemporary Macroeconomic Themes, India Studies in Business and Economics, https://doi.org/10.1007/978-981-99-5728-6_8

183

184

G. Rao

8.1 Introduction I am indeed privileged to write this article in honor of Dr. Rangarajan for the volume to celebrate his 90th birthday. It has been a great honor and privilege for me to interact with him for long, as a Member of the Economic Advisory Council to the Prime Minister of which he was the Chairman, and as the Director of the National Institute of Public Finance and Policy of which he was the Chairman. I have learned a great deal during my interactions with him, and this article is a small tribute to his great scholarship and untiring and relentless passion and enthusiasm to achieve both academic excellence and policy leadership. The implementation of the value-added tax (VAT) is perhaps the most important tax reform seen across countries in the history of consumption tax reform. There are over 166 countries out of 193 with UN membership having the VAT in one form or another. All OECD countries except for the US have implemented the VAT, and the reform has spanned both developed and developing countries (Bird & Gendron, 2007). Of the countries that implemented the VAT so far, only five countries have repealed them but reintroduced them with improvements.1 On midnight of July 1, 2017, India joined the club of countries by implementing a comprehensive VAT on goods and services. It combines 11 different domestic indirect taxes levied at Central, State, and local governments. It is a consumption-type destination-based tax on goods and services with a standard invoice-credit method. The implementation of the goods and services tax (GST) was not easy and took considerable time as it required building consensus among the Union, and 29 States and two Union Territories with the legislature (28 States and 3 Union Territories after converting Jammu & Kashmir into a Union Territory). International experience shows that the implementation of GST takes a considerable time period to stabilize, particularly when the tax is levied at different levels in a federal polity. The only other examples of GST levied at multiple levels are in Brazil, Canada, and the European Union and even after several years of experience, the reform in these countries is still a work in progress. In the case of Brazil, the inter-state trade is subject to the origin based principle and there is no conceptual and administrative clarity in the federal and the State versions of the VAT. Besides the problems with cross-border trade and inter-state tax exportation, it has very high compliance, administration, and distortion costs (Varsano, 2000; Bird & Gendron, 2007). In the European Union, even as the tax is destination-based, by zero-rating exports, the evasion from cross-border trade continues to be a matter of concern (Keen, 2009; Cnossen, 2010). Besides, the structure of the tax is not harmonized with different countries having varying exemptions and tax rates. In Canada as of 2011, six out of the 10 provinces accounting for 80% of the population have imposed some form of VAT or the other. But there are four different systems in Canada.2 1

Bird and Gendron (2007) cite the cases of Belize, Ghana, Grenada, Malta, and Vietnam which repealed the VAT after implementing it, but introduced it in an improved form. 2 For details, see Rao (2022).

8 Goods and Services Tax in India: A Stocktaking

185

The general experience is that the VAT is considered to be a “money machine”.3 However, in India, besides the revenue objective, the reform was motivated by the desire to simplify and unify a number of domestic trade taxes, reduce distortions arising from cascading, and harmonize indirect taxes between the Center and States and among the States inter se. The immediate objective of the reform was to reduce the distortions, minimize impediments to trade across the country, digitize the tax to enhance compliance with the tax, and improve export competitiveness, though in the medium term, it was supposed to increase revenue productivity through better compliance with the tax. By making the tax system destination-based, the reform was intended to minimize inter-state tax exportation as well. However, even before the structure and operation of the tax could stabilize, the devastating effects of the pandemic on the revenue pushed the discussion on the reforms to the back burner. As the economy recovers, the time is opportune to review the performance of the tax so far and identify the measures needed to make it a modern competitive consumption tax. Despite initial setbacks, the GST has taken firmer roots in India by ironing out some shortcomings but, more importantly, adjusting to the complex structure. With the stabilization of the technology platform, the tax can perform well in terms of revenue productivity. This paper attempts to assess the revenue impact of GST including the effect of the pandemic on the revenues and identify further reform areas and the strategy to implement them to make the consumption tax system in the country simpler, less distorting, compliant, and revenue-productive. In many ways, it complements and extends the analysis of the two earlier papers on GST reform (Rao, 2017, 2022 ). The next section briefly lays out the distinct features of the GST in India, Sect. 8.3 analyses the impact of the reform on revenue productivity, and gains in terms of reducing distortions and transaction costs including the freer movement of goods across the country. Section 8.4 identifies the reform areas. The concluding remarks are discussed in Sect. 8.5.

8.2 Salient Features of Indian GST The international experience with VAT implementation of GST shows that there is a “no one size fits all” system. Each country adopts a structure and operational details to suit its political acceptability and convenience (Bird & Gendron, 2007). However, the general principles recommended by most experts are (i) aim for a global tax with few exemptions, credits, rebates, or deductions; (ii) do not use the tax system to achieve too many goals; (iii) keep the threshold at a reasonably high level to focus on the “whales” rather than “minnows”. This will serve to minimize administrative costs and helps to make the tax more equitable (Keen & Mintz, 2004); (iv) avoid multiple rates to evolve a simple structure to minimize administrative, compliance, 3

Keen and Lockwood (2010) test the “money machine” hypothesis in a cross-country analysis and find that the adoption of VAT is associated with an increase in revenues and improved effectiveness in a large majority of countries.

186

G. Rao

and distortion costs; (v) do not collect more information than that is required and can be processed; (vi) actively encourage good record keeping and aim at the long-term goal of self-assessment. In fact, as Bird and Gendron (2007; p. 4) state, “….some ‘bad’ features—such as too high or too low thresholds, overly extensive exemptions, or multiple rates—may be essential to successful adoption in the first place”, but it may not be easy to remove them later.4 India implemented the GST on July 1, 2017, by combining the domestic indirect taxes levied by the Union, State, and Local governments except the State excise duties on alcoholic products, taxes on petroleum products, electricity, and real estate. There are some distinct features in the levy of GST in India. First, the GST combines domestic consumption taxes levied by Central, State, and Local governments. Second, the tax is levied by the Center and States as determined by a joint tax authority—the GST Council. The Council is chaired by the Union Finance Minister, and the Union Revenue Secretary is the Secretary of the Council. The decisions taken in the Council should have at least 75% votes. The Central government has a one-third vote, and the States have the remaining two-thirds, equally shared by them. Third, the collection of the tax is done by providing a seamless input tax credit mechanism. GST comprises a Central GST (CGST), State GST (SGST), and Inter-state GST (IGST). The IGST revenue is credited to a separate account and finally allocated according to the State where the final transaction takes place through a clearing house mechanism. Fourth, all exports are zero-rated. Finally, considering the need for a strong technology platform for managing inter-state transactions, refunds to exporters, and a 100% check on the genuineness of the input tax credit, a specialized agency—the GST Network (GSTN)—was created under the GST Council. The threshold for registration is kept at Rs 40 lakh for goods and Rs. 20 lakhs for services. For the Northeastern and Himalayan States, the threshold has been kept at Rs. 10 lakhs. The administration of the tax is divided between the Center and respective States. All taxpayers below the turnover of Rs. 1.5 Crore are to be administered by the concerned State where they have registered and are responsible for administering the taxpayers above Rs. 1.5 Crore turnover is divided between the Center and the respective States equally based on random selection. The dealers with less than Rs. 1.5 Crore have the choice of paying a simplified tax at a compounded rate of one percent of the turnover in the case of manufacturers and traders, 5% on restaurants, and 4% on service providers without any provision for the input tax credit. The tax is levied at four rate categories equally divided between the Center and the States at 5, 12, 18, and 28% besides exempted goods and services. There is no standard rate as such, but most services are taxed at 18%. A special rate of 0.25% is applicable on precious and semiprecious stones, and gold is taxed at 3%. The items classified as “demerit” and “luxury” items of consumption such as aerated drinks, cars, air conditioners and washing machines, construction materials such as cement, 4

Bird & Gendron (2007) refer to the recommendation of a committee in Sweden to switch over to one rate of tax which was not found to be politically feasible.

8 Goods and Services Tax in India: A Stocktaking

187

paint, marble, and similar goods, and tobacco products are taxed at 28%. In addition, a separate cess is levied on “demerit” goods to compensate the states for any loss of revenue. The compensation to the states was promised to persuade them to join the reform. The Central government agreed to compensate for the shortfall in revenues from the revenues that would have accrued from the merged state and local taxes. The potential revenue was estimated by applying 14% growth annually on the revenue from merged taxes in 2015–16 as certified by the Comptroller and Auditor General for a period of five years. The compensation was to be financed by the cess on “demerit” and luxury items mentioned above at varying rates. The taxpayers are required to electronically file a single return for CGST, SGST, IGST, and the compensation cess. Initially, a fully automated system with 100% matching of invoices for input tax credit was envisaged with taxpayers required to submit three returns GSTR-1, GSTR-2, and GSTR-2 every month and a final annual return at the end of the year. GSTR-1 was required to furnish the details of outward supplies, GSTR-2 was to contain the information on inward supplies, and GSTR-3 was supposed to be auto-populated with information from the two forms. However, the system could not function, and after the repeated postponement and simplified self-assessed summary form providing information and outward and inward supplies were designed as a temporary measure and ITC was settled based on these summary returns. In July 2018, the GST Council announced that by January 1, 2019, a simplified new return will be rolled out, but was postponed repeatedly. In the absence of a proper return, the self-declaration done in GSTR-3B has continued to be the basis for determining the tax liability. A later modification included that while the GST payment has to be monthly, the GSTR-3B could be filed quarterly for taxpayers with less than Rs. 5 Crore turnover. The introduction of GST has led to the abolition of check-posts. However, to keep a vigil on the movement of goods, a web-based e-way bill was introduced, and that must be carried compulsorily on all inter-state supplies by the person carrying the goods exceeding Rs. 50,000 from April 2018. Even for intra-State supplies, the option was given to the States to choose any date on or before June 3, 2018, and all the States have notified e-way bill rules for intra-State supplies. To ensure that the benefits of ITC and any reduction in the rates of GST are passed on to the consumers, a three-tier structure for investigation and adjudication of complaints regarding profiteering at national and state levels and a directorate of anti-profiteering have been set up to address the complaints by the consumers.

188

G. Rao

8.3 Impact of GST on Minimizing Distortions and Revenue Productivity (a) Gains from reform It is easy to identify the important gains from the implementation of GST. The first is the fact that it has successfully unified and simplified several consumption taxes to reduce both administration and compliance costs. Second, it has succeeded in harmonizing domestic trade taxes. The reform has not only minimized the overlap in indirect taxes between the Center and States but has also eliminated tax competition in terms of the race to the bottom among the States by having a uniform rate structure across the country. Although inter-state suppliers must deal with the laws in different States, the laws and rules have been substantially harmonized to reduce the compliance burden. By eliminating the Central sales tax and providing seamless credit for the taxes paid earlier, it has reduced cascading from the tax on inter-state transactions. By making this a destination-based tax, the tax has avoided inequitable inter-state tax exportation. Equally important gain is the formalization of the trade in goods and services as the tax has to be paid online. The most important gain from the tax has been the elimination of check-posts to enable unhindered movement of goods across the country. The abolition of checkposts has helped to reduce transportation time and more importantly, eliminate rentseeking at the check-posts, which was particularly rampant in the States where Octroi was levied. This has paved the way for a nationwide market for goods and services. According to the Ministry of Road Transport, post GST, the long-distance travel time for trucks has been reduced by 20%.5 Another important gain is the cost savings from better supply chain management. Earlier, most businesses created branch offices all over the country and sent their supplies as consignment transfers to avoid the interstate sales tax. With the introduction of GST with a seamless credit mechanism, there are no gains to be had by having branch offices and consignment transfers. The electronic administration envisaged under the GST was to eliminate the interface between the taxpayer and tax collector. Right from registration, payment of the tax by availing ITC, filing of returns, and assessment, the entire process was designed to be electronically managed without any personnel interface. This was also supposed to ensure faster refunds to the exporters. The mechanism was required also to ensure better compliance with the tax as well. A strong technology platform was critical also for ensuring seamless credit on cross-border supplies. However, glitches in the technology platform have constrained the full realization of these benefits, and this is discussed later in the paper. The linking of the GST numbers with the permanent account number (PAN) of the income tax makes it possible to closely monitor both GST and the income tax to enhance voluntary compliance with the tax and increase the revenue productivity of both taxes.

5

See, https://economictimes.indiatimes.com/news/economy/policy/post-gst-travel-time-of-truckshas-reduced-by-a-fifth-government/articleshow/59831749.cms?from=mdr.

8 Goods and Services Tax in India: A Stocktaking

189

Another unique feature of Indian GST is the creation of the GST Council. This is an important innovation in cooperative fiscal federalism where both the Union and the State governments forgo their tax autonomy to achieve a harmonized domestic consumption tax. In a situation in which there is an institutional vacuum for intergovernmental bargaining, promoting coordination and cooperation, regulating competition, and ensuing conflict resolution, the GST Council provides a model for achieving this in the levy of domestic indirect tax. This can be a useful model for adoption in other areas in which Center-state cooperation is important. However, it is necessary to raise a word of caution at this stage as a large number of actors and the attempt to forge unanimity in decisions could result in the tyranny of status quo. (b) Impact on revenues Although the GST was designed to be revenue-neutral, it was expected to generate a significant increase in revenue productivity over the medium and longs term due to improved compliance with the tax. However, the experience in the first four years has not been encouraging and a number of reasons can be attributed to this. The analysis shows that after implementation, there has been a deceleration in the growth of revenues. For 2017–18, the C&AG estimates that the GST revenue was actually lower than the revenue collected from the subsumed taxes in the previous year by almost 10% (India, 2019, p. 28–29). There was a sharp decline in revenues in 2020– 21 due to COVID-19 restrictions. However, the next year saw a “V”-shaped rebound, and in 2022–23, the average mothy revenue recorded a 29% increase. In 2022–23, there was a significant increase in revenues to record an average 21% growth in average monthly collections over the previous year (Fig. 8.1 and Table 8.1). The main reason for the low revenue productivity in the first three years was the inability to stabilize the technology platform. This led to the continuous postponement of the annual returns and the inability of the system to check the correctness of the input tax credit. The result was tax evasion on a massive scale by claiming input tax credits by producing invoices from fictitious firms. Just as the technology showed 200000

150000 100000 50000 0 0

2 2018-19 2022-23 Linear (2021-22)

4

6

2019-10 Linear (2018-19) Linear (2022-23)

8

10

2020-21 Linear (2019-10)

Fig. 8.1 Trends in monthly revenue collections 2019–20 to 2022–23.

12

14

2021-22 Linear (2020-21)

190

G. Rao

Table 8.1 Average monthly collections and growth rates of GST 2018–19 to 2022–23 Year

Average monthly collections (Rs. Growth rate (%) Growth GDP (%) Buoyancy Crore)

2018–19 97,964







2019–20 101,896

4.01

5.10

0.79

2020–21 95,542

−6.24

−1.36

4.57

2021–22 123,715

29.49

19.51

1.51

2022–23 150,587

21.72

15.39

1.41

Source Author’s estimates

signs of stabilizing, the pandemic struck, and with severe restrictions on economic activities, the revenues showed a sharp decline in 2020–21. The quarterly effective rates of GST since its implementation presented in Fig. 8.2 show some important findings. The effective rate is estimated as a percentage of GST revenue in the quarter to the sum of private final consumption expenditure and government consumption expenditure during the quarter. The results show some interesting features. First, the revenue as a ratio of the tax base does not show any clear trend in spite of a significant slowdown in economic activities. Second, the fluctuations clearly reflect the economic impact of the lockdowns and restrictions imposed to contain COVID-19 from time to time. Third, after sharp declines in the effective rate during both the first and the second waves of the pandemic, the rate rebounded to the previous levels in the next quarter during both the first and second waves. Fourth, after the second wave of the pandemic and particularly after the third quarter of 2021–22 the revenues increased sharply to raise the effective rate. Finally, the revenue productivity of the tax has shown a steady improvement in 2022–23 and in each of the three quarters, the effective rate is close to, or higher than, 9%. As the technology platform stablized, the compliance showed a steady improvement, and this is also seen in the fact that the average of monthly collections in 2022–23 is of three quarters is seen to the compliance of the effective rate has been showing a steady increase and this is also seen by the fact that average monthly collections of the tax has risen to more than Rs. 1.5 lakh Crore showing the growth of 21% over the previous year’s. Thus, the trend in revenue collections in successive years since its implementation shows a steady improvement, barring the low collections due to the lockdowns during the first and second waves of the pandemic as shown in Fig. 8.2. The higher collections not only reflect economic recovery after the second wave of the pandemic but also the firming up of the technology platform. With better technology and enforcement, and pent-up consumption demand during the festival season, the revenues have steadily increased to reach Rs. 1.68 lakh Crore in April 2022 and remained higher than Rs. 1.4 trillion thereafter. Even so, revenue productivity has not been high, and it is hoped that with better application of technology and structural reforms, the tax will show significant improvement both in economic efficiency and revenue productivity. These are discussed at a greater length in the following sections.

8 Goods and Services Tax in India: A Stocktaking

191

12.0 10.0 8.0

8.5 8.6 8.5

9.8

8.6 8.6 9.0 8.2 9.1 8.2

7.6

8.9 8.1 8.4 8.0

8.4 8.3

6.0

9.5

8.8 8.6 8.8

9.9 9.5 7.9

8.9 9.4

6.3

4.0 2.0 0.0

Fig. 8.2 Quarterly effective rates of GST (Percent)

(c) Reasons for low revenue productivity The 2019 report of the C&AG made a detailed compliance audit of the technology platform and pointed out a number of shortcomings (India, 2019). It had observed the postponement of the originally envisaged GST returns (GSTR-1,2,3) due to technical glitches and the inability to undertake the originally envisaged universal verification of invoices to match input tax credit (ITC) using the simplified returns (GSTR-3B) as principal reasons for evasion of the tax. The CAG’s report concluded, “….. On the whole, the envisaged GST tax compliance system is non-functional” (India, 2019; p. 22). Further, the settlement of IGST to the States also could not be done properly as the system failed to generate the required modules such as appeals and refunds from the returns. The failure of the technology platform to verify invoices for ITC can lead to false claims and refunds. Similarly, the inability to validate the registrations has led to the creation of several shell companies (some of them within the group) to issue fake invoices which eventually disappear leading to evasion of the tax. The fact that the annual return filing date was repeatedly postponed due to technical glitches for almost two and a half years did not permit the detection of wrong claims on the ITC. In the absence of a clear paper trail, the assessment will be based entirely on trust, and this provides an opportunity for unscrupulous businessmen to evade the tax. In fact, the annual return filing for 2018–19 was repeatedly postponed till December 2020. Similarly, the inability to validate and debar the ineligible taxpayers from availing composition levy has also led to misuse of the option. Based on the information made available from the All India Enforcement Drive by the CBIC through the Directorate of Revenue Intelligence and other CBIC formations from November 2020, it was reported that more than 5700 cases involving an amount of Rs. 40,000 Crore were detected in the fiscal year 2021.6 Indian GST with a complex structure with the simultaneous levy of Central and State level destination-based tax, wide-ranging exemptions, and multiple rates requires a robust technology platform, and the failure to erect that had adverse effects on the compliance of the tax. As the platform gets stabilized, revenue collections show improvements. 6

https://economictimes.indiatimes.com/news/economy/finance/gst-evasion-of-40000-crore-det ected-from-fake-invoicing-itc-laims/articleshow/88603551.cms?utm_source=contentofinterest& utm_medium=text&utm_campaign=cppst.

192

G. Rao

Table 8.2 Turnover range-wise number of taxpayers and tax paid in Karnataka. 2018–19 Turnover range (Rs. million)

Number of taxpayers

Taxable turnoverRs. Million

Tax paid Rs. Million

Percent of taxpayers

Percent of turnover

Percent of tax paid

Percent of tax paid to turnover

10,000

18

21,627,367

187,728

0.00

57.30

21.06

0.87

Total

574,034

37,744,989

891,471

100.00

100.00

100.00

2.36

Source Office of the Commissioner of Commercial Taxes, Government of Karnataka

An important concern in tax administration relates to the need to focus on large taxpayers. The issue must be dealt with in two ways. First, as argued by Keen and Mintz (2004), the threshold should be kept high enough to focus on “whales” rather than minnows. This not only helps the administration to focus on the large taxpayers but also serves the cause of equity as most of the low-income earners buy their requirements from small traders. Bird and Gendron (2007) after reviewing the experiences of different countries suggest a thumb rule of having a threshold of USD 100,000 for developing countries. In India, the threshold was kept at Rs. 20 lakhs when the tax was implemented, but later it was raised to Rs. 40 lakhs for goods, but for services, the threshold remains at Rs. 20 lakhs. The analysis based on the information collected from Karnataka shows that over 92% of the taxpayers are in the turnover range of below Rs. 50 lakhs, and they account for just about 12% of the GST paid (Table 8.2). Considering the low proportion of turnover and the tax paid up to Rs. 50 Lakh turnover, it may be desirable to raise the threshold to Rs. 50 lakhs without making the distinction between goods and services. Should the tax administration undertake 100% matching of invoices? The review of early experience of this in Korea shows that the practice was neither efficient nor effective (Choi, 1990).7 Therefore, the system was modified in 1988 to match invoices above a threshold (approximately £175) and when the discrepancies between input and output invoices or between invoices and VAT returns were higher (approximately £0.2,875), and later, the practice of e-invoicing was adopted. Krever’s (2014) analysis showed that a comprehensive invoice matching system imposed high compliance costs on taxpayers who are least likely to evade VAT and divert administrative

7

Krever, (2014) provides a very useful insight in the Korean experience. See also Bird and Gendron (2007; p. 170–171).

8 Goods and Services Tax in India: A Stocktaking

193

resources from audit processes aimed at uncovering suppression of sales or claims to ITC related to outputs of large taxpayers. Considering the heavy demands on technology, in India, it may be appropriate to review the strategy of 100% matching and confine the practice to those suppliers with a turnover of more than some threshold (say Rs. one Crore, and for invoices more than Rs. 10,000). The GST Council has already introduced an e-invoicing system for B2B transactions for taxpayers with a turnover of more than Rs. 100 Crore and that should help to minimize the evasion through fake invoicing. A detailed audit can be carried out in cases where there are discrepancies unless the technology permits this without casting much administrative and compliance burden. Even if the threshold is not increased, confining the invoice matching to taxpayers with a turnover of more than Rs. 10 Crore will limit the invoice matching requirements to just 8% of the existing taxpayers (Table 8.2). (d) Compensation mechanism to the States as insurance against revenue shortfall To secure agreement from the states for implementing GST, the Center agreed to compensate the States for any shortfall in revenues calculated by applying 14% annual growth on the 2015–16 actual revenues from the subsumed taxes as attested by the C&AG. The analysis of state-wise compensation requirements estimated by projecting the C&AG certified revenues from the taxes subsumed in GST in 2015–16 is shown in Table 8.3. This brings out some important findings. First, the compensation requirement has steadily increased over the years from about Rs. 20,015 Crore in 2018–19 to Rs. 98,954 Crore in 2019–20. Second, in the first full financial year after GST implementation (2018–19), very few low-income and Northeastern and Hill States had to seek compensation due to moving over to the destination-based tax. Third, the compensation cess collections were adequate to meet the requirements from 2017–18 to 2019–20 (Tables 8.3 and 8.4), but thereafter, there was a severe shortfall in the cess collections due to the decline in the revenues due to the severe lockdown. The pandemic brought out the crisis of compensation to the fore. With a sharp decline in revenue collections, the amount of compensation required was large, and that led the Center to breach the agreement plunging the Union-State relationships to a new low. In the 41st meeting of the GST Council, the Center presented the total compensation requirements at Rs. 3 lakh Crore. The amount of cess available for compensation was estimated at Rs. 65,000 Crore. Of the remaining Rs. 2.35 lakh Crore, the Center attributed the loss of revenue due to the “Act of God” at Rs. 1.28 lakh Crore and suggested that the States may either borrow the entire amount of Rs.2.35 lakh crore from the market at market interest rates or borrow Rs. 1.1 lakh Crore under a special window to be opened by the RBI and both the principal and interest payments are to be adjusted against future revenue collections from the cess. After considerable controversy, the States finally accepted the second option. The actual compensation payments to the States increased from Rs. 49,286 Crore in 2017–18 to 1,65,302 Crore in 2019–20 and marginally declined to Rs. 154,365 even though the shortfall was much more, and the States were made to borrow to meet the requirements.

Potential rev Act rev

2018–19

16,738

18,775

8201

23,763

39,137

24,206

48,801

16,402

9561

2835

37,502

19,794

8331

46,973

21,861

19,922

78,632

14,360

18,808

22,299

38,710

20,935

43,391

26,119

Bihar

Chhattisgarh

Goa

Gujarat

Haryana

Jharkhand

Karnataka

Kerala

Madhya Pradesh

Maharashtra

Odisha

Punjab

Rajasthan

Tamil Nadu

Telangana

Uttar Pradesh

West Bengal

28,166

13,510

12,639

83,181

19,751

21,390

42,663

35,351

2529

8665

21,257

18,030

Andhra Pradesh

I. General Category States

State

–2047

–5410

–3271 –7.8

–12.5

–15.6

–1.1

−6.6

−1464

–427

28.2

5298

12.0

−5.8

−4549

1721

0.9

2.2

9.2

1.6

5.1

5.7

10.8

171

471

4310

130

1019

2151

306

9.4

−2.1

−336

896

−17.9

−3227

Shortfall Percent

29,776

49,466

23,866

44,130

25,421

21,441

16,370

89,640

22,711

24,922

53,549

9497

22,565

42,752

3232

10,900

18,698

20,554

27,308

27,232

23,517

38,376

21,954

12,751

13,204

82,602

20,447

20,447

42,147

8418

18,873

34,106

2438

7895

15,801

20,227

Potential rev Act rev

2019–20

2468

22,234

349

5754

3467

8690

3166

7038

2264

4475

11,402

1079

3692

8646

794

3005

2897

327

8.29 38,696

44.95 64,285

1.46 31,016

13.04 57,351

13.64 33,037

40.53 27,864

19.34 21,275

7.85 116,496

9.97 29,515

17.95 32,388

21.29 69,592

11.37 12,343

16.36 29,325

20.22 55,561

24.56 4200

27.57 14,165

15.49 24,300

25,074

50,355

23,600

42,300

37,663

11,522

13,010

88,000

20,448

19,000

37,835

9064

20,350

41,827

2373

7754

20,800

18,671

Potential rev Act rev

2020–21

1.59 26,712

Shortfall Percent

Table 8.3 The shortfall in GST revenue from compensation eligibility in States (Rs. Crore)

13,622

13,930

7416

15,051

–4626

16,342

8265

28,496

9067

13,388

31,757

3279

8975

13,734

1827

6411

3500

8041

(continued)

54.3

27.7

31.4

35.6

–12.3

141.8

63.5

32.4

44.3

70.5

83.9

36.2

44.1

32.8

77.0

82.7

16.8

43.1

Shortfall Percent

194 G. Rao

344

460

1074

4960

27,589

496,312 −3172

451

827

246

333

319

1025

6448

28,677

493,140

Manipur

Meghalaya

Mizoram

Nagaland

Sikkim

Tripura

Uttarakhand

All Special Cat States

All States

Source Author’s estimates

Compensation requirements

504

6194

902

787

5412

−105.3 280 −3.4 −44.2 −4.8

−258

−11

−141

−49

20,015

1088 −0.6

3.8

23.1

−9.1

−75

1488

−74.5

−336

562,180

32,692

7350

1169

364

379

943

514

12.6

7061

5385

8868

782

27.0

−14.3

Jammu and Kashmir

3450

1273

4723

−1111

7779

Himachal Pradesh

8890

529,488

−142.2 379

–0.9

2019–20

5511

2419

142

−91

−234

−252

33

−339

2341

1835

80

–423

98,594

464,924 97,256

27,181

4931

1027

455

613

532

910

853

4720

3550

8788

802

2020–21

17.30 725,391

16.86 37,268

32.91 8379

12.11 1332

−25.12 415

−61.56 433

−90.08 319

3.45 1074

−65.89 586

33.16 8050

34.07 6138

0.90 10,109

–111.43 432

7725

4770

347

–16

–30

–185

–145

–414

973

2687

215

-478

207,470

519,189 206,202

29,543

3609

985

431

463

504

1219

1000

7077

3451

9894

910

39.7

26.1

132.2

35.2

–3.8

–6.6

–36.7

–11.9

–41.4

13.8

77.9

2.2

–52.5

40.5

Shortfall Percent

489,646 198,477

Potential rev Act rev

17.33 688,123

Shortfall Percent

437,743 91,745

Potential rev Act rev

-473

Assam

806

333

Arunachal Pradesh

II. Special Cat States

464,463

All Gen. Cat States

Shortfall Percent

468,723 –4260

Potential rev Act rev

State

2018–19

Table 8.3 (continued)

8 Goods and Services Tax in India: A Stocktaking 195

196

G. Rao

Table 8.4 GST compensation payments to states (Rs Crore) S. no

State/UT-wise Compensation released in 2017–18

Compensation released in 2018–19

Compensation released in 2019–20

Compensation released in 2020–21

1

Andhra Pradesh

382

0

3028

4627

2

Arunachal Pradesh

15

0

0

0

3

Assam

980.39

455

1284

2149

4

Bihar

3140

2798

5464

4493

5

Chhattisgarh

1589

2592

4521

2827

6

Delhi

326

5185

8424

6931

7

Goa

281

502

1093

987

8

Gujarat

4277

7227

14,801

11,563

9

Haryana

1461

3916

6617

5841

10

Himachal Pradesh

1059

1935

2477

1623

11

Jammu & Kashmir

1160

1667

3281

2104

12

Jharkhand

1368

1098

2219

2475

13

Karnataka

7669.59

12,465.14

18,628

13,763

14

Kerala

2102

3532

8111

7077

15

Madhya Pradesh

2668

3302

6538

5863

16

Maharashtra

3077

9363

19,233

22,485

17

Manipur

24

0

0

53

18

Meghalaya

140

66

157

255

19

Mizoram

0

0

0

6

20

Nagaland

0

0

0

27

21

Odisha

2348.08

3785

5122

3633

22

Puducherry

387.29

681

1057

564

23

Punjab

5108.94

8985.06

12,187

6959

24

Rajasthan

2899

2280

6710

6312

25

Sikkim

6

0

0

69

26

Tamil Nadu

1018

4824

12,305

11,269

27

Telangana

169

0

3054

5424

28

Tripura

149

172

293

259

29

Uttar Pradesh

2432

0

9123

11,742

30

Uttarakhand

1432

2442

3375

2235 (continued)

8 Goods and Services Tax in India: A Stocktaking

197

Table 8.4 (continued) S. no

State/UT-wise Compensation released in 2017–18

Compensation released in 2018–19

Compensation released in 2019–20

Compensation released in 2020–21

31

West Bengal

1608

2615

6200

7750

Total

Total

49,276.29

81,887.2

165,302

151,365

Source Ministry of Finance, Government of India (as reported in the answers to questions in Rajya Sabha session 252)

(e) Structure of the tax and revenue impact The best practice approach to GST reform is to keep the list of exemptions in GST small to have a more comprehensive trail of transactions. The rationale for exempting many items or levying low rates on items considered necessities is to ensure equity in the distribution of tax burden. However, while exemptions or low rates of tax on such items may confer larger proportional benefits to the low-income groups, they could result in larger absolute benefits to higher income groups. As argued by Keen (2013), this is an inefficient way of achieving equity as the exempted items are consumed also by the rich and therefore, the better way to achieve equity is by making cash transfers and spending on items like education and health care.8 In India, as many as 148 commodities under four-digit HSN classification and having almost 50% weight in the consumer price index have been kept on the list of exemptions mainly to minimize the impact of the tax on prices. Apart from most unprocessed agricultural goods, processed items are not packaged and branded and some items are exempted for social reasons. The glaring examples of exempted services are the transportation of goods and people by road, railways (except travel in first or air-conditioned class), and inland waterways and courier services. As petroleum products are not included in the GST, except for road and water transportation and upper-class train travel too is exempted. Selectivity in taxation creates other anomalies. While the flour is not taxed, toasted bread and rusk are taxed at 12%, and malt, biscuits, cakes, and pastries are taxed at 18%. Given that the technology involved in their supply is not advanced, the high rate of tax drives the process of production and sale to the informal sector. While items like coffee beans and fresh tea leaves are exempt, coffee and its substitutes, tea, and dry ginger are taxed at 12%. Unbranded food items such as condiments and savories are not taxed, and branded items are subject to a 12% tax. De-oiled cakes used as cattle feed are exempt, but other uses are taxed at 5%. There are many such examples.

8

Munoz and Cho (2004) using micro data on Ethiopia conclude, “……..even very poor countries can sometimes deliver the expenditure goods more effectively than poorly targeted exemption”. Quoted in Bird and Gendron (2007) footnote.13; p. 77.

198

G. Rao

Rate differentials done for reasons of equity simply based on the perceptions about the consumption pattern may not only result in lower tax collections but may also have an adverse impact on employment and equity. Many items such as building materials including cement and its products, marbles and granites, veneer and plywood, paints and varnishes, tiles, sanitary ware, motor cars and parts, refrigerators, airconditioners, vacuum cleaners, and even chocolates and razor blades are taxed at 28% on the perception that these are luxury items. In fact, in the case of some of the items, in addition to the basic rates, compensation cess at varying rates is levied on these items over and above 28% tax. These rates vary widely creating enormous rate differences among the commodities and services subject to the 28% category. As shown in Table 8.5, on motor cars alone, the cess varies depending on the engine capacity, length, and fuel use of the vehicle. In some cases, the total incidence of the tax works out to 50%. The automobile industry with its ancillaries and repair and services has enormous employment implications. The problem is high tax rates create an incentive for evasion of the tax by creating a gray market for such goods. This is particularly true of building materials in which the rate of tax levied is 28%, but the final tax on affordable houses is one percent and on other houses, 5% without ITC. Since there is no paper trail on the input providers for the builders of these properties, they can purchase their inputs from the gray market and merely pay the compounded tax or not pay the tax at all as there is no paper trail. This can result in a huge loss of revenue. In the case of motor cars, high taxes could lead to higher prices resulting in lower demand. The problem is reinforced by the fact that motor spirit and diesel are not included in the base of GST, and this increases the cost of transportation. The resulting lower demand can cause Table 8.5 Tax rates on motor cars Type of vehicles

GST Compensation cess Total rate (%) (%) tax rate (%)

Petrol/CNG/LPG car less than 1200 cc/length less than 4 m

28

1

29

Petrol/CNG/LPG car less than 1200 cc/length more than 4 m

28

15

43

Petrol/CNG/LPG car over 1200 cc (irrespective of length)

28

22

50

Diesel car less than 1500 cc and length less than 4 m

28

3

31

Diesel car less than 1500 cc and length more than 4 m 28

20

48

Diesel car over 1500 cc engine capacity, greater than 28 4 m length, and ground clearance of 170 mm or more

22

50

Electric Cars (all sizes including 2− and 3−wheelers) 12

Nil

12

28

Nil

28

Vehicles fitted for use as an ambulance

The Rates are as decided by the 31st GST Council meeting held in December 2018. Source https:// www.paisabazaar.com/tax/gst-on-cars/

8 Goods and Services Tax in India: A Stocktaking

199

Table 8.6 VATs with a single rate at time of introduction Years

Number of new VAT countries

Percentage with single rate

Before 1990

48

25

1990–1999

75

71

1999–2011

31

81

Source Keen (2013)

a loss of employment. Thus, the tax that is intended to fall the rich may actually hurt the poor more. It is, therefore, important to assess the general equilibrium effects of the tax system while designing the tax structure.9

8.4 Multiplicity of Rates: Complications, Distortions, and Inverted Duty Structure (a) Multiple Rates Keen (2013) convincingly argues that levying the tax at multiple rates and particularly levying a lower rate of tax on the items predominantly consumed by low-income groups for equity reasons is an inefficient way of targeting. The objective is better achieved by the expenditure side of the budget. Jenkins et al. (2006) also recommend the levy of tax without rate differentiation for reasons of simplicity in administration, reducing compliance burden, and to minimize distortions. International experience shows that, over the years, the trend has been increasingly to levy the tax at a single rate (Table 8.6). Levying GST at four rates complicates the structure. In addition, when the cesses are considered, it is difficult even to count the number of rates. In the case of motor vehicles, rate differentiation is done on the basis of engine capacity, length of cars, and fuel base. There are instances where differentiation for the same commodity or service group is done based on the values (footwear, apparel, quilts) for equity reasons, Differentiation is also made based on the nature of the commodity as in the case of fibers used in textiles (natural or man-made), based on whether the commodity is in the nature of input or an output and use of the commodity (oil cakes used as cattle-feed and others). The problem with multiple rate structure is that it is easily prone to misclassification. For example, silk fiber is exempt, cotton and natural fibers are taxed at 5%, and man-made fibers are taxed at 18%. Food served in restaurants in hotels having room tariff less than Rs. 1000 per night is exempt, and in other restaurants, it is taxed at 5% without the provision of ITC whereas restaurant services in hotels with over Rs. 7500 room tariff per night and those in clubs and guest houses are taxed at 9

The automobile industry has been claiming that there have been massive layoffs on account of various factors including high rate of GST.

200

G. Rao

18%.10 Similarly, catering services were taxed at 18% which was brought down to 5% without ITC in the 37th meeting of the Council. Surely, it is not very difficult for the restaurants to classify catering as sales from the restaurants, nor is it difficult for a professional caterer to open a small restaurant to misclassify the same. (b) Cascading element Although the implementation of GST has helped to reduce the cascading element in taxes, it falls well short of the desired. This is mainly due to the exclusion of motor spirit, high-speed diesel, real estate, alcohol, and electricity from the GST. The revenue from these cascading taxes constitutes a substantial proportion of revenues collected from internal indirect taxes. At the Central level, the revenue from excise duty collected on petroleum products constituted 29.3% of the taxes on goods and services in 2019–20. The States’ revenue from consumption taxes not included in GST constitutes 53% of the revenue from total consumption taxes in 2019–20 (Table 8.7), and the ratio varied from 36% in Bihar to 61% in Telangana. The cascading element in taxes is tantamount to imposing a penalty on exports. In fact, because of excluding petroleum products from the GST base, the entire transport sector except air travel and air-conditioned and first-class train travel is exempted, rendering the tax base narrower. In addition to the exclusion of these consumption taxes from the GST, ITC is not available on items exempted from GST, the compounded tax on the dealers with less than Rs. 1. 5 Crore turnover, and compounded tax on services in restaurants and housing. A large list of exemptions and extending the benefit of compounding to cases such as supplies in restaurants and housing also deny ITC and add to cascading. Large-scale exemptions and extending the benefit of compounding to cases such as supplies in restaurants and housing also deny ITC and add to cascading.

8.5 The Reform Issues and Strategy for Implementation The implementation of GST is, perhaps, the most important consumption tax reform in India. However, it is still evolving, and considering the various shortcomings pointed out above, it will be quite some time before it stabilizes to achieve the desired objectives. The pandemic has thrown the reform agenda to the back burner and as the economy recovers; it is an appropriate time to embark on the reform agenda. However, it is by no means an easy tax because, as stated by Bird ad Gendron (1997) once some bad elements get included to make the reform acceptable, it is very difficult to remove them. Besides, it takes considerable effort to forge consensus among different states and Union governments. This section identifies specific areas of reform and the strategy that may be adopted to implement them. The important reform areas identified in the foregoing analysis are (i) broadening the base by minimizing exemptions and including petroleum products and electricity 10

Brought down to 12% in the 37th Council meeting on September 20.

8 Goods and Services Tax in India: A Stocktaking

201

Table 8.7 Revenue from cascading taxes after GST implementation in 2019–20 State

GST (Rs. Million)

Cascading taxes (Rs. Million)*

Total (Rs. Million)

Share of cascading Taxes in Total (Percent)

Andhra Pradesh

20,227

31,656.3

51,883.3

61.01

Bihar

15,805

9278

25,083

36.99

Chhattisgarh

7894.8

12,036.2

19,931

60.39

Goa

2438.5

1832.5

71

42.91

Gujarat

34,106.7

34,252.6

68,359.3

50.11

Haryana

18,873

17,918

36,791

48.70

Jharkhand

8417.7

7371.5

15,789.2

46.69

Karnataka

42,147.2

47,563

89,710.2

53.02

Kerala

20,447

25,733

46,180

55.72

Madhya Pradesh

20,447.8

28,215.9

48,663.7

57.98

Maharashtra

82,601

134,933.9

727,807

46.8

Odisha

13,203.5

16,739.6

29,943.1

55.90

Punjab

12,751.2

14,780.9

27,532.1

53.69

Rajasthan

21,954

32,690.4

54,644.4

59.82

Tamil Nadu

38,376.2

58,022.9

96,399.1

60.19

Telangana

23,516.7

36,682.3

60,199

60.94

Uttar Pradesh

47,232.4

59,018.8

106,251.2

55.55

West Bengal

27,307.5

34,018.9

61,326.4

55.47

All States

484,881.9

564,925.1

1,049,807

53.81

Note Cascading consumption taxes include sales taxes on petroleum products, motor vehicle tax, passengers and goods tax, electricity duty and entertainment tax, and other consumption taxes excluding state excise duty Source State finances: A study of budgets of 2021–22. Reserve Bank of India

in GST; (ii) rationalizing the rate structure by reducing the number of rates to avoid the problems arising from misclassification, disputes, and inverted duty structure; (iii) creating a multidisciplinary technical unit in the GST Council to analyze economic, administrative, and legal implications of the tax design and reform; (iv) the compensation to the States to the loss of revenue has ended in July 2022, but the States have been clamoring for its continuation as the revenue from the tax is yet to stabilize due to the pandemic. The recent buoyancy in the tax is encouraging and may obviate the need, but some decision has to be taken in the GST Council to settle the issue. (i) Broadening the base The best practice approach to tax design and reform is to broaden the base, reduce the rates, and reduce rate slabs to evolve a simple tax system to reduce administrative, compliance, and distortion costs while ensuring revenue productivity (Rao, 2017).

202

G. Rao

This requires substantial pruning of the exemption list. This will not only expand the tax base but also help in formalizing the transactions. A broader base would enable levying the tax at lower rates to collect the same amount of revenue. The basic exemption list should be confined to unprocessed food items and perishables. The forty-seventh meeting of the GST Council has recommended substantial pruning of the exemption list, but the issue needs to be revisited to prune the list to the bare minimum. A major anomaly in the GST system in India is the exclusion of petroleum products and electricity from the GST base. Petroleum products have been kept out mainly for revenue reasons as they constitute a substantial proportion of domestic consumption taxes in both Union and State fiscal systems. It may be noted that in 2019–20, the revenue collected from excise duty and sales tax on petroleum products constituted over 40% of internal indirect taxes collected by the Center and the States. The adverse effects on the competitiveness of Indian manufacturing on account of this have been significant due to the cascading caused by the high taxes on the transportation of goods and services. As argued earlier, it is desirable to amend the State List to exclude electricity duty and include taxation of electricity in the GST. The inclusion of both petroleum products and electricity is important to reduce cascading and to make the tax system competitive. Naturally, there is a revenue concern in including petroleum products in the GST base. However, a part of this can be recoupled by including all forms of transportation. Presently, only upper-class rail and air transportation are taxed and when petroleum products are included in GST, and all forms of transportation are taxed, the sector gets formalized and the tax base is likely to expand. Also, a preferable way is to include petroleum products in the GST base, tax them at the general rate, and levy an environmental excise in addition.11 Also, there is no case for exempting the services of advocates and doctors. (b) Rationalizing the rate structure An important reform like the unifification of tax rates is best done in the recovery phase of the economy and revenues are rising. The revenue from GST has shown a buoyancy of 1.4 in 2022–23. What is important is the revenue increase has not come about only due to economic recovery. The more important reason is the stabilization of the technology platform. E-invoicing for all businesses above Rs. 100 Crore has enabled better invoice matching and detection of fake invoices that were used to claim the input tax credit. This has helped to improve the compliance of the tax and has also enabled its better enforcement of the tax. With the passage of time, GSTN should be able to enforce e-invoice requirements to all businesses above Rs. 10 Crore, and that will cover more than 95% of the taxpayers. With this development, we can expect high revenue buoyancy. The main focus of reforms should be to unify the tax rates in the structure of the tax to make the system simple to reduce collection, compliance, and distortion costs, ideally levying the GST at a single rate besides a minimum list of unprocessed food 11

Ahmad and Starn (2011) while discussing the carbon taxes make such a recommendation.

8 Goods and Services Tax in India: A Stocktaking

203

items in the exemption list. But that may not be politically acceptable at present, and the attempt should be to reduce the number of rates to two and eventually move toward a single rate. The rate category-wise revenue collection in Karnataka in 2021– 22 shows that 67% of the tax is collected from the supplies taxed at 18% alone and together with the 12% category, 75% of the tax is collected (Table 8.8). Perhaps, it is desirable to merge the two rates to 16% without much loss of revenue. Third, there are some items in the 28% category which can hardly be called demerit goods, and these can be brought down to the general rate of 16%. The most important set of items are building and construction materials which are now taxed at 28%, and the revenue collected from this category is just about 14% (excluding the cess). Construction is a labor-intensive activity and reducing the rate would not only increase employment but also provide much-needed relief to the sector. Thus, it would be preferable to merge the 12% and 18% categories into 16%, move the 5% items to 8%, and bring down the 28% category into the general rate of 16%. As the Cesses will cease to exist after 2026, the structure of the tax will be transformed into two rates making it really a “good and simple tax”. (iii) The Question of Compensation for Revenue Loss As mentioned earlier, the States were promised compensation for any loss of revenue for a period of five years to make them agree to join the reform process. However, the pandemic brought in a revenue crisis. The total loss of revenue to the States in 2020–21 was estimated at Rs. 3 Lakh Crore of which Rs. 65,000 Crore was expected Table 8.8 Tax Rate-wise revenue collections in Karnataka Tax Turnover in S. no Total number rates 2021–22 (in of crores) taxpayers (As of March 2022)

CGST

SGST

IGST

CESS

5%

2,63,213.8

3,907.0

3,907.7

5,334.0

2

12%

1,42,127.6

6188.23

6,187.7

4,672.4

3

18%

7,97,010.7

43,571.2

55,038.7

4

28%

1,05,520.7

7,253.9

7,253.9

15,033.9

5

All

1,307,872.8

60,923.0

60,924.0

80,079.0

1

8,69,595

Revenue collections for 2021–22 (| in crores) Total

43,573.8

215.7 2.97

13,364.4 10,860.2

1464.75 1,42,183.6 9595.32 11,279.0

29,541.8 2,13,205.0

Percent of total 1

5%

20.13

6.41

6.41

6.66

1.92

6.27

2

12%

10.87

10.16

10.16

5.83

0.03

8.00

3

18%

60.94

71.52

71.52

68.73

12.99

67.38

4

28%

8.07

11.91

11.91

18.77

85.07

18.36

5

All

100.00

100.00

100.00

100.00

100.00

100.00

Note The total number of Taxpayers reflects regular taxpayers (Other than Composition taxpayers)

204

G. Rao

to accrue from the compensation cess. Of the remaining Rs. 2.35 Lakh Crore, the loss due to the pandemic described as an “Act of God” was estimated at Rs. 1.28 lakh Crore. The Center borrowed Rs. 1.1 Lakh Crore under a special window opened by the RBI and forwarded it to the States and both interest payments and repayments will be made from the future collections of the cess, the levy of which was extended up to 2016–17. The compensation scheme ended in June 2022 and some of the states have been demanding its extension for another five years on the ground that, due to the pandemic, the tax is yet to stabilize, and considering the fragility of their finances, this insurance is necessary. Ideally, increased buoyancy of the tax witnessed in recent times should provide comfort to the States obviating the need for compensation. However, extending the compensation for a couple of years more could be used as a strategy to build a consensus on the measures to rationalize and simplify the tax system. However, it would be unwise to continue with the formula of applying 14% annual growth because the tax has considerably stabilized. In fact, even the application of 14% was overly generous to the States because, the past performance of the taxes included in GST has not shown such high growth in any of the States (Rao, 2017). Considering the high buoyancy of the tax seen in recent times, linking the growth of potential revenue to GSDP growth, while giving some degree of comfort with minimum revenue, would obviate the need for making actual compensation. This revenue certainty may help to build a consensus to evolve a well-structured GST in the country.

8.6 In Conclusion Implementation of GST in India is a far-reaching tax reform involving both the Union and State governments. It is not surprising that some bad features entered into the tax system to get the reform accepted by all governments. The reform has taken firm roots, and GST is here to stay. As the technology platform settles, compliance will improve, and the higher buoyancy should provide enough confidence and impetus to undertake the reforms to make it truly a “Good and Simple Tax”. The GST implementation has led to significant gains. However, the full potential of the reform in terms of gains in market integration, formalization of the economy, improved supply chain efficiency, free movement of goods across the country, and, more importantly, revenue productivity is yet to be fully realized. The abolition of inter-state check-posts erected to keep track of cross-border transactions and intrastate check-posts erected to collect octroi and entry tax by local bodies. Harmonization of the rate structure has helped to reduce the race to the bottom and minimize inter-stage tax exportation. The tax is designed to be destination-based which has helped to improve inter-regional equity. The requirement for online payment of the tax and for availing input tax credit has led to significant formalization. Abolition of the sales tax and octroi check-posts has led to the removal of a formidable source of hindrance to the free movement of goods and has helped to reduce the transaction cost and ensure faster movement and reduced cost of transporting goods. Furthermore,

8 Goods and Services Tax in India: A Stocktaking

205

the GST administrative system has provided a useful example of providing an institutional mechanism to replicate in areas requiring intergovernmental cooperation, bargaining, and conflict resolution in Indian federalism. Although the GST reform has taken a firmer root and is here to stay, much remains to be done to make it simple, efficient, and productive and that should be addressed by the next generation of reforms. These should include broadening the base by pruning the list of exemptions in goods and services, the inclusion of petroleum products, electricity, and real estate in the tax base, simplification of the rate structure by reducing the number of rate categories, firming up the technology platform and its periodic upgradation, and creation of a technical unit comprising experts from various fields such as administration, taxation, accounting, economics, law, and big data analysis to undertake continuous research. One of the major constraints in undertaking detailed research is the reluctance of the GST Council to share the data even for research. There is considerable hesitancy on the part of the Council to share the data even with the C&AG and the Finance Commission. The report of C&AG is unequivocal in stating, “after much pursuance, CBIC has shared only the MIS reports which give aggregate statistics at Commissionerate level (for Central data) and State level (for State data)” and, “….Unhindered and full access to pan-India data is crucial for meaningful audit and to draw required assurances needed, otherwise certifying revenue receipts may become difficult. DoR’s offer of providing data based on CAG’s queries is not workable, as without the full data, it is neither possible to formulate queries, nor run the required algorithms on the data. The CAG sought data through the Application Programme Interface (APIs) already designed by GSTN. It needs hardly be stated that providing such data as CAG may require is a constitutional and legal requirement”. When a Constitutional body like the C&AG itself has difficulties in securing the data required for conducting the audit and analysis, it is not surprising that independent researchers find it impossible to secure access to the information required to undertake quality analysis. Hopefully, the Council will wake up to the need for scientific analysis and not try to “shoot the messenger”.

206

G. Rao

References Ahmad, E., Stern, N. (2011). Effective carbon taxes and public policy options. in M. Govinda Rao & Mihir Rakshit (Eds.), Public economics: theory and policy. Sage Publications Bird, R., & Gendron, P.-P. (2007). Value added taxes in developing and transitional Countries, (Cambridge and New York). Cambridge University Press. Choi, K. (1990). Value-added taxation: experiences and lessons of Korea. In R. M. Bird & O. Oldman (Eds.), Taxation in Developing Countries (pp. 269–287). Johns Hopkins University Press. Cnossen, S. (2010). VAT coordination in common markets and federations: lessons from the European experience. Tax Law Review, 63, 584–603. India. (2019). Report No. 11 of 2019: Compliance audit—Department of revenue: Goods and Services Tax, Comptroller and Auditor General, Government of India. Jenkins, G. P., Jenkins, H. P., & Kuo, C. Y. (2006). In the VAT Naturally Progressive? Queens University. Keen, M. (2013). Targeting, cascading and indirect tax design (IMF Staff Paper: 13/57). International Monetary Fund. Keen, M., & Mintz, J. (2004). The optimal threshold for a value-added tax. Journal of Public Economics, 88(3/4), 559–576. Keen, M., & Lockwood, B. (2010). The value added tax: ots causes and consequences. Journal of Development Economics, 92, 138–151. Krever, R. (2014). Combating VAT fraud: Lessons from Korea? British Tax Review, (3). Muñoz, M.S.,& Cho, S.S. (2004). Social impact of a tax reform: The case of Ethiopia. In Sanjeev Gupta et. Al (Eds.), Helping countries develop: The role of fiscal policy, International Monetary Fund, Washington D.C. Rao, G. M. (2017). Goods and services tax in India: Progress, performance and prospects (Working Paper No. 2019–02). SIPA—Deepak and Neeraj Raj Centre, Columbia University. Rao, M.G. (2022). Evolving issues and future directions in GST reform in India (Madras School of Economics Working Paper 221). Chennai. Varsano, R. (2000). Subnational taxation and treatment of interstate trade in Brazil: problems and a proposed solution. In J. Burki and Guillermo Perry metal (Eds.) Decentralization and accountability of the public sector (pp. 339–355). The World Bank, Washington D. C.

Chapter 9

Recent Reforms in India’s Corporate Income Tax Regime: Rationale, Impacts, and Improvements Supriyo De

Abstract In a recent innovative policy reform, India’s corporate income tax system was overhauled with optional lower rates in lieu of giving up complex deductions. However, official data reveals a puzzle wherein larger companies have opted more for the lower optional rates, while smaller ones appear reticent in switching to the optional regime. This paper explores this issue using empirical methods. The evolution of tax rates is tracked through reforms simplifying the tax system in the 1990s, the subsequent conundrum of zero-tax companies leading to the introduction of minimum alternate tax, and the persistence of lower effective tax rates for larger companies. This provides the rationale for a simpler tax regime with lower rates but fewer deductions. The user cost of the capital approach is used to examine the economically relevant tax impact across various sectors and ownership types. The results indicate that in terms of user cost, the various lower tax options are not attractive, and under certain situations may be worse for younger and smaller companies. In light of the analysis, policy options are suggested to improve the scheme so as to achieve the laudable objective of implementing a simple tax regime with lower rates and minimal deductions. Keywords Corporate income tax · User cost of capital · Minimum alternate tax JEL Codes D21 · E22 · H25

Thanks are due to Adam Hussain, Prachi Jain, and Neeti Gupta for research assistance and to Usha Mathur for secretariat support functions. The chapter is a re-publication of the author’s working paper and is being re-used here with permission. De, S. (2023). Recent Reforms in India’s Corporate Income Tax Regime: Rationale, Impacts and Improvements, NIPFP Working Paper Series, Working Paper No. 393. National Institute of Public Finance and Policy, New Delhi: India. S. De (B) National Institute of Public Finance and Policy, New Delhi, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 D. K. Srivastava and K. R. Shanmugam (eds.), India’s Contemporary Macroeconomic Themes, India Studies in Business and Economics, https://doi.org/10.1007/978-981-99-5728-6_9

207

208

S. De

9.1 Introduction The prosperity of the country depends on the amount and type of investment that businesses and investors make for earning returns. Taxes are one of the factors that impact investment decisions. Low rates of taxes and high tax incentives can encourage entrepreneurs to set up or expand their businesses. Alterations in the tax rate can impact the scale and composition of investment and business asset creation. Complex tax provisions and different tax treatments across different forms of investment can distort investment decisions. In a recent innovative policy reform, India’s corporate income tax system was overhauled with the introduction of optional lower rates in lieu of giving up complex deductions. With effect from assessment year (AY) 2017–2018 (corresponding to income arising in the previous year or fiscal year (FY) 2016–2017), per section 115BA, domestic manufacturing companies were given the option of switching to a lower basic tax rate of 25% instead of the prevailing basic rate of 30%. For this, they had to forgo certain deductions and incentives. For AY 2020– 2021, two further lower rate options were introduced. Section 115BAA applicable to all domestic companies offered a lower rate of 22% for forgoing specified deductions. Section 115BAB offered domestic manufacturing start-up companies an even lower rate of 15% for forgoing a wider set of specified deductions. With these major tax reforms, it is a propitious time to examine how the different tax rates impact different industries and companies. At first glance, the simplified regime with lower rate options seems very attractive and should have seen many companies opting for them. However, data from the government reveals a puzzle wherein larger companies have opted more for the lower optional rates while smaller ones appear reticent in switching to the optional lower tax regime. According to the receipt budget document for 2022–2023, in FY2019-20, only 15.85% of companies opted for 22% tax without exemptions/incentives (s. 115BAA), and only 0.14% of companies for 15% tax meant for new manufacturing units (s. 115BAB).1 Of the companies with a total income of Rs. 0–1 crore, 748 opted for s. 115BAA, 256 for s. 115BAB while the vast majority, 2,48,934 remained with the 30% rate. On the other hand, for companies with total income over Rs. 500 crore, 246 opted for s. 115BAA and only 130 remained with the 30% tax rate. This conundrum motivates our research questions: (1) Why did such few companies opt for the simpler tax regime with lower tax rates and fewer deductions? (2) Why were larger companies more willing to take up the simpler tax regime with lower tax rates and fewer deductions than smaller ones? (3) Why was the take-up of the scheme for manufacturing start-ups low? Official corporate tax data in India is reported at a very high level of aggregation which makes detailed analysis difficult. Corporate accounts datasets (such as CMIE Prowess) report the accounts according to company law accounting standards which 1

https://www.fortuneindia.com/opinion/did-corporate-tax-cut-of-2019-lead-to-tax-gdp-boom/109 294#:~:text=He%20further%20claims%20that%20this,and%20the%20GDP%20by%2033%25.

9 Recent Reforms in India’s Corporate Income Tax Regime: Rationale …

209

differ from income tax norms. Moreover, they report data for mostly publicly traded companies and therefore miss out on some of the smaller entities. In this paper, we circumvent these issues by relying on the user cost of capital methodology to examine why different types of corporate entities may not have been very enthusiastic to adopt the concessional tax rates. In particular, we delve into detailed aspects of the deductions to be forgone and model them in an empirical manner under certain assumptions. Targeted tax deductions lower the cost of capital for firms. For small firms and start-ups, losing them may be increasing the user cost of capital inordinately. Moreover, in the initial phases, start-ups are usually unable to earn positive returns and need to set off losses and carry them forward. Yet if companies desire to avail lower tax rate benefits under sections 115BA, 115BAA, or 115BAB, they must forgo the opportunity of setting off losses. They have to face a difficult trade-off between the concessional tax rates and the opportunity of claiming deductions and setting off losses. On the above argument, we postulate the following. Hypothesis 1: In terms of user cost of capital, the new tax regime does not provide significant advantages over the old regime. Hypothesis 2: Hypothesis 1 is particularly pronounced for small and younger firms. CMIE Prowess data is utilized for the calculation of user cost of capital, based on industry asset composition and ownership classification. The research finds that in terms of user cost of capital, the various lower tax options are not attractive, and under certain situations may be even worse for younger and smaller companies. Moreover, due to its irreversible nature and inability to carry forward losses, start-ups and small companies may be hesitant in committing to the new scheme. To make matters worse, the complexity of the provisions may also be deterring adoption. In light of the analysis, certain policy options are suggested to improve the scheme so as to achieve the laudable objective of implementing a simple tax regime with lower rates and minimal deductions. This paper expands the research frontier in several ways: (1) To the best of the author’s knowledge, this is a pioneering appraisal of the new optional tax regime. (2) It is also one of the few applications of the user cost of capital approach to the analysis of tax policy in substantial depth thereby demonstrating its utility in prospective tax policy formulation. The remainder of this paper is organized as follows: Sect. 9.2 summarizes past literature papers on Indian tax reforms and the user cost of capital theory; Sect. 9.3 assesses tax rate evolution in India, discusses certain tax reforms, and provides stylized facts and analyses of effective tax rates; Sect. 9.4 describes recent optional tax rates for companies and number of companies who availed it; Sect. 9.5 provides the analysis using the user cost of capital approach; and Sect. 9.6 concludes the paper enumerating certain recommendations, providing caveats, and suggesting future research on the theme.

210

S. De

9.2 Literature Review This section highlights literature on various tax reform measures undertaken by the Indian government over a period of time as well as the neoclassical underpinnings of the user cost of capital approach. The papers related to Indian tax reforms are limited in number due to the complex nature of the topic. Acharya (2005) opines that the wave of tax reforms that started in the 1970s (1974–1984) was inadequate and emphasized increasing tax rates to finance government-led companies. Even the tax slab rate was more bizarre. In 1973–1974, there were 11 different tax slab rates ranging from 10% and climbing as high as 85%. With a surcharge of 15%, the top-bracket taxpayers had to pay 97.75%! The consequence of higher tax rates was widespread tax evasion and avoidance of such taxes. Even corporate income taxes were at around 60% with complicated tax systems of tax administration. The second phase of tax reform started under then Finance Minister VP Singh (1985– 1987) was a modern tax reform. Then in the 1990s, the Chelliah Committee Report constituted the finest treatment of tax policy and reform in India. The report was comprehensive, empirically supported and provided clear recommendations. Post2000 reforms simplified personal taxes substantially by reducing the number of rates, raising the exemption limit and widening the tax brackets. Rao and Rao (2006) showed progress in tax reforms which raised the tax to GDP ratio over the years. However, they mention that tax reforms are a continuing exercise for improving equity and productivity and reducing distortions. Ahluwalia (2022) mentioned that tax reforms involving lowering tax rates, broadening the tax base, and lowering all loopholes were expected to raise tax to GDP ratio. Samantara (2021) examined the tax reforms undertaken from 1991 onwards and concluded that still more tax reforms are needed to ameliorate tax revenue while enhancing equity. He also mentions that while revenue from direct taxes increased over the years, there is still space for increasing revenues by removing complex tax structures and rationalization of taxes to reduce tax avoidance and evasion. The study also states that tax reforms in most transitional and developing economies are often necessitated by the need to be in sync with international compliance practices. Alagappan (2019) studied India’s recent tax reforms and stated that the direct tax rate is inflated. Sankarganesh and Shanmugam (2021) analyze the effect of corporate income tax on the investment of Indian manufacturing firms during FY2005-2019. It was found that the effective corporate tax rate has a negative and significant impact on corporate investment. Moreover, the estimated effective tax elasticity is relatively low compared to the magnitude found in other countries. The effect of corporate taxes on investment and entrepreneurship is one of the central questions in economic theory. The neoclassical theory of investment relies largely on two approaches: user cost of capital theory and Q-theory. User cost of capital was initially developed by Jorgenson (1963) and Hall and Jorgenson (1967) to study the impact of tax policies on investment. The theory of user cost of capital was later extended by several economists especially Abel (1983), who incorporated adjustment costs in the user cost of capital model. Past studies using user cost of

9 Recent Reforms in India’s Corporate Income Tax Regime: Rationale …

211

capital have shown that generous deductions and allowances for capital expenditure increase the present value of depreciation allowances and consequently bring down the user cost of capital, which in turn increases investment incentives (Hall & Jorgenson, 1967; Hayashi, 1982, Chirinko et al., 1999; Dwenger, 2014). Similarly, the Q-theory of investment emerged out of Tobin’s (1969) revolutionary conjecture that the rate of investment should be related to the market value of invested capital with respect to the replacement cost of invested capital. Q-theory was later extended substantially by Lucas and Prescott (1971) by using a dynamic investment model with convex adjustment costs to capture the dynamics of investment. Subsequently, Hayashi (1982) equated the marginal value of investment to its average value by imposing the assumption that an investor is a price taker, and production and cost of installation of capital are both homogeneous so that a simple regression of investment on q should have a strong fit. Studies that used Q-theory (see, e.g., Summers, 1981; Feldstein et al., 1983; Cummins et al., 1996; Auerbach, 2002; Devereux et al., 2002) to analyze the impact of tax policies on investment have found a negative association of effective corporate tax rates with aggregate investment. In the context of our research, user cost of capital is more appropriate since Q-theory relies on the stock market value of companies which is not available for closely held companies.

9.3 Tax Rate Evolution in India: Stylized Facts and Analysis of Effective Tax Rates 9.3.1 Basic Tax Rates Tax policy revolves around deciding on the taxable entity, the tax rate, the tax base, and the inclusions or exceptions to these. It needs to strike a delicate balance between competing goals: equity, efficiency, and simplicity. The Indian corporate tax rate structure is also caught in this balancing act. In FY 1989–1990, there were five rates of corporate tax for (1) closely held domestic companies engaged in trading or investment, (2) other closely held domestic companies, (3) widely held domestic companies, (4) foreign companies’ income from certain fees and royalties, and (5) other incomes of foreign companies. A process of rationalization and simplification followed (Fig. 9.1). In 1991–1992, the distinction between closely held trading and investment companies and other closely held companies was removed leaving only 4 rates. In 1995–1996, the distinction between widely held and closely held companies was also removed leaving one rate for all domestic companies. This simplified structure continued till FY 2016–2017 when a lower tax rate was brought for smaller domestic companies (with taxable income below Rs. 5 crore) bringing the number of rates up to 4 (2 for domestic companies and 2 for foreign companies). From FY 2019–2020, this norm was modified to make smaller size category being domestic companies with turnover less than Rs. 400 crore in the preceding year. For FY 2020–2021, a new set of special rates

212

70

S. De

Statutory corporate tax rates

60

50 40 30

10

1988-89 1989-90 1990-91 1991-92 1992-93 1993-94 1994-95 1995-96 1996-97 1997-98 1998-99 1999-00 2000-01 2001-02 2002-03 2003-04 2004-05 2005-06 2006-07 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 2016-17 2017-18 2018-19 2019-20 2020-21 2021-22

20

Corporate tax small companies Corporate tax general Corporate tax widely held Corporate tax closely other than trading or investment Corporate tax closely held trading or investment Corporate tax non-domestic (other than fees from govt.etc.) Special rate 115BA Special rate 115BAA Special rate 115BAB

Fig. 9.1 Evolution of statutory corporate tax rates. Source Author’s analysis. Data collated from various ministry of finance budget documents

was instituted for an alternative simplified regime for companies willing to opt for giving up complex tax incentives, deductions, and MAT credits for lower upfront rates of tax. This brings the number of rates to 6 (7 if the rate of 25% for small companies is considered different from the 25% special rate for s. 115BA). For a deeper dive into corporate income tax data, we use the “CMIE Prowess Database”. This database gives details drawn from annual financial statements (profit and loss accounts, balance sheets) of publicly listed corporations, foreign entities, and even some large private corporations. The provisions and reporting of direct taxes in the corporate accounting system (including the database) when compared with direct tax law need some clarifications. Most tax systems account for incomes in a manner that results in taxable profit being different from profit before tax (PBT) per corporate accounts. These differences arise from differences in depreciation calculations for various assets, additional allowances and incentives given by the tax system, and allowances or disallowances of certain expenses. Tax laws also allow for amortization of preliminary expenses (s. 35D of Income Tax Act) and carry forward and set off losses that may make corporate accounts and tax accounts dissimilar. Besides divergences in the computation of profits, this could also lead to the rather peculiar situation where a company with positive profits as per corporate accounts would have no or very low taxable income per tax law. Creative accounting and tax planning allowed some companies to deliberately pay low or no tax. To address this, many tax systems put in place alternative minimum tax or minimum alternate tax rules. With minimum alternate tax (MAT) provisions in place, there is an additional layer of complication. Per Indian tax laws,

9 Recent Reforms in India’s Corporate Income Tax Regime: Rationale …

213

if the corporate income tax calculated is less than a certain threshold, an additional liability arises that requires payment up to that threshold (MAT is discussed in detail below). This additional liability can be accumulated for a specified period of time for adjusting against future tax liabilities.

9.3.2 Minimum Alternate Tax MAT stands for Minimum Alternate Tax and AMT stands for Alternate Minimum Tax.2 Initially, the concept of MAT was introduced for companies and progressively it has been made applicable to all other taxpayers in the form of AMT. MAT was introduced by the Finance Act of 1987 with effect from the assessment year 1988– 1989. Later on, it was withdrawn by the Finance Act of 1990 and then reintroduced by Finance (No. 2) Act 1996, with effect from 1st April 1997. The objective of the introduction of MAT is to bring into the tax net “zero-tax companies” which, despite having earned substantial book profits and having paid handsome dividends, do not pay any tax due to various tax concessions and incentives provided under the Income-tax Law. As per the concept of MAT, the tax liability of a company will be higher of the following: (1) Tax liability of the company computed as per the normal provisions of the Income-tax Law, i.e., tax computed on the taxable income of the company by applying the tax rate applicable to the company. Tax computed in the above manner can be termed as normal tax liability. (2) Tax computed @ x% (plus surcharge and CESS as applicable) on book profit (manner of computation of book profit is discussed in a later part). The tax computed by applying x% (plus surcharge and CESS as applicable) on book profit is called MAT. This rate x is usually less than the normal corporate tax rate in cardinal terms.3 As per section 115JB, every taxpayer being a company is liable to pay MAT, if the income tax (including surcharge and CESS) payable on the total income, computed as per the provisions of the Income-tax Act in respect of any year is less than x% of its 2

https://www.incometaxindia.gov.in/tutorials/10.mat-and-amt.pdf. “Book profit” for this purpose is net profit as per profit and loss accounts prepared as per corporate law PLUS: income tax paid/payable; amounts carried to reserves; provisions for unascertained liabilities and losses of subsidiaries; dividends paid/proposed; expenditures related to certain exempt incomes; expenditures related to tax-free AOP or BOI; for foreign company expenditures related to capital gains on securities or interest; royalties, etc.; notional loss on transfer of specified capital assets; expenditure related to specified patent royalties; depreciation debited to P&L a/c; deferred tax and provision thereof, etc. LESS (if the amounts are credited to P&L accounts): Amounts withdrawn from reserves or provisions; specified exempt incomes; depreciation credited to P&L a/c; incomes related to tax-free AOP (Association of Persons) or BOI (Body of Individuals); for foreign company incomes related to capital gains on securities or interest; royalties, etc.; notional gain on transfer of specified capital assets; income related to specified patent royalties; profits of sick industrial company till its net worth becomes zero/positive; deferred tax if credited, etc.

3

214

S. De

book-profit + surcharge (SC)+health and education CESS. The MAT rate was 30% for AY 1996–1997 to AY 2000–2001, 7.5% for AY 2001–2002 to AY 2006–2007, 10% for AY 2007–2008 to AY 2009–2010, 15% for AY 2010–2011 and AY 2020– 2021 to AY 2022–2023, 18% for AY 2011–2012 and 18.5% for AY 2012–2013 to AY 2019–2020. However, the provisions of MAT are not applicable on4 (a) The domestic companies which have opted for tax regimes under section 115BAA or section 115BAB; (b) Any income accruing or arising to a company from the life insurance business referred to in section 115B; (c) Shipping company, the income of which is subject to tonnage taxation. A company has to pay the higher of “normal tax liability” or “liability as per MAT provisions”. If in any year the company pays liability as per MAT, then it is entitled to claim credit of MAT paid over and above the normal tax liability in the subsequent year(s). The provisions relating to carrying forward and adjustment of MAT credit are given in section 115JAA. Provided that where the amount of Foreign Tax Credit (“FTC”) allowed against the MAT exceeds the amount of such FTC admissible against the tax payable by the assessee under normal provisions of the Income-tax Act, then, while computing the amount of FTC under this sub-section, such excess amount shall be ignored. As discussed earlier, a company is entitled to claim MAT credit, i.e., excess of MAT paid over the normal tax liability. The credit of MAT can be utilized by the company in the subsequent year(s). The credit can be adjusted in the year in which the liability of the company as per the normal provisions is more than the MAT liability. The brought forward MAT credit shall be allowed to be set off in the subsequent year(s) to the extent of the difference between the tax on its total income as per the normal provisions and as per the MAT provisions. The company can carry forward the MAT credit for adjustment in subsequent year(s), however, the MAT credit can be carried forward only for 15 years (10 years for AY 2010–2011 to AY 2017–2018, 7 years for AY 2005–2006 to AY 2009–2010) after which it will lapse. In other words, if MAT 4

Further, as per Explanation 4 to Section 115JB as amended by Finance Act, 2016, with retrospective effect from 1/4/2001, it is clarified that the MAT provisions shall not be applicable and shall be deemed never to have been applicable to an assessee, being a foreign company, if. (i) the assessee is a resident of a country or a specified territory with which India has an agreement referred to in sub-Section (1) of Section 90 or the Central Government has adopted any agreement under sub-Section (1) of Section 90A and the assessee does not have a permanent establishment in India in accordance with the provisions of such agreement; or. (ii) the assessee is a resident of a country with which India does not have an agreement of the nature referred to in clause (i) and the assessee is not required to seek registration under any law for the time being in force relating to companies.

Further, as per Explanation 4A to Section 115JB as inserted by Finance Act, 2018, MAT provisions shall not be applicable to a foreign company, whose total income comprises profits and gains arising from business referred to in Section 44AB, 44BB, 44BBA, or 44BBB and such income has been offered to tax at the rates specified in those sections.

9 Recent Reforms in India’s Corporate Income Tax Regime: Rationale …

215

credit cannot be utilized by the company within 15 years (immediately succeeding the assessment year in which such credit was generated), then such credit will lapse. No interest is paid to the taxpayer in respect of such credit. The provisions of MAT are applicable to corporate taxpayers only. The provisions relating to AMT are applicable to non-corporate taxpayers in a modified pattern in the form of Alternate Minimum Tax (AMT). Thus, it can be said that MAT applies to companies and AMT applies to a person other than a company. Intentionally no attempt is made to over-simplify the legal complexities of MAT/AMT as per the tax law as reflected above. MAT/AMT adds another layer of complexity to an already complex tax system. Moreover, it seeks to take away from the right hand (through additional taxes) what was earlier given by the left hand (in the form of deductions, allowances, and incentives). The original sin, in this case, is excessive deductions, allowances, and incentives. In the long run, this issue needed to be addressed. Typically, company balance sheets report provisions for the taxes including amounts set aside for past liabilities (or drawing upon past reserves) and those set aside for future tax liabilities (or expected refunds from past payments). Also due to differences in accounting methods of corporate law and income tax law, deferred tax assets or liabilities are reflected in the accounts. For a more precise estimate of the effective corporate tax for a company, we use the “corporate tax” variable which includes MAT paid (but not the utilization of earlier MAT credits). Both metrics, “direct tax provisions” and “corporate tax”, track each other closely, but in a postMAT scenario with additional direct taxes (such as dividend distribution tax and securities transaction tax), the latter reflects the immediate corporate tax liability of a company in an economic sense. If the company’s computed tax is less than the MAT rate, this metric will reflect the corporate tax plus MAT paid to be credited for future claims. If the company’s computed tax is more than the MAT rate, this metric will reflect just the corporate tax (and not MAT credited utilized since that would amount to double-counting MAT paid earlier).

9.3.3 Effective Corporate Tax Rates The average corporate tax as% of PBT is a fair approximation of the effective corporate tax rate. In 1988–1989, this stood at around 25% for the full sample. By 1991– 1992, this had climbed rapidly to nearly 36%. The rate then dropped back to slightly below 25% in 1994–1995 (Fig. 9.2). The decline continued till 2001–2002 (to just below 20%) before slowly climbing back to around 31% in 2014–2015. Subsequently, the rate declined driven partly by corporate rate cuts. Given the differential tax rates for foreign entities, and for some time the differences between tax rates for public and private companies, analyzing the data across entity types is instructive (Fig. 9.3). Prior to FY 1995–1996 (AY 1996–1997), widely held (Public Ltd.) companies had a lower corporate tax rate than closely held (Private Ltd.) companies. Foreign entities continue to have a higher corporate tax rate than domestic companies. This is reflected in the average effective corporate tax rates of the three types of entities.

216

S. De

However, even after having the same corporate tax rate since FY 1995–1996 (AY 1996–1997), closely held companies continue to have a higher average effective corporate tax rate than widely held ones. In either case, domestic effective average corporate tax rates peaked once around FY 1991–1992 and FY 1992–1993 and then again in FY 2014–1915. Average corporate tax rates can also be analyzed using deciles based on turnover and assets as provided by the Prowess database (Fig. 9.3). Until 2016–2017, average corporate tax rates for the highest turnover and asset group, Deciles 1 was lower 40 35 30 25

20 15 10

0

1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020

5

Direct Tax as % of PBT

Corporate Tax as % of PBT

Fig. 9.2 Average effective corporate tax rates. Source Author’s analysis based on prowess data

in percent 45 40 35

30 25 20 15 10 5 1988-89 1989-90 1990-91 1991-92 1992-93 1993-94 1994-95 1995-96 1996-97 1997-98 1998-99 1999-00 2000-01 2001-02 2002-03 2003-04 2004-05 2005-06 2006-07 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 2016-17 2017-18 2018-19 2019-20

0

Decile 1

Decile 10

Fig. 9.3 Average effective corporate tax for highest and lowest deciles. Source Author’s analysis based on prowess data

9 Recent Reforms in India’s Corporate Income Tax Regime: Rationale …

217

than that of the lowest turnover and asset group, Deciles 10. This shows that more endowed companies are better able to plan their tax affairs. This has been noted in prior research (Guha, 2007). Only after the tax rates for smaller companies (with turnover less than Rs. 5 crore in the preceding fiscal year) were lowered (to 29%) from FY 2016–2017 and then again (to 25%) in FY 2017–2018 did the effective rate for the largest companies become more than that of the smallest companies in the sample. To that extent, though the differential tax rates complicated the tax code, the change has enhanced vertical equity in the corporate tax system. The corporate tax rate distribution across deciles and across time is also interesting. By 1999–2000, with the lowering of domestic statutory corporate tax rates to 35%, most firms in the larger categories (Deciles 1 to Deciles 8) had effective tax rates of 10–15% (Fig. 9.4). Moreover, the prevalence of low or zero-tax companies (0–5% effective rates) was relatively high at over 10% each for Deciles 1 to Deciles 7. This was despite the (re-)introduction of MAT in FY 1997–1998. This prevalence was less among smaller companies (Deciles 7 to Deciles 10) indicating partly that larger companies were able to plan their accounts and tax matters better. Further, the larger companies with more investable retained profits and the ability to raise new equity or borrowing may have been more adept at purchasing new capital to use generous tax depreciation and other tax incentives. It may also indicate major deviations between “book profits” per tax law for MAT computations and profit before tax (PBT) per corporate accounts. By 2019–2020, the domestic corporate tax rates were split to 25% for smaller companies and 30% for larger ones (Fig. 9.5). The equitability impact of this is clearly visible. The smaller companies (Deciles 8 to Deciles 10) have a prevalence of average rates of 20–25% close to the relevant statutory tax rate. The larger companies 30 25 20 15 10 5 0

Decile 1

Decile 2

Decile 3

0%-5%

5%-10%

10%-15%

Decile 4 15%-20%

Decile 5

Decile 6

20%-25%

Decile 7

25%-30%

Decile 8

30%-35%

Decile 9

35%-40%

Decile 10 40%-45%

Fig. 9.4 Average tax rate categories by turnover/asset deciles—1999–2000. Source Author’s analysis based on prowess data

218

S. De

Average tax rate categories by turnover/asset deciles - 2019-20

35 30 25 20 15 10

5 0

Decile 1 0%-5%

Decile 2 5%-10%

Decile 3 10%-15%

Decile 4

Decile 5

15%-20%

Decile 6

20%-25%

Decile 7

25%-30%

Decile 8

30%-35%

Decile 9

35%-40%

Decile 10 40%-45%

Fig. 9.5 Average tax rate categories by turnover/asset deciles—2019–2020. Source Author’s analysis based on prowess data

(Deciles 1 to 7) have a prevalence of effective rates of 25–30% close to the relevant statutory tax rate. Nevertheless, there remains a large share of companies at lower tax slots of 15–20% and 0–5%.

9.4 Recent Optional Lower Tax Rate Regime Recent reforms in the corporate tax rates and structures can be analyzed in light of the above discussion. The effective tax rates reveal that many companies still manage to pay much lower taxes than the statutory rate through the creative use of a combination of deductions and incentives. Attempts to address this using MAT or lower tax rates for smaller companies further complicate the tax code. The policy objective then should be to slowly wean the corporate sector away from complex deductions and high statutory tax rates to fewer deductions and lower tax rates. The new optional tax rates specified in Sects. 115BA, 115BAA, and 115BAB are a laudable step in this direction. These sections are discussed below in simplified terms (Table 9.1) while complex provisions of these sections are given in Annexure 1. 1. Section 115BA Section 115BA specifies tax rates only for manufacturing businesses in India under specific circumstances. The tax rate is 25% (plus surcharge and cess) for manufacturing enterprises that adhere to all Chapter XII requirements (apart from sections 115BAA and 115BAB). There is no time limit granted to domestic manufacturing enterprises for selecting a reduced income tax. Once they have adjusted the brought-forward losses, they can utilize the section 115BAB benefits whenever they

9 Recent Reforms in India’s Corporate Income Tax Regime: Rationale …

219

Table 9.1 Summary of major aspects of s. 115Ba, s. 115BAA, and s. 115BAB Particulars

Section 115BA

Section 115BAA

Section 115BAB

Effective date

AY 2017–2018

AY 2020–2021

AY 2020–2021

Eligible entities

All domestic companies engaged in the manufacturing and production of articles

All domestic companies

All domestic companies engaged in the manufacturing and production of articles

Date of introduction

Incorporated and commenced on and after March 1, 2016

No specific requirement

Incorporated on and after October 1, 2019, and commenced on or before March 31, 2023

Not allowed

Not allowed

Allowance for specified Not allowed deductions Basic rate of tax

25%

22%

15%

Surcharge

7% if income above Rs 10% 1 crore up to Rs 10 crores, 12% for income above Rs 10 crores

10%

Cess

4%

4%

4%

Mat applicability

Applicable@15%

Not applicable

Not applicable

Provision for specified domestic transactions

Not applicable

Not applicable

Applicable

Restriction on use of second-hand plant and machinery building used as a hotel or convention center

No such restriction

No such restriction

Restriction is applicable

Note Author’s compilation based on various ministry of finance budget documents

want. There are multi-requirements that must be met to guarantee that the tax rate stays at 25%. Moreover, a lot of conditions need to be fulfilled by the manufacturing companies. 2. Section 115BAA Section 115BAA provides a lower corporate tax rate for all domestic businesses (that is, it covers non-manufacturing sectors also). Domestic enterprises have an alternative tax rate of 22% plus a surcharge of 10% and a cess of 4% with an effective tax rate of 25.17%. This section is effective from FY 2019–2020 (AY 2020–2021) after fulfilling the requirements described under it. Moreover, if the corporation chooses section 115BAA, it is not required to pay tax under MAT. The following requirements must be met for all domestic businesses and manufacturing firms to elect to pay income tax at the rate of 22% (plus any surcharge and cess):

220

S. De

(1) Such businesses should not take advantage of specified incentives provided by various income tax regulations. (2) As a result, the company’s total income is calculated by not allowing various deductions. Moreover, if the company chooses to be taxed under section 115BAA, the opportunity to claim set-off is permanently lost. Such companies will have to exercise this option to be taxed under section 115BAA on or before the due date of filing income tax returns mentioned under section 139, which is typically September 30 of the assessment year. The aforementioned losses shall be deemed to have been allowed and shall not be eligible for carry forward and set-off in subsequent years. 3. Section 115BAB A domestic corporation opts for section 115BAB if it satisfies the requirements listed in section 115BAB (2) mentioned in Annexure 1. A company established and registered in India is considered to be a domestic corporation. The benefit is accessible beginning with the fiscal year 2019–2020 (AY 2020–2021). A new manufacturing business might choose to be taxed in accordance with section 115BAB at the rate of 15%. Once the corporation selects section 115BAB for a certain fiscal year, it cannot change it subsequently. Provisions of taxation should be straightforward to comprehend and apply. Unfortunately, most of the incentive deductions and exemptions in the Income Tax Act of 1961 are extremely complicated due to strict restrictions. These factors have resulted in legal disputes about various incentives. Other additional provisions related to sections 32 and 32AD are exceedingly complex and subject to a variety of conditions. There are a host of conditions that new manufacturing firms must meet in order to qualify for the benefit under the section. For example, a manufacturing firm should not use any plant and machinery that was previously used in any of their businesses; such machinery or plant was not, at any time prior to the date of the installation, used in India; such machinery or plant is imported into India as these assets have already availed the benefit of depreciation before. The above conditions deter new firms from availing the benefit of depreciation, which is the major component for a firm to reduce its net income. As noted, sections 115BA, 115BAA, and 115BAB have associated opportunity costs. The main opportunity cost is not claiming set-off losses and unabsorbed depreciation carried forward from any earlier years. Start-ups are usually unable to earn profits in the initial stage. Since they are unable to carry forward losses into the future, start-ups and newly incorporated entities are reluctant to avail these sections. Data from Budget 2022–2023 also shows that only 8.35% loss-making companies availed section 115BAA benefits, 0.19% of such companies availed section 115BAB losses, while 91.47% exercised normal tax rate provisions due to the high costs of availing the benefits (Table 9.2). This trend reduces almost monotonically as firms grow larger. In effect, the optional tax provisions intended to simplify the tax regime for small firms and start-ups inadvertently became a tax cut for larger and more established firms. Probably for the larger more profitable companies, the lack of various deductions was more than compensated by the non-imposition of MAT. Since for

9 Recent Reforms in India’s Corporate Income Tax Regime: Rationale …

221

Table 9.2 Percentage of companies opting for the concessional tax regime and those under earlier tax regime by total income classification (financial year 2019–2020) S.no.

Slabs of Total Income (in crores)

1

2

Percentage of companies who availed 115BAA (%)

Percentage of companies who availed 115BAB (%)

Percentage of companies who availed earlier tax rate of 30% (plus surcharge and CESS) (%)

Percentage of Total Income (under 115BAA) (%)

Percentage of Total Income (under 115BAB) (%)

Percentage of Total Income (under earlier tax rate of 30% plus surcharge and CESS) (%)

Less 8.35 than zero and zero

0.19

91.47

0.00

0.00

0.00

0–1

0.08

76.80

32.55

0.06

67.40

23.12

3

1–10

50.30

0.02

49.68

52.32

0.01

47.67

4

10–50

60.78

0.00

39.22

61.73

0.00

38.27

5

50–100

64.71

0.00

35.29

64.63

0.00

35.37

6

100–500 68.44

0.00

31.56

68.51

0.00

31.49

7

>500

0.00

34.57

62.41

0.00

37.59%

65.43

Source Author’s calculation based on receipt budget, 2022–2023

companies with turnover below Rs. 400 crore the basic tax rate is 25%, section 115BA is essentially superfluous for them. That only leaves sections 115BAA (22%) and 115BAB relevant for smaller firms (15%). Even they have very limited adoption among smaller entities. This phenomenon is examined empirically in the ensuing sections.

9.5 User Cost of Capital Method and Analysis 9.5.1 Methodology The impact of the tax burden on investment decisions has been explored by various neoclassical approaches. According to these theories, the investment must generate sufficient returns to offset depreciation and the real rental rate of capital. In other words, the returns from the investment should be greater than the user cost of capital. User cost of capital (UCC) is specified as (Hall & Jorgenson, 1967; Fabling et al., 2014; Rosenberg & Marron, 2015): UCC =

{1 − τ(Z + k)}(r ∗ + δ) (1 − τ)

222

S. De

where r is the real rental rate, δ is the economic depreciation rate, Z is the present value of depreciation allowance, τ is the effective tax rate, and k is the investment tax credit rate. The intuition behind this formula is straightforward. Consider an investment of Rs. 100. The cost of deploying this investment will be the rate of interest and economic depreciation. But with taxes, the before-tax returns have to (r +δ) . . But along with giving taxes, the company will also get some allowances rise by (1−τ) like depreciation allowance which will bring the user cost of capital down. For each class of asset (plant and machinery, land and building, intangible assets, software, and computer and information technology), the user cost will vary depending on rates of economic depreciation and tax depreciation allowances. These are calculated for each asset class. Prowess data is used to calculate the weightage of various asset classes industry-wise and then compute the empirical industry-wise UCC. The calculations and basic data sources are provided in Annexure 2.5 Our choice of the user cost of the capital method is driven by several factors. Official corporate tax data in India is reported at a very high level of aggregation which makes detailed analysis difficult. Corporate accounts datasets (such as CMIE Prowess) report the accounts according to company law accounting standards which differ from income tax norms. Moreover, they report data for mostly publicly traded companies and therefore miss out on some of the smaller entities. The UCC method allows us to impute broad industry and ownership group characteristics using Prowess data. This is then used to evaluate possible responses to the tax regime based on the values of the UCC. Thus, the method does not need the use of detailed firm-level tax data which is difficult to come by in the Indian context. We examine the distribution of UCCs among Indian firms to address the following questions: 1. To what extent do UCCs differ across sectors such as manufacturing, service, mining, construction and real estate, and electricity for domestic and foreign companies? 2. How have tax reforms such as changes in the rate of corporate tax and depreciation allowances affected these measures? 3. Do foreign firms and domestic firms have different user costs of capital? 4. What are the implications of the new optional tax rates on the user cost of capital? We also lay out the empirical framework, documenting how tax provisions affect the industry-wise user cost of capital. Estimation of the user cost of capital required assumptions regarding which variables are treated as common across all industries and which are industry-specific. For example, different sectors face different costs of borrowing on different creditworthiness. This, in turn, may be correlated with firm or industry-specific factors. But due to data constraints, we restrict ourselves to a common cost of borrowing rate based on RBI policy and SBI prime lending rates. The key assumptions are 1. Inflation rate (CPI), economic depreciation rate, and real rental rate are the same across all firms. 5

The economic and tax depreciation rates, and asset-wise UCCs can be made available on request.

9 Recent Reforms in India’s Corporate Income Tax Regime: Rationale …

223

2. For the real cost of funds, data from two sources are used since one source does not provide data for all financial years (2000–2022). Firstly, we use data from SBI (prime lending rate) from the year 2000–2004. Thereafter, data are collected from RBI (lending rate) which provides a range. The mid-point of the range is assumed to be the real rental rate. 3. Investment tax allowance under section 32AC at the rate of 35% is assumed to be equivalent to an investment tax credit after calculation of the implied tax effect. 4. For analytical tractability and data consistency, we consider only five broad asset categories. Depreciation allowances provided on them under section 32 are mapped as closely as possible. The asset categories are plant and machinery, land and building, intangible assets, software, and computer and information technology. 5. In the Indian tax structure, there are different implied rates due to varying surcharge slabs, in other words multiple τ exist. We assumed that all companies, whether domestic or foreign, are taxed at the highest slab and surcharge rate. 6. There are multiple provisions under sections 115BA, 115BAA, and 115BAB regarding depreciation allowances. To make the UCC calculations tractable, we assumed depreciation allowances are provided at 100%, 50%, and 25%, respectively. This reflects the scaling back of some deductions and limitations on the use of old capital goods. All the variables are then used to compute the UCC based on the asset composition of the industry and ownership. Firm-level data is sourced from CMIE Prowess database, resulting in a sample of 19,105 firms in the year 2000, rising to 31,718 in 2020. For analytical ease, we aggregated companies into five broad industries: mines, manufacturing, service, electricity, and construction and real estate. The manufacturing industry dominates and a number of foreign-ownership companies in mining and electricity are in single digits. We restrict ourselves to firms given in Prowess data which only included widely held firms, as the closely held companies’ (private companies) data is not reliably available. As mentioned earlier, the marginal tax rate depends on the company’s taxable total income which is not readily available. A common treatment is to use the top marginal tax rate of companies since this rate is thought to be the relevant rate for the majority of companies. The top effective tax rate was around 38.5% in FY 1999–2000 and reduced to 34.94% in 2019–2020. In addition, under sections 115BA, 115BAA, and 115BAB effective tax rates are 29.12%, 25.168%, and 17.16%, respectively. Hence, we explore three more alternative tax rates applicable to different sectors as per the provisions of the sections.

224

S. De

0.4 0.35

UCC(FM)

0.3

UCC(MD)

0.25

0.2 0.15 0.1

2021

2019

2020

2018

2017

2016

2015

2014

2012

2013

2011

2010

2008

2009

2007

2006

2005

2004

2003

2002

2001

0

2000

0.05

Fig. 9.6 User cost of capital of manufacturing industry (2000–2021). Source Author’s analysis based on prowess data 0.4 UCC(EF)

0.35

Electricity_ucc(D)

0.3 0.25 0.2 0.15 0.1

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2007

2008

2006

2005

2004

2003

2002

2001

0

2000

0.05

Fig. 9.7 User cost of capital of electrical industry (2000–2021). Source Author’s analysis based on prowess data

9.5.2 Industry-Level User Costs of Capital for Basic Tax Rates Using Prowess data, the industry-specific user cost of capital is calculated for the basic (non-optional) rates. This varies across different assets due to the different tax depreciation allowances and across industries due to differences in asset composition. Based on this, Figs. 9.6, 9.7, 9.8 and 9.9 show the user cost of capital of different industries comparing domestic and foreign ownership across time. This reveals that user cost of capital declined due to the lowering of corporate tax rates.6

6

Due to the small sample size, we do not report mining industry figures. However, the UCCs are available in Annexure 2.

9 Recent Reforms in India’s Corporate Income Tax Regime: Rationale …

225

0.35

Cons_UCC(D) UCC(CF)

0.3

0.25 0.2 0.15 0.1

2021

2019

2020

2017

2018

2016

2014

2015

2012

2013

2010

2011

2008

2009

2006

2007

2004

2005

2002

2003

2001

0

2000

0.05

Fig. 9.8 User cost of capital of construction and real estate industry (2000–2021). Source Author’s analysis based on prowess data 0.4 0.35 0.3

UCC_SERVICE (D)

0.25 0.2 0.15 0.1 0.05 2020

2021

2019

2017

2018

2016

2015

2013

2014

2012

2011

2010

2008

2009

2006

2007

2004

2005

2003

2002

2001

UCC (SF) 2000

0

Fig. 9.9 User cost of capital of service industry (2000–2021). Source Author’s analysis based on prowess data

9.5.3 Impact of Optional Tax Rates Several changes to the tax provisions under optional tax rates for companies affect UCC calculations. These are shown in Figs. 9.10, 9.11, 9.12 and 9.13 comparing UCCs immediately before and after implementation of the new regime. The major change occurs due to the reduction in depreciation allowances, removal of set-off of loss, unabsorbed depreciation carried forward, and irreversibility. Due to complex provisions and data constraints, only the first aspect can be modeled explicitly. For this, we assume 100%, 50%, and 25% of depreciation allowance availability under the respective sections 115BA, 115BAA, and 115BAB. The results show that the new tax regime does not provide a significant advantage over the old regime. UCCs for concessional rates are paradoxically higher than the normal rate even without taking into consideration the non-availability of unabsorbed depreciation carried forward and irreversibility. This is especially true for

226

S. De

Manufacturing (Domestic) 0.24 0.22 0.20 0.18 0.16

2017

2018

2019

2020

No tax incentive Manufacturing (Domestic)

Section 115BA

Section 115BAA

Section 115BAB

2021

Fig. 9.10 UCCs of concessional rates and normal rates for domestic manufacturing industry. Source Author’s analysis based on prowess data

Construction and Real Estate (domestic) 0.20 0.19 0.18 0.17 0.16

2017

2018

2019

2020

No tax incentive Construction and Real estate (Domestic)

2021 Section 115BAA

Fig. 9.11 UCCs of concessional rates and normal rates for construction and real estate (domestic). Source Author’s analysis based on prowess data

Service (Domestic) 0.20 0.19 0.18

0.17 0.16

2017

2018

2019

No tax incentive Services (Domestic)

2020

2021

Section 115BAA

Fig. 9.12 UCCs of concessional rates and normal rates for service (domestic). Source Author’s analysis based on prowess data

section 115BAB designed particularly for small and younger start-ups. Given our assumptions, section 115BAB providing a 15% tax rate to manufacturing start-ups has nearly the same UCC as the non-concessional 30% basic rate (Fig. 9.10). For manufacturing entities, section 115BAA with a 22% concessional rate for all types of companies has a UCC which is the highest. section 115BA yields the lowest UCC, but

9 Recent Reforms in India’s Corporate Income Tax Regime: Rationale …

227

Electrical (Domestic) 0.26 0.25 0.24

0.23 0.22 0.21

2017

2018

2019

No tax incentive Electricity (Domestic)

2020

2021

Section 115BAA

Fig. 9.13 UCCs of concessional rates and normal rates for electrical (domestic). Source Author’s analysis based on prowess data

this rate is identical to the lower rate of 25% for sub-Rs. 400 crore turnover companies. Therefore, there is no great incentive for smaller firms to give up tax incentives and switch to the simplified system. Non-manufacturing entities only have a choice between the basic rates and section 115BAA. For Construction (Fig. 9.11) and Electrical (Fig. 9.12) industries, section 115BAA UCCs are higher than the basic rates. Only for Service industries (Fig. 9.13) are the two UCCs nearly identical.

9.5.4 Loss-Making Companies It is empirically complex to take losses into account in the UCC calculations. Nevertheless, Figs. 9.14 and 9.15 show that companies earning negative profits over a period of ten years continuously are overwhelmingly young and small. The number shrinks monotonically from 218 for 10–20-year-old companies to 20 for 40–50-yearold companies. In a similar manner, classifying by turnover, 233 small companies with turnover from Rs. 0–5 crores are loss-making, while only 13 larger companies with turnover from Rs. 15–20 crores incurred sustained losses. This implies that the lack of ability to carry forward losses would be a major disincentive for small and young firms to switch to the optional simplified tax regimes. For such entities, the non-imposition of MAT would not be an incentive either since due to their zero or negative book profits they would not be liable to pay MAT.

9.6 Conclusions, Caveats, and Policy Recommendations Tax policies have a significant effect on business growth. In a recent innovative policy reform, India’s corporate income tax system was revamped with the option of availing lower rates in lieu of giving up complex deductions. India’s tax rate reforms

228 Fig. 9.14 Number of loss companies by age. Source Author’s analysis based on prowess data

S. De

Number_of_firms(Age) 250 200 150 100 50 0

Fig. 9.15 Number of loss companies by sales. Source Author’s analysis based on prowess data

0-10

10-20

20-30

30-40

40-50

Number_of_firms (Turnover) 300 200 100 0

are placed in the context of the move to simplify the tax system in the 1990s, the subsequent dilemma of dealing with zero-tax companies leading to instituting the minimum alternate tax (MAT), and the continuance of lower effective tax rates for larger companies. This provided the rationale for moving toward a simpler tax regime with lower tax rates but fewer deductions. In theory, such a move should lower the cost of investment and bring gains from procedural simplicity. However, data from the government revealed a conundrum wherein larger companies have opted more for the simplified lower optional rates while smaller ones appear reticent in switching to the optional lower tax regime. Also, the overall adoption of the simpler scheme was low. This paper explored this issue using empirical methods. The user cost of the capital approach is used to examine the economically relevant tax impact across various sectors and ownership types. The UCC method allows us to impute broad industry and ownership group characteristics using Prowess data. This is then used to evaluate possible responses to the tax regime based on the values of the

9 Recent Reforms in India’s Corporate Income Tax Regime: Rationale …

229

UCC. Thus, the method does not need the use of detailed firm-level tax data which is difficult to come by in the Indian context. The results indicate that in terms of user cost, the various lower tax options are not attractive, and under certain situations may be even worse for younger and smaller companies. Moreover, due to its irreversible nature and inability to carry forward losses, start-ups and small companies may be hesitant in committing to the scheme. To make matters worse, the complexity of the provisions may also be deterring adoption. In light of the analysis, certain policy options are suggested to improve the scheme so as to achieve the laudable objective of implementing a simple tax regime with lower rates and minimal deductions: 1. To make the scheme simple, there should be just one optional lower tax rate say at 20% applicable to both manufacturing and non-manufacturing entities. The diverse lower rates of 25%, 22%, and 15% could be merged into this. 2. The inability to claim depreciation allowances and loss carry forward is a major disincentive, especially for younger and smaller companies like start-ups. These should be re-instituted in the simplified scheme. 3. The tax system should be made simple and not overburdened with multiple policy objectives. Policymakers should take steps to provide comprehensive incentives to innovative start-ups outside the tax system. Such measures could include productivity-linked incentive (PLI) schemes, provision of seed-finance, and appropriate incubation facilities. Certain caveats are in order. Due to the complexity of tax laws, certain simplifying assumptions were relied upon. These include clubbing of asset classes and industry groups, assumptions regarding depreciation rates for the new regime, and use of common real rates of return. However, individual companies may have unique circumstances leading to deviations from these assumptions. Also, more data is needed to precisely identify start-ups. Despite these limitations, our analysis provides new insights. It highlights that due to the inherent complexities of the legislation, the new optional tax regime may not be favorably impacting the intended taxpayers. In effect, the optional tax provisions intended to simplify the tax regime for small firms and start-ups inadvertently became a tax cut for larger and more established firms. Future research could concentrate on using additional data as it emerges. This also highlights the need for the availability of corporate and tax data to cover not only publicly held companies (as in Prowess) but also privately held ones.

9.7 Annexure 1 Income-tax rates applicable in the case of domestic companies for the assessment years 2021–2022 and 2022–2023 are as follows7 :

7

Income Tax Department, Government of India. https://www.incometaxindia.gov.in/_layouts/15/ dit/mobile/viewer.aspx?path=https://www.incometaxindia.gov.in/charts++tables/tax+rates.htm& k&IsDlg=0.

230

S. De Assessment year 2021–2022 assessment Year 2022–2023

• Where its total turnover or gross 25% receipt during the previous year 2018–19 does not exceed Rs. 400 crore

NA

• Where its total turnover or gross NA receipt during the previous year 2019–20 does not exceed Rs. 400 crore

25%

30%

• Any other domestic company

30%

Add: (a) Surcharge: The amount of income tax shall be increased by a surcharge—at the rate of 7% of such tax, where the total income exceeds one crore rupees but not exceeding ten crore rupees and at the rate of 12% of such tax, where total income exceeds ten crore rupees.8 (b) Health and Education Cess: The amount of income tax and the applicable surcharge, shall be further increased by health and education cess calculated at the rate of 4% of such income tax and surcharge. As mentioned earlier, the separate tax rates for small and large companies bring about vertical equity and remedy the persistently lower effective tax rates that larger companies managed to maintain when uniform domestic corporate tax rates prevailed. 1. Special tax rates applicable to a domestic company The special income-tax rates applicable in the case of domestic companies for the assessment years 2021–2022 and 2022–2023 are as follows:

8

Assessment year 2021–2022 (%)

Assessment year 2022–2023 (%)

• Where it opted for section 115BA

25

25

• Where it opted for section 115BAA

22

22

• Where it opted for section 115BAB

15

15

The surcharge shall be subject to marginal relief, which shall be as under. (i) Where income exceeds Rs. 1 crore but not exceeding Rs. 10 crore, the total amount payable as income-tax and surcharge shall not exceed total amount payable as income-tax on total income of Rs. 1 crore by more than the amount of income that exceeds Rs. 1 crore. (ii) Where income exceeds Rs. 10 crore, the total amount payable as income-tax and surcharge shall not exceed total amount payable as income-tax on total income of Rs. 10 crore by more than the amount of income that exceeds Rs. 10 crore.

9 Recent Reforms in India’s Corporate Income Tax Regime: Rationale …

231

Surcharge: The rate of surcharge in case of a company opting for taxability under Section 115BAA or Section 115BAB shall be flat 10% irrespective of the amount of total income. Health and Education Cess (HEC): The amount of income tax and the applicable surcharge shall be further increased by health and education cess calculated at the rate of 4% of such income tax and surcharge. MAT: The domestic company that has opted for the “special taxation regime” under sections 115BAA and 115BAB is exempted from the provision of MAT. However, no exemption is available in the case where section 115BA has been opted. In that case, the provisions of Minimum Alternate Tax (MAT) apply, tax payable cannot be less than 15% (+HEC) of “Book profit” computed as per section 115JB. However, MAT is levied at the rate of 9% (plus surcharge and cess as applicable) in the case of a company, being a unit of an International Financial Services Centre and deriving its income solely in convertible foreign exchange. The special tax rates were first instituted from AY 2017–2018 with the introduction of section 115BA. It provides for a tax rate of 25%. Subsequently, it was extended to domestic companies set up and registered on or after March 1, 2016, that are engaged in the business of manufacture or production of any article or thing and research in relation to, or distribution of, such article or thing manufactured or produced by it. No deductions are available under section 10AA (special provisions in respect of newly established Units in Special Economic Zones) or clause (iia) of sub-section (1) of section 32 (additional depreciation) or section 32AC (investment in new plant or machinery) or section 32AD (investment in new plant or machinery in notified backward areas in certain States) or section 33AB (tea development account, coffee development account, and rubber development account) or section 33ABA (Site Restoration Fund for petroleum or natural gas) or sub-clause (ii) or sub-clause (iia) or sub-clause (iii) of sub-section (1) or sub-section (2AA) or sub-section (2AB) of section 35 (expenditure on scientific research) or section 35AC (expenditure on eligible projects or schemes) or section 35AD (deduction in respect of expenditure on specified business) or section 35CCC (expenditure on agricultural extension project) or Section 35CCD (expenditure on skill development project) or under any provisions of Chapter VI-A under the heading “C.—Deductions in respect of certain incomes” other than the provisions of section 80JJAA (deduction in respect of employment of new employees). Carrying forward of loss related to the above non-permissible deductions is not allowed. Depreciation under section 32 is available for additional depreciation under clause (iia). Section 115BAA came into effect from AY 2020–2021. It provides for a tax rate of 22%. This applies to any domestic company. No deduction is permissible under section 10AA (special provisions in respect of newly established Units in Special Economic Zones) or clause (iia) of sub-section (1) of section 32 (additional depreciation) or section 32AD (deduction in respect of expenditure on specified business) or section 33AB (tea development account, coffee development account, and rubber development account) or section 33ABA (Site Restoration Fund for petroleum or natural gas) or sub-clause (ii) or sub-clause (iia) or sub-clause (iii)

232

S. De

of sub-section (1) or sub-section (2AA) or sub-section (2AB) of section 35 (expenditure on scientific research) or section 35AD (deduction in respect of expenditure on specified business) or section 35CCC (expenditure on agricultural extension project) or section 35CCD (expenditure on skill development project) or under any provisions of Chapter VI-A other than the provisions of section 80JJAA (deduction in respect of employment of new employees) or section 80 M (deduction in respect of certain inter-corporate dividends). Carrying forward of loss or depreciation related to the above non-permissible deductions is not allowed. Section 115BAB is also in effect from AY 2020–2021. This applies to any domestic company that has been set up and registered on or after October 1, 2019, and has commenced manufacturing or production on or before March 31, 2023, and (i) the business is not formed by splitting up, or the reconstruction, of a business already in existence: (ii) does not use any machinery or plant previously used for any purpose. The company should be engaged in the business of manufacture or production of any article or thing and research in relation to, or distribution of, such article or thing. The provision cannot be availed by companies engaged in (i) (ii) (iii) (iv) (v) (vi)

development of computer software in any form or in any media; mining; conversion of marble blocks or similar items into slabs; bottling of gas into cylinder; printing of books or production of a cinematograph film; or any other business as may be notified by the Central Government.

It provides for a tax rate of 15%. But where total income includes any income, which has neither been derived from nor is incidental to the manufacturing or production of an article or thing and in respect of which no specific rate of tax has been provided separately, such income shall be taxed at 22% and no deduction in respect of any expenditure or allowance shall be allowed in computing such income. No deduction is permissible under section 10AA (special provisions in respect of newly established Units in Special Economic Zones) or clause (iia) of sub-section (1) of section 32 (additional depreciation) or section 32AD (deduction in respect of expenditure on specified business) or section 33AB (tea development account, coffee development account, and rubber development account) or section 33ABA (Site Restoration Fund for petroleum or natural gas) or sub-clause (ii) or sub-clause (iia) or sub-clause (iii) of sub-section (1) or sub-section (2AA) or sub-section (2AB) of section 35 (expenditure on scientific research) or section 35AD (deduction in respect of expenditure on specified business) or section 35CCC (expenditure on agricultural extension project) or section 35CCD (expenditure on skill development project) or under any provisions of Chapter VI-A other than the provisions of section 80JJAA (deduction in respect of employment of new employees) or section 80M (deduction in respect of certain inter-corporate dividends). Carrying forward of loss or depreciation related to the above non-permissible deductions is not allowed.

9 Recent Reforms in India’s Corporate Income Tax Regime: Rationale …

233

To clarify, section 115BA which came into effect from AY 2017–2018 for domestic manufacturing companies was set up and registered on or after March 1, 2016. It offers a lower tax rate of 25% in lieu of giving up certain deductions and the ability to carry forward and set off losses. Section 115BAA came into effect from AY 2017–2018 for any domestic company. It offers lower tax rates of 22% in lieu of giving up certain deductions and the ability to carry forward and set-off losses and claim depreciation in relation to the non-permissible deductions. Section 115BAA which came into effect from AY 2017–2018 for domestic manufacturing companies (not formed by splitting up, or the reconstruction, of existing business and not using any previously used machinery or plant) was set up and registered on or after October 1, 2019, commencing manufacturing or production on or before March 31, 2023. It offers a lower tax rate of 15% in lieu of giving up certain deductions and the ability to carry forward and set off losses and claim depreciation in relation to the non-permissible deductions. Income from non-manufacturing activities of such companies is to be taxed at 22% and no deduction in respect of any expenditure or allowance shall be allowed in computing such income.

9.8 Annexure 2 Variables for Measuring User Cost of Capital (See Tables A1 and A2) Variable

Source

1. Real rental rate (r)

Prime lending rate of SBI (2000–2004) and RBI (2005–2022)

2. Economic depreciation rate (δ)

Hulten and Wykoff, US Bureau of Economic Analysis

3. Present value of depreciation allowance Z

From section 32 of Income Tax Act, 1961, depreciation rate (present value based on CPI inflation rate)

4. Effective tax rate (τ)

Tax rates are highest marginal slab rate as per Income Tax Act, 1961

5. Investment tax credit rate (K)

As per section 32AC investment tax allowance made equivalent to the tax credit

234

S. De

Table A1 UCC of domestic companies Year

Cons_UCC (D)

Electricity

UCC (MD)

Mining

Service

2000

0.276414

0.304322

0.294133

0.305606

0.281193

2001

0.276719

0.299664

0.290973

0.301184

0.281835

2002

0.258578

0.282672

0.274217

0.280561

0.260921

2003

0.25414

0.277384

0.269819

0.275631

0.25668

2004

0.25165

0.271097

0.264459

0.270999

0.25175

2005

0.260732

0.277267

0.270963

0.277576

0.261194

2006

0.262829

0.29729

0.291078

0.296645

0.277161

2007

0.295663

0.329419

0.320262

0.326622

0.308061

2008

0.305066

0.336401

0.326107

0.332299

0.311159

2009

0.306882

0.33734

0.326752

0.330583

0.309569

2010

0.289204

0.322952

0.306244

0.276799

0.290858

2011

0.221988

0.254926

0.240741

0.210744

0.222526

2012

0.2417

0.275505

0.260853

0.231319

0.24057

2013

0.22802

0.27048

0.256101

0.223059

0.233662

2014

0.221873

0.264451

0.24905

0.214564

0.227381

2015

0.218051

0.265247

0.248459

0.212434

0.220609

2016

0.211294

0.266121

0.245923

0.187626

0.208873

2017

0.187514

0.244122

0.2219

0.161619

0.184599

2018

0.182968

0.244529

0.220669

0.157602

0.184439

2019

0.187136

0.251725

0.22825

0.164633

0.191057

2020

0.183208

0.242852

0.219466

0.154531

0.183878

2021

0.171705

0.22916

0.204733

0.142483

0.174569

Source Author’s analysis based on prowess data

9 Recent Reforms in India’s Corporate Income Tax Regime: Rationale …

235

Table A2 UCC of foreign companies Year

Electricity

UCC (FM)

UCC (SF)

UCC (MIF)

UCC(CF)

2000

0.333346

0.318095

0.292805

0.335115

0.307545

2001

0.322394

0.311468

0.29139

0.32653

0.287933

2002

0.30152

0.281737

0.257611

0.285463

0.262373

2003

0.2958

0.273736

0.257186

0.290294

0.258874

2004

0.291875

0.270011

0.247673

0.286311

0.250938

2005

0.28763

0.278131

0.272074

0.284514

0.254908

2006

0.324188

0.312275

0.301807

0.309522

0.288334

2007

0.362119

0.347001

0.3396

0.25368

0.307418

2008

0.367637

0.350784

0.338674

0.256703

0.311605

2009

0.298786

0.353618

0.344346

0.371476

0.319933

2010

0.347828

0.342432

0.336576

0.36874

0.312111

2011

0.271504

0.258146

0.22552

0.291425

0.243212

2012

0.294591

0.282705

0.230635

0.180771

0.255711

2013

0.293486

0.277684

0.233131

0.175448

0.280984

2014

0.283178

0.258931

0.231939

0.170167

0.247367

2015

0.283129

0.262167

0.219484

0.170297

0.244898

2016

0.28982

0.262493

0.197455

0.170461

0.253577

2017

0.259724

0.23762

0.186772

0.1462

0.208207

2018

0.260172

0.23645

0.187807

0.144608

0.199069

2019

0.266367

0.244193

0.224458

0.151913

0.203191

2020

0.255235

0.233479

0.209827

0.141894

0.200411

2021

0.240387

0.217778

0.200556

0.128143

0.181185

Source Author’s analysis based on prowess data

References Abel, A. B. (1983). Optimal investment under uncertainty. The American Economic Review, 73(1), 228–233. Acharya, S. (2005). Thirty years of tax reform in India (Vol. 40). Ahluwalia, M. S. (2022). Economic reforms in India since 1991: Has gradualism worked? Alagappan, S. M. (2019). Indian tax structure-an analytical perspective. International Journal of Management (IJM), 10(3), 36–43. https://d1wqtxts1xzle7.cloudfront.net/60342000/004libre.pdf?1566295806=&response-contentdisposition=inline%3B+filename%3DINDIAN_ TAX_STRUCTURE_textendash_AN_ANALY.pdf&Expires=1693150629&Signature=C9tr10 elW9I3vhcYlAyWo037VsZx8uba2hjDTN-bdTquT8ld-7ZrsKvyLuXXt40CLBCUjY5gXJDV 67ThaZcMiqvIdkb77DF6FOoQ8DSX7XtDFZwfaKjv7RVMLLYcnONvkXj36ErtuG6MFq sOagXF6b9byfLVgbqLr11h3yAOzCbzUP4hA9j1XxFCzm3K2sJOrOh2q~RmxGTWcywez4 C~qvaD16U42xn38Pf6m0-MxrnYcxpL9un0nkBR7Hh2V-Y8mrTT1D8F~mr1i130PQ__& Key-Pair-Id=APKAJLOHF5GGSLRBV4ZA Chirinko, R. S., Fazzari, S. M., & Meyer, A. P. (1999). How responsive is business capital formation to its user cost?: An exploration with micro data. Journal of Public Economics, 74(1), 53–80

236

S. De

Devereux, M. P., Griffith, R., & Klemm, A. (2002). Corporate income tax reforms and international tax competition. Economic Policy, 17(35), 449–495. Dwenger, N. (2014). User cost elasticity of capital revisited. Economica, 81(321), 161–186. Fabling, R., Gemmell, N., Kneller, R., & Sanderson, L. (2014). Estimating firm-level effective marginal tax rates and the user cost of capital in New Zealand. Guha, A. (2007). Company size and effective corporate tax rate: Study on Indian private manufacturing companies. Economic and Political Weekly, 42(20), 1869–1874. Hall, R. E., & Jorgenson, D. W. (1967). Tax policy and investment behavior. The American Economic Review, 57(3), 391–414. Hayashi, F. (1982). Tobin’s marginal q and average q: A neoclassical interpretation. Econometrica: Journal of the Econometric Society, 213–224. Jorgenson, D. W. (1963). Capital theory and investment behavior. The American Economic Review, 53(2), 247–259. Rao, M. G., & Rao, R. K. (2006). Trends and issues in tax policy and reform in India. Rosenberg, J. W., & Marron, D. B. (2015). Tax policy and investment by startups and innovative firms. Available at SSRN 2573259. Samantara, R. (2021). Tax reforms In India : A critical analysis-journal ijmr.net.in(UGC Approved). http://ijmr.net.in Sankarganesh, K., & Shanmugam, K. R. (2021). Effect of corporate income tax on investment decisions of Indian manufacturing firms. Journal of the Asia Pacific Economy, 1–20.

Chapter 10

The Anatomy of Public Debt Reductions: Case of India Prachi Mishra and Nikhil Patel

Abstract India has witnessed an unprecedented surge in sovereign debt on the back of the pandemic. We outline the costs and risks associated with high debt in India, drawing on experiences across countries as well as India’s own past. We conclude with a discussion of the benefits of reducing debt and possible scenarios that could achieve a sizable reduction over the next decade.

India’s sovereign debt reached unprecedented levels in 2020, partly driven by the policy response to COVID-19, but also by low growth and high interest rates (Fig. 10.1). Some have argued that high levels of debt may be less concerning in an environment of low interest rates (Blanchard, 2022). But there is also a significant body of evidence that points to several mechanisms through which high levels of sovereign debt can have negative effects on the economy. Against this background, we document stylized facts on the recent evolution of sovereign debt and fiscal deficits in India and ask the following questions: What are the costs of high debt levels in India? Are there any silver linings? And what lies ahead? We do so in part by analyzing macroeconomic outcomes during and after debt “surges”, “stabilization” and “reduction” periods in India and other countries This paper is a revised version of the article published by the Center for the Advanced Study of India based on a larger project at Systemic Issues Division in the Research Department at the IMF, to understand the evolution of public debt in the post-pandemic world across countries with a team including Sakai Ando, Josef Platzer, Adrian Peralta Alva, and Andrea Presbitero. We thank Santiago Acosta-Ormaechea, Elif Ceren, Nada Choueiri, Adrian Peralta-Alva, Dinar Prihardini, Francisco Roch and Jarkko Turunen for helpful comments, and Swapnil Aggarwal, Chenxu Fu and Manzoor Gill for excellent research support. All errors remaining are our own. Views expressed here are those of the authors and do not necessarily represent the views of the IMF, its Executive Board, or IMF management. P. Mishra (B) · N. Patel International Monetary Fund, Washington, USA e-mail: [email protected] N. Patel e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 D. K. Srivastava and K. R. Shanmugam (eds.), India’s Contemporary Macroeconomic Themes, India Studies in Business and Economics, https://doi.org/10.1007/978-981-99-5728-6_10

237

238

P. Mishra and N. Patel

Fig. 10.1 Surge in general government debt in India is unprecedented, driven partly by unprecedented rise in deficit. Source Global debt database, IMF world economic outlook database, actual data till FY 2020, WEO projection (for deficit) for 2021

and ask whether past experiences with surges and reductions shed light on different policy options and the tradeoffs for India during the post-pandemic recovery. The Covid-induced surge in debt in India was unique compared to its own history, but also bigger than that for the average emerging market (EM) economy. The drivers of the debt surge were different too. Both fiscal expansion and the collapse in growth played a proportionately larger role in India compared with the average EM, even as higher inflation played a greater role in reducing debt in India. Notwithstanding the high level of sovereign debt, there are a few silver linings for India. The share of sovereign debt held by foreigners—an important predictor of crises in the literature—is low. Moreover, although global waves of debt surges have been followed by restructuring or default, India has not had any such episode so far. Furthermore, long-term real rates remain low in India, comparable to the median EM. That said, we find substantial heterogeneity across countries. While India was closer to the 25th percentile during the last decade, it has now caught up with the median. We document substantial costs of high debt. A major one is foregone resources on account of strikingly high interest payments, which at almost 30% of overall revenues during Covid are close to three times higher for India than the typical EM (Fig. 10.2). High expenditures on interest payments reduce the resources available for countercyclical fiscal policies in the event of negative shocks such as Covid, as well as for social spending in critical areas such as health and education where India’s public spending remains markedly below peers. Indeed, our analysis suggests that business cycle fluctuations explain a smaller fraction of the variation in debt in India compared with peers, reaffirming the limits to countercyclical fiscal policy on account of high debt levels.1 Simple calculations suggest that reducing India’s interest payments to revenue to the EM average of 10% would release resources of close to INR 6– 8 trillion, a figure comparable to India’s pre-Covid general government education expenditure, and about three times its health expenditure. 1

This inference is based on a five variable structural vector autoregression identifies using narrative sign restrictions (see Ando et al. 2022).

10 The Anatomy of Public Debt Reductions: Case of India Interest Payments to Revenues (%)1

239

Health Spending (% of GDP)2

35 30 25 20 15 10

AEs

EMs

LIDCs

21

20

19

18

17

16

15

14

13

11

12

09

10

0

2008

5

India

Fig. 10.2 High interest payments, less space for countercyclical policies, crowding out of social spending. 1 PPP—GDP weighted averages across country groups. Source IMF world economic outlook database. 2 Source World bank world development indicators, 2019

Another cost of high public debt in India is its impact on borrowing costs. Although real rates in India are low and in line with the median EM, we find that they have increased over time, and that the elasticity of borrowing costs to a unit increase in debt is higher for India than the typical EM (Fig. 10.3). For example, on average, an increase in debt to GDP by 1% point (pp) increases long-term borrowing costs by 0.19 pp in India, while for a median EM it increases by only 0.01 pp. Finally, public debt exemplifies an important factor in the assessments of rating agencies too, where India’ debt and deficits stand out as being markedly higher than similarly rated peers. In order to understand where to go from here, we look at India’s own history and also draw on cross-country experiences. Since 1913, India has had nine episodes of debt surges, five episodes of reduction, and six episodes of debt stabilizations. Surges have typically ended in stablizations in India, whereas in an average EM, 75% of surges end in reductions; in other words, India has been able to sustain debt at high levels without default or restructuring. Across reduction episodes, India reduced debt ratios by 2 pp per year, compared to more than double the figure for the average EM. We also find that debt surge episodes are associated with worse macroeconomic outcomes—low economic growth and public investment—compared with debt reduction episodes. Moreover, cross-country evidence suggests that the greater the magnitude of the rise in debt, and longer lasting the episode, the greater the associated reduction in growth around the surge.2 How much debt could India reduce? One way to approach this question is to look at interest payments and additional budgetary resources that could be generated by lower sovereign borrowing. For example, getting interest payments down to say 22% (still much higher than the EM average of 10%) would require reducing the debt ratio to 70%, bringing it closer to the median for similarly rated peers.

2

Note that these are associations and we do not intend to make any causal claims here.

240 CDS spreads and Debt 1

P. Mishra and N. Patel Long term real rates more sensitive to Debt levels in India2

Fig. 10.3 High spreads, especially since the pandemic; and high elasticity of borrowing costs to debt. 1 Each arrow represents a country. The beginning of the arrow represents the debt and spread pre pandemic, and the end of the arrow their values in 2022). Source IMF world economic outlook database, Bloomberg (for CDS spreads). 2 Based on an unbalanced panel of 19 EMs. Simple scatter plot not controlling for fixed effects. Numbers reported in the text are based on regressions with country fixed effects. Source IMF world economic outlook database, IMF international financial statistics (IFS) database, Bloomberg, Haver, Jorda-Schularick-Taylor Macro history database, OECD, Mauro et al (2015), Global debt database

What could be a possible path, and how long would it take to get there? The higher the growth rate and the lower the borrowing costs, the lower the need for fiscal adjustment. Simulation exercises suggest that if we assume constant values for real GDP growth rate at 7% and real rate at 2% in line with the IMF World Economic Outlook (WEO) assumptions, a general government primary and fiscal deficit of lower than 1.7% and 5.9% of GDP respectively would be needed every year even to reduce debt ratios to 70% in the next 10 years (and interest payments to 22% of revenues). This would require a sharp adjustment when compared with the FY 2022–2023 primary and fiscal deficit at a projection of 4.5% and 9.9% respectively according to the World Economic Outlook. Importantly, the higher the growth rate and the lower the interest rate, lesser would be the required adjustment. For example, a growth rate of 9% or a real rate of 0% would open up more space with a primary deficit of more than 3% of GDP instead of 1.7% still ensuring the same debt reduction (Fig. 10.4). While the calculations above assume constant primary and fiscal deficits, allowing for some transitional dynamics and smoothing the adjustment path, we report in Fig. 10.5, illustrative scenarios for debt and fiscal consolidation for India over the next five years. Indeed, evidence across emerging economies suggests that primary balance consolidations outside of recessions could, in fact, be successful in reducing debt, and do not tend to be detrimental to growth as multiplier effects roughly balance the positive impulse from other channels such as higher confidence (Ando et al., 2022).

10 The Anatomy of Public Debt Reductions: Case of India

241

Fig. 10.4 More India grows out of debt, lesser the required adjustment. Possible Scenarios to reduce debt by 20% points in 10 years. Notes Simulations based on debt dynamics accounting identities in Escolano (2010)

Total Fiscal Deficit

Debt

Per cent of GDP

Per cent of GDP

Per cent of GDP

14

95

6

12

90

4

10

2

8

75

0

6

70

85

2026

2027

2025

2024

2023

2022

2021

80

2020

2020 2021 2022 2023 2024 2025 2026 2027

8

2020 2021 2022 2023 2024 2025 2026 2027

Primary Fiscal Deficit

WEO Alternate Scenario with Added Consolidation Alternate Scenario with Added Consolidation and Higher Growth

Fig. 10.5 Possible trajectories for deficits and debt. Notes The alternate scenario with added consolidation assumes a primary deficit of 0.5% points below the WEO projection for years 2022–2027. The alternate scenario with higher growth and added consolidation assumes, in addition to the above consolidation, a 9% growth rate for years 2022–2027. The alternate scenario with only consolidation reduces debt to 80% by 2030, whereas the scenario with consolidation and high growth reduces debt to 68%. Sources IMF world economic outlook database, author calculations

References Ando, S., Mishra, P., Patel, N., Platzer, J., Alva, A. P., & Presbitero, A. (2022). Anatomy of public debt surges and reductions. Systemic Issues Division, Research Department IMF, memo. Blanchard, O. (2022). Fiscal policy under low interest rates. MIT Press. Escolano, J. (2010). A practical guide to public debt dynamics, fiscal sustainability, and cyclical adjustment of budgetary aggregates. No. 2010/002. International Monetary Fund. Mauro, P., Romeu, R., Binder, A., & Zaman, A. (2015). A modern history of fiscal prudence and profligacy. Journal of Monetary Economics, 76, 55–70.

Chapter 11

Measuring Tax Impact on Corporate Dividend Behavior in India J. V. M. Sarma

Abstract The non-government joint stock corporations have been dominating the Indian economic scene for quite some time now. Yet, the extent of the impact of taxation on corporate behavior remains ambiguous. Whatever be the objective, the way taxation is used for the purpose is simply to alter the relative tax burden between dividends and retained profits of companies. The differential tax burden can be injected through numerous elements in a tax system either at the company level or at the shareholders’ level. In India also the Income tax system contained several elements of tax differentiation aimed at discouraging excessive dividend payments by public limited companies. The present time-series empirical study is an objective attempt to analyze the tax differentiation underlying the Indian income tax system as well as to measure the response of public limited companies to such differentiation. Starting with a relatively simpler version of the general model, the coefficients are estimated by different methods with a view to identifying the correct version of the model. The Cobb-Douglas version of the model does not fit the data satisfactorily. Attempts to correct the lagged dependent bias, and serial correlations have not improved the situation. On the other hand, the CES version, though proved to be a better specification, yielded coefficient estimates which seem to be less stable. Specifically, the manufacturing sector proved to be very sensitive to tax changes. The quantification of the effect by means of simulating the best estimated equations for each group show that much of the effect ln India has been due to the adaptation of the ‘Classical’ system. The effect was highest in 1960–1968 during which time the Classical system was just introduced, the excess dividends taxes were levied, and the rates of personal income taxes had been higher compared to the other sub-periods. An important indication is that the effect of excess dividends taxes by itself is very low, compared to

This paper is an updated version of my original paper—Sarma, J. V. M., “Taxation and Corporate Dividend Behaviour in India” (1982). Discussion Papers. 414. elischolar.library.yale.edu/ egcenter-discussion-paper-series. J. V. M. Sarma (B) Centre for Public Finance, Madras School of Economics, Chennai, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 D. K. Srivastava and K. R. Shanmugam (eds.), India’s Contemporary Macroeconomic Themes, India Studies in Business and Economics, https://doi.org/10.1007/978-981-99-5728-6_11

243

244

J. V. M. Sarma

that of personal income tax, which is not fully in agreement with the prevalent view regarding these taxes.

11.1 Introduction Private sector has been playing an important role in Indian economy. The nongovernment joint stock corporations have been dominating the Indian economic scene for quite some time now. Nearly two-thirds of the total capital in the private sector is in the hands of corporations and roughly three-fourths of the output generated in the private sector is by corporations. Their contributions to total savings, capital formation as well as to employment in the economy are by no means negligible. To regulate and control the private corporate sector, the government uses taxation, particularly income taxation. The subtle and persuasive, rather than coercive, nature of taxation offers a wide scope for effective controlling of the sector without disturbing the institutional set-up. Yet, the extent of the impact of taxation on corporate behavior remains ambiguous. Studies focusing on the market distortions caused by income taxation on corporate behavior are fewer and necessary academic attention is overdue. This study is an attempt in this direction. Considering the complex nature of corporate behavior because of numerous management decisions at firm level and the equally diverse tax structure, the study concentrates on tax effects on dividend behavior of the corporations. Of all the policy decisions taken at the corporate level, few are as important as the dividend decision. Dividend decisions have implications not only at the level of the individual firms but at the macroeconomic level as well. Indeed, the culmination of all the objectives of a modern joint stock company is to generate a steady stream of dividends for its shareholders. Higher and regular dividend payments are sure to enhance the market value of the firm and the reputation of its management. On the other hand, such a policy may mean less availability of internal funds and more dependence on external sources for reinvestment and expansion purposes. Thus, while determining dividend payments a prudent management strikes a balance between shareholders’ preferences and firm’s longterm interests while safeguarding their control of the firm. From the point of view of the economy, dividend policies of individual firms when added together, play a significant role in determining overall rates of savings and investment as well as patterns of flow of funds in the economy (Bandyopadhyay, 2008). Whatever the objective, the way taxation is used for the purpose is simply to alter the relative tax burden between dividends and retained profits of companies. The differential tax burden can be injected through numerous elements in a tax system either at the company level or at the shareholders’ level. The extent of such tax differentiation as well as the designing of tax system for the purpose, differ from country to country depending upon their specific needs and circumstances.

11 Measuring Tax Impact on Corporate Dividend Behavior in India

245

At one extreme lies’ Classical system’ under which dividends are taxed twice, once in the hands of companies as profits and later in the hands of shareholders as personal incomes. The differential tax burden can reach a maximum under this system. The Classical system is currently followed by countries such as Australia, Denmark, Luxembourg, Netherlands, Spain, Switzerland, and so on (IBFD, 1972), (OECD, 1973). At the other extreme lies the ‘Full Integration system’ where the differential tax burden is fully neutralized. This system is as utopian as ‘pure competition’ and is not practiced anywhere though attempts are made in Canada, Greece, and West Germany. Between these two extremes several systems are possible and exist by partially neutralizing the differential tax burden. The partial neutralization is achieved either at the company level or at the shareholders’ level. At the company level it is usually affected by following a ‘Split rate system’* under which distributed profits are taxed at company tax rates different from undistributed profits. This practice is found in Austria, Finland, West Germany, Japan, and Norway. The ‘Imputation system’ where a credit is given to shareholders for taxes paid at company level is in force in France, Ireland, Italy, United Kingdom, Belgium, and Canada. In India also the Income tax system contained several elements of tax differentiation aimed at discouraging excessive dividend payments by public limited companies. Till 1959–60, an ‘Imputation’ type of Income tax system was adopted which was replaced later by ‘Classical system’ and thereby increasing the over-all tax discrimination against dividends. Also, from time to time, additional taxes were levied at company level to accentuate the relative tax burden on dividends.

11.1.1 Recent Developments Pertaining to the Dividend Taxation There have been a few developments pertaining to the dividend taxation in recent times. An amendment to Income-Tax Act was made to the effect that a domestic company shall pay tax on all dividends at the company level and such dividend is not taxable in the hands of shareholder (income Tax Act, Section 115-O). However, there are changes in the subsequent years. For example, 1997 introduced dividend tax at the company level, through the Finance Bill with a rate of 10%, but in 2002 it was removed, making dividend taxable on the hands of the shareholders. However, in 2003 dividend taxation at the company level was reintroduced with the rate of 12.5%, and the rate was raised to 15% in 2007. However, there have been exemptions in the years, 2008, 2012, 2016. This is supplemented by cesses like education cess etc. Except for a few passing remarks and scattered comments, the tax differentiation in the income tax system and its associated effects escaped serious academic attention. Whatever writings exist, concentrated only on the additional differentiation caused by way of the occasional levy of excess dividends taxes. A comprehensive analysis

246

J. V. M. Sarma

of the income tax system from the dividend point of view is long overdue. Government’s concern for stepping-up of Investment and reduction of incomes’ inequality is obvious through measures such as excess dividends taxes, discrimination against unearned incomes, or granting tax deductions to dividend incomes and so on. The present time-series empirical study is an objective attempt to analyze the tax differentiation underlying the Indian income tax system as well as to measure the response of public limited companies to such differentiation. The purpose being to provide an example of the sensitivity of Indian corporate sector to such regulatory tax measures.

11.2 Literature Though taxation has been recognized as one of the important factors affecting dividends ever since dividend decision itself was identified as “the primary and active decision variable in most situations” collection of empirical evidence on tax impact had not gathered momentum until late sixties. Interestingly, in later years the trend in the dividend literature has been clearly in favor of studying the tax aspects. So much so, many of the studies treated factors other than taxes as secondary. Much of this later preoccupation with tax impact owes its origin to some attempts in the United Kingdom on the part of Government to interfere with the existing tax balance between dividends and retentions.

11.2.1 The Lintner Study The literature is vast on this subject. Based on interviews and detailed analysis of a selective sample of 28 US companies over a seven-year period, Lintner (1956, p. 97) postulated a rather simplistic behavior model in which current dividends were primarily determined by past dividends and by current earnings. According to Lintner, the problem facing managements of firms in determining dividends is not as to how much of current profits should be distributed as dividends, nor at what rate (in terms) of capital). The problem is primarily about whether there should be a change in dividend payments at all, and if there should be, how large the change should be. The reason, according to Lintner, seems to be a belief on the part of managements that most stockholders prefer a reasonably stable dividend rate. Management, therefore, try to avoid ‘temporary’ revisions in dividends even under favorable conditions unless they are sure of the sustainability of such conditions. Lintner also specified the mechanism of dividend change as follows: “The principal device used to achieve this consistent pattern was a practice or policy of changing dividends in any given year by only a part of the amounts which were indicated by changes in the current financial figures. Further partial adjustments in dividend rates were then made in the subsequent years if still warranted.” As a result, dividend distributions tend to be

11 Measuring Tax Impact on Corporate Dividend Behavior in India

247

less erratic compared to other ‘financial variables’. Lintner was also very specific about the other ‘financial variables’ or ‘favourable’ conditions. He recognized that current earnings as the only factor which is ‘reasonably persuasive if not compelling’, to induce a change in dividends from past practices. He argued that managements generally “believe that unless there were compelling reasons to the contrary their fiduciary responsibilities and standards of fairness required them to distribute part of any substantial increase in earnings to the stockholders.” Nevertheless, he believed that the other ‘compelling reasons’ are less regular and less understood and therefore, dividend changes are induced by only net earnings. The model specified by Lintner is as follows: First, a ‘target’ or ‘desired’ level of dividends is determined as Dt∗ = r Pt

(11.1)

where D is the current year’s dividends, P is profits after taxes and r, the target pay-out ratio. The ‘inertia’ mechanism by which current dividends are partially adjusted to desired dividends is [ ] ΔDt = a + c Dt∗ − Dt−1 + u t

(11.2)

where ΔDt is the change in dividends in the years indicated by the subscripts t and t − 1. The parameter c indicates the fraction of actual dividend change in the desired change. The constant a was assumed to be positive to reflect greater reluctance to dividend reductions than to dividend raises, a tendency observed by Lintner in his sample companies. Lintner found further empirical support to his model by testing it against time-series aggregate data pertaining to all US companies over the period 1918–51 and observing robust fit. The phenomenon of current dividend payments being dependent upon past dividend practices was, it should be noted, observed even earlier by Dobrovolsky (1951). But the basic contribution of Lintner was the glorification of this aspect by considering past dividends as ‘the primary’ determinant of dividends as well as the rationalization of’dividend inertia’ mechanism in terms of ‘partial adjustment’. The Lintner hypothesis was extensively tested later (Britain, 1966; Sastry, 1968; Tarshis, 1956) emphasizing the role of factors other than profits or past dividends. Lintner himself was aware of this fact but he was convinced that other considerations are “less generally known, less widely understood and sympathetically recognized by stockholders as factors which should have an important bearing upon dividend distributions. Moreover, no other consideration was important year by year and company by company”, Nevertheless, the later studies did prove that several other factors, particularly firms’ investment needs, liquidity position, financing pattern, taxes as well as growth prospects of the firm also play a significant role in determining dividend changes. Studies such as (Krishnamurty & Sastry, 1975; Sastry, 1968) tested for interdependency between dividends, investment and external financing decisions of firms and employing alternative mechanisms and rationalizations such as (Darling,

248

J. V. M. Sarma

1957). Prais (1959) proposed an adaptive expectations mechanism by which dividend change depends upon a change in firms’ expectations regarding future earnings. However, these finer differences in rationalizing intertemporal adjustment behavior of dividends could not be resolved so far as many of the commonly used econometric tools are found to be inadequate to bring out these differences. Several alternative methods are now available which help to minimize the errors in estimating the model. The other points relate to such matters as to whether ‘cashflows’ or ‘gross profits’ would be a better approximation than ‘net earnings’ used by Lintner (with or without inventory gains) to represent firms’ ‘capacity to pay’ dividends, or whether the relevant rate of dividends should be in terms of paid-up capital or net worth, and so on. These, in our opinion, by and large depend upon particular circumstances and situations and cannot be generalized. It would be useful to examine the applicability of Lintner’s model in India, by glancing through some of the empirical studies that used the model to explain the dividend behavior of Indian corporations, as we too will be using this model though modified in a different way.

11.2.2 Studies Testing Tax Impact on Dividend Behavior Taxes, both on companies as well as on shareholders, can induce a change in the preference patterns for dividends, whether there is a need or not. The primary effect noticed was what Brittain described and estimated as the ‘tax depression effect’. The effects of taxation, however, is not limited to only the depression effect. The other effect due to ‘tax differentiation’ between dividends and retained profits is increasingly being utilized by governments to encourage corporate savings, even though increased savings may or may not result in higher Investment efforts. Thus, there is a need to separate out from the target pay-out ratio, another set of factors affecting the mere preferences of companies for dividend changes. Realizing the importance of taxes and the varied ways in which they affect dividend behavior, later developments in literature entirely concentrated on studying the tax impact. Among these studies we shall consider Brittain, Feldstein, Moerland, and King, which marked the development of literature in this direction. One of the earliest studies to recognize tax effects on dividends explicitly/was that of (Britain, 1966). His was a time-series analysis of US companies over a 40-year span, 1920–1960 (excluding WorId War IX years. He used Lintner’s equation but retained in its original first difference form, presumably to minimize the lagged dependent bias. The equation specified was ] [ Dt − Dt−1 = a + c r Yt − Dt−1 + u t

(11.3)

where Y, the capacity variable (cash-flow net of taxes), c and r, the parameters representing the speed of adjustment and long-run target pay-out ratio, and ut , the random factor. Also, to isolate the impact of other factors, Brittain considered depreciation, Investment demand, individual taxes, interest rates, Sales change, and corporate

11 Measuring Tax Impact on Corporate Dividend Behavior in India

249

liquidity as the main factors affecting dividends. Of particular interest are the tax factors. Brittain considered two types of effects taxes can exert on dividends. One is a ‘depressing’ effect because of an overall increase in the corporation tax, and the other, due to a ‘shelter’ provided by retained earnings from individual income taxes. He realized that an increase in corporation tax that is not completely shifted reduces the pay-out base and thus undoubtedly has a depressing effect. This is in line with Lintner’s observation cited above. But the tax depressing effect may not be equally borne between dividends and retentions. In Brittain’s words; “… such a tax increase could actually raise the after-tax pay-out ratio since it cuts the profitability of investment and rules-out marginal projects; in this situation dividends might be maintained without additional resort to external financing. Finally, it is also possible that a large tax increase might be passed along in the form of dividend cut by firms insistent (perhaps irrationally) upon a fixed level of savings or of resources. To test the ‘depression’ effect, Brittain included pre-tax income and taxes, separately in his equation but found no evidence for the ‘substitution’ effect. Therefore, he concluded that an increase in taxes reduces only the level of dividends but not affect the pay-out ratio. More important is Brittain’s analysis of the other effect through individual taxes. Retained earnings provide a tax ‘shelter’ from personal income tax and therefore, a rise in the personal Income tax rate would induce shareholders to prefer lower dividends. An offsetting effect could be due to personal capital gains tax since corporate retentions when realized as capital gains may also be liable to tax. Brittain hypothesized that the dividend pay-out ratio varies inversely with the differential between personal income tax and capital gains tax. However, he realized that the hypothesis to hold, several conditions need to be satisfied. Important are; (a) that a substantial number of shareholders should be sufficiently ‘sophisticated’* to recognize the tax savings via retentions, (b) that capital gains tax rate should be sufficiently lower than income tax rate so that the lag in dividends will be more or less compensated later, and (c) managements should have sufficiently close liaison with their shareholders to take their preferences into account in making dividend decisions. To test this hypothesis, Brittain proxled the ‘differential’ as T = (1 − t g )/ (1 − t y ), where t g and t y represented the rates of capital gains and Individual taxes. The specification of Brittain’s dividend equation containing T is ] [ Dt − Dt−1 = a + c (α + βT )Yt − Dt−1 + u t

(11.4)

Brittain tested this equation with alternative definitions of T j with and without t g , as well as on the basis of several approximations of t g and t y . He found that the shelter variable carried the expected negative coefficient and was highly significant. However, equations containing capital gains tax gave generally poorer performance than those excluding it. Thus, he was inclined to conclude that this might be the ‘economic reality’. “It is possible that individuals and corporate official who think in terms of tax avoidance via low pay-out may be unconcerned with the tax bite attending ultimate realization”. Also of interest was his re-specification of Lintner model to assess the tax impact. Brittain’s extension of the same type of specification even for other demand-side

250

J. V. M. Sarma

factors is, in our view, rather exaggerating the role of pay-out ratio. Further, Brittain’s study implicitly assumed a ‘Classical’ type of income tax system and thus, ignored the other component of tax differential caused by taxation at company level itself. Later studies filled this gap as well. Feldstein (1970) concentrated on this aspect and attempted to take into account the corporation tax differential as well, in particular that caused by the British differential profits tax on dividend policies of British companies. In doing so, Feldstein used a slightly more generalized version of Lintner model. The desired or ‘optimum’ dividend function was specified as β

Dt∗ = A · Ytα · θt · u t

(11.5)

where the desired level of dividends is a function of income, and tax differential variables respectively. Thus the target pay-out ratio of Lintner model is split now into A, a constant component of Y and θ, which varies with tax differential whose ‘response’ or elasticity defined as ‘tax opportunity cost of retained earnings In terms of dividends’ which is similar to Brittain’s definition of ‘tax shelter’ variable. The partial adjustment equation was also suitably modified to be ‘compatible’ with the above equation, which is ]λ [ Dt /Dt−1 = Dt∗ /Dt−1 · vt

(11.6)

The λ was the response elasticity similar to Lintner’s ‘speed of adjustment’ and vt a stochastic term. The main findings by Feldstein were that, the impact elasticity with respect to a tax-induced change in the opportunity cost was less than unity and highly significant. The later studies by Moerland and King endeavored to provide a rigorously theoretical backing as to how exactly taxes should enter the dividend model and what modifications are needed and so on. Moerland (1975) entirely concentrated on theoretical aspects of dividend behavior. He assumed an objective ‘utility’ function containing dividends and retained profits as the elements and that maximization of the function was assumed to be equivalent to maximization of the market value of a firm.

11.2.3 Other Recent Studies There are many other studies examining the tax impact on dividend distribution behavior mostly empirically. They include (Adwani & Joshi, 2017; Jabbour & Liu, 2003; Mario & Chao, 2008; Obayagbona & Ogbeide, 2018; Rhee, 1990) and so on. The study (Adwani & Joshi, 2017) attempted to make a critical evaluation of the present dividend tax system and tries to provide suggestions for a better alternative tax system. It shows that substantial progress has been made in rationalizing and theorizing the dividend behavior and associated tax impact, as well as in improving the estimation methods. This study also tries to find the existence of any significant

11 Measuring Tax Impact on Corporate Dividend Behavior in India

251

relationship between the state revenue from this tax and the shareholding patterns of the corporate entities of the country. With special reference to the shareholding pattern of 30 companies of BSE Sensex, it is an attempt to make a critical evaluation of the present tax system and revenue generated from this. The study also tries to provide relevant suggestions for a better alternative tax system. The study observes that taxing the dividends on the hands of the company has actually increased the burden on the corporate sector and significantly decreased the Government revenue that actually goes against the present globalization policies of India. The study by Mario and Chao (2008) focuses on the distortionary effect of progressive dividend taxation on investment decisions under the dividend taxation. It uses a stochastic general equilibrium model to study the qualitative and quantitative importance of the distortion. The study finds that the dividend taxation introduces a wedge between the marginal cost and benefit of investment the magnitude of distortion critically depends upon the progressivity of the tax system. The study (Obayagbona & Ogbeide, 2018) examines the relationship between corporate taxes, and dividend policy of non-financial companies in Nigeria. Specific changes in dividend payout were judged based on the relative impacts of taxes, agency costs and transaction costs on firms. The study focuses on 48 active nonfinancial companies listed in Stock Exchange for a period of 8 years (2008–2015). Econometric techniques were adopted and the results from the Panel data technique reveal that corporate tax liabilities of non-financial firms do not have significant impact on their dividend pay-out, suggesting that taxes may not affect how firms plan their dividend policy. The study by Jabbour and Liu (2003), find size to be the most important factor related to dividends when taxes are not taken into account. In addition, their empirical evidence suggests that profitability is the only factor related to dividends when tax rates are included. In other words, the more profitable the firms are, the more likely they pay higher dividends as applicable tax rates decline. The study by (Rhee, 1990) tests the hypothesis that firms with high payout ratios tend to be debt financed, while firms with low payout ratios tend to be equity financed. Therefore, it can be predicted that dividend payout ratio and leverage ratio are positively correlated. A study by Agarwal (2021) attempts to capture the impact of dividend distribution tax removal since 2020, in India. It explores the factors associated with changing payouts in general. They observed that out of the top 1000 firms, the dividend payout behavior of 509 nonfinancial firms from 2015 to 2019 are analyzed. It observes that COVID’s impact on the firm’s financial performance and sentiments seems to dominate over the suppressing investors’ expectations of heightened payouts associated with dividend distribution tax advantages, with considerable reductions in payouts and omissions shown by regular and irregular payers in 2020 and 2021 as against the preceding years. The findings signify that the dividend payouts of sample firms are positively associated with the firms’ size and past dividends, The study lends credence to the conservatism in the dividend behavior of Indian corporate firms. Alzahranii and Lasfer (2009) analyzed dividends tax systems across 24 OECD countries and assessed their impact on dividend distributions. They mainly found that the dividend payout is ‘monotonically’ distributed across tax regimes as firms

252

J. V. M. Sarma

of countries with classical tax system have significantly lower payouts and speed of adjustment to target dividends than companies in partial or full imputation tax system countries. They also reported that the type of dividend tax system affects the size of dividend payout while the tax rate differential between dividends and capital gain affects the propensity to pay and the decision to change dividends.

11.3 The Model and the Methodology Following these studies, it is now clear that the starting point in analyzing dividend behavior is an objective function representing a firm’s preferences regarding dividend-retention mix. The objectives of corporate management, over the years, changed to private management, agency management, and so on. Shareholders’ view is different from managing agency. Then there is subsidiary treatment. Following Moerland and (King, 1971), consider a typical firm having a map of indifference dividend preference curves, each indicating a unique level of ‘utility’ obtained by alternative combinations of net dividends and net retained profits. The dividend preference function can be denoted as U = F(Dn , R)

(11.7)

where Dn and R are net of all taxes at all levels. The utility level as given by each curve can be viewed as monotonically related to the motivations of management, which also consider the shareholders’ preferences. The shape of the utility curves might be an outcome of a process of weighing their relative preferences, as well as a result of a number of factors influencing such preferences. However, following (Denny, 1974), a minimum set of reasonable conditions can be imposed. They are (a) monotonicity, and (b) quasi-concavity. Condition (a) ensures that utility derived from higher levels of and R is greater than that from lower levels. Condition (b) ensures that the relative marginal utility declines as the firm moves along the curve. By imposing a further condition of homotheticity, Denny derived a generalized quadratic form for production functions, which was later used by King to describe his ‘indirect managerial function’ of dividends and retained profits. The function is ⎤1/β



∑∑

U =⎣

i

βγ

(1−γ ) ⎦

αi j X i X j

(11.8)

j

where i and j = d and r, the subscripts for Dn and R respectively, and X d = Dn and X r = R. The distribution parameters αij > 0, and the substitution parameters β ≤ 0 and 0 ≤ γ ≤ 1. In order to find the tax effects, we first assume that the

11 Measuring Tax Impact on Corporate Dividend Behavior in India

253

specification of the dividend preference function is independent of taxes, and then introduce the tax elements via a budget constraint. The analytical strategy is similar to consumers demand theory. The constraint in this case is the firm’s total income net of expenses and interest on debt. Depreciation for the time can be assumed as ‘economic’ depreciation and is exactly equal to year-to-year capital consumption so that it can be treated as an expenditure item. It is also useful to deduct dividends paid to preference shareholders, as they can neither be regarded as equity dividends, nor can be treated as interest on debt since their payment is conditional on the existence of sufficient profits base. Certain types of preference shares also lay claims on future profits if current profits are not sufficient. Thus we define the base for dividend payments as total profits net of interest and preference dividends. There is also the question of compulsory provisions specified by company law and tax law. Though these provisions are compulsory the marginal utility of additional non-compulsory retentions very much depend upon the number of compulsory provisions. Therefore, it is essential to include all those provisions. The profit allocation function or the budget constraint can be written as Y − D + R + T

(11.9)

where Y denotes the total profits base, D and R, net dividends and retentions as defined, and T, the total tax liability as a result of income taxes both at the company level as well as at shareholder level. Alternatively, it can be written as Y − Dn pd + Rpr

(11.10)

by defining the * tax prices’ pd as Dg/Dn and py as Rg/R, where Dg and R are ‘gross’ devidends and ‘gross’ retentions, respectively. For example, Dg denotes the amount of profits to be allocated to realize one rupee of net dividends, Dn . The prices, of course, depend upon the prevailing type of tax system and the tax rates. For constrained maximization, define the Lagrange function, L as ) ( L = U + μ Y − Dn pd − Rpr

(11.11)

The first order conditions, ∂ L/∂ D = Ud' − μpd = 0

(11.12)

∂ L/∂ D = Ur' − μpr = 0

(11.13)

∂ L/∂μ = Y − Dn pd − Rpr = 0

(11.14)

Ud '/Ur ' = pd / pr

(11.15)

yield the equation

254

J. V. M. Sarma

where ( )β−1 [ βγ −1 β(1−γ ) ] β−1 αii · β X i + αi j βγ · X i · Xj 1 Ui ' = −βγ βγ β +αi j · β(1 − γ ) · X ·X i

(11.16)

j

The only way for a solution is to impose another restriction on the shape of the objective function namely, that αij = 0 when I /= j so that the cross-terms will disappear, and the function degenerates into a CES function. The imposition of this condition is arbitrary, but needed in order to obtain a manageable solution for Dn. The first order condition so trimmed would be αdd /αrr (D/R)β−1 = pd / pr

(11.17)

Denoting Pd /Pr . as F, l/(β − l) as σ,… αrr /αdd = A, and substituting for R and rearranging terms, the optimal solution for Dn * can be obtained as ] [ Dn∗ = Aσ · φ σ / 1 + Aσ · φ σ , Y / pr

(11.18)

It can be easily seen that σ denotes the elasticity of substitution between D and R. The second order condition requires to be negative. The distribution parameter β → 1, 0, or ∞ as σ → ∞, 0, or 1, and the utility function degenerates into linear, Cobb–Douglas, or Leontief types respectively.

11.3.1 Inter-Temporal Adjustment As we noted above, both shareholders as well as managements seek a regular and less fluctuating dividends which will result in a lagged adjustment of dividends to changing conditions. Two types of the lagged adjustment processes are conceived in the literature; (a) partial adjustment process, and, (b) a process of adjustment to ‘permanent’ levels of factors affecting dividends.and the optimal function of a firm as ( )σ +1 σ +1 .qr Aσ θ σ 1 + B η δ η+1 Y ( ) σ +1 ] · D∗ = [ σ +1 η η σ σ +1 η η+1 pr 1+ B ·δ · qr + A ·θ (B · δ )

(11.19)

From (11.19) it can easily be seen that the target pay-out ratio in Lintner model is not a constant, but varies with the tax factors: θ, δ, qr , and pr . Also, the tax effect is split into two parts analogous to ‘income’ and ‘substitution’ effects in consumer demand theory. While θ and δ represent the ‘substitution’ effects of the relative ‘tax prices’, qr and pr represent the income effects respectively at the shareholders’ level and company level.

11 Measuring Tax Impact on Corporate Dividend Behavior in India

255

The advantage of the extended model is that it not only takes account of the tax differentials both at the company level as well as at the shareholders’ level, but it also allows for the separation of the two influences in a flexible way. In doing so it explicitly allows for the response of management to personal taxes to be exactly same as the combined response of shareholders.

11.3.2 Interpretation of the Model in the Indian Context The optimal dividend equation can be interpreted in terms of different tax systems that prevailed in India as follows. As can be noted the system, prior to 1959–60 was of ‘Imputation’ type which was later changed to ‘Classical’ type. Added to these broad systems were the occasional dividend taxes. For the purpose of interpreting the model we shall leave-out the details and consider only the main features. Under the ‘Imputation system, let t a and t b be the income tax and super tax on a company, and t i and t g be the personal income tax on dividends and an equivalent tax on capital gains respectively. As a result of ‘grossing-up* practice, the personal income tax liability on dividends received by the shareholders would be(t i − t a )/(l − t a ). Further, let t d be an additional tax on dividend distributions at the company level. The tax prices can be determined as follows. Consider the over-all budget constraint of a company as Y = Dg + R g

(11.20)

where Dg and Rg are gross profits assumed to have been allocated for dividends and retentions. The ccnstraint can be written in terms of net dividends and the retentions as Y =

R D n · ps + (1 − ta − tb ) · (1 − td ) (1 − ta − tb )

(11.21)

In a similar way shareholders’ budget constraint would be ( D = Dn ·

1 − ta 1 − ti

)

C ) +( 1 − tg

(11.22)

Comparing these constraints with their counterparts in (11.4) and (11.15), we obtain the ‘prices’ in terms of taxes as pr = 1/(1 − ta − tb )

(11.23)

pe = Dg /D = 1/(1 − ta − tb ) · (1 − td )

(11.24)

256

J. V. M. Sarma

qd = (1 − ta )/(1 − ti )

(11.25)

) ( qr = 1/ 1 − tg

(11.26)

θ = 1/(1 − td )

(11.27)

) ( δ = (1 − ta ) · 1 − tg /(1 − ti )

(11.28)

)η+1 ( (1 − ti )η+1 + B η (1 − ta )η+1 · 1 − tg ps = ( )η+1 B η 1 − tg (1 − ta )η (1 − ti )

(11.29)

)η+1 ( (1 − ti )η+1 + B η (1 − ta )η+1 · 1 − tg φ= ( )η+1 B η 1 − tg (1 − ta )η (1 − ti ) · (1 − td )

(11.30)

Further,

And,

The optimal dividends equation can be obtained accordingly. Under the Classical system, the only change required for deriving the dividend equation is in the interpretation of qd , which now becomes l/(l − t i ). Correspondingly, the term (1 − t a ) disappears from the interpretation of δ, ps , and F. Further, it is also clear that under both the systems, in the case of no additional tax on dividends (t d = 0), pg would be simply 1/(l − t a − t b ) and the tax differential at the company level will be unity. Thus, under the simple classical system, the only tax differentiation would be through the individuals’ taxes.

11.3.3 Influence of Other Factors The utility function we assumed is a result of a number of factors analogous to tastes and preferences in consumer demand theory, which are considered to be relatively constant in the short run. But these factors are less likely to have remained the same over the period under study.

11 Measuring Tax Impact on Corporate Dividend Behavior in India

257

11.4 Empirical Analysis The empirical analysis involves two steps: First, to identify the parameters by fitting the model to time-series aggregate data, and second, to simulate the equation so estimated in order to quantify the tax impact. Accordingly, we shall devote Sect. 11.1 to estimation of the dividend’s equation and Sect. 11.2 to simulate and quantify the tax effect. Also, we will do a disaggregated analysis is to see if industries in different sectors react differently to the tax policy.

11.4.1 Regression Analysis (Cobb-Douglas Assumptions) The first hypothesis we wish to test by regression approach is that profits and tax variables can adequately explain all the movements in dividends rate. Also, initially we hypothesize that the objective ‘utility* function is of Cobb-Douglas type, so that tax differential variable and the elasticity coefficients involved are unity. We expect unitary elasticities for at least one reason. As we are going to use aggregate data rather than firm-wise data, there is a chance that the firm-specific over- and underresponses might be evened-out producing unit elasticity of substitution. However, if the linear regression Indicates that the elasticities are other than unity then we must estimate the more general form by non-linear estimation precedures. The equation implied by the Cobb-Douglas assumption is, λβ1

Dt = Aλ Yt

−λβ2

· φt

β

u · πt 3 · DT(1−λ) −1 · e

(11.31)

where we expect that all β are zero and π = 1/pr . Initially, the estimated equation indicated that the coefficients are affected by the ‘lagged dependent bias’ in the regression. In fact, there are three kinds of bias that could possibly be associated with this regression. The other two types are: a bias due to absence of other relevant factors affecting dividends, and a bias due to misspecification of the dividend equation by restricting to Cobb-Douglas frame. As noted by many studies, the Eq. (11.31) is a typical case of lagged dependent variable appearing as independent variable. As a result, OLS estimates of parameters might be biased, the extent of bias being unknown. Further, the problem is worse in time-series analysis if there is serial correlation in errors, making the estimates inconsistent. It is also well-known that in such cases the standard indicators of serial correlation such as Durbin-Watson (DW) statistic or the serial correlation coefficient (p) cannot be relied on. Even the Theil’s H-statistic is not applicable in our case due to smallness of the sample. Fortunately, several alternative procedures are now available in econometric literature dealing with equations containing 2 lagged dependent variables. However, the superiority of many of these procedures over OLS is not yet clearly established.

258

J. V. M. Sarma

For our purpose, we shall choose the following alternative procedures: (a) Generalized least squares (GLS) with iteration, (b) GLS with grid-search and iteration, (c) Residual adjusted GLS suggested 3 by Durbin and later by Hatanaka, and (d) Wallis’ combined instrumental variables and GLS. Procedure (a) yields maximum likelihood estimates by iteration process in which an estimate of serial correlation coefficient of OLS residuals is used to transform the variables. Procedure (b) also yields maximum likelihood estimates, but instead of using the OLS residuals, it searches for an appropriate estimate of the serial correlation coefficient and conducts the iterations. Both procedures (a) and (b) yield estimates which are asymptotically efficient. Procedures (c) and (d) combine the GLS with the instrumental variable method. The Hatanaka procedure requires no iteration. It consists of instrumental variables to compute the serial correlation coefficient of errors followed by the OLS method with all the variables. The final selected regression estimates are as follows (Table 11.1). Or, 0.86 Dt = −0.21 · Y 0.21069 · φ 0.21 · π .78 · Dt−1

(11.32)

The over-all significance is reassuring. Among the independent variables only Y and Dt − 1 are significant, and the tax variables and n are not significant at any acceptable level of significance (Table 11.2). However, the coefficients of the tax variables have expected signs. The regression supports the applicability of Lintner’s partial adjustment theory. The lag adjustment parameter λ is 0.86, and the only convincing reason for managements to deviate from the past dividend rate seems to be changes in profits variable, though not very compelling, judged by its statistical significance. The estimated elasticities with respect to ϕ and π work out to be 0.91 and −0.94 and subsequent t-tests prove that Table 11.1 Ordinary least squares Ordinary least squares Sample 1951–2015 Number of observations

64

Explanatory variables

Estimated coefficints

Standard error

1

Const

−0.02

0.51

2

INCM (Y)

3

PHI (φ)

4 5

T-statistic 0.00

0.21

0.12

1.01

−0.21

0.19

−0.62

PI (π)

0.78

0.21

0.36

DIV(t − 1)

0.86

0.21

4.15

R2

0.77

DW

1.52

SSR

0.16

SEE

0.02

11 Measuring Tax Impact on Corporate Dividend Behavior in India

259

Table 11.2 Auto-regression results with other variables Sample 1951–2015 Command line

OLS

Const

−0.1137

GLS with iteration 0.03482

−0.0175

GLS with grid search 0.6493

Hatanaka residual adjustment 0.3775

Wallis instrumental variables 0.1875

0.2545

0.003

0.1753

0.3287

0.235

0.28

0.143

0.1786

Y

0.1054 0.6753

2.38

1.87

0.732

φ

−0.12333

0.015

0.0048

0.00378

−0.0023

−0.523

−0.083

0.0043

0.0011

−0.004

−0.3425

−0.072

π

−0.032 D(t − 1)

0.8694

0.2012

−0.00745

−0.0035

−0.00032

0.0233

1.15132

1.3215

−0.00213

0.4235

0.4325

0.7891

0.1097

2.4532

2.4671

3.783

1.568

2.456

−0.0043

−0.8983

−0.0087

−0.00032

−0.00042

−0.023

−0.0542

−0.0872

0.0874

0.5429

0.2138

0.1252

0.3012

0.4591

0.2384

0.00145

1.2804

0.0825

0.0027

0.00397

L

0.0023

0.00041

0.00527

0.00238

0.00765

1.03

1.2458

1.0003

0.0946

0.00247

S

0.0086

0.0045

0.032

0.08232

0.0043

0.0187

0.0483

0.0934

0.0463

0.092

I E

ρ

1.243

Note φ

Tax differential

π

Tax depression

I

Investment demand

E

Capital structure

L

Liquidity demand

S

Sales change

they are not different from unity. Inclusion of other financial variables has not altered the OLS estimates.

11.4.2 Regression Analysis (CES Assumptions) To make sure, there is no specification bias, we shall attempt to estimate the general form of the model,

260

J. V. M. Sarma

[ Dt =

]λ ( ) )σ +1 ( Aσ θ σ · 1 + B η δ η+1 Y λ 1−λ · · Dt−1 ( ) σ +1 pr (B η δ η )σ +1 + Aσ θ η+1 1 + B η δ η+1

(11.33)

The general form will be estimated by a non-linear procedure combining the Gauss-Newton method with the method of steepest descent. These are iterative methods in which prespecified initial values are used to find the vector in the parameter space, along which the sum of squared residuals decreasing most rapidly. If the steepest descent fails then another set of values may be initiated. This process was repeated until the sum of squared residuals is minimum. To minimize the computation burden, the financial factors were excluded. A number of initial parameter sets have been tried, but better estimates are obtained by using the OLS estimates to specify the initial set, and the coefficient of Y and p is kept constant at unity. Also, the parameter was initiated at 0.9 (Table 11.3). The nonlinear results showed significant improvement in R as well as the statistical significance of the coefficients only show that the linear regressions involve a serious specification bias. In fact, the lagged dependent bias can be regarded as negligible when compared to the other. But standard error is higher in the non-linear case, which could be due to a serial correlation. The elasticity of substitution at the firm-level is −3.24 and is well above unity. Shareholders’ elasticity of substitution is also higher than unity at −1.64, but not as high as the firms’ elasticity. The estimates reveal a high sensitivity on the part of firms to tax differentiation. Also noticeable is the dividend lag which is now estimated at 0.3. It means that current dividends reflect only one-third of the changes in the desired dividends, which in turn would mean that two-thirds of a change in the tax policy in t’th year will prolong its effect on Table 11.3 Non-linear estimates Estimation by Gauss-Newton method of steepest descent Convergence achieved by Gauss Sample

1951–2015

Final sum of squares

0.23145

Explanatory variable

Estimated coefficient

Standard error

T-statistic

λ

0.58023

0.045

0.8653

σ

−2.4365

0.7639

−3.7648

η

−7284

0.97834

−2.2864

A

1.4629

0.48329

1.543

B

1.9635

0.78219

1.786

R-sq

0.9875

Mean var

2.2154

DW

1.762

F-stat

54.2916

Rss

0.0005482

11 Measuring Tax Impact on Corporate Dividend Behavior in India

261

dividends. It must be stated however, that too much reliance cannot be placed. On these estimated also, for the simple reason that the non-linear estimation by GaussNewton and the method of steepest descent has always been tricky. In our case, the final estimates showed considerable variation whenever the initial parameter set is changed. And many a time convergence is not achieved, though the sums of squares changed only slowly. The possible reasons could be that firstly, the sample is too small, and secondly, the equation, though ‘identified’ in the statistical sense, is too complex for the estimation. The elasticity estimates are far higher than those obtained by the linear methods. In view of these difficulties, only the linear equation estimates are relied for simulation purposes. [

]0.122 )0.58023 ( 1.46290.58023 θ 0.58023 · 1 + 1.9635B −0.7284 δ −1.7284 Dt = ( )−1.72843 ( )0.58023 B −0.7284 δ η + A0.58023 θ 1.58023 1 + 1.96350.72843 δ 1.72843 ( )0.209 Y 0.5878 · Dt−1 · pr There is improvement as well as the statistical significance of the coefficients. It shows that the linear regressions involve a serious specification bias. In fact, the lagged dependent bias can be regarded as negligible when compared to the other. But standard error is higher in the non-linear case, which could be due to a serial correlation. The elasticity of substitution at the firm-level is 0.96 and is close to unity. Shareholders’ elasticity of substitution is also higher than unity at 1.64, but not as high as the firms’ elasticity. The estimates reveal a high sensitivity on the part of firms to tax differentiation. Also noticeable is the dividend lag which is now estimated at 0.88. It means that current dividends reflect over 80% of the changes in the desired dividends, which in turn would mean that almost a change in the tax policy in tth year will prolong its effect on dividends. It must be stated however, that too much reliance cannot be relied upon these estimates, as they may differ with the type of company, industry and other characteristics.

11.5 Summary An attempt was made to fit the dividend model postulated in the earlier chapter to the time-series data on the Indian corporate sector and thereby measure the dividendsensitivity to variation in the tax differentiation, as well as quantify the tax effects by simulating the estimated model. Starting with relatively simpler degeneration of the general model, the coefficients are estimated by different methods with a view to identifying the correct version of the model. The Cobb-Douglas version of the model does not fit the data satisfactorily. Attempts to correct the lagged dependent bias, and serial correlations have not improved the situation. On the other hand, the

262

J. V. M. Sarma

CES version, though proved to be a better specification, yielded coefficient estimates which seem to be less stable. The disaggregated analysis by six broad industry categories brought the relationships in a clearer way. Specifically, the three industry groups, III, IV, and V, comprising the manufacturing sector proved to be very sensitive to tax changes. The quantification of the effect by means of simulating the best estimated equations for each group show that much of the effect in l has been due to the adaptation of ‘Classical’ system. The effect was the highest in 1960–1968 during which time the Classical system was just introduced, the excess dividends taxes were levied, and also the rates of personal income taxes had been higher compared to the other sub-periods. An important indication is that the effect of excess dividends taxes by itself is very low, compared to that of personal income tax, which is not fully in agreement with the prevalent view regarding these taxes.

References Adwani, V., & Joshi, R. S. (2017). A critical evalueation of corporate tax. Journal of Commerce, Economics, & Computer Science (JCECS), 3(4), 167–174. Agarwal, A. (2021). Impact of elimination of dividend distribution tax on Indian corporate firms amid COVID disruptions. Journal of Risk and Financial Management. Alzahranii, M., & Lasfer, M. (2009, Februar 15). The impact of taxation on dividends: A crosscountry analysis. Retrieved from SSRN: https://ssrn.com/abstract=1343826, https://doi.org/10. 2139/ssrn.1343826 Bandyopadhyay, A. K. (2008). Dividend decisions in Indian firms with special reference to engineering sector. Aligarh Muslim University. Retrieved from https://core.ac.uk/download/pdf/144 520921.pdf Britain. (1966). Corporate dividend policy. The Brooking Institution-Studies in Government Finance. Darling, P. (1957). The influence of expectations and liquidity on dividend policy on dividend policy. Journal of Political Economy. Denny, M. (1974). The relationship between functional forms of the production system. Canadian Journal of Economics, 7. Dobrovolsky, S. (1951). Corporate income retention, 1915–43. National Bureau of Economic Research, Financial Research Program. Feldstein, M. (1970). Corporate taxation and dividend behaviour. Review of Economic Studies. IBFD. (1972). A comparative analysis of classical, dual rate and imputation tax system in Belgium, France, Germany, Italy, Netherlands and United Kingdom. European Taxation. Jabbour, G. M., & Liu, Y. (2003). The effect of ta rate change on dividend payout. Journal of Business & Economics Research, 2(10). King, M. A. (1971). Corporate taxation and dividend behaviour: A comment. Review of Economic Studies. Krishnamurty, K., & Sastry, D. (1975). Investment and financing in the corporate sector in India. Tata McGraw Hill Publishing Co. Lintner, J. (1956, May). Distribution of incomes of corporations among dividends, retained earnings, and taxes. American Economic Review (Papers and Proceedings), 46. Mario, I. M., & Chao, D. I. (2008). The impact of progressive dividend taxation on investment decisions. In Working Papers, Congressional Budget Office. Moerland, P. (1975). Optimal dividend policy and taxes; another approach. Institute for Fiscal Studies discussion paper no.7502. Series on public economics.

11 Measuring Tax Impact on Corporate Dividend Behavior in India

263

Obayagbona, J., & Ogbeide, D. (2018). Corporate taxes, agency costs and dividend policy of non-financial firms in Nigeria. Amity Journal of Finance, 3(1), 61–86. OECD. (1973). Company tax systems in OECD member countries. OECD. Prais, S. J. (1959). dividend policy and income distribution in the U K. In B. a. Tew. In Studies in company finances. National Institute of Economic and Social Research, Cambridge University Press. Rhee, R. P. (Summer 1990). The impact of personal taxes on corporate dividend policy and capital structure decisions. Financial Management, 19(2), 21–31. Sastry, V. (1968). Dividends, investment and external financing behaviour of the corporate sector in India. Tarshis, L. (1956, May). Economic growth; income distribution—comments. American Economic Review (Papers and Proceedings), 46.

Part IV

Banking and Monetary Policy

Chapter 12

Non-performing Assets of Indian Banking: An Evolutionary Journey Rakesh Mohan and Partha Ray

Abstract This paper narrates the story of the evolutionary journey of nonperforming assets (NPA) in the Indian banking sector. Three distinct phases of the intertemporal behavioral of NPAs of the Indian banking sector can be discerned. First, since the initiation of financial sector reforms till about the beginning of the North Atlantic Financial Crisis (NAFC), NPAs showed a consistent downward trajectory. Second, during 2008–09 through 2017–18 the NPAs showed a distinct spurt. Third, since then, NPAs marked by a downward trend till 2019–20 until the economic disruptions caused by Covid 19. Contrary to the popular perception of treating the second phase of rising NPAs as one emanating exclusively from governance issues in public sector banks (PSBs), four factors have been identified: (a) falling commodity prices; (b) regulatory forbearance; (c) initial exuberance in infrastructure projects punctured by a downward phase of business cycles (leading to substantial debt accumulation of select big corporates); and (b) governance failure in select PSBs. Moving forward, while the pandemic and some of the associated policy measures could reverse the recent downward trends in NPA, more durable policy initiatives like bankruptcy reforms are expected to make significant positive changes in the NPA situation of Indian banks. Disclosure: Rakesh Mohan was a Deputy Governor of Reserve Bank of India from 2002 to 2004 and from 2005 to 2009. Partha Ray was a staff member at RBI from 1989 to 2013. The paper reflects personal views of the authors. The authors are indebted to Shankar Acharya, Sajjid Chinoy, Jaimini Bhagwati and Anoop Singh for their comments on an earlier working paper version of the essay (CSEP Working Paper No. 22). An earlier version of the paper was presented in a Conference on “India’s Contemporary Macroeconomic Themes” at Madras School of Economics, April 21–22, 2023. The authors are also indebted to the participants of the Conference and in particular to C. Rangarajan, S. Mahendra Dev, M. Govinda Rao, and N. R. Bhanumurthy for their comments. The usual disclaimer applies. R. Mohan Centre for Social and Economic Progress (CSEP), New Delhi, India e-mail: [email protected] P. Ray (B) National Institute of Bank Management (NIBM), Pune, India e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 D. K. Srivastava and K. R. Shanmugam (eds.), India’s Contemporary Macroeconomic Themes, India Studies in Business and Economics, https://doi.org/10.1007/978-981-99-5728-6_12

267

268

R. Mohan and P. Ray

12.1 Introduction Dr. C. Rangarajan is a leading macroeconomist, a seasoned policy maker par excellence, and a statesman of our times. He had a long association with the Reserve Bank of India, first as a Deputy Governor during from 1982 to 1991, and then as the Governor between 22 December 1992 and 21 December 1997. He has always been a thought leader, particularly in the realm of monetary policy, exchange rate management and on the financial sector during his tenure in the Reserve Bank (Rangarajan, 2022). As early as on October 5, 1997, in a lecture on “Indian Banking—Second Phase of Reforms—Issues and Imperatives”, he commented, “Reduction in NPAs has acquired more focussed attention. The level of NPAs is no doubt high but the percentages are showing a decreasing trend” (Rangarajan, 1997). More recently, in several opinion pieces, he expressed his concern on and diagnosis of the nonperforming assets (NPA) of the Indian banking sector. Illustratively, in an article in the Business Line of December 18, 2019, he (jointly with B. Sambamurthy) noted: “Credit booms are generally succeeded by stress in the banking system. This has happened in India also. Credit expansion was phenomenal,….. It is true in a period of economic expansion every project looks rosy. Exuberance becomes all embracing, partly rational and partly irrational. It is also seen that there was a substantial flow of credit to certain sectors like infrastructure (roads, power, and telecom), iron and steel, mining and aviation. While all these sectors are important for growth, these were also subject to severe output fluctuations” (Rangarajan & Sambamurthy, 2019). Therefore, we thought it would be apposite to dedicate this essay on NPAs in his honour. The inter-temporal trajectory of the NPAs in Indian banking during 1992–2018 followed a distinct three-phase pattern. Data on the extent of NPAs on Indian banking are hardly available before the mid-1990s as appropriate classifications of NPAs did not exist then. Following the economic reform measures of the 1990s, in which financial sector reforms occupied a key position, significant improvement began to take place in the extent of NPAs of Indian banking. This falling trend in NPAs continued till around 2010, after the advent of the North-Atlantic financial crisis (NAFC) of 2008.1 NPAs started rising again after that, initially at a slow pace (perhaps reflecting several measures of regulatory forbearance) and then from 2014 at a faster rate till about 2018. Various factors are held responsible for the unabated rise in NPAs during 2010–2018; prominent identifiable reasons are: (a) fall in commodity prices; (b) prolonged regulatory forbearance; (c) failure of the public–private partnership projects in certain key infrastructure areas; and (d) possible governance issues in commercial banks (Mohan & Ray, 2019). Later, after 2018, coinciding with the initiation of progressive bankruptcy measures, there have been improvements in trends in NPAs once again, until the Covid 19 pandemic hit in 2020. 1

Following Mohan (2011), we use the term North Atlantic Financial Crisis (NAFC), in contrast to the more widespread usage of the global financial crisis (GFC). It has been conscious and has been prompted by (a) the origin of the crisis, and (b) its lack spread across the globe beyond the North Atlantic.

12 Non-performing Assets of Indian Banking: An Evolutionary Journey

269

A popular narrative about the sharp rise in NPAs during 2010–2018 is that banks suffered due to bad governance in public sector banks (PSBs). However, the issue with this narrative is that it is not consistent with the sharp fall of NPAs recorded by the same banks during the period 1996–2010. How could such an improvement take place in PSBs over a decade and a half and then decline so sharply after that? What was different in the governance of public sector banks in that period? Did the governance of the PSBs change drastically after around 2010? Was it accidental that the PSBs did well for so many years? It is here that the role of two specific factors needs to be highlighted. First, there was a significant expansion of large corporate sector lending after the late 2000s, including for lumpy infrastructure projects under public–private-partnership as a matter of public policy. Such lending could have led to unforeseen problems from the fall in some key commodity prices along with the slowing down of the economy. Second, the exercise of regulatory forbearance from 2008, in the wake of the North Atlantic Financial Crisis (NAFC), camouflaged the real dimension of the problem and could have engendered a false sense of complacency and related low measurement of NPAs. Later credit growth started slowing down along with a deceleration of the GDP growth rate. Against this context, the present paper investigates the intertemporal behavior of NPAs over the period, 1992–2018.2 For expository convenience, the structure of the paper is as follows. Section 12.2 discusses the broad trends in NPAs over time; Sect. 12.3 is devoted to a short discussion of the improving trends in NPAs during 1993–2010; Sect. 12.4 enumerates the major reasons for the emergence of NPAs in Indian banking during 2010–2018. While Sect. 12.5 discusses the recent improving trends during April 2018–March 2020, Sect. 12.6 is devoted to the way ahead.

12.2 Broad Trends of NPAs of Indian Banks 12.2.1 Some Definitional Issues At the level of popular folklore, an NPA is a bad loan. Thus, if a loan is not repaid on time, it becomes a bad loan, and hence it needs to be written off and a provision has to be made in the bank’s book of account. In India, recognition of such bad loans was largely opaque before the 1990s. The Committee on the Financial System (Chairman: M. Narasimham), which laid down the blueprint for financial sector reforms, acknowledged the need to recognize the NPAs in the banking sector (Narasimham Committee I; RBI, 1992). In the Report, assets were classified into four categories, namely: Standard, Sub-standard, doubtful and loss assets. Regarding these different categories, general provision has to be created to the extent of 10%, 2

For the bulk of our analysis, we consciously avoid 2020–21 because of the Covid19-related complications.

270 Table 12.1 Evolution of the definition of NPAs

R. Mohan and P. Ray

Year ending March 31

Specified period for “Past Due”

1993

Four quarters

1994

Three quarters

1995 onwards

Two quarters

Source RBI

20% to 50%, and 100%, respectively, of the total outstanding security shortfall in these categories. The crucial question is then: how long can the bank wait to classify a loan to be a bad loan? In terms of regulatory classification, the notion of NPA has undergone quite a bit of evolution.3 In particular, a ‘non-performing asset’ (NPA) used to be defined as a credit facility in respect of which the interest and/or instalment of principal has remained ‘past due’ for a specified period of time. The specified period was reduced in a phased manner over the 1990s (Table 12.1). Due to improvements in payment and settlement systems, recovery climate, and up-grading of technology in the banking system, the concept of ‘past due’ was dispensed with from March 31, 2001. Accordingly, from April 2001, a Non-performing Asset (NPA) was defined as an advance where: (a) “interest and/or instalment of principal remain overdue for a period of more than 180 days in respect of a Term Loan; (b) the account remains ‘out of order’ for a period of more than 180 days, in respect of an Overdraft/Cash Credit (OD/CC); (c) the bill remains overdue for a period of more than 180 days in the case of bills purchased and discounted; (d) interest and/or instalment of principal remains overdue for two harvest seasons but for a period not exceeding two half years in the case of an advance granted for agricultural purposes; and (e) any amount to be received remains overdue for a period of more than 180 days in respect of other accounts” (RBI, 2001). Later, to move towards international best practice and to ensure greater transparency, the ‘90 days’ overdue’ norm for identifying NPAs was adopted in 2003–04. Accordingly, with effect from March 31, 2004, a non-performing asset (NPA) was defined as a loan or an advance where the 180 days timespan was replaced by 90 days for all the criteria (a), (b), (c), (d), and (e), mentioned above. Put simply, in most of the cases, any loan which is due for 90 days or over is typically classified as an NPA.4 3

This discussion follows RBI Circular on “Prudential Norms on Income Recognition, Asset Classification and Provisioning—Pertaining to Advances” of August 30, 2001 (DBOD No. BP.BC/20/ 21.04.048/2001-2002) (RBI, 2001); available at https://m.rbi.org.in/scripts/BS_ViewMasCircular details.aspx?Id=449&Mode=0 (accessed in May 2021). 4 Treatment of an agriculture loan is, however, slightly different and in the case of an advance granted for agricultural purpose, it is classified as an NPA if “interest and/or instalment of principal remains overdue for two harvest seasons but for a period not exceeding two half years”.

12 Non-performing Assets of Indian Banking: An Evolutionary Journey Table 12.2 Provisioning requirements on various categories of loans

Asset category

NPA duration

Provisioning rate (%)