Outward Foreign Direct Investment of Chinese Enterprises (Contributions to Economics) 981194718X, 9789811947186

Citation preview

Contributions to Economics

Wei Tian Miaojie Yu

Outward Foreign Direct Investment of Chinese Enterprises

Contributions to Economics

The series Contributions to Economics provides an outlet for innovative research in all areas of economics. Books published in the series are primarily monographs and multiple author works that present new research results on a clearly defined topic, but contributed volumes and conference proceedings are also considered. All books are published in print and ebook and disseminated and promoted globally. The series and the volumes published in it are indexed by Scopus and ISI (selected volumes).

Wei Tian · Miaojie Yu

Outward Foreign Direct Investment of Chinese Enterprises

Wei Tian School of Economics Peking University Beijing, China

Miaojie Yu Liaoning University Shenyang, China Peking University Beijing, China

ISSN 1431-1933 ISSN 2197-7178 (electronic) Contributions to Economics ISBN 978-981-19-4718-6 ISBN 978-981-19-4719-3 (eBook) https://doi.org/10.1007/978-981-19-4719-3 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022 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

Contents

1 The Exceptional Performance of Chinese Outward Direct Investment Firms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Data and Empirics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.1 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.2 Empirical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.3 Benchmark Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.4 Alternative Measure of Firm Performance . . . . . . . . . . . . . . . 1.3.5 Robustness Checks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4 Policy Suggestions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Firm Productivity and Outward Foreign Direct Investment: A Firm- Level Empirical Investigation of China . . . . . . . . . . . . . . . . . . . 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Data and Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Determinants Affecting Enterprises’ Entry into the FDI Market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Enterprise Productivity and the Enterprise FDI Decision . . . . . . . . . 2.4.1 Impact of the Firm’s Productivity on Its OFDI Decision . . . 2.4.2 Endogeneity Between OFDI and Firm Productivity . . . . . . . 2.4.3 Relationship Between Firm Productivity and Income Level of the Destination Country . . . . . . . . . . . . . . . . . . . . . . . 2.4.4 Enterprise OFDI and Industry Labor Intensity . . . . . . . . . . . . 2.5 Impact of Firm Productivity on the Volume of OFDI . . . . . . . . . . . . . 2.5.1 Benchmark Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.2 Endogeneity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.3 Additional Robustness Tests: Analysis Based on the Gravity Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 1 2 3 3 5 5 7 7 9 11 12 13 13 16 20 25 25 27 28 30 32 32 33 35

v

vi

Contents

2.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

36 38 39

3 Distribution, Outward FDI, and Productivity Heterogeneity: China and Cross-Countries’ Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Data and Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 FDI Decision Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.2 FDI Flow Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.3 Firm-Level Production Data . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.4 Data Merge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4 Extensive Margin of FDI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1 Descriptive Analysis on Productivity Differences . . . . . . . . . 3.4.2 Extensive Margin of FDI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.3 Estimates with Rare Events Corrections . . . . . . . . . . . . . . . . . 3.4.4 Multinomial Logit Estimates with Distribution FDI . . . . . . . 3.4.5 Endogeneity of Firm Productivity . . . . . . . . . . . . . . . . . . . . . . 3.4.6 Discussions of Fixed Costs Ordering . . . . . . . . . . . . . . . . . . . . 3.5 Type-2 Tobit Estimates of Intensive Margin . . . . . . . . . . . . . . . . . . . . 3.6 Investment Destination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6.1 Communication Costs in Destination Markets . . . . . . . . . . . . 3.6.2 Investment Decision by Destination Income . . . . . . . . . . . . . . 3.6.3 Threshold Estimates of the Linder Hypothesis . . . . . . . . . . . . 3.7 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix: Supplementary Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

41 41 45 48 49 49 50 50 53 54 56 60 62 65 66 68 71 72 75 77 78 78 79

4 Outward FDI and Domestic Input Distortions: Evidence from Chinese Firms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Motivation and Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Data and Stylised Facts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.1 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.2 Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.3 Stylised Facts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.1 Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.2 Domestic Production, Exporting and FDI . . . . . . . . . . . . . . . . 4.3.3 Domestic Distortion and Patterns of Outward FDI . . . . . . . . 4.4 Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.1 FDI Decision and Firm Ownership . . . . . . . . . . . . . . . . . . . . . 4.4.2 Input Market Distortions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.3 Channels and Sectoral Heterogeneity . . . . . . . . . . . . . . . . . . . 4.4.4 Capital Intensity and Pattern of Outward FDI . . . . . . . . . . . . 4.4.5 Estimates at the Intensive Margin . . . . . . . . . . . . . . . . . . . . . . .

81 81 86 86 88 90 97 97 99 101 105 105 108 112 115 117

Contents

vii

4.4.6 Outward FDI Data Between 2000 and 2013 . . . . . . . . . . . . . . 120 4.5 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 5 Does Outward FDI Generate Higher Productivity for Emerging Economy MNEs?—Micro-level Evidence from Chinese Manufacturing Firms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Literature Review and Hypothesis Development . . . . . . . . . . . . . . . . 5.2.1 OFDI and EMEs’ Productivity Growth . . . . . . . . . . . . . . . . . . 5.2.2 State Ownership and OFDI’s Productivity Effect on EMEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.3 Absorptive Capacity and OFDI’s Productivity Effect on EMEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.4 Investment Destination and OFDI’s Productivity Effect on EMEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.1 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.2 Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.3 Econometric Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4 Estimation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.1 Results At the Overall Manufacturing Level . . . . . . . . . . . . . . 5.4.2 State Ownership and OFDI’s Productivity Effect on EMEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.3 Absorptive Capacity and OFDI’s Productivity Effect on EMEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.4 Investment Destination and OFDI’s Productivity Effect on EMEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5 Robustness Check and Further Analysis . . . . . . . . . . . . . . . . . . . . . . . 5.5.1 An Alternative Measure of Total Factor Productivity . . . . . . 5.5.2 Investment Destination Measured by Patent Application Per Capita . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5.3 One-Step System GMM Approach to Estimate the OFDI’s Productivity Effect . . . . . . . . . . . . . . . . . . . . . . . . . 5.5.4 Absorptive Capacity and OFDI’s Productivity Effect in Non-technology-Intensive Industries . . . . . . . . . . . . . . . . . . 5.6 Discussion and Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.1 Theoretical Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.2 Policy and Managerial Implications . . . . . . . . . . . . . . . . . . . . . 5.6.3 Limitations and Future Research Directions . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

125 125 128 136 139 141 142 143 143 145 146 149 149 150 153 153 154 154 154 154 155 155 156 157 158 158

6 Outward Direct Investment, Firm Productivity and Credit Constraints: Evidence from Chinese Firms . . . . . . . . . . . . . . . . . . . . . . . 165 6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 6.2 Data and Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168

viii

Contents

6.2.1 Firm-Level Data in Zhejiang Province . . . . . . . . . . . . . . . . . . 6.2.2 Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3 Firm Heterogeneity and ODI Decision . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.1 Model Specification and Basic Results . . . . . . . . . . . . . . . . . . 6.3.2 Interaction Effect Between Productivity and Financial Constraint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.3 Robustness Check: First Time Exporting or Outward Direct Investment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4 Firm Heterogeneity and Outward Direct Investment Value . . . . . . . . 6.4.1 Model Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 The Potential Impact of China–US BIT on China’s Manufacturing Sectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2 Assumptions on the Scenarios of China–US BIT, Focusing on Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3 BIT’s Open Market Requirements to China’s Manufacturing Sector and Its Impacts on Relevant Industries . . . . . . . . . . . . . . . . . . . 7.3.1 Impacts of FDI on Domestic Firms . . . . . . . . . . . . . . . . . . . . . 7.3.2 FDI’s Overall Impacts on Performance of Domestic Firms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3.3 FDI’s Impacts on Specific Industries . . . . . . . . . . . . . . . . . . . . 7.3.4 The Effects of Changes in Policies on Scale or Shares of FDI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4 Suggestions on Negotiation Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4.1 Make It Firm and Steadfast That China Is Serious in Joining BIT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4.2 Protection Measures in the Long-Run . . . . . . . . . . . . . . . . . . . 7.4.3 Gradual Lifting Process of Protection for Certain Vulnerable Sectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4.4 Cooperating in BIT Negotiation with the Domestic Reform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5 Suggestions to Manufacturing Firms Regarding How to Face the Challenges of BIT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5.1 Suggestions for Domestic Firms . . . . . . . . . . . . . . . . . . . . . . . . 7.5.2 Suggestions for Government . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

168 169 171 171 172 174 175 175 175 176 179 181 182 183 183 183 184 186 191 191 191 191 192 192 192 192 193 193 193

8 China’s Opening-Up Policies: Achievements and Prospects . . . . . . . . . 201 8.1 Expanding the Extensive Margin, 1978–2000 . . . . . . . . . . . . . . . . . . . 202 8.1.1 Setting up Special Economic Zones and Industrial Parks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203

Contents

8.1.2 Relaxing Market Access for Foreign Direct Investment . . . . 8.1.3 Reducing Import Tariffs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1.4 Encouraging the Processing Trade . . . . . . . . . . . . . . . . . . . . . . 8.1.5 Comments on the Stage of External Opening . . . . . . . . . . . . . 8.2 Internal Opening up, 2001–2017 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2.1 Accession to the WTO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2.2 Expanding Market Access for FDI . . . . . . . . . . . . . . . . . . . . . . 8.2.3 Relaxation of Outward FDI . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2.4 Establishment of Pilot Free Trade Zones . . . . . . . . . . . . . . . . . 8.2.5 New-Economy Pilot Cities Experiment . . . . . . . . . . . . . . . . . . 8.2.6 Comments on the Stage of Intensive Margin of Opening up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3 Features of the All-Around Opening up . . . . . . . . . . . . . . . . . . . . . . . . 8.3.1 Belt and Road Initiative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.2 Free Trade Ports Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.3 Greater Bay Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4 Policy Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

203 204 205 206 207 207 208 209 211 212 212 214 214 214 215 215 217 218

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China and Cross-Countries’ Evidence” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 Bibilography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

Chapter 1

The Exceptional Performance of Chinese Outward Direct Investment Firms

Abstract This chapter finds that Chinese manufacturing firms that engage in outward foreign direct investment (ODI) have better economic performance than nonODI manufacturing firms. Overall, ODI firms are more productive and have higher pro fit ability than non -ODI firms. The sector analysis shows that the exceptional performance is significant for labor intensive industries. Finally, the ODI activity can raise the productivity of other firms in an industry. The larger the ODI within an industry, the higher the productivity of all firms in that industry. The chapter suggests that domestic firms set up their firm’s global strategy and reallocate the firm’s resources according to the changing investment environment, taking advantages of profit opportunities outside of domestic markets and invest abroad to get new markets and new technology.

1.1 Introduction China’s outward foreign direct investment (ODI) has increased dramatically in the new century. With a 50% annual growth rate, China’s outward foreign direct investment (FDI)has become economically significant to affect international investment (Huang & Wang, 2011). China’s non-financial outward FDI increased from $29.9 billion in 2002 to $326.5 billion in 2011, a nearly ten-fold increase during the period.1 In 2014, China’s outward FDI flow accounted for 7.6% of global FDI flow and ranked third in the world, following the United States and Japan, and first among developing countries (Tian & Yu, 2015). As documented by Chen et al. (2016), there are more than 15,000 Chinese multinational corporations (MNC) now, which is almost equal to the number of MNCs of any developed economy in the world. In 2014, ODI flows from China were USD 140 billion, surpassing the amount of USD 119 billion of inward FDI flows to China. 1 China’s non-financial investment (i.e. ‘green-field’ investment) outweighs the financial investment (i.e. investment from mergers and acquisitions). In 2011, China’s non-financial investment accounted for 91.8% of its entire foreign investment.

This chapter is published in China Economic Journal by Tian Wei, Miaojie Yu and Fan Zhang. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022 W. Tian and M. Yu, Outward Foreign Direct Investment of Chinese Enterprises, Contributions to Economics, https://doi.org/10.1007/978-981-19-4719-3_1

1

2

1 The Exceptional Performance of Chinese Outward …

The fast increase in Chinese outward direct investment (ODI) in recent years has also become a controversial topic in many countries, especially in the U.S. The China–U.S. bilateral investment treaties (BIT) is focused on market access of foreign investment, which will give Chinese firms more opportunities to invest overseas. The present chapter finds that Chinese manufacturing firms that engaged in ODI have better economic performance than non-ODI manufacturing firms. We present three interesting findings. First, ODI firms are more productive and have higher profitability than non-ODI firms. Second, the exceptional effect is more for light industries or labor- intensive industries. Finally, ODI activity can raise the productivity of other firms in an industry. The larger the ODI within an industry, the higher the productivity of all firms in that industry. To fully understand the role of China–U.S. bilateral investment treaties, this chapter seeks to understand whether outward investment activity is good for Chinese manufacturing firms. Put another way, whether Chinese ODI firms are exceptional in terms of economic indicators such as firm productivity and firm profitability. If so, it would be beneficial for Chinese firms to go aboard to invest to other countries. Correspondingly, it would be wise for the Chinese government to actively engage in the ongoing Sino–U.S. BIT negotiation given that the United State is one of the most important current ODI destinations for Chinese firms. The rest of the chapter will be organized as follows. The next section reviews the most important literature related to the current chapter. Sect. 1.3 describes the data used in the chapter, followed by the empirical specifications to explore the possible exceptional performance of Chinese ODI firms. Sect. 1.4 delivers some policy suggestions and, finally, the last section draws some conclusions.

1.2 Literature Review The present study contributes to a small but growing literature on outward FDI and productivity, including the works by Head and Ries (2003), Helpman et al. (2004), Eaton et al. (2011), and Damijan et al. (2007). Moreover, the present chapter further contributes to the literature from the following perspectives. First, different from other related studies that use aggregate industry-level data, this chapter aims to provide micro-level evidence of outward FDI for the largest developing country in the world. Previous works usually examine the industrial characteristics of outward FDI but abstract away the role of firm activity due, possible, to data restrictions. The effect of firm activity on outward FDI is still treated as a ‘black-box.’ In contrast, our firm-level analysis finds that outward FDI by Chinese firms also exhibits a positive correlation between firm productivity and outward FDI. Second, we use a panel of firm-level data to measure the effect of ODI on productivity, which is different from previous studies that only used cross-section data, such as Head and Ries (2003) and Helpman et al. (2004). By gauging outward FDI in the model of firm heterogeneity on productivity as described by Melitz (2003), Helpman et al. (2004) predict that only highly productive American firms will engage

1.3 Data and Empirics

3

in outward FDI, whereas low-productivity firms will serve in the domestic market only. Firms with intermediate productivity sell products domestically and export, but they cannot serve the foreign market through foreign affiliates. However, because of the data restriction, the previous study is not able to capture firms’ dynamic response of outward FDI to productivity growth. Such empirical pitfall is overcome in the present chapter using a careful measure of total factor productivity (TFP) over the years. The third related strand of the literature is research on China’s ODI. Huang and Wang (2011) argue that Chinese ODI firms have different objectives for their investment. In echoing this, Kolstad and Wiig (2012) find that Chinese ODI is attracted to three destinations: Countries with lower institutional quality, countries that are rich in natural resources, and large markets. Most recent related works tend to explore what determines the ODI of Chinese firms. Chen and Tang (2014) also found that firm productivity and the probability of firm ODI are positively correlated, yet, because of lack of data, they remain silent on the intensive margin of firms’ ODI. Tian and Yu (2015) intensively explore the pattern of distribution-oriented ODI in China and its correlation between export and horizontal ODI. Finally, our work is also related to the work of Liu et al. (2015) which also explored whether outward investment boosts firm productivity, though we also focus on the analysis of sectoral heterogeneity. Finally, Chen et al. (2019) examines how domestic input distortions provide an institutional arbitrage for Chinese private firms to invest abroad. Similarly, Wang et al. (2016) explore how the credit constraints faced by private firms affect their ODI behavior. Our present chapter instead aims to explore whether or not the ODI firms have better economic performance.

1.3 Data and Empirics This section starts from a careful data description, followed by our empirical analysis.

1.3.1 Data The data set we used included investment categories by the Ministry of Commerce of China and firm data set by the National Statistical Bureau of China. Firm’s ODI data are from the Ministry of Commerce. The data set includes the following variables: The firm’s name, the names of the firm’s foreign subsidiaries, the type of ownership (i.e. private firm or state-owned enterprise), the investment mode (e.g. trading-oriented affiliates, mining-oriented affiliates), and the amount of foreign investment (in U.S. dollars). The data set is on an annual base and can be traced back to 1980.

4 Table 1.1 Summary statistics of data set, firm level

1 The Exceptional Performance of Chinese Outward … Variable

Mean

Standard Deviation

Log (Labor)

4.71

1.12

Log (Asset)

9.73

1.45

Firm Productivity

1.88

0.57

Whether to Export

0.25

0.43

Whether SOE

0.10

0.30

Whether FIE

0.20

0.40

Notes 2000–2008. Productivity is calculated using an augmented Olley-Pakes approach. FIE includes FIE from Hong Kong, Taiwan and Macao

The second data set that we relied on was the firm-level production data compiled by China’s National Bureau of Statistics in an annual survey of manufacturing enterprises. The data set covers around 162,885 firms in 2000 and 410,000 firms in 2008 and, on average, accounts for 95% of China’s total annual output in all manufacturing sectors. The data set includes two types of manufacturing firms: Universal SOEs and non-SOEs whose annual sales are more than RMB 5 million (or equivalently $830,000 under the current exchange rate). The data set is particularly useful for calculating measured TFP, since the data set provides more than 100 firm-level variables listed in the main accounting statements, such as sales, capital, labor, and intermediate inputs. As highlighted by Feenstra et al. (2014) and Yu (2015), some samples in this firm-level production data set are noisy and somewhat misleading, largely because of misreporting by some firms. To guarantee that our estimation sample is reliable and accurate, we screened the sample and omitted outliers by adopting the following criteria. First, we eliminated a firm if its number of employees was less than eight workers, since otherwise such an entity would be identified as self-employed. Second, a firm was included only if its key financial variables (e.g. gross value of industrial output, sales, total assets, and net value of fixed assets) were present. Third, we included firms based on the requirements of the Generally Accepted Accounting Principles (GAAP).2 After this rigorous filter, around half the observations are deleted from the sample. The first row in the upper module of Table 1.1 reports the number of manufacturing firms in 2000–08; the second row reports the number of ODI manufacturing firms after the filtering process. Table 1.1 reports some key summary statistics of firms’ characteristics for nationwide manufacturing firms. Although the firm-level production data set is useful to understand the firm’s production behavior, the data set does not provide information on the firm’s ODI 2

In particular, an observation is included in the sample only if the following observations hold: (1) total assets are higher than liquid assets; (2) total assets are larger than the total fixed assets and the net value of fixed assets; (3) the established time is valid (i.e. the opening month should be between January and December); and (4) the firm’s sales must be higher than the required threshold of RMB 5 million.

1.3 Data and Empirics

5

behavior. We do not know whether the firms engage in ODI activity. Therefore, to understand a firm’s ODI behavior, we need to merge the firm-level production database with the ODI data set. However, the data merge is a non-trivial task. Although the two data sets share a common variable—the firm’s identification number—their coding system is completely different. We thus use the following approaches to merge the data. First, we match these two data sets by using each firm’s Chinese name and year. If a firm has an exact Chinese name in a particular year, it is considered an identical firm. Still, this method could miss some firms since the Chinese name for an identical company may not have the exact Chinese characters in the two data sets, although they share some common strings. Our second step is to decompose a firm name into several strings referring to its location,3 industry, business type, and specific name, respectively. If a company has all identical strings, such a firm in the two data sets is classified as an identical firm.4 Finally, to avoid possible mistakes, all approximate string-matching procedures are double-checked by eye.

1.3.2 Empirical Results To explore the performance of ODI firms, we first ran a regression model in which the dependent variable is the log of total factor productivity(TFP), and the key interest explanatory variable is the firm’s ODI indicator, a dummy showing whether the firm engages in overseas direct investment. Note that the TFP measure is based on the augmented Olley and Pakes (1996) approach, which strictly follows the method proposed by Yu (2015).

1.3.3 Benchmark Estimates Column (1) of Table 1.2 abstracts away other control variables but only includes the ODI dummy as the independent variable. The positive and statistically significant coefficient of the ODI dummy suggests that ODI firms tends to more productive. As our sample was spread over the period 2000–2008, we thus included both yearspecific fixed effects and firm-specific fixed effects in the estimates of column (2) and still have similar findings. 3

For example, ‘Ningbo Hangyuan communication equipment trading company’ shown in the ODI data set and ‘(Zhejiang) Ningbo Hangyuan communication equipment trading company’ shown in the National Bureau of Statistics of China production data set are the same company but do not have exactly the same Chinese characters. 4 In the example above, the location fragment is ‘Ningbo,’ the industry is ‘communication equipment,’ the business type is ‘trading company,’ and the specific name is ‘Hangyuan.’.

6

1 The Exceptional Performance of Chinese Outward …

Table 1.2 Effects of ODI on firms’ productivity Dependent Var. ln TFP ODI

(1)

(2) 0.257*** (18.66)

(3) 0.013** (2.16)

(4) 0.054*** (3.91) 0.086***

FIE

(73.15) 0.037***

SOE

(26.46) Export

0.042*** (38.36) − 0.050***

Log(labor)

(−95.13) 0.113***

Log(asset)

(280.78)

0.011* (1.72) 0.008* (1.93) − 0.023*** (−7.72) 0.024*** (16.23) 0.015*** (14.39) 0.017*** (18.52)

FE Year

N

Y

N

Y

FE Firm

N

Y

N

Y

2,096,406

2,096,406

1,661,369

1,661,369

Number of Obs R-squared

0.00

0.06

0.07

0.05

Notes TFP is calculated using an augmented OP approach, t statistics are reported in parentheses. ODI is the dummy indicating whether the firm engages in ODI; FIE is the dummy indicating whether a firm is foreign invested; Export shows whether a firm engages in export; SOE indicates whether a firm is state owned; Labor is the total number of employees in a firm; Asset is the total value of assets in a firm. ***, **, and * show significance at 1, 5, and 10%, respectively

Estimates in the remaining two columns of Table 1.2 include two additional sets of variables which are included in the regressions. The first set is related to the type of firm’s ownership. The variable of foreign firm (FIE) is a dummy indicating whether a firm is foreign invested. The SOE dummy indicates whether a firm is state-owned enterprise. The second set of controlling variables is the firm’s size. We thus include the number of firm’s employees (labor) and firm’s total asset (asset) to the regressions. In addition, previous works like Blonigen (2001) show that export and FDI are strongly related. Ignoring a firm’s export behavior may generate some estimation bias, we thus also include the export indicator to the estimations. With other controlling variables, the fixed-effects estimate in the last column of Table 1.2 presents the positive effect of OFDI on the productivity of the parent company, suggesting that ODI firms are more productive. Compared to non-ODI firms, ODI firms have 0.11 higher productivity. As the mean of the firm’s Olley-Pakes TFP is around 1.88, ODI firms thus, on average, exhibit 6% greater productivity than non-ODI firms. In addition, we also find that SOEs are less productive whereas foreign firms are more productive. Larger firms, measured by both logarithm of the

1.3 Data and Empirics

7

number of employees and the logarithm of firm assets, are more productive. Exporters are also more productive. Such findings are highly consistent with other related works such as Dai et al. (2016), Feenstra et al. (2014) and Yu (2015).

1.3.4 Alternative Measure of Firm Performance The estimates in Table 1.2 use firm productivity as the measure of firm performance as it is a widely-accepted measure. We now turn to see whether our benchmark finding is still robust using other alternative measures. The first alternative measure of firm performance is the profit-sales ratio which is defined as a firm’s profit over sales. The economic intuition is that high productive firms are usually more profitable. Column (1) of Table 1.3 uses firm’s profit-sales ratio as the regressand and controls both firm-specific and year-specific fixed effects. It turns out that the key coefficients of the ODI indicator is positive and statistically significant. Column (2) of Table 1.3 includes other control variables but still finds a positive, though insignificant, coefficient of the ODI indicator. Previous works have recognized that exceptional firms usually export more products. To check this out, we used a firm’s export intensity, defined as export value over total sales, as the regressand in the rest of Table 1.3. By abstracting away other control variables, the key coefficient of the ODI indicator in the fixed-effects estimates of column (3) is still positive and significant, indicating that ODI firms tend to have a higher export intensity. After including more controlling variables, the coefficient of the ODI dummy in column (4) is still positive but insignificant. We suspect that this is partly due to the inclusion of the export dummy as the regressand.

1.3.5 Robustness Checks After 2004 it is generally believe that China has already passed the Lewis turning point in the sense that China’s labor supply is no longer infinite. Studies like Cai (2010) also suggest that China’s labor cost increased quickly in the first decade of the new century. If so, one would expect that China’s comparative advantage in the labor-intensive industry will shrink over time. Accordingly, Chinese manufacturing firms in the labor-intensive industry will engage more in outward FDI. Estimates in Table 1.4 pick up the task of examining such a hypothesis. Table 1.4 first separates the entire sample to the group: Labor-intensive sectors and capital-intensive sectors. In particular, all manufacturing sectors with Chineseindustrial classification (CIC) 2-digit level higher than 24 are classified as the capitalintensive industries, whereas the rest of the manufacturing sectors are classified as labor-intensive industries. Estimates in column (1) only include firms in the textile and garment industry. The key coefficient of the ODI indicator is positive but insignificant. We suspect that this is due to the fact that not many firms in the textile and

8

1 The Exceptional Performance of Chinese Outward …

Table 1.3 Effects of ODI on firms’ profitability and export propensity Dependent Var ODI

Profit-Sales Ratio 8.550** (2.20)

Profit-Sales Ratio 1.890 (0.81)

Export Propensity 1.473*** (2.79)

− 4.889**

FIE

Export

(0.38)

-17.648***

− 0.144

(-4.10)

(−0.50)

40.951

38.726***

(1.42) Log(Labor) Log(Assets)

0.088 (0.19) 0.105

(-2.41) SOE

Export Propensity

(58.57)

11.059

− 0.084

(0.65)

(−0.69)

16.150

− 0.079

(1.51)

(−0.57)

FE Year

Y

Y

Y

Y

FE Firm

Y

Y

Y

Y

Number of Obs

2,251,355

1,684,364

2,195,895

1,654,682

Notes t statistics are reported in parentheses. Sales profit rate = operating profit/sales × 100, Export propensity = export/sales × 100. ODI is the dummy indicating whether the firm engages in ODI; FIE is the dummy indicating whether a firm is foreign invested; Export shows whether a firm engages in export; SOE indicates whether a firm is state owned; Labor is the total number of employees in a firm; Asset is the total value of assets in a firm. ***, **, and * show significance at the 1, 5, and 10% level, respectively

garment industry engage in oversea investments. We thus include all labor-intensive industries in the estimates of column (2). It turns out that the coefficient of the ODI dummy now is positive and significant, suggesting that, overall, Chinese manufacturing ODI firms in labor-intensive industries tends to more productive compared to those non-ODI firms in labor-intensive industries. However, we do not see such a corresponding finding in the capital-intensive industries as shown in columns (3) and (4). Our last empirical exercises in Table 1.5 are to examine whether our previous findings are still robust if we aggregate the firm-level data up to the CIC 2-digit industry level. We first calculate industrial ODI share as the number of ODI manufacturing firms dividing by the entire number of manufacturing firms within an industry. Columns (1)–(4) regress firm productivity on industrial ODI share. It turns out that the key variable of industrial ODI share is positive and significant, suggesting that ODI behavior can raise the productivity of other firms in an industry. The larger the ODI within an industry, the higher the productivity of all firms in that industry. A possible explanation is that ODI raises the productivity of the parent firm, which in turn intensifies the competition in that industry and the technology spillovers, and raises the productivity of other firms in the industry.

1.4 Policy Suggestions

9

Table 1.4 Effects of ODI on firms’ productivity in light and heavy industries Dependent Var. ln TFP

ODI

Labor-intensive Industry

Capital-intensive Industry

Textile

General Equipment Manufacturing

All Labor-intensive Industries 0.005

(0.49)

0.030** (2.49)

0.015

0.008

(1.18)

(0.81)

0.016

0.010

FIE

0.002 (0.15)

(1.85)

SOE

− 0.019**

− 0.012**

− 0.029***

− 0.019***

(−2.33)

(−2.09)

(−3.10)

(−4.19)

Export

0.026*** (6.78)

Log(Labor)

0.008*** (3.04)

Log(Asset)

0.012*

All Capital-intensive Industries

0.031*** (11.88) 0.021*** (12.33)

(1.10)

0.017*** (3.77) 0.009** (2.26)

0.019*** (7.11) 0.017*** (8.22)

0.015***

0.018***

0.008***

0.021***

FE Year

Y

Y

Y

FE Firm

Y

Y

Y

Y

133,998

515,635

120,245

60,0781

Ob R-squared

0.03

0.03

0.027***

(1.15)

(10.06)

0.14

0.019*** (10.33) Y

0.05

Notes TFP calculated using an augmented OP approach, t statistics are reported in parentheses. ***, **, and * show significance at the 1, 5, and 10% level, respectively. The light and heavy industries are defined according to firms’ registration

1.4 Policy Suggestions Generally, the Sino–U.S. bilateral investment treaty (BIT) will be beneficial for Chinese firms to invest overseas and become global players, which will be beneficial to Chinese domestic firms (for a detailed discussion see Yu & Zhang, 2016). Current U.S. foreign investment review procedure is less transparent and limited the outward direct invest-ment (ODI) from other countries, including investments by Chinese firms. Currently, the investment is investigated and approved by the Committee on Foreign Investment in the United States (CFIUS), which is an interagency group led by the Treasury Department to ensure acquisitions of U.S. firms do not harm U.S. national security. In certain industries foreign investment is explicitly limited and prohibited by the U.S. government. A number of Chinese investment deals in the U.S. are facing obstacles currently. A BIT is unlikely to change the CFIUS process but will increase the trans-parency of the process, which gives Chinese firms more opportunities to enter the U.S. market. The Sino–U.S. BIT will also be beneficial to China’s domestic reforms of ODI framework. China’s regulatory framework has moved from restricting, to facilitating,

10

1 The Exceptional Performance of Chinese Outward …

Table 1.5 Indirect effect of ODI on firms’ productivity Dependent Var. ln TFP ODI share

(1)

(2) 1.301*** (137.68)

(3) 1.697*** (32.61)

(4) 1.083*** (99.70) 0.075***

FIE

(63.52) 0.054***

SOE

(38.61) Export

0.022*** (20.17) − 0.054***

Log(Labor)

(−103.26) 0.115***

Log(Asset)

(288.00)

1.703*** (28.16) 0.007* (1.74) − 0.023*** (−7.70) 0.024*** (16.22) 0.015*** (14.28) 0.017*** (18.28)

FE Year

N

Y

N

FE Industry

N

Y

N

Y

2,096,406

2,096,406

1,661,369

1,661,369

Ob R-squared

0.01

0.06

0.07

Y

0.06

Notes In this table, the division of industry is according to the 2-digit industrial code in ‘Classification of national economic industries’ (GB/T 4754—2002, 2nd edition). The division of industry before 2003 is according to the ‘Classification of national economic industries and code’ (GB/T4754) and is transferred to the second edition. TFP was calculated using an augmented OP approach, t statistics are reported in parentheses. ***, **, and * show significance atthe1, 5, and 10% level, respectively. ODI share = share of ODI in 2-digit industry × 100

to supporting, to encouraging ODI; but there are still strong elements of administrative control that do not give Chinese firms enough incentives to invest overseas. To allow domestic regulations to be consistent with the BIT, China has to make the domestic regulations on ODI more transparent and simplify the process. The reform of the domestic regulations will also increase the ODI by Chinese firms. In the long-run, BIT and ODI are beneficial for Chinese firms to improve their productivity and profitability. Therefore, Chinese negotiators must make it firm and steadfast that China is serious in joining the BIT with the U.S. through negotiations. In the short run, however, joining BIT will hurt some of the Chinese firms or industries, even though it is beneficial for Chinese firms to improve their productivity and profitability in general. Studies on industries show that firms in some of the manufacturing industries, especially those with large gaps in technology, will be harmed in the shortrun. Therefore, it may need a gradual lifting of the protection in a small number of industries. The Chinese negotiators should insist that the U.S. side needs to make its national security investigation more transparent and the U.S. government should create a more favorable investment climate for foreign investments.

1.5 Conclusions

11

The Chinese government will use the requirement of BIT to reform domestic administration, the judicial system and the state-owned enterprises, especially domestic regulations on ODI. To do that, domestic laws and regulations need to cooperate with the foreign economic policies. First, China needs to institute easier procedures for firms undertaking ODI, by granting firms greater independence in their decision making. Second, the Chinese government needs to clarify the ODI policies. Third, government services related to ODI should be strengthened. Domestic firms need to make contingency plans to meet the challenges of BIT. They need to update their technology, reduce costs, and learn management skills from their foreign competitors. Domestic firms need to do researches on BIT and gain benefit from it, to learn how to use legal means, including the dispute settlement clauses in BIT, to protect their interests. Chinese investors also have to learn how to deal with public opinion. Domestic firms need to set up their firm’s global strategy and reallocate its resources according to the changing investment environment, taking advantage of profit opportunities outside of domestic markets. When the domestic competition becomes more intensive, they must go abroad to find new markets and new technology by outward investment. Like a country can do it on the macro level, firms can set up a global strategy on the micro level, which includes a wide range of activities such as overseas manufacturing, outward foreign investing, and importing. In the times of BIT, firms need to make much broader decisions about whether or not to engage in foreign trade and foreign investment, what specific foreign markets should be served, and how to participate in chosen markets. The government should provide more detailed information to firms about the changes made by BIT and provide financial supports to assist firms to make their structural adjustments.

1.5 Conclusions Using various econometric models and a large firm-level data set, we find the overall effect of ODI and thereby BIT on the Chinese manufacturing sector is positive for firms in their outward investment. As the evidence shows, the ODI raised the productivity and profitability of the firms significantly in the manufacturing sector. Such findings have rich policy implications. Chinese domestic firms need to update their technology, reduce costs, and learn management skills from their foreign competitors. They need to learn to gain benefits from BIT, using the national treatment terms in BIT to enter the foreign markets. Also, Chinese domestic firms also need to set up firms’ global strategies and reallocate their resources according to the changing investment environment, taking advantage of profit opportunities outside of domestic markets.

12

1 The Exceptional Performance of Chinese Outward …

References Blonigen, B. A. (2001). In search of substitution between foreign production and exports. Journal of International Economics, 53(1), 81–104. Cai, F. (2010). Demographic transition, demographic dividend, and Lewis turning point in China. China Economic Journal, 3(2), 107–119. Chen, C., Tian, W., & Yu, M. (2019). Outward FDI and domestic input distortions: Evidence from Chinese firms. The Economic Journal, 129(624), 3025–3057. Dai, M., Maitra, M., & Yu, M. (2016). Unexceptional exporter performance in China? The role of processing trade. Journal of Development Economics, 121, 177–189. Damijan, J. P., & Rojec, M. (2007). Foreign Direct Investment and Catching Up of New EU Member States: Is There A Flying Geese Pattern? Applied Economics Quarterly, 53, 91–118. Eaton, J., Kortum, S., & Kramarz, F. (2011). An anatomy of international trade: Evidence from French firms. Econometrica, 79(5), 1453–1498. Feenstra, R. C., Li, Z., & Yu, M. (2014). Exports and credit constraints under incomplete information: Theory and evidence from China. Review of Economics and Statistics, 96(4), 729–744. Head, K., & Ries, J. (2003). Heterogeneity and the FDI versus export decision of Japanese manufacturers. Journal of the Japanese and International Economies, 17(4), 448–467. Helpman, E., Melitz, M. J., & Yeaple, S. R. (2004). Export versus FDI with heterogeneous firms. American Economic Review, 94(1), 300–316. Huang, Y., & Wang, B. (2011). Chinese outward direct investment: Is there a China model? China & World Economy, 19(4), 1–21. Kolstad, I., & Wiig, A. (2012). What determines Chinese outward FDI? Journal of World Business, 47(1), 26–34. Liu, X., Li, L., Yu, M. & Yuan, D. (2015). Does outward FDI generate higher productivity for emerging economy MNEs? Micro-level evidence from Chinese manufacturing firms. International Business Review, 26 (2017), 839–854. Melitz, M. J. (2003). The impact of trade on intra-industry reallocations and aggregate industry productivity. econometrica, 71(6), 1695–1725. Tian, W., & Yu, M. (2015). Processing trade, export intensity, and input trade liberalization: Evidence from Chinese firms. Journal of the Asia Pacific Economy, 20(3), 444–464. Wang, B., Tan, Y., Yu, M., & Huang, Y. (2016). Outward direct investment, firm productivity and credit constraints: Evidence from Chinese firms. Pacific Economic Review, 21(1), 72–83. Yu, M. (2015). Processing trade, tariff reductions and firm productivity: Evidence from Chinese firms. The Economic Journal, 125(585), 943–988. Yu, M., & Zhang, F. (2016). The potential impact of China–US BIT on China’s manufacturing sectors. China Economic Journal, 9(1), 47–64.

Chapter 2

Firm Productivity and Outward Foreign Direct Investment: A Firm- Level Empirical Investigation of China

2.1 Introduction Foreign direct1 investment (FDI) has long been a driving force of China’s economy. China is the world’s largest country in absorbing FDI, accounting for one-third of the total FDI flows to developing countries. However, China is also a major capital supplier in the world. Although the volume of outward FDI (OFDI) is relatively small compared with inward FDI, its growth rate is considerable. In 2005 alone, the growth rate of OFDI reached 32%, and many large multinational enterprises in China are playing an increasingly important role in international business. According to the United Nations estimates, China’s OFDI exceeds 3% of gross domestic product (GDP). In the 1980s, only state-owned enterprises and some supervised enterprises were permitted to invest abroad. Since the deregulation in the 1990s, China’s “going out” strategy as well as the corresponding policy incentives and information assistance have promoted a large number of Chinese firms that are not state-owned enterprises to expand markets overseas, especially in the fields of home appliances, electronics, and communications. Most of these firms are located in China’s eastern coastal cities and in the manufacturing, commercial, and mining industries. These foreign investments are mainly directed to areas such as Hong Kong and other “tax havens” (see Poncet, 2007). Figure 2.1 depicts the macro trend of China’s OFDI since the twenty-first century. According to the Statistical Bulletin of China’s Foreign Direct Investment in 2009, by the end of 2009, more than 12,000 domestic enterprises in China had made direct overseas investments in 177 countries and regions, with a cumulative net investment of approximately US$245.75 billion, with nonfinancial investment accounting for 84.5%. In 2009, China’s FDI accounted for 5.1% of the global total flows and 1.3% of the stock, ranking fifth in the world in terms of flows. Investments mainly went to services, mining, finance, wholesale and retail, manufacturing, and transportation, which together accounted for 93.8% of total investment for the year. Investment in 1

This chapter is published in China Economic Quarterly by Tian Wei and Miaojie Yu.

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022 W. Tian and M. Yu, Outward Foreign Direct Investment of Chinese Enterprises, Contributions to Economics, https://doi.org/10.1007/978-981-19-4719-3_2

13

14

2 Firm Productivity and Outward Foreign Direct Investment …

Fig. 2.1 Macro trend of China’s OFDI since the twenty-first century

manufacturing industries accounted for about 4%, and the number of manufacturing enterprises accounted for about 32% of the total number of enterprises investing overseas. In terms of the destination of investment, 71.4% went to other Asian countries.2 Local investment has increased significantly, and the number of local enterprises accounted for about 95% of the total, among which the eastern and central regions, such as Hunan, are still the main sources of local direct investment. The number of central enterprises accounted for only about 5%, gradually giving way to local enterprises, but the flow of investment from central enterprises accounted for more than 60% of the total investment. There are many determinants of China’s FDI. At present, there is no systematic empirical literature on China’s OFDI. However, as shown in the questionnaire launched by the Foreign Investment Advisory Service (2005), factors in the Chinese market, policies in the destination countries, and policies in China have an impact on enterprises’ OFDI decisions. First, insufficient domestic demand and overproduction motivate enterprises to seek other markets. However, exports in many industries are restricted by quotas. Direct investment in factories is an effective way to avoid the quotas. If this is true, countries with large exports and large trade surpluses will also see greater growth in OFDI. Second, on the one hand, the direct investment flowing into China intensifies the competition among domestic firms, forcing them to turn to other markets; on the other hand, it brings technology and capital, which effectively helps Chinese firms to build factories overseas. Third, some large firms have established research and development (R&D) centers in developed countries to learn their scientific research techniques. Fourth, some enterprises in the oil refining, natural gas, and mineral industries invest in resource-rich countries to obtain sufficient raw materials. These are mainly state-owned enterprises. Fifth, the preferential policies of the destination countries are also very attractive to Chinese enterprises. For example, the 2

Among other regions, 13% went to Latin America, 5.9% to Europe, 4.4% to Oceania, 2.7% to North America, and 2.6% to Africa.

2.1 Introduction

15

United Kingdom’s investment promotion policy has attracted many Chinese enterprises. Sixth, the Chinese government’s policy changes have also played a significant role. For example, China has signed more than 100 bilateral investment agreements, which also promote China’s OFDI to a certain extent. The analysis shown in the survey by the Foreign Investment Advisory Service is based on the macro level. A further question is what kinds of enterprises contribute to the increasing OFDI. Controlling for the macroeconomic environment, industry factors, and the influence of different destination countries, what kinds of enterprises are more likely to choose OFDI? What is the relationship between the firm’s productivity and its foreign investment behavior? These complex differences at the enterprise level are difficult to explain clearly using information from a general survey. Therefore, it is necessary to use microscopic data at the enterprise level and perform rigorous analysis. The related literature, such as Montagna (2001) and Melitz (2003), has made theoretical predictions on these issues through rigorous argumentation: within the same industry, firms with the lowest productivity only serve the domestic market; firms with higher productivity both serve domestic markets and export; and the most productive firms serve domestic markets, export, and invest overseas. Empirically, Helpman et al. (2004) study the impact of firms’ productivity on their export and FDI decisions. They find that for U.S. firms, exporting firms were about 39% more productive than those that sold only domestically, and firms investing overseas were about 15% more productive than exporting firms, consistent with the classical theoretical prediction of Melitz (2003). Head and Ries (2003) use data on Japanese enterprises to conduct an empirical test. They find that the enterprises with the highest productivity were still investing overseas, the enterprises with lower productivity chose to export, and the enterprises with the lowest productivity only served domestic markets. More interestingly, they also find that low-productivity firms are attracted to low-income countries, while high-productivity firms mainly invest in high-income countries. Eaton et al. (2004) analyze French enterprise data. They find that the number of destination countries is proportional to enterprise productivity in terms of exporting and FDI, indicating that the entry cost and degree of competition in different countries have a significant impact on the decision making of enterprises. Damijan et al. (2007) analyze Slovenian data. Their results support the conclusion of Helpman et al. (2004) that only the most productive firms invest abroad, but they find that the entry barriers for investment in high- and low-income countries were not significantly different, which again differs from the findings of Head and Ries (2003). For Chinese enterprises, the current research mostly focuses on the impact of enterprise productivity on export decisions. Feenstra et al. (2011), for example, find that firms’ exports are affected by the firms’ productivity and credit constraints. High-productivity firms have a greater probability of exporting and a larger volume of exports. Lu (2010) also finds that the costs of Chinese enterprises entering domestic and foreign markets depend on industry characteristics. For labor-intensive sectors, competition in the domestic market is much greater than that in foreign markets, so the cost of entering foreign markets is lower than the cost of selling domestically.

16

2 Firm Productivity and Outward Foreign Direct Investment …

Therefore, in labor- intensive sectors, high-productivity firms sell at home, while lowproductivity firms sell abroad. Lu et al. (2010) also find that apart from productivity, firm ownership is an important factor. Among foreign companies investing in China, those with high productivity are more likely to sell in China rather than export to other countries. There are currently few studies on OFDI by Chinese firms, probably due to lack of data, and most of them stay at the macro level (Huang & Wang, 2010). As far as we know, one paper, by Yu (2011), focuses on the micro level of Chinese firms’ OFDI. That study uses data on enterprises in Zhejiang province investing overseas and finds that Chinese enterprises “going out” does not reduce their exports. OFDI and exports are complementary. However, Yu and Xu do not examine the impact of the firms’ productivity on their foreign investment decisions and investment volume from the perspective of firm performance. This chapter aims to answer three questions. First, do enterprises with higher productivity have a higher probability of engaging in FDI? Second, do enterprises with higher productivity invest more overseas? Third, does the income level of the destination country have a significant impact on the enterprises’ investment? Compared with the cross-sectional data adopted by Helpman et al. (2004) and Head and Ries (2003), our study uses enterprise-level panel data from 2006 to 2008. One of the main reasons why the above two studies have been questioned as to their empirical methods is that they only use cross-sectional data, failing to solve the reverse causality problem. This chapter effectively avoids this shortcoming by using panel data and the appropriate instrumental variables to control endogeneity issues. The structure of the chapter is as follows: Sect. 2.2 provides a detailed description of the data used in the regression and how to measure firm productivity accurately; Sect. 2.3 discusses the main differences between enterprises with and without FDI; Sect. 2.4 discusses the influence of firm productivity on the OFDI decision; Sect. 2.5 studies the influence of firm productivity on the volume of OFDI; and Sect. 2.6 concludes.

2.2 Data and Measurement Two main data sets are used in this chapter. The first is the industrial firm’s database, which contains a wealth of firm-level variables on industrial firms with sales volume greater than 5 million yuan and above the average level, including hundreds of variables such as geographical location, industry, capital composition, staff composition, major operating items, profitability, export value, and so forth. The currently available years are 1998 to 2008. To match the years of the enterprise OFDI data set, the sample from 2006 to 2008 is used in this chapter. The second data set is the large manufacturing enterprise OFDI data set for Zhejiang province, which contains data for three years.3 The OFDI data set contains important indicators of Zhejiang 3

The data were provided by the Zhejiang Provincial Department of Foreign Cooperation.

2.2 Data and Measurement

17

Table 2.1 Amount and proportion of Zhejiang’s OFDI from 2006 to 2010 Year

Rank

Investment (10 thousand US$)

Proportion of total local OFDI (%)

2006

4

19,165

8.52

2007

2

45,898

10.22

2008

2

50,558

8.23

2009

5

78,207

8.36

2010

1

262,139

16.06

Source Cooperation Department, Ministry of Commerce

province’s OFDI, such as outflow Chinese cities and inflow foreign countries and cities, investment amount, industry, and so forth. These are indispensable variables in the following analysis. Zhejiang province plays a very important role in China’s OFDI. The eastern coastal cities have always been the main force of local investment, among which Zhejiang province is the most important. The number of enterprises investing overseas in Zhejiang province accounts for 21.4% of the country’s total, ranking first in China. Its investment volume is the highest in the local ranking every year. As shown in Table 2.1, Zhejiang province’s investment accounted for more than 16% of the total local OFDI in 2010, ranking first. Zhejiang’s OFDI started in 1982, about the same time as the national OFDI. As of June 2007, Zhejiang had 2,809 approved overseas enterprises, with a total investment of US$1.64 billion. The number of domestic entities and overseas institutions ranked first in the country. Zhejiang’s OFDI is very representative: about 70% of total investment was created by private enterprises, including machinery, textiles, electronics, and light industry, mainly in Asia, Europe, and North America, mostly in the form of overseas processing enterprises, resource development, overseas marketing networks, real estate development, and R&D institutions. The forms of OFDI are more and more diversified, from solely investing in enterprises to cross-border equity participation, mergers and acquisitions, and overseas listing, gradually changing to cluster-based scale development. Therefore, using the data on Zhejiang province first ensures the diversification of OFDI forms, enterprise entities, and sectors in the manufacturing industry. Second, the major investors in Zhejiang province are private enterprises, avoiding the historical and political factors in some countries in transition, that is, low-productivity state-owned enterprises that play an important role in OFDI today due to their historical status before the economic transition. A shortcoming of the data set is that some information provided by certain enterprises is inaccurate, making some of the samples potentially misleading.4 Similar to the studies by Xie Qianli et al. (2008) and Yu (2010, 2011), the following criteria were used to remove abnormal information: first, the samples with missing important financial indicators (such as total assets, net fixed assets, sales, and gross industrial 4

For example, some family businesses did not establish a formal accounting system, and their accounting statements.

18

2 Firm Productivity and Outward Foreign Direct Investment …

output) were removed; second, samples with fewer were often based on “yuan,” while the standard requirements were based on “thousand yuan” as the than 10 employees were excluded.5 As in the research by Cai and Liu (2009) and Feenstra et al. (2011), following the Generally Accepted Accounting Principles, this chapter also excludes enterprises with the following: (1) current assets exceeding fixed assets, (2) total fixed assets exceeding total assets, (3) net fixed assets exceeding total assets, (4) without identification numbers, and (5) an invalid time of establishment (for example, after December or before January). After a rigorous and reliable screening of the data, we then measure the productivity of the enterprises accordingly. The existing literature on firm productivity and firm behavior mostly adopts labor productivity. This measurement has certain limitations. Changes in the proportion of labor input will lead to changes in labor productivity. Therefore, for the same industry, capital-intensive enterprises have higher labor productivity than laborintensive enterprises. When Chinese enterprises seek markets overseas, if they use more labor, they naturally become low-productivity enterprises, and foreign enterprises show higher productivity. However, productivity in this sense does not necessarily affect the survival and development of enterprises. Therefore, this definition is somewhat misleading. A better measure is total factor productivity (TFP), which treats all inputs equally and measures the remaining technical and efficiency factors without affecting productivity due to changes in the proportion of inputs. Therefore, we use TFP as the measure of enterprise productivity. The standard method for calculating TFP is to use ordinary least squares (OLS) to calculate the Solow residual, but the traditional OLS method has two defects: reverse causality and selection bias. On the one hand, firms may choose both output and capital stock, or add a specific amount of investment to achieve a certain output, that is, the decision about capital stock is affected by output, rather than exogenous, thus causing the reverse causality problem. Therefore, although researchers initially adopted two-way firm and time fixed effects, they only reduced the co-occurrence bias without overcoming the reverse causality problem. On the other hand, sample selection bias often exists in panel data, that is, only high-productivity firms can stay in the sample, and lowproductivity firms are naturally excluded. This problem is even more prominent here, because one of the main data sets only contains firms above a certain scale, and firms with gradually shrinking scale are likely to exit. This problem cannot be solved using the traditional OLS method. For this reason, the early research usually eliminated enterprises that withdrew from the sample, to balance the panel, but there is no doubt that this wastes a large amount of information and fails to depict the dynamic behavior of the firms. The method of Olley and Pakes (1996), which has made a great contribution to this issue, is a very popular method for calculating TFP at present, and they solve the selection bias problem by setting up semiparametric equations. Based on Olley and Pakes (1996), Van Biesebroeck (2005) and Amiti and Konings (2007) introduce firms’ export and investment decisions into the calculation

5

Levinsohn and Petrin (2003) examine all Chilean enterprises with more than 10 workers in their study, and this chapter follows their standard.

2.2 Data and Measurement

19

Table 2.2 Statistical characteristics of key variables Variable

Mean

Sandard Deviation

Min

Max

Productivity

4.08

0.94

−4.13

8.59

OFDI (dummy)

0.0025

0.05

0

1

Export (dummy)

0.42

0.49

0

1

In (OFDI)

3.27

1.53

0

8.61

Export-domestic sales ratio

0.27

0.40

0

1

Capital

7.96

1.57

−0.29

15.99

Labor

4.45

0.98

0

10.16

Note Total number of observations: 100,999. Productivity, capital, and labor are reported in logarithms

and solve the problem of sample selection bias by establishing a probability model of enterprise survival. Similar to Yu (2010, 2011), we extend the calculation of Olley and Pakes productivity as follows. First, we use industrial-level deflated prices to measure TFP. For the measurement of the production function, Felipe et al. (2004) stress that monetary measurement should be used for the estimation error caused by the output, which is actually an estimation of the accounting identity.6 Second, to capture the importing and exporting behavior of enterprises in the calculation of TFP, we construct two dummies, indicating export and import, respectively. Therefore, compared with previous studies, we can further characterize the impact that enterprises’ foreign trade behavior may have on productivity. The Appendix provides the specific calculation method. In the following section, we refer to the above TFP calculated by the Olley and Pakes method as productivity. The compiled data include a sample of more than 30,000 large and medium-sized manufacturing enterprises in Zhejiang province from 2006 to 2008, with a total of 100,999 observations, among which more than 100 enterprises participate in OFDI on average each year, and 257 enterprises have OFDI behavior in total. Table 2.2 lists the statistical characteristics of some important variables used in the regression. Figure 2.2 shows the distribution of the logarithm of OFDI volume, which basically follows a normal distribution. Most of the enterprises’ investment volume is within US$10 million; over the three years, only a dozen enterprises invested more than this amount. To avoid the possible deviation of the regression results caused by the existence of outliers, we removed these observations.

6

To calculate TFP accurately, “real productivity” should theoretically be calculated according to the prices of specific products (Foster et al., 2008). However, as in many other studies, it is difficult to get the prices of all the products in each firm. Therefore, as a compromise, this chapter uses industry-level prices to deflate the output of enterprises.

20

2 Firm Productivity and Outward Foreign Direct Investment …

Fig. 2.2 Distribution of the logarithm of OFDI volume

2.3 Determinants Affecting Enterprises’ Entry into the FDI Market Before performing a rigorous econometric analysis, it is necessary to discuss whether there are significant differences between non-OFDI enterprises and OFDI enterprises in the key variables of labor, capital, and TFP. The theoretical model of Helpman et al. (2004) predict that firms with the highest productivity will directly invest overseas, while firms with low productivity will not. However, in transition countries such as China, the opposite phenomenon may occur; for example, some state-owned enterprises invested in foreign countries very early because of the policy monopoly. After the reform and opening up, the productivity of state-owned enterprises is relatively low (Wu, 2005), but they can still be active abroad by virtue of the existing overseas markets. Therefore, we may observe the phenomenon that low-productivity enterprises invest abroad. However, only 1% of the foreign investment enterprises are state-owned enterprises in our sample, so the historical factor of OFDI by state-owned enterprises will not change the whole story. To test whether high-productivity enterprises are more likely to invest abroad, we first divide the enterprises into two groups: with and without foreign investment. Table 2.3 presents the average productivity, capital, and labor force of these two groups of enterprises, and calculates the difference between them. As can be seen from Table 2.3, compared with the enterprises without OFDI, the enterprises with OFDI have higher productivity, higher capital scale, and a larger labor force, consistent with the theoretical expectation. Of course, it is not conclusive that high- productivity enterprises are more inclined to invest abroad, because it is possible for enterprises to invest first and gradually increase productivity. Therefore, we present the difference in enterprise productivity year by year in Table 2.4. If there is a learning effect for enterprises, the productivity advantage of

2.3 Determinants Affecting Enterprises’ Entry into the FDI Market

21

Table 2.3 Difference in productivity, labor force, and capital between non- OFDI enterprises and OFDI enterprises Productivity

Labor

Capital

Observations

Non-OFDI firms

92.85

155.41

12,379.01

100,742

OFDI firms

174.20

592.54

62,790.58

257

Difference

−81.34*** (−9.92)

−437.13*** (−19.62)

−50,411.57*** (−10.00)

***, **, * show significance at the 1, 5, and 1% level, respectively Table 2.4 Difference in productivity between non-OFDI enterprises and OFDI enterprises (year-on-year)

Productivity

2006

2007

2008

Non-OFDI firms

91.02

99.17

86.07

OFDI firms

270.09

170.31

70.51

Difference

−179.07*** (−12.10)

−71.14*** (−5.35)

15.57 (1.08)

Observations: non-OFDI

34,371

39,087

27,284

Observations: OFDI

77

113

67

***, **, * show significance at the 1, 5, and 1% level, respectively Table 2.5 Comparison of never-OFDI firms and first-time OFDI firms Productivity

Labor

Capital

Observations

Never OFDI firms

92.58

153.72

12,203.3

100,316

First-time OFDI firms

171.99

572.38

60,996.71

239

Difference

−79.41*** (−9.37)

−418.66*** (−18.46)

−48,793.41*** (−9.38)

***, **, * show significance at the 1, 5, and 1% level, respectively

an enterprise with OFDI relative to an enterprise without OFDI should increase over time. However, as can be seen from the table, the advantage does not increase over time but tends to decrease. Moreover, in 2008, there was no significant difference in productivity between the two groups of enterprises, indicating that the productivity advantage of an enterprise with OFDI is predetermined rather than accumulated through the investment. Table 2.5 reconfirms this inference. Table 2.5 reports the difference between the enterprises that never participated in OFDI and those participating in OFDI for the first time. Enterprises newly entering the OFDI market have higher productivity and more labor and capital. Table 2.6 illustrates the difference between the enterprises always participating in OFDI and the enterprises newly entering the OFDI market in the current year. Obviously, always-participants have no significant advantages from their experience.

22

2 Firm Productivity and Outward Foreign Direct Investment …

Table 2.6 Comparison of non-first-time OFDI firms and first-time OFDI firms Non-first-time OFDI firms

Productivity

Labor

Capital

Observations

203.33

860.05

86,609.17

18 239

First-time OFDI firms

171.99

572.38

60,996.71

Difference

31.33 (0.36)

287.66 (1.11)

25,612.46 (0.68)

Next, we examine whether OFDI volume is related to the income of the destination country. We divide the enterprises with OFDI into two groups based on the income levels of the destination countries: enterprises that invest in middle- and high-income countries and enterprises that invest in low-income countries. Specifically, according to the World Bank’s classification of national income levels in 2008, we define low-income countries as those with per capita GDP less than US$3,855 and highincome countries as those with per capita GDP above that.7 It may be argued that such classification has potential disadvantages, as the same enterprise may invest in multiple countries at the same time, which will result in overlapping classification. However, in our sample, this is not an issue, because only one enterprise in the sample invested in two countries in the same year, which is negligible. The results in Table 2.7 show that most of the enterprises invest in rich countries, mainly located in Europe, North America, Oceania, and Southeast Asia, and a few enterprises invest in countries in the Middle East and Latin America. Only a small number of enterprises invest in poor countries in Africa and other regions. There is no significant difference in productivity, scale, or capital between enterprises investing in rich and poor countries. There are significant differences between enterprises investing in poor countries and enterprises without OFDI in all aspects. These findings are in line with those of Helpman et al. (2004) and different from those of Head and Ries (2003). Of course, a possible explanation may be that poor and rich countries have different comparative advantages and resources in different industries. For example, poor countries tend to have advantages in labor and rich countries tend to have advantages in technology or resources, thus attracting enterprises from different industries. However, this simple comparison alone cannot eliminate the differences in industries; more rigorous econometric support is needed. If an enterprise belongs to an industry with relatively high labor intensity, it is possible that the enterprise faces higher entry costs in China due to the fiercer competition compared with foreign markets. Therefore, the enterprises that only serve domestic markets should have the highest productivity, which is higher than that of enterprises that export (Lu, 2010). Similarly, the productivity of domestic firms should be higher than that of firms investing overseas. To test this inference, following the method of Lu (2010), we classify manufacturing firms into 2-digit sectors, find out the average required capital-labor ratio for each sector, and then 7

According to the World Bank’s classification in 2008, per capita gross national product is less than US $975 for low-income countries, between US $975 and US $3,855 for lower-middle-income countries, between US $3,855 and US $11,906 for upper-middle-income countries, and greater than US $11,906 for high-income countries.

2.3 Determinants Affecting Enterprises’ Entry into the FDI Market

23

Table 2.7 Comparison of enterprises investing in low-income countries and high-income countries Productivity

Labor

Capital

Observations

Non-OFDI firms

92.84

155.40

12 379.01

100 742

OFDI low-income firms

132.58

402.94

45 481.18

35

Difference

−39.74* (−1.80)

−247.53*** (−4.14)

−33,102.17** (−2.43)

OFDI low-income firms

132.58

402.94

45,481.18

35

OFDI high-income firms

130.75

556.50

48,168.99

166

Difference

1.83 (0.06)

−153.56 (−0.81)

−2687.813 (−0.11)

***, **, * show significance at the 1, 5, and 1% level, respectively

(Distribution density)

calculate the difference in productivity between non-OFDI and OFDI enterprises for each sector. Figure 2.3 presents the distribution of the industry capital-labor ratio (by 4-digit code), showing that most firms are in labor-intensive sectors. According to Lu (2010), this difference should be positively correlated with the industry’s capitallabor ratio, but the results show that the correlation coefficient is −0.128, which is a weak negative correlation. However, since the number of foreign-investing enterprises in the sample is small, a 2-digit industry classification would result in too few enterprises investing in each industry, which would affect the effectiveness of the test. Therefore, we adopt a rough classification method, dividing the half of the enterprises with higher capital-labor ratios in their sectors into one group and those with lower capital-labor ratios into

(Industry capital-labor ratio)

Fig. 2.3 Distribution of the industry capital-labor ratio

24

2 Firm Productivity and Outward Foreign Direct Investment …

Table 2.8 Comparison of non-OFDI enterprises and OFDI enterprises in labor- intensive industries (grouped by median) Productivity

Labor

Capital

Observations

Non-OFDI firms

87.82

166.52

6882.529

50,296

OFDI firms

111.61

468.88

24,919.84

129

Difference

−23.78** (−2.36)

−302.35*** (−9.83)

−18,037.31*** (−7.38)

***, **, * show significance at the 1, 5, and 1% level, respectively Table 2.9 Comparison of non-OFDI enterprises and OFDI enterprises in capital-intensive industries (grouped by median) Productivity

Labor

Capital

Observations

Non-OFDI firms

97.84

144.32

17,859.15

55,402

OFDI firms

237.25

717.15

100,957.2

135

Difference

−139.41*** (−10.77)

−572.83*** (−17.79)

−83,098.03*** (−8.51)

***, **, * show significance at the 1, 5, and 1% level, respectively Table 2.10 Foreign investment and productivity of foreign invested enterprises in labor-intensive and capital-intensive industries (15% samples at each end)

Labor-intensive

Capital-intensive

Non-OFDI

91.54

115.01

OFDI

123.08

289.86

Difference

−31.53* (−1.74)

−174.84*** (−5.81)

Obs.: non-OFDI

14,386

15,181

Obs.: OFDI

40

36

***, **, * show significance at the 1, 5, and 1% level, respectively

another group, and calculating the difference in productivity and other indicators between non-OFDI enterprises and OFDI enterprises for each group. If the inference is correct, the differences in the group with low capital-labor ratios should be positive and those in the group with high capital-labor ratios should be negative. Table 2.8 reports that even in labor-intensive sectors, the productivity of OFDI enterprises is still significantly higher than that of non-OFDI enterprises (Table 2.9). However, OFDI in capital-intensive sectors has even greater advantages, possibly because the criterion of capital intensity is only arbitrarily selected based on the median industry, which will affect the test results. Therefore, we finally select the enterprises with the lowest and highest 15% of the capital-labor ratios and take these two extreme sectors as an example and retest. The results are presented in Table 2.10. In the labor- intensive enterprises, there is no significant difference in productivity between the OFDI enterprises and non-OFDI enterprises, but in the capital-intensive industries, the OFDI enterprises are indeed more efficient and larger than the non-OFDI enterprises.

2.4 Enterprise Productivity and the Enterprise FDI Decision

25

2.4 Enterprise Productivity and the Enterprise FDI Decision 2.4.1 Impact of the Firm’s Productivity on Its OFDI Decision It can be preliminarily inferred that there is a significant difference in productivity between enterprises with and without OFDI. More strictly, we can further investigate whether foreign-investing enterprises have higher productivity after including control variables in the regression. Consider the following regression equation: lnTFPit = β1 DitE X P + β2 DitO F D I + β3 ln klratioit + εit ,

(2.1)

where lnTFPit is the productivity of enterprise i in year t, DitOFDI and DitE X P indicate whether enterprise i engaged in OFDI and exporting in year t (yes is 1, no is 0), and ln klratioit indicates the capital-labor ratio of the enterprise. Of course, there are other variables that affect firm productivity. We absorb them into the error term and decompose them into the following three aspects: first, each enterprise’s own fixed effect ⏀i , which controls time-invariant factors; second, time fixed effect ηt , which controls factors that do not change with the enterprise; and third, the specific effect μijt , which is subject to the normal distribution, μi jt ∼ N (0, σi2j ), to control for other remaining factors. Table 2.11 reports the regression results. In the first column, we control the twoway fixed effect and find that enterprises with OFDI have higher productivity than enterprises without OFDI. Exporting firms also have higher productivity than nonexporting firms. In columns (2) to (4), we regress by years and the conclusions are similar. The only exception is that in 2008, neither firms with OFDI nor exporting firms had higher productivity, possibly due to the negative impact of the financial crisis (Feenstra et al., 2011). However, in general, our results are consistent with the findings of Helpman et al. (2004). Further, to capture the learning-by-doing effect of enterprises engaged in exporting or OFDI, and the self-selection effect caused by entry cost, we first assume that it is historically relevant for enterprises to choose to enter or not to enter the FDI market each year. To eliminate the learning effect of enterprises that have ever engaged in OFDI, we separate the enterprises that participated in OFDI for the first time in that year and those that never participated in OFDI and estimate their decision behavior based on the subsample, using a probit probability model. Pr(DOFDItt = 1|X it ) = β0 + β1 ln TFPit + β2 ln K it + β3 ln L it| + β4 DEXP + β5 DSOE + β6 DFIEit + εit .

(2.2)

This model also controls to some extent the unobservable heterogeneity that leads to differences in corporate historical behavior. As shown in the first column in Table

26

2 Firm Productivity and Outward Foreign Direct Investment …

Table 2.11 Impact of OFDI on firm productivity Productivity

(1) Full sample

(2) 2006

(3) 2007

(4) 2008

0.35*** (3.29)

0.02 (0.20)

OFDI dummy

0.38*** (6.53)

0.66*** (4.82)

Export dummy

0.11*** (21.90)

0.14*** (13.02) 0.16*** (15.76) −0.05*** (−4.58)

Capital intensity −0.08*** (−33.37) 0.00 (0.75)

−0.00 (−0.38)

−0.29*** (−63.88)

Year FE

Y

N

N

N

Firm FE

Y

Y

Y

Y

Obs

100 999

34 448

39 200

27 351

R2

0.01

0.01

0.01

0.15

Note In the regression, productivity, capital, labor, and capital intensity are presented in logarithmic form ***, **, * show significance at the 1, 5, and 1% level, respectively

2.12, the higher is the productivity of an enterprise that never participated in OFDI, the more likely it is to start investing abroad. To rule out the impact of capital scale on enterprise decision making, and the impact of the information and market advantages of exporting enterprises over non-exporting enterprises, these control variables are added in the second column, and the results are still significant. Therefore, adequate capital and larger scale are conducive to the creation of an FDI market. At the same time, exporting enterprises are more likely to start investing abroad, thanks to their knowledge and familiarity with foreign markets. However, the fluctuations between different years and the differences between different industries are not taken into consideration, and the entry thresholds of overseas markets of different industries are very different. Therefore, we added dummy variables for years and industries. Because the data used in the regression here include all enterprises, the industry classification using the 2-digit codes will not cause the problem of too few observations or too many variables. The results are reported in the third column and are similar to those of the previous regression. Based on the above analysis, to describe firms’ decision in greater detail, we add the type of owners to the control variables, including foreign-invested enterprises (including Hong Kong, Macao, and Taiwan enterprises), state-owned enterprises, and three other categories. Enterprises with different types of ownership have quite different OFDI decisions. The major OFDI investors are state-owned enterprises in China, accounting for 69%. In Zhejiang province, the main investors are private enterprises, which have weaker market power and fewer resources than state-owned enterprises but are more flexible in operation. The investment volume and number of Hong Kong, Macao, and Taiwan enterprises and foreign-invested enterprises are small. However, compared with state-owned enterprises and private enterprises, foreign-invested enterprises have better overseas resources and markets. Therefore, these types of enterprises are quite different, in terms of investment volume and the nature of the enterprises; thus, it is necessary to control their effects. The results after controlling ownership of the enterprise are presented in the fourth columns in Table 2.12. The findings show that the impact of productivity on the enterprise decision is

2.4 Enterprise Productivity and the Enterprise FDI Decision

27

Table 2.12 Impact of firm productivity on firms’ OFDI decision

Productivity

(1) Probit

(2) Probit

(3) Probit + FE

(4) Probit

0.13*** (5.02)

0.05** (2.17)

0.09*** (2.93)

0.09*** (2.86)

Lagged productivity

(5) Probit + FE

0.16*** (5.07)

Capital

0.09*** (4.77)

0.09*** (3.87)

0.09*** (3.93)

0.08*** (3.17)

Labor

0.12*** (4.38)

0.13*** (4.21)

0.13*** (4.25)

0.13*** (3.90)

Export dummy

0.68*** (10.27)

0.68*** (10.38)

0.69*** (10.53)

0.72*** (9.88)

Foreign-invested dummy

−0.09* (−1.66)

SOE dummy

−0.34 (−0.89)

Obs

100,555 R2

100,555

100,450

100,450

84,537

0.0104

0.127

0.133

0.134

0.142

Year FE

N

N

Y

Y

Y

Industry FE

N

N

Y

Y

Y

pseudo

***, **, * show significance at the 1, 5, and 1% level, respectively

still significantly positive, and the impact of the other major control variables does not change. The ownership structure of the enterprise has no additional impact on the enterprise’s decision to invest abroad.

2.4.2 Endogeneity Between OFDI and Firm Productivity The reverse causality problem still exists because the learning-by-doing effect is not excluded. In addition, even if the learning-by-doing effect is controlled, the endogeneity problem still exists: the above results may be caused by unobservable missing variables that affect both firm productivity and entry cost. For example, the spillover effect brought by the increase of foreign-owned enterprises spreads new foreign technologies, improves firm productivity, and expands the overseas channels of enterprises, reducing the cost of foreign investment. Therefore, it is insufficient to solve the endogenous problems by selecting samples and controlling fixed effects, and the time span of three years fails to control unobservable firm heterogeneity. Without loss of generality, if the error term does not have too strong serial correlation but is determined by the disturbance of the current period, then the current period productivity can be replaced by the productivity lagged one period (Wooldridge,

28

2 Firm Productivity and Outward Foreign Direct Investment …

2002),8 that is Pr(DOFDIit =1|X it ) =β0 + β1 ln TFPit−1 + β2 ln K it + |β3 ln L it + β4 DEXPit + β5 DSOEit + β6 DFIEit + εit .

(2.3)

Assume lnTFPit = lnTFPit −1 + εit , E(lnTFPit −1 εit ) = 0, thus ruling out the impact of the current period’s OFDI on firm productivity. The last column in Table 2.12 reports that the regression results are robust and significant. Productivity still has a significant positive effect on the firm’s OFDI decision, and the magnitude is similar to the previous results. The impacts of capital stock, firm size, and the export dummy are still significantly positive. Of course, it is necessary to test for serial correlation of the error term. First, we found that there is a strong positive correlation between TFP lagged one period and TFP in the current period, indicating that this substitution is reasonable. Second, to test whether TFP lagged one period is exogenous, we first estimate the residual term from the regression results, and then regress it on TFP lagged one period. The results show that the residual has no impact on firm productivity in the previous period, and the correlation coefficient is very low, only 0.004. Therefore, it can be ruled out that some time-persistent factors affect both the current period’s firm decision and the previous period’s firm productivity.

2.4.3 Relationship Between Firm Productivity and Income Level of the Destination Country The Head and Ries (2003) model finds that countries with low income levels have lower entry cost, thus attracting low-productivity firms to invest in them. Therefore, the average productivity of domestic firms investing in low-income countries is lower than that of domestic firms not investing abroad. To test the relationship between income level of the destination country and attractiveness to firms, we perform two regressions by destination. In the first regression, we use a sample consisting of firms that first invested in low-income countries and firms that never invested abroad, to estimate the probability of investing in low-income countries. In the second regression, we use firms that invested in high-income countries and enterprises that did not invest, to obtain the corresponding estimates. We then test the on firms’ differences between the two regression coefficients and compare the impacts of productivity investment in different countries. Pr (D OP OF DOIRit =1|X it ) 8

Another useful alternative is two-stage regression with instrumental variables, which we use to deal with endogeneity in the next section.

2.4 Enterprise Productivity and the Enterprise FDI Decision

29

=β0 + β1ln T F Pit + β2 ln K it + β3ln L it + β4 D E X Pit + β5 D S O Eit + β6 D F I Eit + εit , Pr(D OR IFCDHIit

=1|X it ) =β0 + β1 ln T F Pit + β2 ln K it + β3 ln L it + β4 D E X Pit + β5 D S O Eit + β6 D F I Eit + θit .

(2.4)

For simplicity, we assume that εit and θit are independent; then, we can estimate the two probit equations. Table 2.13 reports the results, which show that firm productivity has no significant impact on firms’ OFDI decision in the two regressions, possibly due to the limited numbers of observations in the samples after separation. However, the results of the joint T-test based on the two regressions show that the coefficients of the two TFPs are the same at the 10% level, which indicates that TFP has no effect on selecting destination countries of different levels of income. Of course, there is still room for improvement in this benchmark estimation method. This is because if we separate firms’ investment decisions into two stages– invest in poor countries or not, invest in rich countries or not–the two stages are clearly not independent. That is, in the above regressions, εit and θit are related. Although an enterprise can invest in both low-income countries and high-income countries at the same time, the total amount of resources is fixed; thus, the two decisions affect each other. Few firms invest in both kinds of countries at the same time. Therefore, Table 2.13 Income levels of destination countries and investment choices of firms Dependent var OFDI dummy

(1)

(2)

By income level

(3)

(4)

(5)

By income level + joint

poor/rich

Sample

First-time OFDI in poor cty. and non-OFDI

First-time OFDI in rich cty. and non-OFDI

First-time OFDI in poor cty. and non-OFDI

First-time OFDI in rich cty. And non-OFDI

First-time OFDI

Productivity

0.03 (0.59)

0.02 (0.86)

0.04 (0.87)

0.02 (0.71)

−0.13 (−1.22)

Capital

0.08** (2.12)

0.07*** (2.99)

0.08** (2.20)

0.07*** (3.00)

−0.04 (−0.48)

Labor

0.04 (0.75)

0. 15*** (4.35)

0.04 (0.78)

0.14*** (4.12)

0.26* (1.75)

Export dummy

0.28*** (2.67) 0.75*** (8.41)

0.27*** (2.61) 0.75*** (8.41)

0.76** (2.44)

Obs

100,555

100,555

100,555

100,555

188

Industry FE

Y

Y

Y

Y

Y

Year FE

Y

Y

Y

Y

Y

Bivariate Probit N model

N

Y

Y

N

Pseudo R2

0.124

–

–

0.115

0.0621

***, **, * show significance at the 1, 5, and 1% level, respectively

30

2 Firm Productivity and Outward Foreign Direct Investment …

based on the above econometric model, we further use the bivariate probit model for estimation. The subsamples and regression equations selected for the above two regressions are unchanged, but assuming that their error terms are correlated with each other, the correlation properties of the error terms can be used to enhance the effectiveness of the estimation. The regression results are reported in the third and fourth columns in Table 2.13. The differences between the two regression coefficients are tested simultaneously, and the P-value is about 0.8, meaning that firm productivity does not affect investment in different destination countries. For a direct comparison of the attractiveness of countries with different income levels, we use a subsample of firms that invested abroad for the first time in that year to estimate the probability of choosing a high-income destination country among all countries. The results are shown in the fifth column in Table 2.13. It is obvious that firm productivity, capital, and size have no significant effect on whether the firm chooses a high-income country or a low-income country. The choice of destination country has nothing to do with firm productivity, which indicates that there is no significant difference in entry cost between countries with different levels of income. Or even if there is a difference in entry costs, it is not the most important factor when a firm decides where to invest.9 As before, the endogeneity problem cannot be ruled out in the above regression. Therefore, we repeat the regression in Table 2.13 by replacing the current productivity with the productivity lagged one period, and the results are presented in Table 2.14.10 From the results in the first and second columns, after dealing with endogeneity, the productivity increase of domestic firms has significantly increased the possibility of OFDI, no matter in poor countries or rich countries. Moreover, contrary to Head and Ries’s (2003) predictions, the productivity of enterprises investing in poor countries is higher. In the third column of the regression with first-time OFDI enterprises, the results are consistent with those in Table 2.13. Whether enterprises invest in poor countries or rich countries is not significantly affected by firm productivity, which once again proves that the Head and Ries (2003) proposition is invalid for Chinese enterprises.

2.4.4 Enterprise OFDI and Industry Labor Intensity Next, we test whether there is a reverse phenomenon of low-productivity enterprises engaging in OFDI and high-productivity enterprises not engaging in OFDI in laborintensive industries, with the results reported in Table 2.15. We first classify the samples into labor-intensive and capital-intensive, and then perform OLS and twoway fixed effect regression, respectively. Here, our classification of capital intensive 9

A multinominal probit model is also used to evaluate robustness. The p-value close to 0.8 is obtained to reject the assumption that the two groups have the same coefficients. The results are similar to those in Table 2.9 and are not listed separately to save space. 10 We thank the anonymous reviewers for their constructive comments on this section.

2.4 Enterprise Productivity and the Enterprise FDI Decision

31

Table 2.14 Income level of destination countries and investment choices of firms (controlling endogeneity) Dependent var OFDI dummy

(1)

(2)

Sample

First-time OFDI in poor cty. And non-OFDI

(3)

By income level

poor/rich First-time OFDI in rich cty. and non-OFDI

First-time OFDI

Lagged productivity

0.10* (1.83)

0.06** (2.14)

−0.17 (−1.37)

Capital

0.05 (1.17)

0. 07*** (3.03)

−0.00 (−0.00)

Labor

0.07 (0.98)

0.13*** (3.73)

0.13 (0.92)

Export dummy

0.32** (2.48)

0.80*** (8.25)

0.93*** (2.62)

Obs

84,423

84,548

179

Industry FE

Y

Y

Y

Year FE

Y

Y

Y

R2

0.0559

0.125

0.0579

***, **, * show significance at the 1, 5, and 1% level, respectively

and labor intensive is consistent with Sect. 2.3. The median of the average industrycapital intensity is used as the cutoff point. Industries with higher capital intensity than the median are classified as capital-intensive industries, and industries with lower capital intensity than the median are classified as labor-intensive industries. Although this classification is somewhat subjective, it is still very representative. The regression results for each subsample strongly support that firm productivity has a significant positive effect on firm OFDI decisions. Through these three groups of analysis, we can draw the conclusion that even in labor-intensive sectors, the firms that invest abroad have higher productivity than the firms that do not invest abroad, indicating that the main driver that affects the investment decision is not the strength of market competition. Table 2.15 Impact of firm productivity on OFDI, by the labor and capital intensity of the industry (1) Labor-intensive

(2) Labor-intensive

(3) Capital-intensive

(4) Capital-intensive

Productivity

0.06* (1.93)

0.10** (1.99)

0.04 (1.21)

0.07** (2.03)

Capital

0.11*** (3. 57)

0.12*** (3.62)

0.06** (2.05)

0.06* (1.85)

Labor

0.08* (1.95)

0.08* (1.77)

0.18*** (4.24)

0.18*** (4.34)

Export dummy

0.73*** (6.50)

0.75*** (6.72)

0.65*** (7.89)

0.64*** (7.77)

Obs

50,184

49,381

50,371

50,027

Year FE

N

Y

N

Y

Industry FE

N

Y

N

Y

Capital-labor ratio

L

L

H

H

0.110

0.116

0.145

0.157

Pseudo

R2

***, **, * show significance at the 1, 5, and 1% level, respectively

32

2 Firm Productivity and Outward Foreign Direct Investment …

Table 2.16 Impact of firm productivity on OFDI, by the capital intensity of the industry (controlling endogeneity) (1) (2) (3) (4) Labor-intensive Labor-intensive Capital-intensive Capital-intensive Lagged productivity 0.07** (2.10)

0.10** (2.32)

0.14*** (3.64)

0.21*** (4.79)

Capital

0.11*** (3.10)

0.11*** (3.13)

0.05* (1.67)

0.04 (1.16)

Labor

0.09** (2.06)

0.09* (1.89)

0.17*** (3.76)

0.18*** (3.88)

Export dummy

0.71*** (6.19)

0.72*** (6.31)

0.74 (7.69)

0.74*** (7.60)

Obs

41,782

40,093

42,842

42,563

Year FE

N

Y

N

Y

Industry FE

N

Y

N

Y

R2

0.104

0.110

0.166

0.181

Capital intensity

L

L

H

H

***, **, * show significance at the 1, 5, and 1% level, respectively

Similar to the previous analysis, to control the endogeneity problem, we replace the current productivity with firm productivity lagged one period and perform the regression in Table 2.15 again, and the results are reported in Table 2.16.11 The results show that firm productivity is a significant determinant of whether a firm engages in OFDI, no matter in labor-intensive industries or capital-intensive industries. Even in labor-intensive industries, high-productivity firms are more likely to invest abroad, contrary to Lu (2010).

2.5 Impact of Firm Productivity on the Volume of OFDI After analyzing whether the firm decides to invest abroad or not and, if so, to which country, the last question to be studied in this chapter is whether firm productivity affects investment volume if the firm invests abroad. Figure 2.4 shows the relationship between the level of OFDI and firm productivity, and it is obvious that there is a positive relationship between them.

2.5.1 Benchmark Regression We consider the following regression: ln O F D Iit =β0 + β1 ln T F Pit + β2 ln k lrati Oit + β3 D S O Eit + β4 D F I Eit + X it + γi + μt + εit 11

We thank the anonymous reviewers for their constructive comments on this section.

(2.5)

2.5 Impact of Firm Productivity on the Volume of OFDI

33

Fig. 2.4 Relationship between firm OFDI volume and firm productivity

where DSOEit and DFIEit , respectively, represent whether enterprise i is a stateowned enterprise or a foreign-invested enterprise in year t. The other variables are as mentioned above. The first column in Table 2.17 includes only the productivity of the enterprise. The coefficient is significantly positive, indicating that high-productivity enterprises have both the willingness and the ability to invest more.

2.5.2 Endogeneity Analysis The above analysis can only explain the correlation between productivity and investment volume. To identify the causal effect, it is necessary to control the endogeneity, that is, to ensure that the enterprise does not learn from experience and gain technology from investment, thus improving productivity. It is also necessary to ensure that there are no missing variables that affect the productivity and investment volume of the enterprise. Therefore, we use the instrumental variable method to control the endogeneity that firm productivity may cause. Specifically, we use last year’s R&D investment as the instrumental variable. This not only ensures a significant correlation between the instrumental variable and firm productivity, but also ensures the exogeneity between the instrumental variable and the firm’s OFDI. Intuitively, on the one hand, an enterprise’s R&D can obviously promote the technological progress of the enterprise; on the other hand, in our sample, the enterprise’s R&D has no direct impact on OFDI: the main investment purpose of most of China’s foreign-investing enterprises is to explore markets rather than to learn the technology. From the perspective of the categories of foreign-investing enterprises, only 5 of 257 in our sample set up R&D institutions overseas, and the vast majority of enterprises’ foreign investment takes the form of establishing trading

34

2 Firm Productivity and Outward Foreign Direct Investment …

Table 2.17 Impact of firm productivity on OFDI volume Dependent var Firm OFDI level

(1) OLS

(2) OLS

(3) OLS

(4) 2SLS

Productivity

0.20** (2.12)

0.19** (2.05) 0.27** (2.65) 1.68** (1.98)

(5) 2SLS 1.43** (2.17)

Capital intensity

0.15 (1.32)

0.12 (0.94)

0.14 (1.25)

SOE

3.57*** (109.53)

3.88* (1.72)

3.98** (2.04)

0.15 (0.82)

0.22 (0.66)

0.09 (0.31)

−0.39** (−1.62)

−0.31** (−2.32)

0.02 (0.04)

−0.14 (−0.37)

N

N

Foreign-invested Proportion of export in total sales IV

Y

Y

Kleibergen-Paap Wald rk F statistic

N

5.459

7.469

Kleibergen-Paap rk LM statistic

5.471

7.486

Year FE

N

N

Y

N

Y

Industry FE

N

N

Y

N

Y

R2

0.02

0.03

0. 09

0.62

0.73

Obs

256

256

256

246

246

***, **, * show significance at the 1, 5, and 1% level, respectively

companies or offices. Therefore, it can be considered that the R&D activities of the enterprises in our data set affect enterprises’ foreign investment through productivity. To confirm the validity of this instrumental variable, several necessary tests are carried out. First, we test whether the variable is correlated with an endogenous regression factor (current firm productivity). The test results are reported in columns (4) and (5) in Table 2.17. We assume that the error term is heteroscedastic: εit ~ N(0, σ2). Therefore, the usual Anderson and Gerbing (1984) likelihood ratio test is no longer applicable as it only applies to the hypothesis that the error term is independent and identically distributed. We use the Wald statistic proposed by Kleibergen and Paap (2006) to test whether this instrumental variable is related to endogenous regression factors. The rejection domain of the null hypothesis is determined by the significance level of 1%. Second, we test whether there is a weak correlation between the instrumental variable and the firm’s current productivity. If this is the case, the instrumental variable approach is invalid. The null hypothesis is that the first stage is weakly identified, rejected by the Kleibergen and Paap (2006) F statistic at a high level of significance.12 In addition, the adoption of R&D investment in the previous period (rather than the current period) can avoid the endogeneity between current productivity and current R&D. In short, these statistical tests provide solid evidence that the instrumental variable is valid. 12

The Cragg and Donald (1993) F statistic only applies to the independent identically distributed hypothesis and is therefore invalid here.

2.5 Impact of Firm Productivity on the Volume of OFDI

35

Table 2.18 Impact of industry capital intensity on firm OFDI volume (sample of developed countries) Dependent var Firm OFDI level (1) OLS

(2) OLS

(3) 2SLS

(4) 2SLS

Productivity

0.19* (1.85) 0.27** (2.48) 1.74** (2.10)

1.74** (2.34)

Firm capital intensity

0.13 (1.28)

0.14 (1.54)

0.25* (1.68)

Industry capital intensity

0.00 (1.45)

0.00 (1.14)

0.18 (1.21)

−0.00 (−0.35) −0.00 (−0.80) 3.83* (1.66)

SOE

4.13* (1.91)

Obs

221

221

215

215

Year FE

N

Y

N

Y

IV

N

N

Y

Y

R2

0.06

0.10

0.61

0.66

***, **, * show significance at the 1, 5, and 1% level, respectively

The fourth and fifth columns in Table 2.17 report the second-stage results of the two- stage least squares regression. Higher firm productivity leads to higher OFDI. Specifically, from 2006 to 2008, for every 10% increase in productivity of foreign-investing enterprises, their OFDI will increase by about 14%. In addition, after controlling endogeneity, there is no significant relationship between the share of exports in total sales volume and the firms’ OFDI, meaning that whether the enterprise is aimed at the domestic market or the foreign market will not affect its OFDI. Finally, we examine whether the investment volume is affected by the industry capital intensity. We control the average capital-labor ratio (capital intensity) of both the firm itself and its industry. If in labor-intensive industries the competition in China is fiercer than in developed countries, but in capital-intensive industries, the competition in China is less, then from the sample investing in developed countries, the higher is the industry capital-intensity, the smaller is the investment volume. Therefore, we use the data on enterprises investing in developed countries for the regression. However, as shown in Table 2.18, the capital intensity of the industry does not affect the investment volume. Therefore, the capital intensity of the industry is not the main factor that affects the enterprise’s survival, nor does it determine the development of the enterprise.

2.5.3 Additional Robustness Tests: Analysis Based on the Gravity Equation13 The gravity equation is a classical model for analyzing bilateral trade at the national and industry levels. The gravity equation says that the geographical distance and economic level of a country are important determinants of bilateral trade. To test 13

We thank the anonymous reviewers for their helpful suggestions on this section.

36

2 Firm Productivity and Outward Foreign Direct Investment …

the robustness of our previous main conclusions under the framework of the classical gravity equation, we introduce the main variables of the gravity equation into Eq. (2.5). Although there is no mature theoretical model to explain the OFDI version of the gravity equation at the enterprise level, we think that geographical distance, GDP of the investment destination country, and so forth all have important impacts on OFDI. Therefore, in the following regression, we use the gravity equation model to test the robustness of the above conclusion. Among the variables, we use the time it takes to sign a contract in the destination country and the cost of obtaining a license to measure the local investment environment. The data were derived from the World Bank’s Doing Business data set. The Doing Business data include other measures, such as the cost of starting a business and the procedures for signing a contract. The results obtained with different measures are basically the same. Table 2.19 reports the regression results, which are consistent with the previous results. Firm productivity is a significantly positive determinant of OFDI. China’s GDP has a significantly positive impact on OFDI, while the income levels of destination countries and the capital intensity of industries have no steady significant impact on OFDI. Different from the classical gravity equation of exports, the influence of geographical distance on OFDI is uncertain. This is because on the one hand, geographical distance directly affects the difficulty of OFDI and on the other hand, it affects the cost of exports to a greater extent. Because exporting and OFDI are alternatives according to the previous analysis, the influence of geographical distance in the regression becomes insignificant.

2.6 Summary Based on data on manufacturing enterprises’ OFDI in Zhejiang province, this chapter investigated several main conclusions in the theoretical literature about the impact of firm productivity on enterprises’ OFDI. We found that, first, productivity has a significantly positive impact on firm OFDI motivation and volume. The higher is the productivity, the greater is the probability of OFDI. Moreover, the higher is the productivity, the higher is the investment volume. Second, generally speaking, it is not less costly to enter low-income countries compared with high-income countries. The level of income in the destination country has no significant impact on firms’ OFDI decision, nor on firms’ investment volume. Third, the capital intensity of the industry has no significant impact on the location of the enterprise. That is, in laborintensive industries, the survival cost of the enterprise abroad is no less than that of the enterprise at home, and the productivity threshold of the enterprise investing abroad is very high. Therefore, there is no reverse phenomenon that high-productivity enterprises in labor- intensive sectors do not invest abroad and low-productivity enterprises invest abroad. The capital-labor intensity of the industry does not have a significant impact on the firm’s OFDI decision or OFDI volume. There has been little research on the micro-level OFDI of Chinese enterprises, probably because of data issues. The data set we use is not perfect in that the sample

2.6 Summary

37

Table 2.19 Impact of firm productivity on OFDI volume (gravity equation model) Dependent var Firm OFDI level

(1) FE

(2) 2SLS

(3) FE

(4) 2SLS

Productivity

0.21** (2.58)

1.53** (2.00)

0.23*** (3.37)

1.44** (2.17)

SOE

2.92*** (12.78)

3.15 (1.54)

2.50*** (8.04)

3.19 (1.63)

Foreign-invested

0.35 (1.24)

0.27 (0.83)

0.23 (0.77)

0.15 (0.50)

Capital intensity

0.11 (1.06)

0.27* (1.91)

0.05 (0.64)

0.23* (1.70)

Proportion of export in total sales

−0.36* (−1.74)

−0.04 (−0.08)

−0.57*** (−8.74)

−0.16 (−0.35)

Geographical distance

−0.002 (−0.02)

0.32 (0.94)

−0.09 (−0.80)

0.24 (0.83)

4.11** (2.30)

China GDP Destination GDP

0.05 (0.63)

0.09 (0.76)

Time to sign contract

−0.57*** (−3.16)

−0.39 (−0.69)

Cost of obtaining a license IV

N

N

4.27*** (2.74) 0.02 (0.33)

0.10 (0.89)

−0.20** (−2.50)

0.06 (0.30)

N

Y

Kleibergen-Paap Wald rk F statistic

6.303

8.045

Kleibergen-Paap rk LM statistic

6.422

8.124

Year FE

N

N

Y

N

Industry FE

N

N

Y

N

Obs

207

200

208

200

Note Productivity, capital intensity, China’s GDP, destination country’s GDP, time needed to sign a contract, and cost of obtaining a license are present in logarithmic form ***, **, * show significance at the 1, 5, and 1% level, respectively

size is relatively small and the time span is relatively short. However, to a certain extent, this study makes up for the academic gap in micro-level research on Chinese enterprises’ “going out” to engage in OFDI. To overcome the defects of the data, we conducted a detailed empirical analysis and used various methods to test the robustness of the results. The analysis effectively controlled the endogenous problems, such as reverse causality and missing variables, and the potential sample selection bias problem of enterprises entering and exiting the FDI market, thus ensuring the reliability of the results. Of course, if we can obtain more detailed data on enterprises’ foreign investment in the future, we would be able to obtain more reliable and rigorous results, and we would analyze additional problems in foreign investment, such as the flow of overseas enterprises and the dynamic behavior of enterprise investment.

38

2 Firm Productivity and Outward Foreign Direct Investment …

Appendix Estimation of TFP Using the Olley and Pakes (1996) Method Olley and Pakes (1996) considered a conventional Cobb–Douglas production function βm

βk

βl

Yit = πit Mit K it L it , βm

(A1) βk

βl

where Y it is the output of enterprise i in time t; Mit , kit , and L it are capital, labor force, and intermediate inputs, respectively; and πit is firm productivity. Assuming that the realized value of the previous period’s productivity determines the enterprise’s expectation of the latter period’s productivity (vit), then the investment of enterprise i can be written as a monotonically increasing function of the logarithm of productivity and capital stock, k it = lnK it . Following the development of the Olley and Pakes method by Amiti and Konings (2007), we also added the export decision of the enterprise into its investment function. Iit = I˜(ln K it , vit , EFit ),

(A2)

where EFit is a dummy variable that indicates whether an enterprise i exports in time t. Therefore, the inverse function of Iit can be written as vit = I˜−1 (ln K it , Iit , EFit ).

(A3)

Therefore, the unobservable firm productivity depends on the firm’s capital stock, investment, and export status. The estimation equation for total factor productivity (TFP) can now be written as ln Yit = β0 + βm ln Mit + βl ln L it + g(ln K it , Iit , EFit ) + εit ,

(A4)

where g(ln K it , Iit , EFit ) ≡ βk ln K it + I˜−1 (ln K it , Iit , EFit ). Referring to the methods of Olley and Pakes (1996) and Amiti and Konings (2007), the fourth-order polynomials of the capital (logarithm), investment (logarithm), and export dummy variables are used to approximate g(·).14 g(kit , Iit , EFit ) = (1 + EFit )

4  4  h=0 q=0

q

δhq kith Iit .

(A5)

After βˆm and βˆl are estimated, the residual of Eq. (2.4) is calculated as Rit ≡ ln Yit − βˆm ln Mit − βˆl ln L it . 14

Approximating g(·) with higher order polynomials does not substantially change the estimates.

References

39

Next, we estimate the coefficient of capital stock,βk . To correct the aforementioned sample selection bias, Amiti and Konings (2007) proposed to estimate a probability model of enterprise survival (the independent variables are higher-order polynomials of capital and investment) and put the estimated survival probability into the equation as a control. Therefore, we estimate the following equation:   Rit = βk ln K it + I˜−1 git−1 − βk ln K i,t−1 , pr ˆ i,t−1 + εit ,

(A6)

where pr ˆ i,t−1 represents an estimate of the enterprise’s exit probability in the  next  year. Since we do not know the exact functional form of the inverse function I˜−1 of the investment function, we approximate it with the fourth-order polyphase form of git−1 and ln K i,t−1 . In addition, formula (A6) requires that the coefficients in front of the two capital items in the formula are the same. Therefore, we use the nonlinear least squares method for the estimation (Arnold, 2005; Pavcnik, 2002). Finally, after obtaining the capital coefficient βˆk , we calculate the TFP of each enterprise i in each industry j. T F PiOjt P = ln Yit − βˆm ln Mit − βˆk ln K it − βˆl ln L it .

References Amiti, M., & Konings, J. (2007). Trade liberalization, intermediate inputs, and productivity: Evidence from Indonesia. American Economic Review, 97(5), 1611–1638. Anderson, J. C., & Gerbing, D. W. (1984). The effect of sampling error on convergence, improper solutions, and goodness-of-fit indices for maximum likelihood confirmatory factor analysis. Psychometrika, 49(2), 155–173. Arnold, J. M. (2005). Productivity estimation at the plant level: A practical guide. Unpublished manuscript, 27. Cai, H., & Liu, Q. (2009). Competition and corporate tax avoidance: Evidence from Chinese industrial firms. The Economic Journal, 119(537), 764–795. Cragg, J. G., & Donald, S. G. (1993). Testing identifiability and specification in instrumental variable models. Econometric Theory, 9(2), 222–240. Damijan, J. P., Polanec, S., & Prašnikar, J. (2007). Outward FDI and productivity: Micro-evidence from Slovenia. World Economy, 30(1), 135–155. Eaton, J., Kortum, S., & Kramarz, F. (2004). Dissecting trade: Firms, industries, and export destinations. American Economic Review, 94(2), 150–154. Foster, L., Haltiwanger, J., & Syverson, C. (2008). Reallocation, firm turnover, and efficiency: Selection on productivity or profitability? American Economic Review, 98(1), 394–425. Head, K., & Ries, J. (2003). Heterogeneity and the FDI versus export decision of Japanese manufacturers. Journal of the Japanese and International Economies, 17(4), 448–467. Helpman, E., Melitz, M. J., & Yeaple, S. R. (2004). Export versus FDI with heterogeneous firms. American Economic Review, 94(1), 300–316. Huang, Y., & Wang, B. (2010). An analysis of the pattern and causes of Chinese investment abroad”, miemo. Peking University.

40

2 Firm Productivity and Outward Foreign Direct Investment …

Kleibergen, F., & Paap, R. (2006). Generalized reduced rank tests using the singular value decomposition. Journal of Econometrics, 133(1), 97–126. Levinsohn, J., & Petrin, A. (2003). Estimating production functions using inputs to control for unobservables. The Review of Economic Studies, 70(2), 317–341. Lu, D. (2010). Exceptional exporter performance? University of Chicago. Lu, J., Lu, Y., & Tao, Z. (2010). Exporting behavior of foreign affiliates: Theory and evidence. Journal of International Economics, 81(2), 197–205. Melitz, M. J. (2003). The impact of trade on intra-industry reallocations and aggregate industry productivity. Econometrica, 71(6), 1695–1725. Montagna, C. (2001). Efficiency gaps, love of variety and international trade. Economica, 68(269), 27–44. Pavcnik, N. (2002). Trade liberalization, exit, and productivity improvements: Evidence from Chilean plants. The Review of Economic Studies, 69(1), 245–276. Poncet, S. (2007). Inward and outward FDI in China, Panthéon-Sorbonne-Economie, Université Paris 1 CNRS and CEPII. working paper. Van Biesebroeck, J. (2005). Exporting raises productivity in sub-Saharan African manufacturing firms. Journal of International Economics, 67(2), 373–391. Wooldridge, J. M. (2002). Econometric analysis of cross section and panel data MIT press. Cambridge, MA, 108(2), 245–254. Wu, J. (2005). Understanding and interpreting Chinese economic reform. Texere. Yu, M. (2010). Trade liberalization and productivity: Evidence from Chinese Firms. Economic Research Journal, 7(3), 809–826. Yu, M. (2011). Processing trade, firm productivity, and tariff reductions: Evidence from Chinese products. China Trade Research Group CTRG Working Paper Series, 14(4), 1151–1181.

Chapter 3

Distribution, Outward FDI, and Productivity Heterogeneity: China and Cross-Countries’ Evidence

Abstract This chapter examines distribution-oriented outward FDI using Chinese multinational firm–level data. Distribution outward FDI refers to Chinese parent firms in manufacturing that penetrate foreign markets through wholesale trade affiliates that resell exportable goods. Our estimations correct for rare-events bias and show that distribution FDI are more productive than non-FDI firms but less productive than non-distribution FDI firms. As cross-border communications costs (transportation costs) increase, there is a higher the probability that firms engage in distribution FDI (non-distribution FDI). Our endogenous income-threshold estimates show that highproductivity Chinese firms invest more in high-income countries, but not necessarily in low-income countries.

3.1 Introduction Distribution-oriented outward foreign direct investment (FDI) refers to the phenomenon of home parent manufacturing firms that penetrate foreign markets through wholesale trade affiliates that resell exportable goods. Distribution-oriented outward FDI is an important phenomenon in developed countries like the United States (Hanson et al., 2001), and in developing countries like China. However, there is relatively scant research on this topic, the present chapter aims to fill this gap. Outward FDI includes two main categories: distribution-oriented and nondistribution production-oriented FDI. Distribution FDI includes the business-service foreign affiliates and the wholesale foreign affiliates. The business-service FDI mainly refers to building overseas business office to explore foreign market, to promote sales, and to serve customers in the hosting countries. Similarly, the wholesale FDI refer to oversea intermediaries of parent firms to help exporting and sales in the host Countries. The wholesale foreign affiliates in the United States accounted for over 20% of total foreign sales by multinationals even in a decade ago. The number of wholesale foreign affiliates is around 50% of that of production foreign affiliates. In a developing This chapter is published in Journal of International Financial Markets, Institutions, & Money by Tian Wei and Miaojie Yu. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022 W. Tian and M. Yu, Outward Foreign Direct Investment of Chinese Enterprises, Contributions to Economics, https://doi.org/10.1007/978-981-19-4719-3_3

41

42

3 Distribution, Outward FDI, and Productivity Heterogeneity …

country like China, the proportion of distribution FDI is even higher. In 2017, China’s outward FDI flow accounts for 11.1% of global FDI flow and ranks second in the world, just following the United States. The share of China’s distribution outward FDI increased from around 28% in 2004 to more than 51%. Within distribution FDI, the wholesale FDI flow accounts for 17% of the total FDI flow. By comparison, production FDI only accounts for around 18%. Previous pioneering works such as Hanson et al. (2001) and Horstmann and Markusen (1995) make significant efforts for us to understand the characteristics of distribution FDI. Horstmann and Markusen (1995) argue that firms have two options in foreign markets: export or distribution FDI. Exporters need to find a local agent, which has private information advantage on its own effort and foreign market. Thus, home exporters have to pay additional information rent. By contrast, building a wholly-owned distribution affiliate requires extra fixed costs. So firms will make decisions by considering the trade-off between the two. By contrast, Hanson et al. (2001) implicitly assume that firm export and distribution-oriented FDI are complementary, as distribution FDI is set up to promote exports. They compare the trade-off between distribution- and production-oriented FDI and find that firms operating in countries with high income tax would prefer distribution FDI rather than production FDI, to avoid paying the high corporate tax. This chapter is mostly motivated by Hanson et al. (2001), in searching for the trade-off among exporting, distribution FDI, and production FDI. However, we are not still entirely clear why some firms choose distribution FDI while others do not, and why distribution FDI is more popular in some countries like China than in other countries. What causes some firms to engage in distribution outward FDI? Moreover, which investment characteristics in the host country matter for firms to engage in distribution FDI? the present chapter seeks to answer such questions. We argue that distribution FDI plays an important auxiliary but significant role to boost China’s exports. In accompany with China’s fast productivity growth in the new century (Feenstra et al., 2014), distribution FDI provides a cheaper alternative for a bunch of Chinese exporting firms to realize the cost-saving effects in reducing the cross-border communication costs. The current chapter presents four main findings. First, firms with distribution outward FDI are found to be more productive than non-FDI firms, but less productive than non-distribution FDI firms. These findings imply that the popularity of distribution outward FDI may be attributed to the fact that most Chinese exporting firms are insufficiently productive to set up overseas production lines. As a compromise, they set up a service or distribution center abroad to promote exports. This finding echoes the stylised fact that China’s exports have increased rapidly in the new century. In addition, we find strong sorting behavior between production FDI, distribution FDI, and non-FDI exports. To explain these findings, inspired by Oldenski (2012), we extend the model of Helpman et al. (2004) to understand this sorting behavior. Our estimates are based on a comprehensive FDI decision data set covering all Chinese FDI manufacturing firms during 2000–2008. However, it is important to stress that only a very small proportion of firms in our large sample engaged in FDI activity. Thus, the standard nonlinear binary estimates would have downward estimation bias

3.1 Introduction

43

(King & Zeng, 2001). We thus correct for such rare-events estimation bias in the chapter. Second, we distinguish the cross-country (i.e., cross-border) communications costs that occur during distribution and sales (like the costs of import procedures, promoting goods, and services before and after sales) from the usual transportation costs (i.e., iceberg transportation costs and tariffs) to demonstrate the importance of distribution outward FDI for exporting firms. We find that the higher are the cross-border communications costs, the higher is the probability that firms engage in distribution outward FDI. By contrast, the higher are the iceberg transportation costs, the higher is the probability that firms engage in non-distribution (or production) outward FDI. These findings are intuitive in the sense that, by setting up a business office or wholesale and retail subsidiary, the firm can largely reduce the asymmetric rent charged by local agents (Horstmann & Markusen, 1995). By contrast, firms can save on transportation costs when exporting is replaced by production FDI. These findings are also highly consistent with our theoretical predictions. Third, by allowing for firm heterogeneity in choosing cross-country host destinations, we find that the role of a firm’s productivity in its FDI flow differs by destination income. Highly productive firms are more likely to invest in rich countries, but not necessarily in poor countries. This finding persists when we check the intensive margin of the Linder hypothesis that rich countries receive more FDI flows. By estimating an endogenous threshold of income in host countries, our threshold regressions find support for the Linder hypothesis on FDI volume to high-income countries. Fourth, we find strong evidence on the intensive margin of distributionoriented FDI. We find that firm productivity significantly boosts distribution FDI flow once firms self-select into distribution FDI. Different from previous studies on Chinese outward FDI, we were able to obtain confidential information on the outward FDI flow for total FDI flow and distribution FDI flow in Zhejiang province, one of the most important FDI provinces in China. This is a novel finding in the literature on understanding China’s outward FDI, as the publicly released nationwide FDI decision data set has the substantial pitfall that data on firms’ FDI flows are unavailable. The chapter makes the following three contributions to the literature. First, it enriches the understanding of distribution outward FDI. As documented by Boatman (2007), as distribution FDI does not save production costs, distribution FDI has received little attention in the literature from theoretical and empirical works, except a few exceptions, such as Horstmann and Markusen (1995), Hanson et al. (2001), and Kimura and Lee (2006). We show that distribution FDI is complementary to firm export as a type of downward vertical FDI. As illustrated in our theoretical framework, firms face a trade-off between variable cost and fixed cost. Firms engaged in exporting without FDI, regardless of distribution or production orientation, bear an additional variable cost of cross-border communications (Oldenski, 2012).1 However, firms engaged in distribution FDI have a larger fixed cost. The 1

Oldenski (2012) finds evidence that firms would prefer exporting if the activities require complex within-firm communication. Instead, firms would prefer FDI if the goods and services require direct

44

3 Distribution, Outward FDI, and Productivity Heterogeneity …

trade-off between variable cost and fixed cost can be interpreted as a new form of the standard concentration-proximity trade-off. Thus, productivity heterogeneity plays an important role in understanding distribution FDI. Only highly productive firms would self-select into distribution FDI. Second, the chapter enriches the understanding of China’s distribution FDI. Different from China’s exports, on which there is already a fairly large micro-level literature (Qiu & Xue, 2014), few papers have investigated China’s FDI. Kolstad and Wilg (2012) find that Chinese FDI is attracted to three destinations: countries with lower institutional quality, countries that are rich in natural resources, and large markets. Using the same universal nationwide FDI decision data set, Chen and Tang (2014) find that firm productivity and the probability of firm FDI are positively correlated. Wang et al. use China’s firm-level data and find that access to external finance increases the probability that firms engage in outward FDI. Chen et al. (2019) explore how domestic distortions affect firms’ outward FDI decision. We take one step forward to examine a large and important part of China’s FDI—the distribution FDI. Our binary estimates find that the sorting predictions among non-FDI, distribution FDI, and production FDI work well in China. Thus, different from the mixed findings on Chinese exports and firm productivity,2 we confirm that the sorting behaviors among domestic sales, exporting, and FDI proposed by Helpman et al. (2004) apply to Chinese FDI firms. Third and more importantly, we explore the intensive margin of firm FDI flows (on all FDI and distribution FDI), which is almost completely absent in previous studies because of the unavailability of data. As introduced in detail in the next section, although the Ministry of Commerce of China released the list of FDI firms (henceforth, the FDI decision data set), the data set does not report each firm’s FDI volume in all years. To overcome this data challenge, we accessed a confidential FDI data set compiled by the Department of Commerce in Zhejiang province, which reports firms’ FDI volume in addition to all other information covered in the FDI decision data set. Thanks to this novel data set, we are able to explore the intensive margin of firm FDI in China. The rest of the chapter is organized as follows. Section 3.2 extends to show sorting equilibrium by productivity heterogeneity. Section 3.3 describes our data sample, followed by a careful scrutiny of measures of firm productivity. Section 3.4 examines the role of firm productivity in the firm’s FDI decision. Section 3.5 explores the intensive margin of FDI flows. Section 3.6 discusses the firm’s investment destination and Sect. 3.7 concludes.

communication with consumers. Based on Russ (2007), Ramondo et al. (2013) find that countries with less volatile fluctuations are served relatively more by foreign affiliates than by exporters. 2 Lu (2010) finds that Chinese exporters are less productive. However, Dai et al. (2016) and Yu (2015) argue that finding was because of the presence of China’s processing exporters, which are less productive than non-exporters and non-processing exporters. Once processing exporters are excluded, Chinese exporters are more productive than non-exporters, in line with the theoretical predictions of Melitz (2003).

3.2 Model

45

3.2 Model We construct a cross-country theoretical framework by extending Helpman et al. (2004) to capture the behavior of distribution FDI. We assume that each country has a representative constant elasticity of substitution utility function as follows:  U=

x(ϕ)

σ −1 σ

 σ σ−1 dϕ

Ώ

where x(ϕ) is the consumption of product ϕ, and σ > 1. Each firm in country i produces one product using labor as the only input, and the firm has a random  k labor productivity ϕ following Pareto distribution, where Pr(ϕ > x) = bx , k > ρ−1, b > 1, So ϕ1 is the variable production cost for each unit of goods produced. The firm first decides whether to enter the market. If entry, a sunk cost of f E is required to set up the production plant. After the entry, the firm observes his productivity ϕ. If the firm would like to serve foreign countries, there are three possible ways: (1) export without any foreign investment, (2) export and also set up a foreign affiliate to promote exports, and (3) set up a foreign plant to produce and sell overseas. The firm must pay a fixed cost f X + f S for the second choice, where f S is the up-front cost to set up a foreign affiliate; and a fixed cost f M for the third choice to build a foreign plant. Here we assume that the fixed costs satisfy the following ranking f M > f X + f S > f X > f D .3 We will validate these assumptions in the empirical part of the chapter as well. An iceberg transportation cost τij > 1 is needed for export, which means τ,- units of product are required for one unit sold in country j. But if the firm builds a distribution affiliation, the transportation cost may be reduced to μτ ij ;0 < μ < l, μτ ij > 1. The discount factor μ captures the cost reduction of investing in a trading subsidiary, which allows firms to distribute their products independently. Oldenski (2012) points out that the expenses incurred during communications between the domestic firm and foreign customers are crucial when firms are making the decision whether to export or build an overseas plant. It is important to distinguish cross-border communications costs from transportation costs. Most of cross-border communications costs are incurred after the goods are transported to the destination and can be reduced by setting up a local business office, that is, distribution FDI, which makes the import procedure and service more effective.4 However, transportation costs can be only phased out when the goods are no longer imported but 3

Note that fixed  costs  for production FDI can be decomposed into two components: fixed cost for production f MP and fixed cost for setting up the firm’s own distribution center ( f S ) which is similar to$ &the fixed cost of distribution FDI. As an usual assumption in the literature, fixed in production-type FDI is assumed to be higher than its counterpart for exports:  costP for production f M > f X We thus have f M = f MP + f S > f X + f s . 4 In practice, some communication costs such as consulting and negotiation could occur even before a trade deal. We thank a referee for pointing this out.

46

3 Distribution, Outward FDI, and Productivity Heterogeneity …

produced locally, that is, via production FDI. Another difference is that most of crossborder communications costs are irrelevant to firm productivity, since those costs are incurred after the transportation. Transportation costs are iceberg costs, which vary across firms with different productivity. To capture these aspects, similar to Berman et al. (2012), we introduce a linear cross-border communication cost in our model. We assume that firms that only export have to pay η- units of labor for the communications costs additional to production costs, but those who build an overseas distribution foreign affiliate do not. The value of η- captures the cost-saving effects from establishing a business office, which helps firms to serve foreign customers by promoting sales and improving after-sales services. In this way, a destination country with a poor doing-business environment may be associated with a poor record in enforcing contracts, which would generate more communications costs. Different from production FDI, which saves transportation costs, distribution FDI mainly reduces the cross-border communications costs incurred. As in Helpman et al. (2004), wages (w) are equal to unity across countries by introducing a homogenous good sector in which one unit of labor is used to produce one unit of output. The homogenous good can be traded freely and an exogenous fraction of income is spent on it. The marginal cost for each product sold, MC d = τ w μτ w w w , MC e = iϕj + η j w, MC s = ϕi j , MC m = ϕj , represents the marginal cost ϕ for selling in the domestic market, exporting without foreign investment, exporting as well as distribution investment, and building a foreign production plant. The derived demand for product ϕ is −σ  X j (ϕ) = L j P jσ −1 p cj (ϕ) where L j is labor income in country j; p cj (ϕ) = σ σ−1 MC c , c = d, e, s, m is the price of product uif it is sold domestically, exported with a production affiliate, respectively. P- is the aggregate price level in which its exact expression is shown in Appendix A. Inspired by Berman et al. (2012), the profits for domestic sales, exports, distribution FDI, and production FDI are as follows:  1−σ 1 Bi − f D ϕ 1−σ  τi j + ηj Bj − fX πiej = ϕ   μ j τi j 1−σ πiSj = B j − fX − fS ϕ  1−σ 1 m Bj − fM πi j = ϕ πid =

(3.1)

(3.2)

(3.3)

(3.4)

3.2 Model

47

 1−σ where B j ≡ σ1 σ σ−1 L j P jσ −1 . The productivity cut-off points satisfy πid = 0, πiej = 0, πiSj = πiej , πimj = πiSj explicitly: 

  1−σ   μ j τi j 1−σ τi j 1 1−σ fD fs = ; − + ηj = ϕdj Bi ϕsj ϕs j Bj  1−σ 1−σ  1−σ  τi j μ j τi j fX 1 f M − fs − f X + ηj = ; − = ϕej Bj ϕmi ϕmj Bj



















where ϕdj , ϕej , ϕsj , ϕmj is the productivity cut-off point for each mode, respectively. As free entry, the expected profit of firm entry is zero. The expected profit after entry equals the entry cost f E :  ∞

ϕdt

πid dG(ϕ) +



N

j=1, j/ =i



ϕsi j

ϕei j

 ϕ mi j



πiej dG(ϕ) +



ϕsi j

πiSj dG(ϕ) +

 ∞ ϕmi j

πimj dG(ϕ) = f E





(3.5)



Equations (3.1)–(3.5) jointly solve the equilibrium ϕdj , ϕeij , ϕsi j , ϕmij , and B j for each country i, j. Note that the equilibrium is irrelevant to the market size (L-). For simplicity, countries are assumed to be symmetric following Melitz (2003). As η j = η, τi j = τ, μ j = μ, every country has the same productivity cut-off points ϕd , ϕe , ϕs , ϕm , and B. We thus have following findings.







1

Proposition 4.1 When every country is symmetric, ff DX > τ 1−σ , f X1−σ −   σ −1 1 μσ −1 1 − 1 where Δ is any upperbound f + f S ) 1−σ > η△, f M > f X + f S 1−μ σ −1 τ μ( X 1









of B 1−σ , we have ϕd < ϕe < ϕs < ϕm . Proof See Appendix A for details. Proposition 4.1 suggests that the most productive firms engage in production FDI, the next most productive firms engage in distribution FDI and export, the even next most productive firms only export, the further next productive firms do not export but only sell in the domestic market, and the least productive firms exit. The intuition is straightforward: only the most productive firms can overcome the highest fixed costs to build an overseas production plant and benefit from the cost-saving effect of crossborder communications costs and transportation costs. Less productive firms, like most of the Chinese FDI firms, can only afford the fixed costs of building international business services or distribution centers to reduce cross-border communications costs to promote their exports. The sorting equilibria for different cutoff points are shown in Fig. 3.1. Proposition 4.2



i. An increase in export-specific communication cos η raises ϕe , ϕs . but does not affect ϕm .

48

3 Distribution, Outward FDI, and Productivity Heterogeneity …

Fig. 3.1 Firm productivity, export, distribution and non-distribution FDI





ii. An increase in iceberg transportation cost τ increases ϕe , ϕs and decreases ϕm . Proof See Appendix B for details. Proposition 4.2 implies that higher cross-country cross-border communications costs η and lower foreign tariffs (lower τ ) increase the probability of distribution foreign investment. This is because most of the cross-border communications costs can be reduced via distribution FDI. Thus, a higher increase the attractiveness of distribution FDI compared with exporting only, but does not alter the benefit of production FDI. However, the transportation costs still exist as long as goods are exported. So a higher tariff imposed by importing countries promotes production FDI and hampers export and distribution FDI. We now turn to test these theoretical predictions.

3.3 Data and Measures To investigate the impact of firm productivity on distribution FDI, we rely on three disaggregated data sets. The first data set provides the list of FDI firms in China. This data set is crucial for understanding firms’ FDI decision. However, the data set does not report any FDI values. To examine the role of the intensive margin, we rely on another firm-level FDI data set, which contains information on the universal firm-level FDI activity in Zhejiang province of China. Finally, we merge the firmlevel manufacturing production data with the two FDI data sets to explore the nexus between FDI and firm productivity.

3.3 Data and Measures

49

3.3.1 FDI Decision Data The nationwide data set of Chinese firms’ FDI decisions was obtained from the Ministry of Commerce of China (MOC). MOC requires every Chinese FDI firm to report its detailed investment activity since 1980. To invest abroad, every Chinese firm is required by the government to apply to the MOC and its former counterpart, the Ministry of Foreign Trade and Economic Cooperation of China, for approval and registration. MOC requires such firms to provide the following information: the firm’s name, the names of the firm’s foreign subsidiaries, the type of ownership (i.e., state-owned enterprise (SOE) or private firm), the investment mode (e.g., tradingoriented affiliates, mining-oriented affiliates), and the amount of foreign investment (in U.S. dollars). Once a firm’s application is approved by MOC, MOC will release the information mentioned above, as well as other information, such as the date of approval and the date of registration abroad, to the public. All such information is available except the amount of the firm’s investment, which is considered to be confidential information to the firms. Since 1980, MOC has released information on new FDI firms every year. Thus the nationwide FDI decision data indeed report FDI starters by year. The database even reports specific modes of investment: trading office, wholesale center, production affiliate, foreign resource utilization, processing trade, consulting service, real estate, research and development center, and other unspecified types. Here trading offices and wholesale centers are classified as distribution FDI, whereas the rest are referred to as non-distribution FDI. However, since this data set does not report firms’ FDI flows, researchers are not able to explore the intensive margin of firm FDI with this data set.

3.3.2 FDI Flow Data To explore the intensive margin, we use another data set, which is compiled by the Department of Commerce of Zhejiang province. The most novel aspect of this data set is that it includes data on firms’ FDI flows (in current U.S. dollars). The data set covers all firms with headquarters located (and registered) in Zhejiang and is a short, unbalanced panel from 2006 to 2008. In addition to the variables covered in the nationwide FDI data set, the Zhejiang data set provides each firm’s name, city where it has its headquarters, type of ownership, industry classification, investment destination countries, and stock share from its Chinese parent company. Although this data set seems ideal for examining the role of the intensive margin of firm FDI, the disadvantage is also obvious: the data set is for only one province in China.5 Regrettably, as is the case for many other researchers, we cannot access 5

To our knowledge, almost all previous work was not able to access nationwide universal outward FDI flow data. An outstanding exception is Wang et al. (2012), who use nationwide firm-level outward FDI data to investigate the driving force of outward FDI of Chinese firms. However, the

50

3 Distribution, Outward FDI, and Productivity Heterogeneity …

similar databases from other provinces. Still, as discussed in Appendix C, we believe that Zhejiang’s firm-level FDI flow data are a good proxy for understanding the universal Chinese firm’s FDI flows. In particular, the FDI flows from Zhejiang province are outstanding in the whole of China; the distribution of both types of ownership and that of Zhejiang’s FDI firms’ destinations and industrial distributions are similar to those for the whole of China.

3.3.3 Firm-Level Production Data Our last database is the firm-level production data compiled by China’s National Bureau of Statistics in an annual survey of manufacturing enterprises. The data set covers around 162,885 firms in 2000 and 410 thousand firms in 2008 and, on average, accounts for 95% of China’s total annual output in all manufacturing sectors. The data set includes two types of manufacturing firms: universal SOEs and non-SOEs whose annual sales are more than RMB 5 million (or equivalently around $746,000 under the current exchange rate). The data set is particularly useful for calculating measured total factor productivity (TFP), since the data set provides more than 100 firm-level variables listed in the main accounting statements, such as sales, capital, labor, and intermediate inputs. As highlighted by Yu (2015) and Chen et al. (2019), some samples in this firmlevel production data set are noisy and somewhat misleading, largely because of misreporting by some firms. To guarantee that our estimation sample is reliable and accurate, we screen the sample and omit outliers by adopting the following criteria (Feenstra et al., 2014). First, we eliminate a firm if its number of employees is less than eight workers, since otherwise such an entity would be identified as self-employed. Second, a firm is included only if its key financial variables (e.g., gross value of industrial output, sales, total assets, and net value of fixed assets) are present. Third, we include firms based on the requirements of the Generally Accepted Accounting Principles.6

3.3.4 Data Merge We then merge the two firm-level FDI data sets (i.e., nationwide FDI decision data and Zhejiang’s FDI flow data) with the manufacturing production database. Although study uses data only from 2006 to 2007; hence, it cannot explore the possible effects of the financial crisis in 2008. 6 In particular, an observation is included in the sample only if the following observations hold: (1) total assets are greater than liquid assets; (2) total assets are greater than the total fixed assets and the net value of fixed assets; (3) the established time is valid (i.e., the opening month should be between January and December); and (4) the firm’s sales must be higher than the required threshold of RMB 5 million.

3.3 Data and Measures

51

the two data sets share a common variable—the firm’s identification number—their coding systems are completely different. Hence, we use alternative methods to merge the three data sets. The matching procedure involves three steps. First, we match the three data sets (i.e., firm production data, nationwide FDI decision data, and Zhejiang FDI flow data) by using each firm’s Chinese name and year. If a firm has an exact Chinese name in a particular year in all three data sets, it is considered an identical firm. Still, this method could miss some firms since the Chinese name for an identical company may not have the exact Chinese characters in the two data sets, although they share some common strings.7 Our second step is to decompose a firm name into several strings referring to its location, industry, business type, and specific name, respectively. If a company has all identical strings, such a firm in the three data sets is classified as an identical firm.8 Finally, to avoid possible mistakes, all approximate string-matching procedures are done manually. Row (1) of Table 3.1 reports the number of manufacturing firms and row (2) reports the number of FDI starting firms by year during 2000–2008. Row (3) reports the number of matching FDI manufacturing firms.9 The share of FDI manufacturing firms over total manufacturing firms shown in row (5) suggests that FDI indeed is a rare event—the share is less than 1% each year. The number of FDI manufacturing firms increased dramatically after 2004. More importantly, row (6) shows that the share of distribution FDI manufacturing firms over total FDI manufacturing firms increased from around 14% in 2000 to 55% in 2008, suggesting that distribution FDI has become more and more important over time. By using these two methods, we match Zhejiang’s manufacturing firms with Zhejiang’s FDI flow firms. As shown in the lower module of Table 3.1, of 1270 FDI firm-years in Zhejiang province from 2006 to 2008, 407 FDI firms are engaging in manufacturing sectors, suggesting that around two-thirds of Zhejiang FDI parent firms are from service sectors or are trading intermediates (Ahn et al., 2010). Table 3.2 reports the summary statistics of firm characteristics for nationwide manufacturing firms and Zhejiang’s manufacturing firms, respectively. The small mean of FDI indicator in both samples ascertains that FDI is a rare event during the sample periods. Finally, as the main interest of this chapter is how firm productivity affects distribution FDI, we carefully measure TFP. The augmented Olley-Pakes TFP is constructed following Brandt et al. (2012) and Yu (2015). Appendix D provides the detailed steps of our measured TFP. In particular, we estimate the production function for 7

For example, “Ningbo Hangyuan communication equipment trading company” shown in the FDI data set and “(Zhejiang) Ningbo Hangyuan communication equipment trading company” shown in the National Bureau of Statistics of China production data set are the same company but do not have exactly the same Chinese characters. 8 In the example above, the location fragment is “Ningbo,” the industry is “communication equipment,” the business type is “trading company,” and the specific name is “Hangyuan.” 9 Note that we merge FDI data and manufacturing production data by firm name rather than by name-year. Number of FDI manufacturing firms in row (3) reports not only FDI starting firms, but also FDI continuing firms. Thus, it is possible that there are fewer FDI starters than matched FDI manufacturing firms, as shown in 2007 and 2008.

0.34

0.23

14.2

(5) FDI share (%)

(6) Distribution FDI share (%)

0.40

0.45

0.48

0.49

0.43

0.44

0.41

0.31

(10) FDI share (%)

0.48

131

427

27,433

55.4

0.46

656

1,183

1,018

222,312

2008

Note Data are from the Ministry of Commerce of China and the authors’ calculations. FDI share (% = (3)/(1)) is obtained by dividing the number of FDI starting firms by the number of manufacturing firms nationwide and in Zhejiang province, respectively. Distribution FDI share (= (4)/(3)) is obtained by dividing the number of distribution FDI manufacturing firms by the number of total FDI manufacturing firms. For Zhejiang firms, the FDI decision data are available every year during the sample, but the FDI flow data are available only for 2006–2008, for which there are 1,270 FDI firms in Zhejiang and 407 of them are manufacturing firms

163

39,465

52.7

113

53.5

616

1,168

1,140

258,246

2007

(9) FDI mfg. firms

51.9

407

761

1,081

248,601

2006

419

38.8

224

431

984

198,285

2005

35,887

13.3

40

103

972

200,989

2004

424

15.0

4

30

587

129,720

2003

(7) Mfg. firms

11.7

3

20

444

110,522

2002

(8) FDI firms

Zhejiang’s FDI flow data

2

2

(4) Distribution FDI mfg. firms

340 17

197

14

(2) FDI starting firms

100,091

2001

(3) FDI mfg. firms

84,974

2000

(1) Mfg. firms

Nationwide FDI decision data

Firm type

Table 3.1 FDI share in number of manufacturing firms (2000–2008)

52 3 Distribution, Outward FDI, and Productivity Heterogeneity …

3.4 Extensive Margin of FDI

53

Table 3.2 Summary statistics of key variables Sample covered

Nationwide (2000–2008)

Zhejiang (2006–2008)

Variable

Mean

Std. dev

Mean 3.27

1.53

Firm FDI indicator

0.004

0.066

0.003

0.05

Firm TFP (Olley-Pakes)

3.61

1.18

4.08

0.94

Firm log FDI

Std. dev

Firm export indicator

0.29

0.451

0.42

0.49

SOE indicator

0.05

0.219

0.002

0.047

Foreign indicator

0.20

0.402

0.16

0.366

Firm log labor

4.78

1.115

4.45

0.983

exporting and non-exporting firms separately in each industry. The idea is that different industries may use different technology; hence, firm TFP must be estimated for each industry. Equally important, even within an industry, exporting firms may use completely different technology than non-exporting firms. For example, some exporters, like processing exporters, only receive imported material passively (Feenstra & Hanson, 2005) and hence do not have their own technology choice. We hence estimate TFP for exporters and non-exporters separately. We now turn to describe distribution FDI in our merged data set. In both FDI data sets, there is a variable used to describe the type of firm FDI, which includes mining, construction, R&D, production, processing trade, market seeking, wholesale, business service, and product design. As our main interest is of distribution FDI, both wholesale FDI and business-service FDI are classified to distribution FDI, following the official definition of MOC of China. Appendix Table 1 reports the proportion of distribution FDI in our sample. In the nationwide FDI data, the number of distribution FDI firms accounts for roughly half of whole FDI firm. Such a proportion even increases to 60% after merging with the production data set. Similarly, nearly 76% samples are distribution FDI in Zhejiang FDI data. The percentage also rises to 80% after merging with production data. All these suggest that distribution FDI is important in China today.

3.4 Extensive Margin of FDI This section discusses how a firm’s productivity affects the firm’s decision to engage in FDI (i.e., the extensive margin). Before running the regressions, we provide several preliminary statistical tests to enrich our understanding of the difference in productivity between distribution FDI and non-FDI firms (and non-distribution FDI firms), following a careful scrutiny of the effect of firm productivity on the decision to engage in (distribution) FDI.

54

3 Distribution, Outward FDI, and Productivity Heterogeneity …

Fig. 3.2 Firm productivity by firm type

3.4.1 Descriptive Analysis on Productivity Differences Proposition 4.1 suggests that firms’ sales decision can be sorted by their productivity. Low-productivity firms serve in domestic markets, high-productivity firms export, higher-productivity firms engage in distribution FDI, and even higher-productivity firms participate in non-distribution FDI. Figure 3.2 exhibits the productivity distributions for non-FDI firms, distribution FDI firms, and non-distribution FDI firms, respectively. Overall, firm productivity for distribution FDI is clearly higher than for non-FDI firms, but lower than (though not obvious) for non-distribution FDI. Eaton et al. (2011) Find that higher-productivity firms are usually larger. If so, we would observe that, compared with non-FDI firms, FDI firms on average are larger, more productive, and export more. Table 3.3 checks the difference between nonFDI and FDI firms on their TFP, labor, sales, and exports. Compared with non-FDI firms, distribution FDI firms are found to be more productive, hire more workers, sell more, and export more. By sharp contrast, compared with non-distribution FDI firms, distribution FDI firms are found to be less productive, hire fewer workers, sell less, and export less. The t-values for these variables are strongly significant at the conventional statistical level.10 However, the simple t-test comparisons may not be sufficient to conclude that distribution FDI firms are more productive than non-FDI firms, since FDI firms are very different from non-FDI firms in terms of size (number of employees and sales) and experience in foreign markets, as already seen. 10

Note that the sample size of non-FDI is 1,137,907 which is lower than the number of observation (i.e., 1,553,740) shown in Table 3.1 due to the fact that some firms’ TFP are missing if any of following data of firm’s capital, labor, intermediate inputs are unavailable.

1,773

4.411 −0.086 (−1.39)

4.411 −0.231*** (−5.73)

0.442*** (8.44)

0.576***

4.180

(21.55)

4.180

0.478*** (9.54)

0.685***

4.289

3.811

(35.41)

4.289

3.604

(3)

Matched

(−8.45)

−0.505***

6.109

(34.83)

0.862***

5.604

(61.34)

1.102***

5.844

4.742

(4)

Log employees

(−9.19)

−4,391,884***

5,637,258

(50.37)

1,148,815***

1,245,375

(150.3)

3,238,112***

3,334,672

96,560

(5)

Sales

(−7.63)

−1,733,972***

2,194,766

(39.22)

442,178***

460,793

(120.1)

1,267,060***

1,285,675

18,615

(6)

Export

Note Numbers in parentheses are t-values. *** Denotes significance at the 1% level. Columns (2) and (3) report the TFP comparison by using the nearest-matching propensity score matching (PSM) approach. Column (2) is unmatched whereas column (3) is the average treatment on the treated (ATT) approach. The treated group is FDI firms, whereas the control group is non-FDI firms. Firm size (in log labor), exports, and sales are used as covariates to obtain the propensity score. Since there are observations with identical propensity score values, the sort order of the data could affect the results. The sort order is made to be random before adopting the PSM approach

Difference = (iii) − (iv)

(iv) Non-distribution FDI firms

Difference = (iii) − (i)

(iii) Distribution FDI firms

1,954

3,727

(ii) FDI firms

Difference = (ii) − (i)

1,137,097

(i) Non-FDI firms

FDI firms versus non-FDI firms

(2)

Unmatched

(1)

Propensity score matching

Mean TFP

# of obs

Variable

Category

Table 3.3 Difference between non-FDI and FDI firms

3.4 Extensive Margin of FDI 55

56

3 Distribution, Outward FDI, and Productivity Heterogeneity …

We thus follow Imbens (2004) and perform propensity score matching (PSM) by choosing the number of firm employees, firm sales, and firm exports as covariates. Each FDI firm is matched to its most similar non-FDI firm. Since there are observations with identical propensity score values, the sort order of the data could affect the results. We thus perform a random sort before adopting the PSM approach. Column (3) in Table 3.3 reports the estimates for average treatment for the treated (ATT). The coefficient of ATT for distribution FDI manufacturing firms is 0.442 (compared with non-FDI firms) and highly statistically significant, suggesting that, overall, productivity for distribution FDI firms is higher than that for similar non-FDI firms during the period 2000–2008. Strikingly, compared with non-distribution FDI firms, the coefficient of ATT for distribution FDI is insignificant. To check this out, we examine productivity difference by year for each type of firm: non-FDI, distribution FDI, and non-distribution FDI firms. Table 3.4 shows that FDI firms are more productive than non-FDI firms by year during the sample period 2000–2008.11 The productivity difference between distribution FDI firms and nonFDI firms is significantly positive before 2003. However, this might be purely due to the fact that only few FDI firms engage in distribution in the early years. Interesting, the gap roughly declines over the period (especially after 2004), also suggesting that distribution FDI firms indeed enjoy much productivity gain via learning from investing.

3.4.2 Extensive Margin of FDI To examine whether firm productivity plays a key role in the firm’s decision to engage in distribution FDI, we start by checking whether productivity affects the firm’s FDI decision, as distribution FDI is a type of FDI. In particular, we consider the following empirical specification:   Pr OFDIi jt = 1 = β0 + β1 ln T F Pit + θ X + j + ηt + εit ,

(3.6)

where OFDI i jt and lnT F Pit represents FDI indicator and the log productivity of firm i in industry j in year t, respectively. X denotes other firm characteristics, such as firm size (produced by firm’s log of employment) and types of ownership (i.e., foreign invested firms or SOEs).12 For instance, private firms are more (equivalently, SOEs might be less) likely to invest abroad because of domestic input distortion in 11

Note that TFP in 2008 is calculated and estimated differently. As in Feenstra et al. (2014), we use deflated firm value added to measure production and exclude intermediate inputs (materials) as one kind of factor input. However, we are not able to use value added to estimate firm TFP in 2008, since it is absent in the data set. We instead use industrial output to replace value added in 2008. Thus, we have to be cautious in comparing TFP in 2008 with TFP in previous years. 12 Here, a firm that has investment from foreign countries or Hong Kong/Macao/Taiwan is defined as a foreign firm, following Feenstra et al. (2014).

(3.49)

5.514

4.210

2.404***

(3.23)

(3) Distribution FDI firms

(4) Non-distribution FDI firms

TFP difference = (3) − (1)

1.557***

(2.75)

1.304***

(2.03)

4.007

5.564

4.190

(3.55)

1.778***

(4.50)

2.670***

4.109

5.888

4.376

3.218

2002

(−1.44)

(−1.90)

(4.56) −0.502***

(2.43)

0.790***

4.358

3.855

4.163

3.065

2004

−0.917

1.231***

5.432

4.514

5.309

3.283

2003

(−4.34)

−0.458***

(3.09)

0.213***

4.093

3.634

3.855

3.421

2005

(−4.45)

−0.344***

(0.73)

0.038

3.923

3.578

3.738

3.540

2006

(−4.35)

−0.282***

(2.01)

0.085**

4.026

3.744

3.738

3.659

2007

(−1.79)

−0.098*

(5.14)

0.184***

5.248

5.150

5.194

4.966

2008

FDI and non-distribution FDI firms and firm TFP difference between distribution FDI and non-FDI firms

Note Numbers in parentheses are t-values. *** (** , * ) Denotes significance at the 1% (5%, 10%) level. The table reports firm TFP difference between distribution

TFP difference = (3) − (4)

2.562***

4.396

(2) All FDI firms

3.001

3.109

(1) Non-FDI firms

2001

2000

Firm TFP

Table 3.4 TFP difference between distribution FDI firms and other firms by year

3.4 Extensive Margin of FDI 57

58

3 Distribution, Outward FDI, and Productivity Heterogeneity …

China (Chen et al., 2019; Hsieh & Klenow, 2009). In addition, larger firms are more likely to invest abroad because they may have an additional advantage to realize increasing returns to scale (Helpman et al., 2004). Inspired by Oldenski (2012), we also include a firm export indicator in the estimations, since an exporting firm could find it easier to invest abroad, given that it would have an information advantage on foreign markets compared with non-exporting firms. Moreover, as the measured TFP cannot be compared over industries, we normalize TFP in each industry to a range between zero and one, following Arkolakis and Muendler. Finally, as stressed by Ishikawa et al. (2010), a host country has some regulations for foreign investment. Such a concern may be relevant and important for Chinese FDI, in particular in the mining industries. Although Chinese parent firms in mining industries are not covered in our data set, foreign investment regulation may be present for some manufacturing industries. To this purpose, the error term is decomposed into three components: (1) industry-specific fixed effects, (2) year-specific fixed effects ηt to control for firm-invariant factors such as Chinese RMB appreci ation, and (3) an idiosyncratic effect εit with normal distribution εit ∼ N 0, σi2 to control for other unspecified factors. Industry and year fixed effects are used to capture possible industry heterogeneity due to foreign regulations and other possible industry-variant and year-variant factors. We start from a simple linear probability model (LPM) to conduct our empirical analysis. It is worthwhile to stress that it is inappropriate to perform firm-specific fixed effects here, given that our nationwide outward FDI data are pooled crosssection data, as we only know the year that firms start to engage in FDI but do not know the year that firms continue or cease FDI. Table 3.5 (except the last column) thus only includes observations with FDI starters and non-FDI firms. We include the two-digit Chinese industry classification (CIC) level industry-specific fixed effects in the LPM estimates in column (1) in Table 3.5. The key coefficient of firm TFP is positive and significant, although its magnitude seems very small.13 We suspect that this is due to the well-known drawback of using the linear probability model, which is that there is no justification for why the specification is linear. In addition, the predicted probability could be less than zero or greater than one, which does not make sense. We therefore perform the probit and logit estimations using twodigit CIC-level fixed effects in columns (2) and (3), respectively, and the result is confirmed.14

13

Note that one must be cautious on the statistical significance in the LPM estimation since the LPM model has serious problem of heteroscedasticity. We thank a referee to point this out. 14 Note that the coefficients shown in the Probit estimates are not marginal effects, which are not reported here given that it is straightforward to calculate the marginal effects in the rare event Logit estimates in column (4).

Complementary

Logit

Logit

Yes

Yes

Yes

Yes

1,138,450

Year fixed effects

Industry fixed effects

Number of observations 1,137,309

Yes

Yes

No

No

(19.64)

1.373***

(4)

1,138,450

Yes

Yes

No

No

(19.63)

1.373***

(28.30)

0.614***

(−8.16)

−0.575***

(−4.96)

−0.876***

(5.68)

1.086***

(5)

1,138,450

Yes

Yes

No

No

(20.46)

1.372***

(29.31)

0.610***

(−8.76)

−0.573***

(−4.85)

−0.878***

(5.61)

1.084***

(6)

896,958

Yes

Yes

Yes

No

(19.19)

1.460***

(25.56)

0.642***

−

(−5.43)

−0.963***

(4.66)

1.079***

(7)

1,097,537

Yes

Yes

No

Yes

(19.40)

1.370***

(27.13)

0.605***

(−8.18)

−0.578***

−

(5.11)

0.988***

(8)

1,139,683

Yes

Yes

No

No

(27.60)

1.096***

(42.90)

0.577***

(−15.55)

−0.704***

(−5.96)

−0.613***

(7.69)

0.893***

Note The regressand is the FDI indicator. Numbers in parentheses are t-values. *** Denotes significance at the 1% level. Column (6) drops foreign firms from the sample whereas column (7) drops SOEs from the sample

1,137,309

No

No

No

No

SOE dropped

(19.91)

(15.27)

(28.33)

(26.18)

0.431***

(15.59)

0.002***

0.615***

0.208***

0.001***

−0.577*** (−8.18)

−0.183***

(−7.77)

−0.001***

(−4.95)

−0.874***

(5.62)

1.076***

(−5.26)

Foreign firms dropped

Export indicator

Log firm labor

Foreign indicator

−0.310***

(−5.17)

−0.001***

(−3.91)

(5.11)

(4.54)

SOE indicator

(2)

0.321***

(1)

0.001***

Variable

Firm relative TFP

(3)

Ever FDI Rare event logit

Logit

Logit

Econometric method

Probit

FDI starter

LPM

Regressand: indicator of

Table 3.5 Effects of firm productivity on FDI decision (2000–2008)

3.4 Extensive Margin of FDI 59

60

3 Distribution, Outward FDI, and Productivity Heterogeneity …

3.4.3 Estimates with Rare Events Corrections Our estimations above may still face some bias. As observed from Tables 3.1 and 3.2, of the total 1,138,450 observations, on average only 0.44% of firms engage in FDI. Thus, our sample exhibits the features of rare events that occur infrequently but may have important economic implications. As highlighted by King and Zeng (2001), standard econometric methods such as logit and probit would underestimate the probability of rare events, although maximum likelihood estimators are still consistent. To see this, consider a simplified logit regression of the FDI dummy on firm TFP. Pr(OFDIit = 1) = ⋀(β1 ln TFPit ) =

exp(β1 ln TFPit ) 1 + exp(β1 ln TFPit )

(3.7)

where ⋀(·) is the logistic cumulative density function (henceforth CDF). Since β 1 > 0, as shown in columns (1)–(3) of Table 3.5, the probability of OFDIit =1 is positively associated with firm TFP; most of the zero-FDI observations will be to the left and the observation with OFDIit =1 will be to the right with little overlap. Since there are around 1.5 million observations with zero FDI, the standard binary estimates can easily estimate the illustrated probability density function curve without error, as shown by the solid line in Fig. 3.3.15 However, since only 0.44% of the observations have positive FDI, any standard binary estimates of the dashed density line for firm TFP when OFDIit =1 will be poor. Because the minimum of the observed rare OFDIit sample is larger than that of the unobserved FDI population, the cutoff point that best classifies non-FDI and FDI would be too far from the density of observations with OFDIit =1. This will cause a systematic bias toward the left tail and result in an underestimation of the rare events with OFDIit =1 (See King and Zeng (2001) for a detailed discussion). As recommended by King and Zeng (2001), the rare-events estimation bias can be corrected as follows. We first estimate the finite sample bias of the coefficients, ˆ to obtain the bias-corrected estimates βˆ − bias(β) ˆ , where βˆ denotes the bias(β), coefficients obtained from the conventional logistic estimates.16 Column (4) in Table 3.5 reports the logit estimates with rare-events corrections. The coefficient of firm TFP is slightly larger than its counterpart in column (5), suggesting that the estimation bias is not so severe. An alternative approach to correct possible rare-events estimation errors is to use the complementary log–log model.17 The idea is that the distributions of standard 15

To illustrate the idea in a simple way, the distribution curves are drawn to be normal, although this need not be the case. 16 Chen (2015) also adopts this method to explore how negative climate shocks (e.g., severe drought, locust plagues) affected peasant uprisings.      17 The CDF of the complementary log–log model is C X ´ β = 1 − exp − exp X ´ β with margin effect exp(− exp(X ´ β)) exp(X ´ β)β. The complementary log–log model also has an additional advantage to avoid a strong assumption of normal distribution in the Probit model, which seems.

3.4 Extensive Margin of FDI

61

Fig. 3.3 Rare events of FDI firms. Note Samples are sorted by firm TFP. The short vertical line represents rare observations with FDI = 1 whereas the many observations with FDI = 0 are not drawn. The solid (dotted) curve refers to the probability density with FDI = 1 (FDI = 0). The cutoff points that best classify FDI = 0 and FDI = 1 would be too far to the right as argued in the text.

binary nonlinear models, such as probit and logit, are symmetric to the original point. So the speed of convergence toward the probability that OFDIit =1 is the same as that for OFDIit =0. This violates the feature of the rare events, which exhibit faster convergence toward the probability that OFDIit , = 1. The complementary log-log model can address this issue, since the model has a left-skewed extreme value distribution, which also exhibits a faster convergence speed toward the probability that OFDIit =1 (Cameron & Trivedi, 2005). The complementary log–log model in column (5) in Table 3.5 shows that the coefficient of firm TFP is fairly close to its counterparts in conventional logit estimates and rare-events logit estimates, suggesting that the estimation bias caused by the property of “rare events” is not so severe in our estimates. One possible reason is that we still do not control for possible reverse causality of FDI on firm productivity, which will be addressed shortly. So far, we include foreign multinational firms in the regressions. But there may be a concern that such foreign firms do not really fit with our analysis for two reasons. First, we only observe a selected sample of foreign firms that have already chosen to be present in China. Second, it is possible that some Chinese domestic firms invest in Hong Kong and Macao and hence should be treated as “multinational” firms, which in turn invest back in China. To avoid such bound-back behavior, we drop foreign firms in column (6) and still find similar results. Another issue is about our FDI decision data per se. As we only observe firms that engage in new FDI, it is good enough for us to examine firms that transition from non-FDI to (any type of) FDI. However, as we do not have information on firms exiting FDI, we are not able to control for this. A possible concern is that some firms were SOEs but then were privatized. Since these firms may have made FDI decisions in the past that were not profit maximizing, once privatized, the firms may decide to unload assets that are not profitable. We indeed observe some indirect evidence from the regressions. The coefficients of SOEs in our previous tables are negative and significant, suggesting that SOEs are less likely to engage in FDI activity.

62

3 Distribution, Outward FDI, and Productivity Heterogeneity …

To address this concern, we run two experiments. First, we drop SOEs from the sample to see whether our main result is affected by SOEs. The estimates in column (7) in Table 3.5 show that our main results are not changed by doing so. Finally, for the sake of completeness, we include firms that ever-employed FDI and non-FDI firms in the last column in Table 3.5.18 In any case, our benchmark findings are insensitive to such robustness checks. Highly productive firms are more likely to engage in FDI.

3.4.4 Multinomial Logit Estimates with Distribution FDI We now examine whether this finding—that highly productive firms are more likely to engage in FDI—applies to distribution FDI firms. Table 3.6 is our first key table. The regressands in columns (1)–(2) are FDI mode, in which zero refers to non-FDI, one is distribution FDI, and two is non-distribution FDI. We again use firm relative TFP to measure firm productivity in all estimates. As firms’ decisions to engage in non-FDI or distribution FDI or non-distribution FDI are made simultaneously, we adopt the multinomial logit model in which the regressand in column (1) is distribution FDI, whereas that in column (2) is non-distribution FDI. The positive and significant sign of firm productivity in column (1) suggests that highly productive firms are more likely to engage in distribution FDI than non-FDI. The coefficient of firm productivity in column (2) is again positive and significant. More importantly, its magnitude is larger than its counterpart in column (1), suggesting that even higher productive firms are more likely to engage in non-distribution FDI.19 Finally, we find that larger firms are more likely to invest abroad, whereas SOEs are less likely to do so. Exporting firms are more likely to engage in distribution FDI by employing their information advantage (Oldenski, 2012), which is in line with the intuition that distribution FDI serves trade. There are four important caveats for these key findings. First, our theoretical model and empirical regressions discuss three options of firm choice: non-FDI, distribution FDI, and non-distribution FDI. However, it is possible that some firms do not directly export their products and hence have no incentive to set up their own distribution center abroad. Instead, they may rely on domestic trade intermediaries to sell their products abroad (Ahn et al., 2011).20 As we have already dropped those firms, our current estimates would not suffer from such a concern. 18

If a firm ever engaged in FDI, we assume that it always engages in FDI afterward during the sample. 19 To ensure that the findings above are not driven by the mass of non-FDI firms, we also drop non-FDI firms and perform the logit (probit) estimates in which the regress and is the distribution FDI indicator (i.e., zero refers to non-distribution FDI and one refers to distribution FDI). It turns out that firm TFP has negative and significant coefficients in all the experiments. Such findings, which are not reported here to save space although they are available upon request, thus are consistent with the sorting behavior illustrated by our theoretical model above. 20 As in Ahn et al. (2011), about 20% of observations are identified and dropped in the customs data.

(−4.76)

(−6.74)

(10.27) No No Yes

(18.42)

No

No

Yes

No

Yes

Yes

1,138,450

Foreign firms dropped

Pure exporters dropped

Non-FDl firms with foreign M&A

Most contractable industries dropped

Year fixed effects

Industry fixed effects

Number of observations

1,021,633

Yes

Yes

Yes

Yes

No

No

(17.00)

1.792***

(17.63)

0.549***

(−5.42)

−0.504***

(−4.10)

−1.594***

(3.64)

0.779***

Distri. (3)

Yes

Yes

Yes

Yes

No

No

(9.86)

0.982***

(22.09)

0.695***

(−3.71)

−0.392***

(−2.23)

−0.475**

(6.79)

1.726***

Non-Distri. (4)

898,068

Yes

Yes

No

Yes

Yes

Yes

(17.88)

1.961***

(16.38)

0.539***

(−4.70)

−1.631***

(2.67)

0.677***

Distri. (5)

Yes

Yes

No

Yes

Yes

Yes

(10.35)

1.076***

(21.73)

0.728***

(−3.62)

−0.756***

(6.21)

1.798***

Non-Distri. (6)

897,034

Yes

Yes

No

No

No

Yes

(26.94)

1.570***

(21.67)

0.456***

–

(−6.81)

−1.324***

(8.56)

1.178***

Distri. (7)

Yes

Yes

No

No

No

Yes

(12.13)

0.745***

(32.10)

0.666***

–

(−3.54)

−0.430***

(12.71)

2.012***

Non-Distri. (8)

Note Numbers in parentheses are t-values. *** (**, *) Denotes significance at the 1(5, 10)% level. The omitted group is non-FDI firms. Columns (1) and (2) are multinomial logit estimates in which the regressand in column (1) is distribution FDI whereas that in column (2) is non-distribution FDI. The regressands in columns (3) and (4) are multinomial logit estimates in which 20 three-digit CIC industries with high contract intensity a là Nunn (2007) are dropped. The regressands in columns (5) and (6) are multinomial logit estimates in which both foreign firms and pure exporters (i.e., firm exports equal firm sales) are dropped. Estimates in columns (7) and (8) drop both foreign firms and non-FDI firms that have ever merged or acquired foreign firms

Yes

Yes

No

0.967***

1.824***

Export indicator

Log firm labor (23.44)

−0.497***

−0.601***

Foreign indicator

(18.44)

(−3.13)

(−4.61)

0.698***

−0.653***

−1.588***

0.533***

(6.75)

(3.05)

SOE indicator

1.625***

0.616***

Firm relative TFP

Non-Distri. (2)

Distri. (1)

Regressand: indicator of

Table 3.6 Multinomial logit estimates of firm productivity on distribution FDI decision (2000–2008)

3.4 Extensive Margin of FDI 63

64

3 Distribution, Outward FDI, and Productivity Heterogeneity …

Second, another possible option buried in firms’ non-FDI choice is that firms contract with outside firms to undertake distribution for them. This is particularly true when firms export intermediate inputs.21 Of course, in an Antràs and Helpman (2004) type of setting like ours, such firms will be less productive than firms that undertake distribution themselves. However, our ranking of firm productivity will only hold where there are incomplete contracts in distribution that make integration an attractive option. There may be a concern about whether the ranking is still valid if some industries are more or less perfectly contractible (Feenstra & Hanson, 2005). To address this concern, we first identify the 20 most contract-intensive three-digitlevel Chinese industries strictly following Nunn (2007).22 Such industries mainly concentrate on equipment manufacturing and electronic components. By dropping from the sample industries in which firms almost always contract out distribution, the estimates in columns (3) and (4) with our restricted sample confirm that our previous findings are still strongly robust. The third caveat is the striking finding (in columns (1)–(4)) that foreign (i.e., nonChinese) invested firms are less likely to engage in FDI activity. One possible reason is that most foreign firms engage in processing trade, as found in Dai et al. (2016). Usually, processing exporters are less productive and enjoy special tariff treatment in China (Yu, 2015). Such firms do not fit with our story and need to be dropped. Since the firm-level production data do not include firms’ processing status, we instead drop from the sample pure exporters, that is, firms that sell all their products abroad, by taking advantage of the fact that processing firms have to export all their products by law. The multinomial logit estimates in columns (5)–(6) without foreign firms and pure exporters show robust evidence. Another possibility that foreign firms are less likely to engage in distribution FDI is due to the fact that foreign invested firms actually have clear foreign customers and hence they do not need foreign distribution subsidiaries. By directly dropping foreign firms, the estimates in columns (7) and (8) yield similar results. The last caveat is on mergers and acquisitions (henceforth M&A). There may be a concern that non-FDI firms may acquire a (domestic FDI or foreign) firm to use its distribution center as well. If so, our previous regressions may suffer estimation bias as even low-productive non-FDI firms can have their own foreign distribution network. However, this is not a problem if a non-FDI firm acquires a domestic FDI firm. In this case, the firm indeed has to report such an activity next year to the MOC and is classified as an FDI firm.23 By contrast, there would be some estimation bias if a non-FDI firm directly acquires foreign firms. To rule out this situation, we use the nationwide M&A data compiled by Bloomberg to identify Chinese non-FDI 21

By contrast, firms that export final goods and have no own distribution center, by default, have to find local agents to distribute their products. 22 We first make concordance between North American IO six-digit and HS eight-digit codes following another concordance between HS eight-digit and Chinese Industries Classification (CIC) three-digit codes. 23 Our FDI decision data set includes M&A activities, although it does not have a variable to stand this out.

3.4 Extensive Margin of FDI

65

manufacturing firms with complete foreign acquisition deals.24 Columns (7) and (8) drop foreign-invested firms and non-FDI firms with foreign acquisitions and still find robust results.

3.4.5 Endogeneity of Firm Productivity Table 3.4 shows that the productivity mean of firms engaging in (distribution) FDI is increasing over time, suggesting that firms may have learning effects from investing. Firms that engage in investment may be able to absorb better technology or gain managerial efficiency from host countries (Oldenski, 2012), which in turn boosts firm productivity. To exclude this effect, the sample we use only includes non-FDI firms and FDI starters, which means as long as the firm starts to invest abroad, it will no longer appear in the sample the next year. But the potential spillover effect of existing FDI firms may also lead to a possible endogeneity problem. To mitigate the endogeneity issue, we adopt an instrumental variable approach. Admittedly, it is an empirical challenge to find an ideal instrument. Here we use the lag of firms’ on-the-job training expenses as the instrument of firm productivity. The economic rationale is straightforward. As highlighted by Acemoglu and Pischke (1998) and Yeaple (2005), firms with more on-the-job training expenses usually are more productive. However, firms with more training expenses will not necessarily have more FDI. A one-year time lag is also helpful to avoid that possibility that firms’ FDI decision reversely affects last year’s on-the-job training. The simple correlation between firms’ FDI decision and firm’s lagged training expenses is close to nil (0.06), as shown in the sample. Note that we only have training data for 2004–2007. Thus, our IV estimates cover observations during 2005–2008 only. We perform IV probit estimates in column (1) in Table 3.7. In column (2) we once again use the rare-events logit estimates with endogenous TFP. This is done in two steps. In the first-stage estimation, we regress the lag of firms’ training expenses as an excluded variable on firm TFP, as well as other included variables such as indicators of SOE, foreign, exporter, and log labor. The standard errors of all the coefficients are bootstrapped with 100 replicates. The bottom module of Table 3.7 shows that the coefficient of log firm training expenses is positively correlated with firm TFP and strongly significant at the conventional statistical level. The F-statistic is greater than 10, which suggests that the IV is not weak in the statistical sense. After correcting for rare-events estimates bias, the coefficient of fitted firm TFP in the rare-events logit in column (2) is found to be much larger than the regular logit estimates, suggesting that regular binary estimates face a severe downward bias once correcting for endogeneity bias. Columns (3) and (4) report the IV multinomial logit estimation results. Once the fitted firm TFP is obtained from the first-step IV estimates, we regress the multinomial

24

We thank Cheng Chen of HKU to kindly share us with such data.

66

3 Distribution, Outward FDI, and Productivity Heterogeneity …

Table 3.7 IV estimates of firm productivity on FDI decision (2005–2008) Econometric method

Probit

Regressand: FD1 Indicator for All FD1 (1) Firm TFP SOE indicator

Rare events logit Multinomial logit All FD1 (2)

Distribution Non-distribution (3) (4)

0.854***

4.575***

1.376***

1.487***

(31.24)

(7.40)

(3.54)

(3.62)

0.149***

0.799***

−0.969***

0.160

(3.43)

(2.80)

(−2.41)

(0.62)

−0.720***

−0.467***

Foreign indicator

−0.091*** −0.524*** (−5.33)

(−6.34)

(−6.73)

(−3.83)

Log firm labor

0.094***

0.532***

0.510***

0.685***

(8.31)

(18.40)

(14.89)

(17.52)

Export indicator Year-specific fixed effects

0.260***

1.562***

1.858***

1.021***

(10.89)

(18.92)

(15.84)

(9.07)

Yes

Yes

Yes

Yes

Industry-specific fixed effects

Yes

Yes

Yes

Yes

Number of observations

484,212

484,212

484,212

484,212

IV: Log firm training expenses First-stage regression 0.028*** (30.40) Notes The regressands in columns (1), (2), and (5) are the FDI indicator whereas those in columns (2) and (6) are the distribution FDI indicator and those in columns (4) and (7) are the non-distribution FDI indicator. Numbers in parenthesis are bootstrapped t-values. *** (**) Denotes significance at the 1(5)% level. The estimates in column (1) are probit IV estimates, whereas those in columns (2) and (5) are IV rare-events logit estimates. Columns (3)–(4) are IV multinomial logit estimates. The omitted group is non-FDI firms. The instrument is the one-year lag of firm’s log training expenses from the period 2005 to 2008

logit estimates in which the regressand is one for distribution FDI and two for nondistribution FDI. Again, the coefficients of firm TFP for distribution FDI and nondistribution FDI are positive and significant. The magnitude of firm TFP for nondistribution FDI is even larger, which confirms our sorting equilibrium.

3.4.6 Discussions of Fixed Costs Ordering Our theoretical model is built on the assumptions on the ordering of fixed costs for non-FDI firms, distribution FDI firms, and non-distribution FDI firms. Although such assumptions are standard and used in other research, such as Helpman et al. (2004), it is still curious whether the ordering of various fixed costs can be validated by the data pattern. Table 3.8 picks up this task.

3.4 Extensive Margin of FDI

67

Table 3.8 Advertising expenses by different types of firms (2004–2007) Regressand: log firm advertising expenses

(1)

(2)

(3)

(4)

Non FDI indicator

−1.32***

−0.56***

−0.62***

−0.77***

(−8.33)

(−5.53)

(−4.27)

(−5.23)

Non-distribution FDI indicator

0.57**

0.32**

0.32

0.22

(2.33)

(2.14)

(1.38)

(0.97)

−0.54***

−0.53***

−0.56***

(−19.13)

(−13.69)

(−14.52)

Foreign indicator

0.22***

−

−

Log firm labor

0.70***

0.69***

0.72***

(138.07)

(88.20)

(92.69)

SOE indicator

(16.32)

Export indicator Export share ×2-digit industry fixed effects

No

−0.02

0.08***

(−1.58)

(4.75)

No

No

Yes

Year fixed effects

Yes

Yes

Yes

Yes

2-Digit industry fixed effects

Yes

Yes

Yes

Yes

Number of observations

128,028

128,026

99,679

98,424

R-squared

0.06

0.22

0.23

0.23

Notes The regressand is firm log advertising expenses. Numbers in parentheses are t-values clustered at the firm level. *** (**, *) Denotes significance at the 1 (5, 10)% level. The omitted group is distribution FDI firms. Columns (3) and (4) drop foreign firms from the sample. Column (4) includes the interaction terms between export intensity (i.e., firm exports over firm total sales) with two-digit industry fixed effects

In general, it is challenging to check directly the validity of the fixed-costs ordering, as data on the fixed costs for non-FDI firms and (non-)distribution FDI firms, to our best knowledge, are unavailable. Still, Table 3.8 attempts to offer some indirect evidence to validate the ordering assumption. As suggested by Dai et al. (2016), we use firms’ log advertising expenses to proxy for firms’ fixed costs.25 The idea is that FDI firms spend more on advertising fees to understand the environment in foreign markets and market penetration. We thus construct two indicators: (1) a non-FDI indicator that equals one if a firm has no FDI and zero otherwise, and (2) a non-distribution FDI indicator that equals one if a firm has non-distribution FDI and zero otherwise. The default omitted group is distribution FDI firms. Our underlying assumption is that distribution FDI firms have higher fixed costs than non-FDI firms but lower fixed costs than non-distribution FDI firms. If this ordering is supported by the data, it should be observed that the nonFDI indicator has a negative and significant coefficient, whereas the non-distribution FDI indicator has a positive and significant coefficient. 25

Note that the Chinese manufacturing firm-level production data set only provides firm advertising expenses during 2004–2007.

68

3 Distribution, Outward FDI, and Productivity Heterogeneity …

These outcomes are exactly what we observe in Table 3.8. The estimates in column start from a simple regression with two indicators as well as year-specific and twodigit industry fixed effects. Column (2) includes several firm-characteristic control variables to control for firm size (proxied by log firm labor), firm type of ownership (foreign firms or SOE), and firm export status. Column (3) drops foreign firms from the sample and, more importantly, includes an additional export dummy to distinguish the difference between domestic advertising and foreign advertising, as our data only report firms’ whole advertising expenses but do not report market-specific advertising expenses. It is also possible that a firm’s advertising share in foreign countries would increase with the number of countries that it served. If so, it is possible that the firm’s export intensity would increase with the number of investing destinations. We thus include a dummy for firm export intensity and its interaction with industries in column (4) and still find similar results. In any case, the anticipated signs of the non-FDI indicator and non-distribution FDI indicator strongly validate our assumption of firms fixed-cost ordering discussed in the theoretical framework.

3.5 Type-2 Tobit Estimates of Intensive Margin Thus far, we can safely conclude that high-productivity Chinese manufacturing firms are more likely to engage in distribution FDI. We now turn to explore the role of firm productivity in FDI flow. Since we only have Zhejiang province’s FDI flow data, we start by examining whether our previous findings based on nationwide FDI decision data hold for Zhejiang’s FDI manufacturing firms, as discussed carefully in Appendix C. The estimates in Appendix Table 3 and their associated discussions in Appendix E clearly suggest that all our previous findings on the extensive margin of FDI hold well for the Zhejiang subsample. To examine the intensive margin of firm productivity in FDI flow, we start from the simple OLS estimates in columns (1)–(2) in Table 3.9 by using different measures of firm productivity. We see that highly productive firms have more FDI flow regardless of the measure of firm productivity. Replacing the regressand with log FDI of distribution FDI firms yields similar results as shown in columns (3) and (4). However, there may still be a concern that the FDI decision and FDI flow are strongly correlated. To address this question, we appeal to a bivariate sample selection model, or equivalently, a Type-2 Tobit model (Cameron & Trivedi, 2005). The Type-2 Tobit specification includes: (i) an FDI participation equation where OFDID denotes distribution FDI:

0 if Uit < 0 D (3.8) OFDIit = 1 if Uit ≥ 0

Year fixed effects

Inverser mills ratio

Firm tenure

Export indicator

Log firm labor

Foreign indicator

SOE indicator

Firm TFP

Firm relative TFP

−0.600**

(−2.03)

−0.603**

(−2.04)

Yes

0.315***

(4.04)

0.314***

(4.03)

(0.39)

(0.39)

(7.08)

0.080

(7.09)

0.081

2.488***

2.492***

Yes

(−2.05)

−0.647**

(4.53)

0.339***

(0.46)

(−2.05)

−0.647**

(4.54)

0.340***

(0.46)

0.097

0.193* (1.88)

0.204*

(1.90)

0.097

0.866* (1.88)

0.922*

(1.92)

(4)

(3)

(1)

(2)

Log FDI of Dist. FDI firms Dist. FDI dummy

Log FDI of all FDI firms

Regressand

Heckman

Yes

Yes

Yes

5.207* (1.82)

7.506**

(1.33)

1.754

(2.61)

0.986***

(−1.40)

−0.643

(2.58) Yes

(1.00)

(0.39)

(5.71)

0.502***

(2.35)

0.149**

(−2.69)

−0.150***

(1.40)

1.141

−0.329 (−0.85)

(1.92)

3.429*

(8)

Log FDI

2nd-step

(4.12)

0.703***

(7)

FDI dummy

1st-step

0.003

(2.10)

2.987**

(3.44)

1.162***

(−2.15)

−1.158**

(7.77)

2.689***

(2.66)

4.968***

(6)

Log Dist. FDI

2nd-step

0.002

(5.45)

0.527***

(2.03)

0.133**

(−3.06)

−0.185***

(4.01)

0.720***

(5)

1st-step

OLS

Econometric method

Table 3.9 Intensive estimates of distribution FDI versus total FDI flow for Zhejiang firms (2006–2008)

(continued)

Yes

(2.79)

7.372***

(2.15)

2.645**

(3.58)

1.294***

(−2.00)

−0.880**

(2.86)

4.751***

(9)

Log Dist. FDI

3.5 Type-2 Tobit Estimates of Intensive Margin 69

0.13

First-stage estimates

R-squared

IV: log firm training –

0.15

0.15 –

210

210

Yes

(21.86)

0.049***

–

60,198

Yes

(5)

–

0.16

251

Yes

(6)

Log Dist. FDI

2nd-step

(22.29)

0.011***

–

60,358

Yes

(7)

FDI dummy

1st-step

–

0.15

251

Yes

(8)

Log FDI

2nd-step

–

0.18

210

Yes

(9)

Log Dist. FDI

Note Numbers in parentheses are (bootstrapped) t-values. *** (**, *) Denotes significance at the 1% (5%, 10%) level. The regressand in the OLS estimates in columns (1)–(2) are total FDI flow whereas those in columns (3)–(4) are distribution FDI flow. The rest of the table is the three-step Heckman IV estimates: (i) The first stage regresses firm TFP on the one-year lag of firm’s log training expenses. (ii) The second stage is the probit estimates, which regress the firm distribution FDI indicator on the fitted variables of TFP obtained from (i). Firm tenure serves as the exclusive variable, which is only included in the second-stage but not in the third-stage. (iii) The third-stage estimates in column (6) regress log distribution FDI on fitted TFP and other control variables. The inverse Mills ratio calculated from the second-stage is inserted in the third stage as an additional regressor. The estimates in columns (7)–(9) are similar to the estimates in columns (5) and (6) except the regressand of the second stage is firm the FDI indicator and that of the third stage is firm’s log FDI flow in column (8) and firm’s log distribution FDI flow in column (9)

–

0.13

251

Number of observations

–

251

Yes

Industry fixed effects

(4)

(3)

(1)

(2)

Log FDI of Dist. FDI firms Dist. FDI dummy

Log FDI of all FDI firms

Regressand

Heckman 1st-step

OLS

Econometric method

Table 3.9 (continued)

70 3 Distribution, Outward FDI, and Productivity Heterogeneity …

3.6 Investment Destination

71

where U it denotes a latent variable faced by firm i; and (ii) an “outcome” equation whereby the firm’s distribution FDI flow is modeled as a linear function of other variables. In particular, we use a logit model to estimate the following selection equation:   D = 1 = Pr U ≥ 0 Pr OOFDIit it   = ⋀ γ0 + γ1 ln TFPit + γ2 SOEit + γ3 FIEit + γ4 FXit + γ5 ln L it + γ6 Tenure it + ξ j + λt

(3.9)

where ⋀(·) is the logistic CDF. In addition to the logarithm of firm productivity, a firm’s FDI decision is also affected by other factors, such as the firm’s ownership (whether it is an SOE or a foreign firm), export status (FX equals one if a firm exports and zero otherwise), and size (measured by the logarithm of the number of employees). Our estimations here include three steps. Because FDI firms may improve their productivity via investment abroad, in the first step, firm TFP is instrumented by the lag of log training expenses, as introduced above.26 In the second step, our Type-2 Tobit model requires an excluded variable that affects the firm’s FDI decision but does not affect its FDI flow. Here the firm’s tenure (Tenureit ) serves this purpose, since the literature finds that a firm’s tenure is highly correlated with the firm’s export decision (Amiti & Davis, 2012). It was shown in our previous estimates that the export decision and the FDI decision are highly correlated. By contrast, the simple correlation between FDI flow and export status is close to nil (0.07), which confirms that tenure can serve as an excluded variable in the third-step Heckman estimates. For the third step, we include the two-digit CIC industrial dummies ξ j and year dummies λt to control for other unspecified factors. Table 3.9 reports the estimation results for the bivariate sample selection model. As shown in column (5), high-productivity firms are more likely to engage in distribution FDI. We then include the computed inverse Mills ratio obtained in the third-step Heckman estimates in column (6) with the log distribution of FDI flow. The positive and significant coefficient of firm TFP suggests that high-productivity firms have more distribution FDI. Finally, columns (7)–(9) perform another robustness check of the Heckman estimates in which the regressand in the first step is the indicator of total FDI and that in the second step is log total FDI flow in column (8) and log distribution FDI flow in column (9). It turns out that our previous findings are not changed at all in such robustness checks.

3.6 Investment Destination Thus far, we have found evidence that high-productivity firms are more likely to invest abroad. Once a firm invests, the higher is its productivity, the more the firm 26

Note that standard errors in Table 3.9 are bootstrapped with 100 replicates.

72

3 Distribution, Outward FDI, and Productivity Heterogeneity …

invests abroad. The firm’s investment decision follows the sorting behavior predicted by Proposition 4.1. High-productivity firms engage in distribution FDI and even higher productive firms participate in non-distribution FDI. As argued before, the importance of distribution FDI is that it can reduce the cross-border communications costs of exporting firms for service and distribution overseas. We now check whether the investment environment and income in the destination country affect the firm’s distribution FDI decision.

3.6.1 Communication Costs in Destination Markets Proposition 4.2 of our theoretical model states that an increase in cross-border communications costs (iceberg transport costs) would increase the probability of distribution (non-distribution) foreign investment. We now turn to examine whether this theoretical prediction is supported by the data Table 3.10. To measure cross-border communications costs, we use data from the World Bank’s Doing Business project. We first use the host country’s days of import document preparation as a proxy for cross-border communications costs. It is important to stress that these import costs are destination-country-specific, independent of industries (or firms), but depend on the import volume. For each unit of a given product exported to a given country, such costs are roughly the same across different exporting firms, regardless of firm productivity. And such costs for distribution FDI are much lower than for non-FDI firms. These features are consistent with the characteristics of cross-border communications costs sketched in our theoretical model. To make a further distinction between communications costs and transportation costs, we include the destination country’s simple average import tariffs as a proxy for transportation costs. In addition, we control for log bilateral distance. These data are all publicly available from the World Bank.27 Table 3.10 is our second key table. Columns (1) and (2) present the multinomial logit estimates; the regressand in column (1) is distribution FDI and that in column (2) is non-distribution FDI. Several interesting findings merit special attention. First, the coefficients of firm relative productivity in columns (1) and (2) are all positive and significant. The magnitude of firm TFP in column (2) is higher than its counterpart in column (1). These findings are similar to our above findings and consistent with our theoretical predictions. Second and more importantly, the coefficient of days of import document preparation in column (1) is positive and significant, whereas its counterpart in column (2) is insignificant, indicating that an increase in cross-border communications costs raises the probability of distribution FDI but not necessary that of non-distribution 27

Note that, in all regressions in Tables 3.10, 3.11 and 3.12, we drop all tax-haven destinations, such as Hong Kong and Virgin Islands, from the sample, as Chinese FDI firms usually do not really invest in such regions but only use them as exprot instead. Similarly, it is very likely that firms will switch their FDI type from distribution FDI this year to non-distribution FDI next year, as shown in Appendix Table 3.

3.6 Investment Destination

73

Table 3.10 Multinomial logit estimates by host investment environment (2006–2008) Regressand: indicator of

Dist. (1)

Firm relative TFP

3.41***

4.84**

3.41***

4.84**

4.24***

4.38*

(3.17)

(2.08)

(3.17)

(2.08)

(3.02)

(1.94)

0.26*

−0.19

0.26*

0.03

0.30*

0.05

(1.76)

(−0.83)

(1.70)

(0.12)

(1.65)

(0.20)

Days of import document preparation Import tariffs SOE indicator

Non-Dist. Dist. (2) (3)

Non-Dist. (4)

Dist. (5)

Non-Dist. (6)

0.08

0.13**

0.07

0.17**

0.04

0.17*

(1.51)

(2.04)

(1.43)

(2.18)

(0.66)

(1.95)

−10.85***

3.22***

−14.92***

2.82**

−10.86***

3.21***

(−72.12)

(3.00)

(−75.15)

(2.44)

(−68.68)

(2.94)

0.43

0.52*

0.39

–

–

Foreign indicator 0.60** (2.24)

(0.74)

(1.94)

(0.66)

0.97***

0.33

0.97***

0.65

0.81**

0.60

(3.36)

(0.92)

(3.37)

(1.56)

(2.48)

(1.19)

Foreign firms dropped

No

No

No

No

Yes

Yes

Log bilateral distance controlled

Yes

Yes

Yes

Yes

Yes

Yes

Year-specific fixed effects

Yes

Yes

Yes

Yes

Yes

Yes

Year dummy x log bilateral distance

No

No

Yes

Yes

Yes

Yes

Industry-specific fixed effects

Yes

Yes

Yes

Yes

Yes

Yes

Number of observations

1,999,441

Log per-capita GDP

1,999,441

1,679,207

Note Numbers in parentheses are (bootstrapped) t-values. *** (**, *) Denotes significance at the 1% (5%, 10%) level. The regressands in columns (1), (3), and (5) are the indicator of distribution FDI, whereas those in columns (2), (4), and (6) are the indicator of non-distribution FDI. The omitted group is non-FDI firms

FDI, since higher cross-border communications costs attract more exporting firms to establish a foreign business office to reduce such costs, exactly as predicted by our theoretical model. Third, our theoretical model also predicts that an increase in iceberg transportation costs would increase the probability of firms engaging in non-distribution FDI but is ambiguous on the probability of firms participating in distribution FDI, since distribution FDI does not reduce iceberg transportation costs as long as the firm exports. If this prediction is supported by the data, the iceberg transportation costs variable should exhibit a positive coefficient in column (2). We hence use the import

74

3 Distribution, Outward FDI, and Productivity Heterogeneity …

Table 3.11 Multinomial logit estimates of the distribution FDI decision by destination income (2006–2008) Predetermined income threshold regressand: distribution FDI decision

GDPPC = $3,885

GDPPC = $10,000

Dist. FDI to poor

Dist. FDI to rich

Dist. FDI to poor

Dist. FDI to rich

(1)

(2)

(3)

(4)

Firm relative TFP 0.058 (0.04)

1.371***

0.377

1.420***

(3.06)

(0.36)

(3.03)

0.500***

0.656***

0.668***

0.637***

(3.05)

(10.44)

(5.49)

(9.46)

Export indicator

1.569***

2.302***

1.974***

2.254***

(3.05)

(8.47)

(4.28)

(8.00)

SOE indicator

−15.33***

−15.67***

−17.70***

−17.44***

(−25.21)

(−55.99)

(−35.42)

(−61.60)

Log of labor

−0.773

−0.177

−1.016**

−0.086

(−1.25)

(−1.04)

(−2.12)

(−0.49)

Year-specific fixed effects

Yes

Yes

Yes

Yes

Industry-specific fixed effects

Yes

Yes

Yes

Yes

Number of observations

100,806

Foreign indicator

100,806

Note The regress and is the distribution FDI indicator in which zero denotes not any type of FDI, one refers to distribution FDI to poor countries, and two denotes distribution FDI to rich countries. Numbers in parentheses are t-values. *** (**) Denotes significance at the 1% (5%) level. The sample in this table covers Zhejiang manufacturing firms during 2006–2008. Two-digit Chinese industry classification industry-specific fixed effects are included in all the estimations

country’s simple-average tariffs as a proxy for iceberg transportation costs. The coefficient of import tariffs has a positive and significant sign in column (2) in Table 3.10. There may be curiosity about whether these results are driven by the income level of the destination country, as high-income countries usually have more transparent and efficient customs processes. And the probability of firms engaging in outward FDI would decrease as countries are further apart. We hence include per capita gross domestic product (GDP) and log bilateral distance as control variables in the multinomial logit estimates in all estimates. To control for other unspecified factors, in addition to the standard year-specific fixed effects and two-digit industry-specific fixed effects, we include the interaction between the year dummy and log bilateral distance, given that bilateral distance is time-invariant, in columns (3) and (4) in Table 3.10. Finally, as a robustness check, the last two columns in Table 3.10 drop China’s multinational firms (i.e., firms with foreign indicator equal to one). The results are similar to those found earlier.

3.6 Investment Destination

75

Table 3.12 Threshold estimates by income of host countries (2006–2008) Estimated threshold

Without threshold

Log GDPPC = 10.726 Regressand: firm log FDI (1) flow Firm TFP Constant

(2)

With threshold Low

High

Low

High

(3)

(4)

(5)

(6)

0.209*** 0.254*** 0.067

0.395*** 0.107

0.460***

(2.29)

(2.43)

(2.68)

(2.61)

(0.67)

(1.00)

2.500*** 2.303*** 2.945*** 1.977*** 2.767*** 1.588*** (6.37)

(5.24)

(6.69)

(2.78)

(5.50)

(2.04)

Year fixed effects

No

Yes

No

No

Yes

Yes

Industry fixed effects

No

Yes

No

No

Yes

Yes

Number of observations

251

251

165

86

165

86

(Joint) R-squared

0.023

0.038

0.061

0.082

Note The regressand is firm log FDI flow. Numbers in parentheses are t-values. *** Denotes significance at the 1% level. Estimates in this table are threshold estimates a la Hanson (2000) by using FDI destination income as the threshold. The estimates in columns (1) and (2) are standard OLS estimates without considering the heteroskedasticity of the threshold. Columns (4) to (6) are estimated by using the estimated threshold (log per capita GDP is 10.726). Joint R-squares are reported in columns (4) to (6). Columns (5) and (6) include CIC two-digit industry-level fixed effects and year-specific fixed effects. The 95% confidence interval estimates for each variable are not reported to save space, although they are available upon request

We now discuss the economic magnitude of our findings in Table 3.10. As shown in columns (5) and (6), a 1 percentage point increase in firm relative TFP increases by 4.2% the probability of distribution FDI rather than other modes. By contrast, a 1 percentage point increase in firm relative TFP increases by 4.4% the probability of non-distribution FDI. Similarly, one more day spent on import documentation increases the probability of distribution FDI by 0.3%. Finally, a 1 percentage point increase in the destination’s import tariffs increases the probability of non-distribution FDI by 0.25%.

3.6.2 Investment Decision by Destination Income As seen in Table 3.10, destination countries’ income plays an important role in FDI decisions. It is worthwhile to take a step forward to consider whether firm productivity matters for host countries’ income. Interestingly, the literature offers divergent answers to this question. Head and Ries (2003) use Japanese data and find that firms investing in poor countries have even lower productivity than do non-FDI firms. However, studies like Damijan et al. (2007) find that the income level of the host country has no significant effect on Slovenian firms’ FDI decision. We consider a multinomial logit model to explore the role of firm productivity in the decision to engage in FDI in different income destinations. We first split our

76

3 Distribution, Outward FDI, and Productivity Heterogeneity …

Zhejiang sample into two groups, low-income destination countries and high-income destination countries, by using a predetermined income threshold suggested by the World Bank. The base category of our multinomial logit regression is non-FDI firms, so the probability that firm i chooses to invest in country j (poor or rich) is as follows28 :   Pr O F D IitD = j|Xit ⎧ 1 ⎪ ( j is without FDI ) 3 ⎪  ⎪ ⎨ 1+ exp(Xit βk ) k=2 = exp(xit β j ) ⎪ ( j is distribution FDI to poor or to rich countries ) ⎪ 3 ⎪ ⎩ 1+  exp(xit βk )

(3.10)

k=2

where the regressors Xit include firm productivity and other control variables, such as export status, firm size (i.e., log firm labor) and firm ownership. We start our regressions with a predetermined income threshold in Table 3.11. According to the World Bank’s classifications in 2008, a country with per capita GDP less than $3,855 is classified as a lower-middle-income country, whereas a country with per capita GDP greater than $10,000 is classified as a high-income country. We hence start our multinomial logit estimates by defining FDI destination countries with income less than $3,855 as poor countries. After controlling for year-specific fixed effects and industry-specific fixed effects, the coefficient of firm productivity is positive and statistically significant for firms investing in rich countries and negative and insignificant for firms investing in poor countries. These findings hold when we increase the income threshold to $10,000, as shown in columns (3) and (4) in Table 3.11. The economic rationale is straightforward. Chinese FDI firms have different motivations for their FDI behavior. Some firms seek foreign markets, whereas some seek foreign sourcing for natural resources (Huang & Wang, 2011). As high-income foreign markets are usually highly competitive, only high-productivity firms are able to invest and seek local markets there. By contrast, firms investing in poor destinations are not mainly seeking foreign markets; instead, the firms may be interested in natural resources or cheaper labor in the host countries. The latter is especially true for firms in labor-intensive industries, such as textiles and garments. For instance, Chinese FDI firms that invest in Africa (e.g., Ethiopia and Madagascar) mostly are low-productivity firms in labor-intensive industries. The economic rationale is straightforward. Chinese FDI firms have different motivations for their FDI behavior. Some firms seek foreign markets, whereas some seek foreign sourcing for natural resources (Huang & Wang, 2011). As high-income foreign markets are usually highly competitive, only high-productivity firms are able to invest and seek local markets there. By contrast, firms investing in poor destinations are not mainly seeking foreign markets; instead, the firms may be interested in natural resources or cheaper labor in the host countries. The latter is especially true 28

Note that only two firms invest in both rich countries and poor countries. For simplicity we drop those two firms from our sample.

3.6 Investment Destination

77

for firms in labor-intensive industries, such as textiles and garments. For instance, Chinese FDI firms that invest in Africa (e.g., Ethiopia and Madagascar) mostly are low-productivity firms in labor-intensive industries.

3.6.3 Threshold Estimates of the Linder Hypothesis Beyond the conventional wisdom that FDI is determined by a proximityconcentration trade-off, Fajgelbaum et al. (2015) argue that the per capita income of host countries is positively correlated with market size. With trade costs, firms are more likely to engage in FDI rather than export when host markets are large (Markusen, 1984). Thus, the standard Linder hypothesis, which stresses that highincome countries have relatively large trade volume, applies to firm FDI behavior: high-income countries usually absorb more FDI. Our final exercise is to see whether the Linder hypothesis for FDI works in China. The first necessary step to perform this task is to classify destination country groups by income. A common and simple way to do this is to use the World Bank’s classification, as in Table 3.10. However, this classification suffers from two pitfalls. First, the threshold varies by year. There are no clear and time-invariant cutoffs for the income groups. Second, even if the cutoffs are fixed, the effect of firm productivity on firm FDI may not exactly correspond to the predetermined income cutoffs. That is, host countries’ per capita GDP is an endogenous threshold for FDI firms in response to productivity movement. To overcome these empirical challenges, Hansen (1999, 2000) provides an econometric approach that considers endogenous threshold regressions. To motivate this, consider an empirical specification with a country’s per capita GDP (pcgdp) as a threshold variable:

O F D Iit = β X it + ∈it if pcgdp pit < T (3.11) O F D Iit = θ X it + ∈it if pcgdp it ≥ T where T is the threshold parameter to be estimated. OFDI ,= is firm i’s FDI flow Without loss of in year t. X it refers to all regressors, including firm productivity.  generality, ∊it is i.i.d with normal distribution: ∈it ∼ N 0, σi2 . By using an indicator function I(·), we can re-express Eq. (3.11) as:     OFDIit = βXit · I pcgd pit < T + θ Xit · I pcgdpit ≥ T + ∈it

(3.12)

As this is a nonlinear regression, we can use the nonlinear squares approach to minimize the sum of squared residuals. Since the estimators also include the threshold parameter Tˆ , the most convenient computational method to obtain the linear squares estimate is via concentration. Thus, the optimal threshold parameter Tˆ is chosen to minimize the concentrated sum of squared errors function so that

78

3 Distribution, Outward FDI, and Productivity Heterogeneity …

Tˆ = arg min SS(β(T ), θ (T ), T ). Based on this, Hansen (1999, 2000) developed an asymptotic distribution theory for the threshold regression estimates. Table 3.12 presents the threshold regression results. By comparison, we start from a regression without considering the threshold effect in columns (1) and (2). By abstracting away all other variables, we see that firm TFP is positively correlated to firm log FDI flow, as shown in column (1). The specification in column (2) yields similar results by controlling industry-specific fixed effects and year-specific fixed effects. The threshold regression results are reported in columns (3) and (4). The estimated threshold parameter of host countries’ log per capita GDP is 10.73 (or equivalently, per capita GDP is $45,524). As before, the coefficient of firm productivity is positive and statistically significant for high-income FDI destinations. However, for low-income host countries, where per capita GDP is lower than the estimated threshold, the effect of firm productivity on firm FDI flow is statistically insignificant, suggesting that firm productivity is not a crucial determinant of firm FDI flow. This finding is robust even when we control for year-specific fixed effects and industryspecific fixed effects in columns (5) and (6), suggesting that the Linder hypothesis for FDI volume to high-income destination countries holds but may not exist for FDI volume to low-income countries. This result confirms that Chinese firms’ investment in poor countries may not be labeled as “horizontal” FDI: tariff-jumping motivation or seeking foreign markets may not be a top priority for these firms (Kolstad & Wilg, 2012).

3.7 Concluding Remarks This chapter is one of the first to explore how firm heterogeneity influences the volume of Chinese distribution FDI. The rich data set enabled us to examine not only the extensive margin, but also the intensive margin of Chinese outward FDI. We found that firms with distribution FDI are more productive than non-FDI firms, but less productive than non-distribution FDI firms. These findings reflect the fact that many Chinese exporters are insufficiently productive to build up production lines in foreign markets. As a compromise, such firms set up foreign distribution centers to promote their exports. We also found that with higher cross-border communications costs (iceberg transport costs), there is a higher probability that firms engage in distribution (non-distribution) FDI. Finally, high-productivity firms invest more in high-income countries but not necessarily in low-income countries.

Appendix: Supplementary Material Supplementary data associated with this chapter can be found, in the online version, at https://doi.org/10.1016/j.intfin.2020.101218.

References

79

References Acemoglu, D., & Pischke, J. S. (1998). Why do firms train? Theory and evidence. The Quarterly Journal of Economics, 113(1), 79–119. Ahn, J., Khandelwal, A. K., & Wei, S. J. (2011). The role of intermediaries in facilitating trade. Journal of International Economics, 84(1), 73–85. Amiti, M., & Davis, D. R. (2012). Trade, firms, and wages: Theory and evidence. The Review of Economic Studies, 79(1), 1–36. Antras, P., & Helpman, E. (2004). Global sourcing. Journal of Political Economy, 112(3), 552–580. Berman, N., Martin, P., & Mayer, T. (2012). How do different exporters react to exchange rate changes? The Quarterly Journal of Economics, 127(1), 437–492. Boatman, K. T. (2007). What determines distribution FDI. Journal of Business and Public Affairs, 1(1). Brandt, L., Van Biesebroeck, J., & Zhang, Y. (2012). Creative accounting or creative destruction? Firm-level productivity growth in Chinese manufacturing. Journal of Development Economics, 97(2), 339–351. Cameron, A. C., & Trivedi, P. K. (2005). Microeconometrics: Methods and applications. Cambridge University Press. Chen, Q. (2015). Climate shocks, state capacity and peasant uprisings in North China during 25– 1911 CE. Economica, 82(326), 295–318. Chen, W., & Tang, H. (2014). The dragon is flying west: Micro-level evidence of Chinese outward direct investment. Asian Development Review, 31(2), 109–140. Chen, C., Tian, W., & Yu, M. (2019). Outward FDI and domestic input distortions: Evidence from Chinese firms. The Economic Journal, 129(624), 3025–3057. Dai, M., Maitra, M., & Yu, M. (2016). Unexceptional exporter performance in China? The role of processing trade. Journal of Development Economics, 121, 177–189. Damijan, J. P., Polanec, S., & Prašnikar, J. (2007). Outward FDI and productivity: Micro-evidence from Slovenia. World Economy, 30(1), 135–155. Eaton, J., Kortum, S., & Kramarz, F. (2011). An anatomy of international trade: Evidence from French firms. Econometrica, 79(5), 1453–1498. Fajgelbaum, P., Grossman, G. M., & Helpman, E. (2015). A Linder hypothesis for foreign direct investment. The Review of Economic Studies, 82(1), 83–121. Feenstra, R. C., & Hanson, G. H. (2005). Ownership and control in outsourcing to China: Estimating the property-rights theory of the firm. The Quarterly Journal of Economics, 120(2), 729–761. Feenstra, R. C., Li, Z., & Yu, M. (2014). Exports and credit constraints under incomplete information: Theory and evidence from China. Review of Economics and Statistics, 96(4), 729–744. Hansen, B. E. (1999). Threshold effects in non-dynamic panels: Estimation, testing, and inference. Journal of Econometrics, 93(2), 345–368. Hansen, B. E. (2000). Sample splitting and threshold estimation. Econometrica, 68(3), 575–603. Hanson, G. H., Mataloni, R. J., Jr., & Slaughter, M. J. (2001). Expansion strategies of US multinational firms. Head, K., & Ries, J. (2003). Heterogeneity and the FDI versus export decision of Japanese manufacturers. Journal of the Japanese and International Economies, 17(4), 448–467. Helpman, E., Melitz, M. J., & Yeaple, S. R. (2004). Export versus FDI with heterogeneous firms. American Economic Review, 94(1), 300–316. Horstmann, I. J., & Markusen, J. R. (1995), Exploring new markets: Direct investment, contractual relations and the multinational enterprise. International Economic Review, 37(1), 1–19. Hsieh, C. T., & Klenow, P. J. (2009). Misallocation and manufacturing TFP in China and India. The Quarterly Journal of Economics, 124(4), 1403–1448. Huang, Y., & Wang, B. (2011). Chinese outward direct investment: Is there a China model? China and World Economy, 19(4), 1–21. Imbens, G. W. (2004). Nonparametric estimation of average treatment effects under exogeneity: A review. Review of Economics and Statistics, 86(1), 4–29.

80

3 Distribution, Outward FDI, and Productivity Heterogeneity …

Ishikawa, J., Morita, H., & Mukunoki, H. (2010). FDI in post-production services and product market competition. Journal of International Economics, 82(1), 73–84. Kimura, F., & Lee, H. H. (2006). The gravity equation in international trade in services. Review of World Economics, 142(1), 92–121. King, G., & Zeng, L. (2001). Logistic regression in rare events data. Political Analysis, 9(2), 137– 163. Kolstad, I., & Wiig, A. (2012). What determines Chinese outward FDI? Journal of World Business, 47(1), 26–34. Lu, D. (2010). Exceptional exporter performance? University of Chicago. Markusen, J. R. (1984). Multinationals, multi-plant economies, and the gains from trade. Journal of International Economics, 16(3–4), 205–226. Melitz, M. J. (2003). The impact of trade on intra-industry reallocations and aggregate industry productivity. Econometrica, 71(6), 1695–1725. Nunn, N. (2007). Relationship-specificity, incomplete contracts, and the pattern of trade. The Quarterly Journal of Economics, 122(2), 569–600. Oldenski, L. (2012). Export versus FDI and the communication of complex information. Journal of International Economics, 87(2), 312–322. Qiu, L. D., & Xue, Y. (2014). Understanding China’s foreign trade: A literature review (I). China Economic Journal, 7(2), 168–186. Ramondo, N., Rappoport, V., & Ruhl, K. J. (2013). The proximity-concentration tradeoff under uncertainty. Review of Economic Studies, 80(4), 1582–1621. Russ, K. N. (2007). The endogeneity of the exchange rate as a determinant of FDI: A model of entry and multinational firms. Journal of International Economics, 71(2), 344–372. Wang, C., Hong, J., Kafouros, M., & Boateng, A. (2012). What drives outward FDI of Chinese firms? Testing the explanatory power of three theoretical frameworks. International Business Review, 21(3), 425–438. Yeaple, S. R. (2005). A simple model of firm heterogeneity, international trade, and wages. Journal of International Economics, 65(1), 1–20. Yu, M. (2015). Processing trade, tariff reductions and firm productivity: Evidence from Chinese firms. The Economic Journal, 125(585), 943–988.

Chapter 4

Outward FDI and Domestic Input Distortions: Evidence from Chinese Firms

Abstract We examine how domestic distortions affect firms’ production strategies abroad by documenting two puzzling findings using Chinese firm-level data of manufacturing firms. First, private multinational corporations (MNCs) are less productive than state-owned MNCs, but they are more productive than state-owned enterprises overall. Second, there are disproportionately fewer state-owned MNCs than private MNCs. We build a model to rationalize these findings by showing that discrimination against private firms domestically incentivizes them to produce abroad. The model shows that selection reversal is more pronounced in industries with more severe discrimination against private firms, which receives empirical support.

4.1 Motivation and Findings Foreign direct investment (FDI) and the emergence of multinational corporations (MNCs) are dominant features of the world economy.1 Therefore, understanding the behaviour of MNCs and patterns of FDI is important for the analysis of the aggregate productivity and resource allocation. The sharp increase in outward FDI from developing countries in the past decade has been phenomenal, and this is especially true for China. The United Nations Conference on Trade and Development (UNCTAD) World Investment Report (UNCTAD, 2015) shows that outward FDI flows from developing economies have already accounted for more than 33% of overall FDI flows, up from 13% in 2007.

1

MNCs refer to firms that own or control production of goods or services in countries other than their home country. FDI includes mergers and acquisitions, building new facilities, reinvesting profits earned from overseas operations and intra-company loans. This chapter is published in The Economic Journal by Chen Cheng, Wei Tian and Miaojie Yu. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022 W. Tian and M. Yu, Outward Foreign Direct Investment of Chinese Enterprises, Contributions to Economics, https://doi.org/10.1007/978-981-19-4719-3_4

81

82

4 Outward FDI and Domestic Input Distortions …

Furthermore, despite the fact that global FDI flows plummeted by 16% in 2014, MNCs from developing economies invested almost US$468 billion abroad in 2014, an increase of 23% over the previous year.2 As the largest developing country in the world, China has seen an astonishing increase in its outward FDI flows in the past decade. In 2015, China’s outward FDI reached the level of 9.9% of the world’s total FDI flows, which made China the second-largest home country of FDI outflows globally. In addition, manufacturing outward FDI from China is becoming more important in China’s total outward FDI flows. Its share in China’s total outward FDI has increased from 9.9% in 2012 to 18.3% in 2016.3 In sum, patterns of China’s manufacturing outward FDI flows need to be explored, given its importance for the world economy. This study investigates the patterns of China’s outward FDI of manufacturing firms through the lens of domestic input market distortions. It has been documented that discrimination against private firms is a fundamental issue for the Chinese economy. For instance, state-owned enter-prises (SOEs) enjoy preferential access to financing from state-owned banks (Dollar & Wei, 2007; Khandelwal et al., 2013; Manova et al., 2015; Song et al., 2011). Moreover, Khandelwal et al. (2013), and Bai et al. (2017) document that private firms had been treated unequally by the Chinese government in the exporting market. In short, it is natural to link the behaviour of Chinese MNCs to domestic distortions in China. To the best of our knowledge, little work has studied how institutional distortions at home affect firms’ investment patterns abroad. The reason is that developed economies have been the home countries of outward FDI for many decades, and their economies are much less likely to be subject to distortions compared with developing economies. By contrast, various distortions are fundamental features of developing countries. For instance, size-dependent policies and red tape have been shown to generate substantial impacts on firm growth and resource allocation in India (Hsieh & Klenow, 2009, 2014). The government discriminates against private firms in China (Brandt et al., 2013; Huang, 2003, 2008). Moreover, there is already anecdotal evidence documenting how private firms in China circumvent these distortions by doing business abroad. For instance, the key to the success of the Geely automobile company (a private car maker in China) was to expand internationally even at early stages of its development (e.g., the purchase of Volvo in 2010). Thus, distortions in the domestic market do seem to affect firms’ decisions concerning going abroad. We document three stylised facts of China’s MNCs in manufacturing sectors to motivate our theory. First, although private non-MNCs are more productive than stateowned non-MNCs on average, private MNCs are actually less productive than stateowned MNCs on average. Moreover, when looking at the productivity distribution of state-owned MNCs and private MNCs, we find that at each percentile, stateowned MNCs have higher (normalised) TFP compared with private MNCs. Second, 2

The UNCTAD World Investment Report also demonstrates that FDI stock from developing economies to other developing economies grew by two-thirds, from US$1.7 trillion in 2009 to US$2.9 trillion in 2013. It also reports that transition economies now represent nine of the 20 largest investor economies globally (UNCTAD, 2015). 3 See Statistical Bulletin of China’s Outward Foreign Direct Investment (2015 and 2016).

4.1 Motivation and Findings

83

compared with private firms, the fraction of firms that undertake outward FDI is smaller among SOEs. Finally, the relative size of MNCs (i.e., average size of MNCs divided by average size of non-exporting firms) is smaller among private firms than among SOEs. These findings are counterintuitive. First, SOEs are much larger than private firms in China, and larger firms are more likely to become MNCs. Furthermore, it has been documented that SOEs receive substantial support from the Chinese government for investing abroad. Thus, why did so few SOEs actually invest abroad? Finally, it has been documented that SOEs are less productive than private firms in China (e.g., Brandt et al., 2012; Khandelwal et al., 2013). Our data also show this pattern when we look at non-exporting and exporting (but non-multinational) firms. Why is this pattern reversed when we focus on MNCs? To rationalise these puzzling findings, we build a model based on Helpman et al. (2004; henceforth, HMY) and highlight two economic forces: institutional arbitrage and selection reversal. Two key departures we make from HMY are the addition of capital (or land) used in the production process and asymmetric distortions across borders. Specifically, we assume that private firms pay a higher capital rental price (and land price) when producing domestically (compared with SOEs), while all firms pay the same input prices when they produce abroad. The existence of the input price wedge comes from the capital and land markets, as the banking sector is dominated by state-owned banks and land is largely owned by the government in China. In our data, private firms pay higher interest rates and unit land price than SOEs, which is equivalent to an implicit input tax levied on private firms. When firms produce abroad, at least a part of the input price wedge ceases to exist, as the capital and land markets in foreign economies are not controlled by the Chinese government. In other words, the domestic input price (relative to the foreign input price) private firms face is higher than that of SOEs.4 As a result of this asymmetry, there is an extra incentive for private firms to produce abroad, since they can circumvent the input market distortion that exists only domestically by becoming MNCs (i.e., institutional arbitrage). Absent the domestic distortion, there would be no difference in the selection into the (domestic and) FDI market, since SOEs and private firms face the same domestic (and foreign) market environment. When there is a domestic distortion, selection into the domestic market is tougher for private firms. However, since they receive an extra benefit from producing abroad (i.e., not just the saving on the variable trade cost), the incentive of becoming an MNC is higher for them. This leads to less tough selection into the FDI market for private firms, which is termed’selection reversal’ in this chapter. This reversal rationalises why there are disproportionately fewer MNCs among SOEs than among private firms and why private MNCs are less productive than state-owned MNCs. 4

It is plausible that the distortion in the input market shows up as a subsidy to SOEs. In this scenario, SOEs have less of an incentive to undertake FDI, since the relative domestic input price they face is lower, which is the same as in our main model. This situation results in tougher selection into the FDI market for SOEs as well, which leads to the same empirical predictions.

84

4 Outward FDI and Domestic Input Distortions …

In addition to explaining the stylised facts, our model yields several additional empirical predictions. First, conditional on other firm-level characteristics, a private firm sells disproportionately more in the foreign market because of the non-existence of distortion abroad. Second, as the distortion exists in the capital and land markets (rather than in the labour market), the selection reversal for state-owned MNCs is more pronounced in capital intensive industries and in industries in which the (industry-level) interest rate (or land price) differential between private firms and SOEs is larger. We present supporting evidence for these additional predictions. It might be true that Chinese firms can borrow money from domestic banks to finance a part of their outward FDI projects. Thus, the discrimination against private firms in the credit market might still exist even when private firms invest abroad. However, even if this is true for a fraction of firms in our study, the discrimination against private firms in the land market is limited to the domestic market as firms cannot move land abroad to do investment. Importantly, we find evidence that the selection reversal is more pronounced in industries in which the unit land price differential (between private firms and SOEs) is larger. Therefore, the asymmetric distortion in the land market across border plays a role in affecting Chinese firms’ FDI decisions. The data set of outward FDI used in this chapter is a representative sample of China’s outward FDI projects, as the ministry of commerce of China requires all outward FDI deals whose investment amounts are higher than US$10 million to be reported to the ministry. Admittedly, our data set loses small outward FDI projects which are most likely to be conducted by private firms.5 Naturally, the exclusion of small private outward FDI deals would prevent us from finding the selection reversal. Given that we do find the selection reversal, this pattern should be more pronounced if we used an universal data set which includes small outward FDI deals. It is important to stress that Chinese firms have different motives to undertake outward FDI. In this chapter, we focus on manufacturing FDI and exclude outward FDI projects in the construction and mining sectors for two reasons. First, manufacturing firms’ investment behaviour is more related to firm performance and profitdriven.6 Second, the canonical model of FDI (i.e., HMY) and asymmetric distortions across border fit well into the case of manufacturing MNCs from China. In particular, the share of manufacturing FDI in total outward FDI is much larger within China’s investment into developed economies than within its investment into developing economies.7 As developed economies probably have fewer distortions than developing economies, our story fits better into manufacturing MNCs from China. We also use several sub-samples of our data sets to exclude alternative hypotheses for the selection reversal pattern. First, we find that the selection reversal does not 5

Shen and Chen et al. draw the same conclusion that outward FDI projects conducted by private firms are substantially underreported in the data set provided by the ministry of commerce. 6 Shen and Chen et al. find empirical support for this argument. 7 According to Statistical Bulletin of China’s Outward Foreign Direct Investment (2015), the share of manufacturing FDI in total outward FDI is 26.3% for China’s investment into the United States and 19.7% for China’s investment into the EU. Note that the average share of manufacturing FDI in total outward FDI is 13.7% across all countries.

4.1 Motivation and Findings

85

hold when we compare private exporting firms to state-owned exporting firms. Given that exporting does not allow private firms to escape from domestic input distortions, this finding excludes an alternative hypothesis related to discriminations in the output market. Second, we find that the selection reversal pattern still exists, even after we exclude merger and acquisition (M&A)-type FDI projects or FDI projects to tax heaven economies from our analysis. As the motive of acquiring better technologies and brands is more pronounced for M&A-type FDI (compared with greenfieldtype FDI) and the motive of shifting profits is more pronounced for FDI into tax heaven economies, these two types of motives cannot explain our empirical finding of selection reversal. Although we focus on how a particular type of asymmetric institutional treatment affects economic outcomes, the insights of this study apply to other circumstances as well. For instance, a rising number of talented and wealthy French people moved abroad because of the increasing tax rates in France.8 This serves as a perfect example of institutional arbitrage. In India, red tape has forced many talented entrepreneurs to leave the country and start their businesses abroad as well.9 This study aims to speak to the literature on FDI and MNCs. In research on vertical FDI, Helpman (1984) insightfully points out how the difference in factor prices across countries affects patterns of vertical FDI. Antràs (2003, 2005) and Antras and Helpman (2004) emphasise the importance of contractual frictions for shaping the pattern of FDI and outsourcing. In research on horizontal FDI, Markusen (1984) postulates the concentration-proximity trade-off, which receives empirical support from Brainard (1994). More recently, HMY (2004) develop a model of trade and FDI with heterogeneous firms. Our study contributes to this literature by pointing out another motive for firms to engage in FDI. This study is also related to the literature that substantiates the existence of resource misallocation in developing economies. Hsieh and Klenow’s (2009) pioneering work substantiates substantial resource misallocation in China and India. Midrigan and Xu (2014), Moll (2014), and Gopinath et al. (2017) study the aggregate impact of financial frictions on firm productivity and investment. Guner et al. (2008), Restuccia and Rogerson (2008), and Garicano et al. (2016) explore the impact of size-dependent policies on aggregate productivity.10 Our work contributes to this research area by showing a link between domestic distortions and firms’ behaviour in the global market. The third related literature is the research on distortions in China and FDI decisions of Chinese firms (Brandt et al., 2013; Bai et al., 2015). Using a similar data set to ours, Shen and Chen et al. study the motives and consequences of China’s outward FDI into Africa. Using the same data set, Tian and Yu (2015) document the sorting pattern of Chinese MNCs, but abstract away from the difference between state-owned MNCs and private MNCs. Compared with the existing work, our chapter links firms’ outward FDI decisions to domestic distortions. 8

See http://www.france24.com/en/20150808-france-wealthy-flflee-high-taxes-les-echos-figures. Readers interested in studying anecdotal evidence of this can find it at http://www.thehindu.com/ news/national/red-tape-forces-top-indian-entrepreneurs-to-shift-overseas/article7367731.ece. 10 For a synthesis of work on misallocation and distortion, see Restuccia and Rogerson (2013). 9

86

4 Outward FDI and Domestic Input Distortions …

4.2 Data and Stylised Facts 4.2.1 Data Four main data sets are used in the present chapter, which we introduce as follows. Detailed discussions of the four data sets can be found in Online Appendix A. Annual Survey of Industrial Firms Data. Our first data set is a production data set of Chinese manufacturing firms from 2000 to 2013, which comes from the Annual Survey of Industrial Firms (ASIF) complied by the National Bureau of Statistics (NBS) of China. All SOEs and ‘above-scale’ non-SOEs (i.e., private firms) are included in the data set.11 FDI Decision Data. The nationwide data set of Chinese firms’ FDI decisions was obtained from the Ministry of Commerce of China (MOC). MOC requires every Chinese MNC to report its investment activity abroad since 1980, if it is above US$10 million. To invest abroad, every Chinese firm is required by the government to apply to the MOC for approval, or for registration if no approval is needed.12 In addition, the nationwide FDI decision data report FDI starters by year. Firm Land Price Data. To explicitly show the price discrimination against private firms in input factor markets, we use a comprehensive and novel firm-level data set of land price, which is collected from the official website of China’s land transaction monitoring system operated and maintained by the Ministry of Land and Resources. This monitoring system contains detailed information of land transactions, including land area, deal price, assigner and assignee. Orbis Data. Finally, we use the Orbis data from Bureau Van Dijk from 2005 to 2014, since they contain detailed financial information on foreign affiliates of Chinese MNCs. For the data before 2011, we merge our ASIF data with the Orbis data by matching the names in Chinese. For the data after 2011, we merge our ASIF data with the Orbis data using (Chinese) parent firms’ trade registration number which is contained in both data sets after 2011. We use the merged data set to study how Chinese MNCs allocate their sales across border. Data Merge. We merge the firm-level FDI and land price data sets with the manufacturing production database. Although the three data sets share a common variable—the firm’s identification number—their coding systems are completely different. Hence, we use alternative methods to merge the three data sets. The matching procedure involves three steps. First, we match the three data sets by using each firm’s Chinese name and year. If a firm has an exact Chinese name in all three data sets in a particular year, it is considered an identical firm. Still, this method could miss some firms since the Chinese name for an identical company may not have the exact Chinese characters in the three data sets, although they share some common 11

The above-scale’ firms are defined as firms with annual sales of RMB5 million (or equivalently, about US$830,000) or more before 2010 and with RMB10 million afterward. 12 Note that the SOEs directly controlled by central government are also required to report their FDI deals. This is why our data samples include such firms as CNPC (China National Petroleum Corporation), CPCC (China Petroleum Chemical Corporation), and China FResource Corporation.

4.2 Data and Stylised Facts

87

Table 4.1 FDI share in Chinese manufacturing firms (2000–12) Firm type

2000

(1) Manufacturing firms

83,579 110,498 199,873 194,201 158,220 306,366 283,018

(2) FDI mfg. parent firms 5 (in our sample)

2002

2004

2006

2008

2010

2012

9

56

562

867

1945

5501

0.01

0.03

0.29

0.55

0.64

1.94

(4) FDI mfg. parent firms – (in the bulletin)

–

–

2670

3650

4654

6744

(5) Matching percentage (%)

–

–

21.1

23.8

41.8

81.6

(6) FDI mfg. SOEs share 20.0 (%, in our sample)

22.2

5.35

3.02

1.49

1.23

1.81

(7) FDI SOEs share (%, in the bulletin)

–

–

26.0

16.1

10.2

9.1

(3) FDI share (%)

0.01

–

–

Notes Data on China’s MNCs were obtained from the Ministry of Commerce of China. FDI share in row is obtained by dividing the number of FDI manufacturing firms in row (2) by the number of manufacturing firms in row (1). That is (3) = (2)/(1). Matching percentage in row (5) equals the number of FDI manufacturing parent firms in our sample divided by number of FDI manufacturing parent firms in the bulletin (i.e., [5] = (2]/[4]). Numbers of FDI manufacturing parent firms in the bulletin before 2006 in column (4) are unavailable. FDI manufacturing SOEs share in row (6) reports the percentage share of state-owned manufacturing MNCs among all manufacturing MNCs in our sample. FDI SOEs share in row (7) denotes the percentage share of state-owned MNCs among all MNCs in the bulletin

strings. Our second step is to decompose a firm name into several strings referring to its location, industry, business type, and specific name. If a company has all identical strings in the three data sets, such a firm is classified as an identical firm. Finally, all approximate string-matching procedures are double-checked manually. We show the matching quality of our data in Table 4.1, and detailed discussions can be found in Online Appendix A. In short, we are able to match 21 42% of manufacturing MNCs reported in the statistical bulletin to our ASIF data between 2006 and 2010. Furthermore, the matching quality has improved substantially afterwards. In addition, our matched sample exhibits the same trend as in the statistical bulletin: The proportion of state-owned MNCs is decreasing over years. Although our firmlevel data set covers 2000–13, we use data for 2000–08 to conduct our main empirical analysis, as the data after 2008 lack information on (parent) firm’s value-added and use of materials, which disenables us to estimate firm productivity (a key variable in our empirical analysis). We instead use data after 2008 for robustness checks in Online Appendix. As highlighted by Feenstra et al. (2014), some observations in this firm-level production data set are noisy and misleading, largely because of misreporting by some firms. To guarantee that our estimation sample is reliable and accurate, we screen the sample and omit outliers by adopting the criteria a la Feenstra et al. (2014).13 13

For details, see Online Appendix A.

88

4 Outward FDI and Domestic Input Distortions …

4.2.2 Measures The SOE indicator and measured firm productivity are the two key variables used in this chapter. This subsection describes how we construct these two measures.

4.2.2.1

SOE Measures

We define SOEs using two methods. The first one is to adopt the official definition of SOEs, as reported in the China City Statistical Yearbook (2006), by using information on firms’ legal registration. A firm is classified as an SOE if its legal registration identification number belongs to the following categories: state-owned sole enterprises, state-owned joint venture enterprises and state-owned and collective joint venture enterprises. State-owned limited corporations are excluded from SOEs by this measure. As this is the conventional measure widely used in the literature, we thus adopt such a measure as the default measure to conduct our empirical analysis. Table 1 of the Online Appendix provides summary statistics for the SOE dummy used in this study.14 Recently, Hsieh and Song (2015) introduce a broader definition of SOEs and suggest defining a firm as an SOE when its state-owned equity share is greater than or equal to 50%. Along this line, we introduce an alternative way to define SOEs following their suggestion. As a result, a firm is defined as an SOE if either (i) it is classified as an SOE using the conventional measure; or (ii) its state-owned equity share is greater than or equal to 50%. We use such a broadly defined SOE dummy in our robustness checks.

4.2.2.2

TFP Measures

First and foremost, we estimate firm TFP using the augmented Olley–Pakes (1996) approach as adopted in Yu (2015). Compared with the standard Olley–Pakes (1996) approach, our approach has five new elements. First, we estimate the production function for MNCs and non-MNCs in each industry separately, since these two types of firms may adopt different technology.15 Second, we use detailed industry-level input and output prices to deflate a firm’s input use and revenue in our productivity estimation. As the revenue-based TFP may pick up differences in price–cost markup and prices across firms (De Loecker & Warzynski, 2012), an ideal method is to use firm-specific price deflators to construct quantity-based TFP. However, such data are not available in China. To mitigate this problem, we follow Brandt et al. 14

For details, see Online Appendix A. As a robustness check, we also pool MNCs and non-MNCs together and, in the inversion step of the productivity estimation, re-estimate the production function by including a dummy variable for MNC status. The results generated by this alternative method do not change our subsequent empirical findings, shown by Table 2 in the Online Appendix.

15

4.2 Data and Stylised Facts

89

(2012) to use four-digit Chinese Industrial Classification (CIC)-level input and output prices to deflate firm’s input use and revenue. Once industry-level price deflators are well defined and the price–cost markup is positively associated with true efficiency, revenue-based TFP captures the true efficiency of the firm reasonably well (Bernard et al., 2003). Third, we take the effect of China’s accession to the WTO (on firm performance) into account, as Chinese firms may export more or do more outward FDI due to the expansion of foreign markets after 2001. We thus include a WTO dummy in the inversion step of our productivity estimation. Fourth, and similarly, we also include a processing export dummy in the inversion step as processing exporters and nonprocessing firms may use different technology (Feenstra & Hanson, 2005). Last and most important, we also add an SOE indicator and an export indicator into the control function in the first-step Olley–Pakes estimates. In particular, we include the SOE indicator (and the export indicator) and its interaction terms with log-capital and log-investment to approximate the fourth-order polynomials in the inversion step of the TFP estimates. As stressed in Arkolakis (2010), firm TFP cannot be directly comparable across industries. We thus calculate the relative TFP (RTFP) by normalising our augmented Olley–Pakes TFP in each industry. As suggested by Ghandi et al. (2016), the correct identification of the flexible input elasticities should be based on the estimation of gross output production functions. Thus, all the TFP measures are conducted by using gross output production function.16 Although we control for the SOE indicator in the productivity estimation described above, it might still be unclear whether the TFP difference between SOE and private firms is caused by input factor distortions (or any other factors). If input factor distortions play an essential role in determining firms’ input use, it should be observed that SOEs are more capital intensive even within each narrowly defined industry (after controlling for firm size and other year-variant factors), as SOEs can access working capital at lower cost. Inspired by this intuition, we first regress the capital– labour ratio of the firm on its size (proxied by firm sales), industry fixed effects (at the finest four-digit CIC level), and year fixed effects, to obtain firm-level clustered residuals. We then interact these residuals with log-capital and log-investment as additional variables in the fourth-order polynomials used in the inversion step of the TFP estimates. We thus re-estimate our augmented relative TFP, taking into consideration the input distortions (RT F P Distort ). Finally, we also consider another specification (RT F PSDistort O E ) by including the firm-level clustered residuals and the SOE indicator (with interactions with log-capital and log investment) in the inversion step of the TFP estimates for robustness checks.

16

We thank a referee for pointing this out.

90

4 Outward FDI and Domestic Input Distortions …

4.2.3 Stylised Facts The main purpose of this subsection is to document three stylised facts using the merged data sets. As our interest is to explore how resource misallocation (across firm type) at home affects Chinese firms’ outward FDI behaviour, we compare state-owned MNCs with private MNCs when stating these stylised patterns.

4.2.3.1

Stylised Fact One: Productivity Premium for State-Owned MNCs

Table 4.2 reports the difference in our augmented Olley–Pakes TFP estimates between SOEs and private firms. Simple t-tests in columns (1) and (3) show that, among non-MNCs and non-exporting firms, private firms are more productive than SOEs. To confirm this finding, we perform nearest-neighbour propensity score matching, by choosing a dummy variable for capital-intensive industries and the year as covariates.17 To avoid the case in which multiple observations have the same propensity score, we perform a random sorting before matching. Columns (2) and (4) present the estimates for average treatment for the treated for private firms. Again, the coefficients of the productivity difference between SOEs and private firms are highly significant, suggesting that non-multinational (and non-exporting) SOEs are less productive than non-multinational (and non-exporting) private firms. The findings for non-MNCs are consistent with other studies, such as Hsieh and Song (2015). By contrast, a selection reversal is found when we focus on MNCs only. That is, private MNCs (i.e., private parent firms) are on average less productive than state-owned MNCs (i.e., state-owned parent firms), which is shown in column (5) in Table 4.2. To confirm this finding, we focus on the productivity difference between private and state-owned MNCs that are engaged in FDI and exporting as well.18 Column (7) reveals the same pattern. In columns (6) and (8), we perform nearestneighbour propensity score matching by choosing a dummy variable for capitalintensive industries, a dummy variable for rich destination economies, outward FDI mode (i.e., horizontal, vertical or R&D seeking) and the year as covariates. The results reported in these two columns confirm our finding that private MNCs are on average less productive than state-owned MNCs, even after we have controlled for aggregate-level factors. For more details, see Online Appendix B. The lower module of Table 4.2 presents evidence of the selection reversal using a broadly defined SOE indicator à la Hsieh and Song (2015). Compared with the numbers of MNCs and SOEs shown in the upper module, there are more SOEs 17

In Melitz-type models (Helpman et al., 2004; Melitz, 2003), firm size is a sufficient statistic for productivity. The model we will present is an extension of Helpman et al. (2004). Therefore, we do not use firm sales or employment as our covariates in the propensity score matching. 18 In reality, some Chinese MNCs engage in outward FDI and exporting. This is especially true for firms that undertake distribution FDI by setting up trade office abroad to promote exports. See Tian and Yu (2015) for detailed discussions.

3.56

0.50***

(47.85)

3.62

3.06

0.56***

(94.19)

(i) Private firms

(ii) SOE

Difference = (i)–(ii)

0.50***

(50.05)

0.56***

(97.56)

Difference = (iii)–(iv) (28.35)

0.33***

3.27

3.60

(27.20)

0.33***

3.27

3.60

(28.60)

0.50***

3.27

3.77

(27.75)

0.51***

3.26

3.77

(4)

Matched

(−2.11)

−0.25**

4.53

4.28

(−1.67)

−0.20*

4.48

4.28

(5)

Unmatched

All firms

MNCs

(−1.84)

−0.57*

4.66

4.09

(−1.68)

−0.52*

4.62

4.10

(6)

Matched

(−3.53)

−0.50***

4.78

4.28

(−3.30)

−0.48***

4.76

4.28

(7)

(−2.17)

−0.78**

4.90

4.12

(−2.08)

−0.75**

4.90

4.15

(8)

Matched

With exports only Unmatched

Note Columns (1)–(4) show that private firms have higher TFP than SOEs among non-MNCs with all firms and with exports only, respectively. Industrial capitalintensive indicator and year are used as covariates to obtain the propensity score for Non-MNCs in columns (2) and (4). Columns (5)–(8) show that private MNCs are less productive than state-owned MNCs with all firms and with exports only, respectively. Industrial capital-intensive indicator, year, rich-destination indicator, and MNC motivation (i.e., horizontal, vertical, or R&D seeking) are used as covariates to obtain the propensity score for MNCs in columns (6) and (8). In Rows (iii) and (iv) private firms and SOEs are defined by using the state share in firm’s ownership a l? Hsieh and Song (2015). The numbers in parentheses are t-values. *** (**, *) denotes the significance at 1% (5, 10%)

3.06

3.07

(iv) SOE

3.56

3.63

(iii) Private firms

Ownership defined by state share

3.06

(2)

(1)

(3)

With exports only Unmatched

Matched

PSM matching

All firms

Non-MNCs

Unmatched

Category

Table 4.2 Selection reversal: state-owned MNCs and private MNCs

4.2 Data and Stylised Facts 91

92

4 Outward FDI and Domestic Input Distortions …

engaged in outward FDI and more firms classified as SOEs when we use the broadly defined SOE dummy. In order to further validate our finding, we run simple OLS regressions. Specifically, we first regress the estimated TFP on the SOE indicator, the interaction term between SOE indicator and MNC indicator, and the firm fixed effects. Detailed discussions can be found in Online Appendix. A. In short, we find that the selection reversal holds in the regression results, as the own coefficient of the SOE indicator and its interaction term with MNC indicator are negatively (and positively) significant, respectively. Table 4.3 reports number of MNCs by types of ownership and the consequent fraction of MNCs during the sample year. There are 566 broadly defined state-owned MNCs in our sample between 2000 and 2013, which double its counterpart when SOEs are measured in a conventional way. Still, the evidence shows that private MNCs are less productive than state-owned MNCs, although private non-MNCs are more productive than state-owned MNCs. Our first stylised fact is robust to different TFP measures as shown in Table 4.4. Columns (1), (4)and (7) report relative TFP for all firms, non-MNCs and MNCs, respectively. A firm’s relative TFP is obtained by scaling down firm TFP in each industry after normalising the TFP of the most productive firm in that industry to one (see Arkolakis, 2010; Groizard et al., 2015). After normalisation, we calculate the relative TFP of firms in each industry. The TFP measure used in columns (2), (5) and (8), RT FP distort, takes firm’s input factor distortions into account when we estimate Table 4.3 Selection reversal: disproportionately more private MNCs 2000–8

Category

2000–13

# of MNCs # of all firms Fraction of # of MNCs # of all firms Fraction of MNCs MNCs MNCs

(1)

(2)

(3)

(4)

(5)

(6)

(i) Private firms

3,623

1,335,514

0.27%

21,426

2,287,915

0.94%

(ii) SOE

104

40,612

0.25%

270

66,192

0.41%

Ownership 3,622 defined by state share (iii) Private firms

1,097,322

0.33%

21,130

2,273,486

0.93%

(iv) SOE

43,512

0.24%

566

80,621

0.70%

105

Notes Column (3) reports the fraction of MNCs that is obtained by dividing column (1) by column (2) for years 2000–8. Similarly, column (6) reports the fraction of MNCs that is obtained by dividing column (4) by column (5) for years 2000–13. Clearly, the share of MNCs is smaller among SOEs than among private firms, which is consistent with part 3 of Proposition 4.1. In rows (iii) and (iv) private firms and SOEs are defined by using the state share in firm’s ownership a la‘ Hsieh and Song (2015). Note that we lose some observations when defining SOEs using the state share in firm’s ownership, as some firms did not report their state shares in our data. Refer to the texts for details

4.2 Data and Stylised Facts

93

t a firm’s relative TFP. The alternative firm TFP measure, RT F PSdistor O E , reported in columns (3), (6) and (9), puts the SOE dummy, distortion residuals and their interaction terms with other firm-level key variables into TFP estimations, as discussed above. Again, our findings are robust to the different TFP measures. Our data clearly exhibit selection reversal in the sense that private MNCs are less productive than state-owned MNCs. Equally interestingly, we then look at the productivity difference between stateowned and private MNCs industry by industry. To do so, we separate all industries into two categories: capital-intensive and labour-intensive, according to the official definition adopted by the National Statistical Bureau of China.19 The lower module of Table 4.4 shows that a productivity premium for state-owned MNCs exists in capitalintensive industries. This finding is important, as it shows that selection reversal exists in industries with more severe distortions in the input market.20 To verify that input distortion plays an essential role in interpreting the productivity premium of state-owned MNCs (compared to private MNCs), we need to make sure that both SOEs and private firms have similar productivity dispersions (also implied by our model in the next section). Admittedly, the productivity distribution of SOEs might have a different level of dispersion compared with that of private firms, and the productivity distribution may change during the era of SOE reforms (see, e.g., Hsieh & Song, 2015). However, we show that the productivity distribution of stateowned MNCs first-order stochastically dominates that of private MNCs in Online Appendix A (i.e., state-owned MNCs are more productive than private MNCs at each percentile of the distribution). Finally, as all of the TFP estimates are essentially based on the Olley–Pakes approach, which uses investment as a proxy for TFP. A possible concern with the Olley–Pakes approach that uses investment as a proxy in the first stage is that investment in developing countries like China is lumpy and the existence of too many zeros can create bias. However, this is not a problem in our estimations as discussed in Feenstra et al. (2014). In particular, we have already dropped those bizarre observations in our sample following the General Accepted Accounting Principle criteria. Also, the most recent advances on TFP estimation and identification such as Ghandi et al. (2016) suggest that adopting the gross output production functions can better mitigate this potential problem. Therefore, all TFP estimates in the present chapter adopt the approach of the gross output production function. Still, for the sake of completeness, we report simple labour productivity (defined as value-added per employee) and Levinsohn-Petrin (2003) TFP in Online Appendix Table 3. Once again, we see that state-owned non-MNCs are less productive than private non-MNCs. But the opposite is true for MNCs: State-owned MNCs are more productive than private MNCs. In short, our first empirical finding is robust.

19

In particular, among the 28 CIC two-digit industries, the following industries are classified as labour-intensive sectors: processing of foods (code: 13), manufacture of foods (14), beverages (15), textiles (17), apparel (18), leather (19) and timber (20). 20 Section 4.4 shows that the input price wedge mainly exists in the credit (i.e., capital) market.

0.016***

(46.42)

0.094***

(93.95)

Difference = (i)–(ii)

0.477

0.023***

(59.05)

0.422

0.087***

(78.03)

(iv) SOE

Difference = (iii)–(v) (59.54)

0.023***

0.480

0.503

(46.29)

0.016***

0.481

0.497

(3)

Distor t RTFPsoe

(78.28)

0.087***

0.422

0.509

(97.07)

0.094***

0.411

0.505

(4)

RTFPOP

(59.14)

0.023***

0.477

0.500

(46.53)

0.015***

0.479

0.494

(5)

RTFPDistort

Non-MNC firms

(59.62)

0.023***

0.480

0.503

(46.40)

0.016***

0.481

0.497

(6)

Distor t RTFPsoe

(−2.39)

– 0.052***

0.676

0.624

(−1.69)

−0.034*

0.650

0.616

(7)

RTFPOP

MNC firms

(−1.65)

−0.020*

0.525

0.505

(2.69)

−0.028***

0.528

0.500

(8)

RTFPDistort

(−1.64)

−0.020***

0.529

0.509

(2.73)

−0.029***

0.532

0.503

(9)

Distor t RTFPsoe

Notes Number in parenthesis are t-value. *** (**, *) denotes the significance at 1(5, 10)%, respectively. Columns (1)–(3) show that private firms have higher relative TFP than SOEs for all firms. Similarly, columns (4)–(6) show that private non-MNC firms have higher relative TFP than SOE non-MNC firms. Columns (7)–(9) show that private MNC firms are less productive than state-owned MNCs. Columns (1), (4) and (7) are relative Olley-Pakes TFP. Columns (2), (5) and (8) are relative TFP featured with input factor distortions. Columns (3), (6) and (9) are relative TFP featured with input factor distortions and interacted SOE dummy with other polynomials. The upper module includes all sample whereas the bottom one includes capital-intensive industries only, which account for around three quarters of the entire sample

0.500

0.509

(iii) Private firms

Capital-intensive industries only

0.478

0.412

(ii) SOE

0.494

(1)

Measures of RTFP

0.506

(2)

RTFPop

(i) Private firms

RTFPDistort

All firms

Category

Table 4.4 Productivity premium of state-owned MNC and relative TFP (2000–08)

94 4 Outward FDI and Domestic Input Distortions …

4.2 Data and Stylised Facts

4.2.3.2

95

Stylised Fact Two: Smaller Fraction of State-Owned MNCs

Columns (3) and (6) of Table 4.3 present our second stylised fact. That is, the fraction of MNCs is larger among private firms than among SOEs. Again, this finding is robust to different definitions we use to construct the SOE indicator and the different time periods we focus on. When using a broadly defined SOE indicator, we find that more firms are classified as SOEs whereas the number of state-owned MNCs does not change much for the sample of 2000–8. For the period of 2000–13, the share of MNCs increases both among SOEs and among private firms compared with the period of 2000–8. In all four cases (two time periods and two definitions of SOEs), there are always disproportionately more MNCs among private firms than among SOEs. On the one hand, this finding is puzzling, since SOEs are larger firms that should be more likely to invest abroad. Furthermore, the Chinese government has supported its SOEs’ investing abroad for many years, known as the Going-Out strategy. On the other hand, such an observation is consistent with our first finding. Namely, as state-owned MNCs are more productive than private MNCs, the fraction of SOEs engaged in FDI should be smaller (i.e., tougher selection).

4.2.3.3

Stylised Fact Three: Larger Relative Size Premium for State-Owned MNCs

Our last stylised fact is related to the relative size premium of state-owned MNCs. The conventional view is that SOEs are larger in size, which is usually measured by log employment or log sales. Our data also exhibit such features, as shown in Table 4 of the Online Appendix. More importantly, the size premium for state-owned MNCs holds in the relative sense as well. Table 4.5 shows that the ratio of average log employment of multinational parent firms to that of non-exporting firms is larger among SOEs than among private firms. Table 4.5 reports the result obtained from the comparison between the relative size of state-owned MNCs and that of private j j j j MNCs. The relative size is measured by jo / jd where jo and jd are the average log employment of MNCs and that of non-exporting firms for firm type j (i.e., private or state-owned). The year-average ratio in the first column shows that the relative size of private MNCs is significantly smaller than that of SOEs. As few SOEs were engaged in outward FDI before 2005, we report the year-average ratio up to a particular year in Table 4.5 as well. All columns suggest larger relative size for state-owned MNCs. To sum up, our third stylised fact states that the absolute and relative sizes of private MNCs (compared with non-exporting firms) are smaller than those of state-owned MNCs. Thus far, we have established three interesting empirical findings. In what follows, we will present a model to rationalise these findings. Furthermore, the model yields several additional empirical predictions, which will be shown to be consistent with the data.

Avg.

≤2001

5.65

−1.06***

(−234.0)

5.48

−0.97***

(−488.1)

(2) SOE

Size difference = (1)–(2) (−283.5)

−1.05***

5.64

4.59

≤2002

(−329.0)

−1.02***

5.58

4.56

≤2003

(−374.1)

−1.01***

5.55

4.54

≤2004

(−400.1)

−1.00***

5.53

4.53

≤2005

(−430.4)

−0.99***

5.51

4.52

≤2006

5.49 (−445.5)

−0.98***

4.51

≤2007

(−466.6)

−0.98***

5.48

4.50

≤2008

Notes This table reports the difference in relative firm size between private MNCs and state-owned MNCs. Firm size is measured by log employment. The table shows that the relative size of FDI firms to non-exporting firms is smaller for private firms than that for SOEs. This finding is consistent with part 1 of Proposition 4.3, that relative size of MNCs is smaller for private firms than for SOEs. The numbers in parentheses are t-values. *** (**, *) denotes significance at the 1% (5, 10%) level

4.59

4.50

(1) Private firms

Relative size of MNCs to non-exporting firms (l o /l d )

Year coverage

Table 4.5 Relative size premium for SOEs

96 4 Outward FDI and Domestic Input Distortions …

4.3 Model

97

4.3 Model We modify the standard horizontal FDI model proposed by HMY (2004) to rationalize the empirical findings documented so far. We study how discrimination against private firms in the input market affects the sorting pattern of MNCs and their size premium at the intensive margin. At the same time, we investigate how the difference in foreign investment costs impacts the investment behaviour of private MNCs and state-owned MNCs at the extensive margin.21

4.3.1 Setup There is one industry populated by firms that produce differentiated products under conditions of monopolistic competition á la Dixit and Stiglitz (1977). Each variety is indexed by ω, and Q is the set of all varieties. Consumers derive utility from consuming these differentiated goods according to ⎡ U=⎣

σ ⎤ σ−1

 q(ω)

σ−1 σ

dω⎦

(4.1)

ω∈

where q(ω) is the consumption of variety ω, and σ is the constant elasticity of substitution between differentiated goods. Entrepreneurs can enter the industry by paying a fixed cost, f e , in terms of the unit of goods produced by the firm.22 After paying the entry cost, the entrepreneur receives a random draw of productivity, ϕ, for her firm. The cumulative density function of this draw is assumed to be F(ϕ). Once the entrepreneur observes the productivity draw, she decides whether or not to stay in the market as there is a fixed cost to produce, f D , (in terms of the units of the goods produced by the firm). After entering and choosing to stay in the domestic market, each entrepreneur also chooses whether to serve the foreign market. There are two options for doing this, the first of which is exporting. Exporting entails a variable trade cost, τ (≥1), and a fixed exporting cost, f X . The second way is to set up a plant in the foreign country and produce there directly. The cost of doing this is fixed and denoted by f I .

21

Major predictions of the canonical horizontal FDI model a la HMY (2004) are consistent with our empirical findings documented in Table 2. For instance, average productivity of MNCs is higher than that of non-multinational firms (see columns 3 and 5 of the table). Moreover, after the propensity score matching, we find that average productivity of non-multinational firms (domestic firms plus exporting but non-multinational firms) is higher than that of domestic firms (see columns 2 and 4 of the table). 22 We follow Bernard et al. (2007) to choose this specification in order to make various fixed costs have the same capital (or land) intensity as the variable cost.

98

4 Outward FDI and Domestic Input Distortions …

Both fixed costs of serving the foreign market are in terms of the units of the goods produced by the firm. Similar to Bernard et al. (2007), there are two factors of production, capital (or land) and labour, and the production function takes the following constant-elasticityof-substitution form:  μ−1  μ−1 μ−1 μ q(k, l) = ϕ k μ + l μ

(4.2)

where k and l are capital (or land) and labour inputs respectively, and ϕ is the productivity draw the firm receives. Parameter μ(≥1) is the elasticity of substitution between capital and labour. We assume that there are two types of firms in the economy: private firms and SOEs.23 The key innovation of the model is to introduce a wedge between the input price paid by SOEs and by private enterprises when they produce domestically. Specifically, it is assumed that private firms pay a capital rental price (or the unit land price) c(>1) times as high as what SOEs pay when they produce domestically. However, firms pay the same wage and capital rental price (or the unit land price) when producing abroad.24 Based on Eq. (4.2), we derive total variable cost and total fixed cost as TVC(q, ϕ) =

qr

1 ϕ 1 + ωμ−1 μ−1

and fi r FC(r, w) =

1 1 + ωμ−1 μ−1 where r and w are the capital rental price (or the unit land price) and the wage rate and i ∈ {D, X, I}. Variable ω = wr is relative price of capital (or land). Capital (or )w land) intensity in equilibrium is given by l(w,r = ωμ−1 . As long as μ > 1, a higher k(w,r )r relative price of capital leads to lower capital (or land) intensity. This property is utilised in our productivity estimation.

23

We do not take a stance on why some firms become SOEs (or private enterprises), since the predictions of the model do not depend on this. 24 We will show that there is evidence for the existence of an input price wedge in the credit land markets, but not in the labour market. Since buying capital usually requires a substantial amount of borrowing, we assume that private firms pay a higher capital rental price than SOEs.

4.3 Model

99

4.3.2 Domestic Production, Exporting and FDI We derive firm profit and revenue as follows. Based on Eq. (4.1), the demand function for variety ω can be derived as p(ω)−σ E P 1−σ

q(ω) =

where E is the total income of the economy and P is the ideal price index and defined as ⎡ ⎢ P≡⎣

1 ⎤ 1−σ



⎥ p 1−σ (ω)MdF(ω)⎦

,

(ω)∈

where M is the total mass of varieties in equilibrium. The resulting revenue function is R(q) = q

σ−1 σ

E σ Pβ , β ≡ 1

σ−1 σ

We derive the SOE’s operating profit of domestic production and exporting first. Since both types of production use domestic factors only, their operating profits are given by DH π S D (ϕ) = σ

βϕ rH

σ −1   σ −1 μ−1 μ−1 1 + ωH

and DF π S X (ϕ) = π S D (ϕ) + σ

βϕ τr H

σ −1   σ −1 μ−1 μ−1 1 + ωH

where Di ≡ Piσ −1 E i and i ∈ {H, F} Subscripts S, D, X, H and F refer to SOE, domestic production, exporting, home country and foreign country, respectively. For private firms, the operating profits are π P D (ϕ) = and

DH σ

βϕ cr H

σ −1



σ −1 1 + (cω H )μ−1 μ−1

100

4 Outward FDI and Domestic Input Distortions …

DF π P X (ϕ) = π P D (ϕ) + σ

βϕ τ cr H

σ −1



1 + (cω H )μ−1

σ −1

μ−1

Firm’s revenue is Rij (ϕ) = σ π ij (ϕ) where i ∈ {S, P} and j ∈ {D, X}. We can derive the exit cutoff and the exporting cutoff for SOEs and private firms respectively: 1

ϕSD =

1

r H (σ cr H f D /D H ) σ −1 r H (σ cr H f X /D F ) σ −1 ; ϕSX = τ  ; σ   (σ −1))μ−1)  σ μ−1 μ−1 (σ −1))μ−1) β 1 + ωH β 1 + ωH 1

1

ϕPD =

cr H (σ cr H f X /D F ) σ −1 cr H (σ cr H f D /D H ) σ −1 ; ϕPX = τ σ

σ

β 1 + (cω H )μ−1 (σ −1))μ−1) β 1 + (cω H )μ−1 (σ −1))μ−1)

Note that ϕ P D > ϕ S D and ϕϕ P X = ϕϕ S X . SD PD Now, we discuss the case of FDI. Following HMY, we assume that the firm uses foreign factors to produce after setting up a plant in the foreign country.25 In addition, foreign factors are used to pay for the fixed FDI cost.26 Thus, the operating profit of firms that engage in FDI is:

 σ −1  μ−1 βϕ σ −1  ; 1 + ωμ−1 F rF

  σ −1 D F βϕ σ −1  μ−1 μ−1 . π P O (ϕ) = π P D (ϕ) + 1 + ωF σ rF π S O (ϕ) = π S D (ϕ) +

DF σ

When both SOEs and private firms produce abroad, they face the same factor prices. The FDI cutoffs are pinned down by the following indifference conditions (between exporting and engaging in FDI):  fI rF



1+

μ−1 ωF

− 1  μ−1

fX rH 1+

μ−1 ωH

DF

= 1  μ−1 σ

 − 25

βϕ S O

σ −1

μ−1

1 + ωH

[

μ−1

1 + ωF

σ −1  μ−1

r Fσ −1

σ −1  μ−1

(τ r H )σ −1

],

In our data set of Chinese MNCs from Zhejiang, we checked whether firms increased their foreign investment after the initial investment and ended up with few cases. The finding is that at least a substantial fraction of factors used in foreign production (including capital and land) is sourced from the foreign country. 26 It is worth stressing that our theoretical predictions will hold well independent of this assumption. In Online Appendix D, we allow for FDI fixed cost to be paid using domestic factors, and private firms do not face discrimination when they pay the FDI fixed cost using domestic factors. In both cases, our theoretical results are still preserved.

4.3 Model

101

and f I rF



μ−1

1 + ωF

− 1  μ−1

σ −1 DF [ = βϕ Po 1

σ 1 + (cω H )μ−1 μ−1 DF

 σ −1  μ−1 μ−1 1 + ωF r Fσ −1



σ −1 1 + (cω H )μ−1 μ−1 ] − (cτ r H )σ −1 It is evident that selection into FDI is tougher for SOEs than for private firms (i.e., ϕ S O > ϕ P O ), as the opportunity cost of engaging in FDI is smaller for private firms than for SOEs.

4.3.3 Domestic Distortion and Patterns of Outward FDI In this subsection, we discuss how the existence of domestic distortions in the capital and land markets affects the patterns of outward FDI at the extensive and intensive margins. Proposition 4.1 Sorting Patterns of Private Firms and SOEs (Extensive Margin): (1) The exit cutoff and exporting cutoff are higher for private firms than for SOEs. However, the cutoff for becoming an MNC is lower for private firms than for SOEs (i.e., selection reversal). (2) Conditional on the initial productivity draw (and other firm-level characteristics), private firms are more likely to become MNCs. (3) Assume that the truncated distribution of the productivity draw for private firms (weakly) first order stochastically dominates (FOSD) that of SOEs, or the two conditional probability density functions (PDFs) satisfy the (weak) monotone likelihood ratio property (MLRP) with: ∂ ∂ϕ

f P (ϕ|ϕ ≥ ϕ0 ) f S (ϕ|ϕ ≥ ϕ0 )

 ≥ 0, ∀ϕ ≥ ϕ0

where f P (ϕ|ϕ ≥ ϕ0 ) and f S (ϕ|ϕ ≥ ϕ0 ) are the truncated probability density functions of the productivity draw for private firms and SOEs, respectively. Then, the fraction of MNCs is larger among private firms than among SOEs. Furthermore, simple average productivity of private firms is greater than that of SOEs overall. (4) Assume that both types of firms draw productivities from the same distribution (which trivially satisfies weak FOSD property). Then the (simple) average productivity of private MNCs is smaller than that of state-owned MNCs (i.e., productivity premium for state-owned MNCs).

102

4 Outward FDI and Domestic Input Distortions …

Proof See Online Appendix C. The intuition for Proposition 4.1 is as follows. First, since there is discrimination against private firms at home, it is more difficult for private firms to survive and export. As a result, the exit cutoff and the exporting cutoff are higher for these firms. Absent the choice of exporting, the FDI cutoff would be the same for SOEs and for private firms, as they would face the same benefit and costs of doing FDI in the relative sense. However, since the firm at the FDI cutoff compares exporting with FDI, the (opportunity) cost of engaging in FDI is smaller for private firms than for SOEs.27 As a result, the FDI cutoff is lower for private firms than for SOEs. Online Appendix Table 8 shows that the selection reversal holds, as the estimated productivity at the 1% (and 5%) percentile is higher for state-owned MNCs than for private MNCs. If we make assumptions on the distribution of the productivity draws, the selection reversal leads to an average productivity premium for state-owned MNCs, and the above theoretical results rationalise the first two stylised facts.28 Finally, Table 4.6 and Online Appendix Tables 5, 6 in the next section show the lower probability of becoming an MNC for SOEs. We next discuss how a variation in the level of domestic distortion affects the sorting pattern of private MNCs and state-owned MNCs differently using the following proposition. Proposition 4.2 Cross-Industry Variations: In industries with more severe distortion (i.e., c↑), the productivity premium of stateowned MNCs is larger. Moreover, SOEs are less likely to produce abroad in industries with more severe distortion than SOEs in industries with less severe distortion. Assume that the production function is Cobb–Douglas with capital and labour. Then, the productivity premium of state-owned MNCs is more pronounced in capital intensive industries. Furthermore, SOEs are less likely to engage in FDI (compared with private firms) in capital intensive industries. Proof See Online Appendix C. The intuition for the Proposition 4.2 is straightforward. Since the asymmetric distortion disincentives SOEs to produce abroad, the selection into the FDI market becomes more stringent for SOEs (than for private firms) in industries with more severe discrimination against private firms. Furthermore, as the distortion exists in the capital market, we expect a more stringent selection into the FDI market for SOEs (than for 27

Exporting does not eliminate the distortion private firms face in the domestic market. The selection reversal holds irrespective of the distribution of the initial productivity draw. The average productivity premium for state-owned MNCs exists, if SOEs and private firms draw productivity from the same distribution. However, the assumption of the same productivity distribution is not required. What we need is that a lower cutoff on the productivity draw implies a smaller average productivity (i.e., a relationship between the marginal productivity and the inframarginal productivity). This why we need MLRP for part 3 of the above proposition.

28

No

No

No

No

No

No

No

No

Foreign firms dropped

Tax haven dropped

Distr. FDI dropped

Switching SOE Dropped No

M&A deals dropped

No

Yes

Yes

Yes

Yes

Industry fixed effects

(4.45)

(7.42)

(10.85)

0.900***

(6.55)

0.004***

0.589***

0.003***

1.333*

(1.80)

0.009***

(4.14)

Year fixed effects

Export indicator

Log firm labour

Firm TFP 1.716**

No

No

No

No

Yes

Yes

Yes

(6.03)

1.142***

(9.78)

0.623***

(2.00)

No

No

No

No

Yes

Yes

Yes

(26.01)

1.102***

(38.49)

0.588***

(18.50)

4.237***

(−12.63)

No

No

No

No

Yes

Yes

Yes

(6.03)

1.145***

(8.86)

0.587***

(2.25)

1.838**

(−2.81)

(5)

No

No

No

No

Yes

Yes

Yes

(6.03)

1.145***

(8.86)

0.587***

(2.26)

1.843**

(−2.88)

(6)

No

No

No

Yes

Yes

Yes

Yes

(5.43)

1.167***

(7.86)

0.565***

(1.66)

1.552*

(−3.26)

(7)

2004–8

(4.25)

2.360***

(−2.68)

(9.37)

No

No

Yes

No

Yes

Yes

Yes

(3.81)

No

Yes

No

No

Yes

Yes

Yes

(5.82)

0.736*** 1.145***

(8.21)

Narrow (10)

(continued)

Yes

Yes

No

No

Yes

Yes

Yes

(6.49)

1.174***

(11.03)

0.567***

(4.96)

2.500***

(−2.56)

−0.703*** −0.662**

(9)

Narrow

0.734*** 0.574***

(1.28)

1.603

(−1.73)

(8)

−0.757*** −1.306*** −0.693*** −0.682*** −1.179*** −0.532*

(4)

(−2.88)

(−1.85)

(3)

−0.002** −0.454*

Narrow

(−2.41)

Narrow

SOE indicator

Broad

(2)

Narrow

Complementary log–log

(1)

Narrow

2000–8

Rare event Logit

Variable

Narrow

Logit

Narrow

Logit

Narrow

LPM

SOE defined

Year coverage

Regressand FDI indicator

Table 4.6 Private firms are more likely to undertake FDI

4.3 Model 103

1,136,603 1,135,467 895,209

Observations

896,314

(4)

Narrow

2000–8

Rare event Logit

895,209

(5)

Narrow 895,210

(6)

Broad 894,815

(7)

Narrow 893,754

(8)

Narrow

Complementary log–log 2004–8

701,277

(9)

Narrow

701,204

(10)

Narrow

Notes The regressand is the FDI indicator. All columns include industry-specific fixed effects and year-specific fixed effects. The numbers in parentheses are t-values clustered at the firm level. *** (**) denotes significance at the 1% (5%) level. Columns (1)–(2) include foreign-invested firms whereas all other columns drop those firms. Columns (1)–(8) cover data over the period of 2000–8 whereas columns (9)–(10) cover data over the period of 2004–8, Column (6) uses broadly defined SOE. Column (7) drops outward FDI to tax haven destinations. Column (8) drops distribution-oriented FDI (i.e., Distr. FDI). Column (9) drops the switching SOEs (i.e., switching from SOEs to private firms). Column (10) drops both switching SOEs and merger and acquisition deals. In all columns, TFP is measured by augmented Olley-Pakes controlling for input price distortion and SOE status

(3)

(2)

(1)

Variable

Narrow

Logit

Narrow

Logit

Narrow

LPM

SOE defined

Year coverage

Regressand FDI indicator

Table 4.6 (continued)

104 4 Outward FDI and Domestic Input Distortions …

4.4 Evidence

105

private firms) in capital intensive industries. We will provide empirical evidence for these two predictions in what follows. Finally, we discuss how domestic distortion affects the sorting patterns of MNCs at the intensive margin. Proposition 4.3 Sorting Pattern of Private Firms and SOEs (Intensive Margin): (1) Suppose the initial productivity draw follows a Pareto distribution with the same shape parameter for private firms and SOEs. Then, the relative size of private MNCs in the domestic market (i.e., compared with private non-exporting firms) is smaller than that of state-owned MNCs (i.e., compared with non-exporting SOEs). (2) Conditional on productivity and other firm-level characteristics, the ratio of foreign sales to domestic sales is higher for private MNCs than for state-owned MNCs. Proof See Online Appendix C. The intuition for Proposition 4.3 is straightforward. Since there is an extra benefit for private firms to produce abroad, they produce and sell more in the foreign market. This effect is another key result of our model, for which we provide empirical support in the next section. The first part of Proposition 4.3 receives strong statistical support from Table 4.5. As the table shows, the relative size of private MNCs is smaller than that of state-owned multinational firms. We will provide evidence for the second part of the above proposition in what follows.

4.4 Evidence Our theoretical model yields three empirical propositions. Some of the predictions of the propositions have already been shown to be consistent with the stylised facts presented in Sect. 4.2; others are still waiting for empirical examination, which is the purpose of this section.

4.4.1 FDI Decision and Firm Ownership Most of the predictions of Proposition 4.1 have been shown to be consistent with the empirical results in Tables 4.2, 4.3, 4.4, 4.5. Only part 2 of Proposition 4.1 needs further empirical examination. The nationwide FDI data only contain the information of the first year when firms began to undertake FDI in a given country (i.e., no information on whether firms exited from FDI in a particular country after entry). Therefore, the estimations in Table 4.6 and the other tables include non-MNCs and current MNCs every year.

106

4 Outward FDI and Domestic Input Distortions …

Table 4.6 reports the estimation results starting from a linear probability model (LPM) in which the regressand is an indicator of outward FDI. As the outward FDI data set only reports the first year when a firm engages in outward FDI in a given country, we assume that a firm will continue to engage in outward FDI afterward.29 That is, the FDI indicator equals one once a firm engages in FDI and zero otherwise. To explore whether SOEs are less likely to engage in FDI, we include an SOE indicator in the regression, as well as several key firm characteristics, such as firm size (i.e., log employment), firm-level TFP and exporting status. Equally important, we include industry-specific fixed effects and year-specific fixed effects to control for unobservable time-invariant and industry-invariant factors.30 The SOE indicator is shown to be negative and statistically significant in column (1), suggesting that SOEs are indeed less likely to engage in outward FDI. The magnitude of the SOE indicator is too small, which is probably due to a well-known pitfall of LPM: The predicted probability could be greater than one or less than zero. To overcome this drawback, we report the logit estimates in column (2) by controlling for a rich set of fixed effects with interactions of industry and year dummies, which yield qualitatively the same results as for the LPM model. Particularly, compared with private firms, SOEs are less likely to engage in outward FDI. For such a nonlinear probability model, firm-specific fixed effects cannot be included in the regression. Instead, we control for year-specific and industry-specific fixed effects in all the rest of the regressions. Our estimates include foreign-invested enterprises (FIEs), which are firms that receive direct investment from foreign entities. However, if an FIE has a dominant share of foreign stakes, it is directly controlled by its foreign headquarters. Our model does not consider such firms, as FIE’s headquarters are not located in the home country. Thus, we drop FIEs from the sample in all regressions, and columns (3)–(10) in Table 4.6 report the results. After dropping the FIE sample, the logit estimates in column (3) still show that SOEs are less likely to engage in outward FDI, conditioning on other firm-level characteristics. There are two important caveats here. First, as shown in row (3) of Table 4.1, less than 1% of manufacturing firms undertook FDI in most of the sample years. Within MNCs, a small fraction of them are SOEs. As highlighted by King and Zeng (2001), standard binary nonlinear models, such as logit or probit models, underestimate the probability of rare events. To address this concern, King and Zeng recommend using the rare-event logit approach, which corrects for possible downward bias.31 Column (4) in Table 4.6 reports the logit estimates with rare-event corrections. The 29

It is important to note that our findings remain unchanged even without imposing such an assumption and with only FDI starters being examined. This can be seen from Table 5 of the Online Appendix. 30 In principle, we can control for firm-specific fixed effects. However, since there are only few ODI deals per Chinese MNC over the period 2000–13, there are not enough degrees of freedom to identify the coefficient on the SOE indicator with firm-specific fixed effects. We thus add industry-specific fixed effects in the LPM estimates. 31 Rare-events estimation bias can be corrected as follows. We first estimate the finite sample bias ˆ where βˆ denotes the ˆ to obtain the bias-corrected estimates βˆ − bias(β), of the coefficients, bias(β), coefficients obtained from the conventional logistic estimates.

4.4 Evidence

107

key coefficient of the SOE indicator is much larger than its counterparts in columns (2) and (3) in absolute value. Equally important, the coefficient is still negative and statistically significant, ascertaining that SOEs are less likely to engage in outward FDI. The rare-event feature of our FDI data also generates another problem, that the probability distribution of state-owned MNCs engaging in FDI exhibits faster convergence toward the true probability that SOEs engage in foreign investment. Standard logit or probit estimates cannot deal with this problem. We thus run complementary log–log regressions in the rest of Table 4.6, which allows for faster convergence toward rare events. Column (5) of Table 4.6 reports the complementary log–log regression by dropping foreign firms. Column (6) adopts the broadly defined SOE indicator in the regression. Clearly, our key results are robust regardless of different SOE definitions. There may be a worry that some Chinese firms may invest in tax haven destination economies, such as Hong Kong and the Cayman Islands, due to the motive of tax evasion or profit shifting.32 Consequently, our model and its underlying story cannot be applied to those firms. Column (7) of Table 4.6 thus drops observations of outward FDI in tax haven destinations.33 Similarly, it is also possible that some Chinese firms may establish trading offices in their exporting destinations to promote marketspecific exports (Tian & Yu, 2015). Such distribution-oriented outward FDI is not the focus of our model, and our theory does not apply to this type of FDI. Column (8) thus drops the sample of distribution-oriented FDI. Next, as shown in row (2) of Table 4.1, China’s outward FDI increases rapidly after 2004, when the government adopted policies to encourage firms to go abroad. It is also true that a large wave of privatisation of SOEs took place after 1998 (Hsieh & Song, 2015). We thus drop SOE switching firms from the sample and focus on observations from 2004 to 2008 in columns (9) and (10) of Table 4.6. The coefficient of the SOE indicator in column (9) is larger than its counterpart in column (8), suggesting that private firms were more likely to go abroad after 2004. Still, there may be a worry that our story fits better into the case of greenfield FDI rather than M&A-type FDI, as the latter usually targets better technology or seeks famous brand names of the targeted firms. We thus drop the M&A–type FDI in column (10) and the estimation still yields the same results: SOEs are less likely to engage in outward FDI.34 Finally, we provide two additional robustness checks. First, we rerun all the regressions in Table 5 of the Online Appendix by setting the FDI indicator to one only in the first year when the Chinese manufacturing firm became an MNC (i.e., the indicator for starting FDI). The results reported in Appendix Table 5 show that our findings in Table 4.6 are not driven by subsequent entries into the outward FDI market. Second, the inclusion of destination-specific fixed effects and affiliates’ industry-specific 32

See Garetto et al. (2017) for this point. The tax haven regions include the Bahamas, Bermuda, the Cayman Islands, Hong Kong, Luxembourg, Macao, Monaco, Panama, the Virgin Islands and Switzerland. 34 To identify M&A–type FDI, we manually merge the outward FDI data set with the M&A–type FDI data compiled by Thomson Reuters, by using the identical names of Chinese parent firms. 33

108

4 Outward FDI and Domestic Input Distortions …

fixed effects does not change our findings in Table 4.6 either, though the value of the estimated coefficients changes somewhat. The results are reported in Appendix. Table 6 of Online Appendix. In total, our finding of a lower probability of SOEs’ conducting outward FDI is robust to different estimation methods, various specifications and different time spans.

4.4.2 Input Market Distortions Our theoretical model is built on the premise that, compared with SOEs, private firms have to bear higher input costs in the domestic market. Although this assumption seems to be widely accepted, we provide direct evidence for it in this subsection. Previous work suggests that Chinese SOEs access working capital by paying a lower interest rate than what private firms pay (Feenstra et al., 2014). Similarly, SOEs acquire land at a lower market price than private firms, which is especially true in the manufacturing sector (Tian et al., 2016). To investigate whether these conjectures are supported by the data, we first construct a measure of firm-level interest rate by dividing the firm’s interest expenses by its current liabilities (in each year), both of which are available in the ASIF data set over the period of 2000–8 (but not after 2008). We then regress this measure on the narrowly defined SOE indicator in columns and (2) in Table 4.7. If our underlying assumption that SOEs access external working capital at a lower cost than private firms is supported by the data,35 we should observe that the SOE indicator has a negatively significant coefficient. This outcome is exactly what we observe in Table 4.7. Estimates in columns (1)– (3) include year-specific and industry-specific fixed effects. In addition, columns (2)–(3) control for province-specific fixed effects and other key firm characteristics, such as firm TFP, log employment of the firm, foreign indicator and export indicator. The key coefficient, the SOE indicator, is always negative and statistically significant, when we use the short-term liabilities to calculate the interest rate (in columns 1 and 2). It is possible that SOEs may have higher long-term liabilities than private firms. Thus, using the short-term liabilities as the denominator to measure interest rate might underestimate the interest rate of the SOEs. To address this concern, column (3) of Table 4.7 measures the firm-level interest rate using the ratio of interest payment to total liabilities, which include both short-term loans and long-term loans. It turns out that the SOE indicator is still negative and statistically significant under this alternative specification of firm-level interest rate. Its absolute magnitude is around 0.09, suggesting that private firms pay annual interest rate 9% higher than SOEs, and hence bear higher capital cost than SOEs.36

35

We find similar results when SOEs are measured in a broad way a la Hsieh and Song (2015). The relative interest rate differential between SOEs and private firms is plausible if the firms’ informal finance is taken into account. Private firms usually have to finance their working capital from informal financial markets due to severe credit constraints (see Lardy 2014). Our magnitude

36

2000–13 208,320

1,119,454

Number of Obs

1,119,446

1,136,049

No No

2000–08

Yes No

Year coverage

No

Yes

Yes

No

Yes

No Yes

No

Yes

Yes Yes

Province-specific fixed effects

Yes

Industry-specific fixed effects

Yes Yes

−9.39*** (−4.43)

City-specific fixed effects

No

Yes

Other firm factors controls

(−2.28)

−0.089**

−0.174*** (−3.34)

−0.134*

(−1.90)

157,810

No

No

Yes

Yes

Yes

(−3.02)

−6.78***

(5)

Firm-level unit land price (4)

(2)

(3)

Measured firm interest rates

(1)

Year-specific fixed effects

One-year Lag of SOE intensity

One-year Lag of SOE indicator

SOE intensity

SOE indicator

Regressand

Table 4.7 Distortions in input factors markets

103,826

No

No

Yes

Yes

Yes

(−4.47)

−11.43***

(6)

1,489

2000–08

Yes

No

Yes

Yes

No

(−1.84)

−54.53*

(continued)

1,306

Yes

No

Yes

Yes

No

(−1.67)

−48.97*

(8)

City-level unit land price (7)

4.4 Evidence 109

(2) 0.01

(1)

0.01

Measured firm interest rates 0.01

(3) 0.07

(4) 0.07

(5)

Firm-level unit land price 0.08

(6) 0.13

(7) 0.15

(8)

City-level unit land price

Note Columns (1)–(3) and (7)–(8) cover the period of 2000–08 whereas columns (4)–(6) cover the period of 2000–13. The regressand in columns (1)–(3) is the firm-level interest rate calculated as the ratio of firm interest expenses to current liabilities in columns (1) and (2) and to total liabilities in columns (3). The regressand in columns (4)–(6) is the firm-level price of land purchased from the government. This is defined as the ratio of the firm’s total spending on land acquisition to the area of land it purchases. The regressand in columns (7)–(8) is the prefectural city-level price of land purchased by firms from the government. This is defined as the ratio of government’s total land revenue to its land area in each prefectural city. The SOE intensity in columns (7)–(8) is defined as the number of SOEs divided by the total number of manufacturing firms within each prefectural city. All columns control for year-specific and industry-specific fixed effects, respectively. Column (2) and (3) add other controls of firm-level characteristics such as log firm labor, foreign indicator, and export dummy, and province-specific fixed effects. Columns (5)–(6) add other controls of firm-level characteristics such as firm’s capital-labor ratio, foreign indicator, and export dummy. Columns (6) uses the lag of SOE indicator whereas column (8) uses the lag of SOE intensity in the regressions. The numbers in parentheses are t-values. *** (**, *) denotes significance at the 1% (5, 10%) level

R-squared

Regressand

Table 4.7 (continued)

110 4 Outward FDI and Domestic Input Distortions …

4.4 Evidence

111

Still, one may have a concern that private firms could use domestic credits to finance costs associated with FDI projects. If so, they would still face discrimination even when investing abroad. Admittedly, we cannot rule out this possibility without further information on firm’s credit allocation. However, a better investigation is to check whether private firms acquire land at a higher unit cost than SOEs. If so, whether the land market distortion plays a role when private firms engage in outward FDI. Thus, columns (4)–(8) in Table 4.7 go further to check whether SOEs acquire land at lower costs. By controlling for industry-specific and year-specific fixed effects, respectively, we first regress firm-level unit land price on the SOE indicator over the period of 2000–13 in column (4). Columns (5) and (6) add other firm-level controls such as firm TFP, log employment, the foreign indicator, and the export indicator. Since there may be a concern that land market discrimination could reversely induce firm churning (i.e., switching from private firms to SOEs or vice versa), column (6) regresses the unit land price on the one-year lag of SOE indicator to avoid possible simultaneous bias. The regression results in Table 4.7 show that the coefficient of SOE indicator is always negatively significant. Thus, private firms do seem to pay higher unit land price than SOEs. Compared with the interest rate estimates in columns (1)–(3), the numbers of observations drop dramatically in columns (4)–(6) due to the usual imperfect matching between the production data set and the land price data set. To overcome such a problem, we use the prefectural-level data of land purchase for robustness checks.37 We first construct a variable of SOE intensity, which is defined as the number of SOEs divided by the number of total manufacturing firms in the city. Our theory predicts that a city with a higher SOE intensity is expected to have a lower average price of land, conditioning on other prefecture-level characteristics. Estimations whose results are reported in columns (7) and (8) regress average land price at the prefectural city level on SOE intensity. Particularly, columns (7) and (8) control for both year-specific and city-specific fixed-effects, respectively. Still, it is possible that aggregate demand for land in each city affects the price of land in the city. Columns (7) and (8) thus control for cities’ total land sales. In all cases, the coefficient of SOE intensity is negative and statistically significant, suggesting that SOEs pay lower unit land price on average and hence bear lower land costs than private firms.

is close to Song et al. (2011) and Midrigan and Xu (2014), which find that private firms pay roughly 9–10% higher interest rate than SOEs. 37 Data are from China’s Land and Resources Statistical Yearbook (various years). As in Tian et al. (2016), we only use data on land sales for land that is sold or granted by market channels, including agreement, auction, bidding and listing. We exclude land transfers to SOEs through direct government leasing and allocation. Thus, the coefficients in the last two estimates in Table 7 shall be understood as the lower bound of the measured distortion.

112

4 Outward FDI and Domestic Input Distortions …

4.4.3 Channels and Sectoral Heterogeneity Part 1 of Proposition 4.2 hints that the selection reversal is heterogeneous across sectors. Specifically, the productivity premium of state-owned MNCs will be more pronounced in industries with severe distortions. Similarly, we should observe that SOEs are less likely to produce abroad in these sectors compared with SOEs in sectors with less severe distortions. To verify these predictions, we regress the firm outward FDI indicator on the SOE indicator and its interaction with the industry-level input price differential that is first measured by the interest rate differential, followed by the difference in the unit land price. For this difference-in-differences regression, our model predicts a negatively significant coefficient of the SOE dummy and of its interaction with the (positive) industry-level input price differential. The economic rationale is evident: compared with SOEs in other industries, those in industries with more severe credit and/or land market distortions against private firms are less likely to undertake outward FDI projects, as they receive even better treatments in the domestic input markets than their counterparts in other sectors. Indeed, the interaction between SOE and industrial input price differential is crucial to testing our hypothesis, as it shows the effect of input price distortions on the likelihood of SOEs’ investing abroad directly. To check whether the credit market plays an important role in shaping the pattern of a firm’s selection reversal, columns (1)–(6) of Table 4.8 report the estimation results using the interest rate to measure the input price. The industrial input price differential here is calculated as the difference between the (high-level) average interest rate paid by private firms and the (low-level) average interest rate paid by SOEs within the same industry-year cell. After controlling for industry and year fixed effects, separately, and other key firm-level variables, we find that the SOE indicator and its interaction term with the industry-level interest rate differential are both negative and statistically significant. These results suggest that the credit market distortion is an important reason why SOEs are less likely to go abroad. Such a key finding is robust to different specifications. Particularly, we drop the sample of FIEs in column (2) and the sample of outward FDI to tax haven destinations in column (3). We also drop the sample of distribution-oriented FDI in columns (4) and (5) and narrow the time window to 2004–8 in column (6).38 All the estimation results suggest that the credit market distortion is an important factor for us to interpret the pattern of Chinese firms’ selection reversal. As mentioned above, one may worry that private firms still suffer from domestic credit market distortions if they use domestic credits to finance FDI fixed cost and/or variable cost. To mitigate such a concern, we go further to examine the channel of land market distortions, which would be a cleaner experiment as firms cannot move their domestic land input abroad. To do so, we first obtain firms’ total cost of acquiring land during the sample period of 2000–13, which is time-invariant, as firms could acquire land unevenly 38

In column (5), we use the ratio of firm’s interest payment to its total liabilities to calculate firm-level interest rate.

Yes

Yes

Yes

Foreign firms included

Tax haven included

Distribution FDI Yes Included

2000–08

Yes

Yes

Industry fixed effects

No

Yes

Yes

Yes

Yes

Yes

(1.56)

(0.54)

Year fixed effects

0.057

0.019

Ind. input price diff

Other controls

(−2.41)

(-1.69)

(−4.22)

Ind. input price diff

(−2.33)

−0.886**

Year coverage

(3)

(4)

(5)

(6)

(7)

(8)

Firm-level land price (9)

(10)

(11)

Yes

No

No

Yes

Yes

Yes

(1.84)

0.079*

(−1.90)

−0.948*

(−6.36)

(−1.97)

(−2.11)

No

Yes

No

Yes

Yes

Yes

(1.30)

0.090

(−2.19)

No

Yes

No

Yes

Yes

Yes

(0.80.0

0.127

(−2.10)

2004–08

Yes

Yes

Yes

Yes

Yes

Yes

(0.53)

0.019

(−1.89)

−1.033** −0.817** −0.767*

(−2.09)

2000–13

Yes

Yes

Yes

Yes

Yes

Yes

(1.35)

0.001

(−2.10)

−0.003**

(−7.67)

(−10.46)

(−6.61)

(−6.60)

Yes

No

Yes

Yes

Yes

Yes

(3.27)

0.001***

(−4.01)

Yes

No

No

Yes

Yes

Yes

(3.40)

0.001***

(−4.35)

2000–08

Yes

No

No

Yes

Yes

Yes

(2.34)

0.001**

(−2.76)

(continued)

2004–08

Yes

No

No

Yes

Yes

Yes

(2.24)

0.001*

(−2.36)

−0.006*** −0.006*** −0.006*** −0.006***

(−8.42)

−0.264** −0.488*** −0.994*** −0.290** −0.250** −0.254** −0.638*** −1.069*** −1.343*** −1.850*** −1.855***

SOE indicator × −0.638*

SOE indicator

(2)

Measured firm-level interest rate

Regressand: FDI (1) indicator

Measure of input price

Table 4.8 Logit estimates on channels

4.4 Evidence 113

1,121,845 879,003

873,150

(3) 829,655

(4) 832,741

(5) 883,712

(6) 2,278,062

(7) 2,200,723

(8)

Firm-level land price

1,750,939

(9) 1,005,294

(10)

739,082

(11)

Note The regressand is the FDI indicator. The numbers in parentheses are t-values. *** (**) denotes significance at the 1% (5%) level. Input prices in columns (1)–(6) are measured by firm-level interest rate whereas those in columns (7)–(11) are firm-level unit land price. The measured interest rate is calculated as the ratio of firm’s interest expenses to current liabilities in columns (1)–(4) and (6) and to total liabilities in column (5). In particular, the industry interest rate (or unit land price) differential (i.e., Ind. Input Price Diff.) is measured by the average industry-level interest rate (unit land price) paid by private firms minus that paid by SOEs in each 3-digit CIC industries by year. Columns (1)–(5) and (9)–(11) drop foreign firms. Columns (3) and (8)–(11) drop FDI to tax haven destination countries. Column (4) drops distribution FDI. All columns include other firm-level controls such as firm TFP (in columns (1)–(6) only), log employment and export indicator. All regressions include industry-specific fixed-effects and year-specific fixed-effects whereas column (11) even controls the industry-year specific fixed-effects

Observations

(2)

Measured firm-level interest rate

Regressand: FDI (1) indicator

Measure of input price

Table 4.8 (continued)

114 4 Outward FDI and Domestic Input Distortions …

4.4 Evidence

115

and infrequently across years. We then calculate the three-digit CIC industry-level difference in the unit land price paid by private firms and by SOEs. Both its own term (of the industrial land price differential) and its interacted term with the SOE indicator are included in columns (7)–(11) of Table 4.8. Similar to columns (1)–(6), columns (7)–(11) include important firm-level control variables such as firm’s log employment and the export dummy. Columns (7)–(9) cover the entire sample period of 2000–13. Columns (8)–(11) drop observations with tax-haven FDI destination countries, and columns (9)–(11) drop foreign firms from the regressions. Column (10) includes the sample of 2000–8 so that it can be directly compared with column (3) with the measure of interest rates.39 We also control for industry and year fixed effects respectively in all regressions. In particular, we include the interacted industry-year fixed-effects in column (11) of Table 4.8. All the specifications in columns (7)–(11) of Table 4.8 yield the same findings: SOEs are less likely to engage in outward FDI. More important, SOEs in industries with more severe land market distortions are less likely to undertake outward FDI than SOEs in industries with less severe land market distortions (i.e., the negative interaction term). This is the key evidence we provide for our explanation for the selection reversal pattern.

4.4.4 Capital Intensity and Pattern of Outward FDI Part 2 of Proposition 4.2 implies that, compared with private firms, SOEs are less likely to engage in outward FDI in capital-intensive industries. This subsection provides evidence for this prediction. By definition, firms in capital-intensive industries have higher demand for working capital. Accordingly, domestic input distortions against private firms favor SOEs more in such industries. If domestic input distortions are the fundamental driving force for explaining the behaviour of Chinese firms’ outward FDI, SOEs in capital-intensive industries should be unlikely to undertake outward FDI. By contrast, such a phenomenon may not exist in labour-intensive industries. To check this out, we run the following difference-in-differences regression and focus on the difference between capital-intensive sectors and labour-intensive sectors. Specifically, we interact the dummy variable for the firm’s being in the capitalintensive sector and the dummy variable for the firm’s being in the labour-intensive sector with the SOE dummy in our Logit regressions.40 The interacted coefficient between the SOE indicator and the labour-intensive indicator shows how the state 39

The number of observations in column (10) is substantially higher than that in column (3), as the land price differential used in column (10) is time-invariant and hence exists in all three-digit CIC industries whereas the interest rate used in column (3) is time-variant and some industry-year pairs have missing observations. 40 Note that our specification is the same as the specification of O F D I = β S O E + β S O E × it 1 it 2 it K dummy + . . ., since it is equal to O F D Iit = (β1 + β2 )S O E it × K dummy + β1 S O E it × (1 − K dummy) where K dummy denotes capital intensive indicator, which equals one minus

116

4 Outward FDI and Domestic Input Distortions …

ownership affects the likelihood of investing abroad for the parent firms from labour intensive industries. Similarly, the interacted coefficient between SOE indicator and labour-intensive shows how the state ownership affects the likelihood of investing abroad for the parent firms from capital intensive industries. Column (1) in Table 4.9 shows that SOEs are less likely to engage in outward FDI in capital-intensive industries. By contrast, the SOE indicator is insignificant for parent firms from labourintensive sectors. This finding is robust to different specifications, such as dropping foreign firms, dropping SOE switching firms, dropping FDI to tax haven destinations or using a shorter time period (2004–8). It is worth discussing why the key coefficient of the SOE indicator is insignificant in labour-intensive sectors. In China, the cost of labour has increased dramatically Table 4.9 Logit estimates by sectors Sectoral category

2000–8

2004–8

Regressand: FDI indicator

(1)

(2)

(3)

(4)

(5)

SOE indicator X

−0.276

−0.290

−0.690

−0.180

−0.170

Labor-intensive indicator

(−0.68)

(−0.70)

(−1.36)

(−0.44)

(−0.35)

SOE indicator X

−0.475*

−0.754*** −1.257*** −0.834*** −0.646**

Capital-intensive indicator

(−1.73)

(−2.70)

(−2.98)

(−3.09)

(−2.35)

Firm relative TFP

1.305*

1.837**

1.151*

2.328***

2.069***

(1.81)

(2.25)

(1.66)

(4.20)

(3.82)

Log firm labour

0.582

0.587

0.565

0.570

0.539

(11.18)

(8.84)

(7.85)

(9.61)

(19.46)

Export indicator

0.896

1.146

1.167***

1.152

1.297***

(4.44)

(6.03)

(5.43)

(5.93)

(18.71)

Year fixed effects

Yes

Yes

Yes

Yes

Yes

Industry fixed effects

Yes

Yes

Yes

Yes

Yes

Foreign firms dropped

No

Yes

Yes

Yes

No

Tax haven destinations dropped No

No

Yes

No

No

SOE switching firms dropped

No

No

No

No

Yes

Observations

1, 135,468 895,210

894,816

707,154

554,768

Notes The regressand is the FDI indicator. All columns include industry-specific fixed effects and year-specific fixed effects. The numbers in parentheses are t-values clustered at the firm level. *** (**) denotes significance at the 1% (5%) level. Columns (1)–(3) cover observations during the years 2000–8, whereas columns (4)–(5) cover observations during the years 2004–8. Column (1) keeps foreign invested firms whereas the other columns drop foreign invested firms. Column (3) drops outward FDI to tax-haven regions. Column (5) drops SOE switching firms. The relative TFP in columns (1)–(5) are measured by augmented Olley–Pakes controlling for input price distortion and SOE status labour-intensive sectors indicator equals one if the firm’s Chinese industrial classification is higher than 20 and zero otherwise

labour-intensive indicator. As we don’t include parent-firm fixed effects in the regression, the perfect collinearity problem does not arise here.

4.4 Evidence

117

since 2004. Accordingly, some firms in labour-intensive sectors established foreign affiliates in other least-cost, labour-abundant countries, such as Bangladesh, Ethiopia and Vietnam. Such firms sought global sourcing instead of global markets (Antràs, 2016), which is out of the scope of the current chapter.

4.4.5 Estimates at the Intensive Margin We now provide evidence for Proposition 4.3. Table 4.5 provides evidence for part 1 of the proposition. Part 2 of Proposition 4.3 states that the ratio of foreign sales to domestic sales is higher for private MNCs than for state-owned MNCs. Data on sales of foreign affiliates are unavailable in the Chinese firm-level ASIF data set. Thus, we merged the ASIF data set with the Orbis data set, which contains information on sales and revenue of (domestic and foreign) affiliates of Chinese MNCs. Unfortunately, the matching rate for the two data sets between 2005 and 2010 is extremely low, as we merge our ASIF data with the Orbis data by matching firms’ names in Chinese.41 For the data between 2011 and 2013, we merge our ASIF data with the Orbis data using the trade registration number of the (Chinese) parent firms, as this information is contained in both data sets. In Orbis data, 11,000 observations (i.e., affiliateyear pairs) have non-missing values for sales, revenue, and employment. Among these affiliates, roughly 30% of them are in manufacturing sectors. Since we can only identify FDI starters (between 2011 and 2013) in our outward FDI data set, we managed to match roughly 750 affiliate-year observations for the two data sets between 2011 and 2013. The data between 2011 and 2013 are the data we use to implement following empirical analysis. Table 4.10 regresses log sales (or log revenue) of each domestic or foreign affiliate of Chinese MNCs in a given year on a dummy variable for being a private (parent) firm, a dummy variable for being a foreign affiliate, and characteristics of the parent firm. Importantly, we add an interaction term between the two dummy variables: Privatei,t × Foreign j,t where i, j and t refer to parent firm (private or state-owned), affiliate and year respectively. As expected, the regression results show that private parent firms have smaller affiliates on average, and foreign affiliates are smaller than domestic affiliates on average. What is interesting is that the size difference between the domestic affiliate and the foreign affiliate (of the same parent firm) is smaller among private MNCs than among state-owned MNCs, as the coefficient of Privatei,t × Foreign j,t is positively significant. This is exactly what part 2 of 41

The main reason for the low matching rate is that firms’ names are in Chinese in our ASIF data, while they are in English in Orbis data. As English translations of a firm’s Chinese name can be multiple, it is extremely challenging to match observations from the two independent data sets (e.g., the company whose English name is Lenovo currently should be translated into ‘Legend’, based on its Chinese name). We identify a matched observation, when the English translation of the firm’s name exactly matches the characters of its Chinese name, the non-matched firms are probably random and should not affect our empirical results.

−2.323*** (−2.94)

−2.518***

(−3.81)

par ent

Yes Yes Yes Yes 733

Yes

Yes

Yes

Yes

713

Year fixed effects

Country fixed effects

Industry fixed effects

Observations

par ent

Parent firm fixed effects

Log current liabilityparent

Log exportsparent

Log total assetsparent (10.05)

(−5.48)

(−7.35)

(11.39)

−1.405***

−1.546***

Private indicator

0.906***

(7.91)

(9.92)

Foreign indicator

0.855***

1.103***

1.223***

Private indicator ×

par ent

log(revenue)

log(sales)

Foreign indicator

(2)

(1)

Regressand

586

Yes

Yes

Yes

604

Yes

Yes

Yes

Yes

(0.44)

Yes

0.0767

(1.04)

(6.25)

0.901***

(−6.82)

−4.190***

(−3.38)

−1.380***

(5.78)

1.111***

log(revenue)

(4)

0.181

(7.48)

0.802***

(−8.95)

−4.362***

(−5.01)

−1.567***

(7.51)

1.338***

log(sales)

(3)

Table 4.10 Ratio of foreign sales to domestic sales is higher for private MNCs

678

Yes

Yes

Yes

698

Yes

Yes

Yes

Yes

(1.02)

(1.80) Yes

0.160

(4.82)

0.749***

(−2.64)

−2.013***

(−5.03)

−1.356***

(7.57)

1.088***

log(revenue)

(6)

0.168*

(7.76)

0.688***

(−3.55)

−2.217***

(−6.97)

−1.496***

(9.72)

1.213***

log(sales)

(5)

559

Yes

Yes

Yes

Yes

(1.38)

0.131

(0.79)

0.133

(5.84)

0.670***

(−7.82)

−4.032***

(−5.12)

−1.541***

(7.65)

1.326***

log(sales)

(7)

(continued)

577

Yes

Yes

Yes

Yes

(0.74)

0.129

(0.16)

0.0283

(4.34)

0.777***

(−5.66)

−3.816***

(−3.30)

−1.342***

(5.58)

1.088***

log(revenue)

(8)

118 4 Outward FDI and Domestic Input Distortions …

(2) 0.875

(1)

0.896

0.897

(3) 0.865

(4) 0.903

(5) 0.881

(6) 0.900

(7) 0.868

(8)

Note Observation are affiliate-year pairs between 2012 and 2014, and log(sales) and log(revenue) are log sales and log revenue of each (domestic or foreign) affiliate of Chinese MNCs in a given year. Orbis data of affiliates between 2012 and 2014 are merged to ASIF data of parent firms between 2011 and 2013 (i.e., one year lag). Specifically, we merge our ASIF data with the ORBIS data using (Chinese) parent firms’ trade registration number (in China) whose information is contained by both data sets after 2011. Both MNC starters and incumbents (after 2011) are included into the regression. Standard errors are clustered at the parent firm level. *** (**, *) denotes significance at the 1% (5, 10%) level, and t statistics are reported in parentheses

R-squared

Table 4.10 (continued)

4.4 Evidence 119

120

4 Outward FDI and Domestic Input Distortions …

Proposition 4.3 predicts: The ratio of foreign sales to domestic sales is higher for private MNCs than for state-owned MNCs.

4.4.6 Outward FDI Data Between 2000 and 2013 In this subsection, we expand the time horizon of our sample to 2000–13. The major reason is the number of state-owned MNCs is small before 2008, as the total number of manufacturing outward FDI projects is not too big before 2008. As there are many more manufacturing-outward FDI projects after 2008, the inclusion of outward FDI data until 2013 can alleviate the concern that the small number of state-owned MNCs might affect our estimation results.42 Moreover, the quality of our data matching becomes better after 2010, and the importance of China’s manufacturing outward FDI (in China’s total outward FDI) is increasing with time. Therefore, using outward FDI data after 2009 provides crucial robustness checks for our previous empirical findings. However, the drawback of using the longer time-series data set is that we cannot estimate firm productivity accurately. The reason is that China’s firmlevel production data report neither value added nor purchase of intermediate inputs after 2008. Without knowing these key variables, we cannot precisely estimate TFP or calculate labour productivity (i.e., value added per worker). Because of these substantial restrictions on the data, we do not use the data for 2000–13 as our main data set. Instead, we use the longer time-series data for robustness checks only and relegate detailed discussions into Online Appendix E. The empirical findings based on the sample from 2000 to 2013 are qualitatively the same as our previous findings. In particular, we still find that SOEs are less likely to invest abroad after we control for important firm-level characteristics and a variety of fixed effects. In addition, this pattern is more pronounced in sectors that have larger interest rate differentials (between SOEs and private firms) and in sectors that are more capital-intensive. In total, both the short sample and the long sample of our outward FDI data lead to the same empirical findings emphasised in this chapter.

4.5 Concluding Remarks In this study, we utilise data on Chinese MNCs to investigate how distortions (i.e., discrimination against private firms) in the domestic market affect firms’ FDI decisions. We document three puzzling stylised facts. First, private MNCs are less productive than state-owned MNCs, although private non-MNCs are more productive than state-owned non-MNCs. Second, SOEs are less likely to undertake FDI, although they are larger and receive various supports from the government for investing abroad. 42

We have taken into account this issue in Table 4.5 and Table 5 of Online Appendix by using the rare event Logit regression.

References

121

Third, the relative size of state-owned MNCs (compared with that of non-exporting firms) is larger than that of private MNCs. We then build a model to rationalise these findings and highlight a key channel through which distortions affect firms’ FDI decisions. Distortions in the domestic market incentives private firms to invest and produce abroad, which results in less tough selection into the FDI market for them. In addition, compared with stateowned MNCs, private MNCs allocate output disproportionately more in the foreign market, and their size increases disproportionately when they become MNCs. Finally, the selection reversal and productivity premium for state-owned MNCs are more pronounced in capital-intensive industries and in industries with more severe discrimination against private firms. All the empirical predictions of the model receive support from the data. We argue that alternative hypotheses rather than the input market distortion cannot be used to rationalise our finding of selection reversal. First, a price wedge in the domestic product market (between private firms and SOEs) would lead to no difference in the selection into FDI market between the two types of firms, as the cost and benefit of doing FDI would be the same (for the two types of firms) in this case. Next, profit shifting motives of private firms cannot be used to rationalise the selection reversal, as our result holds for the sub-sample that drops Chinese FDI projects into tax-haven economies. Finally, technology-seeking and brand-seeking motives of private firms cannot be used to rationalise the selection reversal, as our result holds for the sub-sample that drops M&A-type FDI.43

References Antràs, P. (2003). Firms, contracts, and trade structure. The Quarterly Journal of Economics, 118(4), 1375–1418. Antràs, P. (2005). Incomplete contracts and the product cycle. American Economic Review, 95(4), 1054–1073. Antras, P., & Helpman, E. (2004). Global sourcing. Journal of Political Economy, 112(3), 552–580. Arkolakis, C. (2010). Market penetration costs and the new consumers margin in international trade. Journal of Political Economy, 118(6), 1151–1199. Bai, X., Krishna, K., & Ma, H. (2017). How you export matters: Export mode, learning and productivity in China. Journal of International Economics, 104, 122–137. Bernard, A., Jensen, J. B., & Schott, P. (2003). Falling trade costs, heterogeneous firms, and industry dynamics. NBER Working Papers, 167(4), 613–626. Bernard, A. B., Redding, S. J., & Schott, P. K. (2007). Comparative advantage and heterogeneous firms. The Review of Economic Studies, 74(1), 31–66. Brainard, S. L. (1994). An empirical assessment of the proximity-concentration tradeoff between multinational sales and trade. American Economic Review, 87(4). Brandt, L., Van Biesebroeck, J., & Zhang, Y. (2012). Creative accounting or creative destruction? Firm-level productivity growth in Chinese manufacturing. Journal of Development Economics, 97(2), 339–351. 43

We discuss the role played by the fixed FDI cost in our model and relegate the proof of the result to Online Appendix D.

122

4 Outward FDI and Domestic Input Distortions …

Brandt, L., Tombe, T., & Zhu, X. (2013). Factor market distortions across time, space and sectors in China. Review of Economic Dynamics, 16(1), 39–58. De Loecker, J., & Warzynski, F. (2012). Markups and firm-level export status. American Economic Review, 102(6), 2437–2471. Dixit, A. K., & Stiglitz, J. E. (1977). Monopolistic competition and optimum product diversity. The American Economic Review, 67(3), 297–308. Dollar, D., & Wei, S. J. (2007). Das (wasted) kapital: Firm ownership and investment efficiency in china. NBER Working Papers. Feenstra, R. C., & Hanson, G. H. (2005). Ownership and control in outsourcing to China: Estimating the property-rights theory of the firm. The Quarterly Journal of Economics, 120(2), 729–761. Feenstra, R. C., Li, Z., & Yu, M. (2014). Exports and credit constraints under incomplete information: Theory and evidence from China. Review of Economics and Statistics, 96(4), 729–744. Garetto, S., Oldenski, L., & Ramondo, N. (2017). Life-cycle dynamics and the expansion strategies of US multinational firms. Garicano, L., Lelarge, C., & Van Reenen, J. (2016). Firm size distortions and the productivity distribution: Evidence from France. American Economic Review, 106(11), 3439–3479. Gopinath, G., Kalemli-Özcan, S, ¸ Karabarbounis, L., & Villegas-Sanchez, C. (2017). Capital allocation and productivity in South Europe. The Quarterly Journal of Economics, 132(4), 1915–1967. Groizard, J. L., Ranjan, P., & Rodriguez-Lopez, A. (2015). Trade costs and job flows: Evidence from establishment-level data. Economic Inquiry, 53(1), 173–204. Guner, N., Ventura, G., & Xu, Y. (2008). Macroeconomic implications of size-dependent policies. Review of Economic Dynamics, 11(4), 721–744. Helpman, E. (1984). A simple theory of international trade with multinational corporations. Journal of Political Economy, 92(3), 451–471. Helpman, E., Melitz, M. J., & Yeaple, S. R. (2004). Export versus FDI with heterogeneous firms. American Economic Review, 94(1), 300–316. Hsieh, C. T., & Klenow, P. J. (2009). Misallocation and manufacturing TFP in China and India. The Quarterly Journal of Economics, 124(4), 1403–1448. Hsieh, C. T., & Klenow, P. J. (2014). The life cycle of plants in India and Mexico. The Quarterly Journal of Economics, 129(3), 1035–1084. Hsieh, C. T., & Song, Z. M. (2015). Grasp the large, let go of the small: The transformation of the state sector in China (No. w21006). National Bureau of Economic Research. Huang, Y. (2003). Selling China: Foreign direct investment during the reform era. Cambridge University Press. Huang, Y. (2008). Capitalism with Chinese characteristics: Entrepreneurship and the state. Cambridge University Press. Khandelwal, A. K., Schott, P. K., & Wei, S. J. (2013). Trade liberalization and embedded institutional reform: Evidence from Chinese exporters. American Economic Review, 103(6), 2169–2195. King, G., & Zeng, L. (2001). Logistic regression in rare events data. Political Analysis, 9(2), 137– 163. Lardy, N. R. (2014). Markets over Mao: The rise of private business in China. Columbia University Press. Levinsohn, J., & Petrin, A. (2003). Estimating production functions using inputs to control for unobservables. The Review of Economic Studies, 70(2), 317–341. Manova, K., Wei, S. J., & Zhang, Z. (2015). Firm exports and multinational activity under credit constraints. Review of Economics and Statistics, 97(3), 574–588. Markusen, J. R. (1984). Multinationals, multi-plant economies, and the gains from trade. Journal of International Economics, 16(3–4), 205–226. Melitz, M. J. (2003). The impact of trade on intra-industry reallocations and aggregate industry productivity. Econometrica, 71(6), 1695–1725. Midrigan, V., & Xu, D. Y. (2014). Finance and misallocation: Evidence from plant-level data. American Economic Review, 104(2), 422–458.

References

123

Moll, B. (2014). Productivity losses from financial frictions: Can self-financing undo capital misallocation? American Economic Review, 104(10), 3186–3221. Restuccia, D., & Rogerson, R. (2008). Policy distortions and aggregate productivity with heterogeneous establishments. Review of Economic Dynamics, 11(4), 707–720. Restuccia, D., & Rogerson, R. (2013). Misallocation and productivity. Review of Economic Dynamics, 16(1), 1–10. Song, Z., Storesletten, K., & Zilibotti, F. (2011). Growing like China. American Economic Review, 101(1), 196–233. Tian, W., & Yu, M. (2015). Processing trade, export intensity, and input trade liberalization: Evidence from Chinese firms. Journal of the Asia Pacific Economy, 20(3), 444–464. Tian, W., Sheng, L., & Zhao, H. (2016). State capitalism: A new perspective on land sale in China. Pacific Economic Review, 21(1), 84–101. Yu, M. (2015). Processing trade, tariff reductions and firm productivity: Evidence from Chinese firms. The Economic Journal, 125(585), 943–988.

Chapter 5

Does Outward FDI Generate Higher Productivity for Emerging Economy MNEs?—Micro-level Evidence from Chinese Manufacturing Firms

Abstract This chapter investigates whether emerging economy multinational enterprises (EMEs) that undertake outward foreign direct investment (OFDI) become more productive, controlling for the self-selection into the global investment market. Particularly, we focus on the moderating effects of firm heterogeneity on the OFDI-productivity nexus. A theoretical framework incorporating the resource-based views and institutional theory is established and the propensity-score matching and difference-in-difference (DID) approaches are combined to test the framework, utilizing unique data on Chinese manufacturing firms over the sample period 2002– 2008. We find that EMEs turn to be generally more productive after they conduct OFDI, but this productivity effect varies depending on the parent firm and investment strategy heterogeneity. Our results suggest that EMEs without state ownership but with stronger absorptive capability gain higher and more sustainable productivity effects and such gains are higher for EMEs investing in OECD than in non-OECD countries. Policy and managerial implications are discussed.

5.1 Introduction As an indicator of efficiency, productivity1 has been argued to be a determinant of firms’ survival and sustained competitiveness (Lieberman & Dhawan, 2005; Syverson, 2011) and is crucial for emerging economies to catch up with the rest of the world (Kharas & Kohli, 2011). So far, the research into the productivity differences across firms has come a long way (Bartelsman & Doms, 2000), and the discovery of persistent, large and ubiquitous productivity variations across businesses has shaped the agenda of a couple of research fields seeking to identify the factors affecting

1 Firm productivity is a component of a country’s production efficiency, which plays an essential role in shaping a country’s GDP growth. Therefore, we choose firm productivity as our dependent variable, to some extent, to shed some light on a country’s growth.

This chapter is published in International Business Review by Li Linjie, Xiaming Liu, Dong Yuan, Miaojie Yu. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022 W. Tian and M. Yu, Outward Foreign Direct Investment of Chinese Enterprises, Contributions to Economics, https://doi.org/10.1007/978-981-19-4719-3_5

125

126

5 Does Outward FDI Generate Higher Productivity for Emerging …

productivity, especially the levers that firms can utilize to increase their productivity (Bertrand & Capron, 2015; Syverson, 2011). Among those levers, MNEs’ outward FDI (OFDI) has been touched upon as a mechanism by which firms can not only exploit ownership advantages, but also access new resources, realize resource reallocation stimulate competition, and enhance productivity (Frost, 2001; Bertrand & Capron, 2015). Emerging economy MNEs (EMEs) are believed to be able to gain more productivity premium from OFDI, as they are based in less innovative developing institutions, possesses less knowledge competencies, and thus have more learning opportunities (Buckley et al., 2007). Given the dramatic growth of OFDI flows from emerging economies, it is critical for both scholars and EME managers to know whether there exists OFDI-led productivity growth for EMEs, and conditions under which an EME can gain more OFDI-led productivity benefits (Chen et al., 2012; Li et al., 2016). Yet, as a sizable literature has gone to the impact of OFDI on employment, exports, investment, and productivity in developed economies (Bitzer & Kerekes, 2008; Chen & Yang, 2013; Chuang & Lin, 1999; De La Potterie & Lichtenberg, 2001; Herzer, 2008, 2010, 2011; Kogut & Chang, 1991; Pradhan & Singh, 2008), studies on the crucial OFDI-productivity link in EMEs are very limited (Herzer, 2011; Zhao et al., 2010). Even though some studies have referred to this linkage, their results are inconsistent (Chen & Tang, 2014; Chen et al., 2012; Cozza et al., 2015; Lee et al., 2013; Li et al., 2016; Masso & Vahter, 2008; Yang et al., 2013; Zhao et al., 2010), probably because of a lack of careful consideration of the moderating effects of firm-level heterogeneity (Herzer, 2011), proper control for the endogenous self-selection bias and suitable productivity measurement techniques (De Loecker, 2007; Hijzen et al., 2007, 2011). This study aims to address the above-mentioned research gaps and will contribute to the existing literature in the following ways. Firstly, given EMEs’ lack of capabilities and their strong resource-dependence on home country governments (Buckley et al., 2007, 2008; Deng, 2007; Ramasamy et al., 2012; Wang et al., 2012a, 2012b), we develop a novel theoretical framework which incorporates both the resource-based view (RBV) and institutional theory (IT) to explain the mechanisms for EMEs’ OFDIproductivity nexus. Empirical studies about the impact of OFDI tend to be based on a general literature review or “international business theory”. To our best knowledge, this is the very first study that looks at EMEs’ productivity gains from OFDI at the firm level in line with an analytical framework explicitly incorporating both resource- and institution-based lenses. Compared with developed economy MNEs, EMEs are more resource seeking, and are strongly influenced by their home country institutions. As a result, an application of both RBV and IT to analyse OFDI by EMEs would be more appropriate. Secondly, our study contributes by explaining and testing whether and how firm heterogeneity in terms of state ownership, absorptive capacity and internationalization strategy moderates OFDI’s productivity effects on EMEs. The nexus between OFDI and productivity is complex and the productivity effect of OFDI is by no means automatic (Kokko & Kravtsova, 2008). Suggested by Helpman et al. (2003),

5.1 Introduction

127

firm heterogeneity drives their diversity in strategy and performance. In line with both institutional and resource-based perspectives, our study contributes to existing literature by identifying state ownership, absorptive capacity and internationalization strategy as three important moderators and explaining the mechanisms with which these moderators affect the OFDI-led productivity growth nexus (Bertrand & Capron, 2015; Choudhury & Khanna, 2014; Cohen & Levinthal, 1990; Cui & Jiang, 2012; Wang et al., 2012a, 2012b; Zahra & George, 2002). Methodologically, we augment Olley and Pakes’ (1992) semi-parametric approach to measure total factor productivity (TFP), via introducing the OFDI dummy and export dummy in the production function, allowing for various production estimation functions for EMEs with different OFDI and export status. This enables us to not only efficiently control for the possible simultaneity and selection biases (Olley & Pakes, 1992), but also successfully remove the potential productivity estimation bias from omitting influential variables in the production function estimation. In addition, a method combing the propensity-score matching and differencein-difference (DID) approaches will be employed, to examine the ‘ real’ OFDI-led productivity change for EMEs via careful control for the possible endogeneity of productivity change (Arnold & Javorcik, 2005). China’s drastic changes in OFDI orientation and rapid growth in OFDI flows since 2002 provide us with a natural setting for analyzing the relationship between OFDI participation and firm productivity variations. Based on an integrated dataset from 1516 Chinese firms with 2033 foreign subsidiaries from China’s National Bureau of Statistics, Ministry of Commerce, local government reports and firm-level official websites for the period 2002–2008, we examine the instantaneous and future productivity gains upon OFDI entry controlling for the self-selection process. We find positive productivity premiums for EMEs with OFDI, but this productivity effect varies significantly according to EMEs’ heterogeneity in state ownership, absorptive capacity, and investment destination. The estimation results indicate that EMEs without state ownership gain positive productivity premium via OFDI, while this effect is insignificant for those with state ownership. We also find that EMEs with stronger absorptive capability and OFDI in OECD countries gain higher and more sustainable productivity premium than in non-OECD countries. The rest of the chapter is organized as follows. In line with RBV and IT, the next section introduces our literature review and hypothesis development In Sect. 5.3, we describe the dataset measures of variables and econometric model Our estimation results will be presented in Sect. 5.4. Section 5.5 presents our robustness check via different TFP and investment destination measures, one-step system-GMM estimation and re-estimation of absorptive capacity’s moderating effect in both technology intensive and other industries. Finally, Sect. 5.6 offers discussions and conclusions.

128

5 Does Outward FDI Generate Higher Productivity for Emerging …

5.2 Literature Review and Hypothesis Development Recent research in management has stressed the role of productivity as an indicator of firm performance as it is a representative of a firm’s general resource efficiency (Datta et al., 2005), sustained competitive advantage (Lieberman & Dhawan, 2005) and competitiveness (Causa & Cohen, 2004; Koch & MaGrath, 1996). OFDI has been touted as an important determinant of the firm’s productivity growth because it helps increase firm size and access new knowledge, making the firm more competitive in its home market (Bertrand & Capron, 2015). As later comers, EMEs, in contrast to developed country MNEs, are more likely to pursue productivity enhancement via OFDI as they are based in less innovative developing regions, possess a relatively narrow range and intensity of knowledge competencies, and hence more urgently engage in asset-seeking FDI in order to address their competitive disadvantages and improve their global competitiveness (Buckley et al., 2007). However, so far existing literature has generated inconsistent estimation results about the productivity effect of OFDI on EMEs (Bitzer & Kerekes, 2008; De La Potterie & Lichtenberg, 2001; Driffield & Chiang, 2009; Herzer, 2011; Hijzen et al., 2007; Masso & Vahter, 2008), which challenges the direct OFDI-productivity growth linkage based on traditional international business theory (Frost, 2001; Shan & Song, 1997; Teece, 1992), and stimulates us to ask why the testing results of the OFDI-led productivity growth hypothesis based on similar datasets and estimation techniques are so mixed. Among the early country studies, Herzer (2012), Goodarzi and Moghadam (2014), and Bitzer and Görg (2009) generate opposite empirical results about the impact of OFDI on domestic productivity. Driffield et al. (2009) find that both technologysourcing and efficiency-seeking FDI increases domestic productivity. Herzer (2011) also confirms that OFDI has, on average, a robust positive long-run effect on in developing countries. Herzer (2012) further reports a positive relationship between OFDI and domestic output and productivity. But De la Potterie and Lichtenberg (2001) assert that OFDI’s productivity effect happens only if a country invests in R&D intensive countries. For industry-level research, Braconier et al. (2001) find neither evidence of FDIrelated R&D spillovers, nor any correlation between OFDI per se and domestic productivity. Bitzer and Kerekes (2008) indicate that FDI receiving countries benefit strongly from inward FDI-related knowledge spillovers, but positive OFDI-led technology souring effects have not been found. Deploying a similar dataset Bitzer and Görg (2009) even find that a country’s stock of OFDI is, on average, negatively related to productivity. However, Driffield and Chiang (2009) report a positive association between OFDI to mainland China and labor productivity in Taiwan. In contrast with country- and industry-level analysis, firm-level study is argued to be better for investigating OFDI’s productivity effect, as it avoids the aggregation bias, and provides channels for identifying firm heterogeneity (Helpman et al., 2003), assisting in explaining firm-level variations in OFDI-led productivity effect. Kimura and Kiyota (2006) show that firms engaging in OFDI experience a 1.8 percent higher

5.2 Literature Review and Hypothesis Development

129

productivity growth than domestic firms not engaging in OFDI in Japan. However, Hijzen, et al. (2007) cast doubt on the positive results generated from the above firmlevel studies as they fail to control for the endogeneity bias that arises when more productive firms self-select into investing abroad. To deal with this endogeneity problem, Hijzen and et al., (2007) apply propensity score matching and differencein-difference analysis to data of Japanese firms for the period 1995–2002, and they find insignificant impact of OFDI on firm productivity in Japan. Barba Navaretti and Castellani (2004) apply the same methods as Hijzen and et al.’s (2007) to Italian firm-level data, and find that multinational firms have higher total factor productivity growth after investing abroad than national counterfactual firms. Employing cross-section data from a sample of French acquiring firms and non-acquiring firms, Bertrand & Capron (year of publication?) find positive relationships between OFDI via M&A and the acquirers’ productivity at home. Branstetter (2006) also confirms that OFDI is a channel of technology spillover for Japanese MNEs undertaking OFDI in the United States. Although a number of empirical studies deal with OFDI-related productivity diffusion in developed countries, less attention has been paid to productivity changes induced by OFDI in emerging economies. There are some empirical attempts based on firm-level data from Taiwan, mainland China and Estonia, but the estimation results of these studies are still mixed (Chen & Tang, 2014; Chen & Yang, 2013; Chen et al., 2012; Chuang & Lin, 1999; Cozza et al., 2015; Lee et al., 2013; Li et al., 2016; Masso & Vahter, 2008; Yang et al., 2013; Zhao et al., 2010). A summary of the past research results on OFDI-led productivity at the firm-level2 has been illustrated in Table 5.1. So why are the testing results of the OFDI-led productivity growth hypothesis based on similar micro data and estimation techniques so mixed? , Helpman et al. (2003) suggest that firm level specific heterogeneity makes firms’ investment strategy and performance diversity possible (Grossman et al., 2006). Kokko and Kravtsova (2008) also emphasize that technology diffusion and productivity premium are not automatic. Higher productivity gains should be expected for EMEs which have higher R&D and absorptive capability and invest in relatively developed regions (Cantwell & Janne, 1999; Cohen & Levinthal, 1989, 1990; Dosi, 1988). Additionally, variations in parent firms’ ownership (Ramasamy et al., 2012) and OFDI investment location (Branstetter, 2006; Li, 1995; Nocke & Yeaple, 2007) may significantly moderate EMEs’ learning-by-OFDI effect. However, so far there has been no report of any comprehensive investigation of whether OFDI’s productivity effect varies depending on firm heterogeneity, and how the firm-level heterogeneity moderates the OFDIproductivity change nexus (Bitzer & Kerekes, 2008; Branstetter, 2006; K. M. Chen & Yang, 2013; Chuang & Lin, 1999; De La Potterie & Lichtenberg, 2001; Herzer, 2008, 2010, 2011; Kogut & Chang, 1991).

2

Due to space limitation, a summary of country and industry level studies is not included in the table, but available upon request.

Topic

Longitudinal panel data Regression model on Japanese firms from 1991 to 1994

NA

Exports, FDI, and productivity: dynamic evidence from Japanese firms

Kimura and Kiyota (2006)

Analytical techniques

Firm-level data of Propensity score matching Chinese multinationals’ and DID OFDI into advanced European countries

Data and context

The impact of outward NA FDI on the performance of Chinese firms

Theoretical approaches

Cozza et al. (2015)

Firm-level studies with positive results:

Authors

Table 5.1 OFDI-led productivity growth: summary results of previous firm-level studies

(continued)

Exports and foreign direct investment appear to improve firm productivity once the productivity convergence effect is controlled for. Firms that retain a presence in foreign markets, either by exports or foreign direct investment, show the highest productivity growth, which contributes to improvements in national productivity

China’s OFDI has a positive impact on domestic activities. Specially, OFDI via M&A facilitates early access to intangible access, but is detrimental to financial performance, while greenfield investments have stronger impacts on productivity and scale of Chinese multinationals investing in Europe

Main findings

130 5 Does Outward FDI Generate Higher Productivity for Emerging …

International business theory

The contribution of outward direct investment to productivity changes with China, 1991–2007

Outward foreign direct investment and technical efficiency: evidence from Taiwan’s manufacturing firms

Technological innovation and productivity in late-transition Estonia: econometric evidence from innovation surveys

Zhao et al. (2010)

Yang et al. (2013)

Masso and Vahter (2008) NA

NA

Theoretical approaches

Topic

Authors

Table 5.1 (continued)

Chinese OFDI has had beneficial spill-over effects on total factor productivity growth over the period of the study, and gains in efficiency have been the chief reason for this

Main findings

(continued)

Horizontal OFDI is positively related to the parent firm’s productivity growth

Stochastic frontier OFDI raises firm productivity analysis (SFA) model and through its effect on both the propensity score matching firm’s technological endowments and technical efficiency

Vector auto regression (VAR) decomposition analysis

Analytical techniques

Firm-level data from Structural model Community Innovation Surveys (CIS3 and CIS4), combined with frim-level financial data from 1998–2000 and 2002–2004 in Estonia

Firm-level panel data from Taiwan’s manufacturing industries from 1987 to 2000

China’s ODI in eight developed countries during the period 1991 to 2007

Data and context

5.2 Literature Review and Hypothesis Development 131

International business theory

International business theory

Productivity enhancement at home via cross-border acquisitions: the roles of learning and contemporaneous domestic investment

Investments abroad and performance at home: evidence from Italian multinationals

International reverse International spill-over effects on parent business theory firms: evidences from emerging-market MNEs in developed markets

Bertrand and Capron

Barba Navaretti and Castellani (2004)

Chen et al. (2012)

Theoretical approaches

Topic

Authors

Table 5.1 (continued)

Ordinary least squares (OLS) model with robust standard errors (Huber–White–Sandwich estimator of variance)

Analytical techniques

A panel dataset of 493 Emerging market MNEs over the period 2000–2008

Panel lagged Tobit estimation model

A dataset includes both Propensity score matching Italian multinationals and DID and a random sample of Italian national firms during 1993–1997

Pooled cross-section data from a sample of French acquiring and non-acquiring firms from 1993–2004

Data and context

(continued)

Emerging market MNEs that have subsidiaries in host developed markets richer in technological resources (measured by R&D investments and R&D employment) exhibit stronger technological capabilities at home

OFDI improves growth of total factor productivity and output, but generates no significant effects on employment

Cross-border acquisitions can enhance the acquirers’ productivity at home, and these domestic productivity gains will be greater when there are learning opportunities in the targets’ host countries and when contemporaneous domestic productivity-enhancing investments are made by the acquirers in connection with the acquisitions

Main findings

132 5 Does Outward FDI Generate Higher Productivity for Emerging …

Theoretical approaches NA

International business theory and regional innovation system theory

International business theory

Topic

The dragon is flying west: micro-level evidence of Chinese outward direct investment

Outward foreign direct investment and domestic innovation performance: evidence from China

Outward FDI and knowledge flows: a study of the Indian automotive sector

Authors

Chen and Tang (2014)

Li et al. (2016)

Pradhan and Singh

Table 5.1 (continued) Analytical techniques

Panel data of 436 Indian Tobit model regressions automotive firms during 1988–2008

Balanced panel dataset GMM regression for 30 provinces and municipalities in China during 2003–2011; OFDI and R&D data for Chinese multinationals

A comprehensive Propensity-score dataset that covers close matching to 10,000 Chinese OFDI deals from 1998 to 2009

Data and context

(continued)

Positive effects of OFDI on domestic R&D are found for both investment in developed and developing regions, but stronger in the former case OFDI via IJV or WOS both tends to encourage domestic R&D, but the effect is stronger via IJV

OFDI has a very significant impact on domestic innovation. Contingent factors—absorptive capability, foreign presence, and the competition intensity of local market moderate the impact of OFDI on innovation performance

OFDI is associated with better firm performance, including higher total factor productivity, employment, and export intensity, and greater product innovation

Main findings

5.2 Literature Review and Hypothesis Development 133

NA

NA

Foreign direct investment, R&D and spill-over efficiency: Evidence from Taiwan’s manufacturing firms

Is foreign direct investment a channel of knowledge spill-overs? Evidence from Japan’s FDI in the United States

Chuang and Lin (1999)

Branstetter (2006)

Hijzen et al. (2007)

The effects of multinational production on domestic performance: evidence from Japanese firms

NA

NA

Impact of outward foreign direct investment on domestic R&D activity: Evidence from Taiwan’s multinational enterprises in low-wage countries

Chen and Yang (2013)

Firm-level studies with insignificant results:

Theoretical approaches

Topic

Authors

Table 5.1 (continued) Analytical techniques

Main findings

A large panel of Japanese firms for the period 1995–2002

Firm-level data set on Japanese MNEs’ OFDI and innovative activity

Taiwanese firm-level data

OFDI is a channel of technology spill-overs for Japanese MNEs undertaking direct investments in the United States

OFDI substitutes to R&D activity, and has positive impact on firm productivity due to its significant effect of industry-wide technology spill-overs

(continued)

Propensity score matching Japanese OFDI tends to and DID strengthen the economic activities of Japanese firms in Japan in terms of both output and employment. However, no significant positive effect of OFDI on productivity has been found

Fixed effects negative binomial regressions

Heckman two-stage estimation method

Panel data on Taiwanese Propensity score matching There is a positive relationship manufacturing firms between OFDI and domestic from 1992–2005 R&D activities

Data and context

134 5 Does Outward FDI Generate Higher Productivity for Emerging …

Topic

Theoretical approaches

NA Lee et al. (2013) Do local industrial agglomeration and foreign direct investment to China enhance the productivity of Taiwanese firms?

Authors

Table 5.1 (continued) Analytical techniques

Data for 578 Taiwanese Cross-sectional manufacturing econometric model multinationals and Taiwan industrial agglomeration indicator

Data and context

Local industrial agglomerations exert a positive contribution to firm productivity, but OFDI in China has no significant effects on Taiwanese multinationals’ TFP growth

Main findings

5.2 Literature Review and Hypothesis Development 135

136

5 Does Outward FDI Generate Higher Productivity for Emerging …

Fig. 5.1 Theoretical framework and hypotheses

Given that emerging economies’ OFDI growth rate has exceeded that from developed countries in the past decade (Buckley et al., 2007; Wang et al., 2012a, 2012b), more systematic research on OFDI from EMEs is required. Our work will contribute to the literature by introducing and testing an extended learning-by-OFDI model, which takes into consideration EMEs’ heterogeneity. RBV and IT will be utilized as the theoretical underpinning for this model. The theoretical framework is demonstrated in Fig. 5.1 and we believe that it can better explain and predict OFDI’s productivity effect. In the rest of this section, based on the theoretical lenses of RBV and IT, we first examine how OFDI affects EMEs’ productivity in their home markets as compared with domestic firms that have not conducted OFDI. We then focus on the mechanisms with which state ownership, pre-OFDI absorptive capability and investment destination moderate this OFDI-led productivity gains.

5.2.1 OFDI and EMEs’ Productivity Growth In this study, we examine the relationship between OFDI and EMEs’ TFP growth. Theoretically, the determinants of TFP growth include the creation, transmission and absorption of knowledge, factor supply and efficient allocation, efficient institutions, and effective market competition (Isaksson, 2007). As RBV has suggested, specialized, rare, and inimitable resources (e.g., technology, marketing resources, human capital, intermediaries and management capabilities) can not only derive from the firm itself, but also could be assembled and transferred across national boundaries (Barney et al., 2011; Meyer et al., 2009; Sirmon et al., 2011), and OFDI is a key channel for the transfer, mobility and reallocation of resources across boundaries

5.2 Literature Review and Hypothesis Development

137

(Kogut & Chang, 1991). An MNE’s productivity change thus can arise not only from the ownership of proprietary assets, but also from the ability to secure, or efficiently coordinate, the complementary assets possessed by other firms in host countries via OFDI (Frost, 2001; Shan & Song, 1997; Teece, 1992). As later comers, EMEs, in contrast to developed country MNEs, are more likely to gain productivity premium via OFDI as they are based in less innovative developing regions, possess a relatively narrow range and intensity of knowledge competencies, and hence more urgently engage in OFDI to seek for resources and learning opportunities (Buckley et al., 2007). EMEs’ OFDI, especially that in technology-intensive countries, provides them with channels for accessing advanced technology and human capital, offering EMEs the possibility of gaining productivity spillover via reverse technology flows, linkages with suppliers and clients, employee training programmes, and learning from nearby firms (Fosfuri et al., 2001). Mathews’ (2002) linkage-leverage-learning (LLL) model explains how EMEs obtain access to advanced intangible assets and gain productivity enhancement via OFDI. Specifically, the linkage via joint venture or strategic alliance in global value chains with foreign companies represents a fast and efficient way to access the resources that EMEs desire. Once linked, “latecomer” EMEs could utilize the global connections to leverage their own specific resources and learn about new resources. The greater the technological gap between the leading and backward countries, the greater the potential for technological progress of the latecomer MNEs (Wang et al., 2014). Second, OFDI’s productivity effect derives from the technical efficiency progress via economies of scale in not only manufacturing, but also R&D, sales, and administration (Bertrand & Capron, 2015). OFDI facilitates increased specialization, which is beneficial for the parent firms as it reduces sunk costs and allows reallocation of resources to their best utilization (Görg et al., 2008). OFDI also produces EMEs’ productivity growth via bringing in lower-priced intermediaries, helping them acquire global capital (Frost, 2004; Jong-Sung & Khagram, 2005) and reorganizing global production. Third, Bitzer and Görg suggest that through OFDI, EMEs are able to improve their parent firm productivity as they get exposure to fierce international competition and best practice, similar to the idea of “learningby-exporting” advocated by Clerides et al. (1998). As a complementary and interdependent construct of RBV in MNE study (Wang et al., 2012a, 2012b), IT’s basic assumptions are that institutional isomorphic pressures that stem from industry, government, and societal exceptions (e.g., the norms, rules and standards on product quality, occupational safety, or environmental management) define firm activities, and those pressures applied to all firms in the same institution cause firms to exhibit similarity (DiMaggio & Powell, 1983). Following this, firms that operate over time with partners embedded in the same institutions may turn to be self-reinforcing and more difficult to change (Rangan, 2000; Rosenkopf & Almeida, 2003), but OFDI helps overcome the institutional constrains associated with geographically bounded search (Alcacer & Chung, 2002) and facilitates firms to gain new resources and economies of scales across multiple geographic settings (Salomon & Shaver, 2005). A firm’s capability to evaluate, recognize and learn about resources in a given institution declines with geographic distance from that location,

138

5 Does Outward FDI Generate Higher Productivity for Emerging …

and OFDI is viewed as a channel to access resources which are often embedded in local knowledge clusters, from distant markets (Contractor, 2012). Second, defined as the rules and organs that drive the production climate (Ulubasoglu & Doucouliagos, 2004), more efficient institutions assist firms’ productivity growth via the enforcement of property rights, supply of a developed financial system, and effective innovation system (Isaksson, 2007). OFDI thus provides EMEs with the opportunities to gain productivity effect via taking institutional advantages in host countries. Third, OFDI provides EMEs with channels for institutional arbitrage. This term has been commonly used in the international business literature as a practice of arranging activities in various institutional contexts in order to benefit from differences in regulatory systems (Chacar et al., 2010). In the context of OFDI, gains from institutional arbitrage opportunities with legal and tax optimization could be expected. Thus, in line with RBV and IT, OFDI provides EMEs with channels to. (a) (b) (c) (d)

create, transfer and absorb knowledge; reallocate resources and realize economies of scale; access developed institutions and institutional infrastructure; get exposed to international competition, and all of which could contribute significantly to EMEs’ productivity premium. Based on the above arguments, we hypothesize:

H1. OFDI generates a positive effect on EMEs’ productivity. However, as existing literature has stressed, OFDI not only yields benefits, but also is always associated with increased complexity, coordination needs, and resource trade-offs (Bertrand & Capron, 2015; Levinthal & Wu, 2010). It is not always the case that productivity benefits could outweigh the cost of foreign expansion, and the real story is that the productivity premium EMEs gain from OFDI could be moderated by resource and institutional conditions. We thus move forward to examine moderators of OFDI’s productivity effect on EMEs. Three variables are chosen because they can significantly influence outcomes of EME investment activities based on our framework. Firstly, we choose state ownership as one moderating variable because this status represents EMEs’ affiliation with government, and is the most important institutional factor influencing EMEs (Buckley et al., 2008; Cui & Jiang, 2012; Wang et al., 2012a, 2012b) due to state provision of institution-based resources to these firms ( Hoskisson et al., 2013; Peng, 2003). Secondly, in line with Cohen and Levinthal (1990), Dai and Yu (2013), and Griffith et al. (2004) who suggest that a firm’s valuable resources derive from intangible knowledge, our next moderating variable is absorptive capacity, measured by R&D to capture an EME’s key valuable resources. This variable moderates the OFDI-productivity linkage as it can assist EMEs in recognizing, exploiting acquired global assets and making further innovations. Thirdly, investment destination, as a very important internationalization strategy determined by EMEs’ resources and institutional background (Buckley et al., 2007; Cui & Jiang, 2009; Ramasamy et al., 2012), and can significantly moderate outcomes of EMEs’ OFDI. Thus, we next describe in more detail the three moderators, and our framework is represented by Fig. 5.1.

5.2 Literature Review and Hypothesis Development

139

5.2.2 State Ownership and OFDI’s Productivity Effect on EMEs A key assumption of RBV is that although managerial decisions are constrained by information asymmetry and causal ambiguity, they are driven by motives of efficiency and competitiveness. Following this, EMEs’ OFDI decisions should be economically justified, to get the maximum use of OFDI-led resource-based advantages (e.g., get access to technology, international markets, lower-priced intermediaries, economies of scale, and efficient institutional infrastructure), and positive productivity effects of OFDI can be expected. However, the state ownership, which turns the EMEs with it to be assets or parts of their home country institutions (Cui & Jiang, 2012), stimulates state-owned EMEs (SO-EMEs) to perform far from economically optimal, but as a serve of political goals (Wang et al., 2012a, 2012b). Under this condition, state ownership, as a paramount institutional factor, produces significant effects on EMEs’ (1) resource dependence on governments and OFDI objectives and strategies; (2) resource endowments, utilization and international competitive capability; and (3) political reputation and confronted pressures in host markets (Chen & Young, 2010; Cui & Jiang, 2012; Rugman & Li, 2007; Wang et al., 2012a, 2012b). Therefore, based on RBV, the OFDI-productivity linke can be moderated by state ownership. IT suggests that firms are under institutional pressures to adhere to the formal and informal rules in their institutional fields (DiMaggio & Powell, 1983; Scott, 1995), and their responses to institutional pressures vary according to firms’ levels of resource dependence on the institution that exerts the pressure (Oliver, 1991). With high resource dependence, a firm is more likely to conform to the institutional pressures to avoid negative consequences (Salancik & Pfeffer, 1978). Thus according to IT, as state-owned (SO)-EMEs, in contrast with private EMEs, are politically affiliated with home-country governments and are highly dependent on the home-country institutions for critical resource inputs (Liang et al., 2015), they are under more pressures to conform to, rather than resist the political and strategic purposes home country governments specify for OFDI. While pursuing their business objectives, SO-EMEs are always required to serve the political mandates of the state and align their interests with the home institutions rather than challenge these interests (Scott, 2002; Zhang et al., 2011). Under this condition, when internationalizing, SO-EMEs turn out to be a political actor seeking for political goals, but not profit-maximizing agencies, which goes against to RBV’s assumptions (Buckley et al., 2007). On the contrary, private EMEs, with pressures of survival, turn out to be eager for profits and efficiency. OFDI will be conducted to serve their corporate strategies and economic success, and once linked, private EMEs make all efforts to get access and utilize global resources. Empirical studies confirm that while private EMEs are likely to focus on seeking technology and markets, SOEs favor in investing in seeking natural resources, serving as a political actor for the nation’s sustainable development (Buckley et al., 2007; Ramasamy et al., 2012; Wang et al., 2012a, 2012b; Luo & Thug, 2007 Furthermore, different from private EME managers who

140

5 Does Outward FDI Generate Higher Productivity for Emerging …

formulate internationalization strategy to pursue global assets and markets, many SO-EME managers are often directly appointed by the state after serving as government officials (Brockman et al., 2013; Fan et al., 2007) and their companies go global following the guidance and capital control by the home state (Cui & Jiang, 2012). Correspondingly, SO-EME managers are incentivized not just by the prospect of increasing economic performance but also by satisfying the state’s political and social objectives in making OFDI strategies (Cuervo-Cazurra & Dau, 2009). Thus in conclusion, different from private EMEs’ profit-driven OFDI, SO EMEs’ OFDI are motivated not solely by self-interests, but also by the interests of the institutions they are affiliated with. Private EMEs thus tend to be more incentive to exploit OFDI to access high-tech, efficiency, and international market, and then gain more productivity premium. In addition, state ownership affects EMEs’ resource endowments and thus influence the competition pressures they confronted with and the productivity effect thay gain from competition in host markets (Wang et al., 2012a, 2012b). As Bitzer and Görg suggest, international competition stimulates MNEs to transfer and absorb technologies, management skills, and produce productivity gains. However, with superb resources and unfair competitive advantages (Meyer et al., 2014), SO-EMEs, compared with private counterparts, have less incentive to international competition. SO-EMEs tend to be endowed with monopolistic resources from home governments, like capital from state-owned banks, and extra business chances provided by national corporations (Amighini et al., 2013; Buckley et al., 2007; Cui & Jiang, 2009; Luo et al., 2010). But the low-cost and easy-accessibility of public resources result in SO-EMEs’ less sensitivity to market competition, and to the risk perception during OFDI (Buckley et al., 2007; Cui & Jiang, 2012). With perceived government backing combined with below-market cost of capital, SO-EMEs are able to bear short-term loss and misleading OFDI strategies (Ahmed et al., 2002). When making strategic decisions, SO-EMEs may seek the possibility of further government support, which may be available in unexpected adverse circumstances. Thus, the inequity of resource endowments generate negative effect on SO-EMEs’ efficiency in competition, and further buffer the productivity gain SO-EMEs can get from global competition. At the same time, OFDI from SO-EMEs often encounter highly burdensome and bureaucratic administrative OFDI approval procedures as governments at various levels, seek to affect the direction, amount and scope of outward capital flows (Buckley et al., 2007). These will result in SO-EMEs’ less incentive and ability to gain productivity benefits from OFDI. Furthermore, because of SO-EMEs’ affiliation with home institutions, when they invest overseas, they might be perceived by host-country institutions not simply as business entities, but also as political actors (Globerman & Shapiro, 2009; He & Lyles, 2008). Such a perception can pose challenges and more stressful institutional pressures to SO-EMEs’ institutional processes in host countries (Luo & Rui, 2009; Peng et al., 2008). There can be concerns about the political rationale of SO-EMEs in attempted foreign acquisitions, such as CNOOC’s failed acquisition of Unocal (Wan & Wong, 2009). From the host-country aspect, the state-driven objectives of SO-EMEs are often perceived as non-beneficial, or even harmful, to the host country

5.2 Literature Review and Hypothesis Development

141

(Globerman & Shapiro, 2009). Consequently, the institutional barrier for SO-EMEs to assume ownership and control in their investment in the host country will be high, which decreases the likelihood of productivity gain (Cui & Jiang, 2012). Thus we conclude that as a representative of EMEs’ affiliation with the government, state ownership buffers EMEs’ productivity gains from OFDI as it hinders EMEs’ incentives to pursue profits and economic efficiency through OFDI, reduces EMEs’ sensitivity to competition and produces higher institutional pressures and hazards in host countries, impeding the function of OFDI-led productivity growth mechanisms: H2. State ownership moderates the effect of OFDI on an EME’s productivity as such positive gains will be smaller for EMEs with state ownership than those without.

5.2.3 Absorptive Capacity and OFDI’s Productivity Effect on EMEs RBV suggests that existing resources enable firms to develop dynamic capabilities, just as previous learning facilitates the learning and application of new, related knowledge (Barney, 2001; Deng, 2007; Teece, 2014). It is evident that EMEs invest overseas because they wish to acquire knowledge and learn new skills and capabilities in order to enhance their competitive advantages and productivity. But as Kokko and Kravtsova (2008) emphasize, technology diffusion and productivity premium are not automatic. Higher productivity gains should be expected for EMEs with higher R&D and absorptive capability, which helps them to better recognize the value of new information, assimilate and apply it to commercial ends (Cohen & Levinthal, 1990), or build the “ability to make effective use of technological knowledge in efforts to assimilate, use, adapt and change existing technologies”. Deng (2007) and Rui and Yip argue that the existing stock and quality of R&D influence the extent to which reverse transfer and spillover of knowledge take place within MNEs. Sawada (2010) also indicates that the productivity effect through technology spillovers depends on MNEs’ absorptive capacity. Therefore, EMEs with strong absorptive capability are more likely to capitalize on their assets, recognize and absorb valuable knowledge, build up new resources via OFDI, and gain higher productivity premium (Zahra & George, 2002). In addition, EMEs’ absorptive capability serves in making the maximum utilization of the intermediaries and facilities in host countries. With strong R&D-based absorptive capability, EMEs are more capable of recognizing valuable intermediate inputs, reallocating resources optimally, coordinating efficient international production, exploiting well-developed host country institutional infrastructure (e.g., financial system, human capital, suppliers and clients, innovation center) and achieving economies of scale and productivity upgrade (Bertrand & Capron, 2015; Chen et al., 2012; Görg et al., 2008). At the same time, according to IT, EMEs need particular firm resources and capabilities to deal with host country institutional hazards (Cantwell

142

5 Does Outward FDI Generate Higher Productivity for Emerging …

et al., 2010). As a kind of intangible firm specific resource, EMEs’ absorptive capability assists them in dealing with host country institutional conflict, succeeding in host market competition, producing virtuous cycles for EMEs’ resource utilization and regeneration, and thus stimulates efficiency and productivity enhancement. Based on the above arguments, we hypothesize that: H3. Absorptive capacity moderates the effect of OFDI on an EME’s productivity as such positive gains will be greater, if an EME’s pre-OFDI absorptive capacity is stronger.

5.2.4 Investment Destination and OFDI’s Productivity Effect on EMEs RBV suggests that valuable resources are tacit and likely to be sticky or embedded in geographically-bounded clusters (Barney, 2001; Buckley, et al., 2008; Cantwell & Iammarino, 2000). Here, resources include not only R&D capabilities, but also organizational processes, diverse functional skills (e.g., marketing, commercial), managerial best practices, as well as learning opportunities from competitive interactions and institutional systems (Alcacer & Oxley, 2014). Such resources generate positive impacts on a firms’ productivity growth. As an important channel for accessing resources, OFDI’s location is therefore essential in determining the learning opportunities and the extent to which EMEs will access those resources and gain productivity premium (Bertrand & Capron, 2015). A more significant productivity premium can be expected via conducting OFDI in more developed countries for the following reasons. International business scholars have stressed the role of technological or competitive gap between the home country of investment firms and the targeted host country. There are more opportunities to benefit from knowledge and resources that do not exist in the home country when an MNE invests in a country that is more advanced than its own (Cantwell & Janne, 1999; Kogut & Chang, 1991). While developing countries, as latecomers, are cheap labor and natural resource abundant, developed countries are rich in technology, as R&D has always been considered a domain of firms in technologically advanced and economically developed countries. Thus compared with OFDI in developing countries which probably provides EMEs with economies of scale, OFDI in developed countries are expected to offer EMEs more productivity premium via access to and reverse transfer of technologies. Empirically, Griffith et al. find that U.K. firms with a greater R&D presence in the United States enjoy higher productivity. In addition, networks that are conducive to innovation (e.g., research labs, researchers, technology-generating facilities) are geographically bounded in developed markets and cannot be easily replicated in other locations (Almeida & Kogut, 1999), which indicates that OFDI in developed countries can better stimulate EMEs’ creation of new knowledge.

5.3 Methodology

143

Additionally, as OFDI in both developing and developed countries assists in reallocating resources and generating economies of scale, OFDI in developed countries is expected to offer EMEs more productivity premium, via providing a high level of local density of specialized resources, agglomeration economies, specialized labor and intermediate inputs (Head et al., 1995). At the same time, recent economic studies suggest that productivity premium is likely to be greater when investing in a country whose market is more competitive than EMEs’ home market (Herrerias & Orts, 2012). In developed economies, institutional frameworks foster and stimulate market-based competition and firms’ strategic innovation. Firms’ primary challenge is to develop competitive resources and capabilities to outperform competitors in the market place (Peng, 2003). Well-developed institutional systems and institutional infrastructure in developed countries also facilitate the productivity enhancement. As shown by Alcacer and Chung (2002), OFDI provides investing firms with opportunities to access not only resources in specific firms, but also those embedded in firms’ broader institutional environment and ecosystem. In contrast, developing countries often lack sufficient market-supporting political, legal, and economic institutions (Hoskisson et al., 2013), and this works as a location-disadvantage restricting the formulation of local firms’ capabilities (Khanna & Rivkin, 2001). We thus posit that EMEs’ OFDI in developed countries provides them with more opportunities to access technology, specialized intermediate inputs, competitive markets and well-developed institutional infrastructure, and hence better capabilities and higher productivity premium gains can be expected. Formally, we have our fourth hypothesis. H4. Investment destination moderates the effect of OFDI on an EME’s productivity as such positive gains will be greater if the firm invests in an OECD country compared with a non-OECD country.

5.3 Methodology 5.3.1 Data We test the OFDI-led productivity change hypothesis relying on two disaggregated datasets. One is derived from the Annual Manufacturing Enterprises Survey conducted by China’s National Bureau of Statistics. This dataset covers all SO-EMEs and non-SO-EMEs whose annual sales exceed RMB 5 million.3 The data used in this chapter ranges from 2002 to 2008,4 covering more than 180,000 firms in 2002 and 3

In fact, the aggregated data on the industrial sector in the annual China’s Statistical Yearbook and China’s Industry Economy Statistical Yearbook are compiled from this data set. 4 When estimating the probability to invest abroad in one year, we would use firms’ production data in the previous year, namely when the previous year is 2001, we would also utilize firms’ production data in 2001. Thus more exactly speaking, the production data utilized in this chapter range from 2001 to 2008.

144

5 Does Outward FDI Generate Higher Productivity for Emerging …

more than 410,000 firms in 2008. This dataset includes three major accounting statements—the balance sheet, cash flow statement and income statement, incorporating more than 100 useful variables.5 However, noisy observations still exist in this dataset because of non-standardized financial statements or report errors from some firms. We clean the sample and remove the outliers by using the following filtering criteria. First, following Feenstra et al. (2014), observations with missing primary financial variables (such as total asset, gross industrial output value and net fixed assets) are omitted. Second, firms with fewer than 8 workers are excluded from the sample because they are under a different legal regime (Brandt et al., 2012; Yu, 2015). Furthermore, observations satisfying the following criteria are excluded according to the basic rules of the Generally Accepted Accounting Principles: (a) Liquid assets are greater than total assets; (b) Fixed assets are greater than total assets; (c) Net fixed assets are greater than total assets; (d) An invalid founded time exists (i.e., the opening month is earlier than January or later than December.); (e) The firm’s identification number is missing. The second dataset used in this chapter comes from China’s Ministry of Commerce. It covers rich information of EMEs that have conducted non-financial OFDI, including parent firm names, registration addresses, investment destinations, foreign subsidiary names, approval dates6 and business scopes. All the EMEs engaging in non-financial OFDI from 2002 are covered in the dataset.7 This enables us to investigate the OFDI-led productivity change at the country level. Based on whether a firm invested abroad during the sample period, we divide the whole sample into two subsamples, i.e., one with firms having OFDI and the other with firms having no OFDI.

5

Because the intermediary input variable is missing in 2008, we impute this variable using a conventional method. According to China’s Statistical Yearbook, value added = total output − intermediary input + value added tax payable. We can impute the missing data by the equation, intermediary input = total output − valued added + value added tax payable. Here we assume a firm’s valued-added rate in 2008 equals to that in 2007. Depending on the value-added rate in 2007 and the total output in 2008, we can obtain firms’ value added in 2008, and then firms’ intermediary inputs in 2008 can be straightforwardly derived. Imputing the intermediary input data in 2008 helps extend the sample. If we only utilize the sample 2002–2007, our findings do not change significantly. 6 Considering that the date on which an investment was approved differs from that on which the subsidiary was established, we spared no effort to search the internet (the information from the parent firm’s website is labeled top priority) to confirm the exact date of subsidiary establishment. If the establishment date is unavailable, the year in which the investment was approved is used to approximate the year in which the subsidiary was established. 7 In 2002, the former Ministry of Foreign Trade and Economic Cooperation of China and China’ National Bureau of Statistics jointly developed China’s first Outward Foreign Direct Investment Statistical System.

5.3 Methodology

145

5.3.2 Measures Following previous literature, we deploy TFP to capture EMEs’ productivity change (Damijan et al., 2007), but we avoid the simple Solow residual approach as it is not reliable enough and would cause biased productivity estimation results (Olley & Pakes, 1992). Instead, following Dai and Yu (2013), De Loecker (2007), De Loecker et al. (2012), Yu (2015), we have augmented the traditional Olley and Pakes (1992) approach by introducing OFDI and export dummies when the production function is estimated, because EMEs with or without OFDI and export may confront with different production environments and resource allocation processes. Compared with the traditional simple OLS method, our augmented Olley-Pakes approach in TFP measurement has many advantages. First, this method utilizes the function of real current-period capital stock and investment8 as the proxy variable for current-period productivity,9 which effectively controls for the unobservable productivity shock in the production function estimation, and reduces the simultaneity bias. Second, firms’ survival probability is considered in the estimation of the production function, and thus the selection bias that only productive enterprises could survive in the markets can be effectively corrected. Third, we take into account the role of exporting and OFDI status when estimating production function, which helps to alleviate the production function estimation bias arising from omitting influential variables in the production function (De Loecker, 2011). An EME’s OFDI status is measured based on whether the firm has conducted OFDI. To estimate the impact of state ownership on the EME’s productivity gain from OFDI, we distinguish between EMEs with state ownership and those without for further comparison (Amighini et al., 2013; Cui & Jiang, 2012). Following Cohen and Levinthal (1989), we measure an EME’s absorptive capability by its R&D. It is suggested that R&D not only generates innovations, but also develops an ability to identify, assimilate, and exploit knowledge from the environment (Zahra & Hayton, 2008). The investment destination dummy is composed by non-OECD countries (coded “0”) and OECD countries (coded “1”) (Buckley et al., 2008; Pradhan & Singh, 2008). As suggested by the literature, we include firm features as our control variables, including the firm productivity (Helpman et al., 2004), inputs (capital, labor and intermediate inputs), firm-level strategies (pre-OFDI export and R&D decision), firm ownership (foreign-invested firm or state-owned enterprise), firm age (Wang

8

We adopt the perpetual inventory method as the law of motion for real capital and real investment. The nominal and real capital stock constructed following Brandt et al. (2012). We depend on the firm’s own information in the dataset to construct firm’s real depreciation ratio. 9 Here is an implicit assumption, namely, firms which are more productive now would have higher expected return rates, and hence those firms would invest more in that period. Under a few assumptions of production technology, (Pakes, 1996) has verified this implicit assumption.

146

5 Does Outward FDI Generate Higher Productivity for Emerging …

Table 5.2 Summary statistics and correlations 1

2

3

4

5

6

7

8

9

10

1. Log (number of 1 employees) 2. Log (capital stock)

0.65 1

3. Log (intermediate inputs)

0.57 0.70 1

4. Log (TFP)

0.13 0.25 0.20

1

5. Firm age

0.26 0.21 0.04

0.04 1

6. OFDI dummy

0.03 0.04 0.04

0.01 0.00

7. Export dummy

0.27 0.16 0.19

0.08 −0.02 0.03

1

8. R&D dummy

0.20 0.26 0.21

0.08 0.10

0.09

9. SOE dummy

0.14 0.15 −0.10 0.08 0.43

10. FIE dummy

0.17 0.21 0.17

0.12 −0.13 0.01

0.41

0.02 −0.16 1

Mean

4.67 9.75 9.74

1.91 10.90

0.001

0.25

0.11 0.08

0.20

Standard deviation 1.11 1.44 1.38

0.56 11.33

0.30

0.43

0.32 0.27

0.40

1 0.03

1

−0.00 −0.10 0.05 1

Note We impute the missing R&D variable in 2004 with the average values in 2003 and 2005. R&D data in 2008 are not utilized in the following analysis, and hence not included here. All the variables summarized here range from 2002 to 2008 except for the R&D dummy. TFP presented in this table is calculated using an augmented Olley-Pakes approach

et al., 2012a, 2012b), and the dummy variables for year and industries.10 Head and Ries suggest that firms can adopt OFDI and export as substitutive or complementary strategy to engage in foreign markets. We therefore treat pre-OFDI export decision as one firm strategy that may influence the firm’s OFDI decision. The R&D decision is another firm strategy that may affect the firm’s OFDI decision. As Lu et al. (2011) illustrate, firms in industries with higher levels of R&D intensity have higher probability to conduct strategic asset-seeking OFDI. Table 5.2 provides the correlation matrix of independent variables and associated summary statistics.

5.3.3 Econometric Model Disentangling correlations and causality in the OFDI-productivity growth nexus faces numerous challenges. As high productive firms are more likely to invest abroad, productivity growth may be endogenous and self-selected, and simple least squares estimation is invalid. Inspired by former literature (Bascle, 2008; Hamilton & Nickerson, 2003), we use propensity score matching to assess the causal effect of OFDI 10

In fact, to alleviate the influence of business cycle and control for the industry heterogeneity, we estimate the propensity score on a year-by-year and industry-by-industry basis.

5.3 Methodology

147

on parent firm productivity change. The matching technique creates the missing counter facts of firms that have foreign subsidiaries. It does so by pairing up a firm that conducts OFDI with a domestic plant (or several plants) with similar observable characteristics operating in the same sector and year, where similarities are determined on the basis of those plant features that have explanatory power in the OFDI decisions. Following De Loecker (2007) and Hayakawa et al. (2013), propensity score matching is employed combining with a difference-in-difference approach. The OFDL-led productivity effect is hence inferred from the average divergence in the productivity paths between each firm having OFDI and its matched control plants, starting from the pre-OFDI year. This strategy allows us to control for observable and time-invariant unobservable differences between OFDI firms and their control plants (Heckman et al., 1997). The basic idea of propensity score matching is to take OFDI as a “treatment”, and then the productivity effect of OFDI can be captured by the average treatment effect on the treated (ATT). We rescale the year that a firm just starts to invest abroad as period 0, and employ s 0 to represent the number of years after a firm starts to invest abroad. Variable “ start i = 1” represents that firm I invest abroad. Then the productivity effect of starting to invest abroad could be expressed as:   1   0   1 0 |start = 1 = E ωis |start = 1 − E ωis |start = 1 E ωis − ωis

(5.1)

The productivity is denoted by ω1 if a firm starts to invest abroad, and by ω0 if it does not. Equation (5.3) shows the average treated effect on the treated group (firms that start to invest abroad). The key point of getting the ATT is to find out the counter facts of the treated group, i.e. the control group. To achieve this purpose, propensityscore matching approach is utilized to construct the control group, following some previous studies (De Loecker, 2007; Imbens, 2004; Rosenbaum & Rubin, 1983). Based on the information prior to the year when the firm started to invest abroad, we have constructed the following model to estimate the propensity score:      Pr star ti,0 = 1 =  h X i,−1

(5.2)

where ϕ represents the cumulative density function of a normal distribution. Xi,-1 refers to firm i’s characteristics as our control variables which could predict whether this firm would invest abroad in the next period. Firm pre-OFDI productivity, inputs (capital, labor and intermediate inputs), firm-level strategies (pre-OFDI export and R&D decision), firm ownership (foreign-invested firm or state-owned enterprise),

148

5 Does Outward FDI Generate Higher Productivity for Emerging …

firm age,11 and the dummy variables for year and industries serve this purpose, as our above discussion in Sect. 5.4. According to Becker and Ichino (2002) and De Loecker (2007), the following algorithms are employed to find out the control groups. Firstly, the observations are split into k equally spaced intervals depending on the propensity score.12 Secondly, within each interval, we test whether the average propensity score of the experimental group (treated group) differs significantly from that of the control group. If the test fails in one interval, the interval would be split in half and be tested again until the average propensity scores of treated and control groups do not differ significantly in each interval. Thirdly, we test whether the means of the covariates do not differ significantly between treated and control groups within each interval, and this is to check whether the balancing condition is satisfied. If the balance condition is rejected, we will alter the functional form of the propensity score by adding higherorder covariates and interaction terms and redo the above steps. Fourthly, the nearestneighbor matching method is employed to find out the counterfactual observations after the balancing condition is satisfied.13 After obtaining the control group, we pool all the years and industries together and calculate the average TFP difference between the treated and control groups. C(i) denotes a set of firms that are matched to firm i, and NiC refers to the number of firms in C(i). The weight of firm that is matched to firm i is denoted as wi j = N1c . i ω1 and ωc are the productivity of the treated firm and the firm in the control group respectively. Then the average treatment effect for year on the treated can be written as follows: s AT Tlevel =

 1  1 ωi,s − wi j ωcj,s Ns i j∈C(i)

(5.3)

The year-by-year productivity growth effect can be expressed as follows:

11

Although there is an argument that the newness of the subsidiary could explain the improvement in firm productivity, our analysis still holds. Frist of all, it’s true that subsidiaries started in different years probably take different technologies, but our estimation results still can show the positive productivity spillover effect through the backward linkage if parent firms benefit from engaging in OFDI. Furthermore, our analysis is to compare the productivity changes of parent firms with their counterfactuals (firms that operate in the same year and industry with the treatment group, but not engage in OFDI) rather than directly compare firms with OFDI in different years. Moreover, our sample ranges from 2002 to 2007 (mainly between 2004 and 2007), and hence technologies used in a given manufacturing industry may be relatively similar during such a short time period. More importantly, our results still hold if we restrict the estimation sample to 2004–2007. 12 The initial value of is set to 2. 13 After sorting the sample by the propensity score, we search the counterfactual observations for the treated group by searching upward and downward. In fact, we find two firms for each treated one. Some other matching methods are also utilized, such as finding out one or four counterfactual observations for each treated firm, but our main results do not change significantly.

5.4 Estimation Results

s AT Tgrowth

⎤ ⎡     1 ⎣ 1 1 − ωi,s − ωi,s−1 = wi j ωcj,s − ωcj,s−1 ⎦ Ns i j∈C(i )

149

(5.4)

It is obvious that the productivity effect estimated from Eq. (5.4) is actually the average difference of productivity growth between the treated group firms and the matched control group firms.14 Next we combine the propensity score matching and DID approaches to produce a more precise estimation of the productivity effect of OFDI (Blundell & Dias, 2009). We compare an EME’s productivity with its pre-OFDI level (s = −1), where ABA denotes the productivity growth difference in period & compared to the pre-OFDI level, for the treated and control groups. ⎤ ⎡       1 1 1 ⎣ ωi,s − DIDs = − ωi,−1 wi j ωcj,s − ωcj,−1 ⎦ Ns i j∈C(i )

(5.5)

Considering the advantage of controlling for the pre-OFDI level of productivity after propensity-score matching, we rely on the DID measure to produce our main estimation results.

5.4 Estimation Results 5.4.1 Results At the Overall Manufacturing Level Table 5.3 demonstrates the estimation results at the overall manufacturing level. Panel (1) describes the impact of OFDI on the parent firm’s level-value of productivity change over time, while panel (2) indicates the year-to-year productivity premium the new investor gathered over time. The results show that the productivity premium for EMEs that started to engage in OFDI increased gradually. Based on the DID approach, EMEs’ average productivity gains from the first year to the third year after starting OFDI grew from 4.9 to 14.5%, which is similar to the pure level effect. Panel (3) shows that firms’ year-to-year productivity growth after OFDI is significant except the second year after OFDI.15 Thus our H1 is to a large extent supported.

14

We adopt firms that never invest abroad in the sample period as the control group. There is an alternative way to choose the control group, i.e., treating firms that just do not invest abroad in the given year as the control group. However, the latter approach inevitably neglects the lagged effect of investing abroad in the previous years. Therefore, our estimation results are based on the former approach. 15 This finding is similar to the conclusion about the productivity effect of exporting by De Loecker (2007).

150

5 Does Outward FDI Generate Higher Productivity for Emerging …

Table 5.3 Productivity effect of OFDI—at the overall manufacturing level16 s

0

1

2

3

(1) Results: TFP level

0.015

0.051***

0.080***

0.147***

Standard error

(0.011)

(0.015)

(0.023)

(0.039)

(2) Results: TFP growth: DID measure

0.014

0.049***

0.079***

0.145***

Standard error

(0.011)

(0.018)

(0.025)

(0.042)

(3) Results: TFP: year-to-year growth

0.014

0.030*

0.027

0.053*

Standard error

(0.011)

(0.017)

(0.021)

(0.029)

Number of treated units

1024

657

346

114

Note This table reports the estimation results of OFDI’s impacts on parent firms’ productivity change. An augmented Olley-Pakes approach has been used here, and standard errors are reported in the parentheses. * , ** , *** indicate significance level at 10%, 5% and 1%, respectively

5.4.2 State Ownership and OFDI’s Productivity Effect on EMEs To test our first hypothesis that state ownership moderates OFDI’s productivity effect on EMEs, we split the sample into four groups based on state ownership17 and EMEs’ OFDI status and test whether there is a difference in productivity effects between private EMEs and SO-EMEs. We treat SO-EMEs (private EMEs) with no OFDI in the sample period as the control group for SO-EMES (private EMEs) that conduct OFDI in that year, and our matching approach is on a year-by-year and industry-by-industry basis. The estimation results are listed in Table 5.4. Table 5.4 and Fig. 5.2 show that OFDI indeed significantly contributes to the productivity growth for private EMES. Their productivity benefits increase from 1.8% in the first year to 15.2% in the third year after conducting OFDI. While for SO-EMEs, the productivity growth effect is not significant. Therefore H2 is supported.

16

In this table, the number of treated units is less than that of the treated ones after matching. There are several reasons for this situation. First, production information prior to the year when firms started to invest abroad is needed for matching, and firms with missing pre-OFDI information are omitted. Second, firms that cannot be matched are dropped because of the violation of the balance condition hypothesis. Third, when calculating TFP with the augmented Olley-Pakes approach, firms with missing covariates are deleted. These are also the cases for later estimations. 17 By the official definition reported in China Statistical Yearbook (2008), SO-EMEs include firms such as domestic SO-EMEs (code: 110), state-owned joint venture firms(141), and state-owned and collective joint venture firms(143), but exclude state-owned limited corporations (151), based on the registration type.

5.4 Estimation Results

151

Table 5.4 Instantaneous and long-run productivity effect of OFDI 0

1

2

3

(1) Results: TFP: DID measure

−0.015

0.025

0.036

0.11

Standard error

(0.026)

(0.041)

(0.147)

(0.131)

(2) Results: TFP: year-to-year growth

−0.015

0.046

−0.014

0.073

Standard error

(0.026)

(0.063)

(0.94)

(0.109)

Number of treated units

54

37

29

11

(1) Results: TFP: DID measure

0.018*

0.053***

0.086***

0.152***

Standard error

(0.010)

(0.018)

(0.024)

(0.041)

(2) Results: TFP: year-to-year growth

0.018*

0.026*

0.031

0.058**

Standard error

(0.010)

(0.014)

(0.020)

(0.029)

Number of treated units

916

598

302

97

(1) Results: TFP: DID measure

0.035**

0.066***

0.122***

0.169***

Standard error

(0.014)

(0.022)

(0.030)

(0.050)

(2) Results: TFP: year-to-year growth

0.035**

0.028

0.048*

0.049

Standard error

(0.014)

(0.019)

(0.027)

(0.036)

Number of treated units

447

283

146

51

S By state ownership (A) Results for SO-EMEs

(B) Results for private EMES

By Pre-OFDI R&D status (C) Results for firms with pre-OFDI R&D

(D) Results for firms without pre-OFDI R&D (1) Results: TFP: DID measure

0.007

0.045*

0.072*

0.079

Standard error

(0.012)

(0.026)

(0.037)

(0.061)

(2) Results: TFP: year-to-year growth

0.007

0.031

0.029

0.011

Standard error

(0.012)

(0.019)

(0.030)

(0.039)

Number of treated units

521

321

174

55

By Investment Destination (E) Results for firms starting to invest only in OECD countries (1) Results: TFP: DID measure

0.022*

0.068***

0.106***

0.178***

Standard error

(0.012)

(0.018)

(0.025)

(0.041) (continued)

152

5 Does Outward FDI Generate Higher Productivity for Emerging …

Table 5.4 (continued) S

0

1

2

3

(2) Results: TFP: year-to-year growth

0.022*

0.039*

0.033

0.055*

Standard error

(0.012)

(0.021)

(0.022)

(0.032)

Number of treated units

394

201

102

25

(F) Results for firms starting to invest only in non-OECD countries (1) Results: TFP: DID measure

0.008

0.037*

0.068**

0.135**

Standard error

(0.010)

(0.021)

(0.028)

(0.059)

(2) Results: TFP: year-to-year growth

0.008

0.024

0.026

0.052*

Standard error

(0.010)

(0.016)

(0.019)

(0.028)

Number of treated units

509

332

201

85

Note This table reports the productivity effect of starting to invest abroad grouped by ownership, absorptive capacity and investment destination of parent firms. An augmented Olley-Pakes approach has been used here, and standard errors are reported in the parentheses. * , ** , *** indicate significance level at 10%, 5% and 1%, respectively

Fig. 5.2 Moderating effect of state ownership, absorptive capacity and investment destination

5.4 Estimation Results

153

5.4.3 Absorptive Capacity and OFDI’s Productivity Effect on EMEs To test whether a firm’s absorptive capability matters, we split the sample into four groups according to whether firms have conducted OFDI and whether they have had pre-OFDI R&D.18 The matching approach is conducted on a year-by-year and industry-by-industry basis. We treat firms with (without) positive R&D expenditure before year t as the control group for firms that starting OFDI in year t and with (without) positive R&D before year t respectively19 . Table 5.4 and Fig. 5.2 show on average OFDI promotes the parent firm’s productivity growth no matter whether it has pre-OFDI R&D or not. However, the productivity effect differentiates significantly according to firms’ absorptive capability. Based on the estimation results of TFP, for EMEs that have positive pre-OFDI R&D expenditure, the productivity premiums are highly significant, and OFDI engagement brings in 2.1% higher productivity growth in the first year, than firms without OFDI. Till the third year after (s = 3) starting OFDI, the productivity premium for EMEs with strong absorptive capacity turns to be larger and reached 16.9%. However, for firms without pre-OFDI R&D, the productivity growth benefits brought by OFDI are only significant in early years (the first and second year after OFDI, i.e., only significant when s = 1 & 2), and their productivity growth rate is noticeably lower than EMEs with positive pre-OFDI R&D expenditure. Thus H3 is supported.

5.4.4 Investment Destination and OFDI’s Productivity Effect on EMEs We test H4 by distinguishing EMEs that conduct OFDI in OECD countries only20 from those in non-OECD countries only.21 Our results in Table 5.4 and Fig. 5.2 support H4 by demonstrating that OFDI’s productivity effect significantly exists no 18

To be accurate, for those that have never invested abroad, we split them based on whether they had R&D prior to that year within each industry for each year. 19 An alternative method to test the role of absorptive capability in moderating the productivity effect of OFDI is to directly split the matched results from Sect. 5.1 into two groups by firms’ pre-OFDI R&D status. But it may overestimate the productivity effect for firms that had pre-OFDI R&D, compared to our approach. 20 Members of OECD countries used in this chapter are restricted to those that had joined OECD before 2009, because of our sample period. For more information about the list of OECD members, please refer to http://www.oecd.org/about/membersandpartners/list-oecd-membercountries.htm. 21 In order to get rid of the mixed effect generated by firms that invest both in OECD countries and in non-OECD countries during the starting year, we drop all the observations of those firms in this section.

154

5 Does Outward FDI Generate Higher Productivity for Emerging …

matter EMEs invest in OECD or non-OECD countries, but this OFDI-led productivity gain is clearly higher if investing in OECD countries. Hence, H4 is supported.

5.5 Robustness Check and Further Analysis 5.5.1 An Alternative Measure of Total Factor Productivity In order to check whether our above estimation results are robust to different productivity estimation methods, we have re-estimated firms’ productivity effect using the Levinsohn and Petrin (2003) approach (LP method). This method employs an EME’s intermediate inputs as a proxy for unobservable productivity, to control for the correlation between firms’ inputs decisions and invisible productivity, thus solving the simultaneity bias when estimating the production function. After the firm-level productivity estimation, we test the OFDI-led productivity effect using the same method in Sect. 5.3, and the results are similar to those in Sect. 5.4.22

5.5.2 Investment Destination Measured by Patent Application Per Capita To test whether our estimation results of Sect. 5.4 are robust to different criteria of the host country division, we have divided our sample into two groups based on host counties’ technology levels measured by patent applications per capita. We have averaged each host country’s patent applications per capita during the period 2002– 2008, and then compared them with the overall average, to evaluate whether a country is high-tech or low-tech (a country with above-total-average patent application per capita will be labeled as a high-tech country and otherwise a low-tech country). The estimation results23 are again similar to those in Sect. 5.4.

5.5.3 One-Step System GMM Approach to Estimate the OFDI’s Productivity Effect Given the flexibility of one-step system-GMM, according to Blundell and Bond (1998) and Yu (2015), we have examined OFDI-led productivity growth directly

22 23

The detailed results are not reported due to space limitation, but are available upon request. The detailed results are not reported there due to space limitation, but available upon request.

5.6 Discussion and Conclusion

155

without the pre-estimation of EMEs’ productivity.24 Thus the coefficients of inputs and OFDI in production function are estimated simultaneously, as an extra robustness check. The results25 show that at the overall manufacturing level, OFDI indeed promotes EMEs’ productivity growth. But the productivity effect is moderated by firm heterogeneity and investment strategy. EMEs that have pre-OFDI R&D and non-SOE ownership gain more OFDI-led productivity growth. At the same time, investing in OECD countries helps EMEs gain higher productivity premium. In all the estimation specifications, SO-EMEs are less efficient.

5.5.4 Absorptive Capacity and OFDI’s Productivity Effect in Non-technology-Intensive Industries To check whether firms’ absorptive capacity moderates OFDI’s productivity in nontechnology-intensive industries, we have conducted another test for OFDI’s productivity effect in non-technology-intensive industries. Based on the same estimation methods above, we find that firms with pre-OFDI R&D become more productive than those without after investing in non-technology-intensive industries abroad.26 This indicates that absorptive capacity does play a role in productivity improvement not only for firms seeking for advanced technology but also for those seeking for resources.

5.6 Discussion and Conclusion Given the mixed empirical results about the impact of OFDI on EMEs’ productivity change, we contribute to the literature by establishing a novel theoretical framework combining RBV and IT, and assessing whether there exists a positive OFDI-EMEs’ productivity growth nexus. The moderating effect of firm heterogeneity in terms of state ownership, absorptive capacity and investment destination has been considered. An augmented Olley and Pakes’ (1992) semi-parametric approach has been used as the TFP measurement to control for omitted variable bias and the propensity-score matching and difference-in-difference (DID) approaches have been combined to test out conceptual framework. We feel that this study has the following theoretical, policy and managerial implications.

24

In fact, this is to treat OFDI as a component of TFP, and explicitly test whether starting to invest abroad can promote parent firms’ productivity keeping other production factors unchanged. 25 The detailed results are not reported here due to space limitation, but available upon request. 26 The results are available upon request.

156

5 Does Outward FDI Generate Higher Productivity for Emerging …

5.6.1 Theoretical Implications First, we have focused on an emerging economy context and added to an underresearched area by combing RBV and IT to predict EMEs’ productivity gain from OFDI (Hoskisson et al., 2013; Peng et al., 2008). Existing research about OFDI’ productivity effect on EMEs tends to be based on a general literature review or “international business theory”, and the empirical results are mixed (Cozza et al., 2015; Hijzen et al., 2007; Lee et al., 2013; Yang et al., 2013; Zhao et al., 2010). Given the distinctive resource-based and institutional characteristics of EMEs (Hoskisson et al., 2013; Wang et al., 2012a, 2012b), we argue that OFDI-led productivity growth can be expected for EMEs as OFDI helps EMEs (1) create, transfer and absorb knowledge; (2) reallocate resources and realize economies of scale; (3) access developed institutions and institutional infrastructure; (4) get exposed to international competition. This argument has been supported by our estimation results, and our study thus contributes by confirming the existence of a positive OFDI-EMEs’ productivity growth nexus (Chen & Tang, 2014; Chen et al., 2012; Lee et al., 2013; Li et al., 2016; Masso & Vahter, 2008). Unlike previous studies that treat OFDI-EMEs’ productivity growth as a direct linkage (Herzer, 2011), our second contribution lies in identifying and documenting the role of firm heterogeneity in moderating. OFDI’s productivity effect. Our overarching argument is that although EMEs turn to be generally more productive after they conduct OFDI, this productivity effect varies depending on EME heterogeneity: (1) An EME without state ownership gains more positive productivity premium via OFDI than that with state ownership; (2) The stronger the EME’s absorptive capacity, the more positive productivity premium it can get from OFDI; (3) An EME investing in developed countries gains more from OFDI-led productivity enhancement. Our new theoretical framework extends the existing literature as it does not just look at the direct impact of OFDI on firm productivity, but also examines how the OFDI-productivity relationship is altered when firm heterogeneity is introduced. This makes an original contribution to the ongoing debate about OFDI’s productivity effect. Third, we find evidence that private EMEs can gain positive productivity premium via OFDI while SO-EMEs cannot. This result challenges RBV which indicates that SO-EMEs with more institution— based resources should perform better in global markets (Wang et al., 2012a, 2012b), and supports our argument that IT is needed to explain EMEs’ productivity gains from OFDI. With affiliation to governments, SO-EMEs are confronted with more stressful home country institutional pressures as their high resource—dependence on home country governments pushes them into serving for national politic goals. At the same time, being recognized as political actors, host country institutions exert huge pressures on SO-EMEs, preventing them from performing resource-augmenting activities effectively (Cui & Jiang, 2012). Fourth, we enrich the existing literature related to absorptive capability by recognizing its positive moderating effect on the OFDI-EMEs’ productivity linkage, based on both RBV and IT. We confirm the role absorptive capability plays in shaping

5.6 Discussion and Conclusion

157

EMEs’ recognition, assimilation and application to commercial ends of external valuable knowledge and information (Barney, 2001; Deng, 2007). Apart from that, our work suggests that EMEs’ absorptive capability works as resource -based capabilities, assisting EMEs in dealing with host country institutional pressures and surviving in asset-intensive developed institutions. Finally, the moderating effect of OFDI destination has also been identified. In line with both RBV and IT, we demonstrate that developed countries with agglomerated high-tech and well-developed institutional infrastructure (Hoskisson et al., 2013), offer EMEs with more possibilities for productivity enhancement.

5.6.2 Policy and Managerial Implications Our findings have important practical implications for EMEs’ productivityaugmenting OFDI activities, as well as emerging economies’ OFDI and R&D policies. Firstly, emerging economy governments need to realize that government intervention may sometimes be counter-productive. State ownership often implies that EMEs are supplied with institution-based resources. While this may offer EMEs specific advantages when they internationalize, this support may lead to low productivity if SO-EMEs behave as political actors, and are hence insensitive to market competition. It may be more useful for emerging economy governments to unfasten the political shackles for SO-EMEs, helping and encouraging EMEs to compete effectively in the global market via supplying market and network information, rather than providing excessive financial support. Secondly, emerging economy governments need to pay more attention to the development of their institutional infrastructure including the construction of R&D centers and cultivation of human capital, to enhance emerging economy firms’ absorptive capability. This not only facilitates domestic innovations, but also enlarges OFDI’s productivity effect on EMEs. Thirdly, our findings send EME managers a clear message that productivity premium EMEs can gain from OFDI is by no means automatic, and it varies significantly with firm-level heterogeneity, including firm specific resources, institutional background and corresponding OFDI strategies. Following this, when formulating OFDI strategies and decisions, managers should be aware of their firms’ features, strengths, and their investment purposes (asset-, market-, resource-, or efficiencyseeking), to maximize the benefits they can achieve from OFDI. For instance, an EME that has neither pre-OFDI R&D nor significant human capital may not acquire high-tech firms.

158

5 Does Outward FDI Generate Higher Productivity for Emerging …

5.6.3 Limitations and Future Research Directions As with all studies, our work has several limitations, which provide opportunities for future research. First, our dataset lacks detailed information on EMEs’ entry strategies. This hindered our ability to investigate the role of entry strategy as a moderator for the OFDI-productivity nexus. Existing literature indicates that entry strategy differences affect subsidiaries’ managerial pattern, corporate culture, techniquelearning channels, and their OFDI results (Nocke & Yeaple, 2007). However, the impact of entry strategy on EME productivity is not straightforward. OFDI via either greenfield or M&A brings in costs as well as channels for EMEs’ productivity growth (Dikova & Brouthers, 2016; Pradhan & Singh, 2008), but quantitative studies which simply measure entry strategy as greenfield or M&A cannot fully uncover the real productivity effect of OFDI entry strategy, let alone thoroughly explore the mechanisms with which different entry strategies moderate this effect. Future research of this topic via a qualitative method is sorely needed and strongly encouraged. Second, as our study is conducted based on data from 2002–2008, we cannot figure out the impact of the 2008 financial crisis on the relationship between OFDI and EMEs’ productivity growth. The financial crisis, with its associated credit crunch, has affected institutional environments, economic entities and EMEs’ OFDI abilities (Sauvant, Maschek & McAllister, 2010). For future research, it would be interesting to find out how the financial crisis affects EMEs’ OFDI trajectories and results. Another limitation of this study lies in the absence of detailed subsidiary level data. Thus, further subsidiary-level studies in our topic are highly encouraged, to clearly track the mechanisms with which EMEs’ OFDI enhance parent firms’ productivity (Rugman et al., 2011).

References Ahmed, Z. U., Mohamad, O., Tan, B., & Johnson, J. P. (2002). International risk perceptions and mode of entry: A case study of Malaysian multinational firms. Journal of Business Research, 55(10), 805–813. Alcacer, J., & Oxley, J. (2014). Learning by supplying. Strategic Management Journal, 35(2), 204–223. Almeida, P., & Kogut, B. (1999). Localization of knowledge and the mobility of engineers in regional networks. Management Science, 45(7), 905–917. Amighini, A. A., Rabellotti, R., & Sanfilippo, M. (2013). Do Chinese state-owned and private enterprises differ in their internationalization strategies? China Economic Review, 27, 312–325. Arnold, J. M., & Javorcik, B. S. (2005). Gifted kids or pushy parents? Foreign acquisitions and plant performance in Indonesia. Barney, J. B. (2001). Is the resource-based “view” a useful perspective for strategic management research? Yes. Academy of Management Review, 26(1), 41–56. Barney, J. B., Ketchen, D. J., Jr., & Wright, M. (2011). The future of resource-based theory: Revitalization or decline? Journal of Management, 37(5), 1299–1315. Bartelsman, E. J., & Doms, M. (2000). Understanding productivity: Lessons from longitudinal microdata. Journal of Economic Literature, 38(3), 569–594.

References

159

Bascle, G. (2008). Controlling for endogeneity with instrumental variables in strategic management research. Strategic Organization, 6(3), 285–327. Becker, S. O., & Ichino, A. (2002). Estimation of average treatment effects based on propensity scores. The Stata Journal, 2(4), 358–377. Bertrand, O., & Capron, L. (2015). Productivity enhancement at home via cross-border acquisitions: The roles of learning and contemporaneous domestic investments. Strategic Management Journal, 36(5), 640–658. Bitzer, J., & Görg, H. (2009). Foreign direct investment, competition and industry performance. World Economy, 32(2), 221–233. Bitzer, J., & Kerekes, M. (2008). Does foreign direct investment transfer technology across borders? New evidence. Economics Letters, 100(3), 355–358. Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics, 87(1), 115–143. Blundell, R., & Dias, M. C. (2009). Alternative approaches to evaluation in empirical microeconomics. Journal of Human Resources, 44(3), 565–640. Braconier, H., Ekholm, K., & Knarvik, K. H. M. (2001). In search of FDI-transmitted R&D spillovers: A study based on Swedish data. Review of World Economics, 137(4), 644–665. Brandt, L., Van Biesebroeck, J., & Zhang, Y. (2012). Creative accounting or creative destruction? Firm-level productivity growth in Chinese manufacturing. Journal of Development Economics, 97(2), 339–351. Branstetter, L. (2006). Is foreign direct investment a channel of knowledge spillovers? Evidence from Japan’s FDI in the United States. Journal of International Economics, 68(2), 325–344. Brockman, P., Rui, O. M., & Zou, H. (2013). Institutions and the performance of politically connected M&As. Journal of International Business Studies, 44(8), 833–852. Buckley, P. J., Clegg, L. J., Cross, A. R., Liu, X., Voss, H., & Zheng, P. (2007). The determinants of Chinese outward foreign direct investment. Journal of International Business Studies, 38(4), 499–518. Buckley, P. J., Cross, A. R., Tan, H., Xin, L., & Voss, H. (2008). Historic and emergent trends in Chinese outward direct investment. Management International Review, 48(6), 715–748. Cantwell, J., & Iammarino, S. (2000). Multinational corporations and the location of technological innovation in the UK regions. Regional Studies, 34(4), 317–332. Cantwell, J., & Janne, O. (1999). Technological globalisation and innovative centres: The role of corporate technological leadership and locational hierarchy. Research Policy, 28(2–3), 119–144. Cantwell, J., Dunning, J. H., & Lundan, S. M. (2010). An evolutionary approach to understanding international business activity: The co-evolution of MNEs and the institutional environment. Journal of International Business Studies, 41(4), 567–586. Causa, O., & Cohen, D. (2004). Overcoming barriers to competitiveness. OECD Development Centre Working Papers. Chacar, A. S., Newburry, W., & Vissa, B. (2010). Bringing institutions into performance persistence research: Exploring the impact of product, financial, and labor market institutions. Journal of International Business Studies, 41(7), 1119–1140. Chen, W., & Tang, H. (2014). The dragon is flying west: Micro-level evidence of Chinese outward direct investment. Asian Development Review, 31(2), 109–140. Chen, K. M., & Yang, S. F. (2013). Impact of outward foreign direct investment on domestic R&D activity: evidence from Taiwan’s multinational enterprises in low-wage countries. Asian Economic Journal, 27(1), 17–38. Chen, Y. Y., & Young, M. N. (2010). Cross-border mergers and acquisitions by Chinese listed companies: A principal–principal perspective. Asia Pacific Journal of Management, 27(3), 523– 539. Chen, V. Z., Li, J., & Shapiro, D. M. (2012). International reverse spillover effects on parent firms: Evidences from emerging-market MNEs in developed markets. European Management Journal, 30(3), 204–218.

160

5 Does Outward FDI Generate Higher Productivity for Emerging …

Choudhury, P., & Khanna, T. (2014). Toward resource independence–Why state-owned entities become multinationals: An empirical study of India’s public R&D laboratories. Journal of International Business Studies, 45(8), 943–960. Chuang, Y. C., & Lin, C. M. (1999). Foreign direct investment, R&D and spillover efficiency: Evidence from Taiwan’s manufacturing firms. The Journal of Development Studies, 35(4), 117– 137. Chung, W., & Alcácer, J. (2002). Knowledge seeking and location choice of foreign direct investment in the United States. Management Science, 48(12), 1534–1554. Clarides, S. L., & Tybout, J. (1998). Is learning by exporting important. The Quarterly Journal of Economics, 454, 903–947. Cohen, W. M., & Levinthal, D. A. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly, 128–152. Cohen, W. M., & Levinthal, D. A. (1989). Innovation and learning: The two faces of R & D. The Economic Journal, 99(397), 569–596. Contractor, F. J. (2012). Why do multinational firms exist? A theory note about the effect of multinational expansion on performance and recent methodological critiques. Global Strategy Journal, 2(4), 318–331. Cozza, C., Rabellotti, R., & Sanfilippo, M. (2015). The impact of outward FDI on the performance of Chinese firms. China Economic Review, 36, 42–57. Cuervo-Cazurra, A., & Dau, L. A. (2009). Promarket reforms and firm profitability in developing countries. Academy of Management Journal, 52(6), 1348–1368. Cui, L., & Jiang, F. (2009). FDI entry mode choice of Chinese firms: A strategic behavior perspective. Journal of World Business, 44(4), 434–444. Cui, L., & Jiang, F. (2012). State ownership effect on firms’ FDI ownership decisions under institutional pressure: A study of Chinese outward-investing firms. Journal of International Business Studies, 43(3), 264–284. Dai, M., & Yu, M. (2013). Firm R&D, absorptive capacity and learning by exporting: Firm-level evidence from China. The World Economy, 36(9), 1131–1145. Damijan, J. P., Polanec, S., & Prašnikar, J. (2007). Outward FDI and productivity: Micro-evidence from Slovenia. World Economy, 30(1), 135–155. Datta, D. K., Guthrie, J. P., & Wright, P. M. (2005). Human resource management and labor productivity: Does industry matter? Academy of Management Journal, 48(1), 135–145. De Loecker, J. (2007). Do exports generate higher productivity? Evidence from Slovenia. Journal of International Economics, 73(1), 69–98. De Loecker, J. (2011). Product differentiation, multiproduct firms, and estimating the impact of trade liberalization on productivity. Econometrica, 79(5), 1407–1451. De Loecker, J., & Warzynski, F. (2012). Markups and firm-level export status. American Economic Review, 102(6), 2437–2471. Deng, P. (2007). Investing for strategic resources and its rationale: The case of outward FDI from Chinese companies. Business Horizons, 50(1), 71–81. Dikova, D., & Brouthers, K. (2016). International establishment mode choice: Past, present and future. Management International Review, 56(4), 489–530. DiMaggio, P. J., & Powell, W. W. (1983). The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review, 147–160. Dosi, G. (1988). The nature of the innovative process. Technical Change and Economic Theory, 2, 590–607. Driffield, N., & Chiang, P. C. (2009). The effects of offshoring to China: Reallocation, employment and productivity in Taiwan. International Journal of the Economics of Business, 16(1), 19–38. Driffield, N., Love, J. H., & Taylor, K. (2009). Productivity and labour demand effects of inward and outward foreign direct investment on UK industry. The Manchester School, 77(2), 171–203. Fan, J. P., Wong, T. J., & Zhang, T. (2007). Politically connected CEOs, corporate governance, and Post-IPO performance of China’s newly partially privatized firms. Journal of Financial Economics, 84(2), 330–357.

References

161

Feenstra, R. C., Li, Z., & Yu, M. (2014). Exports and credit constraints under incomplete information: Theory and evidence from China. Review of Economics and Statistics, 96(4), 729–744. Fosfuri, A., Motta, M., & Rønde, T. (2001). Foreign direct investment and spillovers through workers’ mobility. Journal of International Economics, 53(1), 205–222. Frost, T. S. (2001). The geographic sources of foreign subsidiaries’ innovations. Strategic Management Journal, 22(2), 101–123. Frost, S. (2004). Chinese outward direct investment in Southeast Asia: How big are the flows and what does it mean for the region? The Pacific Review, 17(3), 323–340. Globerman, S., & Shapiro, D. (2009). Economic and strategic considerations surrounding Chinese FDI in the United States. Asia Pacific Journal of Management, 26(1), 163–183. Goodarzi, H., & Moghadam, S. E. (2014). Foreign direct investment and its effects on home country: Evidence from developing countries. Journal of Knowledge Globalization, 7(1). Görg, H., Hanley, A., & Strobl, E. (2008). Productivity effects of international outsourcing: Evidence from plant-level data. Canadian Journal of Economics/revue Canadienne D’économique, 41(2), 670–688. Griffith, R., Redding, S., & Reenen, J. V. (2004). Mapping the two faces of R&D: Productivity growth in a panel of OECD industries. Review of Economics and Statistics, 86(4), 883–895. Grossman, G. M., Helpman, E., & Szeidl, A. (2006). Optimal integration strategies for the multinational firm. Journal of International Economics, 70(1), 216–238. Hamilton, B. H., & Nickerson, J. A. (2003). Correcting for endogeneity in strategic management research. Strategic Organization, 1(1), 51–78. Hayakawa, K., Matsuura, T., Motohashi, K., & Obashi, A. (2013). Two-dimensional analysis of the impact of outward FDI on performance at home: Evidence from Japanese manufacturing firms. Japan and the World Economy, 27, 25–33. He, W., & Lyles, M. A. (2008). China’s outward foreign direct investment. Business Horizons, 51(6), 485–491. Head, K., Ries, J., & Swenson, D. (1995). Agglomeration benefits and location choice: Evidence from Japanese manufacturing investments in the United States. Journal of International Economics, 38(3–4), 223–247. Heckman, J. J., Ichimura, H., & Todd, P. E. (1997). Matching as an econometric evaluation estimator: Evidence from evaluating a job training programme. The Review of Economic Studies, 64(4), 605–654. Helpman, E., Melitz, M. J., & Yeaple, S. R. (2004). Export versus FDI with heterogeneous firms. American Economic Review, 94(1), 300–316. Herrerias, M. J., & Orts, V. (2012). Equipment investment, output and productivity in China. Empirical Economics, 42(1), 181–207. Herzer, D. (2008). The long-run relationship between outward FDI and domestic output: Evidence from panel data. Economics Letters, 100(1), 146–149. Herzer, D. (2010). Outward FDI and economic growth. Journal of Economic Studies, 37, 476–494. Herzer, D. (2011). The long-run relationship between outward foreign direct investment and total factor productivity: Evidence for developing countries. The Journal of Development Studies, 47(5), 767–785. Herzer, D. (2012). Outward FDI, total factor productivity and domestic output: Evidence from Germany. International Economic Journal, 26(1), 155–174. Hijzen, A., Jean, S., & Mayer, T. (2011). The effects at home of initiating production abroad: Evidence from matched French firms. Review of World Economics, 147(3), 457–483. Hijzen, A., Inui, T., & Tod¯o, Y. (2007). The effects of multinational production on domestic performance: Evidence from Japanese firms. RIETI. Hoskisson, R. E., Wright, M., Filatotchev, I., & Peng, M. W. (2013). Emerging multinationals from mid-range economies: The influence of institutions and factor markets. Journal of Management Studies, 50(7), 1295–1321. Imbens, G. W. (2004). Nonparametric estimation of average treatment effects under exogeneity: A review. Review of Economics and Statistics, 86(1), 4–29.

162

5 Does Outward FDI Generate Higher Productivity for Emerging …

Isaksson, A. (2007). Determinants of total factor productivity: A literature review. Research and Statistics Branch, UNIDO, 1, 101. Jong-Sung, Y., & Khagram, S. (2005). A comparative study of inequality and corruption. American Sociological Review, 70(1), 136–157. Khanna, T., & Rivkin, J. W. (2001). Estimating the performance effects of business groups in emerging markets. Strategic Management Journal, 22(1), 45–74. Kharas, H., & Kohli, H. (2011). What is the middle income trap, why do countries fall into it, and how can it be avoided? Global Journal of Emerging Market Economies, 3(3), 281–289. Kimura, F., & Lee, H. H. (2006). The gravity equation in international trade in services. Review of World Economics, 142(1), 92–121. Koch, M. J., & McGrath, R. G. (1996). Improving labor productivity: Human resource management policies do matter. Strategic Management Journal, 17(5), 335–354. Kogut, B., & Chang, S. J. (1991). Technological capabilities and Japanese foreign direct investment in the United States. The Review of Economics and Statistics, 401–413. Kokko, A., & Kravtsova, V. (2008). Innovative capability in MNC subsidiaries: Evidence from four European transition economies. Post-Communist Economies, 20(1), 57–75. Lee, Y., Chyi, Y. L., Lin, E. S., & Wu, S. Y. (2013). Do local industrial agglomeration and foreign direct investment to China enhance the productivity of Taiwanese firms? The Journal of International Trade & Economic Development, 22(6), 839–865. Levinsohn, J., & Petrin, A. (2003). Estimating production functions using inputs to control for unobservables. The Review of Economic Studies, 70(2), 317–341. Levinthal, D. A., & Wu, B. (2010). Opportunity costs and non-scale free capabilities: Profit maximization, corporate scope, and profit margins. Strategic Management Journal, 31(7), 780–801. Li, J. (1995). Foreign entry and survival: Effects of strategic choices on performance in international markets. Strategic Management Journal, 16(5), 333–351. Li, J., Strange, R., Ning, L., & Sutherland, D. (2016). Outward foreign direct investment and domestic innovation performance: Evidence from China. International Business Review, 25(5), 1010–1019. Liang, H., Ren, B., & Sun, S. L. (2015). An anatomy of state control in the globalization of state-owned enterprises. Journal of International Business Studies, 46(2), 223–240. Lieberman, M. B., & Dhawan, R. (2005). Assessing the resource base of Japanese and US auto producers: A stochastic frontier production function approach. Management Science, 51(7), 1060–1075. Lu, J., Liu, X., & Wang, H. (2011). Motives for outward FDI of Chinese private firms firm resources, industry dynamics, and government policies. Management and Organization Review, 7(2), 223– 248. Luo, Y., & Rui, H. (2009). An ambidexterity perspective toward multinational enterprises from emerging economies. Academy of Management Perspectives, 23(4), 49–70. Luo, Y., & Tung, R. L. (2007). International expansion of emerging market enterprises: A springboard perspective. Journal of International Business Studies, 38(4), 481–498. Luo, Y., Xue, Q., & Han, B. (2010). How emerging market governments promote outward FDI: Experience from China. Journal of World Business, 45(1), 68–79. Masso, J., & Vahter, P. (2008). Technological innovation and productivity in late-transition Estonia: Econometric evidence from innovation surveys. The European Journal of Development Research, 20(2), 240–261. Mathews, J. A. (2002). Dragon multinational: A new model for global growth. Oxford University Press Melitz, M. J. (2003). The impact of trade on intra-industry reallocations and aggregate industry productivity. econometrica, 71(6), 1695–1725. Meyer, K. E., Wright, M., & Pruthi, S. (2009). Managing knowledge in foreign entry strategies: A resource-based analysis. Strategic Management Journal, 30(5), 557–574.

References

163

Meyer, K. E., Ding, Y., Li, J., & Zhang, H. (2014). Overcoming distrust: How state-owned enterprises adapt their foreign entries to institutional pressures abroad. Journal of International Business Studies, 45(8), 1005–1028. Navaretti, G. B., & Castellani, D. (2004). Investments abroad and performance at home: Evidence from Italian multinationals. Available at SSRN 527562. Nocke, V., & Yeaple, S. (2007). Cross-border mergers and acquisitions vs. greenfield foreign direct investment: The role of firm heterogeneity. Journal of International Economics, 72(2), 336–365. Oliver, C. (1991). Strategic responses to institutional processes. Academy of Management Review, 16(1), 145–179. Olley, S., & Pakes, A. (1992). The dynamics of productivity in the telecommunications equipment industry. Econometrica, 64(6), 1263–1297. Peng, M. W. (2003). Institutional transitions and strategic choices. Academy of Management Review, 28(2), 275–296. Peng, M. W., Wang, D. Y., & Jiang, Y. (2008). An institution-based view of international business strategy: A focus on emerging economies. Journal of International Business Studies, 39(5), 920–936. Potterie, B. V. P. D. L., & Lichtenberg, F. (2001). Does foreign direct investment transfer technology across borders? Review of Economics and Statistics, 83(3), 490–497. Pradhan, J. P., & Singh, N. (2008). Outward FDI and knowledge flows: A study of the Indian automotive sector. International Journal of Institutions and Economies, 1(1), 155–186. Ramasamy, B., Yeung, M., & Laforet, S. (2012). China’s outward foreign direct investment: Location choice and firm ownership. Journal of World Business, 47(1), 17–25. Rangan, S. (2000). The problem of search and deliberation in economic version: When networks really matter. Academy of Management Review, 25(4), 813–828. Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70(1), 41–55. Rosenkopf, L., & Almeida, P. (2003). Overcoming local search through alliances and mobility. Management Science, 49(6), 751–766. Rugman, A. M., & Li, J. (2007). Will China’s multinationals succeed globally or regionally? European Management Journal, 25(5), 333–343. Rugman, A. M., Verbeke, A., & Nguyen, Q. T. (2011). Fifty years of international business theory and beyond. Management International Review, 51(6), 755–786. Salancik, G. R., & Pfeffer, J. (1978). A social information processing approach to job attitudes and task design. Administrative Science Quarterly, 224–253 Salomon, R., & Shaver, J. M. (2005). Export and domestic sales: Their interrelationship and determinants. Strategic Management Journal, 26(9), 855–871. Sauvant, K. P., Maschek, W. A., & McAllister, G. (2010). Foreign direct investment by emerging market multinational enterprises, the impact of the financial crisis and recession, and challenges ahead. In Foreign direct investments from emerging markets (pp. 3–29). Palgrave Macmillan. Sawada, N. (2010). Technology gap matters on spillover. Review of Development Economics, 14(1), 103–120. Scott, W. R. (1995). Institutions and Organizations (Vol. 2). Sage Thousand Oaks Scott, W. R. (2002). The changing world of Chinese enterprise: An institutional perspective. In The management of enterprises in the People’s Republic of China (pp. 59–78). Springer. Shan, W., & Song, J. (1997). Foreign direct investment and the sourcing of technological advantage: Evidence from the biotechnology industry. Journal of International Business Studies, 28(2), 267–284. Sirmon, D. G., Hitt, M. A., Ireland, R. D., & Gilbert, B. A. (2011). Resource orchestration to create competitive advantage: Breadth, depth, and life cycle effects. Journal of Management, 37, 1390–1412. Syverson, C. (2011). What determines productivity? Journal of Economic Literature, 49(2), 326– 365.

164

5 Does Outward FDI Generate Higher Productivity for Emerging …

Teece, D. J. (1992). Competition, cooperation, and innovation: Organizational arrangements for regimes of rapid technological progress. Journal of Economic Behavior & Organization, 18(1), 1–25. Teece, D. J. (2014). A dynamic capabilities-based entrepreneurial theory of the multinational enterprise. Journal of International Business Studies, 45(1), 8–37. Ulubasoglu, M. A., & Doucouliagos, C. (2004, February). Institutions and economic growth: A systems approach. In Econometric Society 2004 Australasian Meetings Paper (Vol. 63, pp. 1–33). Wan, K. M., & Wong, K. F. (2009). Economic impact of political barriers to cross-border acquisitions: An empirical study of CNOOC’s unsuccessful takeover of Unocal. Journal of Corporate Finance, 15(4), 447–468. Wang, C., Hong, J., Kafouros, M., & Wright, M. (2012a). Exploring the role of government involvement in outward FDI from emerging economies. Journal of International Business Studies, 43(7), 655–676. Wang, C., Hong, J., Kafouros, M., & Boateng, A. (2012b). What drives outward FDI of Chinese firms? Testing the explanatory power of three theoretical frameworks. International Business Review, 21(3), 425–438. Wang, J., Liu, X., Wei, Y., & Wang, C. (2014). Cultural proximity and local firms’ catch up with multinational enterprises. World Development, 60, 1–13. Yang, S. F., Chen, K. M., & Huang, T. H. (2013). Outward foreign direct investment and technical efficiency: Evidence from Taiwan’s manufacturing firms. Journal of Asian Economics, 27, 7–17. Yu, M. (2015). Processing trade, tariff reductions and firm productivity: Evidence from Chinese firms. The Economic Journal, 125(585), 943–988. Zahra, S. A., & George, G. (2002). Absorptive capacity: A review, reconceptualization, and extension. Academy of Management Review, 27(2), 185–203. Zahra, S. A., & Hayton, J. C. (2008). The effect of international venturing on firm performance: The moderating influence of absorptive capacity. Journal of Business Venturing, 23(2), 195–220. Zhang, J., Zhou, C., & Ebbers, H. (2011). Completion of Chinese overseas acquisitions: Institutional perspectives and evidence. International Business Review, 20(2), 226–238. Zhao, W., Liu, L., & Zhao, T. (2010). The contribution of outward direct investment to productivity changes within China, 1991–2007. Journal of IntErnational Management, 16(2), 121–130.

Chapter 6

Outward Direct Investment, Firm Productivity and Credit Constraints: Evidence from Chinese Firms

Abstract China is currently the third largest country in terms of outward direct investment (ODI), with the investors mainly being state-owned enterprises. This presents a question: What inhibits private enterprises from increasing ODI ? Using a firm-level panel data set for Zhejiang Province in China, we examine the impact of firm heterogeneity on private firm ODI. We have three main findings: first, a higher productivity level contributes to better access to ODI, and increases ODI value as well; second, lowering a firm’s financial constraint level can increase both the probability and volume of ODI; third, productivity cannot offset the negative effect of financial constraint on private firm ODI.

6.1 Introduction China is currently the third largest country in terms of outward direct investment (ODI). Since China’s Ministry of Commerce started to report annual data on ODI in 2003, the flows of China’s ODI have successively increased. The average annual growth of ODI from 2002 to 2013 was 39.8%. While ODI in the world decreased by 18% in 2012, ODI value from China grew by 17.6%, hitting a record US$84bn, and, for the first time, China became the third largest country in terms of ODI value, right behind the United States (US$329bn) and Japan (US$123bn). In 2013, China’s ODI grew even higher, reaching its highest level of $107.8 bn. Another significant feature of China’s ODI is that state -owned enterprises (SOEs) play an important role, especially central SOEs, which, by definition, are controlled by the central government. For instance, central SOEs’ non-financial ODI flows amounted to US$43.524bn in 2012, accounting for 56% of China’s total non-financial ODI flows. This raises some interesting questions: why are SOEs the primary overseas investors from China and what inhibits the ODI of private enterprises?

This chapter is published in Pacific Economic Review by Wang Bijun, Yuyan Tan, Miaojie Yu, Yiping Huang. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022 W. Tian and M. Yu, Outward Foreign Direct Investment of Chinese Enterprises, Contributions to Economics, https://doi.org/10.1007/978-981-19-4719-3_6

165

166

6 Outward Direct Investment, Firm Productivity and Credit Constraints: Evidence …

Recent published literature shows that overseas market entry decisions are closely related to firm heterogeneity. The most studied kind of heterogeneity is firm productivity. Krugman (1980) first introduced firm heterogeneity into international trade models. Later on, Melitz (2003) built a model with firm heterogeneity to prove that only firms with high productivity enter the exporting market, while less productive firms remain in the domestic market, and the least productive firms exit the market altogether. Helpman et al. (2004) further introduced firm heterogeneity into a foreign direct investment model. Their empirical work on US firms found that lowproductivity firms serve the domestic markets, firms with higher productivity choose to export goods, and the most productive firms choose to invest in foreign markets. However, productivity is not the only determinant in internationalization behaviour of enterprises. Many productive firms only serve the domestic market and, likewise, some low productive firms export and invest overseas as well (Bernard et al., 2003). Figure 6.1 depicts productivity (total factor productivity and labour productivity) distributions of ODI firms and non-ODI firms in China. It confirms that the average productivity level of ODI firms is higher than that of non-ODI firms, but there is a large overlap between the distributions. Bernard et al. (2003), Mayer and Ottaviano and Todo (2011) found a similar phenomenon, respectively, in the USA, Belgium and Japan. Besides productivity, a firm’s financial constraints might influence its decision to enter overseas markets. More and more micro evidence shows that imperfect credit markets seriously restrict firms’ export capacity. Manova et al. identify a significant negative effect of credit constraint on firm exports. More specifically, financial friction limits the range of export products, the number of destinations and the total value of each bilateral export flow. Feenstra et al. (2014) develop a model to examine why credit constraints for domestic and exporting firms arise when banks do not observe firms’ productivity levels, and they find that export enterprises are facing tighter

Fig. 6.1 Productivity distributions of outward direct investment (ODI) firms and non-ODI firms in China

6.1 Introduction

167

credit constraints than purely domestic firms in China. Muuls, Berman and Héricourt (2010), Minetti and Zhu (2011), Li and Yu also observe similar results. Amiti and Weinstein examine how banks in Japan transferred negative shocks of financial crises to exporters during the 1990s, which overcame the endogeneity issues and established the casual link between financial constraint and firms’ exports. Besides the decision to export, Todo (2011) finds a negative effect of credit constraints on firms’ decisions regarding ODI. However, most of the related papers have only examined export behaviour (see Bernard et al., 2003; Girma et al., 2005). Besides, they are mainly based on the experience of developed countries, such as the USA (Helpman et al., 2004), Japan (Todo, 2011; Tomiura, 2007) and Korea (Lee, 2010). Studies on ODI from developing countries are limited. Damijan and Rojec (2007) use manufacturing data for Slovenia and find that the productivities of export and ODI firms are, on average, 20% higher than for domestic firms. Because Slovenia is a transition economy with large inefficient overseas projects, the authors do not find that ODI firms have higher productivity levels than export firms. Tian and Yu (2012) use a data set of Chinese industrial enterprises and show that firms’ total factor productivity positively contributes to their choice of ODI, and also increases their ODI value. There is even less research concerning the impact of financial constraint on firm ODI, let alone related studies on China. One reason is that although ODI from China has been growing fast since the financial crisis in 2008, the phenomenon is quite new, and China is still mainly a recipient of foreign capital. Another reason is that firm-level data on ODI is not publicly accessible. The existing published literature is mainly focused on macro-level cross-country analysis (Cheung & Qian, 2009; Contessi & De Pace, 2012; Wang & Huang, 2012). For example, Wang and Huang (2012) use annual ODI data for 22 industrialized countries and 44 developing countries (including China) during 1981–2005, and demonstrate that financial repression significantly improves ODI in developing countries. This chapter attempts to reveal the impact of productivity and financial constraints on firms’ internationalization behaviour, focusing on small and medium private enterprises. Financial constraints of small and medium private firms in China reflect institutional issues like financial repression and capital control. The present chapter provides some micro evidence calling for further reforms and structural changes in the economy. It also sheds light on the development of other emerging and developing countries. The rest of the chapter organized as follows. Section 6.2 describes the data and variables that we use, Sects. 6.3 and 6.4 examine the effect of firm heterogeneity on firms’ ODI decisions and ODI value, respectively. Section 6.5 concludes.

168

6 Outward Direct Investment, Firm Productivity and Credit Constraints: Evidence …

6.2 Data and Variables 6.2.1 Firm-Level Data in Zhejiang Province Outward direct investment in Zhejiang Province is representative of local private firm ODI throughout China. Although central SOEs were responsible for 83% of the value of non-financial ODI in China during 2003–2009, 92% of the value of ODI projects was through local private firms, with ODI projects from Zhejiang Province forming the largest part. Moreover, 70% of the value of private firm ODI in China is from Zhejiang and Fujian Provinces. Because investment decisions made by private firms are more driven by market demand, using firm-level data for Zhejiang Province can largely overcome the influence of policy issues, with results more comparable to international experience and the existing literature. Our main data set is firm-level ODI from Zhejiang Province during 2006–2008. It includes important information about firm ODI value, ODI destination, firm location and industry.1 In order to obtain other firm-level information like financial statements and export value, we merge the data set with data from the Chinese Industrial Enterprises Database. A few samples in this database are noisy due largely to misreporting by firms. Following Tian and Yu (2012), we clean the sample and delete outliers using the following criteria: (i) some key financial variables cannot be missing (such as total assets, sales and employment); and (ii) variables should not vio-late the general accepted accounting principles, such as the liquid assets exceeding the total assets or an invalid date of establishment. After the outlier filter, we obtain a sample of more than 40 000 manufacturing firms in Zhejiang Province during 2006–2008. The final data set sums up to 135 247 observations, which includes 526 observations with non-zero ODI value and 55 185 observations with zero ODI value and non-zero export value. Table 6.1 summarizes the ODI project amounts and the ODI value from Zhejiang Province by industry. Most of the small and medium-scale ODI projects are in manufacturing sectors. In the case of Zhejiang Province, 76.93% of the ODI projects are from manufacturing sectors, and they contribute 64.44% of the total value. Among them, ‘electronics, machinery and appliance’ and ‘textiles, clothing, footwear and leather’ are two main industries with high levels of ODI, accounting for 81.47% of the total amount of ODI projects and 87.41% total investment value in Zhejiang Province.

1

This dataset is provided by International Cooperation Office of Zhejiang Province.

6.2 Data and Variables

169

Table 6.1 Outward direct investment (ODI) summary by industry Projects Primary industry

63

ODI value (10 000$) 4.96%

22 326

12.77%

Agriculture

34

2.68%

8330

4.77%

Mining

29

2.28%

13 996

8.01%

Manufacturing industry

977

76.93%

11 2634

64.44%

Electronics, machinery and appliance

423

33.31%

42 835

24.51%

Textiles, clothing, footwear and leather

373

29.37%

55 624

31.82%

Chemical and pharmaceutical

52

4.09%

6077

3.48%

Others

129

10.16%

8098

4.63%

Service industry

163

12.83%

28 436

16.27%

Construction and real estate

39

3.07%

11 723

6.71%

Trade and business services

111

8.74%

11 786

6.74%

Other services

13

1.02%

4928

2.82%

Others

67

5.28%

11 384

6.51%

6.2.2 Variables Firm productivity is one of the most important determinants of firms’ decision to enter overseas markets (Greenaway & Kneller, 2007; Head & Ries, 2003; Helpman et al., 2004; Tian & Yu, 2012). We use output per worker to measure firm productivity, and keep firm capital intensity controlled. Labour productivity is a widely adopted measure of productivity (e.g. Helpman et al., 2004), and its use makes our results more comparable to existing studies. Financial constraint is our key variable. Although there is no perfect measure, the literature on corporate finance discusses several ways to measure financial constraint, including investment-cash flow sensitivity (Fazzari et al., 1988; Sun & Yamori, 2009), the Kaplan and Zingales index (Lamont et al., 2001), the Whited and Wu index (Whited & Wu, 2006) and the size–age index (Hadlock & Pierce, 2010). The construction of all the indexes relies on a pre-ranking of all the firms based on their characteristics. We use a synthetic index constructed by firm performance in several aspects (Bellone et al., 2010; Musso & Schiavo, 2008), including both internal funds and external funds (Myers, 1984). All of the variables we choose are perceived as important in determining financial constraint in existing published studies: 1. Cash reserves, as a large part of retaining earnings, reflects internal funds available for enterprises to invest. Therefore, the cash ratio (cash over total assets) is our first variable considered in the synthetic index. When this metric is higher, the firm has more internal funds to invest, and it also shows firms’ ability to pay back debts, which means a lower level of financial constraint.

170

6 Outward Direct Investment, Firm Productivity and Credit Constraints: Evidence …

2. Firm size is used to capture firms’ constraint on external funds. Firm size is measured by the logarithm of total assets. Larger firms usually have better access to external funds. 3. Solvency is also used to capture firms’ constraint on external funds. Solvency is calculated as equity over total liability, which shows the robustness of firms’ equityliability structure. A higher solvency index means lower financial constraint. To summarize, our synthetic index includes three sub-indicators of firm performance; that is, the cash ratio, firm size and firm solvency. Following Bellone et al. (2010) and Bottazzi et al. (2014), for each sub-indicator we sort the firms, then place them in one of the quintiles with a score ranging from one to five, with a score of one representing the smallest value. We sumup the five scores and standardize them to [0, 10], and then obtain our synthetic financial constraint index. Except for all the key variables we discussed above, some other firm characteristics are controlled in our regression: (i) tax ratio, measured by value-added tax over total sales; (ii) FDI dummy, equal to one if it is a foreign company, and zero otherwise; (iii) firm age, counted from its established year; and (iv) capital intensity, measured by net fixed assets over employment. Moreover, we employ year and industry dummies to control for time variance and industry differences. Table 6.2 presents the sample statistics of all the variables used. Table 6.2 Summary of statistics Variables

Domestic

Export

ODI (including ODI & export)

ODI value

—

—

3.346 (1.464)

Export value

—

9.488 (1.565)

10.927 (1.717)

Log (labour productivity)

4.078 (0.793)

3.939 (0.754)

4.281 (0.818)

Financial constraint

4.723 (2.175)

5.387 (2.073)

6.499 (1.703)

Sub-indicators of financial constraint Cash index

2.881 (1.420)

3.166 (1.389)

3.510 (1.250)

Size index

2.783 (1.374)

3.300 (1.413)

4.266 (1.123)

Solvency index

3.004 (1.430)

2.998 (1.393)

3.023 (1.287)

Tax ratio

3.426 (1.847)

2.662 (2.015)

2.430 (2.093)

FDI dummy

0.083 (0.276)

0.294 (0.455)

0.290 (0.454)

Firm age

8.550 (6.451)

8.696 (6.140)

9.397 (5.672)

Capital intensity

0.789 (2.620)

0.705 (4.793)

0.938 (1.787)

The table reports mean and standard deviation (in parentheses) of all the variables by firm type

6.3 Firm Heterogeneity and ODI Decision

171

6.3 Firm Heterogeneity and ODI Decision 6.3.1 Model Specification and Basic Results We employ a multinomial logit model to analyse how firm heterogeneity (productivity and financial constraint) affect firms’ choices of market entry: ODI (might also export), export and domestic. Our model is as follows: exp(α + β1 j pr odit + β2 j FCit + γ j Cit + Y d + I d) k=D,E,F ex p(α + β1 j pr odit + β2 j FC it + γ j C it + Y d + I d) (6.1)

Pr[yit = j] = 

Here, yit is firm i’s choice in year t. j stands for three available choices: ODI (F), export (E) and domestic (D). prod it and FC it indicate productivity and financial constraint, respectively. They are our key explanatory variables. C it includes all the other control variables listing in Table 6.2, Yd and Id are year-fixed effect and industry-fixed effect, respectively. Table 6.3 reports our estimated results. Firm productivity has a totally opposite effect on the decision to export and ODI. More productive firms choose ODI, which is consistent with the finding in Tian and Yu (2012), but less productive firms in Zhejiang Province choose to export. This is a different outcome than that in Helpman et al. (2004). There might be two reasons for the difference. First, processing trade might play a significant role here,2 which means export behaviour in Zhejiang Province is mostly related to low value added production. Second, the finding in Helpman et al. (2004) is based on bilateral trade data, but our result is based on firm-level data from the exporting country. The coefficients of the financial constraint indexes on exports and ODI are significantly positive; the marginal effects are approximately 2.73% and 0.12%, respectively. When the index is higher, a firm’s financial constraint is lower, and it has a higher possibility of entering overseas markets. Therefore, it indicates that financial constraint will inhibit firms’ exporting and firms’ ODI because additional costs are involved in entering a new market. Todo (2011) also finds a negative effect of financial constraint in Japan, but the result is not significant. It suggests that inhibition effect of financial constraint on firms’ internationalization behaviour is more serious in China. For the control variables, the tax ratio has a negative coefficient; that is, lowering taxes encourages firms to enter new markets. Foreign companies have accumulated overseas experience. As we can see from Table 6.3, they are more likely to be involved 2

Since there is no detailed information in the dataset allowing us to classify export firms into general trade and processing trade groups, we are not able to directly test this interpretation. However, some findings from other studies could be good supports of our argument. Dai et al. (2012) used custom data and found that processing trade accounts for nearly half of China’s exports. More importantly, those firms are 4% to 30% less productive than non-exporters. Yu (2015) showed low-productivity firms self-select to engage in process sing trade. These evidence are consistent with our results.

172

6 Outward Direct Investment, Firm Productivity and Credit Constraints: Evidence …

Table 6.3 Multinomial logit estimates of basic model

Variables Prod FC Tax ratio FDI dummy Firm age

(1)

(2)

Export

ODI

−0 179***

0.238***

(0.009)

(0.06)

0.144***

0.394***

(0.003)

(0.025)

−0.228***

−0.299***

(0.004)

(0.025)

1.387***

1.087***

(0.019)

(0.106)

0.022***

0.029***

(0.001)

(0.006)

−0.116***

−0.152***

(0.006)

(0.034)

Year fixed effect

Yes

Yes

Industry fixed effect

Yes

Capital intensity

Yes

Observations

132 902

Pseudo R2

0.152

Standard errors are in parentheses. *** p < 0.01 ** p < 0.05 * p < 0.1

in exporting or ODI. Firm age also has a significantly positive effect on firms’ decision to enter overseas markets. In addition, less capital-intensive firms get more involved in internationalization behaviour.

6.3.2 Interaction Effect Between Productivity and Financial Constraint When a firm is making decisions about overseas market entry, will a higher productivity level release the negative effect of financial constraint? In order to answer this question, we put the cross-term of the productivity variable and the financial constraint index into the multinomial logit model, and regress Eq. 6.2. If the release effect, β2j exists, it should be significantly negative:   Pr yit = j

  exp a + β1 j pr odit + β2 j pr od × FCit + β3 j FCit + γ j Cit + Y d + ld   = k=D,E,F ex p a + β1 j pr od jt + β2 j pr odit × FC ik + β3 j FC it + γ j C it + Y d + ld (6.2)

6.3 Firm Heterogeneity and ODI Decision

173

From Table 6.4, it is evident that there is neither a ‘release effect’ on the exporting decision nor on the ODI decision. What might violate our intuition is that productivity strengthens the impact of financial constraint on exporting, because the coefficient β2E for the export entry decision is positive and significant at the 1% level. Consistent with our previous finding, we think it might still be due to processing trade. Firms usually have less cost in processing trade because they mainly focus on low valueadded production, like assembly, and they also have advantages in taxation and tariff exemption (Yu and Tian, 2015). Hence, processing trade firms might be less restricted by financial conditions. Given that low productivity exporting firms take part in processing trade, it makes sense that financial constraint has a larger effect on firms with higher productivity. Meanwhile, the positive coefficient of the cross-term also shows us that for firms with higher financial capacity (lower financial constraint), productivity has a less negative effect or even has a positive effect on the exporting decision. Table 6.4 Multinomial logit estimates with interaction effect

Variables Prod FC Prod*FC Tax ratio FDI dummy Finn age

(1)

(2)

Export

ODI

−0 379***

0.235

(0.022)

(0.199)

−0.004

0.376***

(0.015)

(0.121)

0.038***

0.005

(0.004)

(0.028)

−0.229***

−0.299***

(0.004)

(0.025)

1.385***

1.087***

(0.019)

(0.106)

0.022***

0.029***

(0.001)

(0.006)

−0.123***

−0.156***

(0.006)

(0.035)

Year fixed effect

Yes

Yes

Industry fixed effect

Yes

Capital intensity

Yes

Observations

132 902

Pseudo R2

0.152

Standard errors are in parentheses. *** p < 0.01 ** p < 0.05 * p < 0.1

174

6 Outward Direct Investment, Firm Productivity and Credit Constraints: Evidence …

6.3.3 Robustness Check: First Time Exporting or Outward Direct Investment To deal with the potential endogeneity problem, and exclude the learning effect of existing exporting or ODI enterprises, and bias caused by fixed entry cost they have already paid (Lee, 2010), we reduce our sample to only contain first-time exporting or ODI observations and non-exporting/ODI observations during 2006–2008. We re-estimate Eq. (6.1), and obtain the result in Table 6.5. It shows an outcome much like our basic result. Therefore, for first-time exporting or ODI firms, the coefficients of the financial constraint index are significantly positive, and financial constraint has a negative impact on both firms’ export and ODI choice. Productivity negatively correlates with the exporting decision, but promotes firms to choose ODI. Table 6.5 Multinomial logit estimates of first time export/outward direct investment (ODI) sample

Variables Prod FC Tax ratio FDI dummy Firm age Capital intensity

(1)

(2)

Export

ODI

−0.103***

0.676***

(0.016)

(0.133)

0.193***

0.405***

(0.006)

(0.061)

−0.200***

−0.190***

(0.006)

(0.055)

1.279***

0.701***

(0.034)

(0.26)

0 103***

0.087***

(0.003)

(0.019)

−0.091***

−0.054

(0.008)

(0.038)

Year fixed effect

Yes

Yes

Industry fixed effect

Yes

Yes

Observations Pseudo

R2

48 492 0.158

6.5 Conclusion

175

6.4 Firm Heterogeneity and Outward Direct Investment Value 6.4.1 Model Specification Following the analysis of the extensive effect of firm heterogeneity on firms’ exporting and ODI choice, we also want to examine its intensive effect on exports and ODI value. The model specification is as follows: lnvalueit = α + β1 pr od it + β2 FCit + γ Cit + Y d + I d + εit

(6.3)

Ln lnvalueit is either firm i’s export value or ODI value in year t, measured in logarithmic form. All the explanatory variables are the same as in Eq. 6.1, and εit is the error term. The estimated coefficients for productivity and financial constraint (β1 , β2 ) are our main concern.

6.5 Conclusion Using a rich firm-level panel data set for Zhejiang Province in China, we investigate firm heterogeneity and overseas market entry decisions, and we provide strong firmlevel evidence on the inhibition of financial constraint for outward direct investment from China. We find that financial constraint not only lowers firms’ possibility of establishing a foreign affiliate, but also reduces their ODI value. Hence, a better financing environment would effectively improve both the prevalence of ODI and the total investment value of ODI. We also find that high productivity promotes firms to choose ODI, but it cannot offset or release the negative effect of financial constraint on firms’ choice of ODI. That is, even if a firm largely improves its productivity, because the financial constraint is still tight, it might not be able to invest overseas. Hence, those more competitive private firms with higher productivity in China might not have the opportunity to pursue ODI. While for China, as has been the case elsewhere, encouraging ODI is an important strategy for upgrading the economic structure, transferring excess production capacity, improving competitiveness and opening up export markets to the outside world.

176

6 Outward Direct Investment, Firm Productivity and Credit Constraints: Evidence …

References Bellone, F., Musso, P., Nesta, L., & Schiavo, S. (2010). Financial constraints and firm export behaviour. World Economy, 33(3), 347–373. Berman, N., & Héricourt, J. (2010). Financial factors and the margins of trade: Evidence from cross-country firm-level data. Journal of Development Economics, 93(2), 206–217. Bernard, A., Jensen, J. B., & Schott, P. (2003). Falling trade costs, heterogeneous firms, and industry dynamics. Nber Working Papers, 167(4), 613–626. Bottazzi, G., Secchi, A., & Tamagni, F. (2014). Financial constraints and firm dynamics. Small Business Economics, 42(1), 99–116. Cheung, Y. W., & Qian, X. (2009). Empirics of China’s outward direct investment. Pacific Economic Review, 14(3), 312–341. Contessi, S., & De Pace, P. (2012). (Non-) resiliency of foreign direct investment in the United States during the 2007–2009 financial crisis. Pacific Economic Review, 17(3), 368–390. Damijan, J. P., & Rojec, M. (2007). Foreign direct investment and catching up of new EU member states: Is there a flying geese pattern? Applied Economics Quarterly, 53, 91–118. Feenstra, R. C., Li, Z., & Yu, M. (2014). Exports and credit constraints under incomplete information: Theory and evidence from China. Review of Economics and Statistics, 96(4), 729–744. Girma, S., Kneller, R., & Pisu, M. (2005). Exports versus FDI: An empirical test. Review of World Economics, 141(2), 193–218. Greenaway, D., & Kneller, R. (2007). Firm heterogeneity, exporting and foreign direct investment. The Economic Journal, 117(517), 134–161. Hadlock, C. J., & Pierce, J. R. (2010). New evidence on measuring financial constraints: Moving beyond the KZ index. The Review of Financial Studies, 23(5), 1909–1940. Head, K., & Ries, J. (2003). Heterogeneity and the FDI versus export decision of Japanese manufacturers. Journal of the Japanese and International Economies, 17(4), 448–467. Helpman, E., Melitz, M. J., & Yeaple, S. R. (2004). Export versus FDI with heterogeneous firms. American Economic Review, 94(1), 300–316. Krugman, P. (1980). Scale economies, product differentiation, and the pattern of trade. The American Economic Review, 70(5), 950–959. Lamont, O., Polk, C., & Saaá-Requejo, J. (2001). Financial constraints and stock returns. The Review of Financial Studies, 14(2), 529–554. Lee, H. (2010). Decision to invest abroad: The case of South Korean multinationals. Pacific Economic Review, 15(2), 281–302. Melitz, M. J. (2003). The impact of trade on intra-industry reallocations and aggregate industry productivity. Econometrica, 71(6), 1695–1725. Minetti, R., & Zhu, S. C. (2011). Credit constraints and firm export: Microeconomic evidence from Italy. Journal of International Economics, 83(2), 109–125. Musso, P., & Schiavo, S. (2008). The impact of financial constraints on firm survival and growth. Journal of Evolutionary Economics, 18(2), 135–149. Myers, S. C. (1984). The Capital Structure Puzzle. The Journal of Finance, 39, 574–592. Sun, J., & Yamori, N. (2009). Regional disparities and investment–cash flow sensitivity: Evidence from Chinese listed firms. Pacific Economic Review, 14(5), 657–667. Tian, W., & Yu, M. (2012). Outward foreign direct investment and productivity: Firm-level evidence from China. China Economic Quarterly, 11, 383–408. Todo, Y. (2011). Quantitative evaluation of the determinants of export and FDI: Firm-level evidence from Japan. The World Economy, 34(3), 355–381. Tomiura, E. (2007). Foreign outsourcing, exporting, and FDI: A productivity comparison at the firm level. Journal of International Economics, 72(1), 113–127. Wang, X., & Huang, Y. (2012). Outward direct investment: A financial repression perspective. In International Workshop on Chinese Outward Direct Investment, Peking University, Beijing, China.

References

177

Whited, T. M., & Wu, G. (2006). Financial constraints risk. The Review of Financial Studies, 19(2), 531–559. Yu, M. (2015). Processing trade, tariff reductions and firm productivity: Evidence from Chinese firms. The Economic Journal, 125(585), 943–988.

Chapter 7

The Potential Impact of China–US BIT on China’s Manufacturing Sectors

Abstract This chapter finds that the overall effect of the foreign direct investment (FDI) and there by the China–US bilateral investment treaties (BIT) on Chinese manufacturing sector is positive, which raises the productivity and profitability of the firms, using various econometricmodels and other evidence. The manufacturing sector as a whole has already opened up to the world economy and needs to continue this process. The industries in the manufacturing sector do not need to be protected, except for in limited fields related to national security, scarce natural resources and well-defined strategic sectors. Gradual lifting of the protection maybe needed in the short-run for a small number of vulnerable sectors. A moderate relaxing of the current restrictions will increase FDI in manufacturing from all countries by 4–8% under different assumptions. This effect will be small when only considering FDI from the USA. Domestic firms need to update their technology, reduce costs and learn management skills from their foreign competitors, while using the national treatment terms in BIT to enter the fields that are not open to domestic firms under current regulations. Domestic firms also need to set up firm-level global strategies and reallocate firms’ resources according to the changing investment environment, taking advantage of profit opportunities outside the domestic markets.

China has been the world’s leading manufacturer of steel, garments, cement, chemical fertilizers and many other products in the past 30 years. At the same time, China has become a preferred destination for the relocation of global manufacturing facilities and manufacturing has been the most important field of foreign direct investment (FDI) in China. The cumulated FDI in all sectors, actually used, reached $117 0.6 billion in 2013 and $119 0.6 billion in 2014. However, FDI in the manufacturing sector has grown much more slowly than that in other sectors in recent years (Fig. 7.1). The USA has been an important investor in China’s manufacturing industry in the past 20 years. Table 7.1 gives data of FDI from the USA to China in the past 14 years, which did not change much since 2005. The shares of US FDI decreased dramatically from 10.8% in 2000 to 2.0% in 2014. This chapter is published in China Economic Journal by Yu Miaojie and Fan Zhang. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022 W. Tian and M. Yu, Outward Foreign Direct Investment of Chinese Enterprises, Contributions to Economics, https://doi.org/10.1007/978-981-19-4719-3_7

179

180

7 The Potential Impact of China–US BIT on China’s …

Fig. 7.1 FDI in the manufacturing sector, actually used, China, 2000–2014, $ million. Note No manufacturing data in 2004 and 2014. Source National Bureau of Statistics of China

Table 7.1 US direct investment in China, actually used, in all sectors, 2000–2014, $ million

Total

US

US/total (%)

2000

40,715

4384

10.8

2001

46,878

4433

9.5

2002

52,743

5424

10.3

2003

53,505

4199

7.8

2004

60,630

3941

6.5

2005

60,325

3061

5.1

2006

63,021

2865

4.5

2007

74,768

2616

3.5

2008

92,395

2944

3.2

2009

90,033

2555

2.8

2010

105,732

3017

2.9

2011

116,010

2369

2.0

2012

111,716

2598

2.3

2013

117,586

2820

2.4

2014

19,562

2371

2.0

Source National Bureau of Statistics of China

The China–US bilateral investment treaties (BIT) are focused on market access of foreign investment, which requires a raft of China’s domestic reforms. Since manufacturing is more open to FDI than that in the service sector, BIT will have less overall impact on China’s manufacturing sector.

7.1 Literature Review

181

Based on the assumed scenarios of China–US bilateral investment treaties, this chapter will explore the possible impacts of BIT on China’s manufacturing sector, and then give suggestions on the Chinese Government’s strategies in BIT negotiations and counter- measures in practice, as well as the strategy domestic firms could adopt to deal with the competition that comes with BIT.

7.1 Literature Review A large body of literature of international economics focuses on the fundamental factors that drive FDI behavior. Melitz (2003) developed a theoretical model of monopolistic competition with heterogeneous firms to explain the decision of FDI, which becomes the cornerstone of the literature examining the role of heterogeneity in FDI. Helpman et al. (2004) generalized Melitz (2003) to explain horizontal FDI. Melitz’s (2003) framework has been used in explaining different issues, including trade liberalization, technology adoption (Bustos, 2005), complex integration strategies (Yeaple, 2003), variable markups (Ottaviano et al., 2002) and so on. The empirical literature on FDI examines the internal and external factors that determine the FDI by multinational enterprises. Firm characteristics are used to explain FDI activities (Morck & Yeung, 1992). External factors are also examined by different researchers, including the effects of exchange rates (Blonigen, 1997; Campa, 1993; Froot & Stein, 1991; Lipsey, 2001), taxes (Hartman, 1984, 1985; Swenson, 1994; De Mooji and Ederveen 2003), institutions (Wei, 2000; Wei & Shleifer, 2000), trade protection (Grubert & Mutti, 1991; Kogut & Chang, 1996) and so on. Many empirical researches are about the effects of FDI on the performants of Chinese firms, including researches on the effects on different industries (Cheung, 2010; Zhang & Zheng, 1999), different source countries (Shiau et al., 2013), location choices (Sharma et al., 2014), inward FDI (Liu et al., 2014) and outward FDI (Buckley et al., 2007; Deng, 2013). Based on data of 220 cities in China from 2003 to 2009, using threshold panel regression estimation, Li and Liu (2012) find significant threshold effects of FDI on China’s environment. Their tests from income threshold show that, in the middle-income stage, FDI generates degradation of local environment quality. Although there are a large number of researches on FDI in general, the literature specifically about BIT is relatively limited. Cui introduces the content of the 2012 version of the US BIT template and gives suggestions on China’s strategy in US– China BIT negotiation. Huang and Zhou (2013) discuss China’s new strategy of open-up to the world, including making progress in China–US BIT negotiations. Liang and Yan (2013) point out the core issues of the China–US BIT negotiations and difficulties of the negotiations due to the institutional factors. Pan and Tang (2013) introduce the national security investigation mechanism of the US Government. Wang (2013) examines the new version of the US BIT template and discusses the

182

7 The Potential Impact of China–US BIT on China’s …

difference between the two sides. Yao (2013) discusses the risks of China–US BIT for China in the long run, including the internationalization of the domestic policy and the liberalization of the capital account. On the US side, Bergsten (2005) argues that, to respond to the challenges and threats in the world, the USA should alter its priorities in the Doha Round to an important degree by placing even greater emphasis on reduction of agricultural subsidies and extensive liberalization of service markets. The Peterson Institute for International Economics and the China Development Research Foundation (2015) has a discussion on the US–China BIT, including the US national security investigation process, implications of the China–Japan–Korea investment agreement for US policy (Schott & Cimino, 2015), China’s state-owned enterprises and competition policy (Miner & Hufbauer, 2015), the USA’s service export (Jensen, 2015) and China’s manufacturing industry (Zhang & Yu, 2015).

7.2 Assumptions on the Scenarios of China–US BIT, Focusing on Manufacturing BIT aims to extend the principle of national treatment, under which foreign firms are treated the same as domestic firms. In China’s World Trade Organization (WTO accession agreement, that goal was met with a list of industries in which national treatment would apply. The great progress in BIT is to switch to a “negative list” approach, which allows only the businesses specified in the treaty to be exempted. Our baseline assumption assumes that BIT will be a moderate revision of the current catalogue of industries for guiding foreign investment prepared by the Chinese Government. The current list of restricted and prohibited industries in this catalogue is shown in Appendix Table 7.7. The current list of restricted and prohibited industries in manufacturing is the result of considering the following factors by the Chinese Government: (1) national security and political reasons (e.g. arms production); (2) scarce resources (e.g. rare earth metal smelting); (3) natural monopoly (e.g. gas and water production); (4) control of low-scale firms’ entry or over-capacity in some industries; and (5) toxic, harmful and environmental pollution. Some of the protection measures reflect the interests and pressure of domestic firms. In this chapter, our baseline scenario for forecasting includes: (1) national treatment is given to all foreign investors in all sectors except sectors in the negative list; (2) moderate reductions in the products or sectors in the negative list and moderate improvements in investment revenue transfer and implementation terms; and (3) moderate improvement in the dispute settlement clauses.

7.3 BIT’s Open Market Requirements to China’s …

183

7.3 BIT’s Open Market Requirements to China’s Manufacturing Sector and Its Impacts on Relevant Industries The impacts of BIT on China’s manufacturing sector are quite similar to that of China’s joining the WTO in 2001. Although some of the industries face negative impacts, China’s experience of joining the WTO tells us that overall the opening up to the outside world will induce higher growth and productivity in the manufacturing sector, even for those industries that were predicted to become losers before joining the WTO, for example the automobile manufacturing industries. In this chapter, we will analyze the impacts of BIT on Chinese domestic firms and the impacts of policy changes on FDI, using firm-level data and econometrics models.

7.3.1 Impacts of FDI on Domestic Firms We first investigate the impacts of FDI on domestic firms’ productivity, profitability and export propensity. FDI has two major impacts on firms in the domestic market. The positive impacts are that FDI will bring technology, management skills and capital to China, which will increase Chinese firms’ productivity and profitability. The negative impacts are that FDI will intensify competition and may crowd out some Chinese firms in the field. This research will do econometrics analysis to find out the effects of FDI on firms’ performants in domestic markets, using a large firm-level dataset created by China’s National Bureau of Statistics (NBS) 2000–2008. The main dataset includes over 2 million firm-level observations from 2000 to 2008. The dataset is collected by the National Bureau of Statistics of China. It includes all state-owned firms and non-stateowned firms with sales over RMB 5 million per year. The dataset provides 80 to 150 firm-level financial indicators, for example, output, sales, fixed investments, number of workers, exports and so on. We clean the dataset following Feenstra et al. (2014) by eliminating observations in which: (1) firms have less than eight employees; (2) fixed assets exceed total assets; (3) current assets exceed total assets; (4) there is no identification number; and (5) no starting time.

7.3.2 FDI’s Overall Impacts on Performance of Domestic Firms First, we find that, overall, FDI has a positive impact on firms’ productivity and profit. We first used a regression to estimate the effects of FDI on firms’ productivity, profitability and export propensity, using the full sample of China’s firm-level data in

184

7 The Potential Impact of China–US BIT on China’s …

manufacturing from 2000 to 2007. Specifically, in the model the dependent variable is ln (TFP), profit/sales and export/sales, where TFP is total factor productivity. The explanatory variables are foreign invested enterprise (FIE, dummy showing whether the firm is foreign invested), and share of the sales of FIE in four-digit industries. The estimated coefficients show how much, on average, FIE performance is higher than the non-FIE. Fixed effects of year and firm are controlled. The results of estimation are presented in Table 7.2. The coefficients of FIE and FIE share are all positive and significant as more variables are controlled. We also estimate the regressions with TFP estimated by other methods, which support the results presented here. Next the performance of FIEs from the US investments is estimated. We calculated the share of FDI from the USA in total FDI in China in 2013 (a), and the share of FDI from the USA in total FDI in China in the sampling period (b). Assuming after signing BIT, the share of FDI from the USA increases by t%, then the effects of the US FDI on manufacturing firms’ TFP (EUS ) will be where α is the coefficient of FIE estimated in our previous regression in Table 7.2. Using Eq. (7.1), we estimated EUS under different assumptions as shown in Table 7.3. a EUS = t α b

(7.1)

Since the share of FDI from the USA has decreased dramatically in recent years, the effect of BIT on firms’ productivity is lower than the average of all foreign investors. We also estimated the effects of FDI on firms’ profitability and export propensity. The results are presented in Table 7.4. The results in Table 7.4 show that FIEs have a positive effect on firms’ profit rate, when export, SOE, labor and asset are controlled. On the other hand, the relation between export propensity and FIE is not significant. To check the robustness of the relation between productivity and FIE, we also estimated models using alternative methods (Levinsohn-Petrin Approach; Levinsohn & Petrin, 2003) to calculate TFP, which received similar results. The results show that, overall, FDI improves the performance of Chinese manufacturing firms.

7.3.3 FDI’s Impacts on Specific Industries Figure 7.2 and Table 7.5 show the basic conditions of selected manufacturing industries, from which readers can see the size of the industry, the FIE and non-FIE production, and the share of foreign capital. The last column in Table 7.5 gives us the number of current restricted and forbidden industries/products in the catalogue for the guidance of FDI. Figure 7.2 shows that FIE production accounts for a large percentage in computer and communication equipment (industry code 39) and automobile (36).

2,096,406

2,096,406

0.01

Ob

R-squared

N

0.07

1,661,369

N

Y

0.05

1,661,369

Y

0.07

1,272,206

N

N

0.023*** (15.09)

0.109*** (233.48)

0.04

1,272,206

Y

Y

0.010*** (7.65)

0.016*** (14.41)

0.016*** (12.93)

−0.018*** (−5.32)

0.025*** (14.49)

0.012*** (3.12)

(6)

0.05

1,661,369

Y

Y

0.002*** (32.32)

0.017*** (18.55)

0.015*** (14.23)

−0.022*** (−7.60)

0.024*** (16.24)

0.007** (2.37)

(7)

0.04

1,272,206

Y

Y

0.002*** (25.12)

0.010*** (7.51)

0.016*** (14.38)

0.016*** (12.80)

−0.018*** (−5.18)

0.025*** (14.54)

0.011*** (2.88)

(8)

Notes FIE in this regression does not include investment from Hong Kong, Taiwan, and Macao. TFP calculated using OP approach, robust t statistics are reported in parentheses. Columns (5), (6) and (8) use sample 2001–2007 (since no data on research and development expenditure in 2000, 2004 and 2008, we use the average of 2003 and 2005 for the research expenditure in 2004). FIE is dummy indicates whether a firm is foreign invested, FIE share is the share of FIE in total sales for 4-digit industries, Export shows whether a firm engaged in export, SOE indicates whether a firm is state owned, labor is total number of employee in a firm, asset is total value of asset in a firm. ***, **, and * shows significance at the 1, 5, and 10% level, respectively. FE—fixed effect

0.06

Y

N

N

E year

FE firm

FIE share

Research

0.113*** (283.67)

Log(Asset) 0.017*** (18.56)

0.015*** (14.39)

−0.049*** (−93.66)

Log(Labor)

−0.046*** (−76.02)

0.060*** (37.75)

−0.023*** (−7.81)

0.029*** (20.81)

0.101*** (60.03)

(5)

0.060*** (51.47)

0.008*** (2.67)

(4)

0.024*** (16.24)

Y

0.096*** (64.67)

SOE

0.013*** (4.42)

(3)

0.054*** (52.55)

0.191*** (148.21)

FIE

(2)

Export

(1)

Dependent var: ln TFP

Table 7.2 Effects of FDI on performance of firms, productivity, China

7.3 BIT’s Open Market Requirements to China’s … 185

186 Table 7.3 Estimates of the effects of BIT on firms’ productivity

7 The Potential Impact of China–US BIT on China’s … Assumed increase in US FDI after

BIT assignment

Coefficient from Table 7.2

EUS (%) 1.5 times

0.013

0.7

2 times

0.013

0.9

Notes The share of US FDI in China reduced from an average of 6.8–2.4% from 2000–2007 to 2013. The coefficient from Table 7.2 is the sum of coefficients of FIE and FIE share in last column of Table 7.2

Econometrics analysis shows that the direct impact of FDI on firms’ productivity is positive for almost all industries. However, the indirect impacts of FDI (the effect of the share of FDI in an industry on firms in that industry) on some industries are negative. (Lack of space forbids further discussion on these regressions for each industry.) Most of the industries in the manufacturing sector are already opened up to the world market and do not need protection in BIT. However, some of the industries are vulnerable when facing the competition of FIEs, including: (1) China’s comparative advantage is changing as labor costs are rising in China. Therefore, there may be less FDI over time in labor-intensive industries, as global firms shift to lower-wage economies elsewhere. This is not the result of the signing of a BIT. For example, the textile industry is one of these laborintensive industries. As China’s comparative advantage changes, labor costs increase and these labor-intensive industries have lost their advantages gradually as the economy transfers to capital- and technology-intensive sectors. FDI helped this structural change. The labor-intensive firms are vulnerable to FDI competition. What the government needs to do is to help these firms upgrade their products, but not to set protection measures in BIT. (2) The second group of industries or sub-industries is vulnerable to the competition of the foreign investments due to the large technology gaps. These industries or sub-industries are in capital- and technology-intensive sectors but have large technology gaps compared with the world technology frontier. These firms may need some temporary protection measures in BIT in the short-run. (3) The third group of industries will face environmental problems caused by FDI and need some regulations to control the size of damages, for example the chemistry industry. BIT or domestic regulations are needed to prohibit FDI in the chemistry industry which causes serious environmental pollution.

7.3.4 The Effects of Changes in Policies on Scale or Shares of FDI To explore the impacts of protection policies on FDI in China, we did a regression in which the dependent variable is the level of foreign capital in the industry and the

16.120 (1.50)

Log(Asset)

1,684,364

0.001

2,251,355

0.001

Ob

R-squared

0.001

1,684,364

Y

Y

0.191 (1.02)

16.118 (1.50)

0.00

1,278,782

Y

Y

12.783* (1.74)

0.099 (0.41)

0.00

2,195,895

Y

Y

0.00

1,654,682

Y

0.00

1,654,682

Y

Y

0.012* (1.86)

−0.073 (−0.51)

−0.072 (−0.51)

8.485* (1.84)

0.072 (0.57)

0.073 (0.57)

−0.078 (−0.26)

38.517*** (37.63)

−0.599 (−1.57)

−3.104 (−0.12)

38.517*** (37.63)

−0.594 (−1.55)

−0.080 (−0.26)

Y

0.078 (0.22)

−9.655*** (−2.78)

41.652 (1.11)

4.185*** (2.59)

Export propensity

0.00

1,278,782

Y

Y

−0.635** (−2.24)

0.012* (1.88)

−0.063 (−0.45)

0.083 (0.63)

−0.079 (−0.26)

38.529*** (37.46)

−0.600 (−1.57)

Notes FIE in this regression not include investment from Hong Kong, Taiwan, and Macao. In this table, the division of industry is according to 2-digit industrial code in “Classification of national economic industries” (GB/T 4754—2002, the 2nd edition). The division of industry before 2003 is according to “Classification of national economic industries and code” (GB/T4754) and is transferred to the second edition. TFP is calculated using augmented OP approach, t statistics are reported in parentheses. ***, **, and * shows significant at the 1, 5, and 10% level, respectively. Sample period used in column (4) and (8) is 2001–2007 (since no data on research expenditure in 2000, 2004 and 2008, we use the average of 2003 and 2005 for the research expenditure in 2004). Sales profit rate = operating profit/total industry output * 100. See notes under Table 7.2 for meanings of variables

Y

Y

Y

Y

FE year

FE firm

Research

FIE share

11.041 (0.64)

Log(Labor)

11.026 (0.64)

−17.181*** (−4.09)

−17.229*** (−4.10)

SOE

2.303** (2.29) 40.912 (1.42)

2.375** (2.38)

40.916 (1.42)

26.644 (1.24)

FIE

Export

Sales profit rate

Dependent var

Table 7.4 Effects of FDI on performance of firms, profitability and export propensity, China

7.3 BIT’s Open Market Requirements to China’s … 187

188

7 The Potential Impact of China–US BIT on China’s …

Fig. 7.2 Size of production, manufacturing, 2007. Source Firm dataset by NSB. See Table 7.5 for names of the industries

explanatory variables are total sales in the industry and the dummies for protection policies. We use a four-digit industry-level dataset collected from the manufacturing firm database. In our model, the dependent variable is foreign capital. Explanatory variables are the policy variable Forbidden (dummy, restricted and forbidden products/industries in the catalogue for the guidance of FDI), total sales in the industry and the concentration level of the industry or Herfindahl–Hirschman Index (HHI). We transfer the products in the catalogue into four-digit industries. The estimated coefficients of the variable Forbidden shows how much more foreign capital will be created if China eliminates one restriction item. The last fixed effect regression in Table 7.6 shows that if the FDI in a four-digit industry is forbidden or restricted, the total foreign capital will be reduced by RMB 1176 million. Using the coefficients estimated in our regression, we calculated the scale of extra FDI from the USA into China when BIT reduced the protection under different scenarios. The regression results show that if 10 out of 51 restricted and forbidden items are lifted (scenario I), total FDI from all countries in manufacturing will increase 4.2%. Since the USA accounts for only 2.4% of total FDI, the effect of China–US BIT on FDI from the USA will be trivial. These estimate effects may increase if considering other possible contents in BIT, for example, tax law changes and improvement in the dispute settlement clauses. To sum up, the basic logic behind the models above is that policy determines the inflow of FDI (Table 7.6), and FDI improves the performance of Chinese firms (Table 7.2).

7.3 BIT’s Open Market Requirements to China’s …

189

Table 7.5 Shares of foreign capital and restrictions, manufacturing, 2007 Industrial code

Name of industry Share of foreign capital (%)

FIEs’ share, in numbers of firms (%)

Number of restrictions and forbidden items, 2011

13

Processing of food from agricultural products

29.0

12.7

2

14

Manufacture of foods

35.4

19.5

1

15

Manufacture of beverage

34.1

13.9

3

16

Manufacture of tobacco

0.4

4.8

1

17

Manufacture of textile

24.4

15.5

3

18

Manufacture of textile wearing apparel

43.0

35.8

0

19

Manufacture of leather

47.8

30.7

0

20

Processing of timber

19.7

12.8

0

21

Furniture

42.5

27.1

0

22

Paper

32.0

14.9

0

23

Printing

30.1

13.4

1

24

Products for culture and education

45.3

34.5

4

25

Processing of petroleum

14.1

8.7

2

26

Chemical materials and products

28.7

14.5

10

27

Medicine

27.0

15.9

7

28

Chemical fibers

42.3

19.6

2

29

Rubber and plastics

37.2

22.1

1

30

Non-metallic 18.3 mineral products

10.7

0

31

Ferrous metals

20.5

7.4

4

32

Non-ferrous metals

18.0

10.8

4 (continued)

190

7 The Potential Impact of China–US BIT on China’s …

Table 7.5 (continued) Industrial code

Name of industry Share of foreign capital (%)

FIEs’ share, in numbers of firms (%)

Number of restrictions and forbidden items, 2011

33

Metal products

31.1

17.2

2

34

General purpose machinery

31.5

15.7

2

35

Special purpose machinery

25.6

19.0

1

36

Automobile

51.4

20.6

0

37

Transportation equipment

27.4

15.0

0

38

Electrical machinery

36.8

21.7

0

39

Computer and communication equipment

79.2

45.6

1

40

Instruments

43.8

27.0

0

41

Other manufacturing

41.6

29.3

0

42

Recycling of waste

21.1

17.7

0

Ave

32.6

19.1

Total

1.7 51

Table 7.6 Regression results: effects of policy variables on FDI Dependent var: foreign capital

OLS

Total sales

0.0346**

Forbidden

−962,826**

FE 0.0334**

0.0300**

−1,069,518**

−1,164,730**

−1,176,522** −739,410

2,025,509**

HHI

0.0300**

FE two-digit industries

Y

Y

FE year

Y

Y

Obs

2747

2747

2747

2747

R2

0.4265

0.4424

0.4327

0.4333

Notes *, ** significant at 10% and 5% level. Sample: 2002–2007. Total sales—total sales of the four-digit industry. Foreign capital—total amount of foreign capital in the four-digit industry. Forbidden—dummy which is 1 if the four-digit industry is restricted or forbidden from FDI

7.4 Suggestions on Negotiation Strategy

191

7.4 Suggestions on Negotiation Strategy The negative list in BIT can be based on current the Foreign Investment Industry Guidance Catalogue established by the Chinese Government, with moderate revisions. Base on the above research, we have some suggestions for China on its negotiation strategies of BIT with the United States.

7.4.1 Make It Firm and Steadfast That China Is Serious in Joining BIT In the long-run, BIT is beneficial for Chinese firms to improve their productivity and profitability. Therefore, Chinese negotiators must make it firm and steadfast that China is serious in joining BIT with the USA through negotiations.

7.4.2 Protection Measures in the Long-Run The Chinese side also needs to negotiate to keep some of the protection measures in the treaty in the long-run for industries with a natural monopoly and for national security reasons. However, the number of these protected industries should be limited. Protection measures should be in limited fields, including: (1) fields related to national security, for example, production of military weapons; (2) fields related to scarce natural resources, for example, processing of rare metals.

7.4.3 Gradual Lifting Process of Protection for Certain Vulnerable Sectors In the short-run, however, joining BIT will hurt some of the Chinese firms or industries, even though it is beneficial for Chinese firms to improve their productivity and profitability in general. Studies on industries show that firms in some of the manufacturing industries, especially those with large gaps in technology, will be harmed in the short-run. Therefore, there may need to be a gradual lifting of the protection in a small number of industries.

192

7 The Potential Impact of China–US BIT on China’s …

7.4.4 Cooperating in BIT Negotiation with the Domestic Reform BIT and domestic regulations have different functions, and needed to be treated separately. Some of the restrictions on FIEs’ activities can be done by domestic regulations, when the FIEs are given national treatment. Therefore, items already restricted by domestic regulation can be removed from the negative list of BIT, for example, capacity requirements for petroleum refinery equipment. The Chinese Government will use the requirement of BIT to reform domestic administration, the judicial system and the state-owned enterprises. To do that, the domestic law and regulation need to cooperate with the foreign economic policies.

7.5 Suggestions to Manufacturing Firms Regarding How to Face the Challenges of BIT The numerous entries of multinationals have threatened local Chinese companies’ existing market positions. China’s joining BIT may make domestic firms face more intensive competition from FIE in certain fields.

7.5.1 Suggestions for Domestic Firms (1) Domestic firms need to make contingency plans to meet the challenges of BIT. They need to update their technology, reduce costs and learn management skills from their foreign competitors. (2) Domestic firms need to do research on BIT and gain benefit from it, for example, using the national treatment terms in BIT to enter the fields that are not open to domestic firms under current domestic regulations. Domestic firms also need to learn how to use legal means, including the dispute settlement clauses in BIT, to protect their interests.

7.5.2 Suggestions for Government The government should provide more detailed information to firms about the changes made by BIT and provide financial support to assist firms to make their structural adjustments. In the fields of gradual lifting of protection, the government should make it clear that the protection will be gradually lifted and the firms need to prepare for the competition from FIE in the near future.

References

193

7.6 Conclusions Using various econometric models and a large firm-level dataset, we find the overall effect of FDI and thereby BIT on Chinese manufacturing sector is positive for firms in the domestic market. As evidence shows, FDI raised the productivity and profitability of the firms significantly in the manufacturing sector. A moderate relaxing of the current restrictions on FDI will increase FDI in manufacturing from all countries by about 4%. This effect will be smaller when only considering FDI from the USA in manufacturing. China’s manufacturing sector as a whole has already opened up to the world economy and that process needs to be continued. The industries in the manufacturing sector do not need to be protected, except for fields related to national security, scarce natural resources and well-defined strategic sectors. Gradual lifting of protection may be needed in the short-run for a small number of vulnerable manufacturing industries. Domestic firms need to update their technology, reduce costs and learn management skills from their foreign competitors. They need to learn to gain benefits from BIT, using the national treatment terms in BIT to enter the fields that are not open to domestic firms under current domestic regulations.

Appendix See Table 7.7.

References Bergsten, C. F. (2005). A new foreign economic policy for the United States. The United States and the World Economy: Foreign Economic Policy for the Next Decade, 3–61. Blonigen, B. A. (1997). Firm-specific assets and the link between exchange rates and foreign direct investment. The American Economic Review, 447–465. Buckley, P. J., Clegg, L. J., Cross, A. R., Liu, X., Voss, H., & Zheng, P. (2007). The determinants of Chinese outward foreign direct investment. Journal of International Business Studies, 38(4), 499–518. Bustos, P. (2005). Rising wage inequality in the Argentinean manufacturing sector: The impact of trade and foreign investment on technology and skill upgrading. Paper, Harvard University. Campa, J. M. (1993). Entry by foreign firms in the United States under exchange rate uncertainty. The Review of Economics and Statistics, 614–622. Cheung, K. Y. (2010). Spillover effects of FDI via exports on innovation performance of China’s high-technology industries. Journal of Contemporary China, 19(65), 541–557. De Mooij, R. A., & Ederveen, S. (2003). Taxation and foreign direct investment: A synthesis of empirical research. International Tax and Public Finance, 10(6), 673–693. Deng, P. (2013). Chinese outward direct investment research: Theoretical integration and recommendations. Management and Organization Review, 9(3), 513–539.

194

7 The Potential Impact of China–US BIT on China’s …

Table 7.7 Current list of restricted and prohibited industries, manufacturing, 2011 Fields

Farm and sideline food processing

Alcohol, drinks and refined tea manufacturing

Regulations

Restricted

Prohibited

Equity controlled by Chinese party

Processing of edible oils

x

x

Production of biological liquid fuels (ethanol fuel, biodiesel)

x

x

Production of yellow rice wine (“huangjiu”), famous and high-quality Chinese spirits

x

x

Processing and production of green tea with Chinese traditional handicraft, processing and production of special tea (including white tea, yellow tea, oolong tea, dark tea, pressed tea etc.)

x

Tobacco products manufacturing

Processing and production of leaf tobacco (i.e. threshing and redrying)

x

Printing and reproduction of recorded media

Printing of publications, equity controlled by Chinese party with minimum registered capital of RMB 10 million

x

x

Oil processing, Atmospheric and vacuum x coking and nuclear distillation (≤10 million fuel processing tons/year), catalytic cracking (≤1.5 million tons/year), Continuous catalytic reforming (containing aromatic extraction, ≤1 million tons), hydrogen cracking production (≤1.5 million tons) Chemical raw material and chemical products manufacturing

Production of calcined soda, x caustic soda, and sulphuric acid, nitric acid, and potassium carbonate with limited capacity and backward technology Production of photosensitive x materials (continued)

References

195

Table 7.7 (continued) Fields

Fields

Regulations

Restricted

Benzidine production

x

Production of precursor chemicals

x

Regulations

Restricted

Prohibited

Equity controlled by Chinese party

Prohibited

Equity controlled by Chinese party

Hydrogen fluoride and other x low-end chlorofluorocarbon or chlorofluoro-compounds

Pharmaceutical manufacturing

Production of butadiene rubber (excluding high cis-butadiene rubber), emulsion polymerized styrene butadiene rubber and thermoplastic styrene–butadiene–styrene rubber

x

Acetylene PVC and below scale ethylene and processing production

x

Pigment and paint production using backward technology, containing harmful substances, and below-scale

x

Boron magnesium ore processing

x

Inorganic salt production with high resource usage, serous environmental pollution, and backward process

x

Production of x Chloramphenicol, penicillin G, lincomycin, gentamicin, dihydrostreptomycin, amikacin, tetracycline, oxytetracycline, midecamycin, leucomycin, ciprofloxacin, norfloxacin, ofloxacin (continued)

196

7 The Potential Impact of China–US BIT on China’s …

Table 7.7 (continued) Fields

Regulations

Restricted

Prohibited

Equity controlled by Chinese party

Production of analgin, x paracetamol, vitamin B1, vitamin B2, Vitamin C, vitamin E, multi-vitamin preparation and oral calcium agent Vaccines which fall inside the scope of the national immunization plan

x

Production of narcotic drugs x and active pharmaceutical ingredients for first class of psychotropic drugs

Fields

Chemical fiber

Production of blood products

x

Regulations

Restricted

x

Prohibited

Processing of materials for traditional Chinese medicines as listed in the “Regulations on Protection of Wild Medical Resources” and the “Catalogue of Rare and Endangered Chinese Plants”

x

Processing of traditional Chinese medicines (through steaming, frying, moxibustion, calcination, etc.); and production of traditional Chinese medicine patent drugs with secret formulas

x

The conventional slice spinning chemical fiber spinning production

x

Viscose fiber production

x

Equity controlled by Chinese party

(continued)

References

197

Table 7.7 (continued) Fields

Regulations

Restricted

Non-ferrous metal smelting and rolling processing

Tungsten, molybdenum, tin (tin compounds except), antimony (including antimony oxide and antimony black) and other rare metals smelting

x

Smelting and processing of radioactive minerals

Equity controlled by Chinese party

x

Electrolytic aluminum, copper, lead, zinc and other non-ferrous metal smelting

x

Rare earth metal smelting and separation

x

General equipment All kinds of ordinary level manufacturing (P0) bearing and parts (ball bearing, retainer), blank manufacturing

Prohibited

x

x

Production of wheeled or crawler cranes (400 tons below)

x

Specialty equipment manufacturing

General polyester filament, short fiber equipment manufacturing

x

Fields

Regulations

Restricted

x

Prohibited

Equity controlled by Chinese party (continued)

Feenstra, R. C., Li, Z., & Yu, M. (2014). Exports and credit constraints under incomplete information: Theory and evidence from China. Review of Economics and Statistics, 96(4), 729–744. Froot, K. A., & Stein, J. C. (1991). Exchange rates and foreign direct investment: An imperfect capital markets approach. The Quarterly Journal of Economics, 106(4), 1191–1217. Grubert, H., & Mutti, J. (1991). Taxes, tariffs and transfer pricing in multinational corporate decision making. The Review of economics and Statistics, 285–293. Hartman, D. G. (1984). Tax policy and foreign direct investment in the United States. National Tax Journal, 37(4), 475–487. Hartman, D. G. (1985). Tax policy and foreign direct investment. Journal of Public Economics, 26(1), 107–121. Helpman, E., Melitz, M. J., & Yeaple, S. R. (2004). Export versus FDI with heterogeneous firms. American Economic Review, 94(1), 300–316. Huang, H., & Zhou, C. (2013). China’s strategy under new situation in opening-up. International Economic Review, 4. Jensen, J. B. (2015). Role of a bilateral investment treaty in increasing trade in services between China and the United States. Toward a US-China Investment Treaty, 24.

198

7 The Potential Impact of China–US BIT on China’s …

Table 7.7 (continued) Fields

Regulations

Restricted

Prohibited

Equity controlled by Chinese party

Manufacturing of bulldozers x (≤320 horsepower), hydraulic excavator (≤30 tons), wheeled loaders (≤6 tons), graders, rollers, folk lifts (≤220 horsepower), electric drive off-highway self-dumping trucks (≤135 tons), hydraulic mechanical transmission and off-highway self-dumping trucks (≤60 tons), asphalt concrete mixing and paving equipment and aerial work machinery, garden machinery and equipment, and concrete machinery (pump, mixer vehicle, mixing station, pump vehicle) Arms and ammunition manufacturing

x

Electrical machinery and materials manufacturing

Manufacturing of vented lead-acid batteries (i.e. direct acid mist discharge), silver oxide batteries with mercury button, alkaline zinc-manganese batteries with mercury button, paste zinc-manganese battery and Cd-Ni battery

x

Transportation equipment manufacturing

Repairing, design and production of ships

x

Communications equipment, computers, and other electronics manufacturing

Manufacturing of ground satellite TV broadcasting receiving facility and key parts

x

Arts and crafts and Ivory carving other Tiger bone processing manufacturing Production of bodiless lacquer ware Production of enamelwork

x

x x x x (continued)

References

199

Table 7.7 (continued) Fields

Regulations

Restricted

Prohibited

Equity controlled by Chinese party

Fields

Regulations

Restricted

Prohibited

Equity controlled by Chinese party

Production of Chinese art paper and ink ingot

x

Production of carcinogenic, teratogenic, mutagenic products and persistent organic pollutants

x

Source Ministry of Commerce

Kogut, B., & Chang, S. J. (1996). Platform investments and volatile exchange rates: Direct investment in the US by Japanese electronic companies. The Review of Economics and Statistics, 221–231. Levinsohn, J., & Petrin, A. (2003). Estimating production functions using inputs to control for unobservables. The Review of Economic Studies, 70(2), 317–341. Li, Z. H., & Liu, H. H. (2012). Are there threshold effects of FDI on environment?: Evidence from 220 cities in China. Finance & Trade Economics, 8, 101–108. Liang, Y., & Yan, D. (2013). China-US BIT negotiations: Institutional factors, core terms and corresponding strategies. Research Institute of the World Economy and Politics, Chinese Academy of Social Sciences, IIS International Investment Working Paper (201305). Lipsey, R. (2001). Foreign direct investors in three financial crises (No. 8084). National Bureau of Economic Research, Inc. Liu, X., Luo, Y., Qiu, Z., & Zhang, R. (2014). FDI and economic development: Evidence from China’s regional growth. Emerging Markets Finance and Trade, 50(sup6), 87–106. Melitz, M. J. (2003). The impact of trade on intra-industry reallocations and aggregate industry productivity. Econometrica, 71(6), 1695–1725. Miner, S., & Hufbauer, G. C. (2015). State-owned enterprises and competition policy: The US perspective. Toward a US-China Investment Treaty, 16. Morck, R., & Yeung, B. (1992). Internalization: An event study test. Journal of International Economics, 33(1–2), 41–56. Ottaviano, G., Tabuchi, T., & Thisse, J. F. (2002). Agglomeration and trade revisited. International Economic Review, 409–435. Pan, Y., & Tang, J. (2013). New trend in national security investigation of the US foreign investment committee. International Economic Review, 5. Schott, J. J., & Cimino, C. (2015). The China-Japan-Korea trilateral investment agreement: Implications for US policy and the US-China bilateral investment treaty. Toward a US-China Investment Treaty, 50(24.9), 6. Sharma, S., Wang, M., & Wong, M. S. (2014). FDI location and the relevance of spatial linkages: Evidence from provincial and industry FDI in China. Review of International Economics, 22(1), 86–104. Shiau, H. L., Huang, C. J., & Chen, F. C. (2013). International involvement, target market selection, and consolidated performance: A firm-level analysis of Taiwan’s FDI in China. Emerging Markets Finance and Trade, 49(sup4), 184–196. Swenson, D. L. (1994). The impact of US tax reform on foreign direct investment in the United States. Journal of Public Economics, 54(2), 243–266.

200

7 The Potential Impact of China–US BIT on China’s …

Wang, B. (2013). China-US FDI: Challenges and deadlock. International Economic Review, 5. Wei, S. J. (2000). How taxing is corruption on international investors? Review of Economics and Statistics, 82(1), 1–11. Wei, S. J., & Shleifer, A. (2000). Local corruption and global capital flows. Brookings Papers on Economic Activity, 2000(2), 303–354. Yao, Z. (2013). How to respond to the China-US BIT substantial negotiations? International Economic Review, 6. Yeaple, S. R. (2003). The complex integration strategies of multinationals and cross country dependencies in the structure of foreign direct investment. Journal of International Economics, 60(2), 293–314. Zhang, F., & Yu, M. (2015). The potential impacts of BIT on China’s manufacturing industries. Toward a Sino–US Investment Treaty, 67–90. Zhang, F., & Zheng, J. (1999). The impacts of multinational companies on China’s economic structure and efficiency. Economic Research, 1, 45–52.

Chapter 8

China’s Opening-Up Policies: Achievements and Prospects

Abstract China began its era of reform and opening up in 1978, which profoundly changed China and has deeply influenced the world. As a result of its openingup policies, China has become the largest country in the world in goods trade and the second-largest country in services trade. Both inward and outward foreign direct investment (FDI) grew fast, reaching US$134.9 billion and US$129.8 billion, respectively, in 2018. Those figures represent 7.9% and 19.3% of the world FDI. Over the past four decades, China’s volume of trade in foreign goods has increased 204-fold, whereas its GDP has increased only 34-fold. In this respect, China has successfully pulled off a miracle in foreign trade. This chapter reviews the major practices and achievements of and lessons learned from China’s opening-up policy and forwards policy suggestions for China’s continued opening in the future.

China’s opening-up process can be regarded as having occurred in three stages: a stage characterized by expanding the extensive margin of opening, that is, a stage focused on increasing the quantity of resources, labor, or production, where “margin” denotes the range (before 2001); a stage of expanding the intensive margin of opening, which focused on improving the quality or output per unit of resource or labor (2001– 2017); and a general opening (since 2017) after the Chinese Communist Party (CCP) announced a new era of all-around opening up in China at its Nineteenth National Congress. Before opening, China had followed an import substitution strategy, which was characterized by a lesser reliance on imports and exports. China had also implemented a heavy industry–oriented development strategy, which entailed lowering interest rates, depreciating the renminbi, and distorting the prices of labor, raw materials, and even agricultural products—all to support the development of heavy industry. This strategy led to surplus labor, low incomes, and hence slow economic growth. After 1978 the Chinese government abandoned the heavy industry–oriented development strategy and began to employ an export-oriented strategy, which lasted until the global financial crisis of 2008–2009. Before China’s 2001 entry into the This chapter is a joint work with Dr. Tenglong Zhang and originally published in China 2049 by David Dollar, Yiping Huang, and Yang Yao, The Brookings Press, 2019. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022 W. Tian and M. Yu, Outward Foreign Direct Investment of Chinese Enterprises, Contributions to Economics, https://doi.org/10.1007/978-981-19-4719-3_8

201

202

8 China’s Opening-Up Policies: Achievements and Prospects

World Trade Organization (WTO), the major force driving export growth was a comparative advantage based on the country’s factor endowment. China is a laborabundant country and its labor costs are relatively cheap. Accordingly, China exported labor-intensive products and served as the largest “world factory.” Since China’s accession to the WTO, however, the cost of labor has increased significantly, and China no longer realizes a cost-saving advantage compared with such countries as Vietnam and other East Asian countries. Indeed, the fundamental force driving China’s export boom today is the country’s increased market size. The global financial crisis that began in 2008 had significant negative impacts on the global economy, especially in developed economies. In the context of weakening demand from the developed countries, China’s reliance on exports to drive economic growth was no longer tenable. China has begun switching from labor cost advantages to quality, brand, and service to drive exports. At the Nineteenth National Congress of the CCP, held in October 2017, the State Council made it clear that China’s economy was in the process of shifting from a focus on accelerating growth to a focus on high-quality development. As a result, improving the quality of export goods and services has become a top priority. At the same time, trade protectionism and antiglobalization sentiments are on the rise around the world. Against this background, the Chinese government has proposed implementing a new strategy that features opening up across the board, with the twin goal of promoting the development of both the Chinese economy and the global economy. What’s more, the China-U.S. trade conflict provoked by the United States has been escalating since the beginning of 2018. This conflict has had negative effects on both the Chinese and the U.S. economy and contributes to uncertainty about global economic development. Therefore, the external environment China faces has become increasingly complex. China intends to deal with the challenges calmly and confidently. Even as China must fight back against the attacks launched by the United States, the country plans to further open up to the world by broadening market access, improving the investment environment for foreign investors, strengthening the protection of intellectual property (IP) rights, and expanding imports, among other steps. This chapter explores the historical development of the extensive margin of the opening-up process, reviews the dynamic evolution of the intensive margin of opening up, considers the recent development of the generalized opening up announced by the State Council, and offers a set of policy recommendations.

8.1 Expanding the Extensive Margin, 1978–2000 China’s opening-up strategy before its accession to the WTO included the following important actions: setting up special economic zones and industrial parks, relaxing market access for foreign investors, reducing tariffs on imports, and encouraging trade processes. The following discussion elaborates on each of these four aspects,

8.1 Expanding the Extensive Margin, 1978–2000

203

with comments particularly on China’s expanding the extensive margin during the opening-up process.

8.1.1 Setting up Special Economic Zones and Industrial Parks China’s establishment of special economic zones (SEZs) was one of the most important changes resulting from the opening-up policy. Before 2000, the establishment of SEZs can be thought of as occurring in three waves. The first wave, also known as the “point” phase, entailed the adoption of four cities in 1981 as the first SEZs: Shenzhen, Zhuhai, and Shantou in Guangdong province and Xiamen in Fujian province. Shenzhen was chosen because of its excellent geographic location; it is a small village near Hong Kong. A similar rationale applied to the selection of Zhuhai, a small town on the western Pearl River near Macau. Shantou was chosen because of its strong network of connections with Chinese immigrants in East Asian countries. Similarly, Xiamen was chosen because it is close to Taiwan province. The second wave of the establishment of SEZs brought in many coastal cities, from Dalian, a northern coastal city in Liaoning province, to Beihai, a southern coastal city in Guangxi province. The cities are connected along a “line.” The third wave of economic zones extended to include zones and parks from the cities along the east coast to the central and western provinces. In particular, the government set up twenty-five high-tech industrial development zones in Shenyang, Tianjin, Wuhan, and Nanjing. Thus, by 1992 China had established six SEZs (the original four point zones, plus Hainan, and Pudong in Shanghai), fifty-four national-level economic and development zones, fifty-three high-tech industrial parks, and fifteen bonded zones (Naughton, 2018). All these zones—SEZs, economic deltas, economic and development zones, and even high-tech development zones—had very similar policy designs. In particular, foreign firms located in the zones were exempt from corporate tax in their first three years. In their fourth and fifth years, they had to pay a corporate tax rate of only 17%, which was half the tax rate levied on Chinese domestic firms. This policy lasted for around three decades. After 2007, foreign firms had to pay the same 25% corporate tax rate as domestic firms. In addition, wholly owned foreign firms and subsidiaries were permitted in all types of zones.

8.1.2 Relaxing Market Access for Foreign Direct Investment After 1978, the Chinese government began gradually relaxing access to the Chinese market for foreign investment. In 1979 the Sino-Foreign Equity Joint Venture Law was promulgated, making foreign-invested enterprises legal economic organizations.

204

8 China’s Opening-Up Policies: Achievements and Prospects

Later on, more laws and regulations were issued, all of which provided a legal basis for the supervision of foreign investment and protected the rights and interests of foreign investors. Foreign investment grew from almost nothing in 1978 to a cumulative amount of US$23.4 billion in 1991. In 1992, at the Fourteenth Congress of the CCP, China’s leadership announced a market economy would be set up, which markedly improved the willingness of foreign companies and individuals to invest in China. The open areas were extended from the cities along the east coast to the central and even western provinces. To facilitate foreign investors and improve the quality of foreign investment, the Chinese government released its Catalogue for the Guidance of Industries for Foreign Investment in 1995. The catalogue assigns all industries to one of four categories: encouraging, allowing, limiting, and forbidding foreign investment. China began to open some service sectors to foreign investment, including its retail, financial, freight, and software industries. After 1992, foreign investment flowed quickly into China, reaching a new high of US$38.9 billion in 1993. From 1994 to 1999, foreign investment topped US$40 billion each year.

8.1.3 Reducing Import Tariffs Before the economic reforms of 1978, China adopted the import substitution strategy by setting high import tariffs and other nontariff barriers against foreign products. In 1992, at the Fourteenth Congress of the CCP, the State Council announced it would set up a market economy. Thereafter China began actively cutting its import tariffs. The simple import tariff, still 42% in early 1992, was reduced to 35% in 1994 during the Uruguay round of WTO negotiations. During the next three years, China cut its import tariffs another 50% or so. At the end of 1997, the simple import tariff was reduced to around 17%. The most important reason for taking such a major step in opening up was that China hoped to accede to the WTO as soon as possible; in 1994, China was still only an observer, rather than a formal member, of the WTO. After its accession to the WTO, China’s simple import tariff was reduced from around 17% to around 10% in 2006. Since 2006, China’s import tariff rate has remained stable. Trade liberalization has been significant for the Chinese economy. To understand the country’s economic development, it is essential to understand the realization of firm productivity, since “productivity is not everything, but almost everything,” as Paul Krugman has said. It is widely accepted that trade liberalization fosters firm productivity. Yu (2015) finds that trade liberalization has contributed to around 14.5% of Chinese firm productivity growth in the twenty-first century.

8.1 Expanding the Extensive Margin, 1978–2000

205

8.1.4 Encouraging the Processing Trade The processing trade is a key for understanding China’s trade development over the past four decades. “Processing trade” denotes the process by which a domestic firm initially obtains raw materials or intermediate inputs from abroad and, after local processing, exports the value-added final goods (Feenstra & Hanson, 2005). As Dai et al. (2016) note, the iPhone is a perfect example of China’s processing trade. Foxconn, a famous iPhone assembler in Shenzhen, imports 155 intermediate components of smart phones from Japan, Korea, and the United States. After producing the final products domestically, Foxconn exports back to the United States and other foreign markets. Governments typically encourage processing trade by offering tariff reductions or even exemptions on the processing of intermediate goods. Processing imports and exports developed quickly after 1978, especially after 1992, when at the Fourteenth Congress of the CCP it was announced that the state would set up a market economy. Just before China’s accession to the WTO, the government decided to establish export processing zones (EPZs) to promote processing trade. Since then, China has set up more than sixty EPZs. Different from the earlier SEZs and industrial parks, such EPZs are spread across the country. They are located not only in eastern coastal cities but also in western inland cities, such as Urumchi in Xinjiang province. Only processing firms are allowed to enter the EPZs. They also enjoy special tariff treatment. In particular, firms in the EPZs are treated as “inside the territory but outside the customs,” as they are exempted from import duty. Figure 8.1 shows China’s export revenues by trade regimes (including ordinary trade, processing trade, and others) after 1978. As can be seen, export revenues in processing regimes increased to US$137.65 billion in 2000 from $1.13 billion in 1981. Processing export revenues increased to $675.1 billion in 2008 from $147.4 billion in 2001, with an annual average growth rate of 29%, far outpacing the annual average growth (12.3%) during 1992 and 2001. Processing export revenues exceeded ordinary export revenues in 1993, and this situation held until 2010. Figure 8.1 also shows China’s import revenues by trade regimes (including ordinary trade, processing trade, and other) after 1978. Import revenues in processing regimes increased to US$92.56 billion in 2000 from $1.5 billion in 1981. Processing import revenues increased to $378.4 billion in 2008 from $93.9 billion in 2001, with an annual average growth rate of 24.6%, also far outpacing the annual average growth (16.1%) during 1992 and 2001. Both lines were close before ordinary import revenues began to increase quickly after 2007. Industries that are intensively engaged in the processing trade are also laborintensive. The processing trade thus absorbs many workers. In particular, the four major processing industries are household appliances; toys; clothing, footwear, and hats; and leather goods. The four industries employed 13.2 million workers in 2009 and 16.2 million workers in 2014, accounting for more 10% of all employment in manufacturing. The processing trade has also contributed to China’s deep integration into global value chains. Accordingly, China has functioned as a nonsubstitutable

206

8 China’s Opening-Up Policies: Achievements and Prospects

Fig. 8.1 China’s export and import revenue by trade regimes, 1981–2017. Source Data from China’s Customs Statistics Yearbook (various years)

“world factory.” Of course, this also implies that China has naturally maintained a high global trade surplus. Indeed, around two-thirds of China’s trade surplus is generated from its processing trade.

8.1.5 Comments on the Stage of External Opening Before the era of economic reforms, China had adopted a heavy industry–oriented strategy and import substitution strategy. These strategies led to a surplus agricultural labor force, a low urbanization rate, and low resident income. All these factors laid a strong foundation for the export-oriented strategy implemented after the start of the reforms and opening-up policy. As one of the most important steps in opening up, China gradually established a series of economic zones and industrial parks between 1978 and 2000, providing preferential terms for firms locating within the zones or parks and promoting export growth. Meanwhile, the Chinese government also relaxed market access for FDI. With an abundant and cheap labor force but lacking in technological know-how and equipment, China’s choice of the processing trade as a way to participate in global trade was almost inevitable. The processing trade increased rapidly, and China became a new “world factory.” During this period, China applied to join the General

8.2 Internal Opening up, 2001–2017

207

Agreement on Tariffs and Trade (GATT) in 1986, and, beginning in 1992, started actively cutting import tariffs, which decreased by nearly 60% between 1992 and 1997. This move shows China’s determination to expand its opening up. However, China was not a member of the WTO during the stage of external opening, which posed a serious threat to China’s international trade. As an example, the United States granted China only temporary normal trade relations (temporary NTR) status. The U.S. Congress reviews that status annually. Chinese exporters would have faced the punishing non-NTR tariffs had the NTR status not been renewed. Therefore, China’s exports still faced great uncertainty. China hoped that accession to the WTO would solve the problem.

8.2 Internal Opening up, 2001–2017 Since the turn of the twenty-first century, China has focused more on opening internally than externally, as exemplified by the expansion of various special zones or industrial parks. In the twenty-first century, before the Nineteenth Party Congress of the CCP, five important events characterized the features of the intensive margin: accession to the WTO, expanding market access for FDI, the relaxation of outward FDI, establishment of free trade pilot zones, and new-economy pilot cities.

8.2.1 Accession to the WTO This section introduces each of the above five actions and comments on the intensive margin of opening up. Meeting the Tariff Reduction Commitment [end] As one of the conditions of joining the WTO, China committed to reducing its import tariffs to a certain level before 2006; the highest tariff rate China could set was known as the bound tariff rate. From 2001 to 2006, the applied average tariff rate was always less than the bound average tariff rate set in 2001. In other words, China met its tariff reduction commitment. Removal of Export Trade Policy Uncertainty [end] The nominal value of Chinese exports increased more than 16-fold between 1992 and 2008. An obvious acceleration of export growth took place around 2001, when China officially entered the WTO. Before WTO accession, from 1992 to 2001, the annual nominal export growth rate was about 14.1%, whereas after accession to the WTO, from 2002 to 2008, the annual nominal export growth rate reached as high as 27.3%. A substantial proportion of the soaring Chinese exports could be attributed to the elimination of the high tariff threat after its WTO accession. According to U.S. law, imports from nonmarket economies such as China are subject to relatively high tariff rates originally set under the Smoot-Hawley Tariff Act of 1930. These rates, known as “non-NTR” or “column 2” tariffs, are typically substantially larger than

208

8 China’s Opening-Up Policies: Achievements and Prospects

the “MFN” (most favored nation) or “column 1” rates the United States offers fellow members of the WTO. When China joined the WTO on December 11, 2001, the United States effectively assigned China permanent NTR status on January 1, 2002, which completely removed Chinese exporters’ concern over possible sudden tariff spikes (Pierce & Schott, 2016). Elimination of Nontariff Measures [end] China issues nontariff measures (NTMs) mainly focusing on ensuring food safety, human and animal health, product quality and safety, and environmental protection, which together account for about 90% of the total NTMs taken. China is actively following good international practices such as International Organization for Standardization (ISO) and International Electrotechnical Commission (IEC) standards in preparing its own national standards. Of the 1,448 mandatory standards related to NTMs, 555 standards (about 38%) are directly adopted from ISO, IEC, and standards promulgated by other international organizations. China is putting increasing effort into streamlining its national standards with international best practices and seeking international cooperation in the standardization process. Especially with China’s new Standardization Law, China intends to foster trade, economic, and social development by reducing restrictions. There are in total 2,071 NTMs identified by the General Administration of Quality Supervision, Inspection and Quarantine (AQSIQ). Of these 2,071 NTMs, only 646 measures apply unilaterally to all countries around the world. The remaining 1,425 measures (about 69%) apply bilaterally or plurilaterally to only a certain group of countries. Specifically, 896 (around 63%) measures out of the total measures that are bilaterally or plurilaterally applied by AQSIQ were implemented after 2010. This shows that over the years, China is increasingly moving from a unilateral relationship with countries—that is, one in which it applies the same measure to all countries—to a bilateral relationship.

8.2.2 Expanding Market Access for FDI After the Catalogue for the Guidance of Industries for Foreign Investment was first issued in 1995, it has been updated seven times, in 1997, 2002, 2004, 2007, 2011, 2015, and 2017. The number of encouraged categories increased from 115 in 1995 to 153 in 2017; the number of restricted categories decreased from fifty-six in 1995 to seven in 2017; and the number of prohibited categories decreased from ten in 1995 to seven today. This shows that the market access for foreign investment has been constantly expanding. And foreign investment could be guided to the industries China prioritizes for development. For example, foreign investment was expanded from labor-intensive industries to capital- and technology-intensive industries. Meanwhile, the Chinese government has also introduced preferential policies to encourage more FDI in Central and Western China. In addition, at this point almost all service sectors are accessible by foreign investors. Some of them have a limitation of holding ratio. Thanks to this policy, China’s import in services also grew quickly,

8.2 Internal Opening up, 2001–2017

209

Fig. 8.2 China’s annual utilized inward FDI, 1982–2018. Source Data from China’s Customs Statistical Yearbook (various years)

from US$248 billion in 2011 to US$514 billion in 2018, for an annual growth rate of 11.4%. Figure 8.2 plots China’s annual utilized inward FDI from 1983 to 2018. FDI grew to US$134.9 billion in 2018 from US$40.7 billion in 2000, with an average annual growth rate of 7%. It is worth noting that the growth rate has slowed in recent years. Specifically, the average annual growth rate was only 3% between 2011 and 2018.

8.2.3 Relaxation of Outward FDI Before 1991, only state-owned enterprises were allowed to invest abroad, and caseby- case approval was required, regardless of the investment amount (Voss et al., 2008). In the 1990s, China’s foreign reserve holdings increased fast as a result of trade surpluses and high FDI inflow, which promoted outward FDI (OFDI). But OFDI was still stringently restricted in this period by the government issuing the “Opinions on Strengthening the Management of Overseas Investment Projects” in 1991. As can be seen from Fig. 8.3, China’s OFDI had remained at a very low level from 1982 to 1999. The average OFDI from 1982 to 1999 was US$1.5 billion. In the twenty-first century, the government’s policy with respect to OFDI has switched from restriction to relaxation. In 2000 the Chinese government proposed

210

8 China’s Opening-Up Policies: Achievements and Prospects

Fig. 8.3 China’s annual flow of outward FDI, 1982–2018. Source Data from Ministry of Commerce of China; UNCTAD statistics

the “Going Out” strategy for the first time. OFDI increased dramatically, to US$6.9 billion in 2001 from US$1 billion in 2000, but fell back in the following three years. In 2004, many implementing rules were introduced by the government, such as the “Decision on Reforming the Investment Systems,” issued by the State Council, corresponding detailed policies to simplify the approval procedures, delegate approval authority, and increase approval efficiency promulgated by the National Development and Reform Commission (NDRC) and the Ministry of Commerce (MOFCOM). As a response, China’s OFDI more than doubled, to US$12.3 billion in 2005 from US$5 billion in 2004, and grew continuously in the next years. As a result of the 2008 global financial crisis, global FDI inflows fell by 14% in 2008 (UNCTAD 2009), while China’s 2008 OFDI flow was double that in 2007. The Chinese government further encouraged OFDI by raising the approval threshold to more than US$300 million for resource development projects and more than US$100 million for other projects. In 2014, China’s OFDI entered a new stage of “registration-based and approvalsupplemented” FDI. In this stage, only projects involving sensitive industries or with a value of more than US$1 billion are required to obtain official approval in advance. All other projects need only register with the provincial Development and Reform Commission (DRC). Established Chinese overseas enterprises are exempt from the approval and registration procedures. As a result, China’s OFDI increased further, to US$196 billion in 2016 from US$73 billion in 2014. Afterward, China’s OFDI fell in

8.2 Internal Opening up, 2001–2017

211

the ensuing two years because of escalating trade protectionism and antiglobalization behavior.

8.2.4 Establishment of Pilot Free Trade Zones In 2013, the government set up a pilot free trade zone (FTZ) in Shanghai. The area of the pilot is not large (its initial area is only around 29 km2 ), but its economic impact is potentially huge. The whole landscape of the establishment of the pilot FTZ can be summarized as 1 + 3 + 7, achieved in three steps. First, the government set up the first pilot FTZ in Shanghai in September 2013. Second, the pilot was extended to three other FTZs in large coastal provinces in April 2015: Guangdong, Tianjin, and Fujian. Third, in September 2016, the government set up seven more coastal and inland pilot FTZs, in Liaoning, Shaanxi, Henan, Hubei, Chongqing, Sichuan, and Zhejiang provinces. The objective of setting up the pilot FTZs is to replicate them in other non-FTZ places if those places are ready for the reform. Drawing on the country’s earlier experiment with SEZs, the government decided on four roles for the pilot FTZs. First, the pilot FTZs aim to promote further trade and investment facilitation. This goal is consistent with the development of EPZs. For goods within the pilot FTZs, the government requires “release the first line, but hold up the second line.” The idea is that the imported intermediate goods used in the FTZs will be exempt from tariffs (that is, released from the first line), but the final products that use such intermediate inputs cannot be sold outside the zones to China’s domestic market (hold up the second line). Second, the pilot FTZs are intended to promote China’s “negative list” investment mode. Differing from the previous “positive list,” the new negative list investment mode has fewer regulations or restrictions on foreign investment in China. If the products or sectors are listed, foreigners are not allowed to invest in those areas. In other words, foreigners may invest in anything that does not appear on the list. This gives foreign investors huge room to invest in new industries or sectors. It turns out that this policy design has been the most successful policy reform. And because it was so successful in all eleven FTZs, in 2018 the Chinese government decided to expand the policy to the whole country. Third, the pilot FTZs are intended to give a further push to China’s financial reform. In particular, the FTZs aim to promote financial innovation with convertible capital projects and by offering more financial services. This reform so far has had only limited effects. Its limited impact is not difficult to understand, as FTZs make up only a tiny proportion of China’s area. Unlike the trade reform, the financial reform cannot be clearly applied separately to entities located inside FTZs but not outside them. It is difficult to undertake a financial experiment in a small area and then apply it to the rest of the country. Fourth, the pilot FTZs require local governments to reduce bureaucratic procedures and simplify the process of doing business within the zones. Specifically, the FTZs emphasize after-event supervision rather than before-event approval.

212

8 China’s Opening-Up Policies: Achievements and Prospects

8.2.5 New-Economy Pilot Cities Experiment Different from the establishment of the pilot FTZs, the opening up via the neweconomy pilots in twelve cities is less well known. In 2015 the government decided to choose twelve cities located in five city groups, as well as some coastal cities, to experiment with the so-called “new-economy pilot.” Jinan, the capital of Shandong province, is the largest of the twelve cities. Zhangzhou, a coastal city in Fujian province, was chosen mainly because of its strong connections with Taiwan province. Fangchenggang, a coastal city in Guangxi province, was chosen because it neighbors one of the SEZs in Vietnam. The government hopes to develop a bilateral border trade by setting up this new-economy zone. The other eight cities are located around China’s five major metropolitan areas. Dalian, in Liaoning, and Tangshan, in Hebei, are the two northern cities closest to the Beijing-Tianjin-Hebei metropolitan area. Xian was chosen because it is the largest city in Northwestern China. Chongqing was chosen because it is one of the core cities among the Chengdu-Chongqing megacities. Wuhan and Nanhang were chosen because they are two large cities in Central China. Finally, Pudong, in Shanghai, and Suzhou, in Jiangsu, are connected to the Yangtze River Delta economic belt. The experiment of these new-economy zones focuses on six areas. First, it aims at exploring the new management mode of the government. Second, it aims at exploring the coordination of various industrial parks. Third, it hopes to explore new ways to encourage FDI. Fourth, it aims at promoting high-quality exports of domestic products. Fifth, it seeks significant improvement of financial services. Sixth, the zones should focus on promoting all-around opening-up in the regions. After three years of the experiment, the independent evaluation panel, led by the National School of Development at Peking University, expressed satisfaction with the twelve cities’ new-economy reforms. Particularly, the panel recognized that the reforms have successfully improved local economic development and been helpful in the supply-side structural reform.

8.2.6 Comments on the Stage of Intensive Margin of Opening up After accession to the WTO, China realized its commitment to reduce import tariffs, and the trade policy uncertainty faced by Chinese exporters was drastically reduced. These moves facilitated the rapid growth of imports and exports. As shown in Fig. 8.4, China’s foreign trade dependence ratio, which is the ratio of total international trade volume to GDP, rose quickly between 2001 (38.1%) and 2006 (63.95%). This means that the contribution of foreign trade to GDP growth increased significantly. Meanwhile, China further expanded market access for FDI and relaxed restrictions on OFDI, both of which also increased quickly during this period.

8.2 Internal Opening up, 2001–2017

213

Fig. 8.4 Foreign trade dependence ratio in China, 1978–2018 (%). Source Data from China’s Customs Statistical Yearbook (various years)

After the financial crisis broke out in 2008, foreign markets were badly shocked, facing weak demand and trade protectionism. Since the turn of the twenty-first century, and especially after 2004, the cost of labor in China has increased dramatically and the country’s population dividend has shrunk fast. Compared with many East Asian countries, China no longer has a significant comparative advantage on the labor- intensive margin. A not small part of China’s foreign markets has been taken over by countries such as Vietnam and Bangladesh. Thus strategies that continue to rely on a simple processing trade to promote China’s exports are no longer feasible. Indeed, both processing import and export volumes decreased dramatically after the global financial crisis broke out (see Fig. 8.1). At the Eighteenth National Congress of the CCP, China’s leadership proposed deepening the reform of China’s foreign trade system, promoting openness through institutional reform, and fostering new competitive advantages in foreign trade. To this end, the Chinese government began to establish pilot FTZs in 2013 and to implement the new-economy pilot cities experiment in 2016. These two policy initiatives have succeeded in improving local economic development and have been helpful in the supply-side structural reform. At the Nineteenth Party Congress, held in 2017, it was emphasized that China’s economy has shifted from a high-speed, increasing-volume stage to a high-quality development stage. The Chinese government has proposed the next step, an allaround opening up, that would promote the development of both the Chinese

214

8 China’s Opening-Up Policies: Achievements and Prospects

economy and the global economy. The specific measures to implement an all-around opening up are discussed in the next section.

8.3 Features of the All-Around Opening up China’s strategy of an all-around opening up heavily relies on rolling out the Belt and Road Initiative (BRI), and developing China’s free trade ports and the Guangdong– Hong Kong–Macau Greater Bay Area.

8.3.1 Belt and Road Initiative The BRI, which was initiated by the Chinese government in 2013, is devoted to improving regional connectivity on a transcontinental scale. The initiative aims to strengthen infrastructure, trade, and investment links between China and other countries through which the BRI will pass. Currently, sixty-four countries are actively involved in the BRI. These include ten ASEAN countries, eighteen countries in Western Asia, eight in South Asia, five in Central Asia, seven in the Commonwealth of Independent States, and sixteen in Central and Eastern Europe. The scope of the initiative is still taking shape—recently, the BRI has been interpreted as open to all countries as well as to international and regional organizations. In addition to trade and FDI, the BRI also concentrates on infrastructure projects. As one of the largest infrastructures and investment projects in history, the BRI addresses the “infrastructure gap” and thus has the potential to accelerate economic growth across the involved countries. The initiative calls for integration of the countries into a cohesive economic area through building infrastructure, increasing cultural exchange, and broadening trade and investment.

8.3.2 Free Trade Ports Experiment The notion of free trade ports was first brought up in a report to the Nineteenth Party Congress. President Xi Jinping explicitly said that the country will allow more freedom to reform the pilot FTZs and explore the establishment of free trade ports. The main feature of a free trade port is that, from the perspective of administrative supervision, it is outside the customs jurisdiction of the country. A free trade port combines the features of a port and an FTZ, with many trade-related functions, including product processing, logistics, and warehousing. But it is a more open platform than an FTZ. The construction of free trade ports will help FTZs advance toward the goal of being a more transparent institutional environment, like Singapore and Hong Kong. Meanwhile, breakthroughs in the areas of trade facilitation measures,

8.4 Policy Recommendations

215

ship fuel prices, financial support, customs supervision, and inspection and quarantine are necessary for free trade ports. As a result, free trade ports will be able to respond better to the profound changes in the global environment.

8.3.3 Greater Bay Area The construction of the Guangdong–Hong Kong–Macau Greater Bay Area (GBA) is one of the most urgent tasks facing China in its opening-up process. The GBA includes nine municipal areas and the Special Administrative Regions of Hong Kong and Macau. The nine municipal cities are on the east and west banks of the Pearl River: Shenzhen, Dongguan, Huizhou, Guangzhou, Foshan, Jiangmen, Zhaoqing, Zhongshan, and Zhuhai. In this way the GBA coincides with the Pearl River Delta, which is a top EDZ in China, along with the Yangtze River Delta. The development of the GBA should focus on the following perspectives. First, to build the GBA, the real economy and the financial economy should be combined, focusing on the real economy. The services industry can play an auxiliary role to boost the GBA economy. The second goal in constructing the GBA is to facilitate innovation. China is transforming from a manufacturing power to an innovation power. The GBA should play a key role in the formation of an innovative country. The third objective is to achieve institutional innovation. Fourth, the GBA should pay more attention to its ecological environment. Other all-around opening measures include (1) further widening market access, (2) improving the investment environment for foreign investors, (3) strengthening protection of IP rights, and (4) taking the initiative to expand imports.

8.4 Policy Recommendations This chapter has reviewed the practices and achievements of China’s liberalization of international trade and investment and the country’s integration into the international economy since 1978. China has done well so far. However, the internal and external environments in which China must operate have become increasingly complex, with rising labor costs, an aging population, and, in the international arena, accelerating unilateralism and the China-U.S. trade conflict. As emphasized by President Xi, China’s opening door will not be closed and will only open even wider. Below are listed several concrete policy recommendations to promote China’s continued opening in the future. Support multilateralism and contribute to the reform of the WTO. Unilateralism and protectionism are on the rise around the world, posing a challenge to the authority and effectiveness of the multilateral trading system. In this context, China should actively participate in WTO reforms. Specifically, China should reinforce the role of the WTO Dispute Settlement Committee and seek to gain a seat on the committee.

216

8 China’s Opening-Up Policies: Achievements and Prospects

Consolidate existing regional free trade agreements (FTAs) and actively promote new FTAs. To date, China has concluded seventeen FTAs with twenty-five partners in Europe, Asia, Oceania, South America, and Africa. China could deepen its cooperation with these partners in various fields such as trade, investment, and cultural exchange. Thirteen other regional FTAs are in the negotiations stage. The most important of these is the Regional Comprehensive Economics Partnership (RCEP). Since 2012, there have been twenty-two rounds of negotiations on establishing the RCEP. China has been actively communicating with the negotiating parties during the current critical round, hoping to address major disagreements and conclude negotiations sometime in late 2019. In addition, China could also consider participating in the Comprehensive Progressive Trans-Pacific Partnership (CPTPP), which is now led by Japan, after the United States dropped out. There are two important reasons for China to participate. First, the CPTPP has lower requirements for openness than the TPP, ones similar to China’s existing policies for openness. Second, joining the CPTPP would expand China’s circle of friends and offset the negative impacts on China’s economy caused by China-U.S. trade frictions. Deal with the China-U.S. trade conflicts by further opening up and pursuing highquality development. China will be obliged to respond to the attacks launched by the United States by setting high tariff and nontariff barriers for U.S. exports. The Chinese government should provide necessary subsidies for industries and workers who bear the brunt of the costs of this trade conflict. China’s GDP is approaching that of the United States and is expected to surpass it by 2027, when the effects of the ChinaU.S. trade friction peak. Once China’s GDP is 1.5 times that of the United States, the United States may accept China’s rise, and China-U.S. relations are expected to shift from competition to cooperation. However, the competition between China and the United States will exist for a long time, potentially for the next thirty years. The Chinese government and Chinese people should be psychologically prepared and concentrate on doing their own jobs well. Expand market access and improve the quality of openness. The Chinese government should try to expand market access for foreign countries by encouraging more foreign investments and imports. Specifically, the negative investment list system should be implemented nationwide, and all companies registered in China should be treated equally. The convenience of doing business in China should also be improved. For example, a secure and effective electronic customs clearance system would speed up the integration of customs and simplify the agency’s procedures. Furthermore, developing a better living environment and better services for businesspersons and international talent is a necessary step for China to improve the movement of personnel as well as the ability to attract talent. China should also strengthen and enforce protections for IP rights. IP rights protection is required by foreign enterprises and should be required by Chinese enterprises. It would provide the biggest boost to the competitiveness of the Chinese economy. To reduce China’s technological dependence on foreign countries, the Chinese government should set up special funds to encourage enterprises to develop core technologies.

8.5 Conclusion

217

At present, China has a deficit in its trade in services, which is consistent with China’s current stage of economic development. However, in the future, China should vigorously develop the service industry and service trade, making service trade another engine of China’s economic growth.

8.5 Conclusion This chapter has described China’s international trade and investment development and opening-up policies the four decades since the country began its economic reform. Overall, the opening-up policy has occurred in three waves: increasing the extensive margin, increasing the intensive margin, and an all-around opening up. International trade is a major focus of China’s opening-up policy. China’s growing role in international trade before the country’s accession to the WTO was driven mainly by the country’s cheap labor pool, which gave it a comparative advantage. This was especially useful for labor-intensive industries, where the cost of labor is one of the most important input factors. Moreover, cheap labor to some extent affected the incremental exports of machinery and transport equipment. This is because China’s foreign trade in machinery and transport equipment is mainly conducted through its processing trade, which takes advantage of the low cost of labor. After China’s accession to the WTO, the main driving force of China’s incremental international trade became the realization of scale economies with the large international market, according to the increasing returns-to-scale theory. One piece of evidence of this is that the share of the processing trade keeps decreasing, whereas total trade volume is increasing (see Fig. 13.1). This phenomenon is not accounted for in traditional comparative advantage theory. In recent years, because of increasing labor costs and an expanding domestic market in China, weak foreign demand and trade protectionism in developed countries, and cheap labor costs in other developing countries and their increasing presence in global trade, China’s economy has pivoted toward high-quality development and away from high-volume production and the processing trade. In international trade, China is striving to shift from a trader of quantity to a trader of quality. The Chinese government has proposed an all-around opening up as the next major phase. The new driving force of China’s international trade will be the dividends of institutional reform achieved by expanding imports, expanding market access for foreign firms, strengthening IP rights protection, and nurturing innovation. Policy recommendations to ensure China’s continuing sustainable and highquality opening up include (1) supporting multilateralism and contributing to the WTO reform; (2) consolidating existing FTAs and actively promoting new ones; (3) dealing with the China-U.S. trade conflicts reasonably, calmly, and confidentially; and (4) further expanding market access and improving the quality of openness.

218

8 China’s Opening-Up Policies: Achievements and Prospects

References Dai, M., Maitra, M., & Yu, M. (2016). Unexceptional exporter performance in China? The role of processing trade. Journal of Development Economics, 121, 177–189. Feenstra, R. C., & Hanson, G. H. (2005). Ownership and control in outsourcing to China: Estimating the property-rights theory of the firm. The Quarterly Journal of Economics, 120(2), 729–761. Naughton, B. J. (2018). The Chinese economy: Adaptation and growth. MIT Press. Pierce, J. R., & Schott, P. K. (2016). The surprisingly swift decline of US M Pierce, J. R., & Schott, P. K. (2016). The surprisingly swift decline of US manufacturing employment. American Economic Review, 106(7), 1632–1662. Voss, H., Buckley, P. J., & Cross, A. R. (2008, April). Thirty years of Chinese outward foreign direct investment. In The CEA UK Conference-China’s Three Decades of Economic Reform (1978–2008) (pp. 1–2). Yu, M. (2015). Processing trade, tariff reductions and firm productivity: Evidence from Chinese firms. The Economic Journal, 125(585), 943–988.

Appendix A

For “Distribution, Outward FDI, and Productivity Heterogeneity: China and Cross-Countries’ Evidence”

Appendix A1: Proof of Proposition 1 Proof: The derived demand for product ϕ is  −σ X j (ϕ) = L j P jσ −1 p cj (ϕ) where L j is labor income in country j, p cj (ϕ) = σ σ−1 MC c , c = d, e, f s, f m is the price of product if it is domestically sold, exported without distribution foreign affiliate, exported with a distribution affiliate or with production affiliates respectively. P j is the aggregate price level, where ⎧ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨

⎡

⎤⎫

⎪ μ f s μ τ 1−σ ϕ 1−σ ⎪ ∞ 1−σ ⎢ sh j τh j ⎥⎪ j hj 1 σ ⎪ ∫ dG(ϕ) + dG(ϕ) + ∫ dG(ϕ)⎦ ⎪ ⎪ ϕ + ηj ϕ ϕ σ −1 L h ⎣ ∫ ⎬ h=1,h = j  ϕeh j ϕ ϕ f mh j j f mh j Pj =

1−σ ⎪ ⎪ ∞ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ dG(ϕ) + σ σ−1 L j ∫ ϕ1 ⎪ ⎪ ⎪ ⎪ ⎩ ⎭ ϕd j



N 













where ϕd j , ϕeh j , ϕ f sh j , ϕ f mh j , are the productivity cut-o¤ point for each mode.  1−σ , ϕd < ϕe . So From Eqs. (1) and (2), we know that when ff DX > τ + ϕe η









when ff DX < τ 1−σ , ϕd < ϕe . 1 Deriving the LHS of (3) with respect to , get ϕ     

−σ    1−σ 1−σ −σ σ −1 μτi j τ d − ϕi j + η j − τ ϕ1 + η /d ϕ1 = (1 − σ )τ μ σ τ ϕ1 < ϕ 0, which means a higher ϕ induces higher relative returns of distribution investment compared with export without FDI. Thus (3) has a single solution. Similarly, we derive the LHS of (4) with respect to ϕ1 to get  

1−σ   1−σ  −σ  μτ d ϕ1 − ϕi j < 0. So the /d ϕ1 = (1 − σ ) 1 − (μτ )1−σ ϕ1

© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022 W. Tian and M. Yu, Outward Foreign Direct Investment of Chinese Enterprises, Contributions to Economics, https://doi.org/10.1007/978-981-19-4719-3

219

220

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

higher is ϕ, the more profitable is building a production a plant relative to a distribution affiliate. Because the LHS of (3) is ϕ, so ϕe < ϕ f s equals





μτ ϕe

 i.e.

μτ



1−σ −



τ +η ϕe

1−σ

1−σ fX B


1, so 0 <

k−(σ −1) 1 ϕˆ

< 1 for any ϕ. ˆ Thus, the above equation

k

satisfies V Profit1 < 1−σ ϕ f s dG(ϕ) is the expected profit from export without FDI, (N − 1) ∫ ϕτ + η ϕe and kb N. k−(σ −1)





1−σ 1−σ ϕ f s ϕ f s τ τ +η dG(ϕ) dG(ϕ) < (N − 1) ∫ (N − 1) ∫ ϕ + η ϕfs ϕe ϕe   1−σ  k   k τ 1 1 k +η − = (N − 1)b ϕfs ϕe ϕfs  1−σ < (N − 1)bk τ + η < (N − 1)bk η1−σ ϕfs

















 fE kbk k 1−σ k−(σ −1) N +(N −1)b η

so B > of B

1 1−σ

, and Δ =

fE kbk k 1−σ k−(σ −1) N +(N −1)b η

1  1−σ

is an upper bound

.

Appendix A2: Proof of Proposition 2



It is obvious from Eqs. (2–4) that an increase in η raises ϕe , lowers ϕ f s , and does not affect ϕ f m , a decrease in μ decreases ϕe , does not affect ϕ f s , and increases ϕ f m ; and a decrease in τ decreases ϕe and increases ϕ f m . We only need to verify a decrease in τ decreases ϕ f s . Derive the LHS of (3) with respect to τ so that we get













222

 d

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China … μτi j ϕ

1−σ

−

τi j ϕ

+ ηj

1−σ 

(1−σ ) ϕ

/d(τ ) =



μ

σ −1 σ

τ ϕ

−σ

−

τ ϕ

+η

−σ 

< 0.:

So a lower τ leads to a higher relative variable profit from engaging in distribution FDI, thus generates a lower ϕ f s . It is worthwhile to note that when μ is sufficiently small, η is large and fIS is not large.

enough, ϕf s could be smaller than ϕe . For example, suppose σ = 2; and f I S <



fX

τ +η μτ



− 1 ; then from (2) and (3) we get 

τ







μτ



ϕfs

ϕe

+η 

−1−

−1

τ



ϕfs

+η

−1 =

fX fI S

Rearrange it to get

   1 1 + ηϕe μτ − τ +ηϕ f s   τ +ηϕe τ +ηϕe fX = fI S − μτ τ +ηϕ f s

ϕe =

fX  τ fI S











  ϕ τ +ηϕe ϕ τ +ηϕe Suppose e < 1, then < 1, then e > ffIXS − 1 > μτ ϕ ϕfs τ +ηϕ f s

 fs ϕ f X τ +η − 1 > 1, contradicts. So in this case, e > 1, ϕe > ϕ f s . (Q. E. D.) fI S μτ ϕfs





















Appendix A3: Distribution of Zhejiang FDI Firms Zhejiang’s firm-level FDI flow data are a good proxy for understanding the nationwide Chinese firms’ FDI flow for the following reasons. First, the FDI flow from Zhejiang province is out-standing in the whole of China. Firms in Zhejiang have engaged in FDI since 1982. Such firms were the pioneers of Chinese FDI activity. As reported by MOC, only around 10 firms began to engage in FDI before 1982. Since then, Zhejiang has maintained a fast growth rate similar to that of other large eastern provinces, such as Guangdong, Jiangsu, and Shandong. In 2008, Zhejiang had 2,809 FDI firms (including greenfield firms and M&A firms), accounting for 21% of all FDI firms in China, and became the largest province in the number of FDI firms. In terms of FDI flow, Zhejiang’s FDI also maintained a high plateau, ranking at the very top in the entire country from 2006 to 2009. Zhejiang’s FDI accounted for 16% of the country’s FDI flow and became the largest FDI province in 2010. Second, the distribution of type of ownership of FDI firms in Zhejiang province is consistent with that across the country. According to the Statistical Bulletin of China’s outward

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

223

Foreign Direct Investment (MOC, 2013) of the Ministry of Commerce, 75% of all Chinese FDI firms are private limited liability corporations in terms of the number of firms.1 In Zhejiang province, 70% of FDI firms are private firms. Third, the distribution of Zhejiang FDI firms’ destinations is similar to that of the whole country. Up to 2009, Chinese FDI firms invested in 177 countries (regimes) and 71.4% of FDI volume was invested in Asia. Hong Kong is the most important destination for Chinese FDI firms.2 This observation also applies to Zhejiang’s FDI firms’ Most FDI firms in Zhejiang invest in Asia, Europe, and North America. Hong Kong and the United States are the two destinations with the largest investments. The most common investment mode is to set up production affiliates and create a marketing network by establishing a trade-oriented office. Finally, the industrial distribution of Zhejiang’s FDI firms is similar to that for the whole of China. According to the Statistical Bulletin of China’s outward Foreign Direct Investment (MOC, 2013), the top sector for Chinese FDI firms investment is retail and wholesale. This is similar to the case of Zhejiang. The lower module of Table 1 shows the number of FDI firms in 2006–08, resulting in a total of 1,270 FDIfirm-year observations in the database.

Appendix A4: TFP Measure The main interest of this chapter is to investigate how firm productivity affects firm FDI. Hence, it is crucial to measure firm productivity accurately. Traditionally, TFP is measured by the estimated Solow residual between the true data on output and its fitted value using the OLS approach. However, the OLS approach suffers from two problems, namely, simultaneity bias and selection bias. Following Amiti and Konings (2007) and Yu (2015) in assuming a Cobb–Douglas production function, we adopt the augmented Olley-Pakes semi-parametric approach to deal with simultaneity bias and selection bias in measured TFP. In particular, we tailor the standard Olley-Pakes approach to t the data for China with the following extensions. First, we use deflated prices at the industry level to measure TFP. Previous studies, such as De Loecker (2011), stressed the estimation bias of using monetary terms to measure output when estimating the production function. In that way, one actually estimates an accounting identity. Hence, we use different price deflators for inputs and outputs. Admittedly, it would be ideal to adopt firm-specific prices as the deflators. Unfortunately, the firm-level data set does not provide sufficient information to measure prices of products. Following previous studies, such as Goldberg et al. (2010), we adopt industry-level input and output deflators for TFP measures. As in 1

In 2013, SOEs accounted for only 8% of the total number of outward FDI firms, although they accounted for 55.5% of total outward FDI volume. 2 Note that it is possible that some Chinese ODI firms take Hong Kong as an international investment exprót. since Hong Kong is a popular “tax haven.” This phenomenon is beyond the scope of the present chapter, although it would be interesting for future research.

224

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

Brandt et al. (2012), the output deflators are constructed using “reference price” information from China’s Statistical Yearbooks, whereas input deflators are constructed based on output deflators and China’s national Input–Output Table (2002). Third, it is important to construct the real investment variable when using the Olley-Pakes (1996) approach.3 As usual, we adopt the perpetual inventory method to investigate the law of motion for real capital and real investment. The nominal and real capital stocks are constructed as in Brandt et al. (2012). Rather than assigning an arbitrary number for the depreciation ratio, we use the exact firm’s real depreciation provided by the Chinese firm-level data set.4 In particular, by assuming that the expectation of future realization of the unobserved productivity shock, vit , relies on its contemporaneous value, firm i’s investment is modeled as an increasing function of unobserved productivity and log capital,kit ≡ lnK it . Following previous works such as Amiti and Konings (2007), the Olley. Pakes approach was revised by adding the firm’s export decision as an extra argument in the investment function since most firms’ export decisions are determined in the previous period: Iit = I˜(ln K it , vit , E Fit ), where E Fit is a dummy to measure whether firm i exports in year t. Therefore, the inverse function of (7) is vit = I˜−1 (lnK it , Iit , E Fit )..5 The unobserved productivity also depends on log capital and the firm’s export decisions. Accordingly, the estimation specification can now be written as: ln Yit = β0 + βl ln L it + g(lnK it , Iit , E Fit )+ ∈it where g(lnK it , Iit , E Fit ) is defined as βk lnK it + I˜−1 (lnK it , Iit , E Fit ) Following Olley and Pakes (1996) and Amiti and Konings (2007), fourth-order polynomials are used in log-capital, log- investment, firm’s export dummy, and import dummy to approximate g(·).6 In addition, since the firm dataset is from 2000 to 2006, we include a WTO dummy (i.e., one for a year after 2001 and zero for before) to characterize 3

In the literature, the Levinsohn and Petrin (2003) approach is also popular in constructing TFP in which. In the literature, the Levinsohn and Petrin (2003) approach is also popular in constructing TFP in which materials (i.e., intermediate inputs) are used as a proxy variable. This approach is appropriate for firms in countries not using a large amount of imported intermediate inputs. However, such an approach may not directly apply to China, given that Chinese firms substantially rely on imported intermediate inputs, which have prices that are significantly different from those of domestic intermediate inputs (Halpern et al., 2011). 4 Note that even with the presence of exporting behavior, the data still exhibit a monotonic relationship between TFP and investment. 5 Olley and Pakes (1996) show that the investment demand function is monotonically increasing in the productivity shock V it , by making some mild assumptions about the firm’s production technology. 6 Using higher order polynomials to approximate g(‘) does not change the estimation results.

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

225

the function g(·) as follows: g(kit , Iit , E Fit ) = (1 + E Fit )

4 4  

q

δhq kith Iit

h=0 q=0

After finding the estimated coefficients βˆm and βˆl , we calculate the residual Rit which is defined as Rit ≡ lnYit − βˆl ln L it . The next step is to obtain an unbiased estimated coefficient of βk . To correct the selection bias, Amiti and Konings (2007) suggest estimating the probability of a survival indicator on a high-order polynomial in log-capital and log-investment. One can then accurately estimate the following specification:   Rit = βk ln K it + I˜−1 gi,t−1 − βk ln K i,t−1 , pr ˆ i,t−1 + ∈it .

where pri denotes the fitted value for the probability of the firm.s exit in the next year. Since the specific “true” functional form of the inverse function I˜−1 (·) is unknown, it is appropriate to use fourth-order polynomials in gi,t−1 and lnK i,t−1 to approximate that. In addition, (10) also requires the estimated coefficients of the log-capital in the first and second term to be identical. Therefore, non-linear least squares seem to be the most desirable econometric technique. Finally, the Olley-Pakes type of TFP for each firm i in industry j is obtained once the estimated coefficient βˆk is obtained: T F PitO P = lnYit − βˆk lnK it − βˆl lnL it

References Amiti, M., & Konings, J. (2007). Trade liberalization, intermediate inputs, and productivity: Evidence from Indonesia. American Economic Review, 97(5), 1611–1638. Yu, M. (2015). Processing trade, tariff reductions and firm productivity: Evidence from Chinese firms. The Economic Journal, 125(585), 943–988. De Loecker, J. (2011). Product differentiation, multiproduct firms, and estimating the impact of trade liberalization on productivity. Econometrica, 79(5), 1407–1451. Goldberg, P. K., Khandelwal, A. K., Pavcnik, N., & Topalova, P. (2010). Imported intermediate inputs and domestic product growth: Evidence from India. The Quarterly Journal of Economics, 125(4), 1727–1767. Levinsohn, J., & Petrin, A. (2003). Estimating production functions using inputs to control for unobservables. The Review of Economic Studies, 70(2), 317–341.

226

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

Appendix A5: Extensive Margin Estimates of Zhejiang Sample Table A3 examines whether our previous findings based on nationwide FDI decision data hold for Zhejiang’s FDI manufacturing firms. The linear probability model estimates in column (1) confirm that Zhejiang’s high-productivity manufacturing firms are more likely to engage in FDI during the sample period 2006.08. The probit estimates in column (2) yield similar findings with a slightly larger coefficient of firm TFP. Of 100,743 manufacturing firms during the sample period, there are only 407 FDI manufacturing firms, as shown in the lower module of Table 1. That is, the probability of FDI is only 0.39%, suggesting that firm FDI activity is also a rare event in Zhejiang province during the sample period and the standard logit estimation results may have a downward bias. In Table 8, we again correct for such bias by using rare-events logit estimates in column (3) and complementary log–log estimates in column (4). The estimated coefficients of firm productivity are much larger than their counterparts in columns (1) and (2), indicating that the downward bias in the regular estimates is fairly large. The increases in the odds ratio caused by firm productivity are similar to their counterparts in Table 5. Columns (5) and (6) perform multinomial estimates in which the regressand is distribution FDI in column (5) and non-distribution FDI in column (6). The coefficient of firm productivity in column (5) is still positive and significant whereas that in column (6) is positive but insignificant. A less important but interesting finding is that the SOE control variable turns to be positive and significant. We suspect such striking findings are due to the inclusion of processing FDI in the category of nondistribution FDI. By dropping such processing FDI in the multinomial estimates in columns (7) and (8), the coefficient of firm productivity in column (8) once again is positive and significant; more importantly, its magnitude is larger than that of distribution FDI, indicating that high-productivity firms are more likely to engage in non-distribution FDI. The coefficients of SOE variable turns to be negative, though still insignificant, as in other previous estimates. Tables A1, A2 and A3. Table A1 Summary statistics of distribution FDI Firm number

Nationwide FDI data 2000–2007 Distribution FDI other FDI

Zhejiang FDI data 2006–2008 Distribution FDI other FDI

Before merge

203,948%2205

96,776%304

After merge

20,359%142

33,783%68

Note The FDI type in Zhejiang data is classified to 12 types including wholesale, business office, production, processing trade, R&D, construction, mining, market seeking, agriculture, housing, and product design

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

227

Table A2 Transitional matrix of FDI mode for FDI firms Year t + 1 Year t

Distri. FDI

Non-Distri. FDI

Distri. FDI

0.91

0.09

non-Distri. FDI

0.25

0.75

Note Each number in the table is the probability of the firm’s outward FDI mode in t + 1, conditional on the FDI mode in t. Non-FDI firms in both periods are dropped

Appendix B: For “Outward FDI and Domestic Input Distortions: Evidence from Chinese Firms” Online Appendix A: Proofs Proof of Proposition 1 Proof: The first two parts have already been proved. Here we prove the last two parts. Because the monotone likelihood ratio property (MLRP) implies first-order stochastically dominance (FOSD), we only need to prove the part 3 under the assumption of FOSD. First, the fraction of MNCs among each type of firm is f racimnc

  1 − Fi ϕ i O  , = 1 − Fi ϕ i D

where i ∈ {P, S} and Fi (ϕ) is the cumulative probability density function (CDF) of the productivity draw. Note that since ϕ P D > ϕ S D , a sufficient condition for to hold is f rac Smnc < f rac Pmnc     1 − FS ϕ S O 1 − FP ϕ P O < 1 − FP (ϕ P D ) 1 − FS (ϕ P D ) Since the FOSD property holds for the truncated productivity distributions and ϕ S O > ϕ P O , it must be true that       1 − FS ϕ S O 1 − FP ϕ P O 1 − FS ϕ P O , < < 1 − FP (ϕ P D ) 1 − FS (ϕ P D ) 1 − FS (ϕ P D ) which leads to the result that the fraction of MNCs is larger among private firms than among SOEs. Second, average productivity of active private firms is ∞

∫

ϕPD

∞ ϕ f P (ϕ) 1 − FP (ϕ) dϕ = ϕ P D + ∫ dϕ 1 − FP (ϕ P D ) 1 − FP (ϕ P D ) ϕPD

0.244*** (11.88) −0.343 (−0.91) −0.065 (−1.23)

0.006*** (6.56)

0.003*** (14.38)

−0.003 (−0.77)

−0.002*** (−3.92)

0.003*** (8.60)

Finn relative TFP

SOE indicator

Foreign indicator

Log firm labor

Export indicator

Probit

Rare events

Comp

100,847

Number of observations

Yes 100,847

Yes

100,847

Yes

Yes

No

Yes 100,847

Yes

100,847

Yes

Yes

No

100,832

Yes

Yes

—

0.004 (0.41)

2.188*** (9.09)

0.660*** (10.53)

100,832

Yes

Yes

Yes

0.034*** (2.81)

3.112*** (3.03)

0.666*** (3.94)

0.398 (0.93)

−12.791 (−0.01)

−12.695 (−0.03) −0.222 (−1.34)

1.648* (1.66)

(2.77) 0.962***

Non-Dist (8)

Multinomial logit Distribution (7)

Notes The regressands in columns (1)–(4) are the FDI indicator. Numbers in parentheses are t-values. *** (“,*) denotes significance at the 1(5, 10)% level. The multinomial logit estimates in columns (5)–(6) include all non-FDI and FDI firms whereas those in columns (7)–(8) include all firms except processing FDI firms. The base types in all multinomial logit estimates are non-FDI firms

100,743

Yes Yes

Yes

Yes

Year fixed effects

Industry fixed eflects

No

—

No

Processing FDI dropped

No

0.502 (1.36)

0.178 (0.50)

0.627 (0.57)

0.702*** (5.11)

0.500 (0.65)

0.328** (2.01)

−0.221 (−1.33)

−13.078 (−0.03)

0.660*** (10.53)

0.961*** (2.77)

Non-Dist (6)

Multinomial logit Distribution (5)

Firm tenure

2.127*** (9.86)

−0.175 (−1.19)

−0.854 (−0.84)

0.672*** (12.22)

0.879*** (2.80)

log–log (4)

1.866*** (3.82)

2.111*** (9.67)

−0.172 (−1.16)

−0.387 (−0.39)

0.678*** (12.74)

0.882*** (2.73)

logit (3)

2.187*** (9.09)

0.664*** (10.43)

0.292** (2.55)

(2)

LPM

(1)

Econometric method:

Regressand: FDI indicator

Table A3 Extensive margin estimates for Zhejiang firms (2006–08)

228 Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China … ∞

> ϕSD + ∫

ϕSD ∞

> ϕSD + ∫

ϕSD

229

1 − FP (ϕ)   dϕ 1 − FP ϕ S D 1 − FS (ϕ)   dϕ 1 − FS ϕ S D

where the first line comes from integration by parts, and the second line is true as ϕ S D < ϕ P D , The last step is true because the truncated distribution of the productivity draw also satisfies the FOSD property. Furthermore, as ∞

∫

ϕSD

∞ 1 − FS (ϕ) ϕ f S (ϕ)   dϕ   dϕ = ϕ S D + ∫ 1 − FS ϕ S D ϕ S D 1 − FS ϕ S D

we have the result that average productivity of private firms is greater than that of SOEs overall. For the proof of part 4, we have to impose a stronger assumption that both types of firms make productivity draws from the same distribution (i.e., f (ϕ) = f P (ϕ) = f S (ϕ)), although this is not a necessary condition for the result to hold. Under this assumption, we have ∞

∫

ϕPO

ϕ f (ϕ) dϕ 1−F (ϕ P O )

∞

= ϕPO + ∫

1−F(ϕ) dϕ 1−F (ϕ P O )

< ϕSO + ∫

1−F(ϕ) dϕ 1−F (ϕ S O )

ϕPO ∞

∞

= ∫

ϕSO

ϕSO

ϕ f (ϕ) dϕ, 1−F (ϕ S O )

which implies that (simple) average productivity of private MNCs is smaller than that of state-owned MNCs.

Proof of Proposition 2 Proof: Comparing Eq. (4.13) with Eq. (4.14) in the main text, we know that the productivity premium of state-owned MNCs increases with the level of domestic distortion (i.e., selection into the FDI market becomes much less stringent for private firms compared with SOEs), or ψϕ so (> 1) increases with c. Furthermore, selection into fo the domestic market becomes more stringent for private firms compared with SOEs when c increases, as ψψ P D (> 1) increases with c. Therefore, the first part follows. 5D For the second part, since we have μ = 1 now, the production function becomes  q(k, l) = ϕ and TVC and FC (for SOEs) become

k 0.5

0.5 

l 0.5

0.5 (1)

230

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

T V C(q, ϕ) =

qr ϕ 0.5

(2)

and FC(q, ϕ) =

fi r ω0.5

(3)

where i ∈ {e, D, X, I}. Repeating the procedure as before, we obtain ϕ ϕPX = S X > 1; ϕ S O > ϕ P O ; ϕ S D < ϕ P D . ϕSD ϕPD Furthermore, it is straightforward to establish that both ψϕ S O (> 1) and ψϕ F D (> 1) PO SD increase with c. Therefore, the productivity premium of state-owned MNCs is more pronounced in capital intensive industries. And, SOEs are much less likely to engage in FDI (relative to private firms) in capital intensive industries.

Proof of Proposition 3 Proof: For the first part, the relative size of private MNCs (i.e., compared with private non-exporting firms) is 

k    k  ϕPD σ r −1 ϕPO 1 − ϕ πP D ϕ P O 1 − ϕ P D PX    =

k−(σ −1)  k−(σ r −1) ϕ σPrD−1 1 − ϕϕ P D π P D (ϕ P D ) 1 − ϕϕ P D PX

PX

under the Pareto assumption. Similarly, for SOEs, the relative size is 

k      k σ −1 π S D ϕ S o 1 − ϕϕ S D 1 − ϕSD ϕ SX So  

k−(σ −1) 

k−(σ −1)  =   ϕSD 1 − ϕ σS −1 π S D ϕ S D 1 − ϕϕ S D D ϕ SX

SX

Since ϕ ϕPX = S X > 1, ϕ S O > ϕ P O , ϕ S D < ϕ P D ϕSD ϕPD the relative size of private MNCs (i.e., compared with private non-exporting firms) is smaller than that of state-owned MNCs. We now prove the second part. Comparing Eq. (4.12) with Eq. (9) in the main text and noting that overall sales are proportional to the operating profit, we conclude

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

231

that the ratio of foreign sales to domestic sales is higher for private MNCs (than for state-owned MNCs), conditioning on ϕ. This is because domestic sales are smaller for private firms than for SOEs, conditioning on the productivity draw, ϕ. For the third part of the proposition, there are three cases to consider. The first case is that both types of firms are non-exporters before the reduction in f I . Equations (7), (9), (11) and (4.12) in the main text together imply that π S O(ϕ) π P O (ϕ) > π P D(ϕ) π S D(ϕ) which is what we need to prove (remember that overall sales are proportional to the operating profit). The second case is that both types of firms are exporters before the reduction in fI. In this case, Eqs. (8), (10), (11) and (4.12) in the main text together imply that π P O (ϕ) π S O (ϕ) > π P X (ϕ) π S X (ϕ) Therefore, after two firms with the same φ undertake FDI, the increase in overall firm size is greater for the new private MNC than for the new state-owned FDI firm. The final case is that the SOE is an exporter and the private firm is a non-exporter before the reduction of the fixed FDI cost. In this case, we have π P O (ϕ) π So (ϕ) π P O (ϕ) > > π P D (ϕ) π P X (ϕ) π S X (ϕ) since π P X (ϕ) > π P D (ϕ). Therefore, after two firms with the same φ undertake FDI, the increase in overall firm size is larger for the new private MNC (than for the new state-owned MNC). In total, the third part of this proposition is true for all possible cases.

Online Appendix B: Outward FDI Between 2000 and 2013 In this appendix, we use the new sample with the longer time span to check the extensive margin of outward FDI. For 2000–13, the MNC ratio for private firms is 0.93%, whereas that for broadly defined SOEs is 0.70%. This finding suggests that the fraction of MNCs is larger among private firms than among SOEs, which is consistent with our theoretical prediction and our finding using data for 2000–08. Since firm productivity cannot be precisely estimated using the new data set, we do not check the productivity premium of state-owned MNCs. Instead, we focus on examining whether SOEs are still less likely to engage in outward FDI, even after we include data after 2008.

232

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

Table 9 picks up this task. Similar to the estimates in Table 5, the regressand is the firm’s outward FDI indicator, whereas the SOE indicator is the key regressor. In all estimates, we control for the log of employment and log of firm size as well as the firm’s export indicator. Column (1) is the simple linear probability model, and columns (2) and (3) are logit estimates. It turns out that, once again, the coefficient of the SOE indicator is negative and statistically significant, suggesting that SOEs are less likely to undertake outward FDI. Column (4) uses rare-event logit to correct for rare-event bias; the rest of the table uses complementary log–log regressions. In particular, column (6) uses a broadly defined SOE indicator, and column (7) drops observations with outward FDI to tax haven destinations. Column (8) drops observations before 2004, and columns (9) and (10) only include observations after the global financial crisis (2010–13). Finally, column (10) drops the switching SOEs (to private firms) from the sample. In all respects, our previous key finding that SOEs are less likely to engage in outward FDI is shown to be robust. As further robustness checks for our previous findings, we use observations until 2013 to run the diff erence-in-differences regressions with emphasis on the industrylevel interest rate differential and the di fference between capital-intensive industries and labor-intensive industries. The results are reported in Tables 10 and (4.14). Similar to our findings using the sample of 2000–2008, SOEs are still less likely to engage in outward FDI when industry-level interest rate differential (between SOEs and private firms) becomes larger. Furthermore, they are still less likely to engage in outward FDI when they come from capital intensive industries (compared to SOEs coming from labor intensive industries). In all respects, it is still true that SOEs are less likely to engage in outward FDI in sectors that experience more severe distortion distortion (in terms of the cost of borrowing). Furthermore, it is still true that SOEs are less likely to engage in outward FDI in sectors that have higher demand for working capital, since the mag- nitude of the interacted coefficient of the SOE indicator and capital-intensive indicator is larger than that of the SOE indicator and labor-intensive indicator.

Online Appendix C: Variants of the Model Fixed FDI Cost In this subsection, we assume that the fixed FDI cost is paid using domestic factors. Under current specification, we derive FDI cutoffs as ( f I − f X )r H

 1 μ−1 β−1 1 + ωH and

⎡

 σ −1  σ −1 ⎤ μ−1 α−1 μ−1 β−1 1 + ω 1 + ω σ −1 ⎢ F H DF  ⎥ βϕ So − = ⎣ ⎦ σ −1 σ (τ r H )σ −1 rF

(4)

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

⎡ σ −1 ⎢ DF  ( f I − f X )cr H βϕ P O = ⎣ 1   σ 1 + (cω H )μ−1 μ−1

μ−1

1 + ωF

 α−1 α−1

 −

r Fσ −1

1 + (cω H )

μ−1

233 α−1  β−1

(cτ r H )σ −1

⎤ ⎥ ⎦. (5)

Denote the inverse of domestic marginal cost (after normalizing φ to one) as

 1 μ−1 μ−1 1 + ωH

  x H r H, w H =

(6)

rH

and the inverse of foreign marginal cost as

x F (r F , w F ) =

 1

μ−1 μ−1 1 + ωF rF

.

(7)

Note that the existence of the input price wedge increases the domestic marginal cost, or. x H (r H , w H ) > x H (c H , w H ). A sufficient and necessary condition for ϕ S O > ϕ P O (for any c > 1) is that. τ σ −1 x F (r F , w F )σ −1 (x H (r H , w H ) − x H (cr H , w H )) < x H (r H , w H )σ − x H (cr H , w H )σ r ,

which puts an upper bound on the marginal production cost in China (i.e., “H”).7 The above condition is more likely to hold in the case of China (especially before 2008), as China enjoyed relatively low production costs compared with developed economies. Another variant of the above model is that both types of firms use domestic resources to pay for the fixed FDI cost, and private firms do not face discrimination when they pay for this fixed cost. This assumption receives some empirical support, as the Chinese government is actively seeking to support the “Going-Out” strategy of Chinese firms which include private firms. For this variant of the model, FDI cutoffs can be derived as ( f I − f X )r H 1

 β−1 1 + ωμ−1 H

⎡

 σ −1  σ −1 ⎤ μ−1 β−1 μ−1 σ −1 1 + ω 1 + ω   F H DF σ −1 ⎢ ⎥ βϕ S o = − ⎣ ⎦ σ (τr H )σ −1 r Fσ −1

and 7

Note that since σ > 1,

x H (r H ,w H )σ −x H (cr H ,w H )σ x H (r H ,w H )−x H (cr H ,w H )

.

(8)

234

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

fI rH

μ−1

1 + ωH

=

f X cr H

− 1  p−1

1 + (cω H )μ−1

⎡

σ −1 ⎢ DF  βϕ P O ⎣ σ

1 + ωμ−1 F

1  μ−1

−1  σβ−1

 −

r Fσ −1

1 + (cω H )

μ−1

−1  σα−1

σ −1

(cτr H )

⎤ ⎥ ⎦

(9)

Obviously, the selection reversal result holds irrespective of parameter values ( i.e. ϕ S O > ϕ P O ), since there is no difference in the fixed cost of engaging in FDI between SOEs and private firms.

Variable FDI Cost In this subsection, we modify our basic model to allow SOEs to use domestic factors when producing abroad. SOEs would have incentive to do so, if x H (r H , w H ) > x F (r F , w F ) > x H (cr H , w H ) and firms are allowed to bring domestic factors to the foreign country to produce. Under this specification, FDI cutoffs can be derived as fI rF

μ−1

1 + ωF

=

− 1  μ−1

fX rH 1 + (ω H )μ−1

⎡

σ −1 ⎢ DF  βϕ S o ⎣ σ

μ−1

1 + ωH

1  β−1

σ −1  μ−1

−

r Hσ −1

μ−1

1 + ωH

(τr H )

σ −1 ⎤  μ−1

σ −1

⎥ ⎦

(10)

and fI rF

μ−1

1 + ωF

=

1 + (cω H )μ−1

1  μ−1

⎡ ⎤  σ −1   α−1 μ−1 β−1 σ −1 ⎢ 1 + ω F 1 + (cω H )μ−1 μ−1 ⎥ − βϕ P O ⎣ ⎦ (cτr H )σ −1 r Fσ −1

DF  σ

f X cr H

− 1  μ−1

(11)

A sufficient condition for the selection reversal result to hold (i.e., ϕ S O > ϕ P O ) is.   τ σ −1 x H (r H , w H )σ −1 − x F (r F , w F )σ −1 < x H (r H , w H )σ −1 − x H (cr H , w H )σ −1 ,

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

235

where x H (., .) and x F (., .) are defined in Eqs. (6) and (7) respectively. Note that this condition is a sufficient but non-necessary condition for the selection reversal result to hold. Absent general equilibrium feedback, the above inequality holds if the distortion is more severe (i.e., x H (r H , w H ) is small enough) or the difference in the undistorted factor prices across countries is small (i.e., x H (r H , w H ) is close enough to x F (r F , w F )).

Online Appendix D: Propensity Score Matching for Productivity Comparison Regarding the matching of MNCs, since there are not enough observations to match based on destination economy-industry-investment mode pairs, we group MNCs into: (i) developed and developing destination economies according to the World Bank’s classification; (ii) capital-intensive and labor-intensive sectors; (iii) three types of investment motives: horizontal, vertical, and R&D seeking. Moreover, since a firm could switch from SOE to non-SOE, we also include dummy variables for each year as covariates in the PSM matching. Thus, we have four covariates in our new PSM for the MNC sample: capital-intensive indicator, year dummies, destination country indicator (developing or developed), and the variable of in- vestment mode. We group investment modes into three categories. Specifically, horizontal FDI includes: production affiliate (code: 4), processing (5), market (6), wholesale (7), and trading centers (8). Vertical FDI includes foreign resource utilization (1) and real estate (2). The R&D seeking FDI include both research and development (3) and consulting service (9). Regarding the non-MNCs matching, we use capital-intensive indicator and year dummies as new covariates to perform the PSM matching. The results are shown in columns (2) and (4) of new Table 2. As expected, private non-MNCs are more productive than state-owned non-MNCs. Moreover, in the standard Melitz-type models, firm size is a sufficient statistic for productivity. Therefore, we do not use firm sales or employment as our covariates in the PSM. However, adding firm-size variables such as employment and sales as new covariates do not change our results.

Tables for Online Appendix In this section, we report summary statistics of our main sample (2000–08) in Table 1. Next, we report input elasticities of our TFP estimates (for non-MNCs and MNCs respectively) in Table 2, Table 3, In Tables 4 and 5, we report average productivity of all firms, non-MNCs and MNCs in our TFP estimations using pooled sample (i.e., MNCs and non-MNCs). In Table 6, we report absolute size premium for state-owned MNCs. In Table 7, we show that private firms are more likely to start FDI. In Table

236

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

Table 1 Summary statistics of key variables (2000–08) Variable

Mean

Std. dev

Min.

Max.

Firm TFP (Olley-Pakes)

3.61

1.18

0.61

6.57

Firm FDI indicator

0.003

0.066

0

1

Firm export indicator

0.29

0.451

0

1

SOE indicator

0.04

0.191

0

1

SOE indicator (broader)

0.07

0.252

0

1

Foreign indicator

0.20

0.402

0

1

Firm log employment

4.78

1.115

1.61

13.25

8, we show that private firms are more likely to conduct and start FDI, even after we have controlled for the industry fixed effects and destination market fixed effects. We report estimation results for the longer sample (2000–13) in Tables 9, 10 and 11.

Reference Hsieh, C. T., & Song, Z. M. (2015). Grasp the large, let go of the small: The transformation of the state sector in China (No. w21006). National Bureau of Economic Research.

0.064

0.032

28

0.071

0.048

0.046

31

32

0.048

0.089

0.070

0.065

0.083

0.064

29

30

0.085

0.026

0.078

26

27

0.120

0.031

25

0.004

0.070

0.054

0.086

0.052

23

0.060

22

0.041

0.068

0.042

0.043

0.056

0.249

0.123

0.053

0.084

(2)

Capital

24

0.080

0.077

20

0.071

19

21

0.063

0.109

17

0.049

16

18

0.042

0.118

14

15

(1)

0.086

13

Labor

Chinese

2-digit

TFPsoe

Measures

0.900

0.823

0.826

0.745

0.893

0.744

0.821

0.832

0.820

0.809

0.847

0.867

0.819

0.855

0.780

0.854

0.784

0.754

0.867

0.778

(3)

Materials

Table 2 Estimates of Olley-Pakes TFP by non-MNCs only

0.047

0.048

0.029

0.009

0.011

0.010

0.013

0.323

0.058

0.079

0.058

0.013

0.072

0.014

0.246

0.066

0.044

0.120

0.011

0.056

(4)

Labor

TFPDistort

0.047

0.086

0.071

0.075

0.080

0.135

0.074

0.136

0.045

0.004

0.079

0.040

0.084

0.064

0.092

0.057

0.089

0.126

0.067

0.103

(5)

Capital

0.900

0.823

0.823

0.740

0.889

0.742

0.819

0.820

0.819

0.809

0.846

0.865

0.817

0.842

0.778

0.854

0.783

0.754

0.865

0.776

(6)

Materials

0.046

0.049

0.030

0.008

0.011

0.011

0.013

0.322

0.058

0.080

0.056

0.014

0.072

0.014

0.248

0.066

0.006

0.121

0.012

0.057

(7)

Labor

TFPDistort soe

0.046

0.081

0.071

0.070

0.081

0.112

0.077

0.126

0.075

0.001

0.076

0.071

0.076

0.063

0.094

0.056

0.286

0.124

0.065

0.105

(8)

Capital

(continued)

0.900

0.823

0.823

0.740

0.890

0.742

0.819

0.822

0.819

0.809

0.846

0.865

0.818

0.842

0.778

0.854

0.766

0.753

0.866

0.776

(9)

Materials

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China … 237

0.049

0.131

0.088

0.049

0.087

40

41

42

0.050

0.856

0.854

0.770

0.801

0.833

0.814

0.820

0.829

(4)

0.089

0.024

0.070

0.012

0.018

0.023

0.010

0.001

0.006

0.049

0.087

0.043

0.072

0.084

0.101

0.071

0.068

0.070

(5)

Capital

0.854

0.768

0.801

0.825

0.812

0.820

0.829

0.855

0.902

(6)

Materials

0.092

0.024

0.071

0.012

0.019

0.023

0.010

0.001

0.006

(7)

Labor

TFPDistort soe

0.049

0.087

0.047

0.071

0.067

0.078

0.072

0.062

0.070

(8)

Capital

0.855

0.767

0.801

0.825

0.811

0.820

0.829

0.855

0.902

(9)

Materials

Notes This table reports the input ealasticities for three types of augmented Olley-Pakes TFP measures for non-MNCs only. Columns (1)–(3), (4)–(6) report the input ealasticities for the Olley-Pakes TFP controlling for SOE dummy (TFPsoe ), for input price distortions (TFPDistort ), and for both SOE dummy and ), respectively. The Chinese industries and associated codes are classified as follows: Processing of foods (13), Manufacture of input price distortions (TFPDistort soe foods (14), Beverages (15), Tabacco (16),Textiles (17), Apparel (18), Leather (19), Timber (20), Furniture (21), Paper (22), Printing(23), Articles for cultures and sports (24), Petroleum (25), Raw chemicals (26), Medicines (27), Chemical fibers (28), Rubber (29), Plastics (30), Non-metallic minerals (31), Smelting of ferrous metals (32), Smelting of non-ferrous metals (33), Metal (34), General machinery (35), Special machinery (36), Transport equipment (37), Electrical machinery (39), Communication equipment (40), Measuring instruments (41), and Manufacture of artwork (42). We do not report the standard errors for each estimated coefficient to save space, although they are available upon request

0.056

0.093

0.074

0.067

37

0.078

0.042

36

39

0.052

0.067

0.046

0.055

34

35

0.903

(3)

(2)

0.056

(1)

0.036

2-digit

33

Labor

Capital

Labor

Chinese

Materials

TFPDistort

TFPsoe

Measures

Table 2 (continued)

238 Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

0.064

0.027

28

0.067

0.048

0.046

31

32

0.046

0.091

0.060

0.046

0.083

0.065

29

30

0.074

0.026

0.075

26

27

0.122

0.030

25

0.002

0.020

0.055

0.086

0.046

23

0.059

22

0.033

0.063

0.066

0.046

0.060

0.054

0.088

0.055

0.081

(2)

Capital

24

0.081

0.076

20

0.072

19

21

0.062

0.110

17

0.047

16

18

0.043

0.116

14

15

(1)

0.086

13

Labor

Chinese

2-digit

TFPsoe

Measures

0.900

0.824

0.826

0.745

0.891

0.746

0.822

0.829

0.822

0.808

0.848

0.868

0.819

0.856

0.781

0.855

0.792

0.757

0.868

0.780

(3)

Materials

Table 3 Estimates of Olley-Pakes TFP by MNCs only

0.044

0.048

0.029

0.016

0.011

0.006

0.013

0.349

0.058

0.081

0.066

0.008

0.076

0.011

0.228

0.065

0.029

0.118

0.013

0.049

(4)

Labor

TFPDistort

0.045

0.091

0.054

0.064

0.083

0.089

0.068

0.114

0.022

0.006

0.079

0.047

0.065

0.063

0.096

0.055

0.070

0.080

0.058

0.095

(5)

Capital

0.901

0.824

0.824

0.741

0.889

0.744

0.820

0.815

0.821

0.808

0.847

0.866

0.819

0.843

0.779

0.855

0.791

0.757

0.867

0.778

(6)

Materials

0.042

0.049

0.029

0.014

0.012

0.004

0.013

0.357

0.059

0.082

0.064

0.007

0.077

0.011

0.228

0.065

0.034

0.118

0.014

0.050

(7)

Labor

TFPDistort soe

0.049

0.091

0.054

0.068

0.089

0.088

0.068

0.113

0.027

0.006

0.080

0.047

0.063

0.066

0.096

0.055

0.087

0.085

0.057

0.097

(8)

Capital

(continued)

0.900

0.824

0.824

0.741

0.889

0.743

0.820

0.815

0.821

0.807

0.847

0.866

0.819

0.843

0.778

0.855

0.779

0.756

0.867

0.777

(9)

Materials

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China … 239

0.088

0.146

0.088

0.041

0.081

40

41

42

0.044

0.856

0.854

0.774

0.800

0.835

0.815

0.821

0.829

(4)

0.085

0.041

0.071

0.017

0.016

0.020

0.007

0.006

0.012

0.051

0.158

0.037

0.069

0.041

0.065

0.052

0.052

0.083

(5)

Capital

0.856

0.771

0.799

0.826

0.813

0.821

0.829

0.855

0.903

(6)

Materials

0.085

0.041

0.071

0.017

0.016

0.021

0.008

0.006

0.010

(7)

Labor

TFPDistort soe

0.049

0.157

0.036

0.069

0.051

0.068

0.053

0.051

0.079

(8)

Capital

0.855

0.770

0.799

0.826

0.813

0.821

0.829

0.855

0.902

(9)

Materials

Notes This table reports the input ealasticities for three types of augmented Olley-Pakes TFP measures for MNCs only. Columns (1)–(3), (4)–(6) report the input ealasticities for the Olley-Pakes TFP controlling for SOE dummy (TFPsoe ), for input price distortions(TFPDistort ), and for both SOE dummy and input price ), respectively. The Chinese industries and associated codes are classified as follows: Processing of foods (13), Manufacture of foods (14), distortions (TFPDistort soe Beverages (15), Tabacco (16),Textiles (17), Apparel (18), Leather (19), Timber (20), Furniture (21), Paper (22), Printing (23), Articles for cultures and sports (24), Petroleum (25), Raw chemicals (26), Medicines (27), Chemical fibers (28), Rubber (29), Plastics (30), Non-metallic minerals (31), Smelting of ferrous metals (32), Smelting of non-ferrous metals (33), Metal (34), General machinery (35), Special machinery (36), Transport equipment (37), Electrical machinery (39), Communication equipment (40), Measuring instruments (41), and Manufacture of artwork (42). We do not report the standard errors for each estimated coefficient to save space, although they are available upon request

0.050

0.093

0.074

0.062

37

0.078

0.042

36

39

0.038

0.058

0.046

0.054

34

35

0.903

(3)

(2)

0.068

(1)

0.037

2-digit

33

Labor

Capital

Labor

Chinese

Materials

TFPDistort

TFPsoe

Measures

Table 3 (continued)

240 Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

0.020*** (54.91)

0.495

0.016*** (47.01)

0.479 0.016*** (46.72)

0.482

0.498

(3)

Distor t RTFPsoe

0.019*** (54.83)

0.484

0.503

(4)

RTFPop

0.016*** (47.11)

0.478

0.494

(5)

RTFPDistort

Non-MNC firms

0.016*** (46.83)

0.481

0.497

Distor t (6) RTFPsoe

0.011 (1.00)

0.514

0.525

RTFPop (7)

−0.027*** (−2.67)

0.528

0.501

RTFPDistort (8)

MNC firms

−0.028*** (−2.70)

0.531

0.503

Distor t RTFPsoe (9)

Notes Number in parenthesis are t-value. *** (**,*) denotes the significance at 1(5, 10)%, respectively. Columns (1)–(3) show that private firms have sshigher relative TFP than SOEs for all firms. Similarly, columns (4)–(6) show that private non-MNC firms have higher relative TFP than SOE non-MNC firms. Columns (8)-(9) show that private MNC firms are less productive than state-owned MNCs. Columns (1), (4) and (7) are relative Olley-Pakes TFP. Columns (2), (5) and (8) are relative TFP featured with input factor distortions. Columns (3), (6) and (9) are relative TFP controlling for input price distortions and interacted SOE dummy. All types of TFP measures pool all firms within an industry and control in the first stage for firm’s MNC status

Difference = (i)–(ii)

0.504

0.484

(2)

(1)

(ii) SOE

RTFPDistort

RTFPop

All firms

(i) Private firms

Category measures

Table 4 Productivity premium of state-owned MNC by different types of relative TFP (2000–08)

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China … 241

All firms

0.39*** (58.82)

Difference = (i)–(ii) 0.007*** (6.49)

0.518

0.525

RTFPLevPet (2)

Non-MNC firms

0.40*** (59.08)

10.29

10.69

Labor productivity (3)

0.006*** (6.57)

0.519

0.525

RTFPLevPet (4)

MNC firms

−0.588* (−4.46)

11.72

11.14

Labor productivity (5)

−0.088*** (−2.80)

0.684

0.596

RTFPLevPet (6)

Notes Columns (1)–(2) show that private firms have higher log labor productivity and relative TFP (measured in Levinsohn-Petrin) than SOEs for all firms. Similarly, columns (3) and (4) show that private non-MNC firms have higher log labor productivity and relative TFP than SOE non-MNC firms. Columns (5) and (6) show that private MNC firms are less productive than SOE MNC firms. Number in parenthesis are t-value, *** (**,*) denotes the significance at 1(5, 10)%, respectively

10.69

10.30

(i) Private firms

Labor productivity (1)

(ii) SOE

Category measures of RTFP

Table 5 Robustness checks fbr productivity premium of state-owned MNCs (2000–06)

242 Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

5.19

6.88

−1.69*** (−140.8)

(i) Private firms

(ii) SOE

Difference = (i)–(ii)

0.550*** (163.30) Yes Yes 323,397 0.21

0.068*** (21.56)

Yes

Yes

323,397

0.07

Firm TFP

Year-specific Fixed Effects

Firm-specific Fixed Effects

Number of Observations

R-squared

0.15

1,375

Yes

Yes

0.180*** (4.41)

1.795*** (4.78)

−1.82*** (−7.85)

6.55

4.73

Lnl (3)

0.33

1,375

Yes

Yes

0.683*** (15.03)

1.701*** (4.07)

−367,772** (−2.26)

549,485

181,713

Sales (4)

FDI non-exporting firms

0.16

2,352

Yes

Yes

0.345*** (7.61)

2.400*** (7.68)

−2.52*** (−14.14)

8.29

5.77

Lnl (5)

MNCs

0.31

2,352

Yes

Yes

0.807*** (15.95)

2.841*** (8.14)

−8,019,798*** (−5.49)

11,130,681

3,110,883

Sales (6)

0.21

2,058

Yes

Yes

0.938*** (11.51)

3.727*** (6.84)

−8,472,556*** (−8.48)

10,347,231

1,874,675

of MNCs (7)

Domestic sales

Note Columns (1)–(6) of the upper module show that private firms have lower sales and employment than SOEs for non-FDI exporting firms, FDI non-exporting firms, and MNCs, respectively. Column (7) in the upper module shows that domestic sales of private MNCs are smaller than those of state-owned MNCs. The lower module regresses firm size (in log employment) and firm sales on the SOE indicator while controlling for firm TFP, year-specific fixed effects, and firm-specific fixed effects. All the regressions show that SOEs are larger than private firms among non-FDI exporting firms, non-exporting MNCs, and MNCs. The numbers in parentheses are t-values. *** (**, *) denotes significance at the 1% (5%, 10%) level

1.491*** (70.83)

1.566*** (79.35)

−69,535*** (−26.71)

130,238

60,703

SOE indicator

Regressions

Lnl (1)

Variable Sales (2)

Non-FDI exporting firms

Category

Table 6 Absolute size premium for SOEs

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China … 243

1.347*** (16.08)

No

Yes

No

Yes

0.002*** (15.77)

No

No

Yes

No

Foreign firms dropped

Tax haven dropped

Year fixed effects

Distribution FDI No dropped

No

Export indicator

Industry fixed effects

Switching SOE dropped

No

No

0.583*** (21.84)

No

Yes

No

Yes

No

Yes

1.621*** (16.98)

0.603*** (18.87)

2.975*** (5.40)

No

Yes

No

Yes

No

Yes

1.624*** (19.09)

0.557*** (21.23)

3.820*** (9.44)

(9)

0.001*** (14.56)

2.641*** (5.19)

(8)

Log firm labor

Narrow

0.004*** (6.10)

(6)

(10)

Narrow

No

Yes

No

Yes

No

Yes

1.620*** (16.98)

0.598*** (19.02)

2.933*** (5.40)

No

Yes

No

Yes

No

Yes

1.620*** (16.97)

0.596*** (18.98)

2.934*** (5.40)

No

Yes

No

Yes

Yes

Yes

1.708*** (16.39)

0.569*** (16.65)

2.773*** (4.63)

No

Yes

Yes

Yes

No

Yes

1.159*** (9.00)

0.720*** (16.41)

3.629*** (4.98)

Yes

Yes

No

Yes

No

Yes

1.621*** (16.92)

0.598*** (18.88)

2.871*** (5.20)

(continued)

Yes

Yes

No

Yes

No

Yes

1.520*** (16.14)

0.587*** (18.94)

2.691*** (4.70)

−1.071*** −0.981*** −1.402*** −0.818*** −1.114*** −1.212*** (-4.79) (-4.29) (−5.06) (−2.99) (−4.68) (−4.59)

(5)

(7)

Firm TFP

Narrow

−0.001*** −0.878*** −1.079*** −1.579*** (−5.46) (−3.97) (-4.79) (−7.54)

(4)

Broad

SOE indicator

(3)

Narrow

Narrow

(2)

Narrow

(1)

Variable

Narrow

Narrow

SOE defined

Narrow

2004–2008

Rare event logit Complementary log–log

2000–2008

Logit

Year coverage

Logit

LPM

Regressand: Starting FDI indicator

Table 7 Private firms are more likely to start FDI

244 Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

1,135,468

896,315

859,096

No 859,096

No 858,705

No 857,641

No

707,154

No

554,760

Yes

(10)

Narrow

Note The regressand is the starting FDI indicator. All columns except column (1) include industry dummies at the 2-digit level and year dummies. The numbers in parentheses are t-values clustered at the firm level. *** (**) denotes significance at the 1% (5%) level. Columns (1)–(2) include foreign-invested firms whereas all other columns drop those firms. Columns (1)–(8) cover data over the period of 2000–2008 whereas Columns (9)–(10) cover data over the period of 2004–2008. Column (6) uses broadly defined SOE. Column (7) drops outward FDI to tax haven destinations. Column (8) drops distribution-oriented FDI. Column (9) drops the switching SOEs (i.e., switching from SOEs to private firms). Column (10) drops both switching SOEs and merge & acquisition deals. In all columns, TFP is measured by augmented Olley-Pakes controlling for input price distortions

859,096

No

(9)

1,136,604

No

(6)

(8)

Observations

Narrow

No

(5)

(7)

No

Narrow

M&A deals dropped

(4)

Broad

Narrow

(3)

Narrow

(2)

(1)

Variable

Narrow

Narrow

Narrow

SOE defined

Narrow

2004–2008

Rare event logit Complementary log–log

2000–2008

Logit

Year coverage

Logit

LPM

Regressand: Starting FDI indicator

Table 7 (continued)

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China … 245

1,136,604 1.00

1,136,604

1.00

Observations

R-squared

Yes

0.32

1,136,604

Yes

No

Yes

Yes

−0.000 (−0.73)

Starting FDI indicator

0.33

1,136,604

Yes

Yes

Yes

Yes

−0.000 (−1.54)

Note The regressand in columns (1)–(2) is FDI Indicator as in Table 11, while the regress and in columns (3)–(4) is the indicator of starting FDI as in Table 12. All columns include firm-level controls such as firm’s relative TFP, log employment and the exporting indicator. Parent industry-specific, year-specific and affiliate industry-specific fixed effects are included into all columns. The numbers in parentheses are robust t-values. *** (**,*) denotes significance at the 1% (5%, 10%) level

Yes

No

Yes

Destination FEs

Year FEs

Yes Yes

Yes

Yes

Parent industry FEs

−0.000*** (−3.96)

−0.000*** (−9.55)

2000–2008

FDI indicator

Affiliate industry FEs

Regressand: SOE indicator

Table 8 Regressions with destination-specific and affiliates’ industry-specific fixed effects

246 Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

No Yes

No

Yes

No

0.004*** (37.66)

0.001*** (10.26)

0.004*** (22.24)

No

No

Yes

Yes

Log firm sales

Log firm labor

Export indicatcr

Foreign firms dropped

Tax haven dropped

Year fixed effects

Firm fixed elfects

2,529,449

Observations 2,028,733

Yes

Yes

No

Yes

0.721*** (24.50)

0.268*** (17.93)

0.455*** (39.50)

−0.699*** (−6.64)

1,820,515

Yes

Yes

No

Yes

No

Yes

1.109*** (52.99)

0.142*** (14.23)

0.548*** (71.84)

−1.413*** (−16. 88)

(4)

Narrow

Rare event logit

2.028.733

Yes

Yes

No

Yes

No

Yes

0.713*** (24.45)

0.258*** (17.89)

0.443*** (39.80)

−0.695*** (−6.63)

(5)

Narrow

2.028.733

Yes

Ye

No

Yes

No

Yes

0.712*** (24.42)

0.262*** (18.06)

0.444*** (39.86)

−0.756*** (−7.99)

(6)

Broad

Complementary log–log

2,022,589

Yes

Yes

No

Yes

Yes

Yes

1.168*** (27.23)

0.276*** (13.57)

0.443*** (28.33)

−0.951*** (−6.50)

(7)

Narrow

2004–13

l,696,358

Yes

Yes

No

Yes

No

Yes

0.712*** (24.42)

0.258*** (17.91)

0.442*** (39.79)

−0.692*** (−6.58)

(8)

Narrow

2010–2013

547,719

Yes

Yes

No

Yes

No

Yes

0.367*** (13.18)

0.261*** (15.57)

0.430*** (35.94)

−0.399*** (−3.78)

(9)

Narrow

Narrow

545,306

No

Yes

No

Yes

No

Yes

0.366*** (13.12)

0.263*** (15.60)

0.430*** (35.85)

−0.425*** (−3.93)

(10)

Note The regressand is the FDI indicator. All columns except column (1) include industry dummies at the 2-digit level and year dummies. The numbers in parentheses are t-values clustered at the firm level. *** (**) denotes significance at the 1% (5%) level. Columns (1)–(2) include foreign-invested firms whereas all other columns drop those firms. Columns (1)–(7) cover data over the period of 2000–2013, whereas Column (8) cover data from 2004–2013. Columns (9)–(10) cover data over the period of 2010–2013. Column (6) uses broadly defined SOE. Column (7) drops outward FDI to tax haven destinations. Column (10) drops the switching SOEs (i.e., switchers from SOEs to private firms)

2,529,074

Yes

Yes

No

Yes

Industry fixed effects

With switching SOEs

No

0.671*** (26.63)

0.216*** (16.88)

0.431*** (44.35)

−0.541*** (−5.23)

−0.002*** −6.35

SOE indicator

(3)

(2)

(1)

Narrow

Variable

Narrow

2000–2013

Narrow

Logit

SOE defined

Logit

Year coverage

LMP

Regressand: FDI indicator

Table 9 Private firms are more likely to undertake FDI (2000–2013)

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China … 247

2004–13 2,178,413

-0.001* (−1.79) 0.684*** (87.33)

−0.000 (−1.53)

0.609*** (89.18)

0.752*** (41.55)

Yes

Yes

Yes

Yes

2000–13

2,586,369

SOE indicator × Ind. rate differential

Log firm labor

Export indicator

Year fixed effects

Industry fixed effects

Foreign firms included

Tax haven included

Year coverage

Observations

(4)

Yes

Yes

Yes

Yes

0.751*** (41.43)

0.608*** (88.81)

−0.000 (−1.52)

−0.000* (−1.95)

2,056,130

2000–13

No

No

Yes

Yes

1.212*** (42.18)

0.706*** (70.15)

−0.001** (−2.38)

0.000 (0.89)

−1.004*** (−10.02)

(5)

1,735,448

2004–13

No

No

Yes

Yes

1.217*** (42.37)

0.701*** (69.73)

−0.001** (−2.54)

0.000 (1.05)

−1.002*** (−9.96)

2,586,369

2000–13

Yes

Yes

Yes

Yes

0.753*** (41.56)

0.610*** (87.93)

−0.000 (−1.03)

−0.000** (−2.15)

−0.254*** (−5.53)

2,066,377

Yes

No

Yes

Yes

0.819*** (39.85)

0.690*** (86.04)

−0.000 (−1.54)

−0.000** (−2.39)

−0.449*** (−9.20)

(7)

2,178.,413

2004–13

Yes

Yes

Yes

Yes

0.751*** (41.43)

0.609*** (87.62)

−0.000 (−1.05)

−0.000** (−2.09)

−0.262*** (−5.65)

(8)

Note The regressand is the FDI indicator. The numbers in parentheses are t-values clustered at the firm level, *** (**) denotes sigificance at the 1% (5%) level. Columns (1)–(5) use coventional definition of the SOE indicator whereas the SOE indicator in column (6)–(8) is broadly defined as in Hsieh and Song (2015). Industry interest rate differential is measured by average industry-level interest rate paid by private firms minus that paid by SOEs in each 3-digit industry level. Columns (4) and (5) drop FDI to tax haven destinations. Columns (2), (4), (5) and (7) drop parent firms that are foreign firms. Colums (1), (2), (4), (6) and (7) cover data over 2000–13 whereas the rest of the table covers data over 2004–13. All regressions include 3-digit industry fixed-effects and year fixed-effects

2.066,377

Yes

No

Yes

Yes

0.819*** (39.90)

−0.000** (−2.43)

−0.000** (−2.00)

Industry rate differential

−0.591*** (−8.44)

−0.728*** (−10.39)

−0.599*** (−8.59)

(6)

(3)

(2)

(1)

Broad

Narrow

SOE indicator

SOE defined Regressand: FDI indicator

Table 10 Logit estimates on channels (2000–13)

248 Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China …

(3)

Yes No

−0.582*** (−4.73) 0.607*** (49.28) 0.758*** (28.10) No

SOE indicator × Capital-intensive indicator

Log firm labor

Export indicator

Foreign firm dropped

Yes No 2,080,027

Yes No 2,602,602

Industry fixed effects

SOE switching firms dropped

Observations

2,074,328

No

Yes

Yes

Yes

Yes

1.248*** (27.57)

0.696*** (38.46)

−1.164*** (−6.79)

1.747,652

No

Yes

Yes

No

Yes

0.835*** (26.37)

0.680*** (49.55)

−0.754*** (−6.03)

−0.638*** (−2.93)

(5)

1,056.652

Yes

Yes

Yes

No

Yes

1.016*** (26.47)

0.738*** (47.31)

−0.995*** (−5.57)

−0.598* (−1.81)

Note The regressand is the FDI indicator. All columns include industry dummies at the 2-digit level and year dummies. The numbers in parentheses are t-values clustered at the firm level. *** (**) denotes sigificance at the 1 rercent (5%) level. Columns (1)–(3) cover observations CXS during years 2000–13 whereas columns (4)–(5) cover observations during years 2004–13. Columns (1) keeps foreign invested firms whereas the other columns drop foreign invested firms. Columns (3) drops outward FDI to tax-haven regions. Columns (5) drops SOE switching firms. Labor intensive sectors indicator equals one if firm’s Chinese industrial classification is higher than 20 and zero otherwise

Yes

No Yes

Tax haven destinations dropped

Year fixed effects

0.836*** (26.40)

0.681*** (49.59)

−0.753*** (−6.07)

−0.443 (−1.50)

(4)

(2) −0.638*** (−2.97)

(1) −0.567*** (−2.61)

Regressand: FDI indicator

2004–13

2000–13

SOE indicator × Labor-intensive indicator

Sectoral category:

Table 11 Logit estimates by sectors (2000–13)

Appendix A: For “Distribution, Outward FDI, and Productivity Heterogeneity: China … 249

Bibilography

Aldrich, H. (1999). Organizations evolving. Sage. Amiti, M., & Weinstein, D. E. (2009). Exports and financial shocks. Quarterly Journal of Economics, 126(4), 1841–1877. Andersson, U., Cuervo-Cazurra, A., & Nielsen, B. B. (2020). Explaining interaction effects within and across levels of analysis. In Research methods in international business (pp. 331–349). Palgrave Macmillan, Cham. Arkolakis, C., Ganapati, S., & Muendler, M. A. (2021). The extensive margin of exporting products: A firm-level analysis. American Economic Journal: Macroeconomics, 13(4), 182–245. Aulakh, P. S., & Kotabe, M. (1997). Antecedents and performance implications of channel integration in foreign markets. Journal of International Business Studies, 28(1), 145–175. Brandt, L., Van Biesebroeck, J., & Zhang, Y. (2012). Creative accounting or creative destruction? Firm-level productivity growth in Chinese manufacturing. Journal of Development Economics, 97(2), 339–351. Bernard, A. B., Eaton, J., Jensen, J. B., & Kortum, S. (2003). Plants and productivity in international trade. American Economic Review, 93(4), 1268–1290. Bhattacharya, R., Patnaik, I., & Shah, A. (2010). Productivity and outbound FDI in software services: A reversal of the HMY model. NIPFP-DEA Research Program. Birkinshaw, J., Bresman, H., & Nobel, R. (2010). Knowledge transfer in international acquisitions: A retrospective. Journal of International Business Studies, 41(1), 21–26. Blonigen, B. A. (2005). A review of the empirical literature on FDI determinants. Atlantic Economic Journal, 33(4), 383–403. Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics, 87(1), 115–143. Blundell, R., & Dias, M. C. (2009). Alternative approaches to evaluation in empirical microeconomics. Journal of Human Resources, 44(3), 565–640. Brouthers, K. D. (2002). Institutional, cultural and transaction cost influences on entry mode choice and performance. Journal of International Business Studies, 33(2), 203–221. Brouthers, K. D., & Hennart, J. F. (2007). Boundaries of the firm: Insights from international entry mode research. Journal of Management, 33(3), 395–425. Buckley, P. J., & Casson, M. (1976). The future of the multinational enterprise (Vol. 1). Macmillan London. Ca Ntwell, J. (1989). Technological innovation and multinational corporations. The Economic Journal, 100(401). Chan, C. M., & Makino, S. (2007). Legitimacy and multi-level institutional environments: Implications for foreign subsidiary ownership structure. Journal of International Business Studies, 38(4), 621–638. © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022 W. Tian and M. Yu, Outward Foreign Direct Investment of Chinese Enterprises, Contributions to Economics, https://doi.org/10.1007/978-981-19-4719-3

251

252

Bibilography

Chen, F., Hope, O. K., Li, Q., & Wang, X. (2011). Financial reporting quality and investment efficiency of private firms in emerging markets. The Accounting Review, 86(4), 1255–1288. Chen, W., Dollar, D., & Tang, H. (2018). Why is China investing in Africa? Evidence from the firm level. The World Bank Economic Review, 32(3), 610–632. Cheng, L. K., & Ma, Z. (2010). China’s outward foreign direct investment. China’s Growing Role in World Trade, 545–578. Child, J., & Rodrigues, S. B. (2005). The internationalization of Chinese firms: A case for theoretical extension? 1. Management and Organization Review, 1(3), 381–410. Conconi, P., Sapir, A., & Zanardi, M. (2016). The internationalization process of firms: From exports to FDI. Journal of International Economics, 99, 16–30. Damijan, J. P., Glaézar, M., Prašnikar, J., & Polanec, S. (2004). Export vs. FDI behavior of heterogenous firms in heterogenous markets: Evidence from Sovenia (LICOS Discussion Paper No. 147) De Loecker, J., Goldberg, P. K., Khandelwal, A. K., & Pavcnik, N. (2016). Prices, markups, and trade reform. Econometrica, 84(2), 445–510. Delgado, M. A., Farinas, J. C., & Ruano, S. (2002). Firm productivity and export markets: A non-parametric approach. Journal of International Economics, 57(2), 397–422. Dunning, J. H. (2015). The eclectic paradigm of international production: A restatement and some possible extensions. The Eclectic Paradigm, 50–84. Dunning, J. H., & Lundan, S. M. (2008). Institutions and the OLI paradigm of the multinational enterprise. Asia Pacific Journal of Management, 25(4), 573–593. Ekeledo, I., & Sivakumar, K. (2004). International market entry mode strategies of manufacturing firms and service firms: A resource-based perspective. International Marketing Review. Fazzari, S., Hubbard, R. G., & Petersen, B. C. (1987). Financing constraints and corporate investment. Brookings Papers on Economic Activity, 1, 141–195. Federico, S. (2008). Outsourcing versus integration at home and abroad. In 2nd FIW-Research Conference “International Economics”, Vienna. Felipe, J., Hasan, R., & McCombie, J. S. L. (2008). Correcting for biases when estimating production functions: An illusion of the laws of algebra? Cambridge Journal of Economics, 32(3), 441–459. Frost, T. S. (1997). Imitation to innovation: The dynamics of Korea’s technological learning. Gandhi, A., Navarro, S., & Rivers, D. (2011). On the identification of production functions: How heterogeneous is productivity? (CIBC Working Paper No. 2011-9). Garetto, S., Oldenski, L., & Ramondo, N. (2017). Life-cycle dynamics and the expansion strategies of US multinational firms. Görg, H. (2000). Analysing foreign market entry—The choice between greenfield investment and acquisitions. Journal of Economic Studies. Griffith, R., Huergo, E., Mairesse, J., & Peters, B. (2006). Innovation and productivity across four European countries. Oxford Review of Economic Policy, 22(4), 483–498. Grossman, G. M., Helpman, E., & Szeidl, A. (2006). Optimal integration strategies for the multinational firm. Journal of International Economics, 70(1), 216–238. Halpern, L., Koren, M., & Szeidl, A. (2015). Imported inputs and productivity. American Economic Review, 105(12), 3660–3703. Head, K., & Ries, J. (2004). Exporting and FDI as alternative strategies. Oxford Review of Economic Policy, 20(3), 409–423. Insead, L. C., & Chatain, O. (2008). Competitors’ resource-oriented strategies: Acting on competitors’ resources through interventions in factor markets and political markets. Academy of Management Review, 33(1), 97–121. Ishikawa, J., Morita, H., & Mukunoki, H. (2016). Trade liberalization and aftermarket services for imports. Economic Theory, 62(4), 719–764. Jefferson, G. H., Rawski, T. G., & Zhang, Y. (2008). Productivity growth and convergence across China’s industrial economy. Journal of Chinese Economic and Business Studies, 6(2), 121–140. Jovanovic, B. (1982). Selection and the evolution of industry. Econometrica: Journal of the Econometric Society, 649–670.

Bibilography

253

Khanna, T., & Rivkin, J. W. (2001). Estimating the performance effects of business groups in emerging markets. Strategic Management Journal, 22(1), 45–74. Kim, Y., & Gray, S. J. (2008). The impact of entry mode choice on foreign affiliate performance: The case of foreign MNEs in South Korea. Management International Review, 48(2), 165–188. King, G., & Zeng, L. (1999). Estimating absolute, relative, and attributable risks in case-control studies. Harvard University. Lecraw, D. J. (1993). Outward direct investment by Indonesian firms: Motivation and effects. Journal of International Business Studies, 24(3), 589–600. Li, L., Liu, X., Yuan, D., & Yu, M. (2017). Does outward FDI generate higher productivity for emerging economy MNEs?–Micro-level evidence from Chinese manufacturing firms. International Business Review, 26(5), 839–854. Li, W., He, A., Lan, H., & Yiu, D. (2012). Political connections and corporate diversification in emerging economies: Evidence from China. Asia Pacific Journal of Management, 29(3), 799–818. Li, Z., & Yu, M. (2009). Exports, productivity, and credit constraints: A firm-level empirical investigation of China. Available at SSRN 1461399. Linder, S. B. (1961). An essay on trade and transformation (pp. 82–109). Almqvist & Wiksell. Liu, L. (2005). China’s industrial policies and the global business revolution: The case of the domestic appliance industry. Routledge. Markusen, J. R. (2002). Multinational firms and the theory of international trade. MIT press. Masso, J., & Vahter, P. (2008). Technological innovation and productivity in late-transition Estonia: Econometric evidence from innovation surveys. The European Journal of Development Research, 20(2), 240–261. Meyer, K. E., Ding, Y., Li, J., & Zhang, H. (2014). Overcoming distrust: How state-owned enterprises adapt their foreign entries to institutional pressures abroad. Journal of International Business Studies, 45(8), 1005–1028. Morck, R., & Yeung, B. (1991). Why investors value multinationality. Journal of Business, 165–187. Morck, R., Yeung, B., & Zhao, M. (2008). Perspectives on China’s outward foreign direct investment. Journal of International Business Studies, 39(3), 337–350. Muûls, M. (2008). Exporters and credit constraints. A firm-level approach (NBB Working Paper No. 139). Pakes, A. (1994). Advances in econometrics. Cambridge University Press. Powel, W. W. (1990). Neither market nor hierarchy: Network forms of organization. Research in Organizational Behavior, 12(3), 295–336. Priem, R. L., & Butler, J. E. (2001). Is the resource-based “view” a useful perspective for strategic management research? Academy of Management Review, 26(1), 22–40. Rauch, J. (2016). Global production: Firms, contracts, and trade structure. CREI Lectures in Macroeconomics Series, 595–598. Schmookler, J. (2006). The level of inventive activity’. Review of Economics and Statistics, 36(2), May, 183–90. International Library of Critical Writings in Economics, 197(1), 174. Shaver, J. M. (1998). Accounting for endogeneity when assessing strategy performance: Does entry mode choice affect FDI survival? Management Science, 44(4), 571–585. Shen, X. (2015). Private Chinese investment in Africa: Myths and realities. Development Policy Review, 33(1), 83–106. Taylor, I., & Smith, K. (2007). United Nations Conference on Trade and Development (UNCTAD). Routledge. Tian, W., & Yu, M. (2020). Distribution, outward FDI, and productivity heterogeneity: China and cross-countries’ evidence. Journal of International Financial Markets, Institutions and Money, 67, 101218. Tian, W., Sheng, L., & Zhao, H. (2016). State capitalism: A new perspective on land sale in China. Pacific Economic Review, 21(1), 84–101. United Nations Conference on Trade and Development. (2013). In World investment report 2013: Global value chains: Investment and trade for development. UN.

254

Bibilography

Vermeulen, F., & Barkema, H. (2001). Learning through acquisitions. Academy of Management Journal, 44(3), 457–476. Wang, J., Liu, X., Wei, Y., & Wang, C. (2014). Cultural proximity and local firms’ catch up with multinational enterprises. World Development, 60, 1–13. Wesson, T. (2000). Foreign direct investment and competitive advantage. Elgar. Witt, M. A., & Lewin, A. Y. (2007). Outward foreign direct investment as escape response to home country institutional constraints. Journal of International Business Studies, 38(4), 579–594. Yiu, D. W., Lau, C., & Bruton, G. D. (2007). International venturing by emerging economy firms: The effects of firm capabilities, home country networks, and corporate entrepreneurship. Journal of International Business Studies, 38(4), 519–540. Zhang, F., & Yu, M. (2015). The potential impacts of BIT on China’s manufacturing industries. Toward a Sino–US Investment Treaty, 67–90. Zhou, N., & Delios, A. (2012). Diversification and diffusion: A social networks and institutional perspective. Asia Pacific Journal of Management, 29(3), 773–798.