PATENT POLICY, R&D AND ECONOMIC GROWTH by Chi Ho Angus Chu A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Economics) in The University of Michigan 2008 Doctoral Committee: Associate Professor Dmitriy L. Stolyarov, Chair Professor John P. Laitner Assistant Professor Christopher L. House Assistant Professor Rosemarie H. Ziedonis © Chi Ho Angus Chu 2008 DEDICATION In memory of mom and grandma ii ACKNOWLEDGMENTS I am especially thankful and truly indebted to Dmitriy Stolyarov for his guidance and patience. I am grateful to Rudi Bachmann, Chris House, John Laitner, Linda Tesar and Rosemarie Ziedonis for invaluable advice and insightful comments and to Bob Barsky, Juin-Jen Chang, Benjamin Chiao, Oli Coibion, Kathryn Dominquez, Yuriy Gorodnichenko, Osborne Jackson, Isao Kamata, Owen Kearney, Miles Kimball, John Leahy, David Levine, Chang-Ching Lin, Stephan Lindner, Zoe Mclaren, Ryan Michaels, Matthew Shapiro, Doug Smith, Susan Yang, Jing Zhang and seminar participants at the University of Michigan, Academia Sinica and the Canadian Economics Association 2007 Conference for very helpful suggestions. For the chapter “Special Interest Politics and Intellectual Property Rights” that has been accepted for publication at Economics & Politics, I am grateful to Devashish Mitra (the Editor) and the anonymous referee for insightful and detailed comments that have greatly improved the manuscript, and to Robert Barro for his generous advice to me regarding journal submission. iii TABLE OF CONTENTS DEDICATION ii ACKNOWLEDGMENTS iii LIST OF APPENDICES v LIST OF ABBREVIATIONS vi CHAPTER I. Introduction 1 II. Quantifying the Effects of Blocking Patents on R&D 3 III. Quantifying the Effects of Patent Length on R&D 38 IV. Special Interest Politics and Intellectual Property Rights 63 V. Conclusion 94 APPENDICES 95 BIBLIOGRAPHY 142 iv LIST OF APPENDICES Appendix A. Proofs for Chapter II 96 B. Transition Dynamics for Chapter II 104 C. The Social Rate of Return to R&D for Chapter II 107 D. Tables and Figures for Chapter II 108 E. Proofs for Chapter III 123 F. The Social Rate of Return to R&D for Chapter III 130 G. Tables and Figures for Chapter III 131 H. Proofs for Chapter IV 134 I. Tables and Figures for Chapter IV 140 v LIST OF ABBREVIATIONS DGE Dynamic General Equilibrium FDA Food and Drug Administration GDP Gross Domestic Product NIHCM National Institute for Health Care Management PAC Political Action Committee PhRMA Pharmaceutical Research and Manufacturers of America R&D Research & Development SIC Standard Industrial Classification SIG Special Interest Group TFP Total Factor Productivity TRIPS Trade-Related Aspects of Intellectual Property Rights US United States vi CHAPTER I Introduction The literature on R&D-driven economic growth suggests that technological innovations result from entrepreneurial activities such as R&D investments. In the next two chapters of this dissertation, I develop tractable frameworks to characterize and quantitatively evaluate the effects of patent policy on R&D and economic growth. As Acemoglu (p. 1112, 2007) writes, “… we lack a framework similar to that used for the analysis of the effects of capital and labor income taxes and indirect taxes in public finance, which we could use to analyze the effects… of intellectual property right polices… on innovation and economic growth.” In Chapter IV, I analyze the political economy of patent policy. What are the effects of blocking patents on R&D and consumption? To answer this question, in Chapter II, I develop a generalized quality-ladder growth model with overlapping intellectual property rights to quantify the inefficiency in the patent system. The analysis focuses on two policy variables: (a) patent breadth that determines the total profit received by a patent pool; and (b) the profit-sharing rule that determines the distribution of surplus between innovators. To quantify the inefficiency arising from blocking patents that are generated by these two policy variables, the model is calibrated to aggregate data of the US economy. Under parameter values that match key features of the US economy and show equilibrium R&D underinvestment, I find that eliminating blocking patents would lead to a conservatively estimated increase in R&D of 12% and long-run consumption of 4% per year. I also quantify the 1 transition-dynamic effects of patent policy and show implications that are different from previous studies in important ways. Is the patent length an effective policy instrument in stimulating R&D? To answer this question, in Chapter III, I develop a generalized variety-expanding growth model and calibrate the model to aggregate data of the US economy to quantify the effects of patent extension. At the empirical patent-value depreciation rates, extending the patent length beyond 20 years leads to a negligible increase in R&D despite equilibrium R&D underinvestment. On the other hand, shortening the patent length can lead to a significant reduction in R&D and consumption. The calibration exercise also suggests that about 35% to 45% of the long-run TFP growth in the US is driven by R&D. Finally, I identify and analytically derive a dynamic distortion of patent policy on saving and investment in physical capital that has been neglected by previous studies, which consequently underestimate the distortionary effects of patent protection. What are the welfare implications of pharmaceutical lobbying? Since the 80’s, the pharmaceutical industry has benefited substantially from the strengthening of patent protection for brand-name drugs as a result of the industry’s political influence. To analyze this phenomenon, in Chapter IV, I incorporate special interest politics into a quality-ladder growth model to consider the policymakers’ tradeoff between the socially optimal patent length and campaign contributions. The welfare analysis suggests that the presence of a pharmaceutical lobby distorting patent protection is socially undesirable in a closed-economy setting but may improve global welfare in a multi-country setting, which features an additional efficiency tradeoff between monopolistic distortion and international free-riding on innovations. Finally, if the SIGs have asymmetric influences across countries, then the country, in which the government places a higher value on campaign contributions, would gain by less or even suffer a welfare loss. 2 CHAPTER II Quantifying the Effects of Blocking Patents on R&D “Today, most basic and applied researchers are effectively standing on top of a huge pyramid… Of course, a pyramid can rise to far greater heights than could any one person... But what happens if, in order to scale the pyramid and place a new block on the top, a researcher must gain the permission of each person who previously placed a block in the pyramid, perhaps paying a royalty or tax to gain such permission? Would this system of intellectual property rights slow down the construction of the pyramid or limit its heights? … To complete the analogy, blocking patents play the role of the pyramid’s building blocks.” – Carl Shapiro (2001) 1. Introduction What are the effects of blocking patents on R&D? In an environment with cumulative innovations, the scope of a patent (i.e. patent breadth) determines the level of patent protection for an invention against imitation and subsequent innovations. This latter form of patent protection, which is known as leading breadth in the literature, gives the patentholders property rights over future inventions, and the resulting overlapping intellectual property rights may dampen the incentives for R&D. This phenomenon is referred to as blocking patents. The main contribution of this chapter is to develop an R&D-driven endogenous growth model to quantify this inefficiency in the patent system. To the best of my knowledge, this chapter is the first to perform a quantitative analysis on patent policy by calibrating a DGE model combining the following features: (a) overlapping intellectual property rights that are emphasized by the patent-design literature; (b) multiple R&D externalities that are commonly discussed in the growth literature; and (c) endogenous capital accumulation that leads to a dynamic distortionary 3