Competition, Innovation, and Regulation: Accounting for Productivity Differences

• Main Author: Bento, Pedro
• Other Authors: Restuccia, Diego
• Language: en_ca
• Published: 2013
• Subjects: macroeconomics; competition; regulation; innovation; 0501
• Online Access: http://hdl.handle.net/1807/43484

The relationships between competition, innovation, and regulation have long been studied in an attempt to understand and evaluate the effect of regulation on the wealth and growth of nations. Recent empirical work has emerged taking advantage of the still ongoing proliferation of ever more disaggregated data to shed more light on these relationships and at the same time uncover new puzzles in need of explanations. This thesis is an attempt to address the discrepancies between some of these newly discovered phenomena and current theory. In Chapter 1 I introduce an insight of Friedrich Hayek - that competition allows a thousand flowers to bloom, and discovers the best among them - into a conventional model of Schumpeterian innovation. I show how the model can account for two seemingly contradictory empirical phenomena, a positive relationship between competition and industry-level productivity growth, and an inverted-U relationship between competition and firm-level innovation. In Chapter 2 I extend the model to investigate the effects of patent protection on competition and innovation, and to understand the interaction between patent policy and product-market regulation. I calibrate the model to show that patent protection in the U.S. is depressing competition, innovation, growth, and welfare. Using patent and citation data, I further provide empirical evidence supporting the implications of the model. In Chapter 3 I investigate the impact of regulatory entry barriers to new firms on aggregate output and total factor productivity. Following recent work by Thomas J. Holmes and John J. Stevens, I extend a standard model of monopolistic competition to account for the existence of both niche markets and mass markets within industries. Calibrating the model using U.S. manufacturing data, I show this extension goes a long way towards explaining the large gap between empirical estimates of the impact of barriers to entry and the quantitative predictions of current models.