On the Economics of the Direction of Research

This dissertation considers the problem of research lines. Of particular interest are technological areas with many research projects available today, whereby successful invention affects the nature of research targets available tomorrow. I use theory and history to investigate the intersection of research lines with optimal patent policy, inefficient firm R&D behavior, path dependence, and the early histories of the aviation and nuclear industries.;In Section 1, joint with Jorge Lemus, we develop a novel and tractable theory of research direction. We show firms will pursue inefficient research lines because they only capture a portion of the value of downstream innovations, they only care about surplus marginal to what they receive from doing nothing, and they are incentivized to "race'' toward relatively easy projects since they do not account for how their effort affects the time at which various downstream projects become available to other firms.;In Section 2, I use a simplified version of the theory from Section 1 to examine the early history of the nuclear industry. It has long been conjectured that the dominant technology in nuclear power, light water reactors, was both an ex-ante and an ex-post inefficient research target. I consider the potential welfare benefits of counterfactually decreasing targeted subsidies to nuclear plant builders, of allowing stronger patents, and of reacting to apparent "lock-in'' of light water at GE and Westinghouse. I suggest that, in the presence of limited planner knowledge about the value and difficulty of research targets, optimal policy is either always second-best or else involves "picking winners'' in a broad sense.;In Section 3, I investigate the collapse of the early aviation industry in the United States. Despite being the home of the Wright Brothers, by World War I the US aviation industry was nearly extinct. I separate the airplane into its constituent microinventions, and show that even in 1903 the US lagged France and Germany in the development of components that would prove critical to a commercially viable airplane. I suggest that strong pioneer patents can provide early innovators an incentive to shut down sources of technology transfer.