Manuel Arellano

Revisiting the Relationship between Competition, Patenting and Innovation

A thought-provoking book by Boldrin and Levine (2009) argues that patents are always detrimental to competition and thereby to innovation. To provide support to their analysis these two authors built a growth model where innovation and growth can occur under perfect competition. The model is then used to argue that monopoly rents and therefore patents are not needed for innovation and growth: on the contrary, patents are detrimental to innovation because they reduce competition. That reducing competition can be detrimental to innovation, is a sound idea which could not be accounted for by the new growth models of Romer (1990) and Aghion and Howitt(1992). In these models, competition is detrimental for innovation and growth for exactly the same reason for why patent protection (IPRs) is good for innovation: namely because competition reduces (post innovation) rents whereas patent protection increases these rents. In this note we argue that this latter conclusion is not robust to considering more elaborated models of competition and innovation. In particular we show below that in a step-by-step innovation model, where a laggard firm needs to catch up with the current leader in its sector (and therefore go through a neck-and-neck stage) before it could later become a leader itself, not only can competition enhance innovation as in Boldrin and Levine’s model, but also and perhaps more importantly competition and IPRs become complementary policies. Why? Because the incentive to innovate depends on the gap between the post-innovation rent and the pre-innovation rent, call it the net innovation rent. And typically, what competition does is to lower pre-innovation rents, and also maybe post-innovation rents, although the difference between postand preinnovation rents may still increase with competition, and all the more so with stronger patents to protect post-innovation rents. In contrast, in our earlier Schumpeterian models where innovations are made by outsiders who then leapfrog incumbent firms, the pre-innovation rent is always equal to zero, thus all ∗Harvard University †Brown University ‡University of Cologne

Sample Selection in Quantile Regression: A Survey

Nonrandom sample selection is a pervasive issue in applied work. In additive models, a number of techniques are available for consistent selection correction. However, progress in the development of non-additive selection corrections has been slower. In this survey we review recent proposals dealing with sample selection in quantile models.