Experimentation in Dynamic R&D Competition

Abstract

This paper studies a two-stage R&D race, in which, at the outset, firms are uncertain regarding the viability of the project. Learning through experimentation introduces a bilateral (dynamic) feedback mechanism. For relatively low-value products, the equilibrium cutoff belief coincides with the socially efficient cutoff belief, although firms might experiment excessively in equilibrium; for relatively high-value products, firms might reduce experimentation and stop rather prematurely due to the fundamental free-riding effect. Perhaps surprisingly, a decrease in the value of the product can spur experimentation.

Keywords: Experimentation; Learning, Dynamic R&D Competition; Inefficiency

JEL Classification: C73; D83; O31; O32