Recovering the Sunk Costs of R&D: The Moulds Industry Case

Even in narrowly defined industries R&D firms coexist with non-R&D firms. In this paper I investigate the existence of sunk costs which rationalize the behavior observed in the data. I develop a dynamic competition model where firms can decide to invest on physical capital and R&D. When firms have incomplete information about the state of their competitors, the industry state can be summarized by the aggregate (payoff relevant) state. The advantages are twofold:(i) it avoids the curse of dimensionality which occurs when there are either several firms in the market or several states per firm and; (ii) it allows me to deal with firms unobserved in the data. I apply the model to recover the sunk costs of R&D for the Portuguese Moulds Industry. The sunk costs are estimated at 1.7 times the yearly average firm sales and they rationalize actual behavior observed in the data. Finally, I evaluate the impact of a 25% reduction in the sunk costs of R&D. In equilibrium, average productivity increases by 11% and average capital stock by 18%. This corroborates the idea that sunk costs of R&D play an important role and have implications for the targeting of R&D subsidies. In the presence of large R&D sunk costs, subsidies should promote R&D start-ups. 

Enrico Pennings