Entry, spending and firm size in a stochastic R&D race

A model is developed to explain participation and spending on R&D as a function of firm size. The R&D process is represented as an n-participant race with a Poisson incidence of success, where the winner takes all during some "protection period". Four effects of scale are taken into account: a sunk fixed threshold cost of entry; a flow cost of expenditure for the duration of the race, which affects both the profitability of winning and the speed of development (the Poisson parameter), both with diminishing returns; allowance for an effect of firm size on the effectiveness (profit/cost) of development. The operational decision concerning the level and intensity of commitment in case of participation is modelled in a traditional fashion as the maximization of expected returns. The strategic decision whether or not to participate (at an optimal level and intensity) is modelled as a stochastic process of deliberation between different makers and influencers of decisions in the firm. The latter is to be seen as an introduction of the "political" and "resource dependence" views of organisations. The resulting model of R&D participation as a function of firm size is tested empirically on data from an R&D survey in the Netherlands.