Innovation, avantages concurrentiels et concurrence. Une analyse expérimentale des incitations à innover et de l'efficacité des marchés en présence de chocs endogènes

We present a series of laboratory experiments run in order to analyse cost-reduction innovation behaviour in competitive double auction markets. The experiments allow us to study both efficiency of markets in which the process of costs improvement is endogenous, and investment behaviours in R?D in such an environment. The general design of the experiment is based on the repetition of ten periods, each of them consisting of two stages : (1) a market stage and (2) an investment stage. During the market stage, four producers have the opportunity to exchange with four buyers some units of a homogenous product. The value of each unit depends on the private costs or redemption value schedules of respectively each seller and each buyer. The demand is stable. The supply shifts according to the individual cost-reduction innovations made during the first stage of each period. Actually, during the investment stage, every producers decide individually and simultaneously how much to invest in R?D. Producers gain or not a costs reduction, depending on a common knowledge function of process innovation. Innovations are not exclusive and additive from period to period that means that the gains of productivity by one seller are cumulative over the experimental session. We analyse the anticipations of market prices by sellers and their impact on investment behaviours. The experimental treatments allow us to study the effect of three variables on market efficiency and incentives to innovate. These variables are, (i) the technological uncertainty, (ii) the diffusion of innovation and (iii) the information about the R?D function and process. The main results are the following. The competitive markets are globally efficient with endogenous process innovations. However, the speed of convergence to the competitive equilibrium is reduced in periods of chocks on the supply curve, particularly when buyers ignore the R?D process at work, and when subjects have experience. Producers continually invest in R?D and innovate. They do not efficiently co-ordinate their investments in R?D. This failure is wide when investments are certain and when innovations are diffused. However the diffusion of innovations do not produce the expected free-rider behaviour. The technological leaders are the only profitable producers of the market competition. Finally, according to a dynamic program, we show the producers not optimally invest in R?D in a social and individual point of view.