Lately several authors have tried to explain the emergence of Stackelberg leadership as an endogenous feature of duopoly models. This article describes two models where assumptions of incomplete information about the rival firm’s marginal cost of production creates incentives for duopolistic firms to end up in a sequential choice situation. In the first model the duopolists can commit themselves to be a Stackelberg leader or follower at the time when they know the distributions, but not the actual values, of their own and the rival’s cost. The central result from this model is that under quantity (but not price) competition the duopolists may agree on who should be the leader and who should be the follower. In the second model the firms cannot commit to a specific role. In the quantity setting version of this model they must decide both when and how much to produce. It turns out that there can never be an equilibrium where they both always produce in the same period. However, there are equilibria where one firm produces in the first period and the other in the second, i.e. Stackelberg-type equilibria. Similar results are found in a price setting version of the model.

The finiteness condition of vertical product differentiation models is translated into the taste distribution model first analyzed by Mussa and Rosen. For a utility function linear in quality, the necessary and sufficient condition for finiteness is that the cost function with respect to quality is strictly concave. Furthermore, for these cost functions, in duopoly, higher quality always implies a higher market share at the Nash equilibrium in prices. The n-firm case is briefly discussed, and some implications for marketing strategy of new products are presented.

Cette étude se place dans la lignée de modèles qui combinent externalités et structures de marché imparfaite. Sa spécificité provient de l’introduction d’externalités de production dans un duopole de Cournot. Une seule firme est cause de la pollution, mais les effets de celle-ci affectent les performances des deux firmes, tout en se cantonnant au sein de l’industrie. Deux moyens de corrections des externalités sont introduits séparément dans le modèle: la taxation des effluents, d’une part, et leur traitement dans une installation appropriée, d’autre part. Les conséquences de l’introduction de ces moyens de correction sur l’équilibre de Cournot sont alors étudiées dans chaque cas.

This study explores the impact of non-constant returns to scale and imperfectly competitive product markets on the empirical performance of the q-model of investment. We present an econometric specification linking the unobserved shadow price of capital to financial market data when returns-to-scale are not necessarily constant and product markets are not necessarily competitive. Furthermore, unlike most Q-studies, we work within a discrete-time framework, and give explicit attention to the relations between inventories, sales, and production. Our new specification allows us to recover parameters characterizing returns-to-scale, product market, and adjustment cost elasticities. Estimates are reported for 12 mature industries in the United States drawn from VALUELINE.