A Note on Partial Vertical Integration

Empirical analysis [see, e.g., McDonald (1985)] suggests that partial vertical integration is a widespread phenomenon. However, there are hardly any theoretical explanations for this composition of industries. Both Arrow (1975) and Green (1986) explicitly allow for partial integration to occur. The former author takes the need for information to be the driving force behind vertical integration, whereas in the latter paper market advantages due to integration versus decreasing returns to an expansion of the scope of activities carried out by a single firm are the opposing forces. In both models, however, the industry equilibrium tends to either no integration at all, or complete vertical integration. Carlton (1979) obtained an intermediate outcome by having risk averse buyers secure their high probability demand. In the model of Quirmbach (1986), scale and/or substitution effects are responsible for partial vertical integration. If a firm switches from market exchange to vertical integration, then this will change the probability distribution of the spot market price. This externality is responsible for partial vertical integration in the model of Hendrikse (1988). The present note is in the spirit of Green (1986). However, we will obtain partial vertical integration as an equilibrium market structure, mainly by taking the risk attitudes of the traders into consideration.