This paper provides an integrative analysis of the drivers of vertical scope, using analytical and computational methods. I propose a model with two vertical segments (upstream and downstream), firm populations with heterogeneous capabilities, and an intermediate market subject to transaction costs in which firms can choose whether to be integrated or vertically specialized. By varying the level of transaction costs and changing the structure of the correlation between upstream and downstream capabilities in the industry, learning curves, and the way in which profitability leads to capability improvement in the upstream and downstream segments, I generate numerical results to explain how vertical integration evolves over time. The results suggest that (a) with no capability differences, even if transaction costs are nil, firms remain integrated; (b) transaction costs catalyze the underlying capability differences to drive scope; (c) dynamic factors, such as learning curves, returns to investment in capabilities, or limits to expansion, exacerbate small, random capability differences and as such promote specialization. These dynamic factors can by themselves lead to substantial specialization when they differ between the upstream and downstream segments. The model also provides a rationale for mixed governance (i.e., concurrent use of both the market and integration), as well as for the initial period of vertical integration, followed by specialization.
