Persistent heterogeneity and sustainable innovation

One of the concerns with strategy as a field is that its goal (durable profits) appears at odds with that of microeconomics (homogeneous firms with zero profits). Thus even if we can find sources of durable profits, maybe we shouldn't - because doing so comes at the expense of social welfare. In this sense, strategy is an 'economic bad.' Probably the most vivid evidence of this tension is the fact that Porter's Competitive Strategy (1982) is Industrial Organization turned on its head - exploiting indications of anti-competitive behavior to create prescriptions for durable profits. This study presents a parallax view in which the field of strategy is an economic good. In this view, heterogeneity stimulates innovation and economic growth. The underlying logic is that heterogeneity fuels diffusion; diffusion erodes leaders' shares; the loss of shares stimulates innovation, which in turn fuels new diffusion. I develop this view through a simple model of firms embedded in an industry, and then evaluate the model via simulation. The key issue for the model is whether there are steady states that sustain heterogeneity and innovation. Results indicate that heterogeneity persists in roughly half the industry conditions I examined. Furthermore innovation and growth are sustained under those conditions. What is comforting about the set of conditions is that they accommodate (and indeed require) all the things we hold dear: industrial policy, and strategic as well as routine behavior by firms. The new insights are as follows: (1) heterogeneity fuels growth (thus 'strategy' is an economic good, rather than anathema to allocative efficiency); (2) persistent heterogeneity is feasible even in the presence of spillovers (without isolating mechanisms); (3) but this heterogeneity (while persistent) requires managers to actively preserve the inherent value of their resource advantages - firms can't rest on their laurels; (4) policy prescriptions differ from those of prior innovation models.