We consider a model in which appropriate organization fosters innovation, but because of contractibility problems, this benefit cannot be internalized. The organizational design element we focus on is the division of labor, which as Adam Smith argued, facilitates invention by observers of the production pro- cess. However, entrepreneurs choose its level only to facilitate monitoring their workers. Whether there is innovation depends on the interaction of the mar- kets for labor and for inventions. A high level of specialization is chosen when the wage share is low. But low wage shares arise only when there are few en- trepreneurs, which limits the market for innovations and therefore discourages inventive activity. When there are many entrepreneurs, the innovation market is large, but the rate of invention is low because there is little specialization. Rapid technological progress therefore requires a balance between these oppos- ing effects, which occurs with a moderate relative scarcity of entrepreneurs and workers. In a dynamic version of the model in which a credit constraint limits entry into entrepreneurship, this relative scarcity depends on the wealth distri- bution, which evolves endogenously. There is an inverted-U relation between growth rates driven by innovation and the level of inequality. Institutional im- provements have ambiguous effects on growth. In light of the model, we offer a reassessment of the mechanism by which organizational innovations such as the factory may have spawned the industrial revolution.