Endogenous growth requires that non-reproducible factors of production be either augmented or eliminated. Attention heretofore has focused almost exclusively on augmentation. In contrast, we study factor elimination. Maximizing agents decide when to reduce the importance of non-reproducible factors. We use a Cobb-Douglas production function with two factors of production, one reproducible ("capital") and one not ("labor"). There is no augmenting progress of any kind, thus excluding the standard engine of growth. What is new is the possibility of changing factor intensities endogenously by spending resources on R&D. The economy starts with no capital and no knowledge of how to use it. By conducting R&D, the economy learns new technologies that use capital, which then is built. There are two possible ultimate outcomes: the economy may achieve perpetual growth, or it may stagnate with no growth. The rst outcome is an asymptotic version of the AK model of endogenous growth, and the second outcome is the standard Solow model in the absence of any exogenous sources of growth. Which outcome is achieved depends on parameter values of saving and production, and there always is a feasible saving rate that will give the perpetual growth outcome. The model thus provides a theory of the endogenous emergence of a production technology with constant returns to the reproducible factors, that is, one that is capable of supporting perpetual economic growth. The model also allows derivation of the full transition dynamics, which have interesting properties. One especially notable feature is that the origin is not a steady state. An economy that starts with pure labor production becomes industrialized through its own e¤orts. The theory thus o¤ers a purely endogenous explanation for the transition from a primitive to a developed economy, in contrast to several well-known theories. Several aspects of the transition paths accord with the evidence, suggesting that the theory is reasonable. In contrast to almost all the existing endogenous growth literature, neither monopoly power nor an externality is a necessary condition for endogenous growth. It is su¢ cient that rms be able to appropriate the results of their research and development e¤orts.
