This paper presents evidence supporting endogenous growth models that emphasize public structural capital. We apply a simple test of endogenous vs. exogenous growth models. In exogenous growth economies temporary innovations to policy variables lead only to temporary changes in GNP levels, while in endogenous growth economies the innovations can lead to permanent changes in GNP levels. Of the seven U.S. policy variables we examine, only non-military equipment capital and non-military structural capital have a statistically and economically significant effect upon long-run GNP levels. Further estimation suggests that the non-military equipment capital result is not robust and that several disaggregate components of structural capital contribute significantly.