History versus Expectations: an Empirical Investigation

This paper provides the first empirical test of the role of history versus expectations in U.S. urban development. Starting from Paul Krugman's theoretical work in new economic geography, we test whether or not a modern city develops because of either advantageous initial conditions or by way of a self-fulfilling prophecy based on expectations of development. Using the methology developed by Granger to establish causality between two variables, but adapted to a cross-section with four time lags, we test whether asset values, that is, farmland values and housing values, anticipate urban development or vice versa. In the case of the former, we would conclude that expectations drive urban development in the U.S., and in the case of the latter we would conclude that history does. The results indicate very strongly that initial conditions, that is history, dominate the process by which one city becomes a metropolis and another languishes in the periphery.