A General Model for the Distribution of Public Investment among Jurisdictions

This article proposes a comprehensive theoretical framework to bridge the gap between the theoretical and empirical literatures regarding the criteria followed by a policymaker to distribute public investment across jurisdictions. A social welfare function that considers, not only output per capita, but also output density, is proposed. We derive the optimal levels of public investment as a share of the total public capital stock, which turns out to be a function of the main criteria suggested by the previous empirical literature. A tractable equation for empirical and policy implementation is obtained. Furthermore, the model allows us to cast theoretical predictions to be tested empirically. For this purpose, we use data on the Spanish regions (autonomous communities) over the 1986-2010 period. Our findings show that most of the coefficients of the conventional criteria are significant and with the expected signs. In addition, the results provide the weights of the output measures in the social welfare function, and strongly suggest that output density should not be neglected, as has been often done in the previous literature. The results are quite robust across estimation methods, different alternate dependent variables and the inclusion of control variables.