Interstate spillovers of private capital and public spending

The analysis estimates the economic returns on public spending by transportation and non-transportation functions vs. private capital, using a panel data set for 48 contiguous states from 1989 through 2002. These actual spending dollars are used as a more precise measure compared to apportioned state public capitals used in the existing literature. For each type of capital/spending, the interstate spillovers were constructed in such a way that different states are weighted by commodity flows across the states to reflect different degree of inter-state dependence. We find that when spending data rather than capital stock is used, all of the interstate spillover effects are negative and statistically significant, suggesting that infrastructure investment does not contribute to economic growth (at least not directly). Therefore, crowding out effects exist among states competing for both private and government funds, in particular if states are highly dependent on allocation of federal funds. These results confirm the finding that previously estimated positive coefficients reflect spurious correlation based on capital stocks and output.