Demand and Distribution in a Dynamic Spatial Panel Model for the United States: Evidence from State-Level Data

We estimate a modified demand-and-distribution system for the 48 contiguous US states and the District of Columbia (DC) employing spatial dynamic panel data for 1980–2019. We allow for endogenous regressors, test for the presence, significance, and magnitude of spatial spillovers, and estimate both immediate and cumulative effects on our endogenous variables of interest. Without testing for spatial dependence and spillovers, we estimate that output growth and capacity utilization in the sample US states and DC rise in response to an increase in their own wage share. Yet when we test for spatial dependence and spillovers as required by the state-level nature of the data, we estimate that a higher state wage share lowers output growth and capacity utilization in the own state, but raises output growth and capacity utilization in neighboring states. The former direct effect is larger (smaller) in absolute value than the latter indirect effect in the case of capacity utilization (output growth). Meanwhile, we find that a higher state output growth or capacity utilization reduces the wage share in the own state, but raises the wage share in neighboring states. The former direct effect is larger in absolute value than the latter indirect effect.