Regression Analysis Results

Using regression analysis the Mackinac Center examined per capita state expenditures on economic development programs and changes in per capita GSP to see if there is a strong correlation between the two for all 50 states. The Center regressed this data for a number of time periods. It also tried to find correlation between economic development expenditures and changes in state per capita income for different periods. In every instance the analysis showed a correlation that was insignificant.

The graph on the left shows the results of the former regression. Each dot represents GSP and economic development spending for a state, and the distance between them and the least squares regression line (the line through the center) represents the variation in y (percentage change in GSP) that is not explained by the regression and x (per-capita economic development spending).

A number of qualifications must be offered. First, the data published by NASDA is not perfectly uniform. The NASDA gathers data and attempts to “identify [economic development] programs in a consistent manner,” but the system is not able to capture every economic development dollar perfectly. Second, other factors, such as labor climate, employment rate, and the state’s overall tax burden, were not included in the analysis. Including them may have influenced the results, but probably not to a degree that would have shown demonstrable effects of economic development spending on changes in GSP. The activities of millions of self-interested people, labor unions, and businesses probably swamp the MEDC’s powers to create real economic growth.