Yesterday, USA Today examined the average pay advantage of state and local government workers over workers in the private sector. The article found, for example, that government workers in California earned an average $71,385 in compensation in 2009, or $7,977 more than the average for private sector workers in that state.

Are these government worker pay advantages related to union shares in the states? I performed a simple regression analysis to find out. The answer is “Yes”—states with more unionized government workforces tend to have a higher government compensation advantage.

The chart below shows the regression results, including the raw data (one blue dot for each state) and the regression-fitted line (pink dots). As the union share increases, the pink line indicates that the average government pay advantage increases. In particular, the pink line indicates that for every 10 percentage point increase in a state’s union share, the average government worker pay advantage increases by $765 annually.

Let’s apply these results to Wisconsin, which has a union share of about 50 percent in its state and local workforce. Suppose that Wisconsin makes legal reforms resulting in the union share falling to 10 percent. The regression indicates that the decline would save Wisconsin taxpayers about $3,060 annually for every state and local worker, or about $872 million in total annual workforce costs for the state. (The state has 285,000 state/​local workers).

(The regression was statistically significant at the 95% level, with an F‑statistic of 5.9 and a T‑statistic on the union share variable of 2.4. The R‑square was 11 percent, which indicates that union share explains only a modest portion of the overall government pay advantages between the states).