We use US county level data (3,058 observations) from 1970 to 1998 to explore the relationship between economic growth and the extent of government employment at three levels: federal, state and local. We find that increases in federal, state and local government employments are all negatively associated with economic growth. We find no evidence that government is more efficient at lower levels. While we cannot separate out the productive and redistributive services of government, we document that the county-level income distribution became slightly more unequal from 1970 to 1998. For those who justify government activities in terms of equity concerns – perhaps even trading off economic growth for equity – the burden falls on them to show that the income distribution would have widened more in the absence of government activities. We conclude that a release of government-employed labor inputs to the private sector would be growth-enhancing.