The size of government is a commonly used variable in many analytical studies on the effects of fiscal policy. An accepted practice is to measure it as the ratio of government spending to GDP. However, this is not the correct metric when computing the stabilization effects of nondiscretionary fiscal policy. Intuitively, public spending does not react to cyclical conditions as much as taxes do - as reflected in the standard zero-one elasticity assumptions for spending and revenue, respectively. This paper shows that the revenue to GDP ratio is the appropriate indicator of government size for the purpose of assessing the stabilization effects of nondiscretionary fiscal policy.