Same, But Different: Testing Government Expenditure Shock Measures

Abstract

In this study we test whether different government expenditure innovation measures proposed in the literature are good proxies for government spending shocks. We compare five government spending shock measures: forecast error of survey of professional forecasters (SPF), forecast error of forecast of Federal Reserve published in Greenbook (GB), military news, Fisher-Peters measure, Ben Zeev-Pappa measure. To compare these measures, we utilize the reliability estimator of a proxy, R², and F statistics which are proposed by [1]. We find that SPF and GB constitute the best proxy variable for government spending shocks with regard to different reliability estimators. SPF and military news proxies have the highest multipliers comparing to the others. Their both multipliers are greater than 1. Multiplier for Ben Zeev-Pappa and Fisher-Peters measures are less than zero. GB also has a positive multiplier but less than one.