Government spending shocks are frequently identified in quarterly time-series data by ruling out a contemporaneous response of government spending to other macroeconomic aggregates. We provide evidence that this assumption may not be too restrictive for annual time-series data.