We use a simple model to show why previous empirical studies of budget policy effects are flawed. Due to an identification problem, those studies' findings can be shown to be consistent with policies either mattering or not. We argue that this problem is difficult and not likely to be resolved soon.

Doan, Litterman, and Sims (DLS) have suggested using conditional forecasts to do policy analysis with Bayesian vector autoregression (BVAR) models. Their method seems to violate the Lucas critique, which implies that coefficients of a BVAR model will change when there is a change in policy...