We study panel data regression models when the shocks of interest are aggregate and there are omitted macro and micro-level shocks of any relative size. This speaks to a large empirical literature that targets impulse responses via panel local projections. We show how to interpret the estimated coefficients when responses are heterogeneous and that a simple recipe leads to uniformly valid inference over the macro–micro composition of the errors: including lags as controls and then clustering at the time level. Finally, we use our methods to reassess the role of firm financial frictions in shaping the transmission of monetary policy.