This paper argues in favour of empirical models for the analysis of the effect of fiscal policy built by including in fiscal VAR models structural shocks identified using information independent from the VAR---i.e. the shocks constructed using a narrative method. We first show that "narrative" shocks are valid shocks in a fiscal VAR, i.e. they are orthogonal to the relevant information set. We then show that the direct inclusion of narrative shocks in a fiscal VAR delivers estimates of the tax multiplier that are similar to those obtained within the traditional fiscal VAR approach. The use of narrative shocks has a big advantage: it does not require the inversion of the moving-average representation of a VAR for the identification of the relevant shocks. Therefore, within this framework, fiscal multipliers can be identified and estimated even when the MA representation of the VARs is not invertible---the relevant case in the presence of fiscal foresight, i.e. when agents receive signals on the tax changes they will face in the future.