Fiscal Multipliers and the Choice of Zero Lower Bound Modeling

In this note we elaborate on the effect of the modeling choice of the zero lower bound on the size of the fiscal multiplier. To this end we contrast two different ways to implement the ZLB in a New Keynesian model: the ZLB modeled as an endogenous central bank reaction to a contractionary demand shock as in Christiano, Eichenbaum, Rebelo (2011) and the ZLB modeled as an exogenous monetary shock as in Cogan, Cwik, Taylor, Wieland (2010). We find that only the former treat the ZLB as an appropriate constraint for policy. We show that the economic significant differences in the size of the fiscal multiplier are not only due to differences in the timing assumption of government spending but also driven by the choice of the ZLB modeling. Ceteris paribus, the impact multiplier is higher if the ZLB is modeled appropriately as a constraint.