The Local Multiplier: Theory and Evidence

I show that 1) the policy-relevant “global multiplier” can be written as the sum of a spending component and a taxation component, all scaled up by spillover effects, 2) the “local multiplier” is exactly the spending component, and 3) if trade is anonymous, the local effects of a shock to federal government purchases in a county will be identical to the effects of a shock to consumer demand for the exports of that locality. I estimate a bound for the local multiplier and consider spillover effects to contiguous counties. I find that a shock of $48,000 creates at least one job-year locally. Analysis at a monthly frequency suggests that these jobs are more persistent than previously estimated. Evidence of higher multipliers in recessions is mixed.