Tax multipliers: Pitfalls in measurement and identification

To analyze the pitfalls in measuring and identifying tax shocks, we build a novel value-added tax rate dataset for the period 1980-2009. The problem of identification (i.e., changes in tax policy not triggered by output fluctuations) is clearly disentangled from the problem of measurement (i.e., finding a tax policy variable under the direct control of policymakers). On the identification front, our results favor the use of narratives à la Romer and Romer (2010). On the measurement front, our findings support the use of tax rates as the true measure of tax policy as opposed to revenue-based measures, such as cyclically adjusted revenues.