Association for Public Policy Analysis & Management

The extent to which a public policy has benefits or costs that spillover from one governmental jurisdiction to another jurisdiction has broad implications for public policy decisions.  Whether it is best to base a policy decision at the local, state, federal, or international level, is determined in part based on the extent of these spillovers (Tiebout, 1956, “A Pure Theory of Local Expenditures,” Journal of Political Economy). 

Minimum wage policies are one such policy arena where the extent of spillovers is crucial for the determination of the best locus of control.  If spillovers are great, then it would be best to have the minimum wage set at a higher level of government; yet if these spillovers are small, then local minimum wages that reflect local cost of living and local preferences may be warranted.  Recently, a large number of local jurisdictions in the United States have adopted local minimum wage ordinances in response to a declining federal minimum wage in real dollars.  For example, the City of Seattle enacted an ordinance that raised the local minimum wage from $9.47 per hour (i.e., the State of Washington’s minimum wage) in 2015 to $15.00 per hour in 2017, which now far exceeds the federal minimum wage of $7.25 per hour.

Our team (The Minimum Wage Study at the University of Washington, http://evans.uw.edu/policy-impact/minimum-wage-study) has begun an investigation of the impact of this ordinance on employer adaptations to higher wage costs; expectations and worries of low-wage householders; impacts on employment, earnings, and hours worked; and the distribution of gains and losses across workers or industries.

To conduct such analyses and to estimate a causal effect of the ordinance on these outcomes, it is necessary to define both the treatment region and a control region.  The control region is the area that is not effected by the policy, and is identified so as to produce an estimate of counterfactual (i.e., what would have happened to the treated region in the absence of the treatment).  In many difference-in-differences analyses, including a large number of studies of the impacts of minimum wage laws, the control region is defined using a political jurisdiction that is adjacent to the treated jurisdiction in which the policy is enforced.  Yet, if the policy effects spillover to this adjacent political jurisdiction, then the control group will be contaminated by exposure to the treatment.

In this paper, we describe our method for identifying and separating the spillover region from the remainder of the State of Washington, which thus allows us to create a control region for identifying policy impacts.  Our research will yield policy impacts that are of interest for scholars of labor market policy.  And, our methodological contribution will have broad implications for scholars that use difference-in-differences analyses to estimate policy effects.  This work fits with the conference themes, particularly for scholars and policymakers interested in “Employment in a Global Market” and “Multilevel Governance, Federalism, and Regional Cooperation.”