Association for Public Policy Analysis & Management

The City of Seattle passed a local minimum wage ordinance in 2014, which raised the minimum wage from $9.47 to as high as $15 per hour in 2017. Our previous research on the impact of the ordinance, using repeat cross-sectional data, attributed to the ordinance a 3 percent wage increase and a 9 percent decline in hours worked in low-wage jobs in Seattle (Jardim et al., 2017). While these results reveal the ordinance’s aggregate impacts on Seattle’s labor market, they do not illuminate the distributional consequences critical to discussions of its welfare effects.

This paper examines the medium term effects of Seattle’s Minimum Wage Ordinance at the individual level using employer-employee matched data from Washington’s Employment Security Department. As Washington is one of four states which mandates that employers report both earnings and hours each quarter, it is possible to track hours as an outcome and impute average hourly wages. The paper exploits the longitudinal nature of the employer-employee matched data to estimate three distinct models:

1) A model of employment spell survival, and conditional on survival, wages and hours. This model will reveal the impact of the minimum wage ordinance on conditional job separation probabilities, wages, and hours for workers initially employed in Seattle. We will pay particular attention to how these impacts vary according to initial wage and spell duration, with theory suggesting that the least-paid and/or least-experienced workers may experience less favorable impacts than those with wages closer to the new floor.

2) A model of non-employment spell duration for workers separating from observed jobs, which will show the impact of the minimum wage for initially employed workers. It will also reveal whether workers became more likely to end a non-employment spell by accepting work outside the City of Seattle. We will pay particular interest to effect heterogeneity by final wage in the previous job.

3) A model of employment spell initiation, which will examine the impact of the minimum wage on the rate at which new employee/employer matches emerge. While the second model will analyze the experiences of previously attached workers, the model of spell initiation will incorporate the rate at which new workers appear in the labor force.

Each model will employ a variant of difference-in-differences or synthetic control analysis, utilizing employment records over the period 2005-2014 as “pre-treatment” records and parts of Washington State outside Seattle as a comparison region. Results of these models will have the capacity to distinguish among competing narratives for why low-wage employment has declined in Seattle, and specifically whether they translate into negative impacts for workers observed in low-wage Seattle jobs at baseline.

References

Jardim, E, Long, MC, Plotnick, R, van Inwegen, E, Vigdor, J, Wething, H. “Minimum Wage Increases, Wages and Low-Wage Employment: Evidence from Seattle.” National Bureau of Economics Research Working Paper No. 23532, Issued in June 2017, Revised in October 2017.