Dissertation Research

From School Boards to State Politics: De-Unionization and the Labor Market for Teachers (Job Market Paper)

How does the labor market for teachers react to a prohibition of collective bargaining? I employ a differences-in-differences design, comparing teachers in Tennessee who lost bargaining rights in 2011 to colleagues within the state whose districts never negotiated. De-unionization modestly reduced teacher compensation, but had no effect on student achievement in the short run. Salaries grew one percentage-point less, cumulatively over four years, while employer-paid health insurance premiums grew five percentage-points less. The law had no impact on either class size or student test scores. In contrast, teachers' unions lost 30 percent of pre-prohibition revenue because contracts no longer stipulate automatic payroll deduction of union dues.

Full Text PDF

Pay for Seniority: Do Civil Servants Value Deferred Compensation?

I estimate the impact of seniority pay on workers’ career trajectories in the U.S. public sector. State Government employees in Michigan participate in either a defined-benefit or a defined-contribution pension plan depending on their date of hire. Defined-benefit members must remain at least 10 years in State Government to receive any benefits, while defined-contribution members vest immediately. Regression discontinuity estimates show that workers with deferred compensation are substantially less likely to leave State service, particularly at older ages and higher pay grades. Governments contemplating pension reform should consider the benefits and costs of a younger workforce with less firm-specific human capital.

Full Text PDF

Tax-Cap Overrides and Teacher Labor Markets (In Progress)

Do local referenda to override property-tax caps affect the wages, benefits, or employment of public-school teachers? School districts in Wisconsin cannot increase property-tax revenue above a per-pupil level, but may circumvent the limit with a majority vote of district residents. Data on the override votes are linked to 1995-2010 teacher personnel records. A dynamic regression discontinuity design compares districts where referenda were determined by a close vote.