Association for Public Policy Analysis & Management

The improvement of working conditions is commonly put forward as the main benefit of unionization in education. Unionized teachers are often paid higher salaries (e.g. Brunner & Squires, 2013; Winters, 2011), have lower class sizes (e.g. Brunner & Squires, 2013; Hoxby, 1996), and work fewer hours (Eberts, 1984; Zigarelli, 1996) than their non-unionized counterparts. Perhaps not surprisingly, these working conditions may also have profound effects on the teacher labor market. Working conditions are associated with lower teacher turnover and higher teacher job satisfaction, which could improve the supply of effective teachers to schools (e.g. Boyd et al., 2011; Johnson, Kraft, & Papay, 2011).

Notwithstanding a growing literature on the effect of unions and collective bargaining agreements (CBAs) in education, we know little about how the working conditions ensconced in CBAs affect teachers’ decisions on where to teach and whether to remain in the profession. Given the well-documented shortage of teachers in specific high-needs teaching areas (McVey & Trinidad, 2019), coupled with recent policy changes that could weaken the ability of unions to advocate for teacher working conditions (Marianno & Strunk, 2018), research on the relationship between CBA working conditions and teacher mobility is particularly timely. If the working conditions negotiated into CBAs matter significantly for attracting and retaining teachers, then changes to union power, and in particular, their ability to effectively represent the interests of teachers at bargaining table, could have important implications for the supply of teachers to schools.

Drawing on administrative records from the Michigan Department of Education (MDE) for the 2007-08 through 2015-16 school years linked to information from school district CBAs from two bargaining cycles, I answer the following research question: What is the relationship between CBA working conditions and teacher hiring, transfers, and exits? To measure teacher working conditions I draw on a unique dataset of CBA content from 96 percent of Michigan school districts. This dataset contains information on over 600 different CBA provisions from two successive district contracts, one before the 2011-12 school year and one after. Using information from these contracts, I employ a partial independent item response model to estimate the latent strength of the working conditions for each school district contract in each time period (Strunk & Reardon, 2010). I link this measure of working condition strength to administrative data on teacher hiring and assignments and estimate a series of fixed effect logistic regression models that estimate the probability of a given teacher being hired, transferring to a new district, and exiting the profession as a function of teacher and school characteristics and the strength of districts’ CBA working conditions.

I find that teachers are more likely to stay (versus transferring to another school district or exiting the profession) in districts where teacher working conditions are growing stronger. A disproportionate number of transfers between school districts are from districts with low working condition strength to districts with high working condition strength. These findings have implications for districts’ abilities to attract and retain teachers in an era of shifting union power.