Collective bargaining agreements (CBAs, or contracts) are negotiated by local teachers’ unions and school districts approximately every three years. These documents contain policies and regulations that dictate aspects of education policy in every facet of the school and district workplace, and govern the interactions between teachers, students, school personnel, administrators and parents. Some researchers and policymakers think that CBAs inherently restrict district administrators from enacting policies and changes that will help to improve student achievement.  Others, however, believe that CBAs serve an important role in protecting teachers and enhancing working conditions in ways that will help attract and retain teachers, especially to hard-to-staff districts.

Research on both ends of the debate is limited because most of the extant work has used data on contracts and contract restrictiveness gathered from only one point in time. This inhibits researchers from assessing how CBAs change over time and how these changes are associated with differences in important outcomes and district characteristics. In addition, the cross-sectional nature of the data has made it difficult for researchers to unpack the direction of these relationships: do more or increasingly restrictive CBAs lead to lower student achievement, or do difficult working conditions within the district (as indicated in part by low student achievement) bring about more restrictive contracts? To this end, it is possible that districts with lower student achievement, in an effort to improve working conditions, negotiate CBAs with greater teacher protections, which translate into more restrictive contracts.

Using a dataset of 400 California school district contracts negotiated in three separate time periods married to administrative district- and school-level data, this paper begins to explore these open questions. I ask: 1) How does CBA restrictiveness change over time?; 2) What district characteristics are associated with changes in contract restrictiveness?; and 3) What is the relationship between changes in contract restrictiveness and student achievement? To answer these questions, I use data from teachers’ union contracts collected in the 2005-6, 2008-9 and 2011-12 school years. I transform data on individual contract provisions into measures of contract strength using a Partial Independence Item Response (PIIR) model to assess CBA’s inherent level of constraint on administrators’ discretion (Strunk & Reardon, 2010). I use first-differenced regressions to examine how changes in contract restrictiveness are associated with changes in district characteristics, as well as with ensuing changes in student achievement. 

I find that, although much has been made of negative cross-sectional relationships between contract strength and student achievement (Moe, 2009; Strunk, 2011; Strunk & McEachin, 2011), there is no relationship between changes in contract strength and ensuing changes in student achievement, on average or at any point in the performance distribution. These results suggest that contracts that grow increasingly restrictive may, in fact, not harm student achievement. I then show that districts that grow in size generate stronger contracts over time, suggesting that contracts may become more restrictive due to bureaucratic bloat and/or attempts to mange increasingly difficult working conditions.