Teachers' early career attrition, retirement, and wage determination: Evidence from panel data

This dissertation consists of three essays that investigate different aspects of teachers' labor markets. The first empirically models early career teacher attrition. The second develops and empirically tests a model in which cost-minimizing school districts can employ early retirement incentives to affect the experience composition of their labor forces. The third develops and estimates a structural model of wage and employment determination.The first essay focuses on whether teacher's early career teacher quit behavior is consistent with occupational choice models. In addition, it models both teacher movements out of the teaching profession and to other school districts. Teacher attrition is modeled using a competing-risks discrete-time hazard model. The model is estimated using data on 5049 teachers who began teaching in New York State between 1978 and 1983. The study finds empirical support for the importance of occupational choice characteristics on teacher mobility. In particular, school districts with higher rates of wage growth have lower teacher attrition.The second essay presents a model in which cost-minimizing school districts may enact policies, such as early retirement policies, that affect the quit rate of experienced workers. A key prediction of the model is, all else equal, school districts that pay experienced teachers higher wages are more likely to have early retirement incentives. The empirical implications of the model are tested using panel data from nearly 500 New York State school districts. The empirical results are consistent with the predictions of the model: school districts appear to use early retirement policies in a manner consistent with cost-minimization.The third essay develops a public sector bargaining model of wage and employment determination. The model's solution generates structural wage and employment equations that are then estimated using panel data from 688 New York State school districts. This approach permits me to distinguish empirically among the major models of wage and employment determination in a unionized market: "monopoly union", "right to manage", "efficient contracting", and "strongly efficient contracting" models. The empirical results support the efficient contracting model. In addition, separate measures of union bargaining power over wages and employment are obtained. Finally, the paper proposes a method of measuring the level of public services that flow to a community when a unionized work force is present. That is, it emphasizes that variations in union rents across districts should be controlled for when one makes judgments about the variability of educational service levels being provided across districts.