Doctoral Dissertation Research in Economics: What's in a Name? The Effect of Changing Definitions of 'Employer' on Worker Outcomes award

This doctoral dissertation research in economics (DDRIE) project will use machine language and modern economic methods to investigate court decisions on who is an “employee” affect worker’s labor outcomes, such as employment tenure, wages, benefits, and unionization. There has been a large increase in “contingent workers” – workers who are on temporary contracts, contracted out, or are independent contractors. These workers have weaker protections in terms of compensation tenure and collective action. In spite of these developments in the labor market, there is relatively little research on the causes of the increase in the proportion of temporary work or independent contracting in the labor market. A major determinant of the increase in the growth of these “contingent workers” is the legal definition of employee as defined in court rulings. This study will use machine learning and random assignment of judges to specific cases to determine whether changing the legal definition of employee affect the growth of the size of “contingent workers” and how this impacts workers’ wages, tenure loss, unionization rates, and inequality. The results of this DDRIE project will shed light on how economist can combine machine learning tools and economic theory to investigate important labor market and other social issues. The results of the research project will also provide guidance on how to develop policies to improve the functioning of labor markets as well as how to increase the living standards of workers, especially those at the lower end of the earning spectrum.

This research project will estimate the causal effects of changing legal definition of an employee. It will collect data on all cases in which a court made a decision about whether workers could be considered “contractors”, and use machine learning to analyze case text to determine the direction of decision and the set of affected workers. It exploits the random assignment of judges to specific cases to obtain exogenous variation for identification. Judges vary systematically in their decisions on employment definitions, and that variation can be predicted by judge characteristic, such as age, race, political preferences, and education. Building on work using machine learning and variable selection, we train a regularized regression model to predict case decision from characteristics of judges assigned to a case. The cross-validated model produces an exogenous instrument for use in a two-stage least squares regression. This approach is leveraged to analyze the causal effects of changing definitions of employee on worker outcomes. The project then investigates whether a legal opinion that workers can be considered contractors affects unionization rates, remuneration, and inequality. This project directly examines the impact of legal precedent on worker outcomes and contracting rates, as well as the role of legal institutions in determining contracting outcomes, and its effect on employment and wages. The results of the research project will provide guidance on policies to improve the functioning of the labor market as well as how to increase the living standards of workers at the lower end of the earning spectrum.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.