An Equilibrium Model of Wage and Hours Determination: Labor Market Regulation in the Retail Sector (Job Market Paper)

A recent push to limit the discretion large retailers have over their employee's schedules aims to increase predictability but at an unmeasured cost to market efficiency. Understanding the implications such policies have for a labor market where hours vary and employees have limited control over their own scheduling requires a model of why such jobs exist and also why individuals accept them. This paper formulates and estimates an equilibrium search model with firm and worker heterogeneity that endogenously generates labor contracts, hiring decisions, and search behavior that matches observed patterns in wages, hours, and employment for the U.S. retail sector. I use a novel approach to separately identify the primitives of the supply and demand side optimization problems that incorporates a mixture of stated preference data collected from workers, data on equilibrium retail jobs, and data on employment flows. The empirical results indicate a counterfactual policy that restricts the extent to which hours may vary in a given week does reduce average variability but also results in a two percent decline in aggregate production and leaves most workers worse off in equilibrium.
See the latest version here

A Model of Human Capital Formation and Contractual Unpredictability (with Flávio Cunha)

We study the importance of contractual unpredictability on labor market participation and outcomes using a dynamic partial equilibrium search model. We develop a model of human capital accumulation where jobs have multiple attributes including variability in hours and earnings, scheduling flexibility, and advanced notice of schedule. We highlight the difficulties of using standard panel data to estimate and distinguish between worker preferences and the primitives of the offer function when jobs have non-monetary attributes and collect hypothetical data that allows us to trace out worker preferences over a variety of job attributes. These preferences then serve as inputs when estimating the distribution of job offers while allowing for dependence on accumulated human capital and the selection effect of only observing accepted offers. This approach provides a rich characterization of why individuals accept jobs with contractual unpredictability and permits us to quantify the effect of restricting the offer space.

The Value of Predictability in the Workplace

Separately identifying the valuations of non-monetary attributes within job contracts requires disentangling their potentially inter-dependent relationship within employee preferences. I use data from a unique vignette-style survey where respondents rank hypothetical job offers to estimate the structural parameters underlying preferences in a random utility model. The model allows for correlated and individual-specific valuations over multi-dimensional job contracts. The results highlight the importance of accounting for these attributes when modeling an individual's decision to accept or reject a job offer and provide a distribution of willingness to pay estimates for a variety of common job attributes.

Hesitating at the Altar: An Update on Taxes and the Timing of Marriage (with Margaret McKeehan)

This paper provides evidence that U.S. tax codes dependence on marital status continues to generate an implicit marriage tax and distort marital decisions. By looking at the timing of marriage rather than the decision to marry (see Alm and Whittington (1997)) we capture a specific distortion while allowing for heterogeneity in other costs of marriage. Using data on couples from the Panel Study on Income Dynamics between 1986 and 2011, we find that a one percent rise in the size of the marriage tax relative to a couple’s income increases the probability of delay by 1.2 percentage points. We further demonstrate the robustness of this result across a variety of alternative specifications and assumptions regarding tax-filing behavior. (Forthcoming, Public Finance Review)
Code available on Github

Data Issues in the Affine Term Structure Models

I show that the choice of data in estimation of a Gaussian affine term structure model affects the qualitative conclusions about the relevance and effectiveness of monetary policy. In particular, the paper introduces the use of real-time data to their estimation which imply different dynamics than using the standard revised series.