- Working Papers
- Multidimensional Skill Accumulation and Mismatch Over the Lifecycle (Job Market Paper)
- This paper investigates the relationship between worker skills and occupation skill requirements with a focus on skill mismatch over a worker's career. Using data from the NLSY, I first document patterns on skill mismatch in workers' initial jobs that suggests workers, on average, are not in occupations that primarily use their comparative skill advantage. As workers gain experience, on average, they do not correct this initial skill mismatch by transitioning occupations. Considering workers' initial and final jobs observed in the sample, 65% of workers use the same skill intensively over their career and 62% of those workers are using a skill in which they do not hold an initial comparative advantage. Two possible explanations for this skill immobility is that workers are involuntarily constrained or, due to skill accumulation, are rationally staying in an initial mismatch. To explain observed skill immobility, I develop a model of occupational choice with heterogeneous worker skills, matching frictions, and skill accumulation where occupations differ in their skill intensities. The model is calibrated to match occupational skill requirements and is used to decompose worker mismatch into voluntary and involuntary mismatch. Finally, considering a counterfactual model in which skill accumulation is tenure-based, the model predicts the average wages of occupation switchers to fall by 18\%, leading to near complete skill immobility.
- Occupational Wages and Skill Portfolios (joint with Derek Stacey)
- In this paper, we develop a method for estimating occupation-specific skill portfolios that exploits variation in both wages and multidimensional skill ratings. We first present a model of occupational sorting to derive relationships between skills and wages across occupations. We then use these relationships to construct skill measures from wage and O*NET skill and ability data using interpolating polynomials. The resulting skill measures use a common interval scale in log wage units. This allows for straightforward comparisons across occupations, linear aggregation across skills, and an intuitive interpretation of the various components of each skill portfolio. We find that requirements related to literacy and cognitive skills appear particularly important in explaining wage differentials across occupations. We then apply our skill measures to worker- level data to ascribe skill portfolios to individual workers. In doing so we are able to explain skill differentials across worker characteristics. Typically, the literature defines two distinct sets of skills for their analysis - cognitive and manual. We find that it is important to further decompose cognitive skills into two subcategories - analytical and interpersonal. We demonstrate why decomposing skills into three categories better explains the heterogeneity inherent in an occupational skill portfolio. We then compare our results to other common skill estimating methodologies.
- The Gender Wage Gap Using Evidence From Occupational Skill Composition
- In this paper, I investigate the Current Population Survey from 2010-2018 for current evidence of the gender hourly wage gap. I document a gender wage gap of roughly 23% after controlling for observable characteristics. This gaps shrinks 21% if I also include the skill characteristics of the occupation the worker is employed in - indicating that occupational skills are important in explaining the wage gap. I use a Lasso algorithm to choose a parsimonious set of skills to represent the entirety of the skill portfolio. I find that returns to skills vary significantly with gender. This difference is particularly stark for cognitive skills. Women earn $1.15 less per hour purely from a male equivalent set of problem solving skills, for example. I also show this is not primarily due to occupation selection. Instead, I present results for observationally identical workers varying only in gender that show different returns to equivalent skill bundles in selected occupations. For instance, an observationally identical female CEO earns 12% less per hour compared to their male counterpart. This provides evidence that differing skill remuneration is an important factor when discussing the aggregate gender pay gap.