Working papers
The Origins and Evolution of Occupational Licensing in the United States (with Jason Hicks, Andrew Karch, and Morris Kleiner)
Download paper (updated March 2025); NBER Working Paper
The analysis of occupational licensing has concentrated largely on its labor market and consumer welfare effects. By contrast, relatively little is known about how occupational licensing laws originated or the key factors in their evolution. In this paper, we study the determinants of U.S. licensing requirements from 1870 to 2020. We begin by developing a model where licensing arises as an endogenous political outcome and use this framework to study how market characteristics and political incentives influence regulators’ choices. Our empirical analysis draws on a novel database tracking the initial enactment of licensing legislation for hundreds of unique occupations, as well as changes to the specific qualifications required to obtain a subset of licenses over time. We first show that, consistent with the predictions of our model, licensing requirements are more common and were adopted earlier for occupations whose tasks plausibly pose some risk to consumers. Second, large, urbanized states are significantly more likely to produce new policies. Third, among occupations regulated before 1940, licensing requirements appeared earlier in states with more practitioners and where incumbent workers likely experienced greater labor market competition. After 1980, state-level factors are more strongly associated with the timing of policy adoption. Finally, political organization, as measured by the establishment of a state professional association, significantly increases the probability of regulation. Together, our findings suggest that both public and private interests have contributed to the diffusion of licensing requirements across states and occupations.
The Labor Market Effects of Occupational Licensing Policies
Download paper (Under review, updated April 2025)
This paper offers new evidence on the labor market effects of occupational licensing using a novel dataset documenting the history of licensing laws across U.S. states, which I link to earnings and employment outcomes from the Occupational Employment and Wage Statistics survey, the Current Population Survey, and a database of online job postings from Lightcast. Across these samples, I find robust evidence that the typical policy I study is associated with a moderate wage premium and an increase in employment. These results are driven by laws that have been in effect for at least two decades, implying that policy changes are slow to show through to state-level averages. Additionally, I find that licensing may substitute for private certification and increase the uniformity of credential requirements across employers. Taken together, my findings contradict the view that licensing primarily reduces labor supply without an offsetting increase in labor demand, at least for most occupations covered by my data. That said, I also document significant heterogeneity in the wage and employment effects of licensing, and find some evidence that employment falls when the consumer protection rationale for regulation is weak. Finally, I show how differences in sample definitions and measurement can reconcile my findings with previous research.
Historical Data on Occupational Regulation in the United States
The data described in this paper are now available on my GitHub page: (ncarollo/licensing-data)
Download paper (updated January 2025)
This paper describes the construction and validation of a novel dataset that compiles over one hundred and fifty years of occupational licensing, certification, and registration requirements in all fifty states and the District of Columbia. It identifies the enactment and effective dates of state and federal policy changes for more than 300 distinct occupations that cover the vast majority of regulated jobs in the United States. Relative to existing data on occupational licensing and regulation, my dataset is more comprehensive, reduces measurement error in enactment dates, and offers an internally-consistent taxonomy of regulatory methods. Linking my policy data to independent measures of occupational licensing and certification from recent household survey data and the text of online job postings, I show that state-level policy differences have an economically significant effect on the credentials workers hold and employers require. I discuss the interpretation of estimates based on these measures and conclude with recommendations and directions for future research.
Other work in progress
New Work in the Second Industrial Revolution (with Jingyi Huang and Elior Cohen)
This paper introduces a new measure of emerging occupations in the late 19th and early 20th century United States. Rapid technological and procedural innovation dramatically altered the task content of production, leading to the emergence and diffusion of entirely new work activities and occupations. We use the original write-in occupation responses recorded on Census enumeration forms between 1850 and 1940 to identify specific, granular, job titles such as “comptometer operator” at the individual level and track their growth over time. Using this data, we first document a set of novel facts about new jobs during the Second Industrial Revolution including the locations where they first appeared, how quickly they diffused across local labor markets and industries, and the demographic characteristics of the workers who held them. Next, we match job titles observed in the Census to textual descriptions of their tasks and tools derived from the first edition of the Dictionary of Occupational Titles. We link these job descriptions to a comprehensive corpus of U.S. patent texts to study the relationship between the adoption of new production methods, the skill content of new work, and local labor market power.
Occupational Licensing in the U.S. Progressive Era (with Jason Hicks)
The first major wave of occupational licensing and regulation in the United States occurred during the Progressive Era around the turn of the 20th Century, concurrent with rapid urbanization, technological innovation, and the emergence of entirely new occupations in fields ranging from healthcare to engineering. In this paper, we reassess the effects of occupational licensing, certification, and registration laws during this time period using a variety of novel data. First, we use recently-compiled data on the history of state and federal licensing policies, which significantly improves upon the accuracy of previously-available datasets. Second, we use data from full-count Census microdata. In addition to covering 100% of the U.S. population and providing fine geographic detail, we leverage the original write-in job titles recorded by Census enumerators to identify workers in licensed occupations. This reduces measurement error in workers' occupational affiliations and allows us to identify dozens of occupations that are not separately identifiable in the Census Bureau's public classification. Finally, we leverage new data on the qualifications needed to obtain a license for a subset of occupations including barbers and beauticians to study the response of employment to changes in mandatory training hours.