John Fallon

Competitive Occupational Licensure: Doctors Versus Chiropractors

This paper provides the first analysis of competitive occupational licensure, where substitute professions maintain separate licensing boards that set entry requirements strategically. I develop and structurally estimate a model where professional organizations choose licensing stringency to maximize industry profits while accounting for competitive responses, as workers with heterogeneous abilities select occupations based on expected returns and consumers observe only average quality within each profession. Testing this theory using historical competition between medical doctors (MDs) and chiropractors (DCs) from 1907–1960, I exploit digitized American Medical Association records and state-by-year variation in chiropractic board adoption. Medical boards responded strategically by increasing college requirements by 10 percentage points, mandating internships (10+ percentage points), and reducing pass rates by 5 percentage points. These regulatory changes generated substantial economic effects: doctors' home values rose by 26% while their numbers declined by 17–40 practitioners per 100,000 population, and chiropractors saw 44% higher home values with increased market presence of 2–12 practitioners per 100,000. Structural estimation reveals that observed equilibria closely approximate profit-maximizing sequential competition rather than welfare-maximizing behavior.

Fewer Licenses, Similar Teachers: Changing Licensing Tests in Indiana

with Marcus A. Winters, PhD

We use longitudinal administrative data from Indiana to examine changes in teacher quality following the state's shift to a more difficult licensure test. The number of unique individuals issued an initial Indiana teaching license fell by about 25% under the new standard. Despite this substantial drop, the overall quality of incoming teachers and the relative quality of licensed teachers compared to unlicensed teachers remained largely unchanged. We find some heterogeneity by subject and school setting, with urban schools experiencing a measurable decline in teacher quality, particularly in math. We document a clear shift toward emergency-permit teachers following the introduction of the more difficult test. This shift may have moderated the average-quality response and concentrated effects in markets with binding teacher-supply constraints.

Extracting Value from Coworkers: Who Should Work with Whom and When?

Workers learn from their peers, but which collaborations build lasting skills? I study co-teaching, where general and special education teachers share a classroom, using administrative records on Indiana teachers, 2012–2019. I compare the same teacher before and after co-teaching. Those paired with an experienced partner (16+ years) score 0.10σ higher after returning to solo instruction; estimates for inexperienced partners are near zero. Longer collaborations add nothing. Teachers with two years of experience gain most. Schools do not exploit this: conditional on staffing, partner assignment is indistinguishable from random. Reallocating partnerships toward experienced mentors implies fiscal benefit-cost ratios of 3.6:1–7.0:1.

The Diminishing Returns to Human Recruiting in Online Labor Markets

with Emma Wiles and John Horton

Employers often rely on outside recruiters to find workers, but it is unclear whether human intermediaries add value when employers already have access to algorithmic screening tools. We study a randomized experiment in a large online labor market that assigned human recruiting assistance to job postings. Employers of treated job posts received 5% more applications and shifted their interviewing toward recruiter-sourced candidates and away from applicants who found the post on their own, yet were no more likely to hire than employers of control posts with access only to algorithmic tools. On average, treated and control posts were statistically indistinguishable in the hours their hires worked, the wages they paid, and the feedback ratings the engagements earned; what employers spent on those hires fell over a 30-day window, significantly so on the full sample though not on our first-post main specification, and the noisier 90-day window can neither confirm nor rule the effect out. Match quality thus shows a null-to-mildly-negative effect rather than any improvement. We develop a model of delegated recruiting in which recruiters and employers rely on a common noisy signal; it predicts that recruiting adds value when the recruiter brings information the algorithm lacks and can subtract value when the two parties' assessments are correlated. Consistent with the model's screening mechanism, recruited applicants are positively selected on observable characteristics yet do not improve hiring outcomes. These findings suggest that as algorithmic screening improves, the scope for intermediaries to add value shrinks because it becomes harder to access independent information.