Research

We present a unified technological explanation of both the movement of workersacross jobs using different skills and the changes in skill use within jobs. An envelope-theorem approach allows us to estimate relative skill-productivity growth from worker mobility using OLS while making minimal assumptions on each occupation’s production function. Using six decades of data, we conclude that routine-cognitive skill productivity grew rapidly and abstract- and social-skill productivity grew slowly - a form of “skill bias.” These effects, along with our estimated relations between skill inputs, also explain changes in skill use within occupations. 

We study how firms and workers adjust when previously restricted migrants gain full and portable work rights in a labor market with substantial informality. We exploit the 2007 EU accession of Bulgaria and Romania, which granted unrestricted work rights to Italy's largest migrant group. Using matched employer–employee administrative data and an IV-DID design, we find that firms suddenly and persistently shift employment composition toward EU07 workers, compressing the native employment share without reducing native hiring or increasing separations. We don't detect any significant change in wages for either natives or EU07 workers. For migrants, this null effect reflects offsetting compositional shifts as newly observed and incumbent EU07 workers enter the formal workforce with different wage trajectories. Consistent with a shift in bargaining power toward workers, EU07 migrants experienced significant gains in job mobility and job security. Overall, the evidence suggests that removing legal restrictions reshaped firms’ personnel choices and altered migrants’ employment relationships, improving their outside options, bargaining position, and access to more secure jobs.

The Black–white male wage gap narrowed in the 1960s and 1970s but stopped convergingafter 1980. I show that routine-biased technological change contributed to this shiftthrough different occupational responses: white men left routine-intensive occupationswhile Black men entered them. The divergence is driven by non-college workers andis stronger among younger workers. It arises through job-to-job entry rather thanjob-to-job exit and persists among workers with the same measured ability. Routine-intensive occupations experienced almost no median wage growth, and their durable paydispersion fell relative to that of abstract occupations. Black men disproportionately entered occupations with weak subsequent wage growth, and occupational sorting offset forces toward racial wage convergence. I rationalize these findings in a novel model in which technological change lowers routine occupations’ return to ability and employers observe Black workers’ ability less precisely, generating opposite racial sorting responses to the same shock .

We study how labor-market tightness affects minority earnings gaps using administrative data for U.S. metropolitan areas from 2010 to 2025. After accounting for national shocks, local seasonality, and local trends, a one-percentage-point unemployment decline reduces the white-minus-minority earnings gap by about 0.2 percentage points for all jobs and 0.3 percentage points for new stable hires. These responses strengthen as slack declines, reaching 0.4 and 0.6 percentage points, respectively, in hot markets. Neither observed worker and vacancy com-position nor concurrent policy changes account for these effects. Increased mobility and upgrading across broad job categories explain little of the narrowing, although hiring distribu-tions become more similar between the two groups. Where recruiting is concentrated among fewer employers, the earnings-gap response varies less with slack. Together, the findings point toward relative earnings improvements within labor-market segments and employer competition as a mechanism through which tightness narrows minority earnings gaps.