Labor Studies Program Meeting, Fall 2021

Acemoglu and Restrepo document that between 50% and 70% of changes in the US wage structure over the last four decades are accounted for by the relative wage declines of worker groups specialized in routine tasks in industries experiencing rapid automation. The researchers develop a conceptual framework where tasks across industries are allocated to different types of labor and capital. Automation technologies expand the set of tasks performed by capital, displacing certain worker groups from jobs for which they have comparative advantage. This framework yields a simple equation linking wage changes of a demographic group to the task displacement it experiences. Acemoglu and Restrepo report robust evidence in favor of this relationship and show that regression models incorporating task displacement explain much of the changes in education differentials between 1980 and 2016. Task displacement captures the effects of automation technologies (and to a lesser degree offshoring) rather than those of rising market power, markups or deunionization, which themselves do not appear to play a major role in US wage inequality. The researchers also propose a methodology for evaluating the full general equilibrium effects of task displacement (which include induced changes in industry composition and ripple effects as tasks are reallocated across different groups). Their quantitative evaluation based on this methodology explains how major changes in wage inequality can go hand-in-hand with modest productivity gains.

Is government guiding the invisible hand at the top of the labor market? Gottlieb, Polyakova, Rinz, Shiplett, and Udalova combine the administrative registry of US physicians with tax data, Medicare billing records, and survey responses to measure physicians' earnings and estimate the influence of healthcare policies on physicians' incomes and labor supply. Annual earnings average $343,600, with major differences across specialties, and comprise 8.5% of national healthcare spending. The age-earnings profile is steep and business income comprises one-quarter of earnings. Health policy has a major impact on the margin: 30% of incremental Medicare spending on physician care goes to physicians personally, and 60% for Medicaid. Physicians gain 5-10% of public money spent on insurance expansions. Policies that increase reimbursement also increase physicians' supply of care and reduce retirement. Gottlieb, Polyakova, Rinz, Shiplett, and Udalova find that government plays a major role in determining the value of physicians' human capital, but that its power is constrained by the market. This limits the scope for policy instruments aiming to reduce US spending on physician services.

Youth employment has been near historic lows in recent years, and racial gaps persist. Heller and Kessler test whether informational frictions contribute to these outcomes by running a large-scale field experiment involving over 43,000 youth in New York City aimed at providing credible information on young job applicants. The researchers build software that allows employers to quickly and easily produce letters of recommendation for randomly selected youth who worked under their supervision during a summer youth employment program. Heller and Kessler then send these letters to nearly 9,000 youth. Being sent a letter generates a 3 percentage point (4.4 percent) increase in employment the following year, with both employment and earnings increases persisting over the two-year follow-up period. By posting their own job advertisement, the researchers document that while treatment youth do use the letters in applications, there is no evidence of a supply-side response (i.e., no increased job search, motivation, or confidence); effects appear to be driven by demand-side responses. Labor market benefits accrue primarily to racial and ethnic minorities, suggesting frictions may contribute to racial employment gaps. But improved employment may also hamper on-time high school graduation. Evidence that letters help improve job match quality suggests that expanding the availability of credible signals about young workers, especially non-White applicants who are no longer in school, could improve the efficiency of the youth labor market.

How do firms set wages across space? Using vacancy data with detailed job-level information supplemented with a survey of HR managers and self-reported data on workers’ wages, Hazell, Patterson, Sarsons, and Taska show that, within the firm, 30-40% of posted wages for a given job are identical across space. Compared to differences between firms, nominal posted wages within the firm vary relatively little with local prices, a pattern that is present for realized wages as well. Using the pass-through of local shocks to wages in other locations of the firm, Hazell, Patterson, Sarsons, and Taska argue that the limited variation of wages within firms is due to national wage setting, meaning that firms choose rigid pay structures in which they set very similar nominal wages for the same job in different regions. Their survey suggests that one reason firms set wages nationally is that nominal, rather than real, wage comparisons matter to workers.

What is the value of workplace dignity to low-wage workers, and is it supplied efficiently? Dube, Naidu, and Reich present a methodology to elicit valuations of difficult-to-measure job characteristics under imperfect competition, and use it to measure the value of workplace dignity to workers and its relationship to wages. They draw on extensive ethnographic work conducted with 87 Walmart workers to design and implement a survey experiment with over 10,000 Walmart workers recruited online. The researchers first validate their experimental design by showing that their hypothetical quit elasticities are close to other estimates in the literature, and that subjects’ behavior (their likelihood of clicking on an outside job link) is consistent with their stated preferences and together imply a non-trivial degree of monopsony power. Next, Dube, Naidu, and Reich estimate workers’ valuations of different workplace amenities, from commuting and scheduling to more subjective measures of workplace dignity, including supervisor respect and self-expression at work. The researchers find that workers value workplace dignity at approximately six percent of their current wage, making it comparable to amenities like commute time and more valuable than widely discussed amenities like control over one’s schedule. Dube, Naidu, and Reich find that workers' experience of dignity at work is higher for white, older, and Southern workers, and is correlated with a measure of job rents. High dignity jobs are also associated with a lower quit elasticity, but a higher bargaining elasticity, with no such heterogeneity by the non-dignity job amenity values. Third, they use geographic variation in the bite of Walmart’s 2014 voluntary minimum wage policy to estimate the causal impact of higher wages on changes in dignity. Contrary to the theory of compensating differentials, Dube, Naidu, and Reich find that workplace dignity not a substitute for wages, as reported dignity values do not decrease among those workers likely to have experienced a raise as a result of the voluntary minimum wage.

Card, Rothstein, and Yi use longitudinal data from the LEHD to study the causal effect of location on earnings. They specify a model for log earnings that includes worker effects and fixed effects for different commuting zones (CZs) fully interacted with industry, allowing us to capture potential impacts of local specialization. Building on recent work on firm-specific wage setting, the researchers show that a simple additive model provides a good approximation to observed changes in log earnings when people move across CZ’s and/or industries, though it takes a couple of quarters for migrants to fully realize the gains of a move. Card, Rothstein, and Yi also show that the earnings premiums for different CZ-industry pairs are nearly separable in industry and CZ, with statistically significant but very small interaction effects. Consistent with recent research from France, Spain and Germany, the researchers find that two thirds of the variation in observed wage premiums for working in different CZs is attributable to skill-based sorting. Using separately estimated models for high and low education workers, Card, Rothstein, and Yi find that the locational premiums for the two groups are very similar. The degree of assortative matching across CZs is much larger for college-educated workers, however, leading to a positive correlation between measured returns to skill and CZ average wages or CZ size that is almost entirely due to sorting on unobserved skills within the college workforce.