Session 16: The Micro and Macro of Labor Markets

Should workers be concerned about the climate transition and the pace at which it occurs? As policymakers choose a timeline to achieve zero net carbon emissions, it is important to know how the pace of this transition will affect the labor market outcomes of workers employed in polluting sectors. We develop a life-cycle model with directed search, skill heterogeneity and retraining in which labor demand expands in a set of clean sectors and declines in polluting sectors. We show that the set of workers who win or lose from the transition is sensitive to its pace. Workers’ lifetime earnings are strongly non-linear and even non-monotonic as a function of the pace at which the clean sector takes over. We structurally estimate the model using US data and use it to evaluate the distributional effects of climate transitions of various speeds. A baseline transition ending in 2060 induces sizable earnings losses, with a large dispersion among workers. Accelerating the transition to end in 2050 is preferable for most workers but thickens the left tail of losses.

Managers often report that labor constraints – defined as inability to find workers – are a

major obstacle to firms’ growth. This phenomenon is puzzling, because economic theory offers a simple remedy: increase wages until the worker is found or hiring is no longer profitable. We explore why firms report labor constraints instead of pre empting them by increasing wages using administrative data from Germany. We confirm that quasi-exogenous variation in labor constraints slows down firm growth. Wages play a role consistent with basic theory: firms that report constraints initially underpay their workers, increase wages later, and a quasi-exogenous increase in wages alleviates their problems. Why then do firms not increase wages earlier to avoid the problem to begin with? Unlike financial markets, labor markets do not have an easily observable price process. Firms set wages based on their beliefs, and when they underestimate market-clearing wages, labor constraints arise. Consistent with this mechanism, labor constraints increase after quasi-exogenous wage increases in other parts of the economy and are more prevalent in settings where firms are less informed.

Wages vary substantially between and within cities. While wages are on average higher in larger cities, the real earnings of low-wage workers are lower. Using French matched employer-employee data, I document two novel facts that highlight the role of employers in shaping between- and within-city inequality jointly. First, high-paying jobs are concentrated in large cities whereas low-paying jobs are present throughout France. Second, the wage gains offered by large cities materialize over time as workers reallocate from low- to high-paying jobs. I propose a spatial framework that rationalizes these facts through two ingredients: heterogeneous employers and frictional local labor markets with on-the-job search. Productive employers agglomerate in large cities to hire more workers. Fiercer competition for workers arises. A higher average wage, faster growth, and greater within-city inequality follow. I estimate the model and quantify that local TFP gaps are minimal once I account for employers’ incentives to sort by size. The steeper ladder of large cities implies higher lifetime real earnings for every local worker, including those with lower real wages.

Pakistan ranks in the lowest decile in female labor force participation, and even in sectors where women are more prevalent, such as teaching, they earn 70 cents for each dollar men earn. While we have extensive evidence on the prevalence of gender bias in hiring, promotions and wages, we know less about the mechanisms underlying this bias and the extent to which certain personnel policies may mitigate or exacerbate these biases. To test this, I conduct a large scale field experiment with 3,600 employees in 250 schools and randomly vary i). how often managers observe a given employee and ii). whether manager evaluations affect employee’s pay or are just used for feedback. First, I find when there are no financial stakes associated with performance evaluations, there is minimal difference in scores between men and women. This holds even when controlling for a rich set of controls of teacher productivity, such as value-added, clock in and out time, time use, and pedagogy measured via classroom observations. In contrast, when principals’ evaluations determine teachers’ end of year raise, we see that female teachers receive 20% lower raises, controlling for productivity. However, when principals are randomly assigned to conduct more frequent classroom observations of the teacher, this increases the evaluation of female teachers and closes two-thirds of the gender gap under financial stakes. To understand mechanisms, I conduct a follow up vignette survey to test whether our results are due to differential manager expectations about employee reactions to low raises (e.g. higher turnover by men) or differential perceived deservedness of scarce financial resources. The results favor this second explanation as managers favoring single-earner (lower income) households over dual-earner (higher income) households, which is highly correlated with employee gender. Combined this suggests that improving the accuracy of manager information could close the gender gap in performance evaluations, even in high stakes settings.

Legal regulations of hours have become increasingly common, yet less is known about how workers’ earnings and hours are set in employment contracts. In this paper, I document that workers commonly report changes in their hours, but the correlation between changes in earnings with changes in hours differs depending on whether their explicit contract pays them by the hour or not. To understand these patterns, I develop and estimate a frictional labor market model with implicit contracts motivated by risk-sharing and constrained by limited commitment. Although the model has no explicit hourly and non-hourly contracts, implicit contracts generate a larger correlation between changes in earnings and hours for workers with lower earnings, hours, and tenure as in the data. I find that hours variability plays a quantitatively important role to increase workers’ output, which yields caution to policymakers when restricting hours variability. This is especially true for workers with high earnings and tenure who benefit the most from risk-sharing in implicit contracts.

I establish a fundamental equivalence between search-theoretic and preference-based approaches to modeling monopsony. When time discount-ing vanishes, the employment distribution from optimal job search in fric-tional monopsony models mirrors the labor allocation chosen by an agent with non-homothetic preferences, with the mean-min wage ratio emerging as a sufficient statistic for how wage inequality shapes employer substitutabil-ity. This equivalence challenges the conventional distinction between "fric-tional" and "taste-driven" monopsony, and yields practical insights for mea-surement and policy design. As an application, I derive and quantify a for-mula for welfare gains from employer entry that requires standard labor market statistics in search-based models but demands residual elasticity es-timates in taste-based models. This formula reveals that job creation can re-duce welfare below a critical unemployment threshold that rises with on-the-job search efficiency, with gains showing counter-cyclical patterns that peak during labor market downturns.

This paper demonstrates two distinct and quantitatively important biases introduced by using an “incorrect” definition of market boundaries when attempting to make inferences about labor market power. The first source of bias, long recognized in the antitrust literature, stems from mismeasurement of relative firm size: the same firm will appear artificially dominant when markets are drawn too narrowly and artificially competitive when they are drawn too broadly. We derive a novel second source of bias, which we term elasticity bias, that generates statistical attenuation of estimates of key parameters that govern model-based conclusions about the size and distribution of markdowns across employers and markets. In simulations calibrated to Brazilian administrative data, we show that the second channel is an order of magnitude more important than the first. Further, we show that market definition bias can be large in empirically-relevant cases where the relative rate of misclassification may be modest, as with administrative labor–market boundaries such as industry/occupation–region cells adopted by virtually all existing studies. We propose an alternative network-based procedure for defining labor market boundaries that extends the algorithm of Fogel and Modenesi (2022). Drawing upon the empirical strategy of Felix (2022), we show that relative to using administrative market definitions, using network-based market definitions yields estimates with 40% larger markdown dispersion and overturns several qualitative conclusions about which workers are harmed by monopsony power. Finally, we propose a simple diagnostic that allows practitioners to pick among off-the-shelf classifications when using a data-driven one is infeasible.

Employment protection legislation (EPL) typically varies across firms and establish- ments according to their size and age. This paper develops a search-and-matching model with multi-establishment firms which enables to separate direct effects of such rules from the general-equilibrium spillovers they induce. We apply this model to a 2009 Portuguese reform that restricted temporary contracts in young establishments of firms with more than 750 employees. Reduced-form estimates indicate: i) that temporary employment dropped in targeted firms without compensation via permanent contracts, causing direct employment losses; ii) positive effects benefiting non-targeted firms. Our model shows that these general equilibrium effects—albeit modest at the firm level—offset the direct negative impact at the aggregate level. Counterfactual analysis indicates that stricter EPL for large firms or establishments tends to raise total employment via indirect effects, whereas targeting older establishments reduces aggregate employment.

Who will gain and who will lose as AI automates tasks? While much of the discourse focuses on job displacement, we show that job transformation—a shift in the task content of jobs—creates large and heterogeneous earnings effects. We develop a quantitative, task-based model where occupations bundle multiple tasks and workers possessing heterogeneous portfolios of task-specific skills select into occupations by comparative advantage. Automation shifts the relative importance of tasks within each occupation, inducing wage effects that we characterize analytically. To quantify these effects, we measure the task content of jobs using natural language processing, estimate the distribution of task-specific skills, and exploit mappings to prominent automation exposure measures to identify task-specific automation shocks. We apply the framework to analyze automation by large language models (LLMs). Within highly exposed occupations, like office and administrative roles, workers specialized in information-processing tasks leave and suffer wage losses. By contrast, those specialized in customer-facing and coordination tasks stay and experience wage gains as work rebalances toward their strengths. Our findings challenge the common assumption that automation exposure equates to wage losses.

Student loan forgiveness has been proposed as a means to alleviate soaring student loan burdens. This paper uses administrative credit bureau data to study the distributional, con-sumption, borrowing, and employment effects of the largest event of student loan forgive- ness in history. Beginning in March 2021, the United States federal government ordered $132 billion in student loans cancelled, or 7.8% of the total $1.7 trillion in outstanding student debt. We estimate that forgiven borrowers’ predicted monthly earnings were $115 higher than borrowers who did not receive forgiveness and $193 more than the general population. We find that student loan forgiveness led to increases in mortgage, auto, and credit card debt by 9 cents for every dollar forgiven. Borrowers’ monthly earnings and em-ployment fell, at increasing rates for each month post forgiveness. The implied Marginal Propensities for Consumption (MPC) and Earnings (MPE) are 0.27 and -0.49, respectively.