Working Papers

How much wage dispersion is attributed to workers, firms, and their sorting depends on how wages are adjusted for observed characteristics. Standard AKM decompositions impose a known linear adjustment. We develop Generalized AKM, a framework that permits an unknown smooth covariate function and group-specific nonlinear interactions while preserving the original variance components. We prove consistency and asymptotic normality with heteroskedastic errors and many fixed effects, and characterize the stronger smoothness required for quadratic forms. In Portuguese employer--employee data, adding worker and firm-input controls lowers the bias-corrected worker-effect variance from 0.551 to 0.474 of total wage variance, firm-effect variance from 0.144 to 0.121, and sorting from 0.080 to 0.047. Across three group-specific nonlinear bases, firm-effect variance remains between 0.114 and 0.117 and sorting between 0.041 and 0.042, while worker-effect variance ranges from 0.474 to 0.491. Which controls enter matters more for firm variance and sorting than how flexibly they enter; worker variance remains more sensitive to the basis.

We develop a model of staged entry: to operate, monopolistically competitive firms must pay two sequential entry costs, each time acquiring a more informative signal of future performance. This model yields two implications for fiscal policy. First, the equilibrium outcome is constrained-efficient if preferences are CES and entry costs are exogenous. Second, when entry costs depend on how many firms pass any entry stage (due to positive knowledge spillovers or negative congestion effects) the resulting externalities can be offset via Pigouvian taxes or subsidies that are timed around the relevant entry decisions. A calibration exercise based on U.S. firm-entry data shows that these policies would raise welfare through a wider pool of entrants and sharper selection of productive firms.

Men experience workplace injuries at twice the rate of women. This paper shows that compensating differentials for non-fatal injury risk contribute to gender differences in firm pay policies. We combine Italian matched employer--employee data with individual injury records and a search model that decomposes firm wage effects into productivity and injury-risk components. Women receive 86 percent of men's wage response to firm-level injury risk. Injury-related channels explain 8 percent of the gender gap in firm wage effects, rising to 17 percent in manufacturing---comparable to other estimates for single amenities. Within firm--occupation cells, men and women price a given task hazard similarly, so the firm-level injury channel operates mainly through sorting across firms and gender differences in exposure to risky tasks within firms.

What happens when employers value worker welfare in frictional labor markets? We show this “responsibility” creates an endogenous wedge in the marginal labor cost—akin to a hiring subsidy—altering wage and vacancy incentives rather than only changing the surplus split. The wedge is strongest when outside options are weak and separations rare, implying larger wage premia in slack, low-mobility markets. In a wage-posting model with on-the-job search, responsible firms may occupy the high-wage segment even when less productive. In a DMP model, responsible firms commit to higher worker bargaining power, raising the value of unemployment and thereby wages at regular firms.

Do workers consider a firm’s “springboard” value in terms of future job opportunities when choosing an employer? Using a search model of the labor market, I introduce the idea that firms differ in enhancing their employees’ chances of receiving external job offers. The model informs a firm-level proxy for outside job offers received by workers. This measure empirically aligns with key model predictions: 1) it negatively correlates with both firm-specific tenure and young workers’ entry salaries, revealing a compensating differential; and 2) it suggests that workers enjoy a salary premium upon leaving such firms, indicative of faster career progression. The model is estimated on administrative data from Italy and successfully captures key aspects of labor market dynamics. My channel explains 10% of the overall job-to-job transitions and shows how firm-induced variation in job search can be a significant driver of inequality, especially at the bottom of the wage distribution.

We document how a reduction in labor costs led to heterogeneous effects on manufacturing firms’ total factor productivity (TFP). Leveraging an Italian labor legislation reform and unique institutional features of the local collective bargaining system, we show that such effects vary along the TFP distribution. Relative to the counterfactual, TFP markedly declines on the left tail, which we explain via selection mechanisms; on the right, TFP mildly increases as firms are able to expand and reallocate their workforce. We develop a general equilibrium model featuring firm selection and frictions in input markets to guide the evaluation of welfare implications.

How much do worker skills, firm pay policies, and their interaction contribute to wage inequality? Standard approaches rely on latent fixed effects identified through worker mobility, but sparse networks inflate variance estimates, additivity assumptions rule out complementarities, and the resulting decompositions lack interpretability. We propose TWICE—Tree-based Wage Inference with Clustering and Estimation—a framework that models the conditional wage function directly from observables using gradient-boosted trees, replacing latent effects with interpretable, observable-anchored partitions. This trades off the ability to capture idiosyncratic unobservables for robustness to sampling noise and out-of-sample portability. Applied to Portuguese administrative data, TWICE outperforms linear benchmarks out of sample and reveals that sorting and non-additive interactions explain substantially more wage dispersion than implied by standard AKM estimates

What drives lifetime income inequality? We estimate a continuous-time lifecycle model incorporating human capital accumulation, innate ability, search frictions, job wage premia, and non-wage amenities. Workers make consumption-savings decisions, and their asset position affects job acceptance. We estimate the model on Danish matched employer-employee data for 2008–2023. Counterfactuals show that innate ability is the dominant source of lifetime earnings inequality, accounting for over 90 percent of the variance. Search frictions and amenities matter for welfare rather than earnings: removing frictions reduces lifetime utility inequality by nearly 80 percent while leaving earnings inequality unchanged. Finally, several mechanisms have opposite effects on cross-sectional versus lifetime inequality, showing that the two measures provide different information.

Which workers receive the smallest share of their contribution to production? We estimate individual markdowns in matched Portuguese worker–firm data for 2004–2023 by comparing production claims, estimated without pay data, with annualized earnings. Firm-year averages miss almost one third of log-markdown variation. Persistent firm differences account for 72 percent of log-markdown variance and worker differences for 28 percent. Markdown differences relative to coworkers persist after workers change employers. Worker and firm components partly offset: workers with high markdown components tend to work at firms with low ones. Grouping firms into broad classes loses most of this negative sorting. A rise in firm-specific export demand initially increases claims more than pay, temporarily widening markdowns.

We develop an OLG model with heterogeneous agents and aggregate uncertainty to study optimal Ramsey taxation when the government can use a credible set of social security instruments. Social security mitigates the income effect in optimal labor tax smoothing and, together with heterogeneity, adds new redistributive motives to both labor and capital taxes while crowding out others. We calibrate the model on three different economies: the US, the Netherlands, and Italy. We argue that the three countries would experience heterogeneous gains, in redistributive and efficiency terms, by moving from the status-quo allocations to those prescribed by a utilitarian Ramsey planner. Our simulations show that retirement benefits in the current economies are higher than their Ramsey-optimal level while we argue that the use of funded social security schemes, neglected in current actual policies, could be welfare improving.

In local labor markets, a pattern often emerges where workers transition during the early stages of their careers from lower-paying firms that provide comprehensive training to higher-paying specialized firms that predominantly employ already-trained workers. We refer to this mechanism as the “Human Capital Value Chain” (HCVC). We study and document its impact on the trajectory of workers' wages and local agglomeration externalities, thereby highlighting its role in the broader labor market dynamics.