We show that monopsony power maps directly to the strength of workers' comparative advantage with their employers.  Provided absolute advantage at an employer aligns with comparative advantage, employers mark down wages especially for their more productive workers, reducing wage dispersion. We show evidence this is indeed the case for the Brazilian formal labor market. Calibrating the model to Brazil, we examine the implications of monopsony power for wage inequality, the allocation of workers between formal and informal work, and the effectiveness of minimum wage policy.

This paper studies firm diversification over 6-digit NAICS industries in U.S. manufacturing. We find that firms specializing in fewer industries now account for a substantially greater share of production than 40 years ago. This reallocation is a key driver of rising industry concentration: specialized firms have displaced diversified firms among industry leaders---absent this reallocation concentration would have decreased. We then provide evidence that specialized firms produce higher-quality goods: specialized firms tend to charge higher unit prices and are more insulated against Chinese import competition. Based on our empirical findings, we propose a theory in which growth shifts demand toward specialized, high-quality firms. We conclude that one should expect rising specialization in a growing economy.  

Do larger firms operate in larger markets? This paper studies the sorting between firms and locations in U.S. retailing. I decompose retailers' production in a particular location into firm- and location-fixed effects (FE). A firm is assigned a larger FE if it produces more than other firms in the same markets; a location is assigned a larger FE if firms tend to produce more in this market than they do in other markets. I find a U-shaped relation: large- and small-FE retailers tend to operate in large-FE markets, while middle-FE retailers tend to operate in small-FE markets. I argue that the U-shaped sorting points toward models where profits depend on firms' market share in their operating locations, such as Kimball demand or oligopolistic competition. By incorporating the U-shape sorting, these models alter previous studies' predicted implications of the spatial expansion in retailing and placed-based policies that target small markets.

We argue that a substantial portion of the downward historical trend in U.S. research productivity is a composition effect resulting from self-selection in researcher ability and the dramatic expansion of the research sector. We estimate said effect with a Roy-like model of researcher supply using 1960-2021 U.S. microdata. Our results suggest that average researcher ability has decreased by half as a result of selection. The decline in efficiency-unit adjusted researcher productivity is therefore half of that documented in the literature. Embedding our model of researcher supply within a standard semi-endogenous growth model, we show that accounting for selection can lead to substantial upward revisions in model-predicted long-run growth rates.

Given worldwide trends in education, wage premium for schooling, and real GDP, we derive a lower bound for the long-run elasticity of labor substitution across schooling groups of around 4, which is far higher than values commonly used in the literature. We exploit our bound to reexamine the importance of human capital in cross-country income differences, including the roles of school quality versus the skill bias of technology in the greater efficiency gains from schooling in richer countries.