Firms’ Wage-Setting Power: A New Take on Monopsony in the Labor Market

Why do so many firms seem to pay workers less than what their labor is truly worth? Economists call this phenomenon “monopsony power”—the ability of firms to set wages below the marginal product of labor, which is the increase in output from adding one additional worker. Traditionally, this has been explained in two main ways: either because workers face costly and time-consuming job searches, or because jobs themselves differ in ways that matter to workers, like location, schedule or prestige.

But what if these explanations are actually two sides of the same coin?

In a new paper, Anton Cheremukhin from the Federal Reserve Bank of Dallas and I developed a unified theory of monopsony that brings these perspectives together.For more details, see our May 2025 Dallas Fed working paper “An Information-Based Theory of Monopsony Power.” We built a model in which both search frictions and job differentiation emerge from a deeper source: limited information. In this framework, workers and firms don’t know everything about the other side of the market; they make strategic, probabilistic choices about whom to search for and how much effort to spend doing so. The result is a much richer and more realistic picture of how wage-setting power arises.

A Smarter Search Theory

We drew inspiration from models of rational inattention that we originally developed in the context of dating and marriage markets and applied them to the labor market.Models of rational inattention are characterized by targeted search under information constraints. For example, see the Anton Cheremukhin, Paulina Restrepo-Echavarria and Antonella Tutino’s January 2020 article, “Targeted Search in Matching Markets,” in the Journal of Economic Theory. Here, both firms and workers decide how precisely to target their ideal matches, constrained by the cost of acquiring and processing information. The more targeted the search, the more effort it takes. These frictions are modeled mathematically using tools from information theory (specifically, the Kullback-Leibler divergence), but the intuition is simple: Workers can’t search everywhere, and firms can’t evaluate everyone.

This approach allowed us to seamlessly connect two classic ideas in labor economics. On one end of the spectrum, we replicated models in which search is entirely directed and strategic. On the other end, we recovered the classic random search world. But in between lies something new: a continuum of imperfect targeting, where firms must guess how many—and which—workers will apply, and workers weigh the costs of precision.

Five Sources of Monopsony Power

One of our key contributions is a new wage equation that reveals five distinct sources of monopsony power:

1. Search costs on the worker side, which reduce how responsive workers are to higher wages
2. Screening costs on the firm side, which limit how many applicants firms are willing to evaluate
3. Labor market tightness, which magnifies the effects of search and screening frictions
4. Sorting patterns, where high-quality workers match with high-paying firms, reducing competition
5. Sequential search, where firms can strategically post wages before workers begin searching

These forces interact in subtle ways. For example, even when workers are highly responsive to wages (i.e., labor supply is elastic), firms may still exercise wage-setting power if they face little competition for those workers because of screening bottlenecks or sorting.

Implications of Monopsony Power for Wages

We calibrated the model with realistic parameters and found that even moderate information frictions can produce large wage markdowns—30% to 40% lower than workers’ marginal product. This is consistent with recent empirical findings. But from a societal perspective, this outcome is inefficient. Based on our model, a social planner who takes into account congestion effects and limited information would recommend a mix of higher wages and less search effort, leading to wage markdowns of 10% to 15%.

When More Sorting Means More Market Power

The model also sheds new light on a longstanding debate in economics: When is it efficient to sort workers and firms by type? While positive assortative matching (i.e., matching high-skilled workers with high-productivity firms) tends to increase output, we found it can also increase monopsony power. Why? Because it reduces wage competition; if every firm gets “its own” type of worker, it no longer has to outbid rivals. For example, the tech industry in Silicon Valley has access to a concentrated pool of high-skilled workers with limited options outside that area. As a result, these firms may exert a degree of monopsony power.

Interestingly, more competitive wage outcomes arise in equilibria where matching is mixed or less sorted. These outcomes may not always be efficient in terms of total output, but they often distribute income more fairly. This result challenges the conventional wisdom that more sorting is always better for the economy.

The Big Picture: A Richer Model of Monopsony

Our work offers a fresh, integrated take on monopsony that goes beyond the classic supply elasticity story. We argue that wage-setting power is fundamentally about information and strategy: how much workers and firms know, how much it costs to know more, and how they behave under uncertainty.

In doing so, we respond to recent calls for richer models of monopsony that reflect the realities of modern labor markets, where matching is imperfect, firms are strategic, and information is scarce.

Notes

1. For more details, see our May 2025 Dallas Fed working paper “An Information-Based Theory of Monopsony Power.”
2. Models of rational inattention are characterized by targeted search under information constraints. For example, see the Anton Cheremukhin, Paulina Restrepo-Echavarria and Antonella Tutino’s January 2020 article, “Targeted Search in Matching Markets,” in the Journal of Economic Theory.