Sorting along Business Cycles

Paweł Gola, Haozhou Tang · 2026

Read Sorting along Business Cycles by Paweł Gola, Haozhou Tang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-09323.txt.

Introduction

Paweł Gola and Haozhou Tang build a tractable model of labor markets with varied workers and firms, where hiring sorts one firm to many workers. Firms choose both headcount and staff mix, shaping worker incomes and firm productivity together. They study cycles driven by market efficiency shocks that favor more productive firms the most. The model's predictions match known facts about how wage and productivity distributions move over the cycle.

Catalog record

CREATORPaweł Gola, Haozhou Tang
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE80858
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.09323

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Sorting along Business Cycles Thanks: We thank seminar and conference participants at Dongbei University of Economics and Finance, and Liaoning University. Gola gratefully acknowledges the support of the UKRI Horizon Europe Guarantee (Grant Ref: EP/Y028074/1). The usual disclaim applies. Paweł Gola, Haozhou Tang Abstract We develop an analytically tractable model featuring heterogeneous workers and firms, where labor markets clear through a one-to-many sorting mechanism. Firms determine both the number and composition of their employees, shaping (1) the income distribution among workers and (2) the productivity distribution across firms. We study business cycles driven by market efficiency shocks that disproportionately benefit more productive firms. The model’s implications are consistent with empirical regularities on the cyclical behavior of wage and productivity distributions. Keywords: sorting, wage inequality, productivity, business cycles JEL Classification: C78, J21, D24, D31, E32 1 Introduction A prevailing convention in business-cycle studies is to treat firm-level productivity as exogenous, determined outside the firm’s recruitment decisions. Firms are typically modeled as passive takers of their productivity draws, making hiring and investment decisions conditional on those primitives. Yet this view abstracts from a fundamental aspect of how firms actually operate: productivity is shaped by whom firms hire. A firm’s capabilities hinge on the talents it brings in—whether a CEO who sets strategic direction, a CFO who manages financial structure, R&D personnel who drive innovation, or sales teams who determine market reach. In many cases, hiring decisions are not responses to productivity; they are determinants of it. This perspective highlights that productivity is an equilibrium outcome of how heterogeneous firms and workers sort in the labor market. Recognizing productivity as an equilibrium sorting outcome raises natural questions: how are the distribution of wages and firm-level productivity connected, and how do they evolve over the business cycle? A salient feature of the productivity distribution is that its dispersion is countercyclical---widening in recessions and narrowing in booms. 1 1 1 See, for example, 4 , 27 , 6 , and 15 . This variation is typically taken as exogenous, often attributed to uncertainty or volatility shocks. In contrast, wage inequality, measured as the variance of log wages, exhibits a distinct cyclical pattern. As shown in Figure 1 , the left panel reveals a long-run upward trend in wage dispersion from 40 , along with clear cyclical movements: wage inequality tends to rise in economic expansions and contract in recessions, highlighted by the NBER-dated shaded regions. The right panel, which plots detrended wage inequality and detrended GDP, makes the comovement more transparent---inequality tends to track the business cycle. Notably, the patterns differ across episodes. For instance, in the late 1990s, this comovement was less pronounced: wage inequality remained below its own trend despite the growth of real GDP. By contrast, wage inequality dropped sharply during the Great Recession. 2 2 2 Throughout this paper, we focus on wage inequality rather than income inequality, which includes all sources of income. The figure uses wage income, i.e., labor earnings, and thus only accounts for employed individuals, whereas income inequality also includes the unemployed—potentially leading to very different patterns. Additionally, different definitions and measures of income inequality can produce substantially different results. See, for example, 34 and 24 . Overall, the data suggest a potential contrast between the two dispersions: while the dispersion of wages—reflecting worker performance—tends to increase in expansions, the dispersion of firm-level productivity—reflecting firm performance—tends to rise in downturns. These patterns highlight an important feature for understanding firm-worker interactions over the business cycle. Figure 1 : Wage inequality over business cycles Notes: Shaded areas indicate NBER recessions. Wage inequality is measured by variance of individual log wages. Both series in panel (b) are detrended. Source: 40 and authors’ calculations. Motivated by these patterns, we introduce a framework with heterogeneous firms and workers in which labor markets clear through a one-to-many matching mechanism. Firms differ in their underlying types, and workers differ in their skill levels. Firms choose both the number and type of their employees, shaping the distributions of wages and productivity, which exhibit Pareto tails in equilibrium. Apart from the introduction of worker heterogeneity and one-to-many sorting, the model is entirely standard, featuring CES demand and decreasing-returns Cobb-Douglas production technology. Crucially, productivity emerges from the equilibrium sorting between firms and workers rather than being imposed exogenously. In this sense, the model generalizes the standard heterogeneous-firm framework: without worker heterogeneity or sorting, it reduces to the canonical setting where productivity is given. Sorting between firms and workers is sensitive to aggregate conditions: changes in the macro environment can shift the composition of matches and thus alter the distributions of wages and productivity. We introduce a reduced-form market efficiency shock that operates through time-varying firm-specific wedges on labor and capital prices, representing distortions that fluctuate over the business cycle. This reduced-form shock may capture the compound effects of several underlying structural shocks and disproportionately affects high-type firms. It is designed to generate business-cycle fluctuations that reproduce procyclical reallocation of labor to more advanced firms, a key cyclical pattern observed in the data. 3 3 3 See, for example, 32 , 22 , 13 , and 23 . We analytically characterize how the reduced-form market efficiency shocks shape the cross-sectional distributions of wages and firm-level productivity. A positive shock induces high-type firms to expand more aggressively, shifting the composition of jobs toward higher-quality positions and thereby widening wage dispersion. Simultaneously, due to this relative expansion, these firms are matched with relatively lower-type workers, pulling their realized productivity closer to that of low-type firms, which shrink in relative size, and thus dampening the dispersion of firm-level productivity. This mechanism demonstrates how shifts in the allocation of talent across firms can generate opposing cyclical movements in wages and productivity, even when underlying technologies are unchanged. It is consistent with the findings in 13 , who document that during downturns high-rank workers are more likely to work at low-rank firms, while low-rank workers are less likely to be employed by high-rank firms. While the mechanism described above can generate opposing cyclical movements in wage and productivity dispersions, not necessarily all recessions exhibit this disparity. To understand the broader range of patterns, we also examine the effects of several structural shocks. In particular, we consider aggregate productivity shocks and second-moment shocks that alter the variance of firm types and firm-level distortions. This allows us to analyze how different shocks shape the joint dynamics of wage and productivity dispersions over the business cycle. Literature review The main conceptual contribution of this work is the finding that, under reasonable assumptions, changes in the matching function have opposing effects on the distributions of wages and measured productivity. In particular, the result that improvements in workers’ matches decrease the inequality of measured productivity is, to the best of our knowledge, completely novel. Combined with the well-known fact that, in a model with worker-firm type complementarities, improvements in workers matches increase wage inequality ( 11 ) , this implies our main result: cyclical fluctuations in sorting cause the procyclical fluctuations of wage inequality and the counter-cyclical behavior of productivity dispersion. Our contributions to the literature on labour market sorting extend beyond these novel comparative statics. First, while frictionless one-to-many sorting models (see 16 for the most comprehensive

Read the complete text