1 Introduction
Shopping for jobs is an integral part of workers’ careers. Young workers enter the labor mar-
ket poorly matched, and gradually relocate across firms to find better matches. Topel and Ward
(1992) find that this process of employer-to-employer (EE) reallocation toward higher paying jobs
accounts for a third of workers’ wage growth during the first 10 years of their careers. More re-
cently, macroeconomists have stressed that such reallocation also plays a critical role for aggregate
economic performance, by reallocating workers from less to more productive firms (Lentz and
Mortensen, 2012; Moscarini and Postel-Vinay, 2017). Yet, despite its importance for both micro
and macroeconomic outcomes, little is known about long-run trends in such reallocation.
The reason is twofold. First, the data required to measure the frequency at which workers
move from one employer to another without an intervening spell of nonemployment are only
available since the mid-1990s. Furthermore, the series available from the main labor force sur-
vey in the U.S., the Current Population Survey (CPS), suffers from bias arising from changes in
survey methodology and non-response over time (Fujita, Moscarini and Postel-Vinay, Forthcom-
ing).
1
Second, many EE transitions in the raw data are toward lower paying jobs (Tjaden and
Wellschmied, 2014; Sorkin, 2018). Of particular interest for macroeconomic performance, how-
ever, is EE mobility toward higher paying, more productive jobs—what we henceforth refer to as
allocative EE mobility (Bilal et al., 2022; Elsby and Gottfries, 2022).
2
This paper proposes a methodology that overcomes these dual challenges, and uses it to docu-
ment three trends in allocative EE mobility over the past half century in the U.S. based on publicly
available micro data from the CPS. First, allocative EE mobility fell in half between 1979 and 2023.
Second, ceteris paribus the decline reduced annual wage growth by over one percentage point.
Third, the fall was particularly pronounced for women, those without a college degree, and young
workers. In terms of potential explanations, we find little support for the notion that allocative EE
mobility is lower today because workers are better matched with their current employers (Mer-
can, 2017; Pries and Rogerson, 2022) or the labor market is worse at matching workers and firms.
Instead, based on long-run variation across U.S. states, we provide evidence consistent with the
view that greater labor market concentration reduced workers’ opportunities to transition toward
higher paying employers (Bagga, 2023; Berger et al., 2023; Jarosch, Nimczik and Sorkin, 2024).
Our point of departure is a prototypical job ladder model in the spirit of Burdett and Mortensen
(1998). In each period, non-employed and employed workers receive job offers with some exoge-
nous and potentially different probability. A job offer is a draw of a wage from an exogenous
wage offer distribution, which may also vary by employment status (Faberman et al., 2022). If the
worker accepts the job, she supplies a unit of labor at the specified wage until either she finds a
1
An alternative, arguably superior, series is available from the Survey of Income and Program Participation (SIPP), but
high-quality data are only available for 1996–2013.
2
We borrow this terminology from the literature, but caution that it should be interpreted from the perspective of a
social planner who does not face any frictions and who aims to maximize aggregate output.
2