How Substitutable Are Workers? Evidence from Worker Deaths

The fluidity of labor markets depends on the ease with which one side of the market can fulfill the needs of the other: whether workers can find employment that suits their skills and firms can find adequate substitutes for workers who leave. Today much is known about the worker’s perspective. A large body of empirical literature documents that workers suffer persistent earnings losses after they have been displaced from their job – in line with Becker’s (1962) idea that human capital has firm-specific components (see, e.g., Topel 1991; Jacobson et al. 1993; and Dustmann and Meghir 2005).

The other side of the market has, however, received a lot less attention in the empirical literature so that important questions remain.  What happens, for example, when a worker leaves his or her position? How easily can the firm replace the worker? How are other workers on the same team or in the same department or other divisions of the company affected? Which types of workers are hard to replace?

In my recent working paper, How Substitutable Are Workers? Evidence from Worker Deaths, I shed light on these issues by estimating how worker exits affect a firm’s demand for incumbent workers and new hires. In particular, I analyze so-called quasi-experiments by estimating the effect of unexpected deaths of workers on their firms’ hiring decisions as well as the wages of co-workers. This empirical strategy builds on previous studies that have used unexpected deaths as a source of variation (see, e.g., Bennedsen et al. 2006, Isen 2013, and Becker and Hvide 2017). My focus is on deaths that are arguably premature and unexpected, and so I exclude individuals who, e.g., were hospitalized or on long sick leave before their deaths.

I estimate the causal effect of unexpected worker deaths on hiring and on the remaining workers’ wages and retention rates based on the entirety of German Social Security records since 1975. I compare roughly 34,000 small firms that experienced the death of a worker in a given year to a group that has similar characteristics but did not experience a worker death that year. My analysis of the effect of worker exits focuses on the so-called “Mittelstand,” in particular small firms, which make up a large share of the German labor market.

To access the interdependencies between workers inside the firm and to understand heterogeneity in the effect of worker exits, I leverage detailed data on the deceased workers’ and the remaining incumbent workers’ occupations. In order to identify groups of workers inside a firm who work in jobs with a similar or distinct range of tasks, I use a horizontal classification into occupation groups based on the thematic focus, e.g., production or accounting. A vertical categorization based on the skill requirements of the occupation is used to identify workers in managerial and supervisory roles.

The data show that, from a broad perspective, worker deaths affect firms’ demand for the labor of their remaining workers. On average, incumbent workers experience a statistically significant increase in earnings of 170 Euros in the first year after the death of a colleague, which corresponds to an average increase of real earnings of about 0.6 percent. Over the course of the five years after the death, the average cumulative effect on the earnings of all incumbent workers in the same firm is close to 6,000 euros (2010 CPI), corresponding to about 18 percent of an average deceased worker’s annual earnings. Moreover, incumbent workers have a 0.5 percentage point higher probability of remaining employed at the same firm and are less likely to be employed at other firms.

However, there is a large amount of within-firm heterogeneity. A statistically significant increase in wages is only concentrated among incumbent workers in the same occupation group as the deceased. Their earnings increase on average by 240 Euros in the short run and by 170 Euros per year in the longer run. In contrast, when a manager passes away, I find that the wages of workers in other types of jobs actually go down in the short term. My results also show that longer-tenured workers and workers in more specialized occupations are harder to replace with outsiders.

Nevertheless, the intensity of these effects heavily depends on the characteristics of the labor market, specifically the size of the pool of potential replacements for the open position. For example, I find that incumbents’ wages respond less and external hiring responds more to a worker’s death when the external labor market in the deceased’s occupation is thick (when many workers in the local labor market possess the right qualifications)—and vice versa in cases of thin labor markets. For example, a biomedical firm seeking to replace one of their biomedical engineers has a higher chance of finding a replacement in a local labor market with a concentration of other biomedical firms. Importantly, my findings imply that the substitutability of incumbent workers and outsiders decreases the more specific the skills and experience of the exiting worker are.

From a theoretical perspective, the findings of positive average effects and of negative effects of manager and high-skilled worker deaths are consistent with bargaining models which allow for positive and negative wage effects depending on the degree of substitutability of different worker types in the firm’s production function (Stole and Zwiebel, 1996a,b). The bargaining position and wages of individual employees, in particular those with hard-to-replace skills, depend on the presence or absence of other workers at the same firm. The empirical findings are in line with predictions from the canonical Lucas (1978) model as it features substitutability of lower-skilled workers as well as a complementarity between managers and higher-skilled employees with the skills of other employees in the same firm

Taken together, my results show that firms face frictions in replacing workers externally as shocks to the firm’s labor supply affect the firm’s labor demand for the remaining workers. Based on the pattern of effects inside the firm, coworkers in the same occupation appear to be substitutes (as seen by the increase in their wages), while high-skilled workers and managers appear to be complements to workers in other occupation groups (wages go down for these workers). Furthermore, the heterogeneity in the effect by labor market thickness suggests that replacement frictions are strongest when worker’s human capital is highly specialized and firm-specific.

Some have hypothesized that labor market frictions arise primarily from the standard costs of searching for new employees (recruitment costs, time a vacancy is open, etc.), which effectively also temporarily increase the demand of current employees. This claim assumes that once hiring is complete, new workers immediately become insiders and the internal demand for other workers diminishes.

My findings, specifically the long term increase in wages for incumbent workers and the impact of labor market thickness on replacement, instead point to firm-specific human capital as the source of these frictions. Thus, the evidence suggests that it takes time for newly hired workers to accumulate specialized skills and become substitutes for their longer-tenured co-workers, as predicted by Becker (1962) or Lazear (2009). In particular, my findings show that a combination of specialized skills, and a thin local labor market for those skills, make workers hard to replace from the perspective of their employer.

Conceptually, my analysis therefore contributes to our understanding of the factors that make workers hard to replace. While my empirical analysis considered the effects of worker deaths, it seems plausible that my findings could be used to discern what happens in other situations such as when a worker quits or another firm hires away the worker. Moreover, for firms considering whether to invest in a new technology that is complementary to specialized skills, my results suggest that having access to a pool of appropriately skilled workers is vital.


Becker, Gary S. 1962. “Investment in Human Capital: A Theoretical Analysis.” Journal of Political Economy 70 (5):9-49.

Becker, Sascha O. and Hans K. Hvide. 2017. “Do Entrepreneurs Matter?”. Working Paper.

Bennedsen, Morten, Francisco Pérez-González, and Daniel Wolfenzon. 2006. “Do CEOs Matter?” NYU Working paper No. FIN-06-032.

Dustmann, Christian and Costas Meghir. 2005. “Wages, Experience and Seniority.” The Review of Economic Studies 72 (1):77-108.

Gibbons, Robert and Lawrence F. Katz. 1991. “Layoffs and Lemons.” Journal of Labor Economics 9:351-380.

Isen, Adam. 2013. “Dying to Know: Are Workers Paid Their Marginal Product?” University of Pennsylvania Working Paper.

Jacobson, Louis S., Robert J. LaLonde, and Daniel G. Sullivan 1993. “Earnings losses of displaced workers.” The American Economic Review: 685-709.

Jäger, Simon. 2015. “How Substitutable Are Workers? Evidence from Worker Deaths.” Working Paper.

Lazear, Edward P. 2009. “Firm-specific human capital: A skill-weights approach.” Journal Of Political Economy 117, no. 5: 914-940.

Lucas, Robert E. 1978. “On the Size Distribution of Business Firms.” The Bell Journal of Economics 9 (2):508-523.

Marshall, Alfred. 1890. The Principles of Economics. New York. Macmillian.

Stole, Lars A. and Jeffrey Zwiebel. 1996a. “Intra-Firm Bargaining under Non-Binding Contracts.” The Review of Economic Studies 63 (3):375-410.

-. 1996b. “Organizational Design and Technology Choice under Intrafirm Bargaining.” American Economic Review 86 (1):195-222.

Topel, Robert. 1991. “Specific Capital, Mobility, and Wages: Wages Rise with Job Seniority.” Journal of Political Economy 99 (1):145-176.

This post comes to us from Professor Simon Jäger at the Massachusetts Institute of Technology and the Institute on Behavior and Inequality (briq). It is based on his recent paper, “How Substitutable Are Workers? Evidence from Worker Deaths,” available here.