Displacements (e.g. layoffs) in the US affect many participants in the labour market. As an example, during the height of the Great Recession, around seven million workers with at least three years of tenure experienced a job loss due to layoff (Bureau of Labour Statistics 2010). In conjunction with this high incidence of displacement, there exists long and distinguished literature documenting large and persistent earnings losses associated with displacement (see, for example, Jacobson et al. 1993, and Davis and von Wachter 2011, henceforth DV). Although estimates differ, economists typically find that even 20 years after displacement, annual earnings are 10 to 20% below pre-displacement earnings.
Classical labour market theories fall woefully short in explaining this phenomenon as all workers receive the market-clearing wage, and there is no involuntary unemployment. This implies that displaced workers’ earnings should recover immediately. The standard workhorse model of frictional unemployment (Mortensen and Pissarides 1994) also implies small earnings losses as observed average job-finding rates in the US are high, and all workers receive identical wages when employed (DV). Furthermore, although DV show that a model that features cross-sectional wage dispersion (Burgess and Turon 2010) can produce nontrivial earnings losses, this framework cannot quantitatively account for the depth and persistence of observed displaced worker earnings losses. The literature sees the inconsequential nature of job loss in these models as a major shortcoming, hindering economists’ understanding of why high unemployment creates concern among policymakers.
My job market paper proposes a model in the spirit of Jovanovic’s (1979, 1984) matching model where workers displaced from relatively stable jobs take time to find similarly high-quality matches and encounter substantial risk of subsequent job loss following an initial displacement. These simple features can account for the vast majority of the depth and persistence of displaced worker earning losses (Krolikowski 2014).
Reduced employment and lower wages
As far as post-displacement employment goes, the framework captures the following intuition. Compared to their stable job prior to the job loss, workers might not be as well matched in their first job coming out of non-employment. This results in tentative new employment relationships that are subject to a high probability of ending quickly. This leads to workers’ ‘spinning their wheels’ for a few years as they face repeated separations into non-employment after the initial displacement. This serial correlation in separations coincides with previous empirical work, where multiple additional job losses are found to be an important part of the workers’ post-displacement experiences (Stevens 1997), and I document the same phenomenon using data from the Panel Study of Income Dynamics (PSID).
The second part of the explanation for protracted post-displacement earnings dynamics is the presence of match-specific human capital and on-the-job search. This means that workers randomly get outside employment offers while working, and can move to better paying jobs over time. These features of the model deliver a ‘job ladder’ whereby newly hired workers start out in jobs with relatively low wages and move up the wage distribution via search on the job. This concept prolongs earnings recovery after displacement as non-employed workers enter poor employment relationships and search for better matches while employed.
When choosing the length of this job ladder, I use data on the amount of residual wage dispersion in the economy. Hornstein et al. (2007) measure residual wage dispersion within local geographic areas and 3-digit occupations, controlling for gender, race, education, and experience.1 Although estimates vary, these authors find that, even after controlling for various factors, similar workers can be paid very different wages. In the model, I attribute all of this residual wage dispersion to frictions in the labour market.
Matching worker outcomes
Figure 1 shows the average probability of separating into non-employment by tenure.2 Individuals with relatively little tenure experience monthly separation probabilities of around four percent. Those with high levels of tenure have markedly lower separation probabilities, around one-half percent. The model delivers these observed separation probabilities, and this is crucial to the model’s success. Displaced workers finding new jobs (with low tenure) experience markedly higher separation probabilities, and this hampers their earnings recovery.
Figure 1. Average employment-to-non-employment probabilities by tenure
Figure 2 presents a comparison between the earnings of displaced workers from the baseline model and the results from DV. The outcome is very encouraging, with the baseline model delivering an earnings trajectory that closely resembles the empirical counterpart.
Figure 2. Annual earnings around displacement
Aside from matching the earnings losses of displaced workers, the model accurately captures the decomposition of lost earnings into reduced employment and lower wages. In particular, on impact, the model predicts that the majority of earnings losses are because workers are out of employment. Much of the short-run recovery in earnings can be attributed to workers finding new jobs, while wages, through the job ladder, explain all the long-run earnings losses. This is entirely consistent with the available evidence on displaced workers (e.g. Topel 1990, and Bender et al. 2009).
Figure 3. Earnings decomposition: Employment and wages
Economists are well aware that worker displacement generates large and persistent earnings losses. This column provides evidence that a theory of job matching, with workers finding poor-quality matches after non-employment and taking substantial time to find high-quality employment, can account for the vast majority of displaced worker earnings losses. The framework implies two policy levers for mitigating individual earnings effects of displacement:
- Getting workers re-employed quickly to jump-start the process of finding a good match. This could be implemented using hiring tax credits or wage subsidies to stimulate the demand side, alongside traditional active labour market policies like disseminating vacancy information, and providing guidance with interviewing and writing quality CVs to buttress the supply side.
- Helping those out of work to find better matches coming out of non-employment. This suggests support for increasing unemployment insurance (UI) generosity, which should generate more stable post-separation employment (see Centeno 2004, for empirical evidence that greater UI generosity leads to longer job tenure).
Bender, S, J F Schmieder, and T M von Wachter (2009), “The Long-Term Impact of Job Displacement in Germany During the 1982 Recession on Earnings, Income and Employment”, Columbia University, Department of Economics Discussion Paper Series DP0910-07.
Bureau of Labor Statistics (2010), “Worker Displacement: 2007-2009”, News release. USDL-10-1174.
Burgess, S M, and H Turon (2010), “Worker Flows, Job Flows and Unemployment in a Matching Model”, European Economic Review, 54: 393-408.
Centeno, M (2004), “The Match Quality Gains from Unemployment Insurance”, Journal of Human Resources, 39(3): 839-863.
Davis, S J, and T M von Wachter (2011), “Recessions and the Costs of Job Loss”, Brookings Papers on Economic Activity, Fall: 1-72.
Hornstein, A, P Krusell, and G L Violante (2007), “Frictional Wage Dispersion in Search Models: A Quantiative Assessment”, NBER Working Paper #13674.
Jacobson, L S, R J Lalonde, and D G Sullivan (1993), “Earnings Losses of Displaced Workers”, The American Economic Review, 83(4): 685-709.
Jovanovic, B (1979), “Job Matching and the Theory of Turnover”, Journal of Political Economy, 87: 972-990.
Jovanovic, B (1984), “Matching, Turnover and Unemployment”, Journal of Political Economy, 92: 108-122.
Krolikowski, P (2014), “Job Ladders and Earnings of Displaced Workers”, Working Paper, University of Michigan.
Mortensen, D T, and C A Pissarides (1994), “Job Creation and Job Destruction in the Theory of Unemployment”, Review of Economic Studies, 61(3): 397-415.
Stevens, A H (1997), “Persistent Effects of Job Displacement: The Importance of Multiple Job Losses”, Journal of Labor Economics, 15(1): 165-188.
Topel, R H (1990), “Specific Capital and Unemployment: Measuring the Costs and Consequences of Worker Displacement”, Carnegie-Rochester Conference Series on Public Policy, 33: 181-214.
1 They restrict their analysis to full-time workers (35-45 hrs/week) and full-year workers (48-52 weeks/yr) to get at measurement error issues. They attempt to look at very low-skilled occupations where occupation and firm-specific skills are arguably not very important to address the issue of individual-specific differences in cumulated skills not perfectly correlated with experience.
 The analysis uses monthly data from the PSID for the years 1988-1997. These data provide observations on tenure, how often workers switch jobs, how often they find jobs when out of work, and how often they lose jobs.