Determinants of postdisplacement reemployment outcomes and occupation changes

Abstract

We investigate the roles played by agglomeration (proxied by population density) and occupational distribution on postdisplacement employment outcomes for US workers, focusing on the probability of being employed, the likelihood of changing occupations, and the occupational distance between the old and new occupations (where occupational distance is defined as the difference in job tasks between two occupations). Overall, we find that the local occupational distribution has a greater impact on our outcome variables compared with our agglomeration measure. Our findings suggest policy makers should look closely at occupational distribution when designing job training programs and proposals for relocation vouchers.

Occupational concentration and outcomes for displaced workers

“…A two standard deviation increase in LLM employment decreases the probability of changing occupations by approximately 6.5 percentage points. These findings are qualitatively consistent with existing literature (Bleakley & Lin, 2012;Kosteas, 2019) and support the need to control for LLM size when examining the effect of occupational concentration on occupation switching. The present exercise also indicates that LLM size is not the only, or perhaps even the key characteristic of CBSAs affecting occupation changes.…”

“…Conversely, if the propensities to move and change occupations are negatively correlated, this will lead to an overestimation of the underlying causal effect. While existing research shows a negative relationship between local density and occupation changes for displaced workers (Bleakley & Lin, 2012;Kosteas, 2019) we do not have a clear indication, controlling for LLM size and/or density, whether workers with a greater underlying willingness to change occupations tend to move to more or less occupationally concentrated labour markets (or if there is any underlying relationship at all). Similar concerns hold for the employment equation.…”

“…While urban economists have traditionally focused on measures of total labour market size (generally proxied by population) or density (proxied by population density) when examining occupational switching, recently published work indicates the local distribution of occupations may be a more important determinant of postdisplacement employment outcomes (Kosteas, 2019). The present paper presents evidence that the occupational concentration of the local labour market (LLM) is an important determinant of post-displacement employment outcomes.…”

“…Since the main focus of the estimation section is to identify whether occupational concentration affects postdisplacement employment outcomes, it is important to build an empirical model that accounts for other aspects of the LLM that may be correlated with both the outcome variables and occupational concentration. Given the strong correlation between agglomeration and occupational concentration (ρ = -0.56 in the DWS data) and the established connection between agglomeration and both employment (Di Addario, 2011;Kosteas, 2019) and occupation changes (Bleakley & Lin, 2012;Kosteas, 2019), all models include the log of CBSA employment as a control variable. Given the established relationship between the relative importance of occupational employment in the LLM and employment outcomes (Bleakley & Lin, 2012;Kosteas, 2019), each model also controls for the occupation share of CBSA employment for the pre-displacement job to ensure that the estimates are not picking up effects based on the correlation between occupational distribution and occupation shares (ρ = 0.098 in the DWS data).…”

“…In general, existing studies focus on industry or occupation specific measures of the LLM. Using the 2004-2018 waves of the Displaced Worker Survey (DWS), Kosteas (2019) finds that occupational distribution in the MSA of employment has a greater impact on occupation changes compared with agglomeration (where the latter is proxied by MSA population density). That paper introduced the average "distance" between the worker's previous occupation and all jobs in the metropolitan area (where distance is defined as the difference in job tasks) as a measure of occupation distribution.…”