Turnover Costs: Evidence from Unexpected Worker Separations
with Pierre Cahuc, Simon Jäger, and Rune Vejlin
(Draft under revision, old draft available upon request)
with Pierre Cahuc, Simon Jäger, and Rune Vejlin
(Draft under revision, old draft available upon request)
Abstract
We analyze the effects of shocks to firms’ employment to provide a novel estimate of turnover costs drawing on data for Danish firms and workers. Unexpectedly losing an employee due to death lowers profits in the subsequent four years by EUR 41,000.
Our estimates of turnover costs are an order of magnitude larger than previous evidence based on firms’ surveys of hiring and training costs. The larger estimates are driven by the imperfect replaceability of incumbent workers with outsiders. These manifest themselves as negative effects on value added even after employment has recovered as well as in higher labor costs to retain the remaining workers. Our evidence helps to resolve several crucial puzzles in search and matching models, which struggled to reconcile seemingly low turnover cost estimates with several other empirical regularities.
Our estimates of turnover costs are an order of magnitude larger than previous evidence based on firms’ surveys of hiring and training costs. The larger estimates are driven by the imperfect replaceability of incumbent workers with outsiders. These manifest themselves as negative effects on value added even after employment has recovered as well as in higher labor costs to retain the remaining workers. Our evidence helps to resolve several crucial puzzles in search and matching models, which struggled to reconcile seemingly low turnover cost estimates with several other empirical regularities.
Employment Reallocation over the Business Cycle: Evidence from Danish Data
with Rune Vejlin and Henning Bunzel.
Pdf
with Rune Vejlin and Henning Bunzel.
Pdf
Abstract
We present new evidence on how employment growth varies across firm types (size, productivity, and wage) and over the business cycle using Danish data covering almost 30 years. We decompose net employment growth into two recruitment margins: net hirings from/to employment (poaching) and net hirings from nonemployment. High-productivity firms are the most growing firms due to poaching. High wage firms poach almost as many workers, but shed an almost equal amount to non-employment. Large firms do not poach workers from smaller firms. In terms of employment cyclicality, we find that low-productive and low-wage firms shed proportionally more jobs in recessions. We relate our findings to recent models of employment fluctuations that jointly analyze worker and firm dynamics.
Abstract
This article studies reasons for internal hiring, i.e., re-assigning or promoting an employee instead of recruiting an external candidate. We exploit a representative survey of establishments covering all European Union countries to measure employers’
search for internal candidates. Internal search is a widespread practice: 66% of establishments typically search internally. The accumulation of specific skills and the provision of incentives for employees are the main advantages of hiring internally
in the theoretical literature. Ordered probit estimates show that on-the-job training and internal search are positively associated. On the contrary, incentive schemes such as variable pay are not associated with employer search. These results help to assess competing theories, and in particular, suggest that specific human capital is an important driver of internal hiring. Finally, we uncover two interesting facts that need further research: internal hiring is less likely in service firms and in noncompetitive product markets.
search for internal candidates. Internal search is a widespread practice: 66% of establishments typically search internally. The accumulation of specific skills and the provision of incentives for employees are the main advantages of hiring internally
in the theoretical literature. Ordered probit estimates show that on-the-job training and internal search are positively associated. On the contrary, incentive schemes such as variable pay are not associated with employer search. These results help to assess competing theories, and in particular, suggest that specific human capital is an important driver of internal hiring. Finally, we uncover two interesting facts that need further research: internal hiring is less likely in service firms and in noncompetitive product markets.
The Unequal Costs of Job Loss Across Countries
with Edoardo Accabi, Andreas Gulyas, Stefano Lombardi, Cristina Barcelo, and Raffaele Saggio
with Edoardo Accabi, Andreas Gulyas, Stefano Lombardi, Cristina Barcelo, and Raffaele Saggio
Abstract
We document the consequences and sources of job loss across countries using a harmonized dataset and research design. Workers in Denmark and Sweden experience the lowest earnings losses, while losses in Italy, Spain, and Portugal are
three times as high. French and Austrian workers face in-between losses. Key to these differences is that Southern European workers are less likely to find employment following displacement. Loss of employer-specific premiums explains a
significant share of wage losses in all countries. Spending on active labor market policies is the sole labor market institution able to explain a significant portion of the cross-country heterogeneity in earnings losses.
three times as high. French and Austrian workers face in-between losses. Key to these differences is that Southern European workers are less likely to find employment following displacement. Loss of employer-specific premiums explains a
significant share of wage losses in all countries. Spending on active labor market policies is the sole labor market institution able to explain a significant portion of the cross-country heterogeneity in earnings losses.
Abstract
Job-to-job (J2J) transitions play a key role in shaping aggregate labour market outcomes. However, these moves are subject to measurement issues in commonly used data sets. We investigate two data limitations using a unique combination
of administrative data for Denmark from 2008-2019. First, we estimate the extent of time aggregation bias by comparing the J2J transition rate at different observational frequencies. We find that the quarterly frequency overestimates J2J by
approximately 40 % compared to the daily frequency. Time aggregation bias is significantly larger in expansions compared to recessions. Second, we evaluate the extent of employer internal labor markets. We do this by looking at the share of
J2J moves between establishments which are employment reallocation within the same employer by matching firm ownership data to employer-employee data. We aggregate firms with the same owner and define it as one employer (a business
group). We find that 11.5% of J2J transitions are reallocations within the same employer, and 17% of these moves are between firms within the same employer.
of administrative data for Denmark from 2008-2019. First, we estimate the extent of time aggregation bias by comparing the J2J transition rate at different observational frequencies. We find that the quarterly frequency overestimates J2J by
approximately 40 % compared to the daily frequency. Time aggregation bias is significantly larger in expansions compared to recessions. Second, we evaluate the extent of employer internal labor markets. We do this by looking at the share of
J2J moves between establishments which are employment reallocation within the same employer by matching firm ownership data to employer-employee data. We aggregate firms with the same owner and define it as one employer (a business
group). We find that 11.5% of J2J transitions are reallocations within the same employer, and 17% of these moves are between firms within the same employer.