Why are specialised employers, more than generalist firms, interested in preserving good matches? My recent paper looks at adverse exogenous shocks and suggests that specific human capital investments can become a protection for employees. Building on Wasmer (2006) and Pissarides (2011), the hypothesis is that the specialised employer is keen to locate and preserve employees because it is costly and difficult to find suitable individuals for the kind of job the specialised firm offers. Reciprocally, the specialised firm’s employees have fewer incentives to leave the firm, because the outside labour market is thin for his idiosyncratic skills.

An example can be found in a watchmaker production line. Typical stages are the dial assembly, the mounting of the hands, the setting into the case and the manual tests. Each of these steps requires a specialist who is an expert in his role. He is fully responsible for his task and conveys the finished piece to the next work stage. This sequence is repeated until the watch is ready to be sold. The dexterity of an individual worker determines the quality of each step, the quality of the subsequent steps of the process, and ultimately the excellence in quality of the watch.

From a theoretical perspective, this study relates to the literature on the complementarity between employees (Becker and Murphy, 1994) — which leads to superadditivity (Rosen, 1978); as well as on an extensive body of studies on firm-specific human capital (Becker, 1962; Gibbons and Waldman, 2004; Lazear, 2009; Gathmann and Schönberg, 2010) highlighting aspects such as low substitutability of employees. The study is also intersected by monopsony — implying that in the extreme there is only one employer, who exhibits market power over his employees (Sullivan, 1989; Manning, 2003Borjas, 2008Ashenfelter et al., 2010; Staiger et al. 2010) and search models (Mortensen and Pissarides 1994).

The study relies on two main assumptions. The first one is that working under high levels of division of labour fosters labour productivity (conditioned, for instance, on coordination costs). The second one is that employers can contract with employees on a single basis, attaining optimal matches (Jovanovic, 1979). Once the match is achieved, the specialised employer is interested in preserving it.

The logic is that the single employee is precious for the specialised firm for at least two reasons: (i) he raises the productivity of his coworkers, (ii) he works on a fairly limited number of specific and closely related tasks, being the only one who possesses information about the role (see Molina-Domene, 2018) among others). Crucially, the employee becomes more important over time: the specialised firm needs him on the grounds they have already developed firm-specific human capital, obtained by learning through specialisation or by training (Acemoglu and Pischke, 1998).

The empirical evidence is based on matched employee-employer data (LIAB 1993-2010) from the German Social Security, which covers all workers employed in one of the surveyed establishments. This data feature is important, because it allows us to build the specialisation proxy — based on the entire occupation distribution within each firm, which measures occupation concentration.

The first part of the study looks at specialised firms’ response to downturns. To assess the business cycle, the paper relies on changes in the National Account industry gross value added in Germany as a source of variation, and focuses on downturns. Consistent with an extensive literature (Burnside et al., 1993; Bernanke and Parkinson, 1991; Hall, 1988; Bernanke, 1986; Fair, 1985; Fay and Medoff, 1985), the results show that more specialised firms preserve their labour force during slumps, appearing less productive. Within our sample, a one-standard-deviation increase in specialisation decreases labour productivity by around 0.03 per cent during downturns. Additionally, specialised firms slightly increase wages by almost 0.02 per cent — to preserve, and potentially to attract new employees.

The second part of the study assesses other frictions related to the matching market. The story is about employers who face the unexpected absence of an employee. To circumvent endogeneity concerns (e.g. an employee quits because he is unsatisfied with the job experience), the paper follows previous research (Jaravel et al., 2018; Fadlon and Nielsen, 2017Jäger, 2016Isen, 2013Becker and Hvide 2013Oettl, 2012Azoulay et al., 2010Bennedsen et al., 2007Jones and Olken, 2005) and uses employee deaths as a source of variation. The setting is a quasi-experimental research design, which compares the response of firms of different levels of specialisation to the death of an employee, in terms of different labour outcomes.

The results show that more specialised firms require more time to externally find a substitute for the deceased, via hiring — compared to generalist firms. Nevertheless, there is some underlying heterogeneity: specialised firms fill the vacancy of high-tenured employees more promptly, conditioned on the labour market availability. Additionally, the death of an employee hits specialised firms hard: a standard deviation increase in specialisation decreases firm productivity by around 0.04 per cent in the second year and around 0.03 per cent in the third year. This is true even if specialised employers decrease the wages of the remaining co-workers more than their counterpart generalist firms. The effects are stronger for smaller firms and extend until the third year after the death, suggesting monopsonistic power.

The findings point to the fact that specialised employers appear to rely on their employees, fostering slow job reallocation. This suggests that these firms’ bargaining power is limited (adverse demand shock), albeit providing a signal of some employers’ monopsonistic power (adverse labour supply shock). Given the employee-employer mutual dependence, a bilateral monopoly is plausibly in place. All in all, specialised firms emerge from this study as not quite resilient to adverse shocks.



Maria Molina-Domene works in the labour programme at LSE’s Centre for Economic Performance (CEP). Her research focuses on labour economics and big data.