Where to go? Workers' reasons to quit and intra- vs. interindustry job mobility

Abstract

We show that workers' reasons for quitting their job affect their decision to stay in or leave their industry, using survey data among public sector employees in the Netherlands. Workers quitting for e.g. pay, work pressure, or job duties move relatively often to another industry, in contrast to workers quitting for commuting time or the atmosphere at work. This suggests that workers use their experience in the initial job to update their expectations on other jobs in the industry, as the first set of job aspects is more likely to be related among jobs within an industry than the latter. Furthermore, it is shown that workers' reasons to quit fully explain the differences in wage growth between intra- and interindustry job movers. Lastly, we find that workers who quit for pay or management often leave the public sector altogether.

†The data used in this article are proprietary.

Acknowledgements

I am grateful to Robert Dur for guidance and encouragement. I would like to thank Andrew Clark, Silvia Dominguez Martinez, Chaim Fersht-man, Richard Paap, Otto Swank and attendants of seminars in Rotterdam and Amsterdam for useful comments and suggestions.

Notes

†The data used in this article are proprietary.

¹ See, among many others, Topel and Ward (1992), Light and McGarry (1998), Campbell (2001), McLaughlin and Bils (2001) and Lima (2004).

² The literature on displaced workers is surveyed by Kletzer (1998) and Kuhn (2002).

³ Employees who moved from one job in the public sector to another may have received two questionnaires. However, there is no evidence of duplicate cases in the dataset.

⁴ The data also distinguish nine different industries in the private sector.

⁵ The focus on job-to-job mobility removes 2904 respondents from the sample who did not have a job before entering their public sector job, as well as 3234 respondents who did not take up another job after leaving their public sector job.

⁶ We have set the floor and the ceiling for both the former and the new hourly wage at 3 euro and 60 euro, respectively. The floor is slightly below the legal minimum wage for 18-year-old employees and the ceiling is (in terms of monthly income) slightly above a Minister's wage in the Netherlands. The cut-off levels for relative wage change have (arbitrarily) been set such that workers whose hourly wages more than halved or more than tripled were excluded. Inspection of the data reveals that most of these cases involve typo's, misspecifications, or misinterpretation of the questions (for instance, some respondents appear to report yearly rather than monthly income).

⁷ This specification imposes arbitrary weights on the questions regarding the importance of job aspects, which is also acknowledged by Mathios (1989). We have checked the robustness of our results by using different specifications. Specifications including only the most important reason for quitting or the three most important reasons for quitting (equally weighted) yield qualitatively similar results, but perform worse than the 0-1-2-3-4 specification in terms of explanatory power. Furthermore, we used a specification which imposes no structure of weights, by inserting a dummy variable for each level of importance of all job aspects. Again, qualitatively similar results emerge, although for several job aspects, the magnitude of the coefficients is not monotonically increasing in the level of importance. Still, the estimated effects of job aspects being ‘most important’ relative to being ‘not important’ closely resemble the estimations reported in the article.

⁸ Note that this need not imply that employees care little about the level of their wage. Due to the rather compressed wage structure in the Netherlands, employees may rationally expect a job change to yield little financial gain.

⁹ Unfortunately, the survey among employees who entered a public sector job did not explicitly ask whether the respondent quit their previous job. Hence, there may be some actual layoffs and discharges in the sample. The survey among employees who left a public sector job did ask whether the employee voluntarily left the job or had been displaced. Given the evidence in the literature that the consequences of a separation differ between workers who quit and workers who are displaced, we have checked whether elimination of all respondents who considered threats of involuntary separations important in their decision to quit would affect our results. It turns out that all results are qualitatively similar in this restricted sample, except for the effect of the reason-to-quit variable ‘contractual hours’ on wage growth, which becomes smaller and insignificant (see Table 3).

¹⁰ Elimination of constant characteristics from the set of explanatory variables turns out not to affect the results.

¹¹ The lack of data on employees who change jobs within the private sector hinders a similar estimation for industries in the private sector.

¹² The difference with the estimations in Table 3 is that the dummies for the new industry of employment have been replaced by dummies for movers and leavers, with stayers as base category.