The impact of matching mission preferences on well-being at work

A worker’s job satisfaction is often considered as part of a worker’s general subjective well-being or life satisfaction (Clark and Oswald 1996; Praag et al. 2003).

Closely related to this study are a number of (field and lab) experiments in economics on mission motivation (Tonin and Vlassopoulos 2010, 2012; Gerhards 2013; Carpenter and Gong 2013; Fehrler and Kosfeld 2014). These studies invariably find that participants with aligned mission preferences exert more effort in a chosen or real effort experiment as compared to participants with conflicting mission preferences.

The 2010 election in Britain provides a perfect test case to examine this distinction as, for the first time since the start of the BHPS, a coalition government took office in Britain.

For more information on the LISS panel study see www.​lissdata.​nl. For more information on CentERdata and their data projects see www.​centerdata.​nl/​en.

The Netherlands has a parliamentary political system with proportional representation in the house of representatives. In a parliamentary system, political parties that have a (coalition) majority in the house of representatives are able to take office as a government. Such a majority is sufficient to implement policies. Political parties in office after the elections of November 2006 are Christian Democrats (CDA), Labor party (PvdA), and Christian Union party (CU). Political parties in office after the elections of June 2010 are Liberal party (VVD) and Christian Democrats (CDA) with support of the Freedom party (PVV) in parliament.

A possible bias would occur if people are dishonest about revealing their vote in the foregoing parliamentary elections. While it is not possible to observe whether a respondent reports honestly, respondents were given the opportunity to indicate that they “do not know” what they had voted or indicate that they “prefer not to say” on which party they voted. The number of observations in these categories is very low, namely between 1 and 5 % each year.

Results show that a match of mission preferences has no effect on a worker’s satisfaction with wages, hours worked, career, or atmosphere at work. Moreover, a match of mission preferences has only a small positive effect on a worker’s overall job satisfaction, as measured by the question: “How satisfied are you, all in all, with your current work?” A possible explanation for this diluted effect on general job satisfaction could be that many different facets of work influence a worker’s overall job satisfaction while matching mission preferences only affect one facet of work, namely work itself.

Other answer categories include: Agriculture, Forestry, Fishery and Hunting, Mining, Industrial Production, Utilities, Construction, Retail Trade, Catering, Transport, Storage and Communication, Financial, Business Services, Education, Healthcare and Welfare, Environmental Services, Culture, Recreation and Other Services, and Other.

Results show that a match of political preferences has no effect on job satisfaction of public sector workers who are not employed by the government (e.g., workers in education or health care). This result is in line with a large part of the public sector organizations having more discretion, as compared to government organizations, in setting their objectives.

In contrast to the Hausman test, the test proposed by Wooldridge (2002) allows for clustering of errors at the individual level.

See, for instance, Winkelmann and Winkelmann (1998) who suggest to choose a cut-off point and estimate a fixed effect binary logit, Ferrer-i-Carbonell and Frijters (2004) who allow cut-off points to differ over individuals, and Das and Soest (1999) or Baetschmann et al. (2015) who combine estimates for each possible cut-off point made possible by the data.

In the main fixed effects regression, the effect of a match for non-government workers (denoted by \(\theta \)) is identified on 432 individuals and the difference in the effect of a match between non-government and government workers (the interaction effect denoted by \(\psi \)) is identified on 82 individuals.

A number of studies have examined the political preferences of government workers. These studies find that government workers, as compared to the general population, are more likely to be left-wing orientated (Rattso and Sorensen 2013), although this does not always translate into a higher likelihood to vote for left-wing or socialist parties (Knutsen 2001, 2005; Jensen et al. 2009).

Note that all of the reported main results are robust to performing the analysis using an ordinal fixed effects method (see Table 7 in “Appendix”). This result is in line with Ferrer-i-Carbonell and Frijters (2004) who find evidence that the bias of the linear fixed effects estimator is generally small when using ordinal data. The ordinal fixed effects estimator used in comparison is the ‘blow-up and cluster’ estimator (see Baetschmann et al. 2015 for an extensive discussion). Geishecker and Riedl (2012) show that this method performs as well as or better than the other available ordered fixed effects methods.

One may wonder whether the effect of a match differs between workers who switch between sectors and workers who do not switch between sectors. Excluding workers who switch between sectors from the analysis does not significantly change the main results.

An alternative explanation for the positive effect of matching mission preferences could be that government workers who voted for one of the political parties in office are rewarded with better jobs. While it is not possible to directly observe whether workers with matching preferences are rewarded with better jobs, it is possible to examine whether matched workers are rewarded with better pay or promotion opportunities. To do so, both a worker’s pay and self-assessed prospects for career advancement are regressed on the matching variable. Results show that workers who voted for a political party in office are not rewarded with better pay and do not rate their prospects for career advancement differently than workers who voted for a political party in the opposition.

The size of this effect is comparable to the effect of a match (0.400), which is significant. The insignificance of this estimate could also be influenced by the relatively small number of workers who switch between sectors in the data.

There may be additional dimensions along which the effect of matching mission preferences differs, for instance, the effect may differ given a worker’s tenure. More tenured workers may react differently because they have experienced more changes in government and different government compositions. Results show that the effect of a match decreases for more experienced workers, however, this effect is very small and highly insignificant.

The low education category contains all workers who have completed only primary school or intermediate secondary education, the intermediate education category includes all workers who have completed higher secondary or intermediate vocational education, and the high education category includes all workers who have completed higher vocational or university education.

A worker’s opinion of a political party is measured on an 11-point scale using the question: “What do you think of [party name]?”. The respondents’ answers to these questions are closely related to their answers to the voting question. There is a strong positive correlation between having voted on a political party that takes up office and how positive a respondent rates the coalition parties in office.