The Role of Contract Types for Employees’ Public Service Motivation

Abstract

The intention of “doing good for society” is regarded to be a crucial motivator for employees in the public sector in order for them to perform well. Recent research in the public sector literature calls for a deeper understanding of how this specific public service motivation (PSM) is shaped. In our paper, we analyze how different degrees of inclusion in the public sector matter for PSM. We investigate how prospects of employment relations (fixed-term versus permanent contracts) and temporal differences (part-time versus full-time employment) moderate PSM in the public and in the private service. Our findings show that aspects of PSM are related by these employment characteristics in various ways, suggesting that the factors influencing PSM are multifaceted and that actual employment conditions need to be taken into consideration when assessing PSM.

Appendix

Fig. 3

Distribution of engagement for others and political/social involvement

Fig. 4

Marginal effects. Note: Estimated marginal effects and 95% confidence intervals are reported

1.1 Linear Estimation Results

The binary probit model does not account for possible general and stable individual levels of PSM. That is why we complement our analysis by individual fixed effects estimations here, which takes unobserved heterogeneity with regard to PSM into account. However, the fixed effects option is not available for probit models. We therefore complemented our analysis with linear fixed-effect models and kind of pretend that our dependent variables are metric. We make use of the original ordinal variables with four response categories here (see Fig. 3) and have coded the response options from 1 (not important) to 4 (very important) and call this variable PubSerMot. The linear individual fixed effects model can then be written as

$$\begin{aligned}\textit{PubSerMot}_{i,t}=&\,\beta_{1}\textit{Public sector}_{i,t}+\beta_{2}\textit{Temporary work}_{i,t}\\&+\beta_{3}\textit{Part-time work}_{i,t}+\beta_{4}\textit{Public sector}*\textit{Temporarywork}_{i,t}\\&+\beta_{5}\textit{Public sector}*\textit{Part-time work}_{i,t}\\&+\beta_{6}\textit{Controls}_{i,t}+a_{i}+u_{i,t}\,.\end{aligned}$$

(3)

Note, however, that one has to interpret the results with caution due to the nature of our dependent variables. Although, we take unobserved heterogeneity into account by considering individual fixed effects, endogeneity may be left to some extent. It is still possible that we have only partly captured sorting effects into the public sector in our analysis, which is argued to be relevant (Georgellis et al. 2011; Leisink and Steijn 2008; Vandenabeele 2008; Wright and Christensen 2010). We therefore further complement our analysis by using the lag of being there for others and political/social involvement as instruments. We report the corresponding Arellano-Bond GMM estimations with one lag of the dependent variable as additional independent variables in Table 3:

$$\begin{aligned}\textit{PubSerMot}_{i,t}=&\,\alpha\textit{PubSerMot}_{i,t-1}+\beta_{1}\textit{Public sector}_{i,t}+\beta_{2}\textit{Temporary work}_{i,t}\\&+\beta_{3}\textit{Part-time work}_{i,t}+\beta_{4}\textit{Public sector}*\textit{Temporarywork}_{i,t}\\&+\beta_{5}\textit{Public sector}*\textit{Part-time work}_{i,t}\\&+\beta_{6}\textit{Controls}_{i,t}+a_{i}+u_{i,t}\,.\end{aligned}$$

(4)

Table 3 Linear estimations on PSM characteristics

There are both significant positive coefficients for the public-sector dummy and negative coefficients for its interaction with the temporary contract dummy. Employees therefore seem to adapt their (attitudes to) engagement for others accordingly their own situation. Changes in the employment situation are less associated with changes in attitudes to political/social involvement, though. The corresponding fixed-effects estimates for this PSM dimension are not significant, although there is more variation in reported attitudes for political/social involvement than for engagement for others. Again, we should stress, however, that an underlying assumption of this linear fixed-effects model consists in equal differences between categories. Since this is not ensured, we should interpret results very cautiously and avoid interpreting size effects. The results of the Arellano-Bond estimations indicate that reported attitudes towards political/social involvement and being there for others are significantly related over time within individuals. Most importantly, we still get significant results for the public-sector dummy and its interaction with temporary and part-time contract, which indicate further evidence in line with our hypotheses. In sum, the results further indicate that at least some socialization for PSM is relevant next to previous sorting effects into sectors.