The impact of subjectivity in performance evaluation practices on public sector managers’ motivation

Abstract

We conduct an explorative study to investigate the effect of subjectivity in performance evaluation practices on managerial motivation in public sector organisations. Increased subjectivity can enhance motivation if supervisors are able to provide better informational feedback. However, subjectivity is likely to reduce motivation if it reduces perceived mission clarity or negatively affects relations between supervisors and subordinates. Our analysis is based on a survey among 94 public sector managers in the Netherlands. We predict and find that subjectivity in performance evaluation practices reduces perceived mission clarity, which in turn decreases motivation. We also find that subjectivity negatively affects subordinate managers’ trust in their supervisor, which also reduces motivation. Jointly, these results indicate that the negative effects of subjectivity in performance evaluation practices outweigh its potential positive consequences, suggesting that New Public Management's focus on more objective performance measures can indeed be beneficial. By itself, however, this does not automatically imply that more objective systems in general are optimal in all public sector organisations as such systems may have dysfunctional side effects such as distortion of performance measures, gaming or manipulation. In addition, we find that the effects of subjectivity are moderated by organisational characteristics.

Keywords:

• mission clarity
• motivation
• performance evaluation
• public sector
• trust

Acknowledgements

Comments from Victor Maas, Sven Modell, Paolo Perego, Roland Speklé and workshop participants at the 5th EIASM International Conference on Accounting, Auditing and Management in Public Sector Reforms are appreciated. We would also like to thank two anonymous reviewers for helpful suggestions. Financial assistance for this project was provided by Deloitte.

Notes

We do not exclude the possibility that employee motives may change as a function of employment sector choice (cf. Wright 2001).

All students in the programme participated and surveyed at least one manager in their organisation; some students surveyed up to five managers of separate, independent units in their organisation. All approached organisations participated in the research project.

Re-coded for further analyses, so that higher scores represent higher task uncertainty.

Measured on a five-point scale as follows: (1) ‘my decision’: a decision without superior involvement; (2) ‘two-person decision’: a decision together with supervisor; (3) ‘multi-person decision’: together with superior, decision is discussed with other supervisors outside the unit; (4) ‘organisational decision’: decision is taken higher in the organisation, at top level; (5) ‘initiative by others’: decision is taken outside the organisation, organisation implements the decision. The scale is re-coded for further analyses so that higher scores represent a higher degree of decentralisation.

Range, mean, median and standard deviation for the variables are calculated based on the simple average of the original item scores. The two items dropped based on the reliability analysis are not included. Reliability statistics in Table 1 are based on the PLS analysis.

Results are inferentially identical with or without these two items.

If we exclude this item with the lower loading and re-run the PLS analysis, the AVE of decentralisation increases to 0.55. All other results are inferentially identical.

Including size reduces the sample size to 82 observations due to missing responses on this question.

Measured by five items: (a) ‘There is a logical order in the steps that must be followed when performing activities in my unit’; (b) ‘There is a logical order in the steps that must be followed when executing tasks in my unit’; (c) ‘The most important processes in my unit can only be executed in one, clearly described, way’; (d) ‘In my unit, it is completely clear in what way the tasks in my unit should be performed’; (e) ‘To execute tasks in my unit, employees can rely on set guidelines and procedures’ (based on Abernethy and Brownell 1997, Withey et al. 1983).

We wish to thank an anonymous reviewer for this recommendation.

Items with a low loading are never deleted from a formative construct, because the construct is determined by definition by those items (Rodgers and Guiral 2011).

Thanks to two anonymous reviewers for this suggestion.