Exploring performance paradox in public organizations: Analyzing the predictors of distortive behaviors in performance measurement

In: International review of administrative sciences: an international journal of comparative public administration, Band 89, Heft 2, S. 501-518

Abstract

Scholars have argued that utilization of quantitative performance indicators by public organizations could generate unintended consequences that might outweigh the benefits of performance measurement. We examined the relationships between goal ambiguity and external control, and distortive practices in performance measurement in public organizations. The data from 47 agencies of the central and local governments in Korea were analyzed using fixed-effect ordered logistic regression methods. The results showed that goal ambiguity was positively and significantly associated with distortive practices in performance measurement. Goal ambiguity may incentivize public employees to misuse performance information to obtain rewards for higher performance ratings. Contrary to our expectation, reinforced external control was negatively and significantly related to distortive behaviors in performance measurement. Higher work autonomy was not significantly associated with manipulation of performance information. Employees with higher intrinsic motivation were less likely to distort performance measures, whereas the ones with higher extrinsic motivation were more likely to misuse performance information to achieve higher performance ratings. Points for practitioners To avoid the potential distortions in performance measurement, public managers should adopt a refined measurement model of performance customized to the unique characteristics of public services. Multiple sets of measures need to be developed and managed properly to respond to complex political environments which involve different key values and the stakeholders in the policy processes, and the nature of public service products and outcomes. Reinforced behavioral control in the process of measuring performance will also be necessary.