Better than their reputation? How public managers use performance information to improve organizational performance

The last decades have seen a widespread tendency to use performance management as a mean to increase accountability and performance in public organizations (Pollitt & Dan, 2013). Yet, very little is known about how these... [ view full abstract ]

The last decades have seen a widespread tendency to use performance management as a mean to increase accountability and performance in public organizations (Pollitt & Dan, 2013). Yet, very little is known about how these management systems actually fare in the context of the public sector (Sanger, 2013). This uncertainty is not only related to the question of whether the management system improves performance, but also concerns the performance-enhancing mechanisms in the system (Ammons & Rivenbark, 2008). Therefore, the aim of this paper is to examine whether performance management affects organizational performance, and furthermore to test whether purposeful managerial use of performance information could be a source for improvements.

In the line with the general notion in the literature, the theoretical point of departure is the assumption that managers’ use of performance information is a prerequisite for performance management to work (Moynihan, 2009). To expand on this notion, I develop a theoretical framework, which specifies how managers could use performance information to improve performance (Behn, 2003). More specific, the theoretical argument is that managers, in line with a problem-solving approach to management (Simon, 1960; Mintzberg et al., 1976), should use performance information to identify problematic performance-areas within their organization, and improve performance by correcting problems in these areas.

I test this argument with data from a performance management system used on 48 public schools in a municipality in Denmark. In this system, each school is measured with the use of thirty different performance indicators (e.g. the GPA, truancy, parental satisfaction, and student wellbeing). After the measurement, the principal prioritize which performance-areas should be the school’s priorities for the following years. The data consists of three cycles (beginning in 2009, 2011 and 2013), and since the unit of analysis is performance information on the different performance-areas, each year has approximately 1400 observations.

The analyses are organized in two parts. In the first part, I use a fixed-effects within school estimation to investigate whether principals prioritize the performance-areas where their school is performing the worst. In the second part, a difference-in-difference estimator is used to examine whether there is any performance-enhancing effects of having prioritized performance-areas relative to non-prioritized performance-areas. Methodologically, the paper makes two contributions. First, I develop a new approach to conceptualize and measure managers’ use of performance information. Instead of relying on survey-measures, “use” is measured in a more direct and objective way with an estimation of the relationship between performance information and principals’ decisions. Second, the two estimation strategies makes it possible to handle some of the general methodological issues in the literature, such as reverse causality, omitted variables, common source bias, and social-desirability bias.

The preliminary results show that principals’ decisions are influence by performance information, as they prioritize the performance-areas where their school is performing the worst. Furthermore, prioritized performance-areas have a positive influence on performance, but only if the performance-area is consistently prioritized (through more than one cycle). The results implicate that managers are able to use performance information to produce results.