Panel Paper: What Policies Increase Bureaucratic Expertise: Evidence from a Field Experiment

Friday, November 3, 2017
San Francisco (Hyatt Regency Chicago)

*Names in bold indicate Presenter

Ulrik Hvidman and Simon Calmar Andersen, Aarhus University


What Policies Increase Bureaucratic Expertise: Evidence from a Field Experiment

One-sentence summary
Using a randomized field trial, we evaluate different methods of increasing bureaucrats' investments in expertise.

Abstract

The implementation of performance information systems has been a core component of public sector reforms that have spread across many Western countries since the 1980s (Pollitt and Bouckhart 2011; Moynihan 2008). Promoting learning and expertise within bureaucracies often constitute a central promise of these reforms. The aim of this study is to understand how policy makers can induce bureaucrats to invest effort in acquiring policy-relevant information and develop expertise. Research has been dominated by formal models of rational agents that do not take into account humans’ limited cognitive capacity. An emerging empirical literature has studied how different factors correlate with the use of performance information (for a systematic literature review, see Kroll 2015). However, as most of these studies are cross-sectional, it is questionable whether the results have a causal interpretation.

The present study makes two main contributions. First, we discuss theoretically—using literature from behavioral economics and psychology—why bureaucrats’ limited capacity to attend to all relevant factors may affect their decision to invest effort in performance data. Second, to evaluate different methods of increasing investments, we study a field experiment in which the Ministry of Education in Denmark encouraged public schools to invest time and effort in acquiring information about the socio-emotional skills and wellbeing of their students—an aspect of education that has been shown to be as important for long-term outcomes as cognitive skills (Heckman 2007). In the experiment, schools were randomly assigned to four groups: a control group that received an invitation to introduce the information system, an economic incentive group that received a payment if they accepted, a behavioral group that were reminded to decide whether or not to invest in the performance data, and a combined group that received both the incentive and the behavioral treatment. Results show a significant interaction effect between the incentive treatment and the behavioral treatment, which supports the notion that bureaucrats’ investment decisions are affected both by incentives and by behavioral barriers.