Measuring results in development: the role of impact evaluation in agency-wide performance measurement systems

Abstract

The need to demonstrate results has been growing over the last two decades. There is a growing recognition that outcome monitoring only tells us what happened, not why it happened, and so outcome monitoring cannot give useful information about the effectiveness of development interventions. Many agencies are turning to impact evaluations to assess the change in outcomes which may be attributed to their interventions. As more impact evaluations are produced, the question arises as to how these studies can be used to assess how the agency as a whole is performing. We discuss the ‘triple A’ principles for an Agency Wide Performance Measurement System (AWPMS) – alignment, attribution, and aggregation – as well as several other challenges faced in designing and implementing such systems: the balance between lesson learning and accountability, independence versus influence, modifying incentive structures for results, determining how many impact evaluations to conduct and of what, and ensuring the use of these studies. We conclude by urging greater cooperation between those charged with measuring results and those conducting impact evaluations in order to develop AWPMS which can meaningfully report on agency performance.

Keywords:

• impact evaluation
• development effectiveness
• performance measurement

Notes

1. For discussion of debates around the need for a counterfactual, see White (2010).

2. A properly designed impact evaluation of MVP is now underway; see Masset et al. (2013).