From Per Pinstrup-Andersen, Science Council Chair

One of the most important goals of the Science Council is to help assure that the funds allocated to the CGIAR are well spent. Given the funds available, we must seek the greatest reduction in poverty while improving food security, nutrition and natural resource management. We will never have enough money to do everything we want to do, so we must prioritize and allocate funds to those Centers and programs we believe will have the greatest impact. But how do we identify these Centers and programs? One way would be to evaluate past performance. In fact, during the last couple of years, increasing attention has been paid to performance measurement and performance indicators. Such measurements and indicators may be a useful input into judgments about future allocations of funds but only if they relate closely to what we in the CGIAR are trying to accomplish. In other words, which Centers, programs and projects achieved satisfactory progress towards meeting the goals of the particular research and what would be the impact of achieving each of these goals? But to answer these questions, Centers, programs and projects must have well defined goals, timelines, and annual milestones. If the goal is not appropriately defined, we will not know if or when it is achieved and if we do not have clearly defined timelines and periodic milestones, we will not know whether we are moving towards achieving the goal at an acceptable speed. This is the logic behind the logframes now used by all Centers.
So, what's the problem?

The problem is that it is very difficult to measure progress towards achieving research goals. Unlike other production processes, one cannot simply count the research contributions and compare them across programs. The performance of a factory producing toothpaste may be counted as the number of tubes produced, the net profits obtained, or some other easily measured result. This can then be compared to results from other toothpaste factories. Not so for research. Attempts to focus performance measurements on what can be counted and compared across programs or Centers tend to give misleading results, because those things that can be counted may not be good indicators of progress towards achieving impact on poverty. For example, should a Center that produces twice the number of refereed journal articles produced by another Center be allocated twice as much money?

If the final goal of the CGIAR is to produce the largest number of refereed journal articles, the answer might be yes, but it is not. Refereed journal articles are at best a means to an end, not an end in themselves. It gets even more misleading when we try to aggregate quantifiable indicators such as number of visitors coming to a particular Center, number of improved lines released, number of people trained, number of publications, etc. The program or Center with the largest total gets the most money! We can obviously do better than that. A focus on what can easily be counted and compared across Centers and programs may also reduce the incentive to take risks in breaking new ground and seeking new solutions, using a learning and feedback mode, which is so important in innovative applied research.

So what am I proposing? I am proposing that we measure progress towards agreed upon goals by integrating performance measurement into the logframes. For this to work, we must strengthen the logframe presentations to include clearly defined goals, timelines, annual milestones, and the extent to which the milestones were achieved. The comparisons across Centers, programs, and projects would then be based on the progress achieved towards a clearly defined set of milestones for each Center, on the relevancy and quality of the science, and on the congruence of Center goals with system priorities instead of the sum of a set of quantifiable indicators that may be more or less irrelevant to the achievement of the goals. Those donors who wish to allocate funds on the basis of relative performance in moving towards agreed upon goals, would be on much firmer ground. Furthermore, everybody would benefit if donors would agree to a common performance measurement framework and share a common set of indicators.

Performance should be measured on the basis of achievement of outputs, outcomes and, what we are ultimately interested in, impact. However, we should recognize that Centers and research programs have less control over impact and outcome than they have on output. The Science Council and other CGIAR stakeholders should work together to strengthen the pathways from research output to impact on poverty. Judging the performance of individual Centers and research programs solely on impact might be misleading because of so many other factors influencing impact that the individual Center or program has little or no control over.

The Science Council is fully involved with the Centers, the Working Group on Performance Management and other stakeholders in moving towards the integration of performance measurement in the logframes of Centers and across Center programs. Hopefully we will have the approach in place in time for the Centers to strengthen next year's medi-um-term plans and logframes so they can serve as the basis for future performance measurement.

Per Pinstrup-Andersen