Although the decision to concentrate on rice biotechnology for developing countries was in itself an example of priority setting, the Rockefeller Foundation faced another challenge upon beginning the biotechnology program. Which areas of rice biotechnology were most worthy of support? Would the payoffs be greatest in disease and pest resistance? In breeding for abiotic stress tolerance? In the development of tools useful for breeding? In the identification of molecular markers that could be used to test for genetic diversity?
In the face of such questions, the Rockefeller Foundation undertook a careful and deliberate program of priority setting. This analytic framework has continued to guide the Rockefeller Foundation's investments in biotechnology since the mid-1980s. The priority setting experience has been documented effectively in Herdt.14 The exercise involved an eight-step process, summarized below:
1. The target environments and regions were defined. These covered six geographic regions and four cross-cutting rice agroecologies. A critical step in any priority-setting process is to decide how to break down the different potential categories of research into RPAs. This decision in itself requires expert opinion to delineate the boundaries of specific RPAs.
2. The importance of different research problem areas was estimated in three different ways. First, "knowledgeable scientists" estimated the yield losses due to different problems in each environment and region. As an alternative measure, a group of scientists "scored" each problem in each region, giving the relative importance of the various problems; the scientists were also asked about the maximum benefits that could be obtained from "solving" all of the relevant problems. Third, yield losses in specific regions were compared to a "reference region", allowing for some consistency in estimates of importance across regions. The challenge here is to arrive at useful relative measures of the importance of different RPAs while keeping sight of the aggregate plausibility of the assumptions. In other words, it is not credible to find that six insect pests each cause an average of 15 percent yield losses annually in wheat. Some adding-up constraint must be imposed if the results are to be believed.
3. Yield losses were converted into monetary terms by multiplying by areas and prices. This is perhaps the most straightforward step in the process.
4. Environmental effects were taken into account by creating a set of weights for each problem in each region and agroecology. These weights were designed to reflect the added benefits that could be gained from using genetic methods to alleviate a problem that currently required pesticide use, herbicide use, or other environmentally harmful practices.
5. Equity considerations were added by making an assessment of which RPAs would have the greatest benefits to the poor. Net benefits were then adjusted to yield an "equity- weighted" measure of the payoffs from achieving success in different RPAs. The point of this step was to allow planners to place greater weight on research that might be expected to benefit the poor.
6. The net present value of research benefits was computed for each problem and each agroecology. This depended critically on the assumed time lag until the problem was solved. Time lags were based on the elicited responses of scientists, who were asked how long they would expect it to take to find a solution for a problem-region-agroecology bundle, given a rate of research investment of $0.2 million per year. Although the figures elicited in this way may not be accurate in absolute terms, they probably offer a decent index of likely time lags. In other words, all else being equal, we would be inclined to believe that scientists might do a decent job of guessing which RPAs might take longest to "solve" and which might take less time.
7. The problems, regions, and agro-ecologies were ranked for susceptibility to a biotechnology solution. This was based on subjective probability estimates of knowledgeable scientists. This issue arises because some RPAs may be very important but may be solvable with conventional breeding techniques more readily than through biotechnology. Since the goal here is to set priorities for biotechnology research, there seems to be little point in selecting RPAs that are equally well addressed through conventional breeding or other techniques.
8. An investment rule was developed. This was a rule converting the overall rankings into actual funding decisions. Having established "priorities", in the sense of a ranking of the importance of alternative research problems, there is still a need to convert these priorities into some action.
The Rockefeller exercise led to a detailed priority listing for rice biotechnology research. The top priority after adjusting for equity issues, environmental impacts, and the likely usefulness of biotechnology techniques, was the search for resistance to the rice tungro virus. Other top priorities were submergence tolerance, gall midge resistance, and a source of cytoplasmic male sterility.
Some of the problems that were ranked as most serious, in terms of raw crop losses, did not emerge as high priorities after the rankings were adjusted to reflect environmental and equity concerns and their susceptibility to biotechnology solutions. For example, weed problems were the biggest single source of crop losses in the data, but they ultimately ranked 15th among RPAs in terms of priorities for research. Conversely, submergence tolerance, which ranked seventh in terms of crop losses, ended up as the second highest priority after all the requisite adjustments.
This suggests that a formal priority-setting process can lead research planners to identify priorities that would not otherwise have been obvious. Moreover, it can lead researchers to discard problem areas that seem unlikely to be worthwhile.
In the case of the Rockefeller undertaking, within a relatively brief time from the beginning of the priority-setting exercise, Herdt was able to report that "rice plants transformed with various gene constructs for resistance to rice tungro virus have been produced and are being evaluated at IRRI."15 This suggests that the priority assigned to tungro research—based in part on the expectation that research could be effective—may have been well founded. Rice tungro virus is one of the most damaging crop diseases in the world, destroying as much as 7 million tons of rice output annually.15
Such claims of success warrant careful analysis. Although it is too soon to evaluate the priority-setting exercise formally, Evenson has undertaken an interim evaluation of the rice biotechnology program. One conclusion is that the benefits appear large. A second conclusion is that interim results support the original prioritization of biotechnology research across categories. Finally, Evenson concludes that there are high payoffs to continued rice biotechnology research for developing coun-tries.11
Was this article helpful?