While the Consultative Group on International Agricultural Research (CGIAR) has been long considered a driving force behind the successes of the “Green Revolution,” no prior study has attempted to develop an aggregate estimate of the value of the System’s impacts. However, economic ex-post impact assessments (IAs) have been conducted for some of the most outstanding individual innovations of the System. This study synthesizes these specific benefits estimates, and sets such against total investments in the CGIAR, with the benefits from all other CGIAR activities omitted, so as to derive estimates for five different aggregate benefit-cost scenarios.
Economic impact studies for inclusion in the meta-analysis were selected based on a literature survey of publications databases, examination of reference lists from prior studies and scrutiny of International Agricultural Research Centre publications. Since impact assessment has been pursued in a largely decentralized manner, standards and approaches differ significantly among studies, and, hence, a critical review process was necessary for determining the reliability of generated results. To develop the conceptual grounding for the review process, best practices were identified for economic impact assessments.
Two overarching principles for evaluating study reliability- 1) transparency and 2) demonstration of causality, as well as accordant criteria and indicators, were developed from the identified best practices. Transparency was represented by three criteria: 1) clearly derived key assumptions, 2) comprehensive description of data sources, and 3) full explanation of data treatment. Demonstration of causality was represented by five criteria: 1) representative data set utilized, 2) appropriate disaggregation, 3) adequate consideration of mitigating factors, 4) plausible counterfactual developed, and 5) precise institutional attribution.
Using these criteria, five benefits scenarios were developed. These scenarios include 1) a scenario only including highly rated “significantly demonstrated” studies that empirically attribute benefits to specific activities of the CGIAR, rather than arbitrarily partitioning benefits from efforts in collaboration with partners, 2) a conservative scenario of only highly rated “significantly demonstrated” studies, 3) a selection of “plausible” studies meeting minimum standards for the criteria described above, 4) a “plausible, extrapolated to the present” scenario in which benefits for the crop genetic improvement studies were assumed to continue from the study period to the present (end of 2001) and 5) a “plausible, extrapolated through 2011,” which assumes that the products of current research will continue to be realized at present rates through 2011.
Against an aggregate investment of 7,120 million 1990 US dollars (6,900 million of investment in the CGIAR, plus relevant pre-CGIAR costs) from 1960 through 2001, all scenarios produced benefit-cost ratios in substantial excess of one, based on benefits accruing from 1972 - 2001. Including only “significantly demonstrated” studies that empirically attribute CGIAR derived contributions to collaborative efforts results in a ratio of 1.94, while if all “significantly demonstrated” studies are considered, with assumed attributive coefficients applied, the ratio rises to 3.77. The “plausible” scenario results in a ratio of 4.76, while when extrapolated to 2001 this rises to 9.00, and extrapolated through 2011, this becomes 17.26. Since costs are distributed over the benefit period, and many benefits peaked in the early 1990s, the discount rate applied only significantly affected generated ratios in the extrapolative scenarios.
The true value of benefits arising from the CGIAR is probably in excess of even the upper bounds of the results demonstrated here, as only a small subset of System impacts have been assessed. To illustrate this point, 98.1% of “significantly demonstrated” and 93.4% of “plausible” benefits were generated by just three research areas - cassava mealybug biocontrol, breeding of spring bread wheat and modern varieties of rice. Anecdotal evidence suggests that these are not the only areas of CGIAR research success, so there is substantial scope for expanded impact coverage, and better illustration of how System activities influence target beneficiaries. Furthermore, even where economically assessed there still remain significant opportunities for improving the methodological rigour, comprehensiveness and transparency of System assessments.
The diversity of methods employed among Centres and research programmes appears to indicate that additional resources for impact assessment leadership at the System level would offer considerable potential to improve consistency and raise analytical standards. However, for such leadership to be highly effective, it will be necessary for the “clients” of impact assessments to articulate expectations for substantiating different types of impact claims. In the absence of such, it is difficult to select one of the six scenarios as most “accurate,” and the “true” benefit-cost ratio of the CGIAR investment will remain unresolved.
