6.1. Increasing share of CGIAR resources allocated to social science research
6.2. Some research gaps
An increasing share of CG resources is being devoted to social science research. As shown in Table 2.1, according to the survey conducted for this study, the social sciences absorb 14.8% of the total Centres' budgets and 18.7% of the Centres' scientists. In the CG budget projections for 1996, the "policy" research category alone would exhaust 16% of the System's budget. This raises the important and difficult question of whether the share of System's resources devoted to this area of research is excessive or not.
In attempting to answer this question, a first observation is that a rising share of resources allocated to social science research reveals the fact that there has been a rising demand for this information and that allocating resources to this subject has been deemed competitive by scientists, Centre DGs, Centre boards, donors, and TAC itself. There is a priori no reason to doubt the wisdom of these many layers of decision making in allocating resources efficiently. In chapter 3 of this study, the complexity and fluidity of the emerging problems that the CG needs to address were used to justify the importance of social science research in confronting these issues in synergy with technological advances. Clearly, successful advances on the technological front without simultaneous advances on the socioeconomic, policy, and management fronts are impossible in such complex domains as poverty reduction and protection of the resource base. In a 1992 meeting of CG social scientists, Steve Vosti observed that "the tradition of agricultural research is to mobilize policy research when a policy issue inhibits dissemination of a new technology. In the ecoregional approach, however, policies may often be the initial point of intervention... The principal research product may be a modified policy rather than a technology" (Vosti, 1992).
The observed, carefully reasoned, increase in the share of resources allocated to social science research reveals that the CG system has strong comparative advantages in this field of science, even if it was initially introduced as an expost appendix to technological research. This System advantage derives from ability to carry simultaneously social science and biophysical research in an integrated multidisciplinary approach, in particular in the context of multi-centre and collaborative projects. The increasing complexity and ambitiousness of the problems addressed has transformed the essentiality of social science research from an instrument to enhance the diffusion of innovations to an integral component of the solutions sought.
The second observation is that the percentage of the CG budget allocated to social science research may not be superior to that which is assigned by the United States Land Grant Colleges and the USDA. At the University of California at Berkeley, the share of social science FTEs (Full Time Equivalent) in the College of Natural Resources is 19%, the same as in the CGIAR. These social scientists are not only in Agricultural and Resource Economics, but also in Conservation and Resource Studies, Forestry, and Nutrition.
Using the Current Research Information System (CRIS) of the USDA does not allow an easy separation by disciplines. An underestimation is to consider as social science the following two research programmes:
- People, communities, and institutions, including rural development- Competition, trade, adjustment, price, and income policy, and to exclude from the social sciences the following research programmes:
- Natural resources
- Forest resources
- Crops (field and horticultural)
- Animals
- Food science and human nutrition.
Doing, this, the budget allocated to "social science" by the USDA and the State Experiment Stations is 10% (USDA, 1993). This is an underestimation since the other research programmes clearly contain a social science component. We should also recall that it would be normal for the USDA to have a lower share of its budget in social science than the CG since the CG can "save" on fundamental science research by being a "user of science, not a doer of science". In addition, the type of problems addressed by the CG occur in a much more complex context of market failures, institutional gaps, weak and unreliable governance, and extensive poverty, all of which require a stronger social science component than if all markets worked and users of technology were living at a comfortable distance above the poverty line.
A scientific determination of optimum resource allocation to social science research would require comparing the internal rate of return from investment in this type of research to competing options. This is, however, rigorously impossible. Rates of return to research can be calculated by commodity, but not disaggregated by science components, particularly for the social science input. Hence, it is illusory to expect that a rigorous "scientific" answer can be provided. However, three observations follow:
1. The budget share allocated to social science research has evolved over many years through a series of demanding criteria and filters, and there is little reason to doubt the wisdom of the many decentralized decisions that have led to this aggregate figure. Until proven wrong, there is consequently no reason to change course of action. To the contrary, the rising figure allocated to the social sciences should be taken as an indicator of the revealed increasing worth of investments in the social sciences, given the changing mandate of the CGIAR.2. An important partial assessment of returns from social science research is to attempt to quantify and qualify impact. Monitoring progress in the current effort at impact assessment and using the results to form oneself an opinion about the value of social science research, however imperfect the information may be, is consequently important to give at least a subjective answer to the question. For as long as impact analysis shows existence of positive synergies between social science and biophysical research, expansion of social science research should be encouraged.
3. Cost reduction, without diminishing the value of the effort, should be pursued. This suggests in particular seeking greater collaborations with non-CG centres of excellence in social science research applied to problems of food, rural poverty, and the environment. In general, it seems that these collaborative initiatives with non-CG actors could be significantly expanded.
The Panel consequently recommends that the current share of CGIAR resources invested in socioeconomic, policy, and management research be maintained, that a thorough analysis of returns from investment in social science research be conducted through impact assessment, that avenues to reduce costs through increased collaborations with external partners be further explored, and that ever more stringent criteria for the selection of research budgets among competing alternatives be applied.
6.2.1. Research on research policy and research management
6.2.2. Research on institutions
6.2.3. Political economy of policy and management decisions
6.2.4. Research on common property resources
6.2.5. Generic policy analysis: a loss of opportunities?
6.2.6. Microdata
There has been a clear increase in the role of the private sector in agricultural R&D, both profit and non-profit. This is due to two forces. On the demand side, it has resulted from democratization and decentralization of governance, increasing organization of civil society, and structural adjustment policies which have reduced the role and the budgetary capacity of the public sector. This has led to the privatization of many agricultural services including variety development, seeds, animal health services, marketing boards, water delivery, and technology transfer. On the supply side, advances in biotechnology, stronger intellectual property protection, improved information systems that facilitate access to scientific and market information, and more open market policies have expanded the range of technologies whose benefits can be appropriated by the private sector. For technologies which remain public goods, the supply side has also frequently been privatized under funding by foundations or national and international grants. Transfer of services to the private sector has been unevenly successful and the distribution of benefits from the change has been unequally distributed across the farm population, with some sectors benefiting while others have been left unattended. The CG focus on research management consequently needs to be broadened to understand the whole spectrum of public, non-profit, and private actors involved. Steps have been taken in this direction, with ISNAR working with NGOs, and looking at the complementary roles of the private and public sector, particularly in Latin America.
The panel consequently recommends that the CG mandate on research policy and research management be broadened to capture not just the traditional public management dimensions of R&D, irrigation, and forestry but also the public-private interface, the co-production of public goods by public and private sectors, and the role and management problems of NGOs and other non-profit organizations of civil society.
An area of economic theory where there have been major advances in the last fifteen years is the "new" institutional economics. These theoretical advances have allowed to bring behavior into traditional institutional economics and to thus understand the determinants of the choice of institutions for efficiency purposes. This body of theory postulates that technological and institutional changes are brought about to reduce transactions and transformation costs (North and Wallis, 1994). In particular, market failures due to transactions costs or to problems of imperfect information induce the quest for institutional alternatives that allow to achieve the objectives sought (production, consumption, or exchanges) more efficiently than through markets. This theory has been highly successful in the field of agrarian institutions, for instance to explain the choice of land tenure contracts (sharecropping versus fixed rent), optimum farm size (family versus commercial farms), labor relations (permanent versus seasonal labor; time versus piece rate wages), vertical integration, the Grafting of informal financial institutions (group lending, savings and loan cooperatives), the organization of cooperative users associations (water users' associations, service cooperatives, social forestry schemes), etc. It has also been important in the field of the environment, for instance for the definition of rules for the management of common property resources and the definition of incentive contracts for the conservation of resources. Finally, it has also been highly successful in the field of industrial organization and management to answer such questions as why firms are organized the way they are, how to design incentive compatible regulatory agencies, and how to design incentive systems in firms and organizations.
In general, use of this approach has been seriously lagging in CG research. Notable exceptions include work on intra-household food distribution and inequality (e.g., Haddad and Kanbur, 1990), and work on mutual insurance schemes and consumption smoothing using the ICRISAT panel data (but largely done by non-CG scientists). We are aware of the fact that some of the research on institutional economics has been highly academic, with little of relevance for normative policy purposes and risks of engaging in marginal issues for the CGIAR. Yet, this is not the rule. Very profitable uses have been made of this approach, ranging from socioeconomic research on household behavior under imperfect markets, to the Grafting of agrarian institutions in response to the vacuum left by state contraction, and to the design of research and management institutions for technology and water among others. In management research, institutional issues arise in the design of the optimum interface between private and public roles. For germplasm conservation, questions of incentive structures for in-situ conservation require the design of implementing and regulatory agencies.
This Panel recommends that all Centres make a serious efforts to get better acquainted with the potential that the "new" institutional economics offers for their socioeconomic, management, and policy research. It suggests that scientists concerned in different centres collaborate in systematically exploring areas where worthwhile advances could be made using this approach.
In describing the logic of Figure 2.1, we observed that most CG policy research had addressed the question of policy impact and policy advice, while little had been done on policy making and policy-making advice. Yet, the full payoff from policy impact analysis requires that it be imbedded within policy making analysis. This is because the political feasibility of policy recommendations requires assessing policy making. For instance, recommendations of decentralized governance for participatory rural development may be vacuous, as shown by the experience with community development in India, without an understanding of the decision-making process at the local level. In policy research, as we discussed it in chapter 2 of this report, understanding the policy-making game is fundamental for both positive analysis and for policy-making advice. Equally, important is to assess the credibility of policy in the sense of time consistency. If policy making institutions are not credible, for instance a Central Bank which is not autonomous from government and need to assume deficits created by soft budgets constraints, neither is monetary policy credible.
The need for political economy analysis also applies to management research. It is important in water management since much of the planned allocation of water is perverted by rent seeking. Patterns of rent seeking need to be understood to improve the efficiency of existing systems, to facilitate the transition to water markets, and to design effective water users' associations. It is also important in research management since rent seeking pressures distort the socially optimal or CG-optimal allocation of resources. In addition, not all resources that enter a research system contribute to research, with extensive deviations of funds in response to rent seeking forces in some countries.
In chapter 2, we also observed that it is not necessarily easy for the CG to go into this difficult and conflictual field of research. This suggests that cooperative ventures between CG and local research institutions may be the appropriate way of organizing this research.
This Panel is informed of previous recommendations made by internal and external reviewers as well as of past attempts to develop this important area of research and of modest ongoing initiatives. It recommends that every effort be made to continue to explore possibilities of making headways in research on the political economy of policy and management decisions.
The theme of common property resources appears in the research agenda of many Centres. Indeed, this is an important issue for many resources with which the System is dealing: open access and community crop and range lands, irrigation systems, community forestry, and open access fisheries. The topic has become very fashionable in universities, in part because it is quite interesting from a theoretical standpoint. However, there is little good empirical work. Like for village and household studies, the CG could make a distinct contribution by organizing and sustaining data collection on the impact of common property on performance. Elinor Ostrom's work has attracted attention because she was able to collaborate with strong theoreticians while keeping the theoretical exercises focused on relevant questions (Ostrom, Gardner, and Walker, 1994). The CG could pursue this example, seeking collaborations with universities where there is good theory and helping make the link to empiricism and policy on the types of questions relevant for the CG.
The current large emphasis given to research on common property resources in the System seems fully warranted and this Panel suggests that this effort should be pursued, particularly in terms of solid empirical research.
There is somewhat of a contradiction between the mandate given by the CG to its Centres, namely to give priority to strategic research, i.e., to research which is producing international public goods, and policy research since policy is fundamentally an issue for which the field of application is the nation state, with all its specificities and idiosyncracies. Cross-country policy studies in which IFPRI engages are appropriate to isolate the common dimensions of policy to a number of countries and to recommend a set of generic policy adjustments. In general, IFPRI's research has shied away from close intermingling with particular country governments, except at a fairly abstract level. If generic research ends up extracting generic policy conclusions and making generic policy recommendations, much of the potential value of research for policy advice may be lost. This is because, the bulk of policy advice is highly country specific as well as specific to a particular contextual instance (conjoncture), for example in terms of the particular coalition of forces that prevails at the moment when policy advice is made.
There are several solutions to this dilemma. One is to consider that IFPRI's role is more to demonstrate how to do policy research and to suggest an array of alternative policy options than to make country specific policy recommendations. If this is the case, IFPRI must be careful that the research teaching function is indeed fulfilled and that specific countries are able to extract the country level policy recommendations that derive from the analysis. Leaving this to the initiative of the reader of a generic study is too much to expect. The second solution is to engage in collaborative research, as IFPRI does, explicitly giving to country counterparts the responsibility of extracting country-specific policy conclusions. This requires more than simple co-participation of nationals to the research effort. It also means that IFPRI must assist in the national effort at extracting specific policy conclusions. This can be done in the context of preparing materials for national policy fora, as has indeed sometimes been done. We shall discuss this further in chapter 8 where institutional strengthening is considered.
A related issue is that many of the policy studies recommend partial policy adjustments without resetting them in the country's global budgetary and political processes that establish the opportunity cost of these recommendations. For instance, most of the research on infrastructure simply recommends that more is better than less, but without establishing the opportunity cost of additional investment in infrastructure, and without asking from where funds would be coming. Again, correcting this would require developing a comprehensive view of the country's options and the processes whereby these options are selected and implemented. It would consequently require getting more immersed in the country's policy problems and the constraints under which policy options must be established.
The implication is that more comprehensive country-level policy analyses may be needed. This would significantly recast the definition of research priorities, for instance focusing on countries engaging in significant policy reforms, say Mexico or Vietnam, and proceeding with methodologies for comprehensive assessment of the set of policy options open to them in a particular time frame.
The Panel does not claim to have the answer to these difficult questions. It recommends that IFPRI and other Centres involved in policy and management research devote some thinking to the issue of generic versus country-specific policy studies before the next EPR and that this issue be addressed by the EPR itself.
Many CG Centres and NARSs have conducted household-farm or farm surveys, for instance ICRISAT for household-farm and IRRI for farm surveys. These data sets vary in quality and in comparability. Low-quality data are essentially useless. Sometimes, low quality is due to limited design and scope. For example, the T&V Extension Systems all collect Monitoring and Evaluation data. To this Panel's knowledge, none of these data bases has led to published work, even though millions of dollars have been spent in Africa. Sometimes, the low quality is due to poor management and control (you cannot manage these from Washington). We are not advocating that existing data be centralized. On the other hand, there clearly exists a need to better document existing data bases, both biological and socioeconomic, and to advertize opportunities for collaborative analysis, particularly for post-graduate students seeking data for their theses.
There is also a need for a more systematic process of regular surveying through the CG-NARSs system, including and especially through panel data. This is again an instance where the outposting of IFPRI senior staff for joint projects involving primary data would be mutually beneficial. There is now enough experience to proceed to a case format in panel data collection - with local variations and special studies. Agricultural research and extension systems have a stake in measuring farmers. So far, this has not been possible, but it should be possible. In general, economists are running out of data and theory has made advances which have not, for this reason, been followed by empirical validation.
Centres have made selective significant contributions to the generation of primary data sets that have had fundamental influence on CG research and on socioeconomic, management, and policy research in general. This Panel recommends that, whenever consistent with the research effort, the very important contributions of the Centres to primary household and farm-level data generation be sustained.
