4.1. Size of the CGIAR Commitments
4.2. Duplication and Gaps
4.3. Organization and Governance
4.4. An Alternative Organization?
4.5. Devolution
4.6. Devolution of Specific Commodities
4.7. Production Systems And Management Research
4.8. Institution Building
4.9. Training
4.10. Relations with Partners
4.11. Impact Analysis
The IARCs 1992 commitments in West Africa were about US$55 million or some 25-30% of the total of NARS plus IARCs. This is perhaps the highest regional share in the world and is higher than can be justified on allocative efficiency grounds. Reasons for this overinvestment are: (i) weak national capacity; (ii) unusual importance of livestock; (iii) a complex set of technical problems (section 2.1); (iv) inability to transfer wheat, rice, and potato germplasm from abroad on the same wide scale that has been done elsewhere, making it imperative to concentrate on difficult crops like millet, sorghum, and rainfed maize; (v) rice is mainly rainfed, grown in disparate environments, and hence more costly to improve than irrigated rice in Asia; (vi) yam is rare elsewhere making West Africa necessarily the focus with attendant high fixed costs; (vii) almost complete dependence on a single food grain legume, cowpea, the others (beans, chickpea, pigeon pea, lentil, faba bean, mung bean and the other grams) being of no importance in the region. Cowpea is a multipurpose crop-grain and leaf for food, leaf for fodder, a soil cover and a weed suppressor - that has been particularly difficult to improve; and (viii) political Balkanization, which raises costs of political and economic interaction.
We expect these commitments to stay high. Reasons iii (unusually complex problems), iv (inability to transfer results), v (rainfed rice), and vii (cowpea) are unlikely to change much in the medium-term and will accordingly justify more agricultural research in West Africa than might be indicated by congruence analysis.
The failure of Nigeria's national program to thrive prevents the IARCs from devolving yam and cowpea research fully (reasons (vi) and (vii)). This is also unlikely to change in the medium-term.
Duplication within and across Centers can occur among research on commodities, themes or climates at one time, among facilities at one time, and on commodities over time. Overlap with national programs is discussed in section 4.5 on devolution and in section 4.10.1 on NARS views on IARC interactions.
Duplication on commodities. The Desk Study, the written responses of the Centers to the Desk Study, and the Panel's visit to West Africa permit the conclusion that there is no significant duplication of research on commodities. Adequate collaborative agreements exist between CIMMYT and IITA on maize, between CIAT and IITA on cassava, and sweet potato was transferred from IITA to CIP some time ago. IITA has limited rice research to pre-breeding activities and a comprehensive agreement exists between WARDA and IRRI 9. ICRAF is the lead institution in a Sahel agroforestry network that includes ILCA and ICRISAT in addition to national and regional partners. The substance of the Centers' responses to the Desk Study was that charges of duplication were exaggerated and we accept that.
9 We say nothing further about the WARDA/IRRI relation given that there has been an Inter-Center Rice Review.
Duplication on themes and climates. Duplication occurs on striga, some plant diseases and insects, soils, agroforestry, and economics and public policy studies. This is inescapable from several perspectives. From a practical point of view, one example is that ICRISAT has agroforestry research at Niamey because trees are a necessary part of the ecoregional approach to the semiarid zone. It is unavoidable from a commodity and an ecoregional perspective because different themes manifest themselves diversely in different crops; eg, striga biology and control in sole crop millet in the Sahelian zone differ from striga biology and control in intercropped maize in the moist savanna. Serious and constructive mechanisms exist to reduce the costs of unnecessary duplication. In the natural sciences. Center staff are in close contact on common themes, through networks, joint experiments, scientific meetings, exchange of breeding materials, field visits, use of common methods, and sharing of laboratories, farms, and village research sites. The problem in the social sciences is not so much duplication on themes and climates as lack of leadership. The Panel has made a recommendation to develop this leadership (Section 5).
Duplication of facilities. The Panel reviewed possible duplication at two sites: near Kano in northern Nigeria (ICRISAT, IITA, IIMI) and near Yaounde in the Cameroon (IITA and ICRAF). Near Kano, IITA and ICRISAT now have an effective working agreement to collaborate in administration and communications. ICRISAT has merged its Kano city office with that of IITA. The two work together at Bagauda. The Panel concluded that there was no duplication worthy of mention near Kano and that relations between IITA and ICRISAT staff there looked excellent.
The Panel did not meet IIMI staff in northern Nigeria (there were none at post during the Panel's visit after the departure of a principal staff member). From discussions with IITA, ICRISAT and IAR staff about IIMI's role, the latter did not seem well integrated into the activities of IAR or into those of the other IARCs in northern Nigeria. We have no recommendation other than to endorse the IIMI External Review's observation that "there is a need to rationalize and justify IIMI's West African program so that more knowledge generation and institutional strengthening can be demonstrated." (TAC Secretariat, 1994, p. 44).
There is some duplication between IITA and ICRAF at Nkolbisson and Mbalmayo. This is explained by the history of IITA in Cameroon, where it had a large bilateral program that closed in 1992. The quick expansion of ICRAF within the national program (IRA) has caused some duplication. ICRAF has apparently decided that facilities within IRA at Nkolbisson, partially independent of those of IITA at Mbalmayo, are needed for the time being so as to be more closely integrated with IRA. Although it is obviously better that IITA and ICRAF work together in the Cameroon, it would be hard to argue that this duplication is a serious cost. The Panel estimates, based on information provided by IARC staff in Cameroon, that the cost of duplicate facilities (labs, offices, and fencing) is about US$500,000, or some US$12,500 in annual depreciation. This issue, including the eventual disposition of the physical facilities, should be taken up by the current IITA review.
Duplication over time. The Panel is concerned about duplication of research over time. IARCs sometimes repeat earlier studies; examples are intercropping agronomy in northern Nigeria, crop residue and rock phosphate management in Niger, Striga research, the proposed farming systems characterization in northern Nigeria, and the proposed IFPRI/ILRI research on livestock economics. We suspect, though we cannot prove it here, that there is duplication over time in crop improvement, notably in work on biotic stresses.
The standard remedy for duplication is to recommend yet another coordinating mechanism. But those mechanisms - networks, scientific meetings, joint trials, field visits, literature reviews - exist and do not seem to have uprooted the problem of temporal duplication. The problem is one of management of scientists, not an institutional one. We consider it improbable that there are any broad recommendations that can solve this problem across Centers. It has to be solved within Centers through aggressive review of research proposals before they begin and of research results as they become available.
Gaps. The list of gaps is surprisingly short. Our interviews with NARS scientists and managers suggested that the most serious gaps they perceived in the Centers' portfolio were on cotton, tree crops (mainly coffee and cocoa), and research on irrigation wand water management. The question of work on cotton, coffee and cocoa has been raised often in the past and we have nothing to add; there appear to be good sources of research from national, bilateral, and commercial sources on these crops.
4.3.1. The Ecoregional Approach and the Commodity Focus
4.3.2. Centralization, Administration Costs, and Staff Composition
4.3.3. Administration Costs
4.3.4. Staff Composition
The key innovation of ecoregional research has been said to be "linking the natural resources base to commodity production research". In judging Center conformity with the approach, our operational definition was that it should: (i) focus on the possibilities of the whole environment (the system), not just on selected commodities; (ii) emphasize interactions among subsectors and resources, such as crops and livestock or trees and soil fertility; (iii) examine the growth of agricultural systems over time; and (iv) account systematically for natural resource and environmental costs. We cite only salient examples in what follows.
Do crop specific mandates impede the ecoregional approach? A past answer may have been "yes" but it is now "no". The best evidence is the programs of IITA and ICRAF. IITA's cassava studies include biological pest control, a long-term collaborative effort on production systems, crop improvement and crop management. At Ibadan, IITA and ILCA study interactions among cattle grazing regime, maize cultivar, and crop management. ICRISAT works on agroforestry at Niamey and Bamako, as ILCA has for some time at Ibadan. The leading technology proposed by ILCA for central Nigeria, fodder banks, directly exploits links between crops and livestock in improving the land resource, while easing the transition to settled stock raising.
Should the Centers move farther away from a commodity focus?
There are unacceptable costs in moving farther away from a commodity focus. The basic incentive reason for a commodity focus is that farmers produce commodities after all, not environmental goods. A second reason is that commodity research has positive environmental effects by creating technical incentives to use high potential areas and not low potential ones. A good example is the selective intensification hypothesis - the idea, being tested by WARDA and CIAT, that intensification of the main technologies for the major crops (e.g. rice, pastures, vegetables) on high productivity areas will reduce population pressure on lower productivity, more environmentally fragile, areas. Another good example is ICRAF's MTP for the humid lowland tropics and the semiarid lowlands of West Africa, which gives cogent reasons for its choice of commodity species and shows how those species fit into the ecoregional model. Third, it is possible to achieve benefits of the ecoregional approach through commodity research. For example, there are environmental gains from the biomass effect of higher crop yields through: (i) longer soil cover into the hot, dry season, thereby reducing wind erosion; (ii) longer supply of crop residues into the dry season, thereby reducing grazing pressure on marginal pastures and browse; (iii) heavier soil cover when rainfall is intense, thereby reducing water erosion; and (iv) more crop residues restored to the soil, thereby replacing organic matter lost to cultivation.
TAC has recently discussed alternative institutional arrangements for the Centers in West Africa. The discussion appears to be a response to criticisms that: (i) the current structure is too centralized; (ii) its administration costs are too high; and (iii) its interactions with partners are not as efficient as they could be.
How centralized are the IARCs in West Africa?
The major IARCs in West Africa were more centralized in 1992 than in 1986, as defined by a diversity index 10. The index was 0.66 in 1992 and 0.87 in 1986, as calculated from the data in Table 4. Desk Study data show that 71.2 % of all West Africa CGIAR scientists were in Niger and Nigeria in 1992, compared to 29.2 % in 1986. (It is understood that not all IITA staff in Nigeria are at Ibadan, but most are). The closure of the IITA project in the Cameroon explains some of the difference between 1986 and 1992.
10 The diversity index is (1-(S i ((xi/X)2))), where x, is the number of scientists in a country and X is the total number of scientists in a region. An index of 0 means no diversity, or all staff at one site, and an index of 1 means full diversity.
What are the costs and benefits of centralization?
Costs. One cost of centralization is excess effort on the commodities and farming systems where scientists live. Its possible significance is indicated by the publications of IITA and ICRISAT, by far the most important West African IARCs; they show that staff location affects research location, as indicated by where the published trial was conducted. The ICRISAT Sadore site represents perhaps 5 % of the West Africa semiarid tropics, but produces a much higher share of publications. The Panel talked about this cost with ICRISAT groundnut scientists at Niamey. They note, correctly, that they have trials elsewhere in Niger and outside that country, which do not necessarily lead to published papers but which are a valuable counterweight to centralization.
A second possible cost of centralization is the neglect of information, embodied in local plant materials for crop improvement, in data about hot spots for resistance screening, in the knowledge of national scientists and farmers, and in the characteristics of farming systems. Such costs may be high because of the variability in regional farming systems and in the relative paucity of existing information about West African agriculture before about 1970. Every national program visited by the Panel stated that international research ignored some characteristics of local materials. ICRISAT millet improvement is said, by some NARS staff interviewed by the Panel, to have neglected local materials, who argued that this neglect explained that Center's lack of impact (see section 4.10.2). IITA admitted that most farmers in northern Nigeria could not use its cowpea bred for solecropping with high pesticide inputs (IITA 1993a) and that this was partly due to a misperception of how those materials fit into the cropping system 11.
11 One cultivar was TVX3236, which the Impact Study counted as having "made a widespread impact" (Jahnke et al, 1985, p. 71).
Benefits. The main benefit is that the facilities and staff at a central site - what is sometimes called critical mass - stimulate scientific output above what can be achieved by small groups of scientists. A second benefit is economies of scale in support facilities and social amenities. Many Center products - e.g., training and methods - are largely independent of where they are developed, but cannot be generated without support facilities. A third benefit - which we admit to be vague and hard to distinguish from the economies of scale effect - is that scientists in larger central programs have better incentives for productivity than do those in smaller orbital programs.
Centralization is an issue only for IITA and ICRISAT as the other Centers are too small in West Africa to be affected by it. We do not think that it is a problem for IITA given its work in Benin, Cameroon, Côte d'Ivoire, and the spread of its activities across the varied climates of Nigeria itself; the last point is crucial and is often forgotten by IITA's critics. With due regard for the benefits of centralization, we do think that ICRISAT's effort in the Sahel has been too centralized, not because of size in and of itself, but because of the principal site location at Niamey.
One assertion is that administration costs are too high in some Centers relative to the others. The Desk Study found significant differences among Centers in relative administration costs. The minima were IIMI and ILRAD (none) and the maxima were WARDA, IFPRI, IITA, and ICRISAT (28, 33, 38, and 43 % respectively). The range might indicate savings were the expensive centers to adopt the management practices of the cheap ones.
The high administration cost centers are those with fixed assets, with the exception of IFPRI. How much do fixed assets contribute to administration costs? We first divided the cost of administration (column (1) in Table 8) by the total cost of the other three components (research, research support, and institution building), giving column (3); we then subtracted the annual amortization of each Center's fixed assets from its administration costs (column (2)) and divided the remainder by the total cost of the other three components, giving column (4) below. Though this adjustment does not affect the rankings, it does make the burden of administration at ICRISAT, IITA and WARDA look a bit lighter.
Table 8. Effect of fixed asset amortization on administration costs
|
|
Total |
Less Amortization |
Relative to total of other 3 categories
|
|
|
(US$ thousands) |
||||
|
|
(1) |
(2) |
(3) |
(4) |
|
ICRISAT |
4,032 |
3,169 |
76% |
59% |
|
IITA |
8,620 |
7,131 |
62% |
52% |
|
IFPRI |
800 |
800 |
50% |
50% |
|
WARDA |
2,000 |
1,455 |
38% |
28% |
|
CIP |
400 |
387 |
29% |
28% |
|
ICLARM |
28 |
28 |
23% |
23% |
|
ICRAF |
210 |
138 |
16% |
10% |
|
IPGRI |
55 |
44 |
15% |
12% |
|
ILCA |
628 |
546 |
15% |
13% |
|
ISNAR |
170 |
170 |
14% |
14% |
|
CIMMYT |
46 |
46 |
9% |
9% |
|
IRRI |
52 |
27 |
7% |
4% |
|
Total/average |
17,042 |
13,942 |
45% |
37% |
Note: CIAT, IIMI, ILRAD and INIBAP reported no administrative costs.
Source: Table 4.
The Panel believes that some of those savings are spurious because some Centers provide services to others that are not reflected in the Desk Study averages. At Ibadan, IITA provides administrative services to ILCA, IRRI, ICRAF, and CIP. At Niamey, ICRISAT does the same for ICRAF, IFDC, and IFPRI. We expect that the same will eventually occur in the Cameroon as other Centers join ICRAF and IITA. At Bamako, Kano, and Cotonou, ICRISAT or IITA give administrative aid to individual Center staff. The Panel concludes that a more accurate accounting of inter-center transfers would reduce some of the apparent dispersion in administrative costs 12.
12 The Panel is not recommending more work to estimate these costs.
While part of the range among Centers is an artifact of different reporting practices - e.g. putting research costs under administration and vice versa - and another part is due to hidden transfers among Centers, it appears that administration costs of ICRISAT, IFPRI, and IITA are still high. In the instance of IFPRI, which has no fixed assets in West Africa, the high value is caused by its practice of calculating full overhead costs, a practice which is apparently not adhered to by some of the other Centers.
Another assertion is that administration costs are higher than those of like institutions in the industrial countries. TAC has remarked (1994, p. 14): "The share of center governance and management costs in several centers seems to be high relative to comparable institutes outside the CGIAR." According to this reasoning, savings could be had by adopting the management practices of such like institutes.
The Panel does not have the resources to evaluate this argument in detail, as it would require a careful sampling of relevant like institutes, but it will say the following. (i) Some CGIAR administrative costs are dispensations to partners. The Centers manage donor funds for collaborative research with national or regional programs. They organize foreign travel, stipends, local transport and visas for trainees beyond what is accounted as training expenditures. The Centers organize various meetings and are, inevitably, the donor of last resort for minor expenses of all kinds. These costs appear as administration, yet they cannot be eliminated without harming partner relations. IITA training costs, in particular, are reported in its core expenditures for administration and general operations but include information services, which that Center contends is research support, not administration. (ii) Other administration costs are a fact of life in the tropics because of the high unit costs of transport, communications, information, banking, water, and power. High unit costs swell the share of administration because that activity uses more of those services than does research. (iii) With respect to infrastructure costs, TAC (1994, p. 17) notes that some Centers incur zero costs of physical plant operation because they use rented facilities in comparison to six others having infrastructure costs of more than 10% of total cost. This is not a just comparison unless the rented facilities are free.
Another governance issue is staff composition. Some Panel interviewees recommended that the Centers have an explicit policy of hiring more regional scientists.
The Panel strongly discourages an explicit regional staffing policy because it would interfere with the scientific independence of the IARCs. Moreover, no policy is necessary as regionalization of internationally recruited staff is occurring anyway, for good reasons. Those reasons include the greater numbers of qualified regional staff who, in the francophone countries, are becoming more competent in English. That regional staff stay longer in the region and in one center is a significant benefit because it reduces learning and other fixed costs per scientist. Regional staff are sometimes more effective because they have better language skills. These market forces will promote a progressive regionalization of scientists without an explicit staffing course.
TAC has outlined an alternative "to an ecoregional approach for research in West Africa [that] would consist of a decentralized network of the CGIAR activities" (TAC Secretariat, 1994: p. 26). A "coordinated set of decentralized but focused programs"... would allow for a CGIAR presence in all the major agroecological zones of west and central Africa, a better integration of CGIAR research activities, and for a coordinated network program in both the francophone and anglophone countries of the sub-region" (Ibid).
The Panel contends that the "decentralized network" alternative would be inefficient for three reasons. We assume that a decentralized network would consist of administratively separate partners (eg, international scientists, NARS, non-governmental organizations, the private sector) with access to common funds in place of the center model.
The Panel proposes a common Board of Trustees for WARDA and IITA as a means of harmonizing research between the two institutions 13. It should not be yet another regional organization because of the risks of political interference. The Panel recommends that (i) a Common Board be created from the existing Boards of IITA and WARDA; (ii) two ICRISAT, one IRRI, and one ICRAF Board members serve as ex-oficio members of the Common Board; (iii) the size of the Common Board not exceed (iv) the WARDA Council of Ministers serve in an advisory capacity to the Common Board, without specific recognition in the constitution of any Center except WARDA.
13 In an earlier draft of this Report, we committed the error of justifying this common Board in terms of cost savings, which we now recognize to be unimportant, half that of the combined existing Boards of IITA and WARDA; and
The evolution of some national programs logically shifts the comparative advantage of the Centers in germplasm enhancement and breeding, production system and management, and institution building.
Germplasm enhancement and breeding (CGIAR activity category 2) took about 19.8% of Center commitments in 1986 and 22.0% in 1992 (Table 4) 14. This compares to 49.3% and 35.6% for production systems and management (Category 3) in 1986 and 1992, respectively. Resources allocated to Category 2 are insufficient despite the shifts among categories from 1986 to 1992.
14 The shares for all Centers in 1992 (Table 5) do not differ.
There are two other major issues in Category 2 in addition to the quantity of resources: overall research strategy and the degree to which commodity commitments can be devolved to national or regional programs.
Research Strategy
Opportunity costs of hybrids vs varieties. Hybrid and variety development share costs - farm overhead, labs, computing, support staff - that can be reallocated among research products and the intermediate products of differing strategies can be adapted to variety or hybrid breeding. Hence the tradeoffs of one strategy against the other in terms of short-run financial cost may not be very large.
The tradeoffs in terms of long-run economic cost may be much greater. The assumption that hybrids are too expensive for small farmers has retarded improvement of maize, sorghum, and perhaps millet. Yet the yield gain from hybrids may now be too large to ignore. ICRISAT staff at Kano argue that a stable yield gain of 25 percent of sorghum hybrids over varieties is feasible in farmers' conditions, a gain which would easily pay for the additional seed and, if projected on even a small area, would pay aggregate research costs.
The assumption that hybrid seed is too risky for small farmers has also been used to justify emphasis on varieties. However, the risks of seed supply are best judged by the farmers themselves, not by international researchers. It is inefficient for the IARCs to neglect hybrids as a research strategy on the grounds that seed supply is uncertain given that they are necessarily less well-placed than farmers to judge the costs of uncertainty over time.
While we understand the argument that private sector companies will generate hybrids in the long-run - and have specifically said what we think the system can do to encourage private sector development for all types of agricultural technology (section 3.2) - in the short run we contend that there should be more emphasis on hybrids.
Finished vs intermediate products. There is a broad consensus that Center crop improvement programs should not produce finished materials. At the same time, the small national programs are too weak to do so. That weakness, which also occurs sporadically in many countries that are otherwise strong, puts pressure on the Centers to help the weaker countries by providing them with finished materials (and other assistance) that they would not normally furnish. This is a specific instance of the small country problem.
The longer-term way to help weak countries is through bilateral programs that emphasize academic training, which is how the stronger national programs developed their capacities in the first place. The shorter-term solutions are through regional networks on the SPAAR model (eg, millet, cowpea, and sorghum) which give materials at various levels of testing and through relations with larger national programs (eg, Nigeria with Togo and Benin, Senegal or Mali with Mauritania, the Gambia, and Guinea-Bissau, Ghana with Sierra Leone). The need to assist the weaker national programs by providing finished materials should not broadly justify provision of finished products to national programs.
Devolution of research to national programs has long been a system objective. An example is the transfer of faba bean to Morocco from ICARDA after the 1988 EPR of the latter. What are the possibilities for devolution of germplasm enhancement and breeding in West Africa?
Candidates for devolution should: (i) be important in West Africa, but less important elsewhere, to maximize the benefits to regional research while minimizing the costs of research foregone outside the region; (ii) have existing regional research capacity, to avoid a lengthy transition to new institutions with high initial costs; and (iii) have existing regional research results, to promote confidence among the donors providing the necessary transitional finance.
Taking the major Center crops of the region - millet, sorghum, rice, maize, groundnut, cowpea, cassava, and yam - rice, sorghum, groundnut, and cassava can be excluded because of their extraregional importance. Maize has been devolved to IITA, which has small programs in crop improvement and management of the crop. IITA has already considered abandoning humid forest zone maize research (but not savannah maize research) and the Institute's next external review should investigate this possibility (section 5 on recommendations about individual Centers). Millet is not a good candidate for devolution partly because of its extraregional importance but mainly because years of effort have not produced really significant field results, suggesting that more basic research is needed. A revised role for the IARCs in millet is already evolving in the SPAAR framework.
Arguments for devolution of cowpea. Cowpea is a more logical candidate. West Africa produces much of the world's cowpea and Nigeria is perhaps the leading world producer. There is regional research capacity and there have been some (limited) results. Cowpea is almost never the major crop in a farming system. The institutional basis for devolution exists in the SPAAR framework for the Sahel. The Panel interviewed NARS and IARC staff about the possibility of devolving cowpea research to a national program in the region. The chief barrier to devolution was said to be lack of money.
Arguments against devolution of cowpea. The most telling financial argument against devolution is that the cost of international cowpea research is small. IITA staff also report that the NARS are unable to carry out strategic biotechnological research on insect problems in cowpea and that significant cowpea germplasm exchange outside West Africa might be disrupted if full responsibility for the crop were to be abandoned by IITA. Cowpea is said to be growing in importance as a legume in cereal-legume (green manure) rotations in irrigated farming areas of Asia and international research is thought to be needed for those areas. Panel interviewees noted that the attempt to devolve yam research from IITA to Nigeria had failed; while this unsuccessful precedent does not mean that the effort is impossible, it does suggest that the essential condition of a strong and stable national program has not been met.
The Panel recommends that cowpea germplasm research not be devolved to the region. The cost savings - 3.25 SSYs annually in IITA's 1994-98 MTP in a total of 99 SSYs (IITA 1992, p. 117) - are small compared to the foregone benefits of research on the crop within and outside West Africa. IITA's cowpea responsibilities do not interfere with national or regional activities in West Africa. They produce benefits outside the region that could not be captured by other existing arrangements. The Panel's recommended devolution of production systems and management research (category 3) to the region would in effect devolve much cowpea research because that crop is by far the most important grain and dual purpose legume.
The leaders in production systems research (CGIAR activity category 3 - Production System Development and Management) are IITA, ICRISAT, and WARDA who contributed about 385 SSM to the 1992 total of 456 SSM. The 1992 ecoregional breakdown was 74% humid and 26% semiarid (62% and 38%, respectively, in 1986). The 1992 total was less than half of the 1986 figure of 912 SSM (Table 9).
Should more of activity 3 be devolved to the national programs? The case for devolving this activity is stronger than that for devolving a commodity. First, many efforts in category 3 have failed to produce appreciable benefits in output: prominent examples are ICRISAT operational-scale trials at Niamey, IITA alley farming research at Ibadan, and ICRISAT intercropping work in Mali 15, ILCA systems characterization in Mali, and the SAFGRAD farming systems effort in West Africa. This argues for shifting the work to where it is cheaper. Second, activity 3 is often site-specific and does not produce international benefits. Third, the national program can now do the work. In many instances, what the Centers do is not different from what national programs do.
15 There are reports of some adoption of improved intercropping techniques in central Mali.
The Panel recommends that activity category 3 be the subject of an explicit devolution policy in IITA and ICRISAT. WARDA's size imposes on it a mode of operation that amounts to such a devolution in practice; were WARDA to expand in this category, then this recommendation ought to apply explicitly to WARDA as well. ICRAF and ILCA are special cases because they work on trees and livestock, in which it is recognized that the national programs are weaker and in which, moreover, there is less that the Centers can do in activities 1 and 2. The other centers either have no work in category 3 (IPGRI, IFPRI, ISNAR), or do little in West Africa (ICLARM, IIMI, CIP, CIAT, CIMMYT).
We earlier anticipated several objections to this recommendation.
TABLE 9: Ecoregional and Country Distribution of CGIAR Activity 3 in West Africa, 1992 (Senior Staff-Months)
CGIAR ACTIVITY CATEGORY 3
|
Country
|
Production systems and management |
Production systems and management |
Total all categories
|
% of AEZ total
|
% of West Africa total
|
|
|
All Centers |
IITA, ICRISAT, WARDA |
|||||
|
HUMID LOWLANDS (HULWA) |
||||||
|
|
Benin |
60.0 |
|
84.0 |
6.3% |
4.6% |
|
Cameroon |
|
|
30.0 |
0.0% |
0.0% |
|
|
Ghana |
|
|
12.0 |
0.0% |
0.0% |
|
|
Côte d'Ivoire |
38.4 |
|
116.0 |
4.0% |
3.0% |
|
|
Nigeria |
244.0 |
|
704.2 |
25.5% |
18.8% |
|
|
Sierra Leone |
5.2 |
|
12.0 |
0.5% |
0.4% |
|
|
Sub-total |
347.6 |
290.3 |
958.2 |
|
|
|
|
% of HULWA |
36.3% |
30.3% |
100.0% |
|
|
|
|
% of West Africa total |
26.7% |
22.3% |
73.7% |
|
|
|
|
SEMIARID LOWLANDS (SALWA) |
||||||
|
|
Burkina Faso |
12.0 |
|
12.0 |
1.3% |
0.9% |
|
Gambia |
2.2 |
|
3.1 |
0.2% |
0.2% |
|
|
Mali |
6.0 |
|
81.5 |
0.6% |
0.5% |
|
|
Niger |
77.7 |
|
221.1 |
8.1% |
6.0% |
|
|
Senegal |
10.4 |
|
24.0 |
1.1% |
0.8% |
|
|
Sub-total |
108.3 |
94.6 |
341.7 |
|
|
|
|
% of SALWA |
31.7% |
27.7% |
100.0% |
|
|
|
|
% of West Africa total |
8.3% |
7.3% |
26.3% |
|
|
|
|
TOTAL |
455.9 |
384.9 |
1,299.9 |
|
|
|
|
% of WA |
35.1% |
29.6% |
100.0% |
|
|
|
Source: Desk Study, Table 3A
What should be the mechanism to devolve Category 3? The cuts in Center funding for category 3 naturally mean that there is less to devolve, but the principles need to be stated clearly. The Panel proposes that:
The root institutional weakness of the NARS is flimsy political commitment to research and extension. It manifests itself in: (i) inability to nurture strong national institutions to replace the departed colonial ones; (ii) lack of qualified staff (though this is improving quickly); (iii) arbitrary political interference in national institutions, for example rapid turnover of unqualified managers and institutions larded with administrators; (iv) weak and variable funding, especially of operating costs, as seen in the decline in average spending per scientist at a time of rapid growth in numbers of scientists; and (v) indifference or hostility to non-public institutions, such as private seed companies and farmers' organizations, that elsewhere promote agricultural technology generation and transfer.
What are the Centers doing directly about institutional development? With the evident exception of ISNAR, the IARCs do little in institutional building as such in West Africa. Much of what is termed capacity building is really training and information. IITA's 1994-98 MTP proposes US$200,000 annually for organization and management counseling, or 0.8% of its 100% base envelope; the remaining US$2.075 million in Category 5 (8.4% of the base envelope) is for training, conferences, and information services (IITA 1992, p. 117). ICRISAT's 1994-98 MTP proposes nothing for organization and management counseling but 15.6 % of the MTP total for institutional building, most of which is training, information, and networks (ICRISAT 1992, Table 3). The Panel expects no significant savings from cutting Centers' activities in institution building as such.
What should the Centers do about institutional development? The IARCs neglect of institution building as such is wholly justified. With the evident exception of ISNAR, the Centers have no comparative advantage in institution building, which requires greater resources, a wider perspective, and political reforms that they cannot effect. Examples of what the IARCs cannot do are in the Senegal - Second Agricultural Research Project, which is supported by the World Bank and is representative of the deeper institutional changes needed to make national research more effective. The project: (i) provides major financial support to the NARS; (ii) strengthens administrative and financial management of the NARS; (iii) strengthens linkages among research, extension and farmers; (iv) establishes a commercial production company to manage revenue-earning activities of the NARS; (v) funds special projects that arise unexpectedly; and (vi) prepares a detailed staff development plan.
A rare exception is research station development. In three instances of which the Panel is aware (Cinzana, Mali; Bengou, Niger; and Kano, Nigeria) such assistance has been: (i) specifically requested by the national program; (ii) the beneficiary of special funding; and (iii) done at a site useful to the Center's research. These individual opportunities will necessarily become less common as the facilities in each country mature. They should only be part of broad Center assistance to station development through ISNAR.
Several reviews have detected possible savings in consolidating similar training activities. This argument has merit and it is one on which the Centers are acting. As noted in the Desk Study, there have been some joint training courses, and a Training Directors proposal to make IARC training more cost-effective. The Training Directors of IITA and ILCA have written a proposal for an "Inter-Center Training Program for sub-Saharan Africa". The proposal seeks to: (i) establish common procedures for managing CGIAR training in West Africa; (ii) provide training on integration themes (eg, crop-livestock-tree interactions); and (iii) share facilities for training and publications. The subject seems adequately treated by these initiatives and we have nothing to add.
4.10.1. Summary of NARS' Views on IARC Interactions
4.10.2. Observations of other NARS in Panel Interviews
The principal partners are the national agricultural research institutes, always the main branch of the NARS, farmers, private companies, NGOs, and universities, both national and foreign, and independent research institutes. Are IARC relations with partners efficient? We define efficient as: (i) informing the Centers about what the NARS do and need in full and timely fashion; (ii) creating a partnership in which the presence of the Centers does not smother the national associate or displace what the national programs can and should do on their own; and (iii) maintaining scientific freedom and standards in the Centers without excessive bureaucracy.
The Panel met NARS administrators and scientists in Mali, Burkina Faso, Niger, Nigeria, Benin, Ghana, and the Cameroon to have their views on relations with the IARCs. We first report, in a slightly altered paraphrase, written remarks from the Institute of Agricultural Research and Training (IAR&T) of Nigeria, because they represent many other comments, before summarizing Panel interviews with other NARS representatives.
Observations of IAR&T, Nigeria
1. Geographic mandate of IARCs is too broad (example of IITA).2. IARCs fail to recognize contributions of national programs.
3. IARCs fail to use farming system research (eg, earlier IITA work on mechanization and high-input using cultivars).
4. IARC staff are sometimes unreceptive to views of national program scientists.
5. There is a lack of formal interaction between IARCs and NARS.
6. Discriminatory conditions of service between national and international scientists in the IARCs are a disincentive for the former to work in the IARCs.
7. Incursion of IARCs into the national extension system creates problems for the NARS, e.g. in providing free inputs in an unsustainable manner.
8. Donor politics affect research (e.g., reduction of soybean research due to pressure from a donor).
9. There are duplicative efforts because of donor influence (eg, CORAF and SAFGRAD).
10. There are high administration charges levied by the IARCs on donor funds used in collaboration with NARS.
We summarize these remarks by a few chief themes, defining areas in which one might expect minor or major disagreement between the IARCs and the NARS.
Role and mandate of IARCs
Minor or no disagreement expected
1. IARC research should be basic/strategic or applied.2. An appropriate role of the IARCs is to develop new research methods.
3. Extension (eg, varietal release) is a national role.
4. Exchange of scientists between NARS and IARCs is necessary to strengthen collaboration.
5. Active involvement of NARS is needed in formulation of the research agenda of the IARCs.
6. IARCs should assist the institutional development of national programs.
Major disagreement expected
1. Agronomy, including adaptation trials and crop utilization, should be handled largely by NARS with IARCs collaborating. The Panel believes that the highly site-specific character of agronomy makes it a candidate for devolution to the NARS and has proposed a mechanism to do this (see section 4.5).2. There has been a lack of local partnership in breeding programs (eg, in Burkina). Finished products, not breeding materials, are sent to NARS for testing. There are conflicts over materials to include in trials. IARC breeders are not open enough to local materials in collaborative trials. The Panel makes no recommendation about this point. We believe that the NARS view is not well justified or is something of minor importance that it can be dealt with in the existing IARC-NARS consultation mechanisms in West Africa.
3. IARCs too often duplicate what the national programs do. The national program of Niger cited the example of crop residue management, and cropping practices agronomy.
Collaboration with national programs
Major disagreement expected
1. The IARCs have failed to heed national needs. Examples cited in Mali were impact evaluation and natural resources research.2. Competition for research funds with the IARCs has damaged the national programs.
Organization of the IARCs in West Africa
Minor disagreement expected
1. Bilateral programs are necessary at an initial stage in the growth of the NARS.2. Program instability has damaged the IARCs (Mali, citing case of ILCA).
Governance
Minor or no disagreement expected
1. Regional centers should be established to address problems across ecological zones (eg, West Africa).
Major disagreement expected
1. The regional centers should be manned by nationals. There is a need to strengthen research by involving nationals.2. IARCs are too independent.
3. There is no functional means of linking IARCs to national priorities. Board representation is inadequate. External reviews every 5 years are too infrequent. Annual program reviews and other seminars and workshops are partial. Malian scientists state that they had too little input into ILCA research programs.
4. IARC recruitment of regional scientists sometimes favor inexperienced over experienced staff.
Other issues
Minor disagreement expected
1. CGIAR should pressure NARS governments to allocate adequate funds to research.
Funding
Minor disagreement expected
1. The funding of the IARCs should be conditioned on results.2. A more competitive funding system is required.
3. The CGIAR should work on new crops, such as cotton, other cash crops in the humid zone, or novel species.
Impact
Major disagreement expected
1. There has been little effect of IARC-generated varieties. The examples of IITA and ICRISAT (Burkina) and ICRISAT research on millet in Niger were cited.2. Livestock research has been very weak (Burkina and Mali).
Relations with other NARS partners
Major disagreement expected
1. Contacts with farmers. Center contacts with farmers should not compete with national research and extension. But it is undesirable to set strict and general rules that would not unnecessarily interfere with good science by center staff. The main reason is that farmer contacts, usually through surveys and on-farm experiments, provide valuable information to the IARCs. After all, if one really accepts the criticism that the IARCs have neglected local farming systems, the only just rebuttal is to study those systems.
2. The private commercial sector. National programs in some instances have sought to limit IARC contacts with the private commercial sector on the grounds that the IARCs are working for the NARS, not for profit-making companies.
Panel's comments on those views
The Panel has tried to make a careful evaluation of these observations without going over the details of each instance, and without taking sides. We present here only our comments on the major disagreements (previous section), referring to other sections where we have made recommendations. We have at each point tried to synthesize recommendations that would be broadly applicable to all Centers in West Africa.
Major disagreements about the Role and Mandate of the Centers in West Africa
1. Agronomy/production systems research. The Panel has recommended a mechanism for devolution of category 3 (section 4.7.).
2. Partnership in breeding. Conflicts in breeding are perhaps more frequent than in other fields because of competition for scientific rewards. That competition is the root of this disagreement. But there is a wide variety of existing mechanisms to eliminate it or manage it at any rate and we see no reason to propose others. There is, in particular, enough commonality between the WARDA model and those used by IITA and ICRISAT that nothing would be gained by imposing the former model on the other Centers.
3. Duplication. The national program of Niger cited the examples of crop residue management and cropping practices agronomy, which have been discussed in the comments on Category 3 research.
4. Institutional development. The Panel has made specific recommendations about training and other actions in the general rubric of institutional development (section 4.4.3).
Major disagreements about collaboration with national programs
1. The IARCs have failed to heed national priorities. The principal criticism made by national program staff to the Panel is that the Centers do not consider national priorities when establishing their programs. After serious consideration, we have decided that the criticism is not generally valid enough to warrant a generic change in the operation of the Centers.
2. Competition for research funds with the IARCs has damaged the national programs.
Major disagreements about Governance
1. The regional centers should more often be staffed by nationals at the scientific level.
2. There is no functional means of linking IARCs to national priorities.
Impact
Major disagreement expected
1. There has been little effect of IARC-generated varieties.
2. Livestock research has been weak and has had no production impact. This observation was made in Burkina, Mali, and Nigeria. IER in Mali condemned the ILCA/Mali semiarid program (1975-1985) for having produced only publications in lieu of results that might lift productivity.
Relations with other NARS partners
Major disagreement
1. Contacts with farmers. There is in fact no major disagreement with the principle that Center contacts with farmers should not compete with national research and extension, but there are many specific squabbles. We conclude strongly that it is inefficient to set strict and general rules about contacts with farmers because they would interfere with good science by center staff. Farmer contacts, usually through surveys and on-farm experiments, do provide valuable information to the IARCs; after all, if one really accepts the criticism that the IARCs have neglected local farming systems, their only just rebuttal is to study those systems.
2. Contacts with the private commercial sector. National programs in some instances have sought to limit IARC contacts with the private commercial sector on the grounds that the IARCs are working for the NARS, not for profitable companies. The Panel makes two recommendations here, one about markets for intellectual property and the other about genetic resources exports (section 5).
Impact analysis (IA) is any technique used to measure the consequences of a center's work on commodity output, information, or its partners. Ex-post IA gauges impact of known technology or information on actual output at the farm or aggregate level. A common application is to estimate the fate of planting materials. The standard procedure is to gather farm data to: (i) estimate the area covered by new material; (ii) calculate the additional yield of the new material compared to traditional materials or to previous generations of improved materials; (iii) value the calculated incremental output at market or international prices; and (iv) compare that value to the costs of the research and extension needed to generate it. A variant is a study of cassava mealybug control under the auspices of IITA (Norgaard, 1988) which used more qualitative estimates of yield effects.
Gathering base data. The main techniques are household surveys and rapid appraisal. The former usually: (i) has fewer sample units; (ii) employs more formal sampling procedures; (iii) asks more questions and in more detail; (iv) is less fixed on agricultural technology; (v) employs more quantitative methods, notably econometric, optimization, and simulation; (vi) may last for several years in the same households; and (vii) may cost more per sample unit because of the longer lag to produce results. IITA, ICRISAT, IFPRI, and WARDA have collected farm data at varying sites and locations, and have analyzed some of it with a view to estimating ex-post research impact.
Trials and information analysis. Economic analysis is done of station or farm trials. This can include analysis of breeding programs, distribution of intermediate materials, and publications analysis. Statistical analysis is done of the production function for publications and of the effect of publications on the rate of commodity or total output growth.
Analyses of scientific impact. Scientific impact, defined as the growth in the production of knowledge, is distinguished from production impact, for two reasons. First, scientific impact offers a check on progress in technical variables that cannot be easily measured on farm for cost or statistical reasons. Such variables could include average disease scores in a crop improvement program, yield potential, and estimates of genetic heritabilities of animal traits. Second, there are sometimes good reasons outside the research domain that some truly profitable farm technologies are not adopted. Where those reasons exist, they impede the research system from judging the true rate of technical change and justify scientific impact as a legitimate scale.
Ex-ante IA measures the weight of a technology that has not been adopted, but whose cost, productivity effect, and adoption can be projected with some precision. It is typically an aggregate technique because of the need to calculate the adoption parameters required to repay the costs of research and extension. There has been, however, at least one farm-level application by ICRISAT to the demand for striga control methods on sorghum in Mali.
The Panel perceives some weaknesses in this area.
16 ISNAR series on structural adjustment and agricultural research, with chapters on Burkina Faso and Ghana (Tabor, forthcoming 1995).
17 IFPRI had village studies in Niger with a different purpose, requiring a discrete sample.
It is difficult to separate observations about impact analysis from those about economics research in general and so we may be accused of exceeding our terms of reference in this matter. For that reason, we add several remarks that seem pertinent before presenting our recommendations.
In light of these weaknesses, we recommend that IFPRI be named as a strong convening Center for socioeconomics, policy and public management research (CGIAR activity category 4) in West Africa in order to integrate the microeconomic focus of the commodity centers with its own policy focus. The mechanism is that: (i) all Center social science programs in West Africa, including those of ISNAR, would be prepared in a common process; (ii) an IFPRI staff member, one of two to be based preferably in Nigeria, manage the process; (iii) all special project funding for those programs, including post-doctoral fellows and students, be prepared jointly; and the regional effort not compromise the size of IARC efforts in Category 4. We understand that relevant collaboration already exists between IFPRI and ILCA on livestock policy research, and between IITA and its COSCA partners, but do not believe that this collaboration has produced the quality and coverage of research needed. The proposal creates the potential for a conflict of IFPRI's interest as the convenor and as a research institute, but the other participating Centers will just have to fight for their interests in this category.
In opposing the preceding recommendation, IFPRI argued that "pulling all social science work into a system-wide effort with IFPRI as the convenor is unlikely to likely to result in insufficient microeconomic analysis needed to be undertaken in close interaction with biological sciences." IFPRI did suggest it could be a "convenor of social and policy research that does not fall into the category of micro work... in close collaboration with biological scientists". ICRISAT contended that "social science leadership is developed within each Center to respond to Center priorities and the demands of multidisciplinary research".
We maintain the original recommendation. We add that: i) much of the basis of policy research is in fact microeconomic work in collaboration with the natural scientists; after all, if one wants to estimate the effects of fertilizer subsidies, it is imperative to know the response functions and if one wants to incorporate micronutrients as a crop improvement factor, then it is imperative to know about the systems in which those crops are grown; ii) it is entirely possible to maintain the microeconomic and multidisciplinary focus of social science work in the commodity Centers under the general responsibility of IFPRI; in fact, one of the weaknesses of current arrangements is the lack of such integration; iii) making IFPRI a limited convenor is not likely to have an effect on the behavior of the commodity Centers and bears the risk of encouraging the latter to abandon this area altogether; and iv) what is needed is a series of studies - ranging from economic analysis of experiments to policy analyses across countries - developed and managed jointly under the leadership of IFPRI that would make up a regional program.
