An Ex Post Economic Impact Assessment of Planted Forages in West Africa ILRI Impact Assessment Series 2 E. Elbasha,1 P.K. Thornton1 and G. Tarawali2 1. Systems Analysis and Impact Assessment Project International Livestock Research Institute (ILRI), Nairobi, Kenya 2. Consultant Agronomist, Ibadan, Nigeria International Livestock Research Institute P.O. Box 30709, Nairobi, Kenya Table of Contents Acknowledgements ........................................................v Summary ................................................................vi 1. Introduction ...........................................................1 2. Description of the technology.............................................2 Attractiveness of the technology ...........................................2 Yield ..............................................................2 Affordability ........................................................3 Benefits............................................................3 3. Methods for estimating social returns.......................................4 4. Data and sources .......................................................7 Estimating benefits ......................................................7 Milk yield ..........................................................7 Weight gain or reduced body weight loss..................................8 Calving rate and calf survival ...........................................8 Cow survival ........................................................8 Crop yields .........................................................9 Residue yield .......................................................11 Fertiliser saved......................................................11 Elasticities of supply and demand .........................................12 Research costs......................................................12 Costs of establishing fodder banks ......................................13 Commodity prices, exchange rates, and national production figures ..............14 5. Adoption rates and patterns .............................................16 Benin ...............................................................16 Burkina Faso .........................................................17 Cameroon ...........................................................17 Chad................................................................18 Côte d’Ivoire .........................................................19 The Gambia..........................................................20 Ghana...............................................................20 Guinea ..............................................................21 iii Guinea Bissau ........................................................21 Mali ................................................................22 Mauritania ...........................................................23 Niger................................................................23 Nigeria ..............................................................24 Senegal..............................................................25 Sierra Leone ..........................................................26 Togo ................................................................26 Summary of the adoption data ...........................................26 6. Application of the economic surplus model to estimate impact ..................29 Baseline analysis .......................................................29 Sensitivity analysis .....................................................31 Future adoption ....................................................32 Analysis by agro-ecological zone ...........................................34 7. Discussion ...........................................................35 References ..............................................................38 Appendix 1. Changes in producer surplus and elasticity of demand ................41 Appendix 2. Sample commodity prices, Ghana (US$) ...........................42 Appendix 3. FAO production data for West Africa, 1977–1995 ...................43 Appendix 4. Adoption of fodder bank technology by country and agro-ecological zone .59 (A) Humid zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 (B) Subhumid zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 (C) Semi-arid zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Acknowledgements Many people contributed to this study. In particular, we would like to acknowledge the input of the following collaborators in each country: Republic of Benin : Dr M. Ehouinsou, Mr T. Aimé and Mr P. Vissoh Burkina Faso: Mr G. Godet and Dr J.B.M. Kamuanga Cameroon: Dr J. Kouonmenioc, Dr R.M. Njwe, Dr T. Sippowo and Dr S. Yonkeu Chad : Dr I. Alfaroukh Côte d’Ivoire: Mr B. N’Guessan, Dr E. Mill, Dr C. Kouamé, Dr J. Cesar and Dr M. Becker The Gambia : Mr O. N’Jai, Prof D. Dempfle, Mr J. van Wingham and Mr Y. Akinbamije Ghana: Prof E.O. Otchere and Dr P. Barnes Guinea : Mr M. Diallo Guinea Bissau : Mr R. Daniel and Mr Q. Bancessi Mali: Dr L. Diarra, Mr E. Dembele, Mr M. Bengaly and Dr S. Debrah Mauritania: Dr D.B. Cissé Niger: Mr M. Adamou, Mrs M.S. Dicko, Dr E. Tielkes, Dr S. Fern « ndez–Rivera, Dr T.O. Williams and Dr P. Hiernaux Nigeria: Dr J. Maina, Dr S.A. Ingawa, Dr O.S. Onifade and Mr P. Chionuma Senegal: Mr C. Floret, Mr R. Manley and Dr A. Ickowicz Sierra Leone: Dr S. Maturi Togo: Mr P. Agbemolo-Tsomafo and Mr A. Defly Other people contributed in various other ways. In particular, and without implicating them in any way, we would like to thank Ralph von Kaufmann, M.A. Mohamed-Saleem, Mohammad Jabbar, Jimmy Smith, Shirley Tarawali, Gerard O’Donoghue, Patti Kristjanson and Peter de Leeuw for data, information and comments on earlier versions of the analysis and report. We alone are responsible for remaining errors and omissions. Summary In the late 1970s and early 1980s the International Livestock Centre for Africa (ILCA) and its national research partners were instrumental in developing and promoting the concept of planting forage legumes in an attempt to help agropastoralists in West Africa alleviate the feed stress experienced by their ruminant animals during the dry season. The ‘fodder bank’ was one (but by no means the only) method developed, whereby an area of the farmer’s land was fenced and planted to Stylosanthes or other legumes, which could be used for strategic feeding during the early dry season. Some cropping systems were also developed in which a cereal was subsequently planted in the fodder bank to make use of the nitrogen fixed by the legume. Research funds were spent on this activity until 1993, and various studies have shown adoption of this technology in a number of countries of West Africa, although no comprehensive study of the adoption of this technology had yet been done. In 1995, ILCA and ILRAD (the International Laboratory for Research on Animal Diseases) combined to form a new institute, the International Livestock Research Institute (ILRI). An ex post assessment of the fodder bank technology in West Africa was carried out to document the impact that ILRI, its national partners and other organisations have had and continue to have at the farm level, which could be used to demonstrate the value of investment in agricultural research programmes. Accordingly, a study was designed with two main activities: a literature survey to quantify production impacts of the fodder bank technology, and the commissioning of a consultant to travel extensively in the region to collect up-to-date information on the number of adopters of the technology from national agricultural research and extension programmes in the region. To date, about 27,000 adopters have been identified growing forage legumes on some 19,000 ha in the 15 countries for which we currently have information. Using modest estimates of production impact of forage legumes on meat and milk production from a herd simulation model and on maize, millet and sorghum grain and residue from the literature, commodity price data, elasticities of supply and demand, and estimates of research costs were combined in an economic surplus model with the number of adopted hectares of forage legumes that could reasonably be attributed to the activities of ILRI and its national partners. The baseline analysis indicates that on an expenditure of research resources of just over US$ 7 million, the total net benefits to society that had accrued up to 1997 amounted to US$ 16.5 million, with an internal rate of return of some 38%. These figures may be conservative, for the adoption data are likely to be conservative and the estimated production impacts are modest. Various sensitivity analyses are carried out to test the robustness of these estimates of impact. Partly because research resource expenditure by ILRI on this technology is now zero, projecting adoption trends to the year 2014 results in at least a doubling of the estimated total benefits realised to date. In many places, the problems facing farmers wishing to adopt forage legumes are serious. Planted forage legumes will undoubtedly occupy niches in the farming systems of West Africa, but for the future farmers will increasingly make more use of crop residue material in their quest for feed resources. Despite this, the impact of adopted fodder banks has paid for the research that went into their development at least three times over, and this will increase substantially in the next few years, given current adoption trends. A further lesson from this work is that the lag associated with the diffusion of this technology is considerable—at least 15 years—and may be much longer than is generally anticipated. 1 Introduction The importance of livestock as a means of sustenance, traction power and transport, as a substantial source of nutrients as manure for agriculture, and as a means of hedging against risk and uncertainty in the lives of the people of sub-Saharan Africa, has been appreciated for centuries. However, livestock development in Africa faces many constraints, among which the most widespread is shortage of feed supply (Winrock International 1992). A sustainable solution to feed deficiencies is essential for the huge livestock potential of the continent to be realised. With this understanding, the Subhumid Programme of the International Livestock Centre for Africa (ILCA) at Kaduna, Nigeria, in the late 1970s developed the concept of a fodder bank as one solution to the problem of inadequate nutrition, especially during the dry season, in West Africa (Mohamed-Saleem 1986). A fodder bank is a small area of forage legumes established and managed by an agropastoralist near the homestead as a feed supplement for livestock during the dry season. Inadequate nutrition during this season (arising from ingested forage with a crude protein content of less than 7%) causes animals to produce less milk and lose weight, and increases calf mortality and reduces conception rates. For a large part of the dry season, fodder banks can maintain a crude protein content of more than 9% (Mohamed-Saleem and de Leeuw 1994). As a result, animals with access to a fodder bank perform better than those kept on natural pasture. In addition, the legume (commonly Stylosanthes spp) accumulates soil nitrogen through biological fixation in the root nodules. The legume can also have beneficial impacts on the physical properties of the soil such as bulk density, infiltration rates and field moisture capacity. As a result, crops grown in plots previously planted as a fodder bank commonly produce higher yields than those cultivated outside such areas. The major benefits of fodder banks can be summarised as follows: increases in crop and crop residue yields (for this analysis, we consider maize, millet and sorghum), increases in milk yield and weight gain (or reduced body weight loss), increased calving rates, decreased age at first calving, and increased calf and cow survival rates. Estimates of most of these benefits can be obtained from the substantial literature on fodder banks. This study assesses, in an indicative way, the impact of fodder banks and forage legume technology in countries of West Africa. The economic surplus method is used to estimate the social rate of return to public investment in research on fodder banks and the distribution of that return amongst producers and consumers. Subsequent sections of this report describe the technology of fodder banks, the economic surplus method, sources of data and the results of the analyses. 2 Description of the technology A fodder bank is supposed to be established, managed and utilised as follows (Tarawali and Mohamed-Saleem 1994): a farmer (i) selects and fences an area of land (the recommendation is 4 ha, but this area could be more or less depending on needs and herd size) using either metal posts or live poles; (ii) prepares land for planting by confining animals overnight in the fenced area, by grazing down for 1 or 2 weeks following seed broadcast, by burning, and by using 150 kg/ha of superphosphate fertilisers; (iii) broadcasts scarified seeds at a seeding rate of 10–15 kg/ha; and (iv) at the beginning of the dry season during the labour-slack period, constructs peripheral fire traces to protect the bank from burning. Good management of a fodder bank consists of (i) allowing animals to graze the fodder bank early in the wet season to control fast growing grasses until the legume is well established, and withdrawing animals when Stylosanthes starts flowering to promote high seed production; (ii) allowing forage to bulk up by deferring grazing until the dry season; and (iii) ensuring sufficient seed drop and stubble for regeneration in the following season. To utilise the fodder bank according to extension recommendations, a farmer allows pregnant and lactating animals (up to a maximum of 5 per hectare) to graze the fodder bank for 2.5 h/day during the dry season. Only three cultivars are recommended for use in a fodder bank: Stylosanthes guianensis cv Schofield , S. guianensis cv Cook and S. hamata cv Verano. Of these, Schofield and Cook are susceptible to anthracnose disease while Verano is less susceptible. Other promising species have been identified in evaluation trials over the years such as Centrosema pascuorum , Chamaecrista rotundifolia and Aeschynomene histrix. Attractiveness of the technology Local tropical grasses are the only available alternative herbaceous cover to compare with Stylosanthes. Various features make Stylosanthes legumes superior to grasses (Bayer 1986): they grow on relatively infertile soils with the low nitrogen and low phosphorus contents that are common in West Africa; they secure soil nitrogen through biological fixation in the root nodules; and they have higher dry-matter digestibility and voluntary intake by animals. Yield Several fodder banks consistently produced 4–6 t dry matter (DM)/ha over several years with a 50–70% legume content. Mohamed-Saleem and Suleiman (1986) reported an average DM yield of 6–8 t/ha in some places in Nigeria during the first year of establishment. Mani (1992) reported that after 1, 2 , 3, 4 and 5 years of establishment the respective average DM yield was 5.2, 4.6, 5.5, 5.2 and 5.3 t/ha and average Stylosanthes content was 59, 66, 64, 61 and 62%. Tarawali and Mohamed-Saleem (1994) found the average DM yield and Stylosanthes composition of fodder banks in ILCA’s case study areas to be 6.3 t/ha and 60%, respectively. Affordability Compared with other supplementary feeds such as oilseed cake and other agro-industrial products, S tylosanthes is less costly, more abundant and more readily available. Benefits The benefits of Stylosanthes are achieved through three main channels: increase in total herbage, extension of the period of production of the pasture and increase in nitrogen (N) content (Mani et al 1993). Mohamed-Saleem and von Kaufmann (1995) estimated the internal rate of return of establishing a fodder bank in the subhumid zone of Nigeria to be 22% using 1989 market prices and incorporating the benefits of improved herd productivity alone. When the benefits of reduced forced sales and increased crop yields are included, their estimate of the internal rate of return reached 36%. In the mid-1980s in the subhumid zone of Nigeria, about 75% of the costs of establishment of a typical 4 ha fodder bank were estimated to be associated with the costs of fencing (Otsyina et al 1987). 3 Methods for estimating social returns A widely accepted procedure for economic evaluation of benefits and costs of a technological change is the economic surplus method (e.g. see Alston et al 1995). The basic idea behind the economic surplus method is that technology adoption reduces the per unit cost of production, and hence shifts the supply function of the commodity down and to the right. If the market for the commodity is perfectly competitive, this will lead to an increase in the quantity exchanged in the market and a fall in price. As a result, consumers benefit from the price reduction and producers may benefit from selling a greater quantity. When computing the economic surplus of a successful research activity in a closed economy, economists usually refer to a diagram of the type shown in Figure 1. The demand for the commodity is denoted by D , the pre-research supply curve is S, while the post- research supply curve is S’. The initial equilibrium is denoted as (P 0, Q 0) while the post- research equilibrium is (P 1, Q 1). Figure 1. Effects of technological change on the supply curve. The change in the Marshallian consumer surplus (?CS) is the area P1P0BC . This is given by: where D(P) denotes the demand function. The initial producer surplus is given by the difference between total revenue (area OP 0BQ 0) and variable cost (the area below the supply curve between O and Q 0 ). The new producer surplus is the difference between the new ? CS D P dP P P= ? ( ) 1 0 Q S S’ Q 0 P 0 D Q 1 P 1 F G B C 0 P total revenue (area OP1CQ 1) and the new variable cost (the area below the new supply curve between O and Q 1). The change in the producer surplus (? PS) is the difference between the new and the initial producer surpluses. Formally, the change in the producer surplus is given by: where S(Q) denotes the pre-research inverse supply function and S’(Q) is the post-research inverse supply function. The change in total economic surplus (? TS) is the sum of the changes in producer and consumer surplus (the shaded area GFBC). Thus: To apply this method empirically, the forms of the supply and demand functions have to be specified. A fairly flexible functional form that is widely used for estimation of supply and demand is the constant elasticity (CE) specification. An inverse supply function of the CE type is written as: where P and Q are price and quantity, respectively, ? measures the elasticity of supply, and a is a constant supply shifter. The flexibility of this functional form stems from the fact that it is generated by any production function that is homogenous of any degree. It is also attractive for estimation because a logarithmic transformation makes it linear and hence amenable to estimation by ordinary least squares. The establishment of a fodder bank enhances productivity and hence reduces the per unit cost of production. The after-establishment supply curve can be written as: where k measures the pivotal shift in supply and is related to the productivity change, h, according to the formula: The way in which we estimate changes in productivity as a result of fodder bank establishment is described below. Similarly, a CE demand function is assumed: where ? is the price elasticity of demand. ?/1aQP = ?/1)1( QkaP ?= ?? /)1()1(1 +?+?= hk ??= aPQ dQQSQPdQQSQPPS QQ ?? +??=? 01 000011 )()(' ?TS D P dP P Q S Q dQ P Q S Q dQ P P Q Q= + ? ? +? ? ?( ) '( ) ( ) 1 0 1 0 1 1 0 0 0 0 Plugging these functional forms in the above formulae, the changes in consumer surplus, producer surplus, and economic surplus are obtained: The new equilibrium price and quantity are related to the initial equilibrium according to: )/( 01 )/( 01 )1 ( )1 ( ???? ??? +? + ?= ?= kQQ kPP PSCSTS QQkQPQPQPPS PPQPCS ?+?=? ??++?=? ??=? + ? ])/)(1(1[1 ])/(1[1 1 /)1( 01000011 1 0100 ?? ? ? ? ? 4 Data and sources To calculate the social benefits of fodder bank adoption using the economic surplus method, it is necessary to estimate the parameters of the supply and demand curves for the various products benefiting from the establishment and use of a fodder bank. This includes estimating elasticities of supply and demand and the rates of shifts in the supply functions of milk, meat, maize grain and residue, millet grain and residue and sorghum grain and residue. In the absence of specific data we have to rely on the literature to obtain such estimates. In addition, we require data on the number of fodder banks established in each year, the prices of all these products and the research costs involved. The following data were obtained from a survey in the countries in West Africa where fodder banks were introduced and adopted: 1. The number of fod der banks es tab lished each year in each coun try in each agro- ecological zone. 2. Data on prices of maize, sorghum, millet, milk and meat in each year in the areas where fodder banks were adopted and data on the prices of maize, sorghum and millet residues. To calculate the rates of shift in the supply functions, the change in the level of production per hectare is estimated. This number is then multiplied by the total number of hectares of each product inside the fodder bank in each year. The result is then compared with the total production of the country in each year to obtain the percentage change in the level of production attributed to the establishment of fodder banks. Estimating benefits Milk yield Tarawali and von Kaufmann (1987) reported that lactation yields in Nigeria were 300 kg without supplement and 312 kg with supplement (an increase of 12 kg or 4%). Otchere (1986) estimated the total amount of milk taken from dams supplemented with a fodder bank in Nigeria to be 9.6 litres (9.3%) more than that from the control group (113.2 and 103.6 litres, respectively). Because the recommended management of a fodder bank consists of grazing 5 lactating cows per hectare, the total increase in milk yield as a result of fodder bank supplementation is 5 × 9.6 = 48 litres/ha per year. This excludes the milk consumed by calves, which is about 24 litres more with fodder bank supplementation. If this is included, total milk yield increases to 169 litres/ha per year. In this study we use Otchere’s (1986) estimates because those of Tarawali and von Kaufmann (1987) are less conservative and less comprehensive in the sense that they did not include milk consumed by calves. Weight gain or reduced body weight loss Tarawali and von Kaufmann (1987) reported weight gains of animals at one year as 98 kg without supplement and 103 kg with supplement (an increase of 5%). Bayer (1986) compared weight losses of a control group to a herd grazing a fodder bank. He found that by the end of the dry season, the two groups differed by 20 kg/head and this difference was statistically significant at the 5% level. If we use Bayer’s estimates, we arrive at an increase of live weight of 100 kg/ha as a result of fodder bank establishment and use. Fodder banks are also used for feeding traction animals and supplementing small ruminants. For example, Ikwuegbu et al (1994) showed that West African Dwarf goats on Stylosanthes-based pastures gained 1 kg more weight during the wet season than those on fallows (at 4 does per ha). Because of lack of data on number of traction animals and goats and other key production parameters, we restrict our analysis to benefits outlined in this section only. We recognise, however, that there is strong evidence that the use of ‘mini’ fodder banks for feed supplementation has substantial potential for improving the growth performance of goats in the subhumid zone (e.g. see Ikwuegbu et al 1995). Calving rate and calf survival Tarawali and von Kaufmann (1987) reported calving percentages of 54% without supplement and 58% with a fodder bank supplement (an increase of 8%). They also reported an increase of 20% in calf survival for animals on fodder bank supplement (72% without supplement vs 86% with supplement). Cow survival Tarawali and von Kaufmann (1987) reported 92% cow survival without and 96% with supplement (an increase of 4%). To convert the benefits from increased calving rate and reduced mortality rates into milk and meat, we used the ILCA Herd Model (von Kaufmann et al 1990). The model uses data on herd structure and offtake rates and simulates the dynamics of a herd over a period of 10 years using two scenarios: one with the calving and mortality rates without a supplement and the other with a supplement. A summary of input data is shown in Table 1, taken from survey data from 1984 in Kaduna, Nigeria (R. von Kaufmann, ILRI, personal communication). The results of the model run showed that the increase in calving rate and reduction in mortality rates as a result of supplementation translate into 4 kg of live weight per hectare and 12.6 kg of milk per hectare. Table 1. ILCA herd model input and output data. Age class Herd structure Live weights (kg) Slaughter offtake (%) Breeding offtake (%) A. Base herd structure, live weights and offtake rates Females 1 year 4.0 51.0 1.0 0.0 Females 2 years 3.7 75.0 1.0 0.0 Females 3 years 3.3 112.0 1.0 0.0 Females 4 years 3.0 212.0 1.0 0.0 Females 5 years 4.0 260.0 5.0 0.0 Females >5 years 16.6 260.0 5.0 0.0 Males 1 year 3.9 53.0 2.0 0.0 Males 2 years 3.5 79.0 3.0 0.0 Males 3 years 3.2 140.0 6.0 0.0 Males 4 years 2.1 220.0 26.0 0.0 Males 5 years 2.2 240.0 51.0 0.0 Males >5 years 0.2 260.0 75.0 0.0 Total 49.7 B. Female production data and mortality rates Without fodder bank supplement With fodder bank supplement Calving rate 53.80% 58.10% Age at 1st calving 60 months 60 months Lactation offtake 244 litres 244 litres Lactation length 270 days 270 days Grazing area 50 ha 50 ha Cattle 0–1 years mortality 29.20% 13.70% Cattle 1–2 years mortality 7.80% 4.00% Cattle 2 years of mortality 7.80% 4.00% C. Live weight and milk offtakes, results of 10-year non-steady state simulations Impact of fodder bank supplement (‘With’ minus ‘Without’) Mean annual carcass offtake rate (kg/ha) 4.1 Mean annual milk offtake rate (litres/ha) 12.6 Crop yields Tarawali (1991) estimated that 45 kg of nitrogen per hectare is transferred from the Stylosanthes legume to the maize crop. Stylosanthes also improves the physical properties of the soil through increasing the organic matter content, resulting in lower bulk density and higher infiltration rates and field moisture capacity. As a result, crops grown on fodder banks or experiment stations commonly produce higher yields than those cultivated outside fodder banks. Tarawali and Mohamed-Saleem (1994) have shown that once a productive Stylosanthes pasture is established, it can subsequently be cropped for at least two years and revert to a similar Stylosanthes pasture without requiring re-seeding. We will assume that all countries under study adhere to this system of rotation. This means that for each hectare of fodder bank in the analysis that follows, 0.75 ha is used for Stylosanthes legumes and 0.25 ha is used for crop production. We further assume that the 1 ha is allocated to the production of each of the three crops according to the share of land that each occupies nationally. Estimates of increases in crop yield as a result of fodder bank establishment varied considerably from one study to another. However, benefits to crops were confirmed at various sites and across agro-ecological zones, with effects on yields lasting up to two years (Mohamed-Saleem 1994). Fodder bank management practices had a strong influence on yields of subsequent crops. The increase in yields of maize grain as a result of fodder banks located in northern Nigeria (Hassane 1995) are shown in Table 2. In this study, no significant effects were found on sorghum yields. Tarawali and von Kaufmann (1987) also found positive effects of fodder banks on maize grain yield. The average maize grain yield inside and outside a fodder bank was 4659 and 2545 kg/ha, respectively, implying a productivity effect of 2114 kg/ha (83%). Tarawali (1991) found average maize grain yields of 820 kg/ha outside fodder banks and 1720 kg/ha inside fodder banks, an increase of 900 kg/ha (109.8%) in yield. Table 3 shows positive effects on grain and residue yields for three cereals in subhumid Nigeria (data from Ikwuegbu et al 1994). Table 2. Maize grain yields (kg/ha) under different management in two zones of Nigeria. Adjacent natural fallow with fertiliser Fodder bank Fodder bank with fertiliser (A) (B) (B)–(A) (C) (C)–(A) Zone - - - - - - - - - - - - - - - - - - - - - - - - - - - - kg/ha - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Subhumid 633 739 106 2700 2067 Semi-arid 327 413 86 611 284 Source: Hassane (1995). Table 3 . Grain and residue yields in subhumid Nigeria. Product Maize (kg/ha) Sorghum (kg/ha) Millet (kg/ha) Grain Previous Stylosanthes pasture 2245 1533 724 Old fallow 1654 1476 475 Difference 591 (+36%) 57 (+4%) 249 (+52%) Residue Previous Stylosanthes pasture 3330 4780 3660 Old fallow 3100 4550 3340 Difference 230 (+7%) 230 (+5%) 320 (+10%) Source: Ikwuegbu et al (1994). The estimates of Ikwuegbu et al (1994) of the productivity effects of fodder banks on sorghum grain yield are consistent with those of Mohamed-Saleem et al (1986). We use these in our analyses, as well as those for millet. However, the estimates of the productivity effects on maize grain yields are on the low side when compared with those of Tarawali and von Kaufmann (1987), Tarawali (1991) and Hassane (1995). For this study we use the mean of the productivity effects on maize grain yield of the four studies (1418 kg/ha). We calculate the productivity effects on residue yield in the same way. Residue yield Fodder banks can have significant effects on maize stover yields. Tarawali and Kaufmann (1987) reported average maize residue yields of 7.4 t DM/ha inside the fodder bank compared with 4.3 t DM/ha outside, a difference of 3.1 t DM/ha (75%). However, Mohamed-Saleem et al (1986) showed the effects on fodder yield of sorghum to be less than those on sorghum grain yield. For sorghum and millet residue we use the results of Ikwuegbu et al (1994) and we take the average of the difference reported in Tarawali and Kaufmann (1987) and Ikwuegbu et al (1994) in order to obtain the productivity effect on maize residue yield. Improved soil structure makes tillage easier, and could be expected to have an impact in terms of reducing farm work. Such benefits are not included in this analysis because we do not have ready estimates of the reduced number of hours from easier tillage. The productivity effects used in the analysis are summarised in Table 4. Table 4 . Productivity effects used in the analysis. Product Productivity effect Milk (litres/ha)1 182 Live weight (kg/ha)1 104 Maize grain (kg/ha) 1,418 Sorghum grain (kg/ha) 57 Millet grain (kg/ha) 249 Maize residue and herbage (kg/ha) 1,679 Sorghum residue and herbage (kg/ha) 230 Millet residue and herbage (kg/ha) 320 1. Milk and liveweight estimates include the values obtained from the herd model calculations outl ined in the text (and see Table 1). Because there are no published data on crop residues, we need to estimate their baseline production levels in the absence of fodder banks. From the studies cited above, residue and herbage yields of maize, sorghum and millet, when sown on fallow land, are 3685, 4550 and 3340 kg/ha, respectively. Multiplying these by the areas under crop production gives us the baseline levels of production that are used in this analysis. Fertiliser saved Tarawali (1991) estimated that maize planted outside the fodder bank required 45 kg/ha of nitrogen to produce the same yield as that of maize on unfertilised plots within the legumes pastures. This effect is included in the higher crop yields used in the analysis below. However, using less fertiliser, apart from its possible positive environmental effects, saves foreign exchange (since it is imported) which is in short supply in most of West African countries. The total amount of fertiliser saved, and hence the foreign exchange saved in each year as a result of fodder banks, are likely to be substantial. Elasticities of supply and demand Estimates of elasticities of supply and demand for Nigeria were obtained from Singh and Subramanian (1986). The elasticities of demand for millet and sorghum are 0.08 and 0.05, respectively. The respective estimates for the elasticities of supply for these two commodities are 0.25 and 0.3. An ACIAR (Australian Centre for International Agricultural Research) database (G. Lubulwa, ACIAR, personal communication) reports the same elasticities of supply and demand for sheep and goat meat for several countries of West Africa. These are 1.76 (supply) and ? 1.0 (demand), respectively. In the absence of a better alternative we use the estimate of the price elasticity of supply of meat in this study. Tambi (1996) estimated the elasticity of demand for meat in Cameroon at 1.8. This estimate was used in our study. The ERS/USDA (Economic Research Service/United States Department of Agriculture) database (G. Lubulwa, ACIAR, personal communication) gives elasticities of supply and demand for milk in sub-Saharan Africa as 0.7 and ?0.5, respectively. Delgado and Reardon (1991) estimated demand elasticities for grain (sorghum, millet and maize) from aggregate data in Mali, Senegal and Burkina Faso at ?0.07,?0.11, ?0.50, respectively. We take the estimates for Senegal to represent the elasticity of demand for maize for all West Africa. Jaeger (1986) estimated a Cobb–Douglas farm production function for West Africa and arrived at an average rate of return to scale of 1.02, implying an elasticity of supply of 0.98. We assume that this is an approximate measure of the elasticity of supply of maize. The elasticities used in this study are summarised in Table 5. Elasticities of supply and demand for crop residue are assumed to be the same as those for grains. Table 5. Elasticities used in the analysis. Elasticity Milk Meat Maize Millet Sorghum Supply 0.7 1.76 0..98 0.25 0.3 Demand –0.5 –1.8 –0.11 –0.08 –0.05 Research costs Research costs associated with developing and testing of fodder banks were estimated from annual summary expenditure budgets going back to 1975, when ILCA started operations (G. O’Donoghue, ILRI, personal communication). The total costs, with an overhead of 18%, of the West African programmes from 1977 to 1997 are shown in Table 6; note that total ILRI expenditures for Niger and Nigeria (1995–97) have been split arbitrarily between the two programmes for these years. Table 6. Research costs (US$ ’000s) of ILCA’s (1975–94) and ILRI’s (1995–97) West African research progra mmes. Year Semi–arid Niger Subhumid Mali Subhumid Nigeria Humid Nigeria Overhead (18%) Total costs Costs attributed to fodder bank research 1 1978 0 300 0 0 54 354 18 1979 0 345 0 0 62 407 20 1980 0 345 0 0 62 407 20 1981 0 345 9 0 64 418 22 1982 0 349 169 249 138 905 195 1983 79 838 795 1,062 499 3,273 863 1984 106 781 1,066 857 506 3,316 1,103 1985 102 728 749 812 430 2,821 798 1986 125 941 735 693 449 2,943 790 1987 125 997 498 451 373 2,444 556 1988 183 973 611 497 408 2,672 664 1989 405 924 606 696 474 3,105 668 1990 529 1016 776 579 522 3,422 827 1991 565 819 686 528 468 3,066 241 1992 683 668 768 554 481 3,154 253 1993 707 411 571 498 394 2,581 188 1994 622 0 498 368 268 1,756 0 1995 700 0 700 0 252 1,652 0 1996 700 0 744 0 260 1,704 0 1997 600 0 552 0 207 1,359 0 Total 6,231 10,780 10,533 7,844 6,370 41,758 7,226 1. See text for details of calculation. The proportion of these research costs that had been expended on fodder bank research were estimated as follows (obtained from a consensus of personal communications from G. Tarawali, J. Smith, M.A. Mohamed-Saleem and R. von Kaufmann, ILRI): • 80% of the budget for sub hu mid Ni ge ria was spent on fod der bank re search for the years 1982 to 1990, while 20% was spent in 1981, 1991 and 1992; we as sumed that 10% of the re search costs for the years 1978 to 1980 were re lated to fod der bank ac tivi ties • none of the work in Ni ger dur ing these years was as so ci ated di rectly with fod der banks • a small pro por tion of the budget (5%) in Mali and hu mid Ni ge ria was spent on fod der bank re search dur ing these years • 1993 was the last year of any di rect ex pen di tures on fod der bank re search. These proportions are applied to the expenditures in Table 6 to give, in the last column, the research costs used in the analysis that are directly attributable to fodder bank activities, totalling some US$ 7.2 million. In the analyses presented in Chapter 6, research costs of ILRI’s partners and other institutions are not taken into account. Costs of establishing fodder banks Fodder banks can require substantial capital investment to establish, depending on how this is done, and the analysis needs to take account of these costs. For this analysis, all cost changes are embedded in the supply function. It is assumed that the technical change of fodder banks is neutral, in the sense that it does not affect the way that inputs are mixed in production and that technical change does not affect input prices. The costs of the technical change, however, include the research costs, to which we add the costs of establishing fodder banks. From the work done on adoption rates and patterns (Chapter 5), it is clear that establishment costs vary widely throughout the region, depending on the materials used for fencing off the fodder bank and the source of labour used. Otsyina et al (1987) estimated the capital costs of establishment of a 4-ha fodder bank with metal posts and strainers in Nigeria to be in the region of US$ 145/ha, with an annual recurrent maintenance cost of about US$ 21/ha, in 1987 prices. G. Tarawali (ILRI, personal communication) reported some establishment costs much higher than these levels, situations where fodder banks have been established with minimal input costs using local fencing materials and household labour, and other situations where the legume is simply undersown with no fencing at all. In the absence of detailed country- and system-level data, a single establishment cost of US$ 150/ha is used, which is taken to represent a weighted average of establishment costs in the region, with a small amount of maintenance being charged (US$ 20/ha) in subsequent years. Commodity prices, exchange rates, and national production figures Some costs and prices for the commodities of interest (maize, sorghum, millet grain and residue, milk and meat) were collected by country for the years 1977 to 1997. In some countries these were broken down by agro-ecological zone. These commodity prices were converted to US dollar equivalents using exchange rates extracted from appropriate issues of the United Nations Statistics Department’s Monthly Bulletin of Statistics. Many countries in West Africa have gone through substantial economic upheaval over the last 20 years, with occasional wild fluctuations in product prices in both local currency and, once converted using local exchange rates, dollar equivalents. To complement the prices obtained so far, we are still searching for prices for some countries for some commodities for the years in question. Where we have adequate price information for a country, we use this in the analyses; where price information is incomplete, we use prices collected for Ghana for 1977 to 1997 for all commodities (including crop residues), expressed in US dollars (see Appendix 2). As noted above, prices for commodities were collected for countries where this information is available. Estimated costs of cereal residues are rare because, until recently, these feed resources were free in most countries. Unlike cereal residues, grain legume residues fetch higher prices than their grains in semi-arid areas, especially in the peak dry season. For instance, in Kano, farmers are adopting cowpea for dry season cultivation in irrigated areas. Fodder is sold after grain harvest in situ, i.e. animals enter the field upon payment of about N 500 (about US$ 6 in late 1997) per 0.5 ha. Uptake of this practice has increased from 1 farmer in 1993 to 1500 in 1997. Time series data of national production of maize, sorghum, millet and cattle meat and milk were downloaded from the FAOSTAT database and supplemented with the appropriate issues of the FAO Production Yearbook. All these data are shown in Appendix 3 in terms of total production, hectares planted, animal numbers and average production per hectare or per animal for the 17 countries of West Africa for the years 1977 to 1997. 5 Adoption rates and patterns Adoption of fodder bank and legume technology has been addressed for various regions in particular countries by a variety of authors such as Tarawali and von Kaufmann (1987) and Ajileye et al (1994). According to Mohamed-Saleem and von Kaufmann (1995), the total number of fodder banks in the region in 1990 was 530. Between 1987 and 1991, a total of 637 fodder banks were established under farmer-managed supervised loans of the World Bank Second Livestock Development Project (Ajileye et al 1994). What was missing was a comprehensive effort to gather data from extension and government sources for all the countries of West Africa in which significant plantings of legume forages have taken place. This work was carried out by G. Tarawali from March to August 1997, and this chapter summarises findings by country. A map showing the countries of the region appears as Figure 2 Figure 2. Countries and climatic zones of West Africa. Benin In Benin farmers are adopting Stylosanthes spp, Centrosema, Pueraria and Mucuna for soil improvement, fodder and as cover crops. Other species such as Aeschynomene histrix have M a u r it a n i a M a li N i g e r N i g e r i a C h a d C a m e r o o n S e n e g a l G a m b i a G u i n e a B i s s e a u G u i n e a S i e r r a L e o n e I vo r y C o a s t G h a n a B u r k i n a F a s o S N EW A r i d S e m i - a r i d S u b - hu m i d H u m i d H i gh l a n d R o a d s R i v e r s been tested on station and are now ready to be taken to farmers for adoption through the extension service. In the north of Benin, fodder banks varying between 0.5 and 2.0 ha are used mainly for feeding traction animals and those that are too weak or sick to graze. In the south, farmers are given Stylosanthes seed to plant at 100–200/m 2 . They are expected to multiply these seeds and expand their fodder banks with time. This scheme, which started in 1988, has attracted about 10,000 farmers and covers 1500 ha of land. With this type of system, no credit facilities were required to promote the technology since the capital requirement is very low. Fencing is done using local materials and family labour is provided for this operation. Mucuna, one of the herbaceous legumes promoted by the non- governmental organisation (NGO) Sasakawa Global 2000 in the West African region for soil improvement and controlling noxious weeds such as Imperata cylindrica, raised the interest of 10,000 farmers (1000 ha) between 1988 and 1996. The legume is usually planted as a short fallow crop in exhausted farmlands and in areas infested by Imperata . After a season's growth, the Mucuna suppresses the weed and enriches the soil with nitrogen and organic matter, thereby making the land suitable for crop production. The foliage and pods are judiciously used as livestock feed. These 1000 ha of Mucuna are excluded from the analysis, as there was no direct ILRI involvement in this activity. Burkina Faso In the subhumid zone 30 to 900 fodder banks (45–1350 ha) were established between 1985 and 1989. Data on the evolution of fodder banks from 1985 to 1989 were not available, nor was there information from 1989 onwards. The number of adopters should be greater than 1000 by 1997, because ORSTOM (Office de la recherche scientifique et technique Outre-Mer) and other institutions such as CIRDES/EMVT (Centre international de recherche–développement sur l’élevage en zone subhumide/Elevage et médecine vétérinaire des pays tropicaux) are currently promoting the use of herbaceous legumes as improved fallows and livestock feed. The forages used are mostly Stylosanthes, Desmodium, Centrosema, Panicum and Brachiaria. They are usually incorporated in the farming system either as sole crops or as intercrops for fodder and reactivating degraded land; some have also been tested in upland and lowland rice-based systems. Cameroon In Cameroon, the highest adoption of improved forages in the farming system is in the humid zone where HPI (Heifer Project International) is promoting this technology to farmers for dairying, soil improvement and seed production. There was an increase in adoption from 10 farmers on 5 ha in 1986 to over 2000 farmers covering 420 ha of land in 1997. Forage adoption has not been as successful in the northern subhumid and semi-arid zones of Cameroon, where fodder bank adopters increased from 1 in 1979 to about 12 in 1989. The huge difference in adoption betwee n the ecoregional zones is attributable to an NGO that is promoting the adoption of forages in the humid zone; in the subhumid and semi- arid zones this role is played by national agricultural research systems (NARS) which have limited resources. In the humid zone of Cameroon, herbaceous legumes such as Desmodium, Lablab and Stylosanthes can take 2 to 3 years before re-establishment. Chad In Chad, pasture development programmes with farmers started in the 1980s with funding from the World Bank and the French government. On-farm/on-station tests of grasses and legumes consisting mainly of Panicum maximum, Cenchrus ciliaris, Chloris gayana, Macroptilum atropurpureum, M. lathyroides, Clitoria ternata, Cajanus cajan, Leucaena leucocephala, Lablab purpureus, Stylosanthes hamata etc were conducted between 1987 and 1990 by the Laboratoire de recherches vétérinaire et zootechnique (LRVZ). The results obtained in 1987 prompted the establishment of an extension programme in 1988. This activity promoted mainly Lablab purpureus, which proved to be the most adaptable amongst the candidates tested. Eleven plots varying in size from 0.5 to 1 ha were established and this number increased to 54 in 1990. The herbage was used for dry season supplementation of cattle especially in the months of March, April and May. Promotion of Lablab slowed down in 1991 because of poor follow-up by a weak extension service. This resulted in the whole extension system in Chad being restructured between 1992 and 1993, leading to the emergence of a new body called Direction de l’organization pastorale (DOP). With this new image, promotion of forages such as Lablab, Mucuna, cowpea etc was reactivated in 1994. About 900 active associations and groups have been formed with the assistance of DOP, whose primary role is to provide facilities (including forage production) to livestock keepers. In addition to introducing improved materials, DOP is also encouraging agropastoralists to establish and protect natural pastures especially in drier areas in a programme called ‘Perimetre pastoral pilote’ (PPP). The concept here is that huge areas of rangelands are demarcated with local trees such as Acacia and the vegetation within this perimeter is divided into paddocks to facilitate controlled rotational grazing. These communal paddocks are exploited by both sedentary and transhumant cattle thereby helping to solve the problem of competition for feed resources that used to exist between the two groups. In the late dry season, when the pasture is nearly exhausted, the transhumants move their animals to wetter areas while the remaining herbage is adequate to see the neighbouring sedentary herds through to the rainy season. Four sites have been developed in the Massaguet, Darqi, N'gora and Dobali regions with 5600, 6500, 6200 and 6000 ha, respectively. This strategy could be improved by including adaptable exotic legumes in the mixtures. ILCA/ILRI contributed some of the materials tested in Chad. There were also many ILCA/ILRI publications in the LRVZ library, including a copy of the 1993 research protocols, signifying that there was some positive interaction between ILCA/ILRI and the NARS in Chad. One of the consultants (Dr Lassine Diarra) employed by the World Bank to assist in organising the pasture development programme in Chad is an ex-ILCA/Mali staff member. The establishment of fodder banks recommended to the GTZ (Deutsche Gessellschaft für Technische Zusammenarbeit) project in Abeche by another World Bank consultant, Dr Boubakarr Hassane (former ILCA research student), is yet to start. Collaborators from LRVZ and DOP in Chad made a request for more germplasm and support for the forage seed multiplication unit which is currently addressing the seed demands of farmers and NGOs (mainly Lablab, Cajanus cajan, Stylosanthes, Andropogon etc). Training in all aspects of pasture development is also one of their needs. Côte d’Ivoire Fodder banks consisting of Stylosanthes grown in association with other cover crops such as Andropogon gayanus, Macroptilium atropurpureum and Panicum maximum were introduced to agropastoralists at several sites in north and central Côte d’Ivoire starting in 1985, when an estimated 42 adopters participated in the programme. The active period of this work in the north was from 1985 to 1989 but adoption came to a standstill from 1990 following the closure of SODEPRA (Société de développement des productions animales), Côte d’Ivoire’s national extension agency. A GTZ-sponsored ecofarm project which operated mainly in central Côte d’Ivoire increased the number of adopters by 8 in 1989. Fodder banks in Côte d’Ivoire were reactivated after the formation of ANADER (Agence nationale d’appui au développement rural) in 1995, the extension agency which replaced SODEPRA. ANADER is now working closely with GTZ to establish 56 Stylosanthes-based fodder banks (10 ha each) by 1998. Observations made during field visits showed that a good stand of Stylosanthes in the ecofarm fodder banks succeeded in suppressing the persistent weed Imperata in the same way Mucuna does in Benin. The total cost of an ecofarm that includes 10 ha of Stylosanthes-based pasture, 10 crossbred cows (plus a shed), fencing made up of Gmelina and barbed wire, and a house for the farmer, is about 10 million CFA (about US$ 18,000 in late 1997). In this system, the farmer provides the land and all the facilities are provided on a loan recovery basis. There is no direct integration of cropping into the fodder bank system amongst most of the farmers but the concept is being developed in a feasibility study which recommends the introduction of a cereal such as maize into a well-managed Stylosanthes guianensis fallow after only one year of establishment of the legume. The Gambia The visit was targetted at the International Trypanotolerance Centre (ITC), Development Livestock Services (DLS) and the National Agricultural Research Institute (NARI). DLS and ITC have been establishing feed gardens consisting of Pennisetum purpureum , Leucaena and Gliricidia at both on-station and village levels since 1993. These materials are trimmed, dried and fed in the form of hay mainly to small ruminants. Currently, The Gambia is benefitting from a grant offered by the Islamic Development Bank of Saudi Arabia to establish 1000 ha of improved pastures in 15 villages. Hopefully, about 1000 farmers will benefit from this generosity. The fodder banks in this scheme consist of Andropogon , Cenchrus ciliaris, Stylosanthes hamata , Centrosema pascuorum, Chaemacrista rotundifolia etc sown in strips. NARI is also evaluating herbaceous legumes on station for adaptation with the hope of introducing promising candidates to farmers in rice- based systems. ITC is keen on pasture development and has even recruited a forage agronomist. Plans for the immediate future include the development of year-round feeding strategies for crossbred cows and small ruminants. The centre is soliciting inter-institutional collaboration with ILRI/IITA on feed resources. The Director General of ITC suggested developing a joint proposal for funding. Ghana Forages were introduced in the humid and subhumid zones of Ghana as early as 1977, largely on institutional farms. In 1977, 10 organisations (250 ha) were practising the concept in the humid zone and 22 (660 ha) in the subhumid zone (these early adopters were omitted from subsequent analysis as ILCA’s work had barely started then). There was a general increase in the adoption rate in all zones by 1997. In order to promote their activities, the Animal Production Department in collaboration with the World Bank imported about 2000 kg of Stylosanthes seeds from Australia in 1993, and this germplasm has been progressively distributed and multiplied, mainly to peri-urban dairy farmers. In view of the usual recommendation to sow 10 kg/ha (the general recommendation for good- quality seed), this quantity of seed could cover 200 ha of land. Assuming that the World Bank is adopting the 4–ha size fodder banks as in Nigeria, this could lead to another 50 potential adopters in Ghana. One of the activities of the Animal Production Department is to improve the natural range by introducing Stylosanthes; in Tamale (northern Ghana) such a system has succeeded in introducing the legume on a 5-km radius. The cost of establishing a fodder bank in Ghana, excluding fencing, is in the range of US$ 200 to 300. In the humid zone, the forages are undersown in tree–crop plantations. On arable farms, the legumes (mainly Centrosema, Lablab and Stylosanthes) are grown in rotation with maize and need to be resown every 2 years. The forages are fed to dairy cattle, traction animals and small ruminants. Some seed production activities are also practised. In the slightly drier subhumid zone, all three cereals (maize, sorghum and millet) are grown in rotation with a wide range of legumes (Stylosanthes, Lablab , Cajanus cajan , Centrosema and Macroptilium atropurpureum ). Annual and biennial legumes are replanted every 2 years while the perennial legumes are renewed after 3 years. The fodder banks and crop residues are grazed by cattle and small ruminants. Guinea Fodder banks were promoted in the Boke region in the subhumid zone of Guinea through RABAOC (Réseau de recherche sur l’alimentation du bétail en Afrique occidentale et centrale) in 1992, although in the mid-1980s some forage screening had been done on station using seeds from Bouaké (Côte d’Ivoire) and CSIRO (Commonwealth Scientific and Industrial Research Organisation, Australia). Adoption of fodder banks has been progressing well in the subhumid zone of Guinea since their introduction in 1992 from 1 (5 ha) in the first year to 82 in 1997 with a total area of 559 ha (the average size of each fodder bank is about 7 ha). These improved pastures consist mainly of Panicum maxiumum C1 and Stylosanthes guianensis CIAT 184. The P. maximum is usually planted as a sole crop (12 kg/ha) while the legume (4–7 kg/ha) is introduced into natural pastures without applying fertiliser. The strategy which was developed on station and recommended to all smallholder farmers is to improve the natural range by incorporating Stylosanthes or other leguminous species, and to set aside other areas of improved pasture (fodder banks) for the strategic feeding of sick animals, suckling calves etc. The work at the Boke Centre is also supported by a large seed multiplication exercise involving a wide range of forages (Panicum , Brachiaria , Stylosanthes and Centrosema). A total of 3 ha is under seed production with an output of 850 kg per year. The average cost of establishing fodder banks in Guinea using local fencing materials is between US$ 300 and 400/ha, most of which are labour costs. Introduction of a crop phase in rotation with the Stylosanthes-based pasture has not been fully developed, probably because of a lack of awareness of the high yield increases that could be derived from such a system. Guinea Bissau In Guinea Bissau, three different bodies were found to promote the use of forages in agropastoral systems. The Holland–Bissau joint project ‘Projet agro-silvo-pastoral’ (PASP) has been introducing plots of Leucaena to villagers since 1996 in addition to testing species such as Stylosanthes, Cajanus cajan, Lablab etc for adaptability and relative performance. The project started with 10 farmers in 1996, each acquiring a 50 m × 50 m plot; this number increased to 30 in 1997. Before these activities started, some accessions and guidance were received from scientists in a Dutch-sponsored project in Sikasso (Mali) which had interacted extensively with ILCA/ILRI over the years. PASP plans to introduce Andropogon /Stylosanthes mixtures in forest areas in order to improve the quality of the natural pasture. In intensively cultivated areas, PASP hopes to enhance the adoption of Lablab/maize intercrop. Scientists in the project are looking for a small number of best-bet forage material (with background information) that could be transferred directly to farmers’ fields. Fa Madinga, a Belgian-sponsored project situated near Bafata, was boosting small ruminant production by encouraging small-scale farmers to establish fodder banks of Andropogon, Hyparrhenia, Leucaena and Stylosanthes. These activities started in 1992 and by 1995 a total of 30 plots, each about 0.25 ha, was recorded. The Institut national de recherche agricole (INRA) also introduced fodder banks of Andropogon for small ruminant production amongst farmers in 1994 and the number rose to 12 by 1996 but these activities ceased because of withdrawal of funds. However, farmers continued to keep their fodder banks and even supplied materials to others. INRA also multiplies seeds of Stylosanthes, Cajanus cajan , Sesbania etc. A total of about 1 ha is set aside for seed production and this germplasm is usually distributed to NGOs and farmers. Colleagues in Guinea Bissau requested more germplasm for diversification and in- country training on pasture development and crop–livestock interactions. Mali Adoption of herbaceous legumes in the cropping systems in Mali started in 1977 with 10 (2 ha) farmers through an FAO/CMDT (Food and Agriculture Organization of the United Nations/ Compagnie malienne pour le développement des textiles) project; the species promoted then were cowpea (Vigna unguiculata ), Dolichos and Siratro. The number of participating farmers rose to 951 (381 ha) in 1987, then to 1317 (648 ha) by 1991 and to 1421 (700 ha) in 1996. During 1989 and 1991, CMDT promoted the use of leguminous crops such as Stylosanthes, Lablab and Panicum amongst 280 farmers. The cost of establishing a fodder bank in Mali depends on the species. Stylosanthes-based pastures cost about CFA 50,000 (about US$ 100)/ha while Lablab pastures cost about CFA 25,000 (about US$ 50)/ha. Local materials are commonly used for fencing. Farmers in Mali also try to avoid the need to fence fodder banks by growing Stylosanthes between crop fields. This may have encouraged an increase in uptake since fencing has been a serious deterrent to adoption because of the high cost and labour demand. After 2 or 3 years of a leguminous phase, the area is cropped with a cereal such as maize, sorghum or millet. The forage is commonly used to feed traction and dairy animals. Mauritania In Mauritania, the concept of exploiting cultivated pastures for livestock production started in 1989 according to records and responses from key collaborators during intensive interviews. Traditionally, the local herds of cattle are fed from rainfall-dependent natural pastures consisting mainly of Cenchrus. However, over the years, the diminishing amount of rainfall (200 mm per annum) in this Sahelian country has caused cattle owners to cultivate irrigated pastures along the River Senegal in the southern part of the country. This effort is being promoted by both small- and large-scale farmers through the formation of co- operatives. To date, there are 39 livestock co-operatives in Mauritania and each consists of about 100 participants, meaning that currently 3900 people are benefiting from pasture development programmes. Each beneficiary is expected to exploit about 0.25 ha for his or her herd. Thus a total of 975 ha is being exploited by co-operatives. The government of Sweden provided a grant of US$ 5 billion to support 108 co-operatives by 1998, but this target is far from being reached because of some implementation problems. It is the inverse relationship between the growing animal numbers and the diminishing natural pasture that has forced co-operatives and research and development agencies to exploit cultivated pastures usually grown along the River Senegal. For instance, the National Research Centre of Agronomy and Agricultural Development (CNRADA) based in Kalde (south of Mauritania, near the Senegal river) has been testing forages such as Panicum maximum , Pennisetum purpureum , Stylosanthes guianensis, Macroptylum lathyroides, cowpea, sorghum, Lablab purpureus, Clitoria ternata, Cajanus cajan etc on station since 1990. Extension of the technology to farmers started in 1994 with about 5 participants and this number grew to about 30 in 1997. Each farmer was encouraged to grow about 0.25 ha of forages consisting mainly of a cowpea/sorghum intercrop and Lablab purpureus for feeding their animals. Because of a low animal population in Mauritania, most of the livestock needs in the country are being supplemented by larger herds from Mali. Similarly, some of the research staff in animal science and other areas were trained in Mali, where ILCA/ILRI had a country programme. Such interactions with ILCA/ILRI may have had a positive effect on the feed resources programme in Mauritania. Even some of the germplasm was supplied by ILCA/ILRI. Niger Traditionally, cowpea is the main fodder crop in Niger, the world's third highest producer of this crop. Cowpea is grown primarily for grain (human consumption) and the haulms (animal feed). However, in areas south of Niamey where there is a higher density of livestock because of the River Niger, the legume is grown mainly for fodder and it is sometimes intercropped with millet. Since 1990, INRAN (Institut national de recherche agronomique du Niger), the national research institute, has developed an efficient structure for distributing cowpea germplasm directly to farmers and development organisations. A total of 5600 kg of cowpea seed is being distributed each year. It is expected that every end user benefits from 1–3 kg (average 2 kg) of the grain. This means that a total of 2800 farmers benefit each year and since the recommended seed rate is 15 kg/ha, an area of 373 ha is cultivated annually. Cowpea hay is usually used as supplementary feed to range grazing, an area that is excluded from the analysis as ILRI had no or very limited direct input to this activity. Other local species such as Commelina, Cenchrus, Alysicarpus and Zornia are commonly harvested from the range and stored or sold for feeding, especially traction animals and donkeys. A reasonable number of projects in Niger are working with farmers to introduce pastures and trees in degraded soils. One such group is DED (Deutsche Entwicklung Dienst) a German-sponsored project which has been adopting participatory approaches since 1990. Basically, the group encourages farmers to plant traditional and adaptable trees such as Acacia and Aristida and local grasses such as Pennisetum pedicellatum and Cenchrus biflorus for feeding animals, and for soil conservation and fuel. The grasses are either grazed directly or cut and carried as hay. Farmers weed and protect more valuable species such as Alysicarpus in the same way Stylosanthes is nurtured in fodder banks. No fencing is required by the farmers and labour to maintain trees or pastures is provided free of charge by the farmers themselves. The farmers frequently collect seeds to expand to new areas and resow in patches within the existing fields. Since 1990, a total of 62 sites covering an area of over 15,000 ha has been developed in Niger (this area is omitted from the analysis, as these sites involve mainly managed natural pasture). Other projects working in similar areas are Project Keita (an FAO project sponsored by Italy), the GTZ-sponsored project ‘Projet protection intégrée des ressources agrosilvopastorales’ (PASP), and the World Bank–Niger-sponsored PRSAA (Programme de renforcement des services d’appui à l’agriculture), but these development agencies have only recently introduced livestock components into their activities. Nigeria In Nigeria fodder banks were being exploited for livestock and crop production in all the agro-ecological zones since around 1982. By 1993 adoption was higher in the subhumid zone (589 smallholder farmers on 2467 ha) and semi-arid zone (3539 farmers on 1220 ha) than in the humid zone (38 adopters on 228 ha). Nationally, by 1997 a total of 4166 adopters and an area of 3915 ha were covered in Nigeria. Until 1992, the promotion of forages (fodder banks) in Nigeria was mainly under the supervision and sponsorship of the National Livestock Projects Division (NLPD) and the World Bank. Under Nigeria’s unified extension system, this responsibility was transferred to the Agricultural Development Project (ADP), which was basically an extension mechanism for crops but currently incorporating a livestock component. This change-over of activities probably accounted for the sharp decline in the rate of adoption between 1992 and 1997, but it is anticipated that after this slack period, adoption rates will escalate again. The semi-arid zone has the highest number of adopters, while the subhumid zone has the larger area. This is attributed to the fact that the promotion of the 4 ha fodder banks using mainly Stylosanthes in the subhumid zone is for cattle production, while in the semi-arid zone, Lablab (one of the most adaptable species in the zone) is established in smaller plots of less than 1 ha within the farm for both soil improvement and livestock feed. The species used in Nigeria in the fodder bank system include Stylosanthes, Lablab, Centrosema and Panicum . The method of establishment ranges from sole cropping in the natural range to intercropping and undersowing. The cost of fodder bank establishment in selected states in Nigeria using wooden posts was about N 14,000 (some US$ 600) in 1994. In a typical 4-ha fodder bank in Nigeria, re-establishment of the legume is done using a 2- to 3-year crop–forage rotation to avoid invasion of Stylosanthes by fast growing nitrophilous grasses. Integrating cropping into the fodder bank system usually leads to a legume- dominant pasture and serves as a cheap way of renovating leguminous pastures for both labour and inputs. Land preparation and weeding usually meant for growing crops could benefit the legume; likewise the phosphorus commonly applied to crops. Being self-seeding, no additional labour is required for resowing nor is any further purchase of Stylosanthes seed necessary. Senegal To alleviate the dry season feed constraint on livestock, CIRAD (Centre de coopération internationale en recherche agronomique pour le développement), ORSTOM and various NGOs have been advising agropastoralists to grow mainly Andropogon and Panicum (0.25 ha plots) and a limited quantity of Stylosanthes for late dry season supplementation. In addition, the pastoralists are also encouraged to sow large areas of the forest with Andropogon/Panicum to improve the natural range. This concept is similar to that used in Guinea. The number of fodder banks has increased to 31 (14.8 ha), and the effort to promote the concept is continuing. Leguminous crops such as Stylosanthes are not widely used in these agropastoral systems simply because seeds are not available. The rate of adoption is also limited by lack of Panicum seeds and Andropogon cuttings. Seeds or cuttings are supplied free and no fertiliser is applied. Fencing is done entirely using local materials (bamboo) and almost all the costs of establishment are taken up by labour. Sierra Leone Baseline surveys are currently being conducted in Sierra Leone to understand the livestock production systems with a view to promoting research, which has remained dormant for several decades. In the government’s new agenda, forage research and development are a priority, but testing and extending forage interventions to farmers lie mainly in plans for the future in this country. Togo Exploitation of legume-based technologies started in the humid zone of Togo in 1983 with 23 adopters (40 ha), and the concept diffused to the subhumid zone from 1984 onwards. At the last count in 1992, a total of 80 farmers (10.6 ha) had incorporated forage legumes into their cropping systems in the subhumid zone, while adoption in the humid zone had stagnated at 28. The species exploited in the humid zone are mainly Leucaena, Gliricidia and Cajanus cajan . These shrubs are grown in association with maize and part of the foliage is incorporated into the soil to improve soil fertility; the other portion is used to feed small ruminants. The stems provide firewood for the farmers. This technology could be related to alley farming. In the northern subhumid zone, Stylosanthes has been successfully introduced amongst the smallholder farmers to alleviate nutritional stress in traction animals, especially in the dry season, and to improve soil fertility. The legume is re-established every 2 years in Togo. Records on the further adoption of forages beyond 1992 could not be accessed, but the adoption trend is likely to continue or increase in view of the growing awareness of the exploitation of this concept in the region. Summary of the adoption data In summary, the forage legume technology promoted originally by ILCA is diffusing slowly in the whole of West Africa, including in countries such as Senegal and Guinea, where ILCA has had no direct influence (country adoption data are shown in cumulative form in Appendix 4). The national and international organisations mentioned in this chapter, other than ILCA, ILRI and national agricultural research and extension systems, as being involved with the development and dissemination of forage legume technology in the region (the list is clearly not exhaustive, but only indicative) are listed in Table 7. The data suggest that the fodder bank concept is increasingly being accepted as one option for solving the problems of livestock nutrition and soil conservation. The highest number of forages were in Nigeria (4166), Cameroon (2200) and Mali (1421), whereas the largest areas were recorded for Nigeria (3915 ha) and Burkina Faso (1380 ha). The large number of adopters but small area in Cameroon reflects that only small areas of land were planted with the legume partly due to land shortage and intensive cultivation, typical of the humid zone environment. In Nigeria, adopters exploit larger areas of land mainly for forage production, especially in the subhumid and semi-arid zones where land availability does not seem to be a serious problem. Adoption was lowest in The Gambia, Senegal and Guinea, largely because forage production and adoption have only been recently introduced in these countries. For the whole of West Africa, a total of about 27,000 smallholder farmers covering an area of about 19,000 ha were recorded (Figure 3). Most of the adoption is concentrated in the subhumid zone and uptake of forage technology in all three agro-ecological zones was only readily apparent in Nigeria and, to a lesser extent, in Cameroon. In the impact analysis a number of hectares were omitted (described above), notably in Benin, Ghana and Niger, where ILRI was clearly not involved in the uptake of these forages. These hectares were also omitted from the analysis because the research costs and benefits associated with work carried out by ILRI’s partners and other institutions are not taken into account. The area used in the impact analysis presented was thus just under 16,500 ha. Figure 3. Adoption patterns for fodder bank and legume forage technology in countries of West Africa. 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 28,00 0 24,00 0 20,00 0 16,00 0 12,00 0 8,000 4,000 0 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 15,00 0 12,00 0 3,000 0 Total Humid Subhumid Semi-ari d 9,00 0 6,00 0 18,000 Adopter s Area (ha) Year Table 7. National and international organisations and projects mentioned in the text, excepting national agr icultural research and extension systems, that have been involved in development and delivery of forage legume technol ogy in West Africa. Organisation or project Country Centre de coopération internationale en recherche agronomique pour le développement (CIRAD) Senegal Centre international de recherche–développement sur l’élevage en zone subhumide (CIRDES) Burkina Faso Commonwealth Scientific and Industrial Research Organisation (CSIRO) Guinea Deutsche Entwicklung Dienst (DED) Niger Deutsche Gessellschaft für Technische Zusammenarbeit (GTZ) Chad, Côte d’Ivoire, Niger Development Livestock Services (DLS) The Gambia Direction de l’Organization Pastorale (DOP) Chad Elevage et médecine vétérinaire des pays tropicaux (EMVT) Burkina Faso Fa Madinga Guinea Bissau Food and Agriculture Organization of the United Nations (FAO) Mali Heifer Project International (HPI) Cameroon International Centre for Tropical Agriculture (CIAT) Guinea International Trypanotolerance Centre (ITC) The Gambia Islamic Development Bank The Gambia Laboratoire de recherches vétérinaire et zootechnique (LRVZ) Chad Office de la recherche scientifique et technique Outre-Mer (ORSTOM) Burkina Faso, Senegal Programme de renforcement des services d’appui à l’agriculture (PRSAA) Niger Projet agro-silvo-pastoral (PASP) Guinea Bissau Project Keita (FAO, Italy) Niger Projet protection intégrée des ressources agrosilvopastorales (PASP) Niger Réseau de recherche sur l’alimentation du bétail en Afrique occidentale et centrale (RABAOC) Guinea Sasakawa Global 2000 Benin World Bank Chad, Ghana, Niger, Nigeria 6 Application of the economic surplus model to estimate impact The previous chapters have specified a complete set of data with which to run an economic surplus model. With this, we can make an estimate of the producer and consumer surplus arising from the adoption of fodder bank technology in West Africa, and we can also estimate the returns to the preceding research investment. There are four steps to the process: Step 1. For each of the eight commodities (milk, meat, maize grain, millet grain, sorghum grain, maize residue, millet residue and sorghum residue), we calculate the productivity impacts on national production in each year of the analysis from 1977 to 1997 arising from adoption of the appropriate number of hectares of fodder bank. It is assumed that once a fodder bank is established it is not abandoned and hence does not revert to natural pasture. Step 2 . For each commodity, given the percentage productivity impact on national production as a result of fodder bank adoption as calculated in step 1, we calculate the changes in producer, consumer and total surplus in each year, using the equations in Chapter 3 and the elasticities in Table 5. Step 3. We then sum the surpluses for each commodity to give the total annual surpluses arising from adoption of fodder bank technology for the countries considered. Step 4. Finally, we take the total surplus for each year, subtract from it the appropriate research costs (as in Table 6) and establishment costs, and discount the net benefit stream, if appropriate, then calculate the net present value of that stream and the internal rate of return of the ‘investment’ in fodder bank research and development. Baseline analysis For the standard baseline analysis, the productivity impacts are rather small, on a regional basis (as might be expected). Thus for the 15 countries used in the analysis, for maize grain, an estimated 1400 t of the transnational maize grain production in 1997 is attributable to fodder banks, or 0.012% of regional production. Similarly for milk, an additional 2250 t is attributable to fodder banks, or 0.14% of regional production, as is an additional 1300 t of meat, or 0.18% of regional production. Time series data for producer, consumer and total surpluses for maize, meat and milk are shown in Figure 4. Note that all of the producer surpluses, except for meat, are negative, i.e. consumers are reaping most of the benefits that accrue to society through cheaper commodities. Meat is the exception, where both consumer and producer surpluses are positive; this follows directly from the magnitude of the elasticity of demand for meat shown in Table 5 (also see Appendix 1). The net benefit stream from 1977 to 1997 is shown in Figure 5. For this stream of benefits, the internal rate of return was calculated to be 38%, with total net benefits of about US$ 16.5 million arising from total research costs of US$ 7.2 million. Negative producer surpluses at the societal level contrast somewhat with the fact that fodder banks are profitable at the individual farm level, even after the small commodity price reductions modelled in the analysis (see Figure 1). The negative producer surpluses are a direct consequence of the form of the supply and demand functions used (the constant elasticity specification, see Chapter 3). Other than using a different specification of the essentially unknown functional form, it may be noted that the economic surplus is a static model, and that in the adoption process, early adopters may reap more benefits than later adopters before the new supply-and-demand equilibrium is actually reached. Figure 4 . Baseline fodder bank impact analysis: Consumer surplus (CS), producer surplus (PS) and total surplu s (TS) for three commodities in countries of West Africa. 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 3.0 2.5 2.0 1.5 1.0 0.5 0.0 –0.5 Milk C S T S P S Surplus (US$ × 1 0 6 ) –0.1 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 3.0 2.5 2.0 1.5 1.0 0.5 0.0 –0.5 Meat T S C S P S 1997 Surplus (US$ × 1 0 6 ) –0.1 19971977 1979 1981 1983 1985 1987 1989 1991 1993 1995 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Maiz e C S TS PS Year Surplus (US$ × 1 0 6 ) –0.5 –0.1 Figure 5 . Baseline fodder bank impact analysis results. Sensitivity analysis A number of sensitivity analyses to investigate the robustness of this estimate of impact and investment returns were carried out. In particular, we looked at what effect different numbers of adopters would have on the analysis. We assumed that the number of hectares reported as being under fodder banks was (i) overestimated by 20% and then (ii) underestimated by 20% (Table 8). A 20% increase in the number of adopted hectares increases the internal rate of return (IRR) by 15% and the total net benefits by 29%; reducing the number of adopters by 20% decreases the IRR by 12% and the total net benefits by 29%. While the results of the analysis are clearly sensitive to the extent of adoption, a wide range of adoption leads to the conclusion that investment in fodder bank research has been profitable to date and in view of future adoption, will become even more so. The 20% changes made above were also applied to the productivity estimates, and resulted in very similar figures and the same conclusions (because of the small percentage changes in regional productivity, the calculated impacts are close to linear, whether applied to the number of hectares of adoption or to the productivity changes arising as a result of adoption). Of course, it is the case that all the data used in the analysis have uncertainty attached to them, but adoption numbers and productivity impacts are particularly uncertain. A sensitivity analysis was also carried out with respect to the elasticities of demand, for which information is certainly lacking. Increasing the elasticities of demand for milk, maize, millet and sorghum to 2.0 resulted in negligible changes in net present value, from which we concluded that the results of the analysis are relatively insensitive to the values used for the elasticities of demand. 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 5 4 3 2 1 0 –1 Net benefits Research and establishment costs Total surplu s Year US$ ( × 10 6 ) Table 8 . Summary of baseline analysis and sensitivity analyses. Baseline analysis 20% increase in number of adopters (ha) 20% decrease in number of adopters (ha) Internal rate of return (%) 38 53 26 Total net benefits by 1997 (US$ × 106) 16.51 21.28 11.76 Consumer surplus, 1997 (US$ × 106) 5.34 6.41 4.27 Producer surplus, 1997 (US$ × 106) ?0.71 ?0.86 ?0.57 Total economic surplus, 1977 (US$ × 106) 4.62 5.55 3.7 Net present value in 1997 dollars (US$ × 106) 11.82 14.7 8.95 Benefit:cost ratio 3.3 3.6 2.93 Future adoption The aggregate adoption curves shown in Figure 3 mask a great deal of country-by-country variation (see Appendix 4). However, the adoption lag is considerable. ILRI had little to do with diffusion in Guinea, Cameroon and Mali, but adoption started in 1992 in Guinea and in the late 1970s in Mali (Figure 6). A question of some importance is, what is the likely equilibrium number of adopters of forage legumes in the region? There are no more research costs associated with the technology, so for future years the net benefit stream is equal to the gross benefit strea m (Figure 5), discounted at an appropriate rate. Figure 6 . Adoption of fodder banks in Mali, Cameroon and Guinea. There are at least two ways to approach this question, both somewhat crude but providing information on possible future adoption levels. One way is to consider the potential area of forage legumes in West Africa, and then take a ‘reasonable’ percentage (perhaps 10%) as adopters, in view of the difficulties faced in adopting forage legumes in many places (see Chapter 7), and then another ‘reasonable’ percentage (perhaps another 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 0 800 700 600 500 400 300 200 100 Adoption (ha) Year 10%) of land actually planted to legumes, given the fact that they are a component of a cropping system etc. Thus an upper limit might be 1% or less of viable land that is planted to forage legumes. Land areas highly suitable to Stylosanthes can be extracted and broken down by agro-ecological zone and by country (van Velthuizen and Fischer 1995). For West Africa as a whole, these authors define 5.3 million ha as being highly suitable for Verano Stylosanthes under low input conditions. One per cent of this is 53,000 ha, indicating that the 19,000 ha of adopted legumes reported to date amounts to a little over a third of what could be expected, assuming that 1% of highly suitable land is planted. A second approach is to study individual country adoption curves, to see if extrapolations can be made into the future to the point where all country adoption curves are flattening off, assuming that a standard ‘S’ shaped adoption curve is appropriate for every country. From the aggregate adoption curve of Figure 3, it is difficult to extrapolate. If 1997 is ignored, the numbe r of adopters appears to be in the middle of the S-shaped curve (the more-or-less linear part), indicating that if 1997 is near the point of inflexion, then about 40,000 adopters could be expected by 2010 or so. Assuming that the number of hectares planted per adopter remains constant, then this implies about 32,000 ha or so. This is probably not unreasonable; from the figures from Nigeria and Côte d’Ivoire, for example (Appendix 4), the rate of increase in the number of hectares adopted has slowed down substantially, while for countries such as Guinea, substantial further adoption could be expected, given that the lag time by country (time from first adoption to maximum number) can be 15 years or so. The impact of both these adoption scenarios (the resulting adoption curves are shown in Figure 7) was quantified. For simplicity, the rate of increase in production of all commodities was fixed at 1.5% per year for the period 1998 to 2014, and for these years 1997 commodity prices were used. The future portion of the net benefit stream was discounted at 10% to give a net present value in 1997 dollars. The analysis results from the high and low adoption projections (Figure 7) are summarised in Table 9. Even taking the low projection, the total returns to fodder bank research are greatly inc reased compared with Table 8, and result in substantial increases in the total net present value. Figure 7. Two adoption scenarios for forage legumes in West Africa to 2014. 197 5 198 5 199 5 200 5 201 5 60 50 40 30 20 10 0 High projectio n Low projection 1997 level Adoption ( '000 ha) Yea r Table 9. Summary of baseline analysis and adoption projections to 2014.1 Baseline analysis ‘Low’ adoption projection to 2014 ‘High’ adoption projection to 2014 Total net benefits (US$ × 10 6) 16 .5 138 188 Net present value of past and future benefits in 1997 dollars (US$ × 106, discount rate 10%) 11 .8 65 .3 81 Benefit:cost ratio 3.3 7.1 7.2 1. ‘Low’ adoption projection (Figure 7) finishes at 32,000 ha in 2014. ‘High’ adoption projection finishes at 53,000 ha in 2014. Analysis by agro-ecological zone Rather more refined impact analyses could be carried out on the basis of agro-ecological zones, in addition to the regional analyses performed above. The adoption data have been collected on this basis (Appendix 4), but considerably more effort would be needed on adjusting the productivity impacts to use in each zone (humid, subhumid and semi-arid). Local variations in commodity prices would also need to be taken into account, and these data are also hard to come by. These analyses could then be run by agro-ecological zone in the future, and the results mapped. 7 Discussion The analyses described indicate that the fodder bank technology, even though not taken up in great numbers by the agropastoralists of West Africa, has more than paid for the research resources expended on its development through increased meat, milk and cereal production. These analyses should certainly be viewed as being indicative rather than definitive, in a number of respects. In Chapter 4, it was noted that there are still substantial data shortages, particularly with respect to commodity prices and elasticities of supply and demand. Lack of country-level and production system-level information on costs of fodder bank establishment and methods of utilisation (and their relative preponderance), meant that broad-brush assumptions had to be made concerning regional-level establishment and utilisation methods. In addition, there is a sense in which the analysis performed is likely to result in fairly conservative estimates of the true (but unknown) benefits of fodder bank technology. We certainly have not identified all the adopters, and we have taken account of only some (although probably the most important) of the benefits that arise from fodder bank utilisation by large ruminants; the benefits of fodder bank utilisation by small ruminants have not been dealt with at all in this study. Many institutions and organisations have been involved with forage legumes in West Africa for many years, for a wide variety of purposes. The total expenditure of research and development resources on forage legumes, going back to the 1930s, may be large, and ILRI’s investment is clearly only a part of this total. However, in contrast to some other forage legume initiatives, the ILCA/ILRI forage legume ‘recipe’ was always clearly aimed at strategic supplementation of livestock during times of stress, with the soil and crop benefits being regarded as desirable bonuses for the farmer (M.A. Mohamed-Saleem, ILRI, personal communication). The adoption data collected during this study were screened to omit those data that could reasonably clearly be traced to the activities of other organisations and/or that involved obviously different methods of utilisation, but inevitably there will still be overlap. It is true, however, that the attribution problem (what proportion of fodder bank adoption can be directly attributed to the activities of ILRI) is extremely difficult to deal with adequately, and highlights both the importance of using conservative estimates of productivity impacts and the importance of performing sensitivity analyses on the results. Given the apparent benefits of planted forage legumes, even allowing for the weaknesses of the analyses conducted during this study, it may well be asked why adoption has not been even more widespread than it has. Various authors have identified a number of constraints to adoption in West Africa: • Lack of ex ten sion in for ma tion. The fact that farm ers are un aware of the fod der bank tech - nol ogy and pro grammes has been iden ti fied as an im por tant fac tor hin der ing adop tion of the tech nol ogy (Mohamed- Saleem and von Kauf mann 1995). • In ap pro pri ate land ten ure . This is an other im por tant fac tor hin der ing the adop tion of the tech nol ogy (M ohamed- Saleem and von Kauf mann 1995). Land fa cili tates ac cess to credit be cause it can be used as col lat eral for a loan. In ad di tion, se cure land rights are a pre - req ui site for any long- term in vest ment, and their ab sence will mili tate against adop tion of rela tively capital- intensive tech nolo gies such as fod der banks. • Fenc ing ma te ri als . Fenc ing can rep re sent nearly 80% of to tal in put costs for es tab lish ing a fod - der bank (Ot sy ina et al 1987). In some coun tries fenc ing ma te ri als are ex pen sive and the pro - cure ment and de liv ery of in puts for es tab lish ing a fod der bank are dif fi cult. The use of lo cal ma te ri als re quires ad di tional la bour, which some farm ers will have dif fi culty fit ting in to al - ready busy sched ules. • Short age of la bour. Dur ing the early rainy sea son la bour is re quired for crop pro duc tion, and there is a short age of ag ri cul tural mecha ni sa tion (in clud ing ani mal trac tion) in man y West Af ri can coun tries. Ad di tional la bour is re quired for in clud ing fod der banks in the farm ing sys tem and this is of ten not avail able, es pe cially for farm ers with young fami lies . In fact, Tho mas and Sum berg (1995) fol low Berry (1993) in iden ti fy ing lim ited la bour avail abil ity, in terms of to tal quan tity and sea sonal avail abil ity, as proba bly the sin gle most im por tant fac tor af fect ing the course of agrar ian change (in clud ing ag ri cul tural in - ten si fi ca tion) in sub- Saharan Af rica in gen eral. Analy ses of the adop tion of leg ume tech - nol ogy in terms of changes in la bour re quire ments are not well- developed to date. • Dis ease con straints. Kraal ing ani mals over night for sev eral nights is some times re jected by farm ers for fear of nema todes in the dung in fect ing cat tle (Mohamed- Saleem 1994). In some zones, Sty lo san thes is not par ticu larly re sis tant to dis eases and pests. An im por tant dis ease is an thrac nose and im por tant pests are ter mites, leaf- eating bee tles, leaf hop pers (eat ing flow ers) and blis ter bee tles (Adeoti et al 1994). • Credit and seed avail abil ity. Un like in Ni ge ria and Côte d’Ivoire, where credit fa cili ties have aided the dif fu sion of legume- based tech nolo gies, in coun tries such as Cam er oon, Sene gal, Guinea and Mali, such in cen tives have not been found. In ad di tion, high in ter - est rates with the many re stric tive con di tions of ten im posed by credit pro grammes pre - vent farm ers from ex ploit ing loan schemes. Lack of suf fi cient seed in some coun tries is also re garded as an im pedi ment to adop tion, and ef forts to en cour age lo cal seed pro duc - tion are needed (Ta raw ali et al 1998). • Land scar city. In in ten sively cul ti vated ar eas farm ers can not leave land fal low for even one year; farm ers thus find it dif fi cult to in clude planted leg umes in their crop ping sys tems un less they adopt other prac tices such as in ter crop ping or se quen tial crop ping. • In va sion by grasses and weeds. When es tab lished in as so cia tion with grasses such as Pa ni - cum maxi mum , the leg ume in the fod der bank, un less well man aged, will be sup pressed lead ing to a grass- dominant pas ture of lower nu tri tional value. In ad di tion, ag gres sive and nox ious weeds, no ta bly Im per ata cy lin drica and Sida acuta, some times in vade fod der banks lead ing to a com plete dis place ment of the de sired le gu mi nous spe cies. Such prob - lems have in the past led to many fod der banks be ing aban doned. • Fires. Burn ing, es pe cially dur ing the dry sea son, is a very com mon prac tice in West Af ri - can range lands. There are con cerns that fod der banks re served for sup ple men ta tion of cat tle in the late dry sea son face the risk of be ing wiped out by fire at the time they are needed most; this can con trib ute sub stan tially to the per ceived riski ness of in vest ing in a fod der bank. Given this array of problems, it is perhaps not surprising that the diffusion of fodder banks in the region has been slow and modest; what is surprising is the indication that the lag in the adoption of fodder bank technology in individual countries seems to be of the order of 15 to 20 years. This is a considerable length of time, and probably much longer than may have been expected. ILRI is no longer investing research resources directly into fodder bank technology research and development. In the absence of direct research effort, the question of whether there is a role for ILRI and other institutions and organisations to promote this technology in the coming years is certainly relevant, but it is not particularly easy to answer directly. As Thomas and Sumberg (1995) note, considerable resources have been devoted to screening and testing forage legumes throughout sub-Saharan Africa over the last 40 years; given the large picture of rather limited use of legumes by smallholders, they wonder about the justification for this expenditure. There may certainly be a role in information dissemination and training of national research partners and those working in extension services. There is also likely to be an important role in helping NARS to secure and multiply germplasm. ILRI still has substantial commitments to research on legumes in mixed farming systems through research programmes in the semi-arid and subhumid zones of West Africa, however. The role of legumes may also change dramatically in West African farming systems over the next 40 years. The agropastoral systems of West Africa will come under enormous pressure primarily from population growth, and the question of adequate forage resources for a rapidly expanding cattle population is very complex. Solutions are unlikely to come from any one source, but are more likely to involve greater integration of crop and livestock enterprises, particularly through increased use of crop residues. Within this milieu, planted forage legumes may occupy an expanding niche. Even if this niche never becomes very large, the research that ILCA carried out on fodder banks in the 1980s and early 1990s has already more than paid for itself, particularly in terms of benefits for consumers in West Africa. References Adeoti A., Tarawali S. and Mohamed-Saleem M.A. 1994. Development of management strategies for stylo with emphasis on pests and diseases. In: de Leeuw P.N., Mohamed-Saleem M.A. and Nyamu A.M. (eds), Stylosanthes as a Forage and Fallow Crop. Proceedings of the Regional Workshop on the Use of Stylosanthes in West Africa, held in Kaduna, Nigeria, 26–31 October 1992. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 123–128. Ajileye E.O., Uza D.V. and Farooqi M.A. 1994. An assessment of farmers’ adoption rate and potential impacts of stylo-based feed production systems. In: de Leeuw P.N., Mohamed-Saleem M.A. and Nyamu A.M. (eds), Stylosanthes as a Forage and Fallow Crop. Proceedings of the Regional Workshop on the Use of Stylosanthes in West Africa, held in Kaduna, Nigeria, 26–31 October 1992. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 311–316. Alston J., Norton G. and Pardey P. 1995. Science Under Scarcity: Principles and Practice for Agricultural Research Evaluation and Priority Setting. Cornell University Press, Ithaca, New York, USA. 585 pp. Bayer W. 1986. Utilization of fodder banks. In: von Kaufmann R., Chater S. and Blench R. (eds), Livestock System Research in Nigeria’s Subhumid Zone. Proceedings of the Second ILCA/NAPRI Symposium held in Kaduna, Nigeria, 29 October–2 November 1984 . ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 351–364. Berry S. 1993. No Condition is Permanent: The Social Dynamics of Agrarian Change in Sub-Saharan Africa. University of Wisconsin Press, Madison, Wisconsin, USA. 258 pp. Delgado C.L. and Reardon T. 1991. Cereal Consumption Shifts and Policy Changes in Developing Countries: General Trends and Case Studies from the West African Semi-Arid Tropics. IFPRI Reprint 255. IFPRI (International Food Policy Research Institute), Washington DC, USA. FAO (Food and Agriculture Organization of the United Nations). 1990–1998. FAOSTAT-PC. FAOSTAT Statistics Database. Accessed in June 1997. FAO, Rome, Italy. Hassane B. 1995. On-farm Evaluation of Fodder Bank Adoption among Agropastoralists in Northern Nigeria . Unpublished PhD dissertation, Department of Range Science, Utah State University, Logan, Utah, USA. 155 pp. Ikwuegbu O. Tarawali G. and Iji P. 1994. Sustaining a crop-livestock farming system in the subhumid zone of Nigeria by matching feed from improved Stylosanthes-based pastures and livestock production. In: de Leeuw P.N., Mohamed-Saleem M.A. and Nyamu A.M. (eds), Stylosanthes as a Forage and Fallow Crop. Proceedings of the Regional Workshop on the Use of Stylosanthes in West Africa, held in Kaduna, Nigeria, 26–31 October 1992 . ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 167–174. Ikwuegbu O.A. Tarawali G. and Rege J.E.O. 1995. Effects of fodder banks on growth and survival of West African Dwarf goats under village conditions in subhumid Nigeria. Small Ruminant Research 17:101–109. Jaeger W.K. 1986. Agricultural Mechanization: The Economics of Animal Draft Power in West Africa. Westview Press, Boulder, Colorado, USA. 199 pp. von Kaufmann R., McIntire J. and Itty P. 1990. ILCA Bio-Economic Herd Model For Microcomputer. User’s Manual. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. Mani R. 1992. Optimal Utilization of Legume Fodder Banks and Low Quality Roughages by Grazing Cattle. PhD thesis, University of Melbourne, Parkville, Australia. 192 pp. Mani R.I., von Kaufmann R.R., Egan A.R., Oyedipe E.O., Dixon R.M. and Holmes J.H.G. 1993. Optimal utilization of fodder banks for the reproductive efficiency of Bunaji cattle: Effect of dry season forage legume supplementation on weight and reproductive performance. In: Adamu A.M., Mani R.I., Osinowo O.A., Adeoye K.B. and Ajileye E.O. (eds), Proceedings of a Workshop on Forage Production and Utilization in Nigeria, Zaria, Nigeria, 11–14 February 1991. National Livestock Projects Division, Federal Ministry of Agriculture and Water Resources, Kaduna, Nigeria. pp. 135–150. Mohamed-Saleem M.A. 1986. The establishment and management of fodder banks. In: von Kaufmann R., Chater S. and Blench R. (eds), Livestock System Research in Nigeria’s Subhumid Zone. Proceedings of the Second ILCA/NAPRI Symposium held in Kaduna, Nigeria, 29 October–2 November 1984 . ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 326–350. Mohamed-Saleem M.A. 1994. Stylosanthes for pasture development: An overview of ILCA’s experience in Nigeria. In: de Leeuw P.N., Mohamed-Saleem M.A. and Nyamu A.M. (eds), Stylosanthes as a Forage and Fallow Crop. Proceedings of the Regional Workshop on the Use of Stylosanthes in West Africa, held in Kaduna, Nigeria, 26–31 October 1992. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 17–23. Mohamed-Saleem M.A and von Kaufmann R. 1995. Fodder banks: improving the nutrition of cattle in the subhumid zone of West Africa. In: Ruiz M.E., Sere C. and Li Pun H. (eds), Global Workshop on Animal Production Systems. Research and Development Collection 26. IICA (Inter- American Institute for Co-operation on Agriculture) and RISPAL (Latin American Research Network for Animal Production Systems), San José, Costa Rica. pp. 25–54. Mohamed-Saleem M.A. and de Leeuw P.N. 1994. Stylo-based pasture development for agropastoral production systems. In: de Leeuw P.N., Mohamed-Saleem M.A. and Nyamu A.M. (eds), Stylosanthes as a Forage and Fallow Crop. Proceedings of the Regional Workshop on the Use of Stylosanthes in West Africa, held in Kaduna, Nigeria, 26–31 October 1992. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 139–145. Mohamed-Saleem M.A. and Suleiman H. 1986. Nigeria and West Africa fodder banks: dry season feed supplementation for traditionally managed cattle in the sub-humid zone. World Animal Review 59:11–17. Mohamed-Saleem M.A., Otsyina R. and von Kaufmann R. 1986. Some methods for improving fodder by incorporating forage legumes in cereal cropping systems in the Nigerian subhumid zone. In: Haque I., Jutzi S. and Neate P.J.H. (eds), Potentials of Forage Legumes in Farming Systems of Sub-Saharan Africa. Proceedings of a Workshop held at ILCA, Addis Ababa, Ethiopia, 16–19 September 1985 . ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 363–378. Otchere E. 1986. The effects of supplementary feeding of traditionally managed Bunaji cows. In: von Kaufmann R., Chater S. and Blench R. (eds), Livestock System Research in Nigeria’s Subhumid Zone. Proceedings of the Second ILCA/NAPRI Symposium held in Kaduna, Nigeria, 29 October–2 November 1984. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 204–212. Otsyina R.M., von Kaufmann R., Mohamed-Saleem M.A. and Suleimann H. 1987. Manual on Fodder Bank Establishment and Management. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. Singh I. and Subramanian J. 1986. Agricultural household modelling in a multicrop environment: Case studies in Korea and Nigeria. In: Singh I., Squire L. and Strauss J. (eds), Agricultural Household Models. Johns Hopkins University Press, London, UK. Tambi E. 1996. The dynamics of household beef consumption in Cameroon. Agricultural Economics 14:11–19. Tarawali G. 1991. The residual effect of Stylosanthes fodder banks on maize yield at several locations in Nigeria. Tropical Grasslands 25:26–31. Tarawali G. and von Kaufmann R. 1987. Fodder Banks: Benefits to Ruminant and Crop Production. Paper presented at the Regional Seminar on Forages and Ruminant Nutrition, held at Ngaoundere, Cameroon, 16–20 November 1987. 15 pp. Tarawali G. and Mohamed-Saleem M.A. 1994. Establishment techniques for Stylo-associated cropping systems. In: de Leeuw P.N., Mohamed-Saleem M.A. and Nyamu A.M. (eds), Stylosanthes as a Forage and Fallow Crop. Proceedings of the Regional Workshop on the Use of Stylosanthes in West Africa, held in Kaduna, Nigeria, 26–31 October 1992. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. pp. 183–192. Tarawali G., Dembélé E., N’Guessan B. and Youri A. 1998. The use of Stylosanthes by smallholder farmers for sustainable food production in subhumid West Africa. Manuscript in revision. Thomas D. and Sumberg J. 1995. A review of the evaluation and use of tropical forage legumes in sub-Saharan Africa. Agriculture, Ecosystems and Environment 54:151–163. van Velthuizen H.T. and Fischer G.W. 1995. Potential of Forage Legumes in Land-use Intensification Towards Sustainable Crop–Livestock Production Systems in West Africa. IIASA Working Paper 95–80. IIASA (International Institute for Applied Systems Analysis), Laxenburg, Austria. 109 pp. Winrock International. 1992. Assessment of Animal Agriculture in Sub-Saharan Africa. Winrock International Institute for Agricultural Development, Morrilton, Arkansas, USA. 125 pp. Appendix 1 Changes in producer surplus and elasticity of demand Under the assumptions of the economic surplus model used in the analysis, whether producers gain or lose from fodder bank technology depends on the magnitude of the elasticity of demand for the various commodities affected by the technology. If the elasticity of demand (?) is greater (less) than one, the change in producer surplus is positive (negative). This can be shown by substituting equilibrium quantity and price into the equation for the change in producer surplus and rearranging. Thus: Since 0