Development of Sorghum and Millet Production Systems in Mali and Burkina Faso

Supported by The Bill and Melinda Gates Foundation

Photo on front cover: Farmer and AMEDD NGO Agent with Grinkan, Summer 2008 Development of Sorghum & Millet Production Systems in Mali and Burkina Faso

John H. Sanders, Purdue University, West Lafayette, IN [email protected] Botorou Ouendeba, 3N Program, Government of Niger, Niamey, Niger [email protected] Ababacar Ndoye, retired, Formerly Director of ITA, Dakar, Senegal [email protected] Soungalo Traore, retired. Formerly Head of Research, National Extension Service, Mopti region, Mali [email protected] Niaba Teme, Crop Breeder, Senior Scientist, IER-Institute of the Rural Economy, National Agricultural Research Institution, Bamako, Mali [email protected]

This report was financed by the Gates Foundation with earlier work upon which it is based, financed by INTSORMIL, USAID, and the McKnight Foundation We especially appreciate the support of IER, INERA and the regional office of the national extension service for the Mopti region.

Content pages:

Recommendations for Moving from the Experiment 1 to 9 Station to the Farm with New Technology and Marketing Strategies

I. New Technology-Marketing Systems, Pages 11- 90 Introduction of a New Sorghum Technology into Mali 12 to 33 Sorghum Production Systems in Burkina Faso 34 to 46 Introducing New Millet Production Systems in Mali 47 to 76

The Mopti Field Program, 2015-2016 Crop Year 77 to 90

II. Secondary Markets, Pages 91 - 138 Demand for Processed Millet Products in Mali & Burkina Faso 92 to 112 Feed Grains as a Secondary Market for Sorghum in Mali &Burkina Faso 113 to 138 Acknowledgements:

We are grateful to the many scientists who made available their personnel and oriented us and to the specialists who identified and interviewed farmers, chicken producers, vets, and millet food processors for us. Farmers and the entrepreneurs were also especially gracious and patient with us and our detailed questions. Clearly without them it would have been impossible to accomplish this work. Below we specifically identify those most involved but we are also grateful to the many people interviewed and those who participated in the four workshops.

 Dr. Boureima DEMBELE, Directeur General de l’IER - Mali;  Dr. Ibrahima NDIAYE, Directeur Scientifique de l’IER - Mali ;  Mr. Oumar MAIGA, Directeur Regional, DRA-Mopti, Mali.  Dr. Abou BERTHE , Directeur, Sasakawa 2000, Mali.  Mr. Oumar SANDINAN, du Sasakawa 2000, Mali  Dr. Mamourou DIOURTE, Chef du Programme Sorgho de l’IER - Mali ;  Mr. Mamadou SOUFOUNTERA du Laboratoire de Technologie Alimentaire de l’IER – Mali ;  Ms. Salimata COULIBALY du Laboratoire de Technologie Alimentaire de l’IER – Mali ;  Mr. Soungalo TRAORE, Direction Regionale de Mopti  Mr. Ibrahima KONATE, Gérant AIID-Conseil - Coordinateur programme de développement des chaines de valeur mil et sorgho – Koutiala - Mali ;  Mr. Gaoussou DIAWARA, Coordinateur de l'ONG Centre d'Appui à l'Autopromotion pour le Développement (ONG-CAAD/KOUTIALA-MALI)  Mr. Mamadou KEBE, Chef secteur Agriculture de Koro  Mr. Issa DAOU, Chef sous secteur Agriculture Bankass.  Dr. Hamidou TRAORE, Directeur General de l’INERA – Burkina Faso ;  Dr. Bréhima DIAWARA, Directeur du Département de Technologie Alimentaire (DTA/INERA)- Burkina Faso ;  Dr. Laurencia OUATTARA/SONGRE du Département de Technologie Alimentaire (DTA/INERA)- Burkina Faso ;  Mr. Michel COMBARI du Département de Technologie Alimentaire (DTA/INERA)- Burkina Faso ;  Dr. Fabrice BATIONO du Département de Technologie Alimentaire (DTA/INERA)- Burkina Faso ;  Mr. Pierre Paul SANON Ingénieur d’élevage Burkina Faso ;  Ms. Gisele Kazoni, ONG Afrque Verte, Burkina Faso.  Dr. Donald LARSON, retired Professor, Ohio State  All the members of the farmers’ organizations and entepreneurs and other scientists that we interviewed. Recommendations for Moving from the Experiment Station to the Farm with New Technology and Marketing Strategies John H. Sanders, Botorou Ouendeba, and Ababacar Ndoye Introduction From 2004-2013 we implemented a pilot program of new technology and marketing strategy introduction in up to four Sahelian counties. We also coordinated with a USAID supported project to scale up our production-marketing system in Mali from 2010 to 2013.This project (IICEM) incorporated the banking sector for fertilizer loans for the farmers’ associations and facilitated contracts between the farmers’ associations and wholesalers to buy the sorghum and millet. From September 2014 to March 1, 2016 the Gates Foundation sponsored us in reviewing the progress and performance of these programs and requested us to evaluate the potential of the secondary markets for millet and sorghum to develop and thereby reduce a price decline from successful technology introduction or favorable weather conditions. Here we consider our five recommendations from this work, three for new technology- marketing system introduction and two for accelerating the growth of the secondary markets: A. Introducing New Sorghum and Millet Technologies and Marketing Strategies 1. Identify and Support Research-Extension on Second Generation Problems; 2. Insure Access to Higher Analysis Inorganic Fertilizer and Regularly Renew the Seed; 3. Strengthen Farmers’ Associations for Sorghum and Millet as Marketing Cooperatives. B. Accelerate the Growth of the Secondary, Domestic Markets for Sorghum and Millet 4. Reduce Costs in Millet Processing to Expand Sales to Middle and Lower Income Consumers; 5. Substitute Sorghum for Maize in the Feed with Increasing Sorghum Yields. Second Generation Problems Many developing countries1 have succeeded in training agricultural scientists and in implementing good agricultural research in national experiment stations. The primary question here is why there have not been more successes with the two basic food crops in these Sahelian countries, sorghum and millet. We propose another explanation besides the traditional ones. Our

1 We have worked in the Sahelian countries of Mali and Burkina Faso for this contract and previously in Niger and Senegal. So the discussion here draws principally upon our recent work in the first two but also includes some consideration of the second two.

1 explanation is the difficulty (lack of available funds) of either national research or extension agencies to respond to second generation problems. There is a litany of reasons given to explain why farmers and donors are not interested in sorghum and millet: these crops do not respond to fertilizer; fertilizer use is not profitable for them; farmers are too risk averse to spend money on these inputs; increasing productivity would lead to price collapses due to lack of secondary markets. To summarize. New sorghum and millet cultivars respond to fertilizer and are profitable when implemented along with improved marketing (Coulibaly et. al., 2015, pp. 55, 57; Sanders, J.H., B. Ouendeba, A. Ndoye, S. Traore, N. Teme [This bulletin], 2016, references, pp. 27, 28). Farmers are willing to take risks and the increased incentive for more effective fertilizer use is the availability and promotion of higher analysis fertilizers. Besides the higher analysis fertilizer recommendation seed renewal with certified seed was given little attention in the scaling up process. So these two production factors are important but the basic problem is an institutional factor involved in the financing of research and extension. The narrow definition of Second Generation Problems is the response of the biological system to the changes incurred with the introduction of new technologies. When the new shorter season dwarf rice was introduced by IRRI (International Rice Research Institute) in the sixties, there was an increase in the number of rice crops per year thereby substantially increasing insect problems. So research on insect resistance and chemical control was undertaken by this international center. Agriculture is a continually changing biological system so new problems are going to arise over time. If the technology is then rejected when these problems occur, the discovery and extension systems have to start over. Moreover, a broader definition of Second Generation problems would include the elements that the breeder (or other technology developer) did not realize were important. For example, with Grinkan, a high yielding sorghum with several other desirable properties, the softness of the variety meant that following traditional processing methods gave a “to”, the basic food commodity in rural areas, that people did not like. In food science terms this is the consistency problem. As a result women refused to buy Grinkan in the village markets. This information was communicated up to regional and national markets. So the markets collapsed for a cultivar offering 50 to 100% yield increases with moderate fertilization over local cultivars (This bulletin, 2016, PP. 19-22, 36-38). Breeders or other technology developers cannot be expected to anticipate all potential problems. So responding to problems identified over time especially with an initially successful technology needs to be an important activity for the research and extension services. So why does not this processing problem become a research and/or extension issue? The main problem with national research and extension agencies is that most of their funding is needed for personnel making them dependent upon outside sources for the operating budget. So developing country management, individual scientist, and extension agent attention are focused on obtaining outside funding. Sometimes programs will build on or extend successful programs or the technology will be so successful that it no longer requires donor support. However, the usual pattern is that when one donor leaves, another comes in and does not continue the program

2 of the previous funders. Where is the incentive of a new donor or a new program to resolve a problem leading to the decline of someone else’s initiative? So the high yielding Grinkan deserves to be thrown out if the primary food resulting from it is not acceptable. This a serious complaint. But what if it is not difficult or expensive to resolve this processing problem? Who will be concerned with that among national management, scientist or extension agents when all are focused on the present cash inflow and responding to the next donor program? It turns out that either by not soaking overnight or by not separating the bran the consistency problem can be resolved. With the first alternative of not soaking overnight the “to” of Grinkan was considered to be better than the local sorghums. We did a short video presentation and got this information on TV in Mali in Bambara. Will this be sufficient to encourage a revival of Grinkan? We will know over the next few years. The concept of identifying technologies that show initially very good results or that are expected to do so after eliminating a specific problem or two is important. Developed country research systems have the technical and economic capacity to make these kind of responses to emerging problems as did IRRI in the sixties with rice. Good impact analysis of released technology would identify areas for further research if undertaken by combined teams of economists, food and agricultural scientists. Unfortunately, impact evaluation is regarded principally as a public relations activity to generate support for research rather than a diagnostic tool to support technology introduction. What needs to be done in developing countries? A. Identification of the problems. There should be more on-farm evaluation of technology performance to identify and to anticipate problems. As noted above Impact Evaluation tends to focus on success stories to generate support from public policy makers and donors. When there is no clear case, there is pressure to twist or massage the data to avoid the fate of the messenger of bad news. A greater emphasis on the more complicated task of analyzing the whole system of production and sale as well as anticipating problems is an urgent need to supplement the many innovations attempted and then quickly discarded. The inclusion of agricultural scientists with specific crop experience on these teams as well as social and food scientists would help identify and even anticipate second generation problems. B. A Funding Mechanism. Second generation problems are not nuclear science. Researchers and extension agents know about them. Food scientists are familiar with the consistency problem encountered with Grinkan. But where is the money or incentive to focus on resolving these problems? In the initial years of the CGIAR (Consultative Group on International Agricultural Research) system the World Bank collected money from many donors and then gave block grants to individual centers to pursue research to increase yields for their specific commodities. The international centers were discouraged from chasing funding except for their designated commodities. Moreover, they were evaluated for their success in attaining productivity goals on farms in developing countries on these designated crops as well as on performing quality research. Could funds be set up for developing countries and

3 allocated to maintain research gains by identifying and responding to emerging problems with new technologies? Access to Inputs (Inorganic Fertilizer, Certified Seed, Credit) Some basic comments about soil fertility. It is well known that Sub-Saharan African soils are often deficient in nutrients especially N and P and that much of the last two decades have been lost searching for cultivars that require minimum nutrient levels or that there existed a magic combination of rotations with legumes, manure and other things that could substitute for moderate levels of inorganic fertilizer. So the first things to do is to develop recommendations for moderate inorganic fertilizer use and then combine that with a fertilizer responsive cultivar. This is well accepted for the “cash crops’ such as cotton and maize and substantial yield gains have been made this way. The argument that this will not work with “subsistence crops,” such as sorghum and millet, has been shown to be without an empirical basis and refuted in our 2 fieldwork (see the previous section for references). Organic and inorganic fertilizers are complements at low fertility levels rather than substitutes. The organic fertilizers will be especially important where soils are so light that there is substantial leaching. The organic fertilizers improve the structure of the soil and provide microbial activity but their basic nutrient levels (N and P) are extremely low. When implementing a large program, it is often easier to utilize the most accessible inputs as donors are in a hurry to increase the scale. Moreover, program managers scaling up often want to make management decisions about program components. The problem of moving from a pilot project3 to the scaling up comes when those implementing the scaling up take short cuts in this case on both the fertilization and the variety choice. Since neither the fertilizer recommendation nor the appropriate cultivar were consistently followed, it is not surprising that the scaling up yields were substantially lower than in the pilot project.4 This points out the need

2For sorghum and millet it is necessary to be concerned with marketing as well as production since with the occasional marketing of small surplus or in very good years with a price collapse at harvest farmers pressed by liquidity needs receive very low prices. In poor rainfall years farmers lose a potential large price increase from selling at harvest as prices can increase 50 to 100%. We return to storage and marketing in the next section on the roles of the farmers’ organizations in the marketing process. 3 From the pilot project there is now substantial yield and profitability evidence as well as farmers’ preferences for the moderate fertilization recommendation of one sack of DAP and one sack of Urea per ha across a wide range of regions for both sorghum and millet. Soil specific fertilizer recommendations are most important where soils are not very low in initial fertility and actual fertilizer use. As moderate use is expanded we need to pay more attention to region and soil specific differences. 4 DAP is a high analysis fertilizer (64% nutrient) as compared with the standard NPK (compound fertilizer- 45% nutrient for the 15-15-15). From calculation of the levels of N and P from the bags, asking farmers, or the economic analysis it is clear that it takes two bags of NPK to get the equivalent levels of one bag of DAP. So there is a general preference among farmers, who have tried DAP, to utilize DAP instead of NPK. DAP does not contain K, which is generally initially adequate in Sahelian soils. At some point probably five to ten years it will be necessary to use K but high analysis K is available internationally.

to continue the pilot project even with the initiation of the scaling up as a measure of what the potential yields are from following the recommendations and utilizing good management. Certified seed is very expensive due to the extension service needing to visit several times during the crop season and to do laboratory analysis. However, if the extension service could work with the associations of seed producers to visit several farms the same day and adjust their pricing mechanism to reflect this cost saving, this could reduce certification costs. As the Sahelian counties have been trying in 2015 to push sorghum hybrids onto farmers’ fields there has been a big increase in resources going into on-farm seed production. It is unlikely that these hybrids will out-yield and be more profitable than Grinkan, so this investment should benefit Grinkan (Niaba Teme, sorghum breeder, IER, PhD in Agronomy from Texas A&M, conversation, February 2016).

Finally we need to consider the credit problem for obtaining the inputs. To obtain bank loans with the scaling up project it was initially necessary for the farmers’ associations to show contracts of wholesalers to buy the Grinkan sorghum. This is several stages in the marketing chain above the collectors with whom the farmers normally dealt. Wholesalers had agreed to pay the market price plus a premium for clean cereal. The premium was low for quality as it was less than the average increased value of the sorghum after eliminating the impurities (Toure, 2007, pp. 13-17, 37, 38). Nevertheless, the need of farmers for input credits meant a substantial bargaining advantage for the wholesalers in dictating the conditions and the timing of the repayment. Wholesalers wanted the cereal during the post-harvest price collapse period. Moreover, the wholesalers wanted the farmers to sell more of their sorghum than just enough to repay the bank loan. Many farmers’ associations ignored both of these conditions by selling the cereal later after some of the price recovery and only selling enough at those prices to repay the bank. The farmers’ associations need to get better at negotiating and be prepared to use arbitrage, ie sell outside the immediate market area if there is cartel price fixing among the wholesalers. Another alternative is to develop with donor assistance or member savings revolving funds to pay for inputs each year. In the Mopti region USAID introduced this financing for input purchase by creating revolving funds and in the south the Dutch government did the same during the economic disaster year of 2012.5 These revolving funds are still functioning in the farmers’ associations we interviewed. By self-financing the farmers’ associations now still deal with the wholesalers, who they now know, but their bargaining power will be increased if they maintain clean, uniform cereals and do market searches if there is price fixing by the buyers. So this combination of access to the higher analysis fertilizer, increased incentives for the supply and the demand for certified seed, and credit at favorable terms are the components for a rapid introduction of new technology. As Grinkan makes a comeback from the “to” processing problem, there needs to be more interaction with public policy makers on the fertilizer 5 This included the ousting of one Malian President and the beating by a mob of his successor. It also was the year of the rebel attacks in the north of Mali, which then attempted to move down from near Sevare to Bamako before the French attacked with planes and helicopters and decimated them. The banks in the north including our Mopti site pulled out their resources and came south.

5 requirements and with the extension service to reduce their costs of supervision of the certified seed production. Finally the farmers’ associations can attain more bargaining ability with the wholesalers either in or outside of the region. With the introduction of PAM (UN program to attack the malnutrition program)6 this cartel problem has not been serious in the Mopti region and PAM is increasingly intervening in the Koutiala region with sorghum. But after PAM and P4P? We have already started looking at marketing issues so let’s continue this in the next section. Institutional Innovation of Developing Strong Farmers’ Association for Sorghum/Millet Farmers in a farmers’ association are more likely to imitate the behavior of successful neighbors in the association than to believe outside recommendations. So if a technology is recommended to an association, some farmers will follow it well and those that do not will often imitate the successful farmers the next year. The retention of the revolving funds from the pilot project by the farmers’ associations was excellent and in many cases expanded. Then on a much large scale USAID and the Dutch government followed this model in Mali with the farmers’ associations they helped as part of the scaling up when the economy collapsed in 2012. Moreover, the farmers associations receiving the revolving funds continued dealing with the wholesalers thereby skipping two stages in the marketing chain by selling in quantity (see the figures of the marketing structures for Mali and Burkina Faso-This bulletin, 2016, pp. 99-104). Besides fertilizer and new varieties farmers need the capacity to avoid the post-harvest price collapse and to be able to increase their bargaining ability by selling at higher levels of the marketing chain or outside the region. The farmers associations can still get better at negotiating and they need to avoid fragmenting into separate associations. However, the concepts of quality control for a price premium, avoiding sales during the post-harvest price collapse unless there is a separate premium for selling early, and skipping stages in the marketing chain are now well known and generally acted upon in the farmers’ associations involved in the scaling up (This bulletin, 2015, pp. 95, 96, 102-104). Now the farmers’ associations need better performance in terms of transparency with their members and training in marketing, accounting, and business management. The farmers’ association is a marketing coop for sorghum or millet that handles revolving funds, does storage and quality control of clean cereal, and sells in quantity at higher stages of the marketing chain. This is an excellent institutional innovation to insure that the use of higher inputs with sorghum and millet remains profitable. However, for this model to work it needs the introduction of excellent production technologies including new cultivars that motivate members and drive the system. This is why it is important to get Grinkan (or another high yielding variety) going and expanding again the size of these farmers’ associations. Toroniou is working fine in

6 PAM is now combined with P4P, the Gates supported project to make sure that small farmers selling the basic food staples also benefit from this program of the UN to help the malnourished. An agency aimed at the malnourished would be expected to buy the cereals as cheaply as possible in the post-harvest price collapse period so as to maximize coverage of the malnourished from their budget allocation.

6 the farmers’ associations in the Mopti and Segou regions7 but needs to be renewed with certified seed (This bulletin, 2016, pp. 46-90). Reduce Costs in Millet Processing to Expand Sales to Middle and Lower Income Consumers Millet is the principal cereal produced in Mali and there are many traditional dishes in the villages. As women have moved into urban areas there have been consumption shifts to rice as the processing and preparation demands are much less than for millet or sorghum and the value of women’s time is higher in urban areas. Food scientists have responded by producing processed millet products that require a simple cooking technique as with rice. In activities operating initially out of their homes women have begun millet food processing in many urban areas. They make products that are ready for cooking as with rice. Donors have helped entrepreneurs with equipment and hygiene training. The millet processing sectors produces quality products and sells overseas to the diaspora, to high income urban consumers and for special holidays. However, they are having difficulty being cost competitive with either rice or with women buying the millet and having it milled locally and producing themselves the traditional dishes (This bulletin, 2016, pp. 96-99). Consequently, millet sales are not expanding as fast as they could to middle and lower income consumers. The millet processing sector is still very small compared with the quantity of millet produced. There are changes in the sector already on-going, which can be facilitated and would reduce the cost of processed millet products. The sector needs greater quantities of clean millet. This is an interaction process with the famers’ associations since a premium for clean cereal needs to be paid and there needs to be quality control as with branding by the farmers’ organizations. This process is slowly advancing but in the scaling up process discussed above the premium offered was less than the value of the additional millet from removing impurities. Secondly, as the farmers’ associations deal with higher levels of the marketing chain, the margins for the lower levels are no longer needed. These margin savings formerly received by collectors and regional merchants will be competed over by the farmers’ associations and the wholesalers. Some of these cost savings will be passed on to the millet processors. Finally and perhaps the most important innovation is to introduce techniques to extend the shelf life in millet as is done with wheat. There is a simple machine that with centrifugal force separates the grains from the weevils and their eggs. Since all the millet products start with the millet flour, this machine, the entoleter, changes everything. The size and through-put of the machine is too large in the US and Europe but small size machines are available from China and India and are increasingly being purchased by millet food processors. Sorghum as a Feed Grain The dramatic productivity gains of maize especially in the last decade have facilitated the rapid growth of the intensive poultry industry in both Mali and Burkina Faso. However, maize production is concentrated in the high rainfall south and has more difficulty moving north into the lower rainfall and low soil fertility regions than sorghum. As the intensive poultry sector

7 These are the two principal regions of millet production in Mali.

7 continues its’ accelerating growth cereals will need to provide 50 to 60% of the feed composition. Much of the maize research to facilitate this northward movement has been oriented to make maize more like sorghum by selecting for earlier maturity for drought escape and searching for more basic innovations for drought resistance. High yielding sorghums are increasingly available especially in the high rainfall zone. In contrast maize has already reached high yields there and has less potential for further yield increase. Increasing sorghum productivity requires more public policy support especially making high analysis, inorganic fertilizer more available and reducing the costs of certified seed production. These innovations will also help reduce the costs of maize production. However, sorghum will be much more competitive moving north and can compete with maize by extending the area there in higher yielding cultivars without necessarily attaining the same high yields of maize as in the south. Besides increasing sorghum yields the use of sorghum in the feed requires some basic information being provided to the producers and their feed advisers. There is a common misconception that all sorghums have tannin. Tannin reduces digestive efficiency in poultry especially in chicks and adolescent chickens. But non-tannin sorghum has 95 to 97% of the nutritional efficiency of maize. So at sorghum prices below 95% of those of maize, profit maximizing intensive chicken producers should switch their feed source to sorghum (This bulletin, 2016, pp. 115-120, 130-134, p. 129 Abdoulaye et al., 2006, pp. 13-17, 37, 38). Unfortunately, with the rapid productivity growth of maize and the lack of growth of sorghum productivity, the prices of sorghum have been much higher for sorghum than for maize (see This bulletin, 2016, pp. 117,118 for the relative prices for the last decade in both countries). Moreover, most intensive chicken producers and veterinarians or others advising on feed components have heard of digestive problems with sorghum in chicken. But they often do not know the details nor are they aware of the non-tannin sorghums. So sorghum remains as a potential feed as productivity is raised and as it becomes more difficult to further raise maize yields8 to respond to the rapid increases in feed demand expected. However, it will be necessary to provide more information about non-tannin sorghums to the intensive chicken producers and their feed advisers. Later in the marketing to feed producers the non-tannin sorghums would need to be identified and not mixed with other sorghums. References: Abdoulaye, T. and J.H. Sanders, 2006. Sorghum or Maize in West African Poultry Rations, INTSORMIL Bulletin No. 4, Department of Agricultural Economics, Purdue University, West Lafayette, IN. 47907, 17 pages. Also available in French Coulibaly, J. Y., J.H. Sanders, P.V. Preckel, T. G. Baker, 2015. “Will cotton make a comeback in Mali?” Agricultural Economics, 46, 53-67 Sanders, J. H., B. Ouendeba, A. Ndoye, S. Traore, and N. Teme, 2016. Development of Sorghum and Millet Production Systems in Mali and Burkina Faso. Bulletin Produced for the Gates

8 Removing the restrictions on maize exports in Mali would be expected to result in higher maize prices.

8 Foundation, Department of Agricultural Economics, Purdue University, West Lafayette, IN, 47907, 138 pages Toure, A., T. Abdoulaye, J. Sanders et B. Ouendeba, 2007.Transformation Commerciale du Mil et du Sorgho au Mali, INTSORMIL Bulletin No. 7, Department of Agricultural Economics, Purdue University, West Lafayette, IN. 38 pages.

9

10

New Technology-Marketing Systems

11 Introducing a New Sorghum Production System in Mali John H. Sanders, Botorou Ouendeba, and Ababacar Ndoye, Abstract Introducing a Caudatum sorghum into zones of principal production of Guineas has been an activity with a long history in Mali.1 Guineas have open heads hence are less subject to mold with late rains but this same characteristic results in little potential for high yields. Farmer selected Guineas have stable but low yields. Caudatums have compact heads hence high yield potential and are the predominant sorghum type in developed countries. However, this same compactness of the heads results in susceptibility to mold with late rains. Mold reduces yields and in the absence of thorough drying leads to very poor germination in the next production season. To make a substantial yield increase without increasing too much the mold risk Malian breeders have been crossing Guineas and Caudatums in the late ‘80s and ‘90s. An open pollinated cross between the two sorghum races was identified in Mali in the early ‘90s that was successful on farmers’ fields in Mali with average yields of 1.5 to 2 ton/ha in 2008 and 2009 with the better farmers getting 2 to 3 tons/ha. After the second year of high sorghum yields USAID/Mali financed a scaling up approach for sorghum and millet technologies of the pilot project. This scaling up program was successful in incorporating bank lending and developing contracts between the farmers’ associations and the wholesalers. However, with the difficulty of renewing the seed and obtaining the recommended fertilizer, there was substantial technology erosion leading to disillusionment with Grinkan. This was combined with a failure of village women to modify traditional processing methods for preparing “to”. As the word of poor “to” quality spread, village, regional and major urban markets declined or disappeared. However, in some villages the women adapted different methods of processing to avoid the lack of stickiness of the “to.” We discuss the general process of rapid increase, decline and then recovery we observe with Grinkan in Mali. We dedicate substantial attention to the problems of scaling up technology. Then we consider several second generation problems of introducing a new sorghum cultivar. Finally we make some projections of the future potential of Grinkan and Grinkan types. The Path of Technology and Product Market Introduction New technologies and new products have similar diffusion paths. With the initial push from sponsors as with demonstrations/subsidies or of products with advertising or discounted prices, there is a rapid increase in area or sales. Then the sponsor or advertiser withdraws. At this point (A) problems with the new technology are often noted and there is a decline in the

1INTSORMIL (Texas A&M) and IER scientists had worked on this in the ‘80s and ‘90s.

12 introduction process. These problems are often called second generation problems. 2 For products the appeal of the newness often declines resulting in sales decrease (AB). It is typical of a new technology to have a rapid burst of activity as it is promoted, then to decline or even collapse once the innovation promoters leave. After this the question of revival and long run growth is what we are looking for with Grinkan given its high yields and excellent taste acceptance except for the lack of consistency of the “to” with the traditional food processing techniques. As the processing problem is resolved, do we expect Grinkan or a Grinkan type cultivar to resume the rapid diffusion process (BC)? Figure 1. Technology and New Product Diffusion Process

A AREA IN NEW TECHNOLOGY /NEW PRODUCT SALES C

TIME

Introduction: From 2004 to 2009 we worked in Mali with IER (the national agricultural research system in Mali) and the NGOs, AMEDD and Sasakawa 2000. During that time we tested many IER new cultivars and some ICRISAT cultivars in our village level, pilot plot activities. The local cultivars were Guineas. We were looking for a Caudatum (densely packed head) or at least a partial (crossed with a Guinea) Caudatum to get high yields. We also excluded over time the taller cultivars that produced mainly stalk and leaves and lodged with moderate fertilization. We recognized that the down side of a Caudatum or partial Caudatum was the susceptibility to head

2 The first generation problems for a new sorghum technology are to increase the yields over local cultivars in an economical manner (often including resistance or tolerance to disease and insects) and to have an acceptable or preferred taste as compared with the local cultivars. Second generation problems of Grinkan are food preparation techniques, storage and field insect control, seed and fertilizer quality and accessibility, and market price differentiation, credit market access, and marketing power of the farmer or farmers’ association. We discuss these later in the paper.

13 bugs and mold with late rains. But believed that the high grain yield potential of Caudatums was worth that risk for most farmers. The higher rainfall regions in the Sahel are primarily cotton regions. Who cares about raising cereal yields when white gold is available? Unfortunately there have been several problems with cotton in the Sahel. Historically cotton zones have problems with declining soil fertility over time. Cotton yields have been stagnating and even declining in the Sahel in the last decade. Secondly, the introduction of Bt cotton in major cotton exporters (China, India and the US) has reduced the costs of controlling insects and/or given yield advantages to adopting producers. These producers then increase the supply and that results in declining international prices for cotton and increasing difficulty for non-adopters of Bt to compete internationally. This is especially noticeable in the more marginal cotton regions with lower rainfall and greater soil fertility problems such as in Koutiala. From 1998 to 2008 cotton production and area declined by over 60% in the Koutiala sub-region of Sikasso, Mali (Figure 2; Coulibaly et al, 2015, p.55).3 In the 21st Century increasing amounts of cotton fertilizer allocations provided by the cotton parastatal (CMDT) had been diverted to maize and sorghum. Maize yields have substantially increased there. So there has been increasing interest in raising cereal productivity in the cotton zones.

Figure 2. Cotton Area and Production in the Koutiala Area from 1998-2008.

Source: J. Coulibaly et al., 2015, p. 54.Data adapted from the Malian Ministry of Agriculture.

3 There was a recovery in 2011 with a 38% price increase after the 80% world price increase of 2010. This cotton price stayed high in Mali for several years before beginning another decline (Coulibaly et al., 2016, p. 54)

14 Scaling Up with the IICEM Program, 2010-2013: USAID/Mali got excited with the pilot project results for Grinkan and associated production and marketing innovations (See Pictures 1and 2). With the demonstrated high yield potential of Grinkan with moderate fertilization USAID contracted IICEM to scale up this pilot program nationally by engaging the national development bank (BNDA) to finance the inorganic fertilizer. IICEM arranged the contracts between the farmers’ associations and the BNDA and put up a guarantee fund for 1/2 of the value of the bank loans made to farmers’ associations in the first year, 2010-11.

Picture 1.

15

Picture 2. Grinkan in Garasso, Koutiala region of Mali, 2008 To increase credibility of the farmers’ associations’ ability to repay the loans, IICEM worked with wholesalers making contracts to buy the sorghum from the farmers’ associations. The contracts were for the market price at sale plus a premium for quality of 10 cfa/kg. The contracts specified that sale was to be made soon after harvesting when threshing was done. However, most farmers’ associations delayed threshing to take advantage of some of the price recovery normally occurring after the harvest. Many farmers’ associations had been engaging previously in storage construction with NGO or donor support so that they could obtain some of the marketing margin from the seasonal price variations, undertake the collection operation, and sell a clean cereal in quantity. On the financial side the program was very successful as the credit guarantee required by the BNDA was reduced to 1/3 the second year and was eliminated in the third year of the program. This was an amazing success story of the IICEM program in working with a large

16 number of farmers’ associations4 obtaining and then repaying credit over four years. Moreover, IICEM established contacts between the upper levels of the marketing system, ie the wholesalers, and the farmers’ associations. Financing from the BNDA was provided for fertilizer but not for seed. The neglect of seed renewal became a critical weakness of the program as there were no systematic efforts to produce and distribute improved Grinkan seed.5 Some farmers were encouraged to produce seeds. But with the food preparation problem with the “to” there was a loss of interest in Grinkan quality seed (see the next two sections for further discussion of this problem). In 2012 the IICEM program was disrupted by the war in the north, a coup throwing out the President shortly before the next election, the beating of the transitional President by a mob and French intervention against the rebels as they were preparing to attack south from Sevare on Bamako. The Banks in the north including the Mopti region shut down. In the Mopti region and to a lesser extent in the south, where the banks continued to operate, donors (USAID in the Mopti region and the Dutch in the south) stepped in to maintain the financing of the inorganic fertilizer for some of the farmers’ associations. The donors provided loans for the fertilizer purchases to be repaid to the farmers’ associations in kind.6 Then the farmers’ associations were expected to hold onto the sorghum until shortly before the next planting season to sell and then provide fertilizer loans in that crop season. The revolving funds eliminated the need for bank loans and the contract obligations with the wholesalers.7 The tendency to fix prices in cartels among the wholesalers was reduced by the activities of PAM (UN program to improve nutrition) in buying cereals and giving price premiums for quality to the farmers’ associations. From the second year of the IICEM program (2011-12) there was technology erosion as the program stopped promoting either Grinkan or DAP, two basic technology components of the pilot project. This technology erosion is a chronic problem in moving from pilot projects to scaling up. Program administrators pressed to expand rapidly were faced with input shortages (certified Grinkan seed in this case) or public policy decisions. In government sponsored programs including fertilizer subsidies the higher nutrient level DAP was set aside for rice production. For the millet and sorghum producers only subsidized NPK was available and no subsidized Urea. Even that NPK subsidy was not available to all producers but rotated annually the areas that could obtain the subsidy. To replace the nutrient levels of one sack of DAP two

4 In the IICEM Koutiala area program there were 174 farmers associations in 2011-12 and this was reduced to 134 as the Feed the Future program attempted to focus more their activities. Unfortunately there was a tendency for these associations to split into smaller units thereby reducing their economic power to negotiate or sell larger quantities. Probably in retrospect IICEM should have done more to avoid this fractionalization and to assist the farmers’ associations in gaining market power with a larger share of the marketing margin. The institutional development and improved marketing ability are two of the most important components of the technology introduction process.. 5 The costs of the certified seed are still very high as the extension service requires repeated farm visits at high costs. 6 In the north (Mopti for millet) the USAID provision was that the full value of the fertilizer loans be repaid in kind to the famers’ association. In the south the Dutch provision was that only one-half of the value of the fertilizer loans was to be repaid to the farmers’ associations. 7 In our interviewing in 2015 in the farmers’ associations in Mopti and to a lesser extent in Koutiala these revolving funds from 2012 had been maintained in the farmers’ associations.

17 sacks of NPK would be needed.8 As only one sack of NPK was substituted per ha this seriously compromised the yield levels of Grinkan (see Figure 3). This shift from Grinkan and DAP in the IICEM program of 2011-12 is expected to be a principal explanation for the abrupt decline in yields in that year. This also indicates the importance of continuing the pilot program even when the scaling up proceeds in order to flag the technology erosion when it occurs and to demonstrate the yield potential of following well the pilot program recommendations An important objective of both the pilot and the scaling up project was to strengthen the farmers’ associations especially their ability to claim a share of the marketing margin. The farmers’ associations became the collectors of cereal and began selling in larger quantities to regional merchants or wholesalers. The farmers’ associations held the cereal and waited for the price recovery as typically the regional merchants and wholesalers do. In the IICEM contracts with the wholesalers the farmers’ associations were pressured to supply the threshed grain quickly after the harvest to benefit the wholesalers from the price recovery after the harvest. However, in the third year those farmers’ associations receiving the emergency funding could hold their cereals longer. Moreover, by then they already had contacts with the wholesalers and more bargaining power when there was no bank involvement but continuing use of the revolving fund.

Figure 3.

Evolution of Grinkan yields for 33 interviewed villages in Koutiala 1500 1600 1420 1387 1400 1307 1206 1161 1191 1200 1000

Yields 800 Yields average (in kg/ha) in

Kg/ha 600

400

200

0 2008/2009 2009/2010 2010/2011 2011 /2012 2012 /2013 2013/2014 2014/2015 Years

Source: unpublished data from village interviews in the summer and fall of 2015, Koutiala.

8 Not only does the substitution of one sack of NPK for one sack of DAP substantially lower phosphorous levels but costing by bag is not even appropriate. DAP is 66% nutrient whereas the predominant type of compound fertilizer is only 45%. Costing needs to be by the cost of providing the essential nutrients.

18 So what happened to prices and marketing margins? We can consider two periods before and after the war and planned election year of 2011-12. Before this year there were substantial marketing margins of 20 to 30 fcfa/kg between the collectors and the regional wholesaler (Figure 4).

Figure 4

Prices paid to farmers associations by regional merchants and wholesalers

250

192 200

143 135 135 133 132 Prices paid for 148 Collectors Grinkan in 137 F CFA/Kg 129 123 Wholesalers 100 115 105

50

0 2009 - 20102010 - 20112011 - 20122012 - 20132013 - 20142014 - 2015

Source: unpublished data from village interviews in the summer and fall of 2015, Koutiala.

In the months preceding the Presidential election of 2012 the government made contracts with the wholesalers to buy at 220 cfa/kg according to local sources with the understanding that farmers or farmers’ associations would be well paid. So early in the year (January, February) before the normal price recovery period wholesalers were buying at 190 cfa/ha. The apparent objective was to also sell the cereal at a subsidized price for consumers. Therefore farmers and consumers would benefit from millet and sorghum prices shortly before the elections. Unfortunately for the President there was a coup before the elections. After 2011-12 the difference between wholesalers and collectors disappeared. The two principal Koutiala sorghum wholesalers regularly visit the farmers’ associations and the village markets to buy directly from farmers’ associations and farmers. So now the wholesalers and the regional merchants, who previously financed local collectors, are doing the same function as larger quantity purchases can be obtained from the farmers’ associations. Prices received by farmers’ associations were forced up by more competition from various levels of the marketing chain and by the marketing functions assumed by the farmers’ associations. Taste and Marketing Problems in Introducing Grinkan: In the pilot project stage Grinkan consistently out-yielded local sorghums by 50%. Local sorghums are tall and have loose grain formation in the heads. Hence their maximum yield potential is expected to be about 1.2 tons/ha with the expectation that following cotton or maize

19 (both heavily fertilized) in the rotation average yields would be 800 kg/ha to 1 ton/ha (Coulibaly et al., 2015). Farmers following recommendations during normal and good rainfall years 9 expected Grinkan yields of 1.5 to 2 tons/ha with very good farmers getting 3 tons and higher. The combination of poor “tô” and the effect of that on local and national markets appears to be the major factor for the rejection or price discounting of Grinkan. In the local markets, where farmers sell small quantities of sorghum to get a little cash on market day and make their small purchases, village women started avoiding Grinkan because of the “tô” quality with traditional processing methods.10 In turn this information was passed on to larger markets so that merchants began to avoid or price discount Grinkan. Then most individual Grinkan producers began focusing on their own and livestock consumption of Grinkan rather than sales through the farmers’ associations. So these two related complaints of the food quality and markets resulted in a sharp contraction of Grinkan area. Of the 33 villages interviewed in 2015 in the Koutiala area the area in Grinkan declined from a peak level of 1184 ha to 710 ha in 2014-15, a decline of 40% 11 (Figure 5). In most villages in Mali Grinkan production has shifted from a group activity to private individuals producing and marketing individually. This private rather than collective activity makes it more difficult for farmers to earn any of the marketing margin by skipping stages in the 12 marketing process and being able to negotiate collectively.

Figure 5

Evolution of total cultivated area of Grinkan for 33 interviewed villages in Koutiala

1400 1184 1200 1000 739 711 800 652 Total 524 600 Total cultivated area… area 400 200 100 200 0 2009/2010 2011 /2012 2012 /2013 2014/2015 Years Source: unpublished data from village interviews in the summer and fall of 2015, Koutiala.

9 Grinkan needs moderate levels of inorganic fertilization (one sack of DAP and one sack of Urea or two sacks of NPK and one sack of Urea). Also especially in the lower fertility regions some organic fertilizer is necessary to improve the structure and better retain water and nutrients. For a more complete set of agronomic and marketing recommendations see Ouendeba et al, 2013. 10 This meant a lack of stickiness of the “tô“ To food scientists this is the consistency problem. A section follows explaining the scientific basis of this problem and how it is being overcome in some villages in Mali. 11 With the late rains there was a further decline in 2015-16.Grinkan is normally planted late to avoid being mature with the late rains. Cotton and maize are planted earlier. With the late commencement of the rains cotton and maize were planted late and Grinkan largely dropped out of the rotation except for a few individuals committed to Grinkan These individuals followed the recommendations better unless they were just producing Grinkan as a forage. Hence, yields jumped up in 2014-15 (Figure 3). 12 Insuring that farmers gain part of the marketing margin by selling through the farmers associations is a principal objective to increase the profitability of production and thereby afford to pay for higher input levels. The farmers’ organizations need to be responsible for storage, quality control, and acting as marketing agents for the farmers. The revenue increase from these marketing functions then helps cushion shocks from weather variability and other production-income shocks.

20

In some villages of Mali women discovered how to adjust their traditional processing methods to produce a good “tô.” Is it realistic to believe in the recovery (BC) of Grinkan? In the next section we review the processing operations and the science behind handling the consistency problem. The Consistency of the “Tô “ and Farmer Acceptance of Grinkan If women are unhappy with the basic dietary staple, the cultivar will be rejected. So let’s consider the processing process and the characteristics of Grinkan, which are different from the traditional sorghum:

Threshing-Separating the grain from the panicle. Important in producing clean cereal to keep the grain off the ground in the process.

Dehulling-Removing the bran (“son”) now optional with Grinkan processing

Milling into flour The traditional West African sorghums are Guineas with hard (vitrosity range from 2.5 to 3.5) grains. In contrast the Caudatum sorghums are soft and floury (vitrosity range from 0 to 2). Grinkan is 25% Guinea and 75% Caudatum. After dehulling the traditional sorghums are soaked overnight to facilitate the grinding of the hard grain in the milling process. When Grinkan is soaked overnight, the “tô” becomes too watery or loses consistency.13 So the overnight soaking needs to be eliminated and the dehulling and milling done in the same day. An alternative is to eliminate the dehulling process as is often done for cereals emphasizing the health food characteristics. In the Koutiala region of Mali where Grinkan has been grown over five years some women have developed one or the other of these two methods of preparation. Either one overcomes the consistency problem and now these women are happy with the “tô.” These techniques were taught by the extension service (DRA) in Sikasso in 2015. Hence, this consistency problem is now more of a communication problem than a technical barrier to Grinkan introduction. However, the damage to the marketing process had already been done. So a recovery period is necessary. Is this likely to happen? Markets and Grinkan Adoption

13 In more technical terms this soaking may reduce the amylose content of Grinkan which needs to be above 20.5% on a dry basis .A. Ndoye, Notes on Preparation of “Tô,”mimeo, June, 2015.

21 Grinkan still out-yields the local sorghums and outside of the less consistent “to” the taste of other traditional foods made from sorghum and millet (bouillie, couscous) was very much appreciated. Animals also favored the stalks and leaves over other sorghums (the ever green gene and the midrift gene). As women began adapting the processing methods either from their own experience or with extension help the demand situation has begun to shift. Evidence for the beginning of the shift, BC, with a movement back to Grinkan comes from the main wholesaler of Koutiala (Siriki Badjan Doumbia), who reports that his contacts in Bamako and Kayes are now requesting Grinkan whereas in previous years they did not want it at all or price discounted it. The most basic objection to the quality of Grinkan for everyday food has been resolved and this will also enable expansion of the local and national marketing. However, there is a difference between the technical resolution of a problem and more widespread knowledge leading to substantial increases in demand. So the communication question is how to get this information more generally known in urban and rural areas. The Malian extension service has begun this process in Sikasso. Moreover, a video showing how women resolved the processing problem in several villages, the readiness of a principal wholesaler to buy large quantities of uniform clean Grinkan to sell in the major markets and the resumption by certified seed producers of Grinkan, was produced in Bambara, French and English (video adaptation for UTube, 2016). This video in Bambara was shown to 19 farmers’ associations in the Koutiala region and appeared on Malian TV three times the first week of March 2016. Returning to other production problems. We have emphasized earlier that people14 and animals especially appreciate the taste of Grinkan. Not surprisingly field and storage insects also are really attracted to Grinkan. There are two alternative effective storage methods (Table 2). Both the PICs sacks and the polypropylene sacks combined with Phostoxin are very effective. Farmers complain of the high price for the PICs sacks of 1,200 fcfa/sack in fall 2015. Phostoxin needs to be undertaken in a storage facility not the house. But both provide good control so the problem is increasing the demand. In contrast one field insect problem, Cecidomyie, can cause serous losses and is difficult to control but not a problem every year.

Table 2. Second Generation Problems Largely Resolved

Constraints Status Future Activity 1. Non sticky “tô ” Two methods for processing No longer a technical problem. (Consistency) now becoming generally known Now a communication problem 2. Village women not Now a communication issue Continuing need for extension buying Grinkan in local markets 3. Wholesalers avoiding or Now a demand from Bamako Improve communication of price discounting and Kayes markets for Grinkan preparation techniques for Grinkan (S. B. Doumbia, Conversation, “to”in urban areas. Make October 2015) farmers’ associations aware of resolution of this problem and new markets for Grinkan.

14 Children break off pieces of Grinkan and suck on it like sugar cane

22

4. Storage Insects Two good solutions available: PICs sacks expensive; for PICs sacks; Phostoxin need separate storage Polypropylene sacks and facilities, not in the home Phostoxin 5.. Field insects Cecidomyie; Difficult but periodic; Aphids; Decis others Source: unpublished field interviews, October, 2015.

Continuing Agronomic and Institutional Issues in Introducing Grinkan or Similar Plant Types The mold problem reflects the basic tradeoff of more risks to get higher yields. In developed countries where Caudatums predominate, late rains are also a problem. There is now a range of choices of the Grinkan type with higher levels of Guinea including Grinkan Yerewolo (a hybrid) developed by IER), Tiandougou, and Tiandougou-koura (all three developed by IER), and Soumba. Since the open pollinated varieties, Tiandougou, and Tiandougou-koura, are more open and with harder polycarp they provide more resistance to mold as well as to storage insects. But this openness makes it difficult to compete with Grinkan in yields during years without the late rain problem

Table 3. Continuing Problems for Introducing Grinkan

Constraints Status Future 1. Mold from late rains Plant later; Avoid poor seed Stagger planting dates; quality by drying or renewing Continuing trade-off of less seed; Use more open variations tolerance to late rains for of Grinkan, as Grinkan greater yield potential from Yerewolo (hybrid), Tiandougou, openness and Tiandougou-koura (all three varieties developed by IER), 2. Renovate seed Increased farmer training in seed Seed sector experience at farm production provided by large level. Still need regular good scale testing and demonstration foundation seed production from trials focused on hybrids in 2015 the public sector (IER) or the by ICRISAT and IER. private sector. 3. Access to DAP and Public sector recommendations Need better response of public quality control of of micro-fertilization and NPK sector to farmers’ demands for fertilizers instead of DAP are not DAP. Quality control by public adequate. Have good field data sector of inputs especially of and farmer demand for our fertilizer and other chemicals (a recommendation of one sack of serious problem for NPK in DAP and one sack of Urea. An 2015). alternative is 2 sacks of NPK and one sack of Urea. 4. Scale of Farmers’ Larger size (50 to 150 members) Need consolidation of farmers’ Associations critical for performing associations into larger units. efficiently marketing functions, Training for storage, marketing storage, quality control, market functions and management. By quantity to exercise market getting more of the marketing power. margin increase profitability and accelerate diffusion.

23

Probabilities estimated by farmers for late rains and the occurrence of the mold problem range from 10 to 20%. Covering and sun drying are generally attempted but their effectiveness depends upon the extent of the late rainfall and resulting mold. The drying and the separation of the moldy cereals is especially important for seed producers. Seed renewal is an important issue in Mali. In 2015 ICRISAT and IER have been implementing widespread demonstration trials of hybrids and training farmer seed producers of hybrids in the Sikasso region. This training will also effect Grinkan seed producers making more frequent seed renewal an option at lower costs. It will be difficult for these hybrids especially the tall Pablo and Fada to outyield Grinkan. We already observed lodging in the field with both. Several other hybrids are related to Grinkan or with similar characteristics including Grinkan Yerewolo and Sewa (intermediate height).These have a much better chance of competing but both have more open heads than Grinkan so they are unlikely to out-yield Grinkan unless there is substantial hybrid vigor15 (Niaba Teme, sorghum breeder at IER, PhD at Texas A&M, conversation fall 2015.) The last two issues (Table 3) involve inputs, public policy, and institutional development to accelerate the introduction of Grinkan. During the scaling up program for Grinkan and associated technologies (2010-2013) it was difficult for most farmers’ associations to obtain DAP. Malian public policy has been promoting NPK by itself without the Urea and even micro-fertilization for sorghum. However, there is evidence of the yield and profitability of moderate levels of DAP and Urea (Coulibaly 2010; Coulibaly et al, 2013; Coulibaly et al., 2014 ; Ibrahim, 2014; Ibrahim et al., 2012 ; Ouendeba et al., 2013).16 Clearly getting the input markets functioning well and responding to farmer demands for DAP are important concerns to benefit both low income farmers and consumers.

The farmers’ associations need a minimum size to arrange for their own storage facilities and to be a force in the market by selling large enough quantities of clean cereals to interest wholesalers or food processors. Yet in the villages due to personal rivalries or family connections we often observed two to five different farmers’ organizations. Once attaining sufficient size these farmers’ associations can benefit from training in marketing, management and transparent management behavior.

15 Note that hybrid seeds must be bought annually and generally are substantially more expensive than open pollinated cultivars such as Grinkan. 16 The substitution of NPK for DAP or even worse micro fertilization have both been shown to substantially reduce yields compared with the above two sack treatment of DAP and Urea. In 2015 a new low in fertilizer distribution was achieved with the poor quality of the NPK imported from Eastern Europe (Farmer interviews, October 2015).Maintaining farm input quality is a critically important public sector function, which is frequently difficult in developing countries.

24 Sorghum in the rotation of the cotton system

White gold has dictated most of the terms of what is done in the cotton zone. CMDT will provide fertilizer for cereals only if certain yield goals are achieved in cotton and the reimbursement must be in cotton. There is pressure from CMDT to use all the farm available organic fertilizer on cotton. Before the 40% price increase in 2011 for cotton, area and production of cotton in the Koutiala area had been decreasing. Also prior to 2011 there was much discussion of the importance of cereals and diversification. In 2011 and thereafter for several years there was a rejuvenation of cotton but now prices and production are falling again (field interviews with farmers in the spring of 2015). In the 21st Century maize has been cheaper than sorghum and millet. Moreover, the availability of fresh maize for roasting during the hungry season (“soudure”) has established itself as essential during the scarcest time of the year for food availability. In the cotton zone the lower cost maize has become the cereal of choice in the cotton zone for “to.” Also many new maize varieties and hybrids have come out of the international centers (CIMMYT and IITA) and national agricultural research centers. Consequently, maize is the predominant choice to follow cotton and is being fertilized with 2 to 5 sacks of fertilizer. The maize varieties with three sacks of fertilizer (1 sack of NPK and 2 sacks of Urea) gave yields of 1.5 to 2.5 tons/ha (J. Coulibaly, 2015, p. 57; also I. Konate, Director of the IICEM scaling up project in the Koutiala region, conversation, December 2015).Grinkan yields can compete with maize yields and the price of sorghum is generally 20 % higher than that of maize. With the new maize hybrids and five sacks of fertilizer yields can be 4 to 5 tons and have even been reported higher. There are also a number of sorghum hybrids being field tested and seed being produced, some of which like Sewa come from the IER breeder of Grinkan and have many similar characteristics to Grinkan. So cotton continues to decline again as Mali has not yet even begun to incorporate the Bt gene into locally adapted cultivars and soil fertility problems are getting worse in the marginal cotton areas such as Koutiala. Hence, we expect renewed interest in cereal production here with more Grinkan producers following cotton in the rotation or following a heavily fertilized maize. In either case Grinkan or a Grinkan type will do well in increasing production and area. With more communication about the yields, ‘to” preparation, tastes and preference for animals this 17 demand for Grinkan or Grinkan types will accelerate.

17 We would also expect increasing diversification in the cotton zone with increased production of soybeans (domestic demand), sesame (foreign demand), and millet (response to increasing soil fertility problems). 25 Conclusions:

New technologies have second generation problems. First generation characteristics of a new cultivar, Grinkan, include higher yields, better tastes, and a preference of animals for the crop residuals. This is a substantial achievement for a breeder to attain all of these characteristics. However, second generation problems arose including the need to change preparation processes of “to” and improved control of storage insects. If neither extension nor research focus on the continuity of investment and extension in resolving new problems, the successes of the first generation are not maintained. It is expected to be more efficient to build upon these first generation successes by investing research and extension in the continuing diffusion of these new cultivars. With Grinkan this includes disseminating the preparation methods for producing good “to” and utilizing better storage techniques.18Unfortunately, the principal objective of national research and extension agencies need to concentrate more on the next source of outside funding rather than on maintaining yield gains with strategic investments.

Grinkan is an impressive innovation from both the perspective of potential yield increases and tastes. In spite of the recent focus on hybrids in Mali, Grinkan is expected to out-yield them with the possible exception of Sewa (conversations with farmers harvesting the hybrids in 2015 in trials and demonstrations), a Grinkan type cultivar. So we are optimistic about the recovery and expansion of Grinkan and/or Grinkan types.

The technical problems are largely resolved especially the different processing techniques to avoid non sticky “tô.” The diffusion of knowledge about the “to” preparation depends upon communication and extension but has already begun. This will lead to market expansion for Grinkan in local and regional/national markets.

The technical solutions to storage insect control are also well known now. We need to understand now why farmers are not adopting them (PICs sacks or polypropylene sacks combined with Phostoxin) more rapidly. Public concern with the functioning of the input markets especially the seed and fertilizer markets are essential for accelerating diffusion. Regular seed renovation and the availability of DAP make a big difference in yields of Grinkan. Grinkan can be in the 1.5 to 3 ton range with these inputs. NPK is not equivalent to DAP and DAP is important for sorghum and millet not just rice. Another critical factor is developing the management potential of the farmers’ associations. They need to be finding financing for storage construction or better use of present storage facilities, increasing the supply of clean sorghum by controlling quality at the storage level, and becoming better at the marketing and management functions.

18 IER produced a harder polycarp sorghum to resist storage insects but the cost was not being able to out- yield Grinkan with these new cultivars. Not everything requires a breeding solution

26 There is a particular problem in developing countries due to the financial problems of the national extension and research services and the availability of many NGOs often with funding to support activities that these services cannot support. As a consequence new technologies that experience problems are often rapidly discarded once the period of outside support ends as village farmers move on to the next program of another funding agency. This is unfortunate because the identification of second generation problems is a natural phenomenon even of a successful innovation as biological systems are dynamic and have many interactions and changing requirements. So breeders are pushed to continually start over again when they could be building upon previous successes. Recommendations: 1. Grinkan is an excellent, high yielding cultivar with good tastes for people and of the stover for animals. Develop mechanisms to encourage extension and research to concentrate more on second generation problems of successful technology. Second generation problems always arise in a biological system. Specific suggestions below (2 and 3). 2. Accelerate the training of the preparation techniques for “to” of the soft cultivars. These techniques are known in several villages and the extension service already gave a course in 2015. 3. Increase responsiveness of the input supply sector to farmers’ demands for DAP and Urea on sorghum and millet 4. Coordinate better between agencies so they are not working at cross purposes especially in increasing the demand for certified seed and moderate fertilization with DAP and Urea. Different fertilization treatments are good competition. 5. More demonstrations of the PICs sacks for cereal storage. Some pathology research on the maximum humidity level of the sorghum in PICs sacks for seed production. 6. Increase marketing and business management training of the farmers’ associations to perform better as economic and marketing institutions. References: Abdoulaye, T., J.H. Sanders and B. Ouendeba, 2008. Evaluation of Sorghum and Millet Technology and Market Strategy Introduction: 2006-2007 Crop year, INTSORMIL Bulletin No 8 Lincoln Ne 22 pages Abdoulaye, T., and J.H. Sanders, 2006. New Technologies, Marketing Strategies and Public Food Policies for Traditional Food Crops: Millet in Niger, Agricultural Systems 90:279-292. Baquedano, F., J.H. Sanders, and J. Vitale, 2010. Increasing Incomes of Malian Farmers: Is Elimination of US Subsides the Only Solution? Agricultural Systems, 103: 418-432 Coulibaly, J., 2010. Evaluation des Technologies de Production et de Commercialisation du Sorgho et du Mil dans le Cadre du Projet IER-INTSORMIL/Mali, Campagne Agricola 2008- 2009. Bulletin IER-INTSORMIL n° 10, Purdue University, Department of Agricultural Economics, 36 pages.

27 Coulibaly, J. J.H. Sanders, P. Preckel, and T. Baker, 2015 “Will Cotton Make a Comeback in West Africa-Mali?” Agricultural Economics, 46:53-67 Coulibaly, J., Gautam Kumaraswamy and John H. Sanders, 2013.Economic Impact of Sorghum and Millet Technologies in Mali Agricultural Campaign, 2010-11, Bulletin IER- INTSORMIL n° 11, Purdue University, Department of Agricultural Economics, 31 pages.

Ibrahim, A. D., John H. Sanders, and Botorou Ouendeba, 2012.The Economic Impact of New Sorghum and Millet Technology Adoption in Niger: Performance and Challenges, 2010- 2011 Cropping Year, INTSORMIL Bulletin No. 12, Department of Agricultural Economics, Purdue University, 42 pages.

Ibrahim, A., 2014. “The Economics of an Extension Program in Maradi, Niger,” unpublished PhD thesis, Department of Agricultural Economics, Purdue University,

Ouendeba, B.,T. Niaba, and J. H. Sanders, Mai 2013, Fiche de Production et de Commercialisation de Sorgho Grinkan, Production-Marketing Project of INTSORMIL, Purdue University, Department of Agricultural Economics, West Lafayette, IN Video, 2016. “To” Preparation Techniques for Grinkan and New Markets and Seed Producers,” YouTube version in English. Bambara and French versions available from the authors. https://www.youtube.com/watch?v=qIYmlAZXtk4

28

TABLE A-1: Cultivated Areas, production and price of Grinkan in the villages Interviewed in KOUTIALA (MALI)

Years VILLAGES INTERVIWE 2008 - 2009 - 2010 - 2011 - 2012 - 2013 - 2014 2014 - Fertilizer D & Data 2009 2010 2011 2012 2013 2015 Used collected. 1 Garasso DAP /urea area ha 50 150 117 10 with Production t 75 206 136 11.5 Price Frs/kg - - - - OPs or NPK/ urea Buyer - - - - 2 Sougoumba (Oumarbougou 13 19 15 100kg de ) NPK Area ha 10.4 13.3 15 Production t - - - and 50kg Price Frs/kg - - - urea Buyer 3 Kaniko 100kg de Area ha 100 50 50 50 50 NPK and Production t 144.3 90.1 51.4 43.7 98.8 50kg Price Frs/kg - - - - - Buyer - - - - - urea 4 Klé Area ha 50 25 19 19 15 DAP, Production t 60 17.5 16.1 13.3 10.5 NPK and Price Frs/kg - - - - - urea Buyer - - - - - 5 Sangaba Area ha 51 57 53 50kg of Production t 61.2 45.6 13 DAP and Price Frs/kg - - - 50kg urea Buyer - - - 6 Karangana Area ha 60 115 sacs (DAP, Production t 72 d’engrais urea and Price Frs/kg - donnés à NPK) Buyer - l’OP 7 N’Golonianass o 0.5 ha / 40 Area ha 0t NPK and Production t 35.2 (seed did urea Price Frs/kg 100 not Buyer C germinate)

29

8 Koloto Area ha 60 55 16 5 2 Production t 63 63.2 16 4t 700k NPK and urea Price Frs/kg - - 100 125 130 Buyer - - B A A 9 Ngounthina Area ha 10 0.5 10 10 12 1 ha NPK and Qtty sold t - - 3.232 7 2.430 - urea Price Frs/kg 115 300 300 135 750 120 Buyer A A A A A A 10 Kapala Tiandougou Soumba Area ha 25 5 Qtty sold t 2 3 18 - - NPK and Price Frs/kg 1 5.755 8.470 135 135 urea Buyer - 190 140 A A A A A 11 NTorlani Area ha 126 81 120 NPK and Qtty sold t 7 8.5 1.1 urea Price Frs/kg 155 140 135 Buyer A A A 12 Molobola

Area ha DAP or 67 105 177 214 236 Qtty sold t NPK and Price Frs/kg - 8.981 18.562 0.712 2.897 urea Buyer 135 190 135 135 125 A A A A A 13 Faraola Area ha Qtty sold t 5 3 4 4 NPK and Price Frs/kg - - - - urea Buyer 125 135 140 125 A A A A 14 Nangorola Area ha Qtty sold t 45 100 60 5 NPK and Price Frs/kg 4 3 2.5 3 urea Buyer 125 150 125 120 D D D D

15 NTobougou NPK and

Area ha urea in 35 25 18 10 Qtty sold t Price Frs/kg 13,35 13 22.47 14.7 2011/201 Buyer 190 135 125 135 A A A A 2 16 NTola Area ha Qtty sold t 60 60 23 18 NPK and Price Frs/kg 13.341 5.048 7.148 14.043 urea Buyer 190 135 130 132.5 A A A A

30

17 Togoba Area ha Qtty sold t 71 59.5 50 55 43 NPK and Price Frs/kg - 1,8 - - - urea Buyer - 190 135 130 132.5 - A A A A 18 Mamarila Area ha 6 6 9 6 Qtty sold t 4 4,4 5 5,4 Not often Price Frs/kg 200 190 130 130 Buyer A A A A 19 MPessoba Area ha 25 21.5 20 12 10 NPK and Qtty sold t - - - - - urea Price Frs/kg 125 200 135 135 135 Buyer D D D D D 20 Signe Area ha 5 3.5 1 DAP or Qtty sold t 1 1 - NPK and Price Frs/kg 80 180 - urea Buyer D D - 21 Zanzoni Area ha 50 16.5 19.5 20.5 19 7 Qtty sold t 11 3,8 3,06 1,827 1,3 1 NPK and Price Frs/kg 135 120 190 135 110 135 urea Buyer A - D A - D A - D A - D A - D A - D 22 Baramba Area ha 4.75 3 1 1 NPK and Qtty sold t - - - - urea Price Frs/kg 120 150 - - Buyer D D - - 23 Zangasso Area ha Qtty sold t 10 6 38 42 55 DAP and Price Frs/kg - 2.16 5.213 4.542 4.607 urea Buyer 110 190 132.5 135 135 A A A A A 24 Sinkolo NPK and Area ha 5 3.75 urea / or Production t - - No Price Frs/kg - 135 Buyer - A-D fertilizer

A. Mr. Siriki Bodjan DOUMBIA, Wholesaler at Koutiala B. Mr. Mamoutou BERTHE, Wholesaler at Koutiala C. Mr. Madou COULIBALY, Collector D. Village Collectors

N.B. : Quantity sold not indicated when Ops not involved

Source: unpublished data collected in Koutiala, Mali, October/November 2015

31 APPENDIX 2: Evolution of Grinkan yields for 33 interviewed villages in Koutiala

Evolution of Grinkan yields for 33 interviewed villages in Koutiala

1500 1600 1420 1427.1 1357.63 1359 1318.4 1400 1283 1500 1420 1386.7 1200 1307.1

1000 1206.15 1161.4 1191.2 Yields 800 in Yields average ajusted (in kg/ha) 600 Kg/ha Yields average (in kg/ha) 400

200

0 2008/2009 2009/2010 2010/2011 2011 /2012 2012 /2013 2013/2014 2014/2015 Years Source: unpublished data from village interviews in Koutiala, fall 2015.

In the adjusted yield measurement we omitted the extremely low yields on the grounds that they did not follow recommendations for one reason or another. Or the higher yields could be considered as the potential yield average. The low yields may also indicate that the farmer was not motivated sufficiently to put the sorghum in a good area such as following cotton or to use the correct fertilizer dose. We used in the text the actual means from the village observations. Note that there were a limited number of yield observations taken from the farmers who came to the village meeting and could recall the yield data for Grinkan over the period. See the number of observations for each year below.

Years Yields average : Number of Yields average adjusted observations considering yields over 800 kg/ha 2008 - 2009 1 1 2009 - 2010 2 2 2010 - 2011 12 11 2011 - 2012 19 15 2012 - 2013 26 22 2013 - 2014 21 18 2014 - 2015 18 17

32 N.B.: Yields less than 800 kg/ha are considered to be explained by:  Inappropriate fertilization  Excess of water;  Poor soils ;  Objective of using Grinkan only for animal feed.

33 Sorghum Production Systems in Burkina Faso Abstract In 2010, Grinkan was introduced into the cotton region of Burkina Faso in a four year pilot project. Over the period 2010 to 2014 high yields were achieved but there were two years of late rainfall especially affecting farmers, who did not plant later as recommended. There was no scaling up program1 in Burkina in contrast with Grinkan in Mali. Moreover, there was an especially rapid decline once the pilot project terminated. So below we consider first this collapse in Burkina as pilot project farmers were intensively interviewed about this. We explain the technical responses to the series of factors especially processing and the marketing once the processing problems were known. Then we consider the other components of technology introduction in Burkina. The principal factors as in Mali were the “to’ quality and the resulting loss of markets. This issue is treated in detail in the previous paper in this bulletin. This is a story principally about Grinkan in Burkina but also about the other components of program introduction and activities in the cotton zone of Burkina. Rejection of Grinkan With the successes of Grinkan in Mali the prospects looked good for Grinkan in a very similar agro-climatic region, the Bobo region. 2010 was a year of demonstrating Grinkan in a nearby plot while farmers adopted the technology package but put it mainly in their locally improved Guinea cultivar Saraisso 11.2 In 2011 and 2012 pilot project farmers began switching to Grinkan. 2012 turned out to be an excellent year for high yields of Grinkan (Picture 1). With early cessation of rain in 2013 Grinkan yields were adversely affected.

In the local markets, where farmers sell small quantities of sorghum to get a little cash on market day and make their small purchases, village women started avoiding Grinkan because of the “to” quality. 3 Moreover, farmers were unhappy with the price discounts for Grinkan from merchants so this village effect spread upward to the larger markets.

Table 1 documents the brief rise and fall of Grinkan in our pilot project sites in the cotton zone of Burkina. 2012-13 and 2013-14 were our last two years and already there was a 50% decline in the last year before Grinkan essentially disappeared one year after the project.

1 The Ministry of Agriculture did make substantial quantity of seed available to the extension service for distribution in the cotton zone in 2013. 2 The cultivars released by the Faroka-Ba station of INERA are all named Saraisso and given a number. 3 To the women this meant a non-sticky “to“. To food scientists this is the consistency problem. See the Mali paper for more details on processing and overcoming this “to” preparation problem.

34 Picture 1. New sorghum production system Grinkan, Kouakole, Burkina Faso November 2012.

Table 1. Introduction of Grinkan in the Bobo Region

Associations 2012 - 2013 2013 - 2014 2014 - 2015 Kouakole Grinkan : 50 ha Grinkan : 40 ha Grinkan: 2 ha Sariasso 11+ Sariasso 11+ local= Sariasso 11+ local= local=20 ha 30 ha 68 ha Missidougou Grinkan : 50 ha Grinkan : 20 ha Grinkan: 0 ha Sariasso + local=20 Sariasso 11+ locale= Sariasso 11+ local= ha 50 ha 70 ha Soukouralaye Grinkan : 2 ha Grinkan : 1.5 ha Grinkan: 0 ha Sariasso 11+ Sariasso 11+ local= Sariasso 11+ local= local=58 ha 58.5 ha 60 ha

35 The reasons given by Kouakole and Lena farmers for dropping Grinkan can be summarized: 1) “To” quality. Women were unhappy with the “to” quality. “To” is supposed to have a certain stickiness and the Grinkan “to” did not have this.4 For food scientists this is consistency and is a trait well known and measured by nutritionists so can be used in sorghum and millet evaluation. Normally sorghum is not tested for consistency and consistency has been a leading reason for the failure to adopt cultivars that were actively promoted such as NAD 1, L30 and SRN 39. NAD 1 was a highly touted hybrid. SRN 39 has Striga resistance and L30 had high yields. All three were promoted in Niger and then just disappeared (B. Ouendeba, conversation, 2015). Food scientists can do consistency tests so the “to” quality could be an important qualification criteria along with yields, 5 specific resistances, and taste. 2) Markets. In the small sorghum sales at local markets that farmers use to finance other necessary purchases the women then refused to buy Grinkan. Then this moved up the marketing chain to regional and urban markets with price discounts or lack of interest of buyers. Both 1) and 2) here are concerned with “to” quality. 3) Susceptibility to the mold-head bug complex from late rains. The compact heads of Caudatum enable high potential yields. Unfortunately they do not dry quickly hence late rains on the mature heads lead to yield declines and poor seed germination (if not stored very dry)6 in the succeeding season. Thorough drying in late rainfall years would be expected to reduce the germination problems. The short season Grinkan needs to be planted late in the cotton zone. In doing that farmers can either plant too late or the farmers may have already allocated their best land to other crops. 4) Insects. Grinkan is generally considered to taste better than most sorghums for humans and animals but it also appeals to insects. Seedlings are attacked by insects (“chenilles”). During the crop season when the heads are formed, the cecidomaye can lay their eggs and the larvae later destroy the grain. This cecidomaye problem does not occur every year but is a serious problem all over the world. Storage insects can be handled with the three layered PICs sacks but these are more expensive than the usual sacks.7 Polypropylene sacks with Phostoxin also give good control of insects but Phostoxin is dangerous. Further development of Grinkan types with harder polycarp for more insect resistance in storage has already taken place in IER of Mali. 5) Increased agronomic labor requirements. Farmers claim that Grinkan is much more sensitive to the lack of sufficient weeding in the early stages than Saraisso 11. Thinning

4 In the Bobo region substantial advances have been made in maize production and maize is consistently cheaper than either millet or sorghum and is now the predominant cereal used for “to” in the cotton zone. For the other traditional products, “bouillie” (a thick porridge), couscous and “zoum-koum” (a thin porridge) there is a preference for Grinkan. For forage, animals like Grinkan very much. 5 The more different characteristics the breeder insists on having in his new cultivars, the less likely is he to produce a new cultivar. This is one reason why improving primary staple productivity is difficult. 6 The PICs sacks for storage insect control make deterioration hence germination worse if the sorghum is not adequately dried. 7 The evidence is mixed on the use of PICs sacks for seed storage. Wholesalers and seed dealers use the polypropylene sacks with Phostoxin

36 (not traditionally done by farmers), density, and time of planting are all important with Grinkan. But this is typical of a higher yielding cultivar to be more demanding of soil fertility and of improved agronomic practices. It is a necessary labor investment for high cereal yields. 6) Weight per volume. Women also complain about Grinkan weighing less for the volume than Saraisso 11. For women this appears to be more work for the mortar and pestle operation. There were adverse rainfall conditions late in the season in 2013 and a late commencement of rain in 2014 and this weather factor was the principal reason given for dropping Grinkan in Kouakole. In Lena farmers complained about the quality of the “to” and asserted that there were problems in marketing associated with this and that the merchants preferred Saraisso 11. In summary, the quality of the “to” and the marketing problem seem to be the two most 8 important factors presently. Yields and Consistency of Grinkan Grinkan has out yielded local sorghums by 50 to 100%.Local sorghums are tall and have loose grain formation in the heads. Hence their maximum yield potential is expected to be about 1.2 tons/ha with the expectation that following cotton in the rotation yields would be 800 kg/ha to 1 ton/ha. Following fertilizer and other recommendations for normal and good rainfall years Grinkan would be expected to give average yields of 1.5 ton/ha with very good farmers getting 2 9 to 3 tons. So as is obvious from the farmer feedback in the Bobo region that Grinkan adoption and sales (1 and 2 above) were reversed by this problem of the "to" quality. Women discovered this problem of “to” and then that affected the small sales in the 10 villages and ultimately even the regional and urban sales. Are Grinkan and Caudatum types dead in Burkina Faso? The decline of the pilot program was rapid in Burkina with only two farmers (would be seed producers) still producing Grinkan in 2014. NAFASO, a private seed company in Bobo with which AGRA (donor agency) had been working, stopped producing Grinkan in 2014 and had no plans to do so in 2015 because they could not obtain foundation seed of Grinkan from the INERA research stations. The extension program, which provided Grinkan and fertilizer in 2013

8 On consumption of “to” maize is easier to prepare because of various mills set up for maize turning this cotton regions into maize centers. Maize also gets larger allocations of the fertilizer subsidy than sorghum. Subsidized fertilizer is 12,500 fcfa/sack plus transportation costs while the market price in 2015 was 17,500 fcfa/sack. One of our two village sites received the subsidized fertilizer and the other did not. 9 Grinkan needs moderate fertilization (one sack of DAP and one sack of Urea or two sacks of NPK and one sack of Urea). Also especially in the lower fertility regions organic fertilizer is necessary to improve the structure and better retain water and nutrients. 10 An adaptation of processing eliminating the overnight soaking has begun to be practiced in several villages of Mali and now taught by the extension service in Sikasso. Women report that this now gives “to” superior to the locals. One Koutiala wholesaler has offered to buy as much uniform clean Grinkan as is produced. Certified seed producers are re-engaging in the production of Grinkan.

37 for the Bobo region in a special program of the Ministry of Agriculture, had no data to report on performance. There are simple methods to avoid the non-sticky “to” but they just have to be communicated to women in the village. An alternative is that collaborative breeding programs with food scientists doing consistency checks could respond to the consistency problem. We are even optimistic about a comeback for Grinkan itself. In Kouakole the president of the farmers’ association did several land preparations for better water retention and Grinkan yields were 2 t/ha. We asked other farmers for their yields of Saraisso 11 in 2014 and got a range of 1 ton/ha (2 observations) s to 1.3 tons (2 observations) for Saraisso 11. So the yield advantage of Grinkan has been clearly demonstrated. Presently, some Grinkan characteristics are being developed for future Sariasso cultivars (Hamido Traore, DG of INERA,conversation Dec 7, 2015). Meanwhile we need to evaluate the other impacts of the pilot project in Burkina Faso. Village Level Impact of our Project: The principal impacts of the project were: 1) The acceptance of the need to fertilize directly the sorghum with a moderate level of inorganic fertilizers. In the higher rainfall regions sorghum often follows cotton or maize and takes advantage of residual effects of cotton fertilizer. Outside of these regions where rainfall limitations were more pressing it was argued that either organic fertilization or micro-fertilization were sufficient. Moreover, the conventional wisdom was that there would be little response to fertilization or farmers would not fertilize or it would not be profitable to use inorganic fertilizer. Farmers in the lower rainfall region (Central Plateau) are now convinced of the need for moderate levels of inorganic fertilizers directly on sorghum. Farmers in the cotton region (Bobo area) are more accustomed to inorganic fertilizer so putting it on sorghum was not a big change in practices as in the drier regions. The yield effects and profitability of sorghum have been documented at the farm level (see T. Abdoulaye et al, 2008; F. Baquedano et al, 2009; J. Coulibaly, 2010; and J. Coulibaly et al., 2013). 2) Market strategy. The concept of storing and selling later to avoid the price collapse at harvest was firmly accepted with storage facilities being widely built with part of the labor being supplied by farmers’ associations. Also the need to search for markets higher in the marketing chain and even outside the region became part of the activities of the farmers’ associations. They just are not very good at negotiating and identifying these markets as yet. But this is a skill that has a learning by doing effect and the farmers’ associations understand and appreciate the concept. There was not as much emphasis on clean cereal as with millet as the major market for food processors is millet and in the cotton zone maize is increasingly relied on for “to”, the basic food staple in rural areas. 3) Development of the farmers’ associations and the revolving funds. A basic objective was to create strong farmers’ organizations that could accelerate diffusion, organize the fertilizer purchases, identify markets for the cereals, sell opportunely, and finally help members get access to bank credit. The best indicators of the evolution of the farmers’

38

associations are the repayment rates on the input credits for fertilizer purchases. With some initial difficulties in Lena in personnel selection both the Kouakole and the Lena sites are now experiencing 100% repayment rates for the input credits financed with the revolving funds. When we started working with Kouakole in 2010 we financed 2.25 million cfa in inputs for 50 members to be paid back to the farmers’ association in sorghum at harvest. The farmers’ association then would store and sell before the next planting season thereby both creating a rotating fund and avoiding the post-harvest price collapse. In Febuary 2015 in Kouakole there were 220 active members (defined as those receiving input credits from the revolving fund to buy inorganic fertilizers) and the revolving fund had increased to 13.4 million cfa. So no bank financing is required and the farmers’ association has been increasing in area and farmer membership. Kouakole is presently an independent organization servicing its farmers. In 2014 OXFAM an English NGO, built a cement storage facility (50 ton capacity) for Kouakole. So other agencies have facilitated activities here, which is also a sign of a functioning farmers’ association. Lena’s history was very different going through periods of poor reimbursement and throwing out non-reimbursing members. We began Lena with 50 members and putting 2.25 million cfa into a revolving fund in 2010.11 In Febuary 2015 there were 52 members and a revolving fund of 2.5 million cfa. The rotating fund and membership have not expanded but they still exist. Moreover, with their bank accounts the Lena farmers’ association has been able to get loans for warrantage (loans with the grain in the storage as the guarantee for the loans) during the last four years. OXFAM built for them a new storage facility in 2014. The farmers’ associations The main program accomplishments to date were to convince farmers to fertilize sorghum with two to three sacks, a step upward from the zero or micro fertilization recommendations. Our recommendation was one sack of DAP and one sack of Urea or two sacks of NPK and one sack of Urea. There was difficulty finding DAP in 2013 and 2014 and sometimes even Urea so farmers’ substituted NPK. So there was good following of the fertilizer recommendations in general given the difficulty in finding DAP. The first four sites (different groups within the two farmers’ associations) show the cotton system here with maize being more important in area terms than sorghum and a variety of sorghums being produced. Some farmers’ groups received the subsidized fertilizer and some did not. This apparently depends upon connections with the extension service. The difference of 5,000 cfa/50 kg sack is large between subsidized and unsubsidized fertilizers. The prices received for the sorghum are very low in the Bobo region as compared with the Central Plateau in Burkina (Tables A-! and A-2). Moreover there was a discount for the private

11 The program paid for fertilizer and seed. The farmers agreed to repay the loan to the farmers’ associations in kind before the next planting season.

39 sale of Grinkan at 9,000 cfa/sack while the average prices of sorghum sold by the association in 12 the program were 11,000 to 12,500 in 2013 and 2014. There are many crops in this cotton system. The increased world prices for cotton in 2010 and the introduction of Bt cotton in Burkina undoubtedly influenced the decline in interest for Grinkan by providing more cash income and encouraging area shifts to more cotton (Coulibaly, 2015). Moving to the Central Plateau in Korsimoro (Table A-2) some farmers were already using an improved Caudatum, ICSV 1040 from ICRISAT. They were using organic fertilizer at reasonably high levels for semi-arid regions given their lower animal density and plant residue availability. Moreover, the use of a traditional water retention device similar to tied ridges, the “zai”, was widespread.13 The local sorghum Kapelga, a Guinea, was low yielding but farmers appreciated very much the taste and the price premium. There was also a great variety of crops and a much smaller total area in crops than in the cotton zone. The “zai”and the production (compost heaps) and delivery of organic fertilizer are very labor intensive. There is also more pressure on the available crop land in the Central Plateau than in the cotton zone with its higher rainfall and more fertile soils. Areas involved in the farmers’ association sorghum program were smaller here than in the cotton zone and membership in the farmers’ associations stayed constant over the period in Korsimoro and Pissili. The price per sack of 18,000 fcfa/sack was 5,500 to 7,000 fcfa/sack higher on the Central Plateau than in the cotton zone. Production conditions are much harsher on the Central Plateau. Why is there not more sorghum moving from the cotton zone to the Central Plateau? Already the farmers’ association of Kouakole produced 800 sacks or 80 tons in 2014. Conclusions: Conceptually (using a Caudatum to get high yields) and based upon field performance, Grinkan is an excellent innovation. Grinkan follows the physiological innovations of the ‘50s that led to the successes of the Green Revolution in rice and wheat. The basic idea is to convert the plant so that it produces more grain and less stalk and leaves. Secondly, the plant is built with a sturdier stalk and shorter to take higher fertilizer levels without lodging. So we demonstrated the yield potential of an open pollinated cross between Guinea and Caudatum with high yield potential and excellent taste characteristics. However, the adaptation of “to” processing was not generally known so women rejected the new cultivar in the villages. This led to the rejection of Grinkan on the village markets and ultimately on the larger markets. Now it is known how to make good “to” from Grinkan and this information could be widely 14 communicated. This communication process has begun in Mali. The problems that we documented need to be considered as second generation problems because Grinkan is an outstanding new cultivar. Grinkan progeny or sister lines have already been developed in Mali with a harder endosperm for greater resistance to storage insects. A

12 The government program (SONEGESS) to acquire grain from farmers and distribute it to the malnourished paid 15,000 cfa/sack in 2013 in Lena. 13 This technique consists of digging small holes in the field and then putting the organic fertilizer and the seed in the holes. 14 This adaptation of processing for maize “to” is also necessary for some of the new maize hybrids.

40 longer cycle progeny would be able to be planted earlier without as much risk from late season rains. The Burkina experiment stations (in both Faroka-Ba and Fada) have incorporated Grinkan into their breeding activity. Grinkan characteristics have been included in cultivars that will come out with the Sariasso station designation but the main emphasis of the national sorghum breeding program as in Mali is in producing hybrids(Hamido Traore, DG of INERA, Dec 7, 2015). This sorghum hybrid emphasis has been the focus of the national programs in Mali and Burkina Faso with AGRA financing and ICRISAT research support over the last five years. Moreover, in the cotton zone of Burkina the focus of development attention is on the new Bt cotton and the high yields of maize cultivars and hybrids. The other components of our pilot program especially the introduction of new marketing practices and the development of the farmers’ associations are now known. However, a successful new technology introduction on farmers’ fields drives this process of improving storage and marketing. Many of these associations have split into smaller units. So even if the funding and several associations remain, they need this dynamic incentive from the new agricultural technology. In contrast with Mali the farmers’ associations have not changed very much their level of the marketing chain to which they sell except to handle directly the collection functions. Increasing interest in the marketing activity will occur once the farmer’ associations get back to a higher yielding cultivar with Grinkan characteristics including intermediate height, compact head, study stalk, high yields and good tastes for humans and animals. Recommendations 1. With Bt cotton and new maize open pollinated cultivars and hybrids it is difficult to get farmers’ attention on the potential of new sorghum technology. As cotton prices fall again and the successes of Grinkan in Niger and Mali are known, reintroduce Grinkan or similar cultivars in demonstration trials in Burkina Faso. 2. With the demonstration trials teach the processing techniques for ‘to” to women in the village. 3. Then move to scaling up by focusing on the input markets for DAP, Urea, certified seed and input credits. 4. Maintain pilot projects to assure that the costs of the technology erosion resulting from the scaling up process are estimated and communicated. 5. Increase attention to developing farmers’ associations of sufficient size and managerial capacity to play an efficient role in higher levels of the marketing chain

41

Table A-1. Characteristics of the Participating Farmers’ Associations in the Bobo (Cotton Zone) Region, 2012-2015

Zones : Cotton zone 2012-2013 2013-2014 2014-2015 Traditional Production System 1. Kouakale - Area (ha) 80 100 120 Maize : 1.5 ha - First year 2011 - - Sésame : 1 ha - Members 200 200 220 Red sorghum 1 ha - Reimbursement 483 sacks 500 sacks 600 sacks Sariasso 11 :2,5 (5 sacks/ha) - ha - Sorghum Price 12,500 f en Mai 12000 f en Mai - Cowpeas: 0.25 a (ASS.) ha - Inorganic 3 sacks NPK 3 sacks NPK 3 sacks NPK Fertilizer/ha 20-10- 05; 1 20-10- 05; 20-10- 05;1 a (unsubsidized) NPK=17500 (1NPK=17500 NPK=17500f f) 7.5 t/ha - Organic Fertilizer/ha

2. Missibougou (part of Kouakale) - Area (ha) 25 35 40 Cotton : 3.5 ha - First year 2008 - - Maize : 2 ha - Members 35 - - Millet : 2 ha - Reimbursement 125 sacks (5 175 sacks 200 sacks Sariasso 11 :1 ha - sacks/ha) - Price 12,500 fcfa 12,000 fcfa - Inorganic 2 sacks NPK fertilizer/ha (un- (20N 10P 5K) subsidized) - 5.4 tons - Organic fertilizer/ha

3. Lena Monemeta Cotton : 4 ha - Area (ha) 50 52 52 Maize : 5 ha - First year 2009 - - Sesame : 2 ha - Members 54 with 4 - - Red sorghum 2 ha - Reimbursement women 208 (4 208 sacs Sariasso 11: 1 ha 200 sacks (4 sack/ha) - - Price sacks/ha) b - Inorganic Fertilizers 15,000 11,000 fcfa 2NPK (14- /ha 2NPK (14-23- 2NPK (14-23- 23-14 ) + 1 14 ) + 1 Urea 14 ) + 1 Urea Urea

42

(subsidized: Urea=12500 f and NPK=13500f) 0 (but planted - Organic fertilizer/ha after cotton)

4. Lena : Korodiadi and Relwindi - Area (ha) 8 8 10 Cotton : 3 ha - First Year 2012 - - Maize : 2.5 ha - Members 8 - - Millet : 1 ha - Reimbursement 32 sacks 32 sacs 40 sacs Sariasso 11 : 1 ha - Price assoc. 4sac/ha) 11,000-fcfa - Peanuts: 0.25 ha sorghum 15,000-fcfab 9,000 fcfa - Cowpeas : 0.25 - Price individual 9,000 fcfa 2 NPK (14-23- 2 NPK (14- ha (sorg) 2 NPK (14-23- 14) + 1 Urea 23-14) + 1 - Inorganic 14)+ 1 Urea Urea fertilizer/ha (subsidized) 5.4 tons

- Organic fertilizer/ha

a. In Kouakale, the unsubsidized NPK (20-10-05) was: 17,500 fcfa. In Lena farmers were able to get the subsidized NPK (14-23-14) at 13,500f/sack and Urea at 12,500 f/sack. DAP is (18-46-0) and Urea is (46-0- 0) b. Purchased by SONEGESS, the Burkina agency charged to acquire and distribute cereals to the malnourished.

43 Table A-2.Interviews with Farmers’ associations on the Central Plateau (February 2015)

Zone of Korsimoro 2012-2013 2013-2014 2014-2015 Traditional Production System 1. Relwende 2009 Maize: ¼ - ½ - Area (ha) ICSV 8 8 8 ha - Members 8 8 8 Millet: ½ - 1 ha - Reimbursement 24 24 16 Sesame: ½ - 1 - Yields (sacks/ha) NA 23-24 23-24 ha - Price/sack 18,000 fcfa 18,000 fcfa NA Sorghum: - Inorganic NA 2NPK (14-23-14) 1 DAP + 1 Urea ICSV: 1 – 5 ha fertilizer/ha + 1 Urea Local Sorghum: Subsidized : Urea ½ - 1 ha 12,500f ; 3 T with 3 T with zai 3 T with zai Peanuts ¼ - 1 NPK=13,500f zai 6 T without zai 6 T without zai ha - Organic fertilizer/ha 6 T Cowpea: ½ - 1 without zai ha Upland rice: ¼ - ½ ha 2. Nabouswende - Area (ha) 16 16 16 - First year 2009 - - - Members 26 26 26 - Reimbursement 48 sacks 24 sacks 32 sacks total 18,000 fcfa 18,000 fcfa NA - Price/sack NA 2 NPK + 1 Urea 1 DAP+1 Urea - Inorganic NA 3 T with zai 6 T 3 T with zai fertilizer/ha Organic without zai 6 T without zai fertilizer/ha

3. Tegwende - Area (ha) 25 25 25 - First Year 2010 - - - Members 43 43 43 - Reimbursement 33 sacks 32 50 total 18,000 18,000 NA - Price (fcfa)/sack NA 2 NPK + 1 Urea 1 DAP + 1 Urea - Inorganic fertilizer/ha (subsidized : Urea 12500 f et NPK=13500f) 3 T with 3 T with zai 6 T 3 T with zai 6 T zai without zai without zai - Organic fertilizer/ha 6 T without zai

44

4. Wendwaoga - Area (ha) 13 13 13 - First Year 2011 - - - Members 22 22 22 - Reimbursement 15 14 24 - Price/sack 18,000 18,000 NA association - Inorganic NA 2 NPK + 1 Urea 1 DAP + 1 Urea Fertilizers/ha (subsidized) NA 3 T with Zai 3 T with Zai - Organic 6 T without Zai 6 T without Zai fertilizers/ha

Zone de Pissila 2012-2013 2013 - 2014 2014-2015 Systèmes de production 1. Tegwende Maïze: ¼ ha - Area (ha) 15 15 15 Millet : 1 – 2 ha - First Year 2011 - - Sesame : 1 ha - Members 15 15 15 Sorghum- - Rembursement Reimburse Reimbursed in NA Kapelga : 1-4 d in cash cash ha - Price/sack paid by No sale No sale NA Peanuts : ¼ - 1 association - ha 2 NPK + 1 Urea Cowpea: 1 ha - Inorganic NA 2 NPK (14-23- fertilizer/ha 14) + 1 Urea (subsidized) - Organic fertilizer/ha 1.75 t 1.75 t

2. Namanegb This farmers - Area (ha) - 10 10 association was - First Year - 2013 - not financed by - Members - 10 10 the project. - Rembursement - Crédit reimbursed NA Farmers in cash imitated the - Price/sack paid by - No sale NA farmers’ association association - implementing - Inorganic 2 NPK + 1 Urea 2 NPK + 1 Urea the project fertilizer/ha activities. (subsidized) -

45

- Organic 1.5 t 1.5 t fertilizer/ha - 3. Sougrimane This farmers’ - Area (ha) - 10 10 association was - First Year - 2013 - not financed by - Members - 10 10 the project. - Rembursement - Sold cowpeas to NA Farmers - reimburse credit imitated the - Price/sack paid by - No sale farmers’ association association implementing - Inorganic - the project fertilizer/ha 2 NPK + 1 Urea 2 NPK + 1 Urea activities (subsidized)

- Organic 2.25 t 2.25 t fertilizer/ha -

46 Introducing New Millet Production Systems in Mali John H. Sanders, Jean Harman, Botorou Ouendeba, and Soungalo Traore Abstract In developing a pilot project and then an action agency broadening it (2004-2012-pilot; 2010-2013-IICEM scaling up) we were concerned with the continuing development in three areas, technology introduction, marketing performance, and institutional evolution. 1 On technology the concept of using inorganic fertilizers and better agronomy with improved seed of a well-known cultivar, Toroniou, was accepted for millet in the two principal regions of millet production, Segou and Mopti. Farmers recognized and appreciated the yield differences with common practices. Not surprisingly once implemented on a larger scale there was some technology erosion. Fertilizer recommendations were not followed well and the seed not renewed sufficiently. Nevertheless, farmers consistently recognized the advantages of DAP over the NPK. The availability of DAP has become a macro policy issue in Mali. Creating a sector of seed producers is critical to most crop program continuity. On marketing the farmers’ associations developed in our pilot and with IICEM activities have taken the collecting and storing to obtain higher prices by avoiding the post-harvest price collapse. Farmers’ associations are also focusing on clean grain to obtain a price premium from that. Changing market structures to obtain part of the marketing margin for the farmers’ associations is another marketing strategy to increase incomes of farmers. The farmers’ associations are becoming significant players in the marketing chain often dealing directly with wholesalers rather than with local collectors or the smaller regional “commercants” (merchants). This obviously stimulates the margins of both collectors and regional merchants.

The farmers’ associations have developed excellent repayment behavior of input credit thereby maintaining the revolving funds. They have set up bank accounts, hold regular meetings for information and group decisions, and have rules about non-compliance that are being implemented. However, our main metric of the quantities farmers allowed the associations to sell for them voluntarily besides the input repayments was still extremely low. As farmer confidence increases in the farmers’ associations and farmers’ production increases we expect these voluntary sales through the association to increase. This is a nice metric combining farmer confidence in their associations and the associations’ ability to market better and to be transparent so that the farmers know they will be well compensated for additional quantities that the farmers’ associations sell for them.

So we conclude that there was good continuing performance in all three areas but several activities that now need to be focused on to continue the momentum. In the 2015-16 crop year we report the results of the implementation of a program in the Bankass and Koro “cercles” to

1 We are grateful to Abou Berthe and oumar Sandinan and other members of Sasakawa 2000 for implementing our earlier pilot project in Segou and their continuing scalling up project and help in conducting these interviews in the Segou region.

47 use the revolving funds to respond to the technology erosion on approximately 3000 ha insuring that DAP and certified Toroniou seed return to the production mix. (See the following report in this bulletin, "The Mopti Field Program, 2015-2016 Crop year)

Introduction

Improved seed of a well-known millet cultivar (Toroniou) was introduced to farmers’ associations along with moderate fertilization and a series of marketing measures in a pilot project (the Production-Marking project of INTSORMIL) in four Sahelian countries focusing on Mali from 2008-2013. From 2010-2013 another project, IICEM, was commissioned by USAID to extend this pilot project on to a larger scale in Mali and to involve the banking sector to finance the inorganic fertilizer. In the movement from a pilot to a larger scale diffusion there are often various adjustments made. For example, the recommended fertilizers may be difficult to obtain. One field research problem then is how much these adjustments have affected productivity and incomes and what can be done to return to the productivity levels of the pilot. On the marketing side, there is a learning by doing aspect as the farmers’ associations learn how to find better markets, improve timing of their sales, add value, and market larger quantities. We were concerned with how much of the marketing strategy has been implemented and what else can be done to raise the marketing margin obtained by the farmers’ associations. Finally, we evaluate the farmers’ associations’ evolution with several measures reflecting their performance in introducing the technologies and marketing strategies and in gaining the confidence of the farmers.

Before evaluating the continuing performance we lay out the methods and review the project objectives. Evaluation method

This project was initially undertaken to respond to the criticism that agricultural researchers continue to do research on the station and with regional trials but do not make the effort or do the research to actually get their improvements onto farms especially with the staple crops often referred to in a derogatory fashion as subsistence crops. So we defined a technology based upon our analysis of the constraints including farmer acceptance of a new cultivar with associated technology. With the national agricultural researchers we identified a moderate fertilizer dose, better agronomic methods, and with the farmers a potential high yielding cultivar, that they appreciated. Other components included new market strategies to pay for the use of higher input levels and a new (or improved)3 institutional structure to implement the diffusion process.

Then we began this project in 2004 in collaboration with extension services and some NGOs for the farmer contacts. Ultimately, we had programs in villages in four Sahalian countries (Niger, Mali, Senegal and Burkina Faso). In 2010, USAID-Mali requested Abt Associates then running the Economic Growth project, IICEM in the AID portfolio, to scale up our pilot project

3 In the villages we generally found farmers’ association just not associations concentrating on millet (or sorghum).

48 by following our technologies and strategies and adding in their bank contacts to provide input 4 credit for farmers’ associations. For the quantitative analysis here, we interviewed in the primary millet production regions of Mopti and Segou. The focus of the interviewing was with the implementing agency officials of the farmers’ associations and their farmers who attended the meetings. There were 46 farmers’ associations involved. The interviewing was deliberately done in the priority regions of the coordinating agencies (IICEM in Mopti and Sasakawa in Segou) as these were our models for the rest of the country. The survey was an appraisal of continuity and changes in the technologies, market strategies and institutional approach of our model. We evaluated the present status of the farmers’ associations with which IICEM or we had worked for performance and for their adherence to project recommendations. Then we made recommendations for continuing development. There were a number of features of our project, which the farmers’ associations had continued well and even expanded on. There were also some major erosions in the technical recommendations as would be expected as development agencies experience practical problems in implementation or make their own modifications. On the marketing side there was some evidence of price fixing or cartel behavior among the wholesalers as the systems responded to new entrants, the farmers’ associations, wanting part of the marketing margin. Model Program The objectives of our field pilot program were: 1).Introduce a new technology for millet (and sorghum) based upon inorganic fertilizer and a variety responsive to fertilizer. Combine these two with water retention techniques and improved agronomy to reduce the risks from rainfall variability and to further increase yields; 2)Train farmers in the farmers’ associations in marketing strategies to offset the two price collapses (annual harvest time and good year) suffered by primary product producers, increase value added of the millet, and improve the bargaining power of the farmers in the marketing system including eliminating stages of the marketing chain; 3) Develop the farmers’ associations as a vehicle for facilitating a more rapid introduction of the technologies and marketing strategies. a) The principal yield constraint for millet and sorghum is soil fertility. Sahelian soils are often deficient in N and P. Unfortunately, popular wisdom is that millet does not respond to fertilization and even if millet did, it would not be profitable or farmers would not do it. One primary objective of the Production-Marketing project of INTSORMIL was then to disprove these myths. So we introduced a moderate dose of inorganic fertilizer higher than the micro- fertilization being promoted by many agencies including AGRA but lower than the physical maximum often recommended by experiment stations. The initial recommendations were two

4 Neither project continued after 2013.Both projects were prohibited from working with governmental agencies after the coup in March 2012.

49 sacks of NPK and one sack of Urea. Over time we shifted to one bag of DAP and one bag of Urea. This provided the nitrogen and phosphorous but not potassium.5 b) The second basic project component is a series of marketing strategies. The objective of these marketing strategies is to assure that farmers can pay for the increased inputs and to moderate some of the price collapses suffered by producers of basic staples. With increasing fertilizer and other input prices, many have argued for low inputs, reducing fertilizer levels, or searching for low input solutions. This has been tried with no significant yield effects for the past twenty years and it is now time to return to the basic needs of the plant. You would not ask a malnourished kid to try to save money by eating less. Do not try to reduce essential nutrient levels of plants. Rather than reducing fertilizer costs increase revenues with the marketing strategies. How? First, reduce or eliminate the cereal price collapses, the annual collapse at harvest time and the collapse in good rainfall years. With storage facilities get the farmers’ associations to hold the cereals at harvest until there is price recovery thereby responding to the harvest price collapse.6For the good year price collapse, develop the secondary markets, i.e. millet food processing and sorghum in the animal feed. There is a value added possible at the farm and farmers’ association levels. Millet food processors and the relief agency, PAM, are prepared to pay a price premium for cleaner millet. This protects the machines of the processors and reduces the cleaning costs. There is also a concern with the deliberate adulteration of the cereals. Over time, a price premium is being paid for clean cereals usually in the 15 to 25 fcfa/kg range. This premium can be greater as in Burkina Faso where there has been much slower technology introduction for millet than in Mali hence a slower rate of expansion of clean millet. PAM has created its own system for insuring clean millet by screening, sacking and identifying the producers. By selling in quantity and holding the cereal until there is a better price the farmers’ associations perform many of the functions of the collector and regional merchant enabling them to sell to the big wholesaler). The farmers’ associations still need to invest more in market information and negotiation as is indicated by the substantial difference in PAM7 prices and those paid by the wholesalers in Mopti (Tables 1 and 2). But the farmers’ associations are gradually learning from their experience. One response of merchants when farmers’ associations refuse to sell at harvest, the low price period, is to fix prices at low levels among themselves for the later period when prices normally go up so that the farmers’ associations (and hence the farmers) do not benefit from the seasonally higher prices. Over time, farmers will get better at the tough negotiation process and lean to overcome cartels with arbitrage. Arbitration means moving the cereal outside the area where the cartel is operating. Many farmers’ associations and certainly the collaborating villages around a pole or central farmers’ association are big enough

5 African soils generally are not deficient in potassium. Over time as higher cereal yields are attained, they will become deficient but presently we need to get N and P levels up and keep the costs down. 6 Because there is a rotating fund for purchasing inputs for the next crop season part of accumulated cereals will need to be sold by the end of May 7 PAM argues that they pay the market price at the time of purchase plus a premium. The premium has been substantial.

50 to hire their own truck and to find better markets with more market search. So we will evaluate here how the farmers’ associations have done with respect to both technology introduction and the use of these marketing strategies.

c).The third component of our program was developing strong farmers’ organizations. The first objective is to use the associations for rapid diffusion of the new technology and better practices. Farmers often do not believe in technology recommendations from outsiders. Involving 50 (first year) to 150 people (third year) in the village was the objective of the pilot. We expected 40 to 60% of the farmers would follow the technology recommendations to make sure that they continued to have access to the fertilizer with credit from the revolving fund. Then in the second year, those farmers not following well the recommendations would learn from these successful farmers. This assumes that the rainfall conditions and other stochastic factors affecting yields are not adverse in the first years of the program.

Secondly, the farmers’ association is responsible for implementing the marketing strategies. The farmers’ associations control the storage, the quality checks for clean cereal, the tracking process through labeling of the sacks, the market search, the handling of the revolving fund, the bank account of the revolving fund, the search for input credit, and most importantly the negotiation process for selling the cereal including timing and search for better offers. We will discuss various elements of the progress of the farmers’ associations below. But, we will focus on reimbursement of the input credits from the revolving fund and the quantity members allow the association to sell for them once they have paid off the reimbursement. The latter is our best measure of the confidence farmers have in the association to sell well for them and to make reasonable charges for storage.

Evolution of the Program in 2010-2013

Our project was implemented on a small pilot or demonstration scale. For example we only had seven village sites in the Mopti region in the period 2010 to 2013. In 2010 IICEM, would work with over a hundred farmers’ associations in the Mopti region. To assure bank interest IICEM made contracts with the wholesalers. IICEM provided a guarantee fund to the banks for the first two years of the program. With good monitoring and training IICEM was successful in getting high repayment rates and their farmers’ associations’ fulfilling the contracts in 2010 and 2011 in the Mopti and Sikasso regions. The repayment rates were so high in the first two years that the guarantee requirement was programmed to be zero in 2012. Then in 2012 there was the war in the north, a coup, a beating of the President by a mob, and the French intervention to stop the rebels from advancing south. Banks closed in the north. USAID intervened with a grant through the farmers’ associations for fertilizer in the Mopti region to cover 5,500 ha at two sacks per ha (Goita, former field supervisor for IICEM in the Mopti region, March 2015). As in our program, the farmers were asked to repay the loans in cereal at harvest to the farmers’ association to set up a revolving fund. The Dutch government made a similar grant in the south including the Segou and

51 Sikasso regions through the farmers’ associations. Here the reimbursement requirement was only half of the loan to be paid to the farmers’ associations. Revolving Fund The most significant factor in the surveying was the continuing effects of the rotating funds of the farmers’ associations from the emergency relief of the US and the Netherlands in 2013. The other important factor was the evolution of direct sales of the farmers’ association to the large wholesalers in Mopti and to the food relief agency of the United Nations, PAM, 8 in Segou. In 2010 and 2011, IICEM had obtained input credits for the farmers’ associations by obtaining contracts between wholesalers and the farmers’ associations and showing those to the bank (BNDA). As is not too surprising given the differences in bargaining power these contracts were more advantageous to the wholesalers than to the farmers. The contracts insisted on farmers paying soon after harvest and selling more at these low prices than just to pay off the loans. With the grants of 2012, the farmers’ associations became independent from the wholesalers’ conditions as the bank loans were no longer necessary to obtain input credit. The farmers’ associations provided input credit to their members with these revolving funds in 2013 and 2014. Nevertheless, many of the farmers’ associations continued dealing directly with the wholesalers in Mopti rather than local collectors or the regional merchants in Koro and Bankass. In the winter of 2015, these rotating funds from 2012 were still operating in both the Mopti and the Segou regions. The structure of this funding was exactly from our project of an initial loan to the farmers’ association for inputs. This was extended to farmers, who were then required to repay at harvest in kind. The farmers’ association would then hold these cereal stocks and sell them before the next planting season. The sale of these cereals would then pay for another round of input purchases. Additional profits could be used for expanding membership or for other objectives to be decided by the farmers’ association. So how much of the marketing margin did these farmers’ associations get from selling to the higher levels of the marketing chain? We compare here the prices received from the wholesalers with the local/regional prices at the same time. The big gains were selling to PAM. Seventeen percent of the farmers’ associations sold to PAM in the Mopti region and PAM is in the process of expanding its activities in there region (Table 1). However, the wholesalers in Mopti paid a much lower price than PAM. 48% of the farmers’ associations in Mopti sold to the wholesalers (Table 1). The average price gain was 7.7 fcfa/kg from dealing with the big wholesalers rather than the local or regional markets with a range of 0 to 15/kg. For a sack worth 16,000 fcfa farmers were only getting an additional 770 cfa or 5%. The wholesalers did have additional costs of transportation (included in calculation) and

8 PAM’s designated concern from its UN mandate is to improve the nutrition of the poor. So initially, they were buying from big wholesalers or at lower levels in the marketing chain acting like the standard “commercants” attempting to buy at the lowest post- harvest prices to minimize their acquisition costs. With the initiation of P4P within PAM they became concerned also with low income farmers. They began buying from farmers’ associations, obtaining cleaner millet, and eventually setting up a new organizational system among farmers’ associations.

52 some paid the expenses to bring farmers from the village to Mopti to make sure that the cereal arrived and to see the evaluation and weighing. Nevertheless, the bargaining power of the farmers’ associations with the wholesalers was still minimal. Moreover this substantial price difference with PAM and the uniform prices paid by the wholesalers, which were often the same or only marginally different than the regional merchant, indicates price fixing and cartel behavior among the wholesalers. Over time, the farmers’ associations will be able to negotiate tougher since bypassing marketing stages can also save costs for the wholesaler especially if the cereal is cleaner and more uniform. However, clearly a market structure with an old boy network among the different levels resists changes from newcomers looking for a share of the marketing margin. When facing a cartel and fixed prices, the farmers’ associations need to learn to use arbitrage, selling outside the region. They would need market information, to arrange transportation, and for several farmers’ associations to work together to sell larger quantities.

53 Table 1. Crop Area, Markets and Marketing Margins in the Mopti Region

Village Area Markets and prices Price of Regional or Price Difference (2013-14) 2013-2014 Local Market (Local or Regional v. (ha) (transport cost/sack Wholesaler) to Mopti) Adjusted for transport costs 1.Logo (IICEM) PAM (205f/kg) et 2 NA 250 wholesalers (Moulaye et Transport : 1000/sack Mamoudou Guindo)Mopti

2.Dimbale (IICEM) PAM through the farmers’ Price : 130 45 fcfa/kg 45 association of Logo (185f/kg) Transport : 1000/sack 3.Sadia (Projet Production- Wholesaler Mopti 160 160 fcfa/kg Local salea Marketing) 120 Members 155

4.Kanikombole (Projet Merchant Bandiagara (160 Price : 160 Local salea Production-Marketing) 73 fcfa/kg) 5.Telly (IICEM) 103 Merchant (160fcfa/kg) Price : 160 Local salea

6.Djanwelly (IICEM) Guindo Wholesaler Mopti 160 fcfa/kg Local salea 28 160 fcfa/kg 7.Ogotana (IICEM) Mamoudou Gindo Mopti Price 160 10 cfa/kg 82 (170f/kg) Transp: 0 8.Ogossagou/IICEM 105 Mamoudou Guindo Price: 155 8 fcfa/kg Mopti(165f/kg) Transp: 200/sack 9.Sokoura (IICEM) 84 Mamoudou Guindo(170f/kg) Bankass 160 10 fcfa/kg Transp: 0 10.Tinto/Barwe (IICEM) 130 Mamoudou Gindo (165f/kg) Bankass 165 0a Transp: 0 11.Ogodire (IICEM) 94 Mamoudou Gindo (165f/kg) Bankass 160 5 fcfa/kg Transp: 0 12. Pel (IICEM) 60 Moulaye Wholesaler Mopti Prce 150 15 fcfa/kg (165f/kg) Transp : 0 13.Temegolo (Production- 126 Moulaye Mopti(175f/kg) Price: 160 14 fcfa/kg Marketing) Transp : 100 14.Pomorododiou Na 88 Moulaye Mopti (155f/kg) Price: 155 0 (IICEM) Transp : 0 15.Pomorododiou Begne Baba Wholesaler Mopti Price: 140-145 10 fcfa/kg (Production-Marketing) 125 (155f/kg) Transp : 0 16.Yadjenga (IICEM) 105 Wholesaler Moulaye (IICEM) Koro: 160 5 fcfa/kg (165f/kg) Transp : 0

17.Tere (Production- Village merchant, collector en Price : 162,5 Local salea Marketing) 63 2013 (162,5f/kg) Transp : 0 18.Tinasasogu Merchant Koro 155 fcfa/kg Price: 155 Local salea (IICEM) 80 Transp : 0 19.Togo Tina (IICEM) Wholesaler Guindo a Mopti Price: 160 15 fcfa/kg 120 (175f/kg) Transp : 0 20.Birga Dogon (IICEM) Merchant Birga 170 fcfa Price: 170 Local salea 60 Transp : 0 21.Kountogoro (Production- Merchant Pei (150f/kg) Price: 150 Local salea Marketing) 73 Transp : 0 22. Balirou (IICEM) PAM en 2013 (205,25f/kg) et Price: 160 35 fcfa/kg 290 2014 Transp :1000/sack 23.Tendely (IICEM) PAM en 2013 (205,25) et Price: 160 35 fcaf/kg 450 2014 Transp : 1000/sack Mean 126 ha 170 cfa/kg

54 a. Indicates no advantage between selling to the regional market and the wholesaler so sold locally. Markets are networks set up over time and their participants often do not want to change them. So this lack of difference in prices appears to be cartel behavior. Note the much higher prices paid by PAM in Mopti and Segou. PAM claims to be paying local prices plus a small price premium for quality. In Segou almost the entire sample received the PAM prices. Hence, a strong stimulant for program participation and reimbursement was created by the PAM purchases (Table 2).

Table 2. Size of Farmers’ associations, area, buyer and price paid, Segou, cercle de Baraoueli.

Village Gender of Areas (ha) Yields (tons/ha) Markets and Prices Paid Members 2012-2013, 2012-13, (2013-14) F H 2013-2014 2013-14 1.Merabugu 4 33 4 0 (2013) 2 t in 2013 PAM (200f/kg) 2.Gurele-were 0 38 91 (2013) 1.6 t in 2013 PAM (200f/kg)

3.Konobugu 20 39 60 (2012) 2.85 t in 2012 PAM (200f/kg) 90 (2013) 2.7 t in 2013 4.Badinantu 30 70 100 (2013) 1.2 t in 2013 PAM (200f/kg)

5.Tigui 5 41 60 (2012) 1.6 t in 2012 PAM (200f/kg) 96 (2013) 1.57 t in 2013

6.Koduguni 59 0 15 (2012) 0.75 t in 2013 PAM (190f/kg) 20 (2013) 7.Welingra 0 25 30 (2012) 1.48 t in 2012 PAM (200f/kg) 32.5 (2013) 1.26 t in 2013

8.Diawarala 16 24 80 (2012) 70 1.17 t in 2013 PAM (200f/kg) (2013) 9.Nugula 25 40 67 (2012) 52 1.11 t in 2013 PAM (2013) 10.Kenema 12 50 80 (2012) 65 1.18 t in 2013 PAM(190f/kg) (2013) 11.Diarabugu 2 18 30 (2012) 27 1.4 t in 2012 PAM (190f/kg) (2013) 1.8 t in 2013

12.Wentibugu 23 23 60 (2012) 50 1.8 t in 2013 PAM(200f/kg) (2013) 13.Kamba 10 42 57.5 (2012) 0.46 t in 2013 PAM(200f/kg) 21 (2013) 14.Soungola 1 31 50 (2012) 40 1.7 t in 2012 PAM(190f/kg) (2013) 0.8 t in2013

15.Niontobugu 4 28 10 (2012) 65 1.30 t in 2013 PAM(200f/kg) (2013) 16.Zana 5 30 110 (2012) 1.8 t in 2013 Regional 60 (2013) merchant(155f/kg) 17.Kokribugu 6 76 55 (2012) 60 2.36 t in 2013 PAM(200f/kg) (2013) 18.Tingona 0 24 40 (2012) 45 1.35 t in 2013 PAM(200f/kg) (2013)

55

19.Tingoni 45 85 185 (2012) 1.95 t in 2012 PAM, Two millet 200 (2013) 1.90 t in 2013 processing firms in Bamako (200f/kg) 20.Wela-cura 0 45 30 (2012) 40 0.88 t in 2013 PAM(200f/kg) (2013) 21.Bomoti-1 3 29 30 (2012) 35 1.5 t in 2012 PAM(195f/kg) (2013) 1.40 t in 2013

22.Bomoti-2 5 9 25 (2012) 20 0.6 t in 2012 PAM(200F/kg) (2013) 0.69 t in 2013

23.Segemba wili 1 31 17.5 (2012) 1.36 t in 2013 PAM(190f/kg) and 27 (2013) members(150f/kg)

Technology performance With an average of 7.5 tons/ha of organic fertilizer and two sacks of inorganic fertilizer Segou producers had average yields of 1.64 tons/ha (Table 3). That should be compared with 800kg/ha to 1 ton/ha for farmers not in the program. So these were good yield achievements in a good rainfall year especially compared with those without the technology package. However, comparison with Tingoni (No. 19), our initial pilot project indicates the potential to obtain another 200 kg/ha if the fertilizer recommendations were followed.10 These recommendations were either 2 sacks of NPK and one sack of Urea or one sack of DAP and one sack of Urea. When DAP is difficult to obtain or farmers’ associations are trying to keep costs down, it is not surprising to see short cuts taken.11 Unfortunately, this short cut of one sack of NPK for one sack of DAP has serious yield and profitability consequences as is recognized by the most experienced farmers’ association in our pilot program, Tingoni. The use of organic fertilizer is reasonably high and has been increasing, and the Segou farmers are marketing very well to PAM and receiving a good marketing margin. Hence areas in the project have been increasing and incentives to follow recommendations are high (Table 3). So it should not be difficult to get the farmers’ associations back on the higher yielding fertilizer recommendation. Now we move to the second most important millet region, where the rainfall is less and the soils are poorer, Mopti. Here there is less organic fertilizer at 6 tons/ha but not a lot less as the predominant Dogon work intensively to use their resources. Overall yields of both genders are lower than in Segou but at 1.1 tons/ha still substantially better than those outside the program of 400 to 600 kg/ha-unfertilized millet (Table 4). The farmers in the farmers’ associations are doubling yields in this good rainfall year. However, the inorganic fertilizer utilization is even further from the recommended dose than in the Segou region. In the last two years (2013 and

10 Since there was an upward bias of yields from having the same agency doing the interviewing as did the previous program administration and extension we expect that the real yield difference was 400 to 500 kg from the yield erosion of not following the fertilizer recommendations. 11 From the beginning of the IICEM program in Mopti in 2010 one sack of NPK was substituted for one sack of DAP. This reduces the P level from 23 to 7.5 units per ha. So it is important to get back to the higher level of P.

56 2014) IFDC picked up many of the IICEM farmers’ associations and reduced the fertilizer levels to their new definition of micro-fertilization- 35 kg of NPK and 35 kg of Urea. Then the availability of Urea ran out. Moreover, the farmers’ associations have not been able to obtain sufficient access to the improved seed of Toroniou as the program of IICEM was expanded rapidly here. Unfortunately, millet outcrosses rapidly so programs of seed renewal are especially important for millet. So the combination of the appropriate fertilization and improving access to good seed of Toroniou would increase yields another 400 to 500 kg/ha. This is especially important here where the marketing margin was squeezed by the greater bargaining power of the big wholesalers and cartel behavior apparently fixing the prices substantially below the level of PAM. Moreover, there is a serious disparity between the levels of organic fertilizer used by the women and the men in the Mopti region and probably also in the Segou region (Table 3). The women are using much less organic fertilizer per ha. The women always mention that they are the last ones to get access to the carts to bring the organic fertilizer to the field. They also have less access to the organic fertilizer available. Since the women generally follow better than the men the agronomic recommendations, giving them equal access to resources especially organic fertilization will enable them to compete with and push the men for increasing millet yields. On the areas in each association, the Mopti farmers’ associations had substantial variation but on the average was much closer to our ideal size of around 150 ha. The Mopti average was 131 ha. The men were supposed to have just one ha but this was not implemented by IICEM. Women obtaining land had access to either one quarter or one-half ha. The average area per association in Segou was much smaller at 71 ha. There was some collaboration between these smaller size farmers’ associations in storage and marketing. However, management talent for keeping books and pursuing markets is limited so a bigger area and membership is useful.

57

Table 3. Area, fertilization, and yields in Segou, cercle de Baraoueli

Village Area (ha) Inorganic Organic Variety Yields 2014-15 Fertilizer Fertilizer (t/ha) a (t/ha) (t/ha) 2014-15 1.Merabugu 50 1 NPK 7.5 Toroniou 1.5 1Urea 2.Gurele-were 50 1 NPK 0 Toroniou 1.9 1 Urea 3.Konobugu 118 1 NPK 12.6 Toroniou 3.05 1 Urea 4.Badinantu 100 1 NPK 4.2 Toroniou et 1.5 1 Urea locale 5.Tigui 120 1 NPK 6.72 Toroniou 1.5 1 Urea 6.Kodougouni 25 1 NPK 0 Locale/sanio 1 1 Urea 7.Welingra 50 1 NPK -- Toroniou 1.35 1 Urea 8.Diawarala 60 1 NPK 13.8 Toroniou 1.27 1 Urea 9.Nugula 82 1 NPK 7.5 Toroniou 1.12 1 Urea 10.Kenema 60 1 NPK 7.5 Toroniou 1.33 1 Urea 11.Diarabugu 45 1 NPK 6 Toroniou 2.7 1 Urea 12.Wentibugu 45 1 NPK 7.68 Toroniou 2.4 1 Urea 13.Kamba 58.5 1 NPK 6 Toroniou 2.26 1 Urea 14.Soungola 45 1 NPK 1.02 Toroniou 1 1 Urea 15.Niontobugu 75 1 NPK 5.46 Toroniou 1.16 1 Urea 16.Zana 55 1 NPK 2.4 Toroniou 1.8 1 Urea 17.Kokribugu 85 1 NPK 4.2 Toroniou 2.6 1 Urea 18.Tingona 80 1 NPK 10.98 Toroniou 1.9 1 Urea 19.Tingoni 215 2 NPK 21.6 Toroniou 1.85 1 Urea 20.Wela-cura 45 1 DAP 14.46 Toroniou 1.05 1 Urea 21.Bomoti-1 45 1 NPK 15.96 Toroniou 1.7 1 Urea 22.Bomoti-2 25 1 NPK 5.22 Toroniou 0.71 1 Urea 23.Segemba wili 30 1 NPK 3.36 Toroniou 1.17 1 Urea Mean 70.6 ha 7.5/tons/ha 1.64 tons/ha a.Yield data are biased upward. We interviewed the leaders of the associations and asked for the mean yields in the association but probably they often give their own mean yields. Generally these are the best farmers in the group.

58 Table 4. Crop Area and Yields (2014-15), Levels of Organic and Inorganic Fertilizers in the Farmers’ Associations in Mopti with a Gender Division Villages Area Gender Organic Fertilizer Inorgnic Fertilizer (t/ha) Variéty Yields (2014- M F (t/ha) (2014-15) 2014-15 (t/ha) 15) Number 2014- (ha) 2015 Men Women Fertilizer levels (kg/ha) DAP Urea NPK Logo 0 275 - 3 0 50 50 Toroniou 1.0 and local 230

Dimbale 100 0 6 - 0 35 35 0.8 150 Sadia 130 40 9 2 50 50 0 1.0 120 Kanikombolé 59 25 7 2,50 50 50 0 Toroniou 1.0 90 and local Telly 50 4 8 2,50 0 35 0 1.4 131,5 Dianwély 50 20 7 2,50 35 35 0 1.1 28 Ogotena 50 12 7,50 1,50 0 35 35 1.6 82 Ogossagou 105 106 31 6,50 2 0 35 35 Toroniou 1.1 Local Sokoura 105 50 30 5 2 0 35 35 0.9 Tintoa. barwé 60 37 75 6 3 0 35 35 Toroniou 1.3 Local Ogodiré 47 75 26 7,50 4 0 35 35 Local 1.5 Pel 80 30 70 5 2 0 50 100 Local 1.4 Témégolo 126 66 60 9 5 0 50 50 Toroniou 1.0 Pomorododiou 88 72 28 6 6 0 50 50 Toroniou 1.0 Na Begné 114 80 4,50 3 0 0 100 Toroniou 1.0 185 Yadianga 95 35 6 3,50 0 50 35 Toroniou 1.0 130 local Téré 60 45 3 1,50 0 0 100 Toroniou 0.9 in 91 2013 1,2 in 2014 Tinassansagou 60 20 1,50 0,4 0 50 50 Toroniou 1.2 T in 90 New bobo 2013 1.0 T in 2014

Togo-tina 60 20 5 3 0 50 50 Toroniou 2013 : 120 New bobo H=2 ; F =1.5 ; 2014 : H=1.8 and F=1.5b Briga-dogon 142 6 4 2 0 35 35 Toroniou 0.9 in 70 New bobo 2013 1.0 in 2014

59

Kountogoro 100 60 7,50 1 0 0 100 Toroniou 0.85 in 146 2013 1,2 in 2014 Balirou 53 67 6 1 0 50 50 H= 1,2 145 F= 1,0 Tendely 200 225 5 5 0 35 35 Toroniou 1,4 580 Local (2) Means 131 ha 6t/ha 2.65 tons/ha 1.16 tons/ha a. Same comments as for Table 3. b. H and F: separate yields obtained by the men and the women respectively

Marketing Strategy Performance: There is widespread adhesion to several of the strategies of the original program.13 This includes storage and later sales, producing clean cereal, and insisting on a price premium for the clean millet. Certainly, PAM has helped to increase the price paid. One farmers association in the Segou region (Tingoni) continues to work with the Bamako millet food processors in spite of the availability of PAM purchase. But the dominant phenomenon in this Segou sample is PAM. PAM has largely changed from buying solely on a large scale from wholesalers to also being concerned with low-income farmers producing millet as P4P14 within PAM pushes to benefit farmers’ associations by buying directly from them or pushing wholesalers to give a better price. P4P sets up a structure with one principal farmers’ association in the center and then ten to twelve other farmers’ associations in a circle around this pole or center. The outer farmers’ associations bring their millet to the central association. There it is screened several times to clean it and put into PAM sacks with the name of the farmers on them. Then the central farmers’ association pays for transportation and the PAM sacks and ships the millet to the PAM central facilities in Bamako or Mopti. There are two indicators of progress of the farmers’ associations. First the re- imbursement rates. There were only two cases of less than 100% repayment in Mopti and none in the Segou region. In these two cases, the members were thrown out of the association until they repaid. They were expected to repay in 2015. IICEM did an excellent job in instilling the need for repayment. Given the historic problems with repayment in the Sahel, this is evidence of institutional evolution. To pay for reimbursement required from 5 to 22% of production (except for one case of yield collapse) with an average of 14% in the Mopti region (Table 5). These rates were higher in the Segou region at 9 to 36% (again with one case of poor yields leading to 54%) with an average of 18% (Table 6). Even though it was more expensive in grain production to repay for the fertilizer in Segou the substantial price premium from PAM resulted in a strong incentive to maintain the program and repay. Our preferred indicator for the farmers’ association performance is the sales the farmers allowed the association to make for them after repayment of the input credit. This indicates

13 Many programs including AGRA and OXFAM have been promoting village storage systems. Our program linked storage facilities to the revolving fund to set aside funding for input credits. With the establishment of the importance of repayment, this revolving fund can be substituted for bank financing or used as leverage to get further bank funding in those farmers’ associations that have established accounts in the banks with their revolving funds. 14 An addition to PAM concerned with getting low income farmers good prices, insuring clean grain, and further development of the farmers’ associations.

60 farmer confidence in the farmers’ association to get them a good price and not charge high costs for the storage or other services. Note that neither region did well by this indicator. In Mopti, these sales ranged from zero (52% of the observations) to 6% with an average of only 1% of production. In Segou closer to the main commercial center, Bamako, these sales took up more of production but were still rarely over 5% (Table 6).The average was only 2% of production. As farmers can sell more they can also achieve more food security. In many sites farmers reported that they did not even know about the potential of selling more of their cereal through the association. Larger quantities sold by the association would result in more bargaining power. Hence, these low quantities of cereals relegated to association sale also indicates an important 15 weakness in the interaction between those running the associations and their membership.

Table 5. Importance of Reimbursements and Other Sales through the Farmers’ Associations in the Mopti Region. Area Reimbursement Member Sales Additional Sales Villages 2013-14 Production Reimbursement as a Percent of through the through the (ha) (t) (2013-14) (t) Production Associationa Association as a Quantity (t) Percent of % Production Quantity (%) Logo 250 300 50 16,66 0 0 Dimbale 45 22,5 9 40 1 4,44 Sadia 120 120 14,8 12,33 0 0 Kanikombolé 73 91,25 15,2 16,65 0 0 Telly 102,85 133,70 7,2 5,38 0 0 Dianwély 28 35 5,6 16 0 0 Ogotena 82 106,6 13,2 12,38 0 0 Ogossagou 105 115,5 21 18,18 5,5 4,76 Sokoura 84 75,6 5,8 7,67 1 1,32 Tinto-barwé 130 169 12 7,10 5 2,95 Ogodiré 94 137,05 9,4 6,83 0,3 0,21 Pel 60 84 12 14,28 2 2,38 Témégolo 126 151,2 25,2 16,66 0 0 Pomorododiou Na 88 88 17,6 20 0,5 0,56 P. Begné 126 151,2 18,75 12,40 0 0 Yadianga 105 105 21 20 6 5,71 Téré 63 56,7 12,6 22,22 0 0 Tinassansagou 80 96 16 16,66 0 0 Togo-tina 120 240 24 10 0,3 0,125 Briga-dogon 60 54 12 22,22 0 0 Kountogoro 72,5 61,62 11,70 18,98 0 0 Balirou 290 348 29 8,33 4 1,14 Tendely 450 630 50 7,93 2 0,31 Mean (excluding Dimbale on reimbursement) 14.1 1.0

15 Farmers asserted that their principal benefit from the program especially in the Mopti region was obtaining larger quantities to store longer for food security. So another interpretation of the lack of contribution of farmers for sales by the association is that the members needed the rest of the millet for food security or later village sales to be able to buy necessities at the local markets.

61 Table 6. Importance of Reimbursements and Other Sales through the Farmers’ Associations in the Segou Region. Village Area Production Remburse Reimbursemen Other Other Sales as (ha) Tons ment ts as a % of Sales a % of 2013 2013 Tons Production (tons) Production (%) (%) 1.Merabugu 40 80 10 12.5 0 0 2.Gurele-were 91 145.6 22.75 15.6 0 0 3.Konobugu 90 243 22.5 9.3 32 13.2 4.Badinantu 100 120 12.5 10.4 10.5 8.7 5.Tigui 96 150.72 24 15.9 0.7 0.5 6.Koduguni 20 15 5 33.3 0 0 7.Welingra 32.5 40.95 8.12 19.8 1.2 2.9 8.Diawarala 70 81.9 17.5 21.4 1 1.2 9.Nugula 52 57.72 13 22.5 0.2 0.03 10.Kenema 65 76.7 13 16.9 1.8 2.3 11.Diarabugu 27 48.6 6.75 13.9 2 4.1 12.Wentibugu 50 90 12.5 13.9 3 3.3 13.Kamba 21 9.66 5.25 54.3 4 41.4 14.Soungola 40 32 10 31.2 1.4 4.4 15.Niontobugu 65 84.5 13 15.4 0 0 16.Zana 60 108 12 11.1 0.5 0.5 17.Kokribugu 60 141.6 15 10.6 1 0.7 18.Tingona 45 60.75 11.25 18.5 2 3.3 19.Tingoni 200 380 45 11.8 16 4.2 20.Wela-cura 40 35.2 10 28.4 0 0 21.Bomoti-1 35 49 7.5 15.3 0.8 1.6 22.Bomoti-2 20 13.8 5 36.2 0 0 23.Segemba wili 27 36.72 4.37 11.9 0 0 Means (excluding Kambaa) 18.% 2.3 % a. It was dry early and late. Kamba planted later, waited too long, and got decimated by the drought at the end of the season. Kamba rebounded with yields over two tons in 2014.

Program Summary In this table the major issues are summarized and recommendations made.

Table 7. Review of the Millet Technology Introduction Process Corrections to Observed Positives Negatives Strengthen Outcome Fertilizer was made The popular myths Not always combined with Public policy support available to all farmers’ about the lack of sufficient organic fertilizer to for making DAP and associations though not response or profitability improve the structure while Urea available to millet necessarily of the type or of fertilizing millet the inorganic fertilizer producers. Training in Inorganic quantity desired. DAP was were definitely refuted. improves soil fertility. improved agronomic Fertilizer generally not available and This was found to be the methods to insure one sack of NPK rather case most often in women’s payoff to inorganic than two often used. fields. fertilizers. (Tables 3 and 4) Substantial variation in Pervasive use. Farmers Women discriminated Finance carts for organic fertilizer use and in recognizing importance against in access to organic women. Organic yields. No clear relationship fertilizers and carts to move Fertilizer observed. it to the field Discuss access issue for organic fertilizer for women.

62

Corrections to Observed Positives Negatives Strengthen Outcome Use of traditional varieties Toroniou responded Need for regular infusion of Improve identification and older “new varieties.” well to the other new new certified seed and training of seed Toronio, sometimes three technologies used. producers in the years out from new farmers’ associations Seeds seeding.

Revolving fund system Generally revolving Bank credibility was not How quickly should working well for providing funds kept in banks or used for leveraging of more these associations use input credit and being microcredit institutions. funds to expand faster and to leverage of their capital Credit reimbursed. Reimbursement at or become involved in other to borrow more? close to 100%. village activities needing financing. Program technology: new Technology as key Technology erosion away DAP and more Urea seed, inorganic fertilizers, program component. from fertilizer made available to millet Technical organic fertilizer, water Increasing yields drives recommendations. Failure to producers. Input and retention techniques, better all the rest of program renew the seed and to More systematic Oversight agronomy, storage activities. adequately train farmer seed training and marketing technology. producers. . help for small farmer seed producers 4 to 6 farmers’ association Maintaining strong More pro-active marketing Marketing courses for officers usually observed.. organizations in period by leadership. Need to win leadership especially for Governance of public chaos (war, member confidence for responding to cartels coup). increased voluntary sales and price fixing. Techniques for clean Clean grain is not an As quantities of millet Poor thresher history handing at harvest were issue with the farmers’ harvested increase, the with machines breaking also clearly understood, associations. The pounding will become down quickly in the though not always used. process of handling untenable in Mopti (the field. Identify and the millet heads as amount of time) for the import better models. Harvesting/han they are harvested women and girls responsible Also grinders on the dling/cleanlines through the for the pounding. The next village level Financing s of the grain pounding/threshed/and step as in Segou, driving for above? winnowing processes. over the grain, leads to dirty The linkage to millet increased prices for clean millet is clear.

Almost all the millet in 100% reimbursement. Fixing of pricing with cartel In Mopti need to use Segou was purchased Plus small additional in Mopti. Little private arbitrage. Sell or be directly by the UN World quantities marketed entrusting of millet for sale able to sell outside Food Program (WFP). through farmers’ by farmers’ association after region. More advance Farmer sales to wholesalers associations. repaying credit contacts and in Mopti often received transportation needed. poor prices. Cartel behavior Joining together of Marketing various farmers’ associations to increase quantities and marketing competence. More work by associations on gaining member confidence

63

Corrections to Observed Positives Negatives Strengthen Outcome Where women were Women interviewed in Women with little access to Alternate forms of allocated land to produce Mopti region were land and to carts for hauling threshing (small scale millet the majority of pleased to learn how to organic fertilizer. As the and large) need to be a women interviewed used manage and cultivate production of millet very high priority. the program practices and millet from planning to increases in Mopti, the Where are the good produced millet marketing. The fields burden on women to hand threshers that hold up a approximately equal to the allocated to women thresh the millet will become few years? Continual men. allowed the women untenable. breakdowns observed. additional income. Need individual ownership of threshers Gender rather than group Considerations ownership to avoid machine abuse.

Farmers’ association commitment to support women’s participation in the program insuring access to all needed inputs.

Source: Adapted and modified from Jean Harmon, Consultants Report, 2015

Conclusions: Millet yield increases of 50 to 100% from traditional practices is good performance and we have shown elsewhere the substantial income benefits from doing this (T. Abdoulaye et al, 2008, F. Baquedano et al, 2009; J. Coulibaly, 2010; and J. Coulibaly et al., 2013). However, the fertilizer recommendations have been eroded with the difficulty of getting DAP in Mali. The fertilizer proven recommendation is either one sack of DAP or two sacks of NPK, both combined with one sack of Urea. One sack of NPK is not equal to one sack of DAP. Moreover, in Mopti and in Segou renewed Toroniou seed needs to be made more widely available. These two returns (DAP and certified seed) to program recommendations offer the potential for another 200 to 400 kg/ha of millet per ha. The use of organic fertilizer is also very important. Farmers know this and intensively use organic fertilizer but reminding them that inorganic and organic fertilizer are complements and not substitutes is important. The large differences between Segou and Mopti in the use of organic fertilizer indicate the larger numbers of livestock in the Segou region. The revolving fund has been well established in both millet regions. The IICEM program of the farmers’ associations bypassing levels in the marketing chain to deal with wholesalers has been continued. The rotating funds restore some bargaining power to the farmers’ associations, as the farmers’ associations no longer need the contracts with the wholesalers to pay for the fertilizer. Even with the availability of the revolving fund of the fertilizer credits from the US and Dutch governments, most farmers’ associations in Mopti have maintained the wholesalers as their principal markets. However, as yet these direct sales have not resulted in much increase of the marketing margin for the farmers’ associations. By working with other farmers’ associations and arranging transportation, the farmers’ associations can break cartel behavior of the wholesalers. The increasing role of P4P within

64 PAM will also push up the prices received by the farmers’ associations. The threat of selling outside Mopti is often sufficient as there are advantages to Mopti wholesalers from larger quantities of clean, uniform millet that the farmers’ associations can provide. Finally, the farmers’ associations in Mopti did well with the other components of the marketing strategy including storage and later sales, producing cleaner millet, and selling in quantity. The farmers’ associations need to do a better market search and to press harder for a premium price for the clean millet if necessary in Bamako. Mopti millet already has a better reputation for being clean while Segou has a reputation for dirty millet The P4P of the PAM is working well for the Segou producers but farmers’ associations there should anticipate at some time dealing again with the normal market channels. They need to be investing now in contacts and information about alternative markets. Specifically for women greater access to organic fertilizer was the primary complaint especially being able to obtain carts and rights to obtain more of the available organic fertilizer. Improved threshers are also needed in the farmers’ associations and would especially benefit women. We have consistently been disappointed with the breakdown of these machines especially those with joint ownership by the farmers’ association. Recommendations: 1. As for sorghum, the access to DAP fertilizer and to certify Toroniou needs to be facilitated. Opposite problems of restricted availability of DAP and inadequate recognition of advantages of certified seed of Toroniou by farmers. 2. Training and mentoring on fertilization, organic and inorganic, timing and placement, advantages of certified seed. 3. Improve marketing ability by guaranteeing clean cereal to processors and being able to negotiate tougher with wholesalers in the post P4P era including the ability to break a wholesalers’ cartel by arbitrage (selling outside the region if necessary. 4. Carts for women

Literature Cited Abdoulaye, T., J.H. Sanders and B. Ouendeba, 2008. Evaluation of Sorghum and Millet Technology and Market Strategy Introduction: 2006-2007 Crop year, INTSORMIL Bulletin No 8 Lincoln Ne 22 pages Baquedano, F., M. Diarra, and A. Ahmoudou, 2009. Economic Analysis of the 2007 Crop Year, INSORORMIL Bulletin No 9 Purdue University, Department of Agricultural Economics, November 2009, 34 pages Coulibaly, J., Evaluacion des Technologie de Production et de Commercialis du Sorgho et du Mil dans le Cadre du Projet IER-INTSORMIL/Mali, Campagne Agricola 2008-2009, Bulletin No. 10, Purdue University, Department of Agricultural Economics, Aout 2010, 36 pages.

65 Coulibaly, J., G. Kumaraswamy and J. H. Sanders, 2013.Economic Impact of Sorghum and Millet Technologies in Mali Agricultural Campaign, 2010-11, Bulletin IER- INTSORMIL n° 11, Purdue University, Department of Agricultural Economics, February 2013, 31 pages. Harmon J., 2015. Consultant’s Report on the Introduction of Millet Technology in the Mopti and Segou Regions of Mali, Washington D.C.

66 Background data from the field interviewing in Bankass and Koro, winter 2016 Table 8. Production- Marketing Results in 2014-2015 Crop Year and Technology Performance in 2015/16 Crop year in Koro, Mopti region

Number of Area Organic Inorganic Production Quantity for Price the Variety Fertilizer Fertilizer 2014 OP sale after Received Farmers’ (2015) (tons) (sacks) Consumption reimbursement (CFA/kg) Association (2015) (2015) (2014/15) (2015) by OP from Buyer (2015) 1. 1 ha. 190 kg 1 DAP, 1 800 kg. 100 kg at 164 164 PAM a Toroniou Urea 26 parents, (April) (April) 200 consumed a 2. 1 ha. Tor 3 tons 35 kg 900 kg 200 kg at 164 164 PAM DAP, 35 20 kg (April) (April) kg Urea parents, 300 kg consumed 3. 1 ha local 3.3 tons 1 sack 800 kg 160 DAP, 1 20 parent, wholesaler- sack Urea 600 Guindo consumed (May) 4. 1 ha local 2.5 tons 1 sack 5 parents, 160 DAP 350 Sold to consumed Guindo (May) 5. 3 ha 5 tons 3 sacks 4725 kg 155 Toronioub, DAP, 3 60 to wives, sold to 2.5 ha local sacks Urea 60 to parents, Guindo 900 (June) consumed 6 3 ha. 21 tons 3 sacks 5966 kg, 300 162.5 to Toronioua, DAP, 3 kg to wives, Guindo 3 ha local sacks Urea 300 to (June) parents, 3500 consumed 7. 2 ha 15 tons 1 sack 1975 kg, 150 1000 kg at 150 150 to PAM Toronioua, DAP, 1 to parents, to PAM (June) (June) 1 ha local Urea. 960 Applied consumed equally to 2 ha 8. 1 ha 2 tons 1 sack 2000 kg 50 500 kg at 180 150 to PAM Toronioua. DAP, 1 kg parents, (May) (April) 10 ha of sack Urea 500 kg local consumed 9. 2 ha Tora 8 tons 2 sacks 2375 kg, 200 160 to and local DAP and 2 kg parents, Guindo sacks Urea 700 kg (May) consumed

67

10. 1 ha Tora, 2 688 kg 1 and ½ 900 kg, 200 kg at 160 160 ha local sacks DAP, 600 kg (April) regional 1 and ½ consumed merchant sacks Urea Poudiougo (april) 11. 1 ha 1.5 tons 1 sack 600 kg, 00 to 200 kg at 160 160 from a Toroniou DAP, 1 wives, 300 (April) Poud. sack Urea consumed (April)

a 12. 2 ha Tor 4.1 tons 2 sacks 2700 kg, 100 300 kg at 125 160 top and 2 ha DAP and 2 kg parents in March. wholesaler local sacks Urea and 600 Maiga consumption (May) a 13. 2 ha Tor 3 tons 2 sacks 3825 kg, 100 200 kg at 115 160 and 5 ha DAP and 2 to parents, in January not wholesaler local sacks Urea 2050 through OP Dodo in consumed May 14. 1 ha 11.4 tons 1 sack 3825 kg with 600 kg at 160 160 in May Toroniouc for Tor and DAP and 1 1200 kg to Dodo, Dodo. and 3 ha 21 tons for sack Urea consumed wholesaler not local locals through OP c 15. 1 ha Tor 12 tons Tor 1 sack 4550 kg with 300 kg at 170 160 to and 5 ha and 60 tons DAP and 89 kg to to Dodo (May) Dodo (May) local local one sack women for Urea local sales 1890 consumed a 16.. 1 ha Tor 3.5 Tor 1 sack 2700 with 50 160 to and 2 ha and 3.5 DAP and 1 to women for Guindo local local sack Urea village sale (May) 1782 consumed a 17. 2 ha Tor 15 tons Tor 2 sasks 3050 with 150 to and 3 ha and 18 tons DAP and 2 500 kg to Guindo local local sacks Urea parents , 415 (April) to women for local sale and 600 kg consumed 18. 3 ha 10.4 tons 1 sack 3500 kg with 150 to Toronioua and 19.2 DAP and 1 1100 regional and 3 ha tons sack of consumed merchant local Urea Poud. (April) 19. 1 ha of 5 tons and 1 sack of 2800 kg and 150 to Poud Toronioua 6 tons DAP and 2 1500 (March) and 4 ha of sacks of consumed local Urea

68

20. 3 ha of 12 tons 3 sacks of 4950 kg with 400 kg in Oct 175 to PAM a Toroniou DAP and 3 200 kg to 125 cfa in March sacks of parents and Urea 2000 kg consumed 21. 3 Ha of 2 tons on 3 sacks of 2700 kg with 175 PAM in Toronioua Toroniou DAP and 3 300 kg to March and 5 ha of and 3 tons sacks of parents and locals on locals Urea 1800 kg consumed 22. ¼ ha Tora 0.4 ton Tor 1 sack 1100 kg 100 kg to OP 175 to PAM and ½ ha and 1.2 DAP and 1 55 kg sold in for 160 (April) (March) local tons local sack Urea village and 100 kg sold 200 kg privately consumed 157 23. 2.5 ha 1 sack 2950 kg and 175 to PAM Toronioua DAP and 1 1200 (March) and 20 ha sack Urea consumed 500 kg to locals buyers at 120 (Feb) a. Toroniou from the certified seed production of 2015 b. Toroniou from own plot probably degraded c. Toroniou from an old INTSORMIL plot. Also probably degraded. OP: Farmers’ association from the French.

Table 9. Marketing Results in 2015 and Technology Performance in 2015/16 Crop year in Bankass, Mopti region

Number of Area Organic Inorganic Production Quantity for Price the Variety Fertilizer Fertilizer 2014 OP sale after Received Farmers’ (2015) (tons) (sacks) Consumption reimbursement (CFA/kg) Association (2015) (2015) 2014/15 (2015) Buyer (2015) 1. 3 ha local. 10 tons 2 sacks of 3600 kg, 300 160 to Toronioua DAP, 1 and for parents, wholesaler received in ½ sack of 1200 Guindo July Urea consumed (May) 2. 1 ha 2.25 tons 1 sack DAP 900 kg, 100 200 kg to OP 160 to a Toroniou and 1 sack kg parents for sale at 160 Guindo Urea and 300 kg (May) (May) consumed 3. ½ ha Tora 12 tons 2 sacks 3600 to 4800 200 kg at 165 165 PAM and 1 and DAP and 2 with 200 to (May (May) ½ local sacks Urea widows and needy and 200 to women for buying condiments, 1800 consumed

69

4. 1 ha local 1 ton 1 sack 1600 kg, 200 for 165 cfa 165 to DAP, 1 with 200 (May) PAM sack Urea parents, 1200 (May) consumed

5. ¼ ha 300 kg 2 sacks 3000kg, 150 160 to Toronioua DAP and 2 parents and Guindo and local sacks urea widows, (May) 1260 consumed, 120 sold in village (women)

6. ¼ ha 0 ¼ sack Not a a Toroniou DAP, ¼ member in sack Urea 2014. New member 7. 2 ha local 6 tons 5 sacks 2000 kg, 500 160 to DAP and 5 kg parents, Guindo sacks 1300 (April) Urea.Used consumed, fertilizer on 200 kg sold other in village activities (women),

8. 2 ha local 15 tons 2 sacks 3600 kg.,100 160 to DAP and 2 kg parents, Guindo sacks Urea 216 kg sold (May) in village, 2200 kg consumed a 9. ¼ ha Tor 75 kg for 1 sack 800 kg, 50 200 for 165, 165 to Tor. 300 DAP, 1 kg parents, PAM (May) PAM for local sack local 300 kg (May) consumed 10. ¼ ha Tora 90 kg for 1 sack DAP 1500 kg., 15 200 for 165, 165 to And 2 ha Tor and and 1 sack for parents, PAM (May) PAM local 600 kg for Urea 400 (May) local consumed 11. ¼ ha Tora 120 kg for 1 sack of 1500 kg, 50 300 kg at 165 165 to and 1 ha of Tor and DAP and 1 for parents, (May) PAM (may) local 675 kg for sack of 400 kg local Urea consumed 12. ½ ha of 4.2 tons 2 sacks of 3600 kg., 160 to Toronioua DAP and 2 200 for Guindo and 1.5 ha sacks Urea parents, 3240 (May) of local for consumption

70

13. 2 ha 4 tons 1 sack of 4200kg,200 169 to Tora and DAP and 3 for parents, wholesaler local sacks of 50 sold in Moulaye Urea village (May) (women) 14. ½ ha 2.5 tons 1 sack of 700 kg, 6 kg 300 kg at 160 160 to b Toroniou DAP, ½ parents, 100 (May) Guindo sack Urea kg sold in (May) village, 100 kg consumed b 15. 2 ha Tor 12 tons 2 sacks 3000 kg, 200 200 kg at 162.5 160 to DAP, 2 kg parents, (may) Guindo sacks Urea 100 sold in ((May) village, 1000kg consumed

16. 1/2 ha of 250 kg for 2 sacks of 700 kg.,50 160 to Tora and 2 Tor and DAP, 2 for parents, Guindo ha of local 500 for sacks Urea 350 (May) local consumed 17. 1 ha local 3 tons 1 sack of 600 kg, 400 160 sold to DAP, 1 kg consumed people in sack Urea village (May) 18. 2 ha of 20 tons 1 sack of 1800 kg., 160 sold to local DAP, 2 100 kg people in sacks Urea parents,1100 village consumed, (May)

b 19. 3 ha Tor 7.5 ton 2 sacks of 3600 kg.,500 160 to DAP, 2 kg parents, wholesaler sacks Urea 270 sold in Timbo village, 1400 (May) kg consumed 20 2 ha local 8.75 tons 2 sacks 2100 kg., 160 to (mixed DAP and 200 kg Timbo with two sacks parents (May) Tor) Urea 21. 2 ha local 5 tons 1 sack 2100 kg., 160 to DAP, 1 Timbo sack of (May) Urea 22. 2.25 Tor 9.9 tons 1 sack of 3000 kg.,400 200 kg for 125 165 To and local DAP and 1 kg parents, (Feb) Guindo sack of 240 sold in (June) Urea village960 consumed

71

23. ¼ ha of 2.6 tons 3 sacks of 3000 kg, 150 160 to Toronioua DAP and 1 parents, 750 Guindo and 3 ha of sack of sold in local Urea village, 2000 consumed 24. 1 ha 2.4 tons 1 sack of 800 kg. 50 165 To a Toroniou DAP kg parent, wholesaler 200 Mopti consumed a. Toroniou from the certified seed production of 2015 b. Toroniou from own plot probably degraded c. Toroniou from an old INTSORMIL plot. Also probably degraded

72 Appendix A-2 Questionnaires Mopti (Interviewing in winter 2015)

Enquête auprès des associations de producteurs du mil : 2015

1. Quel est le nom de votre association ?

2. Quand avez-vous démarré vos activités ?

3. Avec combien de membres avez vous maintenant ?

CRITÈRES POUR L'ORGANISATION DE L' OP UNE BONNE (ET FONCTIONNEMENT ) PRODUCTEUR

1. Est-ce que vos membres rembourser leurs prêts ? Quel represente le pourcentage de remboursement ?

a. 2013 b. 2014 2. Y-a-t-i l des sanctions prévues pour ceux qui ne remboursent pas le crédit ? 3. Est-que vos membres vous permetent de vendre plus de céréales après le remboursement du prêt ? Combien de sacs Qu'est-ce pour cent de la récolte est ce montant ?

2013

4. 4.Que faites-vous pour trouver des marchés pour votre mil ? Qui vendez-vous à ( collecteur , marchand régionale , grossiste, processeur ) .

2013 2014

5. Comment contrôlez-vous la qualité du grain de vous membre?

6. 6.. Que faites-vous quand le prix baisse rapidement dans le prix offert dans la région ?

7. Que faites-vous pour assurer que les membres aient confiance dans votre prise de décision ?

73 Enquete - Ag Performance - Les agriculteurs 1. Au cours des deux dernières années, quelles ont été vos rendements de mil des hommes, des femmes ? 2013

2014

Les niveles de fumiere organic/ha, hommes, femmes?

Les niveles de engrais chemique/ha hommes, femmes?

Les varietes, hommes, femmes? Les superfeis, hommes, femmes ?

Le rendement de pratiques traditionelles (sans engrais chemique), hommes, femmes?

2. Quel prix avez-vous reçu pour votre mil rembourser votre prêt au fonds de roulement ou pour le commerçant ? 2013 2014

3. Quelle quantité avez-vous vendu pour rembourser votre prêt ? 2013 2014

4. Quelle est la quantité de mil avez-vous tous les OP de vendre pour vous ( après remboursement ) ? 2013

5. Quel est le prix que vous ont été donnés pour cette quantité vendue par l' OP ? 2013

6. Avez-vous vendre un de vos mil vous meme ? 2013

7. Quel prix avez-vous obtenu pour les ventes de ce mil ? 2013

74 STRUCTURE DU MARCHÉ 8. Pour les ventes de mil qui achete le mil ? ( négociant, distributeur d'entrée , la transformatrice, collecteur ou quelqu'un d'autre )? 2013 2014 9. Conditions du contrat – 2014 Quelle est la date de vente ? Prix du marché ? Prix du marché majoré d'une prime ? Valeur de la prime ?

2013 Quelle est la date de vente ? Prix du marché ? Prix du marché majoré d'une prime ? Valeur de la prime ?

10. Avez-vous payé les frais de transport ?

2013 - Combien? 2014 Combien?

Cereale Propre- Les Paysans et les OPs 1. Comment est-que vous nettoyez votre cereale ? Quand vous coupez les epis de mil vous mettez sur les tiges Battage sur bache Ou vous utilez le tami Autre 2014 Qualite 2. Avez-vous recu un surprix pour le mil propre ? Combien ? 2013 2014

75 3. Comment est-que le OP peut controller le qualite (cereal propre) de son membres 2014

Le Cout de Credit-OPs (avec et sans un personne tierce, commerçant, vendeur de engrais, transformatrice)

Avec Contrat Sans contrat

Quand vendu

Prix de vente

Quantité vendu Seulement pour payer le prêt

Quantité vendu Plus que le prêt. Combien Globalment 1.Qui dans l’association a été chargé l’année passée de revendre le sorgho/Mil collecté ? 2 Quelles sont les raisons évoquées par ceux qui n’ont pas encore remboursé leur crédit ?

3. Comment encouragez-vous les producteurs à vendre leur sorgho/Mil par l’association ? 4. Quelles sont les difficultés majeures rencontrées par l’association? 5. Quelles ont été les points positifs qui ont renforcé l’association ? 6. Quelles sont les résolutions prises par l’association pour améliorer le fonctionnement le future ? 7. Quels critères pensez-vous sont nécessaire pour une très bonne OP ?

76 The Mopti Field Program, 2015-2016 Crop Year John Sanders, Soungalo Traore, Niaba Teme, and Ababacar Ndoye Introduction To respond to technology erosion and encourage institutional evolution of the farmers’ organizations involved with millet production the project provided funding to a field program in the summer of 2015. Our farm level work has shown a substantial yield and economic response to one sack of Urea and one sack of DAP per ha. There was widespread farm level demand from the farmers’ associations involved in the scaling up of millet technology in the Mopti region to obtain DAP to substitute for NPK.1 There was also farmer support for renewing the Toroniou seed. So in this Gates supported field program of 2015 the extension service identified farmers’ association (2,898 farmers; (Table A-1) that wanted to use this combined fertilization level (one sack of DAP and one sack of Urea) and to buy certified Toroniou seed at the expected seed price of 500 cfa/kg. The inorganic fertilizer was to be financed from the revolving fund of the farmers’ associations and the seed to be paid for directly by the farmers. Program participants in the farmers’ associations agreed to buy this fertilizer combination with the revolving funds and repay in kind after the harvest at the price of millet to be set by the farmers’ association. The subsidized fertilizer would be available at one ha per male farmer and ¼ to ½ ha per female farmer. The fertilizer request was for 3,186 ha. So some farmers would use more than one ha in the program. Women in the program had complained that their main problem on the small areas they were allocated for personal use was their inability to obtain organic fertilizer and the difficulty of transporting it as they were the last to have access to the carts. Moreover, their allocated areas were often far from the village.2 So the program also included financing carts for women’s associations as well as mentoring and training in the improved agronomy (Picture 1).

1 After the first year of this IICEM program there was substantial substitution of one sack of NPK for one sack of DAP because the farmers’ associations could not obtain the DAP. Public policy attempted to reserve DAP for rice in publicly supported programs 2 Women are given annually small areas to farm for themselves after they have worked on the communal area and done their other household and child caring activities. On their private plots the women control the returns.

77

Picture 1. Alternative to Carts for Carrying Manure, Millet Fields. Mopti Region 2015

Program Implementation The fertilizer arrived at the “cercle” extension regions of Bankass and Koro for distribution to the farmers’ associations. Almost all of the highly valued DAP was received and distributed by the farmers’ associations. However, 56% of the Urea requested by Bankass and 11% 3requested by Koro “disappeared.” This is a problem with a subsidized input that it can be very profitable to resell. The regional director of DRA in Mopti arranged for an increased supply but it was late in the crop year when this additional fertilizer became available. Several farmers’ associations in Bankass no longer wanted as much Urea. Others utilized the fertilizer too late for it to be very effective (N. Teme, 2015, pp. 10, 11) In Koro the combined final use of DAP and Urea was 199 tons and in Bankass 106 tons as compared with 338.6 requested in interviews with the management of the different farmers’ associations (Tables A-1. Table A-2 and Table A-3).In general the program functioned very

3 The situation in Koro was different. 140 sacks were not originally provided to Briga Dogon due to missing papers about eligibility for the subsidized fertilizer.

78 efficiently in providing the promised fertilizer levels by obtaining and distributing the fertilizer and responding to the “disappearance” of the Urea. Millet outcrosses rapidly so Toroniou needs to be replaced regularly (at least every three years). IICEM introduced some certified Toroniou in the first year of the program and encouraged Toroniou seed production from local seed producers4 in the following three years. But many of the project farmers continued with their local millets especially in Bankass. In 2015 there was a supply expansion of certified Tornoiou produced in the region as several agencies began promoting the use of certified seed. In the Koro and Bankass “cercles” 66 tons of certified Toroniou was produced in 2014 (Table A-4). In nine villages of Koro and one in Bankass there was certified seed production. This is sufficient seed to plant 11,025 ha (at 6 kg/ha) to 13,320 ha (at 5 kg/ha). The expected demand was for 19.1 tons for the project. Rather than the individual farmers or associations buying the certified seed it was primarily bought by a seed company and by various NGOs. Faso Kebe bought 48% and CRS bought 25%. Both the seed company and the NGO operate nationally. Certified Toroniou was available in the region and the farmers were informed. They had committed to buying it. Now the plot becomes complicated because both the Catholic Relief Service (CRS) and the international fertilizer development corporation (IFDC) became involved with their own programs. The CRS wanted to do small scale demonstration of technology but was offering free DAP. The IFDC fertilizer recommendations included both DAP and Urea but at lower levels than our recommendations. Not surprisingly most farmers opted for the donations. In Bankass 520 kg of certified Toroniou was received and in Koro 1250 kg from CRS and IFDC (N. Teme, 2015, pp. 6-8). In Koro two OPs (Organizations of Producers) bought 780 kg of the certified seed produced in the region. From the fieldwork only 9% of the average village area in Bankass was planted in certified seed. In Koro the average village area estimate was for 36% in certified Toroniou seed but our fieldwork only included 7 of the 12 OPs as with the flooding it was difficult to get access to all the OPs. (Tables A-5 and A-6) So the problem is not input supply but demand. The use of certified seed was much more erratic than that of DAP and Urea. In Bankass of the ten farmers’ associations 7 obtained the seed late and only 3 on time. Only four of 12 farmers’ associations interviewed in Koro got certified Toroniou seed early (N. Teme, 2015, p.6). Toroniou is an early cultivar so it can be planted later. However, much of the seed was obtained later in July so Toroniou would have been relegated to the poorer land area and labor requirements put first on the earlier planted local millets and other crops. The West African Productivity Program (WAPP) brought 39 tons of certified Toroniou to facilitate this program but this seed arrived even later in August.If stored properly it will be useful for 2016 (N. Teme, 2015, p. 6). The use of certified seed was much greater in Koro than in Bankasss (Tables A-5, A-6). The common feature of these estimates is a disappointing use of certified seed especially in Bankass.

4This is selection of the best plants rather than certified seed production.

79 The reasons given for some farmers not buying the certified Toroniou was that they were waiting for free seed from CRS and others felt that 500 fcfa/kg was too high a price. So supply of quality seed has been responded to in the region. Moreover, the high prices for millet grain with continuing purchases by PAM should have increased the farm level demand for this improved seed. However, farmers still need to be convinced that this is a good investment. The yield data from this project for the three types of seed, certified Toroniou, degraded Toroniou, and the local varieties do not show much difference.5In Bankass there was not much difference between the three with 1508 kg/ha for the certified Toroniou, 1445 for the degraded Toroniou, and 1293 for the local (Table A-7).6 In Koro the differences were 1117 kg/ha for certified Toroniou, 966 for degraded Toroniou, and 947 kg/ha for the local (Table A-8). Taking out the cases of very low yields attributed to flooding raises the yields of certified Toroniou by only 60 kg/ha for the certified seed area in Koro. So there are other reasons for the yield disparities besides the seed quality. In both regions all three treatments were using the organic and inorganic fertilizers and mostly following the other agronomic recommendations. The big difference between local practices and all three treatments in Bankass is that the fertilization and improved agronomy roughly doubled yields of the traditional millet production (600 kg/ha). The effects of fertilization and improved agronomy were still substantial in Koro even with the flooding. So first it is necessary to increase fertilization of both inorganic and organic fertilizers. Then the improved seeds make a difference but a much smaller one than the fertilization and agronomy. If the only systematic difference between these three treatments were the type of seed, we can make the simple profitability comparisons of the differences from the seeds alone. As illustration we take the very small yield difference between certified and degraded Toroniou in Bankass of 63 kg. Additional seed cost is 6 kg/ha of seed times 500 to 750 fcfa/kg (the range of seed prices asked and sold in the region).This gives incremental costs of 3,000 to 4,500 fcfa/ha. Farmers’ associations received around 160 cfa/kg for their millet in 2015. At this price the incremental benefits of certified seed are 63 kg times 160 cfa/kg or 10,080 fcfa/ha. Even at a substantially lower price of millet in 2016 the profitability of utilizing better seed is clear with the qualification of first applying the two fertilizers and using the improved agronomy. With the differences in soil quality and access to organic fertilizers the men consistently out-yielded the women. In Bankass the overall difference was 130 kg/ha and for the certified seed 214 kg/ha. In Koro the overall was 165 kg/ha and for certified seed 148 kg/ha. This is very good performance for the women who have poorer soils and less access to inputs including both

5 Note that yields also depend upon a number of other agronomic factors that are not held constant here. This includes actual use of organic and inorganic fertilizers including timing and placement, crop history, soil quality, planting time, and accuracy of measurement of the crop cuts. The usual practice with crop cuts is to ask the local agents to collect very large numbers of observations so the data quality is reduced. Here we asked for four farmers per farmers’ association and three cuts per farm. We also asked for sampling of male and female farmers and sampling by the three types of seed utilized in each village (see Tables A-7 and A-8). 6 In Bankass there were significant differences between yields of certified seed and the locals at 99% but no significant difference between certified and degraded Toroniou. There was a small sample of 19 observations.

80 organic fertilizer and the mentoring. When they receive training and/or mentoring, the women follow better the agronomic recommendations. The principal request of the women to improve their productivity was for carts to carry the manure to the field. Very simple carts were provided for the womens’ associations with the condition that they be paid off after four years so another cart could be purchased for their association. Twenty-five carts were provided in the winter of 2016 in Bankass (10) and Koro (15) with Gates Foundation funding (Picture 2). Each cart was to be shared by five women and the larger womens’ associations got more carts.

Picture 2. New Cart Being Presented by Soungalo (in white scarf) to the Womens’ Association in Koro, Mopti region, 2016.

Conclusions: The scaling up was successful with the fertilization with the qualification about the disappearing Urea. The Toroniou seed demand is still an important concern as the supply of certified Toroniou in the region has been substantially increased. Toroniou is early and sensitive to low soil fertility but certified Toroniou has doubled yields of traditional millet when combined with moderate fertilization and improved agronomy.

A shift in public concern to the basic staple cereals of sorghum and millet has only occurred in Mali since 2010 and the official recommendation still only includes a minimal amount of fertilizer (35 kg of NPK) when the response to more fertilizer and especially DAP has been demonstrated all over the country for millet and sorghum. How to change long standing conventional wisdom that millet and sorghum do not respond to or are not profitable with moderate doses of inorganic fertilizer is the continuing dilemma.

81 A related problem is the collaboration between development agencies. Now there is a demonstrated successful program in the Bankass and Koro “cercles” with farmers’ associations repaying bank credit and establishing revolving funds for input purchases. Hence, many agencies have become involved in working with these farmers’ associations including IFDC, ICRISAT, CRS, and now AGRA. This should be a good thing with more capital and human resources being involved. Unfortunately, there is little coordination and as discussed sometimes the programs actually impede the objectives of other programs.

A continuing problem in development in sub-Saharan Africa has been the expansion of the extension services with outside funding. Hence, the extension service became dependent upon NGOs and donors for operating funds. When one outside supporting agency departs, another or several enter and there is little concern with program continuity as each new agency involvement has their own priorities.

Aiding the women has been a principal priority of this Mopti project. However, the women still have little access to good land near the village or much time to undertake their farming activities. But they did get the fertilizer inputs from the project and the carts were their main request. They need more help to reduce the time demands of their household activities and more training and mentoring to accompany the fertilizer.

Recommendations:

1. With the appropriate fertilizer now being utilized on a large scale and large supplies of certified Toroniou in the region farmers need to be again reminded of the importance of combining the renewed seed with the soil fertility-improved agronomy combination. 2,859 ha is a good base for a further rapid expansion of this technology-marketing program in this low income, poor resource base Mopti region. 2. The training on adequate fertilization and the potential for good income earning with the combined seed-fertilizer-marketing package for millet needs to be communicated better to public officials. 3. Funding sources shift regularly and the dependence of the extension service upon outside operating funds makes program continuity difficult. This is not helped by the usual competition or lack of communication between agencies with field programs. 4. Increased focus on the specific requirements of the women for increasing productivity especially increasing access to resources (good land, organic fertilizer, training-mentoring, labor saving innovations for household activities).

82 References: Teme, Niaba, 2015. “Adoption of Toroniou Millet Technology, Production and Marketing in Bankass and Koro, Mopti, Mali,” mimeo, IER, 10 pages.

83 Table A-1. Initial Demand Indicated for the Fertilizers

Villages Bénéficiaires Fertilizers (sacks of 50kg) Toroniou Seed Men Women DAP Urée (kg) Koro Togo Tina 73 39 142 142 852 Pomorododiou B 102 48 146 146 876 Pomorododiou Na 72 33 178 178 1068 Temegolo 148 13 156 156 936 Kountogoro 150 30 180 180 1080 Briga Dogon 143 06 148 148 888 Pel 52 60 100 100 600 Balirou 69 51 242 242 1452 Yadianga 92 45 142 142 852 Tinassansagou 60 40 100 100 600 Tendeli 197 100 300 300 1800 Didia (Amaga) 73 33 162.5 162.5 975 Didia (Sindiéré) 40 15 39.5 39.5 237 Bankass Logo 0 275 281,5 281,5 1669 Dimbal 81 79 160 160 960 Sadia Dogon 100 80 120 120 720 Telly 49 5 67 67 402 Tinto Barwé 32 80 117 117 702 Dianwely 54 26 35 35 210 Ogodiré 75 26 70 70 420 Ogossagou 106 31 125 125 750 Ogotena 50 25 75 75 450 Kanikombolé 100 100 600 Totals 1818 1080 3186 3186 Source: unpublished data from the regional extension offices of Koro and Bankass. 84 Table A-2. Requests and Distribution of the InorganicFertilizer in Bankass, Mopti. DAP (sacks 50 kg) Urea (sacks 50 kg) Supply of Area in 85 Villages the late Total Urea the Urea project Received P R P R (ha) Dianwely 35 66 35 66 66 35 Ogodire 70 130 70 10 60 70 70 Ogossagou 125 80 125 0 80 80 80 Ogotena 75 160 75 0 40 40 75 Tinto-barwe 117 130 117 130 130 130 Dimbal 160 60 160 60 60 60 Logo 281.5 300 281.5 0 160 160 160 Sadia Dagon 120 120 120 0 80 80 100 Telly 67 94 67 94 94 67 115 115 Kanikombole -a 115 - a 115 Total 1050.5 1255 1050.5 475 420 895 879 a. Not included in the original surveying for participation P: Requested; R: Received Source: Unpublished data provided by Soungalo Traore from interviews with the regions extension offices, summer 2015 Table A-3. Requests and Distribution of the Inorganic Fertilizer, Koro, Mopti

Villages DAP Area in (sacks 50 kg) Project(ha) P R P R Togo Tina 142 143 142 125 190 Pomorododioun-Begne 126 126 126 126 54 Pomorododioun Ma 178 172 178 172 59 Temegolo 156 150 156 150 208 Kountogoro 130 130 130 130 173

Briga dogon 148 140 148 0 (140)a 197 Pel 100 100 100 100 107 Balirou 242 200 242 200 210 Yadianga 142 140 142 140 93 Tinassansagou 100 100 100 100 50 Didia (Amaga) 162.5 130 162.5 130 108 Didia (Singlere 39.5 70 39.5 70 52 Tere 100 100 100 100 100 Tendeli 300 300 300 300 373

Total 2066 2001 2066 1843 (1983)a 1974 a .Papers authorizing subsidized purchased were lost but problem resolved in August, 2015. P: Requested R: Received Source: Unpublished data provided by Soungalo Traore from interviews with the regions extension offices, summer 2015

86

87

Table A – 4 Production and sale of certified Toroniou seed in Koro

Quantity (t) sold in 2014 to Production Villages Afrique In Stock {t}2014 Faso Kaba AGECOM(NGO) CRS Verte ICRISAT Others Tendely 21,3 10 1,7 3,55 0 0 6,05 Balirou 17 12 4 1 Djidia 0 Bondé Tena 20 Odjioutanga 0 Nambrin 0 Tagare 7,85 4,15 3 0,7 Koporo Pen 0 Koporo Na 0 Sogara 0 Total 66,15 22 1,7 11,7 3 1 6,05 0,7 Source: Unpublished data from field interviews in Koro, August 2014 Table A-5: Structure of the Farmers’ Associations by Number of Producers, Quantity and Area in Certified Toroniou in the Project in Bankass.

Number and Gender of Number of ha Percentage of ha Certified Toroniou Number of ha producers per Farmers’ of certified of certified seed in Seed received by In the Farmers’ Villages/OP Association seed in the farmers’ Participating Association in the projecta associations in the Total Men Women Farmers project projectb Ogodire/Densa 101 75 26 30 70 6 9 Dimbal/Ogokana 160 81 79 98 160 19.6 12 Dianwely/Amakene 80 54 26 54 130 10.8 8 Tinto /Barwe- Degoubere 112 34 78 93 130 18.6 14 Logo/ Badjina-Bara 275 0 275 79 282 15.8 6 Telly/Amaga 54 50 4 20 70 4 6

Ogossagou/AmakoKene 137 106 31 0 68 0 0 Sadia Dogon/Kantiemu 180 100 80 73 120 14.6 12 Kanikomole/ Gombere Gue 215 135 80 0 115 0 0 Ogotena/Sabougnouma 75 52 23 73 75 14.6 20 Total 1389 687 702 520 1220 104 9 Average 138.9 68.7 70.2 52 122 10.4 9 Source: Niaba Teme, “Adoption of Toroniou Millet Technology, Production, and Marketing in Bankass and Koro, Bamako,Mali, mimeo, IER, Bamako, Mali, September 2015, p. 22. a. Calculated on the basis of 5 kg/ha b.Column 7 divided by column 6

88 89 Table A-6: Structure of the Farmers’ Associations by Number of Producers, Quantity and Area in Certified Toroniou in the Project in Koro Number of producers in Number of ha Percentage of ha Certified Number of total the farmers’ association of certified of certified seedb Tororniou seed ha in the participating in the seed in the received by the farmers’ Villages/OP project farmers’ farmers’ association associationa association participating in Total Men Women project; Djidia / Amaga 105 60 45 680 165 136 82 Briga Dogon 8 5 /Monrinbima 149 143 6 40 149 Kountogoro /Merekounon 180 150 30 130 Pel 60 46 Maoudé/Amadomon 110 50 60 300 130 Yadianga/Merebara 160 50 110 300 160 60 38 Pomorododiou- Na/Amadjigue-Juru 105 72 33 300 172 60 35 Pomorododiou- Begné/Kamonon 186 106 80 n.a. 72.5

Temegolo/Doumnon 156 148 8 n.a. 156 Togo Tina/Amagana 143 121 22 n.a. 143 Tina Sassogou/Amakene 100 60 40 n.a. 100 Balirou/Deguebere 120 69 51 310 216.5 62 29 Tendely/Merebara 425 200 225 100 325 20 6 Total 1939 1229 710 2030 1919 406 21 Average 162 102 59 290 160 58 36 Source: Niaba Teme, “Adoption of Toroniou Millet Technology, Production, and Marketing in Bankass and Koro, Bamako,Mali, mimeo, IER, Bamako, Mali, September 2015, p. 22. a.Calculated on the basis of 5 kg of seed per ha. b.Column 7 divided by column 6 Table A – 7 Farmer Yield Estimates by Seed Type with Fertilizers and Planting Time Date (Bankass, 2015 – 16 Crop Year)

Available on request as listed by farmers’ names for our cross checking and would need to take that off

Table A – 8 Farmer Yield Estimates by Seed Type with Fertilizers and Planting Time (Koro, 2015 – 16 Crop Year)

Available on request as listed by farmers’ names for our cross checking and would need to take that off

90 Secondary Markets

91

Demand for Millet Processed Products in Mali and Burkina Faso Ababacar Ndoye,1 Botorou Ouendeba, and John H. Sanders Abstract Cereals pass from staples to processed foods and feeds with economic growth. The demand for cereals and tubers substantially declines with economic growth as high value products including animal products, fruits and vegetables become much more important in the diet. However, for the short run (next five years) rapid population growth is expected to be the principal factor maintaining demand growth of the cereals. Moreover, as the prices for the processed traditional millet products are reduced 30% from their present high levels, the increased consumption by lower and middle income consumers would double the demand growth for these processed products over the next five years. On the farm technological change enables a reduction in costs per output unit and therefore prices can decline with many farmers still making money. However, with a too rapid price decline farmers are discouraged from making the investments and increased expenditures necessary for the added inputs associated with technological change. Moreover, the substantial rainfall variation in the Sahel combined with the lack of secondary markets can lead to price collapses in good rainfall years. Therefore, an important component of introducing new technology for millet and sorghum is developing the secondary markets. The primary market is the unprocessed threshed grains. The secondary market reduces or avoids these price collapses. We are investigating the actual and the potential growth of two of these secondary markets, millet food processing and the use of sorghum in animal feed. In this paper we are focusing on the first with the provision that there is some substitution of sorghum for millet especially in poor rainfall years but a general preference for millet over sorghum in the urban 2 areas of the Sahel. In spite of the small size of the millet food processor sectors in the urban areas in both Burkina and Mali relative to the extent of millet food production in these countries there has been substantial investment by donors in machinery and hygiene training.3 So this sector has evolved and produces high quality products for the diaspora of people from the Sahel in Europe and the US, for high income domestic consumers, and for special occasions for middle and lower income consumers in Mali. For this sector to evolve rapidly millet food processors need to reduce costs so that they can obtain increased consumption by low and middle income consumers. Secondly, expansion of the number of farmers’ associations producing clean millet

1 Former Director General of “Institut de Technologie Alimentaire” Dakar - Sénégal , Member of the National Academy of Science and Techniques (ANSTS) , B.P.4344 Dakar - Sénégal 2 With the substantial increases in maize yields during the last two decades with new technologies in the cotton zone many consumers are shifting to the lower cost maize for their “tô”. 3 This sector predominantly includes female entrepreneurs and substantial local female employment for the cereal cleaning.

92 can enable a reduction in the number of stages in the marketing process and thereby reduce costs 4 and prices. So we document the potential for these two sources of cost reduction and then estimate how much this could facilitate the growth of this millet processing sector. Then we compare the effects of these price declines with the effects of population and income growth over the next five to ten years. There are substantial income, nutritional, and health benefits to low income consumers from the evolution of this secondary market as they increase access to lower cost processed millet products. Moreover, the shift from disk mills to hammer or roller mills would have important health effects on consumers by reducing or eliminating the metal residues in the millet flour. Finally, on the horizon is a potential major shift in this demand from bakers for millet flour with the availability of technology improvement for increased millet flour shelf life. In the conclusions we review the various measure to facilitate this evolution of the millet food processing sector in the Sahel. Introduction The objective of this report is to estimate the present and the potential impacts that this sector can have on the demand for millet in Mali and Burkina Faso. We will first consider the importance of millet products in the diet and then summarize the present state of this sector. An important determinant of the growth of the millet processor sector is the increasing availability of clean millet. An increased supply of clean millet with the evolution of the farmers’ organizations can lead to both reduced purchase costs and to a consolidation of the market structure with farmers and processors receiving larger shares of the marketing margin. More efficient firms and improved market structures can result in cost and price decreases. These price declines can enable market expansion, increase the demand for millet, and thereby raise farmers’ and processors’ incomes. A quantitative section will estimate the effect of the price reducing changes on demand growth expected from the cost reductions made possible from technological and marketing changes. Moreover, the demand analysis will include the effects of income and population growth. The conclusions outline the strategy to obtain these gains in the production and marketing system. A final section proposes some measures to achieve this objective of reducing the costs in the systems enabling price declines and thereby increasing the ability to compete with rice and with neighborhood women buying less processed millet and having the bran removed and the milling into flour done locally. Millet Consumption in the Neighborhoods Millet is not only a basic cereal staple it is also made into a series of other products for which there is substantial demand in Burkina Faso and Mali. So here we review the millet consumption pattern in the neighborhoods for three income classes. Rather than going into a full consumer survey we interviewed the millet processors about consumption patterns in their

4 Note that this consolidation of the marketing sector would also enable farmers through the farmers’ associations to obtain a larger share of the marketing margin with further incentives for increased technology use. But there would be some contesting between processors (or wholesalers) and the farmers’ associations over these reduced costs in the marketing process.

93 neighborhoods. They should know present consumption trends because their sales depend upon this. The low income sector eats millet products 1 to 3 times per day (see Table A-1). But it was estimated that only one-half of the low income households in the sample could buy these processed products. There are also a range of other products made from millet that people eat occasionally. So this occasional consumption of other millet products and the preferences that people have for the processed products, such as various kinds of “grumeaux,” “degue” and couscous (see product descriptions in A-2), indicates a strong potential demand as incomes increase. To facilitate this demand growth the processors need clean millet. Demand for Clean Millet Mali Millet processing requires an increasing supply of clean millet. Consumers want clean millet without sand, dirt, or rocks. Dirty millet results from cutting the heads and putting them on the ground, threshing on the ground,5 or deliberate activities to adulterate the grain. Processors want clean millet to protect their machines, to save time and expense in the cleaning operation, and to maintain their reputations for producing a quality product. In Mali and Burkina there is a focus on the demand for higher quality millet cereal and this is changing the marketing system. The sector of processed millet is still small6 but is having a role in these changes. Moreover, there is an important evolution in the demand for processed millet, traditional products. Presently these products are marketed principally for high income consumers especially the diaspora in France and the US. Also in the neighborhood (“quartiers”) there is a demand for these products for special occasions. Now in 2014 in Bamako, there is a focus on buying a higher quality product (fewer impurities) and processors are willing to pay a price premium, over 20 cfa/kg. Even the merchants (“commerçants”) interviewed expressed concern about a quality product and one merchant noted that he gave refunds to his processor buyer for the impurities she separated out in the cleaning process. Moreover, in the rest of the value chain for millet marketing there are adjustments reflecting the greater appreciation of quality of the millet supply. For example, one of the collectors for Misola (infant nutrition firm) and OPAM (Office des Produits Agricoles du Mali- the national grain storage program to moderate the price swings of the staples) has been

5 The traditional activity of mortar and pestle can also keep the millet clean. But as production and productivity increases the women doing this operation get overwhelmed and there have been shifts as in the Segou region to threshing by running over the millet with trucks or tractors. 6 Of the 28 Malian firms interviewed in 2014 they were only processing 797 tons of millet. Adding in the infant food maker (Misola) and the food aid buyer for the United Nations, (PAM), gives another 10,492 tons (Ndoye, Ouendeba, and Sanders, April 2015, pp.2,14-16). The sector is very small compared with the 1.4 million tons of millet produced annually in Mali (Jazayeri, 2014, p. 16). From this total supply 20 percent (280,000 tons of millet) is marketed to urban areas (Jazayeri, 2014, p. 2). But only 3.7% of the millet entering the urban market is passing through these millet processors and that is also including infant food and the relief food purchases.

94 providing new sacks and riding in the trucks transporting the grain to prevent the theft and 7 adulteration often occurring in the transportation process.

Maximum quality premium when buying clean grains (fcfa/kg)

50

40

30 Bamako 19 Firms

20

Segou 7 Firms

0 19 N° de l'entreprise

Figure 1. Premium prices (fcfa) per kg paid for cleaner millet by Bamako and Segou millet processors Source: Field interviews September 2014

Burkina Faso In Burkina Faso the millet processing sector is much smaller and less developed than in Mali. The enterprises are almost entirely of the local neighborhood variety.8 The concept of clean seed is known in the farmers’ associations and by the processors. However, it is much more expensive for the millet food processors to obtain clean millet than in Mali. The principal hypothesis for the difference is the failure to introduce new millet technologies in the principal production zones of millet in Burkina in contrast with the rapid introduction of this technology in the principal millet production zones of Mali.9 Without an increased marketed surplus there is little clean millet available hence higher prices for it. Moreover, the lack of the clean millet means that the market structure still stays more complicated with a lower share of the marketing margin for farmers. In Burkina Faso the collectors had an important role in aggregating sufficient quantities of clean millet for the processors (see the market structure section below).

7 In another illustration of the importance of preventing adulteration and theft in the transportation process we note the example of Thiare, Senegal where the farmers’ association always sends a farmer to accompany the truck to Dakar with their clean sorghum to prevent both. This farmers’ association is a leader in both modernizing the production and building marketing ties with the food processors. They have regular sales of 20 tons/month of millet to the food processors’ association of Dakar. 8 See the Mali country paper (#1) on millet processing for a separation of types of firms. 9 In Segou and Mopti, Mali an intermediate height millet was combined with two to three sacks of inorganic fertilizer to obtain yields 400 to 600 kg/ha higher than local cultivars. This was combined with various marketing techniques of the farmers’ associations. These included storage and later sales, selling in larger quantities to later stages of the marketing chain, getting a price premium for clean cereal. In Burkina the same tall cultivars with little or no inorganic fertilizer do not give an increased marketed surplus.

95 Improvements in the agricultural technology are expected to drive changes in the marketing and in the processing sector. If we compare the price processors had to pay to get clean millet with the price in the villages for ordinary millet (Figure 2), we get a price premium similar to what the processors in Mali had to pay to the farmers’ association to get clean millet. This price premium was generally in the 15 to 25 cfa/kg range in Mali. In contrast in Burkina Faso half the millet processors bought either from collectors or other merchants in the marketing chain. So the price processors had to pay for quality ranged from 37cfa/kg at harvest to 50 cfa/kg four months later and 51 cfa/kg eight months later.

Figure 2 : Prices of clean and dirty millet during 2014 in Burkina Faso

250

236 200 217

187 185 150 Purchase 167 Producers price 150 Processors CFA F/Kg 100

50

0 Harves0t price Price 40months Price 80months after harvest after harvest

a Mean prices for the clean millet were from 17 firms. The prices for the dirty millet bought in the village were obtained from eight firms. Source: unpublished field data from the interviews

Competition for the Millet Processors: What are the products that the millet food processors need to compete with? There are two sources of competition for the millet processors. First millet products have to compete with other cereals and secondly they have to compete with consumers buying the unprocessed millet and making the processed products themselves using neighborhood dehullers and milling machines. The lowest cost cereal staple is the broken rice at around 300 cfa/kg in both countries (Figures 3 and 4). All millet products begin with the flour including the basic dishes of most of the rural population, “tô”, and “bouillie”, a thick porridge for a breakfast cereal. Note that the millet flour presently retails at about 2 and 1/2 times the price of rice in both countries.

96 Figure 3. Average retail prices of broken rice and millet products in Bamako, Mali, 2014

1000 Product 400 750 750 prices 500 CFA F/kg 200 305 312

0 LOWER COST LOWER COST RICE 25% MILLET FLOUR PROCESSORS PROCESSORS FLOUR

Figure 4. Prices for broken rice and the principal millet products of the millet processor in Burkina Faso

1400 1266 1126 1020 1073

Retail 750 prices in F CFA/kg 300

Hence, substantial price decreases are necessary to move millet processed products from a higher income and special occasion products to capture more of the regular market for low and middle income consumers. But note that in Mali some of the millet processors have been able to reduce their costs of production by 1/3 for millet flour and 25% for couscous. In Burkina Faso we did

97 not observe this type of cost savings, nor did we observe a sub sector of processors trying to produce at lower costs for the domestic market in Burkina Faso. Millet processors also have to compete with the households buying the raw millet. Households can buy and process (dehull and mill) millet into flour for about the same price as the broken rice. Again there is the comparison of 500 cfa/kg for the flour in Mali produced by the lower cost processors but still 750 cfa/kg in Burkina Faso (Figures 5 and 6). Hence, it is clear why there are few purchases of millet products by low and medium income consumers. These techniques for cost savings in Mali can be followed by other processors in both Mali and Burkina Faso. Nevertheless, the millet flour price is still double that of the housewife, who purchases the raw millet and has it processed in the neighborhood. So we need to search for further cost reduction when we consider the market structures. First let’s look in detail at where the cost savings come from.

Figure 5. Malian Millet processing Prices of the Cost Reducing Processors v. Housewife Purchased Millet

Millet Purchase Neighborhood cost for Flour Retail Flour Price Couscous

700

500

260 240 200 180 Third Classs of Firm Housewife Purchase

98

1200

1000

750

Millet Purchase

CFA/kg

F

600 Cost for flour

en

Retail price millet flour

Prix 400 299 260 252 217 Retail price millet couscous 200

0

Firms Housewife Purchase

Figure 6. Burkina Millet Processors’ Prices Competing with Household Buying the Raw Millet and Processing it Themselves

Cost Reductions and Market Structures: Here we consider first the method used by these Low Cost Processors in Mali to reduce their costs and prices (Ndoye et al. 2014 for the Mali background paper). Then we give an example of the potential cost savings from changes in the market structure. Mali has a small group of low cost processors producing for the domestic market and often furnishing the neighborhood processors with the dehulled millet. These processors are able to do this with the greater availability of clean millet and the recognition of the need to reduce costs to be competitive with broken rice and with households doing their own purchases In Segou a new machine has been introduced that can clean 30 tons/day so it offers the potential of providing millet processors with very clean and uniform millet in large quantities (Picture 1). Note the rocks and other debris separated by this machine (Picture 2). The extent and

Picture 1: The cleaning machine in Segou, 2014

99

Picture 2. Debris from the cleaning operation, Segou. 2014 the size of the rocks indicates the other problem of reducing impurities when the collectors or farmers adulterate the millet. Cleaner millet from the farmers’ association sold for 125 cfa/kg (interviews with processors from Segou, Sept 2014). This wholesaler with the cleaning machine paid his collectors 175 cfa/kg). A saving of 50 cfa/kg could have been partially passed on to processors and then partially passed on to consumers. This is the type of market structure reform that needs to be accelerated to attain the second cost savings in the system by removing the number of stages of marketing and enabling the farmers’ associations and processors to share parts of these marketing . The second cost reducing factor is the simplification of the marketing process. The marketing process has been Farmers to Collectors to Wholesalers to Processors or to Retailers and then to Final Consumers (adapted from Jayazeri, 2014,). With the development of Farmers Associations introducing new millet technologies, aggregating the cereal into large volumes and cleaning it, the skipping of the collectors and regional markets and selling directly to the wholesalers is the natural evolution of the market structure (for evidence of this process in Mopti and Segou see Sanders et. al., 2015, pp. 8-10). The biggest gains though are when the millet food processors buy directly from the farmers’ associations. Then the wholesaler can also be avoided. This requires that the processors expand more rapidly to take advantage of higher millet yields and greater supplies of clean millet. Moreover, the farmers associations will need to assure the processors the quality of their clean millet. In summary the new marketing system can be Farmers’ Associations to either Wholesalers, who perform critical functions of cleaning on a large scale or the clean millet can be provided by the farmers associations directly to processors. So the marketing structure in Mali has various options. The small sector of the Lower Cost Producers in Mali for the domestic market has been able to expand by taking advantage of the increasing availability of clean millet. The high prices paid by the wholesaler with the cleaning machine for the adulterated millet shows the potential for further gains from improving the quality of the millet received. With the increasing

100 importance of the farmers’ associations in providing clean millet the role of the collectors will be decreased. This will enable cost savings for the processors and higher prices for farmers in the farmers’ associations. Obviously there will be competition over these gains from the cleaning and the market consolidation process.

In Burkina Faso the market structure is very similar except for the absence of the low cost processors and a much larger role for the collectors due to the lower supplies of clean millet. Some of the retail merchants even became collectors going to the village markets where farmers 10 sell small quantities during the year.

10 Farmers use their stored millet and sorghum as a checking account taking small quantities to local markets to exchange in order to finance the household purchases.

101

Figure 7. Present Structure of Millet Marketing in Mali

With the failure of clean millet supplies to expand more rapidly in Burkina Faso it is not surprising that there are no low cost processors emerging. The fundamental difference is the failure to introduce productivity increasing measures (moderate levels of inorganic fertilizers and fertilizer responsive millet cultivars) for millet at the farm level, which would produce a larger market surplus. The failure to increase farm level productivity results in a more elaborate market structure with a greater dependence upon collectors to search for and aggregate small quantities of clean millet (Figure 8).Farmers’ organizations aggregating larger quantities of clean millet can sell directly to the processors. But both farm organizations (some exceptions) and processors were reported to have little investment in storage facilities as compared with Mali.

102 The upper part of Figure 8 of consumers illustrates the choice of buying from retailers the dirty grains and processing them in the neighborhood or purchasing the cleaner more finished millet products from the neighborhood merchant. The neighborhood consumers can opt for the cleaner millet of the neighborhood processors but neighborhood processors also use the same disk mills. As processors obtain more and cheaper millet some will shift to more efficient milling equipment, which reduce or eliminate the metal residues.

Figure 8. Current millet market structure in Burkina Faso with collectors playing a key role in aggregating larger quantities

One large food processor in Burkina has begun specializing in producing maize flour with a hammer mill and he promises to begin producing millet flour on a large scale for bakeries and school feeding programs as part of a World Bank program, WAAP. He could provide the same role of cost reducing processor as is being performed by several processors in Mali. Figure 9 shows the potential emerging market structure with increasing farm level productivity. With technology introduction farmers have more to sell and famers’ associations can produce more clean grain. Farmers Associations also do some of the cleaning plus training in production on keeping the cereal clean. Then the Farmers’ Associations can sell the clean grain directly to the processors for the premium price. The Farmers’ Associations can also sell clean cereal to the wholesalers, who have contracts with the processors for clean cereal. Note that in this market evolution case the role of the collector and the regional market is taken over by the farmers’ associations (Figure 9). As in Mali some wholesalers will invest in 103 new cleaning operations and then sell this clean millet to the processor or retailer. The big change will be when some of the processors become lower cost processors. At this point the neighborhood customers will have access to millet flour from 350 to 400 cfa/kg. This flour will be cleaner and without metal residues as it will be milled with hammer or roller mills. The cost reduction and the quality of the flour and the other products made from it will increase the substitution of the neighborhood processors’ products for the home made products. Many lower and middle class women can shift from previously buying the unprocessed millet and having it dehulled and milled in the neighborhood to the processors’ products. This flour and other products from the low cost processors is cleaner and without metal debris. It can be elaborated into final products of grumeau (monikuru) and degue (tiakry) or sold as flour. The new sector of lower cost processors will need to invest in the better mills and to get their costs down. But we have shown that some processors are already reducing their costs by 30% with access to increased supplies of clean millet and we estimated that with the on-going changes in market structure another 30% price decrease would be possible.

Figure 9. Future improved millet market structure with strong farmers’ associations producing large quantities of clean millet grain.

Improved Millet Flour as Key All millet products, “tô”, “bouillie”, couscous, are made from flour. There is a serious shelf life problem from the weevils that infest the flour.11 Normal shelf life of flour not well

11 It is also necessary to control the moisture content of the flour to avoid molds and oxidative rancidity. Also preventing contamination from bacteria is important. But the chronic and difficult problem is the insect one.

104 dried is 24 to 72 hours in the neighborhood processing of women. In the low cost firms a thorough washing and drying of the dehulled grain before milling is undertaken often with bleach to eliminate contamination in handling the raw material. The washing and drying would not be expected to remove all the eggs. Even a very small number of eggs of the weevils would then find an excellent environment to expand rapidly and with the short life cycle of the weevil would then cause problems. Three months would then be a good estimate of the flour being relatively insect free (A. Ndoye, correspondence, Feb 2015). There is now a machine, the entoleter, that with centrifugal motion can destroy all forms of insect life including eggs. With the entoleter, neither washing nor drying are necessary and the shelf life is prolonged indefinitely, depending on the handling of the flour.12 Freedom from insect infestation during storage can be ensured only if flour is free from insect life when put into bags. The entoleter helps to extend shelf life of flour by killing larvae in it with a fast rotating rotor pin. It is used between production and storage as well as between storage and packaging of flour. The expected shelf life of entoleted normal flour packed in paper bags and stored in cool and dry condition is 2-3 years This entoleter is an important step in the handling of millet flour potentially benefiting substantially not only the millet processors focused on in this paper but also the bakeries. Millet food processors from Senegal have already been buying this machine and processors from Mali have been looking at various potential suppliers. An improvement in the shelf life of the flour is expected to make a substantial difference for the introduction of this flour to bakeries and to other processors of millet products (A. Ndoye, written communication, Feb 18, 2015). With longer lasting flour the processors can supply flour to the bakeries as the basis for cakes, cookies and even bread made from composite millet/wheat flour. In Senegal this type of bread is already common in some bakeries, known as “pain riche.” Demand growth for millet processed products The per capita demand for the millet in these processed products13 is a function of the millet price, and per capita income growth (1).

퐶 푌 =∝ 푃−푛( )∈ (1) 푁 푁 (휕퐶/휕푡)⁄ 퐶 휕푃/휕푡 휕(푌⁄ 푁)/휕푡 휕푁 = −푛 +∈ (2) ( ) 푁⁄ 푃 푌⁄ 푁 휕푡 휕퐶/휕푡 휕푃/휕푡 휕(푌⁄ 푁)/휕푡 휕푁 = −푛 +∈ + ( )⁄ 푁 (3) 퐶 푃 푌⁄ 푁 휕푡

12 The dehulling process has to be contaminant free for this cleaning process to be successful. 13 We did not include here the possible structural shift of a large expansion of the demand by bakeries if the millet flour can attain these longer shelf lives.

105 Taking the log and then the derivatives with respect to time gives (2). On the left hand side (2) is the growth rate of the per capita demand for millet food products. The first term on the right hand side is the contribution to this demand growth from the declining prices. The second term is the effect of per capita income growth on the demand for millet per capita. Then converting this from millet per capital demand growth to total demand growth for millet in these finished products gives Equation (3).The third term in (3) is then the effect of population growth. The price elasticity of demand for millet products is n. ∈ is the income elasticity of demand for millet products. A 30% price reduction has a big effect on the demand growth for millet products even larger than the high population growth rate (Table 1). Transformed millet products have higher price and income elasticities than those for unprocessed millet. Our annual demand growth for millet products results in an estimate of 71% growth over the five years. If there are no price decreases, demand growth is cut almost in half. Using IFPRI estimates (Nelson et al.,2013; also see Balarabe and Chikwendu, 2011 for estimates for northern Nigeria) of the price and income elasticities for unprocessed millet reduces the growth rate to 7.2% and the increase in demand over the period to 54%. Even with a rather extreme assumption of no difference between unprocessed millet and these transformed millet products, this is still a substantial increase in demand. For the ten year estimates and the 60% price decrease the annual demand growth is a respectable 8.6% with a 128% increase over the decade.

Table 1. Estimating Demand Growth for Millet Products over the Next Five and Ten Years in Mali and Burkina Faso. Name Characteristics Price Economic Population Annual Percentage Effect Growth Effect effect on Increase in Effect growth of consumer consumer demand in five demand (ten) years Best Bet Our price and 4.3% 1.8% 3.0% 9.0% 71%% Estimate (5 income elasticities years) (-0.8 and 0.6); 30% price reduction over five years Lower IFPRI price and 3.2 % 1.15% 3.0% 7.2% 54% Consumer Income elasticities preferences (-0.58 and 0.354) for millet(five years) Best Bet (10 More real price 3.8 % 1.8 % 3.0 % 8.6% 128% b years) decrease (60 %) but otherwise same assumptions as Five Year Best Bet

106 Notes: Price Effects. In the Mali millet food processing report (2015) and earlier here we showed the effect of the larger cost reducing firms in reducing prices by 30% for flour and couscous. Over a longer ten year period we think that another 30% reduction would occur as the market structures change. Population growth. Burkina Faso is growing at 2.9% and Mali at 3.1% (World Development Report 2014, p. 296).We used 3% here. Per capital economic growth. Burkina is growing at 6.9% annually and Mali with the war, a coup, and continuing stagnation at –0.44% (World Development Report 2014, p. 296). Adjusting both for expected long term normal growth at 3%.Price elasticity of demand. “Bouillie” and “to” are basic staples but there are a series of other products from millet that people like but that have been high priced from the processors. With lower prices and economic growth there would be substantial expansion of demand for these combined products. Hence, our price elasticity of demand was - 0.8 as compared with the unprocessed millet estimate of IFPRI of -0.58. Note that IFPRI estimates are for millet and our estimates are for the processed, convenient but traditional food products made from millet. Hence we would expect higher elasticities than for millet. We have been fairly conservative in estimating the differences. Income elasticity. We used 0.6 rather than the IFPRI estimate of 0.354 for the same reason as above.

Conclusions: Based upon surveys in Mali and Burkina Faso in the fall of 2014 the millet processing sectors are still very small activities given the extent of millet (and sorghum) production in these countries. Presently, the millet processing sectors produce quality products and many processors are highly mechanized. However, their products are high cost compared with alternative cereals and neighborhood processing by housewives. Therefore sales are concentrated for higher income individuals, including in Mali exporting to the higher income (as compared with Malians) diaspora, and for special occasions for all income classes in Mali. By increasing the efficiency of processing and consolidating the market structure with less middle men between the farmers’ associations and the processors millet product costs can be substantially decreased. The keys are a cheaper production of millet flour as the basic component of all millet products, the further development of farmers’ organizations, and a unified buying system by the processors of the clean millet. These changes could then reduce the marketing margins paid by the processors and increase the share of the marketing margin received by the farmers’ associations. Secondly, as with cowpeas before the introduction of the PICs sacks a principal barrier to increased production of millet flour has been its short shelf life due to the inability to control the weevils. The entoleter gives long term control and changes the whole industry by extending the shelf life of the millet flour, and reducing its cost. Greater availability of clean millet and of longer shelf life flour will enable cost reduction in the market structure and demand expansion for the flour accelerating the growth of the market for processed millet. In the next section, Recommendations, strategies for both countries to accomplish this process of reducing product costs by first 30% and then another 30% are laid out in more detail. Mali is the leader in this process having already substantially increased millet yields in the principal supplying region and now evolving a small sector of high quality processors. Several of the larger processors are now reducing costs for selling to a wider domestic market. The market consolidation, by eliminating many of the collectors and even some of the next stages of regional merchants14 has the potential to reduce the costs another

14 And even some of the wholesaler activity with direct sales of the Farmers’ Associations to the Processors.

107

30%. Using these price reduction estimates combined with population and economic growth we estimate the effects on demand for millet transformed products. The regional effects will be in those regions producing for processors especially the Segou and Mopti regions. There will be income effects for the small farmers in these regions, for the urban processors and for urban consumers especially low income ones. An important additional benefit for low income consumers will be the shift from disk to hammer or roller mills thereby eliminating the metal residues in the flour milled in the neighborhoods (Anon., 2002, p. 24). This combination of potential benefits to low income farmers, processors and consumers makes them especially interesting for further developmental activities to facilitate these changes. These changes will become relevant to Burkina once they increase productivity of the millet system. Higher yields with inorganic fertilizers and fertilizer responsive cultivars will expand the supply of millet with more production surplus to sell. Then further promotion of clean millet, the emergence of low cost processors, and the availability of longer shelf life millet flour will enable the Burkina millet processors to follow the same growth path as in Mali. Recommendations: 1. As the farmers’ associations have increased productivity and produced clean cereals the growth of the processors’ capacity to purchase and integrate this increased supply of clean millet has not accompanied the progress in the farmers’ associations. The processors joining together in an association would enable group buying and thereby facilitate the contacts of many processors with the associations. The farmers’ associations need a price premium for the higher quality millet. The farmers’ associations need to take responsibility for quality control and sell their clean millet in marked bags indicating the origin by association and farmer. Then the processors association could easily take collective action against poor suppliers. There would be a rapid response in quality control by the farmers’ associations and farmers. 2. As the productivity of millet increases the demands on women’s time become too much for the traditional mortar and pestle technique. This technique does result in very clean millet and Mopti is known for this millet quality. In Segou as millet yields were increased farmers switched to running over the millet on the ground with tractors or other vehicles and quickly acquired a reputation for dirty millet and a market price discount for inferior quality. A principal innovation for value added is to switch to methods to produce clean millet and to demand a price premium for the clean millet especially when selling to the millet food processors. One innovation was to put tarps (“bache”) on the ground to reduce the dirt and pebbles picked up with this rolling over technique.15Unfortunately, the tarps soon develop holes and need to be replaced. PAM introduced screens for farmers to

15 Farmers cutting the heads at harvest also need to avoid then putting them on the ground but should put them on the stalks.

108 use especially when putting the cereal into bags in the storage facilities of the farmers’ associations. Ultimately the farmers’ associations will need to have their own threshing machines or to have local farmers own and rent out these machines. Many threshing machines have been tried in Mali. With group ownership they break down very quickly as there is a tendency to let all the members use the machine. Unfortunately, then no one controls the abuse of a member introducing dirty cereal and damaging the machine. So more serious than financing the machine has been the management and maintenance of these machines. 3. A principal constraint on the sales of millet flour has been its short shelf life. Weevils can be a devastating storage pest and are very difficult to control with the conventional washing and drying methods. However, there is a machine available that destroys the weevils and their eggs, the entoleter. This machine offers the potential to open up new markets and products, ie bakeries and composite flour for bread. This type of bread (composite of millet and wheat) is already marketed in some bakeries in Dakar (“pain riche”). Some investments in entoleters could accelerate this process and turn flour into the principal product of these food processors. Low cost entoleters are available from India and China.

Literature Cited Anon., 2002. Promotion du Mil par l’Amelioration des Technologies de Transformacion, Brochure de ROCAFREMI (Millet Network of West and Central Africa), Niamey, Niger, 28 pages Balarabe, A. Ahmed B. and Chikwendu D.O, 2011.“Analyses of Price and Income Elasticities for Cereals Food Crops in an Urban Town of Kaduna, Nigeria,” Agrosearch, Vol. 8 No. 1 & 2: 63. http://dx.doi.org/10.4314/agrosh.v8i1.39439. Ndoye, Ababacar, Botorou Ouendeba, and John H. Sanders, 2015. Demand for Processed Millet in Mali, mimeo produced for the Gates Foundation, Purdue University, West Lafayette, IN Nelson, Gerald C. and Amanda Pa, 2013, African Agriculture and Climate Change: A Comprehensive Analysis, ed. Adulai Jalloh et al. Washington, DC: International Food Policy Research Institute, 2013).

Jazayeri, Ahmad, 2014. Value Chain Analysis (VCA) in Mali : Millet, Sorghum and Rice, draft for USAID-Mali, Bamako, Mali, 36 pages.

Sanders, John H., Jean Harman, Botorou Ouendeba and Soungalo Traore, 2015. Introducing New Millet Production Systems in Mali, mimeo produced for the Gates Foundation, Purdue University, West Lafayette, [In this bulletin pp.47-76] World Bank, 2015. World Development Report of 2014, Washington, D. C.

109 Table A-1. Consumption of Millet Products in the Neighborhoods of Burkina Faso No. Principal No. of times per day No. of Other millet (Refer Products consume products of times/day Products; ence the neighborhood consume frequency of No.) processor millet consumption products bouille a and tô 1. G, D,C,F, Low Income 1 3 Couscous 2/week (3) Gnon Medium Income 1 1 C 2/week; Degue 1/month; Zoum- Koum 1/one to two months High Income 1 1/week 2 (4) G,D,F,C Low Income 1 3 Couscous 1/day Medium Income 1 2 Couscous 1/week High Income 1 1 Couscous 1/week 3 (6) G,D,C Low Income 2 to 3/day Medium Income 1 (to) High Income 1 1/day (to) 2 to 3/week 3 (7) G,D Low Income 3/ day Degue 1/two weeks Medium Income 1/ 1 to 2 to 3/week Degue 1/week 2 days High Income 2 to 3/week Degue 1/day 4 (8) G, D,C Low Income Bouillie Medium Income 1 1/day High Income 2 to 3/week 1/week 5 (9) G, D,C, B Low Income 2to 3/day Zoum-Koum 1/week Medium Income 2 2/day 1 to 2/week High Income 2 1/day 1/month 6 (10) Dehulled Low Income 1 2/ day Degue 1/day millet; G, Medium Income 1 2/day 1/week B High Income 1 1/week 1/ day

7 (11) D,G,C Low Income 1 3/day Gnon, Couscous 1/week Medium Income 1 2/week 1/ month High Income 1 1/month 8 (13) F, Zoom- Low Income 2 1/ day Boire-boire, couscous 1/ month kom Medium Income 2 1/day 1/week High Income 2 1/ day 1/month 9 (15) G, D,B Low Income 2 1/ day Gnon1/day Medium Income 2 3/week 3/week High Income 1/month 10 G,D,C Low Income 1 2/day Couscous, Zoum-Koum 1/month (16) Medium Income 1 2/ day 1/month High Income 1 1/day 1/week

110

11 G,D, Low Income 1b 1/ day Zoum-Koum,others 1/day (17) cookies Medium Income 1 2 to 3/ day 1/day High Income 1 2 to 3/day 2/ week 12 G,D,C,B Low Incomec 3/day Couscous 1/month (18) Medium Income 1 1/day 1/ month 1/month High Income 1 2/month 13 G,D,C,B Low Incomed 1/daye Zoum-Koum (19) Medium Income 1 1/daye e High Income 1 1/day

a. “Tô” often made from maize because the flour is cheaper but millet is expected to regain some of this market if millet flour price can be reduced. b. Low income sector buys more at Ramadan (respondent). c. Low income sector buys less due to high costs of the products (respondent). d. Low income consumers prefer to buy the bouillie sold on the street in this neighborhood (respondent) e. If add in the “tô” from maize, this would double the traditional dishes G: Grumeaux; D: Degue; C:Couscous;B: Bassie; F:Millet flour Source: Field interviews

A-2. Definition of Millet Products in Mali and Burkina Faso Surveys

Tô : Is a thick porridge from cereal (maize, millet or sorghum) flours eaten with sauce of okra, tomato or peanuts; process: boil water; add flour and stir slowly until you get thick paste.

Gnon: prepared with millet flour, powder of cowpea leaves and spices. The mixture is then steam cooked.

Bassi or Sweet Couscous. Process: a mixture of fine flours of millet and peanut is granulated, steam cooked, sugar added and the product is dried then packaged and labeled for marketing.

Zoom-koom : Is a non steamed product. Millet grain milled into very thin flour; add water; stir to get a millet beverage widely used in Burkina Faso.

Grumeaux or Monikuru séché or Bouillie: is a non steam cooked granulated flour; large granules used to make a thick porridge consumed morning and evening. Process: Screen to get thin flour; add water and hand granulate (large granules). Sun dry and do packaging.

Tiakri or Degue: This product is widely used for dinner by some communities in Mali, Burkina and Senegal. Process: wash the dehulled millet grain; add spices and aromas then grind and screen; add water and agglomerate the flour into small granules then steam cook ; screen for the second time then sun dry. Packaging in plastic bags for marketing.

Couscous sec: The millet couscous is eaten with milk or stew.The flour is first screened; put small quantity of water and agglomerate; second screening; steam cook; break the large granules to get the desired size; steam cook for a second time; break the large granules and screen; sun dry and package in plastic bags.

111

Segoudégué: is a spicy granulated flour used to make thin porridge. Consumed any time during the day. Process: wash the decorticated millet; grind and screen the flour; add spices; add water and agglomerate into small granules; dry and pack.

Farine infantile (infant flour): infant food enriched with peanut and soybeans and supplemented with vitamins and minerals; distributed in regions where malnutrition is chronic. This infant food is often sold in pharmacies. Process: Wash the non dehulled millet; dry and roast the grains ;add roasted peanut and soybeans ; mix and mill; screen the product and package in 100, 250 and 500 g bags.

Bénédèguè : enriched flour with sesame ; flour used to make thin porridge consumed by the family (infant and adults). Process: wash the decorticated millet; clean the sesame; roast millet and sesame grains; mill the mixed grains and add spices; for marketing package in plastic bags (500 to 1000 g/bag). Boire-Boire; Probably the same as Mougou Dji: Millet drink made from millet flour, water and milk

112 Feed Grains as a Secondary Market for Sorghum in Mali and Burkina John H. Sanders, Abou Berthe, Botorou Ouendeba and Channa Hira1 Abstract The substantial increases in production and productivity of maize over the last two decades have facilitated the rapid growth of the poultry sector in Mali and Burkina Faso. Maize has consistently been sold below the price of sorghum. Sorghum without tannin has been demonstrated to be an excellent feed but slightly inferior to maize. To substitute sorghum for maize large decreases in the relative price of sorghum to maize are necessary. As the demand expands for feed, cereal production will increasingly need to come from regions north of the prime cotton zones. In these drier conditions sorghum’s greater tolerance to low and variable rainfall and to lower soil fertility will give sorghum a productivity advantage over maize. Given the larger area over which sorghum is produced sorghum costs can become competitive with maize without reaching the same high yield levels as attained in maize in the south. When incomes increase in the development process consumers shift from a predominant consumption of cereals and tubers to fruits, vegetables, meat, milk and cheese. The most rapid growth in this process is for broilers as chickens pass from an expensive meat for wealthy people and special occasions to a staple for lower and middle income consumers. The broiler sector in Mali is increasing rapidly but is incurring high costs now with poor quality protein sources, irregular quality domestic concentrate, and expensive imported concentrates. Moreover, there is concern with improving the quality of the maize (50 to 60% of the ration) by avoiding maize 2 with impurities and molds. Introduction Sorghum and millet are the predominant cereals for human consumption in Mali and Burkina Faso. As productivity is increased it is possible to avoid or at least moderate the price collapses that occur both in good rainfall years and as technology is successfully introduced. The development of secondary markets can play this function. In millet there is a small but increasing sector of food processors of traditional millet products in urban centers. For sorghum there is the potential demand for feed grain. For most of the 21st Century in both countries the cost advantage in feed has been with maize rather than sorghum. In Mali maize has benefitted from substantial gains in yields over the period with new cultivars, higher fertilization and improved agronomy. Yields increased from slightly over one ton to almost 3 tons in 2014. In contrast, sorghum yields have been stagnant

1 Professor of Agricultural Economics, Purdue University; Director of Sasakawa, Mali, PhD in Animal Science from the University of Florida; Director of the 3N (Nutrition) Program in Niger; graduate student, Purdue University. 2 This can be an especially serious problem for those chicken producers importing maize from the coastal countries with their wetter, hotter conditions.

113 (Figure 1). In Burkina both maize and sorghum began with higher yields but neither appreciably increased in the decade (figure 2). In semi-arid regions all over the world sorghum is produced as a substitute for maize in the feed. The first requirement for sorghum to be price competitive with maize is then to increase sorghum yields.

Figure 1. Maize and Sorghum Yields(tons/ha) in

3.5 Mali, 2006-2014

Sorghum

Source: ADA (Directorate of National Agriculture), unpublished data, 2015, Bamako, Mali.

Figure 2. Maize and Sorghum Yields (tons/ha) in Burkina Faso, 2004-2013

2.50

2.00

ha)

\ 1.50

1.00

Yield(Tons 0.50

0.00 2013 Source: Departament d’Agriculture, Annuaire des statistiques agricoles, Ouagadougou, Burkina Faso

Besides the substitution potential of sorghum for maize in the ration this paper is also concerned with the growth of the broiler sector and the modernization process of the intensive poultry sector in Mali and Burkina Faso.

114 The Success of Maize and the Expansion of the Poultry Sector From the early ‘90s Malian maize production accelerated from less than 200 tons to almost 2 million tons in 2014, a tenfold increase (Mas Aparisi et al, 2013, pp. 7).In West Africa the most rapid increase in cereal production since 2000 has been in Mali (Mali Best Report, 2015, p.11). From 2000 to 2013 maize production has increased at a 17% annual growth rate, faster than any other cereal and is now consumed at 35 kg/person year, the forth cereal in importance in Mali (Mali Best Report, 2015, p. 28; Figure 7, this paper)). In the 21st Century maize has been the “go to” crop in the cotton sector. When farmers were unhappy with the low prices or late payments for cotton, they planted less cotton and/or diverted more of the cotton fertilizer to maize and later sorghum. CMDT (the cotton monopsony) made this diversion of their cotton credits to the cereals into a virtue by advocating diversification. Maize production is concentrated in the higher rainfall south with an estimated 80% in the Sikasso region (Mali Best Report, 2015, p. 28). In both Mali and Burkina Faso there have been substantial gains in production of both sorghum and maize from 2008-2014 (Figures 3 and 4).

Figure 3. Maize and Sorghum Production (tons) in Mali, 2006-2014 2500000

2000000

1500000

1000000

500000

0 2014

Sorghum

Source: ADA, Directorate of National Agriculture, Bamako, Mali, 2015.

115

Figure 4. Maize and Sorghum Production (tons) in

Burkina Faso, 2004-2013

2500000

2000000

tons) 1500000

1000000

Production( 500000

0 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

sorghum

Source: Departament d’Agriculture, Annuaire des statistiques agricoles, Ouagadougou, Burkina Faso

Maize production and productivity gains in the last fifteen years have made available low cost maize and facilitated the rapid growth of the Malian and Burkinabe poultry sectors.3 The association of poultry producers (FIFAM) states that there were 2,670 intensive chicken producers in 2014 as compared with 216 in 1999 (1999 estimate of Chemonics; cited from Berthe, 2015 , p.2). Maize is 50 to 60% of the feed volume and feed costs; thirty percent of maize production goes for feed principally for poultry (Mali Best Report, 2015, p. 29). In contrast Malian sorghum yields stagnated at one ton/ha (Figure1). But after 2008 sorghum production was substantially increased reaching 1.5 million tons in both 2009 and 2014 though falling below that in the intervening years (Figure 3).Attaining these high production levels without a complete price collapse indicates the continuing importance of the demand for 4 sorghum as a human food and exports to other countries especially Niger and Mauritania. The Potential Demand Growth of Sorghum as a Feed Maize substantial yield gains have been concentrated in the higher rainfall cotton south of Mali and this includes the introduction of hybrids. Sorghum is just beginning this process of productivity increase in Mali and Burkina Faso but has made rapid yield gains in other countries 5 based upon these same inputs as with maize productivity increases.

3There has been little trade in maize in comparison with the total production. In peak years for trade maize exports were 3% of production in 2004 and imports were 2% of production in 2005. After 2006 export bans led to international exchanges of maize of less than 0.5% of maize production according to official data though some smuggling continued (Mas Aparisi et al, 2013, p.11). So both rapidly expanding productivity and trade restrictions reduced the domestic price of maize and facilitated the growth of intensive poultry production. 4 In earlier trips we learned of Nigerien and Mauritanian traders buying and exporting sorghum from the main export site of sorghum, Koutiala. 5 From the late ‘50s in 14 years national sorghum yields were tripled in the US demonstrating a substantial response to new cultivars, increased fertilization, and better agronomy (Miller, F.R. and Y. Kebede, 1987).

116 Fortunately for the prospects of increasing sorghum productivity there is a substantial yield gap that still exists for sorghum between farm level and experiment station yields. Yield increases of 50% for sorghum and millet technology with open pollinated cultivars and inorganic fertilizer have already been demonstrated in pilot projects and these yield gains have been shown to be profitable both in the cotton zone and in drier regions over several years (F. Baquedano et. al., 2010; J. Coulibaly et.al., 2015; J. Coulibaly et. al., 2013; Vitale and Sanders, 2005). So if there are good prospects for increasing productivity and decreasing the price of sorghum relative to maize, what is the substitution relationship of sorghum for maize as a feed? Sorghums without tanins have a feed efficiency of 95 to 97% of that of maize (J. Hancock, conversation, 2014; also see Tandiang et al., 2014; Clement et. al., 2010; Carmencita and Nelia, 2006; Parthasarathy et al., 2005; and Dowling et. al., 2002; and Hancock, 2000). Therefore at a price of 95% of that of maize equal expenditures for the two cereals would result in equivalent productivity gains (weight gain in broilers and rate of egg production in layers). At a lower price for sorghum than 95% of the maize price the advantage in the feed would shift to sorghum. With rapid supply increase in broilers and falling prices of chicken relative to other meats, poultry producers are expected to be very sensitive to small price changes if they know that productivity can be maintained (Appendix A). When maize prices reached 118,000 to 131,000 fcfa/sack (100 kg) in 2008 and stayed high in the winter of 2009, there was a cost advantage to using non-tannin sorghum from June until September of 2008 (Figure 3). In July and August the sorghum price was only 90% of the cost of maize. Unfortunately, almost all the chicken producers interviewed in Mali did not even know about this potential of substituting for maize with sorghum. This also included the veterinarians and other feed specialists making recommendations to the poultry producers on feed composition. So the problem is not only with the sorghum relative cost but also with the knowledge even among specialists about the feed value and substitution potential of sorghum.

Figure 5. Relative Pricesa (Sorghum/Maize) in Mali

1.85

1.65

1.45

1.25

1.05

0.85 2015 Price Ratio 0.95 Ratio Line

Source: OMA (Observatoire du Marché Agricole ), monthly data. a. Note that it would have been more appropriate to use wholesale prices as most of the chicken producers bought from wholesalers though some did buy from farmers. So we had to assume that marketing margins did not appreciably change over this period.

117

Figure 6. Relative Pricesa(Sorghum/Maize) in Burkina Faso

1.15

1.05

0.95

a. Prices in the two regions of Burkina Faso where Ouagadougou, the principal poultry producer for broilers is found. There is substantial egg production in Bobo. However, including Bobo did not appreciably change the relationship above. See also footnote a in Figure 5 above. Source: Departament d’Agriculture, Annuaire des statistiques agricoles, Ouagadougou, Burkina Faso

After this brief period in Mali with lower costs in 2008 sorghum has again been losing the competition with maize as the lowest cost cereal for feed. Moreover, with another production and productivity jump for maize from 2010 to 2014 this price ratio reached and stayed above 1.2 (Figure 5).The sorghum price was 20% more than that of maize in most of 2014 rather than the goal of 5% percent below. To expand the demand for sorghum this cost advantage of maize needs to be reversed along with making information on sorghum substitution potential available to chicken producers and their technical advisers on feed composition. Implementing productivity increases of sorghum with improved seed production and credit for inputs as has been done with maize and rice is feasible. With the further rapid expansion of the poultry industry there will be a strong continuing demand for cereals as feed. Meanwhile maize breeders search for early and very early (drought escape) cultivars and do more basic research on drought resistance on maize while sorghum and millet already have drought resistant6 characteristics. The Economics of Sorghum Competing as a Feed Returning to the original problem of what happens if the productivity of sorghum is substantially increased. Fortunately, there are two favorable outcomes from increasing the productivity of sorghum. First, as sorghum yields are increased but the price ratio continues to favor maize: This could result if there were continuing productivity gains in maize. In this case consumers will benefit from the lower sorghum prices. Mali still has substantial malnutrition

6 Drought tolerance is a characteristics of early cultivars to escape late season drought. Early drought can be escaped by just replanting or planting later. Drought resistance refers to plant characteristics giving a plant the ability to tolerate drought during the season and at the end of the season.

118 problems and population growth is very high at 3.6% (Malian estimate of population growth quoted in Mas Aparisi A., Diallo F., Balié J., 2013, p. 9). The World Food Program (PAM) and the government acquisition and distribution program (OPAM) concentrate on purchasing sorghum and millet to stimulate production of low income farmers while also distributing to the nutritionally needy (PAM) or responding to drought or other staple food crisis (OPAM). In 2013 PAM purchased 20,262 tons of sorghum and millet and in 2014 OPAM was planning to acquire 35,000 tons of these two cereals (Mali Best Report, 2015, pp. 26, 27).7 The role of food security is still of primary importance to Mali. In the second case sorghum productivity is increased and the sorghum price falls to 0.95 or less of the maize price: Below a price ratio of .95 (sorghum price/maize price), there will be an economic incentive to substitute sorghum for maize in the feed. An increased demand for sorghum of 9.7% annually is estimated stemming primarily from population and economic growth (Appendix A). As a result the price decline from rapid technology introduction of sorghum or from good weather will be reduced. By raising the expected price for sorghum (decreasing the price collapse) farmers gain higher incomes from the increased demand for sorghum in the feed ration. Also note that with increased productivity costs decline so that at least initially before many farmers adopt the higher productivity innovations, farmers can still profit even with lower prices. In both of the above outcomes the increased consumption of 8 sorghum is beneficial to Malian consumers and increases farmers’ incomes. The second case is more likely and more consistent with the factor endowment of Mali as it is difficult to expand maize productivity outside of the high rainfall zone of the Sahel. Another factor making maize less expensive than sorghum is government policy especially the export prohibitions intensively used after 2006 (Mas Aparisi et al, 2013, p. 11). Moreover, there has

Ps/Pm

1.95 a

b Quantity of sorghum

Figure 7. Conceptual Demand for Sorghum in the Feed

7 Note how small these levels are compared with the 1.5 million tons of sorghum in 2009 and 2014 (Figure 2). 8 Another adjustment to falling prices from higher productivity is that farmers can reduce their area in sorghum and plant other crops. 119 been more extensive use of subsidies for maize than for sorghum (J. Coulibaly et al., 2015). Public policy eliminating export bans on maize and equalizing input subsidies would make relative price changes more favorable for the competitive position of sorghum in the feed. Once the sorghum price falls to 95% of the maize price equal expenditures on the two cereals result in equal feed efficiency, the same amounts of weight gains in broilers or egg productivity of layers. Before declining to the point “a” the demand curve for sorghum is the vertical axis or zero sorghum as there would be no productivity gain from switching just higher costs (Figure4). Once the price falls below “a” the profit maximizing chicken producer should switch to sorghum. This assumes that the intensive chicken producer is informed about the productivity of sorghum as a feed and that adjustment costs of switching feeds are not large. Information is critical as not even his technical adviser has been alerting him about the potential of sorghum. But the adjustment costs are expected to be minimal with some changes in suppliers and quantity adjustments in the feed (see Appendix A for further elaboration of demand shifts for maize as a feed above .95 and for sorghum below this price ratio). Meanwhile we need to return to several other factors affecting feed choice and the evolution of the poultry industry in Mali and Burkina Faso. Cereal and Meat Consumption in Mali We have noted above the continuing importance of sorghum and millet as human food. So in 2013 sorghum was still more important than maize as a human food. But as with rice and millet this is changing as maize productivity gains lead to lower prices and increased consumption as food and feed (Figure 8).

Figure 8. Cereal Consumption (food) in kg/capita-year in Mali in 2013

70

60

50

40

30

20

10

0 Rice Millet Sorghum Maize

Source: Mali Best Project, 2015, pp. 15,20,25,28

Note: Rice provides 20% of food calories, millet 16%, sorghum 14% and maize 11%

With the low incomes in the country meat consumption is still very low compared with cereals. Nevertheless, poultry consumption almost tripled from 2008 to 2012 and has ascended to

120 second to beef in 2012 (Table 1). Still beef consumption dominates being five times that of poultry in Mali with its large crop and pastoral regions.

Table 1.Meat Consumption (kg/person/year) in Mali, 2012

Products 2008 2009 2010 2011 2012 Beef kg/person/year 2.60 2.48 2.84 2.78 2.84 (% of meat (65%) (70%) (69%) (63%) (66%) consumption) 0.78 0.28 0.37 0.43 0.35 Mutton (19%) (8%) (9%) (10%) (8%) 0.37 0.47 0.55 0.59 0.48 Goat (9%) (13%) (13%) (14%) (11%) 0.20 0.26 0.32 0.50 0.59 Poultry (5%) (7%) (8%) (11%) (14%) Total meat consumption 3.98 3.51 4.11 4.35 4.29) (kg/person/year) Source: Ministère de l’Elevage et de la Pêche, 2014, Bamako, Mali..

In contrast, pork and chicken are much more important than beef production in Burkina Faso. Why ? Burkina Faso has a higher population concentration hence less land available for pasture and less crop residues. Moreover, 38% of the population is non-Muslim as compared with the 95% Muslim population in Mali (CIA, 2016). This greater importance of poultry in Burkina means that we would expect it to be more advanced and need to look in more detail at the functioning of the broiler production sector there.

Figure 9. Basic Meat Production in Burkina Faso, 2003-2014

350000

300000

250000

tons) 200000

150000

100000

Production(

50000

0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Beef Pork Chicken

Source: Department d’Agriculture, 2015. Annuaire des statistiques agricoles, Ouagadouou, Burkina Faso

121 Shifting the Demand for Sorghum as Feed There are other important factors affecting the shift in demand for cereals in the feed. Maize can get fungi causing myco-toxins in the field. Both sorghum and maize are vulnerable in 9 storage and transport especially if they are not harvested and stored sufficiently dry. Afla-toxin has been an important cost to maize. Producers reported chicken losses to “moisissures” (molds) and said they rejected maize with clear mold evidence. There are effective controls now. T5X costs about 1,200 cfa/kg (sold in 20 kg sacks) and is applied at 2 kg/ton either as a prophylactic or a treatment. This is an important innovation. Unfortunately many chicken producers even though familiar with the molds are not familiar with these latest techniques for control and some are not selective enough in their choice of maize. Tannin is a nutritional inhibitor found in sorghum, grapes, cranberries and dark chocolate. It is an anti-oxidant and reduces primarily protein absorption (L. Rooney, 2005). These tannins in sorghum are problems in the growth of chicks and adolescent chickens (‘poulettes”) but they do not affect adult performance (Sedima10 executives, conversation). Most chicken producers, vets and feed dealers either know about tannins or at least that there are some digestive problems with sorghums. What is not generally known is that with farmer and breeder selection many sorghums11 in Mali and many other countries do not have tannin (Tahirou et al., 2006, pp. 8, 9; Rooney, 2006). Still before recommending the substitution of sorghum for maize a supply of a uniform sorghum cultivar without tannin would need to be available for chicken producers so some labeling and extension information will be important. Qualitative Improvements in chicken production Intensive poultry production is a difficult skill to master so countries are observed to go through a long process of the prices of chicken relative to other meats falling as this learning by doing occurs. How is this introduction process proceeding in Mali? There are serious nutritional problems with the two main sources of protein12 commonly used in the feed. First there is fish meal from Senegal and dried fish from Mali. Fish meal can bring mycotoxins and bacterial infections (salmonella). Cotton seed meal often has too much fiber plus gossypol. Now many broiler producers are switching to the imported concentrate (without the protein sources) or the imported complete feed (without the cereal). In Burkina Faso we did a more complete survey of feed costs as compared with the weight gains and revenue increases from broiler production (Table B-1). On average the decision to import feed reduces costs by 235 cfa per kg of carcass weight of chicken sold (Table B-2 in Appendix B).This significant (10% probability) 29% cost reduction is a substantial savings.

9 The bottom of storage or in transport is generally hotter and wetter and thus more subject to mold infection. 10 Principal feed manufacturer in Dakar, Senegal 11 Of the traditional cultivars found in Malian local markets in 2005 seven of the 13 did not have tannins. Nine of the ten improved cultivars did not contain tannin. There were two laboratory tests in Texas A&M and in the Food Technology Lab of the Agricultural Research Institute (IER) of Mali (Tahirou and Sanders, 2006). 12 Some have been contracting farmers to produce soybeans as a better protein source. But there have been problems with the contracting process and with the necessary treatment of roasting (“torrefaction”).

122 Hence, the domestic producers of concentrate can see the need to improve their quality or reduce their prices. Another complementary response is to improve the quality of the local protein sources. The frequent complaints of Burkina chicken producers (egg producers in Bobo and broilers in Ouaga) were about the irregularity of the quality of some domestic concentrates and 13 about the local protein sources. In analyzing the value of chicken per unit of feed cost the measure now also reflects the marketing ability of the chicken producer. We would expect larger producers to be able to market better and this is reflected significantly here (Table B-3). Selling all your chickens at one time was expected to be more efficient and the sign was correct but no significant effect (Table B-3). However, there was no advantage and a significant loss of revenue to the strategy of many larger producers of selling a smaller carcass size to compete with the backyard chicken of the grilled chicken producers found all over small and large urban areas. Those selling heavier weights (above 1.8 kg or in another regression above 45 days) as to restaurants, hotels or individuals still did better than those focusing on smaller and cheaper birds for the grills (Table B-4). The market for larger carcass size is probably still limited by the higher costs for suppliers to regularly furnish the carcasses to specific hotel or restaurant clients as well as for the slower growth of the larger broiler size. Many producers of broilers introduce five or six rotations. Each batch of broilers takes 35 to 45 days depending upon the weight goal. Often this time length is extended for larger weight gains or because there is difficulty in marketing all the chickens at one time. This reduces the time available for sanitary treatments between rotations. Vets have told us that this can then result in a build up over time of bacterial and other infections which can ultimately decimate production. But we did not have this problem reported in either Malian or Burkinabe interviews. The dirt and small pebbles often found in the maize shorten the life of the feed grinders. Some offering this grinding service to other producers comment that they are reducing this service and just focusing on their own feeds because of the dirty maize they are receiving. Up to 15% impurities have been observed in the maize (field interviews in Bamako). There is a general preference for the yellow maize as the white maize grinds into too fine a powder for a good feed. Improving Efficiency in the Broiler Industry in Burkina Faso The use of sorghum in the feed is a three step process: 1) broiler production needs to increase rapidly as it is now doing in Mali and Burkina; 2) sorghum productivity needs to expand which is not happening; 3) with the productivity increase of sorghum and the increasing demand for maize for export the price of sorghum needs to fall below 95% of that of maize.

13 Two econometric notes. Our size or scale factor, which was the number of birds in the batch was not significant in this same equation hence the cost savings can be attributed more to feed quality than to better management of larger producers. Secondly, running the equation only with the complete feed including protein gave a non-significant response but almost the same parameter value as with the two imported feeds, with and without the protein sources.

123 The most important change is to reduce costs and chicken prices relative to other meats and fish and thereby further expand the demand of the middle and lower income sectors of the population for broilers. This is already occurring in Burkina Faso (see Pictures 1 and 2).Chicken as a low cost fast food depends upon this decline in relative prices in prices and the consumption 14 preference shift away from bicycle chicken. Is there evidence for the start of this process? Some broiler producers have been focusing sales to the grillers of chicken. These grillers have been traditionally supplied with small backyard chicken to keep prices down for the middle and lower classes. Now several intensive producers are producing chickens of 30 to 35 days to sell there. Secondly fast food, low cost chicken production of broilers has now arrived in Ouaga and is expected to expand rapidly. Bicycle chicken from backyard production is very popular in many restaurants and corner grills in Ouaga presently. It is prepared with different sauces and often cooked a long time to reduce the toughness. Broilers, intensively produced, are coming in quickly in spite of the frequent assertion that consumers appreciate the taste of backyard chicken more than the intensively produced broilers. The backyard chicken will ultimately be displaced but in the next five years intensive broiler production is expected to satisfy the rapidly increasing demand for chicken and substitute for the more irregular supply of backyard chicken as it is more affected by disease.

Photo 1.Broiler (“poulet de chair”) for the Mass Market, Ouagadougou, Burkina Faso, December 2015

14 Backyard chicken are taken to market often hanging over a bike or motorcycle.

124

Photo 2.Opening night Picture taken in Ouaga, December, 2015

Vertical Integration and Efficiency in Feed Choice In the modernization of the poultry sector vertical integration is often seen where marketing functions are expanded into the supply chain by the chicken producers or feed dealers. In Central America the large feed dealers even set up their own chicken restaurants in the capitals. Recently in Mali many chicken producers have been selling their chicken carcasses at 1400 to 1600 cfa principally to “revendeurs,” who will buy in quantity and distribute the chickens in smaller quantities to hotels, restaurants, and supermarkets. Some chicken producers develop their own clients as at selling points with refrigeration or sell directly to hotels or restaurants (Figure 10). Then they often need transportation and to sell to order, which can mean having more than one set of broilers being raised at overlapping times. So there are costs associated with skipping stages in the marketing process. But as the supply of producers is expanding rapidly, real prices are coming down so there are increasing pressures to find ways to increase revenues and to reduce costs.

125

Figure 10. 2015 Prices at the Different Marketing Stages for Dressed Chickensa

Hotels/restaurants 3000

Selling point 2100

Revendeurs 1800

Farmer 1600

0 3500 Price at different Markekting levels

Source: Survey data reported in Berthe, 2015. Note: The standard chicken carcass after 35 to 40 days is then 1 to 1.2 kgs

In the modernization process the improved efficiency of the feed with higher dependability is the most important constraint identified here. Increasing numbers of producers are taking the simple and cheaper route of buying imported concentrate. So this reduces costs and avoids the problems with the local protein sources and the irregular quality of some domestic concentrates. Quality control of inputs is an important role of the state. Conclusions: The feed option for sorghum is still out there as a possible floor but it implies a large decrease in the relative price of sorghum to maize. With the substantial cost differences and a twenty five year head start of maize, the sorghum yield gains will need to be widely diffused among farmers for sorghum to become competitive. On the positive side for sorghum are the much larger areas and numbers of farmers in the country who can increase the productivity of their sorghum as compared with the regional concentration of the maize producers in the prime cotton zone. Hence, the yield gains will not need to be as large for sorghum to successfully compete cost wise with maize. The expansion of sorghum and millet production since 2008 even though based upon area increase indicates that both still have important roles to play as human foods. Sorghum production of 1.5 million tons in 2009 and 2014 was impressive. The expansion of the World Food Program and of OPAM have been oriented to sorghum, millet and rice to benefit the low income producers in the acquisition process and those with under and mal-nutrition in the

126 delivery process.15 So in the near term, the next five years, we expect both sorghum and millet to be primarily used as human food and that prices will not fall substantially. If productivity can be increased rapidly, sorghum is expected to be able to compete with maize as a feed especially if export prohibitions are taken off maize. The demand projections for feed use of sorghum indicate 9.7% annual demand growth for sorghum once it reaches the 0.95 threshold (Appendix A). As with maize in the past twenty years this projected demand growth for sorghum is principally driven by the rapid income and population growth in Mali. To improve the productivity of the poultry sector increased concern needs to be addressed to the quality of the components especially the local protein sources and to improved regulation of the quality of domestic concentrates. Development of domestic soybean production for meal as a protein source would be a substantial qualitative improvement in the protein feed options. Also on the marketing side facilitating the skipping of market stages can serve the same purpose as reducing costs thereby providing more incentives to increase efficiency and size. Skipping marketing stages can involve increased costs so it is not always profitable. Recommendations

1. To increase the use of sorghum a price ratio to maize of 0.95 or lower is necessary to equal the feed efficiency of maize. Since maize productivity has increased substantially in the last two decades while sorghum productivity has stagnated, substantial advances in sorghum productivity are necessary. However, sorghum is grown all over the country whereas maize has difficulty moving north where rainfall deficit/variability and soil fertility problems are more severe. The conventional wisdom that sorghum does not respond nor is profitable with higher input levels is false and needs to change especially among public officials as there is empirical evidence to the contrary and successful farm level introduction. 2. With the present nutritional problems and PAM purchases sorghum’s use as a human food has kept demand and prices high and is expected to do so for the next five years. So sorghum research and extension is justified from the perspective of human welfare and with productivity increases enabling lower costs and prices the poultry sector can absorb large increases in sorghum supply while insulating sorghum against a price collapse. 3. In preparation for the availability of lower cost sorghum competitive with maize workshops need to be held for the vets and other nutrition advisers of intensive chicken producers on the feed efficiency of non-tannin sorghum as compared with maize. 4. The lack of a high quality protein source is an important factor reducing feed efficiency presently. The association of chicken producers could organize a number of chicken producers to purchase soybeans jointly from several farmers’

15 OPAM also has an important objective of setting aside a cereal stock to respond to adverse climatic events usually a drought. But note that the purchases by both PAM and OPAM are very small as compared with the total availability of both cereals.

127 associations. One or more of these associations could invest in the facilities to do the roasting of the soybeans. 5. As chicken producers expand in size some are producing specific chicken weights to order and they sell directly to hotels, restaurants and supermarkets bypassing the middlemen (“revendeurs”) and capturing their marketing margin. However, most will need to reduce costs and prices to substantially expand sales to the low and middle income consumers. References: Abdoulaye, Tahirou, J.H. Sanders, and B. Ouendeba, 2009. Evaluacion de l’Introduction de Technologies de Production et de Strategies de Commercialisation du Mil et du Sorgho, Campagne Agricole 2006-2007, INTSORMIL Bulletin No 8, Lincoln, NE 24 pages Abdoulaye, Tahirou, J.H. Sanders and B. Ouendeba, 2008. Evaluation of Sorghum and Millet Technology and Market Strategy Introduction: 2006-2007 Crop year, INTSORMIL Bulletin No 8 Lincoln Ne 22 pages Abdoulaye, Tahirou, and J.H.Sanders, 2006. Sorghum or Maize in West African Poultry Rations, Marketing- Processing Project-INTSORMIL,USAID-West Africa, 17 pages. In French Abdoulaye T., J. Sanders, and B. Ouendeba, 2006. Quelle Cereale pour les Aliments de Volaille en Afrique de l’Ouest: Sorgho ou a Mais, Bulletin No. 4 Projet Marketing-Processing, INTSORMIL, Lincoln, Nebraska

Baquedano, F., J.H. Sanders, and J. Vitale, 2010. Increasing Incomes of Malian Farmers: Is Elimination of US Subsidies the Only Solution? Agricultural Systems 103: 418-432

Berthe, A. 2015. Sorghum Secondary Market Development through the Commercial Poultry Industry in Mali, Consultants Report, Bamako Mali, Clement I.M, I.D. Kwari, J. Igwebuike, I. Nkama, I.D. Mohammed and B. Hamaker, 2010: Performance and economics of production of broiler chickens fed sorghum or millet as replacement for maize in the semi-arid zone of Nigeria. Agric. Biol. J. N. Am., 1(3): 321-325.

Carmencita, D. M and Nelia F. C., 2006: Feeding and Economic Evaluation of Corn, Wheat, and Sorghum Based-Diets in Broilers. Philippine Journal of Science 135 (1): 49-58, June 2006.

CIA (Central Intelligence Agency), 2016. The Fact Book, Internet Coulibaly, J., J.H. Sanders, P. Preckel, and T. Baker, 2015. Will Cotton Make a Comeback in Mali? Agricultural Economics, 46:53-67.

Coulibaly, J., G. Kumaraswamy, and J. H. Sanders, 2013.Economic Impact of Sorghum and Millet Technologies in Mali Agricultural Campaign, 2010-11, Bulletin IER-INTSORMIL n° 11, Purdue University, Department of Agricultural Economics, February 2013, 31 pages.

Dowling, L.F., C. Arndt and B.R. Hamaker, 2002: Economic viability of high digestibility sorghum as feed for market broilers. Agron. J., 94: 1050-1058.

128 Hancock, J.D.,2000. Value of Sorghum and Sorghum Co-products in Diets for Livestock In W. Smith and R.A. Frederickson (edited), Sorghum Origin, History, Technology and Production, Wiley Series in Crop Science, pp. 731-751.

J. Harman, 2015 Introducing New Millet Production Systems in Mali, mimeo produced for the Gates Foundation, Purdue University, West Lafayette, IN. Mali Best Project, 2015. Bellmon Estimation Studies for Title II, USAID-BEST Analysis, Mali, Fintrac Inc, Washington D.C. Mas Aparisi, A., Diallo, F., Balie, J., 2013. Analyse des Incitations et Penalisations pour le Mais au Mali, Serie notes techniques, SPAAA, FAO, Rome, Italy

Mas Aparisi A., Diallo F., Balié J., 2013. Analyse des Incitations et Pénalisations pour le Mil et le Sorgho au Mali. Série notes techniques, SPAAA, FAO, Rome.

Miller, F.R. and Y. Kebede, 1987. Genetic Contribution to Yield Gains in Sorghum, 1950-1980, in W.R.Fehr (editor), Genetic Contributions to Yield Gains of Five Major Plants, Crop Science Society of America Special Publication No. 7, American Society of Agronomy, Madison, Wisconsin,1-14.

Ministère de l’Elevage et de la Pêche, 2014. Développement de l’élevage et réduction de la pauvreté au Mali, Diagnostic de la situation de l’élevage, Equipe Technique Nationale African Iivestock development Initiative (Alive), Bamako, Mali.

Ndoye, Ababacar, Botorou Ouendeba, and John H. Sanders, 2016. Demand for Millet Processed Products in Mali and Burkina Faso, mimeo submitted to the Gates Foundation. Purdue University, West Lafayette, IN

Parthasarathy, P.R., K.R. Gurava, V.S. Reddy and C.L.Hancock, Gowda, 2005. Linking procedures and processors of sorghum for poultry feed: A case study from India. International Crops Research Institute for the Semi- Arid Tropics (ICRISA). New Delhi, India, pp: 12. Rooney, L., 2005.Ten Myths about Tannin in Sorghum, jourlib Journal, (www.jourlib.org/paper/2869243#VSrBGstOz)

Sanders, J.H., Jean Harman, Botorou Ouendeba, and Soungalo Traore, 2015.New Millet Production Systems in Mali, mimeo submitted to the Gates Foundation. Purdue University, West Lafayette, IN

Staatz, J.Valerie Kelly, Duncan Boughton, Niama Nango Dembélé, Miriam Sohlberg, Abdrahmane Berthé, Mark Skidmore, Cheick Oumar Diarrah, Abdoul Murekezi, Robert Richardson, Brent Simpson, Sonja Perakis, Amadou Sékou Diallo, Ramziath Adjao, Mariam Sako, Nathalie Me-Nsopé, and Jacob Coulibaly. 2011.Mali Agricultural Sector Assessment, Food Security Team Department of Agricultural, Food and Resource Economics, Michigan State University, East Lansing, MI

129 Tandiang, D.M.,M.T. Diop, A. Dieng, G.M.L. Yoda, N.Cisse, and M. Nassim, 2014.Effect of Corn Substitution by Sorghum Grain with Low Tannin Content on Broilers Production: Animal Performance, Nutrient Digestibility and Carcass Characteristics, International Journal of Poultry Science, 13 (10): 568-574

Vitale, J. and J. H. Sanders 2005. New markets and technological change for the traditional cereals in semiarid Sub-Saharan Africa: the Malian case, Agricultural Economics, 32, 111- 129.

130 Appendix A. Estimating the demand increases for cereals in the feed Above “a” (Figure 7) there is no substitution of sorghum for maize as the cost of sorghum is greater and there is no advantage in the feed efficiency value from substituting sorghum for maize. At “a” the feed efficiency value for equivalent expenditures on either sorghum or maize are equal. Below “a” moving along the demand curve for sorghum “ab” the quantity demanded of sorghum in the feed at the farm level will be determined by the relative prices of sorghum and maize. 16 Now if we add in per capita income growth as a shifter as we know that the quantity demanded of chicken per capita will increase substantially with income growth, the equation: 퐶 Ps 푌 =∝ ( )−훾( )훽 (1) 푁 푃m 푁 Where consumption per capita of cereal in the feed ration is a function of the prices of sorghum and maize. Once at point “a” a lower price of sorghum relative to maize results in an increased demand for sorghum. This also assumes that chicken producers are informed about the potential for substitution between the two cereals. Taking the logs and differentiating with respect to time turns the equation into rates of growth over time (2). 휕푃 휕푃푠 푀 ( ) (휕퐶/휕푡)⁄퐶 휕푡 휕푡 휕(푌⁄푁)/휕푡 휕푁 = 훾 − 훾 + 훽 (2) ( )⁄푁 푃푀 푃푠 푌⁄푁 휕푡

Finally converting this to total demand growth rate for sorghum as the feed gives us (3):

휕푃푀 휕푃푠 휕퐶/휕푡 ( ) 휕(푌⁄푁)/휕푡 휕푁 = +훾 휕푡 − 훾 휕푡 + 훽 + ( )⁄푁 (3) 퐶 푃푀 푃푠 푌⁄푁 휕푡

The above equation becomes meaningful for sorghum once the price ratio reaches 0.95 (“a”). There might be some substitution earlier but it would not be rational. Once the price ratio is below .95 the rational chicken producer with perfect knowledge of the relative feed values would shift completely sorghum for maize. In reality the acquisition of this knowledge and any adjustment costs would result in a gradual substitution so the estimates here are for a maximum demand growth for sorghum as the information on the feeding efficiency sorghum is made available and is credible to the veterinarians advising the chicken producers on their feed composition. Below “a” the first two terms will be the changes of the growth rate of the demand for sorghum as the prices of sorghum and maize change. The third term on the right hand side picks up the derived demand for sorghum as there is increased demand for chicken in the diet associated with per capita income growth. So the high income elasticities of demand for chicken are reflected here. Finally, the fifth term moves the equation to total rather than per capita growth rates by adding in the effect on demand growth for chicken from population growth.

16 We have already mentioned in the text that chicken producers and their veterinarian or other advisers generally do not know about the potential of non-tannin sorghum to substitute for maize. So even in 2008 when it was briefly advantageous to do so, our sample did not report shifts to sorghum. Many of the producers we interviewed entered the business after 2008.

131 Returning now to above “a” where the cross price elasticity is zero and the sorghum price has not yet become relevant. So first we estimate the recent growth of demand for maize in the feed while sorghum is still not competitive. Taking out 2012 because of the war and the coup prices of maize came down from 120 to 100 over the period 2008-2014.This is an annual decrease of -3% over this period. With a price elasticity of demand for maize we use -1.2 as this is our estimate of the derived demand for broilers. So the falling price resulted in a 3.6% annual increase in the demand for maize as a feed. The effect of population growth was the same as that of the decreased price. Not surprisingly, the income growth had the largest effect at 4.8%. The total estimated annual demand growth for maize was then 12%. So this served as a stimulus for the rapid growth of the poultry sector and moderated the price decline of maize as the production and productivity of maize were rapidly increased. First the prices have to come down to “a”(.95) before it is rational to substitute sorghum for maize. With the rapid expansion of maize production and productivity maize has had the cost advantage. But as the poultry sector continues to expand rapidly maize gains will be constrained moving north out of the prime cotton sector. Moreover, with trade and greater sensitivity of maize to adverse weather especially as maize production attempts to move north, we would expect more yield volatility hence more short term potential for knowledgeable producers to 17 switch cereals source.

17 As in Texas, Kansas and Nebraska we would expect increasing use of sorghum in the feed taking advantage of the ability to produce sorghum in the drier regions with lower soil fertility. At a price historically approximately 10% lower than that of maize and less favorable policy incentives than maize, sorghum has retained its competitive position for feed in the US.

132 Table A-1. Growth in Demand for Cereal in the Feed

Demand for Characteristics Economic Population Annual Maize/sorghum Price Growth Effect effect- Effect Effect growth of demand for maize (sorghum) in the feed a b Annual 3.6% 4.8% 3.6% 12% 훾훾= -1.2;훽훽 =1.6 demand growth Cross price and for maize in income elasticities the feed, 2008- for maize as a 2014. feed. Population growth rate of 3.6. c c Potential Same elasticities 1.3% 4.8% 3.6% 9.7% demand growth as above but now for sorghum the demand below “a.” growth is for sorghum a. The growth effect from the changes in the price of maize. b. This is the annual increase in demand growth for maize over 2008-2014. c. This is the effect of the falling sorghum price relative to maize after “a.” d. Once the ratio fell below 0.95, this became the demand for sorghum as the price ratio continued to fall from .95 to .90 over a five year period. It took five years here to reach 0.95 from a price ratio of 1.21 (average of 2013 and 2014) with a price ratio decline rate of - 4.7%. Once reaching 0.95 we assumed that the rate of relative price of sorghum to the maize price relative to maize would slow up to 1.07% in falling from 0.95 to 0.90 over five years. Then we calculated above the annual growth of the demand for sorghum for feed over this five year period by multiplying by the cross price elasticity.

Futher Notes: Population growth. Mali is growing at 3.1% according to the World Bank (World Development Report 2014, p. 296).The latest Malian survey shows an even higher figure at 3.6% (Mas Aparisi A., Diallo F., Balié J., 2013, p. 9) Per capital economic growth. In Mali with the war, a coup, and continuing stagnation were –0.44% in 2012 (World Development Report 2014, p. 296). Mali rebounded in 2013 and an optimistic growth rate of 6.6% results in in per capita growth of 3%.

After reaching .95 as over the next five years the relative price ratio falls to .9, the rate of annual relative price decline would be 1.07%. With a cross price elasticity of -1.218 this would be a relative price term increasing the demand for sorghum for substitution of maize at a 1.28% annual growth for sorghum demand. Note that this total demand growth is driven principally by the economic and population growth. As with maize the price term has much less effect in increasing the demand for sorghum than the income effect. The sorghum price effect is also only

18 In the interviewing we attempted to informally estimate chicken producers’ cross price elasticities of demand. From their answers we ended up roughly dividing responses into two groups. The first group we called the perfect knowledge group. They would not switch at all unless they were assured that they could attain nutritional equality with maize. Then they would switch entirely with small price changes below the 95%. Since perfect knowledge or large scale shifts are both unlikely we present here the response of the group that gets neutral information from their vet or other specialist. Rather than tell them only to use maize as almost all the vets and other technical support staff do presently, they get neutral information between the two cereals. The vets recognize near equality in nutrients but do not make recommendations. Even with this scenario the cross elasticities were incredibly high due to total switching so we just used a reasonably high cross price elasticity of -1.2 here But interviewing did indicate that these chicken producers are very anxious to make even small cost savings given the pressure from many new entrants into broiler production and the falling prices for chicken meat. Their requirement for switching is that these changes do not have productivity effects. 133 half the population growth effect. However, this 9.7 percent annual growth of the demand for sorghum in the feed is impressive. The critical thing for sorghum is to increase yields and reduce the relative price of sorghum to maize so that the switchover to sorghum is encouraged. The positive potential of this growth results from the advantage of sorghum over maize moving north as the demand for chicken continues to increase rapidly.

134 Appendix B. Measuring Efficiency of Intensive Poultry Production in Burkina Faso There are two measures here that combine production and economic efficiency. The usual production efficiency number is the quantity of feed per kg of weight gain in the batch of broilers. By making the numerator the cost of the feed this takes into account the ability of the producer to find a lower cost but good feed. The measure also reflects marketing ability because if he has to sell his chickens over a longer time period he will continue to have to feed them and his feed costs increase. This is the measure in the second column. The measure in the third column is the value of chicken produced per unit of feed cost and this includes the two components above plus the ability of the broiler producer to find good markets and higher prices. So this is the best total measure combining different economic components, feed efficiency, time of sale and price of the broilers at sale. The highest efficiency with this measure was attained with larger batches of 1,000 to 1,500 birds but not all producers with larger batches were equally efficient. Low mortality rates were generally associated with the highly efficient producers.

135 Table B-1. Economic Measures for Broiler Production in Ouagadougou, Burkina Faso

Number Feed cost per kg Value of chicken Number of % of mortality of Total cost (in CFA Frs) for the of feed of chicken produced per chickens in the the batch production of one batch until a b producer produced unit of feed cost batch or flock sale

1 813 2.0 1500 5 2440000

2 822 2.7 1000 5 1250000

3 490 2.6 500 5 675000

4 448 4.0 1500 2 1106950

5 916 2.3 1000 5 1375000

6 359 5.0 1500 6 962000

c 7 33.3 8 868 2.3 500 4 625000

9 471 4.1 1000 1.5 1140960

10 712 3.0 500 3 622916

11 679 2.7 600 4 875000

12 997 1.1 500 2 950000

13 1372 1.8 500 10 988000

14 1014 2.0 1000 3 1500000

15 1223 1.5 600 3 1050000

16 871 1.5 500 2 950000

17 880 1.2 750 2 1375000

18 421 5.2 1000 1 885000

19 686 3.3 1000 3 1245000

20 748 3.3 1000 5 1171875 Source: Unpublished interviews with broiler producers in Ouaga, Nov.-Dec. 2015 a. This is confidential cost data so we do not list producer names b. This is the number of broilers raised together in the same confined space but they were generally not sold at the same time. Hence, the cost of production data reflects not only feed efficiency in production but also the marketing ability. c. Cost data are not really relevant as lost one third of chickens to disease.

136

Table B-2-Determinants of feeding costs per kg of carcass weight of chickens sold

VARIABLES Cost of Feed/kg of chicken

Imported feeda (Yes=1,No=0) -235.4* (112.7) b Number of Birds(No.) -0.124 (0.162) Constant 1,042 (138.1)

Observations 19 R-squared 0.319 a. Includes concentrate without protein and concentrate with protein. In all cases the cereal is added locally. b. Size or scale factor Standard errors in parentheses ** p<0.05, * p<0.1

Table B-3 Value of chicken sold per unit of feed cost with respect to marketing strategy and size (1) VARIABLES Value of chicken/Cost of Feed

Entirety of stock was sold at one time 0.803 (0.530) Number of Birds(No.) 0.00160** (0.000736) Constant 1.632** (0.659)

Observations 19 R-squared 0.254 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

137 Table B-4 Value of chicken sold per unit of feed cost with respect to weight and scale

(1) VARIABLES Value of chicken/Cost of Feed

Was the average weight above 1.8 kg? 0.842* (0.478) Number of Birds(No.) 0.00130* (0.000680) Constant 1.746** (0.656)

Observations 19 R-squared 0.285 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

138 Top picture on back cover: Millet food processors of Bamako, Mali at a workshop

Bottom picture on back cover: Method of women getting organic fertilizer to their farms. Millet production in Mopti, August 2015