g: ‘9 V‘ a; .. ......... . _ . a ' ' a “b '1 i" v- _ :w~_ ‘ 1 -w'flp.‘ ‘_ . _ “4f. . M‘ : M.- .. ...._.. E... J—J‘; Iv‘. . fl... “£329 . “A," ’1‘1‘" 511 ['1 "'L r: 1‘3”!“ ‘x la 4 . . .. w.“ u #3:”: L;- -.. ‘19-». -..—I "\ n~gl v ~..L 4. v.....-.~.. . .h 'J .., ~w~vas im. . . .“§u a -.-“v 5.. .y- 5.. o m; ’ «as as '7‘? m ”v 11,1.“ 141'.» ‘ "en“:w} ' v ,. :- I ‘ ' 16 - v “3 . . 1 1 . . . a ‘ "1L 631:1 11'1“ 1’. x “in“ 1?)?" he; {Zr‘( 1&1 ’1»;- flag}? . ‘.'1.-4.‘. ’ ' {I 015-”; 3313;1‘rqa‘qR‘i ‘3 ~79)“: ‘ ’4 553-. ---v. . I $—; .. ..¢~‘ ma Ir. I mi". {’1’1 3-1) '1 _ _ . Z {kg/E 11.11.19155183“! 3:; 74'“. u :gL‘igu‘) 155:. ’ QM .\. If! "I D I . 1‘5: . g-gi‘.‘ 11$“ ’3‘." 3%,; ~ .43, “33% 4. gr 3%? W: ~. 5‘. l‘l" 9,. u. “‘45? ‘.‘.‘ . “mm-Aw 9 . 's .J‘. 6‘ ‘ mm Lt 3.230% 3:: 4 “”1111lllllll‘llllI L LIBRARY Michigan State University __.,_. . This is to certify that the thesis entitled EVALUATION OF Sl~1ALLHOLD ER AGRICULTURAL TECHNOLOGY IN MALAWI presented by Odinga Arthur Hudson Jere has been accepted towards fulfillment of the requirements for Master's Agricultural Economics degree in Major professor Date April 2, 1990 0-7639 _ MS U is an Aflirmative Action/Equal Opportunity Institution ,—J PLACE IN RETURN BOX to remove this checkout rrom your rooord. TO AVOID FINES return on or baton duo dun. DATE DUE DATE DUE DATE DUE MSU Is An Affirmdivo Action/Ema! Opportunity Institution EVALUATION OF SHALLHOLDER AGRICULTURAL TECHNOLOGY IN MALAWI BY Odinga Arthur Hudson Jere A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Agricultural Economics 1990 ABSTRACT EVALUATION OF SMALLHOLDER AGRICUDTURAL TECHNOLOGY IN MALAWI BY Odinga Arthur Hudson Jere This thesis contains an analysis of smallholder farmer intercropping technologies from a farming systems and national income perspective. The problem is low adoption rates due to recommendations that are based on biological factors only. Marginal rate of return analysis is used to assess the financial profitability of technology. Primary data were obtained from a farming systems research trial program on maize, common beans and soybean intercropping. Farmers' preferences are used to assess possibilities of adoption. Economic analysis including calculation of social profitability and domestic resource cost ratios is used to assess the value of the program to the nation. The results show that to improve farmer income, it is important to pay attention to financial and economic ii analysis rather than just yield. Intercropping treatments perform better than sole crops indicating that intercropping should be given more emphasis. X45: both soybean and common bean intercropping technologies, Malawi has comparative advantage when measured by net social profit and domestic \ resource COSt ratios. iii To my wife Judith, Mom Catherine and Dad Inkosana Hudson Khwechisa. iv ACKNOWLEDGEMENTS This thesis would not have been possible without the input and support of many others. Eric Crawford, my major professor, committee members Richard Bernsten and Russell Freed never tired of reading my drafts and providing guidance. Other professors particularly Carl Eicher and Michael Weber gave me numerous documents related to my country and work. Michael Weber worked hard for me to go home and conduct a rapid appraisal survey. The years spent on the campus interacting with faculty, secretarial staff and fellow students were a great learning experience. My Malawian colleagues were very helpful. Mention should be made of wyson Mughogho, Thomas Gillard-Byrs, Dick Tinsley, Pickford Sibale, Michael Selenje, Edward Chisala, Steven Kandeya and Lenox Kisyombe. A particular note of thanks is in order for my wife, Judith, who helped me with initial data tabulations and offered moral support. Financial support was provided by Malawi Agricultural Research and Extension Project. TABLE OF CONTENTS LIST OF Tums O O O O O O O O O O O O O O O O O O O O O Xi Chapter 1. INTRODUCTION. . . . . . . . . . . . . . . . . . . 1 1.1 Problem setting and research focus. . . . . . . . . 1 1.1.1 Smallholder development strategies . . . . . . . 2 1.2 Objectives. . . . . . . . . . . . . . . . . . . . . 4 1.3 Research hypothesis . . . . . . . . . . . . . . . . 4 1.3 Overview of methodology . . . . . . . . . . . . . . 6 1.4 Organization of thesis. . . . . . . . . . . . . . . 7 Chapter 2. AGRICULTURE AND AGRICULTURAL RESEARCH IN MALAWI . . 8 2.1 Cropping systems and role of maize, common beans and soybeans. . . . . . . . . . . . . 8 2.1.1 Maize (1gg_may§_L). . . . . . . . . . . . . . . . 8 2.1.2 Common beans (Rngfigglu§_ynlga;1§_L.). . . . . . . 9 2.1.3 Soybeans (Qlyging_max_L . . . . . . . . . . . . . 10 2.1.3.1 Research work on soybeans . . . . . . . . . . . 12 2.1.3.2 Soybean market . . . . . . . . . . . . . . . . 12 2.1.4 Intercropping systems . . . . . . . . . . . . . . 13 vi 2.1.5 Sole cropping systems. . . . . 2.1.6 Farmer circumstances . . . . . 2.2 Agricultural research. . . . . . 2.2.1 Agricultural research with a farming systems perspective . . Chapter 3. CONCEPTUAL FRAMEWORK AND REVIEW OF LITERATURE. 3.1 Assumptions about farmer goal orientation 3.1.1 Food security . . . . . . . . . 3.1.2 Social obligations . . . . . . 3.1.3 Cash-generating activities . . 3.1.4 Profit maximization . . . . . . 3.1.5 Risk avoidance . . . . . . . . 3.1.6 Opportunity cost of labor . . . 3.1.7 Opportunity cost of capital . . 3.2 Marginal analysis as a model for evaluating on-farm trials . . . 3.2.1 Advantages and disadvantages of marginal analysis . . . . . 3.2.2 Farmer opinions . . . . .. . . 3.3 Economic analysis . . . . . . . . Chapter 4. THE EXPERIMENTAL PROGRAM . . . . . 4.1 Year One . . . . . . . . . . . . 4.2 Year Two . . . . . . . . . . . . 4.3 Year Three . . . . . . . . . . . vii 13 14 17 18 20 20 21 21 21 22 22 22 22 24 25 26 27 29 29 30 3O 4.4 Year Four . . . . . . . . . . . . . . 4.5 Trend across years . . . . . . . . . 4.6 Experimental design . . . . . . . . . 4.7 Fertilizer application . . . . . . . 4.8 Statistical inference . . . . . . . . Chapter 5. FINANCIAL ANALYSIS . . . . . . . . . . 5.1 Data . . . . . . . . . . . . . . . . 5.1.1 Description of Lilongwe ADD surveys 5.1.2 Summary of the 1989 rapid appraisal survey . . . . . . . . . 5.1.3 Average yield . . . . . . . . . . . 5.1.4 Adjusted yield . . . . . . . . . . 5.2 Methodology . . . . . . . . . . . . . 5.2.1 Partial budgets . . . . . . . . . . 5.2.2 Dominance analysis . . . . . . . . 5.2.3 Rate of return analysis . . . . . 5.2.4 Returns to labor analysis . . . . . 5.2.5 Food security . . . . . . . . . . . 5.2.6 Minimum returns analysis. . . . . . 5.2.7 Field crop prices . . . . . . . . . 5.2.7.1 Maize prices . . . . . . . . . . 5.2.7.2 Pulse prices . . . . . . . . . . 5.2.8 Produce transportation . . . . . . 5.2.9 Labor time per operation ‘. . . . . 5.2.10 Opportunity cost of labor . . . . viii 31 31 32 33 35 4O 4O 4O 43 44 44 46 46 46 47 47 48 48 49 50 51 51 52 52 5 O 3 Results 0 O O O O O O O O O O O O O 0 5.3.1 Thiwi Lifidzi local maize Intercropping 5.3. 5.3. recommendation domain . . . . . . . 2 Thiwi Lifidzi Mle intercropping recommendation domain . . . . . . . 3 Lilongwe RDP local maize intercropping recommendation domain 5.3.4 Lilongwe RDP MH12 intercropping 5.3. recommendation domain . . . . . . . 5 road 0 O O O O O O O O O O O O O 0 Chapter 6. 6.1 ECONOMIC ANALYSIS . . . . . . . . . . Social profitability and domestic resource costs . . . . . . . . . . . Calculating social profitability . . Social profitability implications . . Calculation of resource cost ratios . Some implications of the DRC analysis Sensitivity analysis . . . . . . . . Chapter 7. CONCLUSIONS . . . . . . . . . . . . . 7.1 7.2 7.3 Biological analysis . . . . . . . . . Financial returns . . . . . . . . . . Food security . . . . . . . . . . . . 7 O 4 RiSk O O O O C O O O O O O O O O O O 7.5 Farmer preferences . . . . . . . . . ix 53 54 57 61 65 7O 72 72 74 82 85 87 87 93 93 93 94 95 95 7.6 Intercropping systems . 7.7 Government policy . 7.8 Underestimating intercropping benefits APPENDIX 1. Marginal analysis tables . LIST OF REFERENCES 97 97 100 101 106 LIST OF TABLES 1321:. Bass 1. Lilongwe and Thiwi Lifidzi RDP maize, common beans and and soybeans intercropping trial treatments . . . 33 2. Thiwi Lifidzi 1989 MH12 intercrop statistical 3. 4. 10. 11. 12. 13. 14. 15. analysis (kg/ha) . . . . . . . . . . . . . . . . . 35 Thiwi Lifidzi 1989 local maize intercrop statistical analysis (kg/ha) . . . . . . . . . . . . . . . . . 36 Lilongwe RDP 1989 M312 intercrop statistical analYSis O O O O O O O O O O O O O O O O O O O O O 37 Lilongwe RDP 1989 local maize intercrop statistical analysis O O O O O O O O O O O O O O O O O O O O O 3 8 Thiwi Lifidzi and Lilongwe RDP recommendation domains (1989). . . . . . . . . . . 42 Soil landscape management units by RDP. . . . . . . 43 Lilongwe ADD 1989 labor requirement per farm operation (1989). . . . . . . . . . . . . 53 Marginal analysis for 10 different sole and intercropping treatments of local maize, vulgaris beans and soybeans tested in Thiwi Lifidzi RDP . . 55 Marginal analysis for 10 different sole and intercropping treatments of local maize, Vulgaris beans and soybeans tested in Thiwi Lifidzi RDP, urban wage rate. . . . . . . . . . . . . . . . . . 56 Minimum returns analysis (1985-1989) for Thiwi Lifidzi RDP. O O O O O O O O O O O O O O O O O O O O O O O 57 Marginal analysis for 10 different sole and intercropping treatments of M312 maize, Vulgaris beans and soybeans tested in Thiwi Lifidzi RDP . . 59 Marginal analysis for 10 different sole and intercropping treatments of MH12 maize, Vulgaris beans and soybeans tested in Thiwi Lifidzi RDP, urban wage rate . . . . . . . . . . . . . . . . . 60 Minimum returns analysis (1985-89) for Thiwi Lifidzi RDP M312 intercropping trials. . . . . . . . . . . 61 Marginal analysis for 10 different sole and intercropping treatments of local maize, Vulgaris beans and soybeans tested in LRDP. . . . . . . . . 63 xi 16. 17. 18. 19. 20. 21. 22. 23 24. 25. 26. 27. 28. 29. 30. Marginal analysis for 10 different sole and intercropping treatments of local maize, Vulgaris beans and soybeans tested in LRDP, urban wage rate Minimum returns analysis (1987-1989) for LRDP local maize trials . . . . . . . . . . . . . . . . . . . Marginal analysis for 10 different sole intercropping treatments of MH12 maize, Vulgaris beans and soybeans tested in LRDP. . . . . . . . . Marginal analysis for 10 different sole and intercropping treatments of MH12 maize, vulgaris beans and soybeans tested in Thiwi Lifidzi RDP, urban wage rate . . . . . . . . . . . . . . . . . Minimum returns analysis (1987-1989) for LRDP M312 intercropping trials . . . . . . . . . . . . . . . On-farm import parity price of maize (1989) . . . . . . . . . . . . . . . . . On-farm import parity price of soybeans (1989) . . . . . . . . . . . . . . . . On-farm export parity price of Vulgaris beans (1989) . . . . . . . . . . . . . . On-farm import parity price of DAP fertilizer (1989). . . . . . . . . . . . . . . On—farm import parity price of urea (1989) . . . . . . . . . . . . . . . . . . Social profitability by recommendation domain (MK/ha). . . . . . . . . . Net farm margins (MK/ha) by recommendation domain . . . . . . . . . . . . . . Domestic resource cost for each recommendation domain . . . . . . .. . . . . . . . Social profitability (MK/ha) sensitivity analysis by recommendation domain (1989) . . . . . Domestic resource cost sensitivity analysis by recommendation domain (1989) . . . . . . . . . xii 64 65 67 68 69 75 76 77 78 79 83 84 87 9O 91 CHAPTER 1 INTRODUCTION MW Malawi is located to the south of the main African Rift valley. The latitude ranges from 9 to 17 degrees south and the altitude ranges from 500 to 3000 meters. Agriculture is the most important sector providing about 40% of the gross domestic product and 89% of the export earnings. About 85% of the people are smallholder farmers who concentrate on food production. The population is 8 million and the growth rate is 3.2% per annum. Suffering from the neo-Malthusian specter of rapidly increasing population in the face of stagnating food production, Malawi could highly benefit from technical change that raises yields. Technology, in this instance, entails a combination of all the management practices for producing or storing a crop or crop mixture and/or livestock and livestock products. Each practice is defined by timing, amount and types of various technological components such as variety, land preparation, fertilizer application or weeding. A subsistence farmer who purchases no input is, however, using technology (CIMMYT, 1985). l 2 Wise Within the smallholder sector the Government has emphasized two broad strategies: the transformation and the improvement approaches. The objective of the transformation approach was to provide focal growth points through the introduction of new technical, social and tenure systems in an attempt to modernize and increase smallholder productivity (Mkandawire, 1988). The transformation approach has been mainly associated with four major integrated agricultural projects1 established in the late 19605. Agricultural development in the nonproject areas and areas not covered by the settlement schemes or estatesz, has followed the improvement approach. This emphasizes the progressive improvement of methods of crop and animal husbandry among smallholder farmers, to induce them to produce without radically changing their traditional social structure and tenure systems (Mkandawire, 1988). Since 1976/77 period, the Government has attempted to extend the experiences of the transformation approach into areas which 1The Integrated Rural Development Projects were launched in 1966 in order to increase agricultural production and improve the standard of living. Four projects were located in areas with high productive potential. The four projects covered about 20% of the population. 2The estate sub-sector produces for export on leasehold land using advanced production technologies. It employs about 60% of the remunerated labor force. 3 were previously associated with the improvement approach, under the National Rural Development Program3. The underlying assumption of these strategies was that through a process of social communication, the relatively better farmers will diffuse innovations to the rest of the farming community, despite differential resources. This justifies a progressive farmer strategy. The main arguments for pursuing this strategy is that production resources are limited and it is not possible to provide resources to every farmer. Rather improved technology should be directed to farmers who already have above average resources and are willing to make intensive use of them. This strategy thus concentrated on the better off farmers only. The low income farmers who have small land holding size, experience peak period labor shortage, periodic food shortage, have no access to credit and a host of other constraints, were excluded. A common precaution taken by the low resource farmers is to diversify the crops they grow through intercropping and/or sequential cropping. Until the early 19803, little or no research was done on these systems because the emphasis was on'the larger smallholders who comprise of only 25% of the smallholder farmers. The 3The National Rural Development Program began in 1976/77 with the aim of spreading agricultural development throughout the country. This program covers the whole nation. 4 strategy adopted contributed to low adoption rate because there was no trickle down effect. mm The major objective of this study is to evaluate smallholder farm technology more comprehensively by including financial and economic analysis, rather than relying on statistical analysis only. Financial analysis, which uses market prices, is useful for determining the profitability of proposed changes to the target group, and consequently for assessing how widely recommendations will be adopted. Financial analysis is distinct from economic analysis (Gittinger, 1984 p. 473). Economic analysis, which uses shadow prices, measures the social profitability at the national level. Economic analysis also measures the comparative advantage of a technology when measured against possibilities of trade. Evaluation of a new technology like soybean intercrop is important because it estimates the cost of domestic currency required to save a unit of foreign currency. lil_B£§§iI£h_h!DQ§h§§i§ The data used in this study were obtained from the Lilongwe Agricultural Development Division4 Adaptive 4Malawi has administratively decentralized its agricultural program into eight contiguous areas called Agricultural Development Divisions (ADD). Each ADD has an Adaptive Research Team (ART). 5 Research trial program. After extensive surveys, it was hypothesized that the farmers would maximize returns to their resources by increasing the density of vulgaris beans and soybeans in the local maize intercrops. Local maize is grown intercropped with Vulgaris beans and to a lesser extent with soybeans. The government is currently encouraging the growing of soybean in order to reduce imports of soybean products. It was also hypothesized that intercropping the pulses with hybrid maize would increase financial returns. Hybrid (M312) is mostly grown in pure stands by a smaller proportion of the farmers. Most hybrid growing farmers get seasonal loans from the government. If pulses increase financial returns, they may help repay loans. In some recommendation domains vulgaris beans intercropping can perform better than soybean intercropping and vice versa. If it proved profitable to grow soybeans in an intercrop with either local maize or uniz, then soybean can be grown at a larger scale and apart from being consumed at home it could be sold for cash. This would contribute to improving the access of families to basic food requirements through improved home production as well as increased incomes. The country can save on foreign exchange by producing sufficient soybeans and soybean products rather than importing soybean products. When this experiment was being designed, a big increase in the official price of soybean was being proposed which 6 made ex-ante analysis of the crop attractive. A 55% price increase was achieved in 1988 for the 1988/89 cropping season. An extra 28% increase was achieved in October 1989 for the 1989/90 cropping season (MDA, 1989b). Since local maize and hybrid maize respond differently to management and are grown by farmers of different categories, they should be tested separately. They are regarded as two different enterprises because hybrids are grown for cash by a small proportion of farmers as sole crops. Local maize is mostly intercropped with pulses. Farmers indicate that they prefer local maize for food because it has superior storage, processing and culinary qualities. Little or no fertilizer is applied to local maize. Interplant and intraplant competition is known to increase with increasing plant population. It is necessary to increase or maintain the yields of maize because it is the staple food. Intercropping local maize or MH12 with either vulgaris beans or soybeans may reduce the yield of maize and consequently returns to resources used. 1W On-farm research data is subjected to statistical analysis, financial analysis and economic analysis. Analysis of variance is used to evaluate the mean yields of the intercropping treatments for statistical significance. Marginal returns analysis is used to evaluate the financial returns to variable inputs. Farmer preferences are included 7 in the financial analysis. Analysis of comparative advantage--social profitability and domestic resource cost ratios--is performed for the best technologies. WW5 Chapter 1 identifies the problem and discusses the objectives and the hypothesis. The problem is low adoption rates because of a progressive farmer approach to development and of the practice of deriving farmer recommendations based entirely on biological and statistical analysis. The objective is to include financial and economic analysis in deriving recommendations. Chapter 2 discusses Malawian agriculture, emphasizing maize, Vulgaris beans and soybeans intercrops. These are the technologies evaluated in this paper. Chapter 3 dwells on concepts, review of relevant literature and discusses the assumption on which the analysis is based. Chapter 4 discusses the experimental treatments, design and the trend of the results over a period of four years, and the statistical analysis of the final season. Chapter 5 discusses the financial evaluation of the treatments including marginal analysis and farmer preferences. Chapter 6 discusses the economic analysis including social profitability and domestic resource cost ratios. Chapter 7 draws conclusions from the analysis. Chapter 2 AGRICULWURE AND AGRICULTURAL RESEARCH IN MALAWI Maize is the most important cereal and food crop in Malawi. It is a dietary staple for over 80% of the population, with an average consumption of 225 to 250 kg per capita annually. Ninety percent of the maize is produced primarily by smallholder farmers who generally use few purchased inputs. Approximately 60 to 70% of maize is grown intercropped with various crops. Maize research dates back to 1947 when agricultural research began in Malawi, but until the mid 1980s no research was done on intercropped maize. Greatest benefits would flow from higher maize yields because maize occupies the majority of the cultivated land. About 63 to 90% of the area is planted to maize (Kangaude, 1988). Much of the balance of the land is given to other 5 food crops including groundnut, pulses and roots. Only 5Pulse is a common name for Vulgaris beans, soybeans and other seeds that grow in pods and are used as food. 9 3.5% of the total smallholder area is allocated to nonfood cash crops: cotton and tobacco (Carr, 1988). Increased efficiency in maize production for the market is also crucial as this could reduce food prices (Kydd, 1989). WWW Vulgaris beans are the most important food crop in Malawi after maize (World Bank, 1989). Vulgaris beans are used dry, although green shelled seed and immature pods are also consumed as food. Vulgaris beans provide a cheap source of protein for the majority of the people. vulgaris beans are also high in calcium, magnesium and iron and contain a large amount of vitamin 8. Green pods and green shelled seeds are a good source of vitamin A and C. Other than being a food crop, surplus beans are sold to neighboring countries to earn foreign exchange. Generally, most of the beans are grown in areas between 1,000 and 2,000 meters above sea level, with well-drained soils and rainfall of about 800-1500 millimeters. Lilongwe is one of the major bean-producing areas (Msuku, 1984). Not much research has been done on intercropping although about 99% of the pulses in Malawi are intercropped with maize (Mkandawire et al., 1989). In Lilongwe Agricultural Development Division (LADD) Vulgaris beans are grown by almost all farmers. But the 10 production is very low in some rural development projects.6 LADD/ART7 (1984) survey report indicates figures as low as 133 kgs per hectare in Thiwi Lifidzi RDP. Trial data indicate yields as low as 24 kg per hectare (LADD/ART, 1989). WWW Soybeans have been grown in Malawi since 1909. Between the early 19403 and mid 1950s about 200 to 900 tons were exported annually mostly to the United Kingdom, Holland and Norway. The soybeans were intercropped with tung in the Southern Region8 (Sibale, 1989). Soybeans have been grown commercially by the estate sector for a long time but its production in the smallholder sector has been negligible. In 1981, the Women's Program and the Food and Nutrition section of the Ministry of Agriculture introduced the preparation of soybeans to women extension agents. The Food and Nutrition Program emphasized the crop because the Malawian diet is particularly deficient in fats and to a lesser degree deficient in protein; soybeans contain about 40% fats and 20% protein. The teaching of recipes for 6ADDs are subdivided into Rural Development Projects (RDPs) whose determination conforms to administrative, political and traditional boundaries, population, settlement patterns and environmental characteristics. 7Lilongwe Agricultural Development Division Adaptive Research Team conducted the trials analyzed in this study. 8Malawi is administratively divided into Southern, Central and Northern regions. Southern region has 3 ADDs, Central region has 3 ADDs and Northern region has 2 ADDs. ll preparing soybean milk, porridge, snacks and relish made soybean popular among women in Lilongwe Rural Development Project (Spring, 1986). Recent studies confirm that whole soybeans are an excellent source of protein and offer a good distribution of amino acids. Soybeans also contain 20% unsaturated fats which are desirable for humans in areas where animal fat is scarce. In addition to providing nutritious weaning foods, soybeans can form an important ingredient in preparation of adult meals. It can be used to enrich porridge or nsimag. The anti—nutritional factors associated with soybeans, e.g., trypsin inhibitor, hemagglutinin and phytic acid can be easily removed under village conditions by heating the beans for about twenty minutes (Sibale, 1989). The characteristic off-flavor due to the enzyme lipoxygenase can also be easily removed by boiling the beans in water for twenty minutes. Cooking to achieve a desirable texture according to most Malawians can be done by adding readily available, inexpensive sodium-bicarbonate (Sibale, 1989). It should be noted however, that these are late developments which have not yet spread through the rural population. Most farmers do not regard soybeans as a relish dish because 9Nsima is the staple food made from maize flour. It is eaten together with vegetable, pulse, fish or meat relish dishes. 12 it does not mash well with nsima (Spring, 1986, and LADD/ART, 1989). W In 1984, soybean trials were conducted in Lilongwe ADD by the Women Programs section to determine smallholder farmer soybean husbandry practices. The results were grouped into three usable grain yield sub-categories of high (2,530 to 2,900 Kg/Ha), medium (1,160 to 1,400 Kg/Ha), and low (320 to 660 Kg/Ha). Farmer yields increased with increasing planting densities. The highest yields were obtained at a population of 30 plants per meter (Spring, 1986). Another survey done by the Women's Program section revealed that 38% of the women who had adopted soybeans were interplanting at a very low planting density: three soybean seeds on one station in between one-meter-spaced maize stations. The farmers were interested in the crop mainly for home consumption. Some farmers used soybeans to feed dairy animals and chickens. Farmers who tried to grow soybean for cash experienced marketing problems. Zililiz_§2!bssn_mszkst A recently concluded survey undertaken by the Soybean Improvement Project, which visited a number of agro- industries in April 1989, has indicated demand for domestically produced soybeans rising up to 10,000 tons by 1992 (Sibale, 1989). The soybeans will be processed into livestock feed and baby food. The current annual 12 it does not mash well with nsima (Spring, 1986, and LADD/ART, 1989). W In 1984, soybean trials were conducted in Lilongwe ADD by the Women Programs section to determine smallholder farmer soybean husbandry practices. The results were grouped into three usable grain yield sub-categories of high (2,530 to 2,900 Kg/Ha), medium (1,160 to 1,400 Kg/Ha), and low (320 to 660 Kg/Ha). Farmer yields increased with increasing planting densities. The highest yields were obtained at a population of 30 plants per meter (Spring, 1986). Another survey done by the Women's Program section revealed that 38% of the women who had adopted soybeans were interplanting at a very low planting density: three soybean seeds on one station in between one-meter-spaced maize stations. The farmers were interested in the crop mainly for home consumption. Some farmers used soybeans to feed dairy animals and chickens. Farmers who tried to grow soybean for cash experienced marketing problems. 2111112_§Q¥h§an_nszkfii A recently concluded survey undertaken by the Soybean Improvement Project, which visited a number of agro- industries in April 1989, has indicated demand for domestically produced soybeans rising up to 10,000 tons by 1992 (Sibale, 1989). The soybeans will be processed into livestock feed and baby food. The current annual l3 utilization of imported soybean products, mostly soybean oil, is in the region of 20,000 tones. The survey also showed that processors are rapidly expanding their capacity to process materials such as whole soybean with good prospects for installation of a solvent oil extraction plant assuming increased soybean production in the country (Sibale, 1989). Currently, the Government of Malawi is encouraging the growing of soybeans by smallholder farmers in order to reduce the importation of soybean products. 2iliA_In§£I§IQERIDQ_§¥§§§m§ Intercropping, a system in which two or more crops are grown in the same field during the same growing season (Mkandawire et al., 1989), is the dominant cropping system for the small scale farmers and will probably remain so for many years to come. Due to the fast rate of population growth and limited land for farm expansion, higher crop productivity per unit area has become necessary. Intercroping Vulgaris beans and maize is known to increase the yield of maize (Reintsma and Lang, 1988). Other advantages are better utilization of land and labor, risk' reduction and soil conservation and improvement (Yiwombe, 1988). MW Since the 19403, the Ministry of Agriculture has recom- mended growing crops in pure stands, because improved maize varieties performed best under sole cropping. Current research recommendations are based on sole crops. These are 14 in line with the smallholder development strategy followed until early 19803 (section 1.1.1). MW Farmer circumstances are all factors that influence farmer decisions about a technology. These can be divided into natural, social, financial and economic factors (CIMMYT, 1988). The financial, economic and environmental factors influence the types of crops and livestock raised. It is easy to make poor recommendations by ignoring factors that are important to the farmers. Some of these factors may not be very evident. A good recommendation can be thought of as the practices which farmers would follow, if they had all the information available to the researcher (CIMMYT, 1988). A good recommendation must at least increase net income or in some way relieve a constraint like labor or risk. It will also take into account the existing infrastructure and institutions and other factors that constitute farmer circumstances. A rapid appraisal survey was carried out in Lilongwe and Thiwi Lifidzi RDPs in 1989, by the author to study some of the farmer circumstances related to the trial programme analyzed in this paper. Some of the survey statistics are discussed below: 15 1. Most farmers preferred two hills of soybean and two hills of Vulgaris beans in between maize hillslo. In Lilongwe RDP (LRDP) 50% of the interviewed farmers preferred planting two vulgaris bean hills in between maize hills due to vigorous vegetative growth and high yields. Vigorous vegetative growth is important because Vulgaris bean leaves are used as relish. Thirty-one and 19% preferred three and one bean hill between maize hills respectively. 'For soybeans, 68% and 32% preferred two and three hills in between maize hills respectively. In Thiwi Lifidzi, 75% preferred planting two vulgaris bean hills in between maize while 25% preferred one. Similar proportions apply to soybeans. 2. In both locations, most ADMARC11 buying points are still operating12 side by side with private traders. Thiwi Lifidzi RDP has 290 while Lilongwe RDP has 127 registered private buyers. At most buying points more than one buyer operates, e.g., there are 29 different buyers at Kupiliatu and 35 different buyers at Gulugulu in Thiwi Lifidzi RDP. In LRDP, Sinyala and Mitundu buying points had 8 buyers 10The intercropping trials were evaluating different densities of pulses in the intercrop by using the number of pulse hills planted in between maize hills (section 4.6) 11ADMARC is a statutory agricultural marketing board which sells inputs and buys output at pan-seasonal and pan-territorial prices. 12In line with the market liberalization policy ADMARC is supposed to stop most of its operations to give way to private traders and let market forces play a greater role in directing economic activity along efficiency lines. 16 each. However, there is little information regarding the private buyer prices and quantities. This makes it difficult to use private buyer figures. 3. The net market price of maize during the food shortage period is almost the same as the ADMARC price at harvest. The storage cost of local maize from the time of harvest to the period of peak food shortage is MK0.042 per kg. The local market price at this time is MKO.31 per kg (MOA, 1989c). The net price is MKO.268. ADMARC offers MKO.26 at harvest time. The difference in prices is minimal and not worth using in a sensitivity analysis. 4. Vulgaris bean seed is obtained from the local market and soybean seed is obtained from official marketing channels. All the farmers in LRDP obtained their seed from the local market while 79% of the farmers that grew soybeans obtained the seed from official channels. The analysis will use market price for Vulgaris bean seed and official price for soybean seed. All farmers interviewed sold their Vulgaris beans to the local market and all sold soybeans to ADMARC. The analysis uses local market price for Vulgaris beans and official price for soybeans. 5. Most farmers are aware of the activities of local money lenders. In LRDP, 50% of the farmers admitted knowledge of local lenders. In Thiwi Lifidzi, 75% admitted knowledge of local lenders. All farmers in both RDPs said the interest rate charged over the growing season, an average of 5 months, is 100%. One hundred percent will be 17 used as the minimum acceptable rate of return in the marginal analysis. 6. Farmers transport their crop by farm carts. The average cost for transporting produce from the field to home is MK1.50 per farm cart trip. An average farm cart carries an equivalent of four 100kg bags. The average cost for carting from home to the market is MK1.77 per 100kg bag. The total transport cost per kg is MK0.02145. 7. The amount of time required to plant depends on the pulse density. Time in minutes was estimated using the stop watch technique and it took farmers 19.38 for planting on the same hill, 23.5 for planting one pulse hill between maize hills, 28.1 for planting two pulse hills, 34.02 and 41.20 respectively for planting three and four pulse hills between maize hills. The net plot size was 4 ridges of 5.4 meters length and the spacing was 0.91 meters between ridges. When converted to a hectare this yields 22.86, 27.81, 33.14, 40.14 and 48.61 hours per hectare, respectively. MW Until the late 19703, agricultural research did not take into account nonbiological factors when developing recommendations for smallholder farmers. Statistical analysis, predominantly analysis of variance to evaluate the statistical significance of biological treatment effects, was the only tool used for analyzing trials (Phiri 1986). Agricultural research has traditionally been organized along l8 biological disciplinary lines without the involvement of social scientists. It has also been conducted on research stations under conditions not representative of the farmer conditions with little or no farmer involvement. This was in line with the transformation and improvement strategies which assumed a trickle down effect regardless of farmer circumstances (section 1.1.1). z12111_A9risultnral_rsssarsh_nitn_a_farnins_sxstsms nsrsnsstixs Research with a farming systems perspective has several objectives, the major one being to increase the productivity of a farming system by generating appropriate new technologies. This approach is often divided into: 1. Location specific research which has short-term objectives of developing improved technologies for target farmer groups. 2. Research conducted with a long-term perspective to overcome major, widespread limitations in the farming systems. Location-specific research is best implemented through on-farm research where farmers are involved in identifying potential technological improvements, which are then tested under their socio-economic and ecological conditions (CIMMYT, 1985) Direct communication between a multi-disciplinary research team and farmers through surveys and continuous monitoring of participating and non-participating farmers 19 increases the understanding of the farmer's decision-making environment and enables the researcher to identify technological options that are more consistent with the environment. Experiments under farmer conditions lead to estimates of yield and cost changes that better reflect what farmers can expect from using the technological options. Farming systems research also focuses on how policies influence the adoption processes. Marketing is one area of public policy addressed in this paper. The degree of government intervention in the market affects adoption. CHAPTER 3 CONCEPTUAL FRAMEWORK AND REVIEW OF LITERATURE WWW Hansen (1982) defined farming systems in Malawi as a complicated interwoven mesh of resources and factors: agronomic, economic, social, cultural, physical, etc. which are managed to a greater or lesser extent by the farmer. The farmer utilizes some of the technology that is available in an attempt to increase farm household utility within a given context of accepted preferences, aspirations, and socioeconomic conditions. Utility refers to a broad range of satisfactions. For a Malawian farmer, utility includes the production of food and relish for home consumption as well as generation of cash income. These, among other objectives, are the central issues in the farmer's planning. They determine which technology he/she adopts. The farming systems concept reflects the empirical complexity of the conditions surrounding smallholder agriculture and the complexity of the decisions that the smallholder farmers have to make. Most of their decisions are compromises in which farmers balance what they want to do against limited resources. Other compromises occur because farmer,s goals conflict, e.g., achieving higher 20 21 incomes versus lower risks. Farmers may fail to adopt an improved technology because they are simultaneously trying to satisfy a range of desires or necessities. Bearing in mind the complexity of the farmer's situation, this study will assume the farmer's decision is based on the objectives discussed below. The analysis will center around these assumptions. W Many farmers are primarily concerned with assuring an adequate food supply for their families. They do this by producing most of what their family consumes, e.g., maize and other cereals, root crops, pulses, and vegetables, and by marketing some of their other crops like tobacco, beans, potatoes or vegetables to obtain cash to buy food and other necessities. 21112.599131_92119§§12n§ The farm family is a part of a wider community toward which it has certain obligations such as to contribute towards festivities and to provide aid such as small gifts of food and money to poorer relations during difficult periods. WW9: The farmers participate in other cash-generating activities like beer-brewing, gin-distilling, baking and 22 selling cakes and other snacks, other rural small-scale industries and off-farm employment for those near the city. 111W Although farmers do not necessarily market all their crops, they are interested in the financial returns. Farmers will consider the cost of changing from one technology to another and the financial returns resulting from that change (CIMMYT, 1988). Thus the farmers weigh and assess the difference in net benefits between technologies and other sources of income and/or food. MW Uncertain markets and climate, since maize intercrops are grown under rainfed conditions, makes risk a big factor in the farmer's decision to adopt a technology. .Farmers may prefer a more stable technology, in terms of both physical and financial output, to one with high output but low stability. 14W Labor may be more productive in one farm activity compared to another or compared to nonfarm activities like rural small-scale industries or, if possible, employment in urban industries. Farmers will use their labor in activities that are most productive. WW1 Capital availability is of crucial importance. Farmers have low per capita annual incomes which makes capital accumulation difficult. A national sample survey of 23 agriculture in the 1980/81 season indicates a rural per capita cash income range of MK7913 to MKlll per annum (Reintsma and Lang, 1989). If farmers are asked to make additional investments in their farming operations, they will consider the cost of the money they invest in relation to the expected returns at a later point. Working capital is the value of inputs, whether purchased or owned, allocated to an enterprise with the expectation of a return at one point in time. The cost of working capital is the benefit given up by the farmer by tying up working capital in an enterprise for a time period (CIMMYT, 1988). This can be a direct cost in the case of someone who borrows money to buy seed and must pay an interest charge on it. Or it may be an opportunity cost, the earnings which are given up by not putting money or an input already owned to its best use (CIMMYT, 1988). Only 15% of the smallholder farmers utilize formal credit (Carr, 1988). The remaining proportion uses informal credit. Some of the reasons are failure to meet eligibility criteria for the formal credit, fungibility of the informal credit, and ease of re-negotiation. Local money lenders are the most common sources of these loans on which they charge 100% interest rate (section 2.1.6). 13 $1.00. Officially 2.47 Malawi Kwacha is equivalent to US 24 The cost of capital can be used as a benchmark rate of return. Any technology which has a rate of return higher than the cost of capital has a high chance of being adopted. This study will assume the minimum acceptable rate of return of 100% (see section 2.1.6). 9: ' lo. 21". 3 i E: '2 [[QA..- h' ‘ ._ .1 1' °!'--.,!!! trials... Economic models are appropriate for modelling the farm situation assuming farmers roughly and indirectly evaluate technology in terms of benefits and costs. The model used in this study is the marginal benefit-cost analysis which computes the rates of increase in benefits to costs at several incremental treatment levels (CIMMYT, 1988). This analysis compares the farmer's practice with other treatment levels. Marginal benefit-cost analysis considers only the cost of inputs which are variable. For noncash transactions like family labor, opportunity costs are estimated (CIMMYT, 1988). The researcher arranges treatments in ascending order according to total variable costs. Dominant treatments are selected by comparing benefits and costs of each treatment. Treatments which yield lower benefits but which are more expensive or cost the same as those yielding higher benefits are eliminated. These treatments are dominated and farmers will not adopt them. Marginal rates of return to variable factors are calculated by dividing the marginal increase in benefit by 25 the marginal increase in costs of adjacent treatments for all nondominated treatments. Experience and empirical evidence have shown that the minimum rate of return acceptable to farmers is between 50 and 100% (CIMMYT, 1988). This analysis can incorporate risk by including minimum returns analysis which examines the variability of the net benefits. Minimum returns analysis compares the average of the 25% lowest net benefits of the proposed technology with the average of the 25% lowest net benefits of the farmer's practice. If the minimum returns analysis shows that the net benefits of the proposed technology, from individual sites and over a period of time, are above the net benefit of the farmer practice, then the recommendation has a high chance of being adopted (Phiri, 1986). The overall aim of the marginal analysis is to derive recommendations consistent with farmer desired and expected food productivity and income generating goals, avoidance of risk and the best possible use of scarce resources (Dillon and Hardaker, 1980). The starting point in on-farm research is the assumption that it is much better to consider relatively small improvements in farmer's practice rather than to propose large-scale changes. The idea is to ask what changes can be made in the present system. The focus is on the difference between two treatments. The advantages of marginal analysis include ease of performing the analysis 26 because most of the needed data can be obtained easily from designed trials. Another advantage is that the method uses only the costs that vary so that only the additional costs and benefits of technologies are needed for the analysis. Marginal analysis, using discrete points can be used for any type of experiment whereas continuous analysis is only applicable to factors that vary continuously such as fertilizer rates or seed rates. The disadvantage is that certain experiments such as those that look at different varieties or modest changes in seed rate, involve changes in costs that may be very small. If the resulting yield differences are substantial, the resulting marginal rates of return can be very large. A good example is the returns to labor analysis in this study. Moving from technology 4 to technology 5 in Lilongwe RDP MH12 intercropping systems, the returns to labor are MK71.43 per hour (table 5 in the appendix). This is extremely high compared to the assumed opportunity cost of labor of MKO.21 per hour because of the slight differences in additional labor requirements between these two technologies. Such marginal returns are not very useful in comparing treatments (CIMMYT, 1988). While marginal analysis is a powerful tool, it must be seen only as part of a research strategy. Wining: It is important to solicit the farmer's reactions to proposed technologies. Alternatives that seem to be 27 promising both agronomically and financially may have other drawbacks which only farmers can identify. However, it is not easy to get the farmers to make straightforward suggestions in the parts of the world where rural people feel outwardly inferior to urban-based college-educated professionals (Kishindo, 1988). Farmers' opinions are summarized in section 2.1.6. WW Comparative advantage-~expression of the efficiency of using local resources to produce a particular product when measured against the possibility of foreign trade (Morris, 1989a)--is used to estimate the value of soybean and Vulgaris bean intercropping technologies to the nation. Knowledge of comparative advantage is important because potential welfare gains from specialization and trade can be used to foster economic growth. National income can often be increased in the short run through policies that encourage farmers to produce commodities which exploit existing patterns of comparative advantage. Over the long run additional welfare gains can be assured if research resources are used to strengthen comparative advantage. Agricultural policies and research resource allocation should be based, at least in part, on comparative advantage considerations (Morris, 1989b). During the 19803, Malawi implemented policy reforms designed to reduce state participation in agriculture, increase productivity, liberalize commodity trade, and free 28 up market forces to play a greater role in directing economic activity along efficiency lines. At the same time increased pressure was put on research administrators to ensure the cost-effectiveness of agricultural research expenditures. These developments are designed to take advantage of comparative advantage. However, comparative advantage is difficult to determine empirically (Morris, 1989a). Simply comparing shadow production costs between two regions or countries is often inconclusive, because comparative advantage is not directly related to absolute levels of production costs. Even if relative production costs are known, frequently these are distorted by policies or market failure. CHAPTER 4 THE EXPERIMENTAL PROGRAM The experimental treatments were selected for testing because an ex-ante analysis showed financially acceptable returns, they are compatible with the farming systems, and they do not present special risk to farmers. The experimental treatments were modified with the acquisition of more information resulting from researchers getting more acquainted with the farmers and the areas. 111.293119ns 1985/86 season was the first year of the experiment. It had six treatments: a sole crop of maize, maize and beans planted on the same station--the most common farmer practice, one and two vulgaris bean stations between maize stations, and one and two soybean stations planted in between maize stations. The farmer practice was included as one of the treatments in the experiment so that both farmers and researchers could compare the farmer's practice with the alternatives. These treatments showed significant yield differences and increasing financial returns with increasing soybean and Vulgaris bean densities. 29 30 $12_1§§I_I!Q In 1986/87, the second year of the experimental program, the number of treatments was increased to allow for higher soybean and Vulgaris bean densities. The soybean and Vulgaris bean stations were increased to three in between maize stations. The seven treatments were a sole crop of maize, maize and Vulgaris beans on the same station, one, two and three Vulgaris bean stations in between maize stations, one, two and three soybean stations in between maize stations. These treatments confirmed an increase in yield and financial returns with increasing densities of both vulgaris beans and soybeans. 512_X£§I_IDI£§ In 1987/88 the experiment was further modified to allow for increasing soybean and Vulgaris bean densities. The stations of soybean and Vulgaris beans were increased to four in between maize stations. The treatments with one soybean station in between maize stations and one Vulgaris bean station in between maize station were dropped so that the experiment would examine relatively few factors at a time. Sole crops of Vulgaris beans and soybeans were introduced. The trend of the results changed. There were no significant differences between different densities of intercropped soybeans and there were no significant differences between different densities of Vulgaris beans. 31 $1A_XEQI_IQBI In 1988/89, the trials were again modified. The pure stands of pulses were dropped and the single stations of vulgaris and soybeans in between planting stations were reintroduced. The pure stands were dropped because they did not yield significantly better than the vulgaris beans or soybeans in the intercrop, yet they demanded more labor and seed. 91W CIMMYT indicates that marginal analysis for a particular experiment should be done on pooled results from at least several locations over one or more years (CIMMYT, 1988). This analysis uses the final year of the experiment results pooled across locations in each recommendation domain. Ideally, it would be more instructive to analyze the pooled results over the period of four years. This was not possible however, because the trial design had undergone several modifications over the years. Some treatments were dropped and some added over the seasons. The modification of the treatments called for the modification of plot sizes. Farmers are not willing to work with trials that take up most of their land. Plot sizes were changed from 4 ridges of 0.91 meters by 7.2 meters net in year one to 4 ridges of 0.91 by 6.3 meters in year two. In year three the net plot size was reduced to 2 ridges of 0.91 meters by 5.4 meters to take care of the increased number of treatments. In year four the net plot 32 size was changed to 4 ridges of 0.91 meters by 5.4 meters. The change in the treatments and plot sizes makes it difficult to do a meaningful across year analysis. The final year retains all the treatments which are likely to be adopted. The treatments that did not perform well were dropped. The variation in yield over time is taken care of by a minimum returns analysis which compares net benefits for each individual location and all seasons (section 3.2). W The experiment was laid out using a complete random block design with three replications at each location for all four seasons. The plot size varied with the number of ,treatments. The ridges were spaced 90 cm apart. Maize was grown 90 cm between plants, three plants per station. The sole crop of dwarf beans was planted on two rows spaced at 30 cm on each ridge or one seed spaced at 10 cm apart along each row. Sole crop of climbing beans was planted on one row, two seeds per station spaced at 15 cm apart. Sole crop of soybeans was planted in two rows per ridge, one seed per station spaced at 5 cm apart. The intercropped Vulgaris beans and soybeans were planted three seeds per station. The treatments, excluding the ones that were dropped are shown in table 1. 33 Table 1 Vulgaris Soybeans beans Immune—niisb—WLWAL 1 -° - - 2 ssd 3700 - 3 1 3600 - 4 2 7100 - 5 3 10700 - 6 4 14300 - 7 1 - 3700 8 2 - 7100 9 3 - 10700 10 4 - 14300 a. The population of maize is constant so it is not included in the table. b. The pulse densities are represented by the numbers of hills in between maize hills. Each hill was planted with three seeds. c. Not included in the treatment. d. Maize and vulgaris beans on the same station. A11.:srtilizsr_annlisatign The recommended levels of fertilizers for maize were applied. M312 hybrid maize was basal-dressed with 200 kg of 20:20:0 (percentage of nitrogen: phosphorus: potassium). This supplied 40 kg of nitrogen and 40 kg of phosphorus. It 34 was top-dressed with 200 kg calcium ammonium nitrate which contains 26% nitrogen. This supplied 52 kg nitrogen. The total amount of nitrogen supplied was 92 kg. Where high analysis fertilizers14 were used 80 kg of diammonium sulphate and 175 kg of urea was applied. Local maize received 150 kg of calcium ammonium nitrate as a top dress. This supplies 40 kg nitrogen. Where high analysis fertilizers were used 20 kg of diammonium sulphate and 40 kg urea were applied. Local maize is known to have a poorer response to fertilizers compared to hybrid maize (MOA,1989a). Fertilizer research in Malawi for intercropping systems has been limited. Consequently there are practically no fertilizer recommendations for intercropping systems. It is important to remember that traditionally these systems used virtually no fertilizers (Mkandawire et al., 1989). Fertilizer use spread with the introduction of improved maize varieties. Even on improved maize varieties, fertilizer is often applied only as a top-dressing at rates considerably lower than recommended (World Bank, 1989 and LADD/ART, 1985). The application of fertilizer levels that are higher than target farmers does not reflect an accurate nonexperimental variable. The fertilizers were applied at a constant rate to all MH12 trials and at a different but constant rate for local 14High analysis fertilizers have higher concentration of chemicals, per unit weight, than standard fertilizers. 35 maize. The data from the experiment can be used to determine the performance of intercropping technologies at a given fertility level. However, the technologies may not perform the same under lower fertility levels. MW There are no significant differences between hybrid maize treatment means in Thiwi Lifidzi (table 2). Table 2 mailman)... W Treatments 1. Sole Maize 4204 - - 2. Maize + V. beans same hill 3457 72 - 3. Maize + 1 V. bean hill 4170 96 - 4. Maize + 2 V. bean hills 4297 99 - 5. Maize + 3 V. bean hills 4163 151 - 6. Maize + 4 V. bean hills 4184 412 - 7. Maize + 1 Soybean hill 4163 - 281 8. Maize + 2 Soybean hills 4348 - 247 9. Maize + 3 Soybean hills 4273 - 342 10. Maize + 4 Soybean hills 4300 - 370 Standard error 241 - 42 83 Coefficient of variation 10 44 47 Significance NSa *b **C a. Not significantly different. b. Significant at 5% level. c. Significant at 1% level. 36 Increasing the density of pulses has no significant detrimental effect on hybrid maize. There are highly significant differences between the means of different densities of soybeans. Increasing the density of soybeans increases the yield of soybeans in the intercrop. There are significant differences among the Vulgaris bean treatment means, increasing the density of Vulgaris beans increases the yield of Vulgaris beans. There are no significant differences between the treatment means of maize (table 3). Table 3 minim W Treatments 1. Sole Maize 1859 - - 2. Maize + V. beans same hill 1430 41 - 3. Maize + 1 V. bean hill 1752 - - 4. Maize + 2 V. bean hills 1567 58 - 5. Maize + 3 V. bean hills 1656 24 - 6. Maize + 4 V. bean hills 1732 41 7. Maize + 1 soybean hill 1536 - 288 8. Maize + 2 soybean hills 1554 - 460 9. Maize + 3 soybean hills 1341 - 487 10. Maize + 4 soybean hills 1622 - 761 Standard error 183 - - Coefficient of variation 20 29 22b Significance NSa NS ** a. Not significantly different. b. Significant at 1% level. 37 Increasing the densities of pulses has no detrimental effect on the local maize in Thiwi Lifidzi. There are no significant differences among the vulgaris bean treatment means, increasing the densities of vulgaris beans does not increase the yield of the beans significantly. The yield of Vulgaris beans are, however, very low (table 2 in the appendix). There are highly significant differences among soybean treatment means, soybean yield increase with increasing densities. There are no significant differences between maize treatments (table 4). Table 4 WWW W Was Treatments 1. Sole Maize 3652 - - 2. Maize + V. bean same hill 2953 235 - 3. Maize + 1 V. bean hill 1 2819 274 - 4. Maize + 2 V. bean hills 3529 473 - 5. Maize + 3 V. bean hills 2900 816 - 6. Maize + 4 V. bean hills 2541 864 - 7. Maize + 1 Soybean hill 3704 - 364 8. Maize + 2 Soybean hills 3093 - 360 9. Maize + 3 Soybean hills 3179 - 566 10. Maize + 4 Soybean hills 3107 - 514 Standard error 336 140 116 Coefficient if variation 18 45 45 Significance NSa *b * a. Not significantly different. b. Significant at 5% level. 38 Increasing the density of pulses has no significant detrimental effect on maize. Vulgaris bean treatment means and soybean treatment means were significantly different from each other, the pulse yield increased with increasing density. There are no significant differences between maize treatment means (table 5). Table 5 W W Treatments 1. Sole Maize 2960 - - 2. Maize + V. beans same hill 2702 62 - 3. Maize + 1 V. bean hill 2363 96 - 4. Maize + 2 V. bean hills 2411 178 - 5. Maize + 3 V. bean hills 2140 199 - 6. Maize + 4 V. bean hills 2222 144 - 7. Maize + 1 Soybean hill 2363 - 148 8. Maize + 2 Soybean hills 2925 - 96 9. Maize + 3 Soybean hills 2253 - 144 10. Maize + 4 Soybean hills 2685 - 219 Standard error 237 31 16 Coefficient of variation 17 40b 22 Significance NSa ** ** a. Not significantly different. b. Significant at 1% level. Increasing the density of pulses has no significant detrimental effect on maize. Both Vulgaris bean and soybean 39 treatment means are highly significantly different from each other: pulse yields increase with increasing densities. In general, the data confirms the hypothesis that increasing the densities of pulses would increase the yield of the pulses and have no significant detrimental effect on the maize yield. All recommendation domains show that, except for Thiwi Lifidzi local maize intercrop show that increasing the density of pulses increase the yield. For Thiwi Lifidzi local maize intercrop recommendation domain, only soybean yield increase with increasing density. The Thiwi Lifidzi local maize intercrop Vulgaris bean yields are, however, lower than the lowest average reported by the Thiwi Lifidzi Project Office (section 2.1.2). This may have been a particularly bad season. CHAPTER 5 FINANCIAL ANALYSIS iil_Di§§ The data used in the analysis were gathered over a period of five years, from 1984, when the LADD/ART started operating at full scale, through 1989 when this study was conducted. Some of the data, especially the background information, dates back further than 1984. Most of these data were gathered through surveys and direct measurements. A rapid appraisal survey (section 2.1.6) was conducted by the author in 1989. W The surveys were conducted from 1984 to 1986 found that most of the farmers that hold less than a hectare of land do not grow enough food to last for the whole year. Their limited access to various resources set limits to the amount they are able to produce for their families. As a consequence they resort to survival strategies to earn their livelihood. Yet most of the adapted strategies entrench these farmers into the cycle of food deficit. Food shortages occur mostly from December to April. In 1985, 65% of the farmers interviewed in Dedza Hills project run out of 40 41 food (LADD/ART, 1985). During this time most farmers have no cash with which to purchase food. Crop diversification through intercropping is a common system employed to try to grow adequate food as well as obtain cash through crop sales. Maize and Vulgaris bean intercrops are the most common. Vulgaris beans can be harvested green and be sold or eaten by February. But Vulgaris beans seem not to be well suited to some RDPs of Lilongwe ADD where they often suffer from diseases. Halo blight and anthracnose are among the most destructive and severe infections can result in heavy reductions of seed and seed quality during the rainy season (Msuku, 1984). Soybeans, which are currently being encouraged, do well in these areas but are mostly grown in pure stands and in small areas. In a few cases, soybeans are intercropped but at low densities. Based on these survey results an experimental program was initiated in Thiwi Lifidzi and later Lilongwe Rural Development Projects of the Lilongwe ADD. Four recommendation domains15 were identified. There were two recommendation domains for each location: one for small smallholders and another for large smallholders (table 6). . 15A recommendation domain is a group of farmers who have similar circumstances and for whom it is likely that the same recommendation will be suitable. 42 Table 6 1 . '0. {’9 01-1 ‘ is" ; 01.0!!100... 0! {out a: Domain Location Cropping Smallholder Smallholder (RDP) 1 T1.“ 124” + Small 95 V. beans or soybeans 2 TL 14312 + Larged 09 V. beans or soybeans 3 LL° LM Small 87 + V. beans or soybeans 4 LL MH12 + Large 13 V. beans or soybeans a. Thiwi Lifidzi RDP b. Local maize c. Lilongwe RDP d. Large smallholders grow MH12 hybrid in addition to local maize, small smallholders grow local maize only. It would be more instructive to delineate recommendation domains based on soil type.’ Lilongwe ADD has abundant information in the form of soil landscape management units. Soil landscape management units (table 7) are areas where specific conditions of elevation, mean annual temperature, mean annual rainfall, landform and soil type occur (Selenje, 1988). 43 Table 7 W802... Lilongwe. Thiwi_Lifidzi Soil unit 1 2 3 4 Soil type Orthic Ferric Ferric Xanthic Ferasols Luvisols Luvisols Ferasols Elevation 1036-1280 1097-1340 1401-1620 1220-1525 Mean annual Rainfall (mm) 769-1016 765-900 900-1200 900-1000 Mean annual temperature(c) 19-21 18-21 17-20 18-20 Landform a b c d a. Gentle to gently undulating plains. b. Gently undulating broad valleys with gentle slopes. c. Dissected moderately undulating plains with broad crests and gentle slopes. d. Dissected gently undulating slopes with high hills and mountains. For each landscape management unit, it is feasible to apply uniform land use and land management practices. Each unit can, therefore, be used as a basis for delineating a recommendation domain. However, the landscape units in each RDP are almost similar and research funds are limited, hence RDPs were used as a basis for recommendation domain. A rapid appraisal survey was conducted by the author in November and December 1989 to update and collect data needed 44 for financial and economic analysis of the trials. Agronomic data explaining how each proposed technology performed were collected from the ADD headquarters. Farmer interviews were conducted to collect data on farmer preferences, markets, knowledge of local money lenders, interest rates charged by local lenders, off-farm job opportunities, land rental, and cost of transportation. Sixteen farmers were interviewed in Lilongwe RDP and 17 were interviewed in Thiwi Lifidzi RDP. Ministry of Agriculture Headquarters provided data on shadow prices of maize, beans, soybeans and fertilizers, opportunity cost of labor and seed, number of private traders operating and number of ADMARC markets operating. WW These are the mean yields for the experimental plot converted from kg per plot to kg per hectare. The plots were small, for example 2 ridges of three-meter spacing in between the ridges and 5.4 meters length. The weaknesses related to converting are discussed in the following section. £111211_Afli§§§§§_¥1§l§1 The adjusted yield is the average yield of a treatment adjusted downwards by a certain percentage to reflect the difference between experimental yield and the yield farmers would expect from the same treatment (CIMMYT, 1988). Experimental yields, even from on-farm experiments under representative conditions, are often higher than the yields 45 that the farmers could expect using the same treatments because of the differences in management, plot size and operation precision. The experiments were farmer-managed but researcher- supervised. The Adaptive Research Team had research assistants located in the experimental vicinity. Their main job was to get cost-route data, but they also reminded the farmer of the expected operations in the experimental plots. Thus the nonexperimental variables were not completely under the farmers' control. The adaptive researchers directly supervised and sometimes got involved in farm operations. The researchers are in general more precise and timely than the farmers in implementing operations such as plant spacing, fertilizer application, weed control and harvesting. Farmer management probably changed to a higher level due to research supervision. Also the mere fact that a farmer was running research experiments on his field was enough to improve the management because of the reward system. It brings high esteem to work with government agents. Yields were estimated from small plots. Yields estimated from small plots often overestimate the yield from an entire field because small plots tend to be more uniform than larger fields (CIMMYT, 1988, and Kearl et al., 1976). There is a tendency to overharvest a small plot by gathering produce from plants bordering the crop. When the data is 46 scaled up to a hectare, any such overharvesting is magnified. A general rule is to reduce experimental yield between 5 and 30% (CIMMYT, 1988). Since this experiment is farmer- managed, but the fertilizer dose does not reflect actual farmer circumstances (see section 4.7), this study will assume an adjustment factor of 15%. £12_H§§hQQQlQ§Y The marginal analysis technique goes through the following steps: 51211.23rtial_budgsts Partial budgets for each recommendation domain are built using data pooled across locations. Partial budgeting is a method of organizing experimental data and information about the costs and benefits of various alternative treatments. The partial budget indicates the net benefit, which is income remaining after subtracting all the costs that vary, from the gross benefit. The partial budget uses currency as the common denominator. This provides an estimate of the cost of investment measured in a uniform manner without implying that farmers are concerned only with money (CIMMYT, 1988). Winn Dominance analysis is performed on the net benefits and associated total variable costs resulting from the partial budget to eliminate inferior treatments, i.e., treatments with net benefits that are less than or equal to a treatment 47 with lower costs (CIMMYT, 1988). In order to improve farmer income, it is important to pay attention to net benefits rather than just yield. Often the value of increase in yield is not enough to compensate for the increase in cost. Wham Marginal rates of return are estimated for nondominated treatments. This analysis reflects the additional profit earned by additional capital and labor invested in the treatments. Labor and capital merit attention because they are the resources that the majority of the farmers have some control over (LADD/ART, 1984). The amount of land cultivated by farmers is controlled by traditional authorities. Marginal rates of return between treatments are calculated, proceeding in step from a lower cost treatment to that of the next higher cost, and comparing those rates of return to the minimum rate of return acceptable to farmers. It is useful both for making recommendations to farmers, where there is sufficient experimental evidence, and for helping select treatments for further experimentation. The minimum acceptable rate of return in this study is 100%, as discussed in section 2.1.6 WWW Returns to labor are estimated for nondominated treatments by dividing the incremental benefits by the incremental labor time. This gives the amount of money labor can earn per hour and separates the effect of 48 additional labor from other variable factors of production. If changing from one technology to another generates more returns to labor than the assumed opportunity cost, the technology has a high chance of being adopted. W The relative ability of each of the nondominated technologies to supply an average family with adequate food supply, assuming all food is produced by the farm family, was determined by how much maize and beans the technology can produce. The technology which produces relatively more maize and beans contributes more towards satisfying the food needs. An average person requires 230 kg of maize and 73 kg of Vulgaris beans per year to supply adequate carbohydrates and protein diet (Chipande, 1987, and Barnes-McConneI,. 1989). In both RDPs, the average household size is 4 adult- equivalents. Adequate food supply for the family requires 920 kg maize and 292 kg Vulgaris beans assuming these are the only sources of carbohydrates and proteins. A technology which produces more than these amounts of maize and beans is most likely to be adopted. On the other hand, a technology which produces less than these amounts but relatively more than other technologies is likely to be adopted too because it contributes towards household food security. We Minimum returns analysis is performed on the nondominated treatments to determine how they would perform 49 in the worst possible situations. This tests for the impact of unpredictable weather, especially erratic rainfall and other unpredictable factors that experimental design cannot account for, on the treatments being considered for recommendation. Malawi suffers from early season dry spells in most of the areas, followed by late season excess moisture (MOA, 1989d). This tends to reduce the yields. W The field price is defined as the value to the farmer of an additional unit of production in the field, prior to harvest. It is calculated by taking the price that the farmer receives or can receive for a crop when he/she sells it, less all the costs associated with harvesting and ' selling that are proportional to yield (CIMMYT, 1988). It is convenient to treat these costs separately from other costs that vary because they are incurred at the time of harvesting. The other costs that vary are invested at the beginning of the season. Farmers have to wait at least five months before recovering their invested inputs in maize and soybeans, and at least three months before they can recover their invested inputs from Vulgaris beans (LADD/ART, 1984). Prices at harvest will be used. Prices at harvest reflect the returns to production while postharvest prices reflect returns to storage (Bernsten, 1989). ADMARC prices for maize and soybean and local market prices for vulgaris beans will be used because these reflect the amounts that farmers actually receive. 50 451211111_N§129_DI12§§ Local market prices rise with time as we move away from the harvest time. The consumer market prices for maize falls moderately after reaching a peak in January. This is due to the normal price movement associated with relative scarce supplies late in the growing cycle. Consumer prices fall below ADMARC selling price at harvest (MOA, 1989c). ADMARC offers the highest prices of maize during the harvesting period. LADD/ART (1986) survey showed that the average price of maize at harvest in the local markets were an average of MK0.09/kg as opposed to MKO.122/kg at ADMARC markets. Price data from agro-economic surveys show a price of MKO.16/kg at the District council market while ADMARC was offering MKO.24/kg (MOA, 1989d). The amount of grain that is sold on the local markets is small compared to the amount that goes through ADMARC at the harvesting time because every farmer has some maize. Despite the market liberalization program, ADMARC still operates in most areas where private traders prefer not to serve due to high operating costs related to transportation. Private traders lack capacity for purchasing large amounts of grain from surplus areas and transporting to deficit regions. ADMARC has the capacity because it is charged with national food security responsibilities and is supported by the government. Data is readily available from ADMARC but there is very little information concerning the activities 51 of private traders such as the exact amounts they purchase and the prices they offer to farmers (MOA, 1989c). mm The price of vulgaris beans is always higher in the local markets than in ADMARC markets. The majority of the smallholder farmers sell on the local markets (section 2.1.6). The agro-economic survey of 1989 indicates an average market price of MK1.89/kg as opposed to the MKO.50/kg ADMARC price (MOA, 1989d). The ADMARC soybean price, which is higher than the local market price, will be used. Most farmers sell their soybeans to ADMARC (section 2.1.6) for processing into animal feed. The rest of the beans are processed into food for the family or snacks to sell. The agro-economic survey of 1989 indicates that no soybeans were sold on the local markets surveyed. In 1990 the ADMARC buying Price for soybean is MKO.60/kg (MOA, 1989c). W The transportation price was estimated assuming the use of a farm cart. The approximate volume of grain the farm cart can carry was converted to weight and the weight was divided by the amount of money paid per trip (section 2.1.6). Farmers hire out their farm carts to other farmers or merchants. The problem is the hiring charge is not dependent on mileage, but vicinity which is a proxy for mileage. The rapid appraisal survey showed that the average total transport cost per kg is MK0.02145 (section 2.1.6.) 52 W The labor time required for the treatment operations was recorded using the stop watch technique. The operators were mostly farmers, and in some cases research laborers. The mean recorded time per treatment was then scaled up to a hectare (section 2.1.6). The precision of this variable may be affected by the fact that the people working were aware that they were being timed and may have worked faster or slower. In most cases they were all adults while the real farm labor force includes children. The labor data is reported in the summary of the rapid appraisal survey in section 2.1.6. W The most common form of employment for the people living in the area under study, apart from farming their own holdings, is working for estates as laborers. The estates offer the minimum wage stipulated by the Government. This comes to MKO.21 per hour. People also work for other smallholder farmers but only during the peak labor seasons. A few people get employed in the city and get a minimum urban wage rate of MKO.27 per hour. This rate will be used in sensitivity analysis. The labor time required per operation and the total cost of labor for the intercropping technologies are shown in table 8. 53 Table 8 .,.L; ,4. ,.. -.L -,;, ; .; -,.u ..; . ., Wu WW). Ridging 70 Planting maize 30 Planting Pulses 33 Basal-dressing 42 Top-dressing 42 Weeding 70 Banking 100 Harvesting maize 38 new 38 Total labor required 463 Total labor cost at rural wage rate 97.23a Total labor cost at urban wage rate 125.01 Sources: Blackie M. J. (1989) and LADD/ART (1986). a. The rural wage rate is MKO.21 per hour. b. The urban wage rate is MKO.27 per hour £11_B§§Bl§§ The financial analysis was performed using average figures for maize in each recommendation domain and average figures for Vulgaris beans in Thiwi Lifidzi (tables 1 to 5 in the appendix). The mean yields for these crops were not 54 significantly different from each other, implying that whatever differences observed were due to chance 9, ‘ ‘._ . ». ._.. ; , -_ ... ,. ; .._,.,,-,._, '. mm For Thiwi Lifidzi local maize intercropping recommendation domain, soybean intercropping treatments dominated the Vulgaris beans intercopping treatments. The farmer practice--maize and beans planted on the same station--is dominated (table 9). Marginal rates of return and marginal returns to labor show that moving to treatment 7--one soybean hill between maize hills--from treatment 1 is worthwhile. The marginal rate of return is higher than the acceptable rate of 100% (section 2.1.6) and the returns to labor are higher than the opportunity cost of labor (section 5.2.4). The farmer would exploit more opportunities for further earning, however, if he/she opted for treatment 8 because this treatment has marginal rate of return that is attractive and is higher than the minimum acceptable rate of return of 100%. Moving from treatment 8 to treatment 9, the marginal rate of return falls below the minimum acceptable rate of return. Treatment 9 is, hence, eliminated from consideration. Marginal rate of return between treatment 8 and treatment 10 is 699%. This is well above the minimum acceptable rate of return. Financially, treatment 10 is the best treatment (CIMMYT, 1988). Sensitivity analysis (table 10) shows similar results. Treatment 'r.v.ca Net henb MRRc MRLd Foode (MK/ha) (MK/ha) (%) (MK/hr) (kg/ha) 1. Sole crop f maize 0.00 300.14 - - 1364 - 7. One soybean hill 11.84 425.38 1058 4.50 1364 - 8. TWO soybean _ hills 18.96 500.14 1050 14.03 1364 - 9. Three soybean hills 26.43 505.52 72 0.77 1364 - 10. Four Soybean hills 34.21 606.74 13109 11.95 1364 - a. TVC is total variable costs calculated by adding up all costs that vary (table 2 in the appendix). b. Net benefit is calculated by subtracting TVC from the gross benefits (table 2 in the appendix). c. MRR is calculated as discussed in section 5.2.3 d. MRL is calculated as discussed in section 5.2.4 e. Food supply is calculated as discussed in section 5.2.5 f. Not relevant g. MRR moving from treatment 8 to treatment 10 is (106.60/15.25) 8 699%. 56 Table 10 9.12 a; .,. ;'- '0 ' o "‘ :9 :0 ‘ 9!! I ‘ 09° 1- ': u‘tt‘ 0’ < H! : -,°e i‘ 0‘:!S e 1 f0 9‘ I ! ! i . Iii i I'Eil i BEE l ! Treatments 'r.v.ca Net Benb MRRC MRLS Foode (Mk/ha) (Mk/118) (’3) (Mk/hr) M(Ks/hat) 1. Sole crop f maize 0.00 286.49 - - 1364 - 7. One soybean hill 13.51 410.07 915 4.44 1364 - 8. Two soybean hills 20.95 484.50 1000 13.96 1364 - 9. Three soybean hills 28.84 489.47 63 0.71 1364 - 10. Four soybean hills 37.12 603.83 13819 14.42 1364 - a. TVC is total variable costs calculated by adding up all costs that vary (table 2 in the appendix). b. Net benefit is calculated by subtracting TVC from the gross benefits (table 2 in the appendix). c. MRR is calculated as discussed in section 5.2.3 d. MRL is calculated as discussed in section 5.2.4 e. Food supply is calculated as discussed in section 5.2.5 f. Not relevant g. MRR moving from treatment 8 to treatment 10 is (119.33/16.17) 8 738%. Treatments 7, 8 and 10 have minimum net benefits that are higher than the farmer's practice (table 11), hence they have a high chance of being adopted. 57 Table 11 Treatments Seassn___fiits 1* 13 8 10 1988/89 1 345.92 412.48 490.60 609.92 1988/89 2 273.77 264.07 359.22 343.48 1987/88 1 450.67 - 589.30 696.78 1986/87 1 354.43 496.83 490.53 - 1986/87 2 469.74 499.60 591.77 561.78 1986/87 1 564.74 - 501.41 647.83 1985/86 1 396.44 445.46 548.67 - Avg 25% lowest net bena 309.85 338.28 424.88 452.63 Avg 25% farmers'b practice_ 326.07 a. The average of the two lowest net benefits in each treatment (section 3.2). b. The average of the two lowest net benefits in treatment two--the farmer practice. The rapid appraisal survey conducted in 1989 revealed that 75% of the farmers in this domain preferred the treatment with two soybeans in between maize hills. This suggests treatment 8, two soybean hills in betweenn maize hills, however, farmers would exploit opportunities for further earning if they adopted treatment 10. u, A ‘ . ,u, ' , - . . . , . - . up“ - , . , . , e . u 3 In this domain, the farmer practice was dominated (table 12). Moving from treatment 1 to treatment 7 is worthwhile. Marginal rates of return fall below acceptable 58 minimum rate of return if we move from treatment 7 to treatment 9, treatment 9 to treatment 10, and treatment 10 to treatment 5. Treatments 9, 10 and 5 are eliminated from consideration. MRR between treatment 7 and 6 is 425. This is well above the minimum acceptable rate of return. In the rapid appraisal survey conducted by the author in 1989, most farmers in this domain preferred the treatments with two soybean and/or two vulgaris beans in between maize hills because of high yields and lower seed costs. However, these treatment is dominated. For low resource farmers--more than 50% of the farmers are categorized as low resource--who cannot afford to invest MK125.41, treatment 7 would be recommended. Adopting treatment 7, h0wever, entails forgoing opportunity for further earning. For large farmers who can afford to invest more than the low resource farmers, treatment 6 would be recommended. Results of sensitivity analysis (table 13) show a similar trend. Minimum returns analysis done on nine locations over a period of four seasons indicates that both treatments 7 and 6 have minimum net benefits that are higher than the farmer's practice (table 14). 59 Table 12 7* e ’. 1.. ; : ‘0 0 e i‘; :g :e ; 2,0 g ; 0007’. u l 0 4!: 1.9-. 1 L °-. . 0'21: 24° —° 0 1‘ ! ! I i Iii I I°E°l . BEE Treatments 'r.v.ca Net Benb MRRC MRLd Foode (Mk/ha) (Mk/ha) (’1‘) (MK/hr) (ks/ha) 1. Sole crop maize 0.00 776.99 - - 3532 - 7. One soybean hill 18.84 898.90 1030 4.38 3532 - 9. Three soybean hills 26.43 913.35 99 g 3532 - 10. Four soybean _ hills 34.21 918.90 71 g 3532 - 5. Three V. bean hills 94.83 923.45 08 0.30 3532 128 6. Four V. bean hills 125.41 1309.95 1264 46.27 3532 350 a. TVC is total variable costs calculated by adding up all costs that vary (table 3 in the appendix). b. Net benefit is calculated by subtracting TVC from the gross benefits (table 3 in the appendix). c. MRR is calculated as discussed in section 5.2.3 d. MRL is calculated as discussed in section 5.2.4 e. Food supply is calculated as discussed in section 5.2.5 f. Not relevant g. Treatment is dominated because it requires more labor but has lower net benefit than the preceding treatment. h. MRR moving from treatment 7 to treatment 6 is (532.96/125.41) - 425%. Treatments 60 Table 13 'r.v.ca Net Ben (Mk/ha) (Mk/ha) b MRRc (’3) MRLc (Mk/hr) 10. Sole crop maize 0.00 One soybean hill 13.51 Three soybean hills 28.84 Four soybean hills 37.12 Three V. bean hills 97.24 Four V. bean hills 128.32 741.67 861.91 875.62 880.66 884.44 1268.22 890 89 61 06 1235 h 45.31 Foode (Kg/ha) 3532 - 3532 - 3532 - 3532 - 3532 - 3532 350 b. c. d. e. f. g. TVC is total variable costs calculated by adding up all costs that vary (table 4 in the appendix). Net benefit is calculated by subtracting TVC from the gross benefits (table 4 in the appendix). MRR is calculated as discussed in section 5.2.3 MRL is calculated as discussed in section 5.2.4 Food supply is calculated as discussed in section 5.2.5 Not relevant Treatment is dominated because it requires more labor but has lower net benefit than the preceding treatment. MRR moving from treatment 7 to treatment 6 is (526.55/128.32) I 410%. Treatments M 1 1 6 88/89 1 786.15 900.40 1315.38 88/89 2 699.57 731.26 709.78 87/88 1 731.17 - 1058.30 87/88 2 1544.62 - 1801.88 86/87 1 769.32 794.15 - 86/87 3 1254.96 1158.30 - 85/86 1 693.77 697.44 - 85/86 2 877.03 977.39 - Avg 25:a lowest net ben 696.67 714.35 709.78 Avg 25%b farmers' practice 551.56 a. The average of the two lowest net benefits in each treatment (section 3.2) and the lowest net benefit in treatment 6. b. The average of the two lowest net benefits in treatment two--the farmer practice. In this domain the farmer practice is among the dominant treatments (table 15). return fall below the minimum acceptable rates if we move The marginal rates of 62 from treatment 7 to treatment 2. Moving from treatment 2 to treatment 10 also results in a marginal rate of return which falls below the acceptable minimum rate of return. Marginal rate of return between treatment 7 and treatment 4 is well above the minimum acceptable rate of return. Moving from treatment 4 to treatment 5, the marginal rate of return again falls below minimum acceptable rate of return. This suggests treatment 4--two vulgaris bean hills in between maize hills. Sensitivity analysis (table 16) shows similar results. Minimum returns analysis (table 17) shows that treatment 7 has higher net benefits and treatment 4 has slightly higher net benefit than the farmers' practice. Fifty percent of the farmers in this domain indicated that-they preferred two hills of vulgaris beans and/or two hills of soybean in between maize hills. This coincides with the financial analysis. The best technology in this recommendation domain is treatment 4--two Vulgaris bean hills in between maize hills. 63 Table 15 O _ .__ I 921: 9; -_!-_ :-‘ 0 '0, z'! :°;-_I°_ 9 z 0.0.- :0 Treatments T.V.Ca Net Benb MRRc MRL? Foode (Mk/ha) (Mk/ha) (%) (Mk/hr) (Kg/ha) 1. Sole crop f maize 0.00 427.86 - - 1938 - 7. One soybean hill 11.84 486.46 494 g 1938 - 2. Maize/V. beans same hill 33.66 493.33 32 2.86 1938 - 10. Four soybean hills 34.21 497.89 51 g 1938 - 3. One V.bean h hills 34.64 542.62 10402 10.77 1938 82 4. Two V.bean hills 64.65 647.74 505 18.79 1938 151 5. Three V.bean hills 94.83 651.03 08 0.47 1938 169 a. TVC is total variable costs calculated by adding up all costs that vary (table 4 in the appendix). b. Net benefit is calculated by subtracting TVC from the gross benefits (table 4 in the appendix). c. MRR is calculated as discussed in section 5.2.3 d. MRL is calculated as discussed in section 5.2.4 8. Food supply is calculated as discussed in section 5.2.5 f. Not relevant 9. Treatment is dominated because it requires more labor but has lower net benefit than the preceding treatment. h. MRR moving from treatment 7 to treatment 3 is (56.16/22.8) - 246%. 64 Table 16 Treatments T.V.Ca Net Benb MRRc MRLg Foode (Mk/ha) (Mk/ha) (%) (Mk/hr) (Kg/ha) Maize____neans 1. Sole crop f maize 0.00 406.98 - - 1938 - 7. One soybean hill 13.51 463.92 421 g 1938 - 2. Maize/V. beans same hill 34.97 471.98 31 2.78 1938 - 3. One V.bean h hills 36.31 532.26 4499 10.65 1938 82 4. Two V.bean hills 66.55 646.75 128 18.78 1938 151 5. Three V.bean hills 97.24 626.05 09 0.38 1938 169 a. TVC is total variable costs calculated by adding up all costs that vary (table 4 in the appendix). b. Net benefit is calculated by subtracting TVC from the gross benefits (table 4 in the appendix). c. MRR is calculated as discussed in section 5.2.3 d. MRL is calculated as discussed in section 5.2.4 e. Food supply is calculated as discussed in section 5.2.5 f. Not relevant g. Treatment is dominated because it requires more labor but has lower net benefit than the preceding treatment. h. MRR moving from treatment 7 to treatment 3 is (68.34/22.80) - 300:. 65 Table 17 MW 11"! i!l.’l Treatments Season Site 1 7 4 5 1988/89 1 553.32 605.58 670.74 332.64 1988/89 2 400.52 414.12 348.97 583.59 1987/88 1 723.69 - 1867.93 1555.87 Avg 33% lowest net hena 400.52 414.12 348.97 332.64 avg 33% farmers b practice 347.07 a. The lowest of the three net benefits. b. The lowest of the three net benefits in treatment ‘ two--the farmer's practice. .,u‘- ;.- a. , - ... ,. - .mu-,.t ., ..ut , In this domain the farmers' practice is among the dominant treatments (table 18). However, moving from the farmers' practice to treatments 3, 4, 5 and 6--one, two, three and four Vulgaris bean hills in between maize hills, respectively--will offer opportunities for further earning. The marginal rates of return for all these treatments are above the minimum acceptable rate of 100%. Sensitivity analysis (table 19) showed similar results. 66 Table 18 Treatments T . v . 0“ Net Benb MRRC mud Foode (Mk/ha) (Mk/ha) (*) (Mk/hr) (Kg/ha) 1. Sole crop maize 0.00 588.62 - - 2675 - 7. One soybean hill 11.84 750.10 1364 g 2675 - 9. Three soybean hills 26.43 831.66 559 g 2675 - 2. Maize/V.beans same hill 33.60 935.40 1447 15.20 2675 202 3. One V.bean hill 34.64 991.82 5432 11.30 2675 233 4. Two V.bean hills 65.64 1278.97 960 54.05 2675 402 5. Three V.bean hills 94.84 1779.97 1657 71.47 2675 693 6. Four V.bean hills 125.41 1843.94 209 8 2675 743 a. TVC is total variable costs calculated by adding up all costs that vary (table 5 in the appendix). b. Net benefit is calculated by subtracting TVC from the gross benefits (table 5 in the appendix). c. MRR is calculated as discussed in section 5.2.3 d. MRL is calculated as discussed in section 5.2.4 e. Food supply is calculated as discussed in section 5.2.5 f. Not relevant 9. Treatment is dominated because it requires more labor but has lower net benefit than the preceding treatment. 67 Table 19 O ‘ ‘ O U 0 ‘ ‘ ‘0 ‘ O ’ .0. O y. o o o - - - -' ‘ - p! ! — Treatments T.V.Ca Net Benb MRRc MRL Foode (Mk/ha) (Mk/ha) (’15) (Mk/hr) (Ks/ha) 1. Sole crop maize 0.00 561.92 - - 2675 - 7. One soybean hill 13.51 721.67 1183 g 2675 - 9. Three soybean hills 25.24 802.50 527 g 2675 - 2. Maize/V.beans same hill 34.79 905.25 1676 15.01 2675 202 3. One V.bean hill 36.31 961.13 4817 11.29 2675 233 4. Two V.bean hills 66.55 1247.70 940 53.67 2675 403 5. Three V.bean hills 97.24 1761.71 1676 73.50 2675 694 6. Four V.bean hills 128.32 1806.92 145 5.34 2675 734 a. TVC is total variable costs calculated by adding up all costs that vary (table 5 in the appendix). b. Net benefit is calculated by subtracting TVC from the ' gross benefits (table 5 in the appendix). c. MRR is calculated as discussed in section 5.2.3 d. MRL is calculated as discussed in section 5.2.4 e. Food supply is calculated as discussed in section 5.2.5 f. Not relevant g. Treatment is dominated because it requires more labor but has lower net benefit than the preceding treatment. 68 Table 20 MW WWW Seasons 1987/88 1988/89 Treatments Sitel Site2 Site3 Site4 lowesta nsLben. 1 460.02 985.49 682.92 400.18 400.18 7 - - 854.07 432.28 432.28 9 527.97 1286.17 777.39 551.43 551.43 2 895.62 1779.62 899.93 1233.19 895.62 3 - - 930.30 937.33 930.30 4 929.38 1705.82 1315.22 1267.53 929.38 5 844.08 1632.17 1751.44 1095.31 844.08 6 1672.55 1929.40 1157.07 1511.01 1511.01 a. Ideally, this analysis takes an average 25% of the lowest net benefits. In this case the lowest figures are used because the data is only for a short period. The farmers' lowest net benefit-- the lowest net benefit from treatment 2--is MK895.62. A minimum returns analysis over two seasons indicates that only treatments 3, 4 and 6 have net benefits that are higher than the farmer practice (table 20). In the rapid appraisal survey conducted in 1989, farmers in this domain indicated they preferred either two soybean hills in between maize hills or two Vulgaris bean hills in between maize 69 hills (section 2.1.6). Since treatment 8 is dominated, treatment 4, which represents two vulgaris bean hill in between maize hills, would therefore be recommended. 51215.2999 On a per hectare basis, all the nondominated treatments produce more quantities of maize than is required to supply adequate carbohydrates to an average family of four adult equivalents per annum (section 5.2.5). Only the Vulgaris beans intercropped with MH12 hybrid are able to supply adequate protein to an average family. Two, three and four Vulgaris bean hills in between MH12 hybrid maize hills are able to satisfy adequate protein in Lilongwe RDP and only four Vulgaris bean hills in between MH12 hybrid maize hills can supply adequate protein in Thiwi Lifidzi (section 5.2.5). However, 50% of the farming families in Thiwi Lifidzi RDP and 40% of the farming families in Lilongwe RDP hold less than one hectare. The smallest holding size category is 0.5 hectares and less. Seventeen percent of the farmers in Thiwi Lifidzi and 15% of the farmers in Lilongwe RDP fall under this category (LADD/ART, 1989). MH12 hybrid maize intercrops in both ecological zones are able to provide adequate carbohydrates to farm families with hectarages as low as 0.5. In Lilongwe RDP, local maize intercrops are able provide adequate carbohydrates to farm families with hectarages as low as 0.5. Thiwi Lifidzi RDP 70 local maize intercrops are, however, not able to provide adequate carbohydrates to an average family. Only Lilongwe RDP MH12 intercropping recommendation domain can provide adequate protein for an average family with 0.5 hectares if four vulgaris bean hills in between MH12 hybrid maize hills is adopted. However, the financial analysis shows that two Vulgaris bean hills in between maize hills is the best technology in recommendation domain. CHAPTER 6 ECONOMIC ANALYSIS a . . .5 ,. ,,. ...;z -:. ; .; ; "Comparative advantage depends on technology, which determines productibn possibilities and influences rates of product transformation, on resource endowment, which determines the value of domestic resources, and on international prices, which determine the value of all inputs and outputs" (Morris 1989a). Domestic resource costs analysis can generate several measures of the relative economic efficiency of production alternatives. One of these measures, net social profitability, indicates the contribution of each production alternative to national income, measured in terms of net social returns to land. A second measure, the resource cost ratio (R.C.R), indicates the efficiency of each production alternative in using domestic resources to earn or save foreign exchange. To determine domestic resource costs (DRC), all inputs must be classified into primary factors or tradable inputs. Primary factors are defined as goods that are not normally 71 72 traded internationally, including land, labor and capital. Tradables are traded internationally or potentially could be traded internationally. Shadow or social prices are used to reflect the true economic value of goods and services in the absence of taxes, subsidies, import tariffs, quotas, price controls and other government policies. Tradables are valued at their world price equivalent adjusted for transport costs and exchange rate anomalies (Morris, 1989a). Primary factors are valued at their returns in the most profitable alternative use expressed in world price equivalents. Social prices can differ substantially from market prices, as when farmers pay less than the full import cost of fertilizer because of government subsidy, or when they receive less than the full value of their output because the official producer price is set below the world price equivalent. When significant discrepancies exist between market and social prices, the interest of the farmers and the nation can diverge. A crop can be profitable to farmers because of high producer prices or subsidies on inputs even though its production does not represent an efficient use of resources from the point of view of the nation's resources. Conversely, a crop can be unprofitable to farmers because of low producer prices or taxes on inputs, even though its production represents an efficient use of the nation's resources. Comparing private profitability with social profitability thus provides important insights into 73 the impacts of government policies on producer incentives (Morris 1989a). One critical aspect of the calculation of comparative advantage is the valuation of inputs and outputs. It is necessary to adjust market prices to eliminate the effects of policy-induced distortions or market failure. This adjustment is accomplished by assigning all inputs and outputs shadow prices reflecting their true value in the economy. W The social profitability of local maize and MH12 hybrid intercrops was calculated for Thiwi Lifidzi and Lilongwe RDPs. The financial analysis indicated that soybean intercropping technologies were the best in Thiwi Lifidzi RDP recommendation domains and Vulgaris bean intercropping were the best in Lilongwe RDP recommendation domains. Two Vulgaris bean hills in between maize hills was the best bet for Thiwi Lifidzi Local maize intercropping recommendation domain. One soybean hill in between maize hills was the best bet technology for Thiwi Lifidzi MH12 intercropping recommendation domain. Two vulgaris bean hills in between maize hills was the best bet for Lilongwe RDP local maize intercropping recommendation domain and Lilongwe RDP, MH12 intercropping recommendation domain. The comparative advantage analysis is based on these technologies. The assumptions underlying these calculations are: 74 1. Gross social returns are calculated by multiplying the world reference price of a commodity by the quantities of the commodity produced. The world reference prices per kg are MK1.01 for soybeans, MKO.91 for Vulgaris beans and MKO.94 for both local and hybrid maize (tables 21, 22 and 23). Table 21 9n:farm_im99rt_2arit¥_nrise_9f_maizei_112§21 Exchange rateb Official Market (Mk/t) (Mk/t) World Price White Maize, _ 0.1.? Rotterdamc 249.96 372.41 ADD: Freight Rotterdam to Durban 148.20 220.80 Port Charges d 35.00 35.00 Road/Rail Durban to Limbe 156.00 232.43 ADMARC Costs 53.10 79.10 On-farm import parity 642.26 939.74 Source: MOA (1989b) and IMF (1989). a. Malawi generally imports maize, sometimes exporting small quantities. b. The official exchange rate is MK2.47 per US$1.00, the market rate is MK3.68 per US$1.00. c. The starting point in the calculation of the parity prices is Rotterdam because there were no data available from neighboring countries. does not reflect the actual trade pattern. This, however, d. Half the transport cost is nontraded and half is traded. This also applies to ADMARC costs. 75 Table 22 Qn:farm_imn9rt_nriss_ef_sexheansi_112821 Exchange rateb Official Market (Mk/t1 (Mk/t1 World price soybeans 0.1.? Rotterdamc 299.36 446.02 ADD: Freight Rotterdam to Durban 148.20 220.80 Port charges 35.00 35.00 Rail or road Durband to Limbe 156.00 232.43 ADD: ADMARC costs 153.10 79.10 On-farm import parity 791.66 1013.35 Source: MOA (1989c) and IMF (1989). a. b. C. Malawi imports soybean products. The official exchange rate is MK2.47 per US$1.00, the market rate is MK3.68 per US$1.00. The starting point in the calculation of the parity prices is Rotterdam because there were no data available from neighboring countries. This, however, does not reflect the actual trade pattern. Half the transport cost is nontraded and half is traded. This also applies to ADMARC costs. 76 Table 23 WW Exchange rateb Official Market (Mk/ti. (Mk/t) ADMARC export price FOB Limbe 763.75 1137.89 ADMARC buying, selling and administrative costs‘3 148.74 221.60 On farm export parity price 651.01 912.29 Source: MOA (1989c) 'a. Malawi exports Vulgaris beans. b. The official exchange rate is MK2.47 per US$1.00, the market rate is MK3.68 per US$1.00. c. ADMARC handling includes both traded and non-traded components: 50% is assumed traded. 2. Cost of seed is calculated by multiplying the shadow price of soybean seed, maize seed and Vulgaris bean seed by the quantities of seed required. The shadow prices of seed and produce are assumed to be the same (tables 21, 22 and 23). The quantities are shown in tables 1 to 5 in the appendix. 3. Fertilizer cost is calculated by multiplying the import parity price of diammonium sulphate and urea with the various levels of fertilizer required per technology. Using the market exchange rate, the import parity price of 77 diammonium sulphate is MK1.84 per kg and urea is MK1.69 per kg (tables 24 and 25). Table 24 Qn:farm_imn9rt_naritx_nzi2e_9f_nA£_fertilizer_112§21 Exchange ratea Official Market (Mk/t) (Mk/t) World Price Durbanb 1000.00 1489.00 Add: Port Charges 35.00 35.00 Rail/ Road Durban to Lilongwe 156.00 232.43 ADMARC costsc 53.10 79.10 On farm import parity price 1244.10 1835.53 Source: Lele (1988) and MOA (1989c). a. The official exchange rate is MK2.47 per US$1.00, the market rate is MK3.68 per US$1.00. b. Malawi imports all its chemical fertilizers through the port of Durban. DAP is diammonium sulphate. c. ADMARC handling includes both traded and non-traded components: 50% is assumed traded. Local maize requires 20 kg diammonium sulphate and 40 kg urea. MH12 requires 80 kg diammonium sulphate and 175 kg urea . 78 Table 25 91""‘21'1 1101-1 .. o ; -' ;. 'Le ; Exchange ratea Official Market (M34317 (MK/t1 World Price Durban 900.00 1340.89 Add: Port Charges 35.00 35.00 Rail/Road Durbanb to Lilongwe 156.00 232.43 ADMARC costsc 53.10 79.10 On-farm import parity price 1144.10 1687.42 Source: MOA (1989c) and Lele (1988). a. The official exchange rate is MK2.47 per US$1.00, the market rate is MK3.68 per US$1.00. b. Malawi imports all its chemical fertilizers . through the port of Durban. DAP is diammonium sulphate. c. ADMARC handling includes both traded and non-traded components: 50% is assumed traded. The calculation of shadow prices included Government controlled foreign exchange rate of MK2.47 per US$1.00 and the market exchange rate of MK3.68 per US $1.00. The shadow price of MK3.68 is used in the analysis. Slow growth in Malawian agricultural exports along with depressed world prices have reduced the nation's export earnings and precipitated a foreign exchange shortage. The government has responded to this by instituting a set of foreign exchange controls, including a system of rationing foreign 79 exchange to essential industries. One effect of this policy is allow the government to maintain an overvalued currency (Morris, 1989a). The official exchange rate does not fully reflect the real value of a unit of foreign currency to the Malawian economy. Using the shadow exchange rate takes care of overvaluation of the domestic currency. The overvaluation of domestic currency is important in comparative advantage analysis because it affects the market price of tradables. Imported goods become cheaper in domestic currency because they can be purchased with fewer units of overvalued domestic currency. Exports become more expensive for foreign buyers because more units of undervalued foreign currency are required to pay for them. Consequently, if adjustments are not made to correct for overvaluation, efficiency analysis will be biased in favor of import intensive activities (Morris, 1989a). 4. Transport is an indirectly traded input comprising both a tradable component and a primary factor component. Such composite goods are decomposed into the tradable component and the primary component (Morris 1989a). Trucks, trains, fuel and spare parts are the tradable component, and labor is a primary input. The analysis assumes that half the transport cost is tradable. The within country transport cost is calculated assuming that the furthest market is 8 km away from the furthest village. The average cost per kg for transporting from the farm to home and from home to market is MK0.01245 (section 2.1.6). 80 5. Labor costs are calculated in table 8. 6. Interest on capital is based on the local lending rate of 100% (section 2.1.6). "Capital” excludes labor costs because the labor is supplied by the farmers themselves. 7. In the sensitivity analysis (section 6.7) average low yields and average high yields in each recommendation domain are used. The average low and average high yields used in the sensitivity analysis were obtained from project reports for the respective rural development project. These figures are estimated by the respective project offices. Even the figures for average high yields are lower than the figures generated by the research program under study. This may be due to differences in fertility and managerial levels: the farmers apply less fertilizer and are poorer managers than the researchers. The average high yield for local maize is 1300 kg per hectare in Thiwi Lifidzi and 1200 kg per hectare in Lilongwe RDP. The average low local maize yield is 880 kg per hectare in Thiwi Lifidzi and 1000 kg per hectare in Lilongwe RDP. The average low soybean yield is 400 kg per hectare in Thiwi Lifidzi. The average low yield of vulgaris beans 125 kg per hectare in local maize intercrops and 300 in hybrid intercrops. The average high yield of MH12 in both Thiwi Lifidzi and Lilongwe RDP is 3000 kg per hectare and the low yield is 2000 kg per hectare. The average high yield of Vulgaris beans is 500 kg per hectare. 81 W The social profitability--returns to management and land--is higher in MH12 intercropping than the local maize intercropping technologies in both Thiwi Lifidzi and Lilongwe RDPs (table 26). One of the shadow costs that stands out in this analysis is the fertilizer cost of MH12 production, which is 4.24 times higher than for local maize intercrops. Despite the high shadow cost of fertilizer, MH12 intercrops dominate the local maize intercrops due to the high yields of hybrid maize. Since social prices reflect the true economic scarcity value of inputs and outputs, MH12 intercrops represent the most profitable production alternative in terms of contributing to national income. Because social prices for primary factors are set equal to their alternative use values, the higher net social profitability of MH12 intercrops indicates that MH12 intercrops are relatively more efficient than local maize intercrops (Morris, 1989a). 82 Table 26 0 -_ 9 0‘ -..° .0 ° ouu‘oqe 0! owl-.1 4|. 9: Thiwi Lifidzi Lilongwe L9Sal_maizS___HHl2___Lesal_maize___nnlz__ Gross social returns 1687.30 3561.08 1939.49 2881.53 Variable costs: Seed” 67.86 59.78 79.00 79.00 Fertilizer 104.10 442.95 104.10 442.95 Transport 21.85 46.93 25.74 38.32 Labor 97.23 97.23 V 97.23 97.23 Interest 193.81 549.66 208.84 560.27 Total variable costs 484.85 1196.55 514.91 1217.77 Total costs 484.85 1196.55 514.91 1217.77 Social profitability 1202.45 2364.53 1424.58 1663.76 a. All the terms used in this table are discussed in section 6. 2. b. The seed cost is higher for Thiwi Lifidzi local maize intercrop and Lilongwe RDP MH12 intercrop because the best technologies for these recommendation domains have two soybean and two Vulgaris bean hills in between maize hills. The best technologies for the other recommendation domains are one soybean hill between maize hills (section 6.2). 83 Farmer net margins are significantly lower than the net social profit for both local maize and MH12 technologies in both RDPs (table 27). Thiwi Lifidzi Lilongwe Local maize MH12 Local maize MH12 Net benefita 500.14 898.90 647.74 991.82 Less: labor” 97.23 97.23 97.23 97.23 Bertc 46.29 125.99 46.29 125.29 356.71 605.77 504.31 698.69 a. Best bet technology net benefits (tables 9, 12 and 15) b. Total labor cost at rural wage rate (table 8) c. Total cost of fertilizer using ADMARC prices. Diammonium phosphate costs MKO.83/kg and urea costs MKO.74/kg. The required quantities are reported in 4.7. Local maize net margins are very low in Thiwi Lifidzi because of low yields of maize and very low yields of Vulgaris beans. The difference between farm net margins and social profitability can be attributed to agricultural policies that tax a large portion of social profits away from the farmers. 84 To date, ADMARC buys most of the farmer produce despite market liberalization (section 5.2.7.1). ADMARC is charged with making a financial contribution to the Governments' development strategies. The contribution comes mostly from maintaining a difference between produce prices and international prices (Christiansen and Southworth, 1988), thus taxing the farmers. Such taxes can provide disincentive to smallholder farmers. However, the calculation of the shadow prices used in the net social profitability assumes that the starting point is Rotterdam. This does not reflect the true trade pattern. Malawi trades with neighboring countries. If the starting point was a neighboring country, the transport costs would most likely be less. This would reduce the net social profitability and the difference between farm net margins and social profitability. W The data used in the calculation of social profitability was rearranged to determine the resource cost ratios of the various technologies. The following formula was used. 2 WP FD RCR= E PoTo - E PiTi 85 Where: RCR - Resource cost ratio for a specific technology. Wpr a Shadow cost of land, labor, 0.5* transport and the interest on working capital. PoTo - Shadow value of maize and vulgaris beans in Lilongwe RDP intercropping recommendation domains and shadow value of maize and soybeans in Thiwi Lifidzi recommendation domains (section 6.2). PiTi - Shadow cost of seed and fertilizer per technology and 0.5* transport costs. RCR for Thiwi Lifidzi RDP Local maize intercrops = (123.5a+97.23b+10.98c+232.43d+232.43e+125.46f)/(1687.309- h-104.101-232.43j)= 0.640759 67.86 a. Opportunity cost of land (section 6.6) b. Opportunity cost of labor (section 6.2) c. Half of within the country transport (section 6.2) d. Half maize transport Durban to Lilongwe (section 6.2) e. Half soybean transport Durban to Lilongwe (section 6.2) f. Interest on working capital--12.5% per annum on items a to h except for b and g. 9. Shadow value of maize and soybeans (section 6.2) h. Shadow cost of seed (section 6.2) i. Shadow cost of fertilizer (section 6.2) 3. Half fertilizer transport (section 6.2) Resource cost ratios calculated for each recommendation domain using similar equations (table 28) showed that all the technologies have comparative advantage. 86 Table 28 hut“: -;. ‘ 0; . 0: ‘0 :a o 1 'l!_'.!’!°.'. 0! 9019-. 112521 Thiwi Lifidzi Lilongwe * Term L9sal_maize______Hnl2____Lesal_uaizs______unlz a 123.50 123.50 123.50 123.50 b 97.23 97.23 97.23 97.23 c 10.98 23.47 12.89 19.16 d 232.43 232.43 232.43 232.43 e 232.43 232.43 232.43 232.43 f 125.46 168.37 124.70 170.24 Sub-total(1) 822.03 877.46 823.18 874.99 9 1687.30 3561.08 1939.49 2881.53 n - 67.86 - 59.78 - 79.00 - 79.00 1 -104.10 -442.95 -104.10 -442.95 j -232.43 -232.43 -232.43 -232.43 Sub-total(2) 1282.91 2825.92 1523.96 2127.15 030(1/2) 0.640759 0.310494 0.540158 0.411344 * The terms are discussed in the Thiwi Lifidzi resource cost ratio equation page. .é151_S9mS_im2li2ati9nS.91.th2.819131_analxsi§ All the technologies have comparative advantage on the previous indicating that the value of domestic resources used in the production is less than the value of foreign exchange earned or saved. Malawi should encourage the production of these intercrops. It should be borne in mind, however, that the trial data used in the analysis received high dosage of 87 fertilizers. The results may be different at lower dosages. §1§_fisnsilixifx_analxsie Technical coefficients used in the construction of enterprise budgets are often calculated from a range of observed values and the prices used in calculating social profitability are often estimated or projected prices. Sensitivity analysis can reveal whether comparative advantage rankings calculated using mean values, estimated values, or projected values for technical coefficients and social prices are likely to change if technical coefficients or social prices eventually differ from expectations (Morris, 1989a). Social profitability and resource cost ratios frequently are sensitive to the following parameters: 1. World reference prices of outputs - A change in the world reference price of an output will have a greater effect on the social profitability and resource cost ratio of an enterprise than a change of similar magnitude in any other parameter. This analysis assumes an arbitrary 10% increase and 10% decrease in the world reference prices. 2. Yields - Comparative advantage tends to be highly sensitive to the level of yield assumed for any fixed quantity of inputs. Improved management can succeed in raising mean yields above current levels and enterprises that appear socially unprofitable at present could become profitable. Average yields experienced over several seasons 88 as reported by the project office, both low averages and high averages, are used in the sensitivity analysis. 3. Wage rate - Conducting sensitivity analysis on wage rates guards against estimation error for shadow wage rate. The sensitivity with respect to wage rate depends upon the labor intensity of the enterprises being compared. The urban wage rate is used in the sensitivity analysis instead of the rural wage rate. 4. Opportunity cost of land is calculated based on the going rate for land rental. Land can be rented at Mk50.00 per acre per season. This gives MK123.50 per hectare. This is a low figure. It would be instructive to base the opportunity cost of land on the returns to land in its most socially profitable alternative use (Morris 1989a), but such data are not available. The land rental will be doubled to see the effect of a higher opportunity cost of land on the comparative advantage. 5. The shadow price of maize is based on Rotterdam freight prices. This does not accurately reflect the trade patterns. Malawi trades with neighboring countries, but the data is unavailable to the author. Maize prices are halved to see the effect of low shadow prices on the resource cost ratios. When the shadow price of maize is halved (table 29), the net social profitability is still higher than the net farm margins (table 27) for all recommendation domains except for Lilongwe RDP MH12 recommendation domain. The 89 MH12 hybrid maize yields in this recommendation domain are not as high as the yields for Thiwi Lifidzi. This is an important point because the shadow prices used in this study assume the starting point is Rotterdam. The assumption, which is not consistent with the trade pattern increases the shadow costs because of high freight costs. Table 29 Thiwi Lifidzi Lilongwe World reference prices increased by 10% 1365.46 2728.64 1656.91 1951.63 World reference prices decreased by 10% 1029.22 2008.42 1230.61 1375.60 Urban wage rate 1174.64 2336.75 1435.18 1635.98 High average yield 1141.15 2027.45 924.47 2088.75 Low average yield 746.35 1087.45 577.22 996.75 Maize price halved 551.14 704.76 548.91 406.13 NB: All the variables reported in this table are discussed in items 1 to 5 of section 6.6 and in section 6.2. 90 An important implication of the social profitability analysis is that existing agricultural policies on soybean and Vulgaris bean intercropping technology provide a disincentive to farmers because the farm net margins are consistently lower than the social profitability, with the exception of the Lilongwe MH12 intercrop. When average high yields, which assumes a lower dosage of fertilizer and poorer management compared to the treatment levels are used, the technologies still exhibit comparative advantage (table 30). Table 30 W W Thiwi Lifidzi Lilongwe WW World reference prices increased by 10% 0.5685 0.2757 0.4791 0.3622 World reference prices decreased by 10% 0.7407 0.3552 0.6189 0.4757 Urban wage rate 0.6624 0.3203 0.5687 0.4244 High average yield 0.6729 0.3525 0.6729 0.3471 Low average yield 1.0163 0.5665 1.2898 0.6124 Maize price halved 1.3015 0.7525 1.3496 1.0063 NB: All the variables reported in this table are discussed in items 1 to 5 of section 6.6 and in section 6.2. 91 This indicates that even at the current lower levels of fertilizer the country has a comparative advantage in using these technologies. Only local maize intercrops loses comparative advantage when the average low yields are used. Three recommendation domains lose comparative advantage when the shadow price of maize is assumed to be half the calculated shadow price. If shadow prices were calculated based on neighboring countries, the transport cost is likely to be lower and hence the shadow prices would be lower. Malawi could then lose comparative advantage. Chapter 7 CONCLUSIONS Wis This study has shown that it is important to do comprehensive analysis of on-farm trial technologies in order to come up with farmer recommendations that are more likely to be adopted. The biological analysis is necessary but not sufficient. Some treatments that give highest biological yield are financially less profitable than some treatments with lower yield levels. Farmers did not indicate preference for any of the highest yielding pulse treatments, although some of them earned high financial returns to investment. This suggests that there are more factors that influence farmer decisions than just yield. 2 2 E' . J ! Financial returns analysis indicates that farmers in the Thiwi Lifidzi RDP local maize intercropping and the MH12 intercropping recommendation domains earn higher financial returns from soybean intercropping than from Vulgaris bean intercropping. In contrast, farmers in the Lilongwe RDP local maize intercropping and MH12 intercropping domains 92 93 earn higher financial returns from vulgaris bean intercrops than from soybean intercrops. In Thiwi Lifidzi, MH12 intercrops have higher net benefits than the local maize intercrops. This is mostly due to the differences in the yield of the maize type because soybeans seem to do better in the local maize intercrops. Similarly, in Lilongwe RDP, MH12 intercrops have higher net benefits than local maize intercrops and yield of vulgaris beans is higher in the MH12 intercrops than in the local maize intercrops. Farmers who can afford to invest in hybrid maize can realize higher financial returns than farmers that invest in local maize intercrops. Win: In both locations, MH12 hybrid maize intercrops give higher maize yields than local maize intercrops. Similarly, in both locations MH12 hybrid intercrops give higher yields of Vulgaris beans than local maize intercr0ps. Assuming most of the maize and Vulgaris beans produced are consumed by the farmers, MH12 hybrid intercrops offer more household food security than local maize intercrops (section 5.2.5). Farmers prefer local maize to MH12 hybrid for food because local maize has better storage and culinary qualities, but in the absence of local maize farmers eat hybrid maize. Most of the maize that farmers buy from ADMARC for food is hybrid maize. It is necessary for the Ministry of Agriculture to emphasize research on factors that will encourage the growing and consumption of hybrid MH12 and 94 other high yielding varieties. Storage and culinary qualities, among other things, should be improved. LLBiak There is more annual variability in MH12 intercrop net benefits than in local maize intercrops (11, 14, 17 and 20). This is a disadvantage of growing MH12 intercrops. Farmers prefer more stable technologies which have lower profitability to technologies with high profitability and variability. MH12 hybrid maize intercrops is a more risky undertaking than local maize intercrops (tables 11, 14, 17 and 20). Local maize is known to tolerate mid-season droughts more than MH12 hybrid maize. Local maize is also known to tolerate late season showers more than MH12 hybrid maize. It is necessary for maize breeders to incorporate mid-season drought and late season showers tolerance in MH12 hybrid maize to reduce yield variability. If the variability can be reduced while maintaining its high yield potential, MH12 can become very attractive. Many farmers including the small smallholders, who are more risk averse, would grow it. This would contribute to household and national food security. 11§_£armsr_2rsfsrsnse§ It is important to include farmer preferences in the derivation of recommendations. This is one important criterion that should be used to supplement marginal analysis because it incorporates factors which marginal 95 analysis cannot capture. It is evident in this study that farmer preferences do not always coincide with marginal analysis and it is up to the researcher to decide which factor to give more weight. Some of the farmer preferences may not be evident to the researcher. Farmers in both locations preferred two soybean and two Vulgaris bean hills in between maize hills (section 2.1.6). In Thiwi Lifidzi MH12 intercropping recommendation domain, two soybean hills in between maize hills were dominated. This suggests some underestimation of the benefits from these treatments by researchers. The possible underestimations are discussed in section 7.8. On the other hand, some treatments that earn high financial returns are not preferred by farmers. In Thiwi Lifidzi local maize intercropping recommendation domain, treatment 10 gave the highest financial returns. In Thiwi Lifidzi MH12 intercoppping recommendation domain, treatment 6 gave the highest financial returns. These treatments represent four pulse hills in between maize hills. None of these treatments were preferred by the farmers. Low cost and lack of seed were some of the reasons the farmers indicated for preferring two pulse hills in between maize hills. This suggests that the farmers' preferences are related to their circumstance. Improvement in institutions serving the farmers will change the farmers' preferences. An increase in the farmers' cash earnings and capital accumulation and seed availability will enable the farmers to invest in high- 96 paying technologies. More research needs to be done in the area of rural cash income and capital accumulation. More efficient seed markets should be encouraged. mm: It is evident that intercropping systems earn higher financial returns than sole cropping systems. In all the recommendation domains there are no statistical significant differences between sole and intercropped maize. Statistically the maize yields are the same. The intercropped treatments, however, have additional yields from the pulses. The net benefits realized from the intercrops are thus higher than those of the sole crop. In order to improve the small farmers' incomes, it is important to place greater emphasis on intercropping systems than on sole cropping. Apart from the financial returns intercropping has other benefits like producing both main and relish dishes in the same field, maintaining soil fertility, etc. (section 2.1.4). WW Malawi's policy is to maintain self-sufficiency in the staple food--maize--and other foodstuffs including vulgaris beans (MOA, 1989a). Malawi's policy also includes reducing imports of commodities, like soybeans, in which it has comparative advantage. Most of the maize that contributes to self-sufficiency is hybrid maize which is grown by large smallholders using subsidized inputs and government loans. Between 1985 and 1988, the production of hybrid maize declined sharply 97 because the government embarked on a fertilizer subsidy removal program while maintaining taxes on the produce sold to ADMARC (Christiansen, R.E., and Southworth, V.R., 1988). To maintain self-sufficiency, Malawi needs to encourage the production of MH12 intercrops. One way of achieving this goal is to remove both taxes on produce and subsidies on fertilizer simultaneously so that the prices of these commodities will come close to reflecting their true values. The social profitability analysis, assuming the starting point for calculating shadow prices is Rotterdam, shows that if the input and output prices of the maize intercrops were closer to the shadow prices, the technologies could be more attractive to the farmers because social profitability is higher than net farm margins. However, this process should be done gradually and cautiously. Transport costs should be given special attention. ( Transport costs are high and contribute to high import and low export parity prices. (This may change because in the late 1989, the Nacala rail route which passes through Mozambique was re-opened. The Nacala route was closed because of the war in Mozambique: All inputs and outputs had to be routed through Durban. This raised the external transportation costs by an equivalent $50 million which was close to 20% of the value of exports and 3% of the GDP by 1984 (Lele, 1989). The re-opening of the Nacala route is likely to reduce the transportation costs. Production costs for crops will 98 most likely decrease because imported inputs would become less expensive. At the same time the value of export commodities would increase due to reduced costs of getting them to the market. The social profitability will most likely increase. If the prices of inputs and outputs come close to reflecting true values, the net farm margins will increase and the technologies will become more attractive. Policy makers can shape comparative advantage. Comparative advantage is not static, although heavily influenced by parameters that can be considered fixed in the short run such as primary factor endowments and technology (Morris 1989b). Comparative advantage is likely to change as these factors change. Technology for example can be subject to deliberate manipulation, e.g., currently, the maize commodity team of the Department of Agricultural Research is screening varieties under lower than recommended rates of fertilizer to identify varieties that could suit low resource farmer circumstances. This is likely to increase the social profitability of maize. Policy makers can thus take an active role in shaping future patterns of comparative advantage by influencing the direction and nature of technological change (Morris, 1989a).; It is important to encourage the production of MH12 hybrid maize and other high yielding hybrids. The MH12 hybrid intercropping are more efficient in terms of net social profitability (section 6.3) than local maize intercrops. While both local maize and MH12 are sold to ADMARC or private traders, hybrid is sold most. MH12 hybrid 99 intercropping production can be encouraged by adopting policies that are designed to take advantage of comparative advantage. MSW It is difficult to include all the benefits of intercropping technologies in an analysis like this one because some benefits, like improving soil fertility, do not have immediate impact. Some benefits, like relish dishes made from Vulgaris bean leaves, can not be easily measured. The financial profitability of the intercropping technologies is based on the grain yield only. While the grain is the most important part of the yield, there are some benefits of intercropping which have not been built into the analysis, e.g., the Vulgaris bean leaf is continuously harvested and used as a relish dish or sometimes the main dish during the food shortage period without significantly affecting the yield of beans (LADD/ART, 1989). The Vulgaris bean leaf is also sold on local markets. If the value of the leaf that can be harvested without affecting maize yields can be determined, it could be included in the analysis. Intercropping maize with pulses has soil-enriching effects because pulses are nitrogen fixers. If the amount of nitrogen fixed and consequently utilized by the maize crop was determined, then the reduction in the amount of nitrogen fertilizers required to maintain the same level of fertility would be included in the analysis. This cannot be easily done. 100 The financial analysis assumes that all the beans are harvested at the end of the season. Vulgaris beans can be harvested green and either be consumed or sold as a horticultural crop. The value of the green Vulgaris beans is higher than the dried Vulgaris beans. Using the dried bean prices underestimates the value of the Vulgaris bean intercrop. This value could be incorporated by determining approximately how much of the Vulgaris beans are sold green and how much are sold or consumed dry. The weighted average price could reflect the vulgaris bean value more closely. If the mentioned benefits among others were included, the analysis would yield different results. vulgaris bean intercrops may perform better than soybeans in Thiwi Lifidzi. APPENDIX Table 1 Partial budgets for ten different sole cropping and intercropping treatments of local maize, hybrid maize, vulgaris beans and soybeans tested in Thiwi Lifidzi and Lilongwe RDPs. DATA PRICES Market Prices Maize MK per Kg 0.26 Vulgaris beans MK per Kg 1.92 Soybeans MK per Kg 0.60 Grain transport to markets MK per Kg 0.02 Field Price (Market prices less harvesting and processing labor, and transportation) Maize MK per kg 0.21 Vulgaris beans MK per kg 1.88 Soybeans MK per kg 0.56 Soybean Seed MK per kg 0.75 SEED RATE (Pulse seed rate assuming one hill between maize hills) Vulgaris beans Kg per ha 15.00 Soybeans Kg per ha 8.00 OPPORTUNITY COSTS Labour for planting pulses On the same station Hrs per ha 22.86 One hill between maize stations Hrs per ha 27.81 Two hills between maize stations Hrs per ha 33.14 Three hills between maize stations Hrs per ha 40.14 Four hills between maize stations Hrs per ha 48.61 Bean harvesting and processing Hrs per kg 0.10 Maize harvesting and processing Hrs per Kg 0.10 Labor MK per ha 0.21 Yield adjustment Percent 15.00 101 mm AVERAGEFEDYEDOW Ava-90mm Magenta-hoe.” WW“ ADJUSTEDYELDM Madam Wang-them,“ Mace-WM eaossraoasuemsvonopm Mun Who-m Soybean Toruonoasmnmm 00313 THAT VARY BY WWW Nether-eh Tantalum.“ mun-am mmusoemm Tm Nata-mm vmwmm WMDWM ABIUTYTOFEEDANAVERAGEFMY What-lawman” “magma-hm Tm “Ila-Mm WWW “.14 ”.14 ON 300.14 eon “.14 0.“ 16mm 800 942 I43 I44 new W1 um heme V. be. V. be. en. II It a. 14mm 13100 1mm 41.1» 800 41 00 I. m n 48.3 ”114 ”.52 11.84 18.” 28.48 48.3 ”.14 “.52 11.84 18.” 28.48 1W.” «50.00 72.02 425.3 “.14 “.82 27.81 $.14 40.14 430 14.” 0.77 «mm amen 180500 800 0.” 0.00 33.” 1O “74 8421 13”.” 10 ”.74 48.81 11.” 1O 1Kfim mfmzhmm'm l-MWUINW In 115.3 10.21 0.00 125.41 10 “.74 84.21 ”.14 0.” 187.“ 437.2 31 .00 “.14 218.“ $18.10 am 12.00 8.“ 18.” M14 418.“ ”.14 281.81 531.” 0.” mac 8.43 28.48 ”.52 270.& mm AVERAGEFELDYELDW William mmmm Maybe-1M ADJUSTEDYIELDOCGIM Womb“ wen-hummu- WWW mmmmmm Mela Vale-shea- Soybean Toruoaoesrabmm MMTVWBYOPERATIONW vuoutebeeneeed Soyeeeneeed Wile-whaling” muesli-9W muvmzoosts new museum Tm WWW haw-lemon.” mp-wmm-W WMTEGWMYSIS Tm Nam Tum” Wtdm WYOMMYSS Tm Mabel-dim VMWWM WMDWW ABUTYTOFEEDMAVEHAGEFMY muwmmmm “unwise-1e) Tm Mun-mm “MW at “2 Pun W Ilene V. beene eeme NI 4156‘» 4156.01 II 7200 an m ”81.78 381.78 000 81 10 0m am 0.00 zeeo 000 000 000 4.” 0m am 000 88.” 778.“ a.” am 11.84 778.“ eeeeo 0.” 11.84 mean 778.“ 39190 0.” 27.81 4.3 1 7 41 &m 41 am 000 0.00 10:1m2lehebmeve'm Tm.“ 10mWhhmm DUO“ ”.14 1.“ 8 41“.” 000 m-MWhhw 1m 10 818.” 71.84 1 81 8.3 48.81 “.27 41“.” 000 fig; 41$.” 181m 381.78 16.88 0.” 778.“ 241 .N 1018.“ 88.” ”8.48 84.88 7.81 11520 10.21 000 18.41 1”.” 10 018.” 84.21 1”.“ 15.41 1288.” 1§E5 41 a.” 211.00 381.78 0.00 778.“ 188.78 810.74 8110 WWW 41 a.” 247.” 381.78 778.“ 1 17.87 “4." 0.00 12.00 8.“ 18.” 878.” 41 “w 8881.78 28070 770.“ 182.78 “78 0.” 18.” 8.48 8.48 018.” 41 “.m “81.78 814.80 778.“ 178.12 “8.11 128.41 mm AVERAGEFELDYELDW mmm wmmm Ava-geeoybealylela WBTEDYIELDM Madam WWW” WWW onosenabmavonopmm Heb Vuloanbean Soybean TOTALGFKBSFELDBENEFUSW ME THAT VARY BY G’ERAT'DNW mmdm mmueoemvss Tm Netbanlbm Wham-em WMDWW MTOFEEDANAVERAGEFMY thnlflneedemhgm “mmmm Tuner-n ”mm “MW 1844.” 800 0.” 427.. 0.00 427.” 0.” 427.“ mm 0.” 11.84 a.” an 2.” 5300 10 84.21 747.84 847.74 88. 14 18.87 moo 181.” MHWZDNUM'M TmTadtheWhOnumm-mnlebam m-mmhum 104 1mm 8.10 Mb 218.” 1044.” 1‘18 Q7.” 104.24 8&10 0.” 2400 10.21 84.21 “7.“ 12541 D mums AVERAGEFIEIDYIEIDMM) wmm ”manual” Weoybeenyiekn ADJUSTEDYIELDMM) Avaeoemm WWW” WWW GKJSSFEDBENEFWBYCFOPNW Mela Vulganbean Soybean TOTALGNFELDWW C(BTSTHATVARYBYOPERATDNW \Mgalebealeeed Soybealeeed Vluganbealmnba Seybeavplalnglfia TOTAL VME (1313 1’67 WIT! WANAUSIS Treannnn Netbeneflnm Tonivanbieooenm mm-WNW-W WHATEOFMANALYSIS Tm Netbanfle Teflvanbleeo‘ Manhattan‘s!» WYOMANALYSIS Tm Netbanflnm Venn-mammals.) WmnlebaM/M MTOFEEDANAVERAGEFMY mummmmm mmmmm Tm ”MW WWW Tile 8 ULONGWEKJPW12NTEWTHAL I41 fl “5 04 Fun new mean HM Mela bean V. bean V. beau sane bl Nb NI 314800 3148.!!! 314100 314300 M 28800 27400 47300 a a III III 2575” 2575.8) 2575.” 2575.0 0.00 mac 232.” 402.06 0.00 0.00 000 am 58880 seen 5‘” seen 0.00 @312 487‘ 755.6 0.00 0.00 am 000 555.” moo 1&5: 1844.58 000 a.” a.” 57.” eon 000 800 000 000 4.” 5.44 4.” 000 0.!” 800 am 000 33” 34.54 54.58 seen an “1.“ 1270.07 1 7 I 0 555.82 750.10 741m 551.” 0.00 11.54 18.“ 25.45 M) 0 M) 1 7 0 2 555.02 750.10 081.” “.40 0.00 11.54 2548 33m 135.” 559.01 1445“ 1 2 3 4 555.52 955.40 “1 .Q 1279.97 0.00 22.51 27.81 33.14 15m 11.30 54m 1 7 2 8 31048 8144.0 2510a M15 0.00 000 20230 23290 Mammal-mam'm 8481.78 1778.87 40.14 71.48 402.08 Tm7ald0aemhvnmw Manhunt“ m-mWhhm 1C5 118.20 10.21 0m 15.41 1848.84 10 790.13 84.21 1278.87 1848.84 48.81 7.“ 24”.” 74300 1888.78 1&41 M18 000 1100 8.48 ”.48 881 .fl 1778.87 84.88 M) 8148.” 51400 LIST 01“ REFERENCES LIST OF REFERENCES Barnes-McConnel, P.M. Personal communication. East Lansing, Michigan State University Bean Cowpea CRSP, 1989. Battacharyya, R.A and Johnson, G.K. ang_Mgthgg§, John Willey and Sons, New York, 1977. Bernsten, R. Personal communication. East Lansing, Michigan State University, 1989. Bernsten, R.H. "Mission Report, Department of Agricultural Economics." East Lansing, Michigan State University,1979. Biggs, S.D. "Resource Poor Farmer Participation in Research. A Synthesis of Experiences from Nine National Agricultural Research Systems." International Service For National Agriculture, 1989. Blackie, M.J "Maize in East and Southern Africa." Lilongwe, Malawi: The Kellog Foundation, 1989. Carr, S.J. "Modification and Extension of the National Rural Development Program.” Paper presented at the Symposium on Agricultural Policy for Growth and Development. Malawi , October 31 - November 4, 1988. Chipande, G.H.R. "Innovation Adoption Among Female-headed Households". Dexslsnment_and_£hanse. 18 (1987): 315-327. Christiansen, R.E., and Southworth, V.R. "Agricultural Pricing and Marketing Policy in Malawi: Implications for a Development Strategy.” Paper presented at the Symposium on Agricultural Policy for Growth and Development, Mangochi, Malawi October 31-November 4, 1988. CIMMYT. n "-_ ‘ 1‘ 10.: 01f 44 . Completely Revised Edition. Mexico, D.F: 1988. 106 CIMMYT. "The Farming Systems Perspective and farmer Participation in the Development of Appropriate Technology.” In Eicher, C.K. and Staatz, J.M.: ' v . Baltimore: The John Hopkins University Press, 1985. Dillon, J., and Hardaker, J.B. . Rome: FAO, 1980. Gittinger,J.P. Baltimore: John Hopkins University Press, 1984. Hansen, A. ”Farming Systems Research in Phalombe, Malawi: The Limited Utility of High Yielding Varieties.” In Jones, J.R. and Wallace, B.J. seeiel_§eienees_end rtfl'l' ’l'.‘ " ‘-. ! v' 71'... 9' -. " 9’ V‘.‘ 9. Agrisulturs_nexeleemen§1 pp145-168- Westview Special Studies in Agriculture Science and Policy. Westview, 1986. Hiwa, S.S. "Agricultural Development Policy and Food Production." In W ' ' . Zomba, Malawi: Center for Social Research, 28 - 31 August, 1988. International Monetary Fund- Internatignal_£inansial Kadyampakeni, J. "Pricing Policies in Africa with Special Reference to Agriculture In Malawi.” Halifax, Canada, Dalhouse University, 1988. Kandoole, B., Kaluwa, B., and Buccola, S. "Market Liberalization and Food Security in Malawi" Sguthsrn ' University of Zimbabwe, 1987. Kangaude, P.M. "Malawi Government Experience with the National Rural Development Program and The Donors." Paper presented at the Symposium on Agricultural policy for Growth and Development, Mangochi, Malawi, November 1-5 1988. Kearl, B., Gafsi, S., Mbilinyi, M., Nabila, J., Norman, D., and Spencer. 0.8- Eield_Data_Qallestign_in_the_figsial selenees; New York: Agricultural Development Council, Inc. 1974. Kingsbury, D.S. ”An Analysis of Price and Non-price Barriers to Agricultural Marketing and Trade in Southern Africa.” Ph.D diss., Michigan State University, 1989. 107 Kishindo, P.A. 1989. "Farming Systems Research In General: Putting the Small Farmer Center Stage in Agricultural Research." lengne1_efi_§e1enee. 1989: 1-6. Kydd, J. "Maize Research in Malawi: Lessons From Failure." I9urnal_f2r_Internatienal_nsxelenment- 1-1 (Jan 1989): 112-114. Lele, U. "Structural Adjustment, Agriculture Development and the Poor: Some Lessons from Malawi." World Bank: Managing Agricultural Development in Africa, 1988. Lilongwe Agricultural Development Division Adaptive Research Team. "Working Papers and Reports." Lilongwe, Malawi. 1984-1989. Malindi, E.S. "Aspects of Agricultural Research in Malawi; An overview." lenznel_efi_seienee. 1989: 7-25. Ministry of Agriculture. 9ui9e_19_Agrieultural_£redustignl Lilongwe, Malawi 1989a. Ministry of Agriculture, Agro-economic surveys. "Input and Produce Output Price List 1989.” Lilongwe, Malawi, 1989b. Ministry of Agriculture, Agra-economic surveys. "Ministry of Agriculture Market Price Bulletin." Lilongwe, Malawi, 1989c. Ministry of Agriculture. "Quarterly Food Security Bulletin." ' . Lilongwe, Malawi, June 1989d. Mkandawire, A.B.C., Ngwira L.D.M., and Edje O.T. "Fertilizer Research in Intercrops." Paper presented at the Workshop on Research Methods for Cereal/Legumes Intercropping in Eastern and Southern Africa, Lilongwe, Malawi, January 27, 1989. Mkandwire, R.M. 1989. "Agricultural Development Strategies in Malawi: Unperceived Problems and Implications, With Reference To Lilongwe Rural Development Project.” lenrnal_gf_§siense- 1989: 26-37- Mkandawire, R.M. and Chipande, G.H.R. ”Smallholder Agriculture in Malawi: The Case For A Target Approach, With Special References to Cases From Selected Agricultural Development Divisions." Paper presented at the Symposium on Agricultural Policy for Growth and Development, Mangochi, Malawi. October 31 - November 4, 1988. 108 Morris, M. "Comparative Advantage and Policy Incentives for Wheat in Zimbabwe.“ Mexico: CIMMYT Economics Working Paper 88/02. 1989a. Morris, M. "Determining Comparative Advantage through Domestic Resource Cost Analysis: Guidelines emerging from CIMMYT's Experience. Mexico: CIMMYT Economics Program Working Paper Draft, July 1989b. Msuku, W.A.B. ”The Potential of Beans in Malawi." Paper presented ar One Day Symposium in Honor of Professor Edje, Malawi, Bunda College of Agriculture, 1984. National statistical Office. u9nthlx_§tatis§isal_nnlletinl Zomba, Malawi: 1989. Office of President and Cabinet Department of Economic Planning and Development (Food Security and Nutrition Unit). E99d_sesuritx_and_Nutritien_nulletinl 1-1. 1989. Phiri, B.S.C. ”Economic Analysis of On-farm Data Using a Microcomputer.” Plan B Paper, Michigan State University, 1986. Reintsma, M. and Lang, P. "The Impact of Economic and Agricultural Policies on Women in Malawi." Washington D.C: Robert Nathan Associates, Inc, 1989. Selenje, M.B. Determination of Growth Prospects for Leuacaena_Leusssennala in Lilongwe Agricultural Development division. Lilongwe, Malawi, 1988. Paper Presented at Spring, A. "Trials and Errors: Using Farming Systems Research in Agricultural Programs for Women in Social Sciences and Farming Systems Research Methodological Perspective on Agricultural Development." in Jeffrey R. Jones and Ben J. Wallace. fles§_ie1_§neeie1 atndies_in_Asrisulturs_§siense_and_zelicx pp 123-142- Westview, 1986. Sibale, P.K. "Importance of Soybeans in Malawi." Paper presented at the In-Country Training Course on Soybean Production and Utilization at Ngabu ADD Residential Training Center, Malawi, 22-27 October 1989. Twyford, I.T. "Development of Smallholder Fertilizer Use In Malawi." Rome: FAO, April, 1988. 109 World Bank. "First Triennial Review of the Malawi Agricultural Research System: Executive Summary." Malawi, April 1989. Yiwombe, D.D. ”Effective Use of Current Intercropping Technologies." Paper Presented at the Workshop on Research Methods for Cereal/Legume intercropping in Eastern and Southern Africa, Lilongwe, Malawi, 1989. 110