Abstract
Résumé
Introduction
Logic of the partial budget approach
Identification of costs and benefits for small ruminants
An empirical example
Concluding comments
References
S. Ehui and B. Rey
International Livestock Centre for Africa
PO Box 5689, Addis Ababa, Ethiopia
Economic constraints and opportunities for improving small ruminant production systems in sub-Saharan Africa must be understood as the basis for developing interventions. However, most national agricultural research systems (NARS) face a shortage of livestock economists, and biological scientists often lack the skills needed to conduct an economic evaluation of the results of their work. This study presents the partial budget analysis (PBA) framework for the economic analysis of small ruminant interventions for use by livestock specialists. The logic of the PBA and the data needs are discussed first. This is followed by an empirical example which analyses the economics of endoparasite control in sheep production system of the Ethiopian highlands. The analysis is based on experimental data from ILCA's Debre Birhan station. The experiment consisted of four treatments: I. Grazing natural pasture; II. Drenching with anthelminthics; III. Supplemental feeding with wheat bran and noug (Guizotia abyssinica) cake; IV. Supplemental feeding and drenching. Results show that net returns per ewe of treatments 11, III and IV exceeded the net return of the control by Ethiopian birr (EB) 4. 86, 8. 56 and 9.35, respectively (US$ 1 = EB 2.07). The increase in cost for treatment 11 relative to the control was EB 1.44; the added net benefit from this treatment was EB 4.86 per ewe, giving a marginal rate of return of 334%. The increase in cost of treatment 111 relative to treatment 11 was EB 19. 60, while the increase in net return was EB 3. 71 per ewe, giving a marginal rate of return on the increased expenditure of 19%. The cost of adding drenching to treatment 111 (treatment IV) was EB 3.58 per ewe, while the increase in net return relative to treatment 111 was EB 0.78 for a marginal rate of return of 22%. The marginal rate of return of treatment IV relative to treatment 111 was 19.4%. Given the high cost of capital, treatments III and IV cannot be recommended. Drenching alone (treatment II) yielded a very high marginal rate of return under experimental conditions and should be tested under on - farm conditions .
Les contraintes et les opportunités économiques relatives à l'amélioration de la production des petits ruminants en Afrique au sud du Sahara doivent être à la base du développement d'interventions technologiques. La plupart des systèmes nationaux de recherche agronomique en Afrique sont cependant fréquemment confrontés à une carence en économistes, et les autres chercheurs ne sont pas toujours équipés pour faire l'analyse économique de leurs résultats. Nous présentons ici la méthode des budgets partiels comme un cadre d'analyse économique d'interventions à la disposition des chercheurs zootechniciens. La logique de la méthode et les données nécessaires seront d'abord présentées. Un exemple empirique d'application sera ensuite discuté. Il se base sur une expérimentation de contrôle de l'endoparasitisme de moutons des hauts plateaux éthiopiens à la station de Debre Berhan du CIPEA, avec quatre traitements: 1. pâturage naturel, 2. déparasitage interne, 3. supplémentation au son de blé et au tourteau de Guizotia abyssinica, 4. supplémentatione et déparasitage interne. La marge nette des traitements 2, 3 et 4 a dépassé celui du témoin de 4,86; 8,56 et 9,35 Birr éthiopien (EB). Le surcoût du traitement 2 par rapport au contrôle est de 1,44 EB, et- la différence de marge nette de 4,86 EB par brebis traitée; donnant un taux de rentabilité marginale de l'intervention de 334%. Par rapport au traitement 2, le traitement 3 coûte 19, EB de plus et procure 3,71 EB de marge nette supplémentaire, ce qui lui donne une rentabilité marginale de 19%. Le déparasitage ajouté au traitement 3 coûte 3,58 EB par brebis et rapporte 0,78 EB de plus, fournissant un taux de rentabilité marginale de 22%. Le taux de rentabilité du traitement 4 comparé au traitement 3 est de 19,4%. Au vu du coût élevé du capital, la supplémentation seule ou associée ne peut être recommandée. Le déparasitage seul offre un taux de rentabilité marginal très élevé en conditions expérimentales et devrait être testé en exploitation.
Small ruminants (i.e. sheep and goats) are essential components of farming systems in tropical Africa. They are raised mainly for meat, milk, and skin providing a flexible financial reserve (Social Security) in bad crop years for the rural population (Sumberg and Cassaday, 1985). Tropical Africa contains about one-third of the world's goats and one-sixth of its sheep. On average there is one goat or sheep on every 10 ha of tropical Africa and there are about 1.1 head of goats and sheep per person employed in agriculture. Available estimates also show that sheep and goats are equivalent in weight terms, to about 17% or the total domestic ruminant biomass of tropical Africa (Wilson, 1988). Sheep and goats are important also because: (1) they require minimal inputs and maintenance costs; (2) they are less susceptible to stress due adverse changes in climatic conditions (e.g. drought); and (3) they have a relatively high reproduction rate and are easy to dispose of (Winrock International, 1983). Thus in face of the declining crop yields due to movement of cropping onto marginal soil types and diminishing fallow periods, improvement in the production of sheep and goats is likely to improve the welfare of smallholders (Peacock, 1987).
A number of national research organisations in partnership with the International Livestock Centre for Africa (ILCA) are conducting research to improve the productivity of small ruminants on the continent. However, although they provide insights into improving small ruminants productivity, some new technologies are not adopted because of lack of economic advantage over current production methods (Amir and Knipscheer, 1987). Increased animal yield is a necessary but insufficient condition for adoption. A new technology must also be profitable (Nagy and Sanders, 1990). For example, the economic benefits of innovation aimed or reducing reproductive wastage and of improved health, nutrition and management must be determined as a basic for recommending such to farmers.
One problem faced by most NARS, however, is an acute shortage of livestock economists, and biological scientists often lack the skills to conduct an economic evaluation of the results of their work. Moreover, most manuals on economic analysis of biological interventions are crop-oriented (e.g. CIMMYT, 1988; MSTAT-C, 1988). Animals, however, have different characteristics making the use of these manuals not immediately accessible to animal scientists. It is therefore necessary that livestock specialists acquaint themselves with basic economic methods for the economic evaluation of small ruminant technologies.
This paper presents the Partial Budget Analysis (PBA) approach for analysing the economic benefits of alternative small ruminant technologies. While there are other types of budgeting (e.g. whole farm and enterprise budgeting), the partial budgeting procedure is the most useful in farming system research (Worman et al, 1990).
The next section presents the logic of the partial budget approach. The third section describes the data-set required to perform the analysis. Special attention is paid to the direct and indirect benefits and cost necessary for a correct evaluation of small ruminant technologies. In the fourth section an example is provided which is based on experimental data collected in the Ethiopian highlands. The last section presents some concluding qualifications and comments.
Partial budgeting is a method of organising experimental data and information about the cost and benefits from some change in the technologies being used on the farm. The aim is to estimate the change that will occur in farm profit or loss from some change in the farm plan (Boehlje and Eidman, 1984). Partial budgets do not calculate the total income and expenses for each of the alternative plan but list only those items of income and expense that change. They measure changes in income and returns to limited-resources, provide a limited assessment of risk and, through sensitivity analysis, suggest a range of prices or costs at which a technology becomes profitable (Mutsaers et al, 1986).
Assuming for simplicity that the farmer's objective is to maximise returns, the method is as follows: Let NI denote the net income from the sale of animals, or animal products, i.e. the amount of money which is left when total costs (TC) are subtracted from total returns (TR):
|
NI = TR - TC |
(1 ) |
Total costs include the costs of all inputs such as feed, non-feed inputs (e.g. veterinary costs, wages of hired labour, etc.). TC can be separated into two categories: Fixed costs (FC) and variable costs (VC):
|
TC = FC + VC. |
(2) |
Fixed costs represent the costs that do not vary when comparing alternative technologies (e.g. land). Variable costs are those that do vary between the technologies, such as the amount of feed or labour used.
In deciding whether or not to adopt a new technology, a farmer will want to know if it will increase his or her income. Similarly in order to properly screen among alternative technologies for further testing, a researcher will want to know which of the interventions is potentially economically more attractive. The increase of changes in net income (D NI) is the difference between the change in total returns (D TR) and the change in total costs (D TC):
|
D NI |
= D TR - D TC |
(3) |
|
|
= D TR - D VC - D FC |
|
|
|
= D TR - D VC, |
|
|
since, |
D FC = 0 |
|
D FC is equal to zero because by definition fixed costs do not vary. Assuming that capital is not a constraint, the technology with the highest D NI will be recommended. However, higher benefits may not be attractive if they require very much higher costs. New technologies typically require a package of increased inputs (thus additional costs), and farmers will want to consider the increase in costs in their decision. Thus, it is necessary to compare the extra (or marginal) costs with the extra (or marginal) net benefits.
Another criterion which takes the cost factor into account is the marginal rate of return (MRR). MRR measures the increase in net income (D NI) which is generated by each additional unit of expenditure (D):
|
MRR = D NI/D VC |
(4) |
In other words MRR measures the effect on net return of additional capital invested in a new technology, compared to the present one (CIMMYT, 1988). It is not necessary to calculate MRR if the new technology costs less than the farmer's present technology, or if the new technology yields a lower benefit than the present one for a comparatively higher cost. When this occurs, the technology is said to be "dominated". For practical purposes Table 1 summarises the methodology for conducting partial budget and marginal analyses.
In making recommendations, three criteria must be observed: First, if net income remains the same or decreases, the new technology should not be recommended because it is not more profitable than the farmer's present technology; second, if net income increases and variable costs remain the same or decrease, the new technology should be recommended because it is clearly more profitable than the farmer's technology; and third if both net income and variable cost increase (this is usually the case), the marginal rate of return should be looked at. The greater the increase in net income and the higher the marginal rate of return, the more economically attractive the alternative technology is.
The most difficult task in performing partial budget analysis, is the proper identification of the costs and benefits associated with the alternative technologies. Using poor data can lead to wrong recommendations. The minimum amount of data which must be collected depends on the design of the trial and the questions for which answer are required. Once it has been decided what questions need answers, it is possible to identify the variables which will provide the necessary data use in conducting the analysis. Generally the following data are required for partial budget analysis: (1) quantities of inputs which vary between alternative technologies; (2) prices of these variable inputs; (3) yields or productivity levels resulting from the alternative technologies; and (4) prices of the outputs.
Table 1. Partial budgeting format
|
1. Additional income: List the items of income from the alternative plan that will not be received from the base plan. 2. Reduced expenses: List the items of expense for the base plan that will be avoided with the alternative plan. 3. Subtotal: 1 + 2. 4. Reduced Income: List the items of income from the base plan that will not be received from the alternative. 5. Additional expenses: List the items of expense from the alternative plan that are not require with the base plan. 6. Subtotal: 4 + 5. 7. Difference: 3 - 6: A positive (negative) difference indicates that the net income of the alternative exceeds (is less than) the net income of the base plan by the amount shown. |
Source: Adapted from Boehlje and Eidman (1984).
All benefits and costs should be calculated using farm-gate prices. That is, the actual price which the farmer pays for the inputs or receives for his products. Thus all inputs prices should account for any cost, such as transportation required to bring the input to the production site. Similarly if the farmer sells his animal off the -farm, then the transportation and, storage charges, and/or marketing costs in delivering it to the sale point should be deducted from the market price of the animal. If a technology affects the quality of the animal (e.g. attractiveness of the colour, fattened tail, appearance), different market prices should be applied for the different qualities. Sometimes it is necessary to value the fixed assets (e.g. facilities, equipment). These are usually valued using the depreciation technique. A simple formula for depreciation is the original cost minus any expected salvage value divided by the number of years of useful life (Worman et al, 1990).
Table 2 presents a breakdown of the benefits for the economic analysis of small ruminants. Four products are generally considered important: (1) reproductive capacity of animals, (2) weight gain; (3) milk yield, (4) meat yield; and (5) manure. The importance of input depends on the technologies being evaluated. But both cash and non-cash costs should be identified. Cash costs include items like feed costs (e.g. grain, wheat bran), non-feed costs (e.g. veterinary charges, wages for hired labour). Non-cash costs include items like family labour, capital costs, depreciation costs (equipment and animals) and other non-market feed costs (e.g. crop residues, household wastes).
Table 2. Breakdown of benefits and costs for economic analysis of small ruminants.
|
Benefits |
Costs |
|
Primary |
Primary |
|
Increase milk yield |
Cash costs |
|
Increase litter size |
Feed costs: hay, salt and |
|
Culls (ewes, rams, does, bucks) |
minerals, concentrates |
|
Goat meat and lamb or mutton |
grains and feeds for the young |
|
Mohair and wool |
|
|
Hides |
Non-feed costs: veterinary, |
|
Manure |
medicine and drugs, vaccine |
|
Horn and hooves for processing |
|
|
into feed supplements and |
Wages for hired labour |
|
other products |
Pasture rent |
|
Meat and milk by products |
|
|
|
Non-cash costs |
|
|
Family labour |
|
Secondary |
Depreciation facilities, (of equipment etc.) |
|
|
Other non-market feed costs |
|
Urine |
|
|
Weight increase (realised |
|
|
when the animal is sold) |
Secondary |
|
Weeding-grazing |
|
|
Farmers' preference |
Carry disease |
|
Attractiveness of colour |
Destruction of crops and trees |
|
and appearance |
Trampling of land (which causes |
|
Pet |
ecological imbalance) |
|
Seed distribution |
Foul odour |
|
Research |
Noise pollution |
Source: Adapted from Amir and Knipsheer (1987).
One of the major problems in performing a partial budget or an economic analysis is what value to assign to the inputs in a trial and output from the trial. Determining the field price of inputs and outputs can become a difficult exercise especially when dealing with non-market inputs or products. In this case, opportunity cost (which is the value of the resource or product in its best alternative use) should be employed. One example is manure. Because it provides a low cost source of fertiliser for crop, it can be given a value equivalent to the reduced use of chemical fertiliser. Another example is family labour. Assuming labour market is competitive, rural wages for hired labour can be used as a proxy. At this point it is important to note that labour (though it is the primary input most farmers make into agriculture), is one of the more difficult items to value. This is because for most farm household members off-farm employment is not really an option. Also for most household members (e.g. women, children, old men and young boys), there are no paying alternative to on-farm work (see Worman et al, 1990 for more discussion on labour valuation). Finally, if the animal or the animal product (e.g. milk) is consumed at home, the opportunity cost is the amount the household would have to pay to purchase that meat or milk instead of using their own product.
This example is based on experimental trials conducted in the Ethiopian highlands. The purpose of the experiment was to assess the impact of drenching and feed supplementation on sheep productivity. The experiment was designed in response to nutritional and health problems (e.g. liverfluke, coenuris, diarrhoea, anthrax, lungworm) faced by sheep in the region. Data used in this study were from several sources. Biological and input price data were adapted from Ngategize and Brokken (1990, ILCA, Addis Ababa, unpublished data) and Ngategize et al (n.d.) which report on the stimulated effects of interventions on flock productivity and the economic of endoparasites control in the Ethiopian highlands, respectively. Animal price data (by head, and sex) were determined using a general linear model which describes the effect of animal characteristics on the price of sheep in the Ethiopian highlands (Andargachew, 1990).
There were four treatments with 50 ewes in each treatment group. In treatment I (the control) animals were under normal grazing practice with no drenching or feed supplementation. Treatment II included drenching with anthelmintics. In Treatment III, animals received feed supplementation with wheat bran and noug cake in addition to normal grazing. Treatment IV included both drenching and feed supplementation. The feed supplements were composed of 300 g of wheat bran and 150 9 of noug cake (Guizotia abyssinica) per ewe per day. The drenching scheme was composed of Ranide (a Fasciola-specific anthelmintic) and Panacur (a broad spectrum anthelmintic which is effective against gastrointestinal nematodes and lungworms). Both were administered three times a year. Experimental results indicated that nutritional supplementation leads to reduced mortality, heavier animals and higher lambing rates. Control of endoparasites leads to reduced mortality but had ho significant effect on lambing rates or weight relative to control (Ngategize and Brokken, 1990, ILCA, Addis Ababa, unpublished data). Production traits parameters under each treatment are summarised in Table 3.
Table 3. Small ruminant production traits parameters under four treatments in the Ethiopian highlands, 1989.
|
Production parameters in traits |
Treatment I |
Treatment II |
Treatment III |
Treatment IV |
|
Annual reproduction rate (%) |
1.02 |
1.02 |
1.35 |
1.35 |
|
Survival (%) |
|
|
|
|
|
0-4 months |
0.75 |
0.85 |
0.85 |
0.85 |
|
4-8 months |
0.85 |
0.90 |
0.90 |
0.90 |
|
8-12 months |
0.95 |
0.95 |
0.95 |
0.95 |
|
Effective lambing rate(%) |
0.62 |
0.74 |
0.98 |
0.98 |
|
Liveweight at 12 months |
17.48 |
17.88 |
18.78 |
18.78 |
|
Liveweight productivity per ewe (kg) |
10.80 |
13.25 |
18.43 |
18.43 |
|
Number of ewes/ram |
16.0 |
16.0 |
16.0 |
16.0 |
|
Average weight/head |
24.77 |
24.77 |
28.67 |
30.37 |
|
Breeding stock mortality rate (%) |
0.18 |
0.16 |
0.16 |
0.16 |
Note: Treatment I is the control. Animals were under normal grazing. Treatment II included drenching with anthelmintics; treatment III included feed supplementation with noug cake and wheat bran, and treatment IV included both drenching and feed supplementation.
Source: Compiled from several sources including Ngategize and Brokken (1990), Gryssels (1988), and Negategize et al (1990).
In order to perform the partial budget analysis, it is necessary to identify the costs that vary among the treatments. The cost components included feeds, anthelmintics, labour for administering the anthelmintics, breeding stock depreciation charges, and capital costs. Drenching required three people at EB 3 per 50 sheep. Radine was administered three times a year costing EB 0.136 per treatment. Panacur was administered at a dosage of one tablet per adult sheep, three times a year at a cost of EB 0.409 per tablet. Wheat bran and noug cake cost about EB 6.38 and EB 16.33 per 100 kg, respectively. Capital cost were computed assuming an interest charge of 30 %. Depreciation charges were derived based on breeding stock mortality rates of 18 % for the base case (treatment 1) and 16 % for alternative scenarios (treatments II-IV).
Based on a predicted price of EB 2.63 per kg for the female and EB 2.00 per kg 2.24 per kg for the male, animal prices by sex were derived using their average weight per head in each treatment. This resulted in animal prices of EB 65.14 per adult ewe for animals under pasture (base case) and animals under drenching alone (treatment ii) and EB 75.40 and EB 79.87 per adult ewe for animals under supplemental feeding (treatment III) and combined supplemental feeding and drenching (treatment IV), respectively. The corresponding for prices adult males were estimated at EB 55.48, 55.48, 64.22, and 68.02 for each treatment, respectively.
Table 4. Gross return (EB/ewe) of feeding and health management experiments for sheep, Debre Birhan, Ethiopia, 1989.
|
|
Treatment¹ |
Treatment |
Treatment |
Treatment |
|
I |
II |
III |
IV |
|
|
Mean price/kg LW (EB) |
2.57 |
2.57 |
2.57 |
2.57 |
|
Mean liveweight productivity at 12 months (kg/ewe) |
10.80 |
13.25 |
18.42 |
18.42 |
|
Average weigh/head (kg) |
24.77 |
24.77 |
28.67 |
30.37 |
|
Gross return (EB) |
91.41 |
97.72 |
121.03 |
125.40 |
1. Treatment I denotes normal grazing; treatment II represents drenching with anthelmintics; treatment III represents feed supplementation; and treatment IV includes both drenching and feed supplementation.
Using these data, the gross and net return for each treatment were computed and reported in Table 5. It is from this that treatment IV (drenching and feed supplementation) yields the highest return over the control (EB 9.35 per ewe) followed by treatment III peed supplementation), IV peed supplementation) and 11 (drenching) with net return over control of EB 8.56 and EB 4.86, respectively. A priori if cash is not a constraint, feed supplementation would be recommended. However, the increase in benefit is accompanied by an increase in cost. It is, therefore, not obvious that farmers would adopt this technology if it were recommended to them.
The first increase in cost (or drenching with anthelmintics) relative to the control was EB 1.44 per ewe. The added net benefits from this treatment were EB 4.86 per ewe for a marginal rate of return of 334%. The increase in cost of supplemental feeding relative to drenching was EB 19.60 per ewe while the increase in net returns was EB 3.71 per ewe for a marginal rate of return on increased expenditure of 19%. The cost of adding drenching to treatment III in combination with supplemental feeding (treatment IV) was EB 3.58 per ewe while the increase in net returns relative to treatment III was EB 0.78 per ewe for a marginal rate of return of 22%. The marginal rate of return of treatment IV relative to treatment III was 19.4%.
Table 5. Results of feeding and health management experiments for sheep, Debre Birhan, Ethiopia, 1989.
|
|
Treatment¹ |
Treatment |
Treatment |
Treatment |
|
I |
II |
III |
IV |
|
|
Gross returns (EB)² |
91.41 |
97.72 |
121.03 |
125.40 |
|
Cash costs |
|
|
|
|
|
Feeds |
|
|
|
|
|
Noug cake |
- |
- |
8.94 |
8.94 |
|
Wheat bran |
- |
- |
6.98 |
6.98 |
|
Veterinary |
|
|
|
|
|
Ranide |
- |
0.41 |
- |
0.41 |
|
Panacur |
- |
1.23 |
- |
1.23 |
|
Labour |
0 54 |
054 |
|
|
|
Non-cash costs |
|
|
|
|
|
Capital cost (at 30% per annum) |
- |
0.65 |
4.78 |
5.43 |
|
Breeding stock depreciation charges |
12.35 |
10.98 |
12.7 |
13.46 |
|
Total cost that vary |
12.35 |
13.81 |
33.41 |
36.99 |
|
Net return |
79.06 |
83.92 |
87.62 |
88.41 |
|
Net return over control |
- |
4.86 |
8.56 |
5.35 |
|
Marginal rate of return (%) |
- |
334.00 |
19.00 |
22.00 |
1. Treatment I denotes normal grazing; treatment II represents drenching with anthelmintics; treatment III represents feed supplementation; and treatment IV includes both drenching and feed supplementation.
2. See Table 4.
Given the high cost of capital, treatments with supplemental feeding and supplemental feeding with drenching cannot be recommended. However, drenching alone yields a very high marginal rate of return under the experimental conditions (334%) and can be suggested for on-farm testing.
Economics constraints and opportunities for small ruminant systems in subSaharan Africa must be understood as a basis for developing interventions. For example, the economic benefits of innovations aimed at reducing reproductive wastage and improved health, nutrition and management must be determined as a basis for recommending such to farmers. The problem is that most national agricultural research systems face an acute shortage of livestock economists and biological scientists lack the skills to conduct an economic evaluation of the results of their work. In order to properly evaluate alternative small ruminant interventions, livestock specialists will need to acquaint themselves with basic economic methods. In this paper, the partial budgeting analysis approach is presented. It is a simple but powerful approach which consists of organising experimental data and information about costs and benefits of various alternative technologies. Partial budgeting also allows the researcher to carry out marginal analysis which is needed for a correct evaluation of alternative interventions. Marginal analysis is important because although the calculation of net benefits accounts for the costs that vary, a technology may not be attractive to farmers because of the extra cost involved.
In performing partial budgeting the first step is the identification of costs and benefits. This requires: (1) proper quantification of the production parameters; (2) proper elicitation of the inputs used, and outputs produced; and (3) proper recording of the (farm-gate) prices of outputs and inputs. The second step is to convert the identified quantities into costs and returns. All other non-cash costs (e.g. family labour, capital cost, and depreciation charges) must be properly valued. Non-market inputs and costs should be valued at their opportunity cost. In the case of small ruminants four products are generally considered important: they include (1) reproductive capacity of animals, (2) milk yield; (3) weight gain; (4) meat yield; and (5) manure. On the input side, cash and non-cash costs must be determined. Cash costs include feed and non-feed costs, wages for hired labour. Non-cash costs include family labour, capital costs, depreciation charges and other non-market feed costs.
Using partial budgeting, an example is provided which compares the economic potential of a sheep breed under alternative health (endoparasitic control) and feeding systems. There were four treatments including: (I) grazing on natural pasture (the control); (II) drenching; (III) feed supplementation; and (IV) feed supplementation and drenching. Despite the fact that treatments (III) and (IV) yielded higher liveweight animal productivity per ewe than drenching alone (treatment II), economic analysis using marginal analysis has shown that endoparasites control with anthelmintics was economically more attractive. This result clearly highlights the importance of economic analysis in the process of technology evaluation. However, it is important to note that before doing economic analysis, the researcher must properly assess the experimental data to verify that the observed response makes sense from an animal science standpoint.
Before concluding, a few words are worth mentioning. Although the partial budget may indicate that the new technology is "better" than the traditional, it will not show that both technologies produce a loss. Also, a partial budget (though it is easy to interpret) is rarely presented with a statement of the farmer's objectives, the farmer's resource base, and important non-cash consideration. For example, the partial budget does not tell US if labour is available to the farmer to feed the animals. Partial budgets ignore the substitutability of inputs and how they are allocated based on fixed endowments and the implicit prices of the resources. Although they yield important insights into the economic attractiveness Of a new technology versus a traditional one, result of partial budgets should therefore be treated with caution.
Amir P and Knipscheer H C. 1987. A conceptual framework for the economic analysis of on-farm trials with small ruminants. In: Devendra C (ed), Small ruminant production systems in South and Southeast Asia. Proceedings of a workshop held in Bagor, Indonesia, 6-10 October 1986. IDRC-256e. IDRC (International Development Research Centre), Ottawa, Canada. pp. 380-391.
Andargachew Kebede. 1990. Sheep marketing in the central highlands of Ethiopia. MSc thesis, Alemaya University of Agriculture, School of Graduate Studies, Alemaya, Ethiopia. 117pp.
Boehlje M D and Eidman V R.1984. Farm management. John Wiley and Sons, New York, USA.
CIMMYT (International Maize and Wheat Improvement Centre). 1988. From agronomic data to farmer recommendations: Economic training manual. Completely revised edition. CIMMYT, Mexico.
Gryssels G. 1988. The role of livestock in the generation of smallholder farm income in two Vertisols areas of the central Ethiopian highlands. In: Jutzi S C, Haque I, McIntire J and Stares J E S (eds), Management of Vertisols in sub-Saharan Africa. International Livestock Centre for Africa, Addis Ababa, Ethiopia. pp. 345-358.
MSTAT-C. 1988. A microcomputer program for the design, management and analysis of agronomic research experiments. Michigan State University, East Lansing, Michigan, USA.
Mutsaers H J W. Fisher N M, Vogel W O and Palada M C.1986. Farming systems program: A field guide for on-farm research with special reference to improvement of cropping systems and technologies in West and central Africa. International Institute of Tropical Agriculture, Ibadan, Nigeria. 197 pp.
Nagy J and Sanders J H.1990. Agricultural technology development and dissemination within a farming systems perspective. Agricultural Systems 32:305-320.
Ngategize P K, Zere Azaz and Brokken R F. (n.d.) Economics of endoparasit control in the Ethiopian highlands. In: Proceedings of the Eighth SC-CRSP scientific workshop, 7-8 March 1990, ILRAD, Nairobi, Kenya. Small Ruminant Collaborative Research Support Program (SR-CRSP), Nairobi Kenya, pp. 340-347.
Peacock C P. 1987. Measures for assessing the productivity of sheep and goats. Agricultural Systems 23: 197-210.
Sumberg J E and Cassaday K (eds). 1985. Sheep and goats in humid West Africa. Proceedings of the workshop on small ruminant production systems in the humid zone of West Africa. International Livestock Centre for Africa, Addis Ababa, Ethiopia.
Wilson R T. 1988. Small ruminant production systems in tropical Africa. Small Ruminant Research 1:305-325.
Winrock International. 1983. Assessment of the current status and potential role of sheep and goats in developing countries. Winrock International, Morrilton, Arkansas, USA.
Worman F. Norman O and Ware-Snyder J (eds). 1990. Agricultural technology improvement project. Farming Systems Research Handbook for Botswana ATIP RP3. Department of Agricultural Research, Ministry of Agriculture, Botswana.
