Livestock Research for Rural Development 20 (5) 2008 Guide for preparation of papers LRRD News

Citation of this paper

Analysis of three goat production systems and their contribution to food security in semiarid areas of Morogoro, Tanzania

J Safari, L A Mtenga, L O Eik*, F Sundstøl and F H Johnsen*

Sokoine University of Agriculture, P.O. Box 3004, Morogoro, Tanzania
*Norwegian University of Life Sciences, NORAGRIC, P.O. Box 5001, N-1432 Ås, Norway   and


Three goat production systems in semiarid areas of Tanzania were characterised in terms of their productivity, contribution to dietary protein and farm economics using Gairo division as a case study. System A involved keeping of dairy goats only while system B involved keeping of goats purely for meat production and system C involved keeping of both milk and meat goats. Information about the production systems was collected with the help of semi-structured questionnaire in a cross-sectional research design. A purposive sampling procedure was adopted and a total of 91 (28, 33 and 30 from systems A, B and C respectively) households were visited and interviewed. Gross margin analysis was used to assess the three goat production systems.


The average herd size was 4.3, 10.4 and 5.3 goats for systems A, B and C respectively. Over 90 % of farmers under system B (meat only) practised outdoor grazing of their animals while more than 50 % of farmers in production systems A and C either confined their goats under zero grazing or practised restricted grazing on fields close to homesteads. Age at first kidding, litter size and mortality rates were higher (P<0.05) in A than in B and C production systems. The per capita consumption of animal protein in households in production systems A and C was four times higher compared to system B (meat only).  Farmers under production system B had gross margin/herd four and six times less as for A and C respectively. Although all the three systems had positive gross margins, the high overall mortality rates of 17.3, 13.4 and 14.0 % observed for production systems A, B and C respectively greatly affected the profitability of the systems. It is concluded that the emerging systems A and C contribute significantly to the livelihoods of farmers in the semi arid areas.

Keywords: Animal protein, farm economics, profitability, smallholder farmers


Livestock play an important role in food security and income generation to most rural dwellers in Tanzania.Twenty- three per cent of meat consumed in the country comes from small ruminants (Mtenga et al 2004). This contribution to livelihoods by small ruminants is also reported in many developing countries (Peacock 2005). In recognition of this importance, the introduction of dairy goats and their resultant crosses with local goats has been supported by governments, NGOs and other institutions in many parts of Africa (Peacock et al 2005). Most of these projects have been targeting high potential areas. In 1999, crosses of Norwegian goats and Small East African goats were introduced in Gairo; a semi arid area in Tanzania where poverty among households is relatively higher compared to other parts of the country. It was of interest to monitor the performance of dairy goats under this harsh environment. With the introduction of these goats in the area, three goat production systems have emerged. One system involves farmers keeping dairy goats only (various blood levels of resultant crosses between Small East African goats (SEA) and Norwegian land race). Another system involves keeping goats for purely meat production while farmers in the third system keep a mixture of dairy and meat goats. Throughout this paper, the three systems are referred to as systems A, B and C respectively. The performance of these systems and their contribution to livelihoods of people in the area has not been studied. The aim of this study was therefore to compare these systems with respect to income generation and household food security.


Materials and methods 

Three goat production systems (systems A, B and C) in Gairo division, (3645E, 630S; 1200 m above the sea level with annual rainfall ranging from 388 to 656 mm), a semi arid area of Tanzania were characterised in terms of their productivity, contribution to dietary protein and simple farm economics. In system A, various blood levels of resultant crosses between Small East African goats (SEA) and Norwegian land race were kept. In system B, farmers kept SEA only. Goats in system C, were a mixture of dairy and meat goats described above. Average herd composition of the three production systems was obtained by physical counting of the animals at each household. Other information about the production systems was collected using a semi-structured questionnaire. The questionnaire was structured in such a way that information on goat management, performance and herd composition dynamics could be obtained. A purposive sampling procedure in a cross-sectional research design as described by Bryman (2001) was adopted to ensure that sampled households represent the three goat production systems. A total of 91 (28, 33 and 30 from systems A, B and C respectively) households were visited and interviewed. Gross margin analysis was used to assess farm economics of the three systems. The reproductive parameters assessed included age at first kidding, kidding intervals and litter sizes while productive parameters were mainly, off-take, milk yield and lactation length. Animal protein output in each system was obtained using the following procedure:

PA and PC = 35q + 0.1(45nw x 8.5e)

PB = 0.1(45nw x 8.5e)


PA, PB and PC = amount of protein obtained in systems A, B and C respectively,
n = number of goats slaughtered,
w = average live weight of goats at slaughter from a given breed (kg),
e = percent edible carcass yield taken as 81 %,
q = quantity of milk in kg produced in the household,
constants 35 and 45 are proportion of protein in g per  kg of milk and dressing percentage, respectively while 8.5 is percentage of protein in edible carcass (Syrstad 1993).


The amount of milk produced per household as stated by farmers was converted into protein equivalent using conversion factor obtained by actual determination of milk protein using AOAC (1990) procedures. In order to estimate daily amount of animal protein made available per person, the amount of beef and chicken meat consumed by each household was obtained through the questionnaire. Contribution of protein from eggs and fish was not included in this study as these commodities are rarely used.


Costs and benefits in each household for the year 2004 were assessed. These included annual births and transfer-in (Transfer-in refers to goats purchased or received as gifts while transfer-out include goats sold or given as gifts) as inflows while slaughters, deaths, losses in the field and transfer-out2 formed outflows. The costs and benefits were converted into monetary terms using either 2004 market prices (US$1 = Tshs 1000) or derived prices as stated by farmers during the study. Production and reproductive data were analysed using GLM procedures of SAS (1998) to study the influence of the three goat production systems on these parameters. Other parameters on the comparison of the systems were analysed using SPSS and MS Excel


Gross margin (total revenue - total variable costs) analysis was carried out using the system described by Nyaribo et al (1990) which involves the use of variable costs only as key components for decision making in day to day operations. The variable cost items of the systems included labour, veterinary services, feeding, purchase of live animals and losses due to mortality. The benefits were consumption of milk and meat, animal sales and skins calculated over year 2004. The capital value of the stock was obtained by multiplying the average herd size with the average purchase and sale price per animal.


Results and discussion 

Description of the systems


Goats were kept in small flocks, ranging from 1 to 24, with an average of nine goats per household. This can therefore be considered subsistence level of goat production which, Ayalew et al (2003) further described it as a low-cost alternative of saving and insurance. In the present study, 70 % of the meat goat farmers (system B) kept goats to meet requirements that would emerge unexpectedly. However, over 80 % of households in system A, kept goats mainly to provide milk for home consumption and sale of live animals. In this system, meat was a by-product through slaughter or sale of excessive males and culls while in system B, goats were not milked. There was also a wide range of agricultural and off-farm activities (e.g. small-scale trading and seasonal labour) but the complimentarity of these activities with goat farming in its broad context was not studied. However, farmers in the study area were of the same opinion that diversification helps to reduce risks in case of failure of one or more activities. The diversification of activities has been reported as a risk reducing strategy between livelihood components in many parts of developing countries (Valdivia et al 1996; Ellis 2000; Chambers 2003).


In the group discussions, feed scarcity in the three systems was identified as one of the main problems. Various management strategies to solve this problem have been developed depending on the intensity of production and flock size. Expectedly, over 90 % of farmers under system B grazed their goats outdoors. Tethering and provision of supplement feeds was not a common practice in this system. During prolonged dry seasons, some farmers moved their flocks far away from their homesteads where goats could graze crop residues and other feed material around these fields. For efficient labour use, goats from different households grazed together and managed by one person on alternating basis, a practice called kiwili. On the other hand, the majority of farmers in systems A and C either confined their goats under zero grazing or practised restricted grazing on fields close to homesteads. Restricted grazing was reported as a strategy to avoid breeding with unwanted bucks, which is crucial for the success of dairy goat production. There were coping strategies in the dry season by farmers in system A, which included conservation of feeds, such as potato vines, tethering and cut-and-carry of fodder as reported by 10, 56 and 60 % of the respondents respectively. Contrary to system B, supplements in form of concentrates and mineral blocks were given to the animals especially the pregnant and lactating ones in systems A and C. The management practices described above have implications on labour demand. On average, 3, 2.5 and 3.5 hours are required daily in maintaining goats in systems A, B and C respectively.


There was great variation in household flock size between the systems; being greatest in system B followed by systems C and A (Table 1).

Table 1.  Average herd composition per household of the three production systemsa


System A

System B

System C

Herd structure, no. of goats




Kids (0-4 months)





0.40 (9)b

1.00 (10)

1.10 (21)


0.30 (7)

1.20 (12)

1.24 (23)

Young (4-12 months)





0.60 (14)

0.90 (9)

0.54 (10)


1.10 (26)

1.30 (13)

0.80 (15)

Adults (above 1- year)





0.10 (2)

1.40 (13)

0.02 (0.4)


1.80 (42)

4.60 (44)

1.60 (30)


4.30 (100)



aIn this and the subsequent tables, systems A, B and C represent keeping of milk goats only, meat goats only and both milk and meat goats on the same farm respectively.

bFigures in parentheses are percentages of total goats in a household

c Twenty percent constitutes meat goats.

In system A and partly in system C, farmers used bucks located in the project centres and this could account for the lower number of males in these systems compared to system B. The main objective of establishing the buck centres is to use animals with good genetic potential and to minimize inbreeding. In system B, it is speculated that inbreeding is a problem as no systematic culling was reported and bucks could stay in the same flock up to six years.


Flock changes in the three production systems during a one-year interval are shown in Table 2.

Table 2.   Average flock inflow and outflow in one-year interval


System A

System B

System C





Stock at the start of the year





2.60 (32.1)

4.60 (24.7)

3.60 (26.5)


0.3 (3.7)

0.03 (0.2)

1.00 (7.4)

Out flow





1.00 (12.3)

1.40 (7.5)

1.60 (11.8)


1.40 (17.3)

2.50 (13.4)

1.90 (14.0)

   Losses during grazing


2.00 (10.7)

1.00 (7.4)


1.40 (17.3)

2.32 (12.5)

3.80 (27.9)

Standing stock at the end of the year

4.30 (53.1)

10.40 (55.9)

5.30 (39.0)

Figures in parentheses are percentages of total inflow or total outflow

The inflows in the three systems are mainly from goats born within the flock where the outflow consists mainly of deaths and slaughter. The transfer-out ranged from 13 to 28 % with highest observation in system C and lowest in system B. The highest (11%) loss observed in system B in this study was not surprising as the animals were more prone to theft in this system compared to the other systems. The stock at the end of the year is lower than that at the beginning of the year, implying that there is higher outflow of animals, which is beneficial to the farmers. The overall mortality rates ranged from 13 to 17 % with highest value in system A. Therefore, interventions must be developed to address deaths and losses of goats during grazing. Production systems A and C are more affected in the decline of stock numbers due to high demand for dairy goats from these systems. This results into artificially high prices (ranging from US$ 40- 60 per animal), which induce farmers to sell more goats. In these two systems, higher financial returns are increasingly being realised from sale of live animals.


During informal discussion with farmers, it was clear that farmers want to sell live animals to get a lump some of money to be able to undertake tangible projects such as buying corrugated iron sheets for building houses. Farmers stated further that money obtained from selling of milk is used to meet daily household needs.


Goat productivity


The mean value for age at first kidding in the three systems ranged from 13.6 to 16.6 months (Table 3).

Table 3.  Productive and reproductive performance of the three goat production*


System A

System B

System C

Age at first kidding, months

13.6 b

16.6 a

15.1 b

Kidding interval, months




Litter size, no.



1.32 a

Annual milk yield/year/doe, kg**




Lactation length, months




*Within rows, values with different superscripts differ significantly (p<0.05).

** Calculated from dairy goats only and excludes milk suckled by kids

Previous on-station studies at Sokoine University of Agriculture Tanzania by Mtenga and Kiango (1992), found age at first kidding ranging from 14 to16 months, which are close to the present findings. Similar values have been found elsewhere in Africa (Karua 1989). Most of the parameters measured in the present study were not affected by production system except for age at first kidding and litter size. Litter size was significantly higher in systems A and C compared to system B. These findings are not surprising as it is well known that exotic dairy goats tend to have higher litter size (Devendra and Burns 1983). In addition, litter size is a parameter which is greatly influenced by nutrition (Sibanda et al 1999) and in the present study the intensity of feeding was higher in systems A and C.


In the present study, milk yield ranged from 220-233 kg/year/doe Table (3). All farmers with crossbred goats practised a partial suckling system in which, kids were allowed to stay with their dams during the day and separated at night. Milking was usually done in the morning. Under system A, average milk yield was 0.95, 0.70 and 0.50 kg/day in the first 3 months, 4 to 6 months and 7 to 8 months, respectively. After two months of rest, the new lactation started, where milk yield is again 0.95 kg/day. These values are slightly below the range of 1.07 to 1.4 kg/day per goat reported by Mtenga and Kiango (1992) and Nordhagen (2003) who studied the performance of the same genotype in Mgeta, Tanzania. This is a highland area, which experiences cool temperatures and has nutritious forages throughout the year (Ingratubun et al 2000). The average milk yield per doe per year in the present study was about 233 kg. With 1.8 milking goats in system A, average milk yield per year per household was 432 kg. In system C, milk production from 1.3 milking goats per household was estimated at 315 kg/year. Milk yield recorded in the present study is encouraging given the harsh conditions in terms of temperature and feeds exposed to these animals.


Protein content of milk samples taken during the study averaged 35 g/kg and was within the range of 25 to 42 g/kg (Hadjipanayiotu 1995; Barrionuevo et al 2002). This value was used to estimate annual milk protein available per household. In relation to protein output from goats (meat and milk), production systems A and C produced approximately 15 and 12 times as much protein per household annually as system B (Table 4).

Table 4.  Available animal protein per household in the different production systems

Sources of protein

System A

System B

System C

Annual protein available, g per household




    Meat from goatsa




    Milk from goats




    Other sources (chicken and beef)








Daily protein available, g per household




    Meat from goats

3.4 (7)

3.0 (30)

4.1 (10)

    Milk from goats

41.4 (81)


30.2 (75)

    Other sources (chicken and beef)

6.4 (13)

6.9 (70)

5.9 (15)





    Protein /personb




aThe number of goats slaughtered were 1, 1.4, and 1.6; 40, 25, and 30 kg were the average live weights at slaughter for systems A, B and C respectively.

bThere were 6, 5 and 5.5 persons per household in systems A, B and C respectively. Figures in parentheses are percentages.

This study shows that the daily animal protein intake per person per household is far below values recommended by FAO (1996) of 21g/person/day. However, when the total protein produced per day is considered, it is found that systems A and C can meet the requirements of 2-3 people while system B can hardly support one person. The per capita consumption of animal protein in households in production systems A and C was four times higher compared to system B (meat only)


Dairy goats have greater potential to improve livelihood through sale and/or consumption of milk. Under normal situations, drinking of milk by poor households is considered a luxury and therefore one would expect such households to sell milk and use income to buy other sources of protein such as beans, which would meet the needs of the whole family (RIPS 1996). In the study area, most of the milk is consumed at home. In systems A and C, 81 % and 75 % of the household animal protein consumed come from the dairy goats, respectively. It is interesting to observe that besides the dismal amount of protein consumed by households in system B, the role of chicken and beef in this system is very pronounced accounting for 70 % of the total animal protein consumed.


Farm economics


Simple monetary input and output relationship is shown in Table 5. Care must be taken in interpreting these results as only variable costs were considered.

Table 5.  Input and output relationship of the three production systems (US$)a


System A

System B

System C






32.0 (46.3)

27.0 (67.5)

41.6 (56.9)


21.0 (30.4)

4.0 (10)

19.0 (26)

    Veterinary care

16.0 (23.3)

9.0 (22.5)

12.5 (17.1)









   Slaughter and sales

48.6 (27.8)

64.6 (98.6)

144 (61)


126 (71.8)


90.7 (38)


0.7 (0.4)

0.9 (1.4)

0.86 (1)





Gross margins/herd




Gross margins/goat




aCalculated on one- year basis;  b as stated by farmers

Price ranges for goats in systems A, B and C are US$ 40-60, 10-27, and 18-54 respectively

In system A and to some extent system C, fixed costs including the construction of goat houses are likely to be important costs of production. In this study, the cost of inputs was highest for labour followed by feed cost and finally veterinary care. The study shows that labour cost decreases with the degree of intensification. The high labour cost in system B was due to the fact that a large proportion of time is spent in grazing the animals. Milk is the most important output accounting for 72 % of the total output in system A. As expected, meat is the major output in system B because local goats are not milked. In system C, the outputs supplement each other with meat contributing 61 %. In addition, the annual gross margins/goat was highest (US$ 30.7) in system C followed (US$ 25.6) by system A. Farmers under production system B had gross margin/herd four and six times less as for A and C respectively.


The values of outputs and hence gross margins in these systems are likely to be higher if edible non-carcass components of the slaughtered animals were considered, which was not the case in this study. Food crop residues gathered from farmers’ fields and grasses collected from roadsides and valleys were not costed although this might be reflected in the labour cost.


Non-marketed benefits were not quantified due to associated difficulties in taking measurements. Such benefits represent additional values to the systems. They raise returns to the systems and are known to contribute to the livelihoods of people (Staal et al 2003). One of these benefits is manure and through its use in crop production, it contributes indirectly to additional revenues to the household. Manure collection is higher in systems A and C because feeding of goats was under total or partial confinement. Another benefit is the use of goats as insurance in case of emergences and other needs. For example, Nordhagen (2003) indicated that introduction of dairy goats in Tanzania, has contributed significantly to the education of children. In future, there is need to study the role of these non-marketed benefits in determining the adaptation of the systems by farmers.




The Norwegian Agency for Development and the Department of International Environment and Development Studies-NORAGRIC of the Norwegian University of Life Sciences (UMB) are acknowledged for facilitating this study.



AOAC 1990 Association of Official Analytical Chemists. Official Methods Analysis. 15th Edition. AOAC, Washington, Volume 1, p. 62


Ayalew W, King J M, Bruns E and Rischkowsky B 2003 Economic evaluation of smallholder subsistence livestock production: lessons from an Ethiopian goat development program. Journal of Ecological Economics 45: 473 – 485


Barrionuevo M, Alferez M J, Lopez I, Aliga M R, Sanz Sampelayo M and Campos M S 2002 Beneficial effect of goat milk on nutritive utilization of iron and copper in Malabsorption syndrome. Journal of Dairy Science 85: 657-664


Bryman A 2001 Social research methods. Oxford University Press Inc., New York, p 540

Chambers R 2003 Whose reality counts? Putting the first last. ITGD Publishing, Southampton Row, London, p. 297


Devendra C and Burns M 1983 Goat production in the tropics. Second Edition CABI Publishing, Wallingford, UK


Ellis F 2000 Rural livelihoods and diversity in developing countries, Oxford University press, p.273


FAO 1996 Food requirements and population growth, FAO, Technical Paper WFS 96/TECH/12 provisional version May 1996, Food and Agriculture Organisation, Rome, Italy p.59


Hadjipanayiotou M 1995 Composition of ewe, goat and cow milk and colostrums of ewes and goats. Small Ruminant Research 18: 255 – 262


Ingratubun G F, Owen E, Massawe N F, Mtenga L A and Mtengeti E G 2000 Effect of upgrading small East African goats on feed resource utilisation in the Uluguru mountains in Tanzania; a farmers' perspective. Livestock Research for Rural Development. Volume 12, article # 3 retrieved December 2, 2007 from


Karua S K 1989 Some performance indictors of Malawi indigenous goats under village and ranch conditions. In: Wilson R T and Azebu M (Editors) African Small Ruminant Research. and Development. ILCA, Addis Ababa, Ethiopia. Proceedings of African research and development conference held at Bamenda, Cameroon. January 18-25, 1989, p. 7-22


Mtenga L A, Chenyambuga S W and Kifaro G C 2004 Carcass characteristics of indigenous goats in Tanzania: A review In Proceedings of the scientific conference of Tanzania Society of Animal Production, held at Kilimanjaro Crane Hotel -  Moshi, 5th-7th October 2004. Volume 31: 202-207


Mtenga L A and Kiango S M 1992 Factors affecting performance of dairy goats and their crosses at the university farm-preliminary results. Journal of the Zimbabwe Society for Animal Production 6: 104-109


Nordhagen Ø M 2003 Productivity of goats in high potential areas of East Africa. A case of Mgeta, Tanzania, MSc Thesis, Norwegian University of Life Sciences


Nyaribo F B, Knipscheer H C and Young D L 1990 Optimal strategies for adoption of dual purpose goats and four forage technologies in Hamis and Siaya clusters. In: Proceedings of the eighth Small Ruminant Collaborative Research Support Programme (SR -CRSP) Scientific workshop, March 7th - 8th, ILRAD, Nairobi, Kenya, p 36-44


Peacock C 2005 Goats – A pathway out of poverty. Small Ruminant Research 60: 179-186


Peacock C, Devendra C, Ahuya C, Roets M, Hossain M and Osafo E 2005 Goats. In Owen E, Kitalyi A,  Jayasuriya N and Smith T (Editors). Livestock and wealth creation. Improving the husbandry of animals kept by resource poor farmers in developing countries, Nottingham University Press, U K, p. 601


RIPS 1996 Small stock project evaluation. Rural Integrated Project Support Programme, Mtwara and Lindi Regions, Monitoring and evaluation report No. 1, Ministry of Agriculture and Livestock Development, Dar-es-Salaam, p. 66


SAS 1998 Statistical Analysis System. SAS/STAT user’s guide. Statistical Analysis Institute, INC. Carry. NC.USA, p. 1028


Sibanda L M, Ndlovu L R and Bryant M J 1999 Effects of a low plane of nutrition during pregnancy and lactation on the performance of Matebele does and their kids. Small Ruminant Research 32: 243-250


Staal S J, Waithaka M, Njoroge L, Mwangi D M, Njubi D and Wokabi A 2003 Costs of milk production in Kenya. Smallholder Dairy Project Research and Development Report 1, Regal Press Kenya Limited, p. 30


Syrstad O 1993 Evaluation of dual – purpose (milk and meat) animals. World Animal Review. 77: 56-59


Valdivia C, Dunn EG and Jette C 1996 Diversification as a risk management strategy in an Andean agro-pastoral community. American Journal of Agricultural Economics 78: 1329- 334

Received 18 January 2008; Accepted 4 March 2008; Published 1 May 2008

Go to top