An analysis of factors affecting smallholder mixed farming activities, performance and interactions in Wundanyi location, Taita district, Kenya

P M Mwanyumba, R G Wahome*, A Mwang’ombe**, E Lenihan*** and M S Badamana*

Ministry of Livestock Development, Department of Veterinary Services, Private Bag - 00625, Nairobi, Kenya
mwahouse2005@yahoo.com
* University of Nairobi, Department of Animal Production, P.O. Box 29053-00625, Nairobi, Kenya
** University of Nairobi, Department of Plant Science and Crop Protection, P.O. Box 29053-00625, Nairobi, Kenya
*** University of Cork, Ireland

Abstract

Principal component analysis was used to analyze data collected from a dynamic study of thirty mixed dairy-crop farmers. The study was undertaken for 11 months over two rainy seasons to analyze the major production factors affecting the smallholder mixed farming in this area. 

The major factors were identified as own and hired labour, cost of inputs, price of milk, milk marketing distance and herd number dynamics. Farm output, sales and income and cattle characteristics and management were the most important factors influencing the activities, performance and interactions of the system.

Key words: Dynamic study; principal component analysis; production factors

Introduction

Smallholder farming is recognized by the Commission for Africa, NEPAD and others as central to rural livelihoods and therefore indispensable to food se­curity and poverty reduction and the achievement of the Millennium Development Goals in Africa (Sabates-Wheeler et al 2008). The Kenya Vision 2030 has earmarked increased productivity of crops and livestock as one of the strategies to increase value in the Agricultural sector (The National Economic and Social Council of Kenya 2007). ILRI (2006) and Moll et al (2007) observed that the availability and values of production factors interact with the smallholders’ choices for production technologies and market forces to characterize the farming system. Steinfeld and Mack (1995) have discussed the livestock production factors: - livestock themselves, capital, feed, land and labour as well as the production process and the consumption end including marketing and processing. Sansoucy (1995), however, noted that in developing countries, most of the increase in animal products has come from an increase in animal numbers rather than in individual animal productivity. This is because the farm production factors are not utilized towards commercialization but subsistence orientation, with the farmers’ objective being mainly food self-sufficiency. Jaleta et al (2009) argued that in the long run, subsistence agriculture could not be a viable activity and was unlikely to ensure sustainable household food security and welfare. An understanding of the production factors and processes that affect animal production is, therefore, a prerequisite for livestock development (Staal et al 1997). This study analyzed the major factors affecting the Smallholder mixed farming in this area to determine their level of influence on the activities, performance and interactions of the system.

Materials and methods 

Description of the study area

Wundanyi location is one of the locations in Wundanyi division of Taita District in Coast Province. The division consists of some of the high potential parts of the District in Agro-ecological Zone 3 i.e. Semi-humid (Jaetzold and Schmidt 1983).  The average rainfall in the study area is 1,400 mm per annum (Ministry of Finance and Planning 2001).

The study was undertaken in Wesu sub-location, in the high agriculture potential zone, where dairy management is the main farm activity in terms of both inputs and outputs. Farming here is low input/low output, with a higher subsistence than commercial orientation, and continuing demographic and land pressures can compromise survival if production is not increased and off-farm income opportunities are not available.

Data collection

Data was collected in 30 farms sampled purposefully from nine villages in the study area using direct measurements and observations by the researchers and six assistants. The survey was dynamic (that is, it evaluated the farming system components over time as discussed by Quijandria 1994) and longitudinal (it gathered information from the same set of respondents through repeated visits over a defined period as discussed by Staal et al 2003). It was undertaken for 11 months over the two rainy seasons. The following information was recorded:

• Land and crop activities, inputs, costs, outputs and values

  
  

• Herd recording, namely housing, equipment and labour;  cattle herd structure and changes and number of other animals

  
  

• Dairy cattle recording, namely condition, inputs (feed, care and services) and outputs (mainly milk)

  
  

• House-hold recording, namely demographic information, farming objectives, crop and livestock products utilization, farm income, off-farm employment income and income from outside remittances  

These four categories of information were considered in the planning stage to best capture the activities, performance and interactions in the farming system.

Data analysis

Data was entered into excel program and analyzed using Statistical Package for Social Sciences (SPSS) for windows, version 16.0. As described by Quiroz et al (1994), Principal component analysis was used to examine the relationships among several quantitative variables, in this case the following:

1. land area in acres, hired labour hours per day on crops, own labour hours per day on crops, cost of inputs and value of outputs

2. hired labour hours per day on herd management, own labour hours per day on herd management, number of animals bought, number of animals died, price of milk, distance milk sold and total tropical livestock units.

3. individual cattle age, number of other animals fed together, weight estimate, kg forage offered, kg concentrates fed, kg crop residues offered, lts water offered, lts milk fed, cost of labour, own hours on cattle management, cost of spray chemical, cost of worm medicine, milk production in lts, kg feed wasted and kg manure output.

4. age of house-hold head, number of adults in the farm, number of school-going children, number of non-school going children, number of people assisting in the farm, kg/lts yield of outputs, kg/lts output sales, value of sales, kg/lts output consumption, value of consumption, total farm income and off-farm employment income.

The means for the different variables for the thirty individual farmers were consolidated together in each month and considered for the 11 months duration of the study (N). The analysis was done for descriptive statistics, correlation, explanation of variance and rotated component matrix. Principal component analysis is a multivariate technique for describing, simplifying and analyzing data sets where many different variables are measured on a set of samples or objects (Mead et al 2003). The analysis uses the statistical/mathematical concepts of mean, standard deviation, variance and covariance which are distribution measurements and Eigen vectors and values which are important properties of matrices in algebra, to convert a set of original intercorrelated variables into a new set of independent variables i.e. the principal components. The Eigen values are the variances of each principal component. The first component contains as much of the variation of the variables as possible, the second contains as much of the remaining variation as possible and so on. In other words, the component with the largest Eigen values is the first principal component and so on.

The principal components are each linear functions of all the original variables i.e.
 

Y₁ = a_{1 1} X _i1 + a_{1 2} X_{i 2} + a_{1 3} X_{i 3}  + ----- + a_{1 p} X_{i p}

Y₂  = a_{2 1} X_{i 1}  + a_{2 2} X_{i 2} + a_{2 3} X_{i 3}  + ----- + a_{2 p} X_{i p}

Where:

- Y₁  and   Y₂  are the first and second principal components and so on.

- a_{1 1,} a_{1 2------} a_{1 p;}  a_{2 1,} a_{2 2----------} a_{2 p}  are correlation coefficients between the first and second principal components (and so on) with the original variables. The coefficients give the weightings or loadings of the original variables on each of the derived components and thus indicate the relative importance of the original variables to the principal components. A high positive coefficient means high correlation i.e. strong relationship.

- X _i1, X_{i 2, ------} X_{i p}  are the original inter-correlated variables, p in number.

Results and discussion  

Principal component analysis by land and crop activities

Land is the main farm component; it occupies the farmer’s thoughts and labour; it is the sink of inputs and source of outputs and it is the medium through which nutrients are recycled. Table 1 shows the variables and their means and standard deviations and the resulting principal components with their Eigen values, contributions to variation and cumulative variation. The first principal component accounts for almost 41% of total variation and together with the second explains 74% of the total variation. This means that for the farmer, these are the first and second most important factors in describing his/her land and crop activities and if it came to choosing what to concentrate scarce resources on, then s/he should choose the variables in these two components.

The first two principal components are also shown in the table in rotated Component Matrix for better interpretation. The most important variables in principal component 1 are hired labour followed by cost of inputs and then value of outputs with the last having a negative correlation. This means for the farmer that land area cannot be used to explain variation in land and crop activities and labour and inputs matter more i.e. bigger land will not necessarily translate into higher productivity in terms of input: output ratio. Therefore, increased intensity of labour use should translate into higher productivity, but this needs a corresponding increase in other inputs such as capital, fertilizer, water, good cattle and crop genetics and knowledge. In principal component 2, own labour is the most important variable followed by land area which has a negative correlation coefficient. This negative correlation between labour and land means that as land decreases, intensity of labour use should increase to realize the same returns.

Principal component analysis by herd characteristics and management

Livestock are the most important factor in livestock development. Their characteristics and management determine the productivity. Table 2 shows the management variables chosen and their means and standard deviations.

The first two components are shown and they explain 68% of the variation with the first alone accounting for over half of the total. The two components are shown in rotated Component Matrix. The first component which accounts for such a large percentage of the variation is also heavily loaded for the distance milk is sold which means that this was a very important variable in herd management compared to the others. The other important variables are hired labour, price of milk and own labour in that order. In the second component the important variables are animals bought, total animals in the herd and animals which died, in that order. There were big differences in the milk marketing distance between farmers as seen from the large standard deviation.  On the other hand, the price of milk was almost equal at around Ksh 20 in the whole area. This means that there is need for intervention in milk collection and marketing. As seen in the earlier participatory analysis study, mean milk production per day was 7 lts. and sales mean was 5 lts, usually of morning milk. Milk that is not delivered by the farmers themselves is collected by individual consumer buyers or by traders, on foot and bicycle, who rely on collecting from many farmers to get sufficient economies of scale. The traders sell in the District Headquarters about 5 Km away and up to the Provincial Headquarters, Mombasa, 200 Km away by public means. Thus, any prospective large scale commercial milk processor would need to construct a collection point at a central place and work with a large milk shed combining this location and others.

Principal component analysis by individual cattle characteristics and management

Options for increasing livestock productivity include improvements in nutrition, disease control, management and breeding (Upton 2000). Table 3 shows descriptive statistics of the individual cattle characteristics and management variables together with the explanation of total variance for the first two principal components. The components had high Eigen values, but contributed relatively low variance with both together accounting for only 56.8% of total variance.

The two components are shown in rotated Component Matrix. Both components are heavily weighted for several variables and this, together with the relatively low variance they contributed,  shows that several variables should be looked at together to influence the good management of a herd. The farmer cannot, for example concentrate on feeding alone and hope to have a ‘good’ herd, but should combine good nutrition with disease control, breeding and management such as housing, cleaning, clean milk production and record keeping. In addition, instead of working hard merely cutting and carrying (fodder, crop residues and weeds), farmers should worker smarter and try to increase feed productivity by conservation during the wet season when there is plenty; appropriate utilization such as chaff cutting to minimize wastage; other methods of fodder production; introduction of other suitable feed crops such as fodder shrubs and ration formulation. These methods have been recommended by Lanyasunya et al (2006) for smallholder areas which suffer the constraint of inadequate land for forage production. The dominance of Napier grass (Pennisetum purpureum) in this study area as in other smallholder areas in the country especially central Kenya (Mwendia 2007), has led farmers to neglect fodder shrubs such as Calliandra (Calliandra calothyrsus) and others with multiple benefits. Wambugu et al (2006) have adequately discussed the advantages, types, properties, growing and utilization of the important fodder shrubs in East Africa.

Principal component analysis by house-hold characteristics and income

The farming household is the originator of all decisions, the center of all activities and the arbitrator of all inputs, outputs and interactions. Table 4 shows the household demographics and chosen livelihood variables and the explanation of total variance for the first two principal components.

The first component has a high Eigen value and alone explains for almost half of the variation. Together with the second, they account for 64% of the variation. The components are shown in rotated component matrix. Both components have many important variables and this again shows that household livelihoods are characterized by many factors, but the most important appear to be sales and total farm income.

As seen in earlier participatory analysis and dynamic studies, in smallholder subsistence systems most product sales are made to meet small and frequent needs rather than for commerce. The primary driving force is food security and thereafter cash to meet other household expenses. This cash is not necessarily surplus and it is sometimes accumulated at the expense of consumption and it is not from one source, but several namely sale of live animals, milk, eggs, horticulture and other crop harvests that can be spared. This is supported by off-farm income and outside remittances, again not by choice or surplus, but of necessity because that support is really needed. Off-farm income can be classified into four categories (Takane 2007) — agricultural wage income, nonagricultural wage income, nonfarm self-employment income, and other income; and agricultural wage income can be earned by working on somebody else's farm as a laborer. Thus, for resource-poor mixed farmers, the balance sheet can be quite complicated and would need consideration of the roles of all incomes.

Principal component analysis by all the variables combined

Table 5 shows the analysis of all the variables from land and crop activities; herd characteristics and management; individual cattle characteristics and management and house – hold characteristics and income. From the many variables, a total of five principal components are derived accounting for 100% of the variation and carrying a different order of variables, positioned differently from the separate analyses. The correlation coefficients of the variables with the principal components are much stronger than with the principal components in the separate analyses and the variances of these principal components are also much larger.  The heavy weightings (strong correlation coefficients) are not in the first principal component alone, but are distributed in the first three. 50% of the variables originated from house – hold characteristics and income (table 4) with most of them being about sales and income. 40% originated from individual cattle characteristics and management (table 3), 10% from land and crop activities (table 1) and none from herd characteristics and management (table 2). All the variables were originally in a first principal component in their separate analyses.

These observations indicate that the variables are as important as they were before, but there is a changed order of importance and a stronger relationship among them. House – hold characteristics and income and cattle characteristics and management appear to be the most important variables to house-hold livelihoods.

Conclusions

• This analysis shows that the major factors influencing the farming activities, performance and interactions of the farming system in this area are labour, cost of inputs, price of milk and marketing distance and herd number dynamics.

  
  

• Overall, farm output, sales and income are the most important factors followed by cattle characteristics and management (feeding and numbers). All these factors indicate the points of intervention to be given priority by the farmers and any assisting development agencies.

Acknowledgements 

This study was supported by funds from Irish Aid through the University of Cork, Ireland. The cooperation of the Research assistants and the farmers is highly appreciated.

References  

ILRI (International Livestock Research Institute) 2006 Lessons from a changing world: Implications for livestock research and development. In Birthal P S, Taneja V K and Thorpe W 2006 (Editors), Smallholder Livestock production in India: Opportunities and Challenges. Proceedings (14) of an ICAR-ILRI international workshop. pp 66. http://www.ilri.org/Infoserv/webpub/fulldocs/SmallholderLivestockPro/SmallholderSMPinIndia_Final.pdf

Jaetzold R and Schmidt H 1983 Farm Management Handbook of Kenya, Volume. 11- Natural Conditions and Farm Management Information, Part C, East Kenya (Eastern and Coast Provinces), Ministry of Agriculture, Kenya.

Jaleta M, Gebremedhin B and Hoekstra D 2009 Smallholder commercialization: Processes, determinants and impact. ILRI Discussion paper No. 18. pp 1. http://dspace.ilri.org:8080/jspui/bitstream/123456789/27/1/SmallholderCommercializationProcesses_DiscPaper18.pdf

Lanyasunya T P, Rong W H, Mukisira E A and Abdulrazak S A 2006 Performance of Dairy Cows in Different Livestock Production Systems on Smallholder Farms in Bahati Division, Nakuru District, Kenya. pp 133,134.

Mead R, Curnow R N and Hasted A M 2003 Statistical Methods in Agriculture and Experimental Biology 3^rd Edition. Chapman & Hall/CRC, Washington, USA. pp 403.

Ministry of Finance and Planning 2001 Taita Taveta District Poverty Reduction Strategy Paper, Consultation Report for the Period 2001 – 2004. pp 6.

Moll H A J, Staal  S J and Ibrahim M N M 2007 Smallholder dairy production and markets: A comparison of production systems in Zambia, Kenya and Sri Lanka. Agricultural Systems 94 (2007) 593–603. pp 601.

Mwendia S W 2007 Impact of Head Smut Disease (Ustilago kameruniensis) on Napier Grass Yields in Smallholder Dairy Production Systems. MSc Thesis, University of Nairobi. pp       30.

Quijandria B 1994 Minimum data required for characterizing systems. In Ruiz M E 1994 Animal Production Systems Research, Methodological and Analytical Guidelines, Inter-American Institute for Cooperation on Agriculture, Costa Rica. pp 55 - 61.

Quiroz  R A, Ame´zquita M C, Guerra P and Quiel J 1994 Using Information Generated in Animal Production Systems Research. In Ruiz M E 1994 Animal Production Systems Research, Methodological and Analytical Guidelines, Inter-American Institute for Cooperation on Agriculture, Costa Rica. pp154.

Sabates-Wheeler R, Devereux S and Guenther B 2008 Social Protection for the Poorest in Africa: Learning from Experience, Building synergies between social protection and smallholder policies. Institute of Development Studies, Sussex, UK. pp 1.

Sansoucy R 1994 Livestock, a driving force for food security and sustainable development.   World Animal Review 84-85 http://www.fao.org/docrep/v8180t/v8180T07.htm

Staal S J, Chege L, Kenyanjui M,  Kimari A,  Lukuyu B,  Njubi D,  Owango M,  Tanner J, Thorpe W and Wambugu M 1997 Characterization of Dairy Systems Supplying the Nairobi Milk Market. A Pilot Survey in Kiambu District for the Identification of Target Groups of Producers. pp 32. http://www.smallholderdairy.org/publications/Collaborative%20R&D%20reports/Staal%20et%20al-1998-Dairy%20systems%20char%20Kiambu.pdf

Staal S J, Waithaka M, Njoroge L, Mwangi D M, Njubi D and Wokabi A 2003 Costs of milk production in Kenya: Estimates from Kiambu, Nakuru and Nyandarua districts. SDP Research and Development Report No. 1. Smallholder Dairy Project, MoLFD, KARI, ILRI,  DFID. pp 9. http://www.smallholderdairy.org/publications/Collaborative%20R&D%20reports/St3/Pages%20from%20Staal%20et%20al-2003-Costs%20of%20milk%20production%20%20pg%208%20-15.pdf

Steinfeld H and Mack S 1995 Livestock Development Strategies.  World Animal Review 84-85 http://www.fao.org/DOCREP/V8180T/v8180T0a.htm#livestock%20development%20strategies

Takane T 2007 Gambling with liberalization: Smallholder livelihoods in contemporary rural Malawi. Institute of Developing Economies, Japan.  https://ir.ide.go.jp/dspace/bitstream/2344/638/1/ARRIDE_Discussion_No.117_takane.pdf

The National Economic and Social Council of Kenya 2007 Kenya vision 2030; New long Term National Planning Strategy. pp 6.

Wambugu C, Franzel S, Cordero J and Stewart J 2006 Fodder shrubs for dairy farmers in East Africa: making extension decisions and putting them into practice. World Agro-forestry Center, Nairobi, Kenya; Oxford Forestry Institute, Oxford, U.K.

Upton M 2000 The ‘Livestock Revolution’ – Implications for Smallholder Agriculture: A Case Study of Milk and Poultry Production in Kenya, Food and Agriculture Organization. pp 4.  http://www.fao.org/ag/AGAinfo/resources/en/publications/sector_discuss/PP_Nr1_Final.pdf

Received 1 January 2010; Accepted 26 January 2010; Published 1 March 2010

Go to top