Livestock Research for Rural Development 30 (4) 2018 Guide for preparation of papers LRRD Newsletter

Citation of this paper

Associations between intensification interventions and negative externalities in smallholder dairy farms in the Kenyan Highlands

F O Agutu, J O Ondiek and B O Bebe

Department of Animal Sciences, Egerton University, P O Box 536, 20115 Egerton, Kenya


Attaining higher productivity levels while sustaining natural resource base and minimizing health risks are emerging issues confronting dairy farming in the Kenyan Highlands. Rapid shifts from low to high levels of intensification within smallholder dairy farms have significantly contributed to negative externalities associated with dairy farming. This study evaluated the associations of intensification interventions with natural resource depletion and human health risks on basis of a set of indicator variables. Data was from a sample of 140 smallholder dairy farms in two Counties (Kiambu and Meru) benefitting from Kenya Market Led Dairy Program. Evaluation was in two stages, starting with Principal Component Analysis to select indicator variables for regression analysis in the second stage which involved selection of optimal models that quantify the contributions of the interventions to negative externalities. The indicator variable of natural resource depletion of significance was the volume of water use for drinking and service in the farms (5.1 litres per Kilogram of milk produced), which represent negative externality. The variations in water use were greatest from socioeconomic interventions (milk sales) and very little from ecological intervention (manure recycling). The depletion of water would increase with sale of more milk and recycling of more manure on the farm. The indicator variable significant for human health risks was the volume of milk rejected (7.7 Kilograms per month), representing negative externality. Though regression model had very low explanatory power (8.3%), socioeconomic intervention had the largest contribution and a little from ecological intervention, represented by the amount of manure recycled on the farm. The volume of milk rejected would increase with sale of more milk, but less when feeding more concentrates and recycling more manure on the farm. It is concluded that continuous monitoring of socioeconomic interventions are essential to provide early warning about negative externalities that may require reversing.

Key words: milk rejections, principal component analysis, regression, water use


Kenyan milk production systems can be divided into two major categories involving the large scale and small scale dairy farmers. The small scale farmers predominates and contributes more than 80 per cent of the total milk produced in Kenya (SNV 2013a). Increase in small scale farmers as well as changes in intensification levels insert strains and huge pressure on the Earth’s natural resources (Descheemaeker et al 2009) as well as posing serious potential threats involving zoonotic health hazards and antibiotic residues through direct consumption of milk (Omore et al 2005). Land degradation closely relates to degradation of natural resources, vegetation, biodiversity and water normally used in livestock production.

Peden et al (2009) defined livestock water productivity (LWP) refers to the ratio of net beneficial livestock-related products and services to the water depleted in producing them. Increased water intensive consumption patterns are amongst the major global concerns for future uncertainties of water resource as livestock production accounts for 20% of agricultural evapo-transpiration (De Fraiture et al 2007). This is mostly contributed by the variations in drinking and service water observed in different smallholder dairy farms in the production of livestock products and services. Studies by (Peden et al 2003) states that water productivity of livestock can be high or low depending on the context within which livestock production is evaluated.

This relates to the level of intensification used as high levels of intensification are associated with increased amounts of water used through drinking and service water. Besides this, studies by Zhang et al (2007) found out that the concerns on water as a natural resource has become more important than ever in relation to the face of the present climate change. Higher milk production levels within dairy farms have corresponding higher water intake for general use.

Besides attaining higher production levels and margins within these farms, human health risks involving zoonotic diseases, antibiotic residues and aflatoxin attacks are negative externalities that arise due to the continued use of the intensification interventions. These are experienced from the direct consumption of unfit milk and meat products as a result of poor quality status and contamination of the products by pathogens. In Kenya, 86 per cent of all milk marketed is sold raw either directly by producers to consumers or through the informal milk market (Omore et al 2005). Due to this increases in direct raw milk sales, increased access and consumption of raw milk by consumers pose serious potential threats involving zoonotic health hazards and antibiotic residues (Omore et al 2005).

Milk safety is a key attribute that assures good quality products are produced which are fit for human consumption. The increasing role of informal non-processed milk pathways in urban areas continues to raise concerns over public health especially zoonosis of brucellosis and tuberculosis (Muriuki et al 2003). Besides zoonosis, antibiotic residues play a major role as an emerging public health crisis of antibiotic resistance with emphasis on agricultural settings (Landers et al 2012). Their use in food animals selects for bacteria resistance to antibiotic used in humans and these spread via the food such as milk and meat to humans thereby causing human infections (Phillips et al 2004).

Inadequate fodder preservation by farmers exposes fodder to mould contamination thereby increasing the risks of aflatoxin upon feeding on these fodder. This in turn is passed on to consumers through direct consumption of products from infected livestock. All these culminates to lowering the quality status of milk thereby increasing the rates of milk rejections at processors level (Ndungu et al., 2016). This study was therefore designed to establish the relationship which ecological, genetic and socioeconomic approaches to intensification have with negative externalities involving nutrient depletion and human health risks indicators within smallholder dairy farms within the Kenyan Highlands.

Materials and methods

Study area

The study was undertaken on smallholder dairy farms benefitting from the Kenya Market Led Dairy Program (KMDP) interventions in Meru and Kiambu Counties in the Kenya Highlands. The farms represent the leading milk sheds in Kenya with a large population of smallholders intensifying their dairy production and favorable climatic conditions for dairy production. Besides these, there is high participation in dairy farmer cooperatives with predominant small land holdings on which dairy is integrated with crops (Bebe et al 2003; Bebe 2004).

Survey methodology

A cross sectional survey of KMDP smallholder dairy farms within the Kenya Highlands was undertaken between February and June 2016. A sample size of 140 farms was determined (Anderson et al 2003).

where z = desired confidence interval level set at 1.96 for 95% confidence interval,

p = the proportion of a characteristic of the population to be sampled, which was set at 0.735 being the proportion of households in the Kenya highlands that keep dairy animals (Bebe et al 2003),

q = (1- p), and e is the error margin allowable for detecting a difference in the sample and was set at 0.1.

The SNV, the Non-Governmental Organization implementing the KMDP program in Meru and Kiambu Counties provided the list of members of the Cooperatives from which the individual sample farms were randomly selected for farm visit and questionnaire administration.

Data collection

Data was collected using structured questionnaires as well as household interviews and observations at farm level to capture primary and secondary data. Data collected at farm level included estimated quantities of water used (drinking and service) and nutrient losses Nitrogen, Phosphorous and Potassium (NPK) through calculations based on balances on amounts of nutrients added through manure use and extracted from the soil through animal feeds and atmospheric losses. Health risks including number of zoonotic cases, estimates of drug residues, mould contamination on feeds and quantities of milk rejected due to poor quality was also recorded.

Herd estimates on total amount of water used for both drinking and service water were taken to represent variations in amount of water used in these dairy farms. These together with the calculated nutrient balances comprised nutrient depletion while data representing variables on intensification approaches were captured using the questionnaires on the basis of applied ecological, genetic, socioeconomics and externalities of intensification as presented in Table 1. This was as per Tropical livestock units (TLU) with measurements used being 1 for bull, 0.7 for cow, 0.5 for heifer and 0.2 for calves (Bebe 2004).

Table 1. Summary of indicator variables defining the intensification interventions and outcomes

Intensification interventions




Stocking density

Insemination cost

Milk sales volume


Health services cost

Extension visits

Leguminous fodder

Herd replacement cost

Distance to markets

Crop residues

% Holstein-Friesian in the herd

Concentrate use

Off-farm sourced feeds

% Animals registered

Externalities of intensification

Herd productivity

Nutrient depletion

Human health risks

Milk yield

Nitrogen (N)

Milk rejected volume

Calving interval

Phosphorous (P)

Zoonotic disease cases

Age at first calving

Potassium (K)


Production costs

Water use (drinking and service water)

Aflatoxin risk (feed with molds)

Gross margins

Statistical analysis

Individual household data was entered in an excel spreadsheet and analyzed through descriptive statistics using the Statistical Package for Social Sciences (SPSS) version 20 (IBM, 2011). All the variables representing natural resource depletion, human health risks and intensification approaches were subjected to Principal Component Analysis (PCA) to identify significant factors for further regression analysis to obtain relationships between natural resource depletion, human health risks indicators and applied intensification interventions. Retained variables with higher factor loadings after Varimax rotations were subjected to Regression procedures of Statistical Analysis System ,(SAS, 2009) version 9.1, to derive optimal predictive model on the basis of fit diagnostic statistics. The best model choice with low Akaike Information Criterion (AIC) or BIC and high Coefficient of determination (R2) was chosen to explain the relationship between nutrient depletion and human health indicators and intensification approaches. All these tests were used to justify selection of optimal predictive model in the final regression analysis.

The multiple linear regression model fitted was in the form:

Where a = the intercept, b1, b2, b3 … bn are the coefficients for variable x1, x2, x3 … xn respectively and eij is the random error.


Table 2 presents a summary of the descriptive statistics for indicator variables defining natural resource depletion and human health risks for the sampled farms.

Table 2. Descriptive statistics of indicator variables of natural resource depletion in sample dairy farms (n=140)

Externalities indicators




Natural resource depletion













Total water use

Litres/Kg of milk



Human health risks  


% positive cases



Milk rejected at market delivery




Zoonotic diseases

Number of cases/year



Aflatoxin risks

% feeds with molds



In Table 3, the PCA fitted for indicators defining nutrient depletion, human health risks and intensification interventions was satisfactory in sampling adequacy (KMO=0.67, KMO=0.55) and the correlation matrix was not an identity matrix (Bartlett’s test Chi square =382.25, 112.631, p=0.000). Two Principal components were extracted that explained 99.38% and 99.47% of the total variance respectively for nutrient depletion and human health risks and applying the rule of thumb (100/2PCs=50%), only variables loading on PC 1 that explained 64.32% and 81.6% of the total variance were retained for subsequent linear modeling. The retained variables were three socioeconomic indicators (credit uptake, milk sales concentrate use and extension visits), one ecological indicator (manure recycling) and two indicator of genetics (insemination costs and disease control) intervention, all of which have positive associations with total water use while quantities of milk rejected had both positive and negative associations.

Table 3. Retained variables for natural resource depletion and human health risks indicators and intensification interventions after Principle Component Analysis

Indicator variables

Natural resource depletion

Human health risks

Principal component 1

Principal component 2

Principal component 1

Principal component 2

Credit (Loans)



Replacement cost



Milk Sales



Milk rejected




Total water use


Insemination cost



Manure recycling



Extension visits


Disease control




Stocking density


Total variance explained (%)





Rotation method: Varimax with Kaiser-Meyer-Olkin Normalisation. Sampling adequacy (KMO =0.670), Bartlett’s test of sphericity (Chi square =382.246, Sig=000).

Table 4 presents the optimal model for explaining water use and quantities of milk rejected selected out of 31 and 63 models evaluated on the basis of smallest AIC, BIC, C(p), SSE values and largest adjusted R2. This model explained 73% of variation in total water use in dairy farm with socioeconomic (milk sales) and ecological (manure recycling) interventions without the genetics intervention. The model explaining quantities of milk rejected had very low explanatory power (7%) and the indicators of significance are socioeconomic (concentrates use, milk sales) and ecological (manure recycling) interventions. The indicators of genetics intervention had no contribution in optimal model that explained the volume of milk rejected and this negative externality has a positive association with milk sales and negative associations with concentrate use or manure recycling. These models are in the form of:

Total Water use = 1847006 + 0.05021(MU) + 3.45141(S)

Milk rejected = 729974 + 0.29974 + 0.29302(S) - 0.93090(MU) - 0.0009081(C)

Where MU= manure recycling and S= milk sales in Kg per herd and C=concentrate use

Table 4. Optimal model selection for dependent variable Total water use


Variables in the model





SSE ‘000’

Water use

Manure recycling, Milk sales






Milk rejected

Manure, Concentrates, Milk sales






The indicator variables of significance in the selected optimal model (Table 4) for total water use are milk sales representing socio economic intervention and it accounted for 63.3% of the total variation and another indicator variable is manure recycling representing ecological intervention, which accounted for only 6.3%. Both manure recycling and milk sales showed positive association with total water use in dairy farms. The optimal model for the volume of milk rejected (Table 5) show that socioeconomic indicators of significance are concentrate use and milk sales, which accounted for the most (6.5%) of the total variance compared to manure recycling (1.8% of variance). Manure recycling and concentrate use are negatively associated with the volume of milk rejected while milk sales is positively associated with the volume of milk rejected.

Table 5. Coefficients and variance contribution (%) by ecological and socioeconomic indicators in the optimal model explaining water use and milk rejected as a proxies for natural resource depletion and human health risks respectively

Intensification indicators

Water use

Litres/Kg of Milk

Milk rejected (Kg/herd/month)


Variance (%)


Variance (%)

Insemination costs (KES/cow)

Concentrate use (Kg/TLU)



Milk sales (KES/herd/month)





Credit uptake (KES/year)

Manure recycling (Kg/year)








Total variance explained (%)




This study evaluated the contribution of different intensification interventions to negative externalities in dairy production in order to inform implementation of sustainable smallholder dairy intensification. The underlying motive is that while smallholders undertake intensification of their dairy production with genetics, ecological and socioeconomic interventions to increase herd productivity and incomes, intensification can result in negative externalities. The evaluation was implemented in two stages, starting with PCA to select from a large number of indicator variables those with significant effects for regression analysis in the second stage of analysis to quantify the contribution of the interventions to externalities.

Natural resources of importance in smallholder dairy farms are soil nutrients (N, P, and K) and water for which the observed descriptive statistics show a status of ongoing depletion compared to reports in 1990 (Stoorvogel and Smaling, 1990) and in 2004 (FAO, 2004). The present study estimated nitrogen balance of -55.9 Kg/ha, representing a deepening depletion from -38Kg/ha in 1990 and -46 Kg/ha in 2004. A similar trend is observed for Potassium balance of -68.3 Kg/ha, deepening from -23 Kg/ha in 1990 to -36 Kg/ha in 2004 and for Phosphorous balance of -6.2 Kg/ha from a balance of 0 to -1 Kg/ha between 1990 and 2004 in smallholder farms in the Kenya highlands. The higher negative balances depict a negative externality through depletion of soil nutrients (N, P, K) thereby raising an environmental concern.

With the model that evaluated the associations between natural resource depletion and intensification interventions adopted, water use was the indicator variable of importance identified in the optimal model selected. The model explained 72.6% of the variations in volume of total water use (drinking and service water) with one socioeconomic intervention indicator variable – milk sales - and one ecological intervention indicator variable – manure recycling on the farm. The socioeconomic indicator accounted for most (66.3%) of the explained variance with (6.3%) of the variations contributed by ecological indicator.

Water use was estimated at 5.1 litres for a Kg of milk produced, a value that is within the reported range of 4.6 to 6.0 litres of water per Kg of milk (Descheemaeker et al 2009; Federation 2009). With this volume of drinking and service water needed for every litre of milk produced, an average farm sample in this study that has two cows, each producing 10 litres a day and milked for 300 days will on average need 30,600 litres of water in a year. This water resource demand demonstrates negative externality of intensification in high utilization and depletion rates of water from either underground or surface water sources in which these small scale farmers might not have the capacity to supply or store the amount of water needed for utilization throughout the year.

Water use was positively associated with volume of milk sold and manure recycled on the farm. This implies that farmers would deplete more water when selling more milk and recycling more manure on the farm. Depletion of water impacts on future water availability and production costs (Haileslassie et al 2010). Farmers who produce and sell large volumes of milk have to observe high standards of hygiene which they attain using water for service cleaning of the dairy units and equipment within the farms. Higher milk sales also involve the use of bigger milk cans for storage and transportation purposes which in turn consumes more water and detergents during cleaning and rinsing. Positive association between water use and manure recycling within these farms, would entail use of more water for cleaning, hygiene maintenance and in slurry manure to ease distribution within the farms for fodder production because feed production is the largest consumer of water in a crop-livestock system (Steinfeld et al 2006; Descheemaeker et al 2009).

The potential negative externalities of intensification of importance to human health risks includes incidences of zoonotic diseases, drug residues in products, aflatoxin risks, mastitis infections and volumes of milk rejected (Byarugaba et al 2008; FAO, 2014). The study estimated the prevalence of mastitis infections at 66.1% which may trigger indiscriminate use of antibiotics in the treatment of sick animals because observing the recommended withdrawal periods means foregoing revenues from milk sales during that withdrawal period. Indiscriminate use of antibiotics is a negative externality that poses public health threats to milk and meat consumers. Other evidence of negative externalities in this study were 1.1 cases of zoonotic disease per year and 4.9% of the animal feeds with molds attack.

In the model evaluating associations between intensification interventions and human health risks, volume of milk rejected was the indicator variable of significance and was estimated in this study at 7.7 Kg/month, slightly lower than the 10 Kg estimated in wet seasons in smallholder dairy farms by (Muriuki 2003). However, the model had very low explanatory power, explaining only 8.3% of variations in volume of milk rejected. The intensification interventions that were associated with the volume of milk rejected were socioeconomics represented by milk sales and concentrates use and ecological interventions represented by the amount of manure recycled on the farm. Genetic intervention had no contribution.

In the derived optimal explanatory model, the volume of milk rejected was positively associated with milk sales but negatively associated with manure recycling and concentrate use. Results indicate that the volume of milk rejected would increase with sale of more milk, but less would be rejected when increasing amount of concentrates fed and amount of manure recycled on the farm. This is contrary to expectations that selling large volume of milk should trigger farmers to practice high standards of hygiene to reduce post-harvest losses. Milk rejection is an indication of failures in hygienic milk production and handling practices which SNV (2013b) and Ndungu et al (2016) attributed to higher bacterial counts and adulterations.

Negative relationship between milk rejections and manure recycling indicates decreased milk rejections with increased manure recycling within these farms. This could mean that when more manure is produced, farmers recycle more manure on the farm either for fodder and crop production or biogas production which thereby reduce manure accumulation within the farms. This in turn would reduce possibilities of milk contamination with manure (dirt or faecal) to explain negative associations with milk rejection. For concentrate use, more usage was associated with less volume of milk rejections, which could mean that farmers feeding more concentrates have better hygienic milking environment that minimize contamination. With high costs of concentrates, farmers tend to efficiently utilize the available quantities hereby minimizing wastes that could contaminate milk from the dairy cows.



The authors are grateful to extension officers and farmers who supported in data collection for this study.

Conflict of interest

The authors declare that they have no conflict of interest.


Anderson D R, Sweeney D J and Williams T 2003 Essentials of statistics for business and economics. Thomson South-Western, Mason.

Bebe B O 2004 Effects of feeding systems and breed of cattle on reproductive performance and milk production on smallholder farms. Uganda Journal of Agricultural Sciences, 9(1 pt2), 558–563. Retrieved from

Bebe B O, Udo H M J, Rowlands G J and Thorpe W 2003 Smallholder dairy systems in the Kenya highlands: Cattle population dynamics under increasing intensification. Livestock Production Science, 82(2–3), 211–221.

Byarugaba D K, Nakavuma J L, Vaarst M and Laker C 2008 Mastitis occurrence and constraints to mastitis control in smallholder dairy farming systems in Uganda. Livestock Research for Rural Development, 20(1). Retrieved from

De Fraiture C, Wichelns D, Kemp-benedict E, Gordon L J, Hanjra M A, Hoogeveen J , Karlberg L 2007 Looking ahead to 2050: Scenarios of alternative investment approaches. In D. Molden (Ed.), Water for Food, Water for Life: A Comprehensive Assessment of Water Management in Agriculture. (pp. 91–145). London, UK: Earthscan; Colombo, Sri Lanka, pp.91-145: International Water Management Institute (IWMI).

Descheemaeker K, Amede T and Haileslassie A 2009 Livestock and water interactions in mixed crop-livestock farming systems of sub-Saharan Africa: Interventions for improved productivity (No. 133). Colombo, Sri Lanka: International Water Management Institute.44p (IWMI Working Paper 133). 978-92-9090-702-2

FAO 2004 Scaling soil nutrient balances. Enabling mesolevel applications for African realities. Rome. Retrieved from

FAO 2014 Impact of mastitis in small scale dairy production. Animal Production and Health Working paper. No. 13. Rome. Retrieved from

Federation I D 2009 Environmental / Ecological Impact of the Dairy Sector:Literature review on dairy products for an inventory of key issues List of environmental initiatives and influences on the dairy sector. Brussels. Retrieved from

Haileslassie A, Peden D and Taddesse G 2010 Livestock water productivity concepts and practices: Experiences from mixed farming system in East African Highlands.IN: Proceedings of the 4th International Conference on Environment and Education, Ahmedabad, India, 24-28 November 2008. (pp. 1–17). Retrieved from on 16/8/2016

IBMC 2011 (International Business Machines Corporation) IBM SPSS Statistics Version 20.0.

Landers T F, Cohen B, Wittum T E and Larson E L 2012 A review of antibiotic use in food animals: perspective, policy, and potential. Public Health Reports (Vol. 127). Retrieved from

Muriuki H G 2003 Milk and dairy products, post-harvest losses and food safety in sub-Saharan Africa and the Near East: A review of the small-scale dairy sector-Kenya. Rome, Italy: . Prevention of Food Losses Programme. Rome, Italy, 60. Food Agricultural Organization.

Muriuki H, Omore A, Hooton N, Waithaka M, Ouma R, Staal S J and Odhiambo P 2003 The policy environment in the Kenya dairy sub-sector: a reviewSDP Research and Development Report No.2 Smallholder Dairy (R& D) Project. Nairobi. Retrieved from

Ndungu T W, Muliro P S, Omwamba M, Oosterwijk G and Jansen A 2016 Quality control of raw milk in the smallholder collection and bulking enterprises in Nakuru and Nyandarua Counties, Kenya. African Journal of Food Science, 10(5), 70–78.

Omore A, Lore T, Staal S, Kutwa J, Ouma R, Arimi S and Kang’ethe E 2005 Addressing the public health and quality concerns towards marketed milk in Kenya. SDP Research and Development Report No. 3. Nairobi (Kenya): Smallholder Dairy (R&D) Project.

Peden D, Taddesse G and Haileslassie A 2009 Livestock water productivity: implications for sub-Saharan Africa. The Rangeland Journal, 31(2), 187–193.

Peden D, Tadesse G and Mammo M 2003 Improving the water productivity of livestock: an opportunity for poverty reduction. IN: In P. G. McCornick, A. B. Kamara, & G. Tadesse (Eds.), Integrated water and land management research and capacity building priorities for Ethiopia. (pp. 57–65). Nairobi, Kenya: ILRI. Retrieved from on 23/08/2016

Phillips I, Casewell M, Cox T, Groot B De, Friis C, Jones R, Waddell J 2004 Does the use of antibiotics in food animals pose a risk to human health ? A critical review of published data. Journal of Antimicrobial Chemotherapy, 53(1), 28–52.

SAS Institute S 2009 SAS/STAT 9.2 User’s Guide, (Cary, Ed.) (2nd ed., Vol. 2). North Carolina: SAS Institute Inc.

SNV K 2013a Dairy Sector Policy Study and Capacity Needs Assessment of Stakeholder Associations. Nairobi. Retrieved from

SNV K 2013b Quality Based Milk Payment Study Kenya Dairy Sector for SNV KDMP Project report. Nairobi, Kenya. Retrieved from explore/download/kmdp_-_quality_based_milk_payment_study_2013.pdf. Accessed on 11/05/2017.

Steinfeld H, Gerber P D, T, W, Castel, V. and de Haan C 2006 Livestock’s long shadow: environmental issues and options. (Food and Agricultural Organization), UN. Rome.

Stoorvogel J J and Smaling E M A 1990 Assessment of soil nutrient depletion in sub-Saharan Africa: 1983-2000. Report 28. Wageningen, The Netherlands, Winand Staring Centre.

Zhang X, Zwiers F W, Hegerl G C, Lambert F H, Gillett N P, Solomon S, Nozawa T 2007 Detection of human influence on twentieth-century precipitation trends. Nature, 448(7152), 461–466.

Received 1 November 2017; Accepted 27 February 2018; Published 1 April 2018

Go to top