Assessing potential for producing dairy replacements under increasing intensification of smallholder dairy systems in the Kenya highlands

B O Bebe

Animal Science Department, Egerton University, Box 536 Njoro, Kenya
obebeb@yahoo.com

Abstract

Smallholder dairy intensification in the Kenya highlands is characterised by a shift from free- to semi-zero- or zero-grazing management in response to inter-generational division of landholdings, keeping of smaller herds of dairy breeds, dependency on external feed resources and poor reproductive performance. As other sources of breeding stock are limited, the changes in herd structure and demographic rates in smallholder herds raise concerns as to whether these herds can produce sufficient replacement stock needed to sustain the continuing intensification. A deterministic model was developed to assess the potential for smallholder herds to be self-sustaining and to generate surplus replacement stock for aspiring dairy farmers as management continues to shift from free- to zero-grazing. The base situation reflected the actual proportion of free, semi-zero- and zero-grazing farms and the size, structure and demographic rates of the herds in representative low-, medium- and high- intensive farming systems in the Kenya highlands.

Model estimates at the base situation showed that the replacement stock available in all the three systems were sufficient for maintaining the breeding population with a surplus remaining for aspiring farmers adopting/or shifting from free- to semi-zero- or zero-grazing. However, the number of replacement stock available decreased with the reduction in farms practising free-grazing, because semi-zero- and zero-grazing farms are unable to produce their own replacement stock. A 4% annual decrease in the proportion of free-grazing farms resulted in insufficient replacement stock to maintain the existing herds in high- but not in low- and medium- intensive systems. In free-grazing farms, reducing the rate of cow mortalities and the proportion of replacement stock sold during the rearing period were the most promising interventions to sufficiently produce the needed replacement stock. Prospects for the continued intensification of smallholder dairying in the Kenya highlands thus depend upon the proportion of free-grazing farms maintained within the farming systems. A rational policy would be to promote intensification of smallholder dairying within a stratified dairy sub-sector.

Key words: dairy intensification; generating replacement stock; landholding subdivision; smallholders

Introduction

Insufficient supply of dairy heifers is a major constraint to the development of smallholder dairy production in many developing countries (De Jong 1996; Afifi-Affat 1998). In Kenya where the integration of dairying into smallholder farming has been relatively successful, public owned large-scale farms used to produce replacement heifers for smallholders at subsidised prices. However, a large majority of large-scale dairy farms including those privately owned have collapsed or have been subdivided for resettlement of smallholders, particularly in the highland areas where dairying is concentrated (Bebe et al 2002).

With human population densities continuing to rise and landholdings to shrink in the Kenya highlands, the trend amongst smallholders is to intensify their dairying from free- to semi-zero- and zero-grazing (stall feeding). This intensification is characterised by keeping smaller herds of dairy breeds often without heifers, dependency on external feed resources and poor reproductive performance (Bebe et al 2002). A recent study of the dynamics of smallholder herds in the Kenya highlands showed that semi-zero- and zero-grazing farms, which comprise over three quarters of all the farms, are unable (or unwilling) to produce the heifers needed to maintain their herd size, whereas free-grazing farms produce surplus heifers (Bebe et al 2003a). Smallholders source heifers mainly from fellow smallholders within their local farming systems and less frequently from outside their farming systems or from large-scale farms because of high pricing and fear of poor adaptability. The trend of shifting from free- to semi-zero- or zero-grazing may result in the need for externally produced dairy heifers for replacement and raises concern about the prospects for maintaining and expanding smallholder dairying in the Kenya highlands.

Deterministic herd models have been used to explore the potential for producing the heifers needed to sustain current herd populations and a surplus for sale to other farmers (Shaw and Hoste 1987; Houterman et al 1993; Affi-Affat 1998). These models projected heifer generation only on the basis of reproductive and survival rates in the herds, ignoring the dynamics in the farming systems. Smallholder farming systems are dynamic, and as exemplified in the Kenya highlands (Bebe et al 2002), they evolve with changes in population growth, land use, economic conditions and development policies. A deterministic model was thus developed to assess whether there is the potential in smallholder farming systems of having herds that are both self-sustaining and generate surplus heifers for aspiring dairy farmers within systems intensifying from free- to zero-grazing dairying.

Materials and methods

Farming systems

Three smallholder farming systems in the Kenya highlands representing low-, medium- and high- intensive dairy farming systems have been described by Bebe et al (2002). Table 1 gives the household population, their dairying management practices, herd size and structure. The land use patterns, population densities and farm sizes, market access, and the proportion of dairy farms practising free-, semi-zero- and zero-grazing within a farming area defined the farming systems. Smallholder farming households in Olkalou division of Nyandarua district, Rongai division of Nakuru district and Limuru division of Kiambu district respectively represented the low-, medium- and high- intensive farming systems.

The low intensive farming systems are found in areas of high agro-ecological potential for cropping and dairying (Jaetzold and Schmidt 1983), but with low market access. Market access was defined on the basis of human population density, local demand for milk, types of roads and the availability of milk marketing institutions (Staal et al 2001). The high intensive farming systems are found in areas of high agro-ecological potential with better market access. Medium intensive systems are in the medium agro-ecological potential areas with medium market access. Proportionately, there are more zero-grazing farms in high- than in low- intensive farming systems in which human population densities are lower. Consequently, on average farm and herd sizes are smaller but cattle stocking rates are higher with the shift from low to high intensive farming systems. As average herd size decreases cows generally form a larger part of the herd, which has fewer or no heifers for replacement. A characteristic pattern of management with the shift from low to high intensive farming systems is increased use of purchased feeds (Bebe et al 2002).

Model design

A dynamic deterministic model was developed to estimate on an annual basis the production of dairy heifers needed to maintain the herd population and surplus heifers for aspiring farmers in the representative low-, medium- and high- intensive farming systems when smallholder farms shift from free- to semi-zero- or zero-grazing dairying. The schematic representation of the model is given in Figure 1 and it is operated in Microsoft Excel ®.

Qualitative data to conceptualise the model relations and the quantitative data to quantify those relationships reflect the dynamics of smallholder dairy systems based on a recent characterisation survey in the Kenya highlands (Staal et al 2001; Bebe et al 2002; Bebe et al 2003a). Table 2 gives the cow age-structure distribution and the annual demographic parameters - calving, mortality, selling and buying rates in the free-, semi-zero- and zero-grazing farms.  

Household input data are the total number of households and the proportions of those practising free-, semi-zero- or zero-grazing in each farming system. Herd input data are the size, structure and annual demographic rates of the herds for free-, semi-zero- or zero-grazing farms. The input values used for the base situation reflect smallholder dairying in the Kenya highlands (C.B.S. 2000; Staal et al 2001; Bebe et al 2003a).

The simulation of the herd dynamics uses a matrix recurrence equation according to Caswell (1989) and Lesnoff (1999). The simulation is run for a period of ten years and is performed separately for each farming system. The availability of dairy heifers needed to maintain the herd population in each farming system and any surplus heifers available for aspiring farmers is calculated on an annual basis using the information on household numbers and their type of dairying with the associated herd size and structure and demographic rates. The model calculates the initial herd population from the number of households with cattle given the size and structure of the herds. The herd projections use the demographic rates overtime and assume that population growth only depends on demographic rates of females. A herd comprises one class of heifer-calves (pre-weaned females), two classes of heifers (post-weaned females below one year and above one year until first calving) and eleven classes of cows (3 to >12 years of age) to reflect the observed herd structures (Staal et al 2001). Of the heifers and cows purchased, 90% originate from within the farming system and the other 10% from outside the farming system (Bebe et al 2003a), linkage with medium- and large-scale farms is thus assumed insignificant.

Herd projections

The distribution of individual female animals over age groups in year t is given by a vector,

(t) = [n₁(t),….., n₁₄(t)]

where n₁(t),…..,n₁₄(t) are the number of females in age group 1 to 14 in year t. 

This vector is linked from one year to the next by an age-transition matrix that contains the maximum likelihood estimates of annual birth rates and the survival rates for each animal class in projecting the population changes from year t to t+1. This population projection matrix A writes as:

n (t + 1) = An (t)          (1).

Survival from year  t  to  t + 1  writes as  P_i+1 = 1 - m – s + b

based on Lesnoff (1999), where m is mortality rate, s is selling rate and b is buying rate associated with each animal class in free-, semi-zero- or zero-grazing farms. Demographic rates apply to the population at the beginning of the year. 

Replacements to maintain herd size and to supply a surplus for aspiring dairy farmers

The number of dairy heifers produced annually is calculated as females surviving to age at first breeding. This number is expressed as dairy heifers available per cow leaving the herd, which includes all deaths and sales. The productive life of a cow is defined by the probability of disposal (death and sales: Table 2); for instance, a disposal probability of 0.26 in zero-grazing farms translates (reciprocal) to a productive life of 3.8 years. Thus, the number of dairy heifers per cow leaving the herd has to be equal or greater than one (1) if the herd size is to be at least maintained on an annual basis. When it is below one (1), it implies that dairy heifers for replacement outside the farming system have to be bought to maintain or expand the herd size. Any available dairy heifers above the numbers needed to replace cows leaving the herd for purposes of maintaining the population is surplus. These surplus heifers are available for potential adopter farmers. These are non-cattle-keeping households or those presently owning free-grazing farms who are more likely to adopt/or shift to semi-zero- or zero-grazing because their holdings become smaller due to subdivision and fragmentation through family inheritance. The number of dairy heifers per adopter farmer has to be above or equal to one (1) for surplus dairy heifers to be available for the potential adopter dairy farmers.

Sensitivity analyses

Sensitivity analyses evaluated the effect of the decrease in the proportion of free-grazing farms and the changes in herd demographic parameters on the number of dairy heifers available for replacing cows leaving the herd and a surplus for farmers potentially adopting/or shifting to semi-zero- or zero-grazing. Relative to the base situation, the proportion of households annually shifting from free- to semi-zero- or zero-grazing was set to vary from 1 to 5 percentage units to reflect the ongoing shift from free- to semi-zero- or zero-grazing (Table 1) observed in these farming systems (Bebe et al 2002). The decrease in the proportion of free-grazing farms results in an increase in the number of semi-zero- and zero-grazing farms, and the probability of a farmer shifting from free- to semi-zero- or to zero-grazing dairying is assumed to have an equal probability because a farmer may adopt either of these.

An annual change of ±3 percentage units relative to the base situation in calving rates, heifer-calf mortality rates, cow mortality rates and proportion of dairy heifers sold during the rearing period, were made in free-, semi-zero- or zero-grazing farms in each of the farming systems. This was to reflect feasible interventions on the basis of experiences with smallholder dairying systems reported in literature. Animal health interventions in intensive smallholder herds in Kagera region of Tanzania reduced overall cattle mortality rate from 11.5 to 7% over a period of nine years (De Jong 1996). The introduction of improved calf-rearing packages for smallholders in Bahati area in the Kenya highlands reduced annual calf mortality rate by 6 percentage units (Lanyasunya et al 1999).

Results

Base situation

The projected number of dairy heifers produced per cow leaving the herd annually in free-, semi-zero- and zero-grazing farms at the base situation was respectively 1.38, 0.89 and 0.78. These rates imply that for the purposes of maintaining the existing breeding herd population on annual basis, free-grazing farms produced surplus heifers whereas semi-zero- and zero-grazing farms produced insufficient heifers. When farms were aggregated at the farming systems level, there was an annual surplus of 7.7, 11.3 and 3.9% heifers in the low-, medium- and high- intensive farming systems, respectively (Figure 2).

The surplus heifers produced in free-grazing farms were sufficient to offset the deficits in semi-zero- and zero-grazing farms, with a surplus remaining. In the high-, medium- and low- intensive farming systems respectively, surplus heifers were sufficient for 1.6, 2.9 and 4.8% of potential farmers aspiring to adopt/or shift to semi-zero- or zero-grazing on annual basis for a ten-year period (Figure 2). Although more surplus heifers were available in the medium- than in the low- intensive farming systems, potential adopter farmers were more (42 vs 15%: Table 1), hence the lower proportion of farmers obtaining at least one dairy heifer.

Effect of decrease in the proportion of free-grazing farms in the farming system

The number of surplus heifers reduced with the decrease in the proportion of free-grazing farms, consequently lowering the proportion of farmers obtaining at least a dairy heifer when aspiring to adopt/or shift to semi-zero- or zero-grazing dairy management (Figure 2). Relative to the base situation, an annual decrease of 4 percentage units in the proportion of free-grazing farms produced insufficient heifers for maintaining the existing breeding herd population in the high intensive system, but not in the low- and medium- intensive systems. With a 3 percentage unit annual decrease in the proportion of free-grazing farms, the proportion of potential dairy adopter farmers obtaining at least a heifer were respectively 0.3, 2.2 and 3.3% in the high-, medium- and low- intensive farming systems. Through sensitivity analysis it was estimated that the minimum proportion of free-grazing farms needed to maintain self-replacing herds was respectively 18, 15 and 12% in the high-, medium- and low- intensive farming systems.

Effect of changes in the demographic rates

Figure 3 shows the percentage change relative to the base situation in the number of surplus heifers resulting from a ±3 percentage unit change in the demographic rates made in free-, semi-zero- and free-grazing farms in the low-, medium- and high- intensive farming systems. The change in the number of surplus dairy heifers was consistently higher for changes made in free- than in semi-zero- or zero-grazing farms.

Thus interventions made in free-grazing farms were the most promising. A decrease in cow mortality followed by a decrease in the proportion of heifers sold during the rearing period had the greatest percentage effect on the number of surplus dairy heifers produced.

Table 3 shows the effect of a ±3 percentage unit change to improve calving rate, reduced heifer selling rate and reduced cow mortality in free-grazing farms on potential dairy adopter farmers (%/y) obtaining at least a dairy heifer in low, medium and high intensive farming systems. Of the potential dairy adopter farmers, those obtaining at least a dairy heifer were highest with the decrease in cow mortality rate followed by a decrease in heifer selling rate.

Discussion

Dairy production by smallholder farmers is a means to achieve multiple objectives: improved food security, supporting crop production, building capital assets and generating cash income. Smallholder farmers in the Kenya highlands pursue dairying intensification to maximise the returns from their limited land and capital. Development agencies encourage intensification of dairying as a sustainable pathway out of poverty for smallholders (MoA 1998; Delgado et al 2001). An important question is whether smallholder herds have the capacity to produce their own replacement heifers in numbers sufficient to maintain the breeding population and to generate surplus heifers for other aspiring dairy farmers.

Projections over time from deterministic models can be a useful basis for exploring the dynamics of livestock populations in different farming systems and with different interventions for development planning or for productivity assessment (Upton 1989; Wakhungu and Baptist 1992; Lesnoff 1999). This study applied a deterministic model with input values reflecting the prevailing herd reproductive performances and the dynamics of smallholder farming systems in the Kenya highlands. The model results suggested that the potential for maintaining the current dairy herd and its continued expansion will be dependent upon the proportion of free-grazing farms maintained in the farming system. With the stratification within the smallholder dairy systems, the surplus heifers produced in free-grazing farms serve as foundation stock for new farmers or as replacement animals for the existing semi-zero- and zero-grazing farms in these farming systems. The valuable complementarity amongst these farming systems provides a market in the semi-zero- and zero-grazing farms for the surplus heifers produced from the free-grazing farms.

With the pressure on land from the continually rising human and cattle populations, the proportion of households practising free-grazing is projected to reduce and with it an increased shortage of internally produced dairy heifers in these farming systems. Consequently, unless alternative supplies become available, the cost of heifer replacements would rise and fewer farmers would be able to afford a dairy heifer for replacement, foundation stock or to expand an existing herd. Insufficient availability of heifers would be particularly marked in the high intensive farming systems where the proportion of households practising zero-grazing is already about two-thirds of all the dairy households (Table 1).

Of the interventions tested for increasing the number of dairy heifers the most promising was in the free-grazing farms through a decrease in cow mortality; followed by a reduction in the same system of the proportion of heifers sold during the rearing period. In the semi-zero- and zero-grazing farms the low impacts of decreased cow mortality and proportion of heifers sold is a consequence of a general high animal turnover in all age classes. To increase the production of dairy heifers in semi-zero- and zero-grazing farms requires a concomitantly improved calving rate and decreased voluntary exit of potential replacement heifers. This may be difficult for households without improved access to affordable credit given that 60 to 85% of the voluntary exits of female animals are for meeting immediate cash needs of the household (Bebe et al 2003a).

The major causes of cattle mortality on smallholder farms in the Kenya highlands are tick-borne diseases and parasitic worm infestations and their interactions with inadequate quantity and quality of feeding (Gitau et al 1997). Despite the adoption of tick control practices by smallholders, losses attributable to tick-borne diseases remain high, regardless of level of intensification in the system (Bebe et al 2003a). Animal health practices, such as acaricide application, are implemented inconsistently because of limited cash flow (Batz et al 1999). Interventions to lower cow mortality rates will require that smallholders have improved access to animal feeds and health services and adequate incentives to adopt these practices and technologies.

Compared to herds on free-grazing farms, those on semi-zero- and zero-grazing farms have lower calving rates and higher voluntary exits of cows and heifers (Table 2) such that a decrease in cow mortality rate alone only marginally improved the number of heifers available for replacement or herd expansion in these farming systems. The high proportions of voluntary exits of heifers are likely the result of decisions by smallholders to reduce competition for the limited feed resources in order to target feeds to cows for milk production (Bebe et al 2003b). On average, free-grazing farms are larger in area and their animals graze on the farmers' own land or on communal lands. Some farms keep a bull for mating. In contrast semi-zero or zero-grazing farms are generally smaller in area, they maintain smaller herds with a higher proportion of cows, which they feed partly on purchased fodder and concentrates.

In a study of herd dynamics of the same population, Bebe et al (2003a) showed that smallholders buy most (90%) of their replacements from fellow smallholders within the farming system. Cattle movements between large-scale and smallholder farms and also between the farming systems are presently minimal. Consequently smallholders practising semi-zero- and zero-grazing face problems of obtaining a dairy heifer of the desired genotype and quality at the required time and at an acceptable price. It is difficult for these smallholders to rear their own heifers because this requires investment in feed, veterinary services, housing and labour while income is only generated later when the animal is sold or is lactating (Mourits et al 1999).

Subsidised heifer-rearing schemes and heifer-in-trust projects to support smallholders in the rearing of dairy heifers have generally proved to be unsustainable for farmers and for projects. Experiences with smallholders in Tanzania and Sri Lanka, for instance, showed that smallholders did not continue with the recommended management practices beyond the period of project support (De Jong 1996; Afifi-Affat 1998). These experiences suggest the need for increased facilities or incentives to access dairy breeding stock from outside the farming system and to reduce the reproductive wastage as more households shift from free- to semi-zero- or zero-grazing dairying. Such facilities or incentives in the form of improved rural infrastructure including water supply need be extended to semi-arid areas to support dairy production in those areas to ease pressure on the medium and high potential agro-ecological zones. In that way, intensification of smallholder dairying can be supported through greater complementarities between the small- and large-scale components of the dairy sub-sector, particularly in the supply of dairy heifers.

Conclusions and implications 

• For smallholder farming systems in the Kenya highlands, the supply of dairy heifers partially depends upon the proportion of free-grazing farms maintained in the system and the stratification within the farming systems. Decreases in the proportion of free-grazing farms leads to a scarcity of heifers, implying that adoption of dairy production by smallholders is likely to decrease in areas of high intensive farming systems where two-thirds of the farms already practice zero-grazing. In the low and medium intensive farming systems the adoption of dairying can continue at the present rates.

• These farming systems respectively corresponded to areas of low and medium market access, and will need improved marketing infrastructures and institutional policies to support marketed dairy production. A rational policy is therefore to promote intensification of smallholder dairying with greater complementarities between the small- and large-scale components of the dairy sub-sector.

Acknowledgements 

The authors acknowledge support for this study from The Netherlands Foundation for the Advancement of Tropical Research-WOTRO, the Smallholder Dairy (R and D) Project (SDP) of the Kenya Ministry of Agriculture and Rural Development, the Kenya Agricultural Research Institute (KARI) and the International Livestock Research Institute (ILRI), and the UK Department for International Development (DFID).

References 

Afifi-Affat K A 1998 Heifer-in-Trust: a model for sustainable livestock development? World Animal Review 91: 13-20 http://www.fao.org/docrep/W9980T/w9980T03.htm

Batz F J, Peters K J and Janssen W 1999 The influence of technology characteristics on the rate and speed of adoption. Agricultural Systems 21: 121-130

Bebe B O, Udo HM J and Thorpe W 2002 Development of smallholder dairy systems in the Kenya highlands. Outlook on Agriculture 31:113-120

Bebe B O, Udo HM J, Rowlands G J and Thorpe W 2003a Smallholder dairy systems in the Kenya highlands: cattle population dynamics under increasing intensification. Livestock Production Science 82: 211-221

Bebe B O, Udo HM J, Rowlands G J and Thorpe W 2003b Smallholder dairy systems in the Kenya highlands: breed preferences and breeding practices. Livestock Production Science 82: 117-127

C.B.S 2000 The 1999 population and housing census Volume I. Population distribution by administrative area and urban centres. Central Bureau of Statistics, Ministry of Finance and Planning, Nairobi, Kenya

Caswell H 1989 Matrix population models. Construction, Analysis and interpretation. Sinauer Associates, Sunderland, MA

De Jong R 1996 Dairy stock development and milk production with smallholders. PhD Thesis, Department of Animal Production Systems, Wageningen Agricultural University, The Netherlands. 303 pp

Delgado C, Rosegrant M, Steinfeld H, Ehui S and Courbois C 2001 Livestock to 2020: the next food revolution. Outlook on Agriculture 30, 27-29  (see also: http://www.fao.org/AG/AGAINFO/resources/documents/lvst2020/20201.pdf)

Gitau G K, Perry B D, Katende J M, McDermott J J, Morzaria S P and Young A S 1997 The prevalence of serum antibodies to tick-borne infections in cattle in smallholder dairy farms in Murang’a district, Kenya: a cross-sectional study. Preventive Veterinary Medicine, 30: 95-107

Houterman J F, Lutalala E M, de Wolf J T and Dummer R M 1993 Quantitative aspects of heifer-in-trust credit schemes: the Kagera experience. 20:168-188 In: Proceedings 20^th Anniversary International Conference on ‘‘Current trends in tropical livestock production’’. Tanzania Society of Animal Production

Jaetzold R and Schmidt H  1983 Farm Management Handbook of Kenya Vol.II. Part B, Central Kenya, pp 510-620. Ministry of Agriculture, Nairobi, Kenya.

Lanyasunya T P, Wekesa F W, de Jong R, Udo H, Mukisira E A and Ole Sinkeet N S 1999 Effects of a calf rearing package introduced to smallholder dairy farms in Bahati division, Nakuru district, Kenya. pp. 450- 457  In: Proc. 6^th Biennial KARI Scientific Conf. held on 9-13 November 1998, Nairobi, Kenya.

Lesnoff M 1999 Dynamics of a sheep population in a Sahelian area (Ndiagne district in Senegal): a periodic matrix model. Agricultural Systems 61: 207-221

MoA 1998 Kenya Dairy Development Policy: Towards the development of a sustainable dairy industry, Ministry of Agriculture, Nairobi, Kenya

Mourits M C M, Huirne R B M, Dijkhuizen A A, Kristensen A R and Galligan D T  1999 Economic optimisation of dairy heifer management decisions. Agricultural Systems 61: 17-31

Paris T R 2002 Crop-animal systems in Asia: socio-economic benefits and impacts on rural livelihoods. Agricultural Systems 71: 147-168

Shaw A P M and Hoste C H 1987 Trypanotolerant cattle and livestock development in West and Central Africa. Vol. I. The international supply and demand for breeding stock. FAO Animal Production and Health Paper 67/1, Rome. 79 pp

Staal S J, Owango M, Muruiki H, Kenyanjui M, Lukuyu B, Njoroge L, Njubi D, Baltenweck I, Musembi F, Bwana O, Muruiki K, Gichungu G, Omore A and Thorpe W  2001  Dairy Systems Characterisation of Greater Nairobi Milk Shed. SDP (Smallholder Dairy (R and D) Project) Research Report, Ministry of Agriculture, Kenya Agricultural Research Institute and International Livestock Research Institute, Nairobi, Kenya. 73 pp

Upton M 1989 Livestock productivity assessment and herd growth models, Agricultural Systems 29: 149-164

Wakhungu J W and Baptist R 1992 Reproductive wastage and mortality components in dairying: an impact analysis method for sustained production systems. Journal of Zimbabwe Society of Animal Production 176: 45-50

Received 7 June 2007; Accepted 10 November 2007; Published 1 February 2008

Go to top