Various previous studies have examined the adoption of dairy technologies and their impacts on smallholders in Kenya. These studies are summarised in Table A1. Although the objectives of the studies differ, many of them share common approaches to assessment of impacts. Most of the studies relied on single-visit surveys with random (or semi-random) samples of households involved in dairy production. That is, only households with dairy cattle were included in the sample, and thus there typically was no 'control' group of households with similar characteristics but not involved in dairy production. The survey methods often relied on comparisons of the situation 'before and after' adoption of the dairying, based on recall of past events by households. Such studies provide indicators of how households perceive the impacts of dairying, but are probably less insightful than studies that use longitudinal monitoring techniques or include a control group in the study design.
Table A1. Previous studies of adoption and impact of smallholder dairying in Kenya.
The technology adoption and diffusion studies emphasise the high variation in adoption rates and factors apparently influencing the adoption of dairy-related technologies and practices. Whereas Irungu et al (1998) reported that 70% of farmers surveyed in Kiambu District planted Napier grass, adoption of Napier grass ranged from 22% to 76% of households in six other districts (Metz et al 1995). The technologies were most often adopted individually, even though the National Dairy Development Project (NDDP) promoted the related technologies as an integrated package (Metz et al 1995). Even when the technologies promoted by NDDP were adopted, some farmers abandoned them after a few years. Maarse (1997) asserted that in Coast Province, the number of farms with pure stands of Napier grass and the amount of Napier grass planted per cow declined by nearly 50% from 1988 to 1993. The number of 'dormant' NDDP farmers (i.e. farmers who did not own a cow, have milking or feeding facilities, produce fodder on farm or practice zero grazing at least 25% of the time, despite previously being registered in NDDP) also varied by district. This indicated that the accessibility and appropriateness of the NDDP 'zero grazing' recommendations vary depending on local conditions. Dormancy was highest in Coast Province compared with other regions of Kenya. Nearly 9% of registered farmers at the coast were dormant in 1993, mostly due to loss of animals due to disease or the 'lack of management input' (Metz 1993).
A potentially large number of impacts could be examined after households decide to undertake dairy production. Impacts assessed in the different studies cited above focus on who performs the tasks associated with more intensive dairying and the perceived benefits to the household. The studies confirm the importance of female household members in dairy production and marketing. Mullins et al (1996) found that in Kilifi District women are frequently responsible for dairy-related tasks other than herding and spraying for ticks. Women in the same district were 'involved' in 30% of the dairy-related tasks performed, more than children (26%) or hired labour (19%) (Price Waterhouse 1990). The MoALDM/NDDP studies (Mugo 1994) found that women contributed between 25% and 41% of the 'relative labour contribution' in the 'main dairy management areas'.
The impacts of dairy adoption, like the prevalence of adoption itself, varied by location in Kenya. The impact of dairying on household labour requirements, perceived responsibilities and the health of household members were outcomes that differed most by district. The percentage of households reporting that the adoption of dairy increased the amount of time devoted to farm work varied from 25% to 75%. (In Kilifi District, however, time devoted to 'other family activities' was not affected by the adoption of dairy, although the amount of work increased; Price Waterhouse 1990.) Households reporting 'more responsibilities' varied from 0% to 72% of respondents in the six districts where the MoALDM/NDDP study was carried out. Between 10% and 78% of households reported that adoption of dairying had improved the health status of the family (Mugo 1994; Mullins et al 1996). Thus, some key effects of adopting dairying appear to be highly variable; this suggests that additional analyses may help to elucidate the underlying reasons for this variability.
The promotion of dairy production is often justified by the assumption that adopting households will consume more milk, but the results of the studies suggest this outcome is not universal. Less than 10% of the households in the MoALDM/NDDP studies for Migori and Nandi districts indicated that more milk was available for consumption. None of the female respondents in those studies indicated that 'more milk for home consumption' was an 'effect of zero grazing', whereas 8% and 17% of male respondents in those districts indicated that milk consumption by the household had increased after adoption. In contrast, Price Waterhouse (1990) and Mullins et al (1996) both found that more than 90% of households reported greater milk consumption after adoption. Launonon et al (1985) reported that over 70% of households in a Meru District survey reported increased milk consumption after the adoption of dairying.
Household perceptions of the impacts of dairying on income and financial status varied less than other impacts assessed by the studies. Fifty-five to ninety per cent of the households reported more income after adoption, and 78% to 100% of households reported improved 'financial status' as a result of dairying (Mugo 1994). Mullins et al (1996) reported that 97% of households said that income had increased after adoption of grade and crossbred animals.
Another hypothesis concerning adoption of more intensive dairying is that it generates employment, because more labour is required to care for Napier grass and grade or crossbred animals. In the MoALDM/NDDP studies, respondents reported that hired labour provided between 28% and 39% of the 'relative labour contribution' for tasks related to zero grazing. Often, hired labourers performed much of the work of weeding Napier grass plots and cutting grass for confined cattle. The results for Kilifi District suggest lower levels of employment generation, but vary depending on the study. Price Waterhouse (1990) found that only 12% of households hired more labour after adoption of dairy production—yet hired labour accounted for nearly one-fifth of dairy-related tasks. This suggests that the households adopting dairying may have already employed hired labour, and some of the additional work was taken up by existing labourers rather than new hires. The study did not examine the total payments to labourers or the amount of time they worked, so the impact on total payments to hired labour is unknown. In contrast, Leegwater et al (1991) found that about half of the NDDP farmers in Kilifi District employed labourers, particularly when the farm owner had off-farm employment. The extent to which dairy cattle per se are responsible for the increase in hired labour was not examined in detail in either of the two studies.
Leegwater et al (1991) was one of the only studies to explore impacts of dairy adoption through examination of adopters and non-adopters. The study examined five groups: NDDP dairy producers, 'independent' dairy producers, extensive livestock producers, households that purchased dairy products from NDDP farmers, and the general rural population in Kilifi District. This study also examined factors in greater detail and more quantitative measurement than most of the other studies. Leegwater et al (1991) found that NDDP farms produced more milk than their 'independent' dairy counterparts, consumed more milk than the other four groups, and purchased a smaller percentage of calories consumed by the household. NDDP households also engaged more frequently in off-farm employment, earned higher total incomes, and enjoyed better nutritional status for pre-school age children.
A limitation of the Leegwater et al (1991) study is that the results rely on tabular summaries for the five groups, and thus do not control for factors other than dairy production, such as land availability or other income, that will influence the reported outcomes. The authors assert, for example, that off-farm income allowed the farmers in NDDP to afford the investments required by the project, yet only one-quarter of the sampled NDDP farmers are reported to have off-farm income. Similarly, the study concludes that the nutritional outcomes, while better for dairy producing and consuming households, are due to 'better child care in general' and thus cannot be attributed specifically to dairy production.
