Research
Activity I: Impact Assessment of the KDPG at the Household

Problem Statement and Approach

The impact on the welfare of households, result of the introduction of the KDPG as an economic enterprise, and asset building important in smoothing income shocks are central research questions of the social sciences project. The effects of the KDPG on farm can be measured through: 1) the output generated, 2) changes in the intrahousehold allocation of resources, and income related to the KDPG that may change the bargaining position of members of the household, and 3) asset building that contributes to economic security, especially food security. Expenditure patterns of men and women have been shown in several studies to differ, especially in Kenya. If KDPGs become the income domain of female heads of household the probabilities of milk and cash from this activity being spent in consumption items for the family will increase. Gender along with wealth and ethnicity must be considered in the analysis. Chaiken and Conelly, in western Kenya found that income and poverty had stronger effects than gender in deriving benefits from the KDPG. Boserup, in her analysis of the changing roles of women with development, states that male migration, in colonial times, resulted in women's greater control of production activities in rural areas, especially food production.

The assessment of the impact of the KDPG focuses on economic viability of the KDPG measuring income generation, cash and in- kind, of the KDPG enterprise, and the diversity of economic activities carried out by the household producers on and off the farm; gender analysis through intrahousehold allocation of resources focusing on labor and income, to determine domains of the economic enterprises; and the flexibility of this technological package as it integrates to the household economy.

Progress

1. Economic Viability of the KDPG

The Kenya Dual Purpose Goat (KDPG) and the accompanying management package were introduced into a number of small holdings in the semi-arid midlands and in the humid Coastal lowlands of Kenya during the short rains season of 1993. These innovations were intended to help improve the welfare of smallholder households through increased income (cash and in kind) and nutrition. Introduction of livestock based innovations usually has significant impacts elsewhere in the farming systems.

A baseline survey of the main characteristics of the study sites was undertaken in 1993, preceding placement of the KDPG innovation. A monitoring survey designed to assess the extent of impacts was carried out in the five clusters between 1994 and 1996. Highlights of on-going research are presented in this report.

A combination of case-studies and questionnaires of all farmers was adopted for this study. Four instruments to capture farm and household information were designed, pretested and applied. The timing and frequency at which these instruments were applied were matched to the production and household activity calendars in the two sites. To keep track of the production, consumption and sales of livestock and products as the KDPG integrates into the farming systems, monthly farm visits were conducted.

Rapid Rural Appraisals (RRAs), one in each of the five clusters during the short rains season of 1993 preceded placement of the KDPGs. Base-line information about the small holdings in the regions was used as bench-mark for impact assessment. Each of the two clusters in the Machakos site (Kitanga and Kimutwa) and the two in the Coast site (Kilifi, Vuga) had 20 smallholder households that had been randomly selected from the farming community. The HPI cluster had been formed previously with the purpose of promoting the keeping of dairy cows by members. Each cluster was allocated 10 does and two bucks, five farmers received two does each. Apart from the requirement that each recipient be a member of the group, undertake to practice acceptable management, facilitate data collection and pass on the first two doelings to members on the waiting list, the individual groups were to regulate themselves. Kitanga, Kimutwa, Vuga and HPI clusters received their KDPGs during the short rains season of 1993, and the Kilifi cluster during early 1994.

Following the placement of the KDPGs on the farms, the Impact Assessment study was planned and implemented. Prior to this implementation, research themes were identified and the required data sets itemized. Finally, data capture instruments were drawn up and pre-tested.

The four instruments used to capture farm and household data were:

• I. Form M was designed to be administered monthly. Items of data recorded on this are: 1) Family labor availability (through changes in the family structure); 2) Family labor use: main occupations; 3) Household non-agricultural labor supply; 4) Wage rates; 5) Milk consumption/sales; and 6) Animal sales.
• II. Agricultural form to be applied 6 times a year, in April, July, October, November, February and March. Land use and cropping activities, access to credit, and remittances from friends and relatives were recorded. Information about production, sales and consumption of permanent crops and fodder was recorded. Difficulties in data collection stages did not allow full year coverage (e.g., monthly and agricultural). Fortunately, few data items were found so sensitive to the frequency at which data were collected.
• III. Captured on Form I were agricultural inputs (costs, source and use), mainly seeds, fertilizers and chemicals. The enumerator had ample opportunity to record pertinent observation. This instrument was designed to be administered at the beginning of each season.
• IV. Form O was designed to record the production of seasonal crops.

In both the Machakos and the Coast sites, data collection started during the short rains season of 1994/5. In the ninth month of data collection at the Katumani site, 37 out of 40 farms (92.5%) had been interviewed concerning input questionnaires and 33 out of the same 40 farms (82.5%) had been interviewed using output questionnaires. Similarly, during the same period, 154 out of the 180 farms (85.6%) had been interviewed with respect to the bi-monthly questionnaire on agricultural operations. This was quite a successful operation. However, the more regular monthly interviews were not equally successful. Indeed, of the 360 interviews to be carried out during the same period, only 148 (41%) were done. This was the year in which SR-CRSP funding was not available for the project.

In the Coast sites interviews started in August 1994 and through out serious disruptions were experienced. Whereas 61.4% of the interviews were carried out in August, only 35% of them were successfully carried out in September, 1994. The following month interviews picked up to 96.5% level but dropped to 52.6% in November. In December, no interviews were done. In January, 1995 modest activity resumed (35.1% of the interviews). The following three months (February, March and April) no interviews were done. In May about a third of the interviews were carried out (36.8%).

Over all, Katumani site did relatively better than Mtwapa site in data collection. Considering that both sites were receiving their financial support from the same source, it is evident that institutional constraints were greater in Mtwapa than in Katumani. Following consultations with the centre management in Mtwapa, arrangements were made to ensure more regular data collection in the future. It was decided that two field assistants be recruited for the highlands (Kitanga); lowlands (Kimutwa) in Machakos and one the Kilifi cluster. The field assistant who was already in place at the Coast would concentrate on the HPI and Vuga clusters in Kwale district. Due to delays in identifying a second field assistant for the Machakos site, it was decided that the one field assistant already in place would cover the two clusters. By January 1995, these measures had already been effected. However, the necessity to follow institutional regulations in channeling resources to the field assistants created some delays. The effects of institute wide shortages of personnel in the socio-economics field were also felt through the inability of collaborating scientists to allocate more time to the supervision of field work and also data management.

Nevertheless, 1995-1996 data represents a significant improvement over the previous year. There are fewer missing observations. In terms of consistency, missing observations, recording errors, etc., the performance of the Machakos field assistant appears reasonable. The Kilifi field assistant also performed satisfactorily, as did the Kwale field assistant, despite delays in dealing with logistical problems the latter was facing.

Estimation of the Impact of the KDPG on Farm Income
The KDPG, introduced to existing farming systems during the short rains season of 1993, were expected to generate in-kind or cash income through the production of milk, bucklings, and possibly manure and skins. Products were multiplied by their respective cash values in Kenya Shillings to calculate total income. Enterprise specific (variable) costs were assessed by estimating the value of inputs which were basically, housing, drugs and treatment, and in some cases water and supplementary feeding. The cost of labor was assumed to be constant and was not priced.

Seasonal Conditions and Performance of the Farm Enterprise
This section is intended to provide a description of the seasonal conditions within which production took place. This has a direct bearing upon the interpretation of production data that is presented in the following sections. Table 1.1 shows rainfall data for the period short rains 1985-long rains 1996 for NDFRC Katumani. There was almost total crop failure during the long rains 1994 season, as in the preceding two seasons. This contributed to the difficulties in interviews because of the unavailability of responsible persons present on the farms during our visits (most would be away seeking means of sustenance).

To this date we are searching for comparable data for the Coast. Interviews with farmers are yielding the following perception regarding climate and cropping seasons: the short rains 1994 were "Poor", the long rains of 1995 were "fair to good", the short rains of 1995 were "poor", and the long rains of 1996, were "poor to fair".

Performance of the Farm Enterprise
To assess impact on farm-household income, cash and in-kind income from other enterprises were measured. Four general classes of activity had been identified. These were: seasonal (mainly food) crops; perennial crops; other livestock and a variety of non-farm enterprises.

These assessments are summarized as follows:

Where Y is non-KDPG income, p_j represents the factor x and product y prices in the _jth enterprise. For expository convenience, these enterprises are grouped as follows:

To the extent that borrowing does in many instances allow smallholder households access to scarce capital resources which could supplement what can be generated on the farm, it was also considered as an inflow into the farm and credit repayments as outflows.

Among the commonly used farm inputs x_i seeds, fertilizers and crop pest and disease control chemicals. The number of miscellaneous inputs such as gunny bags can be very large, depending upon the crop enterprise in question.

Where K is 'net' income generated through operation of the KDPG enterprise and p_i represents factor x and product k prices.

I_t represents impact of the KDPG in the _tth site, cluster or farm household, as the case may be.

Although cash remittances and credit were regarded as inflows and gifts and loan repayments were considered to be outflows, these did not directly enter into the calculations of net income from farm operations.

Main Findings
The impact assessment study was designed to estimate likely impacts of the KDPG technology on household income, division of labor by gender and age groups within the smallholder families, crops and livestock interactions, and potential flexibility of the package.

We summarize the main findings relating to the contribution of the KDPG to income and nutrition in Table 1.2. In addition to these contributions, there were other benefits associated with the KDPG enterprise. These were direct consumption of the KDPG meat and production of manure which must have gone into improving the cropping system.

In the Machakos cluster, seasonal crops contributed well over half of the household income. Other important sources were livestock sales as well as remittances from those working off farm. Perennial crops and credit seem to have contributed little. The KDPG enterprise contributed 2.58 per cent to farm income in the lowlands (Kimutwa) and 4.65 percent in the highlands (Kitanga) cluster. This stands in sharp contrast with the situation at the Coast. The Kilifi cluster income generated from the KDPG represents 20 percent of the income, 10 percent in the case of Vuga and 4.6 percent in the case of the HPI/Matuga cluster. This percentages are sensitive to the size of the income, as it can be seen in table 1.2, because the net value contribution (in Kenya Shillings) does not exhibit those differences. Net contributions are higher in the case of the Kilifi cluster and this relate to milking information. Remittances were five percent of total farm value produced in 1995 by the Kitanga cluster, 4 percent for the Kimutwa cluster, 11 percent in the Kilifi cluster, 12 percent for the Vuga cluster. Absolute value of the remittances did not vary as much as the relative importance result of the greater differences in total value of production. Perennial crops clearly have a major role in their smallholder economy of the Coast. The contribution of the KDPG in the Vuga cluster is between 2 and 5 times that of the Machakos clusters.

Source: Data base monitoring 1995.

The KDPG was intended to enable smallholders for whom dairy cows were out of reach to have access to milk. Milk production was therefore an important aspect of the KDPG impact assessment. This assessment involved the estimation of the numbers of farmers who milked the KDPGs, the amount of milk produced, the amount sold, if any, and the prices obtained. The main findings are presented in Table 1.3. No less than 21 farm families, 2 in Kitanga, 4 in Kimutwa, 3 in Kilifi, 8 in Vuga, 4 in the HPI clusters had some goat milk during 1995. KDPG milk accounted for 20-42 per cent of milk produced.

Source: Monitoring Data Base Socioeconomics.

Table 1.4 highlights farmers characteristics, regarding land use, remittances and credit. Farmers at the Coast have less land in average. This, combined with the fact that Kilifi and Vuga reported no access to credit in 1995, highlights the importance of the added KDPG enterprise, in generating additional sources of income that may be consumed or invested in other activities.

Source: Land size Baseline 1993, Other Data Base Monitoring, 1995. Standard Deviations in Parenthesis

Impact on the Crop Livestock Interactions
Crops and livestock interactions in smallholder farming often involve the feeding of crop residues to livestock; use of organic fertilizer to raise yields through improved plant nutrition, and in the case of oxen, supply of draft for timely field operations. Although the number of goats was small, these interactions were observed in all the five clusters. In all the three clusters at the Coast, all farmers reported that they used maize bran to feed the KDPGs, while in the Machakos clusters, various types of crop residue were used as supplementary feed for the goats. Goats produced manure which farmers claimed was richer than that from local goats. Average production of KDPG manure was 81 and 67 kg per goat in the Kimutwa and Kitanga clusters respectively. At least seven households had some KDPG meat, three in Kilifi, one in Kitanga, one in Vuga and one in the HPI clusters.

2. Gender Analysis

Field research on three specific activities of the household impact assessment workplan were targeted from June through August. A combination of open ended interviews and a questionnaire, were completed. Qualitative and quantitative analyses was conducted. The questionnaire included three areas:

a. Gender Analysis: Gender domains with the introduction of the KDPG to understand technology domains to assist with diffusion and extension. This focused on labor and income to determine the domains of the economic activities related to the KDPG to understand the new household dynamics introduced by the new activity;

b. Adoption of technological practices to determine if there are differences in adoption of elements of the technological package, and analyze the correlation with gender, income, ethnic and religious, distance to markets, availability of inputs, and/or agroecological conditions; and

c. Credit sources available to farmers by gender. Data for the Machakos cluster has been entered and cleaned. Questionnaires for the Coast are currently being entered. This will be analyzed in the first quarter of 1997. Some frequencies from this survey have been completed and incorporated in to the "Of goats groups and gender" report.

The qualitative research took place from June through August. Results are presented in the technical report TR-MU-1 "Of goats groups and gender". It was clear in all clusters that women were the primary caretakers of the goats. In the open ended questions, the farmers were asked to explain what happened to a goat throughout the day. In most cases, it became clear that women were responsible for daily care of the goats and for its general management. Yet there were differences between the two regions. This may be directly related to access to information. In Machakos, the women attend meetings, are often members of other groups, and have time to interact and exchange information. In the Coast, men are seen as better caretakers, because they appear to be more knowledgeable of the management system. Men usually attend the meetings, while women have little time to access information on the care of the animals, because they are more likely to be working at home.

3. Flexibility of the KDPG Technological Package

The tradition of livestock is very important in the adoption of the KDPG (see table 1.5). Farmers in Kwale knew the most about the various technologies that accompanied the KDPG. In contrast the farmers of Machakos district were less able to enumerate the various technologies, and were also less likely to use them, with the exception of the sheds. On the other hand, the farmers in Machakos were more likely to use technologies learned for other purposes, for instance, nappier grass.

Farmers at the Coast, have not been pastoralists, therefore they are more likely to adopt 'the whole package'. On the other hand farmers in Machakos have a tradition of raising goats, and thus are more relaxed, adopting elements that interest them.

Source: Resource Management Labor and Gender Questionnaire, 1996, and "Of Goats Groups and Gender".

[Back to Annual Report 96 Contents]
[Continue to KDPG Activity 2]