Factors influencing farm income in marginal areas of the lower cotton zone in Eastern Kenya (Machakos, Kitui and Mbere Division – Embu
In view of the emphasis given to improving farm productivity in the marginal areas, the main objective of the study was to determine those factors that influence farm incomes in these areas. other minor objectives were to determine the relative importance of these factors, and to establish whether their relationship with farm productivity depended on the method used to derive farm income. The literature review indicated that various factors have been,shown to influence farm income in the semi-arid areas as well as in other environments. Those that are considered in this study are the following; purchased farm inputs, crop area, labour,off farm income, sex of the farm operator, assets, family size and structure, the natural environment as defined by district, and the ease of transportation as specified in terms of the distance from nearest sizeable market to the holding. The source of data was the Integrated Rural Survey 1 (IRS 1) of 1974/75 carried out by the Ministry of Finance and Planning. The analysis involved calculations of percentages, frequency distributions, correlation and regression coefficients and differences between the means of subsamples. Three methods were used to calculate gross farm income and net farm income by taking into account that (a) livestock valuation change may be included or excluded in the calculation of farm income, and (b) as the data was collected during a drought period, an attempt can be made to offset the drought bias. Gross crop output was also considered as an independent variable. It was found that although purchased farm input (fertilizer, seed, spray and machinery expenses) was the most important single factor influencing farm income and crop output, there was limited dependence on fertilizer, seed and sprays. Machinery expenses was the only specific purchased crop i~put significant in explaining the variation in crop output. The importance of this variable seems to stress the impact late planting and weeding have on crop output in marginal areas. Crop area was shown to be positively related to farm income. However this was not indicated by the regression equation that had gross crop output as the dependent variable. Further analysis showed that different groups of farmers operating the same size of crop area had significantly different farm income. This was caused by significant differences in the amount of other factors used especially purchased crop inputs. Both family and hired labour were significant in influencing farm income and crop output. Nevertheless, significant differences in gross crop output was detected between two groups of farmers who had comparable crop labour. Those that had a higher proportion of family labour achieved higher gross crop output than those who operated using a bigger amount of hired labour. Although further analysis in this connection was not possible in the study, the observation seems to imply that family labour contributes more to the success of small-scale farming than the same amount of hired labour. With respect to off-farm income, the correlation and regression coefficients denoted that the relationship between this factor and farm income is small but negative. This finding supports observations made in the study area while it is contrary to other suggestions made on the basis of national sample of IRS 1. The farmers who have high off-farm income, hired significantly more labour but they did not purchase signifi-. cantly more inputs. Off-farm income seems to offer an alternative to farming as the major source of livelihood in addition to competing for labour. Farms operated by women were found to have significantly lower farm income and crop output than those managed by men. This is contrary to what was observed in high potential areas and confirms previous findings from the study area. The analysis implied that this is likely to be the result of the following factors; (a) women managers are associated with much higher off-farm income than men operators, (b) they use lower purchased farm inputs especially those that involve machinery expenses, and (c) they operate with lower family labour. Assets were assumed to be the indicators of the "rich" farmers who could afford to purchase inputs and hire labour. The correlation and regression coefficients with respect to this factor and gross farm income indicated a positive association. Family size and structure, as measured in terms of consumer equivalents, was found to be positively related to farm income. Kitui District was shown to have lower gross crop output than Machakos District, which has higher and more reliable rainfall in general. The variable specifying the distance from the holding to the nearest sizeable market was not significant in any of the regression equations. In general there was little variation in the relationship between the above factors and farm income, whether farm income was calculated including or excluding livestock valuation change, or compensating for the bias caused by the drought year. The gross farm income showed stronger relationship with these factors than net farm income.