Food Poverty Profile Applied to Kenyan Smallholders

that can be used to test conflicting hypotheses. This paper analyzes data from the First Kenyan Integrated Rural Survey (IRS-I) household expenditure data collected by the Central Bureau of Statistics to estimate the magnitude, distribution of, and factors associated with food poverty among Kenyan small farmers.' The analysis uses a new methodology with three improvements over previous empirical studies: (1) food poverty is estimated using an additively decomposable measure that allows poverty to be distributed among population subgroups, (2) the food poverty line is based on observed dietary preferences in response to regional food prices, and (3) the food poverty measure is adjusted to use regional food poverty lines without sacrificing its theoretical properties or decomposability. This paper consists of two main parts. Section II reviews the methodology underlying the setting of the regional food poverty line. On the basis of a detailed integrated rural survey of Kenyan smallholders, the cost of food consumption corresponding to the recommended dietary allowance (RDA) of calories is computed. Since this food poverty line was calculated on the basis of the observed diet consumed, it presumably reflects the prevailing regional food preferences and prices faced by households consuming around the calorie RDA level. The next part of Section II presents a food poverty measure that-in addition to satisfying the major fundamental properties of such measuresis decomposable and allows regional food poverty deprivation to be estimated.