The motivation for this thesis was to interrogate the common view on rural development
in Uganda, which assumes that the rural sector is driven almost entirely by agriculture,
suggesting that rural households depend on production of food and export crops for their
livelihood. However, there is growing evidence that rural households in Uganda and in
many developing countries are involved in non-farm employment as a way of
supplementing income from agriculture hence diversifying their income sources and
improving their income levels. Rural development policies in Uganda need to be based
on a good understanding of the determinants of participation in farm employment, wage
employment and non-farm self-employment, and on the levels of income from these
activities. Such an exploration requires an understanding of the factors determining rural
individuals' participation in the different activities. It also requires an analysis of the
contribution of income from these activities to total household income, and the
determination of whether an activity increases or reduces income inequality. This is the
contribution of this thesis. This analysis is done using the 1999/2000 National Household
Survey data for Uganda collected by the Uganda Bureau of Statistics. The data are based
on a two-stage and in some cases, three-stage sampling procedure. With an exception of
the Tobit and the censored least absolute deviation that do not work with sample survey
commands. the rest of the empirical analysis and descriptive results take into account the
design effects of the survey. Specifically, they consider weights and all standard errors
are adjusted for clustering and stratification.
A non-separable (non-recursive) model was developed to test the partICIpation of
individuals in farm employment, wage employment and non-farm self-employment as a
function of individual characteristics, household characteristics, community
characteristics and location dummies. This analysis is done at individual level. The
findings indicate that age, gender, education, shadow wages, access to electricity,
distance to the nearest headquarters and regional location are key in determining
individual participation in different employment activities. Individuals in rural Uganda
are more likely to participate in wage and self-employment in the early stages of life and
resort to farming, as they grow old. From the gender perspective, men are more likely to
participate in both wage employment and self-employment than women. This is
consistent with the patriarchal nature of Ugandan society, which tends to confine women
to farming and household chores. The results suggest that individual family members
respond significantly to changes in the household's economic opportunities. A higher
shadow wage for males reduces participation in non-farm wage employment and selfemployment
implying that more male labour is supplied to farming activities as the
opportunity cost of farming increases. Higher shadow wages for females increase
participation in non-farm wage employment and self-employment activities leading to a
backward bending female labour supply in farming. Education is a pathway out of the
low-paying farm activities to better paying wage and self-employment activities. Access
to electricity encourages participation in both wage employment and self-employment.
The results further show that the further away the individual is from the district
headquarters (town), the less the likelihood of participating in wage employment.
Xll
University of Cape Town
The inverse of the Herfindahl index is used to measure overall income diversification.
The ordinary least squares estimation was used to tind the determinants of overall income
diversification. Tobit estimation in comparison with the censored least absolute deviation
(CLAD) was used to estimate the determinants of the shares of different income sources.
The shares of four income sources namely farm, wage, self-employment and non-labour
income in total income were considered in the Tobit and the censored least absolute
deviation (CLAD) estimations against household, community and location variables.
Estimation of the models using the two approaches reveals that CLAD approach performs
better against the Tobit model for two income sources, farm share and non-labour
income. The results show that despite the fact that farming is the major income source for
rural households, non-farm activities playa significant role in rural Uganda. Age, sex,
marital status of the household head, household size, education level of adult household
members, ownership of assets, access to public services (electricity and markets) and
geographical location influence income diversification and the income shares from
different sources.
The Gini index and coefficient of variation descriptive approaches to income inequality
are used to determine the contribution of different income sources to income inequality
and whether an income source is associated with an increase or a reduction in income
inequality in rural Uganda. The use of the two decomposition methods was to compare
results given that sometimes, different income inequality decomposition approaches give
different results. The relative concentration coefficients using both approaches show that
in rural Uganda, wage employment income and self-employment income are positively
associated with income inequality. This however does not mean causality. By contrast,
both decompositions agree that farm income and non-labour income are negatively
associated with income inequality in rural Uganda.
From a policy perspective, strategies aimed at increasing incomes of the rural households
in Uganda should consider that income generated from wage employment and selfemployment
is more unequally distributed in favour of the richer households given that
they are the ones already participating. This could explain the worsening income
distribution reported in rural Uganda over time in spite of increasing income levels. In
addition, a rural development policy that seeks to promote not only formal education, but
also sensitize people to the benefits of non-farm employment activities and develop
institutional resources should be emphasized. To address issues of participation in various
activities, policies which promote both public and private investment in infrastructure
especially roads, telecommunications and rural electrification are required. Programs to
support rural women must give greater attention to facilitating their access to wageearning
job markets in agro-industry, trade and other enterprises. In terms of income
diversification, the government should aim at improving the asset endowments of the
poor. Government should create an environment with relatively equitable distribution of
income, well functioning factor markets and a strong emphasis on educational expansion
and improvement.