Kenya faces a high and rising burden of disease with chronic illness becoming an important
contributor to the disease burden. Health as a capital stock can be expected to affect positively
individual’s labor supply and labor productivity, thereby generating economic benefits. High
disease burden would erode such benefits yet like many SSA countries, Kenya is relying on its
labor force to achieve the projected economic growth. Despite the potential negative labor
market effect of illness and the rising disease burden in Kenya, there is limited empirical
evidence on the relationship between self reported illness, labor supply and earnings. The
purpose of this thesis was to investigate the relationship between illness and three labor market
outcomes: labor force participation, employment status and earnings. The data are drawn from
a nationally representative household survey conducted by the Government of Kenya.
In chapter one, the objective was to examine the effect of illness on individual labor force
participation by gender. A standard probit model was estimated. To control for potential
endogeneity of the health variable and unobserved individual heterogeneity, control function
approach was employed. The results do not reveal evidence on endogeneity and unobserved
heterogeneity bias. Standard probit estimates indicate that acute and chronic illnesses reduce the
likelihood of labor force participation. The negative effect of chronic illness is larger than that
of acute illness. Both illnesses reduce the likelihood of labor force participation among women
by a larger magnitude than among men. The findings imply that illness is a major constraint to
labor force participation in Kenya. Effective policies and interventions to reduce prevalence of
chronic and acute illness would bring more Kenyans into the labor force. Moreover such
policies and interventions should be targeted to women.
Chapter two examined the effect of chronic and acute illness on individual’s employment status.
Multinomial probit model was estimated because the IIA assumption underlying the
multinomial logit model did not hold in the data used in this thesis. The employment states
considered were: wage employment, agricultural self-employment, non-agricultural selfemployment
and not working. The results indicated that chronically ill women were less likely
than non ill women to be in wage employment and in agricultural self-employment. For men,
xi
having chronic illness did not significantly influence employment status. The results also
indicated that having an acute illness did not significantly influence employment status of either
men or women. By affecting individual’s choice of type of employment, illness affects labor
allocation and reallocation with implications for economic transformation and growth. The
results imply that policies and interventions to reduce incidence of chronic illness would
increase wage employment among women. This is important since relative to other forms of
employment, wage employment is associated with higher earnings and benefits (such as job
security and health insurance). Such policies and interventions would also increase participation
in agricultural employment.
The objective in chapter three was to investigate the relationship between illness and earnings in
wage employment, agricultural self-employment and non-agricultural self-employment. A
control function approach was used to account for potential endogeneity and unobserved
individual heterogeneity. Sample selection bias was controlled for using a two step approach
suggested by Bourguignon, Fournier and Gurgand (here after BFG). The results reveal evidence
of the three econometric problems, justifying use of control function approach and controlling
for sample selection bias. Full sample results indicated that both acute and chronic illnesses
have negative and significant effect on wage employment earnings. In contrast, both illnesses
do not significantly affect earnings in self employment. When the analysis was done by gender,
only chronic illness significantly affects wage employment earnings and the effect was larger
for men than for women. Public policies and interventions that effectively reduce the incidence
of chronic and acute illness would increase wage incomes and help in poverty reduction