This thesis examines some aspects of the economics of tobacco control in South Africa, Uganda and
Zambia. The first part of the thesis examines whether tobacco expenditure displaces (or “crowds out”)
expenditure on other goods and services within Zambian households. In so doing, I make two
contributions to the literature. Firstly, I use expenditure data from a low-income sub-Saharan African
country where most households are poor. Secondly, I use the standard instrumental variable used in
the literature, the adult sex ratio, to instrument for the tobacco smoking status of Zambian
households. But unlike previous studies, I relax the strict exclusion restriction and allow for the adult
sex ratio to be correlated with the error term. That is, I allow the instrumental variable to be imperfect.
I consider the relaxation of the exclusion restriction to be reasonable given that the adult sex ratio is
just as likely to influence tobacco expenditure as it is to influence expenditure on other goods and
services. Even after relaxing the exclusion restriction, I, however, confirm many findings in the
literature. For instance, I find that smoking households allocate less expenditure towards food,
schooling, clothing, water, electricity, transportation, equipment maintenance and remittances. In
addition, the crowding out patterns I uncover are in some ways related to the geographical location
of households which in turn is related to socioeconomic status in Zambia. In sum, the results in this
part of the thesis show that a broader accounting of tobacco’s costs in Zambia should include other
costs over and above mortality and morbidity considerations.
We know from several studies that tax and price measures are the single most effective policy tool for
reducing tobacco consumption. However, most of this evidence is based on studies conducted in
developed countries with very few published studies on African countries. The second part of my
thesis, therefore, contributes to the recent literature that uses expenditure data to estimate price and
expenditure elasticities of demand for tobacco products in Low- and Middle-Income countries. I use
expenditure data from Uganda and exploit the fact that prices of cigarettes vary across geographical
space. I also adjust my demand elasticity estimates for measurement error and quality heterogeneity.
I find price and expenditure elasticities that are in line with international evidence. For instance, I find
that cigarette demand is expected to decline by between 3% and 4%, at the very least, for every 10%
2
increase in cigarette prices. The authorities in Uganda can, therefore, reduce cigarette consumption
by increasing excise taxes on cigarettes without reducing tax revenues.
The third and final part of my thesis evaluates the impact on per capita cigarette consumption of South
Africa’s consistent excise tax increases that began in 1994. The tax rises have overtime translated into
large increases in the inflation-adjusted price of cigarettes. For instance, the average real price per
pack increased by 110% between 1994 and 2004. The main challenge in conducting policy evaluations
is that of creating a credible counterfactual. That is, we want to know what would have happened to
per capita cigarette consumption in South Africa if the excise tax increases had not occurred. This is
particularly important in the case of South Africa because per capita cigarette consumption had
already started declining by the time the tax rises started. I, therefore, use a transparent and datadriven
technique, the Synthetic Control method, to create a credible counterfactual of South Africa’s
cigarette consumption after 1994. The counterfactual is constructed as a linear combination of the
per capita cigarette consumption of countries that are similar to South Africa but did not engage in
large-scale tobacco control efforts over the period 1994 to 2004. I find that per capita cigarette
consumption would not have continued declining in the absence of the consistent tax rises that began
in 1994. Specifically, I find that by 2004, per capita cigarette consumption was 36% lower than it would
have been had the tax increases not occurred. This result is robust to several falsification (or placebo)
exercises. Based on these results, I conclude that countries in Africa can achieve substantial reductions
in cigarette consumption and prevent uptake from new smokers by consistently increasing excise
taxes in the manner of South Africa.