posted on 2017-02-17, 03:58authored byRoos, Elizabeth Louisa
The thesis is concerned with the construction and application of SAGE-H, a large-scale dynamic computable general equilibrium (CGE) model of the South African economy with a particular focus on HIV/AIDS. A key feature of the SAGE-H model is its detailed treatment of the labour market, with labour force participants distinguished by their labour market activity, age, gender, race and HIV status.
The core theoretical structure of the SAGE-H model is based on the MONASH model (Dixon & Rimmer, 2002). SAGE-H describes the production of 28 commodities by 28 industries. It also defines factors of production, various technologies and preferences, taxes and margins, detailed descriptions of government accounts, the balance of payments accounts and net foreign liabilities. The model recognises a number of dynamic adjustment mechanisms, in particular, capital accumulation, net foreign liability accumulation and sticky wage adjustment. SAGE-H captures the structure and characteristics of the South African economy via the calibration of the model to South African Supply-Use Tables for 2002 (StatsSA, 2006c) and relevant South African parameter and elasticity values.
The detailed modelling of the labour market in SAGE-H is based on that developed by Dixon and Rimmer (2009) to analyse the effects of immigration on the US economy. I adapt their theory of labour supply to enable it to treat the peculiar features of the South African labour market, particularly with respect to the influence of progressive stages of HIV on labour supply. The Dixon approach is to classify the working age population (WAP) at the start of each year into a number of labour market categories based upon their past labour market activities and end-of-year probabilities for transitioning between categories. I adopt the labour market function categories of employment (by occupation), unemployment and “permanently departed from the labour force” and add the HIV relevant additional classifications of age, gender, race and HIV status and stage. At the start of year t, people in each category decide to offer their labour to various labour market activities based upon relative wages and personal preferences/attributes. They make this decision by solving an optimisation problem. One invention of SAGE-H is that the supply of labour depends on the labour force participant’s HIV status and stage. For instance, a person who is HIV negative supplies labour more strongly to employment activities than a person who is HIV positive. Equilibrium in the labour market by occupation is determined by labour demand by all industries and labour supply by all categories.
I carry out two simulations with the SAGE-H model to analyse the economy wide impact of an increase in condom use over the period 2002 to 2045. The first simulation is the baseline or business-as-usual simulation. This simulation models the growth of the South African economy and the HIV epidemic over time in the absence of the policy change under consideration. The second simulation conducted with the SAGE-H model is the policy simulation. This simulation generates a second forecast that incorporates all the exogenous features of the baseline forecast, but now also includes policy-related shocks reflecting the details of the policy under consideration. The impacts of a policy are typically reported in terms of percentage deviations away from the baseline simulation.
In the absence of a cure for HIV/AIDS, preventing the transmission of HIV is the only known way of curbing the spread of the virus. Prevention is also important as it affects adults’ future health prospects. In this thesis, the policy under consideration is an increase in the proportion of sex acts protected by condom use via a publicly-financed program of condom distribution. More specifically, I analyse the impact of an increase in condom use on (1) the number of new HIV cases, (2) labour market adjustment via changes in the level of employment and the real wage and (3) the economy-wide impact of changes in the labour market. To get a better understanding of the impact of an increase in the proportion of sex acts protected from the HIV virus due to an increase in condom use, I first use the Actuarial Society of South Africa epidemiological model (the ASSA2003 model) (Actuarial Society of South Africa, 2005). This model suggests that an increase in the proportion of protected sex acts leads to a fall in the annual number of new HIV infections. In terms of the SAGE-H model, the prevention policy can be viewed as generating a decline in transition probabilities from HIV negative to HIV Stage 1, i.e. a fall in HIV incidence rates. In addition, I model government spending on the program via an increase in government demand which is financed by an equivalent decrease in private consumption in general. Over the simulation period, the policy leads to approximately 200,000 fewer HIV infections relative to baseline. Due to the staged transition path through the HIV stages, the results suggest that the number of people in Stage 1 falls prior to falls in Stages 2, 3 and 4. In the long run the total number of HIV positive adults is 0.82 per cent lower than baseline and the number of HIV negative adults is 0.46 per cent higher than baseline. Based on the underlying assumptions incorporated in the labour-market module, fewer people becoming HIV positive implies that labour market attachment increases i.e. labour offers to all employment activities increase. In the long run employment is 0.23 per cent higher than the baseline value and the real wage is 0.18 lower than the baseline value. The labour offers to employment activities depend largely on the age, gender and race-specific composition of these occupations. While all age, gender and race groups benefit from the prevention policy, the prevention policy has its largest initial impact on labour supply of young African adults. Hence, employment activities with a higher representation of young African adults in their labour force benefit the most from the prevention policy. In the long run the policy also generates an increase in capital stock and together with the increase in employment GDP is 0.17 per cent higher than in the baseline. Real private consumption, net of the cost of financing the program, is 0.15 percent higher than baseline in the long run.
The framework presented in the thesis opens up a number of additional avenues for future research into HIV prevention programmes and treatment options. Two promising future avenues of research with the model are analysis of the labour market consequences of anti-retroviral treatment programmes, and detailed modelling of demand for and supply of finely-defined health services.