Development Policy Review, 2003, 21 (1): 27-49 Overseas Development Institute, 2003. Published by Blackwell Publishing, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA. The Economics of HIV/AIDS: A Survey Edoardo Gaffeo ∗ This article surveys the main economic issues associated with the HIV/AIDS epidemic, paying special attention to sub-Saharan Africa. It explores the economic and behavioural determinants of HIV transmission, the microeconomics of market failures associated with high HIV prevalence, the prospects for regional development from a macroeconomic perspective and the efficient design of policies for coping with the epidemic. In line with the recent appeal by the UN Secretary General, the article argues that, without a decisive effort to halt HIV/AIDS, people living in the region are bound to experience a further fall in their standard of living in both relative and absolute terms. However, to be effective, anti-AIDS programmes must be rooted in sound economic principles. This article attempts to set out a few propositions summarising what is known and largely accepted about the microeconomics and macroeconomics of the worldwide HIV/AIDS epidemic. The purpose in doing so is to put forward a set of presumptions about what should be seen as sound ‘economically oriented’ health policies aimed at limiting damage in the immediate term, and to set the stage for achieving the main long- run target of full eradication of the disease as quickly as possible. In particular, it devotes special attention to sub-Saharan Africa, where contagion has already reached global proportions and the issue has long ceased to be an exclusive health problem and has become a serious threat to long-term social and economic development. The article is organised as follows. The next section gives a brief account of the epidemic so far, and some hints on what can reasonably be expected in the near future. Section three then deals with justifications for public intervention from an economic standpoint, followed by sections analysing the impacts of HIV/AIDS at a microeconomic and a macroeconomic level, respectively. Policy issues regarding the allocation of resources within HIV/AIDS programmes, and the development of effective vaccines, are then treated, followed by a concluding section. An overview of the epidemic The Acquired Immuno-Deficiency Syndrome (AIDS) originally appeared on a sizeable scale at the end of the 1970s, while scientific evidence on the existence of the Human Immunodeficiency Virus (HIV), a retrovirus which, once inside the bloodstream, fatally damages the immune system of the host, 1 dates back to the early 1980s. Two types of ∗ Department of Economics, University of Udine, Italy ([email protected]). 1. Thus, those infected with HIV do not die because of the virus per se, but because their immune system fails to counteract opportunistic infections, which would not otherwise have been fatal. The status of AIDS patients is diagnosed precisely when the immune system of HIV-infected people stops working properly.
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Microsoft Word - Gaffeo.doc Overseas Development Institute, 2003.
Published by Blackwell Publishing, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.
The Economics of HIV/AIDS: A Survey
Edoardo Gaffeo∗
This article surveys the main economic issues associated with the HIV/AIDS epidemic, paying special attention to sub-Saharan Africa. It explores the economic and behavioural determinants of HIV transmission, the microeconomics of market failures associated with high HIV prevalence, the prospects for regional development from a macroeconomic perspective and the efficient design of policies for coping with the epidemic. In line with the recent appeal by the UN Secretary General, the article argues that, without a decisive effort to halt HIV/AIDS, people living in the region are bound to experience a further fall in their standard of living in both relative and absolute terms. However, to be effective, anti-AIDS programmes must be rooted in sound economic principles.
This article attempts to set out a few propositions summarising what is known and largely accepted about the microeconomics and macroeconomics of the worldwide HIV/AIDS epidemic. The purpose in doing so is to put forward a set of presumptions about what should be seen as sound ‘economically oriented’ health policies aimed at limiting damage in the immediate term, and to set the stage for achieving the main long- run target of full eradication of the disease as quickly as possible. In particular, it devotes special attention to sub-Saharan Africa, where contagion has already reached global proportions and the issue has long ceased to be an exclusive health problem and has become a serious threat to long-term social and economic development.
The article is organised as follows. The next section gives a brief account of the epidemic so far, and some hints on what can reasonably be expected in the near future. Section three then deals with justifications for public intervention from an economic standpoint, followed by sections analysing the impacts of HIV/AIDS at a microeconomic and a macroeconomic level, respectively. Policy issues regarding the allocation of resources within HIV/AIDS programmes, and the development of effective vaccines, are then treated, followed by a concluding section.
An overview of the epidemic
The Acquired Immuno-Deficiency Syndrome (AIDS) originally appeared on a sizeable scale at the end of the 1970s, while scientific evidence on the existence of the Human Immunodeficiency Virus (HIV), a retrovirus which, once inside the bloodstream, fatally damages the immune system of the host,1 dates back to the early 1980s. Two types of
∗ Department of Economics, University of Udine, Italy ([email protected]).
1. Thus, those infected with HIV do not die because of the virus per se, but because their immune system fails to counteract opportunistic infections, which would not otherwise have been fatal. The status of AIDS patients is diagnosed precisely when the immune system of HIV-infected people stops working properly.
28 Edoardo Gaffeo
HIV have been identified to date: HIV-1, which is predominant in the United States and throughout the world, and HIV-2, which is primarily found in West Africa.
So far, more than 60 million people worldwide have been infected with the virus and, according to the most recent estimates issued by the Joint United Nations Programme on AIDS (UNAIDS, 2001a), around 40 million people are currently living with HIV/AIDS. In 2001, roughly 3 million people died from the disease, which has nowadays become the fourth largest cause of death at the global level. No curative remedies are currently known for HIV/AIDS, apart from some pharmacological ones, highly active antiretroviral therapies (HAARTs) aimed at prolonging the life of infected people and partially preventing mother-to-child transmission by infected pregnant women. Given the state of the art of bio-medical sciences, the only other workable way of successfully counteracting the epidemic is through prevention, i.e. by adopting safe behaviour, such as protected sex, limiting the number of sexual partners, and using mono-dose needles and, whenever possible, blood auto-transfusion.
Table 1 gives a picture of the epidemic’s geography, from which a couple of striking features emerge. First, every region of the world has been reached by the contagion, although with varying severity. This is reflected in the fact that the main medium of transmission varies from place to place, with unsafe blood donations, needle-sharing among injecting drug users and homosexual relationships being the main modes of transmission in the Northern Hemisphere (i.e., North America, Western and Eastern Europe and Central Asia), while heterosexual transmission is prevalent in the South (i.e., Africa, South and South-East Asia, and Latin America). The second, possibly more impressive, feature is that more than 70% of the entire world population which is currently infected lives in sub-Saharan Africa, where more than 8% of the adult population (i.e., people between 15 and 49 years of age) have HIV/AIDS.2 In 2001, the region was home to approximately 60% of the 5 million newly infected, and to more than 70% of those who died from the disease.
Following a steady rise all over the region since the 1950s, life expectancy started to decline drastically from 1992/93. In countries with high HIV prevalence, like South Africa, Zimbabwe and Zambia, to cite but a few, life expectancy has to date reached an impressive low of 47, 43 and 36 years, respectively. Some estimates forecast that by 2010 about 13,000 people will die every day of AIDS, and that more than 20 million African children will become orphans because of the epidemic. Mainly because of mother-to-child transmission of the virus, child mortality rates are rising as well: in Zimbabwe, for instance, the figure for deaths among children under the age of five is now 70%. Although the prevention efforts exerted by some governments, like those of Zambia, Senegal and especially Uganda, are starting to bear fruit, the high prevalence rates across the region ‘mean that even exceptional success on the prevention front will only gradually reduce the human toll’ (UNAIDS, 2001a: 14).
2. In some countries, like Swaziland and Botswana, the incidence rate is well above 30%.
The Economics of HIV/AIDS: A Survey 29
Table 1: The HIV/AIDS epidemic around the world, 2001
Region People living with
Caribbean 0.42 0.06 2.2 50
Eastern Europe & Central Asia
1 0.25 0.5 20
Australia & New Zealand
Total 40 5 1.2 48
Notes: a) in million people; b) proportion of adults (15 to 49 years) living with HIV/AIDS in total population, 2001. Source: UNAIDS (2001a).
Individual behaviour, market failures and disease transmission
The first step in surveying the economics of HIV/AIDS involves analysing the power of individual and social incentives in avoiding the contagion once at least one member of the population has already been infected with the virus. From this point of view, what really matters is the relationship between individual choices in a situation of incomplete information, and the degree of optimality of social outcomes.
In contrast to infections transmitted through casual interpersonal contacts (for example, tuberculosis) or spread by intermediate vectors (for example, malaria), for which a rationale for welfare-improving government intervention is well established, a sexually transmitted disease (STD) like HIV/AIDS almost always implies a voluntary activity subject to individual control by rational individuals (Gersovitz, 2000a). Hence, one is urged to add some behavioural flesh to the bones of epidemiological dynamics (see, for example, Philipson and Posner, 1993, 1995; Kremer, 1996), particularly if one is interested in analysing the rationale and scope of public intervention in epidemic control (Kremer, 1998).
30 Edoardo Gaffeo
According to this viewpoint, sex between two consenting persons implies a mutually beneficial exchange, so that sexual activity may be conveniently depicted in terms of a Pareto-improving economic trade, regardless of whether monetary transfers are made or not. The problem of AIDS is therefore reduced to a problem of the quality of the goods exchanged, and uncertainty about the health of one’s partner is simply a problem of uncertain quality in consumption.
In particular, two issues are of critical importance in the economic analysis of the HIV/AIDS epidemic: asymmetric information and externalities (Over, 1999). When combined, they imply the emergence of several transmission-enhancing market failures, which in turn constitute the grounds on which the public financing of health-care projects may be justified (Hammer, 1997). A technical exposition is given in Appendix 1. Informally, the argument runs as follows.
Let us start from the simplest case, in which an individual’s behavioural response to HIV prevalence, measured in terms of the desired number of partners, is fixed. For example, take a married man (say, agent A), who regularly has sex with both his wife (agent B), and another partner (agent C), who in turn has a positive probability of being infected. Suppose A has unprotected sex with C. Then, A’s behaviour imposes undesired costs on B – for example, a negative externality – in terms of the positive probability of B’s becoming infected if she chooses to have unprotected sex with A. Note that if actions could be realistically monitored, or if A could credibly commit himself to the use of condoms in extra-marital sex, the famous result of welfare economics known as the Coase Theorem would hold, meaning that A and B could negotiate ‘side payments’ to achieve an efficient outcome. If the affected parties contract with each other the externality will be internalised and there will be no welfare-improving role for government intervention.
Of course, efficient monitoring or credible commitments are practically impossible in the case in hand, so that actions that could help to curb the sequence of the epidemic, such as protected sex, will be under-provided for by the private market. The interaction of externalities and asymmetric information leads to a market failure producing an intolerable aftermath in terms of lost lives.
In addition to the monitoring problem, it has to be stressed that in many cases the power of incentives – expressed in terms of ‘side payments’ – that B can give to A is severely limited for cultural or institutional reasons. This is particularly true for women living in developing countries, who are often refused divorce regardless of their husband’s conduct; or do not have any incentive to put an end to the marriage even if allowed to do so by law, simply because the divorce laws assign women only a tiny fraction of the household’s wealth; or finally are forced by tradition and social customs to submit themselves fully to their husband’s will. Furthermore, in countries where polygamy is accepted and widely practised, heterogeneity in the degree of risk-aversion or free-riding behaviour among wives in their promised side payments might imply that only low-powered incentive schemes could be offered to husbands, so that the overall prevention effort ends up being well below the socially optimal level.
Given the assumption of a null behavioural response to HIV prevalence, this analysis suggests that there is undoubtedly plenty of scope for a welfare-enhancing public intervention through subsidies for private preventive efforts. Because of the public-good characteristic of information on the course of the epidemic, the government could also do a lot to decrease the steady-state prevalence rate by disseminating
The Economics of HIV/AIDS: A Survey 31
knowledge on the risks associated with the infection, and on how to avoid it. Furthermore, the above argument highlights the importance of governments endeavouring to persuade people to adopt safer sexual practices by facilitating the emergence of social norms3 aimed, for instance, at reducing male psychological resistance to condom use, thus modifying the elasticity of substitution between safe and unsafe sexual consumption.
The case for anti-AIDS public interventions should be treated with caution, however, once we admit the possibility that people might respond to changes in HIV prevalence by varying the rate of partner change.4 As emphasised, for example, by Philipson and Posner (1993) and Castillo-Chaves and Hadeler (1995), if the rate of partner change in a homogeneous population declines with prevalence, public efforts for prevention or subsidies to finance imperfect vaccines5 could prove to be partly, or even totally, ineffective because of behavioural response. The final result depends on the elasticity of the rate of partner change to a reduction in the transmission rate associated with the preventive action: if such an elasticity is greater than one, then a reduction in the transmission rate could increase steady-state prevalence.
Along a similar line, Philipson and Posner (1995) maintain that behavioural responses may well lead to perverse effects in the case of anti-AIDS public policies targeted at HIV testing. In particular, if we limit ourselves to a static analysis of the choice ‘safe vs. risky’ sex between two individuals, it emerges that a subsidy for testing can decrease or increase the prevailing incidence rate depending on the pre-test status quo ante, i.e. whether the two individuals had decided to have safe or risky sex before the test was available. If the pre-test sexual exchange is safe sex, testing is likely to increase HIV incidence if only one partner is tested. If, on the contrary, the pre-test status quo is unsafe sex, testing certainly leads to a lower incidence. If we move to a dynamic setting, it appears plausible to suppose that people who test positive could choose to increase their activity without taking any precautions simply because they feel they have nothing to lose, thus helping to widen the difference between private and social welfare, measured in terms of pure infection externality. However, it should be noted that this effect could be mitigated to a certain degree by the adoption of stricter risk-prevention strategies by sero-negative people, who conversely might choose to reduce their rate of partner change.
The issue of how social outcomes emerge as an aggregation of heterogeneous behaviours is addressed by Kremer (1996), who puts forward a model in which low- activity and high-activity people co-exist. He shows that, if increased prevalence leads low-activity people to reduce the number of desired partners substantially, the
3. For anecdotal evidence on the damage likely to be caused by the inaction of Thabo Mbeki’s government in
counteracting the HIV/AIDS epidemic in South Africa, see The Economist (2002). 4. Gersovitz (2000b) presents evidence, based on demographic and health surveys for 4 sub-Saharan African
countries (Kenya, Tanzania, Uganda and Zambia), showing that individual behaviour does indeed change with increased prevalence. In particular, main strategies for risk reduction reported by respondents include: (i) avoiding relations with prostitutes; (ii) reducing the number of partners; (iii) monogamy; and (iv) sexual abstinence, which is actually professed by 20% of the adult population in Kenya, Tanzania and Zambia. By contrast, condom adoption remains in general at a low level, mainly due to the social stigma associated with its use (Cohen and Trussel, 1996).
5. For example, it is estimated that the HIV vaccine working through exposure of the immune system to large quantities of the viral protein gp120, if it were available on a large scale in developing countries, might be roughly 30% effective.
32 Edoardo Gaffeo
composition of the pool of available partners will worsen as soon as high-activity people increase their activity, or at least decrease their activity only slightly, so that a negative externality emerges. This may create positive feedbacks leading to multiple steady-state epidemic equilibria, some of them having the somewhat counter-intuitive characteristic that an increase in the frequency of partner change by low-activity people reduces steady-state prevalence in the population. This result has been confirmed by certain simulations based on survey data on sexual activity in the UK, suggesting that a 5% decrease in the frequency of partner change by low-activity people, that is, individuals who change their partner on average every five years, could increase the long-run prevalence rate by 7% (Kremer and Morcom, 1997). It must be stressed, however, that this finding holds only if the analysis is applied to an ex-ante low- prevalence population, implying that an actual reduction in activity by low-activity people in sub-Saharan Africa, for instance, is not likely to increase HIV prevalence rates all over the region.
As previously maintained with regard to the difficulties of designing suitable incentive schemes to internalise pure infection externalities, it is worth noting that the perfectly rational approach to epidemic transmission discussed so far has probably undervalued other arguments that are likely to be of major importance. As emphasised, for example, by Baylies (2000) and Topouzis and du Guerny (1999), an appropriate way of tackling the problem of explaining and controlling the course of the HIV/AIDS epidemic in developing countries involves recognising that for a large fraction of the population, particularly women and young people, individual behaviour − for instance, in terms of ability to protect themselves during sex − is severely constrained by inequitable gender-related power relations, economic dependence, discrimination and social and cultural conditioning in general. As we shall at times argue below, this calls for a more eclectic approach to sexually transmitted diseases, in which economics, sociology, anthropology, biology and medicine interact in a fully interdisciplinary way.
Microeconomic issues
Thus far we have considered the determinants of disease transmission. What about the consequences?
A useful starting point in assessing the impact of HIV/AIDS on the economic conditions of sub-Saharan people – and a useful organising principle for the discussion which is to follow – consists in analysing in broad terms how sero-positivity interacts with the probability of a household being locked into a poverty trap. In particular, our discussion is based on a slightly modified version of a model by Dasgupta (1997), which investigates the interaction between health and under-nourishment on the one hand, and the dynamics of poverty, on the other. Once again, some basic technicalities are postponed till Appendix 2.
Intuitively, a poverty trap is generally defined as an inescapable situation of poverty, in which individual well-being is subject to a recurring cycle of crisis and partial recovery, if not to a real downward spiral. Several causes of poverty traps have been detected so far, such as insufficient investments in health, nutrition, human capital and productivity-enhancing technologies, limited access to financial services, the degradation of local environmental resource-bases, and social and civic disconnection. What is of real significance, however, is that their interaction gives rise to path-
The Economics of HIV/AIDS: A Survey 33
dependent poverty dynamics, in which bad outcomes are self-enforcing and the poor remain poor even if small interventions or chance improvements occur.
Indeed, the spread of the HIV/AIDS epidemic in sub-Saharan Africa is negatively affecting a wide range of activities at a microeconomic level. Items include agricultural issues such as peasant behaviour, land tenancy and interlinked markets; interactions between personal income, health and productivity; acquisition of education and (dis) investment in human and social capital; and finally access to credit and insurance markets. To understand how positive feedbacks between HIV/AIDS and all these situations can potentially trigger the emergence of countless new poverty traps at the household level, let us look at each of them in turn.
Although HIV/AIDS is generally perceived as an urban problem, in sub-Saharan Africa – where around 70% of the population live in rural areas – incidence and prevalence rates among rural communities are similar to those experienced in cities, so that the absolute number of people living with HIV/AIDS is higher in rural areas. According to estimates reported in Bullock (1994), the main breadwinners for families living in African villages are women, who on average undertake 10 out of 13 tasks in farming and harvesting, and produce between 60% and 80% of the food. The increasing incidence of infection among women6 therefore has serious implications for the food security and general health conditions of rural communities. The impact of the epidemic in terms of reduced food production interacts with other well-known crisis factors, such as desertification and increased costs of inputs, so that in the traditional farming sector the loss of production for several…
Published by Blackwell Publishing, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.
The Economics of HIV/AIDS: A Survey
Edoardo Gaffeo∗
This article surveys the main economic issues associated with the HIV/AIDS epidemic, paying special attention to sub-Saharan Africa. It explores the economic and behavioural determinants of HIV transmission, the microeconomics of market failures associated with high HIV prevalence, the prospects for regional development from a macroeconomic perspective and the efficient design of policies for coping with the epidemic. In line with the recent appeal by the UN Secretary General, the article argues that, without a decisive effort to halt HIV/AIDS, people living in the region are bound to experience a further fall in their standard of living in both relative and absolute terms. However, to be effective, anti-AIDS programmes must be rooted in sound economic principles.
This article attempts to set out a few propositions summarising what is known and largely accepted about the microeconomics and macroeconomics of the worldwide HIV/AIDS epidemic. The purpose in doing so is to put forward a set of presumptions about what should be seen as sound ‘economically oriented’ health policies aimed at limiting damage in the immediate term, and to set the stage for achieving the main long- run target of full eradication of the disease as quickly as possible. In particular, it devotes special attention to sub-Saharan Africa, where contagion has already reached global proportions and the issue has long ceased to be an exclusive health problem and has become a serious threat to long-term social and economic development.
The article is organised as follows. The next section gives a brief account of the epidemic so far, and some hints on what can reasonably be expected in the near future. Section three then deals with justifications for public intervention from an economic standpoint, followed by sections analysing the impacts of HIV/AIDS at a microeconomic and a macroeconomic level, respectively. Policy issues regarding the allocation of resources within HIV/AIDS programmes, and the development of effective vaccines, are then treated, followed by a concluding section.
An overview of the epidemic
The Acquired Immuno-Deficiency Syndrome (AIDS) originally appeared on a sizeable scale at the end of the 1970s, while scientific evidence on the existence of the Human Immunodeficiency Virus (HIV), a retrovirus which, once inside the bloodstream, fatally damages the immune system of the host,1 dates back to the early 1980s. Two types of
∗ Department of Economics, University of Udine, Italy ([email protected]).
1. Thus, those infected with HIV do not die because of the virus per se, but because their immune system fails to counteract opportunistic infections, which would not otherwise have been fatal. The status of AIDS patients is diagnosed precisely when the immune system of HIV-infected people stops working properly.
28 Edoardo Gaffeo
HIV have been identified to date: HIV-1, which is predominant in the United States and throughout the world, and HIV-2, which is primarily found in West Africa.
So far, more than 60 million people worldwide have been infected with the virus and, according to the most recent estimates issued by the Joint United Nations Programme on AIDS (UNAIDS, 2001a), around 40 million people are currently living with HIV/AIDS. In 2001, roughly 3 million people died from the disease, which has nowadays become the fourth largest cause of death at the global level. No curative remedies are currently known for HIV/AIDS, apart from some pharmacological ones, highly active antiretroviral therapies (HAARTs) aimed at prolonging the life of infected people and partially preventing mother-to-child transmission by infected pregnant women. Given the state of the art of bio-medical sciences, the only other workable way of successfully counteracting the epidemic is through prevention, i.e. by adopting safe behaviour, such as protected sex, limiting the number of sexual partners, and using mono-dose needles and, whenever possible, blood auto-transfusion.
Table 1 gives a picture of the epidemic’s geography, from which a couple of striking features emerge. First, every region of the world has been reached by the contagion, although with varying severity. This is reflected in the fact that the main medium of transmission varies from place to place, with unsafe blood donations, needle-sharing among injecting drug users and homosexual relationships being the main modes of transmission in the Northern Hemisphere (i.e., North America, Western and Eastern Europe and Central Asia), while heterosexual transmission is prevalent in the South (i.e., Africa, South and South-East Asia, and Latin America). The second, possibly more impressive, feature is that more than 70% of the entire world population which is currently infected lives in sub-Saharan Africa, where more than 8% of the adult population (i.e., people between 15 and 49 years of age) have HIV/AIDS.2 In 2001, the region was home to approximately 60% of the 5 million newly infected, and to more than 70% of those who died from the disease.
Following a steady rise all over the region since the 1950s, life expectancy started to decline drastically from 1992/93. In countries with high HIV prevalence, like South Africa, Zimbabwe and Zambia, to cite but a few, life expectancy has to date reached an impressive low of 47, 43 and 36 years, respectively. Some estimates forecast that by 2010 about 13,000 people will die every day of AIDS, and that more than 20 million African children will become orphans because of the epidemic. Mainly because of mother-to-child transmission of the virus, child mortality rates are rising as well: in Zimbabwe, for instance, the figure for deaths among children under the age of five is now 70%. Although the prevention efforts exerted by some governments, like those of Zambia, Senegal and especially Uganda, are starting to bear fruit, the high prevalence rates across the region ‘mean that even exceptional success on the prevention front will only gradually reduce the human toll’ (UNAIDS, 2001a: 14).
2. In some countries, like Swaziland and Botswana, the incidence rate is well above 30%.
The Economics of HIV/AIDS: A Survey 29
Table 1: The HIV/AIDS epidemic around the world, 2001
Region People living with
Caribbean 0.42 0.06 2.2 50
Eastern Europe & Central Asia
1 0.25 0.5 20
Australia & New Zealand
Total 40 5 1.2 48
Notes: a) in million people; b) proportion of adults (15 to 49 years) living with HIV/AIDS in total population, 2001. Source: UNAIDS (2001a).
Individual behaviour, market failures and disease transmission
The first step in surveying the economics of HIV/AIDS involves analysing the power of individual and social incentives in avoiding the contagion once at least one member of the population has already been infected with the virus. From this point of view, what really matters is the relationship between individual choices in a situation of incomplete information, and the degree of optimality of social outcomes.
In contrast to infections transmitted through casual interpersonal contacts (for example, tuberculosis) or spread by intermediate vectors (for example, malaria), for which a rationale for welfare-improving government intervention is well established, a sexually transmitted disease (STD) like HIV/AIDS almost always implies a voluntary activity subject to individual control by rational individuals (Gersovitz, 2000a). Hence, one is urged to add some behavioural flesh to the bones of epidemiological dynamics (see, for example, Philipson and Posner, 1993, 1995; Kremer, 1996), particularly if one is interested in analysing the rationale and scope of public intervention in epidemic control (Kremer, 1998).
30 Edoardo Gaffeo
According to this viewpoint, sex between two consenting persons implies a mutually beneficial exchange, so that sexual activity may be conveniently depicted in terms of a Pareto-improving economic trade, regardless of whether monetary transfers are made or not. The problem of AIDS is therefore reduced to a problem of the quality of the goods exchanged, and uncertainty about the health of one’s partner is simply a problem of uncertain quality in consumption.
In particular, two issues are of critical importance in the economic analysis of the HIV/AIDS epidemic: asymmetric information and externalities (Over, 1999). When combined, they imply the emergence of several transmission-enhancing market failures, which in turn constitute the grounds on which the public financing of health-care projects may be justified (Hammer, 1997). A technical exposition is given in Appendix 1. Informally, the argument runs as follows.
Let us start from the simplest case, in which an individual’s behavioural response to HIV prevalence, measured in terms of the desired number of partners, is fixed. For example, take a married man (say, agent A), who regularly has sex with both his wife (agent B), and another partner (agent C), who in turn has a positive probability of being infected. Suppose A has unprotected sex with C. Then, A’s behaviour imposes undesired costs on B – for example, a negative externality – in terms of the positive probability of B’s becoming infected if she chooses to have unprotected sex with A. Note that if actions could be realistically monitored, or if A could credibly commit himself to the use of condoms in extra-marital sex, the famous result of welfare economics known as the Coase Theorem would hold, meaning that A and B could negotiate ‘side payments’ to achieve an efficient outcome. If the affected parties contract with each other the externality will be internalised and there will be no welfare-improving role for government intervention.
Of course, efficient monitoring or credible commitments are practically impossible in the case in hand, so that actions that could help to curb the sequence of the epidemic, such as protected sex, will be under-provided for by the private market. The interaction of externalities and asymmetric information leads to a market failure producing an intolerable aftermath in terms of lost lives.
In addition to the monitoring problem, it has to be stressed that in many cases the power of incentives – expressed in terms of ‘side payments’ – that B can give to A is severely limited for cultural or institutional reasons. This is particularly true for women living in developing countries, who are often refused divorce regardless of their husband’s conduct; or do not have any incentive to put an end to the marriage even if allowed to do so by law, simply because the divorce laws assign women only a tiny fraction of the household’s wealth; or finally are forced by tradition and social customs to submit themselves fully to their husband’s will. Furthermore, in countries where polygamy is accepted and widely practised, heterogeneity in the degree of risk-aversion or free-riding behaviour among wives in their promised side payments might imply that only low-powered incentive schemes could be offered to husbands, so that the overall prevention effort ends up being well below the socially optimal level.
Given the assumption of a null behavioural response to HIV prevalence, this analysis suggests that there is undoubtedly plenty of scope for a welfare-enhancing public intervention through subsidies for private preventive efforts. Because of the public-good characteristic of information on the course of the epidemic, the government could also do a lot to decrease the steady-state prevalence rate by disseminating
The Economics of HIV/AIDS: A Survey 31
knowledge on the risks associated with the infection, and on how to avoid it. Furthermore, the above argument highlights the importance of governments endeavouring to persuade people to adopt safer sexual practices by facilitating the emergence of social norms3 aimed, for instance, at reducing male psychological resistance to condom use, thus modifying the elasticity of substitution between safe and unsafe sexual consumption.
The case for anti-AIDS public interventions should be treated with caution, however, once we admit the possibility that people might respond to changes in HIV prevalence by varying the rate of partner change.4 As emphasised, for example, by Philipson and Posner (1993) and Castillo-Chaves and Hadeler (1995), if the rate of partner change in a homogeneous population declines with prevalence, public efforts for prevention or subsidies to finance imperfect vaccines5 could prove to be partly, or even totally, ineffective because of behavioural response. The final result depends on the elasticity of the rate of partner change to a reduction in the transmission rate associated with the preventive action: if such an elasticity is greater than one, then a reduction in the transmission rate could increase steady-state prevalence.
Along a similar line, Philipson and Posner (1995) maintain that behavioural responses may well lead to perverse effects in the case of anti-AIDS public policies targeted at HIV testing. In particular, if we limit ourselves to a static analysis of the choice ‘safe vs. risky’ sex between two individuals, it emerges that a subsidy for testing can decrease or increase the prevailing incidence rate depending on the pre-test status quo ante, i.e. whether the two individuals had decided to have safe or risky sex before the test was available. If the pre-test sexual exchange is safe sex, testing is likely to increase HIV incidence if only one partner is tested. If, on the contrary, the pre-test status quo is unsafe sex, testing certainly leads to a lower incidence. If we move to a dynamic setting, it appears plausible to suppose that people who test positive could choose to increase their activity without taking any precautions simply because they feel they have nothing to lose, thus helping to widen the difference between private and social welfare, measured in terms of pure infection externality. However, it should be noted that this effect could be mitigated to a certain degree by the adoption of stricter risk-prevention strategies by sero-negative people, who conversely might choose to reduce their rate of partner change.
The issue of how social outcomes emerge as an aggregation of heterogeneous behaviours is addressed by Kremer (1996), who puts forward a model in which low- activity and high-activity people co-exist. He shows that, if increased prevalence leads low-activity people to reduce the number of desired partners substantially, the
3. For anecdotal evidence on the damage likely to be caused by the inaction of Thabo Mbeki’s government in
counteracting the HIV/AIDS epidemic in South Africa, see The Economist (2002). 4. Gersovitz (2000b) presents evidence, based on demographic and health surveys for 4 sub-Saharan African
countries (Kenya, Tanzania, Uganda and Zambia), showing that individual behaviour does indeed change with increased prevalence. In particular, main strategies for risk reduction reported by respondents include: (i) avoiding relations with prostitutes; (ii) reducing the number of partners; (iii) monogamy; and (iv) sexual abstinence, which is actually professed by 20% of the adult population in Kenya, Tanzania and Zambia. By contrast, condom adoption remains in general at a low level, mainly due to the social stigma associated with its use (Cohen and Trussel, 1996).
5. For example, it is estimated that the HIV vaccine working through exposure of the immune system to large quantities of the viral protein gp120, if it were available on a large scale in developing countries, might be roughly 30% effective.
32 Edoardo Gaffeo
composition of the pool of available partners will worsen as soon as high-activity people increase their activity, or at least decrease their activity only slightly, so that a negative externality emerges. This may create positive feedbacks leading to multiple steady-state epidemic equilibria, some of them having the somewhat counter-intuitive characteristic that an increase in the frequency of partner change by low-activity people reduces steady-state prevalence in the population. This result has been confirmed by certain simulations based on survey data on sexual activity in the UK, suggesting that a 5% decrease in the frequency of partner change by low-activity people, that is, individuals who change their partner on average every five years, could increase the long-run prevalence rate by 7% (Kremer and Morcom, 1997). It must be stressed, however, that this finding holds only if the analysis is applied to an ex-ante low- prevalence population, implying that an actual reduction in activity by low-activity people in sub-Saharan Africa, for instance, is not likely to increase HIV prevalence rates all over the region.
As previously maintained with regard to the difficulties of designing suitable incentive schemes to internalise pure infection externalities, it is worth noting that the perfectly rational approach to epidemic transmission discussed so far has probably undervalued other arguments that are likely to be of major importance. As emphasised, for example, by Baylies (2000) and Topouzis and du Guerny (1999), an appropriate way of tackling the problem of explaining and controlling the course of the HIV/AIDS epidemic in developing countries involves recognising that for a large fraction of the population, particularly women and young people, individual behaviour − for instance, in terms of ability to protect themselves during sex − is severely constrained by inequitable gender-related power relations, economic dependence, discrimination and social and cultural conditioning in general. As we shall at times argue below, this calls for a more eclectic approach to sexually transmitted diseases, in which economics, sociology, anthropology, biology and medicine interact in a fully interdisciplinary way.
Microeconomic issues
Thus far we have considered the determinants of disease transmission. What about the consequences?
A useful starting point in assessing the impact of HIV/AIDS on the economic conditions of sub-Saharan people – and a useful organising principle for the discussion which is to follow – consists in analysing in broad terms how sero-positivity interacts with the probability of a household being locked into a poverty trap. In particular, our discussion is based on a slightly modified version of a model by Dasgupta (1997), which investigates the interaction between health and under-nourishment on the one hand, and the dynamics of poverty, on the other. Once again, some basic technicalities are postponed till Appendix 2.
Intuitively, a poverty trap is generally defined as an inescapable situation of poverty, in which individual well-being is subject to a recurring cycle of crisis and partial recovery, if not to a real downward spiral. Several causes of poverty traps have been detected so far, such as insufficient investments in health, nutrition, human capital and productivity-enhancing technologies, limited access to financial services, the degradation of local environmental resource-bases, and social and civic disconnection. What is of real significance, however, is that their interaction gives rise to path-
The Economics of HIV/AIDS: A Survey 33
dependent poverty dynamics, in which bad outcomes are self-enforcing and the poor remain poor even if small interventions or chance improvements occur.
Indeed, the spread of the HIV/AIDS epidemic in sub-Saharan Africa is negatively affecting a wide range of activities at a microeconomic level. Items include agricultural issues such as peasant behaviour, land tenancy and interlinked markets; interactions between personal income, health and productivity; acquisition of education and (dis) investment in human and social capital; and finally access to credit and insurance markets. To understand how positive feedbacks between HIV/AIDS and all these situations can potentially trigger the emergence of countless new poverty traps at the household level, let us look at each of them in turn.
Although HIV/AIDS is generally perceived as an urban problem, in sub-Saharan Africa – where around 70% of the population live in rural areas – incidence and prevalence rates among rural communities are similar to those experienced in cities, so that the absolute number of people living with HIV/AIDS is higher in rural areas. According to estimates reported in Bullock (1994), the main breadwinners for families living in African villages are women, who on average undertake 10 out of 13 tasks in farming and harvesting, and produce between 60% and 80% of the food. The increasing incidence of infection among women6 therefore has serious implications for the food security and general health conditions of rural communities. The impact of the epidemic in terms of reduced food production interacts with other well-known crisis factors, such as desertification and increased costs of inputs, so that in the traditional farming sector the loss of production for several…
Related Documents
