AIDS & Economics[1]

 

 

 

David E Bloom, Ajay Mahal, Jaypee Sevilla, and River Path Associates

 

November 2001

Introduction

More than 22 million people have now died of AIDS, including 3 million in 2000 alone.   Thirty-six million people are currently infected with the virus and, although infection rates are stabilizing in Sub-Saharan Africa (albeit at a very high level), the epidemic is still growing in Asia and Eastern Europe. The lack of an imminent vaccine or cure means that many more deaths are inevitable.

The seriousness of the AIDS epidemic raises questions about the potential impact of the epidemic on national and regional economies. AIDS disproportionately affects people of working age, and is also creating a huge burden of AIDS orphans. Many political leaders have expressed alarm in light of studies showing the potential devastation of their economies, but relatively few businesses have spoken out decisively.

This paper provides an overview of the links between AIDS and economics. Part 1 assesses economic correlates of HIV transmission, using macro, micro and household data. Part 2 examines the effect of AIDS on economies, both directly and as a cause of demographic change.  Part 3 explores the likely economic returns to actions that prevent HIV infections.

 

Part 1 – Economic determinants of HIV transmission

This section reviews existing studies of the association between HIV/AIDS and economic status, at both the macro and micro levels, and assesses the impact of development on AIDS.

The macro level

The link between income levels and AIDS prevalence is complex and poorly understood. Data from the 1980s and early 1990s, mainly from sub-Saharan Africa, seem to indicate that the wealthy were at highest risk from the epidemic. Two of Africa’s richest nations, Botswana and South Africa, are among the most affected nations in the world. Figure 1, based on data from 1997, suggests a continuing disproportionate impact on Africa’s richer countries, possibly reflecting the role that better infrastructure and more mobile populations seem to play in the spread of the disease. Within all other continents, there is neither a positive nor a negative statistical association between income levels and AIDS. Examined continent by continent, therefore, HIV appears to be either affecting the rich more than the poor (as in Africa) or is income neutral (everywhere else).

Between continents, however, the picture looks different. Ninety-five percent of those infected with HIV live in less developed countries, home to 80 percent of the world’s population. As figures 2 and 3 show, at a global level there is a statistically significant relationship between low income and HIV prevalence rates; that is, the poorer the country the greater the HIV prevalence. There is a similar relation between income distribution and HIV prevalence, with countries with greater income inequality facing a more serious epidemic. Absolute poverty rates, defined as income below $1 a day, are strongly associated with HIV prevalence rates (figure 4), as are low rankings on the United Nations Development Programme (UNDP) Human Poverty Index, which takes into account mortality, literacy, malnutrition and access to water, sanitation and health services (figure 5).

Existing data provide some indication that the relationship between poverty and HIV is growing stronger over time, both between and within continents. But it is not possible to infer causality from these data. That is, it is difficult to tell whether poverty causes AIDS or vice versa – or whether another variable, such as war, inadequate health, or poor education, explains the relationship.

The micro level

Compared to existing macro data, micro-studies appear to be better equipped to highlight links between economic status and AIDS.  In this regard, the intuitive link between knowledge and HIV transmission is supported by several studies. School enrolment rates and illiteracy rates in the majority of the developing world, and particularly in Africa, are substantially lower than those in richer countries, and the poor within countries are least likely to receive education.[2] The poor are therefore less likely to be aware of the dangers of HIV/AIDS than the rich.

-        Analysis of household data from Cambodia, Vietnam, Nicaragua and Tanzania (see Appendix 1) shows a strong correlation between both wealth and education and: knowledge that condoms prevent AIDS; knowledge of where condoms can be obtained; and self-reported usage of condoms.[3]

-        Recent research in Cambodia, the country with the most advanced epidemic in Asia, demonstrates that the poorest segments of society have much less knowledge of how AIDS is transmitted and prevented; are more likely to have sex at a younger age; use condoms less frequently; and, in the case of young women, are more likely to turn to sex work as a means of supporting themselves and their families.[4]

-        A study in Brazil showed that three-quarters of people newly diagnosed with HIV in the early 1980s had a secondary or university education, but by the early 1990s this share had fallen to one third.[5]

-        A study in rural Uganda, on the other hand, found that in a cohort of almost 20,000 adults aged 15-59 years followed over 3 and one-half years, HIV-associated mortality was highest among the better educated[6]. However, there is evidence that this pattern may be changing over time. Another study in Uganda (see figure 6) shows that the better educated were hit hardest in the early stages of the epidemic, but that HIV infection rates are now falling quickest among those with more education.[7]

Education is not the only factor highlighted by micro data. There is evidence that poverty forces many people to work in the commercial sex industry, thereby putting them at risk of HIV infection. A series of small-scale studies from sub-Saharan Africa, Haiti and Brazil,[8] shows how poor women can be forced into sex work, or into providing sexual favors in return for money. They are also shown to be less able to insist on condom use than their wealthier clients.

While the micro data is suggestive of a link between poverty and AIDS, many small scale studies are based on non-representative samples in the hardest hit areas, and some larger scale research, such as that conducted in Uganda, shows a negative correlation between HIV and poverty. As with the macro data, there remain many unanswered questions.

The impact of development

While poverty reduction might be thought to reduce HIV/AIDS rates, in some cases the development process may itself strengthen epidemics. Development is associated with infrastructure development, urbanization, increases in disposable income, the growing importance of cash in agriculture, and growing mobility. Furthermore, inequality often grows in the early stages of development, [9] creating increased internal migration (as workers migrate to centers of wealth and employment), a significant risk factor as men travel away to work, but occasionally return to their families in their village of origin. Development is likely to bring greater opportunities for multiple partnering and a growth in the commercial sex industry. [10] Finally, inequality can create changes in gender relations that may facilitate the spread of sexually transmitted diseases. [11]

There is currently little evidence to quantify the extent of the HIV risks caused by the unintended consequences of development efforts. However, a strong case is building for making HIV impact assessment a routine part of programs designed to promote development and poverty reduction.

Poverty to AIDS

In sum, the link between economic status and AIDS is complex. While many micro level studies point to a significant link between poverty and HIV prevalence rates, macro data is unconvincing, particularly in terms of the causality of the link. Some risk factors for HIV, such as a high level of disposable income, are more prevalent amongst the rich than the poor. Others, such as lack of education, are more prevalent among the poor than the rich. Both groups exhibit the kind of mobility that appears to be associated with HIV transmission.

On balance, it seems plausible that the rich are more at risk in the early stages of an epidemic, and that a combination of factors, including lack of education and other economic exigencies, put the poor at increasing risk as an epidemic progresses. One might therefore expect HIV epidemics to become increasingly embedded in poor communities. Although not proven, this hypothesis is broadly consistent with patterns of HIV transmission seen in Africa and other regions, including wealthy industrial countries such as the US.

Part 2: The impact of AIDS on economies 

The humanitarian case for taking action to prevent HIV/AIDS is clear. However, there is also value in exploring the economic case for action. With many problems competing for public sector budgets, governments need guidance on where to devote their resources. Businesses may also need to adjust their strategies to respond to the epidemic. This section therefore explores the epidemic’s economic impact.

Macro evidence

A lack of reliable time series data on poverty rates and AIDS makes drawing macro level conclusions about the impact of AIDS on economies difficult. Inter-continental poverty differences predate the AIDS epidemic, and the hardest hit continent – Africa – is mired in too complex and deep a development trap to make disaggregating the effects of the epidemic feasible or persuasive.

There are many mechanisms through which AIDS may have a potential impact on the economy. Unlike most other deadly illnesses, HIV’s prime target is people of working-age.[12] The result is a potential reduction in savings rates and disposable income, which may have an economic impact.[13] New staff must be trained and recruited, a cost that would not otherwise have been borne. Firms may also suffer a loss of valuable know-how. Moreover, AIDS is debilitating, particularly in the final two years before death,[14] and absenteeism for both those infected and those caring for them may have an impact on businesses and other work organizations. Increases in health spending could mean cuts in investment in other growth-enhancing areas, education and infrastructure, for example. The impact on productivity may also decrease an economy’s attractiveness to foreign investors, and diminish tax revenue.

However, there are other influences that may counter these effects. Workers who die of AIDS may be replaced by people who were previously unemployed and a smaller labor force may even lead to a rise in output per capita. Although HIV/AIDS mortality can reduce overall output, it also reduces population, so per capita productivity may not be reduced.[15]  Even in the hardest hit areas, therefore, it is possible that GDP per capita may not decline.

Macro evidence from the early phases of the epidemic failed to substantiate the hypothesis  that AIDS would have a detrimental effect on growth rates of per capita income. From 1980 to 1992, AIDS had no statistically significant impact on per capita income growth.[16] However, the epidemic has since grown rapidly and has begun to have a significant effect on life expectancy and other human development indicators.[17] In South Africa, for example, life expectancy is expected to fall between 18 and 25 years below its pre-AIDS level.[18] Recent studies of AIDS in South Africa show some evidence of growing macroeconomic impact.

-        In the Caribbean, one study projects that GDP in 2005 could be around 4.2 percent lower than it might have been in the absence of the epidemic.[19]

-        Another study claims that AIDS will reduce output in Botswana by a third.  But since the epidemic may cause the population to fall by about 30 percent, the effect on per capita income may be negligible. 

-        An early World Bank study, of 30 sub-Saharan African countries, concludes that the net effect of AIDS will be a reduction in GDP growth of between 0.8 and 1.4 percentage points per year.[20]

Conversely, however, a recent assessment of AIDS in Asia concluded that the region’s low prevalence rates are likely to mean that the impact of the virus on Asian economies remains minimal.[21]

Data are far from adequate, but calculations made for Thailand may be instructive for understanding the potential economic effect of AIDS in Sub-Saharan Africa. Thailand’s ratio of working-age to total population is projected to be 0.70 in 2015.[22] We estimate that cumulative AIDS deaths by that year will be about 1 million, a relatively small number because risky behaviors have declined as a result of Thailand's highly successful anti-HIV policies. Yet if we simulate cumulative AIDS deaths in the absence of these substantial behavioral improvements, they could be as high as 10 million.  Add to this an estimate of the number of children that would not have been born because of these deaths and the population could be about 11.6 million smaller than it otherwise would have been.  AIDS mortality is disproportionately selective of adults, and we project that of the 10 million deaths, 92% or 9.2 million would be among adults.  To this number, we add the .75 million children these adults would have had, and who would have had the chance to reach working age by 2015, and we find that this high-risk scenario causes the working age population to be smaller by about 9.95 million.[23]  This combined effect on the total and working age population would result in a decline in the working-age share of the population to 0.67.  This difference could reduce the average growth rate of per capita GDP between 1990 and 2015 by about .65 percentage points, such that annual growth rates are projected to be 2.81 percent instead of 3.46 percent.  As a result, the level of GDP per capita in 2015 would be $1272 lower than its projected $8500.  At Thailand’s current prevalence rates, still among the highest outside Africa at an adult rate of 2.15 percent, the impact on GDP is minimal. Nevertheless, the example demonstrates that an unchecked AIDS epidemic – as some African countries are experiencing – can have a substantial effect on the growth of income per capita because it is so highly concentrated in working-age individuals.[24]

In reviewing all the available evidence, UNAIDS states that, “despite incomplete data, there is growing evidence that as HIV prevalence rates rise, both total and growth in national income – GDP – fall significantly.”[25] It is important to emphasize, however, that the data are incomplete; many of the studies forecast rather than report impacts; and the methodology of some studies can be questioned.

The impact on business

The effects of HIV/AIDS on a business are likely to be felt in three areas: a firm’s labor force, its customer base, and the reputation of the company.

There have been a number of studies on the impact of HIV/AIDS on the workforce. In the early stages of the epidemic in Africa, the spread of the epidemic appeared to correlate with wealth, and firms therefore seemed to be losing their most skilled, productive, and expensive-to-replace workers. More recently, however, a study has suggested that overall infection rates will peak at nearly three times the rate for highly skilled workers.[26] Generally, turnover in Africa was minimally affected in the early stages of the epidemic,[27] but as the epidemic has matured, companies in hard-hit areas have begun to feel an impact.[28] Some multinational organizations in South Africa, for example, claim to have hired 3 workers for each position to replace those who die.[29] Outside Africa, there is as yet no evidence that the epidemic is disproportionately targeting the skilled. Companies in certain sectors may suffer – trucking companies in India, for example[30] – but large-scale decimation of workforces is unlikely.

There are other potential effects that are even more difficult to quantify. Studies in Kenya and Thailand, for example, suggest, not surprisingly, that motivation and productivity are adversely affected by AIDS-related illnesses and death, for example.[31] In the future, it is also possible that the quality of available workers will deteriorate, as AIDS orphans (of whom there are currently 13 million) receive less education and are poorly socialized. [32] Studies are not yet able to show the likely economic impact of these effects.

The impacts on the customer base will be felt most keenly in Africa, although countries that trade extensively with African countries (as well as multinationals with franchises in Africa) may feel a ripple effect. Studies in Cote d’Ivoire and Rwanda have shown how health expenditure significantly reduces the household consumption of families living with AIDS.[33] A study of the epidemic in Thailand claims that it may cost Japan as much as 1.2 percent of its gross national product (GNP) due to the weakening of this important market for Japanese exports.[34] The methodological basis of this and similar studies is questionable however.

Even without proof that the virus will devastate labor forces and markets, however, there might be other reasons for some firms to take action against the disease. Pharmaceutical companies have profited from the disease in terms of sales, but many have suffered harmful effects to their reputations as protests by individuals, NGOs, and national governments have attacked AIDS drug patents.[35] Other companies, such as Levi Strauss and MTV, have been widely praised for their prompt action to educate their workers and their customers as to the dangers of the disease. As a recent study on corporate responsibility and HIV/AIDS for the American Foundation for AIDS Research suggested, there are both positive and negative motivations for companies to intervene on behalf of the public interest (see model at figure 7). “Negative” interventions will be a response to demands from employees, customers or shareholders, whereas “positive” actions will be driven by a desire to stand out above competitors and search out new business opportunities. As the paper says, “boosting employee morale, raising corporate profile and contributing to society are powerful drivers for many of today’s most innovative companies.”[36] The high profile of the AIDS virus, particularly among the young generation that has grown up with the disease, presents a powerful opportunity for corporate action.

The effect on households

The effect of AIDS on affected households is substantial. AIDS is an expensive illness to treat, and caring costs are high. Savings rates in affected households are likely to suffer as a result.[37] A series of micro level studies shows the impact of AIDS on households:

-        A large-scale World Bank survey of households in Tanzania, Cote d’Ivoire and Thailand found that household expenditures on AIDS care were much higher than on other illnesses.[38]

-        Household income in the poorest quarter of households in Botswana is projected to fall by 13 percent from current levels in the next ten years as a result of the disease.[39] 

-        A study in Cambodia shows the poor are forced to sell limited family assets to pay for the cost of caring for a family member with AIDS. They are also likely to borrow, at high rates of interest.[40] 

Again, however, the literature is patchy. Many of the micro studies are carried out in extremely hard-hit areas using non-representative samples and none gives a wholly reliable picture of the effect of the virus.[41]

AIDS and economics

The lack of conclusive evidence on the economic impact of the AIDS epidemic reflects the lack of investment in research by governments and donors and a failing in the academic community. Further studies will therefore be needed to provide conclusive evidence of the size and nature of any effects.

Part 3: The Economic Return on HIV Prevention Programs

Irrespective of whether one can (or cannot) measure all of the economic impacts of HIV/AIDS epidemic, it can nonetheless be demonstrated that investments in HIV/AIDS prevention have the potential of yielding high rates of economic return.  This is because, independent of any other consideration, increased numbers of AIDS cases and deaths require medical expenditures for treatment and impose a clear loss to society in the form of lost output. In this section we provide rough estimates of the rate of return to HIV prevention and compare it to returns to other investments, in the health sector and elsewhere.

Rate of return

Without reliable data, it is difficult for governments to set investment priorities. The standard approach adopted by economists advising governments is to calculate the rate of return (ROR) on competing demands for resources and direct the funds to investments that yield the highest return, followed by the second highest, and so on, until the budget is exhausted. If an activity offers higher rates of return than alternative uses, the case for investing in that activity is strengthened. Benefits from HIV prevention accrue from both the medical costs averted (by both private and public sectors) and the value of lives saved on account of the intervention.[42] Research conducted for this paper provides an attempt at assessing the ROR from HIV prevention efforts based on data from Thailand, whose efforts in the 1990s were successful in reducing the number of new AIDS cases, which had doubled to 26,000 from 1994 to 1997, back to 1994 levels in just two years (see Appendix 2 for full details).[43]

The time period from 1990, when Thailand’s prevention activities began, to 2020, when the full effects will be measurable, was chosen for the analysis. Public sector and donor expenditures on HIV/AIDS jumped from US$0.68 million in 1998 to nearly US$10 million in 1991 and $82 million by 1997. It is estimated that roughly 15 percent of these expenditures were on prevention activity.[44] The private sector spent an estimated US$80 million on prevention messages in 1991.[45] Data on changes in behavior suggests that if behaviors had remained unchanged at 1990 levels, there would have been more than 12 million extra deaths due to AIDS in Thailand, cumulatively, by the year 2020 compared to current behavioral patterns.   

Even with conservative estimates of the impact of the prevention campaigns on the changes in behavior, ROR is calculated at between 12 percent and 380 percent annually, depending on the scenario posited:

-        If we focus only on benefits in terms of medical expenditures avoided, the rates of return range from 12 percent to 33 percent over a 30-year period  (the lower bound is an outcome of assuming growth of medical expenditures in tune with per capita income). 

-        If we include averted income losses as additional benefits resulting from the reduced number of AIDS deaths (i.e., in addition to savings in medical expenditures), the rate of return jumps upwards very sharply – with the range now from 37 percent to 55 percent.

-        If, however, instead of income losses, we consider the value of an averted AIDS death to be equal to the statistical value of a life, the rates of return of HIV prevention programs jump to upwards of 380 percent per year.

Estimates of the rate of return for some other health interventions are available. The rate of return (inclusive of income losses due to disability) of the global guinea-worm eradication program, for example, is roughly 29 percent, compared to 37-55 percent for Thailand for HIV prevention, calculated by an equivalent methodology. Our estimates for the rate of return on HIV prevention expenditures (inclusive of income losses) also exceed the range of rates of return from interventions for river blindness eradication in Africa (which the World Bank estimates at 6-17 percent[46]). The World Bank considers an annual rate of return of greater than 10 percent to be acceptable.[47]

Integrated action

It is clear, therefore, that intervention to prevent HIV/AIDS provides a potentially high rate of return on investment (even the lowest estimate is above World Bank criteria). However, there remains the need for further research in this area, to explore the rate of return in different countries, with different intensity of epidemic, at different stages in the development process, and with different intervention portfolios.

It is worth noting that one of the features of Thailand’s success has been the reliability of its data on HIV/AIDS. This has enabled the nature of the virus to be tracked over time and spot emerging changes of profile in the epidemic.[48] In the absence of such surveillance, governments are unlikely to be able to keep up with HIV, and resource allocation will inevitably be inefficient.

Conclusion

The connection between AIDS and economics is complex, and drawing firm conclusions is complicated by the lack of concrete data in many areas. The poor appear to be most vulnerable to AIDS, but it is possible that this is not just because they are poor, but because of the interaction between poverty and other factors such as poor education, migration and weak health systems. Poverty reduction may decrease risk from the epidemic, but it is also possible that ill-planned development efforts will temporarily increase the risk that poor people face.

The impact of AIDS on economies is also hard to measure. It seems clear that there were limited effects early in the epidemic. Although some studies now project increasing impact, they are speculative, even if there is a strong intuition that very badly affected countries will see a significant economic deterioration. It is possible that this effect will only be felt when the prevalence of HIV/AIDS surpasses a certain threshold, however.

Finally, there is a need for better data to help track the development of the epidemic, to judge the most effective interventions, and to help decision-makers decide between competing priorities. Our preliminary estimates suggest a high potential return to investments in HIV prevention.

Our understanding of the epidemic seems to be somewhat weak, given the time since the discovery of HIV, the global nature of the epidemic, and its ferocity. At a time of huge political interest in health as a tool of development, it is clear that further work is badly needed.


Figure 6

                                                Source: UNAIDS (2000)

Figure 7

                        Source: Bloom et al (2001) ibid

Appendix 1

We ran binary logits on each of the following outcome variables:

1.  knowledge of the HIV-preventive benefits of condom use

2.  knowledge of the HIV-preventive benefits of having just one sexual partner

3.  knowledge of the HIV-preventive benefits of avoiding sex with prostitutes

4.  knowing of a source for condoms

5.  knowing about condoms

6.  ever having used a condom

Each logit contained dummy variables for wealth quintile, highest level of education achieved, and 5-year age cohort.  We report odds ratios for the effect of being in the wealthiest quintile and having the highest level of education.  In all countries but Tanzania, the highest education level is higher education.  In Tanzania, the highest education level is secondary education.

v754c  byte     percent18.0g  V754C AIDS:   use condoms during sex

Country

Wealthiest quintile

Highest Education

Cambodia

n=15,351

2.139

(0.259)***

1.953

(1.142)

Vietnam

n=1658

2.684

(1.157)**

6.455

(4.284)**

Nicaragua

n=1334

1.970

(0.377)***

1.876

(0.723)

Tanzania

n=1675

3.031

(0.894)***

3.771

(1.355)***

 

v754d  byte     percent20.0g  V754D AIDS:   only one sex partner

Country

Wealthiest quintile

Highest Education

Cambodia

n=15,351

1.127

(0.095)

1.109

(0.291)

Vietnam

n=1658

1.959

(0.797)

4.144

(2.552)**

Nicaragua

n=1334