http://www.cid.harvard.edu/

For information contact:
CAER II Project Office
Harvard Institute for International Development
14 Story Street
Cambridge, MA 02138
USA
Tel: (617) 495-9776
Fax: (617) 496-9951
Email: caer@ksg.harvard.edu

At a public lecture at Harvard University in 1999, Dr. Peter Piot, Director of UNAIDS, made an interesting comment. He noted that despite having worked on HIV/AIDS for most of the last two decades, he continues to be surprised by the inaccuracy of predictions about the course of the epidemic, including those that he, himself, has made.¹ Dr. Piot singled out the extent to which the economic and social impacts of the disease have been understated.

One. The conceptual model commonly used to trace the dynamic effects of HIV/AIDS on economic growth is seriously flawed. The most common projection models are linear in growth rates.² Emerging evidence suggests that they should be curvilinear with growth rates declining at an increasing rate as the epidemic intensifies.³

Two. The main official responses to the epidemic so far have been focused on prevention and cure. Generally overlooked have been the tasks of training, managing, motivating, and otherwise constructively engaging the large numbers of people who are HIV-positive and whose productive lives are being prematurely shortened.⁴ Because of this bias, maintaining labor productivity has become increasingly difficult. We argue that without a special effort to raise (or maintain) the productivity of those who are HIV-positive, economic growth cannot be maintained let alone increased.

Three. The full implications for economic recovery and subsequent growth and development of the loss of skills and erosion of institutions across Africa remain largely unrecognized.⁵ With the exception of Botswana (where the prevalence of HIV/AIDS is high) and Mauritius (where it is not) all other African countries (including South Africa) have been engaged in efforts to reverse years of economic stagnation and decline. Though some analysts had claimed that Africa was "on the move,"⁶ economic reform was eluding most African countries even before HIV/AIDS began decimating their work forces and eroding the performance of their major organizations. Indeed, a case can be made that the cumulative losses due to HIV/AIDS are preventing recovery in some countries. For countries, such as Zambia, Zimbabwe, Uganda, and Malawi where a high proportion of the adult population is HIV-positive,⁷ barriers to reform and future growth may already exist that are not fully understood.⁸

The literature on HIV/AIDS is truly enormous.⁹ Two aspects of that literature concern us here, namely the economic impact of HIV/AIDS in Africa¹⁰ and the implications of HIV/AIDS for public policy.¹¹

Analyses of the costs of HIV/AIDS have focused primarily on the impact on the labor force, the family, the education system, and national economies. A major area of emphasis has been the rising costs due to the progressive debilitation and loss of workers. These studies highlight other negative effects including: the impact on organizations (public and private) associated with the added cost of training replacement workers, higher wage bills as additional staff are retained to compensate for absenteeism and worker debility, the costs incurred when work schedules are disrupted, the increase in employer provided health and medical costs, and the unanticipated depletion of pension and insurance funds.¹² The early literature suggested that these costs were small. More recent contributions are showing that these costs are rising rapidly. The implication is that both the rate of economic growth and the prospects for future growth have been compromised.¹³

Since the losses from HIV/AIDS are cumulative and data on the effects of the impact of the epidemic are unreliable (see Annex A), the overall impact of HIV/AIDS on the labor force has been subject to considerable uncertainty.¹⁴ As noted in the discussion below, an important source of that uncertainty has arisen because analysts have tended to ignore critical aspects of the underlying dynamics when calculating the costs of HIV/AIDS. A major focus in the literature has been work place prevention. Far less attention has been given to the types of behavioral, social, and structural changes needed within the work place to prevent productivity and growth from falling.¹⁵

This has occurred despite important efforts to identify and advertise "best practices" for dealing with HIV/AIDS.¹⁶ As Annex A shows, such practices have not been widely used. The reality is that, for many organizations and enterprises (and one might argue most of the countries of Southern Africa), the epidemic has elicited few constructive responses let alone the adoption of "best practices." Indeed, a major gap in the research (and practice) has been how to deal with a large and growing number of workers who know (or suspect) that their lives are being dramatically foreshortened.

Under conditions of high AIDS prevalence, as in Africa, conventional approaches to training, managing, organizing, and motivating workers need to be fundamentally re-thought.¹⁷ There is little to suggest that HIV-positive trainees or workers should respond (or can be expected to respond) in the same way as those with a "normal" life expectancy. Moreover, the prospect of large losses of skilled workers from HIV/AIDS makes the value of long-term degree-based training problematic. One question that has not been answered is whether donor agencies are prepared to continue supporting long-term training in the face of rising attrition rates?

For employers, a key issue is the type of (cost-effective) incentives they can devise to induce HIV-positive workers to maintain their productivity. Faced with high and rising incidence of HIV/AIDS, no one should be surprised that employers (alone and collectively) are taking measures (often draconian) to contain their costs.¹⁸

One of the purposes of the review in Annex A is to show how organizations and businesses have sought to maintain productivity (and profitability) when significant numbers of workers will become debilitated and die. Such measures include the "externalization" of costs, the use of expatriate labor, and forced retirement of infected workers. These measures often reflect attempts by employers to shift the costs of HIV/AIDS onto the infected worker's family or onto society at large. Such practices will undoubtedly continue. Nonetheless, it should be recognized that employers are constrained in their ability to shift these costs. Some employers can succeed in the short-run. But, as the HIV/AIDS epidemic intensifies, the macroeconomic effects of these rising costs will feedback to affect the employers' output, sales, taxes, or access to social services.¹⁹

That context is missing from most analyses of the effect of HIV/AIDS on economic growth. The most common approach has been for analysts to provide a model of what they believe is the present structure of an economy. Then, based on their best estimates of trends in the incidence of HIV/AIDS, they derive alternative growth paths reflecting situations 'with' and 'without' HIV/AIDS. Kambou, Devarajan and Over, in one of the earliest examples of this approach, studied the impact of HIV/AIDS in Cameroon.²⁰ Using a Computable General Equilibrium (CGE) model, they derived two sets of projections. The first was a base case that abstracted from the effects of HIV/AIDS. The second allows for those effects. They did that by reducing the supply of labor by an estimate of what the losses to HIV/AIDS were expected to be over the projection period.

There are several reasons for this, but two stand out. First, Kambou et al. could not have foreseen that new, virulent sub-types of the disease would emerge.²¹ Second, using a CGE model to derive the alternative scenarios, 'with' and 'without' AIDS, is inappropriate. As the HIV/AIDS epidemic has intensified, economies have experienced substantial structural change.²² This does not get adequately reflected in CGE models (or other fixed coefficient models) upon which the 'with AIDS' scenario was based.²³ Kambou et al. (and other analysts who have followed their lead) did not modify their basic model to reflect these changes.²⁴

When their research was undertaken (1990), such an approach could be defended as a "rough first cut." At that time, the spread of HIV/AIDS was just beginning to accelerate. What cannot be defended, however, was the continued use of such comparative scenarios. This point, however, has not been widely appreciated. For example, the Sunday Times in South Africa reported that in mid-1997 the impact of the HIV/AIDS epidemic in 2005 would be to reduce overall growth rates of the economy by 1 percent per annum.²⁵ This projection has startling implications. It means that, with more than 20 percent of the adult population infected with HIV and likely to die within five years, the impact on South Africa's economic growth is expected to be incidental. Clearly, something is amiss.

Projection errors like these would be significantly less consequential if officials and the general public across Africa were more forthcoming about the extent of the epidemic. Indeed, with very few exceptions, the official and private response has been denial.²⁶ Events, however, may be changing that. During the last week of May 2000, viewers of. Cable News Network around the world saw Thabo Mbeki, President of South Africa, at the time on a visit to the United States, attempting to justify his reluctance to face the reality of HIV/AIDS in his country.²⁷ Whether Mbeki has come to grips with the problems South Africa is facing or not, others have. The theme of the Durban conference in July 2000 was "breaking the silence." Mbeki gave a keynote speech. His predecessor, Nelson Mandela, whose tenure was notable for its absence of any serious effort to confront the issue, has at last noted that history will judge harshly those who sit idly by. History so far has shown that until mid-2000, both Mandela and Mbeki ducked the issue. In the interim, South Africa (and Southern Africa in general) has borne a terrible cost. Millions of people have died and millions more will die because of their leaders' silence and/or ambivalence.²⁸ Their inaction or, at best, lukewarm response has killed large numbers of their own citizens.²⁹

Unfortunately, there is nothing new about denial and limited recollections of (traumatic) events. History has many examples of societies and groups that exhibit such behavior. Wars and organized mayhem have been particularly fertile periods for generating selective social amnesia. Yet, in most cases, the denial relates to events. By contrast, the HIV epidemic is a process. Rather than dulling its impact, denial allows the epidemic to intensify.³⁰

Scholars and practitioners dealing with HIV/AIDS on a day-to-day basis understand the folly of denial. Yet, there are many reasons why it has persisted. For example, as noted in Annex A, governments were reluctant to openly deal with HIV/AIDS because officials argued it would adversely affect tourism and foreign investment. Others have boxed themselves in by years of silence or cant. African journalists, for example, who failed to write candidly about the disease, and leaders who avoided the issue, face a dilemma. President Chiluba of Zambia provides an example. Though the conference was held in Lusaka, Chiluba failed to attend the "International Conference on AIDS and Sexually Transmitted Diseases in Africa" in September 1999. From his inauguration as president until the time of that conference, more than 500,000 Zambian children have become AIDS orphans.³¹ Attending the conference risked advertising his lack of attention to the toll HIV/AIDS has had and was having on the citizens whose welfare he had sworn to protect. Though his absence from the conference received little attention in Zambia (where observers understood it was consistent with his past behavior), the foreign press took note.³²

Donor agencies, too, have been caught in a web of denial. Although they generously support work by health ministries on prevention and treatment of HIV/AIDS, most donor agencies have failed to scale back their "development" agendas in line with the erosion of skills and institutional capacity in Africa as a result of HIV/AIDS. For example, both the World Bank and the International Monetary Fund have proceeded with their sector investment programs, comprehensive development frameworks, poverty reduction and growth strategies, as though the capacities of African governments remained unimpaired.³³ This directly contradicts the message the World Bank conveyed in its 1999 study Confronting AIDS. That study made referred to the World Development Report of 1997 where it was suggested that developing countries should ensure their development programs were consistent with "State capacity."³⁴

African countries themselves have compounded these problems. Though they lack the relevant capacity to implement the donor-driven agendas, their key policy makers continue to accede to the various World Bank, IMF (and other donor) "initiatives." The (predictable) result has been that most donor-supported adjustment programs are pre-programmed to fail.³⁵

A Conceptual Framework

The basic framework adopted here for thinking about the dynamic impact of HIV/AIDS on economic growth draws on models of economic retrogression. These models were developed in the early 1990s³⁶ to examine the effects of rising levels of corruption and inefficiency on the growth paths of developing countries. The objective was to explain why a large number of countries in Africa had been exhibiting "patterns of growth" that did not conform to the (so-called) "normal" time paths derived by Kuznets and Chenery and his colleagues.³⁷ Economic regression became so pronounced in some countries (such as Zambia, Tanzania, and Senegal) that the relative contribution to GDP of agriculture had risen over time.³⁸

Few of the plausible explanations for retrogression of this nature were 'exogenous.' (Drought "might" be seen as an exception.)³⁹ To capture these effects, the principal features of a regressing economy were modeled by reversing the direction of an endogenous growth model. Corruption, measured as an increase in the unlawful "diversion" of public resources, reduced the investible surplus. As corruption intensified, public confidence declined. There was an increase in capital flight (and other defensive actions by asset holders). This led to further reductions in investment. That, in turn, reduced the growth rate of income and lowered savings and investment even further. The model also incorporated a "backlash" feature that hinged on the reactions of groups most seriously affected by resource diversion. The same effect is produced in outside agents, e.g., donor organizations began forcing the governments they were supporting to observe principles of 'good governance.' This helped boost investor confidence, raising the rate of investment and helping spur income growth. Such a diversion/backlash mechanism generated a wide range of time paths for retrogression and recovery.

The above framework can be readily adapted for examining the dynamic impact of HIV/AIDS on economic growth. The rising prevalence of HIV/AIDS lowers worker efficiency, raises costs, and reduces individual savings and firms' profits. Individuals who are HIV-positive increase their consumption, in part to combat the effects of the disease and, in part, because the prospect of a premature death raises the opportunity cost of time.⁴⁰ These changes lower the supply of investible resources, at both the individual and national levels, and reduce the efficiency with which the existing stock of productive assets is used. Those effects, in turn, lower the rate of growth of per capita incomes, setting off a further cycle of declining savings and investment.

One could imagine that such a process might have a "floor" that provides the basis for economic recovery if measures were devised to prevent the overall efficiency of the labor force from declining. Several mechanisms might suffice. One would be to lower the rate of infection of workers through reductions in the costs of combating HIV/AIDS.⁴¹ Another would be to manage, train, and motivate workers who are HIV-positive in ways that maintain (and even improve) their productivity.⁴² A third would be to use external resources (foreign aid and technical assistance) to compensate for the decline in the local supply of investible resources and loss of domestic skills.

Endogenous Growth: Theories of endogenous growth derive their inspiration from a number of sources.⁴³ The principal idea is that increasing returns to scale (resulting largely from the accumulation of knowledge) boosts the rates of savings and investment. The higher rate of investment leads to an acceleration of the rate of economic growth. There are several "knowledge-based" explanations that are used to support this theory. These include "learning-by-doing", ⁴⁴ "induced innovation", ⁴⁵ the effects of rising population, ⁴⁶ and "learning-by-trading."⁴⁷ In many African countries, particularly those that have had poor economic performance over extended periods, there is also the prospect of "learning-by-reforming," most notably during the transition period.⁴⁸

Whether the resulting growth path involves a "permanent" acceleration of the rate of growth or cycles in the rate of growth (e.g., as in the so-called "new economy") is still subject to debate. The essence of endogenous growth is that there are "spread" effects that raise efficiency more generally in the economy. Adam Smith captured this idea when he described the advantages of the 'division of labor' that was earlier stimulated by the 'expansion of the market.'⁴⁹ The efficiencies resulting from specialization and the division of labor lowered costs thereby further expanding the market. In more recent times, Gunnar Myrdal described this as a process of "cumulative causation."⁵⁰ The basic idea is that existing economic activities benefit from the fact that improving efficiency in other enterprises help to lower their costs. The overall effect is for all enterprises whose costs are falling to take action to expand their output. This sets off another round of spillover effects.

Falling Savings and Investment: The basic problem with attempts to tease out an independent influence for the impact of HIV/AIDS is that the economic performance of many African countries had deteriorated even before HIV/AIDS began to spread.⁵¹ For most African countries, the decline began with the oil and food "shocks" of the mid-1970s. Because of limited internal adjustment and frequent policy reversals, African countries were already heavily in debt and in precarious financial circumstances by the early 1980s. The HIV/AIDS epidemic compounded their problems. The difficulties were evident in adverse trends in key macroeconomic data. The following tables show that for most countries real income was declining, rates of investment and savings were falling, foreign aid flows rose rapidly, and there was a sharp increase in the debt burden. The data in the table covers the period 1970 to 1998 for SSA excluding South Africa and Nigeria.⁵²

The data show that average real per capita income was lower in the 1990s than in the 1970s. Savings rates have declined over time. Investment has fallen by less than savings but only because of a major increase in foreign aid and foreign borrowing. The size and duration of these external flows has been unprecedented. No region of the world has received such extensive support for so long. Yet, even these large and persistent resource flows have been unable to maintain investment at levels that will raise national income on a sustained basis.⁵³ The severity of the economic stress has compromised recovery efforts. The spread of HIV/AIDS has further aggravated the situation.

In this regard, it is useful to recall Harry Johnson's characterization of economic development as a "generalized process of capital accumulation" where capital is broadly defined to include physical and human capital as well as institutions and organizations.⁵⁴

Sub-Saharan Africa: Selected Macroeconomic Indicators, 1970-1998

Year

GDP per capita (1995 USD)

Gross Dom. Invest. (share of GDP) (%)

Gross Dom. Savings (share of GDP) (%)

Net ODA (all donors)(bill. USD)

Net ODA (all donors)share of GDP (%)

Total External Debt (ratio to GDP (%))

1970

612.0

17.3

21.1

1.1

1.6

..

1971

634.0

18.0

20.2

1.2

1.8

14.7

1972

623.3

16.8

22.2

1.3

1.7

15.0

1973

642.2

19.1

25.2

1.6

1.7

15.2

1974

681.8

21.0

28.1

2.4

1.9

14.5

1975

670.3

22.6

23.4

3.3

2.3

15.6

1976

669.5

23.6

25.5

3.1

1.9

16.2

1977

649.6

20.6

26.4

3.6

2.1

21.1

1978

641.0

20.4

24.1

4.9

2.6

24.8

1979

644.7

18.4

25.5

6.3

2.7

24.0

avg. 70s

646.8

19.8

24.2

2.9

2.0

17.9 *

1980

660.9

20.2

28.5

7.4

2.5

22.9

1981

671.7

22.8

23.2

7.3

2.5

26.4

1982

648.0

19.8

21.1

7.5

2.8

31.5

1983

617.3

16.2

20.5

7.3

2.9

37.1

1984

623.0

15.5

21.0

7.6

3.3

41.4

1985

601.6

13.8

21.2

8.5

4.1

54.2

1986

595.4

14.6

20.9

10.5

4.6

56.2

1987

591.8

14.6

20.5

12.1

4.8

58.6

1988

599.2

16.5

19.2

13.7

5.2

56.9

1989

599.0

15.8

19.4

14.5

5.4

58.3

avg. 80s

620.8

17.0

21.5

9.6

3.8

44.4

1990

587.2

14.2

17.9

17.3

5.8

59.7

1991

573.5

17.0

18.2

17.0

5.6

60.6

1992

550.2

14.7

13.0

18.3

5.8

60.0

1993

542.6

16.2

14.5

16.8

5.5

65.5

1994

541.5

17.5

16.5

18.2

6.3

78.1

1995

548.5

18.5

16.2

17.9

5.6

73.6

1996

558.8

17.7

17.7

15.7

4.7

69.5

1997

561.9

17.4

16.4

14.2

4.1

63.6

1998

558.2

17.8

14.8

..

..

54.5

avg. 90s

558.0

16.8

16.1

16.9 *

5.4 *

65.0

Notes: * - Average for the years with observations

Source: World Bank Africa 2000 CD-ROM Database

Sub-Saharan Africa excl. South Africa and Nigeria: Selected Macroeconomic Indicators, 1970-1998

Year

GDP per capita (1995 USD)

Gross Dom. Invest. (share of GDP) (%)

Gross Dom. Savings (share of GDP) (%)

Net ODA (all donors)(bill. USD)

Net ODA (all donors)share of GDP (%)

Total External Debt (ratio to GDP (%))

1970

334.6

15.0

24.9

0.9

3.4

..

1971

344.7

15.7

22.3

1.1

3.6

25.3

1972

347.6

16.1

23.7

1.2

3.7

25.9

1973

349.5

16.9

25.2

1.5

3.8

26.6

1974

363.5

19.4

26.9

2.4

4.9

28.1

1975

362.6

20.2

23.0

3.2

5.6

29.8

1976

372.3

19.9

25.5

3.1

4.8

32.1

1977

370.0

19.5

24.5

3.6

5.0

38.4

1978

367.1

19.1

22.1

4.9

6.0

42.6

1979

366.2

17.6

20.5

6.3

6.6

42.7

avg. 70s

357.8

17.9

23.9

2.8

4.7

32.4 *

1980

357.4

18.1

20.4

7.4

6.5

45.4

1981

362.0

18.8

18.6

7.3

6.3

50.7

1982

359.8

17.7

19.4

7.5

6.7

57.4

1983

353.4

15.3

19.2

7.2

6.9

64.0

1984

348.7

14.2

20.7

7.5

6.8

67.2

1985

346.9

14.7

20.8

8.5

7.3

78.7

1986

349.4

15.2

20.6

10.4

7.7

75.8

1987

347.0

15.4

19.2

12.0

8.4

82.6

1988

349.2

16.7

16.5

13.6

9.1

80.5

1989

348.3

15.3

15.9

14.2

9.5

84.4

avg. 80s

352.2

16.1

19.1

9.6

7.5

68.7

1990

341.5

16.0

15.0

17.0

11.0

91.7

1991

336.4

20.1

16.6

16.7

10.8

96.7

1992

322.7

15.7

7.0

18.0

12.5

109.2

1993

316.3

17.3

10.1

16.3

11.4

120.8

1994

314.1

19.3

14.3

17.7

14.3

135.4

1995

322.6

19.3

12.6

17.3

12.5

126.0

1996

330.9

20.2

13.3

15.1

10.0

113.5

1997

335.5

19.6

14.4

13.5

8.6

105.4

1998

338.3

19.3

13.3

..

..

..

avg. 90s

328.7

18.5

13.0

16.5 *

11.4 *

112.4 *

Notes: * - Average for the years with observations

Source: World Bank Africa 2000 CD-ROM Database

Taken together, the notions of irreversibility and options values are crucial considerations when both locals and foreigners contemplate investing in African countries. Investors are more likely to wait when they have information indicating that the spread of HIV/AIDS will affect adversely the cost structure of any investment they are contemplating. Under these circumstances, waiting provides time to reassess and re-evaluate their options. Having a low-risk, secure (foreign) alternative investment reduces the urgency of committing their resources.⁵⁵ The outcome has been devastating in terms of stimulating new activity across Africa. As the perceived costs of dealing with the spread of HIV/AIDS rises, the rate of investment tends to decline. This has reinforced the decline in the supply of investible resources, already under pressure through falling productivity due to the spread of HIV/AIDS.

Rising Opportunity Cost of Time: One microeconomic mechanism that has reinforced the decline in savings has been the diminished incentive for those who are HIV-positive to save. Although many people who are HIV-positive also have lower incomes from which to save, especially as they become increasingly debilitated, the two effects need to be analyzed separately. The prospect of a premature death raises the opportunity cost of time. There is now a wealth of research, dating back to Irving Fisher, tracing the systematic changes that occur in the ratio of consumption to saving as patterns of time preference change.⁵⁶ For risk averse individuals, the standard result is that time preference is inversely proportional to the length of the decision horizon. The relevant decision horizon for people with HIV/AIDS shortens dramatically. As noted in Annex B, it is normally five to seven years from the diagnosis of HIV to the onset of AIDS. Death comes within one or two years after that.

A similar response can be derived from the 'life-cycle' model of savings.⁵⁷ The underlying theory recognizes that there is a systematic shift in the relationship between a worker's earning potential and the pattern of family formation over his/her life cycle. Early in a person's working life, income streams and resource demands are rising. With the expectation of a "normal" life span of work, the person has the capacity to finance his/her consumption and asset accumulation goals based on the expectation of higher income flows later in his/her productive life. The spread of HIV/AIDS, however, has sharply reduced what individuals can "normally expect" from their (truncated) life span of earnings. Once an individual is HIV-positive, his/her consumption rises while at the same time his/her potential future earnings stream is cut. Thus, when viewed in terms of the 'life-cycle' model, we would expect to see a major reduction in individual savings. To the extent that the individual has accumulated wealth during his/her working life, there will be dissaving.⁵⁸

Much the same result can be derived from the notion of 'overlapping generations.' The typical approach is to assume two generations - one generation produces, the other consumes. (The effect is the same as a continuous two-part 'life-cycle'.) In this formulation, accumulation and growth occurs because the output (or income) produced by the generation of producers exceeds the output absorbed (or expenditure) of the generation of consumers. With the spread of HIV/AIDS, this balance between production and consumption shifts in ways that reduce this inter-generational surplus, thereby reducing accumulation. This effect is reinforced as the dependency ratio (the ratio of workers to persons under 15 years of age) shifts. A major feature of the HIV/AIDS epidemic has been the sharp rise in the number of orphans.⁵⁹ The implication is that the extent to which the generations overlap has undergone a serious, systematic, and adverse change. The net outcome is to reduce the capacity of economies with a high prevalence of HIV/AIDS to save and invest, and ultimately to accumulate and grow.

If finance could be arranged (for example, through the efficient use of foreign assistance), the strategic use of new technology would allow countries to compensate for the loss of labor due to HIV/AIDS. The obvious drawback to such a strategy is that it takes resources to acquire new technologies and skilled personnel to operate them. Though a drawback, this should not be an insurmountable problem, particularly if attention is given to devising goal-oriented practical training.⁶⁰

A further element reducing labor productivity is the reduction in real effective demand associated with the decline in real per capita income. As already noted, real incomes across Africa were declining well before the onset of the HIV/AIDS epidemic.⁶¹ Adding HIV/AIDS to the equation has reinforced the decline. This has occurred through a mechanism described by Adam Smith over two centuries ago. But, while Smith described the benefits (to output, productivity, and growth) of expanding the "extent of the market," the circumstances in Africa have been equivalent to running the mechanism he described in reverse. The decline in real effective demand raises the unit costs of all domestic activities. For African countries where infrastructure is poorly developed, one obvious change is an increase in transport and distribution costs as a proportion of total costs. The relevance of this point becomes increasingly evident when one distinguishes, as T.W. Schultz did, between the economic and the physical supplies of land.⁶² The latter is the land base of a country or region. The former is the land area that, for given technology and market conditions, yields a positive net rent. Declining real effective demand reduces the net rent from all economic activities and by extension reduces the productivity of all factors including the labor that contributes to these activities. The outcome is contraction in the economic supply of land. Furthermore, since land is an important component of wealth in African countries, the declining economic supply of land reduces the value of wealth. This leads to a further contraction in real effective demand, further lowering the productivity of associated factors of production.

Other factors contribute to the decline in labor productivity. The theory of efficiency wages is based on the recognition that, because of fixed costs of employment (hiring, training, settling-in), firms will have an incentive to pay above-market wages in order to keep their employees. A further aspect of the theory is that firms will pay workers above-market wages because of the direct link between wages and worker productivity.⁶³ Annex A reviews what employers have been doing in Southern Africa through training and reorganization of work schedules to maintain productivity. Annex B explains the incentives employers have to providing training.⁶⁴

In these matters, one can readily imagine that both these processes work in reverse as the HIV/AIDS epidemic intensifies. Because of increasing debility and absenteeism, the marginal value product of workers who are HIV-positive is less than the wage being paid by the employer. Faced with this situation, firms have an incentive to reduce employment and/or take steps to reduce labor costs. Doing this, however, reduces the incomes of all their employees whether they are HIV-positive or not.⁶⁵ To the extent that the efficiency theory of wages holds, overall efficiency declines.