1. Introduction

The aim of the World Bank, often referred to in development circles simply as “the Bank”, is “to help each developing country onto a path of stable, sustainable and equitable growth in the fight against poverty”.² Financial services are its core activity, but in support of these the Bank undertakes various analytical and advisory services, and capacity-building. The work of the Bank's health sector—or more precisely, its health, nutrition and population (HNP) sector—is guided by the Bank's broad mission of poverty reduction and by the specific objectives set out in its recent health sector strategy paper.³ (The use of the term “health” in this paper should be interpreted broadly to include nutrition and population issues.) The two overarching objectives outlined there are: (a) to help countries improve the health of the poor, and (b) to help them reduce the impoverishing effects of illness. (The sector has two subsidiary objectives, but these are, in essence, means to achieving the two objectives outlined above. The first subsidiary objective is to enhance the performance of health care systems by promoting equitable access to preventive and curative health, nutrition and population services that are affordable, effective, well managed, of good quality, and responsive to clients. The second is to secure sustainable health care financing by mobilising adequate levels of resources, establishing broad-based risk-pooling mechanisms, and maintaining effective control over public and private expenditure.) Ethical issues arise in the interpretation of both objectives and in helping countries formulate strategies and policies geared towards them. It is these ethical issues—which are often not acknowledged by commentators—that are the subject of this paper. Before addressing them, however, the paper provides a short background section on what and how much the Bank does—in general and in the health sector specifically.

2. What the World Bank does

Chief amongst the Bank's financial services are loans and credits. In general, the poorer the country, the more generous the terms of any loan. Very poor creditworthy countries (in practice, those with a per capita income of $885 or less) are eligible for long term interest-free credits under the International Development Assistance (IDA) scheme. Countries with average incomes between $885 and $1,445 are, in principle, eligible for IDA credits, but in practice, because IDA funds are limited, end up borrowing on less favourable terms. These countries are referred to as “IDA-blend”, the terms on which they borrow being a mix of IDA and those of the Bank's other main scheme—that of the International Bank for Reconstruction and Development (IBRD). Loans from the IBRD carry a positive interest rate but one that is below the rate at which a private bank would lend. International Bank for Reconstruction and Development loans are available to creditworthy countries with a per capita income below $5,225. When countries reach this threshold, the process of “graduating” from the IBRD scheme is triggered. In its IBRD loans, the Bank borrows at preferential interest rates and then on-lends to the borrowing country at a rate that is somewhere between the rate the Bank borrows at and the market rate. International Development Assistance credits are financed in part through the income earned on IBRD loans, and in part through contributions from bank members. Debt relief to IDA countries under the enhanced highly indebted poor country (HIPC) initiative, for example, is financed largely though contributions by the OECD countries to a special HIPC trust fund.

Over the years, overall Bank lending has increased in real terms, though the last few years have seen a downturn (see figure 1). Over the same period, health lending has risen substantially as a proportion of total lending, starting from nothing in the 1960s and rising to 5-6% in the late 1990s (see figure 1). Nearly 60% of bank lending in the health sector is through IDA (see figure 2), but the IDA-IBRD mix varies substantially across Bank regions. International Development Assistance loans make up the entire health portfolio in Sub-Saharan Africa and South Asia, a significant part of the health portfolios of East Asia and the Pacific, and only small parts of the health portfolios of the Middle East and North Africa and Latin America and Caribbean regions (see figure 2).

3. Focusing on the poor

The focus in the Bank's health sector work on the poor and on impoverishment raises a number of ethical issues.

3.1. THE HEALTH OF THE POOR

The obvious way to interpret the Bank's first objective—to improve the health outcomes of the poor—is that a greater weight should be accorded to the health of the poor than to the health of the better-off when choosing between alternative Bank projects. This means taking into account not just the average health improvement associated with a particular project but also the degree to which health improvements are proportionately larger for the poor than for the better-off. It does not mean simply trying to reduce health inequalities between poor and rich. That would imply a complete unwillingness to trade-off the overall average level of health against the level of inequality, which is unlikely to command the support of any right-minded policy-maker. It would, for example, imply rejecting all inequality-increasing policies however small the rise in inequality and however large the rise in the overall average level of health.

This does beg, however, an obvious question, namely how much higher one should weight the health improvements of the poor than those of the better-off. One approach^{4 5} is to think of the problem as one of constructing a distributionally-sensitive measure of population health. The mean is clearly not appropriate, since it weights everyone's health equally, irrespective of how poor they are. One possible set of weights is the person's rank in the income distribution, or some simple function of it. One such scheme is to assign the poorest person a weight of 2 and then let the weight decline by 2/N (where N is the sample size) for each one-person step up the income distribution. Adopting this set of weights produces a distributionally-sensitive measure of population health that is simply equal to the mean level of health of the population times the complement of the concentration index.^{6 7} The latter is a widely used index,^{4, 8–10} that captures the extent to which ill health is concentrated amongst the poor. A policy that resulted in the same proportional improvement in everyone's health would raise the value of the distributionally-sensitive measure of population health, while a policy that led to the same increase in the mean but a larger (smaller) proportional improvement in the health of the poor would produce a larger (smaller) increase in it.

This hinges, of course, on the particular set of weights chosen. But it can be generalised by introducing a parameter indicating the degree of aversion to inequalities in health between the poor and the better-off.¹¹ If one proceeds along these lines, the distributionally-sensitive measure of population health becomes the mean times the complement of the so-called generalised concentration index. The more averse the policy-maker is to health inequalities between the poor and the better-off, or equivalently the bigger the weight the policy-maker wanted to attach to the health of the poor, the more the distributionally-sensitive measure of population health focuses on the health of the poor. In the extreme, the distributionally-sensitive measure of population health reduces simply to the health of the poor, or the poorest group. But this would clearly be a very extreme position to take.

3.2. WHY FOCUS ON THE POOR?

The discussion above begs a key question, namely: why should the Bank focus on improving the health of the poor? Why should it not aim simply to help countries improve the health of their populations? Concerns about equity and justice ought, it might be argued, to be more appropriately tackled either by undertaking to reduce health inequalities across people (whatever their income), or by undertaking to focus on those whose health is worst (irrespective of whether they are poor or rich).

This argument has been made recently by staff of the World Health Organization (WHO),¹² who have argued that whilst health inequalities matter, what matters is the level of inequality between individuals, however poor or rich they happen to be. This view is not inconsistent, in fact, with the view expressed in the Bank's own recent world development report (WDR) on poverty,¹³ which argues that poverty is multidimensional and that people can be “poor” simply by virtue of suffering from bad health, whether or not they happen to be badly off in income terms.

An argument against the view that all health inequalities are equally bad or equally unjust was mounted some years ago by Julian Le Grand¹⁴ and echoed recently by George Alleyne et al.¹⁵ Le Grand argues that inequalities in health are not automatically unjust. They are unjust in so far as they reflect differences in the constraints that people face, but are not unjust if they are the result of people making different choices under the same constraints. What this suggests is that in so far as inequalities in health between poor people and well-off people are due to differences in the constraints they face, rather than to, say, the poor placing a lower value on good health, we can reasonably label these inequalities as unjust.

Income and assets are, of course, one reason why constraints differ between the poor and the better-off. But they are but two. Poor and better-off households may also incur different costs when trying to restore and maintain their health. Health facilities in the developing world vary hugely in their quality. Some have medicines and drugs in stock, are run by well-trained, civil and motivated staff, are well maintained and are easily accessible. But many are not. They are often dilapidated and inaccessible, rarely have medicines in stock, and are run by poorly trained and rude medical staff, who frequently fail to turn up to work because they are too busy running their private practice (often selling drugs “borrowed” from their public facility). What emerges from the Bank's Voices of the Poor consultative exercise,¹⁶ as well as from quantitative studies, is that it is precisely the people who are materially disadvantaged who have to struggle with poor-quality and inaccessible health facilities and many other factors that tighten even further the constraints facing a poor household.

What this suggests is that the inequity of health inequalities between the poor and the better-off stems not simply from the income gaps between them but also from the gaps in the effective “prices” they face when maintaining and improving their health.

3.3. REDUCING THE IMPOVERISHING EFFECTS OF ILL HEALTH

Putting into operation the second of the Bank's health sector objectives—reducing the impoverishing effects of ill health—also presents ethical issues. The concern is with the impact of medical care costs and lost earnings on a household's ability to purchase things other than medical care. In other words, in addition to the desire to ensure that health improvements occur (especially among the poor), there is a desire to ensure that this is not at the expense of excessive drops in the living standards of the households involved.

There are two ways one might interpret this objective. One is that the costs associated with ill health should not drive households into poverty, or drive them further into poverty if they are already there. In other words, the distribution and size of medical care costs and lost income should not be such as to raise the proportion of households or people in poverty (the headcount) or to increase the average depth of poverty (the poverty gap). The impoverishing effects of health care expenses, so defined, can be quantified.¹⁷ An alternative interpretation is that the distribution of costs should not be such as to increase the degree of income inequality. In other words, there should not be more inequality “after” health care payments than there was “before”. Regressive payments (ie payments that absorb a larger share of a poor household's income than of a rich household's) would violate this requirement. The inequality impact, or the so-called “redistributive effect”, of health care payments can be computed.¹⁸ There is, of course, no right answer to the question of whether one should focus on the effect of health care spending on poverty or on its effect on income inequality. Both are legitimate concerns, but they are different concerns.

4. Difficult choices at the global level

As has been seen, the work of the Bank in the health sector takes many forms. In the case of IDA credits and debt relief, resources are limited, at least in the short run, and difficult choices have to be made in deciding between competing claims.

4.1. DEBT RELIEF

Suppose there genuinely is a chance that at least part of the debts being considered for debt relief would, in the absence of debt relief, be repaid. Then any funds used to write off the debts could be used for other purposes, including development assistance to countries that have not accumulated debts. One question that arises, therefore, but which typically gets overlooked in the highly charged debt-relief debate, is whether countries that have accumulated debt should be favoured, as they are under the HIPC initiative, over equally poor countries that have not.

Another question is how to strike a balance between ex ante and ex post considerations. An argument against too generous debt relief is that it may create a “moral hazard” problem in that it will send the wrong signal to borrowing countries, so that in the future they may be less likely to concern themselves with keeping up with loan repayments. The result may be that the international community may be confronted by another debt problem in the years to come.

4.2. WHAT TO DO WITH COUNTRIES WITH BAD GOVERNANCE?

Another key issue which arises in the context of debt relief but also more generally in the context of aid, is how far the likely impact of aid should be taken into account in deciding on the amount of aid to be allocated. There is clear evidence that the impact of a dollar's worth of aid on growth and poverty reduction depends crucially on a number of factors, key among which are the policies being pursued by the recipient of the aid, the quality of governance in the country, and the degree to which the recipient country feels it “owns” its development process rather than having it managed and directed by outsiders such as the Bank.¹⁹ Allocating aid on the basis of the expected productivity of aid might well widen the inequalities between countries, in just the same way as allocating health care between patients on the basis of the ability of care to bring about health improvements can exacerbate health inequalities.²⁰ And yet, not to take into account the productivity of aid would result in worse outcomes than would otherwise be the case.

This is complicated by the fact that the productivity of aid is not fixed. The best form of assistance for countries where the circumstances are such that traditional aid is unlikely to be especially productive is likely to be capacity-building—training civil servants, strengthening the legal system, reducing corruption, and so on. This applies as much to the health sector as to the economy generally.

4.3. OWNERSHIP AND DONOR COMMITMENT

The issue of “ownership” raises the possibility of another potential trade-off. If countries develop more quickly and aid is more effective when countries “own” the development process, this points towards downplaying individual projects, financed and overseen by specific donors, and shifting instead towards a development strategy that involves recipient governments producing medium-term development and poverty-reduction plans and donors pooling their funds in a single pool from which the recipient government can draw without having to seek approval on a project-by-project basis. This is the logic underlying the sector-wide approach (SWAP), programmatic lending, and the use of poverty-reduction strategy papers (PRSPs) in the enhanced HIPC initiative. It is also presumably the logic of European governments pooling at least part of the development aid through the European Union's (EU) development programme.

There is, however, a trade-off here, since as donors pool their funds their development assistance starts to take on a public good character, and the willingness of tax-payers in individual donor countries may wane. If, as is increasingly the case, aid is not tied in the sense that recipient countries are required to purchase goods and services from the donor as part of the aid package,²¹ the benefits to the donor of the aid are non-pecuniary. The motivation is humanitarian and the benefits to the tax-payer in the donor country are similar to those that accrue when the tax-payer gives to a charity such as Oxfam or Save the Children. The difficulty that charities face is the same that donors face in the aid context. What donors benefit from is not the giving itself but the result of the giving—the saving of young children's lives, the reduction in levels of malnutrition, or whatever. The difficulty is that people can benefit psychologically from, say, an Oxfam project that reduces levels of child mortality and malnutrition regardless of how much they themselves give to Oxfam. There is potentially a free-rider problem—donors may rely on other people's contributions safe in the knowledge that they will continue to benefit from the charity's work.²²

The same problem could apply to donors pooling aid at the national level, leading potentially to a decline in the willingness of tax-payers in donor countries to support overseas aid. This will become even more likely if donors and tax-payers in donor countries consider that the benefits associated with pooling (greater country ownership and hence greater development effectiveness) are offset, if only in part, by increased costs and/or reduced efficiency in the process of disbursing funds. This tendency is evident in the UK, where the present secretary of state for international development, Claire Short, has expressed dissatisfaction on a number of occasions with the EU's development programme.

Whether the move towards pooled aid has indeed exacerbated the free-rider problem is hard to say. It is, however, noteworthy, though quite possibly coincidental, that over the period when the pooling of aid became more common, funds for development assistance fell.¹³

5. Difficult choices at the country level

Within countries—whether IBRD or IDA countries—difficult choices arise in choosing between projects and programmes, even if decisions are guided by the broad objectives outlined in section 3.

5.1. USER FEES AND COST RECOVERY

On the face of it, the issue of cost recovery is, from an ethical perspective, an open-and-shut case. User fees deter patients, especially poor ones. And where they do not, they end up absorbing a larger share of the income of a poor household than of a rich household (ie they are regressive). If the aims are to improve the health of the poor and to prevent households from becoming impoverished through ill health, then clearly user fees are to be avoided.

Unfortunately, in practice, things aren't so straightforward. First, one needs to take into account the full cost of health services not just the money cost. Free health services are often associated with long queues, which can often translate into lost income (for example, agricultural workers). Health services also entail transport costs. If revenues from user fees allowed queues to be cut and new facilities to be opened, the overall cost (the lost income and the transport cost) to the poor might actually be reduced. One also needs to take into account the quality of services. In many settings, this is very low—so much so that even the poor choose to bypass free public facilities in favour of facilities where some money payment is involved. For example, free rural clinics often have no drugs in stock and even the poor choose to bypass them and go directly to clinics known to be properly stocked, even if this means further to travel and a higher money payment. If user fees were introduced and the revenues could be used to improve quality, there might well be an improvement in the health of the poor with only a modest increase—if any—in the overall costs associated with ill health. It may well also be the case that user fees could improve accountability and give users a greater voice in the decision making process.

A difficulty with this is that it has proved difficult to link user fees to household income. Exemption schemes have often benefited groups who are easy to identify but who are often not poor, such as civil servants and the military. There are, of course, potential alternative financing sources, such as taxes, social insurance and private insurance. But, in practice, in poor countries these are often not well developed. Social insurance is limited because of the relatively undeveloped formal sector. Private insurance is often non-existent, and would in any case suffer from the well-known equity and efficiency problems. Tax systems are often weak and there would be a good deal of doubt about whether extra taxes would ever translate into improvements in quality in public facilities. All this makes the choice between user fees and other forms of revenue far less straightforward from an equity perspective than it is in the industrialised world.

5.2. WHAT ROLE FOR THE PRIVATE SECTOR?

Another example of an issue that is often far less cut-and-dried, from an ethical perspective, than is often thought to be the case, concerns the appropriate role of the public sector in the provision of health care.

The equity and efficiency issues debated by economists in the market-versus-state debate point towards the government needing to play a key role in the finance of health care but are not usually considered to point decisively to the government provision of health care.²³ In practice, much of the health reform debate in the Organisation for Economic Cooperation and Development (OECD) countries has focused on ways of improving the efficiency of public health services. Although an increasingly important role for the private sector is often envisaged (through, for example, competing with public providers for public contracts), in practice this tends to be a small part of the debate.

In the developing world, by contrast, the issue of who should be providing publicly financed health services is a much debated one. In many poor countries, the efficiency and quality of public health services are very low. By contrast, non-governmental organisations (NGOs) have often proved capable of delivering high quality care relatively efficiently. These organisations are increasingly seen as alternatives to the public sector in the delivery of publicly funded health services. Yet the contracting-out of publicly financed health care is often regarded with suspicion, the fear being that payment schemes will be insufficiently sophisticated to ensure that the poor and very sick will be well treated. Providers might, for example, try to select patients within each payment category who seem least costly to treat. Such concerns are perfectly understandable. However, all discussions about alternative ways of organising health services must be based on comparisons of real-world alternatives, and it also seems sensible to be prepared to trade off increases in inequality against improvements in average health. If it really is the case that the quality and efficiency of the public sector are so low, and that neither can be improved easily, and if it really is the case that NGOs and other providers can genuinely deliver better quality care at lower cost, then it seems eminently sensible for policy-makers to be willing to risk increases in health inequality if the prize is a large increase in average health.

5.3. DOES HELPING THE NON-POOR BENEFIT THE POOR?

There is another difficult ethical issue at the country level, namely whether it is sometimes beneficial in absolute terms for the poor if the better-off benefit substantially from a programme or a technology.

Publicly financed health care provides an example. From an ethical perspective, it seems an anathema that the better-off should benefit substantially from them. After all, if health services were allocated on the basis of need, it ought to be the poor who disproportionately benefit, since they are likely to be in the greatest need.²³ Some argue, however, that including the better-off amongst the beneficiaries is essential to maintain their support for the programme. This argument is often heard, in fact, in the context of the British National Health Service (NHS), from which the better-off do benefit substantially.^24–26 And indeed it seems plausible that health systems and policies aimed primarily at the poor and needy will get the necessary political support only if an appreciable portion of the benefits accrue to the better-off. There may, therefore, be something of a trade-off between targeting publicly financed health spending on the poor and sustaining the programme in the long run. This type of argument is not dissimilar to Rawls's²⁷ argument that in some cases pursuit of equality may be to the disadvantage of the poor in absolute terms, and therefore striving for full equality may not provide the best outcome for the poor. But it also has to be said that the degree of pro-rich bias in publicly financed health services in many developing countries^28–30 is almost certainly far greater than would be required to ensure the continued political support of the better-off.

There may also be a temporal trade-off between benefiting the poor and benefiting the better-off. Cesar Victora et al³¹ argue that new services and new technologies will inevitably be used first by the better-off, and in any case the poor might be reluctant to use services and technologies that are not used by the better-off, since the poor are suspicious of services used solely by the poor and aspire to use services used by the better-off. The temporal trade-off is evident too in David Bloom and Jeffrey Sachs's³² argument that to effect a sustained improvement in the wellbeing and health of Africa's rural poor one first needs to make changes to attract foreign investment and foreign aid, changes that will initially disproportionately benefit the better-off. Examples include the building of infrastructure, improved governance, strengthened legal systems and the reduction in the prevalence of communicable diseases. Bloom and Sachs suggest that although such measures will inevitably benefit the better-off initially, it is unrealistic to think that this stage can be leapfrogged.

6. Conclusions

This paper has aimed to highlight and shed some light on various ethical issues that arise in the context of the economic thinking surrounding the World Bank's health sector work. These arise in the definition of goals and objectives, in the measurement of progress towards these objectives, and in the assistance provided to countries in the formulation of strategies and policies aimed at helping them towards goals. In most cases, there are no cut-and-dried answers, and the best strategy seems to be to articulate fully what the tensions and trade-offs are, and where possible assemble evidence that would allow policy-makers to make informed decisions. This paper has attempted to provide some pointers in this regard for the selection of ethical issues it has highlighted.

Disclaimer

The findings, interpretations and conclusions expressed in this paper are those of the author, and do not necessarily represent the views of the World Bank, its executive directors, or the countries they represent.