Cost-recovery, adjustment and equity in health:
some lessons from Zimbabwe


Kevin Watkins
Kevin Watkins looks at the problems caused by the imposition of inappropriate neo-liberal policies in Zimbabwe and other parts of Africa.

From Soundings issue 7 Autumn 1997


After independence, most governments in sub-Saharan Africa established equity as a fundamental principle of health policy. The World Health Organisation's 1978 Alma Ata declaration on Primary Health Care, with its commitment to achieving adequate health care for all by the year 2000, was embraced across the region (WHO 1995). Most governments continue to reflect the aspirations of the Alma Ata declaration in their national health policy plans. But as the year 2000 approaches, it is difficult not to be struck by the painful contrast between the ambition of the post-independence period and actual achievements. While health indicators have improved overall, they have done so more slowly than for any other region, and in many countries they are deteriorating. Poverty is the underlying cause of Africa's poor human welfare performance. Today, one in two Africans lives on or below the poverty line, resulting in high levels of vulnerability to illness and the highest burden of disease in the developing world.

At the same time, African states have been gripped by a deepening fiscal crisis, with governments unable to provide the basic services needed to address the causes and consequences of poor health. Nowhere is this more apparent than in the health sector. After almost two decades of declining per capita incomes, debt crisis, structural adjustment and reduced public finance for health, there is a deepening crisis in public health policy, with resources diminishing in the face of a resurgence of preventable diseases, the emergence of deadly new strains of viral infection and an HIV/AIDS pandemic.
Southern Africa is no exception to this broad picture. With the partial exception of Botswana, where diamond revenues have enabled the government to maintain growth and expand health care provision, the region faces a deepening crisis in public health policy. Governments and donors have embarked on major reforms in response to that crisis. Among the most important of these reforms has been cost-recovery, or charging for health services. Over thirty countries in Africa (and all countries in Southern Africa) have adopted user charges as a means of raising finance. Countries with a long tradition of attempting - admittedly often with limited success - to provide health services free of charge have introduced fees. Much of the intellectual support and financial impetus behind this policy shift originated in the World Bank and the wider donor community, where private finance solutions to public finance problems gained currency during the 1980s. Arguments advanced in favour of user-fees have placed great emphasis on potential efficiency and equity gains. More broadly, cost-recovery has been seen by many of its more enthusiastic advocates as the first step towards a health policy built on the foundations of neo-classical economic theory, with consumers exercising choice in the market place for health services. In this sense debates over cost-recovery are rooted in a broader approach to economic and social development - and they have implications which go beyond the health sector. Problems in health policy have been widely treated as an indication of the 'failure' of African states as agents for delivering services and achieving efficient outcomes in social and economic policy. An enhanced role for the market, and for private sector operators, has been the central policy prescription to emerge. The prescription has been enthusiastically swallowed by most governments in southern Africa, especially in the health sector. Cost-recovery has been introduced as one element in a broader strategy incorporating support for private insurance, decentralisation and community financing initiatives.

In practice, if not in neo-classical theory, the introduction of user-fees has often had a highly damaging impact on equity, with poor communities being excluded from basic services. Elaborate systems for exempting such communities from payments have foundered. Moreover, the evidence from many countries suggests that, despite their high social costs, cost-recovery schemes have failed to generate the resources needed to address the financing crisis facing public health systems. Given the importance of health care in defining opportunities for human development, and the symbiotic relationship between ill-health and poverty, these are serious public policy problems.
This article reviews the impact of cost-recovery in Zimbabwe. The case is an important one because of Zimbabwe's impressive post-independence track-record in improving human welfare. Health policies designed to promote equity and efficiency in access and service delivery were central to this achievement, with the principle of universal access to publicly financed basic provision a central feature. After 1990, the capacity of government to maintain health services was jeopardised by a chronic budget deficit, with stabilisation targets agreed under IMF-World Bank auspices leading to a sharp decline in social sector spending. Cost-recovery was actively encouraged by the World Bank as an alternative source of health financing. Introduced at a time when poverty and vulnerability levels were increasing as a consequence of drought, falling employment and declining real wages, the new policy had adverse implications for equity in health care, with poor communities excluded from basic services. An exemption system ostensibly designed to protect equity failed as a result of poor design and inadequate implementation. The Zimbabwean case raises serious questions about the uncritical adoption of national user-fee strategies. Greater caution and a commitment to monitoring the impact of policy changes on the poor is required. More broadly, the abandonment of equity goals and embrace of a market-based solution to health policy problems needs to be challenged, with governments seeking alternative financial solutions to maintain publicly-financed basic health provision.

The experience of Zimbabwe
Since independence Zimbabwe has achieved some of the most impressive human welfare advances in sub-Saharan Africa. Life expectancy has increased from 55 years to 64 years. Success in combating preventable disease has halved the infant mortality rate to 50 per 1000 live births (compared to 170 for sub-Saharan Africa as a whole). Parallel improvements have been achieved in maternal mortality rates.
These human welfare gains were the consequence of an explicit commitment on the part of the Zimbabwe government to pursue equity goals through increased public investment and improvements in allocative efficiency. During the 1980s, Zimbabwe pioneered primary health approaches which emphasised maternal and child health services, health education, nutrition, control of communicable diseases, water and sanitation provision, and family planning. Immunisation rates increased from 25 per cent to 80 per cent coverage of children. An extensive primary health structure was developed to meet local needs. In terms of health service delivery, one thousand rural health centres and fifty-eight rural hospitals provided the base of a pyramidical structure with six central hospitals at the top, and a range of government and missionary hospitals in between the two layers. Today, over 85 per cent of Zimbabweans live within 8km of a health centre, one of the most comprehensive levels of coverage in Africa. Political commitment to the creation of a health service accessible to all was reflected in a 1984 White Paper Planning for equity in health, and the Health for all action plan, which was produced three years later. It was also reflected in an increase in the proportion of GDP spent on health from 2 per cent to 3 per cent during the 1980s.
After independence, a directive on user-charges for health was promulgated but not implemented. However, during the second half of the 1980s, as public finances came under growing pressure in the face of slow growth and rising public debt, the failure to enforce cost-recovery came under criticism from the World Bank. One influential report recommended financial targets for user charges, with the objective of quadrupling, from 2 per cent to 8 per cent, the proportion of recurrent health expenditure derived from this source.1 Recognising a potential threat to equity, the report advocated an exemption system, while placing the burden of proof on the patient, who would be required to produce evidence of low income in order to receive free treatment. In an almost surreal indication of the World Bank's distance from reality, it was assumed that health facility clerks would provide the front-line service in differentiating between the poor and non-poor, using a variety of survey data on landholding size and type of dwelling to supplement wage data. To the extent that empirical evidence of ability to pay was gathered, it typically took the following form: 'There is strong anecdotal evidence (high out-of-pocket payments to private doctors, pharmacies and traditional practitioners) to suggest that many Zimbabweans would be willing to pay more for reasonable quality health care' (p5). The central conclusion to emerge from the World Bank's health policy evaluations was that fees were too low, collection efforts too lax, and evasion of fees widespread, with the result that significant financial resources were being lost.
There was little public discussion of the World Bank's health policy recommendations. But their importance became apparent in 1991, when Zimbabwe adopted a structural adjustment programme under IMF-World Bank auspices. At the core of this programme was the objective of reducing the national budget deficit from 10 per cent to 5 per cent of GDP within five years. In the health sector, cost-recovery was allocated a central role in containing public spending. Under the 1991-1995 framework for economic reform, the Government of Zimbabwe committed itself to quadrupling cost-recovery collections. Initially, action to this end focused on more rigorous enforcement and collection of existing fees. Then in 1994 a new fee structure was adopted. This raised the costs of registration at primary clinics from Z$l to $Z6.50, introduced charges for ante-natal services which were previously free, and consolidated an escalating fee structure in which higher charges were applied for access to district level facilities.2
While retaining a commitment to equity in health policy, the health reforms adopted under the structural adjustment package constituted a radical departure from the consensus of the 1980s. In particular, the principle that universal health coverage should be achieved through services financed by general taxation was, for the first time, brought into question. This had implications which went beyond the health sector. During the 1980s, the human welfare achievements outlined above were achieved not as a consequence of increased per capita income (which declined slightly) or the redistribution of assets, but through improved access to health and education financed through a broadly progressive tax system. Thus on both the income and expenditure side, budget spending incorporated important redistributive elements which the stabilisation plans adopted under structural adjustment appeared to threaten. By eroding the state's capacity to maintain a redistributive function, the stabilisation programme posed a direct threat to the human welfare gains achieved after independence.

The policy background
In Zimbabwe, cost-recovery was introduced in a context defined by four interlocking factors.
First, severe droughts in 1991/92 and 1994/95 depleted the assets of rural households. The 1991 drought left the vast majority of poor rural households with no marketable surplus and severely reduced assets. Livestock losses in excess of 50 per cent were reported in arid and semi-arid communal farm areas. When the 1994 drought arrived, households had correspondingly fewer assets and savings to cushion the effects, resulting in reduced consumption and an increase in off-farm labour.
The second factor, related to the impact of drought but exacerbated by structural adjustment, was a catastrophic decline in urban wages and employment. According to the IMF, real per capita incomes in the urban formal sector declined by 32 per cent between 1990 and 1993. Rural wages fell by 40 per cent over the same period, as drought undermined production on commercial farms. By the end of 1995, urban wages were still over 20 per cent below their 1990 level. This slump in wages did not protect employment levels, with over 120,000 jobs being lost during the first three years of the adjustment programme. Inevitably, in a country as urbanised as Zimbabwe, the collapse of formal sector employment and wages was transmitted to the communal farm areas in the form of reduced remittances. In 1991, remittances from urban family members accounted for around 20 per cent of total rural income, rising to over 30 per cent for the poorest households. As these remittances fell, it increased the vulnerability of the poor rural households to sudden shocks caused by losses of output or employment opportunities.
The third defining feature of the background to cost-recovery policies was a collapse in public spending. During periods of stress, poor households depend heavily upon access to public services to protect their welfare. In the case of Zimbabwe, structural adjustment was accompanied by a catastrophic decline in public spending and service provision - especially in the health sector. Between 1990/91 and 1995/96, health spending fell from 6.2 per cent to 4.2 per cent of total government spending. Measured in more relevant per capita terms, health spending reached a peak of Z$55 in 1991, more than any other country in Africa except South Africa, Botswana and Swaziland. By 1995, per capita spending had fallen to Z$32. Budgets for preventative health, the most cost-effective part of the system for protecting the poor, were cut by one quarter. Inevitably, the quantity and quality of health services provided declined, with drugs shortages becoming a recurrent problem.
The fourth factor conditioning the environment into which user-fees were introduced was HIV/AIDS. Zimbabwe has one of the world's highest rates of HIV/AIDS, which affects around 30 per cent of the sexually active population. AIDS is now the leading cause of death in Harare, and an estimated 7 per cent of children are born HIV-positive. Almost all will die before the age of five, most of them after protracted episodes of sickness. The AIDS epidemic has already led to a dramatic increase in cases of tuberculosis, which have risen by 350 per cent since 1990. Reports from City Health departments underline the extent of the public health crisis associated with HIV/AIDS. For instance, in Harare infant mortality rates have increased from 30/1000 to 52/ 1000, while perinatal mortality is increasing at the rate of 18 per cent a year. Nationally, HIV-related illnesses are now the second main cause of death and the single biggest killer in the 25-44 age range. The increasing demands imposed by HIV/AIDS on a health system undergoing deep budget cuts has created intolerable strains. Apart from tuberculosis, the epidemic is increasing the prevalence of respiratory infection, acute malnutrition and associated diseases, increasing the demand for health care at a time when provision is diminishing. Each of the four factors mentioned above had the effect of reinforcing the pervasive poverty which characterises Zimbabwe. Around 17 per cent of rural children are malnourished, exposing children to risk from communicable, infectious and parasitic diseases. 'Poverty' afflictions such as acute respiratory infection and diarrhoea still account for half of all outpatient attendance at health facilities. Against this background, any loss of access to health services for poor people inevitably poses acute risks in terms of vulnerability to diseases.
Despite this neither the Government of Zimbabwe nor the World Bank appear to have considered the implications of cost-recovery. Instead, they assumed that an untested exemption system would protect the poorest and most vulnerable communities from the adverse effects of cost-recovery.

Adverse equity effects
The assumption was fatally flawed. Following the stricter enforcement of fee collections in 1991, it would appear that low-income groups were forced either to delay or reduce health care. One early study found that outpatient attendances dropped by 18 per cent by the end of 1991, while inpatient admissions went up by 12 per cent. One possible explanation is that patients started to seek health care only when it was absolutely necessary, thereby necessitating hospital care. Further evidence of the impact of cost-recovery in hospitals is revealed in the fact that average length of stay in government hospitals fell from nine days to seven days, and from four days to two days for maternity patients. Ante-natal clinic attendance also fell in 1991, whereas maternity admissions increased, along with a 10 increase in unbooked deliveries and babies born-before-arrival. Far from being a temporary adjustment to new market signals, it appears that the decline was part of a catastrophic long-term trend. Although ante-natal care registration started to recover slightly in 1992, by the end of the year attendance figures had still not reached their 1990 levels. The costs of non¬registration are high, with unregistered mothers some four times more likely to die in childbirth.
Use of reproductive health services appear to have been particularly sensitive to the effects of cost-recovery. Admission statistics from Harare Central Hospital for 1991 to 1993 show an increase from 8.8 per cent to 13.6 per cent in the number of babies born to mothers who had not registered for ante-natal care. The associated human costs are not captured by cold statistics. They are reflected instead in the fact that the perinatal mortality rate for unregistered mothers (251 per 100,000 live births) is around five times higher than for those who had registered for ante-natal care. Reduced access to ante¬natal care at the same hospital appears to have been behind a 21 per cent increase in the number of babies born-before-arrival (BBA), often because the mother was attempting to minimise the time and costs incurred in hospital treatment; or because of complications which had not been recorded because of non-registration for ante-natal care. In the six months after the introduction of user-fees, BBAs increased by 20 per cent. Mortality rates among these babies was considerably higher than the average. Evidence from other hospitals pointed to a similar picture of reduced health service use. At the St Paul's Mission Hospital, located in a poor communal farming area in Murehwa district, the number of BBAs increased by half over the six months following the adoption of the new policy.
By delaying or obstructing entry to the health system, cost-recovery imposed enormously high human costs on vulnerable people. Those costs are to be measured in terms of lost lives and the suffering associated with unnecessary sickness. The more immediate efficiency costs for the health system can be captured more simply in the additional costs of treating health problems which could have been resolved more cheaply earlier. Individual accounts of suffering and death are often dismissed by the World Bank and the Zimbabwe government, as 'anecdotal', or a consequence of specific problems in the health facility concerned. At one level, the observation is justified. There have been no large-scale controlled studies designed to isolate the impact of cost-recovery. An additional problem is that demand-diversion responses have not been examined, giving rise to the claim that reduced attendance at one health facility may have been complemented by increased attendance elsewhere. Viewed from a different perspective, the fact that such sweeping policy reforms were initiated by the Government and the World Bank without arrangements to monitor their effects through such studies suggests that some degree of self-criticism may be in order. But whatever the shortcomings of the research, the evidence of adverse social outcomes rests upon considerably more than anecdote.
Since March 1992, UNICEF and the Ministry of Public Service, Labour and Social Welfare have carried out five sentinel-site surveys under their Social Dimensions of Adjustment Programme. These surveys covered forty sites, interviewing 120-125 households in each site. They have produced the most comprehensive data on responses to cost-recovery. Among the most relevant findings, the following suggest particularly powerful causal links between cost-recovery and exclusion from vital services:
? The percentage of children whose diarrhoea was not treated at a clinic because treatment was regarded by parents as too expensive increased from 9 per cent of the total in 1993 to 22 per cent in 1994- This shift coincided with a 117 per cent increase in health fees at rural hospitals and health centres. In a sub-sample of 266 diarrhoeal cases treated at home, 58 per cent of families said they had not taken children to clinics because outside treatment was thought to be too expensive. One of the major reasons given by women for giving birth at home was that the cost of delivering at health services was too high.
In January 1993, cost-recovery was introduced for condoms. In the same month, the number of condoms distributed by survey site health centres fell from an average of 53,033 in 1992 to 28,988 in 1993.
Before-and-after studies at specific health facilities reinforce the broader survey evidence. In January 1993, cost-recovery measures were withdrawn for six months in response to a series of epidemics. At one rural hospital, attendance at outpatient and ante-natal departments doubled in the six months after the withdrawal of user-fees. At the St Paul's Mission Hospital the cost of outpatient visit increased from Z$ 1.50 to Z$17 at the beginning of 1994- The cost of delivering a baby almost doubled to Z$120. By the end of February 1994, the number of inpatients and the number of deliveries had fallen by over 40 per cent each.

Failure of the exemption system
From the outset, the World Bank consistently stressed the importance of combining more rigorous enforcement of cost-recovery in Zimbabwe with the development of an exemption system to protect the poor. In practice, however, there was a considerable time-lag between the more rigorous enforcement of cost-recovery in 1991, and the emergence of a viable exemption system. This sequencing problem resulted in the partial or total exclusion of vulnerable communities from the health system at a time when poverty was becoming more intensive and health needs were growing.
The Government of Zimbabwe introduced the Social Development Fund (SDF) in 1991 with the support of the World Bank and other donors. There were two components in the scheme: and Employment Training Programme and a Social Welfare Component. The latter comprised food money, and assistance with school fees and health fees. Elaborate targeting mechanisms were established, imposing an intensive administrative burden on the Department of Social Welfare. Yet no new budget was agreed until 1992, and no extra staff were employed to implement the new system.
In 1993, an independent report into the operation of the Social Welfare Component of the SDF concluded that the food-money scheme was reaching only about 3 per cent of the target population and the school-fee scheme 20 per cent (Kaseke E and Ndaradzi M, 1993). The report concluded that eligible populations were being systematically excluded as a result of complex and cumbersome bureaucratic procedures devised without regard to the constraints facing poor communities. The following were among the most serious problems identified:
* The high cost of applications compared to the benefits received. Potential beneficiaries were required to go in person to DSW offices with extensive documentation to prove eligibility. In most cases, several trips were required to see an application through, resulting in high time and transport costs. Such costs are closely associated with the accessibility of welfare offices. In rural areas, it is not uncommon for the Social Welfare Office to be up to 100km away, rising to 200km in extreme cases.
* Poor people living in remote rural areas were unlikely to be able to afford bus fares to and from offices located more than 80km away. Even if they could, the cost of transport would outweigh the benefits.
* Time lag in benefit payments. Because approvals required authorisation from Harare, a gap of 6-8 months between applications and payments was normal.
* Separate application procedures. Instead of devising one standardised application for all three benefits, separate forms and procedures were required for each, adding to administrative demands and time/transport costs.
* Lack of resource planning. No attempt was made to link the size of the target population and level of benefits offered with the financial resources available, leading to shortfalls in funding.
* Stigmatisation of applicants. The report noted that many eligible parents were unwilling to apply because of the stigma associated with welfare benefits.
* Local intermediary charges. Confirmation of eligibility from head teachers and village elders was often provided only on payment of fees, further diminishing the potential benefits.

To these problems of administration and targeting can be added another: namely, an inability on the part of potentially eligible people to obtain information. At the end of 1993, only around one-half of the population had heard of the SDF (Government of Zimbabwe 1993). That number increased to around 75 per cent during 1995. However, of the respondents interviewed in UNICEF's sentinel-site survey at the end of 1995, only 21 per cent said they had applied for assistance, even though it was estimated that between 50-60 per cent were eligible. The main reason given for not applying was 'lack of knowledge' of how to apply.
Different procedures in different health facilities added to the general confusion. Some hospitals required letters of exemption from the DSW before waiving fees. Others applied more discretionary approaches. Some refused to waive fees even when presented with exemption letters, on the grounds that payments for treatment from the DSW would take up to eight months to materialise.
Health facilities were also forced to contend with the central paradox of introducing cost-recovery measures for communities characterised by high levels of poverty: namely, the law of diminishing returns. In some communal areas, in excess of 60 per cent of the population might have been exempt from user-fees on income grounds. But as one report noted for rural non-governmental hospitals: 'Exempting every patient would result in very little inflow of revenue given the fact that almost all patients fall below the Z$400 household income per month. Consequently, rural hospitals are encouraging patients to pay'.
The income criteria chosen for exemption was also problematic in at least two respects. First, any means-testing of household income is relevant only to the extent that it reflects the real availability of resources for family members, which is a function of the size and age distribution of families. A fixed income obviously translates into fewer per capita resources in a large family than in a small family. Yet the exemption system took no account of family size.
Second, the threshold levels chosen have not been accurate indicators of poverty. In 1991, when more rigorous enforcement was introduced, the exemption level was the same as it had been in 1981. World Bank estimates suggest that this was around 28 per cent below the rural poverty line and less than half the urban poverty line. In other words, the exemption level selected was weakly related to real needs. Moreover, the Z$150 exemption threshold was kept in place for operation purposes until the end of 1991, when it was raised to Z$400, where it remains today. The figure of Z$400 appears to have been picked almost at random. For instance, it is less than half of the income-tax exemption level, which might have provided an equally valid indicator of ability to pay. By 1994, the threshold level had fallen further below the real poverty line. Adjusted for inflation, the exemption level was around 20 per cent below the rural poverty line and 90 per cent below the urban. In the two years since 1994, inflation has averaged over 20 per cent, but the exemption line has not been adjusted upwards, further eroding the real value of the index.
The upshot is that many poor people have been excluded from claiming exemption because of the threshold criteria selected. Using the World Bank's poverty line data, around 25 per cent of the national population, rising to over 30 per cent of the rural population, do not have access to sufficient income to purchase an adequate consumption basket of food, shelter, education, health, and transport. If poverty is the determinant of ability to pay, this entire group would appear to merit exemption - an option which would impose an enormous administrative burden, while at the same time reducing revenues from cost-recovery. This is one of the central dilemmas facing any exemption system to compensate for the effects of cost-recovery in a situation of widespread poverty.

Alternatives to cost-recovery: the challenge
Considering the high social costs associated with user-fees, the benefits in terms of resource mobilisation in Zimbabwe have been derisory. Despite the introduction of more rigorous enforcement procedures, cost-recovery income has stagnated at less than 3 per cent of the recurrent health budget. Most district-level facilities were unable to raise more than 2 per cent of their budgets through this route. At the clinic level, UNICEF's sentinel-site survey found that fee collection had stagnated in real terms between 1991 and 1993. Taking into account the costs of collecting and accounting for fees collected, it is probable that cost-recovery resulted in a net loss of resources. This would imply that the quality and quantity of services provided deteriorated despite the additional burden being placed on households, calling into question one of the main benefits claimed for cost-recovery.
To its credit, the World Bank was relatively swift in acknowledging the problems associated with cost-recovery. In 1995, a Country Economic Memorandum conceded: 'There is evidence that the fee exemption system under which poor people are entitled to claim free treatment has not worked well, and that fees undermined access to health services for vulnerable sections of the population'.4 The Memorandum went on to recommend the suspension of all fees for ante-natal, maternity and child health services, and free services for basic preventative care, immunisations - and treatment of infectious diseases. After initially rejecting this assessment, the Zimbabwean Ministry of Health conceded that cost-recovery had resulted in adverse outcomes, especially for the rural poor. It suspended cost-recovery for rural health clinics in 1995, except for a nominal registration fee.
Reflection on the Zimbabwean experience would also appear to have prompted a review in some parts of the World Bank. One recent research report on cost-recovery in Africa has recommended the withdrawal of user-fees on basic services, except as a last resort. However, more fundamentalist proponents of cost-recovery retain a powerful policy influence within the World Bank, especially in sub-Saharan Africa. Whatever the World Bank's current ambivalence, it might be argued that, in the absence of ideological blinkers, the outcome of cost-recovery in Zimbabwe would have been entirely predictable; and that more effective monitoring of impact and consideration of the evidence provided by UNICEF would have led to an earlier review of the policy advice which had been given. Such monitoring should form a fundamental part of any major change in health policy.
Future threats
Addressing the problems associated with cost-recovery, and developing alternatives, is a pressing social policy problem in Zimbabwe. The country is about to embark on the second phase of its adjustment programme with the IMF and the World Bank, who stopped releasing funds in 1996 in response to government failure to meet targets for reduction of budget deficit. There is considerable pressure on government to resolve the budget's deficit problem. It remains to be seen whether or not it will be possible to meet the targets set within the time-frame allocated without inflicting irreparable damage on primary health care provision, education, and other priority social sectors, some of which are already on the verge of collapse.
The problem facing the government can be simply stated. In 1994/1995, the fiscal deficit amounted to just over 12 per cent of GDR The aim of the new programme is to reduce it to 5 per cent by the end of the decade. This will require tight control over government expenditure. However, the room for manoeuvre is limited. Interest payments on public debt account for some 30 per cent of government expenditure - more than health and education combined. There is no discretion over these payments. Indeed, the rising cost of financing public debt has reduced the real value of the budgetary resources available for non-interest expenditures by over 40 per cent since 1990/91. Recurrent spending on wages and salaries accounts for another one-third of budget expenditures. When the value of the wages and salaries bill is taken into account, the real value of revenues left for capital spending has fallen by over 80 per cent. If amortisation payments are included, domestic debt servicing now accounts for 48 per cent of domestic revenue.
Between them, education and health account for slightly over one-quarter of total budget spending, making them obvious targets for reduced allocations. However, there is no scope for reducing public spending in priority social sectors without adverse effects for human development and long-term growth prospects. The real wages of health workers and teachers have been reduced by one-third over the past five years. Morale has collapsed in both sectors. During 1996, a major strike by health workers aimed at reversing the decline in salaries and halting the slide in service quality reflected a growing mood of desperation. Social-sector professionals are leaving the country in large numbers, with adverse implications for the quality of service provided. Cuts in the recurrent budget for the health sector have also contributed to chronic shortages of drugs and the virtual collapse of rural ambulance services, with vehicles left non-operational because of shortages of spare parts.
There is no question that the budget situation in Zimbabwe is unsustainable - or that matters will get worse without action to reduce debt and contain future borrowing. Without moves towards a more sustainable budgetary position, it will not be possible to finance the delivery of social services, let alone to undertake the infrastructural investments upon which future growth depends. The question is, how can priority social services be protected during the budget stabilisation programme? The answer to this question is to be found in an three-pronged strategy aimed at increasing revenues, reallocating resources to achieve greater efficiency and equity in the social sector, and changing the sectoral composition of budget spending.
Increasing revenues Measures to increase government revenues are an immediate priority. Under the 1990-1995 adjustment programme, tax structures were reformed in a highly regressive direction, contributing to a decline in revenue collection. This has been reinforced by other measures introduced under structural adjustment which have brought windfall gains for some of the highest-income groups in the country, notably within the agricultural farm sector which is dominated by a small group of white farmers. Measured as a share of GDR tax revenues have declined from 35 per cent in 1989 to 28 per cent in 1995, some five per cent below the target set under the adjustment programme. Comparing 1994/95 with 1989/90, income-tax revenues fell by Z$263m in constant 1990 prices. To put this figure in context, it was equivalent to around 20 per cent of the health budget. It was also many times in excess of the user-fees collected at rural health centres and district hospitals.
There is a strong case for the next phase of adjustment to include a strengthening of revenue administration and the introduction of progressive revenue-raising measures. Higher marginal tax-rates for high incomes are an obvious starting point, along with the withdrawal of generous tax concessions -such as those for company cars and housing loans - aimed at higher-income groups. The introduction of a graduated land-tax in the large-scale commercial farming sector would be justified on grounds of efficiency and equity. It would be justified on efficiency grounds because this sector currently uses land, the country's prime asset, highly inefficiently. In contrast to the densely populated communal farm areas, commercial farms currently cultivate only around half of the land they control. A graduated land-tax, allied to the charging of full costs for water and electricity, would encourage a more efficient allocation of resources.

Such a tax would also promote equity in revenue generation, since it would fall on high-income groups - a fact which helps to explain the reluctance of the government to consider a move in this direction.
Relocating social sector budgets Turning to reallocation within existing social-sector budgets, the challenge is to expand investment in primary level facilities where social and economic returns will be the highest. In the health sector, Zimbabwe performs better than most African countries in allocating resources to the primary level, but its $6 per capita spending is below the $9 minimum which the World Bank estimates is needed to publicly finance a minimum health care package - and the share is declining. Reversing this trend is vital if Zimbabwe is to achieve the goal of equity in health care, cost- recovery at the primary level is not an option consistent with this aim.
Where there is scope for increased cost-recovery is in the central teaching hospitals, and in charging private users, such as insurance companies, full costs for public services. Public spending allocations between the primary and the tertiary sector should also be reviewed.
Reprioritising budgets With regard to the third element of a budget stabilisation strategy, the inter-sectoral allocation of resources, there are tough political choices to be made. Subsidies to loss-making parastatal corporations constitute a major budgetary burden and these need to be reviewed and defence is another item of budget spending which is inflated in relation to need, in 1995/6 accounting for 9 per cent of budgetary expenditure.

Beyond Zimbabwe: issues for the region
Zimbabwe's experience is instructive for other countries in the region and elsewhere in Africa. Cost-recovery approaches to health financing are gaining ground in southern Africa, with the World Bank model of uniform national user-fee schemes remaining the dominant model. Chronic budget problems, the poor quality of state services and familiar arguments about the 'unaffordability' of publicly-financed basic health provision have all contributed to the political momentum behind cost-recovery. To the extent that broad lessons can be drawn from the Zimbabwean case for countries characterised by highly divergent levels of income, human welfare indicators and health problems, the most important would appear to be application of the precautionary principle. Cost-recovery has a poor record in achieving financial and efficiency objectives - and a disastrous record in excluding poor people from basic services. At the very least, governments in southern Africa might wish to review their collective experiences and reflect on alternative financing strategies and it is not only national governments which have a capacity to reduce the fiscal pressures driving increased recourse to cost-recovery. Both the World Bank and the IMF rightly stress the damage to social sector budgets caused by unproductive expenditures in areas such as armaments, loss-making parastatals and corruption. They have been less active in addressing the problems caused by unproductive expenditures made to themselves in the form of debt repayment. By the year 2000, slightly under half of Mozambique's budget expenditure will be directed to debt repayment, with the IMF the single largest recipient. At present, debt servicing absorbs more than twice the amount spent on health. This is in a country where one-in-five children die before the age of five, most of them as a result of infectious diseases which would be easily preventable were the resources available. In Zambia, for every $1 spent on health, another four are spent on debt servicing, with the IMF again the largest recipient. All of which points to a powerful case for providing early and deep debt relief, perhaps developing the Highly Indebted Poor Country initiative to reward governments willing to convert savings from debt into priority social investments.
Despite its well-rehearsed rhetorical flourishes on the importance of primary health and basic education, in practice the World Bank has failed to prioritise equity in social policy under its structural adjustment programmes. Like many of the governments it supports, the Bank - along with the wider donor community - continues to regard social policy as an appendage to market-oriented reforms. However, leaving things to the market is bad economics and bad for human development. Access to health provision is one of preconditions for establishing positive linkages between growth and social equity, creating the conditions in which the poor can realise their productive potential and take advantage of new opportunities. Conscious choices have to be made about how to establish these linkages. The evidence suggests that cost-recovery is the wrong choice.

1. See World Bank, Improving the implementation of cost-recovery for health, Africa Technical Department, Washington 1992.
2. M. Chisvo and L. Munro, A Review of Social Dimensions of Adjustment in Zimbabwe, 1990-94, UNICEF, Harare 1994.
3. A. Renfrew, ESAP and Health, Silvera House, Social Series, No. 3, Harare 1992.
4. World Bank, Achieving Shared Growth, Country Economic Memorandum, Southern Africa Department, Washington 1994
.