11548.pdf
Media
- extracted text
-
DRAFT - for comments; not for circulation
HEALTH SECTOR FINANCING
IN SUB-SAHARAN AFRICA
Queen Elizabeth House
University of Oxford
Preface
Acknowledgements
Contents
A. Introduction
B. Theoretical perspectives
The efficiency-equity debate in health financing
The public good content of health services
Failures in related markets
The problem of poverty and health service provision
Implications for the character of health systems
C. Development of African Health Systems
Pre-colonial health systems
Colonial period
Independence
Health system performance
D. Health Sector Performance in SSA and Reform Options
Resource inadequacy
Failure to use resources efficiently
Government failure
E.
A Case Study of Tanzania
Pre Independence
Independence to Arusha Declaration: 1961-1967
The Arusha Declaration 1967
Economic decline (1974-1985)
Assessment of the Tanzanian health system to 1990
Recent changes introduced by the Government of Tanzania
The future health service in Tanzania
Conclusion
F. Conclusions (to be completed)
ANNEXES
1. References
Executive Summary
This paper presents a literature review on health sector financing in Sub-Sahara Africa
(SSA). It was produced as part of a research project funded by ESCOR. The paper also
incorporates a case study on Tanzania generated during a field-trip in July 1994.
Health consumption by individuals can be financed through a combination of public, private
and donor/charitable funds. The aim of this paper is to establish relative shifts of these
sources over time and to analyse the impact of such shifts. The impact of reductions in
government expenditure on health was a particular concern.
Our review showed that there is almost no information on the relative contribution of private
and aid finance prior to 1990. However, available data suggests that the contribution of the
public sector to total financing has fallen since the Independence period in Africa.
Health reforms have been introduced extensively, partly as a response to this fall in public
finance for health consumption and partly as an aspect of structural reforms during
adjustment. Reforms alter the character of health systems and the terms of access to health
services.
Reforms are intended to combat inefficiency in the public sector. Most reformers associate
inefficiency in the health sector with general bureaucratic deficiencies and government health
policy failures. As a result of the former governments waste resources, while health sector
policies, and in particular public dominance of health markets, are agreed to lead to suboptimal outcomes in the sector. The alternative sought by reformers is a more market based
health system. Indeed market reforms, which are generally associated with an increase in
private finance, are intended to solve the inefficiencies associated with an underfunded public
sector.
While economic theory indicates that in general a market based system has the potential to
generate an efficient allocation of resources, applied to the health sector this is subject to
much debate. As in all markets, potential only translates into actual efficiency when the
standard conditions of perfect competition are in place. In the health sector in SSA, the basic
assumptions of competition in supply and adequate information are absent, while there are
important externalities.
The result of imperfections in health markets is that markets will produce inefficient
outcomes. We recognise that inefficiency exists in the public sector. However if the
movement to a market based system is also subject to inefficiency, the question is whether
there is likely to be any net benefit. Moreover market outcomes usually lead to a different
distributional outcome than public sector allocations, with the poor frequently being affected
adversely.
We argue that these issues can only be explored by research in specific contexts. Reforms
need to be fashioned on the basis of country specific information, in particular evaluating
private and public capacities. Can the public sector regulate a private market? Will the
private market serve rural and remote areas? Only upon completion of such research can
potential inefficiencies and priorities for public action within a country specific health market
be identified.
FINANCING HEALTH IN SUB-SAHARAN AFRICA
A. Introduction
There is considerable concern about the adequacy and sustainability of health sector finance
in Sub-Saharan Africa. As a result, a number of reforms are being implemented and more
have been proposed. The objectives of such reforms, and their theoretical bases, differ
considerably. However, they all envisage changes in the shares of total resources for health
contributed by governments, households, private enterprise and voluntary organizations.
Shifts in responsibility for financing health between the private and public sector are not new
but date back at least fifty years. This paper is concerned with the origins and impacts of such
changes. Its particular focus is the implications for the efficiency of resource use and
allocation within the health sector and the subsequent distribution of benefits that result from
that allocation.
There are three questions which are fundamental to the efficiency and equity outcomes of
health sector financing:
who pays?
who benefits?
what is the relationship between the costs of services and the value of their
outcomes?
Although these questions appear to be straightforward the issues they embody are complex.
In the first place there are widely different views about who should pay. Some say that the
consumption of goods and services which influence health are matters for personal
responsibility and personal arrangements. Others argue that payments for health services
should be raised from taxes because of externalities associated with changes in health status,
and informational and other flaws in health and related markets.
There is also debate surrounding the appropriate distribution of benefits from health services.
On the one hand, some argue that charges at the point of delivery lead to efficiency gains
even if this restricts access to care and that, on the whole, efficiency gains from fees make
up for equity losses. Others argue that access to health services and the option to use them
should be available to all citizens irrespective of their material means.
A third set of difficulties concerns the relationships between health service costs and its
benefits. It is difficult to place a value on an improvement in health, QALYs and DALYs
notwithstanding. In addition, healthiness may be not only, or even primarily, a function of
health service consumption. Income, the state of the environment and knowledge may be
more important in determining health status than the consumption of health care. In addition,
a health service is not homogenous but an industry comprising activities of many kinds
demanding a wide variety of inputs and producing both intermediate and final outputs. As a
result, there are both conceptual and measurement difficulties in assessing technical,
productive and allocative efficiencies in the sector.
Page 1
Section B introduces a discussion of the alternative theoretical perspectives which underlie
the current discussion of health sector financing to provide a background for what follows.
Many are beliefs rather than theory in the true sense, arising as they do from different views
about the proper role of the state and different concepts of equity. From this identification is
the larger observation that fundamental choices about the value of health and the role
expected of government in financing health improvements, have important implications for
the outcomes sought or the priorities pursued in financing reforms.
Section C describes and analyses the background to the alleged ills of health sector finance
in SSA. Alternative explanations of performance are presented which underpin alternative
reform strategies. Of particular relevance is the extent to which neo-classical economic theory
is an appropriate basis for guiding the allocation of public resources in the health sector.
Using the framework developed in Section C, Section D analyses the recorded outcomes of
recent shifts in financing patterns.
Section E is a case study from Tanzania where a brief study of recent reforms in health sector
finance and their implications was carried out.
Section F draws some tentative conclusions. The more important include the paucity of
empirical evidence which might support or undermine reforms being implemented; the lack
of coherent medium term goals which changes are intended to realise; and, in particular,
uncertainties as to the intended role the public sector. Attention is drawn to the costs and
difficulties of reversing trends once established and the capacities which would be required
by the state to regulate and orchestrate efficient private markets in health care and related
services.
Page 2
B. Some Theoretical Perspectives
The efficiency-equity debate in health financing: who should pay and who should benefit?
The desired output of health investment by both private consumers and by the state is an
improvement in health status. However, better health may not come only or primarily from
the consumption of health services. Rich people are healthier than poor people irrespective
of their consumption of health services. Historically, some of the most important
improvements in European health have come from improved urban sanitary conditions rather
than from medical innovations. Indeed, most health outputs can be derived from alternative
combinations of inputs. For example, malaria can be combated by the use of drugs
(preventative and curative); mosquito spraying; or land drainage (Griffin 1990). The questions
which then arise are what inputs should be purchased in order to produce the desired health
output, who should choose them and who should pay for them.
Until recently at least, many believed that both the financing and the provision of health
inputs were primarily the business of the state and that the state also had a role in rationing
the use of available services. Good health was a benefit to society as a whole and,
correspondingly, a citizen’s right. Those who paid their taxes, or contributed to a universal
national health insurance scheme, or even those who did neither, were eligible to obtain the
health and related services on offer free at the point of delivery or at a subsidy. The
responsibility for deciding on the level and type of services available, on the allocation of
resources required for their provision and on their rationing to conform with public budget
constraints were matters for the government.
More recently this approach has been criticised, as part of the wider criticism of ‘government
failure’, on the grounds of allocative inefficiency (rich people over-consume free or highly
subsidised health services forcing under-consumption on poor people); on the grounds of
technical inefficiency (the priorities given to rationing different services do not always
conform to the highest possible social returns from the same expenditure); and on the grounds
of productive inefficiency (public providers do not always use the least cost combination of
inputs to achieve a desired output). In addition, it is argued that the budget assigned by the
state to health may not reflect either optimal health service consumption or the preferences
of society for health inputs.
These criticisms lead on to a spectrum of views which, to a greater or lesser extent, challenge
the sovereignty of the state as the sole financier and provider of health inputs and introduce
the notion of ‘markets for health’. In a market economy, the price mechanism is the means
by which resources are allocated to produce those goods and services which are most valued
by society. The theoretical model of perfect competition, which turns on certain standard
assumptions, suggests that a price mechanism can produce an efficient allocation of resources
for any given income distribution, through the explicit association of price and value. It is
argued that access to health care is like the access to any other good or service and is best
rationed by the price established in a competitive market. For example, Jimenez (1987),
Griffin (1990) and many others argue that health outputs can be valued through the market.
Those of the liberal tradition argue that health is a matter for personal choice and
responsibility in which the state should have no part.
Page 3
However, a number of difficulties confront this approach. As one moves from theoretical to
observed markets, the significance of the assumptions producing the theoretical results
becomes important. This is expressed in the welfarist literature, by a significant debate
concerning the conformity of health markets to theoretical markets. Market supremacy is
challenged on at least four grounds by welfare economists. Some argue in a general sense that
health differs from ‘ordinary’ goods because health outcomes are difficult to measure and, as
a result, difficult to value (Culyer 1989). Others point to the public good quality of many
health services and express doubts about the willingness of private providers to supply them.
Still others argue that failures in health markets are so pervasive that the allocation of health
through an unregulated market would be inefficient (Sugden 1982).
A more fundamental debate concerns whether or not welfare economics based on neo-classical
utilitarian assumptions is a satisfactory theoretical basis for considering these issues. ’Extrawelfarist’ economists, and many from other disciplines, doubt whether it can be said that a
market exists for inputs to a non-tradeable state like health. They emphasize the non-goods
sources of satisfaction including the quality of relationships between individuals (Sen 1980).
Culyer (1989) argues for a more general notion of characteristics of people and their
capabilities and distinguishes between market demand and needs in achieving life potential.
This approach focuses on health as the desired outcome, not health care, and argues that user
prices should not act to discourage the use of care which contributes to the objective of
maximising health. It admits the validity of ‘merit goods’, those which are valued more highly
by society than by individuals. In the context of better health, the impheation is that the state
has a role in subsidising the consumption of health services which fall into this category (such
as vaccinations) and taxing practices which are harmful to health (such as smoking).
Stewart (1994) summarises the basis and implications of three ‘extra-welfarist’ schools of
thought: the basic needs (BN) approach; the capabilities (C) approach and the human
development (HD) approach. She draws distinctions between their philosophical foundations
but shows that they share a common concern with universal access to goods and services
which permit people to achieve their potential ‘functionings’ (in the BN-approach),
‘capabilities’ in the C-approach) or ‘doings’ (in the HD-approach). Moreover all share a
scepticism that, left to itself, a health market will provide these goods and services in relation
to means rather than needs. Healthiness is a priority on the agenda of all three approaches.
Although sometimes unstated, the clear implication is that the state has a role in the ensuring
such access.
Lying behind differences about the financing of health services is a still more fundamental
debate about the proper role and responsibilities of the state with respect to individual
responsibilities and preferences and the notion of equity. Nicholas Barr (1993) usefully
distinguishes between the objectives of the (welfare) state and the means by which such
objectives are pursued. Efficiency objectives result in a largely technical debate about the best
means. This debate currently focuses on the propensities to failure of governments as opposed
to markets and, by extension, whether or not public interventions to correct market failure
result in more or less efficient outcomes. Equity objectives, however, are properly a matter
for political rather than technical definition as, without regard for their costs and feasibility,
a wide range of options exist to redistribute income, wealth and power, some of which lie
inside the consumption frontier but more than one lies on it.
Page 4
The debate about who should pay for and who should benefit from better health reflects this
wider debate and swings around three sets of issues: two relate to efficiency and one to
equity:
the extent to which health inputs are public goods and unlikely to be provided by
private suppliers or paid for by private consumers;
the extent to which failures in health and related markets lead to greater or lesser
allocative inefficiencies than public allocations;
the extent to which concerns for the health of the poor justify state intervention.1
The public good content of health services
The approach of some economists working in the welfare economics tradition is to stress the
unrivalled and non-excludability consumption characteristics of some health services which
give them public good attributes and make it unlikely that private suppliers would provide
them or private consumers pay for them. The problems of such public good qualities can all
be incorporated into a market approach (that is by assessing the divergence between market
and social costs and benefits), without having to sacrifice the market’s allocative properties,
provided collective payments can be arranged. As a result, once health outputs can be
classified by their market attributes, the question of appropriate financing and in particular
the applicability and level of prices can be considered.
The World Bank’s Better Health for Africa (1993) follows such an approach. One
classification of health services according to public or private characteristics is shown in
Table 1. The consumption of Category 1 services is said to be unrivalled in as much as the
consumption of one individual does not diminish the consumption of others. Moreover no-one
can be excluded from their benefits; indeed no-one can exclude themselves. In these senses,
it is argued, Category 1 services have a high public good content and are unlikely to be
provided or paid for privately. The pressing issue for the state is then how much should be
produced. Category 2 services are associated with a high level of consumption externality as
more people benefit than the primary consumer of the service. Yet to ensure the realisation
of the full social benefit, prices would have to be below market prices, with an associated
subsidy required to cover the full (marginal) cost. The benefits of Category 3 services accrue
largely to the primary consumer. Arguably, these services should attract a full fee reflecting
the cost of provision, poverty issues apart.
1.
A related issue is the extent to which private providers are likely to be more efficient
than public providers. Because health services in the past have typically been both
financed and provided by the public sector, institutional efficiencies have become part
of the rationale for financial reforms. However, the financing of health (or any other)
services and their provision need not necessarily be from the same source. It is
perfectly possible for the state to contract with a private supplier and pay for the
services it provides.
Page 5
Table 1: Market Attributes of Health Services
Category 1
Public Health Activities:
Epidemiological data collection
Health System Planning
Health Education
Regulation, Licensing
Environmental Health
Prevention of Communicable Disease
Category 2
Mixed Public Private Activities:
Targeted Health Problems (merit goods)
Family Planning
Maternal/Child Health
Infant Nutrition
Immunizations
Treatment of Communicable Diseases
Category 3
Private Activities
Acute Care
Inpatient Services
Outpatient Services
Laboratory Services
Hospital Hotel Services
Source: Adapted from Griffin (1990)
In summary, this classification suggests that goods and services in Category 1 would be
inefficiently produced in a market system without collective action, which implies either
intervention by the state or collective action by communities. Also market allocation for goods
in Category (2) would be inefficient and require some form of intervention (eg subsidies from
the state) to ensure that consumption was at the socially optimal level. Finally, where goods
and services have mostly private benefits (category 3), the necessity for deviation from market
prices is less clear.
The critics of this line of reasoning make two main observations. First, they attack the degree
to which externalities in health services exist and their importance. They assert that, except
for the provision of information and possibly minimal market regulation, the public benefit
content of many of the services in Category 1 is arguable and that private good characteristics
dominate the public good content of most or all of the services in Category 2. Lal (1994) for
Page 6
example, mounts a vigorous attack on this approach. While accepting a role for public finance
for public good-type services, he argues that there is often less public interest in health
outcomes than appears at first sight. He also points to the flimsy and contradictory nature of
the empirical evidence underpinning the assumptions that externalities are prominent in
Category 2 type services.
It is worth noting that the pattern of disease may itself determine the presence of externalities
in markets for health services and therefore the importance of providing services with more
or less public good-type content. Many economists of the liberal school seem to have in mind
a society where sanitary infrastructure is well-developed and communicable diseases have
been more or less brought under control. However, in countries where they remain common,
or where the health environment is of low quality, the socially optimal mix of services may
have a higher public good-type content than in countries where these problems have been
largely solved. This may be particularly true of an epidemic such as HIV/AIDS which is
already placing an unprecedented burden on both public and private finance in Africa. As
communicable diseases are eradicated from a society, the externalities arising from further
individual protection probably decline. These differences may be crucial in determining the
appropriate financing structure for health services in different epidemiological settings.
Second, critics of a dominant position for public finance in health question the implication
that the presence of externalities automatically justifies state intervention. Lees (1967) quoted
in Culyer (1989) summarises this view as follows:
"many externalities are irrelevant to human action and the achievement of
optimal solutions,
many are dealt with voluntarily,
the costs of government intervention may outweigh its benefits,
the imperfections of government may make an imperfect market solution more
imperfect".
Fallures in related markets
In addition to the possibility of externalities in disease and the benefits of health inputs, those
who concern themselves with the possibility of market failures point to information and
insurance markets as being particularly vulnerable. In relation to information about health and
health services, they point to two sets of problems. The first is that consumers may choose
health services inappropriately because they are unable to make informed judgements about
options. This so-called ’consumer incompetence’ is partly related to a lack of knowledge
about relationships between unhealthy behaviour and its outcomes (smoking and lung cancer
or low-energy diets and child malnutrition), partly related to a lack of knowledge about the
potential seriousness of symptoms and partly to do with an inability to choose knowledgably
between treatments on offer. The last has received considerable attention in the literature.
Such is the asymmetry of information between medical personnel and patients that a
principle-agent problem commonly arises. The result is a tendency to over-consumption of
some services, particularly where the medical attendant has a pecuniary interest in the
treatment, termed ’supplier-induced demand’ (SID). It might be noted in passing that this
over-consumption is not confined to the rich. Poor people also believe (often erroneously) that
an injection or tablet is required to cure many quite trivial conditions (for example of viral
Page 7
origin) for which no cure is known other than time. But doctors and nurses are happy to
comply, particularly if they also dispense the pharmaceuticals for private profit.
Many of the reforms proposed for state-financed health systems in Africa include the
possibility of health insurance (for example, in Tanzania; Abel-Smith et al (1992)) on the
grounds that ill-health is an insurable risk. However, the efficacy and feasibility of raising
health revenues from insurance in Africa is much debated. The critics base their objections
on both theoretical and practical grounds: on the vulnerability of insurance markets to failure
and on the inappropriateness of insurance schemes to Africa. The main theoretical objections
are that health insurance markets are prone to loading (estimated to account for a 10 per cent
deadweight loss in the US), adverse selection and moral hazard. Moreover, there is a lack of
incentive on the part of either patient or doctor to contain costs if insured by a third party.
The US experience is frequently quoted as an example.
Hammer (1993) also argues that most private insurance systems are socially inefficient
because they insure against common but relatively trivial and inexpensive treatments. It would
be better, at least in countries where administrative capacities were limited, for health
insurance to be limited to episodes where the costs of treatment were likely to cause a
substantial perturbation in expenditure or result in the liquidation of assets which had been
built up over a long time.
Culyer (1989) summarises the debate and concludes that with the right regulatory framework
voluntary competitive medical insurance might be technically possible in a large economy like
the US but not in a small economy like the UK. But compulsory universal insurance may
have efficiency advantages as long as they are not offset by X-inefficiencies.
Defenders of health insurance point out that public subsidies, rather than insurance per se.
distorted prices in US health markets. Moreover they argue that adverse selection can be
minimised by experience weighting rather than community weighting; and that moral hazard
can be reduced by ’first pound payments’, co-financing or co-insurance schemes. More
recently attention has turned to local institutional solutions for insurance market failures; for
example. Preferential Patient Organizations and Health Maintenance Organizations (HMOs)
in the US (McGuire et al 1989).
The other line of criticism is directed towards the practicality of introducing health insurance
in Africa such that it would make a major contribution to health finance. Privately provided
insurance can only really be applied to the self-employed and employees in the formal sectors.
These are still the minority in most African countries. As a result, limited insurance schemes
related to access to private care for major episodes in a mixed public-private health care
system may be appropriate but not as a major source of financing for health services for the
majority. A universal system is probably only feasible within the public sector and then as
part of, or related to, a tax system relying on direct income taxation: direct taxes are
[providing a declining share of public revenue and indirect, value-added taxes are becoming
a larger source in many African countries. A related issue is that insurance against loss of
earnings might be as important in many cases as insurance against the costs of treatment.
A third type of market failure which is less extensively addressed in the literature is the
problem of monopolistic behaviour amongst health service providers (Mills, 1994). Some
Page 8
would argue that the medical profession manipulates demand (the SID problem), is secretive
about relationships between treatments and outcomes, is difficult to regulate, and in these
ways effectively protects itself from competition. As a result, even in markets where the state
is not active, prices are not established competitively. In many areas of rural Africa, there is
only one source of medical services available. Price data for different types and qualities of
health service are sparse and an analysis which might throw light on the extent of this
problem could not be found. However, there is anecdotal evidence to suggest that private
medical practitioners in Africa charge what they like, in many cases what they believe the
patient can afford to pay. While this practice in benevolent hands may result in the rich
subsidising the treatment of the poor, this is not an argument which supports the presence of
an efficient, competitive market. The response in the UK has been to encourage internal.
provider markets to develop but it is unclear what scope there is in Africa for this innovation.
A more general point is that, once involved in treatment, a patient has little recourse to a
competitive market even if it existed.
A final difficulty which does not receive great attention in the literature, is the constraints on
the development of health markets in Africa’s remote rural areas. It is not clear how many
private providers would grasp the opportunities for establishing rural clinics for profit if the
state withdrew from financing or provision. Public health services have great difficulty in
persuading doctors and nurses to work in rural areas when more lucrative opportunities and
a greater choice of goods and services exist in the cities. The experience of Ethiopia before
1974, where this would have been legal and a legitimate response to rural demand for health
care, suggests that the response would be small: only 7 per cent of the rural population had
access to health care.
The problem of poverty and health service provision
In addition to failures in health and related markets, economists debate the appropriateness
of a response to poverty via the non-market provision of health care. The debate is divided
along two lines. On the one hand, there is an ethical debate about the rights of citizens to
medical care irrespective of their means. On the other, welfare economists debate the
efficiency and utility of linking poverty-reduction and free health services and whether here
is a role for the state in dealing with poverty.
The ethical debate is based on beliefs not theory. Some argue that good health, a major
contributor to realising human capabilities, should not depend on ability to pay (see above).
They point to the fact that the poor are less healthy than the rich and therefore should have
greater access to health services. They are particularly concerned that fees at the point of
service might reduce the consumption of health services by poor people and jeopardise
improved health outcomes. Others argue that good health, like any other desired human
attribute, is a matter of endowments, enterprise and individual preferences for good (or bad)
health and that ’rights’ do not come into it
The economic debate is divided into two main areas of dispute: whether poverty should be
a concern of the state or of charity; and whether the provision of free or subsidised health
services is an efficient or appropriate response to poverty. Some welfare economists argue that
because poverty offends, causing dis-utility amongst richer sections of the population,
transfers from rich to poor result in a Pareto-efficient welfare gain. Some link this line of
Page 9
reasoning with public choice constructs which estimate the degree of income redistribution
required and permitted by a voting electorate. They may then go on to use econometric
models (somewhat unsatisfactorily) to predict the extent to which tax revenue should be used
to reduce poverty. Linking this proposition with the involvement of the taxpayer in financing
health services, some, but not all, would argue that the state’s involvement in such transfers,
including the coercion involved in raising taxes, is essential to sustain stable levels of poverty
reducing activity particularly in the presence of widespread poverty Barr (1993).
The critics of this approach base their arguments on a theory of the state which assigns the
reduction of poverty to private charitable giving. In addition, they point out that all public
transfers are regressive and therefore inefficient and argue in addition that public transfers
’crowd out’ private transfers (Lal 1994). Donald Cox and Emmanuel Jimenez (1992) have
estimated the extent to which this is so in Peru and conclude that although public (social
security) transfers do reduce private transfers by 16 per cent, there is an overall increase in
transfers from rich to poor if the state is involved. There appears to be no study of the
interactions between public and private transfers in relation to health service consumption in
Africa.
Another line of argument in favour of the state’s involvement in poverty-reducing health
interventions also regards poverty as a public ‘bad’ but, in addition, points to the productivity
gains which arise from better health and which benefit society as a whole. There are critics
of this line of reasoning too. Lal (1994) is one such, arguing that to assign a health-enhancing,
productivity-increasing role to the state is to view it as a rancher only interested in improving
the productivity of the herd.
The other area of economic debate centres on the extent to which free or subsidised health
service is an appropriate and efficient way to tackle poverty. The arguments in favour also
emphasise the importance and ’public-good’ content of human capital and the ’public bad’
quality of poverty and point specifically to the externalities for productivity, education and
inventiveness that good health offers. These arguments are used to justify direct exemptions
from payments for health services by poor people. A related line of argument runs that cash
requirements for health (and education) services compete with other expenditure priorities and,
in the final analysis, divert resources from rural savings and investment. The provision of free
health services is therefore likely to contribute indirectly to both human and physical capital
accumulation. Some critics, while admitting that the reduction of poverty is a valid role for
the state, suggest that there are other ways of achieving this objective more efficiently. While
there may be other arguments which justify public involvement in health financing, this is an
unpromising way to relieve poverty (Hammer & Berman 1994).
One problem with this debate is that there is considerable uncertainty about the effectiveness
and efficiency of alternative ways of transferring resources to the poor. Over the last five
years there has been a general shift away from the subsidy of goods and services (resource
transfers in kind) to the provision of cash or cash-substitute (coupons) transfers on the
grounds of efficiency. However, there is a good deal of evidence that such transfers are
regressive and become more regressive with time and little evidence to support the notion that
stable, progressive transfer systems are possible. In the health sector, a considerable amount
of attention is being devoted to administrative mechanisms for providing targeted exemptions
from health service fees or providing compensating income transfers. Much more research is
Page 10
required before it will be clear that such exemptions and transfers will avoid either inclusion
or exclusion targeting problems and will result in more distributional efficiency than the
provision of free or subsidised services. The analysis of the distribution of benefits from
Malaysia’s publicly-financed health service is summarised in Section D.
Implications for the character of health systems
These widely different views have important implications for the character of health systems
and for the direction of health finance reforms and the sought-for objectives of such reforms.
They also have important implications for the role assigned to the state, or more properly
derive from it, and for the capacities required by public institutions to regulate liberalised
health markets. To provide a framework for analysis, we have devised four categories of
health systems which reflect the diversity of concerns highlighted above. The main features
of these systems are shown in Table 2.
Type A health financing is characterised by the suppression of private markets in health
services and a monopoly by the state in both provision and financing. State monopoly in Type
A systems usually applies to both the direct provision of health care and to related services
such as pharmaceuticals, diagnostic services, health facility construction and contracts for the
supply of equipment. It is based on the ethical belief that citizens have a right to good health
irrespective of their means and that markets are not to be trusted to allocate resources either
efficiently or equitably.
Type B health financing is characterised by a dominant public sector in both financing and
provision but without suppression of private health markets. A feature of Type B systems is
that the state is concerned not only with the financing of public-good type services but also
with the financing, and usually the provision, of subsidised or free services to the poor. In
Type B systems, public finance not only provides for a substantial share of services overall
but finances a sector with a role to innovate, lead and maintain standards. Teaching hospitals
and most research2 are financed publicly as are the most senior and prestigious positions not
only in medicine and nursing, but also in ancillary services, such as diagnostic services and
physiotherapy.
Finance is usually raised from general tax, from an ear-marked tax or from a publicly
administered health insurance system. Institutional options for private financing, usually from
health insurance administered privately and/or by employers are also frequently available.
These typically give preferential access to services (avoid queues), are associated with a
choice of medical attendant, better hotel facilities in hospitals and, although often denied,
superior health care for certain classes of treatment.
Type C health financing restricts the use of public finance to services with a high public-good
content but, distinguishing it from Type D financing, admits the need for the public financing
of free or subsidised health services to the poor and recognises a role for the state in making
up for missing markets and market failure. This means that the state s attention is mainly
2.
In some countries a significant share of research is financed from charitable sources,
often with links to the pharmaceutical industry.
Page 11
focused on the provision of information, on the regulation of health and related markets, on
improvements in environmental sanitation and on the control of communicable disease.
Private finance dominates public finance, user fees are common and health insurance is the
dominant source of financing for private good-type services. A significant share of research
and training is financed by the private sector or by charity.
Type D health financing restricts the role of public finance still further by a more selective
assessment of public benefit in the provision of services and the delegation of responsibility
for poverty to charity. There are few systems which confirm with this theoretical model.
However Type D systems are important as they represent an expression of the classical liberal
school of thought in its purest form. Lal (1994), amongst others, espouse it as the most
efficient and equitable solution to the allocation of health sector resources.
Page 12
TABLE 2: PATTERNS OF HEALTH CARE FINANCING
Private sector dominated
Public sector dominated
Characteristic
Type A
Type B
Type C
Type D
Ethical basis
All citizen’s rights to the same level of
health care irrespective of means
Universal access to health services
Universal access to health services;
public responsibility for poverty
Access determined by means and
charity
View of equity
Equal consumption by all citizens
Minimum consumption guaranteed:
otherwise consumption related to
contributions
Consumption related to means subject
to Pareto-efficient considerations and
provision for the poor
Consumption related to means subject
only to Pareto-efficient considerations
Sources of finance
Taxes and levies
Public finance: taxes & levies; private
finance: fees, employers contributions,
charity and insurance
Public finance: taxes & levies; private
finance: fees, employers contributions,
charity and private insurance
Private finance: fees, charity and
private insurance
public finance: taxes & levies.
Basis for rationing services
Administrative allocation: access to
higher levels by referral
Primarily administrative allocation:
access to higher levels by referral;
prices may be used to rationalise use of
primary and referral levels
Primarily by prices for private good
type services; administrative allocation
for public good-type services
Almost entirely by price: referral
systems not a dominant feature
Role of the public sector
Monopoly in financing & provision of
both public and private good-type
services
Main financier and provider of both
private and public good-type services
(including private market regulation) in
competition with private providers
Financing of public good-type services
including regulation, making up for
market failure, financing services for
the poor; may contract out the
provision of services it finances
Concentrates on financing a limited
number of public good-type services
and market regulation
Role of the state with
respect to the poor
Financier and provider of services
Financier and usually provider of
services but charities also involved
Financier of services; charities
prominent, private sector may be
contracted to provide services
Not the business of the state
Priorities for public finance
Assumes universal target consumption
of health care: state allocates and
rations between priorities
Assumes target consumption of health
care: state allocates and rations
between priorities taking into account
level of private activity
Assumes universal target consumption
of public good-type health service
consumption and minimum
consumption by the poor.
Finances limited number of public
good-type services, primarily
information and market regulation.
Criticisms
Government failure in allocation and
provision: over-consumption by rich/
privileged
Government failure in allocation &
provision: over-consumption by the
rich; under-consumption by the poor.
Uncompensated market failure in
allocation; under-consumption by the
poor.
Market failure in allocation; under
consumption by the poor; mistaken
faith in market efficiency.
Page 13
C. Development of African Health Systems
Health systems in SSA have developed as a result of the combination of pre-colonial, colonial
and present day practice. Whilst this paper is concerned with present day shifts in financing,
an appreciation of the historical antecedents to the present, offers vital clues to understanding
the present structures in many countries. Accordingly we briefly comment on pre-colonial and
colonial systems, before examining and classifying the present day health systems in SSA.
Pre-colonial health systems
The literature on the delivery and financing of health services is sparse and largely anecdotal.
The vast majority of services were supplied by specialist ‘healers’ and practitioners offering
medical services to local communities. The impression gained from the literature concerning
this period is that the terms of access to health services were both dependent upon payment
and upon the relationship individuals and households had with their local communities. For
example, Falola (1992) describes a system in West Africa where health specialists would
administer a variety of services for payment or exchange:
"Services for barrenness needed no more than payment than the goat which
was used or the chicken which would be slaughtered and partly used in
preparing the medicine that the barren woman would take.The remnants of the
chicken would constitute enough payment for the healer. A measure of millet
or corn could also be used as payment... Exchange payment was common.
Here a patient who was physically strong but psychologically or emotionally
ill could offer to clear a plot of land for the practitioner in exchange for the
latter’s services".
Pearce (1992), again on West Africa, claims that due to the influence, importance and position
of healers within the local power structure, they were able with other leaders to direct public
activities to improve health status which included "..the movement of villages from hazardous
locations, the building of irrigation projects; the disposal of waste, or control of deviant
behaviour".
These descriptions highlight a decentralised system, usually based at the local level and
instituted through a formal community relationship. Whilst the above should not be treated
as conclusive, it hints at a system which quite strongly resembles a modem ‘type C’ system.
That is, access to health services depended upon payment at the point of delivery, but that
certain public goods were provided through community action and poverty was dealt with
through means testing at the point of service delivery.
Colonial period
Colonialism extended provision of health services in SSA by both the direct action of colonial
governments and through activity by private companies and religious organisations. The initial
focus of colonial governments was towards providing health care for European administrators
and later for local civil servants. Likewise, private companies sought to provide health
services for their European employees and later for their African employees. This approach
resulted in the bulk of health facilities being located in administrative and commercial centres.
Page 14
For example, at Independence, health services were highly concentrated in the copper belt in
Zambia (Stock et al 1992). However, location was not synonymous with access. In Kenya
"...some of the manual workers on the plantations and estates were among the most wretched
in society...similar accounts can be found...for workers in the former Congo, in Angola, and
in Southern Africa" (Pearce 1992).
Religious organisations were left with providing care to rural areas and to providing relief to
the poor (Itayvyar 1992). This is supported by Gilmurray et al (1979) who suggest that for
Southern Rhodesia the colonial administration "...did not regard the care of the general
population as its responsibility and, ‘it took little note of the mass of the African population’"
(original quotation).
The philosophy of colonial authorities was characterised by the practice of cordoning off
European from African settlements. However, by the 1930’s, increasing awareness of the
negative externalities of disease, fostered by improving medical knowledge prompted
consideration of a wider role for the colonial authorities. Mburu (1992) quotes Dr. John
Gilkes, the Director of Medical and Sanitary Services in Kenya:
"employers of labourers and township or municipality authorities must realize that the
native living under insanitary conditions [is] a danger to the public health of the farm
or township and that proper provision must be made for his accommodation under
sanitary conditions if the health of the other communities [is] to remain satisfactory
and economic progress [is] not to be retarded".
However, the extent to which this appreciation impacted on activities is ambiguous. Whilst
not claiming to be representative for SSA, the case of Nigeria in 1946 may be illustrative. In
that year, a ten year national plan was initiated, which sought to allocate 67 per cent of the
capital budget to urban hospitals, 2.5 per cent to rural health centres, with no specific funds
budgeted for health education, immunization, sanitation or other preventative services
(Itayvyar 1992). It was not until the following decade that public health programmes became
more prominent. For example, following a meeting of the World Health Organisation in
Ghana in 1954, the colonial authorities introduced a programme to eradicate the sand fly from
the Volta Lake because it acted as the vector for onchocerciasis (Twumasi 1992). Similar
programs were initiated in other British colonies (World Bank 1993a). However, the extent
that these programs impacted on health expenditure orientation was small. For example in
Rhodesia (later Zimbabwe) in 1974/5, 86.3 per cent of the health budget was spent on
curative services, whilst only 8.7 per cent were allocated to preventative care (Gilmurray et
al 1979).
The system of financing was by no means universal and furthermore was complicated (see
experience from Tanzania in Section E). Bonsi (1992) suggests that in a number of West
African countries that care was provided and financed by the state, with a nominal charge
at the point of services for European administrators and later African employees of the
colonial authorities.3 In other countries, for example Kenya, the government financed and
3. It should be pointed out that care was provided on separate wards. See discussion of the
Nigerian experience in Itayvyar (1992).
Page 15
provided care for Africans, but Europeans financed their own care and sought provision in
private hospitals Mburu (1992).
That colonial authorities focused their services upon a narrow group of individuals, requires
some comment on how colonial authorities raised their finances. If European and African
employees were the only individuals to pay tax, were they the only groups entitled to
provision? Analysis of tax structures in the early colonial period, suggests that the majority
of government finance was raised from hut and poll taxes levied upon Africans (Stock et al
1992). Furthermore, following the First World War (1914-18), many British colonies were
instructed to move towards self-financing. In Tanganyika, over 50 per cent of the government
revenue was raised from local hut taxes prior to the Second World War (Leubuscher 1949).
It has been argued that in these circumstances, the broad base of African tax payers were
essentially subsidising the care of the employees of the colonial authorities (Timuss et al
1964).
For the bulk of the rural population, ‘...scattered rural health facilities were established by
government authorities and financed by tax levies or by missionaries who combined medical
care and Christian proselytization’ (Stock et al 1992).4 In Nigeria the financing of these
services was supported by colonial authorities through grants and by religious organisations
in Europe (Bonsi 1992). Whilst funding continued from the state, grants from missionary
agencies in Europe began to decline.5 This fostered the introduction of charges for services
(Sorkin 1977).
Due to the variety of experiences, characterising the colonial system requires caution. On the
whole, the colonial period would seem to be characterised by coverage for urban,
administrative and commercial centres, vertical publicly financed programmes aimed at
eradicating communicable disease, and extremely poor rural coverage. The vast majority of
Africans continued to pay for and receive care from traditional sources, those enfranchised
within the colonial facilities received subsidised, mostly curative, care. Whilst there is
evidence that preventative care was incorporated into health plans, this did not form a resolute
change of focus. This suggests a health system, providing and financing selective care for
privileged groups.
In countries where Europeans were left to finance their own care and missions sought to
provide care to Africans, the colonial system strongly suggests association with the Type D
system described above. Where European care was subsidised and access guaranteed through
racial and income status, a Type C system is suggested.
By the end of the colonial period, the health systems of many African countries had been
modified. This is evidenced by: expansion of coverage to African and Asian employees of the
4. Bonsi also suggests that this was to encourage Africans to drop traditional medicine which
was considered as evil and idolatrous (1992).
5. For example, government cash grants to Missions in Tanzania increased from £550 in 1938
to £106,129 by 1956/7 (current values) (Titmuss et al 1964). Support by European mission
agencies dwindled to an average of 15 per cent of total costs by the mid 1960’s (Sokin 1977).
Page 16
colonial state; implementation of public health measures by the state; and subsidies extended
to Mission providers to extend coverage to the poor and in rural areas. To affect these
modifications, public finance and production became increasingly more important. However
whilst this suggests the emergence of a Type B system, the health system was driven by a
requirement to offer expanded health care, not as a right to Africans, but more in response
to the needs of: (i) protection of the settler community; (ii) to serve the altruistic purposes of
European missions; and (iii) to enhance the productivity of the colony. As a result, we
classify the vast majority of health systems at Independence as essentially hybrid Type B
systems directed at and serving the requirements of a particular segment of the population.
Independence
Colonial health policies resulted in the inheritance of health systems which were characterised
by segmented coverage and a weak African skill base.6 As a result, most African countries
were poorly equipped to ensure the health status of the mass of the population, or to fulfil the
aspirations for ‘nation building’ associated with independence.
The response of African countries, irrespective of whether the state was capitalist, populist
socialist or Afro-Marxist in ideology, was to make political and resource commitments to
developing the health sector. The focus of activity was upon extending rural coverage and
medical training. This commitment resulted in extensive state activity in the health sector
(Stock et al 1992). For example, in Ghana between 1957 and 1963, the number of health
centres increased from 10 to 41, whilst the number of doctors and nurses increased from 330
to 904 and 800 to 2366 respectively Twumasi (1992).
In addition, newly independent states sought to continue the system inherited at independence,
which was to finance health services through government budgets and absolve users from the
full direct financial cost of consumption. The major difference following independence was
that this was now extended to the mass of Africans which had been effectively excluded prior
to independence. The responsibility for finance inherent in this widespread choice of policy
, coupled with the early emphasis to extend coverage and access to curative services,
established an implicit contract for ‘health’ service provision and financing vested with the
state. This resulted in the wholesale provision of goods with private, mixed and public good
features.
Yet the expansion in government activity masks significant differences in health sector policy
towards non-government activity. Moreover, within countries policies towards the spectrum
of non-government provision (missions, industrial, traditional and private practice) was not
uniform. For example, in Tanzania, private practice tor Government doctors was banned
whilst missionary activity was encouraged (Titmuss et al 1964). In Zimbabwe, health policy
6. During the colonial era, most senior medical positions had been staffed by expatriates. At
Independence, many expatriates left, with the result that there was a dearth of qualified
physicians and senior medical personnel. In some countries, expatriate medical personnel left
with colonial withdrawal (eg Zaire) leaving a skill gap. In others, expatriates left as a result
of government policy of ‘Africanisation’ (eg Tanzania), again leaving a skill gap. Most
observers agree that the failure to train local staff was a serious failing of colonial authorities.
Page 17
allowed a continuation of all the forms of non-govemment provision that had existed prior
to Independence but was bolstered by increases in government financing (Gilmurray et al
1979). In Angola and Mozambique, private practice was banned in 1975, but private sector
companies were allowed to operate health facilities (Roth 1987). In Benin, Cameroon, Chad
and Togo, non-state care was discouraged, but not made illegal (World Bank 1987).
Moreover, despite the dominance of the state, it is vital not to ignore the impact of nongovemment sources of finance. Griffiths and Mills (1983) indicate that until the adoption of
Health Sector Financing and Expenditure Surveys (HSFES) in the mid 1970’s, the analysis
of health systems in many LDC’s had centred upon government health expenditure as an
explanatory variable and ignored the contributions of the private sector. When the first HSFES
were conducted in SSA they produced some interesting results. For example, in Rwanda in
1977, it was established that only 21.5 per cent of total financing for health service
consumption had originated from government sources. In Botswana a figure of 45 per cent
was established (Ibid 1983). This evidence clearly indicates the need to establish the extent
of the dominance of the state, from actual shares in expenditure. Unfortunately, it was not
until World Bank (1993a), that such data for the region was systematically presented.7
The fact that information on the impact of government expenditure is impaired by a lack of
data on the volume of non-government activity is further compounded by the considerable
difficulty in translating the impact of (even aggregate) figures on public health expenditure
upon health status. As indicated in Section B, consideration of the end use of resources is
paramount to understanding the likely impact on health status.
This diversity in policy and financing makes the classification of health systems as presented
in Section B problematic. What is required is an in depth study of the health policies adopted
at Independence by all SSA countries. However, we have not found studies which go beyond
a classification of health systems according to political ideology. We consider that unless
examination of policy towards all non-govemment health activity is undertaken, it is unclear
whether the state actively seeks to suppress health markets or merely seeks to dominate health
markets.
In spite of this lack of information. We have nonetheless classified countries within the
framework presented in Section B. This basis for our classification is the financing structure
adopted for public services by newly independent governments. We based our classification
on a survey conducted by Nolan & Turbat (1993) which examined the financing of health
services in SSA. From the survey, the authors found no evidence of attempts to recover full
costs from users of public health services, or to differentiate between charges for curative and
preventative services. The findings of the authors effectively delimits the initial starting points
for the countries reviewed to Type A and Type B systems. But differentiation between Type
A and Type B is possible since the degree of subsidisation varied. Not all governments opted
for zero charges to users. Many countries instituted minimal charges for certain services, both
to attempt to recover costs and to act as a rationing device (Nolan & Turbat 1993). Table 3
1. See Table 13, World Bank (1993a).
Page 18
was compiled which categorises countries into Type A (no direct charges for services) and
Type B (nominal direct charges).8 The classification is presented in Table 3.9
Table 3: CATEGORIZATION OF SSA COUNTRIES BY HEALTH SYSTEM
TYPE A
TYPEB
Angola
Benin
Botswana
Burkina Faso
Burundi
Central Africa Republic
Congo
Equatorial Guinea
Guinea
Madagascar
Malawi
Mali
Mauritania
Mozambique
Niger
Nigeria
Sao Tome and Principe
Sierra Leone
Sudan
Tanzania
Uganda
Zambia
Cameroon
Ethiopia
Ghana
Guinea-Bissau
Ivory Coast
Kenya
Lesotho
Namibia
Rwanda
Senegal
Swaziland
The Gambia
Togo
Zaire
Zimbabwe
TOTAL (22)
TOTAL (15)
From the early 1970’s, the majority of SSA governments sought to reorientate their health
interventions towards the financing of public-good type and preventative services (Sorkin
1977). Moving on from the earlier emphasis on extending curative coverage, policy was
refined. For example, ‘the "health goals" in the 1974-78 Kenyan National Development Plan
are stated as "...to control, prevent and ultimately eliminate communicable disease, deficiency
conditions [and] environmental health hazards...These health goals can be achieved more
rapidly if promotive and preventative programs are emphasised rather than curative"’ (original
quotation Sorkin 1977).
It has been argued that the rhetoric for preventative care has not been matched by resources
(Segall 1993). Moreover there is very little evidence to suggest that financing systems were
8. Data is presented on 37 of 44 SSA countries. Data for Somalia, Liberia, Cape Verde, Chad,
Djibouti, Comoros and Gabon was not reported in Nolan & Turbat 1993).
9. The classification leaves much unconsidered, but given the financing focus of this paper,
offers insight into the ramifications of different financing strategies upon health status.
Page 19
adapted to achieve this policy. Even by the mid 1980s, there was virtually no differentiation
of charges between curative and preventative care (Nolan & Turbat 1993).
Health system performance
Notwithstanding the data limitations highlighted above. We present data on the performance
of a sample of countries categorised in Table 3. Our earlier identification of expansion in
government expenditure is reflected in empirical data on the proportion of GNP spent by*
governments on the health sector. The data is reproduced in Table 4 and represented
graphically in Figure I.10
Public Health Expenditure
As Percentage of GNP
rs-r
1.6-z......
1.4-z......
^•2 7
l i-l
z
f ..
0.6---------
04z......
0.2-z......
0K=
I I
Type A Av.
Type B Av.
C 1960 M 1972 M 1980 M 1990~l
Figure 1.
The data indicates that between 1960 and 1970 , both Type A and Type B countries,
increased the proportion of GNP expended on public health. Type A countries actually
managed to surpass that expended by Type B countries. Between 1970-80, both groups
reduced the proportionate expenditure on health, but the fall was more pronounced for Type
A countries. By 1990, Type A countries had failed to stem the fall in proportionate
expenditure which fell from 1.27 to 1.03 percent of GDP. Type B countries, on the other
hand, achieved a substantial increase in average expenditure from 1.25 per cent to 1.76 per
cent of GDP
10. The sample in Table 4 and other tables in the text is determined by data availability as
oppose to statistical selection. For all other SSA countries, historical data on government
expenditure on health was not available from the consulted sources. Data is presented for
three decades to illustrate the progression from the early independence period in SSA up to
the present time. Further, the sample is stratified into ‘type’ groups, as determined by Table
3. For the purpose of comparability this sample is maintained for the majority of the
remaining empirical data presented in the paper.
Page 20
It is useful to compare the data in Table 4 and Figure 1 with the data in Table 5 on trends
in economic and population growth.
Table 5 reveals that, for the sample of countries, the average rate of growth of national
income (GDP) only fell below population growth for Type A countries (when considered as
a group) between 1980-90. This implies that the falls witnessed by both Type A and B
countries in proportionate expenditure between 1970-80, occurred at a time when economic
growth per capita was still rising. For the decade 1980-90, falls in economic growth per capita “
for Type A countries were accompanied by continued falls in proportionate health
expenditure, whilst increased proportionate public health expenditure was associated with per
capita economic growth for Type B countries.
Given the widespread policy of minimal charges at the point of service delivery, it is possible
to assume that the public expenditure figures presented in Table 4 do not contain significant
elements of private finance of public services. However, it is not possible, due to a lack of
data, to estimate whether Table 4 includes aid finance directed through the government
budget. Furthermore, because there is no similar data on private sector expenditure, it is
problematic to conclude that the increases in proportions represent significant additions to
total health expenditures.11
In an attempt to gauge the impact of increased government activity, Table 6 and 7 present
data on changing health status in SSA as measured by Life Expectancy Rate (LER) and Infant
Mortality Rate (IMR) between 1960 and 1990 for the countries sampled in Table 4. Data on
the rate of change between time periods is illustrated in Figures 2 and 3 for LER and IMR.
Percentage Change in Life Expectancy
1960-70,1970-80 & 1980-70
20<
18-/'
I
o
16-z'
...
14-z
O 12-z '
8) io-z '
..
js
§
cj
8-z
6^”
4-/-
2-Z'
o-k-
Type A Av.
II
M 1960-70 M 1970-80
Type B Av.
1980-90
Figure 2.
11. For example, if the volume of private finance grew more quickly, then the increased
proportion witnessed in Table 4, would represent a falling proportion of total expenditure in
health. The opposite case is also possible.
Page 21
TABLE 4: Public Health Expenditure as a Percentage of Gross National Product.
COUNTRY
TYPE
Botswana
Burkina F.
Ethiopia
Ghana
Kenya
Malawi
Rwanda
Sudan
Tanzania
Uganda
Zambia
B
A
B
B
B
A
B
A
A
A
A
Sample Av.
Type A Av.
Type B Av.
1960
1972 *(1)
1980 *(1)
1990 *(2)
1.5
0.6
0.7
1.1
1.5
0.2
0.5
1
0.5
0.7
1
2.02
0.89
0.78
1.22
1.65
1.21
0.66
1.03
1.41
1.15
2.51
1.97
0.81
0.86
0.76
2.03
2.06
0.64
0.27
1.72
0.31
2.44
3.8
0.8
1.6
1.2
1.7
1.7
0.5
0.4
0.7
0.5
2.1
0.85
0.67
1.06
1.32
1.37
1.27
1.26
1.27
1.25
1.36
1.03
1.76
*(1) Data derived from WB tables.
*(2) GDP data.
Sources
UNDP (1991)
World Bank (1987)
World Bank (1993a)
World Bank (1993b).
Page 22
TABLE 5: Average Annual Growth Rates of Gross Domestic Product (GDP) and Population.
1970 - 1980
GDP Pop
1980 - 1991
GDP Pop
COUNTRY
Type
1960 - 1970
GDP Pop
Botswana
Burkina F.
Ethiopia
Ghana
Kenya
Malawi
Rwanda
Sudan
Tanzania
Uganda
Zambia
B
A
B
B
B
A
B
A
A
A
A
n.d.
3.0
4.4
2.1
6.0
4.9
2.7
1.3
6.0
5.9
5.0
n.d.
1.6
2.4
3.0
3.4
2.8
2.6
2.2
2.7
3.7
3.0
14.5
4.4
1.9
-0.1
6.4
5.8
4.7
5.6
3.0
n.d.
1.4
3.8
2.1
2.7
2.2
3.8
3.1
3.4
3.0
3.0
2.7
3.0
9.8
4.0
1.6
3.2
4.2
3.1
0.6
n.d.
2.9
n.d.
0.8
3.5
2.6
3.1
3.2
3.8
3.3
3.0
2.7
3.0
2.5
3.6
4.1
4.4
3.8
2.7
2.7
2.9
4.8
4.0
5.5
3.0
2.8
3.2
3.4
2.7
3.9
3.1
3.0
3.3
Sample Av.
Type A Av.
Type B Av.
Sources
World Bank (1980)
World Bank (1993a)
World Bank (1993b).
Page 23
TABLE 6: Life Expectancy Rate
COUNTRY
Type
1960
1970
1980
1990
Botswana
Burkina F.
Ethiopia
Ghana
Kenya
Malawi
Rwanda
Sudan
Tanzania
Uganda
Zambia
B
A
B
B
B
A
B
A
A
A
A
45
37
34
45
41
37
37
40
42
41
40
49
40
43
49
50
41
44
42
46
50
46
50
39
40
49
55
44
45
46
52
54
49
60
48
45
55
60
48
49
51
54
52
54
39.9
39.5
40.4
45.5
44.2
47.0
47.5
47.3
47.8
52.4
51.2
53.8
Sample Av.
Type A Av.
Type B Av.
n
Sources: UNDP (1991), World Bank (1982), World Bank (1983), World Bank (1984)
TABLE 7: Infant Mortality Rates
COUNTRY
Type
1960
1970
Botswana
Burkina Faso
Ethiopia
Ghana
Kenya
Malawi
Rwanda
Sudan
Tanzania
Uganda
Zambia
B
A
B
B
B
A
B
A
A
A
A
116
252
175
143
138
207
147
168
152
139
151
101
178
158
111
102
193
142
149
132
109
106
78
152
155
100
83
169
172
123
119
113
90
38
134
132
85
67
149
120
102
115
117
82
163
178
144
135
145
123
123
128
118
104
117
88
Sample Av.
Type A Av.
Type B Av.
1980
1990
I
Sources: World Bank (1983), World Bank (1989), World Bank (1992), World Bank (1993b).
Page 24
Percent Change in Infant Mortality Rate
1960-70, 1970-80 & 1980-90
30Z
25-x
o
J 20'Z '
Q
iS
15 •z''
z■
§
e io-z
0-
liJH
Type A Av.
1960-70 o 1970-80
Type B Av.
1980-90 |
_________
Figure 3.
The data contained in Table 6 and 7 indicates that, when considered as a group, both Type
A and Type B countries have registered improvements in LER and IMR. Whilst starting at
different levels in 1960, by 1990, Type A countries had reduced the IMR by 66 per cent,
whilst Type B countries had reduced the rate by 63 per cent. For LER, the corresponding
figures are 30 per cent for Type A and 33 per cent for Type B countries.
The rate of change over the period is illustrated in Figures 2 and 3. The pattern which
emerges from these figures is high initial increases in the indicators followed by declines for
the 1970-80 decade. By 1980-90 however, with the exception of IMR for Type A countries,
the indicators improve on their 1970-80 showing. Whilst Type B countries experienced more
dramatic falls in the 1970-80 decade, the increases registered in the 1980-90 decade are far
higher than for Type A countries.
In spite of the general relationship between per capita growth in income, increased allocations
of GNP to public expenditure and health status, the lack of consideration of the impact of
private expenditure and aid flows the health status data requires some attention.
The volume of aid flows to SSA is presented in Table 8. This data shows aid to the health
sector and public health expenditure for 1986 and 1990. During the period the ratio of health
sector aid to public health expenditure increased for both Type A and B counties, but was far
more greater for Type A as opposed to Type B countries. When considered the data presented
earlier on public finance it is possible to hypothesise that falls in Public finance have been
replaced by aid in Type A countries.
Data on the proportion of private health expenditure within total health expenditure is
presented for as many countries as it was possible to generate two time points in Table 9. Due
to scanty data it was not possible to generate a uniform period of analysis and, as such, the
data presented in Table 9 should be considered as indicative rather than conclusive. The data
Page 25
reveals that of the 13 countries sampled, eight experienced increases in private expenditure
contributions whilst five registered declines in the contribution of private expenditure. When
considered as ‘types’, the majority of Type A countries registered falls in private sector
contributions, whereas for Type B, the majority had increased private sector contributions to
total health expenditure. However, from 1980 onwards, the majority of Type A countries
experienced increases in private sector contributions.
Whilst this data is by no means comprehensive, -when considered together, the data implies
a pattern of substantial increases in aid in Type A countries coupled with increased private
expenditures during the 1980’s. For Type B the pattern is smaller increases in aid allocations
and increased private sector contributions.
The importance of aid flows and the contribution of private expenditure is illustrated in Figure
4, which shows the profile of total health expenditure for 1990 for the sample of countries
in Table 4. For Type A countries, private expenditures are the largest single source of finance
for health care, followed by aid flows, with the smallest contribution from government. For
Type B countries, public and private expenditure share equal prominence as the largest
sources of finance, with aid contributing the smallest proportion to the total.
Health Expenditures
1990
IDO-
90SO7060-
§ 502
al 403020-
10-
0-
Type A Av.
Aid Flows
Type B Av.
Public Expenditure IgaSI Privats Expenditure ||
Figure 4
The lack of data on financing shifts, makes interpretation of the financing profiles indicted
in Figure 4 complex. Has public finance always been so small in Type A countries? If not,
why has it become so diminutive for Type A countries by 1990?
On the former, we would argue that the evidence presented in this paper, indicates that public
expenditure has fallen in Type A countries from previous recorded levels. On the latter, we
would argue that during the 1980s, Type A countries reduced public health expenditure at a
time of negative per capita growth. On the other hand Type B countries, which experienced
positive per capita growth, were able to increase public health expenditure.
Page 26
The data presented in this section exemplifies the fact that attempts to draw inference from
public health expenditure alone misses over 50 per cent of the full expenditure picture. Unless
the full picture is considered, we strongly doubt that accurate inference of impacts on health
status are possible. Moreover, the lack of historical data on relative shifts in finance presents
a serious constraint to analysis. Needless to say both indicate areas in need of research.
Page 27
TABLE 8: Government Expenditure and Aid Flows to the Health Sector (1986-90).
(Millions of Constant US Dollars)
Public
Expend.
(1986 US $)
Aid as %
of Public
Aid
Flow
Public
Expend.
(1990 US $)
Aid as %
of Public
7.9
8.6
45.7
9.9
17.3
86.4
8.6
37.6
6.8
59.8
113.2
25.7
14.4
32.0
158.0
43.1
26.9
33.3
83.1
26.4
120.0
21.3
94.6
71.5
150.1
32.4
26.7
741.8
45.6
37.6
55.4
75.6
265.5
41.3
375.8
254.0
6.3
Aid
Flow
COUNTRY
Type
Botswana *(1)
Burkina Faso
Ethiopia
Ghana
Kenya
Malawi
Rwanda
Sudan
Tanzania
Uganda
Zambia *(1)
B
A
B
B
B
A
B
A
A
A
A
Sample Av.
Type A Av.
Type B Av.
24.1
2.2
11.5
67.3
7.9
26.9
35.7
28.3
42.9
24.5
29.2
13.6
59.0
32.0
4.8
13.4
10.6
44.6
27.6
72.9
35.6
44.0
21.7
46.0
48.7
42.9
18.0
11.0
32.9
15.7
12.6
76.5
58.1
31.9
89.4
175.0
249.1
86.1
Notes: *(1) = data for 1988
Sources:
OECD 1992
World Bank 1993a
Page 28
TABLE 9: Private Sector Contributions to Total Health Expenditure
COUNTRY
TYPE
Botswana
Burkina F.
Ghana
Kenya
Lesotho
Malawi
Mali
Rwanda
Senegal
Sudan
Togo
Zambia
Zimbabwe
B
A
B
B
B
A
A
B
A
A
B
A
B
1970
1977
1978
1979
1980
1981
1982
1984
21
19
73
41
12
23
54
37
39
41
31
26
46
17
1990
22
20
52
39
26
42
47
49
40
84
39
31
49
Sources:
de Ferranti (1985)
World Bank (1993a)
Bloom & Segall (1993)
Lee & Mills (1983)
Page 29
D. Health Sector Performance in SSA and Reform Options
The drastic slow-down in improvement witnessed in Health Status during the 1970’s, the
increased importance of aid in total health financing, and the breakdown in the financing
arrangements estabhshed at independence symbolised by the increasing dominance of private
financing have been associated with three explanations:(i) resource inadequacy; (ii) a failure
to use resources efficiently; and (iii) government policy failure.
Resource inadequacy
There are two elements to the explanation of resource inadequacy. The first relates to the
national income level in SSA. The second relates to the decline in those incomes since the
mid 1970s.
The first element stresses that the level of income in many SSA countries is so low that
governments are unable to adequately fund health interventions. It relies on the assumption
that low-income is the cause of low levels of public expenditure.
That SSA countries have low levels of income is undoubted. Of the worlds poorest thirty
countries in 1993, 23 were from SSA (World Bank 1993b). In the same year the ten poorest
SSA countries had an average per capita income of US$ 164. Examination of public health
expenditure by these ten countries, for the most recent data year (1990), reveals that average
public health expenditure was USS 1.9 per capita, or approximately 0.1 per cent of GNP
(World Bank 1993a). In 1993, the World Bank estimated that the average cost of a basic
package of health care in low income SSA would be US $13.22 {Ibid).
These figures indicate that low income is associated with inadequate levels of public
expenditure, but do not signify why such low levels of expenditure are recorded. Some have
argued that low levels of public health expenditure result from ‘political choice’ rather than
low income (Lal 1994). The argument is that given that public expenditure on health only
occupies on average 5-7 per cent of the national budget, that notions of underfunding and
resource constraints reflect decisions by governments to fund other interventions in preference
to health {Ibid). This is confirmed by research undertaken by the UNDP on a sample of 25
developing countries (5 of which were from SSA), which shows wide variation within income
groups of public social expenditure as a proportion of government budgets (UNDP 1991). For
example, social expenditure as proportion of the total public budget was 39 per cent for Sierra
Leone and 20 per cent for Nigeria in 1988 despite income levels being 300 in the former and
290 in the latter. Zimbabwe with a per capita income of USS 650, managed to allocate 49 per
cent of its budget to social spending, whilst Botswana with a per capita income of USS 1,010
managed less than Sierra Leone at 37 per cent {Ibid). The conclusion following from this
research is that countries should attempt to redirect finance towards social interventions {Ibid).
[Others have argued that public expenditure in SSA is under crisis because of debt
repayments and austerity imposed through structural adjustment...]
Page 30
Whilst the above suggests that decisions about sectoral allocations may be more important
than income level, the experience of declining incomes across SSA has added the second
element to this explanation.12
Data presented in Section C suggested the following relationships with income. First, per
capita national income growth and growth in public health expenditure as a proportion of
GNP can explain the initial increases in 1960-70. Second, although per capita income growth
continued in the 1970-80 decade, it was associated with falls in public health expenditure as
a proportion of GNP, which is associated with dramatic falls in the rate of improvement in
health status. Third, by 1980-90, increased growth in income and the proportion of GNP
expended on public health can explain the dramatic increase in health status of Type B
countries (34 per cent in IMR and 12.5 per cent for LER). Finally, falling per capita national
income, coupled with falls in the proportion of GNP expended on public health, explain the
small improvements registered by Type A countries in health status for 1980-90 (9 per cent
in IMR and 8 per cent for LER).
In addition, data presented in Better Health in Africa, shows that following 1986, only higher
income SSA countries (Botswana, Lesotho, Mauritius, Swaziland and Zimbabwe), were on
average able to increase resource allocations to health, whilst the remainder of other countries,
when considered collectively in middle and low income categories, registered falls in
government allocations to the health sector (World Bank 1993a). This implies that income
level is an important explanation of allocations to public health budgets and that low income
countries in Africa have been unable to maintain allocations to health budgets.
That government allocations have resulted in underfunding of public services is confirmed by
a recent review of World Bank ‘Public Expenditure Reviews’ (PER)s across 30 SSA
countries. In the review 24 of the 30 countries surveyed were considered to have an
underfinanced public health sector (McGory 1993). In the majority of countries the major
problem was the underfunding of recurrent budgets. This is supported by widespread evidence
of underfinancing of maintenance, drugs, and manpower budgets in SSA (World Bank 1993a).
At the centre of this explanation is a belief that the ills of the public sector can be explained
by underfunding. Moreover that the problems experienced in the public sector can be rectified
by increasing resource allocations. Various reforms are open to governments in the pursuit
of increasing funding. The elements of which are presented in Table 10.
The strategy has two components. The first is at the health sector level, which requires the
government to increase resources by a combination of user fees on health services, ‘ear
marked’ taxes, intersectoral reallocations and increased aid bids. At the central budget level,
the government should attempt to improve revenue collection to improve resource availability
at the budget level.
12. Average GNP per capita for SSA declined from 1.9% per annum between 1965-73 to 1%
p.a. between 1973-80 to -1.2% p.a. between 1980-90 (World Bank 1993b).
Page 31
L'BRARY
infokma
TABLE 10: Resource Underfunding
COMPONENT OF STRATEGY
INSTRUMENT/ METHOD
OBJECTIVE
Increase Resources in Health Sector
- levy user fees for health services
- raise resources from users of health services.
- introduce ‘ear-marked’ taxes
(eg taxes on tobacco and alcohol)
- raise resources specifically to finance health services
- reallocate resources from other areas (eg military budget)
- intersectoral reallocations
- increase levels of overseas resources allocated to health services
- call on international aid
Increase Government Resources
- improve tax collection
- increase revenues entering government budget
Assessment of strategy
This strategy seeks to refinance the public sector following falls in national income levels
which have themselves impacted upon health budgets. However, despite considerable evidence
that underfunding has impacted upon health budgets. The strategy of increasing resource
allocations to the health sector as an objective of reform has not been fully embraced by
governments. Indeed of the elements outlined in Table 10, only increased aid and user fees
have been adopted widely. We have found no evidence that although intersectoral
reallocations have been recommended as a solution to underfunding, little progress has been
forthcoming (World Bank 1993a). Moreover, we have found no evidence that improved tax
collection programs have as yet impacted upon government finances (World Bank 1989).
Finally, although recent evidence from Brazil suggest that ear-marking tobacco and alcohol
tax for health expenditure has stabilised the flow of resources to the health sector, there is no
evidence that similar taxes have been earmarked for health expenditures in SSA (World Bank
1993b).
In Section C, we presented data on the importance of aid finance in SSA. There is little doubt
that aid has contributed to sustaining both public services and in providing services (mostly
vertical programs) to households in SSA (World Bank 1993a). However, the extent to which
increased aid represents a sustainable solution to underfunding has provoked concern in the
literature. First, some have argued that increased aid to the health sector introduces
dependency and introduces issues of sustainability of aid’s contribution. Second, there is
concern that the focus of aid expenditure reflects donors priorities more than the priorities of
countries and is biased towards capital intensive projects (World Bank 1993a). With the
corollary that capacity to finance recurrent expenditure may not match donor financed capital
investments. Finally, there is some suspicion that as aid as increased, it has reduced the
pressure on government to raise its own allocations to health (Lawson 1994).
Within this strategy, user-fees have two functions. The first is to increase health resources at
the budget level. The second is to raise resources at the facility level.
Analysis of the overall contributions to Health Budgets (note this excluded non-health
ministry health expenditure), have been reported by Vogel (1990). In a sample of twelve
countries the contribution of user fees to total recurrent budgets for 12 countries ranged from
less that 1 per cent in Burkina Faso to 15 per cent in Ghana, with the average around 5 per
Page 32
cent. Nolan & Turbat (1993), in a review of 37 SSA countries found that in a small number
user fees were contributing between 5-10 per cent of total recurrent expenditure, but the
majority of countries did not approach this figure. The findings are confirmed by Creese &
Kutzin (1994), McPake (1993) and Shaw & Griffin (1994) who confer with a figure of 5 per
cent or less.
Whether these increases at the budget level represent increased allocations to total health
expenditure is unclear. Due to inadequate data it is not possible to estimate whether or not
fee revenues are funds diverted from what would have been expenditure on private health.13
The contribution of fees to particular facilities differs. In a review of impacts on specific
public hospitals and facilities, Creese & Kutzin (1994) indicates that the contributions of user
fees to total recurrent budgets was as high as 78 per cent in a Zairian hospital in 1988, 32 per
cent in a sample of urban hospitals in Ethiopia between 1984-85 and 14.8 per cent in a
hospital in Niger in 1986-7. Shaw and Griffin (1994), also find mixed evidence. They found
that public maternity hospitals in Nigeria raised 82 per cent of recurrent expenditure through
fees, whilst a sample of 42 public hospitals in Zambia in 1989 raised only 3 per cent.
Evidence of the impacts at the primary health care level can be gauged by examination of the
experience of those countries participating in the Bamako Initiative (BI).14 The focus ot the
BI, differs between countries. For example, in Nigeria it is implemented at the district level
and in Benin at the village level (McPake et al 1993). Cost recovery methods also differ,
whilst the majority use a system of user-fees, some use pre-payment schemes (Korte et al
1992). Local conditions, such as affordability and extent of poverty tend to figure highly in
the design of cost-recovery structures. These factors caution against a heavy reliance on
comparative performance analysis. However data on performance has been reviewed by a
number of scholars. Some authors note that cost recovery is as high as 100 per cent in Benin
and 79 per cent in Zaire (McPake et al 1993). Others note that it is over 50 per cent in
Guinea, Mali and Senegal (Creese & Kutzin 1994). Whilst there is no evidence on whether
the level of funds simply diverts funds that would otherwise have been used in private clinics.
13. Data on total expenditure in SSA is only reported for 1990 (see World Bank 1993a).
14. Many LDC governments began to formally recognise the importance of PHC in a strategy
of improving health and these threads led to the UNICEF/WHO Alma Ata Declaration in
1978. The principle focus of declaration was to achieve ‘health for all by the year 2000’
through concentrating resources upon Primary Health Care (PHC). Segall (1983) caiims that
the goals of Alma Ata were never fully supported by resource commitment with the result
that curative allocations still dominated resource use. However the flame of Alma Ata was
rekindled at a meeting of African Minister’s of Health sponsored by UNICEF and WHO in
1987. This meeting sought to address the problems of inadequate resources for PHC through
a new strategy of community financing and participatory management, with somewhat less
emphasis on central government support than in the declaration of 1978. This was called the
Bamako Initiative (BI). The new strategy was African based and sought to use private finance,
through the use of user fees for services, as the main instrument for achieving (after five
years) the goal of self-reliance attainment of universal accessibility to PHC (McPake et al
1993).
Page 33
these figures illustrate the potential that fees can generate funds which can then be used to
sustain and/or improve service delivery. However it has been argued that the target of
sustainability inherent in the BI suggest that figures below 100 per cent temper overall
success of these projects. This argument stresses that given that the target is for recovering
100 per cent of recurrent costs, that vital questions of resource needs for capital investment
(buildings, refrigerators, vehicles, etc) are not addressed within the fees structures (World
Bank 1993b).
In summary the weight of evidence indicates that fees are unlikely to generate the level of
resources required at the budget level to reverse the declines described in Section C.
Moreover, at the facility level, reported evidence is that fees can be a vehicle for raising
significant proportions of recurrent cost funding requirements (Mills 1994). However, even
in programmes administered under the BI, there seems very little evidence that full recurrent
budgets can be met from user fees and the question of funding capital expenditure still
remains (World Bank 1993b).
That the state has been unable or unwilling to allocate its ‘own’ resources to health services
suggests that countries in SSA are unlikely to meet the funding requirements of present health
systems. As public finance has fallen more sharply in Type A countries we would argue that,
as a system, the Type A model is increasingly threatened and unlikely to be sustainable.
Indeed critics now argue that the underfunding argument has been misleading (Lawson 1994).
What is being underfunded? How is underfunding determined?
Lawson (1994) has identified, there are three interpretations of underfunding. Theses are that
public funding is inadequate: (i) for supply to meet demand; (ii) to meet government targets;
and (iii) to sustain existing facilities at the operational levels for which they were designed.
In this schema, the level of underfunding will vary depending on which interpretation is
selected. The choice of interpretation will in turn be related to important questions about the
role of the state: Should the state finance the market in health care or should private finance
prevail? Are health targets commensurate with resource availability or should they be
abandoned in favour of other targets? Should existing facilities be upgraded to reach
operational capacity or rationalised?
It is these concerns which have shifted attention to stress other reform strategies which focus
on: (i) the use of existing resources; and (ii) government health policy.
Failure to use resources efficiently
This explanation stresses the importance of how resources are used to produce health
outcomes. The implication is that health interventions should be judged upon their
effectiveness in improving health status and that costs need to be brought under control by
improved allocative, productive and technical efficiencies (Mills 1994).
It has been argued that in spite of rhetoric aimed at promoting preventative care, resources
have remained channelled to curative care (World Health Organisation 1986). For example
in Nigeria between 1976 and 1985, of all newly trained physicians, only 9 out of 189 trained
by the state specialised in public health (Ubot 1992). Figures recorded for Ghana in 1975/6,
Page 34
illustrate that 75.9 per cent of government expenditure was allocated to curative services
whilst 9.7 was allocated to preventative services (Twumasi 1992). In Chad in 1988, in
Lesotho in 1984 and in Nigeria in 1985 the allocations to curative care were 97 per cent, 70
per cent and 80 per cent respectively (Sahn & Bernier 1993).
A dimension of the allocation of increasing resources to curative care is that it results in
increasing proportions allocated to hospitals budgets. Available data on resource allocations
from central budgets to hospital services illustrate the growth in allocations to hospitals and
is reproduced in Table 11. Of eight countries with available data, six increased allocations to
hospitals. Whilst the size of this sample cautions against reaching conclusions for SSA, it is
interesting to note that, Swaziland excepted, increases in allocations to hospitals occurred
regardless of increases or decreases in real per capita health expenditures by governments
(Table 11 column 3).
TABLE 11: Hospital Health Expenditures
Type
Percentage of Public
Health Expenditure on
Hospitals
Year
(%)
Index of Real
Government Public
Health Expenditure per
capita.
Botswana
A
1979
1984
42
49
100
160
Kenya
B
79/80
85/6
66
73
100 (a)
85 (b)
Lesotho
B
70/1
81/2
64
71
Malawi
A
76/7
85/6
73
81
100 (c)
137 (b)
Swaziland
B
76/7
83/4
69
100 (c)
101 (b)
52
Tanzania
A
71/2
78/9
52
63
Zambia
A
1975
78/9
52
63
The Gambia
B
79/80
85/6
67
45
100
77
(a) 1979; (b) 1983; and (c) 1976.
Data Sources: Barnum & Kutzin (1993); Nolan & Turbat (1993); World Bank (1981); World Bank
(1987); World Bank (1989).
Given the urban location of many hospitals this has prompted some observers to highlight the
inequity in resource flows to rural and urban areas. For example, the three central hospitals
in Zambia consumed 45 per cent of the total Ministry of Health budget during the 1980’s,
leaving the remaining 55 to cover 39 other hospitals (World Bank 1993a). And in Liberia, the
Page 35
JFK Hospital in Monrovia, consumed 45 per cent of the public health budget during the early
1980s (Stock et al 1992). The ramifications of this imbalance is poorer provision in rural
areas. For example, in 1979 in Ogun State of Nigeria, the ratio of doctors to population was
1:11,245 in the Ogun state capital, but only 1:200,861 in the rural area of Ijebu North (Pearce
1992).15
Within the charted shift of resources to hospitals, is evidence of the concentration of resourcesin salaries with an associated decline in resources for complementary inputs and for
maintenance. Korte et al (1992) produce summary data on SSA for the 1980’s, which
suggests that the typical experience in SSA was that 80 per cent of recurrent budgets were
devoted to salaries, whilst drug budgets were cut by up to 50 per cent. A survey in Nigeria
in 1987 found that 30 per cent of medical equipment in hospitals was not in use due to a lack
of maintenance (World Bank 1993a). Similar results were found in Ghana in 1987, where
only 167 out of 660 Ministry of health vehicles were road-worthy and a survey of hospital .
equipment in Uganda in the same year found only 20 per cent of equipment in working
condition (Ibid), Moreover, even salaries have became difficult to finance, with many health
service workers failing to get salaries in some countries(Mills 1994). That medical personnel
depend upon supplies to execute their duties and, like others, are motivated by salaries. It is
clear that the above will impact upon efficiency.
Strategies associated with this explanation are presented in Table 11 and have three
dimensions. First, at the aggregate level, governments should reexamine health interventions
and establish those that carry a clear priority for government financing. Services with
significant ‘social’ benefits should be considered above those with mostly ‘private’ benefits
(eg immunization versus dental care). In broad terms this necessitates concentration of
resources in the financing of public goods and primary health care services.
Second, at the institutional level, the emphasis is upon the management, accountability and
use of resources. Using existing resources more efficiently at the institutional level requires
improving the technical efficiency of existing provision. Many approaches can be used to this
end. For example: decentrahsation of key resource allocation decisions from a central ministry
to a region or to individual facilities can improve flexibility and control; regeneration of
auditing and improved record keeping improve information concerning the flow of resources;
and transparent regulations governing the use of resources can improve accountability.
The final dimension is to institute change at the level of the user. Gone is the commitment
to finance and provide care for all irrespective of means. In its place are attempts to introduce
direct payments for those able to pay, with a residual commitment to finance and provide care
for those unable to pay. The state however maintains the function of financing those services
with recognised social benefits (i.e. those that economic theory would predict would be under
provided in a free market).
15. Twumasi suggests that an explanation for this is the lack of interest in medical
professional to work in rural areas. For example in Ghana, a transfer to the rural area is
generally perceived as a punishment (1992).
Page 36
TABLE 12: Resource Inefficiency
COMPONENT OF STRATEGY
INSTRUMENT/ METHOD
OBJECTIVE
Decentralisation
- autonomy in the raising and use of
finance for health services
- engender responsibility in resource use through attention to cost; improve
local accountability with financiers of public services (users and local tax
payers).
- service provision autonomy
- allow provision of services related to local demand and conditions
- management autonomy
- improve local planning and budgeting
- create sectoral agencies
- to vest responsibility for particular functions outside main bureaucracy or
locate in private sector to foster improvements in technical efficiency.
- introduce management /administrative
training
- improve technical capacity of administrators.
- improve promotion /disciplinary
procedures; introduce performance related
incentives
- improve incentives for professional practice.
- levy charges at tertiary level
- encourage the consumption of preventative services with recognised positive
externalities.
Management Reform
Encourage Appropriate Use of
Government Facilities
- levy charges for curative services, shift
subsidies from hospitals to lower level
facilities and subsidise preventative
services.
Reform of Ministry of Health
- encourage use of more ‘cost-effective’ services at lower level facilities.
- bureaucratic reform
- clarify and (redefine) role and functions of institutions and levels of
provision (Community, district, region, national) within government health
system.
- policy planning and management
training
- define priorities for action and strategies for improving health regulation,
provision and financing.
I
Page 37
Assessment of strategy
This section looks at the evidence of implementation of elements described in this strategy
and the implications of adoption upon the character of the health system.
Introduction of fees
Many countries have already adopted certain elements of the strategy identified in Table 7.
Of particular importance has been the adoption of fees as a first attempt at reform (Nolan &
Turbat 1993). Accordingly, a body of evidence has developed which allows examination on
the extent to which this particular component has impacted on the health sector. The main
impacts associated with fees in Table 12 are that they can foster: (i) a reorientation in the use
of facilities to lower levels (consumer behaviour); (ii) an improvement in the quality of
services offered; and (iii) an improvement in equity.
(i) reoriented use offacilities to lower levels (consumer behaviour). There is limited evidence
on this impact. The World Bank (1993) indicates that in Malawi, fees for different levels of
hospitals have been devised to encourage appropriate use of first level facilities. This is in
contrast to Senegal, where introducing charges at lower level facilities but not at higher level
facilities resulted in an under utilisation of the former and an over-utilisation of the later. This
finding is also supported by Korte et al (1992) who report that in many countries fees are
introduced at the rural and district level but are not introduced in urban areas or at hospitals.
Creese & Kutzin (1994) indicates that Namibia, Zambia, Zimbabwe and Kenya have all
introduced ‘cascading charges’ in public health facilities but as yet there has been evaluation
of this on the referral system. Mills (1994) confers with the ‘lack of information’ hypothesis
on this impact, but also suggests an additional problem that ‘...fees at the hospital level may
not be high enough to discourage excessive use; or even, as in some Bamako Initiative
countries, fees may be charged at the primary level but not further up the system. In addition,
there is little evidence on whether fees are being structured so as to encourage preventative
care, which would conform with the strategy presented in Table 12.
(ii) improve the quality of services offered. Association of quality improvements with the
levying of fees is problematic, since it is necessary to disentangle improvements related to
fee revenue and those from other sources (for example, from donor aid or increased funds
from government). An additional problem is devising a suitable measure of quality. Various
possibilities exist, for example measuring the: (a) numbers of trained personal in a facility;
(b) physical infrastructure; and (c) medical equipment in a particular facility; and (d) supply
of drugs.
Nolan and Turbat (1993) review studies that have attempted to assess the impact of quality
(using the first three variables) on demand. Whilst not strictly related to the use of fees to
improve quality, they identify the interesting result which is that there was no statistically
significant evidence to support the assumption that improved quality increased demand. The
explanation for this rests on the identification made in Section B which emphasised the
ignorance of consumers to the quahty of health treatments.16 On the other hand, Litvack and
Bodart (1992) on Cameroon, identified that when fees were used to improve drug availability,
16. Bennett (1991), suggests that politeness and smiles are more discernable to users than the
qualifications and professionalism of medical practitioners.
Page 38
this had positive effects upon the demand for health services. This is also supported by
evidence from the BI countries, where funds raised from fees have improved drug availability
(Mills 1994).
That drug availability influences consumer demand more than other areas of service
improvement (i.e. perceptions of quality) is confirmed by research in Tanzania. In a survey
across hospitals, health centres and lower level dispensaries, users placed a much higher value
on drug availability than the speed of service or the physical environment of the health
facility (Abel-Smith & Rawal 1992).
(Hi) improve equity. Fees are deemed to foster greater equity if resources raised from those
able to pay can be used to subsidise those unable to pay. However, because fees translate as
an introduction of a price for services which were previously free at the point of delivery,
there is major concern that the price will discourage consumption by those unable to pay the
fee.17
A number of studies have analysed demand following the introduction of fees. Waddington
& Enyimayew (1990) reported substantial decline in Ghana in utilisation at rural health clinics
following introduction of fees. Even after three years the dechne in utilization had not been
reversed, which suggests a strong price effect. The same result was found in Swaziland and
research in The Gambia highlighted substantial falls in outpatient attendance, even after two
years following the introduction of charges (Mcpake 1993).
In a review of the literature Nolan & Turbat (1993) confirm these results. Evidence from
Ghana, Swaziland, Zaire, Lesotho, Niger and Kenya all indicate that utilisation fell as a result
of the introduction of charges. Importantly, the research in Ghana, Kenya, Zaire and
Swaziland, which also considered the utilization of the poor, indicated that their consumption
had fallen back much more than that of affluent consumers. Gertler and van der Gaag (1990)
found that in Cote d’Ivoire demand was responsive to the introduction of fees but that the
poor were more likely to reduce demand following the introduction of fees than the rich.
In spite of these findings there is still significant debate within the literature upon the impact
of user fees upon the demand for health. In particular, Shaw & Griffin (1994) argue that
because the above studies fail to control for quality, falls in utilisation can not be attributed
simply to the introduction of fees. After reviewing the studies above, they conclude that falls
in utilisation resulted from fees being imposed for services which were then subsequently not
improved by the fee revenue. In the Ghana study fees were transferred to the central treasury
(Ibid 1994). This argument they claim, implies that it is not the fees which reduce utilisation,
but the non-use of fees to improve the quahty of services that causes the fall in utilisation.
In support of this argument, the study of Litvack & Bodart (1992) on Cameroon is sited. This
study argues that the link between utilisation and fees is related to the use of the fees and
conclude that if fees are used to improve quality, then utilisation increases.
17. Free at the point of delivery does not actually mean no cost is incurred by users. Travel
cost and time spent waiting for service both have economic costs which must be bom by rich
and poor alike.
Page 39
Others emphasise that the strongest determinant of the negative impacts on the poor derives
from the identification that in many cases, exemption mechanisms are either ineffective or not
in place (Nolan and Turbat 1993). This is an important finding because unless exemptions for
the poor exist, it is likely that either poorer households will stop using services or switch to
alternatives such as traditional, self medication or no medication. Indeed, the World Bank
consider that it is only in the context of certain public subsidies and exemptions that the poor
will benefit most from reforms (World Bank 1993a). Another concern for the poor is that
when fees are collected and retained at facilities, that facilities in poorer areas are likely to
generate less income than those in richer areas. This can have district and even regional
ramifications (see Section E on Tanzania prior to 1967).
In summary, where fees are used to improve drug availability, demand for health services
increase. However, the weight of the evidence suggests that the vast majority of countries
have introduced fees before introducing working exemption systems, so demand
improvements are associated with increased demand of the rich rather than the poor, this is
born out by the evidence on utilisation which clearly indicates that the poor utilisation has
fallen for the countries reviewed in this paper.
Decentralisation
Another element of the strategy widely implemented has been the decentralisation of the
public health system. Examples are: the main hospitals in Zambia becoming parastatals
(World Bank 1987); management boards established for hospitals in Kenya and Uganda (Mills
1994); cost control systems in hospitals in Malawi (Ibid); and decentralisation of functions
to the regional and district level in Tanzania and Ethiopia (World Bank 1993a).
Assessment of the impact of decentralisation measures are severely hampered by a lack of
systematic evaluation reported in the literature. What evidence is presented tends to be
anecdotal rather than quantitative (Mills 1994). However, experience to date suggests that
whilst the objectives are generally well specified, the conditions and means necessary for
success are not sufficiently contemplated. Decentralisation requires (often) drastic changes in
the responsibilities and functions of public servants. These functions ranging from
management, accounting, organisation, integration, operational logistics and planning require
specific skills and capacity which are often assumed rather than assessed.
McPake et al (1993) identify that a major problem in many BI countries has been the lack
of definition of roles and responsibilities for functionaries associated with managing
community financing. As a result of this, serious mistakes have been made in the purchasing
of drugs, control of funds and record keeping in Uganda, Kenya and Nigeria (Ibid 1992).
Parker & Knippenberg (1991) likewise point to problems in the success of a sample of 13 BI
countries in achieving their main aims, but conclude that the most successful element in their
survey was that local management had improved vastly since the introduction of local
autonomy. Other non BI decentralisation experience has also been encouraging:
"In Ghana and Botswana, governments transferred management and
decision-making from the central Ministry of Health level to district
and regional offices, which when combined with additional training of
local level public health workers and administrators, resulted in
Page 40
improvements in management and technical efficiency" (Our emphasis,
Sahn & Bernier 1993).
It has been noted however, that the problems of decentralisation can also run the other way.
In Ethiopia, one report stressed that district level decentralisation was severely impaired by
a lack of technical and management skills necessary at the centre to administer a decentralised
system (World Bank 1987).
Cassell (1993) having reviewed evidence on decentralisation questions the efficacy of
transfers within the government structure. He argues that the limited success in performance
is indicative of the decentralisation of functions and powers which are unlikely to significantly
impact upon the health system. A further ramification is the choice of institutions and
individuals which are subject to decentralisation. Are the pre-reform hospital administrators
likely to be willing and capable of implementing radical reform? Cassell implicitly thinks not,
for he suggests that shifting control from central to local government is not likely to illicite
as significant improvements as shifts to autonomous bodies within the public sector or
agencies in the private sector (Ibid 1994). The World Bank confers with this but adds that
financial autonomy at local levels is likely to improve the prospects for performance (1993a).
Management and Ministry of Health reforms
There is very little documented evidence on the introduction of the managerial and
organisational reforms outlined in Table 12. As a result it problematic to assess the impact
of such reforms within this strategy. However, we would emphasise that the efficacy of such
reforms will be strongly related to the extent to which capacity exists within the Ministry of
Health. This capacity is necessary for the design, execution and monitoring of such reforms.
Of paramount importance to exercising these functions will be information on and an ability
to use cost and performance data. Many authors are concerned that such skills are woefully
lacking at present and moreover require considerable resources to develop and bring on
stream (Bennet 1994, Mills 1994).
Government failure
This explanation emphasises the importance of strategic decisions made by the government
concerning health intervention and policy. Particular attention is given to the range of services
provided, how they are financed, who receives the benefit and policy concerning non
government provision (Shaw & Griffin 1994). The problems identified with government
failure are: (i) overextension and dominance of the state in the health sector due to
inappropriate policy towards the non-govemment sector; (ii) inefficient practice of using the
health sector as a means to transfer resources to tackle poverty concerns; and (iii) unrealistic
expectations of the efficacy of intervention policies, modalities and systems of management.
The elements of the government failure argument are now legend. They are based on the
presence of both inefficiencies and inequities and are summarised in Jimenez (1987). They
include the failure in many Type A (and Type B) countries to achieve or maintain a high
standard of care; over-consumption by the privileged (and/or rich) and under-consumption by
the poor, government failures in resource allocation (in particular, an over-concentration of
resources in therapeutic services and urban institutions despite the rhetoric to the contrary),
poor management of health institutions in the public sector, rising costs in excess of benefits
Page 41
delivered and a disenchantment by society with declining standards and increasing difficulty
in access.
The major elements of the strategy to combat government failure are presented in Table 13.
They are based upon a radical reappraisal of government policy that involve structural
changes aimed at a shifting towards private financing and from government to market-based
resource allocation, involving: (i) the mobilization of private finance to augment public
finance; (ii) the liberalization of markets in health care, pharmaceuticals or other health care
inputs; (iii) the privatization of public suppliers where profitability can be achieved; (iv) the
devolution of responsibility for the provision of private good-type services from public to
private sectors; and (v) the development of improved capacities in the civil service for
regulating private markets and managing publicly-financed contracts with private and
Assessment of Strategy
Whilst the strategy in Table 13 has not been fully embraced by the majority of SSA countries,
important components of the strategy have emerged in on-going reforms implemented across
the continent.
For example, private practice, once illegal has recently been legalised in Tanzania and
Mozambique. In Zambia, Malawi and Uganda, direct subsidies to non-govemment operators
have been extended to cover public health training and treatment of public patients (Sahn &
Bernier 1993). Tax breaks within a package of incentives have been introduced to foster NGO
provision in Tanzania (Kanji et al 1992), Madagascar (Sahn & Bernier 1993) and Zimbabwe
and Nigeria (World Bank 1993a). In Zimbabwe, the government has reduced subsidies by
cutting deductions from income tax for premiums paid to private health insurance (Birdsall
& Hecht 1994). Maintenance and laundry have been contracted out in the main Zimbabwean
hospitals and laboratory services have been contracted out in Nigeria (Mills 1994). In Ghana,
the supply of dings to public facilities was put out to competitive tender (Sahn & Bernier
1993). Finally, between 1987-1994, the number of countries reported as having insurance
schemes increased by 10 per cent (Shaw & Griffin 1994).
Whilst these developments have been reported in the literature, very few have been
systematically evaluated. However, anecdotal evidence is available which is reviewed below,
voluntary providers.
It is reported that liberalisation of drug procurement has yielded significant savings in Ghana
and in Guinea ‘...which in 1985, essentially broke the state importing and distribution
monopolies and allowed eight private drug wholesalers to enter the market. They in turn
supply an even greater number of private retailers. Preliminary indications are that both the
availability of drugs has improved, and prices moderated." (Sahn & Bernier 1993). The
legalisation of private practice in Mozambique is reported to have re-invigorated traditional
medicine and its associated practice of allowing payment in kind has increased the demand
for health services (Southern African Economist 1994). Finally, two new private hospitals are
under construction in Harare reflecting both the confidence of the private sector and
encouragement by the government of Zimbabwe.
Page 42
TABLE 13:Government Failure
COMPONENT OF STRATEGY
INSTRUMENT/ METHOD
TECHNICAL OBJECTIVE
Reduced Role of The State
- liberalisation of markets in health care and
pharmaceuticals
- legalise private practice
- encourage charity
- divest provision (sale, lease or contract).
- allow market to determine use of scarce
resources to foster improved allocative efficiency
Inside Government:
- supply contracts
- user contracts
- performance contracts
- improve technical efficiency
Introduce Competition
- improve technical efficiency, by linking cost
adherence, driven by potential profitability of
provision
- improve responsiveness to consumers
- increase consumer choice
Outside Government:
- legalise private practice
Extend Financing Options
- levy fees for public services
- introduce social insurance
- legalise/encourage private insurance
- encourage community financing
- create conditions for resolution of traditional
market failure and efficiency gains through
leverage of competition either from state to
market (Malaysia) or market to state (UK).
- to provide additional resources for the health
sector and shift burden of financing from state to
user.
- reoriented state towards, regulation and residual
financing role
- create constituency amongst users for higher
quality
- invest responsibility for finance for health
consumption in users which promotes a shift in
the discounting of benefits of preventative
measures and use of lower cost facilities.
Page 43
Not all the experience has been positive. In Tanzania, an analysis of non-profit organisations
revealed that nearly 50 per cent were in fact profit making organisations but were operating
under charitable status to avoid income tax and duties on imports (Kanji et al 1992). In
Lesotho, the lack of suitable private suppliers resulted in a catering contract being awarded
to the private sector, but at a price not considered to have been competitive (Mills 1994). In
Zimbabwe it is reported that in the private sector, providers relied more heavily on untrained
staff and exercised less supervision than in the public sector (Bennett 1994). Sahn & Bernier
(1993) report that liberalisation of health markets has resulted in a flow of doctors to the
private sector. In Zimbabwe it is reported that 40 per cent of the countries doctors work full
time in the private for profit sector covering 10 per cent of the population which has resulted
in shortages in government service (Bennett 1994).
Where elements of the strategy have appeared, they occur by virtue of either an explicit or
implicit belief that non-government financing or non-govemment provision are more efficient
vehicles to achieving improvements in resource use and ultimately in health status. The above
has indicated that it is perhaps too early to judge on whether non-govemment activity will
yield efficiency gains, but the lack of clear evidence has raised some doubts in the minds of
some authors as to whether the initial inefficiencies and ills claimed by proponents of radical
reform were well founded (Bennett 1994).
Two aspects of the assumptions made by radical reformers concerning Type A and B systems
are: (i) the inequity of resource consumption across income groups and (ii) the inefficiency
of public provision.
(i) Inequity in health systems
A charge of radical reformer’s is that the poor are subsidising the consumption of the rich in
public facilities (Lal 1994). Evidence in support of such a claim in provided by Shaw and
Griffin (1994). The authors present data on the distribution of government health by income
quintile for Tanzania and Nigeria. The data shows that a much higher proportion of such
expenditure goes to more wealthy households than to those identified as poor. The method
employed by the authors is measure the resource cost of benefits conferred upon health
facility users compared to the tax or insurance system operating in the country. Having
concluded that the tax system in both countries is ‘regressive’, the authors argue that as the
majority of health facility users are from upper income quintiles that the poor are subsidising
the rich.
Whilst these conclusions sound convincing there have been very few studies which test these
conclusions. Hammer, Nabi & Cercone (forthcoming) report the results of an expenditure
incidence study in Malaysia. They found that the public subsidy share was equally distributed
across income groups and if anything became more progressive with time as richer people
sought more private care. The richest 20 per cent paid more than twice in taxes what they
received in subsidies; while the fourth quintile paid only 20 per cent of what they received
in benefits.
Similar results are found for Cote d’Ivoire, where a descriptive analysis of the delivery and
financing was conducted (Gertler & van der Gaag 1990). The authors conclude that, whilst
Page 44
the majority of health facilities are located in urban areas, the range of taxes (and their
structure therein) used to finance delivery are levied from a ‘mildly progressive’ system, with
the conclusion that contributions and benefits across the system are likely to be equitable.
The only analytical study we found for Africa is Bloom & Segall (1992) on the Kenya health
sector for the World Bank. The upper 12 per cent of income earners paid Ksh 653 per caput
per annum in out of pocket expenses and received Ksh 893 per caput per annum of health
care; a publicly-financed subsidy of Ksh-240 or 27 per cent annually. Their health service
consumption was 15 times that of the rural population and they consumed 50 per cent of
national health resources. The rural poor, on the other hand, paid Ksh 7 per caput annually
and received Ksh 58 each of care; a publicly-financed subsidy of Ksh 51 or 88 per cent. The
contribution of each to taxes was not recorded. These results pose a dilemma which can only
be resolved by taking a particular view of equity. The distribution of publicly-financed health
care is clearly inequitable if equity means an even distribution of health subsidy across
income groups. However, it is not clear from these results that in Kenya that the poor are
subsidising the health consumption of the rich as claimed by proponents of radical reform.
A final question on equity is an analysis of the development of insurance schemes across
SSA. Shaw and Griffin (1994) report that by 1994, 15 countries were covered by such
schemes. This it is argued releases public funds which would otherwise be spent on curative
care for those insured (Ibid). Vogel (1990) has argued that this conclusion can not be assumed
and needs to be researched. He states that public insurance schemes only free up resources
if they are self-financing (i.e resources raised from insurance premiums) and pay the full
economic cost for health services provided by government facilities. The latter consideration
is also relevant for private schemes, since any charge below economic cost would represent
an implicit subsidy to the private scheme. Examination of the 15 countries used by Shaw &
Griffin (1994), reveals that over 50 per cent are state insurance schemes. However, without
supporting information on the concerns raised by Vogel (1993), we can not be sure that public
resources are freed up and used to improve equity.
The above clearly indicates that further detailed research is necessary before ‘assumptions’
about the inequity of publicly financed health systems can be made and the charge of
government failure unequivocally substantiated.
(ii) Extent of government failure: measured inefficiencies
Another criticism levelled at government provision is that it is subject to productive and
technical inefficiencies. Mills (1994) argues that by comparison with non-govemment
provision, unit costs are higher and in particular productivity is much lower in the public
sector. Yet while efficiency criticisms may be justified, a few years ago the success of China,
Cuba, the Soviet Union and Tanzania were generally applauded. These countries achieved life
expectancy and infant mortality rate improvements in excess of what would be expected for
their per capita incomes. Moreover, although public sector inefficiencies may be too selfevident to be questioned, it is notable how few research results are available which document
Page 45
these alleged inefficiencies or, more importantly, which demonstrate the superior efficiencies
of comparable institutions in the private sector.18
For example, Barnum & Kutzin (1993) compare institutional efficiencies across private,
charity and public sector institutions and do not arrive at a conclusion regarding the overall
superiority of one over the other. Bennett (1994), in a review of studies in Malawi (1990),
Lesotho (1989) and Tanzania (1992) finds only the latter country ehibiting inefficiency in..
government provision. Indeed the World Bank found that in a cross section of non
government providers in Zaire, Togo, Ghana and Zambia that:
"Problems [occurred] over different concepts of service provision, often
resulting in patchy coverage and weak support services; poor
accountability to local communities; failure to mobilize communities
for health promotion; and weakness in management systems" (World
Bank 1993a).
These reviews indicate the importance of empirical analysis of provision. They highlight the
dangers of assumptions about the inefficiency of the public sector and of the superiority of
non-govemment provision.
18.
The general lack of data on which to compare public, non-governmental
organisations (NGO)s and private institutional efficiencies or to establish efficiency
bench-marks is a major research gap to which we return later in this paper.
Page 46
E.
A Case Study of Tanzania
Pre Independence19
In 1877 the first European missions arrived in Tanganyika20. In 1891, a medical Department
was established to supervise the health of the German colonial government. At the turn of the
century, the department engaged in campaigns against the communicable diseases of the
plague and sleeping sickness in areas of European occupation (Titmuss et al 1964).
Following World War I, Tanganyika passed over to Great Britain. Having established a public
health department, the new colonial power sought to extend its activities with respect to the
control of communicable diseases and by 1920/21, 23 per cent of the medical budget was
spent on such purposes {Ibid 1964). A ‘mosquito brigade’ was established whose purpose was
to eradicate breeding grounds for malarial mosquitos near urban and administrative areas.
Other campaigns had a national focus. Over one ninth of the population had been treated for
yaws and 100,000 for syphilis. Both campaigns:
" demonstrated vividly to the rural population the effectiveness of
western medicine. This was to have important and lasting consequences
for the future of health services, particularly in creating demand for
curative treatment" (our emphasis, Ibid (1964).
However, by 1930, world depression caused retrenchment in the government budget and the
real levels of expenditure of that year were not reattained until 1943. Titmuss et al (1964)
note that preventative services were the main victim of cuts and indeed that curative services
managed to expand. This occurred in spite of medical knowledge which suggested serious
consequences. In 1938, a ‘Memorandum of Medical Policy’ published in Dar es Salaam
noted:
"The preventative outlook must at all costs be inculcated into the
dispenser [of medicine] and the district sanitary inspector...[they]
should be encouraged to work together to try to keep their people well,
instead of using all the resources of the dispensary to cure the sick”
{Ibid 1964).
Following World War II, Tanganyika became a UN protectorate. Britain then initiated a
policy of self financing for the territory which effectively meant that capital expenditure could
only be financed from current revenue reserves (Leubuscher 1946). However, support for
limited capital expenditure was granted through the Colonial Development and Welfare Act
which typically funded about 35 per cent of capital expenditure during the 1950s (IBRD
1961).
19. The seminal work on this era is Titmuss et al (1964), which is used extensively.
20. Tanganyika was the mainland area of what is now Tanzania.
Page 47
The limit imposed on expansion by the self financing edict was somewhat obviated by the
increasing importance of mission hospitals. Having initially funded their own activities, they
now received increasing allocations from the government to expand and serve rural areas. A
system of grants was introduced to administer ‘...functions which the government accepted
as its own responsibility’, which essentially translated as preventative services and hospital
services for rural areas not covered by government hospitals (Titmuss et al 1964). Whilst
missions still maintained charges for patients, this ‘subsidy’ greatly reduced the costs to users.
In government hospitals, there were no charges for outpatient services, but drugs had to be
purchased. Charges were levied on inpatients, but these were restricted to consumers of
private wards. However as private wards were designed for non-Africans (see below), it is
reasonable to assume that the majority of Africans were not charged in government hospitals.
Fees for drugs and private wards amounted to 9 per cent of recurrent expenditure in 1958
(IBRD 1961). Due to the self-financing requirement the remaining costs of provision were
raised from local taxation sources.
Government accounts reproduced in IBRD (1961), indicate that the major sources of revenue
for 1958 were import duties (28 per cent), income tax (22 per cent) and excise duties (11 per
cent). Ehrlich (1966) notes that for 1958, whilst 8,500,000 Africans contributed half of the
income tax revenue and 124,000 non-Africans contributed the remainder, half of the revenue
raised from import duties were on consumer necessities (such as cloth) which were consumed
by the vast majority of Africans.
As highlighted in the Section D, interpreting tax incidence from this aggregate data requires
caution and is not attempted here. Furthermore no studies exist which comment on the
distribution of the government budget across users during this period or later. However, what
does emerge is a historical antecedent to phenomena witnessed in the 1970’s which is that
preventative care is identified as the first line for economy. Moreover, this occurred despite
knowledge of its detrimental effects. Finally it should be noted that charges (although not full
cost) were levied on those (albeit racially) identified as ‘able to pay’.
The above indicates a system, with subsidies to non-govemment providers and no direct
charge at the point of service for the vast majority of Africans. However, coverage was by
no means universal with rural coverage extremely limited outside areas of mission activity.
It is worth noting that the health system in colonial Tanzania quite closely approximates a
Type C system, rather than the Type D system which we identified with colonialism in
Section C.21
Independence to Arusha Declaration: 1961-1967
21. It has been noted that due to the protectorate status of Tanganyika, the British government
could not deprive Africans of land and forcibly settle an expatriate community. Furthermore,
protectorate status required that the British government engaged in activities designed to
improve the economic status of Africans (Hatch 1972). By contrast refer to the experience of
Kenya cited in Section C.
Page 48
Tanzania gained its independence from Britain in 1961. The country inherited a health system
which had been focused towards urban areas, was heavily staffed by expatriate practitioners
and had limited rural coverage. Kahama et al (1986) note that out of four hundred and three
doctors only twelve were African. Rural coverage was essentially limited to ‘mission’
hospitals, who self-selected locations and were outside any overall framework/national policy
(Mukyanuzi 1994).
Following Independence, the newly invested government introduced the ‘First Five Year Plan’
(1964-69). Within the framework of the market-based economy inherited at Independence, the
plan sought to increase public investment in infrastructure and to extend development into
rural areas. For the health sector, the plan emphasised the need to increase medical training;
intensify curative services; and improve the coverage in rural areas at regional and district
level (Stephens 1968).
Economic growth averaged 6 per cent per annum between 1961-7, which with a population
growth of just below 3 per cent resulted in per capita growth of national income and provided
the basis for development(Netherlands Development Cooperation 1994).22 Resources for rural
infrastructure were raised at the regional level from export and sales taxes. Given quite wide
disparity in the resources positions between regions, this had the effect that in some regions,
infrastructure developed more quickly and was more likely to be maintained than in others
(Amani et al 1987). Little cross-subsidisation was attempted by the government. This outcome
had an important impact upon subsequent policy (see below), but, along with a budgetary
approach of ‘spending based upon resource availability’ enabled the government to record a
budget surplus for this period (World Bank 1989b).
The main thrust of policy was for the state to enfranchise rural areas in the newly liberated
country by improving the availability of health facilities. But this policy was to be achieved
through a transfer of what were previously ‘colonial revenues’ to the regions of origin and
not to the central government for subsequent distribution. Furthermore, the activities of
Missions were extended and encouraged, with the government content to continue to subsidise
non-government provision. Whilst in 1962, private practice for government doctors was made
illegal, occupational and private practice (although very limited) was able to continue offering
services.
Available data on impacts on health status indicate that between 1961 and 1967, the Infant
Mortality Rate (IMR) declined from 215 to 135 (Mukyanuzi 1994) and Life Expectancy (LE)
increased from just below 40 to 42 years for the same period (Kahama et al 1986).
Furthermore, during this period the number of government health centres increased from 22
to 40, whilst lower level dispensaries increased from 746 to 996 (van Ettan 1976). However
it has been suggested that towards the end of this period, plans for country wide coverage had
not been achieved and that resources were increasingly directed to urban facilities and the
training of doctors rather than lower-level cadres suitable for providing rural services (Ibid
1976).
22. See Kahama et al (1986) on export production and export revenues.
Page 49
This system most strongly resembles a Type B system. Free care for the population, but
coexistence with a significant non-govemment presence in the voluntary sector. Whilst there
is no data to substantiate any competition between the two providers, from the point of the
consumer, the possibilities for switching consumption was present, which by default
represents a potential spur to efficiency highlighted in Section C.
The Arusha Declaration 1967
In what has been described as the move to nationhood (Kahama et al 1986), the character and
scope of government intervention in the economy drastically altered following the ‘Arusha
Declaration’ in 1967. The emphasis of the declaration was upon altering the system of
economic relations through a radical shift in the balance of power and influence away from
the individual to the state.
The government asserted the right of individuals to health services; it took up the
responsibility to finance and provide those services; and it embarked upon a programme of
building hospitals, health centres and dispensaries to national target levels. Access to health
was taken out of the market and consumption was explicitly linked to ‘need’ rather than
income or any other factors.
The government introduced the Second Five Year Plan (1969-74) which sought to consolidate
the earlier policy of improving the numbers of medical personnel and facilities, but which
now emphasised the importance of the provision of preventative services, efforts to curb the
spread of communicable diseases, and to further extend rural provision (Mukyanuzi 1994).
In an attempt to ensure regional equality, district councils were abolished in 1972, paving the
way for the national pooling of resources and for the development of a national regional
policy. Also the Ministry of Health bolstered the Maternal and Child Care Health programme
(MCH) in 1974, which sought to provide immunization, nutrition education, postnatal and
antenatal care across the country (Mukyanuzi 1994).
By 1972, the effects of the reorientation were evident in both the flow of public health
finance to rural areas and in the numbers of health centres and dispensaries constructed. These
had increased from 40 to 90 and from 996 to 1400 respectively (Kahama et al 1986). To
bolster the efforts of the government, certain mission hospitals were ‘designated’ as
government hospitals.23 As a result two consultant mission hospitals were designated as
national referral hospitals and seventeen as ‘government district hospitals’.
Furthermore the import, control and distribution of drugs was coordinated through the state,
with the national pharmaceutical company having a monopoly on domestic production and
the Central Medical Stores being the arm of government responsible for importing and
distribution. Payment for drugs was abolished in 1967 and given the continuance of no fees
at the point of service, resources for the health sector were raised through a combination of
central and local taxes and from donor contributions.
23. This effectively annexed the said hospitals into the government tiers and disallowed the
practice of charging.
Page 50
This period represented a significant transition in government policy which explicitly
established the right to health care which was independent of income and which the state took
on the responsibility to finance and deliver. The designation of mission hospitals as
government hospitals did not abolish the ‘market’ in health, but effectively sounded its
inappropriateness in the vision of the Arusha declaration. All in all, it is evident that Tanzania
moved very close to the Type A model.
Economic decline (1974-1985)
Type A status was enforced in 1977, when private for profit practice was criminalised. This
effectively left only those mission hospitals not ‘designated’ as the alternative source of
modern provision to the state.
However, despite an upturn in commodity (chiefly coffee) prices in 1976, the period from
1974 to 1985 was characterised by falling economic growth. Average growth rates were 2.1
per cent from 1974-78 and 1.5 per cent from 1979-85 (Netherlands Development Cooperation
1994). Throughout this period population grew steadily at around 3 per cent, which indicated
on average, GDP per capital fell for the period from 1974-1986 (Mukyanuzi 1994).
Many factors help to explain this downturn in economic growth, amongst them: the oil shocks
of the 1970’s; severe droughts in 1973; the breakup of the East African Community; the war
with Uganda from 1978-80 and the general fall in commodity prices. However, because the
major elements of the tax base in Tanzania were export and sales based taxes, economic
decline impacted on the government budget, which from 1974/75 moved into continual deficit
and was increasingly financed by both domestic and foreign borrowing (World Bank 1989b).
The net effect on the health sector was a dramatic fall in real terms of government
expenditure on health. One source estimates that the 1982/83 expenditure in real terms was
equivalent to only 57 per cent of the allocation in 1977/78 (Mukyanuzi 1994).
This reduced flow of government resources to the health sector created major constraints on
the financing of current activities and of fulfilling planned commitments. The manifestations
of which, became very visible: maintenance and development budgets squeezed and
complimentary inputs (such as drugs) diminished (Abel-Smith & Rawal 1992); rural areas lost
out to urban areas (Mukyanuzi 1994); available resources became increasingly concentrated
in salaries (Bloom et al 1992); there was a concomitant increase in donor activity; and finally
economic dechne diverted resources from government health care to the productive sectors
(Mukyanuzi 1994).
The reduction in resources for the health sector, coupled with the commitment to financing
and provision inherent in the Arusha Declaration limited the extent to which alternative
sources of finance could be tapped. The net effect of which has been widespread
disenchantment with government service and what seems to be a drift towards voluntary
agency (VA) hospitals for those with available income (Abel-Smith & Rawal 1992),24 and
24. Gilson (1993) confers that the availability of drugs is higher in Mission hospitals.
Page 51
a drift towards traditional and or no consumption for those unable to pay (Heggenhougen
1986). Furthermore, the practice of ‘unofficial payment’ at government facilities to both
receive consultations and treatment which became widespread from the late 1970’s onwards
(Doriye 1992).
In summary, the state having committed free access to basic health services for all
Tanzanians, was faced with a series of factors which effectively undermined its ability to
finance the promised provision.
Assessment of the Tanzanian health system to 1990
Gains
The focus on rural areas associated with the Arusha Declaration produced significant gains
in coverage in rural areas. This has resulted in Tanzania being regarded as a principle
architect of the Primary Health Care concept (Gilson 1993), but has also meant that by the
late 1980s, 90 per cent of population lived within 10km of a health facility and 75 per cent
within 5km (Abel-Smith & Rawal 1992). Furthermore, through the network of rural facilities,
extensive preventative services were provided by the state, free at point of delivery (see
Gilson (1993) on immunisation coverage). Recent data from the World Bank indicates that
immunisation coverage in Tanzania is above the SSA average.
The impact of government health policy on health status has also been impressive. Key
indicators such as Infant Mortality Rates, Life Expectancy and Maternal Death Rates are
above the average for SSA and well above the levels expected for income (see Human
Development Index, UNDP 1990). Finally, although concerns for the distribution of
government resources abound (see Shaw & Griffin 1994), tentative evidence presented in a
recent review of the health sector by the World Bank suggests that the overall distribution of
benefits arising from government expenditure (own resources and donor) was fairly evenly
distributed across income groups (World Bank 1994).
Costs
Despite the expansion and investment in the health infrastructure, the constraint of providing
adequate recurrent costs for provision of services began to arise. As economic growth slowed
down, the commitment of the Arusha Declaration, restricted the options for raising finance
from other sources. Historical data on donor finance is extremely scanty, whilst more recent
data is far more complete, the general consensus is that donor contributions have been rising
over time. By 1993/4 donor-finance to the health sector, supported over 80 per cent of
government development expenditure and 81 per cent of all recurrent and development
finance to preventative services (Svantessen 1994).
Donor finance has been directed at important interventions in Tanzania. For example, the
MCH programme was heavily supported by the US Government in its inception (van Ettan
1976). However many are concerned that donor aid also introduces a loss of sovereignty and
questions of sustainability (Lawson 1994). Indeed, donor inputs can present serious problems.
Many authors indicate the lack of choice on the part of recipient governments to ‘aid projects’
(World Bank 1993a); the waste of resources inherent in separate national programmes; the
perverse effect that aid has upon incentives for the state to provide its own finance for the
types of projects funded by donors. Furthermore, because of the interest of donors in rural and
Page 52
preventative programmes, the government is by default left with financing curative-type
services. Indeed the Ministry of Health estimates that for 1993/4 88 per cent of government
recurrent expenditure will go to curative care (Ministry of Health 1993).
Despite the success in extending the health infrastructure across the country, two studies
indicate that quality of service is far from sufficient. Kanji et al (1992) and Gilson (1993),
both pinpoint that ‘process quality’ and ‘structural quality’ are poor due to a lack
complimentary inputs and little maintenance. These factors are also seen to have impacted
upon institutional efficiency, with drug supply problems and lack of equipment impinging
upon the output of medical practitioners. A further dimension to this problem has been the
centralisation of power that resulted from the abolishing district councils in 1972. Whilst the
intension was to foster a national policy, many authors conclude that not only was control
taken from communities (UNICEF 1993), but also that ‘self reliance’, which had existed at
the regional and district level was severely tempered by both confusion in responsibility for
services within the ministries responsible for health care and by poor relations between local
health workers and villagers (Ministry of Health 1993).
A further problem identified in the studies on quality and in other research is that although
the relative share of wages in the government health budget has increased, wage-levels have
declined dramatically in real terms (Doriye 1992). This is seen to both impact on staff morale
and has encouraged absence from work (ibid).
There is general agreement that the Government of Tanzania (GOT) achieved considerable
success in improving the health status of its population. However, falling quality of services
coupled with the emergence of ‘unofficial payment’ created general dissatisfaction with state
provision both on the part of users and the Ministry of Health. It remains difficult to untangle
the specific import of the factors contributing to these developments. The combination of
failing capacity; austerity (planned and not planned); low moral due to poor salaries; and a
failure to adapt policy to the new resource position all impacted upon the health sector.
It should be emphasised that by default ‘unofficial payment’ determined access to government
health services. Moreover, as no management or organisational system was in place to use
funds raised from users, the funds could not be used to improve services or to subsidise the
poor. As a consequence, it has been argued that the Tanzanian health system effectively failed
to serve those unable to pay (Doriye 1992).
Could this situation have been avoided? Did the government over-extend itself? Where its
targets reasonable? There are two themes which arise in the literature. Some authors suggest
that the route of the problem lay in inappropriate budgetary policy which failed to take
account of the full recurrent costs of the expansionary policies in health (Lawson 1994). For
others, it is the decline in resources which have resulted in failing performance, rather than
a failure of planning (Gilson 1993). The former indicates some form of ‘government failure’,
whereas the latter suggests crisis due to unforseen factors.
Whilst many have criticised the government for failure to respond to the changes in resource
availability due to economic decline (see for example World Bank 1989b), the commitments
inherent in the Arusha Declaration firmly wedded the government to a path of development
which offered little scope for manoeuvre. However, even in spite of the rural and preventative
Page 53
emphasis of the Second Five Year Plan (1969-74), available data indicates that the proportion
of the government budget devoted to curative hospital care increased. For example, in 1971/2
hospitals took 52 per cent of budget and in 1978/9 hospitals took 63 per cent of budget (van
Ettan 1976).25 The failure to reorientate expenditure from curative based services is not
solely due to government policy. Curative services are also generally more expensive to fund,
form the focus of careers for medical personnel, and are located in urban areas. The corollary
is that the curative sector is often backed by quite influential political/social forces which
constitutes a barrier to change.
As a result, by the 1980’s, the Tanzanian health system was characterised as follows: donors
were increasingly responsible for financing preventative services; users were effectively
paying for access to government services; and government resources were being concentrated
in curative services. The combination of these factors provoked the onset of reform.
The broadly descriptive analysis of the Tanzanian health system is a reflection of the dearth
of government and non-govemment financial data available for Tanzania. In Section D we
argued that this lack of data was a serious impediment to analysing the impact of shifts in
public finance. However, the following trends are suggested by the description offered above.
First, at independence, over 50 per cent of general hospitals in the country and 25 per cent
of lower level facilities (health centres and dispensaries) were run by missions which charged
users fees. This implies considerable private finance. However, the extent of financing is
uncertain since these facilities did not charge the full economic cost of services to users.
Grants were received from mission organisations in Europe and (increasingly) grants and
subsidies were given by the government. In the remaining government facilities, the vast
majority of services were free and following 1967, all services were free.26
Second, the ‘designation’ of 40 per cent of the mission hospitals as government hospitals,
effectively removed the responsibility for finance from the user to the state. If data existed
we would expect to witness an increased proportion of public finance in the total financing
of health care. As a result we would assume that this would mean that public finance
occupied a greater proportion within total financing than that which existed both prior to
independence and in the immediate period following independence.
Third, following economic decline, evidence suggests that informal payment determined
access to services at government facilities and was necessary to receive government drugs.
This suggests an increase in both the volume of private finance and a relative decrease in
public finance.
Fourth, whilst the literature suggests that even from colonial times aid has been an important
source for public expenditure (see Colonial and Welfare Act contributions above), as public
25. This data might hide ‘preventative services’ provided by hospitals.
26. We again reiterate that there is still an economic cost associated with consumption, that
of travel time and lost employment.
Page 54
finance has fallen back, we would expect to see the proportion of aid in total financing to
have increased.
These ‘suggested’ trends are bom out by data for the only available year. In 1990, the total
financing of health consumption comprised 51 per cent from aid sources; 37 per cent from
private sources; and 12 per cent from the public sector. The magnitude of aid and private
finance accord with data presented earlier for Type A systems. Moreover, the data indicates
the extent both of the decline in public expenditure (assumed by) and described in the text
above, and also why analysis of financing must always consider aggregate data to understand
dynamics in health status.
Recent changes introduced by the Government of Tanzania
In the context of the problems described above, the Ministry of Health has embarked upon
a reform process, which seeks to improve the finance and delivery of health services. This
reform was initially articulated in the National Health Policy document of 1990 and more
recently in the Health Strategy Note of 1993 (Ministry of Health 1993). In September 1994,
the final program for reform will be presented through the ‘Health Sector Reform Group’,
which has
been appraising reforms to date and will present recommendations for the future direction of
reform.
The major objectives of reform are to: reorientate activity towards ‘public health’ activities;
to improve planning, management and organisation; and to concentrate resources on core
programs (Ministry of Health 1993). This is to be achieved by both altering the terms of
finance and delivery of health services.
Measures already introduced
At the sectoral level, the government introduced the Rolling Plan and Forward Budget (RPFB)
from July 1993 which brought together planning and budgeting for government resources. The
major impact of the (RPFB) is that it will curtail the practice of supplementing ministry
budgets following the passing of budget in any given year and it will introduce a three year
time-frame for expenditure plans.27
Within the context of forward planning the government is targeted to eliminate donor
contributions to recurrent ‘health’ costs by the budget year 1994/5 through an inter-sectoral
reallocation of government funds to major public health interventions, which will involve the
reduction in the range of services financed by the state and retrenchment of personnel.28
At the institutional level, the government returned ownership of a number of the ‘designated’
voluntary agency (VA) hospitals back to their original owners. Most notable in this transfer
has been the two national consultant hospitals in Moshi and Mwanza. Early evidence of the
impact of the transfer is that capital budgets have increased markedly. For example in Moshi,
27. See Lawson (1994) for a discussion of budgetary practice pre 1993.
28. Discussion with Chief Medical Officer July 1994.
Page 55
the capital budget has increased by over 100 per cent in real terms. This has followed years
of no development expenditure under government control (Kilimanjaro Christian Medical
Centre 1994). Private practice was reintroduced in 1992, with the provision that public sector
doctors could now practice outside working hours in the private sector.
Through a two-stage process, the government has also abohshed it’s monopoly on the import,
sale and distribution of drugs. The first stage was the legalisation of private dispensaries. The
second stage was through the deregulation of the Cental Medical Stores (CMS). Having
previously enjoyed monopoly powers of import and distribution the CMS is now an
autonomous body within the public sector and must compete with private distributors, through
the progressive reduction of public subsidy.
At the microlevel, the government upgraded fees for inpatient wards.29 Existing fees for
grade I and grade II wards in government referral and regional hospitals were increased from
July 1993. This was followed by the introduction of user-fees for grade III wards from
January 1994 in regional and referral hospitals and in district hospitals from July 1994. With
this package of charges is an exemption-system for those deemed as unable to pay, currently
estimated at 60 per cent of the population (Abel-Smith & Rawal 1992). This is due to be
operational by the end of 1994.
Measures being considered
In September 1994, the Health Sector Reform Group (HSRG) will report on future reform.
Major Reforms currently under discussion will address both the finance and delivery of health
care.
In terms of the finance, the government is set to introduce and administer a Compulsory
National Insurance (CNI) scheme for all formal sector employees (private and public). This
scheme will embrace approximately 10 per cent of the population and is intended to free-up
public resources for the financing of services for the non-insured. Moreover, following
assimilation to charges through the introduction of nominal charges for inpatient care, the
government intends to increase these charges closer to the full economic cost of provision
(Ministry of Health 1993). The government also intends to lease government facilities to the
private sector (eg laboratory tests, X-Ray equipment), to raise resources for their maintenance
and operation in the public sector.30
In terms of delivery, the Ministry of Health is considering the rationalisation of responsibility
between government ministries for health provision. At present, this responsibility is divided
29. In Tanzania referral and regional hospitals have three types of ward: Grade I - (2/3 beds
per ward) - formerly the European Ward; Grade II - (5 beds per ward) - formerly the Asian
Ward; and Grade HI - (10+ beds per ward) - formerly the African Ward. District hospitals
generally have Grade II and III wards. Grade I and Grade II wards are considered as private
wards and have always carried nominal charges.
30. Discussion with Chief Medical Officer July 1994.
Page 56
between the Ministry of Health and the Prime Ministers Office.31 The Ministry of Health
is responsible for national health policy, the four national hospitals, vertical programs and in
administering grants to (VA) hospitals. The Prime Minister’s Office is responsible for
funding, interpreting and implementing national policy at the regional and district level
(Ministry of Health 1993). At the heart of such reforms will be an attempt to improve
coordination between each ministry which is generally recognised as poor (Ibid). Other
delivery reforms include: (i) decentralisation in the form of outright community ownership
of health centres and dispensaries; (ii) the encouragement of private investment in health
through concessions granted in the private investment legislation; and (iii) the contracting out
of public services to private sector practitioners.
The future health service in Tanzania
There is little doubt that the quality of public health services in Tanzania began to decline
following the mid 1970s. Be it that facilities lacked essential drugs (Abel-Smith & Rawal
1992); doctors failed to be at post (Doriye 1992); or hospital equipment along with other
capital items were not maintained (World Bank 1993a), evidence abounds of such declines
in the literature. With the description of decline is usually an explanation of decline which
identifies the importance of underfunding of drugs, maintenance budget or real wages. These
explanations are given credence by data on budget allocations to the health sector. This data
indicates that the government budget allocation (recurrent and development) has fallen from
around 9.5 per cent of the total budget in the early 1970s to approximately 5 per cent in 90/91
(Ministry of Health 1993). These figures are also confirmed by Bloom et al (1992).
In spite of the recognition that underfunding has occurred, there has been little attention at
attempts to refund the Ministry of Health budget through intersectoral reallocations. Instead
the quality problems associated with underfunding are to be combated with a package of
reforms which stress a reduced role for the state and the emergence of market forces in the
health sector. The implication of the reform package implemented by the Government of
Tanzania is that an adequately funded public health system in Tanzania would not be
efficient.
Within the adopted reform strategy, the public role will be to concentrate available resources
on priority interventions. These will include: financing preventative services (to include
amongst others: immunization, public health education and MCH activities); financing a core
programme of basic curative services for common ailments; the subsidisation of curative
services for the vulnerable, disadvantaged and poor; and the training of key cadres of medical
staff. The government will now encourage non-govemment provision of services which will
be regulated by the ministry of Health (Ministry of Health 1993). Finally the public sector
will foster improvements in technical efficiency by introducing management and
organisational reform of public services; by extending the interface between the public and
private provision of services by encouraging private (profit and non-profit) practice and by
contracting out to the private sector.
31. Within the Prime Minister’s Office responsibility is divided between the Regional
Administration and Local Government Department.
Page 57
At the level of the user, the individual responsibility for payment, be it through out-of-pocket
fees at facilities or through compulsory insurance if employed in the formal sector, represent
a considerable shift from state responsibility. Indeed, users having previously had little choice
in consumption, will now be able to purchase services from non-govemment facilities.
Moreover, the introduction of fees in government facilities is assumed to create a constituency for users to demand higher quality services in such facilities.
This reform package sees a leaner and more focused state sector. It also envisages both the
potential and/or actual existence of both non-state finance (private and donor/charitable) and
non-state provision. Within the reform package, the Tanzanian government is to maintain
provision of preventative services. However the reforms have a much greater significance for
curative services in that individual responsibility for finance becomes more important and
non-govemment provision is encouraged. A number of technical issues and potential hazards
arise from this observation, which are briefly reviewed below:
Impact of Charges
Considerable research has taken place and is underway on the impact of user fees in
Tanzania.32 At this stage it is still too early to analyse the impacts of charges. Neither has
the national exemption scheme across referral, regional and district hospitals will not be
operational, or the system for the use of funds raised from fees implemented.33
Notwithstanding this, it is estimated that charges on Grade HI wards (i.e. public wards) will
be waved for an estimated 60 per cent of users (Abel-Smith & Rawal 1992) and that funds
raised through fees will be retained at the facility which levies the charge. Whilst the impact
on utilisation is an empirical issue, it is worthwhile echoing the experience of many countries
in SSA. This experience indicated that utilisation fell after the introduction of charges (Nolan
& Turbat 1993). This occurred because: quality was not improved with fee revenue;
exemption systems failed to target the poor; and management costs were very high resulting
in a minimal net gain of resources.
This raises concern in Tanzania for the following reasons. First, charges are being introduced
without any increase in resources to hospitals. This implies that users will (at least initially)
be paying for the same services that existed prior to charges. Secondly, fee levels have been
set at nominal levels, (even for private wards), with no account of actual unit costs of
services. This latter point implies that the contribution of fees to recurrent cost will be
minimal. Moreover it implies that the ratio of management costs of the scheme to revenue
will be high. The Ministry of Health is currently working with an estimated management cost
of about 18 per cent of the gross yield from charges (Abel-Smith & Rawal 1992). Given
international experience we consider this estimate to be optimistic.
32. See Abel-Smith et al (1992) for background research to the recent introduction of fees.
The Overseas Development Administration funded this research and is currently monitoring
the impacts of fees on efficiency and equity as a second stage to the original project. The
ODA is also providing research funds to the Liverpool School of Tropical Medicine to
monitor the impact of fees and other health sector reforms in the Coast region of Tanzania.
33. Projected date for implementation of both is January 1995.
Page 58
Segmentation
A particular impact of the reforms introduced is that they have the potential to introduce a
degree of segmentation between regions, between rural and urban areas and between those
that will be insured and those that are not. A driving force for the Arusha declaration was the
desire to ensure that all regions had similar quality services. Moreover this policy wasinitiated in part as a recognition that regional inequality in Tanzania had become exacerbated
following the immediate period of independence (Amani et al 1987). With policy emphasising
retention at facility level and decentralisation to community levels, the spectre of regional
income disparities is likely to re-emerge. This concern is bom out by evidence from the VA
sector, with the observation that mission hospitals in relatively affluent areas (eg Kilimanjaro)
have been able to raise nearly 60 per cent of recurrent revenues from charges, whereas those
in poor areas (eg Rufiji) have raised much smaller contributions.34 A similar phenomenon
is likely to occur for those in rural and urban areas and those that are insured and not
insured. Without adequate targeting, it is likely that the reforms will foster the emergence of
significant disparities in access to health care.
Management of Resources
A feature of the pre-reform system, was the lack of information concerning financial flows
within the Ministry of Health. Information about resource use across regions, districts,
facilities, or use within facilities was not produced or analysed. The reforms, bring to the fore
the importance of using resources efficiently and to improve accountability for resource use.
However, with little experience of financial control and or attempts to prioritize services using
cost information, it is likely that reform dependent upon financial information are unlikely to
succeed unless capacity to produce and analyse such information is increased.
Emergence of Market Failure
In Section A we highlighted the debate concerning the extent of market failure in health
markets. Although this debate has not been resolved, even pro-market health economists
recognise the importance of regulation and services whose private benefits (costs) diverge
from their social benefits (costs). The reform package for Tanzania specifies a movement
towards market provision. Within this movement is a shift in the government role from
provider to all to more limited provision and regulation of what remains. Within such an
environment, it is incumbent upon the Ministry of Health to fully anticipate and embrace its
regulatory role if inefficiency associated with market failure is to be prevented.
Inefficiencies at the public-private interface
Reforms already introduced allow publicly trained medical staff to practice in the private
sector and future reform will provide for public purchased equipment to be ‘rented’ to the
private sector. Important questions about the subsidies inherent in these two activities are not
seriously addressed by the reforms. These include the pricing of public equipment and
services to the non-govemment sector. The regulation of use of hospital equipment and
physicians time within the public sector, if these resources can be sold by hospital managers
to the private sector.
34. Communication with the Director of the Christian Medical Board, Dar es Salaam ( July
1994).
Page 59
Capacity issues
All the above issues signal that the new role for the state will require a considerable re
orientation in functions and an improvement in capacity. These will be management and
regulatory based, requiring specific skills and funds to finance the acquisition of such skills.
Management reform will require considerable autonomy at facility or service level in order,
that performance can be correctly attributed to managers and importantly, controlled by them.
Of crucial import within this new role for the state is the generation and use of financial
information.
Conclusion
Prior to independence, the Tanzanian health system was highly segmented. Government
coverage was limited to urban areas with rural coverage limited to mission activity. Moreover
the orientation of health expenditure of the colonial government was consistently towards
curative care at the expense of preventative care. Following independence, government
activity was expanded, especially in rural areas. Following 1967, responsibility for the finance
and provision of health care was embraced by the state. With this responsibility came a
commitment to implement national targets for access.
The funding of the recurrent cost of this expanded access were constrained during the
economic decline which began in Tanzania from the mid 1970’s. During the period of
economic decline the proportion of expenditure on health from the total budget fell. During
this period it has also been suggested that the government failed to reassess national targets
with the result that a falling share of the budget was allocated to an increasing number of
facilities (Lawson 1994).
The decline in quality resulting from dwindling allocations to the health sector has prompted
reform in government activity. Gone is the commitment to finance and provide care for all
irrespective of means. In its place are attempts to introduce direct payments for those able to
pay, with a residual commitment to finance and provide care for those unable to pay. The
state however maintains the function of financing those services with recognised social
benefits (i.e. those that economic theory would predict would be under-provided in a free
market).
These reforms represent a movement from a Type A system to a Type B system, with the
possibility of a movement to a Type C system in the future. As described above, the
movement to Type B has implications for the role of the state. These implications, as well
as the assessment of health systems in Section D are now addressed in the next section.
Page 60
F. Conclusions (to be completed)
Page 61
References
Abel-Smith, B. & P. Rawal (1992) Health Sector Financing Study, first report for Overseas
Development Administration London, mimeographed.
Amani H., S. Hagen, B. Ndulu & K. Stenberg (1987) A Regional and Inter-Regional Study
of Tanzania, Part 1, University of Dar es Salaam.
Barnum H. & J. Kutzin (1993) Public Hospitals in Developing Countries, Johns Hopkins
University Press and The World Bank, Baltimore.
Barr N. (1993) The Economics of The Welfare State, Weidenfeld & Nicolson, London.
Bennett S. (1991) The Mystique of Markets: Public and Private Health Care in Developing
Countries PHP Department Publication No.4, London School of Hygiene and Tropical
Medicine, London.
Bennett S. (1994) The Relationship Between Public and Private Systems in Health Care,
paper for workshop on health financing Institute of Development Studies, Brighton, March
1994.
Birdsall N. & R. Hecht (1994) Strategies for Improving Equity in Health, paper for workshop
on health financing Institute of Development Studies, Brighton, March 1994.
Bloom G & M. Segall (1993) Expenditure and Financing of the Health Sector in Kenya,
Institute of Development Studies Commissioned Study No. 9, Brighton.
Bloom G., G. Singleton & J. Toye (1992) Public Expenditure in the Tanzanian Health Sector
During Structural Adjustment, Report Prepared for the European Community.
Bonsi S. (1992) ‘Foreign Impact on Precolonial Medicine’ in Falola T. & D. Ityavyar (eds)
(1992) The Political Economy of Health in Africa, Ohio University Press, Ohio.
Cassell A. (1993) Health Sector Reform in Practice, draft paper prepared for the Overseas
Development Administration, London.
Cox D. & E. Jimenez (1990) ‘Achieving Social Objectives Through Private Transfers: A
Review’ in World Bank Research Observer, 5(2), pp205-18.
Creese A. & J. Kutzin (1994) Lesson From Cost-Recovery in Health, paper for workshop on
health financing Institute of Development Studies, Brighton, March 1994.
Culyer A. (1989) ‘The Normative Economics of Health Care Finance and Provision’ in
Oxford Review of Economic Policy, Vol 5, No. 1 Spring pp34-58.
Page 62
de Ferranti D. (1985) Paying for Health Services in Developing Countries, World Bank Staff
Working Paper No 721, Washington.
Doriye J. (1992) ‘Public Office and Private gain: an interpretation of the Tanzanian
experience in Wuyts M. et al (eds) Development Policy and Public Action, Open University
Press, Oxford.
Ehrlich C. (1966) ‘Some Aspects of Economic Policy in Tanganyika’ in Smith H. Readings
on Economic Development and Administration in Tanzania, Institute of Public
Administration, Dar es Salaam.
Falola T. & D. Ityavyar (eds) (1992) The Political Economy of Health in Africa, Ohio
University Press, Ohio.
Falola T. (1992) ‘The Crisis of African Health Care Services’ in Falola T. & D. Ityavyar
(eds) (1992) The Political Economy of Health in Africa, Ohio University Press, Ohio.
Ferroni M. & R. Kanbur (1990) ‘Poverty-Conscious Restructuring of Public Expenditure’
Social Dimensions of Adjustment in Sub-Saharan Africa, Working Paper No. 9, The World
Bank, Washington.
Gertler P. & J. van der Gaag (1990) The Willingness to Pay For Medical Care: Evidence
from Two Developing Countries, The World Bank, Washington.
Giffiths A. & M. Mills (1983) ‘Health Sector Financing and Expenditure Surveys’ in Lee K.
& A. Mills (1983) The Economics of Health in Developing Countries, Oxford University
Press, Oxford.
Gilmurray J., R. Riddel & D. Sanders (1979) The Struggle for Health, Rhodesia to
Zimbabwe series No.7, CIIR, London.
Gilson L. (1988) Government Health Care Charges: is equity being abandoned? EPC
Publications No. 15 London School of Hygiene and Tropical Medicine, London.
Gilson L. (1993) ‘Management and Health Care Reform in Sub Saharan Africa’.
Forthcoming in Social Science & Medicine.
Griffin C. (1990) Health Sector Financing in Asia, The World Bank, Washington.
Griffin C. (1988) User Charges for Health Care in Principle and Practice, EDI Seminar
Paper No. 37, The World Bank, Washington.
Hammer J. (1991) ‘A Work Program on Public Economic s of Health in Developing
Countries’, mimeographed.
Hammer J., I. Nabi & J. Cercone (forthcoming) ‘Distributional Effects of Social Sector
Expenditure in Malaysia: 1974 to 1989’.
Page 63
Hammer J. (1993) Prices and Protocols in Public Health Care, Population, health and
Nutrition Department working paper, The World Bank, Washington.
Hatch J. (1972) Tanzania: a profile, Pall Mall Press, London.
Heggenhougen H. (1986) ‘Health Services: Official and Unofficial’ in Bosen J et al (eds)
Tanzania: crisis and struggle for survival, Motala Grafiska, Motala.
International Bank for Reconstruction and Development (1961) The Economic Development
of Tanganyika, Johns Hopkins Press, Baltimore.
Ityavyar (1992) ‘Health in Precolonial Africa’ in Falola T. & D. Ityavyar (eds) (1992) The
Political Economy of Health in Africa, Ohio University Press, Ohio.
Jimenez E. (1987) Pricing Policy in the Social Sectors, Johns Hopkins University Press and
The World Bank, Baltimore.
Kahama G., T. Maliyamkono & S. Wells (1986) The Challenge for Tanzania's Economy
James Curry, London.
Kanji N., P. Kilima & G. Munishi (1992) Quality of Primary Curative Care in Dar es
Salaam, Urban Health Programme, London School of Hygiene and Tropical Medicine,
London.
Kilimanjaro Christian Medical Centre (1994) Annual Report 1992/93, Moshi, Tanzania.
Korte R., H. Richter, F. Merkle & H. Gorgen. (1992) ‘Financing Health Services in SubSaharan Africa: Options for Decision Makers During Adjustment’ in Social Science and
Medicine, Vol. 34, No.l ppl-9.
Lal D. (1994) The Role of the Public and Private Sectors in Health Financing, Human
Resources Development and Operations Policy, Working Papers Series No. 33, The World
Bank, Washington.
Lawson A. (1994) Vndetfunding in the Social Sectors in Tanzania: Origins and Possible
Responses, TADREG Working Paper Series No. Two, Tanzania.
Lee K. & A. Mills (1983) The Economics of Health in Developing Countries, Oxford
University Press, Oxford.
Leubuscher C. (1949) Tanganyika Territory: A study of Economic Policy Under Mandate,
Oxford University Press, Oxford.
Litvack J. & L. Bodart (1992) ‘User Fees Plus Quality Equals Improved Access To Health
Care: Results of A Field Experiment in Cameroon’ in Social Science and Medicine Vol 37.
No.3 pp369-383
Page 64
Mburu F. (1992) ‘The Impact of Colonial Rule bn Health Development: the case of Kenya’
in Falola T. & D. Ityavyar (eds)'(1992) Th& Political Economy of Health in Africa, Ohio
University Press, Ohio.
McGory G. (1993) Health Sector Expenditures and Financing: an analysis of World Bank
PHN sector reports in Africa, mimeographed.
McGuire A., P.Fenn & K. Mayhew (1989) ‘The Assessment: The Economics of Health
Care’ in Oxford Review of Economic Policy, Vol 5, No. 1 Spring ppl-20.
McPake B. (1993) ‘User Charges for health Services in Developing Countries: A Review of
the Economic Literature’ in Social Science and Medicine, Vol 36, No. 11 ppl397-1405.
McPake B., K. Hanson & A. Mills A. (1993) ‘Community Financing of Health Care in
Africa: An evaluation of the Bamako Initiative’ in Social Science and Medicine, Vol. 36, No.
11, ppl383-1395.
Mills A. (1994) ‘Improving the Efficiency of Public Sector Health Services in Developing
Countries: Bureaucratic Versus Market Approaches’, forthcoming in Colclough C. (ed) New
Strategies for Financing Education and Health in Developing Countries: a critical
assessment.
Ministry of Health (1993) Health Sector Reform: Appraisal Mission, Document Report
Prepared for the Government of Tanzania December 1993.
Mukyanuzi F. (1994) Health Sector Overview Paper. Paper prepared for the World Bank,
2nd Draft March 1994.
Netherlands Development Cooperation (1994) Tanzania: Evaluation of the Netherlands
Development programme with Tanzania 1970-1992, Ridderprint BV, Ridderkerk.
Nolan B. & V. Turbat (1993) Cost Recovery in Public Health Services in Sub-Saharan
Africa, Draft October 1993, Human Resources Division, The World Bank, Washington.
Organisation for Economic Cooperation and Development (1992)
Parker D. & R. Knippenberg (1991) Community Cost Sharing and Participation: A review
of the issues, Bamako Initiative Technical Report Series, No. 9. UNICEf, New York.
Pearce T. (1992) ‘Health Inequalities in Africa’ in FalolaT. & D. Ityavyar (eds) (1992) The
Political Economy of Health in Africa, Ohio University Press, Ohio.
Roth G. (1987) The Private Provision of Public Services in Developed Countries, Oxford
University Press, Oxford.
Sahn D. & R. Bernier (1993) Has Structural Adjustment Led to Health Sector Reform in
Africa?, paper presented at conference on health sector reform, Harvard.
Page 65
0
QUEEN ELIZABETH HOUSE
UNIVERSITY OF OXFORD
FOOD STUDIES GROUP
21 St. Giles, Oxford 0X1 3LA
Telephone (0865) 270262
Fax (0865)514468
Position: 1273 (4 views)