Full text of DOI:10.1111/j.1475-5890.2001.tb00037.x

Fiscal Studies (2001) vol. 22, no. 2, pp. 151–183
This is the eighth in a commissioned series of survey articles from distinguished
academics covering the economic issues in public spending.
Expenditure on Healthcare in the
UK: A Review of the Issues
CAROL PROPPER*
Abstract
This review examines the performance of the UK healthcare system. After presenting data on the
level and distribution of resources, three topics are examined. The first is the lessons from
international comparisons of evidence on expenditure, equity and healthcare outcomes. The second
is the lessons from the recent internal market reforms. The third is the lessons from an analysis of
the role for private finance in UK healthcare. The review concludes that economists and
policymakers need to focus more attention on the relationship between healthcare inputs —
expenditure — and health outcomes, and, within this, on the incentives facing suppliers and
demanders of healthcare.
JEL classification: I1, H4.
I. INTRODUCTION
Public expenditure on healthcare in the UK is large. In 2000, just under
£50 billion was spent by the state on healthcare, a sum that accounted for around
6 per cent of GDP. On top of these large public expenditures, individuals buy
*Department of Economics and CMPO University of Bristol, CASE London School of Economics and CEPR.
The author would like to thank Olwen Renowlden for data assistance and Simon Burgess for his very helpful
advice. Funding from the Leverhulme Trust is gratefully acknowledged. Any errors are the author’s.
© Institute for Fiscal Studies, 2001

Fiscal Studies
healthcare and pharmaceuticals directly, and private expenditure on healthcare in
the UK totalled about £9 billion in 2000. Nor is high public and private
expenditure on healthcare just a British phenomenon. Governments and private
individuals throughout the OECD contribute large sums for healthcare
expenditure. In fact, within the OECD and the G7 group of countries, UK
expenditure per capita is low. The UK spends around 7 per cent of GDP per
capita on healthcare, while the G7 average is 9 per cent and the OECD average is
8 per cent. However, in comparison with both the OECD and G7 countries, the
share of expenditure accounted for by the public sector is high. The UK share is
around 85 per cent, while the average for the G7 group is 70 per cent and for the
OECD is 75 per cent.
The large scale of public intervention is justified on both efficiency and
equity grounds,1 but while it is common for governments to intervene in the
market for healthcare, it is also clear that the form and extent of this intervention
vary considerably across countries. This raises questions of whether the UK
spends sufficient sums of either public or private finance on healthcare, whether
the balance between public and private finance is optimal and whether the way
in which those sums are spent achieves either the efficiency or equity goals of
the population. To answer such questions is beyond the scope of a single review.
Instead, this review focuses on a set of topics that are relevant to current debates
about the financing of UK healthcare and also shed light on these broader
questions.
The first topic reflects the debate that other forms of health service
organisation may be an improvement on that currently in operation in the UK. I
examine the relationship between the organisation of the system and three sets of
outputs that are measures of: first, the efficiency of the system; second, the
equity of the system; and third, the outcomes of the system. These are
expenditure per capita, the distribution of payments for, and receipt of,
healthcare resources and measures of health (health outcomes). This examination
is based on comparison of performance across comparable countries, which has
its strengths and weaknesses, which are explored below. The second topic
examines the lessons from the internal market reforms that were implemented in
the UK in the early 1990s. These reforms are part of a general move in
healthcare organisation towards increased competition, and the review examines
the implications for efficiency. The third topic examines the role for private
finance in the funding of UK healthcare. I examine the extent of private finance
in the UK healthcare system and the interrelationships between public and
private finance, with the aim of deriving the implications for the evolution of
public healthcare finance in the UK system.
1See Barr (1998) and Emmerson, Frayne and Goodman (2000) for a review of both efficiency and equity
arguments for government intervention in healthcare markets.
152

Expenditure on Healthcare in the UK
The organisation of the paper is as follows. Sections II and III examine the
performance of the current system. Section II presents the trends in healthcare
expenditure in the UK and compares the UK, in terms of financing and
organisation of the healthcare system, with other OECD countries. Section III
presents the performance of the UK system in terms of the distribution of
resources. Section IV presents evidence from recent international comparative
studies to examine the relationship between organisational design and the
performance of the healthcare system in terms of per capita expenditure. Section
V examines the relationship between system design and equity in the payment
for, and in the receipt of, resources. Section VI examines the relationship
between system design and health outcomes and the distribution of these
outcomes. Section VII examines the internal market reforms and Section VIII
examines the role of private finance and issues in the dynamics of UK healthcare
expenditure. The final section draws out the lessons for the finance of UK
healthcare expenditure and for future research.
II. THE UK LEVEL OF EXPENDITURE ON HEALTHCARE
1. Trends in UK Healthcare Expenditure
Healthcare expenditure in the UK can be broadly divided into public
expenditure, which is mainly spent on the National Health Service (NHS), and
private expenditure, which is defined as expenditure by the household and
corporate sector on healthcare (including dentistry and over-the-counter
medicines). Estimates of public expenditure are somewhat more reliable than
estimates of private expenditure, but whichever estimates are used, the bulk of
expenditure in the UK (around 85 per cent in 2000) is public.
Figure 1 presents real spending on the NHS as a proportion of GDP from
1960 to 1999. The graph shows that the share of the NHS in national income has
grown from under 3.5 per cent to just under 6 per cent in 40 years. But it is also
clear that this increase has not been constant and that there have been years in
which the share has fallen, to be then followed by years in which expenditure has
increased faster than the long-run trend. Figure 2 presents the percentage year-
on-year increase in real NHS expenditure. This shows considerable variation
around the average of 3.4 per cent annual growth for the period. There were
periods of negative growth during the mid-1950s and the early and late 1970s,
and there have been periodic large increases of a size comparable to, or even
slightly larger than, the settlements made by the Labour administration for 2000
and 2001 and forecast through to 2004.
Figure 3 presents private expenditure as a share of GDP. The scale indicates
the smaller share of private expenditure and the graph also shows that private
spending remained small, and indeed even fell, up to the late 1970s. Figure 4
153

Fiscal Studies
FIGURE 1
Real Spending on the NHS as a Percentage of GDP
6.0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1960
1965
1975
1980
1985
1990
1995
1970
Source: Emmerson, Frayne and Goodman, 2000.
FIGURE 2
Real Year-on-Year Growth in Spending on the NHS
10
8
6
4
2
0
1950
1955
1965
1970 1975
1980
1985
1990
1995
1960
–2
–4
–6
Source: Emmerson, Frayne and Goodman, 2000.
154

Expenditure on Healthcare in the UK
FIGURE 3
Real Private Health Spending as a Percentage of GDP
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
1960
1965
1975
1980
1985
1990
1995
1970
Note: ‘Private’ = Total – Public.
Source: OECD Health database 2000.
FIGURE 4
Real Year-on-Year Growth in Private Expenditure
25
20
15
10
5
0
1961
1966
1976
1981
1986
1991
1996
1971
–5
–10
–15
–20
Note: ‘Private’ = Total – Public.
Source: OECD Health database 2000.
155

Fiscal Studies
FIGURE 5
Private Health Spending as a Percentage of Total Health Spending
20
18
16
14
12
10
8
6
4
2
0
1960
1965
1975
1980
1985
1990
1995
1970
Source: OECD Health database 2000.
presents the percentage year-on-year growth in private expenditure in real terms.
The large falls first in the mid-1960s and then again in the mid-1970s are
apparent. The 1980s show a pattern of strong growth, but this tailed off during
the early 1990s, and there has been considerable volatility in growth rates in the
second half of the 1990s. Figure 5 presents private expenditure as a percentage
of total expenditure on healthcare from 1960 to 1999. This shows that the share
that was private rose in the early 1960s, fell from a peak in 1964 to a low in
1975, rose again up to 1991, then fell as NHS expenditure increased rapidly with
the advent of the internal market in 1991, then rose from 1993 to 1997 and fell
again in the late 1990s.
In summary, the picture that emerges is one of long-run growth in both public
and private expenditure but, around this long-run trend, of considerable year-on-
year volatility in both public and private annual growth rates and also in the
share that is private.
2. The UK Compared
Tables 1 and 2 present the UK’s position, in terms of total expenditure and the
share that is public, in comparison with two groups of countries — the OECD
and the G7 countries. The tables indicate substantial differences in the level of
healthcare expenditure across countries, even within the relatively homogeneous
industrialised countries of the OECD. Table 1 illustrates health expenditure per
capita as measured in purchasing power parities (PPPs) in the OECD. In 1998,
these ranged from less than $1,000 (for example, Hungary and Korea) to more
156

Expenditure on Healthcare in the UK
than $2,500 (Switzerland $2,794 and the US $4,178).2 Within these countries,
the position of the UK is towards the bottom end. There is also a range in the
share that health expenditure accounts for in GDP. The highest share is in the
US, the lowest in Korea. Again, the position of the UK is near the bottom of the
OECD group.
TABLE 1
Total Expenditure on Healthcare and Share that is Public: OECD Countries, 1998
Total healthcare
expenditure
per capita
Total healthcare
expenditure
as a share of GDP
(%)
Public finance
as a share of total
(%)
(US$ PPP)
Australia
Austria
2,043
1,968
2,081
2,312
930
2,133
1,502
2,077
2,424
1,167
705
2,103
1,436
1,783
1,822
730
2,215
2,070
1,424
2,425
496
8.5
8.2
8.8
9.5
7.2
8.3
6.9
9.6
10.6
8.3
6.8
8.3
6.4
8.4
7.6
5.0
5.9
8.6
8.1
8.9
6.4
7.8
7.1
8.4
10.4
6.7
13.6
69.3
70.5
89.7
69.6
91.9
81.9
76.3
76.4
74.6
56.8
76.5
84.3
75.8
68.0
78.3
45.8
92.3
70.4
77.1
82.8
73.3
66.9
76.9
83.8
73.4
83.7
44.7
Belgium
Canada
Czech Republic
Denmark
Finland
France
Germany
Greece
Hungary
Iceland
Ireland
Italy
Japan
Korea
Luxemburg
Netherlands
New Zealand
Norway
Poland
Portugal
Spain
1,237
1,218
1,746
2,794
1,461
4,178
Sweden
Switzerland
UK
US
Source: OECD Health database 2000.
2In comparison, in low-income countries, these amounts are much smaller — $10 per capita in many African
countries, and less than $100 per capita in most of Asia and Latin America. The share in GDP is also smaller.
157

Fiscal Studies
TABLE 2
Total Expenditure on Healthcare and Share that is Public: G7 Countries, 1998
Total healthcare
expenditure
per capita
Total healthcare
expenditure
as a share of GDP
(%)
Public finance
as a share of total
(%)
(US$ PPP)
Canada
France
Germany
Italy
Japan
UK
US
2,312
2,077
2,424
1,783
1,822
1,461
4,178
9.5
9.6
69.6
76.4
74.6
68.0
78.3
83.7
44.7
10.6
8.4
7.6
6.7
13.6
Source: OECD Health database 2000.
Table 1 also shows the proportion of healthcare that is publicly financed. The
unweighted average across all OECD countries is 75 per cent, and only in the US
and Korea is the proportion less than 50 per cent. In some countries (for
example, the UK, the Nordic countries and the Czech Republic), almost all
healthcare expenditure is paid for by the public purse.
Table 2 presents these data for the more homogeneous group of G7 countries.
Again, the UK is somewhat of an outlier, with the lowest average expenditure
per capita on healthcare and the highest public share.
The crude share of public finance hides important differences in the finance
and delivery of healthcare. Finance for public healthcare is raised by a mixture
of general taxation, social insurance (payroll tax) contributions and local
taxation. Examples of systems in which general taxation dominates are the UK
and Denmark. An example of a country that uses earmarked social insurance
contributions is the Netherlands. Private finance is usually raised by insurance
premiums and/or out-of-pocket payments. Private insurance can be used to
provide supplementary cover to public cover where those covered are offered no
tax breaks (for example, the UK) or to provide cover for those without
comprehensive public cover (for example, the Netherlands) or to provide cover
against public sector co-payments levied on prescription medicines, dental care
etc. (for example, France and Denmark). Out-of-pocket payments can be
predominantly co-payments, with the third party usually paying the major share
of the bill (for example, the UK, the Netherlands, the US and Denmark) or can
amount to substantial use of the private sector on a fee-paying basis (for
example, Italy, Spain and Portugal).
While the public sector plays a key role in finance, the provision (i.e. supply)
of healthcare varies widely in terms of public/private mix. In the OECD, delivery
spans from being nearly 100 per cent in the public sector (for example, Sweden)
158

Expenditure on Healthcare in the UK
TABLE 3
Classification of Countries by Public/Private Mix
of Provision and Finance in Healthcare
Public/private mix
Mainly public provision, public finance
Country
Ireland
Italy
Norway
Portugal
Denmark
Finland
Greece
Spain
Sweden
UK
Iceland
Mixed provision, public finance
Australia
Austria
Belgium
France
Germany
Japan
Luxemburg
New Zealand
Mainly private provision, public finance
Mixed provision, mixed finance
Canada
Netherlands
Switzerland
Mainly private provision, private finance
Source: OECD, 1994.
US
to being substantially the role of the private sector (for example, Switzerland and
the US). Even where finance is predominantly social-insurance- or tax-based,
countries may use mainly private providers. Canada, for example, has
predominantly public finance but private providers. Table 3 illustrates the
mixture of public and private finance and provision in the OECD.
Given the importance of healthcare expenditure as a share of GDP and of
public expenditure, and the differences in institutional arrangements for
financing and delivering healthcare, an obvious question is the extent to which
differences in expenditure and outcomes are associated with the organisation of
the healthcare system. This is examined in Section IV.
III. THE UK ALLOCATION OF PUBLIC HEALTHCARE
EXPENDITURE
It is clear that successive UK governments have had a strong commitment to
equity in healthcare finance and delivery. In this section, I examine the extent to
which this commitment has been realised and whether it has translated into an
equitable distribution of health outcomes.
1. The Allocation of Resources
Since 1974, an explicit goal in the allocation of public funds for healthcare has
been that these funds should be allocated according to need. Throughout the
1960s, it had become increasingly obvious that the distribution of resources
within the NHS bore little resemblance to the principles of equal allocation for
159

Fiscal Studies
equal need and more to the historic distribution of hospitals. In 1976, the
Department of Health decided to allocate funds to the 14 regions of the NHS on
the basis of need, as measured by standardised mortality ratios (SMRs),
themselves a measure of morbidity and a proxy for need. A region’s need, and
therefore its allocation, would be calculated on the difference between national
and regional SMRs. The resource allocation (RAWP) formula covered about
three-quarters of the NHS current expenditure budget but did not cover either
capital expenditure or payments to general practitioners (family doctors — GPs)
or dentists. At its inception, the scheme meant monies were allocated according
to the RAWP formula to regions, but, within regions, allocations to areas, and
below them districts, were decided by each region.
Over time, there was both greater devolvement of monies according to
RAWP-type principles down to district level3 and criticism of the use of SMRs
as a measure of need. However, despite this criticism, the basic approach was not
modified till the mid-1990s. In part, this was because there was general
agreement over the principle, and, in part, it was because movements to target
allocations (those generated by the RAWP formula) were gradual and dictated
by politicians. Movement to target was achieved by allocation of ‘new’
healthcare resources. Regions kept the same real budget as last year and, if above
target, received no new funds. Extra real resources, if made available at Treasury
level, went to below-target regions. During the 1990s, the formula was adjusted
to incorporate better measures of need and to deal with the fact that changes in
the way the NHS delivers care meant the primary units of allocation were units
of around 100,000 people.
Le Grand, Winter and Woolley (1990) and Le Grand and Vizard (1998)
provide an assessment of the extent to which movements to target reduced
inequalities in allocations across regions. The average distance from target fell
from the inception of the scheme to the late 1980s. Table 4 shows that variation
TABLE 4
Meeting of RAWP Targets
Coefficient of variation between regions
1985–86
1987–88
1990–91
0.093
1992–93
0.091
1993–94
0.140
0.103
0.097
Range of distance from target: districts
1993–94
1994–95
1995–96
27.60
1996–97
20.96
1997–98
14.39
36.80
25.70
Source: Glennerster, Hills and Travers, 2000.
3A district covers around 100,000 people.
160

Expenditure on Healthcare in the UK
in average healthcare spending per person at regional level fell from the mid-
1980s to 1992–93, though it rose in 1993–94 (the last year reported by the
authors) and that variation in the difference from targets at district level fell
during the 1990s.
Recent evidence on the allocation of expenditure at below regional level is
provided by a study of the distribution of health expenditure at ward level within
three urban local authorities in England for 1995–96. Bramley et al. (1998) seek
to allocate all sources of healthcare funding in three relatively deprived cities to
wards.4 In practice, exact allocation of healthcare expenditure to individuals
within small geographical areas is not possible from the administrative sources
used in this study, and for some of the expenditure allocated, allocation is on the
basis of location of the provider (family doctor or hospital) rather than of the
individual user of service. In addition, for hospital services, detailed costs of
services are not known and instead the researchers have to allocate the average
to all users. Given these caveats, Bramley et al. conclude that the distribution of
healthcare expenditure by ward within the three cities was generally less
dispersed than other forms of public expenditure, and was pro-sick and hence
generally higher in more deprived wards. When allowance is made for the
population age structure of the wards, the spending becomes more pro-deprived.
Broadly, it appears that the allocation of healthcare spending at region,
district and possibly also ward level has been such that variations in allocations
have fallen and spending is close to being allocated according to the measures of
need used in the allocation formulae.
2. The Receipt of Healthcare
Equity in the geographical distribution of resources does not necessarily imply
equity in the distribution of healthcare across individuals with different incomes.
In a series of studies, Propper and her co-authors examine equity in the delivery
of healthcare standardised for medical need. These studies estimate the extent to
which violations of the principle ‘equal treatment for equal need’ are
systematically related to income in the UK. The studies use individual-level data
from household surveys that contain data on individuals’ access to resources,
their use of healthcare services and their self-reported health status. Answers to
the self-assessed health questions used in these surveys have been found to be
good predictors of subsequent mortality in a variety of industrialised countries
(van Doorslaer et al., 1997). Individuals’ incomes are defined as their
equivalised household incomes. Medical need is defined as self-assessed health,
and various measures of this are used. Several forms of care (GP care, out-
patient care and in-hospital care) are examined. The approach of using data on
4Wards are small geographical areas with populations of around 5,000 people.
161

Fiscal Studies
TABLE 5
Percentage Shares of NHS Expenditure Standardised for Need
1974
1982
1985
1987
Income quintile
Bottom
24.6
21.6
19.3
17.9
16.6
22.5
20.3
21.1
21.7
14.5
22.7
22.7
19.7
18.9
16.1
22.7
21.2
19.9
19.8
16.3
2nd
3rd
4th
Top
Concentration index
–0.083
–0.092
–0.070
–0.062
Source: Propper and Upward, 1992.
users and non-users contrasts with many other studies of equity in the allocation
of healthcare, which look only at users of healthcare (Propper (1998) reviews
these).
O’Donnell and Propper (1991) examine whether there are departures from
‘equal treatment for equal need’ that are systematically related to income using
data from the mid-1980s and conclude that the distribution of NHS care is
weakly pro-poor. Propper and Upward (1992) examine the distribution of
healthcare utilisation standardised for need by income group for the UK for the
years 1974, 1982, 1985 and 1987. These results are presented in Table 5. The
columns present the share of NHS expenditure received by each income quintile
after standardising for self-reported morbidity. In 1974, the lowest income
quintile received nearly 25 per cent of total NHS standardised expenditure. The
comparable figure in 1987 was just under 23 per cent. The last line in each
column presents the concentration index. This is a measure of the departure of
the cell averages from proportionality and thus is a measure of the departure
from equal treatment for equal need. The value of the index ranges from –1 to
+1, where a negative (positive) number indicates pro-poor (pro-rich) inequity.5
The results indicate a mildly pro-poor distribution in all four years. Propper
(1998) finds a similar distribution for 1991 and 1994.
3. Inequalities in Outcomes
However, whilst the mechanisms for allocation become ever more complex,
evidence on the distribution of health has raised fundamental issues about the
effectiveness of the allocation formula in achieving equitable health outcomes.
While RAWP allocations have reduced regional disparities in funding, there is a
growing body of research that suggests that inequalities in health across
geographical areas have not fallen since the 1970s and may indeed be rising.
5For details of the methodology, see van Doorslaer and Wagstaff (1992) and van Doorslaer et al. (2000).
162

Expenditure on Healthcare in the UK
The Acheson Report (Department of Health, 1998) concluded that, in many
respects, inequalities in health had not improved since the 1970s when RAWP
began, and in some cases had worsened. For example, health status within
London had improved in the better-off (in terms of income) local areas whilst it
had fallen in the poorer areas. Shaw et al. (1999) found that standardised
mortality ratios for the under-65-year-olds were 2.6 times higher in those local
constituencies6 with the worst health outcomes than in the constituencies with
the best health.7 Infant mortality patterns were similar — infant mortality in the
poorest-health constituencies was 2.0 times the level in the best-health
constituencies.
There is also evidence of increasing health inequality from studies of
variation across individuals. Using data from repeated (cross-sectional)
household surveys, Propper and Upward (1992) calculate the concentration
index of self-reported health status for individuals ranked by their equivalised
household income. They find the rich to have better health and that income-
related inequalities increased between 1974 and 1985, but then fell in 1987, to
below the levels of 1982.
Many of those who study inequalities in health in the UK (for example, Shaw
et al. (1999)) argue that the solution lies not in improving healthcare but in
changing individuals’ access to resources. From their detailed review of resource
allocation in the NHS, Glennerster, Hills and Travers (2000) conclude that
allocating resources according to need will be inherently limited in the extent to
which it can overcome differences in health status. Part of the reason is a
technical one: making equal allocations for equal need assumes equal efficiency
of productive units. But they see the more important reason as being that
equalising resources at an area level will not eliminate the differences in
individual behaviour that lead to differences in actual expenditure across
individuals. The problem is compounded by the well-known fact that variations
in health seem to be largely determined by factors outside of the healthcare
system, some well known and others less so.8
IV. DOES THE ORGANISATION OF THE HEALTHCARE SYSTEM
DETERMINE EXPENDITURE?
International comparisons have been used to attempt to answer questions such as
‘does the overall organisation of the healthcare system have any impact on health
expenditure?’, ‘does the use of a gatekeeper to the hospital sector result in lower
6A constituency is the small area used for elections to the national parliament.
7Health outcome of the constituency defined in terms of SMRs for the under-65s.
8In one of the few studies of variation in expenditure across a reasonably homogeneous population, Cremieux,
Ouellette and Pilon (1999) find a positive association between expenditure on healthcare and better life
expectancy and lower infant mortality across Canadian provinces.
163

Fiscal Studies
expenditure?’, does the method of remunerating doctors affect expenditure?’ and
‘do increases in the supply of doctors result in increases in health expenditure?’.
Such international comparisons have also examined the impact of income and of
demographic factors on total expenditure.
These international comparisons tend to be beset by several problems
(Gerdtham and Jonsson, 2000). The first is the weak theoretical base for the
determinants of aggregate health expenditure. There is no clearly accepted model
for the macroeconomic analysis of health expenditure. Few of the estimated
models make clear the causal relationships between aggregate expenditure and
the organisation of the healthcare sector. The importance of the physician as the
agent for the patient and the key supplier of healthcare, the fact that health
services are often provided on the basis of ‘need’ rather than willingness to pay
and the use of non-price rationing (for example, waiting-lists) make the usual
separation of demand from supply difficult. Second, the data may vary in their
reliability. It can be difficult to capture the precise institutional details of a
healthcare system. It can be difficult, for example, to determine which
expenditures to include in healthcare when the same care may be covered in
different countries by the social security system, the health sector or the social
services sector (an obvious example being care for the elderly). The specifics of
the healthcare sector within countries mean that classification of financing and
delivery systems can often be somewhat arbitrary. For example, variables
representing the public fraction of healthcare expenditure, the use of high-cost
procedures and the type of reimbursement system are often only approximations
to the underlying influences of interest. The distinctions between institutional
arrangements of different countries are usually captured by the use of dummy
variables, but this means they often cannot be distinguished from country effects.
Third, many of the extant studies rely on fairly small samples, and increasing the
sample size increases the heterogeneity of the sample. Fourth, many of the
estimates have not incorporated dynamics, even though observed differences in
health expenditure are likely to be the result of both permanent and transitory
differences.9
Given these problems, it is perhaps not surprising that relatively few robust
results emerge from cross-country comparisons of the determinants of healthcare
expenditure. A recent survey (Gerdtham and Jonsson, 2000) concludes that
results must be treated with caution, but identifies the following results. With
respect to the non-institutional variables (those factors that are not features of the
healthcare financing and delivery system), a common and extremely robust
finding is that the effect of per capita income on expenditure is positive and
often close to unity. On the other hand, the effect of population age structure is
generally insignificant. Institutional features appear to have the following impact
9One commentator on early studies in this area concludes: ‘We have had crude data, misspecified equations,
contentious theory and cavalier history’ (Culyer, quoted in Gerdtham and Jonsson (2000)).
164

Expenditure on Healthcare in the UK
on expenditure. The use of primary-care gatekeepers (the UK’s General Practice
system whereby individuals must be referred for treatment and diagnosis in the
hospital sector by their family doctor), the use of direct patient payment for care
followed by reimbursement from a public or private insurer, the use of capitation
payments for physicians in the ambulatory sector and the public provision of
health services (proxied by the ratio of public beds to total beds) are all
associated with lower health expenditures. Use of in-patient expenditure is
associated with higher healthcare expenditures. Evidence on other features of the
healthcare system — for example, the use of budget ceilings on in-patient care,
and whether doctors are salaried or paid on a fee-for-service basis — is less
clear-cut.
These results indicate that both income and institutional factors determine
healthcare expenditure, suggesting that, as countries grow, so will their
healthcare expenditure, but the extent of growth will be determined by the
precise nature of the healthcare system. Looked at another way, these studies
point to the importance of micro-incentives — incentives at the level of the
supplier and the demander of healthcare. These issues are discussed in more
detail below for the UK.
While these studies give some indication of the features of a healthcare
system that may matter, two further caveats need to be borne in mind. The first is
that cross-country studies to date have examined expenditure and not efficiency.
Whilst governments may be concerned with expenditure and expenditure
growth, the real issue that is important for the design of healthcare systems is
efficiency. However, without a way of adjusting expenditure for quality,
efficiency cannot be measured. It is argued that this focus on expenditure distorts
international comparisons. For example, it is argued that looking only at
expenditure ignores the costs imposed by waiting-lists, which tend to be used in
countries that have lower expenditure per capita, and so erroneously equates
higher expenditure with inefficiency.10 More generally, the position of the US as
an outlier with both high expenditure and higher (unmeasured) quality means
that results are often not robust to leaving the US out of the analysis.
Second, none of these studies has fully addressed the issue that the
organisational form and financing of healthcare expenditure are endogenous. For
example, centralised control of, or influence on, health budgets is itself a
response to low income, budget deficits and a desire to control costs. Such
endogeneity will bias the estimates of the coefficients on the institutional design
factors in the estimates. Finally, even income may not be exogenous in the long
run, as the level of healthcare affects productivity.
10For an example of this argument, see Danzon (1992).
165

Fiscal Studies
Age and Expenditure
Perhaps one surprising result from the studies is the lack of an impact of age or
of other measures of need (for example, unemployment) on healthcare
expenditures. The impact of ageing on government expenditure has been a major
concern for both governments and academic writers (for example, Bos and von
Weisacker (1989)). Severe consequences were expected for the healthcare sector
when population ageing was argued to be the cause of rises in government
expenditure up to the mid-1980s (OECD, 1988).11
However, recent analyses of cross-country expenditure data conclude that the
age effect is correlated with income (richer countries in the sample have older
populations) and show that, holding income constant, age per se has little effect.
Recent analysis of country-specific micro-data on healthcare expenditure
indicates that expenditures on healthcare are concentrated in the last few months
of life. An analysis of micro-data for a sample of Swiss individuals in the last
eight quarters of life during the period 1983–92 (Zweifel, Felder and Meier,
1999) indicates that the amount of healthcare expenditure depends on remaining
lifetime but not on calendar age, at least for those over 65. From this, Zweifel et
al. infer that the positive relationship between age and healthcare expenditure
that can be observed in cross-sectional data may be caused by the fact that, at age
80, for example, there are many more individuals living in their last two years
than there are at age 65. The impact of ageing will thus be to push the high levels
of healthcare expenditure to later in individuals’ lives rather than to increase per
capita expenditure.
Cutler and co-authors (Cutler and Meara, 1998 and 1999; Cutler and Sheiner,
1999) carry out detailed analysis of the relationship between ageing and health
using data on healthcare expenditures in the US under the Medicare programme
for elderly persons. Cutler and Meara (1999) analyse the increase in Medicare
spending during the 1980s and 1990s. During these decades, Medicare spending
doubled in real terms despite the fact that the health of Medicare beneficiaries
improved over the period. In attempting to reconcile these two facts, the authors
show that most of the growth has been amongst the oldest old. Between 1985
and 1995, spending among those aged 65–69 rose by 2 per cent annually. In
contrast, spending for those over 85 rose by 4 per cent. However, the source of
this increase is not increased use of acute care (for which the growth rate
amongst the young elderly is in fact higher than that amongst the old elderly) but
increased use of post-acute services.
Cutler and Meara (1999) speculate that this increase in use might reflect three
factors. The first is ‘gaming’ in response to the reimbursement mechanisms used
in Medicare. In-patient care is reimbursed prospectively, so hospitals receive the
same payment regardless of whether they provide rehabilitative care or not.
11Emmerson, Frayne and Goodman (2000) provide projections of the impact of ageing on expenditure for the
UK.
166

Expenditure on Healthcare in the UK
However, post-acute care is reimbursed on a cost basis when services are used.
So, if hospitals unbundle the post-acute care from the in-patient setting, they
collect additional revenues at no extra cost. The second factor is that some of the
additional services might reflect real extra services, perhaps substituting for care
previously provided by a family member. Getting reimbursement for such care
was made easier in the late 1980s. Finally, Cutler and Meara argue that some of
the increase in expenditure might be fraud. The nature of their data means that
they are unable to establish the relative strength of these factors.
This analysis illustrates the importance of the interaction of demography,
health status and the institutional arrangements for reimbursement and delivery
of care. Ageing per se does not necessarily mean that individuals are more in
need of healthcare; in fact, in the US data at least, older groups are healthier than
in the past. However, even though healthier, they use more resources. This is in
part due to changes in family choice (due to changes in taste and in income) and
the availability of better rehabilitative care, but also due to the particular nature
of the reimbursement arrangements for publicly funded care in the US. In other
words, again the literature points to the importance of micro-incentives and the
responses of agents to these incentives in determining the level and composition
of expenditure.
It is also possible that the impact of age will be affected by endogenous
government budgetary responses. Getzen (1992) finds no empirical relationship
between ageing and expenditure and argues (as an ad hoc justification of his
findings) that, during the 1980s, the existence of budget deficits meant that
governments acted to damp down the increased demand that arose from an
ageing population. He concludes therefore that ageing per se does not
automatically mean greater healthcare expenditure: the effect will depend on the
overall fiscal position of the government. This contention remains to be tested.
V. DOES THE ORGANISATION OF THE HEALTHCARE SYSTEM
DETERMINE EQUITY?
There is considerable evidence that policymakers within OECD countries are
concerned about equity in healthcare. As shown in Table 1, the share of the state
in healthcare expenditure is high, and many countries have achieved close to
universal coverage of their population for the majority of healthcare services
(OECD, 1993).
The simple share of public finance in total healthcare expenditure hides the
fact that OECD countries finance their healthcare through a mixture of taxes,
social insurance contributions, private insurance premiums and direct payments
at point of use (known as out-of-pocket payments). These various payment
sources have very different implications for both vertical and horizontal equity in
the payment for healthcare, and may also have an impact on equity in the receipt
of healthcare.
167

Fiscal Studies
1. Equity in the Finance of Healthcare
Wagstaff et al. (1999) examine the progressivity of the healthcare financing
system in 13 OECD countries. Progressivity is measured using the Kakwani
index, which is equal to the difference between the concentration index for
payments and the Gini coefficient for gross (i.e. pre-healthcare-payment)
income. This index measures departures from proportionality. A zero value of
the index indicates proportionality in payments (i.e. all individuals pay the same
proportion of their income for healthcare), and a negative (positive) value
indicates a regressive (progressive) structure. Indices are calculated for the five
sources of payment for healthcare: direct taxes, indirect taxes, social insurance,
private insurance and direct payments. Indices for total taxes (direct and
indirect), total public (taxes plus social insurance), total private and total
expenditures are computed as weighted averages of the relevant separate indices.
The results indicate that direct taxes used to finance healthcare are
progressive. This is particularly true in the UK, Ireland and Germany, but less
true in Sweden and Finland, where a relatively high share of direct taxation for
healthcare finance comes from local income taxes, which are less progressive
than national taxation. Indirect taxes are regressive in all of the countries in the
study. Social insurance is generally progressive, apart from in the Netherlands
and Germany, where higher income groups are excluded from the social
insurance schemes used to finance healthcare. The progressivity of private
insurance depends upon what private insurance buys. In countries where private
insurance buys cover against public sector co-payments (for example, Denmark
and France), progressivity is lower where co-payments are higher (because
poorer individuals spend a higher proportion of their income on insurance
against co-payments). In countries where private insurance is taken out as (often
‘double’ cover) supplementary to cover provided by the state (for example, Italy,
the UK, Portugal and Spain), insurance is generally progressive. This is because
those who pay for cover twice are richer individuals. Where insurance is the sole
form of cover for the majority of the population (the US and Switzerland), it is
highly regressive. Where it is the sole form of cover for richer individuals
(Germany and the Netherlands), it is progressive. Finally, out-of-pocket
payments are generally a highly regressive means of raising revenue, though
their regressivity varies across countries, reflecting the differences across
countries in exemptions from out-of-pocket payments.
The net effect of these different payment mechanisms on the progressivity of
healthcare finance across 12 of the 13 countries is given in Table 6. The
countries are ranked in terms of progressivity of healthcare finance, with the
most regressive at the top. The position in the ranking depends on the weight of
the different payment mechanisms in the total healthcare financing package.
Broadly, it can be seen that countries that rely most heavily on private insurance
168

Expenditure on Healthcare in the UK
TABLE 6
Progressivity of Healthcare Financing
Year of analysis
1992
Progressivity of all payments for healthcare
Switzerland
US
–0.1402
–0.1303
–0.0703
–0.0452
–0.0445
–0.0158
–0.0047
0.0004
1987
Netherlands
Germany
Portugal
Sweden
Denmark
Spain
1992
1989
1990
1990
1987
1990
France
1989
0.0012
Finland
Italy
1996
0.0181
1991
0.0413
UK
1993
0.0510
Note: Progressivity measured as Kakwani index, so a negative (positive) value indicates a regressive
(progressive) structure.
Source: Wagstaff et al., 1999.
have the least progressive healthcare financing system. Tax-financed systems are
proportional or progressive. Countries that rely heavily on social insurance have
less regressive systems than those that rely most on private insurance, but have
less progressive systems than those that use tax finance (the exception being
France, where over 70 per cent of revenues are raised by social insurance).12
2. Equity in the Delivery of Healthcare
The principle of equity in the delivery of healthcare is widely adopted in OECD
country policy documents (van Doorslaer and Wagstaff, 1992). But differences
in the mix of public and private financing, and in the delivery systems, across
these countries may mean that these equity goals are not met in practice. van
Doorslaer et al. (2000) estimate the extent to which violations of the principle
‘equal treatment for equal need’ are systematically related to income in 13
OECD countries. The study uses individual-level data from household surveys.
This contrasts with many other studies of equity in the allocation of healthcare,
which have looked only at users of healthcare. Individuals’ incomes are their
equivalised household incomes, and various forms of care are examined.
12van Doorslaer et al. (1999) examine the impact of healthcare payment sources on vertical and horizontal
equity and redistribution. They find that the vertical effect (unequal treatment of unequals) is far more
important than horizontal inequity (unequal treatment of equals). Public finance systems tend to have small
positive redistributive effects and less differential treatment of equals, while private financing sources generally
have (larger) negative redistributive effects that are, to a substantial degree, caused by differential treatment.
169

Fiscal Studies
TABLE 7
Extent of Progressivity in Delivery of All Medical Care
Year of analysis
1992
Progressivity of delivery of all medical carea
Switzerland
US
Netherlands
Sweden
Denmark
Finland
UK
0.040
0.009
–0.038
–0.014
–0.060
–0.029
–0.016
1987
1992
1990
1994
1996
1989
aA negative (positive) value indicates a progressive (regressive) system.
Note: Countries are ranked in order of increasing progressivity in finance (from Wagstaff et al. (1999)).
Source: van Doorslaer et al., 2000.
Departures from proportionality (equal treatment for equal need) are measured
by an index that ranges in value from –1 to +1, where a negative (positive)
number indicates pro-poor (pro-rich) inequity.13
In all the countries studied, individuals in lower income groups are more
intensive users of the healthcare system. The poor use more services of the
general practitioner, the medical specialist and the hospital. But after
standardising for the level of need for care, there is little or no evidence of
significant inequality in the delivery of healthcare overall. Interpreted literally,
these results indicate that, in the late 1980s and early 1990s, the healthcare
systems of these countries appeared to perform reasonably well on the horizontal
equity criterion of equal treatment for equal need.
Furthermore, there is no evidence of a link between the healthcare financing
system and equity in delivery of care. Table 7 shows the extent of departures
from equal treatment for equal need across countries, where the countries are
ranked in terms of their progressivity in finance (from Wagstaff et al. (1999)).
As can be seen, countries that are the most progressive in terms of finance are
not those that have the most pro-poor distribution of healthcare. For example, the
UK has the most progressive financing system of the group in the table, but
equity in delivery is more pro-poor in Finland, Denmark and the Netherlands.
While this study finds no evidence of significant inequity in the delivery of
healthcare overall, it does find that significant pro-rich inequality arises for
physician contacts. This seems to be due mainly to a higher use of medical
specialist services by higher income groups and a higher use of GP care amongst
lower income groups. This finding appears to be fairly general and emerges in
different countries with very diverse characteristics regarding access to
specialists and provider incentives. It occurs in countries with universal coverage
13For details of the methodology, see van Doorslaer et al. (2000).
170

Expenditure on Healthcare in the UK
TABLE 8
Percentage Saying Government Should Definitely Be Responsible
for Healthcare for the Sick
West Germany
Britain
85
Italy
87
Sweden
n.a.
1985
1990
1996
54
57
51
85
88
n.a.
82
n.a.
71
Source: Taylor-Gooby, 1999.
(Scandinavia) as well as those without (US), in countries with (Denmark and the
Netherlands) and without (Belgium) a GP gatekeeper role and in countries with
(Belgium) and without (Denmark) substantial cost sharing by patients. On the
other hand, this finding does not emerge in two other countries that have few
features in common (the UK and Switzerland).
The lesson that may be drawn from the studies to date of equity in the finance
and delivery of healthcare is that changes to the finance side would appear to
have more impact on equity than changes to the delivery side. However, such
conclusions may be premature. The analysis on the delivery side was not able to
adjust for differences in quality of care across individuals and relies upon
analysis of data drawn from a variety of country-specific surveys. Second, the
results on the finance side are not incompatible with a position in which
countries whose citizens care about fairness in healthcare finance adopt more
progressive finance (i.e. use taxation rather than social insurance or private
insurance and use less co-payment). In other words, the financing system might
be endogenous to the beliefs about fairness.
A study of attitudes towards the role of the state in financing healthcare,
carried out by Taylor-Gooby in 1999, shows that there is a considerable
difference in the level of support for government responsibility for the sick.
Table 8 presents the answers to a question asked in four countries over time, and
shows considerable variation in the level of support for the principle. In addition,
the ranking of these countries in terms of support for the statement is similar to
their ranking in terms of progressivity of healthcare financing. Again, further
research using comparable datasets is needed to ascertain the link between equity
beliefs and financing arrangements.
VI. DOES THE ORGANISATION OF THE HEALTHCARE SYSTEM
DETERMINE HEALTH OUTCOMES?
1. The Level of Health
The above analyses have been concerned with expenditure on the healthcare
system and the level and distribution of healthcare. Health expenditure is,
however, an input; arguably, what individuals care about is the output of the
171

Fiscal Studies
healthcare sector — in other words, health. Obviously, there are many factors
other than healthcare that determine an individual’s health status, but one
measure of performance of healthcare systems is the extent of health of the
population. Tables 9 and 10 present two commonly used measures of outcomes
— life expectancy and infant mortality — for the G7 countries. These tables
make it clear that the UK performs relatively badly on these summary statistics,
though not as badly as the US, which is the biggest healthcare spender. When
countries are ranked in terms of their expenditure, the performance of the UK is
perhaps a little better, in that it is the lowest spender but has outcomes better
than the lowest (though the comparison is less favourable to the UK if the US is
excluded).
TABLE 9
Life Expectancy in G7 countries, 1996
Life expectancy at birth
(years)
Ranking
Life expectancy
Spendinga
Females
83.6
Males
77.0
74.2
75.7
74.9
73.6
74.3
72.7
Females
Males
Japan
France
Canada
Italy
Germany
UK
US
1
2
3
4
5
6
7
1
5
2
3
6
4
7
6
3
4
5
2
7
1
82.0
81.4
81.3
79.9
79.5
79.4
aSpending is ranked according to share of health spending in GDP.
Source: Emmerson, Frayne and Goodman, 2000 (from OECD Health database 1999).
TABLE 10
Infant Mortality in G7 countries, 1996
Infant mortalitya
Ranking
Infant mortalitya
Spendingb
Japan
France
Canada
Italy
Germany
UK
US
4.3
4.9
5.3
6.0
6.0
6.2
8.0
1
2
6
3
4
5
2
7
1
3
4=
4=
6
7
aInfant mortality is defined as the number of deaths at age under 1 year per 1,000 live births.
bSpending is ranked according to share of health spending in GDP.
Source: Emmerson, Frayne and Goodman, 2000 (from OECD Health database 1999).
172

Expenditure on Healthcare in the UK
Examination of outcomes for specific illnesses and conditions tends to show
that the UK performs badly compared with the rest of the G7 group (Emmerson,
Frayne and Goodman, 2000; Coleman, 1999). The UK has high death rates from
ischaemic heart disease and relatively poor survival rates for some common
cancers (lung, heart, breast and prostate; Sikora (1999)). It is argued that these
poor survival rates are indicators of failures in the medical system, although
these failures may be due to lack of money or poor organisation of services (for
example, lack of concentration of cancer services) or both.
While Tables 9 and 10 suggest that the UK is not performing that well, more
detailed cross-country analyses of the relationships between expenditure and
mortality tend to find little relationship between expenditure and mortality. In
early research, Leu (1986) finds no relationship between medical care
expenditures and lower mortality after controlling for per capita income. Hitiris
and Posnett (1992), using OECD data, find limited evidence of a relationship
between healthcare expenditures and mortality rates, despite using a large pooled
time series of countries and allowing for heteroscedasticity and autocorrelation.
Interestingly, the relative position of the UK on the measures has worsened
during the 1990s: prior to this date, the UK position was generally above average
for the G7 group. More formal cross-country analyses indicate some
convergence within EU countries over time in healthcare spending and outcomes
(Nixon, 2000). Those countries that had lower-than-average healthcare
expenditure and lower-than-average outcomes in the 1960s appear to have
caught up with higher spenders and those with better outcomes. Conversely,
those countries with good relative positions in the 1960s and 1970s appear to
have had a fall in their relative position. So countries with worsening trends in
infant mortality over the period 1960–95 include Belgium, Denmark, France,
Finland, Greece, Ireland, the Netherlands, Sweden and the UK. Belgium,
Germany (females only), Denmark, Spain (males only), the Netherlands, Sweden
and the UK experienced worsening trends in terms of life expectancy. Viewed
this way, the relatively poor current position of the UK might simply reflect
regression towards the mean.
2. Inequality in the Distribution of Outcomes across Countries
van Doorslaer et al. (1997) examine income-related inequalities in health across
nine European countries. Using data from household surveys, they calculate the
concentration index of the distribution of self-reported health status.14
Individuals were ranked by equivalent household income. Ill health was
measured by the responses to questions in which respondents were asked to rate
their health. Answers to these questions have been found to be good predictors
of subsequent mortality in a variety of industrialised countries. In all countries,
14For details of the calculation of the concentration index, see van Doorslaer et al. (1997).
173

Fiscal Studies
income-related health inequality is found to be significant. The authors explore
the statistical association between these inequalities and two measures of
healthcare expenditure and the level and distribution of income for the nine
countries. They find that there is little association between health inequalities
and GDP per capita, health spending and the percentage of healthcare spending
that is public. On the other hand, they find evidence of a positive association
between health inequalities and income inequality, as measured by the Gini
coefficient for after-tax income.
VII. LESSONS FROM THE INTERNAL MARKET REFORMS
The UK, along with several other OECD countries, implemented a major reform
of its healthcare system in the 1990s. In 1991, the internal market reforms were
introduced. These separated the functions of provision and of purchase of
hospital-based healthcare, creating one set of agents responsible for provision
and another responsible for purchase. NHS hospitals were to be given greater
autonomy from central and local control, and purchasers were created from area-
based authorities that had previously administered hospitals and from a self-
selected group of primary-care physicians (Culyer, Maynard and Posnett, 1990).
The intention of the reforms was to create competition on the supply side of the
market. Purchasers were free to buy hospital-based healthcare from any provider,
including the small private sector. Providers would compete for contracts to
supply care, which, it was argued, would encourage efficiency (Maynard, 1991).
These reforms are part of a more general healthcare reform process intended
to increase competition in healthcare markets.15 It is generally agreed that such
reforms have brought about less change than their architects hoped.16 However,
within a UK context, the reforms have highlighted the importance of incentives
for providers and purchasers of healthcare. An assessment of the reforms
concludes that ‘the incentives were too weak and the constraints too strong’ (Le
Grand, Mays and Mulligan, 1998). The constraints come, in part, from the
controls imposed by central government (Propper, 1995).
One area over which there was less central control was the behaviour of
General Practice Fundholders. In the internal market, two main classes of buyer
were created: a self-selected group of family physicians (general practitioners, or
GPs), called General Practice Fundholders (GPFHs), and health authorities. The
15Culyer, Maynard and Posnett (1990) and Maynard (1991) provide an overview of the NHS reforms. Propper
(1995) provides an assessment of the limits to reform. Glied (2000) provides a review of the US literature on
managed care. Chalkley and Malcomson (2000) review issues in government contracting for healthcare
services. Le Grand and Bartlett (1993) provide a review of the wider set of ‘quasi-market’ reforms enacted in
the UK.
16Le Grand, Mays and Mulligan (1998) provide an assessment of the UK reforms. Saltman and Figueras (1998)
review the experience of managed competition. A more sociological assessment of several reforms is provided
in Social Science and Medicine, volume 52 (2001).
174

Expenditure on Healthcare in the UK
reforms gave GPs choice, and GPFHs the ability to pay for this choice, for a
subset of treatments.
The fundholding scheme ran from 1991 to 1999. The scheme was voluntary,
and practices joined in different years over the scheme’s life. By its end,
nationally about half of all general practices were fundholders. GPFHs held a
budget from which they were expected to pay for only particular, specified types
of care, including a specified subset of all procedures that an elective hospital
patient might have. These included common elective procedures and accounted
for about 70 per cent of all elective admissions. GPFHs placed contracts for
these procedures with hospitals. These contracts commonly included information
about price and some dimensions of quality, including, in some instances,
waiting times (Glennerster, Matsaganis and Owens, 1994). The contracts were
likely to link payment to activity, with GPFHs basically paying hospitals for
each case treated.
Any surplus from the fund could be retained by fundholders to use in their
practices. As GPFHs were self-employed contractors subject to relatively little
financial monitoring, exactly how these funds were spent was not subject to
detailed scrutiny, and, at the very least, practice improvements translate into
higher income when the GP exits from the practice. Research on the behaviour
of GPFHs shows that GPs were generally active in making changes in the
internal market (Glennerster, Matsaganis and Owens, 1994), and detailed
analyses of the financial incentives embodied in the scheme suggest that GPFHs
responded to financial, as well as non-financial, incentives. The implementation
of the reforms was undertaken in such a way that GPFHs had budgets based on
their referrals in the year immediately prior to entry into the scheme. This gave
them incentives to increase their use of hospital services prior to entering the
scheme in order to inflate their budgets (which they could then hold at this level
for the life of the scheme). Croxson, Propper and Perkins (2001) show that
fundholders did respond to these financial incentives by increasing their referrals
to hospital prior to entry into the scheme, so inflating their budgets upwards for
the duration of the fundholding scheme. Gravelle, Duskeiko and Sutton (2001)
show that fundholding practices responded to positive prices by admitting fewer
patients than non-fundholding practices. They also responded to changes in
waiting times and patient characteristics in a way that was consistent with the
positive financial costs of making referrals.
Evidence on the prices charged by NHS hospitals to GPFHs and the District
Health Authority buyers indicates that external incentives — competition — also
influenced behaviour. The Department of Health set regulations such that price
was meant to equal average cost. However, this rule was not (and probably could
not be) monitored. A limited number of studies of the impact of supply-side
competition on prices were undertaken. These show that higher levels of
competition were associated with lower prices, particularly for services that had
lower costs (Propper, 1996; Propper, Wilson and Soderlund, 1998). These
175

Fiscal Studies
studies also suggest that NHS hospitals gave greater discounts to those buyers
who were more able to move their contracts between sellers.
So, while, in general, the regulatory activities of central government might
have limited the responses of agents in the internal market, it does appear that,
even with such regulation, agents did respond to financial incentives.
VIII. THE ROLE OF PRIVATE FINANCE AND THE DYNAMICS OF
HEALTHCARE EXPENDITURE
While the size of the private sector relative to the public is small, private finance
accounts for around 15 per cent of healthcare expenditure in the UK. This
private expenditure includes co-payments for pharmaceuticals, direct payment
for care provided outside the NHS and payments for health insurance. Setting
aside co-payments for pharmaceuticals (which are free for the elderly, the young
and those in receipt of social security), private payments are made by individuals
who are richer. Besley, Hall and Preston (1999) find the privately insured to be
more likely to be wealthier, more likely to live in prosperous areas of the UK,
more likely to be more educated, more likely to be in work and more likely to be
supporters of the Conservative Party. Propper, Rees and Green (2001) also find
evidence of a cohort effect in purchase: younger individuals are currently more
likely to purchase than their older counterparts. Not surprisingly, adding in the
expenditure on private care to NHS expenditure makes the UK distribution of
health expenditure less progressive (O’Donnell, Propper and Upward, 1993).
There are a number of essentially static arguments in favour of a private
sector that operates alongside a public sector system. First, there are a limited
number of economic models that show that systems in which there is private
provision alongside public are welfare-increasing. Besley and Coate (1991)
argue that systems in which there is de jure universal provision, but in which
richer individuals are de facto allowed to ‘opt out’, can be redistributive, even
when public provision is financed by a non-redistributive mechanism such as a
head tax. Gouveia (1997) argues that allowing the rich to ‘opt out’ maintains the
existence of a public system by keeping tax payments down to levels supported
by a coalition of the rich and the poor. The second, often-advanced, argument is
that the private sector allows individuals choice. The third is that, for a given
budget, individuals who use the private sector may actually benefit those who
remain in the public sector: if richer individuals use private care but pay taxes
for public care, then this will reduce demand in the public sector, so increasing
the resources available per capita in the public sector. For example, in the UK, it
is argued that those who use the private sector for treatment reduce waiting times
for those who remain in the public sector.
But the dynamics of the process may be somewhat different. If the wealthier
receive less of their healthcare through the public system, their commitment to
contribute taxes to the system may decrease. In addition, without the ‘sharp
176

Expenditure on Healthcare in the UK
elbows’ of the middle classes to keep up quality, the quality of the public sector
may fall. A fall may lead to lobbying for lower taxes for the public scheme and a
reduction in budgets. If opting-out does depend on the quality of the public
sector, reductions in the budget will lead to less use of the system by richer
individuals. This will, in turn, lead to further calls for reductions in the budget.
So attempts to cut public expenditure may lead to the public sector becoming a
‘poor service for the poor’.
The relationship between private financing and the evolution of the public
system turns on the relationship between the quality of the public sector, use of
the public and private sectors and political support for the public sector. The
argument that private finance will lead to a downward spiral towards a ‘poor
service for the poor’ depends upon the premisses that support for public sector
financing is negatively associated with private demand and that the demand for a
privately financed alternative is affected by the quality of the public sector.
In practice, there are several possible relationships between support for public
finance and use of private healthcare. Users of private healthcare services may
be less supportive of public services on ideological grounds or because they
simply see little personal return from public services they do not use.
Alternatively, users of private services may be frustrated with the level of service
available from the public sector, although ideologically they may prefer higher
state spending to achieve higher service levels and quality for all. Use of the
public or private sector may affect attitudes. Users dissatisfied with the level of
service in the public sector may switch to the private sector. Users of the private
sector may like the quality of service they receive and no longer see themselves
as potential beneficiaries of the public service, and in turn lessen their support
for state provision.
At an EU level, Mossialis (1997) finds a positive relationship between
expenditure per capita and satisfaction with the healthcare services, though there
are two outliers (Denmark and Italy). In addition, he finds a north–south divide
in citizens’ satisfaction with healthcare systems, with those in the north being
more satisfied, those in the south being less so and the UK (and Ireland) lying
between the two. In an examination of the dynamics of expenditure across
countries (using OECD data), Globerman and Vining (1998) find no relationship
between the current share of public expenditure in a country and past shares of
public expenditure on healthcare in that country. In other words, low current
levels of public expenditure do not necessarily imply low levels of expenditure
in the future.
At a national level, evidence from the UK suggests that those who use the
private sector are, in general, less supportive of public financing of healthcare.
Users of private healthcare services and, more particularly, users of private
medical insurance are less supportive of increases in expenditure on the NHS or
the equity goals of the NHS (Burchardt, Hills and Propper, 1999; Hall and
Preston, 1998). There is also evidence from the UK that the quality of the NHS
177

Fiscal Studies
is associated with use of the private sector: longer waiting-lists have been found
to be linked with higher levels of demand for private medical insurance (Besley,
Hall and Preston, 1999), though not with use of the private sector (Burchardt,
Hills and Propper, 1999). Besley et al. and Calnan, Cant and Gabe (1993) also
find evidence of a link between dissatisfaction with the quality of the NHS and
private insurance purchase in the UK. However, both studies stress that it is
dissatisfaction with the quality of service, rather than the concept of public
provision, that drives people into the private sector.
But there is considerably less evidence that private use leads to change in
attitudes. Burchardt and Propper (1999) find that use of private healthcare in the
UK in the early 1990s did not lead to clear-cut changes in attitudes towards the
NHS over the following five years. Furthermore, over this five-year period, the
attitudes of both private sector and NHS users changed, and the changes in
attitudes of both sets of users were very close. It appears that, in the UK at
present, it is use of a service that leads to attitude change and not whether that
individual uses a public or a private service.
These studies point to an interaction between NHS utilisation, private
utilisation and attitudes, which suggests that the state of the NHS might have an
impact upon the utilisation of the private sector alternative by the individual. The
evidence is less clear as to whether this will translate into lower support for the
NHS and for taxes for the NHS. Judge, Mulligan and New (1998) find that levels
of dissatisfaction with healthcare fluctuated between 1983 and 1996 but rose
towards the end of the period to be higher than during the previous 13 years. A
strong association exists (at national level) between actual levels of expenditure
on the NHS, spending priorities and dissatisfaction with the NHS. Public support
for higher spending and levels of dissatisfaction with the NHS rose in the second
half of the 1980s when funding increases were relatively small, fell back again in
the early 1990s when spending rose, then rose again as spending slowed down in
the mid-1990s. However, the same study finds little support for major changes in
the traditional method of tax finance for the NHS during this period. In a study
of attitudes to public sector funding, Brook, Hall and Preston (1997) find that the
use of private sector alternatives for healthcare is an important determinant of
what individuals think is a funding priority for themselves, but is a less important
determinant of what they perceive to be in the national interest.
Given this evidence, it appears that a decrease in expenditure on the NHS
may lead to increased use of the private sector, but this will not lead to a large
shift in support away from the NHS. Increasing NHS expenditure may therefore
paradoxically put greater pressure on the NHS as more individuals use the
higher-quality public service. Conversely, the findings also suggest that there
may be scope for more private finance at the margin without threatening the tax
base of the NHS.
In considering extensions to private finance, it is necessary to take into
account possible supply-side responses. Under the current performance
178

Expenditure on Healthcare in the UK
monitoring arrangements within the NHS, in which hospitals are set waiting-list
or waiting-time targets, Martin and Smith (1999) find that an increase in waiting-
lists results in larger increases in supply than reductions in demand. They
attribute this to the fact that hospital managers are strongly monitored in terms of
waiting-time/waiting-list targets. In addition, physicians who work in the private
sector generally also work in the public sector and are not strongly monitored in
terms of hours of work supplied to the NHS. So any exogenous increase in
demand for private services (say a tax break for private purchase) or a change in
quality of the NHS will affect both sectors.
Measures to stimulate private finance could lead to a worsening of service in
the NHS as staff moved into the private sector to meet the increased private
demand. If the fall in demand for NHS care were sufficiently large, NHS
waiting-lists would then fall. This might stimulate some NHS demand, but it
might also reduce the pressure on managers to exert effort, and, if the Martin and
Smith estimates are taken at face value, would result in less NHS activity. So, in
the short run, giving tax breaks could harm NHS users. The benefit to private
users would depend on whether prices in the private sector rose and on the
difference in the benefits of treatment — including the difference in waiting
times — in the two sectors.
IX. CONCLUSIONS
This review has examined the financing of the UK healthcare sector using
evidence from both international comparisons and within the UK. The
international comparisons suggest the following. The lower expenditure per
capita of the UK can, in part, be attributed to organisation of the healthcare
system. Features of the UK system, such as the use of primary-care gatekeepers
(the UK’s General Practice system whereby individuals must be referred for
treatment and diagnosis in the hospital sector by their family doctor) and of
budgets set at a system-wide level, appear to keep expenditure below what would
be expected, given national income. The evidence on equity in finance and
delivery indicates that, on the finance side, the UK has one of the most
progressive systems. On the delivery side, the NHS is not worse, and indeed
somewhat better, in terms of equity in the use of healthcare resources, than
several other OECD countries that spend considerably more per capita on their
healthcare.
On the other hand, the international evidence on expenditure cannot be taken
as evidence of efficiency, as no adjustment is made for quality. Further, the
evidence also indicates that that these positive properties of the UK system may
not translate into health outcomes. The UK has, on some measures, poorer health
outcomes than similar countries and a poorer distribution of health. While these
outcomes might be better than expected, given the level of funding, the fact that
little relationship between funding and mortality has been found at cross-country
179

Fiscal Studies
level makes such an inference something of a leap of faith. The distribution of
outcomes appears to be more related to the distribution of income than to any
measures of healthcare spending.
What international comparisons also hint at is the importance of incentives at
the level of the supplier and the demander. Funding analyses show that the UK
use of gatekeepers and payment of doctors by salary rather than fee-for-service
mean expenditure per capita is lower than it would be with other design features.
However, cross-country analyses are generally at a level at which it is hard to
isolate the impact of such incentives: to understand these micro-incentives, it is
necessary to examine data in which the incentive structures can be more clearly
identified.
Examination of responses to the financial incentives embodied in the UK
internal market reforms indicates that physicians and healthcare providers do
appear to respond to financial incentives in a tax-financed, predominantly
publicly provided healthcare system. Referral patterns of GPs reflected financial
(and other) rewards. Where competition was stronger, prices appeared to be
lower. However, it is also clear that responses of agents were limited by the
structures established, and the actions taken, by central government to ‘manage
the market’.
In terms of the type of finance, the review suggests that there are no clear
equity grounds for moving away from tax finance. The present tax-financed
system is more equitable than either social insurance or private insurance. On the
other hand, there are growing pressures on the NHS, which are reflected in
public attitudes. The evidence suggests that a marginal extension of private
finance will not necessarily erode public support for the NHS.
Finally, the review highlights the importance of moving on from a focus on
what are basically inputs — expenditure and its distribution — to an
examination of outcomes and the links between inputs and outputs. The UK
appears to meet its equity goals well in terms of how it spends public finance,
but this is not mirrored by an increase in equity in health outputs. The existence
of health inequalities raises the issue that the focus by governments on inputs
and amounts of money spent is somewhat skewed. As important is a focus on
what is being achieved for this money. To assess the extent to which health
outcomes have much to do with expenditure on healthcare, greater research
effort needs to be directed to looking at the impact of expenditure on health. In
other words, economists need to direct effort to the study of the efficiency of
production. This, in turn, suggests a greater focus on the responses of suppliers
and demanders to incentives.
REFERENCES
Barr, N. (1998), The Economics of the Welfare State, third edition, Oxford: Oxford University
Press.
180