1 WPS3790

Health Systems in East Asia: What Can Developing Countries
Learn from Japan and the Asian Tigers?

Adam Wagstaff

The World Bank, Washington DC, USA

Summary

The health systems of Japan and the Asian Tigers (Hong Kong (China), the Republic of Korea, Singapore and Taiwan
(China)), and the recent reforms to them, provide many potentially valuable lessons to East Asia's developing
countries. All five systems have managed to keep a check on health spending despite their different approaches to
financing and delivery. These differences are reflected in the progressivity of health finance, but the precise degree
of progressivity of individual sources and the extent to which households are vulnerable to catastrophic health
payments depend too on the design features of the system--the height of any ceilings on social insurance
contributions, the fraction of health spending covered by the benefit package, the extent to which the poor face
reduced copayments, whether there are caps on copayments, and so on. On the delivery side, too, Japan and the
Tigers offer some interesting lessons. Singapore's experience with corporatizing public hospitals--rapid cost and price
inflation, a race for the best technology, and so on--illustrates the difficulties of corporatization. Korea's experience
with a narrow benefit package illustrates the danger of providers shifting demand from insured services with
regulated prices to uninsured services with unregulated prices. Japan, in its approach to rate-setting for insured
services, has managed to combine careful cost control with fine-tuning of profit margins on different types of care.
Experiences with diagnosis-related groups in Korea and Taiwan (China) point to cost-savings but also to possible
knock-on effects on service volume and total health spending. Korea and Taiwan (China) both offer important
lessons for the separation of prescribing and dispensing, including the risks of compensation costs outweighing the
cost savings caused by more `rational' prescribing, and cost-savings never being realized because of other
concessions to providers, such as allowing them to have onsite pharmacists.

Corresponding author and contact details: Adam Wagstaff, The World Bank, 1818 H Street NW, Washington,
D.C. 20433, USA. Tel. (202) 473-0566. Fax (202)-522 1153. Email: awagstaff@worldbank.org.

Keywords: Health systems; Kong Kong; Japan; Korea; Taiwan; Singapore.

World Bank Policy Research Working Paper 3790, December 2005

The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange
of ideas about development issues. An objective of the series is to get the findings out quickly, even if the
presentations are less than fully polished. The papers carry the names of the authors and should be cited
accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors.
They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they
represent. Policy Research Working Papers are available online at http://econ.worldbank.org.

Acknowledgements: My thanks to Soonman Kwon and Gabriel Leung for help in locating resources, and for
sharing with me their unpublished manuscripts.

I. INTRODUCTION

Developing countries--understandably--often show more interest in learning from the
health systems of industrialized economies, rather than from one another's. For the developing
countries of East Asia, one especially relevant group of industrialized economies are Japan and
the Asian Tigers (Hong Kong (China), the Republic of Korea, Singapore and Taiwan (China)). It
is true that even countries such as China, Malaysia and Thailand (the richer developing countries
in East Asia) lag far behind Japan and the Tigers in terms of per capita income, and that on
present trends they may well never catch up (Figure 1).* But in terms of health system
challenges and solutions, the gaps are smaller than per capita income gaps suggest--both within
the Japan-Tiger grouping, and between the Japan-Tiger grouping and the developing countries of
East Asia.

Figure 1: Per capita incomes in selected East Asian economies, 1960-2003

40,000
China
Hong Kong, China
35,000 Japan
Korea
s
30,000 Malaysia
Singapore
price
0 Taiwan, China
25,000 Thailand
200
S
$U 20,000

capita 15,000
per

GDP10,000

5,000

0
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

All health systems face the problem of how to pay providers. Japan, richer than the
Tigers, has been most successful in devising a non-inflationary and non-distortionary fee
schedule, but it is Korea and Taiwan (China) that are ahead in introducing prospective payment
methods. Policymakers in many systems are conscious of the pitfalls of the shared approach (a
legacy of Chinese medicine) of having the same person both prescribe and dispense medicines.

*In terms of GDP per capita, China today is roughly where Korea and Taiwan (China) were in 1960. Malaysia today is roughly
where Taiwan (China) was in 1997, and Korea was in 1984. And Thailand today is roughly where Korea and Taiwan (China)
were in the early 1970s.
Data are from the World Bank's World Development Indicators. GDP per capita is at constant 2000 prices, expressed in US
dollars, not purchasing power parities (PPP) due to the shorter time-series for PPP GDP per capita data.

But again it is Korea and Taiwan (China) that are making the fastest progress on the issue. Most
systems in the region have opted for some form of universal coverage, or aspire to achieve it.
Korea introduced universal health insurance (UHI) in 1989, six years ahead of Taiwan (China)
despite being poorer than Taiwan (China), and despite being poorer than Japan was when it
introduced UHI (1961). China, The Philippines, Thailand and Vietnam are all currently taking
major steps towards universal coverage, and on present trends look set to achieve it well before
they reach even the per capita income Korea had when it did so.

In short, despite the wide income differences within the Japan-Tiger grouping and
between that grouping and the developing economies of East Asia, the opportunities for lesson-
learning in the health sector would seem to be considerable. This paper provides an analytical
overview of the health systems of Japan and the Asian Tigers, and suggests some lessons for the
developing countries of the region. The focus is not on describing the five systems, but rather on
linking their policies and institutional arrangements to outcomes of interest, such as health
spending, unit costs, and use of services. The paper tries to pull out lessons both from looking
across the systems, and from looking at the impacts of reforms within the various systems. Over
the last few years, the availability of quality comparable data across the systems has improved.
The efforts by OECD1 and WHO2 on health spending and sources of financing have helped
improved comparability of data for Japan, Korea and Singapore, and Hong Kong (China)* and
Taiwan (China)3 have complemented this work with their own health accounts. And four of the
five systems--Singapore is the exception--have also been represented in a multi-system Asian
comparative study on equity in health care finance and delivery, known as the Equitap project.
In addition, the last few years have seen several interesting policy reforms in the systems, some
of which have been the subject of careful evaluation. These evaluations, as well as the data
emerging from the health accounts and Equitap project, constitute the evidence base for this
paper.

The paper begins in section II with the financing of health care, comparing the shares of
GDP spent by the different systems on health care, and the revenue sources used in the financing
of it. Also reviewed is evidence on the progressivity of health financing in the different systems,
and the extent to which households--especially poor ones--are protected from catastrophic
health expenses. The results are interpreted in the light of schemes aimed at cushioning the poor
from copayments. The paper then turns in section III to the delivery and use of health services,
reviewing the roles of the public and private sectors, and the ways providers are paid and their
influence on provider organization. Also reviewed are recent reforms concerning provider
payments, and the separation of drug prescribing and dispensing. The paper reviews in the light
of this the evidence on utilization rates, comparing rates to OECD averages, and the extent of
inequalities across income groups in use of different types of services. The final section--section
IV--contains the conclusions.

*Hong Kong (China)'s `domestic' health accounts are available at http://www.hwfb.gov.hk/statistics/en/dha.htm.
The project's homepage is at http://www.equitap.org/.

II. HEALTH CARE FINANCE

How much do Japan and the Tigers spend on health care? How do they finance their
spending? Who bears the financing burden? Do the poor pay proportionately more or less for
their health care than the better off? And how far are people protected from catastrophic health
payments? These are the questions addressed in this section.

How much is spent on health care?

Across health systems, it is well known that the share of GDP devoted to health tends to
rise with GDP per capita. Japan--a good deal richer than the richest Tiger economy (Hong Kong
(China))--actually spends somewhat more than is predicted from a cross-section regression of
total health expenditure (as a share of GDP) on per capita income (Figure 2). The Tigers, by
contrast, spend less than predicted by the regression, the biggest `under-spender' by a long way
being Singapore, whose share of GDP devoted to health is barely half the predicted figure. (The
other three Tigers all spend 80-90% of their `expected' shares.)

Figure 2: Health spending and GDP*

14
Actual

PDG 12 Predicted

% 10
as

Exp 8 Japan

altheH 6 Taiwan, China Hong Kong, China
Korea
Total 4 Singapore

0
0 10,000 20,000 30,000 40,000 50,000
GDP per capita ($US)

*Data refer to 2001 the latest year for which complete data are available. The predicted figures are from a regression of the
natural logarithm of total health expenditure (expressed as a share of GDP) on the natural logarithm of GDP per capita. Data are
from the 2005 World Health Report annexes in the cases of Japan, Korea and Singapore. The per capita income data for Hong
Kong (China) and Taiwan (China) are from the World Bank's World Development Indicators. Data on health spending are from
http://www.hwfb.gov.hk/statistics/download/dha/en/table1.pdf in the case of Hong Kong (China) and from
http://www.doh.gov.tw/ufile/doc/200411_Statistic%20of%20Expenditure%20for%20Health,%201991-2003.pdf in the case of
Taiwan (China).

Between 1998 and 2001, all four economies increased their share of GDP devoted to
health, with Hong Kong (China) increasing its share most slowly and Taiwan (China) most
quickly (Table 1). Over the same period, Singapore's share fell further below its expected share,
but the other Tigers all moved closer towards their expected share, Taiwan (China) most quickly.
It was over this period that Japan moved from being on the regression line to above it. In sum,
while the degree to which the Tigers `under-spend' on health care has--with the exception of
Singapore--diminished somewhat recently, this trend would have to continue much longer
before one could begin to question the ability of these systems to keep aggregate spending under
control.

Table 1: Changes in actual and predicted share of GDP spent on health, 1998-2001*

Share of GDP spent on health Actual share as percentage of predicted share
Percentage
1998 2001 point Percentage 1998 2001 % change
change p.a. change p.a.

Hong Kong (China) 5.6 5.7 0.02 1% 78% 82% 4%
Japan 7.2 7.8 0.12 3% 99% 109% 11%
Korea 4.3 5.1 0.16 6% 68% 79% 16%
Singapore 4.2 3.9 -0.06 -2% 60% 57% -5%
Taiwan (China) 5.5 6.0 0.11 3% 82% 91% 12%

How is health spending financed?

The five health systems vary in how they raise the revenues used to finance health care
(Figure 3). At one extreme is Hong Kong (China), which finances just over half of its health
spending through general tax and non-tax revenues. At the other extreme is Singapore, which
finances as much as two thirds of its health spending through out-of-pocket payments. In the
middle are Japan, Korea and Taiwan (China), all of which rely on social insurance for half (or
nearly half in the case of Korea) of their health spending. Noteworthy is the fact that out-of-
pocket payments account for over 30% of health expenditures in all systems, with the exception
of Japan, and that private insurance accounts for a fairly small share in Hong Kong (China) and
Taiwan (China), and a negligible share in the other three systems.

*The predicted figures are from the same regression as used in Figure 2. Data for 1998 are from the same sources as in Figure 2.

Figure 3: Health financing mixes in Japan and the Asian Tigers*

100%

90%

80%

ng 70%

pendis 60% OOPs
thl
50% Priv. Ins.
heal Soc. Ins.

ta 40% Taxes
to
of 30%
%

20%

10%

0%
Hong Kong, Japan Korea, Rep. Singapore Taiwan,
China China

Figure 4 shows the evolution of the out-of-pocket share from 1960 onwards, as well as
the dates where key reforms occurred, notably the introduction of universal health insurance
(UHI) in Japan (1961), Korea (1989) and Taiwan (China) (1995), and the introduction of medical
savings accounts (MSAs) in Singapore (1984). The high pre-UHI out-of-pocket shares in Japan,
Korea and Taiwan (China) reflected in part the uncovered sections of the population, but also the
services that existing plans left uncovered. For example, in Taiwan (China) prior to UHI, 43% of
the population was uncovered, because existing schemes did not cover dependents of formal
sector workers, the elderly, the disabled and the young.5 In Korea, many services were--and
continue to be--uncovered by health insurance plans, for which patients pay out-of-pocket.6

The out-of-pocket share dropped markedly following the introduction of UHI in Korea
and Taiwan (China), and after a lag started falling in Japan too. By contrast, the introduction of
MSAs in Singapore appears to have had little effect on the out-of-pocket share there. Indeed, in
contrast to the systems that went down the UHI route, which have all seen downwards trends in
the out-of-pocket share rather than simply one-off reductions, Singapore's out-of-pocket share
has, if anything, drifted upwards since MSAs were introduced. MSAs themselves have
consistently accounted for only 8-10% of total health spending in Singapore.7-9 This stems from
the fact that they are mainly for inpatient spending, and there is an upper limit on the amount that
can be withdrawn per day.10 Singapore's two complementary schemes, Medishield (a low-cost
catastrophic illness insurance scheme, introduced in 1990) and Medifund (a means-tested public
safety net of last resort for the poor, introduced in 1993), have absorbed even smaller shares,
currently less than 2% combined.9

*Data are from the 2005 World Health Report in the cases of Japan, Korea and Singapore, and refer to 2003. For Hong Kong
(China) and Taiwan (China), they are from O'Donnell et al.4

Figure 4: Trends in the out-of-pocket share in Japan and the Asian Tigers*

100%
Hong Kong
90% Japan
Korea
80% Singapore MSA
Singapore
Korea UHI
70% Taiwan
spending

althehl 60%
Japan UHI Taiwan UHI

taot 50%

in 40%
erahs 30%

OOP 20%

10%

0%
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010

Who bears the burden of health finance?

How is the burden of health finance distributed across income groups? Figure 5 shows
the progressivity of the various finance sources as well as of total payments for four of the five
systems. Only in Hong Kong (China) are total payments progressive. They are most regressive
in the case of Japan, but the differences with Korea and Taiwan (China) are not marked.

*Data for Hong Kong (China) are from http://www.hwfb.gov.hk/statistics/download/dha/en/table2.pdf. Data for Japan and Korea
are from OECD Health Statistics. Data for Singapore are from Liu and Yue7 for 1960-95, and from the 2005 World Development
Report for 1998-2002. Data for Taiwan (China) are from Lu11. UHI denotes universal health insurance. MSAs denote medical
savings accounts.
Distributional data of the type in the next two charts are not available for Singapore, which is a pity given its uniqueness and
the often strong claims that are made for its success.

Figure 5: Progressivity of health financing sources in Japan and the Asian Tigers*

0.40

0.30

0.20

xedini Taxes
0.10 Soc. Ins.

an Priv. Ins.

kw 0.00 OOP

Ka Total

-0.10

-0.20

-0.30
Hong Kong, Japan Korea, Rep. Taiwan, China
China

The differences in the progressivity of total payments reflect the different shares raised
from the various sources (Figure 3) and the progressivity of each source (Figure 5). Taxes are
most progressive in Hong Kong (China), which is also the health system that relies most on them
in the financing of health care. They are regressive in Japan (largely because of the
regressiveness of indirect taxes there4), but progressive--albeit less so than in Kong Hong
(China)--in Korea and Taiwan (China).

Social insurance contributions are regressive in all systems where they are used. This is
not unusual and typically reflects the effects of contributions ceilings.13,14 Any exemptions or
reduced contribution rates among the poor would reduce the degree of regressiveness, and could
in principle make contributions progressive. In Korea15 and Taiwan (China)11 there are
exemptions or reduced contribution rates for the poor or otherwise disadvantaged (in the case of
Taiwan (China) the poorest 1% of households are fully exempt, and a further 4.5% have their
contributions paid in full or subsidized). In Japan, contributions emerge as proportional despite
the fact that different insurance plans (there are over 5000 plans under the UHI umbrella) have
different contribution rates.16-18

The progressiveness of private insurance--especially so in Taiwan (China)--reflects the
fact that it is the better off who buy it. In both Hong Kong (China) and Taiwan (China), private
insurance buys supplementary cover. Out-of-pocket payments emerge as proportional in Hong
Kong (China) and Korea, mildly regressive in Taiwan (China), and regressive in Japan. The

*A positive value of Kakwani's12 index indicates a progressive payment structure, a zero value proportional payments, and a
negative value a regressive structure. The data shown are from O'Donnell et al.4
The Kakwani indices for total payments are a weighted average of the Kakwani indices for the individual payments sources
where the weights are the revenues shares.

progressivity (or regressiveness) of out-of-pocket payments reflects in part the fraction of health
services covered by the benefit package, and in part the extensiveness of exemptions and reduced
copayments for insured services, and the degree to which these are linked to income. If all
services are covered by insurance, and everyone is liable to the same copayments, out-of-pocket
payments will inevitably emerge as regressive. If, on the other hand, the benefit package is fairly
narrow, and the better off are those paying for the uncovered services, out-of-pocket payments
are likely to emerge as progressive. This will be offset--at least in part--by any linking of
copayments to income for the insured services. Korea is an example of the second situation. The
UHI benefit package is narrow, and half of all out-of-pocket payments are for uncovered
services.19 However, UHI exempts people on Medical Aid from copayments on covered services,
and there are reduced copayments for the elderly, patients requiring long-term care for
catastrophic and/or chronic conditions. Japan and Taiwan (China), by contrast, have near-
comprehensive benefit packages, and in Japan's case, extra-billing (charging patients extra on
top of the insurer's fee schedule) is strictly prohibited, except for nine well defined areas.5,11,16-18
Taiwan (China)--in contrast to Japan--exempts very poor households from cost-sharing.5,11

These progressivity results do not shed light on the question of horizontal inequities in
the finance of health care.20 For example, a major issue in all three social insurance systems has
been the different contribution rates across insurance plans. In Korea over the period 1989-2000,
350 separate insurance plans coexisted under the UHI umbrella. In Japan, even now over 40
years after the introduction of UHI, more than 5000 plans coexist under the UHI umbrella. All
offer the same benefits, but their contribution bases and risks vary. In both countries, the
equalization funds managed to reduce but did not eliminate the effects of these differences. The
result was that prior to the merging of funds in Korea in 2000 and in Japan even today, different
insurance schemes had (and continue to have in Japan's case) different contribution rates.15,18

Are households protected from catastrophic spending?

Also of interest--especially for out-of-pocket payments which are unpredictable and
potentially large--is the extent to which payments for health care turn out to be catastrophic, in
the sense they absorb an unduly large share of income or household spending.21,22 In addition to
being interested in the fraction of the population incurring catastrophic out-of-pocket expenses,
one might want to know the extent to which catastrophic payments are concentrated among the
poor.21

Figure 6 shows the incidence of catastrophic payments for different thresholds
(annualized per capita payments exceeding 5%, 10, 15% and 25% of total annual per capita
household spending), as well as the extent to which catastrophic payments are concentrated
among the poor (a negative concentration index) or among the better off (a positive
concentration index). Whatever the threshold, Korea emerges as having a higher incidence of
catastrophic out-of-pocket payments than Hong Kong (China) and Taiwan (China). For a
threshold of 10% and higher, Hong Kong (China) and Taiwan (China) have similar fractions of
people recording catastrophic out-of-pocket expenses. However, what sets Taiwan (China) apart
is that as the threshold is raised, catastrophic payments become more and more concentrated
among the poor, whereas the opposite is true in Hong Kong (China) and Korea. This is despite
the fact that in Taiwan (China)--as in Japan and Korea--out-of-pocket payments for insured

services are capped.5,11,16-19 The reason why in Taiwan (China) catastrophic payments become
increasingly concentrated among the poor as the threshold is raised but do not in Korea is likely
to be due to the fact that Korea's benefit package is so limited and that of Taiwan (China) so
broad. Those incurring especially large out-of-pocket payments in Korea are likely to be people
whose out-of-pocket spending reflects at least in part payments for uncovered services, the
demand for which is likely to rise strongly with income. In Taiwan (China), by contrast, they are
likely to be people incurring large copayments for covered services, and the demand for these, it
seems, is not higher among the better off (more on this below).

Figure 6: Catastrophic out-of-pocket payments in Japan and the Asian Tigers*

0.20

25% Hong Kong, China
Korea, Rep.
0.15 Taiwan, China
ents
m
15%
pay 0.10 25%

phico
str 0.05
ta
car 5%

fo 10%
0.00 15%
5%

kwani 10% 10%

Ka-0.05 5%
15%

25%
-0.10
0% 5% 10% 15% 20% 25%
% with catastrophic payments

The introduction of UHI in Japan, Korea and Taiwan (China) all led (albeit after a lag in
the case of Japan) to a reduced emphasis on out-of-pocket payments. It is plausible that exposure
to the risk of very large out-of-pocket payments also fell. If this is the case, households affected
by UHI can be expected to have reduced precautionary savings. In the case of Taiwan (China),
this appears to have happened--estimates derived by comparing before-and-after changes in
saving behavior of those gaining insurance as the result of UHI suggest that the introduction of
UHI reduced savings in Taiwan (China) by 9-14%.24 The largest proportionate impacts were
observed for those saving least before the reform.

*The figure shows the fraction of households with annualized per capita out-of-pocket payments exceeding 5%, 10%, 15% and
25% of annual per capita total household spending. Also shown is the concentration index for the incidence of catastrophic
payments. A positive index indicates catastrophic payments were more common among the better off, while a negative index
indicates they were more common among the poor. The data shown in Figure 3 are from van Doorslaer et al.23

III. HEALTH CARE DELIVERY AND UTILIZATION

How is the delivery of health care organized? What role do the public and private sectors
play, and how far are first- and second-level providers differentiated from one another? How are
providers paid, and how does this influence the organization and delivery of care? What reforms
have there been to provider payments and provider organization, and with what results? Who
uses health care and how much? These are the questions considered in this section.

Ownership and provider payments

Ambulatory care in the five systems is provided largely by the private sector (Table 2).
Differences emerge, however, in the hospital sector: the hospital sectors of Hong Kong (China)
and Singapore are both largely public, while in the three social insurance systems--Japan, Korea
and Taiwan (China)--hospitals are largely private. In terms of the private-public mix, then, the
five systems span all four possible combinations or ownership and finance. The first-level clinics
in Hong Kong (China) and Singapore are mainly privately owned and operated, and are mainly
privately financed, mostly out-of-pocket. The ambulatory care clinics in Japan, Korea and
Taiwan (China), like the hospitals in these three systems, are also largely privately operated, but
receive large amounts of public finance through UHI reimbursement. The hospital sector in
Hong Kong (China) is largely publicly owned and operated, but is financed largely (80%) out of
general revenues.* By contrast, the hospital sector in Singapore, which is also still largely
publicly owned and operated (more on this below), relies heavily (though increasingly less so25)
on private finance.

Table 2: Percentage of consultations, admissions, physicians and facilities in public sector

Doctors Beds per Ambulatory First-level Hospital Hospital Hospit
per 10,000 10,000 consultations physicians admissions beds % als %
persons persons % public % public % public public public
Hong Kong
(China), China 15.8 52 28% 45% 82% 86% 76%
Japan 20.2 129 36% 33% 20%
Korea, Rep. 19.4 66 10% 10% 10%
Singapore 15.0 28 20% 74% 81%
Taiwan (China),
China 14.0 59 3% 32% 15%
OECD average 28.0 60

The five systems all share a common feature, namely that their ambulatory facilities offer
a wider range of services than a typical ambulatory facility in the West. In Singapore, for
example, ambulatory clinics offer X-rays, clinical laboratory tests, and some rehabilitative
services.27 In Japan, many physician offices (one third of the total) even have a few outpatient
beds.17 And in Japan, as in Hong Kong (China)28, clinics offer specialist care, not just primary

* See Hong Kong (China)'s `domestic' health accounts at http://www.hwfb.gov.hk/statistics/en/dha.htm.
No national health accounts are available for Singapore26, and I have been unable to establish the share of hospital costs borne
by the taxpayer.
Data on doctors and beds from WHO http://www.wpro.who.int/information_sources/databases/core_indicators/, except Taiwan
(China) where the numbers are from Taiwan (China)'s statistical yearbook
http://ecommerce.taipeitimes.com/yearbook2004/P243.htm and the OECD averages which are from OECD Health Data1.

care--indeed, the notion of primary care in these systems is not one of general practice or family
medicine.28 At least until recently, many clinics also prescribe and dispense drugs (see below).

In part, the expanded scope of first-level clinics compared to Western clinics reflects
financial incentives. Fist-level clinics are paid fee-for-service (FFS) in all five systems, whether
directly by patients in the cases of Hong Kong (China) and Singapore, or by insurers (and
sometimes patients themselves) in the cases of Japan, Korea and Taiwan (China). In Hong Kong
(China) and Singapore, the fees are largely unregulated, and there is clearly scope for clinics to
offer services that might elsewhere be provided in a hospital setting. In Japan, where payments
are by the insurer (extra-billing is prohibited), first-level providers and hospitals have
traditionally been paid FFS. Moreover, the fee schedule is the same for physician offices and
hospitals, and the range of services first-level providers can claim reimbursement for is
apparently largely unrestricted.17 This is reinforced by the fact that physicians working at the
first level have no hospital admission rights, and as a result rarely refer patients to hospital. The
result is a markedly undifferentiated delivery system, where the distinctions between the first and
second levels are increasingly blurred. First-level clinics end up offering services typically
associated with a hospital, and hospitals end up setting up large outpatient clinics. In Korea, the
story is much the same--an undifferentiated structure caused by a common fee schedule for first-
and second-level facilities, and first-level providers having no hospital admission rights.6 These
two contributory factors in Korea's case are reinforced by the fact that, in contrast to Japan, the
benefit package in Korea is fairly limited (half of all out-of-pocket payments are payments in full
for uncovered services19) and clinics and hospitals compete with one another in a lucrative and
unregulated market for medical services left uncovered by UHI. It is speculated that providers
exploit the information asymmetry in the health care market and try to shift demand from
covered services whose prices are regulated to uncovered services whose prices are determined
entirely by the market.19 In Taiwan (China), in contrast to Japan and Korea, the fee schedule is
differentiated between different types of facilities, and indeed within the hospital sector a more
generous fee schedule applies to better-equipped hospitals. This, unsurprisingly, has resulted in
hospitals upgrading themselves in order to achieve higher status in the accreditation process.11

The setting of fees for covered services has a potentially large impact on which type of
services end up being delivered to whom, through their influence on patients' willingness and
ability to pay for different types of service, and on the profit margins providers can earn on them.
In China, through a well-intentioned policy of underpicing ambulatory and basic hospital care
and overpricing high-tech hospital care and drugs, the authorities inadvertently ended up
encouraging providers to shift demand to the latter and to under-provide the former.29 In Korea a
similar problem has been noted.19 Korea has a Caesarian section rate of 43%, compared to
Japan's 15% and the United Kingdom's 20%.6 Korea also has a stronger demand for residency
training in specialties with generous margins such as ophthalmology and dermatology.19 Japan,
by contrast, has evolved a highly successful fee schedule.17 The authorities undertake a biennual
review of each of more than 3000 items, rather than making across-the-board proportionate
changes in all. They selectively reduce the fees of procedures that show "inappropriate" volume
expansion. They deliberately price high-tech care below cost and basic ambulatory care above
cost--the exact opposite of China's current practice. They keep a tight control on fee increases,
and increasingly bundle items together. The results have been interesting. Fees for lab tests and
diagnostic imaging in Japan have fallen over the years. Despite the fact that high-tech care is
unprofitable in Japan, hospitals nonetheless do deliver it, perhaps because they feel that a

reputation for delivery of high-tech care helps attract all types of patients, and perhaps because
doctors find the delivery of high-tech care professionally rewarding.17 Japan's fee schedule is
often argued to be one of the reasons it has been able to keep health spending checked despite 40
years of UHI and FFS.17

Of course, FFS is far from ideal, and unsurprisingly shifts away from FFS have started in
all three social insurance systems. In 1993, Japan introduced an all-inclusive per diem
reimbursement rate for its geriatric hospitals which bundled drugs, lab tests, etc. together. Costs
of medications fell by one third, and the costs of lab tests fell to one tenth of their previous
level.17,18 In 1998, Japan piloted a DRGs in 10 hospitals with 183 DRGs, and the experiment is in
the process of being evaluated.18 In 1997, Korea launched a pilot DRG program for inpatient
care, covering nine disease categories (25 codes) which together accounted for 25% of inpatient
cases.6 The pilot, which was voluntary, built in generous margins over FFS to overcome provider
opposition. It also built in outlier payment features to reduce skimping. Some previously
uncovered services were bundled in the DRGs, so that the DRG pilot in effect involved an
expansion of insurance coverage. Comparing changes in reformed and unreformed hospitals, it
has been estimated that DRGs led to a 14% reduction in cost, a 6% reduction in average length
of stay, and a substitution towards outpatient care away from inpatient care. No adverse effects
on quality were found. In 2001, the pilot program was extended to all hospitals for the same
disease categories. Taiwan (China) has also taken steps to move away from FFS.11 A DRG
system was phased in for the 50 most common diseases. Comparing changes between reformed
and unreformed hospitals, studies for two of the covered procedures found that DRGs reduced
both cost per case and length of stay, but increased volume, so much so that total payments in
respect of the two procedures increased.29 Taiwan (China) is set to extend the use of DRGs in
2005, and is also experimenting with global budgets.

Singapore has also launched major supply-side reforms, notably in the corporatization of
its hospital sector.25 By making hospitals more autonomous, it was envisaged that exposure to
market forces would make them more responsive and efficient, and that they would improve
quality faster than would otherwise be the case. Rather than corporatizing individual hospitals,
the Singapore government created a single corporation, the Health Corporation of Singapore
(HCS), under whose control around one quarter of Singapore's hospitals were placed, accounting
for nearly half of Singapore's national bed stock. HCS, a holding company, is governed by
private law, but is wholly owned by the government and answerable to the Ministry of Health.
HCS can hire and fire staff, set terms of remuneration, and has considerable power to set prices.
It is the residual claimant, but this is muted by the fact that although after-tax surpluses can be
kept, the government continues to cover any deficit through a subsidy, albeit on the
understanding the subsidy will diminish over time.

There does not appear to have been a rigorous evaluation of the impacts of Singapore's
hospital corporatization program. Trends in various indicators for the corporatized hospitals have
been reported for the period after corporatization25, but what would like to see is whether key
outcomes changed differently before and after the reform between the corporatized and
uncorporatized hospitals. Some points do, however, emerge. Cost recovery grew in the
corporatized hospitals, from 15% of revenues before the reform to 55% in 2002. This was
achieved by raising prices. The government was forced, in the end, to impose limits on average
charges per patient day. One newly corporatized hospital took the opportunity of the reform to

eliminate C-class beds, whose (typically low-income) occupants receive similar medical care but
have only basic hotel facilities and pay only 80% of the cost. The costs of these beds are met in
part from profits earned on more expensive (B- and A-class) beds. Physician earnings increased
in the corporatized hospitals, and in order to stem the brain drain from unreformed hospitals to
reformed hospitals, the government was forced to allow unreformed hospitals to set their own
fees as well. The reforms precipitated considerable upward pressure on doctors' salaries
throughout the system, but especially on those of surgeons. Concerns have also been voiced
about a reform-induced scramble to acquire the latest technology, and dumping of patients on the
uncorporatized sector.

Separating prescribing and dispensing

As in China and several other developing countries is East Asia, doctors in Japan and the
Tiger economies have traditionally both prescribed and dispensed drugs, and drug sales have
traditionally been a major source of income for them.30 This is often argued to be a factor behind
the high rates of drug use in these systems. In the OECD countries other than Japan and Korea,
drug expenditures account for 16.8% of total health expenditures. In Japan the figure is 18.4%,
while in Korea the figure is 28.2%.1 The suspicion of over-use of drugs in such settings (a fact
rather than merely a suspicion in the case of China31) led to moves in some systems to separate
the functions of prescribing and dispensing. Korea introduced such a reform in 2000 against stiff
opposition from physicians, who went on strike and secured substantial concessions, including a
compensatory increase in medical fees (some rising by as much as 40%).32 Taiwan (China) has
started implementing a separation policy on an incremental basis, and has been the subject of a
study that compares changes over time across reformed and unreformed providers.30*

Several lessons emerge from Korea's and Taiwan (China)'s experiences with separating
prescribing and dispensing. First, there is a risk that that compensation for those who lose out
from the policy (doctors but potentially too pharmacists, who lose prescribing rights) is so
generous it offsets any cost-savings from reduced over-consumption of drugs. In Taiwan
(China), where consultation and dispensing fees were both raised (the latter were paid to
pharmacists after the reform and to physicians before), it is estimated that while more `rational'
prescribing resulting from the separation did lead to cost-savings, they were absorbed entirely by
the compensation costs.30 Second, there is a risk that health providers find ways--legal or not--
to continue to earn profits from dispensing. In Taiwan (China), clinics were allowed to hire
onsite pharmacists, and unsurprisingly many did so--the share of clinics with on-site
pharmacists jumped from zero to 60% following the separation policy.30 Since, these onsite
pharmacists were paid a salary by the provider, the provider's incentives remain unchanged.
Unsurprisingly, the separation policy had little or no impact on costs and prescribing behavior in
clinics with onsite pharmacists. Even if the hiring of an onsite pharmacist is illegal, providers
might be able to collude with dispensing pharmacists (some of whom may be former colleagues
and friends) and channel patients to them; they may receive financial or in-kind `tokens of
gratitude' for their efforts.32 A third lesson is that providers may substitute diagnostic tests and
examinations for drugs to maintain their incomes.30,32 In Taiwan (China), it does not seem they
did so30, and there is no evidence on the issue for Korea. The final point worth making is that

* A suggestion along these lines in early 2005 from a journalist in Singapore led to an outcry from Singapore's Medical
Association http://www.sma.org.sg/sma_news/3701/hobbit.pdf.

while the separation of prescribing and dispensing ensures the provider has no incentive to
prescribe unnecessary drugs, it does not him a positive incentive to engage in cost-consciousness
prescribing32, as does a policy where doctors (rather than patients) have to pay for some or all of
the costs of the drugs they prescribe out of their practice income.33,34

Who uses health services and how much?

How do the financing and delivery features of the systems play out in terms of utilization
rates and inequalities in service use?

Factoring in use of traditional Chinese medicine (TCM) and other traditional providers,
Japan and the Tigers have higher utilization rates of outpatient care than the OECD as a whole
(Table 3). Hong Kong (China), for example, has almost twice as many general practitioner visits
per year as the OECD as a whole. Inpatient care, by contrast, is less heavily used in Japan and
the Tigers. This likely reflects the aforementioned extra services offered by ambulatory facilities
in Japan and the Tiger economies.

What of inequalities in service use between the poor and better off? Outpatient visits in
Korea and Taiwan (China) are pro-poor, but pro-rich in Hong Kong (China). This may reflect the
reliance on out-of-pocket payments for outpatient care in Hong Kong (China), and the fact the
population is covered by public insurance for ambulatory as well as inpatient care in Korea and
Taiwan (China). In all three systems, inpatient care use is higher among the poor. It is least pro-
rich in Korea, perhaps reflecting the fact that many services--including some high-tech services
and the specialist surcharge--are not covered by the UHI package.19,29 Comparing inequalities in
the Tigers with inequalities in the OECD countries is not straightforward for ambulatory care,
because of different definitions. Western doctor visits are less pro-poor in the Tiger economies
than general practitioner (GP) visits are in the OECD, but are more pro-poor than specialist visits
are there. In none of the three Tiger economies is inpatient care as pro-poor as it is in a typical
OECD country.

Table 3: Use of services and inequalities in service use by income*

Traditional Chinese
Western doctors medicine Inpatient care
All western doctors General practitioners Specialists
Average Average Average Average No. per 100
per capita CI per capita CI per capita CI per capita CI persons CI
Hong Kong 6.99 0.0573 1.27 -0.1580
(China) 8.27 0.0241 2.55 0.0237 14.00 -0.0954
Japan 16.00 9.57
Korea, Rep. 9.76 -0.0234 1.38 -0.0321 7.00 -0.0734
Singapore 12.00
Taiwan (China) 8.44 -0.0534 2.02 0.0401 10.00 -0.1470
OECD average 7.15 3.62 -0.0904 1.62 0.0105 16.40 -0.1516

* Data are from Lu et al.35, except in the case of Japan and Singapore, where the numbers are from Imai18 and the World Bank's World Development Indicators respectively, and the
OECD averages, which were computed from the OECD's Health Statistics database in the cases of Western doctor visits and inpatient admissions, and Masseria and van Doorslaer36 in
the case of GP and specialist visits. OECD averages are based on 15-19 countries, depending on the variable. Figures for Hong Kong (China) refer to 2002, Japan's and Korea's to
1998, Singapore's to 1994, and those for Taiwan (China) to 2001. Japan's figure for Western physician consultations may include visits to traditional providers. The concentration
index measures the extent to which use is higher among higher income groups: a positive value indicates that, on average, the better off make greater use of services; a negative index
indicates the opposite.37 The OECD concentration indices are from Masseria and van Doorslaer.

IV. CONCLUSIONS

The differing health systems and reform experiences of Japan and the Asian Tigers
provide many learning opportunities for developing countries, especially those in East Asia who
face similar challenges today to those faced by Japan and the Tigers during the latter part of the
20th century.

In none of the five systems examined is there any compelling evidence of excessive or
runaway health expenditures. Indeed, the only serious outlier internationally is Singapore which
spends barely half of the share of GDP one would expect of a country with its per capita income.
This is despite the fact the five systems span the full spread of financing systems, ranging from
the tax-financed system in Hong Kong (China), through the social insurance systems of Japan,
Korea and Taiwan (China), to Singapore's largely privately financed system. One lesson, then,
seems to be that all types of financing system can be adapted to keep health spending in check--
there is no right or wrong financing mix from the point of view of macroeconomic expenditure
control.

Another lesson that emerges in respect of financing is the limited role in practice of
medical savings accounts. Despite the enthusiasm for them in some quarters, one cannot help but
be struck by the fact that they have never accounted for more than 10% of total spending in
Singapore. Even more striking is the fact that the complementary insurance and safety net
schemes (Medishield and Medifund) have played an even smaller role, accounting for just 2% of
total spending together. Singapore's system is more accurately classified as a largely out-of-
pocket system with limited tax subsidies.

In terms of who bears the burden of health financing, Japan and the Tigers offer similar
general messages to the OECD countries, but provide some useful pointers with respect to
specifics. The progressivity estimates reviewed above point, as in the OECD countries, to tax-
financed systems being the most progressive, and social insurance systems being somewhat
regressive or close to proportional. Beyond that, the lesson that emerges is that the precise degree
of progressivity of each source of finance depends on the design features of the system--the
emphasis on direct versus indirect taxes, the degree to which luxury goods are taxed at a higher
rate, the height of any ceilings on social insurance contributions, whether the poor are exempt
from social insurance contributions, the fraction of health spending covered by the benefit
package, the extent to which the poor face reduced copayments, and so on. Likewise the degree
to which households, especially poor ones, suffer from catastrophic payments also depends on
system features--whether there are caps on copayments, how large they are, and whether they
are lower for the poor. In terms of who bears the burden of health finance, therefore, just as
much may hinge on the design features of each source as on the mix between them.

On the delivery side, too, Japan and the Tigers offer some interesting lessons. While all
rely predominantly on the private sector for the delivery of ambulatory care, some opt for the
public sector to deliver hospital care while others favor the private sector. Singapore provides
some evidence on the consequences of making public hospitals more private without actually

privatizing them. Unfortunately, hard evidence on the impacts of Singapore's hospital
corporatization initiative is not available, but the (sometimes anecdotal and selective) evidence
that does exist points to a number of (largely predictable and mostly negative) consequences,
including rapid increases in costs and prices charged to patients, rapid growth in technology and
high-tech interventions (especially surgical procedures), and attempts by hospitals to dump high-
cost patients and to avoid offering services to low-income patients.

Japan and the other Tigers also offer some useful lessons on paying providers. Korea's
experience with a narrow benefit package provides an illustration of the danger of providers
shifting demand from insured services whose prices are regulated to uninsured services whose
prices are unregulated and highly lucrative. The social insurance systems also offer some useful
lessons vis-�-vis reimbursement of insured services. The experiences of Korea and Taiwan
(China) are less positive than that of Japan, which has managed to combine careful cost control
with fine-tuning of profit margins on different types of care. Experiences with DRGs in Korea
and Taiwan (China) have also proved useful, both pointing towards cost-savings, but in the case
of Taiwan (China) drawing attention to the possible knock-on effects on service volume and total
health spending. Korea and Taiwan (China) both offer important lessons for the separation of
prescribing and dispensing. For example, compensation costs in both cases were substantial and
in the case of Taiwan (China) were so large they offset the cost savings caused by more
`rational' prescribing. The risk of cost-savings never being realized because of other concessions
to providers--such as allowing them to have onsite pharmacists--is also illustrated by the pilot
in Taiwan (China).

It would be a mistake to view the health systems of Japan and the Tigers as `finished
business'--all are likely to carry on exploring ways of financing and delivering health care more
efficiently, if not also more equitably. But in the meantime, their existing arrangements and
recent reform experiences all provide potentially useful lessons for East Asia's developing
countries, if not for one another.

References

1. OECD Health Data [program]. Paris: OECD, 2005.

2. World Health Organization. World Health Report 2005: Make every mother and child count.
Geneva: World Health Organization, 2005.

3. Lu J-FR. The Application of National Health Account Approaches in Estimating National
Health Expenditures: Taiwan (China) Experience. Tao-Yuan: Department of Health Care
Management, Chang Gung University, 2001.

4. O'Donnell O, Van Doorslaer E, Rannan-Eliya R, et al. Who pays for health care in Asia?
EQUITAP Working Paper # 1, Erasmus University, Rotterdam and IPS, Colombo, 2005.

5. Lu JR, Hsiao WC. Does Universal Insurance Make Health Care Unaffordable? Lessons from
Taiwan (China). Health Affairs 2003;22(3 (May/June 2003)):77-88.

6. Kwon S. Payment system reform for health care providers in Korea. Health Policy Plan
2003;18(1):84-92.

7. Liu E, Yue S. Health care expenditure and financing in Singapore. Hong Kong (China):
Research and Library Services Division, Hong Kong (China) Legislative Council
Secretariat, 1999.

8. Hanvoravongchai P. Medical Savings Accounts: Lessons Learned from Limited International
Experience. Geneva: WHO, 2002.

9. Lim MK. Shifting the burden of health care finance: a case study of public-private partnership
in Singapore. Health Policy 2004;69(1):83-92.

10. Nichols L, Prescott N, Phua K. Medical savings accounts for developing countries. In:
Schieber G, ed. Innovations in Health Care Financing. Washington DC: World Bank,
1997.

11. Lu JFR. Social health insurance experience in Taiwan (China), 2005.

12. Kakwani NC. Measurement of tax progressivity: An international comparison. Economic
Journal 1977;87(345):71-80.

13. Wagstaff A, van Doorslaer E, Calonge S, et al. Equity in the finance of health care: some
international comparisons. J Health Econ 1992;11(4):361-87.

14. Wagstaff A, van Doorslaer E, van der Burg H, et al. Equity in the finance of health care:
some further international comparisons. J Health Econ 1999;18(3):263-90.

15. Kwon S. Health care financing reform and the new single payer system in the Republic of
Korea: social solidarity or efficiency? International Social Security Review
2003;56(1):75-94.

16. Ikegami N, Campbell JC. Medical care in Japan. N Engl J Med 1995;333(19):1295-9.

17. Ikegami N, Campbell JC. Health care reform in Japan: the virtues of muddling through.
Health Aff (Millwood) 1999;18(3):56-75.

18. Imai Y. Health care reform in Japan: OECD, 2002.

19. Kwon S. Billing of medical services and the financial burden of patients in Korea, 2005.

20. Wagstaff A, Van Doorslaer E. Progressivity, horizontal equity and reranking in health care
finance: a decomposition analysis for The Netherlands. Journal of Health Economics
1997;16:499-516.

21. Wagstaff A, van Doorslaer E. Catastrophe and impoverishment in paying for health care:
with applications to Vietnam 1993-98. Health Econ 2003;12(11).

22. Xu K, Evans DB, Kawabata K, Zeramdini R, Klavus J, Murray CJ. Household catastrophic
health expenditure: a multicountry analysis. Lancet 2003;362(9378):111-7.

23. Van Doorslaer E, O'Donnell O, Rannan-Eliya RP, et al. Paying out-of-pocket for health care
in Asia: Catastrophic and poverty impact. EQUITAP Working Paper #2, Erasmus
University, Rotterdam and IPS, Colombo, 2005.

24. Chou S-Y, Liu J-T, Hammitt JK. National Health Insurance and Precautionary Saving:
Evidence from Taiwan (China). Journal of Public Economics 2003;87(9-10):1873-94.

25. Phua KH. Attacking Hospital Performance on Two Fronts: Network Corporatization and
Financing Reforms in Singapore. In: Preker AS, Harding A, eds. Innovations in Health
Service Delivery: The Corporatization of Public Hospitals. Washington DC: World Bank,
2003.

26. Rannan-Eliya RP. Status of National Health Accounts in Asia-Pacific Region: Results of the
APNHAN Survey 2001. Colombo: Health Policy Programme, Institute of Policy Studies,
Sri Lanka, 2001.

27. Lim MK. Health care systems in transition. II. Singapore, Part I. An overview of health care
systems in Singapore. J Public Health Med 1998;20(1):16-22.

28. Leung GM, Wong IO, Chan WS, Choi S, Lo SV. The ecology of health care in Hong Kong
(China). Soc Sci Med 2005;61(3):577-90.

29. Eggleston K, Hsieh C. Health Care Payment Incentives in East Asia: Taiwan (China), South
Korea and the PRC in Comparative Perspective. Applied Health Economics and Health
Policy in press.

30. Chou YJ, Yip WC, Lee CH, Huang N, Sun YP, Chang HJ. Impact of separating drug
prescribing and dispensing on provider behaviour: Taiwan (China)'s experience. Health
Policy and Planning 2003;18(3):316-29.

31. Zhang X, Feng Z, Zhang L. Analysis on Quality of Prescription of Township Hospitals in
Poor Areas. Journal of Rural Health Service Management 2003;23(12):33-35.

32. Kwon S. Pharmaceutical reform and physician strikes in Korea: separation of drug
prescribing and dispensing. Soc Sci Med 2003;57(3):529-38.

33. Stewart-Brown S, Surender R, Bradlow J, Coulter A, Doll H. The effects of fundholding in
general practice on prescribing habits three years after introduction of the scheme. Bmj
1995;311(7019):1543-7.

34. Maynard A, Bloor K. Dilemmas in regulation of the market for pharmaceuticals. Health Aff
(Millwood) 2003;22(3):31-41.

35. Lu J-fR, Leung GM, Kwon S, Tin KY, Van Doorslaer E, O'Donnell O. Horizontal equity in
health care utilization: evidence from three high-income Asian economies, 2005.

36. Masseria C, van Doorslaer E. Income-Related Inequality in the Use of Medical Care in 21
OECD Countries. Paris: OECD, 2004.

37. Wagstaff A, van Doorslaer E, Paci P. Equity in the finance and delivery of health care: Some
tentative cross-country comparisons. Oxford Review of Economic Policy 1989;5:89-112.