Pay i ng Fo r Car e

Financing Health
Improvements In India
Resources should be focused tightly on interventions that are of
importance to the health and financial protection of the poor.
by Anil B. Deolalikar, Dean T. Jamison, Prabhat Jha, and Ramanan
Laxminarayan
ABSTRACT: India faces major challenges in sustaining the health gains achieved in the
better-performing states and ensuring that the lagging states catch up with the rest of the
country. In this paper we examine the current status of health financing in India, as well as
alternatives for realizing maximal health gains for the incremental spending. A principal
conclusion is that public expenditures of an additional US$6–US$7 per person per year
(about 1 percent of gross domestic product) would, if focused on about sixteen key inter-
ventions, provide universal access to those interventions and have a favorable affect on
population health. [Health Affairs 27, no. 4 (2008): 978–990; 10.1377/hlthaff.27.4.978]

I
n r e s p o n s e to t h e c h a l l e n g e o f sustaining the health gains in better-
performing states and ensuring that lagging states catch up with the rest of the
country, the Indian government has launched the National Rural Health Mis-
sion (NRHM). A central element of that effort is to increase public spending on
health from the current 1.1 percent of gross domestic product (GDP) to roughly
2–3 percent within five years. The NRHM focuses on rural areas, especially in the
eighteen states with weak health outcomes and infrastructure, including nine
particularly disadvantaged states referred to (using Government of India termi-
nology) as the Empowered Action Group, or EAG states.1
In this paper we examine the current status and future prospects of health fi-
nancing in India. Much has been written on this issue; we synthesize what is
known, draw attention to the benefits of public health spending, analyze why
public spending has improved health outcomes in some regions more than in oth-
ers, and apply lessons from the Disease Control Priorities Project–India (DCPP-

Anil Deolalikar is a professor of economics in the Department of Economics, University of California, Riverside.
Dean Jamison is the T & G Angelopoulos Visiting Professor of Public Health and International Development,
Kennedy School of Government and School of Public Health, Harvard University; he is located in Bethesda,
Maryland. Prabhat Jha is a professor in the Centre for Global Health Research and St. Michael’s Hospital, in
Toronto, Ontario. Ramanan Laxminarayan ([email protected]) is a senior fellow at Resources for the Future in
Washington, D.C., and a visiting scholar at the Princeton Environment Institute, Princeton University, in
Princeton, New Jersey.

978 J u l y /A u g u s t 2 0 0 8
DOI 10.1377/hlthaff.27.4.978 ©2008 Project HOPE–The People-to-People Health Foundation, Inc.
 Fi nanc i ng H e alth

India) so that the resources made available by the NRHM can be applied cost-
effectively.2
The large health and institutional disparities between EAG and non-EAG
states create different financing challenges. In EAG states, such as Bihar, Madhya
Pradesh, Orissa, Rajasthan, and Uttar Pradesh, which together account for 45 per-
cent of India’s population, the health challenge is primarily high levels of infant
and child mortality and child malnutrition. In non-EAG states, such as Kerala,
Tamil Nadu, and Gujarat, noncommunicable diseases are fast replacing infectious
diseases and malnutrition as the leading causes of morbidity and mortality. We fo-
cus on the EAG states.
Health systems have two broad objectives: to improve the level (and distribu-
tion) of health outcomes, and to provide people with financial protection, from
both unanticipated high health spending and income loss. They generally have
two types of resources: financial and system capacity. The latter refers to the peo-
ple, infrastructure, and procedures available at a point in time. Some elements of
capacity cannot be increased with financial resources in the short run, so capacity
constraints can exist alongside financial ones.
The DCPP-India has collected quantitative information on the financial costs of
achieving health gains using different interventions.3 Financial protection is typi-
cally discussed without a sense of budget constraints or the cost-effectiveness of
alternative instruments for acquiring it. Nonfinancial capacity constraints are of-
ten ignored. Here we attempt, in a preliminary way, to address both shortcomings.
We review India’s progress and challenges in health, summarize the current status
of state and central government health spending, evaluate criteria for an interven-
tion package for the NRHM, and discuss the mechanics of financing these health
interventions and the implications for center-state financial responsibilities.

Progress And Problems In Health

To many observers, the 1990s stood out for economic reforms, unprecedented
GDP growth, and marked improvement in living standards for many Indians.
However, the similarly remarkable gains in health and longevity are much less ap-
preciated. Over the course of the decade, under-five mortality dropped at an an-
nual rate of 2.2 percent, and life expectancy at birth increased from fifty-nine to
sixty-four years. Some have argued that when these improvements in health are
taken into consideration, the sharp differences between India and China in eco-
nomic performance during the 1990s are much less striking.4
However, child health improvements in India, although remarkable, are modest
compared with those in some of its neighbors in South Asia. India achieved a 2.2
percent annual decline in infant mortality, but Bangladesh achieved a 5.7 percent
decline over the same period, and Nepal, 4.2 percent.5 India’s reductions in child
malnutrition, maternal mortality, adult mortality, and prevalence of communica-
ble diseases are similarly lower than those of its neighbors.

H E A L T H A F F A I R S ~ Vo l u m e 2 7 , N u m b e r 4 979
Pay i ng Fo r Car e

Improvements in health have not been shared equally.6 Infant mortality de-
clined by 68 percent (from 37 to 12 per 1,000) in Kerala but by only 52 percent in
Uttar Pradesh (from 150 to 72) between 1981 and 2004. Uttar Pradesh alone con-
tributes one-quarter of all infant deaths in the country and, along with Madhya
Pradesh, Bihar, and Rajasthan, accounts for slightly more than half of the total.
During 1994–1999, just 10 percent of India’s 602 districts accounted for nearly 30
percent of all infant deaths, and 23 percent of the districts accounted for approxi-
mately half of the total.7 Furthermore, there are variations across social groups.
Scheduled Tribes have an under-five mortality rate of 117 deaths per 1,000 live births,
and Scheduled Castes have a rate of 108, compared with an all-India rate of 93.8
An important policy question is whether the tremendous growth of the Indian
economy will, by itself, improve health outcomes. Recent evidence suggests that
the association between health outcomes and per capita GDP, which was rela-
tively strong in the 1980s, has weakened.9 If economic growth alone is unlikely to
bring about major health improvements, there is a need for continuing—and even
greater—public action in health.10

Health Spending In India

At about $20 (in purchasing power parity, or PPP), India’s per capita public
spending on health—currently about 1.1 percent of GDP—is much lower than
that of most other countries with comparable per capita GDP. Private health
spending far exceeds the public spending. Indeed, India’s total health spending—
relative to its GDP—is higher than that of every other country in South Asia ex-
cept Afghanistan (but not higher than in other low-income countries outside
South Asia).
At about 80 percent of total health spending, India has among the highest
shares of private (household) spending in the world. Although association be-
tween the public share in health spending and per capita GDP is weak, the norm
for countries at India’s level is a private share of only 50 percent. Much of the bur-
den of out-of-pocket health spending falls on the quarter or third of households
with incomes below the poverty level. Increasing public spending is likely to
crowd out private spending on health, benefiting the poor.11
Increasing public spending on health can be justified only if it yields health
benefits. Comparisons across countries are notoriously difficult and fraught with
methodological problems. Nonetheless, infant mortality in India is lower (by
about 50 percent) than would be expected for a country with only PPP$20 per ca-
pita in public spending on health, which indicates, perhaps, that spending effec-
tiveness in India is better than average.
Three other observations are relevant. First, other low-income countries (such
as Bangladesh) have lower infant mortality rates than India despite their lower
public spending on health, which indicates that government health spending in
India could be more effective. Second, there is much heterogeneity in spending ef-

980 J u l y /A u g u s t 2 0 0 8
 Fi nanc i ng H e alth

fectiveness within India. Although public health spending effectiveness in the

EAG states may appear low, given their weak administrative capacity, poor gover-
nance, and service delivery failures, the marginal effect of an additional rupee of
public spending on health could be high. Third, the effectiveness of public health
spending also depends on the choice of health interventions, target population,
and technical efficiency. As more public resources are directed to health, the qual-
ity of this spending could increase, if directed appropriately.
n Financial implications. Achieving NRHM financing targets will require
much cooperation from state governments. They are responsible for three-quarters
of public spending, and many are in a perilous financial situation, with unsus-
tainably high current account deficits and debt ratios. It is not clear how they can fi-
nance increases in public spending on health. International financing for health has
been increasing but must remain negligible, given the size of the Indian economy.12
Before the economic reforms in the mid-1980s, public spending on health in In-
dia peaked at about 1.6 percent of GDP and 4 percent of the government budget.13
During the 1990s, government health spending failed to keep up with the expand-
ing economy, and by 2001 it constituted 0.9 percent of GDP and 2.7 percent of the
government budget. These numbers fell to 0.8 percent and 2.4 percent, respec-
tively, by 2005. The 2006–07 budget reversed this trend, greatly increasing alloca-
tions to social sectors (education, health, and woman and child development).
Most of the additional allocations to the health sector were channeled through the
NRHM. In the Common Minimum Program (CMP) of 2004, the present United
Progressive Alliance (UPA) government committed to spending 3 percent of GDP
on the health sector before its term expired in 2008–09. Assuming that the Indian
economy continues to grow at its current rate implies a commitment of Rs 1,500
billion to health. Budget estimates for 2006–07 call for central government health
spending of Rs 130 billion (including central grants to states) and state spending
of Rs 285 billion. The commitment to spend 3 percent of GDP means that even an-
nual increases of 30–50 percent in central government health spending over the
next few years will not meet the CMP goal. In short, the 3 percent spending goal is
unlikely to be realized by 2008–09.
n State-level variations in health spending. The National Health Accounts,
2001–02, show wide variations in per capita health spending across states (Exhibit
1). Private spending accounted for most of health spending in nearly every state and
at least 90 percent of health spending not just in poor states like Bihar and Uttar
Pradesh but also in relatively affluent states like Haryana, and in Kerala, which has
good public health facilities.
Exhibit 2 shows no pattern in the ratio of public spending on health to gross
state domestic product (GSDP). Public spending on health is low in relation to
state income (about 0.6–0.8 percent of GSDP) in relatively affluent states like
Haryana and Gujarat as well as in a poor state like Uttar Pradesh. Yet public
spending on health accounts for 1.6 percent of GSDP in Bihar, which is very poor.

H E A L T H A F F A I R S ~ Vo l u m e 2 7 , N u m b e r 4 981
Pay i ng Fo r Car e

EXHIBIT 1
Total (Public Plus Private) Per Capita Spending On Health In India, In Selected States,
2001–02
Assam (69% private)
Orissa (77% private)
Rajasthan (70% private)
Karnataka (71% private)
West Bengal (77% private)
Bihar (88% private)
Gujarat (82% private)
Tamil Nadu (76% private)
Madhya Pradesh (85% private)
Maharashtra (81% private)
Sikkim (20% private)
Andhra Pradesh (83% private)
Delhi (59% private)
Jammu and Kashmir (75% private)
Uttar Pradesh (93% private)
Himachal Pradesh (62% private)
Punjab (83% private)
Haryana (90% private)
Kerala (87% private)

0 500 1,000 1,500 2,000

Per capita health spending (rupees)
SOURCE: National Health Accounts, 2001–02, Ministry of Health and Family Welfare in collaboration with World Bank India
Country Office.

EXHIBIT 2
Public Spending On Health In India, Per Person, As A Percentage Of Gross State
Domestic Product, In Selected States, 2001–02
Himachal Pradesh
Jammu and Kashmir
Bihar
Assam
Rajasthan
Orissa
Kerala
Karnataka
Madhya Pradesh
West Bengal
Andhra Pradesh
Punjab
Tamil Nadu
Uttar Pradesh
Maharashtra
Gujarat
Haryana

0.0 0.5 1.0 1.5 2.0

Percent of gross state domestic product
SOURCE: National Health Accounts, 2001–02, Ministry of Health and Family Welfare in collaboration with World Bank India
Country Office.

982 J u l y /A u g u s t 2 0 0 8
 Fi nanc i ng H e alth

Private spending on health is greater in states that have relatively low public
spending on health.14 Because such spending is very low in India, households gen-
erally finance health expenses themselves. The high out-of-pocket spending on
health care is an important cause of impoverishment.15 Therefore, increasing pub-
lic spending on health could directly help the poor and enhance welfare.
n Benefits of increased availability of public resources. Empirical studies on
the link between public health spending and health outcomes are inconclusive.
Some have observed a strong inverse association between public spending on health
and infant mortality; others have found no significant effect.16 Still others have found
that the inverse association between child mortality and public spending on health
is stronger for the poor than for the nonpoor.17 An econometric study based on
twenty years of data found that public spending on health has larger effects in the
EAG states than in the non-EAG states, since the high-mortality EAG states have
very low coverage—and hence much room for expansion—of low-cost child sur-
vival interventions such as immunization and oral rehydration therapy.18 In the low-
mortality non-EAG states, these basic interventions are already widespread, and
more infant mortality occurs in the first month of life. This neonatal mortality re-
quires more-costly interventions, such as deliveries in medical institutions and
postdelivery and emergency hospital-based care.
Public spending on health can be more cost-effective than private spending. In
the largely unregulated private sector, especially in the rural areas of EAG states,
health workers have an incentive to prescribe more-costly treatments instead of
the standard, recommended drug regimens, since they both prescribe and sell
drugs, as in the case of TB treatment.19
The benefits of increased public spending on health are not restricted to health
outcomes; such spending protects both poor and nonpoor households from the fi-
nancial impact of health shocks. Income redistribution to those most likely to be
impoverished by out-of-pocket payments for health is important in a country
with such a high incidence of catastrophic health payments.20 Nearly 25 percent of
Indian households report spending 5 percent or more of their total annual outlays
on out-of-pocket health payments; for 10 percent of households, out-of-pocket
health payments exceed 10 percent of total outlays.21

What Services Should The NRHM Purchase?

An important question for the NRHM is how to use the additional resources
available to improve health in the eighteen EAG states. First, it may be instructive
to see how the NRHM’s resources have actually been allocated since it was
launched (Exhibit 3). In recent years, more than 40 percent of NRHM funding has
been in the category “Flexible Pool of State Project Implementation Plans”—un-
tied funds allocated to the states to implement the NRHM. The PULSE polio im-
munization program as well as rural family welfare (family planning) services
have each accounted for 14 percent of NRHM funds, while the seven major Na-

H E A L T H A F F A I R S ~ Vo l u m e 2 7 , N u m b e r 4 983
Pay i ng Fo r Car e

EXHIBIT 3
Expenditure On The National Rural Health Mission (NRHM) In India, By Activity,
2005–2008

Revised estimates Revised estimates Budget estimates

2005–06 2006–07 2007–08
Percent Percent Percent
Budget head Rupees of total Rupees of total Rupees of total
National Vector-Borne Disease
Control Program 309 5.1 353 4.9 368 3.7
National TB Control Program 166 2.7 207 2.9 249 2.5
National Leprosy Eradication Program 26 0.4 35 0.5 35 0.4
National Trachoma and Blindness
Control Program 86 1.4 126 1.8 126 1.3

National Iodine Deficiency Disorders

Control Program 11 0.2 14 0.2 24 0.2
National Integrated Disease
Surveillance Program 50 0.8 33 0.5 72 0.7
National Drug De-addiction Control
Program 0 0.0 0 0.0 10 0.1

Total—National Disease Control Programs 649 10.7 741 10.3 884 9.0

Direction and administration 228 3.8 153 2.1 246 2.5

Rural family welfare services (subcenters) 1,259 20.7 982 13.7 1,824 18.5
Urban family welfare services 122 2.0 72 1.0 120 1.2
Contraception 302 5.0 390 5.4 351 3.6

Reproductive and child health project 33 0.5 5 0.1 196 2.0

Routine immunization 156 2.6 266 3.7 301 3.1
PULSE polio immunization program 807 13.3 1,007 14.0 1,289 13.1
Information, Education, and
Communication (IEC) 122 2.0 143 2.0 156 1.6

Training institutions under states and

center 110 1.8 77 1.1 125 1.3
Research institutes 43 0.7 51 0.7 58 0.6
Flexible pool of state project
implementation plans (PIPs) 1,781 29.3 3,089 43.0 4,159 42.3
Area projects 378 6.2 179 2.5 50 0.5
Other family welfare schemes 85 1.4 35 0.5 81 0.8

Total NRHM 6,075 100.0 7,190 100.0 9,839 100.0

SOURCE: Ministry of Finance, Government of India, Union Budget 2005–06, 2006–07, and 2007–08.
NOTE: All rupees are in nominal terms.

tional Disease Control Programs have taken up 10 percent. The remaining 20 per-
cent is spread across several activities, including Information, Education, and
Communication (IEC); research; reproductive and child health; administration;
and, surprisingly, urban family welfare services.
The NRHM should focus on a smaller subset of interventions that can be fi-
nanced by the government and scaled up effectively, for several reasons.22 First,
public spending improves health typically through a few interventions. Although
the individual ranking by burden may differ slightly, similar health conditions be-
set both EAG and non-EAG states. Second, formulating a minimum package can
be more effective than paying for a large range of interventions without regard to
joint costs or shared inputs.23 Such packages have other advantages: simplifying

984 J u l y /A u g u s t 2 0 0 8
 Fi nanc i ng H e alth

planning of new investments in building and staffing, defining minimum neces-

sary inputs, estimating financial needs, and establishing boundaries so that citi-
zens know exactly what the government will pay for. This last concern arises be-
cause bribes are often necessary to elicit services, and the poor do not know what
free services they are entitled to. Specifying a minimum package also simplifies the
identification of bottlenecks to expanding access. Finally, delivering many inter-
ventions can disrupt fragile delivery systems, whereas success with the initial se-
lection of interventions could expand institutional capacity.
Minimum packages have been used effectively in Tanzania, Mexico, and other
countries.24 The interventions that the NRHM should provide may be influenced
by political and social considerations but should also be guided by epidemiologi-
cal and economic rationales. The NRHM should focus on interventions that gen-
erate maximum levels of health gain and financial protection, especially where
these objectives can be jointly realized. A benefit-incidence analysis found that
wealthier populations had better access to high-quality care because they have
greater bargaining power and easier geographical access to medical facilities.25
Since health care is a normal good, the rich typically consume more of it—another
reason why public subsidies for health are generally skewed toward them.26
A more equitable distribution of subsidies would both increase spending to im-
prove quality (and hence use) of primary health care facilities, which tend to be
predominantly used by lower-income groups, and ensure greater geographical ac-
cess.27 More-equitable distribution of outcomes requires explicit universalization
of services that are important to the poor. The favorable distributional conse-
quences of not targeting the poor per se but rather providing universal access to
services important to the poor (and primarily used by them) are highlighted by In-
dia’s apparently successful universalization of elementary education.
The interventions to address the greatest burden of illness at lowest cost in
EAG and non-EAG states include institutionalization of births to prevent mater-
nal mortality (largely by hemorrhaging); expanded immunization coverage for
newborns; access to oral rehydration salts, rotavirus vaccination, and vitamin A
supplementation; case management of acute respiratory infections for children
under age five; expanded coverage of the directly observed treatment, short-
course (DOTS) strategy for TB; promotion of condom use and voluntary counsel-
ing and testing to prevent transmission of HIV; use of human papillomavirus
(HPV) vaccine and cervical cancer screening; spraying and vector control for dis-
eases such as malaria, dengue, and chikungunya; and taxation and regulation of
tobacco through bans on advertising and promotion and use of prominent warn-
ing labels (which are self-financing). Our estimate of the annual cost of this pack-
age, including spending to upgrade the human and infrastructural capacity to de-
liver these interventions, is $6–$7 per person.28 The annual avertable burden
associated with this spending is fifty-eight million DALYs. They represent about
23 percent of the total disease burden in India.

H E A L T H A F F A I R S ~ Vo l u m e 2 7 , N u m b e r 4 985
Pay i ng Fo r Car e

The Mechanics Of Financing Health

n Central-state issues. Although states have most of the responsibility for
health interventions and spending, the central government is important, too, espe-
cially in prevention programs. There is also emphasis on devolving health services
delivery to elected local governments, called panchayati raj institutions (PRIs).29
The central government accounts for about a quarter of all public spending on
health, with most of the remainder from state governments. Central-state propor-
tions vary greatly, with central government grants accounting for just 14 percent
of total government health (revenue) spending in West Bengal but 23 percent in
Andhra Pradesh in 1999–2000 (Exhibit 4). Overall, the central share of govern-
ment health spending between 1991–1992 and 1999–2000 remained constant.
Although the central government’s role in overall health spending is modest, it
dominates financing of public health and family welfare activities (Exhibit 4) and
centrally sponsored communicable disease control programs (such as the Na-
tional Vector-Borne Disease, Leprosy Eradication, TB Control, and AIDS/STD
Control Programs).
n Outcome-oriented financing. Most health spending by states is for medical
services—between 50 and 75 percent for curative services in hospitals in 1999–
2000.30 Some of this spending supports facilities providing primary care, but most
goes for urban health care, teaching hospitals, administration, and family planning.31
Health outcomes, especially infant mortality, respond more to public health
and local clinical interventions than to hospital care.32 Evidence from several
countries, including India, indicates that immunizations and preventive care help
the poor much more than hospital-based curative services do.33 An important
policy challenge is to shift the responsibility for public health and family welfare
funding from the central to state governments. Central government matching

EXHIBIT 4
Share Of Central Grants In Total Government (Revenue) Spending On Health, By
Function, In Selected States, India, 1991–1992 And 1999–2000
Percent 1991–1992 1999–2000 Andhra Pradesh
60 Assam
Orissa
45 Rajasthan
Tamil Nadu
30 West Bengal

15

0
Public health and Total government Public health and Total government
family welfare health spending family welfare health spending

SOURCE: Various state budget demand documents for 1991–1992 and 1999–2000.

986 J u l y /A u g u s t 2 0 0 8
 Fi nanc i ng H e alth

grants could induce states to spend more of their own resources on those activi-
ties.34 Although matching grants could create incentives for states to increase their
absolute and relative spending on communicable disease control and preventive
activities, unless carefully designed, they could disadvantage the poor high-
mortality states that cannot greatly increase their public health spending.
“Outcome budgeting” is another way to motivate states to perform better and
convert financial outlays into outcomes with measurable targets. The idea—revo-
lutionary by Government of India standards—is for ministries and state govern-
ments to be accountable for their budget allocations and to justify them by their
performance. Finance minister P. Chidambaram presented the first outcome-
contingent budget in Parliament on 25 August 2005. The Registrar-General of In-
dia can already measure health outcomes objectively and can build on plans for
district-level surveys and the ongoing Million Death Study.35

Discussion And Conclusions

In this concluding section, we draw together several important findings.
n Focus on a few selected interventions. The NRHM should focus on a small
set of highly cost-effective interventions. This facilitates scaling up and learning by
doing, improving the quality and effectiveness of care. Focusing on priority inter-
ventions permits universal access in targeted states or districts, with important po-
litical benefits. The DCPP-India provides a rigorous assessment of interventions to
include in an initial health entitlement program.36 We estimate the cost at US$6–$7
per capita per year or, for an all-India program, about 1 percent of GDP. Although
substantial, this is less than is envisioned for the NRHM.
n Focus more sharply on states and districts of greatest need. The NRHM
proposes to focus on eighteen states. This might be too many, and some states might
not belong on this list. The initial phase could target just the four states (Uttar
Pradesh, Madhya Pradesh, Bihar, and Rajasthan) that together account for more
than half of India’s infant deaths. Likewise, it is not necessary to target NRHM inter-
ventions to all districts in these states from the outset; universal coverage can be
phased in. A program that initially focuses on just the 100 districts with the largest
absolute number of infant deaths would be more effective than one spread thinly
over all 602 districts. Rational targeting of the NRHM may be politically challeng-
ing, but policymakers should beware of spreading resources too thinly.
n Introduce incentives to motivate state governments to achieve desired
outcomes. Given their lead role in financing health, the states—particularly the
EAG states—must increase their health spending concomitantly with the NRHM.
Creative incentives are necessary, both matching grants and performance- and out-
come-based grants from the central government. This is particularly relevant given
the government’s stated shift toward outcome budgeting. The Sarva Shiksha
Abhiyan (SSA) education project provides a valuable precedent for increasing state
budgetary commitments: the incentive came from increased commitments by the

H E A L T H A F F A I R S ~ Vo l u m e 2 7 , N u m b e r 4 987
Pay i ng Fo r Car e

Government of India (and external donors) and, hence, a substantial increase in its
share in financing elementary education. Recognizing the states’ limited fiscal ca-
pacity, the center provided the bulk of incremental education funding in SSA. For
the poorer states, the percentage from the center will need to be higher still, and per-
formance-based grants may also require technical assistance from the center.
n Use centrally sponsored schemes where they are now working. Although
funding for many NRHM interventions, particularly clinical interventions, will
come from states (or flow through states from federal matching grants), several na-
tional programs already deliver immunizations, TB treatment, and AIDS prevention.
For these interventions, existing organizational and financial structures should re-
ceive additional funding and become the vehicles for NRHM implementation. To-
bacco control via higher central taxation and regulation can also avoid strain on
state health infrastructures.37
n Improve information availability and local oversight. An important way to
improve public-sector delivery performance is to generate and distribute timely, lo-
cally specific information on financial transfers, effective coverage of interventions,
and health outcomes. The Registrar-General is developing capacity for this. PRIs
and service providers should be encouraged to use this information to increase ac-
countability for performance and to monitor progress in improving it.
n Explore options for improving the quality of publicly financed services.
Increasing public-sector spending on health in India is crucial, but so is improving
the quality of additional spending. The NRHM should focus on the reason for the
low use of primary health facilities in the EAG states—such as the unavailability of
medicines and other inputs and poor incentives for personnel to report for work and
be held accountable for performance. Although some problems can respond to in-
creased finance, others must be addressed by specific capacity-building efforts, in-
cluding, initially, greater central government involvement in implementation in
weaker states. It is important to bear in mind that public finance of services need
not entail public delivery. The private and NGO sectors may each have an important
role in delivering some NRHM-financed services.
To conclude, we recommend focusing resources tightly on interventions of im-
portance to the health and financial protection of the poor. Initially, in the poorest
states and districts, all of these services would be publicly financed, because se-
lecting the poor for fee exemptions is difficult or very costly, whereas broad access
engenders political support.

Preparation of this paper was supported by the Health Financing Task Force at the Brookings Institution and the
Disease Control Priorities Project–India. In revising this paper, the authors benefited from comments by
participants at a workshop in New Delhi on 8 January 2007 (chaired by Rajiv Misra) and from comments by Anna
Abreu, Peter Berman, David de Ferranti, Meenakshi Datta Ghosh, Amanda Glassman, Charles Griffin, Satish
Jha, Rajiv Misra, William Savedoff, Julian Schweitzer, Paramita Sudharto, and Viroj Tangcharoensathien. Paolo
Belli provided detailed input, and Jeffrey Chow and Sarah Darley provided valuable research support.

988 J u l y /A u g u s t 2 0 0 8
 Fi nanc i ng H e alth

NOTES
1. Empowered Action Group (EAG) states designated by the Government of India are Bihar, Chattisgarh,
Jharkhand, Madhya Pradesh, Orissa, Rajasthan, Uttar Pradesh, and Uttaranchal. The eighteen focus states
include the EAG states as well as Arunachal Pradesh, Assam, Himachal Pradesh, Jammu and Kashmir,
Manipur, Mizoram, Meghalaya, Nagaland, Sikkim, and Tripura.
2. The Disease Control Priorities Project–India follows on related but more global efforts. See D.T. Jamison et
al., eds., Disease Control Priorities in Developing Countries, 2d ed. (Oxford and New York: Oxford University
Press, 2006). See also R. Laxminarayan et al., “Advancement of Global Health: Key Messages from the Dis-
ease Control Priorities Project,” Lancet 367, no. 9517 (2006): 1193–1208.
3. The DCPP-India report, “Choosing Health: An Opportunity for India,” preliminary materials from a work-
shop in January 2007, is available from the Centre for Global Health Research at http://www.cghr.org/
dcpp.htm (accessed 20 March 2008).
4. D.E. Bloom et al., “Why Has China’s Economy Taken Off Faster than India’s?” Working Paper (Boston:
Harvard School of Public Health, 2006).
5. In Sri Lanka, infant mortality declined from 18 deaths per 1,000 live births in 1991 to 13 in 2001—a rate of
decline that exceeded the rates experienced by both Bangladesh and Nepal.
6. P. Jha, “Avoidable Mortality in India: Past Progress and Future Prospects,” National Medical Journal of India 15,
no. 1 Supp. (2002): 32–36.
7. Note that these data are merely indicative, since the NFHS data are not representative at the level of districts.
8. Scheduled Castes (SCs) and Scheduled Tribes (STs) are Indian communities that are explicitly recognized
by the Constitution of India as requiring special support to overcome centuries of discrimination.
9. A.B. Deolalikar, Attaining the Millennium Development Goals in India: Reducing Infant Mortality, Child Malnutrition,
Gender Disparities, and Hunger-Poverty and Increasing School Enrollment and Completion? (New Delhi: Oxford Uni-
versity Press, 2005).
10. T.W. Croghan, “Road Less Traveled: Four Developing Countries Blaze New Trails to Better Health,” RAND
Review 30, no. 3 (2006): 26–30.
11. Of course, public spending on health will reduce private spending by the poor only if it actually improves
service delivery to the poor. Also, it is assumed that public spending is not financed out of taxes, which fall
disproportionately on the poor.
12. Note that external assistance does constitute a sizable share of national disease control programs for TB,
HIV/AIDS, and malaria.
13. By “public spending on health,” we mean spending by the central and state governments on hospitals and
dispensaries; medical education, training, and research; public health; family welfare (family planning);
and the National Rural Health Mission (NRHM). It excludes government spending on the health care of
its own employees, including those in Railways and Defense.
14. Deolalikar, Attaining the Millennium Development Goals.
15. Nearly 25 percent of Indian households report spending 5 percent or more of their total annual expendi-
tures on out-of-pocket health payments, representing among the highest incidences of catastrophic health
payments in Asia. See E. van Doorslaer et al., “Effect of Payments for Health Care on Poverty Estimates in
Eleven Countries in Asia: An Analysis of Household Survey Data,” Lancet 368, no. 9544 (2006): 1357–1364.
16. See, for example, S. Anand and M. Ravallion, “Human Development in Poor Countries: On the Role of Pri-
vate Incomes and Public Services,” Journal of Economic Perspectives 7, no. 1 (1993): 113–150; B. Bidani and M.
Ravallion, Decomposing Social Indicators Using Distributional Data (Washington: World Bank, 1995); D. Filmer
and L. Pritchett, “The Impact of Public Spending on Health: Does Money Matter?” Social Science and Medicine
49, no. 10 (1999): 1309–1323; and V. Swaroop and A.S. Rajkumar, Public Spending and Outcomes: Does Governance
Matter? (Washington: World Bank, 2002). Rajkumar and Swaroop find that public spending on health is
significantly associated with child and infant mortality reduction, but only for countries with good gover-
nance (as measured by a corruption index and a variable reflecting the “quality of the bureaucracy”).
17. S. Gupta et al., Does Higher Government Spending Buy Better Results in Education and Health Care? (Washington: In-
ternational Monetary Fund, 1999). These authors find that the association between child deaths and pub-
lic spending on health is twice as strong for the poor as for the nonpoor.
18. R. Duggal, Cost of Health Care: A Household Survey in an Indian District (Bombay: Foundation for Research in
Community Health, 1989).
19. M. Uplekar, “Implications of Prescribing Patterns in Private Doctors in the Treatment of Pulmonary Tu-

H E A L T H A F F A I R S ~ Vo l u m e 2 7 , N u m b e r 4 989
Pay i ng Fo r Car e

berculosis in Bombay, India,” Research Paper no. 41 (Cambridge, Mass.: Takemi Program in International
Health, Harvard School of Public Health, 1989); Deolalikar, Attaining the Millennium Development Goals; K. Van
der Veen, “Government and Private Health Care: Two Competing Institutions,” in Managing Rural Develop-
ment: Health and Energy Programmes in India, ed. H. Streefkerk and T.K. Moulik (New Delhi: Sage Publications,
1991); and R. Bhat, “The Private/Public Mix in Health Care in India,” Health Policy and Planning 8, no. 1 (1993):
43–56.
20. Van Doorslaer et al., “Effect of Payments for Health Care on Poverty Estimates.”
21. Ibid.
22. DCPP-India, “Choosing Health.”
23. J.L. Bobadilla et al., “Design, Content, and Financing of an Essential National Package of Health Services,”
Bulletin of the World Health Organization 72, no. 4 (1994): 653–662.
24. J.F. Finlay et al., “A New Canadian Health Care Initiative in Tanzania,” Canadian Medical Association Journal
153, no. 8 (1995): 1081–1085; and J.L. Bobadilla et al., “National Package of Health Services,” in Macroeconom-
ics and Health: Investing in Health for Economic Development, ed. J. Sachs (Geneva: World Health Organization,
2001).
25. A. Mahal et al., “Who ‘Benefits’ from Public Sector Health Spending in India? Results of a Benefit Inci-
dence Analysis for India,” Background Paper prepared for D.H. Peters et al., Better Health Systems for India’s
Poor: Findings, Analysis, and Options (Washington: World Bank Human Development Network, 2002).
26. A. Gumber, “Burden of Disease and Cost of Ill Health in India: Setting Priorities for Health Interventions
during the Ninth Plan,” Margin 29, no. 2 (1997): 133–172.
27. The related issue of providing incentives to health care workers to work in remote areas is addressed in an-
other paper in this series.
28. A.B. Deolalikar, D.T. Jamison, and R. Laxminarayan, “India’s Health Initiative: Financing Issues and Op-
tions,” Discussion Paper no. 07-48 (Washington: Resources for the Future, October 2007).
29. There has been considerable devolution of health service delivery to panchayati raj institutions (PRIs) in
Karnataka and Kerala. In Kerala, local health workers are not only employed by local governments but also
accountable to them. In other states, PRIs receive funds to pay district- and lower-level health workers’
salaries but have little control in their hiring or termination.
30. These data are only broadly indicative, since the “public health and family welfare” category includes sev-
eral expenditures (for example, health education and drug control) unrelated to public health and ex-
cludes expenditures on water and sanitation by the PRIs. In some states, health workers’ salaries account
for nearly 90 percent of government health expenditure.
31. Deolalikar, Attaining the Millennium Development Goals; and R. Duggal, “Healthcare in India: Changing the Fi-
nancing Strategy,” Social Policy and Administration 41, no. 4 (2007): 386–394.
32. DCPP-India, “Choosing Health.”
33. Mahal et al., Who ‘Benefits’ from Public Sector Health Spending in India?”; and A. Mahal et al., “The Poor
and Health Service Use in India,” Background Paper prepared for Peters et al., Better Health Systems for India’s
Poor.
34. Some of the disease control programs are partially funded on matching grant basis, but most are fully
funded by the central government. A useful example is the Local Government Performance Program
(LGPP) in the Philippines. See M. Palabrica-Costello, N. Ogena, and A.N. Herrin, “Policy and Program Im-
plications of the Matching Grants Program in the Philippines” (Manila: Population Council, 2003).
35. Registrar-General of India and Centre for Global Health Research, “Causes of Death in India: Special Sur-
vey of Deaths 2001–3” (New Delhi: Registrar-General, 2007); and P. Jha et al., “Prospective Study of One
Million Deaths in India: Rationale, Design, and Validation Results,” PLoS Medicine 3, no. 2 (2006): e18.
36. J. Chow, S. Darley, and R. Laxminarayan, “Cost-Effectiveness of Disease Interventions in India,” Discussion
Paper no. 07-53 (Washington: Resources for the Future, December 2007).
37. R. John, “Price Elasticity Estimates for Tobacco Products in India,” Health Policy and Planning 23, no. 3 (2008):
200–209.

990 J u l y /A u g u s t 2 0 0 8