Effects of Social Policy On Domestic Demand

Effects of Social Policy
on Domestic Demand
Annual Conference 2009
Edited by Masahiro Kawai and Gloria O. Pasadilla

Kasumigaseki Building 8F
3-2-5 Kasumigaseki, Chiyoda-ku
Tokyo 100-6008, Japan
www.adbi.org
©2010 Asian Development Bank Institute
ADBI Publishing
adbipubs@adbi.org
ISBN: 978-4-89974-035-3
Freely available electronically at:

http://www.adbi.org/effects.social.policy.annual.conference.2009/
The views expressed in this work are the views of the authors and do not necessarily reflect the views or
policies of the Asian Development Bank Institute (ADBI), the Asian Development Bank (ADB), its board of
directors, or the governments they represent. ADBI does not guarantee the accuracy of the data included
in this work and accepts no responsibility for any consequences of their use. Terminology used may not
necessarily be consistent with ADB official terms.
Contents
Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Presenters and Discussants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii
I Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Masahiro Kawai and Gloria O. Pasadilla
1
II Keynote Address . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Haruhiko Kuroda
25
III Social Protection in Developing Asia and the Pacific . . . . . . . . . . . . . . . . .

Bart Édes
29
IV Household Savings Rates and Social Benefit Ratios:

Country Comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Charles Yuji Horioka and Ting Yin
Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
Mukul Asher
V Influence of Age Structure on Savings and Social Spending . . . . . .

Davide Furceri and Annabelle Mourougane
81
Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Hyungpyo Moon
VI Household Savings and Social Protection Policies in the PRC . . . .

Ming Yan and Pan Yi
109
Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Yasuyuki Sawada
VII Consumption, Income Distribution, and

State Ownership in the PRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Yuqing Xing
Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
Yvonne Sin
VIII Public Expenditures on Social Programs and Household

Consumption in the PRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Emanuele Baldacci, Giovanni Callegari, David Coady, Ding Ding,
Manmohan Kumar, Pietro Tommasino, and Jaejoon Woo
Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
Akiko Terada-Hagiwara
IX Social Policy Reforms and Growth Rebalancing in ASEAN . . . . . . . .

Gloria O. Pasadilla and Prayoga Wiradisuria
205
Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
Hiroshi Yamabana
X Panel Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

Foreword v
Foreword
The role of social policy in improving domestic demand will remain a relevant
issue in Asia for many years. The issue was highlighted at the ADBI’s 12th
Annual Conference. I am delighted to see the important contributions of
the conference now compiled in this volume. As I noted in my address at
the conference, the global financial and economic crisis has forced Asia to
re-examine its development strategy and explore sources of growth from both
domestic and regional demand. Moreover, enhanced social policies are an
important centerpiece to achieve a socially inclusive Asia and Pacific region.
ADB is committed more than ever to supporting regional economic
integration as a way to boost regional demand and to lessen the region’s
dependence on Western economies for growth. It is high time for the Asia and
Pacific region to be a locomotive of global growth instead of being highly reliant
on external demand, particularly from American and European markets. For
domestic demand growth—especially in countries with large current account
surpluses and domestic savings—improved social protection policies will play
a vital role in reducing excessive savings caused by precautionary motives.
As the chapters in this volume indicate, improved social policies have value
beyond their role in stimulating domestic demand. At ADB, enhanced social
protection policies are inextricably linked with our inclusive growth agenda.
Indeed, even before the crisis, ADB has long been involved in strengthening
social protection programs in developing member countries. But the global
economic and financial crisis has made our organization more acutely aware
that exposure to various sources of risk, such as economic shocks and natural
disasters, can leave a profoundly damaging impact on economic well-being
and human development when institutional structures such as those for social
protection are weak. Thus, ADB is even more resolute in its belief that there
is a need to strengthen institutional structures to help reduce vulnerability
and respond effectively to various risks.
The chapters in this volume are a welcome addition to the current discussions
on how to help countries in Asia and the Pacific improve their social protection
policies and make growth more inclusive.
Haruhiko Kuroda
President, Asian Development Bank
Preface vii
Preface
In 2009, ADBI organized a series of conferences on the global financial

and economic crisis that analyzed the crisis’ impact on Asian economies;
developed scenarios for growth rebalancing away from exports to developed
countries; and articulated recommendations on macroeconomic policy,
regional economic integration, and financial and regulatory reform. ADBI’s
12th Annual Conference pursued the same objective of formulating appropriate
long-term responses to the global financial crisis, focusing on social policies
and their impact on domestic demand. The conference examined the extent to
which improved social protection policies can help reduce household savings
in Asian countries with very high savings rates and induce a long-term increase
in consumption spending. More specifically, the annual conference tackled
these questions:
t 8IBU BSF UIF UIFPSFUJDBM BOE FNQJSJDBM SFMBUJPOTIJQT CFUXFFO TPDJBM
protection, household savings, and income?
t 8IBUJTUIFFYQFSJFODFPGTPDJBMQSPUFDUJPOQPMJDJFTJOEFWFMPQFEBOE
emerging economies?
t )PXEPGBDUPSTMJLFEFNPHSBQIJDEFWFMPQNFOUTBOEJODPNFEJTUSJCVUJPO
between labor and capital affect savings and demand?
t 8IBUJTUIFJNQBDUPGJODSFBTFETPDJBMQSPUFDUJPOTQFOEJOHPOIPVTFIPME
savings and consumption in Asian countries?
On behalf of ADBI, I would like to express my gratitude to the distinguished
paper writers, moderators, and discussants who shared their insights on this
important topic, as well as to the conference attendees for their very stimulating
discussions and contributions.
Masahiro Kawai
Dean, Asian Development Bank Institute
viii Effects of Social Policy on Domestic Demand
Presenters and Discussants
Mukul G. Asher is a professor of applied public sector economics and economic

reasoning for public policy at the National University of Singapore. He was educated
in India and the United States (US) and is regarded as the leading authority on social
security arrangements in Southeast Asia. He has authored or edited several books
as well as published numerous articles in national and international journals. Mr.
Asher has been a consultant to the World Bank, International Monetary Fund (IMF),
ADB, United Nations Economic and Social Commission for Asia and the Pacific
(UNESCAP), and Oxford Analytica, and he was a visiting professor at the Fiscal
Affairs Department of the IMF. He has served as a resource for policymakers in India,
Indonesia, Viet Nam, People’s Republic of China (PRC), and Sri Lanka.
David Coady is deputy division chief of the Expenditure Policy Division, Fiscal Affairs
Department, IMF. Previously, he was a senior economist and technical assistance
advisor in the IMF’s Fiscal Affairs Department. He holds a bachelor of commerce
degree and a master of economic science degree from University College Dublin.
He also earned a PhD in economics from the London School of Economics in 1992.
Mr. Coady has written numerous papers and books on social issues and his teaching
experience includes development economics, public economics, cost-benefit analysis,
and environmental economics. He has also acted as referee for many academic journals.
He has served several institutions, including the World Bank, as a consultant.
Bart W. Édes is director of ADB’s Poverty Reduction, Gender, and Social Development
Division. Previously, Mr. Édes oversaw communications for SIGMA, the Paris-based
joint initiative of the European Union and the Organisation for Economic Co-operation
and Development (OECD), assisting central and eastern European countries with
public governance reform. Mr. Édes has a bachelor’s degree in government from
Georgetown University and a master’s degree in public policy from the University
of Michigan.
Davide Furceri is an economist in the Macroeconomic Analysis Division of the

OECD. Prior to joining the OECD, he was an economist in the Fiscal Policies Division
of the European Central Bank. He graduated from the University of Palermo in
2001, where his thesis was recognized as the best Italian thesis in economics. In
2007, he earned a PhD in economics from the University of Illinois, where he
received a prize for the best thesis in the social sciences. Mr. Furceri specializes in
macroeconomics and fiscal policy and his works have been published extensively
in established international journals.
Presenters and Discussants ix
Charles Yuji Horioka is a professor of economics at the Institute of Social and

Economic Research at Osaka University and a research associate at the National Bureau
of Economic Research and at the Center for Japan-US Business and Economic Studies,
Leonard N. Stern School of Business at New York University. He received his bachelor’s
degree and PhD from Harvard University. He has taught at Stanford, Columbia, and
Kyoto universities. In 2001, Mr. Horioka won the 7th Japanese Economic Association/
Nakahara Prize, awarded annually to the most outstanding Japanese economist
aged 45 or younger. He specializes in macroeconomics and the Japanese economy
and has written numerous scholarly articles on household saving, consumption,
bequest, and co-residence behavior and on parent-child relations in Japan, the US,
and the PRC.
Masahiro Kawai is dean and chief executive officer of ADBI. He was an associate
professor in economics at Johns Hopkins University and professor of economics at the
University of Tokyo. He served as chief economist for the World Bank’s East Asia and
the Pacific Region, deputy vice minister of finance for international affairs in Japan’s
Ministry of Finance, and special advisor to the ADB president, in charge of regional
economic cooperation and integration. He has written books and numerous academic
articles on international economics, economic globalization and regionalization, and
regional financial integration and cooperation in East Asia. He holds a bachelor’s
degree in economics from the University of Tokyo, a master of science in statistics
and PhD in economics from Stanford University.
Haruhiko Kuroda has been ADB’s president since 2005. Previously, he was special
advisor to the cabinet of the Japanese prime minister and a professor at the Graduate
School of Economics at Hitotsubashi University. He represented Japan’s Ministry of
Finance at a number of international monetary conferences as vice minister of finance
for international affairs. Under his leadership, Japan supported Asian economies hit
by the 1997–1998 financial crisis and helped Asian nations establish a network of
currency swap agreements to avert another crisis. He holds a bachelor’s degree in
law from the University of Tokyo and an MPhil in economics from the University
of Oxford.
Hyungpyo Moon is a senior research fellow and managing director of the Economic
Information and Education Center of the Korea Development Institute. He was a
visiting scholar at the University of California at Berkeley and assistant secretary of
the Health and Welfare Division in the Office of the President of the Republic of Korea.
He earned a PhD in economics from the University of Pennsylvania in 1989. Moon
has written on numerous socio-economic issues, such as tax and fiscal policy, social
welfare policy, population aging, and pension systems in the Republic of Korea.
x Effects of Social Policy on Domestic Demand
Gloria O. Pasadilla is a research fellow at ADBI. She served as convener of the Asia-
Pacific Economic Cooperation Group on Services, and Philippine lead negotiator
for services for the Association of Southeast Asian Nations (ASEAN)-Australia-New
Zealand free trade agreement negotiations. She has advised the Philippines Bureau
of International Trade Relations of the Department of Trade, and the Office of the
Undersecretary for International Economic Relations of the Department of Foreign
Affairs on trade-related matters. She was senior fellow at the Philippine Institute for
Development Studies and assistant professor at the University of Asia and the Pacific.
She had brief stints at ADB, World Trade Organization, Southeast Asian Central
Banks Center, IMF, and Goldman Sachs. She holds a PhD in economics from New
York University and a master’s degree in international law and economics from the
World Trade Institute.
Yasuyuki Sawada is an associate professor in the faculty of economics at the University

of Tokyo; a faculty fellow of the Research Institute of Economy, Trade, and Industry
(RIETI); and a visiting researcher at the Japan International Cooperation Agency
(JICA) Research Institute. His research interests include econometric investigations of
household and firm behavior under uncertainties using micro-data from developing
and developed countries, poverty in Africa, impact evaluation of education policies,
and economic analysis of suicidez in Japan. Recently, he has been investigating the
impacts of natural disasters on household welfare. Mr. Sawada completed his master’s
degree in food research and PhD in economics from Stanford University.
Yvonne Sin is head of investment consulting and director of PRC market development,
employee benefits, and actuarial services at Watson Wyatt. She was head of the global
pensions practice in the World Bank’s social protection department. Ms. Sin specializes
in social insurance and employee benefit issues, and conducts research on pensions and
investments. She has provided technical consulting to the PRC Ministry of Finance
and Ministry of Social Security since 1999. She is also an appointed adjunct professor
at the Nanjing University of Finance and Economics. She received her bachelor of
science degree in mathematics from the University of Toronto and graduated from
the World Bank’s pensions fellowship program.
Akiko Terada-Hagiwara is an economist in the Macroeconomics and Finance

Research Division at ADB. She was previously an economist at the Institute for
Monetary and Economic Studies of the Bank of Japan. Prior to joining ADB, she
served as a consultant at Global Insight and the World Bank. Her areas of expertise
are international finance, development economics, and macroeconomic policies. She
has authored publications on international financial issues in a developing country
context. She holds MPhil and PhD degrees in economics from George Washington
University and a bachelor’s degree from Osaka University.
Presenters and Discussants xi
Yuqing Xing is a professor of economics at the National Graduate Institute for

Policy Studies in (GRIPS) Tokyo and a research associate at the East Asian Institute,
National University of Singapore. Formerly, Mr. Xing was a professor and director
at the International University of Japan. He served as sabbatical fellow at the World
Institute for Development Economics Research and visiting professor at the Institute
of Advanced Studies, both at the United Nations University; visiting research fellow
at the Bank of Finland; and consultant to ADB. Mr. Xing earned his bachelor of
science and master’s degrees from Peking University, and a PhD in economics from
the University of Illinois at Urbana-Champaign. He has published numerous articles
on regional economic issues.
Hiroshi Yamabana is a social security actuary in the International Labour

Organization (ILO)’s social security department. Previously, he has served as senior
actuary and deputy director of the Actuarial Division, Pension Bureau, in Japan’s
Ministry of Health and Welfare; social security actuary at the ILO; and social security
specialist at the ILO’s subregional office in Bangkok. He earned a master’s degree in
mathematics from Kyoto University. Mr. Yamabana has provided policy and technical
advice on social security to over 20 ILO member countries, served as a focal point for
quantitative research and modeling on social security, lectured at universities, and
delivered training courses to social security policy planners and administrators.
Ming Yan is a professor at the Institute of Sociology and the Center for Social Policy
Studies, Chinese Academy of Social Sciences. She served as senior adviser to the China
Program at the University of Southern California, and as assistant/associate professor
in the Social Science Department, City University of New York–LaGuardia. Ms. Yan
was a visiting research scholar for the German Academic Exchange Service (DAAD)
and recipient of an honorary international visiting fellowship from the University of
Surrey. She holds a PhD in sociology from New York University. Ms. Yan has written
on numerous social issues, including housing policies, urban redevelopment, and
sociology in the PRC and has delivered lectures and presentations on migration,
urban issues, Chinese sociology, and social instability.
xii Effects of Social Policy on Domestic Demand
Abbreviations
ADB Asian Development Bank

ADBI Asian Development Bank Institute
ASEAN Association of Southeast Asian Nations
BISP Benazir Income Support Program
CASS Chinese Academy of Social Sciences
CCT conditional cash transfer
CHIP Chinese Household Income Project
CICC China International Capital Corporation Limited
DFID Department for International Development
EPF Employees Provident Fund
GDP gross domestic product
ICBC Industrial and Commercial Bank of China
ILO International Labour Organization
IMF International Monetary Fund
IOM International Organization of Migration
ISSA International Social Security Association
IZA Institute for the Study of Labor
Korea Republic of Korea
Lao PDR Lao People’s Democratic Republic
NBER National Bureau of Economic Research
NGO nongovernmental organization
NPL nonperforming loan
NRCMC New Rural Cooperative Medical Care
NREGA National Rural Employment Guarantee Act
OECD Organisation for Economic Co-operation and Development
PBC People’s Bank of China
PRC People’s Republic of China
SASAC State-owned Assets Supervision and Administration Commission
SME small or medium enterprise
SOE state-owned enterprise
SPI Social Protection Index
UK United Kingdom
UNESCAP United Nations Economic and Social Commission for Asia and the
Pacific
UNICEF United Nations International Children’s Emergency Fund
US United States
Overview 1
I
Overview
Masahiro Kawai and Gloria O. Pasadilla
1. Introduction
One of the most important insights from the 2007–2009 global financial and
economic crisis is that the world should realistically and bravely face the fact
that the global imbalance cannot go on ad infinitum. New sources of growth
other than external demand from developed countries need to be found. One
important source is domestic demand, erstwhile unnoticed while all efforts
have been geared towards the outside market. Particularly in East Asia where
many economies have been successful net exporters and have accumulated a
significant amount of current account surpluses mirrored in high domestic
savings, the strategic re-thinking asks: how can they be enticed to spend more? In
particular, to the extent that social protection impacts saving behavior, is there
a role for better social protection policy to help stimulate domestic demand?
Against the backdrop of the global rebalancing need, ADBI organized its
th
12 Annual Conference in December 2009 with this key question in mind:
Can improved social policies reduce savings in high-saving Asian economies
and induce a long-term increase in consumption spending? The conference
aimed to better understand the channels and linkages of social policies and
domestic demand. That domestic demand needs to adjust to address the
global imbalance—i.e., to rise in countries with high external surplus and
to decline in countries with high deficit—is a generally accepted conclusion
from the global financial crisis; reliance on external demand for growth is
no longer sustainable. But can social policies help change domestic demand
and, if so, how?
2 Effects of Social Policy on Domestic Demand
Social policies have important developmental goals that go beyond merely

boosting consumption and demand. Improved social policies in education and
health, for example, help ensure that citizens gain better access to economic
opportunities. Social protection mitigates risks of social instability that rapid
economic growth and urbanization can induce in the short term. Furthermore,
over the long term, better health care and access to education produce healthier
and more productive human capital that elevates the economic growth path.
Therefore, improving social policies can help create a more stable domestic
society, and thus, has value over and above boosting consumption for
growth rebalancing. Nevertheless, social policies also perform an important
countercyclical role, which, at this juncture, should be highlighted in view of
the need to address low global growth.
The conference dealt with the general issue of savings and consumption,
but because the People’s Republic of China (PRC) is the country with the
highest amount of external surplus and is seen as key to global rebalancing,
the conference included a large number of contributions analyzing PRC’s
social protection policies, savings structure, and impact of social policy on
consumption. Other contributions in this volume review the savings and social
security literature, provide an overview of social protection policies in Asia,
assess the impact of aging and demographic factors on spending, and analyze
social protection policies in selected Association of Southeast Asian Nation
(ASEAN)-member countries.
Our question of how and whether better social protection can help rebalance
global demand, however, also reveals a host of other related policy issues. In
the case of the PRC, along with social policies that can help boost consumption
demand, there are tangled policies that affect its financial system, industrial
strategy, corporate governance, labor market, etc., reforms of which could
induce dramatic domestic demand growth. Likewise, in terms of social
protection, security coverage, pension adequacy, and financial sustainability
are policy challenges that will put pressure on government resources over the
medium and long term. This overview briefly tackles these related issues.
Section 2 discusses the underlying reasons behind the under-consumption
phenomenon in the PRC. Section 3 highlights the links between the demographic
aging of the population, household savings, and social spending. Section 4
discusses other structural reform issues that are closely connected to the global
rebalancing strategy for the PRC, as well as challenges in social protection
policy for the Asian region. The final section summarizes the points raised
at the conference and in this book.
Overview 3
2. Macroeconomics of Social Policy and Global

Rebalancing
The context for the heightened interest in social policy is its potential role in
global rebalancing. The persistent external surplus in some parts of the world,
particularly the PRC, which is then mirrored by the external deficit found in
other countries, namely the United States (US) (Figure 1.1), has been deemed
unsustainable, and hence, calls for a solution. While some countries have
also had rising surpluses (e.g., oil-exporting Middle Eastern countries), these
generally pale in comparison next to the PRC’s yawning surpluses, which have
increased especially since 2001, when it joined the World Trade Organization.
The International Monetary Fund (IMF) estimates that the PRC’s 2009 surplus
was about 0.91% of world gross domestic product (GDP), while the US deficit
accounted for 0.72% of world GDP. This explains why, for all the talk about
“global” rebalancing, the issue is ultimately about two countries—the US and
the PRC—while the rest of the world, including Japan, a former major surplus
accumulator, have relatively minor roles.
Economic theory says that the global adjustment to eliminate excessive
surpluses and deficits can be done through expenditure switching (that is, via
real exchange rate adjustment) and/or expenditure changing (i.e., through
Figure 1.1: Global Imbalances (current account balance in % of world GDP)
PRC+EMA = People’s Republic of China; Hong Kong, China; Indonesia; Korea; Malaysia; Philippines; Singapore;
Taipei,China; Thailand. GER+JPN = Germany and Japan. OIL = oil exporters. US = United States. OCADC =
other current-account-deficit countries. ROW = rest of the world.
Source: IMF (2010).
domestic demand adjustment). In discussing social policies’ effects on domestic

demand, the ADBI annual conference focused on the internal balance
route.1
Much has already been said about the need for the US to limit its over-
consumption and to save more as part of the domestic adjustment. The drop in
household wealth (through a decline in asset values) has already helped to cool
US private consumption. The gaping government deficit—at 10% of national
income in 2009—needs to narrow. The poor employment prospects, even after
massive fiscal stimulus in 2009, however, would likely make this challenging.
On the other side of this situation is the PRC’s under-consumption and
over-saving.2 Figure 1.2 shows that compared with other countries, both
developed and developing, the PRC’s 34% household consumption share of
GDP in 2008 is well below the more than 50% world average. Furthermore,
Figure 1.3 shows that this ratio declined by more than 15 percentage points
from a relatively high share of 55% in 1980.3 This low consumption share
does not align with what economic theory calls “the consumption smoothing
behavior of consumers.” Considering the PRC’s rapid economic growth, the
consumption ratio should have increased with the rise in expected lifetime
income, but the data show that Chinese consumption expenditures have not
increased apace with economic growth. On the contrary, consumption has
declined as a share of GDP. A large part of the internal balance solution to the
global imbalance therefore, rests upon encouraging Chinese consumption.
First, however, we need to understand the reasons why Chinese households
under-consume.
2.1 Understanding Household Under-Consumption in the PRC

2.1.1 Excess Savings
To understand the PRC’s under-consumption, we need to understand
Chinese “over-saving.” McKinsey Global Institute conducted a survey of
Chinese consumers to understand their saving behavior (McKinsey Global
1
Although the expenditure switching route is certainly no less important, it will not be discussed in
this volume.
2
Using household savings data by income group, Yan and Pan (Chapter 6) question the idea that PRC
oversaves.
3
It is important to note, however, that having GDP in the denominator, instead of household disposable
income, may overestimate the extent of under-consumption in the PRC because its GDP contains a
large proportion of investments. Future work in this area should take household disposable income
into account.
Overview 5
Figure 1.2: Household Consumption (% of GDP)
Source: World Bank (2010).
Figure 1.3: Household Final Consumption Expenditure (% of GDP)

Institute 2009). The results revealed that Chinese urban consumers are primarily
motivated to save for education, while illness and expenses associated with
caring for elderly parents follow in importance (Figure 1.4). Chinese consumers
also save for investment, especially for home purchase4 or for business, as well
as for retirement and unemployment risks. Four of the major reasons for saving
are related to social safety nets, that is, illness, elderly care, retirement, and
unemployment risks. The modernization of the Chinese economy began in the
1980s and efforts to increase state-owned enterprises’ (SOEs) competitiveness
meant that SOEs were no longer responsible for many social welfare benefits
(Baldacci et al., Chapter 8; Yan and Pan, Chapter 6). This restructuring forced
households to save much more for contingencies, thus partly explaining the
surge in Chinese savings.
Examining urban households in the PRC, Yan and Pan (Chapter 6) found
that, from 1995 to 2007, the average savings rate rose by 10 percentage points.
There are various explanations for the surge in savings: thriftiness (culture),
demographic factors (aging), macroeconomic uncertainty or the “target saving
hypothesis.”5 Even the predominance of males in the population (sex ratio)
Figure 1.4: Why Do Consumers in the PRC Save?
Source: McKinsey Global Institute (2009).
4
Interestingly, houses are the main asset type for households in the PRC, accounting for 71% of net
wealth, while bank deposits is a far second, with only a 16% share (Chen, Li, and Qui 2010)
5
This theory states that savings, brought down by weak returns on deposits, have to keep increasing
to maintain a savings target.
Overview 7
and the competitiveness in the marriage market have been suggested as
reasons for the high rate of Chinese savings (Wei and Zhang 2009). The
most popular of these factors and the one with the most robust empirical
support is the precautionary motive for savings, particularly the risk of large
health expenditures (Chamon and Prasad 2008). This result appears to be
corroborated by McKinsey Global Institute’s survey.
The inadequate public provision of many social sector services has
contributed to the high precautionary motive for savings. The Chinese
consumption profile shows increases in social security-related private spending,
in contrast to the decline in the share of public social spending. In 1992,
the share of these consumption expenditures constituted about 17% of total
urban consumption expenditures; by 2001, this share had almost doubled
(Zhang 2008). In particular, for health spending, McKinsey Global Institute
(2009) reported that out-of-pocket health care expenditures, currently 10%
of urban consumers’ consumption, is projected to reach 14% by 2025. In
contrast, the government’s share in total health spending declined. In 1986,
the share of public health spending was 39%; by 2007, this had dropped to 20%,
while personal spending on health care rose from 26% to 45% of total health
expenditures (Yan and Pan, Chapter 6, Table 6.4). Clearly, if the government
provides more health care funding, this would abate the financial strain from
health expenditures and reduce precautionary savings.
Private education expenditures in the PRC have, likewise, increased due
to ballooning education costs. In fact, education spending per capita is more
than five times greater than health spending (Barnett and Brooks 2010),
suggesting that increased government provision of education services, more
college scholarships for example, would have an even greater impact in reducing
household savings. Current government spending on education is geared
heavily toward primary and secondary education, while household savings
for education are targeted toward higher education.
2.1.2 Savings Structure: Households vs. SOEs Savings

Not everyone, however, subscribes to the idea that the PRC’s household sector
saves excessively. For example, Xing (Chapter 7) argues that the Chinese
household savings rate is not an international outlier but is well within the East
Asian saving norm. Comparing the PRC’s household savings from 1990–2007
with Japan’s household savings from 1960–1977—the period when Japan’s per
capita income was roughly similar to that of present-day PRC—Xing found that
the savings rate of the two countries under similar circumstances were roughly
equal. Japan’s household savings peaked at 23.1% in 1975 from 14.7% in 1960
and then declined slightly to 21.7% in 1977. Similarly, the PRC’s household
savings rose to 27.5% in 2007.6
More than household savings, corporate and government savings are
what actually drive high aggregate savings in the PRC. From a low rate of
12% of national disposable income in 1992, the corporate savings rate rose
to 23% in 2004 before subsequently declining to 18% in 2007. In contrast,
the household savings rate remained almost the same as in 1992 at 22%
(Figure 1.5). High corporate savings is attributed to large profits that
enterprises—particularly SOEs—derive from a lack of private competition
especially in strategic industries. The industrial restructuring in the past
that removed social welfare responsibilities from SOEs, plus state-sanctioned
monopoly power, has helped SOEs to achieve dominance in industries such
as banking, mining, telecommunication, oil refining and distribution, and
utilities (Xing, Chapter 7).
Figure 1.5: Trend in PRC Saving Rates (by sector, as % of disposable income)
% of disposable income
Source: Calculations based on CEIC Data Company, Ltd.
6
However, Xing argues that the high housing burden, particularly commercial housing in urban areas,
makes the de facto household savings rate lower than what the ‘high’ saving ratio implies.
Overview 9
Significantly, Yan and Pan (Chapter 6) also hold that while the household
savings rate in the PRC may appear relatively high, parsing it between urban
and rural savings or among quintile groupings in the population reveal an
important distribution pattern that has significant impacts on consumption
spending. First, in contrast to the steady increase in the saving rates of urban
households, rural savings have, in fact, declined since 1999. Moreover, when
households are grouped into income quintiles, the top 20% income group
contributes the lion’s share of household savings: 64% of both urban and rural
household savings. If this group’s savings were removed, the remaining 80%
of urban households have a savings rate of only 8.8%, while rural households
save at only 9.2%.
Decomposing savings has important policy implications on how social
policy will impact consumption demand. Depending on who actually saves,
policy solutions could be tailored to reduce savings and to expand consumption.
If, for argument’s sake, we accept that household savings in the PRC is not
excessive, then government policies that try to lessen household savings would
have limited effect in boosting consumption.
Instead, policies that try to extract part of SOEs’ profits in the form of
dividends to the state that would then be redistributed to households may
lead to better results in boosting consumption, especially if the redistribution
comes in the form of higher government social sector expenditures.
Targeting these social expenditures on lower income quintile groups that
have a higher propensity to consume could translate to a higher impact on
consumption.
2.1.3 Household Income

Apart from high savings, another reason used to explain the decline in
consumption spending as a share of GDP in PRC is the deterioration in
household income share in GDP. Aziz and Cui (2007) posit that high Chinese
savings explain only a small fraction of the decline in the consumption
share, while the larger portion is explained by the substantial drop in
household income’s share in GDP. In other words, Chinese households
consume little not because they save a lot but because their income is
little.
Of the different sources of household income, wages constitute the biggest
source. The share of investment in household income is not significant in
the PRC because of a lack of generally available investment instruments as
well as a limited number of publicly listed firms. Government transfers have
become insignificant since the Chinese modernization in the 1980s, while

bank deposit earnings have suffered from caps in interest rates. Wages are the
largest contributing source of household income, but they have not increased
with the rise in Chinese productivity (Huang 2010).
Figure 1.6 shows the strong correlation between household consumption
and wage income as a share in GDP. This suggests that more so than excessive
saving, the determining factor for the decline in the consumption ratio may
actually be the fall in the compensation share in GDP (Chapter 7). Addressing
the decline in household income through, for example, tax cuts,7 is thus the
right policy response to buoy up consumption spending.
2.2 Impact of Social Policy on Consumption

Regardless of the root cause of Chinese under-consumption, it is still useful to
ask how social policy would impact consumption. Baldacci et al. (Chapter 8)
identify three different channels by which improved social policy affects
consumption. First, there is an income effect: by lowering households’
Figure 1.6: Household Consumption and Labor Income Shares (% of GDP)

% of GDP
GDP = gross domestic product.

Source: PRC’s Statistical Yearbook.
7
The highest personal income tax rate in the PRC is 45% and the lowest is 5%.
Overview 11
health and education charges or by raising pension benefits, an expansion of
government social expenditures increases lifetime resources and thus boosts
current consumption. Second is the redistributive channel: the consumption
impact is greater if social policy spending is targeted toward households with
relatively high average propensities to consume. This is done, for example, with
pension transfers to the elderly and resource transfers from low to high average-
propensity-to-consume households (i.e., from high-income households to rural
or low-income ones). The third channel is the insurance effect: households
reduce precautionary savings because of lower future health and education
costs or higher retirement incomes.
How much would the impact of social policy be? Using the PRC’s household
survey data and applying a generational accounting framework, Baldacci et
al. (Chapter 8) estimate the income effect of 1% of GDP increase in each of
the social sectors on consumption. Table 1.1 shows that the impact of pension
reform equivalent to 1% of GDP translates to a 1.42% increase in consumption.
Table 1.1: Impact of Government Social Expenditures (% GDP)
Income Impact of 1% Increase in Government Expenditures in:

Pension Health Education
Total 1.42 0.77 0.51
Urban 0.92 0.46 0.24
Rural 0.50 0.32 0.27
Rural/Urban unit impact 67% 56% 55%
Source: Based on Baldacci et al.(Chapter 8).
For health and education expenditures, the impacts on consumption are 0.8%
and 0.5% of GDP, respectively. Moreover, though the simulation appears to
have a bigger impact on urban consumption, when government spending is
adjusted for differences in urban and rural expenditures,8 social spending
in rural areas has a higher per unit impact. In Baldacci et al.’s (Chapter 8)
simulation, the income effect on consumption of pension expenditures in
rural areas is 67% higher than in urban areas ((0.5/0.25) / (0.9/0.75)); for health
8
Baldacci et al.’s (Chapter 8) simulation assumed 75% of government pension spending is for urban
areas and 25% for rural areas; 69% of health spending for urban areas and 31% for rural areas; and
58% of education spending for urban areas and 42% rural areas.
and education spending, rural income effects are 56% and 55% higher than
those in urban areas, respectively.9
When translated into actual yuan and dollar values, the effects found
by Baldacci et al. (Chapter 8) are not very significant. For example, given
the PRC’s 2008 GDP of CNY30 trillion (approximately, US$4.4 trillion
assuming CNY6.835 to the dollar exchange rate), a 1.42 % of GDP increase
in consumption is equivalent to about CNY426 billion additional consumption
(or US$62 billion). Government health and education spending will generate
approximately US$31 billion and US$22 billion of additional consumption,
respectively. Therefore, the combined effect of pension, health, and education
spending, or 3% of GDP increase in government social spending, will increase
consumption by US$115 billion. This amount does not appear to be sufficient
to solve global rebalancing, given that US private consumption expenditure
is more than US$10 trillion.10
Nonetheless, though the additional consumption boost from social
policy improvement appears modest, improving the PRC’s social safety net
is a critical step forward in its economic development for reasons that go
beyond demand growth (McKinsey Global Institute 2009). It will improve
citizens’ living standards and ensure that more people benefit from the PRC’s
economic progress. And, if complemented by other structural reforms that
increase wage share in national income, the total consumption increase would
be even greater.
Some caveats, however, are worth pointing out. One is that Baldacci et al.’s
(Chapter 8) estimates assume that the increased government social spending
is funded by excess government savings or SOE savings. If tax financing is
factored in, the impact on consumption will likely be more modest than the
authors’ estimates.11
9
In addition, using a very rudimentary calculation of the insurance effect, the authors found an additional
0.3% of GDP increase in spending for pensions, 0.18% for health, and 0.2% for education. Similarly,
Wang (2010) found differential impact of social spending among different income groups. An increase
of social security coverage ratio has a marginal effect of CNY13.37 on per capita consumption in the
lowest income quintile group, while its marginal effect in the 75th quintile group is only CNY10.37.
10
Other studies have found slightly higher effects of government social spending on consumption. For
example, Barnett and Brooks (2009) reported a 2% of GDP increase in consumption resulting from
1% government health spending, but little evidence for effect of education spending. McKinsey Global
Institute’s (2009) study, however, was roughly in line with Baldacci et al.’s (Chapter 8) projections.
11
In anticipation of a future tax increase, Barro’s “equivalence” theorem implies that the impact on
consumption would be less because savings will not decline as much (compared to the situation without
tax financing).
Overview 13
Another caveat is the application of the same policy recommendations
of greater social spending in other East Asian countries that do not have the
same macroeconomic conditions as the PRC, i.e., do not enjoy the same high
level of public savings and current account surplus. While there, too, social
protection policies need to be reinforced because current social protection is
relatively low (Édes, Chapter 3), their own macroeconomic conditions could
constrain an expansion of social spending. The current Greek saga caused by
very generous social benefits without adequate financial resources is a specter
that should not be lost to many developing countries aspiring to spur demand
through social policies. If countries want social benefits to be sustainable,
Park (2009) warns against fiscal policy activism, particularly for countries
that have less fiscal space.
Pasadilla and Wiradisuria (Chapter 9) likewise point out various findings
from the literature that shows an attenuated link between safety nets,
particularly social security spending, and savings. As the literature shows,
social security has potential adverse-incentive effects on labor supply, retirement
decisions, and economic growth in general.
Finally, along with demand-boosting government social policies, addressing
supply constraints in the social sector should be part of a comprehensive policy
reform. For example, a mere increase in government spending on health
insurance, will not lead to a decrease in precautionary savings for health
expenditures if there is not enough access to high quality public healthcare
facilities or if there are not enough trained health personnel in the public sector
as consumers will still save in order to afford better healthcare (McKinsey
Global Institute 2009).
3. Microeconomics of Social Protection

On the microeconomics front, the conference sought to understand the
relationship between social protection, demographic factors, and household
and national savings. One important social protection is pension policy, or more
broadly, social security, which covers not only old-age and retirement pensions,
but also, as in most countries, accident and other safety net components like
unemployment, healthcare, maternity leave, and illness. Many cross-country
studies have analyzed the impact of old-age pension on household savings,
but few have analyzed the impact of other social programs (Horioka and Yin,
Chapter 4). Therefore, most analyses present the microeconomics of social
protection in the context of social security and abstract away other components
like active labor market programs, social assistance or safety nets.
Many studies on the subject have found a negative relationship between
social security and aggregate savings, though others also find ambiguous
correlations.12 The negative relationship comes about as a result of a “wealth
replacement effect”, that is, the expectation of income to rely on upon retirement
obviates the need to save more. However, the introduction of a public old-age
pension system could induce people to retire earlier, and this in turn will
encourage them to save more (“induced retirement effect”). The net impact
of public old-age pensions on household savings depend upon the relative
strengths of these two offsetting effects (Horioka and Yin, Chapter 4).
Horioka and Yin’s (Chapter 4) study notes that the relatively stable social
benefit ratio13 compared to the declining trend in saving rates in Organisation
of Economic Co-operation and Development (OECD) countries suggests
a weak link between the two. The authors’ cross-country regressions of
OECD countries’ savings rates indeed show that social benefit ratios explain
savings variation only to a limited extent. Other factors like population aging
and financial market conditions (credit constraints) seem more important
in explaining variations in saving rates. In particular, Horioka and Yin’s
results suggest that as a country’s elderly population increases, its saving
rate decreases; moreover, a country that has more domestic credit flows
as a percentage of GDP (a proxy for fewer credit constraints) has a lower
savings rate.
In the PRC, the evidence appears to support the negative relationship
between social security and household savings. Figure 1.7 illustrates the
difference in saving behavior of urban households with social security coverage
and those without it. Across various quintile groups, aside from the lowest,
households without social security coverage save more than those with coverage.
The highest income group has the largest difference in saving rates, with
households without coverage saving 46% of income compared to 21.5% for
those with social security (Wang 2010). Similarly, an urban consumer survey
suggests that those without health insurance save approximately 1.5 times more
of their disposable income than their neighbors who are insured (McKinsey
Global Institute 2009). These various pieces of evidence indicate that increased
12
For example, Hagiwara (2009) posited that the extended coverage of the social security system that
happened in the 1990s in the Philippines contributed to the decline in the country’s saving rate.
13
Social security receivable income divided by household disposable income.
Overview 15
Figure 1.7: Urban Household Saving Rates with and without Participation
in Social Security
Urban Savings Rate %
Note: Saving = Net Income – Consumption Expenditure.

Source: Institute of Population and Labor Economics, Urban Household Survey Data (2007) as cited by Wang
(2010).
coverage of social security or health insurance by the government may reduce

savings and increase domestic consumption.
Like Horioka and Yin (Chapter 4), Furceri and Mourougane (Chapter 5)
found a negative relationship between old-age dependency ratios with saving
rates among OECD members and countries in the Asia and Pacific region. In
addition, a higher old-age dependency ratio boosts public social spending,
particularly on pension expenditures and, to a lesser extent, active labor market
programs, housing, and healthcare spending. Furceri and Mourougane,
therefore, have suggested policies that temper the effect of aging through
policies like employment of older workers or deepening financial markets
for market instruments like annuities that could reduce the precautionary
motive for (over)saving.
In his comments on Furceri and Mourougane’s paper, Moon (comments on
Chapter 5) notes the difference of the impact of population aging on savings
in OECD countries with those of Asia-Pacific economies. While careful to note
the different dependent variables used in the regressions—OECD regression
used household savings rates, while Asia-Pacific regression used private savings
rates on the left hand side of the equation—Moon asked what could be driving
the apparently high sensitivity of saving rates in the Asia and Pacific region to
demographic factors. Put differently, why do Asia-Pacific economies’ elderly

populations appear to dissave faster than those in OECD? Various factors
like the difference in bequest behaviors, participation of old people in the
labor market, or the generosity of the social security system may explain the
difference. In his comments, Moon posits that the insufficient income security
system, introduced only in 1988 in the Republic of Korea (hereafter Korea),
makes the elderly rely on self-insurance measures, which may explain the rapid
asset decumulation of the elderly cohort in Korea vis-à-vis those in the US.
Another insight from Moon’s comments is that demographic aging should
have a different impact on public and private sector spending depending on
whether social protection institutions are strong or weak. In countries with
strong social protection, the public sector will likely take most of the pressure
from the demographic shocks, while in countries with weak social protection,
private saving behavior would take most of the adjustment.
4. Reform Challenges
While the conference’s focus was on the contribution of social policy to boosting
domestic demand and rebalancing growth, our analysis of Chinese households’
under-consumption has spawned other related policy issues. Social policy
reform is one small contribution to domestic demand growth and global
rebalancing, while a larger set of challenges form part of an “unfinished
agenda.” A few of these are macro-related, while others are more directly
related to social protection policy reform challenges.
4.1 Economy-Wide Reforms

4.1.1 Financial Sector Development
One issue related to the PRC’s oversaving/under-consumption is its financial
sector development. Various papers have attributed excess savings to borrowing
constraints caused by underdeveloped financial and capital markets. Horioka
and Yin (Chapter 4), for example, find that an improvement in private credit
ratio, a proxy for the degree of financial development, lowers household savings
rate. Similarly, Loayza, Schmidt-Hebbel, and Servén (2000), found that in a
sample of developed and developing countries, the ratio of domestic credit
flow to gross national disposable income had a negative and significant impact
on the private savings rate.
Overview 17
These empirical findings are consistent with the fact that households
that lack access to financial markets need to accumulate savings to purchase
durable goods such as television or automobiles (Jha, Terada-Hagiwara, and
Prasad 2009). Likewise, small-and medium-sized enterprises (SMEs) that have
limited or no access to financial markets need to accumulate large internal
funds to finance both their working capital and fixed investment needs. Even
corporations may, in theory, find the need to keep high rates of corporate
savings if the financial and capital markets are not sufficiently developed.14 In
the case of the PRC, the underdeveloped financial market has contributed to
low household consumption by making it necessary for them to accumulate
high savings to be able to purchase durable goods in the future.
Financial underdevelopment is also the reason why there are few alternative
sources of investment-related income for households. The PRC’s investment-
related income from interest on bank deposits, dividends, and real estate
leasing income is less than 2% of average annual household income (McKinsey
Global Institute 2009). This is far lower than investment income of households
in other countries: for example, real return on financial assets in the US and
Korea are 3.1% and 1.8%, respectively, compared to the PRC’s 0.5% (McKinsey
Global Institute 2009). The government’s cap on deposit rates, intended to
help banks regain profitability after the banking crisis in the 1990s, has come
at the expense of household income.
An improved financial sector landscape could help in rebalancing demand,
not only by lessening the need for households to hold very high savings, but
also by increasing household income, either through higher investment returns
accruing to households or through lower financing costs to firms, which could
stimulate firm activity — such as output and employment — and, hence, labor
income.
14
Arguably, the underdevelopment of the PRC’s financial market may have also partly contributed to
the rise in corporate savings shown in Figure 1.5. Because state-owned-banks tightened borrowing
conditions after the banking crisis in mid-1990s, firms (both SOEs and non-SOEs) have stronger incentives
to hold larger retained earnings. Some (e.g., Xing, (Chapter 7) argue, however, that a major part of
high corporate savings, especially SOE savings, has been due to extraordinarily huge profits reaped
from highly subsidized operations and from factor market distortions (e.g., land and energy subsidies)
that are skewed in favor of SOEs. From somewhat different perspectives, Aziz and Cui (2007) argued
that the heavy reliance of Chinese firms on bank borrowing and the banking crisis in mid-1990s that
led to more restrictive borrowing conditions have increased the financing cost of business operations
and, consequently, have put substantial downward pressure on real wages. The induced lackluster
growth in wages resulted in decline in household income and private consumption. IMF (2009) claims
that the rise in corporate savings reflects a combination of rapid growth, limited competition, and
financial underdevelopment.
4.1.2 Corporate Governance

Development in corporate governance is another critical component in the
region’s rebalancing strategy. Poor corporate governance in the region is touted
as one factor for very high corporate savings because the nonalignment of
corporate managers’ and shareholders’ interests impacts on firms’ executive
compensation and dividend payout policies. In Asia, for example, firms
have been sitting on their profits, not investing them, but not paying them
out in dividends either (Singh 2009), giving rise to high levels of corporate
savings. An improved level of corporate governance would allow shareholders
to exercise their rights and prevent firms from hoarding excessive savings.
Activist shareholders and institutional investors would be able to exert a
stronger influence on firms to either increase dividend payout, or use excess
profits for viable investment opportunities. Either way, domestic demand
would improve.
In his discussion, Xing (Chapter 7) notes the large salary difference in
the PRC between employees at SOEs and those at non-SOEs. Comparing
compensations from a recent wage survey, he found that, in eight out of ten
sectors, the average compensation in SOEs is the highest. In the utility industry,
the annual compensation of SOEs is 4.5 times higher than in collectively-owned
firms, which tend to offer the lowest compensation. In finance, state-owned
banks’ average compensation is close to 300% higher than the compensation
in collectively-owned banks. Xing argues that this evidence attests to poor
corporate governance in SOEs where dividends, which are supposed to be
paid to the government, are whittled down by too generous employee benefits
and privileges. To reduce the high amount of retained earnings in SOEs,
in late 2007, the Chinese government introduced a pilot program in which
SOEs disburse limited dividends (about 5–10% of profits, depending on the
industry) to the government. However, according to Xing (Chapter 7), all
collected dividends were reallocated back to various investment projects of
central SOEs, rather than to social welfare programs.15 While the reinvestment
of profits can further enhance SOEs’ profits, Xing argues that this practice may
also lead to overexpansion, overcapacity, and inefficiency, precisely because
retained earnings are a cheap source of funds and, under weak corporate
governance, less subject to market discipline.
15
Central SOEs are a group of SOEs that are controlled directly by the central government, i.e., the State
Council, rather than being controlled by provincial or metropolitan city governments. The dividends
collected by the Ministry of Finance are placed in a special account designed to support reorganization
and development of SOEs (State Council 2007).
Overview 19
IMF (2009) presents econometric estimates of how much impact improved
corporate governance policies in Asia would have. The IMF estimate suggests
that if Asia had the same level of corporate governance that Group of Seven
economies have, corporate savings would decrease by 2.4% of GDP. This
impact shows that reforming corporate incentives should be an integral part
of regional rebalancing policies.
4.1.3 Market Distortions

High government support of SOEs and labor market distortions are other
contributing factors to the global imbalance. Huang (2010), for example, cited
the labor segmentation between rural and urban workers, strict controls on
capital account outflows, regulated interest rates, state influence in credit
allocation, and subsidized prices of intermediate inputs like oil, energy, gas,
and electricity. These distortions have led to lower input costs and resulted
in high competitiveness of Chinese exports in the international market. The
solution, Huang argued, lies in the PRC’s completing market reform, which
would expose Chinese firms to international competition without indirect
subsidies in the costs of labor, credits, land, and intermediate inputs.
4.2 Challenges in Social Protection

Developing social protection itself also presents enormous challenges not only
for the PRC but also for other economies in the Asia and Pacific region. Édes
(Chapter 3) highlights the need for improved social security administration and
governance, technological support, and other challenges emerging from labor
migration, aging population, as well as climate change. Below, we discuss the
challenge of coverage expansion, pension adequacy, and financial sustainability
of social security in the region.
4.2.1 Expanding Coverage

Social security covers about 90% of the labor force in developed countries;
however, in Asia, coverage ranges from only 7.2% (Lao People’s Democratic
Republic [Lao PDR]) to 76.2% (Singapore) (Figure 1.8). Coverage of the working
age population in developed countries ranges from between 60% and 75%,
while it covers only from 6% and 45% in emerging Asia (Park 2009). In most
low- and middle-income countries, the coverage of the labor force is below
25%. Moreover, coverage tends to be skewed toward urban areas and formal
labor, due in part to the limited institutional capacity of Asian pension systems
Figure 1.8: Coverage Rate of Social Security in Emerging Asia

% of total
Lao PDR = Lao People’s Democratic Republic, PRC = People’s Republic of China.
Source: Yamabana (2009).
and the high administrative costs of reaching the informal and rural labor
sectors. In the PRC, the fragmented pension system and absence of a centrally
administered pension system adds to the low coverage of migrant workers.
The coverage challenge can be met by enhancing the administrative
capabilities of pension system institutions and improving programs’ financial
capacities. To increase social security coverage and address old-age poverty,
some low- and middle-income countries have adopted “social pension” schemes,
which are essentially retirement income transfers (Asher 2009), that are either
universal or means tested.16 Others have reorganized the employer-employee
relationship to include membership of informal sector labor in community-
based groups like cooperatives that act as premium collectors and benefit
facilitators, roles that corporations carry out among formal sector labor.
4.2.2 Retirement Income Adequacy and Financial Sustainability

The replacement rate (i.e., the ratio of retirement income to income before
retirement) is a measure used to assess the adequacy of old-age pension benefits.
The recommended replacement rate by pension experts is between 66% to 75%,
adjusted for longevity and inflation risks, but many Asian replacement rates
are below this recommended level (Park 2009), suggesting inadequacy.
16
Providing benefits to “deserving” beneficiaries through resources or means testing is hampered by
poor targeting capabilities in many countries.
Overview 21
It is one thing, however, to have a generous pension program; it is another
to have the capacity to fulfill it. The problem in these cases is whether the
cost of social security is indeed affordable and sustainable. If the country
cannot afford and sustain a generous pension system, the high replacement
rate is illusory.
The rapidly aging population compounds the financial sustainability
challenge. This is not a problem in Asia alone. Even in Europe, the financial
sustainability concern is forcing a re-examination of some pension policies,
e.g., age of retirement and years of contribution, among other things. Current
projections indicate that the required fiscal revenue to address demographic
aging in Europe is unattainable.17
Sustainability-enhancing reforms are needed everywhere, but especially
in countries with relatively generous pension promises and rapidly aging
populations, as in Western Europe. These reforms can encompass parametric
change: e.g., increasing the retirement age, reducing promised benefits, or
increasing the number of years for required contribution, among others. For
confidence in the system, it is important that social benefits and contribution
arrangements are credibly designed and perceived to be sustainable; otherwise,
households’ precautionary savings will not decrease to support consumption
growth.
4.2.3 Low Collection Rate and Investment Efficiency

Another social protection challenge relates to the upgrade of institutional
and administrative capacities of pension institutions to administer programs
with low transaction costs. This entails developing accurate data and record-
keeping systems, effective collection of members’ contributions, and proper
management of funds, including earning higher rates of return on investments.
Higher investment return is closely connected with the allowable investment
policy of pension funds, i.e., the types of financial assets they can hold. It is also
related to the degree of development of the domestic financial system—a more
developed domestic financial system can offer a broader array of investment
instruments to pension funds.
17
The European Commission (2009) pointed out that projected fiscal revenue of 5% to 6.5% of GDP
is necessary to address population aging. This required revenue increase is unattainable as even the
most ambitious tax reforms that countries can undertake generally do not yield more than 1%–2% of
GDP (Asher, comments on Chapter 4).
5. Summary and Conclusion

The recent economic slowdown has forced a re-examination of the export-
dependent growth strategy that has enormously benefited the Asian region
for many years. It is clear that the PRC and the rest of Asia need to lessen
their dependence on extra-regional demand—particularly from American
and European markets—and increase reliance on domestic demand. ADBI’s
12th Annual Conference focused on one component of domestic demand—
private consumption—and one potential policy instrument that may boost
such demand: social policy.
The discussions ranged from a broad overview of social policies in Asia
to the relationships between private saving and social security, and public
spending and social security systems; from household savings to SOE-retained
earnings policy reforms. The major ideas and policy conclusions were:
t TFWFSBM DPVOUSJFT BSF VOEFSUBLJOH NBKPS SFGPSNT UP JODSFBTF TPDJBM
protection coverage and sustainability, but appropriate policy design is
important to mitigate adverse incentive effects;
t TPDJBMCFOFmUSBUJPT UIFBHFTUSVDUVSFPGUIFQPQVMBUJPO BOECPSSPXJOH
constraints are important determinants of household saving rates as well
as of the pattern of social spending across countries;
t JODSFBTFEEJSFDUJODPNFUSBOTGFSTUPMPXJODPNFGBNJMJFT SVSBMBSFBT BOE
older population, because of their higher average propensity to consume,
will not only boost consumption but also mitigate social inequality;
and,
t DPMMFDUJOHEJWJEFOETGSPN40&TBOEVTJOHUIFNUPGVOETPDJBMQPMJDJFT
is an efficient use of excess corporate savings.
However, one policy instrument—social policy—is not sufficient. For
global rebalancing, other related policies are important such as financial
sector development, corporate governance reform, and further factor market
liberalization.
Overview 23
References
Asher, M. 2009. Extending Social Security Coverage in Asia-Pacific: A Review of Good
Practices and Lessons Learnt. International Social Security Association Working
Paper No. 6. Geneva: International Social Security Association.
Aziz, J., and L. Cui. 2007. Explaining China’s Low Consumption: the Neglected Role of
Household Income. IMF Working Paper No. 07/181. Washington, DC: IMF.
Barnett, S., and R. Brooks. 2010. China: Does Government Health and Education
Spending Boost Consumption? IMF Working Paper No.10/16. Washington, DC:
IMF.
Chamon, M., and E. Prasad. 2010. Why Are Saving Rates of Urban Households in China
Rising? American Economic Journal: Macroeconomics 2(1): 93–130.
Chen, Y., F. Li, and Z. Qui. 2010. Accounting for Household Saving Rates in China.
Presentation available at: http://hdl.handle.net/10086/18551.
European Commission. 2009. Sustainability Report 2009 (provisional version). Brussels:
Directorate General, Economic and Financial Affairs.
Hagiwara, A. 2009. Explaining Filipino Household Declining Saving Rate. ADB
Economics Working Paper Series No. 178. Available at: http://www.adb.org/
Documents/Working-Papers/2009/Economics-WP178.pdf.
Huang, Y. 2010. Fixing China’s Current Account Surplus. East Asia Bureau of Economic
Research Newsletter. January.
IMF. 2009. Regional Economic Outlook: Asia and the Pacific. October. Washington, DC:
IMF.
————. 2010. International Financial Statistics Yearbook, Washington, DC: IMF.
Jha, S., A. Terada-Hagiwara, and E. Prasad. 2009. Saving in Asia: Issues for Rebalancing
Growth. ADB Working Paper No. 162. Manila: ADB.
Loayza, N., K. Schmidt-Hebbel, and L. Servén. 2000. What Drives Private Saving across
the World? Review of Economics and Statistics 82(2, May): 165–181.
McKinsey Global Institute. 2009. If You’ve Got It, Spend It: Unleashing the Chinese
Consumer. Beijing: McKinsey Global Institute.
Park, D. 2009. Aging Asia’s Looming Pension Crisis. ADB Economics Working Paper No.
165. Manila: ADB.
State Council, PRC. 2007. Guidelines on the Pilot Program for Preparing the Account of
State Assets. State Council, No. 26, 2007 (in Chinese).
Singh, A. 2009. Asia’s Corporate Saving Mystery. iMF Direct Economic Forum blog.
Available at: http://blog-imfdirect.imf.org/2009/11/08/asia%E2%80%99s-
corporate-saving-mystery/.
Wang, D. 2010. Can Social Security Boost the PRC’s Domestic Consumption? ADBI
Working Paper No. 215. Tokyo: ADBI.
Wei, S. and X. Zhang. 2009. The Competitive Savings Motive: Evidence from Rising Sex
Ratios and Savings Rates in China. National Bureau of Economic Research (NBER)
Working Paper 15093. Cambridge, MA: NBER. Downloadable at: http://www.
nber.org/papers/w15093.pdf.
World Bank. 2010. World Development Indicators Online. Available at: http://publications.
worldbank.org/WDI/.
Zhang, Y. 2008. The Empirical Analysis of Social Security Level and Consumer Trends.
Xiandai Jingji 12: 105. (in Chinese)
Keynote 25
II
Keynote Address
12th Annual Conference on
Effects of Social Policy on Domestic Demand
Haruhiko Kuroda, President, Asian Development Bank
It is my great pleasure to welcome you all to ADBI’s 12th Annual

Conference.
The topic of this year’s conference is “the Effects of Social Policy on
Domestic Demand.” This is a very timely subject as the global financial
crisis has forced many countries, particularly in Asia, to re-evaluate their
traditional growth strategy. For many years, Asia had benefited from the belief
that demand from the global market, especially advanced economies, could
always absorb its rapidly growing exports. The global crisis has shattered
this illusion.
Yes, the region will remain engaged in trade and hence needs vibrant world
demand. But Asia should also explore ways to increase the growth potential of
domestic and regional demand in the medium and longer term. First, the East
Asian region can take steps to increase the amount of genuine intra-regional
trade by advancing economic integration. Second, and more importantly,
East Asian countries should re-examine what scope they have to spur their
own domestic demand. Given Asia’s diversity, each country will have different
capacities to increase domestic demand. But, for some large countries in the
region, with their high savings rates and large current account surpluses, a
greater scope exists.
Today’s conference, therefore, fittingly asks the following questions: Can
domestic demand be enhanced by improved social protection policies? In
a region that is known for high savings, can better social protection help
reduce household savings and induce a long-term increase in consumption
spending?
Let me share some important related research findings from ADB on social
protection and domestic demand that may contribute to your discussions. The
Asian Development Outlook 2009, released in March this year, argued for two
kinds of policies that can strengthen domestic consumption. First are policies
that transfer more corporate savings to households; and second are policies
that reduce the precautionary motive for savings among households.
Studies of the PRC, the Philippines, and Taipei,China show that the
precautionary saving motive has been important in explaining household
saving behavior. Development of the social security system, including social
health insurance, has been seen to substantially weaken this motive for saving.
The availability of health insurance is particularly important for the aging and
elderly because of rising health care costs. The availability of a credible pension
system is likely to boost household consumption once individuals become
certain about their future income stream in the post-retirement period. And
increased provision of public education could also lower household saving by
reducing a family’s need to save for their children’s education.
Hence, policies that rationalize public spending to increase social transfers,
reform pension systems, and provide universal health care insurance and
education are top potential contributors to domestic demand growth. These
policies will not only generate short-term demand for education and health
care services, but also ensure long-term human capital investment, promote
lifetime earnings, and create greater economic potential.
The effect is stronger if more support is targeted to those with higher
spending propensities, including the poorest people. Policies to shore up
domestic demand should therefore include the poor through targeted transfers.
When directed to the poor, such funds will not be saved, but instead used to
buy goods and services, thereby supporting the broader economy.
Our research also pointed out that the buildup of savings in Asia might
have been smaller if income prospects were more stable and financial markets
had been more developed. Uncertainty in labor markets and income flows
strengthen precautionary saving motives. Lack of access to finance can also
play an important role in driving up saving rates despite rapid income growth,
especially among younger households. Those who cannot borrow against their
lifetime income are likely to save much more to provide for contingencies
than they would if credit were more easily available.
Thus, labor market reforms to increase employment of young adults,
improve flexibility and relevance of skills training systems, and create more
secure jobs with adequate social protection can reduce uncertainty in income
Keynote 27
prospects. Reduced uncertainty would lessen excessive precautionary savings
and increase consumption.
Financial sector reform can also reduce the need for saving by providing
access to borrowing and investment opportunities, through development of
consumer credit and mortgage markets. While not strictly a social policy, this
too is an important determinant of saving.
Today’s conference will present a wide range of topics for discussion and
much food for thought. The analysis of theoretical and empirical research
will provide an important backdrop to the discussions on savings and social
protection. It is indeed important to understand the welfare implications and
costs of social protection expenditures to properly assess how far economies
can and should extend social benefits. The examination of income distribution
policies between capital and labor could assist policymakers in deciding
on possible shifts in policy to bolster household income. Demographic
considerations are, likewise, vital not only for savings but also for long-term
sustainability of social security systems. Finally, the analysis of the impact of
changes in social policies on Asian economies will provide important policy
implications in promoting the growth of domestic demand.
Thank you for your attention and your participation in this comprehensive
dialogue on an issue of much relevance for Asia today. I wish you all the very
best in your discussions.
Social Protection in Developing Asia and the Pacific 29
III
Social Protection in Developing Asia
and the Pacific
Bart W. Édes
1. Introduction
There are a variety of interpretations of “social protection” but all overlap to

a large degree with the definition embraced by the ADB: “Social protection
is a set of policies and programs designed to reduce chronic poverty and
vulnerability by promoting efficient labor markets, diminishing people’s
exposure to risks, and enhancing their capacity to protect themselves
against hazards and interruptions in employment or loss of income” (ADB
2001).1
Social protection covers a wide array of measures, including contributory
social insurance programs, and noncontributory targeted social safety net
programs (also known as social assistance). Social protection also includes
policies and programs designed to promote employment and the protection
of workers. Labor-intensive public works programs provide employment for
the jobless and underemployed. Particularly in times of economic difficulty,
employment guarantee schemes can help workers cope with reduced incomes
caused by job loss or reduction in hours worked.
Social insurance includes programs to cushion the risks associated with
unemployment, ill health, disability, work-related injury, maternity, old age,
The author acknowledges with gratitude the helpful comments of Camilla Holmemo,
Celine Peyron Bista, Davide Furceri, and Karin Schelzig Bloom.
1
The strategy noted here subdivides social protection into five components: labor markets, social
insurance, social assistance, micro- and area-based schemes to protect communities, and child
protection.
and survivorship. Social assistance refers to transfers that are not based on
prior contributions but instead are financed from the general tax system.
Such transfers are intended to assist low income and vulnerable groups, such
as poor children, victims of natural disasters or armed conflict, handicapped
persons, and single-parent households. A range of targeting methods can be
adopted to identify beneficiaries, ranging from self-selection and categorical
targeting, to full means testing. Each targeting method implies trade-offs in
terms of program cost and accuracy.
In addition to their other contributions to a society’s well-being, social
protection benefits act as automatic stabilizers in the economy by minimizing
the fall in consumption during a period of slowing growth.
Countries tend to build up their social protection systems as they rise up
the economic ladder. Not surprisingly, the states with the most extensive,
well-funded social protection systems are industrialized countries, with those
in Western Europe standing out as the most comprehensive.
Having made unprecedented gains in lifting hundreds of millions of people
out of extreme poverty over the past three decades, and boasting the highest
growth rates in the world, the Asia and Pacific region is now giving greater
attention than ever before to social protection. The industrialized countries of
the region with a high per capita GDP have constructed the most substantial
social protection systems, e.g., Japan and Korea. But middle-income countries,
such as Thailand, have also introduced basic elements of social protection to
protect their most vulnerable, and to ensure that poverty reduction gains are
protected through good times and bad.
Virtually all developing countries in Asia and the Pacific have some kind of
public social protection system, and most of the region’s poor people benefit
from at least some social protection. However, overall coverage in the region
is only about 35% (Wood 2009). Provision of social protection in Asia and
the Pacific is highly influenced by the traditionally modest role of the state
in mitigating individual risk and vulnerability, viewing this responsibility
primarily as a matter for one’s self, family, and close community.
India, Indonesia, Philippines, PRC, and Viet Nam are examples of countries
that have begun to integrate contributory benefit programs, basic health-care
provision, and social assistance into their development strategies. Beyond
the benefits that they provide to poorer population groups, these initiatives
help rebalance national economies, making them less vulnerable to external
shocks, and providing more inclusive and broader development (International
Social Security Association [ISSA] 2009).
This chapter reviews the state of social protection in developing Asia
and the Pacific. This comes at a time when the world is recovering from a
serious economic downturn that has forced millions out of their jobs and
into impoverished or vulnerable conditions. The chapter compares the level
of social protection across countries and subregions, and examines the social
protection response of governments to the recent global economic slowdown.
Finally, this chapter looks at three of the key challenges for efforts to strengthen
social protection in Asia and the Pacific: climate change, aging populations,
and coverage of migrants.
This chapter concentrates on developing countries in the region, although
occasional references are made to high-income countries for comparison
purposes. The primary focus is on state-provided social protection programs,
though it is recognized that social networks and private social protection
arrangements are often the first or most important measures that people
rely upon to address risk and vulnerability. The paper does not attempt the
tackle the critical issue of financing social protection, a topic well covered in
other recent publications (see, for example, Grosh et al. 2008).
2. Social Protection Across the Region

ADB has developed a unique Social Protection Index (SPI) summarizing the
overall level of social protection activities in a country. It is modeled on the
Human Development Index created by the United Nations Development
Programme. The SPI provides a measure of the extent to which Asian and Pacific
countries provide welfare, a labor market, social security, health insurance,
microcredit, child protection, and targeted education and health support
programs to their citizens, especially those living below the poverty line.
The SPI is a composite of four components: (i) social protection expenditure,
(ii) the total number of beneficiaries of social protection programs targeted at
key groups, (iii) the number of poor beneficiaries of social protection programs,
and (iv) social protection expenditure going to poor people. Thus, two of the
components relate to expenditure, and two relate to coverage. Two involve
general targeting, and two relate specifically to poor people.
Figure 3.1 presents the SPI values and rankings for 31 countries in the Asian
and Pacific region, with 0 being the lowest value and 1.0 the highest. All of the
countries covered by the SPI have low-to-medium income economies, except
for Japan, Korea, and the Cook Islands, which all rank in the top six. Several
republics that were part of the former Soviet Union, as well as Mongolia, appear
in the upper one-third of the rankings due to their reasonable provision of
social protection, a legacy of the communist era. Several Pacific and low-income
countries in different parts of the region appear toward the bottom of the
Figure 3.1: Social Protection Index Values for Asian Countries
Lao PDR = Lao People’s Democratic Republic, PRC = People’s Republic of China.
Source: ADB (2008).
index. Sri Lanka, India, and the PRC are situated very close together (places
9 to 11, respectively), below some of the former command economies.
As mentioned previously, one of the four components of the SPI is social
protection expenditure. ADB has calculated that the region-wide mean for
social protection expenditure as a share of GDP is 4.8% (the median is 4.0%).
Japan spends by far the most (16.0% of GDP), followed by the Marshall Islands
(13.5%), Uzbekistan (11.1%), Kyrgyz Republic (11.0%), Mongolia (9.8%), and
Korea (7.5%). The highest figure for a South Asian country is recorded by Sri
Lanka (5.7%), followed by Bangladesh (5.3%).
In Southeast Asia, Viet Nam spends the most on social protection as a share
of GDP (4.1%). The social protection expenditure for the PRC is 4.6% of GDP,
for India, 4.0%. Countries spending 1.5% or less include (from the highest
spenders to the lowest) the Maldives, Bhutan, Cambodia, Lao PDR, Tonga,
Vanuatu, Tajikistan, and Papua New Guinea (only 0.3% of GDP).
While the pattern of social protection expenditure varies considerably
between countries, in most cases it is dominated by social insurance programs
which, with a few exceptions (e.g., central Asian countries, Japan, Korea), are
largely confined to the public and formal sectors, thus excluding most of the
population.
2.1 Social Spending Compared to OECD Member Countries

Comparisons of social protection spending between developing Asia and the
Pacific, on the one hand, and developed countries generally, on the other, are
imperfect since the SPI includes only two member countries of the OECD,
Japan, and Korea. However, the OECD does publish data on public social
expenditures, which substantially overlap with, but are not identical to, social
protection spending as measured by the SPI.
An OECD study published in 2009 (Adema and Ladaique 2009) shows
that gross public social expenditures in OECD countries, including public
pensions and public health expenditures, total 21% of GDP. The largest
category of public social spending is for old-age and survivor pensions (7%
of GDP) and health services (6% of GDP). Japan, which tops the SPI’s social
protection expenditure table (Figure 3.1), ranks only 20 out of 30 OECD
countries, with just 18.6% of GDP dedicated to public social expenditures.
Korea, which appears in the top 20% of Asian and Pacific countries in social
protection spending, is ranked last among OECD countries on public social
expenditure (6.9% of GDP).
By way of comparison, Sweden’s public social expenditure as a share of

GDP is 29.4%, the highest among OECD member countries. Figures for other
countries include France (29.2%), Germany (26.7%), United Kingdom (UK)
(21.3%), Czech Republic (19.5%), Australia (17.1%), US (15.9%), and Turkey
(13.7%).
The advocacy nongovernmental organization (NGO), ActionAid, has
produced a social protection ranking of developing countries, including
seven in East Asia and South Asia (Table 3.1). ActionAid uses six indicators
of social protection that would have a direct or indirect bearing on food
Table 3.1: Developing Country Indicator—Social Protection (SP)
Subsidized
Minimum Food
Young Employment/ Rations/
Child Free Living Maternity Vouchers/ Old-Age
Feeding/ School Standards Nutrition/ Community Social SPa
Country Nutrition Meals Guarantee Entitlements Kitchens Pensions Grade
Weighting 17% 17% 17% 17% 17% 17%
India Medium High Medium No Low Medium C
Bangladesh Low No Low Low Low Low D
Viet Nam No No Low Medium No Medium D
China,
People’s
Rep. of No No Low No No Medium E
Cambodia Low No No No Low No E
Pakistan No Low Low No No No E
Nepal No No No No Low No E
a
SP Grade is on a scale of A to E, where A is high.
Source: ActionAid (2009).
security: young child feeding and nutrition, free school meals, minimum
employment/living standards guarantee, maternity nutrition/entitlements,
subsidized food rations/vouchers/community kitchen, and old-age social
pensions. ActionAid’s methodology is simple compared to the SPI, and is
inspired by an interest in what governments are doing to address hunger.
Nonetheless, the results for countries that appear in the ActionAid study
are similar to those of the SPI.
2.2 Social Insurance
Virtually every country in the Asia and Pacific region has a formal social
insurance system. However, these systems are generally limited to workers in
the government and formal employment sectors. These systems, therefore,
have little relevance to the informal and rural sectors, where most people
live and work, especially poor people. However, Japan and Korea—and some
of the central Asian republics—have more comprehensive systems. The PRC
has a comprehensive system for its “legal” urban population, but is only now
extending this system to its large rural population (ADB 2008).
In OECD member countries, which include Japan and Korea, the average
share of the labor force covered by mandatory pension schemes is 83.3% (95.3%
in Japan). This compares to 35.6% in Sri Lanka, 27.1% in the Philippines,
20.5% in the PRC, 16.2% in Viet Nam, 9.1% in India, and 6.4% in Pakistan.
Few countries in the Asia and Pacific region have social pensions that provide
safety net retirement income for people who are not members of formal schemes
(OECD and World Bank 2009).
The region is characterized by considerable heterogeneity in the institutional
nature of its national social security systems—including social insurance
programs, national provident funds, tax-financed programs, mandatory
occupational schemes—as well as in the scope of benefits and breadth of
coverage provided. Most countries in the region provide benefits for work injury,
old age, disability, and survivorship. However, the development of programs
for sickness and maternity benefits, family allowance, and unemployment
benefits is patchier (ISSA 2009).
2.3 Social Assistance Measures

Most developing Asian and Pacific countries have put in place social assistance
programs systems that provide cash or in-kind transfers to very poor and
vulnerable groups, such as the elderly, the disabled, and widows. Beneficiaries
tend to be limited to the very poorest (ADB 2008).
As in other regions of the world, the effectiveness of social assistance measures
depends in large part on how they are targeted. The methods used to distinguish
between who is—and who is not—needy have trade-offs in terms of accuracy and
administrative costs. The main methods of targeting include means testing
(and proxy means testing), community-based targeting, geographic targeting,
demographic (or categorical) targeting, and self-targeting (Grosh et al. 2008).
Governments in Asia and the Pacific often use cash transfers to support the
purchasing power of vulnerable populations. The overall impact of these schemes
depends largely on their ability to reach poor people with minimal targeting errors.
One increasingly popular tool is the conditional cash transfer (CCT), which is a
cash benefit provided to a poor family only if the targeted recipient carries out a
specified, socially desirable action, most often involving investment in children’s
human capital. Such an action can include ensuring children’s attendance at
school, taking the children to the doctor for regular check-ups, attending nutrition
seminars, or ensuring that children are fully immunized.
CCTs are used to reduce current poverty, increase access to social services,
and reduce the chances that the next generation will live in poverty. Originating
in Latin America in the 1990s, CCTs have now been adopted by some Asian
governments. For example, in 2007, Indonesia launched Program Keluarga
Harapan (the Hopeful Family Program). This program, financed by savings
from the removal of a universal fuel subsidy, targeted more than 15 million
households, and was later expanded to more than 19 million households,
representing 35% of the country’s population. Implemented by the Ministry
of Social Affairs, Program Keluarga Harapan families can receive up to about
US$220 per year (depending on their size and composition) on the condition
that specific health- and education-related obligations are met. The Indonesian
Post Office’s extensive network is used for the payment of benefits.
Box 3.1: The Plight of Invisible Children
One obstacle to accessing social protection measures is the lack of registered

births. The United Nations Children’s Fund (UNICEF) estimates that around 51
million children born in 2007 were unregistered, nearly half of them in South Asia.
Without a birth certificate a child may have difficulty accessing basic rights such as
education and health care; for example, in a number of countries a child requiring
vaccinations will be turned away from the health centre if they cannot produce a
birth certificate. Those without birth certificates are also vulnerable to exploitation
and abuse. If a child is trafficked across a country border to work, they cannot prove
they are too young if they have no birth certificate. The United Kingdom (UK)-based
NGO Plan International estimates that about 28% of births go unregistered in East
Asia and the Pacific, and 64% in South Asia.
Source: Plan International (2009).

The Philippines operates the Pantawid Pamilyang Pilipino Program, which
provides grants to extremely poor households to improve their health, nutrition,
and education, particularly that of children aged 0–14. The program aims
to break the intergenerational poverty cycle through investments in human
capital. For example, the program provides PHP6,000 (US$129) annually
per household for health and nutrition expenses, and PHP15,000 (US$323)
annually per household with three children qualifying for educational expenses
if certain conditions are met.
In Pakistan, the World Bank-financed Punjab Education Sector Reform
Program provides families with PKR200 (US$3) a month per girl student to
ensure that their daughters attend school. Net enrollment in primary schools
in Punjab increased from 45% to 62% between 2001 and 2007. Female primary
net enrollment during the same period increased from 43% to 59% and for rural
girls from 38% to 55%. Over 350,000 eligible girls receive monthly stipends
pegged to school attendance (World Bank 2009).
Another example of social assistance is the social pension, generally defined
as state-provided noncontributory regular cash transfers to older citizens,
given at specific ages in different countries. Social pensions help to reduce the
poverty of older people and their dependants, while increasing older people’s
status, material security, and access to services. Few Asia-Pacific countries offer
social pensions to provide safety net retirement incomes for people who are
not members of formal schemes (HelpAge International 2006).
Research carried out by HelpAge International, an NGO, provides
comparative data on social pensions in seven low- and middle-income countries
in South Asia and Southeast Asia (Table 3.2).
3. Social Protection by Sub-Region
3.1 Central Asia

Central Asian countries (including Mongolia) spend substantially more
of their budgets on social protection than countries in other parts of the
Asia and Pacific region (6.8% of GDP versus the region-wide average of
4.8%). Comprehensive, state-provided social systems were developed during
the communist period. Despite their relatively high costs, these systems
have to some extent been maintained irrespective of the post-transition
performance of the economies. In some countries, however, the benefits
Table 3.2: Social Pensions in Low- and Middle-Income Countries
% of People
Universal Amount Paid over 60 Low- (L)
(U) or Monthly in % of Receiving Cost or Middle-
Age Means US$ and Local Population a Social as % of Income (M)
Country Eligibility Tested (M) Currency over 60 Pension GDP Country
US$2
Bangladesh 57+ M BDT165 6 16* 0.03 L
US$4
India 65+ M INR200 8 13 0.01 L
US$2
Nepal 75+ U NPR150 6 12 ** L
63+ men US$3
Tajikistan 58+ women M TJS12 5 ** ** L
US$9
Thailand 60+ M THB300 11 16 0.01 M
US$5
Viet Nam 60+ M VND100,000 7 2 0.02 L
US$5
Viet Nam 90+ U VND100,000 7 0.5 < 0.005 L
* Percentage of people over 57 receiving a social pension.

**Not known.
Source: HelpAge International (2006).
provided by these systems are small relative to the national poverty line.
A much larger share of social protection expenditure in Central Asia is
dedicated to social assistance (24%) than in the Asia and Pacific region as
a whole (17%) (ADB 2008).
Under communism, central Asian countries maintained very high
employment ratios. The greatest challenge now for contribution-funded systems
in the subregion is the fall in employment in the formal sector, in particular in
the public sector, which has led to a rise in system dependency rates and the
erosion of what has hitherto been an almost universal entitlement. Retirement
benefits make a far larger contribution to household incomes than social
assistance payments, and are also of greater consequence in fiscal terms than
other social spending. In the course of the 1990s, however, pensions lost much
of their purchasing power, not least because of the sharp drop in contribution
receipts at a time of declining real wages and open unemployment. Many
pensioners therefore continue to be gainfully employed, are supported by
their relatives, or are affected by poverty (Müller 2003).
Although social expenditures are high by an Asia-wide standard, per
capita government spending on health in all central Asian countries (except
Kazakhstan) is too low to guarantee basic medical care for all. Patients and
their families must therefore make “informal payments” to be assured of
medical treatment. Social protection expenditures are so low in Tajikistan that
poverty is widespread (Müller 2003). Tajikistan does not optimize design and
implementation of social expenditures, spending mostly on utility subsidies
that do not effectively reach poor people (World Bank 2008).
Some of the lower-income former Soviet republics have developed elements
of effective social assistance, typically a single well-targeted program focusing
on poor families. One such example is the Unified Monthly Benefit in the
Kyrgyz Republic. Similarly, in former Soviet republics in the Caucasus region,
Azerbaijan and Georgia operate targeted social assistance programs, and
Armenia maintains a family benefit program (World Bank 2008).
3.2 East Asia

Developing countries in East Asia (excluding Mongolia) spend an average of just
2.8% of GDP on social protection, compared to the region-wide norm of 4.8%.
East Asia dedicates a larger share of its social protection expenditures to social
insurance (64%) than the average for Asia and the Pacific (55%) (ADB 2008).
Cook (2009) recently reviewed social protection in East Asia. Except where
noted, the analysis below is drawn from her paper.
The countries of East Asia exhibit a diverse set of welfare institutions, in
part reflecting an uneven colonial legacy, the influence of Confucian, Buddhist,
and Muslim heritage, and, through the 20th century, very divergent patterns
of economic development and political rule. Nonetheless, formal systems
have evolved in remarkably similar ways in post-colonial, communist, and
developmental states. Generous social security has been provided to a minority
of the urban and public sector elites, while social assistance programs have been
minimal, with the most reliance placed on the family. The more recent reform
agenda is also highly variable, in part due to varied histories and contexts, but
also perhaps due to the weaker presence of the international donor community
compared to some other regions (Cook and Kabeer 2009).
The poorest countries of East Asia, such as Cambodia, face chronic
poverty and a shortfall even in basic services and social provisioning. Their
governance structures and capacity, in part the legacy of conflict, are weak,
and international assistance, both financial and technical, will be necessary
in the foreseeable future for the development, funding, and implementation
of effective social protection programs.
Low- to middle-income countries in East Asia, such as Indonesia, are

building on the experience of their more developed neighbors and the lessons
from the 1997–1998 Asian financial crisis, which demonstrated that social
policies can have a role in generating more inclusive growth, and that the
costs of not having adequate protection mechanisms for vulnerable groups
in place at moments of crisis are extremely high.
Formal social security schemes exist across developing East Asia, generally
covering a relatively small number of civil servants and formal sector workers.
These schemes absorb a significant share of total government welfare
expenditures, benefiting a small proportion of the population, generally urban-
based higher income groups, and thus, in some cases, creating a regressive
distribution of welfare spending. Nonetheless, they have an important function
in developing social security systems, incorporating a role for enterprises or
employers and employees in financing, and potentially acting as a mechanism
for social cohesion.
The Philippines, Thailand, and Viet Nam maintain defined benefit pension
plans under which the amount of income received at retirement is dependent
on the number of years of contributions and on the level of individual earnings.
Indonesia and Malaysia operate defined contribution plans whereby the
contributions are saved over time and either paid out as a lump sum or as
pension-income stream at retirement (OECD and World Bank 2009).
Several countries in developing East Asia are moving toward forms of social
insurance that combine elements of contributions and general tax financing,
thus incorporating different levels of redistribution. Health care programs
are probably the major schemes of this type, with varying combinations,
including out-of-pocket payments and public support.
Social assistance programs in developing East Asia primarily involve cash
or in-kind transfers, both conditional and unconditional. In some countries,
these are based on long-standing systems; others were introduced and expanded
in response to crisis, or have subsequently been put into place. In Cambodia
and Lao PDR, large-scale cash-transfer programs are limited. Instead, social
assistance interventions are often project-based with limited coverage. These
include, for example, school feeding programs, scholarships, and other support
to keep vulnerable children in school.
In the PRC, social protection programs are proliferating. These include
expanded pension, health care, and unemployment benefits funded through
employer and employee contributions. Initially limited to a minority in formal
employment, options are being created for self-employed and migrant workers
to pay into such schemes. Basic social insurance or government-funded schemes,
particularly for health care, are being expanded.
After the number of workers laid off from bankrupt SOEs soared into
the millions during the 1990s, the PRC established the Minimum Living
Standard Guarantee (dibao) program. Rolled out nationally in 1999, the scheme
provides a basic cash transfer to households whose per capita income falls
below a locally determined minimum level. Funded through central and local
government revenues, this has since been expanded to rural areas. As of 2008,
about 23 million urban residents and 39 million persons living in rural areas
received support from the program (Yan 2009).
More recently, the PRC has expanded its basic medical coverage scheme in
rural areas (the New Cooperative Medical Scheme), funded through a mixture
of individual, local, and central government funding. For the poorest, though,
this does not provide sufficient support in the event of catastrophic health
events. It has been supplemented by a newly introduced medical assistance
program to support the poorest people with high costs of care, and the PRC
has now set itself the goal of providing universal basic health coverage to all
by 2020.
The PRC intends to spend US$124 billion on the first phase of a 10-year
overhaul of the health care system. Government outlays for health care
amounted to less than 1% of GDP in 2006, ranking the PRC 156 out of 196
nations surveyed by the World Health Organization. According to official
government figures, out-of-pocket payments rose from 20% of health-care
expenditures in 1978 to about 50% in 2006 (Fairclough 2009).
Like the PRC, Viet Nam has taken steps to extend its social security provision
through pension and health-care programs. The main poverty reduction
program is Program 135, which targets ethnic minorities and people living
in mountainous areas. The program supports local infrastructure development
through public works and access to health care. In addition to poverty reduction,
hunger eradication, and employment creation, particularly for disadvantaged
groups or areas, the strategy aims to expand insurance and social assistance
schemes for those in need, including people with disabilities or living with
HIV/AIDS2. Social protection coverage remains limited: in the case of health,
only about half the population is covered, and pension coverage is even more
limited. Health insurance is supplemented by the Health Care Fund for the
Poor, which aims to improve utilization and reduce the financial burden of
2
HIV = human immunodeficiency virus, AIDS = acquired immunodeficiency syndrom.
health care through the provision of free insurance to poor people and ethnic
minorities in disadvantaged parts of the country.
In Cambodia, the limited public resources aimed specifically at social
protection have gone to the formal sector, such as civil servant pensions and
veterans’ allowances. Initiatives to expand access of poor people to education
and health are part of wider poverty reduction efforts. Over 60% of budget
allocations for poverty alleviation go to rural areas, and include attention to
more and better health services, educational facilities, improved incomes
through rural livelihood activities, and upgraded rural infrastructure. However,
recognizing its vulnerability to international financial crises, the Government
of Cambodia is in the process of developing a national social protection strategy
for its poor and vulnerable citizens.
3.3 The Pacific

Social protection expenditure as a share of GDP averages 4.5% in the Pacific,
not far below the average for the Asia and Pacific region as a whole (4.8%).
However, the subregion’s average masks some striking extremes, ranging
from negligible social protection expenditure in Papua New Guinea and
only 1.1% in Vanuatu, to 13.5% in the Marshall Islands and 6.9% in Tuvalu.
Labor market programs account for a larger share of total social protection
expenditures in the Pacific (12%) than in the Asia and Pacific region as a whole
(7%) (ADB 2008).
Pacific Island countries are well-known for their informal community or
family-based systems of social support and reciprocity. Through such systems,
Pacific communities have developed a resilience to adversity (UNICEF 2009a).
A study of five Pacific Island countries—Fiji Islands, Kiribati, Samoa, Solomon
Islands, and Vanuatu—provides a detailed look at state-financed social protection
programs (Naidu and Moharty 2009). The study shows that governments in
each country finance some form of micro-enterprise development; provide
some disability benefits, survivor benefits, work-injury insurance, and maternity
benefits; offer some components of social assistance in the form of medical
rehabilitation, disaster victim assistance, and subsidized medial treatment;
facilitate micro loans; and manage child vaccination programs.
Yet the same study shows that none of the five countries has unemployment
insurance, social health insurance, or cash-transfer schemes (apart from family
assistance allowance in the Fiji Islands). Only the Fiji Islands provides social
funds, child maintenance, and scholarships for the disadvantaged. Of the
Box 3.2: Tonga Economic Support Program
In Tonga, strong family and community networks provide an important informal

social safety net system. Tonga has one of the highest remittance flows in the
world as a percentage of GDP—31% in 2008. Remittances provide a cushion
against shocks to household income, as well as funding important living expenses
including education fees, transport costs, and social obligations. As of June 2009,
remittances were down 14% year-on-year. As a result, the existing informal safety
net system in the country is under increasing stress.
Although the government has promoted universal access to health and education
services as a means to promote equality and improve living standards, it provides few
other formal social safety net programs. Instead, community and church organizations
are very active in providing social services, particularly to vulnerable groups including
people with disabilities, women suffering from domestic abuse, and children from
low-income households. The public sector has no services directed to the disabled
or the mentally ill, or for palliative care.
The government recognizes that areas for improvement remain in the setting of
priorities, targeting, and delivery of services used by vulnerable persons. Among
the planned steps is commissioning an expenditure-tracking system for health
and education to provide information for improved decision-making on targeted
expenditures. In parallel, the Ministry of Health will undertake a targeted review
of existing facilities and equipment to ensure that these support the provision of
services in the outer islands and districts, and community health centers.
In continuing to address the needs of the vulnerable, the government will put in
place delivery structures for a new community development program, the goal of
which is to reduce hardship and poverty by strengthening communities, civil society
groups, and organizations. The government will work in partnership with the program
to identify, prioritize, and address their own development needs. The expenditure
tracking and community development program will provide useful information for
the strengthening of policy formulation.
With the support of the ADB-financed Economic Support Program, an issues paper
on vulnerable persons will be prepared for consideration by the Government Cabinet.
The paper will identify vulnerable groups, the adequacy of social safety net systems,
and options for enhanced social protection policies.
Source: ADB (2009).

countries studied, Samoa is unique in operating state-run sickness insurance

and aged care programs. Only the Fiji Islands and Samoa provide assistance
for the homeless and nutrition programs for children. Assistance for homeless
young people is provided only in the Fiji Islands and the Solomon Islands.
The same study further notes that, across the five countries, national
provident funds are the most common type of social security system. However,
these only cover workers in the formal employment sector, thereby excluding
the many workers who make their living in the informal employment sector.
Most workers in the formal employment sector are men, leading to a large
gender disparity in who has access to national provident funds. Across the
surveyed countries, only a small proportion of workers have retirement income
coverage, ranging from 47% in the Fiji Islands to only 18% in Vanuatu. There
is limited coverage of maternity benefits across the five countries.
A separate report (ADB 2006) on social protection in the Fiji Islands,
Marshall Islands, Solomon Islands, and Vanuatu shows that these four countries
have health systems providing services free of charge or at low cost to all
residents, without the requirement to pay contributions. The main problem
with these systems is the lack of availability and quality of services provided.
The systems suffer from constant shortages of skilled personnel, appropriate
equipment, and drugs. The report also found that only the Fiji Islands has
a government income support safety net, the Family Assistance Program.
However, this program has an insufficient budget, meaning that families
meeting the criteria for benefits do not always receive them.
3.4 South Asia

South Asian countries spend about one-third less for social protection as a
share of GDP than the Asian average (3.1% of GDP, versus the Asian mean of
4.8%). In South Asia, microcredit schemes account for a much larger share
of total social protection expenditures compared with the region as a whole
(26% versus 13%) (ADB 2008).
UNICEF has published a detailed review of social protection measures in
Afghanistan, Bangladesh, Bhutan, India, Maldives, Nepal, Pakistan, and Sri
Lanka (UNICEF 2009b). Selected highlights from this report are reviewed
below, providing a good look at the situation in South Asia (Table 3.3).
All countries in South Asia have established some components of social
protection, even if social protection is not yet a broadly based notion in the
sense of an entitlement that people claim as citizens. The region has a long
Table 3.3: Social Protection in South Asia: Some Commonalities
Social
Security Social Assistance
General Social Sectoral Social Assistance— Emergency

Country Formal Sector Assistance Transfers in Cash and Kind Transfers
Sickness, Health- Education- Transfers to
unemployment, Poverty- related related cope with
old age, health, related: transfers transfers Employment- shocks,
insurance (e.g., (universal Child benefit (e.g., (e.g., meals related transfers conflict,
public service, or means (e.g., girl maternity at school, (e.g., public and natural
formal sector) tested) child grants) benefits) stipends) works schemes) disasters
Afghanistan ¸ ¸ ¸ ¸
Bangladesh ¸ ¸ ¸ ¸ ¸ ¸
Bhutan ¸ ¸ ¸
India ¸ ¸ ¸ ¸ ¸ ¸ ¸
Maldives ¸ ¸ ¸ ¸
Nepal ¸ ¸ ¸ ¸ ¸ ¸ ¸
Pakistan ¸ ¸ ¸ ¸ ¸ ¸ ¸
Sri Lanka ¸ ¸ ¸ ¸ ¸ ¸
Source: UNICEF (2009b).
history of formal sector social insurance. Several countries have adopted

universal entitlements to some forms of social assistance. Bangladesh, India,
Maldives, and Nepal have social pensions, and the latter two are universal.
The rural employment schemes in India, Bangladesh, and Nepal establish
a right to a minimum number of days of employment, although in the
case of Bangladesh, the entitlement applies only to specific geographic
locations.
India’s National Rural Employment Guarantee Act (NREGA) provides a
legal guarantee for 100 days of employment in every financial year to adult
members of any rural household willing to do unskilled manual work at the
statutory minimum wage. The central government funds the cost of wages,
75% of the cost of materials, and part of the administrative expenses; the
remaining costs are funded by state governments. Participants in NREGA
work in projects such as water conservation, flood control, irrigation, and land
development (United Nations Economic and Social Commission for Asia and
the Pacific [UNESCAP] 2009).
The program has many benefits, including the reduction of distress migration,
employment generation in the most distressed areas, and improvements in
the natural resource base of livelihoods in poor communities. NREGA is
an example of a self-targeted program, characterized by low administration
costs and high accuracy in reaching the targeted population. By guaranteeing
income support to anyone willing to participate in unskilled manual labor
at low wages, and only for 100 days in a year, the program ensures that only
the neediest will apply (UNESCAP 2009). Central government outlays for
schemes under NREGA total an estimated US$8 billion for the fiscal year
2009–2010.
Formal, compulsory social security is available throughout South Asia for
government civil service, staff in public sector enterprises, and employees of
large-scale private enterprises. This is in the form of a guaranteed pension and
other coverage such as health care benefits. Such formal social security, however,
excludes the vast majority of people in South Asia, who are not connected to
government or formal employment. In India, for example, 85% of the adult
population work in the informal sector.
In India, the long-time policy approach to social security provision,
particularly for old age and workers in the formal economy, has been
through large-scale administrative institutions of the provident fund type,
supplemented by social insurance created mainly to provide rapid access for
workers to relevant health-care services. More recently, India has partially
transformed its provident fund into a social insurance pension scheme. The
main deficiency of this approach in practice has been the failure of outreach
to the vast majority of workers in the unorganized sector. Further, the level of
health funding through social security in India remains very low (International
Labour Organization [ILO] 2008).
No government in South Asia has established a full-fledged, comprehensive,
and interlinked social protection system. Therefore, in most instances,
households and individuals in South Asia primarily rely on informal social
protection networks of family, community, women’s groups, or savings
cooperatives, such as rotating chit funds (based on money pooled by a group
of people and auctioned at the end of a specified period, rotating among the
group members), and informal credit markets.
Nevertheless, the situation continues to evolve, with many government-
based social protection instruments coming onstream over the past decade
Box 3.3: Social Protection in Pakistan
A 2007 World Bank report observed that Pakistan has many social protection
programs, ranging from cash transfers to pensions. The country’s social assistance
has included two main federal cash transfer programs, Zakat and Bait-ul-Mal, and
small scattered programs that provide social welfare and care services to persons
with disabilities, child laborers, and others. Zakat, a common practice in the Muslim
world, is a welfare contribution to poor or disadvantaged persons. It is based entirely
on voluntary contributions from wealthy individuals, and uses community structures
to deliver benefits. In contrast, Bait-ul-Mal, introduced in the early 1990s, is funded
and administered by the government. In addition, micro finance programs provide
poor people with access to credit.
Although Pakistan previously implemented public works/workfare programs, no

large workfare program is currently in place. To address aggregate economic (price)
shocks, Pakistan has a wheat-subsidy program. Although no permanent program
is in place to help individuals cope with aggregate disasters, Pakistan has used a
combination of cash transfers, housing, and social service programs to help those
affected by the 2005 earthquake.
Pakistan’s social security system offers pension (old-age, survivor and disability)
benefits to formal sector workers. Public sector workers are provided with civil service
pensions, while private sector workers enjoy access to pensions from the Employees
Old Age Benefits, but also provincially based pension and non-pension programs,
such as the Workers Welfare Fund and Employees’ Social Security Institutions.
More recently, Pakistan has introduced a third major federal cash transfer program,
the Benazir Income Support Program.
or so. In fact, the South Asian region has a long history of formal sector social
security, and there is also a set of non-contributory social assistance-type
transfers taking many different forms. Historically, many of these schemes
have their roots in the independence and post-war welfarist tradition, and, in
some cases, in the British-influenced public administrative service.
Social assistance in the form of in-cash or in-kind transfers not based on
prior direct contributions takes many forms. It includes social pensions for
people living below the poverty line; a few child grants; and sector-specific
transfers, such as education stipends, health-related benefits, and food transfers.

There are also a number of ad hoc or program-based emergency transfers to
cope with conflict and natural disasters, and even housing loans.
The genesis of these social protection elements varies considerably. Most
countries have a “building blocks” approach to social protection, with subsets
of social protection as stand-alone interventions. In the past, governments
rarely made concerted efforts to categorize the interventions by program and
type of support provided, and to bring all schemes under a common umbrella.
However, this is now beginning to change as countries start to reorganize
their social protection programs.
Many social transfers are long-term established programs; others are ad hoc
responses to immediate needs and demands, of which some reflect political
Box 3.4: Social Pensions in Bangladesh
Bangladesh is home to the third highest number of poor older people in the world
(after India and the PRC). The Boishka Bhata, or old-age allowance, started in
1998. The scheme is designed to reach the oldest and poorest 20 people in each
ward (rural district).
The old-age allowance is means tested and recipients are selected by the community.
Eligibility starts at age 57, and half of the recipients have to be women. Additional
criteria for selection are a lack of land or an annual income below BDT3,000 (US$43),
chronic poor health, and inability to work.
The pension of BDT165 a month (US$2.35) is paid in quarterly installments. Recipients

collect their pension from local branches of the government-run Sonali Bank. Each
recipient has a passbook and the Social Welfare Office keeps a register with sample
signatures of all recipients.
The scheme is administered by the Ministry of Social Welfare and financed out of
the state budget. The old-age allowance makes up 0.03% of the GDP. In 2006, 1.3
million people received the allowance.
Research shows that the old-age allowance is spent on basic needs such as food
(60%), health care (30%), and income-generating activities (10%), such as tea-stalls,
handicraft businesses, goat-rearing, and growing vegetables.
Source: HelpAge International (2009).

opportunism as opposed to a strategic approach to providing social protection
as a right. Some schemes are localized, whereas others are national in outreach.
In terms of design, some social protection programs are subject to means tests
or proxy means tests; others use geographical or categorical targeting.
Among the programs targeting by category, some are universal for an age
group or gender, such as old-age pensions in Bangladesh, India, Maldives, and
Nepal, or the girls’ secondary school stipends in Bangladesh. Most benefits are
conditional on action by the recipient, usually relating to working or school
enrollment, or to utilizing a health-care facility.
Social protection programs in South Asia often overlap. Indeed, the coexistence
of many different programs of varying scale, scope, benefits package, eligibility,
and duration generates high administrative and transaction costs for both the
government and program beneficiaries, and increases the risk of eligible groups
not being aware of their entitlements and how to claim them.
4. The Economic Slowdown and Social Protection

The global economic slowdown has had major negative impacts in Asia and
the Pacific, exacerbating already serious social problems. A sharp drop in
exports to developed country markets overseas has triggered a surge of plant
closing and layoffs in industries that manufacture electronics, clothing, and
a variety of other goods for foreign markets.
The result has been a rise in poverty spurred by job losses, fewer working
hours, and downward pressure on remuneration. The number of workers in
vulnerable employment could rise by 52 million, amounting to a total of more
than 1.1 billion. Vulnerable employment is a major challenge, underscoring
the need to include social protection measures in crisis response packages. In
South Asia, the number of vulnerable workers is projected to rise to almost
79% of all workers, or 493 million people. In South East Asia and the Pacific,
the share of workers in vulnerable employment could rise to around 64.4%, or
182 million people, and in East Asia to 56.6%, or 458 million people (ILO 2009).
Working poverty is also a concern. According to the most pessimistic scenario,
the number of people working in the Asia and Pacific region but living in
households that survive on less than US$1.25 per person per day could rise to
589 million, or more than two-thirds of the global total. South Asia is home
to the largest number of working poor, 382 million workers, or 61% of the
workforce (ILO 2009).
The ecnomic slowdown has tended to impact the near poor more than
those in chronic poverty, including those working in low-paying industries.
Young urban workers, those in export-dependent industries, and migrants
are among those hard hit. Very poor people are also affected through spillover
effects. Informal workers are in a particularly precarious position as only
a small minority have social security coverage, and they have experienced
greater job competition from workers laid off in the formal sector.
In the PRC, researchers at the Chinese Academy of Social Sciences (CASS)
estimate that up to 41 million workers lost their jobs during the economic
slowdown (CASS 2008a), and the urban jobless rate reached 9.4%, more than
twice as high as official estimates (CASS 2008b). About 670,000 SMEs closed
during the slowdown (Ping 2009). The Philippines lost an estimated 950,000
jobs between the fourth quarter of 2008 and the first quarter of 2009 due to
the slowdown (Son and San Andres 2009).
The social impacts of the slowdown have had different implications for
men and women. The difficult economy has often forced women to take
on additional, informal, and even degrading work to make ends meet. In
Cambodia, for example, 70,000 workers lost their jobs in the garment industry,
most women and typically single females in their twenties whose rural families
depend upon them for remittances. Many of those laid off have sought work
in the entertainment industry, putting them at risk of exploitation and abuse.
Enterprises surveyed in the Indonesia furniture export industry showed quite
distinct gender-based patterns. As women workers were usually involved in
aspects of the production process that were considered less critical, such as
packaging, they were the first to lose their jobs (Durano and Hung 2009).
In Thailand, female workers were particularly hard hit by the slowdown,
with consistently lower employment in industry in 2008 and 2009, whereas
their male counterparts experienced significant reductions in employment only
in the first quarters of 2008 and 2009. This could be because female workers
in industry are more likely to accept contractual jobs and would thus be the
first to be laid off in times of economic trouble (Son and San Andres 2009).
In the Asia and Pacific region as a whole, growth rates have bounced back,
aided by fiscal stimulus packages launched by many governments. Large-scale
public investment programs have provided more fiscal space for infrastructure
investment, tax cuts, and enterprise promotion. However, comparatively
little has gone toward increased social spending. Nonetheless, all of the fiscal
stimulus packages have included some social protection components. The share
of each package varies widely between low- and middle-income countries, from
Table 3.4: Summary of Social Protection Components of Fiscal Stimulus
Package
Social Protection Components Social Protection Components Calculated % of

Country (Local Currency Units, Billions) (US$, Billions) Stimulus Package
Bangladesh BDT3.74 0.05 12.81%
China CNY920 134.61 23.00%
Georgia GEL1.60 0.96 42.11%
Indonesia IDR6,077 0.55 8.77%
Japan JPY7,500 82.64 27.78%
Korea KRW7,700 6.11 11.46%
Malaysia MYR14.40 4.16 34.29%
Philippines PHP50.00 1.05 15.15%
Singapore SGD7.70 5.35 52.38%
Taipei,China NTD234.00 7.15 46.82%
Thailand THB367.57 10.53 23.45%
Viet Nam VND24,009.80 1.41 16.79%
Source: Zhang, Thelen, and Rao (2009).
under 9% in Indonesia to 42% in Georgia. The available fiscal space is fast

disappearing, however, as governments will not be able to maintain massive
increases in public expenditures (Zhang, Thelan, and Rao 2009).
Table 3.4 shows the share of the stimulus packages that have gone toward
social protection in several Asian countries.
In contrast to the 1997–1998 Asian financial crisis, few countries have
reacted to the most recent economic slowdown to undertake major policy
reforms in the social arena, e.g., to expand the breadth of social protection,
embrace vulnerable groups not already covered, improve targeting, or introduce
novel mechanisms. One exception is Mongolia, where ADB is helping the
government to reduce out-of-pocket expenditures of poor households for health
care, and increase micronutrient consumption to combat malnutrition. In
addition, policy measures adopted with the program’s support will improve
the sustainability of social expenditures through better targeting and
rationalization of social transfers in the education and health sectors.
Also in Mongolia, ADB and the Government of Japan are providing a

US$3 million grant to design and implement a medicard program that will
enable poor people to obtain free health services. One component of the
grant program involves the distribution of micronutrients to at least 15,000
children under 3 years of age in eight aimags (provinces). This component
will complement existing government programs in Mongolia’s other 18
aimags, thereby resulting in nationwide coverage of micronutrient supplement
distribution.
In Cambodia, ADB is helping the government address rural food insecurity
issues through the US$35 million Emergency Food Assistance Program. The
loan and grant program includes a range of safety net elements, including
targeted in-kind transfers (rice for poor people), food for students at school,
a rice allowance for volunteer preschool teachers, and a labor-intensive public
works program that has tested giving both cash and food for labor in small-
scale community infrastructures, such as rural roads and culverts.
Among the major social programs launched during the crisis is Pakistan’s
Benazir Income Support Program (BISP). Initially funded with PKR34 billion
(about US$400 million) for the period 2008–2009, the BISP represents the
third largest allocation in the central government budget (0.3% of GDP).
The program was initiated in part to offset the impact of inflation on the
purchasing power of poorer sections of society. BISP aims to cover almost
15% of the country’s population (40% of the population below the poverty
line). A monthly payment of PKR1,000 (US$12) per family would increase
the income of a family earning PKR5,000 by 20%. The BISP is expanding its
current coverage of 3.5 million families to 5 million families in 2010, with
the aim of reaching 7 million families by 2011.
The Philippines has expanded key social protection programs such as
the Pantawid Pamilyang Pilipino Program, which now provides conditional
cash grants to about 1 million poor households. Additional funding was also
designated for the Self-Employment Assistance Kaunlaran, which provides
capital assistance and a capacity building program for livelihood projects, a
rice allocation program for elementary school children, and a micro-finance
lending program.
In Viet Nam, the government has boosted spending on existing social
assistance initiatives, such as Program 135, as well as the health insurance
program for those in poverty or just above the poverty line.
Although the global slowdown has brought few major structural reforms
in the region’s social protection systems, it has given momentum to changes
Box 3.5: How Social Protection Can Help Response to the

Slowdown
Social protection interventions can help address the global economic slowdown
and underpin other investments in development in different ways. In particular:
They can provide effective instruments for reducing poverty and destitution in
many countries.
By supporting consumption by poor people, they can complement macro-economic

policies aimed at tackling the contractionary impacts of the global slowdown.
By maintaining and building human capital and reducing social risk, they promote
long-term human capital development, livelihoods, employment, and economic
risk.
By providing poor people with a stake in the economy, they promote social cohesion
and facilitate the implementation of other necessary reforms.
Source: Samson (2009).
under consideration. It has also stimulated thinking among decision-makers

about what kind of system improvements need to be made to ensure sustainable
financing of social protection programs. The slowdown has highlighted that
certain social protection measures can quickly help those most affected by
economic crisis. There is also growing recognition that increasing social
expenditures can contribute to rebalancing of the economy, and that social
protection is an investment in human capital—a key element of a competitive
economy enjoying sustainable growth.
5. Emerging Challenges
There are a variety of serious challenges confronting the further development
of social protection systems in developing Asia and the Pacific. Administration,
finance, governance, and technological support of social protection schemes
are among such challenges. Others include reaching remote rural and ethnic
minority populations, and ensuring gender equity.
Three additional challenges now confront policymakers considering

improvements to social protection systems: climate change, labor migration,
and an aging population.
5.1 Climate Change and Impact on Vulnerable People

The global economic slowdown poses significant risks for the environments
of poor people. High rates of joblessness and underemployment have forced
millions of urban dwellers in the Asia and Pacific region into poverty. Many have
moved into overcrowded slums, further degrading the natural environment,
and piling pressure on grossly inadequate urban infrastructure. Over a longer
time frame, climate change is making most environments of poor people less
habitable, adding to existing pressure on natural resources.
Protracted episodes of climate stress can contribute to commodity price
increases and volatility, disproportionately affecting poor and vulnerable
people, as was the case in the 2008 food crisis (Ivanic and Martin 2008).
High food prices increase poverty, worsen nutrition, reduce use of health
and education services, and deplete the productive assets of the poor (Grosh
et al. 2008). Climate change effects could displace people from their homes,
prevent the affected from engaging in their usual livelihood activities for
survival, orphan children, and put women and children at risk of exploitation
and abuse. People employed in the agricultural sector will be among the worst
affected (UNICEF 2009a).
Social assistance programs will need to be strengthened substantially to
cope with the challenges posed by climate change. Many low-income countries
cannot afford permanent transfers to their poor people, but scalable safety
nets that provide basic forms of noncontributory insurance can offer a core
social protection that prevents morbidity, mortality, and excessive depletion
of assets in poor countries where these safety nets have not commonly been
provided (Alderman and Haque 2006; Vakis 2006).
5.2 Labor Migration and Pension Arrangements

Asia is the largest source of temporary contractual migrant workers worldwide,
while simultaneously being characterized by very large intraregional flows of
migrant workers, particularly the vast internal movements in India and the
PRC. In all, the Asia and Pacific region hosts more than 61 million migrants
(International Organization of Migration [IOM] 2008).
Cross-border migration makes migrants who lack entitlements based on
citizenship extremely vulnerable. They also tend to find employment in the
informal economy and temporary contractual jobs. Thus, migrants generally
fall outside any system of social protection, however meager, and may lack
legal status. International migration is an important livelihood strategy for
many low-income households in the region, and an important source of income
for the households and the economy they come from. Most migrants move
within the region, but the majority are believed to be undocumented and not
included in the available official statistics. Southeast Asia has a significant
amount of trafficking, particularly within the Mekong region, associated with
the sex and drug trades, making women and children particularly vulnerable
(Cook 2009).
Due to their particular circumstances, especially the duration of their
employment and residence, migrant workers holding jobs abroad often cannot
obtain coverage under social protection schemes. They risk loss of entitlement
to social security benefits in their country of origin due to their absence, and
also may confront restrictive conditions under the social security system of
their host country, leaving them especially vulnerable to social risks.
Migrants’ families—whether migrating with the main breadwinner or left
behind—often face heightened vulnerability to poverty and risks, especially
if the main breadwinners are low-skilled migrants who earn little. Given the
scope and the scale of the problem, social protection for this group is an urgent
policy issue, and shows the interface of lack of decent work opportunities and
social protection in home countries that drive low-income migration in the
first place, and the gap in international, regional, and national provision for
migrants (UNICEF 2009a).
Internal migrants within the PRC who move from rural areas to a city
typically lack medical insurance and thus resort to using inexpensive,
unlicensed clinics for their health needs. These clinics often do not have
qualified doctors and use out-of-date medicines. There are reports of patients
developing serious infections after receiving inappropriate treatment in
unlicensed clinics. An investigation found more than 1,200 illegal and unsafe
medical clinics in Beijing alone (China Daily 2009). Workers who are paying
premiums for health insurance in one county can find themselves denied
benefits after migrating to another country where they are not registered
residents (Fairclough 2009).
There is evidence that the social security needs of foreign workers are
starting to be taken more seriously. In response to pleas from Filipinos
working overseas, the Philippine Social Security System took major steps to
provide voluntary social security coverage for such workers. Through monthly
contribution remittances made at an authorized foreign bank or remittance
center, the covered overseas worker is entitled to benefits granted to voluntary
members of the social security system. In addition, since 2002, the social
security system introduced the Flexi-Fund Program, which is a tax-exempt
provident fund featuring flexible payment terms and easy withdrawal of
savings. In Korea, a 2007 amendment to legislation on national pensions
allows foreign workers who pay into the pension system to receive a lump-
sum refund of their contributions when they return to their home country
(ISSA 2009).
Japan, meanwhile, has agreed with the Philippines and Indonesia on
special arrangements to admit hundreds of nurses and caregivers annually
from each country. These labor arrangements may be replicated by other
countries with aging populations, since they provide an orderly way of
introducing foreign workers needed to address growing social and health-
care needs.
5.3 Aging Population and Pension Schemes

Old-age income support will be one of the biggest social and economic
challenges facing developing Asia and the Pacific in the 21st century. The
growing spotlight on old-age income support is largely due to a seismic
transition that is fundamentally reshaping Asia’s demographic profile. A young
continent reaping the demographic dividend of a large youthful workforce
is giving way to a graying continent where the ratio of retirees to workers is
on the rise. In contrast to industrialized countries, most Asian countries do
not yet have mature, well-functioning pension systems with wide coverage.
As a result, they are ill-prepared to provide economic security for the large
number of retirees in the future. It is evident that the entire region will have
a drastically different, much older, demographic profile by 2050. The region’s
demographic transition is driven by falling fertility and rising life expectancy
(Park 2006).
A constellation of economic and social factors, such as improved female
education and better medical care, is inducing Asians to have fewer children
and enabling them to live longer. Other demographic indicators also point
unequivocally toward a graying continent. The median age of the PRC,
Indonesia, Korea, Malaysia, Singapore, Thailand, and Viet Nam will exceed
the world average by 2050. Furthermore, life expectancy at 60 is already fairly
high, and by 2050, fertility rates will fall below levels required for a stable
population (Park 2006).
In the PRC, the number of persons older than age 60 is expected to
increase from about 166 million in 2008 to 342 million by around 2030.
They will be particularly vulnerable to poverty given the high dependency
ratio that has resulted from its one-child population policy. In India, the
number of persons older than age 60 is expected to pass 180 million by
2030 (ISSA 2009).
The weakening of informal family-based old-age support mechanisms
suggests a greater role for formal pension systems throughout the region.
Asians have traditionally relied upon their children to take care of their
material needs in their old age. The family network has essentially provided
the region with a pension system, especially in rural environments, where
extended families of three generations have often lived together under one
roof, with younger family members supporting older family members.
However, the far-reaching social changes accompanying the region’s
economic progress have given rise to smaller nuclear families, which are less
conducive to intrafamily support. Such changes include rapid urbanization
and a decline in the relative importance of agriculture in the economy. In
short, urbanization, industrialization, and sociocultural changes are creating
a vacuum in Asia’s old-age support, a vacuum that must be filled by formal
pension systems (Park 2006).
Across the region, changes to family structures, working patterns, and
traditional cultural values suggest that innovative, and integrated and coherent
public policy responses are sorely required to address the social security
and health-care needs of aging populations. Increasing average longevity
demands greater financial resources to ensure the payment of adequate old-
age benefits for longer periods. The need for greater financial resources for
existing programs has to be balanced carefully with any extension of coverage
and greater benefits (ISSA 2009).
Financing longer periods of retirement and extending coverage to the elderly
implies a number of possible responses, including delaying access to benefits and
savings, prolonging the period of active contribution, improving the efficiency
of investments and administration, limiting the generosity of defined benefits,
identifying priorities for fiscal expenditure, and complementing retirement
income with innovative and voluntary income sources (ISSA 2009).
6. Conclusion
Developing countries in the Asia and Pacific region have made remarkable
progress in reducing the percentage of people living in absolute income poverty.
However, the economic growth that has successfully reduced the incidence
of poverty has also been accompanied by growing inequality. Further, given
population growth rates, very large numbers of people in the region remain in
poverty, in conditions vulnerable to poverty, or without access to basic social
services. Realization has grown among policymakers that more active and
comprehensive social protection measures are required to forge a sustainable
and inclusive pattern of growth.
The recent global slowdown has made clear that Asian and Pacific countries
need to reduce their economic dependence on importers in Europe and North
America, and boost domestic consumption and intra-regional trade. Yet efforts
to boost consumption will run up against a deeply ingrained tradition of
high savings. It will take time for more comprehensive and dependable social
protection systems to be constructed, and to gain the trust of the people.
In any case, the incentives to strengthen social protection systems have
become greater with time and the region’s development. Recovery from
the global economic slowdown will be accompanied by the introduction of
new social protection initiatives. The effective targeting of social assistance
programs, and long-term financial sustainability of pension schemes, will be
among the issues receiving attention in deliberations on how to reduce social
vulnerability and risk.
References
ActionAid. 2009. Who’s Really Fighting Hunger? Report. Johannesburg: ActionAid.
Available at: http://www.actionaid.org/docs/hungerfree%20scorecard%20-%20
final%20without %20embargo.pdf.
Adema, W., and M. Ladaique. 2009. How Expensive Is the Welfare State? Gross and Net
Indicators in the OECD Social Expenditure Database (SOCX). OECD Social,
Employment and Migration Working Papers No. 92. Paris: OECD.
ADB. 2001. Social Protection Strategy. Manila: ADB.
————. 2006. TA 6207: REG—Pacific Regional Social Protection Systems Assessment. Manila:
ADB.
————. 2008. Social Protection Index for Committed Poverty Reduction, Volume 2: Asia.
Manila: ADB.
Alderman, H., and T. Haque. 2006. Countercyclical Safety Nets for the Poor and
Vulnerable. Food Policy 31(4): 372–383.
CASS. 2008a. Green Book on Population and Labour. September. Beijing: CASS.
————. 2008b. Society Blue Paper. December. Beijing: CASS.
China Daily. 2009. Illegal Clinics Put Patients at Risk. 20 October.
Cook, S. 2009. Social Protection in East and South East Asia: A Regional Review. Social
Protection in Asia Working Paper 02. Brighton, UK: Institute of Development
Studies.
Cook, S., and N. Kabeer. 2009. Socio-economic Security over the Life Course: A Global
Review of Social Protection. Report. Brighton, UK: Institute of Development
Studies.
Durano, M., and S. Hung. 2009. Lessons Not Learned? Gender, Employment and Social
Protection in Asia’s Crisis-Affected Export Sectors. Paper prepared for The Impact of
the Global Economic Slowdown on Poverty and Sustainable Development in Asia
and the Pacific conference, 28–30 September, Ha Noi. Available at: www.adb.org/
Documents/Events/2009/Poverty-Social-Development/gender-impact-crisis-Hung-
paper.pdf.
Fairclough, G. 2009. China Prescribes Fixes for Ailing Health-care System. Wall Street
Journal. 16–18 October.
Grosh, M.E., C. del Ninno, E. Tesliuc, and A. Ouerghi. 2008. For Protection and Promotion:
The Design and Implementation of Effective Safety Nets. Report. Washington, DC:
World Bank.
HelpAge International. 2006. Why Social Pensions Are Needed Now. Available at: www.
helpage.org/download/4c48d8bb4513c.
————. 2009. Social Pensions in Bangladesh. Available at: http://www.helpage.org/
Researchandpolicy/Socialprotection/PensionWatch/Bangladesh.
ILO. 2008. Income Security for the Elderly and Not-so-elderly in the Asia-Pacific Region.
Paper presented at the Asia-Pacific Regional High-Level Meeting on Socially
Inclusive Strategies to Extend Social Security Coverage, 19–20 May, New Delhi.
————. 2009. Global Employment Trends – Update. May. Geneva: ILO.
IOM. 2008. Managing Labour Mobility in the Evolving Global Economy. Geneva: IOM.
ISSA. 2009. Dynamic Social Security for Asia and the Pacific. Geneva: ISSA.
Ivanic, M., and W. Martin. 2008. Implications of High Global Food Prices for Poverty in
Low-income Countries. Policy Research Working Paper 4594. Washington, DC:
World Bank.
Müller, K. 2003. Poverty and Social Policy in the Central Asian Transition Countries. DIE
Working Papers 6/2003. Bonn, Germany: German Development Institute (DIE).
Naidu, V., and Moharty, M. 2009. Situational Analysis of Social Protection Policies,
Services and Delivery Mechanisms in the Pacific. Paper presented at the UNRISD,
USP, UNICEF, Commonwealth Secretariat, and UNESCAP Conference on Social
Policies in Pacific Island countries: Policy Recommendations and Impact of
Current Crises. Suva, Fiji Islands, 22–23 July.
OECD and World Bank. 2009. Pensions at a Glance. Asia/Pacific Edition. Paris: OECD and
World Bank. Available at: http://www.oecd.org/dataoecd/33/53/41966940.pdf.
Park, D. 2006. Ageing Asia’s Looming Pension Crisis. July. ADB Economics Working Paper
165. Manila: ADB.
Ping, Z. 2009. Starved for Cash. China Daily. 15 September. Available at:
http://www.chinadaily.com.cn/bizchina/2009-09/15/content_8694955.htm.
Plan International. 2009. Universal Birth Registration, Facts and Figures. Website.
Available at: http://plan-international.org/birthregistration/the-campaign/facts-
and-figures.
Samson, M. 2009. The Implications of the Global Financial Crisis for Low-income Countries.
Cape Town: Economic Policy Research Institute.
Son, H. H., and E. A. San Andres. 2009. How Has Asia Fared in the Global Crisis? A Tale of
Three Countries: Republic of Korea, Philippines, and Thailand. ADB Economics
Working Paper 174. Manila: ADB.
UNESCAP. 2009. Economic and Social Survey of Asia and the Pacific 2009: Addressing the
Triple Threats to Development. Data survey. Bangkok: UNESCAP. Available at:
http://www.unescap.org/survey2009/.
UNICEF. 2009a. Protecting Pacific Island Children and Women During Economic and Food
Crises. New York: UNICEF.
————. 2009b. Social Protection in South Asia: A Review. New York: UNICEF.
Vakis, R. 2006. Complementing Natural Disasters Management: The Role of Social
Protection. Social Protection Discussion Paper 0543. Washington, DC: World
Bank.
Wood, J. 2009. A Social Protection Index for Asia. January. Conference proceedings of
Association for Public Policy Analysis and Management 2009. Available at:
https://www.appam.org/conferences/international/singapore2009/sessions/
downloads/2301.pdf.
World Bank. 2007. Social Protection in Pakistan: Managing Household Risks and
Vulnerability. Report. Washington, DC: World Bank.
————. 2008. Social Protection and Economic Shocks in ECA: The Social Side of the
Global Crisis. Report. Washington, DC: World Bank.
————. 2009. IDA at Work – Pakistan: Punjab Education Reform Increases Access.
Website. Available at: http://web.worldbank.org/WBSITE/EXTERNAL/
EXTABOUTUS/IDA/0,,contentMDK:21396362~menuPK:3266877~pagePK:51236
175~piPK:437394~theSitePK:73154,00.html.
Yan, M. 2009. Impact of Social Policies on Household Income and Savings: China Case.
Paper presented at the ADBI Annual Conference, 4 December 2009, Tokyo.
Zhang, Y., N. Thelen, and A. Rao. 2009. Social Protection in Fiscal Stimulus Packages:
Some Evidence. UNDP/ODS Working Paper (September). New York, NY: UNDP.
Household Saving Rates and Social Benefit Ratios: Country Comparisons 63
Country Comparisons
IV
Household Savings Rates and Social Benefit Ratios:
Charles Yuji Horioka and Ting Yin
1. Introduction
It is often asserted that social safety nets will reduce household savings rates
because households will not feel the need to save (self-insure) if social safety nets
are adequate. This chapter presents multi-country data on household savings
rates and social safety nets and analyzes the determinants of differences in
household savings rates among countries with an emphasis on the impact
of social safety nets, the age structure of the population, and borrowing
constraints on the population.
The chapter is organized as follows: In Section 2, we survey the theoretical
and empirical literature on the impact of social safety nets on the household
savings rate. In Section 3, we present data on household savings rates for OECD
member countries. In Section 4, we present data on the social benefit ratio
(the ratio of social contributions and social benefits receivable to household
disposable income) for the OECD member countries. In Section 5, we examine
whether and to what extent the household savings rate and the social benefit
ratio are correlated. In Section 6, we conduct a regression analysis of the
determinants of the household savings rate. Section 7 concludes the chapter
and discusses the policy implications of the findings.
This chapter finds that there are considerable and stable differences
among countries in their household savings rates and social benefit ratios
The authors are grateful for valuable comments from Mukul Asher, Davide Furceri, Masahiro Kawai,
and the other participants of ADBI’s Annual Conference.
but that the latter can explain the former only to a limited extent, with
the age structure of the population and borrowing constraints being more
important as determinants of differences among countries in household
savings rates.
2. Survey of the Literature
In this section, we summarize the theoretical and empirical literature on

the impact of social safety nets on household saving. Looking first at the
theoretical literature, most of this literature has focused on the impact of
public old-age pensions on household saving. For example, the seminal
paper on this topic by Feldstein (1974) showed that the impact of public
old-age pensions on household saving is theoretically ambiguous. On the
one hand, the introduction of a public old-age pension system will induce
households to save less because they no longer need to rely as much on
their own savings to finance living expenses during retirement (the wealth
replacement effect). On the other hand, the introduction of a public old-age
pension system will induce workers to retire earlier, and this will induce
them to save more (the induced retirement effect). The net impact of public
old-age pensions on household saving will depend on the relative strengths
of these two offsetting effects.
The literature on the impact of other components of the social safety net
on household saving is much scarcer. The noteworthy paper on this topic by
Hubbard, Skinner, and Zeldes (1995) demonstrated theoretically that social
insurance programs with means tests based on assets discourage saving by
households with low expected lifetime incomes. Thus, not only public old-
age pensions but also other components of the social safety net may have a
negative impact on the household savings rate.
Turning to the empirical literature, we survey previous multi-country
analyses of the impact of social safety nets on household (or private) saving.
The vast majority of these previous studies have focused on the impact of
public old-age pensions on household saving, and most of them have found that
public old-age pensions have a negative and significant impact on household
saving. For example, Feldstein (1977, 1980) and Bailliu and Reisen (1998)
obtained such findings for a sample of developed countries, while Edwards
(1996) and Dayal-Ghulati and Thimann (1997) obtained such findings for a
sample of developing countries. A major exception is Modigliani and Sterling
(1983), who found that the impact of public old-age pensions on private saving
is ambiguous because a smaller than expected wealth replacement effect was
more than offset by a larger than expected induced retirement effect. Another
exception is Horioka (1989), who found the impact of public old-age pensions
on private saving to be insignificant because neither of the two effects was
found to be significant.
To summarize, most previous multi-country studies have analyzed the
impact of public old-age pensions on household (or private) saving and have
found a negative and significant impact, but few multi-country studies have
analyzed the impact of other social programs.
3. International Comparison of Household Savings Rates

In this section, we present data on household savings rates in the OECD
member countries for which data are available. Table 4.1, which shows data
for 1985, 1996, and 2007, illustrates that there has been tremendous variation
in household savings rates among the OECD member countries, with the
household savings rate ranging from –3.3% in Norway to 21.5% in Italy in
1985, from 0.3% in Finland to 17.9% in Italy in 1996, and from –1.6% in
Finland to 12.2% in France in 2007.
However, broad differences among countries in household savings rates
have been relatively stable over time. For example, if we focus on the years
1996 and 2007, the household savings rate was relatively high in many of the
countries in Continental Europe (e.g., Austria, Belgium, France, Germany,
Hungary, Italy, Netherlands, Poland, Spain, and Switzerland) and relatively
low in many of the countries of Scandinavia (Northern Europe) (e.g., Denmark,
Finland, and Norway), many of the Anglo-Saxon countries (e.g., Australia,
Canada, Ireland, UK, and US), some of the countries of Eastern Europe (e.g.,
Czech Republic and Slovak Republic), and Portugal. The only countries showing
wide fluctuations in their household savings rates were Japan, Korea, and
Sweden.
Note, finally, that household savings rates in the OECD member countries
have shown a long-term downward trend, with the average household savings
rate for the OECD member countries for which data are available declining
from 9.1% in 1985 to 8.9% in 1996 and 5.2% in 2007, a decline of some 43%
in just over 20 years.
Table 4.1: Household Savings Rates in OECD Member Countries
Country 1985 1996 2007

Household Household Household
Savings Rate Savings Rate Savings Rate
(% of disposable (% of disposable (% of disposable
income) Rank income) Rank income) Rank
Australia 10.8 7 7.0 15T 2.1 19
Austria 10.5 8 9.5 11 11.7 3
Belgium 11.1 6 14.3 3 8.6 6
Canada 15.8 3 7.0 15T 2.5 18
Czech Republic n.a. n.a. 6.1 18 4.3 13
Denmark n.a. n.a. 0.9 21 –1.0 23
Finland 3.4 14 0.3 22 –1.6 24
France 8.9 10 11.9 6 12.2 1
Germany 12.1 5 10.5 10 10.8 4
Hungary n.a. n.a. n.a. n.a. 4.9 11
Ireland n.a. n.a. n.a. n.a. 2.7 16
Italy 21.5 1 17.9 1 7.9 7
Japan 16.5 2 10.6 9 3.3 15
Korea, Rep. of 14.8 4 16.3 2 2.9 17
Netherlands 5.6 13 12.4 4 7.4 8
New Zealand 1.3 16 n.a. n.a. n.a. n.a.
Norway –3.3 17 2.6 20 0.4 22
Poland n.a. n.a. 11.7 7 6.5 10
Portugal n.a. n.a. 8.3 13 4.6 12
Slovak Republic n.a. n.a. 8.4 12 4.1 14
Spain 7.8 11 12.2 5 7.1 9
Sweden 2.2 15 7.3 14 9.1 5
Switzerland n.a. n.a. 10.9 8 12.0 2
United Kingdom 6.9 12 6.6 17 1.5 20
United States 9.2 9 4.0 19 0.6 21
OECD Mean 9.1 8.9 5.2
OECD = Organisation for Economic Co-operation and Development.

Notes: “n.a.” denotes “not available.” “T” denotes a tie with another country or countries in ranking. The rate for
Switzerland in 2007 is not available, so the rate for 2006 is used.
Source: OECD (2003, 2009a).
4. International Comparison of Social Benefit Ratios

In this section, we present data on the ratio of social benefits (other than
social transfers in kind) receivable to household disposable income (hereafter
referred to as the “social benefit ratio”) for the OECD member countries for
which data are available. Table 4.2 shows considerable variation in this ratio
between 1996 and 2007. In 1996, the ratio ranged from 16.4% in the US to
43.6% in Denmark, while in 2007, it ranged from 8.1% in Korea (and 17.2%
in the US) to 42.9% in Denmark.
Despite such variation, broad differences among countries in the social
benefit ratio have been relatively stable over time. For example, many
Scandinavian (Northern European) countries (e.g., Denmark, Finland, and
Norway in both years) and many Continental European countries (e.g., Austria,
Belgium, France, Germany, Netherlands, and Switzerland in both years and
Hungary and Italy in 2007) showed higher than average social benefit ratios,
while many Eastern European countries (e.g., Czech Republic, Poland, and
Slovak Republic) and Anglo-Saxon countries (e.g., Canada and the US) plus
Japan in both years, Italy, Portugal, and the UK in 1996, and Ireland, Korea,
and Spain in 2007 showed lower than average social benefit ratios. Thus,
it appears that many Scandinavian (Northern European) and Continental
European countries are welfare states whereas many Eastern European and
Anglo-Saxon countries are at the other extreme.
As for trends over time, the OECD mean of the social benefit ratio was
28.5% in 1996 and 27.4% in 2007, suggesting that there has been little or no
change in the social benefit ratio during the past decade.
5. Correlation between Household Savings Rates and

Social Benefit Ratios
In this section, we present data on correlations between household savings
rates and social benefit ratios and scatter plots of household savings rates and
social benefit ratios in the OECD member countries. The correlation between
the two variables was –0.191 in 1996 and +0.186 in 2007, which suggests that
there is not a strong relationship between the two, either positive or negative.
Figures 4.1 and 4.2 show scatter plots of household savings rates and social
benefit ratios for 1996 and 2007, respectively, for all OECD member countries
for which data on both variables are available. The figures show that there
Table 4.2: Social Benefit Ratio in OECD Member Countries
Country 1996 2007

Ratio of Social Ratio of Social
Contributions and Social Contributions and Social
Benefits to Household Benefits to Household
Disposable Income Rank Disposable Income Rank
Austria 33.8 4 31.6 6
Belgium 30.6 9 31.9 7
Canada 19.1 16 17.4 20
Czech Republic 20.2 14 26.7 14
Denmark 43.6 1 42.9 1
Finland 41.5 2 32.3 5
France 31.0 8 30.0 9
Germany 31.2 7 29.7 10
Hungary n.a. n.a. 29.5 11
Ireland n.a. n.a. 23.4 16
Italy 26.4 11 30.2 8
Japan 18.1 17 24.1 15
Korea, Rep. of n.a. n.a. 8.1 22
Netherlands 39.3 3 35.9 3
Norway 31.9 6 33.6 4
Poland 24.2 13 22.9 17
Portugal 19.9 15 27.4 13
Slovak Republic 24.9 12 21.9 18
Spain n.a. n.a. 21.8 19
Switzerland 33.1 5 36.1 2
United Kingdom 28.1 10 27.7 12
United States 16.4 18 17.2 21
OECD Mean 28.5 27.4

Notes: “n.a.” denotes “not available.”
Source: OECD (2009b).
is no clear relationship between the two, which corroborates our contention
that there is not a strong relationship between the two variables.
Note, however, that there are some countries (e.g., Scandinavian countries
such as Denmark, Finland, and Norway in both years) with a low household
savings rate and a high social benefit ratio, and conversely, some countries
(such as Poland in both years, Italy and Japan in 1996, and Spain in 2007) with
a high household savings rate and a low social benefit ratio. These data suggest
that the social benefit ratio has a negative correlation with the household
savings rate, as expected.
However, there is a larger number of countries in which the social benefit
ratio has a positive correlation with the household savings rate, For example,
some countries (e.g., Eastern European countries such as the Czech Republic
and the Slovak Republic, Anglo-Saxon countries such as Canada and the US in
both years, Portugal and the UK in 1996, and Asian countries such as Japan
and Korea plus Ireland in 2007) have a low household savings rate as well as a
low social benefit ratio, and conversely, some Continental European countries
(e.g., Austria, Belgium, France, Germany, Netherlands, and Switzerland in
Figure 4.1: Household Savings Rates and Social Benefit Ratios, 1996
Household Saving Rate
Social Benefit Ratio
AUT = Austria, BEL = Belgium, CAN = Canada, CZE = Czech Republic, DEN = Denmark, FIN = Finland,
FRA = France, GER = Germany, ITA = Italy, JPN = Japan, NET = Netherlands, NOR = Norway, POL = Poland,
POR = Portugal, SLO = Slovak Republic, SWI = Switzerland, UK = United Kingdom, US = United States.
Note: Blue lines represent means.
Source: Tables 4.1 and 4.2.; OECD (2003, 2009a), OECD (2009b).
Figure 4.2: Household Savings Rates and Social Benefit Ratios, 2007
Household Savings Rate
Social Benefit Ratio
AUT = Austria, BEL = Belgium, CAN = Canada, CZE = Czech Republic, DEN = Denmark, FIN = Finland,
FRA = France, GER = Germany, HUN = Hungary, IRE = Ireland, ITA = Italy, JPN = Japan, KOR = Republic of
Korea, NET = Netherlands, NOR = Norway, POL = Poland, POR = Portugal, SLO = Slovak Republic,
SPA = Spain, SWI = Switzerland, UK = United Kingdom, US = United States.
Note: Blues lines represent means.
Source: Tables 4.1 and 4.2.; OECD (2003, 2009a), OECD (2009b).
both years and Italy in 2007) have a high household savings rate as well as a
high social benefit ratio.
It is curious that the negative relationship between the household savings
rate and the social benefit ratio is observed in some countries but not in others.
This is presumably because factors other than the social benefit ratio more
than offset the impact of the social benefit ratio and cause a low household
savings rate in some countries even though the social benefit ratio is high
and a high household savings rate in some countries even though the social
benefit ratio is low.
The OECD publishes asset and liability data on households for the Group of
Seven countries, and these data can shed light on whether borrowing constraints
are an important determinant of differences among countries in household
savings rates. According to the OECD data, the ratio of household liabilities to
household disposable income was highest in the UK (1.857) in 2007, followed
by the US (1.410), Canada (1.389), Japan (1.277), Germany (1.022), France
(1.001), and Italy (0.725). These data indicate that borrowing constraints are
much more strict in France, Germany, and Italy, which can explain why the
household savings rate is relatively high in these three countries despite their
relatively high social benefit ratios (except for Italy in 1996). Conversely, the
aforementioned data indicate that borrowing constraints are least severe in
Canada, Japan, UK, and US, which can explain why the household savings
rate is relatively low in these four countries despite their relatively low social
benefit ratios (except for Japan in 1996). Unfortunately, asset and liability
data are not readily available for the other OECD member countries, but the
evidence from the Group of Seven countries suggests that borrowing constraints
are an important determinant of differences among countries in household
savings rates. We include a proxy for borrowing constraints in the regression
analysis in the next section.
Looking at whether or not trends over time in the social benefit ratio can
explain trends over time in the household savings rate, there was virtually no
change in the mean of the OECD social benefit ratio during 1996–2007 but
the mean of the OECD household savings rate declined sharply during this
same period (from 8.9% to 5.2%, a decline of some 42%). Thus, it appears that
trends over time in the social benefit ratio cannot explain trends over time in
the household savings rate.
6. Regression Analysis
In this section, we conduct a regression analysis of the determinants of the
household savings rate using panel data on the OECD member countries for
the years 1995, 2000, and 2005.
The dependent variable is the household savings rate (HHSR). The
explanatory variables are AGE, the ratio of the population aged 65 or older
to the population aged 20 to 64; the ratio of social benefits receivable (SBR)
to household disposable income; and CREDIT, the ratio of private credit by
deposit money banks and other financial institutions to GDP. CREDIT is
included as a proxy for the degree of financial development or for the prevalence
of borrowing constraints.
The sources of the data on HHSR and SBR are OECD (2009a) and various
issues of the OECD Economic Outlook. The source of the data on CREDIT is
Beck, Demirguc-Kunt, and Ross (1999, 2009).
The results of the Hausman test indicated that the fixed effects model was
the correct model, so we present the results of the fixed effects model, with the
observations being weighted by the population of each country in 1995.
As the results in Table 4.3 demonstrate, the coefficient of AGE is negative (in
the –0.85 to –1.00 range) and statistically significant at a minimum significance
level of 5%, as expected, indicating that a one percentage point increase in
AGE reduces the household savings rate by 0.85 to 1.00 percentage points.
The coefficient of CREDIT is negative (in the –0.033 to –0.036 range) and
statistically significant at a minimum significance level of 10%, as expected,
indicating that a one percentage point increase in CREDIT lowers the household
savings rate by 0.033 to 0.036 percentage points. Finally, the coefficient of
SBR is positive and totally insignificant, indicating that it does not have a
Table 4.3: Determinants of Household Savings Rates
No. of
Model Constant AGE SBR CREDIT SBR*CREDIT R-squared F-stat. obs.
1 27.573 –0.846 0.408 22.590 67
4.250 0.178 0.028 0.000
6.49 –4.75 0.000
0.000 0.000
2 22.789 –0.847 0.197 0.258 2.770 64
8.339 0.371 0.331 0.044 0.075
2.73 –2.28 0.59 0.007
0.009 0.028 0.556
3 35.721 –1.003 –0.036 0.523 7.660 67
7.189 0.263 0.016 0.021 0.002
4.97 –3.81 –2.16 0.000
0.000 0.000 0.036
4 31.079 –0.980 0.151 –0.033 0.388 2.230 64
11.376 0.449 0.266 0.018 0.025 0.099
2.73 –2.18 0.57 –1.81 0.002
0.010 0.035 0.575 0.079
5 39.772 –0.993 –0.232 –0.103 0.0033 0.437 3.020 64
12.225 0.444 0.371 0.039 0.0018 0.006 0.030
3.25 –2.24 –0.62 –2.66 1.89 0.001
0.002 0.031 0.537 0.012 0.067
Note: The first figure indicates the estimated coefficient, the second figure indicates the standard error, the third
figure indicates the z-value, and the fourth figure indicates the p-value. The first R-squared is within countries,
the second R-squared is between countries, and the third R-squared is for the sample as a whole. The figure
below the F-statistic is the p-value.
significant impact on the household savings rate. Thus, it appears that the age
structure of the population and borrowing constraints are more important
than the social benefit ratio as determinants of differences among countries
in the household savings rate.
Our results concerning CREDIT are consistent with the findings of
Loayza, Schmidt-Hebbel, and Servén (2000), who found that in a sample
of developed and developing countries, the ratio of domestic credit flow to
gross national disposable income had a negative and significant impact on
the private savings rate.
7. Conclusion
In this chapter, we used multi-country data to analyze the determinants of

differences among countries in household savings rates with an emphasis
on the impact of social safety nets, the age structure of the population, and
borrowing constraints on the population. Our main findings were that there
are considerable and stable differences among countries in their household
savings rates and social benefit ratios but that the latter can explain the
former only to a limited extent, with the age structure of the population and
borrowing constraints being more important as determinants of differences
among countries in household savings rates.
Perhaps one reason for our failure to detect a significant impact of social
safety nets on household savings rates is that our analysis did not include a
breakdown of social safety nets into various categories. For example, social
assistance aimed at the poor might have a very different impact on household
saving than a universal health insurance or public pension system. One avenue
for further research is to break down social benefits into various components
to analyze their separate impacts on household saving.
Turning to the policy implications of our findings, our finding that
there is not a clear relationship between social safety nets and household
savings rates implies that improving social safety nets will not necessarily
reduce household savings rates and stimulate consumption. Regardless,
improvement of social safety nets may be desirable because it will alleviate
such household worries as unexpected contingencies and retirement security,
thereby enhancing household welfare. Moreover, our finding that borrowing
constraints are more important than social safety nets as a determinant of
household saving implies that the development of capital markets (and the
relaxation of borrowing constraints) will alleviate the need for precautionary

saving (self-insurance), which is very inefficient, and will serve as a partial
substitute for the development of social safety nets, leading to lower saving,
higher consumption, and higher household welfare. Thus, a two-pronged
approach of simultaneously developing social safety nets and private capital
markets may be the most effective way to enhance household consumption
and welfare.
References
Bailliu, J., and H. Reisen. 1998. Do Funded Pensions Contribute to Higher Savings? A
Cross-Country Analysis. Manuscript. Paris: OECD Development Centre.
Beck, T., A. Demirguc-Kunt, and R. Levine. 1999. A New Database on Financial
Development and Structure. Policy Research Working Paper 2146. Washington,
DC: Development Research Group (Finance), World Bank.
————. 2009. A New Database on Financial Development and Structure (updated May
2009). Washington, DC: World Bank. Available at: http://econ.worldbank.org/
WBSITE/EXTERNAL/EXTDEC/EXTRESEARCH/0,,contentMDK:20696167~pag
ePK:64214825~piPK:64214943~theSitePK:469382,00.html.
Dayal-Ghulati, A., and C. Thimann. 1997. Saving in Southeast Asia and Latin America
Compared: Searching for Policy Lessons. IMF Working Paper WP/97/110.
Washington, DC: IMF.
Edwards, S. 1996. Why Are Latin America’s Savings Rates So Low? An International
Comparative Analysis. Journal of Development Economics 51(1): 5–44.
Feldstein, M. 1974. Social Security, Induced Retirement, and Aggregate Capital
Accumulation. Journal of Political Economy 82(5, Sep/Oct): 905–926.
————. 1977. Social Security and Private Savings: International Evidence in an Extended
Life Cycle Model. In The Economics of Public Services: Proceedings of a Conference
held by the International Economic Association at Turin, Italy, edited by M. Feldstein
and R. Inman. Turin, Italy: International Economic Association.
————. 1980. International Differences in Social Security and Saving. Journal of Public
Economics 14(2, October): 225–244.
Horioka, C. Y. 1989. Why Is Japan’s Private Saving Rate So High? In Developments in
Japanese Economics, edited by R. Sato and T. Negishi, 145–178. Tokyo: Academic
Press/Harcourt Brace Jovanovich.
Hubbard, R. G., J. Skinner, and S. P. Zeldes. 1995. Precautionary Saving and Social
Insurance. Journal of Political Economy 103(2, April): 360–399.
the World? Review of Economics and Statistics 82(2, May): 165–181.
Modigliani, F., and A. Sterling. 1983. Determinants of Private Saving with Special
Reference to the Role of Social Security: Cross Country Tests. In The Determinants
of National Saving and Wealth: Proceedings of a Conference Held by the International
Economic Association at Bergamo, Italy, edited by F. Modigliani and R. Hemming.
London: Macmillan.
OECD. 2003. Economic Outlook No. 73. Paris: OECD.
————. 2009a. Economic Outlook No. 86. Paris: OECD.
————. 2009b. National Accounts of OECD Countries: Detailed Tables, Volumes IIA and IIB.
Paris: OECD.
Comments
Mukul Asher
It is always a learning experience to study research papers by Professor Horioka.

This chapter on differences among countries in household savings rate and
their interrelationships with social benefit ratios is no exception. The method
of analysis, presumably as an initial step, is simple correlation between the
above two ratios, without introducing any lagged effects. The technique
employed is too simple and predictably the results derived from it are less
than robust.1 This suggests the need for further refinements of the study,
including incorporating policy implications that can be drawn for meeting
aging challenges in Asia and elsewhere. My remarks are therefore directed
toward suggestions for further refinements.
1. Comments on Definitions
Let me start with the definition of the savings ratio. The authors define the
savings ratio as net household savings to household disposable income. It is
not clear whether it is the net of liabilities of the household, i.e., borrowing,
and the net of depreciation of the household sector. This suggests that the
data that is used for household savings from household surveys need to be
reconciled with the household savings data in the national income accounts.
In the national income accounts, the business savings only include corporate
savings, not private proprietorships and partnerships and so on; these are all
under household savings figures.
The second area is related to what exactly is savings in the household
context. In particular, how housing expenditure is to be treated in the
1
Editors’ note: Chapter 4 contains a regression equation that controls for other factors, taking into
account this discussant’s comment.
consumption-investment dichotomy. The national income account conventions
use new residential construction as an investment rather than consumption,
but there are many variations in the way the housing expenditures are reflected
in the household account. A clarification of this point would be useful. In the
Asian context, this issue is even more important. First, housing expenditure is
a major item for Asian households. Second, a significant proportion of Asian
household savings is not undertaken in the form of financial savings but in
the form of physical savings. Some countries, such as India, publish separate
data on financial and physical savings.2
Third, the social benefit ratio is defined in the chapter as the ratio of
social contributions and benefits receivable to household disposable income.
Including both contributions and benefits in this ratio is a bit puzzling. It
might be useful to clarify why both are included.3
The household savings data for the OECD countries presented in the chapter
exhibit a great deal of volatility. Italy’s savings rate decreases from 21.5% in
1985 to 7.9% in 2007, while there are no indications that the benefit ratio has
changed much. This tends to suggest that the simple technique employed in
the chapter may lead to over-generalized conclusions, especially with respect
to policy implications.
The chapter also discusses the borrowing constraints, but without elaborating
on their nature or policy implications. For example, can policymakers expect
an increase in household savings or a reduction in household savings if rules
for obtaining credit cards are changed?
2. National Savings, Fiscal Spending, Aging, and Policy

Design
Let me make some broader suggestions for further refinements. The study focuses
on household savings but it is increasingly evident that there are important
interconnections between household, corporate, and government savings.
Prasad (2009) found that in the PRC, the share of corporate savings has risen
markedly, accounting for about half of total national savings in 2007–2008,
2
Horioka replied that he used data from the national accounts of each country as compiled by the OECD,
which are based on the conceptually correct concept of saving: they are net of liabilities, and net of
depreciation; they include saving in the form of investment in housing and other physical assets; and
they include the saving of individual proprietors.
3
Horioka replied that he used social benefits receivable, excluding payable.
but household savings have remained nearly stable. In contrast, Korea, an

OECD member, household savings as share of GDP have declined since the
late 1990s (Prasad 2009). Households in the US are currently rebalancing
their balance sheets by increasing their savings. So household savings as a
share of disposable income have become positive, but government savings are
becoming increasingly negative. The overall national savings trend will depend
on the combined impact of saving behavior of all three sectors, household,
corporate, and government.
The additional fiscal expenditure needed due to aging, and its implications
for household savings (e.g., by including demographic variables), should also
be considered. A recent study by the European Commission (2009) on fiscal
sustainability, assuming a public debt to GDP ratio of 60% in the OECD
countries, projected the additional fiscal revenues needed to be raised. The
estimates, depending on the assumptions made about factors such as economic
growth and inflation, range from 5.0% to 6.5% of GDP. This is huge as even
the most ambitious tax reforms that countries can undertake will not yield
more than 1–2% of GDP. So, the above fiscal adjustment is going to create a
kind of uncertainty that needs to be taken into account.
It will be useful to at least briefly discuss the dynamics between the
household savings and savings by the other sectors. The share of national
savings by each sector and the trends should be examined. There is already
considerable literature on Ricardian equivalence, and its relevance needs to
be brought out in the chapter.
The importance of the design of social benefits and contribution
arrangements should also be considered. Thus, social assistance directed at
the poor in all age groups, or social pensions directed at the elderly, may
impact household savings very differently depending on whether an asset test
is mandated. Mere establishment of social protection programs or promised
extension of coverage that the governments, including those in Asia, have
announced in response to the current crisis will not necessarily bring about
the desired outcomes. The actual savings impact is going to depend very much
on whether these programs are perceived to be sustainable, and whether the
design and delivery of these programs instill a high degree of confidence in
their effectiveness. In the absence of these, precautionary savings may not
decrease in countries such as the PRC.
Finally, the rationale for the chapter and policy implications, taking into
account the altered macroeconomic and labor market environments due to
the global economic crisis, could be further elaborated upon.
References
European Commission. 2009. Sustainability Report 2009 (provisional version). Brussels:
Directorate General, Economic and Financial Affairs.
Prasad, E. 2009. Rebalancing Growth in Asia. Finance and Development 46(4): 19–21.
Influence of Age Structure on Savings and Social Spending 81
V
The Influence of Age Structure on
Savings and Social Spending
Davide Furceri and Annabelle Mourougane
1. Introduction
Changes in the age structure in the OECD countries and in some Asia-Pacific
emerging market economies are likely to affect economic developments in the
medium-term. In particular, saving patterns are expected to alter when the
elderly become a larger proportion of consumers and savers, with widespread
implications for capital and good markets. Social spending is also likely to
adapt to the evolving age structure of the population. This will have major
implications for public long-term sustainability.
This chapter seeks to quantify the impact of the age structure on private
saving and social spending, based on a review of the literature and on new
panel data analyses covering OECD countries and Asia-Pacific economies.
The main findings are as follows:
t "HFTUSVDUVSFJTBTJHOJmDBOUEFUFSNJOBOUPGQSJWBUFTBWJOHJOCPUI0&$%
and Asia-Pacific economies. In particular, the old-age dependency ratio
is estimated to be negatively correlated with household and private
saving and this effect is robust to a range of tests. By contrast, the young-
age dependency ratio generally does not have a significant impact on
household savings rates in OECD countries. At the same time, the young-
age dependency ratio is found to influence private saving, but the sign
and the magnitude of the impact depends on geographical coverage and
the model specification.
The authors would like to thank Charles Horioka, Mario Lamberte, Hyungpo Moon, Gloria Pasadilla, and
all the other participants at the ADBI Annual Conference 2009 for useful comments and suggestions.
t "HFTUSVDUVSFBQQFBSTUPJOnVFODFTPDJBMTQFOEJOHJO0&$%DPVOUSJFT
with an increase in the old-age dependency ratio boosting social spending,
and a rise in the young-age dependency ratio having the opposite effect.
Here again, the old-age dependency channel is estimated to dominate.
This finding reflects the marked influence of the old-age dependency ratio
on old-age pension expenditure and, to a lesser extent, on active labor
market programs, housing, and health-care spending. The results on
health-care spending should nonetheless be interpreted with caution, as
specific factors explaining this expenditure are only imperfectly captured
in analyses through country fixed effects.
The chapter starts with a discussion of demographic trends both in OECD
countries and in Asia-Pacific economies. It then turns to the impact of these
trends on private saving. The effect of the age structure on aggregate social
spending and its breakdown by components is then analyzed. The final section
concludes and identifies policies that would help temper the effects of aging
on saving and social spending, and ultimately on economic growth.
2. Demographic Trends in OECD and Asia-Pacific

Economies
Population aging at the global level is rapid and will lead to a substantially
higher proportion of elderly in the world over the next decades. Whereas the
world population had a pyramid shape in 1990, it is expected to have a bell shape
by 2050, with a marked increase in the number of persons aged between 40
and 65 years. This is particularly the case for the OECD countries (Figure 5.1),
where the progressive aging of populations stems from a combination of the
baby-boom in the early post-war period, the subsequent fall in fertility rates,
and an increased life expectancy. Although aging is global, there are marked
international differences in the speed and the extent of the aging process.
OECD population projections to 2050 suggest that most OECD countries
will experience very modest growth or declines in the total population (Dang,
Antolin, and Oxley 2001; OECD 2009a, 2009b). This will be accompanied by
a fall in the working-age population and an increase in the number of elderly,
particularly those over 80, leading to a near doubling, on average, of the ratio
of elderly to the working-age population. For most countries, this ratio is
projected to increase until about 2035–2045, and then to stabilize or decline
slightly. Some countries will experience pressure due to aging populations
Figure 5.1: Population by Age Group and Gender, in 2000 and 2050, as
Percentage of Total Population in Each Group
% of total population in each group

Source: Authors’ estimate based on OECD data.
after 2050. OECD countries will also see a small decline in the ratio of youth
to the working-age population.
However, not all OECD countries will be affected by aging populations to
the same extent (Figure 5.2). Europe has already a much older population than
North America. Within Europe, the populations of Italy and Germany are
aging faster than France. Countries where the population is currently relatively
young, such as Mexico and Turkey, will experience a fast aging process.
Overall, while demographic developments will continue to sustain growth in
the US, aging is likely to slow the GDP per capita growth in Japan and Europe.
Previous evidence based on simulations suggests that aging could depress
GDP per capita by an average of 0.2–0.3 percentage point per year during
the next 50 years (Oliveira Martins et al. 2005). International flows of capital,
goods and services, and labor will be important mechanisms moderating the
effects of population aging in each individual country.
The populations of OECD Asian countries are predicted to age at an
extremely rapid pace. The aging process accelerated in the 1990s in Japan
and the proportion of elderly in that country is now higher than the average in
OECD countries. This trend is expected to continue at a similar pace until 2050.
Korea is predicted to be the country where population aging will occur at the
Figure 5.2: Old-age Dependency Ratios in OECD Countries—

Historical and Projected Values, 1950–2050

Source: OECD Demographic and Labour Force Databases.
fastest pace in the OECD, with the old-age dependency ratio1 reaching 80% by
2050, close to the rate forecast for Japan. The populations of Australia and New
Zealand are also projected to increase in age in the coming decades, but the
old-age dependency ratio is expected to remain below the OECD average.
Significant aging is foreseen for many emerging market economies of East
and Southeast Asia, although on average the share of the elderly in the population
is projected to continue to be lower than in the OECD countries. The timing
and the extent of aging will differ within the regions (Figure 5.3). The East
Asian countries are further advanced in terms of population aging and their
old-age dependency ratio is expected to rise sharply from 2010 to 2050. The
aging pattern of the Chinese population is similar to that of East Asia. After
2010, the old-age dependency ratio in PRC is expected to increase gradually but
would stay below that of East Asia; it is anticipated to be above 40% by 2050.
The old-age dependency ratio is projected to start to rise in the Southeast Asian
countries from 2020, and to stay below or close to 30% by 2050.
3. Does Demography Affect Private Saving?
According to the life-cycle hypothesis, older people tend to have a higher propensity
to be dissavers, consuming beyond their incomes (Ando and Modigliani 1963).
In this framework, a country with a higher old-age dependency ratio will have
a lower household savings rate because there will be more elderly dissavers and
fewer people belonging in the category of the working-age population, who are
generally net savers. Nonetheless, two factors may compensate for this trend.
First, the expectation of greater longevity may encourage younger people to
increase their savings. Secondly, the expectation of a higher lifetime income may
encourage a new generation to increase their saving levels. It is likely, however,
that these two effects would not fully offset the elderly dissaving, resulting in a
negative relationship between saving and the old-age dependency ratio.
The empirical evidence based on microeconomics studies points to the
existence of a partial consumption smoothing.2 Household survey data
suggest that total consumption displays a humped shape across age groups,
but the hump-shaped income profile is even more pronounced for the elderly
1
The old-age dependency ratio is defined as the ratio of the population over 65 years to the 15–65-year-
old population.
2
One important exception is Japan. Micro-evidence (Horioka 2009) shows that the life-cycle model is
highly applicable in this country.
Figure 5.3: Old-age Dependency Ratios in Pacific-Asian Countries—
Historical and Projected Values, 1980–2050

Source: OECD (2009a).
population (Oliveira Martins et al. 2005).3 In addition, there is evidence that

some elderly people may continue to be net savers during retirement, often
through housing assets, for the motives of bequests or insurance. In particular,
it is found that income-rich elderly run down their net worth at a very slow
rate (Dynan, Skinner, and Zeldes 2004; De Nardi, French, and Bailey Jones
2009). Differences in life expectancy related to health, gender, and permanent
income appear to be important factors explaining saving patterns, with the
effect of each factor being of a similar order of magnitude.
At the macro-economic level, a significant link between the age structure
and private saving has usually been found (Table 5.1). For instance, panel
regressions conclude that saving is positively correlated with the proportion
of working-age population in the total population, and negatively correlated
with the old-age dependency ratio, with the latter effect often estimated to be
larger than the former.4 However, estimates based on cross-sectional aggregate
Table 5.1: Effect of Age Structure on Private Saving
Coefficient on Elderly
Authors Scope Dependency Rate
Feldstein (1980) Industrial countries
Pool cross-section time series Private savings -1.21
Masson et al. (1995) High income and developing countries
Weil (1994) 9 OECD countries
Pool cross section time series Household savings -1.36
Disney (1996) 19 OECD countries
Pool cross-section time series Household savings -2.025
Horioka (1991) Japan
Time series Private savings -1.03
Schmidt-Hebbel et al. (1992) 10 developing countries
Cross-section Household savings -0.48
Modigliani and Sterling (1983) Industrial countries
Cross-section Private savings -0.51
Heller and Symansky (1997) Asian Tigers
Oliveira Martin et al. (2005) OECD countries
Pool cross-section time series Household savings -4.267
Sources: Heller and Symansky (1997) and authors’ compilation.
3
This evidence is not a direct test of the life-cycle hypothesis, as part of the results can be explained by
time and cohort effects.
4
Savings encompass either private or household savings.
data have yielded larger and more significant coefficients than those derived
from pooled time series data.
A simple scatter plot of household savings rates against old-age dependency
ratios in the OECD countries suggests a negative relationship between the two,
although there is a large variation across countries (Figure 5.4). In particular, Italy
displays stronger savings rates than suggested by its old-age dependency ratio,
whereas New Zealand and Luxembourg show a much lower rate of saving.
To get more insights into the effect of the age structure on saving, a saving
equation was estimated for two groups of countries/economies, the OECD
countries and Asia-Pacific economies,5 for an unbalanced panel of countries
in 1980–2008.
savings rateit = ai + y1CONTROL + y2DEMOGRAPHYit + εit (1)
where savings rate is respectively the household savings ratio for the OECD
countries and private saving (gross national saving–gross public saving) for
OECD and Asia-Pacific economies; CONTROL is a set of macro-economic
controls including fiscal balance (as share of GDP), social spending (as share of
GDP), GDP per capita, GDP growth, and the unemployment rate. Fiscal balance
Figure 5.4: Household Savings Rates versus Old-age Dependency Ratio

Household saving ratio (%)
Old age dependency ratio

Sources: OECD (2009a) and World Bank World Development Indicators.
5
The Asia-Pacific group incorporates Bangladesh; PRC; Hong Kong, China; India; Indonesia; Malaysia;
Mongolia; Pakistan; Philippines; Singapore; Thailand; Viet Nam; Australia; and New Zealand.
seeks to capture any Ricardian household behavior whereby households increase

their precautionary saving when they foresee tax increases in the future. Social
spending is included, as it may reduce the need for precautionary saving.
The unemployment rate corrects for cyclical movements. DEMOGRAPHY
encompasses the old-age and the young-age dependency ratios.6 Country
fixed effects have been included to account for country heterogeneity. Data
for saving and age-dependency ratios are taken from World Bank’s World
Development Indicators and the OECD.
There are some econometric difficulties in estimating age-structure models
as set up in equation (1), even though the age variable is certainly of better
quality than many other variables used in empirical work. The age distribution
is slow-moving and thus is difficult to discriminate from other potential
secular trends in the data. Different age groups are correlated both within and
between countries, leading to potential multicollinearity problems. The lack
of harmonization prevents the inclusion of other important drivers of saving
in the analysis such as asset prices. Finally, there may be some endogeneity
issues, as income and income growth are unlikely to be fully exogenous in a
saving equation.
Table 5.2 presents the results regarding household savings rates. The
old-age dependency ratio had a significant negative correlation to saving in
OECD countries. This finding appears robust to a range of robustness tests,
including the introduction of additional demographic factors (such as mortality
or birth rate) and social spending, the use of time fixed effects, and different
estimation methods such as the 2-step generalized least squares (to control
for heteroscedasticity and serial correlation) and the generalized method of
moments (to control for endogeneity issues).
Overall, for OECD countries, a 1 percentage point rise in the old-age
dependency ratio would lead to a 0.2–0.6 percentage point decrease in
household savings rates. In contrast, the young-age dependency ratio mostly
does not have a significant impact on household savings rates. Interestingly,
social spending is found to influence saving decisions.7 More precisely, a
1 percentage point increase in social spending would reduce household savings
by about 0.8 percentage point.
6
In theory, foreign capital inflow could influence savings. However, it is unclear why this should be
incorporated in the analysis, as the saving ratio is an after-measure.
7
Barnett and Brooks (2010) report similar results in the case of the PRC, where government spending
on health is found to reduce household saving. Spending on education would not generate a similar
effect. Evidence is also more mixed for rural households.
Table 5.2: Saving Equation—Household Savings Rates
Baseline+ Baseline Baseline +

Baseline + Mortality + Social Time Fixed Baseline - Baseline -
Baseline Birth Rate Rate Spending Effects GLS GMM
OECD
Saving (-1) - - - - - - 0.893
(64.57)***
Deficit -0.243 -0.238 -0.222 -0.254 -0.213 -0.22 -0.111
(-3.31)*** (-3.14)*** (-3.27)*** (-2.92)*** (-2.47)*** (-10.01)*** (-6.26)***
GDP growth -0.036 -0.026 -0.081 -0.124 0.004 -0.105 -0.063
(-0.55) (-0.42) (-1.22) (-2.33)** -0.05 (-4.12)*** (-2.92)***
GDP per capita 1.03 1.326 7.561 0.278 3.421 1.603 0.888
-0.32 (0.43)* (1.79)* -0.08 -0.6 (3.22)*** (4.23)***
Unemployment rate 0.147 0.164 0.237 0.071 0.168 -0.01 -0.053
(1.84)* (2.17)*** (3.08)*** -0.95 -1.55 (-0.23) (-2.44)***
Social spending - - - -0.807 - - -
(-1.86)*
Social spending
squared - - - 0.017 - - -
(1.82)*
Old-age
dependency ratio -0.552 -0.541 -0.412 -0.386 -0.501 -0.223 0.07
(-3.42)*** (-3.29)*** (-2.86)*** (-1.65)* (-2.63)*** (-3.01)*** (-2.99)***

Young-age
dependency ratio 0.159 0.141 0.287 0.144 0.196 0.036 0.041
-0.87 -0.73 -1.6 -0.71 -0.88 (0.63)*** (2.93)***
Birth rate - 0.128 - - - - -
-0.74
Mortality rate - - 0.07 - - - -
(2.20)**
Number of
observations 632 628 576 473 632 632 623
2
R 0.9 0.9 0.89 0.92 0.9 - -
GDP = gross domestic product, GLS = generalized least squares, GMM = generalized method of moments,
OECD = Organization for Economic Co-operation and Development.
t-statistics in parentheses. *,**,*** Significance at 10%, 5%, and 1%, respectively.
Source: Authors’ estimates.
The old-age dependency ratio was negatively correlated with private saving
both in OECD and in Asia-Pacific countries (Table 5.3). In absolute terms,
while an increase of 1 percentage point in the old-age dependency ratio would
decrease private saving in OECD countries by 0.8 percentage point (baseline
estimation), it would decrease private saving in Asia-Pacific economies by
2.6 percentage points. Several factors could explain the behavioral differences
between the two areas, including differences in bequest motives and in
social protection coverage. While it is difficult to test econometrically for
these assumptions given the paucity of data, there are some indications
that the low social protection coverage may be an important factor. Indeed,
social spending in the non-OECD Asian countries was about 10 percentage
point lower, in terms of GDP share, than in the average of OECD countries
(OECD 2009b).
The young-age dependency ratio was also estimated to influence savings,
but this effect varied depending on regional coverage and model specification.
While there is evidence of a negative relationship for OECD countries, the
impact of young-age dependency ratio on private saving was mostly statistically
non significant for Asia-Pacific countries.
4. How Does Age Structure Influence Social Spending?

Private savings are only one facet of national savings. It is also crucial to
look at the impact of demographic trends on public saving, which represents
a sizable share of national saving in OECD countries and Asia-Pacific
countries. Most assessments of the impact of aging on public sector saving
have focused on calculating the specific implications for expenditure in
key social sectors that are likely to be affected by demographic shifts.
This includes old-age pension programs but also programs permitting
early withdrawal from the labor market and social spending, in particular
on health care and long-term care for the frail elderly, child and family
benefits, and education.
An increase in the old-age dependency ratio is expected to increase total
old-age social spending, as it implies a rising number of pensioners. At the
same time, governments may be willing to decrease pensions as the number of
elderly rises, to ensure the sustainability of the system. Finally, the magnitude
of the effect is an empirical issue. Lindert (1996) found a positive and significant
effect of the old-age dependency ratio on old-age spending. In the same vein,
Table 5.3: Saving Equation—Private Savings

Baseline+ Baseline Baseline +
Baseline + Mortality + Social Time Fixed Baseline -
Baseline Birth Rate Rate Spending Effects GLS
OECD
Deficit -0.476 -0.441 -0.481 -0.441 -0.452 -0.444
(-3.96)*** (-3.80)*** (-3.98)*** (-3.20)*** (-3.18)*** (-7.98)***
GDP growth 0.138 0.142 0.159 0.171 0.097 0.034
-1.61 (1.65)* -1.61 -1.54 -0.95 -0.61
GDP per capita -3.784 -3.547 1.67 -1.165 -2.032 2.038
(-0.61) (-0.55) -0.19 (-0.22) (-0.22) (1.88)*
Unemployment rate 0.307 0.386 0.42 0.382 0.261 0.258
(2.80)*** (2.73)*** (2.29)** (2.56)*** (1.73)* (3.38)**
Old-age dependency ratio -0.876 -0.806 -0.737 -1.24 -0.684 -0.196
(-3.33)*** (-3.66)*** (-2.67)*** (-3.29)*** (-1.96)* (-1.27)
Youth-age dependency ratio -0.855 -0.904 -0.936 -0.878 -0.922 -0.423
(-3.04)*** (-3.40)*** (-2.69)*** (-4.48)*** (-2.90)*** (-3.85)***
Birth rate - 0.272 - - - -
-0.9
Mortality rate - - 0.098 - - -
-1.5
Social spending -1.48
(-2.28)**
Social spending squared 0.032
(2.43)**
Number of observations 317 314 277 573 317 316
R2 0.95 0.95 0.95 0.32 0.95
ASIA-PACIFIC
Deficit -0.76 -0.608 -0.466 - -0.951 -0.722
(-3.02)*** (-2.50)** (-0.84) (-4.53) (-3.80)***
GDP -0.068 -0.059 1.181 - 0.037 0.012

(-0.59) (-0.72) -0.61 -0.2 -0.14
GDP per capita 22.767 21.064 22.863 - 35.436 3.901
(3.72)*** (3.14)*** (2.70)*** (3.88)*** (2.96)***
Unemployment rate 0.347 0.36 0.731 - 0.614 0.221
-1.21 -1.02 -1.23 (1.71)* -0.74
Old-age dependency ratio -2.593 -2.344 -1.492 - -3.345 -2.043
(-3.66)*** (-3.77)*** (-1.78)*** (-3.62)*** (-9.25)***
Young-age dependency ratio 0.322 0.203 -0.022 - 0.284 -0.274
(2.37)** -1.33 (-0.06) (1.12)* (-1.57)
Birth rate - 0.121 - - - -
-0.55
Mortality rate - - 0.91 - - -
(1.94)*
Number of observations 107 82 50 - 107 107
R2 0.95 0.96 0.67 - 0.71 -
GDP = gross domestic product, GLS = generalized least squares, GMM = generalized method of moments,
OECD = Organization for Economic Co-operation and Development.
t-statistics in parentheses. *,**,*** Significance at 10%, 5%, and 1%, respectively
most long-term projections point to a rise in old-age spending due to population

aging. Dang, Antolin, and Oxley (2001) found that total old-age spending as a
percent of GDP rose on average by 3.4 percentage point from 2000 to 2050 for
the countries included in their analysis.8 In particular, Spain, Norway, Korea,
and the Czech Republic were predicted to experience the largest increases,
close to or above 7 percentage point.
Past studies provide only limited guidance on the effect of the age structure
on health expenditure, as there are numerous factors at play whose impact
is quite uncertain. Key factors appear to be the rapid introduction of new
technologies and greater demand for health care, itself a reflection of rising
incomes and a more educated population.9 It is not clear which demographic
factors will have the strongest effect on health-care spending. Empirical
studies show that expenditure may be more related to nearness to death
than to age (Zweifel, Felder, and Meiers 1999; Seshamani and Gray 2004a,
2004b).
Household survey data indicate that the share of health-care spending
increases with age (Oliveira Martins et al. 2005). This increase is particularly
pronounced in the US. However, despite large demographic trends, age-induced
changes in the structure of consumption is expected to be relatively moderate
at the aggregate level, as changes in the consumption share of age-sensitive
products tend to offset each other across age groups. At the macro-economic
level, authors have nonetheless found a significant positive effect of the old-
age dependency ratio on health-care expenditure (Khoman and Weale 2007).
However, Breyer (1999) showed that health-care expenditure in Germany was
mostly influenced by technological progress and not by age. A similar result
was found for the US (Okunade and Murthy 2002).
Family spending is another item that is likely to be sensitive to the age
structure, in particular the young-age dependency ratio. Dang, Antolin, and
Oxley (2001) projected a decline in family/child benefit stemming from the
expected fall in the young-age dependency ratio in most OECD countries.
From 1.6% of GDP in 1995, the share of these benefits on average in 11 OECD
countries would steadily decline to close to 1% by 2035, and would stabilize
at this level throughout 2050. However, family spending in Nordic countries
would still be above average.
8
The study covered 21 OECD countries, including all the Group of Seven countries.
9
The role that technological advances play is complex. Some innovations have been cost-saving. Despite
this, much of the impact of technology appears to have increased health-care costs (Jones 2003).
Building on this line of research, this paper focuses on the impact of age
structure on social spending and its components. We first start with a descriptive
analysis detailing the main features of social spending and how they relate to
the age structure. Data for social spending are taken from the OECD Social
and Welfare Statistics database and are available from 1980 to 2005. Nine
social policy areas are identified: old age, survivors, incapacity-related, health,
family, active labor market, unemployment, housing, and others.10
Social spending represents a significant share of government expenditure
and GDP, on average about 43% and 19%, respectively (Table 5.4). There is,
however, a large variation across countries, with total spending ranging from
a bit more than 5% in Korea and Mexico over the sample to more than 25% in
the Nordic and some continental European countries (Table 5.5). In contrast,
country differences in terms of both level and composition of social spending
have not evolved much over time. Social spending as a share of GDP displays
an upward trend for most of the countries in the sample, the Netherlands
being a clear exception.
Table 5.4: Social Spending Shares and Categories

(average for 1980–2006)
% of Total Expenditure % of GDP

Total 43.4 19.2
Old Age 14.3 6.3
Survivors 1.9 0.9
Incapacity-related 5.4 2.4
Health 12.1 5.2
Family 3.9 1.8
Active Labor Market Program 1.4 0.6
Unemployment 2.8 1.3
Housing 0.8 0.4
Other (policy areas) 1 0.5
Source: OECD Social and Welfare Statistics Database.
10
Public spending on Family includes financial support exclusively for family and children and encompasses
child-related cash transfers to family, public spending on services for families, with children and
financial support for families provided through the tax system. Spending for such items as health care
or housing are not exclusively for families and therefore are not included in this item.
Table 5.5: Social Spending Shares and Categories by Countries (% GDP)
T O S I HE F A U HO OP
Australia 14.7 3.7 0.3 1.9 4.9 2.1 0.4 1.1 0.2 0.2
Austria 25.5 11.6 0.5 2.6 6.1 2.8 0.4 1 0.1 0.3
Belgium 25.8 7 2.5 3 6.4 2.6 1.1 3 0.1 0.4
Canada 17.7 3.9 0.5 1.1 6.4 0.9 0.5 1.5 0.6 2.5
Czech Rep. 18.7 6.6 0.4 2.4 6 2 0.2 0.5 0.1 0.6
Denmark 25.8 7.3 0 3.6 5.2 3.1 1.4 3.9 0.6 1
Finland 25.4 7.5 1 4.2 5.7 3 1 2.2 0.3 0.5
France 26.4 9.6 1.7 2.3 6.8 2.8 1 1.5 0.8 0.2
Germany 25.1 10.3 0.6 1.9 7.4 1.9 1.1 1.4 0.2 0.5
Greece 17.2 8.7 0.9 1.2 4.5 0.8 0.3 0.4 0.4 0.1
Iceland 15.9 4 0.1 2 6.2 2.6 0.1 0.4 0.1 0.4
Ireland 16.5 3.6 1 1.8 5.3 1.8 1 1.7 0.6 0.3
Italy 21.4 9.5 2.2 1.9 5.8 1 0.5 0.8 0 0
Japan 14.2 5.5 1.2 0.7 5.4 0.6 0.3 0.5 - 0.2
Korea, Rep. of 5.3 1.6 0.3 0.4 2.3 0.2 0.1 0.1 - 0.3
Mexico 5.4 0.8 0.3 0.2 2.4 0.5 0 - 0.7 0.4
Netherlands 23 5.6 0.6 4.9 5.4 1.7 1.2 2.4 0.4 0.8
New Zealand 19.2 6 0.2 2.4 5.6 2.4 0.7 1.1 0.5 0.2
Norway 22.4 6.8 0.4 4.6 4.1 3 0.8 0.7 0.2 0.7
Portugal 15.3 5.4 1.1 2.3 4.5 0.8 0.5 0.6 0 0.1
Spain 19.4 7 1.1 2.4 5.1 0.6 0.5 2.5 0.1 0.1
Sweden 29.8 9.1 0.7 5 7.1 3.7 1.9 1.4 0.8 0.7
Switzerland 17 6.2 0.4 2.6 4.7 1.2 0.5 0.6 0.1 0.7
United Kingdom 19.1 5.2 0.8 2.1 5.5 2.4 0.5 1 1.3 0.5
United States 14.6 5.4 0.9 1.2 5.3 0.7 0.2 0.5 1.4 0.5
AVERAGE 19.2 6.3 0.8 2.4 5.4 1.8 0.7 1.3 0.4 0.5
T = total social spending, O = old-age, S = survivors, I = incapacity-related, HE = health, F = family, A = active

labor market program, U = unemployment, HO = housing, OP = other policy areas. (–) means missing.
Source: OECD Social and Welfare Statistics Database.
Looking at the social-spending breakdown by category, old age and health
are by far the largest components of social expenditure. These components
represent on average 14.3% and 12.1%, respectively, of total government
spending. Interestingly, the third largest category is Incapacity-related
spending, followed by family spending. Spending on unemployment, active
labor market programs, and housing represents a small share of the total.
Spending items that are sensitive to the age structure of the population are
found to represent more than 30% of total public spending. This spending
appears to be well correlated with the old-age dependency ratio (Figure 5.5).
Nordic countries, and Sweden in particular, spend more on social expenditure
than suggested implicitly by their old-age dependency ratio. By contrast,
demographic factors would suggest larger social expenditures in some Southern
European countries, but also the US, the UK, and Japan, than is in fact the
case. Among the different components of social spending, the correlation
with the old-age dependency ratio is the strongest for old age-related spending
(Figure 5.6). Health spending appears to show a weak positive correlation with
the old-age dependency ratio (Figure 5.7). The graphs in Figures 5.5, 5.6, and
5.7 should, however, be interpreted with caution as they do not control for
the effects of additional variables.
Moving to inferential analysis, an unbalanced panel of OECD countries
was estimated over the period 1980–2005 to examine the relationship
Figure 5.5: Social Spending versus Old-age Dependency Ratio

Household saving ratio (%)

Source: OECD Social and Welfare Statistics Database and World Bank World Development Indicators.
Figure 5.6: Old-age Spending versus Old-age Dependency Ratio

Old age spending (% of GDP)

Figure 5.7: Health Spending versus Old-age Dependency Ratio

Health spending (% of GDP)
between social spending and the age structure. The following specification
is estimated:
spending
GDPit = βi + β1CONTROL + β2 DEMOGRAPHYit + εit (2)
where spending is total social spending and CONTROL is a set of macro-

economic controls: public tax revenues, GDP per capita, GDP growth, and the
unemployment rate. These factors correct for the cyclicality of social spending
and the dependence of spending on the state of economic development.
DEMOGRAPHY encompasses the old-age and the young-age dependency ratios.
Country fixed effects have been incorporated. Equation (2) was estimated
using ordinary least squares, but several robustness tests were carried out.
First, the rate of birth and/or the rate of mortality were added to the set of
demographic variables. Secondly, time fixed effects were considered. Finally,
the equation was estimated using feasible least squares. Baseline results and
robustness tests are reported in Table 5.6.
The old-age dependency ratio is found to be a significant determinant of
total spending, with a rise in the ratio implying an increase in spending. This
result appears to be extremely robust. A 1 percentage point increase in the
old-age dependency ratio would lead to a 0.2 percentage point increase in total
spending as a share of GDP. The magnitude of the coefficient ranges from
0.1 to 0.4 depending on the control variables included and on the estimation
techniques. The young-age dependency ratio is also estimated to have a negative
influence on total social spending in the OECD countries.
In a second step, individual social spending items were analyzed. This made
it possible to identify which spending categories are driving the aggregate
results. Results are reported in Table 5.6.
The old-age dependency ratio is found to have a positive influence on
certain spending categories, in particular old age and, to a lesser extent, health,
active labor market measures, and housing spending.11 By contrast, the ratio is
estimated to have a negative impact on unemployment benefit and incapacity-
related spending. No significant effect of the old-age dependency ratio is found
for survivors, family, and others spending, although the coefficient has the
expected positive sign (Table 5.7).
11
Results on health expenditure are in line with the expansion of the morbidity hypothesis (Gruenberg
1977). This result should nonetheless be treated with caution, as other demand-and-supply factors which
usually drive health-care spending are only imperfectly captured through country fixed effects.
Table 5.6: Total Social Spending Equation
Baseline +
Baseline+ Birth Rate Time Baseline +
Baseline + Mortality + Mortality Fixed Time Fixed
Baseline Birth Rate Rate Rate Effects Effects-GLS
Revenue 0.237 0.243 0.233 0.239 0.565 0.414
(6.42)*** (6.44)*** (5.78)*** (5.98)*** (41.57)*** (21.06)***
GDP growth -0.177 -0.177 -0.226 -0.198 -0.16 -0.02
(-4.34)*** (-4.11)*** (-4.79)*** (-4.32)*** (-3.53)*** (-1.41)
GDP per capita 0.881 1.234 -2.837 -3.206 -0.294 0.134
-1.06 -1.44 (-2.28)** (-2.65)*** (-1.48) -0.62
Unemployment rate 0.283 0.298 0.231 0.241 0.157 0.235
(7.21)*** (8.66)*** (5.60)*** (6.06)*** (6.56)*** (8.13)***
Old-age dependency ratio 0.193 0.193 0.127 0.127 0.287 0.395
(3.77)*** (3.68)*** (2.01)** (2.00)** (7.91)*** (7.90)***
Young-age dependency ratio -0.141 -0.17 -0.19 -0.239 -0.039 -0.055
(-5.08)*** (-4.99)*** (-4.82)*** (-5.51)*** (-2.03)** (-2.39)**
Birth rate - 0.173 - 0.313 - -
(2.88)*** (3.47)***
Mortality rate - - -0.049 -0.064 - -
(-4.00)*** (-5.02)***
Number of observations 526 513 484 481 526 526
R2 0.95 0.95 0.94 0.94 0.9
GDP = gross domestic product, GLS = generalized least squares.

Source: Authors’ estimates
The effect of the young-age dependency ratio varies widely across categories.
It is estimated to have a negative but significant effect on old age, health,
survivors, and housing spending. These results appear to drive the results found
at the aggregate level. By contrast, the young-age dependency ratio would be
positively correlated to family spending, which includes child subsidies.
Given the complexity of social policy and behaviors, this paper only presents
incomplete evidence and is subject to important caveats. As already mentioned
for savings, there may also be some endogenous issues associated with the
Table 5.7: Social Spending Equations by Category
Unemploy-
Old Incapacity- ment
Age Survivors related Health Family Active Benefits Housing Others
Revenue 0.018 0.03 0.046 0.048 0.028 0.02 0.042 -0.011 0.027
-1.06 (4.11)*** (3.80)*** (4.97)*** (3.62)*** (3.79)*** (4.72)*** (-3.03)*** (6.45)***
GDP growth -0.067 0.006 -0.013 -0.064 -0.027 0.001 -0.028 -0.005 -0.003
(-3.82)*** -0.86 (-1.51) (-6.35)*** (-3.63) -0.17 (-3.45)*** (1.86)* (-0.91)
GDP per capita 0.262 -1.618 -0.007 0.011 1.469 -0.286 -0.334 0.115 -0.063
-0.48 (-6.35)*** (-2.15)* (3.89)*** (6.82)** (-2.14)** (-2.24)** (1.89)* (-0.79)

Unemploy-
ment rate 0.097 -0.024 -0.084 -0.022 0.031 0.015 0.14 0.016 0.019
(5.66)*** (-3.34)*** (-0.33) (-2.04)** (3.15)*** (2.39)** (12.08)*** (5.13)*** (4.10)***

Old-age
dependency
ratio 0.156 0.003 -0.041 0.037 0.003 0.036 -0.026 0.016 0.008
(6.98)*** -0.3 (-2.52)** (2.19)** -0.22 (4.09)*** (-2.20)** (2.80)*** -1.28

Young-age
dependency
ratio -0.121 -0.025 0.009 -0.016 0.027 0.002 -0.004 -0.006 0.001
(-5.86)*** (-2.84)*** -1.18 (-1.67)* (3.70)*** -0.7 (-0.46) (-2.20)*** -0.32

Number of
observations 526 526 526 533 525 491 516 439 498
R2 0.94 0.85 0.92 0.87 0.92 0.87 0.9 0.91 0.91

Source: Authors’ estimates
estimation of equation (2), as income level and growth may also be determined
by social spending. One alternative would be to estimate a system combining
social spending and growth equations (Lindert 1996), but this would also mean
disregarding the time dimension of the data. Finally, the relation between
social spending and age could be non-monotonic.12
12
Lindert (1996) found that the squared old-age dependency ratio has a negative and significant effect
on most spending categories, one main exception being health-care spending.
5. Conclusion
The empirical analysis undertaken in this paper points to a significant effect
of the age distribution of the population on savings in both OECD and Asia-
Pacific countries. Old-age dependency ratios were usually found to have a
negative correlation to private or household saving. The effect of the young-
age dependency ratio was less marked and was not always significant. The
age structure also appears to influence social spending in OECD countries
but, as expected, its effect varies across spending categories. In particular, an
increasing aging population is likely to translate into increasing spending on
old-age pensions.
There is no straightforward leap from these conclusions to results on
national savings. Indeed, one would need to assume that the current share
in GDP of other forms of government consumption, revenues, and social
contributions remains unchanged. Moreover, one would have to assume that
there are no offsetting increases in private savings arising from the decline in
public sector savings. This assumption would run counter to empirical studies
that suggest significant Ricardian effects (e.g., Schmidt-Hebbel, Webb, and
Corsetti 1992).
Nonetheless, the results have important policy implications. First,
population aging is likely to depress private saving and thus be detrimental
to private investment and consequently long-term output. Policy should be
put in place to temper the effect of an aging population. Promising avenues
include pension and labor market policies encouraging private saving and
employment of older workers. A prerequisite will be to ensure that financial
markets are consistent with the optimal allocation of investments. For example,
expanding the market for annuities could reduce any tendency to over-save
for precautionary motives, and developing a reverse, or lifetime, mortgage
scheme would make it easier to use non-liquid assets possibly more efficiently
(Oliveira Martins et al. 2005). Secondly, it is important to account for the
effect of age structure and its evolution over time when assessing the impact
of social measures. This could inform the choice of social-spending categories
that should be prioritized and could optimize the cost-efficiency of measures
taken.
References
Ando, A., and F. Modigliani. 1963. The Life Cycle Hypothesis of Saving: Aggregate
Implications and Tests. American Economic Review 53: 55–84.
Spending Boost Consumption? IMF Working Paper 10/16. Washington, DC: IMF.
Breyer, F. 1999. Lebenserwartung, Kosten des Sterbens und die Prognose von
Gesundheitsausgaben. Jahrbuch für Wirtschaftwissenschaften 50(1): 53–65.
Dang, T.-T., P. Antolin, and H. Oxley. 2001. Fiscal Implications of Ageing: Projection of
Age-related Spending. OECD Economics Department Working Paper No. 305,
September. Paris: OECD.
De Nardi, M., E. French, and J. Bailey Jones. 2009. Life Expectancy and Old Age Savings.
NBER Working Paper No. 14653. Cambridge, MA: NBER.
Disney, R. 1996. Ageing and Saving. Fiscal Studies 17: 83–102.
Dynan, K., J. Skinner, and S. P. Zeldes. 2004. Do the Rich Save More? Journal of Political
Economy 112(2): 397–444.
Feldstein, M. 1980. International Differences in Social Security and Saving. Journal of
Public Economics 14: 225–244.
Gruenberg, E. 1977. The Failure of Success. Milbank Quarterly 55: 3–24.
Heller, P. and S. Symansky. 1997. Implications for Savings of Aging in the Asian “Tigers.”
IMF Working Paper 97/136. Washington, DC: IMF.
Horioka, C. 1991. The Determinants of Japan’s Saving Rate: The Impact of the Age
Structure of the Population and Other Factors. Economic Studies Quarterly 42(3):
237–253.
————. 2009. The (Dis)Saving Behavior of the Aged in Japan. NBER Working Paper 15601.
Cambridge, MA: NBER.
Jones, C. 2003. Why Have Health Expenditures as a Share of GDP Risen so Much? Center
for the Economics and Demography of Aging papers. Berkeley: University of
California and NBER Working Paper 9325. Berkeley, CA: University of California.
Khoman, E., and M. Weale. 2007. Development of Scenario for Health Expenditure in the
European Member States. Final report of work package 8 of the Ageing, Health
Status, and Determinants of Health Expenditure (AHEAD) Project, undertaken
by the National Institute of Economic and Social Research, under the EC 6th
Research Framework Programme.
Lindert, P. 1996. What Limits Social Spending? Explorations in Economic History 33: 1–34.
Masson, P., T. Fayoumi, and H. Somei. 1995. Saving Behavior in Industrial and Developing
Countries. In Staff Studies for the World Economic Outlook. Washington, DC: IMF.
Modigliani, F., and A. Sterling. 1983. Determinants of Private Saving with Special
Reference to the Role of Social Security—Cross-Country Tests. In The Determinants
of National Saving and Wealth, edited by F. Modigliani and R. Hemming. London:
Macmillan.
OECD. 2009a. Pension at a Glance. Paris: OECD.
————. 2009b. Society at a Glance Asia/Pacific. Paris: OECD.
Okunade, A., and V. Murthy. 2002. Technology as a “Major Driver” of Health Care Costs:
A Cointegration Analysis of the Newhouse Conjecture. Journal of Health Economics
21: 147–159.
Oliveira Martins, J., F. Gonand, P. Antolin, C. de la Maisonneuve, and K-Y. Yoo. 2005. The
Impact of Ageing on Demand, Factor Markets and Growth. OECD Economics
Department Working Paper No. 420. Paris: OECD.
Schmidt-Hebbel, K., S. B. Webb, and G. Corsetti. 1992. Household Saving in Developing
Countries: First Cross-Country Evidence. World Bank Economic Review 6: 529–547.
Seshamani, M., and A. Gray. 2004a. Ageing and Health-care Expenditure: the Red
Herring Argument Revisited. Health Economics 13(4): 303–314.
————. 2004b. A Longitudinal Study of the Effects of Age and Time to Death on Hospital
Costs. Health Economics 23(2): 217–235.
Weil, D. N. 1994. The Saving of the Elderly in Micro and Macro Data. Quarterly Journal of
Economics 109(1): 55–81.
Zweifel, P., S. Felder, and M. Meiers. 1999. Ageing and Health Care Expenditure: A Red
Herring? Health Economics 8(5): 485–496.
Comments
Hyungpyo Moon
This chapter empirically examines the impact of the age structure on private
savings and social spending using the new cross-country panel data covering
OECD countries and Asia-Pacific economies. In particular, the authors find new
empirical evidence such that: (i) population aging decreases private savings in
both OECD and Asia-Pacific economies and increases social spending in OECD
countries; (ii) the fertility rate, by affecting the young-age dependency rate, is
also a significant determinant of private saving and social spending; and (iii)
population aging has no significant impact on public health spending in OECD
countries. This study will not only add new insight to academic research, but
also provide both developed and developing countries with important policy
implications for attempts to temper the impact of aging populations.
1. Savings and Aging

One interesting empirical finding of the chapter is that savings in Asia-
Pacific economies are more sensitive to population aging compared to
OECD countries. Estimated results show that a 1 percentage point increase
in the old-age dependency ratio would reduce the household savings rate by
0.2–0.6 percentage point in OECD countries while it reduces private savings
rate by close to 2 percentage points in Asia-Pacific economies. Making a
direct comparison between these two groups of economies is difficult as the
dependent variables are different. Nevertheless, it seems that the impacts of
demographic change on savings in the private sector are significantly greater
in Asia-Pacific economies. Although the chapter did not elaborate this finding,
it would be worthwhile to consider what underlying factors could explain
such a difference.
According to the life-cycle hypothesis, population aging lowers the

household savings rate because there will be more elderly dissavers. The finding
in the paper implies that the elderly in Asia-Pacific economies tend to dissave
faster than those in OECD countries. How can this difference in the pace of
dissaving of the elderly be explained? One reason for it may be a difference
in bequest motive, as the stronger the bequest motive, the slower the pace
of dissaving of the elderly. Higher labor market participation rate may be
another explanation, as the elderly with a higher labor income would be less
likely to dissave. In addition, a difference in the institutional backgrounds,
including the social protection system, can cause a difference in the saving
behavior of the elderly.
One interesting example is shown in Figure 5.8, which compares the
pattern of net worth holdings of the average household by age cohorts in
Korea and the US. The asset accumulation patterns of average households
in two countries show a sharp contrast. Until people are in their mid-50s,
US households accumulate their assets relatively faster compared to Korean
households. At a later stage, Korean households decumulate their assets much
more quickly, producing a humped-shaped asset accumulation profile across
Figure 5.8 Comparison of Net Worth Holdings by Age Cohorts

(Korea versus the US)
KRW 10,000s (US$ 10,000s)
Source: Korea National Statistics Office, National Wurvey of Household Income and Expenditure, 2000. Available
at: http://kostat.go.kr/nso_main/nsoMainAction.do?method=sub&catgrp=eng2009&catid1=g02&catid2=g0
2d&catid3=g02da&catid=g02da. US Federal Reseve Board, Survey of Consumer Finances, 2001. Available at:
http://www.federalreserve.gov/pubs/oss/oss2/2001/scf2001home.html.
age cohorts. In contrast, elderly households in the US deplete their net worth
at a very slow rate.
Figure 5.8 also shows that Korean households possess a relatively higher
level of net worth in the early stage of working life. This would imply that the
bequest motive of the average Korean household is at least as strong as, or
even stronger, than that of US households. Also, according to OECD (2009),
the average retirement age of Korean workers was 71.2 years in 2007, which is
much higher than that of the US workers (64.6 years). Hence, the labor market
participation of the elderly cannot explain the faster dissaving of the elderly
in Korea. One possible explanation of the faster asset decumulation of this
population is the immaturity of the public pension plan. The National Pension
Plan in Korea was first introduced in 1988 and is still in its infancy—only
17.6% of those who are aged 65 and over received pension benefits in 2005.
This inadequate old-age income security system would make the elderly rely
heavily on self-insurance measures and to decumulate their net worth faster
than elderly households do in the US.
The above example can also be applied to the findings in this chapter; that
is, different levels of social protection, especially for the elderly, are the major
cause of different responses in private saving between OECD countries and
Asia-Pacific economies. Hence, it would be worthwhile to examine whether
this discrepancy of demographic impact on private saving still persists after
controlling for the differences in the amount and coverage of public pensions
and health insurance programs. The differences in the impact of the young-age
dependency ratio can also be explained in the same way, at least partially.
2. Aging and Social Spending

My second comment on Furceri and Mourougane’s chapter is directly linked to
the first one. In the chapter, the authors estimated the impacts of demographic
changes on private saving and social spending separately. However, these
impacts may be related. Population aging is indeed a global phenomenon,
although its speed differs from country to country depending on the fertility
rate and health status, among others. If the degree of aging is similar, I think
it is reasonable to assume that demographic shifts will influence each country
more or less to the same extent. However, the impacts of demographic changes
on the subsectors of an economy, private or public sectors, can vary across
countries, depending on their institutional settings.
For instance, people in a country with a well-developed social protection

system such as a generous pension and health insurance programs will be
less vulnerable to demographic changes. However, demographic shifts would
still place significant pressure on age-related public spending. In this case, the
impacts of demographic shocks are likely to fall on the public sector more
than on the private sector. On the other hand, people in a country where
the social protection measures are weak have to respond to demographic
change by adjusting their own saving behavior. Thus, the demographic changes
will impact more on the public sector in countries with a high level of social
spending, and more on the private sector in countries with a low level of
social spending. With the gradual expansion of the social safety net in Asia-
Pacific economies, we can expect that the pace of decline in private savings
will decelerate in the future.
The studies in the chapter do not include the demographic impacts on
social spending in Asian-Pacific countries. However, it would be interesting
to examine empirically whether the impacts of demographic shifts on private
and public sectors are negatively correlated.
References
OECD. 2009. Society at a Glance 2009: OECD Social Indicators. Paris: OECD.
Household Savings and Social Protection Policies in the PRC 109
VI
Household Savings and Social Protection Policies in
the People’s Republic of China
Ming Yan and Yi Pan
1. Introduction
This chapter consists of three parts. First, it analyzes existing research on savings
in the PRC, as well as policy implications for social protection policies. Second,
it traces the historical development of social policies in the PRC from 1949 to
2002, focusing on the impact of marketization and the state’s retreat from
social provision. Third, it discusses the recent efforts made by the government
to improve the social protection policies in the PRC. The chapter concludes by
arguing that the PRC’s strategic prioritization of economic growth over the
past several decades has led to unprecedented economic prosperity, but that
the social protection system has been severely jeopardized. Reconstruction of
the social protection system, though challenging, would generate far-reaching
positive effects in achieving more balanced socio-economic development in
the PRC.
We shall first clarify a number of key terms and their delimitations in
this chapter. Social policy as discussed here is understood as government
intervention to provide social services, with complementary support to the
socially disadvantaged and needy through statutory regulation, including a
range of unemployment and social security benefits (Hall and Midgley 2004).
Social protection policies, or social policies, in the PRC context include social
insurance (or social security, shehui baozhang), social assistance, welfare services
The authors acknowledge Tian Feng for his input and technical support during the initial writing of
the chapter. The authors are also grateful for the comments and suggestions made by the reviewers at
ADBI.
and benefits, the special care and placement system, and social aid and charity
(Figure 6.1). In this chapter, we will mostly focus on social insurance and
social assistance.
Figure 6.1: Social Protection Policies in the PRC
“Social Welfare” refers to services and benefits administrated by Civil Affairs government branches to such
groups as children, the elderly, and people with disabilities. The “special care and placement system” consists of
monetary compensation and job placement policies intended for servicemen and their families, and the families
of martyrs. PRC= People’s Republic of China.
2. PRC Savings and Implications for Social Policy

Savings in the PRC have been a major concern in recent years for scholars
and policymakers within and outside of the PRC (Horioka and Wang 2007;
He, Feng, and Sato 2008; Anderson 2009; Zhou, Zhang, and Li 2009). For
developing countries, encouraging private savings for capital formation appears
necessary because of immature capital markets and liquidity constraints
on corporations and households (Schmidt-Hebbel, Servén, and Solimano
1996). Since the economic reform and opening up, the high savings rate has
contributed significantly to the PRC’s economic growth. Prior to economic
reform, from 1952 to 1977, the average household savings rate was 4.9% and
nominal GDP growth rate averaged 7.2%; from 1978 to 2000, the average
household savings rate rose to 21.5% and the average nominal GDP growth
rate reached 16.4% (Zhang 2009).1 Thus, the PRC’s economic growth can be
1
Here, Zhang (2009) cited a study by Modigliani and Cao (2004) on the PRC historical household savings,
which, due to data limitations, measured PRC household savings through the “increase in personal
wealth.” However, the latter provided data from 1953 (not 1952) to 2000.
characterized as high savings, high investment, high export, and high growth.
Nevertheless, the PRC’s economic growth has been increasingly hampered by
such factors as inefficient investment and resource utilization, appreciation
pressure on the yuan, international trade tensions, and the potential decline of
exports due to the international financial crisis. As a result, as weak domestic
demand is being perceived as limiting long-term economic growth, the high
savings rate has turned into a “problem” to be addressed.
The PRC’s overall savings are high by any standard. Its national savings rate
has long been among the highest in the world and reached 49.9% in 2007, up
from 37.5% in 1998 (Zhou 2009). Rising government and corporate savings are
responsible for much of this increase. Table 6.1 shows the changing national
savings rate and sector contributions in the PRC. One study found that the PRC
government savings rate is comparably high by international standards and
has indeed experienced rapid growth since 2000 (Zhang 2009). Another study
Table 6.1: National Savings and Sector Contributions

in the PRC, 1992–2005
Contributions of the Three Sectors

Savings Rate in the Three Sectors (% of GDP) (% of total national savings)
National
Savings
Year Rate Residents Corporations Government Residents Corporations Government
1992 40.29 21.08 13.33 5.89 52.30 33.10 14.60
1993 41.72 19.32 16.15 6.24 46.30 38.70 15.00
1994 42.73 21.49 16.02 5.22 50.30 37.50 12.20
1995 41.62 20.04 16.70 4.88 48.10 40.10 11.70
1996 40.32 21.32 13.57 5.43 52.90 33.70 13.50
1997 40.76 20.75 14.73 5.65 50.00 36.10 13.90
1998 39.98 20.39 14.33 5.27 51.00 35.80 13.20
1999 38.61 18.55 14.31 5.75 48.00 37.10 14.90
2000 38.50 16.50 15.65 6.36 42.80 40.60 16.50
2001 38.89 16.18 15.14 7.57 41.60 38.90 19.50
2002 40.20 18.63 14.32 7.24 46.40 35.60 18.00
2003 42.90 18.11 15.47 9.32 42.20 36.10 21.70
2004 46.10 18.30 21.79 6.02 39.70 47.30 13.10
2005 47.45 21.16 20.04 6.26 44.60 42.20 13.20
GDP = gross domestic product, PRC = People’s Republic of China.

Source: Chen (2009).
focused on the growing corporate savings contributed by stunning expansion

in heavy industry and exports, and recommended corporate ownership reform
to grant profit transfer to shareholders, reduction of excess capacity creation
in key metals and material sectors, and raising of the yuan exchange rate
(Anderson 2009). Still another report also emphasizes the increasing corporate
savings but sees the problem rooted in the economic transition, during which
corporations are generating excessive profits without shouldering sufficient
social security responsibilities for employees (Zhou 2009).
Research on household savings in the PRC offers various findings and
interpretations. Although all studies acknowledge the consistently high
household savings rate, some attribute it to a distinctive cultural tradition
(Zhou 2009). One study analyzed the social dimension of the savings rate in
terms of the urban–rural divide among different income groups (Chen 2009).
As Table 6.2 demonstrates, Chen found that while the household savings rate in
urban areas rose steadily, in rural areas it increased to a peak of 28.6% in 1999
before declining continuously. Further, when urban households are divided
Table 6.2: PRC Household Savings Rate by Income Group, Urban and Rural
(% of GDP)
Income Level 1995 1998 2000 2001 2002 2003 2004 2005 2006 2007
Urban Households 17.4 20.2 20.4 22.6 21.7 23.1 23.8 24.3 26.0 27.5
Lowest 20% –3.3 7.0 7.4 7.7 6.8 6.9 6.8 7.7 10.2 9.8
Lower-middle 20% 3.5 14.7 14.6 15.1 14.7 15.2 15.4 16.9 19.1 20.0
Middle 20% 6.8 18.4 18.7 19.4 18.1 19.7 20.4 20.5 23.0 24.5
Upper-Middle 20% 10.3 21.8 21.3 23.6 21.8 22.7 24.5 25.3 27.3 29.4
Highest 20% 18.0 27.8 28.0 31.9 28.7 30.9 31.6 32.0 32.9 34.5
Rural Households 16.8 26.4 25.9 26.4 25.9 25.9 25.6 21.5 21.1 22.1
Lowest 20% –21.8 –21.3 –17.4 –23.0 –23.9 –45.1 –37.5 –37.4
Lower-middle 20% 14.4 14.6 15.4 14.3 14.2 5.2 8.2 8.7
Middle 20% 25.1 24.9 24.0 23.8 24.4 18.3 18.5 19.7
Upper-middle 20% 32.2 32.7 31.1 31.7 31.8 28.1 27.4 28.2
Highest 20% 40.5 40.2 40.7 40.8 40.4 40.7 37.7 38.8
GDP = gross domestic product, PRC = People’s Republic of China.

into seven income groups, their proportional contributions to the savings rates
during 1995–2007 can be calculated after weighting population and household
size among the various income groups. As can be surmised from Figure 6.2, the
high-income urban households (high and highest income groups combined)
contributed 63.5% on average to the overall urban household savings rate, even
though they only accounted for 20% of the total households. The remaining
80% of the urban households had an average savings rate of only 8.8%, far below
the 23.5% average for all urban households. Similarly, the high-income rural
households, accounting for 20% of all rural households, contributed 64.5% on
average to the overall household savings rate for rural households (Figure 6.3).
The savings rate of the remaining 80% rural households averaged 9.2%, far
lower than the average of 25.0% for all rural households. The lowest-income
rural households were consistently in debt, having a negative savings rate (Table
6.2). In other words, the critical issue seems to be structural inequality rather
than a high savings rate that may be due to unequal income distribution, as
reflected in the continuously rising Gini coefficient, from 0.454 in 2002 to
0.47 in 2007 (Li and Luo 2009).
Figure 6.2: Proportional Contributions to Urban Household Savings Rate

by Different Income Groups

Figure 6.3: Proportional Contributions to Rural Household Savings Rate by

Different Income Groups
Meanwhile, because of statistical problems in studying savings in the PRC

(described by Kraay 2000), studies on the PRC’s consistently high household
savings have used various analytical models and provide different economic
or demographic explanations. Some argue that the PRC’s high household
savings can be explained in terms of high economic growth and characteristic
life cycle; the latter has been reflected in the declining young-age dependency
ratio resulting from the “one-child policy” (Modigliani and Cao 2004; Horioka
and Wang 2007). Some researchers have found that the household savings
rate has been growing faster than both the economic growth rate and the
household disposable income growth rate, indicating a high precautionary
savings motivation (Zheng 2007). In the larger context of rapid socio-economic
transition, the people of the PRC showed strong motivation of precautionary
savings, due particularly to the unpredictability of employment and income
prospects (Meng 2003; Meng and Hu 2008; Zhang, X. 2009). See figures 6.4
and 6.5 comparing savings and consumption behavior in PRC.
The concept of precautionary savings holds that people save to smooth
current and future consumption when they find their future income and
expenditure unpredictable. Feldstein (1974) revealed that the manner in
which social insurance programs are financed can have major implications
for savings. Some studies have shown social insurance or protection policies
Figure 6.4: Comparison of Savings, Income, and Consumption Per Capita
of Urban Residents, 1990–2006
Source: Meng and Hu (2008).
Figure 6.5: Comparison of the Propensity to Save and the Propensity to

Consume, 1990–2006
Source: Meng and Hu (2008).

to have generally been an effective way to reduce consumers’ uncertainty

so that they may increase consumption and reduce precautionary savings
(Abel 1985; Kotlikoff, Spivak, and Shoven 1986; Hubbard 1986). Yet other
studies have found an insignificant relationship (Hurst and Ziliak 2006) or an
uncertain relationship (Auerbach and Kotlikoff 1984) between welfare policy
and household savings.
The relationship between social policy and household savings in the
PRC is still a new area of study with limited findings. Some have tried to
discern whether and how much the old-age insurance affects household
saving, and have shown that the pension reform in the mid-1990s led to
the reduction of households’ pension wealth, which resulted in increased
saving motivation (He, Feng, and Sato 2008). Others have focused on family
medical expenditure and health care insurance, arguing for well-established
medical insurance to reduce precautionary savings (Chen and Zhang 2007).
One study noted the decrease of household savings during the social security
reform pilot in Liaoning Province and suggested that households may have
taken into consideration the establishment of their individual accounts and
the expected income from the overall public pension program and adjusted
their personal savings behavior (World Bank 2006). Researchers in the PRC
have reached a consensus that, due to the restructuring of the SOEs and
marketization of social benefits provision in the areas of old-age insurance,
housing, medical care, and education, PRC households have increasingly
found it necessary to maximize savings for future uncertainty. Various
researchers have thus offered policy recommendations for social security
improvement (Zhang, M. 2009; Zhang, J.H. 2008; Zheng 2007; Xi, Geng,
and Zhang 2007; Guo 2006).
3. Social Policy in the PRC, 1979–2002

As social policy making and implementation is inseparable from the given
cultural, political, and economic conditions, the development of social policy
in the contemporary PRC has reflected the socio-economic characteristics of
PRC’s two distinct periods: the command economy period (1949–1978) and
the post-reform period (1979–present). It is necessary to distinguish between
these two.
3.1 Social Policy Under the PRC’s Command Economy
The founding of the Communist government in 1949 set the keystone of the
PRC’s reconstruction: state socialism. It promised to cure social problems such
as severe poverty as well as their roots—private ownership; the Revolution
derived its legitimacy from this promise. Measures taken up included state-
ification or collectivization of productive and major consumption assets, a
high priority on industrialization, and the imposition of totalitarian social
control. In order to speed up industrial development, the PRC government
implemented policies to increase capital accumulation from agriculture via
control of grain production and distribution and a strict household registration
policy (known as hukou), which restricted mass migration from the rural areas
to the cities. This gave rise to a gradual urban–rural divide and, hence, a social
protection system divided along urban–rural lines (Hussain 1994).
In the urban areas, social provision was primarily employment-based. In
the early 1950s, priority was placed on war, disaster relief, and social assistance
as the unemployment rate rose to 50% and the unemployed numbered nearly
3.33 million. The unemployment relief fund was set up using contributions
of the state and enterprises, as well as mandatory wage deduction from the
working population (Liu 2004). Soon, attention began to shift toward production
and employment. In February 1951, the Regulation on Labor Insurance was
introduced. It required enterprises to set aside 3% of the total wages for the
insurance of old age, sickness, disability, maternity, and death for their workers
(State Council 1951). The pension was typically 35–70% of the standard wage
depending on the recipient’s number of years in employment. The absence of
unemployment insurance from the Regulation on Labor Insurance was not
due to mere neglect; instead, it implied the then-dominant belief in work as
a right or entitlement. True, open unemployment was practically nonexistent
in the PRC until the 1990s. It was also in the 1950s that state-sponsored
programs of health care and retirement benefits for government employees
were implemented (State Council 1952, 1955).
Although production, employment, and social provision were controlled by
the state, it was danwei in which the state provision was primarily embodied
in urban areas. Danwei, which literally means “unit” in Chinese, is close to
the English term “workplace.” Danwei as workplace refers to the institutional
arrangements through which the individual seeks employment and thus draws
the source of livelihood. However, during the period of the command economy in
the PRC, it was through the affiliation with danwei that the individual was ensured
permanent employment and labor-related benefits such as sick leave, maternity,

disability, and old-age pension. Moreover, as a state employee, one would be
provided with a comprehensive social protection scheme such as subsidized food
service, medical care, recreation, and various allowances (utilities, transport, and
newspaper and books). Danwei as a “mini-society” also provided workers and
their families with childcare, school, and housing. In exchange, the individual
was extremely limited in terms of work mobility. Thus, danwei was more than
a productive and social protection unit; it was integrated with the centralized
redistributive system, and served as an essential tool of totalitarian control in
economic, social, and political terms (Walder 1986; Bray 2005).
In the rural areas, where more than 80% of the PRC population resided,
the state exerted its control through communes or collectives established
around the late 1950s. The distributive principal was “state first, collective
second, and individual last.” Land served as the source of livelihoods and family
served as the provider of basic protection, which was further supplemented by
collective welfare and mutual aid. Collective welfare included public service
institutions such as day cares, nursing homes, and medical clinics (Standing
Committee of the National People’s Congress 1956). In 1956, the government
established a regulation to ensure those who could neither farm nor depend on
familial support the “five guarantees” of food, clothes, housing, heat, and burial
upon death by the commune (Standing Committee of the National People’s
Congress 1956). In addition to those entitled to receive the “five guarantees,”
the commune was obliged to provide financial or human support for family
members of service members or martyrs. A rural cooperative medical service
system was also established in the vast majority of rural collectives. In this
system, individual farmers paid a small portion into a collective fund for
basic health care and emergency service (Deng 1992). The amount of 3–5%
of public funds was to be set aside from the total disposable income by the
brigade (the lower level and basic financial unit of the commune) for social
security and welfare development.2
The Ministry of Civil Affairs, which had been primarily responsible for
disaster relief and social welfare, was abolished during the Cultural Revolution
(1966–1976), during which economic and social development in the PRC was
severely interrupted. The Ministry of Civil Affairs was reinstated in 1978.
In brief, production and social protection under a command economy can
be characterized as full employment, comprehensive welfare, low wages, and
2
See Item 26 in Central Committee of the Chinese Communist Party (1961).
high subsidy. Within such a system, the individual was not personally obliged
to contribute to social insurance, yet the social welfare benefits one received
encompassed hidden labor compensation as the wage level remained low for
many years (Zheng 2002). Some economic growth statistics from 1952 to 1980
bear this statement out: the total industrial and agricultural output increased
8.1 times and industrial fixed assets investment 26.0 times, while the gross
national product increased 3.2 times and per capita consumption only 1.0
time; yet, from 1957 to 1978, the average wage level increased a mere 0.34%
(Leung and Nann 1995). In other words, under the command economy, the
strategic accumulation of industrial investment far exceeded the improvement
of people’s lives, and consumption was kept at a subsistence level.
3.2 Social Policy during the Post-Reform Period (1979–2002)

The economic reform and open door policy officially beginning in 1979 did
not only result in unprecedented economic growth. More importantly, it gave
rise to significant changes in the social protection scheme in the rural and
urban PRC.
The reform initiated in the rural areas in 1979 led quickly to the establishment
of the household responsibility system and the dissolution of the commune,
which allowed more freedom for peasants in entrepreneurial agricultural
production and increased sales surplus at market price. Higher productivity
made for improved living standards and poverty reduction among peasants.
A major achievement of the economic reform was drastic poverty reduction
in rural areas from 250 million rural poor in 1978 to 34.52 million in 2007
(Ministry of Civil Affairs 2007). As the collective organization on which social
protection in the rural PRC was based was severely undermined, the “five
guarantees” were made a state mandate, while the cooperative medical service
collapsed (Zheng 2002). From 1987 to 1997, under the auspices of the Ministry
of the Civil Affairs, experiments were conducted on the reconstruction of rural
old-age insurance, rural cooperative disaster relief, and rural minimum living
security schemes. By the end of 1997, CNY12 billion had been accumulated by
82 million peasants of approximately 2,000 counties nationwide for old-age
insurance (Duoji 1998). However, the endeavor was halted in 1998 when the
urban old-age insurance system suffered from a severe deficit.3
3
Although urban and rural old-age insurances were independent of each other, the fact that the long-
operating urban old-age insurance was found to be increasingly burdensome affected commitment
to the reconstruction of rural old-age insurance in the late 1990s.
Social protection policy in the urban areas during this period underwent
even more dramatic changes, which can be categorized as shifting from
enterprise or state security to social security, or the state’s gradual retreat
from social protection. After the official launching of urban economic reform
in 1984, reform of the social protection policy was intended to “liberalize”
the SOEs, relieving them from the overly heavy social “burdens” in order to
allow them to enhance their market competitiveness. In 1986, social security,
or shehui baozhang, appeared for the first time in the Seventh Five-Year Plan.
Social security was defined to include social insurance, social assistance, social
welfare, and special care and placement policies. Later, in the 1990s, mutual
aid and individual accounts were added to the social protection scheme (Duoji
1998). From the early 1980s and throughout the 1990s, in order to facilitate the
transformation of SOEs, a series of reforms were undertaken, including labor
contract, tax, and wage reforms, and insurance for old age, unemployment,
maternity, occupational injury, and health care. In 1999, the Regulation on
Unemployment Insurance was promulgated. In addition, commercial insurance
programs were reinstated after a long period of ceased practice. Also initiated
were community-based social service, an urban minimum living standard
guarantee scheme (dibao), and pilot projects of housing privatization. Through
the implementation of these programs and reforms, social protection policy was
redefined as a mechanism to accommodate the market transition of enterprises.
Several key sectors in which reforms were made are discussed below.
3.2.1 Old-Age Insurance

Preparations for old-age insurance reform were made from the end of the
1970s to the early 1980s, during which a number of regulations became
effective, setting different pension standards between enterprise workers
and employees of the public sector (Liu 2004). More fundamental reform
of the old-age insurance was undertaken during the 1990s. The process of
the reform can be traced in a series of State Council circulars, which are
summarized in Table 6.3. In 1991, the State Council issued the Decision on
Old-Age Insurance Scheme Reform for Enterprise Workers, which imposed
cost sharing between the government, the enterprise, and the individual for
funding of old-age insurance. It also proposed a multi-level scheme including
basic social old-age insurance, enterprise pension, and individual savings. In
1995, the Notice on Deepening the Reform of the Old-Age Insurance Scheme
for Enterprise Workers specified a combination of social pooling and individual
accounts. This was triggered by the labor market reform which was leading to
Table 6.3: Key Old-Age Insurance Regulations, 1991–2000
Policy Features
1991 Decision on Old-Age Insurance Scheme Reform Imposed cost sharing between the government,
for Enterprise Workers the enterprise, and the individual;
Proposed a multi-level scheme including basic
social old-age insurance, enterprise pension, and
individual accounts
1995 Notice on Deepening the Reform of the Old-Age Specified combination of social pooling and
Insurance Scheme for Enterprise Workers individual accounts
1997 Decision on Establishing a Uniform Basic Old-Age Set a unified contribution rate for both employers
Insurance Scheme for Enterprise Workers and employees; requested social pooling from
city/county to provincial level
2000 Notice on a Pilot Scheme to Improve the Basic Reduced individual contribution rate from 11% to
Old-Age Insurance Scheme in Liaoning Province 8%; gradually substantiated individual accounts
and made them partly accumulative
Sources: Based on Tang (2009) and Zheng and Sun (2008).
increased labor mobility; individual accounts were introduced to make pensions

portable. However, this policy led to the creation of hundreds of schemes
which varied between industries, sectors, cities, and regions. As an attempt
to resolve this problem, in 1997, the Decision on Establishing a Uniform Basic
Old-Age Insurance Scheme for Enterprise Workers set a unified contribution
rate for both employers and employees, and requested social pooling from
the city or county to the provincial level. Even though it was intended as an
accumulative scheme, a vast number of retirees who had not made contributions
needed to be covered by the existing scheme (the so-called legacy costs). As a
result, no real accumulation took place in individual accounts. In 2000, the
Notice on a Pilot Scheme to Improve the Basic Old-Age Insurance Scheme
in Liaoning Province reduced the individual contribution rate from 11% to
8%, and proposed to gradually substantiate individual accounts and to make
them at least partly “accumulative” (World Bank 2006; Piggott and Lu 2007;
Zheng and Sun 2008).
3.2.2 Medical Care Reform

During the command economy era, significant improvements had been
made on the overall health conditions of the population. For example, life
expectancy increased from 35.0 years in 1949 to 67.8 years in 1981. During
the post-reform period, in the context of broader socio-economic reforms and
increased financial pressures, the medical care system underwent gradual

marketization, signified by the reforms of medical care insurance, the health
care provision scheme, and the production and distribution of medicine. Basic
medical insurance reform was launched in 1998. The reform promoted: (i)
wide participation of employees in both formal and informal sectors, as well as
retirees; (ii) coverage by the government for only basic medical needs, with the
individual made responsible for a deductible of up to 10% of the local average
annual wage as well as fees above the “cap” (typically four times the local
average annual wage); (iii) shared contribution responsibilities between the
enterprise (6% of the total wage bills) and the individual (2% of wages); and (iv)
composition of an individual account into which 100% of personal contributions
and 30% of enterprise contributions were deposited, the remaining 70% of
Table 6.4: National Health Expenditure in the PRC, 1978–2007

(hundred million CNY)
Government Social Personal Government Social Personal

Health Health Health Health Health Health
Year Expenditure Expenditure* Expenditure Year Expenditure Expenditure Expenditure
1978 32.2 47.4 20.4 1993 19.7 38.1 42.2
1979 32.2 47.5 20.3 1994 19.4 36.6 43.9
1980 36.2 42.6 21.2 1995 18.0 35.6 46.4
1981 37.3 39.0 23.7 1996 17.0 32.3 50.6
1982 38.9 39.5 21.6 1997 16.4 30.8 52.8
1983 37.4 31.1 31.5 1998 16.0 29.1 54.8
1984 37.0 30.4 32.6 1999 15.8 28.3 55.9
1985 38.6 33.0 28.5 2000 15.5 25.6 59.0
1986 38.7 34.9 26.4 2001 15.9 24.1 60.0
1987 33.5 36.2 30.3 2002 15.7 26.6 57.7
1988 29.8 38.9 31.3 2003 17.0 27.2 55.9
1989 27.3 38.6 34.1 2004 17.0 29.3 53.6
1990 25.1 39.2 35.7 2005 17.9 29.9 52.2
1991 22.8 39.7 37.5 2006 18.1 32.6 49.3
1992 20.8 39.3 39.8 2007 20.4 34.5 45.2
*Social health expenditure mainly includes medical costs covered by health insurance companies and by danwei
of all types for their employees.
PRC = People’s Republic of China.
Source: Ministry of Health of the PRC (2009).
the enterprise contribution being deposited into the collective fund. Clinic
expenses would be paid out of the personal account while hospital bills and
treatment of serious illness would be partly covered by the collective fund (Hu
2009). Yet, the debate on the balance between social and economic roles of
medical care continued (Wang 2008). Most markedly, the shrinking role of
the government in medical care was reflected in its declining share in the total
national health expenditure, which dropped from a high of 38.7% in 1986 to
a record low of 15.5% in 2000; individual expenditure increased from a low of
20.4% in 1978 to a high of 60.0% in 2001 (Table 6.4 and Figure 6.6).
3.2.3 Unemployment Insurance and the Minimum Living Standard Guarantee

Scheme (dibao)
In the late 1990s, the full employment strategy adopted prior to the reform
encountered tremendous difficulties due to the increased number of bankrupt
enterprises and consequently massive number of “redundant” laborers. In
1998, the number of laid-off workers reached 10 million, and later totaled over
28 million (Liu 2004; He and Hua 2006). To ensure the basic means of living
of the unemployed, the “three security lines” were provided in sequence, all
in cash: first, basic living security aid for three years, then unemployment
Figure 6.6: National Health Expenditure in the PRC, 1978–2007

Source: Ministry of Health of the PRC (2009).
insurance for two years, and then dibao. These programs drew financial
resources from enterprises and the central and local governments. From
1998 to 2006, the central government transferred CNY100 billion from the
central government to the local authorities to provide for the basic living and
reemployment training of the laid-off workers. At the beginning of 1999, the
State Council stipulated in the new Regulation on Unemployment Insurance
that the “laid-off insurance” of the SOEs be extended to all the employees of
enterprises as well as public service organizations. It was also mandated that
an unemployment insurance fund be established, into which the enterprises
would contribute 2% of the total wage bills while the individual employees
would contribute 1%. Once unemployed, an individual could draw between 6
and 24 months of unemployment benefits depending on the number of years
that individual had been paying into the system (Hu 2009).
As to the dibao, its implementation was a gradual process. As early as
1993, pilot programs were carried out in a number of cities such as Shanghai
and Dalian. In 1997, the central authorities urged the establishment of dibao
nationwide by the end of 1999. The authorities’ goal was to provide monthly
cash assistance to laid-off workers and those categorized as the “three nos”: no
source of income, no ability to work, and no familial support. By September
1999, when the regulation on dibao for urban residents finally became effective,
not only was dibao adopted in all the 668 cities and 1,638 towns where county
governments were then seated, it was widely extended to all as long as they met
the income criterion. As of October 1999, the total number of dibao recipients
reached 2.82 million, of whom close to 80% were new (Tang 2003).
4. Reconstruction of Social Protection Scheme

As stated above, social policy reform in the PRC until the end of 1990s aimed
at facilitating enterprise restructuring and marketization, which marked a
dramatic turn from the prior ideals of state socialism. Since the beginning
of the 2000s, social policy in the PRC made another shift as the government
began to redefine its strategic goals. In 2003, during the third plenum of the
16th Congress of the Chinese Communist Party, the two key concepts advocated
were “scientific development” and the “five integrations” between urban and
rural areas, between regions, between social and economic development,
between humans and nature, and between domestic development and external
orientation. Along with the emphasis on striking a balance between economic
and social development, more attention and resources have been given to
social well-being. As noted by Hussain (2005: 271):
“Recent years have seen a marked shift away from single-minded
emphasis on economic growth toward the development of a
“harmonious society.” Progress in improving [the PRC’s] social
security system probably will be much quicker over the next 20
years than in the previous 20, but reform will still be piecemeal
and gradual.”
In particular, it was proposed at the 17th Congress of the PRC’s Communist
Party in 2007 that reforms should be made “to accelerate the development
of a social protection scheme to encompass both urban and rural residents”
(Hu 2007). Specified were 12 tasks aimed mainly at coverage expansion and
increase of social pooling of the existing social protection programs in the
urban areas and establishment of social protection in the rural areas:
t 1SPNPUFUIFSFGPSNPGUIFCBTJDPMEBHFJOTVSBODFPGUIFFOUFSQSJTFT QVCMJD
service institutions, and civil organizations, and explore the establishment
of old-age insurance in the rural areas.
t 1SPNPUFUIFEFWFMPQNFOUPGCBTJDNFEJDBMJOTVSBODFGPSVSCBOFNQMPZFFT
basic medical insurance for urban residents, and a new rural cooperative
medical care scheme.
t *NQSPWFVSCBOBOESVSBMdibao and gradually increase the amount of
payment.
t *NQSPWFUIFVOFNQMPZNFOUJOTVSBODF PDDVQBUJPOBMJOKVSZJOTVSBODF
and maternity insurance.
t 3BJTFUIFMFWFMPGSJTLQPPMJOHBOEEFTJHOBVOJmFETDIFNFUPFOTVSF
transferability of social insurance.
t "QQMZNVMUJQMFNFDIBOJTNTUPmOBODFUIFTPDJBMTFDVSJUZGVOE TUSFOHUIFO
surveillance, and preserve the value of the fund and ensure value
growth.
t *NQSPWFUIFTPDJBMBTTJTUBODFTDIFNF
t $POUJOVFQSPWJEJOHUIFTQFDJBMDBSFBOEKPCQMBDFNFOUGPSGBNJMJFTPG
service members and martyrs.
t 5BLFBIVNBOJTUBQQSPBDIBOEJNQSPWFUIFDPOEJUJPOTPGUIFQFPQMF
with disabilities.
t *NQSPWFPMEBHFDBSF
t 4USFOHUIFOEJTBTUFSQSFWFOUJPOBOESFMJFG
t *NQSPWFUIFQSPWJTJPOPGMPXSFOUIPVTJOHBOETQFFEVQUIFSFTPMVUJPO
of housing difficulties for urban low-income families.
What was proposed, it seemed, was a very ambitious plan for a universal
expansion of the social protection scheme. To see what progress has been
made and what obstacles have been encountered, we shall first look at the
recent accomplishments reflected in numbers.
4.1 Basic Old-Age Insurance in Urban Areas

At present, enterprises are required to contribute 20% of the total wage bills
and individual employees are required to contribute 8% of their wages to the
old-age insurance. When the employee reaches retirement age (60 for men
and 50 for women4), payment is to be drawn partly from the “integration
fund” in proportion to the local average wage, and partly from the personal
account based on the amount of contributions. The average monthly payment
increased to CNY914 (roughly US$134) in 2007 from CNY413 (roughly US$61)
in 1998.5 To expand coverage, in 1999, the State Council stipulated that basic
old-age insurance would extend to all types of enterprises as well as the self
employed. Between 2001 and 2003, a pilot project was executed in Liaoning
Province to strengthen the system by establishing funded individual accounts,
separately managed from the pay-as-you-go component (World Bank 2006);
afterwards, 12 additional provinces, districts, or cities participated in the
pilot program. By the end of November 2008, the program covered a total
of 217 million people (165 million working people and 52 million retirees), a
jump from 110 million in 1998, which made the program the largest pension
program in the world. The National Basic Old-Age Insurance Fund increased
from less than CNY150 billion in 1998 to CNY783.4 billion in 2007, an average
annual growth rate of 20%. Still, from 1998 to 2008, the central government
allocated CNY450 billion as a supplementary fund (Hu 2009; Tang 2009).
Experiments with rural old-age insurance are taking place in over 300
counties in the PRC with vastly diverse models concerning design, contributions,
benefits, and management (Zhao 2009).
4.2 Medical Insurance

Several programs have been developed to deal with the needs of various
populations. In October 2002, the central government proposed the
4
In the urban PRC, in general, the retirement age for women employed as blue-collar workers is 50 and
that for women employed as white-collar workers is 55.
5
Here, CNY6.8 = US$1
establishment of the New Rural Cooperative Medical Care (NRCMC), which
involved setting up accounts for individual peasants into which the individual
peasant pays CNY10 and the central government and local government each
contributes CNY20 annually. This modest insurance mainly covers hospital
visits and expenses related to serious illnesses. Between 2003 and 2004, policies
were issued to include in the medical insurance scheme employees in the
non-state and informal sectors, as well as rural-to-urban migrants. In 2007,
a new program was developed, which was intended to cover the non-working
urban population, including the young and the old. While contributions of
the individual and the government vary, the latter contributes no less than
CNY80 per year. Similar to the NRCMC, this program is for serious illnesses,
but benefit payments are higher than those of the NRCMC (Yao 2009; Wang
2008). In addition, there are a number of programs to provide supplementary
aid for medical emergencies or serious illnesses.
4.3 Social Assistance Programs

The “three security lines” discussed above, which were implemented in 1998,
have undergone changes. Since 2006, unemployment insurance has been
playing an increasingly significant role. By November 2008, 123.18 million
employees were covered by unemployment insurance and, each year, between
3 million and 4 million people benefited from it (Hu 2009).
With regard to urban dibao, as of 2008, there were 22.73 million recipients,
constituting 3.8% of the urban population. The number of recipients varied by
region; the more economically developed eastern region had a low percentage
of 1.7% and the western region had a high percentage of 6.0%, while the
central region was in between with 5.7%. The average monthly amount of
the benefit was CNY208 (US$31) per person, while the range was between
CNY150 and CNY400 (US$22–US$59). The financial sources of the program
were the central government (57.8%), provincial government (10.7%), and
local authorities (31.5%).
Between 2007 and 2008, dibao began to be widely adopted in the rural
areas. As of 2008, the number of beneficiaries had reached 38.58 million
people, or 5.3% of the total rural population. The benefit averaged CNY82
(US$12) per person per month, and it varied from CNY154 (US$23) in the
eastern region to CNY61 (US$9) in the central region and CNY31 (US$5) in
the western region. The financial sources were the central government (15.9%)
and the provincial government (23.5%). The old (60 years and above) made
up 27.7% of the recipients, while women accounted for 26.9%, the young (18
years old and younger) 11.0%, and people with disabilities 8.6% (Zhang and
Tang 2008).
Measures for setting the standard for benefits varied locally and could
be based on any of the following considerations: (i) national poverty line; (ii)
minimum wage, e.g., 40% of the given local minimum wage in the urban
areas, and 60% of the urban dibao in the rural areas; (iii) financial capacity
of the local authorities; (iv) the basic living needs and consumer price; or (v) a
proportion (26–30%) of the average gross income in the given area (Center for
Policy Studies of the Ministry of Civil Affairs 2009).
While the number of recipients of dibao in the urban areas seems to be
stable, coverage has been rapidly expanding in the rural areas. As of May 2009,
there were 23.36 million recipients in the urban areas and 44.56 million in
the rural areas; the number of rural recipients had increased by 5.98 million
compared with the number in 2008. In addition, there were 5.52 million rural
people who continued to receive the aid related to the “five guarantees” (Jia
2009). Further expansion of dibao is expected to continue. Some proposed that,
by 2012, dibao recipients should be increased to 8% of the total population
from 4.7% in 2008 (Zheng 2008).
While it is not easy to assess the effectiveness of the programs based on a
descriptive presentation of these numbers, we shall now turn to the challenges
lying ahead.
5. Major Problems and Challenges
The social protection scheme, as characterized by Xiaoyi Hu, an official in

charge of the social insurance division of the Ministry of Human Resources
and Social Security,6 may function in three respects: as a social safety net to
ease labor risks; as an adjustment mechanism of redistribution for lessening
income disparity; and as a shock absorber for social stability—especially during
the fluctuation of economic growth—by providing a minimum living standard
guarantee (Hu 2009). As seen in the discussion above, various social protection
programs in the PRC have indeed served as a social safety net in the sense that
they helped to secure the basic life needs of the most vulnerable groups. The
dibao is a prime example of this. Studies show that during the economic reform
6
Formerly the Ministry of Labor and Social Security.
throughout the 1990s, the poverty rate in urban areas was on the rise, mainly
due to massive unemployment and inadequate social assistance (Xue and Wei
2004; Li and Knight 2002). The dibao was meant to combat these issues, but
research findings indicate that while it has had a significant impact on poverty
alleviation, it has a very limited impact on narrowing income inequality in
the urban PRC. The policy implications are that the level of the dibao should
be raised and its coverage be expanded (Li and Luo 2009).
As many of the programs were established or modified in response to
massive unemployment in the urban PRC in the late 1990s, they surely played
the role of “shock absorber” for social stability at a time of potential turmoil.
Nevertheless, social protection policy has failed to serve as a redistributive
mechanism for lessening income disparity.
Social protection policy has contributed more to the urban–rural divide
for which the PRC has long been known. Until recently, more resources of
social provision have been devoted to urban than rural areas; even the quickly
expanding social assistance or insurance programs in the rural areas are based
on a much lower income maintenance standard. This “urban bias” is partly
due to the practice of regarding peasants as self-employed (Hussain 2005).
It also has to do with the assumed lower cost of living in rural areas. Yet,
while the urban–rural economic disparity reflected in the Gini coefficient is
dramatic and continues to grow, social policy has made the situation more
severe. According to a longitudinal household survey, in the beginning of the
twenty-first century, the social security benefits of urban residents in monetary
terms were equivalent to about 53% of the disposable income per capita, which
was 1.65 times the amount of gross income per capita of rural residents. Thus,
counting the differences in urban–rural social security benefits would enlarge
the urban–rural income gap from 3.1 times to 4.5 times and would increase the
Gini coefficient by about 10% (Li and Luo 2009). The urban–rural income gap
among low-income households is even wider partly due to the lower mobility
of the rural residents and partly due to the relatively better social protection
benefits for the urban residents (Luo 2006). A comparison of transfer income
of the urban and rural residents indicates similar results. Tables 6.5 and 6.6
show that urban households received far more transfer income than did rural
households in absolute terms as well as relative to total disposable income.
Further, even though between 1990 and 2007, transfer incomes received by
both urban and rural households were on the rise, the former increased much
more than the latter. Thus, transfer income contributes positively to urban–rural
income inequality (Yang and Chi 2008; Huang, Wang, and Wan 2003).
The priority accorded to urban areas with regard to medical care is even
more apparent, as reflected in the unequal distribution of health care resources.
Taking 2005 as an example, the gross health care expenditure per capita in
the PRC was CNY662.3, of which the urban residents used an average of
CNY1,123 per capita, while rural residents used an average of CNY319 (3.53
Table 6.5: Income Composition of Urban Households in the PRC
Income Composition 1990 1995 2000 2005 2006 2007
Per capita annual income (yuan) 1,516.21 4,279.02 6,295.91 11,320.77 12,719.19 14,908.61
Income from wages and salaries 1,149.70 3,390.21 4,480.50 7,797.54 8,766.96 10,234.76
Net business income 22.5 72.62 246.24 679.62 809.56 940.72
Income from properties 15.6 90.43 128.38 192.91 244.01 348.53
Income from transfers 328.41 725.76 1,440.78 2,650.70 2,898.66 3384.6
Disposable income 1,510.16 4,282.95 6,279.98 10,493.03 11,759.45 13,785.81

Transfer income/
Disposable income (%) 21.75 16.95 22.94 25.26 24.65 24.55

Source: National Bureau of Statistics of China (2006, 2008).
Table 6.6: Income Composition of Rural Households in the PRC
Income Composition 1990 1995 2000 2005 2006 2007
Per capita annual net income (yuan) 686.31 1,577.74 2,253.42 3,254.93 3,587.04 4,140.36
Income from wages and salaries 138.8 353.7 702.3 1,174.53 1,374.80 1,596.22
Income from household operations 518.55 1,125.79 1,427.27 1,844.53 1,930.96 2,193.67
Income from properties 28.96 40.98 45.04 88.45 100.5 128.22
Income from transfers 57.27 78.81 147.42 180.78 222.25

Transfer income/Per capita net
income (%) 3.63 3.5 4.53 5.04 5.37
Transfer income/
Disposable income (%) 21.75 16.95 22.94 25.26 24.65 24.55

Source: National Bureau of Statistics of China (2006, 2008).
times less the per capita expenditure of urban residents). With regard to the
individual proportion of this expenditure, it amounted to 7.87% of the annual
consumption expenditure of rural individuals, and 7.56% of that of urban
individuals, while the gross disposable income per capita in the rural areas was
only one-third that of the urban areas (Chinese [Hainan] Institute of Reform
and Development Studies 2008). Due to sampling issues, survey results vary
in terms of the absolute numbers of health care expenditures, yet patterns are
consistent. Researchers have found that the concentration of public medical
subsidies is even more severe than the income disparity reflected in the Gini
coefficient, and public medical subsidies are disproportionally allocated to the
better-offs (the urban population) rather than the poor and disadvantaged
groups (the rural population) (Wei and Gustafsson 2008). As for medical facilities
and professional staff, the severe unequal distribution is not surprising.
The focus of government spending on cities has been reinforced by large
geographic inequalities in health expenditures. Thus, government health
spending in Gansu, one of PRC’s poorest provinces, amounted to just CNY46
per person in 2003, while per capita spending in Shanghai and Tianjin, two
of PRC’s richest provinces, amounted to CNY218 and CNY153, respectively.
This is despite the fact that Gansu had far worse health indicators than either
Shanghai or Tianjin. These inequalities stem from the fact that the PRC’s
intergovernmental fiscal system weakened—but failed to break—the link
between local governments’ per capita income and their available resources
for spending (Wagstaff, Lindelow, and Wang 2009).
Within the city, the Matthew Effect 7 is no less serious. Social insurance
programs have been rapidly expanding due to policy promotion, but their
development is highly fragmented and uneven among sectors and regions.
Taking old-age insurance as an example, despite the fact that PRC pensioners
are a relatively small proportion of the total population, spending is nearly
two to three times the average for the Asia-Pacific region, and approaches
the average level in North Africa and the Middle East and Latin America and
the Caribbean. An index of standardized spending—the average pension as a
percentage of GDP per capita—suggests that spending in the PRC was nearly
the highest of all world regions, apart from sub-Saharan Africa. This indicates
that the pension system accords relative generosity to a narrow share of the
older population (Whiteford 2003). Among the pensioners, retirees of the
civil service and public service institutions are especially privileged, as their
7
The phenomenon in sociology where the rich get richer and the poor get poorer.
average pensions are respectively 109.1% and 86.6% more than that of the
enterprise retirees (Zheng and Sun 2008).
The relatively low number of contributors as a percentage of the labor
force has meant that the contributions required to finance these benefits are
relatively high, compared to other world regions. Slightly over half of the
urban employees participate in the old-age insurance scheme.8 As a result,
the PRC has the highest level of social security contributions among the lower-
income Asian countries. There is also a relatively high level of reliance on
employer rather than employee contributions in the PRC (Whiteford 2003).
This has partly led to the particular barriers to the efforts of expanding social
insurance coverage among the so-called floating population, i.e., rural-to-
urban migrants.
Rural-to-urban migrants have so far been a vulnerable group who have yet
to benefit from the social protection programs. Prior to 2005, enterprises were
required to treat the migrants equally with urban employees, which meant
mandated participation in the five social insurance programs. This policy
apparently met significant resistance from enterprises employing high numbers
of migrants; their objections were based on concern about the consequent rising
labor costs. As a result, local authorities in the Pearl River Delta region, where
large number of migrants are concentrated, allowed enterprises to disentangle
the five social insurances bundle and prioritize insurance for occupational
injury and serious illness with lower required contribution rates. Still, social
insurance coverage rates among migrants remain low compared with urban
employees: 20% of them have old-age insurance, 42% have occupational injury
insurance, and 35% have insurance for serious illness (PRC Department of
Population and Employment Statistics, Department of Planning and Finance,
Ministry of Labor and Social Security 2008).
Resistance to the migrant old-age insurance comes not only from employers.
Migrants themselves are not keen on the old-age insurance because of a built-in
transferability problem. As the program only allows benefits within the given
city or province after 15 years of accumulation, rural–urban migrants cannot
receive the pension if they leave the city or province. In addition, the policy
permits cashing out of personal contributions if current employment is ended,
but employer contributions remain in the social pooling account. Two problems
thus arise: (i) a proportion of migrants is always joining and exiting the old-age
8
As of the end of 2007, about 52% of urban employees participated in old age insurance (PRC Department
of Population and Employment Statistics, Department of Planning and Finance, Ministry of Labor
and Social Security 2008).
insurance as they change jobs, leading to increased management costs and
loss of benefits; and (ii) enterprise contributions to the social pooling account
for migrants end up benefiting urban employees.

This chapter has taken a broad perspective on PRC savings and social protection
policy. It first examined savings and argued that even though the overall PRC
savings are high, it was the savings by the government and corporations that have
been rising in recent years. The seemingly high household savings, which have
been shaped by economic growth and favorable demographic characteristics were
shown to be highly unevenly distributed once broken down in terms of household
income. Regardless of the pattern of savings, PRC scholars and policymakers
agree that social protection programs must be improved.
We then outlined the historical development of social policy in the
PRC, focusing on the main features of the old-age insurance, medical care,
unemployment insurance, and social safety nets as they were adapted to
address the pressing labor risks during marketization. We further discussed the
most recent moves by the PRC authorities of reinvestment in social protection
programs toward social harmony and balanced development. We argued that
despite the evidently effective poverty reduction function of some programs
such as unemployment insurance and the minimum living standard guarantee,
social policy in the PRC overall has been pro-urban and pro-state sector, and
has thus been counterproductive as a redistributive mechanism.
In conclusion, while unprecedented economic growth in the PRC has
generated a vast amount of wealth, socio-economic disparity has grown
increasingly stark. It remains a challenge for the PRC to reform its social
protection provision to address the needs accompanying the looming
structural changes, such as the rising living standards, aging and rising old-
age dependency ratio, fast pace of urbanization, and continued growth of the
private and informal sectors. The reform will definitely have to involve the
transfer of wealth from the state and enterprise sector to households, especially
those of lower income. It will also require financial reform between the central
and local authorities, especially in their responsibilities of social provision.
If the PRC can achieve equitable and sustainable social provision, it would
potentially benefit more people than any other national social protection
system in the world. Its bound-to-be innovative approach would be instructive
for both the developed and developing countries.
References
Abel, A. B. 1985. Precautionary Saving and Accidental Bequests. American Economic
Review 75(4): 777–791.
Anderson, J. 2009. The Myth of Chinese Savings. Far Eastern Economic Review 172(9):
24–30.
Auerbach, A., and L. Kotlikoff. 1984. An Examination of Empirical Tests of Social Security
and Savings. NBER Working Paper 730. Cambridge, MA: NBER.
Bray, D. 2005. Social Space and Governance in Urban China: The Danwei System from
Origins to Urban Reform. Palo Alto, CA: Stanford University Press.
Center for Policy Studies of the Ministry of Civil Affairs. 2009. Construction of the
Social Policy Support System of the Urban and Rural Poor Families in the PRC.
Unpublished report. Chinese.
Central Committee of the Chinese Communist Party. 1961. A Tentative Regulation on
the Operation of the Rural People’s Commune. 15 June. Available: http://news.
xinhuanet.com/ziliao/2005-01/24/content_2500797.htm. Chinese.
Chen, C. S. 2009. On Structural Features of China’s Household Savings and Expansion of
Consumption. China Development Review 11(3): 6–12. Chinese.
Chen, X. B, and Y. Zhang. 2007. The Mechanism Simulation of Medical Insurance System’s
Impact on the Household’s Consumption and Saving Decision Making. Journal of
Shanghai University of Finance and Economics 9(6): 55–62. Chinese.
Chinese (Hainan) Institute of Reform and Development Studies. 2008. People and
Livelihood, 100 Questions Concerning Shared Basic Public Service. Beijing: Chinese
Economic Press. Chinese.
Deng, D. S. 1992. A Comparative Study on Social Insurance. Beijing: Chinese Financial
Press. Chinese.
Duoji, C. R. 1998. Prospect of China’s Social Security in the Twenty-First Century. In
China’s Social Welfare and Social Progress Report, edited by Z. X. Shi and Y. Zhu.
Beijing: Social Sciences Documentary Press. Chinese.
Feldstein, M. S. 1974. Social Security, Induced Retirement, and Aggregate Capital
Accumulation. Journal of Political Economy 82(4): 905–926.
Guo, Y. T. 2006. Theoretical Analysis and Empirical Testing on the Precautionary Savings of
the Households in our Country. PhD dissertation, Jilin University. Chinese.
Hall, A., and J. Midgley. 2004. Social Policy for Development. London and Thousand Oaks,
CA: Sage Publications.
He, L .X., J. Feng, and H. Sato. 2008. Pension Reform and Household Saving: The Chinese
Case. Economic Studies 10: 117–130. Chinese.
He, P., and Y. F. Hua. 2006. Study on the Social Security Policy and Implementation
Measures of the Urban Poor Groups. Beijing: Chinese Labor and Social Security
Press. Chinese.
Horioka, C. Y., and J. M. Wang. 2007. The Determinants of Household Saving in China: A
Dynamic Panel Analysis of Provincial Data. Journal of Money, Credit and Banking
39(8): 2077–2096.
Hu, J. T. 2007. Report of Hu Jintao at the Seventeenth Congress of the CCP. Available: http://
news.xinhuanet.com/newscenter/2007-10/24/content_6938568.htm. Chinese.
Hu, X. Y. 2009. Social Security. In Social Security System in Contemporary China, edited by
X. Y. Hu. Beijing: Chinese Labor and Social Security Press. Chinese.
Huang, Z. H., M. Wang, and G. H. Wan. 2003. On Household Income Inequality in Our
Country: Analysis from the Transfer Income Perspective. Management World 3:
70–75. Chinese.
Hubbard, G. R. 1986. Pension Wealth and Individual Saving: Some New Evidence. Journal
of Money, Credit and Banking 18(2): 167–178.
Hurst, E., and J. Ziliak. 2006. Do Welfare Asset Limits Affect Household Saving? Evidence
from Welfare Reform. Journal of Human Resources 41(1): 46–71.
Hussain, A. 1994. Social Security in Present-Day China and Its Reform. American Economic
Review 84(2): 276–280.
————. 2005. Repair China’s Social Safety Net. Current History 104(683): 268–271.
Jia, C. F. 2009. Research Report on Legal Construction of the Ministry of Civil Affairs.
Internal circulation. Chinese.
Kotlikoff, L. J., A. Spivak, and J. B. Shoven. 1986. The Effect of Annuity Insurance on
Savings and Inequality. Journal of Labor Economics Supplement 4(3-2): s183–207.
Kraay, A. 2000. Household Saving in China. World Bank Economic Review 14(3): 545–70.
Leung, J. C. B., and R. Nann. 1995. Authority and Benevolence: Social Welfare in China.
Hong Kong, China: Chinese University Press.
Li, S., and J. Knight. 2002. Three Types of Poverty in Urban China. Economic Studies 10:
47–58, 95. Chinese.
Li, S., and C. L. Luo. 2009. Impact of Public Policy on Income Distribution in Our Country.
Chinese Social Sciences (newspaper). 12 November. Chinese.
Liu, G. X. 2004. A Brief History of Labor Security in the New China 1949–2003. Beijing:
Chinese Labor and Social Security Press. Chinese.
Luo, C. L. 2006. Dynamic Changes of the Income Disparity between Urban and Rural
Residents, 1988–2002. Finance and Economics Studies 32(9): 103–112. Chinese.
Meng, Q. P., and J. Y. Hu. 2008. Insufficient Social Security: A Factor Analysis of
Households Saving Growth in China. Journal of Shandong University 3: 113–119.
Chinese.
Meng, X. 2003. Unemployment, Consumption Smoothing, and Precautionary Saving in
Urban China. Journal of Comparative Economics 31(3): 465–485.
Ministry of Civil Affairs. 2007. Report on the Social Welfare of the Fourth Quarter 2007 of
the Ministry of Civil Affairs. Chinese.
Ministry of Health of the PRC. 2009. China Health Statistics Yearbook 2009. Available at:
http://www.moh.gov.cn/publicfiles/business/htmlfiles/zwgkzt/ptjnj/200908/42635.
htm. Chinese.
Modigliani, F., and S. L. Cao. 2004. The Chinese Saving Puzzle and the Life-cycle
Hypothesis. Journal of Economic Literature 42(1): 145–170.
National Bureau of Statistics of China. 2006. China Statistical Yearbook 2006. Available at:
http://www.stats.gov.cn/tjsj/ndsj. Chinese.
————. 2008. China Statistical Yearbook 2008. Available at: http://www.stats.gov.cn/tjsj/

ndsj. Chinese.
Piggott, J., and B. Lu. 2007. China-Pension Reform and the Development of Pension
Systems: An Evaluation of World Bank Assistance. World Bank Working Paper
39147. Washington, DC: World Bank.
PRC Department of Population and Employment Statistics, Department of Planning and
Finance, Ministry of Labor and Social Security. 2008. Chinese Labor and Social
Security Statistics Yearbook 2007. Beijing: PRC Department of Population and
Employment Statistics, Department of Planning and Finance, Ministry of Labor
and Social Security. Chinese.
Schmidt-Hebbel, K., L. Servén, and A. Solimano. 1996. Saving and Investment: Paradigms,
Puzzles, Policies. World Bank Research Observer 11(1): 87–117.
Standing Committee of the National People’s Congress. 1956. Model Regulations on
the Advanced Agricultural Production Collectives. Available at: http://news.
xinhuanet.com/ziliao/2004-12/30/content_2393677.htm. Chinese.
State Council. 1951.Regulations on the Labor Insurance of the People’s Republic of China.
Available at: http://news.xinhuanet.com/ziliao/2004-12/17/content_2347271.htm.
Chinese.
————. 1952. Instructions on the Medical Care of the Employees of Public Service
Institutions and Civil Organizations by the State Council. Available at: http://www.
fx120.net/zcfg/zc-yygl/jhycw/200404261255154087.htm. Chinese.
————. 1955. Tentative Rules Concerning the Retirement of Employees of the Public
Services. Available at: http://wenda.tianya.cn/wenda/thread?tid=019adc9d3
10a6c31. Chinese.
Tang, J. 2003. Poverty and Anti-Poverty Report in Urban China. Beijing: Huaxia Press.
Chinese.
Tang, Q. S. 2009. Old-Age Insurance in Our Country. In The Social Security System in
Contemporary China, edited by X. Y. Hu. Beijing: Chinese Labor and Social
Security Press. Chinese.
Wagstaff, A., M. Lindelow, and S. Y. Wang. 2009. Reforming China’s Rural Health System.
Washington, DC: World Bank.
Walder, A. 1986. Communist Neo-traditionalism: Work and Authority in Chinese Industry.
Berkeley, CA: University of California Press.
Wang, H. F. 2008. Thirty Years of the Health Care Reform in Our Country. In Report on
Economic Development and Institutional Reform in China No. 1, edited by D. T. Zou.
Beijing: Social Sciences Documentary Press. Chinese.
Wei, Z., and B. Gustafsson. 2008. Analysis of the Inequity of Medical Expenditures of
Chinese Households. In Income Distribution of Chinese Households III, edited by S.
Li et al. Beijing: Beijing Normal University Press. Chinese.
Whiteford, P. 2003. From Enterprise Protection to Social Protection Pension Reform in
China. Global Social Policy 3(1): 45–77.
World Bank. 2006. China-Evaluation of the Liaoning Social Security Reform Pilot. Working
Paper 38183. Washington, DC: World Bank.
Xi, Z. X., L. N. Geng, and W. T. Zhang. 2007. The Impact of Change of the Social Security
System on Household Savings. Economic Issues Exploration 5: 85–89. Chinese.
Xue J. J., and Z. Wei. 2004. Unemployment, Poverty and Income Disparity in Urban
China. In The Cost of Economic Transition: An Empirical Analysis on the
Unemployment, Poverty and Income Disparity in China, edited by S. Li and H. Sato.
Beijing: Chinese Finance and Economics Press. Chinese.
Yang, Y. Y., and Z. H. Chi. 2008. Thirty Years of Reform on the Income Distribution of
China. In Report on Economic Development and Institutional Reform in China No. 1,
edited by D. T. Zou. Beijing: Social Science Documentary Press. Chinese.
Yao, H. 2009. Reconstruction of the Medical Care System and Current Tasks in Our
Country. In Social Security System in Contemporary China, edited by X. Hu. Beijing:
Chinese Labor and Social Security Press. Chinese.
Zhang, J. H. 2008. Impact of Social Security on Consumption and Savings of Chinese Urban
Households. Beijing: Chinese Social Sciences Press. Chinese.
Zhang, M. 2009. High Savings in China. Beijing: Chinese Financial Press. Chinese.
Zhang, S. F., and J. Tang, J. 2008. New Stage of the Urban and Rural Minimum Living
Standard Assistance. In 2009 China’s Society: Analysis and Forecast. Beijing: Social
Science Documentary Press. Chinese.
Zhang, X. 2009. Research on the Savings Behavior and Consumption Behavior of the
Chinese Based on the Precautionary Savings Theory. Haixia Science 2: 27–30.
Chinese.
Zhao, D. G. 2009. On the Issues of the Rural Social Old-Age Insurance in Our Country. In
Social Security System in Contemporary China, edited by X. Y. Hu. Beijing: Chinese
Labor and Social Security Press. Chinese.
Zheng, B. W., and S. J. Sun. 2008. Thirty Years of Reform of the Social Security System in
Our Country. In Report on Economic Development and Institutional Reform in China
No. 1, edited by D. T. Zou. Beijing: Social Science Documentary Press. Chinese.
Zheng, G. C. 2002. Change and Assessment of the Social Security System in China.
Beijing: Chinese People’s University Press. Chinese.
————. 2008. Reform and Developmental Strategies of the Chinese Social Security: Ideas,
Goals and Action Scheme. Beijing: People’s Press.
Zheng, Q. M. 2007. Study of the Precautionary Savings of the Urban and Rural Residents
in China. Master’s thesis, Zhejiang Industry and Commerce University. Chinese.
Zhou, S. J., S. Zhang, and H. B. Li. 2009. Income, Consumption and Saving Behavior of
Chinese Urban Households: A Cohort Analysis. Jingjixue (quarterly) 8(4): 1197–
1220. Chinese.
Zhou, X. C. 2009. Reflections on Issues Concerning Saving Rate. Xinhuanet. 23 March.
Available at: http://news.xinhuanet.com/fortune/2009-03/25/content_11067056.
htm. Chinese.
Comments
Yasuyuki Sawada
Given that the PRC is becoming the world’s savings capital, I enjoyed reading
this chapter, which attempts first, to describe savings in the PRC second,
to trace the historical development of social policies from 1949 to 2002,
and finally, to discuss the recent social protection policies. The method
employed was largely descriptive with some aggregate statistics in figures and
tables. No econometric analyses were conducted. There are three concluding
remarks: first, on strategic economic growth policies, the PRC has achieved
unprecedented economic prosperity; second, the social protection system has
been severely jeopardized in the quest for growth; and third, reconstruction
of the social protection system is needed to generate more balanced socio-
economic development.
1. The PRC’s Savings Puzzle

First I would like to ask how the authors can solve the “PRC saving puzzle”
of Modigliani and Cao (2004), i.e., the sudden spurt in the savings rate in the
mid-1970s and the PRC’s unusually high savings rate given its income level
(Kraay 2000). The lack of sufficient social policies can be a plausible explanation;
the savings rate jumped immediately after the reform because precautionary
saving is induced by the lack of unemployment insurance and/or social safety
nets and the old-age insurance motive is driven by the underdeveloped social
security system. Other possible reasons include that high income growth may
be interpreted as a temporary rise in income; a large young population and
longer life expectancy will raise the saving rate; underdeveloped credit markets
lead to liquidity constraints and a high down-payment ratio to purchase a
house; there may be bequest motives within the extended family or dynasty;
and finally, due to cultural factors of Confucianism, people in the PRC may
be risk averse or more patient and thus high savings may be accumulated by
sophisticated hyperbolic discounters. In these respects, the Japanese experience
is useful (Hayashi 1989; Horioka 1990; Iwamoto 1996)
2. Beyond Descriptive Analysis

If the authors wish to go beyond descriptive discussions of the PRC’s savings
and to analyze the impact of social policy on savings, I think looking at the
drastic reform in 1978 would be useful. The reform can be considered as a
natural experiment, i.e., a “serendipitous” situation where individuals are
unintentionally and randomly assigned to a treatment and control group. In
the reform, “before” is the situation with a treatment of social policies and
“after” is the situation of control. In any case, having micro-level data would
be much more desirable.
The authors should be more careful about discussing the reasons behind the
income disparity increase. Deaton and Paxson (1994) have shown clearly that
the life-cycle permanent income hypothesis implies that aging will generate
inequality in consumption and saving automatically. Social protection policy
has failed to serve as a redistributive mechanism for lessening income disparity
especially between urban and rural households. Also, there is a lack of effective
risk-sharing mechanisms across households. Having said this, I think we need
careful micro-level empirical studies to identify the determinants of income
disparities in the PRC. In any case, again, the 1978 natural experiment may
be utilized.
Finally, there is the important issue of handling data problems because
substantial statistical difficulties arise when measuring saving in the PRC. For
example, I do not understand why the authors found negative savings rates
for the rural poorest. The rural poorest should be liquidity constrained and
exhibit more precautionary saving. Park (2006) found that in the poor regions
of the northwest PRC, households hold sizable grain stocks and devote much
land and labor to grain cultivation despite grain’s inferior asset and production
return. Moreover, Lee and Sawada (2007, 2010) and Sawada et al. (forthcoming)
found that the poor in the US, Pakistan, and Japan are liquidity constrained
with higher precautionary motives. Kraay (2000) also indicated that there are
broader categories of saving. These measurement issues should be carefully
considered because measurement errors would generate attenuation bias and
biases arising from non-classical measurement errors.
References
Deaton, A., and C. Paxson. 1994. Intertemporal Choice and Inequality. Journal of Political
Economy 102(3): 437–467.
Hayashi, F. 1989. Japan’s Saving Rate: New Data and Reflections. NBER Working Paper
3205. Cambridge, MA: NBER.
Horioka, C. Y. 1990. Why is Japan’s Household Saving Rate So High? A Literature Survey.
Journal of the Japanese and International Economies 4(1): 49–92.
Iwamoto, Y. 1996. Japan’s Saving Rate Is Indeed Lower than Professor Hayashi Revealed.
Japan and the World Economy 8(1): 35–41.
Kraay, A. 2000. Household Saving in China. World Bank Economic Review 14(3): 545–570.
Lee, J.-J., and Y. Sawada. 2007. The Degree of Precautionary Saving: Re-examination.
Economics Letters 96(2): 196–201.
————. 2010. Precautionary Saving under Liquidity Constraints: Evidence from Rural
Pakistan. Journal of Development Economics 91(1): 77–86.
Modigliani, F., and S. L. Cao. 2004. The Chinese Saving Puzzle and the Life-Cycle
Park, A. 2006. Risk and Household Grain Management in Developing Countries.
Economic Journal 116: 1088–1115.
Sawada, Y., K. Nawata, M. Ii, and J.-J. Lee. Forthcoming. Did the Financial Crisis in Japan
Affect Household Welfare Seriously? Journal of Money, Credit and Banking.
Consumption, Income Distribution, and State Ownership in the PRC 141
VII
Consumption, Income Distribution, and State
Ownership in the People’s Republic of China
Yuqing Xing
1. Introduction
Global imbalances have been frequently cited as a structural problem causing

the current financial crisis. With a gross national savings rate of more than
50% and more than US$2 trillion in foreign exchange reserves, the PRC
unambiguously belongs to the camp with excess savings, and is expected to
reduce its savings in order to help resolve the imbalance. Chinese consumers
are expected to open their wallets, transform their wealth in savings accounts
into consumption, and offset the global demand shortage left by distressed
American consumers. Encouraging greater domestic consumption would
not only contribute to the recovery of the world economy, but would also
mitigate the over dependence of the Chinese economy on external demand
and facilitate its transition to balanced growth.
To achieve this objective, government policy initiatives are necessary, at
least in the short run. However, whether the objective could eventually be
accomplished depends on the capacity and consumption propensity of Chinese
households, not on the government and enterprises of the PRC. Therefore,
how to encourage Chinese consumers to spend more and save less seems to
be a challenging policy issue. A plethora of studies are available attempting
to explain the low consumption–high savings puzzle of Chinese households.
Precautionary motives, due to the rising private burden of health, education,
and housing costs are argued to be one of the major reasons the Chinese
save so much (Chamon and Prasad 2008; Meng 2003). Using the life-cycle
theory, Modigliani and Cao (2004) attributed the PRC’s high savings rate to
significant income growth and demographic changes. The rigid one-child

policy implemented since the late-1970s resulted in substantial demographic
changes. As a consequence, the PRC began the transition to an aging society
at an early stage not compatible with the level of its economic development.
Workers have strong incentives to save more to prepare for retirement. The
pension reform associated with the reform of SOEs further enhanced the
saving propensities of the Chinese (Feng, He, and Sato 2009).
The rising propensity to save is in fact just one factor contributing
to the decline in the share of consumption in the PRC’s GDP. It explains
a very small portion of the decline since the 1990s (Aziz and Cui 2007).
The fundamental factor, household income—which ultimately determines
household consumption—has been given less attention in academic analyses
and debates on the PRC’s low consumption–high savings puzzle. Without
scrutinizing the dynamic changes in household income relative to national
income, it is difficult to infer whether consumption’s low share of GDP was
due to more conservative consumption behavior—i.e., saving more—or simply
a result of the decrease in the share of GDP.
In addition, as a transitional economy, the structure of national savings
in the PRC differs from that in traditional market economies. In particular,
corporate savings, of which SOEs are the main contributors, exceeded 20%
of national income and emerged as the major source of the PRC’s savings
glut. Identifying who are the savers is crucial for making effective policies to
encourage converting savings into consumption.
In this chapter, we examine the PRC’s high savings myth at both macro
and micro levels, and analyze the relationship between low consumption
and income disparities. We also examine the disparity between capital
and labor, and between state and non-state sectors. At the macro level, we
study the composition of PRC national savings, investigate the relationship
between labor’s income and the share of consumption in GDP, and compare
household saving behavior in the PRC with that in Japan. At the micro level,
we review the advance of SOEs and state banks from the verge of bankruptcy
to profitability and monopoly. We argue that profits in the state-owned sector
contributed substantially to rising corporate savings and benefited mainly
managers and employees of SOEs, not their real owners—the general public,
thus widening income disparity between state and non-state sectors. The
empirical results based on regression analysis show that state ownership is one
of the significant factors driving inter-sector income disparity. Our analysis
suggests that reducing corporate savings should be given high priority for
boosting domestic consumption. Collecting dividends from SOEs for either
direct income transfers or for funding social safety nets would not only boost
consumption, but more importantly, enhance social equity.
2. Savings, Consumption, and Income Disparity at the

Macro Level
2.1 The Role of Corporate Savings

In 2008, the national savings rate in the PRC rose to 51.4%, 13.7 percentage
points higher than in 2000. Compared with past records, the gross national
rate surged substantially. For the world as a whole, the savings rate has been
relatively stable and has ranged from 21% to 27% since 1970 (Wiemer 2009).
It is fair to say that the PRC’s gross national savings rate is extraordinarily
high and definitely beyond the international norm.
In a typical economy, national savings is composed of household savings,
corporate savings, and government savings. In spite of 3 decades of economic
reform, the structure of the PRC economy differs substantially from market
economies. The PRC remains a transitional economy. Imperfect competition,
various government regulations and interventions, and the dominance of
SOEs in many profitable industries often favor the government and SOEs in
the allocation of national income. Corporate and government savings play a
much more important role in the PRC economy than in traditional market
economies. Kuijs (2005) argues that most of the difference in national savings
between the PRC and other countries is due to corporate savings.
Understanding who does the saving is crucial when proposing measures
aimed at reducing savings and expanding consumption. In order to identify
major factors driving the increases in the national savings rate, it is imperative
to scrutinize each component of savings. Simply focusing on household savings
would not yield a full picture of savings dynamics, nor would it help find
effective means of reducing the savings rate. The trends and dynamic changes
in the composition of national savings in the PRC from 1992 to 2007 are
shown in Figure 7.1, based on data presented in a speech in 2009 by Zhou
Xiao Chuan of the People’s Bank of China (PBC). During the period shown,
household savings first followed a declining trend—accounting for 20.3% of
the national income in 1992, then gradually decreasing to 17.0% by 2001.
Household savings then started to recover and reached 20.0% of national
Figure 7.1: Distribution of Saving Among Corporations, Households, and

the PRC Government (% of national disposable income)
Source: Zhou (2009).
income in 2007, almost the same level as in 1992. In other words, household
savings changed very little during the period.
In contrast, both government and corporate savings as a share of national
income grew rapidly. In 1992, corporate savings accounted for 11.3% of national
income and government savings for 4.4%. Corporate savings grew almost
steadily after 1997, rising to 22.9% in 2007, more than double the level of 1992.
Government savings fluctuated around 4.0% before 2002, then climbed rapidly
to 8.1% of national income in 2007. The structural change in national savings
from 1992 to 2007 indicates that the significant increase in corporate savings
contributed most to the rise in the gross national savings rate. Household
savings’ share in national income actually dropped 0.3 percentage points from
1992 to 2007, and thus could not make any positive contribution to increases
in the gross national savings rate. In other words, of the three components
of national savings, household savings made the smallest contribution to the
high national savings rate. Analysis of the composition of national savings
based on flow-of-funds data1 in the country’s national accounts also suggests
that corporate savings became more significant than household savings in
the last decade (Tyers and Lu 2009). Hence, in order to promote domestic
1
Financial in-flows and out-flows among all sectors of an economy.
consumption, policy prescriptions simply focused on household savings would
actually miss the real target and would not be effective.
Imperfections in the labor market—where excess supply remains—are among
the reasons PRC enterprises have continuously generated a rising share of
national savings in recent years. An almost unlimited supply of labor from
rural areas undermines the bargaining power of workers and constrains wage
increases. Empirical studies show that wages in the PRC grew much slower
than workers’ productivity. Workers were actually paid about 25% to 30% of
marginal labor revenue (Cai, Wang, and Qu 2009). The PRC economy has
grown about 10% annually in the last three decades. The declining share of
wage income in GDP implies that enterprises have gained relatively more
from economic growth compared to workers.
Another critical factor responsible for the rise in corporate savings was a
series of SOE reforms, and government policy to retain state control in strategic
industries such as mining, petroleum refining, steel, and telecommunication.
State control of these industries serves as a means of maintaining socialist
characteristics within a market economy—the so-called “socialist market
economy.” The enterprise reforms implemented since late 1990 have
significantly reduced the number of SOEs and eliminated millions of redundant
workers, thus converting surviving SOEs into leaner, more competitive, and
profitable firms. Moreover, supported by government policies, SOEs have
monopolized many industries which are very profitable in an economy that
is growing rapidly on the way toward industrialization.
The huge monopolistic rents and retained profits of these enterprises
comprise a major source of corporate savings (Lu et al. 2008; Tyers and Lu
2009). In 2007, state-owned-and-controlled companies earned net profits of
CNY1.62 trillion, equivalent to 6.5% of the PRC’s GDP and 31.8% of the
government’s fiscal revenue. A handful of central SOEs (about 140 large firms),
which are under the jurisdiction of the central government, accounted for
CNY1.1 trillion of net profits, or about 68% of the record profits of SOEs in
2007 (Ministry of Finance 2008).
2.2 Comparing Household Saving Behavior in the PRC and Japan

In addition to the high national savings rate, the rising household savings rate
estimated on the basis of household survey data is also cited as evidence that
Chinese consumers save too much. However, a high savings rate was a common
phenomenon in the so-called East Asian Miracle. Following the success of
Japan and the newly industrialized economies—Hong Kong, China; Korea;

Singapore; Taipei,China—the PRC adopted export-led growth strategies at the
beginning of their economic reforms. The use of savings from current accounts
surpluses to support the investment needed for rapid economic growth was
the means of implementing an export-led growth strategy.
Do Chinese households save more than their peers in other east Asian
countries? To investigate whether Chinese households are really an outlier
in this regard in the international perspective, the household savings path
in the PRC from 1990 to 2007 was compared with that in Japan from 1960
to 1977. A rationale for this comparison is that both economies grew rapidly
in the periods examined. Along with economic growth, household savings
rates in each economy rose sharply. Further, in 1990, the level of economic
development the PRC roughly matched that of Japan in 1960. GDP per capita
of the PRC at the beginning of the 1990s was about the same level as that in
Japan in 1960.
Figure 7.2 shows the trends of household savings rates in the PRC and
Japan. In 1960, Japan’s household savings rate was 14.7%. It grew steadily and
reached a peak of 23.1% in 1975, then decreased slightly to 21.7% by 1977.
The trend of household savings in the PRC was similar. In 1990, households
Figure 7.2: Comparing Household Savings Rates between the PRC and
Japan
Source: Japanese Cabinet Office (2005), China Statistics Yearbook, and author’s calculations.
there saved about 15.1% of their disposable income. Their saving propensity
gradually increased to 24.3% in 2005, almost the same level that Japanese
households achieved within the same period of time (15 years). The savings
rate in the PRC continued to rise to 27.5% by 2007, however. The evolution
of household savings rates in the PRC basically followed the path of Japanese
households during that country’s period of rapid economic growth from 1960
to 1977. Compared with Japan, it is not evident that PRC households are an
outlier and save excessively.
An obvious difference is that Japan’s household savings rate peaked in 1975,
while that of the PRC continued to grow after 2005. However, in PRC household
surveys, consumers’ purchases of houses are classified as investments. The
welfare allocation for housing was abolished by the government in 1998. Faced
with rising housing prices, mortgage payments have comprised a large portion
of monthly expenditures of individual households in urban areas. Monthly
mortgage payments are usually considered savings, not consumption. In
addition, pitfalls in the survey methodology may underestimate household
propensity for consumption. An official of the Chinese Statistics Bureau
acknowledged that survey methodology was designed in the 1980s and failed
to reflect subsequent structural changes in the economy (Peng 2009). Many
consumption expenditures, such as school selection fees and under-the-table
payments for visits to doctors in hospitals, are not covered by the survey.
Rents for houses are based on house construction cost rather than on market
value, thus underestimating household consumption. Given these facts, the
4-percentage-point difference between the peaks of household savings rates
in the PRC and Japan may not support the view that Chinese consumers save
more than their counterparts in Japan.
2.3 Low Consumption and the Shrinking Share of Wage Income

In addition to savings parameters, macroeconomists examine the share of
household consumption in GDP to evaluate whether private consumption
is sufficient. Chinese household consumption as a share of GDP has been
decreasing steadily since 1995. It shrank to 35% of GDP in 2007, lower than
the averages for the world and for high-, middle-, and low-income country
groups (Xing 2009). Declining consumption seems to be convincing evidence
that Chinese households save too much. They should consume more in order
to boost domestic demand and alleviate the global imbalance. This reasoning
is valid only if the household income as a share of GDP remains constant.
The consumption share is the ratio of household consumption to GDP. If

household income as a share of GDP decreases, the consumption share will
decrease correspondingly regardless whether saving propensity increases
or not. In other words, changes in national income distribution between
households and enterprises also affect the share of consumption in GDP.
Specifically, for any given consumption propensity, a decrease in the share of
household income in national income will lower consumption proportionally.
Therefore, in order to explore the low consumption phenomenon and test the
hypothesis that conservative savings behavior resulted in low consumption,
it is essential to analyze whether income distribution has been constant, and
to what extent the decrease in the share of household consumption in GDP
was due to shrinking income.
Due to underdeveloped financial markets and the limited availability of
investment opportunities, wage income remains a major source of household
income in the PRC. The share of wage income in GDP, however, dropped
sharply in recent years. It accounted for 53.0% of GDP in 1995 and exceeded
half of GDP before 2002. It dipped to 49.5% in 2003, the first time in the last
3 decades that wage income accounted for less than half of national income.
Since then, the share of wage income in GDP decreased continuously and in
2007 fell to 39.7%, more than 13 percentage points lower than in 1995.
Without a doubt, the falling share of wage income in GDP contributed
to the decline in household consumption, as wage income accounts for 89%
of household income in urban areas. The trends of household income and
consumption from 1995 to 2007 are compared in Figure 7.3. Each is measured
as a share of GDP. During the period examined, household consumption
dropped about 10 percentage points from 45% to 35% of GDP, while household
income fell 13 percentage points. Rigorous analysis may be needed to accurately
quantify to what extent the decrease in household consumption was due to the
shrinking share of household income. The simple comparison, on the other
hand, unambiguously suggests that falling income, rather than changes in
savings behavior, is a major factor driving down consumption.
The descriptive analysis of the composition of savings and dynamic changes
in the share of household income in GDP suggest that rising corporate savings
contributed most to the increase in the PRC’s national savings rate. Despite the
increase of the household savings rate, at the macro level, household savings
as a share of national income was constant, thus contributing little to the
growth of the gross national savings rate. Measures aimed at reducing high
national savings should focus on the right target: corporate savings.
Figure 7.3: Household Consumption and Labor Shares in GDP
(%)
Source: Lu and Gao (2009) and China Statistics Yearbook.
Economic reforms implemented in the last three decades have mainly

emphasized efficiency and benefited capital much more than labor. They gave
rise to income distribution biased toward enterprises and a decrease in the share
of household income. Biased income distribution and the shrinking share of
household income constrained the consumption capacity of Chinese households
and brought about a continuous decline in consumption as a share of GDP. A
few studies pointed out the negative impact of income disparity on aggregate
consumption in the PRC (e.g., Wan 2007). Low-income families tend to have
a higher consumption propensity than high-income families. Redistributing
income from the latter to the former would enhance aggregate consumption.
The income distribution problem discussed in this chapter, however, differs
from the more conventional discussions of income disparity in these studies.
The distribution of income between enterprises and households—in particular,
between SOEs and their owners, the general public—is discussed here, not the
distribution between high- and low-income households.
According to the deputy director of the research institute of State-owned
Assets Supervision and Administration Commission (SASAC), on average,
a citizen of the PRC nominally owns CNY13,000 in assets of non-financial
SOEs (21st Century Business Herald 2009). Under current policy, ordinary
citizens benefit neither directly nor indirectly from this ownership. Shifting
savings from SOEs to households would not only stimulate consumption,
but more importantly, would enhance social justice; in theory, all SOEs are
owned by citizens of the PRC.
3. Reform of State-Owned Enterprises, State Monopoly,

and Income Distribution
3.1 The Transformation of SOEs: From Losses to Profitability and

Monopoly
Reforming SOEs has been one of the major tasks on the agenda of economic
reform in the PRC since 1978. A variety of measures and approaches have been
tried to restructure inefficient, mismanaged, and unprofitable SOEs. At the
early stage of enterprise reform, state ownership was left untouched. Reform
policies were guided by questions about how to strengthen the competitiveness,
improve the efficiency, and increase the profits of SOEs. Various incentive
mechanisms were introduced, such as bonuses and profit retention systems,
and managers’ autonomy in enterprise operations was expanded. To enhance
the vitality of SOEs, tax reform was launched in the mid-1980s, aiming to
replace the profit remittance system with a corporate tax system. Starting from
1987, enterprise reform concentrated on restructuring the SOE management
mechanism. The primary objective was to strengthen the responsibility of
management. Measures adopted included leasing in small SOEs, a contract
system in large SOEs, and the shareholding system (Lin, Cai, and Li 2003).
These limited reforms did not solve fundamental problems: corporate
governance and ownership that was not clearly defined. Reforming the
ownership structure of SOEs and instituting standardized corporate governance
began in the early 1990s. To facilitate corporatization of SOEs, the Company
Law was passed in 1993. Under this law, SOEs could be restructured into one
of three types of companies: a wholly state-owned company, a limited liability
company, or a company owned by shareholders.2 Corporatization actually
represented the beginning of the privatization process for SOEs.
Radical reforms occurred in 1997 after the government formulated a new
enterprise reform strategy: “grasping the large and letting the small go.”
Grasping the large referred to cultivating large SOEs that were strong and
2
For the definitions, please see Explanatory Notes on Main Statistical Indicators in Chapter 13 of
National Bureau of Statistics (2008).
competitive, and eventually developing them into multinational companies.
Letting the small go implied that the government would give up control of
SMEs. The government employed various methods to spin off SMEs, such as
bankruptcy, merger and acquisition, leasing, and sale. The ultimate purpose
of the strategy was to allow the state to retreat from competitive sectors by
privatizing SMEs, and dominate a few strategic industries. The latter were
industries that were suitable as natural monopolies, or companies with
economies of scales realized through control of a limited number of large,
central and local SOEs (Zheng and Chen 2007).
To achieve the objective of these reforms, the government recapitalized the
targeted large enterprises by converting state loans to equities, authorizing
state banks to write off bad loans, selectively listing firms in stock markets, and
allowing the firms to form joint ventures with foreign investors. In addition,
redundant workers were laid off, and all social welfare functions such as
provision of housing, schools, hospitals, and pensions were removed from
these enterprises.
The downsizing of SOEs was dramatic. It is estimated that about 30 million
workers have been laid off since 1998 (Garnaut et al. 2005). The number of state-
owned and state holding3 industrial enterprises dropped sharply from more
than 64,000 in 1998 to about 21,000 in 2007. The total number of employees
shrank 54% to 17.4 million—corresponding to about 22% of the total labor
force employed in all industrial enterprises, down from 60% in 1998 (Table
7.1). Industrial SOEs as a whole became leaner in terms of number and work
force, but stronger and more competitive. Measured by assets, the average
size of industrial SOEs rose to CNY765 million, more than five times larger
Table 7.1: Selected Indicators of State-Owned and State Holding

Industrial Enterprises
Average Size Owner’s

Employees Assets by Assets Profits Equity
No. of Firms millions % of labor force (billion yuan) (billion yuan) (billion yuan) (billion yuan)
1998 64,737 37.5 60.5 7,491.6 115.7 52.5 2,675.9

2003 34,280 21.6 37.6 9,451.9 275.7 383.6 3,828.1
2007 20,680 17.4 22.1 15,819.0 764.9 1,079.5 6,856.9
Source: China Statistics Yearbook and author’s calculations.
3
The former means wholly state-owned; the latter means that the state controls more than 50% of
enterprise shares.
than in 1998; owners equity almost doubled to CNY6.6 trillion, indicating

that the strategy of grasping the large and letting the small go was successful.
Moreover, with fewer firms and employees, the aggregate profits of SOEs in
2007 grew to about 20 times the level of 1998, exceeding CNY1.0 trillion.
While the assets of SOEs continue to grow, a series of enterprise reforms
loosened government control over enterprise investment and management. The
government realized that it was imperative to have an agency to supervise and
manage state assets. In 2003, the SASAC of the State Council was established.
SASAC was mandated to consolidate functions previously scattered over
various government agencies, and to regulate and supervise central SOEs.
SASACs at the local level were established to supervise SOEs owned by local
governments.
The strategy of grasping the large and letting the small go not only
indicated that small and medium SOEs could be privatized, but also served
as a rationale for retaining state control of strategic industries considered
indispensable to the economy. Unlike other transitional economies, such as
the Russian Federation, complete privatization of SOEs has never been pursued
in the PRC. The president of SASAC, Li Rongrong, stated that the state had
to maintain absolute control of defense, power generation and distribution,
telecommunication, oil and petrochemical, coal, civil aviation, and shipping.
State control of these industries could insure that state ownership remained
a pillar of the PRC economy (Xinhua News 2006).
After years of effort, state monopoly of these industries has been achieved.
Central enterprises produce 70% of both hydro and power electricity generators,
and control 82% of the civil aviation market and 60% of high-value-added
steel products. China Mobile has 457 million customers, about 65% of the
national mobile phone market. China Petrol accounts for 57% of national
crude oil output and 80% of natural gas production (Nanfang Weekly 2009).
With government support and monopoly power, 118 central enterprises under
the jurisdiction of the central government earned CNY696 billion in profits
in 2008, accounting for 80% of the profits of all non-financial SOEs. China
Mobile’s profit topped CNY146 billion, making it the most profitable telecom
company in the world. With profits of CNY135 billion, China Petrol was the
most profitable oil company in Asia. Table 7.2 lists the 10 most profitable
central enterprises in 2008. The profitability of these enterprises was hardly
affected by the global economic recession, unlike the profitability of their peers
in Europe, Japan, and the US. Among the 10 central enterprises, only three
experienced reduced profits. Each of these giant enterprises belongs to a sector
Table 7.2: Profits of Central SOEs in 2007 and 2008 (CNY billion)
Revenues Profits
Enterprises 2007 2008 2007 2008
Total 118 Central Enterprises 10,028.2 11,870.5 1,005.6 696.2
1 China Mobile Communication Corp. 397.2 451.9 127.4 145.8
2 China National Petroleum Corp. 1,000.7 1,273.0 192.0 134.8
3 China National Offshore Oil Corp. 162.0 194.8 56.5 67.8
4 Shenhua Group Corp. Limited 107.1 144.0 29.8 38.3
5 China Unicom 197.7 188.1 16.7 31.0
6 China PetroChemical Corp. 1,227.9 1,462.4 75.7 26.4
7 Baosteel Group 227.7 246.8 35.7 23.8
8 China Ocean Shipping 158.5 190.6 34.1 17.4
9 China Three Gorges Corp. 16.8 21.3 12.7 11.4
10 China Coal Energy Company 57.7 71.3 8.0 11.3
Source: SASAC.
with high entry barriers. It is likely that their strong financial performance
was attributable more to their monopoly position and preferential treatment
by the government than to the efforts and capabilities of their employees.
3.2 Public Funds and Reform of State-Owned Banks

In addition to the reform of enterprises, reform of the banking system began
at the end of the 1990s. As a result, the four large state banks advanced from
the verge of bankruptcy to high profitability. Hit by the financial crisis, US
banks have been struggling with survival. Banks in the PRC, on the other
hand, have weathered the financial crisis much better than their counterparts
in Europe and the US. Once almost-bankrupt, the four largest state-owned
banks—Agriculture Bank of China (ABC), Industrial and Commercial Bank
of China (ICBC), Construction Bank of China (CBC), and Bank of China
(BC)—joined the club of global 500 in 2008 (Fortune 2009)
These four large banks owe much of their success to the injection of
public funds. They were bailed out by the government with tax money
several times after writing off nonperforming loans (NPLs). In 1998, NPLs
of these banks amounted CNY2,170 billion, accounting for 40% of their total
outstanding loans (PBC 2002). The Ministry of Finance subsequently issued
CNY270 billion in bonds to raise funds to increase the capital of the four
banks. Later, CNY1,400 billion in NPLs from the four banks were transferred
to four state-owned asset management companies, which were funded with
CNY40 billion from the Ministry of Finance and CNY192 billion from the
PBC for the purchase of NPLs.
In 2003, PBC provided US$45 billion to help CBC and BC write off their NPLs
and prepare for listing in stock markets. The public fund injections raised the
capital adequacy of both banks and reduced their NPL ratios significantly. ICBC
and ABC also received US$15 billion and CNY130 billion, respectively, from
PBC. With a series of public fund injections and internal restructuring—such
as laying off redundant workers, diversifying ownership, and strengthening
corporate governance—profits of all four of the large banks improved
dramatically. From 2004 to 2007, profits of ICBC grew to CNY82 billion from
CNY30 billion, and the profits of BC jumped to CNY46 billion, more than
double the level of 2004 (Figure 7.4). In 2008, ICBC emerged as the most
profitable bank in the world with profits of US$16 billion (Fortune 2009).
Although a publicly listed bank, the government owns 70% of ICBC shares.
ICBC, however, has yet to pay any dividends to the government.
Using tax money to bail out failed banks is not a unique practice. There
are many predecessor cases. In the late 1990s, the Government of Japan used
Figure 7.4: Profitability of the Four Large State Banks of the PRC,
2004–2007
( Billion Yuan)
ABC = Agriculture Bank of China, BC = Bank of China, CBC = Construction Bank of China, ICBC = Industrial and
Commercial Bank of China
Source: China Financial Statistics.
tax money to save private banks dragged down by NPLs accumulated after
asset bubbles burst in 1989. Recently, the Government of the US injected
more than US$800 billion to rescue the largest US banks hit by the subprime
loan crisis. In Japan and the US, however, injected public money is not free.
All banks receiving public funds in these countries have a legal obligation to
repay the money or face the risk of nationalization if they fail to survive. In the
case of the PRC, however, it is not clear to the general public what obligations
the four large banks have after receiving enormous amounts of tax money.
It appears that these banks have no obligation to repay public funds. The
ambiguity may be due to the assumption that these banks, despite being
listed in stock markets, remain state-owned banks. If so, the citizens of the
PRC, the real owners of these banks, are entitled to share their profits. No
mechanism exists by which the general public, who shouldered all costs of
bailing out these banks, could share in their profits.
3.3 Who Benefits from the Profits of State-Owned Enterprises?

Following tax reform in 1994, SOEs were exempted from paying dividends to
the government and granted the right to retain all their post-tax profits. SOEs
listed on stock markets do pay dividends to their wholly state-owned parent
companies not listed on stock exchanges. The latter are holding companies,
and retain all profits rather than pass them on to the government. Pressured by
public criticism, the State Council began experiments in 2008 with collecting
dividends from SOEs at rates of 10% and 5%, depending on the industry.
Firms in the defense industry or belonging to research institutes, and financial
companies such as banks and insurance companies, were exempted. About
CNY55 billion in dividends were collected in 2008. All dividends collected,
however, were returned to central SOEs for various investment projects. No
dividends were allocated for social welfare programs, which are severely
underfunded.
Who benefits from the record profits of SOEs? Managers and employees
of these firms are the major beneficiaries. One superficial justification for
leaving huge amounts of profits to these SOEs is that they support the
development of these firms. This practice, in fact, may not serve the best
interests of shareholders. Theoretical studies show that, compared with bank
loans and funds raised from issuing corporate bonds, retained profits are
much cheaper (e.g., Gertler 1988; De Meza and Webb 1987). Investment
with retained profits would be less efficient. Firms with excess cash tend to
expand beyond optimal scales and invest in projects with low rates of return.
Moreover, due to asymmetric information and weak supervision, it is highly
likely that managers use retained profits for their own benefit at the cost of
shareholders.
Current arrangements for allocation of SOE profits seriously undermine
social equity. Only a small portion of employees working at SOEs enjoys high
wages and handsome fringe benefits—notably workers in sectors with a state
monopoly. Excluding compensation from stocks and options, the average
income in 2007 of executive managers in state holding companies listed on
stock markets was CNY340,000 (about US$50,000)—more than 10 times
higher than the average salary of workers (Su 2009). The recently uncovered
“shadow stock” plan of China International Capital Corporation Limited
(CICC) revealed how managers and employees of SOEs utilized state assets
and guaranteed monopoly power to maximize their personal interests. CICC
is controlled by China Investment Corporation, the sovereign fund of the PRC,
with 43.4% of its shares. With the support of the government, CICC almost
monopolizes the initial public offering business of SOEs in overseas stock
markets. CICC granted 20% of its shares to its managers and employees in
the form of shadow stocks, substantially diluting the ownership of the state.
In 2007, total employee compensation by CICC amounted US$435 million,
about half of company revenue (Wall Street Journal 2009).
Providing incentives—such as high salaries and stock options—to executive
managers is argued to be an effective means of solving the principal agent
problem in corporate governance and improving share values. In the PRC,
however, SOE executive managers differ from private sector professional
managers in a typical market economy. First, most SOE executive managers are
appointed by the government, rather than selected through open competition
in the labor market. Secondly, they are government officials. Their status and
benefits as government officials remain even after they become SOE managers.
Regardless of the performance of firms under their management, their status
as government officials rarely changes. In other words, they are under no risk
of pay cuts or being dismissed, as are professional managers working in private
firms. Finally, the government has been using various measures to create a
favorable environment for SOE growth, such as easy credit and deliberate
accommodation of the monopoly power of SOEs. It is difficult to determine
whether rising SOE profits are due to the contribution of managers or to
monopoly power and government support (Su 2009). A few studies suggested
that high manager compensation played a significant role in affecting the
value of company shares and correlated with the performance of listed firms
(e.g, Firth, Fung, and Rui 2007; Kato and Long 2005). These studies not only
ignored the reverse causality—i.e., high profits leading high compensation—but
also incorrectly assumed that SOE managers face the same risks as professional
managers in a typical market economy.
A recent survey conducted by the Shanghai Human Resources and Social
Security Bureau (2009) of aggregate labor compensation (including basic
salary and all benefits) indicated that state ownership is highly correlated
with the level of compensation. The bureau surveyed 6,210 companies in
10 sectors in Shanghai. Within each sector, firms were classified according
to ownership structure: state; collective; foreign; private; and Hong Kong,
China; Taipei,China; and Macau, China considered as one group. The
results showed that banking and finance companies provided the highest
compensation—on average CNY221,000 (US$32,500) per employee or about
five times higher than compensation in the hotel and restaurant sector, the
sector with lowest compensation. Within the financial sector, state-owned
financial enterprises paid CNY240,000, the highest compared with non-state
financial companies. Yet, both return on assets and return on equity of state-
owned banks were lower than such returns of foreign banks in the PRC. Using
marginal labor productivity as a rule for determining labor compensation,
the higher compensation offered by state-owned banks was not justifiable
economically.
Higher compensation by state-owned financial firms compared to non-state
financial firms is not a practice unique to the financial sector. In eight out of the
ten sectors covered by the survey, average compensation by state-owned firms
was highest. In the utility industry, annual compensation of SOEs averaged
CNY138,000, about 4.5 times higher than in collectively owned firms, which
offered the lowest worker compensation in the sector. In the financial sector,
compensation by state-owned banks was CNY156,000 higher than the lowest
compensation offered by collectively owned banks (Figure 7.5). It is well known
that both the utility and banking sectors are largely monopolized by SOEs.
It is not difficult to understand differences in compensation by sector.
Industry-specific factors such as relative market competition, specific labor
skills, average firm size, and technology affect worker compensation across
sectors. Within a sector, on the other hand, labor mobility tends to equalize
compensation across firms. The survey, however, suggested that compensation
within the same sector varies substantially. In particular, SOEs paid more
than other types of firms in most sectors. The significant correlations between
Figure 7.5: Average Employee Compensation by Sector and Ownership in

1,000 YuanShanghai, 2008
Source: Shanghai Human Resources and Social Security Bureau (2009)
state ownership and compensation imply that SOEs either enjoy monopolistic
profits, or their employees benefit at the cost of shareholders and the general
public due to weakened supervision of compensation decisions. No similar
survey is available for the whole of the PRC. In that the development of the
market economy in Shanghai is the most advanced compared with other
parts of the country, the survey provides insightful information about
the relationship between state ownership and workers’ compensation in
contemporary PRC.
4. State Ownership and Interindustry Wage Disparity
4.1 State-Owned Enterprises and Wage Disparity

Rising income disparity has emerged as a major social problem in the PRC. The
country’s Gini index rose to 0.46 in 2006 from 0.35 in 1990, a fundamental
factor undermining consumption growth (Chen et al. 2010). Income disparity
has been widening not only between rural and urban areas, but also across
industries. Interindustry disparity, notably the disparity between industries
monopolized by a handful SOEs and industries facing intensive competition,
has recently attracted public attention. In 1998, the average annual salary
in the highest paying sector was 1.6 times that in the lowest. By 2008, the
ratio had jumped to 4.5. A recent study showed that interindustry income
disparity contributed about 10% to income disparity in urban areas (Wan and
Lu 2008). The survey in Shanghai examined in the previous section revealed
indirectly that workers in sectors with relatively strong dominance of SOEs
receive high income.
As a first step to test the hypothesis that state ownership contributed to the
increase in interindustry income disparity, weighted coefficients of variation
were used to measure inter-sector wage disparity. The coefficient of variation
is a measure of dispersion magnitude, indicating how widely the average wage
in each industry fluctuates around the mean wage of all industries. Higher
values of the coefficient indicate larger disparity. Since industry size varies
in terms of number of employees, employment in each industry is used as a
weight for calculating the coefficient of variation. Specifically, the employment-
weighted wage coefficient of variation is defined as
(1)
where li denotes the number of employees in industry i L , is the total

number of employees in all industries, Wi is the average wage of industry
i, W is the average wage of all industries as a whole, and m is the number
of industries. The coefficients from 2001 and 2007 were estimated and the
results illustrated in Figure 7.6. In 2001, the coefficient of variation was
0.18. It gradually increased to 0.24 in 2007, about 33% higher. It is true that
wages usually differ across industries due to many sector-specific factors. The
estimated coefficients of variation indicate not only that wages vary among
sectors, but also that the variation increased over time. To what extent the
increase was due to the concentration of state ownership is the interest of the
empirical analysis that follows.
Efficient wage theory provides an explanation for interindustry wage
differences. To reduce turnover rates and stimulate employee productivity,
some industries pay higher than the market equilibrium wage (Krueger and
Summers 1988). Rent sharing theory ascribes wage differences to institutional
structures of the labor market and wage negotiations between employers and
labor unions. It emphasizes the influence of bargaining power between firm
Figure 7.6: Interindustry Wage Difference
Source: Author’s estimates.
owners and labor organizations on wage determination. If workers are able

to capture some share of the surplus of firms, this results in higher wages
in industries with higher profits. Bell and Freeman (1991) concluded that
rent sharing behaviors mainly accounted for rising US interindustry wage
dispersion since the 1970s.
These theories, however, explain very little of why interindustry wage
disparity rises continuously in the PRC. There is no collective bargaining in
Chinese firms. All labor unions are part of corporate management in the PRC.
No firms face pressure to share their profits with their employees. Moreover,
with about 9.4% unemployment in urban areas and an enormous number
of rural migrants, Chinese workers are in effect paid much lower than their
marginal productivity. Generally, efficiency theory does not apply to the PRC
labor market, except for the markets for some highly specialized workers.
4.2 Empirical Analysis

The descriptive analysis of wage differences according to ownership structure
and the monopoly power of SOEs in section 3 suggests that state ownership
may be a significant factor in determining wage disparity across sectors. To
test this hypothesis, wage determination based on a standard production
function is used as a starting point.
It is assumed that the output of industry i is determined by a production
function with constant return to scale as shown below:
Qi= Q ( Ki ,hi Li ) (1)
where K denotes capital stock, L labor input, and h human capital

embedded in one unit of labor. Thus, unit labor output in the industry can
be written as
(1) qi= Q ( ki ,hi ) (2)
where qi= Qi /Li is the unit labor output and ki= Ki /Li is capital stock per
worker. The unit labor output i is a standard measure for labor productivity,
which determines labor wages. Equation 2 indicates that physical capital and
human capital are two major variables determining labor productivity and
wage. Following equation 2, a testable wage equation can be derived as
ln (wi ) = α + β1 ln (ki ) + β2 ln (hi ) + ei (3)
Equation 3 could be used to estimate wage level in each sector. However, the
hypothesis concerns interindustry wage differences, not wage level. In other
words, the fundamental purpose of the empirical test is to examine whether
state ownership plays a significant role in interindustry wage differences. Using
equation 3, wage difference can be defined as
ln = α + β1 ln + β2 ln + ei (4)
where w – is the average salary of all industries under consideration, k– is the

–
average capital per labor, and h is the average human capital. The dependent
variable actually measures the wage difference between industry i and
the average wage.
How to quantify state ownership in an individual industry is a challenging
task. It is suggested that the share of sales of SOEs in an industry could be a
close proxy to measure the concentration of SOEs in that industry. Data on
SOE sales in individual industries, however, are not available. Instead, the
share of SOE employment in an industry is used to measure the concentration
of state ownership in that industry. The OECD (2005) used the ratio of SOE
employment in the economy to compare the involvement of the state in the

PRC economy with state involvement in industrialized countries. Relative
state control in an industry is defined as , where si represents the share
of employees working in SOEs in industry i, and s– stands for the share of
SOE employees in all industries. Including into equation 4 as an additional
independent variable gives rise to the following regression equation:
ln = α + β1 ln + β2 ln + β3 ln + eit (5)
Equation 5 is the final model estimated for testing the hypothesis that
state ownership contributes to rising interindustry wage disparity. The model
is estimated with panel data covering 12 industries in 2003–2007. A random
effects model is selected for the estimation. The original data were collected
from various issues of China Statistical Yearbook (National Bureau of Statistics
various years). Human capital i is computed as the ratio of professionals to
all employees in an industry; physical capital i is calculated as fixed capital
investment per employee.
The estimates are reported in Table 7.3. The estimated coefficient of
is 0.103 and significant at 1%, indicating that an industry with relatively more
concentration of SOEs tends to pay higher salary. The estimated coefficient
of is 0.041 and significant at 10%, and the estimated coefficient of is
0.333 and significant at 1%. The positive and significant coefficients of both
physical capital and human capital differences are consistent with the theory
that relatively higher physical capital and human capital lead to relatively
higher salary.
Since both physical capital and human capital differences are included in
the model as independent variables, the significance of , the relative SOE
Table 7.3: State Ownership and Interindustry Wage Differences
Constant Adj. R2 Sample Size

0.072 0.041* 0.333*** 0.103***
Estimates 0.255 60
(0.021) (0.022) (0.051) (0.020)
***: significant at 1%; * significant at 10%; the numbers in parentheses are standard errors.
Source: Author’s estimates.
concentration, suggests that state ownership explains partly the residual wage
difference, which cannot be attributed to either physical or human capital. The
empirical result supports the hypothesis that SOEs pay higher salaries than
non-SOEs, in general. It is difficult to find economic rationales to explain why
state ownership could increase salary levels given that both physical capital
and human capital are constant. Institutional reasons—e.g., state monopoly;
weak corporate governance in SOEs, particularly of employee compensation
decisions; and enormous amounts of retained profits—may contribute to the
relative high compensation in SOEs.
5. Concluding Remarks
Income is the primary source of household consumption. Most studies of the

high savings–low consumption myth in the PRC focused mainly on saving
behaviors of Chinese households. This paper, however, argues that it is income
not saving behavior that fundamentally constrains Chinese household
consumption. Comparison of Chinese and Japanese households showed that
the dynamic saving pattern of Chinese households is not so peculiar. Wage
earnings remain a major source of household income in the PRC. Its share of
national income declined steadily to 39% of GDP in the last decade, resulting
in a reduced share of consumption in national income. Hence, measures aimed
at improving the share of consumption in national income should start from
raising the share of wage earnings in national income—a fundamentally more
logical policy choice than reducing household savings rates.
Measured by the gross national savings rate, the PRC is unambiguously a
country with exceptionally high savings. In order to reduce national savings
with effective policies, it is imperative to understand who owns the savings.
Decomposition of the structure of national savings reveals that corporate
savings accounted for 23% of national income, and the growth of corporate
savings contributed most to the significant increase in the national savings
rate. How to reduce corporate savings rather than household savings should
be the policy target for boosting aggregate consumption and lowering the
gross national savings rate.
How to balance efficiency and equity has been a delicate and challenging
issue for policy makers. In the last three decades, economic reform in the PRC
has concentrated solely on efficiency and ignored social equity—a complete
reversal of the utopian equity pursued under the centrally planned economy.
Reform has greatly enhanced the bargaining power of capital and nurtured
income distribution biased toward capital, which is in part responsible for

surging corporate savings.
SOE monopolies in profitable and strategic industries, and dividend policy
with regard to SOE profits, are both determined by government development
strategies and have played an important role in widening the income distribution
between labor and capital. While the number of employees in industrial SOEs
dropped sharply to less than 22% of all industrial employees, retained profits
by SOEs benefit only managers and employees in these firms. This is despite
the fact that all PRC citizens are the real owners of SOEs. Hence, the current
arrangement has seriously undermined social justice and further amplified
income disparity.
To stimulate domestic consumption, the bias of income distribution
toward enterprises should be tackled. In particular, high priority should be
given policy that tackles how the general public could benefit from the record
profits of both financial and non-financial SOEs. Using part of SOE profits
to fund social welfare services such as education, health care, and pensions,
or even direct income transfers to lower-income households, would create
tangible benefits to the general public and demonstrate their ownership of
SOEs. Such measures would not only facilitate domestic consumption, but
more importantly, alleviate social inequality. Efficiency has dominated the
PRC’s development agenda for the last 30 years. Putting social equity ahead
of efficiency can be justified now.
References
21st Century Business Herald. 2009. Citizens May Benefits the Profits of Central SOEs from
Three Channels. 27 October.
Aziz, J., and L. Cui. 2007. Explaining China’s Low Consumption: The Neglected Role of
Household Saving. IMF Working Paper, WP/07/181. Washington, DC: IMF.
Bell, L., and R. Freeman. 1991. The Causes of Increasing Inter-industry Wage Disparity in
the United States. Industrial and Labor Relations Review 44(2): 275–287.
Cai, F., M. Wang, and Y. Qu. 2009. Industrial and Labor Relocations among Chinese
Regions. Paper presented at the ADBI conference, Labor Market in the People’s
Republic of China and its Adjustment to Global Financial Crisis. Tokyo. 18–19
June.
Chamon, M., and E. Prasad. 2008. Why are Saving Rates of Urban Households in China
Rising? The Institute for the Study of Labor (IZA) Working Paper No. 3191. Bonn:
IZA.
Chen, J., D. Dai, M. Pu, W. Hou, Q. Feng. 2010. The Trend of Gini Coefficient of China.
Brooks World Poverty Institute Working Paper 109. Manchester: University of
Manchester.
De Meza, D., and D. C. Webb. 1987. Too Much Investment, a Problem of Asymmetric
Information. Quarterly Journal of Economics (May): 281–292.
Feng, J., L. He, and H. Sato. 2009. Public Pension and Household Saving: Evidence from
China. Bank of Finland’s Institute for Economies in Transition (BOFIT) Discussion
Paper No. 2. Available at: http://www.bof.fi/bofit_en/tutkimus/tutkimusjulkaisut/
dp/2009/dp0209.htm
Firth, M., P. Fung, and O. Rui. 2007. How Ownership and Corporate Governance
Influence Chief Executive Pay in China’s Listed Firms. Journal of Business Research
60(7): 776–785.
Fortune. 2009. Fortune Global 500. Available at: http://money.cnn.com/magazines/
fortune/global500/2009/industries/192/index.html.
Garnaut, R., L. Song, S. Tenev, and Y. Yao. 2005. China’s Ownership Transformation:
Process, Outcomes, Prospects. Washington, DC: World Bank.
Gertler, M. 1988. Financial Structure and Aggregate Economic Activity: An Overview.
Journal of Money, Credit and Banking 20 (August): 559–588.
Japanese Cabinet Office. 2005. Annual Report on the Japanese Economy and Public Finance
2003-04. Tokyo: Government of Japan.
Kato, T., and C. Long. 2005. Executive Compensation, Firm Performance, and Corporate
Governance in China: Evidence from Firms Listed in the Shanghai and Shenzhen
Stock Exchanges. IZA Discussion Paper No. 1767. Bonn: IZA.
Krueger, A., and L. Summers. 1988. Efficiency Wages and the Inter-Industry Wage
Structure. Econometrica 56(2): 259–293.
Kuijs, L. 2005. Investment and Saving in China. World Bank Policy Research Working
Paper 3633. Washington, DC: World Bank.
Lin, J., F. Cai, and Z. Li. 2003. The China Miracle: Development Strategy and Economic
Reform. Hong Kong, China: Chinese University Press.
Lu, F., G. Song, J. Tang, H. Zhao, and L. Liu. 2008. Profitability of Chinese Firms, 1978–
2006. China Economic Journal 1(1): 1–31.
Lu, M., and H. Gao. 2009. When Globalization Meets Urbanization: Labor Market
Reform, Income Inequality and Economic Growth in China. Paper presented
at the ADBI conference, Labor Market in the People’s Republic of China and its
Adjustment to Global Financial Crisis. Tokyo. 18–19 June.
Meng, X. 2003. Unemployment, Consumption Smoothing, and Precautionary Saving in
Urban China. Journal of Comparative Economics 31(3): 465–485.
Ministry of Finance, Government of the People’s Republic of China. 2008. State-Owned
Enterprises with a Record Profit of 1.62 Trillion Yuan. Beijing: Government of the
PRC.
Modigliani, F., and S. Cao. 2004. The Chinese Saving Puzzle and the Life-Cycle
Nanfang Weekly. 2009. The Big Changes of Central SOEs after Ten Years: the Origin and
the Future. 20 August. Chinese.
National Bureau of Statistics, Government of the PRC. Various years. China Statistical
Yearbook. Beijing: China Statistics Press.
OECD. 2005. Economic Survey of China 2005. Paris: OECD.
Peng, Z. 2009. Some Opinions on China’s Consumption Statistics. Available at: http://
www.stats.gov.cn/tjyj/tjggyj/t20091201_402604787.htm.
PBC. 2002. The Development Report of China’s Financial Market 2002. Beijing: PBC.
Shanghai Human Resources and Social Security Bureau, Shanghai Municipal People’s
Government. 2009. Survey on Employee Compensation by Sectors. Shanghai, PRC:
Shanghai Municipal People’s Government. Chinese.
Su, H. 2009. Some Thoughts About How To Regulate Executives Pay in State-Owned
Enterprises. Institute of Labor and Wages, Chinese Ministry of Personnel. Beijing:
Government of the PRC.
Tyers, R., and F. Lu. 2009. Competition Policy, Corporate Saving and China’s Current
Account Surplus. Working paper in Economics and Econometrics No. 496.
Canberra: The Australian National University.
Wall Street Journal. 2009. The Secret of CICC Shadow Stocks. 20 November. Available at:
http://chinese.wsj.com/gb/20091120/ffe091249.asp.
Wan, G. 2007. Understanding Regional Poverty and Inequality Trends in China:
Methodological Issues and Empirical. Findings, Review of Income and Wealth. 53(1):
25–34.
Wan, G., and M. Lu. 2008. Inter-industry Disparity: An Important Factor Causing Income
Disparity in Urban Areas. Department of Economics Working Paper. Shanghai:,
PRC Fudan University.
Wiemer, C. 2009. Understanding China’s High Saving. East Asian Institute Background
Briefing No. 476. Singapore: National University of Singapore.
Xing, Y. 2009. The Global Economic Recession and Challenges to China’s Export-led
Growth Strategy. Paper presented at the Conference on Global Economic Crisis:
Impact and Implications for Industrial Restructuring in Asia. Taipei,China. 19–20
August.
Xinhua News. 2006. SASAC: The State-Owned Economy Should Maintain Absolute
Control in Seven Sectors. 19 December.
Zheng, Y., and M. Chen. 2007. China’s Recent State-Owned Enterprise Reform and Its
Social Consequences. China Policy Institute, Briefing Series Issue 23. Nottingham,
UK: University of Nottingham.
Zhou, X. C. 2009. Issues Related to the Global Financial Crisis. Speech delivered at the
Global Think-Tank Summit. Beijing. July.
Comments
Yvonne Sin
The chapter shows how an in-depth analysis of the relationship between income,
savings, and consumption can contribute to our understanding of what drives
domestic demand. This is a challenging task, as key indicators such as savings
rate and income disparity can take on different operational definitions and
be subject to varying interpretations. The implications are even greater in
the case of the PRC because of the country’s size and diversity in terms of
geography, climate, and demographics, as well as its rapid and far-reaching
socio-political-economic transformation in recent history. A commendable
effort has been made to close the loop on consumption, income distribution,
and state ownership. However, this may have led to oversimplification of
mechanisms of influence as well as the legacy of SOEs in the context of the
PRC’s reform and opening up policy. Furthermore, while the choice of a rather
narrow scope allows closer examination of one particular dimension, it also
runs the risk of impoverishing the discussion on a very complex topic.
1. Analysis of State-Owned Enterprises

In many respects, SOEs represent a unique phenomenon in the PRC’s transition
from a planned economy to a socialist market economy. They carry the legacy
of past policy decisions, while at the same time they are required to adopt
new measures associated with reform and opening up policies. Therefore,
expecting SOEs to operate like any other modern firm in a mature market
economy system and evaluating them according to efficient wage theory, rent
sharing theory, or some such theoretical model is not quite appropriate. The
case of NPLs is one example. The central government bailed out a number
of state-owned banks, rescuing them from large portfolios of NPLs. Though
the problem with NPLs is now resolved, the government-injected funds were
never repaid. Even though these banks are now showing operating surpluses/
profits (roughly the size of the injections of government funds), these cannot
be deemed true surpluses that can be arbitrarily redistributed in the form of
wages. On the contrary, there would be justification for allocating portions
of these surpluses proportionately—as repayment to the government or as
retained earnings to the SASAC. On another front, as SOEs are gradually
being transformed into global enterprises with many currently listed in stock
exchanges outside of the PRC, corporate compensation policies must stay
competitive.
Since they make up the country’s largest industry sector, the behavior of
SOEs cannot help but have an impact on national savings and consumption
patterns. But those effects will stem from the interplay of national policy
decisions—including fiscal policy, stimulus packages, pension reform, and
policies to stimulate investment or real estate—and not from independent
board decisions, as might be the case with typical private sector corporations.
As such, their ultimate effects on savings and spending trends may not be that
predictable and will vary depending on the prevailing policy environment.
For instance, in a roundabout way, SOEs may have brought about the opposite
effect of a recent stimulus package to boost spending. By getting involved in
property development and speculation, SOEs have been contributing to the
inflation of the property market bubble, leading to further increases in house
prices that are already rising, and indirectly forcing households to save more
in order to be able to afford their own homes at some future time (China
Daily 2009).
2. Profit Redistribution by State-Owned Enterprises:

The Solution?
Clearly, the fact that workers are finding it increasingly difficult to purchase
homes is not simply a problem of wage growth. In fact, absolute wage growth
in the PRC has been above 10% annually on average in the last decade—not
a low growth rate by the standards of many developing countries. But in a
heated real estate market, the median price of a 100-square-meter apartment
in Beijing or Shanghai, for example, is about US$100,000. If we assume the
average annual income per capita to be about US$2,000 in Beijing or Shanghai,
the average Chinese worker would need to save his or her entire annual income
(representing a 100% savings rate) for about 50 years in order to buy such an
apartment. Financing the purchase with a mortgage would still require a 30%
down payment or a one-time payment of US$30,000—the equivalent of about
15 years of income (without counting interest payments). Merely raising wages
or handing out more dividends—assuming the funds are legitimately classified
as true surplus earnings for distribution—will likely not solve the problem.
Too many unknown and uncontrolled variables would remain.
Many experts believe that fundamental problems in the financial system
must be addressed first before distortions in the savings rate, consumption
pattern, income distribution structure, and other areas (including the labor
market, wage growth, enterprise profitability, employment rate, and employee
compensation) can be effectively adjusted. Judging by media releases and
reported speeches, the government is aware of the complexity and implications
of these issues and is working to mitigate the savings imbalance, reduce income
inequalities, and boost domestic consumption. Many economists would also
agree that only a strong domestic financial sector can support the necessary
policies to truly sustain a dynamic domestic economy.
References
China Daily. 2009. Blowing up a property bubble. 4 November.
Public Expenditures on Social Programs and Household Consumption in the PRC 171
VIII
Public Expenditures on Social Programs and
Household Consumption in the
People’s Republic of China
Emanuele Baldacci, Giovanni Callegari, David Coady, Ding Ding,
Manmohan Kumar, Pietro Tommasino, and Jaejoon Woo
1. Introduction
Household consumption in the PRC is low. The ratio of household consumption

to GDP is only 37%, compared to close to or above 50% of GDP in industrial
and emerging market countries. This in part reflects the PRC growth model,
and a high level of precautionary savings. The latter may be due largely to
inadequate social protection programs relating to health and old age, and
the elevated private cost of higher education. Additional factors relate to
demographic trends and inadequate access to credit for a significant share
of the population. The low propensity to consume and the associated high
savings rate have received significant attention in domestic and international
policy circles, and are viewed as key elements in the PRC’s large current account
surpluses and associated global imbalances.
This chapter examines the likely impact of expanding social programs on
household consumption in the PRC. It identifies a variety of channels through
which higher government social spending can impact household consumption,
and quantifies the likely effects. Specifically, it explores the following three
channels: (i) household age-specific propensities to consume out of (lifetime)
disposable income, (ii) the distribution of household disposable income across
different income groups (with different propensities to consume), and (iii) the
overall level of household disposable income. Using household income survey
The authors have benefitted greatly from comments received at the 2009 ADBI’s 2009 Annual Conference,
at a 2010 IMF Workshop in Beijing, and from Benedict Clements, Carlo Cottarelli, Sanjeev Gupta, and
other IMF staff members.
data, this chapter applies a generational accounting framework to estimate

age-specific marginal propensities to consume for different income groups,
and the lifetime amount of resources available to each cohort. The model is
used to simulate the effects on aggregate consumption of alternative reforms
of government social expenditure.
The remainder of the paper is organized in the following manner. Section 2
provides an overview of trends in consumption and savings in the PRC and
discusses the underlying determinants. Section 3 presents new evidence of
the impact of public expenditures on health, education, and pensions on
savings rates for a panel of industrial and emerging market countries. Section 4
describes the methodology and data used in estimating the impact of social
program reforms in the PRC. Simulation analysis exploring the impact of a
range of social expenditure reforms is presented in Section 5. The final section
provides a summary of the results and discusses the policy implications.
2. Consumption Rates in the PRC

Household consumption as a share of GDP in the PRC is low and has been
falling. The ratio of household consumption to GDP was 37% in 2008, having
fallen almost continuously from about 55% in 1981 (Figure 8.1). The current
ratio ranks the PRC at the bottom in the Asia region as well as among emerging
markets. With regard to its evolution over time, in Asia, only India witnessed
a similar decline in the consumption ratio, but started from a much higher
level (Figure 8.2). The level of household consumption in the PRC is low also
in comparative historical terms. For instance, the PRC’s consumption ratio
is lower than that in the US in the 1850s, when GDP per capita in the US was
comparable with that of the PRC today (McKinsey Global Institute 2009).
The decline in the PRC household consumption ratio can be divided into
two components, reflecting changes in the savings rate and in the share of
household income in GDP:
t 5IFIPVTFIPMETBWJOHTSBUFJOUIF13$JTIJHIBOEIBTCFFOHSBEVBMMZ
increasing. The average savings rate of households out of disposable
income rose from 11% in 1990 to 25% in 2007, 12 percentage points higher
than the Asia average. This contrasts with the decline observed in other
Asian countries, with the exception of India (McKinsey Global Institute
2009). The PRC’s household savings rate is also high when compared
with industrial countries. In 2007, the average household savings rate of
Figure 8.1: Private Consumption Expenditure in the PRC (% of GDP)
(%)

Source: International Monetary Fund, World Economic Outlook Database. Available at:
www.imf.org/external/pubs/ft/weo/2010/update/02/index.htm
Figure 8.2: Private Consumption in Seletected Asian Countries (% of GDP)

(%)

Source: International Monetary Fund, World Economic Outlook Database. Available at:
www.imf.org/external/pubs/ft/weo/2010/update/02/index.htm
European Union countries was about 11%, while in the US it was below
2% (Eurostat 2009).
t )PVTFIPMEEJTQPTBCMFJODPNFBTBTIBSFPG(%1JOUIF13$JTMPX BU
about 54%, and declined by about 8 percentage points between 1990
and 2007.
Overall, the increase in the household savings rate accounts for about
9 percentage points of the approximately 13 percentage points of GDP decline
in the household consumption ratio between 1990 and 2007. The rest can
be explained by the fall in the share of household disposable income in GDP
over the same period.
The decline in household income as a share of GDP can be explained by
weak wage growth and the limited redistribution of firms’ profits. Weak wage
growth may be attributed to a variety of factors. The high level of internal
migration from rural to urban areas has maintained a large supply of labor
which outstrips increases in demand, resulting in a sizeable share of labor
that is underemployed or unemployed. This is combined with the absence
of effective union organizations and some degree of monoposonistic power
in the hands of employers. Investment income has languished because of the
virtual absence of profit redistribution to the public. This is due to the limited
number of publicly listed firms and the tendency of SOEs to not pay dividends
to the government, which could be redistributed to the public.
The increase in the household savings rate reflects a number of factors.
According to the life-cycle hypothesis, consumption and saving depend mainly
on the lifetime resources (net present value of income plus net current wealth)
and the demographic structure of the population. These two factors impact
the savings rate differently. Expected future growth of income reduces current
savings. Conversely, an increase in the ratio of the working to the nonworking
population can increase the savings rate, given the higher average income
resulting from a greater proportion of population being employed. In contrast
to the life-cycle hypothesis, the precautionary savings explanation postulates
that higher income and expenditure risks may increase savings as households
save to deal with adverse shocks. This would be consistent with a savings rate
increasing with age.
Each of the above factors appears to have characterized the PRC economy
at various times during the last two decades. Rapid economic growth has taken
place throughout the entire period together with an increase in the share of
the working population. At the same time, reform of SOEs at the beginning
of the 1990s substantially reduced the coverage of the PRC’s welfare services.
Prior to the reform, SOEs were responsible for the social and economic welfare
of workers and their families. After their reform, the burden of health and
education expenditures essentially shifted to the private sector, effectively
reducing households’ lifetime incomes (as income in-kind was lowered by
the reforms), and leading to a perception of higher income and expenditure
risk. The increased risk faced by households of incurring significant health
or education expenditures is thus likely to have played a role in the rise in
the savings rate.
Recent analysis confirms that savings for health and education costs play
an important role in explaining the increase in the household savings rate.
Earlier studies of the relationship between saving and lifetime income lent
some support to the role of demographic factors in explaining the dynamics
of the savings rate (Kraay 2000; Modigliani and Cao 2004; Horioka and Wan
2006) (Box 8.1). These conclusions, however, have been questioned by more
recent estimates on the age profile of the savings rate by Chamon and Prasad
(2008), which showed a U-shaped age profile of savings in which younger
and older households saved relatively more. The same study showed that
households featuring high expenditure risk regarding health (typically older
households) tended to have a savings rate 20 percentage points higher than
households not facing these risks. Similarly, households with small children
tended to have a savings rate up to 5 percentage points higher than households
without children—the additional savings to finance future education spending.
These effects may have been amplified by financial underdevelopment, as
reflected in constraints on borrowing against future income and low returns
on financial assets.1
Other country-specific evidence also supports the premise that extending
social services is likely to increase the household consumption rate. A number
of studies have been undertaken on the impact of extending social services
(including health, education, and pension insurance) on consumption in the
US. These studies suggested a positive relationship between the extension of
social services and household consumption rates. For instance, Gruber and
Yelowitz (1999) found that among the population eligible for Medicaid in
1993, each US$1,000 of added coverage increased household consumption by
US$538. Chou, Liu, and Hammitt (2006) examined the extension of health
1
In line with the age profile estimated by Chamon and Prasad (2008), Wei and Zhang (2009) found that
half of the increase in household savings rates could be explained by the rising share of males in the
population, experienced after the adoption of the one-child policy.
Box 8.1: Determinants of Household Savings Rate:

Survey of Evidence
Data Source or
Authors Country Sample Empirical Findings
PRC: Determinants of Household Savings Rate
Kraay Rural and urban 1978– Expectations of future income growth and the role of
(2000) household survey of 1995 subsistence consumption played an important role
National Bureau of in determining the level of rural household savings.
Statistics of the PRC No significant relationship explaining levels of urban
household savings.
Modigliani Aggregate data from 1953– Savings rate increased with the share of working-
and Cao National Bureau of 2000 age population, in line with the life-cycle hypothesis.
(2004) Statistics’ China
Statistical Yearbook
Horioka Provincial level data 1996– Demographic factors affected savings in line with
and Wan from National Bureau 2005 the life-cycle hypothesis. Savings rate across time
(2006) of Statistics’ China and provinces determined mainly by the lagged
Statistical Yearbook savings rate, income growth, interest rate and, in
some cases, inflation.
Chamon National Bureau of 1990– Virtual absence of consumption smoothing over
and Prasad Statistics’ Urban 2005 time. Savings rates of younger and older households
(2008) household surveys grew relatively more. The factor that best explained
these patterns was the rising private expenditure on
health, education, and housing.
Wei and PRC population 2002 Half of the increase in the household savings rate
Zhang census, county Social- could be explained by the increasing share of males
(2009) economic Statistical in the population, experienced after the adoption of
Yearbook 2000, and the one-child policy. The premise was that males
China Household saved to accumulate assets that would put them at a
Income Project competitive advantage when searching for a spouse.
Barnett PRC provincial data 1994– Spending on health, but not education, had an
and Brooks from CEIC Data Co. 2007 impact on household behavior. A 1-yuan increase in
(2010) government health spending was associated with a
2-yuan increase in urban household consumption.
Other Countries: Impact of Extending Social Safety Nets on Savings
Kotlikoff US 1950– Savings rate was negatively correlated with the
(1989) 1987 availability of public health insurance.
Kantor and US 1917– The introduction of workers’ compensation following
Fishback 1919 injuries at work (gradually introduced from 1917
(1996) to 1919) reduced private savings by approximately
25% of their baseline value.
Gruber and US 1984– Among the population eligible for Medicaid in 1993,
Yelowitz 1993 each US$1,000 of added coverage would increase
(1999) household consumption by US$538.
Chou, Taipei,China 1992– The extension of health insurance coverage
Liu, and 1997 decreased the household savings rate by 3%–10%;
Hammitt this meant that an increase in health expenditures
(2006) of 1 percentage point of GDP increased current
household consumption by 0.4%–0.6% of GDP.
insurance coverage in Taipei,China from 57% of the population in 1994 to
96% in 2000. They concluded that the reform decreased savings by 3%–10%,
and that a US$1 increase in medical care transfer payments reduced savings
by US$0.4–US$0.6. This result is interpreted as indicating specifically the
impact of health insurance on savings, as changes in household income or in
the age of the households were explicitly controlled for in the analysis.
3. Impact of Public Social Expenditures on Household

Savings: Panel Analysis
This section presents new evidence on the relationship between household
savings and government social expenditures. The empirical analysis is based
on a panel of 24 OECD countries over 1990–2008.2 As noted previously, the
empirical evidence from household budget surveys suggests that government
spending on health care, education, and pensions may be associated with lower
household savings. However, there are few studies that assess the significance
and size of these effects for a panel of countries (the main exception appears to
be Barrell, Hurst, and Kirby [2009]). This section fills this gap in the literature.
The results can be used to obtain illustrative estimates of the potential impact
on household consumption of increases in social spending in the PRC. More
specific results for the PRC based on simulations using household budget
survey data are presented in the next two sections.
The analysis that follows draws on well established theoretical and empirical
literature on the determinants of savings. Real per capita income and its growth
rate, demographic factors, and the level of financial system development are
the main determinants of cross-country variations in household savings over
time (e.g., Loayza, Schmidt-Hebbel, and Servén 2000; IMF 2005; Bosworth
and Chodorow-Reich 2007; Park and Shin 2009).3 Building on this literature,
a standard set of explanatory variables is considered, including economic and
demographic factors, and financial system variables. It is in this framework
that the impact of government spending on health care, education (primary,
2
The focus is on the long-run relationship between saving and its potential determinants. Data are
5-year averages; the availability of public expenditure data dictated the length of the time period.
Household savings data obtained from OECD (2009).
3
Most cross-country studies examine national or private savings, rather than household savings. As a
test of robustness, we also estimated models using data on gross domestic savings (World Bank 2009)
and private savings (Bosworth and Chodorow-Reich 2007). The results were qualitatively similar to
those presented in this paper.
secondary, and higher), and social security (pensions and social assistance)
on the savings rate is explored.
3.1 Regression Specification

The baseline regression specification for household savings as a percent
of disposable income is:4
Savingit = α + Xit’β + γZit + νi + εit ,
where i and t denote the country and time period; νi is the country-
specific fixed effect; εit is an error term; Xit is a vector of economic, financial,
and demographic variables; and Zit is the social spending variable (in percent
of GDP).
Xit includes the following:
Real per capita GDP and its growth. Richer countries tend to save more
than poorer countries and fast-growing countries tend to save more than
slow-growing countries.
Demographic structure. This is captured by two dependency ratios: old-
age dependency (ratio of the population aged 65 years and above to that aged
15–64 years), and young-age dependency rates (ratio of the population under
15 years to that aged 15–64 years). Existing empirical evidence suggests that
a high old-age dependency ratio is associated with lower aggregate saving,
as the number of dissavers is greater than the number of savers based on the
life-cycle theory. On the other hand, high dependency rates can have adverse
implications for public savings under a pay-as-you-go system.
Financial development. This is measured by credit extended to the private
sector from banks and other financial institutions (as a percent of GDP). An
underdeveloped financial system can lead to higher savings (Prasad 2009).5
In a growing economy where the desired consumption bundle shifts toward
4
The two main estimation methods reported are fixed-effects panel and dynamic panel generalized
method of moments regression. The results from different estimation methods, including pooled
ordinary least squares (OLS) (not reported), are broadly similar. Focusing on different time periods
or including the time trend or time fixed effects does not alter the results.
5
Caballero, Farhi, and Gourinchas (2008) identified financial underdevelopment as an important
determinant of rising savings rates, and hence a driver of global imbalances. However, Edwards
(1996) argued that financial deepening induces higher savings rates by creating more sophisticated
financial systems.
durable goods, the inability to borrow against future income streams could
lead households to save more in order to self-finance their purchases.
3.2 Empirical Results

Higher government social spending is generally found to be associated with
lower household savings. The estimated effect is nonlinear, implying that the
marginal reduction in savings in response to an increase in social spending
is largest when the level of social spending (as a percent of GDP) is low (Table
8.1). As spending levels rise, the marginal decline in savings in response to
increased public spending becomes smaller. The coefficients for social spending
are statistically significant and economically meaningful. Not surprisingly, the
magnitude of the marginal impact of social spending on household savings
differs across spending items and depends on the initial spending level (reflecting
the nonlinear relationship). Public spending on health care has the largest
negative impact. The effect of spending on social security is somewhat smaller,
but significant. Spending on education also seems to have a sizable negative
impact, but is only significant in regressions when individual social spending
components are considered separately (Appendix Table A8.1).
The coefficients on total social spending and its squared term suggest a
nonlinear U-shaped relationship: the minimum level of household savings
(with respect to total spending) is reached when total social spending is around
31.0%–36.0% of GDP (Table 8.1, columns 1, 2, and 4). The sample mean of
social spending is 28.1% of GDP. As an illustrative example, consider a country
where social spending is at this sample average. Results show that household
savings will fall by 0.26%–0.45% of household disposable income in response
to an increase in total spending by 1.00% of GDP.6
As the average household disposable income is about 54% of GDP, columns
1, 2, and 4 of Table 8.1 suggest that a 1.00% of GDP increase in total social
spending is likely to reduce household saving by 0.14%–0.24% of GDP.
Alternatively, the impact on savings of a simultaneous increase in spending
on each of the three components individually can be computed. This is based
on column 3 of Table 8.1, and yields a similar impact of around 0.13% (of
GDP) on household savings for a 1.00% of GDP increase in total spending
(assuming ⅓ of outlay on each component).
6
Table 8.1, column 1: ΔSaving/ΔZ = 0.06Z –1.95 because Saving = 0.03Z2 – 1.95Z + other terms, where
Z is total social spending.
Table 8.1: Household Saving Panel Regression for 24 OECD Countries,

1990–2008
(1) (2) (3) (4)

Explanatory Variables FE FE FE System GMM
-0.41*
Household saving (lagged) (-1.73)
-0.40** -0.35 -0.27 -0.06
Growth, per capita GDP (-2.19) (-1.66) (-1.13) (-0.18)
48.19 -84.88 190.9** -67.29
Initial GDP per capita (log) (0.73) (-0.97) (2.42) (-0.39)
-2.75 3.93 -9.44** 3.5
Initial GDP per capita (log), squared (-0.85) (0.91) (-2.42) (0.40)
-1.5 -0.88 -2.19 -4.13**
Private credits (-1.62) (-0.83) (-1.63) (-2.56)
-0.75*** -0.71*** -0.46 -1.29**
Old-age dependency ratio (2.92) (-3.14) (-1.31) (-2.48)
-0.27 -0.21 0.15 0.26
Young-age dependency ratio (-1.18) (-0.97) (0.49) (0.51)
-1.95*** -2.14*** -3.66***
Public social spending, totala (-4.17) (-4.43) (-2.92)
0.03*** 0.03*** 0.06***
Public social spending total, squared (4.14) (4.12) (2.9)
-0.33**
Government saving (-2.52)
-6.84***
Public health spending (-3.45)
0.44***
Public health spending, squared (4.09)
2.33
Public education spending (0.67)
-0.10
Public education spending, squared (-0.52)
-2.15*
Social protection spending (-1.81)
0.05*
Social protection spending, squared (1.84)
Arellano-Bond test for AR(2), p-value 0.47
Hansen Test of Joint Validity of instruments 0.95
No. of instruments 14
No. of observations 78 68 78 74
No. of countries 24 20 24 24
R2 0.76 0.8 0.8
FE = fixed effects estimation, GDP = gross domestic product, GMM = generalized method of moments.
a
Public social spending total is the sum of the public expenditures on health, education, and social protection.
Notes:
1. Dependent variable: household saving (% of household disposable income).
2. The panel consists of four 5-year periods for 24 countries. Heteroskedasticity and country-specific
autocorrelation consistent t-statistics are reported in parentheses. Levels of significance are indicated by
asterisks: *** = 1%, ** = 5%, and * = 10%. An intercept term is included in each regression.
Application of these results to the PRC can provide some illustrative estimates.
Given the PRC’s current level of social spending (around 6.00% of GDP), the
marginal reduction in household savings for a 1.00% of GDP increase in social
spending could be in the range of 0.56%–1.03% of GDP.7
Depending on the composition of social spending, however, increases of
such spending will have different impacts. In the sample, for a country with
public health expenditure at the sample mean of 6.30% of GDP, household
savings will fall by 0.70%–0.78% of GDP in response to an increase in health
expenditure of 1.00% of GDP (column 3 of Table 8.1, and column A of
Appendix Table A8.1).8 Using PRC spending levels and ratio of disposable
income to GDP yields an impact of about 2.00% of GDP on household savings
for each 1.00% increase in government spending on health. Similarly, a
1.00% of GDP increase in social security spending reduces household
savings by 0.22%–0.29% of GDP in the sample (evaluated at the sample
mean), and 0.68%–0.72% of GDP in the PRC (column 3 of Table 8.1 and
column C of Appendix Table A8.1). However, public spending on education
is only significant when other social spending components are excluded
(column B, Appendix Table A8.1).9 With this caveat, it appears that a 1.00%
of GDP increase in public education spending leads to a decline in household
savings by 0.79% of GDP in OECD countries and by 1.26% of GDP in the
7
Needless to say, some caution is needed when extrapolating results based on the OECD country sample
to the PRC. The level of economic and institutional development in the sample countries is higher than
in the PRC. The coefficient on health expenditures (about 2.1) is nearly identical to that obtained by
Barnett and Brooks (2010) for the effect of public health spending on the saving of urban households
in the PRC.
8
Since column A in Appendix Table A8.1 includes public health spending only in the regression, the
coefficient of health spending may pick up the residual effects of other categories of spending as well.
Given the positive correlation among public expenditure categories and the negative correlation between
the error term (unobservable determinants of saving) and public social spending, the omitted variable
bias will tend to be negative. The results from Table 8.1, column 3 and Appendix Table A8.1, column
A, however, are very similar. Including government savings as an additional explanatory variable in
column A does not have a significant effect on the results (not reported).
9
Multicollinearity may affect these results. Public expenditures on health and education are correlated
with a coefficient of 0.65. Nevertheless, we obtain very similar results for public health and social
security spending, regardless of whether all three items are simultaneously included (column 3 of
Table 8.1), or each spending category is estimated separately (columns A and C in Appendix Table A8.1).
As column B in Appendix Table A8.1 shows, therefore, it seems reasonable to view public education
spending as having a significant, independent, negative effect on saving. Nonetheless, the result of
public spending on education should be interpreted with caution. One reason for these poor results
on the education variable could be that as most countries in the sample provide full access to basic
education, the spending variable picks up the effect of the latter while precautionary saving increases
mostly as a result of the insufficient availability of affordable higher education.
Table 8.2: Impact on Household Saving of a 1% of GDP Increase in

Government Expenditure (as % of GDP)
Total Social Health Social Security

Spending (Table 1, column Education (Table 1, column 3
(Table 1, columns 1, 3 and Appendix (Appendix and Appendix Table
Item 2 and 4) Table1, column A) Table 1, column B) 1, column C)
Marginal reduction
in household saving 0.14 – 0.24 0.70 – 0.78 0.79 0.22 – 0.29
in OECD
at 28.10% of GDP at 5.80% of GDP at 16.10% of GDP
Measured at OECD for total social at 6.30% of GDP for education for social protection
average spending for health spending spending
Marginal reduction
in household saving 0.56 – 1.03 2.09 – 2.12 1.26 0.68 – 0.72
in PRC
at 6.00% of GDP at 0.90% of GDP at 2.90% of GDP at 2.20% of GDP
Measured at current for total social for health spending for education for social security
levels in PRC spending spending spending
GDP = gross domestic product, OECD = Organisation for Economic Co-operation and Development,
Source: Table 8.1 and Appendix Table 8.1.
PRC. Table 8.2 provides a summary of the illustrative estimates of the

impact on household savings of a 1.00% of GDP increase in government
expenditures.
4. Assessing the Impact of Expanding Social Programs in

the PRC: Methodology and Data
This section and the next estimate the potential impact on household
consumption of higher social expenditures in the PRC using a generational
accounting framework and building on the findings above.10 According to this
framework, there are three channels through which an increase in government
social expenditures can raise household consumption (Box 8.2).
In the first channel, social expenditures will increase the aggregate level of
household lifetime resources and, hence, current consumption. This is because,
when households make decisions about the level of current consumption,
they are assumed to factor in not only current income, but also the expected
10
This approach is based on the analysis in Gokhale, Kotlikoff, and Sabelhaus (1996). See also Kirsanova
and Sefton (2007) for a very similar approach.
Box 8.2: Methodology for Estimating Consumption Impact

of Social Expenditures
A standard decomposition formula is used to capture various channels through which
an increase in social expenditures can boost current household consumption. The
starting point for the decomposition formula is a consumption function defined over
relevant socioeconomic groups (e.g., young/old, urban/rural, and poor/non-poor).
For each group, i, the level of consumption in time t ( cit ) is defined as:
cit = αirit
where αi is the average propensity of each group to consume today out of lifetime
resourcesa and rit is the net present value of the group’s remaining lifetime resources
(i.e., net lifetime disposable income). This approach is consistent with a basic life-
cycle model where households base current consumption decisions on lifetime
resources and reallocate changes in resources at any point in their lifetime across
years. Total consumption can then be decomposed as follows:
where Pit is the number of individuals in the group and is the

average level of lifetime resources in the population.
Based on this equation, changes in consumption can be decomposed into changes
in each of these four components of total consumption. These changes can be
interpreted as follows:
changes in total lifetime household income, Pt rt ;
changes in the distribution of total lifetime household income across groups,
;
changes in group propensities to consume out of group lifetime income, αt ;
and Pit
changes in the distribution of the population across groups, P .
t
The first two channels together constitute the income effect (reflecting the increase
in household real incomes), and the third captures the insurance effect (reflecting the
decrease in precautionary saving). Whereas the income effect can be estimated by
applying generational accounting to household survey data (Box 8.3), the insurance
effect needs to be estimated using alternative micro-econometric techniques. In
the current version of this paper, the final demographic effect is kept fixed, but it is
straightforward to impose projected demographic trends onto the household survey
data to project the likely impact of demographic change on the average propensity
to consume, and thus on total consumption.
a
Note that the average propensity to consume out of lifetime resources will be substantially lower than
the consumption rate (i.e., consumption out of current income), since lifetime income is distributed
across future years. Consistent with this, the average propensity to consume will be higher for the
elderly compared to the young.
Source: Authors’ based on analysis in Gokhale, Kotlikoff, and Sabelhaus (1996) and Kirsanova and
Sefton (2007).
future stream of income. An expansion of government social expenditures

(e.g., lower health and education charges or higher pensions) to cover a share of
expenditures until then borne by households will, therefore, increase household
lifetime resources and household current consumption. Even households not
currently incurring such expenditures (or receiving such transfers) can be
expected to increase current consumption since they will expect to incur such
costs (or receive such benefits) in the future.11 The consumption impact of
increasing social expenditures will, of course, be higher if these expenditures
are financed by diminishing a budget surplus rather than with taxes.
These impacts, in a second channel, will be amplified to the extent these
expenditures are concentrated in households with relatively high average
propensities to consume out of lifetime income. For example, pension transfers
to the elderly can be expected to have a higher consumption impact than
transfers to younger age groups, since the elderly tend to have a relatively
high propensity to consume out of remaining lifetime income. This holds
more generally for social expenditures that are positively correlated with age,
such as health expenditures.
Social expenditures can generate an additional—or third channel—impact
on current consumption by diminishing precautionary saving. Higher social
expenditures mean that households are faced with the possibility of lower future
health and education costs and higher retirement incomes and, therefore,
may save less to cover these expenses.
The first two channels capture the income effect of higher social expenditures,
and can be estimated by applying the generational accounting framework to
household-level data. The third channel represents the insurance effect of
these expenditures; the magnitude of this impact is more difficult to identify
and needs to be assessed separately using alternative estimation techniques.
Below, the potential insurance effect is inferred from a study for Taipei,China
and the results from the above panel regressions. They should, therefore, be
interpreted as illustrative of the potential magnitude of this effect. Both the
income and insurance effects are then combined to gauge the overall savings
impact.
Implementing the above framework entails three steps: (i) calculation of
lifetime resources for various socioeconomic groups, (ii) estimation of their
average propensities to consume, and (iii) translation of increases in public
11
Note that a temporary or one-off increase in expenditures (e.g., as part of a stimulus package) can be
expected to have a much smaller impact on current consumption, since households will distribute the
associated increase in real income over their life cycle.
expenditures into changes in lifetime resources. Individuals are first allocated
to groups according to three characteristics: current income (based on income
quintiles), residence in urban or rural areas, and age:
t Calculating lifetime (disposable) resources requires data on
individual wealth, incomes, taxes, and transfers. These data are taken
from household budget survey data. Each individual is allocated to a
group and then assigned the group mean for each of the lifetime income
components. Lifetime resources for each group member can be estimated
based on group averages and the assumption that future flows for each
group member are given by the group averages for successive age groups.
For example, one can assume that next year’s income for a rural individual
currently aged 20 and in the bottom consumption quintile is the current
income for the same type of individual aged 21 years.12 In other words,
the growth rate of incomes for a group member is given by the growth
in average incomes with age for that group. Net present values are then
calculated by assuming a common discount rate. Box 8.3 presents a more
detailed discussion of the household data used in the analysis.
t The group average propensities to consume out of lifetime income
are then calculated as the ratio of current consumption to lifetime
income. Figure 8.3 presents the pattern of these propensities across age,
urban/rural, and income groups. As expected, they increase with age and
are also higher for rural and lower-income groups. The relative importance
of the age effect is clear and reflects the simple fact that remaining lifetime
incomes of the elderly are allocated over fewer years.
t Expenditure reforms involve increasing the flow of resources to the
different groups over time. A given expenditure reform is translated into an
additional flow of resources to each group over time. The impact on current
consumption is derived as the increase in each group’s net present value of
lifetime resources times their average propensity to consume. For example,
increases in public health expenditures associated with reduced health care
charges can be expected to lead to a larger proportional increase in the net
present value of lifetime resources of the elderly (who have larger expenditures
today), compared with the increase for the young (who incur these large
expenses in the future). Appendix Table A8.2 presents the consumption
shares of various consumption categories across income groups.
12
A more rigorous approach would also take into account the existence of cohort and time effects based
on repeated cross-sections of household data. Such an analysis was undertaken by Chamon and Prasad
(2008), and yielded results qualitatively very similar to the results obtained in this paper.
The implicit assumption that changes in household consumption depend

on changes in household lifetime incomes means that estimated income effects
should be viewed as upper bounds. For example, credit-constrained households
will be unable to increase current consumption in line with higher lifetime
incomes, thus reducing the consumption impact of higher social expenditures.
Chamon and Prasad (2008) found that, contrary to what would be expected
from the life-cycle theory in the absence of credit constraints, both young and
old households had higher savings rates compared to other households. For
young households, this was attributed to their inability to borrow to finance
future lumpy consumption—e.g., for the purchase of a house or a consumer
durable, or for saving to finance future education costs for children. For the
elderly, it was attributed to high levels of precautionary savings due to the
risk of incurring high health costs in the future. Of course, the expenditure
reforms themselves may help reduce these savings rates for the young and
the elderly. However, to the extent that credit constraints persist, realization
of upper-bound consumption impacts will be dependent on the introduction
of supporting financial sector reforms.
5. Assessing the Impact of Expenditure Reforms

Assessing the impact of social expenditure reforms on household consumption
requires the identification of specific expenditure measures and the outlays
on them. The consumption impact of an increase in public expenditures on
pensions, health care, and education are assessed in this section. These reform
measures have been discussed by the government as part of recent reform
proposals, and identified in the literature as important for reducing the high
household savings rate. The consumption impact of such reforms is explored
by simulating the following reforms within the generational accounting
framework previously described:
t Pension transfer. A cash transfer proportional to current income is
given all individuals over 55 years old (i.e., universal coverage). Each
individual starts to receive this transfer upon reaching this age threshold.
According to McKinsey Global Institute (2009), pension coverage in
2009 was extended to around 90% of urban households and 20%–25%
of rural households and migrants. The reform therefore should be
interpreted as significantly expanding coverage, including in rural areas.
Disaggregating the simulated reform by urban and rural areas allows
Figure 8.3: Average Propensities to Consume Out of Lifetime Income
Average Propensity to Consume
Age
Age
Age
Note: Q1 represents the quintile with the lowest income, and Q5 the quintile with the highest income.
Source: Authors’ estimates based on 2002 Chinese Household Income Survey data.
Box 8.3: Constructing Lifetime Income and Average

Propensities to Consume
The data for the analysis are taken from the Chinese Household Income Project
(CHIP). The survey questionnaire was designed by the Institute of Economics of the
Chinese Academy of Social Sciences, and interviews were implemented by the National
Bureau of Statistics. The sample is a subsample of that used by the National Bureau
of Statistics for its own nationally representative annual survey, which is not publicly
available. CHIP covers households in 12 urban provinces (6,800 households and
20,600 individuals) and 22 rural provinces (9,200 households and 38,000 individuals),
as well as migrant households in the 12 urban provinces (2,000 households and 5,300
individuals). The data are for 2002, the latest survey available.
CHIP contains information on assets, incomes, and consumption. All consumption

data are for households, so per capita consumption levels are calculated by dividing
total household consumption by household size. The urban sector survey provides
income data at the individual level. The rural sector survey provides nonagricultural
income at the individual level and agricultural income at the household level. Individual
income levels in the rural sector are thus calculated as individual nonagricultural
income plus per capita agricultural income, calculated as total household agricultural
income divided by household size. Per capita net asset levels are calculated as
the value of net household assets, including the market value of privately owned
houses, divided by the household size.
Aggregated survey data for income and consumption for socioeconomic groups
are merged with macroeconomic national data and demographic data. CHIP data
are used to calculate the distribution of average propensities to consume across
different age and income groups in both the urban and rural sectors. Individuals
between 18 and 73 years old (the life expectancy in the PRC) are allocated to age
groups, and the net present value of individual income is calculated by assuming
future wages follow that of age-specific wages augmented by a growth rate of
7.7% and discounted based on an interest rate of 11.7%. Survival rates applied
are taken from the latest 2006 World Health Organization Life Table for the PRC.
The resulting discount rate is 0.96. The calculation of lifetime resources also
includes the present value of net transfers, net pension benefits, in-kind benefits
for education and health, and financial assets. The average propensity to consume
is calculated as the current consumption level, including education and health
spending, divided by lifetime resources. These propensities are then applied to
2007 national accounts and population data (the latest available) to calculate
group consumption levels.
Although the use of 2002 household survey data reflects primarily the constraints
of data availability, these data are likely to provide a good approximation of the
consumption impact of expanded social expenditure programs. According to the
summary of household survey data presented in Chamon and Prasad (2008), the
household savings rate increased from 16.7% in 1992 to 24.7% in 2006, with most
of this increase occurring by 2002 when the savings rate reached 23.1%. Similarly,
most of the increase in household expenditures on health and education had occurred
before 2002. The share of health expenditures in total consumption increased from
2.5% in 1992 to 7.4% in 2004, compared to 7.1% in 2002. Education expenditures as
a share of total consumption increased from 8.8% to 14.4% over the same period,
compared to 15% in 2002. The study of Chamon and Prasad (2008) was restricted
to urban households, while CHIP data also covers rural households.
Source: National Bureau of Statistics. 2002. Chinese Household Income Survey. Beijing: National
Bureau of Statistics; and Chamon, M., and E. Prasad. 2008. Why are Saving Rates of Urban
Households in China Rising? IMF Working Paper No. WP/08/145. Washington, DC: IMF.
estimation of the impact of a reform that extends pension coverage to

rural areas.13
t Education transfer. A transfer proportional to current education
expenditures is given all individuals. The time profile of this transfer
mirrors the age profile of education expenditures. This year, for example,
20-year-old individuals receive a transfer proportional to their current
share of total education expenditures. Next year they will receive a transfer
equivalent to that received by individuals currently 21 years old. The
amount received will depend on the size of the budget allocation of the
government.14
t Health transfer. A transfer proportional to current health expenditures
is given all individuals. As for education, the time profile of this transfer
mirrors the age profile of education expenditures. (McKinsey Global
Institute [2009] also evaluated the impact of increasing public health
spending).
Appendix Tables A8.3 and A8.4 present the average transfers received by
different age and income groups under the different simulated expenditure
reforms.
13
McKinsey Global Institute (2009) estimated the impact of a fully-funded scheme.
14
McKinsey Global Institute (2009) evaluated the impact of an expansion of student loans for tertiary
education but not the consumption impact of higher public education expenditures.
The expenditure reforms simulated are assumed to be permanent in

terms of their continued existence over the lifetime of each group, and their
financing through a permanent decrease in the fiscal surplus. The budget for
these reforms is assumed fixed at 1% of GDP each year, so that the absolute
size of the budget allocation for these expenditures increases with GDP. They
are assumed to be financed from existing budgetary surpluses—i.e., they are
assumed to entail no increase in taxes.
The financing of expenditure reforms by reducing the fiscal surplus results
in an increase in total household net lifetime resources. The net consumption
impact reflects the transfer of resources from government savings to households,
with the increase in household incomes in turn leading to an increase in current
household consumption. If expenditure reforms were budget neutral (e.g.,
financed from income taxes), then the net consumption impact would obviously
be much smaller and reflect simply a redistribution of resources across households
with different average propensities to consume, as well as some impact via the
insurance channel. Similarly, temporary reforms (akin to stimulus packages)
would have substantially lower impacts on current consumption since the change
in lifetime resources would be much smaller.
As indicated earlier, expenditure reforms will have both an income effect
and an insurance effect on current consumption. The income effect reflects the
increase in lifetime resources, with education and health transfers reducing
the net cost of these services to households. The insurance effect reflects the
increase in the average propensity to consume out of lifetime resources due
to a decrease in precautionary savings.
Table 8.3: Income Impact of Expenditure Reforms on Household

Consumption (% of GDP)
Expenditure Reform Pension Health Education

Simulation
Total 1.42 0.77 0.51
Urban 0.92 0.46 0.24
Rural 0.50 0.32 0.27
Budget shares
Urban 0.75 0.69 0.58
Rural 0.25 0.31 0.42

Note: Simulations assume a 1% of GDP increase in each expenditure category annually.
Source: Authors’ estimates based on Chinese Household Income Survey 2002 data.
5.1 Income Effect
The variation in income effects reflect the age profile of expenditures. The
income effect for the three types of government social expenditures under
Simulation 1, disaggregated by urban and rural population, are presented
in Table 8.3. Since the lifetime budget for the rural population is smaller, so,
too, will be the consumption impacts. The results suggest that the impact on
current consumption is highest for pension expenditures (at 1.4% of GDP),
followed by health (0.8% of GDP), and education (0.5% of GDP).15
The impact of a unit of government social spending on consumption is
substantially higher in rural areas. This reflects the higher propensities to
consume in rural areas, which in turn partly reflects lower income levels. For
example, under Simulation 1 the impact of pension expenditures in urban areas
is 0.92 compared to 0.50 in rural areas. However, only 25% of expenditures
go to rural households. Adjusting for the difference in pension expenditures,
the income effect of pension expenditures in rural areas is about 67% higher
than in urban areas—i.e., (0.50/0.25) divided by (0.92/0.75). The equivalent
differences for health and education are 56% and 55%, respectively. Targeting
expenditures to rural areas will clearly result in a higher impact on household
consumption. Similarly, targeting transfers to low-income households is likely
to be a more cost-effective approach to increasing household consumption.
5.2 Insurance Effect

The insurance effect is more difficult to assess given the paucity of empirical
evidence on the determinants of precautionary saving. A relevant study by Chou,
Liu, and Hammitt (2006), noted earlier, analyzed the impact of Taipei,China’s
extension of health insurance coverage (from 57% of the population in 1994
to 96% in 2000) on household precautionary savings. Their estimates suggest
that an increase in health expenditures of 1 percentage point of GDP increased
current household consumption by 0.4% to 0.6% of GDP due to the insurance
effect. To estimate the equivalent insurance impacts of education and pension
expenditures for this paper, the ratio of this health insurance impact is taken
15
Under the smaller lifetime budget in Simulation 2, the income effects are lower, at 0.8% for pensions,
0.5% for health, and 0.4% for education. Improved targeting could increase these further—for example, if
pensions were given only to those without pensions (e.g., by increasing coverage). Since the beneficiaries
under this scenario would more likely be rural, poorer than existing pension beneficiaries, and therefore
have higher average propensities to consume, the impact on current consumption would be larger.
to the total (income and insurance) health impact estimated in Section 3.

This yields a ratio of 0.24—i.e., 0.5 divided by 2.1. Applying this ratio to the
estimated total consumption impact of education and pension expenditures
(estimated at 1.3 and 0.7, respectively) yields insurance impacts of 0.31% and
0.17% of GDP, respectively.
5.3 Total Effect

Adding the income and insurance effects suggests that the household
consumption impact of social expenditures may be substantial. Total household
consumption impact across expenditures is calculated by adding the income
effects from Table 8.3 to the insurance impacts estimated above. The resulting
total consumption impacts range from 1.6% of GDP for pensions,16 to 1.3% for
health,17 and 0.8% for education (Table 8.4).18 This implies that a 1 percentage
point of GDP increase in social expenditures allocated evenly across these
expenditure categories would result in a permanent increase in household
consumption of 1.2% of GDP. Allocating a greater share of expenditures
to health and pension expenditures can be expected to generate a higher
household consumption impact.
The results suggest that increasing government social expenditures can
make an important contribution to increasing household consumption.
For example, assuming government expenditures are evenly distributed
across education, health, and pension expenditures, a 3.0% of GDP increase
in the household consumption ratio would require a 2.5% of GDP increase
in total social expenditures maintained over the medium term. Targeting
a higher proportion of the expenditure increase to health and pensions,
or to rural and low-income households, would reduce the required budget
increase. For example, the same 3 percentage point of GDP increase in the
16
McKinsey Global Institute (2009) estimated that the consumption impact of a fully funded pay-as-
you-go pension scheme would range from –0.20% to 0.50% of GDP. Our estimates of such a scheme
suggest that financing through income taxation would decrease our estimate of consumption impact
by 0.57% of GDP, giving a net impact of around 1.10%.
17
The health impact is substantially higher than that reported by McKinsey Global Institute (2009). That
report estimated that household consumption could increase by 0.40% to 0.60% of GDP through a
combination of increasing government health expenditure by 1.10% to 2.80% of GDP, and expanding
health insurance coverage (presumably self-financed). This suggests that a 1.00% increase in public
health expenditures would increase consumption by around 0.25% of GDP.
18
McKinsey Global Institute (2009) estimated that the consumption impact of increasing access to student
loans from the current 10.0% to between 33.0% and 50.0% would range from 0.4 to 0.7 percentage
points of GDP.
Table 8.4: Household Consumption Impact of Expenditure Reforms
(% of GDP)
Expenditure Reform Pension Health Education

Total 1.6 1.3 0.8
Income Effect 1.4 0.8 0.5
Insurance Effect 0.2 0.5 0.3

Note: Income effects assume a 1% of GDP increase in each expenditure category annually.
household consumption ratio would require an increase of only 1.9% in pension

expenditures or, alternatively, a 2.2% increase in health expenditures. These
reforms could be complemented by structural reforms that increase the share
of wages in national income by rebalancing domestic growth toward domestic
consumption. For example, simulating a 1.0% additional growth in labor
income in the PRC in the above framework would generate a 0.7% increase
in current household consumption.
Consumption impacts would obviously be lower if the increase in social
expenditures was financed with taxes. A 1.0% of GDP increase in expenditures
in each expenditure category financed by an increase in income taxes would
lead to an offsetting decrease in consumption of nearly 0.6% of GDP. This
would imply a net increase in household consumption limited to 0.2% of
GDP for education, 0.7% for health, or 1.0% for pensions—much smaller than
when social expenditures are financed by a reduction in the fiscal surplus.
The positive net impact reflects the redistributive effect of the combined tax
and expenditure reforms, with resources being redirected from those with
low to those with high propensities to consume.
6. Summary and Policy Implications

This chapter has explored a number of issues related to the low and declining
household consumption ratio in the PRC. It first examined explanations
advanced for developments in this ratio, including the marked increase in
the household savings rate and the decline in the share of household income
in GDP. An underlying cause of the higher savings rate is apparently the
greater uncertainty facing the household sector following structural changes
in the economy, and the reduced support by the government for education
and health services as well as old-age pensions. The paper then undertook an
analysis of the determinants of the household savings rate using a panel of
advanced and emerging market countries. The results support the premise
that government social expenditures (on education, health, and pensions) can
be an important determinant of the household savings rate.
Based on analyses in the preceding section, simulation analyses were used
to explore how and through what channels an increase in government social
expenditures in the PRC could lead to an increase in the consumption ratio.
Three channels were identified: (i) a direct income channel, (ii) a distributional
channel, and (iii) an insurance channel. Simulations using household survey
data showed that the effect of an increase in public spending on the three
categories could be significant. In particular, a sustained 1 percentage point
of GDP increase in government spending would likely lead to an increase in
the household consumption ratio of up to 1.25 percentage points of GDP.
These results also underline the likelihood of broader benefits of expenditure
reforms in connection with the PRC’s large external current account surplus
and global imbalances. The adjustment required to reduce the current account
balance to a more sustainable level is large. Estimates suggest that household
consumption would have to increase by some 3 to 4 percentage points of
GDP—assuming the savings rates of the corporate sector and the government
remain unchanged—to help rebalance world demand. This suggests that other
structural reforms would also be important, such as measures to increase the
availability of credit, retail distribution, and the share of wages in national
income by greater emphasis on domestic consumption in domestic growth.
Nonetheless, the findings of this paper suggest that greater public expenditure
on health, education, and pensions could make an important contribution.
Appendix Table A8.1: Individual Component of Public Expenditures
Household Saving Panel Regression, 1990–2008 for
OECD Sample
(A) (B) (C)

Explanatory Variables FE FE FE
-0.06 -0.11 -0.42**
Growth, per capita GDP (-0.28) (-0.55) (-2.09)
207.9*** 98.73 41.36
Initial GDP per capita (log) (3.24) (1.45) (0.48)
10.01*** -4.83 -2.55
Initial GDP per capita (log), squared (-3.12) (-1.43) (-0.61)
-1.92 -2.24 -0.97
Private credits (-1.69) (-1.60) (-1.10)
-0.61** -0.92*** -0.79***
Old-age dependency ratio (-2.22) (-3.02) (-3.05)
0.46** 0.40* -0.19
Young-age dependency ratio (2.68) (1.75) (0.67)
-6.74***
Public health spending (-4.92)
0.42***
Public health spending, squared (5.09)
-5.76***
Public education spending (-3.43)
0.37***
Public education spending, squared (3.23)
-2.31**
Social security spending (-2.59)
0.06**
Social security spending, squared (2.57)
No. of observations 78 78 78
No. of countries 24 24 24
R2 0.78 0.72 0.71
FE = fixed effects estimation, GDP = gross domestic product, OECD = Organisation for Economic Co-operation
and Development.
Note: The panel consists of four 5-year periods for 24 countries. Heteroskedasticity and country-specific
autocorrelation consistent t-statistics are reported in parentheses. Levels of significance are indicated by
asterisks: *** 1 percent, ** 5 percent, * 10 percent. An intercept term is included in each regression.
Appendix Table A8.2: Consumption Shares of Various Consumption

Categories
Income
Quintiles Urban Rural
Food Non-food Housing Education Health Food Non-food Housing Education Health
Q1 0.405 0.274 0.175 0.071 0.075 0.465 0.242 0.145 0.091 0.057
Q2 0.387 0.309 0.153 0.078 0.073 0.468 0.266 0.125 0.088 0.053
Q3 0.371 0.334 0.150 0.077 0.068 0.443 0.285 0.160 0.069 0.043
Q4 0.346 0.343 0.180 0.071 0.059 0.417 0.292 0.165 0.080 0.046
Q5 0.311 0.371 0.186 0.075 0.057 0.348 0.323 0.205 0.088 0.036
Average 0.364 0.326 0.169 0.074 0.066 0.428 0.282 0.160 0.083 0.047
Source: Authors’ estimates based on 2002 CHIP household survey data.

Appendix Table A8.3: Per Capita Transfers under Different Social
Expenditure Reforms
Age Urban Rural
Assets Income Education Health Assets Income Education Health

18 39743 372 290 27 8325 1402 334 31
19 43187 414 354 29 7873 1260 336 39
20 37324 649 337 30 8072 1181 243 33
21 44410 1397 261 35 8204 1290 226 37
22 39042 1690 214 34 7805 1386 160 39
23 45992 2889 111 41 7704 1343 116 40
24 47072 3387 78 44 8286 1349 80 43
25 56797 4625 75 37 7641 1421 89 42
26 40406 4861 75 35 7472 1270 50 37
27 43441 4880 58 36 8197 1363 41 40
28 39415 5699 96 41 8443 1564 45 37
29 39922 6542 106 43 8514 1626 54 35
30 38313 7628 148 33 9136 1904 67 31
31 46182 8377 187 34 10258 1809 83 49
32 37348 7515 184 57 10607 1930 124 32
33 39333 9284 202 35 9954 1909 141 27
34 52636 9774 252 48 11502 1998 200 34
35 47233 9481 235 39 11411 2296 196 33
36 38349 9284 199 35 10684 2266 278 37
37 50309 9926 318 36 12687 2325 312 26
38 51734 9560 268 36 12007 2267 345 26
39 49219 9762 305 32 13176 2461 424 26
40 63633 10169 327 38 12018 2050 481 35
41 55964 9661 310 41 12880 2413 432 49
42 53693 9706 341 41 12390 2435 391 34
43 62401 10527 393 38 12629 2227 423 34
44 51913 9210 382 35 11584 1490 402 42
45 47947 9603 345 37 9810 1593 333 32

Appendix Table A8.3: Per Capita Transfers under Different Social

Expenditure Reforms (continued)
46 49309 9471 335 29 9627 1502 291 45
47 47387 8839 334 34 9511 1458 268 38
48 50545 9026 287 34 9987 1419 281 56
49 49896 9472 230 33 8754 1565 222 49
50 43669 8854 182 36 9595 1400 178 41
51 52642 9006 163 41 9838 1245 131 37
52 54751 8932 162 46 9809 1174 168 45
53 51701 8571 112 47 10505 1250 95 66
54 45751 8614 92 50 9739 1428 111 53
55 55705 9218 105 48 9074 1459 87 54
56 52896 8318 93 78 8470 1352 59 103
57 61208 8739 60 88 10698 1357 57 81
58 52418 8040 69 94 8833 1095 56 88
59 51825 8893 171 86 10332 1218 44 89
60 48368 8062 82 87 8892 1112 103 132
61 42882 7475 67 88 9993 1086 70 86
62 41134 7427 70 99 8537 1075 53 96
63 49095 6765 66 99 9895 1038 70 75
64 48824 7982 91 89 8753 1120 78 92
65 45892 6948 53 95 8504 883 84 84
66 38477 5623 80 78 9559 1217 91 65
67 52930 6996 95 103 7219 1054 130 86
68 48985 7275 96 86 7160 1160 91 85
69 53534 7012 128 86 6971 1180 74 54
70 40391 8487 70 100 7600 723 120 98
71 53443 6319 63 73 9373 1181 150 89
72 40568 6870 88 113 7737 1101 204 71
73 41888 7657 120 100 8046 1035 137 90
Average 47662 7353 179 55 9505 1495 177 54
Source: Authors’ calculations.

Appendix Table A8.4: Per Capita Transfers and Consumption
Income Quintile Consumption Pension Health Education
Urban
Q1 11284 244 386 286
Q2 12594 325 423 349
Q3 14537 425 449 399
Q4 18148 539 490 457
Q5 26942 783 703 722
Rural
Q1 3668 13 47 58
Q2 3991 32 48 62
Q3 4543 52 44 55
Q4 5370 84 55 75
Q5 8472 147 69 131
Source: Authors’ estimates based on 2002 CHIP household survey data.

References
Barrell, R., I. Hurst, and S. Kirby. 2009. How to Pay for the Crisis or Macroeconomic
Implications of Pension Reform. National Institute of Economic and Social
Research (NIESR) Discussion Paper No. 333. London: NIESR.
Spending Boost Consumption? IMF Working Paper No. 10/16. Washington, DC:
IMF.
Bosworth, B., and G. Chodorow-Reich. 2007. Saving and Demographic Change: The
Global Dimension. Center for Retirement Research at Boston College Working
Paper No. 2007-2. Boston: Center for Retirement Research at Boston College.
Caballero, R. J., E. Farhi, and P. Gourinchas. 2008. An Equilibrium Model of Global
Imbalances and Low Interest Rates. American Economic Review 98: 358–393.
Rising? IMF Working Paper No. WP/08/145. Washington, DC: IMF.
Chou, S., J. Liu, and J. K. Hammitt. 2006. Households’ Precautionary Behaviors—the
Effects of the Introduction of National Health Insurance. Review of Economics of
the Household 4: 395–421.
Edwards, S. 1996. Why Are Latin America’s Saving Rates So Low: An International
Comparative Analysis. Journal of Development Economics 51(1): 5–44.
Eurostat. 2009. Household Saving Rate Higher in the EU than in the USA despite Lower
Income. Statistics in Focus 29/2009. Available at: http://epp.eurostat.ec.europa.eu/
portal/page/portal/product_details/publication?p _product_code=KS-SF-09-029
Gokhale, J., L. J. Kotlikoff, and J. Sabelhaus. 1996. Understanding the Postwar Decline
in US Saving: A Cohort Analysis. The Brookings Papers on Economic Activity 1.
Washington, DC: Brookings Institute.
Gruber, J., and A. Yelowitz. 1999. Public Health Insurance and Private Savings. Journal of
Political Economy 107(6): 1249–1274.
IMF. 2005. World Economic Outlook. Washington, DC: IMF.
Kantor, S. E., and P. V. Fishback. 1996. Precautionary Saving, Insurance, and the Origins
of Workers’ Compensation. The Journal of Political Economy 104(2): 419–442.
Kirsanova, T., and J. Sefton. 2007. A Comparison of National Saving Rates in the UK, US
and Italy. European Economic Review 51(8): 1998–2028.
Kotlikoff, L. J. 1989. What Determines Savings? Boston: The Massachusetts Institute of
Technology Press.
Kraay, A. 2000. Household Saving in China. The World Bank Economic Review 14(3):
545–570.
the World? The Review of Economics and Statistics 82(2): 165–181.
McKinsey Global Institute. 2009. If You’ve Got it, Spend it: Unleashing the Chinese
Consumer. Beijing: McKinsey Global Institute.
Modigliani, F., and S. Cao. 2004. The Chinese Saving Puzzle and the Life-Cycle
Hypothesis. Journal of Economic Literature 42(1):145–170.
National Bureau of Statistics. 2002. China Statistical Yearbook. Beijing: China Statistics
Press.
OECD. 2009. OECD Social Expenditure Database. Paris: OECD. Available at: www.oecd.
org/els/social/expenditure.
Park, D., and K. Shin. 2009. Saving, Investment, and Current Account Surplus in
Developing Asia. ADB Economics Working Paper Series No. 158. Manila: ADB.
Prasad, E. 2009. Rebalancing Growth in Asia. Institute for the Study of Labor (IZA)
Discussion Paper No. 4298. Bonn: IZA.
Wei, S., and X. Zhang. 2009. The Competitive Saving Motive: Evidence from Rising Sex
Ratios and Saving Rates in China. NBER Working Paper No. 15093. Cambridge,
MA: NBER.
World Bank. 2009. World Development Indicators. Washington, DC: World Bank.
Comments
Akiko Terada-Hagiwara
As ADB President Kuroda noted in his keynote speech, the Asian

Development Outlook 2009 that was released earlier this year looked at
issues surrounding rebalancing Asia’s growth. The section on the role of
consumption and saving identified the important role of social security
systems in affecting saving and consumption in developing Asia. In a
similar vein, the chapter presented by David Coady investigates the policy
impacts of public expenditures on social programs, with particular attention
to the PRC.
In particular, the chapter looks into the likely impact of expanding
social programs on household consumption in the PRC. It first examines
factors determining the household saving rate using aggregate sample data
from the OECD. It then uses household budget survey data to compute
lifetime income and average propensity to consume, and undertakes policy
simulations.
It is important to note that unlike the widely observed increasing corporate
savings in the region, rising household saving rates are unique to only a few
countries, including India and the PRC. These are large economies in Asia,
and the implications of this paper could be applied to India as well to the
PRC. According to the paper, the increasing saving rate in the PRC accounts
for two thirds of the total change in the national consumption rate. The rest
is due to the decline in the share of household income in GDP. Recent studies
suggest that precautionary saving for health and education outlays plays an
important role in explaining the increase in the PRC household saving rate.
This chapter makes a contribution by quantifying the impacts of possible
policy changes.
1. Panel Analysis
My comments are divided into two parts—first on the panel analysis, and then
on the generational accounting framework analysis. In the panel analysis
section, the analysis suggests a negative nonlinear relationship between
government social expenditure and the household saving rate. This is an
interesting exercise and result, given the fact that public spending in the
PRC is currently small as a proportion of GDP. However, no explanation is
offered as to why we should expect this nonlinear relationship. This result of a
nonlinear relationship clearly warrants more discussion. Further, in order to
better link this panel analysis to the analysis that follows, I would suggest that
the estimation include interactive terms with household disposable income
(or assets held), along with the public spending variables. We would naturally
expect that lower-income households would be more sensitive to a change in
public expenditures, and hence the nonlinearity in the relationship.
Additionally, a number of variables that are included in the analysis turned
out to be insignificant. Before concluding that this is the case, more analyses
could be done. As for per capita GDP growth, the results suggest it is generally
negative (although insignificant). The sign is puzzling given that economic
relationship, such as habit formation, predicts a positive sign. The authors
could instead try to include growth of household disposable income, which they
already have. Likewise, replacing the initial GDP variables with the household
income is a natural avenue to take as we are testing for the household saving
rate and not national saving. For financial sector development, I believe that a
variable such as consumer credit, rather than credit to the private sector (which
is probably mainly to the corporate sector), would be more appropriate.
2. Generational Accounting
In the section on generational accounting framework analysis, the chapter
computes average propensity to consume using household income survey data
in order to carry out policy simulations. (The survey year needs clarification—is
it 2002 or 2006?) In the model construction, household consumption and
income are divided by family size, except for urban and rural nonagricultural
income. If households include children, however, bias is introduced. This is
because individuals with income would be adults, while consumption per capita
includes children who do not consume as much as adults. This would result in
underestimation of the average propensity to consume. The computed average

propensities to consume are still interesting, though the variations seem to be
driven by the income level—which would be so almost by construction.
The chapter then presents simulations for public expenditures on pension,
education, and health transfers using the computed average propensity to
consume. The first point that I would question is the assumption that the
average propensity to consume is fixed, even after the policy changes. I suspect
that there would likely be a feedback effect—i.e., the policy changes would
affect anticipated lifetime income and also the average propensity to consume
if households know the policy changes are permanent. For example, a cohort
of age 20 in 2010 would behave differently from that of age 20 in 2008.
Another point concerns the relevance of rural households. The analysis finds
that the income effects are higher in rural areas, perhaps due mainly to lower
income levels. If the decomposition is done separately for rural households,
the contribution of low income to declining consumption might be much
more than one third for this group. Additionally, Chamon and Prasad (2008)
indicated that they did not look at the rural area as their saving rate is not very
high. The problem there seems to be more about the low income level.
The last point I would raise is the relevance of other social infrastructure,
which would have an impact on saving and consumption in the PRC. Appendix
Table A8.2 of the chapter suggests the importance of consumption on housing,
which is another important social infrastructure. Past studies tell us that home
ownership does not encourage more consumption in the PRC, even though
households owning homes are free from saving for a mortgage down payment,
for example. Chamon and Prasad (2009) argue that households continue to
save due to the poor quality of housing. In this regard, public expenditure to
improve dwellings is important for the reduction of household saving.
Reference
Rising? NBER Working Paper 14546. Cambridge, MA: NBER.
Social Policy Reforms and Growth Rebalancing in ASEAN 205
IX
Social Policy Reforms and Growth
Rebalancing in ASEAN
Gloria O. Pasadilla and Prayoga Wiradisuria
1. Introduction
Outward-oriented growth strategies have served East Asia well for many
years. But the global financial crisis, having highlighted the persistent global
imbalance—high trade and fiscal deficits in one part of the world and high
trade surpluses and domestic savings in another—is forcing a reexamination
of such growth strategies.
No one is convinced that abandoning external orientation is the solution
to the global imbalance. But might there be scope for increased domestic
demand in the region to lessen dependence on external demand? Some
scholars suggest that, in view of East Asia’s very high savings, the region should
increase consumption and reduce saving to perk up global demand. One
suggested option to improve domestic demand is to increase social protection
expenditures to entice consumers to forgo excessive precautionary savings
and increase current consumption.
This chapter will analyze the appropriateness of this policy prescription
in the context of ASEAN members. The chapter first assesses whether
ASEAN countries have excess savings. Second, it looks at the region’s
social protection policies to see potential scope for improvement. Finally,
it considers the potential economic effects of social protection spending,
drawing from results of studies that have linked elements of social
protection with labor supply, economic growth, private transfers, savings,
and investments.
2. Savings in ASEAN
How big are ASEAN’s savings? How much room is there for savings reduction
to boost demand? Relative to other regions such as Latin America or Europe,
ASEAN-5’s1 savings as a percentage of GDP are fairly high at 31% in 2008. The
savings rate for East Asia and the Pacific, of which ASEAN is a part, is much higher
at 50% because of very high savers in the group, notably the PRC. Compared to
other countries by income group, ASEAN-5’s savings rate is around the same
level as the average for middle-income countries (see Figures 9.1a and 9.1b).
Individual countries in ASEAN-5 have followed different savings paths
(Figure 9.2). Thailand was the highest saver until its savings started dropping
in the mid-1990s. Since then, Thailand had not yet recovered its savings/GDP
rate in the early 1990s. The Philippines, the historically lowest saver, had a
remarkably slow but steady rise in its savings rate starting in 1992, such that as
of 2008, its savings rate exceeded that of Thailand. Viet Nam’s savings rate has
increased rapidly since the late 1990s, making it second to the highest saving
country, Malaysia. Like Thailand, Malaysia’s savings rate dropped during the
Asian crisis but has since recovered and, as of 2008, Malaysia’s savings rate is
36.6%. Indonesia, in contrast, is still slightly below its pre-crisis savings ratio.
Are these savings rates excessively high? Conversely, are ASEAN-5 households
consuming enough? Since the figures above are ratios of national savings to
GDP, they do not show how much households save. A breakdown of savings into
government, corporations, and households indicates that household savings is
not responsible for the overall high savings in the Philippines and Thailand (the
only two ASEAN-5 countries for which such data was available). Figure 9.3 shows
that household savings as a percentage of GDP in the Philippines amounted
to only 2.2% of GDP in 2008 compared to Thailand’s 7.3%. In the Philippines,
corporations (private and state-owned) have dominated savings, averaging
more than 12% from 2002 to 2008, while household savings have consistently
declined since 2003. In contrast, government savings have markedly improved
since 2003. In Thailand, corporate saving has increased over the years and
was responsible for one third of total savings in 2008.2 Though household
savings declined after the Asian crisis, they have been creeping back toward
the precrisis savings rate.
1
ASEAN-5 comprises Indonesia, Malaysia, Philippines, Thailand, and Viet Nam.
2
The disaggregated data on savings is slightly different from the gross savings reported in the World Bank’s
World Development Indicators Database. In this database, savings are derived from the national income
accounts, while household savings from CEIC have come from national household survey data.
Figure 9.1a: ASEAN-5 Savings Rates in Comparison with Other Regions
% of GDP
Figure 9.1b: ASEAN-5 Savings Rates in Comparison with Income Groups

% of GDP
ADB = Asian Development Bank, ASEAN = Association of Southeast Asian Nations, GDP = gross domestic
product.
Notes: ASEAN-5 comprises Indonesia, Malaysia, Philippines, Thailand, and Viet Nam.
ASEAN-5 savings refer to gross national saving as a percentage of GDP downloaded from ADB’s Statistical
Database System on 7 November 2009. Data were missing for Viet Nam for 1990–1995 and Malaysia for
1992–1998.
Other countries’ savings refer to gross saving as a percentage of GDP downloaded from the World Bank’s World
Development Indicators Database on 1 November 2009.
Gross national savings is computed as gross national product minus total consumption, thus this value includes
the net factor income from abroad (ADB). Gross savings are calculated as gross national income less total
consumption, plus net transfers (World Bank’s World Development Indicators definition).
Sources: Authors’ calculations based on World Bank’s World Development Indicators Database and ADB’s
Statistical Database System.
Figure 9.2: ASEAN-5 Countries Savings Rates

% of GDP
ASEAN = Association of Southeast Asian Nations, GDP = gross domestic product.

Note: ASEAN-5 comprises Indonesia, Malaysia, Philippines, Thailand, and Viet Nam.
Source: Authors’ calculations based on ADB’s Statistical Database System, downloaded 1 November 2009.
With the low household savings rates shown in Figure 9.3, households appear
unlikely to be able to contribute significantly to increasing consumption and
boosting domestic demand during the global financial crisis. This is particularly
so for the Philippines, where households savings is a meager 2% of GDP. The
simple graph casts doubt on the argument that, in view of the region’s high
savings, social protection should be improved so that households can reduce
savings and increase consumption.
To compare household savings conditions in ASEAN with those in other
countries, we obtained households savings rates (ratio of household savings
to disposable income) for selected countries (Table 9.1). Among the three
selected ASEAN countries, Indonesian households save the most from their
disposable income, while households in the Philippines save the least, roughly
on par with the US. This suggests that households in Indonesia and Thailand
have relatively more space than those in the Philippines to increase their
consumption by saving less.
Concerning private consumption, its ratio to GDP in ASEAN-5 is already
comparatively high. Breaking down ASEAN-5’s consumption between
households and government reveals that household consumption expenditure
Figure 9.3: Saving Composition of Selected ASEAN Countries
Philippines
% of GDP
Thailand
4.3
5.0 2.7
5.5
2.8
2.8
% of GDP
4.6
2.0 3.7 4.8 4.7
6.1 6.4
2.7
5.8 5.4
4.8 4.3
5.7 4.7
10.1
8.5
7.2 7.3
5.0 4.5 4.2 5.0
3.6 3.9

Note: Corporate savings include those of both private and government corporations.
Sources: Authors’ calculations based on CEIC Data Company, Ltd., downloaded 2 November 2009.
Table 9.1: Household Savings as a Percentage of Household Disposable

Income for Selected Countries
Average Household Saving as a Portion of Household

Country Disposable Income (%)
Thailand 9
Indonesia 17
Philippines 4
Korea, Rep. of 10
Australia 0
United States 3
Notes: Averages of yearly data for 2004–2008 for Korea, Philippines, Thailand, and United States. Averages of
2003 and 2004 for Indonesia. Australia experienced negative household savings rates in 2004 and 2005 and
positive from 2006 onwards, hence the zero average for the period.
Australia’s and Korea’s yearly data are averages of quarterly data. United States yearly data is average of monthly
data.
Sources: Authors’ calculations based on CEIC Data Company Ltd, downloaded 2 November 2009. Indonesian
data was obtained from Santoso and Sarie (2007).
is responsible for the majority of total consumption (Figures 9.4a and 9.4b).
Unlike the PRC, which shows relatively low household consumption compared
to other countries, ASEAN-5’s household consumption is on par with the high
rates of developed countries. Malaysia is the only ASEAN country that has
private consumption of less than 50% of GDP. The Philippines has the highest
consumption rate of 76%, even higher than that of the US, while Thailand,
Viet Nam, and Indonesia are somewhere in between the levels of Malaysia
and the Philippines.
Given the large share of household consumption in ASEAN-5 GDP, it is
unclear whether further growth of consumption is desirable. Excessively high
consumption may be dangerous, and sufficient savings are important for
long-term economic development. Increasing aggregate consumption in some
ASEAN-5 countries might amount to a cure that is worse than the disease,
bringing high vulnerability to external shocks such as currency exchange
fluctuation.
It is noteworthy that, even among supposedly “homogeneous” economies
like the ASEAN-5 countries, differences in consumption ratio, and hence
in savings behavior, prevail. Culture may play a role, but social protection
Figure 9.4a: Household Final Consumption Expenditure and General
Government Expenditure in 2007
% of GDP
Figure 9.4b: Household Consumption, 1990–2007

% of GDP
ASEAN = Association of Southeast Asian Nations, GDP = gross domestic product, PRC = People’s Republic of
China.
Source: World Bank’s World Development Indicators Database, downloaded 1 November 2009.
mechanisms may also be a factor (Box 9.1). Table 9.7 (in the next section),
for example, shows that the two countries relying on defined contribution
schemes, Singapore and Malaysia, have higher savings rates than countries
operating under defined benefit systems.
If aggregate consumption in most ASEAN-5 countries has little room for
growth, can investments help as an alternative source of growth? Figure 9.5
shows how aggregate investments in ASEAN dwindled to an average of slightly
Box 9.1: The Employees Provident Fund (EPF) and Malaysia’s

Household Savings and Consumption Behavior
The share of Malaysia’s household consumption in GDP is below those of their

ASEAN-5 counterparts. This suggests that, of the ASEAN countries, Malaysia has
the most room for increasing household consumption. However, a closer look at
its savings component shows that this may be easier said than done.
One possible factor for Malaysia’s low consumption rate is the high mandatory savings
imposed on the employed labor force through the EPF. The substantial coverage and
contribution rate of the EPF distinguishes Malaysia from other ASEAN-5 countries.
In 2006, the EPF covered 11.36 million members (of which 5.39 million are active
members constituting 52% of the labor force) with a high contribution rate of 23%
of wages (11% from employees and 12% from employers). This contribution rate is
high compared, for instance, to the 3% employee contribution in the Philippines. EPF
assets have grown more than tenfold from RM24.55 million in 1985 to RM290.3 million
in 2006. Such growth in accumulated assets has not only reflected high mandated
household saving but has also enabled the EPF to be a significant funding source,
especially for Malaysia’s public sector.
This extensive mandatory saving scheme seems to have crowded out private voluntary
savings. An empirical study by Ang and Sen (2009) found that in Malaysia, private
saving (voluntary saving in the private sector, which is the sum of household and
corporate savings excluding EPF contributions) appears to be discouraged by
compulsory saving in the form of provident and pension funds. The Central Bank
of Malaysia stated that household savings through the EPF consistently accounted
for about 4–5% of gross national product from 1981 to 1990. On the other hand, the
residual “voluntary” part of private sector savings was on a declining trend. This could
be an indication that Malaysian households will be hard put to increase consumption
significantly unless the mandatory savings through the EPF is reduced.
Source: Plan International (2009).

over 20% of GDP, down from close to 35% before the Asian financial crisis. While
there might be room for investment demand to grow, the policies required to
increase it are related less to social protection policies, and have more to do
with credit availability. After the Asian crisis, most businesses have become
more cautious with their investments and financial institutions more careful
in their lending policies. We leave the discussion of these policies for another
paper and proceed to evaluate the region’s social protection policies.
3. Social Protection in ASEAN
3.1 Existing Social Protection in ASEAN

Social protection takes many forms. Perhaps the best known are the pension
and other social insurance schemes, as well as safety nets or social assistance.
Referring to ADB’s social protection framework, other forms of social protection
Figure 9.5: ASEAN-5 Investment, Saving, and Current Account Balance

Source: World Bank’s World Development Indicator database, accessed 1 November 2009
can be categorized into labor market programs, micro- and area-based schemes,
and child protection programs. Table 9.2 provides some examples of social
protection programs that various countries have adopted.
ASEAN countries have many different social protection policy programs.
What are widely common are disaster management schemes, old age, disability,
and death insurance, as well as work injury, but different social assistance
programs exist in the region as well. Some countries like Thailand, Philippines,
and Myanmar have comprehensive medical coverage, while others address
health care in other ways. For example, Singapore allows borrowing from the
EPF for medical purposes. Appendix 2 provides more examples of projects
or initiatives related to social protection in selected ASEAN countries. The
projects serve as either independent ad hoc initiatives or part of more structured
long-term social protection provision. The list of programs shows how social
protection provisions in ASEAN-5 are varied based on individual country
situations. Table 9.3 summarizes available social protection provisions in
ASEAN countries in different categories using the framework of the ADB
(excluding labor market and child protection).
Table 9.2: Forms of Social Protection
Form of Social
Protection Intended Beneficiaries Examples of Policy Instrument
Active labor force, including new Pre-employment training, employment
entrants, laid-off workers of state- skill upgrading, public works, job
Labor market programs owned enterprises brokerage
Workers and their dependents facing First, second, and third pillar pensions;
loss of income, low-income laborers, unemployment benefits; health
laborers with no access to proper health insurance; sickness and disability
Social insurance care benefits
Most vulnerable groups (disabled
elderly), poorest persons, those unable Cash allowances, food subsidies,
to participate in labor markets, those domestic shelters, health subsidies,
who have multiple disadvantages, upgrading , public housing, community-
are affected by crises, or are socially based social services, institutionalized
Social assistance excluded care
Micro- and area-based Those employed in the formal sector, Micro insurance, small-farmer
schemes rural and urban communities agricultural insurance, social funds
Immunization, school fee waiving,
Infants and children, pregnant women, health fee waiving, programs to combat
children involved in work, soon-to-be child labor, programs for children with
Child protection new entrants into labor market disabilities
Source: Ortiz (2001).

Table 9.3: Existing Social Protection Schemes in ASEAN
Social
Country Assistance Social Insurance Micro- and Area-Based Schemes
1 2 3 4 5 6 7 8
Cambodia s s
Indonesia s s s s
Lao PDR s s
Malaysia s s s s
Myanmar s s s s
Philippines s s s s s s s s
Thailand s s s s s s s s s
Viet Nam s s s s s
ASEAN = Association of Southeast Asian Nations, Lao PDR = Lao People’s Democratic Republic.
Note: 1) Old age, disability, death insurance; 2) sickness, maternity insurance; 3) medical care; 4) work injury;
5) micro insurance; 6) agriculture insurance; 7) disaster management; 8) social fund.
Source: Soeharto (2007).
Despite the programs and initiatives taken by ASEAN governments for

social protection, the existing programs leave much to be desired. ADB’s
Social Protection Index (SPI) reflects that most ASEAN countries need to
catch up with other Asian economies. The SPI is an index that systematically
and consistently quantifies national-level social protection activities in Asian
and Pacific countries.3
The SPI values, however, come with a caveat. Two countries may have similar
SPI values but opposite priorities in social protection; the index may reflect high
coverage and good targeting of beneficiaries but low total social protection
expenditures, or vice versa.4 Nevertheless, the SPI provides a simple way to
evaluate social protection provision. Based on the index, a simple ranking
shows Malaysia, Indonesia, Viet Nam, and Philippines are situated far below
Japan and Korea and slightly below India and the PRC (Figure 9.6). The figure
3
The SPI covers the five areas of ADB’s Social Protection Strategy: labor market, social insurance, social
assistance, micro- and area-based schemes, and child protection. However, it focuses more on policies
and programs that enable vulnerable groups to prevent, reduce, and/or cope with risks, and that (i)
are targeted at the vulnerable groups; (ii) involve cash or in-kind transfers; and (iii) are not activities
usually associated with other sectors such as rural development, basic infrastructure, health, and
education (see Baulch, Wood, and Weber 2006). Subsequent discussions of SPI in this section draw
heavily from ADB (2008).
4
ADB (2008) provides further illustrations of when misinterpretation of SPI values can occur.
Figure 9.6: Social Protection Index of Selected Asian Countries
Source: ADB (2008).
also indicates that ASEAN countries lag behind other developing countries in
South and Central Asia (e.g., Mongolia, Uzbekistan, and Sri Lanka).
Factors that cause ASEAN countries to sit at the bottom of the selected
countries include the following. Indonesia and the Philippines have low coverage
and poor targeting of beneficiaries (the extent to which social protection
programs reach the poor). In the Philippines, only 30% of the poor receive some
form of social protection. Viet Nam and Malaysia perform better compared
to the rest of the ASEAN countries. Viet Nam is the highest-ranking country
in ASEAN in terms of SPI; 38% of its target population receives some social
protection, and 56% of its poor receive some social protection or other social
protection benefits. But Viet Nam’s performance is far below Korea, where
77% of the target population receives some social protection and 100% of the
poor receive some social protection or other benefits.
Focusing on social protection expenditure (from both public and private
sources) as a percentage of GDP, Figure 9.7 shows that ASEAN countries
provide comparably fewer resources for social protection. Japan, with social
protection expenditure of 16% of GDP, sits at the top of the ranking. Countries
like Sri Lanka, Mongolia, and PRC have also considerably higher figures for
social protection expenditure compared to ASEAN countries.5
5
These countries have previously had (and the PRC still has) centralized and totalitarian governments.
This largely explains the relatively high share of social spending as a percentage of GDP.
Figure 9.7: Overall Social Protection Expenditure
Source: ADB (2008).
Considerable variation among ASEAN countries can be observed. Overall

social protection expenditures in the Philippines and Indonesia account for
a meager 1.9–2.2% of GDP, while Malaysia and Viet Nam spend nearly twice
that much. ASEAN countries also differ in terms of the distribution of their
overall social protection expenditures (Figure 9.8). Though social insurance
is predictably the biggest component, its magnitude varies across the region.
Viet Nam and Malaysia, two ASEAN countries with higher overall social
protection expenditures at above 4% of GDP, show a significant difference
in how the expenditure is allocated. In Malaysia, social insurance makes up
90% of the expenditure, while in Viet Nam, it is responsible for only for 50%
of the total. A relatively large portion of the expenditure in Viet Nam goes to
social assistance—a significant contrast with Malaysia. It is also interesting
to observe that the labor market component of social protection is almost
missing or is insignificant as a share of overall social protection expenditure in
Indonesia, Malaysia, and Philippines, while PRC, India, Korea, and Viet Nam
seem to provide more attention to labor market policies.
A closer examination of ADB’s SPI components (Figure 9.9) shows that
social protection expenditures in the four selected ASEAN countries received
between 12% and 26% of GDP, as reflected in the social protection expenditure
variable, compared to about 30% on average for East Asia.6 The number of
6
This refers to the scaled value of social protection expenditures relative to Japan’s.
Figure 9.8: Distribution of Overall Social Protection Expenditure
Source: ADB (2008).
poor beneficiaries as a percentage of total beneficiaries in the region appears

to be fairly good, based on the high value of the poverty targeting variable
(SPDIST). But the low social protection impact (SPIMP) values of ASEAN
countries of below 0.2 indicate that, in terms of value or benefits, the social
protection that the poor have received may be inadequate. Similarly, the low
value of the social protection coverage (SPCOV) variable implies that the
social protection programs in the region may not be reaching all those who
need them. This may be explained by the high proportion of expenditures
on social insurance (more than 50% of social protection expenditures), yet
these programs cover only the formal sector. In terms of reaching the targeted
beneficiaries (the poor) as reflected by the poverty targeting (SPDIST) variable,
Viet Nam and Indonesia scored 71%, Malaysia came in second with 56%, and
the Philippines needs to catch up—currently its social protection provisions
are enjoyed by only 30% of the targeted poor beneficiaries (ADB 2008). Using
Japan as the benchmark for social protection, the four ASEAN countries have
much room to improve.
Figure 9.9: Social Protection Index Data for Selected ASEAN Countries
SPCOV = social protection coverage, SPDIST = social protection distribution (poverty targeting), SPEXP = social
protection expenditure, SPIMP = social protection impact.
Note: East Asia includes Cambodia, People’s Republic of China, Indonesia, Japan, Korea, Lao People’s
Democratic Republic, Malaysia, Philippines, and Viet Nam.
Source: Authors’ compilation from ADB (2008).
3.2 Social Security System and Reform Challenges
Of the different social protection programs, social insurance or the social security
system is perhaps the most complex, but is also the one that receives the most
resources. Because of the importance of social insurance, this subsection focuses
on social insurance systems in ASEAN.
3.2.1 Institutions and Programs

Social security systems in ASEAN can be classified into two groups: provident
fund systems (Singapore, Brunei Darussalam, Malaysia, Indonesia) and social
security-type systems (Philippines, Thailand, Viet Nam) that are either unfunded
or partially funded on a pay-as-you-go basis. Except for Singapore and Brunei
Darussalam, ASEAN countries have separate social security arrangements
for government employees and for private sector workers. Provident funds
typically work on a defined contribution basis while the non-provident schemes
are on a defined benefit basis.6 Table 9.4 summarizes the programs and the
different institutions within ASEAN.
In all countries, the social security system covers not only old-age pension
but also other insurance for contingencies like medical expenditures, accidents,
maternity, or unemployment. Not all countries cover the same contingencies,
but certainly, the social security system in ASEAN is not exclusively for
retirement purposes.
Coverage of social security also varies in each country. In Indonesia and
Thailand, social security contribution is compulsory only for larger enterprises,
while small businesses are exempt. Malaysia and the Philippines have extended
coverage (on a voluntary basis) to particular segments of the population such
as overseas workers and non-working spouses (in the case of the Philippines)
or self-employed and domestic helpers (in Malaysia).
3.2.2 Coverage of the Labor Force

Except for Singapore, social security coverage in ASEAN is relatively low
(Table 9.5). This is especially so in Indonesia and Viet Nam; these two countries
also have the lowest percentage of wage earners in the total working population,
while a large portion of the population works in the informal sector, which is
not covered by any form of social security. In Viet Nam, around 60% of the
working population is employed in agriculture, which is largely considered
under the informal sector and, thus, not eligible for the social security scheme.
Malaysia, on the other hand, has more than 76% of its working population in
the formal sector because a large percentage of Malaysia’s workers work in non-
agricultural sectors. This labor structure helps the extension of social security
coverage in Malaysia, with 50% of the labor force covered by the EPF or Social
6
Defined benefit is a scheme where the promised benefits are delivered regardless of the accumulated
social tax in their account, but also under certain minimum conditions (e.g., number of years of
contribution). Under defined contribution, on the other hand, benefits are based on the individual’s
social security contributions.
Table 9.4: ASEAN Social Security Profiles
Country Organization Benefits Type of Coverage

Scheme
Indonesia TASPEN (1963) Lump sum payment: Provident Government civil servants
disability , death, retirement, fund/ defined
separation contribution
Pension (1981): retirement, Defined

survivorship, disability benefits
JAMSOSTEK Health care benefits, work Provident Compulsory for enterprises with 10 or
(1977) accident and death insurance fund/ defined more employees or payroll of at least
contribution Rp1 million per month
Malaysia Employees Retirement benefits, Defined Compulsory for private employees

Provident Fund pre-retirement benefits, contribution and non-pensionable public sector
(1951) death, disability, members’ employees
investment program
Voluntary for self-employed, domestic
helpers, foreign employees, pensionable
public sector
Social Security Employment injury, Defined Compulsory for all employers employing
Organization occupational diseases, benefits one or more worker
(1971) disability, death
Compulsory for workers who earn no
more than RM2,000 per month
Philippines Social Security Pension and insurance Defined Compulsory for private employees and
System (1954) covering sickness, maternity, benefits self-employed below 61 yrs old
disability, retirement, death,
medical care, accident; allows Voluntary for overseas workers,
borrowing privilege non-working spouses of members,
employees under foreign government;
separated members from employment
Government Pension and insurance: Defined All public employees
Service Insurance retirement, disability and benefits
System (1937) death, unemployment, Voluntary for former public employees
sickness, loan windows, that have resigned or retired early from
optional life insurance government service
Singapore Central Provident Retirement, health care, Provident All private and public employees and
Fund (1953) homeownership, family fund/ defined self-employed
protection, asset enhancement contributions
Thailand Social Security Social security (medical, Defined Compulsory for large private enterprises
Office (1990) sickness, maternity, disability, benefits (10 or more employees)
death); survivorship, disability,
funeral
Pension (added in 1999)
Viet Nam Social Security Pension and death benefits, Defined Mandatory for all enterprises (private
Organization sickness, maternity, and benefits and state-owned) Voluntary for small
(1995) occupational accidents/ enterprises (less than 10 employees)
diseases
ASEAN = Association of Southeast Asian Nations, PT Taspen = PT Tabungan Asuransi Pensiun, Jamsostek = PT
Jaminan Sosial Tenaga Kerja.
Source: Authors’ compilation. See Appendix 1 for more details.
Table 9.5: Coverage Rate of Social Security Pensions
Percentage of Coverage Rate

Non-Agricultural Percentage of
Workers in Wage Earners Percentage of Percentage
Total Working in Total Working Population Aged Percentage of of Urban
Population Population 15–64 Labor Force Population
Cambodia 29.6 12.9 — — 20
People’s Rep.
of China 53.1 — 17.2 20.5 40
India 33.2 — 5.7 9.1 29
Korea 91.9 66.4 — — 81
Mongolia 60.1 39.3 23.6 39.6 57
Nepal 32.7 — 2.5 — 16
Sri Lanka 59.6 58.2 22.2 35.6 15
Lao PDR 20.3 10.4 5.7 7.2 27
Singapore 99.6 86 45.2 76.2 100
Malaysia 85.2 76.2 32.2 49.6 68
Indonesia 56 32.8 11.1 15.5 48
Philippines 63 50.4 18.7 27.1 63
Thailand 57.3 43.8 24.4 30.4 32
Viet Nam 42.1 25.6 12.9 16.2 26
Lao PDR = Lao People’s Democratic Republic.

Source: Yamabana (2009).
Security Organization. In Thailand, where 44% of the workers are in the formal
sector, about one third of the labor force is covered by the Social Security Office.
In the Philippines, 27% of the labor force is covered by the Social Security
System or Government Service Insurance System. In short, existing social
insurance coverage in ASEAN is limited to a fraction of wage earners. While
social security organizations have implemented programs to extend coverage
to informal sector workers, overseas employees, and unemployed spouses,
overall coverage remains small compared to that in developed countries.
Another policy concern is the quality of coverage, for example, whether
the social security benefits allow retired persons to live with relative security
and afford good quality medical care. In the case of Indonesia, a background
study on social security systems in the region commissioned by ADB (2007)
reported that the social health insurance programs provide poor quality
services that force workers to pay out of pocket outside the scheme and
encourage private sector workers to purchase private health insurance on
top of the social health insurance.
3.2.3 Contribution Arrangements

Since adequate financing will be one of the main obstacles to improving the
coverage of social security programs, ASEAN countries will have to re-examine
their existing contribution arrangements between employee, employer, and
government. Table 9.6 shows a variety of contribution arrangements for six
ASEAN countries in 2006.
Table 9.6: Contribution Rates for Social Security Schemes in ASEAN

Countries (% of wages), 2006
Family
Coverage Sickness Maternity Old Age Disability Work Injury Unemployment
Allowances
Government
Government
Government
Government
Government
Government
Government
Employee
Employee
Employee
Employee
Employee
Employee
Employee
Employer
Employer
Employer
Employer
Employer
Employer
Employer
Financing
Source
Indonesia * 3* — — s — 2 3.7 o o — — — — — —
Malaysia 11** 12** — — — 11 12 o o — — — 1.25 — — —
Philippines o o o o 3.3 6.1 o o o — — — 1 — — —
Thailand 1.06 1.06 1.6 s s s 3 3 1 s s s o o o 0.6 0.5 0.5 0.25
Viet Nam 5 s 5 10 o o — — — s — — —
Singapore 20 16 o o — — — — — —
Notes:
* 2003 data.
** 1999 data.
— Does not apply
No contribution
Discretionary / irregular contribution
o Discretionary / irregular contribution under old-age
# Regular / fixed contribution
o Regular / fixed contribution under old-age
s Regular / fixed contribution under sickness
Whole cost
ASEAN = Association of Southeast Asian Nations.
Source: Authors’ compilation based on International Labour Organization’s Social Security Expenditure Database.
Except for Thailand, the overall picture indicates that government

relies heavily on the contributions of employee and employer in virtually
all benefit categories. The government, however, has contingent liabilities,
particularly in the case of defined benefit arrangements and in view of the
decline in population growth that would limit the number of contributors
in the future. Thailand’s social security members seemed to be covered by
regular or fixed government contributions in all benefit categories except
for work injury.
Allocation of the contributions may reflect another issue. Much of the
financial resources of existing social insurance programs in ASEAN countries
are directed toward old-age benefits, as reflected in Table 9.6. A further study to
assess if such allocation is optimal, including how efficient the fund is invested
domestically or abroad, is particularly important given limited financial
resources in most ASEAN countries.
In terms of the old-age benefit category, employee and employer in Singapore
are, respectively, obliged to contribute 20% and 16% of wages for retirement
savings. This is by far the highest in ASEAN countries, followed by Malaysia
with 11% for employee and 12% for employer. This high mandatory savings
tends to be a key feature of provident fund schemes, which, in turn, is reflected
in generally high savings in the economy (Table 9.7). Yet, even with this high
Table 9.7: Savings and Social Security Schemes
Pension
Savings*/GDP Pension scheme Contribution Coverage
Country (average 2000–2009) (DB or DC) (% of wages)*** (% of labor force)
Indonesia 24.71 DB 2 15.5

Malaysia 35.22 DC 11 49.6
Thailand 29.92 DB 3.33 30.4
Philippines 28.12 DB 3 27.1
Singapore 43.87** DC 20 76.2
Viet Nam 32.94 DB 5 16.2
DB = defined benefits, DC = defined contribution, GDP = gross domestic product.

Notes:
*Gross national saving.
** Average excludes missing data for 2001, 2002, and 2003.
***May include contribution for other benefits such as sickness and maternity (one package).
Sources: ADB’s Statistical Database System, International Labour Organization’s Social Security Expenditure
Database, and Yamabana (2009).
contribution, there are doubts that the old-age benefits will be sufficient to
prevent poverty among the old, considering longer life expectancies.
3.2.4 Need for Reform in Social Security Arrangement

While the current global financial crisis has provided some impetus for
efforts to strengthen ASEAN’s social protection systems, other pressures
have also been helping push for changes, particularly in social security
arrangements in the region. These pressures include demographic trends as
well as poverty, urbanization, migration, income inequality, and changes in
family structure.
ASEAN still has a higher proportion of young population to total population
compared to developed countries (Figure 9.10). But with declining fertility rates
Figure 9.10: Population Structure, Dependency Ratios, and Others

Notes:
a. Average of 2005, 2006, and 2007 data.
b. Average of 2006, 2007, and 2008 data. Those aged <15 and 65+ in proportion to those 15–64 years old.
c. Average of 2006, 2007, and 2008 data. Those aged 65+ in proportion to those 15–64 years old.
d. Average of 2003, 2004, and 2005 data.
e. Average of 2007, 2008, and 2009 data.
Source: Authors’ calculations based on World Bank’s World Development Indicators Database, downloaded 9
October 2009.
and population growth, the financial sustainability and funding of the social
security system have increasingly become a concern. The concern is that there
would be fewer workers to contribute and support an increasingly large old
population because of longer life expectancies. Today, the old-age dependency
ratio is still relatively low, but with longer life expectancies, it is not inconceivable
that the ASEAN old-age dependency ratio might reach the same rate as in
Germany (30%) or Japan (32%). In just a decade, the growth of the dependency
ratio in ASEAN has increased by about one percentage point (Figure 9.11).
Other trends such as urbanization and industrialization have also caused
profound cultural changes that affect family finances. Before, people’s social
security depended on strong family and neighborhood support systems, but
this traditional support mechanism has frayed over the years. If decreasing
informal social security is not answered by increasing coverage and quality
of the formal social security system, old-age poverty is likely going to be a
looming challenge.
Meanwhile, existing poverty, income inequality, lack of access to basic
needs, and other pressing social problems of ASEAN-5 countries remain.
Short-term solutions for these would include redistribution through social
assistance programs and other social policies (e.g., access to health care and
education).
Figure 9.11: ASEAN-5’s Increasing Old-Age Dependency Ratio

% of Working-Age Population
ASEAN = Association of Southeast Asian Nations, PRC = People’s Republic of China.

Source: International Labor Organization Social Expenditure Database, downloaded 1 November 2009.
3.3 Summary
To summarize, we have discussed the social protection schemes in ASEAN-5

and found that, based on an index of social protection, ASEAN-5 lags behind
many other Asian countries. Analysis of specific components of the SPI also
indicates relatively few resources allocated to social protection, and low
coverage, with a significant portion of the population outside the social
security net. We then focused on a specific aspect of social protection—social
security—and assessed the existing arrangements and the building pressures
for social security reform. While there are “natural” pressures such as the aging
population and the specter of old-age poverty if the social security system
does not become sustainable, the global financial crisis has added another
impetus for changes in social protection arrangements, namely improved
social protection as a way to boost consumer confidence and domestic demand.
Social protection’s role in increasing demand implies use of social protection
reform as a possible element of “countercyclical” policy responses to the crisis,
but it puts less emphasis on the other objectives of social protection, namely,
poverty relief, income redistribution, and consumption smoothing.
Our discussion of savings in ASEAN should discourage the use of social
protection merely as a countercyclical tool because, unlike the PRC, savings
rates in ASEAN are not actually all that high. Household savings, in particular,
are low in some countries like the Philippines. Moreover, compared to the
PRC’s low consumption ratio, consumption expenditures in ASEAN are already
relatively high. However, the need for social protection improvement remains
valid in view of social pressures, not only from demographic trends, but also
from poverty, inequality, and the generally low “initial condition” of social
protection provision in the region.
The next section further analyzes multifaceted considerations in thinking
about social protection reforms.
4. Economic Effects of Social Protection Spending

The policy advice to increase social protection spending has to be evaluated
in terms of its overall impact on the economy. It is important not to forget
the trade-offs that social policies bring to bear on the rest of the economy
because resources spent on the poor and vulnerable, for example, may be
siphoned off from other sectors. Choices must be made, for example, between
security and the vitality of the economy. Of course, the choice need not always
be “either-or” because some evidence indicates that social policies have
assisted economic growth (see, for example, Department for International
Development [DFID] 2006). But in the short term, the resource constraint
needs to be factored in.
In studying the effects of social protection spending on various aspects
of the economy, it is important to note that much of the literature has been
devoted to the effects of social insurance, pension systems in particular, and
less attention has been given to the other elements of social protection. There
are likewise some studies on the effects of social assistance programs like food
stamps and other welfare programs, but there are not many for other elements
of social protection such as micro- and area-based schemes, child protection,
or labor market policies (e.g., retraining). Thus, this section focuses more on
the effects of social insurance; where we have found some impacts of social
assistance, we will mention so explicitly.
4.1 Labor Supply

To the extent that social insurance is usually financed through contributions
or “payroll tax,” social insurance can give rise to labor supply distortion. For
retirement schemes, in particular, there is significant evidence of adverse
incentives in labor force participation, especially as contributors approach
the mandatory retirement age. Studies on the OECD economies in which an
early retirement option is possible and benefits (albeit partial) are available
show a pattern whereby older workers stop working earlier than the mandatory
retirement age. While in some countries, early retirement was part of the
pension design to create employment for new entrants in the labor force,
the scheme led to the adverse effect of people retiring earlier than is efficient
(Barr and Diamond 2009), thus effectively reducing the potential productive
capacity of the labor force. Yet, while the effective retirement age dropped over
the years, rich countries saw no corresponding reduction in unemployment
rates resulting from the early exit of “old” workers.
Economists explain the incentive effect of early retirement as being
due to the presence of leisure (along with consumption) in the individual’s
utility function, presenting workers with a trade-off between consuming
more (and continuing to work) and consuming less but also enjoying more
leisure. In addition, studies have shown that in some countries such as France
(see Blanchet and Pele 1999), the marginal change in an individual’s social
security wealth (the present discounted value of net social security benefits)
becomes negative as the worker nears retirement age, due to the continuing
increase in contribution (in present value terms) and future taxes. In this
case, continuation in the labor force actually implies a reduction in the
present discounted value of pension benefits. Early retirement therefore
becomes an optimal decision for the individual. For society as a whole, this
may not be the optimal outcome as the labor force’s potential productive
contribution is cut short. The right policy mix to keep people in the labor
force longer could include such solutions as limiting the minimum age for
early retirement, reducing social security contribution of those nearing
retirement, or increasing pension benefits.
For some social assistance programs like Aid to Families with Dependent
Children and food stamps, the evidence in the US shows that higher potential
benefits from the social assistance scheme induce greater participation in
these programs (Moffitt 1992). That is to say, beneficiaries of social assistance
programs do not “graduate” to become self-sufficient (i.e., enter the labor
force), but rather stay on as “permanent” recipients of what ought to have
been temporary assistance.7 While this finding is context-specific, it also shows
the need for good policy designs for social assistance programs so that the
incentive to work is not supplanted, even as the public tries to assist those
who are truly in need.
4.2 Private Transfers and Private Insurance

In countries where state-based (social security system) or market-based (formal
credit markets) economic support is limited, family transfers constitute a
major source of household income. Private transfers, however, were found to
have been crowded out by public social welfare expenditure. The implication
is that the government’s attempts to alter the distribution of economic well-
being can be thwarted by private behavioral responses (Cox, Hansen, and
Jimenez 2004). In other words, the government’s intention to increase or set
a minimum income for households in a certain population (income) group
through some form of public assistance is offset by the disappearance of private
transfers. These private transfers (from families, friends, and neighbors), which
7
The evidence also points to the impact of the social safety net on the family structure. In the case of
Aid to Families with Dependent Children, where part of the condition for aid eligibility is that the
household is headed by a single female, many beneficiary households remain single-female headed.
The decision to marry, therefore, is affected by the conditionality attached to the social assistance.
have supported the poor households, are pulled back as the government is
perceived to be “taking over” the care for their needs.
On private insurance, a similar crowding out effect has been observed
from the coverage expansion of Medicaid to include pregnant women and
children with higher incomes as well as households with two parents.8 The
reduction in private insurance coverage came from workers that dropped
their own insurance coverage, particularly the coverage of their dependents
(Cutler and Gruber 1996).
4.3 Savings, Consumption, and Demand

The question we are considering concerning savings and consumption is
whether reforms in social protection in ASEAN will cause consumption to
increase. The reforms, in the case of pension schemes, can take various forms
including reduced benefits, increased coverage, increased contribution, and
parametric changes in the accrual rates. Reform of pension schemes can also
mean introduction of fully-funded retirement programs (particularly, defined
contribution schemes), multi-tiered pension programs, or other combinations
of pension reform programs. Effects on savings and consumption will vary
depending on the type of reform. For example, reform that takes the form
of increased contribution rates (to improve the scheme’s sustainability) will
have a different impact from an effort to increase social security coverage of
informal sector workers. The former may have a negative short-run impact
on consumption due to reduction in disposable income, while the latter may
help increase consumption, particularly among informal sector households
due to the rise in expected social security wealth. For other types of social
protection, especially social assistance programs directed at the poor and those
with a high propensity to consume, the impact on consumption of increased
government intervention is unambiguously positive.
With regard to social security, the theoretical literature provides unclear
guidance on the potential impact of social insurance on aggregate savings. The
impact of pension9 on aggregate savings varies depending on the motivation
of individuals, the characteristics of the financial market, assumption of labor
supply, and other factors (Blake 2006). In a basic life-cycle model, for example,
8
Previously, Medicaid covered only very low-income women and children in single-parent families.
9
This may refer to existence of a pension system in a country, increase in benefit ratios, etc., depending
on the empirical study.
where the motivation is only consumption smoothing over one’s lifetime,
pension will not cause any increase in aggregate savings because of the 100%
displacement effect of private savings by pensions. But if the labor supply
decision is not fixed and workers can take early retirement, in response to
the relative importance of income and substitution effects, the increase in
private savings (to prepare for longer longevity risk) increases aggregate savings.
Similarly, when bequest or precautionary motives10 are added to the simple
consumption-smoothing purpose of savings, then pension can further result
in an increase in aggregate savings.11
The empirical literature on the relationship between pension and savings
is just as nebulous, describing a wide variety of positive and negative effects
of pension on savings. However, when the expected change in pension wealth
is large, a negative relationship with savings is more perceptible (Kohl and
O’Brien 1998). There is a clear private sector savings offset of pension, but
because private savings is an imperfect substitute for public pension, the offset
is likewise imperfect. The result is a net increase in national savings of varying
magnitudes for different countries.
Cross-sectional empirical work on the pension–savings relationship
provides interesting insights on this degree of variation.12 For example,
from the literature review by Kohl and O’Brien (1998), we find the following
interesting relationships:
t 5IFSFMBUJWFMZXFMMPíBOECFUUFSFEVDBUFEIBWFTUSPOHFSTBWJOHSFEVDUJPOT
in response to public pension.
t -PXJODPNF IPVTFIPME TBWJOHT BSF OPU JNQBDUFE IJHIMZ CZ QFOTJPO
provision (because these households save little to start with), so the degree
of income inequality in the population has an effect on how much offset
there is in aggregate saving.
t %JíFSFOUHFOFSBUJPOTIBWFEJíFSFOUTBWJOHQSPQFOTJUJFTBUUIFTBNFQPJOU
in their life cycle (vintage effect). For instance, households that have
experienced major income shocks, e.g., World War II generation retirees,
tend to save more.
10
The bequest motive increases savings because, besides the consumption-smoothing objective, the
individual wants to have something left over to give to his children. The precautionary motive—i.e.,
saving in the face of uncertainties like death, income disruption, and health expenditures—likewise
contributes to the increase of savings.
11
For an accessible discussion of pension theory, see Blake (2006).
12
The downside of cross-sectional empirical results is that the differences across individuals at a given
time may change when an entire life cycle is considered. Hence, the results may not necessarily apply
to changes in wealth of the overall system across time (Kohl and O’Brien 1998).
t 0UIFSGBDUPSTUIBUNBUUFSJODMVEFBHF IJHIFSEJTQMBDFNFOUPGQSJWBUF
saving with age), size of household and number of children (bigger
households have smaller displacement), existence of occupational pensions
(displacement is higher for state employees and self-employed and lower
for managers), and urban residence (higher displacement).
t -JRVJEJUZDPOTUSBJOUTBíFDUUIFTJ[FPGUIFJNQBDUPGQFOTJPO FWFOUIPVHI
its sign is sometimes positive and other times negative.
A shift from a pay-as-you-go system to a fully-funded scheme, usually in
the form of tax-favored retirement accounts, may have a positive impact on
private savings depending on the design of the scheme. In particular, where
defined contribution schemes provide ease of access (e.g., payroll deduction
schemes) or the public is sufficiently aware of its benefits, these schemes have
met considerable success in increasing savings. For example, the success of the
US 401(k) in raising savings is attributed to these factors, rather than to the tax
incentive per se. Still, other results show that the specific effects of tax-favored
retirement accounts on savings are sensitive to aspects of the scheme design,
such as the ceiling on contributions, eligibility limitations, marginal tax rates,
and whether it is mandatory or voluntary (Kohl and O’Brien 1998). Mandatory
schemes stimulate saving by forcing even poor households to save.
The implication of these results for social insurance reform in ASEAN is
that its specific impact on private and aggregate savings will depend on the
specific design of the reform. For example, if any of these countries shift from
a pay-as-you-go system to defined contribution schemes that have significant
tax costs, national savings may remain unchanged despite many countries’
experience of a rise in private savings. Another implication is that any marginal
tweaking of the existing social security scheme may not result in a perceptible
change in savings. Empirical results using periods in which changes in pension
schemes were sufficiently gradual and in which public awareness was low
showed little effect on savings.
We conclude, then, that social insurance, especially through social security
schemes, is not an effective countercyclical policy because of its ambiguous
effects on savings and, thus, on consumption. Reforms in pension schemes
should be undertaken not in response to short-term macroeconomic policy
concerns but rather to fulfill other objectives of social security provision
like income redistribution, poverty relief, insurance, and consumption
smoothing.
Other types of social insurance, particularly unemployment insurance,
have an important function of maintaining consumption levels during
economic downturns. Along with active labor market policies, unemployment
insurance has a more direct and effective countercyclical effect than do pension
schemes.
Similarly, social assistance or safety net expenditure is different from
social insurance. Studies relating savings with other public transfers (not
old-age benefits) show a clear displacement of savings. This result suggests
the importance of the precautionary aspect of savings, which, in turn, implies
that increased public sector provision of some social services will likely lead
to a reduction in savings and an increase in consumption. Social assistance,
therefore, may be a useful countercyclical policy instrument, provided that its
fiscal costs do not significantly offset its positive impact on consumption.
4.4 Investment, Growth, and Welfare

In a pure life cycle model where the young have higher marginal propensity to
save than the old, pay-as-you-go schemes that redistribute income from young
to old result, in theory, in lower savings and lower steady state capital stock.
Furthermore, the labor supply distortion caused by the pension financing adds
to the lower growth path of an economy under a pay-as-you-go system.
Tran and Jung (2007) studied the impact of introducing pension reform,
through extension of coverage to the informal sector, on savings, labor supply,
output, and welfare using a model calibrated to a developing country condition.
We can consider such pension reform almost equivalent to social assistance,
especially if the informal sector beneficiaries do not contribute to the pension
fund. The results show the same saving behavior and labor supply distortion
as discussed above. Furthermore, the pension program crowds out capital
stock and reduces output. But the magnitude of output reduction depends
on the type of financing, that is, whether the pension reform is financed by
capital tax (greatest effect) or consumption tax (lowest effect).
On welfare, Tran and Jung (2007) found an unequal impact, with informal
sector households generally benefiting from pension reform and formal sector
households being worse off (because of higher tax). The greater the inequality
among households, the greater will be the net welfare effect of these reforms.
That is, the insurance and redistribution function of social assistance programs
(in this case, extension of security coverage to informal households) surpasses
the negative effects arising from labor and saving distortion.
This result shows the policy trade-offs in social safety nets or social assistance
programs. While the crowding out effect on capital stock leads to low growth,
the insurance role played by social security or its redistributive benefit may
dominate to result in a net positive welfare change (Imrohoroglu, Imrohoroglu
and Jones 1995; Tran and Jung 2007).
Is redistribution through social security or social assistance the only way
to address inequality? Are there other redistribution schemes that do not have
as much distortionary effect on the economy? Deininger and Squire (1998)
found that, while inequality has a long-term negative relationship with growth,
it is inequality in assets rather than inequality in income that has the greater
influence on growth.13 Hence, asset redistribution would be more effective
than income redistribution. But if asset redistribution would cause investments
to drop, the authors recommend accumulation of new assets rather than
redistribution of existing assets. For example, education and human capital
investments or microfinance are investment-oriented policies that are beneficial
to the poor while not being detrimental to long-term growth. In general, policies
that increase aggregate investment and facilitate acquisition of assets by the
poor are beneficial for both growth and poverty. Thus, redistribution through
social sector expenditures (such as education) seems to be more attractive, in
a growth and welfare sense, than other means of redistribution.

This chapter has considered the desirability of using reform of social protection
policies to help boost domestic demand in ASEAN as a contribution in global
growth rebalancing. The chapter finds that, based on the savings history
of the region, not all countries in ASEAN are in a position to boost private
consumption because (i) their savings ratios are not really that high compared
to other countries in East Asia, and (ii) private consumption is already a big
component in total domestic demand, so that further increase may render
these countries vulnerable to currency fluctuations. Careful consideration of
the conditions in individual countries and specific social protection programs,
given resource constraints, is necessary.
The chapter also considered the factors that are contributing to the pressures
for reform of the social protection schemes in East Asia. More than short-term
countercyclical policy considerations, factors such as the aging population,
13
One cause of the asset inequality link with growth is that the asset poor have difficulty making
economically profitable investments (e.g., in schooling) that affect long-term growth.
low population growth, urbanization, and the fraying of family ties that had
been the source of informal social security support in years past will motivate
changes in social protection schemes.
Finally, the chapter considered the multifaceted trade-offs that social
protection expenditures impose on an economy. These trade-offs include the
possible labor supply distortion, the crowding out of private (family) transfers
and private insurance, the uncertain impact on savings, and the impact on
capital stock, growth, and welfare. The design of any social protection reforms
must pay attention to these factors.
Social protection programs have grown in importance in the eyes of
policymakers. But considering limited resources and the possible adverse
incentive effects on labor supply decisions and on the economy as a whole,
the need for good program designs cannot be emphasized enough.
Appendix 1: Social Security Institutions in ASEAN

Indonesia established PT Taspen, a central provident fund, in 1963 to provide
government employees with a lump sum cash benefit at retirement age, death
benefits, and cash value before retirement. In 1981, it broadened the benefit
to include old-age pension, as well as survivorship and disability pension.
Indonesia has separate social security arrangements for the armed forces
through ASABRI (PT Asuransi Sosial Angatan Bersenjata Republik Indonesia).
Private sector employees contribute to a separate provident fund called
Jamsostek, which provides insurance for accidents as well as health care, old
age, and death. Indonesia also has comprehensive health insurance called
ASKES (PT Asuransi Kesehatan Indonesia) for government employees and their
families, which private sector employees can also join on a voluntary basis.
Malaysia established the EPF, a centralized provident fund like that in
Indonesia, in 1951 to provide retirement benefits for the private and non-
pensionable public sector. Membership is compulsory for private sector
employees and voluntary for self-employed, pensionable public sector
employees, foreign employees, and domestic helpers. Malaysia also has a
Social Security Organization to provide insurance for accidents, occupational
injuries or illness, disability, and death.
The Philippines has a Social Security System for private sector employees
and a Government Service Insurance System for public sector employees. A
separate scheme covers members of the armed forces. The Social Security
System is a defined benefit social insurance for sickness, maternity, old
age, disability, survivors’ pension, health, and work-related accidents and
illness. It is compulsory for private employees and self-employed persons,
and voluntary for overseas workers, employees under foreign governments,
temporarily unemployed persons, and non-working spouses of members. The
Government Service Insurance System is the counterpart of the Social Security
System for public sector employees, but allows voluntary membership for
public employees that have retired from public service but are below 65 years
old, or have resigned from government jobs.
Singapore has a Central Provident Fund, a social security savings scheme
established in 1953. It provides benefits for retirement, health care, home
ownership, family protection, and asset enhancement. The Central Provident
Fund excludes foreign workers from membership. Like Singapore, Brunei
Darussalam also has a provident fund, the Employees Trust Fund, which
caters to both public and private employees and self-employed workers.
Thailand’s Social Security Office manages three funds: the Social Security
Fund for health care benefits; the Workmen’s Compensation Fund for
employment injury; and the pension fund, established in 1999. Membership
in the Social Security Office is mandatory for private enterprises with 10 or
more employees. Thailand has separate social security benefits for government
employees.
Viet Nam’s Social Security Organization was established in 1995 to cover
retirement, survivorship, and medical care. It is mandatory for private and
public sector employees, except for those in small enterprises (less than 10
employees), which can join on a voluntary basis.
References
Agrawal, N. 2000. Employment Generation and Social Safety Net (Social Safety Net
Component). ASEM Trust Fund for the Asian Financial Crisis Implementation
Completion Report. Available at: http://siteresources.worldbank.org/ASEM/
Resources/TF1-ICMs/Vietnam/Employment+Generation+and+Social+Safety+Net
+Safety+Net+Component.pdf.
Ang, J., and K. Sen. 2009. Private Saving in India and Malaysia Compared: The Role of
Financial Liberalization and Expected Pension Benefits. Monash Economics
Working Papers 13/08, Monash University, Department of Economics. Available
at: http://ideas.repec.org/p/pra/mprapa/14413.html.
ASEAN Social Security Association. n.d. Good Practices of ASSA Member Institutions.
Available at: http://www.asean-ssa.org/GoodPractices.asp.
ADB. 2007. Financial Governance and Social Security Reform: Preparatory Studies on
National Social Security System in Indonesia. Manila: ADB.
————. 2008. Social Protection Index for Committed Poverty Reduction, Volume 2: Asia.
Manila: ADB. Available at: http://www.adb.org/Documents/Books/Social-
Protection/Volume2/default.asp.
————. n.d. Social Protection Project Documentations. Available at: http://www.adb.org/
SocialProtection/projects/assist.asp.
Barr, N., and P. Diamond. 2009. Reforming Pensions: Principles, Analytical Errors, and
Policy Directions. International Social Security Review 62(February): 5–29.
Baulch, R., J. Wood, and A. Weber. 2006. Developing a Social Protection Index for Asia.
Development Policy Review 23(1): 5–29.
Blake, D. 2006. Pension Economics. West Sussex, UK: John Wiley and Sons, Ltd.
Blanchet, D., and L. Pele. 1999. Social Security and Retirement in France. In Social Security
and Retirement around the World, edited by J. Gruber and D. Wise, 403–435.
Chicago, IL.: Chicago University Press.
Bolt, R., and M. Fujimura. 2002. Policy-Based Lending and Poverty Reduction: An
Overview of Processes, Assessment, and Options. Economic and Research
Department Working Paper 2. Manila: ADB. Available at: http://www.adb.org/
Documents/ERD/Working_Papers/wp002.pdf.
Cox, D., B. Hansen, and E. Jimenez. 2004. How Responsive Are Private Transfers to
Income? Evidence from a Laissez-faire Economy. Journal of Public Economics
88(9-10): 2193–2219.
Cutler, D., and J. Gruber. 1996. Does Public Insurance Crowd Out Private Insurance?
Quarterly Journal of Economics 111(2, May): 391–430.
Damrongplasit, K., and G. A. Melnick. 2009. Early Results from Thailand’s 30 Baht Health
Reform: Something to Smile About. Health Affairs 28(3): w457–w466. Available at:
http://content.healthaffairs.org/cgi/content/abstract/28/3/w457.
DFID. 2006. Social Protection and Economic Growth in Poor Countries. March. Available
at: http://www.gsdrc.org/docs/open/SP16.pdf.
Deininger, K., and L. Squire. 1998. New Ways of Looking at Old Issues: Inequality and
Growth. Journal of Development Economics 57(2): 259–287.
Imrohoroglu, A., S. Imrohoroglu, and D. Jones 1995. A Life Cycle Analysis of Social
Security. Economic Theory 6(1): 83–114.
Kohl, R., and P. O’Brien. 1998. The Macroeconomics of Ageing, Pensions, and Savings: A
Survey. OECD Working Paper 200. Paris: OECD.
Manasan, R. G. 2009. Reforming Social Protection Policy: Responding to the Global
Financial Crisis and Beyond. Philippines Institute for Development Studies
Discussion Paper Series 2009-22. Available at: http://dirp4.pids.gov.ph/ris/dps/
pidsdps0922.pdf.
Mohd, S. 2009. Social Protection in Malaysia. United Nations Economic and Social
Commission for Western Asia (ESCWA), Arab Forum on Social Policy, Beirut,
28–29 October 2009. Available at: http://css.escwa.org.lb/sdd/1035/SP_Malaysia_
ver4.pdf.
Moffitt, R. 1992. Incentive Effects of the US Welfare System: A Review. Journal of Economic
Literature 30(1, March): 1–61.
Ortiz, I. D. 2001. Social Protection in Asia and the Pacific. Manila: ADB.
Santoso, W., and A. K. Sarie. 2007. The Household Survey and Monetary Policy in
Indonesia. IFC Bulletin 26, July 2007. Available at: http://www.bis.org/ifc/publ/
ifcb26q.pdf.
Soeharto, E. 2007. Social Protection System in ASEAN: Social Policy in Comparative
Analysis. Presented at the 15th Symposium of the International Consortium
for Social Development, Hong Kong Polytechnic University, Hong Kong,
16–20 July 2007. Available at: http://www.policy.hu/suharto/Naskah%20PDF/
HongkongSocProASEAN.pdf.
Tran, C., and J. Jung. 2007. The Extension of Social Security Coverage in Developing
Countries. CAEPR Working Paper 2007-026. Available at: http://ssrn.com/
abstract=1029611.
UNDP Malaysia. n.d. UNDP, MECD and MCCM Launch Micro Credit Project to Empower
Rural Women with Entrepreneurial Skills. Available at: http://www.undp.org.my/
index.php?option=com_content&view=article&id=185.
Yamabana, H. 2009. Pension System in Asia and the Pacific: ILO Perspectives. 4th Pensions,
Benefits and Social Security Colloquium, Tokyo, 4–6 November 2009.
Appendix 2: Social Protection in ASEAN-5:

Selected Sample of Some Recent Social
Protection Programs
Examples of Projects or Initiatives Fully or Partially Containing
Indonesia Malaysia
Labor Market Community and Local Retraining program (2008)
Government Support Sector A special program was developed
Development Program (2001) under the Labor Department, Ministry
To mitigate the effects of the 1997–1998 of Human Resources, to retrain the
crisis, the government and ADB unemployed and to assist them with
supported this program, designed to new employment. The program helps to
help the poor, including newly laid-off reduce the effect of the global economic
workers in urban areas and associated crisis on the labor market. During this
labor-supplying rural areas, through training, the trainees receive a monthly
the provision of job opportunities and allowance to assist them with their
improvement of basic infrastructure. expenses. A similar program in 1997
It aided about 460,000 households targeted fresh graduates, providing
(equal to 2.3 million people), mostly them with temporary jobs at government
poor, who obtained employment during agencies for a few months with a
construction, providing them with much- minimum monthly allowance in order
needed help in the to ease the effect of the Asian financial
aftermath of the crisis. crisis.
Social Insurance Social Security Reform (2002) “Return-To-Work” program

The Indonesian government started the (2005)
initial phase of social security system The Malaysian Social Security
reform to increase its coverage and Organization introduced this program
extension of the scheme. New legislation with the objective of assisting insured
and a dedicated agency were prepared. persons with employment injuries as well
as those claiming for disability pension
to be able to return to work through a
multidisciplinary approach. This program
aims to achieve a competitive and
healthy workforce, increase productivity
at the workplace, and improve Malaysia’s
economy by reducing down time.
Social Protection Programs

Philippines Thailand Viet Nam
Pangulong Gloria Scholarship Rural area employment Economic slowdown safety net
(2006) program program (1999)
This scholarship is designed to provide Thailand’s government, with funding Following the Asian crisis during
skills and competencies to job seekers from the World Bank, implemented 1997–1998, which had a severe
through appropriate training programs employment generating programs impact on employment generation,
that are directly linked to existing jobs in the rural areas, where many Viet Nam’s government, with
and immediate employment. Initially, the jobless were expected to return assistance from the Asia-Europe
program targeted a 90% employment from urban areas. With regards Meeting Trust Fund, conducted:
rate among recipients, but later placed to mitigating the effect of the (i) analysis of the labor market
the target rate at 50%. This scholarship global financial crisis, Thailand impact of the economic slowdown
accounted for 43% and 56% of the also announced the allocation of and SOE reform, (ii) design of
total national government spending on B10 billion to the Bank of Agriculture severance packages and training for
active labor market programs in 2007 and Agricultural Cooperatives for retrenchment, and (iii) mobilizing of
and 2008, respectively. The program is the Rural Labor Repatriation Fund. donor funding for social safety nets.
being implemented by the government Another B6.9 billion is allocated for
(TESDA) in partnership with private the unemployed, providing 1 month
sector organizations and various training of training and 3 months of wage
institutions. subsidy.
New initiatives in extending Thailand’s 30 Baht Health Developing social security

insurance coverage to overseas Reform (2001) system in Viet Nam (2003)
workers In 2001, Thailand extended Viet Nam’s government studied the
The Social Security System of the government-financed coverage to all feasibility and financial sustainability
Philippines has long been striving to uninsured people with little or no cost of expanding the coverage of
extend its coverage toward overseas sharing. Thailand has added nearly Viet Nam’s social security system.
Filipino workers. Its more recent 14 million people to the system and A pilot project was also launched
initiatives, which include relaxed achieved near-universal coverage to introduce electronic health
document requirements, establishment without compromising access for insurance cards to facilitate health
of foreign offices, granting of special those with prior coverage. insurance expansion in rural and poor
privileges to overseas workers, and communities.
formation of value chains, have increased
coverage from 54,948 overseas workers
in 1998 to 456,965 overseas workers in
2004. Collections from overseas Filipino
workers have also increased eightfold
between 1999 and 2004.
Appendix 2: Social Protection in ASEAN-5:

Selected Sample of Some Recent Social
Protection Programs (continued)
Examples of Projects or Initiatives Fully or Partially Containing

Indonesia Malaysia
Social Assistance Public health and nutrition (2003) Price subsidies for the poor
The Indonesian government, assisted As many poor families were facing
by ADB, developed urban nutrition increased food prices because much
management, focusing on children of their food is imported, the Malaysian
under five and pregnant women One of government has introduced subsidies for
the components aims to examine the staples such as cooking oil, flour, bread,
effectiveness of formal and non-formal and rice.
social safety nets to protect nutrition
status and food security.
Micro-Area Based Poor farmer income improvement Empowering Rural Women

project (2002) project (2005)
This project aimed to improve farmers’ The United Nations Development
income by enhancing their capacity to Programme, in partnership with
adopt innovative agricultural production the Malaysian government and
and marketing methods by better the Malay Chamber of Commerce
targeting village-level public investments Malaysia, launched a campaign for the
based on location-specific needs. Empowerment of Rural Women project
in the states of Kelantan, Terengganu,
Penang, and Melaka. The project, which
promotes “sustainable microfinance” and
a holistic approach to poverty reduction,
targets rural low-income communities, in
particular women, with businesses in the
food sector.
Child Protection Community-based basic Child abuse and neglect

education for the poor (2002) prevention (2006)
A project was launched to explore Under the leadership of the Ministry
ways to integrate schools for the of Women, Family and Community
poor in remote areas. One of the four Development, UNICEF formed strategic
components of the project is piloting partnerships with the Malaysian
community-based scholarship programs. Association for the Protection of Children
and the Royal Malaysian Police to
strengthen the professional framework
to prevent child abuse and neglect in
Malaysia.
ADB = Asian Development Bank, UNICEF = United Nations Children’s Fund.

Sources: ADB (n.d.), ASEAN Social Security Association (n.d.), Agrawal (2000), Bolt and Fujimura (2002),
Damrongplasit and Melnick (2009), Manasan (2009), Mohd (2009), United Nations Development Programme
(UNDP) Malaysia (n.d.), and Ortiz (2001).
Social Protection Programs

Philippines Thailand Viet Nam
Social protection for poor women Subsidies for the poor (2008) Pro-poor health policies (2003)
vendors in Mindanao cities In facing the economic downturn and Viet Nam’s government issued and
(2002) to provide safety nets for low-income started implementing Decision 139
The Philippines government, working people, Thailand’s government (Health Care Fund for the Poor),
with donors, introduced innovative introduced a program providing which will provide health subsidies for
interventions designed to assist poor subsidies for the poor for cooking low-income people.
women vendors in selected public gas, electricity costs, free water,
markets in Mindanao. It will also provide and free transportation on trains and
sustainable and gender-sensitive social buses. This temporary program is
safety nets for Mindanao’s poor vendors. subject to review based on improving
economic conditions.
Development of poor urban Crop insurance scheme for Sustainable livelihood and
communities (2003) farmers rural infrastructure in Central
A project was launched to reduce The Bank for Agriculture and Region (2004)
income poverty in urban areas Agricultural Cooperatives of Thailand A project was launched to reduce
through asset reform (land distribution conducted feasibility studies and rural poverty through (i) intensification
and microfinance). The Philippines drew up the operating guidelines and and diversification of agricultural
government, assisted by ADB, is procedures for a new crop insurance production and non-farm employment
establishing a conducive policy, scheme. The Thai parliament has and income-generating activities;
institutional, and regulatory environment passed the enabling law for this. and (ii) strengthening the capacity of
to meet the housing need of urban poor local communities, NGOs, and local
communities governments to implement broad-
based and participatory development
activities.
Philippines Against Child Child Protection Act and Early childhood development
Trafficking Campaign (2003) Committee (2003) for the poor (2003)
In 2003, the Anti-trafficking in Persons The Child Protection Act was passed The Viet Nam government, assisted
Act was introduced to reduce human in 2003 and the Child Protection by ADB, developed a safe and
trafficking including children. The Committee was introduced healthy environment for children
Philippines Against Child Trafficking subsequently. The Thai government, from poor and vulnerable households
Campaign is a nationwide drive to assisted by UNICEF, organized by (i) reviewing the current early
raise awareness on child trafficking multidisciplinary teams made up of child development, (ii) developing
in the country. The campaign aims to police, prosecution officials, social appropriate survey instruments,
encourage protection of children against and health workers, and other and (iii) increasing stakeholders’
trafficking. professionals to investigate and deal awareness.
with cases of abuse.
Comments
Hiroshi Yamabana
This chapter presents a good overview of the rates of macro savings and private
consumption and the social protection and social security schemes of ASEAN
members, especially of ASEAN-5 (Indonesia, Malaysia, Philippines, Thailand,
and Viet Nam). It also summarizes theories on the economic effects of social
protection spending.
I agree with the concluding remark that attention should be paid to potential
impacts social protection can have on a country’s economy (e.g., growth).
However, I would like to stress that the emphasis should be put on the primary
objective and the role of social protection. Too much emphasis should not be
put on economic impacts when these impacts are uncertain and cannot be
considered as substantial (in solid quantitative terms).
1. Savings and Consumption

It is interesting to see that household consumption shares in GDP differ
substantially from one country to another (Figure 9.4b). It may be difficult
to conclude why they are substantially different. Some of the differences
may be explained more by cultural practices, attitudes, and preferences of
people toward savings or consumption rather than the prevalence of other
mechanisms for unexpected contingencies, such as social protection, as seen
in the considerable overall high savings ratio of the PRC and the preference
toward savings of urban dwellers in that country despite their having more
comprehensive social insurance mechanisms than those in other Asian
developing countries.
It may be too hasty to conclude, only by looking at the global household
consumption share of GDP, that many ASEAN countries do not have much
room to increase consumption. Therefore, it may be interesting to investigate
it in more detail, country by country.
In the case of Malaysia, where the coverage of the EPF is fairly extensive
(around 50% of the labor force) and the contribution to the EPF is substantial (23%
of wages), it is interesting to see the role of the EPF in savings and consumption
in the overall economy of Malaysia. It is also interesting to see whether the
(partial) conversion of the saving scheme into a social insurance scheme could
affect the savings and consumption of the insured and their families.
In the case of the Philippines, as seen in Table 9.1, it may be difficult to
decrease savings and to increase consumption because household savings
are already so low. However, there could be policy room to rectify the high
proportion of corporate savings (Figure 9.3).
In addition to more analysis on country-specific situations, it may be also
interesting to see not only the macro consumption and saving patterns but also
more detailed analysis on the saving and consumption patterns of different
groups of the population. The saving and consumption patterns classified by
income and asset levels and the analysis of basic needs are of particular interest
because the redistribution of income and the provision of basic needs (e.g., in-kind
health care, education) through social security provisions may create additional
consumption on the part of low income earners and their families.
2. Social Protection in ASEAN

The chapter provides a good and in general balanced overview of the
social protection system in ASEAN. The description could be enriched by
emphasizing that:
(i) the existing programs and financing are mainly targeted at the formal
sector employees through social insurance programs (Figure 9.3 and
Table 9.5),
(ii) social protection and social security are not reaching those who most
need them (as shown by looking at some other indices of the Social
Protection Index), and
(iii) more resources should be allocated to social protection (Figure 9.2).
The analysis will be further strengthened by looking at the allocation
of resources by different branches of protection (e.g., old age, health care).
Especially in the case of provident fund countries such as Singapore and
Malaysia, it is interesting to see that the majority of the resources are allocated
to old-age provisions, in the form of savings, even more than is being done in
OECD countries, mainly in the form of social insurance pensions, including
Korea and Japan, where populations are aging much more rapidly.
As for the need to reform social security arrangements, too much emphasis
is put on the aging issue and the financial sustainability of social security in
the long run. The important and urgent issue of poverty reduction and the
medium- and long-term goal of achieving an equitable society with proper
redistribution should be also emphasized, not only for achieving social justice
and social cohesions, but also for achieving better and more stable economic
growth through building a healthier and more productive labor force.
As for the aging “crisis” for financing pensions, it is obvious that the
financing method (e.g., defined benefit and defined contribution) will not
solve the problem (whether through transfer or investment, the productive
segment of the population should pay). The ultimate solution is to gradually
establish longer years of work in the society or increase the retirement age.
3. Economic Effects of Social Protection Spending
This chapter is a summary of existing research, much of which was carried out
based on developed country experiences, except for Tran and Jung (2007). It is
important to learn from experience, especially “failures” of other countries,
so as not to repeat mistakes. But contexts may differ and there may be ways
to rectify existing problems rather than giving up the fundamental idea
of protecting people (e.g., the conversion from defined benefit to defined
contribution).
Even when trade-offs exist, without quantifying their scale and judging
whether they are substantial, it would be misguided to prioritize one policy
over another (e.g., defined contribution to defined benefit).
3.1 Labor Supply

It is important that social security provisions should not provide any major
wrong incentives for people not to work. Policies exist to achieve these objectives
(e.g., proper minimum age for retirement, actuarial reductions, increase of
pensions).
3.2 Crowding Out

The discussion of crowding out should relate more to the roles of social security
rather than the killing of the roles of family transfers or private insurance.
As for the basic role of social security, because universality is essential, the
system cannot rely on “voluntary” mechanisms such as family support or
private insurance in the long run. The social security mechanism should
gradually replace the roles that others have partially played. So long as social
consensus has been gradually built, people may feel happier to pay social
security contributions or taxes than to directly support their family members
or to rely on imperfect private insurance mechanisms.
3.3 Savings, Consumption, and Demand

The prime objective of pension policy is poverty alleviation and the guarantee
of adequate income in old-age through consumption smoothing. The saving
itself (or the impact of savings on the economy) should not be treated as the
prime objective.
It is justifiable to conclude that social security reforms should not be done
only as countercyclical policy. However, social insurance plays an important
role as an automatic stabilizer during economic downturns by helping maintain
consumption levels through pension benefits. Some other social insurance
benefits, such as unemployment insurance and active labor market policies,
play a more direct and effective role for countercyclical policies.
3.4 Investment, Growth, and Welfare

The degree of “distortion” on savings by pay-as-you-go schemes and hence
the impact on growth depends on the model and the parametric assumptions
on growth. The labor distortion argument through pay-as-you-go schemes is
a classical argument and it is true that distortions could happen in a poorly
designed scheme (e.g., early retirement incentives to be provided with a low
retirement age and/or a low or no actuarial reduction in the case of early
retirement). However, a well-designed scheme (e.g., retirement age gradually
and properly adjusted in line with the aging and the higher labor force
participation of the aged population, a proper actuarial reduction provided
to early retirement pension) could overcome these distortion problems. Bearing
in mind the important role of the insurance scheme in stabilizing the life and
the income pattern of the aged population, the distortion issues should not be
overemphasized without taking into account the principal role and the overall
effectiveness of the social insurance system. The final discussion and conclusion
require careful examination because the degree of the “negative” impacts on
growth is not shown in the chapter. Investing in the poor, in terms of health,
education, or providing business capital, is very important. In addition, we
should recognize that:
(i) the urgent issue of poverty alleviation cannot wait for long-term
investment to materialize; and
(ii) a portion of the population may not be helped through lending such
as microfinance, where there is an associated risk of business failure
and bankruptcy.
Therefore, it may be questionable to support social sector expenditures
solely for macro growth and welfare without paying equal attention to
redistribution.
References
Tran, C., and J. Jung. 2007. The Extension of Social Security Coverage in Developing
Countries. Center for Applied Economics and Research Working Paper 2007-026.
Available at: http://ssrn.com/abstract=1029611.
Panel Discussion 249
X
Panel Discussion
A panel discussion summarizing the day’s discussion capped the Annual

Conference. Masahiro Kawai, ADBI Dean, chaired the session and asked the five
panelists to provide their views on whether and how social protection can stimulate
demand on a permanent basis, and whether such programs are affordable by
Asian countries. He also inquired about sources of social programs financing
as well as other constraints for developing social policies in the region.
The panelists for the session were Davide Furceri, David Coady, Charles
Horioka, Ming Yan, and Hiroshi Yamabana. Each was charged with discussing
policy issues related to social protection development in Asia.
Davide Furceri:
“Impact of Social Protection Is Higher in Developing
Economies”
Most of the analysis so far on social protection has focused on OECD countries.
But thus far, we know that the effect of public funding (including social
protection) depends on: (i) the economic development of the country, and
(ii) the size of government expenditure. First, social protection could be more
effective in boosting demand in countries where social protection share is
relatively lower. Put differently, there is a higher effect or higher consumption
multiplier in some Asian economies such as Korea or the PRC where the level
of social protection is quite low. Secondly, this increase in social protection
can have a permanent effect on demand through an “accumulation” process.
For example, in Nordic countries, some social safety nets for people exiting
the labor market allow them to go back to school to study. This is particularly
important during the current crisis where we see a mass exit from the labor
force. It is clear that if you have a safety net, this can help people increase their
human capital, even though they temporarily leave the labor market, which
enhances medium-term growth.
In terms of financing, there are several ways to finance social protection
programs. Here we are mostly focused on increasing tax. But I want to stress
that it is preferable to cut expenditures in other components (of government
expenditures) rather than to simply increase tax.
The problem of excessive saving in the PRC in the context of global
imbalance will not, in large part, be solved by social protection. Of course,
the PRC needs to improve social protection policies, but the reduction of the
global imbalance is not the ultimate goal of social protection. If you want to
reduce global imbalance, issues such as increasing competition or liberalization
of some key markets in the PRC should be addressed.
David Coady:
“Emphasis Should Be on Program Design”
Do social expenditures increase consumption? If you take money from people
who are not spending it, or spend less of it, and give it to people who do spend
it, consumption will go up. We are talking about the elderly who consume
more out of additional income. We are talking about low-income people, who
also consume more out of additional income. The best way to make sure that
such expenditure increases consumption is to focus on program design. In
the context of the PRC where the starting point (of social protection) is very
low, it is clear there is room both for more assistance and for more insurance.
The PRC can run better and more programs. In other countries, perhaps, it
would be a case of “do what you are already doing, but better.” In summary,
the consumption impact is biggest when social protection is well designed.
However, we should be very careful about discussing social expenditure only
in the context of its effect on consumption. We have to keep track of why we
have social expenditure. In fact, the increase in consumption is just a useful
side effect in the short term.
For finance, over time, social insurance will have to be run on a contributory
basis in one way or another, because of demographic considerations. In the short
run in the PRC, social protection may be financed by running down surpluses,
but this cannot last forever. In the end you will have to move to a well-designed
social insurance system that has a more sustainable financing.
The motivation should be greater efficiency and equity. When we are
developing a social assistance program, maybe we should start promoting
such programs not in terms of just social equity impact but in terms of their
efficiency impact. The policy disruption caused by not having access to cost-
effective social assistance programs has a great effect on growth. Most of
the time, governments have a “stop–go” policy: when there is a shock to the
economy, the government needs to spend money to overcome the shock,
and then, often, it is back to square one. I think we should start promoting
good social assistance programs from the perspective of efficiency as much
as equity.
Charles Horioka:
“Increased Demand Is Not the Primary Goal of Social
Policies”
I think there should be basically two broad objectives of social policies. One
is to guarantee a minimum standard of living and get people out of poverty.
And the second is to protect people from unforeseen contingencies regardless
of their income—whether they are poor or wealthy.
I think both objectives of social policy will lead to an increase in demand,
perhaps as an unintended consequence. If you redistribute resources to poor
people who have a higher propensity to consume, of course aggregate spending
will increase. And if you protect people from risks, they will feel less need for
precautionary saving, and again they will consume more.
Policies of both types will lead to greater demand, but I do not think that
this should be the primary goal of social policies. The primary goals should
be to eliminate poverty and also to protect people from risk.
Ming Yan:
“Social Protection Is Not Only for the Poor”
I would like to advocate social protection policies because they do not just help
the poor. I am specifically speaking in the context of the PRC. Health-care
insurance, education, housing—all of these are needed even by the middle
class. I think the reason that the Chinese government in many ways failed to
provide social protection for the Chinese people during the economic reform
was because it had a one-dimensional goal of economic growth. It did not
concern itself with human and social well-being because its objective was
different.
The quality of policy design and implementation are very important.
Otherwise, even with good intentions, such as that of the PRC government
to begin developing social protection, it could become as problematic as
before. For example, without a good program design, the provision of social
insurance to migrant workers could cause more problems for both the migrant
workers and local authorities. Another example is that, recently, the PRC
has implemented a program called “sending electronics to rural areas,” with
the goal of boosting domestic demand. But many rural areas in the PRC
do not even have electricity. Once the people buy a television, refrigerator,
or washing machine with government subsidies, what are they going to do
with these electronic items besides using them to store things? Even in areas
where there is electricity, few can afford to pay for it. Thus, the program of
boosting demand through social protection with government subsidies can
sometimes be wrong-headed.
The other problem with boosting domestic demand is its unintended
consequences such as change of values (consumerism), as well as for the
environment. People are competing to buy large houses and spending millions
of dollars on cars; already, we are competing with the US on car ownership.
This year, the PRC produced 12 million cars. But what about the environment?
What about the air pollution that these cars have been causing? I think the PRC
has a rich cultural tradition that should be valued more, rather than simply
learning from the American lifestyle, which often harms the environment.
On the incentive issue, the concern is, once the government provides
social protection beyond a basic livelihood, how can you prevent people from
becoming passive or dependent on government provision?
Hiroshi Yamabana:
“More Room for Consumption Growth in Asia”
In many parts of Asia, there is still huge unmet global demand. It is not true that
there is no more room for consumption growth. Substantial parts of the population
in these countries have no money or cannot afford basic consumption demands.
Transferring resources in the forms of social assistance, basic health care, and basic
education from richer segments of the population with a lower consumption rate,
to poorer segments of the population with a higher consumption rate, is a short-
term solution and an immediate way to boost consumption demand. Long-term
social insurance systems with good governance could flourish once the economies
grow and more people join the labor force as wage earners. So the challenge is
how to manage strategically the climb up from lower to higher levels of economic
development and, in parallel, gradually replace tax-based social assistance with
social insurance like that available in many developed countries.
Masahiro Kawai:
Summary
The discussion of social protection and domestic demand growth was motivated
by the fact that the global payments imbalance needs to be corrected. It is in
the best interests of Asia to correct this imbalance through higher domestic
demand, because the external demand coming from the US and Europe is
expected to be weak over the medium-term. Asian countries which have been
heavily dependent on external demand from the US and Europe have to create
their own demand in Asia.
The PRC, in particular, can use this opportunity to boost domestic demand
by improving social sector protection, which it has to do anyway. So there
is a good synergy between the need to create demand in Asia, in particular
in the PRC, and the imperative to strengthen social sector protection. Of
course in rebalancing demand, other, complementary, measures will need to
be pursued, such as various types of structural reforms to further marketize
the Chinese economy and strengthen competition. Although the growth in
the PRC has been the fastest in the world, and is expected to continue for some
time, maintaining productivity growth and coping with the aging population
will be a challenge.

Effects of Social Policy On Domestic Demand

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Effects of Social Policy On Domestic Demand

Загружено:

Авторское право:

Доступные форматы

Effects of Social Policy

Annual Conference 2009

Edited by Masahiro Kawai and Gloria O. Pasadilla

©2010 Asian Development Bank Institute

Freely available electronically at:

III Social Protection in Developing Asia and the Pacific . . . . . . . . . . . . . . . . .

IV Household Savings Rates and Social Benefit Ratios:

V Influence of Age Structure on Savings and Social Spending . . . . . .

VI Household Savings and Social Protection Policies in the PRC . . . .

VII Consumption, Income Distribution, and

VIII Public Expenditures on Social Programs and Household

IX Social Policy Reforms and Growth Rebalancing in ASEAN . . . . . . . .

X Panel Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

In 2009, ADBI organized a series of conferences on the global ﬁnancial

Presenters and Discussants

Mukul G. Asher is a professor of applied public sector economics and economic

Davide Furceri is an economist in the Macroeconomic Analysis Division of the

Charles Yuji Horioka is a professor of economics at the Institute of Social and

Yasuyuki Sawada is an associate professor in the faculty of economics at the University

Akiko Terada-Hagiwara is an economist in the Macroeconomics and Finance

Yuqing Xing is a professor of economics at the National Graduate Institute for

Hiroshi Yamabana is a social security actuary in the International Labour

ADB Asian Development Bank

Masahiro Kawai and Gloria O. Pasadilla

Social policies have important developmental goals that go beyond merely

2. Macroeconomics of Social Policy and Global

Figure 1.1: Global Imbalances (current account balance in % of world GDP)

domestic demand adjustment). In discussing social policies’ effects on domestic

2.1 Understanding Household Under-Consumption in the PRC

Figure 1.2: Household Consumption (% of GDP)

Source: World Bank (2010).

Figure 1.3: Household Final Consumption Expenditure (% of GDP)

Source: World Bank (2010).

Figure 1.4: Why Do Consumers in the PRC Save?

Source: McKinsey Global Institute (2009).

2.1.2 Savings Structure: Households vs. SOEs Savings

Source: Calculations based on CEIC Data Company, Ltd.

2.1.3 Household Income

become insigniﬁcant since the Chinese modernization in the 1980s, while

2.2 Impact of Social Policy on Consumption

Figure 1.6: Household Consumption and Labor Income Shares (% of GDP)

GDP = gross domestic product.

Table 1.1: Impact of Government Social Expenditures (% GDP)

Income Impact of 1% Increase in Government Expenditures in:

Source: Based on Baldacci et al.(Chapter 8).

3. Microeconomics of Social Protection

Note: Saving = Net Income – Consumption Expenditure.

coverage of social security or health insurance by the government may reduce

demographic factors. Put differently, why do Asia-Paciﬁc economies’ elderly

4.1 Economy-Wide Reforms

4.1.2 Corporate Governance

4.1.3 Market Distortions

4.2 Challenges in Social Protection

4.2.1 Expanding Coverage

Figure 1.8: Coverage Rate of Social Security in Emerging Asia

4.2.2 Retirement Income Adequacy and Financial Sustainability

4.2.3 Low Collection Rate and Investment Efﬁciency

5. Summary and Conclusion

Haruhiko Kuroda, President, Asian Development Bank

It is my great pleasure to welcome you all to ADBI’s 12th Annual

There are a variety of interpretations of “social protection” but all overlap to

2. Social Protection Across the Region

Figure 3.1: Social Protection Index Values for Asian Countries