A well-designed system for fiscal decentralization and subnational levels will yield to uncertain inter-
promises many benefits, including enhanced eco- governmental flows, capricious revenue policies,
nomic growth and macroeconomic stability. If, and a lack of medium-term fiscal planning.
however, decentralization is designed badly or Much is at stake in getting this decentralization
implemented and monitored improperly, such a sys- system right. East Asian countries have recovered
tem can foster “soft” budget constraints—loosened rapidly from the 1997 financial crisis, posting high
controls on the fiscal activities of subnational growth rates. Regional output grew by 6.7 percent
governments—that may undercut the spectacular in 2002 and 7.9 percent in 2003 and an estimated
economic growth occurring in East Asia today. 7.8 percent in 2004. Growth is expected to reach
This chapter highlights the challenges of decen- 7.1 percent in 2005—the strongest record since the
tralization in the context of macroeconomic man- start of the global and regional recession in late
agement, especially in achieving fiscal sustainability 2000. The number of people living below US$2 per
and providing a medium-term environment for day is estimated to have fallen to around 34 percent
sustained growth. The focus is twofold. First, the in 2004, amounting to some 636 million people—
chapter highlights the critical need for central gov- down from 50 percent as recently as 1999, repre-
ernments to monitor subnational fiscal and quasi- senting 890 million people (World Bank 2004, 2005
fiscal activities through well-designed reporting, table 1.1).
auditing, and financial management institutions, as The policy challenge for each country is to
well as information databases. Second, the chapter maintain its high growth rate while strengthening
addresses the need to institutionalize incentives for its fiscal health through intergovernmental coordi-
subnational governments to publicly and regularly nation and monitoring. Unfortunately, this task
report relevant financial information. is far from straightforward. For a start, whether
Most East Asian countries are at the early stages developing countries experience a causal relation-
of this twofold process. Unless the process keeps ship between decentralization, macroeconomic sta-
pace with decentralization, the result will be subna- bility, and economic growth is unclear. The reasons
tional arrears and pressures for central bailouts—the are twofold. First, the data for measuring the extent
latter occurring through ad hoc financial transfers, to which a country has fiscally decentralized are
or accounting “adjustments” that take significant tax simply not available—a worldwide problem that
and spending activity off-budget. Once countries also exists in East Asia. Second, the line is not
abandon the discipline of hard budget constraints, always clear between formal and informal institu-
systematic budget planning and execution at central tional arrangements for enforcing hard budget
54 East Asia Decentralizes
constraints. Even if the right data and definition time frame for assessing the impact of different
were available to measure decentralization, track- types of public spending is also important. Analysts
ing fiscal risks can be difficult when subnational see health and education spending, for example, as
governments have hidden, off-budget expenses and an investment in human capital with a long-term
funds, as in some East Asian countries. payoff, resembling physical investment.4
Still, the elements of responsible fiscal manage- Ebel and Yilmaz (2003) reproduced this analysis
ment, and the criteria for maintaining fiscal disci- by weighing degrees of revenue autonomy. They
pline in a decentralizing environment, are well found that tax autonomy and nontax autonomy have
known. This chapter therefore focuses on empirical a positive correlation with economic growth, while
conclusions regarding the links between fiscal tax sharing has a negative relationship (see chap-
decentralization and growth, the role of budget ter 6).These results suggest that a country’s economic
constraints in fostering fiscal discipline, the rela- performance partly reflects the degree to which sub-
tionship between subnational borrowing and a national governments control their revenues.
country’s overall fiscal sustainability, and the result- In a study of the United States, Xie et al. (1999)
ing need for managing financial risk. showed that further fiscal decentralization with the
aim of boosting efficiency and economic growth
would in fact harm growth. Akai and Sakata (2002)
Links between Fiscal refuted that finding, incorporating a more elabo-
rate definition of fiscal decentralization. Their
study examined the growth impact of giving sub-
In theory, devolving fiscal responsibilities can national governments the authority to raise taxes
improve public sector efficiency, boost competition and spend public funds.5 Since local governments
among subnational governments in delivering pub- do not necessarily spend locally collected revenues
lic services, and stimulate economic growth (Bird locally, the authors found that they may not in fact
and Wallich 1993). These potential benefits reflect have tax autonomy (Zhang and Zou 2001).
the belief that subnational governments can best Martinez-Vazquez and McNab (2001) con-
make growth-promoting public expenditures in firmed the inconclusive relationship between fiscal
sectors such as education, health, and infrastruc- decentralization and economic growth and the
ture, because they have better information on local importance of individual country circumstances—
circumstances and interregional differences (Oates including noneconomic ones—in determining
1993).1 causality. Lin and Liu’s analysis of China (2000),
East Asia’s spectacular precrisis growth and using provincial-level data from 1970 to 1993, sug-
postcrisis rebound have depended significantly on gests that fiscal decentralization has helped spur
prudent macroeconomic management. However, that country’s impressive growth over the past 20 or
empirical research has been inconclusive regarding 30 years.6 However, Zhang and Zou (2001) found a
any causal relationship between decentralization negative relationship between fiscal decentraliza-
and growth in developing countries. tion and economic growth in China.7
In a study of 46 countries from 1970 to 1989,
Davoodi and Zou (1998) found a negative relation-
Subnational Fiscal Discipline
ship between fiscal decentralization and economic
and Budget Constraints
growth in developing countries, and no relation-
ship in developed countries.2 The authors explain Determining the degree of fiscal discipline among
that “conventional wisdom points towards positive subnational governments in East Asia requires
growth effects of capital and infrastructure spend- identifying the budget constraints they face. Doing
ing and negative growth effects of welfare and cur- so in countries like Cambodia—where subnational
rent spending.” This implies that in developing governments have little spending and revenue
countries—where subnational entities spend a authority and rely almost entirely on transfers from
larger proportion of their budgets on wages and the center—is relatively easy. At the other extreme,
salaries and social welfare—decentralized systems about half or more of the resources available
exhibit slower long-term growth per capita.3 The to China’s subnational governments come from
Achieving Fiscal Sustainability 55
off-budget sources, over which they have virtually Information on these quasifiscal transactions
total control.8 In Vietnam, off-budget accounts remains weak, and work on improving this infor-
such as the Social Security Fund, the Health Insur- mation base is just beginning, even in China.
ance Fund, and the National Development Fund Finally, subnational governments in many coun-
constrain the budget autonomy of subnational tries have faced difficulties in planning their activi-
governments. ties and managing their finances owing to delays in
In some East Asian countries, state-owned allocation decisions and a lack of predictability
enterprises at the subnational level receive off- regarding intergovernmental transfers. In Thai-
budget resources in the form of deferred taxes or land, other than knowing with certainty that their
arrears accruals on debt service and other contrac- nongrant revenues will equal those of the previous
tual payments. This is a particular concern in China fiscal year, local governments find it difficult to
and Vietnam. In Thailand, this problem is miti- predict their shared tax revenue allocations. This is
gated by the fact that subnational governments do mainly due to delays in establishing the criteria for
not typically own enterprises, and by the inclusion distributing the allocations from the center to local
of all state-owned enterprises in the “consolidated governments.10
public sector’s” budgetary accounts.
Extrabudgetary revenues and expenditures
Subnational Borrowing and
among subnational authorities make their true
budget constraints difficult to ascertain. This is
especially problematic when data are not reported Fiscal sustainability means that a subnational gov-
in a timely manner and are often incomplete, as in ernment covers its expenditures out of its own rev-
most East Asian countries. Thus, when govern- enues, reducing its dependence on borrowing and
ments rely substantially on extrabudgetary funds, transfers from the center (Bird 2003). To determine
national budgets give only a partial picture of fiscal whether a subnational government’s plans are fis-
realities. cally sustainable, analysts need accurate informa-
Enforcing hard budget constraints among subna- tion on revenues and expenditures at the central,
tional governments requires clear expenditure regional, and local levels. Analysts also need to
assignments, formula-based transfer systems, local understand the interplay between intergovernmen-
revenues, prudent subnational borrowing rules, and tal grants and government borrowing—that is, how
good financial reporting (see box 1.3 in chapter 1). hard the subnational budget constraint truly is.
None of the East Asian countries examined here meet Careful analysis of country-specific intergovern-
these prerequisites. Such a situation can lead to per- mental relations, and the resulting incentive frame-
verse incentives to overspend, accumulate arrears, work, should accompany any analysis of fiscal
and overborrow. Key among such incentives is the sustainability.
prevailing practice of higher-level financial bailouts Subnational governments in most of the East
for subnational governments that are already in or Asian countries examined here rely heavily on
even heading toward default. This is of particular intergovernmental transfers. The large bailouts
concern if the subnational government is large, as is during the 1997 financial crisis and since, such as
often the case with soft budget constraints.9 the recent recapitalization of Chinese state-owned
A country’s system of intergovernmental fiscal banks in 2004, have undermined fiscal sustainabil-
management (or lack thereof) may also motivate ity in East Asia by softening the budget constraints
subnational authorities to keep their transactions imposed on subnational governments. Discre-
off-budget—especially richer provinces that do not tionary transfers to deficit subnational govern-
want to cede part of their revenue base to the center ments in China, Indonesia, and Vietnam, and
for tax sharing. Effective fiscal decentralization unclear assignment of responsibilities throughout
requires an institutional structure that minimizes the region, have compounded this result.
such adverse expectations. In China, “fee-to-tax” Many subnational governments in East Asia also
reforms aim to bring some off-budget subnational have access to onlending from the central govern-
revenues within the budgetary umbrella, but much ment and donors. Such lending is typically in the
remains to be done in most East Asian countries. form of sovereign guaranteed external loans that
56 East Asia Decentralizes
the central government contracts but channels to borrowing, subnational governments have incen-
subnational entities at different interest rates. tives to run unsustainable deficits.
Onlending interest rates and terms vary across Fiscal, political, and financial institutions that
countries. In China, the final borrower bears the strengthen competition at the local level, especially
entire foreign-exchange risk associated with these for capital, can promote hard budget constraints.13
loans, while in other countries the center assumes If institutions directly or indirectly suggest that the
some or all of this risk. Most subnational borrow- central government will step in to cover subna-
ing in Indonesia has occurred through central gov- tional liabilities in the case of default, they may
ernment onlending mechanisms on terms that are encourage subnational governments to “overbor-
highly favorable to the center. However, the repay- row, overspend, or undertax” (Rodden 2000a).
ment record of Indonesian subnational borrowers The Latin American experience suggests that
has been poor.11 Thailand allows subnational gov- subnational governments that are subject to hard
ernments to borrow from domestic banks and budget constraints are more likely to tax and spend
bond markets, subject to legal ceilings in any given prudently (Bird 2001). Fiscal sustainability is also a
budget year. In Indonesia, the Ministry of Finance forward-looking concept, in that it requires accu-
sets the aggregate limits to regional borrowing for a rate assumptions about revenues and expenditures
particular fiscal year in August of the previous year, and key economic variables. Making meaningful
and no direct borrowing from foreign sources is assumptions that are palatable to policy makers and
permitted. Instead, all such borrowing occurs their constituents requires a good understanding of
through onlending arrangements with the min- institutional and country-specific details.
istry. Subnational governments in Vietnam may not In Indonesia, for instance, only after it passed
borrow at all (see table 3.1). Laws 22 and 25 on intergovernmental fiscal rela-
International experience since the early 1980s, tions in 1999 and issued implementing regulations
especially in Latin America, suggests that without (PP107) in 2000 could regions borrow without
appropriate accountability and transparency mech- strict approval from the center.14 Subnational
anisms, decentralization can encourage dangerous domestic borrowing is now subject to a rule-based
opportunistic behavior by state and local authori- approach and central government approval. For-
ties. If left unchecked, such opportunism could eign borrowing is not allowed, except through
undermine macroeconomic stability. The most onlending from the central government or donors.
vivid manifestation of this phenomenon is the soft- In view of Indonesia’s public debt burden, how-
ening of subnational budget constraints (Rodden ever, and as the country establishes local financial
2000a; World Bank 2002). Avoiding this risk management structures, a ministerial decree tem-
depends on the ability of the central government to porarily banned any subnational borrowing until
prevent subnational authorities from passing their the intergovernmental fiscal relations Law 25/99
liabilities to higher-level governments.12 This, in was revised. This ban has been in effect since 2000
turn, requires institutional mechanisms to disci- but was to be lifted at the end of fiscal year 2004,
pline borrowing by state and local governments. after the implementation regulations of the revised
Examining experiences in Argentina, Brazil, and Law 25 are in place. The goal is to maintain fiscal
India, Rodden et al. (2003) show that unsustainable discipline by strengthening both market and rule-
subnational deficits emerge when provinces have based mechanisms—highlighting the fact that
powerful representatives, when they depend heavily countries may sometimes require strong, centrally
on intergovernmental transfers, and when they imposed fiscal constraints, especially in the early
have autonomous access to sources of deficit stages of decentralization. Subnational govern-
financing. The latter can include bonds, loans from ments may rely on short-term borrowing (with
domestic banks (which may themselves be state- maturities of less than 12 months) to manage their
owned, as in China), nonpayment of employee cash flow. In theory, as in China, the Indonesian
wages, and contingent liabilities. If these exist central government can intercept general grant
where the central government cannot commit to funds if a region fails to meet its debt service
a no-bailout policy, or cannot limit subnational obligations.
TABLE 3.1 Subnational Borrowing Practices
of practices Cambodia China Indonesia Philippines Thailand Vietnam
Borrowing power LGs cannot borrow. LGs may not borrow LGs can borrow or LGs can borrow or LGs cannot borrow, Provinces and cities
and practices Stipulated according against general issue bonds in issue bonds; some domestically or with provincial
(excluding bor- to central govern- revenues or issue domestic market prudential from abroad, status can borrow
rowing from cen- ment regulations. bonds. subject to rules restrictions. without prior within prudential
tral government) LGs can and do and central gov- approval limits and as
borrow through ernment approval. from CG. approved by CG.
asset-holding Temporary freeze on
or project all borrowing up
companies. to 2005.
Borrowing from LGs may not Borrowing by LG LGs borrow mainly LGs borrow exclu- LGs not allowed to
the government/ borrow. (mainly special from CG through sively from gov- borrow without
donor onlending service units) donor onlending ernment financial prior approval
from China and from budget. institutions; from CG.
Development 50% of latter loans largely donor
Bank; commercial in arrears; 63% of onlending.
bank loans to loans to LGs in GFIs monopolize
“off-budget” arrears. depository bank
funds. New onlending business, so de
terms being facto IRA intercept
negotiated. has led to good
Under new PP107, LG repayment
CG can intercept history.
DAU. Terms of onlending
loans by GFI to be
in line with those
of local commer-
cial banks, but in
practice GFI often
sets a lower rate.
TABLE 3.1 (continued)
of practices Cambodia China Indonesia Philippines Thailand Vietnam
Banking sector Not much banking LG-owned LGs borrow short Private commercial LGs are only now LG-owned compa-
sector–related companies borrow term from banks bank lending to exploring oppor- nies borrow from
activity; basically primarily from to cover cash flow. LGs virtually tunities to borrow banks to finance
a cash economy. commercial banks, Banking sector still nonexistent. from local com- infrastructure
Most banks mostly 3–5 years; under restructur- Land Bank of the mercial banks and projects.
foreign-owned. sometimes 10. ing; 22 of 27 BPDs Philippines, the public revolving Commercial banks
National Bank of Interest rates recapitalized. largest provider funds. purchased HCMC
Cambodia regulated by PBC. of credit to LGs, LGs mainly borrow private placement
opened dollar- Banking sector is now uses its own from local bond.
denominated burdened by high resources (as well development Gradual liberaliza-
accounts for rate (40%) of as donor funds): funds monitored tion of interest
commercial banks nonperforming 5 years. by Ministry of rate since 1996;
in 1998; deposits loans; concern for Privatized Philip- Interior. State Bank of
with the bank are the ability of LG pines National Vietnam (SBV)
remunerated at companies to Bank still lends to removed lending-
7/8 of SIBOR. depend on LGs (eligible to be rate ceiling.
rollover of bank an LG depository): SBV launched
loans. 4–7 years. banking-sector
New GFI players in restructuring
LG credit. program in 2001,
phasing in SOCB
Capital markets LG-owned compa- Defaults of munici- 6 LG banks: West 13 LG bond issues LGs cannot go HCMC issued GO
nies cannot issue pal bonds in early Java, East Java, since 1991: 4 directly to capital bonds in 1995
bonds. 1990s led to ban Central Java, issues guaranteed markets. (D 30 billion,
Credit to the private on LG bond North Sulawesi, by Home Guaran- Debt financing by 3-year, private
sector is 7 percent issuance. West Sumatra, tee Corp. (govern- LGs has been placement) for
of GDP, loans are LG companies issue and PT Bank DKI ment agency), 8 limited (some toll-road project;
mostly short term. bonds (Shanghai: issued 12 bonds issues guaranteed revenue bonds is preparing
Mainly providing 8 years); strict (1991–2000); by LGUGC; mostly issued). another. (Decree
import/export procedures for the 3–7 years. 7-year. 93/2001 allowed
financing and issuing corporate LGUGC established HCMC bond
working capital to bonds (single A or in 1998, owned issuance.)
trade and service better for issues 51% by Bankers Ministry of Finance
sectors. greater than Y Association and approval required
100 million). 49% by Develop- for bond issues;
ment Bank of the market-rate
Philippines. pricing unlikely.
lished in 2000;
bond is 10% of
ume; the Hanoi
in early 2005.
Monitoring of local Financial accounting Government finan- Regional financial Ministry of Finance Ministry of Informa- Central budget
government centralized in cial information information pilots new system tion monitors sub- guidelines.
credit National Treasury system being system being for LG fiscal/ national borrow- City budget
and provincial piloted. prepared; Ministry financing ing from local information not
branches. LG credit not of Finance starts reporting. development disclosed.
Main financial state- assessed. mapping LG fiscal GFI appraises central funds and public SOCB evaluates
ment for LG Credit-rating capacity? government revolving funds. large-scale
expenditure man- agencies exist. transfer and LG projects.
agement pro- tax base.
duced by Ministry LGUGC has internal
of Finance Budget credit-rating
Source: World Bank.
Note: BPD state-owned commercial banks; DAU dana alokasi unam, the main intergovernmental fiscal transfer mechanism in Indonesia; HCMC Ho Chi Minh City;
GO government order; IRA Internal Revenue Allotment; PBC People’s Bank of China; LG local government; CG central government; GFI government
financial institution; LGUGC local government unit guarantee corp.; SOCB state-owned commercial bank.
60 East Asia Decentralizes
In Thailand, local governments may borrow do not intend such an outcome, or if local gov-
domestically and internationally, with prior ernments do not use the allocated resources effi-
authorization from the cabinet. These governments ciently, the resulting trends will undermine fiscal
may issue debt securities and borrow from official, sustainability.
external bilateral creditors for development proj- For such rules to be credible and sustainable,
ects. In practice, local debt financing is somewhat they must be part of a well-articulated fiscal frame-
limited, including from domestic capital markets. work that improves the government’s position over
The primary source of borrowing has been local the long term. Such a framework includes clear
development funds managed by the Ministry of intergovernmental fiscal relations, appropriate sub-
Interior. Subnational governments have more national tax structures, and public pensions.20
recently borrowed from commercial banks and The central government needs to clearly define
public revolving funds. accountability and establish financial management
In Vietnam, subnational governments may not practices to enforce these rules.21 In fact, the tem-
run fiscal deficits. Provinces may borrow, but only porary ban on subnational borrowing in Indonesia
domestically, by issuing project investment bonds, is a prudent interim measure until the country
or by borrowing from the Development Assistance establishes a more solid framework for regional
Fund.15 Provinces may use these funds only for borrowing and a regional financial information
projects with prior approval from the Provincial system. A more solid framework for subnational
People’s Council under the five-year provincial borrowing must include procedures for handling
Public Investment Plan. The province must also regional default. Otherwise, without a credible legal
allocate funds for debt service in future budgets or regulatory threat, the center will end up paying
until the debt matures.16 Local state-owned enter- the bill.22
prises that provide essential services, such as waste
disposal, water, electricity, and transportation, may
Managing Fiscal Risks
borrow from both external and domestic sources.
All external borrowing is subject to approval by the Rules and administrative controls can help reduce
central government, which provides a sovereign the risks of subnational borrowing. Key measures
guarantee. include strengthening the intergovernmental fiscal
To promote responsible subnational borrowing, system and, when the situation warrants, requiring
some East Asian governments have disseminated ex ante authorization and ex post monitoring. For
well-defined, transparent fiscal rules. The rationale instance, the central government could set annual
is that all borrowing decisions should take into limits on the debt of individual local governments,
account the fiscal implications for future genera- review individual loans, including their terms and
tions.17 The issue of implementation remains to be conditions, and centralize all borrowing, with
addressed, however. onlending to local governments.23 Other measures
In Indonesia, regulations implementing the to encourage fiscally sustainable borrowing include
decentralization framework limit the debt-to- prohibiting subnational authorities from issuing
revenue ratio to 75 percent of the previous year’s guarantees (see table 3.2 for the kinds of explicit
general revenue, and the debt service-to-revenue and implicit guarantee mechanisms that may exist
ratio to 40 percent in any given budget year. Regula- at subnational levels), and imposing ceilings on the
tions also govern onlending to the regions.18 In the net worth or loan portfolio of borrowers. Indonesia
Philippines, the central government limits transfers and Thailand have established ceilings on debt or
to local governments to 40 percent of their internal debt service as a share of local revenues. Other
revenues from three years before.19 Since the ratio countries, such as Vietnam, require local govern-
of revenue to gross domestic product has been ments to balance their budgets and restrict their
falling, this rule implies that some local govern- borrowing to specific purposes, such as capital
ments may receive higher transfers than intended, investment.
or than the central government can afford. Declin- Countries can also rely on the market to regulate
ing trade taxes also build upward bias into the rev- subnational borrowing. A market-based system
enue share of local governments. If policy makers requires minimum legal and regulatory structures,
Achieving Fiscal Sustainability 61
TABLE 3.2 Subnational Fiscal Risks
Liabilities (obligation in any event) (obligation if a particular event occurs)
Government Local government debt. Local government provides
liability is Arrears (if legally binding). guarantees for debt and other
recognized Nondiscretionary budgetary obligations of financial and
by law or spending. nonfinancial enterprises and
contract. other entities.
Local government insurance
schemes (such as crop insurance).
A “moral” Capital and recurrent costs of local Claims arising from local
obligation on public investment projects. government letters of comfort.
the part of the The cost of future benefits under Claims by failing local financial
government local social security schemes. institutions and other entities.
that mainly Claims related to enterprise
reflects public restructuring and privatization.
expectations Claims by beneficiaries of failed
and pressure by social security or other funds,
interest groups. beyond any guaranteed limits.
Claims related to local crisis
management, such as public health,
environment, and disaster relief.
Source: Brixi and Mody 2002.
such as supervision and disclosure practices; guide- Subnational governments that relax their budget
lines for issuing, settling, and repaying debt; bank- constraints contribute to public sector deficits and
ruptcy procedures (including creditor remedies); threaten national solvency. Hidden budget channels
protection against disruption of essential public include off-budget borrowing; arrears on civil serv-
services; and measures to prevent moral hazard. ice wages and payments to suppliers and other levels
East Asian countries are only now establishing of government; indirect liabilities through public
these institutional structures. enterprises or publicly owned banks, which are often
Another market instrument for reducing the insolvent; and other contingent liabilities such as
credit risk of subnational borrowing is regular unfunded pension and provident funds. Soft budget
monitoring of creditworthiness. Private sector enti- constraints and the expectation of central bailouts
ties can help investors by rating the likelihood that contribute to moral hazard. Data on these “hidden
subnational governments will default. Such credit deficits” in East Asia are just becoming available
ratings should reflect both the capacity and the (Kharas and Mishra 2001). Thus central govern-
willingness of debt issuers to make timely payments ments lack the information they need to monitor the
on both principal and interest. Key elements of fiscal risks of subnational governments. In most East
creditworthiness include the subnational govern- Asian countries, the Ministry of Finance typically
ment’s economic base (net worth), revenue auton- receives regular reports on budgetary revenues and
omy and stability, revenue-expenditure balance, expenditures on a cash basis from subnational gov-
intergovernmental fiscal relations, the subnational ernments. The ministry does not, however, have
debt burden, and contingent liabilities. Financial access to timely information on many extrabudgetary
management practices in the region matter, as do and off-budget capital expenditures and borrowing,
guarantees, insurance, and other mechanisms to or on local guarantees, financial institutions, pen-
enhance the credibility of subnational borrowing. sion funds, employment insurance funds, and other
However, credit ratings of subnational entities are transactions that could generate liabilities (Ma and
not yet available in most East Asian countries. Brixi 2002). Anecdotal evidence on provincial-level
62 East Asia Decentralizes
off-budget and contingent liabilities abounds in vehicles that carry explicit or implicit guarantees
some countries. from the central government raise the latter’s risk.
In China—where direct and indirect support of Yet estimating the scope of the contingent liabilities
expanded investment and credit to subnational that subnational spending and borrowing impose
entities has driven much recent growth—reliance on the central government is difficult, especially if
on banking and off-budget funds has been impor- no one government agency is cataloguing these
tant.24 Poor monitoring of such investment implies transactions. East Asian countries are just begin-
that the health of the financial sector is at risk. Sig- ning to undertake this task, at least for large contin-
nificant investment in industries such as cement, gent liabilities. China and Thailand have been
steel, and aluminum seems even riskier, given sub- building the capacity of the central government to
stantial excess capacity. manage overall public debt.
Table 3.2 outlines typical sources of contingent Early warnings, such as those used in Brazil,
liabilities, which can be explicit or implicit. For Colombia, and the United States, can provide a
instance, the explicit fiscal burden from rising good starting point for monitoring the fiscal risks
safety net expenditures may generate liabilities for of subnational borrowing, but such indicators may
the central or provincial government. Also, what not reliably reflect future financing pressures.27 Ma
may be “contingent” for the central government (2001) has proposed a composite indicator that
may be a direct liability for a provincial or subna- reflects both fiscal pressures and the current fiscal
tional government. Cataloguing such liabilities is position of subnational governments. Information
an essential first step in establishing a system to on their assets and liabilities, exposure to market
assess the obligations and fiscal risks of local and rollover risks, capacity for managing these
governments.25 China, Indonesia, Thailand, the finances, and the compatibility between revenue
Philippines, and Vietnam are now embarking on and spending responsibilities will enable informed
this important endeavor. judgment concerning fiscal risks that may need
As provincial and municipal governments gain immediate attention. Without sound fiscal report-
greater authority to tap domestic and international ing and auditing at the subnational level, even the
financial markets, the financial risk at subnational best-designed early-warning system will not be
levels will also grow—and will need managing. effective. Countries need clear rules for dealing
Weaknesses in regulation and oversight have led to with debt-distressed subnational governments.
the proliferation of off-budget financing, govern- Vietnam is examining the fiscal risks of borrowing
ment guarantees, and other contingent liabilities. by state-owned enterprises, but most East Asian
Decentralization has given subnational governments countries are just beginning to tackle this issue.
a greater role in managing and delivering public
services, and more budgetary responsibility for civil
service pensions and provident funds. All these fac-
tors exert a significant impact on the quality of serv- A piecemeal approach to intergovernmental fiscal
ices and expenditure mix of local governments. reform is unlikely to succeed because it will not
Countries need to estimate the costs of deliver- take into account the interdependence between
ing devolved responsibilities so they understand the transfers and revenue assignments, or between
potential fiscal risks. For example, if Thailand allo- expenditure assignments and own-source revenues
cates 20 percent of revenues to local authorities but and transfers. To enforce hard budget constraints,
does not devolve commensurate responsibility for institutions must clearly and credibly convey the
expenditures, then the central government bears a message that local governments will bear the costs
significant risk of assuming the resulting deficit.26 and benefits of their fiscal decisions. One such
These risks will grow if service delivery declines institution is the capital market, which in theory
owing to capacity constraints among subnational rations access to capital among subnational bor-
governments. In China and Indonesia, where sub- rowers based on the soundness of their fiscal deci-
national governments have more expenditure sions. Institutions that ensure that public officials
responsibilities than revenue-raising authority, must answer to the needs and aspirations of the
subnational borrowing and off-budget financing local populace are essential as well. However, such
Achieving Fiscal Sustainability 63
institutions can work only if local governments appropriate early-warning systems for fiscal risks.
have considerable fiscal autonomy, and when the East Asian countries have begun work in these crit-
central government makes a credible commitment ical areas, but success will require consistent effort
that it will not provide bailouts (World Bank over the short and medium term.
2000b). None of the countries in East Asia now
meet these conditions for enforcing hard budget
1. This theory assumes that labor and capital mobility will
East Asian countries need a comprehensive
ensure competition among subnational governments for
approach to intergovernmental fiscal reform, tack- effective public sector service delivery, as well as a match
ling subnational expenditures and revenues at the between the preferences of local citizens and governments.
same time. Subnational governments must have 2. Davoodi and Zou (1998) define fiscal decentralization in
terms of spending by subnational governments as a fraction
enough revenues to implement their spending of total government spending. Fiscal decentralization rises if
responsibilities (see chapters 5 and 6). Toward spending by state and local governments expands relative to
this end, incentives that encourage cooperation spending by the central government. The authors use the
average growth of real per-capita output over 5-year and
between officials at different levels are essential. 10-year periods as a proxy for long-run growth.
China and Indonesia need such incentives, for 3. Davoodi and Zou (1998) did not use disaggregated subna-
example. tional data to determine which province or region spent
more on capital and infrastructure relative to others. The
Central governments need to carefully monitor
authors also noted that countries may not realize the effi-
subnational borrowing and the resulting fiscal ciency gains of fiscal decentralization if central authorities
risks in order to maintain fiscal discipline and pru- constrain subnational revenue collection and spending,
dent macroeconomic management. Before countries and if local citizens do not elect local officials. Labor and
capital mobility may not be as easy as theory assumes.
give subnational governments free rein to borrow— 4. A significant part of spending in these sectors often occurs
domestically or abroad—they need to make infor- under “current expenditure.”
mation critical to analyzing subnational creditwor- 5. See chapter 2 in Litvack et al. (1998) for a discussion of the
distinction between decentralization, deconcentration, and
thiness available to stakeholders. Appropriate checks delegation.
and balances must ensure that these data are reliable 6. Lin and Liu (2000) found that rural reform, development
and consistent across provinces. Domestic and of the nonstate sector, and capital accumulation have also
been driving forces in China.
foreign capital markets play an important role in
7. In the same study, the authors found a positive relationship
diversifying local government funding sources and between fiscal decentralization and provincial economic
tracking subnational creditworthiness. East Asian growth in India. Questions about the appropriateness of
countries—especially those with weak financial the “decentralization variable” again arise. See Ebel and
systems—will need to implement a complex set of 8. See World Bank (2002). Subnational governments in China
institutional reforms to support these efforts. also often engage in commercial activities to supplement
Finally, to achieve fiscal sustainability, countries their revenue, using land resources and enterprise assets as
their investment capital. Under a model pioneered in
will also need to reform governance, public enter-
Shanghai, many municipalities are creating “corporations”
prises, sectors such as power, and the intergovern- to manage public resources and debt obligations. These
mental fiscal structure to ensure hard budget corporations may “enhance their debt servicing capacity”
constraints for subnational governments. Central by engaging in profit-making activities. The General
Corporation of Shanghai Municipal Property Devel-
governments also must monitor contingent liabili- opment is one of the first such vehicles, created to
ties to ensure prudent fiscal management. Over the help finance the city’s enormous need for facilities and
longer term, a database and analytical indicators infrastructure.
9. Wildasin (1997) showed that a local government’s ability to
can form the basis for a credit-rating system for extract a bailout from the central government depends on
local governments. Such a system is critical to the former’s size. Larger subnational governments thus
opening up subnational borrowing and developing tend to operate under softer budget constraints, creating
incentives for overspending and overborrowing if not ade-
a municipal bond market.28 Also important in
quately monitored by the center.
minimizing future surprises from the decisions of 10. For example, although Thailand’s central government set
subnational governments is a clear division of grant allocations well in advance of fiscal year 2002, it did
not disburse these grants until the very end of the fiscal
responsibilities across levels of government, formal
year because of delays in establishing the allocation rule.
channels for reporting contingent liabilities 11. Empirical evidence in Lewis (2003) suggests that local
and analyzing their potential fiscal impact, and governments have borrowed “well within their fiscal
64 East Asia Decentralizes
capabilities to repay,” but that they have largely been 25. Such a system can potentially be used to generate a ranking
unwilling to repay these debts. The key issue is therefore of local governments based on their fiscal health, and to
one of credit risk rather than fiscal sustainability of subna- enable the central government to decide when to intervene
tional borrowing in Indonesia. and provide emergency assistance (Ma and Brixi 2002).
12. For example, state-level defaults on debt payments helped 26. These risks will be compounded as the local share of rev-
trigger Brazil’s financial crisis in 2000. See Rodden 2000a. enues grows to 35 percent by 2006.
13. Rodden et al. (2003) draw on the experiences of seven 27. See World Bank (2002). Brazil had imposed limits on sub-
developing and transition economies and four OECD national borrowing, Colombia had a “traffic light system,”
countries in identifying institutional arrangements associ- and Ohio in the United States has a “fiscal watch program.”
ated with soft and hard subnational budget constraints. 28. See Ma and Brixi (2002) for relevant experiences in
This exercise provides lessons for East Asian countries. Australia, Brazil, Colombia, New Zealand, and the
14. States in Indonesia can now borrow for projects that gener- United States, and a list of risk indicators for monitoring
ate a “direct or indirect” financial return, per Implementing the contingent liabilities of subnational governments.
Regulation 107 of 2000. However, the central government
can restrict borrowing if “general conditions” warrant such Bibliography
a restriction, as in the temporary ban by ministerial decree
in 2001. Akai, Nobuo, and Masayo Sakata. 2002. “Fiscal Decentralization
15. The Vietnamese Budget Law stipulates that, for a given Contributes to Economic Growth: Evidence from State-
year, a province can borrow up if its outstanding debt will Level Cross-Section Data for the United States.” Journal of
not exceed 30 percent of its capital budget. Until 2003, local Urban Economics 52: 93–108.
governments had issued only two project bonds, both by Bird, Richard. 2003. “Fiscal Flows, Fiscal Balance, and Fiscal
Ho Chi Minh City (HCMC). In 2003 the central govern- Sustainability.” World Bank Institute, Washington, DC.
ment approved new HCMC urban infrastructure bonds. In Bird, Richard, and Christine Wallich. 1993. “Fiscal Decentraliza-
September and October 2003, HCMC issued two-year and tion and Intergovernmental Fiscal Relations in Transitional
five-year bonds at annual interest rates of 8.52 percent Economies: Towards a Systematic Framework of Analysis.”
and 9 percent, respectively. In December 2003, HCMC Working Paper. Country Economics Department, World
issued another five-year urban bond at 8.5 percent interest. Bank, Washington, DC.
As of October 2003, provincial borrowing amounted to Brixi, Hana Polackova, and Ashoka Mody. 2002. “Dealing with
0.43 percent of gross domestic product (D 2.5 trillion). Government Fiscal Risk: An Overview.” In Government at
16. The Ministry of Finance closely monitors this commitment Risk, ed. Hana Polackova Brixi and Allen Schick. Washington,
each year, requiring all subnational authorities to report DC: World Bank.
their debt, interest and principal payments, and borrowing Burki, Shahid Javed, Guillermo E. Perry, and William R.
plans by type of instrument. In practice, subnational gov- Dillinger. 1999. “Beyond the Center: Decentralizing the
ernments have delayed providing this information. State.” Viewpoints. Washington, DC: World Bank, World
17. Hence, these rules are often derived using the fiscal sustain- Bank Latin American and Caribbean Studies.
ability approach based on the government’s “lifetime Davoodi, Hamid, and Heng-fu Zou. 1998. “Fiscal Decentraliza-
budget constraint.” tion and Economic Growth: A Cross-Country Study.” Jour-
18. Decree KMK35 of 2003 in Indonesia stipulates that nal of Urban Economics 43: 244–57.
onlending is allowed only for regional government projects Ebel, Robert D., and Serdar Yilmaz. 2003. “On the Measurement
that generate revenue. The decree does not specify how to and Impact of Fiscal Decentralization.” In Public Finance in
handle a default by a subnational entity, except in the case Developing and Transitional Countries: Essays in Honor of
of loans from the center, where deductions from the gen- Richard Bird, ed. Jorge Martinez-Vazquez and James Alm.
eral grant would cover the debt service. KMK35 is currently Cheltenham, UK: Edward Elgar Press.
under review. Kharas, Homi, and Deepak Mishra. 2001. “Fiscal Policy, Hidden
19. This means that the intergovernmental transfer equation Deficits and Currency Crises.” In World Bank Economists’
excludes trade taxes. Forum, ed. S. Devarajan, F. H. Rogers, and L. Squire.
20. See Kopits (2001). Adopting fiscal reporting in accordance Washington, DC: World Bank.
with the IMF’s Code of Fiscal Transparency would signal Kopits, George. 2001. “Fiscal Rules: Useful Policy Framework
credibility of rules. or Unnecessary Ornament?” IMF Working Paper 01/145.
21. These include best-practice accounting standards and effec- Washington, DC: International Monetary Fund.
tive monitoring and reporting of government liabilities. Lewis, Blane D. 2003. “Local Government Borrowing and Repay-
22. Where the central government tightly controls local spend- ment in Indonesia: Does Fiscal Capacity Matter?” World
ing decisions, as in Cambodia, China, Vietnam, and to Development 31 (6): 1047–63.
some extent Thailand, local governments are more likely to Lin, Justin Yifu, and Zhiqiang Liu. 2000. “Fiscal Decentralization
credibly shift blame for a fiscal crisis to the center (World and Economic Growth in China.” Economic Development
Bank 2000b). and Cultural Change 49 (1): 1–21.
23. See Burki et al. (1999). Litvack, Jennie, Junaid Ahmad, and Richard Bird. 1998.
24. The fixed-assets investment is driven by local governments “Rethinking Decentralization in Developing Countries.”
rather than the central government. For the first half of Sector Studies Series. World Bank, Washington, DC.
2003, central government projects amounted to only McKenzie, Kenneth J. 2003. “A Note on the Sustainability of
¥ 184.8 billion (12.3 percent), a decline of 7.7 percent over Fiscal Policy in Subnational Governments.” World Bank
the previous year. However, local government projects Institute, Washington, DC.
amounted to ¥ 1,322.4 billion (87.7 percent), an increase of Ma, Jun. 2001.”Monitoring the Fiscal Risks of Subnational Gov-
41.5 percent over the same period in the previous year. ernments.” In Government at Risk, ed. Hana Polackova Brixi
Most of this investment is in infrastructure projects. and Allen Schick. Washington, DC: World Bank.
Achieving Fiscal Sustainability 65
Ma, Jun, and Hana Polackova Brixi. 2002. “Monitoring the Fiscal Relations.” Policy Research Working Paper No. 1843. World
Risks of Subnational Governments.” PREM Notes No. 6. Bank, Washington, DC.
Washington, DC: World Bank. World Bank. 2000a. “Thailand: Public Finance in Transition.”
Martinez-Vazquez, Jorge, and Robert M. McNab. 2001. “Fiscal Report No. 20656-TH. World Bank, Washington, DC.
Decentralization and Economic Growth.” Working Paper ———. 2000b. “Decentralization and the Challenge of Hard
01-1. Georgia State University, Andrew Young School of Budget Constraints.” PREM Notes No. 41. Washington, DC:
Policy Studies. World Bank.
Oates, Wallace E. 1993. “Fiscal Decentralization and Economic ———. 2000c. “Vietnam: Managing Public Resources Better.”
Development.” National Tax Journal 42 (2): 237–43. Public Expenditure Review Report No. 24469, annex B.
Rodden, Jonathan. 2000a. “Decentralization and the Challenge of World Bank, Washington, DC.
Hard Budget Constraints.” PREM Notes No. 41. Washington, ———. 2002. “China: National Development and Subnational
DC: World Bank. Finance: A Review of Provincial Expenditures.” Report
———. 2000b. “The Dilemma of Fiscal Federalism: Hard and No. 22951-CHA. World Bank, Washington, DC.
Soft Budget Constraints around the World.” Massachusetts ———. 2003a. “Decentralizing Indonesia.” Report No. 26191-
Institute of Technology, Cambridge, MA. IND. World Bank, Washington, DC.
Rodden, Jonathan, Gunner Eskeland, and Jennie Litvack, eds. ———. 2003b. “Philippines: Improving Government Perfor-
2003. Decentralization and Hard Budget Constraints. mance: Discipline, Efficiency and Equity in Managing Public
Washington, DC: World Bank. Resources.” Public Expenditure, Procurement, and Financial
Shah, Anwar. 2004. “Fiscal Decentralization in Developing Management Review Report No. 24256-PH. World Bank,
and Transition Economies: Progress, Problems, and the Washington, DC.
Promise.” Research Working Paper No. 2832. World Bank, ———. 2004. East Asia Update. Washington, DC: World Bank.
Washington, DC. ———. 2005. Global Economic Prospects. Washington, DC:
Ter-Minassian, Teresa, and John Craig. 1997. “Control of World Bank.
Subnational Borrowing.” In Fiscal Federalism in Theory Xie, Danyang, Heng-fu Zou, and Hamid Davoodi. 1999. “Fiscal
and Practice, ed. Teresa Ter-Minassian. Washington, DC: Decentralization and Economic Growth in the United
International Monetary Fund. States.” Journal of Urban Economics 45 (2): 228–39.
Weist, Dana. 2001. “Thailand’s Decentralization: Progress and Zhang, Tao, and Heng-fu Zou. 1999. “Fiscal Decentralization,
Prospects.” Paper prepared for the King Prajadipk Institute Public Spending, and Economic Growth in China.” Journal
Congress III on Decentralization and Local Government in of Public Economics 67: 221–40.
Bangkok, Thailand. ———. 2001.“The Growth Impact of Intersectoral and Intergov-
Wildasin, David E. 1997. “Externalities and Bailouts: Hard ernmental Allocation of Public Expenditure: With Applica-
and Soft Budget Constraints in Intergovernmental Fiscal tions to China and India.” China Economic Review 12: 58–81.