Information about Asia and the Pacific Asia y el Pacífico

V Medium-Term Fiscal Challenges

Eswar Prasad
Published Date:
June 2004
  • ShareShare
Information about Asia and the Pacific Asia y el Pacífico
Show Summary Details
Annalisa Fedelino and Raju Jan Singh 

China’s current fiscal position appears relatively sound. Both its overall budget deficit (3 percent of GDP in 2003) and debt as a ratio to GDP (about 26 percent in 2003) compare well with those of a group of emerging market countries, although differences in coverage and data quality need to be taken into account when comparing data across countries (Table 5.1).1 While China’s primary fiscal position may appear weaker than in comparator countries, strong economic growth and low domestic interest rates have so far partially counteracted the possible adverse impact of sustained primary deficits on debt dynamics.

Table 5.1.Selected Countries: Comparison of Public Debt and Fiscal Positions, Average 2000–02(In percent of GDP)
Public DebtOverall BalancePrimary Balance
EMBI countries164.5-4.01.2
Of which: emerging Asia271.4-3.41.6
Source: IMF, World Economic Outlook.

Includes 27 emerging market countries covered by the JP Morgan Emerging Market Bond Index.

Including India, Indonesia, Korea, Malaysia, Philippines, and Thailand.

Source: IMF, World Economic Outlook.

Includes 27 emerging market countries covered by the JP Morgan Emerging Market Bond Index.

Including India, Indonesia, Korea, Malaysia, Philippines, and Thailand.

Despite its current position of relative strength, however, China faces important fiscal challenges in the course of its transition to a market economy. Over the medium term, the government is likely to have to shoulder various costs related to the restructuring of a still largely state-owned economy; the cost of recapitalizing state-owned banks; the funding of social security for a rapidly aging population; possible liabilities, explicit as well as contingent, that subnational governments are contracting, especially related to large-scale infrastructure projects; and significant government programs to address increasing regional disparities, including the need for increased spending for health and education (Nehru and others, 1997; and Lardy, 2000). China’s strong macroeconomic conditions provide a propitious environment in which to tackle these liabilities. If not addressed, these liabilities will continue to increase and may call for significant government interventions under less favorable macroeconomic conditions sometime in the future, thereby potentially creating serious economic dislocations.

Against this background, this section briefly illustrates the main fiscal developments in the last few years, highlights progress in fiscal reforms, and lays out the important challenges for the future. The main conclusion is that, while China’s current fiscal position appears healthy, its fiscal sustainability outlook could be altered by the possible impact of substantial contingent claims on the central government and deviations from the currently very favorable macroeconomic conditions.

Recent Fiscal Developments

China’s fiscal position in the last five years has been marked by a sharp change from previous trends. Steadily declining trends in both revenue and expenditure from the early 1980s to the mid-1990s have been reversed, but with a faster expansion in expenditure than in revenue, thus resulting in larger budget deficits on average.2 Reflecting the fiscal stimulus policies adopted in the wake of the Asian crisis, the budget deficit “jumped” to a new level in 1998; and it has since remained around 3 percent of GDP (Figure 5.1).3

Figure 5.1.Revenue, Expenditure, and Fiscal Balance

(In percent of GDP)

Sources: Ministry of Finance; and IMF staff estimates.

The revenue gains, from a low of 11.2 percent of GDP in 1995 to an estimated 18.8 percent of GDP in 2003, have largely reflected strong growth in tax collections, in particular in value-added tax (VAT) and income taxes (Table 5.2). In addition to buoyant economic activity, the revenue boost may be attributed to two main factors: a reform package introduced in 1994 and improvements in tax administration. The 1994 reform modified the revenue-sharing system between the center and subnational governments in favor of the center (for a background discussion, see Section VI); simplified the tax system, by replacing the multitier system of turnover taxes with a VAT; and reformed tax administration, by splitting the tax bureau into two levels—the State Administration of Taxation, responsible for collecting central and shared taxes, and a local tax administration in charge of collecting local revenue. This last measure, by removing central and shared taxes from local administration, has largely eliminated opportunities to divert central revenue via manipulation of tax assessments.

Table 5.2.State Budget Revenue1
1997199819992000200120022003 Estimate
(In percent of GDP)
Total revenue12.113.014.315.317.018.318.8
Tax revenue11.111.813.014.115.716.817.1
Of which:
Taxes on income and profits2.
Value-added tax on domestic goods4.
VAT and excises on imports0.
Nontax revenue1.
(As a share of total revenue)
Tax revenue91.390.791.092.092.391.991.3
Of which:
Taxes on income and profits17.215.416.419.223.624.6
Value-added tax on domestic goods36.435.533.133.332.332.233.0
VAT and excises on imports5.65.48.710.910.09.812.7
Nontax revenue8.
Sources: Ministry of Finance; State Administration of Taxation; and IMF staff estimates.

The coverage of these data includes the central government, provinces, municipalities, and counties.

Sources: Ministry of Finance; State Administration of Taxation; and IMF staff estimates.

The coverage of these data includes the central government, provinces, municipalities, and counties.

Improved tax administration has played a crucial role in strengthening revenue performance in China, especially in view of the challenges posed by the transformation from a centrally planned economy to a socialist market economy with a decentralized fiscal structure. Under the central planning system, taxpayers, most notably state-owned enterprises (SOEs) and collective farms, were relatively small in number and could be easily monitored. In a market economy with a decentralized administrative structure, not only are there many more taxpayers, but they also cannot be as easily monitored from the center. For example, it is estimated that, in the rural sector, the government shifted from collecting taxes from about 50,000 communes to more than 200 million households and over one million town and village enterprises (Wong, 1997). Among other reforms, the “golden tax” project (computerization of the VAT collection) has significantly contributed to increased revenue. Finally, steps have been taken to bring off-budget revenue on budget through the extension of budget coverage to previously excluded government units and through the tax-for-fees reform.4

The revenue increase has allowed a gradual expansion in expenditure programs, from a recent minimum of about 13 percent of GDP in 1996 to an estimated 21.6 percent of GDP in 2003 (Table 5.3).5 The largest increases have been recorded in areas targeted by the government’s development policies. For example, expenditure on pension and social welfare programs increased by about 2 percent of GDP over this period, as some social expenditure mandates previously assigned to SOEs were transferred back to the government. Similarly, capital expenditure has expanded by about 1.5 percent of GDP over the last five years. There was also an increase in expenditure for administration and defense, by almost 3 percent of GDP, matched by a similar increase in outlays for culture, education, public health, and science.6

Table 5.3.State Budget Expenditure1
1997199819992000200120022003 Estimate
(In percent of GDP)
Total expenditure and net lending14.016.118.318.920.121.721.6
Current expenditure11.512.613.714.716.117.317.4
Of which:
Administration and defense2.
Culture, education, public health, and science2.
Pensions and social welfare relief0.
Capital expenditure2.
Unrecorded expenditure0.
Memorandum item:
Primary expenditure13.215.117.618.119.321.020.8
(As a share of total revenue)
Current expenditure82.578.374.777.879.980.080.5
Of which:
Administration and defense19.423.522.623.424.625.4
Culture, education, public health, and science19.016.916.716.917.918.317.8
Pensions and social welfare relief1.
Capital expenditure15.615.918.817.717.918.117.5
Unrecorded expenditure1.
Memorandum item:
Primary expenditure94.394.296.095.795.997.096.2
Sources: Ministry of Finance; and IMF staff estimates.

The coverage of these data includes the central government, provinces, municipalities, and counties.

Sources: Ministry of Finance; and IMF staff estimates.

The coverage of these data includes the central government, provinces, municipalities, and counties.

Although not as far-reaching as the 1994 reform package on the revenue side, various reforms have been incrementally introduced on the expenditure side as well. Starting in 1999, a blueprint for reforms in public expenditure management was adopted, resulting in the restructuring of the Ministry of Finance, the creation of a treasury department, and the establishment of a single treasury account at the central level and its extension to the subnational level.7 Work is also under way on improving the budget classification system and extending the budget coverage to extrabudgetary activities. These ongoing reforms should contribute to improved expenditure management and more efficient expenditure policies.

Since the use of central bank overdrafts was discontinued in 1994 and recourse to foreign financing has remained limited, the rising budget deficits have been covered mainly by issuance of government paper. Hence, the stock of government debt has gradually increased over the last few years, reaching about 26 percent of GDP at end-2002 (Figure 5.2). As foreign debt has remained stable at less than 5 percent of GDP over the period, the issuance of government paper accounts for most of the growth in government debt. The sharp increase in debt in 1998 is also due to the issuance of bank recapitalization bonds, equivalent to 3.4 percent of GDP.

Figure 5.2.Government Debt and Its Composition

(In percent of GDP)

Sources: China Securities Regulation Commission; State Administration of Foreign Exchange; Ministry of Finance; and IMF staff estimates.

1 Includes bank recapitalization bonds issued in 1998 and onlending.

Medium-Term Challenges

Despite the broadly favorable fiscal position currently enjoyed by China, the government stock of debt would be significantly larger if a number of liabilities—direct as well as contingent—were to be assumed by the central government.8 These include nonperforming loans (NPLs)—in case of financial distress, state banks may need to be bailed out by the government;9 unfunded pension obligations; and liability-creating activities of SOEs and local governments (Daniel and others, 2003). All these obligations may need to be settled by the government in the future in the absence of sustained reforms or changes in the current fiscal management system.

A preliminary assessment based on officially reported NPLs at end-2003 suggests that potential losses from NPLs could amount to about 13 percent of GDP for the four big state commercial banks alone (see Section VII for more details on NPLs). This figure is based on a reported NPL stock of 16 percent of GDP and an assumed recovery rate equivalent to that achieved so far by asset management companies (21½ percent). However, potential weaknesses in reporting and inevitable problems in implementing new prudential regimes, especially in some of the smaller banks, suggest that the existing stock of NPLs could be significantly higher. In addition, while efforts are being made to reduce the flow of new NPLs through SOE and financial sector reforms, part of ongoing bank lending could become nonperforming over the next few years (see also Karacadag, 2003).

Estimates of the government’s pension liabilities vary a great deal, depending on assumptions regarding reforms to be introduced into the system. The current pension system, reflecting previous reforms, has three pillars comprising a public pension, mandatory individual pension accounts, and supplementary voluntary individual accounts. The system has an increasing cash deficit due to a relatively high replacement rate (ratio of retirement benefit to earned income) and the need to finance pension liabilities of the old system. As a result, the individual accounts are largely notional, with current contributions being used to meet current payment obligations. If this pay-as-you-go practice is maintained, the deficit will widen as the ratio of contributors to beneficiaries declines further as the population ages.

The World Bank estimates that the transition cost of shifting from the present system to fully funded individual accounts could amount to a net present value of 70 percent of GDP (over the next 75 years) if the parameters in the retirement system remain unchanged (Dorfman and Sin, 2001). Even relatively marginal reforms could drastically alter these costs if undertaken in the next few years. For example, the financing gap could be lowered to about 7 percent of GDP through a number of parametric changes, including raising the retirement age to 65 and indexing pensions to inflation rather than wages. The remaining financing gap could be covered from sources other than payroll contributions, for example privatization proceeds.

The restructuring of SOEs could also translate into significant government liabilities. First, while part of the borrowings by SOEs and semigovernmental finance companies does not carry any explicit government guarantee, it could nevertheless become a liability for the central government in the future in case these entities are unable to discharge their financial obligations.10 Second, social obligations of SOEs (such as the provision of education and health services and pensions) have been progressively transferred to local governments (see Section VI). Again, payments for these obligations may become the responsibility of the center in case local governments experience financial difficulties.

While subnational governments are banned by law from borrowing directly, they have effective recourse to bank borrowing and bond issuance through public enterprises, especially to fund large infrastructure projects. This indirect financing carries fiscal risks for both local governments and, ultimately, the central government. For example, public infrastructure is largely financed by bank loans. Where these projects have limited repayment capacity, or where profits are subject to other commitments (e.g., cross subsidization of loss-making highways), difficulties in servicing debts may translate into direct government intervention to bail out these projects. Estimates of the size of these contingent liabilities are, however, not available.

There are also substantial requirements for additional public spending in the coming years to meet social needs. Although infrastructure spending has significantly expanded over the past years, the OECD (2002b) reports that spending on education and other social programs is still falling short of comparable international levels.11 Moreover, expenditure needs over the next decade for health (especially as the population ages), education, and environment protection will increase further. China’s recent accession to the World Trade Organization will also require additional government resources to facilitate the adjustments implied by trade and investment liberalization, such as retraining displaced workers. Although the uncertainties are large, the OECD estimated that government primary expenditure would need to increase by at least an additional 2 percent to 3 percent of GDP over the next five to ten years to accommodate these needs.

Despite these rising spending pressures and the costs of addressing some contingent liabilities, it would appear that only limited fiscal adjustment would be needed to avoid pushing the government debt-to-GDP ratio on to an unsustainable path. However, this conclusion is based on the important assumptions of continued rapid economic growth and no incurrence of new quasi-fiscal liabilities by the government.

At the same time, the government has substantial assets, which it could sell to meet some of its quasi-fiscal liabilities (Box 5.1). In fact, part of the remaining state shares in some of the publicly listed companies have already been earmarked to fund pension liabilities. Privatization of other SOEs would also allow the government to raise funds to offset its liabilities. However, the value of the state assets is highly uncertain; valuations for China’s state equity share in the SOE sector vary greatly, from 25 percent to 100 percent of GDP. While this suggests that the government does have valuable assets, their divestiture may not raise significant amounts in domestic markets that are not deep and have a limited capacity to absorb a substantial volume of sales. The eventual realization value for these assets will depend crucially on strengthening management and implementing reforms in SOEs ahead of actual divestitures.

Box 5.1.An Assessment of China’s State Equity Share in SOEs

In China, the state has an ownership share in over 150,000 state-owned enterprises (SOEs), of which more than 1,000 are listed on stock exchanges in China and overseas. The value of these assets has important implications for fiscal sustainability, as proceeds from some asset sales are already designated to partially fund government pensions and future sales could contribute also to covering additional government obligations.

It is difficult, however, to estimate the value of these assets, due to lack of comprehensive and independently verified financial information on SOEs. Nevertheless, illustrative estimates can be derived using balance sheet and income data for the SOEs and various stock market indicators. These estimates are sensitive to key assumptions and yield a wide range of results on the net value of state equity share in the SOEs, from zero to 100 percent of GDP. Other studies suggest valuations ranging from 25 percent to 65 percent of GDP (see, e.g., Studwell, 2000; and Bottelier, 2002). While the wide range of these estimates highlights the uncertainties, it suggests that the state probably has sizable assets.

Assessing the State’s Stake in SOEs

Various methods can be used to estimate the value of the state’s equity stake in SOEs. Book value data suggest that SOEs are worth about 50–75 percent of GDP. The upper bound estimate is based on Ministry of Finance data, showing a book value of the state’s equity stake in nonfinancial enterprises of about 65 percent of GDP in recent years, with an additional 10 percent of GDP equity in financial and overseas enterprises. However, nonfinancial enterprises reported that “unhealthy assets” (including actual and potential asset write-downs, such as delayed receivables, delayed depreciation, and excess inventories) were equivalent to almost one-third of equity. Therefore, adjusting for “unhealthy assets” gives a lower bound estimate of about 50 percent of GDP.

Book value estimates may, however, understate the market value of the SOE assets. In other emerging market economies, the market value of companies listed on the stock exchanges has generally been larger than their book value. For 19 emerging market economies, excluding China, the ratio of the market to book value was 1.6 in 2001 (see Tenev, Zhang, and Brefort, 2002). Applying this average ratio to the adjusted book value of China’s SOE assets gives an estimated market value of about 80 percent of GDP.

In contrast, some studies suggest that China’s SOEs may overvalue their assets and understate their liabilities, implying a much lower book value for the SOEs. Lardy (1998) and Steinfeld (2000) have argued that SOEs overvalue fixed assets and inventories by using very low depreciation rates and not adjusting the value of excessive inventories to fully reflect market prices. Moreover, exclusion of pension obligations and other contingencies would understate liabilities. For example, a downward adjustment in the book value of gross assets by one-third would wipe out the equity stake.

An alternative approach is to apply different price/earnings (P/E) ratios of Chinese listed companies to profits of SOEs. Assuming a P/E ratio of 20 (about the level used for initial public offerings of SOE shares issued on domestic stock markets in recent years) for listed and unlisted SOEs gives an estimated equity stake of about 50 percent of GDP, based on 2002 data. Assuming a P/E ratio of 15, in line with that prevailing for mainland companies that are listed on the Hong Kong SAR stock exchange, gives an estimate of 35 percent of GDP, while using the P/E ratio prevailing on Shanghai and Shenzhen stock markets of 40 gives an estimate of 100 percent of GDP. The latter estimate, however, probably overstates the state’s equity share since (1) the P/E ratio on the mainland is artificially high because capital controls restrict offshore investment by domestic investors; (2) less than one-third of shares in individual listed companies are tradable, since the state holds the remaining equity; and (3) evidence from a Ministry of Finance survey suggests that more than half of the SOEs covered had overstated their profits by 10 percent or more (Ministry of Finance, 2002).

Using the price/sales ratios prevailing in other stock markets provides another method to estimate the state’s equity share in the SOEs. For example, the price/sales ratio for H-shares and Red Chips in Hong Kong SAR of about 1.25 gives a valuation for SOEs of about 100 percent of GDP. This estimate would be lowered to about 70 percent of GDP if the price/sales ratio from emerging market economies is used. But these estimates may also overstate the value of unlisted SOEs, as these tend to be relatively unprofitable companies. The profits (before interest) of listed companies were around 10 percent of sales in 2002, almost three times those of unlisted companies.

Implications of State Assets for Fiscal Policy

The divestiture of state assets may provide significant funding for liabilities, but their valuation is uncertain and subject to a multitude of claims. In particular, while the book value of equity in SOEs is about evenly split between the central and local governments, the central government carries the responsibility for most of the explicit and contingent liabilities. Local governments have an incentive to dispose of assets and use the funds for their own purposes, rather than for funding central government liabilities. In addition, other parties could lay claim to funds generated from sales of shares in SOEs, including companies themselves wanting investment capital, and redundant and retired workers wanting payouts. Furthermore, the ability of Chinese financial markets to absorb large amounts from asset sales in a short period may be limited. All these factors suggest that the amount available to fund central government liabilities may be considerably less than the total net worth of state assets.

The establishment of the State Assets Supervision and Administration Commission in March 2003 should improve the management and oversight of nonfinancial SOEs. In particular, increased public disclosure of their accounts and the state’s equity claim would also serve to improve public accountability for the management of these assets.

Prepared by Ray Brooks.


China’s strong growth performance, the strength and potential of its domestic market, the availability of considerable domestic savings, and the large endowment of government assets are some of the elements that make its fiscal position, current and prospective, relatively strong. However, these elements do not provide sufficient reasons to avoid addressing past and incipient liabilities that may ultimately fall on the central government. China’s current position of relative strength provides a unique opportunity to start addressing some of these liabilities and create room for their payment, if and when needed. The authorities’ intention to adhere to a path of gradual fiscal consolidation over the medium term is a step in the right direction.

1China’s external debt, at about 5 percent of GDP, compares favorably to the average for emerging market countries of about 36 percent of GDP. The average in emerging Asia is lower, at 24 percent of GDP—but still considerably higher than in China. However, coverage of debt statistics varies across countries. In China, government debt is mostly domestic and denominated in local currency. It includes liabilities of the central government; local governments are not allowed to borrow by law, although they can be recipients of onlending from the center.
2The expansion, however, has also been affected by the inclusion in the budget of previously off-budgeted revenue and expenditure.
3Fiscal data in China remain subject to limitations. Budgetary data exclude spending associated with onlending to local governments, both domestic and external. Data on social and extrabudgetary funds are provided annually and with a long lag. Expenditure is classified by function, and no economic classification is available.
4The tax-for-fees reform was initiated in 2000 as a measure to boost incomes in rural areas. It aims at the elimination of the many unofficial fees levied by local governments while remaining fees are being converted into taxes and subject to caps mandated by the center.
5“Estimated” refers to official estimate, modified for some adjustments applied by the IMF staff, including onlending to local governments, external borrowing excluded from the budget, and subsidies to SOEs.
6The current budget classification, which is being upgraded, does not provide an accurate description of expenditure by economic and functional types. Hence, it is difficult to explain past trends in expenditure programs.
7A standard treasury single account (TSA) is a bank account or a set of linked bank accounts through which the government, including its entities and spending units, transacts all receipts and payments, and consolidates its cash balances. Cash balances held by a government are efficiently centralized through a TSA.
8A contingent liability is an obligation triggered by a discrete but uncertain event. For example, a loan guaranteed by the central government becomes the latter’s explicit obligation when the guarantee is called.
9There may also be pressure on the government to protect depositors, regardless of the ownership of banks, in the absence of a deposit insurance scheme.
10For example, the Guangdong Trust and Investment Corporation was declared bankrupt and closed without a bailout of most of its creditors in 1999. Nonetheless, the provincial government decided to repay individual depositors, although covering only a very small amount of their claims.
11For example, China spends about 3 percent of GDP on education. This is slightly below the average for Asian countries (excluding Japan) at 3.4 percent of GDP, but compares less favorably with higher levels observed among some other countries in the region—for example, in India (above 4 percent of GDP), Thailand (above 5 percent of GDP), and Malaysia (above 6 percent of GDP). In OECD countries, the average is 6 percent of GDP (OECD, 2002b).

    Other Resources Citing This Publication