Japan: Economic and Policy Developments

This paper provides a brief overview of the causes of the poor economic performance of Japan in the 1990s, and a more detailed analysis of developments in the real sector during 1999. The paper highlights that the collapse of the asset price bubble in 1990–91 provided the trigger for the downturn in 1992, and compounded the economic problems thereafter through its effects on the banking system. This paper also analyzes the fiscal policy developments and the monetary developments in Japan.


This paper provides a brief overview of the causes of the poor economic performance of Japan in the 1990s, and a more detailed analysis of developments in the real sector during 1999. The paper highlights that the collapse of the asset price bubble in 1990–91 provided the trigger for the downturn in 1992, and compounded the economic problems thereafter through its effects on the banking system. This paper also analyzes the fiscal policy developments and the monetary developments in Japan.

II. Fiscal Policy Developments1

1. Fiscal policy has been expansionary for most of the past decade. During the mid-1990s, the general government deficit expanded as public works spending and tax reductions were used to support aggregate demand in the face of an unprecedented economic slowdown. In FY1997, policy was temporarily tightened in response to an apparent recovery in economic activity. However, the withdrawal of stimulus was associated with a renewed slowdown and the policy stance was loosened again in late 1997. Two large stimulus packages in April and November 1998 have subsequently led to a substantial widening of the deficit, although such assessments continue to be hampered by limited fiscal transparency. The initial budget for FY1999 has provided further stimulus to the economy and the general government deficit excluding social security is expected to reach close to 10 percent of GDP.

2. The expansionary policy stance has put increasing strains on public finances. Concerns have been raised about the sustainability of fiscal policy as the net debt of the general government (excluding social security) is set to approach 90 percent of GDP by the end of FY1999, Current strains are most evident in the difficult financial situation of local governments that increasingly hampers the implementation of fiscal stimulus measures—the bulk of which (although planned at the national level) are carried out by local authorities—and has led to higher funding demands on the Fiscal Investment and Loan Program (FILP). Faced by increasing loan requests and a prospective weakening in its deposit base, the Ministry of Finance’s Trust Fund Bureau (the administrator of the FILP) announced a partial withdrawal from markets for Japanese government bonds, which has contributed to an increase in long-term yields since late 1998,

3. These issues will be discussed in greater detail in this chapter. The following sections focus on: (i) recent stimulus policies; (ii) the financial situation of local governments; (iii) the interaction between FILP operations and government bond markets; and (iv) fiscal transparency.

A. Recent Stimulus Policies2

4. The general government’s fiscal deficit expanded sharply during the early-1990s. As growth slowed at the beginning of the decade, cyclical factors, a sharp decline in tax elasticity, and a series of stimulus packages together shifted the general government’s fiscal position (excluding social security) from balance in FY1990 to a 6½ percent of GDP deficit in FY1996. The general government’s structural deficit—a measure that excludes cyclical factors—increased from 1 percent of GDP to 7 percent of GDP over the same period, and net general government debt (excluding social security) grew by half to 60 percent of GDP (Figure II.1).



Citation: IMF Staff Country Reports 1999, 114; 10.5089/9781451820584.002.A002

Sources: Ministry of Finance; Economic Planning Agency; and staff estimates and projections.1/ The fiscal year is from April to March.

5. The government’s main policy tools were a series of large mid-year stimulus programs (Table II.1). These packages sought to stimulate economic activity through a variety of measures across all layers of the public sector. Their most important elements have been public works spending and tax cuts financed by general government bond issues (so-called “real water” measures) but the packages have also included other elements, such as financial support and asset transactions:3

  • Public works (consisting mainly of infrastructure-related construction projects) have been a central constituent of stimulus packages, given the swift impact of a rise in public orders on employment and incomes in the construction sector.4 Total public works projects included in the stimulus packages cumulated to about 7 percent of GDP between 1990 and 1996. Many of these measures served to maintain rather than increase investment levels, but the share of public investment in GDP still rose from 6½ percent to above 8 percent over this period.

  • The rising budget deficit also reflected a drop in tax revenue, particularly in the immediate post-bubble years. The revenue losses were partly caused by a series of tax cuts since 1994, including an upward shift in income tax brackets in 1994 and several (often temporary) income tax cuts in subsequent years. The strongest revenue impact, however, originated from a sharp decline in tax elasticity between 1991 and 1994 that appears to be related to income and wealth losses in the aftermath of the bubble years.

Table II.1.

Japan: Summary of Economic Stimulus Packages, 1993–98

(In trillions of yen, unless otherwise indicated)

article image
Sources: Data provided by the Japanese authorities; and staff estimates.

Includes ¥0.3 trillion in welfare benefits.

Later increased to ¥9.3 trillion (1.9 percent of GDP).

Public investment comprises general public works (including land purchases), disaster reconstruction, buildings and equipment, and independent public works projects by local government.

Excludes land acquisition for public works projects, which is included in public works spending.

Includes loans by the Pension Welfare Service Public Corporation.

Includes ¥1.3 trillion in lending by the Japan Corporation for small business.

6. As growth picked up in 1997, concerns over the deteriorating position of public finances prompted a tightening of fiscal policies. Public investment fell sharply as the effect of previous packages wore off, and tax revenue strengthened as a consequence of an earlier planned increase in the consumption tax rate from 3 to 5 percent in April 1997 and a phase-out of temporary income tax cuts, Moreover, the Diet passed the Fiscal Structural Reform Act (FSRA) that committed the government to lowering the general government deficit (excluding social security) to 3 percent of GDP by the end of FY2003. Prospects of a continued tight fiscal policy, however, together with the immediate dampening effect of the consumption tax increase contributed to the subsequent decline in activity.

The 1998 measures

7. With the economy unexpectedly entering into recession in 1997, the government came under renewed pressure to stimulate activity through fiscal measures. The authorities announced some ¥3½ trillion in income tax rebates and other tax cuts in December 1997 (including a 3 percentage point reduction in the corporate tax rate to 46 percent), although the initial budget for FY1998 was still mildly contractionary. However, as the economy continued to deteriorate, two massive stimulus programs with a combined “headline” figure of ¥40 trillion (8 percent of GDP, see Table II.1) were compiled in the course of 1998, aiming at returning the economy to positive growth in 1999.

8. In April 1998, a ¥16,7 trillion stimulus package was announced immediately after the passage of the FY1998 initial budget, including ¥12 trillion (2½ percent of GDP) in “real water” measures. The measures included:

  • Additional public works spending of ¥7.7 trillion, including ¥1.5 trillion in public works autonomously financed by local governments.

  • Temporary income tax cuts of ¥4.3 trillion.

  • ¥4.7 trillion in additional measures, including ¥2.3 trillion in spending to support the real estate market (largely earmarked for land acquisition) and ¥2 trillion in additional lending to small and medium-sized enterprises through the FILP.

9. A second stimulus package was compiled in November 1998 in conjunction with preparations for the 1999 budget. The headline figure was ¥23.9 trillion, with about ¥15 trillion (3 percent of GDP) in “real water” measures, making it the largest-ever such package. The components included:

  • Additional public works spending of ¥8.1 trillion, largely focused on urban projects such as telecommunications (all folly or partly financed by the central government).

  • A commitment to cut income tax rates by the equivalent of ¥6 trillion in FY1999.

  • Job creation measures worth ¥1 trillion, including subsidies for employment creation and retraining schemes, and an extension of unemployment benefits.

  • Spending vouchers of ¥700 billion distributed to families with children and the elderly.

  • Additional measures to ease bank lending constraints worth ¥5.9 trillion, including increased funding for government financial institutions.

  • Support for the rest of Asia of ¥1 trillion as part of the previously announced Miyazawa initiative.

10. At the same time, the FSRA was suspended, pending a sustained recovery of the economy. The strict timetable for consolidation contained in the FSRA had already been loosened in the face of economic weakness. In April 1998, the deadline for achieving medium-term consolidation targets was moved from FY2003 to FY2005, and the FSRA law was also amended to allow limits on government debt issues to be waived in times of weak economic growth (as defined by several possible criteria, including annualized real GDP growth being lower than 1 percent for two consecutive quarters).

11. The central government’s share of the 1998 stimulus packages was incorporated in two supplementary budgets for FY1998. Total central government spending on stimulus measures amounted to ¥13½ trillion, of which ¥1 trillion was offset by savings elsewhere. With tax revenues falling short of initial budget targets by almost ¥6 trillion (net of local allocation tax payments to local governments), the revised bond issue for FY1998 amounted to ¥34 trillion (6.8 percent of GDP), compared to an initial plan of ¥15½ trillion (Tables II.2 and II.3).

Table II.2.

Japan: Tax Receipts of the Central Government General Account, FY1994–99

article image
Source: Data provided by the Japanese authorities.

Percentage changes calculated relative to most recent data of previous year.

Table II.3.

Japan: Central Government General Account Budget, FY1994–99

(In billions of yen, fiscal years)

article image
Source: Data provided by the Japanese authorities.

Includes repayments of principal and running costs.

12. Measures to restore the banking system and boost credit supply have received indirect budget support. A separate supplementary budget, passed in October 1998, authorized bond issues and public guarantees worth ¥60 trillion (12 percent of GDP) for resolving problems in the domestic banking system (see Chapter IV). Government outlays only occur, however, at the time such guarantees are being invoked, e.g., in the case of financial losses experienced by the Deposit Insurance Corporation.5 The government also authorized ¥ 20 trillion in additional loan guarantees extended by regional credit-guarantee associations, aimed largely at alleviating credit constraints for small and medium-sized enterprises (SMEs), which has since been lauded for bringing about a strong turnaround in SME activity. The government has agreed to cover any losses arising from the guarantees, which are currently anticipated at around ¥1 trillion.

The FY1999 budget

13. For the 1999 budget, the authorities announced tax reductions that went somewhat beyond the income tax cuts announced in the stimulus packages. The final tax package featured ¥9 trillion in tax reductions, of which ¥7 trillion (1.4 percent of GDP) were to become effective in FY1999. The impact of the remaining ¥2 trillion (largely related to more generous tax treatment of mortgage payments) will be spread over a period of 15 years. In addition to the personal and corporate income tax cuts included in the stimulus packages, further measures were intended to promote residential and corporate investment. The authorities also announced that they would consider shifting corporate taxation to a consolidated basis:

  • The personal income tax cuts consisted of a permanent reduction in the top marginal tax rate (central plus local) from 65 percent to 50 percent, and a reduction of tax liabilities of close to 20 percent across the board. These measures implied ¥4 trillion in income tax relief, effectively replacing the expiring temporary rebates of FY1998. Compared to FY1998, the tax cut provided some tax relief for higher income brackets, but the effect on average households was marginal.

  • The corporate tax rate (central plus local) was reduced from 46 percent to 40 percent, resulting in a net tax relief of ¥2.3 trillion, and there were additional cuts in tax rates on SMEs.

  • To promote housing investment, the period for which mortgage holders qualify for tax deductions was extended from six to 15 years (for house purchases taking place in the next two years), and the existing ceiling on the deduction was trebled. However, the rate of deduction was reduced from 2 percent to 1 percent, so the net impact provided for only moderately relief.

  • To enhance the attractiveness of Japanese financial markets, the repeal of the securities transactions tax was advanced from December 1999 to April 1999, withholding taxes were suspended for Finance Bills and Treasury Bills (if registered with the Bank of Japan), and nonresidents will be exempted from withholding tax on government bonds from September 1999.

14. Public works spending is expected to increase strongly in FY1999. Besides the tax cuts, the budget incorporated other elements of the November stimulus package, including about ¥1 trillion in additional public investment. Nevertheless, the public works allocation of ¥9½ trillion in the initial FY1999 budget is considerably smaller than in the revised budget for FY1999 (¥15 trillion). However, taking account of funds appropriated in the FY1998 supplementary budgets that are likely to be disbursed in FY1999, the authorities have indicated that public works spending is expected to rise by 10 percent in FY1999.6 Moreover, to ensure a high level of public investment throughout the fiscal year, the government decided in March 1999 to front-load public works spending, with the aim of minimizing the carry-over of public works projects into FY2000.7

15. The budget included a substantial increase in central government transfers to the social security system. The 8½ percent increase reflected a decision to raise government contributions to the basic pension fund (a relatively small component of the overall pension system) from the current one-third of benefits to one-half by FY2004. The government also decided to postpone an increase in contribution rates (from 17.35 to 19,5 percent of base income) to the Employee Pension Insurance—an earnings-related pension scheme for private sector workers—from 1999 to 2004, prompted by concerns over its impact on take-home pay. The foregone increase in pension contributions is estimated at around ½ percent of GDP per year.

16. As a result of the budget proposals, the staff projects that, on a national accounts basis, the general government deficit (excluding social security) is to rise from 8.3 percent of GDP in FY1998 to 9.7 percent in FY1999 (Table II.4). Official projections show that the general government deficit (excluding social security) will fall from 9.8 percent of GDP in FY1998 to 9.2 percent of GDP in FY1999.8 However, these numbers are based on appropriations data and do not fully adjust for the fact that much of the related public works spending will not occur until FY1999. The structural deficit (including social security) is projected to increase by about 1 percentage point of GDP.9

Table II.4.

Japan—General Government Operations, 1996–1999

(in percent of GDP)

article image
Source: Staff estimates.

17. The budget implies a net bond issue of ¥31 trillion in FY1999, contributing to the sizeable accumulation of public debt in recent years. Japan has become the largest issuer of net debt among industrial countries as net bond issues have increased five-fold since 1990 (Table II.5). As a result, the amount of outstanding central government bonds has roughly doubled, reaching an estimated ¥310 trillion (63 percent of GDP) by the end of FY1998. Net general government debt (excluding social security) is expected to more than double from 42 percent of GDP in 1990 to an estimated 90 percent of GDP at end-FY1999.10

Table II.5.

Japan—Central Government Bond Issues, FY1990–1999

(in trillions of yen)

article image
Source: Bank of Japan, Economic Statistics Monthly; and staff calculations.

Including subsidy, subscription, and DIC bonds, and bonds converted from JNRSC bonds.

Staff estimate.

Budget projection.

B. The Financial Situation of Local Governments

18. Local governments have supported the expansionary fiscal stance of the central government in recent years. Reflecting the close relationship between central and local governments in Japan, local authorities have taken on a major burden in efforts to stimulate the economy (Box II.1). The central government’s stimulus programs have affected local governments mainly in two ways. First, the implementation of public works occurs to a large extent at the local government level (some 80 percent of all general government public works are carried out by local authorities), and most stimulus programs have contained a substantial share of projects to be financed by local governments, either independently or jointly with the central government (Table II.6). Second, tax cuts have affected local budgets both through losses in shared taxes and reductions in local taxes.

Table II.6

Japan: General Government Public Works Projects

(In billions of yen)

article image
Source: Ministry of Finance; Local Government White Paper, various issues; and staff calculations.

Calculated as total central government expenditure on public works minus central share of joint public works.

Joint projects are implemented by local governments.

19. As a result, the aggregate fiscal deficit of local governments roughly doubled between FY1990 and FY1997. On the basis of settlement data (available through FY1997), the consolidated local government deficit rose sharply from 1 percent of GDP in FY1990 to 3 percent of GDP in FY1995, before declining to 2 percent of GDP in FY1997 (Table II.7):11

  • The deficit increase was driven by expenditure developments, the most important being a rise in current spending, driven by personnel and interest cost. There was also a sharp increase in public investment in the wake of a series of stimulus packages between 1992 and 1995. Subsequently, however, public works spending was cut back by a full percentage point of GDP within two years, and savings were also achieved in other discretionary spending items.

  • Overall revenue was largely flat, but local tax revenues dropped by almost one percent of GDP in the early 1990s before recovering moderately through 1997 (partly a consequence of the consumption tax increase, the revenues of which were evenly split between central and local government). Tax transfers from the central government remained practically flat, but other revenues (including user fees and other central government grants) largely offset the swings in own tax revenue.

Japan—Central-Local Government Relations1

Financial relations between the national and local governments in Japan are marked by a substantial vertical imbalance, i.e., revenues generally exceed expenditures at the central level and fall short at the local level. The central government receives 60 percent of all total tax revenue but accounts for only 35 percent of total general government expenditure (excluding social security). By contrast, the tax revenue of local governments amounts to less than 40 percent of the funds necessary to perform their functions. To fill the financing gap at the local level, the central government transfers part of its tax revenue to local governments (mainly in the form of “local allocation taxes”, which is distributed according to formulas that seek to maintain uniform living standards across the country) but is also subject to adjustment each year to match local government needs and resources. Overall, these tax transfers amounted to 22 percent of local revenue in FY1997. Prefectures and municipalities raise the remaining funds through nontax revenue and borrowing.

Although considerably more public spending takes place at the local level than at the national level, the national government has been closely involved in most aspects of local public spending. For example, major programs in education, health, or infrastructure were in practice formulated by national ministries and financed directly or indirectly by the central government. The recently enacted Decentralization Law has provided local governments with greater planning autonomy, but local governments continue to be financially dependent on the central government.

The intense involvement of the central government makes the budgets of central and local governments much more closely linked than in other industrial countries. The vertical imbalance is resolved through grants and borrowing linking the central and local budgets in what is effectively a single budget entity. Each year, the central government produces an official estimate of the revenues and expenditures of local governments (including planned amounts of local allocation tax, transfers, and local borrowing). These estimates are combined with an array of national government policies and programs such as the FILP in order to compile income-expenditure plans for each local government, which are combined in an indicative Local Government Finance Plan. The most important feature of this process is how revenues and expenditures are balanced through the local allocation tax and local borrowing. When the program is approved by the Diet, it effectively locks in local governments in terms of their own local taxes, loan programs, employment and general administrative expenses and, most importantly, the levels of funding for almost all public programs, leaving independent public works spending as the main area where discretionary changes can be made.

In recent years, there has been an increase in local borrowing to assist in the implementation of the central government’s fiscal policy. Given the large relative size of the local public sector, the central government relies on local governments to implement parts of its fiscal policy. It is impracticable to alter the local tax system or the local allocation tax to compensate for changes arising out cyclical fluctuations since these—and the proportion of national taxes allocated to local governments—are fixed by statutes, which can be amended only by the Diet. The size of local borrowing, on the other hand, can be changed by executive action. In view of this flexibility, the proportion of bonds and loans in local revenue tends to increase in years when the national government has adopted an expansionary fiscal policy.

However, local borrowing is increasingly subject to statutory limitations. Local authorities risk losing their remaining financial independence if their borrowing or debt levels exceed thresholds established in the Local Fiscal Restructuring Law. Two fiscal indicators in particular trigger central government intervention: (i) a local authority’s bond issues are restricted if its ratio of debt service to local tax revenues exceeds 20 percent; and (ii) a prefecture is mandated to undertake fiscal restructuring under direct national control if its fiscal deficit exceeds 5 percent of a standardized expenditure measure (20 percent in the case of municipalities). Some heavily indebted local authorities have indeed come close to these limits in recent years, and have embarked on severe austerity programs to avoid falling under the central government’s authority.

1This box summarizes and updates material in Mihaljek (1996).
Table II.7.

Japan: Local Government Operations

(in percent of GDP)

article image
Source: Ministry of Home Affairs; Ministry of Finance; and staff calculations.

Using data provided by the authorities, the staff has reclassified financing items which are treated as revenues in the official statistics.

Including classification errors and omissions.

20. The deficit increase has also reflected specific local factors. Although fiscal expansion at the local level has mainly reflected stimulus policies at the national level, other developments have added to the financial burden on local governments:

  • Social welfare spending at the local level has increased in response to an ageing population. Future spending is expected to accelerate with the introduction of a planned nursing insurance scheme in April 2000 that is expected to cost local governments a total of ¥1 trillion (0.2 percent of GDP) in FY2000.

  • Local governments are struggling with devalued land assets and the associated rise in interest payments on purchases of real estate made in the late 1980s. Indeed, the majority of real estate developments undertaken by local governments in cooperation with the private sector during the 1980s (so-called “third-sector” companies) have turned unprofitable and are saddled with potentially large liabilities (Okue 1999a).

  • A number of agricultural cooperatives are heavily indebted and burdened with non-performing assets, requiring financial assistance to avoid bankruptcy.

  • In certain instances, natural disasters have added to fiscal difficulties—e.g., the Kobe city government expects to use half of its tax revenues in FY1999 to repay loans used to provide shelter for the survivors of the 1995 earthquake.

21. The financial situation of local authorities has deteriorated. Tax losses and rising expenditure have been financed largely through increased local bond issuance, and local government indebtedness has correspondingly shot up from 15 percent of GDP in 1990 to 30 percent of GDP by the end of FY1997. This has already contributed to a rise in debt service payments to over 2 percent of GDP in 1997, up from 1½ percent of GDP in 1990.

22. Some large municipalities and urban prefectures have found themselves in a particularly severe financial state. Revenues of local authorities in urban areas—which depend heavily on local corporation taxes—have been most affected by the decline in economic activity. With many businesses reporting losses, the tax base for these local governments has been shrinking rapidly in recent years. Examples for local governments with financial difficulties include a number of urban prefectures (e.g., Tokyo, Osaka) which have been approaching their bond issuance and deficit limits and have now begun to contemplate drastic steps toward fiscal consolidation (see below).

Developments since 1998

23. Financial difficulties have intensified since early 1998. The indicative Local Government Finance Plan for FY1998 foresaw a slight reduction in the aggregate local government deficit, expected to be achieved by a substantial increase in own tax revenue (½ percent of GDP) buoyed by a growing economy, which was to offset rising spending on debt service and personnel costs (see Table II.7). However, the economy fell into recession, leading to weaker than anticipated revenue growth, which was exacerbated by temporary tax cuts contained in the April 1998 stimulus package. Preliminary estimates indicate that local tax revenue in FY1998 was roughly unchanged from the year before, falling short of initial projections by about ¥2½ trillion (½ percent of GDP), of which slightly more than half was accounted for by the tax cuts. A further revenue decline has been projected for FY1999, largely in response to tax cuts contained in the November 1998 stimulus package.

24. The government has taken steps to alleviate the financial burden on local governments, in particular through the FILP. While the bulk of tax allocation grants are financed by central government taxes, additional funds for local governments have been provided through loans from the Trust Fund Bureau (TFB) to the Special Account for Local Allocation Tax (Box II.2). These loans, which should in theory net out over the cycle, have averaged around ¥4 trillion in recent years. For FY1999, owing to the anticipated decline in local tax revenues, an exceptional ¥8½ trillion (1.7 percent of GDP) in TFB loans are planned. As can be seen from Box II.2, debts to the TFB have mushroomed in recent years to around ¥30 trillion (6 percent of GDP) by end-FY1999.

TFB Loans for Local Allocation Tax Grants

The Japanese fiscal system provides for the distribution of shared tax revenue through the Special Account for Local Allocation Tax. The amounts to be distributed are agreed upon during the budget process and become part of the initial budget (see Box II.1). To provide a stable revenue base for local governments, revenue shortfalls are bridged by short-term loans from the Ministry of Finance’s Trust Fund Bureau (TFB) to the Special Account, using funds that could otherwise be invested through the FILP,1 However, in view of the difficult financial situation of local governments, TFB loans have been part of the initial budget in recent years. The loans have a notional maturity of one month (but are rolled over), and carry an interest rate premium of 20 basis points over 10-year JGB yields. Outstanding loans are to be repaid through future surpluses in the Special Account.

TFB Loans to the Special Account for Local Allocation Tax

(in trillions of yen)

article image
Source: Ministry of Finance.
1TFB financing for local governments falls into two categories: (i) loans with a maturity over 5 years have to be included in the FILP budget and can only be extended to finance specific investment projects; and (ii) special account loans with shorter maturities are not part of the FILP and can thus be used more flexibly. Both forms of loans, however, require Diet approval.

25. Despite this support, local governments are cutting back on discretionary spending. The integrated nature of central and local government financial planning leaves local authorities with limited room for adjustments in current expenditure (see Box II. 1). Many local governments have therefore cut back on own-account investment projects, i.e., projects that are financed independently of the central government (which account for almost 60 percent of all local public works). Partly in response to these problems, the November 1998 stimulus package differed from previous packages in that it did not include increases in independent local public works (by contrast, the April 1998 package included ¥1.5 trillion of self-financed local investment). For FY1999, preliminary budget data for 25 prefectures show a decline in locally funded public works by 16 percent compared to FY1998 initial budgets, although the Local Government Finance Plan assumed that such spending would be flat (Okue 1999b).12 A recent budget survey of 670 municipal governments also found cuts in self-financed investment projects.

26. Governments in urban areas—which receive only small shares of local allocation tax—are under particular pressure to implement adjustment measures. With local allocation tax payments intended to minimize differences in living standards, urban prefectures generally receive smaller tax transfers than their rural counterparts (the Tokyo prefecture, as an extreme case, receives no local allocation tax transfers at all). Some urban authorities have thus been forced to draw up particularly strong fiscal adjustment programs, including politically difficult measures such as the delay, reduction or suspension of pay rises and bonus payments, sharp increases in smaller local taxes and-user fees, and sales of land or other assets.

C. FILP Operations and Government Bond Markets

27. The FILP has evolved into an important tool for financial management and quasi-fiscal operations. About one third of Japanese savings are held by public institutions, including the postal savings and public pension systems. Most of these funds are invested through the FILP, which is run by the Ministry of Finance’s Trust Fund Bureau (TFB). Although the FILP is not formally part of the general government sector, its annual investment plan is formulated in close coordination with the budget process, and submitted to the Diet together with the regular government budget. Indeed, owing to its size, the FILP is often referred to as the “second budget”.

28. The bulk of the FILP is used to finance long-term loans for public sector projects. Three-quarters of the TFB’s portfolio—which currently amounts to some ¥440 trillion (90 percent of GDP)—consist of loans provided to the central government (mainly financing operations of special accounts), local governments, and government financial institutions (GFIs) and other public agencies (Table II.8). The remainder is invested in government securities, mainly long-term government bonds. A small portion of TFB funds is also invested in short-term assets, e.g., financing bills and loans provided to special accounts.

Table II.8.

Japan: Trust Fund Bureau Operations

(In trillions of yen)

article image
Source: Ministry of Finance.

29. Demands for FILP funding have been growing. Measures announced as part of the 1998 stimulus packages have included substantial additional FILP loans, mainly to improve credit conditions in the economy and extend financial support to local governments (Table II.9):

  • The revised FILP budget for FY1998 provided an additional ¥3 trillion in funding for GFIs (in particular the Japan Finance Corporation which lends to SMEs), and ¥2½ trillion in additional long-term loans to local governments.

  • Excluding government bond purchases, FILP spending in the initial FY1999 budget is planned to shrink by ¥7½ trillion compared to the revised FY1998 budget, although still exceeding the initial FY1998 budget. The high level of loans to GFIs has been maintained, implying a 15 percent increase over the previous year’s initial budget. Moreover, while loans to local government were kept flat on an initial budget basis, short-term loans to the Special Account for Local Allocation Tax are to rise sharply (see Box II.2).

Table II.9.

Japan: Fiscal Investment and Loan Program (FILP), FY 1994–99

(In billions of yen, fiscal years)

article image
Source: Ministry of Finance.

Reflects the funding of the “lend-back” system under which the postal savings system, public pension funds, and the postal life insurance fund receive funds for portfolio management on their own account.

Excludes portfolio investment.

Compared with revised plan of the previous year.

30. At the same time, the TFB’s deposit base is likely to decline. In FY1998, net deposits with the TFB registered the smallest increase in many years, amounting to only ¥16 trillion compared with inflows of well over ¥20 trillion for the past decade (see Table II.8). While this may have reflected the particularly weak situation of the economy, TFB deposits are expected to shrink in absolute terms in the coming years:

  • About ¥100 trillion in fixed-term postal saving (teigaku) deposits from the high-interest period in 1990–91 are estimated to mature by 2001. At current low interest rates, a considerable portion of these funds is likely to be invested outside the postal savings system and will thus no longer be available to the TFB.

  • Current reform plans, in the context of financial market liberalization, imply that the TFB will no longer receive automatic deposits from either the postal savings system or public pension funds from FY2001, implying a fundamental change in the FILP’s funding pattern.

  • On the pension side, the postponement of the planned October 1999 EPI contribution increase, combined with strong anticipated growth in benefit payments, will reduce the surplus social security funds available for deposit with the TFB.

31. Anticipating these changes, the TFB decided in late 1998 to shift its securities portfolio toward shorter-term maturities, including through a halt of long-term government bond (JGB) purchases. Faced by strong loan demands and weakening deposit flows, the government announced in December 1998 (at the time of the budget presentation) that it would no longer acquire 10-year JGBs either through underwriting or in secondary markets. Instead, the TFB planned to step up investments in government securities with shorter maturities, thus achieving a more balanced maturity structure of its securities portfolio.13

32. In a related move, the government announced its intention to modify its debt management strategy. To accommodate the liquidity needs of the TFB and other major market players, the government decided to issue more 1-year Treasury bills and 2–6 year bonds at the expense of long-term bonds. Since the beginning of FY1999, the government has thus issued only ¥1.4 trillion in 10-year JGBs per month (instead of a planned ¥1.8 trillion), while ¥400 billion per month has been issued at shorter maturities instead. An eventual move to consolidate maturities of medium-term bonds around the 5-year level is also being contemplated.

The reaction of the bond market

33. The benchmark yield on 10-year Japanese Government Bonds (JGBs) was at an historic low of 0.7 percent in October 1998 (Figure II.2). JGB yields had fallen through most of the 1990s, with the mid-1998 level being a record low in Japan and a yield unprecedented anywhere in the historical record. During the first half of the decade, when inflation had declined, real JGB yields remained in line with international yields.14 However, with inflation hovering around zero since 1997, real yields also dropped to unprecedentedly low levels (see lower panel of Figure II.2). Indeed, with nominal yields so close to the floor of zero, the potential benefits from holding JGBs appeared meager compared to the downside risk of a correction in prices. The main reason for the low bond yield appears to have been large purchases of JGBs by public institutions (see below).



Citation: IMF Staff Country Reports 1999, 114; 10.5089/9781451820584.002.A002

Sources: Nikkei Telecom, WEFA, and staff estimates.1/ Japan: 10-year JGB yield; U.S.: 10-year Treasury Bond. Real rates were defined as nominal rates minus 12-month percent changes in CPI. Japan’s CPI was adjusted for consumption tax increase in April 1997.

34. Subsequently, news that the TFB—the largest JGB holder and a major purchaser in recent years—would withdraw from bond underwriting led to a surge in yields following the release of the FY1999 budget.15 Bond yields had already begun to rise earlier, reaching 1¼ percent by mid-December, after the November 1998 stimulus package increased the planned JGB issue for FY1998 by 60 percent. However, a subsequent yield surge of 50 basis points in late December was mainly associated with the TFB’s decision to stop underwriting in the face of an expansionary FY1999 budget.

35. The yield increase since last November marks the end of a period of large bond purchases by public sector institutions that, combined with investor flight-to-quality, pushed down Japanese bond yields to unprecedentedly low levels. Much of the recent increase in long-term interest rates thus represents the unwinding of two temporary factors that signals a return to more normal conditions:

  • Taking account of secondary market purchases, public institutions have accounted for more than 80 percent of the increase in outstanding bonds since 1996 (Figure II.3). While it is unclear to what extent these purchases have been made for explicit policy purposes, this level of market support has clearly been unsustainable.

  • With conditions in the financial sector deteriorating in 1997 and 1998, investors increasingly invested in safe government debt. There are signs that the flight into JGBs may have ebbed as measures to stabilize the banking system taken by the government in late 1998 appear to have reassured investors that the financial system is now less fragile.

Figure II.3
Figure II.3

Japan: Increase in Outstanding Government Bonds by Holder

Citation: IMF Staff Country Reports 1999, 114; 10.5089/9781451820584.002.A002

Source: Bank of Japan, Economic Statistics Monthly.1/ Staff projection based on data provided by the Ministry of Finance.

36. Markets were primarily concerned about an oversupply of bonds in 1999, but concerns over future debt sustainability also played a role in the yield increase:

  • The direct result of the TFB’s market withdrawal was to almost double the amount of fresh government securities to be issued to the private sector in FY1999, compared to levels prevailing in the mid-1990s (Table II.10). Following the large bond issuance in FY1998, there were concerns about the capacity of the market to absorb the increase.

  • These concerns were reinforced by expectations that the buildup in government debt would continue over the medium term. The government has projected an increase in outstanding bonds to ¥430–440 trillion by 2003, implying that gross issues would continue at a level of ¥60–80 trillion in coming years. Markets also feared that existing contingent liabilities could result in an even larger build-up of public debt, e.g., arising from the expansion of public loan guarantees, larger than expected costs of cleaning up the financial sector, or potential losses in public sector agencies.16

  • The rise in FILP loans to local governments at the expense of investment in securities was interpreted as a reflection of the weak financial condition at this level of government, which could result in a further financial burden on the central government.

Table II.10.

Japan: Primary Government Bond Issue by Purchaser

(in trillions of yen)

article image
Source: Bank of Japan, Economic Statistics Monthly; and data provided by the authorities.

Excluding 5-year discount bonds.

Postal savings system and BoJ rollovers.

37. Concerns about a rise in bond supply were temporarily reinforced by an expectation that secondary market purchases by both the TFB and BoJ would also fall. In the recent past, TFB and BoJ purchases amounted to a total of ¥600 billion per month in secondary markets, adding to a combined purchase of ¥7 trillion per year. For some time it was believed by the markets that these purchases were to be substantially reduced:

  • The TFB halted secondary market purchases in January 1999, before the government decided to resume such operations in February.

  • Markets interpreted a speech by BoJ Governor Hayami in December as signaling that the BoJ would severely curtail JGB purchases in secondary markets. In the speech, the Governor noted the large amount of government bond holdings by the BoJ and observed that other central banks held a greater share of private sector assets than the BoJ.17 However, the BoJ has since reaffirmed that it would continue to purchase JGBs in line with the growth in bank notes.

38. In the first half of 1999, bond yields have continued to be volatile as markets responded to policy announcements. After yields peaked at 2¼ percent in early February, rates declined following the monetary easing by the BoJ in February (including a commitment to expand the scope of open market operations) and the government’s announcement that it would switch to issuing more medium-term bonds. These decisions, combined with sales of financing bills held by the BoJ, was interpreted by markets as an effective “twist” operation that contributed to the decline in yields to 1¼ percent by May 1999. However, rates have since increased to around 1¾ percent as the strengthening of the economy has led to speculation that the BoJ might shift towards tightening sooner than previously anticipated, combined with growing market expectations of a further stimulus package toward the end of the present fiscal year.

D. Fiscal Transparency in Japan

39. With fiscal policy playing a crucial role in stimulating the economy, transparency issues have received considerable attention in Japan in recent years. Market participants have repeatedly reported difficulties in analyzing fiscal developments in great detail. Uncertainty about the fiscal stance has also led to skepticism regarding the need, size and effectiveness of public stimulus programs (e.g., Posen 1998). In particular, the lack of information on the scope and timing of public works has complicated economic forecasting and contributed to volatility in both financial and product markets.18

40. Difficulties in interpreting fiscal policies are partly related to the complex arrangement of central and local government accounts. The focal point of the fiscal accounting system is the central government’s General Account, which receives almost all national tax and bond revenues. Expenditure is mostly channeled through 38 Special Accounts (under the control of individual ministries), which are partly co-financed by other borrowing, e.g., from the FILP, Similarly, the accounts of some 3,200 local administrations (prefectures and municipalities) include ordinary and special accounts, and numerous public agencies on the national and local level also have independent bookkeeping status.19 There is a substantial amount of transfers both within and between central and local government accounts (see Box II. 1), but only limited information is provided on a consolidated basis.

41. While complex, the fiscal system operates under well established principles of fiscal responsibility. Japan’s fiscal policy has generally been run on a conservative basis under close parliamentary oversight.20 Although the weak economy has forced the government to adopt a more expansionary policy stance (including through the suspension of the FSRA in 1998), the budget process remains subject to strong principles of legal and financial accountability:

  • Sound legal basis. Budget planning and execution are guided by constitutional and legal provisions that allocate clear responsibilities across the various levels of government. Funds for public expenditure have to be appropriated by the Diet, and detailed tax legislation provides the statutory basis for tax and customs liabilities.

  • Strong administrative accountability. The budget accounting system is well established and accurate, and the government is obliged to report to the Diet on fiscal developments on a quarterly basis. Settled accounts are submitted to the independent Board of Audit for inspection about 6 months after the end of the fiscal year. There is increasing use of cost-benefit analysis for public works project appraisals, and a framework for a wider evaluation of fiscal expenditure programs is planned to come into effect in FY2001.

  • Specification of contingent liabilities. Budget documents contain information on a number of contingent liabilities, including the amount of outstanding government loan guarantees and the maximum amount of debt guaranteed through the deposit insurance mechanism, Funds for liabilities arising from such guarantees are partially appropriated in the budget, which may also include a contingency reserve.

  • High ethical standards. The National Public Service Law provides legislation relating to ethical standards for national public employees. Each government department and agency has established its own code of conduct.

42. However, the accounting system is not well suited for macroeconomic analysis. Despite their high standards of accuracy, the complex arrangement of Japanese fiscal accounts is not well suited to generating the flow of data and projections on fiscal operations needed for an adequate macroeconomic assessment of fiscal policy. It is therefore extremely difficult for outside observers to monitor recent fiscal developments, to gauge the current stance of fiscal policy, and to assess the need for corrective measures during the fiscal year. Moreover, medium-term fiscal forecasting is largely conducted in an information vacuum. The main problems are:

  • Lengthy data lags. The government publishes monthly information on its General Account and, on a more limited basis, also for some of the Special Accounts. More information is provided with the annual budget, but not in a form that readily allows for the construction of consolidated accounts for the central or general governments. Consolidated accounts on past operations are provided on an annual basis through the National Accounts, which are published only with a lag of 12 months from the end of the fiscal year, owing mainly to difficulties in compiling aggregate accounts for local authorities.

  • Limited information on the current fiscal stance. The budget documents provide only limited information on the consolidated fiscal position implied by the budget proposals. The frequent recourse to supplementary budgets further complicates the analysis—particularly since only limited information is provided on the timing of fiscal operations—as does the convention of comparing initial budgets with earlier initial budgets. As a result, cross-year comparisons can be highly misleading.

  • Lack of detailed medium-term fiscal projections. The authorities publish medium-term deficit and debt scenarios under different macroeconomic assumptions. However, except for some projections relating to taxation and debt service, medium-term projections are constructed from mechanical extrapolation under different assumptions, rather than a detailed assessment of the future implication of existing commitments and policies.

    Extensive use of quasi-fiscal operations. The FILP continues to be used to advance fiscal objectives through operations outside the central government budget (including, for example, the provision of 1½ percent of GDP of financing to local governments or, in the past, the temporary funding of losses associated with the privatization of the Japan National Railway). Moreover, agencies funded through the FILP are not exposed to market scrutiny of their operations. While the default rate for current FILP agencies has been low, future budgetary risks from such operations are difficult to assess.

43. While there should be further scope to enhance data provision, in the end achieving best practices in fiscal transparency would require fundamental reforms to the legal and institutional fiscal framework. Some of the largest obstacles include: (i) the important role of Special Accounts under the control of individual ministries, which makes it difficult to track aggregate spending and revenues; (ii) the reliance on large stimulus packages (implemented through supplementary budgets) that diminish the importance of initial budgets; and (iii) the reliance of most FILP agencies on direct financing or explicit government guarantees (which appears likely to continue even after planned FILP reforms in 2001).


  • Bayoumi, T., “The Japanese Fiscal System and Fiscal Transparency”, in B.B. Aghevli, T. Bayoumi, and G. Meredith (eds.), Structural Change in Japan: Macroeconomic Impact and Policy Challenges, (Washington: International Monetary Fund), 1998.

    • Search Google Scholar
    • Export Citation
  • Ishi, H., “Budgets and the Budgetary Process in Japan”, Hitotsubashi Journal of Economics, 37, pp. 119, 1996.

  • Mihaljek, D., “Intergovernmental Relations and Local Public Finance in Japan”, in Japan—Selected Issues, (Washington: International Monetary Fund), 1996.

    • Search Google Scholar
    • Export Citation
  • Okue, K., “Japanese Fiscal Policy: Prospects and Feasibility”, Tokyo: Dresdner Kleinwort Benson, May 1999a.

  • Okue, K., “Filling the Vacuum: Japan Policy Monitor”, Dresdner Kleinwort Benson Research Paper (April 16), Tokyo, 1999b.

  • Posen, A., Restoring Japan’s Economic Growth, (Washington: Institute for International Economics), 1998.


Prepared by Martin Mühleisen (ext. 38686).


The paper by Mühleisen in the accompanying Selected Issues paper examines the impact of stimulus packages in more detail.


For example, the packages included loan commitments by government financial institutions (GFIs), especially for small and medium-sized enterprises, public loan guarantees, employment measures, and projects for other public sector agencies. These have largely been implemented through the FILP without requiring additional general government funds.


Japan’s construction sector accounts for 10 percent of total employment, compared to an average of 6–7 percent in other major industrial countries.


The supplementary budgets for FY1998 contained a provision of ¥1 trillion to cover losses from earlier bank failures, and the FY1999 initial budget provided ¥2½ trillion to account for losses related to the nationalization of the LTCB and the Nippon Credit Bank.


This assumes that a contingency reserve of ¥500 billion for public works would be spent in the course of the year.


The authorities plan to spend 95 percent of funds allocated for FY1999 within the current fiscal year, as opposed to a historic average of around 75 percent.


The staff’s lower cumulative deficit over the two years reflects an expected shortfall in local government spending, discussed further below.


These projections include the June 11 package of measures to alleviate unemployment that is to be financed through a ¥520 billion (0.1 percent of GDP) supplementary budget.


By end-FY1999, gross indebtedness of the general government is projected to have risen to 130 percent of GDP (compared to 70 percent of GDP in FY1990). Net debt including social security will be around 38 percent of GDP, compared to 10 percent of GDP in FY1990.


Under the official classification, bond issues and borrowing are counted among revenues, and the deficit reflects only changes in financial reserves. Based on information in the White Paper on Local Government Finance, the deficit reported here reflects the definition provided by the guidelines for Government Finance Statistics.


Although public works funded by the central government have increased by 5 percent, overall local public works of Prefectures are still projected to drop by 11 percent.


In the first two months of FY1999, JGB holdings by the TFB have indeed declined by ¥5 trillion, while loans to local government organizations have increased by a similar amount (see Table II.8).


Real yields are defined as nominal yields less the annual rate of consumer price inflation.


The TFB currently holds one fifth of all outstanding JGBs. It accounted for 60 percent of the increase in outstanding JGBs in FY1997, and 30 percent in FY1998.


A widely quoted example was the winding up of the Japan National Railway Settlement Corporation, that included the government taking on 4½ percent of GDP in additional debt in 1998.


The speech was delivered on the same day the end of JGB underwriting by the TFB was announced.


For example, many forecasters were surprised by the strong impact of public investment on GDP growth in the first quarter of 1999, with some expressing doubts over the way public construction indicators were incorporated in preliminary estimates for the national accounts.


For details on institutional arrangements, see Ishi (1996) and Bayoumi (1998).


Up until 1975, a version of the “golden rule” limited borrowing in any year to the level of investment spending (“construction bonds”). Since then, with the exception of 1991–93, the government has also issued special “deficit financing” bonds to finance general expenditure. However, the government retains the medium-term goals of reducing the ratio of special bond issues to total expenditure, and keeping the tax-to-GDP ratio below 50 percent.