Chapter

13 The Japanese Fiscal System and Fiscal Transparency

Author(s):
Tamim Bayoumi, Guy Meredith, and Bijan Aghevli
Published Date:
June 1998
Share
  • ShareShare
Show Summary Details
Author(s)
Tamim Bayoumi

Ensuring fiscal accountability is an important objective for any society. This issue has come increasingly into focus in Japan in recent years, as in almost every other industrial country, owing partly to underlying economic and social trends. Slower economic growth since the early 1970s has reduced the rate of increase of new economic resources, thereby focusing interest on how such new resources are used. Aging populations and rising levels of government debt (in relation to output) have increased public concern over the sustainability of existing fiscal policies. Specific events have also played a role. The privatization program of the Thatcher government in the United Kingdom, for example, has heightened awareness of the inefficiencies of many public corporations and the potential gains from privatization.1 An important aspect of fiscal accountability is transparency, as it is only when the public has an understanding of the operations and objectives of government agencies that a clear picture of government policy can be obtained.2 Fiscal transparency covers a wide range of topics associated with providing an accurate assessment of the government’s fiscal operations, ranging from microeconomic questions, such as improving incentives for government employees, to macroeconomic ones, such as evaluating the impact of current policies on future deficits. It encompasses basic issues, such as ensuring that a budget is adequately scrutinized, and more complex questions, such as the impact of fiscal rules on the economy.3

Not all of the issues associated with fiscal transparency are of equal importance in every country. For example, the Japanese budget process is relatively open and subject to parliamentary control, and there is no systemic problem with corruption. To retain focus, this paper organizes a description of the Japanese fiscal system around three issues of particular importance in the debate about fiscal transparency in Japan:

  • Is the budget a good measure of the overall fiscal stance?

  • Is adequate information provided as to likely future fiscal trends and underlying policy objectives?

  • Are implicit fiscal liabilities accounted for clearly?

These topics cover three basic issues involved in any assessment of the fiscal stance. First, given that the budget request is the focus of most public debate of fiscal policy, how good a measure is it of overall spending and revenues? Second, is the government providing clear information as to likely fiscal trends, its own objectives, and hence the probable path of future policies? Finally, is the government being clear about potential liabilities caused by population aging and such (apparently costless) actions as providing loan guarantees?

Specific events in Japan make a review of these issues particularly timely. Fiscal policy has been used actively over the early 1990s to combat weakness in economic activity. This vigorous use of policy for countercyclical purposes has illustrated both the strengths and weaknesses in the existing fiscal system. On the positive side, the speed of implementation of fiscal packages, even when they were quite large, allowed fiscal stimulus to be provided in timely and effective manner. On the negative side, the rapid expansion in government investment spending made it difficult to ensure the efficiency of the underlying projects, and the size and impact of individual stimulus packages were often difficult to gauge.

Even more important, the lack of a well-defined medium-term framework meant that stimulus packages were not accompanied by plans for future consolidation, resulting in a fiscal outlook in FY 1996 that was clearly unsustainable over time.4 Partly in response to the dire state of its finances, the new government of Prime Minister Hashimoto has embarked on a number of initiatives aimed at fiscal and administrative reform. Targets for medium-term fiscal consolidation were legislated in the Fiscal Structural Reform Act, which passed the Diet in late 1997. In addition, government committees are reviewing aspects of administrative reform, including the structure and size of public works spending, the role of public corporations and ministries, and the future of the FILP. Within this wide-ranging debate on the fiscal system, fiscal transparency is clearly an important theme.

The Fiscal System: Structure and Measurement

The most salient feature of the fiscal system in Japan is its complexity. This comes less from the fact that there are a large number of different accounts—indeed, most governments use different accounts to organize their spending—than from their lack of consolidation. Instead, the budget for each account is generally treated independently from other accounts within the system. The term “the budget” usually refers to the general account of the national government, a practice that we will follow, but budgets also have to be submitted for other government accounts. The sections below describe the Japanese fiscal system (the Appendix provides a more detailed description), discuss how the fiscal position is measured, and explore the relationships between alternative measures of the fiscal position.

Fiscal Structure

Figure 13.1 provides a highly stylized diagram of the Japanese fiscal system.5 To focus on the relationship between the government system and definitions used in the National Income Accounts (NIA), the horizontal axis shows the subsectors of the public sector used by the NIA (central government, social security, local government, and public enterprises), while the vertical axis shows current and capital expenditures. The solid lines represent major transfers between accounts; the broken lines show major loans.

Figure 13.1.The Japanese Fiscal System

The focal point of the system is the central government’s general account, which receives virtually all national tax revenues (but not social security contributions). In the budget, general account revenues are made equal to expenditures by including financing items in with current revenues (Table 13.1 shows the initial budget for FY 1995). The deficit of the general account is approximately equal to the sum of these financing items—income from public bond sales and transfers from the previous year’s account.6 Planned general account spending, which was ¥71 trillion in FY 1995 (14 percent of GDP), included spending on national debt service, transfers to social security, local allocation tax grants, public works, education and science, and national defense.

Table 13.1.General Account Income and Spending in Fiscal Year 1995, Initial Budget(In trillions of yen)
RevenuesExpenditures
Taxes and stamp receipts53.7Social security13.9
Miscellaneous receipts4.3National debt service13.2
Sale of government properties0.3Local tax grants13.2
Monopoly profits0.0Public works9.2
Government enterprises and properties0.0Education and science

National defense
6.1

4.7
Public bond issues12.6Pensions and others1.7
Carried over surplus from Fiscal Year 19940.0Transfers to industrial investment special account1.3
Economic cooperation1.0
Energy measures0.6
Major foodstuff measures0.3
Miscellaneous5.1
Reserves0.4
Total revenue71.0Total expenditure71.0
Source: Financial Statistics of Japan, 1996, Tables 13.4 and 13.7.
Source: Financial Statistics of Japan, 1996, Tables 13.4 and 13.7.

While some government expenditures are funded directly through the general account, most items are paid through special accounts. There are 38 special accounts, many of which are quite small. It is useful to categorize them by their locations within the NIA subsectors of the public sector. The central government special accounts are generally conduits for general account spending and are largely financed through general account transfers. The most important accounts are the National Debt Consolidation Fund, which issues and manages government debt; the Allotment for Local Allocation Tax Special Account, which transfers tax grants to the appropriate local governments; and the Road Improvement, National Schools, River Conservation, National Hospitals, and Foreign Exchange Fund Accounts.

There are four special accounts in the social security subsector. Although they receive significant transfers from the general account, their revenues come largely from their own resources through social security contributions and income from assets. The Welfare Insurance and the tiny Seamen’s Insurance Accounts provide pensions and health insurance for private sector employees; the National Pensions Account covers the self-employed and agricultural workers; and the Labor Insurance Account insures against workers’ accidents and unemployment. (Public sector workers are covered by a different set of agencies, called Mutual Aid Associations, operated by the Ministry of Finance.)

The public enterprise special accounts are again largely funded from their own resources. The most important accounts include the Trust Fund Bureau of the Ministry of Finance, Postal Savings, and Postal Life Insurance. Together with social security funds, these accounts provide the backbone of the FILP, The FILP is not a single government account, but rather a series of accounts that take funds accumulated on the good faith and credit of the government (mainly from the pension system and postal savings through the Trust Fund Bureau) and use this money for public investment projects through other special accounts, public corporations, and other public institutions.

A number of government-affiliated agencies, such as the Development Bank of Japan and the Housing Loan Corporation, are included in the NIA definition of public enterprises, which is also where most public corporations are located. These agencies generally receive government finance through FILP loans.

Local governments have ordinary accounts, which cover administrative services and are included in the NIA definition of local government, as well as enterprise accounts for services such as housing, sewerage, and public transport, which are generally located in the public enterprise subsector. Local governments have relatively little financial independence in Japan,7 as in many states with a single centralized government authority, and in contrast to federated political systems such as those in the United States and Germany. Accordingly, it makes sense to include local government when measuring national fiscal policy.

There is considerable double-counting between various government accounts. In FY 1995, general account expenditures of ¥71 trillion were dwarfed by special account expenditures of ¥242 trillion, yet net spending by the national government was barely double that of the general account. Including local government finance plans, net overall public sector spending was about three times that of the general account (¥211 trillion).8

Measuring the Fiscal Position

The partial coverage of the general account means that it provides only limited information on overall fiscal operations. In addition, it is not always easy to calculate the implications of budgets for the current fiscal position. While the implications of initial budgetary plans are relatively straightforward, supplementary budgets are more difficult to analyze, due to uncertainties about the timing of announced spending plans, the convoluted arrangements for some of the measures (particularly those directed through the FILP), and ambiguity about the degree to which the packages represent new projects.9

As a result, most analysts prefer to use data on general government (usually excluding social security) published in the NIA. Indeed, the government itself is starting to use NIA definitions. For example, medium-term fiscal targets contained in the Fiscal Structural Reform Act (which are discussed further below) have objectives for both the general account and for general government excluding social security. The rationale for excluding social security is that this surplus is specifically earmarked to help pay for Japan’s rapidly aging population and hence that excluding the social security surplus abstracts from this intertemporal element in the government accounts.10

A similar situation exists with regard to data on government debt, where again most analysts (and, on occasion, the government) prefer to use NIA data. At the end of 1994, general government gross debt totaled over 80 percent of GDP, while net debt was below 10 percent of GDP, a difference (as a ratio to GDP) over double that for any other major industrial economy. Even excluding social security financial assets for reasons just mentioned, there remains a significant difference between gross general government debt (above 80 percent of GDP at end-1994) and net general government debt excluding social security (50 percent of GDP), largely reflecting the important role of the government in financial intermediation. These two measures indicate the upper and lower bounds of nonsocial security net government liabilities, with the actual figure depending on the marketability of the underlying financial assets. The government has generally preferred the more conservative option of using gross debt, an approach also taken by European nations when defining the entry requirements for EMU in the Maastricht Treaty. However, net debt has the advantage that its path depends only upon the government’s deficit, while the path for gross debt also depends upon the behavior of the government’s financial assets. Given the large size of these assets, net debt is probably a more useful concept.11

A major disadvantage with using NIA figures is that data are only provided with considerable lag. For example, NIA data on the general government deficit for FY 1995 were only released in December 1996, nine months after the end of the fiscal year. Thus, the data are, on average, over a year out-of-date. Data on the subsectors of government are even more delayed; indeed, FY 1995 data were still not released in early June 1997. As a result, analysts often have to estimate significant amounts of historical data.12 This is a major limitation in analyzing Japanese fiscal trends.

Comparing the General Account to NIA Definitions of Fiscal Policy

Given the partial nature of the general account and the delays in obtaining NIA data, an important issue is how useful general account data are for estimating broader measures of the fiscal position. If the general account largely reflects idiosyncratic events unconnected with broader fiscal trends, then the lack of consolidation across different government accounts is clearly a concern. At the other extreme, if the general account provides a relatively accurate guide to general government totals, then the lack of consolidation is less important, although even in this case, there are clear disadvantages, as past relationships may not hold in the future.

To resolve this issue, data were collected on the general account deficit and on three broader measures of the government deficit reported in the NIA: central government, general government, and general government excluding social security. Three general account series were collected: the initial budget proposal provided as part of the budget legislation before the start of the fiscal year; the amended deficit implied by supplementary budgets passed during the year in question; and the final settled account based on actual transactions.13

Figure 13.2 shows the initial and amended general account balance since FY 1972, together with balances for NIA definitions of the central government, general government, and general government excluding social security.14 All balances show the same basic pattern, deteriorating up to 1978, improving from then until 1991, and deteriorating subsequently. The similarity of overall balances implies that the general account deficit has some predictive power for wider definitions of the fiscal stance.

Figure 13.2.Fiscal Balances

(In percent of GDP)

Source: Economic Planning Agency.

This predictive power was quantified using regressions. As the main interest in projecting future fiscal policy involves changes to the overall fiscal stance, the specification relates changes in the general account deficit to changes in NIA definitions of the deficit:

where (NIA_DEF/Y) was the ratio of the particular NIA definition of the deficit (central government, general government, or general government excluding social security) to GDP; GACC_DEF/Y) was the ratio of the particular definition of the general account (initial, supplementary, or settled) to GDP; and β is an estimated parameter.15

Table 13.2 shows the results from estimating these regressions over the full period for which all data series were available (1972/94), and over the second half of the sample period (1983/94), to explore if there has been any change in behavior over time. The initial general account budget has only a weak role in explaining NIA fiscal outcomes over the longer time period.16 However, the explanatory power of the general account rises when supplementary budgets are taken into account; all of the coefficients are significantly different from zero at the 1 percent level and the R2 are generally more than 50 percent. This definition of the general account appears to have some predictive power for wider definitions of the government deficit. Finally, while coefficients on central government are close to unity, changes in the general account deficit appear to be amplified in the general government sphere, where coefficient estimates are greater than one.17

Table 13.2.Deficit Regressions

Δ(NIA_DEF/Y) = βΔ(GACC_DEF/Y)

Central

Government
General

Government
General

Government

Excluding

Social Security
Dependent variable1972–941983–941972–941983–941972–941983–94
independent variable:

Initial budget
β0.35+0.84**0.53+1.22*0.45+0.99+
(se)(0.19)(0.30)(0.27)(0.57)(0.25)(0.54)
R20.120.390.140.290.110.23
DW1.401.221.821.701.841.75
Supplementary budget
β0.93**0.84**1.20**2.00**1.11**1.71**
(se)(0.14)(0.30)(0.25)(0.41)(0.24)(0.43)
R20.670.390.500.680.490.59
DW1.702.042.572.482.602.52
Settled budget
β1.11**1.29**1.38**2.24**1.27**1.93**
(se)(0.17)(0.22)(0.31)(0.35)(0.29)(0.38)
R20.670.750.460.790.460.70
DW1.872.592.692.772.712.75
Notes: Standard errors are reported in parentheses. The symbols +, *, and ** represent significance at the 10 percent, 5 percent, and 1 percent levels, respectively.
Notes: Standard errors are reported in parentheses. The symbols +, *, and ** represent significance at the 10 percent, 5 percent, and 1 percent levels, respectively.

Similar data were collected on government revenues and expenditures, and the results are reported in Figure 13.3 and Table 13.3. 18 Initial budget estimates of revenues have little predictive power for any of the NIA definitions of government. By contrast, data including supplementary budgets are quite good predictors of central government revenues, as might be expected given that most national tax revenues accrue to the general account, but have a more limited role in predicting general government revenues.19 Turning to the expenditure results, both initial and supplementary budgets contain significant information on outcomes for NIA definitions of government, particularly with regard to central government. Strikingly, there is little evidence that supplementary budget data perform better than initial budget data, plausibly reflecting uncertainties as to the timing of supplementary budget spending.

Figure 13.3.Current Revenue and Expenditure

(Ratio to GD

Sources: Economic Planning Agency and Ministry of Finance.

Table 13.3.Expenditure and Revenue Regressions

Δ (NIA_REV/Y) = α + βΔ(GACC_REV/Y)

Δ (NIA_EXP/Y) = α + βΔ(GACC_EXP/Y)

Central

Government
General

Government
General

Government

Excluding

Social Security
Dependent variable1972–941983–941972–941983–941972–941983–94
Independent variable:
Revenue
Initial budget
β0.050.210.200.180.150.24
(Se)(0.13)(0.21)(0.18)(0.37)(0.19)(0.36)
R2-0.030.080.01-0.070.010.04
DW1.301.042.031.651.881.72
Supplementary budget
β0.57**0.61**0.40*0.470.49”0.31
(Se)(0.08)(0.14)(0.20)(0.35)(0.19)(0.36)
R20.670.630.160.160.220.15
DW2.021.632.202.072.352.11
Expenditure
Initial budget
β0.74**0.57+1.11**1.21+0.92**0.84*
(Se)(0.14)(0.29)(0.24)(0.66)(0.18)(0.39)
R20.490.180.480.170.480.24
DW1.981.511.811.602.061.98
Supplementary budget
β0.61**0.46**0.71**0.69+0.63**0.52*
(Se)(0.14)(0.17)(0.25)(0.36)(0.19)(0.26)
R20.410.320.240.190.250.25
DW1.240.841.451.131.671.39
Notes: Standard errors are reported in parentheses. The symbols +, *, and ** represent significance at the 10 percent, 5 percent, and 1 percent levels, respectively. Constant terms, which are only included in the general government regressions, are not reported.
Notes: Standard errors are reported in parentheses. The symbols +, *, and ** represent significance at the 10 percent, 5 percent, and 1 percent levels, respectively. Constant terms, which are only included in the general government regressions, are not reported.

In summary, while the general account deficit adjusted for supplementary budgets has somewhat more predictive power than the initial deficit (owing to the better prediction of revenues), both contain only limited information about the path of future fiscal deficits.

The Budgetary Process: Expenditure Control and Fiscal Targets

Planning for the general account budget generally starts in the summer, eight to ten months prior to the fiscal year (which begins on April 1).20 Ministries and agencies formulate initial budget requests, which are submitted to the Ministry of Finance by the end of August. Over the same period, the Cabinet approves guidelines for these same ministries and agencies (including “ceilings” on spending). The period from September to December is then spent in the compilation of more detailed budget estimates and in negotiations between the relevant ministries and the Ministry of Finance. Toward the end of December, the Cabinet issues the “General Principles of Budget Formulation,” which the Ministry of Finance uses to produce a draft budget. This is followed by a short period of negotiations between the ministries and the Ministry of Finance, after which the final budget is finalized and approved by the Cabinet. As in many parliamentary democracies, the budget proposed by the Cabinet is rarely altered or amended by the Diet.

From the point of view of fiscal transparency, the main interest in the budgetary process is how much control it provides over the path of fiscal policy. In particular, how the budgetary process influences the ability of the fiscal authorities to control future trends in spending and how well the ultimate objectives of budgetary policy are defined.21

Expenditure Control

The Japanese budget exhibits a high degree of short-term fiscal control with virtually no medium-term expenditure planning. The short-term control reflects the inclusion of financing items in the initial general account budget. As the projected revenue from issuing government bonds is submitted as part of the budget legislation, any significant deterioration in the fiscal deficit within the year, whether owing to external circumstances or government policy, requires the passage of a supplementary budget. This level of parliamentary control compares well with many other industrial countries, where borrowing limits are not part of the budgetary legislation.

As a result of the deficit limit, supplementary budgets are quite common, being used to secure funding for wage hikes for civil servants, emergency spending on natural disasters, and (recently) extra investment spending and tax cuts aimed at stimulating the economy. As each supplementary budget includes new estimates of borrowing and spending, they provide a useful update as to current fiscal trends.22 They also provide a high degree of flexibility for fiscal policy, which has been used in recent years to support the economy. At the same time, by raising spending levels above those in the initial budget, the widespread use of supplementary budgets can complicate fiscal policy assessment. In particular, the initial budget for the following financial year is normally compared with the initial budget from the previous year, meaning that changes to spending owing to interim supplementary budgets are excluded from the comparison.

The major weakness of the Japanese budgetary system with respect to expenditure control, however, is that, unlike most other industrial countries, Japan has no regular cycle of medium-term expenditure planning.23 Discussions on expenditures between the Ministry of Finance and the relevant departments focus almost exclusively on spending levels for the next financial year.24 As a result, Japan has alternated between periods with little expenditure control and periods, such as during the 1980s, when fiscal consolidation was largely imposed through relatively ad hoc controls on overall spending.25 In a similar vein, the FY 1997 budget contained medium-term projections for general account expenditures and revenues based on existing policies and on aggregate expenditure restraint, but with no information on how such restraint would affect specific programs.

The government is actively considering options to reform the current budgetary system. The Fiscal Structural Reform Act includes plans for expenditure restraint in FY 1998, consistent with its objective of reducing general account discretionary spending (i.e., total spending less debt repayment and tax allocation grants). However, except for some limited objectives for spending on public works, social security transfers, and defense spending, no detailed expenditure plans were announced for future years. In addition, government committees have been formed to look at various elements of administrative reform, such as the level of public works spending, the efficiency of public corporations, and the future of the FILP, all of which have budgetary implications. This interest in expenditure control reflects the need for fiscal restructuring in the face of large budget deficits. It has also led to renewed interest in articulating fiscal targets.

Fiscal Targets

Up until 1975, the Japanese government operated a version of the “golden rule,” an approach in which government borrowing is limited to investment projects.26 More specifically, general account finance was limited to construction bonds, which were used to finance public works. In FY 1976, however, the deficit became too large to be financed solely out of construction bonds, and the government started issuing special deficit-financing bonds. The government has continued to issue both construction and deficit-financing bonds every year since, except for 1990/93, when only construction bonds were issued.

Even during periods when special-financing bonds were being issued, the differentiation between the two types of bonds provided a potential medium-term target for fiscal policy. Indeed, eliminating the issuance of deficit-financing bonds was the central objective of fiscal consolidation over the 1980s. Hence, even when the golden rule has not been achieved, it has provided a useful fiscal policy target.27

More recently, the government has defined its fiscal objectives more concretely. Specifically, in March 1997, the Cabinet endorsed “five basic principles” of fiscal structural reform. In addition to restoring the golden rule by eliminating the issuance of general account deficit-financing bonds by FY 2003, this decision pledged to reduce the general government deficit excluding social security to below 3 percent of GDP by FY 2003 (from its estimated FY 1996 level of about 7 percent of GDP).28 These targets have been recently legislated as part of the Fiscal Structural Reform Act. While such targets are welcome, the deadline of FY 2003 appears somewhat unambitious given the projected increase in government debt over the next few years, particularly as the 3 percent of GDP target for the general government deficit excluding social security is not a particularly stringent long-term goal.29

Recent plans to reform the Japanese budgetary process—specifying fiscal targets, increasing control over medium-term expenditure trends, and investigating the efficiency of individual programs—will likely assist in fiscal consolidation. The experience of the 1980s, however, illustrates the limitations of ad hoc policies focusing on current needs. While fiscal retrenchment was achieved in the 1980s, however, the budgetary position was allowed to deteriorate again in the early 1990s, so that a new round of fiscal retrenchment is now necessary. A more comprehensive reform of the budgetary process, involving a regular cycle of medium-term expenditure planning and well-defined objectives, could provide a more favorable environment for maintaining fiscal rectitude over the long term.

Recent legislation in Australia and New Zealand provides examples of such comprehensive fiscal reform.30 The 1994 Financial Responsibility Act in New Zealand and the more recent Charter of Budget Honesty in Australia both aim to ensure that fiscal policy is presented in a medium-term context. Both require the government to provide fiscal projections, policies, and targets for the next few years as part of the regular budgetary cycle. Interestingly, both pieces of legislation refrain from putting specific fiscal targets into legislation and thereby tying the hands of future governments with regard to policy.

Unfunded Liabilities

Budgets and the NIA data contain much useful information about the underlying fiscal position. They do not, however, tell the whole story, as government operations often give rise to a range of unrecorded potential liabilities, which may imply potential future fiscal costs. Information as to their size and significance is clearly necessary for fiscal transparency.

There are three main sources of such unfunded liabilities. The first source of unfunded liabilities are future costs of government entitlement programs. In Japan and other industrial countries, the dominant sources of such future liabilities are government pension and health schemes, which are not adequately funded with regard to increases in costs implied by an aging population. A second source of unfunded liabilities is explicit (or implicit) government guarantees of nongovernment operations. The most important example of such guarantees in Japan involves the FILP, which invests funds largely provided by the postal system and social security in a wide range of public ventures at a rate of return that is guaranteed by the government. A final source of unfunded liabilities are certain public debts that are kept off the government budget, and hence out of government accounts, such as has occurred in the case of the Japan National Railway Settlement Corporation (JNRSC).31

It should be stressed at the outset that there may be good reasons for not fully funding future entitlements, for providing government guarantees, or for moving certain items “off budget.” The appropriateness of such actions is not the main issue in this section, which focuses on the more limited question of whether the potential costs of such behavior are being adequately conveyed to the public.

Unfunded Liabilities Associated with an Aging Population

Aging populations are a social and economic challenge for almost all industrial countries, and this is particularly true for Japan, which is expected to move from having the lowest old-age dependency ratio among the major industrial countries in 1985 to the highest in 2010. Such a demographic change has profound implications for the Japanese economy. The most important fiscal challenge involves the future funding of public pensions and health benefits. In terms of fiscal transparency, the central issue is the availability of information on the size of the future unfunded liabilities associated with an aging population, and government proposals as to how to fund the system in the future.

The Japanese government provides its citizens with health and pension benefits through the social security system (see Table 13.4). Such spending is largely channeled through three types of bodies, which are distinguished by the types of workers covered: Welfare Insurance (also known as the Employee Pension Insurance or EPI), which insures private sector employees; Mutual Aid Associations (MAAs), which cover current (and some former) public sector employees; and the National Pension (NP) and National Health (NH) systems, which service the self-employed and agricultural workers.32

Table 13.4.Social Security in Fiscal Year 1994(In trillions of yen)
PensionsHealthUnemployment1Total
Receipts43.623.12.068.8
Contributions28.317.11.446.7
Of which:
Welfare insurance (EPI)(16.3)(5.5)
Mutual aid associations(7.1)(1.8)
National pension/health(1.7)(3.0)
Other(3.1)(6.7)
Property income, etc.9.70.00.09.7
Government grants5.96.00.612.5
Expenditures29.723.12.054.8
Benefits28.522.21.952.6
Of which:
Welfare insurance (EPI)(13.8)(4.5)
Mutual aid associations(6.9)(1.2)
National pension/health(6.8)(5.1)
Other(1.0)(11.4)
Other costs1.20.90.12.2
Net lending13.90.00.013.9
Sources: National Income Accounts; and IMF staff estimates (staff estimates are italicized).

Includes child benefits.

Sources: National Income Accounts; and IMF staff estimates (staff estimates are italicized).

Includes child benefits.

The core of the pension system is the basic pension (BP), provided to all on an equal basis and operated through the various schemes (EPI, MAAs, and NP), with expenses “pooled” across the schemes through transfers to and from a central account. The individual schemes also provide additional benefits over and above those provided by the BP, so that each provides a mixture of basic pension payments and of “second-tier” benefits. The government provides one-third of all costs of the BP (from the general account) as well as various additional minor subsidies for other parts of the system. The complex nature of the overall system, involving as it does a number of different organizations and levels of benefit, clearly complicates the analysis of future costs.33

The Japanese pension system is partially funded.34 Some of the income from current contributions is used to build up assets to help pay for future liabilities, but the system is not actuarially sound, and an increase in contribution rates will be required in the future to provide the extra funds needed to pay future costs. Future liabilities of the public pension system can be calculated by estimating the path of future pension finances on the assumption that contribution rates remain unchanged. One recent estimate of this type, which can be considered typical, indicates that the future net liability of public pensions in Japan from now until 2050 is 110 percent of GDP, at the upper end of the range among the major industrial countries.35

The Ministry of Health and Welfare provides financial projections of revenues, costs, and contribution rates for the BP, EPI, and NP up to 2060, but not for MAAs, as they are operated by the Ministry of Finance. Under the existing system of benefits, the contribution rate for the EPI, by far the largest of the pension schemes, is projected to nearly double from its current level of 17.35 percent of eligible income to 34.3 percent from 2025 onward.36 The Ministry of Health and Welfare has not published a sensitivity analysis of its projections to changes in underlying assumptions about economic growth and real interest rates (although such calculations are contained in some outside studies of pensions system liabilities). Comprehensive projections of government transfers to the pension system are also unavailable (although as one-third of all costs of the BP are provided from the government, transfers to this scheme can be inferred from projections of future BP costs). Okamura (CHAPTER 15, this volume) projects a relatively modest increase from their current level of 1¼ percent of GDP to 2 percent of GDP by 2020, owing to the anticipated increases in contribution rates.

Health care is provided through a similar set of schemes to public pensions. In the case of health insurance, these divisions are largely irrelevant as people have access to the same basic health services (universal national health insurance was instituted in 1961). Public health benefits are essentially one national system operated through these various schemes for different parts of the population, with the finances of the various schemes being “pooled” together.37 Again, the complexity of the funding process makes it difficult to obtain data on overall government costs.

Employees pay for health insurance through payroll deductions and copayments. The elderly (essentially those over 70) until recently paid a flat fee a month for outpatient care and a daily fee for inpatient care. Both the national government and local governments provide subsidies for part of the costs of medical care, particularly for the elderly and self-employed. The FY 1997 budget proposed significant changes to the system, and a bill containing (slightly amended) reforms was passed by the Diet. The main changes involve raising the copayments of the insured from 10 percent to 20 percent, switching to a fee per outpatient visit for the elderly, and imposing charges on prescriptions.

“Twenty-First Century Welfare Vision,” published by the Ministry of Health and Welfare in 1994, projected that total health care expenditures would increase from about 5 percent of GDP in 1993 to about 10 percent of GDP in 2025.38 As the medical system is operated on a pay-as-you-go basis, this increase in expenditure will have to be financed either by higher payroll taxes, larger transfers from the general account, or higher copayments. No information was given on how the system might be funded over time, however, or on the implications of different underlying economic assumptions on the calculations. Okamura (CHAPTER 15, this volume), writing before the recent health reforms were announced, estimated that national government subsidies to the health care systems would increase by 1¼ percent of GDP by 2020.

The health reforms proposed in the 1997 budget will raise costs to individuals by about ¥1½ trillion (¼ percent of GDP) and, as the system is operated on a pay-as-you-go basis, the long-term reduction in government subsidies will presumably be of similar size. Even with these reforms, it will be difficult to keep the sum of taxes, social security contributions, and the deficit below 50 percent of national income, which is one of the five “basic principles” of fiscal reform. Accordingly, the government recently announced a review to consider further changes to the social security system beyond existing pension and health reform plans.

Current Unfunded Liabilities

Unfunded liabilities in current operations can come from a number of sources. One such source is accounting practices. The Japanese government accounts, like those of many other governments, are calculated on a cost, rather than accruals, basis. This can lead to distortion of government accounts by delaying payments of overdue bills, particularly by governments with severe financial difficulties, as expenditures are only entered into the accounts when such bills are paid. For this reason, several countries have been moving toward using a mixture of cost and accruals accounts (including the United States), although only New Zealand has moved entirely to accruals accounting as part of its more general program of fiscal reform.

In addition to cash accounts, the Ministry of Finance publishes estimates of the government’s “hidden debt,” which includes deferred government payments on existing liabilities, and is a useful check on budgetary manipulation.39 Such “hidden debts” are projected to be ¥45.4 trillion (9 percent of GDP) by the end of FY 1997, although the net figure is presumably much lower as the government is also owed money that has yet to be paid.40

Another source of unfunded liabilities in Japan and other industrial countries is the banking sector. In particular, government guarantees of bank deposits can generate significant potential budgetary costs. The bailout of the savings and loan industry in the United States and the jusen mortgage companies in Japan represent instances in which these costs have been realized. Such costs can be reduced by ensuring that banks are charged the market value of such government guarantees.

The main sources of current unfunded liabilities in Japan come from the provision of loan guarantees by the FILP, and the off-budget debts of the JNRSC. It is useful to discuss the two issues at the same time, as FILP has been a major lender to the JNRSC.

The FILP is not a single government account, but rather a name given to a series of accounts that take funds accumulated on the good faith and credit of the government and use this money to extend loans and make investments in various public institutions.41 The budget for the FILP is compiled annually in conjunction with the compilation of the national budget for the general account, and is debated and approved by the Diet (Table 13.5 shows the budget in FY 1995). Given the size of the FILP (spending is about 10 percent of GDP, roughly two-thirds that of the general account), it is often called Japan’s “second budget.”

Table 13.5.Initial Budget for the FILP: Fiscal Year 1995(In trillions of yen)
RevenuesExpenditures
Trust fund balance36.7Housing12.0
Postal savings10.0Government housing loan corporation10.6
Employees’ pension and national pension7.4Housing and urban development corporation1.4
Others19.3
Small businesses6.3
Postal life insurance fund8.2Of which:
People’s finance corporation3.3
Government guaranteed bonds and borrowing3.3Japan finance corporation for small businesses2.6
Industrial investment special account0.0Environmental sanitation business finance corporation0.4
Other government-affiliated financial institutions4.1
Of which:
Japan Development Bank1.9
Export-Import Bank of Japan1.3
Other public corporations1.9
Of which:
Japan settlement corporation0.9
Local areas8.6
Local government7.3
Japan finance corporation for municipal enterprises1.3
Portfolio investment8.0
Postal savings special account5.0
Pension welfare service public corporation2.0
Postal life insurance welfare corporation1.0
Total48.2Total48.2
Source: The Japanese Budget in Brief, 1996.
Source: The Japanese Budget in Brief, 1996.

The main source of funds for the FILP is the Trust Fund Bureau (TFB), which in turn gets its funds from the Postal Savings Fund and other special accounts, mainly those associated with public pension schemes.42 Other income comes from Postal Life Insurance and government-guaranteed bonds and borrowing. The interest rate paid by the TFB on postal and pension deposits, known as the yotaku rate, is decided by cabinet order based on long-term interest rates. The lending rate of the TFB has been identical to the yotaku rate for many years, making the TFB a simple financial intermediary between government-guaranteed deposits and FILP loans—in effect, the government guarantees the return on FILP investments.

The Industrial Investment Special Account (USA) is the “investment” arm of the FILP, and lends money to projects “contributing to the development of the economy and the improvement of people’s living standards.” The IISA provides loans to various special accounts, public corporations, and other public institutions aimed at enhancing public welfare, particularly in the fields of housing, social infrastructure, and small- and medium-sized enterprises. (The FY 1995 FILP budget, shown in Table 13.5, indicates that ¥48 trillion was lent to various institutions.)

All TFB money is lent for five years or more, and there is a legal requirement that lending be done in a sound and profitable manner to ensure an adequate return on TFB loans. Despite this requirement, the TFB clearly represents a significant potential unfunded liability for the government. The underlying returns on its loans are uncertain, while payments to depositors are government guaranteed, making it, in essence, a large government-guaranteed loan program.43

While the TFB itself has no bad loans, as all its loans are government guaranteed, there are some bad loan estimates for FILP corporations that receive money from the TFB. For example, bad loans at the Housing Corporation and Japan Development Bank totaled about ¥869 billion in March 1996 (about 0.6 percent of all loans by these entities). In other cases, potential liabilities go unreported as the government has guaranteed to cover the costs, so that there is no liability for the FILP. Such is the case with the JNRSC.

In April 1987, when the Japan National Railway (JNR) was privatized and divided into seven companies, ¥25½ trillion of JNR’s substantial debt was transferred to the JNRSC (a public corporation), together with assets of a lesser value largely consisting of land and shares in the newly formed rail companies. In January 1988, the Cabinet decided that the remaining debt after selling off these assets was the government’s responsibility. Much of the JNR debt was owed to the FILP, and the JNRSC inherited almost ¥9 trillion of FILP debt.

Asset sales have gone slower than expected, in part because of the bursting of the asset-price bubble, with the result that the JNRSC financial position is still not resolved. Indeed, because carrying costs have exceeded revenues, the debt level is estimated to have risen to ¥28 trillion in early 1997, while the value of the underlying assets has shrunk to about ¥5 trillion, leaving net debts of more than 4 percent of GDP. Over this period, debts to the FILP have also increased significantly, to almost ¥16 trillion in April 1996, as new FILP money has been used to replace expiring bonds and to pay carrying costs. The JNRSC also owes more than ¥5 trillion in interest-free debt to the general account. Any resolution of the JNRSC will require public money, involving some combination of transferring FILP debt to the government, canceling existing government loans, and increasing taxes or user charges.

The JNRSC represents an intersection of several types of unfunded government liabilities that are familiar across many fiscal systems. The first involves providing government-guaranteed loans to quasi-government agencies or corporations, such as the JNR. Such loans, involving the full faith and credit of the government, can create problems when the underlying borrowers are inefficient, and hence find it difficult to repay the loans without government assistance. As in the case of the JNR, these problems emerge during privatization. Several industrial countries, including the United States and Canada, have government accounting systems in which government guarantees automatically generate a government expenditure aimed at offsetting the potential budgetary costs of these guarantees, a system that makes their implicit costs clear. This would be particularly useful for Japan, given the sheer size of the implicit guarantees provided by the FILP, whose annual spending is about 10 percent of GDP. The JNRSC also illustrates the difficulties created by moving public debts “off budget,” so that they do not appear as part of government accounts. Finally, it illustrates the potential hidden budgetary costs from inefficient public corporations.44

The issues involved here are complex. There are legitimate reasons for moving some items off budget, just as corporations are allowed to have “extraordinary items” in their financial accounts. Similarly, government loan guarantees may well be useful, their costs are difficult to calculate, and the FILP was clearly an important element in the postwar Japanese economic recovery. At the same time, the example of the JNRSC illustrates how loan guarantees and off-budget items can obscure the true nature of the government’s fiscal position. Such lack of transparency does not help to promote informed debate.

Conclusions

This chapter has reviewed the Japanese fiscal system from the point of view of fiscal transparency. While all governments perform the basic tasks—tax and spend—the particulars of every fiscal system are unique, reflecting different historical and economic experiences. It follows that every fiscal system also has relative strengths, weaknesses, and room for improvement, and the Japanese system is no exception. The government has embarked on a comprehensive review of the fiscal system. This chapter provides a critical assessment of the system from the perspective of fiscal transparency.

Seen in this light, the current system has several strengths. There is a high level of fiscal flexibility and of parliamentary control over current levels of spending and borrowing through the use of supplementary budgets. In addition, the Japanese government has confronted the fiscal costs involved with population aging in a relatively straightforward manner. Significant pension and health care reforms have already been proposed, and further reforms are being discussed.

The fiscal system also has some important weaknesses. First, the complexity of the budgetary system, lack of consolidation across accounts, and delays in publishing national income data all make it difficult to gain an accurate picture of current government operations. The most obvious example of this is that the budget, which covers only the central government’s general account, provides a limited view of overall government operations (analyzing the social security sector is also complicated by a lack of consolidation across accounts).

Second, the current system lacks a regular cycle for medium-term expenditure planning and for regularly articulating fiscal targets. The result has been intermittent periods of fiscal laxity and consolidation. The government is in the process of providing such controls and targets for the specific purpose of achieving fiscal retrenchment. Integrating these reforms into the underlying budgetary system could provide a more favorable environment for maintaining fiscal rectitude over the long term.

Finally, the government’s involvement in financial intermediation through postal savings and the FILP creates a large potential fiscal liability, as returns to investors are guaranteed by the full faith and credit of the government. The example of the JNRSC provides a good example of how such loans, when combined with inefficient public corporation, can generate significant fiscal costs. This issue will be one of the factors considered by the government committee reviewing the FILP with a view to its reform.

Appendix. More Detailed Description of the Japanese Fiscal System

This appendix focuses on how the various government accounts correspond to the subsectors of the public sector in the National Income Accounts (NIA), because the NIA definitions provide a comprehensive and consistent way of measuring the public sector across countries. However, one should remain aware of the limitations of NIA definitions. The most notable one is that the NIA general government sector excludes public enterprises. To the extent that public enterprises operate efficiently and involve no costs to the government, it is appropriate to exclude them from the definition of government fiscal policy. In practice, however, such enterprises are often inefficient, and supporting them gives rise to hidden governmental budget costs.

The focal point of the system is the national government’s general account, which receives virtually all national tax revenues (Table 13.1 shows the initial budget of the general account for FY 1995). General account revenues are made equal to expenditures by including financing items (income from public bonds and transfers from the previous year’s account) in with current revenues. The deficit of the general account is approximately equal to the sum of these two financing items, while current revenues, which are dominated by income from taxes and stamp duties, are equal to total revenues minus the deficit.45 General account spending totaled ¥71 trillion (14 percent of GDP) in FY 1995. The largest three spending items in the general account, making up slightly over one-half of total spending in FY 1995, are national debt service, transfers to social security (basically central government transfers to the public pension and health systems), and local allocation tax grants (revenue sharing grants to local governments).46 Other significant areas of spending are public works (which also involve significant transfers to local governments, although these transfers are generally linked to specific projects), education and science, and national defense.

While some government expenditures are funded directly through the general account, many such items are paid through special accounts. There are 38 special accounts, many of which are quite small. While they can be classified in various ways, we will focus on how these accounts fit into subsectors of the public sector identified in the National Income Accounts.47

As can be seen in Table 13.6, taken in aggregate, spending by those special accounts defined as part of the central government, as part of social security, and as part of the public enterprise sector are roughly equivalent. The largest of the central government special accounts (in terms of gross spending) is the National Debt Consolidation Fund, which services government debt. Next comes the Allotment for Local Allocation Tax Special Account, which transfers tax grants to the appropriate local governments. Other significant special accounts in this sector are Road Improvement, National Schools, River Conservation, National Hospitals, and Foreign Exchange Fund Accounts. None of the remaining nine central government special accounts spent over ¥1 trillion in FY 1995 (0.2 percent of GDP).

Table 13.6.Special Accounts Expenditures: Initial Budget, Fiscal Year 1995(In trillions of yen)
Central GovernmentSocial SecurityPublic Enterprises
National debt consolidation fund53.0Welfare insurance52.4Trust fund bureau19.3
Allotment of local allocation tax25.9National pensions20.4Postal savings12.2
Road improvement6.4Labor insurance8.3Postal life insurance12.2
National schools2.9Seamen insurance0.1Postal services7.3
River conservation2.2Foodstuff control4.8
National hospitals1.1Total81.2Industrial investment2.5
Foreign exchange fund1.1Compulsory automobile liability

reinsurance
1.0
Harbor improvement0.8
Measures for energy policy0.7National forestry operation0.9
National land improvement0.6Foreign trade insurance0.8
Airport improvement0.6Agricultural mutual aid reinsurance0.3
Promotion of electric power

resources development
0.5Finance for urban development0.3
Printing bureau0.1
Registration0.2Earthquake reinsurance0.0
Consolidation of specific national

property
0.2Mint bureau0.0
Alcohol monopoly0.0
Patent0.1Forest insurance0.0
Special measures for agricultural management0.1Fishing insurance0.0
Motorcar inspection and registration0.0Total66.3
Total96.4
Grand total241.7
Source: Financial Statistics of Japan, 1996, Table 3.2 and 3.9.
Source: Financial Statistics of Japan, 1996, Table 3.2 and 3.9.

There are four special accounts in the social security subsector, which cover different parts of the populations and functions of government. The Welfare Insurance Account and the tiny Seamen’s Insurance Account deal with pensions and health insurance of private sector employees, National Pensions deals with the self-employed and agricultural workers, and the Labor Insurance Account deals with workers’ accidents and unemployment. (Public sector workers are covered by a different set of agencies, called Mutual Aid Associations, run by the Ministry of Finance.) Unlike the central government special accounts, revenues for the social security accounts come mainly from their own resources, either through contributions or income from assets, rather than from the general account. The structure of the social security system is discussed in the main text.

The largest public enterprise special account is the Trust Fund Bureau of the Ministry of Finance, which collects and invests money mainly provided by postal savings and the social security system. Next come the special accounts associated with Postal Savings and Postal Life Insurance. Together with social security funds, these accounts provide the backbone of the Fiscal Investment and Loan Program (FILP). The FILP is not a single government account, but rather a name given to a series of accounts that take funds accumulated on the good faith and credit of the government (mainly from the pension system and Postal Savings through the Trust Fund Bureau) and use this money to enhance public welfare by financing public investment projects through various special accounts, public corporations, and other public institutions, particularly in the fields of housing, social infrastructure, and small- and medium-sized enterprises. Of the other special accounts in the public enterprise sector, the most important are Postal Services, Foodstuff Control (largely concerned with rice), and Industrial Investment. As with social security, these accounts are not generally funded from the general account.

In addition to the special accounts, there are 11 government-affiliated agencies. Some of these agencies, such as the Development Bank of Japan, are included in the NIA definition of central government, while others, such as the Housing Loan Corporation, are defined as public enterprises.48 Finally, as might be expected, public corporations are generally located in the public enterprise sector. Government-affiliated agencies and public corporations, which are clearly important parts of public sector operations, generally receive government finance through the FILP.

Local governments are organized on two levels, prefectures and municipalities.49 Each of these have ordinary accounts, which cover ordinary administrative services and are included in the NIA definition of local government, as well as public enterprise accounts for services such as housing, sewerage, and public transport, which are in the public enterprise sector. In addition to central government tax allocations and other transfers, local governments are funded by local taxes—including residence, property, and indirect taxes—and by borrowing. However, local government has relatively little financial independence in Japan. Local tax bases and rates are determined by the Local Tax Law (enacted by the national Diet) and local government borrowing requires central government approval.50 In many respects, therefore, local government budgets can be seen as a decentralized part of national government policy. The central government provides an assessment of likely future trends in local government finance for the coming year in its Local Government Finance Program. Settlement data on consolidated local government budgets are only available with considerable lag.

There is considerable double counting between various government accounts. In FY 1995, general account expenditures of ¥71.0 trillion were seemingly dwarfed by special account expenditures of ¥241.7 trillion. Yet net spending by the national government (¥160.0 trillion) was barely double that of the general account, while the remaining ¥160.7 trillion of spending represents transfers between government accounts.51 Examples would be the transfer of general account funds to the National Debt Consolidation Fund and to the Allotment of Local Allocation Tax Special Account. Including local government finance plans, net overall public sector spending is about three times that of the general account (¥211.2 trillion).52 Data on net spending based on initial budget plans of the government is available in Financial Statistics in Japan, published annually by the Ministry of Finance, but similar data are not readily available for revised budgets.

Despite this complicated structure, it is possible to get an approximate correspondence between NIA subsectors of government and specific government accounts or programs. Table 13.7 provides a rough reconciliation, comparing the general account with NIA data on central government, the relevant Special Accounts with NIA data on social security, and the Local Government Finance Program with NIA local government data.

Table 13.7.Partial Reconciliation of Government Accounts and National Accounts(In trillions of yen)
National Income AccountsGovernment Accounts
Central governmentGeneral account
Revenue61.2Revenue55.4
Taxes and fines53.8Tax and stamp receipts50.7
Property income6.3Miscellaneous4.4
Other1.1Other0.3
Expenditure78.1Expenditure78.0
Interest13.3National debt service12.9
Transfers to social security10.4Social security14.5
Transfers to local government25.2Local allocation grants12.3
Education and science6.8
Fixed investment4.9Public works14.2
Capital transfers9.6
Other14.8Other9.3
Borrowing16.9Borrowing22.7
Social security expenditure54.9Relevant special accounts expenditure81.2
Local governmentLocal government finance plan
Receipts60.1Receipts71.2
Expenditures69.7Expenditures82.5
Borrowing9.6Borrowing11.3
Sources: Financial Statistics of japan, various tables, 1996; and National Income Accounts.
Sources: Financial Statistics of japan, various tables, 1996; and National Income Accounts.
References

    BayoumiTamim and BarryEichengreen1995“Restraining Yourself: The Implications of Fiscal Rules for Economic Stabilization,”Staff Papers International Monetary FundVol. 42 (March) pp. 3248.

    CangianoMarco1996“Accountability and Transparency in the Public Sector: The New Zealand Experience,”IMF Working Paper 96/122 (Washington: International Monetary Fund).

    ChandSheetal K. and AlbertJaeger1996Aging Populations and Public Pension Schemes IMF Occasional Paper 147 (Washington: International Monetary Fund).

    Economic Planning Agency1980A System of National Accounts in Japan (Tokyo: Economic Planning Agency).

    IshiHiromitsu1996“Budgets and the Budgetary Process in Japan,”Hitosubashi Journal of EconomicsVol. 37No. 1 pp. 119.

    ItoTakatoshi1992The Japanese Economy (Cambridge, Massachusetts: MIT Press).

    LipworthGabrielle1996“Postal Savings in Japan,” in Japan—Selected Issues IMF Staff Country Report 96/114 (Washington: International Monetary Fund) pp. 13549.

    MackenzieGeorge A. and PeterStella1996Quasi-Fiscal Operations of Public Financial Institutions IMF Occasional Paper 142 (Washington: International Monetary Fund).

    MihaljekDubravko1996“Intergovernmental Relations and Local Public Finance in Japan,” in Japan—Selected Issues IMF Staff Country Report 96/114 (Washington: International Monetary Fund) pp. 93134.

    Ministry of Finance1996The Japanese Budget in Brief 1996 (Tokyo: Budget Bureau of the Ministry of Finance). Also exists for earlier years.

    NoguchiYukio1987“Public Finance,” in The Political Economy of Japan Volume 1: The Domestic Transformationed. by KozoYamamura and YasukichiYasuba (Stanford: Stanford University Press).

    Organization for Economic Cooperation and Development1995Budgeting for Results: Perspectives on Public Expenditure Management (Paris: OECD).

    Organization for Economic Cooperation and Development1996Governance in Transition: Public Management Reforms in OECD Countries (Paris: OECD).

    OkamuraKenji1998“Japan’s Medium- and Long-Term Fiscal Challenges,”Structural Change in Japan: Macroeconomic Impact and Policy Challengesed. by BijanAghevliTamimBayoumi and GuyMeredith (Washington: International Monetary Fund) pp. 21539.

    RoseveareDeborahWilliLeibfritzDouglasFore and EckhardWurzel1996“Ageing Populations, Pension Systems and Government Budgets: Simulations for 20 OECD Countries,”OECD Economics Department Working Paper 168 (Paris: OECD) pp. 169.

    ScottGraham C.1996Government Reform in New Zealand IMF Occasional Paper 140 (Washington: International Monetary Fund).

    Van den NoordPaul and RichardHerd1993“Pension Liabilities in the Seven Major Economies,”OECD Economics Department Working Paper 142 (Paris: OECD) pp. 164.

    von HagenJürgen and IanHarden1996“Budget Processes and Commitment to Fiscal Discipline,”IMF Working Paper 96/78 (Washington: International Monetary Fund).

Japanese privatizations include Japan National Railways, Japan Airlines, and Nippon Telephone and Telegraph. Other specific events with wider ramifications include the savings and loan bailout in the United States and the fiscal rules embodied in the Maastricht Treaty leading to European Monetary Union in Europe.

Interest in transparency is not limited to fiscal policy. In the monetary sphere, it is one factor behind the trend to define central bank objectives more clearly. This has often been associated with increased independence in day-to-day operations, combined with greater public oversight of objectives. The Japanese government has recently rewritten the Bank of Japan Law, increasing the Bank of Japan’s independence and accountability.

The IMF has been doing substantial work on fiscal transparency recently. Microeconomic issues associated with budgetary design are discussed in von Hagen and Harden (1996), while the impact of fiscal rules on countercyclical policy are discussed in Bayoumi and Eichengreen (1995). New Zealand has been a leader in moves toward fiscal transparency, as in fiscal reforms more generally. The New Zealand reforms are discussed in Cangiano (1996) and Scott (1996).

The deficit for general government excluding social security is estimated to have been about 7 percent of GDP in FY 1996, and net general government debt (excluding social security) is estimated to have risen from 45 percent of GDP at the end of 1993 to more than 60 percent of GDP by end-1996.

Other descriptions of the Japanese fiscal system include Ishi (1996), Ito (1992), Noguchi (1987), and Ministry of Finance (1996).

General account debt-servicing expenditures include money to redeem old bonds, which should, in theory, be excluded from the deficit calculation. However, data on such spending are difficult to obtain, so this adjustment is not made in the analysis that follows.

Mihaljek (1996) provides a description.

General government spending, which excludes public enterprises, is somewhat lower.

Analysis has also been complicated until recently by the use of a relatively unconventional statistic by international standards—the dependence rate on national bond issues (i.e., the ratio of bond issues to total expenditure)—as the main measure of the fiscal stance.

Conversely, changes in the government surplus including social security is probably a better measure of the short-term impact of fiscal policy on the economy. As the social security balance is not cyclically sensitive, however, the two measures of the fiscal position generally have similar short-term movements.

This is a less important issue in Europe, where differences between net and gross debt ratios are much smaller than in Japan (data on net and gross debt are regularly published in the OECD’s Economic Outlook).

For example, in the October 1996 World Economic Outlook exercise, the IMF was still estimating the fiscal position in 1995.

The initial budget provides an estimate of likely fiscal developments over the next fiscal year, while the supplementary budget provides a snapshot of fiscal trends in the previous year. The settled budget is of least relevance as a measure of existing fiscal trends as it becomes available only slightly sooner than the NIA data.

To abstract from inflation and real growth, all balances are reported as a ratio to nominal GDP.

Constant terms were initially included in the regressions, but were eliminated as they were universally insignificant.

The coefficient on the initial budget is significant at the 10 percent level in all of the regressions, but not the 5 percent level, and the R2 statistics indicate that the equations explain only 10-15 percent of the underlying variance of the NIA deficits. While the performance of the initial budget improves when using the 1983-94 sample, it remains quite limited.

The results using settled budgets are generally similar to those using supplementary budgets, at least for the longer time period.

Constant terms were not included in the regressions using central government data, but were included in those using general government data (with or without social security), to take account of divergent trends in revenues and expenditures across different levels of government.

The large difference in predictive power between initial and supplementary budgets plausibly reflects the importance of actual economic conditions in determining revenues.

For fuller descriptions, see Ishi (1996) and Ministry of Finance (1996).

Expenditures are the most important aspect of medium-term fiscal planning, as they define the required revenues, and hence tax policy.

As discussed earlier, the timing and size of the extra spending associated with some of the larger recent supplementary budgets has been difficult to ascertain.

OECD (1995) contains a survey of practices across industrial countries.

In addition, the decentralized nature of most of these spending discussions between the Ministry of Finance and individual spending departments may complicate the process of controlling aggregate spending and can tend to limit flexibility across different categories of expenditures. Certainly, there is a widespread perception that allocations within some areas of government expenditure, such as public works spending, have been too rigid in the past.

The government also announces multiyear plans for government investment. However, these plans do not appear to be closely integrated with the annual budgetary process.

The logic for the golden rule is that investment expenditures provide a stream of benefits into the future, and hence should be paid for overtime, while the benefits from current spending are immediate and should be paid from current taxes.

The existence of such a target can, however, reduce fiscal transparency to the extent that authorities use inflated revenue projections and complicated accounting techniques to avoid projecting the issuance of deficit-financing bonds in the initial budget, as such techniques make the initial budget difficult to assess. The FY 1994 and FY 1995 initial budgets had this problem.

The Cabinet also agreed to reduce discretionary general account spending in FY 1998 compared with FY 1997, to name 1998/2000 the “special reform term,” and to ensure that the sum of taxes, social security contributions, and the fiscal deficit does not exceed 50 percent of national income.

For example, if nominal GDP growth were to average 3 percent a year, such a deficit would imply that nonsocial security net debt would converge in steady state to 100 percent of GDP, compared with its current level of about 65 percent.

Cangiano (1996) and Scott (1996) discuss the New Zealand reforms in more detail.

As this book went to press, the government announced plans to incorporate the bulk of the JNRSC debt into the general government’s debt as part of the FY 1998 budget.

In terms of the fiscal structure outlined above, Welfare Insurance and the NP and NH are special accounts, while the MAAs are not included in government accounts.

For example, the concepts for income and expenditure used in Ministry of Health and Welfare projections do not correspond to those in the NIA. The EPI, NP, and MAAs expenditures in the NIA are the sum of second-tier benefits plus payments from the BP to the scheme in question, while in the Ministry of Health and Welfare projections expenditures, are the sum of second-tier benefits plus payments to the BP.

As is the system in the United States. Pension systems in other major industrial countries are basically pay-as-you-go, and have few assets to help offset future increases in costs (Chand and Jaeger, 1996, Table 3).

These estimates take account of the 1994 pension reform in Japan, which provided for more accelerated increases in future pension contribution rates and curtailed future costs through a steady rise in the retirement age and through tying pension increases to the increase in net, rather than gross, wages (Okamura, chapter 15, this volume). By way of comparison, the actuarially fair contribution rate from the beginning of the scheme (i.e., the fixed contribution rate at which the system would be funded) is estimated to be 22 percent.

The benefit package varies with age, with copayments paid by the government being 50 percent of allowable expenses for the elderly compared with 30 percent for others.

Okamura (1996).

For example, part of the estimated ¥2 trillion increase in hidden debts in FY 1997 is a deferment of national government pension payments.

Over one-half of all “hidden debts” are associated with the JNRSC, discussed further below.

See Ministry of Finance (1996) for a fuller description.

The Japanese postal system is the largest financial institution in the world, public or private. For a description, see Lipworth (1996), who also discusses the degree to which the size of the postal system reflects hidden competitive advantages provided by the government to the system compared to its private sector competitors. As discussed earlier, the TFB is a public enterprise special account.

Mackenzie and Stella (1996) provide an overview of quasi-fiscal operations by public financial enterprises in a range of countries.

The true financial condition of public corporations is not easy to establish, as such corporations have not been required to produce the same financial statements as private companies in the past, although this is in the process of being changed.

General account spending includes money to redeem old government bonds, which should, in theory, be excluded from the deficit calculation. Information on such expenditures is difficult to obtain.

Local governments receive about 20 percent of their income through such grants, and a further 15 percent through subsidies, largely derived from the national government.

The assignments to subsectors of government are given in Economic Planning Agency (1980), p. 53. The Japanese Budget in Brief provides a sevenfold division into categories such as enterprise special accounts, insurance special accounts, and so on. Useful as this system is in many respects, it tends to obscure the relationship with NIA data.

As already noted, the Mutual Aid Associations, which provide social security coverage largely for public servants, are part of the social security sector. These are not, however, counted as government-affiliated organizations in the budget.

For a fuller description, see Mihaljek (1996).

This provision is being lifted. However, the central government will continue to have considerable control over local government borrowing.

These figures include ¥8.0 trillion spent by government-affiliated agencies.

General government spending, which excludes the public enterprise special accounts, is somewhat lower.

    Other Resources Citing This Publication