7 Fiscal Decentralization in Japan: Does it Harden the Budgets of Local Governments?

Keimeir Kaizuka, and Anne Krueger
Published Date:
July 2006
  • ShareShare
Show Summary Details
Eiji Tajika and Yuji Yui 


In spite of the well-developed theory of fiscal federalism and accumulated empirical evidence of fiscal decentralization, every country has been facing its own problems in unleashing the capacity of local governments while keeping their budget tight. Oates (1999) reviews theoretical developments since his celebrated decentralization theorem, which basically states that local governments can cater to the needs of the local residents better than the central government, if inefficiencies arising from their limited economy of scale and over-jurisdiction spillover effects are properly coped with by either the central or local governments or both. Inman and Rubinfeld (1997) classify federalism into economic, cooperative and democratic (majority-rule) types, and pay special attention to the Welfare Reform Act of 1996 of the U.S, which has assigned part of welfare policies and much of their implementation to state governments. This is a challenge to the proponents of economic federalism who have claimed it relevant to assign the main role of social security to the central government and rationalized the application of matching grants rather than lump-sum block grants to state governments in order to mitigate possible negative externalities resulting from poor local provision of social services.

On the empirical front of the fiscal federalism, Joumard and Kongsrud (2003a) report an international comparison of fiscal decentralization.1 An interesting fact highlighted in that paper is that whether a nation is federal or unitary does not relate to the relative magnitudes of either revenue or expenditures of local governments compared to the central government. In a comparison of OECD countries, Japan looks rather like a decentralized country; its local governments’ share of revenue to total is 26 percent and the corresponding share of their expenditure is 40.7 percent in 2001. Since average shares of local revenue and expenditure across major OECD countries are 21.9 percent and 32.2 percent, Japan seems to be decentralized in both revenue and expenditure. Joumard and Yokoyama (2005) start their in-depth study about Japanese fiscal decentralization with this result from an international comparison and propose the agenda for further reform. This encompasses improving local governments’ abilities and incentives to manage local public services efficiently, containing costs and adverse effects of the grant system, increasing local government tax autonomy and hardening fiscal rules and strengthening the role of financial markets to ensure fiscal discipline.

Against the background of these theoretical, empirical, and institutional aspects of fiscal federalism, we will present the evidence of fiscal relations between the central and local governments of Japan and identify challenges as well as the direction of reform for fiscal decentralization. This has become a very hot policy agenda since the mid-1990s. Policy makers in Japan, both in the ruling and opposition parties, are very concerned with regional disparities in income. While differences of income among localities are observable almost everywhere in the world, a characteristic aspect in Japan is that these are closely connected with its political system. Simply put, the number of delegates to the national parliament has not been adjusted fully for changes in the geographical distribution of the population: the relative voting power of rural and urban regions have been kept almost unaltered in spite of significant migration from rural to urban regions during the 1950s and the 1960s, Japan’s high growth era. This has created strong political pressures toward revenue transfers from urban to rural areas. And intergovernmental grants, no matter whether they are lump-sum or earmarked, have been extensively used as a means for regional income equalization.

Since outright income support to individuals is mostly restricted to social welfare programs, income transfers to rural regions have sometimes taken such a disguised form as public expenditures on local roads and other infrastructure construction. Hanai et al. (2000) have studied such implicit transfers and shown that local governments have taken them for granted, satisfied with their status of financially dependent regions.2

Thus a consequence of massive intergovernmental transfers in Japan is a disincentive trap in which many local governments have lost their incentives to strengthen their fiscal capacities and increase their own tax revenues. From this angle, the surge of political interest in fiscal decentralization since the mid-1990s seems in contradiction with the dependent state of local governments. That is, unless they are fiscally independent, they would stand more to lose than to gain when they have to self-finance in a decentralized system. A reason behind this puzzle is that local governments are demanding to replace earmarked grants with their own revenues and to have more tax powers transferred from the central government. Moreover, they insist that block grants be preserved to finance the so-called “standard expenditure” of every local government.

Thus decentralization as demanded by Japan’s local governments is a game where nobody loses. Obviously, there is a strong conflict of interest between the central government, whose debt now amounts to more than the nation’s GDP, and local governments. Following this abstract overview, the rest of the chapter is devoted to examining in a systematic way the empirics of local government financing in Japan.

Overview of the structure and budgets of local governments

Overall characteristics and the budgets of local governments

There are three levels of governments in Japan: the central government, prefectures (provinces), and municipalities (cities and towns). There are 47 prefectures, including Tokyo metropolitan government, Osaka, and others; there are about 3,000 municipalities. Sizes of local governments, in terms of both population and economic output, vary significantly. With a population of more than 12 million, Tokyo is the largest of all prefectures, whereas the least populated Tottori prefecture has only 610,000 people. There are big differences in population at the municipality level, too; while Yokohama City has a population of 3.5 million, small towns and villages have populations of a few thousand or fewer. The difference in regional gross product can differ widely as well; for example, Tottori prefecture’s gross product is only about 2.4 percent of that of Tokyo.

We turn next to the budgets of different levels of governments. Table 7.1 shows the three budgets for the year 2002—that of the central government, that of the local governments, and the consolidated budget. Numbers in the tables are all in percent of gross domestic product. There are a couple of salient features in each and consolidated budgets. First, the central government tax revenue is small: only 9.4 percent of the nation’s GDP is collected by the central government. This is partly due to successive tax cuts, in the context of stimulus fiscal packages during the 1990s when Japan experienced a depression. The second point is that a big transfer is made from the center to local governments, amounting to some 6.7 percent of GDP. As a result of this huge transfer, the primary balance of the central government is deeply in the red, to the tune of 3.4 percent of GDP. The third point, which is another side of the coin, is that local governments receive a transfer from the center as big as their own tax revenue. Among expenditures of local governments, spending on pubic investment stands out. This expenditure has an income-redistributing aspect and has been allocated more to rural than to urban areas (Hanai et al., 2000, provide evidence of this aspect of local public investment in Japan). The aggregate budget of local governments has been in a much better shape than the central government’s and it is almost balanced. The last point to note is that after intergovernmental transfers, the expenditure of local governments turn out to be about twice as big as that of the central government, as shown by the consolidated budget of the central and local governments. Thus, notwithstanding its centralized appearance, Japan is much decentralized in the sense that the major providers of public expenditure are local governments.

Table 7.1Budgets of the central and local governments (FY 2002)
Expenditures% of GDPRevenue%
Budget of the central government:
Own expenditure7.0Tax9.4
Local transfers6.7Others0.9
- General-purpose grants4.1
- Specific-purpose grants2.6
Primary balance(deficits) 3.4
Budget of local governments:
Own expenditure14.9Tax7.0
- Wages4.8Transfers from the center6.7
- Administration4.2- General-purpose grants4.1
- Public investment5.0- Specific-purpose grants2.6
Primary balance(surplus) 0.2
Consolidated budget of the central and local governments:
Central government7.0Tax16.4
Local government14.9- Central government9.4
- Local governments7.0
- Central government0.9
- Local governments1.4
Primary balance(deficits) 3.2
Source: Calculated by the authors, using fiscal yearbooks.
Source: Calculated by the authors, using fiscal yearbooks.

We continue our review of the budgets of the central and local governments and look at their time series characteristics. Figures 7.1 and 7.2 show, respectively, the budgets of the central and local governments since 1990. Figure 7.1 shows clearly the impact of fiscal stimulus policies in the 1990s: spending by the central government started to increase at the beginning of the 1990s and reached a peak in 1998, while its revenue declined, most dramatically in 1999. Behind the fiscal expansion of the central government, the transfer from the central to local governments increased, too, most significantly at the end of the 1990s. Now, turning to local governments, Figure 7.2 suggests that their expenditures surpassed revenues in most years during the 1990s, but the deficit relative to GDP was smaller than that of the central government. And local governments started to enjoy fiscal surpluses in 2000. The transfer from the central to local governments is highlighted in this figure to show its magnitude relative to the expenditures of local governments.

Figure 7.1Budget of central government (% of GDP)

Source: Cabinet Office, Annual Report of National Accounts, 2004.

Figure 7.2Budget of local governments (% of GDP)

Source: Cabinet Office, Annual Report of National Accounts, 2004.

We then subtract the transfer from the expenditure of the central government and add the same amount to that of local governments, and set the expenditures in the year 1985 of both the central and local governments at 100. Figure 7.3 shows the result. We can find from it that the expenditures of local governments soared in the 1990s and the gap between the expenditures of the central and local governments was widened. We have already noted that local governments in Japan have spent more than the central government. Figure 7.3 shows the gap between their expenditures increased during the 1990s, when the Japanese government mobilized every fiscally available measure to stimulate its economy.

Figure 7.3Comparing expenditures of the central and local governments (1985 = 100)

Notes: The expenditure of the central government is its total expenditure less transfers to local governments. Local governments’ expenditure includes the transfer received from the central government. The data since 1990 are based on 93SNA, and those before 1989 are based on 68SNA. Two series are connected in 1990.

Source: Cabinet Office, Annual Report of National Accounts.

Background of local governments’ demand for fiscal decentralization

We have reviewed the characteristics and the budgets of the central and local governments. Here, we would like to consider the historical and political background of intergovernmental relations in Japan that has created the current budgetary interlinkages between different level of governments. An asset price bubble started in the late 1980s when the money supply increased sharply as a result of heavy government’s intervention in the foreign exchange market to keep the value of the Japanese yen from rising against the U.S. dollar. The easy money supply triggered speculative investment into land and real estate, and the prices of these assets increased sharply. This raised the collateral value of land and, in turn, made possible further investment in land. Spiraling asset prices went bust in the early 1990s.

Decentralization has become a hot political issue with the end of the asset price bubble. Discussion at times went like this: although local governments are major deliverers of public services, a substantial part of transfer from the central government to local governments has taken the form of earmarked subsidies from the central government, and local governments have to beg the central government for them. Besides, since there are many restrictions and norms about how to use the subsidies from the central government, local governments have not been able to make better use of the resources available to them. Debate also has touched on the adequacy of local own tax revenue. Local governments have not paid much attention to increasing their tax revenue, and to replacing part of subsidies and general-purpose grants with local own taxes. Decentralization as regards the revenue side of the budget has become a contentious issue, too.

Local governments are demanding decentralization of both expenditures and revenue. However, fiscal decentralization is not a one-way street, and implies a difficult trade-off for some local governments. This is because reducing the transfer from the central to local governments and allocating more taxes to local governments will work toward widening the gap of fiscal capacities among local governments. In the context of Japan, Tokyo and other urban economically prosperous regions gain from the reform, and rural ones lose. If losers demand a compensating transfer from the central government, this contradicts the very purpose of fiscal decentralization, which aims at reducing such a transfer and replacing it with locally generated own revenue. So far, the idea of decentralization seems to have been euphoric for every local government and a nightmare for central ministries which lose their budget and control over local administrative agencies. But the day-dreaming will not continue and local governments have to realize that some of them will lose with reform. A well designed decentralization will, however, strengthen the incentives of local governments to increase their own revenue, say by improving the business and living environment, and eventually most local governments should be better off after the reform. But still some will lose, and they will most likely be the ones that have been well supported by transfers from the central government.

In fact, this is one of the most difficult political challenges in Japan. The Liberal Democratic Party, which has dominated over the political scene for almost the entire period since the war, has its strongest political constituency in rural and non-metropolitan areas, where local economies have been stagnating for years. The party has been supporting these constituencies by letting the central government shoulder a massive transfer to them. However, these local governments will be the ones most likely to lose from decentralization. Decentralization is not an easy policy to implement. A new twist, though, is that the “money-scattering” patronage policy which benefits rural areas has been criticized by voters in cities. The government is struggling now to care for city-dwellers as well. These competing political goals complicates fiscal decentralization in Japan.

Activities and roles of local governments in public expenditure

We have seen above that local governments in Japan have a major role in delivering social services. The purpose of this section is to look more closely at the expenditure side of local governments. We will specifically review the activities of local governments in important fields of public services: education, social security, and public investment.

Public spending for education is reported in Table 7.2. Over the years, total expenditure for education has been about 5 percent of the nation’s GDP. The central and local governments spend almost equally. But a noteworthy fact is that 60-70 percent of the budget of the central government is transferred to local governments through either general- or specific-purpose grants. This makes the direct expenditure by the central government very small, less than 20 percent of total public spending on education.

Table 7.2.Financial sources of public expenditure for education
Sources of funds19701975198019851990199520002001
Composition of public expenditure for education
Central government47.146.547343.442.140.747.846.7
- Direct expenditure13.412.213.713.513.515.119.418.9
- Transfer to local governments33.734.333.629.928.625.628.427.7
- Specific-purpose grants19.620.620.217.915.513.913.613.8
- General-purpose grants14.113.713.412.013.111.714.813.9
Local governments52.953.552.756.657.959.352.253.3
Public expenditure for education (¥billion)2,8838,11914,00616,56820,25823,76624,29624,137
% of GDP3.
Source: Ministry of Education, Culture, Sports, Science, and Technology, Survey of Local Expenditure for Education, 2004.
Source: Ministry of Education, Culture, Sports, Science, and Technology, Survey of Local Expenditure for Education, 2004.

As for compulsory education, which consists of six-year elementary and three-year junior high schools, the central and local governments share the total cost equally. The portion paid directly by the central government is in the form of specific-purpose grants. And the other portion is part of the so-called standard basic demand of local governments. If their standard revenue is less than the standard basic demand, the difference is paid by the central government as a general-purpose transfer. Otherwise, this portion is paid by local own revenue. An issue that has been raised by local governments as a step toward fiscal decentralization is that the specific-purpose grants from the central government should be replaced by the general-purpose grants, so that local governments can decide how to use them for compulsory education in their jurisdictions. To put it in another way, local governments are demanding to be freed from the Ministry of Education’s control over the education-earmarked subsidy. Since the subsidy for compulsory education takes on more than a third of the total budget of the ministry, this is a matter directly related to its mission as an agent of the central government. Education is thus at the front line in the battle between the central and local governments in the process of decentralization in Japan.

The financial sources of social security spending (basic social services) are shown in Table 7.3. The bulk of social security is financed by the social insurance systems: health, long-term care, public pension and unemployment insurances. Other remaining public health services and welfare programs are financed by either the central or the local governments.

Table 7.3.Financial sources of social security as a percentage of total revenue (FY 2000)
Social security schemesContributions from insured persons and employersCentral governmentLocal governmentsOther sources aTotal revenue
¥billion% of GDP
1. Social insurance:48.714.93.233.194,97218.4
(1) Health insurance:46.720.35.927.033,9016.6
Health and medical services for the aged0.021.810.367.810,1362.0
(2) Long-term care insurance5.138.526.729.83,8000.7
(3) Public pension52.,14310.1
(4) Unemployment insurance and other related compensations60.,1281.0
2. Public health services0.068.831.20.08540.2
3. Welfare programs b0.,9651.0

Other sources are transfers among insurers.

Welfare programs are minimum-income support and other welfare-related provisions.

Source: National Institute of Population and Social Security Research, Statistics of Social Security, 2004

Other sources are transfers among insurers.

Welfare programs are minimum-income support and other welfare-related provisions.

Source: National Institute of Population and Social Security Research, Statistics of Social Security, 2004

A characteristic of Japanese health insurance is that the insurers of the self-employed are municipal-level governments. Health care for the old and long-term care insurance, which deals primarily with the care for dependent elderly people, are also managed by municipal-level local governments. These commitments by local governments to social security have made their financial shares significant, too. And a portion of the spending of the central government consists of a transfer to local governments, just as in the case of financing education.

Public investment plays a special role in Japan’s pubic spending. Table 7.4 shows public investment as a percentage of GDP. There are a few facts to note. First, the nation’s aggregated gross investment was highest in 1990 and declined thereafter. Second, capital formation by the general government continued to stay high throughout the 1990s, the post-bubble period. This indicates that public investment was used to boost the economy in the recession period. A third point to note is that no less than 80 percent of public investment is made by local governments. (These public projects are often called hakomono, which means “boxes,” “Box” investments refer to the construction of offices for local governments, public recreational buildings, roads, and so on.) Hanai et al. (2000) have shown that public investment by lower-level governments has played a role in boosting the economy, as well as in supporting rural areas through income redistribution.

Table 7.4.Public investment for fixed assets
Share of GDP (%):
Gross domestic investment32.628.028.627.926.626.326.425.224.0
General government5.
Central government0.
Local governments4.
Composition of general government investments (%):
Central government14.715.815.015.517.219.219.919.921.7
Local governments85.384.285.084.582.880.880.180.178.3
Source: Cabinet Office, Annual Report of National Accounts, 2004.
Source: Cabinet Office, Annual Report of National Accounts, 2004.

Sources of revenues

In this section we turn to the sources of revenue for local governments. We start with the taxes of the central government as one of the most important sources of revenue for local governments. Examination of revenue sources of the consolidated local governments, prefecture-level and municipal-level governments then follows.

We have seen in Figure 7.1 that the revenue of the central government declined from the mid-1990s as a result of successive tax cuts to stimulate the economy. At the same time, public expenditure increased and the government’s debt continued to accumulate. Furthermore, the figure shows that the central government increased its transfer to local governments; since local governments are deliverers of public expenditure, the central government had to increase its transfer to local governments in order for its budget to be implemented.

Figure 7.4 shows the composition of the tax revenues of the central government. The shares of personal and corporate income taxes and of the consumption tax (a broad-based tax on consumption equivalent to a value-added tax) are presented in percent. The declining line in the figure is the share of direct taxes, which consist of income-based taxes and inheritance tax. Two facts stand out in the figure. First, personal and corporate income taxes are major taxes in Japan. But their shares in the total tax revenue have declined over the years; most notably in the late 1990s, when big tax cuts of both personal and corporate income occurred. This contributed to the decrease in the revenue of the central government illustrated in Figure 7.1. The second fact is that revenue from the consumption tax has been increasing and has surpassed the revenue from the corporate income tax in 2002 and 2003. Thus the structure of tax revenue of the central government is changing and a shift from income-based to consumption-based taxes has been underway.

Figure 7.4Composition of tax revenues of the central government (% of total)

Source: Ministry of Finance, Statistics of Public Finance, 2003.

We turn now to the sources of revenue of local governments, which refer here to consolidated prefecture and municipal-level governments. Their revenues consist of local taxes (their own taxes), general and specific-purpose grants, and other fee-type revenues. Figure 7.5 shows the magnitudes in percent of GDP and the composition of these revenues. The line closely tracking total revenue refer to the total expenditure of local government.

Figure 7.5Sources of revenues of local governments (% of GDP)

Notes: “Other revenue” does not include borrowing, and “total expenditure” excludes interest payments.

Source: Ministry of General Affairs, Statistics of Local Public Finance, 2004.

A few facts stand out. First, as we have seen in the consolidated budget of the central and local governments in Table 7.1, the budget of local governments has been basically in balance except for a few years in the mid-1990s, when their tax revenue shrunk as the economy headed for a recession. Second, we can observe from the figure that the share of local own tax to total revenue is small; local governments collect only about 30-40 percent of the total revenue from locally collected taxes. Local governments are sometimes called “30 percent governments” in Japan to reflect of the limited role of own revenue in their finance. Third, the amount of transfer from the central government (consisting of general-and specific-purpose grants) is almost as large as local own revenue. Therefore local governments cannot fiscally survive without transfers from the central government.

Figure 7.6 shows the sources of tax revenues of prefectures. An important fact here is that the revenue from the business tax is much bigger than that from the inhabitant tax, which is a locally collected personal income tax. In fact, the largest amount of revenue has been collected through the business tax, whose tax base is mostly the income of corporations. A characteristic of this tax is that its yield is highly volatile and affected by business cycles. This is very clear in the figure: at the height of economic boom in 1989, about half of the total revenue of prefectures came from the business tax alone. Since then, the share of the revenue from this tax has declined sharply. However, the business tax still accounts for almost 30 percent of the total tax revenues of prefectures.

Figure 7.6Sources of prefectural tax revenues

Source: Ministry of General Affairs, Statistics of Local Public Finance, 2003.

A common view of local taxes is that business taxes, especially if on corporate income, are neither desirable nor sustainable for local governments. They are not desirable because they distort the formation and allocation of capital. They are not sustainable because they tend to raise competition among local governments to attract firms to their own turf (Bird, 1999). The reason why prefectures in Japan have nonetheless relied on a business tax as their most important revenue source stems from the nature of fiscal autonomy of local governments, as discussed below.

In Japan, the Ministry of Home Affairs coordinates the expenditure and revenue of local governments, both at the prefecture and the municipal levels. Intergovernmental transfers need some sort of intermediation committees to adjust grants distributed to local governments. The ministry, however, does more than allocating grants: it determines the standard tax rates for locally collected taxes. The business tax of prefectures is no exception and its standard tax rate is determined by the ministry. With the standard (that is, “national”) tax rate for local taxes decided by the central government, local governments start to watch their potential competitors. And the outcome is that most of local governments choose the national standard rates. In a sense, the ministry at the central government facilitates collusion among local governments to set uniform rates for local taxes.

This problem also relates to general-purpose grants allocated to local governments. If a grant-receiving local government sets its tax rate lower than the standard rate, the grants it receives are cut by the ministry, because the revenue from the tax is smaller than what it would have been with the standard rate. Therefore, no local government sets the rate of tax below the standard rate, although setting tax rates below the standard levels does not violate the budget law. There are cases in which tax rates are set above the standard rates. Some local governments do this and are fiscally independent. In most of these cases, fiscally strong local governments “export” their taxes to the people and firms that reside in other jurisdictions than their own.

The mechanism of rate setting for local taxes applies to the business tax of prefectures. No prefecture sets its rate below the standard rate determined by the Ministry of Home Affairs; and it is only those prefectures with strong economies like Tokyo and other metropolitan governments that charge additional rates. This explains why prefectures have corporate income taxes as their largest revenue source. It is an irony that they can tax, because they are not fully free from the coordination (the de facto intervention) of the ministry at the central government.

Let us now consider at the sources of tax revenues for municipal governments, shown in Figure 7.7. Here, major sources of revenue follow textbooks on local government financing. Property taxes levied on land and housing, and other depreciable assets are the biggest source of revenue. In fact, except for very large-scale properties like hydroelectric power plants, essentially all revenue from property tax is collected by municipal governments. Personal inhabitant tax is the second largest revenue source and the corporate inhabitant tax is a distant third revenue contributor.

Figure 7.7Sources of tax revenues of municipal governments

Source: Ministry of General Affairs, Statistics of Local Public Finance, 2003.

So far, we have seen the sources of revenue of local government with emphasis on local own taxes and their differences from those of the central government. It also is relevant to point out the overlap of tax collection by the central and the local governments. Two noteworthy features are piggybacking and shared taxes.

One of the most important tax bases in Japan is income, no matter whether taxes are assigned to the central or local governments. As for their collection, piggybacking systems are well developed: local personal and corporate income taxes are essentially collected using tax returns submitted to the tax offices of the central government. Individual and corporate returns are sent to those offices first, and their copies are sent to local tax offices in the places of residence. There are differences in calculating taxable income and these base adjustments are done by local governments. Local taxes are collected by local tax offices. In fact, this relation is further simplified in some local taxes to the extent that their bases are linked to the amount of taxes charged by the central government. Hence local taxes are added in full to the taxes accruing to the central government.

Another instance where taxes of the central and local governments overlap is the consumption tax. The tax is shared by the central government and local governments in the following manner: the tax rate is 5 percent, of which 4 percent is taken by the central government and 1 percent by the local governments (prefectures and municipalities). The tax collection is completely managed by the central government and the local portion of the revenue from the tax is sent from the central government to prefectures. Each prefecture retains a half of its total receipt and allocates the rest to municipalities within their jurisdiction. The distribution of the tax yield to prefectures is calculated according to the amount of private consumption drawn from statistics on income of the prefectures; and prefectures allocate the remaining tax receipt to municipal governments according to their population.

Besides the revenue from taxes, intergovernmental transfers and borrowing are other major sources of revenue for local governments. We will discuss the outline and problems of intergovernmental transfers separately in the next section. Here, we would like to review some characteristics of local borrowing in Japan.

Regulations on borrowing by local governments have changed in the ongoing process of fiscal decentralization. Earlier, local borrowing were examined by the Ministry of Home Affairs and approved if appropriate. This has now been changed into application basis regulation. Another important feature of local-bond issuance in Japan is that the Ministry of Home Affairs bundles together local bonds and offers them to the market as one package. In that sense, local bonds are guaranteed by the central government.

The intergovernmental transfer system

The determination of general-purpose grants

The local governments receive large transfers from the central government, in an amount almost equal to their own tax revenues. There are two types of transfers: specific-purpose grants and general-purpose grants. We have noted that one of the most important issues regarding specific-purpose grants is the control over their use by the central government. For this reason, local governments are demanding replacing them with their own taxing powers or unconditional (general-purpose) grants. We looked at compulsory education financing to illustrate the main problems of specific-purpose grants. This section discusses the modalities and related problems of general-purpose grants. It argues that one of the most fundamental problems in local government financing in Japan lies with allocation of this type of grants.

The broad outline of the formula for the allocation of general-purpose grants (the so-called local allocation tax) is familiar:

general-purpose grants allocated to a local government = standard basic needs - standard basic revenue

where the standard basic needs is the expenditure necessary to provide basic local services; and the standard basic revenue is 75 percent of the tax revenue of a local government when standard tax rates are applied. If the standard basic revenue exceeds the standard basic needs for a local government, then it does not receive general-purpose grants. But, by the same token, it does not have to contribute the “excess” revenue to the central government, either.

This, however, is rather a superficial interpretation of the formula. A question here is which variables are predetermined and which ones are determined internally by the formula. We will illustrate step by step how actually the allocation of general-purpose grants is determined. The mechanics of this process are illustrated in Figure 7.8 and explained below.

Figure 7.8Determination of general-purpose grants and standard basic needs

Total expenditure of all local governments (including both prefectures and municipalities) is assessed by the Ministry of Home Affairs and appropriated through negotiations among ministries and political groups. This is a predetermined variable in the whole process of determining general-purpose grants.

The total amount of general-purpose grants is then determined as the difference between the total expenditure and the total revenue (Step 1 in the figure). Standard basic needs is then determined in turn by:

Standard basic needs = Total amount of general-purpose grants + Standard basic revenue (Step 2 in the figure).

In a top-down way, though, this rationalizes the amount of general-purpose grants as the difference between standard basic needs and standard basic revenue. Thus standard basic needs is the variable determined endogenously at the final stage of allocating general-purpose grants.

The process means general-purpose grants are not determined as a difference between standard basic needs and revenue from the outset, but by budgetary negotiations initiated by the Ministry of Home Affairs. And once the total amount of the general-purpose grants of all local governments is determined, the aggregated standard needs and revenue are decomposed into amounts for each local government while the implied financing gaps is allocated to every government as a general-purpose grant.

A corollary of this process is that standard basic needs is an ex post budgetary appropriation, and not an ex ante assessment of basic needs. This is therefore the mechanism that bails out local governments when their revenue collections are so poor that they cannot meet their expenses. We can call the Japanese general-purpose grants a fee-for-service payment by the central government instead of a prospective formula-based fund allocation. This may be viewed as an example of the soft budgeting often pervasive in public sector management (Kornai et al., 2003; Wildasin, 2004).

We turn now to a problem that arises at the revenue side of local governments. In short, our claim here is that the allocation mechanism of general-purpose grants discourages the incentives of local governments to increase their own revenue and they remain satisfied to be grant recipients. Technically, the problem arises from a kinked budget line for a local government, and this is illustrated in Figure 7.9.

Figure 7.9Kinked budget line of a local government

The figure shows the budget line that a local government confronts. The amount of local tax collected is shown on the horizontal axis and the total revenue of a local government, consisting here of local tax and general-purpose grants, on the vertical axis. Standard basic needs are shown on the vertical axis as the distance of OA. If local tax is less than the standard basic needs, then the local government’s budget is determined by:

local tax + (standard basic needs - 0.75 x local tax) = (1 - 0.75) x local tax + standard basic needs

and this portion of the budget line corresponds to AC in the figure.

When the amount of local tax collected is more than the standard basic needs, no grants are allocated to the government and the total revenue is the amount of local taxes. This portion of the budget line is a straight line with the slope of 1 starting from C. Combined with the already drawn budget line in the grant-receiving situation, the heavy line in the figure is the budget line for a local government.

The slope is just 0.25 as long as the local government receives general-purpose grants, but it jumps up to unity when it loses the status of grant-receiver. In other words, if it stays as a grant-receiver, it can retain only a quarter of its increased tax revenue. Therefore, if general-purpose grants make feasible the provision of basic services, a local government may not pay as much efforts as possible to raise its own revenue. This, in particular, applies to those local governments with weaker economies and smaller tax bases. It would be a rational decision for these local governments to safeguard their status as grants recipients.

Outcome of the transfer to local governments

We would like to discuss the transfer from the central to local governments to show how extensive and large it is relative to local tax revenue. Our intention here is to underscore that liberal transfers have undermined local governments’ incentives to get out of the grant-receiving status.

Figure 7.10 shows how poorer local governments are better treated by the central government. The figure sets the amount of local taxes at 100, and compares the general and specific-purpose grants and the total expenditure. Tokyo is set aside as a non-grant-receiving government, and the rest of local governments are classified into five groups with the first quintile being the richest and the fifth the poorest on a per capita basis.

Figure 7.10Effects of transfers to the budgets of local governments: the budget of FY 2000

Source: Authors’ estimates.

Tokyo receives a very small amount of specific-purpose grants. General-purpose grants relative to local tax increase as local governments become poorer: in the third quintile, general-purpose grants are almost as big as own tax revenue, and in the forth and fifth quintile groups, the amounts of grants are larger than locally collected taxes. Total expenditure, which is financed by tax, grants and other sources of revenue, shows more straightforwardly how public resources are redistributed in Japan.

Figure 7.11 shows an even more striking evidence of the redistributive nature of the transfer from the central to local governments. Prefectures are ordered from the richest to the poorest according to their per capita income. At the prefecture indexed by zero in the middle of the figure, the average per capita amounts of local tax revenue and grants received are shown. A comparison is then made between per capita tax revenue and grants received over prefectures.

Figure 7.11Per capita tax revenue and the grants received from the central government: the budget of FY 2002

Notes: Prefectures are ordered by per capita income with prefecture 1 the richest and 47, the poorest; average amounts of grants and local tax revenue are shown in the figure at prefecture 0.

Overall, tax revenue declines as per capita income becomes smaller. Some exceptions, where the curve has a positive slope, correspond to the prefectures with large populations relative to their aggregate income. The amount of grants received is the sum of specific- and general-purpose grants. The poorer prefectures receive more grants in per capita terms than their richer counterparts. There are some “dents” in the chart of the grants received and they correspond again to the prefectures with large populations relative to their sizes of economies.

The figure illustrates very clearly the income-redistribution effects of grants in Japan. The grants to poorer prefectures are so large that they more than compensate for their lower tax revenue. In fact, poorer provinces spend far more in per capita terms than their richer counterparts. These facts point to the possibility that excessive transfers from the center to local governments has created incentives for poor local governments to continue to be in the situation that qualifies them for such transfer.

Other mechanisms for transfer and equalization

The system of transfers from the central government to local governments is more complicated than what we have seen above. There are other and more subtle ways for supporting local governments. One among them is to let local governments issue bonds for financing their projects. Usually, specific grants are attached to these projects and the rest of the costs is to be borne by local government. However, a part (sometimes, almost the whole) of interest and amortization payments is permitted to enter into the calculation of standard basic needs of the bond-issuing local government and is paid eventually by general-purpose grants from the central government.

In this case, general-purpose grants are used as leverage for financing public investment by local governments. And the root cause of the problem is that standard basic needs of local governments is not well defined and could even cover the costs of projects already started. This retroactive financing of local projects has made local governments extremely dependent on the central government and their fiscal discipline has been weakened. The expansionary fiscal policies of the late 1990s fostered reliance of local governments on general-purpose grants.

Outstanding issues and the direction of reform

In spite of the appearance of tight fiscal centralization, local governments in Japan provide almost twice as much public services as the central government. They play an important role in the education, social security and public investment. However, it is also true that the size of local governments in total expenditure or revenue does not imply independence from the central government. We have pointed out two aspects that relate activities of local governments to the central government: first, the uses of specific-purpose grants are controlled by the central government; and second, standard rates are set by the Ministry of Home Affairs on local taxes. Thus the size of local governments does not fully represent their autonomy. The tension between the size and degree of autonomy of local governments comes from their dependence on transfers from the central government. We have turned to general-purpose grants to discover that serious incentive problems are present on both expenditure and revenue sides. On the expenditure side, retrospective appropriation of local expenditure has expanded their budgets significantly and weakened the fiscal discipline of local governments. On the revenue side, the budget line of a local government is kinked, with the effect of reducing its efforts to increase its own tax revenue. With disincentives from expenditure and revenue sides combined, local governments in rural and other economically backward regions tend to be content with the status quo. These governments have fallen into a disincentive trap that weakens their efforts to acquire greater autonomy.

On the basis of these findings and our diagnosis of the problems in local public finance in Japan, we would like to suggest a direction for reform. A necessary major step is to change the current retrospective, fee-for-service-type budgeting of general-purpose grants to one based on a prospective formula. Here, hardening the budgets of local governments by a forward-looking assessment by independent experts their basic needs and revenue-raising capacity is the first reform to be undertaken. The difference between the basic needs and revenue capacity is then paid to local governments as a support to their provision of basic public services. Whether sources for these grants are solely financed by the revenue of the central government or some pooled revenue of the central and local governments is a next issue to consider. From the standpoint of removing the kink from the budget line of a local government, it is necessary to have large revenue sources.

The second major reform is to abolish the coordination of local governments by the Ministry of Home Affairs (which has led to collusion), and to let local governments decide their tax rates. Since their revenue capacity is to be assessed by an independent body, a free choice of tax rates will not create moral hazard. With basic needs prospectively assessed and a freer choice of tax rates, local authorities will become more inclined toward increasing their revenue and providing public services of their own.

Besides these fundamental reforms, attention also has to be paid to the purpose of guaranteeing basic needs of local governments. Transfers of resources are made to those local governments that cannot fully finance local basic needs. But this is to enable them to provide the minimum necessary services to the people in their jurisdictions. Hence excessive and (sometimes) disguised grants for the sake of creating jobs and distributing income to local governments have to be discontinued. In short, dual function of the current grant system that supports the provision of basic social services by every local government and redistributes income should be separated out, and these objectives should not be lumped together in the system of grants to local governments.


We conclude by highlighting the implications of our findings for the future course of fiscal decentralization. We have noted from the outset that both earmarked and block grants from the central to local governments in Japan have been politically motivated and that one of their most important functions has been as a means for income equalization. The resulting problems of these grants are threefold: earmarked grants are often too tightly controlled by the central government; the budgets of local governments have been made “soft” due to their retrospective—that is, fee-for-service—nature; finally, incentives of local governments to increase their tax revenue have been weakened by a kinked budget line—that is, the marginal tax surrendered to the central government is very high for grant-receiving local governments.

The fiscal decentralization that local governments are demanding at present focuses only on the first problem of the system of intergovernmental transfer. That is, they want grants with a minimum of strings attached. This is a quite understandable claim and a direction of decentralization so long as they take responsibility for the outcome of their own control over their budgets. However, fiscal decentralization would not be complete unless the remaining two problems related to grants allocation to local governments are solved. In this respect, local governments have been very inactive and made no proposal at all either to harden their budgets or to remove disincentives for increasing their own revenue. On the contrary, they claim that income support from the central government becomes more important under decentralization, to take care of those adversely affected by the reform.

Thus fiscal decentralization as it is advocated by local governments will not contribute much to raising the efficiency of their operation. In an economic environment where the population is aging and competition from neighboring countries is intensifying, Japan cannot continue its present local finance system that puts too much emphasis on guaranteeing the stake that every local government has. To this end, fiscal decentralization has to be designed so as to give strong incentives to the local governments to use their fiscal autonomy to improve equity and efficiency.3


Their paper on fiscal relation among governments is recapitulated in a shorter version in the OECD’s Economic Outlook(Joumard and Kongsrud, 2003b).

Nakazato (1999) studied the effects of public investment on regional economic growth in Japan and showed that there was no significant positive impact. Alesina et al. (1999) studied similar implicit mechanisms of transfer in Italy and concluded that they did not contribute to the development of markets in heavily subsidized regions.

In order to harden the budgets and remove disincentives of local governments, the authors have proposed the introduction of the linear transfer system, which aims at applying the idea of a linear tax system to intergovernmental finance (Tajika and Yui, 2005; Atkinson, 1996), where distribution and the incentive components are separately treated. Briefly, in the linear transfer system, the distribution part is the prospective minimum support and the incentive part is the same marginal revenue-retaining ratio. These two elements contribute to hardening the budgets and removing the disincentives of local governments.


    AlesinaA.; DanninggerS. and RostagnoM.1999“Redistribution through Public Employment: The Case of Italy,”IMF Working Paper 99/177Washington: International Monetary Fund.

    AtkinsonA.1996“Public Economics in Action: The Basic Income/Flat Tax Proposal,”Lindahl Lectures on Monetary and Fiscal PolicyOxford: Oxford University Press.

    BirdR.1999‘Treading the Fiscal Labyrinth: Some Issues in Fiscal Decentralization,” in J.Slemroded.. Tax Policy in the Real WorldCambridge: Cambridge University Press pp. 141-61.

    HanaiK.; TajikaE. and YuiY.2000“Financing Growth and Local Governments in Japan: Why Does the Government Keep Growing?,”Hitotsubashi Journal of EconomicsVol. 41 pp. 153-62.

    InmanR. and RubinfeldD.1997“Rethinking Federalism,”Journal of Economic PerspectivesVol. 11No. 4 pp. 43-64.

    JoumardI. and KongsrudP. M.2003a“Fiscal Relations Across Government Levels,”OECD Economic StudiesVol. 36Paris: OECD.

    JoumardI. and KongsrudP. M.2003b“Fiscal Relations Across Government Levels,” in OECD Economic Outlook 74Paris: OECD pp. 143-60.

    JoumardI. and YokoyamaT.2005“Getting the Most Out of Public Sector Decentralization in Japan,”Economic Department Working Paper No. 416Paris: OECD.

    KornaiJ.MaskinE. and RolandG.2003“Understanding the Soft Budget Constraint,”Journal of Economic LiteratureVol. XLIDecember pp. 1095-136.

    NakazatoT.1999“Social Capital and Economic Growth,” (inJapanese) Japanese Ministry of FinanceFinancial ReviewDecember pp. 67-84.

    OatesW.1999“An Essay on Fiscal Federalism,”Journal of Economic LiteratureVol. XXXVIISeptember pp. 1120-49.

    TajikaE. and YuiY.2005“Hardening the Budgets and Removing Disincentives of Local Governments: A Proposal of Linear-Transfer System,” (inJapanese) Financial ReviewVol. 76May.

    WildasinD.2004“The Institutions of Federalism: Toward an Analytical Framework,”National Tax JournalVol. LVIINo. 2 pp. 247-72.

    Other Resources Citing This Publication