2006/03 Research & Review

Principles and Challenges of the Public Pension System Toward Reform in 2009

NAKATA Daigo
Fellow, RIETI

With nearly two years since the Japanese public pension system underwent major changes in 2004, it may appear that the surge in public interest in the pension system has waned. But a series of recent opinion polls suggests otherwise; public interest in the pension system still remains high and it is fair to say that people continue to recognize this as a major policy issue. The Research Institute of Economy, Trade and Industry (RIETI) has undertaken a project to examine challenges that need to be overcome in enhancing the sustainability of the public pension system. In this column, I will to introduce some of our interim findings in this project.

Underlying principles

The 2004 reform of the public pension system - though little appreciated in terms of media coverage - can be called a very significant achievement compared to a series of previous reforms. In particular, the introduction of a ceiling on contribution rates and the so-called macroeconomic sliding-scale system was a major step forward. It had been anticipated that the burden of pension costs, both on households and companies, would continue to increase; the ceiling is to serve as a brake on the rise in that burden. Meanwhile, the macroeconomic sliding-scale system has provided a built-in mechanism for keeping the total payout of pension benefits at a level not exceeding the sum of premiums paid by beneficiaries and subsidy from the national treasury. Though rather poorly received in Japan, the 2004 reform has been given relatively high marks by foreign governments and international organizations, which I believe is attributable to the inclusion of the aforementioned two schemes.

Of course, this does not mean that all the challenges have been overcome. Having been revised many times over the years so as to adapt to the changing economic conditions and fertility rate, the Japanese public pension system, as it stands today, is extremely complicated in its underlying principles and this remains a major challenge. To date, Japan has managed to maintain the public pension system by making substantial changes to its institutional framework. In the process, the underlying principles of pension benefits and contributions have also changed drastically. When the government first introduced the kosei-nenkin Employees' Pension Insurance (EPI), it started as a funded pension system with premiums paid as a percentage of earnings, whereby the amount of benefits received by retirees was directly linked to the total amount of premiums paid by them and on their behalf. This funded pension scheme, however, collapsed amid inflation and economic turmoil, and then, with the enactment of the new Employees' Pension Law of 1954, it was reborn as a two-tier system wherein the first tier provides fixed-amount benefits and the second tier offers additional earning-related benefits. Because premiums are levied as a fixed percentage of earnings, the shift to the two-tier system can be described as the launch of a new pension system that incorporates an income redistribution function. Meanwhile, the kokumin-nenkin National Pension (NP), for the self-employed and those not covered by other public pension programs, was initially designed so that a married couple would receive a benefit equivalent to what a typical male employee would receive in retirement from the EPI. However, due to changes in the employment structure and the significant imbalance between pension benefits and contributions, it became extremely difficult to maintain the NP as a standalone mechanism. For this reason the program was transformed into a Basic Pension program providing for basic financial security in retirement with its significance and benefit plan design changed accordingly. Subsequent to the introduction of the Basic Pension program (which also constitutes the first-tier of the EPI), income redistribution mechanisms within the Japanese public pension system became even more complicated due partly to the establishment of the "category 3" beneficiary status. Category 3 beneficiaries - unemployed married persons such as fulltime housewives - are covered, without paying any premiums, by the public pension program in which their spouse participates. Such income redistribution has been tolerated because it is generally understood that the pension system is being operated following the social insurance formula with benefits based on the premiums paid, but at the same time that such a social insurance formula, though primarily based on the insurance principle, is being operated so as to partly incorporate the principle of mutual assistance from the viewpoint of income redistribution. If solely based on the insurance principle, no income redistribution, except that based on the use of insurance techniques for the purpose of risk diversification, would have been allowed. But it is also said that the basic principle of the social insurance formula is to guarantee all members of society a minimum standard of living by applying the principle of mutual assistance to those with no or low income.

Even though the underlying principles of the public pension system might be relatively clear in theory, the mechanisms for income redistribution remain unclear in practice and this is why it is so challenging to solve problems in Japan's public pension system. In particular, mechanisms for income redistribution between or within programs or generations have grown so complicated that how they function is difficult to understand. In connection with this point, in 2003 the Cabinet Office conducted an opinion poll on the public pension system. Asked whether they want to see a clearer link between pension benefits and contributions so that the amount of benefits depends on the total amount of premiums paid, 81.1% of the respondents said "yes" whereas only 7.9% said "no." It seems one reason people distrust the public pension system is difficulty seeing just how much of the premiums will be reflected in the amount of benefits and how much redistributed for mutual assistance.

Reform proposals with clearer principles

If the practical ambiguity of the pension system is contributing to public distrust, it is necessary to consider options in a new pension system that incorporate the advantages of both the principles of insurance and mutual assistance (which together constitute the principle of social insurance) and is designed so as to clarify how each of these principles corresponds to specific elements of the pension system. For this, the following two options may be considered:

[Full tax-funding of the Basic Pension]
The provision of Basic Pensions - those paid out of the NP and the first tier of the EPI - would be clearly defined as a scheme to guarantee each citizen a minimum standard of living based on the principle of mutual assistance, whereby the financial resources required for such payments would be fully provided from the government's general account budget or by levying special-purpose taxes.

[Integration of the NP and EPI]
The fixed-amount benefit NP, and the two-tier EPI consisting of the fixed-amount benefit first-tier and the earnings-related second-tier, would be integrated into a single new pension program that primarily operates as an earnings-related scheme but provides tax-funded "minimum security pensions" in cases where the amount of earnings-related benefits is too low to sustain a minimum standard of living.

In order to quantitatively evaluate these reform proposals, our research group developed a new simulation model for pension finance (hereinafter referred to as the RIETI model). The framework of the RIETI model follows the actuarial principle, meaning that the model uses the same framework - though somewhat inferior in terms of data availability and degree of precision - as that used by the Ministry of Health, Labor and Welfare (MHLW) in its Actuarial Valuation of the Employees' Pension Insurance and the National Pension. It is programmed in such a way that enables the pro forma calculation of what could not be calculated based solely on published information, while omitting certain calculation tasks with respect to low priority items. It has been confirmed that the RIETI model provides a close approximation to MHLW's estimation when based on the same assumptions applied in The 2004 Actuarial Valuation of the Employees' Pension Insurance and the National Pension. In what follows, the above two reform proposals will be quantitatively evaluated by using the RIETI model. We assume that socio-economic conditions mirror those applied for the standard case scenario in MHLW's The 2004 Actuarial Valuation and that the transition to a new system takes place in 2010.

Full tax-funding of the Basic Pension

First, let's consider the case in which the Basic Pension is fully funded from the national treasury. The question would arise as to how the benefit formula under the EPI, which is currently providing funds to the Basic Pension account, should be changed. Here, it is assumed that pension benefits are maintained at a designated level (set in terms of benefit multiplier) as set in the 2004 reform and the premium level is adjusted so that it can accommodate the payouts. Because it is no longer necessary to allocate part of the contributions to the EPI set aside for the Basic Pension account, it should be possible to lower the current EPI premium level (14.288%) by the amount saved via eliminating such fund transfers. For the purpose of simulating this model, pension premiums have been calculated not in terms of step-wise contribution rates but as a standardized single rate effective for the years through 2100. Meanwhile, the terminal conditions have been set so that the reserve fund in 2100 is equal to that under the standard scenario of The 2004 Actuarial Valuation with the macroeconomic sliding-scale system put into place to fulfill the conditions.

The simulation results show it is possible to substantially lower the premium rate from its current 14.288% to 11.938% if the survivors' and disability pensions are maintained within the EPI, to 8.688% if the survivors' pension is removed from the system and financed by a separate source of funding, to 8.688% if the disability pension is removed, and to 8.442% if both the survivors' and disability pensions are removed thereby transforming the current EPI into a straightforward old age pension scheme. Enrollment in the existing EPI is not mandatory for part-time employees whose work hours or days are less then three quarters of those of fulltime employees. Any hike in EPI premium rates would increase the cost of employment for fulltime employees (for whom employers are required to bear half the cost of pension premiums), thereby serving as an incentive for companies to increase their reliance on part-time employees for whom they do not have to bear the premium cost. Conversely, it is expected that lowering the premium would induce companies to employ more fulltime workers.

Also, it should be noted that the premium reduction effect of separating the survivors' pensions is extremely large. Having the Basic Pension fully financed by the national treasury would enable clear distinction between the first tier of the EPI which has an income redistribution mechanism and the second tier in which the amount of benefits and contributions paid are directly linked on a one-to-one basis. However, a pension scheme characteristic of such a one-to-one relationship would not be suitable as a mechanism for the survivors' or disability pensions. Under Sweden's notional defined-contribution (NDC) system, which has been well regarded in recent years in Japan as a successful model for pension systems, both the survivors' and disability pensions are separated from the pension system. Given the substantial differences between Japan and Sweden in the female work participation rate and the degree of income gaps, it would be extremely difficult for Japan to emulate the Swedish system. Yet, Japan would be wise to discuss far more actively the question of to what extent the life insurance type mechanism called the survivors' pension should be kept within the pension system.

Next, let me present the estimation results for the amounts of tax revenue required to sustain the Basic Pension in cases where the program is fully financed by the national treasury. Currently, expenditures required for payments of Basic Pension benefits are being financed by funds from the Basic Pension accounts of the three types of pension program - the NP, EPI, and kyosai-nenkin - as well as by funds from the national treasury (the portion financed by the national treasury will be raised to 50% from 2009). Here, for the purpose of simulation, it is assumed that half the cost of Basic Pensions, which has been designated for funding by the national treasury from 2010, would be covered by funds from the general account budget of the central government as planned. Then, based on this assumption, the consumption tax rate required for financing the portion currently funded by premiums collected under the three types of pension program (half the cost of the Basic Pension) was calculated, that is, if the portion is fully funded by additional consumption taxes levied exclusively for the purpose of financing pension programs.

In the initial period following the introduction of the new pension system, additional consumption taxes of less than 4% points would be enough to finance half the cost of the Basic Pension (see figure 1). However, the required consumption tax rate would begin to increase in 2030 and need to be raised to 6% or more in or around 2050. (It should be noted that this tax rate is additional to the existing consumption taxes which would be maintained at the current rate of 5%.) The sharp rise in the required consumption tax rate from 2030 onward is partly attributable to the inevitability that the second generation of postwar baby boomers will exit from the labor market around which time the ratio of vested pensioners to pension subscribers will begin to rise. Another factor is the expected increase in the number of pensioners receiving the full amount of Basic Pensions because, as a result of shifting to the full tax-funded Basic Pension system, the problems of nonpayment and non-subscription would be eliminated and thus the situation would become tantamount to the full payment of premiums by around 2030.

Figure 1: Expected changes in the required consumption tax rate

Figure 2: Expected changes in the reserve fund balance as a percentage of GDP)

Integration of the EPI and NP

I will discuss a proposed scheme for integrating the EPI and NP. (I will refer to it as a "new integrated pension program" to distinguish it from other possible schemes that may be implied by the term "integration of pension programs.") A full earnings-related pension system would be adopted as the basic model of the new integrated pension program. The step-wise premiums applicable to the current EPI would be used, in principle, as premiums under the new program with only exception to this being the treatment of low- or zero-income earners falling under category 1 or 3 beneficiaries. Under the current public pension system, category 1 beneficiaries (NP subscribers) and category 3 beneficiaries are eligible to receive pension benefits without paying any premiums provided they have no income. Meanwhile, the amount of benefits under the full earnings-related pension system (both the survivors' and disability pensions are removed from this system) is proportional to the income level and length of subscription period of each beneficiary, as is the case for benefits paid out from the earnings-related portion of the existing EPI. Under this system, the amount of benefits and contributions correspond on a one-to-one basis. That is, if person A's income (hence the amount of premiums paid) is twice that of person B, that person is entitled to receive twice the benefits.

Then, if the amount of earnings-related pension benefits calculated for a certain pensioner falls below the level of benefits under the existing Basic Pension, the pensioner is provided with a "minimum security pension," a scheme separate from the earnings-related program. The amount of this minimum pension is equivalent to the balance between the amount receivable under the existing Basic Pension and the amount of earnings-related pension benefits calculated by the above method. In other words, those receiving the minimum security pension are uniformly guaranteed the annuity of an amount equivalent to what they would receive under the existing Basic Pension program. The cost of providing minimum security pensions will be fully covered by funds from the national treasury, rather than relying on those collected in the form of pension premiums. Meanwhile, the initial reserve funds for the new integrated pension program will be equal to the total sum of funds reserved under the EPI and NP in 2009 (based on estimates by the MHLW). As a terminal condition, it is also provided that the amount of reserve funds in 2100 must be equal to the total payouts in 2100.

All the insured and pensioners under the EPI, including so-called tsusan-rorei pensioners (those whose subscription periods under a series of different pension programs prior to the establishment of the EPI are counted toward their subscription period under the EPI), will shift to the new public pension system - i.e. the new integrated pension program plus the minimum security pension program - in 2010. On the other hand, among those currently covered under the NP , only those born in 1970 or later will shift to the new system, discarding any records in and prior to fiscal 2009, whereas the remaining NP subscribers will stay with the existing NP. The receipts and payments under the new public pension system include those under the NP that will be maintained, as a transitional measure, for NP subscribers born before 1970. Half the cost of paying benefits to such NP subscribers is to be covered by funds from the national treasury - as would occur from 2009 onward under the existing system with planned changes incorporated - and is added up with the cost of paying minimum security pensions to calculate the total amount that needs covering by additional taxes.

The calculation results show that the benefit multiplier applied to the total amount of remuneration can be set at 11.522, as compared to the current level of 5.481, because the new system is separated from both the survivors' and disability pension. Also, it has been shown that the total amount of national treasury funds required for financing the cost of minimum security pensions and half the cost of the transitional NP can be covered by raising the consumption tax rate a maximum of 3.8%, a substantially smaller increase than would occur under the tax-funded Basic Pension scenario (see figure 1). The reason for this difference is that under the scenario based on the new integrated pension program, tax-funded benefits are provided only to low-income earners, thereby limiting the income redistribution function to the minimum level, whereas the tax-funded Basic Pension program provides tax-funded Basic Pensions to all, even those with high incomes.

However, when we look at the projected changes in the reserve fund balance, shown as a percentage of the GDP, under the new integrated pension program (see figure 2), we can see that the presence of the pension program within the Japanese economy will sharply increase as a result of substantial portions of subscribers to the non-earnings-related NP shifting to the earnings-related new integrated pension program. It is projected that at some point the reserve fund balance under the new system would become as large as the GDP; this is twice the level projected under the existing public pension system and thus it is expected that the new pension system would have a significant impact on the capital market. This enormous increase in reserve funds is a result of adopting the step-wise premium as designated by the 2004 pension system reform. One of the key characteristics of the 2004 pension reform can be seen in the adoption of a "finite balance system," which calls for accumulating as many reserve funds as possible before around 2045 by raising the premium rate at a relatively quick pace so as to manage the cost incurred in the subsequent 50 years. This characteristic would be inherited by the new integrated system provided that the integration between the NP and EPI merges the former into the latter. If we try to equalize the burden across generations in Japan where population is rapidly aging, the premium rate must quickly be raised to and kept at an adequate level. This, however, would result in an enormous amount of reserve funds. If so, then, the next important question would be how and in what sort of portfolio the funds of such an enormous size comparable to the GDP can and should be managed.

In this column, I have presented the results of some proposed options for public pension system reform with clarified underlying principles and considered challenges remaining for each option. Of the pension programs for the employed, the kyosai-nenkin Mutual Aid Pension for government officials, public school teachers and so forth could not be included in the analysis due to the constraint of data availability. In the past several years, however, we have been witnessing rapid improvement in the disclosure of information on the kyosai-nenkin whereby more and more data are being made available for use. Thus, incorporating Mutual Aid Pension into the model is our next challenge.

>> Original text in Japanese

June 8, 2006

Article(s) by this author