33057 The Public Pension System in Japan The Consequences of Papers Rapid Expansion Eiji Tajika orking Funded by: The Policy and Human Resources Development Trust Fund W of the Government of Japan WBI The Public Pension System in Japan The Consequences of Rapid Expansion Eiji Tajika The paper focuses on the public pension system in Japan and traces its development against the overall background of the socioeconomic developments since 1950s. The country's establishment of a public pension system that covers the entire population took place fairly recently. It was only in 1961 that the compulsory pension plan for the self- employed was introduced. Together with the already existing income-compensation scheme for retired civil servants and various employees' pension plans, every Japanese gained access to a public pension plan in one form or another. Just because the public pension was installed late, however, does not mean that its financial status was sound. On the contrary, the social security reform in the early 1970s changed the landscape of Japanese public pension virtually overnight. Put simply, benefits overshot prospective contributions. The paper recommends that since the unfounded pension systems such as the one adopted by Japan are at the mercy of demographic changes, developing countries should adopt pension systems with sound actuarial basis. World Bank Institute Copyright © 2002 The International Bank for Reconstruction and Development/The World Bank 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. The World Bank enjoys copyright under protocol 2 of the Universal Copyright Convention. This material may nonetheless be copied for research, educational, or scholarly purposes only in the member countries of The World Bank. Material in this series is subject to revision. The findings, interpretations, and conclusions expressed in this document are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or the members of its Board of Executive Directors or the countries they represent. The Public Pension System in Japan The Consequences of Rapid Expansion Eiji Tajika 2002. 26 pages. Stock No. 37203 Contents Foreword v The Japanese Public Pension System 2 Projections of Population Aging in Japan 5 The Employees' pension Insurance 9 The National Pension Scheme 14 Issues Related to the Public Pension System 17 Conclusion 20 References 21 Foreword This paper was prepared for a project on Social Development in East Asia. The project was organized by the World Bank Institute under the auspices of the Program for the Study of the Japanese Development Management Experience (The Brain Trust Program), which is financed by the Policy and Human Resources Development Trust Fund established at the World Bank by the Government of Japan (GOJ). We greatly appreciate the generous support from GOJ. The principal objectives of this Program are to conduct studies on Japanese and East Asian development management experience and to share the lessons of this experience with developing and transition economies. The experiences of other countries are often covered in order to ensure that these lessons are placed in the proper context. This comparative method helps identify factors that influence the effectiveness of specific institutional mechanisms, governance structures, and policy reforms in different contexts. A related objective of the Program is to promote the exchange of ideas on development among Japanese and non-Japanese scholars, technical experts and policy makers. The papers commissioned for this project cover a number of important issues related to Japanese Social Policy. These issues include, among other, Japanese pension system, public health insurance in Japan, evolution of social policy in Japan, role of families, communities and government in improving socioeconomic performance, and Japanese policies towards poverty & public assistance. We hope that these papers will provide important policy findings and lessons for policymakers in developing countries and other development stakeholders. Tsutomu Shibata Advisor & Program Manager of the Brain Trust Program World Bank Institute 4 6 The Public Pension System in Japan: The Consequences of Rapid Expansion Eiji Tajika Graduate School of Economics Hitotsubashi University A characteristic of Japanese social security policy throughout its high growth period (1950­73) was that the government was not big. Social expenditure was focused mainly on education and basic health care, and thus it aimed more at social stabilization and economic growth than the establishment of a welfare state. Issues such as when the nation would be able to rebuild after the devastation of World War II, how to raise savings, and how to finance investment preoccupied the minds of those contemplating the future of the country. And the public was busy, too, in working their way to a better life. However, a sharp change in social security policy took place at the beginning of the 1970s. Along with the increase in their incomes, people in Japan became more conscious of quality of life issues, exemplified most clearly by concerns about the environment. Social welfare was not an exception, and a call for better welfare was turning into a movement. Better welfare meant more government commitment to medical care and public pensions. In retrospect, these calls rested on an optimistic view that economic growth would continue and that the government would be able expend considerable energy in the service of its people and the welfare system the public deserved. The year 1973 was declared by the government to be "the inaugural year of welfare," and extremely liberal provisions were offered in every aspect of social expenditure, including medical care and public pension. Overlooked in all this euphoric movement, however, was that the Japanese economy had turned a corner.1 The two oil shocks in the 1970s triggered high inflation and economic growth began to slow down. A problem of social security, at least viewed financially, is that its provisions are typically indexed to price increases; that is, their burden is not eroded by inflation. In addition to this, the decline of the fertility rate and the prolongation of life expectancy had turned Japan into one of the oldest populations in the world. The outcome of these developments is that today the government can only make ends meet by issuing more bonds or by raising more taxes. In view of the difficulty of generating more tax revenues, chances are that the costs of sustaining the present welfare system will be shifted to younger generations. The purpose of this paper is to focus on the public pension system in Japan and to trace its development against the overall background of the socioeconomic developments summarized above. The country's establishment of a public pension system that covers the entire population took place fairly recently. It was only in 1961 that the compulsory pension plan for the self-employed was introduced, and together with the already-existing income-compensation scheme for retired civil servants and various employees' pension plans, every Japanese gained access to a public pension plan in one form or another. 1 In other papers by Tajika and Yui in the World Bank Institute (WBI) Working Paper Series the real growth rates of social expenditure and the gross domestic product (GDP) are compared. The results are that during the high-growth era social expenditure and the GDP expanded at the same rates, but that after the rapid-growth period the rate of social expenditure expansion stayed much higher than the growth rate of the GDP. Consequently, the government expanded after the high-growth era 1 2 Eiji Tajika Just because the public pension system was installed late, however, does not mean that its financial status was sound. On the contrary; the aforementioned social security reform in the early 1970s changed the landscape of Japanese public pension virtually overnight. Put simply, benefits overshot prospective contributions. The financial condition of the Japanese public pension system has been going from bad to worse. The higher-than-expected high speed of aging is another reason for the financial disaster, but the true culprit is the unfunded promises that the government made and its people accepted at the dawn of a welfare state. It was easy to scatter lucrative benefits when the country was young and the pension system was not yet mature. But Japan is now facing the hard reality of coping with a financially unbalanced pension system. In exploring these issues this paper proceeds in the following manner. There are seven sections including this introduction. The second section will touch on the institutional aspects of the Japanese public pension system. The overall structure of the National Pension plan for the self-employed and the Employees' Pension Insurance will be explained first, and then the formulas for determining benefits will be presented. The third section discusses population aging in Japan, and shows how it affects the future of the public pension system. The fourth section will focus on the Employees' Pension Insurance, by far the largest of the public pension plans, and examine how its finances have turned into such a mess over the years. The fifth section will study the National Pension scheme from a similar perspective. The sixth section will consider three issues related to public pension policies in Japan: the cross-subsidization between public pensions of the self-employed and employees' pensions; public pension assets and the public investment and loan program; and the status of women in pension policy. The last section will conclude the paper. The Japanese Public Pension System The Overall Structure Japan's public pension system consists of several plans. They can broadly be classified into the following two schemes: the National Pension plan for the self-employed and various pension plans for employees. The employees' plans are further divided into the Employees' Pension Insurance (EPI) and Mutual Aid Associations; the EPI is an insurance plan for (mostly) private employees and the Mutual Aid Associations are collections of associations for civil and semicivil servants. The employees of formerly state-run, but now privatized enterprises such as Japan Railway Companies and Nippon Telephone and Telegram, and the teachers and staffs of private schools belong to Mutual Aid Associations of their own. The insured are divided into three groups under Japan's public pension plans. Type I refers to those under the National Pension scheme. Type II individuals are members of the EPI and the Mutual Aid Associations. Type III insured persons are dependent spouses of Type II contributors--namely, spouses whose incomes are below the minimum taxable income. This category of the insured was introduced by the 1986 reform. Prior to the reform the employees' dependent spouses contributed either voluntarily to the National Pension or were simply left uninsured (except for survivors' benefits). The 1986 reform treated this last type of people as a separate group and put them under the public pension umbrella, although they are exempt from contributing to public pension plans. An overview of the Japanese public pension system is shown by table 1. Here, various pension plans are compared with respect to their numbers of contributors and beneficiaries, the total amount of benefits disbursed by each plan, and monthly average benefits per beneficiary. Several numbers of the Type III­ insured are missing, because they are included elsewhere in the table and cannot be sifted out. The Public Pension System in Japan 3 Table 1. An Overview of the Public Pension System in Japan, 1995 Type I Type II Type III Types of National pension Employees' Mutual aid Type II's the insured plan pension insurance associations dependent spouses Number of contributors (million people) 19.1 32.8 5.8 12.2 Number of beneficiaries (million people) 11.2 13.6 3.5 n.a. Total amount of benefits (billion yen) 7,745 16,355 7,451 n.a. Monthly average benefits (thousand yen) 44.7 170.1 219.3 n.a. n.a. Not applicable. Note: The categories of the insured are as follows: Type I: The self-employed--for example, small proprietors, farmers, and professionals. Type II: Employees of both private and public institutions. Type III: Dependent spouses of Type II individuals. This group was established in 1986. Before the reform, people in this category paid into the National Pension plan, or went uninsured when they did not. Starting in 1986, they were grouped under Type III. And while they were granted the right to receive pension, they were exempted totally from contributing to public pension plans. Source: Social Insurance Agency (1997). The total number of the insured of all types is about 70 million as of 1995 and this is about 98 percent of the population aged 20­59. The rate of insurance coverage is extremely high; every Japanese over 20 years of age has to belong to some plan of public pension. The EPI is the largest of the public pension plans, consisting of about 32.8 million contributors and 13.6 million beneficiaries in 1995. Its disbursement of benefits in 1995 amounts to 16,355 billion yen2 and constitutes 3.4 percent of the nation's Gross Domestic Product (GDP). The EPI is the least mature of the public pension plans, supported by more young cohorts than in other plans: two-and-a-half contributors support one beneficiary, while in other plans about one-and-a-half contributors have to take care of one beneficiary.3 The benefit per beneficiary is much higher in the employees' plans than in the National Pension scheme; on average the EPI and Mutual Aid Associations' beneficiaries receive four to five times as much in benefits as their counterparts in the National Pension scheme. This is due to a sizeable difference between the respective rates of contribution and the average length of the period over which contributions are made. The contribution to the National Pension scheme is a fixed charge per contributor; the monthly amount is 11,700 yen as of 1995. The contributions to employees' plans are proportional to remuneration; 2 This is equivalent to about US$150 billion. In what follows, the currency unit will be Japanese yen. For convenience, a U.S. dollar in 1995 was exchanged for about 110 yen. 3 The maturity of the Mutual Aid Associations is dispersed among associations. For example, in Japan Railway Company's association, a contributor has to support more than one beneficiary and thus the association cannot sustain itself without financial aid from other associations. 4 Eiji Tajika the rate of contribution (actually, the social security tax rate) is 17.35 percent as of 1995.4 And the average contribution of the EPI amounted to about 47,000 yen per month in 1995. Historically, the benefits of the Mutual Aid Association scheme have been set higher than the EPI and this explains the difference of benefits among employees' pension plans, though this gap is narrowing. The Benefit Structure A complicated feature of the Japanese public pension system is that pension plans have developed independently. Civil servants' plans, like in other countries, are the oldest. Employees' pension plans came next: a scheme was established in 1942 in the midst of World War II, and one of its expected roles was mobilizing savings for fighting the war. The National Pension scheme was established much later in 1961, and at that point the public pension plan was made accessible to every Japanese. Thus, the contribution and benefit arrangements show historical variations from one plan to another. The Japanese public pension system has developed from separate schemes, and as our preceding explanation has shown, the system consists broadly of the self-employed and employees' plans. A major reform, however, was made in 1986 when the "Fundamental Pension" (Kiso-nenkin, in Japanese) was introduced. The idea of this reform that was offered by the government was to provide the same amount of benefits to a beneficiary, no matter which plans he or she had belonged to over the course of his or her working period. While the pension scheme's "Fundamental" portion may look like the first tier of the public pension system and the second tier (in which benefits are proportional to remuneration), appears to be the one that is made available to employees, that is not an accurate picture. A new, universal Fundamental Pension was never established; instead, a part of employees' benefits was notionally defined as the Fundamental portion and made equal to the benefits offered to the self-employed. Thus, the self-employed continued to pay into their National Pension scheme at fixed rates, and employees paid into their plans at a rate that was proportional to their remuneration. Since the two groups of people continue to contribute to their own plans according to different schemes, a special pension-clearing account was set up to deal with Fundamental Pension benefits. This is the Fundamental Pension Account, and the total benefits paid out as the Fundamental Pension are arranged so that they could be shared by all public pension plans according to the number of contributors, not of beneficiaries. Moreover, the Type III­insured (the employees' dependent spouses) are regarded as contributors belonging to the employees' pension plans, although they make no contributions at all to any plan. Thus the Fundamental pension deals only with the benefit side, and no reform for unifying the contribution method has been introduced. As we will see later in this paper, a similar reform helped shift the huge burden of the National Pension scheme to the much bigger and more recently established employees' plans. Now the benefits are determined in the following manner. To begin with, the amount of benefits in the Fundamental Pension portion is proportional to the length of the contribution period. The maximum length one can contribute is 40 years--between ages 20 and 59, and the maximum benefit is determined first as the basis for all other calculations. In 1995 this was about 65,000 yen per month in return for 40 years of contributions; so if one has contributed for 30 years, one can get three-quarters of the maximum, or about 50,000 yen per month. The benefits are indexed not only to consumer prices each year, but every five years or so they are adjusted to match the real living standard of households. 4 There is a remuneration ceiling for the mandatory contribution to pension plans. In Japan, the bonus payments are customary and they constitute about a quarter to a third of total income. Compulsory contribution did not apply to this portion of one's income in the past. However, starting in 1994, a 1 percent contribution out of bonus payments was made mandatory. The Public Pension System in Japan 5 The benefits of employee pension plans consist of two parts: the Fundamental Pension and the proportional remuneration part. The first part was explained above. For the second part, average remuneration over the insurance period is calculated with proper indexation rates applied to past incomes; and the benefit can be expressed as: (0.75 percent) × (average monthly remuneration) x (total months of insurance). Thus calculated, the benefit turns out to be a rather large proportion of current contributors' wages. According to an estimate by the Ministry of Health and Welfare, 40 years of contributing to the EPI will provide the insured individual (with dependent spouse) with almost 70 percent of the average remuneration of workers who are supporting the retirees. Actually this replacement ratio was even higher before the 1986 reform; the prereform formula would have entitled the retiree to more than 80 percent of a contemporary worker's regular income. However, we will argue later in this paper that even the present formula of reduced benefits is too generous, because it pushes the financial burden of the EPI beyond a manageable limit that workers of future generations can afford to carry. Projections of Population Aging in Japan This section examines the aging trend of the Japanese population and shows how demographic changes over the coming years will make pension financing difficult. Since Japanese public pension plans are financed essentially by the pay-as-you-go method and the ratio of pension benefits to the wage of working population (defined as the replacement ratio) is very high, population estimates are crucial for formulating Japanese pension policy. With this in mind, we would like to start with estimates of the total fertility rate--that is, the average number of children a women gives birth to. Table 2 shows three estimates: the 1986 and 1992 estimates reported by the National Institute of Population and Social Security Research (which is part of the Ministry of Health and Welfare) and 1994's IFES estimate reported by the Institute of Fiscal and Economic Studies, an independent research organization. The Ministry reviews pension policy about every five years, and the latest reviews were done in 1989 and 1994. The 1986 estimate and the 1992 estimate, respectively, were used for these two reviews. When examining population trends in Japan, the most striking factor would likely be the extremely sharp decline in the fertility rate. In this regard, the 1986 estimate was too "optimistic" and eventually misled the Ministry to propose too lenient an increase in the contribution rate and only modest cuts in benefits. It even projected the fertility rate for 1990 to be 1.84, though it in fact went down to 1.54. Moreover, the fertility rate was projected to rebound quickly and go over 2. The 1992 estimate gave a much lower fertility projection. The IFES estimate looked at the fertility trend more cautiously and set its level in 2000 at about the one reported in the Tokyo metropolitan area at the end of the 1980s. Each of these estimates anticipates later increases in the fertility rate, but the IFES estimate gives the lowest figure among the three. Although the IFES estimate looks too cautious in reading the future course of the fertility rate, the recent 1994 estimate that was provisionally released by the National Institute of Population and Social Security Research comes closer to the IFES figures. Now, what will be the future population trends according to the three estimates? Table 3 shows their projections for the share of those aged 65 or over in the total population (the numbers for the years up to 1980 are actual data). In conjunction with this table, figure 1 illustrates the movements of the population index over the period 1950­2050. The table and the figure show clearly that Japan as a society is and will be getting older, and that this process is happening extremely fast, and is therefore highly significant. 6 Eiji Tajika Table 2. Total Fertility Rate by Three Estimates The IFES Year Actual The 1986 estimate The 1992 estimate estimate 1950 3.65 n.a. n.a. n.a. 1960 2.00 n.a. n.a. n.a. 1970 2.13 n.a. n.a. n.a. 1980 1.74 n.a. n.a. n.a. 1990 1.54 1.84 1.54 1.53 2000 n.a. 1.96 1.60 1.37 2010 n.a. 1.98 1.78 1.36 2020 n.a. 1.99 1.80 1.42 2030 n.a. 2.01 1.82 1.52 2040 n.a. 2.02 1.86 1.63 2050 n.a. 2.04 1.91 1.71 n.a. Not applicable. Note: The 1986 and 1992-estimates are those reported by the National Institute of Population and Social Security Research (Ministry of Health and Welfare), in 1986 and 1992 respectively. The IFES estimate is the one made by the Institute of Fiscal and Economic Studies, an independent research organization, in 1994. Sources: National Institute of Population and Social Security Research, Ministry of Health and Welfare; IFES. Table 3. The Share of the Population Aged 65 and Over by Three Estimates (percent) Year Actual The 1986 The 1992 The IFES estimate estimate estimate 1950 5.0 n.a. n.a. n.a. 1960 5.7 n.a. n.a. n.a. 1970 7.1 n.a. n.a. n.a. 1980 9.1 n.a. n.a. n.a. 1990 n.a. 11.9 12.0 12.0 2000 n.a. 16.2 16.7 16.9 2010 n.a. 19.9 20.9 21.5 2020 n.a. 23.5 25.2 26.6 2030 n.a. 23.0 25.4 27.7 2040 n.a. 24.0 27.3 31.0 2050 n.a. 23.4 27.4 32.7 n.a. Not applicable. Sources: National Institute of Population and Social Security Research, Ministry of Health and Welfare; IFES. The Public Pension System in Japan 7 Figure 1. The Share of Ppopulation Aged 65 and Over by Three Estimates Source: IFES (1992). National Institute of Population and Social Security Research (1986, 1992). People aged 65 and over were only 5 percent of the total population in 1950, but the share had already more than doubled by 1990. The 1986 estimate projected the share to reach 25 percent, whereas the 1992 estimate foresaw 28 percent and the IFES estimate projected more than 30 percent. Population aging will be thus fast, but we have to note also that it will not peak about the year 2020 when baby boomers reach their seventies. Since the fertility rate will decline sharply, the process of population aging will continue deep into the middle of the twenty-first century. Another population index that will be useful for considering the finance of public pension is the share of the population aged 65 and over to the working population--for example, the people aged 20 to 59. Table 4 and figure 2 show the actual numbers and projections of the index over the period from 1950 to 2050. Here the population estimates relate more straightforwardly to pension problems. According to the 1986 estimate, for every 100 working people, there would be about 50 (maximum) nonworking, aged individuals. However, the actual numbers turned out to be 60 and eventually more than 70 in the 1992 (Ministry) and the 1994 (IFES) estimates, respectively. Now, a simple bit of arithmetic will do to gain a feel about the difficulty of pension finance in Japan. The preceding section outlined the EPI's benefit formula and showed that the benefits would be as high as 70 percent of the wages of the working population. Combining this replacement ratio with the above "65 to 20­59" ratio enables us to do the following calculation: When 100 workers support 50 old people (as in the 1986 estimate) and the replacement ratio is 70 percent, the working people will have to sacrifice 35 percent of their wages to balance the benefits and contributions in the pension account. Similarly, when the same 100 working people have to pay for the pension of 60 and 70 retirees (as in the 1992 and 1994 estimates), corresponding social security tax rates will respectively surge to 42 percent and 49 percent. 8 Eiji Tajika Table 4. The Share of the Population Aged 65 and Over in the Population Aged 20­59 by Three Estimates (percent) Year Actual The 1986 estimate The 1992 estimate The IFES estimate 1950 10.6 n.a. n.a. n.a. 1960 11.2 n.a. n.a. n.a. 1970 12.5 n.a. n.a. n.a. 1980 16.1 n.a. n.a. n.a. 1990 n.a. 21.3 21.3 21.3 2000 n.a. 29.9 29.9 29.9 2010 n.a. 41.1 41.5 41.5 2020 n.a. 48.6 51.9 52.8 2030 n.a. 47.7 52.9 56.6 2040 n.a. 52.0 60.7 69.5 2050 n.a. 48.4 60.0 74.6 n.a. Not applicable. Sources: National Institute of Population and Social Security Research, Ministry of Health and Welfare; IFES. Figure 2. The Share of Population Aged 65 and Over in the Population Aged 20­59 by Three Estimates Source: IFES (1992). National Institute of Population and Social Security Research (1986, 1992). The Public Pension System in Japan 9 Of course, this is only a back-of-the-envelope calculation, and hardly an accurate estimate of the future burden on the younger population. As we get deeper into public pension financing and as real problems unfold, we will find that the actual situation is even more dire than hinted at by the simple projections here. The Employees' Pension Insurance This section studies the financial status of the EPI and shows how its difficulties have developed. The combination of high benefits and the sharp increase in the number of retirees that a working generation has to support will lead to a skyrocketing rise in the contribution rate. In fact, the Ministry of Health and Welfare projects it to rise more than 30 percent by 2020. Thus, the EPI will continue to shift the burden of pension benefits to younger generations. An outcome of this will be an extremely unfair distribution of burden over cohorts. Younger generations will certainly lose by joining the public pension plan. This section will show how serious the problems will become. We start with the benefit-side of the story. Figure 3 compares average benefits of the EPI, average wages of the working population, and the Consumer Price Index (CPI). To make the comparison easier, the value of the three variables are set to 100 in 1970. The figure shows how fast and how high the pension benefits went up; it was 116.7 in 1971, jumped up to 269.7 in 1973, and to 488.7 in 1976. That is, the benefit more than doubled overnight in 1973. In fact, the nominal rate of increase of benefits in 1973 was 131.1 percent, whereas the CPI increase rate was 16.3 percent; hence the benefits ballooned more than twice in real terms. Figure 3. Average Benefits of the EPI, Average Wages and Consumer Price Index (1970 = 100) Wages, too, increased greatly over the years, but pension benefits surpassed their ascent. The numbers in figure 3 show this: while pension benefits and wages started from 100 in 1970, they reached 1190.1 and 518.7, respectively, in 1993. One reason for this increase in benefits was the huge upsurge in the 1970s, but another 10 Eiji Tajika has been that every five years or so, the benefits are readjusted to catch up with the productivity increase. That is, the benefits so far have been indexed to wages, but the adjustments have been lagging by several years. To continue with the story of benefits, the CPI reached 303.6 in 1993, which means that pension benefits increased almost four times as much. However, we have to mention that these benefits would have been even higher if they had not been pruned by reforms. A major reform of this kind was passed in 1986, slashing benefits by about a quarter. However, it took years to fully implement this cut, so there is no clear indication that the rise of benefits will ultimately subside. To compound the effect brought on by the liberal provision of benefits, people in pension plans are getting older. An index that can relate aging to pension finance would be the ratio of pensioners to workers. The proportion of people aged 65 and over to the working population aged 20 to 59 is such an index, and we discussed it in the preceding section (table 4 and figure 2). This section examines a similar index in terms of the EPI. We call the proportion of EPI pension receivers to its contributors the Degree of Maturity, and we report our estimates in table 5. When estimating this index, greater care is necessary (compared with the establishment of previous indexes), because there is more then one type of pension receiver. There are at least two more to consider in addition to the normal old-age pension receivers when estimating the maturity of the EPI. The first of the two types is those who have contributed to the EPI less than 25 years, and who therefore may not be eligible for the Fundamental Pension that requires a minimum 25 years of contribution. The second type is the group of the surviving spouses. We assign a weight of 1 to normal receivers, 0.2 to partial pension receivers, and 0.5 to surviving spouses. These weights correspond approximately to the amounts of benefits claimed by respective pensioners. The Effective Degree of Maturity modifies the Degree of Maturity, taking these factors into consideration. Table 5 reports the two measures of population aging in the EPI, using the aforementioned 1992 and IFES estimates. (The 1986 estimate is dropped here because of its overly optimistic reading of future population trends. Besides, the government itself has switched from the 1986 estimate to the 1992 estimate for its 1994 review of pension policies.) Table 5. The Effective Degree of Maturity of the EPI (percent) The 1992 estimate The IFES estimate Degree of Effective degree of Degree of Effective degree Year maturity maturity maturity of maturity 1995 20.65 28.48 20.69 28.53 2000 24.55 34.92 24.61 35.00 2010 35.88 58.82 35.99 53.97 2020 42.98 68.79 43.68 69.90 2030 48.81 79.60 52.57 85.78 2040 58.10 94.36 67.10 108.98 2050 58.19 96.60 71.85 119.19 Note: The Degree of Maturity is defined as the ratio of the number of EPI pension receivers to the number of contributors. And the Effective Degree of Maturity modifies the Degree of Maturity by giving relevant weights to different classes of pension receivers: the pensioner himself or herself is given a weight of unity, the beneficiaries who contributed to the EPI less than 25 years are assigned 0.2, and survivors 0.5. The weights approximately represent the per person amount of relevant pension benefits. Source: National Institute of Population and Social Security Research, Ministry of Health and Welfare; IFES--from Tajika, Kaneko, and Hayashi (1996: 95). The Public Pension System in Japan 11 The Degree of Maturity corresponds to the proportion of people 65 and over to those from the 20­59 age group in table 4. In fact, the two indexes yield more or less the same results as the 1992 and IFES population estimates. However, the picture gets murky when we count the effective number of pensioners that a working person supports. In either the 1992 or the IFES estimate the projections are startling; a working person will have to support almost one pensioner, and may end up having to support even more. Recall that the full benefit a pensioner collects after 40 years of contribution to the EPI is about 70 percent of the average wage of the working population. And since the Effective Degree of Maturity goes up to 1 (or potentially higher), the resulting burden for the younger generations may reach 70 percent of their wages or more. This is of course the worst scenario for financing the EPI, but it should be clear that the Japanese public pension system has a huge time bomb with precious little time left on its ticking clock. Precautionary steps necessary to cope with this problem are either (a) to slash the benefits, or (b) to start raising the rate of contribution as soon as possible. Indeed the government has been proposing reforms toward these two goals. The age at which a retiree can pick up benefits has been set to be raised gradually from 60 to 65, and the index applied to past wages for assessing the lifetime average wage has been turned into a "net-of-taxes" index--that is, it deals with after-tax, not gross wages. A more straightforward remedy for pension financing is to increase the contribution rate. Figure 4 shows the schedule of the rate increase proposed by the government in its latest public pension review (Ministry of Health and Welfare 1995). In the figure, the rate starts from 6.2 percent in 1970, and it stands at 17.35 percent in 1997. This rate is split into two halves, and they are paid respectively by employers and employees. Figure 4. Contribution Rate of The EPI Source: Ministry of Health and Welfare (1995). 12 Eiji Tajika Line A in figure 4 traces the actual rates from 1970 to 1997 and the government's latest proposals for the succeeding years. The rate will reach 29.5 percent in 2020 and will stay at that level.5 While this rate itself looks very high, our preceding discussion on demographic structure and the level of pension benefits to wages would rather suggest that the proposed rates are rather modest. Line B, which starts in 1994 and jumps up to 26.1 percent in one stroke, is the rate-increase schedule that will collect the same amount of contribution in present value as in the schedule along line A. Thus, the proposed schedule on line A pushes the rate increase farther to future dates, and will increase the burden on younger generations. Line C sets the contribution rate at 22 percent, which would have collected the same amount of contribution as in the preceding two schedules, if the rate had been set at this figure from the beginning of the EPI. In short, it is clear that the speed of the rate increase has been far behind a schedule that would have enabled the balancing of the employees' pension plan budget. Now, what will be the combined outcome of benefit cuts and ever-increasing contribution rates? That should result in shifting more burden to younger generations. In other words, the proportion of the lifetime present-values of benefits (called B) to contributions (called C) will decline over generations. Table 6 gives estimates of this ratio, B/C, for the cohorts in the EPI who had already received pensions at the time of the 1986 reform (when the Fundamental Pension was introduced). The estimates are our own, and they are the average of the B-C ratio for men and women of each cohort. More details about the estimation method are given by Tajika, Kaneko, and Hayashi (1996). Table 6. Lifetime Benefit-To-Contribution Ratio B/C Year of birth Male Female 1901 14.07 7.17 1905 13.96 9.83 1910 11.59 13.57 1915 12.26 16.85 1920 7.70 13.16 1924 5.45 10.74 1925 5.07 10.17 Note: B: Lifetime discounted value of benefits. C: Contributions. Source: Tajika, Kaneko, and Hayashi (1996: 67). The cohorts in the table belong to the generations who saw benefits grow faster than contributions. Yet, the results are surprising; some of the cohorts here received lifetime benefits amounting to more than 10 times as much as what they had paid into the EPI. We could have anticipated this based on our observations so far in this section, but the numbers here were somewhat beyond what we would have guessed. And they show clearly how generous the benefits were in comparison with contributions. Now we have to ask ourselves how excessive benefits to older generations can be financed. We have seen some cuts in benefits and (a much sharper) increase in contribution rates. After all these reforms, how much will the B-C ratio of younger generations fall? There are several estimates on this and they differ from one estimate to another, depending on years of contribution, lifetime wages, and personal variables such as whether one is married or not. Table 7 shows results that were originally supplied by the Ministry 5 The employees used to contribute out of their monthly wages; however, as was mentioned in second section on the structure of the pension system, 1 percent of the bonus was set aside to be contributed to the EPI starting in 1994. The Public Pension System in Japan 13 of Health and Welfare. Assumptions in the example illustrated here are as follows: a man covered by the EPI pays his contributions for 40 years (the maximum length) and receives average wages; when he gets married his wife (who is three years younger) becomes a dependent and collects survivors' benefits after her husband's death. This benefit is three quarters of the remuneration-proportional benefits of her husband's; and both the husband ant the wife have an average life span. In short, the benefits assumed in table 7 are the maximum for the average-wage earners in each cohort. As a comparison I have estimated the B-C ratio for a man who has stayed single throughout his life. Table 7. Lifetime Benefit-To-Contribution Ratio Married Nonmarried Year of birth B/C (B­T)/C B/C 1924 7.78 6.32 5.06 1934 3.78 3.07 2.46 1944 2.08 1.69 1.36 1954 1.40 1.14 0.91 1964 1.11 0.90 0.72 1974 0.94 0.76 0.61 1984 0.85 0.69 0.55 1994 0.81 0.66 0.53 Note: 1. B and C are the lifetime discounted values of benefits and contributions. T is the subsidy granted by the government to beneficiaries. 2. The benefits of pensioners with dependent spouses include their own pensions, the spouses' fundamental pension, and survivors' pension. An assumption made here is that the nonworking spouses are three years younger than the pensioners. 3. Contributions are assumed to be borne fully by the employees, although they are paid into the insurance agency by employers and employees, each contributing a half. This corrects the original numbers reported by the Ministry of Health and Welfare. Source: Ministry of Health and Welfare (1995: 252). The B-C ratio in Table 7 is indeed higher than the corresponding one in table 6, which reports the average of the ratio using the EPI's annual data. For instance, a man born in 1924 receives 7.78 times as much of his contribution over his lifetime in table 7, whereas the corresponding ratio is 5.45 in table 6. The Economic Planning Bureau (1996) presented similar indexes recently in its annual White Paper, and its B-C ratio is lower than in table 7; for example, the B-C ratios estimated by the Bureau for cohorts born in 1960 and 1990 are, respectively, 1.1 and 0.68. Granting that the numbers in table 7 are overestimated as averages for the cohorts listed in the table, we are still able to discern that the public pension scheme is getting to be extremely unfair for younger generations. Cohorts born in the year 1970 or later will have to contribute to this unfair pension plan compulsorily. Moreover, we have to note that a third of the benefits of the Fundamental Pension is subsidized by the budget. We call this portion of the benefits T and have calculated the net benefit to contribution rate as (B­T)/C, which is shown in the second column of the "Married" scenario. The B-C ratio breaks even for cohorts of individuals born in 1959. The cutoff year of birth is pushed farther into the past for a single individual. Those born about 1950 may pay into the plan more than they get from it. Thus, doubling benefits in the early 1970s turned out to be a nightmare the EPI could not get 14 Eiji Tajika rid of, and its costs, so far, have been continuously postponed and passed on to younger generations. In a sense, the children of Japan now are born with a debt that has been accumulated by their grandparents and great-grandparents. The National Pension Scheme We turn now to the National Pension scheme--in other words, public pension for the self-employed. Unlike employees' pension plans, monthly contributions are fixed in this scheme and benefits can only be received if the length of the insurance period is 25 years or longer. Those whose income is below taxation limit are exempt from contribution, but can retain the status of "Type I insured," which means they are insured by the National Pension scheme even though they are not required to contribute. The period over which contributions are exempted is deemed to be a third of the entire period over which contributions are funneled into the system. The share of the insured exempted from contributions is about 15 percent of the total. Moreover, a serious problem for the National Pension scheme is that a large proportion of the self- employed do not contribute to it; one of every six to seven self-employed individuals has reportedly not contributed to the Pension scheme, even though this is mandatory. While there are many self-employed individuals who deliberately stay away from public pension schemes, a peculiar aspect of the National Pension scheme is that it demands everyone over 20 who is not paying into other plans to join. Since 1991, even students who depend on their parents financially have to be registered as "Type I insured," with contributions being paid by their parents. The government seems to desperately want to increase the number of Type I individuals and thus increase the total amount of contributions. How is the National Pension scheme doing financially? As in the EPI, we start by comparing the rates of the benefit increase and the Consumer Price Index. Figure 5 sets average benefits and the price index to 100 in 1971, and tracks their trends thereafter. The overall picture is very similar to that in the EPI diagram (figure 3). The rise of benefits in 1973 was also spectacular--more than doubling overnight. As in the EPI, the benefits are not only completely guarded from inflation, but also every five years or so they are raised to match the increase in the cost of living. Over the period from 1971 to 1994 in the figure, the benefits reach about ten times as much as their initial amount, while the price index only triples. Thus, here we see again that pension benefits have been distributed to retirees in an extremely liberal fashion. The aging process of the members of the National Pension scheme is hard to examine, for the Social Insurance Agency stopped reporting the number of beneficiaries in the scheme after the 1986 reform (which introduced the Fundamental Pension). The Degree of Maturity, which we showed for the EPI earlier, is thus not available here. However, the estimate of the index is available for the entire Fundamental Pension, which consists of the three insured types. The Degree of Maturity in this extended pension scheme is more or less the same as in the results obtained from the population estimate itself; that is, 100 contributors will have to support 50­60 pension recipients in 2020 or later (Ministry of Health and Welfare 1995: 214). The people covered by the National Pension scheme will probably be aging faster than the members of other pension plans, but in later years the population structure will cause these subgroups to converge. The lucrative benefits and the process of rapid aging in the National Pension scheme should have called for cuts in benefits and an increase in the amount of contributions. A brutal cut in benefits indeed took place in 1986 along with the introduction of the Fundamental Pension. The idea behind the reform was to make the full pension available for 40 years of insurance payment period instead of 30; thus, a quarter of the benefits was slashed, though this was of course on paper--the actual process of cutting benefits took many years. The Public Pension System in Japan 15 Figure 5. Average Benefits of the National Pension and Consumer Price Index Source: Ministry of Health and Welfare (1995). The contribution amount has been revised upward at every occasion of financial reform of the National Pension scheme. This is the same as in the EPI; the public has gotten used to it, and their expectations so far have never been spurned. The latest projection of contributions is shown in figure 6. The amounts of the contribution per month are shown in the figure: those up to 1994 are actual numbers, and those projected after 1994 are indexed to the CPI. Line A traces the actual contribution amount up to 1994, and continues with the projections for succeeding years. The amount of contribution will continue to rise, picking up 500 yen every year after 1994 until it reaches 21,700 yen (per person, per month) in 2015. This annual increase for more then 20 years may seem like quite a rapid rise in terms of the additional burden involved, but line B suggests that this schedule would postpone the impact of pension costs to younger generations. To achieve the same amount of contributions in present-value terms (as in the line A schedule), every insured individual would have to pay 19,500 yen per month from 1994 onward. Thus, as was exactly the case in the EPI, public pension costs are shifted to latecomers. We now want to take a look at lifetime benefit-to-cost ratios of various cohorts in the National Pension scheme. Before examining the numbers, we have to note that there was a special plan for people aged 50 to 54 when the scheme was established. When the pension plan started its operation in 1961, an exception was allowed for this group of self-employed individuals so that they could receive their pension even though the maximum period in which they could contribute to it was 10 years. Thus, this special provision was called the ten-year pension. Later in 1969, another special plan was offered to those who belonged to the above cohorts, but had not joined the ten-year pension. The plan was called the five-year pension, for now a five-year contribution to the National Pension scheme made them eligible to receive benefits. While the original packages themselves were already very thick in benefits, no sooner had either the ten-year or five-year 16 Eiji Tajika pensioners started to collect their benefits that the amount they received more than doubled by the aforementioned 1970s reforms. Figure 6. The Amount of Monthly Contribution to the National Pension Plan Source: Ministry of Health and Welfare (1995). Table 8 shows how high the benefit-to-contribution (B-to-C) ratios were for people in the two special plans. In fact, lifetime benefits for men in the ten-year pension plan were more than 100 times their lifetime contributions. The B-to-C ratios of other pensioners of the two plans are just as incredible, whether they are for men or women. The two pension plans should be regarded more plausibly as massive give-aways to the old from working generations than anything called insurance. It would, however, be important to understand the special sympathy of Japanese for the generation that was involved in the War and contributed to the nation's prosperity through their hard work. The cohorts in the ten-year and five-year pensions represented precisely such "hard times" survivors. While being sympathetic toward this national mood, we should still be able to identify its destructive effect: once thick benefits were granted, they constituted a basis for future benefits. And as we discussed above, the consequences were cuts in benefits and a sharp rise in contribution rates. Table 9 reports the B-to-C ratios for younger generations. The numbers here are taken from the estimates of Ministry of Health and Welfare (1995). The insured are supposed to contribute for the maximum 40 years to the National Pension scheme, and they are to qualify for the full pension at age 65. Receiving the full pension, however, will not make the National Pension a good investment for the younger generations. The Public Pension System in Japan 17 Table 8. Lifetime Benefit-To-Contribution Ratio: The National Pension Plan The ten-year pension: The five-year pension: B/C B/C Year of birth Men Women Men Women 1907 62.81 103.82 39.31 52.00 1908 67.33 105.30 45.18 54.40 1909 73.10 117.28 48.52 59.01 1910 76.24 123.17 51.80 60.52 1911 80.64 124.76 55.09 62.51 Note: 1. The ten-year pension was the National Pension for those who were 50 to 54 years old when the National Pension scheme was established in 1961. The five-year pension was the plan for the same class of the people, but for those who did not participate in the national plan at its outset. 2. B and C are, respectively, lifetime benefits and contributions. Source: Tajika, Kaneko, and Hayashi (1996: 129). Table 9. Lifetime Benefit-To-Contribution Ratio: The National Pension Plan Year of birth B/C (B-T)/C 1924 10.63 7.08 1934 4.93 3.29 1944 2.72 1.81 1954 1.70 1.13 1964 1.28 0.85 1974 1.11 0.74 1984 0.94 0.62 1994 0.94 0.62 Note: B and C are, respectively, lifetime benefits and contributions; T is the subsidy from the government. Source: Ministry of Health and Welfare (1995: 252). The cohorts born in the 1980s will receive less from the Pension than what they will have to pay into. As was already mentioned above, a third of the benefits is a subsidy from the government. If we take this part, T, away from the benefits, the net receipts will be less than the amount of contributions for those who were born in the 1960s or later. Thus, just as it was in the case of the EPI, the consequences of excessive benefits hurt the younger generation to the extent of making them worse off for having stayed in the public pension plan. Issues Related to the Public Pension System As we have seen so far, the public pension system in Japan is such a huge institution that it has been closely related to some of the nation's most important public issues. The purpose of this section is to discuss briefly three issues related to the public pension scheme in Japan: cross-subsidization between 18 Eiji Tajika employees' pension plans and the National Pension plan, the management of the assets of public pension plans, and the treatment of women in the pension system. Cross-Subsidization between Public Pension Plans Supporting businesses and maintaining the living standard of the self-employed has been one of most important policy goals in Japan. And this concern has influenced various public policies: for example, special tax treatments have been applied to small businesses, and 50 percent of the medical costs for self- employed individuals has been covered by the government. The public pension system has been no exception and thus massive financial support has been given to the public pension plan for the self-employed--that is, the National Pension plan. This support has two components. The first is a direct subsidy from the government, and the second a cross-subsidization from employees' pension plans to the National Pension plan. We have already touched on the first type of subsidy--that is, a third of the fundamental pension benefits is subsidized by the government, and this is applied universally to all beneficiaries. The other source of support to the National Pension reflects better the nature of public policies in Japan. After the sweeping reform of 1986 various employees' pension plans started shouldering a part of the costs for paying National Pension benefits. In other words, those people paying contributions under their employees' pension plans began to support the self-employed beneficiaries of the National Pension plan. It is not easy to sort out how much transfer has been made from employees' pension plans to the National Pension plan, but (given this caveat) the results of our estimate are shown in table 10. It is startling to see how much transfer has been made from salaried workers to the self-employed. In 1986 the self-employed contributed only 36.6 percent of the total receipt of their own pension. The rest was paid by the government (19.8 percent), and employees' pension plans (43.6 percent). The 1986 portrait does not change much throughout the years that we have obtained relevant data for. Table 10. Breakdown of Revenues for the National Pension Plan (percent) Transfer from Subsidy from employees' pension Year Contribution the government plans 1986 36.6 19.8 43.6 1987 38.4 22.0 39.6 1988 37.8 27.0 35.2 1989 36.1 27.3 36.6 1990 34.2 25.0 40.8 1991 33.9 25.0 41.2 1992 34.3 25.7 39.9 1993 35.9 27.0 37.2 1994 38.2 24.1 37.7 1995 37.4 24.3 38.3 Source: Author's estimate. The Public Pension System in Japan 19 A further problem with this transfer is that the extent of salaried worker support for the National Pension plan has not been made public. The mechanics by which salaried people contribute to the National Pension is that the total amount of the benefits of the fundamental pension is shared based on the proportion of the contributors, as opposed to the beneficiaries. Since employee pension plans are supported more by young people than by the self-employed (who are covered under the National Pension), salaried workers end up supporting the financing of the pension plan that benefits the self-employed. The most important aspect of the 1986 reform is arguably the institutionalizing of the transfer from employees' pension plans to the (then) de facto bankrupt pension plan for the self-employed. Pension Assets and the Public Investment and Loan Program The public pension has played an important role in mobilizing savings throughout the period following World War II. In Japan some of the savings have been managed by the government and its closely linked organizations. Put simply, postal savings and the funds accumulated by the public pension have been managed by public lending organizations and invested for the construction of the nation's infrastructure.6 The amount accumulated by the public pension is huge. Table 11 shows the proportion of the pension's funds to the GDP. The system has continued to expand its funds: they were about 5 percent of the GDP in 1965; the share has been expanding steadily and stood at about a quarter of the GDP in 1995. Table 11. The Proportion of Funds Accumulated by the Public Pension Plans to the GDP (percent) 1965 1970 1975 1980 1985 1990 1995 4.8 6.8 9.2 12.4 16.4 21.9 24.8 Source: Economic Planning Agency (1998). A problem of managing public pension funds this way is that the funds may not have been invested in the most profitable manner possible. The huge asset collected by the public pension system has been almost fully invested in a single public institution and obtained only regulated interest rates. If the interest rate is set below the market rate, there are enormous opportunity costs associated with investing these funds in the public investment and loan program. With the liberalization of or even the "big ban" on financial institutions in Japan, more serious discussion on the use of pension funds is in order. And this will relate eventually to the issues revolving about the privatization of the public pension system. The Status of Women Women in Japan have been granted special status in various ways. However, if one is to speak of privileged treatment, then the beneficiaries as a group are not women in general, but married women without steady jobs. When a man whose wife does not work pays his individual income tax, he can subtract from his income the deduction for a nonworking spouse. Nonworking spouse status is now 6 Since the middle of the 1980s a part of the funds lent to the public investment and loan program was allowed to be invested by the public pension system itself. However, this is only a portion of the pension plan's total assets. Moreover, direct management of the pension funds has not been permitted; that is, pension funds must first be lent to the public investment and loan program, and a part of it can then be borrowed for investment by the public pension system. 20 Eiji Tajika granted to those who earn about a million yen annually, and spouses whose income is below this threshold are exempted from paying either individual income tax or social insurance contributions. Dependent spouses are entitled to other special benefits as well. Even without paying a social security contribution they have a right to receive the Fundamental Pension benefits when they reach the age of 65. Moreover ,when a working spouse dies, the nonworking spouse can receive survivors' benefits consisting of three-quarters of the remuneration-proportional portion of the deceased spouse's pension. Added up, benefits for nonworking spouses increase one's lifetime income significantly--and this seems to have killed dependent spouses' incentives to work. Table 12 offers a glimpse of this problem. We saw earlier (in the section on the institutional aspects of the Japanese public pension system) that of the three types of insured individuals, type III is exempt from paying contributions. Table 12 shows the number of people who are classified as type III--nonworking spouses. While the concept of nonworking spouses itself applies to men and women equally, these spouses in Japan are predominantly women--and the table shows an upward trend in the number of people in this group. However, we need to be more cautious in interpreting the data. What the table shows us is that married women stop working when their income rises beyond the threshold that gives them special status. This implies that the special treatment of women in Japan has had a strong impact on their choice between labor and leisure. Table 12. Number of Nonworking Spouses (1,000 people) Year 1986 1987 1988 1989 1999 1991 1992 1993 Dependent 30 35 38 40 42 42 43 44 married men Dependent 10,898 11,264 11,577 11,748 11,915 12,008 12,069 12,119 married women Source: Ministry of Health and Welfare (1995: 38). If we have identified the effects of special treatment on women correctly, we have to conclude that this has had an adverse effect on the economy. Participation of more women in the labor market should have reduced wages and made the economy more competitive. This does not apply to manufacturing industries alone, but rather various service industries that need more women. Japanese society is aging, and in this process various services that have hitherto been supplied by families need to be replaced by what the market can provide. These services--such as taking care of the elderly--are extremely labor-intensive. Therefore, there are many promising and expanding markets in Japan that call for the full-time participation of women. If current special treatments keep women from joining these markets, we need to reform our system so that women can contribute to solving the problems the country is facing. Conclusion The theme of this paper is that pension benefits in Japan exceed contributions, under the misguided notion that economic growth would be able to sustain lucrative benefits to the elderly. The economy has lost its steam, its high-growth period is now in the past. When this period's liberal benefits were granted, it was more often than not by raising the rate of contribution to balance the pension account. When benefits were The Public Pension System in Japan 21 cut, it usually took many years to fully implement the revised benefit structure. In effect, the burden of liberal benefits has consistently been shifted to subsequent generations. The aging of the population has aggravated the financial situation. The intrinsic cause of this financial disaster, however, is the excess benefit that provides retirees with 70 percent of the working wage. This paper has focused as much as possible on how the pension's financial administration has stumbled and how younger generations eventually will shoulder the burden created by excessive benefits. The implications of the Japanese experience for constructing a workable pension system for developing countries should be apparent. To sum up: the relevance of pay-as-you-go financing is fading. Probably the most vulnerable aspect of this financial method is that it is susceptible to political pressure to start distributing benefits to retirees at an early stage. Japan fell into this trap in the early 1970s, only a decade after it introduced the National Pension scheme--and now the unfunded system is at the mercy of demographic changes. Thus, if a pension system is to be created from scratch there is a better case to be made for a funded system than an unfunded one. A question arises, then, as to how one is to live with an unfunded national pension system that is already in existence. The second implication of the Japanese experience would be that the first and second tiers of the pension system have to be clearly differentiated. Japan has failed in setting up a universal Fundamental Pension that is equally applicable to both self-employed individuals and salaried employees. Instead, it has gone so far into massive cross-subsidization that it has to keep on life support a financially vulnerable self- employed pension plan. It seems to be the case, therefore, that when an unfunded system already exists and is hard to get rid of, it has to apply universally. A reasonable degree of income distribution by this system would help stabilize the country. In addition, the second tier has to be fully reserved for the funded system. There are many ways to implement this so-called upper deck, but the most important part of it is making sure that it functions truly as an insurance system for individuals--one that is actuarially fair for an individual regardless of that individual's life span. References Economic Planning Agency. 1998. Annual Report on National Accounts. Tokyo. Economic Planning Bureau. 1996. Keizai Hakusho (White Paper, 1996 edition). Tokyo. In Japanese. IFES (Institute of Fiscal and Economic Studies). 1992. Koreika Shakai no Grando Dezain (Designing the Old-Age Society). Tokyo. Ministry of Health and Welfare. 1995. Nenkin to Zaisei (Public Pension and Finance, 1994 edition). Tokyo. In Japanese. National Institute of Population and Social Security Research. 1986. Nihon no Shourai Jinko (Population Projections of Japan). Tokyo. ------. 1992. Jinko No Doukou (Population Projections). Tokyo. Tajika, Eiji, Y. Kaneko, and F. Hayashi. 1996. Nenkin no Keizaigaku: Hoken no Shiten (The Economics of Pension: An Insurance Perspective). Tokyo: Toyo-Keizai Shinposha. In Japanese. Social Insurance Agency. 1997. Jigyo Nenpo (Annual Report). Various issues. Tokyo: Shakai-hoken- kyokai. Mission of World Bank Institute This mission of WBI is to help World Bank clients and staff acquire new development knowledge and skills through a variety of courses, seminars, and other learning events. It designs programs on topics related to economic and social development for governments, nongovernmental organi- zations, and other stakeholders. The Institute produces and disseminates publications and electronic information products that support these objectives. For information on WBI publications write to: Publications WBI The World Bank 1818 H Street, NW Washington, DC 20433 Tel: 202-473-6349 Fax: 202-522-1492 Visit us on the World Wide Web at: http://www.worldbank.org/wbi SN 37203