Japan's Pension System -Evaluating the 2004 Reform and Establishing Clear Principles for Further Reforms-

The session began with a presentation titled "The Basic Questions on the 2004 Pension Reforms in Japan" by KOSHIRO Kazutoshi (Emeritus Professor, Yokohama National University).

The 2004 reform was revolutionary in responding to an era of falling population and ranks alongside reforms made in other countries. The initial reaction in Japan was lukewarm, but praise from overseas has prompted an ongoing reevaluation of the reform within Japan.

How did the 2004 reform make the pension system financially sustainable? The answer lies in the following four points:
(1) Increasing the contribution rate ceiling to 18.3% by 2017.
(2) Raising the national government subsidy for the basic pension to one-half by 2009.
(3) Introducing a macroeconomic slide mechanism (by FY2023) to limit future pension benefit increases to 0.2%, an amount calculated by taking the annual rate of increase in real wages (1.1%) and subtracting amounts equivalent to the decrease in the labor force (0.6%) and the increase in average life expectancy (0.3%). The macroeconomic slide will also be applied to price indexing of existing pensions, allowing a drop in actual pension benefits if prices decline.
(4) Adopting a finite horizon balancing mechanism and drawing down the pension reserve fund to one year's reserve by 2100. Note that if the total fertility rate is not restored to 1.39 or if the economy underperforms against projections, the balance will be reviewed at five-year intervals. The macroeconomic slide mechanism significantly mitigates intergenerational inequity. A proposed alternative solution suggested freezing the contribution rate for the Employees' Pension System at 13.58% and keeping benefits within the existing one-third government subsidy, in which case the income replacement rate would be reduced to 35%. Which alternative is better as a policy option is a value judgment, but I believe the method of reform selected was the appropriate choice.

Under a pay-as-you-go system, conceptually a certain amount of pension liability arises even with a stable population. Economists who are critical of maintaining the pay-as-you-go system in the face of shrinking populations in the future generally gave positive marks to the 2004 reform, but they criticized it because net pension liabilities will amount to 570 trillion yen. Yet it doesn't affect the fact that pensions can be maintained under the existing system.

Since the macroeconomic slide mechanism will also be applied to the basic pension, there are concerns that the basic pension could fall below subsistence level, although there is no danger of this provided the full pension is paid. However, in cases where the contribution period is less than the specified number of years because of failure to make contributions, failure to join the pension scheme or exemption from contributions, it is possible that the basic pension could fall below the subsistence level and depending on economic conditions it could also fall below the level of public welfare assistance. This situation will not come about in the near future, but if such a problem arises it should basically be addressed via welfare programs.

Under the 2004 reform, the pension benefit/cost ratio for younger generation model households (taking into account total joint contributions by employees and employers) stays above 1 (using the nominal wage increase rate as the discount rate). In fact, there is debate over what should be used as the discount rate. It is one-sided to consider intergenerational equity simply in terms of the generational gap in the benefit/cost ratio of pensions (i.e. whether an individual gets back the money paid into the system). In the early days of the system contributions were low and benefits were high, but as well as paying contributions the early participants supported their own parents. It is also necessary to consider broader private and social investment (social capital formation, investment in child-raising and education, investment in housing, etc.). While there is a gap in the benefit/cost ratio, it is debatable whether this amounts to inequity.

Looking at the impact on the labor market, firstly the problem of category 3 insured persons still remains. There are 11.24 million category 3 insured persons who benefit from the existing system (twice that number if we include their spouses), making up one third of those eligible for pensions. Since this group has considerable political clout, no conclusion has been reached on the issue. Opinion polls are divided on the merits of the potential solutions, namely (1) leaving the current system unchanged, (2) collecting contributions from category 3 insured persons, and (3) reducing the basic pension portion of their pension benefits. From the theoretical perspective, the "ability to pay" principle has been upheld since the basic pension system was introduced in 1985, but in recent years some economists have argued that if we consider the "imputed income" of domestic labor category 3 insured persons should also pay basic pension contributions.

Secondly, in the course of deliberations on the 2004 reform, it was proposed that pension eligibility should be extended to part time workers who work more than 20 hours a week or earn over 650,000 yen in a year (this would affect approximately 4 million people). This was not implemented due to opposition from industry and business groups, but pension eligibility should at least be extended to part timers who work 20 hours a week or more.

Thirdly, the pension reduction system for aged workers was regarded by many as a barrier to workforce participation by the elderly, but in 2005 the uniform 20% cut was abolished for those in the 60-64 age group, which went some way toward lowering barriers to employment. The reduction of pension benefits (from April 2002) for workers aged 65-69 who earn more than a specified amount is contrary to the stated aims of the social insurance system, but there are many executives and highly-skilled experts aged over 65 who earn considerable incomes, thus this step was deemed unavoidable in order to uphold a system of intergenerational support. Similar benefit restrictions will apply to high-income workers over 70 from April 2007 (only the second-tier pension will be reduced - the full amount of basic pension will be paid).

Since a cabinet decision in 1984, a series of steps has been taken toward consolidation of public pension schemes, and the ruling coalition is discussing consolidation of the Employees' Pension System and mutual aid pensions by 2007. This will be difficult to reconcile with the aims of the opposition Democratic Party, which proposes the merging of guaranteed pensions and earnings-related pensions including the National Pension System, but the merits of that proposal are difficult to comprehend if the feasibility of implementation is considered. The raising of the government subsidy for the basic pension to one-half must be achieved by fiscal 2009. Due to successive amendments, the formulae under the existing system have become extremely complicated, thus there is an urgent need to introduce a points system. While this will entail a large volume of administrative work, such a system is desirable.

At the root of debate over Japan's public pensions lies the question of why Japan has not adopted the Swedish Notional Defined Contribution (NDC) system. While the disparity in the two countries' birth rates is a factor, the fundamental reasons that Japan cannot use NDC are the differences in pension benefit design: (1) In Sweden NDC is applied only to old-age pensions (together with the guaranteed minimum pension, survivors' and disability pensions are paid from the general budget), and (2) the Swedish system does not have a flat-rate benefit. Due to these differences the "turnover duration" (in essence the gap between the average age of pension recipients and contributors), an essential actuarial component for setting up an NDC system, cannot be defined in Japan. Considering this point, the 2004 reform introduced the macroeconomic slide as an automatic balancing mechanism instead of the NDC system.

After the Second World War, employees of major corporations and public servants consistently demanded higher lump-sum retirement allowances and corporate pensions to guarantee their livelihoods in old age. In addition, with the slump in share prices following the bursting of the economic bubble at the end of the 1980s and lower interest rates on government bonds, lump-sum retirement allowances and corporate pensions have come to represent 12% of total labor costs for major companies. Meanwhile, employers' contributions to public pension schemes amount to just 5%. One could go so far as to say that employees of major corporations and public servants receive "double payment" in the form of lump-sum retirement allowances/corporate pensions and public pensions, when the former alone would be sufficient to ensure their livelihoods in old age. On the other hand, only one-third of all workers can enjoy the benefit of such double payment, while employees of small and medium-sized companies receive minimal retirement allowances. I can appreciate that major companies are faced with fierce cost competition amid the trend of increasing globalization, but I believe a higher level of contributions to the Employees' Pension System should be tolerated, even if this means reducing retirement allowances.

In response to the presentation, Olivia S. MITCHELL (Professor of Insurance & Risk Management, The Wharton School of the University of Pennsylvania) offered the following comments.

The main concern regarding pensions is the shortfall in cashflow. Increases in contributions are not keeping pace with the growth in total benefits paid, which results from increasing numbers of recipients and longer life expectancy.

It is interesting to note that opinions vary considerably in Japan on the extent of the pension problem.

More sensitivity analysis is required. There is also a need to conduct more behavioral analysis to see how people's behavior would change if contributions or taxes were increased or benefits were reduced.

More detailed analysis is needed on what happens if economic performance diverges from projections, especially in respect of investment yields. Also required are further consideration of the discount rate, the introduction of an infinite time horizon, and stochastic simulations.

Gathering microdata is essential to appropriate pension system design. The full picture cannot be extrapolated from model households and it is necessary to ascertain specifically where the rest of the government subsidy portion is coming from.

Long term government liabilities are generally understated. For example, if social security and Medicare payments in excess of dedicated revenue streams are taken into account, long-term U.S. government liabilities amount to $63 trillion. However, opinions are divided on whether governments should report these kinds of liabilities. Personally, I think governments' long term liabilities should be made clear to an infinite horizon. Japan has started to reduce long term government liability by using the macroeconomic slide mechanism, and in this sense its reforms are ahead of the U.S.

No matter how big a reserve is amassed, there is no guarantee that returns will live up to expectations. Sometimes returns can be significantly negative. Opinions also vary widely on the appropriate schedule for drawing down reserves and there is much debate over investment methods, for example whether some funds should be invested overseas.

Next, Ole SETTERGREN (Director of the Department of Pensions, Swedish Social Insurance Agency) made the following comments.

The very demanding demographic trends made the 2004 reform necessary, but I am somewhat skeptical about whether the results will be as pessimistic as the projections suggest. In a sense, Sweden may be overly optimistic.

To make the pension system sustainable, earlier entry into the labor force is necessary and greater labor force participation by the elderly and women would be helpful. It is necessary either to reduce the number of years spent in education or raise the retirement age or age of pension eligibility.

The automatic balance mechanism (macroeconomic slide) does not work when inflation is low, which may raise the fiscal risk. Also, the projected increase in average life expectancy should not be built into the automatic balancing mechanism. If the purchasing power of pensions is reduced by an average of 0.9% per annum via the macroeconomic slide, many retirees' incomes will fall below subsistence level. To make the automatic balancing function work, objective and easily understood fiscal information needs to be made available every year.

The Japanese pension system is too complex, and should be simplified.

John PIGGOTT (Professor of Economics, University of New South Wales) made the following comments.

The existing system treats single workers unfairly and hinders labor market participation by category 3 insured persons. To ensure the sustainability of the system, labor force participation needs to be increased and it would be better if part time workers also had to make pension contributions.

The relationship between public pensions and social security needs to be examined carefully.

Prof. Koshiro made the following responses to the comments above.

To Prof. Mitchell:

Concerning the point that sensitivity analysis is required, the results of such analysis are included in the actuarial report. In the most pessimistic case, an income replacement rate of 45% was forecast.

Concerning the relationship between real economic growth and real wage increases, even in an economy where the population, the labor force and the number of contributors to pension schemes are all falling, it is possible to maintain positive economic growth by raising total factor productivity (TFP) and boosting capital spending. Many commentators are pessimistic about whether a society with a shrinking population can maintain positive economic growth, but the Pension Council and the Pension Bureau have conducted their examinations based on the economic assumptions set out in Chapter 4, Section 4 of the fiscal recalculation results.

It is certainly true that even with reserves, there is no guarantee that returns will live up to expectations. Appropriate management of pension funds is required. Ultimately interest rates depend on how high TFP can be raised.

To Mr. Settergren:

On the assumption that government policies will significantly increase labor force participation by women and the elderly, labor force projections for these groups are higher than past trends. I fully agree that labor force participation by women and the elderly should be encouraged, and the reform law was also formulated on the basis of that judgment. The M-shaped graph depicting labor force participation by women in Japan over the last 20 years has become quite similar to the inverted U-shape that describes the situation in the U.S. and Sweden.

I agree that raising the age of pension eligibility to 67 is a commonsense argument, but given that the prospect of having to continue working until 65 has only just been introduced in Japan it is too soon for such a step. This also relates to the problem of young people not in employment, education or training (NEET).

It is correct that actual figures should be used for average life expectancy.

Concerning the criticism that the basic pension may fall below the poverty line if the macroeconomic slide is applied to it, the key question is what should be done to protect the livelihoods of those who cannot receive the full amount of the basic pension. This will have to be carefully judged in a comprehensive manner considering the relationship between all relevant systems including pensions, social security, health insurance and nursing care. If the fertility rate continues to fall and the economy deteriorates, particular care will need to be paid to reexamining the treatment of those on low pensions within the overall review of the pension system.

Supplementary point:

Concerning the setting of the basic rate, Mr. Settergren and Prof. Mitchell seem to be saying opposite things, so I would be very pleased to hear them discuss this point.

Q: Under the finite horizon balance mechanism, the target fiscal year for the balance will change at each fiscal recalculation. What are your views on this point?

A (Prof. Koshiro): Under the finite horizon balance mechanism, the target year for the balance does change at each fiscal recalculation, but to sustain the system fiscal review needs to be carried out properly. To maintain the income replacement rate, it is necessary to draw down the reserve via the finite horizon balancing mechanism.

Q (For Prof. Mitchell): Why did the U.S. switch from a finite horizon for the balancing mechanism to an infinite horizon in 1975?

A: The advantage of an infinite horizon is that the impact of pension reform proposals can be better identified. Since economic performance often diverges from projections, pension systems need to be considered over a longer time period. An infinite horizon can avoid the bias that occurs when the period for balancing is fixed.

Q (For Mr. Settergren): I understand Sweden has taken the opposite step, switching from an infinite to a finite horizon. Is that correct?

A. Sweden is currently using a 75-year time horizon. Sweden's NDC is balanced each year with the calculation of the annual balance sheet. In this sense, you could say that Sweden has a zero projection system. The Swedish method is highly transparent and easy to understand. If Japan feels it will not work, it perhaps should not be introduced.

Q: How should pensions in existence before the universal pension (public service pensions) be handled? Should benefits be cut or should they be dealt with via contributions?

A (Prof. Koshiro): Public service pensions should have been dealt with via contributions when they were consolidated into mutual aid pensions.

Q: What are your views on the idea of simply using the national health insurance mechanism and making the National Pension System earnings-related?

A (Prof. Koshiro): That is an interesting thought, but the problem is that national health insurance has few levels of monthly earnings standards and the upper limits differ.