Taking the Opportunity of the Release of New Population Estimates to Develop a Constructive Debate on Pensions (I)

The new population projections are due to be released December 2006 by the National Institute of Population and Social Security Research (IPSS). They constitute the regular revision of the estimates carried out five years after the previous estimates (2002), and public interest in them is high, as they are used widely for policymaking in Japan, not only for social security policy but also in other spheres such as demand forecasts for public works. Leading newspapers and other media have recently been making a variety of "forecasts of the population forecast" based on sources such as the proceedings of the population subcommittee of the Social Security Council at the Ministry of Health, Labour and Welfare. Those media reports predict that the forthcoming estimates will revise downwardly the figures from the previous estimates.

In past estimates, the IPSS's population projections have repeatedly been revised downwards. On each occasion, it has been necessary to revise the social security and other systems, thus the IPSS population projections have been subject to severe public criticism. It is true to say that, in retrospect, one wonders why the projections could not have been a little more accurate, since they seek to forecast the future population with the maximum possible scientific objectivity, using standard demographic projection methods. Many economists argue that they should incorporate economic factors explicitly, but the IPSS has denied the possibility that its projections could be arbitrary, and has adhered scrupulously to demographic estimation. It is not my task to defend the IPSS projections, but it is undoubtedly worth listening seriously to many of the points the institute is asserting. Paradoxically, downward revisions of estimates in the past could be interpreted as demonstrating the dynamic fluctuation of people's reproductive behavior.

Nevertheless, constant downward revisions to projections certainly cast doubt not only on the reliability of the projections themselves, but also on the reliability of policies, and therefore there are many aspects of the IPSS projections in which it is hoped that improvements will be made. Judging from what is said in the media and from the proceedings of the population subcommittee, it can be inferred that the new population projections will respond to some of the oft-made criticisms. In my view, however, rather than present the population projections as scenario-specific in the form of high, medium, and low variant projections, it would be more convincing if they were presented in a way that explicitly sets out ranges of probability centered on the medium variants. It looks, however, like the new projections will contain scenario-specific projections as in the past.

In any event, the new projections will undoubtedly take into consideration the impact of factors such as the avoidance of marriage, late marriage, and late childbearing, and be revised below the previous projections, though the specific nature of the projections remains unclear. However, if we compare the previous projections with the actual values to date, in many cases the actual values are around midway between the medium and low variants in the previous projections. For example, the total fertility rate (TFR) in 2005 was previously postulated at 1.31076 in the medium variant projection and at 1.22074 in the low variant projection, while the actual value is put provisionally at 1.25. In view of this, many observers believe that the medium variants in the impending new projections will probably lie midway between the medium and low variants in the previous ones.

Social security, particularly the public pension system, is the most likely area to be impacted by the downward revision of the population projections. Given that the sustainability of the pension system is heavily reliant on fluctuations in the ratio of vested pensioners to pension subscribers, a downward revision of the population projections would clearly be a negative factor for the pension system. The 2004 revision of the pension system included the imposition of a ceiling on the burden of premiums and the maintenance of the income replacement ratio of the model household at 50%, but naturally that was premised on the medium variant projection in the previous (2002) population projections, and the outcome will differ if the population assumptions differ. Of particular note is that if the ceiling on contribution rates is maintained, the maintenance of a 50% replacement ratio for the model household will be very difficult. In order to give more specific consideration to this aspect, I will also make a forecast of the population forecast, and calculate a provisional population projection based on the scenarios set out below.

Population forecast Case A
The birthrate by age group after 2006 is assumed to be midway between the medium and low variants in the 2002 population projections. The TFR in 2055 is 1.24532, and from 2055 to 2105 is projected to return to 2.07, which is the population replacement level.

Population forecast Case B
This case is a little more optimistic than Case A, projecting a TFR of 1.30 in 2055. From 2006 to 2049, the medium and low variants in the 2004 projections are distributed proportionately as ratios to the 2055 TFR of 1.30. After 2055, as in Case A, this case postulates that the TFR will return to the population replacement level of 2.07.

Click to enlarge

It should be noted that this calculates the population somewhat roughly, using the IPSS 2002 projections, and therefore does not take into consideration factors such as the revision of the future life table (changes in average life expectancy).

Based on these hypothetical population projections (Case A and Case B), let us now examine the outlook for pension finance by using the simulation model for pension finance (the RIETI model) developed by our research group. The table below gives projections for the balance of kosei-nenkin (Employees' Pension Insurance [EPI]) reserves; the following are the reasons for showing the trend of this reserve balance. In the 2004 revision, the financial balance period for future pension finance was stipulated to be 100 years, with the requirement that in the final year of that period one year's worth of benefit for the relevant fiscal year be secured (finite balance system). Therefore, the level of the reserve balance accumulated during the financial balance period is an important indicator of the sustainability of pension finance. Here, however, in order to make a clear comparison with the reform of the pension system in 2004, various non-population-related assumptions are made to match the standard case scenario in the Ministry of Health, Labour and Welfare's 2004 Actuarial Valuation of the Employees' Pension Insurance and the National Pension (2.1% rate of increase in nominal wages, 3.2% nominal investment yield, 1.0% rate of price increases), and the calculation period is until 2100 (2105 would be acceptable as the final year, given that the next pension reform will be in 2009, but for comparison purposes the final year has been set as 2100).

Click to enlarge

The 2004 reform introduced a so-called macroeconomic sliding-scale system, a fiscal adjustment method for ensuring that pension finance is maintained. Under this system, until the projection of ensuring fiscal sustainability during the balance period is achieved (until the forecast that the reserve balance after approximately 100 years will ensure one year's worth of benefits is met), the sum of the rate of decline in the number of beneficiaries and an aging rate of 0.3% will be temporarily deducted from the rates applicable to the sliding benefit scale (newly instituted: growth rate of average wages; already instituted: commodity price index). This is an important fiscal adjustment mechanism insofar as it automatically keeps the increase in aggregate pension benefits temporarily within the bounds of GDP growth, and indeed was the most important reform measure within the 2004 reform. The 2004 reform, which was premised on the 2002 population projections, provided that this macroeconomic sliding-scale system be applied until 2023.

Among the premises for the 2004 Actuarial Valuation, if the population assumptions are only replaced in the aforementioned Case A and Case B, namely, the macroeconomic sliding-scale system is applied until 2023 and only population assumptions are replaced in Case A and Case B, then in these cases the amount of the reserve would bottom out in 2080 and 2090 respectively, and fiscal sustainability could no longer be maintained (the broken green and purple lines in the chart above). If this fiscal deterioration is addressed by extending the macroeconomic sliding-scale system, then by extending it to 2028 as in the Case A population and to 2026 as in the Case B population, it is estimated that more than one year's worth of benefits can be secured from the reserve fund (the unbroken green and purple lines in the chart above).

However, the extension of the macroeconomic sliding-scale system would have the effect of lowering the level of benefits. If the system were applied until 2023 premised on the 2002 population projections, then the replacement ratio of the model household at the time benefits are first received (age 65) from 2023 onwards would be 50.2%, but if the system were applied until 2026 premised on the Case B population, the replacement ratio of the model household from 2026 onwards would be 48.06%, and if it were applied until 2028 premised on the Case A population, the replacement ratio of the model household from 2028 onwards would be 47.18%.

As can be seen from the above, if the population projections are revised downwards, the pension finance forecast also changes. In particular, in order to maintain pension finances under the present system and keep the ceiling on the burden of pension costs restrained as laid down in the 2004 reform, it is highly likely that it would be necessary to extend the period during which the macroeconomic sliding-scale system is applied and to keep the level of benefits down. If that were to occur, however, it would be necessary to reexamine the 50% replacement ratio (model household).

What was the original ground for guaranteeing a 50% level of benefits in the replacement ratio? There probably was debate about the definition of relative poverty (disposable income of less than 50% of the median for the entire population), but even if that level could be maintained by means of public pensions, that does not guarantee that confidence in pensions will be maintained or increased. This can be inferred from the fact that even today quite a few people are dissatisfied with present pension benefits, even though the replacement ratio is much higher than what it will be in the future. If there are distortions in benefits or the principles of the system are unclear, public confidence in the system will not increase, even if the replacement ratio is high.

High benefit levels are of course better than low levels, but in view of the fact that there will be fewer people to support the system as the population declines, it will be pointless to discuss the matter if it is premised on there being burden levels that households and firms will be unable to bear and that are fiscally unsustainable. It is far more important to discuss the kind of benefit system that would be capable, based on limited financial resources, of giving people the most peace of mind and confidence.

So far, I have considered only the extension of the macroeconomic sliding-scale system, but there are a number of other important issues, for example, making another increase in the initial age to receive benefits (67, 70), revising survivors' benefits, and earmarking consumption tax for social securities. With respect to raising the age at which benefits become payable, Germany has led the way by already deciding on the change, and given that Japanese have the world's longest lifespan, discussion should be given to the appropriateness of 65 as the initial age at to receive benefits. In regard to survivors' benefits, whereas under the present system the surviving spouse of a subscriber to EPI can receive 75% of her husband's earnings-related pension, the level in other countries appears to be around 50%, though the disparate schemes in operation there make a straightforward comparison difficult. Broad debate is essential as to whether this level will continue to be appropriate for Japan.

However, if the age at which benefits become payable were raised, and survivors' benefits revised, or the low rate of payment of the kokumin-nenkin National Pension (NP) premium continues, and the rate of public assistance were forecast to rise sharply to above a level the public considers tolerable, then it would instead be desirable to shift the debate in the direction of strengthening the income redistribution function of pensions.

Nevertheless, it will be very difficult to have a detailed discussion on these matters in Japan, because Japan is conspicuously slow as regards compiling statistical data on middle-aged and elderly people, particularly reliable panel data. A process perfectly normal in Europe or the United States, such as using data to examine in advance the potential impact of a particular change in the pension system on other economic variables, is hopelessly difficult in Japan.

In view of this, RIETI has organized a project to bring together Japan's foremost experts in the spheres of economics, public health, and health care, including SHIMIZUTANI Satoshi (Associate Professor, Institute of Economic Research, Hitotsubashi University), ICHIMURA Hidehiko (Professor of Economics, Graduate School of Public Policy, The University of Tokyo), NOGUCHI Haruko (Associate Professor, Faculty of Social Science, Toyo Eiwa University), and HASHIMOTO Hideki (Visiting Professor, Graduate School of Medicine, The University of Tokyo) and is conducting a large-scale panel survey on middle-aged and elderly people. This research is on a par with the panel surveys on the same age groups already conducted and established as global standards, namely, the U.S. HRS (Health and Retirement Study), ELSA (English Longitudinal Study of Ageing) in the UK, and SHARE (Survey of Health, Ageing and Retirement in Europe) in Continental Europe. The collated data will be suitable for mutual comparison.

The survey forms have been designed with great attention to detail, so that valuable information can be collated on the basis even of only single-year data, while the cumulative data from multiple years will create highly informative panel data. I am convinced that, when the research data have been accumulated, the discussion surrounding Japan's social security policy will grow more detailed and constructive.

A number of issues remain to be discussed with the objective of enhancing public pensions in the future. If these issues can be discussed in greater depth, it may still be possible to fully maintain and improve people's welfare in spite of limited resources. However, a cause for concern is that, as happened at the time of the 2004 reform, pension-system reform will be handled in a highly politicized way that will totally stifle substantive debate. What must be avoided in particular is any single indicator (even if it may be an important yardstick, such as the 50% replacement ratio) being focused on exclusively, leading nowhere except toward a constantly unfolding offensive and defensive battle. It must be made absolutely clear that any debate conducted in circumstances such as those would not be in the public interest.

The release of the new population projections will give rise to resurgent public interest in pensions, and consequently in social security policy as a whole. I hope that will trigger not meaningless witch-hunts, but the spread of more constructive debate.

Taking the Opportunity of the Release of New Population Projections to Develop a Constructive Debate on Pensions (II)- Forecast for pension finance after the release of "Population Projections for Japan (December2006 Estimate)"