Printer Friendly

A universal fully funded pension scheme.


Economists around the world are by now quite familiar with the pitfalls of a pay-as-you-go pension system. An aging society coupled with the prospect of increasing longevity is causing financial strains in public pension schemes from Europe to America (see table 1). In his 1993 Nobel lecture, Professor Robert Fogel presented evidence that the mortality and disability rates for the elderly had fallen for longer than expected, with the result that the Census Bureau likely has underestimated the projected U.S. elderly population in 2050 by about 36 million.

Calls for pension reform are heard everywhere. In the industrial world, quite a few analysts are talking about privatization of public pensions (Roberts, 1995; Dornbusch, 1995). A favorite model of pension reform is privatization Chile style. Actually, however, the Chilean system is more properly represented as based on "an intermediate form of funding." The Chilean government guarantees a minimum benefit payable irrespective of the performance of the funds invested (Mitchell, 1993, pp. 27-28). Another well-known system, the Central Provident Fund, which is in operation in Malaysia and Singapore, is privately and fully funded but publicly administered (see Asher, 1994). The recent report of President Clinton's Advisory Council on Social Security considered the Chilean option (dubbed "personal security" in contrast to "social security") (Business Week, January 20, 1997, pp. 26-27). Yet, the superiority of this model is subject to question. Note, for example, MIT economist Peter Diamond's (NBER Working Paper no. 4510, 1993) conclusion:

We have come to think of privatization as a route to greater efficiency and lower costs. Thus, perhaps the most surprising aspect of the Chilean reform is the high cost of running a privatized social security system, higher than the "inefficient" system that it replaced.

James and Palacios, (1995) point out that the administration costs of public pension plans are likely to be under-reported. However, under-reporting is unlikely to account for the huge gap between the administration costs of public and private plans (Mitchell and Zeldes, 1996, p. 11). The U.S. Social Security Administration reported administration costs at less than 1% of annual benefits while those of the life insurance industry are known to range from 12% to 14% of annual benefits (Diamond, 1993, p. 7). Scale economies of management, and economies of supervision, as well as economies arising from eliminating much of marketing costs, are likely to provide important cost savings attributable to the central administration of public pensions. (Central administration does not necessarily mean government management of the pension funds. It does mean, however, that there may be a monopsony (sole buyer) for fund management services. A board of trustees charged with administration of a public pension may "farm out" funds for management by private pension funds. Under such a setting, the private pension funds would be in a weaker position to pass along marketing costs to consumers.)

Public Pension Expenditure as a Percentage of GDP in OECD Countries

Country 1960 1975 1980 1985

Austria 9.6 12.5 13.5 14.5
Canada 2.8 3.7 4.4 5.4
France 6.0 10.1 11.5 12.7
Germany 9.7 12.6 12.1 11.8
Italy 5.5 10.4 12.0 15.6
Japan 1.3 2.6 4.4 5.3
Sweden 4.4 7.7 10.9 11.2
U.K. 4.0 6.0 6.3 6.7
U.S.A 4.1 6.7 6.9 7.2

Source: Table 1 in Mitchell (1993).

Equally important, a "private social security system" is a contradiction in terms. Under such a system, each worker has his own account. Each worker and his employer are to contribute a percentage of the monthly salary into this account, the funds accumulated are managed privately, and the total amount inclusive of investment returns is repayable upon retirement. The only aspect that is "social" about a purely private system is that it is mandatory. (The Chile model is not a truly private system because the government guarantees a minimum pension and mandates a minimum annuity requirement.) A true Mandatory Private Provident Fund (MPPF) scheme allows individual account holders to opt for lump-sum payment or to buy an annuity in the market place from voluntary underwriters. Such a fund is being recommended for Hong Kong (Hewitt Associates LLC and GML Consulting Ltd., 1995, para. 33) - Yet it neither pools longevity risks (Eckstein et al., 1985), or deals with the high cost of voluntarily purchased annuity plans (Abel, 1986). Also, it is doubtful that workers stand to gain from employers' contributions. In a highly competitive world, the market dictates how much can be spent in compensation for labour. Workers' pay and/or other benefits are likely to fall, thus offsetting employers' contributions. Such pension arrangements will not expand the real opportunity set open to workers.

On the contrary, employees see their opportunity set reduced because they are required to contribute the stipulated percentage of their salaries towards their retirement plans. If the capital market were perfect, individuals could trade future payouts for current income. But if this were the case, the very concept of mandatory provident fund would fall apart, as no individual could be forced to save more than he would like. From this perspective, if the plan has any merit at all, any merit must be founded on the assumption that citizens do not know what is best for them (the assumption of myopia) or the assumption of moral hazard, namely that some citizens over-consume before their retirement, intending to take a free ride on public money to bail them out after they have retired. However, even if the myopia and the moral hazard assumptions are valid, the MPPF is not necessarily the best option.

Abel (1986) demonstrates the welfare enhancing property of "Fully Funded Social Security." With compulsory participation, fully funded social security overcomes the problem of adverse selection. As a result, the rate of return is higher than that on private annuities. With a higher, actuarially fair rate of return based on population average mortality, and assuming exogenously given factor prices, both the steady state consumption of young consumers and the steady state level of bequests can be shown to be higher than would be the case under private, voluntary annuities.

Unfortunately, Abel does not address the problem of actual design for such a "Fully Funded Social Security." The fact is that practically all the public pension plans known today are either pay-as-you-go or partially funded. This paper explains the design of a "fully funded social security" system. The design is consistent with the spirit of the "generational accounting" framework as advocated by Auerbach et al. (1994).


Given the demographic trends of the world's population, economists generally agree that a pay-as-you-go type of pension system ultimately will require a rise in contribution rates, a cut in benefits, a postponement of the retirement age, or a combination of these. To avoid the predicaments of uncertain benefits or uncertain contributions, a fully funded system, like the "Central Provident Fund" model, of Singapore or Malaysia (Asher, 1994), or a "Mandatory Private Provident Fund" model, seems necessary.

The spirit of a fully funded system is that workers should save for their own retirement needs. Post-retirement payouts should be "fully funded" by accumulated contributions rather than from the general tax revenue. According to Auerbach et al. (1994) and the Office of Management of the Budget, "even assuming a discount rate of just 3% and a productivity growth rate of 1.25%, future generations will still have to pay 65% more (taxes) than current newborns" (Auerbach et al., 1994, p. 82). This presents a case of gross generational imbalance in U.S. fiscal policy, a problem underscored by generational accounting. Robert Haveman (1994) raises some questions about the generational accounting framework but concedes that it should serve "as a useful supplement to the annual budget" (p. 110). The OMB presented for the first time in history a tabulation of the lifetime tax rates of current and future generations in the FY 1994 budget.

In general, only a fully funded system is compatible with avoiding arbitrary inter-generation transfers. However, a fully funded system does not require that each individual fully fund his own retirement needs. It is conceivable and desirable that each generation or cohort, rather than individuals, fund its own retirement needs.

The Universal Fully Funded Pension (UFFP) system proposed here would have members of each age cohort contribute to and draw from the same pension fund. Rather than relying on the working population to support the retirees, each cohort finances its own retirement. The basic version of the scheme - with refinements to be introduced later - requires each member to contribute the same amount in each month during the contributing years and to draw the same amount each month during the payout years. The budget constraint is one simple equation that says that funds accumulated prior to the stipulated age for drawing benefits must equal the present value of post-retirement payout at the stipulated retirement age:

(1) [summation of] C[(1 + r).sup.t] where t=0 to N-1 = [summation of] B/[(1 + r).sup.j] where j=1 to n

Here N is the number of contribution instalments, C is the monthly contribution, r is the effective average real rate of return assumed and also is used as the discount rate for the n instalments of payouts at B per month, and life expectancy is equal to the current age plus (N + n) /12.

Because there is little uncertainty over the life expectancy for a specified age cohort, one can readily calculate the size of the required contribution to support a specified payout from the stipulated age through the end of the expected life for the entire cohort, given the long-term real rate of return. If the expected life has increased at the stipulated payout age, say because of a breakthrough in medical knowledge, the budget constraint requires either postponing the payout age or reducing the payout. Each cohort is responsible for itself and thus will have to make the choice. In principle, however, the payout should be large enough to cover a basic standard of living agreeable to the community. Of course, individuals have the choice to top it up as they see fit from private savings.

In general, the standard contribution may be borne by the employee or shared between the employer and the employee. Over the long run, if the market functions efficiently, whether the employee bears the standard contribution or shares it with the employers will make no difference. Self-employed persons, housewives, and the unemployed must be responsible for their own contributions unless they satisfy certain criteria for public subsidy, in which case the contribution is either partly or wholly paid by the government. The scheme also allows the government to subsidize an employee's contributions on grounds of lack of means. This arrangement has the advantage of making otherwise implicit and unclear redistributive initiatives transparent. With a standard contribution and a standard payout, and providing a means-tested subsidy for contributions, the design will reduce the regressivity of existing pay-as-you-go plans arising from the longer lives of the high-income people and their steeper earnings profiles (James, 1995, p. 6). It should be noted, however, that redistribution is not a central feature or an integral part of the scheme.

The UFFP needs to be phased in slowly. Clearly, cohorts that have now reached, say, the age of 50, have only 15 years to pay before reaching 65 (the assumed payout age). Accumulated funds by 65 would be small. The budget constraint would dictate that, unless contributions are larger, this cohort draw a smaller pension than cohorts that are younger when the scheme starts operation. In principle, one can assume that the older people have accumulated more savings than the younger ones at the time the scheme takes effect. In addition, during the phase-in, the government will have to provide assistance to the aged who are needy.


The UFFP is a universal, mandatory plan. It is therefore free from the problem of adverse selection that would characterize a true MPPF with voluntary annuity plans. The UFFP covers everybody, including employees, the self-employed, employers, housewives, and the unemployed. Everyone within the age group has to contribute unless a means test shows one eligible for financial assistance.

The UFFP can easily accommodate any degree of redistribution desired, explicitly and without compromising the fully funded nature of the scheme. Any subsidy to contribution made to the poor is an explicit transfer that funds future payouts.

An interesting question is whether government contribution on behalf of the poor would invite a free-rider problem, as people might attempt to declare an income below the threshold contribution level. Clearly, a system of policing to ensure that the subsidized do meet the requisite criteria, combined with necessary penalties to ward off fraud, will have to be in place. But any problem associated with policing is not unique to UFFP and is part and parcel of any redistributive program.

Moral hazard actually is likely to be less serious under UFFP than under a MPPF that exempts the poor young from payment and makes transfers to the poor old. If transfers are made to the poor old through a means-testing mechanism, the availability of the safety net for those without adequate savings will lure people to stay poor and avoid making contributions. In contrast, under the UFFP, young people with a temporary decline in incomes are subject to a means test that looks at their net worth. They will need to make contributions if they have sufficient savings.

Under the proposed UFFP, all individuals are entitled to the pension for their cohort. Recent immigrants will collect benefits in proportion to the number of years of their contribution - regardless of whether their contributions have been subsidized. A recent immigrant starting to contribute, say, at the age of 50, will be entitled to (65-50)/(65-20) of the standard stipend when he reaches the age of retirement, which is assumed to be 65.

The stipulated contribution is a flat amount, which makes it simple to administer. This contrasts with most public pension schemes in practice. Most public pension plans provide for benefits that rise with contributions and incomes. Such plans may have distributionally undesirable consequences because they accentuate the regressivity of many public pension plans earlier noted. In any case, the direction of the implicit redistribution - progressive or regressive - cannot be easily discerned. In the case of the mandatory private provident fund, contributions are income-related and are akin to a payroll tax. To the extent that low income people have to struggle to survive, a MPPF that does not provide for public subsidization of contributions and does not exempt the poor from contributions may push the poor people to earlier death. If mandatory "private" provident fund schemes are modified to provide for public subsidization of contributions, the schemes would lose their "private" character and resemble the UFFP.

The benefit under UFFP also is a flat amount, with the full amount payable on reaching the stipulated age but discounted if early withdrawal of benefits is deemed necessary because of health reasons. Clearly, its magnitude depends on the size of the contributions and the rate of return to investment, as illustrated in equation (1). In principle, the amount should pass a test of adequacy but should be a basic amount. Individuals would be free to top it up with private savings as they please.

Unlike the laissezs faire regime or the pure MPPF, the UFFP pools longevity risks and minimizes cost. Upon reaching the stipulated age, members of the UFFP draw a monthly stipend as long as they live. The MPPF can, of course, be amended to require the non-discriminatory treatment of buyers of annuities. This will make it look like a UFFP.

As a fully-funded scheme, the UFFP, like the MPPF, is free from the risks associated with changes in the dependency ratio and from the political risks of participants trying to extract a larger payout, which may affect a Pay-as-you-go Pension Plan especially if the government is made an automatic contributor along with the employer and the employee. This scheme is the "Three-pronged Contribution Approach" recommended by the Hong Kong Social Security Association. Any plan with the ultimate bearer of the financial burden hidden away is subject to the risk caused by the "free lunch" mentality. Under UFFP, the requirement that each age cohort obey its own budget constraint eliminates the risk that members of a cohort may attempt to extract larger payouts at the expense of other cohorts. Similarly, by setting up individual accounts and making the benefit-drawer responsible for his own contributions, the MPPF enjoys freedom from political pressures. Diamond (1993, p. 19) finds this a key attraction of the Chilean system: "There is real appeal in individual accounts as insulation of the pension system from political actions to increase benefits without direct financing."

This advantage of the Chilean system, however, will diminish as the distribution of income in society gets more unequal. The more unequal the distribution of income, the greater the pressure for the government to exempt low income employees from the plan, and for the government to provide assistance for the poor old whose accumulated savings may be inadequate. In Hong Kong, the current government proposal exempts employees with monthly income below HK$4000. Political pressure will force the government to raise this exemption level and to support the poor old. The problem of moral hazard and politicization cannot be dismissed. Thus, the inability of the MPPF to provide universal, basic support for the elderly means that there will be demand for an extra tax-financed pillar for the needy. Such a pillar, unfortunately, may lure households into earning below the exemption level of income.


Given the budget constraint listed above, to the extent that the realized rate of return turns out to deviate from the expected rate, actual payouts may deviate from expected payouts. However, while the realized rate of return may fall short of the expected rate for some cohorts, it may exceed the expected rate for others. No one can insure immediate systematic deviations from the expected rate of return for an entire cohort. However, the government has a long time horizon and is in the position to offer, upon payment of the necessary premiums, a guarantee for some minimum real rate of return for funds invested by a cohort that desires to contain investment risks. It can do so by using surpluses for one cohort to tide over deficits for another cohort. This approach to deal with "correlated risks" will achieve Pareto gains for all generations that are averse to such risks, but cannot find commercial insurers to protect them from such risks. Professor Yew-Kwang Ng points out that if we all face a Rawlsian "veil of ignorance," we may prefer some cross-subsidies between different cohorts to reduce uncertainty and increase expected utility. In the same spirit, Ho (1981) introduces the concept of Pareto risk improvement and suggests that an arrangement that led to an ex post redistribution may be Pareto improving if all parties concerned are ex ante risk-averse and if the arrangement reduces everyone's risk exposure.

Compared to a provident fund scheme, which in essence is simply a form of forced savings, the UFFP is likely to result in smaller savings because capital markets are incomplete and imperfect. If capital markets were complete and perfect, the very idea of forced savings would be inconceivable, because those who prefer to save less always could borrow on the security of their future claims on personal "forced" savings. In general, in view of longevity risks, if the provident fund mode of old age security is adopted, more savings will have to be made individually than are necessary to cover the expected life span. Although this means that investment is likely to be lower under the UFFP scheme than under the provident fund scheme, one must not assume that more investment is necessarily desirable, as the benefit of higher current consumption also should be considered.

Because a provident fund scheme alone cannot guarantee a minimum standard of living for the elderly, many of those who subscribe to a mandatory private provident fund scheme also believe that a "mandatory publicly managed pillar" also is needed (James, 1995). In Hong Kong, for example, a motion to establish a pay-as-you-go pension along with the mandatory private provident fund scheme was passed in March 1995 and again in December 1995 (see Mingpao, 14 December, 1995). (However, motions passed in the Legislative Council in Hong Kong represent only recommendations that need not be implemented by the government.) The coexistence of a heavily regulated MPPF with a public pension scheme is, however, administratively demanding. Moreover, the larger this tax-financed pillar is, the greater will be the distortionary effects associated with the related tax burden. Admittedly, the tax-financed subsidies for contribution by the poor under the UFFP also are distortionary. As always, one must find a balance between redistribution and allocative efficiency. Table 2 presents a comparison of the properties of the MPPF and the UFFP.


According to James (1992, p. 50), "the most general recommendation" and one that is "relatively clear-cut and non-controversial" is that old age security should be based on a multipillar system that comprises. (i) a broad based privately managed mandatory savings-annuity pillar to avert the results of myopia and incomplete insurance markets, (ii) a carefully designed public pillar for redistribution to those in long-term poverty and "possibly for insuring against correlated risks," (iii) tax-advantaged private pensions, and (iv) a purely voluntary savings pillar.

A multi-pillar system appears to be both logical and indeed the direction of reform being considered in Europe, America, and elsewhere (Ploug and Kvist, 1996; Business Week, January 20, 1997). Two major questions remain, however. Should the private component of old age security be mandatory and private or should it be voluntary and private? Additionally, if there is still a role for public pension, how should it be reformed so as to be financially viable? The analysis here suggests that the mandatory private option to replace the public pension system is fraught with many problems that would require a heavy close of regulation to avoid. The pay-as-you-go public pension system, however, is subject to uncertainty and financial risks. On the other hand, a mandatory public pension system can deal with longevity risk far better than can individual savings accounts or provident funds.

The fully funded pension system as proposed both provides a mandatory pension that deals with myopia and incomplete insurance markets and offers government subsidies for the poor. It therefore accomplishes by and large the objectives of the first two pillars as envisaged by James. By design, the fully funded pension system is small scale, allowing individuals to "top-up" as they see fit. This recognizes the imperfection and incompleteness of capital markets, which make transforming excessively large future pensions into current consumption difficult for individuals.

A recent article in World Bank Development Brief (August 1995) aptly remarks, "The future course of mortality is a major social risk that must be borne by some group no matter how retirement incomes are organized." The proposed UFFP system is predicated on the assumption that each cohort should be responsible for itself so that no generation is burdened with the uncertainty of supporting another given the fact that mortality and demographics change. This is consistent with a concept of [TABULAR DATA FOR TABLE 2 OMITTED] justice linked to Rawls (1972). (See Ho, 1996, for a discussion.) If everyone faced a "veil of ignorance" and was randomly assigned to different cohorts, everyone probably would prefer a system that requires each cohort to save for its own retirement.

Many countries have different pensions plans for different occupations, and some countries have different pension plans for employees and the self-employed (International Social Security Association, 1987; Noguchi, 1983). The proposed UFFP is based on the assumption that a mandatory plan should be universal, simple, basic, and free from the vagaries of uncertain demographics, without inhibiting each occupation and each company from developing its own private schemes.


FY: Fiscal Year MPPF: Mandatory Private Provident Fund OMB: Office of Management of the Budget UFFP: Universal Fully Funded Pension

This is a revised version of a paper presented at the Pacific Rim Allied Economic Organizations Conference in Hong Kong, January 10-15, 1996. The author thanks Stuart Gillan, Darwin Hall, John Heywood, Yew-Kwang Ng, Stan Siebert, and two anonymous referees for comments on an earlier draft.


Abel, Andrew B., "Capital Accumulation and Uncertain Lifetimes with Adverse Selection," Econometrica, 54, September 1986, 1,079-1,097.

Asher, Mukul G., Social Security in Malaysia and Singapore: Practices, Issues, and Reform Directions, Institute of Strategic and International Studies, Malaysia, 1994.

Auerbach, Alan J., Jagadeesh Gokhale, and Laurence J. Kotlikoff, "Generational Accounting: A Meaningful Way to Evaluate Fiscal Policy," Journal of Economic Perspectives, 8:1, 1994, 73-94.

Diamond, Peter, Privatization of Social Security: Lessons from Chile, NBER Working Paper No. 4516, October 1993.

Dornbusch, Rudi, "And You Thought Social Security Was in Trouble," Business Week, August 28, 1995, p. 7.

Eckstein, Zvi, Martin Eichenbaum, and Dan Peled, "Uncertain Lifetimes and the Welfare Enhancing Properties of Annuity Markets and Social Security," Journal of Public Economics, 26:3, 1985, 303-326.

Haveman, Robert, "Should Generational Accounts Replace Public Budgets and Deficits?" Journal of Economic Perspectives, 8:1, 1994, 95-111.

Hewitt Associates LLC and GML Consulting, Ltd., Report of the Consultancy on the Mandatory Provident Funds System, 1995.

Ho, Lok Sang, Value Capture: Analysis of Policy Alternatives, Ph.D. dissertation, University of Toronto, 1981.

-----, "Institutional Foundations for a Just Society," paper presented at the 8th International Conference on Social Economics, Geneva, 1996.

Ho, Lok Sang, J. J. Lee, and J. P. Leung, Retirement Benefit Programs: Analysis and a Proposal for Hong Kong, HK: CUHK: HK Institute for Asia-Pacific Studies, Occasional Paper, 1991.

International Social Security Association, Relationship between Pensions and Occupations, European Series No. 13, Geneva, 1987.

James, Estelle, "Averting the Old-Age Crisis," Finance and Development, 32:2, June 1995, 4-7.

James, Estelle, and Robert Palacios, "Costs of Administering Public and Private Pension Plans," Finance and Development, 32:2, June 1995, 12-15.

Mitchell, Olivia S., Retirement Systems in Developed and Developing Countries: Institutional Features, Economic Effects, and Lessons for Economies in Transition, NBER Working Paper No. 4424, August 1993.

Mitchell, Olivia S., and Stephen P. Zeldes, "Social Security Privatization: a Structural Analysis," NBER Working Paper No. 5512, March 1996.

Noguchi, Yukio, "Problems of Public Pensions in Japan," Hitotsubashi Journal of Economics, 24, 1983, 43-68.

Ploug, Niels, and Jon Kvist, Social Security in Europe: Development or Dismantlement? Kluwer Sovac Series on Social Security, The Hague: Kluwer Law International, 1996.

Report of the Consultancy on the Mandatory Provident Funds System, Hewitt Associates LLC and GML Consulting, Ltd., April 1995.

Rawls, John, A Theory of Justice, Harvard University Press, Cambridge, 1971.

Roberts, Paul C., "It's Time to Privatize Social Security," Business Week, February 27, 1995, p. 8.

Report of the Consultancy on the Mandatory Provident Funds System, Hewitt Associates LLC and GML Consulting, Ltd., April 1995.

Schwarz, Anita M., "Pension Schemes: Trade-offs between Redistribution and Saving," Finance and Development, 32:2, IMF, June 1995, 8-11.

Ho: Professor and Head, Department of Economics, Lingnan College and Hon. Research Fellow, Chinese University of Hong Kong, Phone 852-2616-7178, Fax 852-2819-7940, E-mail
COPYRIGHT 1997 Western Economic Association International
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1997 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Ho, Lok Sang
Publication:Contemporary Economic Policy
Date:Jul 1, 1997
Previous Article:Fiscal reform in Sweden: what generational accounting tells us.
Next Article:Telecommunications, factor substitution and economic growth.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters