The World Bank has signalled a shift in its position on the pension issue. With the publication in February of ‘Old age income support in the 21st Century’, an international prospective on pension systems and reform it indicated that it had taken on board some of the criticisms that have been levelled against its seminal work on pensions, ‘Averting the old age crisis’, that was published more than a decade ago.
That study introduced the three-pillar model and the concept of mandatory individual funded retirement savings, ideas that have been widely adopted from Latin America, through central and eastern Europe to Kazakhstan.
In addition, it promoted the concept of notional defined contributions (NDCs), the so-called Swedish model that features individual accounts which are funded on a pay-as-you-go basis and which has since been implemented in Latvia, Poland, Italy and China and is being considered by the Czech Republic.
However, this approach provoked considerable criticism from such institutions as International Social Security Association (ISSA) and the International Labour Office (ILO), and from the academic world. The individual accounts did not fit well with traditional patterns of PAYG systems based on solidarity, and notional accounts were seen as inferior to real funding.
But although the ILO criticised the World Bank’s concept, it did so in such a low-key and indirect way that the divergence between their points of view has seldom been registered.
And for many years the World Bank did not appear to be willing to engage in debate with those holding alternative opinions. Peter Orszag, senior director at Sebago Associates and formerly a special White House assistant on economic policy, and Joseph Stiglitz, professor of economics at Columbia University and senior vice-president and chief economist at the World Bank, used a conference marking the fifth anniversary of ‘Averting the old age crisis’ to present their own paper, ‘Rethinking pension reform: Ten myths about social security systems’, which argued that the 10 reasons used by the World Bank to support a switch out of statutory PAYG old age benefits were not valid.
Robert Holzmann, director of the World Bank’s social protection unit, said their paper was interesting but provocative. Nevertheless, although World Bank staff did not agree with the arguments, the conference was one of the first times that the World Bank openly discussed criticism of its pension reform proposals.
But with its new paper, which is by Holzmann and Richard Hinz, a World Bank pension policy adviser within Holzmann’s social protection unit, and was published in February, the World Bank signalled that it has reconsidered its view on both the three-pillar system and NDC. And while the World Bank characterises its shift as an evolution of its prospective on pension reform, for many it marks a revolution.
The original World Bank three-pillar concept encompassed:
q An unfunded, publicly managed defined benefits system;
q Mandatory funded and privately managed defined contribution schemes; and
q Voluntary retirements savings.
Criticism of this approach focused on several key issues, notably that the recommendation of a one-size-fits-all model was a straitjacket for all governments needing support from the World Bank.
At the fifth anniversary conference, Holzmann had denied the accusation, saying that the multi-pillar approach was not a benchmark and the World Bank did not favour a specific approach. He noted that it had helped Lithuania to implement a relatively modest PAYG system with a redistribution formula. Nevertheless, the model was promoted aggressively and World Bank staff advocated it in almost all national pension debates. The new paper makes it clear the World Bank will be less rigid in its policy prescriptions and support a wider variety of reform options.
The World Bank was also criticised for not paying attention to the need for adequacy of pension levels. The ILO recommended what amounts to a four-tier system, including a ‘zero tier’ that would effectively include a social safety net (social assistance) tier outside the pension system to protect the elderly against poverty. Further, ILO conventions contained a replacement ratio for earnings-related pensions.

The World Bank appears to have adopted this zero pillar approach and its new study also hints at minimum levels for pension provision. While these differ from the levels that international organisations like ILO see as adequate, they are a first step. And the new paper addresses criticisms that the World Bank has focused too much on the pre-funding of pensions and had not paid enough attention to the high costs of administration and vulnerability to bad governance, economic shocks and the effects of ageing populations.
It still advocates a higher level of mandatory retirement savings, seeing advanced funding as both useful and necessary, but it concedes that there are limits to funding and mandated participation in pension schemes.
The World Bank has engaged in dialogue with key international trade union organisations but until relatively recently refused to agree that there could be another valid view on the pension topic. But in May 2003 it discussed with the International Confederation of Free Trade Unions (ICFTU) allegations that it paid too little attention to solidarity between generations and the role that collective agreements and unions can play in pension provision. Although they continue to differ on several points there seems to have been an agreement on the need to put coverage and the inclusion of informal workers at the top of the agenda and on the importance of involving the social partners in pension reform.
This meeting is just one example of the World Bank’s dialogue with academics, partner organisations and international institutions, and its process of internal discussions about and evaluation of pension reform policy over recent years.
The result has been a reshaped multi-pillar model, replacing the three-pillar model with a five-pillar pension system which is composed of:
q A non-contributory or ‘zero pillar’ (in the form of a social pension) that provides a minimal level of protection;
q A first-pillar contributory system that is linked by varying degrees to earnings and seeks to replace some portion of income;
q A mandatory second pillar that is essentially an individual savings account but can be constructed in a variety of ways;
q A voluntary third-pillar arrangements that can take many forms (individual, employer-sponsored, defined benefit, defined contribution) but are essentially flexible and discretionary in nature;
q Informal intra-family or intergenerational sources of both financial and non-financial support to the elderly, including access to healthcare and housing.
The new paper makes it clear that the World Bank continues to see advantages in a multi-pillar design that contains funded elements but that it increasingly recognises the importance of first implementing some basic conditions and that a range of choices can help policymakers achieve an effective and tailored implementation of the multi-pillar model. It now also acknowledges that the primary determinants of an appropriate pension reform are the particular conditions and circumstances of the environment in which it occurs and so it must be tailor-made rather than off the shelf.
It stipulates that its model can only be seen as a benchmark and that it must be clearly distinguished from a default option or policy prescription. This may be good news for all those governments meeting internal resistance to pension reform because of opposition to the World Bank’s three-pillar-model with mandatory individual accounts.
The new multi-pillar design still leaves room for first-pillar provisions based on notional accounts, the NDC.
This non-financial defined-contribution structure is appealing because of its simplicity, transparency and adjustment to increasing life expectancy. Italy, Latvia, Poland and Sweden have adopted it and Kyrgyzstan and Mongolia have transformed their statutory defined benefit systems into NDC schemes.
But for some this is old wine in new bottles, as Michael Cichon, head of the ILO’s financial, actuarial and statistical services branch (Soc/Fas) department, entitled a critical 1999 article. In his opinion NDC schemes are mathematically identical to defined benefit schemes with a correct accrual rate and actuarial increments or decrements for late or early retirement. Yet many policymakers see it as an important switch.
John Williamson, research associate of the Centre for Retirement Research at Boston College, Massachusetts, has written that it mirrors the philosophy of a funded system with individual accounts, but with a PAYG financing structure. So while it is attractive to policymakers it is also often confusing for the participants who do not realise that their account is unfunded.
Also it is not a what-you-see-is-what-you-get system. In an NDC system an individual account is established to which contributions by the individual (and his or her employer) are earmarked and on which a notional interest rate (essentially the growth rate of covered wages in a mature system) is paid. At retirement, an annuitisation factor that reflects the remaining life expectancy of the cohort at retirement and the relevant interest rate is applied to the accumulated notional individual account to determine the benefit payout. So participants think they have a secured old age provision, but at retirement age things can look completely different.
The confusion stems from such issues as the annuitisation factor differing from country to country. Sometimes it is based on unisex life expectancy at the time of retirement but mostly it is based on a retirement age of 65. In Sweden, for example, the formula is based on age-specific unisex life expectancies at the time when the worker’s cohort reaches age 65. The procedure used is very similar to that in other countries with NDC schemes, although there is some variation with respect to how often estimates of life expectancy are updated.

Consequently, the NDC option has attracted a good deal of criticism. In an evaluation of notional accounts as a pension reform strategy, Richard Disney, professor of economics at the University of Nottingham, concluded that if the object of a shift to notional accounts from traditional defined benefits schemes was to make it clear that pension claims should be sustainable, then it was unclear why such a reform would be politically superior. Disney noted that there was a real risk of a costly loss of credibility when the reality of the system was revealed.
His paper, which was published on the World Bank website, marked the outbreak of a fierce dispute over the NDC concept that continued at a defined contribution conference organised by the World Bank and the Swedish Social Insurance Board in September 2003. The conference saw experts exchange experiences from countries that had introduced NDC systems and their reports exposed the major pitfalls.
The key problem was that the concept could be hijacked by politicians and could lead to less-than-optimal system designs through, for example:
q A choice of the ‘wrong’ notional interest rate, such as per capita GDP or wage growth instead of growth of the total wage base;
q The use of the remaining cohort life expectancy as a fixed factor, as in China;
q The granting of contributions to the individual account even if they had not been paid, as in Italy;
q The underestimation of the substantial administrative requirements to run such a scheme, as in Kyrgyzstan and Mongolia.
The World Bank recognises these pitfalls but despite such negative experiences it still believes that reforms based on the NDC concept are likely to be the best way to restructure a typical unfunded defined-benefit scheme within a multi-pillar structure.
Yet the questions of what constitutes an ideal notional DC system, how it fares when compared with other benefit options (such as non-financial defined benefit or financial defined contribution schemes) and how it performs in reality and under political stress are still very much open to discussion. The World Bank does not yet have answers to these questions.
Hasko van Dalen is adviser on social security, pensions and employee benefits at ING Nationale-Nederlanden