Reform of pension systems has become a major preoccupation of governments around the world. In 1997, for example, systemic pension reforms were undertaken by three more countries in Latin America (bringing the number of reformers in this region to eight) and three transition countries – Kazakhstan, Hungary and Poland – passed new pension laws or started implementing new pension systems. Many other countries, including China, Russia and Brazil, are also preparing to reform their pension systems. At a lunchtime meeting on 11 December 1997, hosted by the Research Centre of the Faculty of Economics at the University of Ljubljana, Dimitri Vittas (The World Bank) outlined the progress of these pension reform programmes and the lessons which can be drawn from their experiences for Central and Eastern Europe. The meeting was organized jointly by CEPR and the Institute for EastWest Studies under the auspices of the Economic Policy Initiative.

Vittas pointed out that the motives for reform of pension systems are similar across most developing countries. They include serious design faults; poor performance; and (of less relevance) demographic ageing. Design faults include young retirement ages; lax provisions for early-retirement and disability pensions; high targeted replacement rates; high payroll taxes; and widespread evasion. Systemic pension reform programmes are complex, however, and successful implementation requires detailed preparation. The main obstacle is not technical but political, in that reform programmes are highly vulnerable to dilution through the legislative process.

Although no two programmes are identical, they all share two features: the downsizing of public 'pillars'; and the promotion of private pension funds. The multipillar structure is predicated on both theoretical and practical arguments. The theoretical argument is about the alignment of pillars and functions; the practical about the wisdom of diversifying across providers. This is relevant because pension contracts span a period of 60 years or more, while experience with fully fledged and mature pension systems, as well as deep and well-regulated capital markets, is not only shorter but also far from fully satisfactory.

There are many different ways in which either the public or the private pillar can be organized. The choice often reflects local political and historical factors, but it is also shaped by the preferences of reformers and their advisers. For the public pillar, what matters is to reduce the generosity of promised benefits and contain the extent of redistribution. Modest promises are more affordable and hence more likely to be sustained. For the private pillar, the key issue is operational and investment efficiency, rather than the specific form.

Integration of the two pillars is important to avoid inequities and distortions at the margin. The Swiss pension system has an excellently designed public pillar and admirable integration between the two pillars. The efficiency and transparency of the funded private pillar leave much to be desired, but as the pillar is operated with a targeted replacement rate, transparency matters less than it would in other circumstances.

A very important aspect of the reform programme is the regulatory framework for the private pillar. A segmented approach with draconian regulations may be justified at first, but relaxation of regulations and a move towards an integrated system should be undertaken as the system matures. Asset segregation and external custody are, however, essential preconditions for a successful reform.

The main weaknesses in developing countries have been the high marketing intensity, unexpectedly high levels of account switching, and the high operating costs of private pension funds. These seem to be related to uniform pricing rules and the prohibition of loyalty, group and volume discounts, and perhaps also to structural controls in the form of 'one account per worker' and 'one fund per company'. To contain marketing costs and account switching, Latin American countries have started to allow loyalty and group discounts, while also imposing limits on account switching. Kazakhstan is planning to impose low ceilings on fees, while Bolivia has authorized only two management companies with pre-assigned membership. No country, however, is contemplating removal of the uniform pricing rule.

Progress during the transition phase depends upon initial conditions and upon political and social constraints. Countries with low pension costs, low tax rates, low public debts and ample privatizable assets are in a better position to undertake bold and speedy reforms than countries at the opposite end of the spectrum. The dynamic consequences of the transition process are also relevant. Pension reform may stimulate formalization of employment and capital market development as well as spur saving and growth. The impact on labour and capital markets, as well as on saving and growth, is likely to be greater in developing countries where these markets are less well developed, and where unfunded public pillars suffer from far more serious design faults, and cause far greater distortions in economic incentives.

It is often argued, Vittas noted, that the creation of funded pension schemes should not be encouraged in countries that have underdeveloped capital markets and weak regulatory structures. Although widely held, this view is not fully justified, since it disregards the dynamic interaction that is likely to emerge between pension funds, capital markets and regulatory structures. The main pre conditions for successful pension reform are a strong political commitment, effective external custody and access to foreign expertise. International financial institutions and capital markets can therefore facilitate the implementation of reform programmes by making technical expertise available, by providing financial support, and by developing new instruments to encourage risk diversification and to meet the financial needs of retired workers.