Paying for pensions: An international comparison of administrative charges in funded retirement-income systems

High charges for personal pensions were one factor in the personal pensions mis-selling debacle in the United Kingdom. They continued to arouse concern among politicians and commentators. The Labour government, with its new flagship ‘stakeholder’ pension, chose to regulate both the structure of charges and their level. This paper assesses the international experience of charges in funded retirement-income systems, drawing on evidence from fourteen countries with very diverse policies. Measuring the price of financial services is more difficult than comparing the cost of other goods or services. Providers can levy many different kinds of charges. These can include one-off and ongoing charges; proportional and fixed-rate fees; some based on contributions, some on the value of assets in the fund and some on investment returns. These different charges accumulate and interact in Complicated ways over the membership of a pension plan. The most familiar summary measure of charges is the ‘reduction in yield’. This adds together all the charges over the lifetime of an example pension policy and expresses them as a percentage of assets. Measuring charges as a proportion of contributions is the alternative. This turns out to be the same as calculating lifetime charges as a proportion of the balance accumulated at retirement. This second measure is known as the ‘reduction in premium’ or the charge ratio. The fourteen countries surveyed (Section 2) adopt very different approaches. At one end of the spectrum, Australia and the United Kingdom (with personal pensions) have completely liberal policies on charge levels and structures, but require providers to set out the effect of charges in a standard format. Most Latin American countries, including Argentina and Chile, restrict the charge structure: in these cases, allowing a fixed fee plus a charge as a proportion of contributions. Poland, too, limits the types of fee that can be levied, but also limits funds to charging 0.6 per cent of assets, while other charges are uncapped. Sweden, Kazakhstan and the United Kingdom (with stakeholder pensions) restrict both the charge structure and the charge level. In the last two, there is a fixed ceiling while Sweden varies the cap using a complex formula based on the amount that providers charge to manage voluntary savings. Finally, Bolivia auctioned the rights to manage its mandatory pension fund assets to international fund managers. The empirical evidence from these countries charge levels. In countries with systems based on individual accounts and individual choice among competing pension providers, average charges vary from under 15 to above 30 per cent. The paper assesses the options and the arguments for ontrolling charges. Measures to increase transparency comprise requirements for providers to disclose the level of charges, public provision of information in charge ‘league tables’ and allowing charges to be levied on top of rather than out of mandatory pension contributions. If governments choose to restrict charge structures, to facilitate comparisons between different providers, the most important policy choice is between contribution-based levies and asset-based fees. Latin American countries have tended to opt for the former, the United Kingdom has chosen the latter for stakeholder pensions. The main issues in this choice are the time profile of providers’ revenues, fund managers’ incentives to maximise returns and the incidence of the charges on different providers. Restricting charge levels raises some important concerns, particularly about governments’ ability to choose the ‘right’ level for the ceiling and the trade-offs in terms of restricting competition and individual choice of fund. Many of these policies to limit charges are aimed particularly at protecting low-income workers. But some countries have adopted alternative policies: for example, excluding low-income workers from the requirement to contribute and protecting them with safety-net pensions in old-age or cross-subsidising low-income workers directly with a minimum contribution from the government. Some commentators have suggested alternative institutional structures for managing funded pension assets to reduce costs. However, empirical evidence shows that publicly managed pension funds have generated poor returns. Also, the evidence on economies of scale in fund management suggests that the minimum efficient scale is relatively small and does not imply the presence of efficiency gains from a monopoly in managing funded pensions except in small economies. Again, there are important trade-offs in these policies, including corporate governance problems and the restriction of competition and individual choice.