Orszag, J. Michael - Department of Econommics, University of London

Dr. J. Michael Orszag [1]

Individual Accounts: Lessons from the UK Experience

The role of individual accounts in Social Security reform is perhaps the most contentious issue to be debated at this conference. The UK experience may help inform the discussion. In 1988, the UK took the groundbreaking step of allowing individuals to contract out of the earnings-related portion of the state pension (SERPS) by opening up personal accounts (Appropriate Personal Pensions). Since Britain is the only major industrialized country to experiment with these accounts, its experience provides a unique laboratory in which to investigate both the opportunities and problems of a switch to such accounts.
In terms of opportunities, establishment of personal accounts in the UK has had important incentive effects [2] But the UK experience has also indicated a number of practical problems for policy design. In particular, the administrative costs of running individual accounts have proven to be surprisingly high, and misleading sales practices have produced a $15 billion scandal.
To evaluate administrative charges in the UK, it is useful to define the charge ratio, a measure of how much of a pension's value is dissipated due to administrative charges and other costs. In , particular:

1 - pension w/ Charges Charge Ratio = ---------------------- Pension w/ No Charges

The closer the charge ratio is to zero, the lower the costs. The charge ratio can be decomposed into three components, corresponding to losses upon retirement (annuitization), losses from building upon funds before retirement assuming no switching among funds (accumulation), and losses from switching among funds before retirement (transfer):

The annuity ratio reflects the losses from annuitizing an account at retirement. It measures the ratio of private annuity yields to theoretical yields from population mortality tables, and captures both adverse selection and cost loadings on private annuities.

The accumulation ratio captures fund management and administrative costs during the accumulation stage of a worker's career. It assumes that individuals do not switch funds during their careers.

The transfer ratio measures the costs from switching funds during a worker's career. It is computed as the ratio of the amount received at retirement by an individual switching funds a typical number of times to the amount that would have been received at retirement by the same individual if he/she had not switched funds at all.

When there is no lump sum pension payment, these ratios enter multiplicatively:

Pension w/ Charges --------------------- = Annuity Ratio x Accumulation Ration x Transfer Ratio Pension w/ No Charges

In combination, these three sources of costs are substantial. Reasonable figures for the UK are an accumulation ratio of roughly 75% (consistent with a 9% equity yield, a 50 basis point fund management charge, and a 5 percent annual fee), an annuity ratio of 85%, and a transfer- ratio of 70%. The total charge ratio is therefore l- 0.75 x 0.85 x 0.70 = 0.55, so that 45% of the pension is _ lost in charges. In other words, the value of a private account in the UK is cut roughly in half by administrative and other costs.
These high costs are not just hypothetical figures. We have calculated the three component ratios for the UK, using data back to 1988 (when personal pensions were introduced):

Accumulation ratio. Data from Money Management surveys indicate an implied accumulation ratio for a typical worker who works 40 years of about 75% in 1997.

Annuity ratio. For annuities, adverse selection and cost loadings amount to roughly 10-15% of a pension's value. The annuity ratio is thus roughly 85%. [3]

Transfer ratio. The transfer ratio has improved considerably in the UK, but remains relatively low. The Money Management surveys imply that the ratio for one transfer after 5 years has risen from 79% in 1994 to 89% in 1998. But even in 1996, over a quarter of the market had five-year transfer ratios below 80%. [4]

Two other aspects of the UK experience are worth noting:

Providers have not left the market. The number of pension providers in the UK has not changed much over time: in 1992 there were 90 unitised personal pension plans, whereas in 1998 this number was 91. [5]

New disclosure rules have not significantly affected average charge ratios. In 1995, the UK government introduced strict disclosure rules for personal pension costs. These rules have had surprisingly little effect on the average charge ratio, but they do seem to have reduced the variance in charges across providers (which declined by 25% between 1994 and 1997).

In summary, the UK experience provides a number of useful policy lessons for the US debate. [6] First, administrative costs are a substantial issue with private pensions, both because of charges during the accumulation and because of reductions in yield during the annuitization stage. Second, free competition over more than 10 years has not resulted in a substantial reduction in providers -the mutual structure of the industry has perhaps impeded mergers which might have exploited economies of scale to bring costs down further. Third, disclosure has had important effects in reducing the variance of charges across providers.

1 Address: Department of Economics, Birkbeck College, University of London, 7- 15 Gresse St, London W 1 P 2LL, UK. Phone: +44-171-63l-6427, Fax: +44-171-63l-6416, Email: jmo@ricardo.econ.bbk.ac.uk.
2 Stefan Folster, Social Insurance Based on Personal Savings Accounts; J.M. Orszag and Dennis Snower, Expanding the Welfare System: A Proposal for Reform in European Economy, 1997,4.
3 The actual profit loadings on annuities are quite small, and in fact ex post profits recently are negative due to better than expected male longevity.
4 These hidden charges are particularly confusing to consumers, because most providers claim they have no charges for transfer values. This is misleading because most providers front load their annual charges, hence imposing hidden charges on those who do not hold their accounts with the same provider for a long period of time; the difference between a fund with level charges and one that is front-loaded is captured by our transfer ratio. Such front-loading is quite significant: even in 1998, the market average ratio of one year transfer values to a fund with no charges was only 53%.
5 Some uncompetitive life offices have shut down their pensions operations and have been replaced with nontraditional direct channel providers such as Virgin Direct, Marks & Spencer and supermarkets. But competition in the market remains strong, as evidenced by Fidelity's decision to withdraw from the personal pensions market in 1993.
6 A more detailed investigation of UK pensions and annuities costs is currently the subject of a detailed panel data study conducted under the auspices of the World Bank by Dr. Mamta Murthi (Cambridge University), myself, and Dr. Peter Orszag (Sebago Associates, Inc.). The study is correlating charges with pension and annuity plan details and insurance accounting data, to examine in more detail the causes of relatively high costs. The study is also comparing personal pension and employer-provided pension costs, as well as examining annuity prices and cost loadings over time.