No free lunches in providing for an ageing population

Disease, wars and unhealthy work conditions meant that until recently relatively few people could expect to live much beyond retirement. As recently as the 1950s male average life expectancy at birth in the UK was no greater than the statutory retirement age; but as the century draws to a close male life expectancy has risen, so population growth in most developed countries has slowed. The combination of these factors means that in most Western countries a huge burden will be placed on state-financed retirement schemes.

Recent OECD estimates show that if state pensions remain as generous as they now are within Europe then either tax rates in the main economies have to rise sharply or substantial budget deficits will be generated over sustained periods. In the UK the scale of the problem for future governments is likely to be less than in Germany, France or Italy where, on unchanged policies, the present value of all future pension deficits is well in excess of annual GDP. But this apparently healthy UK position depends on a continuation of the strategy of lifting the basic state pension in line with price inflation rather than with the rise in average earnings. Such a policy implies that the state pension will tend to keep falling relative to average earnings; if real wages grow at the long run average of a bit over 2 per cent a year, the ratio of retirement income to earnings for someone relying on the state pension will be only around 10 per cent by 2030.

One of the reasons why reducing the relative value of state pensions may be politically more feasible in the UK than in other European countries is that private pensions - either occupational or personal - are far more important here than in most countries. Greater reliance upon private pensions, especially among the politically influential middle classes, means less hostility towards gradual erosion of the value of the state pension. But the fact that such a policy is more likely to be feasible here does not make it more desirable. The real economic question is whether it is efficient to move away from state pensions financed through the tax system, where one generation of workers provides the funds to pay the pensions of the current elderly, to a system where individuals accumulate a fund of assets from which incomes are paid in retirement.

One way of thinking about this question is to compare the returns on contributions made through a typical working life for a state-run pay- as-you-go scheme to the returns earned in a funded scheme with personal contribution rates levied on the wages of current workers. These are higher than the contributions they themselves paid while at work by an amount depending on how much higher total real wages are "today" compared with "yesterday". In other words the rate of return on contributions paid into an unfunded state pension scheme is the growth in the real wage bill over a working life. Over the long term this is likely to be roughly equal to real GDP growth. In contrast, the return from making pension contributions into a fund is the average rate of real return on the investments.

The table shows for the major economies the average growth in real GDP over the past 35 years and the real return over that period that could have been earned on a fund equally invested in domestic bonds and equities. In the majority of cases the real return on the fund exceeds GDP growth. In some cases the difference between returns on funds and GDP growth is substantial; in the UK pounds 1,000 equally invested in bonds and equities in 1960 would now be worth about pounds 5,000 in real terms; the same investment would be worth under half that amount if it grew in line with GDP. If one took the average over all countries in the table then the returns of a fund would exceed the growth of national income by about 40 per cent over a typical working life.

On the surface it might seem that the figures in the table provide a powerful argument for switching to a funded scheme. But there is a crucial complication to the apparently simple issue of comparing GDP growth with real returns on savings. Any switch from an unfunded to a funded scheme must take account of the past contributions made by those nearing retirement at the time of the switch; those contributions have not been used to accumulate a fund and such people have too little time left to work to build up their own funds. Obviously the current retired have no funds to their name but did make contributions while at work. So there are obligations to the current elderly, but a switch to a funded scheme means these cannot be met from contributions by younger workers, whose payments are now going into a fund. In short there is a transition problem; allowing young workers' contributions to build up in a fund removes what has hitherto been the source of money for paying pensions.

Now the benefits of higher returns earned in a fund would accrue to every future generation, so it seems a shame to lose them because of a one-off transition problem. If you thought it was reasonable to treat the welfare of future, as yet unborn, generations as equivalent to those of the current old then their extra income added up over future millennia obviously outweighs the losses of the current elderly. So, arguably, it is worth sacrificing the welfare of today's old as a cruel solution to the transition problem. This is tough on today's old.

If the future benefits of switching to a funded scheme are so big, then surely we could work out a way of compensating today's elderly so they are no worse off and still generating benefits for future citizens. Surely we do not have to abandon today's old? In fact this is not so easy to avoid. One strategy certainly will not work: the policy of borrowing the money needed to pay the current generation of retired (and to buy out older workers for their past contributions) and paying back this debt from higher future taxes will be self defeating if the interest on that debt matches the returns on pension fund assets. If that were the case the extra return on the assets accumulated from the pension contributions of the current young would just have to be paid out again as higher taxes in the future.

All this is very frustrating. But it is no more than an application of that well-known piece of advanced economic theory about free lunches being a mirage. This is worth remembering the next time someone tells you that switching to funded private pension schemes is obviously the answer to the problems generated by Europe's greying population. Ultimately, it might be a nice place to get to but it is better now to start from here.

David Miles is professor of economics at Imperial College, University of London and an economic adviser to Merrill Lynch.

Returns from 35 years

% GDP growth % annual

1961-94 average real return

1962-94

average

US 2.99 4.31

Japan 5.80 4.67

UK 2.32 4.39

Germany 3.05 4.06

France 3.40 3.60

Italy 3.52 1.89

Netherlands 3.14 3.92

Belgium 3.12 3.32

Ireland 4.27 4.76

Denmark 2.83 6.42

Spain 4.21 3.13

Austria 3.31 5.43

Norway 3.74 4.95

Sweden 2.46 5.34

Average 3.44 4.30

Real return equals average of annual real return on long-dated government bonds and return on domestic equities