With a real rate of return on assets of 6 per cent a pension worth 60 per cent of final salary that then kept pace with typical wage rises could be generated by saving about 10 per cent of income for every year of work.But if the working life is incomplete, or the return on saving lower, things look worse. If assets only yielded 4 per cent – still a good bit more than can now be earned on safe investments like government inflation- proof bonds – the saving rate would need to be about 20 per cent. That is very much higher than the average household saving rate in the UK. And it is dramatically higher than the average (over the working life) saving rate of a high proportion of the population who hold few financial assets at any point in their life.Rather than expect, or force, people to save much more in their working lives it may be more desirable for people to work to greater age.

It is increased life expectancy that is the main factor behind the projected fall in the support ratio. With unchanged retirement, greater life expectancy either requires greater contributions from income during work or lower consumption in (a longer) period of retirement. But to let all the adjustment from greater life expectancy come through lower consumption, with no extension in the typical working life, is strange. Market mechanisms might naturally make the option of working longer more attractive.

As the population ages one can expect wages for increasingly scarce workers to rise. At the same time the incentive to work longer is clear enough for those who have saved little in earlier life and now face the prospect of living for 20 years on a state pension worth only a small fraction of typical wages.There are two powerful reasons why an extension in the typical working life, brought about by a rise in the age at which peoplestop work, is appropriate. First, people are healthier and stronger for longer than in the past while most work is much less physically demanding. Even manual jobs (gardening, making cars) have been revolutionised by machines capable of a wide range of tasks.Second, the proportion of people’s lives that are spent working has fallen to very low levels. Consider the lives of typical males born in the UK 150 years ago, 50 years ago and today. An average boy born in 1847 would have started work before becoming a teenager (and quite probably long before). He would have had a life expectancy of just over 40 years and would probably have worked for close to 60 hours a week for all but one or two weeks of the year.

Assuming eight hours a day spent sleeping, this would imply that about 40 per cent of his waking life would have been spent at work. A boy born 100 years later, in 1947, could expect to live for 64 years (just under statutory retirement age). Typical hours of work are down to about 50 while the typical age at which works begins is probably up to around 17. Now, a bit above 30 per cent of the total waking life is spent at work.What about a boy born today? If there were to be no change to the current retirement pattern, he would not expect to work beyond 60 (and would be highly likely to stop work before that).