Advisers have to tell it like it will be in retirement

To date, the superannuation industry's greatest attention has been paid to the accumulation phase: how much people need to save to ensure an adequate income in retirement. But the boomers' interest is switching to the retirement phase: how their savings should be managed to best effect.

And there are signs financial planners are working to improve the advice they give retirees. This seems clear from a recent speech by Dominic Stevens, of the annuities provider Challenger. Stevens made extensive use of an article by Joseph Tomlinson, "A Utility-based Approach to Evaluating Investment Strategies", published in the US Journal of Financial Planning. I'll be drawing on both sources.

To date, most advice to retirees has focused on "asset allocation" - how their investments should be divided between shares and fixed-interest securities - and on setting a safe rate at which money can be withdrawn and spent without it running out before the retiree dies.

Tomlinson's objective is to provide advice that is less one-size-fits-all, encompasses more eventualities and incorporates the insights of behavioural economics. These days, computers make it easier to provide more accurate advice and deliver it in user-friendly programs.

Remember, no one knows what the future holds. Who knows what will happen to the sharemarket - or any other financial market? So advice is based on reasonable assumptions and on averages, and advisers seek to estimate expected returns.

But more can be done to allow for the personal preferences of the particular retiree and to take account of the range of likely outcomes around the average.

The first issue is the "risk-return trade-off". The higher returns some investments offer - shares versus fixed interest, for instance - usually reflect a higher degree of risk: risk you won't get your money back, and risk that returns will vary a lot from year to year. It's generally accepted that old people who need to live off their savings can't afford to run the same degree of risk as young people with many years to recover from sharemarket setbacks.

These days more attention is being paid to "sequencing risk". Say you need to live off your savings for 15 years and it's reasonable to expect there'll be two bad years for the sharemarket in that time. Just when those two years occur makes a big difference.

If they come late, it won't be so bad if they come early you could be almost wiped out and never recover. This suggests retirees need to hold more of their savings in fixed interest than many do.

In any case, most people are "risk averse". Consider this choice: which would you prefer, the certainly of earning $100, or a 50 per cent chance of earning nothing and a 50 per cent chance of earning $200?

If you were "rational" you wouldn't care either way because both options have the same "expected value" (for the second: 50 per cent of $0 plus 50 per cent of $200 equals $100).

If you much preferred the certain $100, that makes you risk averse (and normal). If you fancied the chance of walking away with $200, that makes you a "risk seeker". Risk aversion is pretty much the only departure from "rational" behaviour that economists regularly allow for.

A vital question in working out how much of your savings you should withdraw each year (a common rule of thumb is 4 per cent) is how long you'll live. You can't know, of course.

The advisers' standard approach is to look up in the government's actuarial life tables the average life expectancy for someone of your sex and age.

If the answer was 20 years, this would be used for your planning. But a lot of people will fall a bit below or a bit above the average, and Tomlinson's more sophisticated calculations take account of this wider range of probabilities. At present, the main objective in setting your withdrawal rate is to ensure you don't suffer "plan failure" - run out of money before you die.

The alternative to running out is to die with money left - the "bequest amount".

Conventional economics assumes that, dollar for dollar, your pain at having your money run out before you're ready to die would be equal to your pleasure at knowing you'll be leaving a bequest to your relos.

But this seems highly unlikely. As Stevens argues, if a retiree was living on $30,000 a year and that dropped to $20,000, it would have a more profound negative effect that the positive effect of income increasing to $40,000.

The two psychologists who pioneered behavioural economics, Daniel Kahneman and Amos Tversky, call this "loss aversion" (as opposed to risk aversion). They found that most people hate losing $100 about twice as much as they like gaining $100. Since running out of money before you die is a much bigger deal than losing small sums while you're working, it's likely retirees' loss aversion is a lot greater than the usual rate of 2:1. Some preliminary surveys suggest it might be as high as 10:1.

If so, this means retirees' desire to avoid running out of money (and having to fall back on the age pension) is a lot stronger than investment advisers' conventional calculations assume. And this, in turn, suggests retirees' choice of investments ought to be a lot more cautious than it often is.

Tomlinson argues that particular retirees' degree of loss aversion ought to be taken directly into account when determining the best investment strategy to meet their needs. When you do so, the bottom line of the calculation is not the average expected return on their savings but the average expected utility from those savings.

His refinement takes account of the possible size of plan failure - whether your savings are gone one year before you die or 10 years - not just whether or not failure is likely.

It also acknowledges the size of bequests is likely to suffer from diminishing marginal utility. Each extra dollar gives you less satisfaction than the one before.

Shifting the focus from expected returns to expected utility could make investment advisers' advice a lot more realistic and thus a lot more helpful.