29 June 2017
Infinite outcomes for a finite life
Seeking a solution to retirement planning
by Frank O’Nomics
Baby boomers face what has been described as the “nastiest, hardest problem in finance” – that of deciding how to use their savings in retirement. This description comes from the pen of Nobel laureate William Sharpe; a man revered by portfolios managers the world over for his previous work on measuring risk via an eponymous “reward to variability ratio”. The point is, if someone as bright as Mr. Sharpe regards something as hard, how can the rest of us cope? According to Mr. Sharpe reaching the optimal answer of what to draw down from your portfolio each year (whether it be from pension savings or elsewhere), depends on balancing six interrelated sets of interrelated variables and, while none of them is particularly complex, when combined the range of outcomes seems almost infinite. Further, one can argue that there is a seventh highly unpredictable variable to add – that of the changeability of government policy – a factor particularly pertinent in the UK since the introduction of pension freedom. To get some idea of the difficulty we all face it is worth considering each of Sharpe’s variables in turn. At least if we have considered the range of outcomes we can start to assess the risks of an impoverished old age.
That rather neatly leads us into the first variable – how long are we going to live? Mortality tables give us some idea. Average life expectancy in the UK is 81 years, but those who have just hit retirement age will typically reach the age of 87 if male and 89 if female (as an aside life expectancy has actually fallen over the past 3 years). For married couples there are then wide ranges of possible permutations, which will mean that standard assumptions of withdrawing 4% of their savings each year are unlikely to be appropriate for many.
Economists like to make assumptions, so lets make one. Lets assume that we know the length of time for which we will need income. The next big variable faced is the vast range of outcomes for our investment and pension portfolios (especially as the number of people in defined benefit schemes dwindles). We might use past performance as a guide, but this could be highly inappropriate given the implications of an aging population. As the economy has to support more and more retired people, the level of growth generated (unless we have a huge increase in productivity) is going to fall steadily and this will have implications for the return on assets. Similarly, if people are steadily selling assets to pay for their retirement, the value of those assets is going to decrease.
We should also be aware that it is the real value of those assets and the purchasing power they can generate that is most important. The third variable then is inflation. The development of central bank inflation targeting may mean that, over long periods, we can be comfortable in factoring in an assumption of 2%. However, deflation in Japan gives us a guide as to what an ageing population can lead to, while the long-term impact of quantitative easing may still be one of stimulating inflation that becomes difficult to suppress. There is here a fourth variable that is closely related, and that is how markets value inflation protected securities. If you want a secure retirement, buying inflation-linked bonds would seem to be a safe route, but if markets insist on valuing such securities at a prohibitively expensive level (and in the UK real yields are currently negative) then this security will be beyond the scope of most.
The fifth element is the amount of income that we will receive in our retirement. The state pension may seem quite predictable currently, but there is widespread debate over how it increases (the triple lock) and the age at which we will receive it. The other major source of income will be workplace pensions. Those lucky enough to be in a defined benefit scheme may have some confidence, but face a risk of the schemes being insolvent and the prospect of a reduction if the pension protection fund is called upon. Those in a defined contribution scheme have to decide whether to use income drawdown or buy an annuity, the former varying according to investment returns and the latter with long-term interest rates and longevity. It is here that I would look to add a further variable to Sharpe’s, which is the future impact of government policy on pensions and savings in general. Further cuts in the amount of tax-free contributions that we can make to our pensions seem likely and there is also a risk that a pension pot may face tax even on values below £1mn. Similarly, cash strapped governments may find the vast wealth held tax free in ISAs as irresistible, and find a way of generating government revenue here – thereby further restricting our retirement income capacity.
Sharpe’s final variable is that of how the utility of income changes as the years goes by. As we get older we may become less prepared to trade consumption in one period for that in another, but equally we will probably need less. As with all the other variables, this will vary significantly between individuals, particularly given different appetites for risk.
When one looks at the vast number of potential outcomes generated by combining these six (or in my view seven) variables the “head in the sand “ approach might seem the most attractive. However, while there is no doubt that this might be the “nastiest, hardest problem”, by identifying the variables you can at least make some headway in working out some plausible outcomes in relation to your own circumstances. William Sharpe has written a book “Retirement Income Scenario Matrices” (available free online) and is hopeful that financial engineers will develop his analysis.n For those of us less comfortable with complex algebra we will have to hope that ultimately a useful tool will be developed for financial advisors.
If you enjoyed this article please share it using the buttons above.
Please click here if you would like a weekly email on publication of the ShawSheet
