The more I think about psychometric risk questionnaires the less useful they appear. My view is advisers would do better to dispose of them and spend more time thinking about risk capacity.
Purveyors of psyquests will tell you their questionnaires are constructed to give consistent results. Your client will emerge with the same attitude to risk next year as today. They claim they are measuring a stable personality trait.
As Keith Armstrong makes clear in a recent paper on risk (available from firstname.lastname@example.org), the designers simply assume the trait exists and then set about creating measures. The academic literature provides no convincing evidence for such a trait. Moreover, if the trait never changes, what can it usefully predict?
Faced with the evidence of clients’ changing tolerance of risk, the psychies say people’s perception of risk changes in response to events and that their appetite for, or tolerance of, risk therefore changes also. But the attitude to risk derived from the psyquest is not predictive of such changes, their direction or magnitude. So what use is it?
The answer is that you can map the five, seven or 10 ATR profiles derived from a psyquest on to a similar number of profiles corresponding to an asset allocation mix. This is what most risk-profiling tools did until the FSA issued its warning last year. Now they do not make such a direct link – or they wrap a lot of cautionary blurb around the process.
As Armstrong asks, what is the logic of this mapping? Why does someone who has an ATR profile of three get a cautious asset allocation? What justification is there for saying this ATR is consistent with a 30 or 40 per cent exposure to equities?
The answer is there is none, the two scales are simply made to fit onto each other. You can divide the ATR into as many segments as you like, likewise asset allocation mixes. Depending on the assets you choose and the way you measure volatility – weekly, monthly or quarterly – you will end up with a different risk-return map. Depending on how you weight answers to the questionnaire, you can allocate people to an ATR of three, four or five on a 10-segment scale. Neither scale is scientifically validated or supported by any convincing practical evidence.
All you can say in favour of psyquests is that they confirm what a good adviser knows as a result of knowing their client but they do not validate any specific asset allocation for a client any more than the adviser’s knowledge of their client validates it.
The best determinant of a suitable asset allocation is not attitude to risk but capacity for loss or for a shortfall from what is needed to meet a client’s goals. I was interested to note the process created for Sanlam by Rick Eling uses the crunch question on what level of loss the client can accept to override any input from the adviser or the psyquest.
This is the correct approach and I expect more advisers to adopt similar methods once they realise psyquests do not protect them from client claims for losses due to taking on more risk than was appropriate.
Chris Gilchrist is director of FiveWays Financial planning. He edits the IRS report newsletter and is the author of the Taxbriefs Guide, The Process of Financial Planning