A few weeks ago, I wrote about the perils of risk profiling, so I was pleased to read the FSA paper that identified lots of short-comings in risk-profiling tools as well as in their users. Some of these are obvious but the FSA’s prescriptions have scary implications for advisers and product providers.
To start with the obvious, someone’s stated preferences about risk and their under-lying attitudes to risk, as revealed by psychometric questionnaires, do not necessarily correspond with their capacity for risk.
This is determined mainly by objective factors such as their resources, age, health and expectations. The FSA rules say advisers must take risk capacity into account in making recommendations. Many do not. As I remarked before, if you go straight from the answers to a psychological questionnaire to a portfolio, you may omit consideration of capacity and therefore give bad advice. The FSA has just given a wake-up call.
In my view, the psychometric questionnaire is an ingenious ancillary tool but it is a recent invention. Before it existed, advisers assessed risk capacity either explicitly or implicitly. They just need to put the horse back in front of the cart. Objective factors should always outweigh subjective ones in the advice-generating process.
A secondary point is that where capacity is the key issue, then probabilistic assessment of loss is inappropriate. To put it another way, if you have limited risk capacity, then it does not matter that the percentage of loss with an investment is small. What the adviser has to focus on is the consequences of a loss. Warren Buffett puts it like this: “A small chance of distress or disgrace cannot be offset by a large chance of extra returns.”
Those who recommend structured products often, in my view, fall into this trap. They do not consider the awful consequences of a loss (as with Lehman-backed structured products) because they are misled by the probabilities.
Back to the main issue. Tools that create portfolios can use attitudes to risk derived from questionnaires but measuring someone’s risk capacity with a few questions is much more difficult. Some creators of risk-profiling tools such as Oxford Research & Analysis are aware of this issue and are working on solutions.
Even if they succeed, the FSA paper makes creating usable tools harder. Quite rightly, it says advisers who use such tools must understand their construction and assumptions. In the case of tools that use stochastic processes to gener-ate portfolios, I would guess that eliminates about three-quarters of advisers. I think you can only really understand these tools and explain their output without the risk of misleading a client if you understand the portfolio theory on which they are based. And portfolio theory is not easy – I would argue that it requires A-level maths or better. So bigger, better black boxes are effectively ruled out. Perhaps simplicity will get a look in.
Since the FSA leans towards interpreting risk capacity as the ability to withstand losses, it may be that drawdown (the maximum historic peak to trough loss over relevant time periods) is a better measure of capacity. That will raise another set of tricky technical issues.
You can argue that trying to make investment advice more scientific is doomed by the uncertainty that prevails but well designed tools could at least reduce the frequency of bad investment advice, which would be a good outcome.
Chris Gilchrist is director of Churchill Investments and editor of The IRS Report