You might not know it, but one of the core components of traditional finance advice is hotly debated, and rightly so. Risk tolerance has a couple of definitions, but all broadly related to how much risk a person is willing to bear in exchange for a slightly higher return on investment. While this should be a core means of tailoring portfolios for customers, the question of if we can accurately gauge it, and how we do so, is one of the most hotly contested issues both in academic circles and between practitioners. Many financial advisors lament that risk tolerance questionnaires aren’t accurate, and many clients resent being put through time consuming quizzes which all seem the same.
It’s easy to both make the controversy and the frustration go away. Traditional definitions of risk tolerance confound two different aspects of a persons psyche, giving inaccurate measurements of both of them. Once you separate risk tolerance from loss-reactivity, you’ll have a much more useful view of the individual.
The first conception of risk tolerance comes from academic economists, and has been around for a long time. For economists, it’s a matter of how you make decisions about future outcomes- it’s all about uncertainty regarding where you’ll end up at the end of an investment horizon, say 20 years from now. In formal decision models, risk tolerance is combined with expected returns & expected risk to reach a decision about whether or not to invest. Using these three inputs, there are three reasons someone may decrease their stock allocation:
Their expectation of benefit (returns) has gone down.
Their expectation of risk has gone up.
Their tradeoff measure (risk tolerance) has gone down.
A key point is that to an economist or decision scientist, a good measure of risk tolerance doesn’t move at all in response to market changes. – option three shouldn’t occur. By definition it shouldn’t – it’s a fixed personality trait which helps you make decisions in varying environments or across different options.
Thanks to psychologist Daniel Kahneman, we have a fairly good answer to these weightings- generally, people weight a loss about twice as much as an equivalent gain when they make decisions. A very small set of people are risk neutral – they don’t weigh losses any more heavily than gains. And some people weight losses much heavier – they are not very risk tolerant. So within academic economists, there isn’t too much controversy over risk tolerance (or aversion) as a measure.
However, what most financial advisors want to know is “how will this person react to a loss in their portfolio of a given size?” At the center of the concern is that emotionally jumping out of the market after a loss, and hopping back in after a period of gains is hardly a winning strategy. This emotional market timing reduces investor returns over time, by around 1.6% a year. And so good advisors want to minimize their clients jumps. At the same time, the reward for bearing risk is higher returns, so they want to have their client bear as much risk as possible, subject to not jumping. Advisors care about the risk of the journey, not just the risk at the destination.
This second measure I call it loss reactivity, and it’s much more in line with what advisors want to know. While research that I did while at Barclays Wealth did find it to be correlated with risk tolerance, it is independent enough that you cannot categorize individuals as the same on these dimensions reliably. You can put a high risk-tolerance individual, who also is highly loss-reactive in a high risk portfolio, and they’re more likely to be stressed and jump than a non-loss-reactive individual.
Loss reactivity may also be more valuable, as it has more practical uses than the academic definition of risk tolerance. Correctly identifying an individual’s risk tolerance is difficult and often results in false precision. A great example is this graphic from Isaac & James (2000), which shows that the way you elicit a risk aversion coefficient is almost completely responsible for if individuals are classified as risk-averse, or risk-seeking – a huge difference. Each arc represent an individual’s assessment using one method (above 1, or risk-seeking), and then using a second method (beneath 1, or risk averse).
And that said, it’s not clear that it’s better to use psychometric risk tolerance rather than financial measures when determining how much risk someone should take.. and optimizing for loss reactivity to ensure they actually get from A to B without jumping.
The good news is that we’re already fairly far along the path to diagnosing it ahead of time, and thinking about how to prevent it from harming investors. More to come.