Should risk be an afterthought?

Risk is a phenomenon frequently spoken about but difficult to manage in practice. From an investment perspective, it is often seen as the danger of losing money, not gaining value from an investment, or not keeping up with the client's benchmark.

Does risk matter?

Of course, the answer to this must be a resounding 'yes'. Although decision-making has been studied for centuries by philosophers, mathematicians and economists, it has not been widely explored in experimental psychology, and is often neglected as a separate concept within investment management. As a result, the modern theory of decision-making and risk emerged from a logical analysis of games of chance, rather than from the psychological analysis of risk and value at stake.

Identifying a client's degree of risk aversion is one of the key questions for financial intermediaries. Needless to say, it is not practical to simply ask a client to assess his or her own level of risk aversion. As we know, humans do not always make rational decisions, or determine their preferences in a rational manner. Evidence shows that when it comes to gaining or winning, people exhibit risk-seeking behaviour if small amounts are involved, but risk-averse behaviour for larger amounts. Likewise, people exhibit risk-seeking characteristics when it comes to small probabilities of winning, but the opposite for larger probabilities of losing.

Loss aversion

Behavioural finance confirms what most people already know: that losses seem more painful than gains are beneficial. Investment managers and consultants should always bear this in mind when making investment decisions or giving advice to their clients.

In addition to seeking risk, people also try to avoid loss. This affects the curvature of the utility function (or UF, see figure one). Loss averse behaviour implies convex (curving outward) UFs that are steeper than the concave ones for gains. A practical example would be a lottery that gives the same chance of winning and losing the same amount of money, that is 100, 50, -100. Most people would refuse to gamble because they feel that a loss causes more pain than an equal gain gives pleasure. This is known as 'loss aversion', as explored by Tversky and Kahneman in 1991.

According to Tversky and Kahneman, people's loss aversion coefficient is typically just over two, which means that in the example of the lottery, the upside potential would need to be more than twice the downside potential in order for people to participate.

risk profiling

We have already established that risk matters to investors. So how do we determine their risk profile? In constructing a successful client-profiling questionnaire it is worth considering some of the biases of expected utility theory and rational behaviour. Empirical research has shown that people exhibit inconsistent preferences when the same choice is presented to them in different forms, assuming that ultimately the same information is provided.

For example:

1. Assume yourself richer by £300 than today. You have to choose between:

A: a sure gain of £100

B: 50% chance to gain £200 and 50% chance to lose £200.

2. Assume yourself richer by £500 than you are today. You have to choose between:

C: a sure loss of £100

D: 50% chance to lose nothing and 50% chance to lose £200.

Despite the fact that the two problems are essentially identical, the large majority of respondents preferred A in the first problem and D in the second one. Thus, contextual differences may lead to essentially different results.

Modern vs traditional

The traditional multi-manager model claims to control risk through fund selection. In other words, by not putting all your eggs in one basket (that is, by having a diverse range of funds within the fund), your risk is supposedly reduced and, therefore, controlled. The problem with this theory is that if you put all those baskets in a single truck that crashes then the diversity of baskets is not much use. This was particularly notable in 2000, when so-called 'diversified' funds of funds fell significantly further than the benchmark because they were on the same 'growth truck'. In that case, risk was a by-product of the funds selected, rather than the starting point of portfolio construction.

A modern approach to portfolio construction is more aligned with investors' concerns about the risk of their investment. Investors need to know how much risk they are exposing their capital to in order to properly assess the size of investment they make. The more traditional model, where risk is the afterthought rather than the starting point, does not allow an investor to make this assessment.

Benchmarks have been a central plank of risk control, with the most commonly used relative risk statistic being the tracking error (TE). TE shows the risk of a portfolio relative to the benchmark, while the information ratio is used to measure outperformance over the benchmark return while accounting for the amount of risk taken.

Recently, there has been an increased focus on absolute performance and absolute risk measurement as the concentration on benchmarks has resulted in managers selecting securities from a narrower opportunity set. This has resulted in investors exploring absolute products that focus on generating absolute return, as opposed to just outperforming the set benchmark. The emergence of these absolute return products has resulted in a focus on absolute risk measures. Absolute risk shows how much capital a portfolio can lose, as opposed to what the risk of underperforming the benchmark would be.

We have now seen that risk does indeed matter to investors and that identifying an investor's risk profile is of paramount importance for financial intermediaries. Utility function and loss aversion theory have shown that people find a loss causes more pain than an equal gain does pleasure.

We have also seen how difficult it is to truly determine an investor's risk profile, as different answers are given for differently structured (though essentially similar) questions. Subjectivity and 'psychological satisfaction' also come into play, making the job of determining risk profile all the harder. For fund managers, benchmarks continue to play a central role in controlling risk, though newer absolute return products require the measurement of absolute risk.

In a practical investment context a true diversification of funds is always a sensible approach, where different styles as well as different stocks and sectors are included in portfolios.

Insight Alpha2

At Insight, we don't just base decisions on performance, we consider why and how outperformance has been achieved. Then we consider whether we believe it is repeatable - will these managers continue to deliver good results for our investors? We won't include a manager in our portfolios if we don't fully understand their process. Of course, things change. So we constantly monitor and review the funds we have chosen, rebalancing and monitoring holdings to ensure our portfolios are positioned to make the most of prevailing market conditions.

At Insight, we implement this by utilising a process called Insight Alpha2, which places risk management at the centre of portfolio construction and uses sophisticated proprietary quantitative tools to support portfolio decision-making.