Sound judgement is a key element in running any business. In banking and finance this skill is seen as vital to be able to manage and price risk successfully. After all, if a bank cannot correctly monitor, judge and manage the risks on its books, what prospect does it have of surviving, let alone prospering?
The issue of management judgement is not new and the modern banker has a huge raft of tools to help in the task of correctly assessing risk. In the past 50 years or so there has been an explosion in so-called quantitative measures of risk. Whether it is the principles of sensible portfolio diversification laid out by Harry Markowitz or the breakthrough in the pricing of options by Messrs Black & Scholes, or the proliferation of Value-at-Risk models in the past 10 years or so, everywhere bankers are offered a huge range of instruments and measures with which to manage, control and judge risk.
Yet, while these developments in mathematical risk models and the availability of cheap and powerful desk-top computing have helped to revolutionise finance, the problem of risk has not been solved. Dreadful mistakes and expensive errors of judgement still come to light, often in organisations that have access to huge resources and brain power. With all this help at hand, why do things still go wrong?
Increasingly creative and complex solutions have been applied to risk inputs and to the creation of more and more sophisticated calculating routines and algorithms. But the banking industry has been slow to look at what is done with the results of these models and black boxes. The risk outputs generated are often misunderstood, or even wilfully ignored or misused. It appears that even when financial institutions are armed with the cleanest and most accurate data and the finest software and risk expertise, they often fail to take into account a vital ingredient: human nature or, more precisely, the way in which people view and treat risk.
New wave of thinking
Until recently, the science of management judgement and our attitudes towards risk and decision making was overlooked. However, research in the past few years in the field of behavioural finance and prospect theory looks set to unleash a new wave of risk thinking. We may be on the verge a second and even more powerful revolution than the current in-vogue quantitative approach.
Many readers will be familiar with some of the broad ideas of behavioural finance, and in particular the running battle between those who believe in efficient market theory and those who doubt this hypothesis, having experienced some of the more capricious and irrational convulsions of financial markets. If we accept that humans are not always entirely rational, this may give us some comfort in accepting the difficulties in analysing future financial trends but, unfortunately, it also opens up a potentially huge problem with the purely input-based approach to risk management. After all, what is the point of spending a fortune on sophisticated and complex risk management regimes if they are flawed because staff and managers misunderstand or misuse their results?
Skewed views
We tend to presume that risk assessment is black and white, yet the way in which we view and interpret risk data is quite subtle and prone to biased judgement. One striking aspect of the new risk research is the observation that we often make choices that fail to maximise our own long-term interests.
This can be illustrated by our attitude to lotteries (the ever hopeful punter in the UK lottery game has less than a one-in-13,000,000 chance of winning!) when contrasted with our hope that cancer always happens to others and not ourselves (here the odds are a rather chilling one-in-three risk of mortality from the disease).
In short, we tend to overemphasise low probabilities and understate high probabilities. As a result we have a skewed view of risk that depends on our reference point. This so-called “framing” effect is one of the primary reasons why we misuse risk output data. In some favourable circumstances, we are often risk averse (we frequently view risk as “bad”) and yet, in situations where we are already facing a loss, we often throw caution to the wind. To illustrate this skew in our behaviour, consider the following example:
Imagine you had to choose between the following risk decisions, options A and B: In option A you can have a sure gain of £240. In option B you have a 25% chance of gaining £1000 and a 75% chance of gaining nothing.
Which would you choose? The expected return of B is greater than that of A by about £10 but the uncertainty of option B means that in tests most people plump for option A – in effect, they are willing to pay a risk premium of £10 for the certainty of the outcome. When offered the choice of two gains, it seems we prefer the certainty of option A rather than the risky option B. In this scenario, we are risk averse. Now consider two further options:
In option C you are faced with a sure loss of £750. In option D you have a 75% chance of losing £1000 and a 25% chance of losing nothing.
In this case, if people were to decide on expected outcomes (both £750), we would expect to see an even split among the test group between the options. However, in tests the vast majority of people opt for option D. This shows that when the two options are phrased as losses most people prefer the risky option rather than the certain loss. Interestingly, the majority of us will adopt a higher risk strategy when we are in this loss domain.
In the first part of the example, the majority of people fail to maximise their risk opportunities – and display a degree of risk aversion that appears commonsense. It would be interesting to see how low option A has to be before people select option B: would they still pick A if it was only £200, for example? In the second set of options, we can see the close parallels with rogue traders who gamble ever more recklessly to try to recover their losses rather than close out the exposure and crystallise the loss.
Best behaviour
How we view and interpret risk outputs is a vital part of efficient risk decision making. In the future, financial institutions will have to pay closer attention to a number of factors that are barely discussed in most organisations, let alone fully understood. As well as the problems of framing and risk aversion, many poor risk decisions are caused by behavioural factors, such as status quo bias and sunk cost bias. Both of these, among a number of other such factors, influence our risk judgement.
Status quo bias, as the name implies, often pulls our risk decisions towards a more risk-averse outcome. Understandably, business managers who are closely monitored and held accountable for risk decisions are often overly cautious in taking risk and adopt the safety first approach of leaving things unchanged. At the overall company level, this often leads to poor and inefficient returns for the business.
The issue of sunk cost bias is potentially one of the most destructive: how do you stop a project or investment spinning out of control? At what point do managers stop throwing good money after bad? The answer is an uncomfortable one: we are likely to become more risk tolerant as we slip into loss situations. The temptation to spend more money or double up the exposure on a trade or investment to get out of the present problem is often easy to rationalise. Then we hit the dreaded point at which we think: “We might as well carry on given that we have spent so much already”.
Connoisseurs of this effect will recall the UK’s Millennium Dome project in London as an example par excellence. In banking, this behaviour often occurs in lending, with the difficult issue of correctly assessing when to demand repayment.
Understanding these issues, and developing practical solutions to counter them, will become increasingly important as banks begin to see the advantages of efficiently judging risk outputs to further reduce unexpected losses, and gain more efficient use of capital. Also prospect-theory-based tools can be used to measure and assess reward structures for staff, and provide more accurate capital allocation and attribution regimes. The actions of regulators, for example capital adequacy and operational risk rules, and the rules on credit ratings are not immune from these considerations.
The financial world will increasingly focus on risk output data and develop tools that manage them rationally, and accurately de-bias management risk pre-conceptions and judgements. The revolution is only just starting.
Gerald Ashley is managing director of St Mawgan & Co Limited, a London-based risk management consulting firm specialising in behavioural finance, decision making and prospect risk in banking and finance
