Risk: From Russian roulette to risk revolution
Risk is not inherently predictable – as the downturn has shown. But human responses are. Nick Bullman investigates how new models are trying to balance the measurable economic data against the human tendency toward exuberance on the one side and trauma on the other.
If there is one lesson that should be learnt from the Great Financial Crash of 2008, it is that our understanding of risk is deeply flawed. The levels of debt incurred during the 90s and this decade were, in effect, like loading a gun; but one with five bullets in the chamber. Risk had changed and yet bankers, regulators, governments and individuals continued to play the equivalent of financial Russian roulette. The question is why. To answer that question we need to understand the very nature of risk; how it works, what our human perception of risk is and most importantly to develop an understanding of when we are being paid to take risk and when not.
Up until 2008 the Value at Risk model (VaR) was the central model for managing risk from banking to fund management, in many ways this model is at the centre of the Great Financial Crash. VaR encouraged its users to believe they could control risk, it allowed them to leverage their bets with a supposed 98% certainty of outcome. The problems with VaR, however, are numerous. Tail risks, the 2% outliers that should only occur infrequently, actually occur with greater frequency than the system predicts, and disastrously tend to be of large scale, greatly increasing potential for loss.
These so-called “fat tails” may be either predictable but unexpected, or “black swans” which implies that they are both unpredictable and unexpected. David Einhorn the hedge fund manager likens the VaR model to buying a car with all the safety equipment, knowing full well that in an accident none of it will work.
VaR encourages greater risk taking and dismisses the tail events as being unlikely at best. It treats risk as being both predictable and measurable with mathematical precision. VaR assumes that life is not random, and that history may be projected into the future to create a reasonable risk model.
In large part, VaR encouraged bankers to leverage their balance sheets to ridiculous levels, regulators to turn a blind eye, and governments to sit back and bask in the economic sunshine. Very few people saw that VaR had two fatal flaws. Namely; no common sense input, and secondly, no rate of change analysis. And it was not just the bankers who got the bug; individuals began to feel comfortable with outrageous levels of debt too. Behavioural finance can offer some reasons as to why it became comfortable to play the equivalent of financial Russian roulette.
The diagram shows a financial asset bubble. For a bubble to occur there needs to be some kind of displacement, a new technology like the internet, and easy credit. In the case of the Irish property boom it was the massive EU economic stimulus and initially Anglo Irish Bank.
At an early stage, an investor may still buy the asset for less than it is truly worth and this gives them a margin of safety. In other words they are being paid to take risk. As the asset price goes up, that margin of safety is eroded, eventually the price exceeds the true value of the asset and from this point onward risk increases dramatically. The blue line depicts our human perception of risk; that is how we feel about owning the asset. At the left hand side of the chart the blue line shows a great deal of anxiety. We are assuming we have come off a previous bubble and investors have been burnt. So at a time when investors should be taking risk they are traumatised. As prices rise they begin to feel more comfortable, they are making money again and so their perception of risk declines. Human perception of risk continues to decline because investors see their wealth increasing. We just want to believe the story no matter what the risks, until the bubble bursts.
New risk systems focus on understanding the relationship between perceived and actual market risks. This creates a “risk heat map” which crucially helps identify if the investor is being paid to take risk, or the inverse. Other research has thrown light on the fact that risk tends to cluster. It is as if risk begets more risk until it is finally released from the system. Like a seismologist measuring pre-earthquake tremors, a risk firm can measure the rate of change of the risk environment. This will not tell you exactly when a financial correction will occur but does give plenty of warning that pressures are building or being released in the system. Critically it will tell you when risk is at dangerous levels. The diagram shows how risk clusters for relatively long periods.
Rate of change analysis has a major advantage where risk factors are concerned; it grounds the risk system into the present. Modern risk systems are not projecting the distant past into the future. That is akin to driving down a motorway looking in the rear view mirror. The best models look at risk in the now. They focus on what is changing and balances economic data against human exuberance.
Risk is not inherently predictable; human responses are. Understanding the bridges of interconnectivity in our modern world unveils chains of risk, and risk relationships that cannot be identified by older systems. Most of us are excellent risk managers in our daily lives, we negotiate a complex and dangerous world; but when it comes to business and investment we are often let down by our lack of risk understanding. A new and open approach to risk is needed. Going forward one can either continue to play Russian roulette or be part of a risk revolution.
The author, Nick Bullman has been in financial services for 29 years, and has worked at investment banks such as Goldman Sachs. Nick is a non-executive director of Zurich Bank and Dunbar Bank, both subsidiaries of Zurich Financial Services. He is a regular commentator on Newstalk, and BBC Radio 4, with regard to the financial crisis. He is also a regular contributor to the Daily Telegraph and other financial reviews.