Source: Economist

Academic view: Underestimating risk

Michael Rockinger, a professor of finance at HEC Lausanne in Switzerland, argues that business schools undervalue the teaching of risk management

RISK is essential in business. Without it there would be no innovation. In the words of Peter Drucker, a management guru, “defending yesterday…is far more risky than making tomorrow.” Yet, risky behaviour is also at the root of the world’s current economic woes. So, given the impact risk can have on our standard of living, why are we not more worried about it?

MBA students are taught how a company should protect itself against financial risks such as exchange rate variations, increases in the price of key commodities or an unexpected rise in interest rates. But risk is multi-facetted, not just something to do with finance. In the operational management class, for example, students discover risks of a completely different type—that late delivery of materials may slow a production line, or that a creaky online payment system will not only lead to financial loss but also a loss of customers’ confidence. These are so-called operational risks. Meanwhile, reputational risk—something that is increasingly important to large corporations—is usually covered in marketing lectures. Yet this is also founded in both operations and finance.

What is needed is a broader view of risk. It should be treated as a discipline in its own right rather than as an adjunct to other areas of business. Some business schools already take this multi-dimensional approach. The first step is to teach the right quantitative skills based on econometric modelling, such as extreme value theory. This is used in hard sciences to model such things as weather patterns or sea level changes, but can be adapted to help with production management.

However, such analytical tools alone are not enough to mitigate risk. Consider the various elements that make up a bank’s risk management strategy. It relies on a robust data reporting system and state-of-the-art methodology to calculate the fair value of financial products. Where a financial product is traded the fair value should be the actual market price. Where the product is not traded, the price has to be set using a mathematical model. But there are several holes in a purely quantitative approach to measuring risk. Modelling the risk of a single time series—that is, one variable over a single sequence of data points—is relatively simple. What is difficult is the aggregation of risks.

In a bid to provide more accurate modelling and aggregate risk, mathematicians have in recent decades developed the concept of copulae—calculations that link the dependence between univariate models. But these become cumbersome if more than three or four univariate models need combining. When this happens, it is difficult to predict real life situations—which is exactly what precipitated the crash of 2007. Before then, markets were relatively independent. Yet suddenly they all acted together in the same way. This highlighted the limitations of the methodology. It was made worse because there was blind faith in the model’s effectiveness, so people stopped being vigilant.

Another problem is how we determine risk. So-called “value at risk” is today’s favoured measure, but it is not a very good one. It indicates the minimal loss we can expect over a certain period of time for a given likelihood. But it does not assess the dangers of a set of combined risks. A better, but less popular measure is the expected shortfall—the average loss incurred if things do indeed go bad. Other risk measures are also available, but there is currently no consensus on which is the best one, which makes quantitative risk management a rather inexact science.

Rogue control

Students must also receive training in ethics and responsible behaviour. The importance of ethics—otherwise known as plain old common sense—is obvious when you think about that classic operational risk, a rogue trader. As long ago as 1965, Citibank faced significant losses in its Brussels branch because of a risk-loving trader. In 1995, unauthorised trading by Nick Leeson brought about the collapse of Barings. More recently, Jérôme Kerviel at Société Générale, Kweku Adoboli at UBS and Bruno Iksil at JP Morgan have all become famous for the billions in losses they incurred as a result of their activities.

These traders have two crucial things in common: they were all supposed to take essentially riskless positions and yet they were able to hide huge unauthorised trades because of an astonishing lack of supervision. For example, if anyone at Barings had questioned Mr Leeson’s requests for cash to pay margins, his subterfuge would have been uncovered. At Société Générale, a high level of managerial turnover meant that Mr Kerviel’s positions went uninspected. And in the case of Citibank, it was only when an experienced trader uncovered inconsistencies after talking with the speculator that an investigation was launched.

So to be a player in modern risk-management education, a business school needs to recognise the importance of these human factors and provide training in a number of different, but complementary, areas. Teaching econometric modelling on its own isn’t enough. In fact, one of the most important messages to convey is also one of the simplest: talk to colleagues about what they are doing, provide clear limits and make it clear that there are sanctions if those limits are broken. This, after all, is a lesson that any good parent could teach.