In these troubled times it is worth taking a careful look at how we deal with risk in our businesses. It can be seen that virtually nobody predicted the kind of economic upheaval which would befall the world once Lehman Brothers was allowed to fail. How can business better deal with risk, and how well prepared is your company?

As baseball coach Yogi Berra said: "It's tough to make predictions, especially about the future." I would recommend reading two books by former derivatives trader Nassim Taleb: Fooled by Randomness and The Black Swan. While Taleb's writing style can be a little irritating, the points that he makes seem quite profound. Essentially, he believes that rare, catastrophic events occur a lot more often than conventional economic models, such as used in banking risk models, predict. He made considerable money in the 1987 stock market crash, and his 2006 prediction of the dire consequences of a bank failure in today's interlocked financial systems suggests his ideas are worthy of consideration.

The interesting thing is that he does not pretend to have any idea when the next major crisis will come, but merely that such major events will happen a lot more often than conventional models allow. I have developed risk analysis tables for projects and even whole companies, yet how often do these lists of risk get translated into action rather than merely taking up Appendix A of the business plan? Not very often.

While working at Shell in the 1990s, I was introduced to the notion of scenario planning. Every year the analysts would predict where the oil price was going, and very serious investment decisions were made based on these projections. However, if you took a look back at the historic five-year projections and compared them with reality, it was pretty shocking: there was essentially no correlation, and lest you think the Shell analysts were unusually dense, many years ago even The Economist predicted a $5 oil price at pretty much the exact bottom of the market in March 1999.

This dilemma led Shell to develop scenario plans. The idea was that rather than sticking to a fixed price prediction, alternative plans would be developed to deal with a range of scenarios. This way, Shell would be more flexible in its response to sudden lurches in the oil price. Although I was unconvinced as to how deeply the scenario planning approach really permeated the business, the basic logic seems sensible.

Most business plans merrily project the future in a linear fashion, extrapolating from today with some modest growth. While I developed a fairly detailed business plan for my previous company, I kept thinking that it was, to an extent, a comfort blanket, since there were so many variables that the projection could easily be swamped by unpredictability.

For this reason I set out three separate scenarios, from conservative to optimistic, and used these when planning our activities. But even this proved problematic: in 2001 our sales projections were quite ambitious, and despite the March 2001 hi-tech meltdown our sales were actually well ahead of even the rosiest of the three scenarios. Yet in 2002, though we knew there was a difficult market, the reality was worse than the most conservative scenario.

As someone who has developed a few computer models over the years, I tend towards a deep scepticism when the subject matter being modelled is inherently complex. As human beings, we tend to crave certainty and patterns, where in reality there may be high volatility. Rather than putting more effort into models that will almost certainly not be accurate, it would be better to consider in detail what we have to do in our businesses to create maximum flexibility for a wide range of scenarios. For example, a practical step could be to invest as little as possible in fixed, capital projects and make heavy use of services from outside firms, which can be easily cut back or ramped up if things do not go as expected. A culture where change is welcomed rather than resisted will help.

This does not help much with making decisions on long-term projects like infrastructure, yet, in many cases I have seen, companies take steps that embed in-house resources rather than admittedly costlier external ones, in order to save money in predictable times. This may be the case, but in a volatile world it means that there is a price to be paid: it is much harder to shed or recruit permanent staff than it is to increase or decrease the service level of a third-party contract. There may be a price to pay for flexibility, but making our businesses more robust to rapid change in whatever way possible seems a better investment than building ever more complex risk management models that appear doomed to fail us.