Risk is part of everyone's life, and, increasingly, people are demanding more control over the many risk-related decisions they face on a daily basis. A recent MORI poll revealed that 94 per cent of Britons believe the government should be more open about how it reaches its decisions. The public's perception of risk is something which cannot be ignored by any government or ruling body, and below you will find a brief selection of the results from MORI's poll entitled Public Attitudes to Risk.

The following tables illustrate the result of a survey which used 1,015 interviews conducted between 919 January 1999 in 87 separate location in Britain. Data are weighted to the known profile of the population, in terms of sex, age, working status, car in household, social background and region. An asterisk (*) represents a figure between zero and 0.5 per cent.

Which of these, if any, have you done over the last two or three years because of any risks involved?

Peter Bernstein

The dominance of rational models of human behaviour has led to the notion that some form of quantification of variables can guarantee safety against risk. The view expressed here is both more elegant and sublimely logical. Drawing inspiration from Pascal, Bernoulli and Gould, Bernstein argues that risk is not only inevitable but valuable and that a predictable world would be the most dangerous of all.

Business management is a process rooted in the future. We do not know the future. Then how in the world can we make decisions? In this article I look at this question from the viewpoint of two great philosophers of the past and one distinguished scientist of the present. The bottom line is that the consequences of the choices we face should outweigh probabilities of the outcomes we expect, but humans do have free will and can make rational choices even when the future is uncertain.

The first approach in both historical order and order of importance dates back nearly 350 years. My source is Blaise Pascal, a Parisian who was one of the most famous mathematicians and philosophers of his time. It was he, together with the lawyer and mathematical genius Robert de Fermat, who first propounded the principles of probability in 1654. Pascal spent most of his life alternating between a life of casinos and fleshpots, followed by intense religious periods marked by total abstinence from these enjoyments. In the end, the religious mode won: Pascal retired to the monastery of Port-Royal which was supported by the earnings of the first commercial bus company in Paris, of which Pascal was the founder.

Scribbled in the manuscript of his autobiography, Pascal set forth what has come to be known as le pari de Pascal, or Pascal's wager. Pascal began by daring to ask: 'God is, or he is not. Which way we should we incline? Reason cannot answer.'

We have to reconstitute the question before we can answer it. Is an outcome in which 'God is' preferable more valuable in some sense than an outcome in which 'God is not'? The issue is not belief in God, for that cannot be reduced to a decision in which you wake up one day and say, 'Today I have decided to believe (or not believe) in God.' Nor do our guesses as to the probability that God is help us very much because we have no reliable method for testing or falsifying the hypothesis. There is a more fruitful approach.

Suppose you act as though God is and therefore choose to lead a life of virtue, unselfishness, and abstinence. Suppose you are wrong and God is not. You will have passed up some fun, but there would be rewards as well. Now suppose that you act as though God is not. You hang out in the casinos and fleshpots and place 'me' above all else. And suppose you are wrong on this side of the bet. You may have had a lot of goodies in the span of your lifetime, but you will suffer damnation into eternity. Nobody wants that. The value of the bet that God is, is infinitely greater than the value of the bet that God is not. The probability that God is or God is not is irrelevant.

Now translate this line of analysis into business decisions. Suppose you are planning to launch a new product into another country where the market for this kind of product is growing vigorously. The probabilities of success appear to be high, but this market is also highly competitive. Differentiating your brand is difficult; price is the dominant variable. On the other hand, given the interest rate spreads and the exchange market's optimistic outlook for this nation's currency, you see no need to hedge the foreign currency bet at this time. Should you hedge or not hedge?

Suppose you decide to hedge and are wrong: exchange rates turn out to be stable. Although you will have incurred the cost of hedging, you will still be able to price your product correctly for this competitive market, no matter what happens to the exchange rate. Now suppose you agree with the forecast of the foreign exchange market and refuse to hedge but the currency surprises the market by depreciating. You will receive less sterling than before for each unit you sell abroad in a market where raising your local currency price even by a little will put you out of business.

You have bet that God is not, but God is. You cannot afford to make that bet. The consequences of being wrong are too serious. Regardless of what you believe the currency is going to do, regardless of the probabilities, maintaining a competitive local price for your product is the single most important element in success. Hence, you must hedge or not take the risk of entering that market.

Fast forward about 50 years to a scientist and mathematician named Bernoulli, a member of a family that was extraordinarily talented but also nasty and mean, even in their relations with one another. For example, when one family member and his son competed for the same prize from the French Academy of Sciences for work in planetary orbits, and the son won, the father was so infuriated that he threw his progeny out of the house.

In 1703, Jacob Bernoulli wrote a letter to Gottfried von Leibniz in which Bernoulli observed that: 'It is strange that we know the odds of throwing a seven instead of an eight with a pair of dice, but we do not know the probability that a man of 20 will outlive a man of 60.' Jacob proposed an experiment in which he would compare a large number of pairs of men of various ages to see if he could deduce the probabilities from that evidence.

Leibniz took a dim view of the suggestion. 'Nature', he wrote, 'has established patterns originating in the turn of events, but only for the most part.' In those days, correspondence of this type was always carried on in Latin, but Leibniz placed such importance on those last five critical words that he set them in Greek. Never forget them.

His message rings true across the centuries. You cannot escape uncertainty. No mathematical model works to perfection. Statisticians are satisfied when they can demonstrate that a model works with only a 5% probability that its results are due to chance, but that still leaves 5% that we do not understand, cannot model, and can cause all kinds of mischief if we mechanically make decisions based on the model. No event is without cause, so ascribing a probability to chance, or luck, is merely assigning a number to our ignorance.

As patterns repeat themselves only for the most part, we must use the tools of risk management the art of survival when our forecasts of the unknown future turn out to be wrong. In the example above, the currency did depreciate, even though the probabilities were against it. The other part matters every bit as much as the most part.

Now I turn to the brilliant American evolutionary biologist, Stephen Jay Gould. In his recent book, Full House: The Spread of Excellence from Plato to Darwin¹, Gould writes about the moment some 15 years ago, when the doctors informed him that he had developed a deadly form of cancer called mesothelioma. The median life expectancy for patients with this disease, they reported, was eight months. How does one react to a diagnosis like that? Give way to devastation? Carry on as though the problem did not exist? Or pray?

In a manner of speaking, Gould chose prayer. His religion is Darwinian evolution, a theoretical structure that describes a reality in which causes have effects but the variety of possible effects from any given cause is virtually limitless. Consequently, the effects are not predictable. In this religion, as Gould describes it, 'variation stands as the fundamental reality and calculated averages become abstractions.'

Taking his cue from this Darwinian keynote, Gould reasoned that it would be an error to view the forecast offered by his doctors as the most likely outcome for any single individual. 'I am not a measure of central tendency, either mean or median. I am one single human being with mesothelioma, and I want the best assessment of my own chances.' His crucial insight was to recognise that there is little room between the absolute minimum of zero life expectancy (dropping dead at the moment of diagnosis) and the median value of eight months: 'Half of the variation must be scrunched up into this left half of the curve between the minimum and the median. But the right half, in principle, extends out forever, or at least into extreme old age.'

From that deduction stemmed the good news. Gould noted that many factors favoured his potential location in the right tail he was young and full of fight, was blessed with a supportive family, lived in a city that offered outstanding medical treatment, and was fortunate in having his illness diagnosed early on. He fought and won. If he had been mesmerised by the central tendency by the median he might well have ended up as a suicide. Yet, 15 years later he is a healthy and vigorous man.

Gould goes on to make the point in a general sense: 'We are still suffering from a legacy as old as Plato & a tendency to abstract a single ideal or average as the 'essence' of a system and to devalue or ignore variation among the individuals that constitute the full population & We have never put aside this distinctive view that populations of actual individuals form a set of accidents, a collection of flawed examples, each necessarily incorrect and capable only of approaching the ideal to a certain extent & We regard variation as a pool of inconsequential happenstances, valuable largely because we can use the spread to calculate an average, which we may then regard as the best approach to an essence.'

In short, all of us depend far too much on measures of central tendency and therefore allow ourselves to be mesmerised by the hole rather than the doughnut. We base our decisions on trends, coefficients of correlation, normal distributions, even more arcane gadgets, and above all on the average.

But the sum and substance of risk management is in the recognition of variety. Losses stem, not from average results, but from deviations from the average or norm, from the outliers at the far reaches of the tails, and from outcomes never even imagined.

Surprise pursues us relentlessly because we can never have all the information we need for a correct forecast every time. But life is even harsher than that, for the incomplete information we do have is overwhelming, and getting more so all the time. We probably could not handle all the information we need, even if it were available to us!

We have no choice but to simplify. In particular, we are addicted to smoothing the data, to removing the bumps and wrinkles in order to identify the essence, the trend, the values that we would like to believe are the true values. We prefer seasonally adjusted data to raw data, we find it easier to understand year-over-year rates of change than sequential monthly or quarterly changes, and we like moving averages and trendlines. The revered discounted cash flow model itself is a smoothing tool that glides us from the present into some distant point in the future.

Accrual accounting is an especially vivid form of smoothing, a true act of faith that we know what the future holds in store. It records as revenues money not yet received, and that may not be received. It excludes from expenses money actually laid out on assets that are expected to produce cash revenues in the future, but they may fail to produce.

Hence, although every form of smoothing assumes that variations from a specified path are transitory or systematic, major risks lurk in the interstices of all those techniques. We are like people who are led to believe they are seeing moving pictures at the cinema when all they are seeing is a rapidly moving sequence of frames of still pictures. Simplification lures us into the trap that Gould warns we set for ourselves with our demand for the 'essence' in preference to the variation, for simplification is impossible without the averages and the other measures of central tendency.

We do concern ourselves with the big risks admittedly when changes are so great that the new environment seems to have had no past at all events like the OPEC shock of 1973, the abrupt arrival of disinflation during 1982, the recent far-reaching transformation in the role of government, and the all-too-frequent powerful smashings of well-established parameters such as former highs and lows in interest rates, equity valuation, or yield spreads. These are the kinds of change where all the old rules go out the window and we have to learn an entirely new game. We characterise these moments as discontinuities, paradigm shifts, or regime changes.

Gould and his fellow Darwinians would laugh at our nomenclature. They refuse to recognise discontinuities, paradigm shifts, or regime changes. Before there can be discontinuity, we must have continuity; before there can be a paradigm shift, we must have a paradigm; before there can be a regime shift, we must have a regime. We must have trends, norms, averages, and other indicators of central tendency before we can have what we perceive as shocks to the system.

But the Darwinian would argue that diversity in nature is so great and every event could lead to such a great variety of outcomes, that there is no such thing as a state of nature that we can identify as a paradigm. The trick in risk management, perhaps, is in recognising that normal is not a state of nature but a state of transition and trend is not destiny.

Yet Gould is sending us a positive message. Variety is the spice of life. Just imagine what life would be like if everything closely resembled its measure of central tendency if we were all clones of just one particular human being, or if all corporations were merged into just USA Inc or, even worse, into Globalisation Inc.

A profound and overriding point is embedded in that suggestion: life would not only be intolerably boring without variety, life would be intolerably risky! Variety is what makes risk management possible, for by definition diversity is the necessary condition for diversification and for hedging. It means that we have more than one basket in which to deposit the eggs we want to carry. As a result, society is willing to take on more risk than if we had just one basket. Indeed, variety is at the essence of survival. We do not all come down with mesothelioma at the same instant, all ships do not sink together, all houses do not burn up simultaneously, all stocks do not go up and down in lockstep together, and different capital assets have different cyclical patterns.

The philosophical issue runs even deeper. Take away variety and nothing remains but averages with zero standard deviations and central tendencies always on the mark. Take away variety, in other words, and we eliminate uncertainty. And what is so good about that? Once you eliminate uncertainty, you have crushed human free will, for then everything is predetermined and decision-making becomes an obsolescent skill.

Peter Bernstein's book, Against The Gods: The Remarkable Story of Risk, was published by John Wiley and Sons in October 1996.

REFERENCES

Gould, S. J. (1996) Full House: The Spread of Excellence from Plato to Darwin. New York: Crown Publishers.