RADICAL UNCERTAINTY: THE PURPOSE OF FINANCIAL MARKETS - The End of Alchemy: Money, Banking, and the Future of the Global Economy - Mervyn King

The End of Alchemy: Money, Banking, and the Future of the Global Economy - Mervyn King (2017)

Chapter 4. RADICAL UNCERTAINTY: THE PURPOSE OF FINANCIAL MARKETS

And what you do not know is the only thing you know

And what you own is what you do not own

And where you are is where you are not.

T.S. Eliot, ‘East Coker’, The Four Quartets

‘You’ve got to expect the unexpected.’

Paul Lambert, Aston Villa manager, press conference, 22 November 2013

Are we really capable of expecting the unexpected?1 In 1998, the hedge fund Long-Term Capital Management (LTCM) failed, although its senior management team comprised two Nobel Laureates in Economic Science, Myron Scholes and Robert Merton, and an experienced practitioner in financial markets, John Meriwether. Their strategy, successful at first, was to create a highly leveraged fund that bought large amounts of one asset and sold equally large amounts of a slightly different asset (for example, government bonds of slightly different maturities), so as to exploit anomalies in the pricing of those assets. The return on each transaction was tiny but done on a sufficiently large scale, it generated huge profits.

The choice of assets was based on sophisticated statistical analysis of high-frequency data over a decade or more. But the data, although voluminous, covered only a short period of history, and when a rare but significant event - the Russian default and devaluation in the summer of 1998 - occurred, past correlations proved a poor guide to asset returns. LTCM failed. Its management argued, in hindsight almost certainly correctly, that given sufficient time they would be able to work their way through the losses to a position of positive net worth. But if you start with a portfolio worth less than nothing, you have to persuade your creditors to allow you to continue. As usually happens in such circumstances, the creditors called a halt. The lesson is that no amount of sophisticated statistical analysis is a match for the historical experience that ‘stuff happens’. At the heart of modern macroeconomics is the same illusion that uncertainty can be confined to the mathematical manipulation of known probabilities. To understand and weather booms and slumps requires a different approach to thinking about uncertainty.

The illusion of certainty

Risk, luck, fate, uncertainty, probability theory - we all have names for the game of chance. Most decisions in life involve risk. Sometimes we embrace it, as when we enjoy a bet on the Grand National or the Super Bowl, and sometimes we avoid it, as when we insure our houses against fire. The playing of the hand we are dealt can be a pleasure in a game of bridge and a burden in life. We accept that Lady Luck has her part to play in our personal lives. But we cling to the ‘illusion of certainty’ in monetary matters. There is a seemingly insatiable demand for economic forecasts. Newspapers and television are only too willing to print the latest forecast of, say, national income with a degree of precision that beggars belief and far exceeds the ability of statisticians to measure it. And at the end of each year prizes are awarded to the forecasters who turned out to be the most accurate. It makes as much sense as it would to award the Fields Medal in mathematics to the winner of the National Lottery.

No economic forecaster has ever been able to match Edmund Halley, who in 1682 made calculations predicting that the comet then visible in the skies would return seventy-six years later. It did - on Christmas Day 1758. Fortunately the length of the economic cycle - the duration of the expansion and subsequent contraction of the economy before it returns to its normal levels of output and employment - is shorter than the periodicity of Halley’s Comet - although if it goes on increasing at its present rate even that might not be true. But Halley was able to rely on scientific laws; economic predictions are inherently less reliable because they depend upon human behaviour.

Despite the repeated inability of economic forecasting models to predict accurately, there is a persistent belief that there is, if only we could find it, a ‘model’ of the economy that will produce forecasts that are exactly right. When giving evidence to the Treasury Select Committee in the House of Commons, I would sometimes respond to questions by saying, ‘I don’t know, I don’t have a crystal ball.’ Such an answer outraged many Members of Parliament. They thought it was my job to have an official crystal ball in order to tell them what the future held. Any attempt to explain that not only could I not forecast the future, but neither could they, and nor for that matter could anyone else, was regarded with disbelief. Down the ages, quack doctors selling patent medicines and astrologers selling predictions have been in strong demand. Added to their number today are economists selling forecasts, reflecting a desire for certainty that is as irrational as it is understandable.

Why are we so reluctant to accept that the future is outside our control? The reluctance to give adequate prominence to risks may reflect the fact that many of us feel uncomfortable with formal statements of probabilities. Probability theory is a relatively recent development in our intellectual history, dating back to a flowering of ideas in Europe around 1660 produced by Blaise Pascal, Gottfried Leibniz, Christiaan Huygens and others. Despite advances since then, statistical thinking remains prone to confusion and is often avoided. Television weather forecasts in Britain rarely employ the language of probabilities used by the meteorologists themselves. Professor Gerd Gigerenzer, a psychologist and Director of the Max Planck Institute for Human Development in Berlin, who studies the mental processes that actually underlie decisions in practice, has demonstrated in a series of studies how poorly doctors, lawyers and other professionals understand probabilities.2 At the start of the crisis in August 2007, the Chief Financial Officer of Goldman Sachs, David Viniar, said that the losses suffered on one of their hedge funds implied that ‘we were seeing things that were 25-standard deviation moves, several days in a row’.3 That certainly is extreme, since such moves should occur even less often than once every 13 billion years, or the time elapsed since the creation of the universe!

In times of genuine uncertainty, even the most hard-bitten financiers become disorientated. And despite Seneca’s maxim that ‘luck never made a man wise’, airport bookshops continue to stock titles on how to become rich written by successful investors and entrepreneurs who are confident that their success is the result of outstanding business acumen rather than good fortune. Matthew Syed, a former table tennis international, argued that sporting success reflected practice more than talent - the result of ‘hitting a million balls’.4 His thesis was widely hailed, in part because it gave us back the feeling that we could be in control of our destiny. Clearly, practice is crucial to success, but as Ed Smith, the writer and former England cricketer, explained, chance, or simply luck, plays a big role in both sporting and personal life - ‘stuff happens’.5 The desire for certainty and control over our destiny is a deep-seated human characteristic.

The difficulty we have in confronting uncertainty, and our strong desire to control our own lives, lead to seemingly irrational decisions. After the terrorist attacks on New York on 11 September 2001, many Americans stopped flying for a period and drove instead. Traffic on interstate highways rose 5 per cent in the three months after the attack, and it took a year before normal patterns of travel were resumed. In that period, around 1600 Americans lost their lives in road accidents because of the switch from flying to driving, some 50 per cent of the death toll incurred on 9/11 itself.6 Such behaviour might appear irrational. After the attacks, airline security was drastically tightened. But how were people to assess the risk of flying in a world of new uncertainties? They opted for a form of transport more directly under their own control, even if it turned out to be more dangerous.

Coping with unquantifiable uncertainties outside our control is a challenge to our mental discipline. We are tempted to put blind faith in experts who claim certainty. We rely on extrapolations of the past. If house prices have risen each year for a long period, it seems natural to conclude that they will go on rising.7 Such beliefs can fuel a continuing rise in prices until some external event confronts those beliefs with reality. Much statistical analysis in economics - the use of econometrics - relies on the assumption that past correlations will continue to hold in the future because the underlying ‘model’ generating observations of economic data remains unchanged. But if the model is wrong, observed correlations will prove a poor guide to the future, as LTCM discovered. Neither house prices nor any other asset price are likely to rise indefinitely, relative to a measure of incomes. Failure to understand the context in which correlations are observed leads to false conclusions. The steady fall in long-term real interest rates since 1990 was always likely to lead to a continuing rise in house and other asset prices. But it was never plausible that such a fall could continue indefinitely. Similarly, before the crisis, banks appeared to be well capitalised and had little trouble attracting funds, until one day they couldn’t. It is difficult to cope with the complexity of the world, and so we fall prey to the illusion of certainty. An example from the stock market illustrates the problem.

The stock market is volatile and difficult, if not impossible, to predict over short periods. At the beginning of any particular week the chance of the market rising over the following week is roughly the same as the chance of its falling. So if I were to predict the direction of the market movement correctly for five successive weeks, you might think that I knew something you didn’t. Indeed, you might be willing to subscribe to an investment service with that sort of track record. How might one create the illusion of clairvoyance? Select around six thousand names and addresses from the London or New York telephone directory. Divide the names into two groups. To the first group, send a letter predicting that the market will rise over the coming week. To the second, write predicting a fall in the market. At the end of the week keep the three thousand or so names who were given the correct prediction and discard the others. Divide those names in turn into two groups. To the first, predict a rise in the market and to the second, a fall. Repeat this process for five weeks, at which point there will be around 200 people to whom the following letter could be sent: ‘You may well have been sceptical when you received our first letter, but by now you will know that we have indeed found the secret of predicting successfully the direction of movement of the stock market. You know that our method really works. To subscribe to our investment service please send £5000 by return.’

My publisher has insisted that I make clear that I am not encouraging any reader to set up such a scheme. But the example illustrates that the interpretation of ex post outcomes depends critically on understanding the context from which the observations were drawn. In complicated situations that may require imagination of a high order. The wish for certainty and the belief that it exists are seductive and dangerous. The desire to resolve the cause of apparently inexplicable events means that people can easily be misled into believing a false story by a failure to appreciate the context from which the events they observe are drawn. Coping with uncertainty is by no means straightforward, even for the most highly trained professionals. As Voltaire put it, ‘Doubt is not a pleasant condition, but certainty is an absurd one.’8

The two types of uncertainty

In coming to terms with an unknowable future, it is helpful to use the distinction between risk and uncertainty introduced in 1921 by the American economist Frank Knight.9 Risk concerns events, like your house catching fire, where it is possible to define precisely the nature of the future outcome and to assign a probability to the occurrence of that event based on past experience. With risk it is then possible to write contracts that can be defined in terms of observable outcomes and to make judgements about how much we would pay to take out insurance against that event. Many random events take the form of risk, and that is why there is a large industry supplying insurance against fire, theft, accidents and death. Uncertainty, by contrast, concerns events where it is not possible to define, or even imagine, all possible future outcomes, and to which probabilities cannot therefore be assigned. Such eventualities are uninsurable, and many unpredictable events take this form. A capitalist economy generates previously unimaginable ideas, new products and new technologies. For example, a friend decides to open a software business and asks you to be an investor. It may be difficult to assess the value of the software product and impossible to assign probabilities to its success or failure.

The distinction between risk and uncertainty can be illustrated by human mortality. ‘In this world nothing can be said to be certain, except death and taxes,’ wrote Benjamin Franklin in 1789. It is clear, however, that there is indeed substantial uncertainty about both the likely date of our death and the contributions or taxes required to pay for our pensions. Longevity risk - the probabilities of dying at different ages - can be assessed by looking at the experience of others. In England and Wales in 2012 the most common age of death for women was eighty-seven, but of course most died at different ages. Because those frequencies of death at different ages are observable, individual risk - that we will die either earlier or later than the average for our peers - can be insured by taking out life insurance against early death or by purchasing an annuity to insure against later death.

There is, however, also genuine or ‘radical’ uncertainty about the average length of life of people belonging to different generations. Average longevity has increased over time, with a remarkable reduction in infant mortality during the twentieth century and a more recent fall in mortality at later ages. A woman who was sixty in 1902, and subject to that year’s mortality rates, would have expected to live for another fourteen and a half years. By 2012 that expectation had increased to over twenty-five years. Changes in average longevity have proved hard to predict. In 1798, the English cleric and scholar Thomas Malthus wrote that ‘with regard to the duration of human life, there does not appear to have existed, from the earliest ages of the world, to the present moment, the smallest permanent symptom, or indication, of increasing prolongation’.10 That past experience was to prove a poor predictor of the future. In 1798, life expectancy in Britain was around forty. Today it is over eighty, and even higher for women. We simply do not know how life expectancy will change in the future.11 Developments in medical science, especially the results of stem-cell research, may enhance the prospects for life expectancy radically, and new infectious diseases may have the opposite effect. Good judgement rather than statistical extrapolation is key to making assessments about changes not only in longevity but in many economic and social variables.

Economists typically think about risk rather than radical uncertainty. They see the future as a game of chance in which we know all the outcomes that might emerge and the odds of each of them, even though we cannot predict the roll of the dice. In that world, because all future outcomes can be defined, it is theoretically possible to hold the grand economic auction described in Chapter 2, leading to efficient decisions about what to produce and consume.12Although such an auction would, of course, be impossible to organise in practice, the real failure of the auction model is more profound. If we cannot imagine the goods and services that may exist in the future, nor conceive of all the eventualities that may befall us, then it is impossible to define the markets that are required by the auction model. Radical uncertainty drives a gaping hole through the idea of complete and competitive markets. Even if the markets that do exist are competitive, many crucial markets for future goods and services are absent. When IBM launched its personal computer (the ‘PC’) in 1981, there were no markets in the products that subsequently displaced it in the consumer marketplace, such as laptop and tablet devices. Neither producers nor consumers can know what options will be available to them in the future, and so they cannot express preferences in markets that might provide a guide to investment decisions. The markets are simply missing. And how tedious it would be if we could imagine what the future holds. Uncertainty - radical uncertainty - is the spice of life.

Coping strategies as rational behaviour under uncertainty

In a world of pure risk, where we can list possible future events and attach probabilities to them, there is a traditional view among economists of what constitutes rational behaviour - the so-called ‘optimising’ model. According to this view, individuals first evaluate each possible future outcome in terms of its impact on their well-being or ‘utility’, and then weight each utility by the probability of the event to which it is attached, so deriving the average or ‘expected utility’ from a given set of actions. People are assumed to choose their actions (for example, how much to save today) in order to reach the highest level of ‘expected utility’.13 Such optimising behaviour in a world of risk has proved a useful tool in analysing the impact of government interventions in markets and the provision of insurance against known risks. Choosing between different models of car, how many hours to work, how much to pay for insurance against an identifiable event - all such decisions, and there are many of them, can be analysed perfectly well within the conventional framework of economic risk and hence the traditional economists’ optimising framework. As the imperial power of the social sciences, economics has extended its reach to theories of marriage patterns, divorce and childbearing. It is striking, however, that such economic analysis is largely concerned with areas of economic choice where risk rather than uncertainty is the norm.14

The main defence of the theory of optimising behaviour is the one provided by Milton Friedman in 1953, when he compared people making economic decisions with billiards or snooker players who do not understand Newtonian mechanics, but play as if they did.15 This ‘as if’ argument has been powerful in persuading economists that people behave as if they carried out immensely complex mathematical calculations - rather as if they were computers playing chess. But we know that when computers play chess, they do so differently from human beings. The former make millions of calculations; the latter make intuitive leaps of imagination. And that should be a warning of the limitations of the economic calculus underlying traditional economic theories. Human capacity for making conscious calculations is bounded, but the ability of the human brain to engage in lateral thinking is well developed. People are better than computers at recognising faces. Most of my economist colleagues have had their deepest insights through the use of intuition, and have deployed logical mathematical proofs to demonstrate to others why that intuition is correct. But the original insight did not come from making a mathematical calculation.

And it appears that cricketers and baseball players are the same, only more so. If, like calculating billiards players, fielders followed the dictates of Newtonian mechanics, they would observe the ball hit high in the air, compute where it would land, run in a straight line, stand still and catch the ball. Careful video observations have revealed that neither cricketers nor baseball players do that.16 Rather, they seem to follow a simple rule of thumb: watch the ball and keep the angle between it and the horizon constant.17 It can be shown that this rule of thumb means that fielders will run, not in a straight line but in an arc, and will end up catching the ball while still running. That pattern of behaviour is exactly what the videos have shown. Traditional optimising behaviour is replaced by a rule of thumb that is simple to follow and robust to the complexities of swirling winds and atmospheric resistance of the ball, a practice sustained by experience.

More generally, in a world of radical uncertainty, where it is not possible to compute the ‘expected utility’ of an action, there is no such thing as optimising behaviour. The fundamental point about radical uncertainty is that if we don’t know what the future might hold, we don’t know, and there is no point pretending otherwise. Right through his life, John Maynard Keynes was convinced that radical uncertainty, as it has become known, was the driving force behind the behaviour of a capitalist economy. As he explained, drawing on Knight’s distinction between the two types of uncertainty, there is an essential difference between a game of roulette or predicting the weather, on the one hand, and the prospect of war or the scope of new inventions, on the other. Of the latter, he wrote: ‘About these matters there is no scientific basis on which to form any calculable probability whatever. We simply do not know.’18

The language of optimisation is seductive. But humans do not optimise; they cope. They respond and adapt to new surroundings, new stimuli and new challenges. The concept of coping behaviour does not, however, mean that people are irrational. On the contrary, coping is an entirely rational response to the recognition that the world is uncertain. There is no need to abandon the conventional assumptions of economists that people prefer more consumption, or profit, to less, and that their choices display a degree of consistency.19 The strength of economics as a social science is the belief that people will attempt to behave rationally. The challenge is to work out how a rational person might cope with radical uncertainty. People aren’t dumb. It is just that in a world of radical uncertainty even smart people do not find it easy to know what it means to behave in a smart manner.

The main challenge to the economists’ assumption of optimising behaviour comes from ‘behavioural economics’, a relatively new field often associated with Daniel Kahneman, Richard Thaler and Amos Tversky.20 It studies the emotional and psychological dimensions of economic choices.21 Behavioural economics has identified an impressive array of cognitive biases in the way people behave in practice. For example, people are observed both to display overconfidence in their ability to judge probabilities and to underestimate the likelihood of rare events. But behavioural economics assumes that deviations from traditional optimising behaviour result from the fact that humans are hardwired to behave in a way that is ‘irrational’. Daniel Kahneman suggested that decisions are made by two different systems in the mind: one fast and intuitive, the other slower, deliberate, and closer to optimising behaviour.22 In this way he was able to explain aspects of behaviour that appear anomalous in the traditional approach.

But simply patching up the optimising model by making the decision process more complicated - adding an intuitive to a rational self - in order to explain particular observed anomalies does not mean that it is likely to perform better in explaining future behaviour. The gold standard of scientific tests is prediction. Consider the problem of trying to explain the movement of the stock market over the past year. By taking into account more variables and more detail from that year, we would be able to ‘explain’ a larger proportion of the movement. But much of that would reflect the accidental quirks of the past year rather than any underlying structure of the stock market. A complicated explanation of the past makes it no more likely that we can predict the stock market over the coming year than simply tossing a coin. Many smart people have tried and failed to beat the market. Information is quickly incorporated into stock prices, and explaining why prices moved in the past is no basis on which to predict the future. The stock market is a good example of the tendency of economists ‘to excel in hindsight (fitting) but fail in foresight (prediction)’.23

The danger in the assumption of behavioural economics that people are intrinsically irrational is that it leads to the view that governments should intervene to correct ‘biases’ in individual decisions or to ‘nudge’ them towards optimal outcomes. But why do we feel able to classify behaviour as irrational? Are policy-makers more rational than the voters whose behaviour they wish to modify? I prefer to assume that neither group is stupid but that both are struggling to cope with a challenging environment. After the crisis, the earlier belief that competitive markets were efficient and yielded rational valuations of assets was replaced by a conviction that financial markets were not merely inefficient but reflected irrational behaviour that produced ‘bubbles’ in asset prices and excessive demand for credit. Both views are extreme. Of course, emotions play an important part in economic decisions, especially in financial markets. Professor David Tuckett of University College, London, interviewed a large number of investment managers in different financial centres to discover what motivated them when making their decisions about how to invest large sums of money. Out of this came a theory of ‘emotional finance’. Rather than viewing unconscious emotions and conscious reasoning as two systems in conflict, Tuckett sees them as engaging in a continuous two-way communication. As he argues, ‘emotion exists to help economic human actors when reason alone is insufficient’.24 In other words, emotions help us to cope with an unknowable future and should not be seen as ‘irrational’.

The problem with behavioural economics is that it does not confront the deep question of what it means to be rational when the assumptions of the traditional optimising model fail to hold. Individuals are not compelled to be driven by impulses, but nor are they living in a world for which there is a single optimising solution to each problem. If we do not know how the world works, there is no unique right answer, only a problem of coping with the unknown. A different way of thinking about behaviour as neither irrational nor the product of a constrained optimisation problem is, I believe, helpful in understanding what happened both before and after the crisis. In other words, we need an alternative to both optimising behaviour and behavioural economics.

What does it mean to be rational in a world of radical uncertainty? Once we are liberated from the view that there is a single optimising solution, rules of thumb - technically known as heuristics - are better seen as rational ways to cope with an unknowable future.25 A heuristic is a decision rule that deliberately ignores information. It does so not just because humans are not computers, but because it is rational to ignore information when we do not understand how the world works. As is clear from the example of trying to explain and then predict the stock market, getting lost in the thickets of the past conceals the big picture. Ignoring information is rational when it is likely to be of little help in solving the problem we confront - sometimes less is more. Heuristics are not deviations from the true optimal solution but essential parts of a toolkit to cope with the unknown.

An eye surgeon of my acquaintance told me how his patients responded when he explained to them, as was his duty, the risks of a surgical procedure. Half of the patients had already made up their mind before entering the room, based on their attitude towards risk, irrespective of the probabilities of different outcomes. And the other half made up their mind by listening carefully to the doctor and then making the decision on the basis not of reported probabilities but on their own judgement of the character and personality of the man in front of them. In the words of Frank Knight, ‘The ultimate logic, or psychology, of these deliberations is obscure, a part of the scientifically unfathomable mystery of life and mind. We must simply fall back upon a “capacity” in the intelligent animal to form more or less correct judgements about things, an intuitive sense of values.’26 A coping strategy is a way of capturing that unfathomable mystery. Humans are not pre-programmed to solve complex mathematical optimising problems, because it is impossible to know in advance which problems they will need to solve. But they are programmed to learn and to adapt. Coping strategies are the natural, even perhaps genetic response to the need to adapt to an uncertain world. They are, in Gigerenzer’s phrase, ‘ecologically rational’; that is, they are decision processes that are well suited to the environment in which they are used.27 In that sense, in a world of radical uncertainty they are more rational than the economists’ assumption of optimising behaviour.

A coping strategy comprises three elements - a categorisation of problems into those that are amenable to optimising behaviour and those that are not; a set of rules of thumb, or heuristics, to cope with the latter class of problems; and a narrative. The set of heuristics may comprise one or several rules of thumb to deal with different problems within the class. Each heuristic is a rule for making decisions which ignores much of the information used in optimising behaviour in order to provide a quick and robust decision. It is specific to the environment in which it is used. The narrative is a story that integrates the most important pieces of information in order to provide a basis for choosing the heuristic and the motive for a decision. Narratives compel us to action, and so play a big part in decisions taken under conditions of radical uncertainty. When we cannot write down a mathematical model with numerical probabilities, we can nevertheless think and talk about the future in qualitative terms.

The heuristic must be operational and the narrative believable. Coping strategies are not universal solutions to all problems in all environments but robust and rational ways of responding to particular problems. When things go wrong, as in the crisis, the cause is not necessarily irrational behaviour, nor an external shock, but possibly a mismatch between the chosen heuristic and the environment. In Chapter 8 I shall use these ideas to describe the continuation of unsustainable levels of borrowing and spending before the crisis.

Two real-life examples from financial markets show what this abstract description of decision-making means in practice. The first concerns J.P. Morgan and its British-born banker Sir Dennis Weatherstone, who started as a bookkeeper at the age of sixteen and rose to become CEO in 1990. The challenge was how to decide which of the many new and obscure financial products suggested by the traders and mathematicians on its staff the firm should sell to its clients. With no past history for the performance of those products, there was no basis for judging which ones were likely to be effective. The new products were an example of radical uncertainty. The strategy Weatherstone employed was to make sure that any new product was understood by senior management. The narrative underlying the strategy was that if the product could be explained in a conversation among senior managers then there was less risk that something might go badly wrong. Weatherstone, I was told, would give the inventors three slots of fifteen minutes to explain the product to him. If at the end of that he still did not understand the product, the firm would not sell it.28 In 2008 there must have been many executives who wished they had followed Weatherstone’s heuristic.

Just before Barings Bank, one of the oldest banks in the world, collapsed in 1995 under the weight of losses of $1.4 billion caused by a rogue derivatives trader, Nick Leeson, in its Singapore branch, the senior managers in London told the Bank of England that they were pleased with the trading results but slightly puzzled as to how its Singapore business had earned such a large profit. A useful heuristic for managers and regulators alike is to probe not only those parts of a business that are losing a lot of money but also those that are making a lot.

The second example is the problem of how to regulate a bank. If a bank fails and is unable to meet its obligations to depositors and bondholders, the ensuing chaos may lead to a loss of confidence in other banks and disrupt the system of payments of wages and bills in the economy as a whole. To limit the risk of failure of a bank, regulators (the Federal Reserve, the Federal Deposit Insurance Corporation and the Office of the Comptroller of the Currency in the United States, the Bank of England in the United Kingdom, and the European Central Bank in the euro area) insist that a bank finances itself with a minimum amount of equity capital contributed by its shareholders. In that way, the bank has some capital that can absorb any losses that might arise, so reducing the risk of failure. The amount of equity capital the bank is required to issue - known as its ‘capital requirement’ - is related to the riskiness of the bank’s activities.

At first sight this seems eminently sensible. The riskier a bank’s assets, the more likely it is that the bank will fail unless it has sufficient equity to absorb losses. Internationally agreed standards set the capital requirement as a proportion of its ‘risk-weighted assets’, where each type of asset on the bank’s balance sheet is weighted by a measure of its riskiness.29 For example, debt issued by governments has a zero weight, meaning that banks are not required to have any equity to support that type of asset. The justification for the zero weight is that governments are assumed not to default on their debt - an assumption that might have looked reasonable when the standards were drawn up after long negotiations among many countries, but looked decidedly odd during the euro area crisis from 2012 onwards. Risk weights derived from statistical studies of the past, moreover, proved highly misleading in the crisis. For example, past data had suggested that mortgages were a relatively safe asset for banks to own, and yet in the crisis they turned out to be the source of large losses. It is extremely difficult, if not impossible, to judge how the riskiness of different assets will change in the future. The appropriate risk weights can change abruptly and suddenly, especially in a crisis, and are an example of radical uncertainty, not risk, despite the words used by regulators.

Risk-weighted capital requirements appealed to many of my international colleagues because risk was explicitly incorporated into the calculation. But if the nature of the uncertainty is unknown, then the use of such measures can be highly misleading. It is better to be roughly right than precisely wrong, and to use a simple but more robust measure of required capital. Heuristics are better than so-called optimising solutions that assume the wrong model. In the case of bank regulation, it is better to use a measure of leverage rather than a ratio of capital to risk-weighted assets. Leverage ratios measure capital relative to total (unweighted) assets. A Bank of England study of 116 large global banks during the crisis (of which 74 survived and 42 failed) found that the simple but robust leverage ratio was better at predicting which banks would fail than the more sophisticated risk-weighted measures of capital.30

The most extreme example was Northern Rock, which failed in the autumn of 2007. At the start of that year, Northern Rock had the highest ratio of capital to risk-weighted assets of any major bank in Britain, so much so that it was proposing to return capital to its shareholders because they had no need of it - under the regulations. At the same time, the bank’s leverage ratio was extraordinarily high at between 60 to 1 and 80 to 1.31 The reason for this remarkable discrepancy between the two measures of capital ratios was that the international standards assumed that mortgages were an extremely safe form of lending and Northern Rock did little else. But Northern Rock had been selling its mortgage loans to other investors rather than holding them on its own balance sheet. When the market for such transactions closed in the late summer of 2007, the bank could no longer obtain funds to finance its assets. The initial threat to Northern Rock did not come from volatility in the value of its mortgages but from the risk that its short-term creditors would withdraw their funds, leaving the bank high and dry. And that is what happened after the shock to markets in August 2007. The bank was left exposed as it tried to find alternative, and more expensive, sources of funding, which were not forthcoming to an institution with such a high leverage ratio. Keeping your eye on the leverage ratio is the banking equivalent of the rule of thumb used by baseball and cricket players. The heuristic of keeping the leverage ratio below some critical level is designed to deal with a situation where we know that we don’t know how to measure the risks facing a bank. That is where the regulators and executives of the big banks went wrong.

Financial markets and derivatives

Coping strategies are especially important in financial markets because these markets are a link between the present and the future. Radical uncertainty is the key to understanding not just money and banks but financial markets in general. Of all the actors in recent economic history, the most infamous, revered and reviled in turn, were ‘the markets’. I wish I had a pound (or even a dollar or a euro) for each time I had to listen to a politician explain, always in an infuriated tone, that ‘the markets’ didn’t understand. But markets are not people. They are an impersonal mechanism by which many different real agents - such as businesses, banks, investors, pension funds and indeed governments - interact in the buying and selling of foreign exchange, loans, stocks and bonds, and, increasingly, a bewildering variety of new financial instruments. What is the purpose of all these financial markets? They channel household savings into business investment, at home and abroad. They make it possible to share risk by giving us the opportunity to insure, hedge, or even speculate, against future events. And they provide continuous valuations - a financial running commentary - of the myriad activities that make up our economy.

Equity, debt and insurance are the basic financial contracts underpinning our economy. A corporate stock is a claim on the earnings of a company in all future eventualities and provides an uncertain stream of dividends over an indefinite future. A loan made today is a claim on repayment of principal and interest over a fixed number of years in all states, except when the borrower defaults. An insurance contract yields payments under specified contingencies at unknown dates in the future. The estimated total value of stocks and bonds around the world is somewhere between $150 and $180 trillion. The total ‘global financial stock’ of marketable instruments plus loans must be well over $200 trillion. Much of this represents the value of the finance raised by governments, households and companies to fund their expenditure, both current and capital. Finance is essential to the ability to invest in real capital assets - houses, factories, railroads, sewers and a host of private and public infrastructure. Even a cursory glance at economic history shows the importance of a banking and financial sector able to channel household savings into investment projects, to share the risks resulting from those investments, to manage our wealth, and to enable us to do mundane things such as pay our bills. The absence of a proper financial sector handicaps development and was well illustrated by the inefficiencies of centrally planned economies.

Over the past twenty years, a wide range of new and complex financial instruments has emerged, expanding dramatically the scale of financial markets. These new instruments are elaborate combinations of the more traditional debt, equity and insurance contracts, and as such they are known as ‘derivative’ instruments. They package streams of future returns on a wide variety of investments, ranging from housing to foreign exchange. They are claims on returns generated by the underlying basic financial contracts and play a valuable role of filling in gaps in markets, offering new ways both to hedge risks and speculate on future price movements of the underlying contracts, such as stock prices. Derivatives typically involve little up-front payment and are a contract between two parties to exchange a flow of returns or commodities in the future. The principle of derivative instruments is simple, but if you want to make it complicated there are many lawyers, and investment bankers who will help you - at a (significant) price.

Examples of derivative instruments include forward and futures contracts (the purchase of a commodity to be delivered at a future date), options (the right to buy, or sell, a basic contract such as a stock at a given price on or before a given date), and swaps (where two parties exchange a stream of cash flows in different currencies or for different profiles of interest payments to hedge their other exposures). Many of these instruments have real practical value. For example, the Wimbledon tennis championships receives payments in dollars for broadcast rights in the United States. But almost all the costs of running the tournament are in pounds sterling, which creates a risk from unknown future movements in the pound-dollar exchange rate. That risk can be reduced, at a price, by contracting today to sell dollars for pounds at a specific date in the future and at a particular exchange rate. Many, if not most, companies benefit from similar transactions.32

During the crisis more complicated derivative instruments, as well as bundles of underlying assets packaged up and sold as ‘securitised’ instruments, acquired a certain notoriety because the failure to understand their true nature brought down banks and even AIG, a large American insurance company. These included credit default swaps (CDS, where the seller agrees to compensate the buyer in the event of default of some named party), mortgage-backed securities (MBS, a claim on the payments made on a bundle of many hundreds of mortgages, sold to the market by the originator of the mortgages, often a bank), and collateralised debt obligations (CDO, a claim on the cash flows from a set of bonds or other assets that is divided into tranches so that the lower tranches absorb losses first and the higher last, with investors able to choose in which tranches to invest). All of these complex instruments were legitimate financial contracts to create and sell. But the buyer needed to be aware of what he or she was getting. In their New Year sale the London store Harrods used to offer socks at half price - provided you bought a fixed package of five. When you got home you would discover at least one pair you would never wear (in my case, orange socks). The set of five pairs was rather like a CDO that bundled socks instead of sub-prime mortgages - a legitimate tactic by a sharp salesman.33

Talented mathematicians were recruited by banks to invent and market even more complicated instruments that often only they really understood. And even the mathematicians did not fully appreciate that all their sophisticated calculations could take into account only observable risk and not unquantifiable uncertainty. Crucially, since derivatives are not basic contracts representing economic activity but synthetic instruments, there is no limit on the size of the exposure - and the potential losses - that can be created. The scale of exposure inherent in derivatives can be many multiples of the value of the underlying equity or debt claims. In Chapter 1, I described how, at the start of the crisis in 2007, regulators drew comfort from the fact that the size of the basic contracts in sub-prime mortgages was too small to bankrupt the banking system. But the magnitude of the derivative instruments built on sub-prime mortgages was many times greater and led to enormous exposures and losses. It is rather like watching two old men playing chess in the sun for a bet of $10, as one can in Washington Square in New York, and then realising that they are watched by a crowd of bankers who are taking bets on the result to the tune of millions of dollars. The scope for introducing risk into the system rather than sharing it around is obvious. And that is why Warren Buffett described derivatives as ‘financial weapons of mass destruction’.34

All of this was pointed out to the financial services industry by central bankers and regulators before the crisis.35 Why, then, did derivatives grow so quickly? One answer is that betting is more addictive than chess, and the trading mentality fed on itself. Derivatives also allowed a stream of expected future profits, which might or might not be realised, to be capitalised into current values and show up in trading profits, so permitting large bonuses to be paid today out of a highly uncertain future prospect. But another answer is that derivatives do have real value when used in the right way - to reduce, not create risk. Many companies and institutions want to hedge (that is, insure against) risk associated with future shifts in the prices of commodities, changes in interest and exchange rates, and other economic variables. Derivatives also create financing options that may not exist in conventional debt markets. Consider investor A, who would like to purchase a bundle of mortgages (an MBS) but can pay for it only by borrowing short term. That creates a risk that when the loan must be repaid it may be either expensive or indeed impossible to take out another loan, as happened to many borrowers in the crisis. Writing a derivative contract with an investor B, under which A pays B fixed amounts over the life of the contract and in return receives the cash flows from the mortgages, amounts to locking in the finance for the purchase of the MBS, which is then used as collateral. Derivatives provide alternative ways to borrow.36

Much of the impetus for the creation of a wide range of derivative financial instruments, including options, was the belief that by adding more and more markets a gain to society would be achieved by the effective ‘completion’ of markets in order to mimic the auction economy described in Chapter 2. In that fictional world of the grand auction, financial markets are redundant. The prices and quantities of every transaction for goods and services, including labour services, are set in the auction, and so financial assets are no more than synthetic packages of those basic components. In reality, where markets for many of those goods and services are missing, as they inevitably are in a world of radical uncertainty, financial markets play a significant role. They can - up to a point - substitute for some of those missing markets. In so doing, they have created the illusion that markets provide almost unlimited ways to cope with uncertainty.

In fact, filling in the gaps between existing markets may serve a valuable purpose, but it cannot deal with the problem of how to create a market in something we cannot imagine. Derivatives do not offer insurance against radical uncertainty. Auctions cannot be held nor contracts written on unimaginable outcomes. As Frank Knight put it, ‘it is this true uncertainty which by preventing the theoretically perfect outworking of the tendencies of competition gives the characteristic form of “enterprise” to economic organisation as a whole and accounts for the peculiar income of the entrepreneur’.37 In other words, radical uncertainty is the precondition of a capitalist economy.

Used carefully, derivatives can reduce risk. But the very complexity and obscurity of derivatives can mislead the unwary into thinking that they are hedging risks while in fact they remain exposed to great uncertainty and huge potential losses in the event of even a small change in underlying asset prices. It all brings to mind the previously untold story of what happened when financial markets arrived on a desert island I visited recently, following in the footsteps of Robinson Crusoe.

The parable of financial markets on a desert island

Sitting on his desert island, Robinson Crusoe would have been astonished to learn of the miracles of modern technology and the importance we attach today to financial markets. His descendants, however, learned the hard way. The effects of the explosion of derivative instruments in the financial sector are well captured by the sad story of what happened to that unfortunate island when it inadvertently allowed banking to overtake fishing as its principal activity.

In the beginning, fishing with rods was the only economic activity on the island. Then nets came into use, produced by specialist net-makers. For fishermen, nets were an investment and a banking system came into being to accept deposits, which were then lent to net-makers. The subsequent sale of nets enabled the loans to be repaid. All was going well on the desert island, until one day a banker had a bright idea. Instead of holding on to the loans and paying interest to depositors, he decided, the bank would package together a number of the loans made to net-makers and sell them as a new financial instrument: net-backed securities (NBSs). By selling NBSs to savers, it was possible for the banks to finance further loans and to stop worrying about whether the fishermen would catch enough fish to buy the nets, so enabling the net-makers to repay the loans. That was now the problem of the people who had bought the NBSs.

Some clever islanders with a mathematical bent realised that it was possible to go one step further. They ‘sliced and diced’ the various NBSs to create new synthetic securities that would allow investors to choose in which quality of net-maker (in the impressive language of the financial advisers, in which tranche of the returns) to invest. These collateralised debt obligations (CDOs) proved highly fashionable. So some of the clever and more mathematically inclined fishermen joined the bank. They created even more complex securities - CDO-squared and even higher orders were not uncommon. Some people were hired to act as ‘rating agencies’ to demonstrate that the securities were not as risky as might have been feared.

All this activity was rather exciting. The clever people involved in creating and trading the new securities worked out that it was possible to make bets on future fish catches without putting up much capital. As trading in these instruments proceeded, and people’s views about the size of the future fish catch changed, so the values of the new securities went up and down. By adopting the modern accounting convention of valuing the new instruments by ‘marking to market’ - that is, valuing assets at the latest observed price and including all changes in asset values as profits - optimism about the future, whether justified or not, created large recorded profits from the trading of these new securities. In effect, anticipated future profits were capitalised and turned into current profits. From those large reported profits, the clever people paid themselves large bonuses. They acquired more and more claims on the catch of fish. As the wages of people in the financial sector rose, the wages of fishermen fell. There was much concern about rising inequality. But it was explained by the result of the market - trading required high skills. This was undoubtedly true.

The financial sector grew in size, the incomes earned by those in it reached levels not seen before on the island, and there was great admiration for the talents of those who had created such a vibrant and expanding sector. Some visionary islanders even suggested that if only they could find a trading partner across the sea, then it would be sensible for the island to abandon fishing and devote itself entirely to financial activity. Ordinary fishermen felt rather left behind, but they too had to admire the ability of financiers to create such apparently profitable activity and be so successful. Even the community leaders were envious of the power that accrued to the financial sector, and there developed a close, and, in the eyes of many, none-too-healthy relationship between the island’s political leaders and the financial sector.

Then one day it all collapsed. The expansion of trading activity had reduced the supply of labour to fishing. Some people started to question whether the old man in the tree, known as the National Statistician, was right to count the profits on these trading activities - which, for a while, had more than offset the reduction in the production of fish - as GDP. As a result, a few people expressed doubts about the underlying value of many of these new financial securities. To meet their obligations, some of the banks started to sell the securities for money. That led to a downward spiral in the prices of these assets, and to concerns about the solvency of some of the banks. Markets in the securities closed, as no one was willing to take the other side. Liquidity disappeared. Banks had nothing with which to create new money. Panic set in, and the demand for liquidity soared. But the supply of liquidity fell. The banks failed and were taken into communal ownership while some of the clever people were employed to disentangle the web of complex interrelationships and contracts between the banks.

Everyone on the island felt let down. The cult of finance had led to the contraction of a successful fishing industry. Too many talented people had been sucked into trading which, with the benefit of hindsight, was little more than a zero-sum activity generating little or no output. How could so many people have been taken in by the new world of finance?

After a painful post-mortem, it was agreed that the banking system should go back to its traditional role of accepting deposits and financing loans to net-makers. More people needed to work in fishing, and to invest in the making of nets to enhance the future catch of fish. Clever people realised that their reputation would in future be enhanced by adding to the social value of production rather than diverting resources from other people’s pockets into their own. Some of them even moved out of finance into teaching. They recognised that they had a responsibility to write the history of this episode, and to convey it to future generations, in order to prevent a repetition of the near disaster.

The illusion of liquidity

Radical uncertainty also leads to another problem in financial markets - the illusion of liquidity. In the world of the grand auction, liquidity is not an issue. Prices are determined in the auction, and then, as time passes, the commitments entered into are fulfilled. Markets don’t reopen. In reality, of course, markets do reopen; financial markets, in particular, remain open almost continuously with buyers and sellers coordinating through an intermediary - a shop-keeper or ‘market-maker’ - who holds some stock to bridge the time between the arrival of a buyer and a seller. Such continuous trading is particularly important in financial markets, and I shall return to it later in the chapter. Market-makers offer the opportunity to transact immediately once a decision to buy or sell has been taken. Many financial centres boast that their markets are ‘deep and liquid’. By that they mean that investors can quickly sell their financial assets, with only a very small reduction, if any, in the price, in order to obtain money.

Liquidity is the quality of ‘immediacy’.38 For liquidity to be valuable it must be reliable. One aspect of the alchemy of financial markets is the belief that markets are always liquid. It is an illusion because the underlying assets (the physical assets and goodwill of a company, for example) are themselves usually illiquid, and liquidity depends on a continuing supply of buyers and sellers on opposite sides of the market. Radical uncertainty can disrupt that supply. Markets can be liquid one day and illiquid the next, as happened on 19 October 1987 (‘Black Monday’) when the Dow Jones Industrial Average fell 23 per cent in a day and the market-makers temporarily disappeared because they were worried about the risk of buying at one price and being able to sell only at a much lower price a short time later.

Liquidity also waxed and waned regularly during the year from September 2007 to September 2008. The liquid MBS market, for example, became highly illiquid when investors realised that house prices could go down as well as up. Their claims were now dependent on the circumstances of the individuals whose mortgages had been bundled into the MBSs rather than the value of the houses taken as collateral. Investors discovered that they knew very little about the borrowers on whose payments they depended. One MBS was no longer a good substitute for another. Liquidity dried up because it became difficult to value bundles of mortgages without knowing much more about the characteristics of the borrowers.

None of this should really have been surprising. Radical uncertainty makes it likely that from time to time there is a revision in the narrative guiding investor behaviour, or in the coping strategy as a whole, leading to sharp changes in traders’ perception of values and willingness to buy or sell financial assets. A market that was previously characterised by a steady flow of trades may see a disappearance of activity. The behaviour of the LIBOR market (in which banks would lend to each other for short periods, say three months, at the London Inter-Bank Offer Rate) in the crisis brought this home forcefully. Banks supplied to a panel daily quotes of the interest rate at which they believed they could borrow from each other. LIBOR was set by the panel broadly as an average of the quotes received.39 During the crisis, LIBOR became unstable, with large swings in the rate from day to day and from month to month. What became apparent was how few transactions there were in the interbank lending market at a number of maturities. Even well after the crisis, in 2011, there was little borrowing between banks at maturities of four months and above in sterling and above six months in dollars.40 With few or no transactions taking place, it was difficult and at times impossible for banks to know what rate to quote. Indeed, once markets realised that different banks had different risks of failure then the whole concept of a single interbank borrowing rate became meaningless. Does this matter? Yes - because LIBOR is used as a reference rate in drawing up derivative contracts worth trillions of dollars. The benchmark interest rate used in those contracts had shallow foundations and in a storm it just blew down.

Amid the smoke created by the inability to define LIBOR during the crisis, it became possible for some traders in banks to collude and submit quotes designed to benefit them directly. Regulators later discovered evidence of manipulation and submission of ‘false’ quotes, and the so-called LIBOR scandal led to the fining of firms and individuals.41 But at times in 2007-8 there were so few transactions in the interbank market that any quote submitted would have been hypothetical. An understandable response by some banks was to withdraw from the panel to which quotes were reported, only to discover that while being investigated by regulators for the submission of false quotes, they were also being told that they had to keep submitting quotes with no basis in reality. LIBOR has had its day.42 Liquidity is not a permanent feature of financial markets. Places that boast of deep and liquid markets, the financial equivalent of an infinity pool, should be aware that their depth is variable, with a long shallow end that is sometimes drained.

In the 1980s, economists debated with some passion whether the stock market was ‘rational’ or ‘irrational’. Three participants in the debate were subsequently awarded Nobel Prizes in Economic Science.43 One of them, Robert Shiller, argued that it was impossible to explain the volatility of stock prices by reference to the volatility of the dividend stream that is the return to stocks. Others argued that expectations of large surprises in the distant future, not captured in the data for dividends in any observable sample, justifies the volatility apparent in stock markets. The issue can never be resolved, for the simple reason that in a world of radical uncertainty it is impossible to know what the future holds and therefore whether or not any particular valuation is rational, only whether it seems to embody a wise or a foolish judgement. Stock prices move around because investors are trying to cope with an unknowable future. Their judgements about future profits can be highly unstable. This instability is fundamental to a capitalist economy.

The Austrian economist Joseph Schumpeter coined the phrase ‘creative destruction’ for the way a capitalist economy promotes investment in new ideas and ventures, undermining investments in earlier undertakings.44Sometimes the message from the markets provides a helpful signal to businesses about when and in what directions to invest. On other occasions, the message tells us about the psychological nervousness, or even panic, among investors. Keynes’s description of the stock market has become famous:

… professional investment may be likened to those newspaper competitions in which the competitors have to pick out the six prettiest faces from a hundred photographs, the prize being awarded to the competitor whose choice most nearly corresponds to the average preferences of the competitors as a whole; so that each competitor has to pick, not those faces which he himself finds prettiest, but those which he thinks likeliest to catch the fancy of the other competitors, all of whom are looking at the problem from the same point of view. It is not a case of choosing those which, to the best of one’s judgment, are really the prettiest, nor even those which average opinion genuinely thinks the prettiest. We have reached the third degree where we devote our intelligences to anticipating what average opinion expects the average opinion to be. And there are some, I believe, who practise the fourth, fifth and higher degrees.45

Narratives play an important role in the coping strategies of investors. Under radical uncertainty, market prices are determined not by objective fundamentals but by narratives about fundamentals.46 Those stories can be influenced by important players, such as central banks and governments, but also by changes in intellectual fashion or a realisation that the existing story is misleading - as happened in the crises of 1914, when the narrative that war was inconceivable was replaced by the narrative that it was here and would be won, and 2008, when the narrative that previous spending paths were sustainable was replaced by the narrative that they were not (see Chapter 8 for a fuller exposition of this point).

In today’s stock market, the competitors trying to second-(or third-) guess market sentiment have been replaced by computers. Over one-half of orders are driven by computer algorithms - mathematical formulae that tell the computer when to buy or sell. Because stock exchanges have made it possible for some extremely large ‘high frequency’ traders to pay for faster access to the exchange, the computers of such firms can watch the order flow and then send in their own orders microseconds ahead of other traders, so jumping the queue and getting to the market before the price turns against them.47 Such behaviour is called ‘front-running’. It imposes a ‘tax’ on the transactions of other investors who are less fleet of foot, and encourages investment in expensive technology, not to incorporate new information in market prices, but to exploit information about other people’s orders.

There is no social benefit in allowing some traders preferential access to knowledge of the overall balance between orders to buy and sell. One way of eliminating the ‘tax’ on ordinary investors would be to change the system of trading on organised exchanges to electronic auctions held once an hour, once a minute, or even once a day, depending on the nature of the stock being traded.48 Moving to auctions separated by intervals chosen to match the likely arrival of relevant news, as opposed to supposedly real-time trading in which some investors can move (very slightly) faster than others, has much to commend it. It is already the case that much trading is carried out close to the opening and closing of the trading day, as investors want to transact when there are many other buyers and sellers so that their own orders have less of an influence on price.49

By their very nature, moreover, algorithms cannot easily change their strategy in the light of new information; so far only humans can rewrite the algorithms, and that requires not raw computing power but judgement.

Under radical uncertainty, investors make judgements, perhaps based on a coping strategy, and with the benefit of hindsight these are sometimes described as ‘mistakes’. But beliefs change, and who is to know which beliefs are correct? The valuations in financial markets are for the moment. They change quickly, and sometimes violently, reflecting uncertain knowledge of the future. Investors are simply people trying to cope with an unknowable future and behave, as we all do in such situations, sometimes cautiously, sometimes erratically, but always in a fog of uncertainty.

We too can learn from the experience of Robinson Crusoe’s descendants. Finance should support, not overshadow, the real economy. Financial markets can help us to cope with an uncertain future provided we do not succumb to the danger of believing that uncertainty has been turned into calculable risk. Central to a capitalist economy is the fact that the future cannot be seen as a game of chance in which the only source of uncertainty is on which number the wheel of fortune will come to rest. The future is simply unknowable. And in a capitalist economy, money, banking and financial markets are institutions that have evolved to provide a way of coping with an unpredictable future. They are the real-world substitute for the economic theorist’s concept of a grand auction.

For that reason, a capitalist economy is inherently a monetary economy. Money has a special role. Provided that there is sufficient trust that its value will be maintained from one period to the next, it offers a means by which one can park generalised purchasing power, to be used in the future when unimaginable events occur. Money gives us the ability to exchange labour today for generalised purchasing power in the future. That is why many savings contracts are denominated in money terms. We expect to earn an interest rate defined as a percentage increase in the amount of money in our account. Money is not just a means of buying ‘stuff’ but a way of dealing with an uncertain future. A rise in the desire for a reserve of generalised future purchasing power lowers spending today, and can lead to a recession or even a depression.

The struggle to cope with radical uncertainty affects not just investors, businesses and households but also the institutions set up to deal with collective problems such as money creation. Central banks, arguably the most important such institutions, need a coping strategy too.