Copycats and Contrarians

by Michelle Baddeley

  What drives speculators to herd together in this way? At first glance, episodes of speculative herding seem to overturn two fundamental and related assumptions that form the backbone of mainstream economic and financial theory. Economists call the first assumption the rational expectations hypothesis.12 Like Homo economicus, which we introduced in chapter 1, economists assume that people generally, and financial traders specifically, are clever and rational. In deciding if they want to buy an asset, they must first decide what it’s worth. They must form, as accurately as possible, an expectation of the asset’s value in the future – if they wanted to sell it in a few years’ time, for example. This expectation should reflect the fundamental value of the asset – what the asset would be worth if a person held on to it forever. We can illustrate the concept of fundamental value with some examples. For a homeowner, the fundamental value of a house, if they rent it out, would be all the rent it would earn its owner over its lifetime, or the rent its owner would save if they decided to live in it. For a stock or share in a company, whether listed on a stock exchange in London, New York, Riyadh or Shanghai, the fundamental value is all the dividends the stock or share would earn for as long as the company was listed on the stock exchange, and these dividends will track the profits of the listed company over time. According to mainstream financial theory, when traders form these expectations of what an asset will be worth in the future, these will track the asset’s fundamental value.

  In capturing the behaviour of speculators, the rational expectations hypothesis complements a second assumption from mainstream economics and finance: the efficient markets hypothesis.13 This is about how the price of a financial asset – whether it is a stock, share or tulip bulb – changes over time as new information arrives. This links to the idea that, if financial markets are working efficiently, then changes in an asset’s price should reflect all information, including the latest news. Share prices will fluctuate in tandem with news, good and bad, about the likely future performance of the underlying company. Fluctuations in BP’s share price after the Deepwater Horizon oil spill in April 1990 illustrate the way in which share prices can change following bad news, reflecting speculators’ adjusting of their expectations of future profits. Various problems with the construction of the Deepwater Horizon oil well in the Gulf of Mexico led to a blowout in the wellhead, spilling millions of barrels of oil into the ocean – with catastrophic consequences for the environment, wildlife and local businesses. As soon as news of the spill broke, speculators quickly guessed that BP’s future profits were likely to be significantly eroded by compensation claims and so rapidly sold their BP shares: by June 2010, BP’s share price had collapsed by over 50 per cent.

  Economists also assume that speculators are acting independently of others, both in their use of information – social learning is precluded – and by looking after their own self-interest. These highly rational agents do not make systematic mistakes, and they efficiently use all the information they come across. In this sort of world, traders will trade away any difference between an asset’s fundamental value and its market price. For example, if traders perceive that the fundamental value of BP shares has fallen but the market price is still relatively high, then they will sell their BP shares. Then the forces of supply and demand kick in. With lots of traders selling the shares and not many wanting to buy them, the market price will fall until it matches the fundamental value. So, profits will not persist for any length of time.

  The problem with the efficient markets hypothesis and its sister rational expectations hypothesis is that they both embed extreme assumptions about markets and people. Economists know well that markets only work smoothly and fluidly when there are no market failures – but key market failures, including imperfect information and uncertainty, are endemic in financial markets. How can an ordinary person know everything they need to know about the value of the assets they buy, especially in a world plagued by uncertainty? People struggle to predict how the price of petrol might change in a day, let alone how the price of exotic, esoteric assets might fluctuate over time.

  Episodes like Tulipmania illustrate that it is not as easy to be clever as mainstream economic theory suggests. That does not mean, however, that there are no good reasons to follow herds of other speculators. If you had found yourself in the middle of the Tulipmania bubble, your best strategy would have been quickly to follow other speculators into the tulip market, but make sure that you quickly followed them out of the market too. The herding heuristics introduced in chapter 3 play an important role in guiding these speculators’ buying and selling choices. As we explored earlier, we use herding heuristics as a form of fast thinking. Heuristics enable us to decide quickly, without having to explore thoroughly all the potential sources of information. Instead, we employ our herding heuristics by copying what someone else is doing, assuming they have done the research already and know all that we need to know. The problem with herding heuristics in financial markets is that those markets are far from simple interactions between small numbers of people. Particularly in the modern, globalised and complex financial system, herding heuristics can trigger systemic crises that spread through financial systems and into macroeconomies more widely, as the 2007/08 US subprime mortgage crisis amply illustrates. Money is liquid and easy to trade and so errors are quickly copied and magnified. To learn more about this we can return to the theories of John Maynard Keynes.

  Keynes on speculators

  Keynes had a range of useful insights about speculative traders. Some foreshadowed economists’ explanations for herding.14 Others focused more on sociopsychological influences: Keynes was a pioneer in analysing the social forces driving financial markets and the macroeconomy. He focused particularly on the role of conventions in trading behaviour. In times of uncertainty, social conventions encourage speculators to believe what others believe and to do what others do. The manifestation of this is that speculators imitate others and follow the crowd.15 Keynes did not argue, however, that social conventions are irrational. From his early A Treatise on Probability of 1921 through to his major masterwork The General Theory of Employment, Interest and Money, Keynes’ view was that conventions are a useful tool that helps us to judge the probabilities of various alternative options. In an uncertain world, expectations about asset prices are volatile because no-one knows what to expect next. Amid this confusion, the conventional opinions we share with others provide an (albeit often unstable) anchor for beliefs, calming our anxieties.16 Keynes’ speculators are chasing short-term profits and making money by quickly buying and selling financial assets. They are focused on the price they can get for the assets they are selling over the day, the week or the month – or even the millisecond, given innovations enabling high-frequency trading today. If speculators are operating in a world where they might have to sell quickly, it makes sense for them to pay very close attention to what everyone else is thinking because they may have to sell to someone else within a short period of time. So they follow conventions and scrutinise others’ actions before deciding what to do themselves.

  Herding and social learning

  Delving deeper into his analysis, Keynes focused on three main reasons why financial investors are so preoccupied with what everyone else is doing and thinking: social learning, reputation and beauty contests. In financial markets, imitation determines whether or not we buy a financial asset and how much we are prepared to pay. We buy these assets not necessarily because we know much about their potential, but because we see others buying them and assume they know something we don’t. People follow the crowd because they think that the rest of the crowd is better informed. Keynes postulated that the same process operates in financial markets. In times of uncertainty, speculators realise that they are ignorant and respond by imitating other speculators. Speculators use social information about what other speculators are buying to guide their own choices, and this tendency intensifies when information is poor and uncertainty is endemic.17 Our de
cision to sell is partly driven by what we hear in the news and partly what we can see the rest of the herd doing. This links to the Bayesian social learning models of self-interested herding introduced in chapter 1. When social information overwhelms our private information, we will join a herd of copycats all choosing the same option. In this Bayesian process, speculators are using sophisticated logic. The difference in Keynes’ analysis is that he focuses more on the social and psychological motivations and less on the application of mathematical tools.

  There are individual differences in susceptibility to these informational influences. One example is the distinctive strategies adopted by professional as opposed to amateur speculators.18 Amateur speculators are more inclined to imitate, but as they acquire more knowledge and private information, they become less dependent on the social signals conveyed in others’ choices. Professional speculators are less likely to follow the crowd because they have a larger stock of private information and expertise. Another example is the small minority of the players in financial markets who ignore social influences almost entirely, making their money out of what seem to many other speculators like excessively risky maverick trading strategies. Famous investors George Soros and Warren Buffett, for instance, have made large fortunes from their distinctive investment strategies. So, speculators are not always copycats. Occasionally a small number of speculators may have the expertise and skills to make their fortunes from contrarian financial investment strategies.

  The economist Richard Topol has constructed a general model that captures this range of speculator behaviours – from imitation driven purely by what others are doing through to the completely independent decision-making associated with the mainstream models. Topol does this by setting out a model in which speculators decide what they are prepared to pay for an asset by balancing the information they have about other traders’ valuations. They have two sets of information: first, what they believe themselves is the right price for an asset, and second, the prices that other traders are willing to pay or accept when they are buying or selling. How speculators weight these different pieces of information will change depending on how confident they are in their own judgements. When copycat speculators have little confidence in their own judgements about the price of an asset, they will focus on how much other speculators are paying. They will assign a zero weight to their own beliefs. Herding will overwhelm their private judgements – in much the same way as social information overwhelms private information in the Bayesian social learning models. At the other extreme, when contrarian speculators ignore all the others then they are effectively assigning a zero weight to other speculators’ prices and focusing entirely on their own judgements. Topol’s model then reverts to the mainstream model in which rational, independent speculators form their judgements independently and do not worry about what the herds of speculators around them are doing.19 In this way, Topol covers the range – from the standard economic model, based around the assumptions of rational expectations and efficient financial markets, through to the pure herding models in which speculators are completely preoccupied with what other speculators think.

  Reputation

  As we have already seen, preserving reputation is another reason for people to copy others. John Maynard Keynes made the astute observation that it is better to be conventionally wrong than unconventionally right. This can explain conventions in financial markets: a trader who loses £1m when his peers are also losing £1m will probably keep his job. A trader who loses £1m while others are losing nothing will almost certainly be fired.

  Keynes’ insight has made its way into modern economic theory, for example in the analysis of the decisions of managers of investment funds – these are the funds invested in portfolios of different financial products. The job of the investment fund managers is to convince their customers that they are investing wisely. Sometimes a fund manager will lose money because markets are inherently unpredictable and not because they made poor decisions. Then their mistakes are only mistakes from the perspective of hindsight. Given this unpredictability, fund managers will therefore rely for their reputations on comparisons with their peers, via a process of benchmarking against other analysts operating in similar markets. Benchmarking and peer comparison lead traders to focus on a different set of goals and incentives. They are being encouraged to compare themselves to others, and this leads them to follow others and disregard their own private information, even if it is more reliable.20

  Economists David Scharfstein and Jeremy Stein use these insights to analyse herding in financial fund managers’ decisions, and they explain financial herding as the outcome of reputation-building.21 In selling their products, investment fund managers have to work hard to convince investors to invest with them. The problem is that potential investors are often more worried about short-term performance than long-term performance. But fluctuating financial markets mean short-term performance is not necessarily a good indicator of skill. Financial markets can exhibit upward momentum in asset prices in the short term, and so just because a fund manager has bought into that rising momentum it does not mean that they have a genuine and unique talent for delivering further gains in the future. Also, if their potential clients are not professionals, and are relatively ignorant, then fund managers may have no clear incentive to worry about complex performance indicators that their clients cannot understand anyway. Instead, they rely for their business on building their reputations and comparing well against their peers. In this, others’ recommendations, whether disseminated via word of mouth or social media, will be a powerful influence on investment managers’ ability to attract and retain their customers.

  Beauty contests

  Financial herding is also driven by speculators’ attempts to second-guess what others are thinking. When we are deciding what we are prepared to pay for an asset, especially if we intend to sell it quickly, what other people are willing to pay for it is a good anchor for our own judgement about what we should pay. Others’ willingness to pay will determine the price we might be able to achieve if we are selling the asset ourselves. Keynes described this phenomenon using the metaphor of a beauty contest.22 He imagined a newspaper competition in which readers are asked to look at some photos of women and then judge not who they personally think is prettiest, but who they think other readers think is prettiest. Keynes argued that a similar process describes financial speculation: speculators buy stocks and shares at seemingly exorbitant prices not because they independently believe that these assets are really worth that much, but because they believe other speculators are prepared to pay similar prices.

  Speculators’ preoccupation with others’ opinions has a reasonable basis. Ultimately, speculators are in the business of buying and selling assets to make a profit. They are also trading in fast-moving, highly liquid markets and they want to be able to sell very quickly, so they need to be able to match the price expectations of other traders around them. Speculators cannot afford to wait too long to find someone whose ideas about the fundamental value of an asset match their own. So, the individual speculator decides that their own convictions and judgements are largely irrelevant. For them, it is more important to know how much others are prepared to pay. How much do others think others are prepared to pay? How much do others think others think others are prepared to pay? How much do others think others think others think others are prepared to pay? And so on and so on. Keynes argued that, with everyone worrying about what everyone thinks everyone else is thinking, financial markets are not founded strongly on people’s careful assessment of the likely prospects of different assets. In fast-moving financial markets, carefully assessing the facts determining the fundamental value of an asset does not help speculators to make money. Predicting what others think might.

  Modern economists have adapted Keynes’ metaphor in their theories of iterated reasoning. We form our beliefs about a collective judgement, for example about the price of a share, by iterating from one person to the nex
t. For example: imagine I try to predict what Abu thinks a share is worth, while Abu is trying to figure out what Bob thinks it’s worth. Bob is trying to figure out what Chandra thinks it’s worth, and Chandra is trying to figure out what Des thinks it’s worth, and so on. As for me, I have to figure out what Abu thinks Bob thinks Chandra thinks Des thinks it’s worth. A lot of cognitive effort is required to figure out what the crowd, as a whole, thinks about the value of a share. We might judge (sensibly) that it’s not worth making all that cognitive effort when we can just copy the next person by paying what they pay. More importantly, if no-one else is thinking very deeply about the problem, then it is pointless for us to think deeply about it. We will do much better if we just copy the herd.

  Experiments based on beauty contests in financial settings have confirmed that many people are not very good at reasoning far into these iterative thinking problems. Some of these experiments analysed decisions by CEOs and other readers of the Financial Times, audiences we might expect to have a relatively sophisticated knowledge of finance. Even the CEOs did not reason deeply about the beauty contest game.23 For those who did try to reason through the example given above, most of them got as far as worrying about what Des was thinking, and then stopped trying to second-guess any further. Their failure to think beyond Des was not necessarily because they were not capable of reasoning more deeply. They may have made the strategic choice not to think too deeply because they guessed that others wouldn’t get very far with it either. Their best guess just needs to match the next person’s.

  The problem is that a world in which everyone is worrying about what everyone else is thinking is a breeding ground for financial instability, and this was one of Keynes’ fundamental points. It is important to emphasise that this is not a stupid strategy for each individual speculator. If a speculator just wants to make money quickly, then it makes sense for them to focus on what the herd is doing and paying – from their own perspective at least. From a collective, social or macroeconomic perspective, however, when this preoccupation with what others think is aggregated across many individuals interacting within complex financial systems, financial markets transform into incubators for financial disaster. No single individual has any incentive to figure out what assets are really likely to generate in real terms in anything beyond the very near future. If no-one is worrying what an asset is likely to deliver in real terms, then there is no guarantee that money will flow towards the most productive and efficient businesses and projects. As Keynes observed: