The occurrence of stock market bubbles and crashes is often cited as evidence against the efficient market hypothesis. It is argued that new information is rarely, if ever, capable of explaining the sudden and dramatic share price movements observed during bubbles and crashes. Samuelson (1998) distinguished between micro efficiency and macro efficiency. Samuelson took the view that major stock markets are micro efficient in the sense that stocks are (nearly) correctly priced relative to each other, whereas the stock markets are macro inefficient.
Macro inefficiency means that prices, at the aggregate level, can deviate from fair values over time. Jung and Shiller (2002) concurred with Samuelson’s view and suggested that waves of over- and undervaluation occur for the aggregate market over time. Stock markets are seen as having some predictability in the aggregate and over the long runBubbles and crashes have a history that goes back at least to the seventeenth century (MacKay 1852). Some writers have suggested that bubbles show common characteristics.
Band (1989) said that market tops exhibited the following features: 1. Prices have risen dramatically. 2. Widespread rejection of the conventional methods of share valuation, and the emergence of new ‘theories’ to explain why share prices should be much higher than the conventional methods would indicate. 3. Proliferation of investment schemes offering very high returns very quickly. 4. Intense, and temporarily successful, speculation by uninformed investors. 5. Popular enthusiasm for leveraged (geared) investments. 6. Selling by corporate insiders, and other long-term investors.
Extremely high trading volume in shares. Kindleberger (1989) and Kindleberger and Aliber (2005) argued that most bubbles and crashes have common characteristics. Bubbles feature large and rapid price increases, which result in share prices rising to unrealistically high levels. Bubbles typically begin with a justifiable rise in stock prices. The justification may be a technological advance, or a general rise in prosperity. Examples of technological advance stimulating share price rises might include the development of the automobile and radio in the 1920s and the emergence of the Internet in the late 1990s.
Examples of increasing prosperity leading to price rises could be the United States,Western Europe, and Japan in the 1980s. Cassidy (2002) suggested that this initial stage is characterised by a new idea or product causing changes in expectations about the future. Early investors in companies involved with the innovation make very high returns, which attract the attention of others. The rise in share prices, if substantial and prolonged, leads to members of the public believing that prices will continue to rise.
People who do not normally invest begin to buy shares in the belief that prices will continue to rise. More and more people, typically people who have no knowledge of financial markets, buy shares. This pushes up prices even further. There is euphoria and manic buying. This causes further price rises. There is a self-fulfilling prophecy wherein the belief that prices will rise brings about the rise, since it leads to buying. People with no knowledge of investment often believe that if share prices have risen recently, those prices will continue to rise in the future.
Cassidy (2002) divides this process into a boom stage and a euphoria stage. In the boom stage share price rises generate media interest, which spreads the excitement across a wider audience. Even the professionals working for institutional investors become involved. In the euphoria stage investment principles, and even common sense, are discarded. Conventional wisdom is rejected in favour of the view that it is ‘all different this time’. Prices lose touch with reality. One assumption of the efficient market hypothesis is that investors are rational.
This does not require all investors to be rational, but it does require that the rational investors outweigh the irrational ones. However there are times when irrational investors are dominant. A possible cause of market overreaction is the tendency of some investors (often small investors) to follow the market. Such investors believe that recent stock price movements are indicators of future price movements. In other words they extrapolate price movements. They buy when prices have been rising and thereby tend to push prices to unrealistically high levels.
They sell when prices have been falling and thereby drive prices to excessively low levels. There are times when such naive investors outweigh those that invest on the basis of fundamental analysis of the intrinsic value of the shares. Such irrational investors help to generate bubbles and crashes in stock markets. Some professional investors may also participate on the basis of the greater fool theory. The greater fool theory states that it does not matter if the price paid is higher than the fundamental value, so long as someone (the greater fool) will be prepared to pay an even higher price.
The theory of rational bubbles suggests that investors weigh the probability of further rises against the probability of falls. So it may be rational for an investor to buy shares, knowing that they are overvalued, if the probability-weighted expectation of gain exceeds the probability-weighted expectation of loss. Montier (2002) offers Keynes’s (1936) beauty contest as an explanation of stock market bubbles. The first level of the contest is to choose the stocks that you believe to offer the best prospects. The second level is to choose stocks that you believe others will see as offering the best prospects.
A third level is to choose the stocks that you believe that others will expect the average investor to select. A fourth stage might involve choosing stocks that you believe that others will expect the average investor to see as most popular amongst investors. In other words, the beauty contest view sees investors as indulging in levels of second-guessing other investors. Even if every investor believes that a stock market crash is coming they may not sell stocks. They may even continue to buy. They may plan to sell just before others sell.
In this way they expect to maximise their profits from the rising market. The result is that markets continue to rise beyond what the vast majority of investors would consider to be the values consistent with economic fundamentals. It is interesting to note that Shiller’s survey following the 1987 crash (Shiller 1987) found that 84% of institutional investors and 72% of private investors said that they had believed that the market was overpriced just before the crash. Shiller suggested that people did not realise how many others shared their views that the market was overpriced.
As Hirshleifer (2001) points out, people have a tendency to conform to the judgements and behaviours of others. People may follow others without any apparent reason. Such behaviour results in a form of herding, which helps to explain the development of bubbles and crashes. If there is a uniformity of view concerning the direction of a market, the result is likely to be a movement of the market in that direction. Furthermore, the herd may stampede. Shiller (2000) said that the meaning of herd behaviour is that investors tend to do as other investors do.
They imitate the behaviour of others and disregard their own information. Brown (1999) examined the effect of noise traders (non-professionals with no special information) on the volatility of the prices of closed-end funds (investment trusts). A shift in sentiment entailed these investors moving together and an increase in price volatility resulted. Walter and Weber (2006) found herding to be present among managers of mutual funds. Walter and Weber (2006) distinguished between intentional and unintentional herding. Intentional herding was seen as arising from attempts to copy others.
Unintentional herding emerges as a result of investors analysing the same information in the same way. Intentional herding could develop as a consequence of poor availability of information. Investors might copy the behaviour of others in the belief that those others have traded on the basis of information. When copying others in the belief that they are acting on information becomes widespread, there is an informational cascade. Another possible cause of intentional herding arises as a consequence of career risk. If a fund manager loses money whilst others make money, that fund manager’s job may be in jeopardy.
If a fund manager loses money whilst others lose money, there is more job security. So it can be in the fund manager’s interests to do as others do (this is sometimes referred to as the reputational reason for herding). Since fund managers are often evaluated in relation to benchmarks based on the average performance of fund managers, or based on stock indices, there could be an incentive to copy others since that would prevent substantial underperformance relative to the benchmark. Walter and Weber (2006) found positive feedback trading by mutual fund managers.
In other words the managers bought stocks following price rises and sold following falls. If such momentum trading is common, it could be a cause of unintentional herding. Investors do the same thing because they are following the same strategy. It can be difficult to know whether observed herding is intentional or unintentional. Hwang and Salmon (2006) investigated herding in the sense that investors, following the performance of the market as a whole, buy or sell simultaneously. Investigating in the United States, the UK, and South Korea they found that herding increases with market sentiment.
They found that herding occurs to a greater extent when investor expectations are relatively homogeneous. Herding is strongest when there is confidence about the direction in which the market is heading. Herding appeared to be persistent and slow moving. This is consistent with the observation that some bubbles have taken years to develop. Kirman (1991) suggests that investors may not necessarily base decisions on their own views about investments but upon what they see as the majority view. The majority being followed are not necessarily well-informed rational investors.
The investors that are followed may be uninformed and subject to psychological biases that render their behaviour irrational (from the perspective of economists). Rational investors may even focus on predicting the behaviour of irrational investors rather than trying to ascertain fundamental value (this may explain the popularity of technical analysis among market professionals). There are theories of the diffusion of information based on models of epidemics. In such models there are ‘carriers’ who meet ‘susceptibles’ (Shiller 1989).
Stock market (and property market) bubbles and crashes are likened to the spread of epidemics. There is evidence that ideas can remain dormant for long periods and then be triggered by an apparently trivial event. Face-to-face communication appears to be dominant, but the media also plays a role. Cassidy (2002) suggested that people want to become players in an ongoing drama in which ownership of stocks gives them a sense of being part of a social movement. People invest because they do not want to be left out of the exciting developments.
The media are an integral part of market events because they want to attract viewers and readers. Generally, significant market events occur only if there is similar thinking among large groups of people, and the news media are vehicles for the spreading of ideas. The news media are attracted to financial markets because there is a persistent flow of news in the form of daily price changes and company reports. The media seek interesting news. The media can be fundamental propagators of speculative price movements through their efforts to make news interesting (Shiller 2000).
They may try to enhance interest by attaching news stories to stock price movements, thereby focusing greater attention on the movements. The media are also prone to focus attention on particular stories for long periods. Shiller refers to this as an ‘attention cascade’. Attention cascades can contribute to stock market bubbles and crashes. Davis (2006) confirmed the role of the media in the development of extreme market movements. The media were found to exaggerate market responses to news, and to magnify irrational market expectations.
At times of market crisis the media can push trading activity to extremes. The media can trigger and reinforce opinions. It has been suggested that memes may play a part in the process by which ideas spread (Lynch 2001). Memes are contagious ideas. It has been suggested that the success of a meme depends upon three critical factors: transmissivity, receptivity, and longevity. Transmissivity is the amount of dissemination from those with the idea. Receptivity concerns how believable, or acceptable, the idea is. Longevity relates to how long investors keep the idea in mind.
Smith (1991) put forward the view that bubbles and crashes seem to have their origin in social influences. Social influence may mean following a leader, reacting simultaneously and identically with other investors in response to new information, or imitation of others who are either directly observed or observed indirectly through the media. Social influence appears to be strongest when an individual feels uncertain and finds no directly applicable earlier personal experience. Deutsch and Gerard (1955) distinguish between ‘normative social influence’ and ‘informational social influence’.
Normative social influence does not involve a change in perceptions or beliefs, merely conformity for the benefits of conformity. An example of normative social influence would be that of professional investment managers who copy each other on the grounds that being wrong when everyone else is wrong does not jeopardise one’s career, but being wrong when the majority get it right can result in job loss. This is a form of regret avoidance. If a bad decision were made, a result would be the pain of regret. By following the decisions of others, the risk of regret is reduced. This is safety in numbers.
There is less fear of regret when others are making the same decisions. Informational social influence entails acceptance of a group’s beliefs as providing information. For example share purchases by others delivers information that they believe that prices will rise in future. This is accepted as useful information about the stock market and leads others to buy also. This is an informational cascade; people see the actions of others as providing information and act on that information. Investors buy because they know that others are buying, and in buying provide information to other investors who buy in their turn.
Informational cascades can cause large, and economically unjustified, swings in stock market levels. Investors cease to make their own judgements based on factual information, and use the apparent information conveyed by the actions of others instead. Investment decisions based on relevant information cease, and hence the process whereby stock prices come to reflect relevant information comes to an end. Share price movements come to be disconnected from relevant information. Both of the types of social influence identified by Deutsch and Gerard (1955) can lead to positive feedback trading.
Positive feedback trading involves buying because prices have been rising and selling when prices have been falling, since price movements are seen as providing information about the views of other investors. Buying pushes prices yet higher (and thereby stimulates more buying) and selling pushes prices lower (and hence encourages more selling). Such trading behaviour contributes to stock market bubbles and crashes. People in a peer group tend to develop the same tastes, interests, and opinions (Ellison and Fudenberg 1993). Social norms emerge in relation to shared beliefs.
These social norms include beliefs about investing. The social environment of an investor influences investment decisions. This applies not only to individual investors, but also to market professionals. Fund managers are a peer group; fundamental analysts are a peer group; technical analysts are a peer group. Indeed market professionals in aggregate form a peer group. It is likely that there are times when these peer groups develop common beliefs about the direction of the stock market. Common beliefs tend to engender stock market bubbles and crashes. Welch (2000) investigated herding among investment analysts.
Herding was seen as occurring when analysts appeared to mimic the recommendations of other analysts. It was found that there was herding towards the prevailing consensus, and towards recent revisions of the forecasts of other analysts. A conclusion of the research was that in bull markets the rise in share prices would be reinforced by herding. Research on investor psychology has indicated certain features about the behaviour of uninformed investors, who are often referred to as noise traders in the academic literature. Tversky and Kahneman (1982) found that they have a tendency to overreact to news.
DeBondt (1993) found that they extrapolate trends, in other words they tend to believe that the recent direction of movement of share prices will continue. Shleifer and Summers (1990) found evidence that they become overconfident in their forecasts. This latter point is consistent with the view that bubbles and crashes are characterised by some investors forgetting that financial markets are uncertain, and coming to believe that the direction of movement of share prices can be forecast with certainty. Barberis et al. (1998) suggested that noise traders, as a result of misinterpretation of information, see patterns where there are none.
Lee (1998) mentioned that a sudden and drastic trend reversal may mean that earlier cues of a change in trend had been neglected. Clarke and Statman (1998) found that noise traders tend to follow newsletters, which in turn are prone to herding. It seems that many investors not only extrapolate price trends but also extrapolate streams of good or bad news, for example a succession of pieces of good news leads to the expectation that future news will also be good. Barberis et al. (1998) showed that shares that had experienced a succession of positive items of news tended to become overpriced.
This indicates that stock prices overreact to consistent patterns of good or bad news. Lakonishok et al. (1994) concluded that investors appeared to extrapolate the past too far into the future. There is evidence that the flow of money into institutional investment funds (such as unit trusts) has an impact on stock market movements. Evidence for a positive relationship between fund flows and subsequent stock market returns comes from Edelen and Warner (2001), Neal and Wheatley (1998), Randall et al. (2003), and Warther (1995). It has been suggested by Indro (2004) that market sentiment (an aspect of crowd psychology) plays an important role.
Indro found that poll-based measures of market sentiment were related to the size of net inflows into equity funds. It appears that improved sentiment (optimism) generates investment into institutional funds, which in turn brings about a rise in stock market prices (and vice versa for increased pessimism). If stock market rises render market sentiment more optimistic, a circular process occurs in which rising prices and improving sentiment reinforce each other. It has often been suggested that small investors have a tendency to buy when the market has risen and to sell when the market falls.
Karceski (2002) reported that between 1984 and 1996 average monthly inflows into US equity mutual funds were about eight times higher in bull markets than in bear markets. The largest inflows were found to occur after the market had moved higher and the smallest inflows followed falls. Mosebach and Najand (1999) found interrelationships between stock market rises and flows of funds into the market. Rises in the market were related to its own previous rises, indicating a momentum effect, and to previous cash inflows to the market. Cash inflows also showed momentum, and were related to previous market rises.
A high net inflow of funds increased stock market prices, and price rises increased the net inflow of funds. In other words, positive feedback trading was identified. This buy high/sell low investment strategy may be predicted by the ‘house money’ and ‘snake bite’ effects (Thaler and Johnson 1990). After making a gain people are willing to take risks with the winnings since they do not fully regard the money gained as their own (it is the ‘house money’). So people may be more willing to buy following a price rise. Conversely the ‘snake bite’ effect renders people more risk-averse following a loss.
The pain of a loss (the snake bite) can cause people to avoid the risk of more loss by selling investments seen as risky. When many investors are affected by these biases, the market as a whole may be affected. The house money effect can contribute to the emergence of a stock market bubble. The snake bite effect can contribute to a crash. The tendency to buy following a stock market rise, and to sell following a fall, can also be explained in terms of changes in attitude towards risk. Clarke and Statman (1998) reported that risk tolerance fell dramatically just after the stock market crash of 1987.
In consequence investors became less willing to invest in the stock market after the crash. MacKillop (2003) and Yao et al. (2004) found a relationship between market prices and risk tolerance. The findings were that investors became more tolerant of risk following market rises, and less risk tolerant following falls. The implication is that people are more inclined to buy shares when markets have been rising and more inclined to sell when they have been falling; behaviour which reinforces the direction of market movement. Shefrin (2000) found similar effects among financial advisers and institutional investors. Grable et al. 2004) found a positive relationship between stock market closing prices and risk tolerance. As the previous week’s closing price increased, risk tolerance increased. When the market dropped, the following week’s risk tolerance also dropped. Since risk tolerance affects the willingness of investors to buy risky assets such as shares, the relationship between market movements and risk tolerance tends to reinforce the direction of market movement. During market rises people become more inclined to buy shares, thus pushing share prices up further. After market falls investors are more likely to sell, thereby pushing the market down further.
Projection bias is high sensitivity to momentary information and feelings such that current attitudes and preferences are expected to continue into the future (Loewenstein et al. 2003). Mehra and Sah (2002) found that risk tolerance varied over time and that people behaved as if their current risk preference would persist into the future. In other words the current level of risk tolerance was subject to a projection bias such that it was expected to continue into the future. Grable et al. (2006) pointed out that this interacts with the effects of market movements on risk tolerance.
A rise in the market enhances risk tolerance, projection bias leads to a belief that current risk tolerance will persist, people buy more shares, share purchases cause price rises, price rises increase risk tolerance, and so forth. A virtuous circle of rising prices and rising risk tolerance could emerge. Conversely there could be a vicious circle entailing falling prices and rising risk-aversion. The Role of Social Mood People transmit moods to one another when interacting socially. People not only receive information and opinions in the process of social interaction, they also receive moods and emotions.
Moods and emotions interact with cognitive processes when people make decisions. There are times when such feelings can be particularly important, such as in periods of uncertainty and when the decision is very complex. The moods and emotions may be unrelated to a decision, but nonetheless affect the decision. The general level of optimism or pessimism in society will influence individuals and their decisions, including their financial decisions There is a distinction between emotions and moods. Emotions are often short term and tend to be related to a particular person, object, or situation.
Moods are free-floating and not attached to something specific. A mood is a general state of mind and can persist for long periods. Mood may have no particular causal stimulus and have no particular target. Positive mood is accompanied by emotions such as optimism, happiness, and hope. These feelings can become extreme and result in euphoria. Negative mood is associated with emotions such as fear, pessimism, and antagonism. Nofsinger (2005a) suggested that social mood is quickly reflected in the stock market, such that the stock market becomes an indicator of social mood.
Prechter (1999, 2001), in proposing a socionomics hypothesis, argued that moods cause financial market trends and contribute to a tendency for investors to act in a concerted manner and to exhibit herding behaviour. Many psychologists would argue that actions are driven by what people think, which is heavily influenced by how they feel. How people feel is partly determined by their interactions with others. Prechter’s socionomic hypothesis suggests that human interactions spread moods and emotions. When moods and emotions become widely shared, the resulting feelings of optimism or pessimism cause uniformity in financial decision-making.
This amounts to herding and has impacts on financial markets at the aggregate level. Slovic et al. (2002) proposed an affect heuristic. Affect refers to feelings, which are subtle and of which people may be unaware. Impressions and feelings based on affect are often easier bases for decision-making than an objective evaluation, particularly when the decision is complex. Since the use of affect in decision-making is a form of short cut, it could be regarded as a heuristic. Loewenstein et al. (2001) showed how emotions interact with cognitive thought processes and how at times the emotional process can dominate cognitive processes.
Forgas (1995) took the view that the role of emotions increased as the complexity and uncertainty facing the decision-maker increased. Information can spread through society in a number of ways: books, magazines, newspapers, television, radio, the Internet, and personal contact. Nofsinger suggests that personal contact is particularly important since it readily conveys mood and emotion as well as information. Interpersonal contact is important to the propagation of social mood. Such contact results in shared mood as well as shared information.
Prechter suggested that economic expansions and equity bull markets are associated with positive feelings such as optimism and enthusiasm whereas economic recessions and bear markets correspond to an increase in negative emotions like pessimism, fear, and anger. During a stock market uptrend society and investors are characterised by feelings of calmness and contentment, at the market top they are happy and enthusiastic, during the market downturn the feelings are ones of sadness and insecurity, whilst the market bottom is associated with feelings of anger, hostility, and tension.
Dreman (2001) suggested that at the peaks and troughs of social mood, characterised by manias and panics, psychological influences play the biggest role in the decisions of investment analysts and fund managers. Forecasts will be the most positive at the peak of social mood and most negative at the troughs. Psychological influences can contaminate rational decision-making, and may be dominant at the extreme highs and lows of social mood. At the extremes of social mood the traditional techniques of investment analysis might be rejected by many as being no longer applicable in the new era.
Shiller (1984) took the view that stock prices are likely to be particularly vulnerable to social mood because there is no generally accepted approach to stock pricing; different analysts use different models in different ways. The potential influence of social mood is even greater among non-professionals who have little, or no, understanding of pricing models and financial analysis. Nofsinger (2005a) saw the link to be so strong that stock market prices could be used as a measure of social mood. Peaks and troughs of social mood are characterised by emotional decision-making rather than rational evaluation.
Cognitive evaluations indicating that stocks are overpriced are dominated by a mood of optimism. Support for one’s downplaying of rational evaluation receives support from the fact that others downplay rational evaluation. The optimism of others validates one’s own optimism. It is often argued that the normal methods of evaluation are no longer applicable in the new era. Fisher and Statman (2002) surveyed investors during the high-tech bubble of the late 1990s and found that although many investors believed stocks to be overpriced, they expected prices to continue rising.
Eventually social mood passes its peak and cognitive rationality comes to dominate social mood. Investors sell and prices fall. If social mood continues to fall, the result could be a crash in which stock prices fall too far. The situation is then characterised by an unjustified level of pessimism, and investors sell shares even when they are already underpriced. Investors’ sales drive prices down further and increase the degree of underpricing. Fisher and Statman (2000) provided evidence that stock market movements affect sentiment.
A vicious circle could develop in which falling sentiment causes prices to fall and declining prices lower sentiment. Taffler and Tuckett (2002) provided a psychoanalytic perspective on the technology stock bubble and crash of the late 1990s and early 2000s, and in so doing gave a description of investor behaviour totally at odds with the efficient markets view of rational decision-making based on all relevant information. They made it clear that people do not share a common perception of reality; instead everyone has their own psychic reality.