Report on Global Financial Crisis
Discussions on psychological factors affecting People’s behaviors in the crisis and their motivations
Qiang Sheng 9th May 2011
Financial Risk Management Lecturer: Bernd P. Leudecke
Macquarie University Melbourne
4.1 Three areas of applications were reviewed and investigated: 1. The pricing of financial assets; 2. The portfolio choice and trading decisions of investors; 3. The behavior of firm managers;
4.2 A “Bubble” is an episode in which irrational thinking or a friction causes the price of an asset to rise to a level that is higher than it would be in the absence of the friction or the irrationality; and, moreover, the price level is such that a rational observer, armed with all available information, would forecast a low long-term return on the asset (Barberis, 2010).
4.3 Two categories of theories explaining “Bubble Formation” (Why an asset class might become overvalued): 1. “Investor Beliefs Based” theories; 2. “Investor Preferences Based” theories;
4.4 Three “Belief-Based” theories of “Bubble Formation” (Barberis, 2010): First theory argues that a bubble forms when investors disagree sharply about an asset’s future prospects and there are short-sale constraints. Second theory argues that bubbles arise because investors extrapolate past outcomes – returns, earnings growth, or default rates – too far into the future. Third theory is based on overconfidence – specifically, on the idea that people overestimate the precision of their forecasts (overconfidence on reliability of favorable information).
1
4.5 Two “Preference-Based” theories in “Bubble Formation”: First theory argues that investors become less risk averse because of “House Money Effect”, which in short, means having experienced gains, they are less concerned about future losses because any losses will be cushioned by the prior gains. Reduced risk aversion leads to increased asset purchase, which pushes up price even further. Second theory argues that bubbles are particularly likely to occur in stocks related to a new technology. Given that people have a strong preference for lottery-like payoffs – perhaps as Kahneman and Tversky (1979) argue, the brain overweighs low probabilities – they may overvalue these stocks (1990s U.S. Technology Stocks Bubbles).
4.6 The argument on lottery-like investments in the article was explained by Prospect Theory and the psychophysics of the probability weighting function. Prospect Theory was introduced by Kahneman and Tversky in their 1979 paper. It suggested that people do not judge outcomes based on their final asset positions that include their wealth, but the changes of wealth, an idea first introduced by Markowitz. Prospect theory uses Value Function to value outcomes, and the probability weighting function, π (p), is used to value probabilities. Contrary to the Expected Utility Theory, weighting of probabilities exhibited unlinearity (Figure 1). Y.H. Cheung, in his paper further explored and demonstrated the shape of the unlinear graph of the weighting function through Monte Carlo simulation.
2
(Figure 1) It was interesting to see that Problem 8 and 14 from Kahneman and Tversky’s paper showed people exhibiting risk-seeking with low probability positive events, whereas Prospect theory showed a general risk-aversion in positive domain and risk-seeking in negative domain. When faced with choosing prospects for specific event, Kahneman and Tversky suggested that decision weights measure the impact of events on the desirability of prospects, and not simply the perceived likelihood of these events. As mentioned above, probability weighting performed unevenly around the ‘end points’, and most of us are neither able to comprehend nor understand the extreme situations presented, hence these possible outcomes were either ignored or overweighted in our decision-making process, and then most probable outcomes will be treated as certainty. Burns, Chiu and Wu (2010) suggested that people are very sensitive to changes in probabilities closer to 0, and that induced the overweighting of events associated with that level of probabilities. However, they suggested, and Kahneman and Tversky also mentioned in their paper, that overweighting is distinct from risk seeking in gains, and it is the overweighting plus the S-shaped value function that holds the risk-seeking in low probability gains and risk-aversion in low probability losses.
3
4.7 As Kahneman and Tversky suggested, over/under-estimation of perceived likelihood of events means differently from over/under-weighting, which is a property of decision weights. Over-extrapolation hypothesis suggests that analysts extrapolated the past performance in underlying asset too far into the future, hence underestimating the probability of security defaults. It coincides with the 2nd “belief-based” theory in “Bubble Formation”, mentioned in Barberis’ 2010 paper. He believes analysts made these wrong judgments by the application of representativeness heuristic, which was introduced by Kahneman and Tversky (1974). Analysts were over-confident in property prices in the future, that they neglected the financial background of some of the sub-prime mortgage holders. Lack of study in potential outbursts in payment default causes absolutely no preparation for its coming; hence AAA rating securities were no longer safe. Kahneman and Tversky’s 1974 paper provided some important psychological explanations to this kind of over-extrapolation, such as insensitivity to predictability, insensitive to prior probabilities and the illusion of validity. In their 1979 paper, their theory on very low probability event being ignored due to people’s inability to comprehend, might explain why the markets did not “insure” themselves from this possible crisis. If given favorable description of numerical predictions, then the reliability of evidence and the statistical expected accuracy of the prediction are no longer sensitive to the case. Prior the crisis, the industry was confident in these securities because of the booming housing market and the low default rate. The outlook was favorable to the investors, as long as everyone keeps buying property, taking out mortgages and paying their loans. Overall, the positive market environment discouraged market participants to look into the probabilities of potential outcomes. Hence any base rate probabilities of default or price plunge were overlooked. The illusion of validity suggested the confidence market participants had depends primarily on the match between current input and predicted outcome, with little or no regard for the factors that limits the predictive accuracy (Kahneman and Tversky, 1974). 4
The bullish market performance in mortgage security was a good match for a continuous good performance of the sector in the future. Hence, analysts ignored the limitations of risks in sub-prime mortgage defaults and risks in the plunging housing prices, for it was reasonable to think so at the time. Moreover, people tend to ignore highly unlikely events if they cannot understand them.
4.8 The 2nd theory in why analysts over-extrapolate past performance is a type of belief manipulation. This is closely connected with the representativeness heuristic and is explained in sections below.
4.9 All of the events discussed below can be explained by the over-extrapolation hypothesis and naïve application of representativeness heuristic mentioned above.
The U.S. stock market in the 1920s: The Roaring Twenties was a phase of high growths economically, socially, and politically, principally in North America
(http://en.wikipedia.org/wiki/Wall_Street_Crash_of_1929). The U.S. stock market was bullish until the Wall Street Crash in 1929, with the famous Black Thursday and Black Tuesday in the following week. Dow Jones Industrial Average did not recover until July 1932, with a total value loss of 89% of the highest value from the peak. Followed by the crisis was the 12-year Great Depression and long bearish market. Academics debated heavily on how the plunge in stock prices happened and whether there were speculative and immoral acts. The U.S. Congress passed the famous Glass-Steagall Act in 1933.
The U.S. stock market in the 1990s: The Dot-Com Bubble was the second boom in U.S. stock market that resembled the crisis in 1929 (http://en.wikipedia.org/wiki/Dot-com_bubble). It formed a 5
speculative bubble between 1995 and 2000, marked by the founding of a group of internet-based companies. The stock market crashed on early March 2000, and it was speculated that the plunge in NASDAQ was accidental due to a simultaneous sell order from major high tech stocks. While many other theories were presented to support the bubble burst between 2000 and 2002, such as 9/11 attacks and excessive business spending for Y2K, The dot-com model was, as the article suggested, flawed: a vast number of companies with homogenous business plans to monopolize their respective sectors and massive spending before any probable income nor sound profit model.
The Japanese real estate and stock markets in the late 1980s: The collapse of the Japanese real estate and stock market occurred gradually, compared with the Wall Street Crash and the burst of Dot-Com Bubble. From decades of savings in banks, Yen appreciated strongly and Japan had large trade surpluses. Since loans and credit were easily obtained and backed by large savings, Japanese banks became speculative and overconfident in market. Some suggested the crash was due to speculative acts from overseas investment. Followed by the crash was the “Lost Decades” of the 90s and 00s.
The South Sea bubble of 1720: “South Seas” means only to South America and surrounding waters at the time. The South Sea Company was established as a trading company where its real purpose was of funding government debts from the War of the Spanish Succession (http://en.wikipedia.org/wiki/South_Sea_Company). It was a scheme to trade high-interest illiquid debts into low-interest liquid debts and shares in South Sea Company. The company was given steady income streams from government and monopolistic trade rights in South America. The company’s P/E increased dramatically along with its share prices, and heads of government and royal families supported this frenzy in the hope of selling their shares for profits in the future. Shares plunged in 1720 when reached GBP1, 000 and the Bubble Act was introduced.
6
The Tulip-mania of the 1630s: Tulips became popular and luxurious items in Netherlands from Mid-16th century. The Tulip-mania was an act of irrationality of crowds of people, which reflected the general cycle of a bubble, as did the South Sea Company
(http://en.wikipedia.org/wiki/Tulip_mania). It was the first large scale crisis that shocked the value system of a large society. Critiques included Mackay’s “Madness of Crowd” and the modern Efficient Market Hypothesis challenged Mackay’s view. The mania was referenced to other major crisis in modern history, including the 2007 credit crisis.
4.10 We put our skin in the game when we invest heavily in the corporations or schemes with our own money, hence achieving aligned interests and comfort for other parties involved. It seemed that the banks held large number of sub-prime securities to give investors confidence in the investments they had in these products.
4.11 Banks, specifically the ones in shadow banking system, was high vulnerable to deep leverage in short term debts, because their capital structures were mainly short term liabilities with long term assets
(http://en.wikipedia.org/wiki/Shadow_banking_system). Firstly, these banks have to constantly return to capital markets for refinances since they don’t have deposits not support from Central Bank. Therefore, during housing market deterioration, they could not obtain funding from investors through mortgage-based securities as they used to. Secondly, high leverage magnified losses and these banks could not obtain funding to cover the losses except from selling their long-term assets. Short-term debts cannot be repaid because of illiquidity in the asset structure and they went bankrupt.
7
4.12 Barberis (2010) suggested three answers. Firstly, labeled the “bad incentives” view, posits that people on the mortgage desks of banks were aware of the significant risks to which their institutions were exposing through their activities, and their compensation schemes encouraged them to do so for they faced no consequences of the risk they took (Acharya et al, 2009). Secondly, labeled the “bad models” view, says that the mortgage people were genuinely unaware of the risk embedded in their sub-prime holdings due to faulty reasoning. Thirdly, labeled the “bad luck” view, disputing the second view, argued that a rational individual with the right incentives would not have anticipated the poor subsequent performance of sub-prime-linked securities.
4.13 Yes, Barberis could not agree on all three conjectured answers suggested to the question, for he thought the logic behind these theories were somewhat naïve. I would agree on his analysis towards these views in some degrees. Firstly, I do believe the people developing all those mortgage-related products are fairly educated in Mathematics and Finance, hence the models should not have big fundamental flaws, unless with faulty technical things. Secondly, how people balance their rewards and responsibilities is of the issue of morality, so we cannot have a clean-cut answer to why or why not some of us do things that are potentially ruinous to others for our own benefits. Nevertheless, pride in our job can cause extreme behaviors, and make us ignore others’ feelings from our actions. Thirdly, luck should not be considered separately, although it is a large possible reason, and I believe information asymmetry existed among even the most rational individuals, which has nothing to do with luck sometimes.
8
4.14 – 4.16 Cognitive Dissonance is an uncomfortable feeling caused when new ideas disagree with the existing perceptions of the world. Some classical examples are “Sour Grapes” and people’s attitude towards smoking. The theory leads to the conclusion that humans tend to rationalize more often than to be rational. The anxiety associated with the possibility of having made the wrong predictions can lead to rationalization, which is the tendency to create positive attributes to support existing predictions. Cognitive Dissonance can lead to Confirmation Bias, which contributes to over-confidence in personal beliefs and causes people to give up searching for a solution for the wrong predictions (http://en.wikipedia.org/wiki/Cognitive_dissonance). Once again, ego plays a large role in inducing people to manipulate their beliefs to reduce dissonance. Representativeness heuristic made this attitude towards change in beliefs easier to embrace. As mentioned above, the rational decision to resign was too costly for mortgage people, and instead they rationalize their beliefs to justify their performance, a less costly alternative, as discussed in Barberis’ 2010 paper about the behaviors of mortgage bankers, mortgage traders and credit rating agencies. People’s motivation to manipulate their beliefs on risks of sub-prime mortgage securities was supported by representativeness heuristic and cognitive dissonance. People chose to believe the wrong perceptions of the risks of sub-prime securities; hence belief manipulation view explained the “bad models” view. It is relatively much more economically efficient and less time-consuming to adapt a new belief than to research what is wrong with the current belief. Benabou (2009) has developed a general model explaining wishful thinking and reality denial spread through organizations and market. In his paper, he suggested many outside forces causing a collective overconfidence in delusion of reality and the systematic ignoring of warning signals. He suggested harmful groupthink bureaucratic pressures influenced individual belief manipulation and led the market to manias.
9
4.17 – 4.18 I prefer to discuss both institutional and psychological amplification mechanisms together, since the two ideas were closely related. Krishnamurthy (2010) defined these two concepts as Balance Sheet amplifier and Information amplifier. When managers of funds and investments experience excessive market mispricing, their return performances were badly affected, investors would have serious doubts in the managers’ capabilities, and they would prefer fund withdrawals and tightening future funding into the market in general. Hence, to resolve this illiquidity, managers will have to liquidate their illiquid assets to reoptimize their portfolio. Because there were so many managers taking very similar strategies in sub-prime related investments, the unwinding of all these positions and the sale of many assets to cover the positions created a shock to the market, sending frightening signals to investors, causing more similar actions occurred, in very short period of time. Creditors wanted their investments back quick and tightened credits in the markets. When prices were driven down to the level of margin calls, it was the time many other managers in the markets start to take the same positions as those unfortunate ones have to, even though they were not adversely affected by the sub-prime securities. Because of the limits to Arbitrages from Performance-based Arbitrage (Shleifer and Vishny, 1997), agency conflicts, derived from market knowledge inconsistencies between investors and managers (arbitrageurs), caused managers to change their strategies and left markets when their participations were needed the most. This in general, is the institutional amplification mechanism. I agree with Krishnamurthy and Caballero’s view (2008) on psychological amplification mechanism, that it was the main driver in the recent credit crisis. Sub-prime products had a relatively short history in financial markets and investors have not yet fully understand them due to their complexities and lack of past performance for analysis. People are very fragile when facing uncertainties, explained by loss aversion and ambiguity aversion. Institutional amplification mechanism explained how actual markets actions would worsen the situations, whereas psychological amplification mechanism drove those actions and magnify the scale and 10
groups involved. When the initial bad performances occurred, people panicked and could not understand why these instruments would failed, causing all of them, including the ones in relatively safe positions, to make decisions based on worst scenarios. Lack of supporting knowledge and agency conflicts contributed to people’s worsening confidence that eventually drove markets into serious illiquidity. Opportunities to recover from bad positions were forgone because of massive investor bail-outs. People thought they will loss everything, and they liquidate all positions and withdraw funds from markets, causing the sale of assets and magnified value loss spirals mentioned above.
4.19 – 4.20 Loss aversion was introduced and explained by Kahneman and Tversky (1979), where they suggested people are more sensitive to losses than gains, through graphical illustration of a steeper loss value function. Ambiguity aversion (Uncertainty aversion), explained by Daniel Ellsberg Paradox (1961), refers to the attitude of choice to understandable risks over unknown risks. Economists and financial practitioners normally limit their understanding of ambiguity aversion to a static choice in decision, that if the events have immeasurable risks or vague probabilities of outcomes, people will turn away from the game. However, they are ignoring the fact that people’s understandings and knowledge will improve towards the situations, hence change the degree of aversion towards ambiguity (Heath and Tversky, 1991, Fox and Tversky, 1995). Competence hypothesis, introduced in Heath and Tversky’s 1991 paper, was a theory contrary to the conventional ambiguity aversion, which was discussed extensively by Ellsberg and others. It suggested that the preference for clear over vague probabilities was not confined to game of chance, but also extends to the uncertain beliefs based on knowledge (Fox and Tversky, 1995). Heath and Tversky thought motivation is the reason behind competence hypothesis, which people with knowledgeable background in the game played will be credited for their success, and people with limited background will be blamed for their losses. The balance of blame 11
and credit was the main psychological drive behind the fact people bet on skills and knowledge, besides the cognitive reason, which people learn from experiences that success would more likely come with knowledge. Hence people will sometimes prefer vague probabilities and do not bet on chances because they are relatively more familiar to the type of situations they are in.
4.21 Fox and Tversky (1995) introduced the comparative ignorance hypothesis and the study involving students from the two universities was one of the original experiments they conducted for analysis purpose. The hypothesis suggested that subjects (people) show relatively more ambiguity aversion when evaluate prospects, clear and vague probabilities, jointly, than when evaluate each prospect separately. In other words, the experiments focus more on the subjects than on the prospects. In this particular study, when students from San Jose State University were told that peers from Stanford were also doing the analysis, their sense of competency dropped, caused them deviate from preference to vague probabilities (paid 150 if correct), and go for the clear probabilities(paid 50 sure). Fox and Tversky’s paper had other experiments, which provided important insights in why people felt less competent after experiencing the plunge in sub-prime assets performance, later turned into a general market panic and triggered the large scale credit withdrawals and market illiquidity.
4.22 I think these concepts explain why people make different decisions, from unique yet connected angles. Ambiguity aversion is perhaps the most important psychological effects upon people’s panicked behavior in the crisis. I would assume that investors felt pretty confident with the sub-prime products they bought through fund managers and banks, because they were told by these agents that the underlying would have prosperous outcome in the future, and the market remained stable for the past years. Agency conflicts made it difficult for investors to fully understand the situations they 12
are in and the risks involved. Belief manipulation came in play and altered investors’ belief, so they became confident in their investments and felt more competent in themselves, hence the build-up of the large position in market before the crisis. Ambiguity aversion increased when investors suffered losses from their investments, and they now felt less competent because they could not understand the losses, hence deviated aggressively from the situations. Loss aversion contributed to the withdrawals because it magnified after the initial losses, causing the bad situation worsen much quicker. Risk aversion explains investors’ attitude towards preferences under known probabilities of outcomes. It is a theory on how much compensations investors want from taking on certain level of risk, which are measurable and understandable. Some managers in this crisis were risk-averse, and they were not arbitrageurs in relative nature. However, they were still affected by the spirals of the crisis, that they exhibited similar attitudes and make withdrawals from the market. I believe risk seeking behavior is positively correlated with ambiguity seeking behavior. A known risk of possible breakdown from the sub-prime products due to low qualities of the underlying was overlooked by belief manipulation, and confident investments in ambiguity seeking nature built up. Loss aversion magnified the downturn in attitude toward risk and ambiguity.
4.23 Prescriptive behavioral finance could mean financial innovations designed in strict laws and rules and intended to achieve better financial results. Current innovations seemed to complex and there are neither regulations nor guidelines to help standardizing them.
5 I do think that that many Finance practitioners are self-delusional most of the time. Today’s financial market is extremely large and complex, compared to 100 years ago. Practitioners are more specialized in their own area of expertise, and sometimes could 13
not fully understand a bigger picture. Information asymmetry and constraints contributes to this situation, despite all the information are seemingly available over internet and paper materials. It would not be possible for one person to fully comprehend and connect all these information and form a really informed idea of the market. I mean one could do that, but without the participation in the market, because it’s just too energy consuming. As a result, one can only understand as much as he can. Having said that, ideas come up from different market practitioners all the time, and many of us formalize these ideas and produce theories and products. Imagine the speed of these ideas flow into the markets, and investors and money managers must try to understand them before choosing the ones that suit their portfolios. It would be a daunting task. Therefore most of us will cut back the work load by giving a subjective belief rather than material belief to certain products. We produce these beliefs based on our expert experiences and personal competence. However, as discussed many times in this paper, how reliable our beliefs are is questionable, and subjective judgments can turn us away from those real valuable opportunities. Our eyes and mind often trick us into believing things that are not real. The concept magical thinking explains the type of associative thinking applied by many practitioners, who make decisions over-confidently. This illusion of control is toxic because in the hierarchy of corporations, senior pressure will force this unreal belief upon those who disagree and hold views which in fact more material. A wide spread of this thinking could lead the market slowly into crisis. Interestingly, when people are panicking, they are irrational, hence they hold on to certain beliefs from those who are strong mentally. Also, gambler’s fallacy tells us that people would expect an unrealistic turn of events under extreme situations, where in fact no correlation existed between the past and the future.
14
6 I found this assignment very useful, because it forced me to understand behavioral finance, which I never stumble upon before. Psychology is very important in finance, because I believe that behind all the analysis are the interactions of minds that drive the market. I found the topic in cognitive dissonance and the history of the five bubbles to be most interesting, because belief manipulation was beyond finance and history of those bubbles brought out lots of other interesting topics, and applied to all general people decisions. I found the most difficult part was when I tried to understand prospect theory and its related discussions. All the detailed discussions on the unique assumptions and hypothesis on human behaviors are the most unfamiliar to me, and they are challenging because some of the concepts sounds really close, yet unique.
REFERENCES Acharya, V., Cooley, T., Richardson, M., and I. Walter (2009), “Manufacturing Tail Risk: A Perspective on the Financial Crisis of 2007-2009,” Foundations and Trends in Finance 4, 247-325. Barberis, N. (2010), “Psychology and the Financial Crisis of 2007-2008,” Financial Innovation and Crisis, MIT Press. Barberis N. and A. Shleifer (2003), “Style Investing,” Journal of Financial Economics 68, 161-199. Benabou, R. (2009), “Groupthink: Collective Delusions in Organizations and Markets,” Working Paper, Princeton University. Burns, Z., Chiu, A. and G. Wu (2010), “Overweighting of Small Probabilities,” Working Paper, University of Chicago. Caballero, R. and A. Krishnamurthy (2008), “Musical Chairs: A Comment on the Credit Crisis,” Banque de France Financial Stability Review 11, 1-3. Cheung, Y. H. (2010), “A Monte Carlo Study of the Probability Weighting Function,” Working Paper, Edith Cowan University, Western Australia. Fox, C. and A. Tversky (1995), “Ambiguity Aversion and Comparative Ignorance,” Quarterly Journal of Economics 110, 585-603.
15
Heath, C. and A. Tversky (1991), “Preference and belief: Ambiguity and Competence in Choice under Uncertainty,” Journal of Risk and Uncertainty 4, 5-28. Kahneman, D. and A. Tversky (1974), “Judgment under Uncertainty: Heuristics and Biases,” Science 185, 1124-1131. Kahneman, D. and A. Tversky (1979), “Prospect Theory: An Analysis of Decision under Risk,” Econometrica 47, 263-291. Krishnamurthy, A. (2010), “Amplification Mechanisms in Liquidity Crises,” American Economic Journal: Macroeconomics 2, 1-30. Shleifer, A. and R. Vishny (1997), “The Limits of Arbitrage,” Journal of Finance 52, 35-55.
