More Than You Know

by Michael J Mauboussin

  —Antonio Damasio,

  Descartes’ Error: Emotion, Reason, and the Human Brain

  Emotions and Decisions

  Neuroscientist Antonio Damasio describes how early in his career he realized that traditional views on rationality had to be wrong. He saw a patient with all the faculties for rational behavior intact—attention, memory, logic. But brain damage had eviscerated the man’s ability to experience feelings, and this had robbed him of the ability to make successful decisions day to day. Damasio saw the link: impaired feelings and flawed decisions go hand in hand.1

  Damasio’s later work confirmed his observation. In one experiment, he harnessed subjects to a skin-conductance-response machine and asked them to flip over cards from one of four decks; two of the decks generated gains (in play money) and the other two were losers. As the subjects turned cards, Damasio asked them what they thought was going on. After about ten turns, the subjects started showing physical reactions when they reached for a losing deck. About fifty cards into the experiment, the subjects articulated a hunch that two of the four decks were riskier. And it took another thirty cards for the subjects to explain why their hunch was right.2

  This experiment provided two remarkable decision-making lessons. First, the unconscious knew what was going on before the conscious did. Second, even the subjects who never articulated what was going on had unconscious physical reactions that guided their decisions.

  When Damasio replicated the experiment on brain-damaged patients, he saw none of the typical reactions. The skin-conductance response and verbal descriptions confirmed that the patients had no idea what was going on—either unconsciously or consciously.3

  Two Follows One

  In his Nobel Prize lecture, Daniel Kahneman describes two systems of decision making.4 System 1, the experiential system, is “fast, automatic, effortless, associative, and difficult to control or modify.” System 2 is analytical and “slower, serial, effortful, and deliberately controlled.” Exhibit 12.1 compares these systems.

  EXHIBIT 12.1 Comparison of the Experiential and Analytical Systems

  Experiential SystemAnalytical System

  1. Holistic 1. Analytic

  2. Affective: Pleasure/pain oriented (what feels good) 2. Logical: Reason oriented (what is sensible)

  3. Associationistic connections 3. Logical connections

  4. Behavior mediated by “vibes” from past experiences 4. Behavior mediated by conscious appraisal of events

  5. Encodes reality in concrete images, metaphors, and narratives 5. Encodes reality in abstract symbols, words, and numbers

  6. More rapid processing: Oriented toward immediate action 6. Slower processing: Oriented toward delayed action

  7. Slower to change: Changes with repetitive or intense experience 7. Changes more rapidly: Changes with speed of thought

  8. More crudely differentiated: Broad generalization gradient; stereotypical thinking 8. More highly differentiated

  9. More crudely integrated: Dissociative, emotional 9. More highly integrated: Cross-context processing

  10. Experienced passively and preconsciously: We are seized by our emotions 10. Experienced actively and consciously: We are in control of our thoughts

  11. Self-evidently valid: “Experiencing is believing” 11. Requires justification via logic and evidence

  Source: Epstein, “Cognitive-Experiential Self-Theory.” Adapted by permission.

  In Kahneman’s model, system 1 uses perception and intuition to generate impressions of objects. These impressions are involuntary, and an individual may not be able to verbalize them. He argues that system 2 is involved in all judgments, whether or not the individual is making decisions overtly. Intuition is a judgment that reflects an impression. Kahneman’s work (along with that of his collaborator, Amos Tversky) shows how impressions can lead to judgments that are suboptimal according to classical economic theory.

  So the evidence suggests that you can’t separate emotions (system 1) from decisions (system 2). In fact, as Damasio showed, system 1 needs to operate normally in order for you to make good judgments. From an investor’s standpoint, two questions become central: What influences our impressions, and how do these impressions shape perceptions of risk and reward?

  The Affect Heuristic

  One of the main shapers of our impressions is what psychiatrists call “affect.”5 Affect is the “goodness” or “badness” we feel based on a stimulus. For example, a word like “treasure” generates positive affect, while a word like “hate” is negative.

  Affect operates in the realm of system 1 and hence is rapid and automatic. And affect often directs our impressions in a reasonable way: most things you feel good about are good. But affect, like other heuristics (or rules of thumb), has biases. Investors need to heed the biases that emanate from affect.

  Affect is a noteworthy extension to prospect theory—which shows that investors are risk averse when facing gains and risk seeking when facing losses. Experiments show that affect—how we feel about a financial opportunity—can amplify the suboptimal biases that arise from prospect theory (see exhibit 12.2).

  Let’s get more concrete. The goal of an investor is to buy an asset below its expected value. Expected value is the weighted-average value for a distribution of possible outcomes. You calculate expected value by multiplying the payoff for a given outcome by the probability that the outcome will occur.

  Research on affect demonstrates two central principles related to expected value. First, when the outcomes of an opportunity don’t have strong affective meaning, we tend to overweight the probabilities. Second, when the outcome does have strong affective meaning, we tend to overweight the outcome.

  Paul Slovic tested the first principle, probability dominance, with a simple experiment. He asked subjects to rate one of sixteen gambles by crossing various probabilities (7/36, 14/36, 21/36, and 28/36) and various payoffs ($3, $6, $9, and $12). He found that even though the subjects wanted to weight the probability and payoffs equally (and thought they had done so), the actual weighting for probability was five to sixteen times higher than for payoff.6

  EXHIBIT 12.2 Affective Psychology of Risk

  Source: Rottenstreich and Hsee, “Money, Kisses, and Electric Shocks.” Used by permission of Blackwell Publishing.

  The researchers posit that the subjects leaned on probabilities because there was no way for them to judge the attractiveness of the payoffs—the payoffs lacked affective meaning. Scientists see examples of this probability dominance in other fields as well, including studies of life-saving interventions.

  In contrast, when payoffs are vivid—that is, when they carry substantial affective meaning—subjects tend to place too little emphasis on probabilities and too much emphasis on outcomes. For example, researchers find that lottery players tend to have the same feelings about playing the lottery whether the probability of winning is one-in-ten million or one-in-ten thousand because the payoff is so affective. This feature of the theory also offers an explanation as to why handicappers consistently overestimate the odds of a long shot at the racetrack and why people fear flying.

  The bottom line is that when investors feel good about an investment idea, they deem the risks low and the returns high irrespective of more objective probabilities.7 And when they dislike an idea, the inverse is true—risk is high and reward is low. Great investors aren’t too swayed by affect. Perhaps this is a result of how their system 1s are wired.

  When the Experiential Fails

  Our experiential systems function well by and large. When do they fail?

  Our experiential system can fail us when outside forces manipulate it. One example is advertising. Advertisers often try to appeal to your affect by providing you with a vivid perception. So whenever you face a probability-and-outcome decision, be very aware of how you feel (or are being made to feel) about the outcomes, and try not to let that feeling cloud the objective probabilities.

  Experiential systems also fail in nonl
inear or nonstationary systems. In nonlinear systems, cause and effect are not neatly linked. As a result, outcomes can be very counterintuitive. In nonstationary systems, the underlying statistical properties of the system change over time, which means that the past may not be a good predictor of the future. The stock market exhibits both nonlinearity and nonstationarity. Accordingly, investors must take a very methodical and self-aware approach to judging expected values.

  Affect: Individual Versus the Collective

  One should be careful about extrapolating the implications of affect to suggest that markets are inefficient. We all have our individual hard wiring and experiences; hence we are all going to feel affect in different ways. As markets are an aggregation of individual views, they can be efficient (or near efficient) provided that affect-driven biases are uncorrelated.

  A dominant idea in Western society is that we should separate emotion and rationality. Advances in science show that such a separation is not only impossible but also undesirable. Yet successful investing requires a clear sense of probabilities and payoffs. Investors who are aware of affect are likely to make better decisions over time.

  13

  Guppy Love

  The Role of Imitation in Markets

  When people are free to do as they please, they usually imitate each other.

  —Eric Hoffer, The True Believer

  Guppy See, Guppy Do

  At first blush, biologist Lee Dugatkin appears to be a guy with way too much time on his hands. The focus of his research is the apparently esoteric question of how female guppies select mates. As it turns out, female guppies have a genetic preference for bright-orange males. But when Dugatkin arranged for some females to observe other females choosing dull-colored males, the observing females also selected the dull males. Surprisingly, in many instances female observers overruled their instincts and chose instead to imitate other females.1

  Why should anyone care about how female guppies pick their partners? The answer gets to the core of a lively debate about whether animal behavior is shaped solely by genetic factors or if culture plays a part. Dugatkin’s work demonstrates that imitation—a form of cultural transmission—is clearly evident in the animal kingdom and plays a central role in species development.2

  Certainly, too, imitation is a vital force with humans. Fashions, fads, and traditions are all the result of imitation. And since investing is inherently a social activity, there is every reason to believe that imitation plays a prime part in markets as well.

  Most investors and businesspeople have fundamental philosophies that are supposed to define their behavior—much like genetics shape guppy mate selection. But we know that for money managers and guppies alike, imitation sometimes has a substantial influence on decision making. So is imitation good or bad for investors?

  Feedback—Negative and Positive

  Well-functioning financial markets, like other decentralized systems, rely on a healthy balance between negative and positive feedback. Negative feedback is a stabilizing factor, while positive feedback promotes change. Too much of either type of feedback can leave a system out of balance.3

  The classic example of negative feedback in markets is arbitrage. Indeed, arbitrage is a central plank in the case for efficient markets. For example, if the price of a security diverges from its warranted value, arbitrageurs buy or sell the appropriate securities in order to close the price/value gap. Negative feedback resists change by pushing in the opposite direction.4

  Positive feedback, on the other hand, reinforces an initial change in the same direction. The snowball effect, cascades, and amplification are all examples of positive feedback. While investors often view positive feedback as undesirable, especially when it leads to a runaway process, it isn’t always bad.

  When is positive feedback good? Well, it can help promote a smart decision. For instance, early investors in a promising new industry may encourage others to invest, sparking the industry’s growth. Positive feedback can also get a system out of a bad situation. In nature, a “follow-your-neighbor” strategy may allow a flock of birds to elude a predator. Analogously, it can help investors flee a bad investment.5

  Follow the Ant in Front of You

  Imitation is one of the prime mechanisms for positive feedback. Momentum investing, for example, assumes that a stock that is rising will continue to rise. If enough investors follow a momentum strategy, the prophecy of a high price becomes self-fulfilling.

  Most investors view pure imitation with some misgiving, belying their often-imitative actions. But imitation often has a rational basis. Consider the following cases, for example:6 Asymmetric information. Imitation can be very valuable when other investors know more about a particular investment than you do. We all routinely use imitation in our day-to-day decision making, allowing us to leverage the specialized knowledge of others.

  Agency costs. Many money-management firms must make trade-offs between maximizing the performance of the investment portfolio (long-term absolute returns) and maximizing the value of the money-management business (by collecting assets and fees). Companies that choose to maximize the value of the business have an incentive to do what everybody else is doing. This imitation minimizes tracking error versus a benchmark.

  Preference for conformity. As Keynes said, “Worldly wisdom teaches that it is better for reputation to fail conventionally than to succeed unconventionally.” Humans like being part of a crowd, as the group often bestows safety and reassurance.

  So positive feedback is desirable under some circumstances and investor imitation can make sense. But positive feedback can also lead to excesses.

  Financial economists describe herding as when a large group of investors make the same choice based on the observations of others, independent of their own knowledge.7 In effect, herding occurs when positive feedback gets the upper hand. Given that markets need a balance between positive and negative feedback, such an imbalance leads to market inefficiency. This is in contrast to the classical view that investors trade solely on the basis of fundamental information.

  Determining exactly how much positive feedback is too much may be an impossible task. Extensive scientific studies of innovation and idea diffusion reveal that there is typically a critical threshold, a tipping point, beyond which positive feedback takes over and the trend dominates the system. The relative frequency of bubbles and crashes strongly suggests that there are consistent discrepancies between price and value.8

  The market is not the only decentralized system that exhibits suboptimal imitation. For example, there is the fascinating case of army ants. A group of worker ants, which are essentially blind, sometimes separates from the colony. Since no individual ant has any idea how to relocate the rest of the colony, all of the ants rely on a simple decision rule: follow the ant in front of you. If enough individuals follow the strategy (i.e., they reach the tipping point), they develop a circular mill, where ants follow each other around in circles until death. One such mill persisted for two days, had a 1,200-foot circumference, and a circuit time of two and a half hours (see exhibit 13.1). Eventually, a few workers created the requisite diversity by breaking away from the trail, and the mill dissipated.9

  EXHIBIT 13.1 A Circular Mill of Army Ants

  Source: Author analysis.

  Of course, for the ants imitation is hard-wired genetic behavior, not cultural. Investors, in contrast, have the ability to think independently. However, Charles MacKay’s famous words from over 150 years ago remind us that avoiding the imitation trap is an age-old problem: “Men, it has been well said, think in herds; it will be seen that they go mad in herds, while they recover their senses slowly, and one by one.”10

  Herding from the Grapevine

  George Soros is the most prominent investor to explicitly cite the role of positive feedback in his investment philosophy. Soros’s theory of reflexivity argues that there is positive feedback between a company’s stock price and its fundamentals, and that this feedback can l
ead to booms and crashes. Soros’s strategy was to take advantage of these trends by either buying or shorting stocks.

  The finance literature also reveals a number of examples of herding among investors:• Mutual funds. Russ Wermers found evidence of herding among mutual funds, especially in small-capitalization stocks and growth-oriented funds. He found that the stocks the herd buys outperformed the stocks the herd sells by 4 percent during the subsequent six months.11

  • Analysts. Ivo Welch shows that a buy or sell recommendation of a sell-side analyst has a significantly positive influence on the recommendations of the next two analysts. Analysts often look to the left and to the right before they make their recommendations.12

  • Fat tails. Econophysicists, using simple herding models, have replicated the fat-tail price distributions that we empirically observe in markets. These models provide a much more convincing picture of market reality than those that assume investor rationality.13

  In markets, a symbiotic relationship between positive and negative feedback generally prevails. If all speculators destabilized prices, they would buy high and sell low, on average. The market would quickly eliminate such speculators. Further, arbitrage—speculation that stabilizes prices—unquestionably plays a prime role in markets. But the evidence shows that positive feedback can dominate prices, if only for a short time.14 Imitation can cause investors to deviate from their stated fundamental investment approach and likely provides important clues into our understanding of risk. Next time you buy or sell a stock, think of the guppies.