Reconciling Efficient Markets With Behavioural Finance

R E C E N T R E S E A R C H

VOL. 7, NO. 2, 2005 21

Abstract

The battle between proponents of the EfficientMarkets Hypothesis and champions of be-havioral finance has never been more pitched,

and little consensus exists as to which side is winning orthe implications for investment management and con-sulting. In this article, I review the case for and againstthe Efficient Markets Hypothesis and describe a newframework—the Adaptive Markets Hypothesis—inwhich the traditional models of modern financial eco-nomics can coexist alongside behavioral models in anintellectually consistent manner. Based on evolutionaryprinciples, the Adaptive Markets Hypothesis impliesthat the degree of market efficiency is related to envi-ronmental factors characterizing market ecology such asthe number of competitors in the market, the magni-tude of profit opportunities available, and the adapt-ability of the market participants. Many of the examplesthat behavioralists cite as violations of rationality thatare inconsistent with market efficiency—loss aversion,overconfidence, overreaction, mental accounting, andother behavioral biases—are, in fact, consistent with anevolutionary model of individuals adapting to a chang-ing environment via simple heuristics. Despite the qual-itative nature of this new paradigm, I show that theAdaptive Markets Hypothesis yields a number of sur-prisingly concrete applications for both investmentmanagers and consultants.

Introduction

Much of modern investment theory and prac-tice is predicated on the Efficient Markets Hy-pothesis (EMH), the notion that markets

fully, accurately, and instantaneously incorporate allavailable information into market prices. Underlyingthis far-reaching idea is the assumption that market par-ticipants are rational economic beings, always acting inself-interest and making optimal decisions by trading offcosts and benefits weighted by statistically correct prob-abilities and marginal utilities. These assumptions ofrationality, and their corresponding implications formarket efficiency, have come under attack recently froma number of quarters. In particular, psychologists andexperimental economists have documented a number ofdepartures from market rationality in the form ofspecific behavioral biases that are apparently ubiquitousto human decision making under uncertainty, several ofwhich lead to undesirable outcomes for an individual’seconomic welfare.

While there is little doubt that humans do exhibitcertain behavioral idiosyncrasies from time to time, thereis still no clear consensus as to what this means forinvestment management. Although a number of alterna-tives have been proposed, no single theory has managedto supplant the EMH in either academic or industryforums. This state of affairs is due partly to the enormousimpact that modern financial economics has had on the-

RECONCILING EFFICIENT MARKETS WITH BEHAVIORAL FINANCE:

THE ADAPTIVEMARKETS HYPOTHESIS1

By Andrew W. Lo, Ph.D.


ory and practice during the past half century. It is diffi-cult to overturn an orthodoxy that has yielded suchinsights as portfolio optimization, the Capital AssetPricing Model, the Arbitrage Pricing Theory, the Cox-Ingersoll-Ross theory of the term structure of interestrates, and the Black-Scholes/Merton option pricingmodel, all of which are predicated on the EMH in oneway or another. Although behavioral versions of utilitytheory (Kahneman and Tversky, 1979), portfolio theory(Shefrin and Statman, 2000), the Capital Asset PricingModel (Merton, 1987; Shefrin and Statman, 1994), andthe Life Cycle Hypothesis (Shefrin and Thaler, 1988)have been advanced, these models have yet to achievethe kind of general acceptance among behavioralists andpractitioners that the EMH enjoys among its disciples.

But another reason for the fragmentary nature ofbehavioral finance is the dearth of fundamental axiomsfrom which all behavioral anomalies can be generated.For example, while Kahneman and Tversky’s (1979)prospect theory can generate behavior consistent withloss aversion (see the section titled “Motivation,”below), their framework cannot generate biases such asoverconfidence and regret at the same time. EMH pro-ponents sometimes criticize the behavioral literature asprimarily observational, an intriguing collection ofcounterexamples without any unifying principles toexplain their origins. To a large extent, this criticism is areflection of the differences between economics andpsychology (see Rabin, 1998, 2002, for a detailed comparison). The field of psychology has its roots inempirical observation, controlled experimentation, andclinical applications. From the psychological perspec-tive, behavior is often the main object of study, and onlyafter carefully controlled experimental measurementsdo psychologists attempt to make inferences about theorigins of such behavior. In contrast, economists typi-cally derive behavior axiomatically from simple princi-ples such as expected utility maximization, resulting insharp predictions of economic behavior that are rou-tinely refuted empirically.

In this article, I describe a new paradigm that rec-onciles the EMH with behavioral biases in a consistentand intellectually satisfying manner, and then provideseveral applications of this new paradigm to some of themore practical aspects of investment management and

consulting. Named the Adaptive Markets Hypothesis(AMH) in Lo (2004), this new framework is based onsome well-known principles of evolutionary biology(competition, mutation, reproduction, and naturalselection), and I argue that the impact of these forces onfinancial institutions and market participants deter-mines the efficiency of markets and the waxing andwaning of investment products, businesses, industries,and ultimately institutional and individual fortunes. Inthis paradigm, the EMH may be viewed as the friction-less ideal that would exist if there were no capital mar-ket imperfections such as transaction costs, taxes,institutional rigidities, and limits to the cognitive andreasoning abilities of market participants. However, inthe presence of such real-world imperfections, the lawsof natural selection—or, more appropriately, “survival ofthe richest”—determine the evolution of markets andinstitutions. Within this paradigm, behavioral biases aresimply heuristics that have been taken out of context,not necessarily counterexamples to rationality. Givenenough time and enough competitive forces, any coun-terproductive heuristic will be reshaped to better fit thecurrent environment. The dynamics of natural selectionand evolution yield a unifying set of principles fromwhich all behavioral biases may be derived.

Although the AMH is still primarily a qualitativeand descriptive framework, it yields some surprisinglyconcrete insights when applied to practical settings suchas asset allocation, risk management, and investmentconsulting. For example, the AMH implies that (1) theequity risk premium is not constant through time butvaries according to the recent path of the stock marketand the demographics of investors during that path; (2) asset allocation can add value by exploiting the mar-ket’s path dependence as well as systematic changes inbehavior; (3) all investment products tend to experiencecycles of superior and inferior performance; (4) marketefficiency is not an all-or-none condition but is a char-acteristic that varies continuously over time and acrossmarkets; and (5) individual and institutional risk pref-erences are not likely to be stable over time.

In the next section, I review several of the mostprominent biases documented in the behavioral litera-ture, and the following section contains a rebuttal ofthese examples from the EMH perspective. To reconcile

22 THE JOURNAL OF INVESTMENT CONSULTING


VOL. 7, NO. 2, 2005 23

these two opposing perspectives, I then present somerecent results from the cognitive neurosciences litera-ture that shed new light on both rationality and behav-ior. These results provide the foundation for the AMH,which I then summarize. Several applications of theAMH to the practice of investment management are out-lined in the following section, and I conclude with a dis-cussion of a new and broader role for consultants, onethat involves a more intimate relationship between con-sultant and investor, as well as a new relationshipbetween consultant and manager.

Motivation

To illustrate the conflict between the EMH andbehavioral finance, consider the following exam-ple of overconfidence drawn from a study by

Russo and Shoemaker (1989). A number of subjectswere asked to provide 90-percent confidence intervalsfor a series of general-knowledge questions with numer-ical answers (see table 1). In other words, instead of pro-viding numerical estimates for each question, subjectswere asked to give lower and upper bounds so that eachinterval bracketed the correct answer with 90-percentprobability. If subjects accurately assessed their degree ofuncertainty with respect to each question, they shouldbe wrong about 10 percent of the time. In other words,each subject should be able to correctly bracket theanswers in nine out of ten questions. In a sample of more

than 1,000 participants, Russo and Shoemaker (1989)found that less than 1 percent of the subjects scored nineor better, and most individuals missed four to seven ofthe ten questions. They concluded that most individualsare considerably more confident of their general knowl-edge than may be warranted; that is, they are over-confident. Overconfidence has been demonstrated inmany other contexts and seems to be a universal traitamong most humans, as observed by Garrison Keillor inhis fictional town of Lake Wobegon, “...where all thewomen are strong, all the men are good-looking, and allthe children are above average.”

A second example involves another aspect of prob-ability assessment in which individuals assign probabil-ities to events not according to the basic axioms ofprobability theory but according to how representativethose events are of the general class of phenomena under consideration. Two psychologists, Tversky andKahneman (1982), posed the following question to asample of eighty-six subjects:

Linda is thirty-one years old, single, outspoken, and verybright. She majored in philosophy. As a student, she wasdeeply concerned with issues of discrimination and socialjustice, and also participated in antinuclear demonstra-tions. Please check off the most likely alternative: • Linda is a bank teller.• Linda is a bank teller and is active in the feminist

movement.

TABLE 1

Self-Test of Overconfidence90% CONFIDENCE INTERVAL

LOWER UPPER

1. Martin Luther King's age at death

2. Length of the Nile River (in miles)

3. Number of countries in OPEC

4. Number of books in the Old Testament

5. Diameter of the moon (in miles)

6. Weight of an empty Boeing 747 (in pounds)

7. Year in which Wolfgang Amadeus Mozart was born

8. Gestation period of an Asian elephant (in days)

9. Air distance from London to Tokyo (in miles)

10. Deepest known point in the ocean (in feet)

(For answers, see “Answer Key to Overconfidence Test” on page 39.) Source: Russo and Shoemaker (1989)


Despite the fact that “bank teller” cannot be lessprobable than “bank teller and feminist” (because thelatter category is a more restrictive subset of the for-mer), almost 90 percent of the subjects tested checkedthe second response as the more likely alternative.Tversky and Kahneman (1982, p. 98) concluded: “Asthe amount of detail in a scenario increases, its proba-bility can only decrease steadily, but its representative-ness and hence its apparent likelihood may increase.The reliance on representativeness, we believe, is a pri-mary reason for the unwarranted appeal of detailed sce-narios and the illusory sense of insight that suchconstructions often provide.”

This behavioral bias is particularly relevant for thecurrent risk-management practice of scenario analysis inwhich the performance of portfolios is simulated forspecific market scenarios such as the stock-market crashof October 19, 1987. While adding detail in the form ofa specific scenario to a risk-management simulationmakes it more palpable and intuitive—in Tversky andKahneman’s (1982) terminology, more representative—italso decreases the probability of occurrence. Therefore,decisions based largely on scenario analysis may over-estimate the likelihood of those scenarios and, as a result,underestimate the likelihood of more relevant outcomes.

A third example of a behavioral bias—one involvingrisk preferences—is a slightly modified version of anoth-er experiment conducted by Kahneman and Tversky(1979), for which Kahneman was awarded the NobelMemorial Prize in Economic Sciences in 2002.2 Supposeyou are offered two investment opportunities: A and B.A yields a sure profit of $240,000 and B is a lottery tick-et yielding $1 million with a 25-percent probability and$0 with 75-percent probability. If you had to choosebetween A and B, which would you prefer? InvestmentB has an expected value of $250,000, which is higherthan A’s payoff, but this may not be all that meaningfulto you because you will receive either $1 million ornothing. Clearly, there is no right or wrong choice here;it is simply a matter of personal preference. Faced withthis choice, most subjects prefer A, the sure profit, to B,despite the fact that B offers a significant probability ofwinning considerably more. This behavior often is char-acterized as “risk aversion,” for obvious reasons. Nowsuppose you are faced with a second choice, between

C and D. C yields a sure loss of $750,000, and D is a lottery ticket yielding $0 with 25-percent probabilityand a loss of $1 million with 75-percent probability.Which would you prefer? This situation is not as absurdas it might seem at first glance; many financial decisionsinvolve choosing between the lesser of two evils. In thiscase, most subjects choose D, despite the fact that D ismore risky than C. When faced with two choices thatboth involve losses, individuals seem to be risk seeking,not risk averse as in the case of A versus B.

The fact that individuals tend to be risk averse in theface of gains and risk seeking in the face of losses canlead to some very poor financial decisions. To see why,first observe that the combination of choices A and D isequivalent to a single lottery ticket yielding $240,000with 25-percent probability and –$760,000 with 75-per-cent probability (choice A yields a sure gain of $240,000,and choice D yields $0 with 25-percent probability, and–$1,000,000 with 75-percent probability, implying a netloss of $760,000). On the other hand, the combinationof choices B and C is equivalent to a single lottery ticketyielding $250,000 with 25-percent probability and–$750,000 with 75-percent probability. The B-and-Ccombination has the same probabilities of gains and loss-es, but the gain is $10,000 higher, and the loss is$10,000 lower. In other words, B and C is formallyequivalent to A and D plus a sure profit of $10,000, yetmost subjects prefer A and D, precisely because they arerisk averse when it comes to gains, and risk seekingwhen it comes to losses. In light of this analysis, wouldyou still prefer A and D?

A common response to this example is that it iscontrived because the two pairs of investment opportu-nities were presented sequentially, not simultaneously.However, in a typical global financial institution, theLondon office may be faced with choices A and B, andthe Tokyo office may be faced with choices C and D.Locally, it may seem as if there is no right or wronganswer—the choice between A and B or C and D seemsto be simply a matter of personal risk preferences—butthe globally consolidated financial statement for theentire institution will tell a very different story. Fromthat perspective, there is a right answer and a wrongone, and the empirical and experimental evidence sug-gest that most individuals tend to select the wrong



VOL. 7, NO. 2, 2005 25

answer. Therefore, according to the behavioralists,quantitative models of efficient markets—all of whichare predicated on rational choice—are likely to bewrong as well.

These examples illustrate the most enduring critiqueof the EMH: individuals do not always behave rationally.In particular, the traditional approach to modeling behav-ior in economics and finance is to assert that investorsoptimize additive time-separable expected utility func-tions from certain parametric families, for example, con-stant relative risk aversion. This is the starting point formany quantitative models of modern finance, includingmean-variance portfolio theory and the Sharpe-LintnerCapital Asset Pricing Model. However, a number of stud-ies have shown that human decision making does notseem to conform to rationality and market efficiency butexhibits certain behavioral biases that are clearlycounterproductive from the financial perspective; forexample, overconfidence (Fischoff and Slovic, 1980;Barber and Odean, 2001; Gervais and Odean, 2001),overreaction (DeBondt and Thaler, 1986), loss aversion(Kahneman and Tversky, 1979; Shefrin and Statman,1985, 2000; Odean, 1998), herding (Huberman andRegev, 2001), psychological accounting (Tversky andKahneman, 1981), miscalibration of probabilities(Lichtenstein et al., 1982), hyperbolic discounting(Laibson, 1997), and regret (Bell, 1982; Clarke et al.,1994). For these reasons, behavioral economists con-clude that investors are often—if not always—irrational,exhibiting predictable and financially ruinous behaviorthat is unlikely to yield efficient markets.3

Grossman (1976) and Grossman and Stiglitz(1980) go even further. They argue that perfectly infor-mationally efficient markets are an impossibility, for ifmarkets are perfectly efficient, there is no profit to gath-ering information, in which case there would be littlereason to trade and markets eventually would collapse.Alternatively, the degree of market inefficiency deter-mines the effort investors are willing to expend to gath-er and trade on information; hence, a nondegeneratemarket equilibrium will arise only when there are suffi-cient profit opportunities, that is, inefficiencies, to com-pensate investors for the costs of trading andinformation gathering. The profits earned by theseattentive investors may be viewed as “economic rents”

that accrue to those willing to engage in such activities.Who are the providers of these rents? Black (1986) gaveus a provocative answer: “noise traders,” or individualswho trade on what they consider to be information butwhich is, in fact, merely noise. But if informed tradersare constantly taking advantage of noise traders, how donoise traders persist? The answer is a slightly modifiedversion of P. T. Barnum’s well-known dictum: “A noise-trader is born every minute.”4

Rational Finance Responds

The supporters of the EMH have responded tothese challenges by arguing that while behavioralbiases and corresponding inefficiencies do exist

from time to time, there is a limit to their relevance andimpact because of opposing forces dedicated to exploit-ing such opportunities.5 A simple example of such alimit is the so-called “Dutch Book,” in which irrationalprobability beliefs can result in guaranteed profits for thesavvy arbitrageur. Consider, for example, an event E,defined as “the S&P 500 index drops by 5 percent ormore next Monday,” and suppose an individual has thefollowing irrational beliefs: there is a 50-percent proba-bility that E will occur and a 75-percent probability thatE will not occur. This is clearly a violation of one of thebasic axioms of probability theory—the probabilities oftwo mutually exclusive and exhaustive events must sumto one—but many experimental studies have document-ed such violations among most human subjects.

These inconsistent subjective probability beliefsimply that the individual would be willing to take bothof the following bets B

1and B

2:

where Ec denotes the event “not E.” Now suppose wetake the opposite side of both bets, placing $50 on B

1

and $25 on B2. If E occurs, we lose $50 on B

1but gain

$75 on B2, yielding a profit of $25. If Ec occurs, we gain

$50 on B1 and lose $25 on B

2, also yielding a profit of

$25. Regardless of the outcome, we have secured aprofit of $25, an arbitrage that comes at the expense ofthe individual with inconsistent probability beliefs.Such beliefs are not sustainable, and market forces—

B1 = { $1 if E ,

–$1 otherwise B2 = { $1 if E c

–$3 otherwise


namely, arbitrageurs such as hedge funds and propri-etary trading groups—will take advantage of theseopportunities until they no longer exist; that is, until theodds are in line with the axioms of probability theory.6

Therefore, proponents of the EMH argue that there arelimits to the degree and persistence of behavioral biasessuch as inconsistent probability beliefs, given the sub-stantial incentives for identifying and exploiting suchoccurrences. While all of us are subject to certain behav-ioral biases from time to time, EMH disciples argue thatmarket forces will always act to bring prices back torational levels, implying that the impact of irrationalbehavior on financial markets is generally negligibleand, therefore, irrelevant.

But this last conclusion relies on the assumptionthat market forces are sufficiently powerful to overcomeany type of behavioral bias, or equivalently, that irra-tional beliefs are not so pervasive as to overwhelm thecapacity of arbitrage capital dedicated to taking advan-tage of such irrationalities. This issue cannot be settledby theory but is an empirical matter that requires high-ly structured data analysis and statistical inference. Thequestion of market rationality can be reduced to a quan-titative comparison of the economic forces of rationalityversus the behavioral tendencies that are endemic tomost market participants.

One anecdotal type of evidence is the series offinancial manias and panics that have characterized cap-ital markets ever since the seventeenth century whentulip bulbs captured the imagination of the Dutch.From 1634 to 1636, “tulip mania” spread throughHolland, causing the price of tulip bulbs to skyrocket toridiculous levels, only to plummet precipitously after-ward, creating widespread panic and enormous finan-cial dislocation in its wake.7 Other examples includeEngland’s South Sea Bubble of 1720, the U.S. stock mar-ket crashes of October 1929 and October 1987, theJapanese real-estate bubble of the 1990s, the U.S.technology bubble of 2000, the collapse of Long-TermCapital Management and other fixed-income relative-value hedge funds in 1998, and the current real-estatebubble in England. These examples suggest that theforces of irrationality can dominate the forces of ratio-nality, even over extended periods of time (seeKindleberger, 1989, for additional examples).

What, then, does this imply for the practice ofinvestment management, much of which is based on theEMH framework? Before turning to this issue, we needto digress briefly to develop a better understanding ofthe ultimate sources of behavioral biases.

A Neurosciences Perspective

Because much of the debate surrounding the EMHand its behavioral exceptions centers on rational-ity in human behavior, we look to the recent lit-

erature in the cognitive neurosciences for additionalinsights. A number of recent breakthroughs link behav-ior with specific brain functions, and this research hasled to a significant reformulation of psychological andneurophysiological models of decision making. Many ofthese breakthroughs have come from new research toolsin the neurosciences such as positron emission tomog-raphy (PET) and functional magnetic resonance imag-ing (fMRI), where sequences of images of a subject’sbrain are captured in real time while the subject is askedto perform a certain task. By comparing the amount ofblood flow to different parts of the brain before, during,and after the task, it is possible to detect higher levels ofactivation in certain regions of the brain, thus associat-ing the performance of the task with those regions.These technologies have revolutionized much of psy-chological research, providing important neurophysio-logical foundations for a variety of cognitive processesand patterns of behavior.8

One example that is especially relevant for financialdecision making is the apparent link between rationalbehavior and emotion. Until recently, the two were con-sidered diametrical opposites, but by studying thebehavior of patients who lost the ability to experienceemotion after the surgical removal of brain tumors,Damasio (1994) discovered that their ability to makerational choices suffered as well. One patient—code-named Elliot—who lost his emotional faculties afterhaving a portion of his frontal lobe removed, experi-enced a surprisingly profound effect on his day-to-dayactivities, as Damasio (1994, p. 36) describes:

When the job called for interrupting an activity andturning to another, he might persist nonetheless, seem-ingly losing sight of his main goal. Or he might inter-rupt the activity he had engaged, to turn to something



VOL. 7, NO. 2, 2005 27

he found more captivating at that particularmoment... The flow of work was stopped. One mightsay that the particular step of the task at which Elliotbalked was actually being carried out too well, and atthe expense of the overall purpose. One might say thatElliot had become irrational concerning the largerframe of behavior...

Apparently, Elliot’s inability to feel—his lack ofemotional response—somehow caused him to makeirrational choices in his daily decisions.

This conclusion surprises many economists becauseof the association between emotion and behavioral bias-es. After all, isn’t it fear and greed, or “animal spirits,” asKeynes once suggested, that cause prices to deviate irra-tionally from fundamentals? In fact, a more sophisticatedview of the role of emotions in human cognition is thatthey are central to rationality (see, for example, Damasio,1994, and Rolls 1990, 1992, 1994, 1999).9 Emotions arethe basis for a reward-and-punishment system that facil-itates the selection of advantageous behavior, providing a

kind of mental yardstick for animals to measure the costsand benefits of the various actions open to them (Rolls,1999, ch. 10.3). Even fear and greed—the two mostcommon culprits in the downfall of rational thinking,according to most behavioralists—are the product of evo-lutionary forces, adaptive traits that increase the proba-bility of survival. From an evolutionary perspective,emotion is a powerful tool for improving the efficiencywith which humans learn from their environment andtheir past. When an individual’s ability to experienceemotion is eliminated, an important feedback loop is sev-ered and the decision-making process is impaired.

What, then, is the source of irrationality, if not emo-tion? The neurosciences literature provides some hints,from which we can craft a conjecture. The starting pointis a basic fact about the brain: it is not a homogeneousmass of nerve cells but a collection of specialized compo-nents, many of which have been identified with particu-lar functions and types of behavior. For example, thebrainstem, which is located at the base of the brain and sits

FIGURE 1

THE TRIUNE BRAIN MODEL OF MACLEAN (1990)

Corpus Callosum

Cerebellum

Brain Stem

Reptilian Complex

(Old Brain)

Limbic System

(Mammalian or Mid Brain)

Cerebrum

(Neo Cortex or New Brain)

Illustration adapted from Caine and Caine, 1990.


on top of the spinal cord (see figure 1), controls the mostbasic bodily functions such as breathing and heartbeatand is active even during deep sleep. The limbic system,which comprises several regions in the middle of thebrain, is responsible for emotions, instincts, and socialbehavior such as feeding, fight-or-flight responses, andsexuality. And the cerebral cortex, which is the tangledmaze of gray matter that forms the outer layer of the brain,is what allows us to think complex and abstract thoughtsand where language and musical abilities, logical reason-ing, learning, long-term planning, and sentience reside.These three areas form the triune brain model, proposedby MacLean (1990) and illustrated in figure 1; he refers tothem as the reptilian, mammalian, and hominid brains,respectively. This terminology underscores his hypothesisthat the human brain is the outcome of an evolutionaryprocess in which basic survival functions appeared first,more advanced social behavior came second, and unique-ly human cognitive abilities emerged most recently (thatis, within the past 100,000 years).

The relevance of the triune brain model for ourpurposes is that it provides a deeper foundation forsome of the behavioral biases that affect financial deci-sion making. In particular, behavior may be viewed asthe observable manifestation of interactions among sev-eral components of the brain, sometimes competitivelyand other times cooperatively. For example, the urge toflee from danger may be triggered in the mammalianbrain by an approaching stranger in a dimly lit corridor,but this urge may be overridden by the hominid brain,which understands that the dim lighting is due to a tem-porary power outage and that the approaching strangeris wearing a policeman’s uniform and therefore unlikelyto represent an immediate threat. However, under othercircumstances, emotional responses can overrule morecomplex deliberations; for example, in the midst of abar fight in which there is likely to be real physical dan-ger. Indeed, neuroscientists have shown that emotion isthe “first response” in the sense that individuals exhibitemotional reactions to objects and events far quickerthan they can articulate what those objects and eventsare (see Zajonc, 1980, 1984).

Moreover, it is well known that the primacy ofemotional reactions—especially fear and the fight-or-flight response—can identify danger well ahead of the

conscious mind (see de Becker, 1997). In fact, extremeemotional reactions can short-circuit rational delibera-tion altogether (see Baumeister, Heatherton, and Tice,1994); that is, strong stimulus to the mammalian brainseems to inhibit activity in the hominid brain. From anevolutionary standpoint, this seems quite sensible—emotional reactions are a call-to-arms that should beheeded immediately because survival may depend on it,and higher brain functions such as language and logicalreasoning are suppressed until the threat is over, that is,until the emotional reaction subsides.

In our current environment, however, many threatsidentified by the mammalian brain are not, in fact, lifethreatening, yet our physiological reactions may still bethe same. In such cases, the suppression of our hominidbrain may be unnecessary and possibly counter-productive, and this is implicitly acknowledged in thecommon advice to refrain from making any significantdecisions after experiencing the death of a loved one ora similar emotional trauma. This is sage advice, for theability to think straight is genuinely hampered physio-logically by extreme emotional reactions.10

The complexity of the interactions among the threedivisions of the brain may be illustrated by two exam-ples. The first involves the difference between a naturalsmile and a forced smile (see Damasio, 1994, pp.141–143 and fig. 7-3), which is easily detected by mostof us, but why? The answer lies in the fact that a natur-al smile is generated by the mammalian brain(specifically, the anterior cingulate) and involves certaininvoluntary facial muscles that are not under the controlof the hominid brain. The forced smile, however, is apurely voluntary behavior emanating from the hominidbrain (the motor cortex), and does not look exactly thesame because involuntary muscles do not participate inthis action. In fact, it takes great effort and skill to gen-erate particular facial expressions on cue, as actorstrained in the “method school” can attest—only by con-juring up emotionally charged experiences in their pastsare they able to produce the kind of genuine emotionalreactions needed in a given scene, and anything lessauthentic is immediately recognized as bad acting.

The second example is from a study by Eisenberger,Lieberman, and Williams (2003) in which they deliber-ately induced feelings of social rejection among a group



VOL. 7, NO. 2, 2005 29

of subjects and then identified the regions of the brainthat were most activated during the stimulus. They dis-covered that two components were involved, the anteri-or cingulate and the insula, both of which are alsoknown to process physical pain. In other words, emo-tional trauma—hurt feelings, emotional loss, embarrass-ment, and shame—can generate the same kind of neuralresponse that a broken bone does. Many who have expe-rienced the death of a loved one have commented thatthey felt physical pain from their losses despite the factthat no physical trauma was involved, and we are nowbeginning to develop a neuroscientific basis for this phe-nomenon. Eisenberger, Lieberman, and Williams (2003,p. 292) conclude that “...social pain is analogous in itsneurocognitive function to physical pain, alerting uswhen we have sustained injury to our social connec-tions, allowing restorative measures to be taken.”

These two examples illustrate some of the manyways in which specialized components in the brain caninteract to produce behavior. The first example showsthat two different components of the brain are capableof producing the same outcome: a smile. The secondexample shows that the same components can beinvolved in producing two different outcomes: physicalpain and emotional pain. The point of specialization inbrain function is increased fitness in the evolutionarysense. Each specialized component may be viewed as anevolutionary adaptation designed to increase thechances of survival in response to a particular environ-mental condition. As environmental conditions change,so too does the relative importance of each component.One of the unique features of Homo sapiens is the abili-ty to adapt to new situations by learning and imple-menting more advantageous behavior, and this is oftenaccomplished by several components of the brain actingtogether. As a result, what economists call “preferences”are often complicated interactions among the threecomponents of the brain as well as interactions amongsubcomponents within each of the three.

This perspective implies that preferences may not bestable through time and are likely to be shaped by a num-ber of factors, both internal and external to the individ-ual; that is, factors related to the individual’s personalityand factors related to the individual’s specific environ-mental conditions. When the environmental conditions

shift, we should expect behavior to change in response,both through learning and, over time, through changes inpreferences via the forces of natural selection. These evo-lutionary underpinnings are more than simple specula-tion in the context of financial market participants. Theextraordinary degree of competitiveness of global finan-cial markets and the outsized rewards that accrue to the“fittest” participants suggest that Darwinian selection is atwork in determining the typical profile of the successfulinvestor. After all, unsuccessful market participants areeventually eliminated from the population after sufferinga certain level of losses.

This neuroscientific perspective suggests an alter-native to the EMH, one in which market forces and pref-erences interact to yield a much more dynamiceconomy, one driven by competition, natural selection,and the diversity of individual and institutional behav-ior. This is the essence of the AMH.

The Adaptive Markets Hypothesis

The application of evolutionary ideas to economicbehavior is not new. Students of the history ofeconomic thought will recall that Thomas

Malthus used biological arguments—the fact that popu-lations increase at geometric rates whereas naturalresources increase at only arithmetic rates—to arrive atrather dire economic consequences, and that bothDarwin and Wallace were influenced by these arguments(see Hirshleifer, 1977, for further details). Also, JosephSchumpeter’s (1939) view of business cycles, entrepre-neurs, and capitalism have an unmistakeable evolution-ary flavor to them; in fact, his notions of “creativedestruction” and “bursts” of entrepreneurial activity aresimilar in spirit to natural selection and Eldredge andGould’s (1972) notion of “punctuated equilibrium.”However, Wilson (1975) was among the first to system-atically apply the principles of competition, reproduc-tion, and natural selection to social interactions, yieldingsurprisingly compelling explanations for certain kinds ofhuman behavior, for example, altruism, fairness, kinselection, language, mate selection, religion, morality,ethics, and abstract thought.11 Wilson aptly named thisnew field “sociobiology,” and its debut generated a con-siderable degree of controversy because of some of itspossible implications for social engineering and eugenics.


These ideas recently have been exported to a num-ber of economic and financial contexts,12 and at leasttwo prominent practitioners have proposed Darwinianalternatives to the EMH. In a chapter titled “TheEcology of Markets,” Niederhoffer (1997, ch. 15) likensfinancial markets to an ecosystem with dealers as herbi-vores, speculators as carnivores, and floor traders anddistressed investors as decomposers. And Bernstein(1998) makes a compelling case for active managementby pointing out that the notion of equilibrium, which iscentral to the EMH, is rarely realized in practice andthat market dynamics are better explained by evolu-tionary processes.

Clearly the time has come foran evolutionary alternative to mar-ket efficiency, and this is the direc-tion taken in Farmer and Lo(1999), Farmer (2002), and Lo(2002, 2004). In this section, I pro-vide a brief summary of the AMH(Lo, 2004).

Contrary to the neoclassicalpostulate that individuals maxi-mize expected utility and haverational expectations, an evolu-tionary perspective makes consid-erably more modest claims,viewing individuals as organismsthat have been honed—throughgenerations of natural selection—to maximize the survival of theirgenetic material (see, for example,Dawkins, 1976). This perspectiveimplies that behavior is not necessarily intrinsic andexogenous but evolves by natural selection and dependson the particular environment through which selectionoccurs. That is, natural selection operates not only upongenetic material, but upon biology (recall the special-ized components of the triune brain model) and alsosocial behavior and cultural norms in Homo sapiens.

To operationalize this perspective within an eco-nomic context, Lo (2004) revisited the idea of “bound-ed rationality” first espoused by economist HerbertSimon, who won the Nobel Memorial Prize inEconomic Sciences in 1978. Simon (1955) suggested

that individuals are hardly capable of the kind of opti-mization that neoclassical economics calls for in thestandard theory of consumer choice. Instead, he arguedthat because optimization is costly and humans are nat-urally limited in their computational abilities, theyengage in something he called “satisficing,” an alterna-tive to optimization in which individuals make choicesthat are merely satisfactory, not necessarily optimal. Inother words, individuals are bounded in their degree ofrationality, which is in sharp contrast to the currentorthodoxy of rational expectations, where individualshave unbounded rationality (the term “hyperrational

expectations” might be moredescriptive). Unfortunately,although this idea garnered aNobel Memorial Prize for Simon, ithas had relatively little impact onthe economics profession untilrecently,13 partly because of thenear-religious devotion to rational-ity on the part of the economicsmainstream but also because ofone specific criticism leveledagainst satisficing: What deter-mines the point at which an indi-vidual stops optimizing andreaches a satisfactory solution? Ifsuch a point is determined by theusual cost–benefit calculationunderlying much of microeco-nomics (that is, optimize until themarginal benefit of the optimumequals the marginal cost of getting

there), this assumes the optimal solution is known,which eliminates the need for satisficing. As a result, theidea of bounded rationality fell by the wayside, andrational expectations has become the de facto standardfor modeling economic behavior under uncertainty.

Lo (2004) argues that an evolutionary perspectiveprovides the missing ingredient in Simon’s framework.The proper response to the question of how individualsdetermine the point at which their optimizing behavioris satisfactory is this: Such points are determined notanalytically, but through trial and error and, of course,natural selection. Individuals make choices based on


Lo argues that an evolutionary

perspective provides the missing

ingredient in Simon’s framework.

The proper response to the

question of how individuals

determine the point at which

their optimizing behavior is

satisfactory is this: Such points

are determined not analytically,

but through trial and error and,

of course, natural selection.


VOL. 7, NO. 2, 2005 31

experience and their best guesses as to what might beoptimal, and they learn by receiving positive or negativereinforcement from the outcomes. If they receive nosuch reinforcement, they do not learn. In this fashion,individuals develop heuristics to solve various econom-ic challenges, and as long as those challenges remainstable, the heuristics eventually will adapt to yieldapproximately optimal solutions.

If, on the other hand, the environment changes,then it should come as no surprise that the heuristics ofthe old environment are not necessarily suited to thenew. In such cases, we observe behavioral biases—actions that are apparently ill advised in the context inwhich we observe them. But rather than labeling suchbehavior irrational, we should recognize that subopti-mal behavior is likely when we take heuristics out oftheir evolutionary context. A more accurate term forsuch behavior might be “maladaptive.” The flopping ofa fish on dry land may seem strange and unproductive,but under water, the same motions propel the fish awayfrom its predators. And the antagonistic effect of humanemotional reactions on logical reasoning described ear-lier is maladaptive for many financial contexts.

By coupling Simon’s notion of bounded rationalityand satisficing with evolutionary dynamics, many otheraspects of economic behavior also can be derived.Competition, cooperation, market-making behavior,general equilibrium, and disequilibrium dynamics areall adaptations designed to address certain environ-mental challenges for the human species, and by view-ing them through the lens of evolutionary biology, wecan better understand the apparent contradictionsbetween the EMH and the presence and persistence ofbehavioral biases.

Specifically, the AMH can be viewed as a new ver-sion of the EMH, derived from evolutionary principles.The primary components of the AMH consist of the fol-lowing ideas:

(A1) Individuals act in their own self-interest.(A2) Individuals make mistakes.(A3) Individuals learn and adapt.(A4) Competition drives adaptation and innovation.(A5) Natural selection shapes market ecology.(A6) Evolution determines market dynamics.

The EMH and AMH have a common starting point inA1, but the two paradigms part company in A2 and A3.In efficient markets, investors do not make mistakes,nor is there any learning and adaptation because themarket environment is stationary and always in equilib-rium. In the AMH framework, mistakes occur frequent-ly, but individuals are capable of learning from mistakesand adapting their behavior accordingly. However, A4states that adaptation does not occur independently ofmarket forces but is driven by competition, that is, thepush for survival. The interactions among various mar-ket participants are governed by natural selection—thesurvival of the richest, in our context—and A5 impliesthat the current market environment is a product of thisselection process. A6 states that the sum total of thesecomponents—selfish individuals, competition, adapta-tion, natural selection, and environmental conditions—is what we observe as market dynamics.

For example, prices reflect as much information asdictated by the combination of environmental condi-tions and the number and nature of species in the econ-omy or, to use the appropriate biological term, theecology. By species, I mean distinct groups of marketparticipants, each behaving in a common manner. Forexample, pension funds may be considered one species;retail investors, another; market makers, a third; andhedge fund managers, a fourth. If multiple species (orthe members of a single highly populous species) arecompeting for rather scarce resources within a singlemarket, that market is likely to be highly efficient; forexample, the market for ten-year U.S. Treasury notes,which reflects most relevant information very quicklyindeed. If, on the other hand, a small number of speciesare competing for rather abundant resources in a givenmarket, that market will be less efficient; for example,the market for oil paintings from the ItalianRenaissance. Market efficiency cannot be evaluated in avacuum because it is highly context dependent anddynamic, just as insect populations advance and declineas a function of the seasons, the number of predatorsand prey they face, and their abilities to adapt to anever-changing environment.

The profit opportunities in any given market areakin to the amount of natural resources in a particularlocal ecology—the more resources present, the less


fierce the competition. As competition increases, eitherbecause of dwindling food supplies or an increase in theanimal population, resources are depleted, which, inturn, eventually causes a population decline, therebydecreasing the level of competition and starting thecycle all over again. In some cases, cycles converge tocorner solutions; that is, certain species become extinct,food sources are permanently exhausted, or environ-mental conditions shift dramatically. However, a keyinsight of the AMH—taken directly from evolutionarybiology—is that convergence to equilibrium is neitherguaranteed nor likely to occur at any point in time. Thenotion that evolving systems must march inexorablytoward some ideal stationary state is plain wrong.14 Inmany cases, such equilibria do not exist, and even whenthey do, convergence rates may be exceedingly slow,rendering the limiting equilibria virtually irrelevant forall practical purposes. The determinants of cycles versusconvergence are, ultimately, the combination of marketparticipants and natural resources in the market ecolo-gy. By viewing economic profits as the ultimate foodsource on which market participants depend for theirsurvival, the dynamics of market interactions and finan-cial innovation can be readily derived.

Under the AMH, behavioral biases abound. Theorigins of such biases are, in many cases, heuristics thatare adaptations to nonfinancial contexts, and theirimpact is determined by the size of the population withsuch biases versus the size of competing populationswith more beneficial heuristics; that is, heuristics thatare more conducive to economic success. During theFall of 1998, the desire for liquidity and safety by a cer-tain population of investors overwhelmed the popula-tion of hedge funds attempting to arbitrage suchpreferences, causing those arbitrage relations to breakdown. However, in the years before August 1998, fixed-income relative-value traders profited handsomely fromthese activities, presumably at the expense of individu-als with seemingly irrational preferences. In fact, suchpreferences were shaped by a certain set of evolutionaryforces and might have been quite rational in other envi-ronmental conditions. Therefore, under the AMH,investment strategies undergo cycles of profitability andloss in response to changing business conditions, thenumber of competitors entering and exiting the industry,

and the type and magnitude of profit opportunities. Asopportunities shift, so too will the affected populations.For example, after 1998, the number of fixed-incomerelative-value hedge funds declined dramatically—because of outright failures, investor redemptions, andfewer startups in this sector—but many have reap-peared in recent years as performance for this type ofinvestment strategy has improved.

Applications

The new paradigm of the AMH is still in its infan-cy and certainly requires a great deal moreresearch before it becomes a viable alternative to

the EMH. It is already clear, however, that an evolution-ary framework can reconcile many of the apparent con-tradictions between efficient markets and behavioralexceptions. The former may be viewed as the steady-state limit of a population with constant environmentalconditions, and the latter may be viewed as specificadaptations of certain groups that may or may not per-sist, depending on the particular evolutionary paths thatthe economy experiences.

Apart from this intellectual reconciliation, there isstill the question of how relevant the AMH is for thepractice of investment management. After all, despitethe limitations of the EMH, it has given rise to a wealthof quantitative tools for the practitioner. Part of thistreasure trove of applications comes from the EMH’smuch longer history—behavioral models have onlyrecently begun to gain some degree of respectability inthe academic mainstream. Moreover, the internal con-sistency and logical elegance of the EMH framework arealmost hypnotic, and it is all too easy to forget that theEMH is merely a figment of our imagination, meant toserve as an approximation—and not always a terriblyaccurate one—to a far more complex reality. Unlike thelaw of gravity and the theory of special relativity, thereare no immutable laws of nature from which the EMHhas been derived.

Also, once we depart from the highly structuredframework of the EMH, there are seemingly endless pos-sibilities for modeling economic behavior. This shouldnot dissuade us from the quest to derive mathematicalembodiments of behavioral research, however; other-wise we risk becoming the drunkard searching for his



VOL. 7, NO. 2, 2005 33

lost keys outside the bar where he left them, just becausethe lighting is better in the street. In particular, quantita-tive implications of the AMH may be derived through acombination of deductive and inductive inference—forexample, theoretical analysis of evolutionary dynamics,empirical analysis of evolutionary forces in financialmarkets, and experimental analysis of decision makingat the individual and group level—and are currentlyunder investigation. But even at this formative stage, theAMH yields several concrete appli-cations for investment manage-ment and consulting.

Preferences Matter

Perhaps the most immediateapplication is to individual and

institutional risk preferences.Despite their usefulness in othercontexts, the heuristics that manypsychologists and economists havedocumented as behavioral biasesare often counterproductive for thepurposes of building and preserv-ing financial wealth. To avoid such pitfalls, one must firstbe aware of them. Therefore, one critical set of applica-tions involves the proper measurement and managementof preferences.

The quantitative measurement of preferences has along history in psychology, economics, operationsresearch, and the new field of marketing science. Eachof these disciplines emphasizes different aspects of indi-vidual decision making: psychological surveys aredesigned to capture broad characteristics of personality(Costa and McCrae, 1992), economists perform choiceexperiments with various lotteries to elicit risk prefer-ences (MacCrimmon and Wehrung, 1986), and market-research consultants conduct field studies of consumerpreferences as inputs to product-design efforts (Urbanand Hauser, 1993). In light of the emerging neuroscien-tific view of preferences as interactions among special-ized components in the brain, the proper measurementof preferences may require a combination of each ofthese approaches.15

In particular, the kind of risk-preference question-naires used by brokerage firms and financial advisers may

be a useful starting point, but a typical subject’s respons-es are not likely to yield a stable estimate of the subject’strue financial decision-making process. For example,suppose a thirty-five-year-old subject fills out such aquestionnaire before experiencing the untimely loss of aspouse and then fills out the same questionnaire a fewmonths after the tragic event—should we expect theresponses from the two questionnaires to be identical?Despite the fact that love, family, and death are typically

not included in standard financialdecision-making paradigms such asportfolio optimization and assetallocation, the human brain doesnot necessarily compartmentalizedecisions in the same way.

More generally, Statman(2004a) observes that investorshave multifaceted objectives inmind when formulating theirinvestment decisions; hence, theeffective consultant will help theinvestor to acknowledge and artic-ulate these objectives before mak-

ing any recommendations. By developing moreencompassing survey instruments—not just risk-prefer-ence questionnaires—we may be able to develop a morecomplete, hence a more stable and predictive, model ofindividual and institutional preferences. One alternativeis to measure more fundamental aspects of an individ-ual’s personality, such as temperament, and relate thesemeasures to risk attitudes and investment decisions.16

While measuring preferences has been well stud-ied, the management of preferences is a politically sensi-tive issue, especially for economists, who tend to shyaway from most normative implications of their ideas.Because of the inherent difficulties in comparing levelsof happiness between two individuals, most economiststake individual preferences as given and actively avoidweighing one consumer’s gains against another’s losses,except in the unit of measurement defined by each indi-vidual: their own utility functions.17 “To each his own,”as the saying goes. However, one of the lessons from therecent neurosciences literature is that preferences arenot immutable or one-dimensional but the sum total ofcomplex interactions between competitive and cooper-

One alternative is to measure

more fundamental aspects of an

individual’s personality, such as

temperament, and relate these

measures to risk attitudes and

investment decisions.


ative components of the brain. What we take to be pref-erences may, in one instance, reflect largely hardwiredautonomic responses of the limbic system and, inanother instance, be the result of careful deliberationmediated by the prefrontal cortex.

This level of complexity, while challenging, holdsout the hope that preferences can be actively managedso as to produce more desirable outcomes. However,what does “more desirable” mean? In the general eco-nomic context, this phrase has little content becauseutility is presumed to be the ulti-mate objective; hence, it makes nosense to say that one can managepreferences to yield higher utilitybecause utility is itself the metricby which success is gauged. In thecontext of investment manage-ment, however, “more desirable”does have independent meaningapart from an individual’s utilityfunction; for example, a minimallevel of real wealth at retirement, aminimum probability of a pensionfund’s assets exceeding liabilitiesover a ten-year horizon, or a con-sistent investment process formaking asset allocation decisionsover time. In each of these cases,one can argue that certain types of preferences lead toless attractive outcomes with respect to these objectives.For example, loss aversion generally leads to lowerexpected real wealth at retirement than a logarithmicutility function. Therefore, if building wealth is a prior-ity for an individual or institution, actively managingthe investor’s preferences is critical.

Asset Allocation Revisited

Another application of the AMH framework toinvestment practice involves asset allocation deci-

sions, the selection of portfolio weights for broad assetclasses. An implication of the AMH is that to the extentthat a relation between risk and reward exists, it isunlikely to be stable over time. Such a relation is deter-mined by the relative sizes and preferences of variouspopulations in the market ecology as well as institu-

tional aspects such as the regulatory environment andtax laws. As these factors shift over time, any risk/reward relation is likely to be affected.

A corollary of this implication is that the equity riskpremium is not a universal constant but is time varyingand path dependent. This is not so revolutionary anidea as it might first appear to be—even in the contextof a rational-expectations equilibrium model, if riskpreferences change over time, then the equity risk pre-mium must vary too. The incremental insight of the

AMH is that aggregate risk prefer-ences are not fixed but are con-stantly being shaped and reshapedby the forces of natural selection.For example, until recently, U.S.markets were populated by asignificant group of investors whohad never experienced a genuinebear market—this fact hasundoubtedly shaped the aggregaterisk preferences of the U.S. econo-my, just as the experience of thepast four years since the technolo-gy bubble burst has affected therisk preferences of the currentpopulation of investors.18 In thiscontext, natural selection deter-mines who participates in market

interactions; those investors who experienced substan-tial losses in the technology bubble are more likely tohave exited the market, leaving a markedly differentpopulation of investors today than four years ago.Through the forces of natural selection, history matters.Irrespective of whether prices fully reflect all availableinformation, the particular path that market prices havetaken over the past few years influences current aggre-gate risk preferences.

A second implication is that contrary to the classi-cal EMH, arbitrage opportunities do exist from time totime in the AMH. As Grossman and Stiglitz (1980)observed, without such opportunities, there will be noincentive to gather information and the price-discoveryaspect of financial markets will collapse. From an evo-lutionary perspective, the existence of active liquidfinancial markets implies that profit opportunities must


What we take to be

preferences may, in one

instance, reflect largely

hardwired autonomic responses

of the limbic system and,

in another instance,

be the result of careful

deliberation mediated by

the prefrontal cortex.


VOL. 7, NO. 2, 2005 35

be present. As they are exploited, they disappear. Butnew opportunities also are being created continually ascertain species die out, as others are born, and as insti-tutions and business conditions change. Rather than theinexorable trend toward higher efficiency predicted bythe EMH, the AMH implies considerably more complexmarket dynamics, with cycles as well as trends, andpanics, manias, bubbles, crashes, and other phenomenathat are routinely witnessed in natural market ecologies.These dynamics provide the motivation for active man-agement as Bernstein (1998) suggests and give rise toNiederhoffer’s (1997) carnivores and decomposers.

A third implication is that investment strategies willalso wax and wane, performing well in certain environ-ments and performing poorly in other environments.Contrary to the classical EMH in which arbitrage oppor-tunities are competed away, eventually eliminating theprofitability of the strategy designed to exploit the arbi-trage, the AMH implies that such strategies may declinefor a time and then return to profitability when envi-

ronmental conditions become more conducive to suchtrades. An obvious example is risk arbitrage, which hasbeen unprofitable for several years because of thedecline in investment banking activity since 2001.However, as the pace of mergers and acquisitions beginsto pick up again, risk arbitrage will start to regain itspopularity among both investors and portfolio man-agers, as it has just recently.

A more striking example can be found by comput-ing the rolling first-order autocorrelation ρ̂

1of monthly

returns of the S&P Composite Index from January1871 to April 2003 (see figure 2). As a measure of mar-ket efficiency (recall that the Random Walk Hypothesisimplies that returns are serially uncorrelated, hence ρ

1

should be 0 in theory), ρ̂1

might be expected to take onlarger values during the early part of the sample andbecome progressively smaller during recent years as theU.S. equity market becomes more efficient. It is apparentfrom figure 2, however, that the degree of efficiency—asmeasured by the first-order autocorrelation—varies

FIGURE 2

ROLLING 5-YEAR SERIAL CORRELATION COEFFICIENT OF THE S&P COMPOSITE INDEX

January 1871 to April 2003

(Data Source: R. Shiller)


through time in a cyclical fashion,and there are periods in the 1950swhen the market was more effi-cient than in the early 1990s.

Such cycles are not ruled outby the EMH in theory, but in prac-tice none of its existing empiricalimplementations have incorporat-ed these dynamics, assuminginstead a stationary world in whichmarkets are perpetually in equilib-rium. This widening gulf betweenthe stationary EMH and obviousshifts in market conditions nodoubt contributed to Bernstein’s(2003) recent critique of the policyportfolio in strategic asset alloca-tion models and his controversialproposal to reconsider the case fortactical asset allocation.

A final implication of the AMH for asset allocation isthat characteristics such as value and growth may behavelike risk factors from time to time; that is, portfolios withsuch characteristics may yield higher expected returnsduring periods when those attributes are in favor. Forexample, during the U.S. technology bubble of the late1990s, growth stocks garnered higher expected returnsthan value stocks, only to reverse after the bubble burst.While such nonstationarities create difficulties for theEMH—after all, in that framework, a characteristic iseither a risk factor or it is not—the AMH places norestrictions on what can or cannot be a risk factor.Whether or not a particular characteristic is priceddepends on the nature of the population of investors at agiven point in time; if a significant fraction of investorsprefers growth stocks over others, a growth-factor riskpremium arises. As the number of growth-orientedinvestors dwindles—either because they retire and with-draw their wealth from the stock market or because anew generation of investors enters the stock market withits own preferences—the growth premium declines, andother characteristics may emerge in its place.

The natural question that follows is this: How do wedetermine which characteristics are priced and which arenot? Although the AMH does not yet offer quantitative

methods for answering this ques-tion, some qualitative guidelinesare available. The main determi-nants of any factor risk premiumrevolve around the preferences ofmarket participants and how theyinteract with the natural resourcesof the market ecology. This suggestsa specific research agenda for iden-tifying and measuring priced fac-tors that coincide with the study ofany complex natural ecosystem:Construct summary measures ofthe cross-section of preferences ofcurrent investors and the popula-tion sizes of investors with eachtype of preference, develop a par-simonious but complete descrip-tion of the market environment inwhich these investors are interact-

ing (including natural resources, current environmentalconditions, and institutional contexts such as marketmicrostructure, legal and regulatory restrictions, and taxeffects), and specify the dimensions along which compe-tition, innovation, and natural selection operate. Armedwith these data, it should be possible to tell which char-acteristics might become risk factors under what kinds ofmarket conditions. Of course, gathering this kind of datais unprecedented in economics and finance and likely tobe a nontrivial undertaking. However, this may very wellbe the price of progress within the AMH framework.

The bottom line is that active asset allocation policiesmay be appropriate for certain investors and under cer-tain market conditions. If the investor populationchanges, if investors’ preferences change, if environmen-tal conditions change, and if these changes can be mea-sured in a meaningful way, then it indeed is possible toconstruct actively managed portfolios that are better ableto meet an investor’s financial objectives. Whether or notthe cost–benefit analysis supports active management is,of course, a different question that depends as much onbusiness conditions in the investment managementindustry—fee structures, total size of assets to be man-aged, and the amount of competition for assets and man-agerial talent—as it does on the magnitude and nature of


The main determinants of any

factor risk premium revolve

around the preferences of

market participants and how

they interact with the natural

resources of the market ecology.

This suggests a specific research

agenda for identifying and

measuring priced factors that

coincide with the study of any

complex natural ecosystem . . .


VOL. 7, NO. 2, 2005 37

the portfolio manager’s skills. Moreover, measuring thekind of changes described here is no mean feat, especial-ly given the lack of relevant data such as investor prefer-ences, demographics, and the cross-sectional distributionof stock holdings. The AMH does not imply that assetallocation is any less challenging, but it does provide arationale for the apparent cyclical nature of risk factorsand points the way to promising new research directions.

The Dynamics of Competition and Market Ecology

One final application of the AMH framework toinvestment management is the insight that innova-

tion is the key to survival. The EMH suggests that certainlevels of expected returns can be achieved simply by bear-ing a sufficient degree of risk. The AMH implies that therisk/reward relation varies through time and that a betterway of achieving a consistent level of expected returns isto adapt to changing market conditions. Consider thecurrent theory of the demise of the dinosaurs from a killerasteroid (Alvarez, 1997) and ask where the next financialasteroid might come from. The AMH has a clear implica-tion for all financial market participants: survival is ulti-mately the only objective that matters. While profitmaximization, utility maximization, and general equilib-rium are certainly relevant aspects of market ecology, theorganizing principle in deter-mining the evolution of mar-kets and financial technologyis simply survival.

The natural tendency ofall organisms to push for sur-vival requires both managersand consultants to maintain acertain degree of breadth anddiversity in their skills and focus. By evolving a multi-plicity of capabilities that are suited to a variety of envi-ronmental conditions, investment managers are lesslikely to become extinct as a result of rapid changes inthose conditions. And by acknowledging that changesin business conditions can influence both investmentperformance and investor preferences, consultants willbe better prepared to provide the kind of advice andsupport that will be of most value to their clients. Man-agers have to innovate constantly to stay ahead of thecompetition, and consultants need to innovate as well.

Like motherhood and apple pie, innovation is aneasy concept to embrace. In the context of the AMH,however, it takes on an urgency that is generally miss-ing from most financial decision-making paradigmssuch as the EMH, modern portfolio theory, and theCAPM. Innovation and adaptability are the primary dri-vers of survival; hence, a certain flexibility and open-mindedness to change can mean the difference betweensurvival and extinction in financial markets.

The Evolving Role of the Consultant

Skeptics have sometimes facetiously discounted therole of consultants as glorified scapegoats for spon-

sors of underperforming pension funds. However, asindependent third parties, investment management con-sultants are ideally positioned to play a central role in theasset management industry. Within the context of theAMH and behavioral finance, the consultant can provideseveral valuable services that are currently not available.

First, the consultant can assist managers andinvestors in dealing with preferences in a more seriousfashion. Instead of simply matching an investmentproduct with a buyer, the consultant can offer three far more valuable services: (1) educating investors andmanagers about preferences, expectations, and poten-

tially detrimental behavioralbiases; (2) assisting investorsin articulating, criticallyexamining and, if necessary,modifying their risk prefer-ences to suit their statedinvestment objectives, con-straints, and current marketconditions; and (3) matching

an investment manager’s preferred investment processwith an investor’s suitably modified risk preferences.

These new services may be more challenging thanthey seem. In many cases, consultants will be the bearersof bad news—no one enjoys being told that one’s prefer-ences are inconsistent with one’s objectives, and that onemay need to alter expectations to avoid being disap-pointed. This is the essence of a fiduciary’s responsibility,however: to have the client’s best interests in mind, andhow else can a client’s best interests be determined exceptthrough a deep understanding of his or her preferences?

The AMH has a clear implication for all

financial market participants: survival is

ultimately the only objective that matters.


Of course, the very best consultants already do this in aninformal and intuitive manner by investing time andeffort in establishing long-term relationships with theirclients and their managers. However, a more systematicapproach using the latest innovations in psychologicaltesting and investment technology is likely to yield evenmore significant benefits and for a broader set of consul-tants and their clients.

Second, by being sensitive to the changing natureof financial markets and the ebb and flow of investmentproducts and market cycles, the consultant will be in a better position to support and advise clients.Unfortunately, this means that traditional tools such asmean-variance optimization and risk budgeting are nolonger sufficient for addressing a pension plan’s con-cerns, and more dynamic models that adapt to chang-ing market conditions are needed (see Ang andBekaert, 2004, for a recent example). Also, the consul-tant must be familiar with a wider spectrum of invest-ment products and services and must monitor not onlycurrent performance characteristics but also the cycli-cal nature of each asset class and how it relates to cur-rent business conditions.

Third, to achieve these lofty goals, the consultantmust undertake continuing education and training to beat the forefront of investment theory and practice. Suchtraining will include not only financial technology butalso the latest advances in the cognitive neurosciences—at least to the extent that such advances are relevant tothe financial decision-making process—and new meth-ods for implementing investment policies under time-varying preferences and business conditions. Althoughsuch training may seem unrealistically demanding—especially because some of the relevant tools and meth-ods for implementing behavioral models such as AMHhave yet to be developed—significant benefits to theconsulting community could accrue almost immediatelyby including some basic information about investor psy-chology and dynamic asset allocation models in existingtraining and certification programs such as the CertifiedInvestment Management Analyst (CIMA® ) programoffered by the Investment Management ConsultantsAssociation. Over time, as some of these ideas provetheir worth in the investment community, the pace ofresearch and development will quicken considerably,

and more applications of behavioral research to invest-ment management will become available.

This ambitious agenda does raise the bar for theinvestment management industry, and it implies a moreactive and intimate role for consultants than the statusquo. In fact, one could argue that the consultant’s newrole is not unlike the role of a psychotherapist, and thisanalogy may not be completely inappropriate. Unlikebasic service providers of a homogeneous commodity,consultants are typically not dealing with one-size-fits-allinvestment products. A much deeper understanding ofeach client’s objectives, constraints, and perspectives—inshort, the client’s thought processes—is necessary for aconsultant to dispense the appropriate advice. Such anunderstanding can be developed only through a rela-tionship of trust, not unlike the relationship betweenpatient and therapist. Of course, investment managersmust have some understanding of these issues as well,but the current division of labor suggests that consul-tants—as disinterested third parties—are in the bestposition to provide independent advice to investors.

We are at the threshold of an exciting new era ininvestment theory and practice, where a number of dis-ciplines are converging to yield a more complete under-standing of how to make sound financial decisions. Andas new challenges arise in the investment managementindustry, those who are better equipped to meet thosechallenges will flourish and those who are not may per-ish, for Tennyson’s memorable phrase, “Nature, red intooth and claw...” applies with equal force to the land-scape of financial competition.

Appendix

This appendix contains a brief guide to the behav-ioral finance literature and the answer key toRusso and Shoemaker’s (1989) overconfidence

self-test.

Suggested Readings

Investment professionals rarely have the luxury ofimmersing themselves in the academic literature, and

the challenges to the uninitiated are even greater in fieldsthat are changing as rapidly as behavioral finance and thecognitive neurosciences. Therefore, I provide a ratherpersonal and incomplete but more manageable guide to



VOL. 7, NO. 2, 2005 39

the most relevant references for the ideas discussed in thisarticle, loosely grouped according to five general themes.

Efficient Markets HypothesisMost investment professionals will need no referencesto the EMH literature because the ideas in this line ofresearch pervade the current investment orthodoxy.Nevertheless, for completeness and for a historical per-spective on the development of these ideas, I recom-mend two references: Bernstein (1992), a fascinatingand highly readable intellectual history of quantitativefinance; and Lo (1997), a reference collection of themost significant academic papers in the market efficien-cy literature. Three recent papers that focus squarely onthe debate between EMH and behavioral finance areRubinstein (2001), Malkiel (2005), and Merton andBodie (2005), all of which provide eloquent summariesof the efficient markets mainstream.

Behavioral Finance The behavioral finance literature contains severalbranches and, unlike the EMH literature, has not yetcoalesced into an integrated whole.

A good overview of mainstream behavioral financeis provided by Shefrin (2000, 2005) and Statman (1999,2004b, 2004c), two of the pioneers of this strand of theliterature. Also, recent interviews with Robert Shiller andWilliam Sharpe in this journal present additional contextfor behavioral theories in the current literature.

A second branch of the literature lies within themore established field of psychology, which has had astrong impact on behavioral finance and economics, asunderscored by the fact that the Nobel Memorial Prizein Economic Sciences was awarded to a psychologist,Daniel Kahneman, in 2002. A number of popular expo-sitions of fascinating psychological experiments docu-menting behavioral biases have been published over theyears, but my two favorites are by Gilovich (1991) andDawes (2001). Plous (1993) provides a more academicand comprehensive exposition of the behavioral biasesliterature and does so in a remarkably readable fashion.And for those interested in the differences and similari-ties between the academic disciplines of economics andpsychology, Rabin (1998, 2002) provides a thought-provoking comparison.

The third branch contains the most recent researchin economic behavior, which incorporates ideas fromeconomics, finance, psychology, and the cognitive neu-rosciences. As a result, it is currently not consideredpart of either the mainstream finance or economics lit-eratures but has begun to develop an identity of its own,now known as neuroeconomics. Camerer, Loewenstein,and Prelec (2004) provide an excellent review of thisemerging discipline.

Adaptive Markets HypothesisThe AMH is a term that I coined in Lo (2004) and hasyet to become part of the standard lexicon of financialeconomists and investment professionals. Evolutionaryprinciples, however, have been applied to many eco-nomic and financial models. Unfortunately, because theapplications are so context specific, the AMH relies onno single literature. In fact, the ideas underlying theAMH have been inspired by several literatures: bound-ed rationality in economics (Simon, 1982, is the classicreference, of course); complex systems (Farmer, 2002);evolutionary biology (Wilson, 1975, and Trivers, 1985,in particular); evolutionary psychology (Pinker, 1997,provides a popular account of this new field, andBarrett, Dunbar, and Lycett, 2002, is an excellent refer-ence text for researchers); and behavioral ecology(Mangel and Clark, 1988).

Cognitive NeurosciencesAs a discipline, psychology has undergone a revolutionover the past few years because of the confluence of newtechnologies such as brain imaging and interdisciplinarycollaborations between neuroscientists and psycholo-gists. The two disciplines are now often considered one,sometimes called brain sciences, and more often calledcognitive neurosciences. Any serious student of behavioraleconomics and finance cannot afford to ignore this liter-ature, as some of the examples in this article haveshown. Damasio (1994) provides a riveting popularaccount of the clinical research that led to his seminalideas about the role of emotion in rationality, and hegives us digestible portions of neuroanatomy and neuro-physiology along the way. But this book is now some-what dated—in this fast-paced field, ten years makes ahuge difference in terms of progress and perspective—


and more recent popular expositions are available. Twofavorites are by Restak (2003), a clinical neurologist andvery experienced science writer, and Ramachandran(2004), a world-renowned neuroscientist. Finally,Gazzaniga and Heatherton (2003) is a comprehensivereference text that does an excellent job of connectingthe traditional psychology literature with more recentresearch in the neurosciences.

Answer Key to Overconfidence Test(1) 39 Years; (2) 4,187 Miles; (3) 13 Countries; (4) 39Books; (5) 2,160 Miles; (6) 390,000 Pounds; (7) 1756;(8) 645 Days; (9) 5,959 Miles; (10) 36,198 Feet

ENDNOTES1. The views and opinions expressed in this article are

those of the author only and do not necessarily represent theviews and opinions of AlphaSimplex Group, the InvestmentManagement Consultants Association (IMCA), MIT, or any oftheir affiliates and employees. The author makes no represen-tations or warranty, either expressed or implied, as to theaccuracy or completeness of the information contained in thisarticle, nor is he recommending that this article serve as thebasis for any investment decision. This article is for informa-tion purposes only. Research support from IMCA and the MITLaboratory for Financial Engineering is gratefully acknowl-edged. I thank Ed Baker, Bonny Brill, Stephanie Hogue,Dmitry Repin, Meir Statman, and Svetlana Sussman for help-ful comments and discussion.

2. Tversky died in 1996, otherwise he no doubt wouldhave shared the prize with Kahneman.

3. It should be emphasized, however, that irrationalitydoes not necessarily lead to violations of efficient markets.Certain forms of irrationality are simply irrelevant for the pricediscovery process, hence they have no impact on the EMH.See, for example, the case of a “Dutch Book,” described in thenext section, in which irrational probability beliefs concerninga particular random event yield arbitrage profits, implying thatsuch irrationality is unlikely to have a material impact on theprices of securities associated with that event. The importanceof a given behavioral pattern for financial market prices ishighly context-dependent and must be considered on a case-by-case basis.

4. The epithet “A sucker is born every minute,” com-monly attributed to P.T. Barnum, is in fact due to DavidHannum, one of Barnum’s competitors. See http://www.histo-rybuff.com/library/refbarnum.html for details.

5. See, for example, Rubinstein (2001) and Merton andBodie (2005).

6. Only when these axioms are satisfied is arbitrage ruledout. This was conjectured by Ramsey (1926) and proved rig-orously by de Finetti (1937) and Savage (1954).

7. According to MacKay (1841), at the peak of this bub-ble in 1636 one particularly rare species of tulip—theviceroy—was purchased for the following bill of goods: twolasts of wheat, four lasts of rye, four fat oxen, eight fat swine,twelve fat sheep, two hogsheads of wine, four tuns of beer, twotons of butter, one thousand pounds of cheese, a completebed, a suit of clothes, and a silver drinking cup. By 1637,bulbs that were worth 6,000 florins at the height of the maniawere trading for less than 500 florins, if they traded at all.

8. See Camerer, Loewenstein, and Prelec (2004) for areview of the neurosciences literature that is most relevant toeconomics and finance. Gazzaniga and Heatherton (2003) isan excellent and comprehensive introduction to the “new”psychology literature.

9. Recent research in the cognitive neurosciences and eco-nomics suggest an important link between rationality in deci-sion making and emotion (Grossberg and Gutowski, 1987;Damasio, 1994; Elster, 1998; Lo, 1999; Lo and Repin, 2002;Loewenstein, 2000; and Peters and Slovic, 2000), implying thatthe two are not antithetical but in fact complementary. Forexample, contrary to the common belief that emotions have noplace in rational financial decision-making processes, Lo andRepin (2002) present preliminary evidence that physiologicalvariables associated with the autonomic nervous system arehighly correlated with market events even for highly experi-enced professional securities traders. They argue that emotion-al responses are a significant factor in the real-time processingof financial risks, and that an important component of a pro-fessional trader’s skill lies in his or her ability to channel emo-tion, consciously or unconsciously, in specific ways duringcertain market conditions.

10. Other familiar manifestations of the antagonistic effectof emotion on the hominid brain include being so angry thatyou cannot see (“blinded by your anger,” both physically andmetaphorically), and becoming tongue-tied and disoriented inthe presence of someone you find unusually attractive. Bothvision and speech are mediated by the hominid brain.

11. See, for example, Barkow et al. (1992), Pinker (1993,1997), Crawford and Krebs (1998), Buss (1999), Gigerenzer(2000), Trivers (1985), and the emerging literature in “evolu-tionary psychology,” which is reviewed in detail in the recenttext by Barrett, Dunbar, and Lycett (2002).

12. For example, economists and biologists have begunto explore the implications of sociobiology in several veins:direct extensions of Wilson’s (1975) and Trivers’s (1985)framework to economics (Becker, 1976; Hirshleifer, 1977;Tullock, 1979); evolutionary game theory (Maynard Smith,1982; Weibull, 1995); evolutionary economics (Nelson andWinter, 1982; Andersen, 1994; Englund, 1994; Luo, 1999);and economics as a complex system (Anderson, Arrow, andPines, 1988). Hodgson (1995) contains additional examplesof studies at the intersection of economics and biology, andpublications such as the Journal of Evolutionary Economics andthe Electronic Journal of Evolutionary Modeling and EconomicDynamics now provide a home for this burgeoning literature.



VOL. 7, NO. 2, 2005 41

Evolutionary concepts also have appeared in severalfinancial contexts. For example, in a series of papers, Luo(1995, 1998, 2001, 2003) explores the implications of nat-ural selection for futures markets. Hirshleifer and Luo(2001) consider the long-run prospects of overconfidenttraders in a competitive securities market. And the literatureabout agent-based modeling pioneered by Arthur et al.(1997), in which interactions among software agents pro-grammed with simple heuristics are simulated, relies heavilyon evolutionary dynamics.

13. However, his work now is receiving greater attention,thanks in part to the growing behavioral literature in econom-ics and finance. See, for example, Simon (1982), Sargent(1993), Rubinstein (1998), Gigerenzer et al. (1999),Gigerenzer and Selten (2001), and Earl (2002).

14. For a concrete example, consider the rolling serialcorrelation of monthly returns of the S&P Composite Indexfrom 1871 to 2003, described in more detail further in thispaper. As a measure of market efficiency, the serial correlationcoefficient should converge to zero over this 133-year periodif markets are becoming progressively more efficient overtime. However, figure 2 tells a very different story—the cycli-cal behavior of the serial correlation coefficient is likely theresult of institutional changes in equity markets as well as theentry and exit of various market participants.

15. Simon’s (1982) seminal contributions to this litera-ture are still remarkably timely, and their implications have yetto be fully explored. More recent research about preferencesincludes Kahneman, Slovic, and Tversky (1982), Hogarth andReder (1986), Gigerenzer and Murray (1987), Dawes (1988),Fishburn (1988), Keeney and Raiffa (1993), Plous (1993),Sargent (1993), Thaler (1993), Damasio (1994), Arrow et al.(1996), Laibson (1997), Picard (1997), Pinker (1997), andRubinstein (1998). Starmer (2000) provides an excellentreview of this literature.

16. For example, using a well-known personalityprofiling instrument—the Keirsey (1998) TemperamentSorter—Statman and Wood (2005) show that risk attitudescan be related to personality types. Lo, Repin, and Steenbarger(2005) use daily emotional-state surveys as well as personali-ty inventory surveys to construct measures of personality traitsand emotional states for a group of eighty day-traders and cor-relate these measures with daily normalized profit-and-lossrecords. They find that subjects whose emotional reaction tomonetary gains and losses was more intense on both the pos-itive and negative side exhibited significantly worse tradingperformance, and large sudden swings in emotional statesseem especially detrimental to cumulative profits-and-losses.

17. The one exception to this general ethos of aversion tointerpersonal comparison of utility is welfare economics, thebranch of economics specifically focused on social welfareissues such as poverty, economic inequality, and related policyissues. Schumacher (1973) provides a highly readable lay-man’s introduction to equality versus efficiency from an eco-nomic perspective.

18. One specific behavioral mechanism by which thispath dependence is generated is the representativeness biasdescribed in the second section of this article.

REFERENCESAlvarez, W. 1997. T. Rex and the Crater of Doom. Princeton,

N.J.: Princeton University Press. Andersen, E. 1994. Evolutionary Economics: Post-Schumpeterian

Contributions. London: Pinter. Anderson, P., K. Arrow, and D. Pines, eds. 1988. The Economy

as an Evolving Complex System. Reading, Mass.: Addison-Wesley.

Ang, A., and G. Bekaert. 2004. “How Regimes Affect AssetAllocation.” Financial Analysts Journal 60: 86-99.

Arrow, K., et al. 1996. The Rational Foundations of EconomicBehaviour. New York: St. Martin’s Press.

Arthur, B., et al. 1997. “Asset Pricing Under EndogenousExpectations in an Artificial Stock Market.” In The Economyas an Evolving Complex System II. Edited by B. Arthur, S.Durlauf, and D. Lane. Reading, Mass.: Addison Wesley.

Barber, B., and T. Odean. 2001. “Boys Will Be Boys: Gender,Overconfidence, and Common Stock Investment.”Quarterly Journal of Economics 116: 261-229.

Barkow, J., L. Cosmides, and J. Tooby. 1992. The AdaptedMind: Evolutionary Psychology and the Generation of Culture.Oxford: Oxford University Press.

Barret, L., R. Dunbar, and J. Lycett. 2002. Human EvolutionaryPsychology. Princeton, N.J.: Princeton University Press.

Becker, G., 1976, "Altruism, Egoism, and Genetic Fitness:Economics and Sociobiology", Journal of EconomicLiterature 14, 817-826.

Bell, D. 1982. “Risk Premiums for Decision Regret.”Management Science 29: 1156-1166.

Bernstein, P. 1992. Capital Ideas. New York: Free Press. ———. 1998. “Why the Efficient Market Offers Hope to

Active Management.” In Economics and Portfolio Strategy,October 1. New York: Peter Bernstein, Inc.

———. 2003. “Are Policy Portfolios Obsolete?” In Economicsand Portfolio Strategy, March 1. New York: Peter Bernstein,Inc.

Black, F. 1986. “Noise.” Journal of Finance 41: 529-544. Buss, D. 1999. Evolutionary Psychology: The New Science of the

Mind. Boston: Allyn and Bacon. Caine, R., and G. Caine. 1990. Making Connections: Teaching and

the Human Brain. Nashville, Tenn: Incentive Publications. Camerer, C., G. Loewenstein, and D. Prelec. 2004.

“Neuroeconomics: How Neuroscience Can InformEconomics.” Forthcoming in Journal of Economic Literature.

Clarke, R., S. Krase, and M. Statman. 1994. “Tracking Errors,Regret, and Tactical Asset Allocation.” Journal of PortfolioManagement 20: 16-24.

Costa, P., and R. McCrae. 1992. “Normal PersonalityAssessment in Clinical Practice: The Neo PersonalityInventory.” Psychological Assessment 4: 5-13.


Crawford, C., and D. Krebs, eds. 1998. Handbook ofEvolutionary Psychology: Ideas, Issues, and Applications.Mahwah, N.J.: Lawrence Erlbaum.

Damasio, A. 1994. Descartes’ Error: Emotion, Reason, and theHuman Brain. New York: Avon Books.

Dawes, R. 1998. Rational Choice in an Uncertain World. NewYork: Harcourt Brace.

———. 2001. Everyday Irrationality: How Pseudo-Scientists,Lunatics, and the Rest of Us Systematically Fail to ThinkRationally. Boulder, Colo.: Westview Press.

Dawkins, R. 1976. The Selfish Gene. Oxford: Oxford UniversityPress.

De Becker, G. 1997. The Gift of Fear: Survival Signals ThatProtect Us from Violence. Boston: Little Brown.

DeBondt, W., and R. Thaler. 1986. “Does the Stock MarketOverreact?” Journal of Finance 40: 793-807.

De Finetti, B. 1937. “La Prévision: Ses Lois Logiques, SesSources Subjectives.” Annales de l’Institut Henri Poincaré7:1-68. Translated into English in Studies in SubjectiveProbability. 1964. Edited by H. Kyburg and H. Smokler.New York: John Wiley & Sons.

Earl, P. 2002. The Legacy of Herbert Simon in Economic Analysis.Chelthenham, U.K.: Edward Elgar Publishing Co.

Eisenberger, N., M. Lieberman, and K. Williams. 2003. “DoesRejection Hurt? An fMRI Study of Social Exclusion.”Science 302: 290-292.

Eldredge, N., and S. Gould. 1972. “Punctuated Equilibria: AnAlternative to Phyletic Gradualism.” In Models inPaleobiology. Edited by T. Schopf. San Francisco: Freeman,Cooper: 82-115.

Elster, J. 1998. “Emotions and Economic Theory.” Journal ofEconomic Literature 36: 47-74.

Englund, R., ed. 1994. Evolutionary Concepts in ContemporaryEconomics. Ann Arbor, Mich.: University of Michigan Press.

Farmer, D. 2002. “Market Force, Ecology and Evolution.”Industrial and Corporate Change 11: 895-953.

Farmer, D., and A. Lo. 1999. “Frontiers of Finance: Evolutionand Efficient Markets.” Proceedings of the National Academyof Sciences 96: 9991-9992.

Fishburn, P. 1998. Nonlinear Preference and Utility Theory.Baltimore, Md.: Johns Hopkins University Press.

Fischoff, B., and P. Slovic. 1980. “A Little Learning...:Confidence in Multicue Judgment Tasks.” In Attention andPerformance, VIII. Edited by R. Nickerson. Hillsdale, N.J.:Erlbaum.

Gazzaniga, M., and T. Heatherton. Psychological Science: Mind,Brain, and Behavior. New York: W.W. Norton & Co.

Gervais, S., and T. Odean. 2001. “Learning to Be Over-confident.” Review of Financial Studies 14: 1-27.

Gigerenzer, G. 2000. Adaptive Thinking: Rationality in the RealWorld. Oxford: Oxford University Press.

Gigerenzer, G., and D. Murray. 1987. Cognition as IntuitiveStatistics. Mahwah, N.J.: Lawrence Erlbaum.

Gigerenzer, G., and R. Selten. 2001. Bounded Rationality: TheAdaptive Toolbox. Cambridge, Mass.: MIT Press.

Gigerenzer, G., P. Todd, and the ABC Research Group. 1999.Simple Heuristics that Make Us Smart. Oxford: OxfordUniversity Press.

Gilovich, T. 1991. How We Know What Isn’t So: The Fallibility ofHuman Reason in Everyday Life. New York: The Free Press.

Grossberg, S., and W. Gutowski. 1987. “Neural Dynamics ofDecision Making Under Risk: Affective Balance andCognitive-Emotional Interactions.” Psychological Review 94:300-318.

Grossman, S. 1976. “On The Efficiency of Competitive StockMarkets Where Traders Have Diverse Information.” Journalof Finance 31: 573-585.

Grossman, S., and J. Stiglitz. 1980. “On the Impossibility ofInformationally Efficient Markets.” American EconomicReview 70: 393-408.

Hirshleifer, J. 1977. “Economics from a Biological Viewpoint.”Journal of Law and Economics 20: 1-52.

Hirshleifer, D., and G. Luo. 2001. “On the Survival of Over-confident Traders in a Competitive Securities Market.”Journal of Financial Markets 4: 73-84.

Hodgson, G., ed. 1995. Economics and Biology. Cheltenham,U.K.: Edward Elgar Publishing Co.

Hogarth, R., and M. Reder. 1986. Rational Choice: The Contrastbetween Economics and Psychology. Chicago: University ofChicago Press.

Huberman, G., and T. Regev. 2001. “Contagious Speculationand a Cure for Cancer: A Nonevent That Made Stock PricesSoar.” Journal of Finance 56: 387-396.

Kahneman, D., P. Slovic, and A. Tversky. 1982. JudgmentUnder Uncertainty: Heuristics and Biases. Cambridge, U.K.:Cambridge University Press.

Kahneman, D., and A. Tversky. 1979. “Prospect Theory: AnAnalysis of Decision Under Risk.” Econometrica 47: 263-291.

Keeney, R., and H. Raiffa. 1993. Decisions with MultipleObjectives: Preferences and Value Tradeoffs. Cambridge, U.K.:Cambridge University Press.

Keirsey, D. 1998. Please Understand Me II: Temperament,Character, Intelligence. Del Mar, Calif.: Prometheus Nemesis.

Kindleberger, C. 1989. Manias, Panics, and Crashes: A Historyof Financial Crises. New York: Basic Books.

Laibson, D. 1997. “Golden Eggs and HyperbolicDiscounting.” Quarterly Journal of Economics 62: 443-477.

Lichtenstein, S., B. Fischoff, and L. Phillips. 1982. “Calibrationof Probabilities: The State of the Art to 1980.” In JudgmentUnder Uncertainty: Heuristics and Biases. Edited by D.Kahneman, P. Slovic, and A. Tversky. Cambridge, U.K.:Cambridge University Press.

Lo, A., ed. 1997. Market Efficiency: Stock Market Behavior inTheory and Practice, Vols. I and II. Cheltenham, U.K.:Edward Elgar Publishing Co.

Lo, A. 2002. “Bubble, Rubble, Finance In Trouble?” Journal ofPsychology and Financial Markets 3: 76-86.

———. 2004. “The Adaptive Markets Hypothesis: MarketEfficiency from an Evolutionary Perspective.” Journal ofPortfolio Management 30: 15-29.



VOL. 7, NO. 2, 2005 43

Lo, A., and D. Repin. 2002. “The Psychophysiology of Real-Time Financial Risk Processing.” Journal of CognitiveNeuroscience 14: 323-339.

Lo, A., D. Repin, and B. Steenbarger. 2005. “Fear and Greedin Financial Markets: A Clinical Study of Day-Traders.”Forthcoming in American Economic Review.

Loewenstein, G. 2000. “Emotions in Economic Theory andEconomic Behavior.” American Economic Review 90: 426-432.

Luo, G. 1995. “Evolution and Market Competition.” Journal ofEconomic Theory 67: 223-250.

———. 1998. “Market Efficiency and Natural Selection in aCommodity Futures Market.” Review of Financial Studies11: 647-674.

———. 1999. “The Evolution of Money as a Medium ofExchange.” Journal of Economic Dynamics and Control 23:415-458.

———. 2001. “Natural Selection and Market Efficiency in aFutures Market with Random Shocks.” Journal of FuturesMarkets 21: 489-516.

———. 2003. “Evolution, Efficiency and Noise Traders in aOne-Sided Auction Market.” Journal of Financial Markets 6:163-197.

MacCrimmon, K., and D. Wehrung. 1986. Taking Risks: TheManagement of Uncertainty. New York: The Free Press.

MacLean, P. 1990. The Triune Brain in Evolution: Role inPaleocerebral Function. New York: Plenum Press.

Malkiel, B. 2003. “Market Efficiency and Active Management:A Non-Random Talk with Burton G. Malkiel, Ph.D.”Journal of Investment Consulting 6, no. 2: 11–19.

Mangel, M., and C. Clark. 1988. Dynamic Modeling in BehavioralEcology. Princeton, N.J.: Princeton University Press.

Maynard Smith, J. 1982. Evolution and the Theory of Games.Cambridge, U.K.: Cambridge University Press.

Merton, R. 1987. “A Simple Model of Capital MarketEquilibrium with Incomplete Information.” Journal ofFinance 42: 483-510.

Merton, R., and Z. Bodie. 2005. “Design Of Financial Systems:Towards a Synthesis of Function and Structure.” Journal ofInvestment Management 3: 6-28.

Nelson, R., and S. Winter. 1982. An Evolutionary Theory ofEconomic Change. Cambridge, Mass.: Belknap Press ofHarvard University Press.

Niederhoffer, V. 1997. Education of a Speculator. New York:John Wiley & Sons.

Odean, T. 1988. “Are Investors Reluctant to Realize TheirLosses?” Journal of Finance 53: 1775-1798.

Pinker, S. 1993. The Language Instinct. New York: Morrow. ———. 1997. How the Mind Works. New York: W.W. Norton

& Co. Peters, E., and P. Slovic. 2000. “The Springs of Action:

Affective and Analytical Information Processing in Choice.”Personality and Social Psychology Bulletin 26: 1465-1475.

Picard, R. 1997. Affective Computing. Cambridge, Mass.: MIT Press. Plous, S. 1993. The Psychology of Judgment and Decision Making.

New York: McGraw-Hill.

Rabin, M. 1998. “Psychology and Economics.” Journal ofEconomic Literature 36: 11-46.

———. 2002. “A Perspective on Psychology and Economics.”European Economic Review 46: 657-685.

Ramachandran, V. 2004. A Brief Tour of Human Consciousness.New York: Pi Press.

Ramsey, F. 1926. “Truth and Probability.” In Foundations ofMathematics and Other Logical Essays. Edited by R.Braithwaite. New York: Harcourt Brace.

Restak, R. 2003. The New Brain: How the Modern Age IsRewiring Your Mind. New York: Rodale.

Rolls, E. 1990. “A Theory of Emotion, and Its Application toUnderstanding the Neural Basis of Emotion.” Cognition andEmotion 4: 161-190.

———. 1992. “Neurophysiology and Functions of thePrimate Amygdala.” In The Amygdala: NeurobiologicalAspects of Emotion, Memory, and Mental Dysfunction. Editedby J. Aggelton. New York: John Wiley & Sons: 143-166.

———. 1994. “A Theory of Emotion and Consciousness, andIts Application to Understanding the Neural Basis ofEmotion.” In The Cognitive Neurosciences. Edited by M.Gazzaniga. Cambridge, Mass.: MIT Press: 1091-1106.

———. 1999. The Brain and Emotion. Oxford: OxfordUniversity Press.

Rubinstein, A. 1998. Modeling Bounded Rationality. Cambridge,Mass.: MIT Press.

Rubinstein, M. 2001. “Rational Markets: Yes or No? TheAffirmative Case.” Financial Analysts Journal 57: 15-29.

Russo, J., and P. Shoemaker. 1989. Decision Traps: Ten Barriersto Brilliant Decision Making and How to Overcome Them.New York: Simon & Schuster.

Sargent, T. 1993. Bounded Rationality in Macroeconomics.Oxford: Clarendon Press.

Savage, L. 1954. Foundations of Statistics. New York: John Wiley& Sons.

Schumacher, F. 1973. Small Is Beautiful: Economics As If PeopleMattered. New York: Harper & Row.

Schumpeter, J. 1939. Business Cycles: A Theoretical, Historical,and Statistical Analysis of the Capitalist Process. New York:McGraw-Hill.

Shefrin, H. 2000. Beyond Greed and Fear. Boston, Mass.:Harvard Business School Press.

———. 2005, A Behavioral Approach to Asset Pricing.Amsterdam: Elsevier Academic Press.

Shefrin, H., and M. Statman. 1985. “The Disposition to SellWinners Too Early and Ride Losers Too Long: Theory andEvidence.” Journal of Finance 40: 777-790.

———. 2000. “Behavioral Portfolio Theory.” Journal ofFinancial and Quantitative Analysis 35: 127-151.

Shefrin, H., and R. Thaler. 1988. “The Behavioral Life CycleHypothesis.” Economic Inquiry 24: 609-643.

Simon, H. 1955. “A Behavioral Model of Rational Choice.”Quarterly Journal of Economics 69: 99-118.

———. 1982. Models of Bounded Rationality, Vols. 1 and 2.Cambridge, Mass.: MIT Press.


Starmer, C. 2000. “Developments in Non-Expected UtilityTheory: The Hunt for a Descriptive Theory of ChoiceUnder Risk.” Journal of Economic Literature 38: 332-382.

Statman, M. 1999. “Behavioral Finance: Past Battle and FutureEngagements.” Financial Analysts Journal (November/December): 18-27.

———. 2004a. “What Do Investors Want?” Journal of PortfolioManagement 30: 153-161.

———. 2004b. “Lessons in Behavioral Finance.” In TheInvestment Think Tank. Edited by H. Evensky and D. Katz.Princeton, N.J.: Bloomberg Press.

———. 2005. “Martha Stewart’s Investment Lessons.” Journalof Investment Consulting 7, no. 2: 52–60

Thaler, R., ed. 1993. Advances in Behavioral Finance. New York:Russell Sage Foundation.

Trivers, R., 1985, Social Evolution. Menlo Park, Calif:Benjamin/Cummins Publishing Company.

Tullock, G. 1979. “Sociobiology and Economics.” AtlanticEconomic Journal 7: 1-10.

Tversky, A., and D. Kahneman. 1981. “The Framing of Decisionsand the Psychology of Choice,” Science 211: 453-458.

Urban, G., and J. Hauser. 1993. Design and Marketing of NewProducts, 2nd ed. Englewood Cliffs, N.J.: Prentice Hall.

Weibull, J. 1995. Evolutionary Game Theory. Cambridge,Mass.: MIT Press.

Wilson, E. 1975. Sociobiology: The New Synthesis. Cambridge,Mass.: Belknap Press of Harvard University Press.

Zajonc, R. 1980. “Feeling and Thinking: Preferences Need NoInferences.” American Psychologist 35: 151-175.

———. 1984. “On the Primacy of Affect.” AmericanPsychologist 39: 117-123.


Documents

Reconciling Efficient Markets With Behavioural Finance