Beauty Contests and Tail Feathers: How the Theory of Mind Feeds Economic Narratives
Psychologists have noted that the human species is unique in the advanced development of its theory of mind—that is, humans’ strong tendency to form a model in their own minds of the activities in others’ minds. We are thinking about what others are thinking, about their individual thoughts. We observe their actions, their facial expressions, and their vocal intonation, which we then relate to their beliefs and intentions.
The contagion of specific narratives may be related to storytellers’ impressions regarding what other people will think. People like to hear stories that they can retell to others who will like the same story, and so storytellers like to tell such stories.
In 1936, Keynes introduced what we now call theory of mind into economic theory with his “beauty contest” metaphor,16 which he put forth to explain speculative markets, such as the stock market. Keynes thought that people deciding which investments to make were basing their decisions on observations of what other investors were thinking or what they were about to do with their investments (which might cause future price changes). In the case of stock market investments, investors look at what other people whom they randomly encounter are saying and emoting, and they look at patterns in stock prices that offer clues regarding what other people are doing or will soon be doing. They are usually not looking at real evidence based on the firm’s technology or management style.
Keynes said he had seen a newspaper contest that displayed a hundred photos, each of a pretty face. But the women in the photos were not the contestants in this unusual form of beauty contest; the readers of the newspaper were. They were asked to mail to the newspaper their list of the six prettiest faces. The person whose list most closely matched the most popular faces as revealed by all the lists together would win the contest prize.17
Keynes pointed out that the optimal strategy is not to pick the six prettiest faces based on one’s own opinion. Instead, it makes more sense to pick the six that one thinks other people would find prettiest. But this strategy is not optimal either, if we carry the model of mind to the next step in the chain. One should pick the faces that one thinks that others think that others find the prettiest. So, in a rational world, one might suppose that investors, trying to gauge what other investors think other investors are thinking, will try to determine the right thing to think about the speculative investments. However, investors do not necessarily follow this strategy, even if all investors are rational and know that all investors are rational.18 In addition, we have to account for the investors’ less-than-perfect rationality and the investor irrationality expected by other investors.
In our 2009 book Animal Spirits, which was in many ways an expansion and elaboration of Keynes’s ideas, George Akerlof and I used the beauty contest metaphor to construct a theory of the emotional foundation of business fluctuations in general. The beauty contest metaphor also applies to the contagion of narratives. When we choose to tell a story to others, we base that choice on our perceptions of how people will react to that story in their own minds. We will likely spread a story, whether it is a story about boom-time thinking or about economic despair, if we think that others will like the story enough to want to spread it further. Even if we are spreading an economic narrative for no other reason than trying to amuse ourselves, we are likely to engineer our story to spread based on our model of others’ minds.
The stories that go viral are essentially random, just as mutations in evolutionary biology are random. Traditional evolutionary theory suggests that the mutations that survive and spread are those few out of many that are in themselves advantageous for survival. But there is another branch to Darwin’s theory, that of sexual selection, and it suggests that the winning mutations may be just as random as the original mutation. Something like this randomness may affect economic narratives going viral as well.
In his 2017 book The Evolution of Beauty, ornithologist Richard O. Prum argues that sexual selection gives rise to fluctuations in the animal kingdom that resemble speculative bubbles in economics. Perhaps the most famous example of sexual selection in biology is the male peacock, which has very heavy tail feathers that inhibit his activities. But these feathers are much favored by the female of the species, which facilitates mating and the reproduction of more beautiful tail feathers. Thus the female sexual choice may create an evolutionary advantage for some useless characteristic in a process called a Fisherian runaway, after theorist R. A. Fisher.19 The mechanism does not even require two distinct sexes, as there is evidence for such sexual selection processes among hermaphrodite species in which each individual has both male and female organs.20 In both evolutionary biology and narrative economics, some kind of ornament or display can become popular for no more reason than the fact that it randomly began to be popular.
Irrational Impulses Inform Economic Narratives
Psychologist Jerome Bruner, who has stressed the importance of narratives in understanding human culture, wrote that we should not assume that human actions are driven in response to purely objective facts:
I do not believe that facts ever quite stare anybody in the face. From a psychologist’s point of view, that is not how facts behave, as we well know from our studies of perception, memory, and thinking. Our factual worlds are more like cabinetry carefully carpentered than like a virgin forest inadvertently stumbled upon.21
That is, narratives are human constructs that are mixtures of fact, emotion, human interest, and other extraneous details that form an impression on the human mind.
Psychiatrists and psychologists recognize that mental illness is often an extreme form of normal behavior or a narrow disruption of normal human mental faculties. So we can learn about the complexities of normal human narrative brain processing by studying dysnarrativia, or abnormal narrative phenomena. Neuroscientists Kay Young and Jeffrey Saver (2001) listed some of its varied forms: arrested narration (the ability to tell only stories learned before a brain injury), undernarration (the telling of vacillating, impulsive stories), denarration (failure to organize a story in terms of an action-generating temporal frame), and confabulation (the fabrication of stories that have little or no relation to reality). Each form of dysnarrativia is related to injury in a specific part of the brain.
Schizophrenia is a serious mental illness that can manifest as a disorder of narrative, as it often involves hearing imaginary voices delivering a fantastic and jumbled narrative.22 Hearing voices as a symptom of schizophrenia is correlated with volume deficits in specific brain areas.23 The narrative disruption found in autism spectrum disorder also is related to brain anomalies.24
Framing, the Representativeness Heuristic, and the Affect Heuristic
Narrative psychology also relates to the psychological concept of framing.25 If we can create an amusing story that will get retold, it can establish a point of view, a reference point, that will influence decisions. Framing is related to the Daniel Kahneman and Amos Tversky representativeness heuristic (1973), whereby people form their expectations based on some idealized story or model, judging these expectations based on the prominence of the idealized story rather than estimated probabilities. For example, we may judge the danger of an emerging economic crisis by its similarity to a remembered story of a previous crisis, rather than by any logic.
George Katona, one of the founders of behavioral economics and author of the 1975 book Psychological Economics, noted an odd phenomenon: when he interviewed common people and asked them about their expectations of key economic variables, he had the feeling that they had no clear expectations, and that they made up numbers on the spot to please him. But I would argue that these ordinary people were thinking about narratives that involved people and prices. If asked in an interview about their expectations for inflation, for example, they might not answer the question directly but rather offer a dramatic story with human interest and with clear moralizing, about politicians’ or labor unions’ act
ivities that might be related to inflation.
Psychologists have also noted an affect heuristic, whereby people who are experiencing strong emotions, such as fear, tend to extend those feelings to unrelated events.26 Sometimes people note strong emotions or fears about possibilities that they know logically are not real, suggesting that the brain has multiple systems for assessing risk. This “risk as feelings” hypothesis holds that some primitive brain system more connected to palpable emotions has its own heuristic for assessing risk.27
In joint work with William Goetzmann and Dasol Kim, George Akerlof and I examined data from a questionnaire survey of investors and high-income Americans since 1989. We found that people have exaggerated assessments of the risk of a stock market crash, and that these assessments are influenced by the news stories, especially front-page stories, that they read. One intriguing finding was that a natural event such as an earthquake could influence estimations of the likelihood of a stock market crash. The respondents in our survey assigned statistically significantly higher probabilities to a stock market crash if there had been an earthquake within thirty miles of their zip code within thirty days, triggering the affect heuristic. It seems reasonable to hypothesize that local earthquakes start local narratives with negative emotional valence. Analogous evidence has indicated that seemingly irrelevant events with strong narrative potential can affect economic or political outcomes: the World Cup competition can affect economic confidence,28 shark attacks at local beaches can affect votes for local incumbents,29 and background music in advertisements can have a strong effect on consumers.30 Wine stores find buyers purchasing more expensive wines if the background music is classical versus Top 40.31
An affect heuristic also operates in generating activity by Internet trolls (people who send nasty or obscene comments on the Internet).32 Trolling behavior appears to be contagious: an experimental group randomly selected from the general population was primed with nasty examples of trolling. Members of that group were then much more likely to post similar comments.
Going Forward
The tantalizing evidence about the impact of narratives from neuroscience and related observations suggests some entirely different explanations of the severity of major economic events. In part II of this book we consider some organizing principles for narrative economics. A key issue is assigning the direction of causality from dispersed and ill-defined narrative constellations to actual economic activity, a topic to which we turn in the next chapter. The chapter after that offers key foundations of narrative economics. Part III then presents a list of nine important perennial narrative constellations, one (or a pair) per chapter.
Part II
The Foundations of Narrative Economics
Chapter 7
Causality and Constellations
The goal of this book is to improve people’s ability to anticipate and deal with major economic events, such as depressions, recessions, or secular (that is, long-term) stagnation, by encouraging them to identify and incorporate into their thinking the economic narratives that help to define these events. Before we can forecast reliably, we need some understanding of these events’ true ultimate causes. The key problem is determining what is a cause versus what is a consequence.
Though modern economists tend to be very attentive to causality, as a general rule they do not attach any causal significance to the invention of new narratives. I want to argue here not only that causality exists, but also that it goes both ways: new contagious narratives cause economic events, and economic events cause changed narratives.
Of course, almost nothing beyond spots on the sun is purely an outside influence on the economy (more on sunspots later in the chapter), but we can think of new narratives as causative innovations, because each narrative originates in the mind of a single individual (or as a collaboration among a few people). Economic historian Joel Mokyr (2016) calls such an individual a “cultural entrepreneur,” and he traces the concept back to philosopher and polymath David Hume, who wrote in 1742:
What depends on a few persons is, in great measure, to be ascribed to chance, or secret and unknown causes; what arises from a great number may often be accounted for by determinate and known causes.1
Understanding the effects of the “few persons” who create contagious new narratives is essential to formulating the foundations of a theory of narrative economics.
The effects of a “few persons” sometimes work through the creation of contagious new narratives. Though narratives are commonly connected with celebrities, the “few persons” who invent a contagious narrative are usually not famous, and often we will never know who they were. Later on, we can look for celebrities attached to them, but we will usually not find their authors.
In this chapter we will consider the causal elements that make economic narratives go viral—especially stories and storytelling—with the aim of developing a better understanding of these narratives’ deep structure.
Direction of Causality
It is not easy to prove direction of causality between a narrative and the economy. For example, did the stories of successful speculators and wild enthusiasm for stocks that characterized the 1920s cause increased stock prices and increased corporate earnings? Or did those increased earnings cause the enthusiasm? Was the similar enthusiasm for Bitcoin after 2009 in any way responsible for the increase in Bitcoin’s price? Or was Bitcoin’s increased value just a logical reaction to news stories and new progress in the mathematical theory of cryptography?
A problem in establishing direction of causality for major economic events is that economists usually cannot run controlled experiments that accurately simulate economic conditions at large. In contrast, laboratory scientists conduct random trials, perhaps by administering a test drug to an experimental group and a placebo to a control group, and then using statistical analysis to determine whether the drug really causes patients to recover. The best economists can often do is to look for events that might be deemed natural experiments. Henry W. Farnam, in his 1912 presidential address before the American Economic Association, addressed economists’ inability to conduct controlled experiments, asserting nonetheless that the study of economic history can allow economists to infer causality because random shocks have occurred through history, as when governments embark on crazy economic policies. In fact, Farnam said, “The economist is really fortunate in having experiments tried for him without expense.”2
In their 1963 Monetary History of the United States, Milton Friedman and Anna J. Schwartz gave three examples of what they called “quasi-controlled experiments” to establish causal impact from monetary policy to the aggregate economy: the large gold discoveries of 1897 to 1914, which expanded the money supply, and the periods during and immediately after World War I and World War II. We can debate whether these events were truly random exogenous shocks (that is, not caused by the economy), but much more discussion on inferring direction of causality with economic data has taken place since 1963. The general conclusion is that it is indeed possible to infer causality even when controlled experiments are impossible. New narratives might be interpreted as exogenous, helping us identify additional quasi-controlled experiments. In fact, the gold discoveries and wars that Friedman and Schwartz emphasized likely were exogenous because they were made possible by innovations in popular narratives, such as gold rush stories or fake news about foreign conspiracy.
We must be wary of many (but not all) economists’ supposition that the causality always runs from economic events to narratives, and not the other way around. There has been a lively debate about the impact of self-fulfilling prophecies in economics. Sociologist Robert K. Merton coined the phrase self-fulfilling prophecy in 1948, intending to apply the concept to economic fluctuations. The term often refers to prophecies stimulated by genuinely extraneous events, with the most popular example being sunspots (spots on the sun, which come and go through time, and are observable through telescopes).
The economist Willia
m Stanley Jevons proposed in 1878 that world economic fluctuations might be driven by “periodic variation in the sun’s rays, of which the sun-spots are a mere sign.”3 If the heat coming from the sun is stronger in some years than in others, then crops and other economic output may be stronger in hotter years, which may lead to major economic fluctuations. There was by 1878 already astronomical evidence on solar activity, going back centuries, in the form of counts of sunspots through time. He thought he discerned a correlation between those sunspot counts and economic events. And the cause of this correlation had to be the sun, for there is no conceivable theory that causality could go the other way, from economic events on earth to spots on the sun. His theory sounded plausible, but subsequent economic research did not support it, and variations in solar output are too small to have any substantial such effect. Sunspots should hardly affect the economy, but they may do so if people mystically believe they should, as economists David Cass and Karl Shell explained in 1983. Now, economists use the term sunspots to refer to any extraneous noise that affects the economy because people believe it will. Economist Roger E. A. Farmer has been a leader in the field of macroeconomic self-fulfilling prophecies.4 To his and others’ work I add the idea that these self-fulfilling prophecies do not come out of nowhere. Rather, they typically come from millions of mutations in narratives, of which a few are contagious enough in the current environment to become major epidemics. As we have seen, this process can be observed and modeled.
Narrative Economics Page 9