narrative economics
Robert J. Shiller
narrative economics
How Stories Go Viral & Drive Major Economic Events
princeton university press
princeton & oxford
Copyright © 2019 by Robert J. Shiller
Requests for permission to reproduce material from this work should be sent to [email protected]
Published by Princeton University Press
41 William Street, Princeton, New Jersey 08540
6 Oxford Street, Woodstock, Oxfordshire OX20 1TR
press.princeton.edu
All Rights Reserved
ISBN 9780691182292
ISBN (e-book) 9780691189970
Version 1.0
British Library Cataloging-in-Publication Data is available
Editorial: Peter Dougherty and Alena Chekanov
Production Editorial: Terri O’Prey
Text Design: Leslie Flis
Jacket Design: Faceout Studio
Contents
List of Figures vii
Preface: What Is Narrative Economics? ix
Acknowledgments xxi
Part I The Beginnings of Narrative Economics
1 The Bitcoin Narratives 3
2 An Adventure in Consilience 12
3 Contagion, Constellations, and Confluence 18
4 Why Do Some Narratives Go Viral? 31
5 The Laffer Curve and Rubik’s Cube Go Viral 41
6 Diverse Evidence on the Virality of Economic Narratives 53
Part II The Foundations of Narrative Economics
7 Causality and Constellations 71
8 Seven Propositions of Narrative Economics 87
Part III Perennial Economic Narratives
9 Recurrence and Mutation 107
10 Panic versus Confidence 114
11 Frugality versus Conspicuous Consumption 136
12 The Gold Standard versus Bimetallism 156
13 Labor-Saving Machines Replace Many Jobs 174
14 Automation and Artificial Intelligence Replace Almost All Jobs 196
15 Real Estate Booms and Busts 212
16 Stock Market Bubbles 228
17 Boycotts, Profiteers, and Evil Business 239
18 The Wage-Price Spiral and Evil Labor Unions 258
Part IV Advancing Narrative Economics
19 Future Narratives, Future Research 271
Appendix: Applying Epidemic Models to Economic Narratives 289
Notes 301
References 325
Index 351
Figures
2.1 Articles Containing the Word Narrative as a Percentage of All Articles in Academic Disciplines 13
3.1 Epidemic Curve Example, Number of Newly Reported Ebola Cases in Lofa County, Liberia, by week, June 8–November 1, 2014 19
3.2 Percentage of All Articles by Year Using the Word Bimetallism or Bitcoin in News and Newspapers, 1850–2019 22
3.3 Frequency of Appearance of Four Economic Theories, 1940–2008 27
5.1 Frequency of Appearance of the Laffer Curve 43
10.1 Frequency of Appearance of Financial Panic, Business Confidence, and Consumer Confidence in Books, 1800–2008 116
10.2 Frequency of Appearance of Financial Panic Narratives within a Constellation of Panic Narratives through Time, 1800–2000 118
10.3 Frequency of Appearance of Suggestibility, Autosuggestion, and Crowd Psychology in Books, 1800–2008 120
10.4 Frequency of Appearance of Great Depression in Books, 1900–2008, and News, 1900–2019 134
11.1 Frequency of Appearance of American Dream in Books, 1800–2008, and News, 1800–2016 152
12.1 Frequency of Appearance of Gold Standard in Books, 1850–2008, and News, 1850–2019 159
13.1 Frequency of Appearance of Labor-Saving Machinery and Technological Unemployment in Books, 1800–2008 175
14.1 Percentage of Articles Containing the Words Automation and Artificial Intelligence in News and Newspapers, 1900–2019 197
15.1 “Housing Bubble” Google Search Queries, 2004–19 226
16.1 Frequency of Appearance of Stock Market Crash in Books, 1900–2008, and News, 1900–2019 232
17.1 Frequency of Appearance of Profiteer in Books, 1900–2008, and News, 1900–2019 243
18.1 Frequency of Appearance of Wage-Price Spiral and Cost-Push Inflation in Books, 1900–2008 259
A.1 Theoretical Epidemic Paths 291
Preface: What Is Narrative Economics?
When I was a nineteen-year-old undergraduate at the University of Michigan over a half century ago, my history professor, Shaw Livermore, assigned a short book by Frederick Lewis Allen, Only Yesterday: An Informal History of the 1920s, about the run-up to the 1929 stock market crash and the beginnings of the Great Depression of the 1930s. It was a best seller when it was published in 1931. After reading it, I came to believe that the book was extremely important, for it not only described the lively atmosphere and massive speculative booms of the Roaring Twenties but also illuminated the causes of the Great Depression, the biggest economic crisis ever to hit the world economy. It struck me that this period’s history of rapid-fire contagious narratives somehow contributed to the changing spirit of the times. For example, Allen wrote an eyewitness account of the spread of narratives throughout 1929, just before the stock market peaked:
Across the dinner table one heard fantastic stories of sudden fortunes: a young banker had put every dollar of his small capital into Niles-Bement-Pond and now was fixed for life; a widow had been able to buy a large country house with her winnings in Kennecott. Thousands speculated—and won too—without the slightest knowledge of the nature of the company upon whose fortunes they were relying, like the people who bought Seaboard Air Line under the impression that it was an aviation stock. [Seaboard Air Line was a railroad, so named in the nineteenth century, when “air line” meant the shortest conceivable path between two points.]1
These narratives sound a bit fanciful, but they were repeated so often that they were hard to ignore. It couldn’t have been so easy to get rich, and the most intelligent people in the 1920s must have realized that. But the opposing narrative, which would have pointed out the folly of get-rich-quick schemes, was apparently not very contagious.
After I read Allen’s book, it seemed to me that the trajectory of the stock market and the economy, as well as the onset of the Great Depression, must have been tied to the stories, misperceptions, and broader narratives of the period. But economists never took Allen’s book seriously, and the idea of narrative contagion never entered their mathematical models of the economy. Such contagion is the heart of narrative economics.
In today’s parlance, stories of fabulously successful investors who were not experts in finance “went viral.” Like an epidemic, they spread from person to person, through word of mouth, at dinner parties and other gatherings, with help from telephone, radio, newspapers, and books. ProQuest News & Newspapers (proquest. com), which allows online search of newspaper articles and advertisements back to the 1700s, shows that the phrase go viral (and variations going viral, went viral, and gone viral) first appeared as an epidemic in newspapers only around 2009, typically in connection with stories about the Internet. The associated term viral marketing goes back only a little further, to 1991, as the name of a small company in Nagpur, India. Today, as a ProQues
t search reveals, the phrase going viral itself has gone viral. Google Ngrams (books.google. com/ngrams), which allows users to search for words and phrases in books all the way back to the 1500s, shows a similar trajectory for go viral. Since 2009, trending now, a synonym for going viral, has also gone viral. These epidemics were helped along by the prominent statistics displayed on Internet sites about numbers of views or likes. Both “going viral” and “trending now” characterize the rising part of the infectives curve, when the epidemic is growing. There isn’t as much popular attention to the process of forgetting, the later falling part of the infectives curve, though for economic narratives that will likely be as important a cause of changes in economic behavior.
Allen was thinking in terms of stories going viral when he wrote his book, though he did not use the term. He wrote about his “emphasis upon the changing state of the public mind and upon the sometimes trivial happenings with which it was preoccupied,”2 but he did not formalize his thinking about the contagion of narratives.
We need to incorporate the contagion of narratives into economic theory. Otherwise, we remain blind to a very real, very palpable, very important mechanism for economic change, as well as a crucial element for economic forecasting. If we do not understand the epidemics of popular narratives, we do not fully understand changes in the economy and in economic behavior. There is an extensive medical literature on forecasting disease epidemics. This literature shows that understanding the nature of epidemics and their relation to contagion factors can help us forecast better than those using purely statistical methods can.
Narrative Economics: What’s in a Phrase?
The phrase narrative economics has been used before, though rarely. R. H. Inglis Palgrave’s Dictionary of Political Economy (1894) contains a brief mention of narrative economics,3 but the term appears to refer to a research method that presents one’s own narrative of historical events. I am concerned not with presenting a new narrative but rather with studying other people’s narratives of major economic events, the popular narratives that went viral. In using the term narrative economics, I focus on two elements: (1) the word-of-mouth contagion of ideas in the form of stories and (2) the efforts that people make to generate new contagious stories or to make stories more contagious. First and foremost, I want to examine how narrative contagion affects economic events.
The word narrative is often synonymous with story. But my use of the term reflects a particular modern meaning given in the Oxford English Dictionary: “a story or representation used to give an explanatory or justificatory account of a society, period, etc.” Expanding on this definition, I would add that stories are not limited to simple chronologies of human events. A story may also be a song, joke, theory, explanation, or plan that has emotional resonance and that can easily be conveyed in casual conversation. We can think of history as a succession of rare big events in which a story goes viral, often (but not always) with the help of an attractive celebrity (even a minor celebrity or fictional stock figure) whose attachment to the narrative adds human interest.
For example, narratives from the second half of the twentieth century describe free markets as “efficient” and therefore impervious to improvement by government action. These narratives in turn led to a public reaction against regulation. There are of course legitimate criticisms of regulation as practiced then, but those criticisms were usually not powerfully viral. Viral narratives need some personality and story. One such narrative involved movie star Ronald Reagan, who became a household name as the witty and charming narrator of the highly popular US television show General Electric Theater from 1953 to 1962. After 1962, he entered politics in support of free markets. Reagan was elected president of the United States in 1980. In the 1984 reelection, he won every state except his opponent’s home state. Reagan used his celebrity to launch a massive free-markets revolution whose effects, some good and some ill, are still with us today.
Contagion is strongest when people feel a personal tie to an individual in or at the root of the story, whether a stock personality type or a real celebrity. For example, the narrative that Donald J. Trump is a tough, brilliant dealmaker and a self-made billionaire is at the core of an economic narrative that led to his unlikely election as US president in 2016. Celebrities sometimes concoct their own narratives, as in the case of Trump, but in many cases the celebrity’s name is merely added to an older, weaker narrative to increase its contagion—as in the story of the self-made man told many times over, each time with a different celebrity. (I discuss many celebrity-based narratives throughout this book.)
Narrative economics demonstrates how popular stories change through time to affect economic outcomes, including not only recessions and depressions, but also other important economic phenomena. The idea that house prices can only go up attaches to the stories of rich house flippers seen on television. The idea that gold is the safest investment attaches to stories of war and depression. These narratives have a contagious element, even if their attachment to any given celebrity is tenuous.
Ultimately, narratives are major vectors of rapid change in culture, in zeitgeist, and in economic behavior.4 Sometimes, narratives merge with fads and crazes. Savvy marketers and promoters then amplify them in an attempt to profit from them.
In addition to popular narratives, there are also professional narratives, shared among communities of intellectuals, that contain complex ideas that subtly affect broader social behavior. One such professional narrative, the random walk theory of speculative prices, holds that prices in the stock market incorporate all information, thus implying that attempts to beat the market are futile. This narrative has an element of truth to it, as professional narratives generally do, though there is now a professional literature that finds imperfections not predicted by the theory.
Occasionally these professional narratives translate into popular narratives, but the public often distorts these narratives. For example, one distorted narrative states that a buy-and-hold strategy in the domestic stock market is the best investment decision. That narrative conflicts with the professional canon, despite the popular idea that the buy-and-hold strategy comes from scholarly research. Like the popular interpretation of the random walk, some distorted narratives have an economic impact for generations.
As with any kind of historical reconstruction, we cannot go back in time with a sound recorder to capture the conversations that created and spread the narratives, so we have to rely on indirect sources. However, we can now capture the arc of contemporary narratives through social media and other tools, such as Google Ngrams.
Better Forecasts of Major Future Events
Most contemporary economists tend to think that public narratives are “not our field.” If you press them, they might suggest you check with other departments of the university, such as the journalism and sociology departments. But scholars in these other fields often find it difficult to tread in the land of economic theory, thus leaving a gap between the study of narratives and their effects on economic events.
No economist gave a credible forecast of the worldwide nature of the Great Depression of the 1930s before it happened, and only a handful predicted the peak of the US housing boom in 2005 or the “Great Recession” and “world financial crisis” of 2007–9. Some economists in the late 1920s argued that prosperity would reach new heights in the 1930s, while others argued the opposite extreme: unemployment would remain high forever, because labor-saving machinery would permanently replace jobs. But there seems to have been no public economic forecast of the actual events: a decade of very high unemployment and then a return to normal.
Traditionally, economists who study data have excelled in creating abstract theoretical models and in analyzing short-run economic data. They can accurately forecast macroeconomic changes a couple quarters into the future, but for the past half century, their one-year forecasts have been on the whole worthless. When assessing the probability that quarterly US GDP growth will be negative
one year in the future, their predictions have had no relation to actual subsequent negative growth rates.5 There have been, according to a Fathom Consulting study, 469 recessions (defined as a decline in a country’s GDP over a year) in 194 countries forecasted since 1988 by the International Monetary Fund in its biannual World Economic Outlook. In only 17 of these did they forecast a recession in the preceding year. They predicted recessions that did not occur 47 times.6
One might think that this forecasting record is good relative to that of weather forecasting, which is accurate for only a few days. But in economic decisions, people typically think years ahead. They plan to send children to high school or college for four years, and take out thirty-year home mortgages. So it is natural to suppose that we would sometimes know that the next few years will be strong or weak.
Maybe economic forecasters are doing the best they ever could do. But it seems that, with economic events coming again and again for no apparent cause, it would be a time to think whether economic theory could stand some fundamental improvement.
It is rare to see a professional economist, in interpreting the past or forecasting the future, quoting what a businessperson or newspaper writer thinks is going on, let alone what a taxi driver thinks. But to understand a complex economy, we have to take into account many conflicting popular narratives and ideas relevant to economic decisions, whether the ideas are valid or fallacious.
Criticism of traditional approaches to macroeconomic research is not new. In a famous 1947 article, “Measurement without Theory,” economist Tjalling Koopmans criticized the then-standard approach of looking exclusively at statistical properties of time-series data like GNP or interest rates to find leading indicators to help in forecasting. He asked for theories based on actual observations of underlying human behavior:
Narrative Economics Page 1