Under the Influence: Putting Peer Pressure to Work

By Robert H. Frank

From New York Times bestselling author and economics columnist Robert Frank, bold new ideas for creating environments that promise a brighter future

Psychologists have long understood that social environments profoundly shape our behavior, sometimes for the better, often for the worse. But social influence is a two-way street—our environments are themselves products of our behavior. Under the Influence explains how to unlock the latent power of social context. It reveals how our environments encourage smoking, bullying, tax cheating, sexual predation, problem drinking, and wasteful energy use. We are building bigger houses, driving heavier cars, and engaging in a host of other activities that threaten the planet—mainly because that's what friends and neighbors do.

In the wake of the hottest years on record, only robust measures to curb greenhouse gases promise relief from more frequent and intense storms, droughts, flooding, wildfires, and famines. Robert Frank describes how the strongest predictor of our willingness to support climate-friendly policies, install solar panels, or buy an electric car is the number of people we know who have already done so. In the face of stakes that could not be higher, the book explains how we could redirect trillions of dollars annually in support of carbon-free energy sources, all without requiring painful sacrifices from anyone.

Most of us would agree that we need to take responsibility for our own choices, but with more supportive social environments, each of us is more likely to make choices that benefit everyone. Under the Influence shows how.

• Language: English
• Publisher: Princeton University Press
• Release date: Oct 19, 2021
• ISBN: 9780691232713

Robert H. Frank

ROBERT H. FRANK is the H. J. Louis Professor of Management and Professor of Economics at Cornell University’s Johnson School of Management. He has been an Economic View columnist for the New York Times for more than a decade and his books include The Winner-Take-All Society (with Philip J. Cook), The Economic Naturalist, The Darwin Economy (Princeton), and Principles of Economics (with Ben S. Bernanke). He lives in Ithaca, New York.

Book preview

UNDER THE INFLUENCE

Under the Influence

Putting Peer Pressure to Work

Robert H. Frank

PRINCETON UNIVERSITY PRESS

PRINCETON AND OXFORD

Copyright © 2020 by Robert H. Frank

Preface to the paperback edition, copyright © 2021 by Robert H. Frank

Requests for permission to reproduce material from this work should be sent to [e-mail]

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

Library of Congress Control Number: 2021939856

First paperback printing, 2021

Paper ISBN 9780691227108

Cloth ISBN 9780691193083

ISBN (e-book) 9780691232713

Version 1.0

British Library Cataloging-in-Publication Data is available

Editorial: Joe Jackson and Jacqueline Delaney

Production Editorial: Lauren Lepow

Cover Design: Matt Avery / Monograph

Production: Erin Suydam

Publicity: James Schneider and Caroline Priday

Cover art: iStock

Example, whether it be good or bad, has a powerful influence.

—GEORGE WASHINGTON, MARCH 5, 1780

CONTENTS

Preface to the Paperbackix

PART I. INTRODUCTION

Prologue3

1 The Argument in Brief11

PART II. THE ORIGINS OF BEHAVIORAL CONTAGION

2 How Context Shapes Perception33

3 The Impulse to Conform48

PART III. CASES

4 It Was, Until It Wasn’t: The Dynamics of Behavioral Contagion63

5 The Sexual Revolution Revisited81

6 Trust97

7 Smoking, Eating, and Drinking113

8 Expenditure Cascades128

9 The Climate Crisis155

PART IV. POLICY

10   Should Regulators Ignore Behavioral Contagion?177

11   Creating More Supportive Environments191

12   The Mother of All Cognitive Illusions218

13   Ask, Don’t Tell232

EPILOGUE257

Acknowledgments265

Notes267

Index293

PREFACE TO THE PAPERBACK

Debate continues about where and when the first human SARS-CoV-2 infection occurred. But the virus was clearly not on anyone’s radar screen in mid-2019, when the hardcover edition of this book went into production. Early reports of the new pathogen, which appeared to surface in a Chinese fish and poultry market in December 2019, were slow to spark alarm in the West. But by the time Under the Influence was released late the next month, infections were spreading rapidly around the globe. On March 11, 2020, the World Health Organization declared COVID-19 a pandemic, and shortly thereafter, the United States and a host of other countries went into lockdown.

This timetable explains why the coronavirus received no mention in the hardcover edition of Under the Influence. But if the book had been published a year later, both the virus and our responses to it would have been among its central themes. As the New York Times columnist Ezra Klein wrote in his introduction to our May 2020 Vox podcast conversation,

Frank’s new book, Under the Influence: Putting Peer Pressure to Work, came out shortly before the coronavirus reshaped daily life. But for that very reason, the book is extraordinarily timely: It’s an effort to show that the economics of social contagion could reshape the world, solving our hardest problems—from climate change to income inequality—and offering new ways to think about the power we have as individuals.… We are watching a version of Frank’s thesis play out right now, in real time. In the wake of coronavirus, social pressure has driven perhaps the single fastest behavioral transformation in human history.¹

That transformation could not have happened nearly as quickly if most of us had not been so strongly predisposed to respond to behavioral cues from one another.

This paperback edition provides a welcome opportunity to sketch the connections between the book’s themes and our responses to the pandemic. It also affords a chance to comment on arguments offered by some who have opposed policy proposals in the book. Many of their criticisms resemble those leveled against recent pandemic public health measures.

---

There are deep similarities between viral contagion and behavioral contagion. For example, people in close or extended proximity to others infected by a virus are themselves more likely to become infected, just as people are more likely to drink excessively when they spend more time in the company of heavy drinkers. But there are also important differences between the two types of contagion. One is that visibility promotes behavioral contagion but inhibits the spread of infectious diseases. Solar panels that are visible from the street, for instance, are more likely to stimulate neighboring installations. In contrast, we try to avoid others who are visibly ill. Another important difference is that whereas viral contagion is almost always a bad thing, behavioral contagion is sometimes negative—as in the case of smoking—but sometimes positive, as in the case of solar installations.

As concerns public policy, perhaps the most important common feature of viral and behavioral contagion is that individual interests and collective interests are often fundamentally at odds. During the pandemic, for example, this conflict was the rationale for social distancing and mask recommendations. Although the personal costs of limiting one’s own social engagement are both immediate and vivid, the benefits accrue mostly to others. Thus, an infected person who ignores social distancing will infect an average of more than four hundred others during the following thirty days, whereas someone who reduces social contact by 75 percent will infect fewer than three others. Similarly, the principal benefit of mask wearing is that others are less likely to become infected.

An analogous conflict between individual and collective interest shows up in many spending decisions. Suppose, for example, that everyone initially owned cars weighing less than 2,500 pounds and someone then bought a 5,000-pound vehicle. That person would face less risk of injury and death than before, while all others would face more. Their best response might be to buy 5,000-pound cars themselves, in which case everyone’s risk would be higher than when all were driving smaller cars. As in the familiar stadium metaphor, when all stand to get a better view, no one sees better than if all had remained comfortably seated. In both instances, individually rational behavior is collectively irrational.

Taxing vehicles by weight would be a simple step to bring individual and collective interests into closer alignment. But that step would inevitably provoke some to complain that all taxation is theft, while others would object that such measures constitute social engineering. I respond to these objections in the book.² But there is a broader point to make here, which is that although dislike of taxes is widespread around the globe, Americans have been uniquely skeptical about all other forms of governmental regulation. As Ronald Reagan put it in his first inaugural address, Government is not the solution to our problem; government is the problem. In recent decades, that skepticism has given way to outright hostility toward government’s role as an instrument for collective action.

This hostility has been enormously costly. For example, it spawned not just the February 2021 Texas electric grid meltdown but also a long string of similar failures, including responses to Hurricane Katrina, the 2008 financial crisis, the COVID-19 pandemic, and the climate crisis. None of these challenges came without warning. For example, a similar cold-induced power grid failure had occurred in Texas in 2011, prompting the Federal Energy Regulatory Commission to warn that unless the state took steps to weatherproof its infrastructure, even more serious failures would be inevitable. Yet Texas not only failed to act, it also took steps to prevent the state’s power companies from connecting to the national power grid, thereby exempting them from federal regulation.³

More serious challenges await us. Referring to the lessons of the pandemic year, the novelist Ian McEwan recently wrote, Covid is our mass tutorial, our dress rehearsal for all the depredations as well as tragedies that the climate emergency could bring.⁴ As David Wallace Wells described in his 2019 book, the record storms, droughts, and wildfires of late offer only a hint of far more serious threats in store if current warming trends continue.⁵ Collective action on a grand scale offers our only hope of parrying those threats.

The intense resistance of many Americans to science-based public health recommendations during the pandemic suggests that this hope rests on a thin reed indeed. As captured on bystander videos posted on social media, for example, collective measures like mask ordinances have routinely provoked violent confrontations. In May 2020, Calvin Munerlyn, a security guard at a retail outlet in Flint, Michigan, was shot and killed by the husband of a shopper angered by Munerlyn’s attempts to enforce the store’s mask requirement.⁶ Measures to limit travel or attendance at certain types of events to people who have been vaccinated against the virus arouse similarly intense pushback.

Unless we can overcome the political challenge posed by such antigovernment hostility, there is no realistic path forward. One fruitful response, I believe, is to scrutinize the arguments of antigovernment activists in their own terms. Many of them appear rooted in a belief that government is inherently inefficient. Milton Friedman, the late economics Nobel laureate and patron saint of small government conservatism, once quipped that if the federal government were put in charge of the Sahara Desert, in five years there would be a shortage of sand. That there are bridges to nowhere and other vivid examples of inefficiency in government goes without saying, of course. But as even a cursory examination of government budgets reveals, the bulk of public spending goes for services that most citizens value. More to the point, there is also substantial waste in the private sector. Much of that waste occurs when individual spending choices generate negative spillover effects. Buying 5,000-pound cars when 2,500-pound cars would be better for almost everyone is an example: it is waste, pure and simple.

What about regulations that limit individual action? Free-market enthusiasts often cite Adam Smith’s Invisible Hand theory in support of their claim that unregulated markets will harness the interests of self-interested individuals in the service of society as a whole. This is a remarkable claim indeed, and Smith’s contribution was to identify the stringent conditions under which it might be expected to hold. But as he also understood clearly, individual and collective interests often diverge sharply.

Resolving such conflicts is by far the most important reason for government regulation. Although former president George W. Bush once said, We don’t believe in planners and deciders making the decisions on behalf of Americans, that is of course exactly what all societies entrust government bureaucrats to do. The reason is straightforward: Even if people are just as rational and markets are just as competitive as Milton Friedman believed, the incontestable fact is that individually rational actions often yield demonstrably bad outcomes. That simple fact has always been, and will remain, a compelling rationale for government involvement in economic life.

We require catalytic converters on cars, for instance, because each individual’s decision to install one would be costly but would yield no measurable impact on air quality. Yet when everyone installs a catalytic converter, the benefit of the resulting air quality improvement far outweighs the corresponding cost. By the same token, as I argue in chapter 7, the most compelling reason for having taxed cigarettes heavily is not, as regulators once claimed, to protect people against exposure to secondhand smoke. Rather, it is to have made smoking less socially contagious. Although the strongest predictor of whether people will become smokers is the percentage of their friends who smoke, few individuals would refrain from smoking out of concern that that they might make others more likely to take up the habit. Yet that side effect is by far the biggest injury caused by any smoker. Cigarette taxation encourages us to behave as if we cared about it.

Are such measures legitimate? As even the staunchest libertarians concede, government may discourage us from inflicting undue harm on others under at least some circumstances. In chapter 10 and elsewhere in the book, I try to explain why regulations that attempt to discourage harmful forms of behavioral contagion are often justifiable under this rationale. But the vehemence with which many have opposed even clearly evidence-based public health measures makes me now wish I had devoted more space to this issue.

Where people who oppose government efforts to change individual behavior and I often part company is in how we think about what constitutes undue harm to others. We all agree that it is legitimate for government to restrain people from stealing others’ property or from committing gratuitous acts of physical violence against them. Most of the cases in dispute concern the more indirect forms of harm that are my focus in the book.

Prohibition against anabolic steroid consumption by competitive athletes is an instructive case in point. Although sprinters who take steroids may make no physical contact with their closest rivals, they nonetheless impose heavy costs on them. Rivals can either abstain from taking steroids, in which case they forfeit their chance to win and any return on their substantial investments of time and effort. (Steroids have been estimated to lower men’s times in a one-hundred-meter sprint by over two-tenths of a second, more than twenty times Usain Bolt’s winning margin in the 2015 Beijing World Athletics Championships final in that event.)⁷ Alternatively, rivals can restore the competitive balance by consuming steroids themselves, thereby courting serious long-term health risks.⁸

Either way, they would suffer substantially greater harm than if they had been lashed with a whip or had a bicycle stolen. Yet many opponents of regulation insist that it should be sprinters’ right to take performance-enhancing drugs. Why should that right trump the right of others to escape the resulting harm? Why shouldn’t harm count merely because it is indirect? Confusion about this point arises in part because indirect harm is often harder to measure than direct harm. But direct harm is often hard to measure, too, and in those cases there is usually no debate about whether it should count. People could, of course, claim for themselves an unfettered right to injure others whenever and in whatever ways it might suit them. But I hope that would be a bridge too far for anyone wanting to be taken seriously in public debate. Good-faith participants in policy deliberation will want to offer carefully considered reasons for any claim that indirect forms of harm should not count.

Even for those who agree that discouraging indirect forms of significant harm to others is a legitimate basis for collective action, there remains the question of how to intervene most effectively. Because there are almost always multiple ways of discouraging harm, my claim is that respect for personal autonomy dictates choosing the option that preserves the broadest latitude for individuals to do as they please.

As in an example discussed in chapter 11, suppose toxic factory emissions could be reduced in either of two ways: by requiring each factory to cut back its emissions by half, or by charging producers sufficiently high emissions fees to achieve the same aggregate result. If, as will almost always be true, some producers can reduce their emissions more cheaply than others, the far better approach will be to charge emissions fees.⁹ That approach lets firms decide for themselves how much pollution to eliminate. Given this freedom, those that can limit emissions cheaply will choose to cut back a lot, while others will find it cheaper to continue polluting and pay the required fees. As a result, we achieve the aggregate emissions target at the lowest possible cost.

With firms, of course, there is less reason than with individuals to worry that someone’s feelings might be hurt by being told what to do. But the analogy is nonetheless instructive. Rules that restrict individual freedom impose greater psychological burdens on some people than on others. Instead of adopting flat prohibitions of harmful behavior, we could impose taxes, fines, or other disincentives to engage in such behavior. The latter approach will generally reduce aggregate harm at the lowest possible total cost and in the most equitable manner. Those who can reduce harm cheaply will alter their behavior a lot. Others, for whom behavioral changes would be far more costly, will continue to cause harm and be penalized accordingly. And each dollar raised by a tax on a harmful activity is a dollar less that we need to raise from other taxes.

In frontier days, there were relatively few demands for government to restrict individual behavior. Because people didn’t bump into their neighbors nearly as often, they had fewer opportunities to cause harm to one another. Society is not only much denser now but also much more highly interactive, and our emissions are far more damaging. Opportunities to harm one another are now far more abundant. Such opportunities will continue rising sharply and will generate additional demands for regulation.

More forceful advocacy of individual liberty will be sorely needed as government’s role continues to expand. Those who insist that restrictions on individual behavior are almost never legitimate have diminished the credibility of others who would argue most forcefully for governance with a light touch. And in so doing, they have imposed high costs on all of us. They have blocked many forms of desperately needed regulation and are indirectly responsible for design flaws in many of the regulations that have been enacted.

As the pandemic recedes, we will face a choice between two starkly different futures. In Society A, we will keep building bigger houses and cars and keep hosting ever more elaborate and expensive celebrations of special occasions; we will continue skimping on investment in climate mitigation and keep giving short shrift to infrastructure, vaccine research, and hospital surge capacity. Alternatively, we could choose Society B, where we would spend less on many forms of private consumption and significantly more on public investment.

Findings from decades of careful research in the behavioral sciences are helpful in thinking about this choice.¹⁰ One of the most robust findings from this large and contentious literature is that beyond a certain point—one that has been long since passed in the West—across-the-board increases in many forms of private consumption do little more than raise the bar that defines what people consider adequate. If all mansions were to double in size, for example, the people living in them would be neither happier nor healthier than before. In contrast, the same literature finds that public investments like the ones listed in the preceding paragraph would create meaningful improvements in most people’s lives.

As discussed in chapters 8 and 12, these findings, which are also consistent with recent experience under the pandemic, support two conclusions: People would be both happier and healthier in Society B than in Society A; and the additional tax revenue to finance the investments necessary to sustain Society B would require fewer painful sacrifices than most people appear to think.

Problems like climate change, economic inequality, and racial injustice strike many as intractable, even in the face of broad and well-funded collective action. Even more people regard individual action in response to such threats as fruitless. That’s a plausible position, since the concentration of CO2 in the atmosphere would not decline measurably even if a large family reduced its emissions all the way to zero. My greatest hope, therefore, is that the evidence I describe in the coming chapters will persuade anyone holding these views to reexamine them critically. The direct impact of individual steps like changes in diet are indeed inconsequential, but behavioral contagion, harnessed intelligently, can amplify that impact by many orders of magnitude. More important, behavioral contagion has great power to affect not only which leaders we choose but also which policies we empower them to enact.

What is clear, in any event, is that continuing to elect candidates who are hostile to government’s role in solving collective-action problems ensures continued failure to confront the multiple serious challenges we face. Fortunately, such candidates will not be the only ones on upcoming ballots.

PART I

Introduction

Prologue

That we are more closely connected to one another than most of us realize is an old idea. An important variation dates from 1929, when the Hungarian writer Frigyes Karinthy published a short story titled Láncszemek (Chains). Two of the story’s main characters speculate that any two living people can be connected by a chain involving no more than five acquaintances. That some such chains exist is hardly surprising, since most of us can think of people in distant locations to whom we are connected through a small number of intermediaries. The man-bites-dog aspect of Karinthy’s conjecture is that almost any two people A and E could be linked by a chain like A knows B, who knows C, who knows D, who knows E.

One of the first systematic attempts to test this claim came in the 1960s with a series of experiments by the psychologist Stanley Milgram. In one, he sent packages containing a small booklet to ninety-six people chosen at random from the Omaha, Nebraska, telephone directory. His cover letter asked them to try to forward the booklet to a specific resident of Boston, Massachusetts, through a chain of personal acquaintances. He told them the name and address of the Boston resident, that he worked as a stockbroker, and that the first person in their chain should be someone they knew on a first-name basis. Milgram also advised them to choose someone they believed (presumably based on the target’s location and occupation) could be socially closer to the target than they themselves were. Subsequent recipients in the chain were asked to forward the same instructions.

Many of the Omaha recipients undoubtedly tossed Milgram’s booklet in the trash, so it is remarkable that the Boston target received eighteen of the ninety-six packages. The average number of links in the eighteen chains was 5.9. But the now-familiar expression six degrees of separation would not gain broad currency until many years later, when John Guare’s play by that name premiered on Broadway in 1990.

The concept became a meme in full when four Albright College students introduced Six Degrees of Kevin Bacon in 1994, a game designed to measure the professional proximity of an actor, living or dead, to the American film star Kevin Bacon. An actor has a Bacon number of 1, for example, if she or he appeared in the same film with Bacon. Someone who appeared in a film with another actor who appeared in a film with Bacon has a Bacon number of 2, and so on. The average Bacon number is 2.955 among actors who have one. In that group, even the actor most distant from Bacon, William Rufus Shafter, has a Bacon number of only 7. Shafter, a Union Army officer during the Civil War, appeared in two films in 1898.

For academics who study social connectedness, the six-degrees concept gained little traction until 1998. That’s when the sociologist Duncan Watts and the mathematician Steven Strogatz published their landmark paper, Collective Dynamics of Small-World Networks, in Nature. In the years since then, this paper has provided the mathematical foundation for the analytical tools that social scientists have been using with such remarkable success to study how ideas and behaviors spread through populations like infectious diseases. It has already been cited by other scholars more than thirty-eight thousand times and is one of the few papers ever published to be among the most widely cited works in multiple disciplines.

In his 1976 book, The Selfish Gene, the evolutionary biologist Richard Dawkins coined the term meme, which Webster now defines as an idea, behavior, style, or usage that spreads from person to person within a culture. The meme is to cultural transmission, Dawkins argued, as the gene is to biological transmission.

One of Charles Darwin’s central insights was that natural selection favors genetic variations that enhance the individual organism’s ability to survive and reproduce. Much of the time, the same variations also benefit larger groups. But often not. For example, an inclination to cheat when no one is looking might benefit individuals, but widespread cheating almost always makes matters worse for groups. It is the same with memes. The memes that propagate most successfully are often ones that benefit both individuals and groups. But here, too, not always. As the legal scholar Jeffrey Stake has argued, Ideas should not be treated as inert products but as living things that sometimes exert some influence over their environment. Some of the ideas are more adept at surviving than others, and the ones that survive will not necessarily be good for humans.¹

It is often hard to evaluate whether a specific behavior even qualifies as a meme, and if so whether its consequences, on balance, are positive or negative. Yet sometimes the evidence is clear.

We know, for example, that the strongest predictor of whether people will take up smoking is the proportion of their close friends who smoke. Smoking is thus clearly a meme. The negative health consequences of smoking are also conclusively documented, and most smokers themselves express regret about having started. By definition, then, smoking unambiguously qualifies as a socially destructive meme.

On the other side of the ledger, we have compelling evidence that the adoption of photovoltaic solar panels is both socially contagious and almost uniformly positive in terms of environmental consequences. Accordingly, few would object to calling the practice a socially beneficial meme.

Adam Smith, widely considered the father of economics, is often cited in defense of the claim that competitive markets produce the greatest good for the greatest number. But that was never Smith’s position. His signature insight was that the pursuit of narrow self-interest often leads to socially beneficial outcomes, but not always. The same holds for competition among ideas. Good ideas often triumph, but there is no presumption that the marketplace of ideas reliably promotes the common good, especially in the short run. My central claim in Under the Influence is that we therefore have a powerful and legitimate public policy interest in encouraging socially beneficial memes and discouraging socially harmful ones.

Many insist it is solely the individual’s responsibility to choose which ideas to embrace, which behaviors to mimic. Some who hold that view are likely to oppose my claim. I am sympathetic to the sentiments that underlie their position. No one wants to live in an Orwellian nanny state. But my aim will be to explain not only why it is compellingly in our interest to exert at least some collective control over the social forces that shape our choices, but also why failure to do so may threaten our very survival.

The existential threat we face is the climate crisis. In October of 2018, the United Nations’ Intergovernmental Panel on Climate Change reported that in the absence of decisive measures to curtail greenhouse gas emissions, we are now on track to catastrophic increases in average global temperatures by 2040. The forecasts of climate models are notoriously imprecise. Temperatures might well rise by less than predicted, but they might also rise by considerably more. And even the current temperature rise of only 1°C (1.8°F) has already produced floods, droughts, and fires on a scale unseen in human history. As David Wallace-Wells put it in the first sentence of The Uninhabitable Earth, It’s worse, much worse, than you think.² If temperature increases come anything close to the IPCC’s projections, many hundreds of millions of people will die, and much of the earth’s wealth will be destroyed.

Reactions to this threat vary greatly from country to country. In the United States, supporters of the Green New Deal have proposed an expansive legislative agenda that attacks climate change and economic inequality simultaneously. Critics, including many from the left, have objected that tackling both problems at once will simply make failure in both domains more likely. Green New Deal proponents counter that unless we include policies to mitigate rising inequality’s impact, it will be impossible to assemble a sufficiently broad political coalition to overcome our current gridlock.

A deeper understanding of the power of behavioral contagion, which psychologists define as a tendency to mimic others’ behavior, suggests that the Green New Deal is less impractical than many critics think. One of the most costly ways we influence one another is by reinforcing individual spending decisions that are highly wasteful. A simple example: we buy heavier cars because driving a relatively light car is more dangerous, yet when all buy heavier cars, everyone’s risk of injury and death goes up, not down. Understanding how contagion amplifies such spending patterns, we will see, helps identify simple policies that would redirect trillions of dollars annually in support of carbon-free energy sources, all without demanding painful sacrifices from anyone. Those same policies would reduce economic inequality and stimulate the creation of good jobs. Or so I will argue.

But before launching that attempt, I digress briefly about my choices of what to include in the pages ahead. Most books on writing celebrate brevity. Rule 17 in the renowned Elements of Style, for example, exhorts writers to Omit Needless Words. For the most part, I try to follow this rule, both sentence by sentence and in my choice of topics to omit. But for reasons I will explain shortly, I have also included material whose omission