Worst-Case Economics: Extreme Events in Climate and Finance

By Frank Ackerman

Why do climate and financial crises pose such extreme risks? And what does it take to respond effectively to those risks? Extreme weather events – storms and sea-level rise, heat waves, droughts and floods – seem ever more common and extreme, while scientists warn of even greater climate risks ahead. Financial failures on the scale of 2008 make a mockery of the supposed efficiency of the market economy. None of this would be possible in the world as imagined by conventional economics – an imaginary land of gradualism, equilibrium, well-informed rationality and the win-win solutions dealt by the invisible hand.

The erratic rhythm of boom and bust in financial markets could be explained either by the patterns of crowd-following behaviour among investors, or by the unequal distribution of wealth (and the impact of the largest investors on the markets). Climate crises reflect the fact that natural systems can reach tipping points or critical transitions, where gradual change gives way to large-scale discontinuous changes. The economics of climate change has lagged behind the science, understating the severity of the problem and the likelihood of a crash.

While the causes of climate and financial extremes are distinct, the implications for public policy have much in common. The frequency of extreme events, of varying sizes, means that there is no way to predict the likely size of future crises. The traditional approach to risk aversion cannot account for longstanding patterns in financial markets. Better theories of risk call for more precautionary approaches to both financial and climate policy. In the frequent cases in which potential outcomes have unknown probabilities, the best policy is based on the worst-case credible scenario. When a single catastrophic risk commands everyone’s attention, a World War II-style, costs-be-damned mobilization is the right response. There is no formula for perfect responses to extreme risks, but there are important guideposts that point toward better answers.

• Language: English
• Publisher: Anthem Press
• Release date: Oct 23, 2017
• ISBN: 9781783087099

Book preview

WORST-CASE ECONOMICS

WORST-CASE ECONOMICS

EXTREME EVENTS IN CLIMATE AND FINANCE

FRANK ACKERMAN

Anthem Press

An imprint of Wimbledon Publishing Company

www.anthempress.com

This edition first published in UK and USA 2017

by ANTHEM PRESS

75–76 Blackfriars Road, London SE1 8HA, UK

or PO Box 9779, London SW19 7ZG, UK

and

244 Madison Ave #116, New York, NY 10016, USA

© Frank Ackerman 2017

The author asserts the moral right to be identified as the author of this work.

All rights reserved. Without limiting the rights under copyright reserved above,

no part of this publication may be reproduced, stored or introduced into

a retrieval system, or transmitted, in any form or by any means

(electronic, mechanical, photocopying, recording or otherwise),

without the prior written permission of both the copyright

owner and the above publisher of this book.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library.

ISBN-13: 978-1-78308-707-5 (Hbk)

ISBN-10: 1-78308-707-2 (Hbk)

This title is also available as an e-book.

To my grandson, Ronan, born while I was writing this book, too young to understand the madness of this moment. May he grow up in a sane and sustainable world.

Economists set themselves too easy, too useless a task, if in tempestuous seasons they can only tell us, that when the storm is long past, the ocean is flat again.

—John Maynard Keynes¹

1 Keynes ( 1923 ).

Contents

List of Figures

CHAPTER 1Introduction

CHAPTER 2Steam-Engine Economics

CHAPTER 3Beyond Homo Economicus

CHAPTER 4Big and Dirty

CHAPTER 5Pictures of Improbability

CHAPTER 6Trillions, or Only Hundreds?

CHAPTER 7Zipf’s Law and Other Stories

CHAPTER 8Ants and Traders

CHAPTER 9Too Big to Ignore

CHAPTER 10Climate Tipping Points and Known Unknowns

CHAPTER 11Predators and Prey

CHAPTER 12Good Enough for Government Work

CHAPTER 13Fat Tails and the Failure of Forecasting

CHAPTER 14Misunderstanding Risk

CHAPTER 15Choices Beyond Calculation

CHAPTER 16Who Won World War II?

CHAPTER 17Conclusion

Acknowledgments

Bibliography

Index

Figures

FIGURE 1Northwest Atlantic cod harvest

FIGURE 2Probability distribution for the sum of two dice

FIGURE 3Probability distribution for the sum of ten dice

FIGURE 4The normal distribution

FIGURE 5Height of US women, ages 30–39

FIGURE 6Daily percentage change in S&P 500, 1950–2013

FIGURE 7Daily percentage change in S&P 500 (bottom of Figure 6, enlarged)

FIGURE 8Daily percentage changes in S&P 500, 2006 and 2008

FIGURE 9Annual variation in S&P 500 and in normally distributed data, 1950–2013

FIGURE 10The tails of normal and fat-tailed distributions

FIGURE 11Normal versus power law distributions, ratio scale

FIGURE 12The same curved line, shown in conventional and ratio graphs

FIGURE 13Absolute value of daily percentage changes in S&P 500

FIGURE 14The 2003 heat wave and the normal distribution

FIGURE 15Frequency of the most common words in Shakespeare

FIGURE 16Distribution of top incomes in the United States and Japan, 2010

FIGURE 17Temperatures in central Greenland, reconstructed from ice core data

FIGURE 18Policy choices as a game with nature

FIGURE 19The maxmin criterion: a numerical example

FIGURE 20The minimax regrets criterion: the same numerical example

CHAPTER 1

Introduction

Most of the time, there are no speculative bubbles bursting or financial markets crashing. The abrupt loss of millions of people’s jobs, homes and retirement savings is not a frequent occurrence.

Most of the time, hurricanes are not battering Houston, Miami, New Orleans and other vulnerable places. It is rare for the world to reach a tipping point into an irreversibly worse climate.

It is quite likely that none of these worst-case events are happening on the day when you are reading this book, or at any other single point in time. But none of these disasters are improbable enough to ignore.

Financial and climate crises may look like an odd couple for combined discussion. While both involve episodic crises, their rhythms of instability are different. Financial crises are fast-moving, repeated and, over enough years, reversible. Climate crisis is slow-moving, cumulative and could become irreversible on any human time scale.

Yet there are also deep similarities. Both involve rare, costly extreme events, with risks of huge losses that individuals cannot escape on their own. Both are created by short-sighted human activity, pursuing immediate economic gain at the expense of long-term consequences.¹ Both are misunderstood by conventional economics, which minimizes their importance and discourages the development of sensible public policy responses. And both face major industries with a vested interest in the status quo, a powerful obstacle to the adoption of policies that could prevent future crises.

This comparison of the two fields grew out of my research on the economics of climate change. In that research, an analysis of financial markets, focusing on the treatment of extreme risks, turned out to be remarkably useful in understanding potential climate outcomes.² In both fields, there is a need for a new and better economics of risk and responses. Protection against both climate and financial crises requires policies that bear some resemblance to insurance, based on a new understanding of risk and innovative ways of thinking about decision making under uncertainty.

The goal of this book is to understand the patterns of extreme events and the policies needed to address the risks related to these events. How should we respond to the fact that high-cost losses in both climate and finance are not as rare as we might have hoped or expected?

Costs of crisis

The Great Recession threw almost nine million people out of work; US employment did not return to its 2007 level until 2014.³ And many who remained employed ended up with lower wages and/or fewer hours of work than before. Youth unemployment was particularly severe, with long-term consequences for those who entered the labor market during the downturn. Research on prior years has found that those graduating from college during a recession experience significant negative effects on wages—effects that may last for as long as 15 years.⁴ Those who were already of working age during the Great Recession, meanwhile, are likely to have permanently lower retirement incomes, as their job losses and slower wage growth ripple through the formulas for future Social Security payments.⁵ For those who had been saving for retirement, the problem is compounded by the collapse in the value of homes and financial assets.

More than nine million households lost their homes between 2006 and 2014; many of them may never be homeowners again.⁶ House values fell by a nationwide total of $5.5 trillion; in 2011, 11 million households owed more on their mortgages than the current market value of their homes.⁷ Housing is the largest asset owned by most people, so the crash in house values had a devastating effect on wealth in general. Between 2007 and 2011, one-fourth of American families lost at least 75 percent of their net worth; more than half lost at least 25 percent.⁸ Losses were worst, in percentage terms, for low-income and minority households.⁹

The United States has had numerous weather disasters for which the country was ill-prepared, including Hurricane Katrina drowning New Orleans in 2005, years of record-breaking drought in California, Hurricane Sandy clobbering the New York area in 2012, Hurricane Harvey inundating Houston in 2017 and floods along the Mississippi and other major rivers. Elsewhere in the world, the impact of extreme weather is even worse. The worst recent climate-related losses in high-income countries occurred in the European heat waves of 2003 and 2010, discussed in Chapter 6. Storm damage is proportionally greater, and resources for protection and recovery are more limited, in low-income coastal countries such as Bangladesh and the Philippines, as well as in small island nations that are threatened with complete inundation.

The cause of these climate trends is well known, even though individual events are unpredictable in detail. Rising greenhouse gas emissions, primarily due to combustion of fossil fuels, are warming the world and destabilizing the climate. Larger, irreversible damage could occur in the not-so-distant future, although the timing and magnitude of such threats remain uncertain. The direction of change, however, is clear: catastrophic climate losses, although still unlikely today, become steadily less unlikely as the world grows warmer.

Beyond whispers and tweets

Imprecise predictions of extreme events are no excuse for inaction. Ample warnings have repeatedly been ignored by policy makers. Four years before Hurricane Katrina, a detailed description of the vulnerability of New Orleans and the need for improved storm defenses appeared in Scientific American.¹⁰ Seven years before the housing, subprime mortgage and stock market bubbles burst, Robert Schiller, a prominent economist, published a widely cited book with a title that tells the story: Irrational Exuberance.¹¹ Alongside the puzzling origins of extreme events, there is the mystery of why public policy remains feckless in the face of such well-anticipated harms.

Both climate and financial risks are misunderstood by the conventional economics of public policy. The policy choices of recent years have failed to provide adequate protection in either sphere. The Obama administration spoke eloquently about both financial and climate crises, but its actions only whispered. The Trump administration, if it understood French, could tweet après moi le deluge, as it attacks the rules that restrain both financial fraud and global warming. Both the earlier, timid, and the later, destructive, responses reflect the power of corporate lobbies that oppose limits on speculative investments or fossil fuels. Theories that dismiss obvious risks, endorsing little or no response, are convenient for (and sometimes subsidized by) those who profit from the status quo.

In terms of solutions, this book argues that cost–benefit analysis, a standard tool of policy evaluation, is not a useful approach to policies addressing extreme risks. Among other limitations, this type of analysis fails because for many extreme risks there is no such thing as a meaningful average or expected value, and because standard theories misunderstand, and badly underestimate, the importance of risk aversion. More appropriate alternatives call for more precautionary approaches to major risks. Decision making without numbers, the realm in which we inevitably find ourselves, calls for policies based on the credible worst-case outcome—an approach similar, though not identical, to the precautionary principle. In the very worst cases, when facing an existential threat, total mobilization of resources is called for; this approach is valuable when valid, but dangerously easy to abuse.

Once upon a textbook

Economics matters, not only to economists. It has become the language of public policy, providing quantitative analysis and intellectual authority on complex political questions. The economics of policy debate, however, often relies on ideas that are simpler and more traditional than leading economic theorists now endorse. The old-time religion of free-market theory seems entrenched in the intuition and practice of politics in America and has powerful advocates elsewhere. Cost–benefit analysis, increasingly seen as a gatekeeper for policy approval, typically starts from the assumption that the current state of affairs is close to being optimal and sustainable. Risks of extreme events are often implicitly excluded in an approach that focuses on average or most likely outcomes.

Although it is no longer state of the art in research, naïve policy economics has a lengthy academic pedigree. The first part of this book explores the origins and limits of traditional economics. There have been other critiques of the subject, raising a number of the points discussed here. But to round out the understanding of what is wrong with conventional economics, two key features deserve greater attention. First, there is a tendency to dismiss extreme events—a blind spot that has been present ever since the nineteenth-century origins of the theory. Second, there is a growing disconnect between advanced academic and pedestrian policy-oriented styles of economics. Understanding public policy debates therefore requires looking back into the origins of the field.

In the beginning, in the 1870s, there was a powerful but flawed analogy to nineteenth-century physics, as explained in Chapter 2. Gas molecules collide, exchanging energy as they approach equilibrium; firms and households encounter each other in the marketplace, exchanging goods and services as they approach economic equilibrium. This analogy created the framework of economics as it was taught and thought about for much of the last century—and as it is often applied to public policy, even today.

The analogy to physics was never a complete success. Economic particles were assumed to be individuals following their own idiosyncratic preferences. This assumption created a satisfying story about political liberty. At the same time, it lost the rules-based predictability and simplicity of the comparable theory of the dynamics of physical particles. The final, rigorous form of Adam Smith’s story of the invisible hand, which emerged in the mid-twentieth century, depended on at least three sets of unrealistic assumptions.

First, the marketplace was assumed to be populated by homo economicus, an imaginary species of completely asocial, selfish, perfectly informed and perfectly calculating individuals. In contrast, real people are inescapably social, incompletely informed and imperfectly rational, as discussed in Chapter 3. Broadly speaking, this helps to explain why there are so many individual and collective decisions that in hindsight appear wrong—in financial markets, in climate and environmental policies and elsewhere. But the ability to transcend selfish, asocial calculation is also part of the solution to our current crises: real people are capable of compassion, and of taking actions that benefit others as well as themselves.

Second, homo economicus inhabits an equally imaginary landscape, where goods and services are produced and sold by enterprises too small to exert any market power. Meanwhile, a problem-free natural environment provides fresh air, forests, tolerable temperatures and other services in abundance. It is a far cry from the modern world of inevitably large firms and pervasive pollution, reviewed in Chapter 4. Climate and financial crises are results of big and dirty reality and would not occur in a small, clean utopia.

Third, the original analogy to physics suggests that fluctuations in the economy, such as the ups and downs of financial markets, might follow a bell curve, or normal distribution. If this were true, it would paint a comfortably subdued picture of uncertainty, one in which extreme events are vanishingly rare and could safely be ignored. Chapter 5 looks at the bell curve and shows that it clearly does not fit the pattern of changes in stock market prices.

Chapter 6 introduces an alternative picture, known as the power law, in which extreme events are much too common to ignore. (In this picture, the probability of an event is inversely proportional to some power, such as the cube, of the magnitude of the event. Mathematical understanding of the power law is not required, except to note that it implies that dangerously extreme events can be dangerously likely.) This unruly picture of uncertainty provides a much better fit to the data, both in financial markets and in extreme weather events such as hurricanes and heat waves.

Stuff happens

The second part of the book asks why extreme events are so common. As Chapter 7 explains, there are at least four possible explanations for the ubiquity of power laws. First, it could be simply a statistical artifact, an error introduced in data analysis. Second, as natural systems approach tipping points and phase changes, such as melting or freezing, these systems are known to flicker in a manner that displays power-law patterns. This could be important for climate extremes. Third, the sandpile story, a modern metaphor, suggests that some simple systems continually return to the edge of instability, with the resulting episodes of instability following a power law. Finally, the story of intermittency—such as the challenge of balancing a stick on your fingertip—leads to similar patterns, in a manner that may be particularly relevant to financial crises.

The next two chapters offer two independent explanations of extremes in financial markets. Either explanation could, on its own, cause the observed patterns of boom and bust. In Chapter 8, social interactions among traders—reflecting the ways in which real people do not behave like homo economicus—are enough to create recurring financial bubbles and crashes. Traders influence each other’s strategies, which leads at times to waves of abrupt change. In this interpretation the instability of the market is analogous to the stick-balancing challenge, and the future will always bring another wobble, if not a fall.

Inequality alone could also explain financial instability, as seen in Chapter 9: the wealthiest traders cannot always avoid rocking the boat as they climb in and out of the market. In this story, extreme market outcomes simply reflect extreme concentrations of wealth, a well-known and longstanding pattern. Those concentrations of wealth call for some explanation: they would arise naturally in a different story about market interactions, a sociological theory of continual conflict over resources.

Climate risks are the subject of Chapter 10. Nature becomes fragile when pushed too hard in the wrong places, as the fossil-fueled economy seems determined to demonstrate. Numerous potential tipping points have been identified, points at which large, irreversible damages are expected to occur. Yet measurement of climate risks is constantly frustrated by known unknowns. Unlike financial risk, which involves recurring events, the most serious climate risks center on the unknown likelihood of future one-time events. Despite this difference, similar questions arise in both fields about the probability of extreme outcomes. The timing and magnitude of climate risks remain unknown, although inaction makes the risks grow more severe. The early warnings of those risks are already more than enough to motivate immediate and vigorous measures to reduce carbon emissions.

Predator–prey models from biology offer a different image in Chapter 11, suggesting an analogy both for Hyman Minsky’s theory of financial instability and for the economics of climate change. The asymmetric rhythms of predator and prey arise when debts devour incomes in Minsky’s world, or when greenhouse gases in the atmosphere prey on output in climate economics.

Standard climate economics models often come with optimism built in, using a framework that rules out a climate-caused slowdown in productivity and growth rates. Newer research, finding that climate affects productivity and growth, predicts deeper, longer-lasting damages. The persistence of carbon dioxide in the atmosphere means that today’s emissions will contribute to climate change far into the future. Thus, a crash rather than a continuing cycle becomes a likely result in the predator–prey picture of the fossil-fuel-burning economy.

Making decisions in the dark

The causes and the dangers of extreme events are not just matters of intellectual curiosity. The final section of the book turns to the appropriate responses, the approaches to public policy that are required in a world of worst-case economics.

The status quo in policy analysis, as seen in Chapter 12, misstates or misunderstands the importance of climate and financial extremes. Complex economic models often make the implausible assumption that full employment is the norm both before and after a policy change, preventing analysis of the most urgent and controversial issues. Cost–benefit analysis, seemingly a matter of common sense, errs in assuming that everything of value has a meaningful price, that all new costs are undesirable and that uncertainty can be dealt with by using average or expected values.

How large are the risks of extreme events? As Chapter 13 explains, there is no well-defined answer. Cost–benefit analysts have attempted to estimate the average size of financial crises—a number that appears to be needed to evaluate policies that address such crises. But, due to the frequency of extreme events, the search for the magnitude of a typical crisis inevitably fails. Likewise, there is no precisely predictable size for future climate damages. Economic analysis of climate change extrapolates to temperatures far beyond actual experience—all too often relying on controversial assumptions that suppress or minimize worst-case risks. One influential study that takes worst-case risks seriously finds that the value of emissions reduction is literally infinite, a conclusion with paradoxical implications.

A different paradigm for responding to worst-case risks arises in insurance, where people routinely buy policies that lose money more than 99 percent of the time. Chapter 14 shows that risk aversion is a subtle phenomenon. The most common story about risk is inadequate in several respects and cannot solve the puzzle of the enduring popularity of low-interest government bonds (compared to stocks, which are almost always much better long-term investments). Three proposed solutions to the puzzle all lead to new insights about risk—and all three call for more precautionary approaches toward both financial and climate policy.

Decision-making about extreme events inevitably occurs in the absence of hard numbers measuring the probability or magnitude of the relevant risks. Approaches to this problem, the subject of Chapter 15, include a rigorous theory of the economics of ignorance, a theory that resembles a revised version of the precautionary principle. The worst-case forecast made by a credible expert turns out to be decisive—a conclusion that avoids the impossible search for the perfect forecast, but leads to difficult debates about who qualifies as a credible expert.