Pandora's Risk

Uncertainty at the Core of Finance

Kent Osband

• Publisher: Columbia University Press
• Year: 2011
• ISBN:
  9780231525411
• Topic: Business
• Subtopic: Financial Risk Management
• Index: Business

Author of the acclaimed work Iceberg Risk: An Adventure in Portfolio Theory, Kent Osband argues that uncertainty is central rather than marginal to finance. Markets don't trade mainly on changes in risk. They trade on changes in beliefs about risk, and in the process, markets unite, stretch, and occasionally defy beliefs. Recognizing this truth would make a world of difference in investing. Belittling uncertainty has created a rift between financial theory and practice and within finance theory itself, misguiding regulation and stoking huge financial imbalances.

Sparking a revolution in the mindset of the investment professional, Osband recasts the market as a learning machine rather than a knowledge machine. The market continually errs, corrects itself, and makes new errors. Respecting that process, without idolizing it, will promote wiser investment, trading, and regulation. With uncertainty embedded at its core, Osband's rational approach points to a finance theory worthy of twenty-first-century investing.

1

Introduction

This is a book about the most important risk we face in finance. It’s the risk that comes from learning about risk. I call it Pandora’s risk in honor of legend’s prime culprit. If she hadn’t opened the box of wealth and woe, we’d have no hunger to learn.

Other fields involve learning too. Since the observer never fully understands the observed, there’s always something to learn. Occasionally, learning overturns some core beliefs. That’s how scientific revolutions occur.

Finance stands out in that the core objects of study are themselves observers. Market participants rarely know the true value of what they’re trading. Markets grope for knowledge by aggregating individual beliefs. But beliefs are constantly shifting.

Twentieth-century finance theory treated learning as a sideshow. It assumed that the market consensus largely captured the true risks. Error was dirt around the edges. If the dirtballs got big enough, speculators should arbitrage the discrepancies and clean things up.

In reasoning this way, theorists missed something that every market practitioner knows. Most speculators don’t trade on changes in risk. They trade on changes in beliefs about risk. Those aren’t the same. Sometimes they’re not even close.

This confusion pervades financial risk analysis. For example, the global banking regulations known as Basel sanctified nearly unlimited leverage for loans to top-rated credits. They brushed off as minor detail the difficulty in rating safety. The disregard helped stoke the greatest debt bubbles in world history.

Most financial analysts incline to downplay Pandora’s risk. They want to impress their superiors with what they know. Their superiors in turn want to persuade investors and regulators that risks are under control. Few want to advertise their uncertainty.

Results betray them. Risk bounds have to be continually reset, even when the nominal investments don’t change. Despite these adjustments, losses breach extremes far more frequently than standard models suggest. When stricken, most standard setters seek comfort in the crowd.

It doesn’t have to be that way. Uncertainty is our friend as well as our enemy. It encourages us to agree with each other, or act as if we agree, even when there’s no objective basis. The consensus encourages real trade and investment, which turns the fictions of belief into material facts.

Better appreciation of uncertainty can also help us think outside the box. As Mark Twain is alleged to have said, “It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.” Every financial crisis brings reminders.

This book is dedicated to changing mindset. I want professional analysts to realize that uncertainty is core to finance, not peripheral. The market isn’t some knowledge machine grinding out approximately ideal prices. It is a learning machine that continually errs, corrects itself, and makes new errors.

The scale of error makes us moan. Societies rend and nations lose their way over mania that later looks idiotic. Long-term inevitabilities get obscured by noise.

The scale of error correction makes us marvel. Capital markets forge consensus vision out of the dreams of millions of people, temper it in the fires of observation, and harness resources and will toward realizing it in practice. No one who has not witnessed huge capital markets would believe they exist.

Never has it been more important to keep both tendencies in mind. The world as a whole is the richest it’s ever been, and growth is nearly the fastest. Yet debt imbalances are outpacing gross domestic product (GDP) and appear unsustainable. Financial analysts who can’t draw the connections endanger those they serve.

On the practical side, this book offers some approximations to help track uncertainty. Mostly they’re just stylizations of what traders already do. They aren’t perfect. They can’t be. The best model would require an infinite number of calculations every instant.

There is a broader lesson here. Every practical application makes do with error-correcting approximations. Sometimes these approximations work terribly and unleash a plague of market demons. Eventually the error correctors kick in and help rectify the mistakes.

From a learning perspective, both apostles and foes of free capital markets should curb their enthusiasm. On the one hand, markets deal in consensus beliefs rather than truth. Following like lemmings can lead us off the cliff. On the other hand, markets are awesome error-correcting mechanisms. Stifling what regulators don’t want to see or hear often transforms downturn into disaster.

Confusion about the strengths and weaknesses of markets has fomented some rotten regulation. It lets some big excesses go unchecked, while exaggerating the importance of minor signals. It encourages statistical fraud. From a learning perspective, we need to simplify the regulatory framework and encourage more fiduciary responsibility.

Other overheated disputes pit orthodox finance against behavioral finance. Orthodox finance tells us how markets should operate, given well informed participants acting in their own rational self-interest. Behavioral finance emphasizes the abundance of counterexamples and links them to human irrationality. Treating the market as a rational learning machine can help transcend the divide.

Last but not least, a learning perspective can improve portfolio management services for retail investors. Standard approaches expose them to far too much risk in crises. As we shall see, this is remarkably easy to mitigate. While it won’t completely level the playing field with wealthy investors, it will reduce the “Dooh Nibor” (reverse Robin Hood) effects that prevail now.

The Context of Finance

To appreciate the importance of learning in finance, let’s go back to basics. Life forces a trade-off between risk and reward. By venturing out to eat, I expose myself to being eaten. Brains evolved in large part to raise the munch-to-munched ratio. Risk analysis is what they do. Financial risk analysis is simply a special case, focused on investments.

Broadly speaking, financial risk analysis makes three kinds of assessments. The first kind rates the pleasure or pain of the possible outcomes. The second kind forecasts the relative likelihood of these outcomes. The third kind estimates the uncertainty fogging our projections.

In practice it is hard to say where valuation ends and risk or uncertainty estimation begins. Imagine, for example, a strawberry. What is eating it worth? On reflection, that depends on ripeness, flavor, and other qualities not evident on sight. Perhaps the strawberry is harboring a stomach bug. Or perhaps a stomach bug one already has will make the strawberry unpalatable. That happened once to my youngster, leaving a red stain on the carpet as memento.

We could, if we wanted, distinguish a host of specific strawberry-eating experiences, assign likelihoods to each, and gauge our uncertainty. Both brain and gut would soon tire of the effort. To simplify, we bundle choices and outcomes, define a bigger action like “buying a strawberry and eating it,” and compare that action to other actions in terms of rewards, risks, and uncertainty.

For more complex risk analysis, consider the joint action of planting a strawberry field, tilling it to harvest, and marketing the crop for profit. Here we have to guess both the strawberry yield and the future price per strawberry. Granted, we might lock in one or both with a futures contract, but that just transfers the risk analysis to someone else.

For still more complexity, consider the purchase of a strawberry farm. Now we must analyze a succession of strawberry fields, one year after another. The further we peer into the future, the more wobbly our estimates will be. We will not know what strawberry-growing innovation might appear, what pests might infest the area, or whether future customers will prefer blueberries. We will not know the future tax regime on strawberry fields or the option value of selling the farm to a real estate developer.

Next imagine a giant agribusiness. It operates many farms producing many different products, runs various processing plants, and engages in a host of trading operations. Diversification and vertical integration lower some product-specific risks but heighten exposure to macroeconomic risks.

Finance weighs risks against rewards in search of higher or more secure profit. When finance gets it right, it raises the average munch-to-munched ratio and encourages productive investment. Occasionally finance gets things spectacularly wrong and wreaks havoc. Sometimes, too, it cannot resist munching on the people it is supposed to serve.

These contrasts between treasure and trouble, between promise and peril, naturally fascinate observers. They inspire envy and dread, admiration and contempt. Finance is awesome and awful.

One of the strangest features of finance is its self-absorption. Its most earthshaking decisions can seem divorced from studies of the earth. Most financiers are most concerned about what other financiers will think. This adds another layer of risk, but a strange, self-referential one. To invoke Keynes, a financial market is like a beauty contest aiming to pick not so much the greatest beauty as the one the judges will pick.

This makes financial markets terribly prone to herding. They can chase after bad ideas or run too far too fast after good ones. Those who resist can get trampled in a stampede.

However, there is a positive side to this. To appreciate it, let’s transform our imaginary herd of cattle into a hive of bees. The bees go out looking for nectar. Not knowing where it is, each starts looking on its own. When one bee finds it, other bees find out. Perhaps they hear directly. Perhaps they notice telltale pollen on the discoverer’s feet. Perhaps they just follow the buzz. Soon a swarm forms. The swarm becomes a better indicator of pollen than any individual bee. That is the wisdom of the hive.

As best we can tell, the wisdom of the hive generally outweighs the horrors of the herd. Economies with financial markets tend to be much more productive and innovative than those without. In the twentieth century, Communist states tried to prove the superiority of market-light planning. They failed miserably. The Soviet bloc collapsed. Only China thrived economically, and only after it reversed course and let markets guide planning.

Note that this is an empirical inference about financial markets, not a theoretical one. Note too that our confidence falls far below the standards of proof in natural science. We have observed a few dozen market-phobic industrializing societies, not a few trillion. We haven’t rerun Soviet experience substituting different planners.

But that is the norm for financial risk analysis. Our inferences come from relatively few experiments, most of them poorly controlled. Consequently our conclusions are more tentative than in natural science and prone to more frequent and bigger readjustments.

Learning About Learning

Market learning is now a mainstream topic in economics. Pastor and Veronesi (2009) have recently surveyed the developments. I too have tried to popularize these themes in the narrower circle of finance practitioners interested in theory (Osband 2002–2005, 2008).

However, financial risk analysis still treats learning risks as peripheral. It is stuck in a mindset akin to that of physics in the early twentieth century, which clung to classical mechanics despite the accumulating refutations. Arguably it’s stuck worse. Whereas physicists never blamed dumb matter for messing up theory, finance theorists frequently blame dumb traders.

How can we best change mindset? I wish I knew. Rigorous treatments get dismissed as dry or incomprehensible. Popular treatments leave professionals unconvinced. This book will try something in between. It will lead a study tour.

Our tour will visit some of the biggest financial risks in the world and explore major uncertainties at their core. We will watch markets unite, stretch, and defy beliefs. We will witness the damage that comes from ignoring uncertainty. We will look for neat ways to rebuild.

As the tour guide, it’s my responsibility to keep things lively and not too long. I will mix topics and approaches that don’t normally get mixed and draw analogies that don’t normally get drawn. I will ignore most caveats and keep intimidating terminology to a minimum. Where models scream out with policy advice, I will let them.

To those prizing neat results, I offer a few delights. We’ll expose the statistical confusion implicit in conventional risk measures and discover superior alternatives. We’ll decouple safety from certainty. We’ll link market turbulence to learning.

In return, I’m going to ask a lot from participants. Readers ought to have a sound grounding in economic history, finance theory, and statistics. They should be interested in economic policy. They should enjoy mathematical modeling. They should love thinking outside the box.

If you’re rusty in these areas, that’s okay. I’ll try to keep the exposition clear and provide references for deeper study. I will route most of the math to the Appendix so that it doesn’t overwhelm the flow.

One thing I won’t ask is complete agreement. Some of what I’m saying is surely wrong; I just don’t know which some. For fairness, I will let two august critics weigh in at the end of every chapter.

If you’re not sure it’s worth investing the time, skip to the Conclusions in Chapter 12. If they all make perfect sense, you don’t need to read the book. If none makes any sense, you won’t want to read the book. I’m aiming for the persuadable middle.

Itinerary

Money (Chapter 2): Every day people sell real goods and services for ciphers simply because everyone else does. It’s never certain that money will keep its magic, and sometimes it doesn’t. Yet modern civilization is unimaginable without it. We’ll uncover Mahserg’s Law and watch it at work.

Wealth (Chapter 3): Financial wealth discounts future earnings in ways that have long puzzled theorists. We will see that the most plausible explanations invoke uncertainty. Forced to look backward into the future, people discount heavily for tiny fears of disasters and prize perceived safety.

Debt (Chapter 4): Debt trades current money for future money with interest. If wealth grows fast enough, debt can potentially be repaid indefinitely out of rollover. This tempts self-financing bubbles of debt. We’ll examine a model in which patently worthless debt stays low-risk for long periods before blowing up. The model is too close to reality for comfort.

Banking (Chapter 5): Banks tend to be more interested in borrowing short to lend long than in facilitating payments. This makes the financial system more fragile and cycle prone. Regulators have inadvertently exacerbated the risks. To better appreciate the dynamics, we’ll model credit spreads and debt stocks as a predator-prey game.

Safety (Chapter 6): Many credit markets behave as if they’re inferring from only a few dozen years of relevant observations. We can best model their beliefs as highly dispersed distributions. The uncertainty calls for much larger contingent reserves on top-rated credits than standard regulations assign.

Regime Change (Chapter 7): Risks often change with little notice,...

Table of contents

• Cover 
• Half title
• Title
• Copyright
• Dedication
• Contents 
• Preface
• Acknowledgments
• Abbreviations
• 1: Introduction
• 2: The Ultimate Confidence Game
• 3: Great Expectations
• 4: Sustainable Debt
• 5: The Midas Touch
• 6: Safety in Numbers
• 7: When God Changes Dice
• 8: Credit-ability
• 9: Insecuritization
• 10: Risks in Value-at-Risk
• 11: Resizing Risks
• 12: Conclusions
• Appendix
• References
• Index