Nassim Nicholas Taleb

A black swan is an outlier, an event that lies beyond the realm of normal expectations. Most people expect all swans to be white because that’s what their experience tells them; a black swan is by definition a surprise. Nevertheless, people tend to concoct explanations for them after the fact, which makes them appear more predictable, and less random, than they are. Our minds are designed to retain, for efficient storage, past information that fits into a compressed narrative. This distortion, called the hindsight bias, prevents us from adequately learning from the past.

From my perspective, Black swans occur when there are significant mismatches between the models people use to understand the world and the subsquent expectations that those models produce and observations. In other words, black swans are model errors – something that I’ve written (Peterson, Carpetner & Brock et al 2003) in the context of ecological management.

In Wired Magazine article Black Swans: expect the unexpected economic writer James Surowiecki talks to Nassim Nicholas Taleb.

From Wall Street to Washington, we’re constantly being told that the future can be forecast, that the world is knowable, and that risk can be measured and managed. Nassim Nicholas Taleb is having none of this. In his new book, The Black Swan, the finance guru and author of the surprise hit Fooled by Randomness argues that history is dominated not by the predictable but by the highly improbable — disruptive, unforeseeable events that Taleb calls Black Swans. The effects of wars, market crashes, and radical technological innovations are magnified precisely because they confound our expectations of the universe as an orderly place. In a world of Black Swans, the first step is understanding just how much we will never understand.—

Wired: If Black Swans are the crucial determining events in history, why do we think we can predict anything at all?
Taleb: After they happen, in retrospect, we think that Black Swans were predictable. We think that if we can explain why something happened in the past, we can explain what will happen in the future.

But with better models and more computational power, won’t we get better at predicting Black Swans?
We know from chaos theory that even if you had a perfect model of the world, you’d need infinite precision in order to predict future events. With sociopolitical or economic phenomena, we don’t have anything like that. And things are getting worse, not better, because the growing complexity of the world dwarfs any improvement in sophistication or computational power.

So what do we do? If we can’t forecast the really important things, how do we act?
You need to ask, “If the Black Swan hits me, will it help me or hurt me?” You cannot figure out the probability of a Black Swan hitting. But if you’re in a business that’s prone to negative Black Swans, like catastrophe insurance, I advise you not to take your forecasting seriously — and to think about getting into a different business. You don’t want to be a sucker. What you want are situations where you can have as much of the good uncertainty as possible, where nothing too bad can happen to you, and where you have what I call free options. All of technology, really, is about maximizing free options. It’s like venture capital: Most of the money you make is from things you weren’t looking for. But you find them only if you search.

Nassim Nicholas Taleb was also profiled by Malcolm Gladwell in the New Yorker in 2002 . There his work at his trading firm describes how his trading focuses on the fat tails of probability distributions:

Nassim Taleb and his team at Empirica are quants. But they reject the quant orthodoxy, because they don’t believe that things like the stock market behave in the way that physical phenomena like mortality statistics do. Physical events, whether death rates or poker games, are the predictable function of a limited and stable set of factors, and tend to follow what statisticians call a “normal distribution,” a bell curve. But do the ups and downs of the market follow a bell curve? The economist Eugene Fama once studied stock prices and pointed out that if they followed a normal distribution you’d expect a really big jump, what he specified as a movement five standard deviations from the mean, once every seven thousand years. In fact, jumps of that magnitude happen in the stock market every three or four years, because investors don’t behave with any kind of statistical orderliness. They change their mind. They do stupid things. They copy each other. They panic. Fama concluded that if you charted the ups and downs of the stock market the graph would have a “fat tail,”meaning that at the upper and lower ends of the distribution there would be many more outlying events than statisticians used to modelling the physical world would have imagined.

In the summer of 1997, Taleb predicted that hedge funds like Long Term Capital Management were headed for trouble, because they did not understand this notion of fat tails. Just a year later, L.T.C.M. sold an extraordinary number of options, because its computer models told it that the markets ought to be calming down. And what happened? The Russian government defaulted on its bonds; the markets went crazy; and in a matter of weeks L.T.C.M. was finished. Spitznagel, Taleb’s head trader, says that he recently heard one of the former top executives of L.T.C.M. give a lecture in which he defended the gamble that the fund had made. “What he said was, Look, when I drive home every night in the fall I see all these leaves scattered around the base of the trees,?” Spitznagel recounts. “There is a statistical distribution that governs the way they fall, and I can be pretty accurate in figuring out what that distribution is going to be. But one day I came home and the leaves were in little piles. Does that falsify my theory that there are statistical rules governing how leaves fall? No. It was a man-made event.” In other words, the Russians, by defaulting on their bonds, did something that they were not supposed to do, a once-in-a-lifetime, rule-breaking event. But this, to Taleb, is just the point: in the markets, unlike in the physical universe, the rules of the game can be changed. Central banks can decide to default on government-backed securities.

According to the article, a simplified version of Taleb’s trading strategy is based on expecting the unexpected – betting that the markets underestimates extreme events – and being resilience to catastrophe. Gladwell’s article continues:

Taleb likes to quote David Hume: “No amount of observations of white swans can allow the inference that all swans are white, but the observation of a single black swan is sufficient to refute that conclusion.” Because L.T.C.M. had never seen a black swan in Russia, it thought no Russian black swans existed. Taleb, by contrast, has constructed a trading philosophy predicated entirely on the existence of black swans. on the possibility of some random, unexpected event sweeping the markets. He never sells options, then. He only buys them. He’s never the one who can lose a great deal of money if G.M. stock suddenly plunges. Nor does he ever bet on the market moving in one direction or another. That would require Taleb to assume that he understands the market, and he doesn’t. He hasn’t Warren Buffett’s confidence. So he buys options on both sides, on the possibility of the market moving both up and down. And he doesn’t bet on minor fluctuations in the market. Why bother? If everyone else is vastly underestimating the possibility of rare events, then an option on G.M. at, say, forty dollars is going to be undervalued. So Taleb buys out-of-the-money options by the truckload.

Related posts:

1. Gelman’s notes on Black Swans
2. Steve Carpenter on Black Swans
3. Tom Peters, Black Swans, and Resilience

Disproportionality is fairly general in many forms of environmental impact. For example, in the Lake Mendota watershed in Wisconsin, most of the phosphorus pollution comes from only a few of the farms in the watershed. However, ecological disproportionality is complicated by the fact that the vulnerability of different sites to human impact also varies – which compounds the disproportionality.

In a recent paper Disproportionality as a Framework for Linking Social and Biophysical Systems (Society and Natural Resources 2006 19:153-173) Pete Nowak, Sarah Bowen, and Perry Cabot write

Early social science was influenced by the work of Adolphe Quetelet, who promoted the idea that the average in a normal distribution represented the ‘‘essence’’ of a social system whereas variance or outliers were viewed as ‘‘accidents’’ in the study of social processes (Kruger et al. 1990). Charles Darwin, on the other hand, viewed variance, or the outlier, as central to understanding evolutionary biological processes. In this article, we have argued that giving more attention to variance across multiple scales can serve as a conceptual bridge between the social and biophysical sciences. Disproportionality is a concept that can bridge disciplines by focusing on the salient interactions between humans and their environments at different spatial and temporal scales.

Nowak et al use the example of farming practices and phosphorus runoff to explore how disproportionality in social and ecological systems intersects. They use the figure below to illustrate how the impact of a behaviour is shaped by place and timing.

Figure (from Nowak’s paper) The combination of typical conservation behavior, exhbiting a skewed normal or log-normal probability distribution, and typical environmental conditions, also exhbiting a skewed normal or log-normal probability distribution of the probability of environmental risk, combines to produce a situation in which a small proportion of inappropriate social behaviors within a particularly vulnerable setting can have a disproportionately large impact on overall environmental quality of an ecological system.

Nowak et al write:

Our conception of disproportionality emphasizes that human contributions to environmental degradation are not normally or randomly distributed, but arrayed in a way that may be strongly skewed and determined by the specific biophysical setting where it takes place. This view of disproportionality is a derivative of Robinson’s (1950) ecological fallacy in that the environmental performance of a unit of social organization does not imply that all individuals within that organization perform in a similar fashion. Because logging, mining, agriculture, or suburban development degrades the environment, for instance, this does not mean that all loggers, miners, farmers, or developers are equal contributors to degradation. The classic scientific caveat, ‘‘It depends,’’ conditions the impact of behavior based on the setting of that behavior. Furthermore, the log-normal world suggested by Limpert et al. (2001) implies that a very small minority within a social group may cause significantly greater environmental degradation, and possibly more than the combined impact of the remaining group. This disproportionate outcome occurs, not because the behavior of the minority is especially egregious or deviant, but because their actions are inappropriate behaviors taking place in biophysically vulnerable settings or times.

Employing this concept of disproportionality suggests that environmental sociologists need to account for both the distribution of behaviors within social organizational boundaries and, equally important, the distribution of the resiliency of the biophysical settings of those behaviors. Acknowledging that disproportionality may drive the improvements or degradation in an ecological system requires social scientists to explicitly account for within-group variation of social actors. Depending on the biophysical setting, only a few cases may have a disproportionate impact on the overall ecological system being studied. The potential for this form of disproportionality implies that the environmental meaning placed on a social behavior requires accounting for where and when it occurs in a biophysical setting. We emphasize that the social interpretation placed on any behavior by environmental social scientists, and any subsequent analysis of the ecological impacts of this behavioral pattern, must take into account where and when the behaviors occur in an ecological setting (Abbott 1999). The interaction between the characteristics of the behavior and the biophysical setting where the behavior occurs should determine the meaning placed on that behavior. For example, the all-too-common term of ‘‘bad actor’’ is only partially correct; both the ‘‘acting’’ and the ‘‘stage’’ for that action need to be used in forming such ill-advised value judgments. In our examination of the Pheasant Branch watershed, we found that the concept of disproportionality offered an approach to addressing the question of why significant changes in the social system did not induce parallel changes in the biophysical system.

Gladwell’s New Yorker article provides other clear examples of social disproportionality. He focuses on homelessness and how policies that try to grapple with its disproportionality face ethical issues:

… fifteen years ago, a young Boston College graduate student named Dennis Culhane lived in a shelter in Philadelphia for seven weeks as part of the research for his dissertation. A few months later he went back, and was surprised to discover that he couldn’t find any of the people he had recently spent so much time with. “It made me realize that most of these people were getting on with their own lives,” he said.

Culhane then put together a database—the first of its kind—to track who was coming in and out of the shelter system. What he discovered profoundly changed the way homelessness is understood. Homelessness doesn’t have a normal distribution, it turned out. It has a power-law distribution. “We found that eighty per cent of the homeless were in and out really quickly,” he said. “In Philadelphia, the most common length of time that someone is homeless is one day. And the second most common length is two days. And they never come back. Anyone who ever has to stay in a shelter involuntarily knows that all you think about is how to make sure you never come back.”

The next ten per cent were what Culhane calls episodic users. They would come for three weeks at a time, and return periodically, particularly in the winter. They were quite young, and they were often heavy drug users. It was the last ten per cent—the group at the farthest edge of the curve—that interested Culhane the most. They were the chronically homeless, who lived in the shelters, sometimes for years at a time. They were older. Many were mentally ill or physically disabled, and when we think about homelessness as a social problem—the people sleeping on the sidewalk, aggressively panhandling, lying drunk in doorways, huddled on subway grates and under bridges—it’s this group that we have in mind. In the early nineteen-nineties, Culhane’s database suggested that New York City had a quarter of a million people who were homeless at some point in the previous half decade —which was a surprisingly high number. But only about twenty-five hundred were chronically homeless.

It turns out, furthermore, that this group costs the health-care and social-services systems far more than anyone had ever anticipated. Culhane estimates that in New York at least sixty-two million dollars was being spent annually to shelter just those twenty-five hundred hard-core homeless. “It costs twenty-four thousand dollars a year for one of these shelter beds,” Culhane said. “We’re talking about a cot eighteen inches away from the next cot.” Boston Health Care for the Homeless Program, a leading service group for the homeless in Boston, recently tracked the medical expenses of a hundred and nineteen chronically homeless people. In the course of five years, thirty-three people died and seven more were sent to nursing homes, and the group still accounted for 18,834 emergency-room visits—at a minimum cost of a thousand dollars a visit.

…

The current philosophy of welfare holds that government assistance should be temporary and conditional, to avoid creating dependency. But someone who blows .49 on a Breathalyzer and has cirrhosis of the liver at the age of twenty-seven doesn’t respond to incentives and sanctions in the usual way. “The most complicated people to work with are those who have been homeless for so long that going back to the streets just isn’t scary to them,” Post said. “The summer comes along and they say, ‘I don’t need to follow your rules.’ ” Power-law homelessness policy has to do the opposite of normal-distribution social policy. It should create dependency: you want people who have been outside the system to come inside and rebuild their lives under the supervision of those ten caseworkers in the basement of the Y.M.C.A.

That is what is so perplexing about power-law homeless policy. From an economic perspective the approach makes perfect sense. But from a moral perspective it doesn’t seem fair. Thousands of people in the Denver area no doubt live day to day, work two or three jobs, and are eminently deserving of a helping hand—and no one offers them the key to a new apartment. Yet that’s just what the guy screaming obscenities and swigging Dr. Tich gets. When the welfare mom’s time on public assistance runs out, we cut her off. Yet when the homeless man trashes his apartment we give him another. Social benefits are supposed to have some kind of moral justification. We give them to widows and disabled veterans and poor mothers with small children. Giving the homeless guy passed out on the sidewalk an apartment has a different rationale. It’s simply about efficiency.

We also believe that the distribution of social benefits should not be arbitrary. We don’t give only to some poor mothers, or to a random handful of disabled veterans. We give to everyone who meets a formal criterion, and the moral credibility of government assistance derives, in part, from this universality. But the Denver homelessness program doesn’t help every chronically homeless person in Denver. There is a waiting list of six hundred for the supportive-housing program; it will be years before all those people get apartments, and some may never get one. There isn’t enough money to go around, and to try to help everyone a little bit—to observe the principle of universality—isn’t as cost-effective as helping a few people a lot. Being fair, in this case, means providing shelters and soup kitchens, and shelters and soup kitchens don’t solve the problem of homelessness. Our usual moral intuitions are little use, then, when it comes to a few hard cases. Power-law problems leave us with an unpleasant choice. We can be true to our principles or we can fix the problem. We cannot do both.

The entire article Millon Dollar Murray: Why problems like homelessness may be easier to solve than to manage is interesting.

Related posts:

1. Exploring Resilience in Social-Ecological Systems – E&S special feature
2. Expansion of social-ecological systems science
3. Resilience of social-ecological forest systems: post-doctoral position at Umeå