The Black Swan: The Impact of the Highly Improbable

Citation

Hannabuss, S. (2008), "The Black Swan: The Impact of the Highly Improbable", Library Review, Vol. 57 No. 2, pp. 152-155. https://doi.org/10.1108/00242530810854080

From time to time, and with revealing regularity, books on statistics and risk appear that capture the interest both of the general reader and of the professional manager. This is one. Knowledge of formal statistical decision‐making is helpful here but not essential as a prerequisite for having such an interest: we are all intrigued by risk and uncertainty, we are all (vaguely and in some cases learnedly) aware that events do not always conform with standard distributions and bell‐shaped curves, we all try to explain things in plausibly narrative ways that may well be wrong, and, if pushed, many of us are sceptical of experts.

With this background, then, The Black Swan is a book that has attracted a lot of current interest and, as such, merits review here. Taleb is a US writer known for his contrarian views about forecasting financial markets, and his earlier book, Fooled by Randomness (2004), attracted and raised debate about risk and uncertainty in much the same way as, earlier, Chris Anderson's The Long Tail (2006) did (this queried standard or Gaussian distributions of events like book sales on the internet). Taleb's approach is to ask, first, whether events – financial and all others – really do conform to such patterns and expectations, and to argue that they do not. His critique is strongest on economists, financial experts, historians (with their narrative fallacy interpretations of past and present events), and, of course, statisticians themselves. For anyone, in information‐related work or not, the arguments are stimulating, infuriating, meandering, challenging, and – at all times – fascinatingly sourced.

The mind‐set of current thinking about risk and uncertainly, Taleb suggests, has been over‐shaped by reliance on traditional ideas about events, distributions and deviance. Not even Zipfian ideas went far enough. The metaphor, taken from Umberto Eco's notion of the library and anti‐library, is that we rely too much on the books we've read, the ideas we think are true (epistemically), rather than what is really much more random. This tendency to shape explanations of events so that they have cognitive coherence is what he calls Platonicity (after Plato's forms) and an epistemic arrogance (a subjectivity which Taleb might, with advantage, have contrasted with arguments about alethiology but does not). Many events, as a result, in financial markets and in organizational decision‐making, occur which do not conform to Gaussian assumptions (Gauss, you will recall, got us thinking about bell‐shaped curves) and so, when unexpected events occur (like market crashes, 9‐11, the sales of the Harry Potter books, the dot‐com bubble, Hurricane Katrina), we are not ready for them, we react to them in panic, and then afterwards expect them to recur in very much the same way. Hence, measures against international terrorism and against environmental disasters.

Such events are “black swans”. The phrase will immediately echo the idea of black and white swans in the quite different context of falsifiability, in the work of Popper, though Taleb is keen to separate his own arguments from those. He is indeed keen also to separate his arguments from the conventional ones about probability in the casino sense, the sense that many of us have about randomness (cards, dice, and the like). For readers further into this field, and because his arguments go that way, Taleb suggests that a more flexible view of randomness – that developed by Mandelbrot and based on fractals (which preserve across scales) – is necessary if we are to provide any credible explanation of black swans. His frame of reference is wide and well‐informed, and some of his reading has been indicated in the references to this review so that readers can follow it up: this is something that is certainly well worth doing.

For many readers of this book, some of the tables summarizing the main ideas (as well as the useful glossary) will be particularly helpful. Early on, for instance, Taleb tabulates the differences between what he calls the two different ways of thinking about and explaining random events – Mediocristan (where, because things are assumed to conform to the bell‐shaped curve, only misleading explanations of risk can be used) and Extremistan (where we accept that there is no typical event and so accept the possibility, even the likelihood, of black swans). Later on, again, he summarizes the two distinct approaches to randomness (on page 284).

So how convincing and relevant is all this? On one level, that of financial forecasting and economic interpretation and statistical analysis, Taleb's arguments distill and popularize familiar ideas, many drawn from research papers (which are cited), and for the expert reader will offer an entertaining but ultimately merely populist reminder of how we should re‐examine our assumptions about randomness, above all in decision‐making (a literature in itself rather too full of dogmatic over‐rationalizations and theorizing). On another level, Taleb has written an intellectual firework of a book, offering, as many such books do and have over recent years, a series of ideas that will impress your friends and dazzle your enemies. On that level, too, it is a book that appeals to the popular imagination, the airport market, the news junkie, and to anyone fascinated by the unexpected yet sceptical about the role of experts.

At yet another level, the thoughtful reader, the experienced practitioner who has come to realise how beguilingly plausible are explanations (organizational and person) of events that happen and how fragile are forecasts of how things might happen (trends, correlationships, chi‐square and t tests for significance, time series, Delphi projections) will read the book and find a lot of common sense there. The extent to which (1) such a reader will and would carry Taleb's ideas forward into applying Mandelbrotian approaches and (2) such a reader will and would find the ideas of randomness and risk technically relevant to actual situations at work remain to be seen.

Taleb does not go there in a literal sense, even though he quotes numerous papers that do. In addition to all this, it is an eccentric and idiosyncratic book with personal asides (his own background and education, for instance, comments on autism, demolitions of fellow scholars, for he is a contrarian), and this can be both a delight and a frustration. It comes as no surprise that Allen Lane published the book in paperback very soon after the hardback, and it will be mainly in paperback format that it will sell in large numbers. Yet the appetite for books that seem or even claim to unpack some of the mysteries of the intellectual universe is both voracious and ephemeral, and I can foresee how this book will perform like a black swan itself, gaining unexpected popularity for a while and then vanishing.

Marketers know this from their bestseller curves, so it is not as unexpected perhaps as Taleb suggests. One message he leaves that, I think, remains constant and credible: that, unless your way of explaining risk and uncertainly is able to deal with the unscalability of black swans, it is not going to be enough. The literature on risk and disaster management at its best is coming to see how pragmatic planning might be far more useful than statistical extrapolations, and, for many information specialists reading this book, other probably that those with an intellectual interest in matters Zipfian (though even those will want to apply black swan thinking to the Darwinian evolution of search engines), the book will be a lively read and one to recommend more to the statistics and finance students who use their library rather than to apply it to their own daily practice.