Book review: Antifragile

Antifragile by Nassim Taleb (Allen Lane 2012)

I first met Nassim Taleb in New York in the spring of 1998 while working for a trade magazine. I had been commissioned to edit a supplement on the 25th anniversary of the publication of the Black-Scholes option pricing formula, while Taleb at the time was a currency option trader at the French bank Paribas which was sponsoring the supplement. Over brunch at a midtown deli, Taleb spoke of his admiration for Robert Merton, and later delivered an article in which he paid tribute to his work.

Following the near-collapse of Long-Term Capital Management, the hedge fund involving Merton and fellow options guru Myron Scholes later that year, I went on to write Inventing Money. Shortly afterwards, Taleb published own first popular book, Fooled by Randomness, and thus began his transformation into the Merton-hating, economist-baiting, bodybuilding, fiery intellectual he is today.

In his latest book, Antifragile, Taleb, goes further than any of his previous works in setting out a ‘theory of everything’. Alongside the book itself, laced as most of Taleb’s writing with personal anecdotes and literary-mytho-historical musings, are technical papers aimed at backing up his claims to have discovered what he calls a ‘philosophers’ stone’.

Antifragile takes the idea of optionality from finance and broadens it out into a universal quantity. From engineering to medicine, to career development to politics, Taleb sees optionality everywhere. How does he do this? In finance, option contracts give nonlinear exposure to changes in some underlying quantity, or what finance experts call convexity.

Now applying the idea of optionality to everyday life isn’t new. Berkeley professor Mark Rubinstein (one of Taleb’s bête noires) told me 14 years ago how he saw personal decisions like marriage as forms of option. Taleb extends Rubinstein’s analogy by considering how natural entities or manmade systems are exposed to random fluctuations, uncertain external forces or errors, in the same way that financial derivatives are exposed to swings in markets.

The health or resilience of such entities or systems, Taleb argues, must be a nonlinear function of external influences because the tails or extreme events have a proportionately much greater impact than day-to-day fluctuations. In other words, outside finance, without having traded contracts that explicitly provide nonlinear exposures, you have options whether you want them or not.

Medicine is one area considered by Taleb in this way, where the impact of medication is uncertain. For a healthy person, medication can’t make them very much healthier than they already are, but there is a possibility it will make them much unhealthier – so a healthy person is like a trader who is short an option with regards to drugs. On the other hand, a seriously ill person has a lot more upside from uncertainty in medication than downside, so they are long an option.

The word ‘optionality’ is too narrow for his purposes, so Taleb coins the term ‘fragility’ for entities or systems that are short options, and ‘antifragility’ for those that are long. Taleb’s point is that when uncertainty increases, culminating in extreme events, ‘fragile’ things die or fail, while ‘antifragile’ ones survive and thrive, like option traders who own out-the-money contracts. He argues that by optimising systems to survive under normal conditions, we’re setting ourselves up for catastrophic failures like the Fukushima nuclear reactor.

I use the words in quote marks because one has to question whether these universal qualities actually exist as Taleb claims. A key issue is Taleb’s objectivist approach to uncertainty. In the technical document co-authored with Raphael Douady, Taleb writes: ‘a coffee cup on a table suffers more from large deviations’. Reading that, I ask, how does a coffee cup suffer? Does it have feelings?

For me, the sensible approach to such questions is to ask the person responsible for the coffee cup how they feel about its fate. Compared with the relative certainty of leaving the cup in the cupboard, how do you feel about the negative consequences of the cup tumbling off the table versus the limited upside of the cup doing its job and successfully conveying coffee to your lips without incident? By stating your relative happiness or unhappiness about these outcomes, you can actually derive your subjective probability that something will happen to the coffee cup.

Now Taleb rules out such approaches from the start, declaring that ‘psychological notions such as subjective preferences…cannot apply to a coffee cup’. That leaves us with Taleb’s universal, suffering coffee cup that I personally care nothing about. You won’t find any discussion of this distinction in Antifragile, which is why if you aren’t a fan of Taleb’s writing style, it’s best to skip the book and download his technical documents for free.

This is more than mere philosophical hair-splitting. The point about subjective preferences is that they lead to decisions, which is the stuff that actually happens, objectively speaking. The modelling of random events is a secondary activity – an intellectual construct that people use to communicate and validate their fundamental, subjective decisions.

That’s why in The Devil’s Derivatives I considered the financial crisis by analysing the subjective preferences of two tribes of players – ‘hate to lose’ versus ‘love to win’ – whose asymmetry with regards to upside or downside amid uncertainty determined the types of investments or trades they were prepared to accept. The risk modelling, the quant models, the credit ratings came after the fact, providing validation that helped accelerate and propagate the consequences of these preferences to systemic proportions.

Take the mortgage agency Fannie Mae, which Taleb uses as an example of where experts gave a misleading risk estimate for what turned out to be a fragile institution. I would argue that Taleb’s example, while embarrassing for those experts, is secondary to the main issue, which is how constituencies in US Congress (that had ‘skin in the game’) expressed their subjective preferences for government intervention in the housing market. The convenient fiction that Fannie was private-sector while enjoying quasi-Treasury funding costs, the tacked-on requirement to promote minority lending; these preferences were simply inconsistent with one another. The false judgement by experts-for-hire that Fannie was risk-free was no more than ex-post validation of a political fantasy.

Of course the models were wrong – or susceptible to Taleb’s famous ‘black swans’. They were set up that way. That’s why Taleb’s demand that experts have ‘skin in the game’ is so misguided and dangerous. Having skin in the game is precisely the mechanism that results in psychological preferences turning into bets, which is why experts need to be independent in order to question the subjective probabilities implied by these bets.

Where does this leave the ‘philosophers stone’? By ignoring the psychological drivers of decisions and risk, Taleb is forced to consider human agents interacting with one another as if they were universal, suffering coffee cups. You might identify embedded options, of convexity lurking in these relationships. The trouble is that for a great many situations, the direction of this optionality isn’t known in advance. And while in principle, all options become more valuable when uncertainty increases, the directionality of options also becomes more unstable as the psychological preferences of players change or reverse themselves.

In the financial crisis, quirks in marking-to-market exposed these issues in a number of places. A bank buys a credit default swap hedge on a counterparty, and when the value of the payoff increased, that has the effect of making the counterparty more distressed, and which worsened investor confidence in the bank that bought the hedge in the first place. A distressed firm whose options went further into the money finds it harder to obtain collateral from counterparties who began assessing the value of the option embedded in ISDA contracts to only close out contracts in their favour.

The Eurozone crisis throws up similar ambiguous options – such as Greece’s option to misstate its accounts, Germany’s option to impose fiscal austerity on others, and the option of banks to enter carry trades on PIIGS debt using ECB repo facilities. Do these options always become more valuable in the same direction as uncertainty increases? Outside of finance in everyday life, even more ambiguities crop up.

When confronted with psychological preferences, the central thesis of Antifragile simply unravels, like a woolly jumper snagged on a barbed wire fence. It leaves Taleb’s book looking, what’s the word?…fragile.