Black swans and dark matter

There’s been a lot of talk about the idea that the GFC (the in-group shorthand for ‘global financial crisis’) is an example of a ‘black swan’, that is, an event that would be treated as impossible on the basis of induction from past experience, and hence that could not be encompassed by formal models of the kind used by risk managers. All this talk has of course been great for Nicholas Taleb who has a book with this title. It’s good in a lot of ways, but I found it ultimately insufferable in the continuous repetition of the message that only Taleb was smart enough to see all this. ( To be fair, Taleb predicted a global financial crisis, and didn’t simply claim it in retrospect as an unpredictable Black Swan).

I spend a lot of my time working on how to think about unforeseen contingencies and I’m not at all convinced that the GFC should be described in this way. Of course, the models used by the risk managers in investment banks didn’t include this as a possibility; if they had, the implication would have been that all sorts of much-desired deals should not go ahead. But as I pointed out a while ago, very simple models based on well-established principles predicted that the bubble economy would end badly.

The crisis then, involved something more like dark matter, the ‘missing’ matter in the universe that must exist if it is to work as it does, but can’t at presented be detected. Given that risk can’t easily be made to disappear*, it was obvious that the risk associated with lending of all kinds (most obviously, mortages offered to people with no capacity to repay) was being borne by someone, and probably someone who was unaware of it.

The big problem for the Cassandras (and we were certainly both correct and disregarded) was that it was easy to see that the bubble could not continue and much harder to foresee how it would end – it’s one thing to say that dark matter must exist and another to work out what it is really like. Like Brad and Brad, I expected that the problems would emerge first in the form of a run on the US dollar, given that holders of US dollar assets were receiving very little compensation for the obvious risk of large capital losses. In fact, the US dollar actually rose in the early stages of the meltdown, though it has been falling more recently.

One ‘black swan’ explanation of the mortgage crisis was that the mortgage derivatives created by Wall Street couldn’t fail except in the event of a simultaneous downturn in all major housing markets in the US, something that had never been observed, and therefore could not be included in the models. But of course the reason such a thing had never occurred was that local housing markets had been separate from each other, with their own sets of banks, S&Ls and other financial institutions. The very banks that were doing the modelling were creating the conditions under which a national bubble and bust could take place. This was both foreseeable and foreseen.

*Unlike matter, risk can be created with ease. Reducing it is more difficult, though pooling and sharing of uncorrelated risks works in many cases.

56 thoughts on “Black swans and dark matter

  1. radish,
    I don’t think so.

    “Although that’s obviously what agents of the state would have you believe.”

    If you are inferring to illegal or black market activities, I doubt that constitutes the majority of economic activity within an established society.

    “But the whole point is that it can’t be necessary, because people are constantly using money that wasn’t issued by a state they personally trust. ”

    Which money is that, which “people are constantly using that wasn’t issued by a state they personally trust”? Please elucidate your position.

    “Anonymous reciprocity is like that.”

    Ah…, but it is anonymous reciprocity within the context of the legal framework established by the state. Further, I don’t really see the connection between market arbitrage and anonymous reciprocity, please elucidate.

    In times where a society morphs into a dysfunctional state or completely collapses, the economic agents revert to a barter system or to a commodity that exhibits the fungible properties of money, for instance gold.

    radish, I actually welcome this discussion :-), although I do wonder, whether we in science often attempt to explain observations in one area of human endeavor with observations in an other that might not be all that relevant. That is not to say, one should not explore such possibilities, but rather be aware of the limitations of such a pursuit. I think the following passage by Taleb might be really worth some consideration– “I urge all you scientists to go take your “science” where it may work—and leave us in the real world without more problems. Please, please, enough of this “science”. We have enough problems without you.”

  2. Well I guess since our host hasn’t yet kicked us out for abusing his comment section…

    Which money is that, which “people are constantly using that wasn’t issued by a state they personally trust”?

    How about US dollars. My position is that an agent’s confidence in the negotiability of dollars does not necessarily depend on the US enforcing legal tender laws, nor did it depend on the agent associating any material utility with the commodity which backed dollars back when dollars were backed by metals. Rather, it has, in general, mostly to do with the agent’s observations about how many other agents are willing to accept dollars today.

    Other than mentioning that I personally don’t trust the US to have a sane monetary policy, I’m not sure how to support this position without doing some rather labor intensive research, the marginal utility of which would be low for me personally. And since nobody has offered to pay me for it, I hereby decline to do it.

    Further, I don’t really see the connection between market arbitrage and anonymous reciprocity, please elucidate.

    It’s not that anonymous reciprocity creates opportunities for arbitrage, it’s that anonymous reciprocity is an unexpected behavior, just as failure to exploit arbitrage is an unexpected behavior. Why that particular behavior is unexpected is a bit complicated, so again, apologies for length.

    Money is a form of information, fundamentally a signaling medium. But agents who use money have a pronounced tendancy to treat these signals as reliable, sort of as though they had inherent material utility, or could only be sent at high cost. I believe this is one of the things that’s made economics into a somewhat schizophrenic discipline. 😉 Anyway, it’s puzzling, because it’s extremely rare elsewhere in nature, and occurs only under very specific circumstances.

    One way explain this is essentially what Ikonoclast suggested. That money is simply a special case of language. We can coordinate reciprocity because we can talk about the future. We can make promises. While this is true as far as it goes, it doesn’t explain why we believe the promises of strangers when they speak with money but not when they speak in other languages.

    More technically, it doesn’t explain how reciprocity signals can be homogenous and universal when identity signals are heterogenous and highly differentiated. For trivial cases, or if you just don’t look too closely, this might not seem like a deal-killer, because there are indeed some “universal” signals associated with reciprocity and trust among humans. Smiling, frex, can trigger forward reciprocity. What it can’t do get you to work umpty hours a week for years on end for someone whom you hate with a passion.

    Ants and bees are also creatures using intraspecific signals which are strong reciprocity triggers, but those are associated with very strong, very homogenous, identity signals. At the end of the day, forward reciprocity ordinarily requires a lot of very strong identity signaling to be in place, in humans, and in all the other creatures which engage in it. The exception being the use of money, which seems to fuction as an identity signal in and of itself. That’s the puzzle. If you really want a decent model this discrepancy has to be dealt with.

    So. Humans have acquired this strictly behavioral adaptation which enables them to engage in lots of forward reciprocity without the perceived risk of having it turn into altruism. They have, somehow, done away with the otherwise consistent requirement for a reliable identity signal, and replaced it with… what?

    You say trust in the state. I say that doesn’t explain the edge cases, and it wouldn’t be strong enough even if it did. The state can coerce people to cooperate with it. It can even discourage them from certain behaviors. What it cannot do is coerce them into cooperating with each other.

    I urge all you scientists to go take your “science” where it may work—and leave us in the real world without more problems.

    It’s funny you should mention that because I just saw that a few days ago, and found it very amusing to see a mathematician/philosopher and sometime quant guy rant about science being detached from the “real” world. I respect Taleb’s opinions a great deal in general, but I have to agree with the folks who say that he doesn’t seem particularly interested in anybody else’s opinions, or in phase changes that aren’t abrupt and dramatic. There are quite a lot of interesting things that Taleb doesn’t seem to like talking about.

    Doug Rushkoff’s comment however, is one that I can endorse wholeheartedly and without reservation, and which I bookmarked when I saw it.

  3. For the record, since posting here my long comment about my history with Taleb, I went back and looked at TBS and found that I had misrepresented his barbell strategy, which calls for more generalized “speculative bets” rather than just puts on major crashes. I have posted on econospeak a public apology to him, with a more extended discussion “under the fold.”

  4. radish,

    “Well I guess since our host hasn’t yet kicked us out for abusing his comment section…”

    This will be my final post on this matter. I will respectfully agree to disagree with you on some of your analysis. 😛

    “How about US dollars. My position is that an agent’s confidence in the negotiability of dollars does not necessarily depend on the US enforcing legal tender laws, nor did it depend on the agent associating any material utility with the commodity which backed dollars back when dollars were backed by metals.”

    The US dollar is a fiat currency. It obtains it’s unique perceived value from the US govt. decree, declaring it as the legal tender of the state of the US. Moreover, the state has the ability to pronounce this decree through it’s unique position and the power of authority bestowed on it by, either the majority of the citizenry of the state, or by force.

    The economic agents of the state have faith in the US govt. to ensure that the US dollar or US dollar denominated financial instruments are accepted as legal tender in commerce.

    The states function is to establish and maintain an environment that ensures the above, and that intrinsic properties of money [1] are established and maintained. If the state fails in it’s essential role to achieve this, then the economic agents will eventually cease to consider this as a form of money, but rather just a piece of printed art work, and nothing more. If the fiat money ceases to serve it’s role in effectively providing the essential characteristics of a efficient mechanism of exchange, then the economic agent will seek substitutes.

    This circumstance may precipitate for many reasons, a few being, if the economic agents loose faith in the US govt. due to it’s poor economic management, or the continued debasement of the currency by the state, or the loss of authority of the state over it’s citizens.

    In any case, the state is an essential actor in the establishment and maintenance of stability in all its forms. From this stability further dynamics manifest themselves that facilitate sophisticated economic pursuits.

    “You say trust in the state. I say that doesn’t explain the edge cases, and it wouldn’t be strong enough even if it did. The state can coerce people to cooperate with it. It can even discourage them from certain behaviors. What it cannot do is coerce them into cooperating with each other.”

    It doesn’t need to coerce them to cooperate. The economic agents will on their own recognise the benefits and accept to use this form of money as a efficient means to facilitate the activity of commerce, with the premise that the this form of money truly does serve that role.

    What I am suggesting is the immensely import property, called trust, and all that which precipitates from it, such as, confidence, integrity, faith et cetera. These properties precipitate further derivatives, such as future reciprocity. The framework which allows for these properties to establish themselves in a sophisticated society is through the essential role that the state plays in establishing a society that has stability and longevity. Without the state, it is difficult to perceive that such sophistication could ever be achieved.

    [1] – These intrinsic properties that any form of money is required to fulfill are:

    1. a store of value.
    2. a unit of account.
    3. a medium of exchange.

  5. According to Galbraith, Veblen and others there is an additional function of money
    “an enormous number of people want money for its own sake, for the self gratification and self assurance its successful pursuit and possession give them”

  6. When it comes to making predictions, there are a few points to keep in mind. One is that if people make enough predictions, it is almost certain that some of them will turn out to be correct. The other is that if one predicts doom frequently enough, it is likely that eventually one will be proved correct. I have spoken to people who have been confidently predicting the imminent collapse of the Aus housing market since the late 1990s. I guess they had to be correct sooner or later.

    Even dart-throwing monkeys are right a significant proportion of the time.

    Therefore, all that really counts is whether an individual has a relatively high average success rate of predicting events. Not whether an individual is right on one prediction.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s