A nice introduction to the subject is a plot of the weight of a turkey against time which rises steadily upwards in a beautiful “predictable” line with an R-Squared of 99.99% (actually it should have been an “S” curve with a third-order polynomial but that’s detail), until suddenly it goes down to zero. Someone ate it.
Another example is a story of a casino in Las Vegas which spent millions on security to manage their risk, but then lost a fortune when a tiger from a show escaped and started eating the audience, and their insurance didn’t cover that.
Taleb’s thesis is that you can’t anticipate Black Swans, they are like the Internet, and who could have predicted the Internet? That’s why he says (in his book) that he doesn’t make predictions, although only a week ago he was “predicting” a cataclysmic decline in the value (rise in yields) of US Treasuries.
I shall have to remember that if ever I achieve a minor “guru” status to resist the temptation to open my big fat mouth about subjects that are outside my particular (and sadly quite limited) area of expertise.
Whilst Taleb’s ideas are new in one sense, they are really just a re-packaging of ideas that have been known about for years. At its core, insurance is about covering for a “tail risk”, and the theory there is that if you insure enough risks that are not inter-connected, although you may take a hit from time to time, in the end you come out ahead.
The trick there is not so much to work out the risk of one policy blowing up in your face, that risk can only be evaluated with a low degree of precision, because you are insuring against something that in many cases no one could have anticipated. A tidal-wave in Indonesia, a bush-fire in Australia, a failed “O”- ring on a Shuttle launch. Sure you assess the probability of an “event” but you know that is not exact, but you also work out what the damage will be “what-if”.
Where insurance goes badly wrong, when it goes wrong, is not its lack of ability to precisely anticipate highly unlikely events, it is that it did not anticipate correctly how much the policies would have to pay out if the highly unlikely “unimaginable” thing happened.
A good example is the CDS that were written by AIG. Those were priced based on historical data that considered the likelihood of any one AAA rated security out of thousands defaulting, assuming life went on as normal.
They were not priced based on an evaluation of the likelihood of the whole financial system imploding or the cost of that. That’s why they were sold so cheap, and that contributed to the problem because people told themselves “well if Moody’s rated this, and AIG insured it, they ought to know, so it’s SAFE”.
But in retrospect, financial systems have collapsed before, that is not something new. The “system” collapsed in 1929, LTCM collapsed, so did the Roman financial system 2,000 years ago, so did the Spanish system thanks to a combination of gold brought back from South America and the introduction of the joys of fractional lending.
Just as if the turkey had taken the time to look back at the longer-term history of turkeys, he would have realised that 99.99% R-Squared was not a great basis for predicting his immediate future, the various collapses of financial systems were “predictable”.
Legend has it that there were just two guys working out of the AIG offices in London who managed that almost unknown, and largely unseen element of the business; what they failed to work out was how much it would cost if their policies went wrong. Two people; and you can trace the whole daisy chain back to them.
Where I disagree with Taleb is that such events are not predictable. Certainly “WHEN” is impossible to predict, as is the probability of “something” happening within five, ten or twenty years.
But the amount of damage “IF” is not. A prudent investor should always bear the cost of the “highly unlikely” in mind, and if he knows he can’t cover that, he should insure against it, preferably with someone who knows what they are doing.
In that regard the mistake that the casino owner made was not to take an all-risks insurance without exclusions, the mistake that AIG did was not to go out into the marketplace and re-insure the risks that they had taken on with the CDS contracts they sold, if only because if they had, they might have found that other people saw that risk differently from them.
The fact that there was possibly a housing bubble in the USA in 2004 going into 2006 was openly discussed, the IMF talked about it, so did Professor Shiller, so did The Economist. People knew there was a risk – like a “tail-risk”, but very few people sat down and figured out what might happen in the “highly unlikely event” that there was a cataclysmic change of direction, and instead of going up and up like that line of the weight of the turkey, they would start to go down.
One of the reasons for the “overconfidence” was corruption, information was hidden and concealed on “off-balance-sheets” in the so-called “markets” for mortgaged backed securities and derivatives, on which pricing decisions were made. But those markets were not regulated, and they were manipulated so that investors were not provided with all of the necessary information to make prudent decisions.
The discovery of corruption is often the event that precipitates a reversal. A recent example is the “discovery” that the budget deficit of Greece for 2009 was not 3.5% of GDP it was 12.7% of GDP and it caused a reversal of the Euro and all sovereign debt.
That was not a slip of a pen, and now we learn that Greece has been consistently cooking its books under the noses of the EMU and the IMF. That “new” knowledge, and the suspicion that perhaps nothing is as it seems, caused a reversal…so ”perhaps there is more, perhaps Italy, Spain, Portugal, Ireland even the UK have been cooking their books?”
By the same token, the sudden reversal of Enron was predictable, but only if you had knowledge that they had been cooking their books; same thing for Made-off.
Case Study: Climate Change:
The United Nations Intergovernmental Committee on Climate Change (IPCC) produced a report in 2007 that concluded there was a 95% probability that sea levels will rise by a maximum of 59cm by 2100, and that the effects of Global Warming will cut 0.12% of GDP growth rates, (i.e. from perhaps 2.5% to 2.38% per year). There are other reports, for example the Stern Report prepared for the British Government, but let’s stick with IPCC:
Two points:
(1) A 59cm rise in sea levels over 100 years is not cataclysmic, and frankly, who cares about the Polar Bears? Sure I’d give $10 a year to “save the polar bears”, but I wouldn’t give $10,000…would you?
(2) The IPCC report and the level of scientific rigor that has been applied to the study of climate change is a subject of much debate; on one side the critics say that the scientists dishonestly exaggerated the dangers, and that politicians and the “green-wash” brigade have of late simply jumped on the bandwagon to pursue their own agendas.
On the other side the vast majority of the scientific establishment agree that there is a problem although they don’t know how big, and they agree that based on the information that is currently available, it is probably impossible to know with any clear certainty how big the problem is. A good debate about that from a “real” scientist can be found here.
That sounds a bit like where the debate about regulation of derivatives, inflation targeting, and house prices (or the danger of a cataclysmic crash in house prices) was in early 2006. A lot of talk and positions on one side of the table or another with anyone who strayed too far from the middle and “predicted” the end of the world (like “Dr. Doom”) being labelled as a bit of a “lunatic”.
BUT:
There is the tail-risk, which if the IPCC report is to be believed, (at least in general terms), has a probability of occurring within 100 years of 2.5% if the distribution of risk is “normal” (100% - 95% divided by two).
A tail risk is typically defined as events that are three standard deviations from the mean. They are very unlikely, but they happen.
For example, after a level of due diligence using data that was a lot more certain than the climate data that scientists analysed, a typical AAA sub-prime-mortgaged backed security was judged to have a risk of default of any kind of 0.26%. Essentially what happened, thanks (allegedly), to a combination of “greed, stupidity, incompetence, and fraud”, plus some other factors that may never be known for sure, a “tail risk” event…Err…happened.
The risk of Global Warming ending up outside the 95% “certainty” that was assigned by IPCC, assuming they got their numbers right, is thus about ten times more than the risk of a credit crunch.
If that happens, both the ice caps could melt and the sea level could rise 30 ft. That would be cataclysmic, and if that happened there is a good chance it would happen all of a sudden. Try putting a lump of ice in a slightly warm room. It just sits there, and sits there for hours and a little puddle forms then suddenly POOF, it’s gone.
Right now the assumption is that if we all do a bit here and do a bit there, and the USA starts to be “good” and China starts to be “good” and we all pull together and create plenty of pork for the politicians (like the program in Germany to subsidise solar panels (made in China) and pay eight times the normal tariff to people who sold that power to the grid), then everything will work out in the end.
That’s nice. Of course the tail-risk is that the ice caps do melt, in which case all the ports in the world will be flooded (won’t help world trade much), plus all of Shanghai, Guangdon, New Orleans, and the Netherlands, half of Tokyo, a good chunk of London and Washington DC, to name a few.
That would cost in direct damage and the present value of the loss of economic output easily $200 trillion.
But no one’s talking about that, well not seriously, not like it could happen. If that unlikely event is an option, basically to mitigate against that right now “the world” needs to start building 7,000 nuclear power stations and to ban the consumption of meat produced by ruminants (cattle produce 8% of the worlds greenhouse gases).
Will that happen? Highly unlikely.
Will getting the world to take 1% of its power requirements from solar energy (the ten-year target) make any difference? Highly, highly, highly unlikely.
Here’s a tip, if you live less than 30 ft above the current sea level, it might be worth checking if your insurance covers the possibility of the sea level rising by 30 feet, and also that in the event that happens the insurance company can cover your loss (making sure they are re-insured at Lloyds (LYG) is a good start).
Or move?
Other Possible Black Swans?
Two obvious potential candidates that come to mind:
1: Housing – and all who sail with it.
There was a decline in the USA, UK, and in many other countries, which was combated by energetic actions by governments to prevent price discovery of the mal-investments that were made in the past.
But the reality is that typically it takes about as long for housing to recover from a bubble as it took for the bubble to form. That took seven years, so if that model works, unless the governments drive themselves bankrupt “stemming the tide” (80% of all home mortgages written these days in the USA are essentially by the government), there could be another lurch down.
That’s the thing about Black Swans; just as you got “comfortable” they sneak around behind you and bite you you-know-where.
2: Euro-Denominated Sovereign debt
Greece has come clean and told everyone that yes it lied about its financial circumstances and it has been lying for years, but as for now it’s telling the truth.
Why don’t I believe them?
And if Greece could do it, so could Italy, Spain Portugal, and Ireland. A Black Swan can also be the phantom that lays all lies to rest.
3: Bombing Iran
But if the USA (or its ally Israel did it), the first thing that would happen is the price of oil would go to $200, which in the current circumstances would not be helpful for either the USA or China (if China is supposed to be pulling the world out of recession).
The logic this time is that Iran would pose a risk to World Peace if it renegged on its commitments under the Nuclear Non Proliferation Treaty.
There are three ironies there. The first is that India, Pakistan and Israel refused to sign that treaty (and have WMDs), and two of those states (Israel and Pakistan) are threats to Iran and are backed by the USA and NATO.
The last time a country got that level of support from the USA and NATO was when Iraq invaded Iran, so you can understand why they are nervous.
The second irony is that if the parties that are nominally involved in the dispute (Iran, Israel and the USA) were signatories to the International Criminal Court, there would be a jurisdiction for resolving the accusations that Iran had "broken International Law" by (allegedly) not complying with a treaty that neither Israel nor Pakistan have signed.
The third irony is the Geneva Convention, which theoretically (at least the 1977 version) would preclude a pre-emptive strike based on "intelligence" about what the other side "might" be thinking of doing. And we all know how reliable that sort of "intelligence" is.
But then nobody, least of all the USA, seems to care a toss about the Geneva Convention anymore (only recently the US Attorney general called it "quaint"), and in any case the jurisdiction for enforcement of the Convention lies with the ICC which the USA and Israel do not acknowledge.
The irony is that if there was a strike against Iran in the name of "International Law", the question is exactly who's International Law did the protagonists have in mind?
Black Swans feed on such ironies.
Disclosure: No Positions
