Don't always bet on black
Don't always bet on black
Nassim Taleb, author of The Black Swan and Fooled by Randomness, is interviewed in this month’s Fortune magazine. The interview is excellent, and so are his books. But when Taleb’s theories go mainstream, it’s time to start thinking about what is likely as well as what is possible.
The crux of his argument is based on Karl Popper’s falsifiability theory. It was once assumed by Europeans that all swans were white but, no matter how many white swans were seen, the theory could never be proved, only disproved (which, in fact, it was with the discovery of black swans in Australia). Just because something has never been seen before doesn’t mean it doesn’t exist.
The risk models used by the world’s largest financial institutions, which rely solely on what has happened in the past to predict what will happen in future are, therefore, useless (and, in Taleb’s opinion, extremely dangerous).
The meltdown of the US mortgage market and the hundreds of billions lost by said financial institutions has thrown Taleb’s theories into the limelight. Although he’s described this financial disaster as a ‘grey swan’ at best (being somewhat predictable), people have thrown the past out the window and are looking for black swans all over the place.
It’s now time to remember that black swans are rare. Flip a coin 10 times and the probability that you flip 10 heads or 10 tails in a row is 1/512. But flip a coin a thousand times and the probability that you’ll get a run of at least 10 heads or 10 tails, at some point, is more than 85% (it’s been a while since I’ve done probability theory, so correct me if I’m wrong – I get 85.6%). Play for long enough, and it's highly likely that you’ll get a freakish run or two.
But they’ll be few and far between. And at the moment many securities, especially in debt markets, are being priced as if there’s a black swan swimming in every pond. Set yourself up so that the rare events won’t send you broke – little or no debt, thoughtful diversification and a thorough understanding of what you own – and small risks can be rewarded with extremely attractive returns.
Custom Search 2
Comments
I recently came accross your blog and have been reading along. I thought I would leave my first comment. I dont know what to say except that I have enjoyed reading. Nice blog.
Tim Ramsey
Steve
I'm not sure it's quite correct to say that institutions' models are useless because they rely solely on what has happened in the past to predict what will happen in future.
I think Taleb's point is that gaussian models don't work in scale-invariant environments. In other words, if you are a cobbler, the odds of you becomming a millionaire are incalculably slim (slim enough to call zero). Your shoe-making capabilities don't scale.
However if you write books for a living, you expend no more effort to sell 10 million books than you do 10.
Accordingly, it is inappropriate to model your chances of success using a normal-distribution curve.
Justin
Justin, I think I'm missing the connection between your points. Could you explain it a little further?
On my reading of the book, the banks' models are a little like the thanksgiving turkey Taleb talks about. Many banks base their risk models off, say, the past five years' market movements which, like the turkey's experience, may be very different from what tomorrow will bring.
The point on the cobbler vs author (or funds manager, say) seems a very different one to me.
I wish I was smart enough to offer something on the merits or otherwise of gaussian models in scale-invariant environments, but - courtesy of the Unitab/Tatts 'merger of equals' - I own a few Tatts shares and I have to go and deal with the big ugly dark grey swan that has just landed on my pond.
Hi Greg!
There is a way to reconcile what Justin said and the basic idea of this article.
What Taleb said is that the real world that we live in is "extremistan". That is, most of the most important events don't follow a linear pattern. Outcomes actually happen more exponentially than we think they are, due to a quirk of the human brain. Part 1 of his Black Swan book lists out the quirks-narrative fallacy, confirmation shmonfirmation, luddic fallacy, etc. In fact, all these brain quirks can be summarised as this: all of us are basically much more superficial than we think we are.
Many financial models are "useless" (to use Steve's words) is because no matter how complex and comprehensive mathematicians can make them to be, they are superficial when compared to the real world. Simplistic extrapolation to the indefinite future is just only one symptom of superficiality.
A very good example is this: economists come up with extremely complex computer-driven models to 'predict' how the economy will go in, say 12 months. But the problem is that economists tend to be too narrowly specialised to see that models can be easily stuffed up big time by events that is completely outside their area of specialisation (e.g. political events, terrorism, bird flu are extreme examples. but you get my drift). The same goes with financial analysts who are extremely good at crunching numbers but who cannot see complex systemic interdependence between their little financial world and the real world. One journalist wrote about Sol Trujillo's ridicule of analysts (emphasis ours):
Trujillo opened up on the sort of long-term group returns he is aiming at when I pointed out that most of the analysts say that Telstra will generate big future cash flows but not substantial profit rises. “Yeah, well, the analysts here in Australia that have written about Telstra have been absolutely wrong almost across the board,” he says. “Look back at what they said in November 2005 and what has happened since then. So that’s why some people are analysts and why some people are managers and leaders.
Now, this weakness applies to value investing philosophy. We know you guys at the Intelligent Investor are doing a great job and so, the following criticism are by no means reflection of you- rather, they are just our critical opinion on the value philosophy. You see, value-investing follows a bottom-up approach. It evaluate businesses based on its own individual merits. As a result, it holds a neutral view on big macroeconomic big picture (e.g. inflation, current account deficits, money supply, credit growth, etc). That's where the weakness lies- because it holds a neutral view of the big macroeconomic picture, that vast area becomes an area where Black Swans lurks.
That's where the blind spot of value-investing is. Some investment decisions makes a lot of sense when you look at it from a bottom-up perspective. But when you look at the 'fundamentals' of the macroeconomic big picture, it does not make sense. Take an example with your recommendation of Infomedia. We understand that it is a fantastic business and we like the economics of the business. But when you look at this businesses in conjunction with the big macroeconomic picture, we knew it is not the time to get into that business yet. The US dollar was losing value and will continue to lose value due to the policies of money-printer Ben Bernanke, which was 'scripted' by the powerful establishments in the US economic and political scene. As such, we took the decision to sell Infomedia at the beginning of last year at around $0.78. That does not mean we do not like Infomedia's business- on the contrary, we think it is still a fantastic business and will come back to it when the time is right.
We agree with many of your bottom-up analysis of other businesses as well, but when we study the macroeconomic big picture, we discern that we are at the turning point of the global business cycle. At turning points, by definition, the status quo changes, and as such, will play havoc on many analysis that comes from the bottom-up view. Therefore, turning points are one of the most hazardous time to make investment decisions and as such, one should take extra care.
Also, we respect Warren Buffett. When it comes to value investing, we listen to him. But Warren Buffett's expertise does not like in macroeconomics. So, when he talk about macroeconomics, we would not like to treat him as if he is at the same level expertise as a value investor.
In response to the above post, this point only holds true if you believe you can pick when the economic cycle changes and in my experience, no one does. It would have been easy to sell out of the market in 1998 during the Asian crisis, believing the cycle had changed. Similarly, it would have been easy not to buy in 2003 before the Iraq war. Both courses of action would have been a mistake.
You're right, turning points are hazardous times in which to make investment decisions. But one has no way of knowing beforehand when a hazard pops up, nor does one know that one is facing a turning point, by definition, until after it has occurred. The thing is, it is this very uncertainty that usually creates the buying opportunity. To wait until there are no hazards and no turning points is to miss these opportunities and probably buy in at the top.
You're whole argument sounds reasonable but is flawed because it is based on the premise that you know what is about to happen, and I don't believe neither you nor anyone else does. As Peter Lynch says, 'if you spend 13 minutes a year thinking about economics, you've just wasted 10 minutes.'
Hi John!
You’re whole argument sounds reasonable but is flawed because it is based on the premise that you know what is about to happen, and I don’t believe neither you nor anyone else does.
We have heard this line, "no one knows" before many times.
The statement, "no one knows what is about to happen" is a very vague statement.
What is the level of ignorance when you say "no one knows"? Is it at one extreme level of complete utter cluelessness? Or is it towards the other extreme of perfect knowledge of exact precise point of turning points?
If you say that "no one knows what is about to happen," then why is it that this does not apply to value investing? There are some practitioners (the trend followers, who made serious claims to have a record that challenges Warren Buffett) who scorn at value investing because they use the same line of "no one knows."
Do you see the contradiction when you says "no one knows"? One one hand, you say looking at the macroeconomic big picture is a waste of time because "no one knows." But on the other hand this reason does not apply when evaluating businesses.
As Peter Lynch says, ‘if you spend 13 minutes a year thinking about economics, you’ve just wasted 10 minutes.’
Peter Lynch says this because economics is outside his area of competence, not because economics itself is the problem.
... and in my experience, no one does
Do you see this area of Black Swans in your realm of knowledge?
The problem with economics is that it is dominated by the mainstream neo-classical school of thought. There is another school of economic thought called the Austrian School. They have an extensive business cycle theory that is not well understood and not taught at uni/MBA.
The mainstream school of economic thought does not have a proper theoretical framework on business cycles. It has the roots on the German Historical school in which all forms of qualitative analysis is done away with and completely driven by statistics and numbers. An analogy would be to have a super-precise radar that tells you what is happening in every square millimeter of water but does not offer any understanding of cause and effects of tides. That's why it gives people the idea that "no one knows."
Also, we don't profess to be able to predict the exact timing of business cycles- no one is able to. But as the Intelligent Investor once said before, it is "better to be vaguely right than to be precisely wrong."
It would have been easy to sell out of the market in 1998 during the Asian crisis, believing the cycle had changed. Similarly, it would have been easy not to buy in 2003 before the Iraq war. Both courses of action would have been a mistake.
When for those who sold off in 1998/2003, what is their basis of reasoning to believe that turning point has turned? Is their reasoning based on hunches, speculations, guestimates? Or they have have a proper framework to understand what is going on?
When people says "no one knows", it shows that this is area of gap in their knowledge, which is where Black Swans lurks. As such, when things happen, it appears 'random' and 'unpredictable', which strengthen their convictions that "no one knows." The mainstream school of economics has gone so astray that it offers completely no insight of cause and effects, which gives people the impression that "no one knows."
You get our point? We are not saying bottom up value investing is useless. Rather, value-investing with complete utter cluelessness about economics makes one more vulnerable to Black Swans. Understanding value-investing in conjunction with understanding of proper economics helps one sees a broader dimension. As Warren Buffett himself says, to be a good investor, you need to read widely and broaden your range of expertise.
Yes, I get your point. You're trying to advertise your supposed expertise to a far bigger potential audience than you attract to your own site.
John:
Just to clarify: we are NOT saying you are clueless about economics as you may take it for us to mean it that way. What we mean is that knowing value investing while at the same time, (1) completely disregarding a gap in one's knowledge and (2) make decisions as if that entire gap does not exist is inviting poorer decisions.
For example, take these 2 line of thinking:
1. "Economics is outside my area of expertise."
2. "I do not know about Economics and will completely ignore this realm of knowledge and I'll base my decisions by assuming that business cycles and whatever macroeconomic factors will not have any impact on my investments."
When a person says, "no one knows", it shows that that person is thinking along the lines of (2).
Let us quote page 140-141 of The Black Swan by Nassim Nicholas Taleb:
Epistemic arrogance bears a double effect: we overestimate what we know, and underestimate uncertainty, by compressing the range of possible uncertain states (i.e. by reduce the space of the unknown).
…
I remind the reader that I am not testing how much people know, but assessing the difference between what people actually know and how much they think they know.
We are trying to drive a point, which somehow was probably misconstrued by you as insulting your intelligence (that's why face-to-face meeting is better form of communication).
Our point is that (and sorry if we're repeating it the 3rd time) is that when one think along the lines of "no one knows" (i.e. number 2), one may actually fall into the danger of making bad decisions by increasing the difference between what one knows and what one thinks he/she knows.
On the other hand, if thinks along the line of number 1, one is actually decreasing the difference between what one knows and what one thinks he/she knows.
John Addis, what a childish behavior!
Regardless of whether you use just the business valuation or macroeconomic assumptions as well, at the end of the day it comes down to whether the assumptions you make end up being right or not. With Infomedia cited as an example, it is possible that there business will grow well enough for the share price to recover without need for the USD to recover. The problem in most cases is that people only think about the explicit assumptions that they need to make to come up with some numbers, and not about the implicit assumptions that they are making and not thinking about.
Matt:
Well said! We certainly agree with you on that! :-)
As investors, it is important to search our own minds and dig out all those implicit assumptions that we make sub-consciously. That's why we still regard Infomedia very highly.
Hi Steve Johnson!
Sorry for this yet another comment. Hope you don't mind.
There is another crucial concept about Black Swans that you've left out. Without that concept, it can seriously undermine understanding about them.
Black Swans (according to Nassim Nicholas Taleb) are:
1. Statistically improbable (which you've covered very well)
2. Massively high impact/consequence (which you've missed)
Therefore, the coin-toss analogy that you used is incomplete. This is because, the consequence of consecutive 10 heads is immaterial. The world is not going to end if one get 10 heads in a thousand coin toss.
A better analogy would be parachute jumping. Why would parachutists carry a secondary reserve parachute? Based on statistical probability, it is rare for primary parachutes to fail. But if it fails, the consequence is unimaginable (death) if the secondary reserve parachute is not available. Therefore, primary parachute failure is a Black Swan event. Packing a reserve parachute is a very prudent measure.
In the context of the financial markets, if you look at pricing of some securities, it may not be that irrational after all if you take into account that Black Swans events, although are improbable, can take on colossal consequences if they happen (touch wood).
To clarify ...
I wanted to make the point that the risk models used by institutions are not useless because (as you suggest) they 'rely solely on what has happened in the past to predict what will happen in future'.
The problem is that these models use historical data *selectively*. In other words, they rationalise-away historical black-swan events.
We all know that the map is not the territory and that models are a selective representation of the essential elements of reality -- that is their very purpose.
However, problems occur when those elements of reality deemed non-essential have massive impact (as another poster points out above).
The solution is NOT to build a better model. Reality itself is the only totally-dependable model! The solution is to explicitly recognise the inevitable shortcommings of our models and the possible impacts of their fallibility.
10 heads in a row is a high impact event for someone offering odds of 1000-1 on it happening (which would be a fair price). Which is exactly the reason you see poker machine manufacturers limit the number of times a gambler can double and insurance companies like Berkshire never write a policy with an unlimited exposure. Don't ever put yourself in a situation where you can go broke.
A Black Swan event might take on colossal consequences if you were stupid enough to put your whole portfolio in one security. But when there's a portfolio of them out there, some yielding in excess of 30% to maturity, you eliminate the chance of any one event wiping you out and let the law of averages play out in your favour.
As for your comments, I've already emailed you about promoting your own site while adding little to the discussion here - I'll moderate comments from now on. There are more ethical ways to build your business than trying to farm traffic from someone else.
Hi Steve:
10 heads in a row is a high impact event for someone offering odds of 1000-1 on it happening (which would be a fair price).
Correct us if we are wrong, but in the original coin example, we did not interpret any mention of a 1000-1 'reward' for a 10 heads in a row run. Of course, if that occurs in a gambling context, then yes, it is a high-impact event.
A Black Swan event might take on colossal consequences if you were stupid enough to put your whole portfolio in one security. But when there’s a portfolio of them out there, some yielding in excess of 30% to maturity, you eliminate the chance of any one event wiping you out and let the law of averages play out in your favour.
We agree with you on that, provided that there's no correlation in the diversified securities. Platinum wrote a report before, the mistake of those who made the CDO models is that they assumed no correlation in the various tranches of the CDO securities. It turned out that in the real world, there is strong correlation and thus, those CDOs blew up.
As for your comments, I’ve already emailed you about promoting your own site while adding little to the discussion here - I’ll moderate comments from now on. There are more ethical ways to build your business than trying to farm traffic from someone else.
We do take note of your email and has changed from what we used to do. Rest assured, for this particular blog post, we received neglible traffic from your site. Some points (you're welcome to disagree with us):
1. We don't consider ourselves a commercial business. So, we have hardly anything to gain from farming traffic (except for a few miserly cents of advertising revenue). If you think we are a commerical business, then we thank you for thinking so highly of us. In that case, it's our vanity that leads to this.
2. We do want to take part in discussions (we do take part in lively discussions from other sites and forums too). If it is our intention to farm traffic, we would not have bothered to spend so much effort to write up such an extensive reply for so little 'reward'.
3. It could be that the discussions that we took part in other sites happens to be non-commercial in nature, and perhaps that explains the difference in response from those blog/forum owners.
We tend to be more talkative and loves debate more than average. So, that explains why we write so much and perhaps get thrown out by commercial blogs (but on the other hand, the non-commercial ones tend to be more accomodative to and even welcome our long-winded rants). Or sometimes, we ended up offending people unintentionally.
But that's the whole nature of blogs isn't it? People rant, people argue and worse still heap insults at each other. But sometimes, the most interesting off-tangent discussions also take place in blog comments.
Sorry for taking much of your time once again. We will shut up now...
The difference between Steve and Contrarian is a clash of philosophy. The coming digital showdown:
We are not even a decade into the digital millennium and already the battle lines have been drawn. Two camps have emerged, each with widely divergent views on the nature of information, who owns it and how it should be distributed.
The forces are at this stage evenly matched, and it is not apparent from the day-to-day squabbling which side will emerge victorious. But one side must, because their views are diametrically opposed and can't coexist in the long term.
On one side are those who believe information is a commodity that can be owned, bought and sold, and its distribution controlled. This naturally leads to a restrictive view of information. This group comprises most of the music, publishing and film industries, and most hardware and software companies.
On the other side are those who believe that information by its nature should be free, and that its distribution should be uncontrolled. This viewpoint naturally leads to an expansive view of information.
This group comprises the open software movement, a few far-sighted computer companies (Google is the best example, but also includes heavyweights such as IBM and Sun), and most consumers. I am firmly in this second camp.
I don't think Contrarian's comment is doing any harm to Steve's blog. I think it helps Steve because people can get to read interesting discussions here.
Contrarian, you've written too much. Tone down a little.
Sensible business practice has the greatest effect on a successful business regardless of the macro-economic conditions. If I had paid attention to every forecast, made over the last 25 years, of a black swan discovery, I would not have ventured into the stock market at all. Recognizing that there are risks in every investment and attempting to minimize those risks is the best strategy. There are lots of companies which go bust during the boom times i.e. read bad business practice. Subscribing to Intelligent Investor is one method I use to try to minimise my risk.
Hi Dougie!
We do agree with you on most point. Investing in sensible business goes without saying. Except on that part of Black Swan discovery. By definition, you cannot find Black Swans because if you can find them before they happen, then they are not Black Swans any more.
You may be surprised to hear that, but we don't listen to economic forecast either. But we do try and understand cause and effects from a macroeconomic perspectives (the mainstream economists are weak in that area) and that acts as a supplement to our investment decision. For example, if one had invested in 1929, it was a period of great discouragement to all investors, not just value investors. But Austrian School economists, Ludwig von Mises, saw the coming catastrophe a few years before the 1929 crash and explicitly warned about it. Of course, no one listened.
Subscribing to the Intelligent Investor is a good idea because it is far better than many of the dodgy tip sheets in the market today. We like the value philosophy of Intelligent Investor and the truth is, many people don't understand that philosophy. For example, some people thinks that a "Buy" recommendation automatically means the stock price is going to go up. Of course, that don't always happen and they blame Intelligent Investor. The problem is, they don't understand value philosophy at all.
[...] pretty strong stuff, but the blogosphere hasn’t taken him to task for it. The Bristlemouth investing blog frets that Tasseb’s theories are going mainstream. The Skeptical CPA grumbles [...]
I think Contrarian Investors Journal has been given bit of tough time.
So let me defended the old liberal values of the freedom of the individaul is the only truly progressive policy and old liberal values are becoming black swans.
Example is the gambling industry government central planning and how one policy change can cause shareholders to lose big time.
We have become dependent on big government and central planning to save us and this will be the real black swan going forward.
Division of knowledge was the greatest discovery of the century and will win everytime against large central planning.
Interest rate cuts in America only going to cause more inflation pain for ordinary people buying food and petrol.
Again government central planning is going to cause a black swan event not any one policy.
Stephen Johnston
I HATE and DETEST that the managing director gets paid (one could hardly say earns) 22 million dollars. That is obscene greed - nobody needs that much money and there is intrinsic dishonesty in him accepting that while a lot of people who struggle to pay their bills have lost a lot of money in these dopey shares. There should be some law to stop people who produce nothing of value for the world being paid huge amounts of money, in fact any money at all. He and his ilk are repulsive, slimy and just plai horrible camkers on the human race and society.
Post new comment