Thursday, January 3, 2008

Looking For Little Black Swans


Click on Table for full-size and details

Nassim Nicholas Taleb's "The Black Swan" (Random House, 2007) has attained cult status with the hedge fund crowd. His book is about the enormous consequences of highly improbable events. The essential points are:
1. Black Swans are the highly improbable events that have had enormous historical impact- like the current credit crunch caused by the sub-prime crisis, 9/11, the Latin American debt crisis of 1982, or the fall of the Soviet Union in 1990. Though very rare, Black Swans have many times more impact than more predictable, more frequent events. Black Swans are by definition unpredictable.
2. Black Swans are not as improbable as conventional statistics theory would have us believe. This is because the way we calculate probability has always been, and continues to be based upon the Gaussian [Bell Curve] distribution, where anything beyond 3 standard deviations is considered to have a probability of less than 1 %. Black Swan distributions are non-Gaussian, with fat tails [extremities] that could be 20 standard deviations long . Technology and modern society has made the world more interconnected resulting in an even greater degree of non-linearity i.e. the impact of events are magnified on an exponential scale, and totally unpredictable. [Think of it this way: there are now more nodes in the global grid whether social, economic or ecological. Network theory has demonstrated that networks tend to have a few very big nodes which have many more times the links of the average node. Everything hums along nicely until something happens to one of these big nodes]. Which means we can expect to see more and bigger Black Swans in the future.
3. Not all Black Swans are bad. There are good Black Swans too. And riding on good Black Swans can lift you to heights unimaginable. Some jobs are more exposed to Black Swans than others. Novelists, song-writers, sportsmen and hedge fund managers are more likely to meet a Black Swan than dentists, engineers, or the neighborhood grocer.
4. The nature of investing makes outperformance more a matter of luck, than expertise. Successful Fund managers, investment bankers, politicians, media journalists, economists, statisticians wrongly attribute their success to skill when it is all due to luck.
5. In the stock market you cannot predict a Black Swan but you can try to put yourself in a position where you are more likely to meet a [good] Black Swan.
6. Since Black Swans are dangerous, expose yourself to many little Black Swans instead of one giant Black Swan. e.g. Only use 10 % of your investment capital for to bet on Black Swans. The other 90 % should be invested in blue-chips or Treasuries. Even if all your supposedly good Black Swans turn out to be bad Black Swans, it wouldn't kill you.
7. The application of Black Swan philosophy need not be restricted to investing. You can apply Black Swan philosophy as a decision-making tool in the conduct of your daily affairs. At any time, when faced with a decision to make, ask yourself two questions: (1) what are my maximum possible losses? (2) what are my maximum possible gains. If (2) far outweighs (1) then go ahead, and vice versa.
While we don't totally agree with Nassim Taleb's 'treatise' let's leave the riposte for another day#, and meanwhile attempt to devise a ValuEngine screen that makes it more likely that you will meet some good little Black Swans. Our screening criteria should be such that anything that smacks of prediction based on 'expert' analysis of fundamentals and past patterns [all based on the Bell Curve of course, since the world has still not found a practicable alternative to it] should be excluded. Forget about valuations, Engine-Rating, expected EPS growth, and track record of returns. The past counts for naught in Black Swan theory. Instead, concentrate on:
1. 1-month forecast return rank >80 < ---- The further out we try to forecast, the more the effect of accumulated errors that arise from the highly non-linear nature of stock markets, so use 1-month forecast return rank. The shorter the term of forecast, the more likely that it will be due to random and not fundamental factors, which is what we want for the purpose of meeting Black Swans. 2. Volatility is the key: Volatility is conducive to the birth of Black Swans. We screen for stocks with Volatility Rank less than 30. 3. Market cap >$0.5 bil and average volume>100000<---- we can't handle too many little Black Swans 4. To increase the Black Swan effect, we only screen for stocks in the most beaten-down sector viz Finance. Then whittle down to the ten most volatile. So lets see how these little Black Swans fare in the next few weeks. * the reason why we use Ranking instead of absolute values for this screen is that a Black Swan must be seen in relation to, and within the context of the current situation, i.e. the market as at now, the statistics of our stock Universe as at now. Ten Little Black Swans picked up on Dec 31 2007 [DJIA 13264] as at today * DJIA is down 1.57 % since Dec 31. But most of our Black Swans are down by > 1.57 % and the portfolio as a whole is down by 3.82 %. So Black Swans are definitely more volatile than than the market on the downside. Let's see if they are also more volatile on the upside.

No comments:

Post a Comment