I never met a rich economist. They are all broke. They always try to figure out what will happen before it does and they end up losing.
I like that quote taken from the swedish army training manual. My point is simply that you can not produce accurate models of risk. Risk of a blowup is always there no matter what. The only way to limit risk is to reduce exposure and this is precisely why most of my hedge fund friends leverage 7 or 8 times capital max. But to reduce exposure is to essentially participate less and thus not put money to its full use in the markets. Significant alpha becomes difficult to attain if money is not being used to the fullest extent possible. From a competitive standpoint, 8000 other hedge funds out there are competing with you and so you must be sure to take enough risk (which, as I stated, is impossible to accurately quantify) to be competitive. With so much competition, hedge funds are being forced into less diversified strategies in order to profit in an increasingly efficient market. Less diversification means less hedging which means much more significant market risk....I hope I didn't just predict a huge blowup the likes of which the united states has not seen before.....since 90% of all hedge fund funds are held by 300 or less hedge funds, this may not be an accurate prediction. However, it does seem to point to a decrease in the number of hedge funds that stay in business... That said, I try to keep my models as simple as possible and often consider risk estimations to be underestimates of what to expect in actual reality. Someone once told me that LTCM had calculated, after the fact, that what happened to them was a 1 in 100,000,000 anomolaic event. They had so many bets on and for the structural relationships between all their hedge pairs to crumble at the same time was calculated to be a statistical impossibility. I never did have much faith in science or mathematics despite having studied mathematics as a major in college. Why would I have any more faith in complex models whose flaws will only be revealed after I have lost money using them? I like to keep it simple. Keeping it simple is no more risky, in my opinion, than relying on models of risk. Please note, that on this issue, very few of my peers agree with me.... But that is okay....i've always been a bit at odds with common consensus....
Here is a joke with implications that ring true..... An engineer, a chemist, and an economist are stranded on a desert island with no food source but the ton of canned goods sitting on the beach. They are trying to figure out how to open it using only the sparse vegetation found around the island. The engineer gets to work making a complicated catapault device to burst the cans on impact with the sagebrush. The chemist gets to work trying to extract acid from the local plants. The economist, meanwhile, just sits on the beach. When the other two ask him what he's doing, he says. "The problem's very simple. First, we assume a can opener, then we . . . "
Although you can't produce accurate risk models with statistics based on the normal distribution, you can if you use other distributions (e.g. Cauchy, Power-Law, etc.). Whereas the Normal distribution gave the academic idiots at LTCM the false confidence that a blow-up is a 1-in-100,000,000 anomaly, other statistical models give much more accurate predictions of the very high likelihood of very bad results. Very perceptive and mostly likely true (YIPES). Greed chases profit to create a nasty positive feedback loop. High returns both attracts more money to a sector AND creates a false impression of low risk -- its a recipe for disaster as you point out. In "Against the Gods: The History of Risk" the author talks about the repeated crashes of the Mexican currency -- that high Mexican interest rates attract too much foreign capital then some minor crisis causes the currency to tank and the cycle starts again. Each time, investors' greed for profit overcomes investors' memory for loss. And clearly their risk models were deeply flawed in ways that could have been easily proven before LTCM took one dime of investor money. That LTCM used such flawed models that underestimate the risk suggests some combination of ego or greed on their part. Math is both more insanely powerful than most realize and more pathetically weak than most realize. As a means of undertstanding the structure of logic systems, math is a very powerful means of generating new insights. But as a means of undertstanding the world, it can be deeply flawed if the real world doesn't behave according to idealized mathematical axioms. (In this regard, the Central Limit Theorem does not apply to markets and invalidates LTCM's analysis of potential risk). It's not a matter of using simple or complex models, its a matter of testing the alternative risk models and not using models that have clear flaws. LTCM (and many economists) have failed this elementary step and continue to fail this elementary step, hence the assertion that economics is rubbish.
Lets say you have two economists, one is named white the other one is named black, both can predict economic performance 5 out of 10 times correctly. Whiteâs thinking is in tune with most economists (conventional thinking), while only a few economist thinks like black. White would probably loose money in the marked while black could probably win, since the benefit of being right when most are wrong is bigger than the loss of being wrong when most are right money wise.