Theory versus Statistics, Financial Economics Edition.

May 4, 2009 7:14 AM

Theory versus Statistics, Financial Economics Edition. "You can almost here the lament of this quant that the real math theory has been dead since 1980, and that it has all been applied and statistics ever since. It’s like Fischer Black was Kool Herc and Myron Scholes was Afrika Bambaataa, and they’d all go plug in their computers into lamp posts and do martingale representations in the streets and at house parties. And, of course, it was all ruined in 1979 when it went commercial." A response to The Last Temptation of Risk by Barry Eichengreen.
posted by chunking express (8 comments total)

Empirical analysis, ten dollar words for "common sense". Nice to see some economists get it. I chuckle too, at the efficient markets theories. Have they forgotten who it is they are modeling?

The problem is, though, that economists, like the rest of us, have 20/20 hindsight. Ask an economist what will happen in the future, and you'll likely get a long winded precisely articulated version of "I don't know". So people ignore economists, and listen to the guys with the "mathematical models", cause the guys with the mathematical models will aslo say "I don't know - but I have a mathematical model that says this."

And, right or wrong, they proceed on that basis, like the Donner party did.
posted by Xoebe at 8:28 AM on May 4, 2009

These look like interesting reads - but for now, I'm just going to savor that image (martingales in the street, etc) for a while with my morning coffee. Mmmmm.
posted by freebird at 8:45 AM on May 4, 2009

For the most part the guys with the mathematical models are the economists. Just a different set.
posted by JPD at 8:56 AM on May 4, 2009

Empirical analysis, ten dollar words for "common sense". Nice to see some economists get it. I chuckle too, at the efficient markets theories. Have they forgotten who it is they are modeling?

Huh? It sounds like the guy is arguing for the opposite: that we should dispense with statistics and in staid use pure models based on whatever crackhead models economist come up with based on their idiotic "rational actor" dogma.
posted by delmoi at 9:26 AM on May 4, 2009

One of the most compelling arguments that Taleb made, for me at least, was that economics is currently where medicine was in the middle ages. He'd probably argue that we are somewhere near Galen and less towards Vesalius, but the fact remains that while we might have some theories that hold well to logical analysis, we have such a dim idea of the fundamentals that trying to apply the science to the technology is completely fruitless. In "Bad Medicine" Wooten makes the claim that up until fairly recently, mid-nineteenth century, that doctors would have done better by practicing homeopathy than anything derived from Hippocratic teachings, which while holding up well to logical analysis, fell resolutely in empirical analysis. Homeopathy's greatest strength is not so much that it is treating the fundamental problem but that it simply isn't doing harm itself.

It is funny that economists are still talking about the massive data sets as proof of their empirical verification, when if anything, the massive data sets prove that there is no predictive qualities at all! I'm sort of reminded by the anecdote in Wooten's book of doctors believing that the urine being nearly all you needed to know about a patient (as if the body is big Markov chain of question marks, and the body outputs innumerable measurements that help describe these unknowables, it is almost scary how similar it is to finance today!). Since the urine was so important, then the input of water was then just as important, with all sorts of measurements deriving from it, like strength and hardness (two separate qualities!) and a complete matrix of good and bad water sources (lakes, streams, hot water, cold water), that almost begins to look like a co-variance matrix, and is all self-fulfilling and logical in a way only an academic or medieval philosopher could love.

Fundamentally the current crisis was caused by a debt bubble and Scholes-Black or anyone with a computer had very little to do with the fact that banks were working in the interest of shareholders and the only way to increase their equity is to increase their liabilities. The higher math and computation was nothing more than astrology, a framework people could believe in. If we had not had computers to crunch the numbers, there'd be a valuation technique (peer reviewed!) that would have come through and enabled the same thing.

The real question is why do I constantly hear about "stress tests" performed on the banks and other nonsense that just perpetuates the current, failed system. It seems as if everyone believes that we are debugging a giant program and just need to get it right. If there was a new, complex regime for AIDs patients and 5 years in 30% of them died and the rest were left much weaker than before the regime was started, would you want to want to continue on the same regime only because tests on mice indicate that it can't make things much worse? No of course not.

I really blame schools for this, because they present math as some sort of Nerd God, have you read course descriptions for hire level math? It is peppered with all kinds of "Difficult" "Demanding" etc., no it is not that hard for a smart person to sit down and learn. But you create this aura that if you can make it work in a structure of mathematical logic then it must be applicable to anything, anywhere, when really the math is good, we just have no idea of what we're talking about.

BTW I'm a big believer in Mutant's cash flow method which he's demonstrated many times, to the point where I use it almost exclusively when investing in anything that is not cash. I use a percentage of the cash flow generated to invest in put and buy options (like when oil was at $150bbl I was buying "cheap" put options for it at $75, $55, etc.) I cannot say how it will work in the long term but I've done a lot better than my parent's 401k. It is also very, very rewarding to hit the lottery and see some cheap option give a huge pay off, even though most of the time I lose. It also reflects pretty accurately how people live (via cashflow), where keeping money in a non-dividend paying equity really doesn't do much for you except perhaps lower your borrowing costs (which is how banks got into this in the first place!).
posted by geoff. at 9:28 AM on May 4, 2009 [2 favorites]

This is more punchline than insight but economists model firms as though they were run but really talented economists and at cocktail parties they ridicule MBAs for being faddish superstitious dummies who aren't nearly as smart as economists.
posted by I Foody at 9:49 AM on May 4, 2009 [1 favorite]

I don't get it. What about statistical theory?
posted by Mental Wimp at 3:12 PM on May 4, 2009

I don't think Kool Herc and Afrikaa Bambaataa appreciate being dragged into this.
posted by altie at 2:17 PM on May 5, 2009