Slashdot Mirror
Why Economic Models Are Always Wrong
mayberry42 writes "Did you ever wonder how and why professional economists often seem to get it wrong in terms of predicting consequences or policies accurately (or even at all)? Or how very few even saw the current economic collapse? This article provides an interesting, if obvious, reason as to why economic models are effectively always wrong."
Most economic models are based on "how we would like people to act" rather than "how people actually act". Much of the time, the model works, but they fail when people act in irrational ways.
Simples.
"She's furniture with a pulse"
I wonder if that could apply.
... is not a science. The legal structure of money, the way prices work in a one way fashion, and private ownershp are all political all the way through. Now this may piss off Americans but there are alternative ways to organize society whether they like it or not. Human beings tend to be people of their era and they often have a profound lack of imagination, the black and white right/left thinking I see from people already disqualifies them for not even having the courage to analyze or think about the structures and societies in which they find themselves, the false notion that it is either THIS/THAT, BLACK/WHITE is having given up critical thinking and analysis for good.
I think the article is missing another key factor - the fact that people abuse the system to their benefit.
Starbucks, Harbuckle of Breath.
So small changes in inputs can produce big, unpredictable changes in the output of complex systems? It's almost as if a butterfly flapping its wings could affect the weather!
They should find a snappy name for this marvelous discovery. Something like "chaos theory".
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
Many, many, many people saw the economic collapse.
I was reading plenty of blogs on the housing bubble, housingpanic.com, et etc, describing the preposterousness of "liar loans", subprime this, and idiocy that, and the crazy valuations.
The New York Times even had a plot of the inflation-adjusted Case-Schiller price index which was enormously above any prior peak. During 2006 and 2007 and 2008.
The notion that "nobody" saw it is simply propagandistic truthiness baloney. I personally didn't profit, because I was much too early shorting the mortgage companies & home builders and got stopped out---the bubble was too powerful.
The real crime is that a small number of very powerful people had an exceptionally lucrative interest in NOT stopping it, because they were getting ginormous paychecks from the continuation of the bubble. And now the notion that nobody could see it is used as excuses for the powerful to excuse themselves from responsibility from fraud and crime.
Down in the guts of banks, there were both risk modeling quants in the fancy banks, and the traditional "ladies with a bun" in the retail banks who processed the paperwork who saw how much outright fraud and insanity there was. Their jobs were threatened when they attempted to speak up and stop the madness, because the business side executives were making shitloads of shekels on volume.
Few people saw the collapse coming? Really?
All you had to do was turn on some form of broadcast radio after about 1995 and listen for a little while. When the commercial break appeared you heard one mortgage mill after another hawking refis, credit lines, etc. Bad credit? No credit? No problem! Interest only mortgage. Balloon mortgage. Jumbo mortgage!
This went on for years and years.
I saw it coming. If you missed it you're a fool. Maybe we just have a lot of fools.
Even if everyone acted rationally, you would then have the instability which is generated because all of these rational people would then change their behavior based on ... the model. It's unclear, and in my eyes rather unlikely, that a "fixed point" exists where all of these rational people start behaving identically and predictably.
The unpredictability doesn't only come out of irrationality. If you look at game theory, you see that many optimal (i.e., rational) strategies are "mixed" strategies where the rational party necessarily behaves probabilistically, not deterministically.
Models aren't equal to models, and even rough models of chaotic phenomena can be very useful and predictive, if they are the right ones. Read this for some acknowledgement of which brand of economics has been right during the last few years. Here is another account, including some pointers to predictions of the current crisis reaching as far back as 1999. Krugman even has a "model" of how good models get out of fashion.
Economics suffers from the manipulation by political interests, and by the wish of many practitioners to project their moral ideals onto the world. Many economists simply go and try to prove that the world works however they want it to work, and find funding for that from rich supporters. That makes the endeavour biased.
But it is an investment.
What the housing bubble became wasn't investment. It was speculation. And every serious investor will tell you: never speculate with what you can't afford to lose.
In a religion, you just tell people what is the Truth. In science, you try to observe and learn.
The models are self fulfilling prophecies.
The high priests of the Economy tell us the Truth. The lower priests spread the word. And the people believe. Without the belief of the people, the system would instantly collapse. And if reality turns out differently, then they/we just invent a New Truth.
I mean, is it really necessary to give trillions of euros/dollars to banks to bail them out? In which pockets is that money disappearing? The bailouts are presented as "The Only Way"... but nobody actually knows.
From the article:
Wait ... you are saying the growing number on my bath scale isn't because the constant of gravity is growing? :-)
The Tao of math: The numbers you can count are not the real numbers.
"Or how very few even saw the current economic collapse"
Y'know, there's an entire school of economics that predicted the collapse. And the collapse before it and the ones before that. It's called the Austrian school. But even though they predicted every single damned collapse because they didn't use shiny models and after the mid 90's shiny powerpoints nobody pays any attention to them.
"Remember that all models are wrong; the practical question is how wrong do they have to be to not be useful." (George E.P. Box and Norman R. Draper, Empirical Model-Building and Response Surfaces (1987), p. 74)
"One of the most insidious and nefarious properties of scientific models is their tendency to take over, and sometimes supplant, reality." (Erwin Chargaff)
I think that says it all, really.
--Bud
Economics tries to take the veneer of science by using a lot of mathematics.
Just like Astrology.
-jcr
The only title of honor that a tyrant can grant is "Enemy of the State."
Many, many, many people saw the economic collapse.
A newsletter from an economics professor and CNBC financial commentator: ...
Any talking head who tells you that this market is a buying opportunity has his/her head screwed on backwards. The only buys are the kind of value plays that the likes of Buffett are pulling off. That is, it is very much a stock picker’s market.
Recession plus inflation plus a credit crisis plus a softening European economy plus an inflation-plagued Chinese economy plus Russian strong-arming in natural gas plus two leading presidential candidates who are ignoramuses on economics plus a rising long bond in the face of Fed rate cuts does not a bull market make."
http://www.peternavarro.com/2008.02.01_arch.html
"Thursday, February 28, 2008
That is his oldest newsletter but I understand he was telling his economics students to "get out" of the market in fall 2007. Plus he was showing them a whole bunch of historical indicators that were all pointing in the wrong direction.
Actually, I'd say Germany and the US are a pretty fair comparison. If you really believe the population of Germany is homogeneous, you should try visiting sometime.
-- Let us endeavor so to live that when we pass even the undertaker shall be sorry. -- M. Twain
Hey! 12A is just a broom closet.
You are entitled to your own opinions, not your own facts.
Quite a few people who had good savings still lost jobs, burned through their savings and retirement funds and in the end lost their homes anyway.
The idea that only bad or irresponsible people lost their homes in foreclosure is magical thinking. You can be a responsible person and still get wiped out
during a deep recession.
There is no spoon.
My first Journal Entry ever, in 8 years! http://slashdot.org/journal/365947/aphelion-scifi-fantasy-horror-poetry-webzine
I know the article claims "Calibration--a standard procedure used by all modelers in all fields, including finance--had rendered a perfect model seriously flawed." but obviously our climate models are simple enough that you can't just calibrate the parameters until it matches the data, right?
Because all economist mentioned are Keynesian economists. Browse around mises.org. Search for articles in 2003-2007, and it is obvious they saw it comming. Here is one notable austrian economist. You would think politicians would be knocking at his door constantly to help them see. If you claim it was a fluke, here is another much more famous guy that follows austrian economy, that predicted every single recession since 83. Heck, you can also predict the next recession, just spend a few hours reading on mises.org, they have courses for free.
A relationship between money and debt is inevitable. Imagine there is no money. Suppose I do something for you in exchange for a promise of reciprocation. You are now in debt to me, in the traditional sense. There's no numeric accounting, but this notion of debt is firmly buried in human psychology and is part of the reason humans are able to build economic systems. Then you do something for me and we're even. Now formalize it: imagine, when I do this something for you, that you create out of thin air (in a ledger, in our heads, in the location of special shells which which mutually agree will represent it) a numeric representation of that debt. Then, when you reciprocate, this accounting is reversed and both debt and proto-money disappear. Or, alternatively, I could instead pass these tokens on to another person in exchange for a promise that /he/ will receive the return of your debt instead. And thus the money begins to circulate, and in effect you are the central bank.
Money IS debt. It's a transferable formalization and extension of one persons social obligation to return a favour to another. Replace it with a 'resource based economy' (I presume you mean replace token money with something like gold) and things won't be much different. People will expect that receiving gold from you entitles them to be given useful products or services in exchange for it. They will be just as unhappy if that is not the case - they'll fell just as much as if a social obligation has been broken - if that doesn't happen and they're left with gold as they would if they were left with tokens.
People did predict the economic Collapse. Ron Paul predicted it all along, as many others did. Those who "failed to predict" it are simply those economists and politicians who stood to gain financially from the whole thing. The issue isn't that it is impossible to predict what is to happen, but rather, that those who are in positions to control and predict are those who stand to gain in the bubbles that end in economic turbulent times [read the Fed, Wall Street, Washington].
Well, Carter's argument is sometimes wrong. I do Bayesian calibration of computer models, and with some models the maximum a posteriori estimate, or the posterior mean, is consistently very different from the "true" parameter values (in a perfect model simulation study). This is basically a combination of non-identifiability in the model combined with insufficiently informative priors. It's hard to do anything about this, and it's a problem if the estimated and "true" parameter values lead to very different predictions. (Sometimes they don't, and if you only care about predictions, it may not matter that your valid predictions are based on "wrong" parameter estimates.)
Game theory always does account for things like that, primarily because the behavior you're describing is not irrational. The very fact that you are predicting that "he gets ahead" is what makes it rational.
Same for your "when they zig, you zag" idea: I have never heard of anyone using game theory that doesn't account for (and in fact, predict) that sort of behavior.
If you want to come up with an example where game theory doesn't work, you're going to have to try a few thousand times harder than that.
The reason game theory tends to disappoint, is that peoples' intuitive hunches for the payoffs of certain actions don't match the theory, but those hunches are what they act upon -- and that in turn changes all the payoffs, sometimes toward causing the hunches to becomes true (!) and sometimes toward causing the hunches to be more false. And that itself can be analyzed and predicted, but only if you just happen to know what other people's hunches are going to be -- and that is never predictable.
Game theory is about finding optimum equilibriums for behavior; it can never tell you what people believe.
BTW, back onto GP's subject.. a few months ago I went on an AdamCurtis-athon with some high expectations. It was a letdown, and not nearly as serious a criticism of the targets as I had hoped, especially since I just assumed some of them (e.g. the neo-cons) would be shooting fish in a barrel. I won't say watching all his docs is a waste of time -- it's not -- but don't get your hopes up. You'll find some good anecdotes, carefully selected interesting trivia, and great quotes like the one about economists and psychopaths .. but that's all.
As copyright owner of this comment, I authorize everyone to defeat any technological measure which limits access to it.
He had two models. The first model produced hypothetical historical data (analogously, the data from 1950 to 2010). He then created a second model, built on part of the 'historical data' (1950-1999) and tested on the remainder (2000-2010). He then used the first model to produce another segment of data (2011-2020), and found that the second model did not predict this 'new' data at all.
You and he are doing the same thing; the fact that your 'historical' data comes from reality and his comes from a model is irrelevant so far as the second model is concerned. The phenomenon described by the article is known as 'over fitting the training data' in Machine Learning circles, and is widely known in other fields, according to the comments. Perhaps if you pursued a degree in something a little less soft than 'Applied Economics' you might understand the Math you're blindly applying.