See Inside
Overreliance on financial software crafted by physics and math PhDs helped to precipitate the Wall Street collapse
Note: This article was originally printed with the title, "After the Crash".
This article was originally published with the title After the Crash.
See what we're tweeting about
scicurious Do we drink too much caffeine? http://t.co/IQlREQ67Lj NO! NO!! YOU CAN'T TAKE IT AWAY FROM ME! *clutches mug*
notscientific How noise pollution can potentially lead to marine mammals having "the bends." By @SaraMynott http://t.co/QHybUmqwXi #scitable
BoraZ Illusion of the week: It Kind Of Looks Like a Building http://t.co/A2ieqe8f3z by @illusionchasers at #SciAmBlogs
Deadline: Jul 15 2013
Reward: $5,000 USD
SciBX: Science-Business eXchange, a joint publication from the makers
Deadline: Jun 30 2013
Reward: $1,000,000 USD
This is a Reduction-to-Practice Challenge that requires written documentation and&
Powered By: 
YES! Send me a free issue of Scientific American with no obligation to continue the subscription. If I like it, I will be billed for the one-year subscription.
38 Comments
Add CommentAren't these economic "quants" using the same kinds of methods "quants" in other fields use to predict weather patterns? Long term and short term?
Reply | Report Abuse | Link to thisMy model predicted a problem in commodities market, housing, market, EFT's, spiders, etc. starting in early 2007. I used the same methods as MAD was developed on.
Reply | Report Abuse | Link to thisshhh bliswell.... you're ruining the punchline...
Reply | Report Abuse | Link to thisWhat I'm more waiting on is admittance on how these models also hindered while it was down turning. I mean, earlier this year, we had many cycles of the economy trying to rebound, until the analysts started to look at their models and creating panic.
I do not think the "Quants" in econ are using the same modeling methods, I would not even try. Weather is science and econ has to include emotion and irrational prediction. This appears to be backed up by the research papers of the modeling techniques these two groups are publishing.
Reply | Report Abuse | Link to thisYou have to remember Phd stands for "Piled Higher and Deeper". I have yet to heard a Phd admit to being wrong or making a mistake.
Reply | Report Abuse | Link to thisYou must not know any Ph.D.s in an engineering or science field, the ones who are most likely to be modeling the environment and not usually the stock market. They are a relatively modest, if sometimes narcissistic bunch, and are the first to admit and sometimes even recognize their own mistakes.
Reply | Report Abuse | Link to thisTo try now and correlate the meltdown with modeling climate change is simply puerile. The analogies that investors use when the markets are up, about investors being sheep or lemmings, are infinitely more applicable when things are spiraling downwards. Even lions turn sheepish and follow the herds then, increasing the negative impact of the downturn. The stock market is created by people, for people and controlled by the actions of people and/or software written by people. It would be hard to make the same case for the environment, especially that part of the environment that existed before people appeared. If you think there's some analogy here I'd be interested in seeing your model of the stock market before 1900.
Didn't anyone learn anything 21 years ago when DJIA dropped about 20% in a single day due to over-reliance on computers? Sure, at the time it was automated trading, but the bottom line is that when you trust your finances on a computer you'll eventually get bit in the ass.
Reply | Report Abuse | Link to thisActually I am an engineer and have worked with many PHd's. I wrote my first model in 1979 on a home brew computer. Did me first work in environmental sciences in 1973. I would not even try to use the same modeling methods for the stock market and the environment. For similar reasons I would not bother with market data before 1960. Tim Masters wrote some interesting material and some of his concepts can be used in non mainstream modeling methods.
Reply | Report Abuse | Link to thisThese quants are trying to do social (probability) science using hard science tools. Jeez, even physicists are having to learn quantum possibilities. It's inductive logic that's needed, not deductive.
Reply | Report Abuse | Link to this"Fairy Tale Economic Models - I"
Reply | Report Abuse | Link to thisThe 'quants modeling approach' that led to the great economic crash (which is still happening) is just a fairy tale designed to inveigle people into believing they can really 'get rich fast' forever as though by magic. In this context, a more realistic approach to modeling would involve building from a basis such as articulated by John N. Warfield, from whom I quote:
The Structural Modeling Approach – and how it is significantly different from any conventional approach
Modeling is a process that begins with human perception. A sequence of the following nature describes the activity of modeling:
1) Perception
2) Storage in the brain
3) Identifying a context within which to place the perceptions, and within which they can potentially be integrated
4) Generating factors associated with that context and with the perceptions that are the focus of attention at the time
5) Identifying types of relations that appear to be associated with these factors in the chosen context
6) Structuring the factors to show how they are interrelated through specific relationships that are representative of the selected types
7) Interpreting the structures produced
8) Associating the factors with algorithms that permit the relationships discovered to be quantified (if they are possible to quantify)
9) Assigning or computing numerical values to/for the factors
10) Interpreting the model-related information for purposes of design or decision-making
(Paraphrased from “Structural Thinking”, J.N. Warfield: 1995-96 Essays on Complexity)
The above sequence describes Structural Modeling. Built into this process is the certainty that we are modeling the real world when we get to our numbers, which is something apparently forgotten by the the quants whiz-kid modelers who drove the world’s financial system to ruin with their fairy-tale models. Robert M. Pirsig, in “Zen and the Art of Motorcycle Maintenance” has some apt ideas that describe the recent (and I guess current as well) economic models that drive much of our modern world. (Follows later, as space is running out here).
"Fairy Tale Economic Modeling - II"
Reply | Report Abuse | Link to thisThe apt quotation from Robert M. Pirsig:
“To speak of certain government and establishment institutions as ‘the system’ is to speak correctly, since these organizations are founded upon the same structural conceptual relationships as a motorcycle. They are sustained by structural relationships even when they have lost all other meaning and purpose. People arrive at a factory and perform a totally meaningless task from eight to five without question because the structure demands that it be that way. There's no villain, no ‘mean guy’ who wants them to live meaningless lives, it's just that the structure, the system demands it and no one is willing to take on the formidable task of changing the structure just because it is meaningless.
”But to tear down a factory or to revolt against a government or to avoid repair of a motorcycle because it is a system is to attack effects rather than causes; and as long as the attack is upon effects only, no change is possible. The true system, the real system, is our present construction of systematic thought itself, rationality itself, and if a factory is torn down but the rationality which produced it is left standing, then that rationality will simply produce another factory. If a revolution destroys a systematic government, but the systematic patterns of thought that produced that government are left intact, then those patterns will repeat themselves in the succeeding government. There's so much talk about the system. And so little understanding.”
It be a lot of work, but it's really not an impossible task, to rebuild our financial and economic structures to be based on foundational realities. More information about Warfield's seminal contributions to systems science are available from http://www.jnwarfield.com and from the "John N. Warfield Collection" of books, papers and presentations held at the library of George Mason University (check out: http://ead.lib.virginia.edu/vivaead/published/gmu/vifgm00008.tp).
Based on Warfield's contributions, there has developed a powerful aid to problem solving and decision making that I call the 'One Page Management System' (OPMS), that enables (individual and group) users to choose a Mission - any Mission - and, from their own current ideas, to develop effective Action Planning to accomplish that Mission. Economic modelers could, for instance work on a Mission: "To create more realistic models on which to base our economic and financial systems".
"Fairy Tale Economic Modeling - II"
Reply | Report Abuse | Link to thisThe apt quotation from Robert M. Pirsig:
“To speak of certain government and establishment institutions as ‘the system’ is to speak correctly, since these organizations are founded upon the same structural conceptual relationships as a motorcycle. They are sustained by structural relationships even when they have lost all other meaning and purpose. People arrive at a factory and perform a totally meaningless task from eight to five without question because the structure demands that it be that way. There's no villain, no ‘mean guy’ who wants them to live meaningless lives, it's just that the structure, the system demands it and no one is willing to take on the formidable task of changing the structure just because it is meaningless.
”But to tear down a factory or to revolt against a government or to avoid repair of a motorcycle because it is a system is to attack effects rather than causes; and as long as the attack is upon effects only, no change is possible. The true system, the real system, is our present construction of systematic thought itself, rationality itself, and if a factory is torn down but the rationality which produced it is left standing, then that rationality will simply produce another factory. If a revolution destroys a systematic government, but the systematic patterns of thought that produced that government are left intact, then those patterns will repeat themselves in the succeeding government. There's so much talk about the system. And so little understanding.”
It be a lot of work, but it's really not an impossible task, to rebuild our financial and economic structures to be based on foundational realities. More information about Warfield's seminal contributions to systems science are available from http://www.jnwarfield.com and from the "John N. Warfield Collection" of books, papers and presentations held at the library of George Mason University (check out: http://ead.lib.virginia.edu/vivaead/published/gmu/vifgm00008.tp).
Based on Warfield's contributions, there has developed a powerful aid to problem solving and decision making that I call the 'One Page Management System' (OPMS), that enables (individual and group) users to choose a Mission - any Mission - and, from their own current ideas, to develop effective Action Planning to accomplish that Mission. Economic modelers could, for instance work on a Mission: "To create more realistic models on which to base our economic and financial systems".
A model is only a representation which captures a certain essence of the subject matter! It is never the same thing. Like in copy cat science fiction serials, if you put the cart before the horse, this is what you get
Reply | Report Abuse | Link to thisI think that it is an unbelievable stretch to blame the mathematicians and physicists for this. All they have done has been to investigate an interesting problem and develop models based on their research. The use to which the models are put is the responsibility first of executives making the decisions, and second of the SEC in allowing them to do so. After all, if a company builds an airplane, and an untrained individual attempts to fly it and crashes, it is obviously not the companies fault; it is first the stupidity of that person, and second that of the airport that allowed him to do so.
Reply | Report Abuse | Link to thisWhen viewing any models mathematical or otherwise derived from scientific study one should always remember that they can be wrong.
Here's a sign of the financial and moral hazard decline of the times. Why oh why does nobody question the world astounding stupidity of computer generated "credit scores"? The formula is kept secret, it obviously does not work, and it has totally replaced the judgment of human loans officers. It has become a way to manipulate unfairly the the lives of consumers, but it also greased the wheel of bogus lending, mortgages and otherwise, more than any other factor. It help to make all mortgages defacto subprime. So if computers can't be trusted along, more accurately the complexity of human life can not be reduced to an algorithm, why does this corrupt and broken mechanism go unquestioned? Maybe it is the "Hoover" effect...question it and find out that really reduces your scores. We need transparency in the financial system and this core mechanism is anything but. The consumer rating agencies are broken as badly as the bond rating ones. And, gee, have even less regulation despite their enormous power, including the power to do ill for the rest of society.
Reply | Report Abuse | Link to thisSoftware models are based on mathematical models. Isn't it silly to blame the software? Didn't the maths people verify and test the software first?
Reply | Report Abuse | Link to thisMaybe these Quants should all be issued with a copy of Nassim Nicholas Taleb's The Black Swan, on day one of Quant Basics 101.
Reply | Report Abuse | Link to thisIt reminds me of Jurassic Park, when the mathematician started a rant about complex systems and how increasingly difficult they were to control, we know the dinos escaped from the park, now let's see how we do dodging the "rent o saurus", the "profiteersaururs rex" and so on...
Reply | Report Abuse | Link to thisDOH!!
Reply | Report Abuse | Link to thisDid anyone ever read the book, Tulipmania about the 1600 investment foolishness? Do I ever bet on horse races? Not unless a certain friend says that his Mob buddies gave him a tip. Just like that type of gambling, the stockmarket is for the large foundations only. Rockefellers, Heinz, etc. families put everyone's money in a huge conglomerate. They have money to spare, dudes with 401K's and Pension Funds should not invest in the Markets. Unions should do very mild CD's that are guaranteed. Less risk is less return but anyone who makes less than $200,000 a year should invest in a different type of stock, i.e., the same ones that exude the power spoken of by Chairman Mao. And who now owns most of the Capitalist Banks being bailed out??? We gave them enough rope and indeed we have been HUNG.
Reply | Report Abuse | Link to thisPhysicists do not know what they are doing. They confuse things that are mathematical with those that are physical and get away with it because nobody understands their babble. Read the books on string theory by Lee Smolin and Peter Woit.
Reply | Report Abuse | Link to thisMuch of mathematics is meaningless. Read "Constructive Analysis" by Bishop and Bridges.
This is the result of the fact that the only people permitted to evaluate math and physics are mathematicians and physicists. Government funding agencies have no idea what they are getting for the taxpayers' money and university administrators haven't a clue as to what any of their professors are doing other than filling up the journals with unintelligible prose.
The stock market crashed recently because the Fed kept money supply too low, leveraging ratios were allowed to climb to 40 to 1, and the surplus money from the foreign trade imbalance had no where to go but into the stock and bond markets. Since bonds were paying low interest rates, most investors opted for buying stocks.
Reply | Report Abuse | Link to thisUsing entropy analysis as defined by Boltzmann, Shannon, and Rost, the equilibrium states of economies can be modeled (Econolibrium Table). Actual economic performance can be compared to econolibrium states and the likelihood of a market advance or decline can be predicted with accuracy. Consult www.llbis.blog spot.com for details.
I have
Reply | Report Abuse | Link to thisIt seems like there is a new villain every week for the crash. I would like to find out more about the SEC meeting referred to in the article.
Reply | Report Abuse | Link to this"There investment bankers from the largest institutions pleaded successfully with Securities and Exchange Commission (SEC) officials during a short meeting in 2004 to lift a rule specifying debt limits and capital reserves needed for a rainy day. This decision, a real event described in the New York Times, freed billions to invest in complex mortgage-backed securities and derivatives that helped to bring about the financial meltdown in September."
Thanks,
The recent stock market crash can be attributed to low money supply rates, free trade imbalances that drove surplus investment funds into the stock markets, and extraordinary leveraging that at times reached 40 to 1 ratios. The best computer programs can do for investors is to assess the status of the economy and and the magnitude and direction of its subsequent change. Consult www.llbis.blogspot.com for details.
Reply | Report Abuse | Link to thisInvestors don't model the markets with computer programs. They just think they do. I have tried for 30 years with modest success and I have 54 years of computer and modeling experience. The markets crashed because of poor government economic policies that totally distorted our economy's equilibrium state. If you want to follow the economy and the markets, try the econolibrium table that models the various equilibrium states of the economy.
Reply | Report Abuse | Link to thisAs someone who works in finance, I have to say that the role of the average quant is a bit narrower than this article implies.
Reply | Report Abuse | Link to thisA quant's job is to develop models showing the sensitivity of a financial product to various inputs; in effect, to compute the first, second and various cross-derivatives. Computing the theoretical value is a by-product. If a firm is hedged (that is, if they buy or sell in the market to offset these derivatives) then they should not need to care about the model's theoretical value.
That said, some firms will keep complex products, such as mortgage and asset-backed securities, on their balance sheets unhedged. These firms have been hurt, deeply in the current crisis. Their models were not wholly to blame, however. The problem lay with their cavalier attitude toward risk. Their models should have told them that they were exposed to increases in defaults, and especially increases in correlated defaults.
To have quants develop models incorporating behavioural elements and/or other speculative forecasting methods may be interesting, but is not the same as the very important role of assessing risk.
As most physical modellers know, it is necessary to verify the model and ensure the implicit assumptions in the model apply to the case under study before placing any reliance on the model.
Reply | Report Abuse | Link to thisWhen we are dealing with humans and their reactions we are not sure what the assumptions really are or should be and in most cases the real world situation is different and far more complex than the model. Our ability to react differently based on our past experience means that we have to account for that past experience and our collective learning process - not a simple task.
The reliance on untested models is the folly but that doesn't mean one shouldn't try and develop models. After all that is how we learn. We build models, test them and only rely on the ones that work. we also don't learn by doing everything right the first time. That we leave for for those who suffer under the delusion that they are gods. I prefer feet of clay
As a developer of spatial data modeling software, I always tell the users to know their data, know the model, and always have ball-park estimates of intermediate and final results. The software spares you the calculations, but not the thinking.
Reply | Report Abuse | Link to thisI was shocked by traders and bankers coming to a dead stop when the market went into serious decline. Obviously, they also didn't comprehend what was happening before then.
Just some things to consider:
Reply | Report Abuse | Link to thisThe United States of America dollar has no real value and hasn't since the value of the dollar was changed from being based on a real asset to being based on the theoretical value of the productivity of the average American worker.
Most years the price of the NYSE, Amex and NASDAQ increase at a rate that exceeds the rate of growth of the American economy. Sometimes by rates of more than double.
Real estate value grows based on population increase vs available land/housing decrease. Real estate pricing changes based on percieved profitability, which has little relation to the growth of value.
Someone set us up the bomb!
All your base are belong to us.
I think this whole thing proves economics and markets are a pseudo science.
Reply | Report Abuse | Link to thisI have some perspective on this as technologist who worked in the pit where much of these instruments were born at J.P. Morgan in the late 90s.
The models were based on a narrow band view based on immediate circumstances done by people with too much self confidence.
The whole thing proves economics and markets are a pseudo science.
Reply | Report Abuse | Link to thisI was at ground zero, J.P. Morgan, working at a technologist on the trading floor when the market for this stuff took off.
The underlying models reflected narrow band thinking based on immediate circumstances generated by the highly educated and overly confident.
This is ridiculous. Any model has conditions where it will fail. It is th responsibility of those who use the models to know when those conditions might be occuring, and when to abort. The fact that the writers of this article advocate "better" models from behavioral economics reveals their ignorance. If you change and "improve" the model, you've only changed the circumstances under which it will fail, but the fundamental problem still exists.
Reply | Report Abuse | Link to thisAnything in finance, a quantitative model, or a non-quantitative strategy will fail under the right circumstances, improving models does not fix the problem. Proper use is the only thing that can.
This article is ridiculous. Any model has conditions under which it will fail. It is up to the user to know when those conditions might be occuring, and when to abort. The fact that the authors recommend better models from behaioral economics reveal their ignorance. Whether that would really make a model better or not, Having better models doesn't fix the fundamental problem, because all models, regardless of how good they are will fail under the right circumstances.
Reply | Report Abuse | Link to thisMy rule of thumb when dealing with computers is that while Artificial Intelligence has not been properly realized, Artificial Stupidity has long been a part of computing.
Reply | Report Abuse | Link to thisI think quoting Robert Pirsig is rather problematic. The truth is, we're not limited by our cleverness, but by our ability to process a gobstopping amount of information. Sometimes it does help to step back and reconsider the situation and the situation that lead to it. He's right in that a system can take on a life of its own, beyond the necessity that should fuel it. But we should be under no illusion that society can run smoothly without certain bureaucratic elements at work.
The real problem, I think, is that they built a system that could not even model a known reality, that is the effects of a severe market downturn, much less warn people about when the risk of that might be approaching.
If you want your information to serve you well, you must be prepared to percieve and deal with uncomfortable information.
All models have points of failure. The trick is to adapt them to fail least in the environment they're employed most.
Reply | Report Abuse | Link to thisBut more to the point, these models were put together with the stunning inability to model a known, all too likely scenario in real world markets: sudden, emergent, irrational panics. To not even have that possibility considered is like having a screen door on a submarine. To call it a design flaw understates the problem.
A bit of criminality at an important portal was the collusion between bond rating companies and the bond issuers. If these junkiest of bonds had been rated correctly, the whole housing bubble would have been stopped early on. The bond rating companies shouldn't be paid directly by the company requesting the rating.
Reply | Report Abuse | Link to thisThe quants seemed incredibly naive expecting the lenders, packagers, bond raters and bond salesmen they were servicing to act like immutable incorruptible laws of nature. They overlooked one of the immutable laws of human nature: if people of easy morals can find a way to make a crooked buck, they'll exploit it.
The reason why they were unable to forecast the crash of the stock market was because the models used do not take into account the cause of the crisis. This is due to the speculation in land values. None of the models used are sufficiently broard in their concept as to include the whole of the macroeconomics system including the effect of the price of land.
Reply | Report Abuse | Link to thisThe Georgist theory of business cycles allows for this vital factor on the behavour. Georgist economists like Fred Foldvery were forecasting the crisis at least 2 years befor it occurred. Its not the maths that is wrong but the models that are not representative of the whole shebang!
REF: "Progress and Poverty" Henry George 1879. see also:
http://www.progress.org