I'm wondering if the EMH and geometric brownian motion is still the best model for financial market behavior? Note, I'm not asking if it is a perfect model, since most of us know the weakness of this theory. I'm no academician in this field, so I'm not up to date on the latest market theories. I know activities like insider trading and front running are baked into price movements, but those aren't of themselves, a market model. If not EMH/GBM, what are some other competing theories of market behavior which can account for these more questionable behaviors? I post this in TA so I can continue the long and tiresome discussion on whether TA is bullshit. I'm becoming more of a believer that it is.
The real question is, what strategy does your model lead to? Seems to me EMH just leads to coin flips, which is unacceptable for trading.
Behavioural economics has a lot of ideas that could be a bit more interesting than EMH. The most famous one is "prospect theory" of Kahneman and Tversky.
Can you recommend a Cliff Notes version? This can't possibly be Econ 101, even at CalTech. http://people.hss.caltech.edu/~camerer/Ec101/ProspectTheory.pdf
Ha, for a general guide to Kahneman's life's work, expressed in everyday terms, you should read his excellent book "Thinking, Fast and Slow".
I think it depends on what you want the model for. Is it do risk management? Then a combination of EMH, brownian motion, if used correctly isn't a bad model and has the advantage of simplicity. Is it to predict the market? Well these models tell you can't predict the market, and shouldn't try. Behavioural finance isn't yet, and perhaps never will be, a consistent complete theory like EMH; but it has the advantage of being able to explain a lot of patterns in the market, so I am a big fan of it. Prospect theory, which explains why trend following works, is probably the kernel of this complete theory. A second vote for "Thinking fast and slow". If you are short of time there is an older book, "Beyond greed and fear" Hersh Shefrin, which is brief and very finance focused. If you're really short of time here's a quote from the chapter in a book I'm writing at the moment.
I know the book was a best seller and recommended by a lot of traders. I was scratching my head when the book description had nothing to do about trading or even economics in general. I'll check it out. Thanks.
It has to do with how people, broadly, make decisions... So it has a LOT of relevance to trading, really.
You need to define "best model". Then there are various questions that you can ask: given a finite realization from a GBM is it possible that a real instrument could have a "similar" (to be defined precisely) realization? The answer is yes. Or if you have a real price trajectory, could I find a GBM which is "likely" to generate something "similar". Of course you can. If you have a finite real price trajectory can you say that it does not exist a GBM which could have generated it ? You can't. In very practical and intuitive terms, the GBM is, imho, highly "unrealistic" for several reasons. Simply being familiar with both simulated data and real trading will make you convinced that the realizations of a pure GBM, observed in a large number, "look and feel" in general quite "unrealistic". In the real world, there is no such a thing like constant volatility or constant drift (as in the GBM). GBM realizations tend to "reverse" much less than most instruments do (especially energy/commodities related). Also they tend to reach "farther" than real instruments do. Of course there are ways to modify or combine GBMs and other generators to make a simulation "feel" a bit more "realistic" (at least to an experienced trader). GBM is just a model that has no many other justifications than being relatively "simple" (say, formally tractable). For the rest, the real world has no obligation to even come close to these human fantasies and simplistic models. Actually, it has no obligation to satisfy any "model" one can think of. >TA is bullshit. I'm becoming more of a believer that it is It's just guessing, based on visual hints and optical illusions. There is nothing "technical" about it, apart the use of the pompous adjective to somehow disguise the guessing activity as something justified by some more "legitimate" reason. Whatever you guess, sometimes it will come true, sometimes it will not, and you can actually "believe" whatever you like, as there is no way to confirm or confute a causality connection. And when things are not even falsifiable, some would argue we are out of the domain of "science". (Check out: http://en.wikipedia.org/wiki/Falsifiability#Falsificationism)
To be sure I won't devote much time to what is certain to be an endless discussion, but here we go for a start. These guys found there was 'some value' within a certain time period. Whether that 'value' can be monetized and if that time period phenomenon still persists I leave to others. I am only addressing a request for continuation (quote 1) and a hint at "non technicality" and falsification (quote 2). ******************************************************************* Foundations of Technical Analysis:Computational Algorithms, Statistical Inference, and Empirical Implementation -ANDREW W. LO, HARRY MAMAYSKY, AND JIANG WANG THE JOURNAL OF FINANCE • VOL. LV, NO. 4 • AUGUST 2000 ABSTRACT Technical analysis, also known as “charting,” has been a part of financial practice for many decades, but this discipline has not received the same level of academic scrutiny and acceptance as more traditional approaches such as fundamental analysis. One of the main obstacles is the highly subjective nature of technical analysis— the presence of geometric shapes in historical price charts is often in the eyes of the beholder. In this paper, we propose a systematic and automatic approach to technical pattern recognition using nonparametric kernel regression, and we apply this method to a large number of U.S. stocks from 1962 to 1996 to evaluate the effectiveness of technical analysis. By comparing the unconditional empirical distribution of daily stock returns to the conditional distribution—conditioned on specific technical indicators such as head-and-shoulders or double-bottoms—we find that over the 31-year sample period, several technical indicators do provide incremental information and may have some practical value