One of the main tenets of the information-based Efficient Market Hypothesis (info-EMH) is that prices adjust instantaneously to reflect all the available information (to differing degrees depending on the weak, semi-strong, or strong form). This is a thesis that directly contradicts the core assumption of Technical Analysis (TA), which states that prices adjust gradually, though not necessarily to the theoretically correct value, to reflect information of the market and the sentiment of the market participants. Adherents to TA believe that it’s possible to use a variety of short-term trading strategies to exploit this gradual price movement, something deemed impossible under the info-EMH.
One slant in thinking of this to further clarify thinking and judge the merit of these views is to consider that prices are a function of the beliefs of market participants, whether they are rational and based on fact and proven theory, or irrational and based on whims and unsubstantiated belief. This would naturally lend itself to interpreting TA as applied social psychology.
One example used to illustrate this point is to ask someone to look at the price movement of two stocks, one on a continually increasing trajectory and the other in a continually decreasing trajectory, and then ask, which one would you rather own? Of course, most people would want the one with the continually increasing stock price, thus revealing the intuition of the TA viewpoint. In contrast, the info-EMH would state that the auto-correlation of the stock prices is exactly zero, meaning that there is no sense in looking at the previous history of each of the stock prices since you can’t use it to predict the future.
My belief is that reality reflects both views, depending on the circumstances. For stocks of companies with market capitalization in the billions of dollars, that are highly traded and researched, price will almost instantaneously reflect available information, as per the info-EMH, since any mispricings are arbitraged away by trading firms, hedge funds, and other institutions of the sort. (There exist other “technicalities” such as past price history [weak form info-EMH], all available public information such as analyst reports and financial statement information [semi-weak form info-EMH], and all available public “and” private information [strong form], but let’s just take this to mean: “Prices reflect information.”) On the other hand, if you are in an emerging market where laws are less strict or even not yet formed, the nature of companies are not known, and the volume of stocks traded is not so high, then perhaps the value of the stock is more of a barometer of investor sentiment than it is with mathematically/theoretically correct prices. In this case, many tools in TA analysis would be valid (for example, looking at the cash reserves in brokerage accounts…if there is a lot of cash available in those accounts, then people are probably scared of the market…in which case you can use contrary-opinion rules to act opposite of what people think.).
So perhaps its simply a matter of understanding the domains of validity of each of these viewpoints and modify our thinking accordingly.
The question then shifts to how useful social psychology is. Is it mathematically precise like the Black-Scholes option pricing model, for which an Economics Nobel was awarded? Of course not, but more importantly does it matter if it is so mathematically precise to n decimal places? If theories in social psychology are very coarse-grained caricatures of reality, yet still have some truth to them, then perhaps this is really enough, and consequently that technical analysis is a “correct” way of viewing the market. Although it might not be possible to distinguish the value of some indicator or parameter as 45.789 or 45.788 in social psychology, it seems that being able to distinguish between “a lot of the time” and “not too much of the time” still has much important predictive value to it. Practically, this just means to make sure the amount of your bet (which is basically what a short-term trading or long-term investing decision is, despite their different names) is larger when your technical analysis indicator states “a lot of the time” then when it states “not too much the time.”
In the end, it’s not whether this theory or that theory is correct, but what approximation of correctness is needed for a model to be built that allows you to be correct even just a tiny bit more of the time than when you are incorrect. Furthermore, this changes as a function of time and the relative sophistication of the analysis and actions of other market participants.