Efficient Market Theory holds that because all investors are rational, this requires that they use all the information they possess in making purchases or trades. Thus, if information exists in the market, it will be reflected in the price. A corollary of this is that only unexpected information can bring about true changes in the market: by the time an expected change happens, the market will have adjusted. This can be seen every time the Fed changes interest rates. Most investors have a reasonable expectation as to the action they will take, and so adjust their behavior to compensate well beforehand, making the effect of the rate change itself virtually nil.
Certain behaviors follow from this theory: first, that since you can’t outperform the market in the long term, buying an index fund and holding it forever is the best investment one can make in the long term. By pegging your earnings to the average growth of the stock market (the S&P 500, for example), one doesn’t incur fees by trading around, and trading around wouldn’t do much long term good anyway except as a function of luck, meaning you’re just as good holding onto it as putting it somewhere else. You cease to collect information because you trust the information that the stock price gives you
There is strong evidence that certain variations of the theory very closely correspond to actual markets in the US and UK. On the other hand, there are certain investors who can consistently outperform the market:
These investors’ strategies are to a large extent based on identifying markets where prices do not accurately reflect the available information, in direct contradiction to the efficient-market hypothesis which explicitly implies that no such opportunities exist.
If all available information is contained in a stock price (and it is), then how do people like Warren Buffet and George Soros exist and consistently profit from the stock market?
The answer lies in trust, and weighted information. EMH requires that all investors use all their knowledge, but it does not require that they know everything, nor that different investors know the same things. Consider a horse race: every better knows the same information about the horses, and the normal distribution of decisions based on risk is perfectly adhered to. Let’s say the odds on one spectacular horse are 70-30 in its favor based on the collective and informed bets of everyone currently there.
Now imagine the crowd doubles due to an influx of newbies. Each one sees the odds on the screen, trusts the wisdom of the crowd before them, and bets on this horse. Suddenly the odds are 85-15 in its favor. Did anything in the horse inherently change? No. Was any information lost or changed in the marketplace? No; the information was simply disproportionately weighted. Suddenly the underdog horse is a much better buy – you get an 85% return (not counting what the establishment skims off the top) for what is still essentially a 70% risk. Nevertheless, you would not know this from trusting the board.
Similarly in stock markets, people have a propensity to buy stocks following someone else’s lead. If Soros picks up a lot of stock in a particular company, people trust that and buy stock themselves, giving himself a short-term boost in the manner of a self-fulfilling prophecy. If Buffet sees this behavior and owns some of that stock himself, he can note that Soros’ information is overweighted in the market, causing a jump in stock price despite no underlying change in the stock’s business itself. By riding this wave and selling at the peak of it, watching the behavior of investors deemed trustworthy (especially big ones like Soros and Buffet) and bidding opposite to them (after a short delay to allow their effects to be felt first – for example, buying a stock newly cheapened after Buffet and followers unload it) is a way to, counter to strict EMH, consistently outperform the market.
In the case of the horse betting example, the reaction of the informed would be quick enough that they would trade their bets and shortly balance out the discrepancy, making the market effectively efficient for all but the quickest and most observant. But that requires that there be quick and observant people that benefit from this. Furthermore, the ability of one in a stock market to reap the returns on his investment by selling at any point he chooses makes it possible for the quick and observant to ride these waves, rather than waiting for the final result and payout, by which time the market will have corrected itself. If there were a substantial secondary market for bets – meaning bets could be bought and sold from peers based on the current odds, with the expectation of a payout on the order of those current odds before they level out – then one could conceivably make a consistent profit on the horse racing tracks just as in the markets by riding waves of misweighted information.
The catch is, of course, that if everyone catches onto this technique, there is no more trust as a heuristic – nobody follows anyone else’s stock picks – so all information is weighted equally, and EMH really holds true in its purest form. Conversely if everyone trusted the market and resigned themselves to matching its growth by investing in index funds, the market wouldn’t have any information to reflect. Stock prices are set by the people who trade, which according to EMH is suboptimal behavior. This is the paradox of EMH: it doesn’t hold true unless no one believes in it – and acceptance of it invalidates it.