Wednesday, September 26, 2012

Efficient Market Hypothesis

In simple terms, the efficient market hypothesis says that there is no better measure of the value of a stock than its market price. In his 1988 letter to shareholders, Warren Buffett ridiculed academics who cling to efficient market theory in the face of decades of market beating returns by Buffett and his mentors, saying “apparently, a reluctance to recant, and thereby to demystify the priesthood, is not limited to theologians.” This debate rages on with one side insisting markets are efficient or nearly so, and the other side dismissing efficient market theory. However, this debate is mostly pointless. On its own, it makes little sense for the stock market to be efficient or not. It can only be efficient with respect to some observer. This means that the market can appear to be efficient to one investor, but not another.

To explain what I mean by market efficiency being dependent on the observer, consider a simple example of a coin toss. To most observers, when the coin reaches its maximum height, the probability that it will come up heads is 50%. However, an observer with a high-speed camera connected to a computer running an algorithm that analyzes the coin’s velocity and spin might be able to determine at the coin’s maximum height that the probability of the coin coming up heads is 55%. So, who is right? Is the probability 50% or 55%? The correct answer is that the probability is different for each observer.

Buffett talks of a stock’s intrinsic value. This is essentially an investor’s estimate of the value of a share. When an investor, such as Buffett, is able to calculate a better estimate of a stock’s value than the market price, he is able to use this information to get market-beating returns. He doesn’t have to do this for all stocks; he just has to be able to do this for a few stocks and to know for which stocks he has a good estimate of intrinsic value.

The vast majority of investors, including me, are not able to produce estimates of a stock’s intrinsic value that are better than the stock’s market price. In a sense, this is saying that we’re not smarter than the collective market of investors. Share prices are a worldwide dollar-weighted consensus of the intrinsic value of stocks. Viewed this way, it isn’t surprising that few investors can calculate a stock’s intrinsic value better than the market can. This doesn’t mean that no investor can have an expectation of beating the market; it only means that this ability must be rare.

So, the market looks efficient to most investors, even if their overconfidence makes them think otherwise. Buffett has proven in the past that the market looked quite inefficient to him. With his huge portfolio and shrinking universe of businesses large enough for him to invest in, my guess is that the market looks more efficient to him with each passing year.


  1. That's a very very good point Michael. I never thought of it that way (despite having my own dissertation on the efficient-market hypothesis) but it makes perfect sense.

  2. @Patrick: Thanks. The dependence of probabilities on a observer was well-known to me, but I just thought of the connection to efficient markets recently.

  3. This consideration is mathematically valid under an assumption that all players "play clean". Which is hardly a probable assumption. General investor population makes decisions based on newspapers and TV, which the likes of Buffet can afford to influence.
    The "big guys" can drive markets in the direction they need through mass media. Likely for "small guys" there are quite a few big guys on the market and, therefore, on average all market is more or less adequately reflects economy. At least we can hope it does. Because otherwise we should all give up and lock funds in GICs.

  4. @AnatoliN: Your comment speaks to why the market may not look efficient to some big players. But this changes little. The market will still look efficient to the thundering herd of most investors.

    I don't agree that the best strategy is to lock in GICs even if there are manipulators out there. GICs are certainly better than day trading, but long-term buy-and-hold investors in broad market index GICs are not harmed much by market manipulation as long as true value eventually wins out.