Fortunately, academics are now starting to get the idea that there is something wrong with "classical" Modern Portfolio Theory, and this has only occurred in the mid to late 1990s. Professors Eugene Fama and Ken French have for many years worked on looking for a replacement to the capital asset pricing model, in particular getting rid of beta. Fama and French studied stock returns from 1926-1997 and concluded that while they still think the market is efficient, volatility doesn't explain much.

The Fama/French Three-Factor Model was based on their survey of stocks, using statistical regression techniques they found a neat model that managed to explain market returns quite nicely. While they are still struggling with identifying the hidden risks that explain these anomalies, an investor can influence his or her return by varying three factors:

1. How much stock market exposure to accept

2. The size ranking of the companies bought

3. The book to market ratio

All investors can earn a risk free rate by investing in cash, if they want higher returns they need to invest in stocks. The first factor in influencing returns is obviously if you want to take higher risk you can get a higher return by buying stocks instead of cash.

The second two factors were the really shocking ones, the second factor is the "size premium", the extra performance you get by buying small companies. As David Dreman points out in chapter 15 of Contrarian Investment Strategies: The Next Generation , the original studies into the size effect were seriously flawed, and that most of the returns from small stocks came in just a few episodes. Using better data, and without the flaws in identifying "small stocks" that Dreman found in his criticism of the first studies, Fama and French reestablished the small company effect, and found that it amounted to 0.20% per month over the 1926-1997 period.

In answer to Dreman's criticism saying that the small company effect is false and pointing out the superior performance of the S&P500 in the time since Dimensional Fund Advisors opened its small company fund, since the US bear market began, the S&P500 has lost much of its outperformance, and small company stocks have actually outperformed in the 7 years since 1995. While the small company premium does assert itself only in spurts, it does seem to exist.

What really bugs the academics is that Fama and French found that value stocks, which they define as stocks with a high book to market ratio (inverse of price to book ratio, to avoid divide by zero errors), also outperform, and that this outperformance is not related to higher volatility.

The value premium appears to be very robust, and has been confirmed in various studies on various markets by various researchers. This premium, which does fluctuate over time, in the 1926-1997 period has been larger than the size premium, at 0.43% per month.

The reason why this premium bugs Fama and French is that value stocks are actually less volatile than the general market, which is a kick in the guts for the idea that risk=volatility and that higher return=higher risk. These two remain dyed in the wool Efficient Market Hypothesis believers, so they have sought to explain the value premium by stating that value companies are inherently higher risk due to their perceived chances of going broke. There may be some truth in this, many of the cheapest companies are quite sick, though this doesn't translate into price volatility. The researchers argue that volatility is not an effective catch all for risk analysis and that fundamental factors do play a role.

They don't, however, explain with any great authority why the most expensive stocks, which are usually the most volatile, earn so much less than value stocks. Value investors claim that "growth" stocks are overpriced because people pay a gambling premium on them, and "value" stocks are underpriced because they are boring. This is where the Efficient Market Hypothesis breaks down, small growth companies are by far the worst performing asset class, and are extremely risky (and volatile). Value investors have been saying this for years, but now finally academia has caught up.

(Though it does annoy me that beta and the CAPM are still taught in many finance courses, despite them being discredited and abandoned as theories by academia as well as industry. As an example, beta is taught in courses like some versions of the Diploma of Financial Planning).

The three factor model has a variety of uses. First of all it tells us that if you want to "beat the market" you can tilt your portfolio away from the standard large company indexes and buy value stocks and small company stocks (value and small company index funds are finally available in Australia, though they've been around in the US for over 20 years).

Secondly, it allows more efficient study of fund manager performance. No longer are academics comparing all funds vs the one benchmark, these days statistical tests of manager value adding take into account the performance of index funds closer to the fund manager's strategy. This has provided a new dimension in analysis, and you may have noticed that ratings houses like Morningstar these days have started to classify funds in a 3x3 square grid of value/balanced/growth and small/med/large companies. This is no doubt a far more efficient method of classifying a fund than the arbitrary old classifications like "income and growth".

Where Efficient Market Hypothesis possibly still lags behind is in their definition of risk with value companies. I tend to agree with Buffett on this one, if the price of a business falls by half it becomes twice as desirable, not twice as risky. People clinging to EMH assert that if a stock falls there must be a damn good reason for it doing so, because the market is efficient and thus probably right.

Actually, EMH does not necessarily say that the market is always right, it does not contradict the proposition that stocks are sometimes (or often) "incorrectly" priced, the theory just says that investors are unlikely to make any profits from these errors because they will be hard to identify at the time. Sometimes it is easy to identify mispricings (few people really believed that Internet stocks were rationally priced in the late 90s, but as nobody knew when the bubble would end it was too risky to rush out and start shorting NASDAQ futures).

All of the debates over the origin of the value premium will keep academics going for at least another decade or two, but in a practical sense investors need not worry too much. The Three-Factor model probably is true, whatever its explanation, investors need not worry too much on the finer points behind it, when constructing your portfolio if you want higher returns you can buy value and small company stocks to improve your long term returns.

I've written a quite detailed article about Fama and French's work, and other claimed sources of higher returns such as momentum and emerging markets at my company website.  If you are interested in this subject, read this: http://www.aifa.com.au/portfolio-construction/investing-for-higher-returns.html