In 1990, John Wiley & Sons, the second largest publisher of business books, asked me to write the first book dedicated solely to the subject of asset allocation for investment professionals. The book, Dynamic Asset Allocation, was a big seller, having been in print for eleven years. The following are some of my thoughts back in 1990 and they still hold true today.
Most people believe the more stocks a person holds in an equity portfolio, the lower the volatility (risk) of the portfolio because of the advantages of diversification. If that were true, then how come the Dow Jones Industrial Average (consisting of only 30 stocks) is no more volatile than the Standard and Poors 500; and why is the S&P 500 even less volatile than the Wilshire 5000? The answer is: the important thing is the correlation of price changes between the stocks, not the number of stocks in the portfolio.
Imagine a portfolio that consists of only two stocks, both of which are expected to double in ten years. But what if (in the short-term) every time Stock A went up by x-percent, Stock B went down by x-percent and vice versa. The portfolio would have an annualized, expected return of over 7% (plus dividends) yet have no volatility (short-term risk). Even if both stocks are expected to double in ten years, most people still care about short-term risk. This is because not everyone is going to buy a portfolio of stocks and close their eyes for the next ten years, even if they don’t plan on touching the money for a long time. Things usually change. The fundamentals of the stocks in the account can deteriorate or the investors’ level of risk-aversion can increase due to the possibility of having to withdraw money from the portfolio sooner than expected or wanting a decent night’s sleep when the market is gyrating all over the place.
At my firm, Hammer Asset Management LLC (Gilford NH), and at Young Investments (Naples FL and Newport RI), we manage stock portfolios using only 16 to 32 securities. Dick Young likes 32 as a good discipline that extends the benefit of using two or more stocks from the same industry; and I use no more than 32 equities, even for my largest institutional clients. Here is why Dick and I use only 16-32 names (compared to dozens, if not hundreds of securities used by many mutual funds).
When you own a stock, there are two types of risk (reasons for volatility). The first is called systematic risk, caused by economic or sociological changes that affect all stocks, such as a change in interest rates, in the economic outlook, in investor confidence, in taxation, etc. Systematic risk cannot be reduced no matter how many stocks you own, which is why it is called non-diversifiable risk. This type of risk accounts for 50-70% of what makes an individual stock go up or down on any given month, week, day or hour.
The other type of risk is called diversifiable risk, the cause of a stock going up or down based on fundamentals affecting only that particular stock or its industry. For example, if the CEO of XYZ Corp. dies, that incident may have no effect on most other companies; or if the price of steel drops, that would be very detrimental to a steel producer but have almost no effect on a toothpaste manufacturer. It is this diversifiable risk that a portfolio manager can attempt to eliminate. If an investor owned nothing but oil stocks in the portfolio, the diversifiable risk would be high. But, if the portfolio owned stocks each in a totally different line of business, the diversifiable risk would be low.
With today’s computer capabilities, a professional portfolio manager doesn’t even need to look at the industries in which the stocks are involved or how closely their businesses are related to determine the diversifiable risk. The manager can measure the cross-correlation of periodic price changes between any two (or more) securities. The idea is to build a portfolio where the correlation of price changes between all the securities that make up the portfolio is at a minimum.
A portfolio of two stocks that have virtually zero correlation to each other would have half the diversifiable risk of a one stock portfolio. A four stock portfolio, where all four stocks had little relationship to each other, would have half the diversifiable risk of the two stock portfolio, and so on. So, if you own 16 stocks that are uncorrelated, the diversifiable risk would be half-of-half-of-half-of-half of a single stock portfolio. At that point, the investor has eliminated 94% of the non-market risk. In a 32 stock portfolio (assuming all 32 stocks were uncorrelated, which may be an unrealistic assumption in the real world) the diversifiable risk would be next to nothing. But with today’s technology, it is quite possible to find a dozen or more stocks that have little cross-correlation. In addition, the more stocks you own in your portfolio, the greater the likelihood that the portfolio will have an expected return no greater than that of the overall market.
So, why is the Dow Jones 30 less risky than the S&P 500 that, in turn, is less risky than the Wilshire 5000? The answer is, because the indices with the fewer components are actually more diversified than the indices with the greater components.
Therefore, I must conclude: a portfolio manager would be unproductive in trying to be an expert in hundreds of stocks if a 16-32 stock portfolio can have an above-average likelihood of beating the market and simultaneously have below-average risk (volatility).