Contrarian Strategies Within Industries

As weve already documented, investors are entirely too confident of their ability to forecast which stocks will win and which will lose. Trends and fashions in the marketplace play a powerful role in drawing people to popular stocks. Similarly, lack of excitement or lackluster outlooks push investors away from others. Does that extend beyond "best" and "worst" stocks as measured by the contrarian indicators in the last chapter? It should.

If trends and fashions exist in the marketplace as a whole, it is reasonable from a psychological perspective to expect that they exist within specific industries. Analyst research, expert opinion, current prospects, and a host of other variables should work on investor expectations almost identically within industries as in the overall market. The result again will be expectations set too high for favored stocks within an industry and too low for out-of-favor companies. Thus, Merck might be a favorite in the pharmaceutical industry, while Glaxo inight be thought of as a laggard. Siinilarly, the American Intemational Group might be a favorite in the insurance industry, while Ohio Casualty is unloved and unwanted. Ove ricing and unde ricing of favored and unfavored stocks within value and growth industries would thus appear to be a natural extension of contrarian strategies. Sounds great in theory, but unlike our academic friends, were going to ask a mde question: does it work?

To get the answer, in collaboration with Eric Lufkin, I examined the 1,500 largest companies on the Compustat tapes by market size between 1970 and 1996.3 The 1,500 stocks were divided into 44 industiies." The most favored stocks consisted of the 20% of companies in each industry witii the highest P/Es, price-to-cash flow, or price-to-book value ratios, or lowest yields. The most unpopular were the 20% of stocks in each industry that had the lowest ratios by the first three measurements and the highest yield by the fourth. Retums were calculated in the same way for the remaining 60% of stocks in the iniddle quintiles.

Using the industry strategy (and taking low P/E as our example), the lowest multiple in one industry, such as banking, might be 10; in another industry, such as biotech, 25. Yet, if we were right, the lowest P/E stocks in both industries should provide well-above-market retums. In this strategy, we speak of relative P/E-the lowest 20% of P/Es (or price-to-cash flow, price-to-book value, or highest yield) within an industry- versus the lowest absolute P/Es in the entire market, the strategy in the past two chapters.

Figures 9-1 and 9-2 give the results of our study for low price-to-

New, Powerful Contrarian Approach 195 Figure 9-1

Industry-Relative Price/Earnings

Dividends, Appreciation & Total Retums January 1, 1970 - December 31, 1996

2 3 4 High

Relative P/E Quintiles

Market

□ Dividend Return □ Appreciation Total Retum

earnings and low price-to-dividend.* Looking at Figure 9-1, for example, we see the lowest 20% of stocks as measured by price-to-earnings for each of the 44 industries in the first column, the second lowest in the second, the highest in the fifth, and the market in the last column on the right. The lowest price-to-eamings group provides an average retum of 17.7% annually over the 27 years of the study. The highest 20% retums 12.2%, the market 15.3%. The lowest price-to-eamings group outperforms the highest by 5.5% annually over the life of the study.

I find Figure 9-2 particularly interesting. It shows that the total annual retum using the industry-relative high dividend strategy is 17.0%, which is higher than the absolute price-to-dividend strategy retum of 16.1% shown in the last chapter. This is because appreciadon is higher using the industry dividend strategy than it is buying the highest yield-

* Results for low price-to-cash flow and low price-to-book value (not shown) are very similar.

Figure 9-2

Industry-Relative Price/Dividends

Dividends, Appreciation & Total Retums January 1,1970 - December 31, 1996

3 4 High

Relative P/D Quintiles

Maricet

□ Dividend Retum Appreciation Total Retum

ing Stocks available in the overall market (10.4% vs. 8.2%). Yield hself is somewhat lower than buying the highest absolute dividend payers (6.5% versus 8.0%), because most of the high-priced industries pay small dividends.

Table 9-1 shows the returns of buying low P/E stocks by this method and holding them for periods of 2, 3, 5, and 8 years. Again, the results of investing in the most out-of-favor stocks within industries are similar to buying the most unpopular stocks overall. As Table 9-1 indicates, the retums of the laggards do not decay over longer periods of time. A buy-and-hold strategy, as noted, enhances your capital by avoiding most commissions and other transaction costs, and by minimizing capital gains taxes. The retums of price-to-cash flow and price-to-book value (not shown) are also similar.

Figure 9-3 demonstrates just how soundly the relative (or industry) contrarian strategies beat the averages over the full 27 years of the study. Low price-to-cash flow does the best. Ten thousand dollars invested by