James O’Shaughnessy’s books

James O’Shaughnessy’s What Works on Wall Street is an excellent book. Many of its results are definitive. It shows that some investment strategies consistently outperform the market as a whole while some other strategies consistently do worse than the market as a whole. O’Shaughnessy demonstrates through examples that accepting an increased risk does not guarantee a reward. In fact, some strategies increase risk and lower returns.

His investigations were extensive. He presents detailed results. He makes use of Standard and Poor’s Compustat database, the most comprehensive available, which extends back to 1950. [His research covers 1951-1996.]

He defines two stock universes. His All Stocks universe consists of all companies in the database with market capitalizations of $150 million or more. This includes 4000 to 5000 companies. His Large Stocks universe consists of 1200+ companies with capitalizations bigger than the average. This translates roughly to capitalizations of $1.0 billion and higher, which extends through the midcaps and down into some small caps.

He did examine the effect of excluding companies with capitalizations smaller than $150 million. He found that it is wrong to include them in studies. They are illiquid. Their prices are unrealistic. Some have bid and asked prices that differ by a factor of 2 or more. The smallest group, with market capitalizations of $25 million and less, produce dramatically better results than all other capitalizations. But this is only an artifact rooted on illiquidity. This is the reason that academic studies have favored small cap stocks.

James O’Shaughnessy establishes explicit rules for each strategy and applies them consistently. He starts with an initial purchase of $10000. He makes no deposits and no withdrawals throughout the entire period. This seems like buy-and-hold. It is not.

Part of his procedure is to select fifty stocks when he applies his rules and then to rebalance his portfolio every year. The problem is that this kind of rebalancing is NOT rebalancing. It is trading.

O’Shaughnessy starts by buying equal dollar amounts of fifty stocks. One year later, he applies his rules to identify a new group of fifty stocks. The new group is selected from either the All Stocks universe or the Large Stocks universe. It is NOT limited to his original selection. He sells all of his previous holdings and buys equal dollar amounts of the new group of fifty stocks.

This procedure can result in an annual turnover very close to 100%.

O’Shaughnessy’s standard procedure assigns no cost to this rebalancing and reallocation. A side effect is that relative strength looks much more attractive than it really is. [Relative strength is the price increase from the previous year compared to a larger group. Another factor such as the price-to-earnings ratio threshold is used as an initial screen. The fifty stocks that are selected come from those that remain. They are selected according to relative strength.]

Some Important Results

O’Shaughnessy reports that investing in the S&P500 index has worked in the past because it is one large cap strategy that is applied consistently. There is no style drift. His arguments are convincing, subject only to the limitations of his study. His blind spot is cost, which he models as zero.

With occasional qualifiers as to whether examining All Stocks or Large Stocks, traditional valuation measures work. Low valuations translate into improved performance compared to the market as a whole. High valuations bring about diminished performance, often dramatically so.

Selecting low single-year price-to-earnings ratios works with the Large Stocks, but not with the universe of All Stocks. [This might be interpreted as indirectly supporting Benjamin Graham’s recommendation to average several years of earning when using the price-to-earnings ratio. The difference in behavior might have to do with the stability of earnings. Larger stocks have more consistent earnings.]

The results warn those relying on low price-to-book ratios that this indicator failed throughout the 1950s and 1960s. Popular growth stocks typically have high price-to-book ratios.

Value wins out eventually but it can take a very long time.

Additional value indicators include low price-to-sales ratios, low price-to-cashflow ratios and high dividend yields. The best single valuation measure was the price-to-sales ratio. High dividend yields worked in the Large Stocks universe but not the All Stocks universe. Considering that Large Stocks are defined as having capitalizations in the neighborhood of $1.0 billion and up, this is not restrictive.

Combining traditional value indicators helps. The same companies usually satisfy several value criteria. There may be trouble when different value indicators tell different stories.

Treated separately, traditional growth indicators fared poorly. Big single-year earnings gains, big five-year earnings gains, high profit margins and high returns on equity all disappointed.

High relative strength worked. Such stocks carry high valuations. But typically, they do not carry the highest valuations. Buying the worst performing stocks, those with the biggest single-year losses, leads to financial ruin.

The combination of value screens and relative strength produced excellent performance. Temper this conclusion because of its tight connection to the annual rebalancing process (with its new stock selection). The relative strength indicator needs to add value for only one year.

Cornerstone Strategies

James O’Shaughnessy identifies Market Leaders as the roughly 570 Large Stocks (i.e., with market capitalizations of $1.0 billion or more) with more common shares outstanding and higher cashflow per share than the averages in the Compustat database with 1.5 times the sales of the average in the Compustat database. Somewhat arbitrarily, he excludes all utilities from this list.

His Cornerstone Value portfolio consists of the 50 Market Leaders with the highest dividend yields. He would have us buy equal dollar amounts of all of the stocks in this portfolio.

O’Shaughnessy selects growth stocks from the All Stocks universe. They must have higher earnings than one year earlier and they must have price-to-sales ratios below 1.5. He calls the 50 stocks with the strongest one-year price gains within this collection his Cornerstone Growth portfolio. He would have us buy equal dollar amounts of all of the stocks in this portfolio.

How To Retire Rich reads like a diet book, generating a lot of enthusiasm and building up confidence in O’Shaughnessy and his strategies.

He glosses over costs in an obvious conflict of interest situation. He has created strategy index funds with annual fees (apparently) of 1.25% (refer to page 120).

His use of the term index fund is highly misleading. The qualifier strategy cancels it out completely. O’Shaughnessy has several mechanical investment strategies, which typically hold 50 stocks at a time and which replace all holdings once per year. Apparently, holding 50 stocks for a year that are selected by previously established rules is sufficient to qualify as a strategy index fund. The turnover can fall below 100% only when the same company meets the same selection criteria. This is from a universe of 500 or more stocks.

James O’Shaughnessy introduces five strategies in How To Retire Rich. Two of them are for holdings of 25 or 50 stocks. What he now calls Reasonable Runaways is the same as Cornerstone Growth except that he requires the price-to-sales ratio to be below 1.0 instead of 1.5. His Leaders with Luster are the same as Cornerstone Value.

He recommends holding equal portions of Reasonable Runaways and Leaders with Luster.

He has new portfolios with holdings of ten stocks. They are the Dogs of the Dow, his Utility Strategy and his Core Value holdings. He continues his version of rebalancing. The turnover is likely to be less than 100% by nature of his selection criteria, especially for the Dogs of the Dow.

With the Dogs of the Dow, you buy equal dollar amounts of the ten highest yielding of the 30 stocks of the Dow Jones Industrial Average. Sell and replace annually. [Even if the same company appears on a newer list, he adjusts the number of shares so that each holding starts out with an equal dollar amount.]

With his Utility Strategy, his approach is very similar to his Cornerstone Value approach. However, he includes utility stocks as well as other stocks. He buys equal dollar amounts of the 10 stocks with the highest dividend yields that have a Value Line safety rank of 1. He sells and replaces all of these annually. [Even if the same company appears on a newer list, he adjusts the number of shares so that each holding starts out with an equal dollar amount.]

With his Core Value Strategy, he starts with those stocks with the highest Value Line financial strength rating of A++. This leaves 40 stocks. Then he eliminates half of these stocks, retaining those with above average dividend yields [along with their A++ financial strength ratings]. That leaves 20 stocks. His final selection is the ten with the highest projected Dividend Growth Rates for the next 3-5 years. He sells and replaces all of these annually. [Even if the same company appears on a newer list, he adjusts the number of shares so that each holding starts out with an equal dollar amount.]

Have fun.

John Walter Russell
I wrote this on May 14, 2005.