The Most Magical Formula of All - Research Results (Updated 28 Jan 2006)

If this is your first visit to Three Grahamian Virtues, please take a moment to read the welcome message and site philosophy: http://3gv.blogspot.com/2006/01/welcome-and-3gv-philosophy.html

Preliminary research is complete and I will cut to the chase -- the proposed Most Magical Formula of All (MMFA) is a screen that ranks and scores all companies on the following criteria:

Absolute valuation
Valuation relative to historical norms
Return-on-Capital
Debt Levels
Long-term earnings/sales growth
Recent earnings/sales growth
Recent relative price strength
Company Size

The fine print: (I would like to preface the details of this experiment by stating that I am trying to hold to the value investor's maxim that it is better to be roughly right than precisely wrong. This is a very early version of the MMFA -- the first, to be exact -- and it may or may not change as deeper study into the source formulas takes place. As of this writing, no backtesting of the MMFA has taken place, though I hope to do so in the next couple weeks using AAII's Stock Investor Pro database. In seeking the characteristics of formulas that do really well in bull markets and/or bear markets, I am not at this time subjecting the criteria to rigorous statistical analysis (or even contrasting the best strategies against the worst). Instead, I am looking for patterns in types/proportions of criteria -- if is there an emphasis on long-term growth and/or recent growth, deep value or relative value, operational excellence, stock price movement, etc. When I have identified the essence of a set of great formulas for both conditions, I will judge how they may be joined together in a complementary way so as to "flex" from one set of market conditions to the next. It's impossible to say whether a given year will be bullish or bearish until the year has passed, but if a formula is adaptive, then one may happily ignore externalities and prognosticators (leaving more free time for the finer things in life). I accept the possibility that there is no formula that is completely adaptable, and/or that one may be better off running separate bull and bear strategies simultaneously to smooth returns while maintaining high returns over time. Finally, I am not discussing the specifics of AAII's material, because I am trying to respect their copyrights and encourage subscription to their products.)

Let's look at the conclusions of this experiment. After this section follows a somewhat more detailed examination of the winningest bull/bear formulas. In the very near future, I will be able to backtest the MMFA; at that time, I will talk about how I converted the criteria categories from above into specific, screenable factors.

In my opinion, the essence of the best bull strategies is this: Small, sustained high growth companies with improving/high-grade operations, strongly rising share prices in recent months, and valuations near to or below market levels.

The essence of the best bear formulas is this: Small, growing companies with improving and/or high-grade operations, and an emphasis on deep to moderate valuation.

You can clearly see that they are not all that different! Bull strategies focus less on valuation, but they are more similar than not.

Here are the features screened for in the formulas and a brief discussion of them.

1. Growth - Far and away, if all the criteria from both formula types is aggregated, growth in earnings and revenues is the most prominent variable. Surprisingly, this factor is even more heavily weighted in formulas that did well in bear markets. Generally, the growth criteria concern annualized growth over five years and/or high/accelerating growth in the last year.

2. Valuation - Great bull market strategies have a range of valuation standards. Some have no valuation criterion. Others look for deep value. Some only require comparable to slight undervaluation relative to the market. Great bear market formulas have a more nuanced view of value and emphasize deep value. Some look for valuations below historical norms and below their industry. Others seek those with very low valuation multiples, period. It is worth saying that great bear market strategies tend to hold up better in bull conditions than do some bull market winners that crumble in bearish conditions. Needless to say, I think this should be a prominent, indispensible feature of a formula one chooses to employ.

3. Recent Stock Price Strength - There is an emphasis on recent strong stock price performance relative to the market. Not every formula has this criterion, but it is present in both bull and bear winners. Though not a Grahamian or Buffettesque precept, to some extent, it makes sense. The bane of value investors are value "traps" that never appreciate. The benefit of this "technical" criterion is that it often accompanies stocks whose underlying economics are strong/improving -- when tied to a valuation component, it may give an investor greater confidence that a company is in the process of having its intrinsic value recognized. If a purist value investor accepts that a stock he buys might continue a downward slide, yet he has conviction in his assessment, what is wrong with waiting for that stock to demonstrate that its price reversal (value reappraisal) is for real? This is about the extent of my ponderings on the subject of technical analysis, and it is only a complement to a value-oriented selection process.

4. Operational Quality - Some formulas include basic levels of profitability (moderate to low). Others emphasize high quality, recently and/or consistently. Obviously, it is today's price to tomorrow's quality that makes or breaks returns. Buying baskets of stocks that appear to be turning the corner from previously poor economics seems to work when prices are low across the board. Buying companies whose economics are consistently better than average at prices below historical norms/below competitors/below market valuation levels also works well. The fatal flaw of growth-oriented strategies that seek high operational performance without regard to the fact that high expectations have already been incorporated into share prices. However, buying high octane companies that have a history of growing quickly, but for whom expectations have deflated, is also a winning gambit (see the John Dorfman formula articles).

7. Debt - There is a surprising lack of debt criteria outside of Graham's enterprising formula. I happen to think that including debt standards is like strapping on a seat belt.

8. Size - Some strategies have minimum size standards (presumably as a quality-seeking measure), while others specifically seek small companies (presumably as a way to find the largest discrepencies between expectations and economics). Due to the empiric proof of average small company outperformance, I definitely lean toward the latter (in direct opposition to Graham).

8. Dividends - The only criterion with no overlap between bull/bear market winners is dividends. These are only sought by more conservative bear strategies (though one might consider them subordinate to the cash flow criteria which are present in bull formulas).

The next step is to put the MMFA into practice. Please check back soon to see these details and thereafter to see backtested results.

- Ian