Sunday, March 5

Week 296 - Testing Our Stock-picking Algorithm

Situation: The “normal” way to invest in stocks is to play the Market as a whole. Typically, that means dollar-cost averaging into index funds such as those offered by Vanguard Group (VFINX, VEXMX, VTSMX, VGTSX) or SPDR State Street Global Advisors (SPY, MDY, DGT). That way, transaction costs are minimized and you can’t miss out on market moves. Or, you can try to beat the Market by building (and managing) a portfolio composed of many stocks representing all 10 S&P Industries. There’s no shortage of books on the subject but one will suffice: “The Four Pillars of Investing: Lessons for Building a Winning Portfolio” by William J. Bernstein, 2002, McGraw-Hill. There you’ll find a mathematical exercise proving that the only logical way to do well from investing in stocks is to focus on dividend growth. 

Mission: Lay out an algorithm for stock selection.

Execution: Start with companies that have grown their dividend annually for 10 or more years, i.e., S&P Dividend Achievers. Select the ones that have grown their dividend faster than our benchmark S&P 500 index fund, VFINX, over the past 10 years, which is 6.8%/yr. Narrow that list down to those companies large enough to be on the Barron’s 500 List. Why? Because large companies a) have multiple product lines, and b) their stock has enough activity on the CBOE (Chicago Board Options Exchange) to facilitate price discovery. Remove any companies that don’t have an S&P Bond Rating of at least A- and an S&P Stock Rating of at least A-/M. Remove any companies that don’t have a 16-yr trading record that has been analyzed statistically by the BMW Method. Exclude companies that rely on long-term debt for more than 1/3rd of total capitalization, or couldn't meet dividend payments from free cash flow (FCF) in the two most recent quarters. Also exclude companies that are over-reliant on short-term debt, i.e., have more than 5% negative Tangible Book Value.

Administration: There are 27 companies that pass the above screen. By using the BMW Method, we have separated those into a group of 16 that has no greater chance of loss in a future Bear Market than the S&P 500 Index (see Column M of the Table under “Non-Gambling”), and a group of 11 that has a greater chance of loss (see red highlights in Column M under “Gambling”). If you do choose to invest in one of the Gambling companies, watch price-action because you’ll likely want to SELL at some point. In Columns N-P we provide data on 3 ratios that assess the overall health of Financial Statements.

Bottom Line: The purpose of stock-picking is to Beat the Market. It is very difficult, expensive, and time-consuming to do so over more than one Market Cycle. We have laid out a system for picking stocks, and back-tested it. It has a Failure Rate of 4%. In other words, 25 of the 26 stocks beat the S&P 500 Index over the past 16 years (see Column K in the Table). Just to be clear, we recommend that you dollar-average into index funds (see BENCHMARKS section of Table), and/or Berkshire Hathaway B-shares (where you would be building a position in over 100 large and mid-cap companies). 

Of the 16 stocks we designate as non-gambling investments, most carry market multiples (or lower) for EV/EBITDA: GWW, UNP, CNI, WEC, APD, NEE, TRV, WMT, TGT. Those stocks are attractive for purchase if no issues arise from your further research, such as reading the Morningstar evaluation.

Risk Rating is 7 for the stock selection system outlined above. Why is that? Because of Selection Bias ( and Transaction Costs (

Full Disclosure: I dollar-average into UNP, NKE, JNJ, PG, NEE and MSFT, and also own shares of CNI, MMM, WMT, HRL, TRV, MKC, ROST, TJX, GD and CAT.

NOTE: Metrics are current for the Sunday of publication. Red highlights denote underperformance vs. VBINX at Line 40 in the Table. Purple highlights denote Balance Sheet issues and shortfalls. Net Present Value (NPV) inputs are described and justified in the Appendix to Week 256: Briefly, Discount Rate = 9%, Holding Period = 10 years (no dividends collected in 10th year), Initial Cost = average stock price over the past 50 days (corrected for transaction costs of 2.5% when buying ~$5000 worth of shares). Dividend Growth Rate is the 5-Yr CAGR found at Column H. Price Growth Rate is the 16-Yr CAGR found at Column K ( Price Return (from selling all shares in the 10th year) is corrected for transaction costs of 2.5%. The Discount Rate of 9% approximates Total Returns/yr from a stock index of similar risk to owning a small number of large-cap stocks, where risk due to “selection bias” is paramount. That stock index is the S&P MidCap 400 Index at Line 45 in the Table. The ETF for that index is MDY at Line 39.

Post questions and comments in the box below or send email to:

No comments:

Post a Comment

Thanks for visiting our blog! Leave comments and feedback here: