The Perfect Database?

by Ethan Haskel ([email protected])

Baltimore, MD. (Oct. 21, 1998) -- Last week we saw that, in order to maximize the investment returns of high-yielding stocks, we need to find a diverse group of large, market-leading companies. Where do we look?

We need to find a list, or database, of companies in which we can find specific stocks that fit our criteria. The database is the heart and soul of any screening strategy. Without it a world-class idea remains unproven and unusable.

What are the characteristics of the perfect database?

-- It should be authoritative and accurate.
-- It should be widely available at a reasonable cost.
-- It should be easy to use.
-- It should contain data for many past years and offer the strong possibility of surviving in present form for many future years.
-- It should contain all essential data to find stocks that meet the chosen criteria.

When searching for the best database to choose stocks for our Beating the S&P (BSP) approach, what essential data do we need?

-- The market capitalization of the companies. (Market capitalization is the stock price times the number of shares outstanding.)
-- A workable classification of the companies into sectors, or similar areas of economic activity.
-- Stock yield.

Because Beating the S&P requires relatively few nuggets of data, it would seem a fairly simple task to find a suitable database. If only 'twere so!

The Wall Street Journal and Investor's Business Daily don't list a stock's capitalization or sector. The Standard & Poor's Stock Guide does list market capitalization, but trying to extract the highest capitalized stocks is the ultimate lesson in frustration. Most of the computerized databases don't contain many years of historical data.

Standard & Poor's will perform personalized database searches, but at prices that are non-standard, guaranteed to make you poor. (James O'Shaughnessy received permission to analyze the exhaustive S&P Compustat database, the first time ever for an outside researcher. That opportunity just doesn't come around very often.)

After many fruitless searches at the local library, a fellow Fool (whose name, alas, is lost) clued me in to the Business Week annual issue detailing the top companies of the S&P 500. There in one glorious package were the largest American companies, complete with yields and sector groupings. All courtesy of Standard & Poor's, an unsurpassed authority in the field.

The Business Week listing fit most all of my criteria for an ideal database, with the notable exception that the usable data was only available for 1987 forward. That's one limitation I had no choice but to accept.

We now have our roster of available players. We still need to develop some basic rules to choose the BSP All-Star Challengers, which will be competing against the perennial champions, the S&P 500 Team. Once the teams are set, let the backtesting games begin!

BSP year to date returns (as of 10-20-98):

 
 Anheuser Busch    +31.2% 
 Emerson Electric  +16.6% 
 Ford              +46.1% 
 Kimberly-Clark     -3.0% 
 Texaco            +15.4% 
  
 Beating the S&P   +21.3% 
 S&P 500            +9.6% 
 Dow 30             +7.6% 
 Nasdaq             +4.4%