The Daily Workshop Report
by Randy Befumo (TMF Templr)

WILLIAMSBURG, VA. (June 13, 1997)

What does it really mean when someone presents a screen that when backtested generates 30%-plus returns? Even starting with relatively small amounts of money, high absolute returns tend to add up really quick. As the size of the assets under management increases, the flexibility of the investor managing those assets becomes more and more limited. This column normally tends to showcase compounding experiments that leave the investor with billions of dollars after a decent amount of time starting with small investments. Although it is important to illustrate to potential investors the wealth creation machine that is the stock market, the reality is that unless the stocks being invested in have multi-billion dollar market capitalizations, these approaches tend to collapse under their own weight fairly quickly if mechanically followed.

Take for example MICRON TECHNOLOGY <% if gsSubBrand = "aolsnapshot" then Response.Write("(NYSE: MU)") else Response.Write("(NYSE: MU)") end if %>, a pretty large company. The manufacturer of commodity chips currently has a market capitalization of $8.5 billion, not too shabby. The investor putting a few thousand dollars in this company over a relatively short period of time really would never expect to cause the share price to move. The investor putting a few million dollars in the stock over a short period of time might actually start to move it quite a bit. The investor putting a few tens of millions into the stock would definitely cause the price to move quite a bit, as it would suddenly become a pretty substantial chunk of the outstanding equity. Even if mechanical models show strong historical returns, when capital actually starts to attempt to achieve those returns, they change the system that was being analyzed. The former inefficiency begins to become more efficient by the very allocation of capital. Capital compounded at high rates begins to have more of an effect on the actual valuation of the stock, meaning success contains the seeds of its own undoing.

How can a model be successful then? With billions of dollars being deployed with a short-term emphasis every single day, if money is invested with a different time frame in mind, all of the "noise" of short-term antics will drown out any effect of long-term money being invested. If billions are invested in a short-term game, long-term investors have to worry only when that money starts to chase longer-term goals. The dissonance between the two time frames creates an inefficiency that can become an opportunity for those investors who rely on the actual fundamentals of the business, most notably the rate of earnings growth, sales growth, or dividend yield.

Models that generate high returns run into problems when they are transactional, i.e. when they trade frequently. While potentially generating 40% per year, having to reinvest 12 times a year quickly extinguishes part of the gain because of the effect the transactions have on the stock itself. However, generating 20% with minimal trades -- preferably none -- increases the efficiency of the assets and does nothing to affect the current pricing of the company being invested in. Additionally, despite the high returns allegedly possible in churning, at a certain point giving up potential 20% returns on each dollar spent on commissions to chase 40% returns -- which will quite possibly become undone by the very act of chasing them -- destroys wealth rather than creating it.

The benefits of deploying capital with a long-term time frame are immense and difficult to discuss in a few paragraphs. The most important for this audience are sustainability of returns and the efficiency with which the money is invested. The fact will always remain that billions are invested with very short time frames, normally less than a year if not less than a month. Competing with this money to find inefficiencies with mechanical models that can be isolated by a few screens becomes difficult, particularly when the models have a high degree of "friction," or trading expenses. As every dollar not invested generates a negative return and every dollar actively reinvested can diminish the expected returns, focusing on "screens" that would identify quality businesses that could be bought and held would seem to be sensible. This balances potential returns against the risks that frequent trading presents to models with very little statistical analysis other than some work on the absolute returns they can generate.

Monthly Growth Screens
(Jan. 3 to present)
 30.21%   Relative Strength   
 19.41%   S&P 500 Index   
 17.84%   Low Price/Sales   
 -3.02%   YPEG Potential   
 -1.12%   Investing for Growth   
-19.88%   Unemotional Growth   
 -9.82%   EPS Plus RS   
-20.31%   Formula 90   

Annual Value Screens
(Jan. 1 to present)
20.41%   Dogs of the Dow   
20.68%   Dow Jones Ind Avg   
14.19%   Beating the S&P   
11.99%   Unemotional Value   
11.99%   Beating the Dow   
10.56%   Dow Combo   
 4.80%   Foolish Four