First off, the time period chosen must be deciphered as an adequate representation of the firm and its movements throughout a full business cycle; choosing a cyclical business and focusing on only half of the cycle will certainly lead to inaccurate conclusions. For this exercise, I’m most comfortable taking no less than ten years worth of data, while consciously considering the time frame in question and how developments during which have adversely or positively impacted the company. A great example would be Lululemon (LULU); in my mind, to grab the last five years of data and simply extrapolate into the future would be a grave miscalculation that fails to account for a fad which will face increased competition and compressed margins due to the lack of any real competitive advantage for the incumbent. Without any barriers to entry, Lululemon will not be able to continue to generate outsized gross margins (56%, compared to 40%, 37%, and 44% for other retailers, such as The Gap (GPS), Aeropostale (ARO), and The Buckle (BKE), respectively), and will revert to the mean as new entrants chip away at excess returns. As Graham noted, “The analyst’s primary role is to anticipate change when others are extrapolating the past.”
Another important concept is the ability to step back from the research, and not to become enamored by the numbers. Academic finance has fallen victim to this desire for precision (but certainly not accuracy), with an emphasis on metrics like WACC, cost of equity, and beta. As Graham notes, “It [the trend projection] should only be used as a rough index of what might be expected from the future.”
As always, the intelligent investor demands a margin of safety, and would not make an investment on the hope that extraordinary growth rates continue to materialize. As James Montier calculated in “Value Investing: Tools and Techniques for Intelligent Investment,” staying away from expectations of high double digit growth would be well advised: From 1951-1998 in the United States, only stocks in the 90th percentile and above were able to achieve growth in operating income (before depreciation) of 15%-plus over a ten-year period.
As expected, trend extrapolations are most reliable and useful for companies with a long history of consistent (hopefully increasing) cash flow generation. Unlike the speculator, who is guessing about the future without a reasonable guide from the past, the investor must “treat the future as a hazard which their expectations must encounter and attempt to guard against.” This explains why someone like Warren Buffett of Berkshire Hathaway (BRK.A, BRK.B) buys companies with sustainable competitive advantages and relatively static business models that won’t look much different 10, 20 and 50 years from now.
Many of the growth stocks that are loved by investors today (like Salesforce.com (CRM), for example), simply do not meet the safety requirements (especially at these prices) for sound investment: “The security analyst is on safest ground when favorable expectations are treated as an added reason for a purchase which would not be unsound if based entirely on the past record and the present situation.” Salesforce.com, at $136 per share, is the antithesis of this idea: Nobody in their right mind would consider paying even half of today’s market price based on the historical record and present economic situation.
People interested in investing rather than speculating should heed the words of Ben Graham and James Montier, and flock to stocks with consistent (proven) earnings power rather than wildly optimistic expectations for future growth; for the intelligent investor, this means treading with caution when implementing trend line extrapolation in security valuation.