Value Trust
"It is odd to watch with what feverish ardor the Americans pursue prosperity and how they are ever tormented by the shadowy suspicion that they may not have chosen the shortest route to get it… They clutch everything but hold nothing fast, and so lose their grip as they hurry after some new delight."
—Alexis de Tocqueville (1835)
How Much is Too Much; or When to Sell?
One of the common questions we get from pension consultants, brokers, and shareholders is "When are you going to sell your Dell or America Online ('AOL')?" In the Value Trust, the average cost of our Dell position, acquired in early 1996, is about $4.00 per share. The shares trade today at about $115. The average cost of our AOL position is about $20; the stock trades now around $125 per share. At the end of the second quarter, Dell made up over 8% of the assets in Value Trust, AOL just over 7%. In the Special Investment Trust, where we've owned AOL for a longer period, the stock makes up a whopping 16% of assets at June 30.
There are several issues implicit in the question of when to sell. First is the generic strategy question: what general criteria drive the process of deciding when to sell? The second has its origin in the usual behavior of money managers, whose average portfolio turnover approximates 100% per year. In a world where investors appear to be frenetically trading shares in response to or in anticipation of short-term price moves, multi-year holding periods are atypical. Investing in a company for many years, instead of speculating on a stock's price action, has become unusual enough to warrant an explanation. Why have you held the stock for such a long time and when are you going to sell? Third is the question of risk: How much of a portfolio's assets should be concentrated in a single holding? How large can a holding become without subjecting the portfolio to undue risk?
A variant of the first question has to do with strategy. We are value investors who invest in shares of companies trading at large discounts to our assessment of the intrinsic value of the business. How can a company's stock go up 25x in a few years and still be trading at a discount to intrinsic value? How can anything selling at 75x trailing earnings (Dell) or 100x estimated earnings (AOL) possibly be considered a value? Haven't the shares gone from value to growth to outright speculation?
Our selling strategy is simple: we sell when a company's share price has reached fair value; we sell to replace an existing holding with a better bargain; and we sell when our original investment case is no longer applicable. "Fair value" means that the market has, in our opinion, correctly priced the stock. It understands and has adequately discounted the future free cash flows of the business so that we can no longer earn an excess return by holding the shares. Put differently, fair value means that the risk-adjusted rate of return available by holding the shares long-term is no greater than that of the market as a whole, and may be less. Second, if we are fully invested and our research uncovers a security we believe has more value than something we already own, we will replace the lesser bargain with the greater. Finally, we sometimes find that our analysis was wrong, either because we uncover new information, or because we interpret our existing information differently, or because the world changes—new legislation is passed, new regulations are promulgated, the competitive landscape changes, and so on. We will sell securities when we believe we are wrong, whether the share price is above or below our purchase price.
In general, we try to keep our portfolios positioned in a diversified, but focused, group of companies that we believe trade at discounts to intrinsic value and therefore offer the opportunity for above-average risk-adjusted rates of return.
There is no time limit on investing. We do not sell because we've owned shares for some period of time. Price and value are two different and independent variables. We do not sell because a stock's price has gone up, or down, a lot. We value businesses, not stocks—we compare the value of the business to what the market says it is worth. If we believe the discount is sufficient, we purchase the shares. We invest in companies; we do not speculate on share price changes. Thus we do not react to earnings reports, or analyst opinion changes, or try to guess whether the Fed will change rates next month, or whether the GDP number will be higher or lower than expectations.
When we look at Dell, for example, which has a market capitalization of over $60 billion, our analysis focuses on whether Dell's business is worth more or less than $60 billion, and not on whether the stock is likely to outperform, or whether computer sales this month will be slower or faster than expected. Elsewhere in this report, Lisa Rapanou, assistant portfolio manager of the Special Investment Trust, takes a more detailed look at how we analyze the value of AOL.
When the market prices two companies at the same price/earnings ratio, then, other things equal (e.g., assuming the same capital structure), it believes the businesses have the same value expressed as a multiple of earnings. But other things are rarely equal. Dell and Gateway compete in the personal computer area. Dell trades at over 40x next year's earnings, Gateway at under 20x. Is Gateway cheaper than Dell? If Dell earns over 200% on each dollar of invested capital, as it does, isn't it worth more than Gateway, which earns 40%? How much more? If it's 5x as profitable, is it worth 5x as much? If its price/earnings ratio is only 2x that of Gateway, is it then a bargain? This is not meant to suggest an answer, just to note that businesses have different values, and those values are hardly captured by a single number on a linear scale such as price/earnings ratios. We are delighted, though, when people persist in thinking price/earnings ratios tell most of the story about value since it gives us an opportunity to make better investment decisions by doing a more thorough job of analysis.
That still leaves the question of how much is too much? This is a question of risk versus reward. When does a stock become so large a portion of a portfolio that the portfolio's returns depend too much on that security's performance? We typically own 40 or so stocks, and our top ten names constitute about 40% of our assets. We think that the fewer the names the better. Even if our largest holding were 10%, that means that 90% of our returns would be sourced elsewhere. The Chicago Bulls have five sources of points operating at any one time, but Michael Jordan's play usually determines their results. Should they trade him for a rookie so they won't be so dependent on him? If he does poorly, in all likelihood, so will they. When will they replace him? Probably when they determine he can no longer produce the results given what they are paying for him. That same criterion is helpful in thinking about stocks.
Warren Buffett has been justly celebrated for his investment genius and has been a consistent proponent of portfolio concentration. Our new colleague Robert Hagstrom, author of The Warren Buffett Way, and manager of the Legg Mason Focus Trust, calculated that Coca Cola, which made up 12% of Buffett's common stock portfolio shortly after he purchased it in 1988, now makes up about 37% of his portfolio. Should he sell it, trading as it is at over 50x next year's earnings?
Is Buffett no longer a value investor because he continues to hold Coke, Gillette, and Disney, all of which sell at price/earnings ratios and price-to-book value ratios far above the market's current levels, which are historically among the highest ever?
You may have read recently about the estate left by Don Othmer, a professor of chemical engineering at a small school called Brooklyn Polytechnic and his wife, Mig. In the late 1950s they put all of their money with Warren Buffett and it all subsequently went into one stock, Berkshire Hathaway. Had they consulted other advisors, they no doubt would have been told they had way too much in one stock, and they would, as many investors do, have continuously cut back their position as it went up to diversify into other investments. They also would not have been worth $800 million when their estate was settled.
In 1956, J.L. Kelly, Jr. published a paper in mathematical information theory that dealt with the transmission of information over a phone line. In it, Kelly devised a formula that subsequently was shown to have general applicability for any problem that seeks to maximize a growth rate per unit of investment. Using Kelly's formula, we can calculate that if Buffett has 37% of his portfolio in Coke, he believes his probability of being right about Coke is 70%. More generally, Kelly's formula would indicate that the growth rate of a portfolio is maximized with no more than five stocks, assuming the returns of each security are independent and one believes the probability of outperformance is greater than 50%. Kelly's formula, which is not well known among investors (it also serves as the basis of fixed fractional betting strategies among gamblers), provides a clue about why the largest fortunes are made by concentrating not by diversifying, and also would suggest that the new category of focus funds is likely to prove quite rewarding to investors, assuming managers can generally distinguish their best ideas from the rest.
The questions of how much is too much and of when to sell have no perfect answers; they depend on imperfect information and on judgments that are often wrong. We cannot promise we will be right about these issues, but we do promise to think carefully, critically, and diligently about them, and to use our best judgment to try to add value to your investment.
We have not attempted to be exhaustive in these remarks on selling and on concentration, but to provide some insight into our thinking on these complicated subjects. As always, we appreciate your support and welcome your comments.
Bill Miller, CFA
Share