Recently I studied the prospectus for Royal Dutch’s 2005 exchange offer. On page 47 it reads:

“The historical trading relationship between Royal Dutch ordinary shares (RDA) and Shell Transport ordinary shares (SHEL) has broadly matched the 60/40 interests set forth in 1907. When this relationship has deviated from parity, it appears to have done so for reasons external to the Royal Dutch/Shell Group, such as index inclusion, relative index performance and taxation changes.”

From 1986 to 2005, the market capitalization of RDA as a percentage of the Royal Dutch/Shell Group averaged 61.72.

This mispricing per se does not prove that security prices are inefficient. The short sale of RDA and simultaneous purchase of SHEL had been consistently profitable, with one exception: in 1998 it cost Long-Term Capital Management several hundred million dollars. In the case of closed-end fund arbitrage, which involves the purchase of fund shares below NAV and the short sale of underlying portfolio securities, the magnitude of mispricing is correlated with the difficulty of finding shares to short sell. These two cases vindicate efficient market hypothesis as I understand it. While mispricings exist, they are either too risky, too costly or too difficult to exploit.

Relative value strategies, however, do not need to be narrowly defined as the type practiced by LTCM, West End Capital (a Buffett investee) or Salomon Brothers. Early this year I effected a relative value hedge by purchasing $3,000 worth of Genesco and $3,000 worth of Finish Line. My initial success has given me a strong interest in specialized operations of this kind—among other things, I have concluded that money can be made both conservatively and plentifully by buying two common stocks which analysis shows to be inconsistently discounting the chance of one major event. This is an unpopular strategy but one that seems to be entirely logical. In the mid-1960s, Warren Buffett practiced a more common variant:

“‘Generals – Relatively Undervalued’ – this category consists of securities selling at prices relatively cheap compared to securities of the same general quality. We demand substantial discrepancies from current valuation standards, but (usually because of large size) do not feel value to a private owner to be a meaningful concept. It is important in this category, of course, that apples be compared to apples – and not to oranges, and we work hard at achieving that end. In the great majority of cases we simply do not know enough about the industry or company to come to sensible judgments – in that situation we pass.

“As mentioned earlier, this new category has been growing and has produced very satisfactory results. We have recently begun to implement a technique which gives promise of very substantially reducing the risk from an overall change in valuation standards; e.g., we buy something at 12 times earnings when comparable or poorer quality companies sell at 20 times earnings, but then a major revaluation takes place so the latter only sell at 10 times. This risk has always bothered us enormously because of the helpless position in which we could be left compared to the “Generals – Private Owner” or “Workouts” types. With this risk diminished, we think this category has a promising future.”

This technique was well suited to the “Nifty Fifty” era, when for instance GM sold at a large premium to Ford, despite nearly identical operating metrics. A great deal has changed since then. First, it is almost impossible to find two corporations similar enough in their operations to be comparable (even Coca-Cola and Pepsi are quite different); and second, the speculative component that caused divergent valuations in the 1960s is no longer present.

Nonetheless I believe that low-risk relative value opportunities will arise from time to time—perhaps once a year.