Ben Graham calls it the Central Concept of Investment:
In the old legend the wise men finally boiled down the history of mortal affairs into the single phrase, “This too will pass”. Confronted with a like challenge to distill the secret of sound investment into three words, we venture the motto, MARGIN OF SAFETY. (The Intelligent Investor, Chapter 20).
Seth Klarman used it as the title of his book (a copy of which used to be, but is no longer, available online – for reasons that escape me).
The Margin of Safety is the cornerstone of intelligent investing. But its meaning is often misunderstood. A common definition is: the difference between the intrinsic value of a stock and its market price. This echoes a famous passage in which – quoting Walter Schloss, one of The Superinvestors of Graham-and-Doddsville – Warren Buffett writes: “If a business is worth a dollar and I can buy it for 40 cents, something good may happen to me”.
This definition is correct but incomplete. It misses a crucial point:
We must recognize, however, that intrinsic value is an elusive concept. […] It is a great mistake to imagine that intrinsic value is as definite and as determinable as is the market price. […]
The essential point is that security analysis does not seek to determine exactly what is the intrinsic value of a given security. It needs only to establish either that the value is adequate – e.g., to protect a bond or to justify a stock purchase – or else that the value is considerably higher or considerably lower than the market price. For such purposes an indefinite and approximate measure of the intrinsic value may be sufficient. (Security Analysis, Chapter 1. Also quoted in Klarman, Ch. 8)
Exact Appraisal Impossible. Security analysis cannot presume to lay down general rules as to the “proper value” of any given common stock. Practically speaking, there is no such thing. The bases of value are too shifting to admit of any formulation that could claim to be even reasonably accurate. The whole idea of basing the value upon current earnings seems inherently absurd, since we know that the current earnings are constantly changing. And whether the multiplier should be ten or fifteen or thirty would seem at bottom a matter of purely arbitrary choice. But the stock market itself has no time for such scientific scruples. It must make its values first and find its reasons afterwards. Its position is much like that of a jury in a breach-of-promise suit; there is no sound way of measuring the values involved, and yet they must be measured somehow and a verdict rendered. Hence the prices of common stocks are not carefully thought out computations but the resultants of a welter of human reactions. The stock market is a voting machine rather than a weighing machine. It responds to factual data not directly but only as they affect the decisions of buyers and sellers. (Security Analysis, Chapter 39)
Many analysts and investors spend an inordinate amount of time and effort calculating intrinsic values with ever higher precision. They build impressively detailed valuation models, which help bolstering their confidence in the robustness of the output. But the Margin of Safety has little to do with the precision of the intrinsic value calculation:
Most investors strive fruitlessly for certainty and precision, avoiding situations in which information is difficult to obtain. Yet high uncertainty is frequently accompanied by low prices. By the time the uncertainty is resolved, prices are likely to have risen. Investors frequently benefit from making investment decisions with less than perfect knowledge and are well rewarded for bearing the risk of uncertainty. The time other investors spend delving into the last unanswered detail may cost them the chance to buy in at prices so low that they offer a margin of safety despite the incomplete information. (Klarman, p. 157-158).
The Margin of Safety does not reside in the precision of the intrinsic value calculation, but in the probability that the market price is wrong. The size of this probability is proportional to the valuation gap, but the relationship is not linear. The gap needs to be ample enough as to allow for a large area of imprecision.