Leithner & Co. Pty. Ltd.

Circulars to Shareholders
Site Map


Part I

15 September 2000

... the bona fide investor does not lose money merely because the market price of his holdings declines, and the fact that a decline may occur does not mean that he is running a true risk of loss... we apply the concept of risk solely to a loss which is either realized through actual sale, or is caused by a significant deterioration in the company’s position – or, more frequently perhaps, is the result of the payment of an excessive price in relation to the intrinsic worth of the security.

Benjamin Graham and David Dodd,
Security Analysis

Most market participants, including institutions, brokers, advisors and private investors, define investment risk in terms of the short-term ups-and-downs of a security’s market price (relative either to comparable securities or that of ‘the market’ as a whole). As a result, the practice of investment risk management is conventionally understood as an attempt to reduce within acceptable bounds the short-term variability – particularly in a downwards direction – of an investment portfolio’s current market worth.

A small and reprobate minority, value investors in the Graham-and-Dodd mould, disregard both the conventional definition of investment risk and the standard practice of investment risk management. Their impertinence – and in some cases scorn – extends to the body of academic literature which underlies this definition of risk and practice of risk management. Prominent within this academic literature is the Capital Asset Pricing Model (CAPM), Efficient Market Theory (EMT) and Modern Portfolio Theory (MPT). Graham-and-Dodders’ scepticism extends to the mathematical and statistical methods of this academic literature and the money managers who pledge allegiance to it.

In the words of Warren Buffett, Benjamin Graham’s most prominent student and colleague, “for owners of a business – and that's the way we think of shareholders – the academics’ definition of risk is way off the mark, so much so that it produces absurdities.” Among the many examples: Berkshire Hathaway, Inc., which Mr Buffett chairs, bought a substantial percentage of The Washington Post Company in 1974. The price paid for these shares implied that TWPC, considered as a whole, would fetch $80m if it were sold on the open market. Yet “if you’d asked any one of 100 analysts how much the [entire] company was worth when we were buying [a portion of] it, no one would have argued with the fact that it was worth $400m.”

By Mr Buffett’s way of thinking, the price volatility of TWPC’s shares enabled him at a propitious moment to buy them at a small fraction of their intrinsic value. In effect, he was buying assets worth $1.00 for only $0.20. According to CAPM, EMT, etc., however, the price volatility of TWPC’s shares made them a risky proposition – indeed, the greater the shares’ price volatility and the lower their purchase price, the riskier the investment. As Buffett concluded incredulously, “with that, [the academics] have lost me.” (Berkshire’s stake in The Washington Post Company, for which it paid $10.6m in 1974, had a market value of $150.0m in 1984 and $419.0m in 1994).

In 1984, at a Columbia University symposium marking the fiftieth anniversary of the publication of Graham and Dodd’s Security Analysis, Mr Buffett delivered a celebrated speech entitled “The Superinvestors of Graham-and-Doddsville.” In his address (which was published in the fourth revised edition of Graham’s The Intelligent Investor) Mr Buffett observed that “needless to say, our Graham and Dodd investors do not discuss beta, the Capital Asset Pricing Model or covariance of returns. these are not subjects of any interest to them. In fact, most of them would have trouble defining those terms.” Elsewhere he has noted that “the disservice done to students who have swallowed EMT has been an extraordinary service to us and other followers of Graham. In any sort of contest – financial, mental or physical – it’s an enormous advantage to have opponents who have been taught that it’s useless to even try. From a selfish standpoint, we should probably endow [academic posts] to ensure the perpetual teaching of EMT.”

Graham-and-Dodders’ criticisms of the conventional definition of risk, the ‘management’ of risk, CAPM, EMT and the like are well known. But derivations of value investors’ actions from first principles about risk are thinner on the ground; and it is even less appreciated that value investors’ actions can be derived from simple premises and expressed in terms of a straightforward ‘risk management’ framework.

This circular sets out three of these premises. Part II, dated 1 October, derives from them a ‘risk management’ framework which generates some counter-intuitive (and, for adherents to the conventional definition of risk and practice of risk management, startling and disconcerting) results. Part III, dated 15 October, applies this framework to a hypothetical series of investment decisions; and Part IV, dated 1 November, uses the framework to summarise Graham-and-Dodd investors’ approach to the allocation of investment capital. Considered as a whole, the four circulars encapsulate Leithner & Co.’s conception of investment risk and practice of ‘investment risk management.’

Back to Top

Premise #1: The Nature of Investment Risk

To invest is to exchange one asset (such as cash) for another (a common stock, bond, title to real estate, etc.) at an agreed price. The ownership of an asset confers the right to receive a stream of earnings which that asset is expected – but not guaranteed – to generate. To invest is therefore to make a number of decisions: is it reasonable to expect that Asset A (which Bloggs owns) will produce a larger or more secure stream of earnings than Asset B (which I own)? Under what circumstances does it make sense to sell B to Bloggs? To buy A from him? At what prices?

These decisions necessarily rest upon analyses (whether conscious or intuitive) of an asset’s past and present ability to produce earnings. They are also based upon assumptions (implicit or explicit, upbeat or downcast) about the asset’s ability to generate a stream of earnings into the future. But one’s theories can be false, one’s logic invalid and one’s evidence unreliable. Accordingly, analyses about the past and present are necessarily imperfect. Further, the future is inherently uncertain: at best it is only roughly foreseeable and not infrequently it is completely unpredictable. Thus the correspondence between the calculations which one makes today and results which they achieve in several years’ time are at best flawed and at worst non-existent. Whether their magnitude is small or large, in other words, entrepreneurs, investors and other decision-makers will inevitably make mistakes.

This point applies not just to decision-makers considered in isolation: it applies with equal force to their interactions. The exchange of one asset for another occurs in the expectation that it will over time produce a financial benefit. It occurs only if each party to the exchange perceives that the (subjective) value of what he receives exceeds the (objective) price that he pays. Clearly, one or both of the parties’ perceptions may be mistaken. Risk thereby unavoidably accompanies any decision to invest.

A very large academic literature, written at a very high level of technical sophistication, has been devoted to the nature, measurement and ‘management’ of risk. Stripped of its complexities, risk is the probability that our decisions do not yield their expected results (“good things”) and instead produce undesirable, unforeseen and unintended consequences (“bad things”). Investment risk is the likelihood that our investment decision(s) cause a particular “bad thing” – relative or absolute financial loss – to occur.

Implicit in this definition of risk – which, not incidentally, has nothing to do with the short-term volatility of an investment’s market price – are analytic (estimation of probability), empirical (observation of outcome) and normative (evaluation of an outcome as desirable or undesirable) aspects. Perhaps a particular investment decision over-estimates an asset’s ability to generate a stream of earnings of a given size (or, indeed, of any size); perhaps we pay too much for those earnings; or perhaps we commit some combination of these two errors. If so, then we reduce our rate of return below ‘acceptable’ levels.

It follows that the key to an appreciation of risk – and hence to the ability to make justifiable decisions in the face of inherent uncertainty – is the ability to do two things. First, one must partition total risk (i.e., the overall likelihood that bad things might occur) into a set of specific risks (i.e., the individual bad things that might occur). Second, one must prioritise these specific risks.

Specific risks typically ‘compete’ with one another. To sell all of one’s shares and bury the cash derived therefrom in the back garden, for example, eliminates the likelihood that one type of ‘bad thing’ (i.e., the bankruptcy of the underlying businesses, evaporation of one’s stream of earnings and total loss of one’s capital) occurs. To take this decision, however, necessarily creates the possibility that other ‘bad things’ occur (e.g., that the cash is stolen, that a pile of $x of cash will over time generate a smaller stream of earnings than a portfolio of $x of common stocks, etc.). The challenge, whether one is investing or making any other decision with imperfect information, is objectively to identify the total risk at hand and subjectively to choose a ‘tolerable’ specific risk(s) among a set of competing risks.

Back to Top

Premise #2: ‘Frequentist’ and ‘Subjective’ Probabilities

If risk is the probability that ‘bad things’ occur (either individually or in their totality), how does one estimate these probabilities?Some risks can be calculated with reasonable (and in some fields high) degrees of retrospective accuracy. Their calculation depends upon large amounts of valid and reliable data and long periods of comparable past experience. Their calculation also presupposes agreement with respect to auxiliary assumptions and mathematical techniques. Notable examples are the theft of property and the incidence of injury, mortality and morbidity – and hence insurance rates. Indeed, the very existence of the life, casualty and property insurance industries is predicated upon the existence of large numbers of relatively uniform events such as deaths and motor car crashes. Frequentist probabilities, based upon mathematical reasoning about the frequency of the past occurrence of large numbers of comparable events, are the bread-and-butter of the actuarial and insurance industries.

Other risks, however, can either be estimated only with far higher degrees of subjectivity or cannot be calculated at all. Several reasons underlie this difficulty: the events under consideration are regarded as unique rather than general and comparable; they may be difficult to observe and record accurately; or they may occur very rarely. For these or other reasons, valid and reliable data from long periods of past experience may not be available. Further, consensus with respect to auxiliary assumptions and appropriate mathematical models with which to calculate the probabilities of loss may not exist. Earthquakes, corporate bankruptcies and business operations more generally are examples. Hence it is much easier to insure your life against loss than it is to insure your business against bankruptcy. In the absence of frequentist probabilities, subjective probabilities must suffice. It is no accident, it seems to me, that Berkshire Hathaway’s remarkable success over the years stems to a significant extent from the ability of its senior executives not only to calculate probabilities of loss accurately but also to price them profitably.

Back to Top

Premise #3: The Law of Large Numbers

The Law of Large Numbers, which links theoretically-derived probabilities to their actual frequency of occurrence in the real world, is a pillar of statistical theory. It tells us that in an infinitely large number of repeated, independent trials of a given event, the actual frequency with which we observe an event will coincide with the theoretical frequency of its occurrence. As a trivial example, if our ‘event’ consists in 500 tosses of a fair coin and we repeat this event a very large number of times under identical conditions, then the average observed frequency of ‘heads’ will approach 250, the average observed percentage of heads will approximate 50% – and the long-run probability of observing heads will approach 0.5. Further, the greater the number of comparable tosses which comprise an event and the greater the number of times we repeat the event the more closely our observed results will approximate their theoretical frequencies and frequentist probabilities.

Two key implications follow. First, if one repeatedly takes ‘bad risks’ (e.g., crosses the Nullabor without water in January or climbs Mt Kosciusko in shorts in June), plays unfavourable games (e.g., buys lottery tickets, bets at the races, plays in casinos) or undertakes unethical or illegal practices, then – although the result on any given occasion is uncertain and need not produce a loss – it is likely that a ‘bad’ outcome will eventually be incurred. Indeed, the more often and longer these actions are undertaken, the greater the likelihood that a loss will eventually be incurred. Second, if one repeatedly takes ‘good’ risks or undertakes ‘good’ practices then over time desired results will be achieved and the losses borne along the way will tend to be relatively small.

Back to Top

The Argument Thus Far

To invest is to use imperfect and incomplete information in order to choose among competing financial risks in the face of an inherently uncertain future. One invests in order to obtain long-term financial benefits. Accordingly, investment risk has nothing whatever to do with the short-term ebb and flow of a security’s market price. Rather, it has everything to do with the likelihood that one’s investment decision will produce absolute or relative financial loss. Risk therefore inheres in any decision to allocate investment capital.

The assessment of risk is to a significant extent a ‘subjectivist’ matter. This most assuredly does not suggest that its assessment is arbitrary. Instead, it means that individuals (whose goals and preferences differ) will identify different specific risks and prioritise them in different ways. Some, in other words, are more averse than others to a specific risks. It also means that some – as indicated by the records over many years of “The Superinvestors of Graham-and-Doddsville” – possess a particularly well-developed ability to make justifiable decisions in light of uncertainty. More generally, Graham-and-Dodd value investors understand implicitly that if one repeatedly takes ‘good’ risks, undertakes ‘good’ practices, avoids ‘bad’ risks and eschews ‘bad’ practices, then one is likely to achieve ‘good’ results and mitigate the impact of ‘bad’ results.

...continued in Part II

Circular 18
Contact Us

Back to Top

Designed & maintained by
Artist Web Design
©1999-2013 All Rights Reserved