Security Analysis

Analysis connotes the careful study of available facts with the attempt to draw conclusions therefrom based on established principles and sound logic — it is part of the scientific method. But in applying analysis to the field of securities we encounter the serious obstacle that investment is not a pure science, and the future is always uncertain.

Intrinsic value was once thought to mean book value; that view proved worthless in practice because market prices showed no tendency to be governed by book value. The concept then shifted to earning power — but ‘earning power’ must imply a fairly confident expectation of certain future results, and that confidence is rarely warranted.

The concept of intrinsic value as applied to security analysis is flexible, not definite. It may be more or less distinct depending on the case; the degree of indistinctness may be expressed by a ‘range of approximate value.’ What matters is whether the price is clearly too low or too high relative to that range — not a precise calculation.

The analyst has three functions: descriptive (marshaling the facts), critical (detecting discrepancies between reported and true conditions), and evaluative (reaching judgments about value, safety, and managerial policy). The third function is the most difficult and the most useful.

Security analysis requires weighing both quantitative factors — the statistical exhibit of earnings, assets, dividends — and qualitative factors such as the nature of the business, the character of management, and the long-run outlook for the industry. Neither alone is sufficient.

The qualitative factors are by their nature elusive and difficult to appraise. Quantitative data, by contrast, can be checked and confirmed. Our preference therefore is to place primary reliance upon what can be measured and to treat qualitative factors as supplementary evidence or as a reason to require a larger margin of safety.

A business with a long and stable earnings record carries with it a strong presumption that its future will resemble its past. A business whose earnings are highly variable or whose competitive position is uncertain requires correspondingly more conservative treatment.

Nearly every issue might conceivably be cheap in one price range and dear in another. A corollary follows: the exclusive emphasis on the choice of the enterprise — buying ‘good companies’ without regard to price — leads to paying too high a price for a good security and must be rejected.

We propose a threefold classification of securities that cuts across the conventional bond/stock division: fixed-value investments (where the owner’s dominant interest is safety of principal and steady income), variable-value investments with an investment component, and speculative securities. These categories are not rigid — the same instrument may shift among them as price and circumstances change.

The investor’s dominant interest in a fixed-value commitment is the avoidance of loss. The return is contractually limited; the hazard is unlimited. Bond selection is therefore primarily a negative art — a process of exclusion and rejection — rather than an affirmative search for the best opportunity.

Prior to the S.E.C. regulations, less than half of industrial corporations supplied even the minimum quota of financial information required for responsible analysis. Railroads and public utilities had long been required to report adequately; industrials had not. The analyst must demand the full income account — operating revenues, depreciation, interest charges, nonoperating income, income taxes, dividends paid, and surplus adjustments — before any conclusions can be drawn.

An investment operation is one that, upon thorough analysis, promises safety of principal and a satisfactory return. Operations not meeting these requirements are speculative. The distinction depends upon the underlying facts including the element of price, not on any easy generalization about the quality of the company.

The failure properly to distinguish between investment and speculation was in large measure responsible for the disasters of the 1929–1933 collapse. The new-era theory led investors to treat common stocks as attractive irrespective of their prices — a notion that is, when examined, incredible in its absurdity, yet one that swept the country.

There are justifiable speculations as well as unjustifiable ones. The speculator who knows he is speculating, who limits his commitment to what he can afford to lose, and who demands a specific reason for expecting a profit is in a sound position. The genuine danger is the investor who drifts into speculation without recognizing that fact.

Our definition of investment is broader than most: it may conceivably be applied to stocks, to margin purchases, and to purchases made chiefly for profit appreciation — provided the operation rests upon thorough analysis and offers specific, demonstrable grounds for expecting a satisfactory result. Form does not determine character; the underlying facts do.

Safety in a bond is measured not by specific lien or other contractual rights, but by the ability of the issuer to meet all of its obligations. A first-mortgage bond of a weak enterprise is not a satisfactory investment; deficient safety cannot be compensated for by an abnormally high coupon rate.

Bonds should be bought on a depression basis. Any bond can do well when conditions are favorable; it is only under the acid test of depression that the advantages of strong issues and the weaknesses of unsound ones are revealed. The question is not whether an industry is exempt from depression but how severely it is affected — and whether even its worst-case performance would still service the debt.

The third principle is that diversification by itself is not a substitute for individual quality. Spreading risk among many second-grade issues does not transform them into first-grade ones. A portfolio of bonds that each individually fail the depression test will collectively fail it as well.

Bond selection is a process of exclusion first. The analyst must establish a list of disqualifying conditions — inadequate earnings coverage, excessive funded debt relative to assets, lack of adequate information, unsatisfactory trend — and reject any issue that fails any test, however attractive its yield may appear.

The traditional threefold classification of enterprises — railroads, public utilities, industrials — reflects real differences in inherent stability. A railroad’s fixed charges were once thought to be almost automatically covered; the depression shattered that assumption. We now require that a railroad’s gross revenues, not merely net income, be measured against its fixed charges over a seven-year period including depression years.

Industrial bonds present greater hazards than railroad or utility bonds of the same nominal quality, because the industrial corporation’s business may change radically, competitive conditions may shift, and management quality is harder to assess from outside. The compensation for this greater uncertainty must be a substantially larger margin of safety — earnings coverage of five times or more — not merely a higher coupon.

Large size is a trait of considerable advantage when dealing with exceptionally unfavorable developments. The evidence from the 1930–1933 depression suggests that investment in industrial bonds should be restricted largely to major companies; the smaller enterprise, however strong its recent record, lacks the financial staying power to endure a prolonged contraction.

Working-capital requirements and sinking-fund provisions are not mere technicalities — they are substantive protections that may determine whether a company can remain solvent during a period when it cannot refinance its debt. An analyst who ignores these protective covenants is ignoring the architecture of the commitment.

The preferred stockholder occupies a contractually weaker position than the bondholder, yet the preferred dividend is treated by investors as though it carried the same certainty as a bond coupon. This is an illusion. The preferred dividend may be omitted without default; the arrears accumulate but there is no mechanism forcing payment. A preferred stock that an investor treats as equivalent to a bond is a preferred stock he does not properly understand.

An investment preferred issue must meet all the requirements of a good bond — plus an extra margin of safety to offset its contractual disadvantages. The technique of analyzing a senior stock issue is therefore essentially the same as that for bonds, but the minimum earnings coverage required is higher, not lower.

The market has repeatedly demonstrated that it does not price this contractual disadvantage correctly. Preferred stocks selling near par in good times have declined catastrophically in bad times because investors failed to apply the depression test. The United Cigar Stores Preferred is a vivid example: seven times coverage in 1928, yet lease obligations drove it to near-worthlessness within a few years.

We suggest that the minimum coverage required for preferred-stock investment should be set materially higher than the standards hitherto accepted as adequate. Our requirements would have disqualified a large part of preferred-stock financing done before 1931; such severity would have benefited the investing public enormously.

Convertible bonds and preferred stocks present a combination of fixed-income protection and equity participation. For the investor, they are attractive only if they first qualify as sound fixed-value investments on their own terms. An issue that fails the bond test is not rescued by its conversion privilege; it is a speculative issue wearing a fixed-income disguise.

The proper policy for an investor who holds a convertible is to ignore the conversion privilege so long as it remains out of the money, and to treat the security as a straight bond. When the conversion privilege moves into the speculative range — when the issue is selling well above par because of a rise in the common — the investor should decide independently whether to own the common stock, not cling to the convertible out of inertia.

Subscription warrants attached to bonds or preferred stocks represent a device that systematically favors the issuer over the investor. The speculative element — the warrant — tends in bull markets to trade at large premiums that have no rational basis, and in bear markets it collapses entirely. The investor who buys warrant-bearing issues for their ‘extra kicker’ is paying for the warrant at a price he would never pay if offered the warrant separately.

Where manipulation of accounts is found, stock juggling will be found in some form as well. Familiarity with the methods of questionable finance — inflated earnings through subsidiary manipulation, manufactured deficits to create future gains, excessive write-downs to boost later reported results — is among the most practically valuable things a securities analyst can acquire.

Prior to the World War, the standard approach to common-stock investment rested on the concept of a stable average earning power that could be capitalized at a conservative multiple. By 1927–1929, a new-era theory had displaced this: value depended not on current or average earnings but on the trend of earnings into the indefinite future, and a good stock was attractive at any price.

The new-era theory contained two fatal weaknesses. First, it abolished the distinction between investment and speculation by declaring that good common stocks were sound investments regardless of price. Second, it rested the case for high valuations on a historical record of earnings that the very act of bidding up prices had already rendered obsolete — using past above-average returns to justify prices that made such returns impossible.

The three propositions of new-era doctrine — that value depends on future earnings, that good stocks have rising earnings trends, and that good stocks will prove profitable investments — sound innocent and plausible. Yet they led directly to the conclusion that no price could be too high for a good stock. The new-era theory was simply the old cynical epigram repackaged: ‘Investment is successful speculation.’

The subsequent depression dissipated the concept of permanently strong corporations. We are back to the older realization that time brings unpredictable changes in the fortunes of business undertakings. The proper response to this uncertainty is not to abandon quantitative standards but to apply them more rigorously, demanding a larger margin of safety precisely because the future is less predictable than the new-era theorists believed.

With the prewar approach obsolete and the new-era approach exploded, we must establish a new set of logically sound principles. The starting point is that an investment in common stock can be justified only on both qualitative and quantitative grounds simultaneously. Neither a good business alone nor a low statistical price alone suffices; both must be present.

We suggest that about 20 times average earnings is as high a price as can be paid in an investment purchase of a common stock. This is not a formula but a limit — the boundary beyond which the element of speculation unavoidably enters, however attractive the enterprise. A higher price may prove profitable, but it will have proved a wise or fortunate speculation, not a sound investment.

For common stocks, the earnings record to be used in computing the multiplier should be the average over a period covering at least a full business cycle — ideally seven to ten years — not the current or most recent year’s figure. Current earnings are governed more by the market level than by long-term value, which accounts in good part for the wide and irrational fluctuations in common-stock prices.

The capitalization structure of a company — the division between senior securities and common stock — has an important bearing on the significance of earnings per share. A heavily leveraged company whose total earnings are barely adequate to cover its fixed charges will produce wildly amplified EPS changes on small swings in total earnings. Such ‘speculative’ capitalization structures throw all the company’s securities outside the pale of investment.

The analyst must approach the income account from three angles simultaneously: the accounting question (what are the true earnings?), the business question (what does the earnings record reveal about future earning power?), and the valuation question (what standards must be applied to arrive at a reasonable value?). Wall Street’s naïve acceptance of reported earnings per share is one of the most dangerous habits in investment practice.

Reported earnings are frequently distorted by arbitrary depreciation and amortization charges, by the treatment of nonrecurring items as ordinary income or vice versa, by the inclusion or exclusion of subsidiary results, and by charges to surplus that would properly belong in the income account. The analyst’s first task is to restate the income account on a basis that reflects economic reality.

Where manipulation of accounts is found, stock juggling will be found also in some form or other. The balance sheet provides a check on the income statement: if reported earnings are genuine, they must be reflected in growing cash or other assets, or in growing equity, after allowing for dividends. Discrepancies between the income statement and the balance sheet are among the most reliable signals of financial irregularity.

Extraordinary charges to surplus — massive write-downs in a bad year that have the effect of reducing future depreciation charges and thus inflating future earnings — are ‘manufactured earnings.’ The analyst must reconstruct the actual charge against income in each year and refuse to accept the management’s arbitrary choice of what belongs in the income account versus the surplus account.

The current-asset value of a common stock — current assets minus all liabilities and claims ahead of the common — is more practically useful than book value, which includes fixed assets that may realize very little in liquidation. When common stocks sell for less than their net current assets, a serious error is being committed somewhere: either in the market’s judgment, in management policy, or in the stockholders’ passivity about their own property.

A large number of common stocks sell for less than their net current asset value. This widespread divergence between price and liquidating value is a comparatively recent development; in the 1921 depression the proportion was quite small, suggesting that the 1932–1938 phenomenon reflects excessive pessimism rather than accurate appraisal of the underlying businesses.

Common stocks selling well below liquidating value represent on the whole a class of undervalued securities. They have declined more severely than the actual conditions justify. Nevertheless, the analyst should exercise discrimination: a company that is persistently losing money is consuming the very asset values that make the stock appear cheap, and patience without a catalyst is not an investment strategy.

The liabilities on a balance sheet are real; the assets must be questioned. In calculating liquidating value, all true liabilities must be deducted at their face amount, while the realizable value of assets must be estimated conservatively — receivables at a discount, inventories at a larger discount, and fixed assets at their likely forced-sale value rather than their carrying value.

It is traditional that a corporation should be run for the benefit of its owners, the stockholders. We now know, however, that a controlling group may hold the power to divert profits to themselves. The extensive separation of ownership and control raises a new and troubling set of problems that security analysis cannot ignore.

The typical investor prefers to have his dividend today rather than to have the company retain the earnings on his behalf. Yet management’s discretionary power over dividend policy is rarely challenged. Stockholders have accepted ‘conservative dividend policies’ in a spirit that is peculiarly perfunctory — accepting what they do not want because they have been told it is good for them.

The discretionary power over dividend policy may be abused in sinister fashion: sometimes to facilitate the acquisition of shares at an unduly low price, at other times to enable insiders to unload at high quotations. An analyst who fails to examine the relationship between a company’s earnings and its dividend disbursement policy is ignoring one of the most important dimensions of the common stockholder’s situation.

Management has a responsibility to prevent, insofar as it is able, both absurdly high and unduly low prices for its securities. The sanctimonious attitude of executives who profess not even to know the market price of their shares deserves no patience — in many cases they have a vital personal interest in those very prices, and at times they have been material participants in their manipulation.

Market analysis — forecasting future price movements from charts, indices, or near-term business prospects — seems easier than security analysis, and its rewards may appear to materialize more quickly. For these very reasons, it is likely to prove more disappointing in the long run. There are no dependable ways of making money easily and quickly, either in business or in Wall Street.

Chart reading cannot be a science: if its conclusions were dependable, everyone would predict tomorrow’s prices correctly, and hence everyone could make money continuously from the market. This is a logical impossibility. The belief that chart patterns recur in predictable ways is not confirmed by any evidence that would survive scientific scrutiny.

The critical difference between security analysis and market analysis is the margin of safety. The securities analyst who buys a stock well below its calculated intrinsic value can be wrong about the exact value and still not lose money. In market analysis there are no margins of safety; you are either right or wrong, and if you are wrong you lose money. The two activities are not equivalent alternatives.

Security analysis is skeptical of the ability of the analyst to forecast the near-term behavior of individual issues with any useful degree of accuracy. More satisfactory results are obtained by confining attention to cases where the price — relative to fundamental value — is so far out of line that the conclusion of undervaluation can be drawn with high confidence, even allowing for a wide range of uncertainty about the future.