Benjamin Graham on Capital Allocation

Western markets, selling at a substantial discount. As the Japanese had no prejudice against these bonds, his back office was inundated with reams of documents. The typical purchase of $100,000 face amount would usually result in the appearance of one thousand separate bonds. The special safe deposit box for these bonds was known, not too favorably, as the "Ben Graham" box. After two years, the Fujimoto Bank set up its own New York office, with Miki in charge, to buy these bonds. Two other Japanese banking firms then became customers and made up for some of the lost business. Ben's main work was in handling an inquiries about security lists or individual issues. He was given an assistant, Leo Stern, later a senior partner in the firm and the father of Walter P. Stern-whose own distinguished career has included terms as President of The Financial Analysts Federation and of The Institute of Chartered Financial Analysts. Periodically, they issued "circulars" analyzing one or more securities in detail. For example, in May of 1921 they recommended the sale of the U. S. Victory 4%'s due in 1923 and selling at 97% and reinvestment in the U. S. 4%'s of 1938 then selling at 87Y2. They believed that the then high level of interest rates would subside and thus the longer term bonds had better appreciation po~sibilities. This circular was advertised in the newspapers under the title "Memorandum to Holders of Victory Bonds."

Ben devised a series of formulas to work back from three items-taxes, pretax income, and the property account-to determine how much of the property account was in the goodwill category. These findings were the basis for an article in The Magazine of Wall Street. Editor Powers said: "Ben, nobody around here can make head or tail of your formulas. It looks as if you've done the whole thing with mirrors. But, we'll publish it anyway." Although the published figures available could have been misleading, Ben's computations proved remarkably correct. The accuracy of his calculations was not publicly available for many years-until most corporations finally started to write off the more imaginary intangibles embedded in their balance sheets. By then, earning power had begun to become the most significant factor affecting a stock's price and asset values were much less important. Ben's computations, for example, revealed that all the $508 million par value of the U. S. Steel common stock and even a good part of its $360 million of preferred had originally been "water." Subsequently U. S. Steel wrote down $769 million of "goodwill" and similar intangibles by using many years of retained earnings. Word of Ben's success with arbitrage and hedging operations spread, and several clients opened accounts that allowed him, as sole manager, a 25 percent share in the cumulative net profits.standard

The investor with a portfolio of sound stocks should expect their prices to fluctuate and should neither be concerned by sizable declines nor become exC£ted by sizable advances. He should always remember that market quotations are there for his convenience, either to be taken advantage of or to be ignored. Sound generalz"zations can be more dangerous than unsound ones because they lure more people into unwarranted actions. The Intelligent Investor Third Edition, 1959 The post-World War II world has been characterized as 'brave' and 'new.' Brave it is, indeed, but we are not positive that it is equally new. We can be skeptical about a complete break with the past. Security Analysis Fourth Edition, 1962 Common stocks have one important investment characteristic and one important speculative characteristic. Their investment value and average market price tend to increase irregularly but persistently over the decades, as their net worth builds up through the reinvestment of undistributed earnings . ... However, most of the time common stocks are subject to irrational and excessive price fluctuations in both directions, as the consequence of the ingrained tendency of most people to speculate or gamble-i.e., to give way to hope, fear and greed. Financial Analysts Iournal September/October, 1976

The 1975 article formalises Graham's long-standing fifty-fifty bond-stock rule, with an explicit adjustment band. Graham recommends that the defensive investor hold a roughly equal split between high-grade bonds and a diversified list of common stocks, and rebalance when the mix drifts by more than five percentage points in either direction. The rebalancing discipline forces the investor to take profits from the asset that has risen and to add to the asset that has fallen — the opposite of what momentum would dictate. Graham's argument for the fifty-fifty rule is not theoretical optimality. He concedes that an investor who could correctly forecast the better-performing asset would do better by tilting toward it. His case is that few investors can forecast reliably, and that the mechanical rebalancing discipline captures the only edge most investors can credibly claim: the willingness to sell what has gone up and buy what has gone down, against the crowd. The 1975 article also introduces a valuation-conditioned variant: when the price-to-central-value ratio is well below one, the investor may hold up to seventy-five percent in equities; when it is well above one, the investor may reduce to twenty-five. Graham's working band is therefore twenty-five to seventy-five percent in equities, with the central case at fifty. The valuation-conditioned band preserves the discipline of the fifty-fifty rule while allowing the investor to act on Graham's view that the broad market can be visibly cheap or dear at long intervals.

The Graham-Newman letters to shareholders, written between 1946 and 1958, lay out a stable operating policy: the partnership purchased securities at prices below their intrinsic value as measured by asset coverage, earning power, or contractual claim, and sought to realise the discount through liquidation, distribution, or market re-rating. Graham-Newman classified positions into three buckets — undervalued common stocks, special situations with a catalyst, and arbitrage or workout positions with a defined timeline. The letters record that Graham-Newman pursued each bucket with a distinct discipline. Common stocks had to trade at a discount to net current assets or to conservatively capitalised earning power; special situations required an identifiable catalyst such as a merger, recapitalisation, or litigation resolution; arbitrage positions required a small gross spread per unit but a high annualised return because the time to resolution was short. Graham-Newman reported the three buckets separately to shareholders so the sources of return could be tracked. Graham-Newman's reporting discipline is itself a contribution of the letters. The partnership did not simply report aggregate return; it reported the components of return by category, the realised and unrealised portions separately, and the net asset value per share on a transparent mark. This reporting standard, well before the modern hedge-fund era, made the partnership's record auditable and reproducible — a discipline Graham regarded as part of the analyst's obligation to his capital.

The 1956 closing letter is the document in which Graham-Newman announced that the partnership would be wound up and its remaining positions distributed to its shareholders. The letter is, in part, a record of the partnership's analytical method over the partnership's life, and in part a reflection on the conditions under which the method had become harder to apply. Graham and Newman record that the partnership's undervalued common-stock category had become harder to find candidates for, because the bull market of the early 1950s had narrowed the universe of structural discounts to a degree the partnership regarded as durable. The closing letter's analytical contribution is to make explicit the partnership's view that the analytical edge the partnership had exploited since 1936 had, by 1956, narrowed to a degree that justified winding the partnership up and returning the remaining capital to the partners who had supplied it. The letter's discussion of the partnership's analytical edge is its most instructive passage. Graham and Newman record that the partnership's working method had been built around the identification of securities whose market prices were below the working-capital value of the underlying businesses, and that the bull market had lifted most of the structural discounts above the working-capital floor at which the analytical edge had been identified. The letter's instruction is that the partnership's analytical method is a feature of a particular market condition, and that the method's edge narrows as the market condition changes. The letter is candid that the partnership does not regard the method as obsolete; the partnership regards the method as appropriate to the market condition the partnership was founded to exploit, and as inappropriate to the market condition the partnership finds itself in by 1956. The closing of the partnership is the candid acknowledgment that the method has outlived the condition that made it pay. The closing letter's other instructive passage is the partnership's discussion of the distribution of its remaining positions. Graham and Newman record that the partnership's remaining undervalued holdings, the special-situations positions, and the arbitrage positions would be distributed to the partnership's shareholders in specie, and that the shareholders would then hold the positions directly rather than through the partnership's vehicle. The letter's instruction is that the partnership's analytical method was always a feature of the partnership's working organization, and that the method's continued application after the partnership's closing would be the shareholders' responsibility rather than the partnership's. The 1956 closing letter is, in this sense, the document in which Graham-Newman's working method is most clearly shown to be the analytical discipline of the partners who applied it, and the closing of the partnership is the partners' candid acknowledgment that the analytical discipline had outlived the market condition that had made it pay.

The 1955 Graham-Newman annual report, the partnership's report for the year ended January 31, 1955, is the document in which the partnership records its most explicit discussion of the difficulty of finding bargains in a rising market. The report notes that the bull market of the early 1950s had narrowed the universe of undervalued common stocks, and that the partnership's undervalued category had become a smaller fraction of the partnership's total assets as the market's re-rating of the structural discounts had proceeded. The 1955 report's analytical contribution is to make explicit the tension between the partnership's analytical method and the market's general direction: in a rising market, the structural discounts the partnership identifies close more quickly, and the partnership's analytical edge is correspondingly reduced. The report is candid that the narrowing of the edge is the analytical wage for the market's advance, and that the partnership must accept the narrowing as the structural condition of a bull market. The report's discussion of the difficulty of finding bargains is its most instructive passage. Graham-Newman records that the partnership's undervalued positions had been bought at prices the partnership estimated to be below the working-capital value of the underlying businesses, and that the rising market had lifted many of the partnership's holdings above the working-capital floor at which the analytical edge had been identified. The report's instruction is that the partnership's analytical edge in the undervalued category is not a permanent feature of the market; it is a feature of the market's particular condition at the time of purchase, and the partnership must accept that the edge narrows as the market rises. The report is candid that the narrowing of the edge is the analytical wage for the market's general advance, and that the partnership cannot expect to find the same density of structural discounts in a market that has already re-rated them away. The 1955 report's other instructive passage is the partnership's discussion of the special-situations category as the partnership's hedge against the narrowing of the undervalued category. Graham-Newman records that the special-situations positions, held for the closing of defined catalysts, were less affected by the market's general direction than the undervalued positions, because the catalyst's closing was the analytical event the partnership had identified, not the market's general re-rating. The report's instruction is that the special-situations category is the partnership's structural response to the narrowing of the undervalued category in a rising market: when the structural discounts close, the partnership shifts its analytical effort to the catalysts whose closing is not dependent on the market's direction. The 1955 report is, in this sense, the document in which the partnership's working method is most clearly shown to be a response to the market's general condition, and not a fixed recipe that the partnership applies regardless of the market's state.

The 1953 Graham-Newman annual report, the partnership's report for the year ended January 31, 1953, is the document in which the partnership's special-situations category receives its most extended treatment. The report records that the partnership's special-situations positions were held for catalysts that the partnership expected to close within a defined horizon, and that the partnership's return on the category came from the closing of the catalysts rather than from the market's general direction. The 1953 report's analytical contribution is to make explicit what the partnership had been doing implicitly since its founding: the special-situations category is, in Graham-Newman's working method, the area in which the partnership's analytical edge is most directly expressed, because the catalyst is the event the partnership's analysis identified and the closing is the verification of the analysis. The category's return is, in this sense, the analytical wage for the partnership's identification of catalysts at the time of purchase. The report's discussion of the special-situations category is its most instructive passage. Graham-Newman records that the catalysts the partnership's positions were held for included mergers, reorganizations, liquidations, and recapitalizations, and that the partnership's analytical work in each case was to estimate the probability that the catalyst would close, the time horizon over which it would close, and the value the position would realize at the closing. The report's instruction is that the special-situations analyst is not forecasting the market's direction; he is forecasting the closing of a defined corporate event, and his return is the analytical wage for the probability-and-timing estimate he made at the time of purchase. The partnership's discipline in the special-situations category is to demand that the expected return on the position, weighted by the probability of the catalyst's closing within the expected horizon, exceeds the partnership's hurdle rate by a sufficient margin. The 1953 report's other instructive passage is the partnership's discussion of the relationship between the special-situations category and the broader portfolio. Graham-Newman records that the special-situations positions were held alongside the undervalued common stocks and the arbitrage positions, and that the partnership's total return was the weighted sum of the three categories' returns. The report's instruction is that the special-situations category is not a substitute for the undervalued category; it is a complement to it. The undervalued category earns the partnership the market's re-rating of the structural discounts the partnership has identified; the special-situations category earns the partnership the closing of the catalysts the partnership has identified. The two categories' returns are not perfectly correlated, and the partnership's diversification across the two categories is the structural protection against the analytical error that any single category's positions may contain. The 1953 report is, in this sense, the document in which the partnership's three-category working method is most fully defended as a portfolio-construction principle.

The 1950 Graham-Newman annual report, the partnership's report for the year ended January 31, 1950, sets out the returns the partnership earned across its three working categories and is the document in which the partnership's working method is most clearly documented. The report breaks the partnership's total return into the contributions from undervalued common stocks, from special situations, and from arbitrage positions. The breakdown is the report's analytical contribution: it allows the partnership's shareholders to see which of the three categories carried the year's return, and it allows the partnership's auditors to verify that each category's contribution reconciles to the underlying positions the partnership held through the year. The 1950 report is, in this sense, the document that most clearly distinguishes the partnership's three working categories from one another in the published record, and it is the document on which subsequent generations of value partnerships have drawn for the working method of categorical return attribution. The report's discussion of the undervalued-common-stock category is its most instructive passage. Graham-Newman records that the partnership's undervalued holdings were bought at prices the partnership estimated to be below the working-capital value of the underlying businesses, and that the partnership's return on the category came from the market's re-rating of the holdings toward their working-capital values as the cycle progressed. The point the report makes is that the return on the undervalued category was not a forecast of the market's direction; it was a return on the structural undervaluation the partnership had identified at the time of purchase. The market's subsequent re-rating was the mechanism by which the structural undervaluation closed, and the partnership's return was the analytical fruit of the identification. The report's instruction is that the undervalued category's return is the analytical wage for the partnership's identification of working-capital discounts at the time of purchase. The report's discussion of the arbitrage and special-situations categories is the report's other instructive passage. Graham-Newman records that the arbitrage positions were held for defined closings such as the completion of tender offers and the settlement of recapitalizations, and that the special-situations positions were held for defined catalysts such as mergers, reorganizations, and liquidations. The report's instruction is that the returns on the arbitrage and special-situations categories were not dependent on the market's direction; they were dependent on the closing of the catalyst. The partnership's return on these categories was, in this sense, the analytical wage for the partnership's identification of catalysts at the time of purchase, and the partnership's discipline in holding the positions through to the catalyst's closing. The 1950 report is, in this sense, the document in which Graham-Newman's working method is most clearly mapped to the returns the partnership earned in each category.

The 1949 Graham-Newman annual report, the partnership's report for the year ended January 31, 1949, was the first full-year report the partnership issued after the SEC's distribution of the GEICO position that Graham-Newman had acquired in 1948. The distribution is the report's most consequential event, and Graham-Newman's discussion of it is the report's most consequential passage. The partnership had been forced by the SEC to distribute the GEICO stock to its own shareholders, because a technicality in the Investment Company Act prohibited an investment fund from holding more than ten percent of an insurance company. The distribution, the report notes, was a forced consequence of regulatory structure rather than a free investment decision; the partnership did not sell the GEICO stake because it wanted to, but because the regulator insisted on the distribution as the price of the partnership's continued operation as a registered investment company. The report's discussion of the GEICO distribution is also the report's most analytical passage. Graham-Newman records that the cost basis of the GEICO stake was approximately seven hundred and twenty thousand dollars, that the distributed value was many multiples of that cost, and that the partnership's shareholders received, in the distribution, a stake whose market value ran into the hundreds of millions of dollars over the subsequent decades. The 1949 report does not forecast the future value of the distributed stake; the analytical point Graham-Newman makes is narrower. The partnership's investment method had identified, in GEICO, a security whose intrinsic value was substantially above the price the partnership had paid for it. The margin of safety the partnership had demanded at the time of purchase was the analytical basis on which the subsequent re-rating could occur, and the report is candid that the magnitude of the re-rating exceeded even the partnership's analytical expectations. The 1949 report also develops the categorization of the partnership's positions that Graham-Newman would use in every subsequent year. The partnership's holdings are divided into three categories: undervalued common stocks held for the market's re-rating of the underlying value; special situations held for a defined catalyst such as a merger, reorganization, or liquidation; and arbitrage positions held for a defined closing such as the completion of a tender offer or the settlement of a recapitalization. The categorization is the report's analytical contribution to the discipline of reporting: it allows the partnership's shareholders to see where the partnership's returns are coming from, and it gives the partnership's auditors a verifiable map of each position's expected catalyst. The 1949 report is, in this sense, the document in which Graham-Newman formalized the working method that the partnership would carry through the rest of its life and that subsequent generations of value partnerships would adopt as their own working discipline.

The 1948 Graham-Newman annual report, preserved in the Columbia Business School archive, sets out the partnership's reporting discipline. The report carries an audited balance sheet as of January 31, 1948, a profit-and-loss statement, and a statement of unrealised appreciation. The partnership distinguishes realised from unrealised gains explicitly, and reports net asset value per share to the dollar. The reporting discipline is itself an analytical contribution. Graham-Newman did not bury losses in unrealised accounts, nor did it report gains before they were realised. The separation of realised and unrealised returns allowed shareholders to see what the partnership had actually earned by closing positions and what it carried in mark-to-market on open positions. The discipline foreshadowed the modern hedge-fund reporting standard by several decades. The 1948 report also confirms the partnership's holdings in three categories: undervalued common stocks held for market re-rating, special situations held for a catalyst, and arbitrage positions held for a defined closing. The categorisation let shareholders assess where the partnership's returns were coming from, and gave the partnership's auditors a verifiable map of each position's expected catalyst.

Graham and Dodd devote a substantial chapter to preferred stock, and the 1940 edition's treatment is sharper than the 1934 version. They argue that the typical preferred stock is a hybrid security with the downside of a bond — no participation in earnings above the dividend — and the call feature of an equity claim. The result is that most preferreds offer bond-like returns with equity-like risk, a combination the analytical investor should approach with scepticism. The book lays out specific tests. The preferred dividend should be covered by normalised earnings several times over, the issue should be backed by a tangible asset position that meaningfully exceeds par, and the company should have a record of earning power through adverse cycles. Graham and Dodd warn that many industrial preferreds fail these tests, while many utility and railroad preferreds pass them — the analytical distinction is by financial structure, not by label. On capital structure, Graham and Dodd argue that a moderate amount of senior debt can enhance the equity's return on the senior's capital without unduly impairing safety, but that there is a clear line beyond which leverage stops adding to equity returns and starts impairing solvency. The analytical investor's job is to find companies whose capital structure sits on the productive side of that line, and to avoid companies whose leverage has crossed into fragility.