Warren Buffett on Look-Through Earnings
All the passages below are taken from the book, “The Essays of Warren Buffett, Lessons for Corporate America, Third Edition.” It was selected and arranged by Lawrence A Cunningham and published in 2013.
When one company owns part of another company, appropriate accounting procedures pertaining to that ownership interest must be selected from one of three major categories. The percentage of voting stock that is owned, in large part, determines which category of accounting principles should be utilized.
Generally accepted accounting principles require (subject to exceptions, naturally . . . ) full consolidation of sales, expenses, taxes, and earnings of business holdings more than 50% owned. Blue Chip Stamps, 60% owned by Berkshire Hathaway Inc., falls into this category. Therefore, all Blue Chip income and expense items are included in full in Berkshire's Consolidated Statement of Earnings, with the 40% ownership interest of others in Blue Chip's net earnings reflected in the Statement as a deduction for "minority interest".
Full inclusion of underlying earnings from another class of holdings, companies owned 20% to 50% (usually called "investees"), also normally occurs. Earnings from such companies---for example, Wesco Financial, controlled by Berkshire but only 48% owned---are included via a one-line entry in the owner's Statement of Earnings. Unlike the over-50% category, all items of revenue and expense are omitted; just the proportional share of net income is included. Thus, if Corporation A owns one-third of Corporation B, one-third of B's earnings, whether or not distributed by B, will end up in A's earnings. There are some modifications, both in this and the over-50% category, for intercorporate taxes and purchase price adjustments, the explanation of which we will save for a later day. (We know you can hardly wait.)
Finally come holdings representing less than 20% ownership of another corporation's voting securities. In these cases, accounting rules dictate that the owning companies include in their earnings only dividends received from such holdings. Undistributed earnings are ignored. Thus, should we own 10% of Corporation X with earnings of $10 million in 1980, we would report in our earnings (ignoring relatively minor taxes on intercorporate dividends) either (a) $1 million if X declared the full $10 million in dividends; (b) $500,000 if X paid out 50%, or $5 million, in dividends; or (c) zero if X reinvested all earnings.
We impose this short-and over-simplified-course in accounting upon you because Berkshire's concentration of resources in the insurance field produces a corresponding concentration of its assets in companies in that third (less than 20% owned) category. Many of these companies pay out relatively small proportions of their earnings in dividends. This means that only a small proportion of their current earning power is recorded in our own current operating earnings. But, while our reported operating earnings reflect only the dividends received from such companies, our economic well-being is determined by their earnings, not their dividends.
Our holdings in this third category of companies have increased dramatically in recent years as our insurance business has prospered and as securities markets have presented particularly attractive opportunities in the common stock area. The large increase in such holdings, plus the growth of earnings experienced by those partially-owned companies, has produced an unusual result; the part of "our" earnings that these companies retained last year (the part not paid to us in dividends) exceeded the total reported annual operating earnings of Berkshire Hathaway. Thus, conventional accounting only allows less than half of our earnings "iceberg" to appear above the surface, in plain view. Within the corporate world such a result is quite rare; in our case it is likely to be recurring.
Our own analysis of earnings reality differs somewhat from generally accepted accounting principles, particularly when those principles must be applied in a world of high and uncertain rates of inflation. (But it's much easier to criticize than to improve such accounting rules. The inherent problems are monumental.) We have owned 100% of businesses whose reported earnings were not worth close to 100 cents on the dollar to us even though, in an accounting sense, we totally controlled their disposition. (The "control" was theoretical. Unless we reinvested all earnings, massive deterioration in the value of assets already in place would occur. But those reinvested earnings had no prospect of earning anything close to a market return on capital.) We have also owned small fractions of businesses with extraordinary reinvestment possibilities whose retained earnings had an economic value to us far in excess of 100 cents on the dollar.
The value to Berkshire Hathaway of retained earnings is not determined by whether we own 100%, 50%, 20% or 1% of the businesses in which they reside. Rather, the value of those retained earnings is determined by the use to which they are put and the subsequent level of earnings produced by that usage. This is true whether we determine the usage, or whether managers we did not hire---but did elect to join---determine that usage. (It's the act that counts, not the actors.) And the value is in no way affected by the inclusion or non-inclusion of those retained earnings in our own reported operating earnings. If a tree grows in a forest partially owned by us, but we don't record the growth in our financial statements, we still own part of the tree.
Our view, we warn you, is non-conventional. But we would rather have earnings for which we did not get accounting credit put to good use in a 10%-owned company by a management we did not personally hire, than have earnings for which we did get credit put into projects of more dubious potential by another management---even if we are that management. 
The term "earnings" has a precise ring to it. And when an earnings figure is accompanied by an unqualified auditor's certificate, a naive reader might think it comparable in certitude to TT, calculated to dozens of decimal places.
In reality, however, earnings can be as pliable as putty when a charlatan heads the company reporting them. Eventually truth will surface, but in the meantime a lot of money can change hands. Indeed, some important American fortunes have been created by the monetization of accounting mirages.
Berkshire's own reported earnings are misleading in a different, but important, way: We have huge investments in companies ("investees") whose earnings far exceed their dividends and in which we record our share of earnings only to the extent of the dividends we receive. The extreme case is Capital Cities/ABC, Inc. Our 17% share of the company's earnings amounted to more than $83 million last year. Yet only about $530,000 ($600,000 of dividends it paid us less some $70,000 of tax) is counted in Berkshire's GAAP earnings. The residual $82 million-plus stayed with Cap Cities as retained earnings, which work for our benefit but go unrecorded on our books.
Our perspective on such "forgotten-but-not-gone" earnings is simple: The way they are accounted for is of no importance, but their ownership and subsequent utilization is all-important. We care not whether the auditors hear a tree fall in the forest; we do care who owns the tree and what's next done with it.
When Coca-Cola uses retained earnings to repurchase its shares, the company increases our percentage ownership in what I regard to be the most valuable franchise in the world. (Coke also, of course, uses retained earnings in many other value-enhancing ways.) Instead of repurchasing stock, Coca-Cola could pay those funds to us in dividends, which we could then use to purchase more Coke shares. That would be a less efficient scenario: Because of taxes we would pay on dividend income, we would not be able to increase our proportionate ownership to the degree that Coke can, acting for us. If this less efficient procedure were followed, however, Berkshire would report far greater "earnings."
I believe the best way to think about our earnings is in terms of "look-through" results, calculated as follows: Take $250 million, which is roughly our share of the 1990 operating earnings retained by our investees; subtract $30 million, for the incremental taxes we would have owed had that $250 million been paid to us in dividends; and add the remainder, $220 million, to our reported operating earnings of $371 million. Thus our 1990 "look-through earnings" were about $590 million. 
In our view, the value to all owners of the retained earnings of a business enterprise is determined by the effectiveness with which those earnings are used---and not by the size of one's ownership percentage. If you have owned .01 of 1% of Berkshire during the past decade, you have benefited economically in full measure from your share of our retained earnings, no matter what your accounting system. Proportionately, you have done just as well as if you had owned the magic 20%. But if you have owned 100% of a great many capital-intensive businesses during the decade, retained earnings that were credited fully and with painstaking precision to you under standard accounting methods have resulted in minor or zero economic value. This is not a criticism of accounting procedures. We would not like to have the job of designing a better system. It's simply to say that managers and investors alike must understand that accounting numbers are the beginning, not the end, of business valuation.
In most corporations, less-than-20% ownership positions are unimportant (perhaps, in part, because they prevent maximization of cherished reported earnings) and the distinction between accounting and economic results we have just discussed matters little. But in our own case, such positions are of very large and growing importance. Their magnitude, we believe, is what makes our reported operating earnings figure of limited significance. 
Within [the] gigantic auction arena [composed of the entire array of major American corporations], it is our job to select businesses with economic characteristics allowing each dollar of retained earnings to be translated eventually into at least a dollar of market value. Despite a lot of mistakes, we have so far achieved this goal. In doing so, we have been greatly assisted by Arthur Okun's patron saint for economists---St. Offset. In some cases, that is, retained earnings attributable to our ownership position have had insignificant or even negative impact on market value, while in other major positions a dollar retained by an investee corporation has been translated into two or more dollars of market value. To date, our corporate over-achievers have more than offset the laggards. If we can continue this record, it will validate our efforts to maximize "economic" earnings, regardless of the impact upon "accounting" earnings.
We also believe that investors can benefit by focusing on their own look-through earnings. To calculate these, they should determine the underlying earnings attributable to the shares they hold in their portfolio and total these. The goal of each investor should be to create a portfolio (in effect, a "company") that will deliver him or her the highest possible look-through earnings a decade or so from now.
An approach of this kind will force the investor to think about long-term business prospects rather than short-term stock market prospects, a perspective likely to improve results. It's true, of course, that, in the long run, the scoreboard for investment decisions is market price. But prices will be determined by future earnings. In investing, just as in baseball, to put runs on the scoreboard one must watch the playing field, not the scoreboard. 
The primary test of managerial economic performance is the achievement of a high earnings rate on equity capital employed (without undue leverage, accounting gimmickry, etc.) and not the achievement of consistent gains in earnings per share. In our view, many businesses would be better understood by their shareholder owners, as well as the general public, if managements and financial analysts modified the primary emphasis they place upon earnings per share, and upon yearly changes in that figure.  (pg 230-235)