Look-Through Earnings

Concept Analysis

Definition & Origins

Look-through earnings is Buffett's alternative measure of Berkshire's true economic earning power: reported operating earnings, plus Berkshire's proportionate share of the retained operating earnings of major investee companies, less an allowance for the tax Berkshire would have paid had those retained earnings been distributed as dividends. Standard accounting forces Berkshire to report only the dividends it receives from companies it owns less than 20% of; the far larger share of earnings those companies retain and reinvest never appears on Berkshire's income statement. Look-through earnings corrects that distortion.

The underlying idea predates the term by more than a decade. In the 1978 letter Buffett was already explaining that Berkshire's share of SAFECO's earnings was $6.1 million while only about 18% of that — the dividends received — showed up in reported operating earnings, and that the unreported balance was "just as real in terms of eventual benefit to us as the amount distributed." Through the 1980s, as the equity portfolio grew to include Coca-Cola, Capital Cities/ABC, GEICO, and Gillette, the gap between what GAAP reported and what Berkshire economically earned widened into the hundreds of millions. In the 1989 letter Buffett gave the idea its name and its quantitative frame: he appended Berkshire's share of investees' retained earnings to reported operating earnings and set the explicit test that look-through earnings must grow about 15% annually for intrinsic value to compound at the same pace. From 1990 onward the letter carried a dedicated "Look-Through Earnings" section with a full calculation table, which ran through 2000.

Core Ideas

GAAP systematically understates a holding company's earnings. When Berkshire owns 8% of Coca-Cola, GAAP lets it book only the dividend check. But Coca-Cola retains most of its earnings and reinvests them at high returns on equity — compounding that accrues to Berkshire's ownership interest as surely as if the cash had arrived in Omaha. The 1990 letter made the magnitude concrete: roughly $250 million of Berkshire's share of investee earnings was retained that year, against $371 million of reported operating earnings. Counting the retained portion (after a hypothetical tax allowance) nearly doubled the earnings figure, to about $590 million.

Retained earnings can be worth more than distributed earnings. Buffett's argument is not merely that retained earnings count, but that they may be worth more than 100 cents on the dollar when the investee's management can deploy capital at high rates of return. His "a-bird-in-the-bush-may-be-worth-two-in-the-hand" reasoning in the 1989 letter is explicit: earnings retained by Coca-Cola or Cap Cities are deployed by talented, owner-oriented managers who sometimes have better uses for the funds in their own businesses than Berkshire would have in its. This is why Berkshire does not demand dividends from its investees — the 1996 letter asks, rhetorically, "So why should we want them paid out?"

The tax allowance keeps the measure honest. Look-through earnings is not simply reported earnings plus the investees' full retained earnings. Buffett deducts the tax Berkshire would owe had the retained earnings actually been distributed — typically a 14% rate on dividends — so the figure approximates after-tax economic reality rather than an inflated gross number.

The measure is deliberately rough, and Buffett says so. Every annual table carried the warning that the figures are "necessarily very rough" and based on judgment calls. Precision was never the point; the point was to give shareholders the right order of magnitude of Berkshire's earning power and to hold management accountable for growing it at the 15% target rate.

Look-through earnings ties earnings growth to intrinsic value growth. Buffett used the measure as his own scorecard: if look-through earnings compound at 15%, intrinsic business value compounds at roughly 15%. The 1991 letter noted that since 1965 look-through earnings had grown at almost the identical 23% rate recorded for book value — evidence that the metric tracked the economics it was designed to capture.

Practical Application

Evaluate a holding company on economics, not on its income statement. For any investor analyzing a company with large minority stakes — Berkshire in the 1990s being the archetype — reported earnings alone are misleading. The correct procedure is Buffett's: take operating earnings, add the share of investees' retained earnings, subtract the tax that distribution would have triggered, and judge management on the growth of that figure. Buffett published his own number annually and reported honestly when it fell short, as in 1991, when look-through earnings declined 14% on falling media earnings, the Gillette preferred conversion, and a break-even result at Wells Fargo.

Use retained earnings to judge capital allocation at investees. The framework doubles as a test of whether an investee should retain earnings at all. Buffett's 1993 letter supplied the proof case with Capital Cities/ABC: Berkshire bought three million shares in 1986 at $172.50, and over eight years the retained earnings attributable to those shares — hypothetically taxed under the look-through method — were $152 million. When Berkshire sold one-third of the holding at $630 per share, the realized gain after a 35% capital gains tax was $297 million. Retention had created value far beyond the retained dollars themselves, confirming that look-through earnings was a conservative representation of true economic earnings.

Watch for the framework's limits. The look-through adjustment only adds value if investees actually deploy retained capital well. A company retaining earnings to fund mediocre projects destroys the value the framework credits it with. Buffett acknowledged the assumption embedded in the measure by restricting his generous assessment to specific managements — "I would not make such a generous assessment of most managements" — which is the practitioner's cue: apply look-through math only where reinvestment skill justifies it.

Common Misconceptions

Misconception 1: GAAP earnings accurately measure any company's performance. For a manufacturer or retailer whose income all flows through its own income statement, GAAP earnings are a reasonable approximation. For a holding company carrying large minority positions on the cost method, GAAP structurally omits the majority of economic value creation. The error is not random noise — it is a systematic understatement that grows with the size and profitability of the equity portfolio.

Misconception 2: Look-through earnings are speculative or non-cash fantasy. Every input is an actual reported number from an actual investee: real operating earnings, real retention rates, a stated tax assumption. Nothing is projected. The only "look-through" element is counting earnings GAAP requires to be excluded from the investor's own income statement — earnings whose effect later surfaces in the investee's share price and dividends.

Misconception 3: The measure is an excuse to inflate results. Buffett built in the opposite bias. He deducted hypothetical taxes on earnings never received, excluded capital gains from both sides of the calculation, and labeled the tables rough approximations. The 1993 Cap Cities comparison showed the framework had understated the value created by retention — the realized gain dwarfed the look-through carrying value.

Misconception 4: Retained earnings are always as good as cash in hand. The entire case for valuing undistributed earnings rests on reinvestment quality. Where an investee lacks high-return uses for capital, retention is worth less than distribution — which is why Buffett's reasoning names specific managers rather than endorsing retention in the abstract.

Thought Evolution

Related Concepts

Case Companies