Materiality Has a Number

A sell-side analyst is building a model the night before your results. She has a customer-concentration shift to assess, a deferred capital project, and a 240 basis-point move in one segment's margin. She is not consulting a "reasonable investor" standard. She is asking a different, blunter question: does any of this move my number by enough to matter? She has a threshold in her head. If the factor shifts her EPS estimate by more than five percent, or her return-on-invested-capital read by more than half a percentage point, it goes in the model. If it doesn't, it goes in a footnote, or nowhere.

That analyst is the reader our disclosure committee was supposedly serving. And she runs materiality as a calculation, with an actual number, every working night of her life.

Which is awkward, because we published an essay arguing she shouldn't be able to.

1The position we took

In Materiality Is a Choice, Not a Calculation, the closing essay of the Bedrock Papers, I argued that the IR profession's instinct to treat materiality as a threshold is the error underneath most disclosure failures. The piece was direct about it. The practitioner tradition "treats materiality as fundamentally a legal question," produces "decisions calibrated to the legal threshold rather than to the information needs of the company's actual investor audience," and in doing so manufactures the information asymmetry that foreign institutions price as a discount.

The essay's title states the thesis outright: choice, not calculation. The recommendation that followed was to lean toward disclosure, to ask what this audience needs to know rather than whether it clears the threshold, and to accept the judgment burden that comes with abandoning the clean number.

I still think the framing question, material to whom and for what decision, is the right one to carry into a committee. But the essay was rude about the number. It treated the quantitative threshold as a crutch the profession reaches for to avoid thinking. That was too fast. The number is not a crutch. It is a tool, and the people who actually price your stock use it every day.

2The counter-case, steelmanned

Here is the case I should have engaged and mostly waved past: materiality can be calculated, the people doing the calculating are exactly the audience you care about, and securities law keeps reaching for bright lines precisely because pure judgment is unworkable at scale.

Start with the sell side, because that is where the number is most explicit. James Valentine's Best Practices for Equity Research Analysts, the codified manual that trains a meaningful share of the analysts covering your stock, gives a flat quantitative test. A factor is material, in his framework, if it will likely move earnings per share or cash flow per share by more than about five percent, or return on invested capital or equity by more than fifty basis points. Valentine even works the arithmetic: enough to move a company's ten percent ROIC to 9.5 or 10.5. He is candid that these are starting points that vary by fund objective, time horizon, and sector. But he does not throw up his hands and call materiality a quagmire. He gives the analyst a filter, because without one the analyst drowns in factors that could easily be mistaken by the broader market as critical even though more research would prove otherwise. The threshold's first job is subtraction: narrowing a long list of things that look important down to the few that move the number.

None of this is a fringe view. Walk over to the corporate-communications literature and you find the same instinct. Westbrook's Strategic Financial and Investor Communication treats ten percent as the generally accepted accounting cut for a "material" variation in earnings, and advises companies to wait until near a balance date "to be reasonably sure that differences will actually be material." Different number, same move: when you need to decide whether to flag a profit warning, you reach for a percentage, not a philosophy seminar.

Now the harder half of the steelman, which is that the law itself once drew the line and meant it. Take M&A disclosure timing, which Mahoney's Professional's Guide tracks across the 1980s. For a stretch, the standard was a genuine bright line. In Heublein (1984) and Carnation (1985), the prevailing standard held there was no duty to disclose merger talks until the parties reached an agreement in principle. Clean, calculable, defensible. It gave a company, in Mahoney's phrase, "a clearly defined and somewhat comfortable start point." You knew exactly where you stood. So this is not a hypothetical about the appeal of a number. The system actually built one, and IR practitioners actually relied on it.

The one time U.S. securities law reached for a materiality bright line, the Supreme Court found it and dismantled it on purpose.

Then the Supreme Court took it apart. In Basic Inc. v. Levinson (1988), the Court ruled that "preliminary merger or acquisition negotiations could be material under certain circumstances," and replaced the bright line with a case-by-case probability/magnitude test: how likely is the deal (are senior management, the board, and the bankers involved?) weighed against how big it is (the size of the companies, the premium on offer). The Court was explicit about why. It acknowledged a bright line is "easier to follow than a standard that requires the exercise of judgment," but held that "ease of application alone is not an excuse for ignoring the purposes of the Securities Acts," and chose the harder, judgment-based test anyway. Read that again, because it is the cleanest statement against my own counter-case that the law itself has ever produced. The one time U.S. securities law reached for a materiality bright line, the Supreme Court found it and dismantled it on purpose, ruling that the rule's ease of application was no excuse for ignoring what the securities laws are for.

That cuts the steelman, but it does not kill it. Notice what the bright line's rise still tells you, even after its fall. The system's first instinct, when it needed disclosure decisions to be consistent and predictable across thousands of companies and lawyers, was to reach for a rule with an edge you could see. Judgment-only materiality is not free. A standard that says "ask what a reasonable investor would want" gives a disclosure committee no way to be consistent quarter to quarter, no way to defend a decision after the fact except by asserting good faith, and no way to train a junior team. The bright line, for all its crudeness, does what judgment cannot. It produces the same answer regardless of who is in the room. It survives a hostile post-mortem, and it can be written into a charter and taught to a junior team that will rotate through the function long after you have left it. Those are not small virtues. They are most of why disclosure committees exist. The court that abolished the bright line did not deny any of this. It simply decided that getting materiality right was worth more than getting it easy — a value judgment, not a proof that the number was useless.

For an Asian issuer the calculable test is doing even more work than it does in New York. The regional regulators (OJK and IDX in Indonesia, MAS and SGX in Singapore, HKEx in Hong Kong) generally operate single-materiality regimes anchored to financial impact, with thresholds that are often narrower and more numeric than the US framework. An Indonesian issuer deciding whether a transaction triggers a disclosure obligation is frequently working a percentage-of-equity or percentage-of-revenue test that the rulebook spells out. That is not the profession failing to think. That is the profession using the instrument the regulator handed it. And when the marginal buyer of the stock is a Singapore- or London-based long-only fund whose analyst runs Valentine's five-percent filter, the company that knows its own numbers against that filter is speaking the analyst's language. The number is the shared protocol. Telling issuers the number is a crutch, while their actual readers run on it nightly, was the weak point in our case.

3The verdict: refine

The counter-case wins a real concession, and it is this: the quantitative threshold is a genuine discipline, and the original essay dismissed it too fast. A five-percent EPS test, a 50bps ROIC test, a ten-percent earnings-variation test: these are not avoidance mechanisms. They are filters that the smartest readers of your disclosure use to decide what enters their model, and a disclosure committee that does not know its own facts against those filters is flying blind. The number is the floor. It tells you the set of items that almost certainly matter to almost everyone. You should always run it. If a factor clears the threshold, the conversation is over; it is material and it is going out.

But here is the concession running the other way, and it is decisive. The one time the law itself reached for a materiality number, it recanted. The Basic court found the bright line and replaced it with judgment, on the record, holding that its ease was no excuse for a worse answer. And eleven years later the SEC said the same thing in the accounting register and said it harder. In Staff Accounting Bulletin 99 (1999), the Commission explicitly rejected mechanical rule-of-thumb percentage tests for materiality. A misstatement well under any percentage threshold is still material, the SEC said, if it masks a change in earnings trend, converts a profit into a loss, hides a failure to meet analyst consensus, affects loan-covenant compliance, or triggers a management bonus. Such misstatements, the SEC staff wrote, "are not immaterial simply because they fall beneath a numerical threshold." That is the regulator, with full authority to bless the bright line, declining to. The five-percent item that flips you from a beat to a miss is material at one percent. The number cannot see that. Only judgment can.

The number disposes of the easy cases. Judgment is for the hard ones, and the hard ones are where the asymmetry hides.

So the sharper synthesis is this: use the number as a floor and a filter, never as the whole answer. Run the calculation first. It is fast, it is defensible, and it captures the items that obviously matter. Then run the second test, the one SAB 99 forces and the one Basic foreshadowed: does this sub-threshold item change the story a reasonable holder is telling herself about the company? Does it convert the narrative the way it would convert a profit to a loss? The number disposes of the easy cases. Judgment is for the hard ones, and the hard ones are where the asymmetry hides.

That is a better answer than either essay gave alone. Materiality Is a Choice was right that judgment is unavoidable and that a committee hiding behind a threshold will systematically under-inform its long-duration holders. But it was wrong to treat the number as the enemy. The number is the first line of the analysis, not the absence of analysis. Mahoney's honest description of materiality as a "quagmire" of evolving legal standards and case-by-case judgment is accurate precisely because both things are true at once. There is a calculable layer, and there is an irreducible judgment layer sitting on top of it, and the whole arc from Heublein to Basic to SAB 99 is the law discovering, the hard way, that you need both.

For the disclosure committee, the practical instruction tightens. Do not ask only "material to whom, for what decision," as the original essay urged; that question, on its own, gives a junior team nothing to hold onto and invites the inconsistency the bright line was built to prevent. Ask two questions in sequence. First: does this clear the number, the five-percent move, the 50bps shift, the regulator's stated threshold? If yes, disclose; the analyst's model needs it. If no, ask the second question, the SAB 99 question: does this small item nonetheless change the trend, the consensus read, the covenant, the story? If yes, it is material despite the number, and the long-only holder in London needs it as much as the analyst does. A committee that runs both tests is neither hiding behind the threshold nor drowning in open-ended judgment. It is doing the actual work.

The number, in other words, has its place. We just put it in the wrong one. It is not the answer to the materiality question. It is the first thing you check before you start answering it. The law tried making the number the standard, once, and abandoned it within four years. The original essay was right that materiality is, in the end, a choice. The refinement is that the choice gets easier, more consistent, and more defensible once you have run the calculation first, and that the calculation is most of the job most of the time. The hard cases, the ones SAB 99 was written for, are the minority. They are also the ones that get companies in trouble. So you cannot let the number make the decision for you. You also have no business starting without it.