Some stocks fall because a quarter disappointed. Others fall because, in a single evening, it becomes clear that the business model rested on a rule somebody else wrote. Since Tuesday, Fair Isaac, the company behind the FICO score, belongs firmly in the second group. The shares closed at $617.87, down $223.02 or 26.5%, the worst trading day in the company’s history. At the low, the stock touched $595.19. That leaves it down roughly 62% year to date and almost 70% below its record high near $2,000.
The trigger was not a profit warning, an accounting scandal or a rival with a better product. The trigger was a table. Bill Pulte, director of the Federal Housing Finance Agency (FHFA), announced on Monday evening that Fannie Mae and Freddie Mac will move to a single pricing grid for mortgages scored with Classic FICO and for mortgages scored with the competing VantageScore 4.0 model. Pulte said the old two-grid structure “makes zero sense.” The stock slipped about 8% after hours on Monday; on Tuesday it collapsed.
What sounds like a technical footnote from the plumbing of mortgage securitization is really a question about whether one of the most admired moats on Wall Street was ever a moat at all, or merely a permit. The market gave its verdict on Tuesday. The more interesting question is why it happened now. The answer sits in FICO’s own income statement.
What the FHFA actually changed
To see how a pricing table can wipe out a quarter of a company’s market value, you have to know how a conforming mortgage is priced. Roughly 70% of US mortgage originations run through Fannie Mae and Freddie Mac, which buy loans from lenders, pool them and sell them on as securities. For every loan, the two agencies charge risk-based fees known as loan-level price adjustments, or LLPAs. Those fees depend heavily on the borrower’s credit score: a higher score means a smaller adjustment and a cheaper mortgage rate.
For decades that grid was built on Classic FICO alone. A lender who wanted to sell a conforming loan to the agencies had to buy FICO scores, typically from all three credit bureaus for every applicant. In early September the FHFA had already opened the door, telling Fannie and Freddie to approve all lenders to use VantageScore 4.0 after a pilot with around 50 lenders. But there was a catch. VantageScore loans were priced on a separate grid that effectively knocked 20 points off the score, on the grounds that VantageScore overstated credit quality by roughly that amount relative to FICO.
That haircut is now gone. A borrower with a 740 VantageScore will be priced like a borrower with a 740 FICO. That removes the only economic reason a lender had to keep paying for the more expensive score. TD Cowen’s Washington analyst Jaret Seiberg was quick to flag the inconsistency: the very same agency had published grids just two weeks earlier concluding that VantageScore overstated credit quality by about 20 points. No effective date for the unified grid has been announced. The direction, however, could hardly be clearer.
$10 versus 99 cents: the price gap behind the crash
The decisive number of the day was not in Pulte’s post but on the price list. A FICO score for mortgage origination now costs about $10 per pull. VantageScore is being offered through TransUnion at 99 cents through December 2028. A typical mortgage with two applicants and pulls from three bureaus means six scores per application, and often several rounds of applications as borrowers shop around or walk away.
For an individual homebuyer those sums are not dramatic. For the industry as a whole they are. VantageScore says it expects nearly $1 billion in savings over the next 12 months. Rocket Mortgage, the largest mortgage originator in the country, has already said it will make VantageScore 4.0 its preferred model on all eligible loans starting in the fourth quarter. Borrowers who came out ahead with VantageScore saved an average of $1,600 at closing, according to Rocket. UWM has reported that about a quarter of its borrowers score better on VantageScore. With the 20-point haircut gone, that share is likely to climb sharply.
Consider the lender’s math. Switch to VantageScore, and you pay a fraction of the cost for the score while some of your customers get better pricing at the same time. There is no longer a reason to stick with the expensive product, other than habit, IT integration costs, and the question of whether the secondary market will follow.
The number that gave FICO away: 97% growth without more loans
To understand why Washington struck now, look at FICO’s fiscal third quarter, reported at the end of July. Revenue came in at $674.2 million, up 26% from a year earlier. The Scores segment grew 41% to $458.9 million. Mortgage origination revenue rose 97%. Loan volumes, by contrast, were up only in the low single digits. FICO itself attributed the growth primarily to a higher mortgage origination unit price.
Put differently, nearly all of the growth in the company’s most important business came not from more people buying homes, but from each homebuyer paying more for the score. Mortgage originations accounted for 71% of B2B Scores revenue and 62% of total Scores revenue. That works out to roughly $285 million in the quarter, or a little over 40% of group revenue, from mortgage scores alone. The company’s operating margin was 53.8%.
That is the textbook economics of a toll booth, and investors adored it for years. Analysts estimate FICO has roughly doubled its prices since the end of Donald Trump’s first term. As long as the rule held that there was no way around FICO, every price increase dropped straight to the bottom line. But a company that doubles revenue in a regulated market through price increases, at a moment when politicians in both parties are campaigning on housing affordability, is sending a signal. Pulte received it. The irony is that FICO’s own earnings release supplied the rationale for its demotion.
Debt-funded buybacks: why the fall was so steep
A 26% one-day drop is unusual for a company with operating margins above 50%. Part of the explanation lies in the balance sheet. In the first nine months of its fiscal year, FICO spent $3.0 billion buying back its own stock. In June it drew a new $1.5 billion term loan and put it straight into an accelerated share repurchase, with an initial delivery of about 1.05 million shares. On that math, the entry price for the tranche was well above $1,000 a share, close to double Tuesday’s close.
Total debt stands at $5.58 billion, stockholders’ equity is negative at roughly minus $4.1 billion, and cash was only about $305 million at last report. Quarterly interest expense has nearly doubled to almost $60 million. As long as the mortgage toll booth delivered rising revenue every quarter, this was an elegant way to lift earnings per share: diluted shares have shrunk to 22.7 million. Now the lever works in reverse. If part of the pricing power disappears, the debt stays exactly where it is.
This is not a solvency issue. Free cash flow was $370 million in the third quarter alone, and full-year guidance calls for $2.53 billion in revenue and non-GAAP earnings of $42.43 per share. But it does explain why the market did not wait for the revenue hit to show up in the numbers. A company that borrows to buy back stock at record prices turns a business risk into a valuation risk for shareholders.
Why the credit bureaus fell too
At first glance, the credit bureaus should have been the winners. VantageScore was created jointly by Equifax, Experian and TransUnion. Yet Equifax fell 6.7%, TransUnion 4% and London-listed Experian 1.2%. It looks paradoxical, but it is logical.
On every mortgage inquiry, the bureaus earn not only on the credit report itself but also on reselling the FICO score, on which they take a markup. When FICO raises prices, the bureaus’ revenue per pull often rises with it. A 99-cent score is a product that wins customers, not one that leaves much margin behind. On top of that, industry reports say the FHFA is also reviewing the requirement for three credit reports per application. Dropping one would be a direct volume loss for the bureaus. On Tuesday the market was not just pricing the loser of a monopoly; it was repricing an entire value chain that had lived off high fees on every home purchase.
Who is exposed on Wall Street
For US investors, the obvious name is Fair Isaac (FICO) itself, which shed roughly $5 billion of market value on Tuesday alone and now carries a market capitalization of about $14 billion. Close behind are Equifax (EFX) and TransUnion (TRU), whose mortgage businesses were among their fastest-growing lines in recent quarters thanks to rising score prices. Experian (EXPN), listed in London, has the least US mortgage exposure of the three, which its modest decline reflects.
On the other side of the trade sit the lenders. Rocket Companies (RKT) and UWM Holdings (UWMC) stand to lower their cost per closed loan, and in a market where origination volumes remain subdued with 30-year Treasury yields above 5.6%, every dollar of cost matters. The savings are small per loan but meaningful at scale, and they give lenders room to compete on price for borrowers.
The broader exposure is in portfolios. FICO has been a staple of quality and momentum funds for years, precisely because of its pricing power and stock performance. Anyone holding a quality-factor ETF should check how much weight sits in businesses whose pricing power is protected by regulation rather than by product. That list goes beyond FICO to include the rating agencies S&P Global (SPGI) and Moody’s (MCO), index providers such as MSCI, and exchange operators. None of them is facing a comparable decision today. But all of them share the same structural feature: their moat is partly written into rules.
The counterargument: why FICO is not going away
As brutal as the day was, there are good reasons FICO will not lose half its mortgage business overnight. First, the secondary market is slow to change. Investors in mortgage-backed securities still expect FICO scores to model pool risk, because FICO has decades of performance data behind it. As long as that remains true, many lenders will keep pulling both scores. Deutsche Bank kept its buy rating but acknowledged the developments are “meaningfully negative.”
Second, the unified grid applies to the conforming market run by Fannie and Freddie. In the government-backed FHA market, FICO 10T is slated for implementation in January 2027. Credit cards, auto loans and personal loans are not directly affected by the decision. Third, FICO has a software business generating $215 million a quarter, with platform ARR growing 62%. That business is about to become much more important to the story.
Fourth, and most interesting, the valuation has already compressed sharply. At $617.87 and guided non-GAAP earnings of $42.43 per share, FICO now trades at roughly 14.6 times earnings. A year ago, the multiple was several times that. Even if mortgage revenue fell by a third, a highly profitable company would remain. The open question, as RBC Capital Markets put it, is whether FICO will have to accelerate a shift away from per-pull origination fees toward other pricing structures, and what margin those new structures can deliver.
The bigger lesson: moats by permission
On Wall Street, a monopoly with pricing power is the holy grail. Warren Buffett has called the ability to raise prices without losing customers the single most important factor in evaluating a business. For years FICO was the textbook example. But there are two kinds of pricing power: the kind that comes from a product customers desperately want, and the kind that comes from a rule forcing customers to buy it. The first is durable. The second is durable only until someone changes the rule.
The FICO case shows how the second kind breaks: not through a technological attack, but through the price that makes it visible. When a FICO score cost a couple of dollars, almost nobody cared. At $10 a pull, with mortgage score revenue up 97% in a single year, it became a political target, in a midterm year in which housing costs are one of the defining issues.
That is a question investors should ask of every quality stock they own. How much of the profit comes from a product that is better, and how much from a regulation that mandates it? And how far has the company already pushed its price? The FICO story is not one of bad management. It is a story about how a moat can be overextended. Every price hike makes next quarter’s earnings bigger, and it also makes it more likely that someone lowers the drawbridge.
Outlook: what to watch
Over the coming weeks, three things will matter. First, the effective date of the unified grid, which the FHFA has not yet announced. Second, how quickly large lenders beyond Rocket Mortgage switch, from UWM to the big banks and regional lenders. Third, how mortgage-bond investors react: as long as they demand FICO data, part of the double-pull volume survives.
The next hard numbers arrive with FICO’s fiscal fourth-quarter report in early November. It will show whether management is already cutting unit prices to defend share, or betting on the market’s inertia. Either path has a cost. If FICO cuts prices, margins shrink. If it holds them, market share shrinks.
Tuesday made a stock a quarter cheaper. It also clarified something that appeared in no valuation model: FICO’s value never lay solely in the mathematics behind the score, but in a table at Fannie Mae and Freddie Mac. That table is now being rewritten. The bill will not be paid by homebuyers. It will be paid by the shareholders who paid a monopoly’s price for a toll booth.
Try TradingView Free for 30 Days
Plus get a $15 discount on your first subscription through this link.
