Fair Isaac (FICO): FHFA Order Ends Score Monopoly, Stock Falls 17%

On Friday 4 September 2026, Federal Housing Finance Agency Director Bill Pulte instructed Fannie Mae and Freddie Mac to approve every mortgage originator to use VantageScore 4.0, effective immediately. Fair Isaac shares closed at $932.26, down $186.67 or 16.7% on the session, after trading as much as 21% lower during the day. The instruction ended a four-month pilot that had been capped at roughly 50 lenders, and with it the effective monopoly Fair Isaac has held over credit scoring in the government-sponsored mortgage channel for decades.

The number that matters most, though, is not the share price. It is 62%. In Fair Isaac's fiscal third quarter, reported on 7 August 2026, mortgage origination revenue accounted for 62% of Scores segment revenue of $458.9m, and 71% of business-to-business Scores revenue, up from 53% in the same quarter a year earlier. That mortgage line grew 97% year on year. It grew almost entirely on price.

What actually happened

Pulte announced the change himself, in his own words: "So, EFFECTIVE IMMEDIATELY, I'm instructing Fannie and Freddie to approve ALL lenders to use VantageScore." VantageScore's own release the same day carried the title "VantageScore 4.0, the Mortgage Credit Score, Now FHFA Approved for All Lenders Originating Fannie Mae and Freddie Mac Mortgage Loans", confirming the scope: all lenders, not a widened pilot.

VantageScore 4.0 is jointly owned by Equifax, Experian and TransUnion. It differs from the incumbent FICO models in that it incorporates rental and utility payment histories, the basis of its sponsors' claim that it scores a broader slice of the population. Until Friday its use at the two government-sponsored enterprises was confined to a limited pilot. Now it is approved for the whole channel.

The pricing gap behind the decision is stark. Fair Isaac raised the wholesale price of a mortgage credit score from $4.95 to $10.00 for 2026, roughly a doubling; five years ago that same wholesale score cost about $0.60. Pulte has characterised the cumulative increase as roughly 1,800% since 2020. Against that, the bureaus have priced VantageScore 4.0 aggressively: TransUnion and Experian at $0.99 a mortgage origination score, and Equifax at $1.00, held through the end of 2027 — down from the $4.50 Equifax had offered a year earlier. Whatever a given lender negotiates, the alternative now sits at roughly a tenth of the incumbent's price.

There was a second limb to Friday's announcement that is easy to miss. Pulte separately posted that "Equifax, Experian, and TransUnion have been overcharging Americans for far too long", and that the FHFA is "seriously considering bi-merge and stronger solutions" for government-backed home loans. A bi-merge framework would allow lenders to pull reports from two bureaus rather than three, removing roughly a third of mortgage report volume across the industry. Equifax closed down 6.4% at $177.05 and TransUnion 5.9% at $79.88, both well off intraday lows near 9%; Experian's London-listed line fell about 5%.

Why the market reacted the way it did

A 16.7% fall in a business with no operational problem, no earnings miss and no change to current-quarter revenue requires explanation. The market was not repricing this year's cash flows. It was repricing the mechanism that produced them.

Consider what the fiscal third quarter actually showed. Scores segment revenue rose 41% year on year, driven by a 49% increase in business-to-business revenue, and the company attributed that primarily to higher mortgage origination score unit pricing. Mortgage origination revenue nearly doubled while US mortgage volumes did nothing of the kind. This was not a company winning share or riding a housing boom. It was a company exercising pricing power inside a captive channel, because a lender originating a conforming loan had no alternative score to buy.

On Friday afternoon, the lender acquired an alternative. That is the whole event. And it matters far more to the forward path than to the current one, because price-led growth is the most fragile kind of growth there is when the price was set by structure rather than by competition.

Here is the asymmetry the sell-off was pricing, and it is worth stating plainly: the earnings damage will be slow, but the pricing power died immediately. Loan origination systems, investor guidelines, mortgage insurance workflows, securitisation documentation and lender pricing grids all reference FICO scores today, and rewiring them takes quarters rather than weeks. Not a single loan needs to switch score for the thesis to break. It is sufficient that Fair Isaac can no longer put through another doubling in 2027, because a lender facing a $10 score with a $0.99 substitute available under an approved framework has a conversation it did not have a week ago.

The Openbook read

Momentum is the factor that moves furthest and fastest, and it moves against the stock. Friday was the worst single session in Fair Isaac's recent history, and at $932.26 the shares sit roughly 53% below the 52-week high of $1,998.01, with the 52-week low of $870.01 now only about 7% away. On roughly 23.6m shares outstanding, Friday's move removed on the order of $4.4bn of market value — derived from the per-share fall rather than reported directly, so approximate. A momentum score built on trend and drawdown deteriorates sharply here and will stay deteriorated for some time, because this is not the kind of drawdown that mean-reverts on a good quarter.

Growth requires the most careful handling, and it is where the mechanical scores will mislead for at least three reporting periods. Trailing growth looks superb: Scores up 41%, mortgage originations up 97%. But a growth factor that reads price increases and volume increases identically will score this business as a compounder right up to the point where the price increases stop. The honest adjustment is to decompose the growth. Roughly 62% of a $458.9m quarterly Scores line is mortgage origination — about $285m a quarter on that arithmetic — and the near-doubling in it came from unit pricing. Strip the price effect and the underlying volume growth is unremarkable. Growth should be marked down not because demand fell, but because the source of the growth has been disclosed as non-repeatable.

Profitability is the factor doing the real work in this repricing, and it is a quality-of-earnings question rather than a level question. Scores is Fair Isaac's high-margin engine, and its margin was regulatory in origin: a captive channel, a mandated score, and a wholesale price the buyer could not refuse. Margins earned that way and margins earned through product advantage look identical in a spreadsheet and behave completely differently under stress. Nothing about Friday changes the reported margin. Everything about it changes what that margin is worth as a forecast.

Solvency is, for once, not where the story is. Fair Isaac remains strongly cash-generative and the balance sheet is not the pressure point in this event. It is worth watching only insofar as the company has historically been an aggressive repurchaser of its own stock, and a management team that believes the market has overreacted has an obvious way to express that.

Reward/Risk is where the picture has changed shape rather than merely direction. Before Friday, the distribution of outcomes was fairly narrow: a dominant franchise raising price into an inelastic channel. It is now genuinely bimodal, and the fork is adoption speed. If lenders move slowly — and the plumbing argument says they will — the reported numbers hold up for several quarters while the multiple stays compressed, and the gap between the two becomes the argument. If a handful of large originators move quickly and publicly, the price-led revenue line unwinds faster than the market currently assumes. What has unambiguously widened is the range of defensible answers, and a reward/risk factor should register that widening even before it registers a direction.

The reader's real question is the one the five factors converge on: has a roughly 53% drawdown priced a monopoly loss that will take years to appear in the profit and loss account, or has it priced it correctly and instantly? The case for "correctly and instantly" is that markets price the terminal state, and the terminal state changed on Friday regardless of the transition path. The case against is that "effective immediately" describes a permission, not a migration, and the company will keep printing very good numbers while the argument is unresolved. Both can be right over different holding periods, which is why reward/risk rather than growth is the factor to lean on here.

The read-across

The sharpest irony in Friday's session is that the owners of the winning product were punished alongside the loser. VantageScore is a joint venture of Equifax, Experian and TransUnion. All three fell, and Equifax and TransUnion each shed more than 5% on a day their jointly owned model was handed the whole government-sponsored channel.

The resolution is that Pulte fired at two targets in a single day. Opening VantageScore to all lenders helps the bureaus, whose joint venture gains volume. Threatening bi-merge hurts them considerably more, because removing one of three bureau reports from every government-backed mortgage file deletes roughly a third of that revenue line outright — and unlike the scoring change, it requires no lender behaviour change at all to take effect. The market weighed the two and concluded the threat outweighed the gift. That is a coherent judgement, and it explains why the read-across from this story runs through regulatory posture rather than through competitive dynamics. Investors screening the credit and data complex on our screener should note that Equifax and Experian now carry a regulatory overhang of their own that has nothing to do with who wins the scoring contest.

More broadly, this is a template event for any listed business whose margin rests on a mandate rather than on a moat. The screen to run is not "which companies have high margins" but "which companies have high margins that a single administrative decision could remove". Interchange, clearing, ratings, index licensing and exchange market-data all sit somewhere on that spectrum.

What to watch next

Three things, in order of how much they will move the argument.

• The bi-merge decision. Pulte said the FHFA is "seriously considering" it. There is no published timetable. It is the larger of the two announcements for the bureaus and entirely unresolved, which makes any headline on it a genuine catalyst rather than a restatement.
• Fair Isaac's fiscal fourth quarter and full-year results, due in the first half of November on the company's usual schedule. The Scores segment disclosure is what matters — specifically whether mortgage origination revenue holds at roughly 62% of the segment and whether management gives any indication of 2027 wholesale pricing. A company that announces no price increase for 2027 will have answered the central question without saying so.
• Lender adoption evidence. Watch for the first large originator to state publicly that it is ordering VantageScore 4.0 on conforming files, and for the government-sponsored enterprises to publish operational guidance covering delivery, pricing grids and representations. Until that plumbing exists in writing, the directive is permission without a path, and the migration cannot begin in volume.

Full profiles and factor scores for Fair Isaac and its data and analytics peers are on Openbook. The factor view will update as the November numbers land; what will not change quickly is Friday's central observation — this was a repricing of the durability of a margin, not of the demand for a product.