Flagstar Moves the Finish Line to 2028 — and Puts the Buyback Back in the Drawer

Flagstar Moves the Finish Line to 2028 — and Puts the Buyback Back in the Drawer

Flagstar's stock is down roughly 29% in ninety trading days and sits about 32% below its mid-July high of $15.36. That drawdown is not a mystery: the bank told investors on Friday that the transformation it has been selling for eight quarters will now be complete in 2028 rather than 2027. The interesting part is how that delay was framed — and how thin the margin for error has become.

Twelve months, not "one or two quarters"

Three months ago, on the July call, CFO Lee Smith was explicit that the downside case was a rounding error on the timeline. Here is what he said then, versus what he said this week:

the worst case scenario is maybe it takes us 1 or 2 quarters longer to get to where we said we were going to be by Q4 '27

That is a 4x-6x expansion of the "worst case" in a single quarter, and management did not pretend otherwise. Joseph Otting's framing — “given a higher for longer interest rate environment, the full realization of our transformation has shifted to 2028 from 2027” — Joseph Otting, Executive Chairman and CEO · 2026-10-23 — is honest, but it also retires the credibility of the earlier timeline.

The keyword shift is stark. Flagstar's own keyword history shows rate environment vaulting to the top of the quarter's list, with higher interest rates right behind it. Both are words the company has used before, but never as the load-bearing explanation for a guide-down. Meanwhile CRE runoff — the good-news engine of prior quarters, all those substandard loans walking out the door at par — slipped down the ranking as payoffs held at $1.2 billion but the credit benefit did not.

The numbers underneath are unflattering to the old plan. EPS guidance is now $0.10–$0.15 for 2026, $0.85–$0.95 for 2027, and $1.60–$1.70 for 2028. More tellingly, the balance sheet forecast has collapsed: year-end 2026 total assets at $82.1 billion, $92.1 billion at the end of 2027, and roughly $97 billion at the end of 2028. In July, Smith was guiding to about $100 billion by the end of 2027; earlier in the year the number was $109.6 billion. A smaller balance sheet with a lower margin does not compound its way to peer profitability quickly.

Where the quarter actually broke

The provision was $76 million. But the reveal in Q&A matters more than the headline: “$34 million was already fully reserved for. So if you take that out, we would have been at $40 million in net charge-offs” — Lee Smith, Co-President, Co-Chief Financial Officer · 2026-10-23 — meaning the bank had to replace reserves it had already built, because higher cap rates on Bronx rent-regulated buildings and rising operating expenses pushed more loans into the criticized bucket. Net charge-offs were $74 million, or 49 basis points, down $25 million sequentially; those two facts sit awkwardly together.

The credit migration is the crux. Migration into criticized and classified rose $670 million, or 6%, and substandard rose $248 million. Smith was candid that this is partly self-inflicted rigor: the 18-month look-forward applies contractual reset rates — SOFR plus 300 or prime plus 275 — against 2025 financial statements that have not yet captured market-rate rent increases. DSCR coverage is the trigger, and it is being run conservative. Nonaccruals held near $2.8 billion, with 42% still current and paying.

Then there is the part competitors started doing this week too: selling the rent-regulated book. Otting said Flagstar is not: “we have currently like nothing on the table to do that” — Joseph Otting, Executive Chairman and CEO · 2026-10-23, arguing loan-by-loan resolution better protects shareholder capital. That is a bet that New York City rent regulations do not get worse — and that the $12.9 billion tranche, of which $8.1 billion is more than half rent-regulated, can be worked down without a portfolio sale.

The one variable that decides everything

Strip out the noise and Flagstar's 2027–2028 bridge rests on two mechanical items and one act of faith. The mechanical items are genuine: $12 billion of sub-3.9% CRE and multifamily loans reset or mature between October 1, 2026 and December 31, 2027, plus another $7 billion in 2028 at a 4.43% weighted average coupon, plus $10.7 billion of floating-rate CRE that reprices upward with the curve. The act of faith is deposits.

And that is where the competitive deposit environment bites. Otting's description of the CD market is the most concrete thing on the call: “30 days ago, it was pretty difficult to find a CD for 1 year over 4%. Some of the competitors are at 4.75% today” — Joseph Otting, Executive Chairman and CEO · 2026-10-23. Flagstar's cost of interest-bearing deposits ticked up just a basis point to 3.09%, but the spot rate of 2.56% and a targeted deposit beta of 60% versus a realized 32% on non-CD interest-bearing balances tells you how much repricing is still in the pipeline. This is the same tension showing up in regional bank calls across the tape this week — deposit pricing and cost of deposit commentary is now near-universal.

is up 5% year-over-year but down 44% from the 2024Q1 peak of $786 million. NIM expanded 4 basis points to 2.17%, with September at 2.23%. A quarter ago Smith called the June NIM "a floor" — “I am looking at that as a floor” — Lee Smith, Chief Financial Officer · 2026-07-24 — and the Q4 guide is now 2.20%–2.25%, so the floor held, barely. But holding a floor is not the same as expanding one, and the deposit assumption was cut from prior expectations to $800 million–$1 billion a quarter.

On credit, the provision went from $18 million in the second quarter to $76 million — a swing of $58 million against a base of almost nothing. Efficiency, at least, is intact: an efficiency ratio of 42.8% is down from the 82.7% disaster of 2024Q3, and full-year 2026 operating expenses are still guided down 9% year-over-year.

The buyback retreat

The sharpest irony sits on capital. Flagstar finished its $250 million repurchase — 19.1 million shares at roughly $13 — and Otting now says “it's prudent until we build our core earnings and ensure that our credit quality forecast is accurate” — Joseph Otting, Executive Chairman and CEO · 2026-10-23, with resumption expected in 2027. Yet a quarter ago, when an analyst asked hypothetically about a $10 stock, Otting said “that would be an incredibly attractive price been for us to execute on our stock buyback” — Joseph Otting, Executive Chairman and Chief Executive Officer · 2026-07-24. The stock trades around $11 today, at 0.8x book, with CET1 at 12.38% — top quartile against peers and well above the 11% target.

So the market is being asked to underwrite a 2028 EPS number of $1.60–$1.70 that management itself has now moved twice, on a balance sheet that keeps shrinking, funded by deposit growth that is the acknowledged swing factor, in a rate environment that has repriced 75 basis points on the 5-year in three months. The balance sheet diversification strategy is real — C&I is up $2 billion to $20.6 billion, 164 revenue-producing bankers versus 124 at year start, a $2.7 billion pipeline — and the fixed asset repricing math is genuinely mechanical.

But "mechanical" plus "one act of faith" is precisely the combination that has cost Flagstar a year of credibility. The stock is telling you it wants proof of the deposit line, not another bridge slide.