Initiated 2026-09-18 · Price $1.26 (as of 2026-09-17 close, stockanalysis.com) · Mkt cap $2.02bn · Financials / mortgage finance · Model: verified (generic, full tier)

Rating: Sell — Conviction: Medium

The common is a thin residual behind ~$5.5bn of net senior claims plus $1.65bn of new preferred accruing 10% cash / 13% accreted ahead of it. Despite an 82% drawdown the market still capitalizes the enterprise at ~10.8x FY2025 Adjusted EBITDA — a full multiple, not a distressed one — so the stock is cheap in price, not in what it pays for the business. Owning it is a leveraged bet on mortgage rates falling hard and soon, taken after a controls failure, not a cycle event, destroyed this year's earnings.

Model value range vs price
Bear $0.00Base $0.00Bull $2.15Price $1.26

Business overview

Largest US residential mortgage originator by volume, selling only through independent brokers rather than a retail loan-officer force. It originated $163.4bn in FY2025 and $39.7bn in 2Q26 (60% purchase / 40% refi), earns a gain-on-sale margin on each loan, and retains servicing — a $247.6bn UPB book carried at $5.3bn. Earnings turn on volume (a direct function of the mortgage rate), gain-on-sale margin (channel price competition), and the MSR mark, which moves against volume. A fourth variable now matters as much: the cost and availability of the secured funding — $8.6bn warehouse, $2.95bn MSR facilities — carrying the balance sheet.

Bull case

  1. Rate relief is worth more here than anywhere in the sector — volume and gain margin expand together on a fixed cost base; Adj. EBITDA went $282mm (FY22) → $697mm (FY25) on a partial recovery. Plays out if the Fed eases with growth intact and the 30-year mortgage falls under ~5.5%. Model: rev_growth, ebitda_margin
  2. The recapitalization removes the near-term funding cliff — $1.65bn of preferred for debt and MSR-facility repayment, plus a $500mm undrawn revolver, buys time without a forced MSR sale. Plays out if the 2027 maturities refinance in the 7s. Model: FY0 net debt and the preferred (credit.debt, other_claims) enter value directly; the refinancing rate (cost_of_debt) and capital_return_pct move EPS and net cash only
  3. Scale and the wholesale channel are durable — broker distribution is structurally cheaper than retail and UWM's share is defensible, so a normalized-volume year should convert above FY2025's 22.1% margin. Model: ebitda_margin, exit_ev_ebitda

Bear case

  1. The senior claims grow whether the business does or not — the preferred compounds at 10–13% ahead of the common while the common dividend is suspended, so cash that would delever now services it. Model: other_claims, exit_ev_ebitda
  2. Leverage is at a ratings-distress level and the cycle has not turned — non-funding debt / tangible equity went 1.2x (YE23) → 3.2x (1Q26) → 6.1x (2Q26); Fitch cut to B+, KBRA to BB+. Equity of $985.3mm sits $242mm above the $743.6mm agency minimum net worth, and $2.375bn must refinance at B+ pricing by July 2027. Model: ebitda_margin, nwc_pct_rev, wacc
  3. The loss was a controls failure, not a cycle — a $603.2mm 2Q26 derivative loss from an MSR hedge put on against stated practice, for a Two Harbors deal abandoned in March; a securities class action followed in September. Model: exit_ev_ebitda, wacc
  4. Dilution is contractual and ongoing — a 200mm-share rights offering (+12.5%) priced ~59% above the last close, so insiders likely backstop it, plus 330mm warrants at $2.00 and $6.00 running to 2036 that cap upside in exactly the scenarios where it appears. Model: not in the value range — share_change moves EPS, net cash or a cross-check only (see model note)

Valuation & balance sheet

Metric (definition) Current Own history Peers Source, as-of
P/TBV (mkt cap on all Class A+D shares ÷ total equity; negligible goodwill) 2.05x Equity $2.47 → $2.05 → $1.59 → $0.99bn, YE23→2Q26 RKT ~1.89x P/B 10-Q 2026-06-30; quote 2026-09-17
EV / Adj. EBITDA (pro-forma EV incl. preferred ÷ FY25 Adj. EBITDA) 10.8x n/a (unverified) n/a (unverified) 10-Q + FY25 release
Gain-on-sale margin 133bps (2Q26) 77 · 92 · 110 · 116bps, FY22–25 n/a (unverified) Q2'26 release, 2026-08-05
Origination volume $39.7bn (2Q26) $127 · $108 · $139 · $163bn, FY22–25 Largest US originator Q2'26 + FY25 releases
ROTCE ~−44% (H1'26 ann.) ~13.4% (FY2025) n/a (unverified) 10-Q + FY25 release

Model-implied value range (model-summary.json; generic module, DCF-Gordon and DCF-exit-multiple, midpoints): Bear $0.00 · Base $0.00 · Bull $2.15, i.e. implied returns of −100% / −100% / +70% vs $1.26. These ranges show how the bull and bear drivers translate into value; they are not price targets. Base and Bear sit at the module's zero floor — enterprise value does not cover the senior claims. A sensitivity changing only exit_ev_ebitda shows that is no artifact of a harsh multiple: Base is $0.07 at 9.0x, $0.21 at 10.0x, $0.33 even at the ~10.8x the market itself pays (−74%). The price is not paying for the base case at any plausible multiple; it pays for the bull recovery, or for option value a DCF midpoint cannot express — the main reason this is Medium, not High, conviction.

Balance sheet: net leverage 5.57x, gross 8.65x corporate debt / FY0 Adj. EBITDA, coverage 1.99x (model); non-funding debt / tangible equity 6.1x (Fitch). Liquidity: $498.4mm cash, $500mm undrawn related-party revolver, $1.65bn preferred proceeds. Nearest material maturities: $500mm 5.750% notes June 2027, $1,875mm MSR facility July 2027 — the model's $11.06bn 2027 ladder bucket is mostly revolving 364-day warehouse capacity; the real task is $2.375bn. Ratings: Fitch B+ (Aug 2026), KBRA BB+ (2026-07-14, stable).

Model note: Built, full tier, 0 failing error checks, no scenario CHECKs. verification.status is verified as of 2026-09-20 (LibreOffice's independent recalculation matches the Python values on every formula cell; it was not_run at initiation because the environment's LibreOffice install was missing its Calc component). Unverified inputs: market.cash only (pro-forma reported cash + committed preferred proceeds; components sourced, the combination mine). No assumptions without a basis. Three deliberate upward biases make implied values generous: Adjusted EBITDA excludes MSR marks and derivative results — what caused the 2026 losses; it treats capitalized retained-MSR value as cash; and preferred is held flat, its dividend charged via capital_return_pct — which moves only the net-debt path, never implied value — and never accreted, so the model understates the preferred's claim twice over; with Base and Bear already at the zero floor this cannot change the call. Floating-facility coupons are estimates tied to the reported $86.8mm of 2Q26 non-funding interest. Ratings come from press summaries (Fitch date n/a (unverified)) and the preferred's 10%/13% terms from a filing summary, not the indenture. Generator note (2026-09-24): in this module share_change, capital_return_pct and cost_of_debt move EPS and the net-cash path only, and peer_pe feeds a cross-check only; none changes the value range, so bull/bear tags no longer cite them. Scenario deltas were rebuilt after multiple_pct shocks were extended to the Gordon method (previously they moved only the exit-multiple half), and EPS now uses average diluted shares.

Scenario stress test

Reasoned from the bull/bear drivers above. The model column comes from the scenario overlays (Base case + shock).

Scenario Effect Mechanism Magnitude Model Δ value vs Base
S1 Fast equity crash − Risk premium on a levered B+ stub spikes; flight-to-quality rate relief helps volume only with a lag High $0.00 (at floor)
S2 Slow bear / recession − Purchase volume (~60% of mix) falls with housing turnover; delinquencies raise servicing advances and repurchases High $0.00 (at floor)
S3 Rapid rate shock − Mortgage rates spike, refi shuts, purchase freezes; MSR marks rise but Adjusted EBITDA and cash do not High $0.00 (at floor)
S4 Slow rate grind − Same channel, longer: origination stays depressed and the 2027 refinancings price wider High $0.00 (at floor)
S5 Soft-landing cuts + The one genuinely good outcome: refi wave plus purchase demand converting on a fixed cost base High +$1.30 (to $1.07)
S6 Recession-driven cuts ± Refi tailwind against weaker purchase demand, rising advance costs, tighter funding for a B+ issuer Medium $0.00 (at floor)
S7a Credit liquidity shock − Warehouse and MSR advance rates cut, secured funding gaps wider — most direct threat to the common High $0.00 (at floor)
S7b Slow default cycle − Servicing advance obligations and repurchase demands rise without offsetting volume Medium $0.00 (at floor)
S8 Stagflation − Rates high and sticky while real incomes fall: affordability and volume crushed as funding costs rise High $0.00 (at floor)
S9a Dollar spike 0 No material effect, not modeled — US residential only; no FX revenue, cost or funding — $0.00 (at floor)
S9b Dollar slide 0 No material effect, not modeled — reaches UWM only via the long end, already S3/S4's channel — $0.00 (at floor)
S10 Melt-up + Tight spreads and risk appetite cheapen the 2027 refinancings and lift the multiple on a levered stub Medium +$0.31 (to $0.08)
S11 Energy supply shock − Oil spike lifts headline inflation and mortgage rates while squeezing household budgets Medium $0.00 (at floor)
S12 Mega-cap/AI derating 0 No material effect, not modeled — no earnings channel; rotation-driven rate moves sit in S3–S6 — $0.00 (at floor)

With Base at the zero floor the model cannot express further downside: every negative scenario reads $0.00 and delta_value_vs_base is 0.0 throughout. The summary's delta_eps_y2_vs_base is a percentage change against a Base Y2 EPS of $0.15, and it carries the ranking — S3 −95%, S8 −90%, S4 −50%, S7b −49%, S2 −48%, S11 −42%; S5 +95%. Active/on watch per the playbook: S3 partially active; S8, S10, S11 on watch. S3 is the worst scenario for this name and is live now.

Model value change vs Base, by scenario
S1 Fast equity crash$0.00S2 Slow bear / recession$0.00S3 Rapid rate shock$0.00S4 Slow rate grind$0.00S6 Recession-driven cuts$0.00S7a Credit liquidity shock$0.00S7b Slow default cycle$0.00S8 Stagflation$0.00S9a Dollar spike$0.00S9b Dollar slide$0.00S11 Energy supply shock$0.00S12 Mega-cap/AI derating$0.00S10 Melt-up+$0.31S5 Soft-landing cuts+$1.30

What would change the call

Upgrades if: quarterly originations exceed ~$55bn with gain margin above 120bps for two consecutive quarters (S5 in the numbers, not the forecast) and preferred terms confirm no conversion with the dividend paid in cash; or a bulk MSR sale at or above carrying value retires the 2027 maturities. Downgrades further if: warehouse or MSR advance rates are cut; the $242mm agency net worth cushion falls below ~$100mm; the class action produces a hedge-accounting restatement; or the preferred accretes at 13% instead of paying 10% cash.

Watch items

Sources