Setup

Ready Capital (RC, RC-PE) reported Q1 2026 on May 8. The print confirmed the bear thesis on the common — but the most actionable insight isn't there. It's the cross-cohort comparison, which surfaces a $201.6M deferred tax asset bomb no peer flagged and a STWD mispricing hidden behind cohort tarring.

What the filing says

The new information is buried in Q&A. KBW's Jade Rahmani directly flagged the $201.6M deferred tax asset (DTA) on the balance sheet, asking about valuation allowance risk given two consecutive years of large losses. CFO Andrew Ahlborn: "we do think that deferred tax asset has value, but certainly we are aware of the magnitude." Non-answer, not denial. Sustained losses meet the GAAP "more likely than not" threshold for valuation allowance — point estimate -$0.97/share latent book value loss; range -$0.65 to -$1.30. Auditor likely forces the issue at FY2026 audit (~Feb 2027).

Surrounding numbers: book value collapsed -15.5% in one quarter ($8.79 → $7.43), nearly matching all of 2025 in 90 days. Distributable earnings (DE) -$1.00/share (annualized -$4). Management guided net interest income (NII) NEGATIVE going forward. $2-$2.5B distressed asset sales planned through Q4 2026; CFO declined a Q2 book value floor when pressed. Common dividend declared at $0.01/quarter (eliminated in all but name); preferred dividends maintained, common equity buffer ≈9.5x preferred liquidation.

What the market thinks

RC common trades at 0.24x book, 17% short interest, P/C 2.07. RC-PE preferred yields 12.30%, RSI 82.9 after a +23% MoM move — defensive yield-grab already happened. STWD RSI 31.6, recently off — cohort-tarred despite multi-cylinder positioning.

Why the gap exists

Three specific gaps, ranked by edge:

1. DTA risk is uniquely unflagged. Cohort cross-check across 8 CRE mREITs (STWD, BXMT, KREF, ABR, FBRT, TRTX, BRSP, ARI) — zero peers mentioned DTA valuation allowance risk in Q1 2026 earnings calls. Sell-side models assume DTA realizable absent explicit guidance. Auditor-forced timing creates a discrete catalyst at FY2026 audit.

2. Cohort tarring on STWD. STWD Q1 print held Bucket 4 (multi-cylinder defender): $0.39 DE vs $0.48 dividend maintained without cut, supported by $1.4-1.5B unrealized multifamily gains, LNR servicer counter-cyclical income, and Fundamental Income spin-out optionality. RC alone in Bucket 6 (terminal) — book value trajectory ≈5x worse than next-worst peer (KREF -9.0%). Market hasn't distinguished the buckets.

3. Microcap mechanics decouple instrument from thesis. RC at $300M cap + 17% SI means tape responds to sector flow even when fundamentals diverge. Crowded short with mediocre net EV after borrow (15-20%) and squeeze risk. The fundamental analysis is right; the standalone short instrument is washed out.

Risks (ranked)

1. STWD/RC pair beta to MORT unmeasured. If beta differential >0.3, pair carries unintended sector exposure. Regression required before deployment.
2. STWD Q2 print Aug 8-12 disappoints. DE <$0.35 breaks the multi-cylinder thesis; 5 active STWD predictions resolve adversely.
3. RC squeeze on cohort relief rally. 17% SI plus microcap dynamics = double-edged on the short side.
4. RC borrow cost quoted 8-15%, unconfirmed. If 20%+, pair EV drops to ~+15% net.
5. Cohort relief lifts both legs equally. Pair edge collapses without conviction realization.

Catalysts

• 2026-08-08 to 2026-08-12: STWD Q2 print (resolves the multi-cylinder dividend coverage test)
• ≈2026-08 mid: RC Q2 10-Q (book value trajectory test)
• ≈2026-11 mid: RC Q3 10-Q (DTA disclosure window opens)
• ≈2027-02-28: RC FY2026 10-K + audit (forces DTA allowance decision)

What would change our mind

• RC Q2 print shows BV > $7.00 — execution materially better than guided
• Auditor signs off on DTA without allowance at FY2026 audit — primary edge collapses
• STWD cuts dividend at Q2 print — multi-cylinder thesis breaks
• Q2 cohort prints show peers also -10%+ BV — trichotomy compresses to sectoral, idio LR weakens
• RC announces capital raise / equity issuance — restructuring path more probable than modeled

Evidence