Bank Vulnerability System

FDIC + BBG Implied Vol + OFR NDFI | 66 US Banks | 2026-03-15 | 3 Pillars: Fundamentals + Price + Vol Signals
Concept: Rank banks by hidden risk — weak fundamentals the market isn't pricing, strong price suggesting complacency, vol signals showing options aren't pricing the stress. Backtest: 20-year test (2005–2026) shows Safe 10 returned 531% vs Danger 10's 185%. Model works for rate/duration crises but inverts during credit crises (GFC).
20-Year Backtest: Safe 10 vs Danger 10

Summary — 84 Quarters, 2005–2026

Safe 10 Total Return
+531%
100 → 631
Danger 10 Total Return
+185%
100 → 285
L/S Spread
2.2x
Safe outperformed 2.2x
12M Ann. Spread
+1.9%
Q1: 11.0% vs Q5: 9.1%
Quarters
84
2005-Q1 to 2025-Q4

Portfolio NAV: Safe 10 vs Danger 10 (rebased to 100)

Key events: GFC 2008 — both crash but Safe recovers faster. COVID 2020 — sharp drawdown, rapid recovery. SVB 2023 — Safe pulls away decisively. The spread compounds through crisis protection, not calm-market alpha.

Long/Short Spread (Safe / Danger)

Rising = model working. GFC (2007-2009) collapsed the spread from 100 to 56. Post-GFC recovery took until 2012. Since then, steady gains.

Model Performance by Market Regime

Built for duration/rate risk. Excels at SVB-type crises (+9.1%) but inverts during credit crises (GFC -2.4%) where off-balance-sheet exposure matters more.

Annualized Returns by Quintile

HorizonQ1 (Safe)Q2Q3Q4Q5 (Danger)Spread
3M3.5%3.0%2.8%3.2%2.1%+1.4%
6M6.1%5.5%5.7%5.5%4.5%+1.5%
12M11.0%9.8%10.9%10.4%9.1%+1.9%

Current Holdings (2025-Q4 Rebalance)

#Safe 10Danger 10
Next rebalance: 2026-03-31. Holdings update with fresh FDIC data, prices, and implied vol.

Quarter-by-Quarter: 6M Forward L/S Spread

QuarterEraQ5 6MQ1 6MSpreadN
Live Vulnerability Screen — March 2026

Summary

Vulnerability Map: Weak Fundamentals vs Price Strength

Top-right = danger zone. Weak fundamentals AND strong prices. Bubble border: purple = high Vol Signals (>70).

Vol Disconnect: IV/RV Ratio (lower = options underpricing)

CMA (IV/RV=0.63) and WBS (0.78) — implied vol far below realized moves. Options sellers asleep.

RV Term Structure: 10d vs 250d Realized Vol

Above diagonal = rising short-term vol. CMA (1.32x) and CADE (1.31x): 10d RV elevated vs 250d norm.

Vulnerability Score Distribution

Most vs Least Vulnerable: What Differs?

Private Credit / NDFI Exposure (Large Banks)

FCNCA 15.8% private credit. PNC (11.5% NDFI) #9 vulnerability.

Securities Exposure vs Price Complacency

Top-right = massive securities books + near highs = SVB pattern. BOH: 2.67x MktCap in HTM+AFS, only 8.5% from high.
Full Vulnerability Rankings
RkTickerName VulnFundPrice VolIVolRV10d RVrIV/RVP/B DDvsHiHTM AFSTCRNDFI MCap
Key Conclusions
1. BOH #1 (79.4%) — triple threat. Weak fundamentals (65: 2.67x MktCap in securities, thin TCE), strong price (82: near 1Y high, 1.93x book), AND high vol signals (91). The quintessential hidden risk bank.
2. CMA #2 (79.1%) — vol disconnect poster child. Vol Signals 99. IV only 28.4% while 10d RV is 45% — IV/RV of 0.63. Options massively underpricing actual moves.
3. CADE #3 (70.3%) — vol stress buildup. RV Ratio 1.31x (10d vol elevated vs 250d), Vol Signals 97.
4. The large bank split. WFC #10, TFC #11, BAC #13 all highly vulnerable with meaningful vol signals. JPM dropped to #48 — its IV/RV is 2.27x (options already cautious).
5. Least vulnerable have high IV/RV. PB (3.31), EGBN (3.56), TOWN (3.42) — options charging 3x+ realized moves. Already priced for trouble.
Data: BBG 3M ATM Implied Vol via Rose + Yahoo Finance + FDIC BankFind API Q4 2025 + OFR NDFI Reports | snow.ai · campbell ramble
snow.ai