Call Report Ledger U.S. Bank Filings by Meru AI
Research Briefing 02 ยท Balance-Sheet Duration & AOCI Accounting

HTM & AFS Unrealized Securities Losses: How Rate-Driven Mark-to-Market Drag Impacts Adjusted Tier 1 Capital Ratios

By Meru AI Published Reporting Period 2026-Q2 Rev 4d6b48f119dcfefe

Reconciling FFIEC Schedule RC-B securities disclosures across 56,703 quarterly Call Reports from 2019-Q4 through 2026-Q2 tracks the full interest-rate duration cycle across U.S. banks, from an aggregate -$11,493.16M unrealized trough in 2023-Q4 to +$5,572.29M in 2026-Q2 while residual underwater bond books persist at individual charters.

GAAP vs. Regulatory Capital: How the AOCI Opt-Out Election Masks Duration Risk

Under U.S. bank capital rules (12 CFR Part 324), only Global Systemically Important Banks and Category I/II institutions are required to flow Accumulated Other Comprehensive Income (AOCI) unrealized gains and losses on Available-for-Sale (AFS) debt securities directly into Common Equity Tier 1 (CET1) and Tier 1 Leverage (RBC1AAJ). Almost all Category III, Category IV, regional, and community banks elect the AOCI neutral filter, excluding AFS mark-to-market fluctuations from regulatory capital ratios, while Held-to-Maturity (HTM) securities are carried at amortized historical cost on Schedule RC unless sold or credit-impaired.

Moreover, 1,814 community banks (42.0% of all active U.S. charters) operate under the Community Bank Leverage Ratio (CBLRIND = 1) framework pursuant to 12 CFR Section 324.12, maintaining a streamlined Tier 1 Leverage Cushion (median 11.00% nationwide) without filing Basel III risk-weighted CET1 schedules. Consequently, statutory regulatory capital ratios remain smooth even when rising Treasury and agency mortgage-backed security (MBS) yields depress the liquidation value of long-duration bond portfolios.

Table 1. Multi-Year U.S. Banking Industry Balance-Sheet, Leverage & Unrealized Securities Trajectory (2019-Q4 to 2026-Q2)
Filing Period Active Banks Total Assets Median Tier 1 Lev. Uninsured Dep. % Net Securities Mark ($M)
2019-Q4 5,244 $18.81T 10.92% 49.14% +$3,977.8M
2020-Q4 5,067 $22.05T 10.14% 51.43% +$8,143.0M
2021-Q4 4,904 $23.94T 9.97% 52.07% +$3,010.3M
2022-Q4 4,773 $23.80T 10.29% 49.97% $-3,904.3M
2023-Q1 4,740 $23.92T 10.37% 47.54% $-2,176.4M
2023-Q4 4,658 $23.90T 10.57% 45.92% $-11,493.2M
2024-Q4 4,560 $24.36T 10.65% 46.90% $-6,184.9M
2025-Q2 4,494 $25.23T 10.84% 47.85% $-2,247.8M
2025-Q3 4,452 $25.37T 10.91% 48.06% $-3,218.0M
2025-Q4 4,411 $25.54T 10.83% 48.34% $-4,079.2M
2026-Q1 4,353 $26.39T 10.90% 48.80% +$45.0M
2026-Q2 4,313 $26.75T 11.00% 50.01% +$5,572.3M

Tracing the 2020-to-2026 Bond Portfolio Cycle Across FFIEC Filings

Table 1 captures the complete arc of post-2020 balance-sheet duration risk across FDIC-insured institutions. Between 2019-Q4 and 2021-Q4, aggregate U.S. deposits surged by $5.18T (from $14.63T to $19.82T). Facing tepid commercial loan demand during 2020 and 2021, bank treasury desks deployed excess liquidity into 15- and 30-year agency MBS, municipal bonds, and intermediate U.S. Treasuries yielding 1.25% to 2.00%.

When the Federal Open Market Committee tightened policy rates by 525 basis points during 2022 and 2023, reported net securities marks swung from +$8,142.98M in 2020-Q4 to -$3,904.26M in 2022-Q4 and bottomed at -$11,493.16M in 2023-Q4. Notice that throughout that entire mark-to-market drawdown, reported median Tier 1 Leverage actually rose from 9.97% (2021-Q4) to 10.57% (2023-Q4) via retained earnings, proving why unadjusted regulatory capital alone failed to warn depositors at Silicon Valley Bank (CERT 24735) before forced AFS sales crystallized losses in March 2023.

Seasoning, Loss-Trade Restructurings & Residual 2026 Drag

Through 2025-Q4 (-$4,079.23M), 2026-Q1 (+$44.96M), and 2026-Q2 (+$5,572.29M), principal paydowns, curve roll-down, and deliberate loss-trade balance-sheet restructurings restored aggregate net marks to positive territory. Yet dispersion across individual charters remains meaningful, as shown in Table 2.

Table 2. Active U.S. Bank Charters Carrying Largest Residual Negative Unrealized Securities Marks (2026-Q2)
FDIC CERT Institution Name HQ Total Assets Securities Book Unrealized Mark Tier 1 Leverage
CERT 24045 Banc of California CA $34,899M $4,484M $-253.1M 9.64%
CERT 12441 Hancock Whitney Bank MS $36,329M $7,891M $-98.6M 10.59%
CERT 24998 Commerce Bank MO $35,017M $8,355M $-88.9M 11.43%
CERT 6548 U.S. Bank National Association OH $705,555M $162,532M $-82.5M 9.49%
CERT 639 The Bank of New York Mellon NY $425,115M $155,520M $-78.0M 6.20%
CERT 131 Seacoast National Bank FL $21,336M $5,752M $-39.6M 9.84%
CERT 58629 Stellar Bank TX $10,405M $1,325M $-34.7M 10.48%
CERT 35583 Pinnacle Bank TN $128,792M $20,629M $-26.6M 8.97%

How Counterparty Analysts Compute Adjusted Tangible Capital

Institutional credit officers stress-test a bank's true loss-absorbing runway by deducting tax-effected unrealized AFS and HTM marks from Tier 1 capital and comparing the resulting adjusted leverage ratio against uninsured deposit flight risk. So long as core retail deposits remain sticky and Federal Home Loan Bank (FHLB) or Federal Reserve discount-window collateral capacity exceeds uninsured balances, negative bond marks merely compress net interest margin (NIMY) rather than threatening solvency.

Conversely, if an institution carries both deep HTM unrealized discounts and heavy reliance on rate-sensitive wholesale or uninsured institutional deposits, rising funding costs squeeze net interest income while restricting balance-sheet flexibility because selling even a small slice of an HTM pool taints the entire HTM classification under ASC 320 accounting standards. Follow our step-by-step screening checklist in Briefing 04 or compare institutions directly in the Compare Banks Workbench.

Methodology & Analytical Scope Limitations

Unrealized securities figures reflect FDIC BankFind IGLSEC / equity-unrealized schedules scaled deterministically from $000s to $M. Fair-value interest-rate swaps hedging AFS portfolios are reported on separate derivative schedules and can partially offset raw bond-price sensitivity; see Methodology for full reconciliation checks.