Bank Reserve Quality: JPM's 5.43x vs BAC's 2.48x Coverage
All three mega-banks maintain similar capital adequacy (TCE ratios 6.7-6.9%), but reserve quality tells a divergent story: JPM 5.43x vs BAC 2.48x coverage, BAC's flat ALLL despite highest NPLs, JPM's hidden deterioration (NPAs up 23%), WFC turnaround validated (lowest NCOs). The 4-metric framework reveals which banks are prepared for credit stress.
TL;DR
All three mega-banks maintain similar capital adequacy (TCE ratios 6.7-6.9%), but reserve quality tells a divergent story:
- Reserve coverage varies 2.2x: JPM 5.43x vs BAC 2.48x
- BAC's misalignment: Flat ALLL ($13.3B for 3 quarters) despite highest NPLs ($5.3B) and improving credit narrative
- JPM's hidden deterioration: NPAs up 23% to $10.6B, NCOs up $506M—but 5.43x coverage provides buffer
- WFC turnaround validated: Lowest NCOs ($944M vs JPM's $3.2B), 4.54x coverage confirms credit improvement
- HTM concentration risk: BAC's $446.5B HTM securities (1.6x JPM, 2.5x WFC) adds layered risk atop lean reserves
The 4-metric framework (Reserve Coverage, ALLL Trend, NCO Intensity, TCE Ratio) reveals which banks are prepared for credit stress—and which are running reserves too lean.
The Question ROIC Can't Answer
Traditional return metrics break down for banks. JPMorgan's ROIC appears negative in standard calculations, not because the business destroys value, but because banking economics don't fit industrial capital models.
Banks transform assets through 10:1+ leverage. A manufacturer invests $100M in PP&E to generate returns. A bank takes $100M in equity, levers it to $1T in loans, and earns on the spread. ROIC denominators explode while numerators stay modest—making the ratio meaningless.
ROE suffers from different problems. High ROE can signal excellent management or dangerous over-leverage. Reserve releases can inflate ROE by 200+ basis points temporarily, pulling forward future losses. Without context, 15% ROE tells you nothing about sustainability.
The real question: Which banks maintain adequate loss reserves while reporting earnings?
This requires a framework grounded in credit risk and earnings quality, not return metrics designed for manufacturers.
The 4-Metric Reserve Quality Framework
Metric 1: Reserve Coverage Ratio (ALLL / NPLs)
What it measures: How many dollars of allowance exist per dollar of non-performing loans.
Why it matters: Banks estimate future credit losses through ALLL (Allowance for Loan and Lease Losses). When loans become non-performing, ALLL should already cover expected losses. Low coverage means either (1) credit quality is genuinely strong, or (2) reserves are inadequate.
| Bank | ALLL | NPLs | Coverage Ratio | What It Reveals |
|---|---|---|---|---|
| JPM | $25.7B | $4.7B | 5.43x | Every $1 NPL backed by $5.43 reserves—conservative even for pristine credit |
| WFC | $14.3B | $3.2B | 4.54x | Strong coverage validates turnaround narrative |
| BAC | $13.3B | $5.3B | 2.48x | Despite highest NPLs, lowest reserves—either confident or lean |
The 2.2x spread between JPM and BAC is massive. BAC carries $5.3B in non-performing loans (13% more than JPM's $4.7B) but only $13.3B in reserves (48% less than JPM's $25.7B).
If BAC's credit quality matched JPM's conservatism, they would need $28.8B in reserves—implying a $15.5B reserve shortfall.
Alternative interpretation: BAC's risk models genuinely show lower expected losses. Their consumer banking exposure (credit cards, mortgages) may have better recovery rates than JPM's wholesale/corporate concentration.
Metric 2: ALLL Trend (Quarterly Momentum)
What it measures: Whether reserves are growing, stable, or declining.
Why it matters: ALLL should track loan portfolio risk. Flat reserves amid rising NPLs suggest under-reserving. Declining reserves during credit improvement signal appropriate adjustment.
| Bank | Q3 2025 | Q2 2025 | Q1 2025 | Direction | What It Reveals |
|---|---|---|---|---|---|
| JPM | $25.7B | $25.0B | $25.2B | Stable (~$25B) | Maintaining reserves despite loan growth—absorbing credit deterioration |
| BAC | $13.3B | $13.3B | $13.3B | Frozen | No reserve build for 3 quarters despite highest NPLs |
| WFC | $14.3B | $14.6B | $14.6B | Declining | Reducing reserves by $300M amid improving credit—appropriate adjustment |
BAC's flat ALLL is the red flag. NPLs sit at $5.3B (highest among the three), yet reserves haven't budged from $13.3B in three quarters. Either:
- Credit quality stabilized perfectly at current NPL levels (unlikely)
- Management is confident provisions will decline faster than NPLs rise
- Reserves are being kept lean to support earnings and capital returns
WFC's declining ALLL ($14.6B → $14.3B) makes sense: NCOs fell to $944M (lowest among three banks), validating reserve release.
Metric 3: NCO Intensity (Credit Stress Indicator)
What it measures: Net charge-offs relative to revenue or loan volume (using revenue as proxy since avg loan balances aren't in XBRL).
Why it matters: Charge-offs are realized losses—actual loans written off after recovery efforts fail. High NCO intensity signals credit stress regardless of reserve levels.
| Bank | NCOs (Q3 2025) | Revenue (Q3 2025) | NCO/Revenue | What It Reveals |
|---|---|---|---|---|
| JPM | $3.2B | $46.4B | 6.9% | Highest credit stress: $3.2B in losses on $46.4B revenue |
| BAC | $1.7B | ~$32B (est) | ~5.3% | Moderate stress: lower losses per revenue dollar |
| WFC | $944M | ~$31B (est) | ~3.0% | Lowest stress: credit quality improving |
JPM's NCO intensity is 2.3x WFC's ($3.2B vs $944M). This aligns with the narrative data: JPM's NPAs jumped 23% to $10.6B, driven by wholesale downgrades and California wildfire consumer loans.
Scale obscures the problem. $3.2B in quarterly charge-offs sounds manageable for a $4.1T balance sheet—until you realize it's up $506M year-over-year. The trajectory matters more than the absolute level.
WFC's $944M NCOs validate the turnaround story. Lowest absolute losses, lowest intensity, declining trend. Post-scandal skepticism should fade when credit metrics confirm improvement.
Metric 4: Tangible Common Equity Ratio
What it measures: Tangible equity (excluding goodwill/intangibles) divided by risk-weighted assets.
Why it matters: TCE represents loss-absorbing capital after stripping out accounting fluff. 6-7% TCE is typical for large banks; below 5% raises regulatory concerns.
| Bank | TCE Ratio | 8Q Trend | What It Reveals |
|---|---|---|---|
| BAC | 6.85% | Stable | Highest capital cushion—regulatory buffer intact |
| WFC | 6.72% | -0.33% (declining) | Adequate but trending down—capital returns outpacing earnings |
| JPM | 6.71% | -0.08% (stable) | Adequate capital despite credit deterioration |
All three banks sit in the 6.7-6.9% range—capital adequacy is not the differentiator. The question isn't "Do they have enough capital?" but "Are their reserves aligned with credit risk?"
BAC's 6.85% TCE is highest, yet they carry the leanest reserves (2.48x coverage). This creates an asymmetric risk profile: strong capital but thin credit cushion.
JPM: Conservative Reserves Absorb Hidden Deterioration
Reserve Quality Rank: #1
JPMorgan's 5.43x reserve coverage is 2.2x BAC's and 1.2x WFC's. This isn't paranoia—it's prudent preparation for deteriorating credit.
The Hidden Deterioration
Non-performing assets surged 23% to $10.6B (Q3 2025 vs Q3 2024). Management attributes this to:
- Wholesale downgrades in "certain industries" (unspecified—likely commercial real estate or energy)
- California wildfire impact on consumer loans (January 2025 event creating Q3 spillover)
- Higher loans at fair value in Corporate & Investment Banking
Net charge-offs climbed to $3.2B (up $506M), driven by Wholesale and Card Services. TTM NCOs hit $11.8B—the highest absolute losses among the three banks.
Why Reserves Provide Buffer
Despite NPAs +23%, JPM's 5.43x coverage means every dollar of NPL has $5.43 in reserves backing it. Even if credit deteriorates further, the cushion absorbs losses without forcing emergency reserve builds that would crater earnings.
Key quote from 10-Q:
"The Firm's nonperforming assets totaled $10.6 billion at September 30, 2025, up 23%, driven by higher wholesale nonaccrual loans, reflecting downgrades in certain industries, and higher consumer nonaccrual loans, predominantly due to the impact of the wildfires in California."
This is management acknowledging credit stress—not dismissing it.
Bull vs Bear Case
| Bull Case | Bear Case |
|---|---|
| 5.43x coverage is highest among peers—prepared for stress | 23% NPA surge signals early-stage credit cycle turn |
| $956B HQLA, $1.5T liquidity sources provide fortress balance sheet | Wholesale downgrades suggest CRE/energy exposure deteriorating |
| Markets/IB revenue +17% offsets NII margin compression | NCOs up $506M despite strong economy—recessionary losses would spike |
| Efficiency ratio 52.3% shows operating leverage | Card Services NCOs rising despite consumer strength |
Honest assessment: JPM's conservative reserves are appropriate given deteriorating NPAs. The 5.43x coverage isn't over-reserved—it's correctly reserved for a bank experiencing credit deterioration.
BAC: The Reserve-Credit Quality Misalignment
Reserve Quality Rank: #3 (Concerning)
Bank of America presents a paradox: improving credit narrative but lean reserves creating structural risk.
The Misalignment
Provision for credit losses decreased, signaling management confidence in credit quality. Consumer Banking net income rose $750M. Earnings quality score is highest among the three (8/10 vs JPM 7/10).
Yet:
- Lowest reserve coverage (2.48x vs JPM 5.43x)
- Highest NPLs ($5.3B vs JPM $4.7B, WFC $3.2B)
- Flat ALLL ($13.3B for 3 consecutive quarters)
The question: Is credit quality genuinely improving, or are lean reserves inflating earnings?
The Hidden Risk Layer: HTM Securities
BAC carries $446.5B in HTM securities (held-to-maturity bonds)—1.6x JPM's $274.9B and 2.5x WFC's $180.5B.
HTM accounting lets banks avoid marking bonds to market. If interest rates rise, unrealized losses sit off the income statement (in accumulated OCI). SVB's collapse proved HTM portfolios can hide capital erosion until liquidity demands force sales.
BAC's dual risk: Lean credit reserves (2.48x coverage) AND largest HTM exposure creates layered vulnerability. If credit deteriorates OR rates spike, capital gets squeezed from both sides.
Why Reserves May Be Intentionally Lean
Capital return commitment: $40B share repurchase program authorized. Running lean reserves frees capital for buybacks, boosting EPS and ROE in the near term.
Consumer exposure thesis: BAC's loan mix skews consumer (credit cards, mortgages) vs JPM's wholesale concentration. Recovery rates on consumer loans may justify lower coverage ratios.
Regulatory confidence: If BAC's stress test models show low expected losses, regulators approve lean reserves. External investors don't see the stress test assumptions.
Bull vs Bear Case
| Bull Case | Bear Case |
|---|---|
| Provisions declining validates credit improvement | Lowest coverage (2.48x) amid highest NPLs ($5.3B) is misaligned |
| Highest earnings quality score (8/10) shows clean accounting | Flat ALLL for 3 quarters suggests reserve freeze, not improvement |
| Consumer banking +$750M net income shows segment strength | HTM exposure $446.5B adds rate risk atop lean credit reserves |
| 6.85% TCE highest among three—strong capital cushion | $40B buyback drains capital while reserves stay flat |
Honest assessment: BAC's lean reserves create asymmetric risk. If credit deteriorates, they'll need emergency reserve builds that would wipe out several quarters of earnings. The 2.48x coverage provides minimal margin for error.
WFC: Turnaround Validated by NCO Data
Reserve Quality Rank: #2 (Strong)
Wells Fargo's turnaround story gets validated by credit data, not just narrative.
The NCO Proof Point
$944M in quarterly NCOs (vs JPM $3.2B, BAC $1.7B) is the lowest absolute losses and lowest intensity (~3% of revenue). TTM NCOs of $4.0B are 66% lower than JPM's $11.8B despite similar asset bases.
This isn't management spin—it's realized losses hitting the income statement.
ALLL declining from $14.6B to $14.3B makes sense in this context. If credit quality genuinely improved, reserves should adjust downward. The 4.54x coverage remains strong even after the reduction.
The Remaining Risks
VIE exposure: $13.4B in assets, $4.0B in liabilities from variable interest entities. Higher than JPM/BAC and reflects off-balance sheet complexity.
Legal proceedings: $6.2B interchange settlement, ongoing regulatory investigations. Hidden liabilities risk rated "high" in 5-pass extraction.
Post-scandal credibility: Will regulators/investors give WFC credit for improvement, or will skepticism persist?
Bull vs Bear Case
| Bull Case | Bear Case |
|---|---|
| Lowest NCOs ($944M) validate credit improvement narrative | VIE exposure $13.4B creates off-balance risk |
| 4.54x coverage provides strong buffer despite ALLL decline | Legal proceedings $6.2B settlement overhang |
| Commercial real estate showing improvement per MD&A | Post-scandal credibility requires sustained performance |
| TCE 6.72% adequate capital for current risk profile | ALLL declining while regulatory scrutiny persists |
Honest assessment: WFC's reserve quality is strong (4.54x coverage). The turnaround is real—NCO data confirms it. But VIE exposure and legal overhang create tail risks not present in JPM/BAC.
The Hidden Layer: HTM Bond Exposure
Held-to-maturity securities represent a second layer of hidden risk beyond credit reserves.
| Bank | HTM Securities (FV) | Unrealized Loss | % of Total Assets | What It Reveals |
|---|---|---|---|---|
| BAC | $446.5B | -$84.9B | ~14% (est) | Largest exposure + largest unrealized loss (19% of fair value) |
| WFC | $180.5B | -$33.7B | ~9% (est) | Moderate unrealized loss (18.7% of fair value) |
| JPM | $274.9B | -$18.5B | ~7% (est) | Smallest unrealized loss (6.7% of fair value) despite larger portfolio |
Why HTM matters: Banks classify bonds as HTM to avoid mark-to-market volatility. Unrealized losses bypass the income statement until bonds mature or are sold. This works until:
- Liquidity stress forces sales (SVB scenario)
- Capital ratios include AOCI (unrealized losses hit regulatory capital)
- Duration mismatch creates run risk (short-term deposits funding long-term bonds)
BAC's $446.5B HTM portfolio is 2.5x WFC's, with $84.9B in unrealized losses. Combined with 2.48x reserve coverage (lowest among three), BAC carries compounding dual risk: lean credit reserves ($13.3B ALLL) AND massive interest rate exposure ($84.9B unrealized loss—6.4x their total reserves).
If rates spike or credit deteriorates, both risks converge on capital—forcing either reserve builds (hitting earnings) or HTM sales (crystallizing $84.9B in losses). This dual exposure makes BAC uniquely vulnerable to simultaneous credit and rate stress.
Recovery & Deterioration Signals
What to Watch for JPM
Deterioration signals:
- NPA growth >10% for 2+ consecutive quarters
- Wholesale NCO acceleration (commercial real estate deterioration)
- Reserve build >$2B in single quarter (emergency provisioning)
Stabilization signals:
- NPA growth <5% quarterly
- Card Services NCO plateau (consumer credit stabilizing)
- Efficiency ratio <50% (operating leverage offsetting credit costs)
What to Watch for BAC
Deterioration signals:
- Reserve coverage falls below 2.0x (critically lean)
- NPL growth while ALLL stays flat (misalignment widening)
- HTM losses crystalized through sales (capital pressure)
Improvement signals:
- Reserve build to 3.0x+ coverage (acknowledging current lean state)
- NPL decline >10% (credit quality truly improving)
- Consumer banking NCO rate <2% (sustainable credit performance)
What to Watch for WFC
Deterioration signals:
- NCO reversal >20% increase (turnaround stalling)
- VIE losses materializing (off-balance risk hits income)
- Reserve coverage falls below 4.0x (undermining turnaround credibility)
Turnaround confirmation signals:
- NCO/revenue staying <3% for 8+ quarters (sustained performance)
- Legal settlements finalizing <$10B total (tail risk capped)
- Regulatory restrictions lifted (operational freedom restored)
Bottom Line: Reserve Quality Hierarchy
The three banks maintain similar capital adequacy (TCE 6.7-6.9%) but reserve quality varies dramatically:
1. JPMorgan (5.43x coverage) - Most conservative
- Appropriate for deteriorating credit (NPAs +23%, NCOs up)
- Provides buffer against further wholesale downgrades
- Highest absolute reserves ($25.7B) for highest risk
2. Wells Fargo (4.54x coverage) - Strong
- Validates turnaround with lowest NCOs ($944M)
- ALLL decline ($14.6B → $14.3B) appropriate given improving credit
- VIE/legal risks create tail exposure
3. Bank of America (2.48x coverage) - Adequate but Concerning
- Leanest reserves despite highest NPLs ($5.3B)
- Flat ALLL for 3 quarters while credit narrative "improves"
- HTM exposure $446.5B adds rate risk atop lean credit reserves
The key insight: Reserve coverage matters more than static capital ratios. BAC's 6.85% TCE looks strong until you see 2.48x reserve coverage—creating asymmetric risk if credit deteriorates.
JPM's 5.43x coverage isn't paranoia. It's prudent preparation for a credit cycle that's already deteriorating (NPAs +23%). BAC's 2.48x coverage isn't confidence. It's lean reserves inflating earnings and capital returns today at the expense of resilience tomorrow.
Investors should monitor reserve coverage direction, not just absolute levels. A bank moving from 4.0x to 3.0x is signaling rising confidence or hidden stress—context determines which.
FAQs
Q: Why can't I use ROIC to evaluate banks?
ROIC breaks down for banks because 10:1+ leverage explodes the invested capital denominator. A manufacturer invests $100M in PP&E to generate returns. A bank levers $100M equity to $1T in loans, making ROIC calculations meaningless. Use reserve coverage and TCE ratios instead.
Q: What is a "good" reserve coverage ratio?
3.0-5.0x is typical for large banks during normal credit cycles. Above 5.0x (like JPM's 5.43x) suggests conservative reserving or deteriorating credit. Below 2.5x (BAC's 2.48x) suggests lean reserves or genuinely strong credit. Context determines if lean reserves are appropriate or risky.
Q: Why is BAC's flat ALLL concerning?
Reserves should track credit risk. BAC's ALLL frozen at $13.3B for 3 consecutive quarters while NPLs sit at $5.3B (highest among three banks) suggests reserves aren't keeping pace with risk. Either credit stabilized perfectly (unlikely) or reserves are being kept lean to support earnings.
Q: How does HTM accounting hide risk?
Held-to-maturity bonds bypass mark-to-market accounting. If rates rise, unrealized losses sit in accumulated OCI (equity section) rather than hitting the income statement. This works until liquidity stress forces sales or regulatory capital includes AOCI—then hidden losses become real.
Q: Is Wells Fargo's turnaround real or narrative?
Real. NCO data validates it: $944M quarterly losses (lowest among three), TTM NCOs $4.0B (66% below JPM). You can spin narratives, but you can't fake realized losses hitting the income statement. WFC's credit quality has genuinely improved.
Q: Should I worry about JPM's 23% NPA increase?
Yes, but context matters. JPM's 5.43x reserve coverage absorbs the deterioration without forcing emergency reserve builds. The NPA surge is an early warning signal of credit cycle turn, but conservative reserves provide buffer. Watch if NPAs grow another 20%+ in Q4—that would signal deeper problems.
Q: Why does BAC have the lowest reserves but highest earnings quality score?
Earnings quality (8/10) measures accounting cleanliness, not reserve adequacy. BAC's score reflects low SBC, clean revenue recognition, and transparent reporting. Reserve coverage is a separate dimension measuring credit risk preparation. BAC can have clean accounting AND lean reserves simultaneously.
Q: What's the biggest risk to BAC?
Dual layer: lean credit reserves (2.48x) AND largest HTM exposure ($446.5B). If credit deteriorates OR rates spike, capital gets squeezed from both sides. A recession forcing reserve builds while HTM unrealized losses hit regulatory capital would pressure BAC hardest among the three.
Q: Can WFC reduce reserves while still in turnaround?
Yes, if credit quality genuinely improved. ALLL should track expected losses, not past sins. WFC's NCOs fell to $944M (lowest among three)—reserves declining from $14.6B to $14.3B reflects appropriate adjustment. The 4.54x coverage remains strong even after reduction.
Q: How do I monitor reserve quality going forward?
Track two metrics quarterly: (1) Reserve coverage ratio (ALLL / NPLs) and (2) ALLL trend direction. If coverage falls below 2.5x or ALLL declines while NPLs rise, that signals potential under-reserving. Use NCO trends to validate if reserve changes align with actual credit performance.
Q: Why isn't reserve coverage regulated like capital ratios?
CECL (Current Expected Credit Loss) accounting forces banks to estimate lifetime losses upfront, but banks retain discretion over models and assumptions. Regulators review stress tests but don't mandate specific coverage ratios. This creates room for banks to run lean reserves during good times—creating risk when credit turns.
Q: Should regional banks use this framework?
Yes, but adjust thresholds. Regional banks often carry 2.0-3.0x coverage due to concentrated portfolios and higher risk. Watch for coverage <1.5x or declining reserves amid rising NPLs. The framework applies universally: compare reserve coverage, ALLL trends, NCO intensity, and capital adequacy within peer groups.
Explore More
This analysis is part of our Earnings Quality Analysis framework.
Related Analyses:
- Earnings Quality: The Complete Framework — Full methodology for assessing earnings sustainability
- Cash Flow Quality Analysis Framework — Understanding cash conversion and accrual quality
- FinTech Earnings Quality: PayPal, Block, Coinbase — Similar reserve analysis for fintech lenders
Disclosure
This analysis uses SEC 10-Q filings (Q3 2025) and XBRL-extracted metrics. All quantitative data is sourced from official filings; no proprietary models or non-public information is used.
Reserve coverage ratios are computed as ALLL / NPLs per filed balance sheets. NCO/Revenue ratios use revenue as proxy since average loan balances aren't standardized in XBRL.
This is not investment advice. The author may hold positions in mentioned securities. Always conduct your own due diligence.
Generated with SEC filing analysis and XBRL data extraction. See filing intelligence pages for detailed source data.
MetricDuck Research
SEC filing analysis and XBRL data extraction