The Stress Test Lies We Tell Ourselves: Why 2026 Regulatory Scenarios Are Already Broken

The Federal Reserve and FDIC released their 2026 stress testing scenarios on February 12. The industry read them, filed them, and moved on.

That's the wrong response.

The regulatory stress scenarios are politically constrained, not economically rigorous. They're designed to show that "banks are safe under reasonable conditions," not to capture actual economic tail risk. Banks that use regulatory scenarios as their primary stress-testing framework are flying blind.

The Regulatory Scenario (What They Assume)

The FDIC's 2026 stress test assumes:

Base Case (Baseline Scenario): - Real GDP growth: 1.2% in 2026, 1.5% in 2027 (weak but positive) - Unemployment: 4.8% (gradually declining) - 3-month Treasury rate: Averages 4.2% in 2026 - 10-year Treasury rate: Averages 4.1% - Commercial real estate valuations: Decline 15% over two years - Commercial real estate delinquencies: Rise to 6.5%

Severe Stress Case: - Real GDP growth: -0.5% in 2026 (mild recession) - Unemployment: 7.5% (elevated) - 3-month Treasury: 2.1% - 10-year Treasury: 2.9% - CRE valuations: Decline 35% - CRE delinquencies: Rise to 12%

Why These Scenarios Are Political Fiction

Assumption #1: Unemployment stays under 8% even in a "severe" scenario.

Historical fact: In the 2008-2009 recession, unemployment hit 10%. In the 2001 recession, it hit 5.5%. In the 2020 pandemic shock, it hit 14.8%.

A "severe" recession that only pushes unemployment to 7.5% is not severe. It's "uncomfortable but manageable."

If you stress-tested your loan portfolio assuming unemployment rises to 10% (a realistic tail scenario), credit losses would be 30-50% higher than the regulatory scenario implies.

Example: A bank with $5B in commercial loan portfolio would assume $75-100M in losses at 7.5% unemployment (regulatory scenario). At 10% unemployment (historical recession), losses would be $110-150M. That's a $35-50M underestimation.

Assumption #2: Rates don't spike.

The regulatory scenario assumes a "severe" case where the 10-year Treasury falls to 2.9%. But recent history: - August 2023 (start of crisis fears): 10-year peaked at 4.3% - September 2024 (Fed pivot): 10-year fell to 3.6% - Current (March 2026): 10-year at 4.1%

What if rates spike to 5.5% (not unprecedented—we hit 4.9% in 2023)? A 100+ bps move in rates creates: - Deposit outflows (savers move to higher-yielding money markets) - Securities losses (banks' bond portfolios mark down) - Loan demand collapse (higher rates kill borrowing appetite)

The regulatory scenario doesn't stress for a rates-UP scenario. It only stresses for rates-DOWN. This is backwards.

Assumption #3: CRE valuations decline 35% in the "severe" scenario, but delinquencies only hit 12%.

This is mathematically inconsistent.

If a commercial property falls from $100M to $65M in value, and the mortgage is $80M, the loan is now underwater by $15M. If the borrower refinances at this point, the bank takes a loss. If the property is owned by a leveraged sponsor (4-5x leverage), the equity is gone and the sponsor stops paying.

A 35% valuation decline should produce 20-25% delinquency rates, not 12%. The gap suggests the scenario is softening the stress to avoid shocking bank capital models.

Assumption #4: Your deposit base is stable.

The regulatory scenario doesn't assume major deposit flight. But we've seen this: - March 2023 (SVB crisis): Uninsured deposits fled at 3-4% per week - Banking stress periods: Deposit/loan flows reverse rapidly

A bank with 30% uninsured deposits could see 15-20% deposit outflow in a stress scenario. The regulatory baseline assumes maybe 3-5%.

If your bank relies on deposit funding and faces a 20% outflow, you're raising rates sharply or selling assets at losses. The regulatory scenario doesn't capture this.

What Your Stress Tests Should Actually Assume

Scenario 1: Mild Recession (2-Year Horizon) - GDP: -1.2% in 2026, +0.8% in 2027 - Unemployment: 7.2% (peaks mid-2026) - 10-year Treasury: Falls to 3.2% initially, then rises to 4.5% as recovery begins - CRE valuations: Decline 25% - CRE delinquencies: Rise to 10% - Deposits: 8% outflow (uninsured depositors move to money markets or larger banks) - Auto/card defaults: Double - Commercial loan losses: 2.5-3.5% of portfolio

Scenario 2: Sharp Recession (12-Month Shock) - GDP: -2.5% annualized - Unemployment: 8.8% (rapid deterioration) - 10-year Treasury: Falls to 2.8% (flight to safety), then rises to 4.2% as Fed cuts - CRE valuations: Decline 40% - CRE delinquencies: Rise to 18% - Deposits: 15% outflow (panic deposit flows to FDIC-insured competitors) - Auto/card defaults: Triple - Commercial loan losses: 3.5-5% of portfolio - Charge-offs: 1.5-2% of retail portfolio

Scenario 3: Systemic Shock (Tail Risk) - GDP: -3% annualized - Unemployment: 10.5% - 10-year Treasury: Falls to 2.2%, Treasury curve inverts sharply - CRE valuations: Decline 50%+ - CRE delinquencies: 25%+ - Deposits: 25-30% outflow (depositors panic, demand exceeds FDIC insurance limits) - Auto/card defaults: 4-5x normal - Commercial loan losses: 5%+ of portfolio - Securities losses: 8-12% of bond portfolio

These aren't hypothetical. The 2008-2009 cycle experienced something close to Scenario 3. The 2020 pandemic triggered Scenario 2. A sharp recession would land you in Scenario 2 territory.

Why This Matters for Capital Planning

Regulatory stress tests determine your Stress Capital Buffer (SCB) requirement. If you pass the regulatory scenario, you meet your SCB. But if you fail your own (more realistic) scenarios, you're under-capitalized relative to actual risk.

Example:

Regulatory scenario says: "Your bank needs 3.5% additional capital (SCB) to survive stress."

Your scenario analysis says: "If unemployment rises to 9% and CRE falls 40%, you need 6.5% additional capital."

The gap (3 percentage points of capital) is real risk you're not carrying. If stress hits, you'll discover it too late.

For a $10B bank, that's $300M in under-capitalization. You'll either: 1. Quickly raise capital at unfavorable terms 2. Cut dividends and damage shareholder confidence 3. Restrict lending and lose market share 4. Sell assets at losses

None of these are pleasant.

What Your Board Needs to Do This Quarter

1. Commission a parallel stress test.

Hire an external firm (Deloitte, Oliver Wyman, McKinsey) to run your own stress scenarios. Don't rely on your internal models alone—they're built to pass regulatory tests, not to capture edge cases.

Cost: $300-500K. Value: Priceless when a crisis hits and you've already positioned capital defensively.

2. Expand your scenario analysis.

Add three custom scenarios tailored to your bank's actual exposures: - If your bank is CRE-heavy: stress CRE valuations to -45% and delinquencies to 20% - If your bank is deposit-heavy: stress deposit outflows to 20% - If your bank has fintech/NBFI exposure: stress those borrowers separately

3. Stress your margin assumptions.

The biggest error banks make: assuming net interest margin holds steady in a stress scenario. It doesn't. - Deposit costs rise (depositors demand higher rates to stay) - Loan volumes fall (fewer borrowers want loans at higher rates) - NIM compresses sharply

Model this explicitly. For every 200 bps deposit cost increase, your NIM might fall 30-50 bps. That compounds.

4. Build contingency funding plans.

Assume you lose 15% of deposits in a stress event. Where will you get funding? - Raise rates further (expensive, erodes NIM) - Reduce lending (impacts revenue) - Sell securities (may realize losses) - Borrow from Fed (signals weakness)

Have actual plans for each. Test them quarterly.

5. Review capital management policy.

If your SCB is 3.5% but you believe actual need is 6.5%, adjust your dividend and buyback policies to build capital faster. Your shareholders won't love it short-term. But they'll appreciate not having a 40% equity drop in a crisis.

The Hard Truth

The regulatory stress scenarios are politically biased toward showing "the system is safe." They're not economically rigorous reflections of real tail risk.

Banks that use them as the primary stress-testing framework are under-capitalized relative to actual risk.

The ones that run their own rigorous scenarios, build margin assumptions from first principles, and stress-test for realistic tail events will: 1. Survive the next cycle with less damage 2. Have capital available for acquisitions when competitors are constrained 3. Build competitive advantage through superior planning

Start now. The next crisis is already being priced into credit markets. Your models should reflect that reality, not regulatory fiction.