Bank Lending Standards: The Recession Indicator Nobody Watches
Every quarter since 1990, the Federal Reserve has surveyed senior loan officers at dozens of banks with one question: "Are you tightening or easing lending standards?" The answers predict recessions with remarkable accuracy—an average lead time of 4 months. Yet almost nobody watches this data. We test 35 years of the Senior Loan Officer Opinion Survey (SLOOS) and find it outperforms the yield curve on timing. But there's a catch: 2 major false signal episodes since 1990, including a massive tightening in 2022-2023 that produced no recession.
What is SLOOS?
The Senior Loan Officer Opinion Survey on Bank Lending Practices (SLOOS) is a quarterly survey conducted by the Federal Reserve. They ask roughly 80 large domestic banks and 24 branches of foreign banks about:
- Commercial & Industrial (C&I) loans: Loans to businesses
- Consumer loans: Credit cards, auto loans, personal loans
- Real estate loans: Mortgages, commercial real estate
The key questions:
- Are you tightening or easing standards for approving loans?
- Is demand for loans increasing or decreasing?
- Why are you making these changes? (economy weakening, capital requirements, competition, etc.)
The Fed reports results as net percentages: (% tightening) - (% easing). A positive number means more banks are tightening than easing. A number above 20% typically signals significant credit contraction.
Why This Matters
Banks are forward-looking. When they tighten standards, they're saying: "We think economic conditions are deteriorating, so we're getting more selective about who we lend to." This happens before the economy actually weakens, because banks have access to real-time data on their borrowers' finances and payment histories.
Once standards tighten, credit growth slows. Businesses can't invest. Consumers can't spend. The economy contracts. It's a self-fulfilling prophecy, but one that's highly predictable.
Test 1: Does SLOOS Predict Recessions?
Let's look at the full history: net percentage of banks tightening C&I lending standards from 1990 to 2025, with NBER recession periods shaded.
Bank Lending Standards vs. Recessions (1990-2025)
Net percentage of banks tightening C&I lending standards (blue line). Gray shaded areas are NBER recessions. Notice how the blue line spikes before each gray bar. Average lead time: 4 months.
Recession Lead Times
| Recession | Peak Tightening | Tightening % | Lead Time |
|---|
Result: SLOOS has predicted all 4 recessions since 1990 with an average lead time of 4 months. Much shorter than the yield curve (12-18 months), but shorter leads are actually better for actionable signals.
Test 2: SLOOS vs. Yield Curve
We've previously tested the yield curve as a recession predictor. It works, but with long and variable lags (6-24 months). How does SLOOS compare?
Comparing Two Leading Indicators
Red: SLOOS tightening %. Blue: 10Y-2Y yield curve spread (negative = inverted = recession signal). Yield curve inversions typically lead recessions by 12-18 months. SLOOS spikes appear 4 months before recessions.
Indicator Comparison
| Indicator | Avg Lead Time | False Signals (1990-2025) | True Signals | Signal Accuracy |
|---|---|---|---|---|
| SLOOS Tightening >20% | 4 months | 2 episodes (1998, 2022-23) | 4 recessions predicted | 67% (4 of 6 signals) |
| Yield Curve Inversion | 12-18 months | 2 episodes (1998, 2022-25) | 2 recessions predicted (2001, 2008) | 50% (2 of 4 signals) |
The Key Difference: SLOOS leads by only 4 months (more actionable timing) and predicted all 4 recessions since 1990, giving 67% signal accuracy (4 true, 2 false). The yield curve leads by 12-18 months (harder to time) and only predicted 2 of 4 recessions, with 50% accuracy (2 true, 2 false). SLOOS missed COVID-2020 briefly but caught it during, while yield curve failed completely on the short COVID recession and hasn't been followed by recession in 2022-2025 despite inversion.
Test 3: The False Signal Problem
Here's the uncomfortable reality: since 1990, there have been multiple distinct tightening episodes where SLOOS crossed above 20%. Of these episodes, only the ones that actually led recessions were useful signals. The rest were false alarms where you'd have gone defensive for nothing.
False Signals: When Tightening Didn't Lead to Recession
Red shaded areas show tightening episodes above 20% that did NOT predict a recession (no recession within 6 months before or 12 months after the peak). Two major false signals visible: late 1990s and 2022-2023.
Notable False Signal Episodes
- Late 1990s (1998): Tightening during LTCM crisis and Asian financial contagion. Fed cut rates and economy recovered without recession.
- 2022-2023: Massive tightening (>60%!) after aggressive Fed rate hikes and regional bank failures (Silicon Valley Bank). Despite record-breaking tightening levels exceeding 2008, no recession materialized by late 2025.
Why These False Signals?
- 1998: Fed acted quickly with rate cuts to stabilize markets before credit crunch spread to real economy
- 2022-2023: Strong labor market, excess consumer savings from COVID stimulus, and low corporate refinancing needs created resilience despite tight credit
Test 4: Does Loan Growth Actually Decline?
When banks say they're tightening standards, does credit actually contract? Let's check by comparing SLOOS to year-over-year growth in commercial & industrial (C&I) loans outstanding.
"Loans outstanding" is the total dollar amount of C&I loans on bank balance sheets. When this number grows year-over-year, banks are extending more credit to businesses. When it shrinks, credit is contracting. We expect a strong correlation: tighter standards → fewer loan approvals → slower loan growth → possible credit crunch.
Lending Standards vs. Actual Loan Growth
Blue: SLOOS tightening % (left axis). Orange: C&I loan growth year-over-year (right axis). When standards tighten, loan growth slows with a lag of 1-2 quarters. But not all tightening produces negative loan growth.
The Lag Structure
Quarter 0 (Survey): Banks report tightening standards
Quarter 1-2: Loan growth begins to slow as new applications get rejected
Quarter 2-3: If severe, loan growth turns negative
Quarter 3-4: Recession begins (if credit contraction is deep enough)
Key Insight: The severity matters. Tightening from 10% to 30% might slow loan growth from 8% to 4% (no recession). Tightening from 10% to 70% might turn loan growth negative (recession likely).
Test 5: What About Consumer Credit?
We've focused on C&I (business) lending, but what about consumer lending standards? Do they predict recessions independently?
Consumer vs. Business Lending Standards
Blue: C&I lending standards (1990-present). Orange: Consumer lending standards (starts mid-1990s - FRED data not available earlier). Both tend to move together, but consumer standards lag slightly and are less volatile.
Consumer vs. Business Standards
C&I standards lead: Banks tighten business lending first because business loans are riskier and more sensitive to economic conditions.
Consumer standards follow: 1-2 quarters later, banks tighten consumer credit as unemployment rises and household finances weaken.
Exception - 2022: Consumer standards tightened MORE than C&I standards, possibly due to rising credit card delinquencies and auto loan defaults. This unusual pattern may explain why recession didn't materialize—businesses stayed healthy even as consumers struggled.
Test 6: The 2022-2023 Mystery
The most important test: what happened in 2022-2023? SLOOS showed the most extreme tightening since the 2008 financial crisis—over 60% net tightening in Q2 2023. Yet no recession came. What went wrong?
2022-2023: Record Tightening, No Recession
SLOOS tightening exceeded 60% in Q2 2023 following Silicon Valley Bank collapse and Fed rate hikes to 5.25%. Historical rule: anything above 40% predicts recession. This time: no recession through end of 2025.
Why 2022-2023 Was Different
- Fiscal support still flowing: Unlike past cycles, government transfer payments to households remained elevated through 2023. Consumers had cushion.
- Strong labor market: Unemployment stayed below 4%. Hard to have recession when everyone's employed.
- Sector-specific crisis: Banking stress was concentrated in regional banks exposed to commercial real estate and tech deposits. Money center banks stayed healthy.
- Pandemic excess savings: Households entered 2022 with record savings. Even as credit tightened, they could draw down cash.
- Corporate borrowing front-loaded: Many companies refinanced at low rates in 2020-2021. They didn't need new loans in 2022-2023 even as standards tightened.
The Lesson: SLOOS measures supply of credit (banks willing to lend). But recessions require both supply contraction AND demand for credit. If businesses/consumers don't need to borrow (because they're flush with cash or already locked in low rates), tight standards don't bite.
Key Takeaways
- SLOOS works for recession prediction: 4 of 4 recessions predicted since 1990, 4-month average lead time.
- But 67% signal accuracy: 2 false episodes (late 1990s, 2022-2023) out of 6 total tightening episodes.
- Beats the yield curve on timing: 4 months vs 12-18 months lead time. Yield curve has similar false signal rate (50% accuracy).
- 2022-2023 was the biggest false signal ever: 60%+ tightening with no recession. Combination of strong labor market, fiscal support, and low refinancing needs.
- Supply ≠ Demand: Tight lending standards only cause recession if businesses/consumers actually need credit. When they're flush with cash, standards don't matter.
- Use as confirmation, not standalone: SLOOS works best combined with other indicators like unemployment, yield curve, credit spreads, and actual loan growth data.
For Your Toolkit
SLOOS is worth monitoring as a confirmation signal:
- If yield curve inverts AND SLOOS tightens >30% → high recession probability
- If yield curve inverts but SLOOS stays loose → wait (like 2022-2025)
- If SLOOS tightens but yield curve is positive → probably false signal (like 2016)
Where to find it: Federal Reserve Board publishes SLOOS quarterly. FRED database has series DRTSCILM (C&I standards) and DRTSCLCC (consumer standards). Reports come out ~30 days after quarter end.
Data & Methodology
Data Sources:
- DRTSCILM: Net % of banks tightening C&I lending standards (Federal Reserve SLOOS, quarterly, 1990-2025)
- DRTSCLCC: Net % of banks tightening consumer lending standards (Federal Reserve SLOOS, quarterly, 1991-2025)
- BUSLOANS: Commercial and industrial loans outstanding (Federal Reserve H.8, monthly, 1990-2025)
- TOTALSL: Total consumer credit (Federal Reserve G.19, monthly, 1990-2025)
- USREC: NBER recession indicator (FRED, monthly, 1990-2025)
Calculations:
- Lead time: Months between peak SLOOS tightening (before recession) and official recession start
- False signal: SLOOS >20% tightening with no recession within 12 months
- Loan growth: Year-over-year % change in loans outstanding
- All data accessed via Federal Reserve FRED database using authenticated API