How Accurate Are Recession Indicators Historically?
No single indicator is perfectly reliable. The yield curve, credit spreads, the Sahm Rule, and the Conference Board LEI each have strong track records but different lead times and occasional false signals. Combining multiple indicators — rather than relying on one — produces the most robust recession probability framework.
Looking for more macro answers?
In this article
The short answer
If there were a single, infallible recession indicator, everyone would use it — and the recession it predicted would never happen (because everyone would adjust their behavior in advance). The fact that recessions still occur tells you that prediction is inherently imperfect.
That said, several indicators have remarkably strong historical track records. The challenge is that each has a different lead time, a different false positive rate, and different structural assumptions that may or may not hold in changing economic environments.
The best approach is not to search for a single perfect indicator but to build a multi-indicator framework that assigns probability based on how many signals are flashing simultaneously. When the yield curve, credit spreads, lending standards, and the LEI all deteriorate together, the probability of recession is very high. When only one is signaling, the evidence is ambiguous.
→ Building your macro toolkit? Explore the Eco3min Financial Education Hub
What the data shows
Using FRED data across multiple indicators and NBER recession dating (1969–2024), here is a comparative scorecard.
Yield curve (2s10s): 8/8 recessions predicted since 1969. Lead time: 12–17 months median. False positives: 1 (1998, brief inversion without recession). Strengths: long lead time, consistent. Weaknesses: timing imprecision, recent 2022 inversion still debated.
HY credit spreads: 5/5 recessions detected since 1996 (data availability). Lead time: 3–9 months. False positives: 1 (2015–2016, energy sector stress without broad recession). Strengths: shorter lead, real-time market signal. Weaknesses: sector-specific stress can trigger without broad recession.
Sahm Rule: 9/9 recessions detected since 1970. Lead time: 0–4 months (detection, not prediction). False positives: 0. Strengths: simplicity, real-time, zero false positives. Weaknesses: detects, doesn’t predict — by the time it triggers, the recession has started.
Conference Board LEI: 8/8 recessions predicted since 1970. Lead time: 7–12 months. False positives: 2 (late 1960s, 2022–2023 when LEI declined for 22 consecutive months without clear recession). Strengths: comprehensive (10 components). Weaknesses: complex, structural changes may reduce reliability.
Bank lending standards (SLOOS): 4/4 recessions detected since 1990. Lead time: 6–12 months. False positives: 1 (2023 tightening without confirmed recession). Strengths: direct measure of credit availability. Weaknesses: quarterly data, smaller sample.
→ Datasets: Yield Curve History · Credit Spreads & Recession · Sahm Rule
Why it happens — the macro mechanism
Each indicator captures a different channel of recession transmission — which is why they have different lead times.
The yield curve reflects monetary policy expectations — the market’s assessment of whether current tightening will break the economy. It leads by 12–17 months because it signals the policy configuration that typically precedes recessions, not the recession itself.
Credit spreads reflect credit market stress — the actual cost and availability of corporate borrowing. They lead by 3–9 months because they respond to deteriorating fundamentals as they happen, not in anticipation.
The Sahm Rule reflects labor market deterioration — the moment when job losses become self-reinforcing. It triggers at or near recession onset because unemployment is a lagging indicator — it rises only after the economy has been weakening for months.
Lending standards reflect bank risk perception — a forward-looking measure of credit availability. They lead by 6–12 months because banks tighten before defaults rise, based on deteriorating loan quality indicators they observe in real time.
The composite framework uses these different lead times constructively: the yield curve provides the earliest warning (12+ months out), lending standards and credit spreads confirm building stress (6–9 months), and the Sahm Rule confirms arrival (0–4 months). When all four are aligned, the signal is about as strong as macroeconomic forecasting allows.
No single instrument tells you the weather. But a thermometer, barometer, and wind gauge together rarely get it wrong.
→ Framework: The Economic Cycle
What it means for different economic actors
Portfolio managers should build a recession probability dashboard using multiple indicators with different lead times. A simple approach: assign a score when each indicator crosses its critical threshold. When 3+ indicators are triggered simultaneously, defensive positioning is warranted. When only 1 is triggered, maintain awareness but don’t overreact.
Business owners can use the multi-indicator framework for operational planning. The yield curve provides the earliest heads-up for strategic decisions (capital spending, hiring plans). Lending standards signal when credit conditions will tighten for your business specifically. The Sahm Rule tells you the downturn has arrived and it’s time to execute the contingency plan.
Policymakers face the challenge that indicators disagree — especially at turning points. The 2023–2024 period saw the yield curve inverted, LEI declining, and lending standards tight — yet GDP remained positive and unemployment was low. This divergence may reflect structural changes (longer transmission lags, fiscal buffers) or may be a temporary reprieve before delayed effects materialize.
The fundamental error is relying on any single indicator — or dismissing indicators because they haven’t triggered a recession within the expected timeframe. Indicators measure probabilities and vulnerabilities, not certainties. A model that correctly identifies recession risk 8 out of 8 times will still appear “wrong” for months or years before the event materializes.
Go deeper
📊 Studies: Yield Curve History · HY Spreads as Leading Indicator
📁 Datasets: Credit Spreads & Recession · Sahm Rule
📖 Related: Why does the yield curve invert?
Related questions
Frequently asked questions
Why do indicators sometimes give false signals?
Because the economy is not a machine with fixed relationships. Structural changes (fiscal policy, demographics, financial innovation) alter the transmission mechanisms that indicators rely on. The 2023 LEI decline without recession may reflect the unusual pandemic recovery dynamics — massive fiscal support, excess savings, and labor market tightness that buffered the economy against tightening. Indicators are empirical regularities, not physical laws.
Can AI improve recession forecasting?
Machine learning models can process more variables simultaneously and identify non-linear patterns. Some research shows modest improvement over simple indicator-based models, particularly for nowcasting (identifying current conditions). However, the fundamental challenge — that recessions are rare events with structural variation between cycles — limits the training data available and makes overfitting a persistent risk.
What’s the simplest reliable approach?
Watch three things: the yield curve (2s10s), high-yield credit spreads, and the Sahm Rule. When all three are triggered, a recession is highly likely. When only the yield curve has inverted (earliest signal), maintain heightened awareness but don’t panic. This three-indicator framework correctly identified every recession since 1990 with minimal false positives.
Go further:
Last updated — 13 April 2026