The main dashboard tells you what regime we’re in. This page tells you how close we are to a turning point. Five leading signals: recession probability, yield curve, financial conditions, credit spreads, and labor momentum. Each pairs a chart, a percentile vs. history, and a clear threshold for what would change the call.
The takeaway from all four signals below, distilled. Detail sections follow.
Twelve-month-ahead recession odds from a term-spread probit (the NY Fed's yield-curve model, recomputed here from the 10Y–3M spread), with the Chauvet-Piger coincident probability for context. The Sahm Rule triggers once the 3-month average jobless rate rises 0.5pp above its low of the prior 12 months. Both have a recent false alarm: the curve model read above 50% through 2023 and Sahm triggered in July 2024, and no recession followed.
The 10Y–3M spread is the NY Fed’s preferred recession model input; 10Y–2Y is the financial-press favorite. Every US recession since 1960 was preceded by a sustained 10Y–3M inversion (12–18 month lead). The dangerous signal isn’t the inversion itself; it’s the bull-steepener that follows (recession starts as the curve un-inverts).
Chicago Fed composites of ~105 risk, credit, and leverage measures. NFCI is the raw reading (z-scored by construction); ANFCI strips out the part explained by current macro. The actionable signal is divergence: when ANFCI tightens while NFCI is flat, conditions are tightening for reasons not justified by the economy. That’s the early warning.
A market-priced gauge of default risk. FRED now carries only about three years of the ICE OAS series, so the chart adds the Moody's Baa-10Y spread for long history. HY thresholds: <300bp = complacent, 400–600bp = normal, >800bp = stress, >1000bp = crisis. When HY widens while IG stays calm, stress is idiosyncratic (single-name). When both widen together, stress is systemic.