62.2/ 100
Stable
US Treasury yield curve
2016-07-29 9.9 yrs · 123 pts 2026-06-26
Top drivers
t10y2y_inverted_zt10y3m_inverted_z
⌁ mcp.call("adw-054") vADW-054-live-1.0 ← All products
Financial/Rates · signal Gold
Yield-Curve Inversion Signal
How deep is the current yield-curve inversion and is it deepening?
- Refresh
- monthly
- History
- 9.9 yrs
- Plan
- Gold
Use cases
What it unlocks
⌁
For an agent
A credit-allocation agent ingests ADW-054 weekly; when inversion_signal_score exceeds 60 (current: 62.2, 78th percentile, backtest range 19.9–73.8) with a falling trend — meaning inversion is deep but starting to normalize, historically the window most correlated with recession onset — the agent reduces allocation to cyclical corporate credit and increases cash equivalents, logging methodology_version v0.1 and source_lineage (FRED T10Y2Y, T10Y3M) for the audit trail. The IOM's composite_z field lets the agent distinguish a true inversion from a mere rate-level move.
📈
For the business
A bank's risk desk uses ADW-054 as a leading indicator for loan-loss provisioning reviews. At 62.2 (78th percentile of a 10-year history), both the 2-year and 3-month spreads are simultaneously inverted at above-average depth — a combination the NY Fed's free yield-curve model captures only at a one-month lag. The desk can front-run the quarterly CECL reserve cycle by flagging elevated recession probability now, rather than reacting after credit deterioration appears in charge-off data.
Forward outlook
Prediction
- Horizon
- —
- Recommended use
- Track yield-curve inversion depth as a recession-risk signal. High score = deep inversion historically associated with future downturns. Descriptive; no predictive IC claimed.
Methodology
How it's built
Negate both spreads → z_10y2y (0.5) + z_10y3m (0.5) vs trailing 130-obs window → composite z → 0-100 (50=neutral, >50=more inverted than hist avg)
FRED (T10Y2Y, T10Y3M)
Version ADW-054-live-1.0 · validated to beat a naive baseline · benchmark: NY Fed yield-curve recession model (free, lagged)