U.S. Department of the Treasury, Daily Treasury Par Yield Curve Rates
One row per Treasury business day of the official U.S. Treasury Daily Par Yield Curve Rates (keyless year-scoped CSV exports, https://home.treasury.gov/resource-center/data-chart-center/interest-rates/TextView?type=daily_treasury_yield_curve; the publisher's tenor grid evolved 2019-2026 and is normalized into canonical columns). Each row: 14 par-yield tenors (1m-30y, percent; honest nulls where a tenor did not exist that day/year), term spreads in basis points (10y-2y, 10y-3m, 30y-10y, 5y-2y), inversion flags + depth and per-episode tracking of consecutive inverted days (start date, day number, episode max depth), curve slope (10y-3m) and curvature (2*5y-3m-10y), 1/5/20-day 10y momentum, and a documented look-ahead-free 0-100 curve_stress score (0.40*inversion-depth(10y-2y) + 0.35*depth(10y-3m) + 0.25*inversion-breadth, each causal-rolling min-maxed over the trailing 10y window, null until 1y of history) with t1-t4 calm-to-severe tiers, plus flat-curve and deep-inversion flags. Fail-loud gates: closed tenor-label set, yields in [0,30]%, core tenors never null, unique dates, >=200 rows per full year. Units: percent, basis points, pp changes; row hashes are deterministic. Coverage: United States, 2019-01-02 -> latest published curve day (~1,900 rows; publisher publishes business days only, ~2 business-day lag at quarter-end). Provenance: keyless Treasury CSV downloads, U.S. public domain - cite the U.S. Department of the Treasury. Primary key: curve_date. Cadence: refreshed daily. Nullability: young tenor series (y_45d, y_4m, y_20y) null before introduction; episode columns null on non-inverted days; momentum null on the panel's first rows; stress_score/stress_tier null until 252 observations. Sample use: filter inv_10y2y_flag=1 to list every 2s10s-inverted day with its episode; order by stress_score desc to find the most strained curve days of the window.
- treasury
- bonds
- yield-curve
- interest-rates