Yield Curve Data

Yield curve data is the set of interest rates across Treasury maturities for a given day, almost always laid out with one column per maturity tenor. Two primary sources supply it.

The Federal Reserve's FEDS staff curve ships as a daily CSV - 17,013 rows by 100 columns, about 17 MB, running from June 14, 1961 to the present - with maturities in 1-year steps from 1 to 30 years, and Svensson parameters alongside zero-coupon, par and forward rates. Treasury's Resource Center data center takes a different shape: one row per business day with one column per maturity tenor, archiving par yields back to 1990, bill rates to 2001 and real yields to 1997. FiscalData's Treasury datasets API adds a monthly Average Interest Rates series for lower-frequency work.

Two official curves that disagree on shape will wreck a merge.

• Tenor grids differ. FEDS steps in 1-year increments from 1 to 30 years; Treasury posts one column per its tenor set. Joining them without an explicit mapping interpolates silently.

• Rate type is a modeling decision. Zero-coupon, par and forward rates answer different questions; the FEDS file carries all three plus Svensson parameters, so pulling the wrong column biases duration and hedging math.

• Archive depth bounds backtests. Par yields from 1990, bill rates from 2001 and real yields from 1997 on the Treasury side - versus 1961 from FEDS - determine how far a strategy can be tested.

1. Map tenor grids before joining. Reconcile FEDS 1-to-30-year yearly steps against Treasury's per-tenor columns explicitly.

1. Name the rate type you need. Zero-coupon, par or forward - the FEDS CSV carries all three, so pin the columns in writing.

1. Check archive depth against your backtest. June 14, 1961 for FEDS versus the 1990, 2001 and 1997 starts for Treasury par, bill and real yields.

1. Plan for file weight. The FEDS daily CSV is 17,013 rows by 100 columns, roughly 17 MB - trivial once, meaningful in automated refresh loops.

1. Add monthly frequency where needed. FiscalData's Average Interest Rates series serves monthly reporting without resampling daily curves.

Adjacent entries cover the delivery mode and the institutional-data families these curves belong to:

Bulk download is the delivery mode both primary yield sources use. central bank statistics belongs to the same data family as this entry. government fiscal statistics extends this entry's coverage. Browse curated sources in diversified-financial-services data.