Data Documentation & Code

Terms and conventions

All yield files begin with date and the three coverage fields, followed by one column per maturity. Daily-maturity files use _days; monthly-maturity files use _months.

Monthly-observation files use the last available published date of each calendar month; monthly-maturity files report maturities in months. For standalone KR in Germany, Japan, the United Kingdom, Canada and Switzerland, max_curve_ttm_days / _months coincides with max_observed_ttm_days / _months. KR-T for those markets may extend further, where an FX-transformed U.S. Treasury reaches beyond the longest domestic bond. For U.S. KR, max_curve_ttm_days / _months follows the publication horizon and therefore need not coincide with max_observed_ttm_days / _months.

Data sources and sample construction

Sources

U.S. Treasury data come from the CRSP Treasuries Time Series. Government bond data for Germany, Japan, the United Kingdom, Canada, and Switzerland, together with the corresponding FX spot and forward rates, come from Bloomberg using the composite BGN pricing source.

Bond filters

• Bonds within about three months of maturity excluded, on liquidity grounds
• Callable / sinkable / puttable / perpetual / convertible / floating-rate / fungible securities excluded
• Outlier quotes excluded, where the yield deviates sharply from that day's cross-sectional median
• U.S. on-the-run issues excluded, for their liquidity and specialness
• U.K. gilts excluded within the ex-dividend window before a coupon date, where accrued interest turns negative, and during partly-paid issuance periods
• Swiss Treasury bills excluded, as the source returns almost no quotes for them

Coverage and date inclusion

International KR panels exclude domestic Bloomberg holidays and dates with fewer than five domestic bonds. KR-T additionally requires U.S. reference data and FX spot and forward data. For date-level bond counts, observed maturity coverage, and the reason a date is included or excluded, see the bond coverage detail files below.

Field definitions

Field	Meaning
date	Observation date (YYYY-MM-DD)
n_bonds	Bonds entering that day's fitted cross-section, after the screens above
min_observed_ttm_days	Shortest remaining maturity among those bonds
max_observed_ttm_days	Longest remaining maturity among those bonds
kr_status	Whether the date is included in the KR panel. Values: published, holiday, few_bonds
krt_status	Whether the date is included in the KR-T panel. Values: published, holiday_domestic, few_bonds, no_us_data, no_fx_spot, no_fx_forward, fx_input_inadequate

• published: the date is in that panel.
• holiday (KR) / holiday_domestic (KR-T): Bloomberg holiday in the domestic market.
• few_bonds: fewer than five domestic bonds that day.
• no_us_data: U.S. reference data unavailable, a U.S. holiday or market closure.
• no_fx_spot: FX spot rate missing although forwards exist.
• no_fx_forward: no FX forward quote for that currency that day.
• fx_input_inadequate: the observed FX quotes do not support the short end (0.06% of KR-T dates). How these quotes are identified →

Updates

The database is updated quarterly. Each update extends the yield curve panels to the most recent quarter end, applies the same estimation parameters and data filters as the published paper, and leaves previously published values unchanged. Corrections, if any, are listed in the release history below.

Code

Papers and how to cite

The data are free to use, provided the corresponding paper is cited.

Stripping the Discount Curve - a Robust Machine Learning Approach

Damir Filipovic, Markus Pelger & Ye Ye · Management Science 72(9), 8057-8120, 2026

Introduces the kernel ridge regression approach to discount curve estimation. Methodology for the KR yield curves and the risk-free return.

DOI · SSRN · Copy citation · BibTeX

Shrinking the Term Structure

Damir Filipovic, Markus Pelger & Ye Ye · Review of Finance, advance article, 2026

Unifies curve estimation with cross-sectional factor modelling and identifies a fourth complexity factor. Methodology for the U.S. discount bond returns, term structure factors and bond market condition measures.

DOI · SSRN · Copy citation · BibTeX

Stripping Discount Curves Across Currencies: Transfer Learning from US Treasuries

Damir Filipovic, Markus Pelger & Rose Wang · Working paper, 2026

Extends KR to six government bond markets and introduces KR-T, which carries information from data-rich to data-poor markets through FX forwards. Methodology for the international KR and KR-T yield curves.

SSRN · Copy citation · BibTeX

Download all three citations as BibTeX (citations.bib) →