0
$\begingroup$
  1. Various central banks publish their fitted nominal yield curve estimates: the Fed, BOE, SNB, BOC, ECB (cf: Bundesbank), RBA, Russia, RBI.
  • (I couldn't find for BOJ; Brazil; BOK; or PBOC. Links for these would be extremely welcome!)
  1. Bloomberg has its own estimates of these zero-coupon nominal curves (e.g. USGG10YR for the US ten-year).

These estimates are not the same (at least not always), and in particular they differ in their historical availability: Bloomberg has less data. Just to pick one example, the Russian 10y on Bloomberg (GTRUB10Y Govt) has data since 2010; the central bank's website estimates back to 2003.

Does any data source have better (i.e. longer) cross-country coverage for yield curves?

Thank you!

$\endgroup$
1
  • $\begingroup$ Jonathan Wright also has estimates for 10 countries here, but it's less than two decades of data $\endgroup$
    – bhalperin
    Commented Jan 27, 2022 at 14:22

1 Answer 1

7
$\begingroup$

First, a quick comment on Bloomberg symbols such as USGG10YR. These are actually yields on "generic bonds"; typically these are benchmark, on-the-run issues. Long story short, these are not zero coupon bond yields.

For fitted, constant maturity zero coupon rates or par yields, you've already compiled some of the major public sources. Here's a slightly augmented list for reference:

Free Sources

  • United States: The Gurkaynak, Sack & Wright dataset, which represents off-the-run Treasury yields fitted using the Nelson-Siegel & Svensson models, go back to 1961. The Fed's H.15 release provides yields fitted to on-the-run issues using cublic splines, which also has a very long history.
  • United Kingdom: Bank of England has fitted yields using a spline model going back to the 1970s.
  • Canada: Bank of Canada publishes their fitted yields computed using the Merrill Lynch Exponential Spline model, with histories going back to the 1980s.
  • Australia: This link and this one have RBA's fitted yields, also computed using the Merrill Lynch Exponential Spline Model.
  • EMU: The ECB publishes fitted yields computed using the Svensson model, with histories going back to the start of Eurozone.
  • Germany: Bundesbank's data is also computed using the Svensson model, with histories going back to the 1970s.
  • Belgium: Here is the estimated yields published by National Bank of Belgium. History goes back to the early 1990s.
  • Switzerland: SNB publishes zero coupon rates going back to 1988.
  • China: ChinaBond has extensive fitted Chinese government yield curves since 2002, computed using a cubic Hermite spline if I remember correctly.
  • Japan: Japan's Ministry of Finance has par yields for JGBs since 1974; these are calculated using a cubic spline as well.
  • Norway: Norges Bank has the latest bond yields, although I'm not sure whether these are constant maturity yields.
  • Russia: Bank of Russia publishes zero coupon rates, computed using the NS/Svensson-class models.
  • Thailand: ThaiBMA publishes constant maturity bond yield, with history going back to 1999.

For completeness, there are a few additional links for benchmark yields (but not constant maturity yields):

Commercial Sources

  • Augur Labs provides fitted zero coupon, par, and forward curves for 36 economies, with very long-term histories (typically 1970s for developed world and as far back as possible for EM economies). [Disclaimer: I'm affiliated with Augur Labs.]
  • J.P. Morgan Markets has global fitted yield curves. This is "commercial" in the sense that it's not in the public domain, but the data is free for JP Morgan clients. The length of history is a bit uneven though, depending on the country of interest.
  • Similar to Morgan Markets, Barclays Live provides many fitted yield curves to clients.
  • ICE Index Platform also has pretty good global yield curves coverage (20+ countries and fairly decent history).
  • Other providers are the usual suspect, such as Refinitiv, YieldBook, etc. I know they have such data, but I haven't used them for ages and can't say much about data quality/length.
$\endgroup$
5
  • $\begingroup$ 1. This is amazing, thank you, I feel like I should actually be paying you; 2. ...well that is extremely good to know about Bloomberg! $\endgroup$
    – bhalperin
    Commented Jan 28, 2022 at 0:38
  • $\begingroup$ I was going to make a separate question asking about real yields, for those countries which issue inflation-linked bonds. The Fed and BOE (fortunately of the two largest relevant markets!) have their own series; but I haven't seen any good sources for other countries. I'm not sure if it would be appropriate for me to make a separate question on this, but at the very least perhaps I'll ask here @Helin: does Augur Labs cover real rates as well? Thanks! $\endgroup$
    – bhalperin
    Commented Jan 28, 2022 at 0:43
  • 1
    $\begingroup$ @bhalperin We do US, UK, and euro area. Other countries, as you pointed out in the other post, don't have enough bonds. Using inflation swaps is probably a better option. We have an internal project to create synthetic linkers and build curves based on them, but this is not a top priority to be honest and likely won't be available for months. $\endgroup$
    – Helin
    Commented Jan 28, 2022 at 2:56
  • $\begingroup$ Thanks! Out of curiosity and if you're willing to share, would you be building the synthetics using inflation swaps + the associated nominal bonds? $\endgroup$
    – bhalperin
    Commented Jan 28, 2022 at 13:28
  • 1
    $\begingroup$ @bhalperin To be honest I haven't decided. We have models that generate the full term structure with just a few bonds. We may do that first and compare the result against inflation swaps for sensibility. Or we may just use inflation swap as the "truth." $\endgroup$
    – Helin
    Commented Jan 28, 2022 at 21:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.