0
$\begingroup$

I was trying to replicate the paper by Gürkaynak

and wondering how I could filter out the "on-the-run" and "first-off-the-run" from CRSP dataset.

$\endgroup$

2 Answers 2

3
$\begingroup$

"On the run" USTs are the most recently issued UST. The previous on-the-run becomes the "first-off-the-run" or the "old".

T-bills are auction weekly (4,8,13,26,17,52 week maturities). 4,8,17 week bills are issued on Tuesdays following the auction week; 13,26,52 week maturities are issued on Thursdays of the auction week. The "on-the-run" is the most recently auctioned T-bill. The "first-off-the-run"s are therefore 1 week shorter in maturity.

The 2,3,5,7 year Notes are auctioned monthly. 2,5,7 are issued on the last calendar day of the month. The 3Yr is issued on the 15th. The "on-the-run" is the most recently auctioned. The "first-off-the-run"s are therefore 1 month shorter in maturity.

The 10Yr Note, 20Yr and 30Yr Bonds are auctioned quarterly (Feb, May, Aug, Nov). The 20Yr is issued on the last calendar day of the auction month. The 10Yr and the 30Yr are issued on the 15th of the auction month. The "on-the-run" is the most recently auctioned Note/Bond. The "first-off-the-run"s are therefore 1 quarter shorter in maturity.

$\endgroup$
1
  • $\begingroup$ All correct, and as another wrinkle, if they choose not to offer some maturity at some auction, then the difference between the on the run and the old may be longer. E.g. there were no 30-year issued between February 2002 and February 2006. $\endgroup$ Commented Apr 14 at 11:12
1
$\begingroup$

Typically, we'd classify a new bond as "on-the-run" as soon as it has been auctioned (i.e., as of the auction date). On that day, the old "on-the-run" becomes "first-off-the-run." Alternatively, you could also introduce new bonds into the estimation set after the "issue date."

$\endgroup$
3
  • $\begingroup$ Since on-the-run bonds carry significant premium in it that could affect the estimation of yield curve, so I want to filter it out to compare the results $\endgroup$
    – D H
    Commented Apr 14 at 11:25
  • 1
    $\begingroup$ @DH Yes that's generally a good idea, although dependent on market conditions. In recent years, this kind of exclusion is not that useful. In earlier periods, you may want to remove even the 2nd off-the-run. $\endgroup$
    – Helin
    Commented Apr 14 at 17:27
  • $\begingroup$ Thanks for your input! $\endgroup$
    – D H
    Commented Apr 14 at 18:38

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.