I am a bit confused on what is going on regarding the new benchmark rate SOFR. My understanding is that SOFR is to replace Libor. However, I also get information on Fed fund OIS discounting is replaced by SOFR discounting .
$\begingroup$
$\endgroup$
1
-
$\begingroup$ saw this in hot network. thank god there's no 'libor rate' here $\endgroup$– BCLCMay 11, 2021 at 13:11
Add a comment
|
2 Answers
$\begingroup$
$\endgroup$
4
The market is using SOFR discounting for all sorts of quotations already (not FF). For example, swaption vol is quoted with SOFR discounting, CME and LCH moved to SOFR PAI and discounting on Oct. 16 2020 on new AND legacy swaps.
For EUR cleared, major CCPs did this since July 27 2020.
The market switched to discounting with the relevant RFR rates on the dates above. Hence, if you have a dual stripped curve (e.g. 3m US libor), you use SOFR and no longer OIS (FF).
ISDA fallbacks will apply from 31 December 2021 for GBP, JPY, CHF and Euro-LIBOR and from 30 June 2023 for USD LIBOR. Note that the FED have issued supervisory guidance encouraging banks to “cease entering into new contracts that use USD LIBOR as a reference rate as soon as practicable and in any event by December 31, 2021”.
That said, it does NOT get rid of FF. It will co-exist, with SOFR being the more widely used (and being the choice for discounting).
-
3$\begingroup$ Would you mind incorporating the following? Commonly, there exists an agreement between two derivatives counterparties (party/ccp or party/party) on whether, and how, to exchange collateral for their bilateral portfolios and at which rate to reimburse the collateral. From first principles we know that in our case here, a curve bootstrapped from its collateral-rate-swaps must be used for proper discounting. Hence, if you have a bilateral CSA with FF (SOFR) as collateral rate, you should use FF (SOFR) for discounting. $\endgroup$ May 11, 2021 at 6:27
-
$\begingroup$ That's a good addition. If you rely on quotes for bootstrapping you will quickly realize you have no choice in some cases (Bloomberg for example only has Libor swap rates quoted with SOFR). So your 3m forward curve will always be using SOFR in dual curve stripping. You can still discount with FF and CSA curves derived from FF though. $\endgroup$– AKdemyMay 11, 2021 at 8:16
-
2$\begingroup$ While CCP's have switched to RFR discounting on SOFR/ESTR many brokers are still trading OTC options with CSA's referencing FF/EONIA, so while what you say is true regarding the CCP's it isn't true in all cases yet as each counterparty with the broker must agree to the change of CSA. This causes the need for an additional curve when calibrating interest rate models for example. For EONIA it is certain the CSA must change by year end but is it true for SOFR? Perhaps not? The Fed has been lazy and watching what Western Europe is doing first too much rather than killing off FF imo. $\endgroup$ May 11, 2021 at 8:49
-
$\begingroup$ agree. I think I got the full picture now. However, it is still not clear to me why SOFR is used to replace OIS as the PAI. Fundamentally, is it because SOFR is a better approximation than FF rate as a risk free rate? And the reason SOFR was not used previously was because it did not exist until a coupe of years ago ? $\endgroup$– PeacefulMay 11, 2021 at 8:51
$\begingroup$
$\endgroup$
SOFR was never meant to take USD LIBOR's role, as USD LIBOR reflects unsecured funding (and is credit sensitive).
An index like BSBY, on the other hand, can. BoA just started issuing FRNs linked to it. A BSBY-SOFR basis swap was also struck a month ago.
