Will OIS be the successor to LIBOR? Lets take a look at the similarities and the differences.
An OIS exchanges a fixed rate for a floating overnight Fed Funds rate geometric average. The rate reflects expectations for Fed Funds over the term of the trade. As a cleared swap, there isn’t an exchange of notional, only IM and VM and a final cash settlement based on the differences between the rates.
LIBOR also reflects interest rate expectations over the term of the trade. Except there is an initial exchange of notional – the cash is actually lent – and there is no mark-to-market on the transaction during the loan. There is counterparty risk in an inter-bank loan that is mitigated on an OIS trade (by running the trade through a CCP).
From the point of view of a financing, a bank trader looking to raise money might borrow cash via a LIBOR-based loan OR put together a combination of an OIS trade for the same term (paying fixed, receiving floating Fed Funds effective) plus borrowing cash in the overnight Fed Funds market. On the OIS, the two Fed Funds legs cancel each other out (more or less), leaving a fixed rate borrowing.
But if the cash flows look the same (on a net basis) and each reflects a fundamental interest rate expectation component, why have the two not always yielded the same results?
If a trader borrows at 6 month LIBOR, come hell or high water they have the cash. Their obligation is simple: to return the money at the end of the trade plus interest. If a trader enters into an OIS trade, they have a different risk. Should the Fed Funds market be disrupted – either because of some sort of systemic problem or idiosyncratic issue with the borrowing bank, they are short cash. In the financial crisis when LIBOR/OIS spiked, it reflected the price banks had to pay to lock in the cash versus rolling the dice each day on Fed Funds. Regulators have often criticized banks for running funding mismatches – borrowing short, lending long – but isn’t that what is happening with a OIS + Fund Funds borrowing?
On earlier occasions we have gone into the other issues with OIS – for technical reasons Fed Funds is an impaired rated due to IOER, etc. etc. That still holds true. But the fundamental differences between OIS and LIBOR show that the risks don’t line up. One is not simply a synthetic version of the other. It is hard to advocate for the market to veer toward OIS as a substitute for LIBOR.
OIS levels can and do inform us about short-term rates in general and LIBOR levels in particular. And OIS levels and volume are more transparent than LIBOR (at least until the LIBOR trade repository is done). But it is still apples to oranges.
2 Comments. Leave new
The difference between OIS and LIBOR is less about fed funds disruption risk and more about bank credit in my view.
To explain: OIS is an overnight loan returned the following day. LIBOR is a term loan returned at the end of the term. Though both encompass counterparty risk, the OIS counterparty risk is limited to the possibility that the bank on the other side goes down the following day only whereas LIBOR encompasses bank credit risk through to term which creates a credit risk premium on the rate (hence the concerns about banks being on the high side of the LIBOR average). OIS in effect strips out nearly all the bank risk (apart from overnight default risk) and focuses much more nearly on the pure cost of funding in the central money market.
Suitability as LIBOR replacement:
Term rates? – there is a term OIS market e.g. a 3 month OIS rate is the rate for overnight loan lent out in 3 months time returned the next day.
Reliability? – because of the 2008 LIBOR problems (TED spread went to about 300 bps at the peak of the crisis), banks are already moving towards making OIS the funding assumption in their valuations (e.g. even a LIBOR trade would be discounted in NPV calculations using an OIS curve).
Actual transactions? – OIS trading (where the swap floating rate is OIS instead of LIBOR) also exists in major currencies but reamins a minority of IRS with LIBOR still the majority. Using SDRs , TRs, it ought to be possible to create an OIS transaction weighted average.
All good points. But at the end of the day, wasn’t LIBOR supposed to have counterparty risk embedded in it? It should not have come as a surprise given the exchange of notional.
I don’t think OIS can simply be dropped in to replace LIBOR in existing contracts (and by that I mean term OIS rates). Over time OIS may replace LIBOR as an index of choice — although repo indices have a shot too — but it will take a decade or two given the existing LIBOR-linked base of swaps. If you remove the counterparty risk component by moving to an instrument that is centrally cleared, you’ll end up with something closer to pure interest rate risk. But that is a different product and not a pure synthesis of LIBOR based swaps.