In a speech a few days ago, the Chief Executive of the FCA confirmed that Libor appears to be "terminally ill." At least there are a few years left to prepare for the end.
If you agree with the fundamental principle that fixings need to be based on executed transactions, perhaps there was no other outcome. Changes in market structure and funding methods since the financial crisis have undermined the basis for calculating Libor.
Why do we have Libor and do we still need it?
Libor came into being to bring standardization to derivative contracts and over the years has been immensely successful. It also makes a convenient reference for lending from an asset & liability management (“ALM”) perspective, so a bank funding desk can look at risk on both sides of the balance sheet in a similar way.
Until 2007/2008, Libor was also a proxy for the risk-free rate. Given the turmoil that bank balance sheets have been through, it became more than apparent after the crisis that Libor conceptually was certainly not risk free. Libor dislocated from what is now accepted to be the true risk free rate – the collateralised overnight funding rate or overnight index swap (the “OIS” rate – e.g. Sonia for GBP, Eonia for EUR, etc).
As a result, from more or less 2009 onwards, the derivative market has gradually accepted that the correct valuation of a typical collateralized interest rate swap is based on cashflow projections using Libor, but to use the relevant OIS discounting curve to find the fair value.
Given the basis on which Libor is defined – effectively short-term interbank borrowing – it gives us with the spread between Libor and OIS as the credit spread of the banking system relative to the risk free rate. If you think of it that way, there should be no problem changing lending spreads to reference OIS rather than Libor. In fact, it’s more satisfying to have a funding spread relative to the risk free rate rather than referencing it to credit spread of the banking system. Having Libor as the reference rate for funding is pro-cyclical.
Changing the lending reference from Libor to OIS would optimally require a bank ALM desk to switch its baseline risk view of funding to OIS as well. The deposit base would still carry a term structure so the Libor-OIS spread problem becomes an implicit issue to be managed inside a bank.
Is it possible to create a replacement for Libor?
Thinking back to the origins of Libor helps define the problem. To replace Libor precisely we would need an index to be based on something that is:
- Based on market activity
An OIS rate such as Sonia meets the requirements, but what should be done with the legacy portfolios of derivatives that reference Libor? If Libor ceases to exist then how can these be valued when there is no index to set the projected cashflows? It will be hard to design a credible replacement.
The idea that has been floated to create a Libor replacement referencing of bank CDS I find to be a bit alarming. Fixing 3, 6, and 12 month lending spreads based on CDS curves would be flawed because these are not liquid contracts – they are accounting marks. CDS liquidity was dreadful during the EU sovereign debt crisis and would inevitably suffer in another banking crisis. Using this approach to build new derivative contracts would be unreliable and, at worst, hazardous.
Using alternative bank funding rates (e.g. deposits or funding transfer pricing “FTP” curves) would be fraught with the same difficulties as Libor and would still be subjective. FTP curves can be set to create incentives for different banking activities so cannot be relied on as a fair market price.
Another suggestion has been be to lock Libor-OIS spreads at a point in time and make a wholesale conversion of derivative contracts to an OIS reference. While this has the benefit of clearing the derivative problem away once and for all, it does have an impact on risk positions for both banks and investors so is not trivial. There would be winners and losers. It also requires other traded markets to change to OIS referencing such as the cross-currency swap market. This in turn, requires floating rate issuers to have migrated to OIS indices.
Perhaps another way out would be to look to the futures market for inspiration. Using quarter days instead of daily calculations would not undermine a reference index and some form of interpolation could take care of legacy Libor swaps with odd dates. Instead of referencing CDS, perhaps a sufficiently large, diverse and tradeable pool of short-dated Commercial Paper in circulation could be used as a benchmark.
In the meantime, market liquidity for longer-dated Libor based derivatives may well suffer in the years running up to its demise. I can imagine a series of protracted negotiations with major fixed income participants over how to migrate portfolios onto new indices and otherwise collapse risk positions. In the absence of an agreed solution, I imagine that a ‘Libor-like’ index will be constructed to keep things limping along until the situation is resolved.
The best solution is often the simplest. Arguably, the sooner that markets can migrate to using OIS references for lending, the easier it will be to move away from Libor. Perhaps we should accept that Libor has served its purpose and should be allowed to go gracefully.
-Chris McHugh is a lecturer at The London Institute of Banking & Finance focusing on risk, banking, markets and derivatives for undergraduate and postgraduate courses.