What are the alternatives to LIBOR?
Sensu stricto: none! LIBOR is a measure of the interbank unsecured lending rates between London banks in different currencies. The reason why the LIBOR benchmarks are replaced, beyond the past history of manipulation, is that the interbank unsecured lending has reduced dramatically. The reduction has reached such an extend that some of the indices do not really make sense anymore. The indices measure daily the average rate experienced by multiple banks for borrowing from their peers at a single point of time (11:00 am). In some cases, the mechanism that is measured take place only once a week or less. If no tree falls in the forest and everybody is listening intently, how do you best describe the average sound? There is almost no way to create a new interbank index due to the lack of underlying market. To the best of my knowledge, nobody is actually trying to propose an alternative interbank term benchmark.
In some sense you could argue that using transaction based fixing would be easier to manipulate than the current mechanism. If there are less transactions than rate setter, it would be easier to agree a (small) transaction at a suitable rate with one person which will be the base of the fixing than to collude with enough rate setters to move the market significantly.
What are the proposed indices?
Different new indices have been proposed recently. All of them are of the overnight type. The indices that have been discussed recently are: SARON (CHF), BFTR (USD), SONIA (GBP). The SARON index is a repo based index that has been indicated to replace the CHF TOIS. The relevant committee has indicated that the replacement will be finalized by year end. As expressed earlier, I have my doubts that the replacement can be fully effective in a so short period of time. For SONIA, this is not a new index, but an adaptation of an existing index. For the USD, the BRTR is based on repo rates. The index will start to be published in 2018. The rate has been selected by the Alternative Reference Rate Committee (ARRC).
All those indices are of the overnight type, this means that they can not replace directly LIBOR which is used mainly in its term deposit versions, with the most used terms being 1M, 3M and 6M.
I have read in a lot of places the suggestion to replace LIBOR by “OIS”. I have to confess that I don’t understand what that actually means. LIBOR is a term deposit benchmark while OIS is a derivative. Does that mean that OIS is understood as its actual meaning of a swap and instead of using a 3M deposit benchmark to fix the rate on a quarterly basis, a 3M OIS benchmark would be used instead? In that case, the ON benchmark is not enough. You need an extra layer of OIS benchmark, which itself refers to the overnight benchmark, with the similar questions on how to derive such a benchmark. The historical IRS benchmarks, like ISDA FIX, used in CMS-like products, have also been accused of manipulation over the recent years. Is the propose replacement by OIS a misnomer and should be read as replacement by overnight benchmark? In that case the benchmark is readily available, but how do you apply a daily benchmark to a quarterly fixing? Do you change the fixing frequency to daily, with all the back-office, risk management and valuation impacts? Do you keep a term fixing with the overnight rate, with the risk management and convexity adjustment impacts?
I will discuss more of these issues in the next episode, “the great pretender”, where I will describe what I thing should be the required features of replacement benchmarks.
ICE appear to be willing to continue with the LIBOR benchmarks, even without the support of the FCA. So there may be a LIBOR after 2021. If this is the case, there will be no forced transition by that date. It will leave more flexibility for the user to do the transition or not and if yes, when to do it.
Where is the throne?
I titled this episode “nobody is running for the throne”, but I have never said what the throne is! We know who is on the throne today: LIBOR. We know who LIBOR is, an interbank rate. But it is not because today we have one unique ruler of a “united kingdom” of interest rate, that the best strategy is a unique ruler or that the ruler has to come from the same familly. There is the interbank market, there is the government market, there is the secured/collateralised/repo market, there is the corporate market with all its sectors, there is the retail mortgage market, and so on. All of them deserve a specific benchmark. This may lead to a fragmented (derivative) market, with plenty of basis, but it is how the reality is. One of the roles of the banking system is to offer risk management services to the users. In last 30 years, the final users have borne the basis risk between the interbank market and their actual risk; maybe it should be reverted to its natural order and the banking system should offer that risk management service to the market in general. This would also mean that banks should be allowed to keep (some of) that risk in their books, run open positions, and be remunerated for it. You cannot do that at large scale without reviewing all the regulations about market risk and capital. Are the current regulation viable if the new benchmarks are introduced and the banking system is in the business of managing (and taking) market and credit risk? A reform of the benchmarks may require other reforms? Can it be that with new benchmarks in place almost everything is becoming a non-modellable risk?
The throne and the court
The pretender to the throne is not the only thing that need to be sorted out. The full court will need to be reviewed. This consists in a long list of swaps, futures, options, collateral remuneration, etc. that rely on the benchmarks. The new products will need to be designed. The design has to take into account the business, risk management and quantitative finance aspects of those products. Luckily, the well design products to match business and risk management requirements have often a simpler quantitative finance impact. The design of those products is worth a forthcoming episode on its own. A nice name for the episode will be “The court”.
In the waiting of its publication and for implementation details don’t hesitate to contact me for advisory work around those products. Some of my recent designs include options on futures bundles (presented 18 months before CME launch), rate based deliverable swap futures (not launched yet) and cap/floor on compounded rates.