The blog Sober Look published a post yesterday called “LIBOR converging with CD rates; CD transparency proves useful” which posited that the LIBOR curve will be based on CD rates. We typically enjoy Sober Look’s take on the world but we disagree with some important points in this article.
Earlier in the month, Sober Look looked at LIBOR and cited some research from JP Morgan which said “JPMorgan: – We expect that requiring a formal process for Libor submissions, together with the explicit requirement to reference a broad set of short term markets, is likely to produce convergence towards observable rates such as bank CD rates.“
In the most recent Sober Look post, they said, “It looks increasingly likely that going forward the dollar LIBOR curve will be set based on term deposit rates (discussed here). There simply is not enough term (longer than one week) interbank unsecured lending to determine LIBOR without the risk of potential manipulation (or perception of manipulation).” Yes, that is true — there isn’t a lot of inter-bank deposit trading going on, so its hard to point to actual trades to benchmark LIBOR off of. And using actual rates to determine LIBOR is the holy grail of benchmarking…so why not use CD rates? Sober Look had a nice graphic illustrating their point about convergence.
The Wheatley Review of LIBOR understood this too. They wrote about CDs (and CP) in their final report:
“…Certificates of Deposit (CDs) and Commercial Paper (CPs) are another way in which institutions can raise short-term unsecured funding. CDs are typically unsecured promissory notes issued by banks to investors in return for depositing funds with the bank for a specified length of time. They may pay fixed or floating rates of interest. As with inter-bank deposits, once the money has been deposited for a period of time, the depositor cannot withdraw the funds without incurring a penalty…”
But in the Wheatley August initial discussion paper, they also wrote:
“…There are low volumes in both the primary and secondary markets for CDs and CP; these markets have been adversely affected by concerns for counterparty credit risk since 2007-08. The low volumes also affect the ability for the rates to be representative of bank funding…”
In the CFTC order which fined Barclays, (that Wheatley praised for their methodology) the CFTC outlined how Barclays would, going forward, determine LIBOR. They ordered that the first factors that Barclays should look at are:
“…Factor 1 – Barclays’ Borrowing or Lending Transactions Observed by Barclays’ Submitters:
a. Barclays’ transactions in the market as defined by the Benchmark Publisher for the particular Benchmark Interest Rate;
b. Barclays’ transactions in other markets for unsecured funds, including, but not limited to, certificates of deposit and issuances of commercial paper; and
c. Barclays’ transactions in various related markets, including, but not limited to, Overnight Index Swaps, foreign currency forwards, repurchase agreements, futures, and Fed Funds…”
Sober Look thought that since CD rates are posted by all banks and tracking on sites like Bankrate.com so easy, that it makes sense to use CD rates as the raw material to benchmark LIBOR. Wait a second…
1.) The CDs that the Wheatley Report was referring to were negotiable CDs, typically issued in million-dollar notionals. This is not what your local Citi branch is posting and bankrate.com is tracking. We doubt this is what JP Morgan’s research was referring to. From the final Wheatley report “…However, a key difference between CDs and inter-bank deposits is that they can be negotiable, which means that they can be sold on in the secondary market before the term of the loan is up…” Apples to oranges.
2.) The “O” in LIBOR stands for “Offered”. From the BBA, the exact definition is “The rate at which an individual contributor panel bank could borrow funds, were it to do so by asking for and then accepting interbank offers in reasonable market size, just prior to 11.00am London time.” Put another way, it is where a panel bank thinks they will be offered money if they went out looking for the cash. While the CD and LIBOR rates look to be on top of each other, they represent two different things. With all due respect to used car dealers, should you expect to get the same price when you are buying a car as when you are selling one?
3.) The “L” in LIBOR stands for “London” and represent trading in offshore markets. No FDIC insurance there. The CD rates in Bankrate.com are US domestic levels. While there may be an equivalent site in UK that track local bank CD rates, we doubt it would very helpful in figuring out US$ LIBOR rates.
4.) Just because a bank is posting a CD rate, it is unclear how much is actually going through. Are there any transparent volume levels available?
So while CD rates are an critical part of the money markets and certainly should be used to inform LIBOR rates, let’s makes sure we use the right CD rates. The debate about how LIBOR should be determined is important and ongoing.
A link to the initial Wheatley discussion paper (August, 2012) is here.
A link to the final Wheatley paper (September 2012) is here.
A link to the CFTC order on Barclays is here.
A link to the British Bankers Association definitions is here.