As of April 13, 2012 the CME has started accepting corporate bonds as collateral for cleared OTC Interest Rate Swaps. We have been thinking about the CME criteria for accepting corporate bonds and, while they are on the right track, we wonder if it is workable.
The paper is held by tri-party clearing agents (JPM and BNYM), who manage mark to markets, margining, and concentration limits. Corporate collateral will be included under the CME’s IEF4 structure. IEF stands for Interest Earning Facility. The name may be more of an artifact than an accurate description of the account structure, but it has stuck. IEF4 accounts allow CFTC Reg 1.25 eligible collateral to be pledged for term into tri-party.
From the CME’s Fact Sheet on Acceptance of Corporate Bonds, the specifics are:
The CME will publish a list of CUSIPs of the eligible bonds at the start of every month.
In addition, according to the CME website (a link to the page is here), IEF4 accounts (which include corporate bonds), foreign sovereign debt, gold, and stock, on a combined basis, are limited to the lesser of 40% of the core requirement or $3 billion.
We wonder where the 20% haircut came from? Equities are haircut at 30%. Maybe the argument was that corporates should therefore be lower? There was likely to have been some back testing to figure out if the 20% would cover the price move for a broad range of bonds in a default scenario. Some additional detail please?
The concentration limits are a good idea: by issue, by issuer, and by industry. But they are tricky to manage and a sophisticated forward-looking collateral management tool will be necessary. To avoid getting caught out, clearers will need to have plenty of paper available to meet the diversification requirements. We question if it will be worth the work to many clearing members versus simply using that paper for generic tri-party repo funding.
Using only TRACE “eligible and disseminated paper” concerns us. We wonder if TRACE is being used as some sort of liquidity marker? But the fact that a bond is on TRACE isn’t enough. TRACE data has to be actually used to benchmark liquidity. On TRACE you can tease out how much of a specific bond has traded over a given time frame. Perhaps there should be some hurdle that needs to be met to be eligible? How about a limit on the amount held as collateral relative to how much paper has actually traded?
Corporate bonds have been known to suffer from long bouts of illiquidity. Unlike equities, which can see volume spike higher when volatility goes up, corporates go the other way. The markets go into hibernation and bid/ask spreads balloon. That is exactly the same moment when pricing on the collateral mark-to-market becomes squirrely. That creates a sort of wrong-way risk. And if the paper is being marked by the pricing service to a model, look out.
One benchmark that is missing is CDS spreads on the underlying paper. Could they be used as a prospective indicator of trouble brewing? Something to consider.
In general, we like the idea of broadening out eligible collateral. A complicated model to manage portfolio limitations can turn into a Rube Goldberg very quickly and, sure, some of our ideas might tend in that direction. For corporates, the risk is sudden, but prolonged, evaporation of liquidity. That has to be managed. CME has the fundamentals right, but there is more to do.
Finadium wrote about the use of corporates as collateral in Corporate Bonds and Equities as High Quality Assets for Collateral Management and Bank Balance Sheets, March 2012. A link to the synopsis is here.