A Better Credit Rating Solution?

Christopher Cox, Chairman of the SEC was criticized in PrefBlog on February 8 for his apparent belief that what this world needs is more rules.

I was very happy to be alerted by Naked Capitalism to some remarks he has made that show a better appreciation of the problem:

So how would the SEC substitute — and diminish — the regulatory reliance on ratings?

Mr. Cox said “one means of substituting … would be to substitute the current definition of the rating currently provided by the rating itself.”

What that means is that the SEC is considering ways of setting criteria that gets away from the ratings but focuses instead on the underlying concept. For example, for some rules the SEC could require bonds to be liquid and then develop some measure by which to sort them other than a credit rating, says one person familiar with the matter.

I posted yesterday in Earth to Regulators: Keep Out! that there wasn’t much point in regulation. All the rules in the world won’t make anybody a better analyst … or, indeed, to avoid blow-ups of any kind in this uncertain world.

The problem the regulators face is that under Basel II there is a good chance that a pro-cyclical bias will be introduced to bank capital requirements:

Under Basel II, the capital requirements for the largest banks would be based on their current assessment of the probability of default of the borrower (ie rating) – Basel Committee of Banking Supervision (2003). There is a live policy debate over whether different rating approaches adopted by the banks would lead to different procyclical outcomes and if they did which approach banks would choose to adopt. We find that less forward-looking bank rating systems, conditioned on the point in the economic cycle, could lead to a substantial increase in capital requirements in recessions. Looking at the 1990–1992 recession, ratings based on a Merton-type model, which reflect the point in the cycle through the use of current liabilities, lead to a 40% to 50% increase in capital requirements. In contrast, Moody’s ratings which are more forward looking, lead to little increase in capital requirements.

The new Accord which will be introduced in 2006 could, however, have a profound effect on the dynamics of bank minimum capital and lending in recessions. In contrast to the current Accord where, for a given quantum of lending to a particular set of borrowers, the capital requirement is invariant over time, under the new Accord the capital requirements will depend on the current assessment of the probability of default (PD) of those borrowers. If borrowers are downgraded by a bank in a recession, then the capital requirements faced by the bank will rise. This would be in addition to the possible reduction in the bank’s capital because of write-offs and specific provisions.

It’s amusing that the authors (Eva Catarineu-Rabell, Patricia Jackson and Dimitrios P. Tsomocos) of the quoted paper prefer the Ratings Agency style of attempting to rate “through the cycle” as opposed to the more dynamic Merton-style approach, but that’s beside the point for now.

There is only on legitimate reason that Credit Ratings might be of importance to regulators of any kind: the effect on bank capital requirements. And it should be noted that the only reason they have any effect on bank capital requirements is because the concept was written into the Basel Accords. And, of course, they were written into the Basel Accords because of the superb track-record (remember that concept? track record?) of the Credit Rating Agencies.

Not perfect, by any means. There are big blow-ups and minor whoopsees, but it’s an uncertain world dominated by human frailty. Get used to it.

So, I suggest, if the regulators wish to improve their legitimate regulation via new and improved credit ratings, it is only right and proper that they do it themselves, rather than imposed ludicrous constraints and requirements on private businesses in a horrific central planning exercise. Cox is on the right track.

Naked Capitalism takes a very dim view of the remarks:

Require bonds to be liquid?. That is the most deranged thing I have heard in a very long time, and this presumably coms from someone within the US’s top securities regulator. It confirms what I have long suspected: that the SEC is preoccupied with the equity market and knows perilous little about debt.

A simple illustration: just about all corporate bonds are illiquid. That’s one reason credit default swaps are popular. Investors can use CDS to create synthetic corporate bond exposures, which unlike the underlying bonds, can be readily traded. But Cox would have us write off the corporate bond market.

Corporate bonds are illiquid relative to, say, stocks. Sure. I discussed this on November 19, in relation to Prof. Cecchetti‘s desire to have all financial instruments traded on an exchange:

Looking at the topography of the financial system, we see several immediate candidates for migration to exchange trading.  I will mention two:  (1) bonds and commercial paper, and other fixed income securities; and (2) interest-rate swaps and other derivatives that are traded in large volume. Bonds as we know them have been around since at least 16th century.  And the quantities outstanding are substantial – in the United States there something like 4000 distinct corporate bond issues with a market value of roughly $10 trillion.  I can see no reason that these “fixed-income” instruments are not traded on an exchange.

How can we encourage the movement of mature securities onto exchanges?  The answer is through a combination of information and regulation.  On the information side, it is important that less-sophisticated investors realise the importance of sticking with exchange-traded products.  The treasurer who manages the short-term cash balances for a small-town government should not be willing to purchase commercial paper, or any security, that is not exchange traded.

However … there are degrees of illiquidity! Assiduous Readers of PrefBlog will be very familiar with the idea that, while the average trading value (however calculated!) of a particular preferred share might be only $100,000, it is generally possible (for MOST issues, MOST of the time) to call a dealer and trade a block worth $10-million … maybe not right away, but, say, within a week.

So, subject to problems with measurement and control that must be addressed, I’m entirely comfortable with the SEC (or other regulators) coming up with some kind of way in which liquidity will be measured and applying some kind of concentration penalty on banks who hold a position that is too large to be regarded as liquid. This would be a little easier in the case of the States, which already has the NASD TRACE system in place.

There is also the potential for so-called liquidity guarantees to be put in place at time of underwriting. So-called? Well, the European Covered Bond market association called for suspension of the agreement on November 21 and there are current problems with liquidity on Auction Rate Municipals, as mentioned yesterday. I’ve mentioned my own problems in not being able to get a bid for less than a million of good quality corporate paper. So, while I would not put too much faith in the ability of the private sector to provide a bottomless pit of liquidity for bonds in general, liquidity could be enhanced … central banks, for instance, could enter into “liquidity provider of last resort” agreements and accept corporate and other bonds as collateral on a routine basis.

I would not advocate formal liquidity guarantees. It’s too much central bank intervention in the economy … in times of stress, the private sector just ain’t gonna want to take long term debt onto its books, especially not if it’s esoteric. But liquidity could be measured and a capital penalty applied to instruments whose liquidity in times of stress was feared to be sub-normal. 

I’m not sure that I agree with Naked Capitalism‘s point about CDSs. They are sometimes more liquid, sure – it’s a lot easier to go long a CDS than it is to short a corporate bond, especially in size. But as I see it, the main attraction of CDSs has been that they make it a lot easier to lever up a portfolio, compared with taking a cash position (long or short) and (financing or borrowing it). That’s related to liquidity, but is not exactly liquidity.

The guts of the problem, however, is step 2 of Mr. Cox’s idea: and then develop some measure by which to sort them other than a credit rating.

Market prices won’t do it. I’ve posted elsewhere about market implied ratings and how dubious I am that such a system will prove to be a better indicator of credit quality than what we already have. The Fed is dubious too!

And, as noted, structural models such as Merton’s (equity implied ratings were briefly mentioned on October 18) (a) have a lot of problems, and (b) are pro-cyclical.

I’ve also noted that any quantitative system performs badly at trend changes … and it is at precisely the time of trend changes that stresses on the financial system become pronounced.

So what’s the Fed to do? The only answer is … if they want to come up with some way of defining credit risk, they’ll have to set up their own in-house credit rating service. Good luck with that!

Leave a Reply

You must be logged in to post a comment.