Tuesday, February 26, 2008

Why Yield Spreads Aren't the Same as Credit Risk

(Felix Salmon) John Carney is one of those people who seems to think that if you're contrarian you must be right. His latest broadside is directed against Jesse Eisinger, of all people; in it he tries to assert that there's really nothing wrong either with the municipal bond market or with the way that the ratings agencies rate it. But he makes an enormous mistake: he assumes that the yield of any bond is basically just the risk-free rate plus a premium for credit risk. And because that simply isn't the case, his entire argument falls apart.

For instance, Carney claims that muni bonds' low yields are a function of their low credit risk:

Markets are very much aware that muni bonds rarely default, regardless of the rating. This is why muni bond investors accept lower yields--they know they are getting less risk. Similarly rated corporate and muni bonds are not similarly priced--the munis have lower yields that reflect their lower risk. The only way to conclude that the muni bonds are rated too low is to ignore pricing differences.

The whole reason why muni bonds have lower yields than identically-rated corporate bonds is very simple: they're tax-exempt. End of story. Their lower yields reflect not their lower credit risk, but rather the fact that bondholders pay no tax on their coupon income.

But Carney doesn't even mention that particular elephant in the room, and blunders on:

Imagine if this were not the case. If munis were consistently rated too low and the market somehow overlooked this error, there would be huge opportunities for risk-free yield in the market. We would expect arbitrageurs to very quickly make these opportunities vanish.

Which reminds me of the old joke about how many Chicago-school economists it takes to change a lightbulb: none, since if the bulb needed changing the market would have done it already. Of course Carney doesn't tell us what these "opportunities for risk-free yield" might be. Presumably they involve in some manner going long municipal bonds while going short identically-rated corporate debt: how one can do that in a risk-free way I have no idea. Especially given the fact that the trade has negative carry. Selling high-yielding debt and buying low-yielding debt? That's certainly not my idea of "risk-free yield".

Carney even has a theory for why munis are rated more stringently than corporates:

Rating munis according to the same criteria as corporate bonds would reduce the amount of information available to the market by obscuring differences in the risks of different municipalities. If two-thirds of munis were rated triple-A, investors would lack guidance about real differences between the issuers.

Real differences? Differences in what, exactly? Not default rates, that's for sure. And why should muni investors be the only buy-siders able to access ratings-agency information about subtle ratings gradations within the world of triple-A? Given the enormous quantity of triple-A structured products out there, don't you think that this service would be even more useful with them?

The truth is that different types of debt are rated on different scales, and the market is very well aware of this. Munis are rated on a scale that compares them to other munis, not corporate bonds. Corporate bonds, in turn, were obviously rated on a different scale than CDOs, and the market reflected its awareness of this by requiring higher yields for highly rated CDOs than it did for highly rated corporate bonds.

If this were really true, the agencies wouldn't use the same letters for different ratings scales, and buy-side institutions wouldn't have rules about investing only in triple-A securities, and Basel II wouldn't have rules allowing banks to zero-risk-weight their triple-A assets.

Carney seems to think that if you see a difference in yields between different fixed-income asset classes, that must mean the market perceives a difference in credit risk. But that's not true at all. In fact, outside the world of relatively liquid corporate bonds, credit risk is very rarely something accurately priced into fixed-income instruments.

I'm not just talking about liquidity, either, although that is crucial: a liquid Treasury bond trades at a lower yield than an illiquid structured product partly becaues it's so easy to get in and out of that Treasury bond whenever you like, without affecting the price. I'm also talking about credit analysis.

Why are there so many triple-A-rated securities out there? From CDOs to SIVs, from wrapped munis to securitized credit-card receivables - whenever the fixed-income markets come up with some bright new idea, it nearly always carries a triple-A credit rating. And there's a good reason for that: most buy-side fixed-income investors don't want credit risk. They have a tough enough time as it is worrying about the future of the yield curve; their risk-averse investors are certainly not the kind of people who are going to be at all happy in the event of a default. If they're long-term buy-and-hold investors, they'll accept higher yields on more illiquid instruments. But the main reason why triple-A debt is so popular is very simple: there is a huge number of investors out there who have neither the ability nor the inclination to do detailed credit analysis on every bond they buy.

How much would it cost to hire someone to look in forensic detail at the bonds issued by some power plant in Wisconsin, and get that person to work out just how risky they were? If you're only buying a couple of million dollars' worth of the bonds in the first place, it simply can't be worth it: you're much better off outsourcing that analysis to the monolines. Let them take the credit risk, it's much easier, and it's much cheaper, and the cost of the labor is embedded in the bond price and spread out over all of the buyers, rather than having to come out of fund-management fees.

Of course, the people who relied on the monolines to take the credit risk also relied on the ratings agencies to give the monolines the crucial triple-A seal of approval. In hindsight, that wasn't so clever.

But never mind all that, just count heads. Ask yourself how many different stocks there are traded in the United States, and how many people there are - equity analysts, managers at hedge funds and mutual funds, newsletter writers, internet pundits, stay-at-home professional investors, amateur punters, etc etc - who are all competing with each other to value those equities. And then look at price volatility in the stock market.

Now, turn to bonds. The number of different fixed-income issues is vastly greater than the number of stocks, especially when you add in all the different tranches, and levels of seniority, and municipal issuers, and structured products, and so on and so forth. Yet the number of people actually doing credit analysis on all those issues is a tiny fraction of the number of people doing equity analysis. If all bonds needed to be acutely analyzed by the market on an ongoing basis, the demand for credit analysis would vastly oustrip its supply.

That's where the ratings agencies come in: they're where the buy side outsources its credit analysis. And that's also the reason for the whole industry of AAAification, where the sell side attempts to render the whole business of credit analysis moot. It's not very pretty, but it gets the job done. Until, that is, the system breaks.

No comments: