Ratings Agencies Incentivized By Incentives

Publish date:
Updated on

Back when "Meredith Whitney says all your munis are belong to default" looked like it could be more like her bank calls than, say, Harold Camping's apocalypseseses, muni issuers and bankers liked to point to the fact that municipal bonds actually defaulted much, much less often than corporate bonds at the same ratings category.

Which is kind of weird since ratings categories are meant to predict default. Moody's explanation was "All of the revenue-producing power of a municipality can be brought to bear to service the debt," but, y'know, if you knew about all that revenue producing power, maybe you should have rated the thing higher? If, as the agencies claim, ratings are meant to be comparable across different types of credits, that's not a very satisfying explanation.

Bloomberg today discusses a study with a plausibly better explanation: munis get worse ratings for the same expected default because they pay ratings agencies less.

“Credit-rating inflation is correlated with the asset classes that generate the most revenue for the raters,” said Jess Cornaggia, who has presented research at the SEC and the Commodity Futures Trading Commission. ...

Bond ratings for municipalities, which typically pay the least, are the most accurate, a measure within the study showed. Those bonds, whose issuers usually have the power to tax constituents, have a 0.44 accuracy ratio, where a higher number means a stronger relation between lower credit ratings and higher defaults. The ratio was 0.4 for company securities, 0.36 for sovereign debt, 0.3 for bank and financial companies and 0.16 for securities backed by assets such as credit cards and mortgages.

Oops! Unsurprisingly the worst performers were bank debt and structured credit, where volumes are highest, everything gets ratings (usually multiple ratings), and issuers are pretty fee-insensitive. (The study authors write "We lack sufficient data to document variation in profit margin by asset class, which would be more compelling than correlation with revenues (i.e., deal volume)," but we can cut them some slack on this one: you'd expect ratings margins to be generally high (it's just people looking at models), so volumes are more important, and you'd also expect banks to be more willing to pay to get deals done than your typical rural county.)

While the paper is pretty embarrassing for the agencies, it's hard to believe that anyone was sitting around saying "let's screw municipalities because they don't pay us much," and the authors don't really claim that they were. Instead, the real explanation probably comes from the fact that different credit types are really in some sense incommensurable. ABS issues are rated on math, using cash-flow models that are highly sensitive to correlation assumptions rather than detailed work on underlying assets. Corporates are rated on old-fashioned corporate finance. Municipal and sovereign creditworthiness is as much a matter of electoral psychology as anything else: if things get bad, or even if they don't (ahem debt ceiling), will the issuer make difficult political decisions to repay bondholders? Incentives matter, but they interact in complex ways.

The paper authors get this:

One explanation for the inflated ratings of structured products is that structured ratings are "noisy" due to the opacity of the underlying assets. The intuition is appealing: synthetic CDOs are opaque relative to corporations with audited financial statements; this opacity results in greater dispersion among credit assessment by competing CRAs; this dispersion results in a greater opportunity for ratings shopping by the issuers of the CDOs....

Issuer opacity is a more compelling explanation for ratings inflation among synthetic CDOs backed by credit default swaps than for traditional RMBS or ABS. Indeed, one could argue that a pool of mortgages or credit card receivables should be less opaque than corporate issuers with synthetic leases and other exotic off-balance-sheet liabilities. However, we find even these more transparent structured products exhibit significant ratings inflation relative to corporate bonds.

Moreover, to the extent that issuer opacity is a compelling explanation for ratings inflation, it should also apply to municipalities and sovereign issuers. Dispersion in the qualitative credit assessment of the government of Indonesia should, like complex structured products, be greater than the dispersion in [agency] assessment of corporations‘ audited financial statements.

That last part can't really be right: Indonesia is not a portfolio of Indonesian companies; it's an entity that can tax its citizens. But the basic point is a good one. It's too easy to say "well, the ratings agencies screwed up correlation models and got MBS ratings almost universally wrong, but that's understandable, math is hard." But structured credit shouldn't really be harder than corporate credit, or at least, it shouldn't be harder than bank credit - it should be easier, because it has fewer humans involved.

Your average mortgage back security is a mess of underlying cash flows and uncertain correlations. So is your average MF Global. But the MBS is a predictable structure full of unpredictable stuff. MF Global is a structure of unpredictable stuff controlled by an unpredictable guy who can just decide to buy a lot of European sovereign exposure. It's not surprising that its ratings are volatile; it's surprising that its' peers ratings are mostly so much less volatile than those of financial issuers that are not subject to the whims of human decisionmaking.

Credit Rating Companies Favoring Firms Paying Most Over Nations [Bloomberg]

Credit Ratings Across Asset Classes: A ≡ A? [SSRN]