The three main firms that rate financial instruments, which failed so badly during the Housing Bubble at rating mortgage-backed securities have notoriously had a conflict of interest since the early 1970s. They used to be paid by buyers of financial instruments, who thus wanted honest evaluations, but Information Wants to Be Free! (In other words, the spread of the photocopier made it hard to make money out of being paid by buyers.)
So, for over 35 years, Moody's, D&B, and Fitch been paid by sellers who have incentives to demand softball evaluations. The rating firms' old school culture of honesty carried them along fairly well, but eventually it broke down under the lucrativeness and complexity of Housing Bubble mortgage-related instruments. Lots of street level petty fraud was going on between borrowers and mortgage brokers, but nobody -- the clients, the feds, the media, the GOP, the Democrats -- wanted to hear that "underserved" borrowers weren't going to be able to pay back their mortgages. So, instead of doing things like hiring private detectives to check for fraud, the ratings agencies just rubber stamped the crud concocted by the people paying their bills.
Here's a simple suggestion for fixing the ratings firms: barter. Only allow them to be compensated by the sellers of financial instruments in those exact financial instruments they rated. Then make the rating firms hold 100% of everything they rate for the lifetime of the asset. (They can borrow against those assets for the cash they need to carry on their businesses.)
Simple?
Posts
17 comments:
They don't need to be fixed. They need to be blown up. If you want to buy a bond, pay your own damn expert to rate it.
No. Because then you have insider trading, i.e. rate a bond high, pump it up, and make various deals to sell it.
The FT had a column suggesting using CDS. Use the relative rates of insuring against defaults to determine the "rating" of a bond. So that banks, and other institutions that are regulated to have low risk, must only hold bonds that have low Credit Default Swap rates.
This does not require massive regulation, or people not to work in their own interest, and relies on sharp-eyed folks in the short market to expose stuff that is junk.
Okay, then make them hold 100% of the instrument for its lifetime. No insider trading.
How about paying financial rating firms only in the securities they rate which they will be required to hold for the duration of a typical investment cycle -- five years or so. Also, buyer and seller should jointly pay for the rating.
That's missing the point. The ratings firms exist by fiat of the regulators. The way to fix them is to say that the regulators no longer believe them.
I was involved, in 2007-2008, in the preparation of Value at Risk reports for an investment bank that no longer exists. The primary purpose of the risk reports was to satisfy the regulators. The VaR on a credit instrument was based on the rating. That satisfied the regulator. If the regulator had not believed the rating firm, nobody would have cared what the rating firm said.
If the rating agencies are still in business today, it is because the regulators still believe in the ratings.
The problem is that the primary cost of running one of the legal oligopoly ratings agency (NRSRO's) is wages and intellectual property. They need cash flow to pay their analysts and to license their modeling software, so being compensated in assets of legally-imposed illiquidity isn't feasible.
Now, theoretically, they could generate cash flow by borrowing against these assets as collateral, but how would creditors value the collateral? You can see that you would quickly end up in a viscous circle.
So, what you could do instead is require the NRSRO's and their clients to put their money where their mouths are by *warranting* the quality of the analysis, and incurring certain liabilities should the underlying asset not perform as anticipated.
The company that sells an asset that gets a AAA rating, for example, could be required to deliver those assets with put options at a strike-price at the lower-bound of the expected AAA variance, and priced at the anticipated risk premium, that allow buyers or creditors to recover should the asset move below that band. The ratings agency could back up the option by being required to issue a credit-default swap as insurance against unexpected default, and they could be required to back up those CDS' with Tier-1 rock-solid Collateral provided, as bond, by the company being rated.
So, the higher the company or asset is rated, the cheaper it is for them to debt-finance, but this would be offset by the burden of insurance - and the ratings agencies could get paid by the client, but never enough to offset the costs of having to pay out the insurance claims of a pumped-up rating.
Bartering is impractical, and I'm sure there's some alternative financial wizardry that could accomplish the effect I suggest, but the bottom line is that warranty liability might solve this entire problem.
Two criticisms would be:
1. Even with insurance, the catastrophe was so widespread that there wouldn't have been enough to pay out all the creditors, and the NRSRO's would have gone broke long before then.
Yes, and that's essentially what happened to AIG - but then again, AIG got bailed out and all their insurance obligation guaranteed by the government precisely for this reason. The NRSRO's would be treated in the same way, and while creditors would be protected, this presents second-order too-big-to-fail moral-hazard implicit-subsidy problems.
2. Wouldn't a better alternative would be to house the NRSRO function within the government as kind of a public-good dissemination of analysis so people can use the data to pursue their interests?
Well, the government actually used to do things like this internal to some agencies, and it simply didn't work out very well for all kinds of well-studied institutional bureaucratic-culture reasons. That's why they essentially blessed-off on the big-3 NRSRO's and no one else can ever break into that market.
A final note - notice the European Central Bank rules also depend on the opinions of the big-3. If two of them agree that an EU sovereign debt-issuer's bonds are "junk" then the ECB can neither buy them nor accept them as tier-1 collateral when it lends to member-country banks (well, supposedly, for the EU is not much of a 'hard-rules' institution).
Greece was downgraded to junk by S&P this week (too late, if that isn't obvious to everyone), and it's only a matter of time before Moody's or Fitch does the same. If they had done so before the bailout plan - European banking would be in utter chaos, so that probably tells us a lot about the particular timing of all this. The EU officials made a show of specifically and publicly (but dog-whistle covertly) asking the other ratings agencies to cool their heels while things got worked out.
Do governments really want NRSRO's to be so honest they won't be solicitous of request for indulgences?
It's a good idea, Steve. The only tweak I'd add would be to adjust the amount they have own by the rating they give the security (i.e., the higher the rating, the more they own). And if they rate something junk they have to short it (or buy CDS on it).
It's useless, because you can still bribe someone with junk bonds as with regular bonds. You just give them a bit more. The real problem is even having ratings at all. the government should not sanction, license and prevent competition with these ratings. The market should set the value of the instruments. A bad signal is sent to retail investors when the government in effect ends up telling people how risky things are. Let them judge on their own.
TCO is right.
Advocates of a government monopoly on economic quality control assume that the monopoly (such as the SEC, the FDA, and the FCC) invariably acts ethically or can be made to do so. Little troups of private watchdogs assume only that their audience has some level of intelligence. Although it's hard to say which of these assumptions is the more reasonable, remember that buying off one organization is more efficient than buying off many.
Also simple would be to create competing rating agencies. Whichever has the better track record can use that to tout itself as the rating agency of choice. They could also publish their own (expensive) newsletter. Shadowstats sells its services.
It would be good to create a new exchange as well. The new exchange should be in a red state like Texas with no income tax. If it could maintain higher ethical standards (not difficult), it could bleed business away from New York.
There already are competing rating agencies. The government expanded the NRSRO oligopoly a year or two ago.
"The real problem is even having ratings at all"
You need ratings because most money managers and pension funds don't have the resources to analyze every single fixed income security. But another approach to this would be for a consortium of pension funds and other institutional investors to create and fund their own rating agency, which wouldn't get paid by issuers.
It seems intuitively dumb for the sellers of securities to pay the ratings agencies. But the problem is that the purchasers know damn well that's what's going on. I had trouble making sense of it until Charles Calomiris explained: the purchasers WANTED misrated securities. They wanted the return that came riskier assets, but didn't want to face higher regulatory hurdles under Basel II.
This could not work with the current ratings agencies for roughly the reasons Indy lays out. If the liquidation value of a ratings agency is far below the value of the securities it is thinking about lying about, then, because of limited liability, you can never provide it with adequate financial incentives not to lie. Indy says roughly this in a very detailed way.
The two potential ways out of this are 1) to have ratings agencies which are privately owned by very rich people (who are willing to post enormous bonds), so that there is someone to go after when the securities go bad, or 2) use non-financial penalties like putting people in jail for fraud.
Method 1 would be expensive since those rich folks would have to be very well compensated to tie up their assets and take on this kind of risk. Method 2 is what we have now --- what is supposed to happen now is that a bunch of people at the ratings agencies and at the investment banks go to jail and forfeit all their money.
"the purchasers WANTED misrated securities. They wanted the return that came riskier assets, but didn't want to face higher regulatory hurdles under Basel II."
If that's true (unlikely, because it would imply a collusion between the rating agencies and the purchasers of the securities for which there should be plenty of evidence) the purchasers were fools. Higher risk doesn't equal higher returns.
Anonymous #1:
The way to fix them is to say that the regulators no longer believe them.
That would require an Act of Congress. And the chances of Congress doing that right are lower than that of the rating on a mortgage-backed security being correct.
Emotionally I agree with polymath, here our pension ( superannuation funds) are big enough by volume to pay for their own rating analysis , just too lazy. I imagine yours would be too. And what's stopping the much touted US free market- tell your government to back off- much good they've done so far. Rating should be a finacial service offered freely in the market and we'' soon see who can cut it and who can't. After all what more harm can they do? MIM
As a Australian lawyer I could never understand how the ratings agenciess could get away with such a blatent conflict of interest for so long- being paid by by the entity requesting the rating. Looks like chickens are finally coming home to roost in the USA wiht one of your big pension funds being the first to sue a rating agency; http://zerohedge.blogspot.com/2009/07/calpers-lawsuit-against-rating-agencies.html
Good comments! I'll repeat a couple: (1) The rating agencies are now selling government- or contract-required Ratings, not Information (sort of like colleges--- you pay it for an A-average, not for an education), (2) If the agencies had to take payment in the securities, they'd just raise their nominal price and happily take them-- those securities do have non-zero value remember, just not the same value as truly AAA securities.
What we should expect to happen is a two-niche system of rating agencies. You go to S+P to buy Ratings, and they only sell AAA. You go to New Agency for Information, and they tell you the true value, which customers also would like to know. But maybe you don't make the New Agency rating public-- you just give it to buyers who ask, or call it an unofficial rating.
Post a Comment