Reform the Market for Ratings Agencies
Kevin Drum wonders what can be done about the ratings agencies:
Beyond that, I’m also a bit flummoxed about what the answer to the ratings agency problem might be. There’s probably a reasonable regulatory solution for fraud and negligence, but there seems to be wide agreement that the real problem is incentives: since issuers are the ones paying for ratings, it’s inevitable that agencies are going to lean into the wind to provide ratings the issuers like. I’ve read dozens of proposals for ratings agency reform, but the only one that really gets at this fundamental conflict-of-interest problem is to simply do away with them and turn debt rating into a government function. I’m a little skeptical of that, though, since it’s not at all clear to me that a government agency could hire the kind of talent it takes to keep up with Wall Street’s rocket scientists. What’s more, it’s not at all clear to me that anyone — Fed regulators included — would have rated SIVs much differently during the boom years than the ratings agencies did.
I’ve been told by people working in finance that in their opinion it would be feasible to use regulation to simply switch the payment scheme around and make it so that buyers of securities rather than issuers were the ones paying the ratings agencies. It’s not totally clear to me that that’s correct (for any given security you have one seller and many potential buyers so it seems it would be much more efficient to have the sellers pay) but that’s what I was told.
But I think the larger issue with the ratings agencies isn’t so much that they’re underregulated as it is that regulations we’ve put on other actors in the marketplace have created a ratings agency cartel. The underlying the premise of the idea that private ratings agencies can work is that agencies that fail to do a good job will fail as businesses. That can’t happen if there are only three ratings agencies and it’s impossible for new competitors to enter the market. As Mark Calabria explains in a paper whose general conclusions I wouldn’t embrace:
The modern regulation of credit rating agencies began with the SEC’s revision of its capital rules for broker-dealers in 1973. Under the SEC’s capital rules, a broker-dealer must write down the value of risky or speculative securities on its balance sheet to reflect the level of risk. In defining the risk of held securities, the SEC tied the measure of risk to the credit rating of the held security, with unrated securities considered the highest risk. Bank regulators later extended this practice of outsourcing their supervision of commercial bank risk to credit rating agencies under the implementation of the Basel capital standards.
The SEC, in designing its capital rules, was concerned that, in allowing outside credit rating agencies to define risk, some rating agencies would be tempted to simply sell favorable ratings, regardless of the true risk. To solve this perceived risk, the SEC decided that only Nationally Recognized Statistical Rating Organizations would have their ratings recognized by the SEC and used for complying with regulatory capital requirements. In defining the qualifications of an NRSRO, the SEC deliberately excluded new entrants and grandfathered existing firms, such as Moody’s and Standard and Poor’s.
In trying to address one imagined problem, a supposed race to the bottom, the SEC succeeded in creating a real problem, an entrenched oligopoly in the credit ratings industry. One result of this oligopoly is that beginning in the 1970s, rating agencies moved away from their historical practice of marketing and selling ratings largely to investors, toward selling the ratings to issuers of debt. Now that they had a captive clientele, debt issuers, the rating agencies quickly adapted their business model to this new reality.
The damage would have been large enough had the SEC stopped there. During the 1980s and 1990s, the SEC further entrenched the market control of the recognized rating agencies. For instance, in the 1980s the SEC limited money market funds to holding securities that were investment grade, as defined by the NRSROs. That requirement was later extended to money market fund holdings of commercial paper. Bank regulators and state insurance commissioners followed suit in basing their safety and soundness regulations on the use of NRSRO-approved securities.
This kind of situation could conceivably work if there were a large number of NRSROs but instead there are only three. And they essentially have the entire market in global finance captive. It’s a recipe for disaster.
15 Responses to “Reform the Market for Ratings Agencies”
Register
RSSLogin
Topic Cloud
ArchivesBlog RollAbout Matt YglesiasContact Me
Donate
September 29th, 2009 at 10:24 am
It’s a recipe for disaster.
Since we’re in a disaster aided by ratings agencies, it’s a “tried and true” recipe for disaster.
Who, btw, let Warren Buffet buy significant positions in a brokerage, a ratings agency, and an insurance company? I get dizzy from reflected ethical vertigo just typing that damn question.
September 29th, 2009 at 10:38 am
A larger pool would just get you the race to the bottom problem again. Already with just three, Fitch is kind of the doggy one whose ratings were not seen on equal footing of the others.
The bigger problem is letting a somewhat subjective rating process have objective benefits. If the regulators need to have some objective measure of safe investing, they should have based in on some objective formula of the financial wherewithal of the counterparty on the security.
On the other hand, there are no privileged ratings for stocks, and they muddle along. Maybe investors just needed to rely on their own due diligence for their fixed income investments as well.
Also Buffett has been selling Moodys
http://dealbook.blogs.nytimes.com/2009/09/04/berkshire-pares-stake-in-moodys-again/
September 29th, 2009 at 11:01 am
I don’t understand this obsession with ratings agencies. Obviously they were in the wrong and had perverse incentives that lead them to label crap as AAA debt. But people investing millions of dollars–billions sometimes–have their own analysts, have access to independent analysts, and could even see credit default swaps that disagreed with the ratings from S&P and Moody’s. I’m sure there were buyers of these toxic waste securities who were naive dupes. But most were sophisticated buyers who would never rely solely on the ratings to make their decisions. They knew what they were buying was riskier than the ratings said. They 1) wanted the higher returns (and were willing to accept the associated risk), 2) didn’t have skin in the game and indeed had incentives to take higher risks with OPM, and 3) used the high debt ratings as a CYA in case the investments blew up. In other words, sure the ratings agencies are culpable for what they did and should be reformed, but the problem was systemic and has to be solved systemically. Buyers were just as culpable as sellers and the ratings agencies.
September 29th, 2009 at 11:12 am
How about a stamp duty on all financial products that need to be rated. Issuers will be randomly assigned a ratings agency and the cost will be paid out of the stamp duty revenue.
If more the rated products default at a rate higher than expected by a certain amount the agencies loose their right to rate.
Perhapse there can be farm team ratings agencies that only rate smaller issues and they could be gromed to replace the larger agencies if they drop the ball again.
September 29th, 2009 at 11:12 am
You can deal with the “fly-by-night” new ratings agencies by only allowing new raters to issue the highest risk ratings. Every 2 or three years, if they have been doing their job well, they are empowered to issue a less risky rating.
Rate the raters. While the government might not have the expertise to rate all bonds, they certainly have the ability to rate the raters. Keep a public record of the performance of each agency with regard to the performance of each rated investment. New ratings agencies will need to outperform existing ones if they want to succeed.
Tax the ratings. The government can tax each rating of a security based on the total value issued. The proceeds of the tax go to creation of a future bail-out fund. Since AAA ratings have the most potential for harm, they would be charged a higher fee. This creates a disincentive for rubber stamping things AAA.
Rate the financial environment. The government can create an independent commission to examine the general value of assets that are traditionally used as collateral. Before the most recent crisis, they might have issued a finding that residential real estate was overvalued compared to historical levels, and should be treated as reduced in value for collateral. They could do the same for things like general stock holdings if they think the stock market as a whole is overvalued. They could do the same for future eanings as collateral if they think a recession is coming. The ratings agencies would still do the specific analysis, but if their actions went against the grain of the government commission’s findings, they would be subject to more scrutiny from investors.
September 29th, 2009 at 11:16 am
(for any given security you have one seller and many potential buyers so it seems it would be much more efficient to have the sellers pay)
So on the one hand we have efficiency, and on the other we avoid a total collapse of the economy necessitating $trillions in bailouts. I think we can sacrifice some efficiency.
September 29th, 2009 at 11:26 am
Njorl @5, if the government is going to separately conduct its own ratings, why not just cut out the middleman? The ratings function could be taken over by the government, and funded by a tax on all securities transactions.
September 29th, 2009 at 11:40 am
I’m shocked.
1. Government regulations have resulted in only three ratings agencies? Who knew?
2. The government’s very own Federal Reservie made the pricing of everything almost impossible to determine due to its mission of monetary dilution. People then rushed to purchase housing in order to have a store of value in the face of monetary dilution. which itself was the source of funding for the housing bubble. Which everyone in a position of authority denied was even occurring except for the Austrian School (and a few mainstreamers here and there).
3. Therefore, let’s put the government in charge of everything and hold our breath when the Republicans return to run the government.
September 29th, 2009 at 11:43 am
This kind of situation could conceivably work if there were a large number of NRSROs but instead there are only three.
I don’t know why Matthew persists in making this statement. It is false. There are ten approved NRSRO.
September 29th, 2009 at 11:53 am
The government isn’t doing any rating in my scheme. The government looks at the state of the real estate market, the raters assess the the value of underlying property. The government predicts the effects that most companies can expect on their future earnings, while ratings agencies look at specific companies. The agencies here would be doing several orders of magnitude more work than the government commission.
The government does all of this work already, I’m just suggesting they package it and use it to set up a system of warning flags which are easily accessible and commonly considered.
September 29th, 2009 at 12:18 pm
I will also note, to reiterate what I wrote yesterday, that some of the approved NRSROs are NOT paid by issuers, and are instead paid by investors. For example, Egan-Jones and Realpoint.
The reforms necessary to open up the credit rating agency market have already occurred.
September 29th, 2009 at 2:16 pm
I never really understood why it was considered wrong to get insurance on securities (hence the CDS work-around), but risk ratings are in some sense mandated by the government.
Effectively an insurance company would rate the risk when it insured something, but as you pay the premiums over time, instead of up front for the risk rating, and the “rating agency” — the insurance company — is on the hook for the losses if it turns out to be wrong, you avoid some of the nonconstructive incentives. But this creates a leverage problem since the financial companies booked CDS (insurance) payouts as covering their losses, and insurance companies never think they are going to have to pay out to everyone they cover (that’s how insurance works).
The best (and simplest) solution is just to deregulate and tax the heck out of any security sale based on a government risk rating so as to pay to clean up the mess when they eventually fail, and tax the heck out of salaries and bonuses so as to continuously “claw back” executive gains — punishing risky behavior while they are engaging in it.
September 29th, 2009 at 3:39 pm
Its a tricky situation. Opening up competition in ratings organizations will just lead us back to companies paying for ratings to game the system. Letting the buyers pay ratings organizations will just invert the problem. Buyers will pay for lower ratings so they can get a better price, which would have the added consequence of increasing market volatility.
Maybe a better solution would be to link ratings organization pay to the performance of the investments they rate. Give them a very small percentage of the profit or losses from investments they rate, but make their return a function of the rating itself, so that if they get the rating right, they get the maximal return, if they get it somewhat wrong, they break even, and if they get it very wrong, they go broke.
The ratings organizations have an incentive to take a deal like this because they’ll be profit without having to use their own capital.
I also tend to think that coming up with a compensation scheme like this wouldn’t actually be too hard. Modern Portfolio Theory, basically provides the equations.
September 29th, 2009 at 9:54 pm
Like you (Matt Yglesias) I don’t think making the purchasers of securities pay for ratings will make much difference. I think the incentive problem related to SIVs was not mostly that SIV issuers effectively bribed ratings agencies. I think the problem is that, if the ratings agencies had done their jobs properly, the whole industry wouldn’t exist.
Ratings agencies benefit if there are more financial instruments to rate. That means that, whoever pays them, they have an incentive to be generous in rating innovative financial products. This generosity can not be sound if the product is so new that there can’t be enough data to rate it.
please please click here
I’d say a proper incentive scheme could be developed in which the ratings agencies were paid an annual lump sum, required to rate traditional instruments (corporate bonds say) and allowed to rate other stuff for the salary of those who do nothing else but rate that instrument plus 20%.
Now one could make an incentive contract of the following
September 30th, 2009 at 7:39 am
There is actually a relatively simple way to re-align rating agencies’ incentives without doing violence to the rest of the world’s regulatory architecture. The regulator should fine them every time a firm/structured instrument goes bust within a certain period of time of the agencies putting an investment grade rating on it. Say three years. The higher the rating, the larger the fine. So if you rate a firm/security BBB and it goes bust, you pay a few peanuts. If you rate a firm/security AAA and it defaults you fork over half your bank account.
The biggest problem I can see is that the regulators would be tempted to use this scheme as a cash machine instead of trying to carefully align the fines to offset the rating agencies’ incentives. After all, if we make the fines too draconian, we end up with no investment grade ratings, which is as bad as having too many. Perhaps fines should go to creditors of the bankrupt firm?