By Douglas Lucas, Managing Director, SEDA Experts LLC
Critics of credit rating agencies’ issuer-pay revenue practice say that it allows debt issuers and arrangers to shop among agencies for the highest rating. They allege that such “rating shopping” incents rating agencies to win business by rating debt higher than warranted by actual credit risk. To eliminate inflated ratings, issuer-pay critics suggest barring debt issuers from selecting and paying rating agencies.
Issuer-pay critics assume that ratings would be accurate if only the issuer-pay incentive for inflated ratings was extinguished. But we document a historical test of this assumption, an instance when issuer-pay and rating shopping were irrelevant concerns and ratings were still inappropriately high. With issuance effectively zero in 2007-09, S&P had no commercial incentive to maintain inflated ratings on subprime mortgage-backed securities and collateralized debt obligations backed by subprime mortgage-backed securities.
Yet, S&P’s ratings were severely inflated when compared to market prices and the credit analyses of market observers. Because 2007-09 conditions still produced inflated subprime ratings, getting rid of issuer pay will also not be sufficient to improve rating agency accuracy. But if issuer-pay critics succeed in abolishing the revenue practice, it would only be the latest in a long history of ineffectual and counter-productive rating agency regulations.
Getting Rid of Issuer-Pay Will Not Improve Credit Ratings
The credit rating agencies’ issuer-pay revenue practice, where debt issuers choose and pay the agency that rates their debt, is under scrutiny (Warren 2019, Podkul 2019b, Grassley et al 2020, Podkul 2020a and Podkul 2020b). Critics claim that the practice allows debt issuers and arrangers to shop among agencies for the highest rating. Such “rating shopping” is said to incent agencies to win business by offering ratings higher than warranted by actual credit risk. Issuer-pay critics also claim that the entry of new rating agencies over the past ten years has exacerbated the competition to supply inflated ratings (Podkul and Banerjo 2019).
To stop rating shopping and eliminate inflated ratings, some issuer-pay critics suggest that rating agencies be randomly assigned to rate debt issues, rather than chosen by issuers. They also suggest rating agency payments be automatically deducted from debt proceeds rather than paid at the pleasure of debt issuers (Kotecha 2012). Issuer-pay critics assume that ratings would be accurate if only the issuer-pay incentive for inflated ratings was extinguished. But eliminating an incentive to produce inflated ratings creates neither the incentive nor the ability to produce accurate ratings. The US subprime mortgage crisis provided a unique opportunity to test rating agency performance in the absence of commercial incentives.
We detail a two-year period, July 2007 to June 2009, when Standard & Poor’s (S&P) had no commercial incentive to inflate the ratings of subprime mortgage-backed securities (subprime bonds) and collateralized debt obligations backed by subprime bonds (subprime CDOs).[1] Because issuance of subprime bonds and subprime CDOs was effectively zero during those two years, angling for new-issue rating fees was not a commercial consideration. No issuer-pay incentive prompted S&P to maintain erroneously high ratings on the subprime bond and subprime CDO ratings it monitored. In fact, given the intense public interest at the time in rating agency performance in the subprime sector, S&P had significant reputational incentives to get ratings right. Yet, we show that S&P’s ratings were inflated by comparing them to market prices and the credit analyses of market observers. Because banning issuer-pay revenue would fall short of creating the conditions that existed 2007-09, we do not think it would improve credit rating accuracy.
But if issuer-pay critics succeed in abolishing the revenue practice, it would only be the latest event in a long history of ineffectual and counter-productive rating agency regulations. From the Great Depression to the Dodd Frank Act, rating agency regulators have misunderstood rating agency incentives.
Rating accuracy is important because regulators and private parties use credit ratings for diverse purposes such as calculating required capital, setting investment restrictions, and collateralizing bilateral credit agreements. The US Department of Justice drew a line from rating agency errors to the 2007-08 financial crisis when it sued S&P in 2013. The department alleged (US Department of Justice 2013) that in years preceding the crisis, “S&P’s concerns about market share, revenues and profits drove them to issue inflated ratings, thereby misleading the public and defrauding investors. In so doing, we believe that S&P played an important role in helping to bring our economy to the brink of collapse.” S&P settled the suit for $1.5 billion in 2015 (Settlement Agreement 2015).
This paper has four sections. “The Demise of Subprime Issuance,” shows how little issuance there was after June 2007, thus eliminating any commercial incentive for S&P to maintain inflated ratings on outstanding debt to gain new issue rating mandates. “Subprime Credit Conditions 2007,” uses market prices and sell-side analyst research to show the deteriorating credit quality of subprime bonds and subprime CDOs. “S&P’s Subprime Ratings 2008-09,” shows that S&P maintained inflated ratings when it had no commercial incentive to mis-rate debt. “Rating Agency Regulation Failures” gives a brief history of efforts to improve rating accuracy that failed because regulators did not understand rating agency incentives.
The Demise of Subprime Issuance
Subprime bond issuance went from $509 billion in 2006 and $203 billion in the first half of 2007 to $27 billion in the second half of 2007 and virtually zero thereafter. There was a $303 million issue in January 2008, a $48 million issue in October 2008, and nothing in 2009. Month-by-month issuance is shown in Exhibit 1 (Flanagan, et al 2004-09). In the second half of 2007, the financial infrastructure to originate subprime mortgage loans was effectively dismantled as originators went bankrupt, sold themselves to prime originators, laid off employees, or otherwise shut down subprime lending (Sharif 2007).
Exhibit 1
Monthly Subprime Bond Issuance 2004-09
Source: JPMorgan’s Global ABS/CDO Weekly Market Snapshot, 2004-09.
Meanwhile, subprime CDO issuance went from $182 billion in 2006 and $113 billion in the first half of 2007 to $916 million in the second half of 2007 and virtually zero thereafter. There was a $20 million issue in February 2008, a $75 million issue in December 2008, and two issues totaling $821 million in February 2009. Month-by-month issuance is shown in Exhibit 2.
Exhibit 2
Monthly Subprime CDO Issuance 2004-09
Source: JPMorgan’s Global ABS/CDO Weekly Market Snapshot, 2004-09.
The $113 billion of subprime CDO issuance in the first half of 2007 overstates the health of that market. Many of those CDOs resulted from broker-dealers preemptively liquidating the subprime bond warehouses they financed for CDO managers. The dealers packaged subprime bonds from various manager warehouses together into unmanaged, static CDOs they issued at a loss. Likewise, the burst of issuance in February 2009 does not indicate a revival of CDO issuance. Those CDOs were not created from newly issued subprime bonds and they were not sold to investors. The CDOs were created so that their AAA tranches could be posted as collateral for loans from the European Central Bank. As these CDOs’ asset portfolio deteriorated in 2009, the banks took the CDOs back from the European Central Bank and dissolved them before they could default.
The amount of new ratings business to be gained in subprime bonds and subprime CDOs after June 2007 was trifling. S&P thus had effectively zero commercial incentive to maintain inflated ratings on outstanding subprime bonds and subprime CDOs in order to gain new rating mandates.
Subprime Credit Conditions 2007
Market consensus on the credit quality of subprime bonds over 2007-09 can be gleaned by ABX index prices. ABX indices were created every six months from 20 subprime bond transactions issued the previous six months. Thus, 2006-1 indices referenced 20 subprime bond deals closing July to December 2005, 2006-2 indices 20 deals closing January to June 2006, 2007-1 indices 20 deals closing July to December 2006, and 2007-2 indices 20 deals closing January to June 2007. Each of these ABX “rolls” was comprised of six indices, each referencing 20 like-rated bonds, one from each of the 20 deals. There were indices for penultimate-pay AAA bonds, last-pay AAA bonds, and ones for bonds rated AA, A, BBB and BBB-.[2]
Each ABX index formed the basis of a credit default swap. Semiannual protection payments, from protection buyer to protection seller, were set at the time of the index’s conception and never varied, not even for new contracts entered into much later. Instead, to account for changes in perceived credit risk, an upfront payment, expressed as a percent of par, was made at the time the swap was entered into.
Loosely speaking, these upfront payments can be viewed as the market’s best guess as to future losses, in a gambling sense, the “over-under.” Thus, an upfront payment of 50% of par means that protection buyers are willing to buy protection at that price because they think losses will be greater than 50% of par. Protection sellers are willing to take 50% of par upfront because they don’t expect protection payments to exceed 50%.[3]
Exhibit 3, shows upfront payments, or, by our explanation, loss expectations on BBB- ABX indices. Loss expectations for 2006-2, 2007-1, and 2007-2 were 58%-62% of par at the end of July 2007. ABX 2006-1 was relatively creditworthy with a 41% loss expectation. By the end of 2007, loss expectations for all four indices were 71%-81%. Loss expectations increased further in 2008, when we begin comparing ABX prices to S&P’s ratings.
Exhibit 3
Upfront Payments for Credit Default Swaps on ABX BBB- Indices
Source: UBS Securitized Product Research Exhibit 4 shows upfront payments, or loss expectations, on last-pay AAA ABX indices. By the end of 2007, loss expectations for all four indices were 6%-25%. Loss expectations increased further in 2008, when we will begin comparing ABX prices to S&P’s ratings.
Exhibit 4
Upfront Payments for Credit Default Swaps on ABX Last Pay AAA Indices
Upfront payments express an expectation of future credit losses.
Source: UBS Securitized Product Research
The pessimism reflected in ABX prices was shared by sell-side analysts at UBS and JP Morgan. In February 2007, UBS named six bonds in the 2006-2 BBB- index it said would default (Zimmerman et al 2007). JP Morgan went further that same month, naming 11 bonds in the 2006-2 BBB- index and nine bonds in the 2007-1 BBB- index it thought would default (Flanagan 2007a). But by September 2007, the view was even more pessimistic. UBS’ opinion that month was for subprime bond losses to extend up to most A tranches and losses on CDOs backed by BBB- to A+ subprime bonds to extend up to senior AAA tranches (US Senate Permanent Subcommittee on Investigations 2010 and Lucas 2007a). Analysts’ pessimism stemmed from their consideration of an increasing body of evidence.
In September 2007, the latest Case-Shiller Home Price Index for 20 US metropolitan areas showed the greatest year-over-year decline ever, 4.0% in July. Among the cities with the largest year-over-year declines were Detroit (9.7%), Tampa (8.8%), San Diego (7.8%), Phoenix (7.3%), and Washington DC (7.2%) (Sharif 2007). Back in December 2006, UBS’ chief economist predicted a 15% decline in home prices because the high proportion of vacant houses among houses for sale portended desperate sales.
The effect of home price depreciation on subprime credit quality is dramatic. Besides creating greater losses in the event of a default, home price depreciation increases default frequency, particularly among investor properties. When JP Morgan analysts predicted in February the default of 20 of the 40 bonds making up the BBB- tranches of 2006-2 and 2007-1, they did so under their base case expectation of 0% home price appreciation in 2007. At the same time, they said that if home prices declined 3% in 2007, 36 of those 40 bonds would default. In fact, assuming a 3% home price decline, JP Morgan predicted 31 of the 40 bonds in 2006-2 and 2007-1 BBB tranches would default (Flanagan 2007b). In the seven months since JP Morgan made those predictions, home prices had already depreciated 3.5%.
Loan delinquencies, defaults and losses kept mounting. Rather than curing, or remaining in the same “delinquency bucket,” a greater percentage of delinquent loans were “rolling” into greater delinquency. For example, a greater percentage of 30-day past due loans were becoming 60-day past due rather than either remaining 30 days past due (the borrower having made one payment and remaining one payment behind) or becoming completely cured. But besides home price depreciation and mounting delinquencies and defaults, another factor suggested that things were going to get worse.
In the past, most subprime borrowers refinanced their loan before its interest rate reset, when the loan came out of its low interest rate teaser period, jumped to a higher interest rate, and incurred higher monthly mortgage payments. Homeowners who couldn’t refinance often defaulted because of increased payments. But new subprime loan origination was effectively shut down and borrowers’ refinancing opportunities narrowed to Federal Housing Administration (FHA) loans or Federal Home Loan Mortgage Corporation (Freddie Mac) and Federal National Mortgage Association (Fannie Mae) conforming loans under the agencies’ expanded subprime programs. However, these alternative refinancings were only available to the most creditworthy among subprime borrowers. Homeowners who could not refinance and who could not afford higher payments were compelled to either negotiate a loan modification with their loan servicer, entailing a loss to the mortgage holder, or default outright (Zimmerman 2008).
Home price depreciation also caused an increase in strategic defaults, where the borrower has the ability to make mortgage payments but decides not to do so for economic reasons, usually because of declining home prices and resulting negative home equity. The social stigma against defaulting had weakened and it was becoming more acceptable to default on one’s mortgage obligation. Consulting services sprang up to help homeowners default with minimal cost and the least damage to their financial reputations. In some cases, delaying tactics could even garner months of free rent, thus creating higher loan losses. The consulting firm whose name most characterized the new attitude toward default was “youwalkaway.com.” Later, politicians started using the term “predatory lending” to explain the mortgage crisis, giving homeowners a moral justification for defaulting.
S&P’s Subprime Ratings 2008-09
The rating history we will relate in this section shows that when S&P had no commercial incentive, it still maintained obviously inflated ratings on subprime bonds and subprime CDOs. This history convinces us that banning issuer pay, in an attempt to eliminate commercial incentives for inflated ratings, is not sufficient to improve rating accuracy.
We’ve shown that subprime bond and subprime CDO issuance crashed in July 2007 and we’ve shown the market’s pessimistic view of subprime credit quality in 2007 as embodied in falling ABX prices. We’ve also related market conditions and sell-side analyst opinion as of September 2007. But let’s generously assume it would take a while for S&P to realize that the subprime market was dead and that there would be no more revenue from new subprime securitizations. Let’s further assume that S&P felt it should move slowly, even though subprime debt was being traded in the secondary market, informed by S&P’s ratings. Finally, let’s assume that it would take a while for S&P to realize that such downgrades were the only way for it to salvage some of its reputation and it would take some time for S&P to actually downgrade credits. So, we won’t begin looking at S&P’s subprime ratings until April 2008.
Subprime Bond Ratings in April 2008
In April 2008, a subprime bond hadn’t been issued since a $303 million issue in January. Upfront payments on ABX BBB- indices were 85%-94%, depending on the vintage. Upfront payments on last pay AAA indices were 5% to 42%. In line with these ABX prices, UBS subprime analysts predicted that 292 of the 400[4] bonds composing the four rolls of the ABX index would be at least partially written down, with some loss of principal (Lucas 2008c). In fact, more bonds eventually defaulted.
S&P rated all the bonds UBS said were going to default and rated some highly. Exhibit 5 shows the then-current S&P rating distribution of these subprime bonds that UBS predicted would be at least partially written down. For example, the second row, second column of the exhibit shows that UBS predicted that 24 bonds rated AAA at 1 April 2008 by S&P would be written down.
S&P downgraded more than 600 subprime bonds in July 2007 and by April 2008 had downgraded 219 bonds underlying ABX indices. It had even downgraded seven ABX bonds originally rated AA all the way down to CCC. But it had not downgraded enough bonds and most of the ones it downgraded were not downgraded far enough. S&P had not downgraded 78 of the 292 bonds UBS said would default and it only rated 29 of the 292 bonds CC, the rating that expresses a future default to be a “virtual certainty.”
Exhibit 5
S&P Ratings of ABX Underlying Bonds That UBS Predicted Would be Written Down,
1 April 2008
Source: UBS Securitized Product Research Exhibit 5’s third and fourth columns quantify S&P’s discriminatory power in another way. The third column shows that the 24 AAA-rated bonds UBS thought were going to default were 8% of the 292 predicted defaults. The fourth column presents a running total going down the exhibit: 10% of bonds UBS said were going to default were rated AA+ or higher, 18% AA or higher and so on. The column shows that 33% of bonds UBS thought were going to default were rated investment grade, BBB- or higher. According to S&P, a bond rated BBB- “exhibits adequate protection parameters.” Forty-three percent of the bonds UBS thought would default were rated B or higher. According to S&P, the obligor of a bond rated B- “has the capacity to meet its financial commitments on the obligation.”
Subprime Bond Ratings in July 2008
By July 2008, a subprime bond still hadn’t been issued since the $303 million issue in January 2008. Upfront payments on ABX BBB- indices were 91%-95%, depending on the vintage. Upfront payments on last pay AAA indices were 11% to 56% and upfront payment on the new penultimate pay AAA indices were 5% to 47%. UBS subprime analysts repeated the April analysis, but this time focusing on the 257 of 480 ABX bonds they believed would be written down completely, with all principal on the bonds lost (Lucas 2008d). In fact, all 257 bonds eventually defaulted with total principal loss, and higher rated tranches above them in the deals’ capital structure also defaulted.
S&P rated all the bonds UBS said were going to default with no principal return and rated some highly. Exhibit 6 shows the then-current S&P rating distribution of these subprime bonds that UBS predicted would be written down completely. For example, the second row, second column of the exhibit shows that UBS predicted complete write-downs for five bonds S&P rated AA at 14 July 2008.
S&P had downgraded 52 more bonds underlying ABX indices since April, rating seven more bonds CC. But it had not downgraded enough bonds and most of the ones it downgraded were not downgraded far enough. S&P had not downgraded 26 of the 257 bonds UBS said would default with no principal recovery and it only rated 36 of the 257 bonds CC, the rating that expresses a future default to be a “virtual certainty.”
Exhibit 6
S&P Ratings of ABX Underlying Bonds That UBS Predicted Would be Completely Written Down, 14 July 2008
Source: UBS Securitized Product Research
Exhibit 6’s third column shows that the five AA-rated bonds UBS thought were going to default with no principal return were 2% of the 257 predicted defaults. The fourth column presents a running total going down the exhibit: 3% of bonds UBS said were going to default were rated AA- or higher, 5% A or higher and so on. The column shows that 15% of bonds UBS thought were going to default with no return of principal were rated investment grade, BBB- or higher. According to S&P, a bond rated BBB- “exhibits adequate protection parameters.” Thirty-three percent of the bonds UBS thought would default on all principal payments were rated B or higher. According to S&P, the obligor of a bond rated B- “has the capacity to meet its financial commitments on the obligation.”
Subprime CDO Ratings in 2008-09
Like subprime bond issuance, subprime CDO issuance crashed in July 2007. In all of 2008, only $95 million was issued. Back in August 2007, UBS had predicted severe losses on the senior AAA tranches of subprime CDOs including ACA ABS CDO 2006-2 (Lucas, Li, Shinozuka, and Mladinich 2007). Exhibit 7 shows a timeline of S&P’s rating actions in 2008. S&P first downgraded the tranche into a distressed rating category in July 2008, 11 months after UBS had declared it destined for severe losses.
Exhibit 7
UBS Default Prediction for the Senior AAA Tranche of ACA ABS CDO 2006-2
and Subsequent S&P Rating Downgrades
Source: UBS Securitized Product Research
Ambac Ratings in 2008-09
In November 2007, UBS estimated $4 billion of losses just on the credit protection Ambac Assurance Corporation had written on subprime CDOs and CDOs of subprime CDOs (CDO squared). Its prediction did not include any other exposure AMBAC had to mortgages. The timeline in Exhibit 8 shows it took 20 months after UBS’ estimate for S&P to downgrade the insurer to speculative grade. But long before that it was obvious that AMBAC could not pay all insurance claims and later, in fact, it did not.
Exhibit 8
UBS Loss Predictions for Ambac Assurance Corporation and Subsequent S&P Rating Downgrades
Source: UBS Securitized Product Research
Subprime Loss Estimates in 2009
S&P continued to hold an optimistic view of subprime credit quality into 2009. Exhibit 9 shows S&P’s base case subprime mortgage loss assumptions at February and July 2009 and predictions from UBS, JP Morgan, Barclay’s, and CitiGroup in November and December 2008. The sell-side analysts made predictions for particular ABX rolls or for all subprime mortgage bonds issued in a particular calendar period. In February 2009, S&P’s loss predictions were substantially less than any of the preceding predictions from sell-side analysts. The rating agency didn’t catch up to sell-side analysts until July 2009.
EXHIBIT 9
S&P’s Subprime Mortgage Loan Loss Predictions versus Those of Sell-Side Analysts
Source: UBS Securitized Product Research
Notes: Barclays predicted subprime losses by ABX index. Citi and S&P predicted subprime losses by calendar year. JP Morgan predicted losses both ways. UBS predicted losses by ABX index, but also for the first six months of 2005.
RATING AGENCY REGULATION FAILURES
We’ve shown that eliminating issuer-pay incentives is not sufficient in itself to eliminate inflated ratings. In 2007-09, S&P had no issuer-pay incentive to inflate subprime bond and subprime CDO ratings, because there was almost zero new issuance. Yet, its ratings on those securities were severely inflated. This occurred despite reputational incentives for S&P to get its ratings right.
Doing away with issuer pay could even harm rating quality. The most well-thought-out schemes to implement a non-issuer-pay revenue structure call for rating mandates to be assigned to rating agencies randomly and for rating agencies to receive regulatory-determined fees. Rating agency quality does not factor into these plans to assign mandates and set fees. Issuer-pay opponents just assume that in the absence of issuer-pay incentives, rating quality will improve. We think it more likely that if rating agency revenue is fixed, agencies will focus on reducing costs rather than on maintaining or improving the quality of their analysis.
But if regulators do ban issuer-pay and institute some alternative revenue scheme, and it does prove ineffectual or counter-productive, it would only add to the long history of flawed rating agency regulation.
The first regulatory blow to credit rating quality occurred decades before rating agencies began charging issuers for ratings. Regulators began appropriating credit ratings for their own use in 1933, when the Office of the Comptroller of the Currency and the Federal Reserve required banks to hold extra capital against speculative-grade bonds. Later in the Great Depression, those regulators banned banks from holding speculative-grade bonds altogether.
The regulatory use of ratings grew over decades. To provide a few examples in the US, the National Association of Insurance Commissioners began using ratings to assess insurance company investment portfolios in 1951. The Securities and Exchange Commission started using ratings to assess broker-dealer capital in 1975. The Department of Labor started using ratings with respect to pension funds in 1989. Congress applied ratings to Savings and Loan regulation in 1989. The SEC applied ratings to money market funds in 1991. Later, the SEC used ratings as the basis for lowering disclosure requirements for investment grade bonds, mortgage-backed securities, and asset-backed securities. Regulators saw credit ratings as a useful tool in their efforts to preserve the safety and soundness of the financial system.
The head of Moody’s corporate and structured rating practices explained how the regulatory use of ratings threatened rating quality in a speech before the SEC in 1995 (McGuire 1995). He argued that if ratings are exclusively used by investors to make investment decisions, rating agencies will strive to make their ratings accurate, even in the presence of issuer-pay incentives. “As long as the product … being sold to issuers [is] credibility with investors, there [is] a natural force, the need to retain investor confidence, to countervail the pressure of rating fees.”
But when ratings are used in regulation, investors’ desires become muddled. Regulated entities like high ratings on the investments they purchase to please regulators and decrease capital requirements. After purchase, debt investors don’t want to see their investments downgraded, and if they are to be downgraded, they want plenty of warning beforehand.
“By using securities ratings as a tool of regulation, governments fundamentally change the nature of the product agencies sell. Issuers then pay rating fees to purchase, not credibility with the investor community, but a license from a government. As a result, officially-recognized rating agencies have a product to sell even when they fail to maintain credibility with the investor community,” said McGuire.
In his plea to stop the use of ratings in regulation, he described the regulatory use of ratings as “a chronic sickness” … “eroding the integrity and objectivity of the credit rating system” ... “like a cancer, it slowly and silently kills the natural defenses the rating agencies need to protect themselves against the economic leverage of issuers on their rating decisions.”
Congress passed the Credit Rating Agency Reform Act of 2006 because “additional competition [in the rating agency business] is in the public interest” (Lucas 2008e). The legislation and subsequent SEC rule-making opened up Nationally Recognized Statistical Rating Organization (NRSRO) status to any firm with a three-year track record, 20 customers, and 10 letters of recommendation. Congress effectively forbade the SEC to apply a rigorous standard to NRSRO designation and now there are nine NRSROs.
Some market observers say the additional competition has not worked as intended and the new rating agencies are competing against older agencies by offering higher ratings and lower credit enhancement requirements. In an analysis of 30,000 ratings on $3 trillion of debt, the Wall Street Journal found that “The challengers tended to rate bonds higher than the major firms. Across most structured-finance segments, DBRS, Kroll, and Morningstar were more likely to give higher grades than Moody’s, S&P, and Fitch on the same bonds. Sometimes one firm called a security junk and another gave a triple-A rating deeming it supersafe” (Podkul and Banerjo 2019).
In 2008, New York Attorney General Andrew Cuomo believed he had stopped the practice of rating agency shopping. His website described the problem and its solution: “The agencies were paid no fees during their initial reviews of the loan pools or during their discussions and negotiations with the investment banks about the structuring of the loan pools. Investment banks were thus able to get free previews of RMBS assessments from multiple credit rating agencies, enabling the investment banks to hire the agency that provided the best rating.” To end this practice, his website announced an agreement with Moody’s, S&P, and Fitch in return for forbearance of further investigation against them: “Credit rating agencies will now establish a fee-for-service structure, where they will be compensated regardless of whether the investment bank ultimately selects them to rate a RMBS.”
The idea was that by providing a conservative rating agency with a pre-rating analysis fee, that rating agency would be less dependent on revenue from actually being chosen to rate the debt. In theory, the income from these analysis fees would allow a conservative rating agency to stick to its rating opinions without suffering economically. But it was never made clear why issuers and arrangers would commission pre-rating analysis from rating agencies they knew to have conservative rating standards.
In 2010, the SEC instituted a rule that data on structured finance transactions must be shared among all rating agencies, even those not asked to rate the transaction’s debt. The idea was to promote rating discipline by encouraging un-hired agencies to make unsolicited ratings. But it has resulted in few, and possibly zero, unsolicited ratings (Podkul 2019a). Rating agencies have no incentive to perform the work required to issue and monitor ratings for free.
From 1933 to 2010, regulators have misunderstood how their regulations incent rating agencies and the entities that buy and use ratings. Regulations have been ineffectual or counter-productive to rating quality.
Conclusion
Critics of credit rating agencies’ issuer-pay revenue practice say that it allows debt issuers and arrangers to rating shop and incents agencies to win business by rating debt higher than warranted by actual credit risk. Some also claim that the entry of new rating agencies over the past ten years has exacerbated the competition to supply inflated ratings. To eliminate this incentive, and concomitant inflated ratings, they suggest barring debt issuers from selecting and paying rating agencies. Rating mandates might be randomly assigned to agencies and rating agency fees might be automatically deducted from debt issuance proceeds.
Issuer-pay critics assume that ratings would be accurate if only the issuer-pay incentive for inflated ratings was extinguished. But eliminating an incentive to produce inflated ratings creates neither the incentive nor the ability to produce accurate ratings. The subprime mortgage crisis provided a unique opportunity to test rating agency performance in the absence of commercial incentives.
We show that market conditions eliminated issuer-pay incentives, yet ratings were still inflated. With issuance effectively zero after June 2007, S&P had no commercial incentive to maintain inflated ratings on subprime bonds and subprime CDOs to gain new issue rating mandates. Given public scrutiny on rating agencies, S&P had a tremendous incentive to get its ratings right. Yet, by comparing S&P ratings to market prices and the credit analyses of market observers, we showed that S&P’s ratings on these products were still inflated. Because the conditions that existed 2007-09 did not produce accurate subprime ratings, we don’t believe that getting rid of issuer pay will improve rating agency accuracy.
But if regulators do ban issuer-pay and institute some alternative revenue scheme, and it does prove ineffectual or counter-productive, it would only add to the long history of flawed rating agency regulation. The regulatory use of ratings prompts investors to desire high ratings, weakening an incentive to get ratings right. Allowing more entities NRSRO status seems to have increased the competition among rating agencies to supply high ratings. New York’s pre-rating analysis fees failed to reward conservative rating agencies and the sharing of structured finance data has resulted in few, if any, unsolicited ratings.
[1] Strictly speaking, subprime mortgage loans are residential mortgage loans made to individuals who, because of their credit histories, are deemed not to be in the best, i.e., “prime,” category of credit risk. By 2005, the term “subprime mortgage” was conflated with “affordability” mortgage products not necessarily made to borrowers with poor credit histories. High loan-to-value loans and loans made simultaneously with a second mortgage lowered required down payments. Loans with introductory teaser interest rates lowered monthly payments, at least until the mortgage rate reset at a higher rate. Finally, “subprime” became conflated with the lack of underwriting rigor applied to some mortgage lending, for example “low doc” and “no doc” loans made with little or no verification of a borrower’s loan application. We refer to securitizations of subprime mortgage loans as “subprime bonds,” the terminology of the 2000s. But instead of referring to securitization of subprime bonds as “asset-backed security collateralized debt obligation” or “ABS CDOs,” the imprecise term used in the 2000s, we use the more descriptive “subprime CDOs.” [2] The rule for a bond to qualify for an ABX rating class was that it had to be rated at least that level by both S&P and Moody’s. Thus, when Moody’s took a more conservative rating approach to subprime bonds, some bonds in the BBB-, BBB, A, and AA ABX indices had higher ratings from S&P. [3] There is certainly room to quibble with this simple, hopefully intuitive, explanation of ABX pricing that among other things ignores present value. But we don’t have to argue that upfront payments exactly equal the market consensus of future losses to support our thesis. We just have to point out the massive difference in perceived credit quality when an ABX index clears with, say, a 50% upfront payment on securities that S&P rates investment grade.
[4] The penultimate AAA index had not yet been launched, so there were only 400 underlying bonds in all the ABX indices.
Douglas Lucas has over 30 years of experience in the financial industry and is a world-class expert in fixed income and structured product credit risk. He created Moody’s CLO rating methodology, including the WARF and diversity score measures of portfolio credit risk. At UBS, he was for 8 years consistently voted onto Institutional Investor’s Fixed-Income Research Team for CLO and CDO research.
Doug Lucas worked at Salomon Brothers, JPMorgan, UBS, and Moody’s. He created Moody’s first default study and its CLO rating approach in 1989 (including WARF and Diversity Score metrics). At UBS, 2000-08, he was voted #1 for CDO research in Institutional Investor’s poll. From 2009 to 2018, he managed Moody’s most-read publication, responsible for 18% of Moody’s total research readership. Some of his articles on CLOs, CDOs, and other structured finance products; default correlation and credit analysis; and rating agency regulation can be found at independent.academia.edu/DouglasLucas2.
Flanagan, Chris, et. al. 2004-09. Global ABS/CDO Weekly Market Snapshot. New York: J.P. Morgan Securities Inc. 2004-09.
Flanagan, Chris. 2007a. “Asset-Backed Securities.” US Fixed Income Markets Weekly. New York: J.P. Morgan Securities Inc., 23 February 2007.
Flanagan, Chris. 2007b. “Asset-Backed Securities.” US Fixed Income Markets Weekly. New York: J.P. Morgan Securities Inc., 2 March 2007.
Grassley, Charles E., John Kennedy, Roger F. Wicker, Sheldon Whitehouse. 2020. “Letter to SEC.” 3 February 2020. https://s.wsj.net/public/resources/documents/Senate%20SEC%20letter%202-3-2020.pdf?mod=article_inline
Kotecha, Mahesh; Sharon Ryan; Roy Weinberger and Michael DiGiacomo. 2012. “Proposed Reform of the Rating Agency Compensation Model.” The Journal of Structured Finance, 18 (1) 71-75.
Lucas, Douglas J. 2007a. “A Break in the Clouds?” CDO Insight. New York: UBS, 3 October 2007.
Lucas, Douglas J. 2008a. “How to Save the Rating Agencies.” CDO Insight. New York: UBS, 11 March 2008.
Lucas, Douglas J. 2008b. “How Bad Are ABS CDOs? How Bad Can CLOs Get?” UBS CDO Conference Handout, 24 March 2008.
Lucas, Douglas J. 2008c. “Rating the Rating Agencies on Subprime.” Mortgage Strategist. New York: UBS, 1 April 2008.
Lucas, Douglas J. 2008d. “Rating Agency Optimism and the ABX.” Mortgage Strategist. New York: UBS, 15 July 2008.
Lucas, Douglas J. 2008e. “Wishful Thinking on Securitized Product Ratings” CDO Insight. New York: UBS, 26 August 2008.
Lucas, Douglas J., Shumin Li, Rei Shinozuka and Charles Mladinich. 2007. “ABS CDO Collateral Losses Version 3.0.” CDO Insight. New York: UBS, 3 October 2007.
McGuire, Thomas J. 1995. Ratings in Regulation: A Petition to the Gorillas. New York: Moody’s Investors Service, June 1995.
Podkul, Cezary. 2019a. “SEC Fix for Conflicts of Interest at Credit-Ratings Firms Has Failed: Postcrisis plan enabled unsolicited ratings, meant to limit bond issuers’ ability to exert influence over ratings firms. Few, if any, such ratings have been published.” Wall Street Journal, 29 October 2019. https://www.wsj.com/articles/sec-fix-for-conflicts-of-interest-at-credit-ratings-firms-has-failed-11572341401?mod=article_inline
Podkul, Cezary. 2019b. “SEC Urged to End Ratings Firms’ Conflicted Business Model: An advisory panel is considering whether to recommend a new dynamic.” Wall Street Journal, 4 November 2019. https://www.wsj.com/articles/sec-urged-to-end-ratings-firms-conflicted-business-model-11572911797
Podkul, Cezary and Gunjan Banerjo. 2019. Inflated Bond Ratings Helped Spur the Financial Crisis. They’re Back: Credit-grading firms are giving out increasingly optimistic appraisals as they fight for market share in booming debt-securities markets.” Wall Street Journal, 7 August 2019. https://www.wsj.com/articles/inflated-bond-ratings-helped-spur-the-financial-crisis-theyre-back-11565194951
Podkul, Cezary. 2020a. “Lawmakers Push for Changes in Credit-Ratings Industry Bipartisan effort focuses on longstanding conflict of interest in industry and could lead to revamp of bond-ratings firms’ business models” Wall Street Journal, 5 February 2020. https://www.wsj.com/articles/lawmakers-push-for-changes-in-credit-ratings-industry-11580908541?mod=searchresults&page=1&pos=1
Podkul, Cezary. 2020b. “SEC Rethinks Approach to Conflicts Among Bond-Rating Firms: Agency is seeking industry input on how to combat rating inflation as 2010 fix falters” Wall Street Journal, 24 February 2020. https://www.wsj.com/articles/sec-rethinks-approach-to-conflicts-among-bond-rating-firms-11582589644?mod=searchresults&page=2&pos=2
Settlement Agreement. “Statement of Facts” beginning page 46, 2 February 2015. The case is United States v. McGraw-Hill Companies, Inc., and Standard & Poor’s Financial Services LLC, No. CV 13-00779- DOC filed in US District Court for the Central District of California on 4 February 2013. https://www.sec.gov/Archives/edgar/data/64040/000006404015000004/mhfi-ex1034x20141231xq4.htm
Sharif, Dipa. 2007. “Subprime Market Events.” Mortgage Strategist. New York: UBS, 25 September 2007.
US Department of Justice. 2013. “Department of Justice Sues Standard & Poor’s for Fraud in Rating Mortgage-Backed Securities in the Years Leading Up to the Financial Crisis: Complaint Alleges that S&P Lied About Its Objectivity and Independence and Issued Inflated Ratings for Certain Structured Debt Securities.” Press release 5 February 2013. https://www.justice.gov/opa/pr/department-justice-sues-standard-poor-s-fraud-rating-mortgage-backed-securities-years-leading
US Senate Permanent Subcommittee on Investigations. 2010. “Exhibits: Hearing on Wall Street and the Financial Crisis: The Role of Credit Rating Agencies.” Washington DC: 23 April 2010, page 149.
Warren, Elizabeth. 2019. “Letter to the SEC.” 26 September 2019. https://www.warren.senate.gov/imo/media/doc/2019.09.%2026%20Letter%20to%20SEC%20re%20inflated%20bond%20ratings.pdf
Zimmerman, Tom et al. 2007. “A Simple ABX Loss Projection Model.” Mortgage Strategist. New York: UBS, 27 February 2007.
Zimmerman, Tom et al. 2008. “ABX Implied Writedowns Similar to Shutdown Scenario.” Mortgage Strategist. New York: UBS, 25 September 2007.