Mortgage Ratings Blamed for Subprime Crisis Still Flawed, Ex-Insider Saysby
Model where issuers pay for ratings still flawed, Esaki says
Tossing out the most lenient ratings can improve incentives
The system for rating mortgage bonds and similar securities is still flawed a decade after loose grades helped trigger a financial crisis, and fixing it may be as simple as making a single rule, a former analyst for S&P said.
Howard Esaki, who retired from S&P Global Ratings in 2015, said he believes that for securities like mortgage bonds, issuers should be required to ignore the most lenient grade they receive. If a bank selling a bond seeks ratings from three firms, for example, it should have to use the two that are based on stricter standards, Esaki wrote in a forthcoming paper to be published by the Milken Institute.
“Efforts to address conflicts of interest haven’t gone far enough,” Esaki said in a recent phone call. Spokesmen for the three biggest ratings firms, S&P, Moody’s Investors Service and Fitch, declined to comment.
Esaki and his co-author, Larry White, a professor at NYU Stern School of Business, are trying to fix a long-standing problem in the ratings business: issuers pay for ratings, so graders have an incentive to go easy on their customers to win more business. The 2008 financial crisis came about in part because competition among ratings companies inflated grades on mortgage bonds, a Senate subcommittee report found in 2011. The failure of those securities contributed to more than $1.9 trillion of losses at financial firms worldwide during the global meltdown.
Moody’s last month settled with the U.S. Justice Department and 21 states for $864 million over its ratings on subprime mortgage securities in the run-up to the crisis. S&P settled similar claims with the U.S. in 2015 for $1.5 billion.
Even after the crisis, the U.S. Securities and Exchange Commission accused S&P of lowering its standards to win more revenue in commercial mortgage bonds, and forced it to sit out from part of the market for most of 2015. The ratings unit of S&P Global Inc has since been allowed to return. Securities like mortgage bonds contribute about a fifth of the annual revenue at Moody’s Investors Service, making it second only to corporate debt in importance to the firms, said Peter Appert, an equity analyst at Piper Jaffray.
Esaki has seen some of these conflicts first-hand. He spent 15 years as a top commercial mortgage-bond researcher for Morgan Stanley, during which time he witnessed the effect of weak ratings on the subprime mortgage market. In 2010, he joined S&P as head of research for global structured finance, a position he left in 2015.
For asset-backed securities, mortgage bonds, and other types of repackaged debt known as “structured finance” securities, changing the rules for winning new business would give bond raters less to gain by lowering their standards, Esaki and White write in their forthcoming paper.
Under their proposed rules, issuers would be allowed to ask as many ratings firms as they want for preliminary grades. The firms would get paid for doing that work, and would have to show the criteria they used to come to their conclusion. The most lenient grade would be tossed out, and the remaining firms would get a bigger fee.
Determining how lenient raters were would be based on how much “subordination” a deal has -- in other words, how much of the loans backing the bond have to default before investors in the safest securities suffer losses. That tends to be the factor that gets weakened when ratings firms are looking for business, Esaki and White said. In the 1990s, about 30 percent of mortgages in a bond would have to fail before the highest-rated bonds suffered credit losses, but by 2008 that figure fell to as low as 2 percent, they wrote.
The U.S market for asset-backed securities and mortgage bonds without government backing was more than $2.7 trillion U.S. at the end of 2016, according to data from the Securities Industry and Financial Markets Association.
The issuer would still choose the graders from those that remain, which means that ratings firms that are too strict won’t win business either, according to Esaki and White. All preliminary findings would be public.
That information could end up being useful for investors, who could find patterns that might make it clear which ratings firms are trying to bend criteria to win market share, said Kim Diamond, a founding employee at Kroll Bond Rating Agency. Diamond said she wonders if Esaki and White’s proposal would encourage ratings firms to focus on engineering criteria to win business instead of honestly assessing credits.
Fixing the incentives that go along with how bond graders are paid isn’t easy. The 2010 Dodd-Frank reform law required the SEC to come up with ways to rein in bond graders’ conflicts. The agency came up with rules to do so in 2014, but stopped short of requiring the firms to change their business models. Kara Stein, then a commissioner, said at the time that “our work is far from complete” in improving incentives for the firms.
Forcing investors to pay rather than issuers may eliminate some conflicts, but has problems too. For one thing, fund managers are reluctant to pay for grades that they are used to receiving for free, as Jules Kroll discovered when he looked at setting up a ratings provider based on investors footing the bill. He ended up instead creating a firm, Kroll Bond Rating Agency, where issuers pay.