Viewpoint: S&P Clubs CMBS Markets
It is widely agreed that woefully flawed rating methodologies contributed materially to the ongoing troubles in the financial system. A quick recap: whether lulled by the demon of home price appreciation (always abbreviated HPA, as basic as USA) or distracted by the pursuit of profits, market share and such, the purveyors of bond ratings dramatically underestimated the risk of default in subprime and other risky residential mortgages. On top of that, the raters blessed the structuring alchemy of CDOs (collateralized debt obligations, in which pools of private MBS, ABS, leveraged corporate loans and derivatives of the same are the source of the underlying cash flows) with excessively optimistic ratings. The enormous appetite for CDOs (offering extra yield per any given rating) effectively sidelined the real “watchmen” of mortgage credit, the investors who specialized in buying the credit support tranches of ABS and private MBS. These investors, whose more rigorous investment process regularly second-guessed the ratings, were simply outnumbered and outbid by CDO managers’ wholesale demand for low rated paper that could be restructured into presumably investment grade paper. (I’ll skip the incestuous linkage between AAA-ratings and bond insurers wrapping dreck to AAA. That’s a story all by itself.) We know what happened, but if you’d like to graze this topic a bit longer, here’s a fascinating S&P insider’s commentary re: one strand in the noose, how mortgage ratings were so far off the mark: Frank Raiter, head of RMBS rating until 2005, commenting last week in an NPR interview and more extensively in testimony to Congress late last year. (The NPR bit was clipped from last weekend’s “The Watchmen” installment of This American Life. That broader inquiry is worth a listen as well.) Now S&P wants to uproot a green shoot Given the calumny subsequently heaped on them, the raters have made some highly-visible attempts to reestablish investor and public confidence in their products. The latest of these efforts comes from Standard & Poors Ratings Services, arguably the largest and best known of the ratings companies (investment guidelines typically specify S&P and Moody’s or require two nationally recognized raters -- the outcome is that S&P and Moody’s historically have roughly split a 80% lock on the bond ratings business). On May 26, S&P published a request for comments (RFC) on proposed changes to its ratings methodology for conduit or “conduit/fusion” CMBS. These deals are backed by loans to a diverse set of borrowers across a number of real estate sectors and constitute about 85% of the outstanding CMBS market. Comments were due yesterday, June 9. (Kind of a rush job, I’d say, especially given the Fed’s efforts at roughly the same point in time to finalize TALF legacy CMBS terms.) At the time it unveiled the new methodology, S&P projected the ratings of classes all the way up through the most senior tranches of existing deals would be affected. On May 26, their preliminary finds indicated about 25% of 2005 vintage super-duper seniors, 60% of 2006 vintage and 90% of 2007 vintage could be downgraded. These were preliminary findings, because they were still evaluating the universe of conduit/fusion deals. S&P promised to follow-up. The follow-up was published last Thursday, June 4, forecasting yet more dire outcomes: if adopted, S&P anticipates the changes would result, for example, in downgrades from AAA of 50% of 2005 super-dupers, 85% of 2006, and 95% of 2007s. Potential downgrades to AM, AJ and lower-rated (but investment grade) tranches were estimated for the first time, ranging from 90% and 95% in a few cases to 100%. Excuse me, but these are material differences in expectations -- especially given the short comment period! Rally short-circuited The RFC announcement derailed the rally in CMBS. The addition of CMBS in TALF had triggered a dramatic rally in the sector: secondary market spreads on AAA tranches tightened on the order of 400 to 900bp (depending on average life). With S&P threatening wholesale downgrades, spreads backpedaled a couple hundred basis points. (That’s a lot in price terms, kids. Try to imagine the pain of carrying that position at market value.) If adopted, the new methodology could eviscerate the Fed’s efforts to get commercial real estate mortgage lending going again (not going into it here, but the ability of borrowers to refinance their balloon mortgages at the appropriate time is one of the big risks facing existing CMBS investors). That is, under current terms, TALF loans will be available on legacy CMBS currently rated AAA. If adopted the TALF universe could be a fragment of what the market had been assuming. S&P says all kinds of right things about the good intentions behind the proposed changes. They’re part of its new look: “a broad series of measures … aimed at augmenting our independence, strengthening the rating process and increasing transparency.” (No mention of giving investors their money’s worth.) The goal is “establishment of an ‘AAA’ credit enhancement level that is sufficient to enable tranches rated at that level to withstand market conditions commensurate with an extreme economic downturn without defaulting.” Sadly, I’m not a CMBS analyst or I could do this on my own. (The last time I was asked to write about this asset class was almost fifteen years ago, when the CMBS market was just beginning jell). But I do have in hand a couple of analyses by experienced analysts, whose efforts have always struck me as balanced. Let’s hear what they say. Analysts at J.P. Morgan object. They state “there is not sufficient evidence to warrant widespread downgrades of recent vintage super-senior ‘AAA’ bonds.” They are speaking, I believe, from their daily task of monitoring the credit performance of the universe of CMBS deals, talking to servicers, special servicers, investors (and CMBS investors tend to be quite knowledgeable, able to evaluate the collateral under their bonds on a loan-by-loan basis. This is a different puppy from securitized consumer and mortgage loans). What bothers them most is exactly what would bother me: “… little or no evidence was seriously presented and seemingly arbitrary standards and thresholds were put forth.” For example, the new methodology applies defined stresses to rents, by property type. Based on the resultant losses, ‘AAA’ subordination levels are defined. The new methodology includes a tidy little table showing worst historical rent declines (worst 3-year cumulative national and average worst 3-year by MSA) alongside the proposed ‘AAA’ stresses. For example, for offices, where the worst 3-year national decline in rentals was 20%, worst average 3-year decline by MSA was 22%, the proposed ‘AAA’ stress is 29%. For Retail (think malls, etc.), respective stress are 2%, 12% and 24%. What distresses the J.P. Morgan analysts (and does me too), is that S&P doesn’t explain is how they how they get from 20%/22% historical worst cases to a 29% decline rent stress in on office properties, while they bounce from 2%/12% historical declines to a 24% stress in retail properties. They just note the new ‘AAA’ stresses are incrementally higher, based on S&Ps “assessment of the type of declines that might occur in a period of extreme stress.” I would say, “What assessment? Show me the beef!” The J.P. Morgan analysts had that opportunity, to which the S&P analysts replied that the stressed cash flow declines were backed out from a 20% subordination level on a prototypical pool. For J.P. Morgan analysts this begs the question, “Why is a 20% enhancement level deemed to be the correct starting point?” Their concern is, if the methodology is based on an arbitrary initial assumption, what prevents S&P from shifting that assumption again? What’s missing, says J.P. Morgan, is transparency and detail. I say, for a purveyor of ratings that is trying to reposition its rating process for pooled assets as trustworthy and reliable, this is more of the same-old same-old. In effect the lack of transparency and detail says, “We’re S&P. That means we’re explicitly or implicitly written into virtually every regulatory investment rule, bond fund prospectus, and financial institution investment guideline out there, so like it or lump it!” More distrust for the update In the June 4 update, “The Potential Ratings Impact of Proposed Methodology Changes on U.S. CMBS”, the projected downgrades are surprisingly more extensive than originally estimated in the May 26 RFC. Also a surprise, the update indicates that the methodology differentiate among ‘AAA’ classes with different average lives. The underlying intuition behind such a distinction would be that shorter bonds are paid down earlier, limiting their exposure to losses in the underlying collateral. However, this treatment was not indicated in the May 26 RFC describing the new methodology. More transparency (NOT). This is a heck of a surprise to spring on market participants so close to the end of the comment period, and with no supporting analysis either. Bear in mind as well that this is a departure from traditional practice. Until this point, a block of senior, ‘AAA’ cash flows would be created subject to the rating companies’ requirements for subordination (along with other forms of credit support). Say 80% senior -- 20% subordinate. Then the ‘AAA’ chunk is time tranched into, for example 3-, 5-, 7- and 10-year bonds. (There is some more complexity in CMBS deals, but this is the basic idea of senior-sub structures.) In other words, in credit terms, the 3- and 10-year bonds are created equal. So the rules are changed after the fact. Changing the rules in the current economic environment can be defended. However, the new methodology does not always produce consistent results. S&P provides a number of tables detailing the percentage of potential downgrades with the universe of 402 deals it devaluated using the new methodology. Inexplicably, 5-year ‘AAA’ tranches in some vintages fare worse than 7-year ‘AAAs.’ Among 2008 vintages, 40% of 5-year ‘AAAs’ would be downgraded to a weighted average rating of ‘A+,’ while 35% of 7-years would drop to a weighted-average ‘A+.’ Something the reverse happens in the 2006 vintage: 10% of 5-years ‘AAAs’ would be downgraded to a weighted-average ‘A+,’ but 25% of 7-years would drop to a weighted average rating of ‘AA,’ a notch better than the 5-years. Still another conundrum -- in the 2005 vintage deals, 90% of AM and 100% of AM and AJ tranches, originally rated ‘AAA’, would be downgraded, but just 95% of A-rated and 90% of BBB-rated tranches would be downgraded. Market impact Do the proposed changes in rating methodology derail legacy TALF? There is considerable speculation that the Fed can simply change its criteria to read “originally rated ‘AAA’.” Or it could change its rating requirements to accept the middle or majority of three ratings, a course it might prefer in general now that it is moving to widen the number of recognized NRSROs. If S&P’s new methodology does disqualify a vast majority of outstanding ‘AAA’ paper, J.P. Morgan analysts expects secondary spreads can still tighten because “investors judge them based on the actual cash flows, subordination levels and relative value versus alternative investments.” So long as they are still rated investment grade, most investors are under little pressure to sell them. Also, “as new issue TALF takes hold,” firming spreads in the new issue market, secondary bonds should benefit as well. The new methodology would have more severe impact on 2007 and 2008 vintage AJs (junior bonds cut from ‘AAA’ rated mezzanine cash flows) and bonds originally-rated AA. On average these bonds would be downgraded to below-investment great and subject to forced selling by investors with regulatory capital requirements or “investment grade only” investment guidelines. As Citigroup analysts bluntly say of the downgrade risk to 2007 and 2008 AJs, “We do not think the market is fully appreciating what S&P intends to do, as many investors will have to sell their bonds, potentially flooding the market.” Great. Just as markets were beginning to firm, S&P yells fire sale. Investors take losses if they sell, and those who mark to market (money managers, mutual funds, hedge funds, etc.) take losses even if they don’t sell. And crushing the market for credit tranches does nothing to improve prospects for issuers of new, TALF or not, deals hoping to place subordinate cash flows. And so on. The old saying used to be, “Pride goeth before a fall.” But now, for these arbitrary raters, “pride goeth before hubris.” Ratings -- for residential mortgage securities or commercial mortgage securities -- should not be based on black boxes and circular assumptions. S&P, show these market participants the beef! Editor’s note: Linda Lowell is a 20-year-plus veteran of MBS and ABS research at a handful of Wall Street firms. She is currently principal of OffStreet Research LLC, and is a columnist for Market News International and HousingWire Magazine.