You will have your own favorite villain of the ongoing financial-and-now-economic crisis. Greedy Wall Street bankers? Not a bad choice. Shady mortgages brokers? No argument there. Unscrupulous speculators? Sure! All of ‘em helped make the mess. But for my money, one group stands out as the low-downdest skunk of all: the rating agencies. Not only did they play a central role in pumping the market full of doomed mortgage paper in the first place; since then, they’ve helped make a bad situation worse with their ham-handed, irrational ratings downgrades. If it weren’t for the rating agencies, a lot of the most severe pain of the credit mess might have been avoided.
The rating agencies have already demonstrated that, as regards subprime mortgage credit, they don’t know what they’re talking about. Even so, they continue to provide misguided, incompetent analysis. Only this time, rather than underestimate the extent of credit problems, they’re overestimating it.
To see just how hapless the agencies are, take a look at the “reasoning” Moody’s offered to justify one of its more recent idiotic pronouncements: the placement of financial guarantors Ambac and MBIA on watch for potential downgrade (Disclosure: Ambac and MBIA are both near and dear to my heart. The fund I manage owns both.)
Anyway, on September 18, Moody’s Structured Finance Group issued a report updating its “base case” loss projections, last issued in January, for 2006-vintage subprime loans. The original loss estimate range of 14% to 18% went to 22%, while the stress case went to 30%.
Why the increase? Moody’s provided a number of reasons.
Reason 1: “The rate of new delinquencies is currently exceeding our initial assessment.” I don’t know what Moody’s' expectations used to be (and it won’t say) but back in January when the agency first put out its loss estimate, the percentage of current loans going into the 30-day delinquent bucket was high and rising. Since then, as you can see in Chart 1, new delinquencies have gone down.
(right click 'view image' to enlarge)
Reason 2: “Moody’s has also revised upward its assumption on average loss severity from 45% to 55%.” No argument there. We boosted our severity assumption, to 55% from 45%, several months ago, as well.
Reason 3: There are no more reasons! That’s it. I kid you not.
It’s hard to convey just how upside-down-and-backwards Moody’s' analytical subprime credit work is. Fact: Since Moody’s put out its January report on the 2006 vintage, the inflow rate of new problem loans has slowed, not increased. Fact: Roll rates from early-stage delinquency buckets to later-stage buckets have fallen, not risen. Fact: The rate at which late-stage delinquent loans become current again has increased. The only number that has gotten worse is loan-loss severity. But in every other material metric, things are better now than they were in January. And Moody’s raises its loan loss assumptions?
It’s crazy. So crazy, in fact, that Moody’s' new loss estimates are a virtual mathematical impossibility. As it happens, the agency doesn’t supply much detail surrounding its assumptions. But it gives enough that one can reverse engineer them to see what has to happen for its expected losses to occur.
It’s not that hard to do. To start, think of the subprime class of 2006, and put its loan losses into three mental piles:
- Realized losses to date
- Losses from loans that are already 60 days or more past due. (We and Moody’s call these “pipeline” losses.)
- Losses from loans that are still current.
OK, now take a look at Chart 2. It shows what’s happened to the $600 billion of subprime loans originated in 2006. So far, $240 billion have been paid down, $24 billion of principal has been written off, and $336 billion are still outstanding.
Now to Chart 3. At first glance, it will look like a plate of spaghetti, but it’s actually a pretty straightforward summary of how Moody’s expects pipeline losses to develop. Using the agency’s own roll-rate and severity assumptions, the $115 billion in loans past due by 60 days or more will generate $38 billion in losses. Realized losses, recall, add up to $24 billion.
Our story this far: Realized losses, $24 billion, add up to 4% of losses. Another $38 billion in pipeline losses account for 6%. Combined loss rate of the two piles: 10%. Moody’s' new loss estimate, recall, is 22%. So the balance of losses will have to come from the $221 billion of loans that are still performing.
From Moody’s' standpoint, this presents two problems: 1) for that 22% loss rate to happen, still-performing loans will have to default at a staggering rate, and 2) that 30% stress case isn’t just highly unlikely, it’s mathematically impossible, given Moody’s' own assumptions!
Chart 4 tells the tale. For Moody’s' 22% loss rate to happen, the performing portfolio would need to kick in $70 billion in losses. At 55% severity, that implies $127 billion of still-performing balances would have to go delinquent and roll all the way through to liquidation. Let’s assume 75% of the performing loan balances that default go to liquidation. That would mean that fully 77% of currently performing loan balances have to go delinquent. (And that assumes a 75% roll rate into liquidation; Moody’s, remember, assumes only 40% of loans 60 and 90 days late go to liquidation)
A 77% delinquency rate isn’t just highly unlikely. It’s literally unbelievable. Remember, the $221 billion of 2006 balances that are still current represent borrowers who’ve been making payments for two years. There aren’t any speculators or fraudsters in the group. Those have long since been shaken out, as have the weakest credits. And yet Moody’s seems to believe that this group will default at a rate almost three times the 27% default rate experienced thus far. That doesn’t begin to make a shred of sense. Doesn’t anybody at Moody’s provide a simple reality check?
Worse, though, is the modelling that appears to have generated the agency’s stress case loss of 30%. At a minimum, it is reckless; at worst, possibly criminally irresponsible.
Chart 5 shows what would have to have to happen to the 2006 performing subprime mortgage loans to get to a 30% loss, even with assumed severity bumped up to 65%. You read that right: Over 100% of the currently performing loans would have to default.
The SEC Needs to Act
Moody’s' methodology and numbers are clearly nonsensical. Yet its judgments aren’t just academic; rather, they can (and do) have real-world repercussions that can be devastating. A ratings downgrade, for instance, can put a company in violation of its debt covenants. Or it can force a company to put up more (sometimes a lot more) collateral with its trading counterparties. So a downgrade can put a company under severe financial stress. If the rating agencies were rational, and knew what they were doing, that wouldn’t be a problem. But they don’t know what they’re doing. Instead, their fecklessness has taken a bad situation and made it even worse.
Both major rating agencies have performed abysmally throughout the subprime mortgage debacle. So abysmally, in fact, that I believe the SEC should consider eliminating the agencies’ privileged status as a Nationally Recognized Statistical Rating Agency (NRSRO).
As is evident with its moves related to the guarantors, Moody’s' irresponsibility isn’t just continuing, it’s getting worse. Moody’s needs to be stopped!
There will be many regulatory reforms in the coming years. One of the most vital will be the elimination of the SEC-sanctioned NRSRO designation. In addition, any agreement or contract with terms that take into account or is based on agency ratings should be amended. No anointed organizations determines the suitability of equity investment, none should pass judgment on fixed-income instruments, either. This should be especially obvious given how poorly the rating agencies have lately done their job.
Tom Brown is head of Bankstocks.com
Disclosure: See declared holdings for fund managed by author (including ABK and MBIA): Second Curve Capital