I love zerohedge. Last week, they posted a story Chris Germain San Francisco whose headline blared “Kanjorski Admits There Is A Growing Bubble In Commercial Real Estate As S&P Observes CRE Losses Could Wipe Out Banking System.”
You’re forgiven if you didn’t know, but Kanjorski is a Congressman from the 11th District of Pennsylvania, and his website prominently features a picture of the beautiful Susquehanna River winding its way through an endless horizon of verdant green forests. Meanwhile, the S&P report never comes close to making such a cataclysmic, categorical observation about the U.S. banking system, and even if they had, who cares??
So what’s really going on in commercial real estate? Defaults are skyrocketing by almost every measure, but that’s hardly news. According to RealPoint’s monthly delinquency report, not only had delinquent, unpaid CMBS principal balances increased by 380%, but loss severity reached an all time high of 52%. For those of you following along at home, this means that many CMBS loans are worth no more than 48 cents on the dollar.
Excluding the Peter Cooper/Stuyvesant Town default from its estimated rate of delinquency growth, RealPoint predicts a CMBS default rate of roughly 8.5% by June 2010. Expressed in terms of delinquent, unpaid principal balance, this would be approximately 20 times higher than the low point set in March, 2007 – just as the market peaked. As CMBS delinquencies increase, specially serviced loans, as a percentage of overall CMBS outstanding, are skyrocketing:
This distress in the CMBS market serves as valuable, eyeball grabbing headlines for sites like zerohedge, but it’s more useful and instructive to compare the distress in CMBS distress to the relatively low default rates seen by other lenders. While Realpoint is dramatically predicting a CMBS default rate of 8.5% by June, Fannie Mae and Freddie Mac have consistently been clocking in with CRE default rates of just about one half of one percent, and insurance companies are reporting default rates of just about half that rate.
Clearly, the old CMBS model doesn’t work well, while the Fannie Mae/Freddie Mac model, which requires third party underwriters to take the first loss risk, is working much better. Having “skin in the game”, as it were, causes Fannie Mae DUS lenders and Feddie MAC correspondents to be much more cautious in underwriting their loans than a bank would be in making a loan that can instantly be turned into someone else’s problem via securitization. And often times, that “someone” isn’t all that smart. It’s all about pay for performance, and we should bring it back for real.
ShareThis
{ 12 comments… read them below or add one }
I'd further argue the bank balance sheet model works less well. Their delinquencies are higher, and historically banks have always been encouraged to make shorter-term construction loans (in the G.Depression era these were considered lower risk because they were shorter-term).
Take a look at MBA's somewhat useless chart where they put the 90+ day rate of banks next to 30+ day delinquencies on CMBS! http://www.housingwire.com/wp-content/uploads/2010/01/delinquency-and-vacancy.png
Good to hear from you again. That chart is totally useless, you're right. I almost published it with this post, but it's so irrelevant that I decided it would be better off buried in the MBA's Q3 delinquency report (where they too candidly admit to its uselessness).
However, aside from the fact that going back to the old balance sheet model is an even worse fix, I would dispute the broad notion that bank balance sheet lending works even less well than securitization.
You already know this, but for the sake of others reading this post and these comments, the CRE default data reported to the FDIC, the OCC and OFHEO by banks and thrifts is not nearly granular enough to be of any use, and here's why:
It's true that all of these loans are secured by commercial real estate, but many of them are actually business loans in which real estate happens to be just one small component of a much larger collateral package. The data simply does not distinguish between a $5 million loan for an office building in Topeka and a $5 million line of credit for a trucking company in Topeka secured by trucks, trailers, forklifts, accounts receivable, and one very small warehouse.
As a result, those bank and thrift default rates are misleading and overstated, and even even Elizabeth Warren, head of the congressional oversight panel, complained about the poor utility of the available data.
Not only that, this time around it was securitization that drove banks and thrifts back into construction loans – which they knew were more risky – only because they couldn't compete with the insanity in whole loans caused by CMBS.
As for the congressional oversight panel, if zerohedge were to post more thoughtfully and insightfully on its overly simplistic populist/conspiratorial theme, they might come to this conclusion: The congressional oversight panel was nothing but well orchestrated and well planned cover for another expensive and unnecessary government-led CRE bailout.
I wrote about this most recently in this post (includes a chart of CRE loan maturities by year, broken down by lender type), and again very early on in this oldy but goody, entitled appropriately enough: "Why Developers Are Desperate For the Dole".
This is not a political blog, but in my earlier life, I attended Senate hearings as a banker, and I know first-hand that they are better at creating theatre than producers on Broadway. I feel another post coming on, but this recent quote from Alan K. Simpson says it all:
“There isn’t a single sitting member of Congress — not one — that doesn’t know exactly where we’re headed,” Mr. Simpson said in a telephone interview [with the New York Times] “And to use the politics of fear and division and hate on each other — we are at a point right now where it doesn’t make a damn whether you’re a Democrat or a Republican if you’ve forgotten you’re an American.”
Where is the truth and who can we trust?
"Clearly, the old CMBS model doesn't work well, while the Fannie Mae/Freddie Mac model, which requires third party underwriters to take the first loss risk, is working much better. "
Huh? FN/FH are at risk for the Stuy Town debt. They lent most of the $3B senior debt which was CMBS-securitized. By current estimates they do not seem likely to recover 100 cents on the dollar on that loan (especially if the loan stays with the special servicer for a long time) unless they provide takeout financing with taxpayer money. It looks to me like they were cowboy lenders. There is a time lag between what they did in multifamily versus their other bad residential originations (5% of prime, etc), but I can not say for sure that they were much better than say Countrywide when it came to CRE lending.
Cowboy lenders? That's a bit too strong. They're not perfect, and they did Archstone too, and like most lenders, they have their gaze fixed firmly on the rear view, but their default rate is demonstrably lower.
I'll have to confirm this, but my understanding is that the Archstone and StuyTown deals were direct deals done outside of the usual DUS process. And I'm sure they were "advised" by some intermediary who would not get paid unless the deal closed…
You would have to be a student of the CMBS market to know that every deal done since the early part of last decade contained a bond backed solely by multifamily loans, typically in the low nine digits. The agencies bought these bonds as it fulfilled their residential lending charter. This robot bid incentivized multifamily lending to the point where any sane underwriting standards went out the window. Interest-only? Fine. DSCR<1 at issuance? Not a problem. British REIT buying NYC low-income housing? Back up the truck for loading. I would argue that the agencies were at the forefront of bad underwriting. The end result that multifamily loans now perform worse than loans on other property types where the agencies were not involved, with Stuy Town as the most egregious example amongst many (Bethany, Riverton Apts, Savoy Park, Lembi, etc.)
crabsofsteel (if you please), you're right, the Fannie Mae aspect of this is interesting. Combined with the explosion of the B piece market, which they may have encouraged with these bids, and the perception that multi-family was/is the safest of the four major food groups, Fannie Mae could have played a big role in blowing up CMBS *and*, ironically, multifamily. Buying huge amounts of AAA multi-family CMBS would obviously have been a great way to satisfy their volume requirements! (aka, "fulfill their public mission")
Care to elaborate or comment?? Any additional detail would be appreciated.
I think that regardless of the data put out by the government, which everyone knows is manipulated considerably, prices across all industries, including commercial real estate, will continue falling because of the reduced demand that accompanies a deflationary credit collapse. And the policy response to this has been and will continue to be money printing by Bernanke and the Fed, which are very bullish for gold. I actually came across a good article on gold stocks that discusses how certain gold companies should continue to benefit from the government's policies:
Gold Stocks (GDX) Fall – Premier Gold Mines Bucks Trend
As such, I think that while the government can keep debasing the dollar, but it can't increase the standard of living in this country over the longer term merely by printing money, as no wealth is created in doing so, therefore prices will continue to fall for all sorts of consumer goods and services.
What more can I tell you than what I already stated? What more do you want to know? It was more Freddie than Fannie pumping worthless loans out the door. They both hold tens of billions of these multifamily-backed loans, not all of which are going to be money good. They are incentivized to refi these loans with public money, and my bet is that's exactly what they will do.
Well, my understanding is that they always bought the AAA piece. Did they cover the whole AAA offer, for example, on any one deal, and if so, how often did they do that?
I know they bought a big chunk of Archstone, among others like it, and they did that direct with Lehman. What was their annual volume in CMBS purchases? If they were just one of numerous bidders and not buying in great size, then it's no big deal.
If not, then I am wondering if it's worthy of a more detailed post.
David – there are some very smart people who believe in deflation. I am not all that smart, so maybe that's why I don't believe we're going to see protracted deflation.
I wrote a post on here on deflation fears, outlining the reasons why deflation fears are overstated. The comparisons to Japan are just flat out false. There's much more in the comments.
Cheers, RW
David, sorry the link was bad. The post on deflation is here
70% of every CMBS deal done since 2006 was "super-senior" AAA with 30% of protection underneath it (including support from other AAA-rateed classes … let's not get into how there could be different flavors of AAA if it's supposed to always mean money good). The agencies took down most of the multifamily portion of that 70%. There were 60-70 fixed-rate conduit deals done in 2006-7, so if you figure an average bond size of $200MM, they bought $26BB of these bonds. I'm going to take a peek at their 10Qs to see if they make any mention of them.
CrabsOfSteel is correct, but the Agencys’ investments in CMBS are double his guesstimate.. There were 120 Conduit deals in ’06 & ’07. The A1A tranches that CoS is referring to were directed tranches with the majority of the multifamily collateral (not all of it, but most) serving just that tranche, which was pre-sold to Freddie and Fannie. The ’06 & ’07 balance for the A1A tranches was $55 billion, but the total outstanding balance (they started this practice in CMBS back in 2003) is $75 billion.
On average, in ’06 & ’07, the A1A tranche was 15.4% of the total deal. So, to CoS’s point, the super senior was 70% of the total deal, A1A is about 10% of that super senior class of bonds (15.4% * 70%). Although it is not directly correlated, if you line up, say, the worst 10 CMBS deals of 2007 and put it next to a list of the 2007 deals with the largest A1A tranche sizes – there is some overlap – C30 (29% A1A tranche, this is the largest Stuy Town exposure), IQ13 (29%), LBCMT 2007-C3 (28%), CSMC 2007-C2 (the winner at 46%), etc.
I came back across your article all these months later because of the article out this week that implies that Fannie & Freddie are the stalwarts in the CRE multifamily space because of their low delinquencies in their portfolio, but the article just ignores their CMBS investments altogether. I look pretty perfect too, you just have to ignore all the mistakes I’ve made, catching me at just the right angle, under just the right light, and from a pretty good distance away!