The latest Big Apple trophy being coveted by oil-rich sovereign wealth funds is the landmark Chrysler Building.
Sources say the super-rich Abu Dhabi Investment Council is negotiating an $800 million deal for a 75 percent stake in the Art Deco treasure that has defined the Midtown skyline since 1930.
The Chrysler assets would be purchased from TMW - the German arm of an Atlanta-based investment fund that's been eager to cash out of its Chrysler stake.
Spurred by the weak dollar and the strong euro, European travelers to the U.S. have been lapping up everything from Gap boxers to iPhones to luxury condos in Palm Beach. Now a top Italian real estate investor has nabbed a crown piece of New York property, a sale that echoes the Japanese purchase of Rockefeller Center in 1989. Valter Mainetti has confirmed to TIME that his company, the Sorgente Group, has acquired a majority share of Manhattan's historic Flatiron building
Tanta this may June 08 release might interst you interest you:
The Office of Thrift Supervision (OTS) announced today the execution of a Supervisory Agreement with AIG Federal Savings Bank (AIG FSB) for its failure to manage and control in a safe and sound manner the loan origination services outsourced to its affiliate, Wilmington Finance, Inc. (WFI). The Agreement addresses loan origination activities by WFI that had a negative financial impact on certain borrowers of AIG FSB.
On page 4: there seems to be an error in the Overall Mortgage Portfolio table: the entry for Oct 2007 (in $ millions) is 3,21,902. However the entries for Nov 2007 onward are 10 times larger. I suspect a missing digit, probably a 7.
They are foreclosing more on prime loans which are seriously delinquent (pg 13) than other types.
So, at this point 1,500,000 homes are in danger of default... I assume this does not include those that have already defaulted. REO's galore.... half way through sub-prime, alt-a and option arms just beginning, prices still spiraling down, where it stops nobody knows. Not to worry, over by 2013.
Nice effort overall - but the one thing that jumped out at me was their strange definition of "Alt-A".
Alt-A, as definded by the OCC: FICO scores 620-659
That's completely misleading and inappropriate. You could have a borrower with a 800 FICO that is ALT-A because the loan was underwritten using ALTERNATIVE documentation.
I assume this does not include those that have already defaulted.
The numbers in this report are as of 3/31/08. Some houses that appear to be at risk of foreclosure may have already happened. Some New Foreclosures may have happened that are not yet reflected in this report.
Tanta, anything jump out at you from this report? Something we didn't really know already? Or was it more of same problem just more thoroughly covered?
I regard the initial numbers as meaningless but the trends will be what I pay attention too. As long as they keep a sound methodolgy and account for their reduction in "others" then this could show us the 'real' vs forecast problems with those loans that have not blown up yet.
Loss mitigation vs delinquency will be interesting.
byzantine_ruins writes:
Good to see you posting and active, Tanta.
I agree with this, although I always hesitate to say anything that might imply you're not in good health and moreover in a fragile condition. Not only that, I'm consumed with a impudent desire to know what the heck you've been up to. Sorry for the presumption but,...
"90% of loans are securitized either through the GSEs or private label issuers.." Let me guess, 80% GSEs and 10% private label issuers?
You have no idea what`s coming guys. It might have taken a long time to see what an idiotic idea (GSEs) can do to a capitalistic sistem, but the results will be horrific.
"The OCC uses a data vendor
to aggregate, validate, store, and generate reports, but
retains ownership and control of the data..."- Anyone know who the vendor is?
jkiss writes:
".... half way through sub-prime, alt-a and option arms just beginning, prices still spiraling down, where it stops nobody knows. Not to worry, over by 2013."
A 5 year descent to a smooth Japanese style landing might be the good news.
A deflationary stock and credit collapse ending financial life as we know it in 2008 or 09 could be the bad news.
To echo Schnaps' comment, I too found this stunning:
The OCC Mortgage Metrics Report uses standardized
definitions for three categories of mortgage
creditworthiness: prime, Alt-A, and subprime. These
are defined using ranges of FICO credit scores at the
time of origination, as follows: prime 660 and above;
Alt-A 620 to 659; and subprime below 620.
There is certainly value in standardizing definitions, but this Alt-A definition completely misses what's meant by the term "Alt-A." It's like saying we're going to track all miles travelled by passengers in cars, boats and planes, and we define boats as vehicles that ride on rails over land and make stops at stations. Yeah, you've standardized a definition but you're going to accurately report the passenger miles traveled by what any non-idiot thinks of as a boat.
Another, arguably lesser, evil of this definition is that it (once again) fetishizes FICO scores.
If you had to choose one dimension, just one dimension, with which to segment the entire portfolio into something more meaningful than stupid FICO buckets, I would have gone with DTI, something like Documented Low, Documented High, and Unknown/Unverified. That's cheating a little b/c I'm insisting on the documented vs. undocumented distinction, but that's much better.
Of if you wanted to act like you had a credit thought in your brain, instead of relying on the FICO buckets, you might have defined the three buckets by LTV. There's a number that every loan has and much more readily correlates to what you'd think of as a "prime" vs. "nonprime" kind of risk.
All that said, kudos, b/c an effort at data collection is useful. Although I'd note that with a subscription to LoanPerformance you get incredibly granular data already.
Okay, now that I've skimmed the rest of the report, the stupid Alt-A definition is really pissing me off, b/c it makes the rest of the report almost completely useless.
The report tells me "Alt-A" is 9% of serviced mortgages (in this portfolio... which by the way since it excludes CFC is missing far and away the largest servicing book of truly Alt-A loans, but whatever).
Then it tells me that Alt-A loans were seriously DQ at a rate 6 times higher than Prime loans were. But b/c of their stupid definition, we're learning anything useful here. Yeah, people in the relatively narrow FICO band of 620-660 are much more likely to go 2 months behind in a mortgage payment than the MUCH LARGER population over 660. I know that. The way you get a 650 FICO is by relatively frequently paying late. If you look at the FICO distribution of the US population with FICO scores, it is heavily skewed to the high end. (I haven't seen the data in a while, but I believe the average FICO is in the 730s, and the mode is higher than that).
This reporting could generate useful information -- not just useful to us lenders or investors but, you know, useful to the GD regulators(!!) if they'd instead reported on something like documentation level or LTV. Those are the two dimensions where lending got completely out of hand and where some standardized data collection and analysis could actually help make the world safe again.
Instead what we get is continued fetishizing of FICOs which will lead to really insipid regs about making loans to borrowers w/ FICOs from X to Y. And yes, I know that the regulatory definitions for "subprime" mortgages requiring 100% risk-weighting already fetishizes FICO, but just b/c we've been lazy in the past -- with disastrous results -- is no reason to insist on continuing to be lazy in the future.
Thus endeth my rant. I'll be curious if someone actually finds something useful and interesting in here.
Interesting data, but the fact that it is taken from nine of the large banks raises questions about how representative the data is. I could not help but notice that Countrywide is not on the list. Covering the largest banks makes sense from the perspective of an agency with limited resources (which I understand all too well), but the largest banks in the past kept some of the most transparent assets for their portfolios while shoveling the questionable assets to other entities.
These statistics are useful (any information is better than none), but I have to wonder what the performance of the other 60% of mortgages looks like in comparison.
Shortcomings aside, I'm just glad to see public data on delinquencies, finally. The Mortgage Bankers Association charged hundreds of dollars for its reports. So much of the market activity is subsidized with federal money but regulators conveniently never produced data about how loans performed. Woulda been helpful, like, 10 years ago.
Price Stout - I find it amazing and appalling that someone doesn't know the FICOs on 20% of the mortgages. Kind of makes it difficult to estimate risk on the portfolio, doesn't it? That, more than anything else, is what is notable about this report. One immediately suspects that these banks have not a clue as to what their risks are.
Also, I have always considered Alt-A to be prime borrowers with riskier loan types, not borrowers with lower FICOs.
One immediately suspects that these banks have not a clue as to what their risks are.
While I agree with the general point you are trying to make, MoM, here's the thing: The credit risk may in some cases not be theirs - "they" being the servicer. Stay with me here.
Strictly speaking, in order to service a loan, one doesn't need to know what the FICO was at the time of origination.
And if your solitary stake in a given loan is the right to perform its servicing, then why pay to store a data element you don't need? Every byte adds up, MoM. Which is exactly the kind of myopic crapthink you might ponder and/or implement when your forte as a servicer is largely based on worshipping at the altar of "efficient operations" and "economies of scale".
The same altar that apparently only Tanta, the Shnapster, and maybe a handful of other CR-dwelling heretics such as yourself are foolish enough to question.
Shnaps - well, one reason I'd want to know the FICO is that it costs more to service subprime loans!!!
Just trying to imagine how this is the best info the OCC could get is making my head hurt. No geographic distributions? Surely they know where they are?
MaxedOut - the 20% surprised me, yes, but it surprised me on the low side. Most banks are much worse at data capture and retrieval than you could possibly imagine. That nine banks can report the FICO score on 80% of the mortgages they service is actually pretty impressive to me. Note the language in the definition section about "best efforts" data collection. That suggests there was a fair amount of effort involved in collecting this data, much more than simply pressing a button and producing a flat file for the OCC.
You are correct in your understanding of Alt-A. That is exactly the point the issue that set me off. The OCC is measuring something w/ that 620-660 bucket, but it ain't what anyone in the industry would call Alt-A. They might call it "near prime."
Here is one way to spread the risk of CRE:
The latest Big Apple trophy being coveted by oil-rich sovereign wealth funds is the landmark Chrysler Building.
Sources say the super-rich Abu Dhabi Investment Council is negotiating an $800 million deal for a 75 percent stake in the Art Deco treasure that has defined the Midtown skyline since 1930.
The Chrysler assets would be purchased from TMW - the German arm of an Atlanta-based investment fund that's been eager to cash out of its Chrysler stake.
wonder how much the statute of liberty would get?
Dont know. But here is another NY landmark:
Spurred by the weak dollar and the strong euro, European travelers to the U.S. have been lapping up everything from Gap boxers to iPhones to luxury condos in Palm Beach. Now a top Italian real estate investor has nabbed a crown piece of New York property, a sale that echoes the Japanese purchase of Rockefeller Center in 1989. Valter Mainetti has confirmed to TIME that his company, the Sorgente Group, has acquired a majority share of Manhattan's historic Flatiron building
Tanta this may June 08 release might interst you interest you:
The Office of Thrift Supervision (OTS) announced today the execution of a Supervisory Agreement with AIG Federal Savings Bank (AIG FSB) for its failure to manage and control in a safe and sound manner the loan origination services outsourced to its affiliate, Wilmington Finance, Inc. (WFI). The Agreement addresses loan origination activities by WFI that had a negative financial impact on certain borrowers of AIG FSB.
OTS Executes Supervisory Agreement with AIG FSB
Tracking financials at financials:
For April, Downey reported more than 13% of its $13 billion in loans were in default. In March it reported 12% of its loans in default.
Lets see if the trend breaks for May?
Barley: "Now a top Italian real estate investor... Valter Mainetti... his company, the Sorgente Group"
?? never heard of him nor of his company.
BTW: last round of 'top Italian real estate investors' are either broke by now or under criminal trial or both.
Thank you for the nice meaty data.
On page 4: there seems to be an error in the Overall Mortgage Portfolio table: the entry for Oct 2007 (in $ millions) is 3,21,902. However the entries for Nov 2007 onward are 10 times larger. I suspect a missing digit, probably a 7.
They are foreclosing more on prime loans which are seriously delinquent (pg 13) than other types.
sidd
So, at this point 1,500,000 homes are in danger of default... I assume this does not include those that have already defaulted. REO's galore.... half way through sub-prime, alt-a and option arms just beginning, prices still spiraling down, where it stops nobody knows. Not to worry, over by 2013.
Nice effort overall - but the one thing that jumped out at me was their strange definition of "Alt-A".
Alt-A, as definded by the OCC: FICO scores 620-659
That's completely misleading and inappropriate. You could have a borrower with a 800 FICO that is ALT-A because the loan was underwritten using ALTERNATIVE documentation.
I assume this does not include those that have already defaulted.
The numbers in this report are as of 3/31/08. Some houses that appear to be at risk of foreclosure may have already happened. Some New Foreclosures may have happened that are not yet reflected in this report.
Good to see you posting and active, Tanta.
Tanta, anything jump out at you from this report? Something we didn't really know already? Or was it more of same problem just more thoroughly covered?
And thanks for reading it so we don't have to...
Thanks Tanta.
Another stat to follow but it could be useful.
I regard the initial numbers as meaningless but the trends will be what I pay attention too. As long as they keep a sound methodolgy and account for their reduction in "others" then this could show us the 'real' vs forecast problems with those loans that have not blown up yet.
Loss mitigation vs delinquency will be interesting.
byzantine_ruins writes:
Good to see you posting and active, Tanta.
I agree with this, although I always hesitate to say anything that might imply you're not in good health and moreover in a fragile condition. Not only that, I'm consumed with a impudent desire to know what the heck you've been up to. Sorry for the presumption but,...
OT- Anyone have a line on whether the LEH financing is truly going to get done tomorrow?
Lehman Could Be In Hot Water - Forbes.com
"90% of loans are securitized either through the GSEs or private label issuers.." Let me guess, 80% GSEs and 10% private label issuers?
You have no idea what`s coming guys. It might have taken a long time to see what an idiotic idea (GSEs) can do to a capitalistic sistem, but the results will be horrific.
sdtfs -
She may have been down for a spell, but I something tells me that she's back to blogging brilliance and shitting excellence.
"The OCC uses a data vendor
to aggregate, validate, store, and generate reports, but
retains ownership and control of the data..."- Anyone know who the vendor is?
jkiss writes:
".... half way through sub-prime, alt-a and option arms just beginning, prices still spiraling down, where it stops nobody knows. Not to worry, over by 2013."
A 5 year descent to a smooth Japanese style landing might be the good news.
A deflationary stock and credit collapse ending financial life as we know it in 2008 or 09 could be the bad news.
To echo Schnaps' comment, I too found this stunning:
The OCC Mortgage Metrics Report uses standardized
definitions for three categories of mortgage
creditworthiness: prime, Alt-A, and subprime. These
are defined using ranges of FICO credit scores at the
time of origination, as follows: prime 660 and above;
Alt-A 620 to 659; and subprime below 620.
There is certainly value in standardizing definitions, but this Alt-A definition completely misses what's meant by the term "Alt-A." It's like saying we're going to track all miles travelled by passengers in cars, boats and planes, and we define boats as vehicles that ride on rails over land and make stops at stations. Yeah, you've standardized a definition but you're going to accurately report the passenger miles traveled by what any non-idiot thinks of as a boat.
Another, arguably lesser, evil of this definition is that it (once again) fetishizes FICO scores.
If you had to choose one dimension, just one dimension, with which to segment the entire portfolio into something more meaningful than stupid FICO buckets, I would have gone with DTI, something like Documented Low, Documented High, and Unknown/Unverified. That's cheating a little b/c I'm insisting on the documented vs. undocumented distinction, but that's much better.
Of if you wanted to act like you had a credit thought in your brain, instead of relying on the FICO buckets, you might have defined the three buckets by LTV. There's a number that every loan has and much more readily correlates to what you'd think of as a "prime" vs. "nonprime" kind of risk.
All that said, kudos, b/c an effort at data collection is useful. Although I'd note that with a subscription to LoanPerformance you get incredibly granular data already.
Okay, now that I've skimmed the rest of the report, the stupid Alt-A definition is really pissing me off, b/c it makes the rest of the report almost completely useless.
The report tells me "Alt-A" is 9% of serviced mortgages (in this portfolio... which by the way since it excludes CFC is missing far and away the largest servicing book of truly Alt-A loans, but whatever).
Then it tells me that Alt-A loans were seriously DQ at a rate 6 times higher than Prime loans were. But b/c of their stupid definition, we're learning anything useful here. Yeah, people in the relatively narrow FICO band of 620-660 are much more likely to go 2 months behind in a mortgage payment than the MUCH LARGER population over 660. I know that. The way you get a 650 FICO is by relatively frequently paying late. If you look at the FICO distribution of the US population with FICO scores, it is heavily skewed to the high end. (I haven't seen the data in a while, but I believe the average FICO is in the 730s, and the mode is higher than that).
This reporting could generate useful information -- not just useful to us lenders or investors but, you know, useful to the GD regulators(!!) if they'd instead reported on something like documentation level or LTV. Those are the two dimensions where lending got completely out of hand and where some standardized data collection and analysis could actually help make the world safe again.
Instead what we get is continued fetishizing of FICOs which will lead to really insipid regs about making loans to borrowers w/ FICOs from X to Y. And yes, I know that the regulatory definitions for "subprime" mortgages requiring 100% risk-weighting already fetishizes FICO, but just b/c we've been lazy in the past -- with disastrous results -- is no reason to insist on continuing to be lazy in the future.
Thus endeth my rant. I'll be curious if someone actually finds something useful and interesting in here.
Interesting data, but the fact that it is taken from nine of the large banks raises questions about how representative the data is. I could not help but notice that Countrywide is not on the list. Covering the largest banks makes sense from the perspective of an agency with limited resources (which I understand all too well), but the largest banks in the past kept some of the most transparent assets for their portfolios while shoveling the questionable assets to other entities.
These statistics are useful (any information is better than none), but I have to wonder what the performance of the other 60% of mortgages looks like in comparison.
That was a pretty good rant, Price Stout. Thanks for saving me the keystrokes.
Shortcomings aside, I'm just glad to see public data on delinquencies, finally. The Mortgage Bankers Association charged hundreds of dollars for its reports. So much of the market activity is subsidized with federal money but regulators conveniently never produced data about how loans performed. Woulda been helpful, like, 10 years ago.
Why is it that this report shows repays outnumbering mods by 4-to-1, while the last MBA report showed a 50-50 split?
Price Stout - I find it amazing and appalling that someone doesn't know the FICOs on 20% of the mortgages. Kind of makes it difficult to estimate risk on the portfolio, doesn't it? That, more than anything else, is what is notable about this report. One immediately suspects that these banks have not a clue as to what their risks are.
Also, I have always considered Alt-A to be prime borrowers with riskier loan types, not borrowers with lower FICOs.
Totally bizarre.
Where is everyone? What's going on? Hold me, Sebastian, I'm scared!
One immediately suspects that these banks have not a clue as to what their risks are.
While I agree with the general point you are trying to make, MoM, here's the thing: The credit risk may in some cases not be theirs - "they" being the servicer. Stay with me here.
Strictly speaking, in order to service a loan, one doesn't need to know what the FICO was at the time of origination.
And if your solitary stake in a given loan is the right to perform its servicing, then why pay to store a data element you don't need? Every byte adds up, MoM. Which is exactly the kind of myopic crapthink you might ponder and/or implement when your forte as a servicer is largely based on worshipping at the altar of "efficient operations" and "economies of scale".
The same altar that apparently only Tanta, the Shnapster, and maybe a handful of other CR-dwelling heretics such as yourself are foolish enough to question.
Shnaps - well, one reason I'd want to know the FICO is that it costs more to service subprime loans!!!
Just trying to imagine how this is the best info the OCC could get is making my head hurt. No geographic distributions? Surely they know where they are?
MaxedOut - the 20% surprised me, yes, but it surprised me on the low side. Most banks are much worse at data capture and retrieval than you could possibly imagine. That nine banks can report the FICO score on 80% of the mortgages they service is actually pretty impressive to me. Note the language in the definition section about "best efforts" data collection. That suggests there was a fair amount of effort involved in collecting this data, much more than simply pressing a button and producing a flat file for the OCC.
You are correct in your understanding of Alt-A. That is exactly the point the issue that set me off. The OCC is measuring something w/ that 620-660 bucket, but it ain't what anyone in the industry would call Alt-A. They might call it "near prime."