Interesting chart. The trend looks like the first 12 months after the end of the issuing year are the key ones, then the delinquency rate stabilizes. Since subprime standards changed considerable over the summer, things are going to get a lot worse between now and summer 08, and then the performance (or lack thereof) of subprime MBSs could stabilize.
Tanta, I'm wondering if we are going to see a 2nd surge in delinquencies for the '05 and maybe the '04 vintages. As prices fall next year - in some places below '04 prices - I think some of those older loans might get in trouble.
This chart only goes out 21 months, but look at how steep the '05 vintage is near the end - it wasn't flattening out as much as earlier years.
hm, good point, CR. I didn't put that together with the 2005 curve. Since prices were going up so quickly, people with 2-year resets that happened up to 2005-2006 could refi easily, probably into another 2/28. Hm... but if that happened, then wouldn't that just transfer the refi over to the 2005 or 2006 bin? I wonder how many of the pre-2005 loans had 3 or 5 year resets.
Interestingly, FICO scores and LTV ratios do not vary significantly between the strongest and weakest performing transactions and on average transaction performance does not appear to have been influenced by these characteristics.
Weird. Maybe it's correlated with the borrowers financial ability to repay the loan? Nah! Must be more complicated than that...
well, the 04/05 vintages mostly have FRMs left in them b/c the hybrids have mostly prepaid away...if you look at a chart going out to say 30m, there wasn't a huge spike in default rates at reset, but there likely will be for 06/07, which is already running way higher anyways.
Peterbob, CLTV is total loan to value, whereas LTV is the originated loan to value. In other words, if you are underwriting 80% and the borrower is getting a 20% behind that, your LTV is 80% but CLTV is 100%.
True DTIs would catch overloads of debt, but stated income loans don't provide meaningful DTIs. As JBR notes, who the heck cares about DTIs anyway! All we need is a good FICO....
Another case in point are home equity loans in which you are behind. Your LTV might be 15%, which sounds great - but if there's another 75% in front of that, the CLTV is 90%.
Tanta, I'm wondering if we are going to see a 2nd surge in delinquencies for the '05 and maybe the '04 vintages.
Well, most of that 2005 vintage was 2/28 ARMs, meaning that they haven't all reset yet. So I expect to see 2005 continue to rise.
I'd guess, actually, that 2004 is probably about as bad as it's going to get. Those folk had enough equity (or "equity") to sell or refi in 2006 when the reset hit. Which is of course another reason why 2006 sucks so much: it contains the "refugees" (refinances) from earlier purchase vintages. I believe that average pool factor (% of original balance remaining) for 2004 subprime is around 23% (as opposed to 53% for 2005, 83% for 2006 and 97% for 2007).
another reason why this doesn't match with other chart is because they're using months after issuance instead of age in the x-axis.
As for the fico/ltv, it's well known that fico doesn't affect subprime's delinquency, but it's surprising that they're claiming ltv doesn't either. probably because the 1st ltv doesn't vary much, since most people who wants to borrow more go for 1st+2nd liens.
if you look at abx 06-1 as a proxy for later 2005 2/28s resetting right now (i think the WAC increased 40 bps last month), the delinquencies really aren't moving higher any faster than 2006 loans that haven't reset yet, and they continue to prepay decently fast.
LTV=Loan to Value Ratio
CLTV=Combined Loan to value Ratio
Example:
$100,000 house
$80,000 first mortgage
LTV: 80%
CLTV: 80%
Example #2
$100,000 house
$80,000 first mortgage
$15,000 second mortgage (or HELOC)
CLTV: 95%
IMO: this is why CLTV matters a lot (because it takes all loans into account) but LTV doesn't as much (because you may be ignoring a significant second loan/line of credit)
lots of those crappy loans out there are 80% LTV (on the first) but also 20+% for other loans giving a 100% CLTV...
In addition to the points about curve slope, future exposures work the curves as a timeseries in you mind. 2000/01 were pretty abysmal but got things got very good indeed in terms of levels and slopes in '03/'04 and then went to hxxx quickly. Think back on the timecurve against the economy - flattening economy in '00 with GDP and employment both tanking in '01 or so and then picking back up.
Taken all together this also tells us why so much of the ratings were so "sanguine" - fascinating story of not looking for the characteristics that drove the obsevable despite Fleck et.al. doing the ATM dance for so long.
Miami Valley becomes year's 3rd failed bank - FDIC
Miami Valley Bank, with $86.7 million in assets, became the third failed insured bank this year when Ohio's industry regulators closed it down on Thursday, the Federal Deposit Insurance Corporation said.
The Citizens Banking Corp took over $62 million of insured deposits and two offices, which will reopen on Friday, the FDIC said in a statement.
The FDIC, which took receivership of the failed bank, said its remaining $14 million in 269 deposits exceeded the federal deposit insurance limit and depositors of that amount will become creditors.
They say that FICO score is irrelevant to the findings but this chart only shows subprime borrowers which is a classification that is FICO-derived. I wouldn't expect there to be much difference in default rates.
What I would like to see is what we're talking in actual dollar value and what it looks like for prime and Alt-A, which I would assume are much bigger pies than subprime.
The steepness is shocking, but the subprime origination peaked in late 2005 and declined thereafter. So, the overall impact of the 2007 vintage might not be as big as that of 2005 or 2006.
I wonder if there is not some kind of Game Theory going on with borrowers trying to anticipate actions by lenders,politicians and other borrowers to decide whether it will be better to be current or delinquent when policies are decided. A form of real estate "Prisoners Dilemma".
very good, very good. my thoughts exactly. probably the calmest actor in this whole scene is the borrower smiling in amusement at all the politicians and lenders running around with their heads cut off trying to "fix" the problem and save their jobs. maybe Joe average isn't so dumb after all.
Gee, Zarley, that kind of sounds like the sky falling to me.
Tanta -- If you'll still be reading these, why is the slope so steep for 2007 loans? I would've thought with subprime coming out of the closet as being problematic this year, and the whole real estate industry slowing down so much, that subprime lending would be slowing down and getting less lethal. Along those lines, wouldn't you think from here on out these figures are going to start cleaning up? My mind is mush right now so maybe I'm all wrong.
To begin with, neg. am. and IO loans were disproportionately distributed in many markets. I seem to recall a 40% figure in San Diego.
Second, rather than asking what percentage of borrowers have subprime loans, might more relevant questions be:
-"What is the average debt to equity ratio of the average homeowner?"
-"If prices do drop by 20%, what percentage of homeowners will be underwater on their mortgages?"
-"When do underwater homeowners start walking away from their homes?" In California, this amounts to "What is the dollar value of your credit rating"
Finally, we're got a long way to go before "prevailing bias" resembles Tanta.
"When do underwater homeowners start walking away from their homes?" In California, this amounts to "What is the dollar value of your credit rating"
That's way too simplistic. If your home is in a good neighborhood where the houses still look nice and the children play in the streets, that's one thing.
But if your neighborhood has turned into weeds, green swimming pools, and crack dealers camping out, that's another.
If you have huge debts aside from your mortgage and your job sucks and people are getting pink-slipped, that's one thing. But if you are debt-free and have tenure or work for the government, that's another.
The report said sterling would soon slip out of the G-10's top trio of high-yield currencies. "This could have dramatic consequences: the "hot money' flow that supported sterling could leave suddenly," said Mr Bloom.
They make it sound "hot money" is a good thing, not some speculative pestilence that swarms from country to country devouring financial infrastructure and real productivity.
Every country wants to keep their easy credit high, no matter how much more damage it does each day.
Maybe we could bring Alan Greenspan back:
He was the best drug dealer the global economy ever had.
Do you believe 40% of all homeowners in the San Diego area have neg am or IO loans. Is that what you believe??? If so I got a bridge for you my friend...
Zarley | 10.04.07 - 11:05 pm |
i did'nt 'read' that implication in his statement...
i read 'in the 04'-'05 vintage, 40% 'may' have been io,NA loans...
that could be true...
A statistically relevant variable has to be current net equity in the property. The newer vintages essentially were soon under water. Combined with the aggressive/lax underwriting, you have a high foreclosure rate.
These loans are/were essentially a house option and the option is/was out of the money.
How can a loan be "60 days or more delinquent" 1 or 2 months after issuance?
There are two ways you can define a "vintage": by the year the loan closes, or the year the security closes. This chart is based on the security closing (or issue) date. In the 2005-2006 years, especially, securities were issuing so fast that loan age at the time of issuance was likely to be 1-3 months on average. One thing we will see as the 2007 numbers get riper is that the loans are somewhat older--as the securitization market has frozen up, you have closed loans sitting around in inventory, and deals finally done when the loans are more like 6 months old.
Stuff you see from the RAs is likely to be based on deal year, not loan year, because, well, they deal with deals. So most of us look at a chart like this and mentally shift back a quarter to get an approximate loan age.
It was certainly possible in earlier subprime vintages to have loans already 30 days down when the deal closed. With the RAs tightening up, I suspect there will be less of that (the early nonperformers will have to go into scratch & dent deals). So yes, you can certainly have loans 60 days or more down in the same quarter they were issued.
That is the real point here: these vintage charts are really useful for showing the speed of delinquencies, not the level.
ISTM that CLTV would be a factor in the chance of default, but LTV would be important in AMOUNT of loss IN a default. Of course total DTI would probably be better than CLTV as a predictor of the chance of default.
Interesting chart. The trend looks like the first 12 months after the end of the issuing year are the key ones, then the delinquency rate stabilizes. Since subprime standards changed considerable over the summer, things are going to get a lot worse between now and summer 08, and then the performance (or lack thereof) of subprime MBSs could stabilize.
I wonder if this data is available for alt-a.
Tanta, I'm wondering if we are going to see a 2nd surge in delinquencies for the '05 and maybe the '04 vintages. As prices fall next year - in some places below '04 prices - I think some of those older loans might get in trouble.
This chart only goes out 21 months, but look at how steep the '05 vintage is near the end - it wasn't flattening out as much as earlier years.
Just a thought ...
All my best!
hm, good point, CR. I didn't put that together with the 2005 curve. Since prices were going up so quickly, people with 2-year resets that happened up to 2005-2006 could refi easily, probably into another 2/28. Hm... but if that happened, then wouldn't that just transfer the refi over to the 2005 or 2006 bin? I wonder how many of the pre-2005 loans had 3 or 5 year resets.
Interestingly, FICO scores and LTV ratios do not vary significantly between the strongest and weakest performing transactions and on average transaction performance does not appear to have been influenced by these characteristics.
Weird. Maybe it's correlated with the borrowers financial ability to repay the loan? Nah! Must be more complicated than that...
What is CLTV as opposed to LTV?
well, the 04/05 vintages mostly have FRMs left in them b/c the hybrids have mostly prepaid away...if you look at a chart going out to say 30m, there wasn't a huge spike in default rates at reset, but there likely will be for 06/07, which is already running way higher anyways.
more charts here:
http://www.fitchratings.com/web_content/sectors/subprime/Subprime_Review_1004.pdf
CR i'm thinking the same thing. The slopes on all those curves could adjust much higher come march.
Peterbob, CLTV is total loan to value, whereas LTV is the originated loan to value. In other words, if you are underwriting 80% and the borrower is getting a 20% behind that, your LTV is 80% but CLTV is 100%.
True DTIs would catch overloads of debt, but stated income loans don't provide meaningful DTIs. As JBR notes, who the heck cares about DTIs anyway! All we need is a good FICO....
Another case in point are home equity loans in which you are behind. Your LTV might be 15%, which sounds great - but if there's another 75% in front of that, the CLTV is 90%.
Tanta, I'm wondering if we are going to see a 2nd surge in delinquencies for the '05 and maybe the '04 vintages.
Well, most of that 2005 vintage was 2/28 ARMs, meaning that they haven't all reset yet. So I expect to see 2005 continue to rise.
I'd guess, actually, that 2004 is probably about as bad as it's going to get. Those folk had enough equity (or "equity") to sell or refi in 2006 when the reset hit. Which is of course another reason why 2006 sucks so much: it contains the "refugees" (refinances) from earlier purchase vintages. I believe that average pool factor (% of original balance remaining) for 2004 subprime is around 23% (as opposed to 53% for 2005, 83% for 2006 and 97% for 2007).
another reason why this doesn't match with other chart is because they're using months after issuance instead of age in the x-axis.
As for the fico/ltv, it's well known that fico doesn't affect subprime's delinquency, but it's surprising that they're claiming ltv doesn't either. probably because the 1st ltv doesn't vary much, since most people who wants to borrow more go for 1st+2nd liens.
Tanta, thanks for the explanation. I didn't realize the pool size (23% for 2004) had declined that much.
Those sure are steep lines for '06 and '07!
Best Wishes.
Does anyone remember why 2000 was so bad? The recession? I don't remember and RRE problems.
probably because the 1st ltv doesn't vary much, since most people who wants to borrow more go for 1st+2nd liens.
Yep. Average LTV for the top-performing bucket is 81%, in the worst bucket it's 81.1%. Big diff.
if you look at abx 06-1 as a proxy for later 2005 2/28s resetting right now (i think the WAC increased 40 bps last month), the delinquencies really aren't moving higher any faster than 2006 loans that haven't reset yet, and they continue to prepay decently fast.
Simple explanation:
LTV=Loan to Value Ratio
CLTV=Combined Loan to value Ratio
Example:
$100,000 house
$80,000 first mortgage
LTV: 80%
CLTV: 80%
Example #2
$100,000 house
$80,000 first mortgage
$15,000 second mortgage (or HELOC)
CLTV: 95%
IMO: this is why CLTV matters a lot (because it takes all loans into account) but LTV doesn't as much (because you may be ignoring a significant second loan/line of credit)
lots of those crappy loans out there are 80% LTV (on the first) but also 20+% for other loans giving a 100% CLTV...
In addition to the points about curve slope, future exposures work the curves as a timeseries in you mind. 2000/01 were pretty abysmal but got things got very good indeed in terms of levels and slopes in '03/'04 and then went to hxxx quickly. Think back on the timecurve against the economy - flattening economy in '00 with GDP and employment both tanking in '01 or so and then picking back up.
Taken all together this also tells us why so much of the ratings were so "sanguine" - fascinating story of not looking for the characteristics that drove the obsevable despite Fleck et.al. doing the ATM dance for so long.
Miami Valley becomes year's 3rd failed bank - FDIC
Miami Valley Bank, with $86.7 million in assets, became the third failed insured bank this year when Ohio's industry regulators closed it down on Thursday, the Federal Deposit Insurance Corporation said.
The Citizens Banking Corp took over $62 million of insured deposits and two offices, which will reopen on Friday, the FDIC said in a statement.
The FDIC, which took receivership of the failed bank, said its remaining $14 million in 269 deposits exceeded the federal deposit insurance limit and depositors of that amount will become creditors.
Business & Financial News, Breaking US & International News | Reuters.com
They say that FICO score is irrelevant to the findings but this chart only shows subprime borrowers which is a classification that is FICO-derived. I wouldn't expect there to be much difference in default rates.
What I would like to see is what we're talking in actual dollar value and what it looks like for prime and Alt-A, which I would assume are much bigger pies than subprime.
I will offer a simplistic explanation for this complex trend.
Americans are losing their phobias about not repaying debts.
Hey, only live once!
Thanks for the graph, Tanta.
The steepness is shocking, but the subprime origination peaked in late 2005 and declined thereafter. So, the overall impact of the 2007 vintage might not be as big as that of 2005 or 2006.
I wonder if there is not some kind of Game Theory going on with borrowers trying to anticipate actions by lenders,politicians and other borrowers to decide whether it will be better to be current or delinquent when policies are decided. A form of real estate "Prisoners Dilemma".
Gaming the system would not be novel in America.
NC Jim
very good, very good. my thoughts exactly. probably the calmest actor in this whole scene is the borrower smiling in amusement at all the politicians and lenders running around with their heads cut off trying to "fix" the problem and save their jobs. maybe Joe average isn't so dumb after all.
*** DELETED BY CR, Name Calling, False information ***
Edited By Siteowner
Gee, Zarley, that kind of sounds like the sky falling to me.
Tanta -- If you'll still be reading these, why is the slope so steep for 2007 loans? I would've thought with subprime coming out of the closet as being problematic this year, and the whole real estate industry slowing down so much, that subprime lending would be slowing down and getting less lethal. Along those lines, wouldn't you think from here on out these figures are going to start cleaning up? My mind is mush right now so maybe I'm all wrong.
HSBC reports on the Kingdom under siege
HSBC warns of hot money exodus from Britain under 'siege' - Telegraph
Britains CRE starting to suffer
British Land pulls crunch casualty - Telegraph
A Fed speech you will want to read.
Fed Dallas
Richard W. Fisher
Inflation Measurement and Price Volatility
Remarks Before the Charlotte Economics Club Oct. 4, 2007
Inflation Measurement and Price Volatility - Richard Fisher Speeches - News & Events - FRB Dallas
OT: I see the Implode-O-Meter has a new Countryfried logo, LOTFLMAO!
That's not a CountryFried logo.
It's a t-shirt! Order some for Xmas presents.
Zarley,
To begin with, neg. am. and IO loans were disproportionately distributed in many markets. I seem to recall a 40% figure in San Diego.
Second, rather than asking what percentage of borrowers have subprime loans, might more relevant questions be:
-"What is the average debt to equity ratio of the average homeowner?"
-"If prices do drop by 20%, what percentage of homeowners will be underwater on their mortgages?"
-"When do underwater homeowners start walking away from their homes?" In California, this amounts to "What is the dollar value of your credit rating"
Finally, we're got a long way to go before "prevailing bias" resembles Tanta.
That's way too simplistic. If your home is in a good neighborhood where the houses still look nice and the children play in the streets, that's one thing.
But if your neighborhood has turned into weeds, green swimming pools, and crack dealers camping out, that's another.
If you have huge debts aside from your mortgage and your job sucks and people are getting pink-slipped, that's one thing. But if you are debt-free and have tenure or work for the government, that's another.
Don't you agree?
Oh, and if the kids playing in the streets have turned into crack dealers...GET OUT!
UK, Ireland, France and Spain RE prices declining
Housing Prices Decline in Europe As Loan Costs Rise - WSJ.com
How can a loan be "60 days or more delinquent" 1 or 2 months after issuance?
The report said sterling would soon slip out of the G-10's top trio of high-yield currencies. "This could have dramatic consequences: the "hot money' flow that supported sterling could leave suddenly," said Mr Bloom.
They make it sound "hot money" is a good thing, not some speculative pestilence that swarms from country to country devouring financial infrastructure and real productivity.
Every country wants to keep their easy credit high, no matter how much more damage it does each day.
Maybe we could bring Alan Greenspan back:
He was the best drug dealer the global economy ever had.
Do you believe 40% of all homeowners in the San Diego area have neg am or IO loans. Is that what you believe??? If so I got a bridge for you my friend...
Zarley | 10.04.07 - 11:05 pm |
i did'nt 'read' that implication in his statement...
i read 'in the 04'-'05 vintage, 40% 'may' have been io,NA loans...
that could be true...
UK, Ireland, France and Spain RE prices declining...
So much for the world decoupling from the US.
OT: rent stagflation/deflation
Renters' luck: Apartment market flooded as condos convert back to rentals
They just need a higher percentage of "good" liar loans.
06 looks like it's gunning for 12% and 07 is even uglier, 15-20%? I look forward to the rating agencies having to downgrade their downgrades.
A statistically relevant variable has to be current net equity in the property. The newer vintages essentially were soon under water. Combined with the aggressive/lax underwriting, you have a high foreclosure rate.
These loans are/were essentially a house option and the option is/was out of the money.
why is the slope so steep for 2007 loans?
How can a loan be "60 days or more delinquent" 1 or 2 months after issuance?
There are two ways you can define a "vintage": by the year the loan closes, or the year the security closes. This chart is based on the security closing (or issue) date. In the 2005-2006 years, especially, securities were issuing so fast that loan age at the time of issuance was likely to be 1-3 months on average. One thing we will see as the 2007 numbers get riper is that the loans are somewhat older--as the securitization market has frozen up, you have closed loans sitting around in inventory, and deals finally done when the loans are more like 6 months old.
Stuff you see from the RAs is likely to be based on deal year, not loan year, because, well, they deal with deals. So most of us look at a chart like this and mentally shift back a quarter to get an approximate loan age.
It was certainly possible in earlier subprime vintages to have loans already 30 days down when the deal closed. With the RAs tightening up, I suspect there will be less of that (the early nonperformers will have to go into scratch & dent deals). So yes, you can certainly have loans 60 days or more down in the same quarter they were issued.
That is the real point here: these vintage charts are really useful for showing the speed of delinquencies, not the level.
ISTM that CLTV would be a factor in the chance of default, but LTV would be important in AMOUNT of loss IN a default. Of course total DTI would probably be better than CLTV as a predictor of the chance of default.