warning: simplexml_load_file(http://gdata.youtube.com/feeds/api/videos/wZ7YedEopp4) [function.simplexml-load-file]: failed to open stream: HTTP request failed! HTTP/1.0 404 Not Found
in /home/dev1/public_html/sites/all/modules/custom/contributed_links/contributed_links.module on line 735.
warning: simplexml_load_file() [function.simplexml-load-file]: I/O warning : failed to load external entity "http://gdata.youtube.com/feeds/api/videos/wZ7YedEopp4" in /home/dev1/public_html/sites/all/modules/custom/contributed_links/contributed_links.module on line 735.
When men destroy their old gods they will find new ones to take their place. - Pearl S. Buck
Oh yeah, that one does it justice.
Reality is that which, when you stop believing in it, doesn't go away. - Philip K. Dick
No wait. That one!
A person will worship something, have no doubt about that. We may think our tribute is paid in secret in the dark recesses of our hearts, but it will out. That which dominates our imaginations and our thoughts will determine our lives, and our character. Therefore, it behooves us to be careful what we worship, for what we are worshipping we are becoming. - Ralph Waldo Emerson
I can feel it! My name is changing! Finally! No longer shackled by "stagflationary" concerns!
Actually, you know, I'd probably prefer subprime myself under the circs.
All this is really saying is that subprime securities had risk baked in to the yields and support levels, and a lot of Alt-A did not. It's apparent by now that subprime wasn't perfectly priced, but we're talking about buying at a hefty discount here.
Or, to put it another way, if you're feeling like taking some risk in search of higher yields, which makes more sense: subprime or Alt-A? Alt-A doesn't pay you enough for the risk.
OT: Can we do something about the "first posters"? This plague of comment sections has arrived at CR and now appears in almost every post. If left unchecked, before long you have to scroll through several comments before you find one with any content. Witness the content free first post above.
If this becomes the consensus view, that credit enhancements are better on subprime paper, and that the coupons are a better risk-adjusted return than AltA, then marked-to-market losses will be much larger than forecast.
The average AltA loan has twice the principal balance of an average subprime loan, so we have ammo for twice as many losses. (the FirstBoston slides from last week show '06 total production shares of subprime & AltA about equal).
Clyde, the other thing that struck me here (and in the RA downgrade announcements lately) is the apparent relative consistencies across subprime shelves as opposed to Alt-A.
Civilians: a "shelf registration" means an issuer can put together all the expensive legal work for a series of deals in one SEC registration; this creates a series of issues all tied back to the original "shelf." So some set of bonds labelled 2007-01, 2007-02, etc. are multiple issues off a single shelf.
This leads me to conclude that subprime is still more granular than Alt-A, and so the models are probably more reliable. Alt-A has so much variation in loan size and characteristics (while having way too much geographical concentration) that the tail problem is enormous.
Or, to put it another way, if you're feeling like taking some risk in search of higher yields, which makes more sense: subprime or Alt-A? Alt-A doesn't pay you enough for the risk.
I'm not much of a risk taker these days, but feel free to substitute Alt-A for subprime in what follows. Subprime just sounds more amusing to me.
I think it is clear that the losses in subprime are significant but some part of this risk was priced it. And Alt-A had razor-thin margins.
Let's not forget that there was no option payment subprime while it was common in Alt-A. And pay-option mortgage is a more potent bomb but with longer fuse.
This is interesting, but I'm afraid I'm not certain I understand what they mean...
Do they mean they over-collaterized the subprime securities, so $1.1 billion in loans is packaged into $1 billion in bonds, but with Alt-A $1b loans = $1b in bonds?
But what does subordination refer to here?
Weve historically been very wary of alt-A because of the decreased levels of subordination in the transactions...
No equity tranche to take risk?
See, this is why I don't rate your and CR's writings with the Bloomberg & WSJ, I read just about everything on either about Alt-A and neither have mentioned this distinction.
Tanta, I guess this answers my question from (last month?) about which is worse: higher tranches of securites backed by really bad mortgages or lower tranches of securities backed my merely bad mortgages. THIS is the "models break down" phrase from the ratings people means. Instead of having bond defaults evenly progressing up the scale from junk to A to AA to AAA we have a more random pattern of defaults. This is why we get a credit crunch: you can't easily "retreat to quality," when you have no reliable indication of quality. The ineffectiveness of bond ratings and the opacity of many of these "funds of funds" will result in a "great de-leveraging" IMHO.
Bob, they're talking here about the size or "depth" of subordination as well as the OC.
Because the rating agencies predicted lower defaults for Alt-A than we're getting, they allowed creation of deals with very thin subordinate tranches. So an Alt-A deal might have, say, 5-6% of the principal balance subordinated to the A-level tranches. A subprime deal might have 10-12% (or more).
The level of subordination creates the "break loss," or the amount of principal that can be written down before a given tranche takes a loss.
The subordinate classes, in turn, are supported by the overcollateralization. A lot of these deals were structured with "target OC" instead of original OC. In other words, they may have issued a $1B bond collateralized by a $1B pool, but they did so with a pool that (is supposed to) throw off more interest than the bond coupon. The excess interest gets applied to the senior bonds as principal. That means that the senior bonds amortize faster than the pool; this creates the OC over time (bond balance becomes smaller than pool balance). Issuers like this because it increases their leverage: they don't have to hold large equity portions of originally-overcollateralized pool balances.
It bites you in the ass if all of the below happen to you at once:
Defaults happen much earlier in the life of the deal than you expected (before OC can build up).
Refi opportunities tighten, such that only your best loans prepay, leaving you with adversely-selected remaining loans. They may have the highest rates, but they're defaulting.
The excess spread was goosed by including 5-10% high-yield second liens in the deal, which default fast and produce 100% loss severities.
What can happen is that those thin subs lose their "loss coverage ratio" because the OC isn't building up to target levels. This can happen to a subprime bond as much as to an Alt-A, but some people are claiming that the subprimes probably predicted early defaults more accurately and had more excess spread/more reliable excess.
1. Defaults happen much earlier in the life of the deal than you expected (before OC can build up).
Refi opportunities tighten, such that only your best loans prepay, leaving you with adversely-selected remaining loans. They may have the highest rates, but they're defaulting.
The excess spread was goosed by including 5-10% high-yield second liens in the deal, which default fast and produce 100% loss severities.
Yeah, but c'mon what are the chances of all three of those things happening at once? ducks
Damn, Tanta, once again you explain something simply which none of the financial journalists seem grasp, even vaguely.
You should set up seminars for financial journalists on mortgage securities. You could make a mint. I knew a lot of people doing this in the Internet-bubble years. This is how it works:
You charge participants a completely absurd rate, because, after all, the company's paying for it.
You charge all sorts of tangentially related businesses for booths, courtesy suites, etc.
You charge people to be on panels and to give keynote addresses, etc.
You get the hotel to give you space for free because of all the business you're bringing them.
Your expenses are next to nothing...
I know one dot-com "guru" still working this game...
if any of u remember, S&P "accidentally" included some alt-a deals in their initial subprime actions, and those alt-a weren't subsequently hit in the alt-a action. if u use subprime type assumptions in an alt-a deal, it blows up, obvs...
I think Tanta's comment on the greater granularity of subprime is very important.here are 2 Alt-A loans I am familiar with.A non-english speaking us citizen Dishwasher whose daughter the Realtor put him in a $680k home with an 80-20 piggyback stated income loan.A self employed Medical professional who bought out the wife in a divorce so the son could stay in the same home until he graduated HS with a 55% LTV I/O hybrid ARM,all he could afford since the home had appreciated so much in the 15 years since he bought it...This man has a 6 figure income and little debt.due to the prevalence of fraud and the total lack of underwriting you can have no clue WHAT the quality is in the Alt-A pools originated in california.
If this is a prolonged buyer's strike or we have massive ongoing liquidity dislocations, it well could be that a very low level of defaults in prime mortgage deals will generate much larger losses than we could ever conceive.
The credit enhancements for AAAs on prime deals were in the 3.25% area until February or so. With all of the mounting liquidity issues, new regulation, alienation of the foreign investor base, deleveraging, etc., resets are going to be tough for prime borrowers as well.
Let's just hope that we see some improvement in the mortgage markets by January. I think we are done for the rest of the this year.
probert, I didn't mean anything sinister by putting LCR in quotes. I really just was indicating that it is an industry term, not mine.
LCR is just a way of measuring how well-protected those subs are given current expectations of lifetime losses on the deal. At any time you have an LCR of less than one, you have a tranche that is going to take a principal write-down sooner or later. AAA tranches should have LCRs of at least 2.50. The LCR will change over time if the loss estimates change.
bacon dreamz, some days I just don't want to clutter things up with too much industry jargon. It's my understanding that the overwhelming majority of subprime and Alt-A issues in the last several years were OCs, not sixpacks. I look at the rating announcements pretty regularly, and I can't remember having seen a downgrade of a six-pack recently. Have I been wit-wandering?
tj, prime deals are more likely to have bond insurance, letters of credit, or agency wraps than OC/excess spread. They are also the real "granular" deals with the highest degree of geographic diversification. As Clyde notes, this doesn't make them bullet-proof, but it makes them way easier to model.
That's really what this is about: Alt-A is just harder to model than prime or subprime.
oh no i haven't seen any 6pack downgrades either, and most deals are OC, what i'm saying is, perhaps they end up being safer than deals that rely on a build-to OC, even though issuers generally want the OC structure for leverage, like u say. also, 7-10 yr step-downs? another way subprime bonds get better protection.
First? First!
"We are much bigger believers in subprime."
Hope springs eternal. Nonetheless, even prime non-jumbo from 2005 and 2006 will be garbage if things get bad enough
We are much bigger believers in subprime.
When men destroy their old gods they will find new ones to take their place. - Pearl S. Buck
Oh yeah, that one does it justice.
Reality is that which, when you stop believing in it, doesn't go away. - Philip K. Dick
No wait. That one!
A person will worship something, have no doubt about that. We may think our tribute is paid in secret in the dark recesses of our hearts, but it will out. That which dominates our imaginations and our thoughts will determine our lives, and our character. Therefore, it behooves us to be careful what we worship, for what we are worshipping we are becoming. - Ralph Waldo Emerson
I can feel it! My name is changing! Finally! No longer shackled by "stagflationary" concerns!
Actually, you know, I'd probably prefer subprime myself under the circs.
All this is really saying is that subprime securities had risk baked in to the yields and support levels, and a lot of Alt-A did not. It's apparent by now that subprime wasn't perfectly priced, but we're talking about buying at a hefty discount here.
Or, to put it another way, if you're feeling like taking some risk in search of higher yields, which makes more sense: subprime or Alt-A? Alt-A doesn't pay you enough for the risk.
Bernanke ordeal has Greenspan written all over it
Bernanke ordeal has Greenspan written all over it
| Reuters
The Legacy
OT: Can we do something about the "first posters"? This plague of comment sections has arrived at CR and now appears in almost every post. If left unchecked, before long you have to scroll through several comments before you find one with any content. Witness the content free first post above.
If this becomes the consensus view, that credit enhancements are better on subprime paper, and that the coupons are a better risk-adjusted return than AltA, then marked-to-market losses will be much larger than forecast.
The average AltA loan has twice the principal balance of an average subprime loan, so we have ammo for twice as many losses. (the FirstBoston slides from last week show '06 total production shares of subprime & AltA about equal).
Clyde, the other thing that struck me here (and in the RA downgrade announcements lately) is the apparent relative consistencies across subprime shelves as opposed to Alt-A.
Civilians: a "shelf registration" means an issuer can put together all the expensive legal work for a series of deals in one SEC registration; this creates a series of issues all tied back to the original "shelf." So some set of bonds labelled 2007-01, 2007-02, etc. are multiple issues off a single shelf.
This leads me to conclude that subprime is still more granular than Alt-A, and so the models are probably more reliable. Alt-A has so much variation in loan size and characteristics (while having way too much geographical concentration) that the tail problem is enormous.
Or, to put it another way, if you're feeling like taking some risk in search of higher yields, which makes more sense: subprime or Alt-A? Alt-A doesn't pay you enough for the risk.
I'm not much of a risk taker these days, but feel free to substitute Alt-A for subprime in what follows. Subprime just sounds more amusing to me.
Monty Python: Spam
Have you got anything without subprime in it?
Well... subprime, TIPS, I-Bonds, and subprime. That's not got much subprime in it.
I don't want ANY subprime!
I think it is clear that the losses in subprime are significant but some part of this risk was priced it. And Alt-A had razor-thin margins.
Let's not forget that there was no option payment subprime while it was common in Alt-A. And pay-option mortgage is a more potent bomb but with longer fuse.
Tanta,
This is interesting, but I'm afraid I'm not certain I understand what they mean...
Do they mean they over-collaterized the subprime securities, so $1.1 billion in loans is packaged into $1 billion in bonds, but with Alt-A $1b loans = $1b in bonds?
But what does subordination refer to here?
Weve historically been very wary of alt-A because of the decreased levels of subordination in the transactions...
No equity tranche to take risk?
See, this is why I don't rate your and CR's writings with the Bloomberg & WSJ, I read just about everything on either about Alt-A and neither have mentioned this distinction.
Thanks, Tanta.
Tanta, I guess this answers my question from (last month?) about which is worse: higher tranches of securites backed by really bad mortgages or lower tranches of securities backed my merely bad mortgages. THIS is the "models break down" phrase from the ratings people means. Instead of having bond defaults evenly progressing up the scale from junk to A to AA to AAA we have a more random pattern of defaults. This is why we get a credit crunch: you can't easily "retreat to quality," when you have no reliable indication of quality. The ineffectiveness of bond ratings and the opacity of many of these "funds of funds" will result in a "great de-leveraging" IMHO.
Bob, they're talking here about the size or "depth" of subordination as well as the OC.
Because the rating agencies predicted lower defaults for Alt-A than we're getting, they allowed creation of deals with very thin subordinate tranches. So an Alt-A deal might have, say, 5-6% of the principal balance subordinated to the A-level tranches. A subprime deal might have 10-12% (or more).
The level of subordination creates the "break loss," or the amount of principal that can be written down before a given tranche takes a loss.
The subordinate classes, in turn, are supported by the overcollateralization. A lot of these deals were structured with "target OC" instead of original OC. In other words, they may have issued a $1B bond collateralized by a $1B pool, but they did so with a pool that (is supposed to) throw off more interest than the bond coupon. The excess interest gets applied to the senior bonds as principal. That means that the senior bonds amortize faster than the pool; this creates the OC over time (bond balance becomes smaller than pool balance). Issuers like this because it increases their leverage: they don't have to hold large equity portions of originally-overcollateralized pool balances.
It bites you in the ass if all of the below happen to you at once:
What can happen is that those thin subs lose their "loss coverage ratio" because the OC isn't building up to target levels. This can happen to a subprime bond as much as to an Alt-A, but some people are claiming that the subprimes probably predicted early defaults more accurately and had more excess spread/more reliable excess.
1. Defaults happen much earlier in the life of the deal than you expected (before OC can build up).
Yeah, but c'mon what are the chances of all three of those things happening at once? ducks
Got any spare space in the Bankerdome?
take this FWIW, but chatter is current bid sheets on alt-a AAAs r in the 80s...liquidations are fun!!!
Damn, Tanta, once again you explain something simply which none of the financial journalists seem grasp, even vaguely.
You should set up seminars for financial journalists on mortgage securities. You could make a mint. I knew a lot of people doing this in the Internet-bubble years. This is how it works:
Your expenses are next to nothing...
I know one dot-com "guru" still working this game...
tanta don't forget about the 6packs! they're not all OC deals!
if any of u remember, S&P "accidentally" included some alt-a deals in their initial subprime actions, and those alt-a weren't subsequently hit in the alt-a action. if u use subprime type assumptions in an alt-a deal, it blows up, obvs...
Okay, since Alt-A MBS have less "protection" than Subprime, do Prime MBS have any at all (besides the presumed quality of the loans themselves)?
It's not just Subprime & Alt-A homes losing value, so I expect those Prime MBS to fall harder & faster once containment spreads far enough.
I think Tanta's comment on the greater granularity of subprime is very important.here are 2 Alt-A loans I am familiar with.A non-english speaking us citizen Dishwasher whose daughter the Realtor put him in a $680k home with an 80-20 piggyback stated income loan.A self employed Medical professional who bought out the wife in a divorce so the son could stay in the same home until he graduated HS with a 55% LTV I/O hybrid ARM,all he could afford since the home had appreciated so much in the 15 years since he bought it...This man has a 6 figure income and little debt.due to the prevalence of fraud and the total lack of underwriting you can have no clue WHAT the quality is in the Alt-A pools originated in california.
Tanta, what does "loss coverage ratio" usually mean, and why did you put it within quotation marks above?
TJ,
I think you have hit on the depression question.
If this is a prolonged buyer's strike or we have massive ongoing liquidity dislocations, it well could be that a very low level of defaults in prime mortgage deals will generate much larger losses than we could ever conceive.
The credit enhancements for AAAs on prime deals were in the 3.25% area until February or so. With all of the mounting liquidity issues, new regulation, alienation of the foreign investor base, deleveraging, etc., resets are going to be tough for prime borrowers as well.
Let's just hope that we see some improvement in the mortgage markets by January. I think we are done for the rest of the this year.
probert, I didn't mean anything sinister by putting LCR in quotes. I really just was indicating that it is an industry term, not mine.
LCR is just a way of measuring how well-protected those subs are given current expectations of lifetime losses on the deal. At any time you have an LCR of less than one, you have a tranche that is going to take a principal write-down sooner or later. AAA tranches should have LCRs of at least 2.50. The LCR will change over time if the loss estimates change.
bacon dreamz, some days I just don't want to clutter things up with too much industry jargon. It's my understanding that the overwhelming majority of subprime and Alt-A issues in the last several years were OCs, not sixpacks. I look at the rating announcements pretty regularly, and I can't remember having seen a downgrade of a six-pack recently. Have I been wit-wandering?
tj, prime deals are more likely to have bond insurance, letters of credit, or agency wraps than OC/excess spread. They are also the real "granular" deals with the highest degree of geographic diversification. As Clyde notes, this doesn't make them bullet-proof, but it makes them way easier to model.
That's really what this is about: Alt-A is just harder to model than prime or subprime.
oh no i haven't seen any 6pack downgrades either, and most deals are OC, what i'm saying is, perhaps they end up being safer than deals that rely on a build-to OC, even though issuers generally want the OC structure for leverage, like u say. also, 7-10 yr step-downs? another way subprime bonds get better protection.
there certainly were alt-a 6packs in 06, though, when the rating agencies didn't like the looks of ur excess spread. they aren't that rare.
just to get all number-y on u, of 39 indicative option arm deals in 2006, 15 were 6pack, 24 were OC. in 2005, 57 were 6pack, 7 were OC.
Walker, the first post plague has been contained.