Securitization of subprime mortgage credit

[shanlane]

Hello,

I am reading this paper, which is long but a pretty easy read, and had a few questions I was hoping you could answer.

1. This statement, from p 32, really confused me:
After the lockout period...the o/c is released.

What does is mean by released? I understand what the lockout implies but saying that the over-collateralization is "released" does not resonate with me at all.

2. The other really confusing part of this is the relationship between tables 18 and 19. None of the numbers add up correctly ("deposits" does not equal sum of the rest of the columns) and it seems strange to have excess spread and a realized loss in the same year. Am I misinterpreting something?

3. Last one: There seems to be some backward logic that I cannot wrap my head around. Why does it keep saying as credit rating ratings increase that loss severity, level of expected loss on collateral and involuntary prepayments (defaults) INCREASE? This seems completely backward.

There are typos all over this paper. Maybe these are just a few examples of this?

Sorry for the long question

Thanks,
Shannon

 

[ibrahim-1987]

hi david,

regarding securitization also, what it does mean by equity tranche?
is it acting like equity to be the first loss absorber, and it is a low quality assets" loans"?
Or it is really equity issued to absorb losses first? and if it is like this way, who issue these equities? SPV? Originator? & is it a cash reserve sited aside?

 

[David Harper CFA FRM]

@Shannon, I have to try and come back to your compound question, I'm having trouble identifying the references (and I need to get notes out tomorrow)

@Ibrahim:
You might think of it like the capital structure of a company: assets on the left must equal debt + equity on the right. Equity = Assets - Liabilities. The SPE/SPV is a company with a capital structure.

In the securitization, the loans ("credit-sensitive assets") are the assets that are purchased from the originator by the SPV. Then securities are issued to investors, most of these will be debt (credit-linked notes). If the Assets are purchased for $100, and tranches of debt (notes) are issued to investors that total $90, the equity will have a value of $10 (but it is risky, if the assets suddenly drop by $10, the equity is wiped out).

So the equity (aka, junior or toxic) tranche is the security with the residual claim in the capital structure (analogous to the holder of common stock in the corporation). By definition, it absorbs the first losses. Due to its residual nature, it can be located various places; e.g., the SPV can issue equity directly to an equity investor; the originator can retain by selling the assets at a discount (if you sell me something worth $100 for $90, you can keep a $10 equity stake); the SVP can just "hold back" by overcollateralization (O/C).

The equity tranche in the case study is created by overcollateralization: the SPV purchased ~ $882 million in principal loan balances (ASSETS) then issued 98.6% of that in notional tranches (DEBT), so the equity tranche of 1.4% (=Assets - liabilities). If the assets go up in value, equity increases without cash. Excess spread (~ retained earnings) will increase. Due to the high leverage, I would imagine that in this case especially, the current value of the equity tranche never had a strong relationship with any cash contributed. In the same way that, for example, that it is almost impossible to value $1 invested into a bank with 96% or 98% leverage: $1 is your cost, but it has upside of ? and can easily be wiped out to zero, so your price and value will be a function of a negotiation.

Thanks, David

 

[shanlane]

No worries. You have been very extremely helpful. A friend of mine told me that you were very good about answering questions and she was right. I am very glad I decided to join for part 2!

Thanks,
Shannon

 

[David Harper CFA FRM]

Shannon, Thanks, I really appreciate that (sometimes good questions stump me ... I've studied the Ashcraft paper for seriously years, but you've come up with some new challenges, so i do look forward to trying this weekend)

 

[shanlane]

Looking forward to seeing what you come up with. I am sure it will help clear things up for me!

Thanks in advance for you help,
Shannon

 

[David Harper CFA FRM]

Hi Shannon,

1. The O/C release. Geez, p 32/37 of Ashcraft does a terrible job of explaining this; I get more confused with each re-reading of this section. Fortunately, this aspect is rather common to MBS. What release means is best conveyed by a single of his sentences sentence, "At this time [dh: i.e., in the non-event of a non-trigger, such that the structure can step-down without a trigger], any excess O/C is released to the holders of the Class X tranche."
... note #1: the trigger a "bad" trigger in the sense that that it includes losses exceeding a threhold (Figure 7). If the trigger is not hit, the credit situation is okay (w.r.t expectations) and the senior bondholders are okay to "step-down," without a trigger, and "release the O/C"
... note #2: keep in mind the Class X is residual equity (equity = loan assets - note liabilities)

So this deal is typical in the sense that the initial O/C is 1.4% (debt notes issued that = 98.6% of the principal value of the mortgage loans) then "builds up" the O/C over the first 36 months to 2*1.4% = 2.8%. Ashcraft doesn't say but this is almost surely by "turboing:" using excess spread to pay down (prepay) the notes/debt (probably senior/class A tranches). This has the same effect, and is much the same as (only the source of the funds is different) as the "shifting interest." For a given value of the assets, to reduce the liabilities, is to increase the equity (including the O/C); of course, the prepayment reduces the assets too, so you've got to payoff liabilities faster than assets. From my library is more on this idea:

"Instead, OC is increased by paying off the bonds’ principal. When OC is being built up early in the life of a deal, the available excess spread is “converted” into principal, and used to accelerate payment of the bonds (following the prin-cipal waterfall at that time). Note that early in a typical deal’s life, only the senior tranches receive signifi cant amounts of principal. After the step-down date, OC is released by diverting principal payments from the bonds to the residuals. At any point in time, if excess interest drops below current losses, then OC is reduced (because the collateral balance decreases faster than the bonds’ balance) and the loss amount is written down from the OC. In effect, OC decreases when the face value of the bonds is not allowed to, or cannot, drop in tandem with the collateral’s face value." -- Subprime mortgage credit derivatives

So, I would draw an analogy: Assume a firm starts with 9X leverage: 100% assets are funded by 90% debt + 10% equity (D/E = 9x). Prior to the step-down, the firm can decrease leverage (~ build O/C) by using excess cash flow to pay off debt. The firm pays down debt until leverage is down to 20% equity. Now the equity holders feel "safe" (~ triggers are NOT hit) and the firm can "step down:" the firm can now use excess cash to raise the dividend. Keep in mind that losses are dropping the asset value without a corresponding drop in liabilities: losses naturally drop the O/C. Now, if the firm diverts excess cash (excess spread) to the equity holders as dividend (release the OC), while losses occur, the OC will naturally head in the other direction and drop.
This is releasing the O/C.

2. I copied the table into Excel, per the last column, the deposits are consistently 98% to 99% of the sum of the other columns. You are correct, and they don't explain it, which is frustrating. (I bet it's explained in the source, if i had time, I'd chase it down .... ).

3. I agree with you: this sentence (on my page 54) vexes me, too. The text agrees with the values in the exhibit, so I have to assume that I don't understand what he's saying. I have to bookmark this, for later investigation, I want to understand this counter-intuitive statement, but I currently do not. Sorry!

Thanks, David

 

[shanlane]

That was very helpful. It is also nice to know that I am not just overlooking something very simple and that there is at least one other person out there that finds this as strange as I do.

Thanks!
Shannon

 

[benj]

Hi David,

I am also reading this paper and can't get my head around the section "Prepayment risk" on my page 48.

I was wondering if I could use your help on the following question:

How one identifies the number of voluntary and involuntary prepayments from

1 - CPR assumptions and
2 - Dollar value of involuntary prepayments

I always find more unknown variables than equations even when I take into account the loss severity assumptions.

Thanks a lot.

Benj.

 

[David Harper CFA FRM]

Hi Benj, per my email, I do have this query bookmarked, it's just that I don't know the answer without research (unknown duration) so I need to return to it after the next content deadline, sorry, thanks for your understanding,