SoA6, TAC vs PAC [Archive] - Actuarial Outpost

01-02-2002, 05:56 PM

SoA6, TAC vs PAC
HFIS, Ch. 26 CMO’s
Page 630.

“Unlike PACs, TACs do not provide protection against WAL extension if prepayments fall below the speed necessary to maintain the TAC schedule.”

Okay, so PACs do provide protection against WAL extension; TACs do not provide protection against WAL extension.

I can see generally that TACs want to offer a higher yield by not having protection against Weighted Average Life extension risk which occurs when interest rates increase, thus causing pre-payments to slow. After all it is the early return of principal that we are worried about, here.

However I am confused as to how the TACs excludes this protection. What I mean is that when principal is coming in much slower than the lower band, does the conduit forward the principal that it (the conduit) has not yet received so that the WAL of the TAC does not increase?

[ This Message was edited by: OT on 2002-01-04 14:22 ]

Exam Slave

01-02-2002, 06:20 PM

"However I am confused as to how the TACs exclude this protection."

If prepayment rate is slow, then the TAC class gets less principal, and the time when all principal has been finally paid gets extended. A TAC's PAC band-equivalent is 0% PSA to 240% PSA, for example (emphasis on the 0%).

A big problem with the exhibits is that each assumes a constant PSA. No word on what happens when the prepayment rate changes drastically from one period to the next.

(Edited to re-answer question.)

[ This Message was edited by: Exam Slave on 2002-01-02 19:16 ]

01-02-2002, 06:46 PM

Thanks, ES, for the correction. I am actually inquiring about the TAC and altered my original inquiry to be clearer, I hope.

But here is another question, how does a P(planned)AC companion provide protection against WAL extension risk? Okay, I see that the accretion of a PAC companion Z-bond provides protection against WAL extension risk, but not a non-accreting companion. When pre-payments are lower than the PAC lower band, a non-accreting companion has nothing to “give” since it does not accrete principal. A non-accreting companion provides protection against the WAL decrease or quicker/greater pre-payments, not WAL increase.

[ This Message was edited by: OT on 2002-01-02 19:02 ]

Exam Slave

01-02-2002, 08:53 PM

A companion bond will stave off WAL increases to the PAC, but a large enough increase in interest rates, which lowers prepayment rates, will increase WAL for PACs.

It is protection, not a guarantee. Lower risk, not no risk.

PACs were made to lower the risk due to prepayment/maturity volatility in regular MBSs.

01-02-2002, 09:54 PM

If prepayment rate is slow, then the TAC class gets less principal, and the time when all principal has been finally paid gets extended. A TAC's PAC band-equivalent is 0% PSA to 240% PSA, for example (emphasis on the 0%).

A big problem with the exhibits is that each assumes a constant PSA. No word on what happens when the prepayment rate changes drastically from one period to the next.

(Edited to re-answer question.)

[ This Message was edited by: Exam Slave on 2002-01-02 19:16 ]

In regard to ES's first paragraph, what is described looks, to me, just like a PAC/Companion. But wait, I may be way off here (1) TAC's have companions also, right? (2) Can actual prepayments for TAC's and PAC's be less than 0% PSA? Is there a required minimum speed for the return of principal for the PAC in the contract/indenture? This speed must be at least the scheduled return of principal of the underlying securities/mortgages.

In regard to ES's second paragraph: yeah, I was just thinking that too. However, it does not seem important to know for the exam since it was not addressed in the text.

Thanks, Exam Slave.

[ This Message was edited by: OT on 2002-01-02 22:06 ]

Exam Slave

01-03-2002, 11:41 AM

Companions serve two purposes for PACs, but only one purpose for TACs.

(Or, I'm totally wrong on this. There isn't as much info on TACs as there is for PACs.)

boohoo

01-03-2002, 12:55 PM

I think you two are missing something.

PAC has three things
1 Upper collar or band (ex: 200%PSA)
2. Lower collar or band (ex: 80% PSA)
3. The pricing band (ex: 120%PSA)
(there are other things such as lockout period payment windows etc, let us forget them for a minute)

Pricing band generate the principal
payment schedule.

TACs does not have a lower band (2 in PAC). So if the two collateral backing two securities are identical (i.e. no Zs) then
TAC do not have the protection against extension of WAL for prepayment in the range
bleow the pricing band but above lower band of PAC. If prepayments fall below the lower band of PAC then PAC will be exposed to extension of WAL as same as TAC.
If you put Zs in one structure then whole things is out of whack!

Remember most likely this would be a MC question. In a written answer question it is safer to write everything you know on PAC/TAC as soon as you see these terms.

01-03-2002, 03:24 PM

Thanks, boohoo, for the post.
-------------------
“PAC has three things
1. Upper collar or band (ex: 200%PSA)
2. Lower collar or band (ex: 80% PSA)
3. The pricing band (ex: 120%PSA)
(there are other things such as lockout period payment windows etc, let us forget them for a minute)

Pricing band generate the principal
payment schedule.” by boohoo.

Yes, I got this – but just in case let’s see. Taking boohoo’s example, the minimum of #1 and #2, let’s call it #4, create a kind of prepayment threshold. Pre-payment amounts above this threshold, #4, get spilled over to the companion. This threshold is defined in the contract/debenture. The absorption of the higher pre-payments by the companion protects the PAC from WAL decrease risk or call risk or pre-payment risk. (Here is where my question about the difference between a PAC and TAC lies – how does the TAC exclude WAL extension risk.) (There is also an effective PAC band which is usually a higher upper band because the companion enables a higher prepayment rate in actuality. So in boohoo’s example, the effective PAC band may be 80% to, say, 230%.)

The pricing band is used for just that, pricing (determining current market values). Because pre-payment behavior is unknown we need some pre-payment assumption to price and determine the current market value of the security. This pre-payment assumption is not part of the contract/indenture of the PAC and TAC and is at the discretion of the individual/party doing the pricing. If principal is returned slower than this pricing “schedule” a default of the security does not occur.

But perhaps here is where I am mistaken. After all the text states, “...the typical TAC can be viewed as a PAC with a lower band equal to the CMO pricing speed and an upper band similar to that of PACs backed by comparable collateral.” (page 630, section on TAC) To me it makes more sense if they say something like, the CMO pricing speed pricing speed is set to the TAC’s threshold or lower band.

-------------------
”TACs does not have a lower band (2 in PAC).” by boo hoo.

So then the threshold I speak of is just #1? I think, actually, there is a “range” for the TAC which to me indicates that there is a #2 or lower band.

-------------------
“So if the two collateral backing two securities [PAC and TAC] are identical (i.e. no Zs) then TAC do not have the protection against extension of WAL for prepayment in the range below the pricing band but above lower band of PAC. If prepayments fall below the lower band of PAC then PAC will be exposed to extension of WAL as same as TAC.” by boo hoo

I think because we have a different idea of the pricing speed/assumption this does not make sense to me.

---------------
“If you put Zs in one structure then whole things is out of whack!” by boo hoo [bracketed addition mine].

No, but the inclusion of a Z is, to me, the only way to pay principal to earlier trenches to meet the principal payment schedule when pre-payments are too low (lower than threshold/principal payment schedule) thus causing WAL extension. The inclusion of a Z would make the TAC or PAC an Accretion-Directed Class, I think.

Sorry for the long post. It seems like we are all benefiting from this thread. But thanks for the pas and future input.

[ This Message was edited by: OT on 2002-01-03 15:51 ]