CHAPTER 12
COLLATERALIZED MORTGAGE OBLIGATIONS AND STRIPPED MORTGAGE-BACKED SECURITIES
CHAPTER SUMMARY
In this chapter we discuss two derivative mortgage-backed securities products: collateralized mortgage obligations and stripped mortgage-backed securities. These securities derive their cash flow from underlying mortgage collateral, such as pass-throughs or a pool of whole loans.
COLLATERALIZED MORTGAGE OBLIGATIONS
Collateralized mortgage obligations (CMOs) are bond classes created by redirecting the cash flows of mortgage-related products so as to mitigate prepayment risk. The mere creation of a CMO cannot eliminate prepayment risk; it can only transfer the various forms of this risk among different classes of bondholders.
CMO Structure
A CMO is security backed by a pool of pass-throughs, whole loans, or stripped mortgage-backed securities. CMOs are structured so that there are several classes of bondholders with varying stated maturities. When there is more than one class of bondholders with the same level of credit priority, the structure is called a pay-through structure, as opposed to a pass-through structure in which there is only one class of bondholders at a given level of credit priority. The bond classes created are commonly referred to as tranches.
Sequential-Pay Tranches
The first CMO was created in 1983 and was structured so that each class of bond would be retired sequentially. Such structures are referred to as sequential-pay CMOs.
A CMO is created by redistributing the cash flow—interest and principal—to the different tranches based on a set of payment rules. There are separate rules for the payment of the coupon interest and the payment of principal, the principal being the total of the regularly scheduled principal payment and any prepayments.
Each tranche receives periodic coupon interest payments based on the amount of the outstanding balance at the beginning of the month. The disbursement of the principal, however, is made in a special way. A tranche is not entitled to receive principal until the entire principal of the preceding tranche has been paid off.
The principal pay-down window for a tranche is the time period between the beginning and the end of the principal payments to that tranche. Tranches can have average lives that are both shorter and longer than the collateral, thereby attracting investors who have a preference for an average life different from that of the collateral.
There is considerable variability in the average life for the tranches. However, there is some protection provided for each tranche against prepayment risk. This is because prioritizing the distribution of principal (i.e., establishing the payment rules for principal) effectively protects the shorter term tranche against extension risk. This protection must come from somewhere, so it comes from the other tranches. At the same time these other tranches are provided protection against contraction risk.
Accrual Bonds
In many sequential-pay CMO structures, at least one tranche does not receive current interest. Instead, the interest for that tranche would accrue and be added to the principal balance. Such a bond class is commonly referred to as an accrual tranche, or a Z bond (because the bond is similar to a zero-coupon bond). The interest that would have been paid to the accrual bond class is then used to speed up the pay down of the principal balance of earlier bond classes. Thus, the average lives for the nonaccrual tranches would be shortened as a result of the inclusion of accrual tranche.
The accrual bond has appeal to investors who are concerned with reinvestment risk. Because there are no coupon payments to reinvest, reinvestment risk is eliminated until all the other tranches are paid off.
Floating-Rate Tranches
Floating-rate tranches can be created from fixed-rate tranches by creating a floater and an inverse floater. We can select any of the tranches from which to create a floating-rate and an inverse floating-rate tranche. We can even create these two securities for more than one of the four tranches or for only a portion of one tranche.
Any reference rate can be used to create a floater and the corresponding inverse floater. There is an infinite number of ways to cut up the monetary value between the floater and inverse floater, and the final partitioning will be driven by the demands of investors.
Unlike a floating-rate note in the corporate bond market, whose principal is unchanged over the life of the instrument, the floater’s principal balance declines over time as principal payments are made. The principal payments to the floater are determined by the principal payments from the tranche from which the floater is created.
Assume that the reference rate is the one-month LIBOR of 3.75%, then the coupon rate on the inverse floater takes the following form: K – L (one-month LIBOR) where K is the cap or maximum coupon rate for the inverse floater, and L is the multiple that determine the coupon rate for the inverse floater. (L is referred to as the coupon leverage .) If K is set at 28.50% and L at 3, then the coupon rate for the month is: 28.50% – 3(3.75%) = 17.25%. The higher the coupon leverage, the more the inverse floater’s coupon rate changes for a given change in one-month LIBOR.
Inverse floaters with a wide variety of coupon leverages are available in the market. Participants refer to low-leverage inverse floaters as those with a coupon leverage between 0.5 and 2.1, medium-leverage as those with a coupon leverage higher than 2.1 but not exceeding 4.5, and high-leverage as those with a coupon leverage higher than 4.5.
As in the case of the floater, the principal pay-down of an inverse floater will be a proportionate amount of the principal pay-down of the bond class from which it is created.
Because the reference rate (e.g., one-month LIBOR) is always positive, the coupon rate paid to the floating rate bond class cannot be negative. If there are no restrictions placed on the coupon rate for the inverse floater, however, it is possible for the coupon rate for that bond class to be negative. To prevent this, a floor, or minimum, can be placed on the coupon rate. In many structures, the floor is set at zero. Once a floor is set for the inverse floater, a cap or ceiling is imposed on the floater.
The cap for the floater and the inverse floater, the floor for the inverse floater, the coupon leverage, and the margin spread are not determined independently. Given four of these variables, the fifth will be determined.
Planned Amortization Class Tranches
The CMO innovations attracted institutional investors who had previously either avoided investing in mortgage-backed securities or allocated only a nominal portion of their portfolio to this sector of the fixed-income market.
Potential demand for a CMO product with less uncertainty about the cash flow increased in the mid-1980s. In March 1987, the M.D.C. Mortgage Funding Corporation CMO Series 0 included a class of bonds referred to as stabilized mortgage reduction term (SMRT) bonds; another class in its CMO Series P was referred to as planned amortization class (PAC) bonds. The Oxford Acceptance Corporation III Series C CMOs included a class of bonds referred to as a planned redemption obligation (PRO) bonds. The greater predictability of the cash flow for these classes of bonds, now referred to exclusively as PAC bonds, occurs because there is a principal repayment schedule that must be satisfied.
The greater certainty of the cash flow for the PAC bonds comes at the expense of the non-PAC classes, called support or companion bonds. It is these bonds that absorb the prepayment risk. Because PAC bonds have protection against both extension risk and contraction risk, they are said to provide two-sided prepayment protection.
Although there is no assurance that the collateral will prepay between selected PSA speeds, a PAC bond can be structured to assume that it will. The two speeds used to create a PAC bond are called the initial PAC collars (or initial PAC bands).
Most CMO PAC structures have more than one class of PAC bonds. From a PAC bond, we can create other bonds with average lives that are stable and also where all average lives are either much shorter or longer. Even if prepayments are faster than the initial upper collar, there may be sufficient support bonds to assure the average life is unchanged. The degree of protection against extension risk increases for shorter PAC bonds. The effective collar can be wider than the initial collar for shorter PAC tranches.
A PAC window can be wide or narrow. The narrower a PAC window, the more it resembles a corporate bond with a bullet payment. PAC buyers appear to prefer tight windows, although institutional investors facing a liability schedule are generally better off with a window that more closely matches the liabilities. Investor demand dictates the PAC windows that issuers will create. Investor demand in turn is governed by the nature of investor liabilities.
As we have emphasized several times, the creation of a mortgage-backed security cannot make prepayment risk disappear. This is true for both a pass-through and a CMO. Thus the reduction in prepayment risk (both extension risk and contraction risk) that a PAC offers must come from somewhere.
Where does the prepayment protection come from? It comes from the support bonds. It is the support bonds that forego principal payments if the collateral prepayments are slow; support bonds do not receive any principal until the PAC bonds receive the scheduled principal repayment. This reduces the risk that the PAC bonds will extend. Similarly, it is the support bonds that absorb any principal payments in excess of the scheduled principal payment that are made. This reduces the contraction risk of the PAC bonds.
Thus the key to the prepayment protection offered by a PAC bond is the amount of support bonds outstanding. If the support bonds are paid off quickly because of faster than expected prepayments, there is no longer any protection for the PAC bonds.
The support bonds can be thought of as bodyguards for the PAC bondholders. When the bullets fly (i.e., prepayments occur) it is the bodyguards that get killed off first. The bodyguards are there to absorb the bullets. When all the bodyguards are killed off (i.e., the support bonds paid off with faster than expected prepayments), the PAC bonds must fend for themselves: they are now exposed to all the bullets.
Busted means that the prepayment protection is reduced. It is the term used in the CMO market when a PAC schedule is broken. The initial collars are not particularly useful in assessing the prepayment protection for a seasoned PAC bond. This is most important to understand, as it is common for CMO buyers to compare prepayment protection of PACs in different CMO structures, and conclude that the greater protection is offered by the one with the wider collar. This approach is inadequate because it is actual prepayment experience that determines the degree of prepayment protection as well as the expected future prepayment behavior of the collateral.
The way to determine this protection is to calculate the effective collar for a seasoned PAC bond. An effective collar for a seasoned PAC is the lower PSA and the upper PSA that can occur in the future and still allow maintenance of the schedule of principal repayments.
The effective collar changes every month. An extended period over which actual prepayments are below the upper range of the initial PAC collar will result in an increase in the upper range of the effective collar. This is because there will be more bodyguards around than anticipated. An extended period of prepayments slower than the lower range of the initial PAC collar will raise the lower range of the effective collar. This is because it will take faster prepayments to make up the shortfall of the scheduled principal payments not made plus the scheduled future principal payments.
The PAC schedule may not be satisfied even if the actual prepayments never fall outside the initial collar. This may seem surprising because our previous analysis indicated that the average life would not change if prepayments are at either extreme of the initial collar. However, our previous analysis has been based on a single PSA speed for the life of the structure.
There are two ways to provide greater protection for PAC bonds: lockouts and reverse PAC structures. One obvious way to provide greater protection for PAC bonds is to issue fewer PAC bonds relative to support bonds. Such a CMO structure with no principal payments to a PAC bond class in the earlier years is referred to as a lockout structure. A CMO structure requiring any excess principal payments to be made to the longer PAC bonds after all support bonds are paid off is called a reverse PAC structure.
Targeted Amortization Class Bonds
A targeted amortization class (TAC) bond resembles a PAC bond in that both have a schedule of principal repayment. The difference between a PAC bond and a TAC bond is that a PAC bond has a wide PSA range over which the schedule of principal repayment is protected against contraction risk and extension risk. A TAC bond, in contrast, has a single PSA rate from which the schedule of principal repayment is protected. As a result, the prepayment protection afforded the TAC bond is less than that for a PAC bond.
The creation of a bond with a schedule of principal repayments based on a single prepayment rate results in protection against contraction risk but not extension risk. Thus, whereas PAC bonds are said to have two-sided prepayment protection, TAC bonds have one-sided prepayment protection.