Don't get caught holding the bag with CMOs.

In the mid- to late 1980s, noninvestment-grade bonds were sold to thousands of community banks eager for higher-yielding investments.

Many if not most small banks did not possess the internal resources or expertise to properly evaluate the underlying credits of junk bonds, and certain broker-dealers often used these institutions as the buyer of last resort for the more dubious issues.

When the junk bond market collapsed at the beginning of this decade, a significant number of small banks were left holding "the junkiest of the junk," resulting in numerous defaults and bankruptcies.

In response to the junk bond debacle and the havoc it wreaked on bank portfolios, regulatory agencies severely restricted the credit risk exposure of banks and thrifts. Indeed, the regulators' success in curbing credit abuse motivated the broker-dealer community to increase its production of credit-worthy, high-yielding mortgage-backed securities.

Implicit U.S. Guarantee

Mortgage securities now account for over 30% of the entire U.S. bond market and have the implicit backing of the U.S. government, providing investors with virtually no credit risk.

The most prevalent type of mortgage-backed security is the collateralized mortgage obligation, which is composed of an underlying pool of mortgages broken into multiple classes or tranches representing different risk and reward characteristics.

These classes have a predetermined order in which they can be redeemed; i.e., a given tranche cannot be redeemed until all earlier-priority tranches are redeemed.

Option Risk

To further complicate matters, CMOs contain complex options that dictate how the instrument will behave, given the rate environment and rates of mortgage prepayments. Like any risk, option risk can be managed if it is well understood by the portfolio manager.

However, even though few small banks are in a position to competently evaluate and handle option risk, many bank investment managers find CMOs attractive because of their lack of credit risk and relatively high quoted yields.

Thus, it is no surprise that certain broker-dealers look "down the food chain" for yield-hungry community banks to absorb some of the more complex and less marketable tranches that are of no interest to larger institutions with broader, in-house expertise.

Once again, small banks and retail investors have become the buyers of last resort for esoteric and tricky instruments.

Accident Waiting to Happen

This is not conjecture. As providers of portfolio modeling software to bond managers, we are intimately familiar with the composition of community bank portfolios.

All too often, the CMO exposure is high.

And so, just three years after the junk bond crisis seriously compromised hundreds of local banks and thrifts, another accident is waiting to happen.

Historically, the risk of loss was closely tied to the amount of credit (default) risk that the portfolio manager was willing to assume. However, the options embedded in CMOs introduce a new kind of risk that is different from the "event risk" associated with default.

At the Mercy of Rates

Option and credit risk are inherently different in nature. Credit risk is specific to the issue, while option risk is affected by overall interest rate changes. Credit risk can quickly dilute the value of a portfolio. Option risk will dilute the value of the portfolio slowly, over time, like undetected high blood pressure.

In a traditional mortgage or pool of mortgages, there is substantial historical information to derive empirical models that allow the holder to establish the prepayment sensitivity of a mortgage under a host of interest rate changes.

The ability to measure the option risk in CMOs (where it is a function of the underlying mortgage collateral) is also widely available. Measuring the risk does not eliminate it, of course, and the unique and diverse nature of CMO structures and tranches, rearranges and often amplifies the risk of a conventional mortgage.

When CMOs were first developed, investors were generally offered a reasonably clear and distinct choice between low-return, low-risk tranches, and high-return, high-risk tranches.

Tranches Multiply

Unfortunately, and perhaps foreseeably, the market's appetite for ever-increasing rewards in a low-rate environment has encouraged an overzealousness in the tranching of CMOs. In fact, the number of tranches in a single CMO deal has multiplied from four, in the original CMOs, to over 100 at times.

With that many tranches, the relationship between risk and reward can become blurred. The nature of the tranches within a given CMO deal, in a sense, creates a zero-sum game. More reward cannot be created without creating commensurately more risk. Tranches that offer an imbalance favoring reward over risk will have to be offset by other tranches offering an imbalance with the opposite effect.

An example of a "simple" CMO is a Fannie Mae Remic trust offered in 1988, which contained four tranches that on each pay date would pay principal sequentially to the classes in order of their final distribution dates. Those with the earlier distribution dates offered lower reward/ lower risk. Those with the later distribution dates offered higher reward/higher risk.

Complex Freddie Mac Offering

But deals like the Freddie Mac Multi-class Mortgage Participation Certificates (PC), Series 1539, issued Jan. 1, 1993, are far more complicated.

For example, on each payment date, Freddie Mac will pay, in "strip classes," 1/7,501 of the PC principal amount concurrently to the R and RS classes, pro rata; and the remainder of the PC principal amount, as follows:

First, to the Type I PAC Classes, until each has been reduced to its "targeted balance" for that payment date, allocated as follows:

Concurrently to the PA and PB classes in the ratio of 143 to 16, while the PA class is outstanding, and then concurrently to the PB and PC classes in the ratio of 89 to 46, and then -- pro rata -- to five more Type I classes before payouts to Type II classes, and their components, commence.

The complexity results from the effort to induce greater option risk. Indeed, a careful analysis of this CMO might lead a dispassionate observer to conclude that the terms are merely there to create the option risk, often in a way that appears to serve no useful purpose.

This does not make life easy for portfolio managers. Moreover, ascertaining the risk-reward tradeoff exposes the portfolio manager to two major pitfalls: misestimation of the collateral sensitivity and misestimation of the terms.

Once each of these modeling conditions is satisfied, the portfolio manager must test the risk-return characteristics of a selected tranche under a set of general interest rate changes.

For reprint and licensing requests for this article, click here.
MORE FROM AMERICAN BANKER