Series: Ask ALM
Investment yield has become top of mind for many financial institutions and their investment managers lately due to the challenges of excess liquidity, net interest margin compression and a low-rate environment.
However, it is critical to remember that simple measures of “yield” can have the unintended effect of keeping investors misinformed as to a portfolio’s true risk/return profile.
As you can see in the chart below, topline yield may be misleading as assets with similar – or even the exact same – yield on the surface may have very different risks, which can impact their actual returns over time. Risk-adjusted spread analysis levels the playing field in the investment universe and allows institutional investors to examine investment selections on a relative value basis.
Expected Performance vs. Realized Performance
Even after you deconstruct a yield into risk/return components and establish a foundation for relative value analysis, it is important to note that much of spread analysis is a model-driven, forward-looking exercise. Therefore, an informed investor should always remember that expected performance may not result in realized performance, or as we used to say “your mileage may vary,” particularly when observed events differ from original inputs and assumptions.
A recent example of this can be seen in very new production mortgage-backed security (MBS) pools trading at high dollar prices as investors struggle to evaluate pre-payments. As uncertainty increases with more complex assets, short-term yields can be drastically different than what models project as the yield to maturity (YTM). While more modeling and better analytics are necessary for more complex assets, diversifying the risk profile of the assets that comprise your investment portfolio is often the best solution as individual securities may vary month-to-month. A diversified portfolio is often able to produce a more consistent rate of return in various environments.
Modeling Option-Adjusted Spread (OAS)
The ability to assess a complex asset’s OAS requires the use of an interest-rate model to value the embedded option and a prepayment model to project future asset cashflows. The appropriate models vary depending on the type of security.
Interest-rate models:
- For mortgages and residential mortgage-related securities, models employ a Monte Carlo simulation of prepayment behavior across potential future interest rate and value the optionality
- For callable bonds, models employ a binominal interest rate tree function to price the option
- Numerous variants of these models exist in the industry, each with its own parameters and assumptions
Prepayment models:
- Multi-factor models based on coupon, size, seasoning, seasonality, burnout, geography, credit aspects of the borrower, etc.
- ZMFS, Bloomberg Prepay Agency Model (BAM), Andrew Davidson, etc.
Projecting yield/OAS for mortgages and mortgage-related assets can be complicated given the significant impact of prepayments on performance, particularly for securities priced well below or above par. That’s why some financial institutions choose to work with an unbiased, strategic partner to assist them with this level of analysis.
Limitations to Expected Return Metrics
As discussed, both yield and OAS metrics are model-dependent and realized returns can vary greatly from expected returns. Yields often don’t translate into actual returns because of changes in market risk factors, most notably prepayments. For example, a high-premium MBS pool’s yield-to-maturity can vary significantly from its short-term dollar return (period coupon – premium amortization) as actual short-term prepayment speeds may differ from expected model speeds. With interest rates falling significantly over the past 6 months, many MBS pools are experiencing faster-than-expected prepayments and corresponding erosion of premiums on those pools.
You Can’t Eat OAS – Dollar Returns will Differ from Model Returns
- A 15yr 3.0% new-production MBS pool can have a model OAS of 50bps and top-line YTM of 1.15%
- However, yield is not linear — aside from the actual return most likely differing from projected, the realization of that return will also be uneven over the life of the pool
- As the pool moves up the WALA ramp, the accelerated amortization of premium can turn realized dollar return negative until the pool’s prepayments slow down and the pool seasons
Variance in Prepayments Significantly Impacts MBS Returns
For a high-premium MBS pool experiencing elevated prepayments speeds, premium amortization can negate much of the coupon collected in the period, resulting in lower-than expected and possibly even negative dollar returns. The model expectation is that prepayment speeds will slow down at some point in the future and that over the life of the investment, the investor is more likely to realize the YTM. However, premium erosion can really hurt an institution’s earnings over the short-term and there is no guarantee that realized prepayment speeds will follow model speeds.
ALM First’s Team of Professionals is Here to Assist You! Contact us to discuss your institution’s investment needs and how partnering with ALM First could benefit your institution.
