In the U.S., hundreds of billions of dollars of mortgages are issued every year. Mortgages that meet certain government requirements (loan size, LTV ratio, etc.) can be sold to or guaranteed by government backed agencies, including Fannie Mae, Freddie Mac, and Ginnie Mae. Pools of these mortgages are packaged together to create “agency residential mortgage backed securities,” or simply “agency RMBS.” Mortgages that are not guaranteed by government agencies can be packaged up and sold privately as customized securities; these are known as “non-agency residential mortgage backed securities” (or just “NARMBS”). NARMBS are not standardized in any way, and each deal can be different from every other deal. However, in general they are securitized and tranched, whereby junior bonds in the securitization take the first losses and senior bonds are more protected from losses because of the subordination structure.
RMBS, which offer a premium over similarly rated corporate bonds, seek to provide Financial Advisors and their clients higher yields and a steady stream of income over the long term. The following article discusses the benefits of RMBS and why the asset class is becoming a popular choice for Financial Advisors and their clients.
Where is the opportunity in NARMBS?
Simply investing in NARMBS can offer investors an attractive yield. In the case of pre-2008 bonds, we believe the yields are very attractive given the real risk that is taken on relative to other fixed income issues. For the reasons stated above, the credit risk associated with NARMBS is much less than what the perceived risk is. This is despite higher yields on a relative basis to other bonds. While a simple yield from smart investing is good, additional opportunities in this space arise from structural issues with bonds that can be exploited for significant upside. We also believe that there is a misconception that, as the pre-2008 market size shrinks, opportunity in the space is also shrinking. In fact, we believe that the opportunity set is increasing as the size of the market is shrinking.
As mentioned above, each NARMBS securitization is governed by its own prospectus. We also mentioned that each of these structures anticipated losses from losses to underlying mortgages. The documents were written to account for these expected losses. However, the realized severity of losses was not contemplated and very little attention in the governing documents was paid to how should things work in such an extreme environment.
Primary Risks in NARMBS Exposure
No pool of NARMBS is issued with the idea that it will not suffer losses; i.e. losses are expected. The question becomes what are the loss assumptions that you should make when analyzing a bond. As a bond holder, you are exposed to the credit of the underlying borrowers and the fluctuations in market value of the underlying collateral (the houses). Which bonds you acquire, in terms of seniority, determines the amount of risk that you take as an investor. A good way to minimize risk is by primarily owning senior tranches of these structures and by screening securities based on, among other things, delinquency rates.
What makes for good credits in this market space?
Loans issued before 2008 correspond to borrowers who survived the housing crisis and either kept up with their mortgage payments or had them modified to the extent that they are capable of making the payments. When you combine this with the fact that they likely have more equity through scheduled principal payments and appreciation in their home values, these borrowers should be viewed as good credits in terms of their mortgage loans.
While credit ratings might be a good gauge for many corporate bonds, we believe that ratings are not a good guide for judging NARMBS, especially ratings on pre-2008 bonds. During the early stages of the 2008 crisis, ratings agencies either dramatically lowered their ratings or pulled their ratings on these bonds altogether. The credit rating paradigm is not particularly useful for distressed MBS. For example, a bond that has incurred a loss from defaulted mortgage loans will be forever rated as defaulted even if hypothetically the remaining loans in the pool have no credit risk at all. The correct way to view credit risk in MBS is to analyze the credit risk of the loans relative to the remaining bonds structures and the bond price.
When investors first familiarize themselves with corporate and municipal bonds, they quickly learn that credit ratings from S&P, Fitch, and Moody’s are of paramount importance when it comes to pricing bonds and assessing risk. However, the relevance of ratings to non-agency residential mortgage backed securities (NARMBS) could not be more different.
Why wouldn’t a NARMBS investor be interested in whether a security is rated investment grade or junk?
There are two main reasons why an NARMBS investor wouldn’t be interested in whether a security is rated investment grade or junk.
First, the ratings for pre-2008 NARMBS are often stale and irrelevant to current conditions. It is fairly common to find an MBS that experienced heavy losses during the financial crisis, got downgraded by the rating agencies, but went on to perform quite well in the 10+ years following. If the delinquencies are low and the borrowers have been reliable for the last several years, do we really care about the fact that S&P either withdrew its rating or downgraded the security back in 2008?
In fact, the situation may be even better than the history suggests. The borrowers that remain in the securitization after the crisis (and after the ratings downgrade) are likely the stronger borrowers, and as time goes on, they typically build equity in their homes and become even stronger borrowers. Thus, the rating from 2008 becomes less and less relevant as time goes on.
Secondly, NARMBS are often expected to take losses, but this does not mean financial ruin for the bondholders. For example, suppose that there is a bond that we fully expect to take some low level of losses, but it is priced at a significant discount. It could still be a low risk investment in this case. Remember that a low rating simply indicates that the rating agency thinks that there is a relatively high probability that the holder will not receive all the money promised to them. But if you are estimating that you’ll receive 90% of the money owed to you and only paying 80 for the bond, the rating agency’s concern that you will not receive 100% is not terribly important.
This is a common situation with NARMBS, and this is a stark contrast to the corporate bond market. If a corporate bond has a 5% chance of default, an investor might only expect to receive 40 cents on the dollar in those 5% of cases. What is the analogous outcome for an RMBS? An RMBS consists of hundreds of underlying mortgages, and a default could simply mean that the delinquency rate is 7% instead of 6%, and the investor will get 99% of the money they expected. This is hardly a disaster. Moreover, it would be foolish to put this situation in the same bucket as a corporate default.
In other words, defaults for NARMBS are not binary events like they are for corporate bonds. It is reasonable to regard a low-rated NARMBS that trades at a sufficient discount much less risky than a high-rated NARMBS that is priced to perfection. For these reasons, NARMBS should be judged on their structure and the quality of underlying loans; the ratings are not really a consideration when it comes to analysis on whether you want to include a particular bond in your portfolio.
In terms of increasing opportunity, the market is underpricing the bonds of many issuers in the space. In addition, as the market size shrinks, few players remain in the space, leaving investors to find and exploit these issues with less competition.
What are returns expected to look like going forward?
Because of the process involved with identifying problems with particular bonds, acquiring those bonds and resolving the issue can take significant time, we expect the returns to be somewhat lumpy. Trades can take anywhere from a few days to a few years to reach their ultimate payoff. In some instances, there will be some increase in the mark-to-market value of the bonds along the way, but in most cases, there is a single point in time resolution that will be reflected over a short period of time as the market adjusts to the appropriate pricing of the bond or the fund liquidates its position.
It is important to bear in mind that with many event-driven trade ideas, the market has not priced in any likelihood of the desired outcome; therefore, the market is effectively pricing in a 0% chance of the outcome. Unlike other trading strategies, there is a very asymmetrical expected return with no downside, just smaller upside. Either the trade will make 25% or the trade will make a baseline yield. Obviously other market factors can cause losses, but the trade thesis being incorrect will not negatively affect performance.
Finally, we believe that the influx of new securitizations will create additional capacity in the market and will be an attractive relative source of yield for investors. More importantly, we also believe that these new securitizations will have problems and will likely give rise to special situation opportunities (investments in securities that may be mispriced due to structural or market driven factors) in the years to come.