White Paper | by Dalton Investments | 12.15.15
As volatility persists in the interest rate environment, structured mortgage-backed securities (“MBS”) backed by adjustable-rate mortgages look attractive to many investors. Take for example discounted senior floating rate securities backed by adjustable-rate mortgages of the 2005-2008 vintage. A simple mathematical calculation implies that as interest rates rise, the coupon on these bonds should see a larger proportional increase due to their discounted nature. Or maybe not.
A lot has changed since 2008, and investors in so called “money good” (paying 100% of principal) and top of the capital structure floating rate bonds, may be disappointed to find that their investments actually perform worse in an increasing rate environment.
Years of intense loan modifications by mortgage holders, especially at the low quality end of the credit spectrum, have made major MBS pools, which at issuance were primarily composed of adjustable-rate collateral, fundamentally more fixed in nature. These loan modifications are greatly affecting the duration, weighted average life and expected spread of these securities. Since the Financial Crisis, the risk profile of these investments has changed to a degree that most managers will be hard pressed to understand. Fortunately, they also provide opportunity for smaller managers, like Dalton, who are nimble and opportunistic, and thus have a better chance to navigate this risk.
Take for example, MABS 2006-AM2. This deal is structured where the AAA bonds start paying pro-rata once losses run through the subordinate bonds and hit the AAA. At issuance in 2006, only 11.27% of the pooled mortgage loans were fixed rate. Today, 74% of the loans are fixed. The fixed rate percentage is artificially high because 44% of loans have been modified – typically to rates that are far below market conventions.1
These fixed-rate modifications cause multiple problems in a higher rate environment; the biggest being the decrease of the available excess spread between the collateral coupon and bond coupons, normally used to pay down principal in the front pay bond (the A3). Therefore, the front pay bond receives a slower principal paydown, creating maturity extension risk, and leaving it with a higher balance, plus leaving open the potential of taking more pro-rata principal losses.
According to Dalton’s analysis, bonds further down the capital structure actually perform much better in an increased rate environment. We believe that the front pay A3 bond that pays LIBOR +23bps, has an expected yield of 5.51%, 11.22yr weighted average life (“WAL”), 4.02yr modified duration, and 379bps spread in our base runs. If we shock LIBOR+100bps, we believe that the yield falls 4.90%, the WAL increases to 14.14yrs, the duration increases to 4.51yrs, and the spread widens to 206bps over.
The last cash flow bond actually performs much better than the current pay bond under the same conditions, according to Dalton’s analysis. Our base case run produces a 5.44% yield, 20.75yr WAL, 8.3yr modified duration and 331 basis point spread. At an L+100 shock, the yield increases to 6.42%, the WAL increases to 21.52yrs, the duration decreases to 7.15yrs, and the spread retreats to 326bps over.
Meanwhile, the M1 bond, which we consider a credit IO because we don’t expect it to receive any principal, is very attractive (at an assumed price level) in an increasing rate environment. We model this bond to a 1.75yr WAL, 1.36yr modified duration, and 53bp spread. At an L+100 shock, the WAL and modified duration decrease, but the spread increases to 5,744bps. We started our strategy at Dalton investing in this type of cashflow and we believe that at some point this could again be a very attractive investment opportunity.
In our research, Dalton has been seeing this phenomenon a great deal in subprime, alt-a, and pay-option arm deals. We also see this demonstrated in floating rate prime deals of later 2006-2007 vintages, as they are more likely to hold a larger percentage of modified loans. We’ve already seen senior bonds and seasoned subordinate “money good” bonds trade significantly off over the past few months. The reason for this is that portfolio managers did not price in the proper risk premium associated with the changes in the characteristics of the underlying mortgages. These deals pose serious risk for larger mandates that are forced to buy a large portfolio, but provide opportunities for smaller, more agile managers who have experience navigating the risks.