After strong performance in the agency mortgage-backed securities (MBS) market in 2014, the recent decline in U.S. interest rates has resulted in material underperformance of MBS relative to U.S. Treasuries. For several years, agency mortgages have been the beneficiary of tremendous Federal Reserve balance sheet expansion, limited interest rate volatility and strong demand for high-credit-quality assets offering a yield spread over U.S. Treasuries. With the Fed having ended net purchases of MBS in October 2014 and a material decline in U.S. Treasury yields, investors need to consider what the shift in supply and demand means for agency MBS.
With the 10-year Treasury note recently falling to 1.64% and now hovering around 2%, the mortgage market is the most negatively convex it has been in years, which has significant implications for mortgage investors. The term negative convexity is classic bond geek jargon, but in simplified terms it means that a bond goes up in price slower than its spot duration implies when interest rates decline and goes down more than its spot duration when interest rates rise. In MBS, negative convexity stems from the option that borrowers have to prepay the underlying fixed-rate mortgages: When interest rates fall, borrowers typically begin to refinance their mortgages and the bonds backed by these mortgages prepay earlier than expected, and when rates rise, borrowers no longer have incentive to refinance, which causes the bonds to extend in maturity. January 2015 was a textbook example of the negative convexity in MBS:
- 10-year U.S. Treasury prices increased by 4.76% and yields declined by almost 50 basis points.
- 30-year FNMA 4.0% MBS began the year with a duration of 2.85 years.
- While spot durations on 31 December 2014 implied that FNMA 4.0% prices should be approximately 1.43% higher in price, they are actually up just 0.02%.
Why is this? We would attribute it primarily to three key developments in MBS: 1) rising prepayment speeds, 2) higher supply and 3) lack of a “value-agnostic” Fed balance sheet. Below, we delve into each topic in greater detail.
Three key trends in MBS
Rising prepayment speeds. As rates move lower, an increasing amount of the mortgage universe can be refinanced. This is contrary to what investors became accustomed to after the “taper tantrum” in mid-2013 sent mortgage rates more than 100 basis points higher and removed the incentive for a majority of mortgage borrowers to refinance. The limited refinancing led many investors to revise their MBS models, reflecting a lower likelihood of mortgages prepaying. We believe models will need to be recalibrated back toward historical levels.
Higher supply. Increasing prepayments not only impact mortgage bond cash flows, but also drive new-issue supply higher. After a year of surprisingly low refinancing and mortgage origination activity in 2014, we believe supply is likely to surprise investors to the upside given current mortgage rate levels. After less than $1 trillion of gross agency MBS issuance in 2014 (the lowest amount since 2007), many investors revised their supply estimates for 2015 to historically low levels. Should issuance merely revert to historically normal levels, the new supply will likely put further pressure on MBS valuations in the coming months. We believe the result could be detrimental to MBS performance: Cash flows will be deteriorating at the same time that supply is increasing.
Lack of a “value-agnostic” Fed balance sheet. While there have been periods of volatility in the MBS market in recent years, most have been short-lived, as the Fed’s balance sheet was expanding during its asset purchase program and supporting prices regardless of valuations or market developments. With the Fed ending net purchases of MBS, this is no longer the case. While reinvestments by the Fed will remain a tailwind for the sector, with each month that passes, we are closer to the end of reinvestment as well. The end of reinvestment will likely leave MBS structurally cheap for an extended period. As a result, the sense of complacency that has dominated the MBS market in recent years could quickly dissipate, and the Fed will not be the buyer of last resort this time around.
Who will buy?
Against this backdrop, the important question becomes: Who can step in to take the Fed’s place in agency MBS in 2015?
All of the new issuance the market will soon face will need to find buyers, and we anticipate fewer buyers at today’s lower yields and tight spreads. For much of the post-crisis environment, banks and, to a lesser extent, mortgage REITs (real estate investment trusts), were net buyers of MBS, while non-U.S. investors have been steady sellers.
Banks and overseas investors are typically what we would characterize as “yield investors,” buying bonds when rates are relatively high and selling when rates are relatively low. Given the low level of interest rates and widespread consensus that the Fed will initiate rate hikes at some point in 2015, buyers who traditionally utilize the agency MBS market for yield are likely to wait on the sidelines for better entry points. This is especially important because the banking sector owns approximately 27% of all agency MBS and was one of the largest net buyers of MBS in 2014. Banks typically purchase fewer MBS and retain fewer loans (increasing securitization rates) when interest rates are on the lower end of a perceived range.
Demand from non-U.S. investors for MBS is also likely to wane at current low interest rates, because these buyers are yield-sensitive. Following the financial crisis, overseas investors have been allowing their MBS holdings to wind down through amortization (receiving principal and interest payments and not reinvesting). As principal payments on MBS increase with refinancing, these investors could effectively become even larger net “sellers” of MBS simply by not reinvesting. Counting on yield-based investors to fill the Fed’s shoes as supporters of MBS seems unlikely at current spreads and yields.
Mortgage REITs, the third largest non-Fed supporter of MBS in the post-crisis environment, are also not in a position to be material buyers in the current environment. Most mortgage REITs are trading below book value in the secondary market, making it difficult for them to raise additional capital to deploy into MBS. In addition, rising levels of implied volatility are causing much of the mortgage REIT sector to keep leverage low, resulting in limited demand for MBS.
The bottom line is that the favorable environment of Fed support, benign prepayment speeds and positive technicals that dominated the MBS market in 2014 appears to be rapidly shifting in 2015. As is typically the case in sectors that have exhibited exceptionally low volatility, many investors have stretched for carry in agency MBS, often using leverage, and the recent volatility suggests these trades are in the early stages of being unwound.
The implications for MBS
We have become accustomed to a mortgage-backed securities market that can remain stable irrespective of valuations due to unprecedented Fed purchases of MBS, but the changing environment should not be taken lightly. With the Fed no longer a net buyer, a modest reversion of issuance to historically normal numbers would result in a surprising amount of new MBS supply in the marketplace – with fewer buyers at current interest rate and spread levels.
“Keep calm and carry on” was 2014’s theme, but “back to reality” might be the theme for 2015. At PIMCO, we are cautiously navigating the changing environment by operating from a core underweight to MBS, but capitalizing on the relative value opportunities that continue to arise as volatility returns to the market.