You are on page 1of 27

Agency MBS Outlook 2014: Uncertainty is Certain

December 5, 2013
FOR US DISTRIBUTION ONLY
Looking Back: 2013 was a very interesting and very bifurcated year for the Agency MBS Market. Many of the favorable post-crisis market trends lower rates, heightened demand for high quality assets, relatively muted speeds (due to credit constraints), and accommodative monetary policy persisted into 1H 2013. Decent demand from both traditional (banks, REITs, money managers) and non-traditional (the Fed) investors generally exceeded supply and narrowed spreads. However, the tides began to turn in May 2013 as concerns of a Fed taper materialized, sending rates and volatility higher and cutting off virtually all non-Fed demand. Looking Ahead: We expect 2014 market dynamics to closely reflect those seen during the latter half of 2013. As we linger in this higher rate range, we forecast light supply (which should be absorbed in large part by the Fed, even in an eventual tapering), and even lighter demand (from traditional MBS accounts that will be primarily focused on short duration paper, despite the recent steepening). Despite this decent technical, the overall extension of the MBS universe is a less compelling fundamental. Moreover, we believe that over the next 12 months, investors will have to navigate through numerous uncertainties, including all forms of policy risk: monetary, fiscal, economic and GSE. Even as we write this outlook, the pivotal November payrolls number (December 6) and December FOMC meeting (December 17-18) outcomes are unknowns. While these uncertainties will likely manifest via increased volatility and wider spreads, we actually think that any underperformance could provide some opportunities down the line. In the pages that follow, we 1) highlight the main issues that will affect investors next year, 2) discuss our issuance and supply forecasts, and 3) conclude with our best ideas in Agency MBS. Specifically:

Invitation to Consider a Derivatives Transaction This communication is prepared by the sales and trading desk and is marketing material and/or trader commentary. It is not a product of the research department. This material constitutes an invitation to consider entering into a derivatives transaction under U.S. CFTC Regulations 1.71 and 23.605, where applicable, but is not a binding offer to buy/sell any financial instrument. The views of the author may differ from others at The Royal Bank of Scotland plc, The Royal Bank of Scotland N.V. and/or RBS Securities Inc. (collectively "RBS").

Other important disclosures can be found on the last page of this publication.

We are neutral on the mortgage basis. Technicals are decent and valuations appear
fair versus historicals but both arguably carry less weight in a policy-driven landscape.

Within TBA, we like adding the belly on weakness and GNMA vs. Conventionals. We
currently find 15 year TBA rich to alternatives (but generally think short duration makes sense in the upcoming environment).

Within Specified Pools, we like the lower pay-up stories (loan balance 3.5% and 4% in
TBA eligible, CQ/CR in non-deliverable), though we acknowledge the extension risk.

In Short Duration space, we prefer stable CMO PACs and SEQs and 20 year paper (if
you can find it) as substitutes for 15 year pass-throughs. ARMs look a little less cheap to us, but still provide decent relative value.

We believe the best relative value can be achieved by giving up some liquidity.
Specifically, we like the highest LTV paper (CQ/CR) and HREMIC floaters and IOs.

Contacts: Jeana Curro, Director Co-Head of Agency MBS Strategy +1 203 897 4650 jeana.curro@rbs.com

Ashley Gam Agency MBS Strategy +1 203 897 1617 ashley.gam@rbs.com

www.rbs.com/mib

Agency MBS Outlook 2014: Uncertainty is Certain

PART I. Uncertainty Is Certain: Key Issues for 2014


Going forward, investors will be challenged by a great deal of uncertainty, the majority of which is policy related. As time passes and some of the following questions are resolved, we anticipate that investors will become more active again. Below are some of the key issues currently perplexing Agency MBS investors: 1. Monetary Policy Uncertainty: As early as May 2013 (when Bernanke stated during a Q&A from his JEC testimony that the Fed could step down asset purchases in the next Fed meetings), taper speculation permeated the MBS market. For most of the summer, market participants dissected Fed language and economic data to try and get a heads up on when the Fed would taper and by how much. Just when investors felt confident that they had the answers to those questions, the market was shocked when no taper was announced in the September meeting. We believe that monetary policy will be the number one uncertainty plaguing the markets into next year. As per our economists,1 the most likely outcome at present is as follows (although there is again no certainty in this). While we cannot rule out a December taper, we still think the odds favor the Fed scaling back its asset purchase program in March. December presents a challenge in that markets are illiquid, and a tapering announcement could lead to outsized market moves similar to the type that unnerved the Fed in the early summer. With the January meeting essentially a changing of the guard as Janet Yellen (most likely, now that she only needs 51 votes) takes over on February 1, as well as the fact that no press conference is scheduled following that meeting, we lean towards March as the most opportune time to taper. Importantly, we think that the Fed will combine a tapering announcement with enhanced forward guidance that pushes expectations for the first rate hike into 2016. As the Fed has stressed the importance of economic recovery in helping determine monetary policy, we think the upcoming payrolls number will be increasingly important in shaping tapering views. With respect to the mechanics of tapering, we continue to think that the Fed will scale back about $20 billion each quarter, evenly divided between mortgages and Treasuries, winding down the asset purchase program at the end of 2014, although we would acknowledge that there is a risk that the program could be scaled back somewhat faster, perhaps winding down by Q3 2014. 2. Fiscal Policy Uncertainty: 2013 was rife with fiscal policy uncertainty and as many of the issues were not then resolved, we expect 2014 to be somewhat similar. Our economists speculate that the odds of another government shutdown have lessened following the budget battles of the fall. That being said, much will depend on the outcome of the budget conference, which is set to present a plan to Congress on December 13 on how to fund the government beyond the January 15, 2014 deadline, when funding expires. Currently, indications are that the budget conference is moving towards a modest deal that would replace portions of the 2014 sequester with other targeted cuts in spending, as well as set a top-line figure for discretionary spending through the balance of Fiscal Year 2014 (which ends on September 30, 2014). If negotiators are unable to reach a deal, then prospects for a shutdown on January 15 would certainly rise. However, both parties need a "win," with Republicans eager to avoid being blamed for another shutdown, and Democrats on the defensive following the botched rollout of the Affordable Care Act. If a deal is reached prior to January 15, then we believe the debt limit will be

With thanks to Omair Sharif, Michelle Girard and Guy Berger

Page 2/27

Agency MBS Outlook 2014: Uncertainty is Certain

raised prior to February 7 and will likely extend past the November mid-term elections. 3. Rate Uncertainty: Our rates strategists2 predict that over the next year we will see 1) gradually higher rates, with 10s at 3.15% by year end, and 2) slightly steeper curves, with 2s10s rising to 270bps. This outlook, however, is expectedly dependent upon various known unknowns such as the timing and composition of tapering, potential changes to forward guidance, how the Yellen Fed will operate, etc. The forecast for steeper curves is also based on doubts over the Feds credibility, another area of uncertainty that necessitates a higher risk premium. Finally, we expect tapering to contribute to heightened volatility in the long end and muted in the front end, at least for as long as the Feds forward guidance remains credible. Supply and Demand Uncertainty: The Fed is by far the largest single holder of Agency MBS securities, currently owning ~$1.35T or 25% of the universe and scheduled to hold $1.55T or 27% by year end (chart below). Naturally, there is a tremendous amount of uncertainty that revolves around when this program will end. In addition to when the Fed is going to taper and by how much (the aforementioned monetary policy uncertainty), the market is also concerned with how to replace them. It is unclear who will provide the backstop bid once the Fed steps away. We anticipate that the higher rates and wider spreads that could result from tapering may lure back some interest from previously sidelined, currently cash heavy accounts, but this remains to be seen and any said demand is unlikely to be supplied by another single investor of Fed size. Another uncertainty (that is less discussed in our opinion) is where the remaining supply will come from while the Fed continues asset purchases. With issuance expected to fall sharply in 2014 (more on this in next section) and several investor classes (banks, REITs, money managers, to name a few) having already reduced their overweights (now flat or underweight MBS), it is difficult to see an obvious source of future supply.

4.

Fed Agency MBS Holdings and Market Share


Source: CPRDR, RBS

1,600 1,400 1,200 1,000 $ Billions 800 600 400 200 Jun-12 Feb-12 Feb-13 Jun-13 Aug-12 Aug-13 Dec-11 Dec-12 Dec-13 Feb-14 Oct-11 Oct-12 Oct-13 Apr-12 Apr-13

30 25 % Fed of Outstanding 20 15 10 5 0

$FedHold (LHS)

%FedHold (RHS)

While these supply and demand uncertainties are certainly prevalent, we actually believe that tapering is already priced in to some extent. We do expect to see a knee-jerk widening, as the market tends to reward for confirmation, but then we would expect spreads to fall back in. Recall back to the widely anticipated QE3 announcement in September 2011, where nominal spreads tightened ~40bps over 10 days before
2

With thanks to William ODonnell, John Briggs and Gabe Mann. Please see their Rates Outlook, out later this month.

Page 3/27

Agency MBS Outlook 2014: Uncertainty is Certain

reversing 35bps in the subsequent 5 days. Moreover, because of supply uncertainty, a tapering may even be welcomed as net Fed volumes ($40B/mo) have stayed steady despite a drop in new issuance; so far this year the Fed has absorbed roughly half of gross issuance and 150% of net issuance. Assuming that tapering is announced in the March FOMC meeting (essentially effective starting April), the first quarter should still benefit from full Fed demand taking out ~90% of gross issuance and over 200% of net given the expected decline in issuance over 2014, and the subsequent months may also be minimally affected. Below we show that in the event of a March taper, even tapered purchase volumes would exceed net issuance for at least a while. If tapering is in line with our economists estimates ($10B/quarter), the Feds presence would persist throughout 2014, with the Fed still absorbing the majority share of net issuance throughout the year. If tapering takes a more aggressive pace (e.g. a gradual reduction of say $5B/mo, completed in October), then the Fed will represent a majority presence through Q3 2014. We elaborate on the factors contributing to our light issuance forecast in the next section. Light Supply in 2014; Effects of Fed Tapering Likely to be Offset by Drop in New Issuance
The Current State: Fed % of Issuance ($Billions, %)
2012 Sep Gross Issuance Fed Gross Purchases Fed % of Gross Net Issuance Fed Net Purchases Fed % of Net 155 24 15% 19 40 Oct 138 46 33% -15 40 Nov 201 64 32% 57 40 Dec 132 74 56% -16 40 Jan 161 71 44% 18 40 Feb 155 80 52% 25 40 Mar 151 82 54% 20 40 Apr 155 77 50% 24 40 May 151 70 46% 18 40 2013 Jun 149 71 48% 29 40 Jul 144 65 45% 33 40 Aug 131 67 51% 41 40 Sep 106 68 64% 34 40 Oct 94 65 69% 25 40 Nov 94 60 64% 25 40 Dec 94 60 64% 25 40 2013 10m 1,396 715 51% 267 400 150% 2013 FY 1,584 835 53% 318 480 151%

208% -261%

71% -250% 226% 162% 196% 164% 224% 138% 122%

98% 116% 159% 159% 159%

Taper Scenario 1: Issuance Down, Taper $10B/qtr


2014 Jan Gross Issuance Fed Gross Purchases Fed % of Gross Net Issuance Fed Net Purchases Fed % of Net 70 60 86% 18 40 222% Feb 70 60 86% 18 40 222% Mar 70 60 86% 18 40 222% Apr 70 50 71% 18 30 May 70 50 71% 18 30 Jun 70 50 71% 18 30 Jul 70 40 57% 18 20 Aug 70 40 57% 18 20 Sep 70 40 57% 18 20 Oct 70 30 43% 18 10 56% Nov 70 30 43% 18 10 56% Dec 70 30 43% 18 10 56% 2015 Jan 70 20 29% 18 0%

167% 167% 167% 111% 111% 111%

Taper Scenario 2: Issuance Down, Taper $5B/month


2014 Jan Gross Issuance Fed Gross Purchases Fed % of Gross Net Issuance Fed Net Purchases Fed % of Net 70 60 86% 18 40 222% Feb 70 60 86% 18 40 222% Mar 70 60 86% 18 40 222% Apr 70 55 79% 18 35 May 70 50 71% 18 30 Jun 70 45 64% 18 25 Jul 70 40 57% 18 20 Aug 70 35 50% 18 15 83% Sep 70 30 43% 18 10 56% Oct 70 25 36% 18 5 28% Nov 70 20 29% 18 0%

194% 167% 139% 111%

Source: New York Fed, CPRCDR, RBS. RBS Estimates in Italics.

As we enter into the next year, we believe that reduced supply will serve as a decent technical, compensating for reduced demand from traditional investors. However, offsetting this market-positive development is the fact that mortgage market fundamentals are arguably less favorable. Weighing the technicals versus the fundamentals will be a common theme that persists throughout our outlook for next year.

Page 4/27

Agency MBS Outlook 2014: Uncertainty is Certain

5.

Policy-Related Prepayment Uncertainty (Watt risk): The recent change allowing certain presidential nominees to be confirmed with a simple majority vote has made Mel Watts appointment as the next FHFA head virtually a done deal. This change in leadership opens the door for borrower-friendly mortgage relief, either in the form of HARP extension or principal forgiveness3. Of the two, HARP expansion would likely have the larger impact on Agency MBS (as it would target current borrowers, i.e. loans that currently lie in MBS pools), but the borrower incentive (and therefore MBS market impact) should be minimized by any backup in rates. Conversely, while a principal forgiveness program would most likely target delinquent loans (i.e. loans that have already been pulled from conventional pools), the MBS market impact could be magnified should such a program trigger a wave of strategic defaults. This is exactly the kind of uncertainty that is currently sidelining many investors and contributing to very light flows, particularly in higher coupon pass-throughs and mortgage derivatives. GSE Reform Uncertainty: The RBS view remains that GSE reform will take a very long time. With the GSEs currently returning sizeable profits to Treasury (the GSEs have repaid a combined ~$185B to Treasury, versus ~$187B in total draws), there is less urgency to quickly wind them down. Moreover, we believe that Reids recent use of the nuclear option to change filibuster rules has created more discord in an already divided Congress, rendering some of the more aggressive GSE reforms highly unlikely in the near term (i.e. those that would require Congressional action, such as completely winding them down altogether, or merging them to form a new GSE). While the next steps for the GSEs are less clear given the upcoming change in FHFA leadership, we expect that attention on GSE reform will continue. We anticipate that the GSEs will most likely in simple terms expand on what works and continue to evaluate what does not4. For example, the success of Freddie and Fannies respective STACR and CAS deals demonstrated that the market has some appetite for selective credit risk; going forward we thus expect the GSEs to continue issuing these or similar risk-sharing securities (i.e. senior/sub deals). Exactly how much they will have to issue is currently unknown as the conservatorship goals for 2014 have not yet been released (the 2014 Scorecard is in the works at present), but the target will likely be greater than it was in 2013 ($30B UPB for each GSE). Finally, we estimate that the completion of more complex goals, such as full development of the Common Securitization Platform, will likely take a much longer time.

6.

PART II. Issuance, Supply and Demand


Issuance Forecasts:
All signs point to lighter Agency MBS Issuance in the near term. For full year 2014, we estimate that issuance will be $840B gross and +$225B net. This decline in new supply represents a sharp contrast to the recent past (time series below). As a refresher, 2013 YTD (through October) gross new issuance has been $1.44T (FN, FH, GN; Fixed and ARM) and is on track to reach roughly $1.6-1.7T by year

The Impact of HARP expansion was explored in depth by our Prepayment Strategy team. For details and estimates, please see Potential Watt Confirmation and HARP Expansion: Impact on Prepayments found here: https://strategy.rbsm.com/Tools/Content/ContentViewer.aspx?ContentID=370925 The most recent GSE progress report can be found here: http://www.fhfa.gov/webfiles/25835/ProgressReportrelease112513.pdf
4

Page 5/27

Agency MBS Outlook 2014: Uncertainty is Certain

end. Net issuance has been roughly $242B YTD and will likely reach ~$260B by year end. Over the first few months of 2013 we saw very similar themes as we did in 2011-2012: positive HPA (we are on track for +10.5% for full year 2013), a continuous decline in mortgage rates, a measureable refinancing cycle (which was the predominant driver of gross new issuance), lenders increasing capacity and the Agency market dominating mortgage finance, with roughly 90% of new originations. Over the second half of 2013, the exact opposite was true: the rise in rates caused refinances to drop dramatically and lenders underwent waves of headcount reductions on expectations of lighter supply. We think these trends are likely to persist into 2014, contributing to a near 50% reduction in gross issuance. Expect Gross Issuance to Fall Substantially, Net to Fall Only Marginally in 2014
2,500 2,250 2,000 1,750 1,500 1,250 1,000 750 500 250 0 -250 2,117

1,721

1,731 1,216

1,429

$ Billions

1,003

961

912

1,146

1,152

1,400

1,729 2013 (est) 2014 (est)


200 150 bps Incentive 100 50 0 -50 -100 -150 Sep-13

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

Gross New Iss

Net New Iss

Source: CPRDR, RBS. 2013 Data known through October 2013 (Gross: $1.441T, Net: +$242B).

For 2014, we envision an environment somewhat similar to the years of 2005-2006. At this time, the aggregate mortgage universe was at a discount, with the average gross WAC roughly 25-75bps below the prevailing market rates; we show a history of the universe with prepayment speeds in discount (shaded) and premium (unshaded) environments below. We note that the most recent data shows near zero rate incentive; should rates move higher, we could potentially enter into an environment where the majority of the MBS market is again at a discount. Prepayment History of Mortgage Universe: Premium and Discount Environments
70 60 50 1-mo CPR 40 30 20 10 0 Mar-95 Mar-96 Mar-97 Mar-98 Mar-99 Mar-00 Mar-01 Mar-02 Mar-03 Mar-04 Mar-05 Mar-06 Mar-07 Mar-08 Mar-09 Mar-10 Mar-11 Mar-12 Sep-94 Sep-95 Sep-96 Sep-97 Sep-98 Sep-99 Sep-00 Sep-01 Sep-02 Sep-03 Sep-04 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-10 Sep-11 Sep-12 Mar-13

Universe at a Discount

CPR (LHS)

Incentive (RHS)

Source: The Yield Book, RBS. As of the close on Thursday, 10/31/13.

Page 6/27

840

Agency MBS Outlook 2014: Uncertainty is Certain

Refi vs. Purchase Share of Conventional Applications


Source: Bloomberg, RBS

100% 80% 60% 40% 20% 00% '02 '03 '04 '05 '06 '07 '08 '09 '10 '11 '12 '13 % refi % purchase

During the 05-06 period, purchase activity was the primary driver of new production. In the conventional space, refinances were roughly 40-45% of new loans whereas purchases were 55-60%5. For perspective, in 2013 the % refi share of applications was as high as 85% in January but is now roughly 65%. In GNMA product during 05-06 (not shown), refinances were roughly 30-40% of total issuance, whereas purchases were 6070% (while this trend seems sensible, it is worth noting that GN market share was very small during this period, as the subprime market was just beginning to flourish). Going forward, we expect net issuance to again be GNMA driven as the Conventional markets are gradually shrinking. Overall, 2014 gross estimates will represent an approximate 50% decline from 2013 levels, and while net issuance estimates are positive, we expect them to also be lower than 2013 aggregate volumes, by roughly 15%. Again, for full year 2014 we estimate that issuance will be $840B gross and +$225B net. The mixed bag of factors guiding our issuance estimates are as follows:

MBA Refi Index Sharply Lower


Source: Bloomberg, RBS

Less Refis / More Purchases: Just as we saw in 2005-06, we think that the share of
8000 6000 4000 2000 0

6.0% 5.5% 5.0% 4.5% 4.0% 3.5% 3.0% '09 '10 '11 '12 '13 Freddie PMMS Rate (LHS) Refi Index (RHS)
Source: Bloomberg, RBS.

purchases will ramp up over 2014. 2104 is also less likely to see the heavy refinance supply that we saw in 1H 2013 as we have rebounded off of historically lower mortgage rates. The MBA refi index has plummeted from its May 2013 highs (chart left), illustrating that borrowers have become generally less responsive to small rallies. Effect on Issuance: Largely Negative For Gross, Slight Positive For Net.

Higher G-Fees: With the last g-fee hike occurring in January 2013, we expect the
GSEs will likely raise fees again in 2014. We estimate a ~10-20bps increase over 2014, bringing fees to the 65-80bps area (chart left). During Q3 20136, Fannie Mae charged on average 59bps g-fee, and Freddie Mac charged roughly 53bps (net of MAP adjustments). Any repricing could gradually transfer supply into the private label market (and thus contribute less to gross and net issuance), although we expect a very gradual shift. Effect on Issuance: Slight Negative for Gross and Net.

'14

Growth of Private Label Market: We still expect the government presence in


Expect Higher G-Fees in 2014
Source: FHFA, RBS. *2013 through Q3, 2014 est.

90 80 70 60 50 40 30 20 10 0 2007 2008 2009 2010 2011 2012 *2013 *2014

mortgage finance to remain high but this role will likely shrink gradually as the private label market exhibits the potential to ramp up. The likely higher g-fees will contribute to private label growth. In addition, eventual decreases in GSE loan limits should help shift some market share (FHFA has decided to leave loan limits (conforming and high cost) unchanged for 2014 although stating Further information on potential future changes in the maximum size of loans that Fannie Mae and Freddie Mac guarantee will be forthcoming.7). Additionally, other factors that should at a minimum help private label execution include: First, more participants have gotten up and running to issue and securitize loans. Second, QM and QRM definitions have been finalized, providing more clarity for lenders. Third, the rating agencies have published their criteria for new issue, which should remove some uncertainty from the securitization process. Finally, the success of the GSE credit deals (STACR, CAS) indicates that the market has appetite for credit sensitive bonds. Some issues still remain in private label land, however, which is why we think growth here should continue to be gradual. Our base case estimates for PL issuance are in the $25B ballpark, which is a measureable increase over this years estimated $13B although far from the $1.1 Trillion peak in

bps

Conventional numbers based on MBA Mortgage Application Data. As per company filings: http://www.fanniemae.com/portal/about-us/media/financial-news/2013/6039.html http://freddiemac.mwnewsroom.com/press-releases/freddie-mac-reports-third-quarter-2013-financialr-otcqb-fmcc-1066084
6 7

http://www.fhfa.gov/webfiles/25847/CLL2014112613Final.pdf

Page 7/27

Agency MBS Outlook 2014: Uncertainty is Certain

Expect Private Label Growth


Source: Inside MBS & ABS, RBS. *2014 RBS estimate

2005-06 (chart below left, 2007 and earlier not shown for scale). Effect on Issuance: Slight Negative for Gross and Net.

Rising HPA: Home prices are projected to increase further next year, although at a
more gradual pace than 2013. RBS estimates +4% for 2014, vs. +10.5% for full year 2013. Rising rates plus rising home values will likely hurt home affordability slightly, which is why 2014 gross estimates are smaller than 2005-2006 numbers. Effect on Issuance: Positive for Gross, Slightly Less Impact on Net.

30 25 20 15 10 5 0 2008 2009 2010 2011 2012 2013*2014* Subprime Alt-A

$ Billions

Increased Loan Retention: Banks will likely continue keeping the best credit loans
on balance sheet (instead of securitizing) because loans generally receive favorable capital treatment when retained on balance sheet; securitizing and holding just the servicing may be less popular as MSRs have stricter capital treatment. Also super prime loans may be worth more if retained on balance sheet, when compared to proceeds from securitization. Effect on Issuance: Slight Negative for Gross and Net.

Prime

Underwriting Standards Have Loosened Somewhat


Source: Ellie Mae, RBS

755 750 745 740 735 730 725 720 May-12 May-13 Aug-12 Aug-13 Nov-12 Feb-13

83 82 81 80 79 78 77 76

Tight Underwriting: As a result of banks retaining the best credit loans, we may see
some adverse selection on what gets pooled into new issuance. Therefore recent trends may reflect a slight easing in credit via lower FICOs and slightly higher LTVs at origination that cannot entirely be attributable to the success of the HARP program. However, we believe that overall underwriting standards remain strict and despite the most recent trend, they are far from the standards used during the housing boom. Effect on Issuance: Slight Negative for Gross and Net.
LTV

FICO

FICO (LHS)

LTV (RHS)

Declining Delinquencies: The chart (left margin) shows that the seriously delinquent
rates for single family conventional loans have been steadily declining. We expect these rates to drop further with rising HPA and any other signs of economic improvement. Lower delinquencies have two important implications for net issuance: first, fewer loans will be removed from securities due to buyouts; second, securitization of reperformers could actually add to MBS outstanding. Effect on Issuance: Slight Positive for both Gross and Net.

Seriously Delinquent Rates for Fannie Mae and Freddie Mac Single-Family Mortgages
Source: Fannie Mae, Freddie Mac, RBS
6% 5% 4% 3% 2% 1% 0% Jul-08 Jul-09 Jul-10 Jul-11 Jul-12 Jul-13

Capacity Constraints: Lenders have also been actively downsizing given the backup
and future rate uncertainty. The mortgage finance employment index is a lagging indicator but shows the start of a recent downturn. Additionally, there have been several headlines confirming headcount reductions among top mortgage lenders. Thus we expect this trend to continue. Effect on Issuance: Slight Negative for Gross and Net.

FNMA

FHLMC

Supply Forecasts:
Mortgage Finance Employment: The Start of a Downtrend
Source: US Bureau of Labor Statistics, RBS

Thousands of Employees

305 295 285 275 265 255 Mar-09 Mar-10 Mar-11 Mar-12 Sep-08 Sep-09 Sep-10 Sep-11 Sep-12 Mar-13 Sep-13

In addition to net new issuance, there are other notable components of supply. Specifically, outright selling as well as not reinvesting pay downs are both contributors. Once weve estimated the net selling, we examine whether demand will be sufficient to absorb that amount plus $225B in net new issuance. We conclude that in 2104, estimated demand of $335B should just exceed the estimated supply of $305B. With demand exceeding supply, this technical looks very similar to what we observed in 2H 2013. However as weve also observed over 2H 2013, favorable technicals may not be enough to override other factors. Below, we show the top holders of Agency MBS over the past few years and include our best estimates for 2014.

Employment in Real Estate Credit + Mortgage and Nonmortgage Loan Brokers (NSA)

Page 8/27

Agency MBS Outlook 2014: Uncertainty is Certain

Top Holders of Mortgage Related Securities


$ Billion Investor Type Commercial Banks US Treasury/NY Fed Mutual Funds Foreign Investors GSEs Life Insurance Cos. REITs Investor Type Commercial Banks US Treasury/NY Fed Mutual Funds Foreign Investors GSEs Life Insurance Cos. REITs Dec-14 1225 1780 675 710 290 265 252 Dec-14 2 1 4 3 5 6 7 Dec-13 1225 1480 650 700 340 250 280 Dec-13 2 1 4 3 5 7 6 Jun-13 1248 1208 710 645 368 245 311 Jun-13 1 2 3 4 5 7 6 Dec-12 1,274 927 724 720 407 257 348 Rank Dec-12 1 2 3 4 5 7 6 Jun-12 1 2 3 4 5 6 7 Dec-11 1 2 4 3 5 6 7 Jun-11 1 2 5 3 4 6 7 Jun-12 1,276 858 638 595 431 315 263 Dec-11 1,209 866 580 605 492 335 202 Jun-11 1,138 1,010 550 615 568 490 185

Source: Inside MBS & ABS, RBS, Federal Reserve, Fannie Mae, Freddie Mac Notes: Mortgage-related securities include all securities or debt obligations collateralized by residential mortgages or MBS. Estimates are in italics and from Inside MBS & ABS and RBS

Mortgages underperformed other high-grade asset classes for most of 2013, even showing negative total returns in the second quarter (chart below). Should MBS performance continue to lag, we would expect the more traditional accounts to continue taking a backseat to the Fed in the early months of 2014 (we assume QE purchases will continue at their current pace until March 2014). Though spread widening on an eventual tapering could potentially lure back some sidelined money, we anticipate that demand (outside of the Fed) will be light overall until some uncertainties are cleared up. Historical Index Total Return: Treasury, Agency, Mortgage and IG Corporate
Source: The Yield Book (Broad Investment Grade Indices), RBS.

12 10 8 6 4 2 0 -2 -4 Q1 10 Q2 10 Q3 10 Q4 10 Q1 11 Q2 11 Q3 11 Q4 11 Q1 12 Q2 12 Q3 12 Q4 12 Q1 13 Q2 13 Q3 13 USBIG, Treasury USBIG, Industrial, AAA/AA USBIG, Agency USBIG, Finance, AAA/AA US BIG, 30-Year FNMA USBIG, Agency, Callable

Page 9/27

Agency MBS Outlook 2014: Uncertainty is Certain

With this in mind, we estimate that total net supply (net new issuance + net selling) will sum to around $303B, which falls short of our net demand estimate for $335B (which is largely based on Fed buying and a small degree of real money adding). In the below table we provide 2014 demand estimates for the key investor classes. We elaborate on each of these estimates in the subsequent paragraphs. 2014 Supply and Demand Estimates ($ Billions)
Net Demand 300 25 10 0 Net Supply

US Treasury/NY Fed Money Managers Foreign Investors Commercial Banks REITs GSEs Net Issuance TOTAL
Source: RBS

335

0 28 50 225 303

The Fed (+$300B): The Feds asset purchase program is expected to trickle into 2014, but the key unknowns are how much and for how long. Over 2013, the Feds gross purchases ($40B/mo + reinvestments) have so far amounted to $734B. Net of reinvestments, they are on track to purchase $480B for the full year 2013. For 2014, if we assume that QE continues as is until March and then gradually tapers by $10B/qtr in MBS, the Fed will take down an additional $300B (net). Here we assume that the Fed will continue to reinvest pay downs and that purchases will likely continue to target production coupons. Of course, it is feasible for the Fed to reduce their presence more sharply by tapering more aggressively; for example, a $5B/mo MBS reduction beginning in March would imply net activity of only ~$260B for 2014 and would culminate in October. Fed Gross Purchases in Base Case: $10B/qtr Taper Beginning Fed Net Purchases in Base Case: $10B/qtr Taper Beginning in in March 2014 March 2014
Source: NY Fed, RBS Source: NY Fed, RBS

90,000 80,000 70,000 60,000 ($mm) 40,000 30,000 20,000 10,000 0 May-13 May-14 May-15 Mar-13 Mar-14 Sep-13 Sep-14 Nov-13 Nov-14 Mar-15 Jul-13 Jan-13 Jan-14 Jul-14 Jan-15 ($mm) 50,000 estimates

90,000 80,000 70,000 60,000 50,000 40,000 30,000 20,000 10,000 0 May-13 May-14 Mar-13 Mar-14 Jul-13 Sep-13 Jul-14 Sep-14 Nov-13 Nov-14 Jan-13 Jan-14 estimates

2.5

3.5

4.5

Page 10/27

Agency MBS Outlook 2014: Uncertainty is Certain

Fed Gross Purchases in Aggressive Case: $5B/mo Taper Beginning in March 2014
Source: NY Fed, RBS

Fed Net Purchases in Aggressive Case: $5B/mo Taper Beginning in March 2014
Source: NY Fed, RBS

90,000 80,000 70,000 60,000 ($mm) ($mm) 50,000 40,000 30,000 20,000 10,000 0 May-13 May-14 Mar-13 Mar-14 Jan-13 Jan-14 Sep-13 Sep-14 Nov-13 Nov-14 Jan-15 Jul-13 Jul-14 estimates

90,000 80,000 70,000 60,000 50,000 40,000 30,000 20,000 10,000 0 May-13 May-14 Mar-13 Mar-14 Jul-13 Sep-13 Jul-14 Sep-14
Jul-13

estimates

Nov-13

2.5

3.5

4.5

Money Managers (+$25B): Our best estimates for 2013 are that money manager Agency MBS holdings are down roughly $75B from 2012. We believe they sold early into Fed demand and then again when faced with redemptions during and after the summer sell-off. As the overall returns of MBS have lagged other asset classes for 2013, we would expect that most money managers are currently flat to the index (versus being overweight going into last year). Spreads for Agency MBS are currently tighter overall versus AAA corporates, but we think a widening could lure some sidelined interest back into the Agency MBS market. The challenges to this theory are 1) if more redemptions occur and 2) if the spread widening trend extends to other asset classes, they may look more attractive than Agency MBS. In the yield grab this year, we saw decent flows out of Agency MBS into Corporates and Equities (which set historic highs several times over 2013). Overall we optimistically estimate that money managers will add roughly $25B. Historical Index Yields: Treasury, Agency Mortgage and IG Corporate
Source: Bloomberg, RBS

Historical Index Spreads: Treasury, Agency Mortgage and IG Corporate


Source: The Yield Book, RBS

4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Jul-10 Jul-11 Jul-12 Oct-10 Oct-11 Oct-12 Jul-13 Jan-10 Jan-11 Jan-12 Jan-13 Oct-13 Apr-10 Apr-11 Apr-12 Apr-13

250 200 150 100 50 0 -50 Jan-10 Jul-10 Oct-10 Jul-11 Oct-11 Jul-12 Oct-12 Jan-11 Jan-12 Jan-13 Oct-13 Apr-10 Apr-11 Apr-12 Apr-13

USBIG, Treasury US BIG, 30-Year FNMA USBIG, Finance, AAA/AA

USBIG, Agency USBIG, Industrial, AAA/AA

USBIG, Treasury US BIG, 30-Year FNMA USBIG, Finance, AAA/AA

USBIG, Agency USBIG, Industrial, AAA/AA

Mortgage REITs (-28B): While steeper curve environments are generally aligned with REIT investment strategies, we expect REITs to be quiet in 2014 and if anything skewed

Page 11/27

Nov-14

Jan-13

Jan-14

Agency MBS Outlook 2014: Uncertainty is Certain

towards better selling. This expectation is in sharp contrast to what we have seen over the past few years (but again, consistent with the most recent activity). Up until mid 2013, REITs were actively raising capital and deploying it out the curve. However in the four months starting in mid-May, REIT stocks fell from multi-year highs to multi-year lows, and have only partially recovered since (below left). Capital raises essentially ceased shortly after the start of the backup (below right). Mortgage REIT Stocks Just Above August Lows
Source: Bloomberg, RBS

Public Residential Mortgage REIT Issuance of Equity


Source: NAREIT, RBS. 2012 data through October 31, 2013.

410 MVMORT Index Value 390

16,000 14,000 12,000 $ millions 10,000 8,000 6,000 4,000 2,000 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Feb-11 Feb-12 Aug-11 Aug-12 Dec-10 Feb-13 Aug-13 Dec-11 Dec-12 Apr-11 Apr-12 Jun-11 Jun-12 Apr-13 Jun-13 Oct-11 Oct-12 Oct-13

370 350 330 310 290

Full Year

Q1

Q2

Q3

Q4

As we expect 2014 market dynamics to closely reflect 2H 2013, we think REITs will continue to reduce leverage and preserve capital. Below we show some indicators for mortgage REITs over 2013; we also include data points for American Capital (AGNC) as it is one of the larger Agency MBS REITs and in our opinion, a good representative of the sector. The patterns indicate declining stock prices, lower price: book ratios and reduced leverage. We believe these trends are apt to continue in 2014. While a spread widening may potentially recapture the sidelined interest of other investors, we expect that REITs will behave more conservatively. REIT Summary: Lower Prices and Decreased Leverage in 2013
2013 MVMORT Index Capital Raises ($ Millions) AGNC Stock Price AGNC Price:Book Ratio AGNC Leverage Ratio Q1 404 4,628 31.7 1.1 8.1 Q2 336 1,073 23.1 0.9 8.5 Q3 323 287 22.6 0.9 7.2 Q4 312 0 19.9 0.8 7.0

Source: Bloomberg, NAREIT, RBS. Capital raises based on publically traded mortgage REITs that invest in MBS. Q4 MVMORT Index value as of 11/29/13, all other Q4 numbers are RBS estimates.

The predominant trend over the past few months has been one of defensive positioning. Specifically, REIT outright selling has subsided, and several REITs have instead been reallocating into shorter-duration instruments (notably ARMs, 15-year paper, and some coupon swaps). The silver lining here is that REITs should be better hedged going into next year so a backup of similar magnitude as what we saw this summer could potentially generate less selling than before. Our best estimates for REIT activity over 2014 is a 10% reduction, or net selling of approximately $28B. Domestic Banks (flat): Bank demand has been the wildcard for some time now as banks have been awaiting regulatory clarity (in addition to monetary, fiscal and GSE

Page 12/27

2013

Agency MBS Outlook 2014: Uncertainty is Certain

reform clarity). MBS allocations have been flat for most of 2013, but recently trended lower. On the contrary, cash allocations have really ramped up over 2013 (chart below). Assets for Large US Banks: Cash Holdings Increase Dramatically Over 2013
Source: The Federal Reserve, RBS

1200 1000 800 600 400 200 0 Jan-02 Jan-03 Jan-04 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-10 Jan-11 Jan-12 Jan-13 Jul-02 Jul-03 Jul-04 Jul-05 Jul-06 Jul-07 Jul-08 Jul-09 Jul-10 Jul-11 Jul-12 Jul-13

$ Billions

Agency MBS

Tsy and Agy Securities

Other (Non-Govt) Securities

Cash Assets

Going into 2014, we expect bank demand to be light to slightly negative in the early year, but could potentially resurface as some of the prevailing uncertainties are solidified. Banks certainly have the cash available to deploy once levels widen, but how they will use their cash is uncertain. We think bank investments will more or less fit the bill of defensive investments that comply with the upcoming Basel regulations. On the defensive front, we could see demand for higher cap floaters and stable CMO cash flows off of good convexity collateral. Currently shorter amortization collateral is well bid and we expect as the 15-year sector cheapens, it may be used more in CMOs. PACs tend to be a natural fit for a volatile rate environment given the dual-sided protection offered by the bands (we discuss the relative value of PACs versus other short duration instruments in the next sector). Regarding regulatory guidance, better credit investments typically benefit from lower capital ratios and more favorable liquidity treatment. The capital requirement (which was announced in July of this year) removed the AOCI filter for large banks that adopt the advanced approach. In short, we anticipate that this will skew bank demand towards shorter-duration securities (e.g. CMO floaters). The liquidity coverage ratio assigns more favorable treatment to top credit securities (i.e. GNMA MBS, US Treasuries, Cash and Reserves) and second level classification to Fannie Mae and Freddie Mac MBS. The Federal Reserve estimates that banks in aggregate are currently shy of being fully LCR compliant by about $200B in high quality liquid assets. While the above arguments certainly support bank demand at wider levels, we acknowledge that said buying may actually be more reflective of re-allocating and less reflective of outright adding. Moreover, the general trend of holding unsecurtized loans or cash may continue. Only if the backup is substantial enough are we likely to see decent interest resume and sidelined cash be put to work. Because of these conflicting factors, our best estimate is for bank activity over 2014 to be unchanged from 2013 levels. Foreign Demand (+$10B): Foreign investors have been more or less quiet the past year with holdings little changed. During 2013, we estimate that holdings fell roughly $20B; for 2014 we could foresee a slight adding of ~$10B should rates back up enough. However, overseas interest is largely dependant on confidence in the US markets. With so much

Page 13/27

Agency MBS Outlook 2014: Uncertainty is Certain

uncertainty currently affecting the Agency MBS market, particularly surrounding rates and monetary policy, it is unclear how overseas demand will play out. GSEs (-$50B): The GSEs are mandated by the Senior Preferred Stock Purchase Agreements to reduce their portfolio holdings. Treasury upped this amount from 10%/year (of the pre-set portfolio limit) to 15%/year. Now for full year 2014, both Fannie and Freddie need to bring their retained portfolios down to $474.03B max. Currently, Freddies portfolio is nearly there (last $482.67B as per the October 2013 monthly summary), while Fannies has a bit more to go (last $504.83B as per October 2013 monthly summary). With FRE currently less than $10B away from the 2014 limit, they could even add to their PC holdings and still meet the 15% aggregate, which could be a very slight positive for gold/fn swaps (very light). Both GSE portfolios (with history) are shown in the table below. GSE Portfolio Trends (Dollars in Billions)
Fannie Mae 2008 2009 2010 2011 2012 2013 2014 limits Jan 13 Feb 13 Mar 13 Apr 13 May 13 Jun 13 Jul 13 Aug 13 Sep 13 Oct 13 Total $787.29 $772.50 $788.77 $708.41 $633.05 $504.83 $474.03 $620.51 $604.29 $597.78 $587.96 $574.85 $565.20 $547.05 $531.28 $516.26 $504.83 FN MBS $287.57 $358.19 $260.43 $220.06 $183.96 $134.17 $121.90 $182.99 $172.49 $170.21 $165.83 $159.17 $151.83 $147.93 $142.73 $137.24 $134.17 Other Agy $34.85 $42.65 $18.76 $15.55 $12.32 $9.62 $9.14 $11.92 $11.64 $12.41 $10.65 $10.81 $11.02 $10.57 $9.86 $9.78 $9.62 Non-Agy $99.62 $90.43 $82.51 $74.53 $65.06 $38.53 $36.60 $64.32 $63.76 $63.16 $62.25 $59.51 $58.61 $52.35 $46.74 $40.83 $38.53 Loans $365.25 $281.24 $427.07 $398.27 $371.71 $322.52 $306.40 $361.28 $356.40 $352.00 $349.23 $345.36 $343.74 $336.13 $331.96 $328.41 $322.52 Total $804.76 $755.27 $696.87 $653.31 $557.54 $482.67 $474.03 $550.04 $542.73 $534.15 $528.26 $518.50 $521.22 $521.25 $511.94 $497.81 $482.67 FH PC $424.52 $374.62 $263.60 $223.67 $186.76 $180.80 $187.25 $183.84 $182.03 $178.66 $177.32 $175.22 $184.64 $188.87 $190.82 $183.28 $180.80 Freddie Mac Other Agy $70.85 $66.17 $40.13 $33.04 $23.77 $18.72 $17.79 $23.03 $22.12 $21.48 $20.81 $20.14 $20.14 $20.48 $20.70 $20.96 $18.72 Non-Agy $197.91 $175.67 $158.39 $142.63 $125.69 $97.78 $92.89 $124.17 $122.95 $121.50 $119.73 $116.91 $114.72 $111.46 $106.87 $102.80 $97.78 Loans $111.48 $138.82 $234.75 $253.97 $221.31 $185.38 $176.11 $219.01 $215.63 $212.51 $210.41 $206.22 $201.73 $200.44 $193.55 $190.77 $185.38 Portfolio Max $900.00 $810.00 $729.00 $656.10 $557.69 $474.03

Source: Fannie Mae, Freddie Mac. RBS estimates are in italics. 2014 Portfolios are hard limits; 2014 estimates for Other Agency, Non-Agency and Loans represent 5% decrease from current levels (and are not hard limits); FN MBS and FH PC represent the difference between portfolio limits and all other categories.

In the past, the GSEs could reach their portfolio limits by run-off alone as pay downs were not reinvested. With rates reversing trajectory, portfolio run-off is likely to decline; however, given how close the GSEs are currently to their limit, we think portfolio run-off could indeed suffice. Additionally, Freddie Mac has been reducing loans on balance sheet by securitizing some new products, which include reperforming (R) pools, fixed rate modified (M) pools, and step up modified (H) pools. To the extent that there is investor demand for these assets, we would expect the GSEs portfolios to shrink well past the max levels.

PART III. Best Investment Strategies in Agency MBS


The recent curve steepening (which our rates strategists predict will hold well into next year, chart below) has not yet motivated investors to move out the curve. Instead, and especially in light of the heightened policy uncertainty, investors prefer short duration, defensive investments.

Page 14/27

Agency MBS Outlook 2014: Uncertainty is Certain

2 year 10 year Curve Expected to Stay Steep into Next Year


Source: RBS. Estimates in Light Blue.

350 300 250 200 bps 150 100 50 0 Jul-10 Jul-11 Jul-12 Jul-13 Apr-10 Apr-11 Apr-12 Apr-13 Apr-14 Oct-10 Oct-11 Oct-12 Oct-13 Jul-14 Jan-10 Jan-11 Jan-12 Jan-13 Jan-14 Oct-14

We think these kinds of strategies should hold up well into next year. For 2014, some of our key investment themes are as follows:

Dont Fight the Fed: Although choppy, the market has been extremely responsive to
Fed activity, outperforming on steady buying and underperforming on hints of tapering. Even as demand is expected to subside, the Fed will likely remain the largest buyer (and holder) of Agency MBS for some time. Also Janet Yellen in particular is a strong advocate of enhanced communications and transparency as tools in the Feds arsenal. That said, the market should have even more language to take direction from once she is confirmed as Fed Chair.

Best Convexity is Key: As market volatility is likely to persist, we like securities that
should perform well over a range of rate scenarios. We also believe that the heightened volatility will present many attractive entry points to add this kind of paper in the upcoming year.

Given Rate Uncertainty, Look to Short Duration: This trade has been favored for
some time now. Although the overall duration of the Agency MBS market is longer than in previous years, we find that there are several pockets of short duration relative value (notably front end CMO cash flows and shorter amortization pass-throughs).

For Yield Seekers, Giving Up Some Liquidity may be Preferable to Extending


Duration: With interest rate risk still palpable, we think many investors would rather stay short than move far out the curve. Furthermore, the prevailing policy risk tends to mostly affect the TBA market. GN HMBS is a great example of a security that is small in market share but is relatively unaffected by monetary policy, GSE reform etc and instead offers increased yield at relatively short duration, complete with government backing. The remainder of this document focuses on investment strategies for the next year, keeping the above themes in mind. We are neutral on the basis but think there are some opportunities within TBA pass-throughs. Next we highlight best-convexity strategies that should do well in a volatile rate environment (specified pools, 20 year paper, short PACs). Finally, for investors looking for enhanced returns, we highlight relative value in asset classes that are notably less liquid (e.g. 105+ LTV paper, GN HREMIC floaters and IOs).

Relative Value In TBA Pass-throughs:


Neutral on Agency MBS Basis: Technicals Favorable, Valuations Mixed and Fundamentals Less Certain: Given the mixed indicators surrounding Agency MBS, we are neutral on the basis overall. We think the technicals are actually favorable, with Fed demand likely to compensate for bank, REIT and money manager inactivity and more

Page 15/27

Agency MBS Outlook 2014: Uncertainty is Certain

importantly, likely to exceed supply even in the event of a tapering. However, fundamentals possess some uncertainty, as the overall duration of the MBS market is longer and prepayments may ultimately be policy driven, creating hedging challenges. Current valuations are also mixed, with nominal spreads tight but LOAS numbers generally in line with 2010-2011. Given the many mixed messages in Agency MBS, we are neutral on the basis. We are hesitant to underweight them (a taper delay could cause some light outperformance) but we do not necessarily want to overweight them either. We like holding a neutral stance until some of the aforementioned uncertainties are removed. CC Nominal Spreads Have Tightened
Source: Bloomberg, RBS

But CC LOAS Valuations Look OK


Source: The Yield Book, RBS

300 250 200 150 100 50 Mar-09 Mar-10 Mar-11 Mar-12 Mar-13 Nov-08 Nov-09 Nov-10 Nov-11 Nov-12 Nov-13 Jul-09 Jul-10 Jul-11 Jul-12 Jul-13

60 40 20 LOAS 0 -20 -40 -60 Mar-09 Mar-10 Mar-11 Mar-12 Nov-08 Nov-09 Nov-10 Nov-11 Nov-12 Mar-13 Jul-09 Jul-10 Jul-11 Jul-12 Jul-13

CC5/10

05-07 "Pre-Crisis" Avg

LOAS

2005-07 "Pre-Crisis" Avg

Coupon Stack: Like Adding Belly Coupons on Weakness: Relative Value in TBA conventional 30 years is proving to be more of a challenge now than in the recent past. Deciding where to be in the coupon stack is no longer synonymous with taking a view on prepayments, as investors must now consider policy as well. For example, the basis and lower coupons are currently most affected by Fed involvement while the higher coupons are essentially independent of monetary policy and are instead the most susceptible to Watt risk. In fact, with the lower coupons out of the money and the higher coupons still credit constrained, one can argue that normal rate response is relatively low on the list of investor concerns. Parametric Analysis: 30 Year TBA Pass-throughs
Cpn FNCL FNCL FNCL FNCL FNCL FNCL 3.0 3.5 4.0 4.5 5.0 5.5 Price Settle 12/11 12/11 12/11 12/11 12/11 12/11 WAC 3.59 4.01 4.46 4.94 5.71 6.08 MBS Collat Info Yield Book WAM Age LTV FICO Yield WAL 354 349 334 353 297 322 3 8 21 45 95 72 72 75 74 73 70 73 765 757 752 756 725 726 3.54 3.44 3.32 2.68 1.51 0.67 9.5 8.7 7.2 4.1 2.7 2.1 1mo 2.4 3.7 8.1 24.3 34.9 41.1 Proj CPR 3mo 2.9 4.2 8.1 21.2 34.3 41.1 LT 6.5 7.6 10.1 19.9 29.7 38.1 E Dur 8.14 7.25 5.96 3.69 2.20 1.35 Optional Measures E Cnvx LOAS -0.11 -0.41 -1.29 -2.81 -1.31 -0.77 32 27 28 20 1 -12 O Cost 20 25 31 47 44 48

3 95-30 3.5 100-11 4 103-30 4.5 106-16 5 108-21+ 5.5 109-15

As of the close on 12/3/13. Source: The Yield Book, RBS.

All things considered, we like positioning in the belly coupons (3.5s, 4.0s and to a lesser extent 4.5s) and we would underweight the discounts (3.0s) and premiums (5.0s and higher). This stance is partially related to positive valuations (parametric analysis shown above) but also heeds upcoming policy concerns. Specifically, our thoughts on the coupon stack are motivated by the following:

Page 16/27

Agency MBS Outlook 2014: Uncertainty is Certain

Belly Coupons (Like 3.5, 4s; Neutral on 4.5s):


3.5s are almost at par and have decent liquidity (though new issuance is lighter than in 4.0s, which have become the true production coupon). 3.5s are fully extended and given the trajectory of rates, unlikely to pose serious contraction risk. As a result, they currently have very low negative convexity which is reflected in its larger LOAS. 4.0s are the current production coupon and thus have the best liquidity. 4.0s also benefit from strong technicals (Fed buying that exceeds supply) which we expect to continue at least in the near term. 4.0s do have some premium risk (103 handle) but essentially no HARP risk, and currently have high LOAS. 4.5s currently receive good Fed demand and liquidity is decent, with monthly issuance volumes now very close to the size we see in 3.5s. However there is a notable inflection point moving UIC from 4.0s to 4.5s: Spreads in 4.5s have recently tightened and the LOAS drop from 4.0s to 4.5s (-8bps) is noteworthy. Also: 4.5s have the most negative convexity across the stack and are currently ~ 50bps in the money (fully rate refinanceable). Finally, 4.5 (along with 5%) are the most susceptible to HARP expansion risk. Given the mixed indicators, we are neutral on 4.5s at these rate levels, however we may be inclined to add in a large enough backup.

Discount Coupons: We are generally negative on 3.0s despite large LOAS numbers.
We expect them to be slow but priced accordingly. Currently 3.0s are modeled at nearfull extension, and have a 9+ year WAL and an 8+ year effective duration in the base case. Supply in 3.0s has essentially shut off and they run the risk of being orphaned.

Premium Coupons: We also like underweighting premiums. 5.0s and higher are
more or less illiquid but still have the tightest LOASes and the lowest yields across the stack. The move up in coupon during the period of heavy taper speculation over the last few months contributed to the richening. Additionally, 5.0s continue to pose the greatest HARP-expansion risk. Underweight 15-year paper: The current environment has really motivated investors to flock towards the most liquid short duration paper: 15 year conventional TBA passthroughs. While we do not necessarily disagree with the preference to be short duration, we think that Dwarf/Fannie swaps are currently fully valued (we make a case for more attractive short duration instruments later in this piece). The chart below shows the relative wides in DW/FN swap prices (left) and tights in current coupon LOAS difference (15yr CC LOAS 30yr CC LOAS, right). Dwarf Fannie Swaps in Lower Coupons Have Widened
Source: RBS

FNCI FNCL Current Coupon LOAS


Source: The Yield Book, RBS

150 100 Ticks 50 -50 -100 Jun-12 Jun-13 Oct-11 Oct-12 Feb-12 Aug-11 Aug-12 Feb-13 Aug-13 Dec-11 Dec-12 Apr-12 Apr-13 Oct-13

bps

50 40 30 20 10 0 -10 -20 -30 -40 -50 8/1/11 2/1/12 4/1/12 6/1/12 8/1/12 2/1/13 4/1/13 6/1/13 10/1/11 12/1/11 10/1/12 12/1/12 8/1/13 10/1/13

DW2.5/FN3

DW3/FN3.5

DW3.5/FN4

Taper concerns were a main contributor to the relative cheapening in 30 year, but we generally feel that a sizeable part of the tapering now is priced in. The Fed is still currently buying 30 year so we could foresee the DW/FN spread narrowing at least in the near term. Moreover, with banks and REITs relatively quiet, it is unclear how much future demand 15 year paper will see at these levels. Finally, the 15 year sector may cheapen

Page 17/27

Agency MBS Outlook 2014: Uncertainty is Certain

as investors find better short duration relative value elsewhere (20 year pass-throughs, front end CMOs, hybrid ARMs we elaborate on these trades later in this document). Ginnie vs Conventionals: GN May Outperform on Regulatory Requirements: With the phase-in periods for several Basel III requirements beginning next year, we expect compliance with regulatory policy to remain at the forefront of bank investment guidelines. Regarding Basel IIIs Liquidity Coverage Ratio8 requirement, the Federal Reserve estimates that currently banks in aggregate have an approximate $200B shortfall in high quality liquid assets. We anticipate that banks will focus on fixing this shortfall by adding Level 1 liquid assets (as opposed to adding Level 2 or reducing net outflows). With banks looking to preserve yield and with GN MBS receiving classification of Level 1 liquid asset, we would expect GN to see gradually growing bank demand over next year. While GN/FN swaps have been improving recently, they still fall largely short of the wides we saw in 2011, 2012 and 1H 2013 (chart below). Most swaps are currently trading around 1 point; in prior years when investor preference for the explicit government guarantee of GNMA paper garnered heavy demand, we saw these swaps trade (and stay) at around 3 points. Ginnie/Fannie Swaps: Have Risen But Still Far from the Wides
120 100 80 Ticks 60 40 20 0 May-11 May-12 May-13 Mar-12 Feb-12 Feb-13 Mar-13 Oct-11 Oct-12 Jun-11 Aug-11 Sep-11 Nov-11 Dec-11 Jan-12 Jun-12 Aug-12 Sep-12 Nov-12 Dec-12 Jan-13 Apr-13 Jun-13 Aug-13 Sep-13 Oct-13 Apr-11 Apr-12 Nov-13 Jul-11 Jul-12 Jul-13

GN/FN 3.0
Source: RBS

GN/FN 3.5

GN/FN 4.0

Two more market positives that should support GN vs CV: 1) To the extent that tapering is not fully priced in, we expect that any lingering effects will more negatively impact FN and FH 30 year MBS and to a much lesser extent GN; 2) As the FHA-to-Conventional Refi trend picks up, GN MBS should see limited extension versus conventional counterparts.

Relative Value in Best Convexity Alternatives: Specified Paper Outlook9


Given higher rates and slower speeds (as evidenced by the past few prepayment reports), extension protection is a growing concern amongst investors. As a result, specified pools may not be an obvious choice going into the new year. However, with several stories now trading at single digit pay-ups, the good convexity of these securities is hard to pass up (especially ahead of increasing volatility). Moreover, the good convexity of certain stories implies faster discount speeds (as well as slower premium speeds). Finally, for those willing to sacrifice some liquidity we think the non-deliverable LTV stories (CQ, CR) are attractive across coupons. Below we show our standard
We have explored the upcoming Liquidity Coverage Ratio at length. For background and details please see our Special Report from November 6, 2013: Spotlight on the Liquidity Coverage Ratio, Effective January 2015 found here: https://strategy.rbsm.com/Tools/Content/ContentViewer.aspx?ContentID=369083 9 Our monthly Specified Snapshot recapping November in full will be out later this month
8

Page 18/27

Agency MBS Outlook 2014: Uncertainty is Certain

relative value analysis for 30 year specified pools (versus TBA). We use Yield Books new model (v20), which among its improvements now projects faster projected speeds (vs TBA, due to less lock-in) for lower loan balance discounts. Parametric Analysis: TBA vs Specified Pools
YB CPR Projections 1mo 6mo 3.73 4.22 3.79 4.33 3.78 4.31 3.77 4.29 3.76 4.27 2.34 2.87 2.05 2.48 1.80 2.15 1.72 2.01 1.84 Life 7.62 7.74 7.71 7.70 7.68 7.45 7.20 6.92 6.65 6.16 Yield Book Even OAS % (Act/ Eff Opt Theor. Theor. Th. Base Base Eff Yield WAL Dur Cnvx OAS Cost OAS Price Payup Payup) 3.44 8.7 7.3 -0.4 25 25 25 100-10+ 3.41 8.6 7.3 0.2 39 15 25 101-17+ 39 21% 3.41 8.7 7.3 0.1 36 17 25 101-10 32 22% 3.43 8.7 7.3 0.0 33 20 25 101-02 24 17% 3.43 8.7 7.3 -0.1 31 21 25 100-27+ 17 17% 3.44 8.8 7.5 -0.2 27 23 25 100-18 8 26% 3.44 9.0 7.7 0.0 28 21 25 100-19+ 9 22% 3.44 9.2 7.8 0.1 28 20 25 100-20 9 21% 3.44 9.4 7.9 0.1 27 19 25 100-19+ 9 25% 3.57 9.8 8.1 8.5 7.3 7.4 7.2 6.0 6.4 6.3 6.3 6.2 6.6 6.8 6.9 7.1 7.3 7.4 6.1 6.2 5.3 3.7 5.5 5.0 4.7 4.3 4.3 4.6 4.8 5.0 5.4 5.9 4.1 4.1 0.3 0.6 -0.2 -0.3 -0.6 -1.3 -0.1 -0.3 -0.5 -0.7 -0.4 -0.3 -0.2 -0.1 0.1 0.4 -0.9 -0.8 -2.9 -2.7 -0.8 -1.4 -1.8 -2.2 -2.7 -2.5 -2.4 -2.3 -2.0 -1.5 -2.6 -2.8 40 42 26 26 31 25 35 33 31 30 31 33 33 33 47 53 26 27 28 17 25 22 21 19 20 22 22 22 28 36 12 19 17 12 24 23 27 31 16 19 22 25 25 24 22 21 18 11 29 28 40 48 22 29 34 39 45 43 41 39 35 28 46 44 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 25 17 17 17 17 17 17 17 17 17 17 17 17 17 100-18+ 100-14+ 100-14+ 100-15 99-23+ 103-31 105-05+ 104-29+ 104-21+ 104-15 104-16+ 104-21 104-24 104-27 104-30 105-05 104-05 104-07 103-04 106-16+ 108-20+ 108-04+ 107-22+ 107-10+ 107-06 107-14 107-18+ 107-25+ 108-02+ 108-13 106-26 106-30+ 8 4 4 4 -19 38 31 23 16 17 22 25 28 31 38 6 8 -27 68 52 38 26 21 30 34 41 50 60 9 14 -367% -1021% 50% 47% 170% 47% 46% 44% 43% 23% 18% 28% 32% -65% -79% 67% 49% 118% 77% 81% 79% 85% 70% 64% 71% 73% 54% 30% 205% 79%

Cpn

3.5
Loan Bal

TBA Hi LTV

Description TBA LLB 85k max MLB 110k max HLB 150k max 175K 175k max MHA90 100% Refi MHA95 100% Refi MHA100 100% Refi MHA105 100% Refi 100% Refi 100% Refi FICO < 700 100% Inv 100% Jumbo 85k max 110k max 150k max 175k max 100% Refi 100% Refi 100% Refi 100% Refi 100% Refi 100% Refi FICO < 700 100% Inv 100% Jumbo 85k max 110k max 150k max 175k max 100% Refi 100% Refi 100% Refi 100% Refi 100% Refi 100% Refi FICO < 700 100% Inv

AO LnSz 304 67 98 129 161 246 230 222 216 206 204 304 304 538 278 65 97 128 161 236 219 215 204 202 197 278 278 543 261 64 97 127 159 227 216 222 208 194 191 261 261

O Wac 4.01 4.01 4.01 4.01 4.01 4.01 4.01 4.01 4.01 4.01 4.01 4.01 4.01 4.46 4.46 4.46 4.46 4.46 4.46 4.46 4.46 4.46 4.46 4.46 4.46 4.46 4.94 4.94 4.94 4.94 4.94 4.94 4.94 4.94 4.94

Wam Age oltv 349 8 69 349 8 69 349 8 69 349 8 69 349 8 69 349 8 90 349 8 95 349 8 100 349 8 105 8 349 349 349 349 334 334 334 334 334 334 334 334 334 334 334 334 334 308 308 308 308 308 308 308 308 308

fico 757 757 757 757 757 757 757 757 757

pay up Price Settle - 100-10+ 12/11 8 100-18+ 12/11 7 100-17+ 12/11 4 100-14+ 12/11 3 100-13+ 12/11 2 100-12+ 12/11 2 100-12+ 12/11 2 100-12+ 12/11 2 100-13 12/11

Non CQ TBA CR LTV LFICO Low FICO Inv Jumbo

4.01 349

115 757 -30 99-12+ 12/11 1.56

4
Loan Bal

TBA Hi LTV

Investor CK TBA LLB MLB HLB 175K MHA90 MHA95 MHA100 MHA105

8 155 757 -42 99-00+ 12/11 1.57 8 69 669 2 100-12+ 12/11 3.53 8 69 757 2 100-12+ 12/11 3.63 8 75 762 -32 99-10+ 12/11 3.19 21 62 752 - 103-31 12/11 8.13 21 62 752 18 104-17 12/11 7.38 21 62 752 14 104-13 12/11 7.42 21 62 752 10 104-09 12/11 7.50 21 62 752 7 104-06 12/11 7.60 21 90 752 4 104-03 12/11 4.36 21 95 752 4 104-03 12/11 3.94 21 100 752 7 104-06 12/11 3.59 21 105 752 9 104-08 12/11 3.48 21 115 752 -20 103-11 12/11 3.30

1.83 5.09 3.62 10.7 4.22 7.17 3.44 9.0 4.14 7.34 3.44 8.9 3.82 6.97 3.59 9.2 8.06 10.13 3.31 7.2 7.91 9.19 3.26 7.7 7.91 9.25 3.28 7.7 7.93 9.37 3.29 7.6 7.96 9.51 3.30 7.5 4.82 9.22 3.34 7.7 4.31 8.78 3.36 7.9 3.90 8.40 3.36 8.1 3.73 8.03 3.37 8.4 3.57 7.47 3.51 8.7 9.4 7.6 7.7 6.3 4.1 6.8 6.2 5.7 5.2 4.8 5.2 5.4 5.7 6.3 7.4 4.5 4.9

Non CQ TBA CR LTV LFICO Low FICO Inv Jumbo

4.46 334

4.5
Loan Bal

TBA Hi LTV

Investor CK TBA LLB MLB HLB 175K MHA90 MHA95 MHA100 MHA105

21 155 752 -30 103-01 12/11 3.73 4.08 6.53 3.58 21 62 667 4 104-03 12/11 6.64 7.19 9.36 3.33 21 62 752 4 104-03 12/11 7.22 7.64 9.11 3.33 21 74 755 -32 102-31 12/11 9.15 7.42 12.15 3.42 45 63 756 - 106-16+ 12/11 24.32 21.19 19.85 2.67 45 63 756 52 108-04+ 12/11 10.42 10.28 10.63 3.05 45 63 756 42 107-26+ 12/11 12.63 11.89 12.04 2.99 45 63 756 30 107-14+ 12/11 14.94 13.61 13.51 2.95 45 63 756 22 107-06+ 12/11 17.26 15.39 15.09 2.88 45 90 756 15 106-31+ 12/11 14.20 11.93 16.50 2.83 45 95 756 19 107-03+ 12/11 11.76 9.85 15.25 2.90 45 100 756 24 107-08+ 12/11 10.52 8.64 14.56 2.93 45 105 756 30 107-14+ 12/11 9.63 7.90 13.42 2.98 45 115 756 45 155 756 45 63 666 45 63 756 27 107-11+ 12/11 8.34 6.90 11.79 3.12 18 107-02+ 12/11 7.14 6.27 9.29 3.34 19 107-03+ 12/11 19.00 16.67 18.12 2.67 11 106-27+ 12/11 19.42 16.59 16.39 2.85

Non CQ TBA CR LTV LFICO Low FICO Inv

4.94 308 4.94 308 4.94 308 4.94 308

Investor

As of the close 12/2/13. Source: The Yield Book, RBS

Within the coupon stack, we still like specified 4s and 3.5s given low nominal payups. In both coupon buckets, specifieds pick decent yield and LOAS to TBA for just slightly more duration. Furthermore, the low nominal pay-ups make specifieds appear very attractive on a percent-of-theoretical basis. In both coupons, the specifieds have substantially better convexity than the TBA; in 3.5s, convexity is actually positive. Although 3.5s and 4.0s are currently not refinanceable at prevailing mortgage rates, we find that this good convexity is priced so attractively it may be hard to give up.

In 3.5s, we prefer Loan Balance to MHA. Loan balance 3.5s generally have shorter
WALs and durations than MHA pools. While loan balance pools generally trade at a higher pay-up than MHA, they pick decent LOAS to MHA. Finally, loan balance buckets are trading at a slightly lower theoretical percent of TBA than are MHA

Page 19/27

Agency MBS Outlook 2014: Uncertainty is Certain

buckets. Aside from the RV figures shown above, we prefer loan balance paper for the following reasons: 1) the call protection of loan balance paper has historically held in rising HPA environments, whereas it erodes somewhat for MHA; 2) loan balance paper is unlikely to be targeted by new policy (Although we are pessimistic that re-HARPing will be permitted, the idea continues to be discussed and MHA paper continuously widens on the speculation); and 3) the good convexity of loan balance paper has traditionally implied faster speeds in a backup due to a move-up effect (while there is little data showing how MHA paper prepays in a discount environment).

In 4.0s, MHA appears just slightly cheap to Loan Balance (on both LOAS and
theoretical comparisons) however we have a slight preference towards loan balance (for all of the aforementioned reasons).

We find that Non-Deliverable LTV Paper (CQ and CR Collateral) remains the most
attractive. This remains our favorite story and may be especially appealing to those investors who do not need the liquidity of the TBA deliverable. CQ (105% < LTV 125%) and CR (> 125% LTV) collateral continue to be the top performers in our parametric analysis and look cheap to TBA and other specified stories. These pools have the best (now positive) convexity, which is captured by the large LOAS numbers. Nominal pay-ups are small or negative. Finally, in line with the thinking that higher LTV buckets will perform better in rising HPA environments, CQ and CR pools make a lot of sense.

Relative Value In Short Duration: Consider 20 year vs 15 year


For investors who want the execution offered by the TBA market, 20 year pass-throughs are an attractive alternative to 15 year. The borrower characteristics (below, left) and prepayment profiles (below right) of 20 year paper are very similar to 15 year paper, but 20 year is priced closer to 30 year paper. The lower loan size of 20 year paper and the slight degree of built in burnout contribute to slower speeds across the board, and we believe the slightly lower FICOs and higher LTVs (vs 15 year) provide slightly more call protection. We do acknowledge that one issue with 20 year paper its relatively small market share (consistent with our theme of giving up some liquidity to earn return). Currently, 20 year represents just 4% of all conventional fixed rate issuance, versus 10% in 2011. As 20 year is traditionally a refi product, production may likely decline. Regarding pricing, 20 year paper trades at a pay-up to 30 year (as it is technically good delivery as 30 year). This convention implies that generally 20 year looks cheap when DW/FN swaps are at their wides. This relationship is exactly what we see now; even though 20 year pay-ups have risen slightly, they are very close to the historical lows of the interpolated spreads (currently ~48%; we consider 50-55% fair value; charts bottom of page).

Page 20/27

Agency MBS Outlook 2014: Uncertainty is Certain

Collateral Characteristics by Origination Year and Term


Source: CPRCDR, RBS

Historical 1mo CPR: 1 year history (sample period 2013)


Source: CPRCDR, RBS

Product 2009 FN30 FN20 FN15 FN30 FN20 FN15 FN30 FN20 FN15 FN30 FN20 FN15 FN30 FN20 FN15 FN30 FN20 FN15 FN30 FN20 FN15

Olnsz ($k) 229 183 172 227 194 177 213 193 175 222 192 184 217 170 176 222 187 177 168 156 145

CLnsz ($k) 179 130 102 188 152 121 185 164 136 210 179 164 215 167 170 200 166 149 174 147 114

Orig FICO 755 759 756 754 761 763 754 761 763 759 763 766 753 752 760 755 759 763 747 755 758

Orig LTV 71 66 61 74 69 63 75 69 64 74 71 65 75 71 65 74 70 64 74 70 64

Cur LTV 62 53 44 64 57 47 63 57 49 64 61 54 70 66 60 65 61 53 65 59 50

%Cur LTV>80
1- mo CPR 1- mo CPR

30 25

2010

2011

2012

2013

2009-13

All

2 20 0 15 0 10 7 5 0 0 0 -125 -100 -75 -50 -25 0 25 50 75 100 125 9 Rate Incentive 0 FN 15 FN 20 FN 30 0 10 Historical 1mo CPR: 5 year history (sample period 2009-13) 4 Source: CPRCDR, RBS 1 12 30 5 25 2 20 9 15 3 10 1 5 11 0 2 -125 -100 -75 -50 -25 0 25 50 75 100 125 1
Rate Incentive FN 15 FN 20 FN 30

20 Year Pay ups Have Recently Risen


Source: RBS

But Look Attractive on an Interpolated Basis


Source: RBS

160 140 120 100 80 60 40 20 0 1/31/13 2/28/13 3/31/13 4/30/13 5/31/13 6/30/13 7/31/13 8/31/13 12/31/12 9/30/13 10/31/13

100% 80% 60% 40% 20% 0% 1/31/13 2/28/13 3/31/13 4/30/13 5/31/13 6/30/13 7/31/13 8/31/13 12/31/12 9/30/13 10/31/13

FNCT 2.5

FNCT 3.0

FNCT 3.5

FNCT 2.5

FNCT 3.0

FNCT 3.5

All things combined, 20 year valuations currently look quite attractive. Below we show a parametric analysis of 20 year paper versus duration matched blends of comparable 15 and 30 years. For both 3.0% and 3.5% coupon, we see that the 20 year picks yield (+1014bps) and LOAS (+9-13bps) to its duration matched 15/30 blend, which is largely reflected by its better convexity.

Page 21/27

Agency MBS Outlook 2014: Uncertainty is Certain

Parametric Analysis: 20 year collateral vs Duration-weighted 30 and 15 year


Payup % FNCI 3 FNMA 3 Combined FNCT 3 30yr Passthrough 51% 15/30 Blend 20yr Passthrough Px to FNCL 96-26+ 96.82813 99-31 100-06+ 0-08 105-14 105.4453 102-25+ 103-05+ 0-12 65 3.500 3.500 3.500 3.500 108 Cpn 3.000 3.000 3.000 3.000 WAC 3.480 3.590 3.536 3.560 WAM 168 354 263 236 Age 11 3 6.92 3 -3.92 30 8 16.47 8 -8.47 FICO 759 765 762 759 -3 762 757 759 747 -12 LTV 66 72 69 70 1 64 75 71 72 1 15yr Passthrough 50% 103-07+ 103.2344 LnSz $k Settle Eff Dur Eff Cnvx 260 320 290 252 -38 227 315 281 235 12/11 -46 12/11 12/11 12/11 12/17 12/11 12/17 12/11 4.34 8.08 6.24 6.24 0.00 2.97 7.14 5.53 5.53 0.00 -0.80 -0.24 -0.52 -0.16 0.35 -1.34 -0.63 -0.90 -0.68 0.21 Yield Book OAS OCost 9 23 16 25 9 6 19 14 27 13 13 20 17 14 -3 16 26 22 19 -3 Yield 2.26 3.42 2.85 2.95 0.10 1.85 3.32 2.76 2.89 0.14 WAL 5.0 9.4 7.2 6.9 -0.4 3.6 8.5 6.6 6.2 -0.4 CPR 9.6 6.6 8.1 7.6 -0.5 16.0 7.9 11.0 9.4 -1.7

Difference (20yr - 15/30) FNCI 3.5 FNMA 3.5 Combined 15yr Passthrough 39% 15/30 Blend

0.024 -27.37 3.900 4.010 3.968 146 349 271

30yr Passthrough 62% 101-04+ 101.1406

FNCT 3.5 20yr Passthrough Difference (20yr - 15/30)

4.060 231 0.092 -39.85

As of the close on 11/25/13. Source: The Yield Book, RBS

Finally, we show that 20 year paper is a nice compromise to 15 and 30 year. It has WALs generally longer than 15 year but the investor is compensated by a substantial yield pick in all scenarios (roughly 60-70bps pick for 1.5-2 years WAL). The 20 year outperformance (vs 15 year) increases in rallies, where the two display similar WALs but 20 year continues to pick yield (36-80bps). 20 year also looks very attractive vs 30 year, with very similar yields (~30-40 bps difference in base case and backups) but roughly a 3 year shorter WAL. Yield and Average Life Profile: 15, 20 and 30 year Pass-throughs
Shift 300 200 100 0 -100 -200 FNCL 3.5, 30 yr Generic Passthrough, Price: 101-04+, Settle: 12/11/13, (4.020/352/8) PSA 84 90 107 146 383 712 Yield 3.36 3.36 3.35 3.32 3.16 2.92 AvgLife 11.2 10.9 10.0 8.5 4.2 2.4 54 55 59 89 210 251 Spread/I FNCT 3.5, 20 yr Generic Passthrough, Price: 103-05+, Settle: 12/11/13, (4.000/235/5) 98 114 137 195 412 642 PSA Yield 3.02 3.00 2.96 2.88 2.55 2.20 AvgLife 8.0 7.6 7.1 6.1 3.8 2.7 Spread/I 70 77 84 111 165 171 FNCI 3, 15 yr Generic Passthrough, Price: 103-07+, Settle: 12/17/13, (3.480/163/17) PSA 95 108 138 215 405 609 2.36 2.34 2.29 2.17 1.81 1.37 Yield 5.8 5.6 5.2 4.4 3.0 2.2 AvgLife Spread/I 70 74 84 104 124 103
Source: Bloomberg, RBS. As of the close on Monday 11/25/13.

-300 950 2.75 1.9 248 1088 1.56 1.8 131 678 1.21 2.0 92

Relative Value In Extension Protected Paper: Front End CMOs Cheap to Alternatives
With rates resetting comfortably off the 2013 lows and potentially poised to reset higher, investors are justifiably concerned over extension. Typically 15 year TBA is the preferred choice to reduce extension risk given its limited term, degree of borrower burnout and strong liquidity. However, heavy demand has richened the sector already. Instead we find that investors can look to structured paper (notably well structured CMO front cash flows and Hybrid ARMs) to find very similar extension protection at decent yield and spread picks to 15 year collateral. We acknowledge that although both CMOs and ARM spreads have tightened relative to collateral since the recent wides, they still represent

Page 22/27

Agency MBS Outlook 2014: Uncertainty is Certain

solid relative value vs 15 year. While valuations for both are attractive (even after coming in substantially), we would expect that the demand outlook is slightly more favorable for CMOs than for ARMs (we elaborate on this later). We first illustrate the relative value of short structured securities in the parametric analysis below. All of our securities fall within the 4-5 year duration bucket. Versus FNCI 3.0, we compare:

FHR 4138 PA, a 4 year PAC [136-239] off of new MHA Gold 3.5% (99 WALTV) FHR 4242 PA, a 4 year PAC [134-201] off of new HLB Gold 4.0% (134k WAOLS) FN AL4112, a 7/1 Hybrid ARM, 5/2/5 Cap, 12m L + 1.683, 15% TPO
From our results below, the cheapness of structure is apparent. All securities pick LOAS to FNCI 3.0: PACs LOAS at 24-27bps and our ARM at 31bps for our ARM vs 8bps for our pass-through. On a base case yield comparison, FHR 4183 PA is in line with FNCI 3.0 (2.28% vs 2.26%, respectively) and FHR 4242 PA picks 15bps (2.41%). Our 7/1 ARM gives nearly 40bps to FNCI as per Yield Books model. Finally, our short CMOs currently have the best convexities of the group; the ARM convexity is slightly worse than DW 3.0 Parametric Analysis for 4 year Alternatives: 15 Year Pass-through, Front PACs and 7/1 ARM
Name FNCI 3.0 Description 15yr Passthrough PRICE 103-07+ 104-27+ 102-24 WAC/WAM/WALA 3.480 / 168 / 11 4.051 / 335 / 15 4.533 / 333 / 22 3.069 / 347 / 14 Settle Date Dec 17 Nov 27 Nov 27 Nov 26 Curr Cpn 3.00 3.50 3.50 2.50 At Pricing YTM 2.20 2.28 2.41 1.75 110/i 105/i 36/z Sprd Speed 215 150 202 15 PSA PSA PSA CPB WAL 4.6 4.5 4.9 3.6 Yield Book Opt Yield WAL LCPR Dur Cnvx LOAS Cost 2.26 2.28 2.41 1.88 5.0 4.5 4.9 5.8 9.6 8.1 9.9 4.35 -0.73 4.67 -0.17 4.32 -0.34 8 27 24 31 12 19 18 25

FHR 4183 PA 4yr PAC x HiLTV FH 3.5% [136-249] 105-1+ FHR 4242 PA 4yr PAC x HLB FH 4% [134-201] FN AL4112 7/1 ARM

13.0 4.13 -1.23

Source: The Yield Book, RBS. As of the close on Thursday, 11/21/13.

The real value of short CMOs (versus 15 year pass-throughs) can be best seen in the yield and average life table below. For these comparisons, we assume Bloomberg consensus speeds for the standard parallel curve shifts for the pass-through and CMOs; we assume a handful of CPB projections for the ARM. Yield and WAL Comparison: 15 Year Pass-through, Front PACs and 7/1 ARM
Shift 300 200 100 0 -100 405 1.81 3.0 -200 609 1.37 2.2 -300 678 1.21 2.0 850 0.11 1.5 671 0.37 1.6 50 0.43 1.4

FNCI 3.0, 15yr Passthrough, price: 103-7+, settle: 12/17/13 95 108 138 215 PSA 2.36 2.34 2.29 2.20 Yield 5.8 5.6 5.2 4.6 AvgLife

FHR 4183 PA, 4yr 3.5% PAC [136-249] x HiLTV FH 3.5%, price: 105-1+, settle: 11/27/13 84 90 108 150 326 658 PSA 2.54 2.50 2.39 2.28 2.05 0.85 Yield 5.9 5.7 5.0 4.5 3.7 2.0 AvgLife FHR 4242 PA, 4yr 3.5% PAC [134-201] x HLB FH 4%, price: 104-27+, settle: 11/27/13 95 112 136 202 425 600 PSA 2.55 2.46 2.41 2.41 1.52 0.72 Yield 5.7 5.2 5.0 4.9 2.6 1.8 AvgLife FN AL4112, 7/1 ARM, price: 102-24, settle: 11/26/13 7 10 12 CPB 1.83 1.77 1.73 Yield 4.5 4.2 3.9 AvgLife
Source: Bloomberg, RBS. As of close on 11/21/13

15 1.75 3.6

25 1.41 2.7

40 0.89 1.8

Page 23/27

Agency MBS Outlook 2014: Uncertainty is Certain

Here we note that the CMOs pick yield to FNCI 3.0 in the base case and all backup scenarios, while having an average life profile almost identical to dwarfs. In rally scenarios, CMOs start to underperform, slightly giving yield and contracting a bit more than the pass-through. For investors less concerned about rate drop scenarios, CMOs appear to be a solid substitute to 15 year paper. We include the 7/1 ARM for completeness, but we see that its WAL is roughly 1 year shorter than the others in the base case and backups; as a result it gives roughly 75-100 bps in yield here. All things considered, while we believe ARMs and front end CMOs both represent attractive short-duration opportunities, we are partial to well structured front-end CMO cash flows. In addition to the relative value advantages demonstrated above (having the best convexity, significant yield and decent spread picks to collateral, comparable extension protection), our preference for CMOs is further motivated by an improving supply/demand technical. Supply is likely to tick higher for both, as a steeper curve lends itself to more ARM issuance and is more favorable for CMO creation. We have seen gradual pick-ups here (charts below), and would expect this trend to continue should rates continue to back up with the front end pinned. We believe that in both sectors, extra supply would be beneficial for market liquidity. However, we think the demand prospects for short CMOs are slightly more favorable than for ARMs. As mentioned earlier, we believe banks will be focused on regulatory requirements so we could see some light interest in front end CMOs (especially if a higher rate range provides more attractive entry points). However, we expect REIT demand to be light and as a result we could foresee ARMs underperforming on a withdrawal of support. CMO Issuance
35 30 25 20 15 10 5 0 May-12 May-13 Mar-13 Jun-12 Jan-13 Jun-13 Oct-12 Aug-12 Sep-12 Aug-13 Sep-13 Nov-12 Dec-12 Feb-13 Oct-13 Apr-12 Jul-12 Apr-13 Jul-13 Nov-13

ARM Issuance
250,000 200,000 150,000 100,000 50,000 0 30% 25% 20% 15% 10% 5% 0%

$ Billions

$ Millions

FHLMC

FNMA

GNMA

Fixed

Source: Bloomberg, RBS

Source: CPRCDR, RBS

Relative Value In Less Liquid Alternatives: HREMIC Floaters and IOs Cheap to Conventional Forward Counterparts
We continue to make the case that investors willing to sacrifice liquidity can find reward via enhanced returns. We also maintain that giving up liquidity does not necessarily mean moving down in credit. We discussed earlier that the highest LTV pass-throughs (CQ and CR) are ineligible for good TBA delivery, however still carry the same Agency guaranty and weighting as FNCL. In this last section of our 2014 Outlook, we show that GN HREMIC Floaters and IOs pick return to Conventional counterparts, and the GN HREMIC investor benefits from the favorable capital treatment and explicit government guarantee that is characteristic of all GN MBS securities. Below we show two examples where HREMICs demonstrate decent relative value. In the first comparison we show an HREMIC floater x Standard ARM collateral (GNR 2013-H16 FA) vs. a FN strip floater x CR 4.5% collateral (FNR 2012-99 FQ). Our HREMIC floater pays a 1mL+54bps coupon (currently set at 0.7158%) with an 11% cap, while the strip

Ja n0 Ja 2 n0 Ja 3 n04 Ja n0 Ja 5 n0 Ja 6 n0 Ja 7 n08 Ja n0 Ja 9 n1 Ja 0 n1 Ja 1 n1 Ja 2 n13


ARM % ARM

Page 24/27

Agency MBS Outlook 2014: Uncertainty is Certain

floater has a coupon of 1mL + 45 bps (currently set at 0.616%) with a 7% cap. Both floaters should be natural fits for banks given their par-ish dollar price, coupons that increase with rising rates, and favorable treatment under upcoming capital regulations. However, we the HREMIC floater benefits even further from a GNMA backing and a higher cap. Our outlook for the HMBS sector is also positive (we elaborate in later paragraph). Finally, at current levels the HREMIC floater offers decent relative value versus the FN CMO floater (yield table below). Yield and WAL Comparison of Floating Rate Comparables
GNR 2013-H16 FA, HREMIC Floater x Standard ARM, Price: 100-3+, Settle: 12/6/13, Cpn: 0.7158% MHP DM YTF WAL 20 53 3.03 8.7 30 52 2.94 8.1 40 52 2.86 7.6 50 52 2.77 7.1 60 52 2.70 6.7 80 52 2.55 6.0 100 52 2.41 5.4
FNR 2013-27 KI, 175k 15yr 3.5% Strip IO, Price: 14-15+, Settle: 12/5/13 300 200 100 0 Yield Shift 102 125 163 248 PSA 7.16 5.73 3.34 -2.13 Yield WAL 5.4 5.2 4.7 3.9

Yield and WAL Comparison of IO Comparables


GNR 2013-H14 HI, HREMIC IO x Standard Fixed, Price: 11-30, Settle: 12/5/13, Cpn: 2.49364% 20 30 40 50 60 MHP 6.93 6.63 6.06 5.23 4.26 Yield 5.7 5.7 5.5 5.4 5.2 WAL

80 2.16 4.9

100 0.13 4.7

FNR 2012-99 FQ, Floater x CR 4.5%, Price: 99-28, Settle: 12/6/13, Cpn: 0.616% Yield Shift PSA DM YTF WAL 300 96 47 3.64 10.2 200 103 47 2.63 9.8 100 115 47 1.63 9.3 0 139 47 0.63 8.4 -100 200 47 0.47 6.5 -200 249 47 0.48 5.5 -300 276 48 0.48 5.1

-100 -200 -300 374 516 587 -10.53 -20.52 -25.74 3.1 2.4 2.1

Source: Intex, Bloomberg, RBS. As of close on 12/2/13.

Source: Intex, Bloomberg, RBS. As of close on 12/2/13.

The HREMIC floater not only exhibits more average life stability but also picks over 200bps in base case yield (2.77% vs. 0.63%) despite being more than a year shorter in average life and ~190-225bps in yield in rallies (~2.55% vs. ~0.5%). The HREMIC floater also picks ~30-125bps in yield over most backup scenarios, and though there is a slight give in yield in the most extreme backup case, it experiences slightly less extension here than the strip floater. The HREMIC floater has a slight DM advantage (~5bps) in all of the above rate scenarios. In the second comparison above (right), we show an HREMIC IO x Standard Fixed collateral (GNR 2013-H14 HI) vs. a FN strip IO x 175k max 15yr 3.5% collateral (FNR 2013-27 KI). The HREMIC IO picks substantial yield to the strip IO comparables in the base case and all rallies and all but the +300 backup. The real value ultimately lies in the stability of its WAL profile. Our HMBS outlook is also positive over the near term. We expect to see two main themes in HMBS off of newly originated HECM collateral: shrinking supply (which should remain predominantly ARM as long as front end rates are pinned) and a general extension of cash flows10. The former is the result of requiring a financial assessment for the first time ever to qualify borrowers, as well as placing limitations on the upfront draw and reducing the principal limits. The latter results from the expectation of lower default related prepayment (ensuring borrowers are able to repay their mandatory obligations from HECM proceeds if need be), and creating a longer time until assignment (via lower principal limits). In addition to all of the above, our HREMIC trade is further supported by the following: Because of the unique nature of the product, HMBS is distanced from the volatility that
For a cohesive summary (including impact) on the changes to the HECM program which were announced in September, please see our past publication HECM Revamp: Smaller Supply and Slower Speeds to Come which can be found here: https://strategy.rbsm.com/Tools/Content/ContentViewer.aspx?ContentID=359606
10

Page 25/27

Agency MBS Outlook 2014: Uncertainty is Certain

currently surrounds forward mortgages (taper talk, rising rates, and speculation over government induced refi or modification). Also HMBS tends to exhibit excellent convexity as borrowers tend to be more age-affected and less rate-affected. Finally, borrower mortality implies that extension risk here should be limited.

Conclusion
Agency MBS investors will likely encounter numerous uncertainties heading into 2014. The majority of these are policy relates: monetary (the most imminent in our opinion), fiscal, economic and GSE-related to name a few. However, we also expect supply/demand dynamic to be non-traditional, as the usual investors will be sidelined by regulations on top of everything else. While these uncertainties will likely manifest via increased volatility and wider spreads, we actually think that any underperformance could provide some opportunities (down the line). Over the next year we believe investors may be find to add good convexity at cheap levels. We also think extension risk should be at the forefront of investor concerns. We recap our favorite trade strategies for 2014 below:

We are neutral on the mortgage basis. Technicals are decent and valuations appear
fair versus historicals (but arguably carry less weight in a policy-driven landscape).

Within TBA, we like adding the belly on weakness and GNMA vs. Conventionals. We
currently find 15 year TBA rich to alternatives (but generally think short duration makes sense in the upcoming environment).

Within Specified Pools, we like the lower pay-up stories (loan balance 3.5% and 4% in
TBA eligible, CQ/CR in non-deliverable), though we acknowledge the extension risk.

In Short Duration space, we prefer stable CMO PACs and SEQs and 20 year paper (if
you can find it) as substitutes for 15 year pass-throughs. ARMs look a little less cheap to us, but still provide decent relative value.

We believe the best relative value can be achieved by giving up some liquidity.
Specifically, we like the highest LTV paper (CQ/CR) and HREMIC floaters and IOs.

Page 26/27

Agency MBS Outlook 2014: Uncertainty is Certain

Invitation to Consider a Derivatives Transaction: This communication is prepared by the sales and trading desk and is marketing material and/or trader commentary. It is not a product of the research department. This material constitutes an invitation to consider entering into a derivatives transaction under U.S. CFTC Regulations 1.71 and 23.605, where applicable, but is not a binding offer to buy/sell any financial instrument. The views of the author may differ from others at The Royal Bank of Scotland plc, The Royal Bank of Scotland N.V. and/or RBS Securities Inc. (collectively "RBS"). Copyright 2013 RBS Securities Inc. All rights reserved. RBS Securities Inc. member FINRA (http://www.finra.org ) / SIPC (http://www.sipc.org ), is a subsidiary of The Royal Bank of Scotland plc. RBS is the marketing name for the securities business of RBS Securities Inc. The author of this material is an economist, desk strategist, salesperson or trader and will be compensated based in part on the authors own performance, the firms performance and the performance of the sales or trading desk for which the author works. This material is informational only. It is not considered research and is not a product of any research department. RBS Securities Inc. is not, by making this material available, providing investment, legal, tax, financial, accounting or other advice to you or any other party. RBS Securities Inc. is not acting as an advisor or fiduciary in any respect in connection with providing this information, and no information or material contained herein is to be construed as either projections or predictions. Past performance is not indicative of future results. RBS Securities Inc. transacts business with counterparties on an arms length basis and on the basis that each counterparty is sophisticated and capable of independently evaluating the merits and risks of each transaction and that each counterparty is making an independent decision regarding any transaction. Counterparties must make their own independent decisions regarding any securities, financial instruments or strategies mentioned herein. This material is intended for institutional investors only and should not be forwarded to third parties. RBS Securities Inc., its affiliates, and/or any of their respective employees, clients and officers, including persons involved in the preparation or issuance of this material, act as a market maker for or deal as principal in the financial instruments mentioned and may have a long or short position in, or engaged in transactions in, these financial instruments. As part of the sales or trading desk, the author may have consulted with the trading desk while preparing this material and the trading desk may have accumulated positions in the financial instruments or related derivatives products that are the subject of this material. Accordingly, information included in or excluded from this material is not independent from the proprietary interests of RBS Securities Inc., which may conflict with your interests. Except as to securities issued by the U.S. government and debentures issued by government-sponsored enterprises, RBS Securities Inc. does not permit the author of this material to hold, or engage in transactions in, the securities or other financial instruments discussed herein in any personal account of the author or a member of the authors household. Certain transactions, including those involving futures, options, derivatives and high yield securities, give rise to substantial risk and are not suitable for all investors. Foreign currency denominated securities are subject to fluctuations in exchange rates that could have an adverse effect on the value or price of or income derived from any investments discussed herein. In addition, investors in securities such as ADRs, the values of which are influenced by foreign currencies, effectively assume currency risk. Please ensure that you have read and understood the current options disclosure document before entering into any options transactions. United States listed options disclosure documents are available from our sales representatives or at http://www.theocc.com/components/docs/riskstoc.pdf . Unless otherwise specifically stated, all statements contained herein (including any views, opinions or forecasts) are as of the date indicated only, and are subject to change without notice. Changes to assumptions made in the preparation of such material may have a material impact on returns. RBS Securities Inc. does not undertake a duty to update this material or to notify you when or whether the analysis has changed. While the information contained in this material is believed to be reliable, no representation or warranty, whether express or implied, is made and no liability or responsibility is accepted by RBS Securities Inc. or its affiliates as to the accuracy or completeness thereof, except with respect to any disclosures relative to RBS Securities Inc. and/or its affiliates and the authors involvement with an issuer that is the subject of this material. RBS Securities Inc. makes no representations that this material or any information contained herein are appropriate for use in all locations or that transactions, securities, products, instruments or services discussed herein are available or appropriate for sale or use in all jurisdictions, or by all investors or counterparties. Investors who receive this material may not necessarily be able to deal directly with RBS Securities Inc. and should contact the RBS Securities Inc. entity or affiliate in their home jurisdiction unless governing law permits otherwise. Those who utilize this information do so on their own initiative and are responsible for compliance with applicable local laws or regulations. This material is made available in the European Economic Area ("EEA") by The Royal Bank of Scotland plc, which is authorized and regulated by the Financial Services Authority (FSA). RBS Securities Inc. distributes this material in the U.S. With respect to the execution of interest rate derivatives, credit derivatives and foreign exchange transactions, RBS Securities Inc. generally acts as agent for RBS which will be the principal in such transactions. This website and its contents have not been reviewed by any regulatory authority in Hong Kong. RBS Securities Inc. does not conduct, nor holds itself out as conducting a business in investment advisory or dealing services in Hong Kong, nor any other regulated activity. It holds no Hong Kong regulatory licenses. This material is distributed in Singapore by The Royal Bank of Scotland plc (Singapore branch), Level 26, One Raffles Quay, south Tower, Singapore 048583, which is regulated by the Monetary Authority of Singapore. In Singapore, this material is intended solely for distribution to institutional investors, accredited investors, and expert investors as defined under the (Singapore) Financial Advisers Act, Chapter 110. Persons receiving this material in Singapore should contact The Royal Bank of Scotland plc (Singapore branch) in respect of any matters in connection with the material.) Both RBS Tokyo and RBSSJ are regulated by Japanese Financial Services Agency. This Material is provided for information only and it does not constitute an offer or a solicitation to deal in any specific products or to enter into any legal relations, nor an advice or a recommendation with respect to such products. Clients should independently evaluate particular investments before dealing in any products herein referred to. Persons obtaining this information through RBSM website are limited to Professional Investors (as defined in Article 2-31 of Financial Instruments and Exchange Law) only. The access to the website is provided by RBS Tokyo Branch (RBS Tokyo) or RBS Securities Japan Limited (RBSSJ) to their respective Clients in Japan who are classified as Professional Investors. However some Material is intended to be provided only to Qualified Institutional Investors (as defined in Article 10 of Cabinet Office Ordinance concerning Definitions provided in Article 2 of Financial Instruments and Exchange Law). RBS Securities Inc. is exempt from the requirement to hold an Australian Financial Services License under the Corporations Act 2001 of Australia in respect of these financial services, and that RBS is regulated by the SEC under US laws, which differ from Australian laws. The views expressed in this publication accurately reflect the authors personal views about the subject financial instruments and issuers, and no part of the authors compensation was, is, or will be directly or indirectly related to the specific recommendations or views contained herein.

Page 27/27