Why Agency Mortgage Buyouts Will Not Prove To Be The Fed's MBS Purchasing Exit Panacea
Goldman's Alec Phillips provides some perspectives on why the recent announcement that the GSEs will purchase 120+ day delinquent loans will not be a major impact to either the overall MBS market or to the Fed's QE tactics when it comes to stabilizing the market. Furthermore, in addition to not being a material mitigant to the Fed's massive balance sheet holdings, some direct implications as pertains to end borrowers include a greater incentive to foreclose, a greater emphasis on non-modification strategies, and more aggressive modification efforts. Yet the major issue, the one addressed by the Fed's Hoenig yesterday when discussing the urgent need to commence selling Fed assets asap, will likely be completely sidestepped, in essence stressing the need to find a legacy replacement through to QE when it expires in just over a month.
Agency Mortgage Buyouts: Hardly an Offset to the End of Fed Purchases
Last week, Fannie Mae and Freddie Mac announced that they will purchase loans that are 120 days or more delinquent out of the mortgage backed securities that they guarantee. This will mean the purchase of $72 billion (bn) in loans out of Freddie-backed securities by March and $117bn from Fannie securities over the next few months. Assuming that the GSEs continue this policy for loans that become seriously delinquent over the coming year, this implies a grand total of $380bn reduction in the outstanding balance of existing GSE MBS, around $300bn of which take the form of prepayments to bondholders.
Two related issues led to this change in policy. First, the Treasury announced on December 24 that it would expand its capital assistance to the GSEs, and more importantly, that it would raise the limit on the retained portfolios that the GSEs can hold, effectively allowing the enterprises to increase their holdings by $110bn by the end of 2010, rather than decreasing the cap to $150bn under their combined holdings as of year end 2009. This freed up room in the GSEs’ portfolios to hold the repurchased loans.
The second and more directly related issue was the implementation of SFAS 166 and 167, which reduced the accounting obstacles to buying loans out of a securitization. The GSEs are required to purchase loans out of securitizations (1) before modifying them, (2) when a foreclosure has occurred, or (3) when the loan is 24 months delinquent. The enterprises also have the option to purchase a loan once it is four or more months delinquent, but aren’t required to do so. Prior to 2010, purchasing a loan out of a security resulted in a charge off against the GSEs’ loss reserves equal to the difference between the par value paid to trust and the fair value of the loan. Such a loss eats into capital, which given the GSEs zero net worth at present, means greater capital infusions from the Treasury. In short, previous accounting rules provided the GSEs with motivation to avoid purchasing loans out of the pool for as long as possible.
Accounting changes that took effect January 1 changed this. The implementation of FAS 166/167, which requires the consolidation of off-balance-sheet entities, has the effect of consolidating the guaranteed securities on the balance sheet, eliminating the need that existed until now to recognize a loss due to the purchase of a loan out of a pool due to modification, foreclosure, or delinquency. In theory, this should encourage not only delinquency-driven buyouts, but also modifications. In practice, public policy considerations rather than accounting rules have dictated modification activity recently, so accounting changes should have the greatest effect on buyouts of delinquent loans, where they seem to have been a greater impediment.
Implications for Foreclosure Mitigation Policies
The fact that the GSEs could soon be large property owners—$380bn in seriously delinquent loans, at an average balance of around $200,000 per loan, means nearly 2 million properties—raises the importance of what policies the GSEs set to deal with these loans. The fact that these loans will be bought out does not necessarily change how the GSEs will deal with them, but it does raise several interesting possibilities:
1. A greater incentive to foreclose? Although accounting changes reduced the obstacles to buying out loans, the Treasury-imposed portfolio cap could still affect how those loans are dealt with once purchased. If the GSEs allow them to remain in the retained portfolio, this will raise the balance toward the cap. In the case of Fannie Mae, this could be an important issue, since it has limited room under its cap (see below) and could be forced to sell MBS out of its portfolio to make room. Foreclosing on the property would be one way out of this situation, since it would liquidate the loan and result in real estate owned (REO) on the balance sheet, which does not count against the Treasury imposed cap. It is worth noting that the loans in question are seriously delinquent, and many are likely to end in foreclosure eventually regardless of the buyout policy.
2. A greater emphasis on non-modification strategies? If many of these loans are put into foreclosure, an interesting possibility is the potential for new policies related to foreclosed property. In November 2009, Fannie Mae announced a ‘deed-for-lease” program, which allows homeowners to relinquish ownership of the property and rent it instead. In January, Fannie Mae took another step in this direction with the announcement of a policy to allow tenants in Fannie-Mae foreclosed property to rent the property on a month-to-month basis after foreclosure. To the extent that the GSEs view property values as bottoming this year—Fannie, for instance, forecasts a 0.2% decline in the FHFA index this year, followed by a 2.3% rise in 2011—holding a greater amount of housing inventory off the market while generating income could reduce their losses.
3. More aggressive modification efforts? The Treasury’s decision in December to provide unlimited capital to the GSEs for the next three years increased speculation that the enterprises could pursue more aggressive modification efforts, including principal reductions. Last week’s buyout announcement seems to have added a little more fuel to this fire, largely because it increases the number of loans held in portfolio and that are therefore able to be modified. However, while it seems likely that the Treasury and the GSEs will pursue some additional targeted modification efforts, we are skeptical that this, or the December announcement, is a precursor to broadly applied principal reduction, for three reasons: (1) The cost would be significant. Fannie and Freddie reported that 14% and 17% of the loans in their respective single family loan books had some negative equity as of September 30, 2009, which combined negative equity of somewhere around $100bn between the two entities. An approach targeting only delinquent borrowers could reduce the cost, but the perceived moral hazard such programs create seems to have kept the Treasury from proposing them in the past. (2) The logistical obstacles are still enormous. Difficulties in qualifying borrowers and working through second liens have stymied previous efforts, and while they may not be quite as challenging among borrowers with conforming mortgages as they have been with non-agency borrowers, they would still present an obstacle. (3) Differentiation based on securitization status could prove unpopular. The Home Affordable Refinance Program (HARP) which allows borrowers with agency-securitized loans to refinance with up to a 125% current loan to value ratio, ran into criticism on these grounds. Making government subsidized principal forgiveness available based on what a borrower’s lender decided to do with the loan could provoke an even stronger reaction from ineligible borrowers. The administration (or financial institutions) would probably come under great pressure to provide a similar benefit for non-agency borrowers as well, which would probably mean reopening TARP. Since the administration has oversight of major GSE decisions, it seems unlikely they would want to go down this road.
Implications for the Transition away from Fed MBS Purchases
When the Treasury announced that it would incrementally raise the cap on the GSEs retained portfolio, this created a potential opportunity for the GSEs to become bigger buyers of their own MBS as the Fed wound down its purchases. However, the agencies are unlikely to become a meaningful substitute for Fed purchases.
1. GSE programs are no replacement for QE. While both the GSEs and the Fed can target mortgage rates through purchase of MBS, the Fed’s strategy goes beyond this. Most importantly, the expansion of the Fed’s balance sheet—a good deal of which is the result of MBS purchases—has been funded by the creation of bank reserves. By contrast, GSE purchases will be funded with agency debt issuance (and possibly the draw down of some of the fairly large cash balances the GSEs currently hold), so neither the amount of federally sponsored debt outstanding nor the amount of excess bank reserves will change as a result of the buyouts. In this sense, the program is much more akin to the Treasury’s purchase of agency MBS than it is to the Fed’s program.
2. Prepayments could boost demand for MBS… Although the GSE purchases won’t serve the same purpose as the Fed’s program, they will put around $300bn back into the hands of investors, much of which will presumably be recycled back into the market to create a bid for agency MBS that should help to offset the decline in Fed purchases. To put this in context, the $200bn in buyouts likely to occur over the next three months or so will result in a monthly pace slightly higher than the Fed’s current pace of purchases.
3. …But will reduce GSE capacity to intervene in the market. The GSEs’ retained portfolios could play an important role in lowering volatility in mortgage rates, and this is where the buyback announcement could change things the most. The GSEs have $92bn in combined portfolio capacity under Treasury imposed caps, which is not nearly enough to accommodate the $200bn in currently 120-day delinquent loans plus another $180bn or so in loans that may become seriously delinquent before the end of the year. However, there are a few offsetting factors that should allow the GSEs to accommodate most if not all the buybacks without resorting to selling other assets to make room. Most importantly, the cap doesn’t apply until the end of the year, and the GSEs will have somewhat more flexibility in the interim. In addition, the GSEs already hold about 20% of their total MBS float in their retained portfolios; disregarding these loans reduces the needed capacity by around $75bn (380x20%). Prepayments and normal runoff should also free up another $225bn on an annual basis at the current liquidation rate. All in, this leaves only about $80bn in needed capacity between the two GSEs, or $12 billion to spare compared with current levels. Unfortunately, their capacity is distributed unevenly: Freddie appears to have more than enough room, while Fannie seems likely to press against its cap if not break it. If it appears likely to do so, Fannie would presumably sell a small amount of the $400bn in MBS it holds in its portfolio to make room.
4. Buyouts should very modestly reduce MBS holdings on the Fed’s balance sheet. The Fed is now a larger holder of GSE MBS than the GSEs themselves, but the impact of the buyouts on its balance sheet should be very limited. The Fed’s balance sheet stands at $2.29 trillion, of which $971 billion is comprised of Fannie, Freddie and Ginnie MBS. (Technically, the Fed had purchased a net $1.18 trillion as of last week, but there is a long delay between the weekly FRBNY report of purchases and the time they show up in the weekly report of Fed holdings). If buyouts were entirely proportional to the amount of securities held, the Fed would see a prepayment of roughly $50bn. To the extent that it had bought any of these securities above par, it could also record a small loss. However, the Fed has tended to purchase lower coupon MBS in its program, which hold far fewer delinquent loans and thus should be less prone to buyouts than the higher coupon securities held in private hands. As a result, the Fed is likely to realize only around $15bn or so in prepayments, or 1.5 weeks of purchases at the current pace.
Readers who wish to learn more are encouraged to read Harley Bassman's take on this topic: "The Reverse Robin Hood"
- advertisements -