Q&A with Alan Boyce: Freddie Mac and Inverse Floaters

rcwhalen's picture

Updated -- Further to my earlier post, Freddie Mac Mortgage Predator | Alan Boyce on Inverse Floaters <http://www.zerohedge.com/contributed/freddie-mac-mortgage-predator-alan-..., below is a Q&A prepared by Alan Boyce, CEO of The Absalon Project, on the uproar regarding the way that Freddie Mac manages its portfolio.

Absalon is a joint venture between George Soros and the Danish financial system.  Previously, Mr. Boyce was the senior managing director for investment strategy at Countrywide Financial Corporation, where he was responsible for secondary markets, the hedging of mortgage servicing rights, and the balance sheet for Countrywide Bank and Balboa Insurance.  Bio: <http://www.milkeninstitute.org/events/gcprogram.taf?function=bio&EventID... -- Chris

Q&A with Alan Boyce: Freddie Mac and Inverse Floaters

Link to folder with all PDFs for inverse floaters <https://docs.google.com/#folders/0B_PWOgU_OFjkMTBiNTRmNmYtMjVhMy00NmJkLW...


1.           Isn’t it meaningless to look at the inverse floaters in isolation? To assess risk, shouldn’t we look at the entire portfolio held by Freddie Mac?

Analyzing Freddie Mac’s portfolio as whole would be the best way to measure its risk, but the enterprise has never disclosed the securities that it is holding, nor has it announced any intentions to change this practice. Without full information, analysts can only look at what Freddie Mac is forced to disclose, which is precisely what led to the discovery of these highly levered inverse floaters. To truly evaluate the risk of the firm, analysts would need to understand the composition of the entire portfolio, the hedges and liability structure. However, until this occurs, commentators must use the information that is available.

The US mortgage market represents $10 trillion, with the vast amount of outstanding loans being callable, 30 year, fixed rate notes. Most of these loans are securitized into Agency MBS. Thus we have a very large bond market that is exposed to long-term interest rate risk and prepayment risk. As a holder of $650 billion of mortgage loans and MBS (both straight pass-throughs and structured REMIC parts), Freddie Mac is heavily exposed to long-term interest rate and prepayment risk. While hedging interest rate risk is possible through the mortgage market, there is insufficient market depth for Freddie Mac or any other large market participant to hedge its exposure to prepayment risk. Over the last two decades, hedges have been done in the US Treasury and Dollar interest rate swap markets, which are just large enough to help.

In at least 29 separate transactions over a 6-month period, Freddie Mac provided the MBS collateral and took back an Inverse IO position. This inverse IO position concentrates long exposure to prepayments. In none of these deals did Freddie Mac disclose taking back a straight PO or deeply discounted bond, which might have provided a hedge against prepayments. These transactions do not appear to satisfy any standard types of hedging demand for a large holder of long-term mortgages such as Freddie Mac.

Some have suggested that the agency may have in fact been over-hedged in swaps and swap options, and thus increased exposure to prepayments would actually reduce it risk. Even assuming that Freddie Mac could have over-hedged its massive position, this argument would not make sense. Instead of investing in inverse floaters, Freddie Mac could simply sell the excess hedges and book the profit. Another principle of risk management is to retain liquidity whenever possible. As noted below, these transactions converted liquid MBS into illiquid and highly leveraged debt instruments. This was what traders would call a “Texas Hedge”.

2.           Did Freddie Mac have any choice besides keeping these inverse floaters? Aren’t they difficult to sell off because the bond market sees them “toxic waste?”

Yes, Freddie Mac had a very simple alternative to retaining these inverse floaters: sell the original collateral on the deep, liquid market that already exists for MBS. Freddie Mac instead voluntarily undertook transactions that transformed billions of liquid MBS into floaters and inverse floaters. These inverse floaters were not the small leftovers of a larger securitization; the vast majority of the transactions referenced in the report did not involve the creation of any “vanilla” MBS, only floaters and inverse floaters. Freddie Mac appears to have gone out of its way to create an illiquid, levered position for itself that pays off when refinancing does not happen. While these securities were not originally “toxic waste,” Freddie Mac would now find it exceedingly difficult to sell them. In part, such sales would cause the market to suspect that the agency has inside information on the collateral prepayment speeds.

These deals were done after the agency was in conservatorship and had a mandate to shrink its portfolio. The creation of illiquid assets likely complicates future portfolio reduction.

3.           $5 billion of inverse floaters are only a small component of Freddie Mac’s $653 billion mortgage portfolio. Could such a small portion of the portfolio be a factor in determining policy?

While $5 billion is technically correct, the par, market, or book value of derivative securities is not the right measure of exposure to risk. Inverse floaters are highly levered securities, and their value thus depends more on how many mortgages are underlying the securities than the current value of how much the inverse floaters are worth. The $3.4 billion of inverse floaters referenced by NPR and ProPublica were levered at approximately 6:1, containing the effective risk of $19.5 billion in MBS. It is not clear how the reported $5 billion of inverse floaters are valued or the value of the underlying mortgages. If the leverage on the $5 billion of inverse floaters is similar, it might represent the equivalent risk of approximately $30 billion in MBS. While this still might not appear to be a large component of Freddie’s overall portfolio, consider that agency MBS accounts for $257 billion of the $653 billion total.  Under the Preferred Stock Purchase Agreement (PSPA) governing Freddie Mac’s conservatorship, the enterprise must reduce its mortgage-related investment portfolio by 10% annually.  By transforming $30 billion of MBS on the balance sheet into $5 billion of inverse floaters, Freddie Mac would have reduced its MBS holdings by nearly 10% with almost zero net change in exposure to prepayment risk. It is unclear whether the sale of mortgage principal while retaining prepayment risk is consistent with Congress’s intent when it mandated portfolio reduction. It is clear that Congress did not intend for Freddie Mac to enter into transactions that would maintain economic exposure while significantly impairing its liquidity.

In regards to policy, Freddie Mac only started investing in inverse floaters on a large scale in October 2010, the same month it began tightening credit on refinancings, and the month before it imposed new post-settlement delivery fees to further restrict refinancings. Announcements of these policies were not made publicly, unlike Fannie Mae which publishes its fees on its website. Fannie Mae did not appear to invest heavily in inverse floaters based on its disclosures, and it did not impose the same lending restrictions at the same time. In the HARP 2.0 guidelines announced in November 2011, Freddie Mac tightened credit relative to its own rules, adding new restrictions on refinancing mortgages with loan-to-value ratios below 80 percent. Fannie Mae did not add the same restrictions. About one-half of mortgages with an interest rate above 6% have a loan-to-value ratio below 80 percent. Many such mortgages have second liens or are taken out by borrowers with credit challenges that might make refinancing impossible, but for whom a refinanced mortgage would significantly reduce default risk.

4.           Are there remaining concerns if Freddie Mac stopped making these transactions at the request of the FHFA?

The FHFA halting these transactions is certainly a step in the right direction but does not explain the existence of at least 29 such deals. If there are risk concerns about these deals, analysts might question the role of the FHFA, as the equivalent of the CEO of Freddie Mac, in effectively overseeing and managing risk.

5.           Hasn’t the FHFA publicly stated that decisions related to Freddie Mac’s post-settlement delivery fees and other such policies are made without regards to the agency’s retained portfolio?

The FHFA has said this on record, but this does not provide an explanation for why Freddie Mac’s refinancing policies are stricter than those of Fannie Mae under HARP 2.0. In November 2011, when the GSEs announced their guidelines for HARP 2.0, Freddie Mac imposed harsher restrictions on refinancing than Fannie Mae did for borrowers with an LTV ratio below 80%. Such a move appears to have been in contradiction to one of the primary motives of the conservator, who has vowed to better align policy across the two agencies. As well, the conservator has stated that the enterprises were tasked with creating refinancing plans without regard to their retained portfolios.

From the perspective of default risk, the enterprise would appear to be strictly better off by encouraging as many refinancings as possible. Even borrowers with LTV ratios below 80% face default risk, especially borrowers with high CLTVs or with other credit impairments. So it is unclear why Freddie Mac would impose additional restrictions on refinancing if it were following a mandate to ignore its portfolio in setting standards for refinancing under HARP 2.0. Nonetheless, from a portfolio perspective, based on the positions with the inverse floaters and any other retained MBS, a policy restricting refinancing might make sense. About half of all mortgages with an interest rate greater than 6.5% have an LTV ratio below 80%, so investors in high-coupon MBS directly benefit if these borrowers cannot refinance.

This is not an isolated example. In November 2010, just before Freddie Mac began investing heavily in inverse floaters, the agency announced a substantial increase in its post-settlement delivery fees (a.k.a. PSDFs), which borrowers must pay up-front in order to refinance. If the portfolio were not a consideration, then it would have made more sense to impose a smaller annual fee on top of the mortgage rate. This would have increased cash flows to Freddie Mac and simultaneously reduced the risk of mortgages that it already guarantees. However, the agency instead chose to impose large up-front fees that impair refinancings by credit-constrained borrowers.

6.           Why should Freddie Mac help borrowers like the Silversteins? Aren’t poor credit controls what got us into this mess in the first place?

Everyone agrees that better risk management in the run-up to the crisis would have been beneficial, but this is a separate issue from setting restrictions to refinancings. Freddie Mac already bears the credit risk of the Silversteins’ mortgage. The Siversteins were prevented from refinancing by a new rule established by Freddie Mac in October of 2010.  Prior to this, short sales were encouraged by the GSEs as the most efficient way to remove a homeowner from a house. By putting long lasting penalties upon the willing participants of a short sale, the GSEs have done great damage to one of the few mechanisms that could help resolve the shadow inventory facing the US housing market.

If the Silversteins (or any other constrained borrower with a Freddie Mac guaranteed loan) stop making their mortgage payments, Freddie Mac is legally obligated to buy back the loan from an MBS pool at par. Taking into account legal fees and other transactional costs involved with foreclosure, it can be difficult to recover even half the value of the original mortgage. A policy that allowed homeowners such as the Silversteins to refinance into a more affordable mortgage might have eliminated hundreds of thousands of defaults, potentially saving American taxpayers tens of billions of dollars. Ignoring the impact of refinancings on the portfolio, lower mortgage payments for borrowers are a windfall for Freddie Mac, dramatically reducing its liability for mortgage guarantees.

7.           Freddie Mac points out that it has undertaken more than 4.3 million refinancings over the past three years, with refinanced mortgages making up as much as 80 percent of its new originations. How could Freddie Mac be refinancing so many mortgages if it were really taking steps to inhibit refinancings to benefit its portfolio?

Mortgage refinancings appear to make up a large portion of Freddie Mac’s recent business. Nonetheless, it is possible for Freddie Mac to refinance mortgages for relatively high credit quality borrowers, often with lower mortgage rates, while simultaneously inhibiting refinancing for some borrowers whose mortgages are in the inverse floaters or are held on Freddie Mac’s portfolio. According to the NPR/ProPublica story, the average mortgage rate in the pools of inverse floaters was around 6.5 percent. Such mortgages were likely taken out prior to 2008, when credit standards became markedly tighter. As well, Freddie Mac has said that the mortgage securities underlying the inverse floaters were already in its portfolio. Such MBS were likely purchased before 2008, when Freddie Mac was put in conservatorship and apparently stopped buying new MBS.

The first attached figure reports constant prepayment speeds (CPRs) for Freddie Mac pools of 30-year fixed-rate mortgages based on the coupon in the pool for September-November 2011.  The figure splits prepayments based on whether they were voluntary (typically a refinancing) or involuntary (bought out of a pool by Freddie Mac). As is apparent in the chart, refinancing speeds peak for mortgages with mortgage rates around 5 percent. Mortgages with rates of 6 percent or higher appear to refinance at much slower speeds. The second figure examines the coupon breakdown between mortgages originated prior to the crisis those originated since 2008 and plots the volume of voluntary refinancings. The results are clear: the vast majority of current refinancings are going to borrowers paying relatively low interest rates who have already been able to obtain a mortgage during the credit crunch of the past three years.

This finding matches recent research from Lender Processing Services that shows that prepayment speeds are especially fast for mortgages originated in 2008, 2009, and 2010.  In other words, many of the 4.3 million borrowers who have refinanced are likely relatively recent borrowers who have mortgages with already low rates are those who are refinancing at the highest CPRs. Even borrowers with very recent mortgages from 2010 are refinancing at higher CPRs than borrowers with mortgages from 2007. Recent borrowers also tend to have very good credit and would not need help from programs such as HARP and HARP 2.0 to refinance their mortgages.

Putting the facts together, it is possible for Freddie Mac to do a large amount of business refinancing recent mortgages (sometimes more than once) while also benefitting its portfolio. Mortgage refinancings for recent vintage mortgages would not impact its portfolio much, since Freddie Mac appears to have stopped purchasing new MBS since 2008. At the same time, older mortgages and those with especially high mortgage rates appear to be refinancing at much lower CPRs. These are the same mortgages that likely make up the bulk of Freddie Mac’s retained portfolio of MBS and also are concentrated in the inverse floaters where Freddie Mac would have large losses if prepay speeds were to pick up.

Figure 1: Voluntary and Involuntary Prepayment Speeds by Bond Coupon

 Freddie Mac 30-Year Fixed-rate Mortgages

Figure 2: Projected Balance Voluntarily Prepaid by Bond Coupon and Vintage

Freddie Mac 30-Year Fixed-rate Mortgages

Figure 3: Voluntary and Involuntary Prepayment Speeds by Bond Coupon

 Freddie Mac 30-Year Fixed-rate Mortgages, Originated in 2006 or Later