Mortgage Market: Some Observations

Here are just some thoughts about a few abnormalities in the mortgage market that I have observed.
Jumbo-Conforming Mortgage Spreads:
Jumbo spreads on 30-year fixed rate mortgages remain elevated at approx 100bps versus historical spread of 12-25bps. According to Bankrate.com, the average rate for a conforming 30yr fixed is 5.62% and 6.60% for jumbo 30yr mortgages. Jumbo mortgages are loan amounts above $417,000 which exceed the requirements for Fannie / Freddie securitization. Since jumbo loans are ineligible for Agency MBS, they are less liquid. see article (august)

Since most prime jumbo mortgages are underwritten according to Fannie Mae guidelines, the credit risk between conforming and jumbo loans are the same. That’s why borrowers could usually get almost same rate on a $416k loan as they could a $418k loan. Any spread reflects the liquidity premium of the loan, not the default risk. Now, we have seen that liquidity premium spike affecting the most creditworthy borrowers.

Conforming mortgage rates currently are in-line with the yield on the 10yr Treasury, thus the mortgage crisis has left Agency MBS relatively unaffected since it’s such a highly liquid market. Jumbo AAA paper commands 100bps premium since investor appetite has diminished.

It appears that in part, investors are suspect of the credit quality of any jumbo loan not guaranteed by one of the Agencies even if the originator claims the loan conforms to Agency underwriting standards. Certainly there is a lack of trust with regards to the composition of the underlying mortgages for private label MBS and CMOs. A MBS may be backed by 1000’s of loans from several originators or even hundreds of independent brokers, thus it’s essentially not economical to thoroughly examine each mortgage backing the security. Investors are so far removed from the point of origination that they have to trust proper procedure was followed and rely on legal recourse if originators took a short-cut.

In addition, it appears elevated spread is due to reduced confidence in private insurers ability to guarantee the principal. Historically this has not been a concern since the backers of private label MBS were financially strong and thought to have the wherewithal to handle defaults on the underlying mortgages. Low mortgage defaults coupled with high recovery rates on collateral resulted in a low number of reimbursements required by private insurers. Now that the above situation is reversed, insurers’ ability to cover losses is now a question.

I am somewhat surprised that jumbo spreads have not shown any signs of tightening. According to the graphs from Bankrate.com, jumbo rates began to decrease mid-October, yet reversed course as spreads rose again in November. This is an indication that the non-conforming mortgage market has virtually not improved since the crisis began in July.

LIBOR – CMT Disparity:
Borrowers with LIBOR ARMs scheduled to reset soon are in for a surprise. ARMs reset to a rate which is determined by adding the pre-disclosed margin to the current rate of a specified index. At the time of origination, a borrower can choose which index he/she prefers. Most common indices are the 1-year CMT (Constant Maturity Treasury) and 1-year LIBOR.

Historically, LIBOR has been roughly 50bps (40 bps since 1990) higher than CMT, thus the margin has been 50bps higher for CMT ARMs, hence equating the combined rate (margin + index). Essentially, the choice between the two is a coin flip since both mortgages will reset to nearly the same rate.

Currently, there is a huge disparity between CMT and LIBOR rates. The CMT is approx 3.10 and LIBOR is 4.45, a difference of 135bps which is 85bps higher that the historical relationship. What does this mean? Well, for borrowers that took out a LIBOR 3/1 ARM, three years ago, their mortgage will reset to a rate 85bps higher than if they had chosen CMT. The new rate for CMT ARMs would be 5.85% (2.75 margin + 3.10) and LIBOR ARMs would be 6.70% (2.25 margin + 4.45). Wow!

The short answer for why the CMT-LIBOR spread has become disconnected is because the CMT is truly risk-free rate since it is based on 1-year Treasury yields. LIBOR is the prime rate European banks charge each other and involves credit risk because banks are not government institutions. The long history of a static spread implied credit risk was not perceived as a dominant factor in the determining components of the LIBOR rate. I say this because the LIBOR spread to Treasuries had been constant, suggesting that the mark-up was more customary as opposed to changes in risk perceptions. Recently this has not been the case whatsoever. The credit risk premium is a dominant factor in LIBOR rates as evidenced by the 1yr LIBOR rate remaining high while the risk-free rate has declined. This is a clear indication that the credit quality of banks has significantly deteriorated due to risky mortgage holdings.

FNMA - FHLMC MBS Price Spread:
Another thing that I have noticed recently is that Freddie Mac MBS are trading about 4 ticks or 1/8 pt worse than Fannie Mae MBS. This is strange because Freddie Mac Gold PC pays on the 15th of the month while Fannie Mae payments are 10 days later on the 25th. Since Freddie MBS CFs occur sooner than FNMA, investors should “pay up” for Freddie MBS because its PV of cash flows is larger than PV of FNMA MBS, all else equal.

Freddie Mac MBS always traded at a premium to Fannie throughout my experience on a mortgage trading desk, just as logic would suggest. Recently, Fannie MBS has garnered a slightly higher price despite the payment delay, which raises the question “Why?” Ostensibly the price inversion is due to concern over Freddie Mac’s ability to back its mortgage bonds. Obviously, fears are not paramount, else the spread would be much greater, but this does illustrate that investors are paying attention to something that historically was just a given.

Generally speaking, Freddie and Fannie MBS have had the perception of being free of credit risk due to the implicit guarantee by the Federal Government. Ginnie Mae MBS carries the full faith and credit of the US Government. Agency MBS are backed by high quality mortgages and 80 LTV (PMI required if LTV >80); Fannie and Freddie haven’t ever incurred significant losses from mortgage securitization. In fact, they make a handsome profit securitizing, because the revenues generated from guarantee fees (insurance premiums) far exceeds the losses covered on mortgage loans.

Now, there are questions about the actual risk inherent to both Freddie and Fannie. Both have engaged in risky side businesses that may have impaired their primary business of backing prime MBS. Another concern is the possibility of a higher than expected default rate on mortgage collateral of MBS they issue. Hence, guarantee fees charged for securitization are way to low with respect to actual losses.

Given the major problems announced at Freddie Mac recently, investors are apprehensive as evidenced by Freddie MBS trading back to Fannie MBS, the reverse of historical norms.