​My interest was recently piqued by a report in National Mortgage News entitled “Confidential Report:  Jumbo Conduits Folding.”  The report stated that “…’several’ conduits formed over the past year have been shuttered, their investment capital returned to investors.”  While arcane developments in the MBS market may seem irrelevant to mortgage lenders, they have actually had a major impact on the markets, and will affect the lending business for years to come.

​In order to understand these developments, a brief primer on non-agency MBS and its differences from agency pooling practices will be helpful.  In an agency pool, the credit support comes from the government agency in question, and is funded by paying a guaranty fee to the agency.  (It’s effectively an insurance policy with the premium being paid by diverting a portion of the loan’s interest payment to the “insurer.”)  By contrast, the typical non-agency (or “private-label”) structure does not use insurance as the primary source of credit support, either from the GSEs or from a private party.  They instead use a technique called “subordination.”  This means that the deal is divided into separate sectors that have different priorities with respect to receiving cash flows from the loan collateral.  Generally speaking, the “senior” sector has priority for receiving principal and interest from the investment entity that holds the collateral; the “subordinate” sector receives cash once the seniors are paid.  (Note that private-label deal structures can be enormously complicated; this discussion is best treated as a simplified introduction.)

​The amount of subordination required for the senior class (or classes) is determined by the rating agencies, and this is an area that’s created great difficulties in the market. Both the amount of subordination and the price it ultimately garners are key factors in the overall execution of the deal; in turn, the execution determines the prices that can be paid for the loans that will collateralize the deal.  (This is directly analogous to how agency-eligible loans are priced, except that the inputs are different.)  Historically, deals have been utilized in order to distribute lenders’ production into the capital markets; getting “sale treatment” for accounting purposes is important to lenders in order to prevent their balance sheets from ballooning.

The rating agencies have been rightly singled out as culprits in the 2007-2008 mortgage debacle by assigning unrealistically low subordination levels to deals, which made the funding of many types of dubious loans and products economical.  In the current market, however, the rating agencies have made their models unduly onerous, to the point where even high-quality loans are treated as very risky assets.  As a result, the amount of subordination required for deals to create triple-A senior tranches is so large that non-agency securitization is virtually uneconomical.  This is especially true for loans with some credit blemishes.  The subordination required for deals backed by purchase loans with, for example, FICO scores between 720-740 and LTVs of 80%, are so onerous that the deals are uneconomical to issue.

This in turned has directly impacted the business of lenders.  While published Jumbo loan rates are currently about 55-60 basis points over conforming rates, these quotes are misleading, since they are only for the very highest quality “super-prime” borrowers (with LTVs <=70% and FICOs of 760 or higher).  Moreover, current jumbo rates (around the 4.60% area) are available only because banks are willing to hold high-quality jumbo loans on their balance sheets.  If they were pricing these super-prime loans based on securitized execution, rates would be at least 40 basis points higher.  More ominously, the rates that fall out when running moderate-credit loans approach predatory levels, by legal standards.  The only way such loans can be made at all is by non-depository originators that are willing to hold either the loans, or the deal’s subordination, on their books.  (Put differently, they are not concerned about getting “sale treatment,” and can treat the securitization as a form of long-term financing.)

In my mind, this has seriously impacted the mortgage and housing markets, especially in high-cost states such as California.  Rather than having a reasonably seamless transition between conforming- and jumbo-balance lending, the current lending environment has broken down into multiple tiers, each with different rates structures and credit demands.