A report from the Federal Housing Finance Agency (FHFA) says that while the average fee charged by the government sponsored enterprises (GSEs) for providing a loan guarantee (g-fee) has more than doubled since 2009, both pricing differences and fee equity have increased.  The FHFA report is required for annual presentation to Congress.

Fannie Mae and Freddie Mac, the GSEs, acquire single-family loans from lenders, some of which they hold in their own portfolios but most of which are securitized in the form of mortgage-backed securities (MBS) and sold to investors.  The GSEs guarantee timely payment of interest and principal from borrowers to investors in these securities and in return charge the lender (seller) a g-fee to cover three types of costs they expect to incur.  Costs include what they expect to lose on average through borrower default, the costs of holding economic capital against possible catastrophic losses from borrower default, and general and administrative expenses.  Cost of capital is by far the most significant of these expenses, according to the FHFA.

Mortgage News Daily's Matt Graham doesn't see that as a complete list, however.  "I think FHFA should be more thorough in explaining the logic behind what looks like arbitrary g-fee padding.  They gloss over the cushion by referring to it as 'fee equity,' but is all that fee equity really needed for the sound functioning of the post-crisis mortgage market?  Some might say this is currently a significant source of revenue for the federal government.  Others might say this is simply in line with FHFA's strategic goals of encouraging private capital participation."  

The GSEs use proprietary models to estimate the amount of capital that needs to be set aside as loss reserves.  These models are built around key macroeconomic assumptions including house price appreciation and volatility, economic stress paths, and target rate of return on capital.  In addition there are loan level variables that affect the cost of guarantees including borrower characteristics such as credit history, property type, occupancy, interest rate, and mortgage type. From these models the GSEs calculate a "gap", the difference between the actual fee charged and the estimated cost of the guarantee.  If the gap is positive or zero the GSE expects to achieve at least its target rate of return.  If the gap is negative it may still make money on the loan but not reach its target rate of return.

There are two types of guarantee fees - ongoing and upfront.  The former is collected each month over the life of the loan while upfront fees are one-time payments made by lenders when the loan is acquired by a GSE.  Fannie Mae refers to upfront fees as loan level pricing adjustments and Freddie Mac calls them delivery fees.  To date the GSEs have relied primarily on upfront fees to reflect differences in risk across loans as opposed to ongoing fees.

Prior to March 2008 g-fees had been primarily based on product type rather than loan attributes but due to the deteriorating housing and mortgage markets the GSEs increased overall fees and introduced upfront fees based on a borrower's downpayment and credit score and a 25-basis point adverse market fee.  Later that year the GSEs refined their borrower-based upfront charges and other fee components and since then has gradually raised fees to better reflect credit risk.

In 2011 Congress passed the Temporary Tax Cut Continuation Act which extended a previous one-year reduction in payroll (Social Security and Medicare) taxes.  This was funded by an increase of 10 basis points in all GSE-backed loans.  This fee accrues to the U.S. Treasury and not to the GSEs.

In 2012 FHFA again directed the GSEs to raise g-fees to encourage more private sector participation, reduce the GSEs' market share, and more fully compensate taxpayers for assuming mortgage risk.  Graham wouldn't quite use those same words.  "Let's please not kid ourselves about the fact that G-fees are intentionally jacked up higher than they need to be to cover costs or to protect taxpayers.  Whether or not the cushion is justified, the direct consequence is higher rates for consumers.  In a real sense, part of what FHFA calls 'compensation for taxpayers assumption of risk' is the same money that taxpayers just shelled out in the form of higher mortgage rates!"

From 2010 through 2012 guaranteeing a 15-year loan was expected to be less costly than guaranteeing a 30 year loan and the 2012 increase sought to remove this difference by a greater increase of the guarantee fee on longer term loans.  The 30-year loan fee was raised by 17 basis points and the 15-year by 9.  The figure below shows that the difference between the loan times diminished in 2013 and that sizeable negative gap exhibited by 30-year loans in 2012 largely disappeared in 2013 when the impact of the 2012 fee increase was fully realized.

The 2012 fee increase was also designed to reduce differences between loans originated by small lenders and large lenders.

An additional fee change was announced in December 2013 which would have increased fees by another 10 basis points, adjusted upfront fees by risk categories, and removed the 25-basis point adverse market charge for all but four states.  When Melvin Watt was confirmed as director of FHFA he suspended implementation of these fees pending further review and has requested industry input on the changes.   The input period ended on September 8, 2014, and FHFA is currently in the process of reviewing the comments that were received.

In its report FHFA summarizes the fee changes and their impact over the past six years:

  • guarantee fees increased to an average of 51 basis points in 2013 compared to an average of 36 in 2012 and 22 in 2009;
  • the g-fee increases in 2012 reduced pricing differences between small and large lenders in 2013;
  • the percentage of loans sold to Fannie Mae and Freddie Mac by extra-large lenders decreased from 60 percent in 2010 to 49 percent in 2013 while the percentage of loans sold by extra-small lenders increased from 8 percent to 19 percent over the same time period