The latest hot issue in the mortgage markets has been the 10
basis point addition to G-fees mandated by the payroll tax cut passed late last
year. Unfortunately, the media has only
recently picked up on the fact that this “tax cut” was financed by a hidden tax
on mortgages. However, since neither
Congress nor the President know much about mortgages, details on how this tax
would be imposed were not included in the bill.
The legislation did not state, for example, whether the 10 basis point
g-fee surcharge was considered part of the guaranty fee or a separate add-on to
the loan’s rate.
Read: Tax Cut Extension Now Officially Raising Mortgage Rates
This ambiguity created a number of issues for lenders. Most servicing contracts specify that the
total amount of interest either bought up or down with the GSEs cannot exceed a
fixed sum, i.e., 37.5 basis points (which is fairly standard language). If the total g-fee (including the surcharge)
is increased by 10 basis points, the amount of excess servicing that can be
“bought up” is thus reduced by the same 10 bps.
While buy-ups used to be relatively rare (since Freddie and Fannie both
assign putrid multiples to the cash flow), the lack of ready buyers for
servicing means that another avenue for selling excess servicing has been
The additional fee also has played havoc with loan
pricing. As I understand it, the GSEs’g-fee
buy-down multiple (i.e., what they will charge to convert the g-fee into a
single payment paid when the loan is pooled) is around 6x. This means that the 10 bp surcharge is worth
around 60 basis points in price (or 0.6% of a loan’s face value). The uncertainty surrounding the surcharge also
threatened to disrupt pooling practices.
To this point, there has been no guidance on whether the entire g-fee
(including the surcharge) actually can be bought down. If it couldn’t, that would mean that 30-year
3.75% loans could not be securitized into agency 3.5% pools, the lowest coupon
with good liquidity and well-behaved pricing.
If these loans can only be pooled into Fannie or Gold 3s, this could
severely disrupt both the MBS market and loan pricing. (As I discuss later, heavy originator selling
will in turn increase the points required for lenders to offer these loans—or,
more simply, make it more expensive for borrowers to obtain low rates.)
I’ve been told that both Freddie and Fannie will have a call
today (Friday, 2/10) to clarify the treatment of the surcharge. As I understand it, the surcharge will be
treated as part of the g-fee, meaning that it can be bought down entirely. This will allow 3.75% loans to be pooled in
3.5% pools; while it will be relatively expensive to buy down the entire g-fee,
it does give lenders the option of either pooling down (into 3s) and holding/selling
the servicing, or pooling into 3.5s by buying down the g-fee.
This in turn raises the interesting and important issue of
how the lower coupons are trading. There
is decent liquidity in 15yr 2.5s, with average monthly issuance of 585mm over
the last three months. However, the
market for 30-year conventional 3s remains relatively illiquid, although
production seems to be ramping up; according to Fannie Mae, roughly 230mm 3%
pools have been created thus far in February.
While the overall price performance of Fannie 3s is better than it has
been in the past (when prices could move up or down by a half-point or more on
minimal volume), the dollar rolls remain pretty volatile. A decent origination day will push prices in
the back months (April and May) sharply lower, while front months (February and
March) remain roughly in line with other coupons, duration-adjusted. This week, for example, saw the March/April
roll (i.e., the price difference between March and April) for Fannie 3s as wide
as 30/32s (equal to a financing rate of -7%); the roll is now 21/64s (for a
more moderate -0.71%).
There is a complex interactive relationship between a
security’s price performance and issuance.
A coupon won’t perform well if its outstanding balance remains small; a
small tradeable “float” means that prices can get pushed around on very limited
volume. However, no originators will
issue pools for securities with erratic performance. The biggest factor inhibiting the growth of
the market for 30-year 3s is the belief that there is no natural clientele for
the security. Its coupon is too low, and
its duration too long, for banks and depositories; insurance companies and
pension funds that would buy their long average lives and durations need higher
returns on their portfolios. In my mind,
the only natural holder is the government (or, more to the point, the
Fed). A program to purchase 30-year 3s
(and 15-year 2.5s, to a lesser extent) would be the most direct way for the government
to positively influence the mortgage markets.