Econintersect: The average thirty-year fixed-rate mortgage is now near its historic low of about 3.3%. However, the mortgage-backed-securities (MBS) market, where bundles of mortgage loans are sold to investors, have been even more dramatic. In fact, all else equal, had these declines passed through to loan rates one-for-one, the average mortgage rate would now be around 2.6%.
The following was published today by Liberty Street Economics / New York Fed:
Why Isn’t the Thirty-Year Fixed-Rate Mortgage at 2.6 Percent?
Andreas Fuster and David Lucca
As of mid-December, the average thirty-year fixed-rate mortgage was near its historic low of about 3.3 percent, or half its level in August 2007 when financial turmoil began. However, yield declines in the mortgage-backed-securities (MBS) market, where bundles of mortgage loans are sold to investors, have been even more dramatic. In fact, all else equal, had these declines passed through to loan rates one-for-one, the average mortgage rate would now be around 2.6 percent. In this post, we summarize some of the findings from a workshop held at the New York Fed in early December aimed at better understanding the drivers behind the increased wedge between mortgage loan and MBS rates.
The chart below shows the recent evolution of the gap between the two rates (the “primary-secondary spread”) as measured by the difference between a representative yield on newly issued agency MBS (the “current coupon,” or secondary rate) and the average thirty-year fixed loan rate (the “primary rate”) from the Freddie Mac Primary Mortgage Market Survey. The spread has increased 70 basis points, on net, from around 45 basis points in 2007 to about 115 today, implying that declines in the primary rate have been smaller than those in the secondary rate. So, why is the primary rate at 3.3 percent and not at 2.6 percent today?
As we noted in a white paper prepared as background material for the workshop, a number of factors may be at play. First, the primary-secondary spread is only an imperfect proxy for the degree of “pass-through” between the MBS market and the primary market. For one thing, we can’t observe the MBS yield directly, but instead must compute it under a number of assumptions that are particularly sensitive to misspecification in the current environment. Also, the spread doesn’t take into account the guarantee fees on loans charged by Fannie Mae or Freddie Mac (the agencies or “GSEs”), which have increased from 20 to 25 basis points before 2008 to about 50 basis points today.
Rising Cost or Rising Profit?
A clearer picture of the link between MBS and primary markets comes from an alternative measure that we call “Originator Profits and Unmeasured Costs,” or OPUCs. The OPUC measure captures the loan originator’s average revenue from selling a mortgage in the agency MBS market (after accounting for the guarantee fee) as well as the revenues from servicing the loan and from points paid upfront by the borrower. As the name implies, OPUCs represent either lender costs (other than the guarantee fee), lender profits, or a combination of the two.
Recent movements in OPUCs, shown in the chart below, are similar to those in the primary-secondary spread. OPUCs have increased significantly, from below $2 (per $100 loan) in the 2005-08 period to a record high of about $5, suggesting that even after accounting for the higher guarantee fees, something unusual may be occurring in mortgage markets today.
Much of the discussion at the workshop, like the analysis in the paper, revolved around the question of why OPUCs have grown so substantially. First, OPUCs could be unusually high in order to compensate originators for increased put-back risk (the risk of having to repurchase some delinquent loans from the GSEs), possible increases in pipeline hedging costs, or other loan production expenses. They could also be overstated due to a decline in the value of mortgage servicing rights. While it’s hard to measure each of these components, we tentatively concluded in the paper that these costs don’t seem to have changed sufficiently to offset the increase in OPUCs, suggesting that profits in mortgage origination have likely increased. Workshop participants overall agreed with this analysis, but also countered that while each cost increase was small on its own, their sum was a sizable amount, offsetting part of the OPUC rise. In addition, some panelists pointed to more time-intensive underwriting procedures as contributing to higher origination costs. Overall, the discussion didn’t refute the alternative explanation of higher profit levels. So what would drive such an increase in the profitability of mortgage origination?
In the paper, we discuss how lenders’ pricing power appears to have increased over the past couple of years, especially on refinancing loans. For example, we document how rates on refinancing loans, in particular those under the Home Affordable Refinance Program (HARP), have been higher than those on purchase loans, for which the market appears to be more competitive. We also note, however, that originators’ capacity constraints are likely the key driver behind the currently elevated profit levels. The chart below plots the OPUC measure against the mortgage application volume index from the Mortgage Bankers Association. OPUCs tend to increase whenever application volume rises rapidly. As discussed at the workshop, a likely interpretation of this relationship is that in such times, lenders can’t handle all the demand they would get if they lowered rates further, and thus they keep their offered rates high.
In principle, when profits are high, one would expect existing firms to expand capacity or new competitors to enter the market. But workshop participants pointed out that these channels may be impaired today because of the decline in third-party originations by mortgage brokers or correspondent lenders, uncertainty about how long the high origination volumes will last, lack of clarity about future regulations, more general fear of future liability risk, or difficulty hiring qualified underwriters. As a result, profit levels and the primary-secondary spread could stay higher than normal.
What’s Next?
Given the discussion thus far, will mortgage rates eventually hit 2.6 percent if MBS prices stay at current levels? This is unlikely, we think, in part due to the increase in guarantee fees, which alone accounts for 25 to 30 basis points of the rise in the primary-secondary spread. Whether or not average mortgage rates could fall more modestly to, say, 3 percent depends crucially on the degree to which the capacity constraints in the mortgage origination industry could be alleviated and the lenders’ apparent pricing power reduced. With these issues in mind, some policy changes may be worth further consideration. For example, extending representation and warranty reliefs to different servicers for streamline refinancing programs (such as HARP) may help increase competition for these loans. More generally, making more agency loans eligible for such programs would likely lower underwriting burdens and loosen the industry’s capacity constraints without increasing the GSEs’ credit risk exposure (as the loans are already guaranteed by these institutions). Relaxing the GSE minimum net worth requirements and volume caps for loan sellers and servicers could also positively affect capacity and entry in the industry. While each of these proposals requires further study, we see these topics as staying at the forefront of the policy discussion. The pass-through from financial market conditions to consumer rates has always been of interest to central banks, but as noted at the workshop by New York Fed President William Dudley in his opening remarks and by Boston Fed President Eric Rosengren in his keynote speech, it’s of even greater interest today as large-scale purchases of agency MBS—aimed in part at lowering mortgage rates—have become a key monetary policy instrument.
Disclaimer
The views expressed in this post are those of the authors and do not necessarily reflect the position of the Federal Reserve Bank of New York or the Federal Reserve System. Any errors or omissions are the responsibility of the authors.
Andreas Fuster is an economist in the Federal Reserve Bank of New York’s Research and Statistics Group.
David Lucca is an economist in the Research and Statistics Group.