The recent actions of the Federal Reserve have pushed fixed mortgage rates to all-time lows.  The Freddie Mac survey rate, a time-honored proxy for consumer rates, recently hit an all-time low of 4.87%, and qualifying borrowers can find sub-5% rates fairly easily.  The Fed has acted to push both intermediate- and long-term Treasury rates lower by buying Treasury securities outright; they have also bought huge amounts of agency debentures and MBS in order to drive down primary mortgage rates.  The Fed is expected to continue these activities, as they are part of the arsenal they are using to fight the financial crisis and the resulting recession.  In turn, this suggests that mortgage rates won't be allowed to rise significantly for the foreseeable future.

Mortgage market participants are obviously wondering whether fixed mortgage rates could conceivably go lower.  Assuming that the Fed's activities keep Treasury yields around current levels, agency MBS spreads over Treasuries will need to tighten significantly in order for rates to drop further.  A cursory look at the MBS market would suggest that agency TBAs are trading at very tight levels by recent historical standards.  The spread of the Fannie Mae current coupon is currently around 170 basis points over the interpolated 5-10 Treasury blend, well inside the 208 basis point spread that prevailed at the beginning of the year.  Moreover, the current coupon spread over the 10-year Treasury yield is the tightest it's been in years.  (A few definitions:  The Fannie Mae current coupon is an interpolated rate based on the dollar prices of Fannie Mae passthrough coupons above and below parity.  It's a more useful tool than looking at the yield spread of any single coupon since, by definition, it is not impacted by different prepayment assumptions.  The 5-10 Treasury blend is the average of the 5- and 10-year Treasury yields, or a proxy for a 7.5-year Treasury.  It is often used as a benchmark for MBS, since mortgages are assumed to move in line with that part of the curve.)

However, any attempt to judge the relative level of MBS spreads must account for recent changes in the shape of the "yield curve," i.e., the rates for Treasuries with different maturities.  The long end of the yield curve has been particularly volatile.  At the peak of the financial crisis in November, the spread between yields on the 5- and 10-year Treasury notes was almost 140 basis points (a very "steep" curve, by recent historical standards).  By the beginning of 2009 the spread had narrowed to around 65 basis points, but immediately began widening again; by the middle of February, it had widened to over 100 basis points.  The spread narrowed (i.e., the curve "flattened") briefly after the announcement of the Fed's Treasury purchase program, as 10-year yields decreased by almost 50 basis points on the day of the announcement (March 18th); however, it has since widened modestly to around 105 basis points.

The steepening of the yield curve since the beginning of the year has interesting implications for observers looking at the relative level of mortgage spreads and, by implication, a sense of where primary mortgage rates could be headed.  I'd argue that MBS spreads are not as tight as they seem when a shorter maturity Treasury is used as a benchmark.  The attached chart shows the current-coupon spread over 5- and 10-year Treasury yields as well as the 5-10 interpolated Treasury blend.  It also shows the average spread of the current coupon MBS over the 10-year Treasury for the period between April 2006 and early August 2007.  (Using the average spread over the 10-year Treasury as the average is arbitrary; given the flatness of the yield curve over the period in question, the average spreads over the different benchmarks differ by only a few basis points.)  The chart indicates that while the spread between the current coupon and the 10-year Treasury is inside its average level during the more tranquil period defined above, the spread over the 5-year Treasury remains quite wide.

Viewed in this way, MBS spreads are not nearly as tight as they initially appear.  Observers looking to gauge the richness or cheapness of MBS relative to recent history will rightly ask which spread is the "right" one to use in the analysis.  I'd argue that the spread versus the shorter benchmark (i.e., the 5-year Treasury) is more useful in assessing MBS spreads.  As noted previously, primary mortgage rates remain around their lowest recorded levels.  This means that the entire mortgage market is highly refinanceable, and even recently funded loans will be able to be refinanced with only a moderate drop in rates.  In turn, this is reflected in the fact the fixed-rate MBS are trading relatively "short," i.e., MBS prices are exhibiting a relatively low sensitivity to changes in the level of interests rates.  Since MBS prices are moving in line with shorter-term Treasuries, the spread of the current coupon over the 5-year is, I believe, a more useful indicator of the relative level of spreads relative to their recent history than spreads over either the 10-year Treasury or the 5-10 blend.

This suggests that MBS may have more room to tighten versus Treasuries, which (if Treasury rates remain in the current area) implies lower primary mortgage rates.  Of course, other forces will collectively impact MBS prices and, in turn, primary mortgage rates.  I expect the Fed's buying programs to continue.  Their buying will be particularly important in the MBS sector, sopping up much of the flood of recent supply.  (Issuance of Fannie Mae 30-year passthroughs hit an all-time high of $72.3 billion last month.)  I also expect renewed buying interest from banks and investors looking to replace mortgage assets that have "run off" their portfolios.  Finally, market levels for "implied volatilities" should continue to moderate, reflecting a more optimistic outlook for the financial system and lessened fears of "tape bombs," i.e., bad headlines.  (Implied volatilities are generated by using option pricing models where price is a known input.  Option buying is a way for investors to reflect (and hedge) expectations of increased market volatility; the buying pressure pushes option prices and, in turn, implied volatilities higher.  MBS valuations are strongly influenced by implied volatility levels.  Since mortgages are essentially short a prepayment option, an uptick in volatility makes the short option more valuable, reducing the value of MBS.)

 Therefore, I'm fairly optimistic that primary fixed mortgage rates can drop further.  There are, of course, a number of factors that could upset this analysis.  These could include slower Fed buying of assets; lack of investor buying of the product; and a spike in implied volatility resulting from renewed fears in the banking sector.  (Put differently, my crystal ball is no more accurate than anyone else's.)  Based on the factors I've outlined, however, I expect mortgage rates to trend lower over the next few quarters.  I expect zero-point conforming fixed rates to reach, and possibly breach, the 4.5% level by the end of the summer.