As refis die, turnover is the new story in MBS. Researchers are trying to pin down the factors driving speeds, particularly in discount prepayments.
In its outlook for turnover rates, Lehman Brothers argues that the next real backup in mortgage rates will probably produce discount prepayment speeds well below those seen in the late 1990s.
There has been no prevailing consistency in prepayments during the major discount episodes over the past decade. Analysts said that fully seasoned turnover in the most recent rate backup in 1999 to 2000 averaged 11% to 12% CPR, which is about 4% CPR higher than what was seen in the 1994-1995 and 1996-1997 backups.
"It is important to know which, if either, of the past prepayment experiences will materialize during the next backup in mortgage rates," said analysts.
Lehman said the best way to know the answer to this is by understanding the primary drivers of prepayment differences across the different episodes and, in turn, use that information to predict future speeds. Lehman stated that based on their analysis, the differences in prepayments in these episodes were the result of variations in the economic scenarios at work and the underlying collateral during those times.
Specifically, analysts have identified the slope of the curve at origination and home price appreciation as the two factors that have the most impact on turnover rates.
Lehman said that equity accumulated through home price appreciation allows borrowers to trade up to bigger houses as well as to monetize their home-equity gains by trading down or cashing out. However, Lehman said that home price appreciation by itself could not explain the overall rise in turnover between the mid-1990s and late 1990s. Differences in borrower horizons, which were caused by the difference in the yield curve's slope at origination, also drove the increase. "Intuitively, borrowers who choose 30-year mortgages during steep yield curve environments tend to have longer horizons than borrowers who choose 30-year mortgages during flat yield curve environments," wrote researchers.
By studying these two factors, Lehman predicted turnover on the biggest outstanding 5% and 5.5% cohorts to be less than that seen in the 1999-2000 bout of heavy refinancing activity. These cohorts were originated in 2002 to 2003 under a steep yield curve and an average 30-year and 5/1 spread of above 100 basis points. Aside from this situation, the housing sector is beginning to exhibit signs of relative weakening. So even with a higher percentage of those refinancing in the largest outstanding cohorts, Lehman predicts seasoned discount prepayments on these to be 2% to 3% CPR slower compared with those in the 1999-2000 backup.
However, Lehman said that one factor that may distort the picture in the near term is the curve continuing to be steep. As long as the Fed remains on hold, even if 30-year mortgage rates rise moderately, significantly lower hybrid rates (e.g. 125bp lower for a 5/1 hybrid versus 30-year fixed rate mortgage) will act to buoy housing prices and mitigate lock-in, potentially adding 1% to 2% CPR to fully seasoned turnover rates.
Over the longer term, once the Fed starts to tighten, slower housing price growth and less attractive hybrid opportunities should push prepayment speeds lower.