From an analysis by J.P. Morgan:

We expect that home prices, as measured by the Case-Shiller 10-city composite index, will decline as much as 15-20% from the June 2006 peak before this cycle hits bottom, potentially two or three years from now. As of April 2007, the index was down 3% from the peak, suggesting there is another 12-15% decline to go. While a total price decline of 15-20% sounds large, many markets saw prices double between 2000 and 2006, making a retreat in prices less extraordinary. Moreover, subprime heavy areas are expected to perform worse than higher end housing markets.

Back in the days of 20 percent down payments, a 15-20 percent decline of home prices would still leave the borrower with positive equity. When borrowers have positive equity, defaults are rare–you’re better off selling the house than letting the lender have the house.

But 20 percent down payments are so-o-o-o last century.

The pointer is from David Leonhardt, who also points out that adjustable-rate mortgages are about to reset (i.e, have their monthly payments jump) in large numbers.

In all, the interest rates on about $1 trillion worth of mortgages, or 12 percent of the nation’s total, will reset for the first time this year or next. A couple of years ago, by comparison, only a marginal amount of mortgage debt — a few billion dollars — was resetting each month.

This Leonhardt paragraph puzzles me:

From 1994 to 2005, some 3.2 million households were able to buy homes thanks to subprime mortgages or other such loans, according to an analysis by Moody’s Economy.com. About 1.7 million of them will probably lose their homes to foreclosure when all is said and done.

I would assume that most of the borrowers who purchased between 1994 and 2004 are safe, because of subsequent price appreciation. Obviously, someone who bought in 1998 and refinanced in 2006, taking out all their equity, might be in trouble. But that should not be huge numbers of people.

In fact, according to the Office of Housing Enterprise Oversight, house prices are still going up. (Their index is based on the same calculation techniques as Case-Shiller, but uses a different sample of homes.) In the first quarter of 2007, their index was 4.25 percent above the same quarter a year ago, which was 12.61 percent above Q1 of 2005. Homeowners who bought in the first half of 2005 or earlier should have built up a pretty fair equity cushion.

Given all this, I wonder how many borrowers out there do not have positive equity in their homes. Until somebody does that calculation, I am not sure whether to predict a trickle of mortgage defaults or a flood.