The burst of optimism in response to the Case-Shiller numbers stems from this chart, which shows that home price increases are starting to appear in the April 2009 index data, released June 30:

Eight metropolitan areas showed home price index increases, including even San Francisco where the home price decline to date has been 45.75%.  These one month changes have to be taken with a grain of salt because there is a strong seasonal component to the indices, there is significant measurement error, and there are frequent revisions.  Nonetheless, this is the first time we have seen a significant number of home price increases in a long time.

Even the one month percentage changes, however, show something very significant for the way we should be thinking about macro factor driven default and prepayment risk.  The differences in monthly price changes are enormous, ranging from a 4.05% monthly decline in Las Vegas to the 2.01% increase reported in Dallas.  Not too long ago, it was a rare financial institution that was measuring and publicly reporting on its macro factor sensitivity to home price risk.  In fact, in prior quotes on www.kamakuraco.com, senior executives at Merrill Lynch, UBS, and Citigroup went on record to say that they did not in fact have limits on home price risk in place prior to the crisis.  The first step that many banks have taken in response to the crisis is to look at home prices as if all home prices are driven by a common factor, very consistent with the U.S. bank regulatory agencies’ mandated stress tests of February-April 2009.  The diversity of changes in April shows that much more granularity is necessary to meet a high standard of risk management.

This phenomenon is the home price equivalent of a common explanation used in silo interest rate risk management 12 or 13 years ago.  In arguing why transaction level risk analysis is essential, interest rate risk managers would point out that a loan portfolio with interest rate caps on half the portfolio at 6% and 8% on the other half does not behave in the same way as a portfolio where all of the caps are at 7%.  The same point is equally true with respect to macro factors like home prices.  The chart below shows the huge differences in home price declines relative to all time highs by metropolitan area:

Assuming all home prices move together is a vast improvement over the traditional assumption that they don’t move at all.  That being said, the assumption that there is a parallel shift in home prices across the country is massively wrong.  The 9.55% decline in home prices in Dallas is light years away from the 54.07% decline that Phoenix has experienced.

A rich multi-factor approach to measuring and monitoring home price risk is an essential component in capitalizing on the expensive lessons to be learned from the current crisis.

Donald R. van Deventer

Kamakura Corporation

Honolulu, July 1, 2009