Jim Hamilton points to new research from Dastrup & Carson indicating that one of the strongest predictors of decline in housing values during the crash is simply the extent of the increase during the boom:
All told, extent of price increases explains over half the variance in the extent of declines. Detroit is a clear exception to this trend, crashing much harder than it ever rose. Something notably about this is that a great many metro areas still haven’t seen declines equal in magnitude to the boom. Will prices continue to decline or stagnate until everything’s back to zero?