Quote from h hubbins:
prior appreciation had nothing to do with it. state unemployment rate did. caveat: state wide data. local data can be different.
Interesting blog. However, I think his conclusions are seriously flawed.
He observes two correlations: a negative correlation between appreciation and mortgage delinquency, and a positive correlation between unemployment and mortgage delinquency.
The thing is, these correlations should typically hold regardless of market conditions. People who are sitting on large price appreciations can sell or refinance their way into solvency. Whereas, people who lose their job often have to foreclose. These aren't so much correlations as they are tautologies.
There's another reason why low income areas are getting hit first. Since housing prices peaked in 2006, the only mortgages that are likely to be underwater and suffering payment shock are those with the shortest reset periods, like 2 year ARMs. These are almost all subprime. The prime and alt-A mortgages in more affluent areas are typically 5 year ARMs or IOs. Any of those mortgages that are underwater are still enjoying teaser rates. (You can find all this data and more in Ivy Zelman's report for Credit Suisse.)
Finally, there is the oft-repeated truism that housing distress starts at the bottom of the food chain and spreads to the top - the plankton theory. Someone who buys a mansion has to sell their old home to a move up buyer, who sells their old home to a first time homeowner. What is happening today is that the entry level homeowners are getting clobbered, but it's only a matter of time for excess inventory and price depreciation to spread up the housing food chain.
Martin