A remarkable interagency group of economists wrote a paper that investigated the NPV test that is at the heart of HAMP: for a borrower to get a loan modification, the value of the modification must be on net greater than zero (or in the case of Fannie-Freddie loans, greater than -$5000). In other words, the losses from expected default must be greater than the losses from modification.
An upshot of this rule is that borrowers who are deeply under water get no modifications--because the chance their loan will in the end cure is very small. This means that borrowers in places that need HAMP most--Las Vegas, Phoenix, Florida, etc.--are least likely to get it, all else being equal.
It also underscores a basic point: in places where values have fallen by more than, say, 50 percent, anything less than principal balance relief just delays the inevitable. When households in, say, Central California want to move or retire, their house sale will not be sufficient to cover their loan balance. Even moving to make short sales easier would help. Otherwise, it is hard to see how we can restart the market anytime soon.