There’s been a lot of criticism over the last year or so about the ability of algorithmic trading strategies to accurately predict market movements. But that isn’t stopping Treasury Secretary Tim Geithner from leaning toward using a computer program called “Mo Mod” to tell him how to use almost half of the remaining $350 million TARP fund to modify mortgages.
The New York Post describes Mo Mod:
“Mo Mod” is an algorithmic mortgage processing program that can rewrite up to 500,000 loans a month, and will be a major part of Treasury’s plan to help repair tattered bank balance sheets.
The 21-day “Mo Mod” program works by structuring a new mortgage that more accurately reflects a home’s worth so that a troubled borrower no longer owes more on their home than the property is worth.
The process then enables a lender to pool these new mortgages together into securities that reflect more accurately a home’s value, which makes them less risky for investors.
Let’s put aside the criticisms of mortgage modification in general. This idea is intriguing because it avoids many of the modification through bankruptcy problems. But can a computer program accurately anticipate the way people will pay back their mortgages or the direction of home prices? So far, many, many people have been crushed based on predictions that the housing market will normalize, revert to the mean, stop being “dislocated.” Will a computer do better?
Also, how will we stop this thing from becoming SkyNet?