|
Mortgage Prepayment Model, Sensitivity Analysis and Model Risk
Bill Burns, Ph.D. and Ying Shen, Ph.D.
Over the past few decades, prepayment modelers have been continuously updating
their models to reflect a wide range of prepayment experiences. Frustratingly,
few of these prepayment models have accurately predicted the market when a
market crisis or other unforeseen event has occurred. Given the complex behavior
of home borrowers who are affected by a large set of nonlinearly correlated
economic and non-economic factors, it is not obvious that there is a perfect
model that can adequately forecast prepayments for the next thirty years. Aided
by decades of extensive research and a wealth of prepayment data on the impact
of important underlying variables, the mortgage community has finally reached
the conclusion that the quest for the perfect model is simply a delusion.
|
|
|
|