![]() |
|
||
|
If the prepayment model is setup properly, the next concern is whether to use or ignore the enhanced dataset (loan size, LTV, geography, etc.). Pools with low loan sizes, or ones that originated in NY, FL or TX are slower and have “longer” than average OADs. The valuation method is closely related to the use of the prepayment model. OAD is a valuation measure that is not generated by the prepayment model itself, unlike the average life or the equivalent CPR. Complaining that an OAS system delivers a dubious OAD assumes that the constant-OAS method must correctly quantify the risk. That is, if interest rates move up or down, OAS levels will remain unchanged. This does not happen in the real world, in part because the OAS levels themselves are associated with bearing prepayment model risk priced by the MBS market (Levin [2004a], Levin and Davidson [2005]). PrOAS, or simply OAS to a risk-neutral prepayment model, can measure both risk and reward better. At the 2006 annual AD&Co conference (Prepay Risk-Neutrality & Model-Free Value To implement AD&Co prepayment risk-neutrality you generally just need to change both the refinancing tuning and the turnover tuning. These can be found weekly on our market analysis WEB pages. Some of our OAS clients have adapted the prOAS views, but many still naively think the standard OAS method must do what they like to see – accurate prediction of price moves.
|
|
|
|
|||||||||||||||||