 |
 |
Valuation Commentary - March '07
The Hunt for Duration
by Alex Levin
Whether practitioners use our analytics to delta-hedge, to report hedge effectiveness under FAS-133, or to compute hedge ratios when taking short TBA positions against a pipeline origination, they focus foremost on a single analytical outcome - Effective Duration (ASA option-adjusted duration, OAD). Most requests for analytical consulting and most questions concerning valuation results also revolve around the reasonableness of this measure.
Much like most of our clients, we pay thorough attention to the OAD derived by our models. We have our own elaborated in-house OAS system that is heavily employed in the AD&Co business regimen as well as some client environments. This OAS system is used to generate our weekly market analysis reports. Last, but not least, it helps in benchmarking the valuation work done by our diverse clientele. Let me postpone my arguments and jump right to the conclusion: AD&Co analytics, when used properly, allows modeling the OAD for various MBS instruments accurately and inline with the market. This “proper” usage involves not only bug-free settings, but also a deep understanding of the limitations and extensions of the OAS method.
What does the OAD depend on?
Obviously, the prepayment model is one of the key analytical ingredients driving the interest rate risk measurement. It is very important that the AD&Co prepayment model setup is correct. The cause of a suspicious OAD could be from an incorrect setup of this model. For instance, the model expects to get the gross coupon of the MBS in question along with the net current coupon market rate that is often quoted by Bloomberg (MTGEFNCL, etc.). This requirement has been commonly misunderstood and confused. Sending a wrong market index (say, gross instead of net) materially slows down prepay speeds and extends OADs. Improper mapping of instruments to loan types can and did cause trouble for some users.
