Value Commentary

Inquiries & Interactions: We're Always Listening
By Alex Levin

AD&Co. often receives questions about particular modeling aspects, parameter tunings, or recommended settings. We appreciate the dialogue with clients, as it provides us with valuable user insight, which helps us provide enhanced products and services. This article focuses on some commonly asked questions concerning our valuation system and results.

The Role of the Slope - Clarified
After writing "Be Wary of the Slope" for the September 2003 Pipeline, I received some feedback. In that article, I stated that the 10-to-2 spread has never been negative over the last 10 years. One reader pointed out that "…the yield curve was negative from February 2000 to December 2000 based on the 10 year and 2 year constant maturity Treasuries." We are both right! I referred to the swap not Treasury curve, though explicitly said so only at the end…

Another reader questioned my claim that "For many valuation exercises with fixed-rate mortgages, a two-factor model will not add or subtract considerable value extracted by a one-factor model - once volatility term structures are made identical (taken from the market) for both models." He correctly pointed to the AD&Co. modeling assumption that forecasts mortgage rates based on a linear combination of the 2-yr and the 10-yr rates. Under such a model, correlation between the 2-yr rate and the 10-yr rate will indeed affect the resultant volatility of the mortgage rate, and, consequently, the value of prepayment option.

Since the 30-yr fixed-rate mortgage current coupon is largely driven by the 10-yr rate, this effect is small. When considering "shorter" instruments (15-yr, balloons, hybrids), inter-rate correlation will matter, as I indicated in the article using ARMs as a counter-example.

Monte-Carlo: How Many Paths to Run?
Many have been asking about the recommended number of paths to specify when using AD&Co. Monte-Carlo and Quasi-Monte-Carlo. It may be worth it to remind readers that the theoretical accuracy of Monte-Carlo is inversely related to the square root of paths. It is difficult to come up with a universal >>>