Prepayment Uncertainty

A measure of the sensitivity of a security’s price to a
change in the forecasted rate of future prepayments. This concept is primarily
applicable to mortgage-backed securities,1 where homeowner prepayments due to
refinancing incentives and other conditions are difficult to predict. To calculate
this measure, alternative sets of future cash flows for a security are generated by
adjusting the current prepayment forecast upward and downward by some percentage,
e.g. 10% (for mortgage-backed securities, prepayment rates may be
expressed using the ‘PSA’ convention, or as SMMs, single monthly mortality rates,
or in terms of CPR, conditional/constant prepayment rates). Holding all other
things constant (including the initial term structure of interest rates, volatility
inputs and the security’s OAS), two new prices are computed using the slower
and faster versions of the base case prepayment forecasts. The average percentage
change in price resulting from the alternative prepayment forecasts versus the
current price is the measure of prepayment uncertainty; the more variable a security’s
cash flows under the alternative prepay speeds versus the current forecast,
the greater its prepayment uncertainty and therefore its ‘model’ risk.
An alternative approach to deriving a prepayment uncertainty measure is to
evaluate the sensitivity of effective duration to a change in prepayment forecasts.
This approach provides additional information to the risk manager and may be
used in place of or to complement the ‘price sensitivity’ form of prepayment
uncertainty. Either method helps to focus awareness on the fact that while
mortgage-backed security valuations capture the impact of prepayment variations
under different interest rate scenarios (typically via some type of Monte Carlo
simulation), these valuations are subject to error because of the uncertainty of
any prepayment forecast.
We now discuss these risk measures in detail, beginning with spread duration.