Security Capital Structure
Macro Factors and
Principal & Interest
tions, and lending standards at financial
institutions; all under significant strain
because of current economic conditions.
There are many potential loss models
that can be used to estimate credit risk/
potential losses over the life of the security. We’ll limit ourselves to four that
are based on models familiar to casualty
actuaries: the paid-loss development factor and the Bornhuetter-Ferguson (BF)
method, along with using incurred losses
instead of paid losses for both methods.
Paid losses can be calculated from
the loan-level data provided by trustees
or servicers. A loss curve is a measure
of the proportion of losses that are paid
by loan age. The anticipated losses can
be calculated using the paid losses to
date and the percent of total losses that
would be paid by a certain age.
FIGURE 3: Illustrative Loan Characteristics
Similar calculations can be performed
for the incurred-loss method. The incurred method will need to incorporate
current delinquencies in the calculation
of a probability of default. Incurred-loss
curves can be used to estimate total or remaining losses. For example, incurred-loss
curves can be estimated by introducing an
acceleration to the paid loss curves.
However, the emergence pattern for
mortgage-credit losses can be severely
distorted by economic conditions, with
calendar-year effects influencing the emergence curve at various maturity points
for different vintage years. Thus, caution
should be exercised in utilizing both paid-and incurred-loss development methods.
BF-based methods, in contrast, lend themselves more readily to such projections.
The BF method requires an a priori
loss calculation for projecting total collateral losses. This can be derived in two
steps using loan-level detail and economic assumptions. The probability of default
is derived first, followed by its severity.
For example, a loan that has very little
documentation is more likely to default
than one that has full documentation.
The graph in Figure 3 depicts frequently mentioned loan characteristics, such as
consumer credit (FICO) scores, loan-to-value (LTV), loan amortization type (fixed
or adjustable rate), interest-only loan, occupancy, loan size, and property type.