2007 Annual Report - AIG.com

American International Group, Inc. and Subsidiaries
Management’s Discussion and Analysis of
Financial Condition and Results of Operations Continued
in 2006, and believes that its reserves as of December 31, 2007 required. Prior accident year development in 2007 was favorable
are adequate for its D&O and related management liability classes by approximately $15 million, an insignificant amount for this
of business.
class.
Loss reserves pertaining to D&O and related management
liability classes of business are included in the other liability Overview of Loss Reserving Process
claims made line of business, as presented in the table above.
The General Insurance loss reserves can generally be categorized
Excess Workers Compensation: This class of business exper- into two distinct groups. One group is short-tail classes of
ienced significant adverse development in 2005, a relatively minor business consisting principally of property, personal lines and
amount of adverse development in 2006, and a minor amount of certain casualty classes. The other group is long-tail casualty
favorable development in 2007. The adverse development in classes of business which includes excess and umbrella liability,
2005 related to 2002 and prior accident years. This adverse D&O, professional liability, medical malpractice, workers compendevelopment
resulted primarily from significant loss cost in-
sation, general liability, products liability, and related classes.
creases, primarily attributable to rapidly increasing medical inflation
and advances in medical care, which increased the cost of Short-Tail Reserves
covered medical care and extended the life span of severely
For operations writing short-tail coverages, such as property
injured workers. The effect of these factors on excess workers
coverages, the process of recording quarterly loss reserves is
compensation claims experience is leveraged, as frequency is
generally geared toward maintaining an appropriate reserve for the
increased by the rising number of claims that reach the excess
outstanding exposure, rather than determining an expected loss
layers.
ratio for current business. For example, the IBNR reserve required
In response to the significantly adverse loss development in
for a class of property business might be expected to approximate
2005, an additional study was conducted for the 2005 year-end
20 percent of the latest year’s earned premiums, and this level of
actuarial reserve analysis for DBG pertaining to the selection of
reserve would generally be maintained regardless of the loss ratio
loss development factors for this class of business. Claims for
emerging in the current quarter. The 20 percent factor would be
excess workers compensation exhibit an exceptionally long-tail of
adjusted to reflect changes in rate levels, loss reporting patterns,
loss development, running for decades from the date the loss is
known exposure to unreported losses, or other factors affecting
incurred. Thus, the adequacy of loss reserves for this class is
the particular class of business.
sensitive to the estimated loss development factors, as such
factors may be applied to many years of loss experience. In order
to better estimate the tail development for this class, AIG claims
Long-Tail Reserves
staff conducted a claim-by-claim projection of the expected
Estimation of ultimate net losses and loss expenses (net losses)
ultimate paid loss for each open claim for 1998 and prior
for long-tail casualty classes of business is a complex process
accident years as these are the primary years from which the tail and depends on a number of factors, including the class and
factors are derived. The objective of the study was to provide a volume of business involved. Experience in the more recent
benchmark against which loss development factors in the tail accident years of long-tail casualty classes of business shows
could be evaluated. The resulting loss development factors utilized limited statistical credibility in reported net losses because a
by the actuaries in the year-end 2005 study reflected an increase relatively low proportion of net losses would be reported claims
of approximately 18 percent from the factors used in the prior and expenses and an even smaller percentage would be net
year study without the benefit of the claims benchmark. In
losses paid. Therefore, IBNR would constitute a relatively high
addition, the loss cost trend assumption for excess workers proportion of net losses.
compensation was increased from approximately 2.5 percent to AIG’s carried net long-tail loss reserves are tested using loss
6 percent for the 2005 study. trend factors that AIG considers appropriate for each class of
For the year-end 2006 loss reserve review, AIG claims staff business. A variety of actuarial methods and assumptions is
updated the claim-by-claim projection for each open claim for normally employed to estimate net losses for long-tail casualty
accident years 1999 and prior. These updated claims projections classes of businesses. These methods ordinarily involve the use
were utilized by the actuaries as a benchmark for loss develop- of loss trend factors intended to reflect the annual growth in loss
ment factors in the year-end 2006 study. AIG’s actuaries
costs from one accident year to the next. For the majority of longdetermined
that no significant changes in the assumptions were tail casualty classes of business, net loss trend factors approxirequired.
Prior accident year development in 2006 was adverse by mated five percent. Loss trend factors reflect many items
approximately $70 million, a relatively minor amount for this including changes in claims handling, exposure and policy forms,
class.
current and future estimates of monetary inflation and social
For the year-end 2007 loss reserve review, AIG claims staff inflation and increases in litigation and awards. These factors are
again updated the claim-by-claim projection for each open claim periodically reviewed and adjusted, as appropriate, to reflect
for accident years 2000 and prior. These updated claims
emerging trends which are based upon past loss experience.
projections were utilized by the actuaries as a benchmark for loss Thus, many factors are implicitly considered in estimating the year
development factors in the year-end 2007 study. AIG’s actuaries to year growth in loss costs.
determined that no significant changes in the assumptions were
52 AIG 2007 Form 10-K