AIG 2006 Annual Report Download - page 91

American International Group, Inc. and Subsidiaries
For the year-end 2005 loss reserve review, AIG’s actuaries assumption for excess workers compensation was increased from
responded to the continuing adverse development by further approximately 2.5 percent to 6 percent for the 2005 study.
increasing the loss development factor assumptions. The loss For the year-end 2006 loss reserve review, AIG claims staff
development factors applicable to 1997 and subsequent accident updated the claim-by-claim projection for each open claim for
years were increased by approximately 4 percent. In addition, accident years 1999 and prior. These updated claims projections
AIG’s actuaries began to give greater weight to loss development were utilized by the actuaries as a benchmark for loss develop-
methods for accident years 2002 and 2003, in order to more fully ment factors in the year-end 2006 study. AIG’s actuaries
respond to the recent loss experience. AIG’s claims staff also determined that no significant changes in the assumptions were
conducted a series of ground-up claim projections covering all required. Prior accident year development in 2006 was adverse by
open claims for this business through accident year 2004. AIG’s approximately $70 million, a relatively minor amount for this
actuaries benchmarked the loss reserve indications for all class.
accident years through 2004 to these claim projections.
For the year-end 2006 loss reserve review, AIG’s actuaries Overview of Loss Reserving Process
determined that no significant changes in the assumptions were The General Insurance loss reserves can generally be categorized
required. Prior accident year loss development in 2006 was into two distinct groups. One group is short-tail classes of
favorable by approximately $20 million, an insignificant amount for business consisting principally of property, personal lines and
these classes. AIG’s actuaries continued to benchmark the loss certain casualty classes. The other group is long-tail casualty
reserve indications to the ground up claim projections provided by classes of business which includes excess and umbrella liability,
AIG claims staff for this class of business. For the year-end 2006 D&O, professional liability, medical malpractice, workers compen-
loss reserve review, the ground up claim projections included all sation, general liability, products liability, and related classes.
accident years through 2005.
Loss reserves pertaining to D&O and related management Short-Tail Reserves
liability classes of business are included in the Other liability
claims made line of business, as presented in the table on For operations writing short-tail coverages, such as property
page 37. coverages, the process of recording quarterly loss reserves is
generally geared toward maintaining an appropriate reserve for the
Excess Workers Compensation: This class of business exper- outstanding exposure, rather than determining an expected loss
ienced significant adverse development in 2005, and a relatively ratio for current business. For example, the IBNR reserve required
minor amount of adverse development in 2006. The adverse for a class of property business might be expected to approximate
development in 2005 related to 2002 and prior accident years. 20 percent of the latest year’s earned premiums, and this level of
This adverse development resulted primarily from significant loss reserve would generally be maintained regardless of the loss ratio
cost increases, primarily attributable to rapidly increasing medical emerging in the current quarter. The 20 percent factor would be
inflation and advances in medical care, which increased the cost adjusted to reflect changes in rate levels, loss reporting patterns,
of covered medical care and extended the life span of severely known exposure to unreported losses, or other factors affecting
injured workers. The effect of these factors on excess workers the particular class of business.
compensation claims experience is leveraged, as frequency is
increased by the rising number of claims that reach the excess Long-Tail Reserves
layers.
In response to the significantly adverse loss development in Estimation of ultimate net losses and loss expenses (net losses)
2005, an additional study was conducted for the 2005 year-end for long-tail casualty classes of business is a complex process
actuarial reserve analysis for DBG pertaining to the selection of and depends on a number of factors, including the class and
loss development factors for this class of business. Claims for volume of business involved. Experience in the more recent
excess workers compensation exhibit an exceptionally long-tail of accident years of long-tail casualty classes of business shows
loss development, running for decades from the date the loss is limited statistical credibility in reported net losses because a
incurred. Thus, the adequacy of loss reserves for this class is relatively low proportion of net losses would be reported claims
sensitive to the estimated loss development factors, as such and expenses and an even smaller percentage would be net
factors may be applied to many years of loss experience. In order losses paid. Therefore, IBNR would constitute a relatively high
to better estimate the tail development for this class, AIG claims proportion of net losses.
staff conducted a claim-by-claim projection of the expected ultimate AIG’s carried net long-tail loss reserves are tested using loss
paid loss for each open claim for 1998 and prior accident years as trend factors that AIG considers appropriate for each class of
these are the primary years from which the tail factors are derived. business. A variety of actuarial methods and assumptions is
The objective of the study was to provide a benchmark against normally employed to estimate net losses for long-tail casualty
which loss development factors in the tail could be evaluated. The classes of businesses. These methods ordinarily involve the use
resulting loss development factors utilized by the actuaries in the of loss trend factors intended to reflect the annual growth in loss
year-end 2005 study reflected an increase of approximately 18 per- costs from one accident year to the next. For the majority of long-
cent from the factors used in the prior year study without the tail casualty classes of business, net loss trend factors approxi-
benefit of the claims benchmark. In addition, the loss cost trend mated five percent. Loss trend factors reflect many items
Form 10-K 2006 AIG 41