2007 Annual Report - AIG.com

American International Group, Inc. and Subsidiaries
Management’s Discussion and Analysis of
Financial Condition and Results of Operations Continued
containing small amounts of asbestos, companies in the distribu- For asbestos, these tests project the losses expected to be
tion process, and parties with remote, ill-defined involvement in reported over the next nineteen years, i.e., from 2008 through
asbestos (Tiers Three and Four). Through its commitment to 2026, based on the actual losses reported through 2007 and the
appropriate staffing, training, and management oversight of asbes- expected future loss emergence for these claims. Three scenarios
tos cases, AIG seeks to mitigate its exposure to these claims. were tested, with a series of assumptions ranging from more
To determine the appropriate loss reserve as of December 31, optimistic to more conservative. In the first scenario, all carried
2007 for its asbestos and environmental exposures, AIG per- asbestos case reserves are assumed to be within ten percent of
formed a series of top-down and ground-up reserve analyses. In their ultimate settlement value. The second scenario relies on an
order to ensure it had the most comprehensive analysis possible, actuarial projection of report year development for asbestos
AIG engaged a third-party actuary to assist in a review of these claims reported from 1993 to the present to estimate case
exposures, including ground-up estimates for asbestos reserves reserve adequacy as of year-end 2007. The third scenario relies
consistent with the 2005 and 2006 reviews as well as a top-down on an actuarial projection of report year claims for asbestos but
report year projection for environmental reserves. Prior to 2005, reflects claims reported from 1989 to the present to estimate
AIG’s reserve analyses for asbestos and environmental exposures case reserve adequacy as of year-end 2007. Based on the results
was focused around a report year projection of aggregate losses of the prior report years for each of the three scenarios described
for both asbestos and environmental reserves. Additional tests above, the report year approach then projects forward to the year
such as market share analyses were also performed. Ground-up 2026 the expected future report year losses, based on AIG’s
analyses take into account policyholder-specific and claim-specific estimate of reasonable loss trend assumptions. These calculainformation
that has been gathered over many years from a tions are performed on losses gross of reinsurance. The IBNR
variety of sources. Ground-up studies can thus more accurately (including a provision for development of reported claims) on a
assess the exposure to AIG’s layers of coverage for each
net basis is based on applying a factor reflecting the expected
policyholder, and hence for all policyholders in the aggregate, ratio of net losses to gross losses for future loss emergence.
provided a sufficient sample of the policyholders can be modeled For environmental claims, an analogous series of frein
this manner.
quency/severity tests are produced. Environmental claims from
In order to ensure its ground-up analysis was comprehensive, future report years, (i.e., IBNR) are projected out nine years, i.e.,
AIG staff produced the information required at policy and claim through the year 2016.
level detail for nearly 1,000 asbestos defendants. This repre-
At year-end 2007, AIG considered a number of factors and
sented over 95 percent of all accounts for which AIG had received recent experience in addition to the results of the respective topany
claim notice of any amount pertaining to asbestos exposure. down and ground-up analyses performed for asbestos and
AIG did not set any minimum thresholds, such as amount of case environmental reserves. AIG considered the significant uncertainty
reserve outstanding, or paid losses to date, that would have that remains as to AIG’s ultimate liability relating to asbestos and
served to reduce the sample size and hence the comprehensive- environmental claims. This uncertainty is due to several factors
ness of the ground-up analysis. The results of the ground-up including:
analysis for each significant account were examined by AIG’s ( The long latency period between asbestos exposure and
claims staff for reasonableness, for consistency with policy
disease manifestation and the resulting potential for involvecoverage
terms, and any claim settlement terms applicable.
ment of multiple policy periods for individual claims;
Adjustments were incorporated accordingly. The results from the ( The increase in the volume of claims by currently unimpaired
universe of modeled accounts, which as noted above reflects the plaintiffs;
vast majority of AIG’s known exposures, were then utilized to ( Claims filed under the non-aggregate premises or operations
estimate the ultimate losses from accounts or exposures that
section of general liability policies;
could not be modeled and to determine an appropriate provision ( The number of insureds seeking bankruptcy protection and the
for unreported claims.
effect of prepackaged bankruptcies;
AIG conducted a comprehensive analysis of reinsurance ( Diverging legal interpretations; and
recoverability to establish the appropriate asbestos and environ- ( With respect to environmental claims, the difficulty in estimatmental
reserve net of reinsurance. AIG determined the amount of ing the allocation of remediation cost among various parties.
reinsurance that would be ceded to insolvent reinsurers or to
After carefully considering the results of the ground-up analycommuted
reinsurance contracts for both reported claims and for sis, which AIG updates on an annual basis, as well as all of the
IBNR. These amounts were then deducted from the indicated above factors, including the recent report year experience, AIG
amount of reinsurance recoverable. The year-end 2007 analysis increased its gross asbestos reserves by $75 million, all of which
reflected an update to the comprehensive analysis of reinsurance was reinsured, resulting in no increase to net reserves. Additionrecoverability
that was first completed in 2005 and updated in ally, during 2007 a reduction in estimated reinsurance recover-
2006. All asbestos accounts for which there was a significant able, partially offset by several large favorable asbestos
change in estimated losses in the 2007 review were analyzed to settlements, resulted in a minor amount of adverse incurred loss
determine the appropriate reserve net of reinsurance.
development.
AIG also completed a top-down report year projection of its
Based on the environmental top-down report year analysis
indicated asbestos and environmental loss reserves. These performed in the fourth quarter of 2007, a minor increase in both
projections consist of a series of tests performed separately for gross and net reserves was recognized, resulting in the relatively
asbestos and for environmental exposures.
minor amount of development shown in the table below.
60 AIG 2007 Form 10-K