AIG 2015 Annual Report Download - page 201

ITEM 7 / CRITICAL ACCOUNTING ESTIMATES

201
Class of Business or Category and Actuarial Method Application of Actuarial Method
Many of our primary casualty policies contain risk-
sharing features, including high deductibles, self-insured
retentions or retrospective rating features, in addition to
a traditional insurance component. These risk-sharing
programs generally are large and complex, comprising
multiple products, years and structures, and are subject
to amendment over time. We generally use premium
and expected loss ratio methods for these programs.
In 2014, we updated our tail factor selections in CA and
NY, our loss development factors in the Construction
class, and refined our segmentation for excess of a large
deductible business. Each of these segments appeared
to have been impacted by specific structural changes in
the portfolio. For CA business, our tail factor increases
were in response to changing long-term medical
development patterns. In NY, there has been a
lengthening of the period between the date of accident
and the classification of non-scheduled permanent
partial injuries. We completed a review of claim
emergence and payouts for our top six states in workers’
compensation and concluded that NY and CA were the
main states where the loss development patterns had
materially changed since our last review. For the
Construction class, we noted that the construction sector
has experienced a comparatively slow recovery in
payroll levels. As a result of the diminished employment
opportunities in this industry sector, injured workers may
experience limited return-to-work opportunities, which
moderate the shortening of claim duration that normally
accompanies a labor market recovery. Lastly, for excess
of large deductible business, we updated our analyses
to consider the impact of changes in the mix of
retentions that has occurred over time as the data by
retention band was becoming more credible.
Commencing in 2014, we also enhanced our analysis by
considering our best estimate expectations of inflation
(principally, the PCE Deflator for Health Care Services)
and loss cost trends and we also reflected the impacts of
enhancements in our claim management and loss
mitigation activities, such as opioid drug management,
fraud investigation and medical management.
As part of the 2015 reserve review related to these
policies, we enhanced our segmentation to better reflect
varying policy features. As a result of the improved loss
segmentation, we were able to give more weight to loss
development methods in addition to premium and
expected loss ratio methods.