AIG 2015 Annual Report Download - page 208

ITEM 7 / CRITICAL ACCOUNTING ESTIMATES

208
Class of Business or Category and Actuarial Method Application of Actuarial Method
Loss Adjustment Expenses
We determine reserves for legal defense and cost
containment loss adjustment expenses for each class
of business by one or more actuarial or structural
driver methods. The methods generally include
development methods comparable to those described
for loss development methods. The development could
be based on either the paid loss adjustment expenses
or the ratio of paid loss adjustment expenses to paid
losses, or both. Other methods include the utilization of
expected ultimate ratios of paid loss expense to paid
losses, based on actual experience from prior accident
years or from similar classes of business. We also
provide reserves for the estimated future costs of our
claims department settling existing reserves.
We generally determine reserves for adjuster loss
adjustment expenses based on calendar year ratios of
adjuster expenses paid to losses paid for the particular
class of business. We generally determine reserves for
other unallocated loss adjustment expenses based on
the ratio of the calendar year expenses paid to overall
losses paid. This determination is generally done for all
classes of business combined, and uses claim
overhead costs as a percent of losses paid. We may
supplement our judgments with an analysis of loss and
legal expense mix change using predictive models that
explicitly represent such mix change and detailed
reviews with the claims department on the methods
used to allocate the costs of the claims initiatives to
new and in-force business and to different classes and
sub-classes of business. In 2015, we increased our
estimates of our claims department costs to reflect
higher expected future costs associated with new
technology and operating models.
Catastrophes and Severe Losses
In response to major catastrophes and severe losses,
special analyses by both Actuarial and Claims are
conducted to estimate the gross and net liability for
unpaid losses and loss adjustment expenses from
such events.
These analyses may include a combination of
approaches, including catastrophe modeling estimates,
ground-up claim analysis, loss evaluation reports from
on-site field adjusters, and market share estimates.
Loss emergence patterns of similar catastrophic
events or aggregated severe losses are used to help
determine the reasonableness of ultimate loss
projections and the resulting IBNR need.
Alternative Loss Cost Trend and Loss Development Factor Assumptions by Class of Business
Two of our most significant assumptions in loss-based methodologies are those for loss cost trends and loss development
factors. There is generally some potential for deviation from our selected best estimate assumptions related to loss cost trends
and loss development factor assumptions. The classes noted below are considered to have a higher sensitivity to changes in
these assumptions than other classes. For example, primary auto liability reserves were impacted by changes in the frequency
and severity of losses in recent years which may not be captured by sensitivities of the parameter modeled below.
The sensitivity analysis below addresses each major class of business for which there is a possibility of a material deviation
from our overall reserve position as a result of changes in these individual assumptions. The analysis uses a range of potential
alternative assumptions for each class based on historical experience. Actual reserve development may not be consistent with
either the original or the alternative loss trend or loss development factor assumptions, and other assumptions made in the
reserving process may materially affect reserve development for a particular class of business.
It should be noted that there are many other assumptions used in the loss reserving process, and the sensitivities shown
below are not intended to be indications of possible loss reserve ranges for the class but merely illustrations of the sensitivity
of the reserves to certain alternative assumptions considered individually. The actual reserve range is a function of many
factors interacting with each other.