Voya 2014 Annual Report Download - page 64

present. We analyze actual experience over that entire period, as we believe that over the duration of the variable
annuity policies we may experience the full range of policyholder behavior and market conditions. However,
Management’s current best estimate of variable annuity policyholder lapse behavior is weighted more heavily
toward more recent experience, as the last three years of data have shown a more consistent trend of lapse
behavior.
We review overall policyholder experience at least annually (including lapse, annuitization, withdrawal and
mortality), and update these assumptions when deemed necessary based on additional information that becomes
available. As customer experience continues to materialize, we may adjust our assumptions. We increased
reserves in the fourth quarter of 2011 after a comprehensive review of our assumptions relating to lapses,
mortality, annuitization of income benefits and utilization of withdrawal benefits. The review in 2011 included
an analysis of a larger body of actual experience than was previously available, including a longer period with
low equity markets and interest rates, which we believe provided greater insight into anticipated policyholder
behavior for contracts that are in the money. This resulted in an increase of U.S. GAAP reserves of $741 million
and gross U.S. statutory reserves of $2,776 million in the fourth quarter of 2011.
During the third quarters of 2014 and 2013, we completed our annual review of assumptions, including
projection model inputs. Annual assumption changes and revisions to projection model inputs implemented
during 2014 resulted in a gain of $102.3 million (excluding a gain of $37.9 million due to changes in the
technique used to estimate nonperformance risk). This $102.3 million gain included a favorable $170.2 million
resulting from policyholder behavior assumption changes partially offset by an unfavorable $40.5 million
resulting from changes to mortality assumptions. The gain from policyholder behavior assumption changes was
primarily due to an update to the utilization assumption on GMWBL contracts, partially offset by an unfavorable
result from an update to lapse assumptions. The 2013 result included a loss of $185.3 million (excluding a gain
of $144.6 million due to changes in the technique used to estimate nonperformance risk) due to annual
assumption changes. This $185.3 million loss included an unfavorable $117.9 million resulting from changes to
mortality assumptions and unfavorable $85.5 million resulting from policyholder behavior assumption changes.
As discussed above, our recent changes in lapse assumptions moved our assumptions to be in line with lapse
experience between mid-2011 to the present. Also as described above, future reserve increases in connection with
experience updates could be material and adverse to the results of operations or financial condition of the
Company.
We will continue to monitor the emergence of experience. If adjustments to policyholder behavior
assumptions (e.g., lapse, annuitization and withdrawal) are necessary, which is ordinary course for interest-
sensitive long-dated liabilities, we anticipate that the financial impact of such a change (either under U.S. GAAP
or due to increases or decreases in gross U.S. statutory reserves) will likely be in a range, either up or down, that
is generally consistent with the impact experienced in the past two years.
Other Risks. Despite the closure of new product sales, some new policy amounts continue to be deposited as
additional premium to existing contracts.Benefit designs do limit the attractiveness of additional premium, but in
some cases these additional premiums may increase the guarantee available to the policyholder.The volume of
additional premiums has diminished since we ceased new product sales in 2010.
On June 2, 2014 we entered into an agreement to outsource the actuarial valuation, modeling and hedging
functions of our CBVA segment to Milliman, Inc. (“Milliman”). Under this agreement, Milliman will perform
the calculation of financial reporting and risk metrics, along with the analytics used to determine hedge positions.
We will continue to oversee and manage the CBVA segment and retain full accountability for assumptions and
methodologies, as well as the setting of the hedge objectives and the execution of hedge positions. This
agreement will allow us to create a more variable cost structure for the CBVA segment.
41