3 Critical accounting estimates and judgment in applying accounting policies

Application of the accounting policies in the preparation of the financial statements requires management to apply judgment involving assumptions and estimates concerning future results or other developments, including the likelihood, timing or amount of future transactions or events. Those estimates are inherently subject to change and actual results could differ from those estimates. Included among the material (or potentially material) reported amounts and disclosures that require extensive use of estimates are: fair value of certain invested assets and derivatives, deferred policy acquisition costs (please refer to paragraph 2.13), value of business acquired and other purchased intangible assets (please refer to paragraph 2.6), goodwill (please refer to paragraph 2.6), policyholder claims and benefits (please refer to paragraph 2.29), insurance guarantees (please refer to paragraph 2.19), pension plans (please refer to paragraph 2.22), income taxes (please refer to paragraph 2.24) and the potential effects of resolving litigation matters (please refer to paragraph 2.25). Accounting policies that are critical to the financial statement presentation and that require complex estimates or significant judgment are described in the following sections.

Valuation of assets and liabilities arising from life insurance contracts

The liability for life insurance contracts with guaranteed or fixed account terms is either based on current assumptions, on the assumptions established at inception of the contract, reflecting the best estimates at the time increased with a margin for adverse deviation or on the valuation assumptions (historical cost), without risk margin. All contracts are subject to liability adequacy testing which reflects management’s current estimates of future cash flows (including investment returns). To the extent that the liability is based on current assumptions, a change in assumptions will have an immediate impact on the income statement. Also, if a change in assumption results in not passing the liability adequacy test, the entire deficiency is recognized in the income statement.

To the extent that the deficiency relates to unrealized gains and losses on available-for-sale investments, the additional liability is recognized in the related revaluation reserve in shareholders’ equity.

Some insurance contracts without a guaranteed or fixed contractual term contain guaranteed minimum benefits. Depending on the nature of the guarantee, it may either be bifurcated and presented as a derivative or be reflected in the value of the insurance liability in accordance with local accounting principles. Given the dynamic and complex nature of these guarantees, stochastic techniques under a variety of market return scenarios are often used for measurement purposes. Such models require management to make numerous estimates based on historical experience and market expectations. Changes in these estimates will immediately affect the income statement.

In addition, certain acquisition costs related to the sale of new policies and the purchase of policies already in force are recorded as DPAC and VOBA assets respectively, and are amortized to the income statement over time. If the assumptions relating to the future profitability of these policies are not realized, the amortization of these costs could be accelerated and may require write-offs due to unrecoverability.

Actuarial and economic assumptions

The main assumptions used in measuring DPAC, VOBA and the liabilities for life insurance contracts with fixed or guaranteed terms relate to mortality, morbidity, investment return and future expenses. Depending on local accounting principles, surrender, lapse, and utilization rates may be considered.

Mortality tables applied are generally developed based on a blend of company experience and industry wide studies, taking into consideration product characteristics, own risk selection criteria, target market and past experience. Mortality experience is monitored through regular studies, the results of which are fed into the pricing cycle for new products and reflected in the liability calculation when appropriate. For contracts insuring survivorship or mortality, allowance may be made for further longevity or mortality improvements. Morbidity assumptions are based on own claims severity and frequency experience, adjusted where appropriate for industry information.

Investment assumptions are prescribed by the local regulator, market observable or based on management’s future expectations. In the latter case, the anticipated future investment returns are set by management on a countrywide basis, considering available market information and economic indicators. A significant assumption related to estimated gross profits on variable annuities and variable life insurance products in the United States and some of the smaller countries, is the annual long-term growth rate of the underlying assets. The reconsideration of this assumption may affect the original DPAC or VOBA amortization schedule, referred to as DPAC or VOBA unlocking. The difference between the original DPAC or VOBA amortization schedule and the revised schedule, which is based on actual and estimates of future gross profits, is recognized in the income statement as an expense or a benefit in the period of determination.

Assumptions on future expenses are based on the current level of expenses, adjusted for expected expense inflation if appropriate. In Aegon the Netherlands, the expense basis makes an allowance for planned future cost savings, which are included in the liability adequacy test.

Surrender and lapse rates depend on product features, policy duration and external circumstances such as the interest rate environment and competitor behavior. For policies with account value guarantees based on equity market movements, a dynamic lapse assumption is utilized to reflect policyholder behavior based on whether the guarantee is in the money. Own experience, as well as industry published data, are used in establishing assumptions. Lapse experience is correlated to mortality and morbidity levels, as higher or lower levels of surrenders may indicate future claims will be higher or lower than anticipated. Such correlations are accounted for in the mortality and morbidity assumptions based on the emerging analysis of experience.

Actuarial assumption and model updates

Assumptions are reviewed periodically, historically in the third quarter, but now in the second quarter for the US and Asia and in the fourth quarter for Europe, based on historical experience, observable market data, including market transactions such as acquisitions and reinsurance transactions, anticipated trends and legislative changes. Similarly, the models and systems used for determining our liabilities are reviewed periodically and, if deemed necessary, updated based on emerging best practices and available technology.

During 2018, Aegon implemented actuarial assumption and model updates resulting in a net EUR 121 million charge to income before tax (2017: EUR 276 million charge). Refer to note 5 Segment information for further details.

For the years 2016 through 2018, Aegon kept its long-term equity market return assumption for the estimated gross profits on variable life and variable annuity products in the Americas at 8%. During the three year period, the long-term assumption for 10-year US Treasury yields was 4.25% and the uniform grading period was 10 years. Aegon’s assumed returns for US separate account bond funds are 4% over the next 10 years and 6% thereafter. The long term credit spread assumption, net of assumed defaults and expenses, on our most common corporate bonds is 114bps. The 90-day Treasury yield was 2.37%, 1.39%, and 0.51% at December 31, 2018, 2017 and 2016, respectively. During the period 2016 to 2018 the 90-day Treasury yield was assumed to have a uniform grading over 10 years to 2.5%. On a quarterly basis, the estimated gross profits are updated for the difference between the estimated market return and the actual market return.

Sensitivity on variable annuities and variable life insurance products in the United States

A 1% decrease in the expected long-term equity growth rate with regard to Aegon’s variable annuities and variable life insurance products in the United States would result in a decrease in DPAC and VOBA balances and reserve strengthening of approximately EUR 148 million (2017: EUR 130 million). The DPAC and VOBA balances for these products in the United States amounted to EUR 2.6 billion at December 31, 2018 (2017: EUR 2.7 billion).

A relative increase of 10% to the mortality assumption was used, dependent on product and characteristics of the block of business, would reduce net income by approximately EUR 207 million (2017: EUR 265 million). A relative 20% increase in the lapse rate assumption would increase net income by approximately EUR 89 million (2017: EUR 78 million).

Any reasonably possible changes in the other assumptions Aegon uses to determine EGP margins (i.e. maintenance expenses) would reduce net income by less than EUR 9 million (2017: EUR 15 million).

Sensitivity on long term care products (LTC) in the United States

Sensitivities of significant product liability assumptions on the LTC IFRS after-tax Gross Present Value Reserve (GPV) margin are indicated below. The GPV is the liability as determined on a best estimate assumption basis. Until an assumption change is of significant amount to breach the current margin of EUR 35 million (loss recognition block of LTC), there will be no IFRS financial impact (2017: EUR 17 million). The GPV margin is the amount by which the IFRS reserve exceeds the GPV liability.

A 5% increase in the incidence rates with regard to Aegon’s long term care products would result in a GPV increase of approximately EUR 170 million. A 5% decrease in the incidence rates with regard to Aegon’s long term care products would result in a GPV decrease of approximately EUR 170 million.

Removing the morbidity improvement, which is a component of the incidence assumption, would result in a GPV increase of approximately EUR 500 million, of which EUR 425 million relates to the loss recognition block.

Reducing mortality 10% would result in a GPV increase of approximately EUR 85 million. Increasing mortality 10% would result in a GPV decrease of approximately EUR 85 million.

Removing future mortality improvement would result in a GPV decrease of approximately EUR 85 million.

Determination of fair value and fair value hierarchy

The following is a description of Aegon’s methods of determining fair value, and a quantification of its exposure to assets and liabilities measured at fair value.

Fair value is defined as the amount that would be received from the sale of an asset or paid to transfer a liability in an orderly transaction between market participants at the measurement date under current market conditions (i.e. an exit price at the measurement date from the perspective of a market participant that holds the asset or owes the liability). A fair value measurement assumes that the transaction to sell the asset or transfer the liability takes place either:

◆

In the principal market for the asset or liability; or

◆

In the absence of a principal market, in the most advantageous market for the asset or liability.

Aegon uses the following hierarchy for measuring and disclosing of the fair value of assets and liabilities:

◆

Level I: quoted prices (unadjusted) in active markets for identical assets or liabilities that Aegon can access at the measurement date;

◆

Level II: inputs other than quoted prices included within Level I that are observable for the asset or liability, either directly (that is, as prices) or indirectly (that is, derived from prices of identical or similar assets and liabilities) using valuation techniques for which all significant inputs are based on observable market data; and

◆

Level III: inputs for the asset or liability that are not based on observable market data (that is, unobservable inputs) using valuation techniques for which any significant input is not based on observable market data.

The best evidence of fair value is a quoted price in an actively traded market. In the event that the market for a financial instrument is not active or quoted market prices are not available, a valuation technique is used.

The degree of judgment used in measuring the fair value of assets and liabilities generally inversely correlates with the level of observable valuation inputs. Aegon maximizes the use of observable inputs and minimizes the use of unobservable valuation inputs when measuring fair value. Financial instruments, for example, with quoted prices in active markets generally have more pricing observability and therefore less judgment is used in measuring fair value. Conversely, financial instruments for which no quoted prices are available have less observability and are measured at fair value using valuation models or other pricing techniques that require more judgment.

The assets and liabilities categorization within the fair value hierarchy is based on the lowest input that is significant to the fair value measurement.

An active market is one in which transactions for the asset or liability take place with sufficient frequency and volume to provide pricing information on an ongoing basis. The judgment as to whether a market is active may include, although not necessarily determinative, lower transaction volumes, reduced transaction sizes and, in some cases, no observable trading activity for short periods. In inactive markets, assurance is obtained that the transaction price provides evidence of fair value or it is determined that adjustments to transaction prices are necessary to measure the fair value of the instrument.

The majority of valuation techniques employ only observable market data, and so the reliability of the fair value measurement is high. However, certain assets and liabilities are valued on the basis of valuation techniques that feature one or more significant market inputs that are unobservable and, for such assets and liabilities, the derivation of fair value is more judgmental. An instrument is classified in its entirety as valued using significant unobservable inputs (Level III) if, in the opinion of management, a significant proportion of the instrument’s carrying amount is driven by unobservable inputs. ‘Unobservable’ in this context means that there is little or no current market data available from which to determine the price at which an at arm’s length transaction would be likely to occur. It generally does not mean that there is no market data available at all upon which to base a determination of fair value. Additional information is provided in the table headed ‘Effect of changes in significant unobservable assumptions to reasonably possible alternatives’ in note 47 Fair Value. While Aegon believes its valuation techniques are appropriate and consistent with other market participants, the use of different methodologies or assumptions to determine the fair value of certain instruments (both financial and non-financial) could result in a different estimate of fair value at the reporting date.

To operationalize Aegon’s fair value hierarchy, individual instruments (both financial and non-financial) are assigned a fair value level based primarily on the type of instrument and the source of the prices (e.g. index, third-party pricing service, broker, internally modeled). Periodically, this logic for assigning fair value levels is reviewed to determine if any modifications are necessary in the context of the current market environment.