Lapse/persistency considerations in the time of COVID-19

By Amal Rajwani, ACIA, and Dan Doyle, FCIA, members of the CIA Committee on Life Insurance Financial Reporting

The experience related to the impact of COVID-19 on lapse/persistency of life and health insurance contracts is just starting to emerge and life insurers are monitoring this experience closely. Broadly speaking, there are offsetting expectations relating to policyholder behaviour in the short term: (i) increased lapsation/policy loans due to increased unemployment, and (ii) decreased lapsation as policyholders recognize the value of having life insurance coverage. 

Early anecdotal reports indicate that little change in lapse rates have been observed to date.  However, these reports are based on the first two months of the pandemic and, as such, may be influenced by the following underlying factors:

• Policyholders making greater use of existing grace periods.
• Insurers offering combinations of extended grace periods and longer re-payment periods. 

Actuaries might pay attention to the use of any extended grace period programs and emerging policyholder behaviour. For example, lapses could reduce during the extended grace period but could spike immediately after. The net result could simply be to move expected policy lapses to later durations.

As insurers consider the appropriateness of policy liabilities during the COVID-19 pandemic the following summarizes key guidance:  

• Section 2.1 of Use of Actuarial Judgement in Setting Assumptions and Margins for Adverse Deviations states that emerging trends in experience would be reflected in the best estimate assumptions while random fluctuations in recent experience would not.  This educational note also states that ‘actuaries would often consider keeping the <best estimate> assumptions unchanged until clear trends emerge.’
• Subsections 2250 and 2260 of the Standard of Practice direct actuaries towards higher margins for adverse deviations (MfADs) for ‘unusually high uncertainty.’

Actuaries might consider differentiating between changes in short-term lapse behaviour from long-term/permanent changes in lapse behaviour when assessing persistency assumptions: 

• The immediate COVID-19 impact to lapsation does not naturally fall into either the ‘random fluctuation’ or ‘emerging trend’ category. As such it is an open question whether changes to best estimate lapse assumptions in the short term would be considered. If changes to assumptions are considered, then the timing of such changes becomes the next open question given that the pandemic may run its course over the next two years.
• Whether COVID-19 ultimately leads to a permanent change in long-term lapse behaviour remains to be seen and it is too early yet to assess this. As such, it may be appropriate to maintain existing long-term best estimate lapse assumptions. 

Notwithstanding the above, actuaries may consider adjusting MfADs (permanent or temporary) or creating a temporary (i.e., bulk) liability that would be released as the environment returns to normalcy.

The following are some broad considerations related to the lapse/persistency assumptions used in the valuation:

• There is a possibility of a short to medium term increase in unemployment rates due to COVID-19. The last such period of increased unemployment was during the 2008 financial crisis and companies could consider evaluating their lapse/persistency experience during that period vs. the preceding and subsequent periods. However, considering the pandemic nature of COVID-19, insureds may want to maintain their insurance coverage as there is more immediate potential benefit compared to the circumstances in 2008. This may temper lapses. Additionally, the impact of economic conditions and increased unemployment rates may vary by policy size and by product type.
• Lapsation impacts related to COVID-19 may differ by age. Older insureds have heightened mortality risk and may opt to maintain coverage. Conversely younger insureds may, if under financial pressure, opt to cancel coverage as the COVID-19 impacts to younger ages has not been as large. A compounding effect is that any economic impacts of COVID-19 may also vary by age.
• For renewable term and other life insurance products – there may be additional anti-selection against insurers as more unhealthy insureds retain coverage.
• For universal life contracts – fund balances are affected by fund returns as well as policyholder deposits and withdrawals. Some impacts may have already been realized (e.g., poor fund returns) but other impacts may not yet have been experienced (e.g., projected premiums or withdrawals may be impacted by their economic circumstances).  The impact on policy behaviours might be reflected through the following:
  • Higher lapses which may not be observed until later periods through the fund exhaustion mechanism. The impact of fund exhaustion depends on the extent that policyholders expect fund values to be a source of funding for future cost of insurance charges. 
  • Fund transfers to “safer” investments options especially those with minimum guarantees. Actuaries would monitor exposures to these minimum interest guarantees.

The actuary is reminded of Section 4.3–4.5 of the Valuation of Universal Life Policy Liabilities educational note that discuss policy lapse, transfer of funds, expected premiums, and partial withdrawal assumptions.

• Policies that include cash surrender values (CSV) likely offer policy loans. Policy loans are often the first non-forfeiture option utilized in order to prevent policy lapsation. However, increased policy loans may lead to increased future lapses if the policy loan balance surpasses the CSV.
• For segregated fund contracts – the utilization of dynamic lapse functions is a common practice within the industry. These functions are usually comprised of a base lapse assumption and a dynamic (in the moneyness) assumption. These functions typically reduce lapse rates as the guarantees become more in-the-money. For these products, insurers would consider the appropriateness of these base lapse assumptions and related margins. The current environment poses a test to these dynamic lapse functions; however, the ‘dynamic’ portion can only be assessed relative to the appropriate base lapse assumption. The actuary is reminded of Section 7 of the Considerations in the Valuation of Segregated Fund Products educational note when considering changes to best estimate assumptions versus MfADs.
• Insurers offering group coverages might consider the impacts of the risk of significant block shrinkage due to plan cancellations or membership deterioration. Insurers would want to consider the impact of cancellations along with the following behaviours:
  • Anti-selection if groups with high drug costs are more likely to keep business in force (particularly smaller employers).
  • Drug stockpiling prior to cancellation (mitigated by provincial guidelines to limit the size of the drug refills).

Government programs will have a material impact on the above. In particular, the Canada Employment Wage Subsidy will enable employers to keep employees on payroll (with benefits) even where business activity is reduced.

Actuaries would consider the impact of COVID-19 on their policyholder experience including the impact on future policyholder experience studies. The actuary may want to disregard the 2020 experience if it is considered a short-term fluctuation in experience.

In conclusion, the environment is unique and uncertain. Actuarial guidance suggests that when uncertainty increases, more prudence would be included in the valuation – which may be reflected though increased short-term provisions for adverse deviations. We hope the considerations in this article are useful in the assessment of the appropriateness of both best estimate and margins for assumptions regarding policyholder behaviour.

CIA publications pertaining to the lapse/persistency assumption

Actuaries are encouraged to review the following excerpts of these CIA publications:

1. Paragraphs 2350.19 to 2350.31 that relate to insured life withdrawal and partial withdrawal, and anti-selective lapse within subsection 2350 of the Standards of Practice.
2. Section 5.3 (Persistency and Lapse Risk) of the Considerations for the Development of a pandemic scenario research paper.
3. Section 620 within the Expected Mortality: Fully underwritten Canadian individual life insurance policies educational note that discusses the factors that influence selective lapsation assumption.
4. Section 2.1 of the Use of Actuarial Judgement in Setting Assumptions and Margins for Adverse Deviations educational note offers several considerations for setting lapse assumptions in times where experience is volatile.
5. Sections 4.3 and 4.4 of Valuation of Universal Life Policy Liabilities educational note lists various factors that actuaries would consider in setting policyholder behaviour assumptions for universal life products.
6. Section 7 of the Considerations in the Valuation of Segregated Fund Products educational note lists various factors that actuaries would consider in setting policyholder behaviour assumptions for segregated fund products.