Longevity Risk Framework

The risk management of longevity is still in its infancy and there are a number of aspects that could improve this:

  • creation of an industry-standard definition of all the risks behaviours underlying longevity risk
  • development of a method to benchmark the output of different risk models used by individual companies using this standard definition
  • identification of new ways to reduce risk using this standard definition to identify the intrinsic and avoidable risks
  • consideration of the differing longevity risk exposures of different kinds of beneficiaries.

The MRSC has published an IFoA Longevity Risk Framework that is intended to address the first bullet.

Summary of Framework

The framework has been designed to achieve the following qualities:

  1. The taxonomy should be applicable to both “economic” and “one-year” assessments of risk.
  2. It should differentiate between the systemic population risk (to which all providers are exposed, albeit to different extents) and the specific portfolio risk (to which individual providers may have materially different exposures).
  3. It should be applicable to all forms of longevity exposure and longevity risk management approaches.

It contains 10 risk components split between those affecting the general population and those that would affect the specific portfolio. The risk components are described in the table below.

  Risk Component Definition Example
Population Risks Event Risk Risk of future longevity events occurring at times or with effects not consistent with the assumptions. Reduction in smoker propensity since the 1970s
Population Modelling Risk Risk that modelling choices or interpretations made regarding the reference population are incorrect (or change) without the data or information changing. Recognition of the cohort effect
Population Mis-estimation Risk The risk that the reference population assumption mis-estimate the correct level of the population mortality rates. Overestimation of mortality improvement in the late 00s (realised following the 2011 census)
Population Volatility Risk Risk of short-term deviations in reference population mortality improvements from the underlying level of mortality improvement owing to systemic effects. Poor performance of 2015 flu vaccine
Portfolio Risks Heterogeneity Risk The risk that lives with materially different longevity profiles are considered homogenous within a classification group. Lack of behavioural or mental well-being information when determining assumptions
Classification Risk The risk that lives are misallocated to a classification group within the mortality basis and utilise assumptions that are not appropriate. Inaccuracy of assumed impairment level based upon individual medical information
Basis Risk The risk that the assumptions derived are not relevant to the lives in the portfolio (including geared impact on the portfolio exposure to future longevity events) Uncertainty in relevance of the slowdown in population mortality improvements since 2011
Portfolio Modelling Risk Risk that modelling choices or interpretations made regarding the portfolio are incorrect (or change) without the data or information changing Uncertainty in the shape and duration of anti-selection effects
Portfolio Mis-estimation Risk The risk that portfolio adjustment assumptions derived from external evidence and/or empirical experience mis-estimate the correct level of the assumption Limited credibility of experience at older ages
Portfolio Volatility Risk The risk that even if the classification and the assumptions are correct the specific mortality and morbidity events that occur cause the pattern of future cashflows to change The number of actual deaths that occur over a period may differ from the expected number of deaths though random variation