Actuarial Predictions and Subjective Survival Expectations: Comparing Two Ways to Foresee the Random Future

Future longevity is a key issue for both life insurance companies and potential insureds, as it affects the
valuation of premiums and future liabilities, for the former, and the individual purchase behavior, for the latter. While life insurance companies rely on extrapolation procedures or on stochastic mortality
models to forecast future longevity patterns (e.g. considering age, period and cohort effects), potential insureds formulate their expectations about their residual lifetime basing on a richer set of information and on heuristics (cf. Tversky and Kahneman 1974). As stressed in (Perozek 2008), individuals can make mortality predictions by exploiting their own genetic background, along with environmental and behavioural factors that often represent private information; their expectations about survival probabilities may thus provide further relevant information to practitioners interested in forecasting longevity.
We aim to present our empirical findings on the relationship between subjective survival expectations and actuarial forecasts, thus between the outcome of human reasoning and the output of mathematical/statistical models, being motivated by the interest in making the life insurance companies' world and methodologies to dialogue with the individuals' universe and expectations. This
study paves the way for the discussion on: (i) whether and how much actuarial forecasting methods (cf. Apicella and Sibillo 2018, Apicella, Dacorogna, Di Lorenzo and Sibillo 2019) can benefit from the wide range of information incorporated in subjective expectations, without sacrificing parsimony, (ii) how much the possible divergence between subjective expectations
and actuarial forecasts impacts on the consumer purchase behaviour towards life insurance products.