Anticipating the Risk of Dependency and Assets

The significant gains in life expectancy at advanced ages that have been recorded over recent decades, combined with the arrival of the baby boom generations at old age, suggest a rapid growth in the costs of long-term care. This perspective fuels debates on the most suitable method for financing the loss of autonomy: in the absence of sufficiently broad insurance coverage and given the difficulty of evaluating both the probability of becoming dependent and the attendant costs, the risk is high that households will not save enough. This article addresses the issue: by drawing on data from the 2020 wave of the Pat€r survey, we evaluate to what extent households adapt their savings effort in function of their estimation of the probability that they will lose their autonomy on the medical level. To this end, we estimate household assets according to their own assessment of the risk of becoming dependent one day, controlling notably for their self-reported health status. To account for the possible endogeneity of expected dependency risk on household savings (people save less if their health is poor), we use instrumental variable estimation methods. Our results confirm that people with a higher self-assessment of the risk of becoming dependent accumulate more assets: measured at the average self-assessed risk, an increase of this risk by one standard deviation is associated with an increase in overall gross wealth of three to eight months of permanent income.