According to the standard life-cycle model, you are supposed to accumulate assets while working and then decumulate them once retired. The optimal outcome is to die with nothing left, but bequest motives and uncertain lifetimes make that difficult. Still, assets should get reduced over time. Does this square with the data? No it does not, as the elderly cling to their wealth, especially their illiquid wealth such as housing, for which they make little attempts to extract any cash from, for example through reverse mortgages or annuities (previous posts I, II).
Agnese Romiti and Mariacristina Rossi try to understand why using European data on retirees. It turns out that financial literacy is a major factor here. Financially sophisticated households (identified by numeracy tests in the survey) have a more balanced portfolio, in particular with less illiquid real estate. They also make ends meet more easily, which the authors interpret as having a consumption path closer to optimal. Finally, they are actually decumulating their housing wealth, thus more likely to actively doing something about the illiquid wealth they are sitting on. As the paper is exploiting a data set that has several cohorts, it would be good to see whether the situation is actually improving, that is, whether younger cohorts do better.