This "annuity puzzle" has challenged economists for years, given that annuities are so effective at smoothing income with little risk of outliving it, yet Finke concludes that ultimately the real problem is that most people aren't economists; they have more complex and nuanced goals, and irrational biases, not typically captured in standard economic models.
Yet notably, for the inflation-adjusted annuity that everyone has by default - their Social Security payments - people exhibit the opposite irrational extreme; when asked what lump sum would be necessary to give up $500/month in Social Security payments in 2008, the average respondent in the Health and Retirement Study responded $250,000 - equivalent to an assumed longevity of 130 years with a 0% discount rate. Our unwillingness to sell something for a fair price once we own it is known as the endowment effect, which explains the overvaluing of Social Security payments and our unwillingness to convert them to a lump sum, but Finke notes that the endowment effect is also likely what causes us to overvalue our liquid retirement sums and become unwilling to convert them to an annuity.
Accordingly, Finke suggests that one of the best ways to consider addressing the issue is by "reframing" their retirement assets as income instead of wealth; for instance, the Department of Labor is considering an approach where retirement plan sponsors would illustrate retirement assets based on the income they would buy rather than just their future lump sum value (especially since most people have no idea how to translate one to the other intuitively).
Another challenge is dealing with loss aversion - even if it's unlikely, we fear the losses associated with dying shortly after purchasing an annuitized stream of income; Finke notes that while economically obtaining a cash refund guarantee is less optimal, in that it reduces income, if it helps to manage the loss aversion fear, this "imperfect alternative" may still be a better course of action.