Annuities Get a Behavioral Finance Makeover

An appreciation for the influences of behavioral finance has had a critical impact on retirement plan designs – so what about applying those principles to the decumulation phase?

To explore the alternatives and implications Allianz worked with Professor Shlomo Benartzi of UCLA to reach out to a number of academics in the field, several of which were on hand at a meeting in New York City to present some of those findings and their implications.

The findings, which Allianz said it submitted as its response to the February request for information by the Department of Labor and Treasury Department on retirement income products, was produced under the title “Behavioral Finance and the Post-Retirement Crisis”, broadly defined as being “about outliving your assets.” 

Indeed, while the issues attendant with crafting an effective retirement income solution are many, much of the panel’s discussion was oriented toward the challenge presented by longevity risk, simply stated the risk of outliving one’s savings.  To kick things off, Benatzi used the example of ten high school friends who retire at age 65.  Of those, Benatzi said that the first of the 10 would die just four years into retirement, at age 69, while the last in the group to die would, according to statistics, not die until age 99. 

This post-retirement crisis is magnified by the poor financial decision-making of retirees, who according to the research presented, pay too much attention to recent stock market performance (those retiring after stock market increases of six to 12 months are much more likely to select the lump sum option rather than lifetime income), have trouble making financial decisions, and are “hyper” risk averse.  

“Nudging” the Annuity Decision

Dr. Alessandro Previtero of UCLA highlighted a finding that would seem to defy “common wisdom” about annuities, specifically the reality that, when given choice to pick a lump sum as well as an annuity in a defined benefit plan, the vast majority opt for the lump sum.  He cited a study that found that in a period from 1999 to 2005, only 2% to 6% of retirees elected guaranteed lifetime income when it was available in their 401(k) plans – much lower than expected, and a disparity he referred to as the “Annuity Puzzle” (the puzzle being why people don’t choose annuities).

However, Previtero recently conducted new research with what was described as “a unique dataset of more than 100 defined benefit plans”, covering more than 100,000 retirees. Each of these individuals had to actively choose between guaranteed lifetime income and a lump sum. Because there was no default, they had to decide themselves how to withdraw funds – and Previtero said that 49% of retirees making an active choice between guaranteed lifetime income and a lump sum actually picked the lifetime income option. 

He went on to note that defined benefit payouts are typically communicated in terms of producing monthly income, and annuity payout options tend to look attractive to participants accustomed to thinking of those benefits in like terms; but he contrasted that with cash balance plans that are often, like defined contribution plans, are communicated in terms of account balances or lump sums.  Previtero said he found that retirees in defined benefit plans were 17% more likely to choose the guaranteed lifetime income than their peers in cash balance plans. 

The recommendation; make retirement income solutions available in 401(k) plans and, somewhat ironically considering the panel’s general affinity for using defaults to help participants make better choices – “nudging” retirees to actively make a choice.


Another study cited was one by Professor Jeffrey Brown of the University of Illinois, who found that when an annuity choice was presented in a “consumption” frame, in other words as providing monthly income of $650 for life, 70% preferred the annuity. But when it was presented in an “investment frame” (an investment with a $650 return for life), only 21% opted for the annuity.  It was noted that defined contribution plan designs have “taught” people to think of these accounts in terms of investments, rather than pension plans, where participants are more likely to think of the monthly benefit they provide.  The suggestion from the researchers; present the programs with an emphasis on the income they will provide, not the return on the investment made.

Professor Eric Johnson of Columbia University pointed out that for most of us, “losses loom larger than gains.” That is, investors experience the pain of a financial loss much more acutely than they feel the pleasure of the same size gain – and by a factor of about two to one.  However, recent research he conducted with ACLI and AARP found that retirees displayed “hyper” risk aversion – which meant that they tended to weight losses about TEN times more heavily than gains.

Now, while Johnson said he assumed that this hyper loss aversion would translate into a preference for products with guaranteed lifetime income, his research revealed that retirees with hyper loss aversion actually responded less favorably to financial products with more protection and guarantees.  Johnson said that it seemed that to this group, giving up control of their money was viewed as just another type of loss.  Consequently, he said that solutions should emphasize that those guarantees and protections were a way to restore, not surrender, control.

Cognitive “Dissonance”

A study by Professor David Laibson of Harvard University reported a significant decrease in “analytic cognitive functioning” as people age, as well as an increase in the occurrence of dementia. For example, after age 60, the prevalence of dementia roughly doubles every five years, and the research suggests that by the time people reach their 80s, more than half will suffer from either dementia or other “significant cognitive deficits”.  The older adults that Laibson studied also showed marked declines in “numeracy”— the mathematical skills needed to cope with everyday life and to understand information in graphs, charts or tables, and they also had “great difficulty” understanding simple measures of risk.  The optimal age for making those decisions?  53.

A suggestion to counter this problem; solutions, “including investment strategies and public policies that encourage people to make binding decisions earlier in life and prior to the onset of cognitive impairment”.

One of the reports included in the handouts was based on an interview with Professor. Brigitte Madrian of Harvard University, who suggested that one-size-fits-all-defaults are ill-suited to helping different groups of participants achieve optimal results.  Before setting defaults, she said that plan sponsors would be well advised to evaluate the potential impact of inertia on different types of retirees, and said that policymakers should make it easier for sponsors to customize decumulation options by eliminating non-discrimination rules that require all retirees—even those with unique needs—be presented with the same default payouts.


Future Self

Professor Daniel G. Goldstein of Yahoo Research and London Business School said that part of the problem was that people have trouble relating to their “future self” when it comes to making decisions and trade-offs regarding retirement savings (Seinfeld fans would have enjoyed his analogy of the “morning guy” and the “night guy” – if you haven’t a clue what that’s about, check out, pointing out that those with a strong connection with the perspective of their future self tend to save more and invest better (in one of the more whimsical parts of the presentation, he suggested the deployment of tools that could age pictures of the individual as a way to better help them visualize that future self).

“Obvious” Decisions

An interview with Professor John Payne of Duke University included in the materials said that for lifetime income solutions, retirees are typically presented with materials highlighting the monthly payouts provided by each option – and for many, the optimal choice is obvious: the highest monthly payout.  That, in turn, tends to lead them toward choosing single life annuities (with larger payouts), rather than joint and survivor.  In fact, 69% of married women and 28% of married men opt for single life annuities rather than joint and survivor annuities, according to the report.  But, that tendency to go for the option that is easiest to understand means that retirees may fail to recognize the implications of their decision on their spouse.

Other challenges highlighted in the report; people vastly underestimate the impact of inflation on their cost of living (leading to a suggestion that some kind of inflation protection be incorporated in retirement income solutions), and that the attractiveness of a retirement income solution depends on its perceived fairness.  Part of the perceived “unfairness” of the annuity is a perception that an early death on the part of the annuitant benefits the financial institution that issued the product.  Professor Suzanne Shu of UCLA suggested countering that perception by positioning it as benefiting other people in the annuity “pool.”