PANC 2022: The Psychology and Practicality of Retirement Income

While automated features have transformed the way participants save for retirement in defined contribution plans, one of the authors of such features says they are not the answer for decumulation and creating income in retirement.

Speaking at the 2022 PLANADVISER National Conference in Scottsdale, Arizona, Shlomo Benartzi, professor emeritus and co-founder of the Behavioral Decision-Making Group at the UCLA Anderson School of Management, said auto-features that harness the power of participants’ inertia to help them save for retirement won’t work to turn DC plan savings into a paycheck in retirement.

“Over a lifetime, people accumulate assets and differences,” said Benartzi, who, along with Nobel laureate Richard Thaler of the University of Chicago, pioneered the Save More Tomorrow program to nudge employees to increase their retirement savings rates gradually over time. “You can’t give the same solutions to such different people.”

Benartzi also cautioned that participants should not be enrolled into income solutions that are irreversible, as circumstances—including health and cognitive ability—can change in retirement, and income may need to as well.

During a panel session titled “The Psychology and Practicality of Retirement Income” with Benartzi, George P. Fraser, managing director at the Fraser Group, Retirement Benefits Group, and Michelle Richter, executive director of the Institutional Retirement Income Council, the speakers outlined the details of a new program from Benartzi, Pension Plus, that uses data points available from a plan recordkeeper and a short one-on-one interview to personalize an individual’s decumulation plan and create a paycheck in retirement.

“Without personalization, you cannot deal with decumulation,” said Benartzi.

Fraser, who has used Pension Plus with participants of plan sponsors that he advises, called it “really simple for people to understand.”

“For the average American, it gives people the ability to know what their [weekly, bi-weekly or monthly] check is going to look like,” said Fraser.

Richter stressed that personalization, or “the need to understand the human,” is a key difference between the asset accumulation phase of retirement planning and the decumulation phase, when people are living off their assets. “Asset management doesn’t require knowing the human at all,” she said. “In the field of asset accumulation, you can rely on modern portfolio theory.”

Conversely, she cited Nobel laureate William Sharpe’s assertion that decumulation is “the nastiest problem that exists in finance.”

The Pension Plus tool uses information from individuals, starting with their gender and the age at which they want to retire, their investment risk tolerance, whether they want to leave any bequests to family or charity and their preferred spending path. (Do they want to spend more when they are younger in retirement or when they are older?)

Fraser indicated that in his experience, it takes about 15 to 20 minutes in a one-on-one meeting with a participant to explain the tool and the concept of creating a retirement paycheck. Benartzi added that the process could be more automated if more information, including gender, could be available from recordkeepers’ systems.

While systems that assume the most popular options for all, as is common in auto-features, work well on the accumulation side of retirement plans, those same assumptions would end in results appropriate for only 4% of people in decumulation, Benartzi said.

Using as examples a man and a woman of the same age, retirement account balance and self-assessment of their health as good, Benartzi demonstrated the tool and how different answers to the questions would result in vastly different paychecks in retirement. In the example, one person’s set of choices led to a monthly paycheck of $4,959, while the other resulted in a check of $761.

“This is personalization in action,” said Benartzi. “If you take the most popular choice [instead of the personalized choice] on many dimensions, at the end you might find that it fits nobody.”