In 2012, an estimated 6.4 million seniors in the United States were living in poverty. Although Social Security has helped reduce late-life poverty in the 20th century, the shifting landscape of retirement plan options has left many seniors using it as their primary or sole source of income. That makes many people’s choice of when to begin claiming benefits one of the most important determinants of their financial wellbeing during retirement. The earliest age at which people can claim benefits is 62, but because benefits grow the longer they wait to claim, claiming at 62 can also be a costly, sometimes shortsighted choice.

To tackle the problem that many people appear to claim benefits too early, researchers from the Social Security Administration and Columbia Business School teamed up to design a subtle, nearly costless intervention in order to nudge people towards later claiming and hence improve their financial outlooks. In their recent study, “Time to retire: Why Americans claim benefits early & how to encourage delay,” people who simply considered the merits of delayed claiming before considering the merits of earlier claiming reported a preference to claim later than subjects who considered the merits of those options in the reverse order.

Why is delayed claiming of Social Security benefits so important?

Less than 30 years ago, Social Security was only one of “three pillars” of financial stability during retirement, along with defined benefit plans and savings. Defined benefit plans guaranteed a monthly income to the majority of US workers. Today, the vast majority of US employees instead participate in defined contribution plans, placing the onus on them to save and budget any funds that would supplement Social Security during retirement.

Yet Americans save notoriously little. The majority of married households have saved less than $50,000 for retirement by the time the head of the household turns 64, and the majority of single individuals have saved nothing for retirement—at any age. The result is that many individuals outlive what savings they have, and Social Security becomes their primary or only source of income.

The problem is especially pronounced among individuals who live to be older than average. The poverty rate for seniors 90 and older is 51 percent higher than it is for seniors 65-89 years old, and Social Security is the largest source of these seniors’ income.

Because the Social Security Administration increases benefits by 8 percent per year up to age 70, a person who begins claiming at age 70 could receive almost double the benefits per year that they would have received if they began claiming at age 62. Yet about half of US citizens begin claiming at age 62.


American consumers are used to comparing cars in terms of miles per gallon, and they know that, all else equal, more is better—the car will use less gas and thereby cost less to run and hurt the environment less.  Unfortunately MPG itself is tricky to use when estimating cost savings and environmental benefits.  Consider a family that owns two cars, one that gets 10 MPG and another that gets 20MPG, both driven the same distance each year (e.g., 10,000 miles). They are considering trading one in for a more efficient vehicle. Which trade-in would save more gas?

A. 10 MPG vehicle for a 20 MPG vehicle

B. 20 MPG vehicle for a 50 MPG vehicle

Intuitively, B seems far better – it is a larger improvement in MPG both in absolute and relative terms.  However, to know the actually gas savings, one has to calculate them by taking a distance, such as 10,000 miles, and dividing it by MPG.  The relationship between MPG and gas consumption is highly curvilinear (technically, a reciprocal) and this has surprising implications.  Dividing 10,000 miles by MPG reveals that the 10 to 20 MPG trade in reduces gas use from 1,000 gallons per year to 500 gallons per year; the 20 to 50 MPG trade in reduces gas use from 500 gallons to 200 gallons.  The family saves more gas by having two 20 MPG vehicles than a 10 and 50 MPG vehicle.  In this example, “doing the calculations” means giving consumers a consumption metric in which the amount of gasoline used is calculated for some meaningful distance, such as 100 miles of driving (Europe, which uses the metric system, uses a consumption metric of liters per 100 kilometers). In this case, choosing Option A saves an estimated 300 more gallons of gas than Option B.


figure 1

Framing is a widespread intervention in the “choice architecture” literature, which seeks to uncover how the design of a decision environment influences choice. Yet, in this case, the intervention had no detectable impact. Even when individuals saw a graph showing how earlier claiming resulted in losing money relative to what they could receive, many still preferred to claim early.

The second intervention varied the order of prompts asking people to consider the benefits of claiming early and the benefits of claiming late. Although it shouldn’t logically make a difference which option people consider first, previous research on query theory  suggests that thinking about one option in a choice set actually makes it harder to think about alternatives; the order in which you think about options matters.

According to the theory, when people think about early claiming first, the benefits of doing so—perhaps the relief of leaving an unpleasant job or using the extra income for long overdue house repairs—are vivid in their minds. Subsequent thoughts of delayed claiming—such as having more money during years when medical expenses might be highest—are simply harder to access when the benefits of early claiming are already top of mind. However, when the order of these considerations is reversed, thoughtful consideration of delayed claiming has a fighting chance.

The authors found that people who first considered the reasons for delayed claiming reported that they preferred to claim an average of 9.4 months later than those who first considered the reasons for early claiming, which would typically lead to a $660 increase in benefits per year. While seemingly modest, every $1,000 increase in annual Social Security benefits leads to a 2 to 3 percentage point reduction in poverty rates among senior households. With almost 3 million US citizens turning 65 every year, if the authors’ low-touch, essentially costless intervention actually led to a 9.4-month average delay in claiming, it could potentially help thousands of seniors every year stave off poverty.

Future research can explore if individuals follow up on their stated preferences to delay claiming. As the literature on present bias demonstrates, stating a preference to make a prudent choice does not mean that people actually make that choice when it is time to act. However, the query theory-based intervention is a low-cost way of nudging individuals approaching retirement to plan a little more wisely for their future.


Shannon is a PhD Candidate in Behavioral Science at the University of Chicago Booth School of Business, where she is studying contextual influences on attitudes, preferences, and behavior. She is a former Senior Associate of ideas42, a New York City-based behavioral design lab, and she earned her MPP at the Harris School of Public Policy.