Increasing Savings Rates

When Sunstein was a young law student, he interviewed for summer jobs at Washington law firms. In one interview, a senior partner asked the classic question: “What would you most like to ask me?” A bit terrified, Sunstein shyly responded, “What’s the best thing about your firm?” The partner answered, “We have a really terrific pension plan!” Sunstein wasn’t sure what a pension plan was. He thought it might be something for old people—maybe it had something to do with retirement—but he would have rather heard about the lunches. He later learned that helping people plan for retirement is crucial.

Saving for retirement is one of the hardest tasks Humans face. It’s computationally difficult, requires sustained self-discipline, and is relatively new for our species. Formal retirement planning is comparatively recent. Older people have often relied on continued work, family support, or other arrangements. Longer lives and geographically dispersed families helped make organized provision for retirement more important. In 1889, Otto von Bismarck’s early social security program in Germany began addressing this problem. Early private pension plans were typically “defined-benefit” plans, where workers were promised a specific stream of payments in retirement, often based on final salary and years of service. From a choice-architecture perspective, defined-benefit plans are forgiving: employees need to make almost no decisions—just provide a Social Security number. The main catch is that those who change jobs frequently might end up with very little if benefits haven’t vested.

In the United States, a new type of employer-based retirement plan, the 401(k), emerged in 1980. These “defined-contribution” plans specify the contributions put into a tax-sheltered account in the employee’s name, but the ultimate retirement income depends on the employee’s decisions about how much to save, how to invest, and how to manage the portfolio over decades. Defined-contribution plans are portable and customizable, but they shift the responsibility—and the risk—to the employee. Workers must enroll, choose a savings rate, manage investments, and decide what to do with the money in retirement. Many people find the process frightening, and many make poor choices.

The 3 percent savings rate used in early automatic-enrollment plans was never meant as a recommendation. Advisory rulings clarifying the legality of automatic enrollment included a phrase like: “Suppose a firm automatically enrolls employees into a retirement plan at a three percent savings rate…” The official meant this only as an example, but anchors are highly influential. For years, nearly every firm using automatic enrollment adopted 3 percent as the default, leaving many workers saving far too little.

In response, Thaler and his collaborator Shlomo Benartzi devised a choice-architecture system they called Save More Tomorrow. Their design was conscious of five important psychological principles:

• Many participants say they think they should be saving more and plan to save more, but never follow through.
• Self-control restrictions are easier to adopt if they start sometime in the future. (Saint Augustine’s prayer: “God, give me chastity… but not yet.”)
• Loss aversion: people hate to see their paychecks go down.
• Money illusion: losses are felt in nominal dollars, not adjusted for inflation. A 3 percent raise feels like a gain even with 4 percent inflation, but a cut in take-home pay is strongly resisted.
• Inertia: after joining a plan, many employees never revisit their choices for a decade or more.

The Save More Tomorrow plan invited participants to commit in advance to a series of contribution increases timed to coincide with pay raises. The design aimed to keep the contribution increase within the pay raise so that take-home pay would not fall and the higher savings would not feel like a loss. Once enrolled, the increases became automatic, using inertia to raise savings rather than prevent them. Combined with automatic enrollment, this approach could achieve both high participation rates and progressively higher savings.

The first implementation occurred in 1998 at a midsize manufacturing firm. About 90 percent of employees accepted an offer to meet with a financial consultant, who used software to suggest a savings rate based on individual circumstances. The software often recommended the maximum allowed (15 percent of pay), but the consultant knew that would be rejected and typically suggested an increase of five percentage points. About 25 percent of participants accepted immediately. The rest said they couldn’t afford the pay cut; they were offered the Save More Tomorrow program, where their savings rate would increase by 3 percentage points with each pay raise (typical raises were around 3.25 to 3.5 percent). Of those reluctant savers, 78 percent joined.

The results were striking. Over three years, employees who had declined the consultant’s meeting maintained a flat savings rate of about 6 percent. Those who accepted the immediate 5-point increase jumped from just over 4 percent to just over 9 percent and stayed there. Those in Save More Tomorrow started with the lowest rate, around 3.5 percent, but after three and a half years and four pay raises, their savings rate almost quadrupled to 13.6 percent—well above the 9 percent of the immediate-increase group. Most Save More Tomorrow participants stuck with the program until they hit the maximum allowed contribution; the few who dropped out merely stopped increasing, rather than reverting to lower rates.

In designing Save More Tomorrow, Thaler and Benartzi used a “kitchen sink” strategy, incorporating as many favorable features as possible. Over time, they learned that two ingredients were attractive but not essential: tying increases to pay raises (which many firms find hard to implement) and letting people decide now to join later. The program was simplified into automatic escalation: savings rates automatically increase annually, usually by 1 percentage point per year. Some firms make automatic escalation part of the default enrollment plan, starting employees at 3 percent and increasing by 1 percentage point yearly up to a cap like 10 percent; others offer it as an option. In the Vanguard study cited in this account, about 70 percent of firms using automatic enrollment included automatic escalation as an opt-out default, and nearly all the rest offered it as a choice.

Automatic escalation does help to increase savings rates, though not as dramatically as the original experiment. In the Vanguard sample, about half of participants who remain at the firm after three years continue with the scheduled 1 percentage point annual increases, while most of the others have taken steps to raise savings even more. These ongoing trends are encouraging, especially because the accidental 3 percent starting savings rate persists. Still, there is a gradual shift toward higher starting rates—either 4 or 6 percent—which, combined with automatic escalation, can raise contributions. Whether the resulting savings are adequate still depends on workers’ circumstances.