Case Study 2 — The 401(k) Default and the Power of Inertia

The most famous and most influential application of behavioral economics to public policy is the way 401(k) retirement plans are designed. The 401(k) plan — a tax-advantaged employer-sponsored retirement savings vehicle — was introduced in 1978 and became the dominant form of private retirement saving in the U.S. by the 1990s. For decades, the standard 401(k) was structured the same way: employees had to actively sign up to participate, choose how much to contribute, and choose how to invest. This is the "opt-in" structure.

In 2006, the Pension Protection Act made it much easier for employers to use "automatic enrollment" — the opt-out structure where employees are enrolled by default and have to actively choose not to participate. The change was inspired directly by behavioral economics research, particularly work by Brigitte Madrian and Dennis Shea (Harvard) and Richard Thaler (Chicago).

The results have been dramatic. They illustrate the power of choice architecture and the practical importance of behavioral economics for policy.

The setup

Before 2006, a typical 401(k) plan worked like this: when you started a new job, your employer's HR department would give you a packet of materials about the retirement plan. The packet would list contribution levels, investment options, and the deadline for enrolling. To participate, you had to read the materials, decide on a contribution rate, decide on an investment allocation, fill out a form (often paper), and submit it to HR. To not participate, you had to do nothing.

The standard model says: rational employees should look at the financial benefits (employer matching contributions, tax advantages, the long-term value of saving) and make an informed decision about whether and how much to contribute. The decision should reflect their actual preferences for current vs. future consumption, their expected lifetime earnings, and their risk tolerance.

In practice, what happened? About half of eligible employees signed up. Half didn't. Among those who signed up, contribution rates were highly correlated with whatever the "default" suggestion was — if the materials suggested 6%, most contributors chose 6%. If they suggested 10%, most chose 10%. If there was no suggestion, contributions clustered around 3% or 4%.

This is not what the standard model predicts. The standard model says that two employees with the same financial situation, the same preferences, and the same retirement plan should make the same contribution decision regardless of how the plan is presented. They should pick whatever is optimal for them. But in real workplaces, the suggested default acted as an anchor, and most people followed it.

Madrian and Shea's findings

In 2001, Brigitte Madrian and Dennis Shea published a study of one large U.S. company that had switched from opt-in to opt-out 401(k) enrollment in 1998. The same company. The same workers (essentially — they compared cohorts hired before and after the change). The same plan. The only thing that changed was the default.

Their findings:

Participation rate before the change: about 49% of eligible workers signed up.

Participation rate after the change: about 86% of eligible workers signed up.

The change was dramatic. The same company, with the same workers in the same labor market, going from less than half to more than four-fifths participation. The only difference was the default.

But there was more. Among those who participated under the opt-in regime, contribution rates were spread across many levels (some 1%, some 3%, some 6%, some 10%). Under the opt-out regime, contribution rates clustered tightly around the default (which the company had set at 3%). Workers who had been free to choose any contribution rate in the opt-in regime mostly defaulted to whatever the company had suggested in the opt-out regime — even when 3% was lower than what they would have chosen actively.

This was a paradox. Auto-enrollment got more workers to participate (good), but it also pushed many of them to a lower contribution rate than they would have chosen on their own (less good). The aggregate effect on retirement savings was positive — the gain from more participants outweighed the loss from lower per-participant contributions — but the effect of defaults on individual choices was striking.

What followed

The Pension Protection Act of 2006 made it much easier for U.S. employers to use auto-enrollment in 401(k) plans. It also encouraged escalation features (the "Save More Tomorrow" plans designed by Thaler and Benartzi) that automatically increased contribution rates over time. By 2015, about 60% of large U.S. employers used some form of auto-enrollment. By 2020, the share was over 70%.

The U.K. went further. In 2012, the U.K. government required employers to automatically enroll their workers in a workplace pension plan unless the worker actively opted out. The result was a participation rate among private-sector workers in workplace pensions that rose from about 47% in 2012 to about 88% by 2018. Millions of new pension savers, simply because the default was changed.

These are some of the largest impacts of any single behavioral intervention in policy history. By changing one default — opt-in to opt-out — millions of workers in the U.S. and U.K. now have substantially more retirement savings than they otherwise would have had.

The mechanism

Why does this work? Several behavioral mechanisms reinforce each other:

1. Status quo bias. Once you're enrolled, doing nothing keeps you enrolled. Doing nothing requires no effort. The same is true in reverse — once you're not enrolled, doing nothing keeps you not enrolled. The cognitive cost of changing the status quo is much higher than the cost of leaving it alone, and most people leave it alone.

2. Loss aversion. Opting out of a retirement plan feels like losing the future benefit. Opting in to a plan feels like losing the current paycheck. The same dollars, viewed from different angles, produce different responses. Loss aversion makes the side that involves "losing" the default option feel more painful.

3. Procrastination from present bias. The decision to think about retirement savings is itself something people procrastinate. With opt-out enrollment, you don't have to decide now — you're already enrolled. With opt-in, the decision feels overwhelming and people defer it.

4. Anchoring. The default contribution rate becomes an anchor. Most people think "the company suggested 3%, so 3% must be right." Many do not actively reconsider whether 6% or 10% would be better for them.

5. The complexity of the decision. Choosing a retirement contribution rate involves projecting future income, estimating retirement needs, calculating tax effects, choosing among investment options, and considering employer matching. It's complicated. Faced with complexity, people fall back on heuristics — and the strongest heuristic is "do whatever the default is."

All of these work together to make the opt-in vs. opt-out distinction enormously consequential for behavior.

Critics and concerns

Not everyone is enthusiastic about default-based retirement saving. Critics raise several concerns:

1. The default may be wrong. A default contribution rate of 3% might be too low for someone who needs to save 10% to be ready for retirement. By making 3% the easy choice, the default may anchor everyone to an inadequate savings rate. (Some companies have responded by raising default contribution rates over time, or by combining opt-out enrollment with default escalation.)

2. Workers are being nudged into saving they may not want. A worker who genuinely prefers to spend now and rely on Social Security in retirement is being nudged into saving against their stated preferences. The standard "respect freedom of choice" framing accepts this as long as opt-out is allowed, but the practical effect is to substitute the architect's preferences for the worker's.

3. The effect on investment quality is unclear. Auto-enrollment increases participation but does nothing to ensure that workers are choosing good investments. Some workers may end up in default investment funds that have high fees or poor risk profiles. The default-on architecture solves the "do they save?" problem without solving the "are they saving wisely?" problem.

4. There are distributional concerns. Workers in lower-paying jobs may be hurt by default contribution rates that take a meaningful share of their take-home pay. Auto-enrollment that's beneficial for higher-income workers may be harmful to workers who can't easily afford the default rate. Some pension experts have argued for income-based defaults, where the default rate scales with income.

These concerns are real. They have led to refinements of auto-enrollment policies — combining default contributions with default escalations, adjusting defaults based on income, providing better investment defaults — but the basic mechanism (changing the default to increase participation) remains widely used and broadly successful.

What this case study illustrates

Lesson 1 — Defaults are not neutral. Every choice architecture involves a default. Choosing the default is choosing what most people will do.

Lesson 2 — Behavioral economics has practical, large-scale policy implications. The 401(k) auto-enrollment story is the most consequential single application of behavioral economics to U.S. policy. Millions of workers have meaningfully more retirement savings as a result.

Lesson 3 — The same behavior can be achieved with very different policies. A traditional approach to retirement saving might have been: tax breaks (incentives), education (information), or mandatory savings (regulation). The behavioral approach is: change the default. The behavioral approach has been more effective than any of the alternatives — and it's cheaper to implement.

Lesson 4 — Nudges can be welfare-improving without being coercive. Opt-out enrollment preserves the choice (you can opt out), but it shifts the path of least resistance to the option that the architect believes is generally beneficial. People who genuinely prefer not to save can still choose not to. People who would prefer to save but have been procrastinating get nudged into doing so.

Lesson 5 — Defaults need to be chosen carefully. A poorly chosen default can do harm. The success of auto-enrollment depends on the default being roughly right for most workers. Setting the default too low fails to capture the benefit. Setting it too high may cause hardship. The right default is a real policy question, not a trivial detail.

Discussion questions

1. Madrian and Shea found that switching from opt-in to opt-out raised 401(k) participation from 49% to 86%. Why is the gap so large? What does this tell you about the assumption of rational decision-making in the standard model?

2. Auto-enrollment also pushed many workers to lower contribution rates than they would have chosen actively. Is this a problem? What policy could fix it?

3. The U.K.'s 2012 mandatory auto-enrollment law took the U.S. approach further by requiring all employers to participate. Should the U.S. adopt a similar mandate? What are the arguments for and against?

4. Some workers actively prefer to spend now and rely on Social Security in retirement. Auto-enrollment nudges them away from their stated preferences. Is this acceptable paternalism, or unacceptable manipulation?

5. The behavioral approach (change defaults) has worked better than traditional approaches (tax breaks, education, mandates). Why? What does this tell you about how to design effective policy in general?