The Effect of Default Options on Retirement Savings
Many workers in the U.S. and around the world have defined contribution retirement savings plans, either through their employers or as part of the social security system. Workers covered by these plans typically must make a number of decisions about the plan, which may include whether to participate, how much to contribute, how to allocate plan assets across various investment vehicles, and how to decumulate assets following retirement.
For each of these decisions, there is usually a default option; for example, it could be that workers are not enrolled in the plan unless they opt in or that plan assets are invested in a money market fund unless workers select a different asset allocation. Standard economic theory suggests that defaults should have little effect on retirement savings outcomes - if the default option is not the best choice for the worker, he will simply switch to his preferred option, as long as switching is not particularly costly.
In practice, however, defaults seem to matter a great deal. This is the conclusion of researchers John Beshears, James Choi, David Laibson, and Brigitte Madrian in "The Importance of Default Options for Retirement Savings Outcomes: Evidence from the United States" (NBER Working Paper 12009). The authors look at the effect of defaults on participation, contribution rates, asset allocation, and postretirement distributions, as well as consider some of the explanations for the persistence of defaults.
Most employer-provided pension plans in the U.S. have standard enrollment, where workers must opt in if they wish to participate. However, some plans have automatic enrollment, where workers instead must opt out if they do not wish to participate. The authors find that having automatic enrollment leads workers to join the plan at a much faster rate. In one firm that switched from standard to automatic enrollment for new hires, the plan participation rate was 35 percentage points higher after three months on the job under automatic enrollment, and remained 25 points higher after two years.
The authors next examine the choice of how much to contribute and find that here too, defaults matter. In one firm with a default contribute rate of 3 percent of salary, more than one quarter of workers contributed exactly that amount to the plan, despite the existence of a dollar-for-dollar employer match on contributions up to 6 percent of salary. Once the firm switched to a 6 percent default, virtually no new hires selected a 3 percent contribution rate.
Defaults affect asset allocation as well. The authors look at the experience of one firm that switched from standard to automatic enrollment, where only workers hired under automatic enrollment had a default investment fund. They find that fully one-third of workers hired under automatic enrollment invested all of their assets in the default fund, a choice made by virtually no workers hired under standard enrollment.
The authors also discuss the evidence on defaults and the decumulation of assets after workers leave their jobs. Defaults have been shown to affect whether a young worker takes his account balances as a cash distribution or keeps it invested in a retirement account, and also whether an older worker takes his account balances as a single or joint life annuity. Thus defaults are extremely influential at every stage of retirement plan decision-making - participation, contributions, asset allocation, and decumulation.
Next, the authors discuss some possible explanations for the persistence of defaults. One explanation is the complexity of making retirement plan decisions. Surveys of financial literacy find that many individuals lack basic financial knowledge about concepts like risk or compounding. The psychology literature has established that individuals are more likely to put off making decisions as the complexity of the decision increases. Defaults that simplify retirement decision-making can thus encourage participation. A second possible explanation is that some individual have problems with selfcontrol, which may lead them to postpone enrolling in the plan or to make other decisions that are inconsistent with their long-term goals. A third explanation is that individuals perceive the default as an endorsement of a particular course of action.
The evidence indicates that all of these factors contribute to the importance of defaults. For instance, "quick enrollment," which requires individuals to opt in to the plan but offers a preset contribution rate and asset allocation, generates a higher participation rate than standard enrollment but a lower participation rate than automatic enrollment. This shows that simplifying decision-making is helpful, but that self-control problems also play a role. There is also substantial evidence for an endorsement effect - for example, workers are more likely to allocate their own contributions to company stock when the employer match is made in company stock.
The importance of defaults combined with the trend towards more defined contribution retirement plans suggests that the actions of policy makers can have a large impact on retirement savings outcomes. There are many examples of existing public policies that encourage retirement savings, such as the welldiversified default asset allocation in the Swedish pension system. But there are also policies that discourage savings, such as the regulation allowing employers to compel a cash distribution of plan assets for departing employees when assets are less than $5,000. As the authors conclude, "defaults are not neutral - they can either facilitate or hinder better savings outcomes."
The authors acknowledge financial support from the National Institute on Aging (grant R01-AG021650 and T32AG00186) and the Social Security Administration (grant 10-P-98363-1).