References

consumer choice in policy and in business applications. In many respects, the use of default options can be considered the most successful contribution of behavioral economics to public policy at the present writing.

Rational Choice and Default Options

Standard models of economic choice consider that people evaluate every option based upon the utility they will derive from the choice and then select the option with the highest utility. Thus, default options should hold no special place in the mind of the individual decision maker. For example, someone choosing between two possible choices x and y simply chooses x if the utility derived from consuming x is greater than the utility derived from consuming y, ux > uy. If more utility is derived from consuming y, then y will be chosen.

Suppose x represents choosing a car with an automatic transmission, and choosing y represents an identical car with a manual transmission. Cars often have an automatic transmission by default unless the consumer requests a manual transmission. Consumers might decide to request manual transmission, for example, if they live in a snowy area where the manual transmission could add to their ability to control their car on slippery surfaces. Unless the customer explicitly enjoys choosing whatever item is named the default, the naming of a default should not change the preference for either choice x or choice y. Thus, naming a default option should have no impact on choice so long as making the choice is free and so long as the customer is not indifferent between the two options. So, if suddenly auto manufacturers named manual transmission the default, we would expect most consumers to persist in choosing the automatic transmission if it were not costly to do so. In this case, we might expect the default option to affect choice only when the options are not very different in terms of outcome or when the cost to switching to the nondefault option is large. In other cases where the decision is relatively costless and where outcomes are relatively dissimilar, we should expect very little impact from switching the default option.

EXAMPLE 4.1 Organ Donors

The lack of suitable organ donors creates a constant problem in the United States and in many other countries. Those who have failing kidneys, livers, or other organs can extend their life significantly if their failing organs can be replaced with healthy organs harvested from those who have recently died. Thousands of people die each year as a result of the dearth of suitable organs. One of the primary reasons for the scarcity of suitable organs is that a majority of Americans have chosen not to donate their organs upon death. Only 28 percent of U.S. citizens sign the donor card necessary to become an organ donor.

Eric Johnson and Daniel Goldstein found that the level of organ donations across countries is very closely related to the default option specified. In the United States, Denmark, the Netherlands, the United Kingdom, and Germany, people are not considered for organ donation upon death unless they explicitly specify their willingness to be an organ donor. In each of these countries the rate of organ donors is less than 28 percent, and in the case of Denmark it is only around 5 percent. Alternatively, Austria, Belgium, France,

Hungary, Poland, Portugal, and Sweden all consider people to be organ donors unless they have explicitly stated their objection. Aside from Sweden, the rate of those selecting organ donation is greater than 98 percent; the rate in Sweden is about 85 percent.

To determine how much of this effect may be due to the cost of choosing, Johnson and Goldstein ran an online experiment asking subjects to make a hypothetical choice. Some subjects were asked to imagine that they had just moved to a new location where one was not considered on organ donor unless one stated one’s preference to donate. Others were given the same scenario but were told that one was considered an organ donor unless otherwise stated. Finally, a third group was asked to state their preference when no default was designated. Donation rates for those who were required to opt in were 42 percent, compared to 82 percent and 79 percent in the opt-out and no-default conditions. We call a preference for the default option the default option bias.

EXAMPLE 4.2 Journal Subscriptions

Members of professional societies often receive research journals as part of their membership. For example, the American Economic Association for many years included subscriptions to three journals in the price of membership: American Economic Review,

Journal of Economic Literature, and Journal of Economic Perspectives. With the increasing printing costs of journals, some associations sought to save money by offering members a discount for eliminating their subscription to some journals. This brings up the question of what the default option would be. Apparently, at one point the American Economic Association considered offering a discount for subscribing to only two of the three journals and receiving the third only electronically. Daniel Kahneman, Jack Knetsch, and Richard Thaler report that many prominent economists involved in this decision believed that more members would subscribe to all three journals if that was presented as the default option, with a discount for eliminating one, rather than a default to receive two of the journals with an optional extra fee for receiving the third.

Preference Formation, Framing, and the

Default Option

The rational model of choice supposes that people have well-formed and consistent preferences. It is conceivable, or even probable, that people have not formed preferences over many of the choices they face. If this is the case, people may be unduly influenced by the framing of the question or other subtle cues that help to determine the value. Johnson and Goldstein propose that default options shape choice by providing an anchor for subsequent decision making. Essentially, in the absence of preferences, a person looks for suggestions. The default option fills the void of preferences.

This is not the only realm where preferences appear to respond to suggestions. For example, Dan Ariely, George Loewenstein, and Drazen Prelec conducted an experiment in which many M.B.A. students were presented with various items with an average retail price of around $70. Each student was first asked whether he or she would be willing to

required for the longest duration when the longest was asked first was about the same as the compensation required for the shortest duration when the shortest was asked first. Although preferences appear coherent in many circumstances, they can result from arbitrary anchoring.

EXAMPLE 4.3 Retirement Savings and Default Options

The use of defaults to shape behavior has often been discussed in the context of retirement savings. There is a documented need to increase the propensity of workers to save more for retirement, preventing future public expenditures. Nearly 50 percent of Americans exhaust their retirement savings before they die. Brigitte Madrian and Dennis Shea examined the enrollment of employees at a particular firm in 401(k) retirement programs. These programs generally require some minimum level of contribution by employees from their regular paycheck. The employer then contributes some additional amount to the retirement account. The money is allowed to accumulate and earn a return on investment that will not be taxed until it is withdrawn from the account after retirement. Before April 1998, this firm had allowed employees to opt into the 401(k) if they wished. After that date, employees were automatically enrolled unless the employee decided to opt out. Controlling for age disparities, they compared enrollment rates of those hired before and after the change in the default, finding the default option increased participation by about 50 percent. Thus, policymakers interested in reducing the number of elderly with inadequate retirement savings may be able to induce savings without eliminating the employee’s participation choice by simply setting up a default option.

EXAMPLE 4.4 Insurance and the Right to Sue

States often set guidelines for the types of auto insurance policies offered in its borders. At one point, many states considered limiting the types of lawsuits that one could pursue following an accident. In particular, they debated allowing one insurance policy that allowed full rights to sue for any damages and another policy that cost somewhat less that limited the amount the policyholder could seek in certain types of lawsuits. Again, this leads to the question of which policy would be considered the default option. A team of researchers led by Eric Johnson conducted an experiment asking subjects to imagine they had moved to a new state and that they needed to decide on the type of coverage. One group was presented the full-rights insurance policy as the default, a second was given the limited-rights policy as the default, and a third group was presented the two policies without any default specified. In each group, a set of fixed prices was associated with both choices. Of those presented the full rights as the default, 53 percent retained these rights, whereas only 28 percent retained these rights when the limited policy was the default. Of the neutral group, 48 percent decided to choose the full-rights policy. Thus, the amount one is willing to pay for the right to sue depends on how the choice is framed. If the default includes this right, giving it up requires more money in compensation than if the default excludes this right.