Policy makers are taking a hard look at so-called "soft paternalism" these days, i.e., the idea that people who are making the "wrong" choice can be guided to the right choice by changing the default option. Soft paternalism is a response to the deviation between predicted and observed behavior. For example, many people say that they want to save for retirement, but then fail to enroll in a retirement contribution program. Instead, they "choose" to contribute nothing, i.e., they do not change their retirement contribution from the default ("contribute nothing") to an alternative ("contribute x% per month").
Note that economists often assume that rational individuals will make the choice that they prefer and that suits them. When they observe low retirement contributions, they conclude that people PREFER low contributions. Psychologists, OTOH, claim that people are boundedly rational, i.e., they do not make calculations for every choice they face. In the past, "too low" retirement contributions would lead economists and psychologists to make different recommendations. Economists would accept the contributions as rational and optimal; psychologists would say that they are boundedly-rational and suboptimal. Their policy recommendations would also differ. Economists would recommend providing more information (for better choices); psychologists would recommend a mandatory minimum contribution. Soft paternalism strikes a compromise -- it changes the default contribution from zero, but it allows people to choose (optimally!) to change it back.
Soft paternalism has had some good success. On the pension side, more people are saving more money. In another example -- donating one's organs when dead -- the results are even more striking. In countries where people must check a box to agree to donate their organs, only 10-20 percent of people participate. When they must check a box to NOT donate, participation is 80-90 percent. Since it's not too hard to assume that people have similar tastes for donation, we can see that most of this difference arises from the default option.
Given the importance of retirement contributions, organ donation, etc., soft paternalism promises to bring a lot of benefits at low costs, and that's why we are going to see a lot more of it.
Are there any drawbacks to soft paternalism? I can only think of one: Say that the default has been x for many years, and I have "chosen" the default for many years. Now say that the default is changed to y. I may just "choose" y without looking at it. Put differently, I will pay less attention when I am looking at the default after 5-10 times than when I am looking at it for the first time. So, soft paternalism is a great idea for that first time I enroll in a retirement plan, but it needs to have many more flags if it is introduced to my retirement plan after 10 years...
Bottom Line: People are boundedly rational, and we have to be careful not to assume that they can process and choose among complicated (even simple!) options. Although we can improve things by changing the recommended option to an "expert's choice," we have to make sure that people know when such a choice has been made for them -- and then make it easy for them to change that choice.