Mark Thoma asks. I answer.

Mark Thoma notes

Most people know about the Earned Income Tax Credit (EITC). Under this program, payments to qualifying individuals are made once a year. There is also something called the Advance Earned Income Tax Credit (AEITC) that allows qualifying individuals to receive the credit with each paycheck. But even though this program exists, most of the payments are made under the EITC and come around the time tax returns are filed.

Quotes Andrew Goodman-Bacon and Leslie McGranahan who argue that this is a good result.

and asks

In some cases, there are market failure arguments that provide the foundation to force people to participate in particular programs (e.g. adverse selection in health care or insurance for drivers), and there are arguments that can be made here, but the particular argument ought to be made explicit. Why is it better to force people to save (we do this with Social Security)? Unless there's some good reason for the government to step in and make choices for people, I'd rather not have the government get in the habit of thinking it knows better than I do what is good for me:

I reply

There is a justfication for forced savings which is not at all based on market failure. It is based on dynamically inconsistent preferences. If people discount future rewards with any function other than the exponential, they may wish to deprive their near future selves of freedom of action in order to protect their more distant future selves. In particular, if a two period discount factor is greater than the square of a one period discount factor, people will want to force their one period future selves to save.

Given standard assumptions including individual rationality and dynamically consistent preferences, a public intervention is desirable only to deal with a market failure or to redistribute from the rich to the poor. The assumptions are standard not because they are plausible but because they make model building much easier. It is possible to gain some insight on whether people have dynamically inconsistent preferences by asking them their discount factors. It is not clear that people really have the answer to that question in their minds, but they do answer the question and generally on average claim to have dynamically inconsistent preferences.

Another way to test would be, say, to give people a choice between the EITC and the AEITC. If people have dynamically inconsistent preferences, they may rationally chose the EITC over the AEITC. Thus the fact that they do so is evidence (not proof given the red tape but evidence) that people are right when they claim to have dynamically inconsistent preferences.

Now, forcing people to save might be paternalistic, but giving people the option to force their future selves to save can't be. A program where people can choose between the EITC and the AEITC gives them more freedom and is less paternalistic than one in which they are forced to accept the AEITC.

More generally, the basic principle that we should have laissez faire (or laissez faire with redistribution) unless we can point to a market failure is based on theory which, in turn, is based on making assumptions that lead to nice simple results like ... we should have laissez faire (or laissez faire with redistribution) unless we can point to a market failure.