Friday, December 22, 2006

Rational and Behavioral Reasons for Gift-Giving

Economists like to tweak normal people’s sensibilities by saying that gift-giving is inefficient, and that it would make more sense for everyone to give money instead (and even more sense to compare the dollar values of any two anticipated gifts and have the net giver hand over the difference to the net recipient). This article by economist Joel Waldfogel, brought to my attention by a former student, is a nice example because – in addition to making the standard theoretical argument – Waldfogel actually provides empirical evidence that people value the gifts they receive at less than their dollar value.

But as I’ve said before, even though I introduce this lesson to my students with the title “why my uncle the economist always gives me money for my birthday,” I do in fact give non-money gifts to my own friends and relatives. And the justification for doing so does not, as Waldfogel implies, necessarily require invoking behavioral economics and denying rational choice. The dollar value placed on the gift item itself is not comprehensive; it does not include all the other benefits of gift-giving. I listed some of the reasons last year; probably the most important is signaling to the recipient that you care enough to spend your time thinking about them. The trust-building that results can have great value independent of the actual gift’s value.

But even if we set aside signaling, there is also the enjoyment of unwrapping the gift on a special occasion, in the company of loved ones. And perhaps most importantly, there is the utility of the gift to the giver. You might give a book to a family member because you look forward to discussing it with him later, a game to friends because you hope to play it with them, a piece of jewelry or clothing to a romantic partner because you want to see her wearing it (or not wearing it), and so on. Such quasi-selfish benefits surely ought to be counted, and I suspect they could easily swamp the 20% loss in value cited by Waldfogel.

Finally, if we do go the behavioral econ route, there is yet another function of gift-giving cited by Alex (damn him for beating me to this point!): it is a device for indulging our fun-loving, spontaneous, immoderate sides. My sister actually chastised me for putting socks on my Christmas list this year because, she explained, the whole point is to get something fun and non-utilitarian. (In my defense, I did have fun stuff on the list; I was just trying to give people more options!) There’s a serious point here that goes beyond the theory of gift-giving. The (often paternalistic) behavioral econ literature typically focuses on the harms imposed by the short-run self upon the long-run self, as well as on the commitment devices employed by the long-run self to reign in those intemperate impulses: resolutions, gym memberships, diets, automatic savings plans, etc. But it’s also possible for the long-run self to impose harms on the short-run self – by creating guilt, curtailing spontaneity, and generally ruining the fun. To overcome the overweening nanny-self, people will sometimes precommit to pleasure. For instance, they will sometimes choose vacation packages and other luxuries over other cash prizes of equivalent value. Gift-giving performs a similar function: it allows recipients to have fun without guilt.

2 comments:

David Friedman said...

"I listed some of the reasons last year; probably the most important is signaling to the recipient that you care enough to spend your time thinking about them."

As stated, I don't think that one works. You could send the same signal by giving an amount of money equal to the cost of the gift plus the cost of the time you would have spent finding it.

To fix that you either need an argument--I have one--implying that it is easier to find a good gift for someone you care about, hence that it works as a signal, or you need some richer account of costs to the giver.

Suppose one believes in the short run/long run internal distinction. The short run me doesn't care about money, since the money spent on the gift reduces consumption in the long term, not immediately, assuming reasonable liquidity. The short run me does care about the time. So spending time signals that the short run me cares about the recipient.

This implied a testable prediction. Someone who is liquidity constrained, who can spend money only at the cost if immediate consumption, will be more likely to give money as gifts.

Glen Whitman said...

I agree, the signaling argument is only in shorthand here. If what we want from signaling is a separating equilibrium -- one that distinguishes those who *really* care about you from those who don't -- then the signal has to be less costly for the more sincere. This could be because someone who knows you has better information than someone who doesn't, or because someone who likes you actually enjoys the process of thinking about you.