Why Do You Lie? The Perils of Self-Reporting

I am always surprised at how easily, and cheaply, we humans lie.

Have you ever been in a conversation about, say, a particular book and been tempted to say you’ve read it even though you haven’t?

I am guessing the answer is yes. But why would anyone bother to lie in such a low-stakes situation?

The book lie is what you might call a lie of reputation: you are concerned with what other people think of you. Of the many reasons that people lie, I have always thought that the lie of reputation is the most interesting — as opposed to a lie to gain advantage, to avoid trouble, to get out of an obligation, etc.

A new paper by the economists Cesar Martinelli and Susan W. Parker offers some fascinating insights into lies of reputation. It is called “Deception and Misreporting in a Social Program,” and will be published soon in the Journal of the European Economics Association.

Martinelli is a Peruvian-born, U.C.L.A.-educated economic theorist who teaches at I.T.A.M. in Mexico City; Parker is an American-born economist, educated at Yale, who also teaches in Mexico City, at C.I.D.E..

Their paper takes advantage of a remarkably rich data set from Oportunidades, a Mexican welfare program. It records the household goods that people say they have when they are applying for the program and then it also records the household goods that are actually found to be in that household once the recipient’s application has been accepted. Martinelli and Parker worked with data from more than 100,000 applicants, representing 10 percent of the applicants interviewed that year (2002).

It turned out that a lot of people underreported certain items that they thought might exclude them from getting benefits. Below is a list of underreported items followed by the percentage of recipients who owned a certain good but who said they didn’t:

Car (83.10 percent)
Truck (81.71)
Video recorder (79.73)
Satellite TV (73.91)
Gas boiler (73.12)
Phone (73.12)
Washing machine (53.46)

That’s not very surprising: you might expect people to lie to gain the advantage of a welfare benefit. But here’s the surprise. Below is a list of household items that were overreported — i.e., which applicants said they had but in fact did not (again, followed by percentages):

Toilet (39.07 percent)
Tap water (31.76)
Gas stove (28.56)
Concrete floor (25.41)
Refrigerator (12.05)

So 4 out of 10 applicants without a toilet said they had one. Why?

Martinelli and Parker chalk it up to embarrassment, plain and simple. People who were desperately poor were also apparently desperate to not admit to a welfare clerk that they lived without a toilet or running water or even a concrete floor. This is one of the most amazing lies of reputation I can imagine.

It should be noted that there is a lot of incentive to lie to get into the Oportunidades program, for the cash benefit equals about 25 percent of the average applicant household’s expenditures. Furthermore, the penalty for underreporting was not very strong: many of the people found to be underreporting goods like satellite TV’s and trucks were were not kicked out of the program. You could argue that the penalty for overreporting, meanwhile, was greater since it might mean being excluded from the program in the first place — which makes the overreporting even more costly.

The Martinelli-Parker paper may have broad implications for not only poverty programs but any kind of project where the data are self-reported. Think about the typical survey on drug use, sexual behavior, personal hygiene, voting preference, environmental behavior, etc.

Here’s what we once wrote, for instance, in an article about the lack of hand hygiene in hospitals:

In one Australian medical study, doctors self-reported their hand-washing rate at 73 percent, whereas when these same doctors were observed, their actual rate was a paltry 9 percent.

We’ve also written about the subjects that online daters are most likely to lie about, and the risky business of election polling — especially when the issue of race is involved.

But as often as we or anyone else writes about the perils of self-reporting, the Martinelli-Parker paper really gives the whole topic a foundation to stand on. Not only does it deliver a surprising insight into why we lie, but it is also a sobering reminder to naturally distrust self-reported data — at least until some scientists enable us to peer into one another’s minds and see what’s really going on there.

I am interested in hearing from readers what kinds of lies you tell, and why.

[Note: I'll be discussing this topic early tomorrow (Tues.) morning on the new public-radio program The Takeaway.]