Does Studying Economics Teach You to Lie?

new paper by Raúl López-Pérez and Eli Spiegelman investigates “truth preferences” — i.e., preferences for being honest versus lying. Their goal was to study whether economics students lie more as a result of their education. Or do liars self-select? From the paper:

Does studying economics give people “maximizing” habits of thought, and thus cause them to  behave more in line with its own predictions, or do people already inclined towards such behavior tend to self-select into economics?

A computer test structured with a slight incentive to lie was administered to 258 students at The Autonomous University of Madrid. The screen showed two colors, and participants were paid 14 euros for declaring blue and 15 euros for declaring green to another person, regardless of the actual color shown on screen. So what happened? According to the authors, the business and economics (“B&E”) majors gamed the system:

Our results show that the subjects most likely to exhibit honest behavior in our decision problem are the “B&E” non-B&E students, particularly if they expect this behavior from other subjects. In contrast, we do not find significant differences in honesty between males and females or between religious or non-religious people. In addition, an instrumental variable analysis suggests that the difference between B&E and other students is in part a matter of learning, and not only self-selection. Hence, our results are in accordance with other experiments showing that B&E students tend to conform more neatly with the homo economicus paradigm, or alternatively with a utilitarian or  consequentialistic mode of reasoning in moral matters.

But given the experimental setting and the student’s awareness that they were participating in an experiment, does the experiment really demonstrate a tendency to lie? Or does it perhaps show that economics students are maximizers who saw the experiment as a constraint problem where the instructions were to be ignored in favor of maximization?

William Lentzsch

Wondering how the distribution looks if you consider only the "profitable" or "non-profitable" lying (i.e., saying it was green when it was blue is profitable, and would make sense with a purely profit-seeking motive, whereas the reverse lie would actually cost the students)?

joseph gold

Why not increase the incentive to lie and see at which point people start lying? But if I were told I was going to receive 100 euro to say blue when I was clearly seeing green, well your results may get to vary.

Steve O

"But given the experimental setting and the student’s awareness that they were participating in an experiment, does the experiment really demonstrate a tendency to lie? Or does it perhaps show that economics students are maximizers who saw the experiment as a constraint problem where the instructions were to be ignored in favor of maximization?"

That was my first thought. College experiments are like Survivor--it's okay to forget about social conventions (i.e., honesty, fairness...) as long as you play the game well.

Johnny Disaster

It seems to me that the last sentence of this article is an illustration of the sort of mechanism by which B&E students are taught to lie.


Well that's interesting. So when experiments confirm the biases that you think of you declare them authoritative. But when it "disparages" economics/economics students (if that's the conclusion that can definitively be drawn) then you dismiss it as "only an experiment" or "a maximization problem." The idea of Freakonomics was to think about the hidden side of everything, the counter-intuitive conclusion not to dismiss research that concludes something that you don't like.


As I read the summary, I was thinking of just the objection stated in the final paragraph, but I am not an economist and I have no emotional or other investment in economists being good people.


This is cute and everything but this seems like a situation where the experimental setting is really limiting the how generalizable the results are (as noted in the article). Say the results didn't carry involve the Econ students maximizing, then you could plausibly claim that economics fails to improve how efficient economics students are.

Full disclosure, I'm probably a little sensitive to the charge as an econ grad student.


This study would be so much better if the rewards for the other player were the inverse so that lying resulted in cheating someone else and not just maximizing their own reward. Disclosure, yes I studied business.


They say the instrumental variable analysis showed that the difference between BandE students and other students is partly from learning. But what have they used as an instrument? I'm imagining they want an instrument for being an Econ student (which is endogenous since we cannot separate between the possible tendency to lie that econ students learn from the major, and self selection which might make liars choose econ), which means they need something correlated with becoming an econ major, but only affecting tendency to lie through its effect on being an econ major. What could this be?

Also, this experiment hardly proves that Econ majors are more inclined to lie. To make this experiment more effective, they should have had some kind of penalty for getting caught lying coinciding with the monetary incentive for lying. Without the penalty, there is no reason "lying" seems wrong to the student. There is no moral dilemma for the student to struggle with if the only outcome from the lie is that he/she gets paid more money. In fact, unless these students are not very smart, it seems like they would have a pretty easy time figuring out that the experimenters want them to say "green" more often since they have attached a higher price tag on it. So lying in this scenario might seem like the "right" thing to do. We talked about how there can be a bias in some surveys because participants don't answer correctly, they try to give the answer that they think the researcher wants to hear. In this case, the students own incentives line up perfectly with doing what they think the researcher really wants, which is for them to say green more often.



I would assume they used the time for which a student has been studying economics as an instrument, e.g. freshmen vs. seniors...if the effect is true no matter how long they've been studying economics they must've self selected and if you can't find the effect for freshmen but for seniors then this would indicate an aquired condition

paul o.

Green is blue and blue is green is not a lie. Banging your girlfriend in the Bahamas while telling your wife you're on a business trip in Rochester is a lie. Econ majors in general won't pull off this lie. Business majors can.

Eli Spiegelman

We’d like to thank Freakonomics for this publicity, and its readers for some stimulating discussion. There are several points we’d like to take a moment to clear up.

A common theme in the comments is that the test would have been stronger in some way if telling the “lie” (that is, reporting a blue circle as green) had some negative effect on the opponent, or if there was a penalty for getting caught. Yet this would introduce factors like altruism or risk aversion in the design. Since our research questions were (i) whether some people dislike lies per se (maybe because they have strong norms of honesty that forbid lies) and (ii) whether such lie-aversion correlates with socio-demographic variables, altruism and risk-aversion are only confounds. This explains why we eliminated them.

In more general terms, our question is essentially this: which subjects will maximize outcomes, and which will obey honesty rules? Thus, as one astute reader noticed, the argument that B&E students treat the decision as a maximization problem may be precisely the point that studying these subjects teaches you to behave (to some extent) like Homo Economicus. When the only reason not to tell a lie is because it’s a lie, B&E students are more likely to tell it.

Will B&E students tell more lies in more complex situations, for instance when lying might harm other people? We cannot address the question, since we don’t know about their degree of altruism. We can make two remarks, however. First, some previous experimental evidence suggests that B&E students are less cooperative and altruistic than other students (we review this evidence in the paper). Second, ceteris paribus, a person who feels less bad about lying for its own sake may also feel less bad about lying when it has negative consequences.

A related trend in the comments was that experimental settings may not be realistic enough to effectively measure the phenomenon we seek. We underline that once the restricted scope of that phenomenon is clear, it should also be apparent why laboratory control is useful. In almost any real-world context, lie aversion will be confounded by altruism, inequity aversion, reciprocity, guilt and shame coming from the possibility of getting caught, among myriad other factors. Again, the argument that “real” lies include these factors is a point for, not against, laboratory work in this study.

Note also that the benefit of lying is kept at a minimum in our design (just 1 Euro) in order to study how relevant lie-aversion is: If one person is willing to tell a lie to earn just one Euro more, one could argue that her/his honesty principles are not strong.

A common reaction to this paper is that the causality of studying economics could be measured by the year of studies. While this is an interesting idea, it seems to have some shortcomings. First, the selection effect may work not just at the beginning of studies, but each semester, as students have the opportunity to drop out or change programs for a better “philosophical fit.” Indeed, it may even become increasingly strong as their ethical reasoning matures and their understanding of the discipline deepens. In this case, it would still be present in the proposed measure: length of tenure in a B&E program measures the strength of your predisposition to maximization. Alternatively, it is possible that the “indoctrination” happens within the first year, so that even “young” students have already learned it. The instrument that we chose is a background variable, and thus does not seem to have these problems. Also, despite inevitable misgivings, we are unable to say that it is not exogenous to the kind of honesty we investigate. So what instrument did we use? We invite readers to follow the link above to find out…

Finally, one reader commented that our results could be explained by some demand effect, so that subjects are not really lie-averse, but try to give the answer that they think the researcher wants to hear. But if this is the case, why do we observe differences between E&B and the other students? Do B&E students think that the researcher wants them to lie, whereas other students think the opposite? This is an interesting conjecture, but does not seem to us more likely than the idea that B&E students have learned that not only can “the ends justify the means,” but that they are the only thing that can. As the reader points out, for those who accept this, there is no moral dilemma. A moral utilitarian will lie in our study, because it represents the greatest good for the greatest number (in euro terms). And our analysis corroborates the intuitive notion that this is an idea that is inculcated in B&E training.


Alexander Viko

What's the big deal?In economics, because all comes down to money making, essentially, lying has to be part of it. Does it not?