In a classic experimental economics paper, Marwell and Ames observed Economists Free Ride, Does Anyone Else?
The paper reports the results of a simple experiment: subjects are given tokens, which they can either put in an individual pool, which benefits only themselves, or a common pool, which benefits all participants in the experiment. The strategy that maximizes a player's personal return (in a one-shot game) is to put all of her tokens into the individual pool, because that way she is the only one enjoying the benefits from her investment.
The common pool, however, offers a higher overall rate of return, even though that return is split more ways. The best strategy for the group as a whole would be to cooperate, and put all of their tokens in the common pool. Yet, because the game is set up so that people cannot communicate or coordinate their actions, every one has an incentive to "free ride" – invest their own tokens in the individual pool, and hope to reap the benefits of other's generosity.
Marwell and Ames ran their experiments on the nearest subjects: students. They discovered a puzzling pattern. Economics students contributed less to the common pool, and kept more for themselves, than others. Marwell and Ames set out two possible explanations of their findings, selection and indoctrination:
Economists may be selected for their work by virtue of their preoccupation with the ‘rational’ allocation of money and goods. Or they may start behaving according to the general tenets of the
theories they study. Confronted with a situation where others may not
behave rationally, they nevertheless behave the way good economic theory
predicts.
The suggestion that we are special and different proved irresistable to economists, and Marwell and Ames' paper spawned others such as Frank, Gilovich and Regan's Does Studying Economics Inhibit Cooperation? and Carter and Irons' Are Economists Different and, If So, Why?
The relative importance of selection and indoctrination has never been definitively established, but there is at least some evidence that teaching students about free riding affects their behaviour. For example, Frank et al find that students' free riding increased more when exposed to a professor who "placed heavy emphasis on the prisoner's dilemma and related illustrations of how survival imperatives often militate against cooperation", and less when students were taught by an economics development specialist who devoted less time to game theory.
I've always found the idea that studying economics inhibits cooperation profoundly satisfying. It suggests that Econ1000 makes people see the world differently, and helps them to think like an economist. What people learn in Econ 1000 changes how they act and behave. Frank et al seem to interpret their findings in this same light, repeatedly using the phrase "training in economics," with the implication that something – a skill, an attitude, a way of thinking – has actually been taught.
Yet recent research in behavioural economics calls this comforting view of economics instruction into question. Dan Ariely reports that, "we’ve found repeatedly that people become more likely to lie and cheat after witnessing the dishonest behavior of others." Are the students who are repeatedly exposed to the prisoner's dilemma actually learning something about game theory and pay-offs, and acting accordingly? Or are they simply modelling their behaviour on the examples of non-cooperation that they have heard about in class, without really thinking about what they are doing, or why?
[I am not aware of any one who has tried to distinguish between these two explanations of the cooperation-inhibiting effects of economics, though it might be possible to do – just expose one group of subjects to an explanation of how to calculate payoffs and dominant strategies prior to the experiment, and the other group to stories about people cheating and defecting.]
Yet it matters. Years ago, I ran the Marwell and Ames experiment in the graduate public finance course. The student who earned the highest personal return – that is, who free rode to the greatest degree – was a newly arrived PhD student from China. Recently I was reminiscing with him about this incident. I said, "I was really impressed by the way you were able to work out the rational strategy. What was going through your mind?" His reply was something along the lines of: "Actually, I just didn't understand what we were supposed to be doing."
This is the question on my mind: what do people actually learn when they study economics?
Worthwhile Canadian Initiative 
Evidence from repeated public goods games suggests that subjects learn their way to the equilibrium over time. This is not inconsistent with the hypothesis that economists are (more) able to find the payoff maximising action through introspection, whilst other students require some experience in the novel environment before they learn how to maximise their own payoffs.
To separate out the differences in preferences between economists and other students you could use, for example, a dictator game experiment. This must have been done before, but I don’t currently have time to check (have to go supervise a midterm).
Evan – it has been – see here http://econ-new.stanford.edu/files/Theses/Theses_2005/Gross.pdf: In conclusion, this study found students with economics to offer less in both the Dictator and Ultimatum Games and to hold lower rejection rates in the Ultimatum Game. Thus, it may be argued that in offering and accepting less, individuals with economics hold a lesser concern for fairness (or a lower notion of what is fair). Additionally, economics training seems to lower both offers and acceptance thresholds and thus, conceptions of fairness overall. Lastly, as subjects were informed that they were randomly matched to another student, individuals should have assumed that they were most likely not paired with an economics major. It follows that non-economics majors exhibited a greater degree of “social intelligence” or “rationality” in being far more likely
to offer half and thus maximize expected value. On the contrary, more often than not, economics students, in generally offering less, played the Ultimatum Game as those they Gross, were playing against themselves, thereby becoming more likely to earn less when paired with an individual from the entire sample.
This, however, does not distinguish between the two pathways to non-cooperation: understanding the basic structure of the game and responding rationally versus thinking “well, other people usually cheat and so I will too.”
Thanks for the link.
There is a case to be made that the dictator game is so simple that everyone should understand the structure of the game.
But we still must ask: Can we ever measure true preferences in the lab? Do true preferences even exist? If they do exist, could it ever be possible to measure them?
I dunno … seems to me the more interesting experiment would be to allow the subjects to communicate and see how well they manage to coordinate (or not).
Evan, Patrick, I think would prefer to bracket out questions of the intrinsic value of experimental economics – there are different ways of setting up the games, and those tell you different things.
Evan, I think your point is that a more sophisticated understanding of game theory couldn’t explain the differences in a dictator game, because the dictator game is so simple that anyone can understand it. And so the reason economists offer less in a dictator game is – well, I’m not sure how you would answer that question. Sure, some is selection, but the rest? Is it the kind of effect that Dan Ariely finds in his research, that is, if you stand up and say “everyone cheats” then people will cheat? Or is there something else going on?
Frances, yes, that is my claim.
As to explaining why economists offer less in the dictator game, I’m not entirely sure what I think the answer is! It could be that economists simply have different preferences (with respect to fairness) than non-economists. Or it could be an Ariely-type effect, or that experimental demand effects are stronger for non-economists. Or it could be that economists have been trained to ignore framing effects and focus more strongly on the outcome than the process that caused the outcome.
“I’ve always found the idea that studying economics inhibits cooperation profoundly satisfying. It suggests that Econ1000 makes people see the world differently, and helps them to think like an economist. What people learn in Econ 1000 changes how they act and behave.”
I beg you to question your sense of satisfaction. A two person zero sum game has a value and a winning strategy for each player in the sense that that player can use that strategy to guarantee that her expected payoff will be at least as large as the value of the game. N-person games and non-zero-sum games do not have values and winning strategies in the same sense. Knowing an equilibrium of such a game does not in itself tell you how to play the game. If economics teaches students how to play such games so that they score worse than naive players, it does not do those students a favor. Claims of rationality are PR.
While I think that in general economics can in average attract more autistic people compared with population – which would explain different way economists play the game – I also think that it is their training in playing those games that gives them the “edge”
For instance any economics student who was exposed to Keynesian beauty contest game: http://en.wikipedia.org/wiki/Keynesian_beauty_contest knows the general dynamics of the game. He knows that the most first-timers will act with their rationality bounded to Level 0 or 1 with a few Level 2 and rare Level 3. When devising his strategy an economist has clear edge over other people not familiar with the game. Let the game run a few times and this edge will be lost.
And then there is a third thing that was shown in several behavioral studies and I actually think that this may be the strongest one. Frequent market exchanges tend to destroy cooperation. For instance if students in university are reminded all the time that they are actually competitors for the best grades and for the best jobs (as it seems to be at least in US), this sentiment may carry over to different aspects – even to games played during lessons. It was shown on several studies that frequent exposure to market exchange is detrimental to incentives for cooperation. There are studies that rich people tend to perform worse at reading faces and emotions of other people – mostly because they don’t need to if they satisfy their needs on the market. Poor people who have to rely more on their social skills to get what they require tend to be more perceptive of social signals of their counterparts. They are more concerned with their reputation and how other people thing about them as this is crucial for their overall well being.
J.V. Dubois: “Frequent market exchanges tend to destroy cooperation. ”
I came across that result while researching this post and found it interesting. I don’t know if it’s just competition, though. Advertising lies, and any child that has grown up in a capitalist society knows that people lie to you all of the time, because they’ve grown up with advertising. I also wonder how much modern capitalism removes our need to negotiate with people – just buy stuff at the supermarket rather than negotiate over its price – and that makes us worse at negotiation games.
Interesting post.
Is there any information on how 2 economists handle a one-time prisoner’s dilemma type situation?
Frances: How can you find it “comforting” that economics teaches people to free ride?
Kevin –
It’s a comforting thought when the alternative possibility is that economics teaches people nothing at all. What truly demoralizing is thinking that all the effort one is putting into teaching has absolutely no impact whatsoever. What makes me seriously uncomfortable is when a student does not do as well in the multiple choice portion of my exam as some random person who has never taken an economics course that I’ve persuaded to proof-read the exam questions.
Also: as an educator, I’m committed to communicating, as far as I am able, something approximating the truth.
You must be familiar with the studies that find religious people are, on average, more happy than non-religious people, and that conservatives are more happy, on average, than liberals? See, e.g. here. I’m not interested in manipulating my students’ values, even if it would make them happy.
Adrien – take a look at the papers cited here, the papers that cite them, and the papers they cite.
This mornings NEP-EXP mailing list from REPEC contains a paper relevant to the discussion. I haven’t read it yet, but a link is here: http://cahiersdugretha.u-bordeaux4.fr/2012/2012-25.pdf
This is in the abstract “Economics students
are, to a large extent, more prone to adopt greedy choices patterns than other social
sciences students.”
Frances: As for market incentives being adverse to cooperation – that is valid in case markets subvert non-market incentives. There are obvious examples, like paying your mother for a good dinner a going rate for a restaurant food acts contrary to some deeply encoded behavioral patterns and our image of ourselves as people with dignity and moral values.
As for “happiness” I would not read too much into it. The research shows that the personal feeling of happiness has homeostatic nature, it tends to return to a stable level after any sudden change (with rare exceptions, such as suffering permanent disability or as a result of drug abuse). What determines how high is that stable level of happines? It seems that 50% is pre-determined, the most important factor is your temperament, especially your level of neuroticism. And it seems that for a person to have low levels of neuroticism it is important to integrate attitudes, values, standards and the opinions of others into one’s own identity. This is closely tied to an overall self-esteem and people who have a strong sense of themselves as good and moral tend to have low level of neuroticism as they are able to internalize severe disorders such as depression, panic anxiety. In my book a married religious person fit this description quite well as it is with any secular person who has a set of moral values that she feels strongly about and who is stable in her psychological profile (less anger outbursts, less depression periods etc.) to be able to have healthy long-term relationship.
Sounds more likely that it would be an example of demand characteristics (See Wikipedia). If economics students are more likely to know that such an experiment would predict free riding, then we would expect them to free ride at a higher rate in an attempt to be “good participants”.