因为一门课程的原因，我写了一段关于硕士论文（现在是一篇working paper)内容的简单陈述。目标就是让受过大学本科教育的人都能够看懂。所以请各位指正啦，帮忙看看哪里写得很confusing的~

Have you ever worked with classmates who never spend a second on the team assignment? Or have you lived with roommates who never clean the house, but instead always count on you to do the job? Those people are called “free riders”, for they consume more than their fair share of a public good (such as a clean kitchen which every house member can use). A big problem generates by free riders is loss of efficiency for the society as a whole. That is why economists, who value efficiency, try their best to find out the most effective mechanism in promoting cooperation and decrease free-riding problem.

One research tool economists refer to is lab experiment, in particular, an experiment called public goods game (PGG). In a public goods game, every player in a group secretly and simultaneously chooses how much of their own experimental money to put into the public pot. The group as a whole does best when everyone contributes all of their money into the public pot. However, any player has the incentive to under contribute. This experiment nicely capture the free rider problem in real life: We will all be better-off if everyone cooperates, but some may have the incentive to hold back their effort.

Researchers have run hundreds of such experiments and do found that cooperation gradually decays when people play the game over and over again. Two researchers, Fehr and Gaechter, discovered a very effective way in preventing the decreasing contribution trend. They added an extra stage, named “punishment”, on top of the public good game. Here is how their game works. Players first played the standard public goods game described above. Then the experimenter provides every group member a list of everyone’s contribution. Based on this information, players can decide whether and how much to reduce the earnings of certain group members from the game. At the same time, however, the punisher has to bear a cost – for in reality, the behavior of punishment is often costly. In the experiment, the punisher’s earnings is reduced together with the punished. Fehr and Gaechter showed that even when people only play the game once and this costly punishment does not generate any future benefits, many are still willing to punish free riders at their own cost.

However, in real life, people are heterogeneous so that their ability to contribute varies from one to another. Imagine that two students, Tom and Mike, are working together for a computer programming project. If they are equally capable, it is very easy for them to split the task. Now suppose Tom can do the task ten times as fast as Mike. Since their productivities very so significantly, they may have different opinions on the “appropriate” contribution level to the project. Mike might think that Tom should do the job since he is an expert, but Tom might think that it is totally unfair for Mike to free ride on him. It that case, it is unclear who should is the one to be blamed if the project fails to due on time.

In this paper with Charles Noussair, we aim at investigating how people decide on punishment mechanism in this kind of situation. In particular, we let players choose the punishment rules, and exam whether these rules selected remain efficient comparing with the previous literature.

The main features of our experiments are as follows. We randomly assign players into groups of four, and different roles. Two members of a group are “capable” players, and the rest two are “weak” players. Given the same amount of contribution, everyone benefits more from capable players. For every several periods, they can express their preferences of the following punishment rules: I allow “capable players” to be punished if they under contribute / I allow “weak players” to be punished if they under contribute/ I allow “capable players” to be punished if they over contribute / I allow “weak players” to be punished if they over contribute. If more members say “yes” than “no” vote to a certain item, this punishment rule will be executed for the following number of votes before the next voting is carried out.

The main result we find is that punishment rule chosen by a mixture of capable/ weak players are not as efficient as the rule chosen b a group of players with the same ability. The reason lies in the fact that players are more likely to vote for allowing punishment of another type rather than their own. This mirrors the reality in that institutional rules may not be efficient as they should be if it is voted out by heterogeneous people who have different belief of the contributions other society members should make.

Along the way you bump into people who make a dent on your life. Some people get struck by lightning. Some are born to sit by a river. Some have an ear for music. Some are artists. Some swim the English Channel. Some know buttons. Some know Shakespeare. Some are mothers. And some people can dance ...

You can be mad as a mad dog at the way things went; you can swear and curse the fates – but when it comes to the end, you have to let go.