The ability of humans to form functioning and cooperative groups with completely unrelated individuals is something of a riddle, and has been the focus of much recent research in fields from economics to genetics. Some of this research has even made its way into the mainstream media in recent months, such as the "dictator gene" story or the different behavior of freeloaders in various countries. One branch of this research looks into altruistic punishment - the willingness of people to pay a personal cost to punish cheaters, freeloaders and others who make cooperation more difficult. One very interesting leaf of that branch looked at the difference between three different configurations of altrustic punishment - one in which no one is able to punish anyone else, one in which everyone is allowed to punish freeloaders, one in which there is a single designated punisher.

The no-punishment group performed badly, as should be expected, as freeloading ran rampant. The main finding of the research is that the one-punisher model works as well as the all-punisher model at preventing freeloading, and results in greater profits because the punishment is meted out with less waste and more efficiency. Of course, games designed by economists are not always going to give us an accurate picture of how things would really work with large numbers of people interacting over a span of years or decades. Also note that these punishers were given their power by random choice, and were not necessarily those most suited to the task or most able to rise to power.

With those caveats, what this paper does demonstrate is that concentrating power in the hands of a single individual clearly does have benefits, and suggests that we may have developed genetic and/or cultural inclinations to organize ourselves in ways which take advantage of them. I am reminded of the people who've told me that the most fun they've ever had playing music was in military bands because there was a single powerful authority and all decisions were made at the top - no time and energy were wasted in arguing.

The authors suggest some avenues for possible follow-up studies. The one which looks most interesting to me is repeating this experiment in several different cultures. It would be very interesting to see the difference between high-trust and low-trust societies.

I would not use this paper to argue that dictatorship works better than democracy unless I were arguing with someone who is not smart enough to bring up the objections I mentioned earlier. I would use it against any attempt to claim that all human beings carry in their hearts a yearning for democracy, though.