http://www.econ.cudenver.edu/beckman/Tiffany/cherry%20-hardnose-aer.pdf

When you raise the stakes in the ultimatum game (say to tokens that can save your life), there are two effects on the proposer’s incentives: First, the responder is a lot less likely to veto a bad offer, so you can afford to be less generous. But second, a veto is now a lot more devastating, so you might want to be *more* generous just to be safe. Vernon Smith once gave me an argument for why you’d expect these two effects to be of roughly the same size, so that the offered splits (as a percentage of the total pie) don’t change. But I’m not able to recall what his argument was.

Without getting into specifics, I think it is a perfectly reasonable moral code to say that you’re willing to be paid for doing useful work, but not to take other people’s money by force in exchange for nothing at all.

To get into one specific, Landsburg thinks it is a perfectly reasonable moral code to have people BE FORCED to pay him to do useful work – that is, where most of the people doing the paying are not also the people evaluating whether Landsburg’s work is useful, or whether it’s worth what he’s paid.

That seems like a reasonable point of distinction to the Ultimatum Game. I still question whether the outcomes of the Ultimatum Game can be explained in terms of people’s aversion to accepting tax dollars without doing useful work.

they might give relatively even offers

fixed my post

Couple more points. Having actually had a look at the Wikipedia article, it seems that some committee did think it worth exploring high value amounts – not quite $1,000,000. They used $100, but in Indonesia where it represented about 2 months wages. Rejections were reportedly independent of the stakes, indicating that the effect is not only valid for small amounts.

If I ever became a mass murderer, I would make people play the ultimatum game for their lives. The proposer divides 100 tokens between him/herself and the responder. If the responder accepts, for each player a random number generator (I personally like the idea of a “Wheel of Fortune” for its tension) produces a number between 1 and 100. If this number is greater than the number of tokens that player has, the player is shot in the head. If the responder rejects the offer, both players are shot in the head.

My hypothesis is that few players would reject say, a 10% chance of living to kill someone who offered them an uneven share. On the other hand, I think proposers may be so scared of getting rejected that they might give relatively uneven offers. Finally, players may question my credibility to not kill them all anyway and defiantly play differently than if they were trustworthy.

Unfortunately it is unlikely that my proposal would pass an ethics committee. I do hope that if mass murder is inevitable, at least one perpetrator will take up the opportunity to perform valuable research.

Now, not everyone is as moralistic as I am on this point (as I am not as moralistic as some others on some other points), but most people are not psychopaths, so most people are going to feel at least a little squeamish about taking money nonconsensually from others.

I assume the University of Rochester is a tax-exempt educational institution, receiving all kinds of public subsidies in the form of government services, students with guaranteed student loans, etc. As a research institution, I expect it receives scads of government grants. I’d guess that particiating in a $10 Ultimatum Game would represent a trivial burden upon taxpayers relative to Landsburg’s professional activities.

How widespread is Landsburg’s aversion to benefiting from taxpayer-financed largesse? Everything I’ve ever heard would suggest that this feeling is not widely held. Indeed, all of political science would seem to be premised on the opposite conclusion. I’d be curious to know how many of Landsburg’s students and fellow faculty members share the view that it’s wrongful to seek or accept government scholarships or grants. I hope Landsburg can expand on this idea.

Huh? Could you please fill in the missing steps from Steve’s post to your characterization of it?

Fake Name:

We could run your experiment with the variable: does responder know that the sum is compensation?

Steve: “This is a great idea. Someone should run this experiment.”

I’m right on it. As soon as I get government funding. :-)