The following similarity occurred to me yesterday, while I was trying to save the planet. "Saving the planet" seems to refer to an assortment of unsatisfying activities, which in my case consisted of attaching an address label to a spent printer cartridge (courtesy of HP Planet Partners) and sending it back to HP. The post office turned out to be closed by the time I reached it. 'Twould have been easier by far to discard the cartridge in the dustbin, like an unwanted and embarrassing corpse. As it is, I still have it.
Now, all these kind of minor environment-saving measures have one thing in common, which is that they do the individual no good at all; quite the contrary, they involve you in some amount of expense and inconvenience. It's remarkable that anyone bothers, but they're motivated by some sort of moral pressure that's hard to ignore. This brings me neatly to the topic of refereeing research papers, especially journal submissions. The motivation for this activity (yes, I'm in the middle of doing one at the moment) is entirely moral, again. You incur these tasks by submitting papers, so you ought to perform such tasks. Trouble is, there's no pay-off. Allegedly you get some respect from the editor, but this doesn't seem to translate into any kind of concrete advancement.
The reason why journal papers are worse than conference papers to referee, is firstly, that they are more work --- you really are supposed to check them in detail. Also, they are yesterday's news; the conference version usually came out about a year previously. Luckily, the one I'm in the middle of is well-written and on a topic that is highly relevant to my research, so I can't complain about it too much.
(added later:) I guess the reason why I find this interesting, is that it contradicts a standard premise of game theory, namely, rational behaviour (i.e. selfish behaviour). No doubt people have tried to model this kind of "irrational" behaviour, but I don't know how well that has worked.