There's an interesting paper in this book which looks at the effects of incentives and punishments on behaviour. The paper is called 'Incentives, punishment, behavior' if you're interested.
In one experiment Israeli schoolkids were given the job of trying to elicit charity donations from members of the public by going door-to-door (apparently 'donation days' are a regular thing in Israel). They were split into three groups, and within in each group divided up into pairs. The first group were given a motivational talk beforehand, and told that results of of the collection would be published so that the amount collected by each pair would be public knowledge. The second group were given the same information, but also told that they would be given a reward of 1% of value of the donations elicited (to be paid separately rather than taken out of the donations). The third group were also given the same information but also told they would receive a reward of 10% of the value of the donations.
So the results - the first group (no incentive) collected the most, the second (small incentive) collected substantially less, the third group (significant incentive) collected more but still some way behind the first group. It does seem to suggest that small incentives may crowd out intrinsic motivation.
One caveat - if we were to try to apply this to the Charity Muggers - is that I have read another bit of research (can't find it immediately) that suggests that intrinsic motivation does drop off if it's a mundane task. Which suggests that actually offering a decent incentive for longer-term campaigns is sensible in terms of keeping chuggers motivated. This doesn't take into account the impact that they have on the subjects of requests for donations stopped on busy streets and whether this affects their desire to make donations... :-)