New research indicates that people may be more likely to lie when a charity benefits from their dishonesty. A group of researchers led by Alan Lewis at University of Bath investigated this in their paper "Drawing the line somewhere: An experimental study of moral compromise" (ungated here). From the paper's abstract:
In a study by Shalvi, Dana, Handgraaf, and De Dreu (2011) it was convincingly demonstrated that psychologically, the distinction between right and wrong is not discrete, rather it is a continuous distribution of relative ‘rightness’ and ‘wrongness’. Using the ‘die-under-the-cup’ paradigm participants over-reported high numbers on the roll of a die when there were financial incentives to do so and no chance of detection for lying. Participants generally did not maximise income, instead making moral compromises.