We run a modified dictator game experiment to investigate the determinants of donation choices to philanthropic organizations. We find experimentally that the adoption of a simple form of accountability such as the disclosure of information on the ranking of aggregate contributions received bythe organizations has important redistributive effects on donations, leading donors to reallocate significantly their giving from top to bottom performers. Our findings support the hypothesis that individuals have preferences on total donations and their “ideal” distribution and not just on their own giving. Policy consequences of our findings in terms of public and private contribution disclosure rules arediscussed.