We have examined how much money participants take for themselves from an amount designated either for a well-known charity or for a state's public budget. For a third of the participants, the decision was real – they were paid the chosen amount afterward, and the donation to a charity or public budget was lowered by this amount. For the rest, the decision was hypothetical, with no consequences. In a follow-up study, a different sample of participants was tasked with an estimation of behavior in both conditions. As expected, participants took more money from the public budget than the charity. However, when the decision was hypothetical, they took less money only from the public budget. Participants who could take money from the charity did not take less in the hypothetical than in the real condition. This was unexpected even for participants in the follow-up study, who significantly underestimated the amount of money taken from charities in the hypothetical condition. The results highlight limited generalizability of findings regarding moral and pro-social choices that use only hypothetical scenarios and suggest that interactions between positive self-presentation and monetary motives may be more complex than expected.