Abstract In the vast majority of laboratory experiments documenting the existence of reciprocity subjects are endowed with windfall funds. In many environments with salient fairness considerations such endowments are known to inflate subjects’ other-regarding behavior, thereby creating a so-called “house money effect.” This suggests that laboratory experiments might also overestimate reciprocal behavior. In this study we identify two reasons why the source of endowment might matter for negative reciprocity: (1) Using earned – as opposed to windfall money – might increase the costs of negative reciprocity due to this money being in a different mental account and therefore lead to less retaliation. (2) Appropriating some of the decision-maker’s endowment consisting of earned money might be considered a stronger violation of property rights and lead to more retaliation. While we find experimental support for the latter conjecture, we also observe that subjects actually retaliate more with their earned money than with windfall money as long as at least a part of their endowment is earned. However, conditional on earning a part of their endowment, subjects do not seem to distinguish between situations when they retaliate using earned money versus using windfall, suggesting that their main motivation is the violation of property rights established by performing a real-effort task. Our results thus point out that endowing subjects with windfall funds, absent of clearly established property rights, diminishes their negatively reciprocal responses.