Does the potential victim of dishonest behavior—a family or a bank, a pensioner or an insurance firm—affect the propensity to engage in such behavior? We investigate the effect of victim type—an individual person or an impersonal institution—on dishonest behavior and test whether it interacts with potential perpetrators' social value orientation (prosocial or proself). In a between-subjects design, we allowed experimental participants (N = 368) to misreport private information in order to increase (decrease) their profit (loss) at the expense of either another participant or the experimenter’s budget. Both prosocials and proselfs engaged in dishonesty, but proselfs did so much more. Furthermore, prosocials reduced their dishonesty when the victim was another person, rather than an institution, but proselfs did not. A direct implication is that the dishonesty of prosocials may be curbed by increasing the salience of the adverse effect their dishonesty has on other individual people but that such interventions will not be effective for proselfs. In contrast with recent results, we did not find a general effect of increased dishonesty under a loss (vs. gain) frame.