The paper inquires into the impact of mistakes of identity (ID errors) on the optimalstandard of proof. A mistake of identity is defined as an error such that an individualis punished for someone else’s crime; and for the same crime, the criminal is falselyacquitted. Therefore, the decision to engage in a criminal activity generates a negativeexternality, as the expected number of ID errors increases. Thus, our objective is tounderstand how public law enforcement can deal with this type of error by means ofthe standard of proof. Our main results are twofold. First, we show that when ID errorsoccur, the under-deterrence issue is exacerbated. Second, we find that the optimalstandard of proof may be higher or lower than without ID errors, depending on thecrime rate at equilibrium and on the impact of the standard of proof on (i) the probabilityof an acquittal error for each crime committed, (ii) the probability of convictingan innocent person when an acquittal error arises, and (iii) the level of deterrence.