The moral foreign-language effect (MFLE) suggests biases present when making moral decisions in the native language are not present in the foreign language. However, the literature using explicit dilemmas shows inconsistent findings. The present study investigates whether MFLE has its origin in the reduced emotion hypothesis. Instead of the typically employed explicit paradigms, we utilize an implicit paradigm, avoiding conscious processing. Chinese–English bilinguals completed an implicit association test (Experiment 1) and an evaluative priming task (Experiment 2) in their native (L1 Chinese) and second language (L2 English). Both experiments found consistent evidence that the self–other moral bias was only observed in the native language. Therefore, we propose that the MFLE has its origin during the automatic associative stage. It results from the reduced emotional reaction in a foreign compared to the native language.