We extend the analysis of self-reporting schemes to criminal teams. When the violators behave non-cooperatively, maximum deterrence can be reached at virtually no cost by designing a prisoners' dilemma. One drawback of such a scheme is that it might induce cooperative behaviour in the self-reporting stage. If the cooperation rate is increasing the benefits from cooperation, it is optimal to impose less than the maximum fine if both individuals self-report. The same result occurs for imperfect self-reporting technologies where the conviction of one agent does not necessarily lead to a conviction of his accomplice.