In a tax compliance experiment with real face-to-face communication between declaring subjects and officers, we analyse the role of both the subject's and the officer's gender for deceptive behaviour. We do not find, first, that the amount of underreporting generally depends on the officer's...