We use Swedish adoption data combined with police register data to study parent-offspring associations in crime. For adopted children born in Sweden, we have access to the criminal records of both the adopting and the biological parents. This allows us to assess the relative importance of pre-birth factors (genes, prenatal environment and perinatal conditions) and post- birth factors for generating parent-offspring associations. We first estimate a simple linear model. We find that both pre-birth and post-birth factors are important determinants of offspring convictions. Mothers contribute approximately equally through both pre-birth and post-birth factors, while fathers contribute mainly through pre-birth factors (genetics). For sons, biological mothers contribute more than biological fathers. This difference may be due to the role played by prenatal environment and perinatal conditions, which work solely through the biological mother. We find no evidence of interaction effects between biological and adoptive parents’ criminal convictions and limited evidence of a gene-environment interaction between adoptive parent income and biological parent criminality.
The full paper by Hjalmarsson & Lindquist is here (2010, published in Labor Economics, November 2012). The title is "The Origins of Intergenerational Associations in Crime: Lessons from Swedish Adoption Data."
Bryan Caplan in his book "Selfish Reasons to Have More Kids," also explores the evidence on the link between genes and criminal behavior. In the long run he finds that what matters most is genes, and much less so the environment.