Research on violent crime has recently seen a renewed interest in the contribution of community social organization to patterns of offending. Specifically, the question of why do some socially organized neighborhoods exhibit relatively high crime rates? This paper develops and tests an alternative model which challenges the traditional social organizational and subcultural approaches or negotiated coexistence model and highlights the tension between dense social networks and informal social control orientations in the neighborhood-level regulation of crime. It is argued that while social networks may contribute to neighborhood collective effectiveness, they also provide a source of social capital for offenders. Competing theories considered are drawn from the systemic and cultural transmission perspectives. To test these theories, 1990 decennial census information with data from the 1994-95 Project on Human Development in Chicago Neighborhoods Community Survey and the 1991-93 and 1995-97 Chicago Homicide Data were combined. The tests of the three models offered strong and consistent support for the negotiated coexistence model. Thus, as the social network interaction and reciprocated exchange increase, the negative effect of collective effectiveness on violence declines in magnitude. Appendix and references