PurposeThe goal of this study is to examine how community policing policies (CPP) can be effective in addressing racial disparities in police killings in the United States.Design/methodology/approachThe study utilized multi-level mixed modeling techniques.FindingsThe study finds that CPP training for in-service officers is effective when the police chief is black, in contrast to the presence of written CPP statements and CPP training for newly recruited officers. This article concludes that the effectiveness of policy implementation is dependent upon policing leaders who manage policy implementation.Research limitations/implicationsThis research is limited in that it only includes data from people who were killed by police. In addition, it was extremely difficult to collect data on the race of the officer. Hence, it reduced the number of viable cases that we could include in the analysis.Practical implicationsThe most significant practical limitation to our research is the ability to generalize to police departments within a city and between cities. In some cases, police killings were confined to one or two areas in a city.Social implicationsDisproportionality in police killings is important in every country where certain groups are overrepresented in the number of police killings. This is particularly true today, where we see groups like Black Lives Matter highlighting higher levels of lethal force in minority neighborhoods.Originality/valueUsing representative bureaucracy theory, this research shows leaders select and emphasize specific goals among a set of organizational goals, seek to build trust rather than fight crimes and support goals to improve policy outcomes, which fills a theoretical gap in the theory.