A rare controlled experiment to test the effectiveness of unarmed civilian approaches to disorder and crime found in Medellín, Colombia, that assigning full-time liaisons to public safety tasks had no impact on crime and government legitimacy unless used in neighborhoods where the government already had a strong, trusted presence. "The divergent results suggest the importance of existing state capacity, plus the dangers of over-promising and under-delivering," researchers concluded in their new paper based on a 2018 randomized control trial. Lead researcher Christopher Blattman, of the University of Chicago, wrote in a Twitter thread on the study that it is believed to be the first community-level RCT of a non-police public safety intervention in the world.
The two-year study started in 2018 to track crime and emergency calls, supplemented by surveys of 2,500 people about their faith in state services. In half of the 80 city sectors chosen, full-time liaisons were assigned to help community groups organize, connect people to city services, refer disputes to resolution officers family services, and serve as a bridge to local police. This intensified, civilian-led problem-solving effort had no effects on crime, numbers of emergency calls, or legitimacy. By sorting results depending on the preexisting quality of neighborhoods' governance, researchers saw the outcome when city leaders didn't show up to meetings or respond to calls for services: a "stark divergence" with neighborhoods that could count on a government response to the civilian organizing. "In initially well-governed sectors, state legitimacy rose about 10%, reported crime fell by 28%, and emergency calls to police dropped 55%," Blattman wrote. "Communities seemed to be solving a lot of these problems on their own now." The study should not be read to say "intensifying non-police street staff 60-fold didn't work," Blattman wrote. Instead, the study shows "what probably matters is common sense: 'do not over promise and under deliver.'"