Brief 88: How the State Discourages Vigilantism: Evidence from a Field Experiment in South Africa
EGAP researcher: Anna Wilke
Key takeaways: A field experiment increasing police capacity to locate households boosted residents’ willingness to rely on police and reduced their willingness to participate in mob vigilantism. The evidence indicates the shift was driven less by residents expecting better police service and more by a heightened fear that police would identify and punish them for taking part in vigilante violence. This suggests that expanding state capacity can curb informal justice by making its costs more tangible rather than by making the state a more attractive alternative.
Geographical Region: South Africa
Type of study: Pre-registered field experiment (RCT)
Preparer: Danielle Bohonos
Executive Summary
This study asks whether increasing police capacity leads citizens to rely more on the state and less on mob vigilantism — the practice of citizens physically punishing suspected criminals — and, if so, why. Working with a South African NGO, the author randomly assigned households in a high-crime township to receive a home alarm linked directly to the local police station, making these households more “legible” and known to police. Using household surveys she finds that treated households became more willing to cooperate with police and less willing to personally participate in vigilante violence. Results point to a specific mechanism: the alarm mainly worked by raising residents’ perceived risk of being caught and punished for vigilantism, rather than by convincing them that police service quality had improved.
Policy Challenge
A large body of research argues that citizens cooperate with the state, rather than turning to informal alternatives such as gangs, or vigilante groups, when state institutions perform well and are perceived as trustworthy or procedurally fair. However, empirical evidence on whether strengthening the state actually crowds out informal alternatives is mixed: some studies find that greater state capacity weakens informal provision, while others find it can leave non-state actors just as strong, or even strengthen them.
One reason for this may be that citizens do not always see state and informal justice as substitutes: vigilante punishment is often harsher and faster than what courts deliver, and it bypasses due process protections that some citizens see as a weakness of formal justice. If citizens simply prefer vigilantism on its own terms, better police services alone may not be enough to displace it. A separate, less-examined possibility is that a stronger state discourages vigilantism not by out-competing it as a service, but by more credibly threatening to punish those who engage in it.
Context
South Africa has one of the highest rates of mob vigilantism in the world, despite having police capacity that is above average for Sub-Saharan Africa; a legacy of violent, distrustful citizen-police relations under Apartheid is often cited as part of the explanation. The study took place in a semi-urban, low-income township characterized by informal settlements, inconsistent house numbering, sparse street lighting, and few street names — conditions that make it difficult for police to locate households, especially in emergencies. Burglary is the most common crime in the area and a frequent trigger of vigilante violence. At baseline, more than half of residents believed police would take over two hours to respond to an emergency call, or would not come at all, and roughly one in seven residents said they would personally participate in beating a suspect caught stealing from a neighbor.
Research Design
The author partnered with a South African crime-prevention NGO to randomly assign 100 of 250 sampled households to receive an in-home police alarm, triggered by a panic button, motion sensor, or cellphone; the remaining households served as the control group. When triggered, the alarm sends the owner’s name, contact information, and location directly to the local police station and Community Policing Forum, and the same information is kept on file with police regardless of whether the alarm is ever used. The alarm was designed to increase both the reality and the perception that police could quickly locate and identify a given household.
Outcomes were measured through household surveys conducted one month and eight months after alarm installation, covering respondents’ willingness to rely on police, their willingness to personally participate in or support vigilantism, and their perceptions of police service quality and the risk of state punishment for vigilantism. To help distinguish mechanisms, Wilke embedded a separate information experiment: some respondents were read a news story about police efforts to convict vigilantism perpetrators, and others a story about police efforts to convict perpetrators of unrelated crimes. By comparing the two stories, she distinguishes the effects of service quality and punishment risk on vigilantism independently.
Results
Households assigned to the alarm became more willing to rely on police — both to report break-ins and to share crime-related information, which persisted through the eight-month survey. Willingness to personally participate in vigilantism fell as well, particularly among respondents who had started out most doubtful that police would punish vigilantism or deliver good service. Alarm recipients also came to see police as more likely to know their household and more motivated, more likely to believe police would imprison vigilantism perpetrators, and somewhat more positive about general police service quality.
The embedded information experiment helped separate the two possible mechanisms. The news story emphasizing the conviction of vigilantism perpetrators raised respondents’ perceived risk of punishment for vigilantism and reduced their willingness to participate in it by roughly 20 percent among initially skeptical respondents. Consistent with this, the alarm’s effect on vigilantism was significantly stronger when paired with the punishment-focused news story than with the service-quality story.
Lessons
This study suggests that expanding state capacity can meaningfully discourage participation in mob vigilantism, but that the mechanism policymakers should expect is deterrence rather than competition: residents did not abandon vigilantism primarily because they came to see police as a better service, but because a more legible, engaged police presence made them more afraid of being identified and punished for taking part. For policymakers designing interventions against vigilantism or other entrenched informal or illegal practices, this implies that investments explicitly signaling the state’s ability and willingness to detect and sanction the informal practice may be the more effective lever, particularly for the historically hardest-to-reach and most skeptical households.