This dissertation explains the puzzling rise of vigilante violence in democratic Indonesia, despite relative success in management of larger security threats. It asks why states that are effective in stemming the tide of violence from terror groups and ethnic insurgents nevertheless display ambivalence towards daily acts of mob violence by ordinary citizens? Unlike prevailing theories that emphasize dissatisfaction with law enforcement as a motivation for citizens to take the law into their own hands, my research locates the causes of vigilantism in factors that allow citizens to get away with it. While public dissatisfaction with the law and its enforcement can motivate citizens to engage in vigilantism, their ability to do so is constrained by fears of legal prosecution and revenge attacks from the victims. I argue that endemic levels of vigilantism are only possible when vigilantes can find a credible way of reducing the uncertainty about the consequences of engaging in vigilantism; and act with a reasonable assurance of impunity. My research shows that vigilantism becomes endemic when street-level police officers can provide vigilantes impunity by withholding legal proceedings and thwarting revenge attacks from the victims’ side. Impunity provided by street-level police is selective and based on assessments about the risk of potential escalation of violence. Impunity is also revocable, as protection offered to the vigilantes may be retrospectively withdrawn if it leads to a public scandal or poses the risk of escalation into communal violence. The central claim of this dissertation is that impunity for vigilantism is generated as the result of state’s dependence on coercive functions of civil society: states allow latitude for extra-legal violence against local threats, in order to seek society’s cooperation in combating national menaces. Existing scholarship describes violence or its absence as a consequence of interaction between formal and informal sources of coercion. My research builds on these studies by examining how parallel realms of order and disorder are produced when grassroots civil society structures are mobilized to complement the state’s coercive power. Civic structures such as neighborhood associations and ethnic councils have long served as a way for the state to organize society and make it legible for purposes of control. My research shows that when faced with threats from insurgents, dissidents and terrorists, states also draw on these structures for fine-grained surveillance, electoral control and provision of manpower for security patrols. Over time, this reliance turns into institutional dependence as a) the state’s access to society is mediated through civic leaders and; b) the state’s own coercive apparatus is built around the expectation of predictable support from them. Within this context of institutional dependence, two changes generate impunity for vigilantism: a) loss of state’s political control over civic structures and; b) expansion of the state’s formal coercive apparatus. In authoritarian settings, compliance with coercive tasks is sought through high levels of political control over the leadership of these grassroots structures. When political liberalization diminishes this control, states manage the legacy of dependence by expanding the presence of their formal coercive apparatus. At the micro level, these changes incentivize street-level bureaucrats to provide selective impunity for extra-legal violence as a way to earn the goodwill of civic leaders that is necessary for performing high priority tasks. At the macro-level, this selective enforcement of the law by street-level bureaucrats cumulatively creates security trade-offs for the state: it has to allow society latitude for violence against local offenders, in order to keep its support for combating national threats. Thus, I claim that vigilantism flourishes not because the state is weak, but because its strength can be leveraged by vigilantes to protect them from the risks of engaging in violence. I support my claims with original quantitative data from Indonesia and qualitative evidence collected during 14 months of fieldwork. Analysis of 240,000 incidents of violence, including 33,000 cases of vigilantism shows that the rapid expansion of formal police presence in Indonesia is associated with higher levels of vigilantism. Studies of 20 specific cases, based on 186 interviews with perpetrators and victims of vigilantism, along with local law enforcement officials and community leaders, provide deep insight into the fears that govern vigilantes’ conduct and the concerns that shape the response of state agents.