The use of military forces for public security tasks has become a common approach to combat crime and insecurity. Advocates of these “mano dura” (iron fist) policies view them as necessary to control crime, while detractors claim they undermine human rights. In Cali, Colombia, researchers conducted a randomized evaluation of a military policing program called “Plan Fortaleza” to rigorously measure the impacts of military policing on crime rates and human rights. They found that the Plan Fortaleza program of intensive, recurring military patrols in crime hot spots significantly reduced the objective prevalence of crime, but only on days and times when soldiers were physically present on the streets. The program did not, however, reduce citizens’ reports of crime victimization, nor did it improve subjective perceptions of safety, except among business owners.
Latin America is the most violent region in the world. In 2018, the homicide rate, a standard indicator for measuring violence, was 22 people for every 100,000 inhabitants; this is equivalent to four times the world average. In addition, 39 percent of worldwide homicides occurred in the region, although the Latin American population represents only 9 percent of the global population. 41 out of the 50 most dangerous cities in the world are located in Latin America.
Insecurity and violence have led many to support “mano dura” (iron fist) policies. These policies generally include repressive measures against low-level criminals and minor crimes, the reduction or suspension of due process guarantees, and the use of military forces for security work. Advocates of these mano dura policies view them as necessary to control crime, while detractors claim they undermine human rights. Examples abound, from Mexico to Brazil to Guatemala. The strategy is not unique to Latin America. South Africa, for example, has relied on soldiers to reinforce police officers in the field for nearly two decades. Military policing is similarly common in Indonesia and the Philippines, among other countries. Despite the military’s increasingly prominent role in law enforcement around the developing world, empirical evidence on the efficacy of these policies remains scarce.
Cali is the third-largest city in Colombia and one of its most violent. In 2018, Cali reported a homicide rate of 47 people for every 100,000 inhabitants, nearly double the rate of Colombia’s second-largest city, Medellín, and more than triple the rate of the capital, Bogotá. In response, Cali’s Security and Justice Secretariat, housed within the Mayor’s Office, deployed the military to patrol crime hot spots as part of an initiative known as “Plan Fortaleza”.
Plan Fortaleza consisted of recurring, intensive vehicular and foot patrols by heavily armed soldiers from the Armed Forces of Colombia. Cali has 22 “comunas” in total; Plan Fortaleza focused on comunas 18 and 20, both hot spots for crime. The two comunas comprise 30 barrios (neighborhoods), which are further divided into “manzanas” (blocks). Their combined population was approximately 215,000 in 2018.
In Colombia, researchers conducted a randomized evaluation of the Plan Fortaleza program in Cali. 1,255 manzanas or blocks were randomly selected to receive military patrols roughly every six days; a comparison group received normal police presence. The average time spent patrolling was short—around 11 minutes per block per day—due in part to the small size of most blocks and the relatively large number of soldiers on patrol. While on patrol, soldiers searched residents for possession of drugs and weapons, checked IDs and business licenses, erected roadblocks, detained suspected criminals, and conversed with residents.
To maximize compliance, researchers hired civilian monitors to accompany each patrol. Monitors used GPS devices and smartphones equipped with a customized Google Maps interface to help soldiers navigate to their assigned treatment blocks. Further, to examine the impact of the intervention, researchers used data from the following sources:
- Administrative crime data from the Colombian National Police, the Mayor’s Office, the Attorney-General’s Office, and the Coroner’s Office, geocoded to the block level; and administrative data on human rights abuses from the Office of the Attorney-General.
- A household survey of 2,096 randomly selected residents of the two comunas between October 17 and December 19, 2019, beginning while the intervention was ongoing and continuing for roughly a month after it was over.
- A household survey of 7,921 randomly selected residents and business owners between January 17 and February 25, 2020, between two and three months after the end of the intervention.
- GPS data and geocoded, timestamped firsthand observations from the civilian monitors hired to accompany the soldiers while on patrol.
- Semi-structured, in-depth qualitative interviews with 49 civil society leaders from the two comunas, selected using snowball sampling. Interviews were conducted between June 26 and July 14, 2020, approximately seven months after the end of the intervention.
Overall, researchers found that Plan Fortaleza significantly reduced the objective prevalence of crime, but only on days and times when soldiers were physical present on the streets. Given that all patrols were conducted on weekday nights, researchers observed a 0.11 standard deviation reduction in crimes committed on weekdays on participating blocks, and a 0.17 standard deviation reduction on adjacent blocks while the intervention was ongoing. There was no evidence of crime deterrence on weekends or during the day. They also found no evidence that this reduction persisted after the intervention ended in November 2019.
Researchers found no evidence that the program reduced citizens’ reports of crime victimization either during or after the intervention. However, they found that the program increased citizens’ reports of witnessing crimes by 0.15 standard deviations on participating blocks and 0.19 standard deviations on adjacent blocks, and also increased citizens’ willingness to report the crimes they witnessed to the authorities. The researchers interpret these findings as evidence of increased vigilance on the part of civilian populations.
Plan Fortaleza improved subjective perceptions of safety among business owners on participating blocks, equivalent to an improvement of 0.28 standard deviations relative to comparison blocks. This improvement was not perceived by residents.
Despite the null effects on crime victimization and subjective perceptions of safety for residents, the intervention improved perceptions of the military and increased demand for military policing. Plan Fortaleza improved perceptions of the military in the final survey by 0.08 standard deviations on participating blocks. The program also increased demand for military policing by 191.2 percent in participating blocks and 115.3 percent in adjacent blocks.
Researchers found no evidence that the program increased support for retributive criminal justice policies. However, they found that the program increased support for military coups by 0.08 standard deviations on participating blocks and 0.07 standard deviations on adjacent blocks.
Finally, researchers found some weak evidence of increased human rights abuses by soldiers and stronger but still suggestive evidence of increased abuses by police officers. This finding is specific to the monitoring data and final surveys. Researchers found no evidence of increased abuses in administrative data from the Attorney-General’s Office, or in the firsthand observations of civilian monitors.