Immigration—especially across the southern border of the United States—is perhaps the dominant political, economic, and cultural issue of Donald Trump’s presidency. While economic incentives to migrate are important factors, widespread violence in migrants’ countries of origin also plays a significant role in explaining the surge in the demand for entry into the United States. In surveys of migrants from Central America’s Northern Triangle—Honduras, Guatemala, and El Salvador— almost 40 percent cited attacks or threats to themselves or family as the reason for leaving home. Violence pushed many to join the Central American migrant caravans that crossed Mexico in autumn 2018 on their way to the United States. This is hardly surprising as these three countries are among the most violent places in the world. In Honduras and El Salvador, homicide rates were at 64 and 109, respectively, per 100,000 in the year 2015, compared with an average rate of 1.2 in high-income OECD countries. In parallel with increasing violence in these countries, the number of refugees and asylum seekers from the three Northern Triangle countries has increased tenfold between 2011 and 2017. Many of the almost 200,000 unaccompanied minors from Central America apprehended in the United States between 2011 and 2016 were fleeing sustained rises in homicide rates in their communities of origin.

The case of the Northern Triangle may be extreme, but it is hardly unique. We demonstrate that a robust and substantively important link exists between violence and the demand for entry into the United States and argue that at least some of this relationship is driven by U.S. policies. Specifically, we provide evidence that deporting foreign-born individuals back to their homelands drives the demand for others to seek entry into the United States.

This paper makes three contributions to our existing understanding of the dynamics of population mobility. First, we focus on unregulated migration into the United States; that is, entry into the United States outside the formal visa system. This allows us to understand the demand for entry, which may exceed the number of visas allocated to residents of a particular country.

Second, we estimate how violence translates into demand for emigration. Since both violence and emigration are affected by a host of social, political, and economic variables that are sometimes hard to observe, we use the number of individuals deported from the United States to a particular country of origin as an exogenous determinant.

Third, we trace a vicious migration cycle whereby deportations from the United States lead to violence in the deportee’s home country; this, in turn, spurs subsequent demand for entry into the United States. This has immediate policy relevance, as one pillar of U.S. immigration policy since the mid-1990s has been a tough stance on immigrants who have committed criminal offenses while in the United States. Between 1996 and 2015, the United States deported 5.4 million individuals back to their homelands. Forty percent of these—2.2 million—had committed a felony while in the United States. It is important to note that unauthorized entry into the United States or overstaying one’s visa does not constitute a felony; these are misdemeanors. A criminal deportation hearing is initiated for a migrant—whether or not they are in the United States legally—if the migrant has committed a felony; not all migrant felons are subject to deportation hearings.

In cases such as Honduras and El Salvador, the cumulative number of deportees with a criminal background corresponds to between 1.3 and 1.5 percent of these countries’ total population measured in 2015. By deporting convicted felons, the United States returns home people likely to have developed connections with transnational organized crime upon incarceration in the United States. Moreover, the literature on parole in the United States points to the existence of a “revolving door” whereby criminal offenders who are incarcerated are more likely to commit crime once paroled than those who are sentenced to probation.

We test our argument using both data from 20 Latin American and Caribbean countries and subnational data from the case of El Salvador. In the 20-country analysis, we estimate a model of homicide rates in those countries as a function of deported felons from the United States. We also show that neither leading values of deported convicts nor the inflow of nonconvicts affects homicide rates at origin. Next we use predicted homicide rates to explain the demand for entry into the United States from Latin American and Caribbean countries.

Municipal-level data from El Salvador complements the 20-country analysis, focusing on the export of gangs from the United States as one specific mechanism of how the vicious migration cycle works. When deported convicts were sent back to El Salvador in large numbers, the gangs spread along migration corridors that had emerged when Salvadorans initially fled civil war in the 1980s. We use migrants’ exposure to violent crime as a predictor for gang-related violence using both administrative and survey data. Then we explain both migration intentions and actual emigration rates as a result of predicted gang-related violence.

We focus on the United States as a destination country as well as a source of deportations for two reasons. First, the United States is the largest destination for migrants in the world, with an estimated population of undocumented migrants approximating 11.2 million. Second, the United States is the country that deports the largest number of migrants or attempted migrants through legal means in the world. The countries in Latin America and the Caribbean are the largest sources of legal and undocumented migration in addition to the origin of most asylum seekers in the United States. Over our period of study from 2004 to 2015, Latin America and the Caribbean countries accounted for 98 percent of annual apprehensions at the U.S. southern border and 43 percent of annual asylum requests and received 97 percent of all deportations.

Our research bears important policy implications. First, the deportation of convicts has huge social, human, and economic costs in migrants’ countries of origin, largely borne by innocent people. U.S. deportation policies feed violence in countries that often lack adequate social and institutional mechanisms to control the spread of violence. The deportation of convicts also falls on fertile ground where deportees face strong social stigmas and limited economic opportunities. Depending on country contexts, the way deported felons relate to criminal networks at home can take different forms. A danger they all share is that they nourish and eventually transnationalize criminal activities and networks. Second, these policies may easily turn into a boomerang that creates new migration cycles. Hence, as a policy intended to disincentivize new migration, it is self-defeating.

NOTE:
This research brief is based on Christian Ambrosius and David Leblang, “Immigration Demand and the Boomerang of Deportation Policies,” August 14, 2019, https://​papers​.ssrn​.com/​s​o​l​3​/​p​a​p​e​r​s​.​c​f​m​?​a​b​s​t​r​a​c​t​_​i​d​=​3​4​91522.