Aaron J. Chalfin
Since 1980, the share of the US population that is foreign born has doubled, rising from just over 6% in 1980 to over 12% in 2010. Compounding this demographic shift, the share of the foreign born population of Mexican origin also doubled, leading to a quadrupling of the fraction of US residents who are immigrants from Mexico. A majority of recent immigrants of Mexican origin living in the United states are thought to be undocumented, leading to a contentious policy debate concerning the collateral consequences of this particular type of immigration. Indeed a narrow majority of US natives indicate a belief that Mexican immigration increases crime in the United States.
It stands to reason that immigration is neither automatically good nor bad. Depending on who the migrants are, it is easy to see how immigration could either increase or decrease US crime rates. Every country has had a different experience with immigration and sometimes the experience varies with immigrants' country of origin. So what has our experience been with Mexican immigration?
Theory offers little guidance in sorting out the effect of either Mexican immigration specifically or immigration, more generally, on crime. On the one hand, Mexican immigrants, on average, possess demographic chracteristics, which, in the US native population, appear to be positively associated with crime. In particular, they are more likely to be young and male and have considerably lower earnings than other US residents. On the other hand, there is a multitude of evidence that immigrants are positively selected on the basis of ability, face higher costs of punishment and bring with them abundant social capital that is protective of participation in crime.
From an empirical standpoint, the relationship between immigration and crime is difficult to disentangle. Data on the nationality of state prisoners are unreliable and municipal police departments do not systematically collect information on the birthplace of individuals they arrest. Another strategy is to see what happens to crime in a US city when more immigrants arrive. This too is difficult because the timing of immigration to the US is hardly random. Migrants come to take advantge of job opportunities and it could well be the case that migrants arrive in the United States when crime markets are in a state of flux for reasons that have nothing to do with the immigrants themselves.
If we could transform the United States into a research laboratory, what we would like to be able to do is assign each US city a different number of migrants in each year and then observe what happens to that city's crime rate in the following years. The idea is motivated by the concept of the randomized controlled trial in medicine - the idea that we can draw inferences about the efficacy of a treatment or a drug by randomly assigning patients to either take the drug or take a placebo pill.
Over the past decade researchers have identified several natural experiments, which can be used to understand the effect of increases in Mexican immigration on crime. By leveraging unanticipated weather shocks in Mexico which wreak havoc on local agriculture and thus spur migration to the United States and taking advantage of the considerable regional variation in Mexico's demographic transition decades ago, scholars have identified conditions under which the receipt of Mexican immigrants in US cities appears to be arbitrary - that is, not a function of local economic or crime conditions.
The results of this research offer little evidence that Mexican immigration increases crime in the United States. If anything, there is some evidence that crime declines after immigrants arrive. These findings are supported by research from the Public Policy Institute of California on the composition of inmates in California prisons, which reveals that Mexican immigrants are dramatically underrepresented in the state prison system.
For those who are skeptical that these findings are true, consider the case of El Paso, Texas a working class city of approximately 700,000 people that sits opposite the Rio Grande river from Ciudad Juarez, one of the most violent and lawless cities in Mexico. More than 80% of El Paso's residents are Hispanic and the vast majority of these individuals are of Mexican origin. A large population of El Paso's Hispanic population are immigrants. In fact, El Paso has one of the highest proportions of immigrants among U.S. cities. Many of these migrants are undocumented. If those who fear Mexican immigration are right, then El Paso should be a hotbed of violence. As it turns out, El Paso is one of the safest cities in the United States with a homicide rate of 2.4 per 100,000 residents. Just a tiny handful of American cities have a lower homicide rate and most of those that do (San Diego, Chula Vista, and Mesa, AZ, for example) also have outsize Mexican populations. Incredibly El Paso's homicide rate is so low that it compares favorably to European capitals like London, Paris and Amsterdam, cities which have rates of lethal violence that are generally an order of magnitude lower than cities in the United States.
To summarize, while it is right for Americans to be concerned about the equality of the immigrants we attract, there is simply no evidence to support that Mexican immigration should be a cause for concern. If anything, there is quite a bit of evidence that the immigrants we attract from Mexico serve to make us safer than we otherwise would be.
Butcher, Kristin f. and Anne Morrison Piehl (2008). "Crime, Corrections, and California: What Does Immigration Have to Do With It?" Public Policy Institute of California Population Trends and Profiles, Volume 9, Number 3.
Chaflin, Aaron (2015). "The Long-run Effect of Mexican Immigrants on Crime in U.S. Cities: Evidence from Variation in Mexican Fertility Rates, "American Economic Review 105(5), p. 220-225.
MacDonald, John and Jessica Saunders (2012). "Are Immigrant Youth Less Violent? Specifying the Reasons and Mechanisms," The Annals of the American Academy of Political and Social Science 641(1): p. 125-147.