Documenting the Undocumented

How do you count unauthorized immigrants?

Immigrant activists participate in a rally on immigration reform in front of the White House on November 8, 2012 in Washington, DC.
Activists rally for immigration reform in front of the White House on Nov. 8, 2012.

Mark Wilson/Getty Images

President Obama outlined his vision for immigration reform in a speech on Tuesday. The president described his plan as a “comprehensive approach that finally deals with the 11 million undocumented immigrants” in the United States. If these people are undocumented, how do we know there are 11 million of them?

Subtraction. The Pew Hispanic Center produces the most-cited estimate of the nation’s unauthorized immigrant population. The methodology is complicated, but it’s based primarily on two data sources. Every March, the federal government surveys between 75,000 and 80,000 households, looking for foreign-born people. That number, scaled up and adjusted for survey biases, establishes the total number of both authorized and unauthorized immigrants living in the country. The researchers at Pew then use decades of data from the Department of Homeland Security, which tracks the total number of immigrants admitted legally to the United States, to estimate the authorized population. Once they have those two numbers, it’s a matter of subtracting the documented immigrants from the total foreign-born population, leaving the undocumented immigrants. The most recent survey suggests that the unauthorized immigrant population stands at 11.1 million people, which is statistically indistinguishable from last year’s count of 11.2 million and significantly down from the 2007 peak of 12 million.

There are some statistical complications, for those truly interested in demographic methodology. One problem is that the government’s annual survey systematically undercounts immigrants, especially those who are recently arrived. Another is that there is a significant foreign-born population that doesn’t fit neatly into the “authorized” or “unauthorized” categories, such as applicants for asylum and people authorized to stay temporarily in the United States during periods of strife in their home countries. There are also so-called “sojourners,” who pass briefly through the country without authorization and are extremely difficult to track. By some estimates, more than one million people may belong to these in-between groups.

Nevertheless, there are indications that the Pew figures are accurate. For example, surveys of households in Mexico suggest there should be around 6.5 million undocumented Mexican nationals in the United States, which is approximately the same number that Pew’s research suggests from that country. (Mexicans account for 58 percent of undocumented immigrants.) Academics who have estimated the undocumented immigrant population the easy way—by simply asking people—have come up with broadly compatible numbers for specific geographic areas. (Reliable nationwide surveys of this kind have never been conducted.) In addition, the number of people who applied for citizenship after the last mass legalization roughly matched the figures that Pew’s method predicted.

Only in recent years has the Pew data achieved general acceptance, giving policy makers and politicians a common number to argue over. As recently as eight years ago, there were widely varying estimates. A 2005 analysis by the now-defunct brokerage firm Bear Sterns claimed that there were more than 20 million undocumented immigrants in the United States. Today, most demographers believe this estimate, based on a surge in apprehensions along the Mexican border, was badly off, because the analysts failed to consider that many arrestees were repeat offenders. Some pundits, such as the arch anti-illegal immigration commentator Lou Dobbs, used the false estimate repeatedly.

Got a question about today’s news? Ask the Explainer.