The number of immigrants in the U.S. illegally fell to 11 million since 2009, largely because of a drop-off in the number of Mexicans without legal status, according to a study released Tuesday.

The report by the nonpartisan Pew Research Center — using survey data from 2015 — showed the number of immigrants lacking legal status was 11.3 million in 2009. The number of Mexicans in the country illegally dropped to about 5.6 million from 6.4 million during the same six-year period.

“The numbers are not going up and, in fact, the numbers for Mexicans have been going down for almost a decade now,” said Jeffrey Passel, senior demographer at Pew Research Center in Washington, D.C. “And that is counter to a lot of the rhetoric you hear.”

Pew didn’t give a reason for the decline. But in earlier reports, it said the U.S. economy was slow to recover from the recession and border enforcement got stricter around that time.

The report is based on data taken by the U.S. Census Bureau during the latter years of the Obama presidency. But it comes amid the Trump administration’s efforts to deport immigrants in the country illegally and build a wall on the U.S.-Mexico border.

Pew did not provide a forecast for this population since Trump took office.

The report shows illegal immigration climbed during the 1990s and into the 2000s and peaked before the recession. Since then, the number of Mexicans in the country illegally has fallen while the number of Asian and Central American immigrants has grown.

The number of Central Americans in the country illegally was 1.8 million in 2015, up from 1.6 million six years earlier, while the number of Asian immigrants without legal status rose to 1.5 million from 1.3 million, the report showed.

In 2014, the number of immigrants in the country illegally was 11.1 million, with Mexicans accounting for about 5.9 million of that population.

Pew also released a preliminary estimate for immigrants in the country illegally in 2016 of 11.3 million. The center says the estimate is not statistically different from the 2015 figure because it stems from a separate data set with a larger margin of error.
…