How do we evaluate border arrests?

The Border Patrol announced that the number of arrests made on the U.S.-Mexico border is down 8% from last year, citing better enforcement. Incidentally, this was reported now—one week before the election—instead of January, which is when it is normally released.

Analyzing the number of arrests is tricky, and in general it’s a pretty blunt instrument. For example, the number of arrests may be down since last year, but it is still up from 2003. In addition, as the article itself notes, yearly arrests are like a roller coaster, going up and down in ways that are not necessarily predictable, or that may be responding to short-term shocks (like 9/11).

In addition, arrests overall mask what is happening in specific border areas. Earlier in 2006, there were reports that arrests were increasing in New Mexico as well as in San Diego. What is particularly striking is that these increases are attributed to more enforcement in Arizona, and widespread reports of people dying in the Arizona desert. However, the original increase in Arizona and other more dangerous areas had been attributed to better enforcement in California in the past (on this, see Massey et al.).

In short, immigrants often respond to short-term stimuli, but clearly the enforcement in California has not deterred anyone from crossing there. Maybe this current dip in arrests is due to more workplace enforcement and the use of the National Guard, but the truth is that we really have almost no idea, and we also can’t predict how the numbers will change in the near future.