Root causes of illegal immigration

Whenever I hear someone use technical Quality terminology in an unexpected venue, my ears perk up. So I sat up sharp a few days ago, when I was reading an article about the upcoming election and it mentioned that Vice-President Harris was given an assignment back in 2021 to address "the root causes that drive migration to the United States." [Emphasis mine.] Really? I thought. We tackled the root causes of something? Cool. I wonder how that turned out? 

[You might guess that I don't always keep up with the news.]

Before I go on, let me be clear. This is not a post about immigration, nor about the election! My interest here is in the application of root-cause analysis as a problem-solving tool, because I think this story highlights a useful point. As for the political questions … you don't need my opinions because you've got your own, and yours are probably better. I'll keep mine to myself.

In some ways we should expect it to be very difficult to identify root causes for a large-scale social phenomenon. The classic understanding of a root cause is that when you turn it on and off, the effect turns on and off as well—like a light switch. But social groups are systems in the technical sense, and systems involve delays. So even if you find a root cause and switch it off, it might take a while for the effect to disappear. 

On the other hand, it is tempting to look for root causes of social phenomena, for the same reason that we look for the root causes of more tractable problems. If you solve a root cause, the problem goes away and you don't have to keep fixing it. This is a powerful inducement. 

So what was done? 

On February 2, 2021, President Biden signed an Executive Order that called for the development of a Root Causes Strategy. (You can find the full document here.) A month later, he asked Vice President Harris to lead this effort diplomatically and organizationally. The analysis did not start from zero, but built on earlier analyses carried out under the Obama administration. (See references in, for example, here and here.)

What root causes were identified in this analysis? 

A Congressional Research Service report by Peter J. Meyer, "Central American Migration: Root Causes and U.S. Policy," acknowledges that "motives [for migration] vary by individual," but identifies four broad categories of root cause: socioeconomic conditions, natural disasters, security conditions (this means crime and violence in the home countries), and governance (which covers such topics as autocratic rule, low public investment, and systemic corruption). 

What was the action plan?

Not only did the responsible team look for root causes for migration, but they carried out a Pareto analysis to identify the countries-of-origin providing the most migrants. The answer was that—after Mexico (with whom other programs were already in place)—the largest numbers of migrants came from Guatemala, El Salvador, and Honduras. (See this document for detailed statistics.) Therefore the Administration focused its efforts on a combination of humanitarian assistance, direct investment, and government-to-government initiatives with these three countries. [source] In addition, the Administration "created more legal pathways of entry for migrants …. [and implemented] harsher punishments … for crossing illegally." [source]

How did it work out?

Not as planned. Initiatives were put in place promptly in 2021. The next year, total apprehensions at the border hit record numbers.

It would be tempting to wave away these results by pointing out (as we noted above) that there are delays built into any complex system, so perhaps the high numbers in 2022 reflected some causal activity from before the program even started. But the details don't support that simple an answer. In fact, migration from Guatemala, El Salvador, and Honduras dropped significantly between 2021 and 2022. But increased numbers from other countries—notably Venezuela, Nicaragua, Cuba and Haiti—more than made up the difference.

Where did the analysis go wrong?

Wait—maybe that question is premature. Did the analysis go wrong?

At the most basic level, yes it did. The analysis predicted that a certain course of action would have a certain result; and while the narrow result fit expectations, the overall result did not. Illegal migration from Guatemala, El Salvador, and Honduras dropped, but overall illegal migration into the United States rose. The empirical data show that something has to have been left out of the original planning.

To be clear, I recognize that these phenomena are enormously complex, and that we cannot expect the kind of precision we would demand in a laboratory experiment. I hope it is equally clear that I am saying nothing either for or against the specific policies that were implemented. Even if the reasons behind a policy are weak, that says nothing about whether the policy itself is helpful or harmful. Sometimes people do the right things for the wrong reasons.

But the reasoning here was weak in at least one spot: the application of the Pareto analysis was premature.

It was a nice idea, to be sure. In industrial manufacturing, if most of your mechanical failures come from the same step in the process, and if you can fix that step, then you have gone a long way to fixing all your mechanical failures. But that only works when the universe of failures is bounded. Then you can nibble away at one cause after another until you've solved them all.

In the unbounded case, eliminating one source of problems simply makes room for another source of problems to elbow its way in. Gardeners are familiar with this. Years ago my mother's garden was being crowded out by some aggressive ivy; once she had the ivy removed, that made room for a ficus vine to do the same thing.

The Administration's strategy on managing immigration recognized that people come to the United States for reasons including socioeconomic conditions, natural disasters, security conditions, and governance. But many countries in the world suffer from these same problems. Therefore a strategy focused on improving conditions in only a few countries while ignoring the others is not really addressing an actionable root cause. And root causes have to be actionable.

Two final thoughts

First, from the pure problem-solving perspective these empirical results are a good thing, because they teach us more about the problem. Remember that in processing an 8D, after you finish the action plan in D6 you have to check whether the implementation was effective. Did you make the problem go away? If yes, you can proceed to D7. If not, go back to D4, using the new information that you have just learned (namely, that your action plan didn't work) to re-do your root cause analysis. In this case, the results of the Administration's strategy at the border help us to figure out (as we saw above) that their application of the Pareto analysis was premature. That information should help make future analyses stronger.

Second, does anyone want me to try to sketch out what a full root-cause analysis of immigration ought to look like? (I mean one that incorporates the lessons learned that we discussed above.) I won't generate anything useful, of course. The issue is far too complex for any back-of-the-envelope scribbling to be practical. But if you think it would be conceptually helpful, leave me a comment.