Does immigration make littering worse?

Philippe Lemoine is a young and outspoken campus conservative.  He is also a very bright boy generally.  He is very much at ease with statistical analysis.  He uses that skill to do a thorough examination of the claim that minorities litter more.  He first takes apart a claim that they do not and shows the defense concerned to be statistical garbage

He then goes on to examine statistically all aspects of the issue and shows that blacks and Hispanice litter more than twice as much as whites.  His analyses are pretty straightforward so should be understandable by many readers.  Because he has to cover so many bases, his article is very long.  I reproduce below therefore just the more important excerpts.

Does immigration make littering worse?  It sure looks that way
Alex Nowrasteh and Andrew Forrester just published a piece which they claim to show that, in the US, immigrants don’t litter more than natives.

Their post has been shared pretty widely and uncritically by pundits, academics and journalists and it’s already being used to argue that anyone who claims that immigrants litter more than natives is a racist. The problem is that, as I will argue in this post, not only does their analysis fail to show that immigrants don’t litter more than natives, but insofar as the data they used show anything, they show exactly the opposite.

Nowrasteh and Forrester claim to have been moved to write about this because, even though the issue seems a bit superficial, the argument that immigrants litter more is made increasingly often by restrictionists:

At the recent National Conservatism conference, University of Pennsylvania law professor Amy Wax argued in favor of restricting non-white immigration to the United States because she said they litter more. My colleague David Bier was heckled at the Conservative Political Action Conference in 2018 and asked about immigrant littering. Fox News Tucker Carlson has been bringing up immigrant littering over the years, most recently with the help of City Journal associate editor Seth Barron.

If you ask me, this paragraph is extremely uncharitable, to the point of being dishonest. It makes his sound as if Carlson and Wax think the fact that non-white immigrants litter more alone justifies restricting non-white immigration, but obviously that’s not what they mean. They just use littering as one example of anti-social behaviors that, according to them, immigrants from certain parts of the world engage in more often than natives.

In fact, if you read what Wax said in her infamous talk (it was actually during the Q&A), she explicitly said as much:

"I think we are going to sink back significantly into Third Worldism. We are going to go Venezuela, and you can just see it happening. I mean one of my pet peeves, one of my obsessions, is litter, and I… If you go up to Stockbridge, Massachusetts, or Yankee territory, right? Or versus other places that are “more diverse,” you are going to see an enormous difference. I’m sorry to report. You know, generalizations are not very pleasant, but little things like that, which aren’t little, they really affect our environment, attitudes towards public space."

Anyway, having explained why they felt the need to jump into this debate, Nowrasteh and Forrester  describe how they allegedly showed that immigrants don’t litter more than natives:

"Fortunately, there are data available to at least partially answer this question. The American Housing Survey (AHS) is a biennial longitudinal housing survey that asks about the amount of trash, litter, or junk in streets, lots, or properties within a half-block of the respondent’s housing unit. The answers are a “small amount of trash,” a “large amount,” and “no trash.” We constructed a scale from zero to one using a min-max normalization for all non-missing observations where a higher value indicates more trash in a neighborhood. We then take a weighted average of these scores using the weighting variable present in the AHS public use file for each metropolitan area.

The smallest geographical unit in the AHS was the Core-Based Statistical Area (CBSA) for 15 major urban areas in the United States that account for about 33 percent of the total U.S. population (around 58 percent of the foreign-born population and 30 percent of the native-born population).

We linked the foreign-born shares of the CBSA populations from the 2017 American Community Survey (2013-2017, 5-year estimates) to the AHS survey responses on the amount of litter and trash. We then ran a regression where the independent variable is the percent of the CBSA’s population that is foreign-born and the dependent variable is the response to the litter question."

As they explain in the rest of the post, when they did that, they didn’t find statistically significant relationship between the proportion of immigrants in a CBSA and the amount of litter.

But does it mean that Nowrasteh and Forrester are right and that immigrants don’t litter more than natives? Well, no, it does not. Not at all. The way they analyzed the data makes absolutely no sense and, when you do it correctly, you find exactly the opposite result. In fact, insofar as the data show anything, they completely vindicate Wax, but amazingly Nowrasteh and Forrester reached exactly the opposite conclusion. The problems with their analysis are so obvious that, when I read their post, I just couldn’t believe it. The dataset they used contains more than 50,000 observations, but by aggregating at the CBSA level, they reduced that to only 15 observations. The only reason to use this approach is that, by doing that, we can precisely determine the demographics of the area where respondents live instead of using their own demographics for the composition of their neighborhood. But the CBSAs in the dataset are huge and contain a lot of people, whereas respondents were asked about the presence of trash in the streets within 1/2 block of where they live, so this is not helpful.

In general, when people complain about “immigration”, even if they don’t use any qualifier, they are only complaining about some groups of immigrants but not all of them, so the average effect of immigrants is irrelevant to their claims. For instance, when people in France complain about immigration, everybody knows they’re talking about African and North-African immigrants. I have almost never heard anyone complain about Asian immigrants, because they tend to do very well and don’t cause many problems.  Thus, if we are going to use the American Housing Survey to assess whether people like Carlson and Wax are right, we must disaggregate by region of origin and/or race, which Nowrasteh and Forrester didn’t do. Again, even if they had, it wouldn’t have shown anything one way or the other, because it would have done nothing to alleviate the problems with their methodology that I described above.

Finally, restrictionists about immigration don’t just talk about immigrants, but also about their descendants, because their central point is that neither immigrants nor even their descendants magically adopt the culture of their country of destination just in virtue of living there. You may think that it’s wrong, although it clearly isn’t, but in any case that’s the claim they make, so you can’t possibly refute it by just looking at immigrants. Indeed, just looking at immigrants and ignoring the effects of their descendants is another trick that pro-immigration advocates often use, but the effects of immigration are not limited to the effects of immigrants themselves. It also includes, among other things, the effects their descendants have. In fact, if you go back to Wax’s answer where she talked about littering, you will see that she doesn’t even talk about immigrants. She claims that littering is worse in areas that are more “diverse”, so she isn’t making a point about immigration per se but about race/ethnicity, although those issues are obviously related. Thus, if we want to use the American Housing Survey to assess whether she is right, we have to examine whether race/ethnicity, not just place of birth, is related to littering.

Even though I have just explained that one should disaggregate between immigrants depending on region of origin, let’s first see whether, when the data are analyzed at the individual level, we find a difference between natives and immigrants.As you can see on this chart, where I represented the confidence intervals, there is a statistically different between immigrants and natives, so the answer is yes.

Next, since Wax singles out non-white immigrants, let’s see what happens when we disaggregate the group of immigrants into white and non-white immigrants.Well, look at that, it turns out that not disaggregating hid a significant difference between white and non-white immigrants. I wonder who could have predicted that? Well, I guess we know at least one person who had predicted it, Amy “the horrible racist” Wax.

But the white/non-white dichotomy is still very crude and no doubt hides a lot of heterogeneity, so let’s focus on immigrants and disaggregate further based on where they come from.As you can see, for most groups of immigrants, there is no way to know for sure whether they are less or more likely than natives to report the presence of trash in the streets near where they live. But this is not true for immigrants from South/Central America and the Caribbean, who are significantly more likely than natives to do so. As it happens, when people say that immigrants litter more than natives, hispanic immigrants are precisely the group they usually single out. Thus, far from undermining this claim, the data seem to support it. It should also be noted that South Asians are significantly less likely to report the presence of trash in the streets near where they live, but I’ve never heard anyone complain that Indian immigrants litter…

Finally, as I have explained above, people like Wax are not just concerned about immigrants themselves, but also about their descendants. Even if immigrants litter more than natives, it may not be a big deal as long as their descendants, having been socialized in the US, had adopted the local norms and behaved similarly to natives with respect to littering. In order to check whether the data supports this hypothesis, let’s look only at natives and disaggregate by race/ethnicity.


No comments:

Post a Comment

All comments containing Chinese characters will not be published as I do not understand them