« Why would Bloomberg run? | Main | What would a 2016 Democratic postmortem look like? »

January 29, 2016


Feed You can follow this conversation by subscribing to the comment feed for this post.

Hi Noel -- I don't see any reason to make mention of my sister. I'm responsible for what I write.

On Qian & Lichter:
I made two claims from Qian & Lichter.

In 2007, Zhenchao Qian of Ohio State and Daniel T. Lichter of Cornell found that over the course of the 1990s, the percentage of Asians marrying whites, and Hispanics marrying whites, fell sharply, a development they attribute to rising immigration.

Qian & Lichter wrote:
The past decade has ushered in unprecedented declines in intermarriage with whites and large increases in marriage between native- and foreign-born coethnics among Hispanics and Asian Americans ... The retreat from intermarriage largely reflects the growth in the immigrant population.

It's possible that you came away from the paper with a different conclusion. I drew on their language.

Second, you wrote:
Qian and Lichter found that the skills gap between immigrants and natives also plays a role. For example, native-born Hispanic women with a college education were more than three times as likely to be married to whites as native-born Hispanics with less than a high-school education.

From Qian & Lichter:
In general, rates of intermarriage go up with increasing levels of education…native-born Hispanic women with a college education were more than three times more likely to be in a marriage with whites compared to their counterparts with less than high school education.

I did not "misrepresent" Qian & Lichter ... I reiterated their conclusions.

As to the points you raise:

If you classify anyone who checks off the boxes as half-white as minority — i.e., you classify people the way the 1990 census did — you would compare column (1) and column (4) and find that between 1990 and 2000 interracial marriages for native-born Asian-American men rose from 50% to 55% and for women from 58% to 67%.

It is not true that the 1990 census "classified anyone who checks off the boxes as half-white as minority." That's what the 2000 census did. The 1990 census only allowed respondents to select one racial category. So the people who were mixed-race in 2000 might have been Asian in 1990, or they might have been white. Qian and Lichter took differences between the two censuses into account when arriving at their conclusion that intermarriage for Hispanics and Asians declined due to immigration. (Page 75: "We use a cohort approach to explore—at the aggregate level—how multiracial individuals in 2000 may have reported their race in the 1990 census.")

Nor do Qian and Lichter believe that intermarriage rates have declined for Latinos. They spell out why on page 85, where they attempt to account for confounding demographic shifts. Bottom line: “Intermarriage with whites increased significantly among native-born Asian Americans and Hispanics, by 36 percent and 13 percent, respectively. Significantly, these multivariate results are different from the modest 1990s’ declines in intermarriage reported in Table 2.”

This is a bit odd. In this section, Qian and Lichter are obviously not saying that everything they have said up to this point in the paper is wrong—that, in fact, intermarriage increased for Hispanics and Asians. What they are saying is if you were to discount the impact of an array of important variables—"the changing size of each racial/ethnic group, sex ratios, and educational compositions" and so on—then you would have seen an increase in the intermarriage rate. The reason they do this is that they are trying to isolate the effect of marital selection—of individuals discriminating against potential marriage partners of a different race. People apparently became less discriminating, all other things being equal, from 1990 to 2000. But that does not mean the intermarriage rate increased for these groups. It decreased, in large part because of immigration.


Your characterization of my understanding of the Italian-American experience seems to reflect a misunderstanding on your part, which is fair enough.

And I don't see how any of the "three conclusions" you highlight from Jiménez's work on replenished ethnicity undermine my paragraph on his excellent book. There is more to his book that his conclusion In fact, the Jiménez passages you quote seem entirely consistent with what I wrote: ethnic replenishment reinforces the salience of ethnic boundaries. (I fully expect that Jiménez does not share my policy conclusions. I would be surprised, however, if he said that the intra-ethnic dynamic he describes has only one expression.)

I absolutely believe that assimilation is proceeding apace. I also believe, however, that we are witnessing segmented assimilation, and that to make generalizations about immigrants as a whole ignore stark differences in entry wages, social networks, rates of wage convergence over time, and in various cultural indicators over time across groups. Much depends on the "mode of incorporation," including whether immigrants are unauthorized or authorized. Yet it is also true that the experience of Puerto Rican migrants (US citizens) has been not entirely dissimilar to that of immigrants streams that are more heavily unauthorized, which suggests that legal status is not the only factor at work.

Hi, Reihan! Thank you for the response and apologies for the delay in my answers to your objections.

There are a couple of different moving parts here, so let me separate them. I’ll start with Qian and Lichter. Their argument is rather complicated. To paraphrase their findings:

In the gross numbers, which I presented in the body of the post, intermarriage between Latinos and native-born non-Hispanic whites declined substantially; intermarriage between Asian-Americans and native-born non-Hispanic whites declined for Asian-American men but rose for Asian-American women. The fall in the number of intermarriages by Latinos and Asian-American men was driven by an increase in the number of first-generation immigrants, who are less likely to marry out of their racial group. On the other hand, the propensity for individual first and second-generation immigrants to marry non-Hispanic whites rose.

They summed up the immigration finding in their abstract thusly: “The past decade also ushered in unprecedented declines in intermarriage with whites and large increases in marriage between native- and foreign-born co-ethnics among Hispanics and Asian Americans.” When you read the paper, it is very clear from the data that the latter drove all of the decrease.

If that were all they had found, then it might be correct to conclude that assimilation is slowing — although it isn’t fair to dismiss the rise in intermarriage among Asian-American women! The complication is that is not all they find. In your response, you nailed exactly what’s going on. To paraphrase your paraphrase of Qian and Lichter: A individual Latino or Asian immigrant with the same level of education and the same generational status was more likely to marry a native-born white person in 2000 than in 1990. This holds across educational levels. Outmarriage fell (slightly) because the proportion of first-generation less-educated immigrants rose, not because those kinds of immigrants became less likely to marry out.

They are in fact saying that it would be wrong to put too much weight on the gross numbers presented earlier in the paper.

So is it correct to say that assimilation slowed down if individual propensities to marry out continued to rise? I would say not. It is, of course, still entirely legitimate to argue that the gross numbers are what matter, but I do not see how your argument about segmented assimilation holds if outmarriage is increasing across all social groups.

As for the effect of the changes in the census form between 1990 and 2000: I apologize for being unclear. As you say, in 1990 you could check only one box; in 2000, you could check several. The general consensus from the microdata is that most of the people who checked multiple boxes in 2000 had previously checked off a non-white box. That consensus is not absolute, however, which is why they ran their calculations multiple ways.

And all of those calculations show increases in Asian-American outmarriage relative to the headline numbers. Outmarriage rates for Asian-American women increased even more than the headline increase of 2%. Outmarriage for Asian-American men either decreased by only 6% (instead of 9%) or increased by 10%. Even if you want to discount the shift-share argument, those numbers are simply not consistent with the argument that Asian-American assimilation is slowing!

As a general rule, it’s not a good idea to cite social science research as gospel, especially when the paper is as accessible as this one. Their own data shows that they are phrasing some conclusions more strongly than the data warrant. (This is sadly common in research; I’ve been caught out by referees for doing the same thing.)

Back over to you, Reihan.

Regarding Jiménez: you’re absolutely right! His passages could be used to reinforce your argument. But you elided over the reason he finds ethnic boundaries to become more salient with more immigration. Is not that Mexican-Americans find it easier or more attractive to retreat into a separate cultural enclave. Rather, it is that non-Mexicans become more racist the more they encounter first-generation Mexicans.

(I mean racist quite explicitly: he concludes that native-born whites initial reaction is against Mexican cultural signifiers, not skin color, but those whites then begin to negatively stereotype everyone of Mexican descent based on phenotype. I.e., the whites may not start out racist, but high levels of immigration prompt them to become racist.)

Recognizing that would not change your policy conclusions! But it does present the problem those conclusions are trying to correct in a different light: one of native-born racism rather than foreign-born separatism. No?

Me, like everyone in Brooklyn of a certain generation, I always wanted to be Italian. According to Vox, that is why I continue to raise the “aw” sound. What, who wouldn’t wanna talk like Al Pacino?

But seriously, what am I misunderstanding about your characterization of the Italian American experience?

The comments to this entry are closed.