First, Yglesias is right to observe that immigration raises the average income of native-born Americans. It’s not clear, however, that immigration raises the average incomes of native-born Americans more than a range of other policy measures, like shifting towards consumption taxation, reforming Social Security and Medicare, or drastically improving the productivity of our education system. Moreover, not all Americans are native-born (roughly 13 percent of the U.S. population is foreign-born, as is 16.3 percent of the U.S. labor force), and there is a broad consensus that, as Yglesias has pointed out, previous immigrants are “the big economic loser” from increased immigration levels. It is easy to imagine a scenario in which increased immigration raises the average incomes of native-born Americans — by, for example, lowering the cost of child care for affluent professionals, who can then work longer hours — while also lowering GDP per capita, assuming that new immigrants lower the average skill level of the U.S. workforce, as has been the case in recent decades. Would this be a desirable outcome? It really depends on what we care about most.

Yglesias finds it odd that anyone would be concerned about the economic impact of increased immigration on previous immigrants. And he has a point, as the people who are most skeptical of increased immigration tend to be native-born Americans, not immigrants, who are generally sympathetic to the idea, not least because many of them hope to have friends and relatives join them in the U.S. Yet there is a fairly straightforward reason to be concerned: given that previous immigrants are poorer and less-educated than other Americans on average, reducing their market incomes will have negative consequences for them and for their children. Let’s stipulate that affluent professionals really do benefit enormously when they have a ready supply of less-skilled service workers to prepare their meals, mind their children, drive them to and fro, and perform various other tasks they’d prefer not to perform for themselves.

Nevertheless, we should keep in mind the life chances of the children of the less-skilled service workers who are helping to make life less expensive for affluent professionals. Can we be confident that they will earn much higher incomes than their parents, despite the evidence that children raised in low-income households and in high-poverty neighborhoods tend to fare worse than their better-off counterparts? Do we have reason to believe that the public schools serving these children can make up for the various deficits they’re likely to experience at home? ​We can mitigate the effects of low market incomes by increasing transfers and by creating or expanding labor-intensive programs designed to improve the long-term economic prospects of their children, hence recent calls for more spending on early childhood education. But of course this is very expensive, and the U.S. already has a large number of children who are in great need of such labor-intensive services. Does it make sense to greatly increase the number of U.S. children living in low-income households to make life easier for affluent professionals? Perhaps we can tax away the higher incomes that affluent professionals earn by outsourcing household production and use this tax revenue to finance the new support services for the children of less-skilled immigrants. Will this solve the problem? To believe that you must first believe that government social programs can adequately substitute for engaged parents with strong social network. The evidence for this proposition is not exactly rock-solid.