Tyler Cowen directed me to a thoughtful essay on immigration by Reihan Salam.  Salam favors an immigration policy that emphasizes integration and cultural cohesion.  At the same time, I believe he misses some recent trends that suggest fears of cultural fragmentation are overrated.  For instance, consider this comment:

In past eras, high immigration levels were matched by high native birthrates. The end result was that, even if immigrants had large families, these second-generation youth were greatly outnumbered by the descendants of the native-born. Investing in the next generation meant investing in the children of immigrants, yes, but also in the children of natives, who, by virtue of their numbers, would set the cultural tone.

Collapsing native birthrates have changed the picture, setting off a cultural panic among the likes of Rep. Steve King, the Iowa congressman who infamously tweeted, “We can’t restore our civilization with somebody else’s babies.”

Immigrant birthrates have recently been falling much faster than native-born birthrates.  While the immigrant birthrate is still slightly higher, the difference is no longer as dramatic.  Here are some recent data:

The birth rate for women in their reproductive years (ages 15-50) declined more than twice as much for immigrants as natives between 2008 and 2015. Between 2008 and 2015 the fertility of immigrant women feel from 76 to 60 births per thousand. In contrast, native fertility declined from 55 births per thousand to 49 births per thousand — a decline of six births per thousand. Although still higher than that of natives, immigrant fertility has only a small impact on the nation’s overall birth rate. The presence of immigrants raises the birth rate for all women in their reproductive years by just two births per thousand (3.6 percent). Even if the number of immigrant women 15 to 50 doubled along with births to this population, it would still only raise the overall national birth rate for women by 2.5 percent above the current level.

Salam also worries about a permanent underclass of low-skilled immigrants.  But recent trends show a sharply rising share of immigrants are coming from Asia, and of course Asians actually out earn white Americans.  Meanwhile, the share of immigrants from Latin America has declined sharply.  Furthermore, there is a very high rate of intermarriage between whites, Asians, and Hispanics, and the children of these marriages (including my daughter) are generally regarded as culturally “white”.  This is why it’s not likely that the share of whites in the US population will fall sharply over the next 50 years.  America today is much like America of 100 years ago–a culturally fragmented society that will become more unified as the children of the immigrants assimilate and intermarriage increases.  Fifty years ago, America was more culturally homogeneous than 100 years ago.  The same trend toward integration is likely to occur over the next 100 years, even if we accept a lot of new immigrants.

This is not to say that Salam has no valid arguments.  His call for a more skill-based immigration approach has some merit, as long as the US has a welfare state.  (If we really cared about the poor, we’d get rid of the welfare state and have lots more immigration of poor people.)  The US could easily shift to a more skills-based approach, and at the same time sharply increase the overall level of immigration—to something like 3 million per year.  These educated newcomers would come from all over the world, indeed we already accept many highly skilled immigrants from regions such as Asia and Africa.  Don’t assume “highly skilled” means “white”. (Unfortunately, the Trump administration has been trying to reduce immigration of highly skilled workers.)

Let me also say a word about the recent worry about AI taking our jobs.  I cannot say that this problem will never occur, but right now we are very far from having a labor surplus.  Over the past six months, I have experienced the tightest labor markets of my entire life.  Service at many businesses has fallen sharply.  I often ask people at restaurants and stores why the service has gotten so much slower, and the answer is always the same, “we can’t find enough workers”.  Recently, we waited 15 minutes to get a table at a restaurant where two thirds of the tables were empty.  They didn’t even have enough workers to seat customers. So while AI might conceivably be a problem in the future, recent trends are going in the other direction. For now, we could use a lot more low-skilled immigrants to serve as cooks, waiters, cashiers, hotel maids and farmworkers.

This reminds me of currency.  People keep saying that electronic money will replace currency, while the actual use of currency keeps rising.  Indeed US currency holdings are now far higher than 90 years ago, even as a share of GDP.  While some of that currency is held overseas, the same is true in many other developed countries.  Even at a global level, currency use doesn’t seem to be declining significantly. Yes, someday currency will probably be replaced, but for the rest of my life it will be an important part of our economy.

PS.  The sharp decline in service quality means that American living standards are rising more slowly than GDP.  However, on balance I’m still pro-tight labor markets.  They are bad for me personally, but it’s very good for America.