Bernie Sanders, the democratic socialist from Vermont, has been getting an unusual number of challenges from the left of late. First it was on race, now it's on immigration. In an interview with Ezra Klein, he declared that open borders would be a "Koch brothers proposal."

Dylan Matthews hit back, making the case that open borders — in other words, much greater levels of immigration — would be splendid policy. Now, I am all for pie-in-the-sky proposals. And on the merits, advocates of open borders have a good case. But it's worth considering the likely downsides to such an approach.

First, let's keep in mind the political context here. No major presidential candidate is anywhere close to supporting open borders — on the contrary, the GOP frontrunner is an unhinged nativist who rants about how Mexican immigrants are rapists. So by American standards, Sanders is quite pro-immigration. He supports comprehensive immigration reform and the DREAM Act, focusing his ire mainly on guest worker programs that are often used to trap workers in low-paying jobs with few prospects to obtain permanent status or citizenship.

As for the economics: The evidence marshaled in favor of open borders is pretty strong, but it is possible to overstate the case. Matthews quotes the libertarian economist Bryan Caplan making the following prediction: "To me, a big point of open borders is just to fast-forward to the world of the future where everyone can enjoy a First-World standard of living rather than making people wait 100 years."

The idea here is that people in poor nations are more productive when they move to rich nations, sometimes to an eye-popping degree. That's definitely true when it comes to individuals. So if all poor nations emptied themselves into rich ones, hey presto, world GDP doubled and poverty eradicated!

I suspect this is a big part of the appeal of open borders for libertarians, who are often criticized for wanting to torch social insurance. It allows them to claim the moral high ground, while continuing their traditional practice of treating human beings like interchangeable widgets.

Still, these migration and productivity studies are solid. But there is a rather enormous scaling problem in projecting such titanic benefits. Let's compare the United States to China. The two nations are about the same geographic size, while China has about four times the population. Assuming a roughly equal capacity for human support, one could pack the entire population of both North and South America in the U.S. (about 950 million), plus Nigeria (184 million), Ethiopia (90 million), and Egypt (89 million), and it would about equal China's 1.36 billion. I don't think it overly cynical to assume that the results shown by Haitian immigrants in America won't scale up 100 million-fold.

What air-dropping a billion random foreigners into the country would do, of course, is create the mother of all nativist backlashes. That is a truly lamentable fact, but it's also undeniable. Even extremely decent, cosmopolitan nations struggle mightily to incorporate a significant uptick in immigration.

That is the legacy of nationalism, which produced many of the greatest evils of the 20th century and is also a key foundation of the modern nation-state. As much as it devils libertarians, as yet no other human institution has proven itself capable of operating a modern economy. The biggest experiment in post-national economics thus far, the eurozone, has been a cataclysmic disaster.

Indeed, capitalism was one of the major forces behind the rise of nationalism, due to its need for large markets and a strong state to enforce property and contract law. Without a basis of mutual trust and fellow feeling, a state simply can't work. Ironically, classical capitalist texts typically ignore the nation altogether. As historian Eric Hobsbawm wrote about the heyday of classical liberalism (the antecedents of modern libertarianism) in the 19th century:

Liberalism was the anarchism of the bourgeoisie and, as in revolutionary anarchism, it had no place for the state. Or rather, the state as a factor in the economy existed only as something which interfered with the autonomous and self-acting operations of "the market"... If it could be shown that the entire industrial production of the world should be concentrated in Madagascar (as 80 percent of the production of watches was concentrated in a small part of Switzerland), or that the entire population of France should transfer itself to Siberia (as a large proportion of Norwegians were actually transferred by migration to the USA), there was no economic argument against such developments. [The Age of Capital]

In libertarian utopia, the entire world population would be constantly shuffled from place to place in search of the highest marginal productivity. That this would obliterate the social basis for capitalism simply goes unmentioned.

At any rate, none of this is to say that immigration should be forbidden or even slowed. My concern is how it can best be increased and accommodated without stoking a nativist furor that would undermine its political support.

As my colleague Jeff Spross points out, full employment and pro-union policies are one obvious answer. Economic weakness and insecurity make people resentful and suspicious of foreigners — anti-immigrant and even straight-up fascist parties are on the rise across Europe. But when jobs are plentiful and labor is scarce, it's much easier to sell arguments that immigration is a net benefit for just about everyone.

One can already see glimmerings of this in the United States. Despite the nonsense about our supposed "exceptionalism," when it comes to friendliness to immigrants America really is quite an outlier — barely any other country has birthright citizenship, for example. Constant material prosperity and a reasonably fast pace of immigration might conceivably lead to a Star Trek-style utopia eventually — but saying we can double world GDP with this one weird trick probably won't.