Robots or Immigrants?

Post by Nathan Smith (regular blogger for the site, joined March 2012). See:

There’s been some buzz lately about falling birthrates in the United States.

The U.S. birthrate plunged last year to a record low, with the decline being led by immigrant women hit hard by the recession, according to a study released Thursday by the Pew Research Center.

The overall birthrate decreased by 8 percent between 2007 and 2010, with a much bigger drop of 14 percent among foreign-born women. The overall birthrate is at its lowest since 1920, the earliest year with reliable records. The 2011 figures don’t have breakdowns for immigrants yet, but the preliminary findings indicate that they will follow the same trend. (via Marginal Revolution)

Here’s Ross Douthat’s take and Megan McArdle’s take, both very eloquent, and thoughtful, and worried, as is Bryan Caplan‘s take. Douthat mentions immigration obliquely but doesn’t think it’s a solution to demographic decline:

But deeper forces than the financial crisis may keep American fertility rates depressed. Foreign-born birthrates will probably gradually recover from their current nadir, but with fertility in decline across Mexico and Latin America, it isn’t clear that the United States can continue to rely heavily on immigrant birthrates to help drive population growth.

This isn’t quite convincing, because the US wouldn’t need high immigrant birthrates to drive population growth. High levels of immigration would suffice to drive population growth. McArdle goes into more detail about the possibility of more immigrants as a solution to demographic decline:

In theory, you just export capital to younger societies, or import young immigrants.  But there are some problems with this theory, the largest of which is that the whole world is getting older almost all at once.  Every country is facing (or soon will) the same looming demographic pressure.

That’s an exaggeration. It’s true that birthrates are falling virtually everywhere in the world, but they’re still pretty high in Africa and many other developing countries (with India, 20.60 births per 1,000 persons, well above the United States, 13.68). There will be plenty of young immigrants to draw in for a long time yet. McArdle argues that there are limits to investing a broad as a strategy for securing the future:

Which means, first, that there are fewer big growth opportunities than there otherwise would be, and second, that if you send a bunch of capital abroad, it may be very hard to get paid when you’re ready to consume.  Nationalism has not gone away, particularly not in the developing world; I find it somewhat unlikely that cash-strapped Argentinians or Brazilians or Chinese governments are going to allow their countries to ship output abroad for the benefit of comparatively rich American retirees.

Well, at any rate there is certainly long-term political risk involved in strategies that systematically lend out capital so as to create a future drain of dividends and loan repayments back to the US. Of course, net investment abroad is not a strategy the US has been pursuing lately! Back to immigration:

And third, that the supply of immigrants may not be what folks are expecting.  All the immigrants that we want to import have parents and grandparents–and unlike in previous generations, those immigrants will increasingly have few siblings and cousins to stay back in the Old Country and take care of Mom.  They’re often in countries with poor and weakening safety nets.  How many will abandon them to their fate in order to come here?  Are we going to import all the old people in order to get access to the labor of their children?  Will we import only the orphans, and the callous?  And given Friedman’s thesis, how will we muster the social capital to welcome all these strangers with open arms?

This isn’t quite persuasive. First, as I said, the supply of young foreigners willing to immigrate to rich countries is far from drying up. Birthrates are falling, but they’re still high enough in many places to supply millions of young migrants who won’t have to abandon parents alone at home in order to come.

Especially since, after years of working for us, those immigrants will themselves become consumers of social security benefits and retirement savings.  That means we do not just need more immigration, but accelerating immigration, even as the rest of the world shrinks.  Meaning that at best, this is a short-term patch for a long-term problem.

That’s not necessarily right, either. First, as one group of immigrants ages, you can let in the next group of immigrants. Maybe there are limits to that, but it’s not obvious what they are. It could be that the global supply of willing young immigrants eventually runs out, but that’s a long way off and may never happen. It could be that the right kind of immigrants are in limited supply, but (a) even high-skilled immigrants have trouble getting in now, and (b) with education levels rising all over the developing world, if we do think high-skill immigrants are better (though I’m skeptical: our erudite commenter BK may be an exception, but I think most people who think this simply misunderstand comparative advantage), the supply of those is probably increasing too. McArdle worries about how well assimilation can work:

Even without the political dynamics, I think there are practical limits to how fast immigrants can be absorbed.  I’m a big fan of immigration, but I think it needs to happen at a speed where they can be assimilated.  Much of the western world’s higher productivity is simply what you might call our “cultural operating system”: a set of shared assumptions about how things are done, what is right (refusing to pay bribes) and wrong (using your government job to procure stuff for your relatives).  If you import new migrants too fast, the operating system breaks down.

Quite possible, though I am not aware of any historical examples of immigration significantly degrading culture and/or institutions, except when it took the form of an explicitly military invasion from the beginning.  Of course, it doesn’t follow that it couldn’t happen, and this is something to watch carefully.

I don’t mean to say that we have too many immigrants now–I think we could take more than we do.  But I don’t think that we could import enough people to make up for, say, Italian or Japanese-style birthrates.  The percentage of the population that is foreign born is already pretty close to its historical peak–at least, the part of our history in which we were a settled, industrialized country.  And if the issues are large for us, they are probably insurmountable for countries like Japan, where a big part of the cultural operating system consists of trusting other Japanese people.

Yes, that’s one reason why you don’t want to be like Japan. Open beats closed. But I think McArdle is missing an important factor here, namely, that globalization has half-Americanized half the world already. 19th-century immigrants may have been racially more similar to America’s white native majority, but they were less familiar with democracy, with the English language, with America via movies and music and TV, with American-style market capitalism, with Coke and McDonalds and Microsoft and Google and many other American firms, with blue jeans and free speech and religious tolerance, than a 19th-century immigrant from the Hapbsurg empire or tsarist Russia. There are lots and lots of foreigners who could show up on an American college campus or in an American corporate office building and fit in, just fine, almost immediately. There’s no reason to think that 15% or so was ever an upper bound on the foreign-born share of the population that America could absorb, and that upper bound is probably much higher today, because of cultural assimilation that has occurs across international borders, with the influences running both from abroad to America and from America to abroad, though the latter direction of influence is surely more important.

The keyhole solution in Chapter 9 of Principles of a Free Society is partly designed with the idea of maximizing assimilation in mind. Through a mandatory savings account withdrawable only in a migrant’s home country, it encourages migrants who are not particularly enamored of life in America to go home, where their savings might help them buy a home or start a business or a family, while those who love America enough to forfeit a lot of forced savings to stay, can stay. By allowing a lot of sojourners to come, America would develop a lot of international connections and, I expect, cultural/educational influence. Also, recognizing a right to migrate is likely to evoke more efforts on the part of young foreigners to become Americans, than any discretionary policy. Acquiring the language, skills, culture, and so on to become a full-fledged participant in the American polity and economy involves a large personal investment. If the payoff to this investment depends on the arbitrary discretion of some bored and resentful customs official doing time so he can get a real job in the diplomatic service someday, much less of that investment will take place. In that sense, while demographic decline may be an argument for opener borders rather than open borders, it would still help to make the process rule-based and rights-based rather than arbitrary, as at present.

I jumped to immigration as a solution and skipped a prior question: Why is demographic decline a problem? First, is there a reason to prefer some particular level of population for the US, or other countries? Probably not, or not much. On the one hand, some resources are inherently scarce, but this isn’t a very important constraint. I recall an estimate that land’s share of national income in the US is probably about 5%. On the other hand, more people allow for a richer division of labor and generate more new ideas. The two effects offset each other. I think more population helps per capita income, but it probably isn’t a huge effect either way.

Second, is there a reason to regard a particular age distribution as desirable? Very likely so, though the reason why is not as obvious as it might appear. A greying population involves an increase in the dependency ratio, or (to put the same thing another way) fewer workers per retiree. That’s bad news for public pension programs: higher taxes or less generous benefit cuts are needed to make it sustainable. But if far-sighted governments and individuals foresaw planned for demographic decline appropriately, they could, in principle, sustain living standards, even without mass immigration, through savings and investment. Capital equipment would make workers of  the future more productive and help older people keep working longer. 

So ultimately, the choice for rich countries in demographic decline, if they want to avoid seeing living standards fall, may be: robots or immigrants? By accelerating the replacement of muscles with machines, it may be possible to sustain high living standards, with a rather feeble population relying on robots to do much of what humans do now. Life will have to adapt more to the needs of the robots as human labor becomes scarce. Maybe rooms and lawns will be reorganized so they can be cleaned and mowed (respectively) by robots. Shopping may be placed by home-delivery systems, perhaps in the form of pipelines/conveyor belts, which would require greater concentration of population. Just to give a couple of speculative examples. The elderly of the future would have their needs met with little of the human element. Alternatively, large populations of immigrants can be admitted and the services provided in a more human fashion.

A greying society dependend on robots makes me nervous for, so to speak, “Burkean conservative” reasons. It’s too much of a departure from a long and universal human tradition of having lots of young people around. Young people make me happy. They symbolize hope for the future. They give the older generations something to work for as they try to maintain a decent society and lead it into the future. I prefer immigrants to robots.

Not that we can’t have both.

10 thoughts on “Robots or Immigrants?”

  1. This might be an argument against open borders. The Unites States has traditionally been multi-cultural. (The arguments of the Federalist Papers apply just as strongly to multi-cultural societies as multi-regional societies.) If the immigrants come from just one culture, they could potentially crush minority groups.

    1. The point I think Nathan was making is that they can quickly adjust to life in the US, not that they are culturally identical to US natives (who are themselves quite culturally diverse to begin with). In other words, they may like different kinds of food, movies, and music, but they can get along just fine in the US.

  2. Technological growth is a substitute for a growing labor pool, but there are limits to how far you can take it. A major component of retirement spending is health care, a sector which has resisted innovation in service delivery for a long time. Agriculture and manufacturing are already highly mechanized. I think capital investment is going to run up against decreasing returns pretty quickly if America tries to focus on the “robot” side instead of the “immigration” side.

    1. Australian agriculture is a lot more mechanized than Californian agriculture. Wonder why that is…

      Also see the third paragraph of http://www.quora.com/Computer-Programming/Are-computer-programmers-essentially-slave-owners/answer/Christopher-Chang . It’s ultimately GOOD when high wages drive down the relative economic viability of certain operations: let the price signal propagate and reduce our collective dependence on uninteresting jobs. The moral way to help still-existing poor populations is trade, not permanently blocking your own country’s development.

      1. Christopher, I think the relation between cheap labor and technological progress was not quite the focus of Nathan’s post, but we have a detailed discussion of these issues on the website here. My main takeaway on this point is that although cheap labor does slow down labor-substituting innovation in the industries that are flooded by that cheap labor, this is in general welfare-enhancing because the high-skilled labor that would be used to mechanize jobs that could be done by low-skilled labor can instead be used for other purposes — including perhaps mechanizing high-skilled labor! Perhaps agriculture would be more mechanized with more immigration restrictions, but the high-skilled labor effort that would go into building cheaper, better, and more reliable agricultural machines would then be diverted away from other areas of innovation. More at the link.

        1. Your preference may differ, but I personally prefer to live in a society with a large and increasing share of high-skilled people. For those with my preference, making an essentially one-time effort to mechanize away many lower-skilled jobs has grossly superior effects compared to *always* “[using the high skilled labor] for other purposes”.

          Again, I recognize that others have different preferences than me, which is one reason why I wish no ill will toward controlled open borders experiments. But my own preference is legitimate, and it’s blatantly obvious to me that *some* sort of restriction (perhaps falling under what you call an “open borders keyhole solution”) is necessary to satisfy it.

  3. Actually, Japan proves that half the cleaning service jobs and even simple task home care cold be done by robots. Immigrants needed has always been exaggerated and Japan still has a lower unemployment than most of the world since folks are old. Young populations do worst folks.

    1. Well of course Japan has a low unemployment rate; they’ve run out of humans so they need to hire robots. The relevant question is not whether various tasks can be done by robots. The question is whether robots equal or outperform humans from a cost-adjusted standpoint; whether, if given the choice right now between hiring a robot and hiring a human (at market prices for both), you would hire a robot. In a consumer’s daily life, the robot currently wins for vacuuming the floor (in some cases, anyway, and primarily in countries where the labour costs are prohibitively high), and little else outside Japan.

      This suggests that either the barriers to trade between Japan and the rest of the world are incredibly high, or that in practice, robots aren’t even close to outperforming humans for most things consumers expect in their daily lives. Given how greedy US corporations (or firms anywhere else, really) are, do you think they would still be hiring Mexicans to clean their offices if they could get the job done just as cheaply or even cheaper by machines?

Leave a Reply