Tag Archives: Garett Jones

Open Borders and the Hive Mind Hypothesis

I recently finished a (fairly advanced) draft of “The Global Economic Impact of Open Borders,” posted to SSRN here. Here’s the abstract:

Open borders, in the sense of the abolition of policies restricting migration, would cause billions of people to migrate, and result in almost a doubling of world GDP. Based on a model that stresses human capital as a determinant of the wealth and poverty of nations, but which also has a spatial element and allows total factor productivity to differ across cities, two openborders scenarios are constructed. In the first, “pure market clearing” scenario, world GDP rises 91% as 82% of the world’s population migrates, mostly to the West, and the living standards of unskilled workers worldwide rise to 26% of the US level. In the second scenario, with several adjustments made to favor greater realism at the expense of some arbitrariness, world GDP rises 85% as 58% of the world’s population migrates, and the living standards of unskilled workers worldwide rise to 31% of the US level.

For more on this paper, see my guest post at Market Monetarist last fall; and my three-part series on “Open borders and the economic frontier” (part 1, part 2, and part 3).

While I plan to do another round of revisions (a sorely needed Acknowledgments section is high on the priority list), I want to bring out some of the main themes of the paper, through blog posts at Open Borders: The Case. One of these is that while the claim that open borders would dramatically raise world GDP (“double world GDP” is the usual, sometimes criticized as over-optimistic but in my view apt, slogan) is robust to many changes in assumptions, it cannot withstand the “hive mind” hypothesis about the determination of GDP.

In technical jargon, the “hive mind” hypothesis is that TFP (total factory productivity) depends on human capital externalities. In non-technical language, the hive mind hypothesis is that people’s productivity and earnings depend, not so much on their own intelligence or skill, as on that of people around them.

(In the academic literature, Jones (2011), “The Hive Mind Across Asia,” seems to be the most prominent paper using the phrase “hive mind” in this sense, but Lucas (1988) is a seminal paper for the idea that human capital externalities are an important determinant of TFP, on which a large literature builds. John Lee comments here on how Docquier, Machado, and Sekkat (2014), the chief outlier among papers estimating the global economic impact of open borders, uses a form of the hive mind hypothesis to arrive at its conclusion that open borders would only raise world GDP by 4%.)

The principal motivation for the hive mind hypothesis is to explain the fact that highly skilled workers do not tend to earn more where skills are scarce. If anything, they earn more where they are abundant. A well-designed study by Michael Clemens uses the randomness of US visa allocations to show clearly that Indian software programmers earn far more in the US than in India, for reasons that seem to indicate higher productivity, since they’re working for the same companies, and the companies would have no reason to sponsor their visas if they didn’t anticipate a productivity increase sufficient to justify the higher salary. Against this, my own experience in Malawi taught me that what Amy Chua (2004) calls “market-dominant minorities” can achieve, in the poorest countries on earth, living standards much superior to those of middle-class Westerners in some respects (land, servants, to some extent leisure) while inferior in others (access to shopping, internet) in such a way that, overall, they might be rated as similar; and Chua (2004) gives very extensive evidence that this is true not just in Malawi but all over the world. My rough assessment is that while Indian software programmers might be far more productive in the US, the living standard earned by human capital is similar all over the world. (More precisely, the additional living standard that a worker with human capital equivalent to that of an average American will enjoy, over and above whatever a local unskilled worker would earn varies by less than an order of magnitude across countries.) Still, even if skills don’t earn less where they are scarce, they ought in theory to earn more, and that they don’t is a mystery demanding explanation.

Normally, a factor of production is most valuable where it is scarcest. Thus, water is intensely valued in California but less so on the rainy East Coast. Land is very expensive in Manhattan, where it is scarce (relative to population), but much less so in Montana, where there’s plenty of it. We should ordinarily expect the same thing with respect to brains, skills, human capital. They should be expensive where they are scarce, cheap where they are abundant. In fact, human capital seems to earn as much or more where it is abundant. To resort to technical language again (sometimes it clarifies) there is a correlation between average human capital and total factor productivity (TFP); and the hive mind hypothesis is essentially that this is causal, with the direction of causation running from average human capital to TFP. There may be a variety of reasons, e.g., smart people vote more like economists (so that democracy => good policy works only as well as the voters are smart), or smart people are better at cooperating, or maybe smart people need stimulation from other smart people to exercise and improve their intelligence. That’s the hive mind hypothesis in a nutshell. If you want more, ParaPundit’s post “Benthamite Libertarian Collectivists Wrong on Open Borders” has a good explanation (though without the term “hive mind”) with further links. Even better are the extensive contributions by anonymous commenter BK at the “Open borders and the economic frontier” posts linked above, and also here and here. Chaper 2 of Collier’s Exodus, which I review here, has what might be called an institutional spin on the hive mind hypothesis.

If the hive hypothesis is true (if average human capital is causally linked to TFP), what does it imply for open borders?

First, if the hive mind hypothesis is true, the impact of open borders on world GDP would probably be far less favorable than most of the existing estimates suggest. In one of the two model extensions that make TFP a function of human capital externalities, I find that world GDP falls by nearly one-quarter under open borders. One critique of open borders is that it would kill the goose that lays the golden eggs. The hive mind hypothesis is probably the form of “killing the goose” argument that has the best support in the literature at the moment.

Second, even in my most pessimistic hive mind scenario, unskilled workers worldwide end up with living standards 15% of the current US level, an order of magnitude above current levels. This is a counter-intuitive result, because in essence, what it shows is that the seemingly most pessimistic open borders scenario, killing the goose, and the seemingly most optimistic open borders scenario, the end of poverty, are actually not inconsistent! Even if the institutional harms/negative externalities/various downsides of open borders were severe enough that world GDP, far from doubling, actually fell substantially, yet billions of the world’s most destitute people would still see their incomes multiplied five-fold, ten-fold, or more, because they could live under institutions that, while much degraded relative to the pre-open borders rich world, were still a good deal better than those they suffer under today; and also because they would benefit from greatly increased opportunities for productive complementarity with skilled workers, which the status quo precludes. Is it worth reducing world GDP by one-quarter to raise the wages of the bottom billion ten-fold? Probably so, if it came to that.

Third– and this is why ParaPundit misses the mark– whatever the normative implications of the hive mind hypothesis may be with respect to immigration, they certainly do not suggest that the status quo is anything like optimal. If one’s goal is to maximize world GDP, the only advantage of the status quo, relative to open borders, is that it involves some human capital stratification, allowing today’s rich countries to be productive “hive minds” of high human capital people. But the migration status quo is a very suboptimal human capital stratification system. First, there should be open borders for smart people– no reason to shut them out– and the admission of people with high IQs and advanced degrees to rich countries could be far more automatic and transparent. We could create schemes to bribe low-IQ, less-educated citizens of rich countries to emigrate and surrender their citizenship, since such emigration should raise average human capital, and TFP, in the source country. We could establish charter cities for low-IQ people to emigrate to, and a bit of ongoing foreign aid might be a fiscal price well worth paying to have them out of the way. Meanwhile, we could establish gated communities for smart people, graduating them into increasing levels of self-government, until eventually they attain full independence as (not philosopher-kings but) philosopher-republics.

Critics of open borders from a hive mind angle, like ParaPundit, can be called on to explain why they don’t advocate a global program of maximal human capital stratification, since that’s what their arguments would really point to.

I’m skeptical of the hive mind hypothesis, and I don’t think open borders would kill the goose that lays the golden eggs; I think it would double world GDP. But I don’t rule the hive mind hypothesis out either, and it’s one of the most respectable reasons to dissent from claims that open borders would raise world GDP dramatically.

More related reading

  • Grappling with the Goose by Paul Crider, Open Borders: The Case, February 17, 2014.
  • How migration liberalization might eliminate most absolute poverty by Carl Shulman, May 27, 2014, making a very similar point. Here is the summary:

    While some estimates that open borders would double gross world product implicitly project the migration of most of the developed country labor force, a much smaller quantity of migration might cut global poverty rates by half or better. The additional income to the poorest required to bring them above extreme poverty lines is in the hundreds of billions of dollars per annum, while doubling world product would approach a hundred trillion dollars of additional annual output. Legal barriers to migration, and blocked desire to migrate, are most extreme for the poorest countries, suggesting extra migrants from those sources. While migrants may receive more income gains than are needed to escape absolute poverty remittances to family, trade, and investment may help to distribute the gains more widely. Overall, the case that migration liberalization for less skilled workers could eliminate most absolute poverty is significantly more robust than the most extreme estimates of global output gains.

  • Intelligence, international development, and immigration by Vipul Naik, Open Borders: The Case, August 19, 2012. See also the follow-up Garett Jones responds to my intelligence post.

International Tiebout competition

A major reason for skepticism about open borders among many who are partially sympathetic is the fear that poor immigrants will vote for redistribution. Natives might benefit, on average, from their interaction with immigration in the market, but in the political arena, poor immigrants with little to lose will vote for higher taxes and government handouts, making natives worse off. To deny immigrants the vote would solve this problem in theory, but raises other philosophical issues about the meaning of consent of the governed, and in any case might not be politically sustainable. Opportunistic parties might hand out citizenship to immigrants who they hope will be their future constituency. Immigrants might also take advantage of their physical presence to agitate in the streets for the vote and/or directly for the government benefits they hope to win by it. In short, open borders will make government more redistributive.

But the logic of Tiebout competition points the other way. Tiebout (1956) famously argues that local governments will provide public goods efficiently if people are free to move among jurisdictions, and concludes that we can expect public goods to be provided more efficiently at the local level than at the national level. Caplan recently explained where Tiebout goes wrong: (1) local governments are not perfectly competitive but face downward-sloping demand curves for residence and so can extract monopoly rents; (2) emigrants can’t take their real estate with them so bad local governments will reduce property values; and (3) local governments aren’t for-profit corporations and don’t face the incentives that for-profit corporations do to give customers/residents what they want. All good points, but there’s still something to be said for “voting with the feet.” Local governments may not be profit-maximizing firms, but they still differ in their performance, and people do get some choice over what local public goods they want by deciding where to live. In the short run, local misgovernment may depress property values more than inducing migration, but in the long run, capital can be adjusted, and misgoverned places can revert to weeds while well-governed places teem with new high-rises. And local governments are still somewhat competitive.

In the Tiebout model, government provides public goods rather than engaging in redistribution. The people who choose to live in a location don’t mind paying for what the government does, because they regard the benefits as greater than costs. If not, they would move. The government might charge the rich more than the poor, if it’s still providing the rich value for money, and especially if the rich value local public goods more, in money terms, than the poor do, which is plausible, since they presumably get less marginal utility from a dollar but might not get less marginal utility from a statue in the park or clean air or good streetlights. But if the government charges the rich too high a price for local public goods, they’ll move to a jurisdiction with lower taxes, and explicit redistribution is ruled out by the exit option.

Garett Jones has pointed out that a comparison of state tax regimes with federal tax regimes seems to support (loosely) the Tiebout model. See “Can Progressivity Survive Exit?”:

I should note though, that while state taxation is regressive in percentages, it’s progressive in dollars.  And that’s the point of my tweet: Higher earners pay more than lower earners even though they could leave.  Perhaps some of that is altruism, but I suspect-without-proof that most of it is just that the rich (and middle class) buy more and better government services than the poor.
It’s possible that progressive income taxation could coexist with voluntary competitive government. Maybe the high-skilled need to be near each other to produce a lot, so the locales preferred by the rich can tax that demand for proximity.  The (not very progressive) New York City income tax comes to mind.But at the national level, I suspect that the reason the rich pay higher total tax rates is mostly because it’s hard to leave the nation.  Easy targets.

Jones also points to the French government’s retreat on capital gains taxes as a victory for “[the] Tiebout [model against] progressivity” (I think that’s what the title of his post means). And here he entertains the suggestion that the fact that European tax systems are much less progressive than America’s is explained by Tiebout competition. Since the EU has open borders, wealthy elites can shop among jurisdictions. I don’t know enough about European tax politics to affirm or deny the causal link here, but it fits the theory. An international comparison of corporate tax rates also suggests that Tiebout competition is at work. From the Tax Foundation’s blog, here are “Corporate Income Tax Rates Around the World.”

Country Corporate Tax Rate in 2000[1] Rank in 2000 Corporate Tax Rate in 2006 Rank in March 2006
Japan 40.9 3 39.5 1
United States[2] 39.4 6 39.3 2
Germany 52 1 38.9 3
Canada 44.6 2 36.1 4
France 37.8 7 35 5
Spain 35 11 35 5
Belgium 40.2 4 34 7
Italy 37 9 33 8
New Zealand 33 16 33 8
Greece 40 5 32 10
Netherlands 35 11 31.5 11
Luxembourg 37.5 8 30.4 12
Mexico 35 11 30 13
Australia 34 14 30 13
Turkey 33 16 30 13
United Kingdom 30 21 30 13
Denmark 32 18 28 17
Norway 28 26 28 17
Sweden 28 26 28 17
Portugal 35.2 10 27.5 20
Korea 30.8 20 27.5 20
Czech Republic 31 19 26 22
Finland 29 24 26 22
Austria 34 14 25 24
Switzerland 24.9 28 21.3 25
Poland 30 21 19 26
Slovak Republic 29 24 19 26
Iceland 30 21 18 28
Hungary 18 30 16 29
Ireland 24 29 12.5 30
OECD Average[3] 33.6 28.7

Note that the US has the second-highest (after Japan) corporate income tax rate in the OECD. This is counter-intuitive, since ideologically the US has a reputation for being free-marketeer and pro-business, with a comparatively thin welfare state and less regulation. But if international Tiebout competition is at work, this pattern is sort of what you’d expect. The US is big and geographically isolated, so corporations based in the US won’t find it very easy to hop the border. European countries face steeper competition from other jurisdictions, so companies are relatively mobile.

Open borders would make international Tiebout competition a more effective force for disciplining governments’ redistributive impulses. But here I don’t mean primarily unilateral open borders, nor open borders with very poor countries. The US government is not likely to cut taxes on the wealthy for fear that they’ll emigrate to Rwanda, or even Mexico. But it might someday cut taxes to keep the wealthy from moving to Hong Kong, or Italy, or some yet-to-be-founded futuristic free charter city. If the right to emigrate were made a global reality, international Tiebout competition might start to matter a lot.

Continue reading “International Tiebout competition” »

Garett Jones responds to my intelligence post

[UPDATE: Check out Nathan’s related post Immigration and Institutions]

In my previous post on intelligence, international development, and immigration, I referenced some of the writings of Garett Jones. I sent an email to Jones asking him for his thoughts on the post, and he replied with the following email:

A very nice, thoughtful post. Thank you for giving my writings a careful reading. Much appreciated.

I’d have one critique of your claim, one that is common to many supporters of freer low-skill immigration.

You claim that institutions are important, something I agree with. And you claim that low IQ populations tend to have bad institutions, partly because of the low IQ population, again something I agree with.

But from there you conclude that low-IQ immigrants should be allowed to come to countries with good institutions. That might be reasonable as a moral case but I’m no expert on morality so I’ll leave that to others.

I would emphasize a different conclusion: That the low-IQ immigrants will tend to worsen the institutions of the higher-IQ countries they move to. Low IQ immigrants will, to some degree, tend to make the country they move to more like the country they came from.

Partly this will be due to MRV and Caplan/Miller reasons: low IQ groups vote for bad policies. Partly it’s because they will tend to elect individuals from their constituencies, which will, on average, tend to lower the average IQ of the legislature. And partly it’s because the bureaucracy will tend to hire individuals from low-skill groups, which will lower government quality.

For these and other reasons, new low IQ citizens impose a tax on the nation’s institutions, and this institutional cost should be counted in a candid cost-benefit analysis.

*Shorter version: Good institutions are rare treasures, and institutions are endogenous with respect to (among other things) citizen IQ. *
Again, many thanks for drawing attention to my work, much appreciated.

I think Jones is correct, particularly for those aspects of institutions that are determined through electoral political processes (for more on the political externalities arguments made by restrictionists, see political externalities). I should have acknowledged more explicitly in my original post. This concern would have less applicability to market processes or to those aspects of the law that are deeply entrenched and less subject to political change. [UPDATE: Nathan Smith’s comment below reminded me that I should mention the following: even for those of you who consider the political externalities case to be serious, there are keyhole solutions to the problem such as guest worker programs that allow people to migrate to work but don’t given them voting rights. The focus of this post, however, is to consider the strength of the concern per se, not to propose remedies.]

I still stand by the key point of the original post, namely, that sustaining high quality institutions is a lot easier than creating high quality institutions, and even low IQ people would be able to discern that institutions in the country they migrate to are better than institutions in the country they migrated from, which would limit (but not eliminate) their desire to recreate the situation of their source country.

To make my point a little clearer, I’m arguing that it’s a lot harder to improve a country’s poor institutions by importing a lot of high IQ people than it is to sustain a country’s good institutions even allowing low IQ immigration (importing the institutions themselves might work — that’s the hope behind charter cities). In a sense, I’m arguing that institutions have their own inertia. This argument can be thought of as a version of status quo bias, and it has been made by Bryan Caplan in his digest version of the political externalities of open borders, where he writes (emphasis added by me):

2. The political effect of immigrants on markets and liberty is at worst modestly negative. The median American isn’t a libertarian, and the median immigrant isn’t a Stalinist. We’re talking about marginal disagreements between social democrats, nothing more. Immigrants’ low voter turnout and status quo bias further dilute immigrants’ negative political effect.

I’m actually arguing something slightly stronger: institutions have their own inertia, but good institutions have more inertia than poor institutions, even with low IQ populations, because people can see the results and tell the difference, at least when it’s sufficiently dramatic. They may misdiagnose the causes, and may even misjudge minor differences. But they’re unlikely to undo all the gains achieved through improved institutions.

In Jones’ language, my framing of his assertions that “institutions are endogenous to (among other things) a country’s IQ” would be that it is changes in institutions that are endogenous to a country’s IQ.

That said, I do agree with Jones that, viewed solely from the angle of the quality of institutions in the target country, immigration of low IQ people could have a negative impact, or, even if not a direct negative impact, an “opportunity cost” (i.e., institutions don’t improve as rapidly as they otherwise might).

Intelligence, international development, and immigration

For background reading on the topic of IQ as an objection to immigration, see IQ deficit.

I recently learned from Arnold Kling’s blog post of a new book by Richard Lynn and Tatu Vanhanen titled Intelligence: A Unifying Construct for the Social Sciences (buy here). The book is an extension of earlier work by Lynn and Vanhanen, including IQ and the Wealth of Nations (Wikipedia page).

In IQ and the Wealth of Nations Lynn and Vanhanen introduce the concept of “national IQ” — the average IQ of a nation — and then attempt to demonstrate that national IQ is correlated with a number of measures of national per capita wealth. They then try to argue that at least part of the correlation is causal from IQ to per capita wealth. Controlling for IQ, they find that the extent to which an economy is a free market economy is the best predictor of national wealth. Roughly, they contend that national IQs explain about 1/3 of the variation in national wealth, market orientation explains another 1/3 of the variation, and the remaining 1/3 is explained by a host of other factors (which they don’t attempt to enumerate in full).

In Intelligence, Lynn and Vanhanen extend the analysis beyond wealth to various other measures of well being including health measures, water access, democratization, crime, and happiness. They argue that IQ can explain a significant portion of each of these (though in some cases it is not as significant) and conventional explanations such as market orientation and specific historical events can account for some of the residual. Their overall thesis is that intelligence should be treated as a unifying construct and explanatory variable across a wide range of social sciences, akin to the way that concepts from physics have explanatory power across all domains of the natural sciences.

While L&V’s thesis is new, I think that they make reasonable arguments and attempt to address all the prima facie objections one may have. How well they succeed, and whether their thesis withstands further empirical assault, is not something I feel confident to comment upon. However, I think that L&V sometimes draw the wrong conclusions from their data on the rare occasions that they try to discuss the implications for international development.

Although I don’t have any credentials in this area, I’ve relied, in addition to L&V, on the research of Garett Jones, who largely agrees with the L&V framework but tries to dig deeper into the mechanisms by which IQ might play a causal role in creating wealth. While the synthesis I present is largely my own, it relies on Jones’ work to quite an extent.

Also note: my critique of some of the conclusions that L&V (and others) draw from their work presupposes, for simplicity’s sake, that L&V’s overall framework is correct. Even if it isn’t, and IQ is not as powerful an explanatory variable as claimed, my arguments may still work in a modified sense (replacing IQ by whatever X factor is driving national differences).

National IQs versus individual IQs

Jones, L&V, and many other students of national IQs have argued that there is a relationship between national IQ and economic measures, and that this relationship is logarithmic: a one point increase in national IQ leads to a fixed proportional increase in productivity, hence also in per capita GDP and other measures. However, one of the remarkable findings is that the effects at the national level are much more salient than the effects at the individual level. Continue reading “Intelligence, international development, and immigration” »