Reading Bryan Caplan and Zach Weinersmith’s new book has made me somewhat more skeptical about Open Borders (from a high prior belief in its value).
Before reading the book, I was already aware of the core arguments (eg, Michael Huemer’s right to immigrate, basic cosmopolitanism, some vague economic stuff about doubling GDP).
I was hoping the book will have more arguments, or stronger versions of the arguments I’m familiar with.
It mostly did not.
The book did convince me that the prima facie case for open borders was stronger than I thought. In particular, the section where he argued that a bunch of different normative ethical theories should all-else-equal lead to open borders was moderately convincing. I think it will have updated me towards open borders if I believed in stronger “weight all mainstream ethical theories equally” moral uncertainty, or if I previously had a strong belief in a moral theory that I previously believed was against open borders.
However, I already fairly strongly subscribe to cosmopolitan utilitarianism and see no problem with aggregating utility across borders. Most of my concerns with open borders are related to Chesterton’s fence, and Caplan’s counterarguments were in three forms:
1. Doubling GDP is so massive that it should override any conservativism prior. 2. The US historically had Open Borders (pre-1900) and it did fine. 3. On the margin, increasing immigration in all the American data Caplan looked at didn’t seem to have catastrophic cultural/institutional effects that naysayers claim.
I find this insufficiently persuasive. ___ Let me outline the strongest case I’m aware of against open borders: Countries are mostly not rich and stable because of the physical resources, or because of the arbitrary nature of national boundaries. They’re rich because of institutions and good governance. (I think this is a fairly mainstream belief among political economists). These institutions are, again, evolved and living things. You can’t just copy the US constitution and expect to get a good government (IIRC, quite a few Latin American countries literally tried and failed).
We don’t actually understand what makes institutions good. Open Borders means the US population will ~double fairly quickly, and this is so “out of distribution” that we should be suspicious of the generalizability of studies that look at small marginal changes. ____ I think Caplan’s case is insufficiently persuasive because a) it’s not hard for me to imagine situations bad enough to be worse than doubling GDP is good, 2)Pre-1900 US was a very different country/world, 3) This “out of distribution” thing is significant.
I will find Caplan’s book more persuasive if he used non-US datasets more, especially data from places where immigration is much higher than the US (maybe within the EU or ASEAN?).
___
I’m still strongly in favor of much greater labor mobility on the margin for both high-skill and low-skill workers. Only 14.4% of the American population are immigrants right now, and I suspect the institutions are strong enough that changing the number to 30-35% is net positive. [EDIT: Note that this is intuition rather than something backed by empirical data or explicit models]
I’m also personally in favor (even if it’s negative expected value for the individual country) of a single country (or a few) trying out open borders for a few decades and for the rest of us to learn from their successes and failures. But that’s because of an experimentalist social scientist mindset where I’m perfectly comfortable with “burning” a few countries for the greater good (countries aren’t real, people are), and I suspect the government of most countries aren’t thrilled about this.
___
Overall, 4⁄5 stars. Would highly recommend to EAs, especially people who haven’t thought much about the economics and ethics of immigration.
If you email this to him, maybe adding a bit more polish, I’d give ~40% odds he’ll reply on his blog, given how much he loves to respond to critics who take his work seriously.
It’s not hard for me to imagine situations bad enough to be worse than doubling GDP is good
I actually find this very difficult without envisioning extreme scenarios (e.g. a dark-Hansonian world of productive-but-dissatisfied ems). Almost any situation with enough disutility to counter GDP doubling seems like it would, paradoxically, involve conditions that would reduce GDP (war, large-scale civil unrest, huge tax increases to support a bigger welfare state).
Could you give an example or two of situations that would fit your statement here?
Almost any situation with enough disutility to counter GDP doubling seems like it would, paradoxically, involve conditions that would reduce GDP (war, large-scale civil unrest, huge tax increases to support a bigger welfare state).
I think there was substantial ambiguity in my original phrasing, thanks for catching that!
I think there are at least four ways to interpret the statement.
It’s not hard for me to imagine situations bad enough to be worse than doubling GDP is good
1. Interpreting it literally: I am physically capable (without much difficulty) of imagining situations that are bad to a degree worse than doubling GDP is good.
2. Caplan gives some argument for doubling of GDP that seems persuasive, and claims this is enough to override a conservatism prior, but I’m not confident that the argument is true/robust, and I think it’s reasonable to believe that there are possible bad consequences that are bad enough that even if I give >50% probability (or >80%), this is not automatically enough to override a conservatism prior, at least not without thinking about it a lot more.
3. Assume by construction that world GDP will double in the short term. I still think there’s a significant chance that the world will be worse off.
4. Assume by construction that world GDP will double, and stay 2x baseline until the end of time. I still think there’s a significant chance that the world will be worse off.
__
To be clear, when writing the phrasing, I meant it in terms of #2. I strongly endorse #1 and tentatively endorse #3, but I agree that if you interpreted what I meant as #4, what I said was a really strong claim and I need to back it up more carefully.
Makes sense, thanks! The use of “doubling GDP is so massive that...” made me think that you were taking that as given in this example, but worrying that bad things could result from GDP-doubling that justified conservatism. That was certainly only one of a few possible interpretations; I jumped too easily to conclusions.
cross-posted from Facebook.
Reading Bryan Caplan and Zach Weinersmith’s new book has made me somewhat more skeptical about Open Borders (from a high prior belief in its value).
Before reading the book, I was already aware of the core arguments (eg, Michael Huemer’s right to immigrate, basic cosmopolitanism, some vague economic stuff about doubling GDP).
I was hoping the book will have more arguments, or stronger versions of the arguments I’m familiar with.
It mostly did not.
The book did convince me that the prima facie case for open borders was stronger than I thought. In particular, the section where he argued that a bunch of different normative ethical theories should all-else-equal lead to open borders was moderately convincing. I think it will have updated me towards open borders if I believed in stronger “weight all mainstream ethical theories equally” moral uncertainty, or if I previously had a strong belief in a moral theory that I previously believed was against open borders.
However, I already fairly strongly subscribe to cosmopolitan utilitarianism and see no problem with aggregating utility across borders. Most of my concerns with open borders are related to Chesterton’s fence, and Caplan’s counterarguments were in three forms:
1. Doubling GDP is so massive that it should override any conservativism prior.
2. The US historically had Open Borders (pre-1900) and it did fine.
3. On the margin, increasing immigration in all the American data Caplan looked at didn’t seem to have catastrophic cultural/institutional effects that naysayers claim.
I find this insufficiently persuasive.
___
Let me outline the strongest case I’m aware of against open borders:
Countries are mostly not rich and stable because of the physical resources, or because of the arbitrary nature of national boundaries. They’re rich because of institutions and good governance. (I think this is a fairly mainstream belief among political economists). These institutions are, again, evolved and living things. You can’t just copy the US constitution and expect to get a good government (IIRC, quite a few Latin American countries literally tried and failed).
We don’t actually understand what makes institutions good. Open Borders means the US population will ~double fairly quickly, and this is so “out of distribution” that we should be suspicious of the generalizability of studies that look at small marginal changes.
____
I think Caplan’s case is insufficiently persuasive because a) it’s not hard for me to imagine situations bad enough to be worse than doubling GDP is good, 2)Pre-1900 US was a very different country/world, 3) This “out of distribution” thing is significant.
I will find Caplan’s book more persuasive if he used non-US datasets more, especially data from places where immigration is much higher than the US (maybe within the EU or ASEAN?).
___
I’m still strongly in favor of much greater labor mobility on the margin for both high-skill and low-skill workers. Only 14.4% of the American population are immigrants right now, and I suspect the institutions are strong enough that changing the number to 30-35% is net positive. [EDIT: Note that this is intuition rather than something backed by empirical data or explicit models]
I’m also personally in favor (even if it’s negative expected value for the individual country) of a single country (or a few) trying out open borders for a few decades and for the rest of us to learn from their successes and failures. But that’s because of an experimentalist social scientist mindset where I’m perfectly comfortable with “burning” a few countries for the greater good (countries aren’t real, people are), and I suspect the government of most countries aren’t thrilled about this.
___
Overall, 4⁄5 stars. Would highly recommend to EAs, especially people who haven’t thought much about the economics and ethics of immigration.
Sam Enright has a longer review here.
Did you ever write to caplan about this? If not, I might send him this comment.
If you email this to him, maybe adding a bit more polish, I’d give ~40% odds he’ll reply on his blog, given how much he loves to respond to critics who take his work seriously.
I actually find this very difficult without envisioning extreme scenarios (e.g. a dark-Hansonian world of productive-but-dissatisfied ems). Almost any situation with enough disutility to counter GDP doubling seems like it would, paradoxically, involve conditions that would reduce GDP (war, large-scale civil unrest, huge tax increases to support a bigger welfare state).
Could you give an example or two of situations that would fit your statement here?
I think there was substantial ambiguity in my original phrasing, thanks for catching that!
I think there are at least four ways to interpret the statement.
1. Interpreting it literally: I am physically capable (without much difficulty) of imagining situations that are bad to a degree worse than doubling GDP is good.
2. Caplan gives some argument for doubling of GDP that seems persuasive, and claims this is enough to override a conservatism prior, but I’m not confident that the argument is true/robust, and I think it’s reasonable to believe that there are possible bad consequences that are bad enough that even if I give >50% probability (or >80%), this is not automatically enough to override a conservatism prior, at least not without thinking about it a lot more.
3. Assume by construction that world GDP will double in the short term. I still think there’s a significant chance that the world will be worse off.
4. Assume by construction that world GDP will double, and stay 2x baseline until the end of time. I still think there’s a significant chance that the world will be worse off.
__
To be clear, when writing the phrasing, I meant it in terms of #2. I strongly endorse #1 and tentatively endorse #3, but I agree that if you interpreted what I meant as #4, what I said was a really strong claim and I need to back it up more carefully.
Makes sense, thanks! The use of “doubling GDP is so massive that...” made me think that you were taking that as given in this example, but worrying that bad things could result from GDP-doubling that justified conservatism. That was certainly only one of a few possible interpretations; I jumped too easily to conclusions.
That was not my intent, and it was not the way I parsed Caplan’s argument.