Yes! It literally just shifts the supply curve! Both producers and consumers pay depending on the elasticity of demand and supply. I don’t know how economists bear the standard of public conversation about this sort of stuff. At least with philosophy the arguments people get tripped up over are usually somewhat hard and/or subtle.
There’s literally a simple picture that explains all of this!
Capital and labor are substitutes in the short run but economic complements in the long run.
The equilibrium wage rate for a particular laborer is the marginal revenue product of that person's labor. In non-economic terms, that mean how much the person can contribute to his employer's productivity. If the wage rate were lower than that, competition would bid it up as competitors hire the person at a higher wage rate. If it were higher than that, the employer would want to reduce production (or ultimately might go out of business).
Capital increases the productivity of labor, and thus increases the wage rate. THIS is the reason the standard of living in the US and a lot of industrialized countries became so much more wealthy over the past (say) 250 years. We have added to the capital stock: physical machines, inventions, roads, electricity, and most importantly human capital. That has increased the wage rate and all the increased productivity increased the standard of living for everyone. Not equally of course, but by a huge amount for everyone.
The short run effect of a change in capital is to substitute capital for labor. But in the long run, lower production costs lead to more production. Imagine a bunch of people with shovels. They can dig a lot. But them someone invents a bulldozer. Now one person with a bulldozer can do the work of (say) 100 people with shovels. Companies hire 1 person with the bulldozer instead of 100 people with shovels. (That person has a higher wage rate though.)
Not only are costs now lower, but the type of digging the bulldozer can do it much greater. So people come up with new digging projects. So a construction company can dig a space for an 8 story underground parking structure that is under a 60-story building with a bunch of people with bulldozers. That project was not feasible in the guys-with-shovels era. So you get an increase in demand for digging in new and unexpected areas. And the benefits accrue to lots of people: bulldozer manufactures, laborers driving the bulldozers, the owners and residents of the new building, etc.
The same will be true with AI. As it is used more and more, people will use AI to do work more efficiently that others do now. There will be a substitution effect. But then people will start finding new uses for AI. And of benefits of both of these will be more broadly spread through the economy. There will be both short-term disruption and long-term benefits for lots of people.
"Capital and labor are substitutes in the short run but economic complements in the long run." OK, let's stop there, as we are back in the realm of abstraction. In the long run, everyone dies and their personal economies stop mattering to them. But in the short run, humans have lives and livelihoods that they really hate being made precarious, e.g. during a 'substitution' period. *If* AI causes lots of unemployment, telling the victims that 'it'll be better in the long run', and not expecting mass discontent in response...well, good luck with that. How do you manage it?
That's a good question. And it is exacerbated by the fact that AI is likely to change things rapidly.
In a dynamic economy, things change. That's the way it is. Businesses are constantly folding. New businesses are opening. Every neighborhood has tons of restaurants that used by be a different restaurant. Blockbuster drove all the little video rental stores out of business, but could not compete with streaming, and is now gone. Pan Am and TWA were powerhouses and are now gone. The American auto industry suffered huge losses including unemployment and has adapted. There are no more mule train drivers pulling barges up the Erie Canal from Buffalo so that NY bakeries can get cheap wheat. Joseph Schumpeter called competition the "gale of creative destruction."
We have some things that ameliorate this. Unemployment insurance, new job training, etc. But there certainly will be disruption, loss of jobs, and unhappy people.
I don't have a plan to manage it, but there really is no alternative. The alternative to one stare of the world has to be another feasible state of the world. not some hypothetical world where something bad does not happen. What do you suggest?Pass a law prohibiting technological improvements to protect jobs? Bring back barges on the Erie Canal?
I really enjoyed the discussion between Steve and Bruce. They landed on the key issue for plants, animals, humans, and societies as a whole. What happens when the rate of change exceeds any organism or groups ability to adapt?
In the anthropocene, we are seeing that there is an acceleration in extinctions.
At the individual human level, we see bad things like unemployment, starvation, depression, and suicide, and good things like incredible individual prosperity for those who adapt quickly or just happen to be well-positioned for the future state. Some humans avoid the worse by pursuing traditional human vocations in craft or small-scale farming, but they for the most part do not participate in the prosperity.
At the societal level, I believe we see a net increase in prosperity as the society approaches its maximum change velocity. When maximum societal change velocity is exceeded, the result will be similar to the result for individual humans.
So, dampening the increase in the velocity of change is an imperative for nature, humans, and societies.
There's a cost to dampening as well. If we lower the productivity growth rate, all sorts of beneficial discoveries are delayed. For example, a previous unknown cure for cancer might happen in 20 years rather than 5 years, and that means 15 years of more people dying from this form of cancer. And more generally, if the real income and wealth of low-income people increase at a higher rate, that can lead to all sorts of actual benefits.
It will take a lot of thinking to determine the right path.
In practical terms, if dampening is warranted, how could we do this? Massive amounts of private equity funds are pouring into AI. Is there a feasible law or regulation that could be passed. (Limit the amount of AI investment, properly defined? Tax AI investment or profits?) None of that seems feasible.
Obsoleting barge-pullers on the Erie canal only put a limited set of skills out of work, and only did it locally. Even nationally then, I doubt barge-pulling ranked where, say, software engineering, does now. (I am not a software engineer). But your point that 'things change' is of course true ...and not uncommonly made. And that is my point.
I would say, especially to 'realists' and 'the rationalist community' (I am not saying you are of them, and I've lost track of where Bentham's B is at the moment, but your arguments ring of them), *instead of* reminding voters that there is 'no alternative' and that they must inevitably ride the tide of history, think about, write about, focus really, really, really hard about those 'things that ameliorate this'. Oh, and, yeah, regulation, that naughty word, we could use some of that.
I get the impression that working to make things concretely 'less bad in the short run' is considered unacceptably at odds with the lofty goal of being 'less wrong' in the long run.
I'm open to any reasonable suggestion to ameliorate problems in the short run. I'm not sure what they are. You suggest "regulation." What would you suggest?
People saying “but aren’t the tariffs paid by Americans?” are not making a distinction between consumers and producers (or in this case importers), but rather objecting to the nonsensical claim that tariffs are a way of getting foreign entities to subsidize the American government. And your first year Econ 101 argument is irrelevant, as the tariff is only applied to a segment of the actual producers' market. It may almost work for a product manufactured exclusively for the US market but is just nonsense for a product which can find alternative consumers elsewhere.
Total US exports in 2024 were about 3.2T while imports were about 4.1T, for a trade deficit of about .9T. In general, we have been seeing our goods trade deficit increase while our services trade surplus also increases.
Globally, the US accounts for roughly 1/4 of global GDP, an even greater share of global profits, and more than 1/2 of total global market caps.
Our services sector is much more profitable than our goods sector, while much of our deindustrialization stems from a focus on high-margin, low-capital industries and the export of low-margin, high-capital ones.
Trump's tariffs may be an effective way to end all of this. But the associated grift is enriching him, his family, and, to a lesser extent, his allies. This is how s***hole countries are made.
And our increasingly lousy distribution of wealth also means that the fruits of our labors accrue to a smaller portion of our population. The wealthiest among us hate wealth redistribution, so they fund politicians who are either useful idiots or grifters. Tariffs appeal to the widely held belief that we were better off when we made more things here. In reality, factory workers used to receive a greater share of the profits.
Among my favorite solutions are public campaign financing, a return to more progressive taxation (even Reagan-era rates would be an improvement), and a much more aggressive death tax. Of course, in 2025, we made things worse in all three areas.
I’ve never once heard someone say it’s Americans paying for the tariffs and thought the speaker meant anything other than Americans will have to pay more money for the same goods because of the tariffs. You are interpreting these speakers literally when you shouldn’t be.
Exactly! But I wonder if you’re maybe too smart to grok just how low a level the debate is being held at. It appears you’re arguing against a strawman that’s a bit more robust than the real thing being argued. There is currently no room for your kind of sophistication in public-facing politics in the US or much of the rest of the world.
I’m not really capable of giving the strongest version of the pro-tariff case, but it goes something like:
Tariffs will make foreign products less competitive, which will incentivize people to buy American instead. And since there aren’t tariffs on American goods, American companies will become more competitive (and they will hire more people) without anything becoming more expensive.
Also, everyone else will still want to sell their stuff in the US, and will have to pay tariffs to do so. In order to keep competitive prices, they’ll eat the increased cost. Again, nothing becomes more expensive for consumers.
Finally, all the extra income that the government collects from foreign companies who care more about market access than sustainable profits will be sent to American citizens in the form of a check, signed in sharpie by DJT.
So the story in favor is that tariffs will benefit consumers massively and directly, and prices won’t go up.
This is so obviously wrong and dumb that anyone who believes it – especially after seeing how it has been implemented – is either an idiot or an ideologue. (Also, some dishonest grifters and climbers may promote this story even if they don’t actually believe it.)
The counter-argument you take issue with makes the same point you’ve made, but in a way they hope will resonate with the idiots who have bought into the pro-tariff argument (not the ideologues and grifters), when they inevitably notice prices going up.
No one thinks regular consumers will get a bill in the mail for tariffs on stuff they buy at the store; and businesses and people who get a bill from customs for their imports know who pays that bill. That’s not the question.
When people argue over who ends up paying the tariffs, it’s an attempt to bring home the most obvious business fact in the world: That all costs in any sustainable business will ultimately have to be paid by the customer.
But it has to be articulated so that it resonates with people who don’t think about the economy, but just feel it. (I’m not sure it’s going great so far.) It’s not really a point about economics or practicalities, but about emotions and narratives.
I am a PhD economist. The only way to understand the reasoning behind the Trump tariffs entails eating paint chips like they are Doritos.
The tariffs are for extortion. They are self inflicted harm in order for Trump to gain some sort of concessions from third parties. The more he does it the more it comes off as crying wolf. He just threatened 200% tariffs on French wine because he felt snubbed by Macron. The failure of this strategy is Trump is going to punish you regardless of what you do, so no one will cooperate.
I'm assuming you mean "increased production" (as in GDP), not productivity, which is a measure of how much production is generated by the same amount of inputs (land, labor, capital). Increased productivity is great, certainly, but it's not the primary goal here, it's only good instrumentally by increasing long run supply, and thus long run production.
1. Politically, it does matter who pays, specifically because people are dumb. Republicans are supposed to be the anti-tax party—ie, less taxes on Americans. Since their leaders know people are stupid, they are putting huge amounts of effort to convince people it’s not taxes that Americans themselves pay. That’s also why journalists, pundits, etc. are pushing back on that nonsense. As you show, that may be a worse strategy than simply focusing on the resultant higher costs, of course. On the other hand, people do hate hypocrisy.
2. “What matters is boosted productivity, not how many people have jobs, especially when unemployment is very low.” That seems like a pretty radical statement.
If you delete “, especially,” it’s obvious. Without that, it seems you’re fine in our current era with high unemployment. That day will presumably come from automation at some point, and hopefully we’ll be ready with the requisite social welfare to support it, but we’re so far from that place at the moment that Great Depression-era unemployment (eg) would be absolutely devastating. It very much matters how many people have jobs—or, rather, what percentage of people do.
Yes! It literally just shifts the supply curve! Both producers and consumers pay depending on the elasticity of demand and supply. I don’t know how economists bear the standard of public conversation about this sort of stuff. At least with philosophy the arguments people get tripped up over are usually somewhat hard and/or subtle.
There’s literally a simple picture that explains all of this!
"What matters is boosted productivity, not how many people have jobs, especially when unemployment is very low)."
OK, but *what if* productivity increases thanks to AI, but so does unemployment? Underneath the abstraction there are humans, and they get angry.
Capital and labor are substitutes in the short run but economic complements in the long run.
The equilibrium wage rate for a particular laborer is the marginal revenue product of that person's labor. In non-economic terms, that mean how much the person can contribute to his employer's productivity. If the wage rate were lower than that, competition would bid it up as competitors hire the person at a higher wage rate. If it were higher than that, the employer would want to reduce production (or ultimately might go out of business).
Capital increases the productivity of labor, and thus increases the wage rate. THIS is the reason the standard of living in the US and a lot of industrialized countries became so much more wealthy over the past (say) 250 years. We have added to the capital stock: physical machines, inventions, roads, electricity, and most importantly human capital. That has increased the wage rate and all the increased productivity increased the standard of living for everyone. Not equally of course, but by a huge amount for everyone.
The short run effect of a change in capital is to substitute capital for labor. But in the long run, lower production costs lead to more production. Imagine a bunch of people with shovels. They can dig a lot. But them someone invents a bulldozer. Now one person with a bulldozer can do the work of (say) 100 people with shovels. Companies hire 1 person with the bulldozer instead of 100 people with shovels. (That person has a higher wage rate though.)
Not only are costs now lower, but the type of digging the bulldozer can do it much greater. So people come up with new digging projects. So a construction company can dig a space for an 8 story underground parking structure that is under a 60-story building with a bunch of people with bulldozers. That project was not feasible in the guys-with-shovels era. So you get an increase in demand for digging in new and unexpected areas. And the benefits accrue to lots of people: bulldozer manufactures, laborers driving the bulldozers, the owners and residents of the new building, etc.
The same will be true with AI. As it is used more and more, people will use AI to do work more efficiently that others do now. There will be a substitution effect. But then people will start finding new uses for AI. And of benefits of both of these will be more broadly spread through the economy. There will be both short-term disruption and long-term benefits for lots of people.
"Capital and labor are substitutes in the short run but economic complements in the long run." OK, let's stop there, as we are back in the realm of abstraction. In the long run, everyone dies and their personal economies stop mattering to them. But in the short run, humans have lives and livelihoods that they really hate being made precarious, e.g. during a 'substitution' period. *If* AI causes lots of unemployment, telling the victims that 'it'll be better in the long run', and not expecting mass discontent in response...well, good luck with that. How do you manage it?
That's a good question. And it is exacerbated by the fact that AI is likely to change things rapidly.
In a dynamic economy, things change. That's the way it is. Businesses are constantly folding. New businesses are opening. Every neighborhood has tons of restaurants that used by be a different restaurant. Blockbuster drove all the little video rental stores out of business, but could not compete with streaming, and is now gone. Pan Am and TWA were powerhouses and are now gone. The American auto industry suffered huge losses including unemployment and has adapted. There are no more mule train drivers pulling barges up the Erie Canal from Buffalo so that NY bakeries can get cheap wheat. Joseph Schumpeter called competition the "gale of creative destruction."
We have some things that ameliorate this. Unemployment insurance, new job training, etc. But there certainly will be disruption, loss of jobs, and unhappy people.
I don't have a plan to manage it, but there really is no alternative. The alternative to one stare of the world has to be another feasible state of the world. not some hypothetical world where something bad does not happen. What do you suggest?Pass a law prohibiting technological improvements to protect jobs? Bring back barges on the Erie Canal?
I really enjoyed the discussion between Steve and Bruce. They landed on the key issue for plants, animals, humans, and societies as a whole. What happens when the rate of change exceeds any organism or groups ability to adapt?
In the anthropocene, we are seeing that there is an acceleration in extinctions.
At the individual human level, we see bad things like unemployment, starvation, depression, and suicide, and good things like incredible individual prosperity for those who adapt quickly or just happen to be well-positioned for the future state. Some humans avoid the worse by pursuing traditional human vocations in craft or small-scale farming, but they for the most part do not participate in the prosperity.
At the societal level, I believe we see a net increase in prosperity as the society approaches its maximum change velocity. When maximum societal change velocity is exceeded, the result will be similar to the result for individual humans.
So, dampening the increase in the velocity of change is an imperative for nature, humans, and societies.
There's a cost to dampening as well. If we lower the productivity growth rate, all sorts of beneficial discoveries are delayed. For example, a previous unknown cure for cancer might happen in 20 years rather than 5 years, and that means 15 years of more people dying from this form of cancer. And more generally, if the real income and wealth of low-income people increase at a higher rate, that can lead to all sorts of actual benefits.
It will take a lot of thinking to determine the right path.
In practical terms, if dampening is warranted, how could we do this? Massive amounts of private equity funds are pouring into AI. Is there a feasible law or regulation that could be passed. (Limit the amount of AI investment, properly defined? Tax AI investment or profits?) None of that seems feasible.
After I wrote this, I saw the following Op Ed in the NYT on AI and medical care entitled "Stop Worrying, and Let A.I. Help Save Your Life".
https://www.nytimes.com/2026/01/19/opinion/ai-health-medical-care.html?unlocked_article_code=1.F1A.Ts_-.KgdkCdkC8NSy&smid=url-share
Obsoleting barge-pullers on the Erie canal only put a limited set of skills out of work, and only did it locally. Even nationally then, I doubt barge-pulling ranked where, say, software engineering, does now. (I am not a software engineer). But your point that 'things change' is of course true ...and not uncommonly made. And that is my point.
I would say, especially to 'realists' and 'the rationalist community' (I am not saying you are of them, and I've lost track of where Bentham's B is at the moment, but your arguments ring of them), *instead of* reminding voters that there is 'no alternative' and that they must inevitably ride the tide of history, think about, write about, focus really, really, really hard about those 'things that ameliorate this'. Oh, and, yeah, regulation, that naughty word, we could use some of that.
I get the impression that working to make things concretely 'less bad in the short run' is considered unacceptably at odds with the lofty goal of being 'less wrong' in the long run.
I'm open to any reasonable suggestion to ameliorate problems in the short run. I'm not sure what they are. You suggest "regulation." What would you suggest?
Certain AI techlords themselves were advocating for regulation, at one point -- what did they suggest?
People saying “but aren’t the tariffs paid by Americans?” are not making a distinction between consumers and producers (or in this case importers), but rather objecting to the nonsensical claim that tariffs are a way of getting foreign entities to subsidize the American government. And your first year Econ 101 argument is irrelevant, as the tariff is only applied to a segment of the actual producers' market. It may almost work for a product manufactured exclusively for the US market but is just nonsense for a product which can find alternative consumers elsewhere.
Total US exports in 2024 were about 3.2T while imports were about 4.1T, for a trade deficit of about .9T. In general, we have been seeing our goods trade deficit increase while our services trade surplus also increases.
Globally, the US accounts for roughly 1/4 of global GDP, an even greater share of global profits, and more than 1/2 of total global market caps.
Our services sector is much more profitable than our goods sector, while much of our deindustrialization stems from a focus on high-margin, low-capital industries and the export of low-margin, high-capital ones.
Trump's tariffs may be an effective way to end all of this. But the associated grift is enriching him, his family, and, to a lesser extent, his allies. This is how s***hole countries are made.
And our increasingly lousy distribution of wealth also means that the fruits of our labors accrue to a smaller portion of our population. The wealthiest among us hate wealth redistribution, so they fund politicians who are either useful idiots or grifters. Tariffs appeal to the widely held belief that we were better off when we made more things here. In reality, factory workers used to receive a greater share of the profits.
Among my favorite solutions are public campaign financing, a return to more progressive taxation (even Reagan-era rates would be an improvement), and a much more aggressive death tax. Of course, in 2025, we made things worse in all three areas.
I’ve never once heard someone say it’s Americans paying for the tariffs and thought the speaker meant anything other than Americans will have to pay more money for the same goods because of the tariffs. You are interpreting these speakers literally when you shouldn’t be.
Exactly! But I wonder if you’re maybe too smart to grok just how low a level the debate is being held at. It appears you’re arguing against a strawman that’s a bit more robust than the real thing being argued. There is currently no room for your kind of sophistication in public-facing politics in the US or much of the rest of the world.
I’m not really capable of giving the strongest version of the pro-tariff case, but it goes something like:
Tariffs will make foreign products less competitive, which will incentivize people to buy American instead. And since there aren’t tariffs on American goods, American companies will become more competitive (and they will hire more people) without anything becoming more expensive.
Also, everyone else will still want to sell their stuff in the US, and will have to pay tariffs to do so. In order to keep competitive prices, they’ll eat the increased cost. Again, nothing becomes more expensive for consumers.
Finally, all the extra income that the government collects from foreign companies who care more about market access than sustainable profits will be sent to American citizens in the form of a check, signed in sharpie by DJT.
So the story in favor is that tariffs will benefit consumers massively and directly, and prices won’t go up.
This is so obviously wrong and dumb that anyone who believes it – especially after seeing how it has been implemented – is either an idiot or an ideologue. (Also, some dishonest grifters and climbers may promote this story even if they don’t actually believe it.)
The counter-argument you take issue with makes the same point you’ve made, but in a way they hope will resonate with the idiots who have bought into the pro-tariff argument (not the ideologues and grifters), when they inevitably notice prices going up.
No one thinks regular consumers will get a bill in the mail for tariffs on stuff they buy at the store; and businesses and people who get a bill from customs for their imports know who pays that bill. That’s not the question.
When people argue over who ends up paying the tariffs, it’s an attempt to bring home the most obvious business fact in the world: That all costs in any sustainable business will ultimately have to be paid by the customer.
But it has to be articulated so that it resonates with people who don’t think about the economy, but just feel it. (I’m not sure it’s going great so far.) It’s not really a point about economics or practicalities, but about emotions and narratives.
I am a PhD economist. The only way to understand the reasoning behind the Trump tariffs entails eating paint chips like they are Doritos.
The tariffs are for extortion. They are self inflicted harm in order for Trump to gain some sort of concessions from third parties. The more he does it the more it comes off as crying wolf. He just threatened 200% tariffs on French wine because he felt snubbed by Macron. The failure of this strategy is Trump is going to punish you regardless of what you do, so no one will cooperate.
"What matters is increased productivity..."
I'm assuming you mean "increased production" (as in GDP), not productivity, which is a measure of how much production is generated by the same amount of inputs (land, labor, capital). Increased productivity is great, certainly, but it's not the primary goal here, it's only good instrumentally by increasing long run supply, and thus long run production.
Agreed with most of this. Two quibbles, though:
1. Politically, it does matter who pays, specifically because people are dumb. Republicans are supposed to be the anti-tax party—ie, less taxes on Americans. Since their leaders know people are stupid, they are putting huge amounts of effort to convince people it’s not taxes that Americans themselves pay. That’s also why journalists, pundits, etc. are pushing back on that nonsense. As you show, that may be a worse strategy than simply focusing on the resultant higher costs, of course. On the other hand, people do hate hypocrisy.
2. “What matters is boosted productivity, not how many people have jobs, especially when unemployment is very low.” That seems like a pretty radical statement.
If you delete “, especially,” it’s obvious. Without that, it seems you’re fine in our current era with high unemployment. That day will presumably come from automation at some point, and hopefully we’ll be ready with the requisite social welfare to support it, but we’re so far from that place at the moment that Great Depression-era unemployment (eg) would be absolutely devastating. It very much matters how many people have jobs—or, rather, what percentage of people do.