Trump’s early 2018 tariffs on washing machines are costing American consumers $1.5 billion a year in higher prices, even after deducting the meager $82 million in customs revenues. The whole tariff was paid by American consumers in higher prices—in fact, even more than the whole tariff if we consider that dryers, a complement good, also increased in price. The total cost to American consumers amounts to more than $800,000 a year for each of the 1,800 jobs created in America by the tariff.
These are the main results of an econometric study by three economists, Aaron Flaaen (Federal Reserve), Ali Hortaçsu (University of Chicago and NBER), and Felix Tintelnot (same affiliations as the latter). Their article was published in the latest issue of the American Economic Review under the title “The Production Relocation and Price Effects of US Trade Policy: The Case of Washing Machines” (the article is not gated). I previously mentioned on Econlog the authors’ preliminary research on the same topic as published in March and April 2019.
Another lesson from the washing machine tariffs, which is not surprising either, is that the tariffs have had the same impact on the price of domestically produced washing machines as on imported ones. I explained in a few previous Econlog posts why a tariff always increases the price of the domestic substitute as much as the price of the targeted imported good (see, for example, “Post Scriptum on Steel Tariffs,” December 5, 2018). A simple way to see this is to realize that the very reason domestic producers ask for a tariff on the competing foreign goods is to be able to increase their prices at the level from which they had to cut them because of this competition.
Flaaen et al.’s estimates are not too far from the results of my back-of-the-envelope calculations the Spring issue of Regulation (“Putting 97 Million Households through the Wringer”). More than a praise for your humble blogger, this is a vindication of the usefulness of economic reasoning. On the legal and institutional background of the “Safeguard clause” invoked by Trump, see Lauren Landsburg, “Taken to the Cleaners,” Econlib, March 5, 2018.
READER COMMENTS
robc
Aug 10 2020 at 7:34am
How many jobs were lost due to the $1.5B lost?
At $800k per job, I am guessing 5-7 lost per job gained?
Pierre Lemieux
Aug 10 2020 at 10:42am
@robc: To avoid Keynesian pitfalls, I prefer to look at this in terms of general equilibrium, as a move on the production possibility frontier. In this perspective, after the initial shock, there is no job jost, just less consumption (through reduced trade). A bit more on this in my post “You Can’t Benefit from Free Trade If You Don’t Have a Job, Right?”
Warren Platts
Aug 10 2020 at 2:09pm
You economists continue to baffle me. Why should a high cost to consumers be a bad thing? “Cost to consumers per job” is literally means the same as “labor productivity.” Thus a high cost to consumers per job means that job is highly productive. Would it be better if instead of producing 1,800 manufacturing jobs, 18 thousand jobs were produced? Then the cost to consumer per job would only be about $80,000 per job created.
Answer: NO! If you divide the country’s GDP by the number of workers, the “cost to consumer” per American job is about $140,000. So if the cost to consumer per new washing machine manufacturing job was only $80,000 per job, then that would entail the new jobs would be less productive than the average job. The fact that 1 washing machine worker can produce $800,000 worth of GDP by theirself ought to be a cause for celebration since their labor productivity is 5.7X the national average!
Pierre Lemieux
Aug 10 2020 at 3:04pm
Warren: I fear you are a bit confused. Labor productivity is real output divided by the quantity of labor services. Following the reasoning you propose, increasing prosperity would be very easy: just let the government impose a 1,000,000% consumption tax.
Warren Platts
Aug 10 2020 at 6:20pm
Yes of course, but how is this different from “cost to consumer per worker”? For example, manufacturing gross output for 2019 was approximately $6.27 trillion and the number of manufacturing workers was about 12.8 million; thus output/worker was $490K. (I use gross output rather than value added because that is what Flaaen et al. apparently did since they were going off consumer prices.)
Now isn’t that $490K/worker not also the “cost to consumer per worker”? Assuming consumers buy all of $6.27 trillion in manufactured goods, then that must be the cost to consumer per worker that also equals the labor productivity of said workers.
Thus the narrative that the higher the cost to consumer is, the worse the effect is, I don’t buy. As you say in your comment above, overall employment is not affected (much). Thus, presumably the new 1,800 workers are coming from another sector. Most likely this would be the service sector, perhaps in the retail trade. The gross output/retail worker/cost-to-consumer was only $124K for 2019. So if the tariff can transform relatively unproductive service workers into productive manufacturing workers, that can only be a good thing–for workers, at least.
Jon Murphy
Aug 10 2020 at 7:23pm
No. Not at all.
You’re confusing inputs and outputs.
Warren Platts
Aug 10 2020 at 8:35pm
Inputs are irrelevant to the question at hand, that is whether ‘cost to consumer per worker’ means the same as ‘labor productivity’. All that matters for that question is output.
It is like GDP: there are multiple ways to calculate it. One can total up the production of all output, or one can total up the expenditures spent on all output. They should be equal. Divide by the number of workers and you have production/worker or cost-to-consumer/worker. They must be the same.
Jon Murphy
Aug 10 2020 at 8:43pm
Right, which is why your equivalence is incorrect.
Jon Murphy
Aug 10 2020 at 7:46pm
Remember, labor is an input, not an output in the production process.
Warren Platts
Aug 11 2020 at 12:52pm
Yes, of course, labor is an input to the production process. But again, that is irrelevant to the point being debated. We have an output: a consumer good, namely, washing machines. They have a certain value. Thus we can add up the value of all produced washing machines, divide by the number of workers, and thus come up with a labor productivity estimate in terms of value-of-output/worker.
Similarly, consumers spend a certain amount of money to buy all the washing machines produced by the production process. That is the total cost to consumers of washing machines. We can then divide by the number of washing machine manufacturing workers and come up with an estimate of the cost-to-consumer per worker.
value-of-output/worker = cost-to-consumer/worker
Jon Murphy
Aug 11 2020 at 2:14pm
Right. That’s my point about labor.
Yes. Also crudely called “labor productivity.”
That is incorrect (especially in the context of a tax). Review the concept of cost in economics, as well as deadweight loss. That is where your error is stemming from.
As Pierre said, it is possible, with strong enough assumptions, to make the two numbers (average labor productivity and total cost to consumers) equal. But they are not the same thing. My gas bill is $30. My whiskey is $30. But that does not imply my gas is the same as my whiskey.
Warren Platts
Aug 11 2020 at 3:53pm
Jon, that is a total red herring. My argument is saying that ‘Morning Star’ and ‘Evening Star’ are two different names that refer the the same object.
The way you calculate the labor productivity per worker of the Jack Daniels distillery is by adding up their sales (aka consumer expenditures) and then dividing by the the number of workers. To call that figure the “cost-to-consumer per worker” instead of labor productivity and then try to imply that the higher the number the worse it is, is a semantic shell game that should not fool anyone.
High labor productivity per worker = good
High cost to consumer per worker = bad
It doesn’t make sense. That’s like saying:
Morning Star = good
Evening Star = bad
Jon Murphy
Aug 11 2020 at 4:17pm
Right, but that argument is wrong, as a rudimentary understanding of the invoked concepts tell us.
In fact, your analogy helps show us the very mistake you are making.
“Morning” and “evening” star refer to the same thing only in very precise circumstances (namely if we are referring to the planet Venus or Mercury during a specific time period). If we use “Morning Star” in one of its other uses, they are no longer the same thing. If one is referring to the Morning Star to mean Venus and Evening Star to mean Mercury, then it is wrong to say they are two different names for the same objects.
Here is something for you to ponder: Literally every single economist says you are wrong here. Even the very models you cite, like Stolper-Samuelson and optimal tariff, show you are wrong here. Now, it is possible that everyone is wrong and you are right. It’s possible the whole profession is wrong or misunderstanding; it’s happened before (see Coase 1960).
But I am willing to make a bet with you that you are misunderstanding. I want you to write up your idea and submit it to the AER (or some other similarly-ranked, peer-reviewed journal). If it gets accepted or an R&R, I will pay your submission fee.
Warren Platts
Aug 11 2020 at 6:07pm
Jon, the Morning Star/Evening Star (they refer to Venus) argument goes back to a famous philosopher you probably never heard of: Gottlob Frege about 120 years ago. If you think he wrong, submit a paper to the Journal of Philosophy and I will pay your page charges if it gets accepted.
Frege’s point, however, was that even though ‘Evening Star’ and ‘Morning Star’ have the same referent (Venus), they have different senses. Evening Star connotes an object visible as the Sun is setting; Morning Star connotes an object visible when the Sun is rising.
It is the same with ‘labor productivity’ and ‘cost to consumers per worker’. Both have the same referent: a number arrived at by dividing output by the number of workers that produced the output. However, they have different senses. ‘Labor productivity’ is something we want more of: the higher the labor productivity, the better. But ‘cost to consumers per worker’ has the opposite connotation: consumers like cheap stuff–the cheaper the better–so if there is a high cost to consumers per worker that sounds bad because the the lower the wage, the cheaper the product.
As I said, this is a mere semantic shell game. These sorts of anti-tariff papers are not simply objective scientific articles: they have a definite and obvious political component to them as well, as is shown by these sorts of rhetorical games they play.
Pierre Lemieux
Aug 10 2020 at 8:15pm
Warren: If a tax can transform unproductive workers into productive workers, as you suggest, the solution is simple, as I mentioned: increase the tax by a factor of one million and you get economic nirvana. Your error may be that your implicit model is an economy with only one good (say, potatoes, not even leisure), only one factor of production (labor–no capital and no land, or perhaps infinite land with the same quantity and a zero price), no investment (of course), no government services, and complete autarky. Then, it may be correct to say that the quantity of cantaloupes produced multiplied by its price in terms of wage good (say, 2) divided by the number of units of labor times the wage rate in terms of wage goods (which must also be 2 cantaloupes) gives both the (average) productivity of labor and the quantity of consumption per unit of labor. But note that, in this model, the trick of boosting productivity by a tax does not work. (And, in the absence of money, the price in terms of wage goods in indeterminate.) I don’t think this model is very useful, though.
Warren Platts
Aug 11 2020 at 12:40pm
Pierre, I understand what you are saying: it is of course ridiculous to think that applying million percent taxes willy-nilly is going to magically increase labor productivity.
But a tariff is not an ordinary tax. It is a tax on imports with the dual goal of raising tax revenue while promoting certain desired sectors of the economy. Not all sectors are created equal. Manufacturing workers are more productive than fast food workers even though they are roughly substitutable. Therefore, we should not sacrifice our manufacturing sector to foreign workers making $2/hour or $2/day. Because of the Baumol effect, the more highly productive manufacturing workers we have, the higher the wages of fast food workers will be. I did the calculation: if the washing machine tariff improved overall working class wages by 1 penny per hour, they are better off on average despite an 11% increase in the retail cost of washing machines (assuming that price increase is real–I have my doubts).
Moreover, the tariffs–at least for a big continental sized economy–do not promote monopolies. The washing machine tariff did not turn Whirlpool into a monopoly. Rather, the tariff increased competition within the USA by the addition of two more firms, LG & Samsung. But at least the competition is now on a level playing field.
The apparent losers under such conditions is the professional & managerial elite class. They will lose out on a bit of consumer surplus. Moreover, higher wages for working class folks entails that they must pay more for nannies & gardeners & dog walkers, arguably making female lawyers less productive because they may have to spend more time at home to take care of their own children and dogs.
However, the higher the wages of working class folks, the more they can spend on consumption. In turn, desired (productive) investment (as opposed to the stock market/real estate casino) will increase to meet the increased aggregate demand. Because of increased productive investment, economic growth will increase, thus benefiting the elite managerial/professional class in the long run.
None of these ideas are original to me–it is the so-called American System that goes back to Alexander Hamilton and Henry Clay. It served our country well for nearly 200 years. But after Friedman’s infiltration, we decided to try the free trade experiment. Unfortunately, the observed results did not match the predictions. Economic growth slowed, wages are flat, death rates are up, birth rates are down. Therefore, if we are good scientists, we must reject the hypothesis that free trade is good for the USA (granted, free trade has worked wonders for Communist China–and, as a result, WW3 is now more likely than ever).
Jon Murphy
Aug 10 2020 at 3:16pm
Warren-
If what you say is true, then an optimal tariff (by driving down domestic prices and increasing imports) would reduce a country’s well-being, not increase it (if you do not believe me, I suggest you draw out both a large country and small country trade model and it will prove it to you).
Warren Platts
Aug 10 2020 at 4:53pm
Jon, I am afraid I don’t follow your reasoning. Ordinarily, the optimal tariff model predicts that domestic prices will rise, just not as much as under the small country assumption, because foreign producers will reduce their prices somewhat in order to maintain their market share.
Jon Murphy
Aug 10 2020 at 5:05pm
As I said, draw it out. In or explanation, you highlight your own contradiction.
Warren Platts
Aug 10 2020 at 7:33pm
Jon, you are shooting from the hip again. You wrote:
This is wrong on both counts: an optimal tariff drives domestic prices up, and decreases imports. Here is a drawing of an optimal tariff for you:
http://www.personal.psu.edu/faculty/d/x/dxl31/ec340/optimal.jpg
As you can see, the domestic price goes up, and imports go down.
Jon Murphy
Aug 10 2020 at 7:44pm
Again, I suggest you review the model. You can add it to your homework that Pierre gave you
Warren Platts
Aug 11 2020 at 12:55pm
No, seriously Jon: you are flat out wrong. Doubling down on wrongness does not make it right. I suggest you follow your own advice and review the drawing that was made by a real professor.
Don Boudreaux
Aug 12 2020 at 11:09am
Mr. Platts:
You write:
In follow-up comments, Pierre, along with my student Jon Murphy, responded well to your misunderstanding. I here emphasize just one fundamental point: You confuse ends with means.
In markets, high wages are rewards paid to means (namely, skilled labor) used to achieve especially urgent ends (namely, satisfaction of especially intense consumer demands). Consumers don’t purchase in order to serve workers; instead, workers work to serve consumers. And society grows more prosperous the greater is the satisfaction of these consumer demands.
You and other protectionists miss this reality by mistaking high wages themselves as being sources of prosperity (instead of what such wages are, namely, rewards for creating prosperity). You are led by your failure to understand that prosperity grows from the satisfaction of consumer desires to the utterly bizarre conclusion that society is made more prosperous the greater is the number of urgent needs – urgent consumer desires – government artificially creates by imposing tariffs and other protectionist obstructions.
Pierre Lemieux
Aug 12 2020 at 12:00pm
@Don Boudreaux: Thanks, Don: a fundamental point indeed. (If somebody decrees that the goal of the economy is to grow trees with blue leaves, his conclusion that government should help and control the economy would be unremarkable and useless.)
Jon Murphy
Aug 12 2020 at 12:18pm
To yours and Don’s point, these baffling conclusions (scarcity is preferable to abundance, costs should be maximized) are the result of treating accounting identities as causal relationships. As the Nobel Prize-winning GMU economist James Buchanan reminded us: order is defined in the process of its emergence. Identities like GDP, worker productivity, etc., lose their economic significance when the underlying market process is suppressed or otherwise manipulated.
Warren Platts
Aug 13 2020 at 4:03pm
Hi Don, I see you wrote an article on my comment on your website; I would have responded there, but I cannot seem to log on to your website for some reason……
Anyways, you are correct that protectionism considers high wages themselves to be a source of prosperity. You are mistaken, however, to consider that a mistake.
Wages are not a “reward”. Wages are a price, and like all prices, they are set by supply & demand. But unlike other commodities, workers buy stuff; workers are consumers. The higher their wages, the more production they can they can purchase. The higher the aggregate demand, the more the country’s producers will produce.
In your article, you of course create a bizarre straw man in order to make your counterargument: since oncologists get high wages, the government should force to get cancer, which is absurd.
The disanalogy is that an entire country sick with disease will drive down productivity. Tariffs designed to bring back manufacturing jobs, on the other hand, improve labor productivity because the new workers will either come from the vast army of the unemployed, or they will come from less productive service occupations.
Regardless, the argument here is not new: it the old, conservative, American System. Michael Pettis describes it very well in his article “High Wages Versus High Savings in a Globalized World“:
Jon Murphy
Aug 10 2020 at 3:14pm
So much for China paying for the tariff!
Warren Platts
Aug 10 2020 at 5:09pm
I guess not, considering their estimate of the tariff pass-through ranged
between 108 and 225 percent! But what does that mean, though? How can a tariff cause a price rise that is 200% of the tariff rate?
Jon Murphy
Aug 10 2020 at 5:53pm
Pgs. 2105-2106. It’s your standard econ story.
Warren Platts
Aug 11 2020 at 1:01pm
I don’t know what those page numbers refer to, nor do I know what you mean by ‘standard econ story’. However, a tariff that causes a price increase that is 200% of the total tariff certainly is not included in standard tariff theory.
Of course there are many possible explanations for the anomaly:
the standard tariff theory is broken
there is something majorly wrong with Flaaen et al.’s methodology
middlemen up & down the supply chain use the tariff as leverage when negotiating price points (“Why did you jack the price?” “Trump’s tariffs, ya know.” “Oh OK..”)
There are other things going on that swamp the effect of the tariff
Jon Murphy
Aug 11 2020 at 1:25pm
The cited study.
All those are possible, yes, but none particularly probable.
For points 1 &2, overturning nearly two centuries of economic theory (including Stolper-Samuelson and optimal tariff theory) calls for something more substantial than “I don’t understand elasticity and demand curves.”
Point 3 is unlikely given that middlemen reduce prices, not increase them
Point 4 is just a meaningless catch-all.
Warren Platts
Aug 11 2020 at 1:41pm
Ah yes. But still, they do not attempt to offer an explanation about how a tariff can cause a price increase that is 200% of the tariff. Of all the literature they cite, none of them found that tariffs increase the price more than the tariff rate.
So you don’t find my possible explanations very compelling. Fair enough. Why do you think the tariff caused a price increase that was substantially higher than the tariff itself?
Jon Murphy
Aug 11 2020 at 2:18pm
Yes they do. Again, see the cited pages. Especially the last two paragraphs. If you are still confused, I suggest you review how elasticity, demand curves, and substitute/complementary goods work.
You are, of course, free to consult an actual economics professor.
Warren Platts
Aug 11 2020 at 3:40pm
No, they don’t. Yes, they found a tariff-elasticity > 1, but that is a strange, unique result that does not appear elsewhere in the literature. They say their finding “contributes to the literature” (and that it certainly does), but they do not attempt to identify the channel that caused their surprising result.
Jon Murphy
Aug 11 2020 at 4:08pm
Uh…it occurs all the time (see, for example, just about any of the papers the cite). Again, review your theory. Any intro book goes through this.
Warren Platts
Aug 11 2020 at 6:28pm
Uhh…. No it doesn’t. Flaaen et al. cite many papers: some find partial pass-through (tariff-elasticity < 1); some find 100% pass-through (tariff-elasticity = 1).
Some even found negative elasticity (tariff-elasticity < 0): a seemingly strange result, but can be caused by “tariff jumping” when production moves to a 3rd country where wages are even lower.
Alternatively, one study found that a tariff can encourage foreign multinationals to relocate production to the home country, thus increasing competition behind the tariff wall ultimately resulting in lower prices than would otherwise be without the tariff. This is of course an old principle of the American System of economics. I am glad to see some contemporary support for their theory.
However, none of the papers cited by Flaaen et al. claim to have found price increases significantly higher than the tariff itself (tariff-elasticity > 1). Nor do they infer a possible explanation. Moreover, I have read the tariff sections of many introductory textbooks. NONE of them ever suggest that tariffs can result in price increases in excess of the tariff itself. (You can easily settle this point and prove me wrong by providing an actual quotation from a textbook that says tariff-elasticities > 1 are to be expected.)
That’s why I was asking if you could think of a possible explanation. Anyways, I think my #3 is most plausible: middlemen up & down the supply chain all want their pound of flesh. Flaaen et al. focus on retail prices; retailers will typically have a standard mark-up on whatever the wholesaler charges; who in turn will have a standard markup on whatever the manufacturer charges. If both mark-ups are 25%, then if a manufacturer marks-up its price by the full tariff amount, say $1.00, then the wholesaler will increase that dollar to $1.25, and then the retailer will in turn make the $1.20 price increase int0 a $1.56 increase.
Thus by the time the widget is bought by the final consumer, the tariff-elasticity has mushroomed from 1 to 1.56; so I can see how Flaaen et al. arrived at their result.
Jon Murphy
Aug 11 2020 at 7:14pm
Assuming this is true, it is your problem: you just jumped to the section without reading the theory it is based upon. Your errors are far more fundamental: you do not understand what costs are, what complementary and substitute goods are (and their effects/reactions from prices), you do not know what scarcity is, etc.
In short, you are trying to build a house without understanding the role the foundation, plumbing, framework, and walls play and then acting like your roof collapsing is a feature, not a bug, because roof collapse is the same thing as having a second floor.
Warren Platts
Aug 11 2020 at 7:23pm
OK, so you cannot find a quotation in a textbook that says that tariffs can result in tariff-elasticities significantly greater than 1. So we agree that textbooks do not say that. Cool!
Jon Murphy
Aug 12 2020 at 11:50am
Ignoring your tortured bit of logic (you never demanded a quotation and, in fact, I pointed explicitly to where you can find your answer multiple times), you actually inadvertently prove your point that cost = benefit wrong.
No textbook will support you. Therefore, you must be wrong.
Pierre Lemieux
Aug 11 2020 at 9:19pm
As noted by Warren, it is not immediately obvious why the price of a complement good (such as a dryer) would increase with the price of its complement (such as a washing machine). I think Flaaen et al. could have treated this problem better. In a competitive market, there is no reason to expect that: if the price of shoes increases, the price of laces should decrease because the quantity demanded of shoes will decrease and so will the demand for laces. (By the way, there is no reason either to expect that a markup, which is just a managerial rule of thumb, would remain unchanged: exchange and competition drive markups, not the other way around.) The way Flaaen et al. respond is to observe (p. 2125) that the washing machine industry (and thus the complementary dryer industry) is “moderately concentrated” (according to the HHI). This is not surprising given the limitation of international competition by tariffs in this industry. I haven’t read Amiti, Itskhoki, and Konings, but I can see how it makes sense that, in an industry characterized by some market power, the profit-maximizing price of a good would increase (producers would push it up) if the price of its complement is pushed up by a tax or a tariff.
Jon Murphy
Aug 12 2020 at 11:54am
I suspect it’s more a bundling thing than necessarily market concentration (though the two go hand-in-hand). A washing machine is relatively less useful without a dryer. Washers and dryers tend to work in tandem with one another. If the price of one goes up, since the other is bundled with it, the price of the other goes up as well.
Furthermore, dryers were likely affected by the other tariffs on steel, aluminum, etc as well.
Pierre Lemieux
Aug 12 2020 at 5:46pm
@Jon: I don’t think the bundling hypothesis provides an answer. Consider a laptop computer, which bundles thousands of components. Suppose that a tariff (or some other tax) is levied on CPUs. Certainly, the price of laptops will only increase by the amount of the tax on CPUs (if the market is competitive).
Jon Murphy
Aug 12 2020 at 5:59pm
Hm, touche.
Capt. J Parker
Aug 10 2020 at 4:01pm
The US Federal Government demands that a big piece of the US social safety net must be baked in to the cost of US manufacturing exports. They do this by taxing labor as opposed to taxing consumption or production via a VAT. South Korea has a VAT and I presume they exempt exports but apply the VAT to imports.
A US made washing machine exported to S. Korea is at an economic disadvantage. The US machine consumer price in S. Korea carries the cost of both S. Korean VAT and US taxes on labor while the S. Korean machine carries only the S. Korean VAT. In the US market, the S. Korean machine, absent tariffs, caries no tax while a US machine carries the baked in US taxes on labor. This means that US manufactures are at a disadvantage in both markets simply because of the way the US tax code is structured relative the S. Korean code.
So instead of repeating the mantra that US tariffs are taxes on US citizens, which most people that come to this blog already understand, why not consider tariff effects in light of the existing trading realities and political realities.
If income from tariffs on washing machines subsidized exports or subsidized domestic purchases of domestically produced machines would this correct some the inefficiencies that must be caused by differential tax treatment of the machines by the two trading partners?
Even in tariff receipts only went to the US treasury general fund, would not the lessened pressure on the US governments need to tax labor be a benefit to all US exporters and would not that benefit need to be assessed to know the true cost of the tariffs?
Jon Murphy
Aug 10 2020 at 4:21pm
Quick point of fact: if US washing machines are exported to Korea, then they are not at an economic disadvantage since people are buying them. Given a choice, it means Koreans are buying those washing machines over other options, so at least for those Koreans and those American manufacturers, those machines have the economic advantage.
So, be careful here. Even with the differential tax treatment, it does not necessarily imply there is an inefficiency here. It’ll depend on what your other assumptions are. “Inefficiency” as compared to what?
Furthermore, even if there is an inefficiency, the costs of moving from the status quo to some desired alternative may be too high as to justify. For example, imposing a tariff and subsidizing exports runs into all sorts of public choice and transaction cost issues. How likely is it the gains from such a scheme outweigh the 800k+ deadweight loss from the tariff coupled with the political costs of the subsidy?
Capt. J Parker
Aug 11 2020 at 5:54pm
@Jon Murphy
To be more precise: US producers of washing machines are at an economic disadvantage when selling machines in S. Korea relative to where they would have been absent a US tax code that increases the cost of the labor component of the US machines.
To your second critique on inefficiency: Assume there were no S. Korean VAT taxes and no US income or payroll taxes. Then assume, under those conditions Koreans buy some number of Korean and some number of American Machines. Likewise for Americans. The relative numbers of Korean or American machines consumed by each consumer group is determined by the consumers of each country maximizing their utility. Now along comes the taxes. If the taxes result is Koreans and Americans consuming more Korean machines relative to American machines than they otherwise would have then the taxes have distorted the market and that distortion is very likely to have created an innefficiency.
Jon Murphy
Aug 12 2020 at 11:47am
Right, but that’s part of my point. You’re actually just assuming in that the taxes are causing some sort of inefficiency. Now, they could be. They probably are; it’s unlikely that the governments imposed taxes in such a way to minimize inefficiencies.
Capt. J Parker
Aug 12 2020 at 3:08pm
@ Jon Murphy
So, we are in agreement on that point. What that should mean is that those who are calculating the cost of Trumps tariffs ought to consider if the tariffs are increasing or decreasing those inefficiencies.
Now, I readily admit that it is unlikely that the Trump tariffs resulted in a big efficiency improvement. But, could they have if designed differently? You need to ask the question and not just view the tariffs through the lens of “tariffs are no more than a tax on American consumers” because there are plently of other distorting taxes to contend with beyond just the tariffs.
Jon Murphy
Aug 12 2020 at 3:21pm
That’s what this study does. It finds it increases the inefficiencies to the tune of 815,000 per consumer per job saved (or, in absolute terms, $1.5 billion annually)
Capt. J Parker
Aug 12 2020 at 11:38pm
It does no such thing. It considers only price changes from the pre Trump tariff market. The pre Trump tariff market was already distorted by South Korean and US taxes on consumption and labor respectively.
Pierre Lemieux
Aug 13 2020 at 11:01am
@Capt. J. Parker: If the arguments in my post “Taking Comparative Seriously” are correct, your argument is not valid, because you would have to add thousands of other public policies in the distortion bag. For example, the fact that American governments heavily subsidize education, which favors the English language, would be seen as creating a distortion against South Korean producers (and even more against North Korean ones!). We could then arrive at the conclusion that if free trade with no “artificial” comparative advantage is impossible, autarky would be better.
Capt. J Parker
Aug 13 2020 at 5:28pm
@Dr. Lemieux,
I agree with you that my argument taken to a reductio ad absurdum could well mean that you would have to add “thousands of other public policies in the distortion bag.” That doesn’t make my argument wrong. It may make my approach analytically intractable unless of course you can discard most of the public policies as having either a diminimus effect or having no practical remedy for the distortions they create. It is my belief that if you were to do so you would find that US income taxes vs trading partner VAT taxes produce significant distortions and that these distortions favor foreign imports and disfavor US exports. These distortions do have practical remedies. Some of these remedies may involve US tariffs and the effects of such remedies in reducing the tax policy distortions deserve consideration in any negative critique of changes in US tariff policy.
I never argued that we should have no trade if no artificial comparative (or absolute) advantage/disadvantage is possible. And it probably is not possible. I do argue, and I’m sure you would agree, that public policy aught to minimize market distortions as much as possible. In order to minimize the distortions you first need to recognize that they are there.
Jon Murphy
Aug 10 2020 at 5:14pm
There’s another question here: rather than dealing with subsidies and the enevitable multiplication of inefficiencies that come along with them, why not just remove the taxes that caused the original inefficiences?
Pierre Lemieux
Aug 10 2020 at 8:50pm
@Capt. J. Parker: Thanks for your contribution to the conversation. A couple of rejoinders, which I will have to make short:
1) It is not sure at all that payroll taxes are “baked into” the price of exports. Only marginal costs affect the supply curve. Plus: an employer’s payroll tax will be at least partly paid by the employees.
2) A VAT is paid only by the final domestic consumer, and what he pays reimburses what producers have advanced. (Try to build a simple example with a farmer, a baker, and a consumer. Assume the farmer buys no inputs.)
3) In passing, there is no reason to favor exports over imports. See, for example, the “Here is the catch…” paragraph in my “A Collectivist Case for Trade, but not FREE trade.” I’ll have a short article on this very point in the Fall issue of Regulation.
4) In a previous post (“Taking Comparative Advantage Seriously“), I explain why domestic taxes that affect comparative advantage simply become a negative part of it. If some producer thinks the tax is discriminatory and disadvantages them, let them persuade their government to change it. (He may have the support of economists and libertarians.)
Capt. J Parker
Aug 11 2020 at 5:32pm
Dr. Lemieux, It was my pleasure to receive your reply. I’d like to raise a new issue in response to your Collectivist Case for Trade but please allow me to respond to each of your points.
I agree. Just how much US tax gets baked in to the consumer price of washing machines sold in South Korea is a matter of the price elasticity of both labor supply and demand in the US. I’m assuming the elasticities are such that the cost of the labor component of US washing machines is greater than it would be, absent the taxes. I don’t think this is difficult to support.
I agree. My point is VAT is applied to both the US and S. Korean washing machines sold in S. Korea. Baked in US taxes are applied as well but only to US made machines by virtue of the US Tax code.
The Collectivist Case for Trade was great. I need to read it a few more time to fully digest it. Here is my own reason for desiring more exports: To buy imports we do one of three things a) sell exports to foreigners, b)sell domestic assets to foreigners or c) give foreigners promissory notes telling them we will give them even more exports or domestic assets in the future. I believe b and c are unsustainable. https://www.cfr.org/report/quarterly-update-foreign-ownership-us-assets. History can describe to us what it means for a country to have large portions of its land and means of production owned by foreign interests. I don’t wish for my children to inherit such a country.
Once again I agree. And I think Tax Policy is in fact hurting our comparative advantage. If Tax Policy is causing us to export less and import more than we should were tax policy neutral then I would think that is something academic economists would want to tell policy makers about.
Jon Murphy
Aug 11 2020 at 10:06am
Here is another thing worth considering: the authors of this study almost certainly underestimate the net costs to Americans. Once retaliatory tariffs, the political costs, and enforcement costs are factored in, I bet the cost-per-job is considerably higher.
Comments are closed.