The argument that tariffs raise imported goods prices is exactly the same as the argument that gasoline taxes raise gasoline prices.
You might object that it’s theoretically possible that a given tariff doesn’t raise prices. That’s true. It is also theoretically possible that a gasoline tax increase doesn’t raise gas prices. In both cases, the seller might absorb 100% of the tax. The chances of that occurring in the real world are vanishingly small, especially for tariffs that apply to all countries.
Matt Yglesias recently retweeted a Scott Lincicome tweet and added a comment:
Here I think Yglesias is greatly overstating the extent of disagreement. That might seem an odd claim, as he uses the phrase “very little disagreement” and points to a study that shows only 5% disagreement. Nonetheless, I still believe he’s massively overstating the degree of disagreement, which I suspect is actually far below 1%.
The poll question asked about the effect of tariffs on “general economic welfare.” A few economist (not many) favor tariffs because they think they might boost welfare. But that’s almost certainly not because they think tariffs will avoid raising prices. For instance, suppose an economist thought that the loss of blue color jobs to imports was a bigger problem than higher prices. It’s not a nonsensical claim, although I happen to think it’s wrong, partly for reasons expressed in my previous post. (I don’t believe it would save jobs.)
The small number of economists that do favor protectionism do so precisely because they believe a tariff would raise prices. If it didn’t raise prices, if it didn’t protect domestic industry from cheaper imports, then it would fail to protect jobs in import competing industry.
You might think I’m making a mountain out of a molehill, making too much of the difference between a 5% minority and something like a 0.5% minority. But I worry that people might assume that a proposition is almost certainly true if 95% of economists believe it to be true. If 50 economists out of 1000 hold a heterodox opinion on a given subject, it’s certainly not all that implausible that they might be correct—certainly much more than 5% odds. Consider a case where 95% of economists thought it was 75% likely that X was true, and 5% of economists thought it was only 25% likely that X was true. If polled, you might see 95% of economists saying they believe X is true, but in fact it would be only 75% likely that X is true, even if those 95% were completely correct.
I am part of a tiny percentage of economists that believe that the Fed caused the 2008 recession with a tight money policy. But even if a poll shows that 99% of economists believe that I am wrong, that would not suggest that there is a 99% chance that I am wrong. Indeed I doubt many of those economists who disagree with me would accept a wager where they could win a measly $102 on a $100 bet, on the question of whether an alternative monetary policy in 2008 could have prevented the big drop in NGDP, especially given that we weren’t even at the zero lower bound! (Yes, this would be hard to test, but imagine if there were a test.)
Polling economists certainly tells us something useful about what experts believe. But it’s important not to overstate the significance of a strong majority of economists lining up on one side of an issue. It’s not meaningless, but also not definitive.
PS. It’s also possible that a poll question on whether tariffs raise prices would also yield the same heterodox 5%, in which case it may be that there are a small number of economists who are simply highly eccentric. But I still believe the figure would be well below 5%, particularly if the two questions were asked back-to-back, reminding the economists being polled that they are two distinct questions.
PPS. Less than an hour after completing this post, I was reading The Economist and came across the following story about the Russian economy:
Russian GDP will rise by over 3% in real terms this year, continuing its fastest growth spurt since the early 2010s. In May and June economic activity “significantly increased”, according to the central bank. Other “real time” measures of activity, including one published by Goldman Sachs, a bank, suggest the economy is accelerating (see chart 1). Unemployment is close to an all-time low. Inflation is too high—in July prices rose by 9.1% year on year, above the central bank’s target of 4%—but with cash incomes growing by 14% year on year, the purchasing power of Russians is rising fast. In contrast with people in almost every other country, Russians are feeling good about the economy.
I’d estimate that back in 2022, far more than 95% of economists (including me) were wrong about how the Ukraine War and the resulting sanctions would impact the Russian economy. More often than not, 95% of economists will be right. But in a disturbing number of cases they are not.
READER COMMENTS
Robert Benkeser
Aug 25 2024 at 10:20pm
Sanctions to date on Russia remain highly circumscribed and targeted at particular entities. This is beginning to change now that secondary sanctions are being applied.
If we applied secondary sanctions to Russian blood oil, the entire Russian economy would implode.
nobody.really
Aug 25 2024 at 10:58pm
“What do Americans think about corporal punishment? We poled our studio audience and–shesh, it looked painful.” David Letterman (quote approximate)
Thomas L Hutcheson
Aug 26 2024 at 2:02pm
I do not know about the percentage of what economists think, but the optimal tariff (more applicable probably to the optimal export duty) is a standard result of orthodox trade theory. It does rely, however, on raising the price to the consumer of the imported item with the tariff revenue costlessly rebated to the consumer to get the net benefit
Jon Murphy
Aug 26 2024 at 5:48pm
According to Paul Krugman, Maurice Obstfeld, and Marc Melitz in their International Economics textbook, while optimal tariff does follow out of the Hecksher-Ohlin model, it’s useless for understanding policy or real world outcomes. I forget their exact phrasing, but they basically its a mathematical quirk rather than a practical result.
So, one can accept the optimal tariff model while rejecting the idea that tariffs improve welfare. At least, thee prominent trade theorists think so.
Jon Murphy
Aug 27 2024 at 10:14am
Here’s the quote from Krugman, Obstfeld, and Melitz:
Warren Platts
Aug 27 2024 at 8:24am
Not how that works: the tax revenue can be used for any purpose as far as the general welfare calculation is concerned.
steve
Aug 26 2024 at 3:51pm
I am not so sure that what the Russians are doing is sustainable, but it would be good if some economists took a shot at explaining how it seems to be working. By my understanding, the central bank has set rates pretty high, 18%-20%. The government is hiring like crazy, outcompeting private sector jobs. The government is financing mortgages at 8%. This is an incredible amount of intrusion into the economy by the government and it doesnt really sound like it should be that conducive to growth.
To your point about Yglesias underestimating, he used a vague term that we dont know the percentages to which it applies. It reminds me of an older paper in the medical literature looking at what people mean when they say never and always. In their study always meant, IIRC, 80%-100% and never meant 0%-10%.
Steve
Warren Platts
Aug 27 2024 at 8:36am
That’s a good point. I would add that poll question doesn’t specify whether general economic welfare means net general economic welfare: I guess one could claim that even where a tariff increases net general economic welfare, the negative consumer surplus and deadweight losses still represent a reduction in general welfare even if netted out.
A
Aug 28 2024 at 11:54am
Could it just be that Russia is now sharing more of its government funds with regular people? Before Russia made a ton of money from oil and gave it all to oligarchs, but now it is using the money to pay soldiers and factories, so essentially it’s like a big tax rebate.
Warren Platts
Aug 27 2024 at 8:17am
There are at least three things going on in the tweets above: raising of consumer prices, effect on [net?] general economic welfare, and whether the alleged negative effect happens some of the time or necessarily all of the time. (I suspect the “consensus” mostly assumes some of the time.) If the tariff raises prices (hence reducing consumer surplus) but raises producer surplus and tax revenue, the reduction to general economic welfare is just the (usually relatively small) deadweight loss associated with the import tax — assuming no terms of trade gain. But according to the “big country” theory (and certainly USA counts as a big country), the foreign producers will lower their prices such that some or all of tax burden falls on the foreign producer instead of the domestic consumer. In that case, the terms of trade gain will tend to outweigh the deadweight loss (assuming the tariff isn’t prohibitory), thus increasing net general economic welfare. (Also why economists refer to tariffs as “beggar-thy-neighbor” policies.) I once did a BOTE study on the Canadian lumber tariffs around 2016. Enough information was given to estimate the various elasticities and price changes, and a modest increase in U.S. general economic welfare was estimated.
Jon Murphy
Aug 27 2024 at 8:38am
Be careful here. The way you’ve written this, it sounds like that any reduction in price (or absorption of some portion of the tax) will lead to a terms of trade gain that offsets the decline in consumer welfare and the del resulting from inefficient resource use. That is not the case. The dead weight loss can still be greater than the tax revenue generated in many cases. The tariff doesn’t need to be “prohibitive.” It just needs to be sufficiently high.
Further, as an empirical matter, if other countries retaliate or the tariffs cause firms to adjust along unanticipated margins, the dwl will exceed any potential terms of trade gain. That’s why the empirical investigations of the Trump/Biden tariffs have found they’ve been a net loss to Americans (I’m unaware of any study that finds a welfare gain).
Scott Sumner
Aug 27 2024 at 12:03pm
Optimal tariff models tend to exclude retaliation, which makes them very unrealistic. Real world tariffs are almost always a net negative.
Jon Murphy
Aug 27 2024 at 1:15pm
Optimal tariff models are similar to optimal price control models (of the Lange variation, or more recently, Isabella Weber):
Are there some theoritical set of circumstances where a price control or tariff could be applied that increases welfare? Sure. But the number of things you need to go right in order for that outcome to happen is extremely high, even in the best case scenerio. In a realistic world, where political considerations happen and policymakers are not national welfare-maximizing computers, those models fall apart.
In the end, they’re more exercises in mathematical cleverness rather than economic models. Those sorts of models are trivially easy to invent. But rarely do they provide any sort of insight into how the real world works.
Warren Platts
Aug 27 2024 at 1:59pm
These sorts of comments are rather beside the point when addressing the concerns of people like Trump and probably a majority of voters that are talking about tariffs, industrial policy, and ending the trade deficit.
Yes, I agree that in a perfect world of free trade, global GDP would be maximised. Ricardo proved that 200 years ago.
But we are not in a perfect world. Thus we want tariffs for reshoring industrial production and jobs. But not numbers of jobs so much as increasing wages for the working class (even if at the expense of reduced real wages for the laptop class).
As for industrial policy: our policy of no policy still results in American industrial policy, only that policy is dictated by the likes of China and other mercantilist states. A world free of industrial policy is a pipe dream; better that we should design our own industrial policies.
And the trade deficit is out of control. It’s a trillion dollar giant sucking sound that’s a huge headwind to GDP growth and it’s because of mercantilist states who force up their national savings (typically by suppressing their own wages) and then export those savings, mainly to the U.S. and U.K. because we have about zero capital controls. Moreover, the NIIP is at 70% of GDP last I checked. Historically, that’s danger territory that no one is addressing. The trade deficit WILL unwind one day. The only question is whether we try to manage the unwinding or let nature take its course.
Thus we would really appreciate it if you economists could steel man these arguments at least before you dispense with them!
Jon Murphy
Aug 27 2024 at 3:06pm
This is one of the annoying things about discussing policy these days. It just becomes a game of whack-a-mole as the defender of the policy searches for justifications:
“Tariffs are good because of the optimal tariff model!”
“Well, no, because…”
“Well, all that’s all beside the point because tariffs are really about industrial policy”
“Fine, but they won’t achieve those goals because…”
“Well, all that’s beside the point because tariffs are really about raising revenue.”
“Ok, but they’re not the most efficient way to raise revenue because…”
“Well, all that’s beside the point because tariffs are really really about setting optimal tariff levels”
And so on. We saw that with Trump’s/Biden’s tariffs. We’re seeing this now with Harris’s price gouging comments. These are policies in search of justifications rather than any serious attempts to solve political or economic concerns.
Warren Platts
Aug 28 2024 at 11:00am
OK, I found a recent paper by Autor et al. that compared the effects of Trump’s tariffs with retaliatory tariffs, and sure enough, they found that Trump’s tariffs did not have (much of) an effect on employment/population ratios in commuter zones that presumably would benefit from Trump’s tariffs, yet agricultural commuter zones targeted by retaliatory tariffs did suffer measurable employment losses that were only partially compensated by subsidies.
The authors say, “The stated goal of the Trump administration’s
trade policy was ‘to bring back jobs to America.’” But that begs the question of what “bringing back jobs” really means. Does that mean the same as ‘increasing the number of employed people in the US’ or ‘increasing the mix of manufacturing jobs in the US’? Clearly, it’s the latter. After all, we are repeatedly told by expert economists that trade policies don’t have (much of) an effect on employment, especially at full employment. Hence Autor et al. spend 55 pages of mind-numbing mathematics attacking a straw man. They didn’t even ask the question that seems obvious to me, whether the thousands of manufacturing jobs created were filled mainly by people in low-wage service jobs rather than from the ranks of the unemployed. If the former, we should expect employment to remain level, but wages should increase. That’s what Autor et al. should have looked at. We do know that nationally, during the Trump tariff era, real wages for the working class increased faster than laptop class wages for the first time in living memory. But it sure would be nice to a fine-grained commuter zone look at the effect on wages.
Jon Murphy
Aug 29 2024 at 11:25am
I like the overall point in your post. But one thing I think we ought to keep in mind for Matt Yglesias’s tweet is that he is responding to recent claims by Donald Trump that tariffs lower prices and JD Vance that there is much disagreement among economists about the effect of tariffs. Yglesias, retweeting Liniciome, gives the lie to those claims.
The 95% may be empirically wrong, but there isn’t much disagreement about what economists think.
Warren Platts
Aug 29 2024 at 1:24pm
Vance is correct: tariffs have lots of effects, and economists disagree on them; the effect on consumer prices is only one such effect. What Vance further claimed is exactly what I said in my post immediately above: that tariffs will raise the real wages of the working class whom will be made better off as a result. (Granted, this has the effect of reducing the real wages of the laptop class). As for prices, they can be reduced to below the full tariff for at least two reasons: (1) the optimal tariff theory we’ve already discussed; and (2) this goes back to Alexander Hamilton: that tariff walls will increase competition among domestic firms that will in turn tend to drive prices down. The latter is exactly what we witnessed with respect to Trump’s washing machine tariffs: prices spiked at first but competition among U.S. firms about doubled, production increased, and prices soon reverted to below their inflation-adjusted pre-tariff price. That Flaaen et al. study that everybody cites that reported an 11% price spike by mid-2018 was obsolete by the time it was published in 2019…
Comments are closed.