« Maybe NAFTA is safer than we think | Main | What can we learn from Huey Long? »

February 14, 2017


Feed You can follow this conversation by subscribing to the comment feed for this post.

Do you think that the GOP will not carve out an exemption for what will be portrayed as an gas tax? Otherwise looks correct.

Also, this effect is reversed if Nafta is exempted, as Canadian & Mexican producers high carbon, as you discuss in the post.

Would be good to divert border-adjustment tax from oil to infrastructure. Good way to raise gas tax indirectly (if it can pass)

I'm opposed to border adjustment as manuf supply chain disruptions will be severe, for autos for ex. so costs will be very large [seems you have written on this. http://noelmaurer.typepad.com/aab/2017/01/border-adjustment-the-mexican-peso-and-hoyt-bleakley.html]

One more point: if Canadian oil subject to border tax, many pipeline projects probably no longer viable. Could kill Dakota Access Pipeline for example

Here's where I'm uncertain. Supply chains would be disrupted, but in the short-term moves in the exchange rate should cancel that out.

Even if not, isn't 20% well within historical swings? In late 2008, the peso went from 10 to 14 and then slowly sank back to 12 by the end of 2011. Didn't cause disruptions.

The Canadian dollar is even wilder. C$1.6 per dollar at the beginning of 2003. Steady rise to parity by 2009. Then craziness with GFC in late 2008--a crash to 1.3, then recovery. Finally, since 2013 a steady fall from 1:1 to 1.3 per USD.

Those movements are gonna swamp a 20% border adjustment tax, no? But supply chains held together fine, with remarkably little pass-through on the American side. (There's some evidence that this is true for Canada as well: http://www.bankofcanada.ca/2005/10/working-paper-2005-29/.)

In the long term, wouldn't you expect a more efficient allocation of resources, since the current U.S. tax system is biased towards imports?

It's true that the fluctuations from exchange rates are larger than the border adjustment tax, but many of these fluctuations are temporary unlike the border tax which would be permanent. Some factories become immediately unprofitable. So reallocations would be more significant, and I would guess that the negative side comes from layoffs while the expanding establishments use more capacity, so immediate impact is more unemployment.

The exchange rate would move partially, but capital account will also move both from border adjustment tax as well as in response to depreciation itself. So uncertainty over magnitude is large.

Traded GDP always more productive, no? Imports more efficient than domestic competition and Exports more competitive than other domestic production (comparative advantage). I've never seen this formally derived, but I would bet that with lump-sum taxes a mild trade subsidy is optimal, even relative to free trade without transfers. So BAT inefficient as basically a tariff. I don't think the US tax system is biased towards imports, but let me explain in the next comment.

The BAT is being sold as a synthetic VAT, as imported inputs would be taxed as is the case in a VAT and is not the case for a sales tax. However, the incidence of a sales tax falls on the whole value added of a sale, while a value-added tax will tax each stage of production. So if there is a choice between guacamole made from California of Mexican avocados, it is true that under the current system Mexican avocados don't pay any tax at the border. However, the value of the guacamole (assuming US processing for both) is still taxed the same, so it's not the case (from what I can tell) that imports are favored, as alleged by BAT supporters. It is true that the VAT is not paid by the Mexican avocados that are exported as would be paid on domestic sales of Mexican avocados, but if this were the case the Mexican avocados would be double taxed, paying tax both in Mexico (directly) and indirectly through US sales tax. So this VAT rebating is portrayed as an export subsidy, when in fact it's avoiding the VAT functioning as an export tax.

What about US corn exported to Mexico to be processed into tortillas? Well the Mexican tortilla company will pay net VAT on the value of its processing, retailing, distribution, etc. But the way VAT works is the VAT is paid on the sale price and then the value of costs/intermediate inputs are rebated so that only the value added is counted. To go on a brief tangent, this makes for an enforcement advantage of VATs because if a companies supplier underreports their value-added to avoid tax, then the firm making the sale will report them to avoid paying their portion of the VAT. So anyway, since the VAT is paid on just the Mexican value-added regardless of whether Meixcan or American corn is used, there is no implicit tariff from the VAT either.

So I would prefer a VAT to a sales tax, but they are roughly equivalent, while the BAT is basically an implicit tariff. I am not a public finance expert, but this is how I see it. If I'm making a mistake here, please let me know, as I'd like to be well-informed on this issue.

Countries with a VAT (like Mexico) don't see the same impacts on domestic vs. foreign oil production, because the border tax is basically a tariff. With a VAT system, both the imported and domestic oil pays the same VAT, with the refining/distribution also paying VAT on their value-added. But your analysis is correct for an oil tariff, which would have environmental benefits.

This was incredibly helpful.

BATs have been discussed a lot recently in the context of carbon taxes, but not in general trade. Your comments got me thinking, however. Turns out that they got a lot of attention around 1970, when the EEC (now E.U.) adopted its current VAT system.

So I looked at that literature in light of your discussion!

You can think of the BAT as a combination import tax/export subsidy. That's why it has a different effect from a straight import tariff. The straight tariff discourages imports and discourages exports. The latter is because export revenues are ultimately used to buy imports or investment overseas; with more expensive imports, however, you get less bang for your export buck. And so they reduce the volume of trade with indeterminate effects on the balance of trade.

A BAT, on the other hand, cancels out the export-discouraging effect with an export subsidy. So ultimately you should have no effect on the volume or balance of trade: the exchange rate will pick up the adjustment. In theory. (If the ER is fixed, inflation will pick up the adjustment. Again, in theory.) Meade 1974 has a good summation in the JPE. (Man, people don't write like that anymore!)

Thinking it through, it seems like the BAT is the equivalent of abolishing the corporate income tax, introducing a VAT, and then cutting payroll taxes.

1. Abolish the corporate income tax.
2. Introduce a 20% VAT. (A VAT hits value-added. Which is more-or-less equal to cash flow over the long term -- yeah, lots of devils in those details. But it more-or-less the same.)
3. Abolish payroll taxes. (The BAT, unlike a VAT, will allow the deduction of labor expenses.)

Since payroll taxes are lower than 20% in the U.S. (lower still when you think of the incidence on employers) then there is an implicit wage subsidy here. But that's the major distortion, no?

Off to work through the tortilla example again, unless I've missed something big above.

I should add that as you say, a BAT does not have the enforcement advantages of a VAT!

I think I got it. You're right, but it really all depends on the resulting exchange rate movements. I think.

For the moment let's pretend that the BAT rate is the same as the U.S. income tax rate. It reduces the number of moving parts.


(1) A Mexican avocado producer exports avocados to the USA. They do pay Mexican income taxes on the profits from the sale. They do not pay Mexican VAT.

(2) The American retailer pays U.S. sales taxes on the full value of the avocado when they sell the avocado. They also pay U.S. income tax on the profits from the sales, less the cost of the imports from Mexico.

(3) An American avocado producer pays U.S. income taxes on the profits from their sale to an American retailer. They do not pay sales taxes.

(I am ignoring incidence for the moment.)

Now the U.S. introduces a BAT.

(1) A Mexican avocado producer exports avocados to the USA. They do pay Mexican income taxes on the profits from the sale. They do not pay Mexican VAT. Nor do they pay the BAT. On the surface, this is the same as before.

(2) The American retailer pays a 20% BAT on the avocados. They pay U.S. sales taxes on the full value of the avocado when they sell the avocado and an additional 20% on the profits from the sales. From their point of view, Mexican avocados have gone up in price by 20%.

(3) An American avocado producer pays the 20% BAT on the profits from their sale to an American retailer. They do not pay sales taxes. On the surface, this is the same as before.

This looks like it is establishing fairness. After all, the Mexican avocado farmer was exempt from U.S. income taxes whereas the American avocado farmer was not. Bad!

But not really, because the Mexican company was not exempt from Mexican income taxes. If it had been, then the BAT proponents would have a case that their proposal is leveling the playing field. But they are, so it isn't, at least from countries that levy higher income taxes than the United States.

But there's a macro wrinkle: this is still not the same as a tariff. The BAT is also, in effect, an export subsidy to American exporters. (To see the logic, switch the above around, replacing "Mexico" and "U.S." You can also switch "corn" for "avocado," for versimilitude.)

That means that the BAT, unlike a straight-up tariff, won't alter the volume of trade. Exchange rates will adjust, leaving both companies in the same position. (Mexican income, however, will be redistributed from importers to exporters via a fall in the peso. US income will not be redistributed, with ER movements counteracting the tax subsidies.)

But that is assuming that adjustment happens. If it does not, for whatever reason, then the BAT will act as a tariff.

Did I make a mistake, Gabriel?

You are right, I didn't think about the export-subsidy effect of the BAT, so it's not totally equivalent to the tariff. Since this is symmetric, the effect on the exchange rate is clearer as the price change on imports and exports is the same.

The exchange rate will adjust to some extent, but will only adjust fully if nothing else changes. But BAT may affect the capital account. Also, not sure what happens to future expectations, so may be exchange rate overshooting, etc. Basically, exchange rate behavior only follow fundamental based models very tenuously (Meese Rogoff I think) so I am hesitant to make strong statements about exchange rates adjusting fully in accordance with theory. Reality with exchange rates always messy relative to the theory.

An update from this earlier discussion on oil and the border adjustment tax: it only took Goldman Sachs a few months to pick up on what you knew back in February:

Goldman Sachs says U.S. border tax would 'blow out' Brent/WTI crude spread


The comments to this entry are closed.