J.D. Foster J.D. Foster
Former Senior Vice President, Economic Policy Division, and Former Chief Economist


May 29, 2017


Download Document

As the policy debate surrounding border tax adjustments rages on, a closely related debate percolates regarding what a border tax adjustment system would really mean for the U.S. economy. One would think the policy debate wherein choices are made might wait until the economic debate describing the implications of those choices had settled — but then this is a debate in Washington, so no such presumption is appropriate. How exchange rates would respond is a good example of the ongoing debate among economists. Many firm opinions are expressed, but the fact is, no one really knows.

Some blithely assert that “most economists believe that prices will adjust through changes in nominal exchange rates.” No basis for this assertion about what “most economists” believe is offered, nor could there be one because economists who have actually studied the issue have yet to be polled. As a word of caution regarding exchange rates and border tax adjustments, one should always remember Kenneth Kasa’s 1995 observation:

If you asked a random sample of economists to name the three most difficult questions confronting mankind, the answers would probably be: (1) What is the meaning of life? (2) What is the relationship between quantum mechanics and general relativity? and (3) What’s going on in the foreign exchange market? (Not necessarily in that order).

What was true in 1995 remains true today (of all three questions). To be sure, international trade theory provides wonderfully comprehensive models of exchange rate determination in a variety of settings. These models operate with marvelous mathematical precision and elegance. The trouble is, exchange rates rarely behave as theory suggests. Not only do those models typically fail to forecast exchange rate movements, but they typically fail to backcast them well.

For example, basic theory suggests that if the U.S. economy strengthens, then imports would rise more rapidly than exports, leading to downward pressure on the dollar to restore the balance of payments. In contrast, what we usually observe is that a strong economy is associated with a rising dollar. Further, contrary to basic theory suggesting exchange rates should adjust to prevent this, most countries tend to run large trade deficits or surpluses for years, if not decades. Through periods of dollar strength and weakness, the United States has run an almost uninterrupted trade deficit for 35 years. Traditional theory is lovely — it just doesn’t seem to hold regarding exchange rates unless one measures these things across geological epochs.

What follows is intended neither to advance the cause of border tax adjustments nor to thwart the option. Proponents and opponents alike have valid points to make regarding border tax adjustments. The intention here is simply to clarify some issues regarding how exchange rates might react if a border tax adjustment regime were enacted in the United States.

A Few Essentials Upfront

A border tax adjustment is a tax mechanism to level the playing field for cross-border sales of goods and services. In the simplest case, suppose the federal government levied only a domestic business cash flow tax with a border tax adjustment system. Goods and services originating within the United States would face the tax, and those imported into the United States would face tax imposed at the border. All goods and services sold within the United States would then face the same level of tax.

Obviously, goods and services originating abroad from foreign producers and sold abroad would not be subject to the U.S. tax, so to ensure U.S.-source products sold abroad can compete on a level playing field, U.S. tax would be lifted from

U.S. exports. In theory, the border tax adjustment acts like a canal lock, raising or lowering the tax burden so goods and services can have tax neutrality in moving from higher- to lower- or lower- to higher-tax countries.

In practice, a border tax adjustment regime would be more complicated than the simple arrangement here described. However, the complications are not our concern. What is of concern is whether, in response to introducing a border tax adjustment regime, the dollar exchange rate would adjust upward in whole or in part, quickly or over many years. Two arguments are offered to suggest exchange rates would likely move little, if at all, and then only over a long period.

Argument 1: Incidence

Something of a consensus exists on the incidence of the federal business income tax. While Treasury and the Joint Committee on Taxation continue to perfect their respective methods and estimates, they agree on one basic point: Business tax generally falls in some proportion on labor and the owners of capital. This proportion likely varies by industry, and over time, but — critically for our purposes — the consensus holds that the tax generally is not passed on in higher prices.

If the business income tax were reformed, the proportions of the tax borne by capital and by labor would likely change systematically, depending on the nature of the reforms. For example, if reform involved abandoning accelerated for economic depreciation, then the tax burden on the normal return to capital would rise, and so the share of the overall tax borne by capital would likely rise.

Most modern tax reform proposals such as the House blueprint take another path, adopting expensing over accelerated depreciation. Expensing eliminates the tax on the normal return to capital, and so the incidence of the resulting tax likely falls somewhat less on capital than before. Critically, whatever path tax reform takes, business tax still would not be passed on to consumers.

To be sure, a foreign producer selling into the United States and now facing a new, substantial border tax levy would be sorely pressed. Many such producers would be forced to change their business practices substantially or accept deeply cut profit margins, and these decisions would in turn affect the U.S. importers and consumers of their products. This is less an unfortunate byproduct of the policy than its intended consequence, and suggests strongly that if a border tax adjustment regime were introduced, then it should be introduced slowly to allow all parties to adjust.

In specific instances, foreign producers facing a new border tax adjustment regime would at least for a time be able to raise prices to reflect the new regime. For example, a foreign producer selling into a U.S. market in which there are no ready competitive domestic-source products would be more likely to successfully raise prices to U.S. customers. However, even under these circumstances, the foreign producer’s ability to raise prices is likely to be severely limited. If prices could be raised so easily, then why would the producer be content to charge a lower price before the border tax adjustment was instituted? And if prices did rise following a border tax adjustment, competing producers and customers would over time adapt investment and business practices in response to these price increases, thereby creating new downward price pressure on those imports.

Tax incidence is crucial to the exchange rate question involving border tax adjustments because the analysis typically begins with the assumption that tax is passed on in higher prices. If the border tax adjustment were passed on in higher prices on imports and lower prices on U.S. exports, then one would expect upward pressure on dollar exchange rates. However, as shown, this price effect is generally absent because the tax is passed backward to capital and labor. If the tax is not passed on in higher prices on imports and lower prices on exports, then the price effect is missing, and so this particular intuition driving exchange rate movements is inapplicable.

It is easy to understand how so many analysts might go wrong in this regard. Historically, border tax adjustments were included as part of credit-invoice VATs (CIVATs). Those CIVATs are presumed to result in an upward shift in the price level to reflect the added tax — that is, because the CIVAT is in the nature of a sales tax, the tax is then presumed to be passed on in higher prices. Unless one thinks carefully about the disparate incidence patterns of the two different tax types (CIVAT vs. business cash flow tax in the House blueprint), it becomes an easy mistake to assume the price and exchange rate effects expected under a CIVAT-based border tax adjustment regime are mimicked under the plan outlined in the House blueprint.

Argument 2: Relative Proportions

Suppose the incidence analysis is incorrect and that border tax adjustments under a domestic cash flow tax would be passed on in the same manner as assumed for a CIVAT. Certainly, exchange rates would be forced to adjust accordingly, right? Probably not.

The exchange rate implications of trade theory on this point are fairly straightforward. If domestic prices rise broadly because of a sudden increase in the tax wedge or other factor such as a surge in inflation, then one would expect the dollar exchange rate to fall to reset the prior relationship between the U.S. price level and rest- of-the-world price level.

Trade theory is fairly straightforward, but in some respects it doesn’t seem to apply, as evidenced by the fact that the United States has run a substantial trade deficit for nearly all of the last 35 years. Trade theory would tell us the dollar should have considerably fallen long ago to restore the balance of trade. Something fundamental is obviously missing from the theory, and that something is “the rest of the story,” as the great radio newscaster Paul Harvey was wont to say.

The rest of the story involves the other side of the balance of payments ledger — the capital account, specifically the net of gross capital inflows from abroad less gross capital outflows. To put the issue into context, about $5 trillion is traded per day in global exchange markets. About $13 billion of that results from the international trade of goods and services, and another $3 billion is needed for foreign direct investment. Consequently, about 99 percent of all foreign exchange transactions effectively have nothing to do with international trade.

Prices are set at the margin. Exchange rate transactions to accommodate international trade flows are not the marginal trades, meaning whatever price effects occur because of the imposition of a border tax adjustment regime would be swamped by the other forces in play in determining exchange rates.

This may be difficult to accept intuitively. We are taught from the first class in Economics 101 that when demand increases or when supplies decrease, prices usually rise. Surely a border tax adjustment regime, even one adding only incrementally to foreign exchange trades, would create some upward pressure on the dollar exchange rate. To see why not, consider the example of a busy downtown street corner. On three of the four corners can be found a coffee shop, and others are sprinkled throughout the immediate neighborhood. You open a new coffee shop on the fourth corner, thereby increasing the supply of coffee to the sleepy masses. Do you or your competitors expect the price of coffee to decline as a result? Possibly, but not likely. Your contribution to local coffee sales is too slight given the market saturation to make much of a difference.

The issue sharpens further when one realizes many exporters to the United States are paid in dollars and are content to hold those dollars rather than convert them into their local currencies. The U.S. dollar is the premier reserve currency in the world, with much international trade and capital flow occurring entirely outside the United States in U.S. dollars. This means that whatever the border-tax-adjustment-induced exchange rate effect that might otherwise be expected, it is further reduced by that reserve currency effect.


Other dimensions of tax reform aside from a possible border tax adjustment regime may have exchange rate consequences that should be considered. For example, if tax reform reduces the tax bias against saving and investment, then domestic saving and investment patterns may change in different proportions or over different time frames. This, in turn, could alter the pattern of net capital inflows, which in turn could influence the net demand for dollars on foreign exchange markets, depending on how much of any change in net inflows involves foreign-held dollars. Anyone considering those effects has to be impressed by the number of ifs, maybes, and contingencies in any such analysis, all of which play out against a background in which the basic theory is elegant and well defined, yet predicts actual exchange rate movements rather poorly.

The bottom line would seem to be best reflected in the observation that exchange rates would adjust under a border tax adjustment regime, if at all, in the fullness of time. Because that time might best be measured in geological terms, it might be better just to set aside the whole exchange rate debate and focus on the underlying tax and economic issues, which are complicated enough without extraneous discussions intruding unnecessarily.

About the authors

J.D. Foster

J.D. Foster

Dr. J.D. Foster is the former senior vice president, Economic Policy Division, and former chief economist at the U.S. Chamber of Commerce. He explores and explains developments in the U.S. and global economies.