MAR. 20, 2017
Timothy Reichert
Perry Urken
In this report, Reichert and Urken demonstrate that the U.S. adoption of a border adjustment would cause both double taxation of import flows and nontaxation of export flows, contrary to the principles of international tax treaties and the recent OECD base erosion and profit-shifting initiative. Under various scenarios, they argue, this would lead to a significant increase in the overall corporate tax burden of U.S.-based companies.
* * * * *
Table of Contents
I. Introduction: The Blueprint
A. Destination-Based Tax
II. Double Taxation and Nontaxation Under BA
A. Imports
III. Tax Treaties and Double Taxation
IV. The Uncertain Effect on Corporate Taxes
A. The BA With Export Credit
B. The BA Without Export Credit
V. The Possibility of Tax Policy Retaliation
VII. Transfer Pricing Under BA
IX. Appendix A: Results With No Export Credit
A. IP Ownership in Destination Country
B. IP Ownership in Source Country
C. IP Ownership in Destination Country
D. IP Ownership in Source Country
X. Appendix B: Results Under BA Equilibrium
In June 2016 House Republicans issued a "blueprint," which proposed a broad array of overhauls to the tax code.1 The blueprint suggests a sweeping redesign of the current corporate tax framework, which it calls a destination-based cash flow tax (DBCFT).
A DBCFT is not intuitive, and the DBCFT framework is not part of the common tax policy lexicon. Given the concept's novelty, and because the main characteristics of the blueprint's corporate tax proposal are a function of it, the analysis below begins by outlining the main principles of a DBCFT-based tax policy.2
The blueprint proposes that corporate tax be assessed exclusively based on where products and services are consumed. As proposed, this means that if a U.S.-based business exports products for sale abroad, value associated with these exports would not be taxable in the United States. In contrast, value associated with products sold by a U.S.-based business in the U.S. market, and the value of products sold by foreign businesses to the United States (that is, imports), would be taxable in the United States. Thus, tax would apply to both domestically produced and imported goods sold to U.S. customers.
While the destination-based nature of the tax is similar to a sales tax, the tax is not assessed on revenue. Nor is it assessed on business profits reported by companies. Rather, it is based on a cash flow concept. Cash flow, as presented in the blueprint, is defined as revenue from domestic sales less a set of deductions limited to corresponding cash outflows for domestic salaries, inputs, and capital investments.3 While cash flow is generally defined as the cash left over in each period after a business deducts all of its cash outlays from its revenue, the destination-based feature of the blueprint means that businesses would not be permitted to deduct amounts paid for foreign-based inputs from domestic revenue. Nondeductible foreign-based inputs include finished goods, components, capital equipment, services, and intangible property, procured from non-U.S. sources.4
This report examines the key characteristic of the DBCFT's destination-based concept -- namely, the nondeductibility of imports and nonrecognition of exports. The blueprint calls it a border adjustment (BA). The border adjustment, as the name implies, is an adjustment made to prices received or paid for exports or imports, respectively. As proposed, the border adjustment adjusts both import and export prices to zero. Other components of the blueprint, such as the proposed reduction in the statutory corporate tax rate and the expensing of capital investments, are not directly addressed in the following analyses.
Generally, double taxation occurs when income earned by a business enterprise is taxed by more than one tax authority.5 It is straightforward to demonstrate that the border adjustment represents double taxation of import value and complete nontaxation of exports. Correspondingly, for non-U.S. trading partners, the border adjustment gives rise to double taxation of exports to the United States and nontaxation of imports from the United States. The resulting double taxation and nontaxation are inconsistent with the principles of international tax treaties.
Examining the border adjustment's effects on imports first, we see that before implementing the border adjustment, import prices would be subtracted from a U.S. importer's revenue when arriving at taxable income. That is, but for the border adjustment, import values (that is, import prices) would be a tax deduction. This deduction is eliminated through the border adjustment.
The imported goods would generally be resold at a markup or profit.6 This resale price is of course taxable. However, because the value (price) of the imported good is not deductible, taxable cash flows are higher by the amount of the import price. Thus, the import's value is contained in taxable cash flow and is therefore taxed in the United States.
Now consider the foreign counterparty selling the imported good to the U.S. importer -- that is, the foreign exporter. The foreign exporter will recognize the export price (the U.S. importer's import price) as revenue and subtract as tax deductions the value of any associated components, labor, and capital used to produce the export good.
At first blush, this seems to imply that what is taxed on the non-U.S. side of the transaction is limited to the profit associated with the export to the United States. That is, the seeming implication is that there is double taxation only of the capital returns embedded in the export to the United States (that is, the profit on the exporter's side is taxed, which is only a portion of the U.S. importer's price, which is taxed in full). However, that implication is incorrect -- what is double taxed is the portion of the price of the product imported into the United States that is attributable to inputs with an origin outside the United States. To see this, we simply need to recognize that the foreign exporter's deductions (costs) include purchased components and capital from other non-U.S. companies, as well as local labor deductions, all of which are recognized and taxable in non-U.S. jurisdictions -- through corporate or personal income taxation. Those components and capital assets, in turn, include the value of non-U.S. manufactured components and capital purchases, which are recognized as taxable abroad -- and so on. Applying that logic to the entire supply chain shows that of the import price paid by the U.S. importer, the entire portion resulting from non-U.S. labor and capital inputs is double taxed.
Thus, we see that in a global tax system in which almost all corporate income is taxed at its source -- that is, income is taxed where it is created -- the unilateral adoption by the United States of the border adjustment to import prices gives rise to double taxation.
Currently, to determine taxable income, a U.S. exporter would recognize the value of exported goods (price received multiplied by quantity sold) as revenue and would deduct the value of any associated components, capital depreciation, labor, and other inputs used to produce the exported product. However, the border adjustment eliminates the requirement to recognize the price of the exported product because the border adjustment does not assess tax on export flows. Thus, implementing the border adjustment in the United States would result in the value of U.S. exports not being taxed by the IRS.
The foreign counterparty (the foreign importer) that purchases the U.S. exporter's product can deduct the price paid for the U.S. exporter's product. Its value is not included in taxable income. Thus, while the border adjustment results in double taxation of imports, it also gives rise to complete nontaxation of exports. Neither the IRS nor the foreign counterparty's tax authority taxes the value of exports from the United States under the border adjustment.
Depending on the details of potential future Treasury regulations, the U.S. exporter might still be able to recognize the costs (deductions) pertaining to capital, labor, and components that it used to produce the exported good in calculating its U.S. tax base, even as the income received from the exports is untaxed. If so, the U.S. exporter would have negative taxable cash flows pertaining to its exports that could, in principle, be applied as an offset to (deduction or credit against) profits realized on domestic and import-related activity of a U.S. taxpayer. However, the availability of this credit is unclear from the blueprint. For example, it is possible that a destination-based framework will be interpreted so that the notional losses on a U.S. exporter's cash flow statement are unavailable for use as a credit or deduction against profits on the company's import and domestic activity, simply because export activity may be interpreted as falling outside the taxing jurisdiction of the IRS in a destination-based framework. Below we address the use of such deductions and the relationship between using losses and U.S. companies' effective tax rates.
As noted, when the same income earned by a business enterprise is taxed by more than one tax authority, there is double taxation, which the international community views as undesirable. This view is directly reflected by the OECD, which states: "Double taxation is undesirable and should be eliminated whenever possible, because it constitutes a potential barrier to the development of international trade and investment flows."7
To prevent double taxation, countries have entered into an extensive network of tax treaties that govern negotiations regarding transfer pricing disagreements that may arise between the countries and the taxpayers that make transfers between them.8 The United States is a party to 68 separate bilateral tax treaties.9 A study published in 2007 found that 2,351 tax treaties were in place globally and that the parties to these treaties encompassed nearly all OECD countries and accounted for a large proportion of global foreign direct investment flows and stocks.10
Many of those treaties have terms that are based on three prominent model tax conventions, which are published by the OECD, the Treasury Department, and the United Nations. These models are intended to serve as templates for other nations to use in drafting their own treaties.
International tax treaties generally contain an extensive framework for defining processes and standards by which international tax disputes can be resolved in a manner that avoids double taxation. All three conventions adopt the arm's-length principle as the standard for that purpose. The arm's-length principle, as reflected in these conventions, holds that taxation of controlled affiliates of multinational groups operating within their respective jurisdictions should be based on the business profits that each affiliate might be expected to earn if it were a distinct and independent enterprise engaged in the same or similar activities under the same or similar conditions.11
In 1928 the United States pioneered basing transfer pricing administration on the arm's-length principle. In 1988 Treasury published a white paper that endorsed the arm's-length principle as an appropriate way to administer transfer pricing, and most countries have adopted that principle. The OECD calls application of the arm's-length standard "the international consensus on transfer pricing." In fact, our research indicates that countries that administer international transfer pricing based on the arm's-length principle account for more than 96 percent of global GDP.
Over decades of application, multinational corporations (MNCs) and tax authorities have developed common understandings of how the arm's-length principle should apply. The OECD observes that "experience under the arm's length principle has become sufficiently broad and sophisticated to establish a substantial body of common understanding among the business community and tax administrations. This shared understanding is of great practical value in achieving the objectives of securing the appropriate tax base in each jurisdiction and avoiding double taxation." While substantive concerns regarding the administrability of the arm's-length principle have been raised, the international consensus reflected in the OECD's recent base erosion and profit-shifting initiative, in which the United States has played a leading role, is that the best course of action is to attempt to build on this "body of common understanding" by improving the application of the arm's-length standard to address these concerns. Largely because of the BEPS initiative, the arm's-length standard has begun to evolve in a manner that targets the elimination of global nontaxation of corporate income along with the avoidance of double taxation.
Putting aside questions whether the border adjustment is consistent with international norms and treaties, a pressing question is how it will affect overall corporate taxes paid. We can begin to understand this using a simple algebraic model that compares corporate taxes under the existing system with corporate taxes given imposition of the border adjustment.
We begin with some assumptions and definitions. First, to focus our analysis solely on the effect of the border adjustment on corporate tax burdens, we assume that cash flow and profit are the same. That is, we isolate the effects of the DBCFT's border adjustment from the effect of its proposal to move from a tax on profits that are calculated using accrual accounting concepts (for example, matching of capital depreciation with revenue) to a tax on cash flows (for example, immediate deduction of capital expenditures).12
Second, we initially abstract from three important questions regarding the prices surrounding imported goods: (1) whether the increased tax cost to shareholders stemming from the nondeductibility of imports is passed on to customers (and if so, to what degree); (2) whether the tax will be partially borne by the exporters from whom imports are purchased (and if so, to what degree); and (3) whether the tax on imports and subsidy on exports are reflected in a more expensive U.S. dollar (dollar appreciation), and if so, to what degree. Once we have modeled the basics of the border adjustment, we incorporate those considerations.
Third, our model considers U.S. enterprises to have three divisions: a domestic division; an export division; and an import division. We assume that cash flow statements can be constructed for each, and refer to this process as "divisionalization."
As noted earlier, depending on how Treasury regulations are written, divisional cash flow statements might consolidate into a single divisional cash flow statement. That is, it may be that Treasury regulations embodying the border adjustment will require separate calculation of the taxable cash flows of the company's export, import, and domestic divisions. It is plausible, for example, that the destination-based features of the DBCFT will be interpreted so that revenue and costs attributable to U.S. exports (that is, all export activity) are deemed to constitute nonrecognized revenue and costs for tax purposes -- that is, neither would factor into the determination of U.S. taxable cash flows. Alternatively, the regulations could be written to effectively adjust export revenue to zero while maintaining the ability to use the costs of the labor, components, and capital inputs used for exports as a credit, or deduction, against the taxable cash flows reported by the import and domestic divisions.
Thus, the ability to use an "export cost credit" represents a key issue for the DBCFT.13 To draw attention to this important issue, we model the border adjustment with and without the export credit.
We begin by defining variables. For each division (domestic, export, and import), the variables fall into six standard economic categories: prices, quantities, variable costs per unit, fixed costs, capital stocks, and rates of return:
PD = Quantity weighted average price of the domestic division's products
QD = Quantity sold by the domestic division
PX = Quantity weighted average price of the export division's products
QX = Quantity sold by the export division
PM = Quantity weighted average price of the import division's products
QM = Quantity sold by the import division
a = Domestic division variable cost per unit
b = Fixed cost, domestic division
c = Export division variable cost per unit
d = Fixed cost, export division
e = Import division variable cost per unit
f = Fixed cost, import division
RD = Economic rate of return on capital stock of domestic division14
KD = Capital stock of domestic division15
RX = Economic rate of return on capital stock of export division
KX = Capital stock of export division
RM = Economic rate of return on capital stock of import division
KM = Capital stock of import division
Those variables allow us to form a simple version of the cash flow statements for each division as shown below.
PDQD- aQD- b (equiv) RDKD, or cash flow for the domestic division
PXQX - cQX - d (equiv) RXKX, or cash flow for the export division
PMQM - eQM - f (equiv) RMKM, or cash flow for the import division
The above expressions are identities because the capital stock for each division is defined as historical capital expenditures less economic depreciation. That is, cash revenue minus cash costs is by definition equal to the company's realized internal rate of return on its in-service capital investments (capital investments less economic depreciation) multiplied by its historical capital investments depreciated using economic depreciation.16
The effect of the border adjustment can be seen, under divisionalization with an export loss credit, by subtracting taxes but for adoption of the border adjustment from taxes given adoption of the border adjustment. The term t stands for the tax rate and the term stands for incremental taxes given adoption of the border adjustment:
Equation 3 tells us that the incremental tax paid after adoption of the border adjustment, given divisionalization with an export division loss credit, is equal to the tax rate multiplied by the difference between the value of imported goods and the value of exports. That is, companies whose import costs are greater than their export revenue will experience a tax increase following adoption of the border adjustment, and vice versa. As is intuitive, equation 3 also tells us that a credit on the export division's cash losses is far more attractive than divisionalization without credit. Crediting export losses means that the relative values of import and export shipments determine the effect of the border adjustment on total taxes paid.
Equation 3 implies that the condition, for each company, under which the border adjustment causes no incremental tax is:
That is, the value of imports -- that is, their weighted average price multiplied by quantity -- is equal to the value of exports.
This, in turn, allows us to derive the following expression:
- Equation 5: QX= (e/PX)QM
Figure 1. Graphical Depiction of Required Export
Quantity for a Firm with Imports of QM
Figure 1 shows a company with imports of QM, a weighted average import price of e, and a weighted average export price of PX. QX(A) in Figure 1 represents the company's actual exports, and QX the level of exports needed in order for the company's tax burden to be unaffected by the border adjustment. This implies that the company's "export gap" -- that is, the additional quantity of exports needed for the company to experience no increase in taxes because of the imposition of the border adjustment -- is equal to QX - QX(A). Companies with positive export gaps will, all else being equal, experience tax increases of tPX(QX - QX(A)).
Figure 1 is intentionally drawn with an import price to export price ratio greater than 1 and a ratio of imports to exports of roughly 3. That is because, as we show in Appendix C, that ratio is consistent with the aggregate relationship between MNCs' related-party imports into the United States and their U.S. exports to affiliates. While the data available to us provide only the relative values of MNC imports and exports (price times quantity), we know from our own practice with large MNC clients that the slope (e/PX) and quantities shown are consistent with the position of many large MNCs. That is, many MNCs have very large export gaps.
Proponents of the border adjustment argue that it will ultimately have a muted effect on taxes -- that is, tPX(QX - QX(A)) will, on average and over the intermediate to long run, be small. Their argument is that market adjustments will occur that will offset the incremental border adjustment tax burden faced by U.S. companies. These include appreciation of the U.S. dollar relative to the currencies of our trading partners as well as increased exports and decreased imports. These effects are shown in Figure 2:
Figure 2. Currency Changes and Changes in Import
and Export Quantities: Effects on the Export Gap
Figure 2 shows a two-step, two-market response. First, there is a currency market response to the border adjustment. That is, proponents and other commentators have argued that there will be an appreciation in the U.S. dollar. This, in turn, decreases the cost of imported products -- as shown by the change in slope of the terms of trade line from a ratio of (e/PX) to , where is the new weighted average import price and KX is the new weighted average export price. This currency market change has the effect of decreasing the export gap. Second, a goods market response is shown, in which imports decline and exports rise. This, too, has the effect of decreasing the export gap.
Other commentators have noted that a portion of the increase in the cost of imports, and a portion of the decrease in the cost of exports, will be passed on to customers. That has the effect of muting the U.S. dollar appreciation shown in Figure 2. That is, to the extent that export cost decreases are passed on, the increase in demand for U.S. dollars is muted. Similarly, to the extent that import cost increases are passed on, the decrease in the derived demand for imports (and therefore foreign currency) is muted.
Figures 1 and 2 also illuminate an important story about the variability and uncertainty of the border adjustment's effect on corporations' tax burden. In short, its effect is highly variable across companies and highly uncertain over time. First, companies can be thought of as differing according to their relative import prices and quantities as well as their relative export prices and quantities. To illustrate, if we simplify by assuming prices and quantities are either high or low, companies can have high or low import prices, high or low import quantities, high or low export prices, and high or low export quantities -- a total of 16 combinations.17
Second, companies differ according to their exposure to different foreign exchange markets. In some cases, companies face foreign exchange markets that float freely and quickly in response to trade flows, and in others they face markets that are arguably manipulated or are more strongly influenced by monetary policies (interest rates) than by trade flows. Third, companies differ in the demand elasticities that they face -- both the downstream elasticity of demand for resold imports or products with high import content, and the elasticity of demand for exports. These relative elasticities differ across products and markets and determine the strength and direction of step 2 in Figure 2 above.
Our experience confirms this. Our firm's clientele includes companies with high export prices and quantities, coupled with low import prices and quantities; companies with "low-low" exports and "high-high" imports; and every combination in between. Similarly, our clientele contains companies whose treasury departments have very different views of the degree to which foreign exchange responses are likely to occur in response to the border adjustment (dependent largely on the countries in which their related affiliates operate), and sales and procurement personnel with divergent views of the degree to which they can or should pass on cost changes resulting from the border adjustment. Thus, our modeling of the border adjustment's effect, and the nature of the dynamics considered by our clients, confirms that the economic uncertainty generated by the border adjustment is extremely high.
One implication of this is that the border adjustment is not simply a new feature of our tax policy; it is also trade policy, given that trade policy can be defined as the set of government policies, regulations, standards, and rules that directly affect international trade. Further, it is trade policy that will affect all companies, rather than being confined to companies operating in markets that protectionists care about. Finally, it is trade policy that has a highly uncertain effect at the level of individual companies -- involving numerous variables with a predictability, especially taken together, that is low.
In light of that, it is puzzling to hear statements that the border adjustment is "a very good deal" and that the border adjustment "will not affect trade because it will be offset by foreign exchange rate changes." We would ask in reply, "How can you know?" Surely the variation across companies' terms of trade, expected foreign exchange rate changes, quantities of imports and exports, and price elasticities together means that the effect of the border adjustment is less certain than the effects of other more traditional and targeted forms of protectionist trade policy (that is, tariffs). Yet precisely because the border adjustment is tax policy rather than trade policy, that uncertainty would be placed on all companies.
B. The BA Without Export Credit
As noted, it is possible that Treasury regulations regarding the border adjustment will not allow the deductions (labor, component, and capital costs) associated with the export division to be taken against import-related cash flows. While the prevailing view is that this credit will be allowed, some have commented that if economists who assert that foreign exchange rate changes will offset the border adjustment are assumed to be correct, those who have a trade-policy-related (that is, protectionist) motivation for the border adjustment will have an incentive to interpret the "destination-based" concept within the DBCFT as liberally as possible. That is, they will have an incentive to disallow the use of losses from the export division against the now higher taxable cash flows of the import division.18
Also, while these credits are incorporated in the analyses of publications from proponents of the border adjustment,19 it is worth noting that the blueprint neither endorses, nor mentions, the availability of export credits. In fact, the blueprint states that under the border adjustment, "tax jurisdiction follows the location of consumption rather than the location of production."20 If interpreted strictly, that could be construed as implying that under the border adjustment, the U.S. government would not have taxing rights for U.S. companies' export operations because they would fall outside the scope of jurisdiction defined by the destination-based framework. In traditional source-based territorial taxation systems, governments do not maintain taxing rights for the profits reported by MNCs' foreign affiliates because they fall outside the defined territory of taxation rights. The profits (or cash flows) of U.S. companies' export activities would similarly seem to fall outside the U.S. taxation territory as defined by the destination-based framework outlined in the blueprint.
If so, the model given above changes in the following way. Once again, the effect of the border adjustment can be seen by subtracting taxes but for adoption of the border adjustment from taxes given adoption of the border adjustment. This leads to a modified version of equation 1:
As seen earlier, equations 7 and 8 expand and then simplify equation 6.
Equation 8 tells us that the incremental tax paid after adoption of the border adjustment, given no export division loss credit, is equal to the tax rate times the difference between the value of imported goods and the cash flow on exports.
It is obvious from equation 8 that disallowance of the export losses (no export loss credit) would give rise to a much steeper "required exports" line than that shown in Figure 1, all else being equal. This is illustrated below.
First, equation 8 tells us that absent the export loss credit, the condition that must be satisfied in order for incremental taxes to be zero is:
That is, the value of imports -- that is, their weighted average price multiplied by quantity -- must be equal to the cash profit earned on exports. We can express export cash profit per unit as = RXKX/QX, implying:
This in turn gives us the slope of the required exports line as before:
All else being equal, equation 11 will have a much steeper slope than equation 5. This is because the quantity weighted average profit per unit of exports will always be lower than the weighted average price per unit of exports. This is shown in Figure 3, wherein the dark line represents equation 11 and the light line represents equation 5 as before. As shown, all else being equal, the export gap will be higher for all companies, given no export loss credit.
Figure 3. Export Loss Credit Versus
No Export Loss Credit
The following table illustrates this using a hypothetical example of an MNC with domestic, import, and export activities in the U.S. market. The example given assumes that the revenue received by the MNC from customers for each of the three segments is equal. For simplicity, the rate of return on capital realized by the company is also assumed to be consistent across these segments. Lines 7 and 8 present the derivation of taxes payable by the company under the current tax regime (as governed by international treaties to which the United States is a party) assuming a 35 percent tax rate. Lines 9 and 10 present the company's taxes assuming the implementation of the border adjustment by the United States given divisionalization and no export loss credit. As shown, the tax base for the import division under the border adjustment is higher than under the current regime by the value of imports, which are nondeductible under the proposal. The export division is not taxed under the border adjustment, and thus the tax faced by the company is lower on these activities than under the current regime.
Line 11 derives the difference between the total taxes due under the current and border adjustment regimes (an increase of $24 in our example). In line 12, we calculate the same value using the formula presented in equation 8 above. Thus, = t(eQM - RXKX) = tax rate (value of imports - cash flow of exports) = 0.35(76-7) = 24.
Table 1. Hypothetical Example: Impact on Corporate Tax Burden
From the Implementation of the BA
______________________________________________________________________________
Line Item Domestic Imports Exports Total
______________________________________________________________________________
1 Revenue 100 100 100
2 COGS 73 76 88
3 Op expenses 18 22 5
4 Business profit 9 2 7
5 Capital 88 22 70
6 Return on capital (line 4/line 5) 10% 10% 10%
Current tax regime
7 Taxable flows 9 2 7
8 Tax (line 7 * 35%) 3 1 2 6
Border adjustment
9 Taxable flows 9 78 0
10 Tax (line 9 * 35%) 3 27 0 30
11 Difference (line 10 - line 8) 24
12 Difference using equation (3) [imports - export 24
profits] * 35%
Equation 8 demonstrates that given divisionalization without credit, companies whose import costs are greater than their export profits (cash flows) will experience a tax increase following the adoption of the border adjustment by the United States, and vice versa. This is consistent with the notion expressed in some of the commentary on the border adjustment -- specifically, that MNCs, being on the average very large net importers, can be expected to experience a material tax increase from the unilateral implementation of the border adjustment by the United States. Importantly, however, equation 8 tells us that without an export loss credit, many U.S. companies that are not net importers will still experience an increased tax burden. In the hypothetical example above, third-party sales of imported product is equal to that of exports, and overall taxes under the border adjustment are still roughly five times higher than under the current regime.
Appendix A provides results for MNCs given a version of the border adjustment without an export loss credit. These show that under reasonable assumptions, MNCs with export flows from the United States that are well more than double that of their U.S. imports would still realize a material increase in their overall tax burden because of the implementation of the border adjustment. The models in Appendix A also show that if we assume that rights to all intellectual property used in cross-border transactions are owned in the jurisdiction of the products' destination,22 even MNCs with revenue from U.S. exports that is more than seven times higher than their revenue from U.S. imports will bear an overall tax increase from the border adjustment.
As referenced above, however, the value of MNC U.S. affiliate imports (that is, imports procured by U.S. affiliates from their non-U.S. affiliates) has consistently been much larger than the value of these U.S. affiliates' exports to non-U.S. affiliates. As summarized in Appendix C, data compiled by the Census Bureau from customs filings on U.S. related-party imports and exports demonstrate that the aggregate value of U.S. imports from affiliates is generally just under three times as large as the value of U.S. exports to affiliates. For example, in 2013 related-party imports into the U.S. market were about $1.1 trillion, while the value of exports made by U.S. entities to foreign affiliates was about $390 billion. By using the same assumptions presented in Table 1 above, but changing the relative scale of import and export flows to reflect the actual aggregate relationship between related-party imports and exports, we also show in Appendix A that the tax burden realized by the hypothetical company increases to a level that is about 11 times higher than under the current regime.
The potentially high incremental tax burden generated by the border adjustment may indeed provide sufficient incentives for companies to relocate some of their foreign manufacturing operations to the United States.23 However, the unilateral imposition of the blueprint by the United States would also create an inconsistency between U.S. tax policy and the terms of existing treaties. It seems reasonable to believe that foreign governments may be motivated to address both the double taxation borne by exporting MNCs based in their jurisdictions and the nontaxation of exports from the United States.
One way that foreign governments could address these concerns is to enact their own version of the border adjustment. In fact, leaders from several countries including China have already threatened to retaliate if the United States implements policies deemed to be protectionist in nature.24
If foreign trading partners implement regulations with similar provisions to the border adjustment, one can conceive of an eventual result that we call "BA equilibrium." In BA equilibrium, all major trading partners subsidize exports and penalize imports through border adjustments. Assuming that all countries adopting a border adjustment adjust import and export prices to zero, a company's effective tax rate will equal the weighted average of the corporate tax rates in each country in which it has import and domestic operations, weighted by the cash flows earned by these operations (net of any applicable credits for the export division's losses).
The effect of this on the corporate tax burden would depend on whether net importing and large domestic market countries have higher or lower tax rates than net exporting and small market countries. If net import and large domestic market countries have higher corporate tax rates than net export and small domestic market countries, corporate effective tax rates will rise. Note that the United States has the largest domestic market, is a very large net importer, and has one of the highest corporate tax rates.
Putting tax rate differentials aside (that is, assuming tax rates are the same across jurisdictions for purposes of isolating the impact of the border adjustment on companies' tax burdens), companies' tax burdens will be unaffected by their respective balances of imports into and exports from the United States. This will be the case when an export credit is available and when it is not.
However, the availability of the export credit is enormously important to companies' tax burdens in BA equilibrium. If an export credit is available, the border adjustment will not have any effect on the overall tax burden relative to the current tax regime.25 This is because the penalties assessed on imports in one jurisdiction are effectively offset by the use of exports as a credit in the other jurisdiction (assuming the same transfer price is respected for a given transaction in both jurisdictions).
However, for a scenario in which an export credit is unavailable, the tax burden for all companies engaged in material degrees of international commerce will likely be extremely high. As demonstrated in Appendix B, under those circumstances the tax burden will vary based on two factors: (1) the consolidated operating cash flow margin realized by the company regarding cross-border transactions; and (2) the magnitude of the importing affiliate's operating expenses relative to revenue that, under the border adjustment, can be deducted in the destination market. All else being equal, nondeductible prices paid for foreign inputs by importers will represent a lower share of revenue and business profits for companies with higher consolidated profit margins for a given cross-border transaction, which reduces the effect of the resulting double taxation on imports. Higher levels of value added contributions in the destination market associated with import activities also could generally be expected to lead to reductions in tax burden under the border adjustment, in contrast to transactions in which most value added is performed by the exporter (which leads to relatively higher nondeductible prices for imports). Appendix B presents analyses that demonstrate the potential for an extremely high tax burden under BA equilibrium without export credit and how this burden varies given the levels of the two variables defined above.
Some proponents of the border adjustment have argued that the tax increases described above are unlikely to affect the dollar value of trade balances because the border adjustment will cause the U.S. dollar to appreciate in response to increased U.S. exports and decreased U.S. imports. That is, exchange rate (that is, price) movements will offset the change in net exports (that is, quantity).
While this argument is not necessarily inconsistent with exchange rate theory, we are not entirely convinced by it. There are several reasons for this.
First, the arguments in favor of this offsetting exchange rate movement have not been entirely clear. That is, proponents of the exchange rate offset argument have not been clear as to whether they are arguing that exchange rates will immediately adjust in anticipation of increased net exports or that they would adjust ex post in response to actual trade patterns (increased net exports).
The argument that exchange rates will fully adjust ex ante, and thus the pattern of trade will be unaffected, assumes that the border adjustment's effect on trade flows can be inferred by the market. Knowing firsthand the complexity of large MNC's trade patterns and tax structures, we believe that this is highly unlikely.
Alternatively, if the argument is that exchange rates will adjust after a change in net exports has been induced by the border adjustment, we find it even less appealing. This argument is effectively saying that the border adjustment is protectionist and not protectionist at the same time. That is, it is saying that the border adjustment will first cause the pain of high effective tax rates for importers; this will then change trade patterns (increase net exports); and this change in trade patterns causes dollar appreciation that will alleviate the pain of higher effective tax rates. This seems to be a sort of intellectual sleight of hand -- someone must bear the cost of tax-induced changes in trade patterns or there would be no change in trade patterns.
We note finally that exchange rates are not well behaved, and their movement is not fully understood by economists. We do know that exchange rate movements are strongly influenced not only by the supply and demand of exports and imports (that is, a country's net exports), but also by monetary policy and other factors that economists have not been able to reliably identify. Thus, it is uncertain that U.S. dollar appreciation will readjust the border adjustment -- ex ante or ex post.26
Proponents of the border adjustment have argued that one of its merits is that it decreases the importance of intercompany transfer prices for MNCs. That is, because the transfer prices of imports from, and exports to, controlled foreign affiliates are adjusted to zero, the border adjustment eliminates or decreases the incentives to manipulate transfer prices.
This argument is true only in a superficial sense. It is true that the transfer prices for imports and exports are adjusted to zero -- that is, they are always equal to zero when deriving the taxable cash flows of a U.S. taxpayer. However, because the border adjustment causes double taxation of imports and nontaxation of exports, the transfer price is the measure of import double taxation and a primary determinant of the level of export nontaxation. This simple fact creates a number of incentives related to intercompany transfer pricing. Assuming the availability of an export credit and that the border adjustment is applied unilaterally by the United States, transfer prices remain central to the determination of MNC tax burdens. Thus, under these circumstances MNCs will have a strong incentive to move any intangible property to the U.S. market that pertains to imported goods to the United States from foreign affiliates. This will have the effect of lowering import prices. Equation 3 shows us the advantage of these transfer price amendments. Multinationals will also have a strong incentive under these circumstances to locate intangible property ownership rights for U.S. exports in the U.S. market, since this will have the effect of increasing the value of nontaxed export revenue.
Under divisionalization with no export loss credit, there will be a strong incentive, all else being equal, to shift costs from the export cash flow statement to the import and domestic divisions' cash flow statements. Also, the incentives present when export credits for the domestication of intangible property rights for U.S. imports are available would also apply in this scenario. Only under the BA equilibrium scenario, in which both the United States and its trading partners adopt border adjustment policies, is the importance of transfer pricing truly eliminated.
This report has demonstrated that the adoption of the border adjustment by the United States will cause both double taxation of import flows and nontaxation of export flows. This is inconsistent with the principles of international tax treaties currently in place.
Our analysis demonstrates that the border adjustment will give rise to a significant and highly variable increase, across companies and over time, in the overall corporate tax burden. This increased burden would be pronounced for companies whose imports into the United States exceed their U.S. exports. However, we show that if losses in companies' export divisions are not credited against positive cash flow in their import divisions, the border adjustment would cause a tax increase even for those companies whose U.S. exports materially exceed the scale of their imports into the United States.
Because the border adjustment is essentially trade policy effected through tax policy, it is conceivable that other countries will adopt similar tax policies -- tantamount to trade retaliation. We show that if all countries adopted a border adjustment and adjusted import and export prices to zero, a company's effective tax rate will equal the weighted average of the corporate tax rates in each country in which it has import and domestic operations, weighted by the cash flows earned by these operations (net of any applicable credits for the export division's losses). If no export loss credits are available in a BA equilibrium, even in some countries, general adoption of a border adjustment will give rise to very high effective tax rates.
Separately, we demonstrate that transfer pricing considerations will remain relevant under the border adjustment, contrary to assertions made by some. We also express skepticism in the argument that exchange rate adjustments will offset the effects of the border adjustment on import-intensive businesses.
In this appendix, we expand our analysis of the conditions under which MNC overall tax burden will be unchanged given adoption of the border adjustment with no export loss credit.27 This discussion assumes a set of transactions involving a U.S. distributor of imports from a foreign affiliate (the import division) and a U.S. manufacturer exporting to a foreign affiliate (the export division).28
We begin with equation 8, introduced in the body of the paper:
Next, we rearrange the equation to derive the following condition whereby the change in tax burden is zero:
(eQM- RXKX)
Next, we use equations that present the value of imports, the cash flow of exports, and other key terms as percentages of revenue. This allows us to think in terms of profit and cost margins when solving for the relationship between exports and imports in the no incremental tax condition.
We define the following variables:
SI= Third-party revenue for imported products
SE = Third-party revenue for exported products
OE%D = Distributor expenses as percentage of revenue
OM%C = Consolidated operating margin (for subject transaction)
MD = Markup for routine distribution functions
MM = markup for routine manufacturing functions
Below, we use these variables to derive the value of imports. We present two alternative formulas for this computation. Both formulas assume, as is commonly the case, that IP rights give rise to any rents in the consolidated MNC system. The first formula assumes that such IP is owned in the country of destination for the product. The second assumes that the IP rights for the exports are owned by the U.S. exporter.
A. IP Ownership in Destination Country
Value of Imports = [Revenue] - [U.S. Distributor Expenses] - [U.S. Distributor Profit]
Value of Imports = [SI ] - [OE%DSI ] - [(OM%CSI) - (SI - OM%CSI - OE%DSI)MM ]
B. IP Ownership in Source Country
Value of Imports = [Revenue] - [U.S. Distributor Expenses] - [U.S. Distributor Profit]
Value of Imports = [SI ] - [OE%DSI ] - [OE%DSIMD ]
We can now use the same variables to derive the cash flow of exports. Again, we present two alternative formulas.
C. IP Ownership in Destination Country
Cash Flow of Exports = [U.S. Value - Added Cost of Exports] * [Markup]
Cash Flow of Exports = [SE - SEOM%C - SEOE%D ] [MM ]
Because in this formula we assume that the IP ownership is in the destination country, the profits for the U.S. exporter are derived by applying a routine return on value added expenses incurred (akin to the cost of capital applied to the economically depreciated capital investments).
D. IP Ownership in Source Country
Cash Flow of Exports = [Consolidated Profits] - [Non-U.S. Distribution Profits]
Cash Flow of Exports = [SEOM%C ] - [SEOE%DMD ]
In Alternative 2, given that the IP ownership is in the source country, the export cash flow incorporates a return to this IP.
Next, we combine the above equations such that the value of imports is equal to the cash flow of exports, beginning with Alternative 1.
Rearranged, this gives us:
Which yields:
The ratio of SE relative to SI as derived above represents the relative scale of the value of third-party sales of exported products compared with that of third-party sales of imported products, which generates no change in tax burden for a given MNC. As presented below, the ratio varies depending on which of the two alternatives is used (that is, where IP is held). With the use of Alternative 2 (which assumes that IP rights are held by the exporter, as opposed to the importer), the ratio varies depending on both the consolidated operating profitability of the cross-border transaction and the level of the importer's deductible expenses relative to sales.
The following table shows that given reasonable assumptions for these two inputs, we derive ratios where exports exceed imports by factors of between roughly 2.5 and 5. Thus, even companies whose exports are more than double their imports will see a significant increase in their overall tax burden under the border adjustment, if no export loss credit is provided:
Table A-1. Ratio of Third Party Revenues of Exports Relative to Import
Revenues at Which the Aggregate Impact of the BA Is Zero
(IP Owned by Exporter)
______________________________________________________________________
Consolidated Operating Margin
_____________________________
Importer Deductions As
Percentage of Sales 20% 30%
______________________________________________________________________
10% 4.9 3.1
30% 4.6 2.7
However, when we assume that the IP rights are owned by the importer, the ratio is unaffected by variations in the consolidated operating profitability of the cross-border transaction or the level of the importer's deductible expenses relative to sales. Given this, under the assumption that IP rights are held in the destination jurisdiction and using the same parameter assumptions used to derive Table A-1, we find that if IP rights are owned in the destination country (that is, by the importer), companies must have export revenue about 7.7 times higher than imports for the border adjustment to result in no increase in the overall tax burden.
But this ratio of 7 to 1 in terms of exports relative to imports is far from the average ratio for MNC activities for the U.S. market. Aggregate U.S. import value from affiliates is about three times the aggregate value of U.S. exports sold to affiliates. Table A-2 illustrates a scenario in which the hypothetical example presented in Table 1 of the report is adjusted so that imports are three times higher than exports.
Table A-2. Example Using Actual Ratio of Aggregate
U.S. Related-Party Imports and Exports
______________________________________________________________________________
Row Item Domestic Imports Exports Total
______________________________________________________________________________
a Revenue 100 300 100
b COGS 73 227 88
c Op expenses 18 66 5
d = a - b - c Business profit 9 7 7
e Capital 88 66 70
f = d/e Return on capital 10% 10% 10%
Current tax regime
g = d Taxable flows 9 7 7
h = g*35% Tax (at 35 percent) 3 2 2 8
Border adjustment
i = d + imports Taxable flows 9 234 0
j = i*35% Tax (at 35 percent) 3 82 0 85
k = j/h Tax burden increase factor 10.8
We see that for a U.S. company whose trade balance mirrors the aggregate U.S. ratio of exports to imports, implementation of the border adjustment will increase the company's overall tax burden to a level that is about 11 times higher than under the current regime, if no export loss credit is given.
In this report, we considered a scenario in which foreign trading partners implement regulations with similar provisions to the border adjustment, that is, a BA equilibrium, and in which the "export credit" is unavailable. In BA equilibrium, all major trading partners subsidize exports and penalize imports through border adjustments. As explained, the ratio of U.S. exports to imports has no effect on the overall tax burden of an MNC under BA equilibrium.
Table A-2 presents a summary of our modeling exercise results. The exercise relied on two primary assumptions: (1) consistent with the incentives resulting from the border adjustment, discussed in this report, we assume that the ownership of IP used for cross-border transactions is in the destination market; and (2) in order to isolate the effect on the tax burden of the border adjustment, we assume that tax rates in all jurisdictions are the same.
The results, presented in Table B-1 above, correspond to alternative assumptions for consolidated operating profitability and the magnitude of importers' deductions relative to revenue. As presented, our results indicate that MNC overall tax burden on cross-border activity will increase by between 56 percent and 1,800 percent depending on variations in the two identified inputs. The modeling in support of these results again assumes no export loss credit.
Table B-1. Increase in MNC Tax Burden Resulting
From BA Equilibrium Given Different Cross-Border Profitability
and Importer Value Added Activity
______________________________________________________________________
Consolidated Operating Margin
_____________________________
Importer Deductions as
Percentage of Sales 5% 15% 30% 45%
______________________________________________________________________
5% 1800% 533% 217% 111%
15% 1600% 467% 183% 89%
30% 1300% 367% 133% 56%
Below we present the computations used to derive the results presented in Table B-1. Rather than presenting support for each value, we present our computations for the minimum resulting value, an intermediate resulting value, and the maximum resulting value.
Table B-2. BA Equilibrium Results:
45% Consolidated Operating Margin, Importer Costs 30% of Sales
Table B-3. BA Equilibrium Results:
30% Consolidated Operating Margin, Importer Costs 15% of Sales
Table B-4. BA Equilibrium Results:
5% Consolidated Operating Margin, Importer Costs 5% of Sales
Table C-1. Summary of U.S. Census Bureau Data (millions of USD)
______________________________________________________________________
U.S. Related-Party
Imports Relative
U.S. 'Related-Party' U.S. 'Related-Party' to Exports
Year Imports (a) Exports (b) (c = a/b)
______________________________________________________________________
2009 $740,481 $261,332 $2.8
2010 $922,202 $314,489 $2.9
2011 $1,056,215 $365,014 $2.9
2012 $1,131,902 $381,695 $3.0
2013 $1,122,570 $390,896 $2.9
1 Tax Reform Task Force, "A Better Way: Our Vision for a Confident America" (June 24, 2016) .
2 This summary is prepared based on a review of the GOP blueprint as well as publications from Alan J. Auerbach, an economist whose proposals influenced most of the corporate tax proposals contained in the blueprint. See Auerbach, "A Modern Corporate Tax," Center for American Progress and the Hamilton Project (Dec. 2010); and Auerbach and Douglas Holtz-Eakin, "The Role of Border Adjustments in International Taxation," American Action Forum (Nov. 30, 2016).
3 Note that as part of the blueprint's proposal, cash outflows for capital investments would be immediately deducted as opposed to the current practice of capitalizing and depreciating these investments.
4 Nor would businesses be permitted to deduct interest expense. This means that the DBCFT emphasizes enterprise cash flow, as opposed to shareholder cash flow.
5 Other forms of double taxation -- such as the taxation of income at the corporate level and at the level of dividends paid to shareholders -- involve different issues and are not addressed in this report.
6 The imported good will be resold with a markup regardless of whether the importer is a retailer, distributor, or manufacturer that uses the imported product as a component in its manufacturing process. This is because capital and labor inputs will in all cases be applied to the imported good by the importing company.
7 OECD transfer pricing guidelines.
8 The titles of many of these treaties directly reflect the importance of avoiding double taxation as a motivating factor for their creation. For example, the tax treaty between the United States and the United Kingdom is titled "Convention Between the Government of the United States of America and the Government of the United Kingdom of Great Britain and Northern Ireland for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion With Respect to Taxes on Income and on Capital Gains."
9Available at https://www.irs.gov/businesses/international-businesses/united-states-income-tax-treaties-a-to-z.
10 Fabian Barthel, Matthias Busse, and Eric Neumayer, "The Impact of Double Taxation Treaties on Foreign Direct Investment: Evidence From Large Dyadic Panel Data," 28 Contemp. Econ. Pol'y 366 (2010) (last revised May 2009).
11See, e.g., article 7, para. 2, of the 1996 U.S. model tax convention.
12 The independent effect of the proposed immediate deductibility of capital expenditures may be significant in many cases given the significant timing differences relative to current rules.
13 The ability to use export-related costs as a credit is equivalent to consolidating the three divisions for purposes of calculating taxes due. We note also that the ability to use an "export cost credit" in this manner could have additional implications. For example, companies that are net importers from a U.S. perspective could have material incentives to acquire companies that export more from the U.S. market than they import, potentially leading to large increases in the market values of those "net exporter" companies.
14 This is the internal rate of return on the company's investments given that K is defined as capital expenditures less economic depreciation.
15 Capital stock less economic depreciation.
16 Economic depreciation is the change in the present value of gross returns (operating profit before reinvestment) to capital investments, using the internal rate of return as the discount rate.
17 In our experience, most U.S.-based MNCs fall into the "high-high" combination for imports and the "low-low" combination for exports. That is, they have high weighted average import prices and high import quantities, relative to their export prices and quantities. The reason for this is that at present, most MNCs currently have a large portion of the productive capital (physical and intellectual) that pertains to goods that they sell in the U.S. market located outside the United States. Similarly, they locate a small portion of the productive capital pertaining to foreign markets in the United States.
18 Note that the ability to use an export cost credit in the manner assumed in the prior section could have additional implications. For example, companies that are net importers from a U.S. perspective could have material incentives to acquire companies that export more from the U.S. market than they import, potentially leading to large increases in the market values of those "net exporter" companies.
19For example, see Auerbach and Holtz-Eakin, supra note 2.
20 GOP blueprint, supra note 1, at 27.
21 Regarding imports from controlled affiliates, the value of imports represents the transfer price.
22 As discussed later in this report, the border adjustment may provide clear incentives for the migration of IP rights for products and services traded in cross-border transactions to destination jurisdictions. Cash flow realized on export transactions (which are not taxed under the border adjustment) can be expected to be materially lower in cases in which the importer rather than the exporter owns the IP rights, minimizing the tax benefit realized from the DBCFT treatment of export flows.
23 The blueprint does not explicitly present the on-shoring of foreign manufacturing operations as a benefit of the DBCFT policy. Rather, it states that the policy will eliminate incentives for inversions and will allow companies to "make location decisions based on the economic opportunities, not the tax consequences."
24See, e.g., Joe McDonald, "China Is Preparing to Retaliate Against Donald Trump's Tough Trade Stance, Warns American Chamber of Commerce," Financial Post, Jan. 18, 2017, available at http://business.financialpost.com/news/economy/china-is-preparing-to-retaliate-against-donald-trumps-tough-trade-stance-warns-american-chamber-of-commerce. The Wall Street Journal recently speculated that this retaliation could take the form of similar border adjustment taxes: Richard Rubin, "Trump Tax Idea for Wall Echoes House GOP Plan," The Wall Street Journal, Jan. 26, 2017, available athttp://www.wsj.com/articles/white-house-says-tax-on-mexican-imports-could-pay-for-border-wall-1485463326.
25 This assumes identical tax rates in each jurisdiction. As mentioned, if tax rates differ, the border adjustment with credit would give rise to an effective tax rate equal to the weighted average of the corporate tax rates in each country in which the company has import and domestic operations, weighted by the cash flows earned by these operations (net of any applicable credits for the export division's losses).
26 For example, Kenneth Roboff and Martin Feldstein say that "the extent to which monetary models (or, indeed, any existing structural models of exchange rates) fail to explain even medium-term volatility is difficult to overstate. The out-of-sample forecasting performance of the models is so mediocre that at horizons of one month to two years they fail to outperform a naïve random walk model." Rogoff and Feldstein, "Perspectives on Exchange Rate Volatility," in International Capital Flows 441 (1999). A Goldman Sachs research report also expresses significant skepticism regarding the extent to which exchange rate adjustments will offset trade distortions resulting from the border adjustment. Jan Hatzius et al., "What Would the Transition to Destination-Based Taxation Look Like?" (Dec. 8, 2016), available athttp://www.usfashionindustry.com/pdf_files/Goldman-Sachs-Destination-Tax-Report.pdf.
27 As with other analyses in this study, this analysis examines the impact only of the border adjustment as defined above, as opposed to other components of the blueprint such as the reduction in the statutory tax rate and the expensing of capital expenditures.
28 We assume away intermediate transactions with third parties.
END OF FOOTNOTES