An Appraisal

Most observers regard a cut in the statutory rate of corporation tax as long overdue. Many of the most significant distortions implicit in the U.S. system—including incentives to shift profits outside the US (including by inverting⁵) and toward artificially high leverage—are ameliorated simply by reducing the statutory tax rate (since the value of those various deductions, incentives, and tax avoidance practices will be proportionally less at a lower statutory rate). The lower statutory rate will also lessen the wide dispersion in effective tax rates (across industries and capital structures) that distorts resource allocation toward those activities and financing choices that face a lower tax rate—a point that is taken up further below.

On profit shifting, which has been a particular focus of interest, applying a consensus estimate of its responsiveness to cross-country tax differentials from the meta-analysis in Beer, de Mooij and Li (2018),⁶ the 14-point cut in the statutory rate implies a 21 percent increase in the reported CIT base in the U.S. This almost entirely eliminates the profit shifting estimated there, in itself leading to additional revenue (at the new low rate) of about US$40 billion (though that is much more than offset by the loss from the rate cut itself.)⁷

Discussion has thus centered around not the direction but the extent of the rate change. As seen above, the reduction simply restores the U.S. to the relative position it had in the early 1990s, at around the OECD norm, and indeed somewhat above that when account is taken of subnational corporate taxes⁸). Certainly, the cut in the CIT rate comes at a significant revenue cost (estimated by the Joint Committee on Taxation at US$1.4 billion over the next 10 years). From the perspective of other countries, the reduction in the U.S. rate may be seen as removing an anchor that has supported CIT rates in the rest of the world, with some risk of further aggravating international tax competition. This is a complex issue, with several other features of the reform being relevant, and is taken up in Section VII.

Expensing and Depreciation

Prior to the TCJA, the U.S. maintained a system of bonus depreciation that allowed for a front-loaded write-off of certain types of corporate capital spending. Specifically, firms that purchased property that was depreciable under the Modified Accelerated Cost Recovery System with a recovery period of 20 years or less could take a deduction for 50 percent of the cost of the investment in its first year in service (with standard depreciation schedules applicable to the remaining 50 percent).⁹ Certain investment properties—typically spending to improve the interior of a building—also qualified for this bonus depreciation. This provision applied only to the original or first use of the asset (that is, newly acquired, but previously used, assets did not qualify).

The TCJA replaces this with more generous treatment. Instead of a 50 percent write-off, firms can fully expense—that is, immediately deduct the full cost of—various forms of both new and used tangible property that has a recovery period of under 20 years. This 100 percent write-off applies until end-2022; from 2023–27 there will be a gradual reduction in the share of investment that can be immediately expensed.¹⁰

Treatment of Interest Expenditure

Businesses subject to U.S. tax have generally been allowed to deduct all interest in the taxable year in which it is paid or accrued, within some limits. This creates an inherent tax incentive to the use of debt rather than equity finance, since there is no comparable deduction for the cost of the latter. Such excess leverage can pose financial stability risk when the borrowing is from third parties (sometimes referred to as ‘debt bias’). It can also incentivize the artificial shifting of profits within multinational groups to affiliates in jurisdictions where the tax rate applied to interest received is below that at which it is deducted (‘debt-shifting’).

The U.S. has though long had ‘earnings stripping’ rules that can limit the amount of interest that can be deducted (as a means to prevent the erosion of the corporate tax base through interest payments to non-residents). Specifically, prior to the TCJA, interest deductions to related parties (where no U.S. tax was imposed on the corresponding interest income) are limited to 50 percent of a firm’s adjusted taxable income¹¹ in cases where (i) a corporation maintains a debt-to-equity ratio that is higher than 1.5 and (ii) some part of the interest paid is not subject to U.S. income or withholding tax. Disallowed interest can be carried forward indefinitely.

In an effort to reduce the debt bias and debt shifting implicit in the business tax, the TCJA introduces new and broader limits on the deductibility of interest¹². (For cross-border payments, the general ‘BEAT’ provision discussed later can also create limits on interest deductibility). From 2018–21, interest deductions in any given year are capped at 30 percent of earnings before net interest expenses, tax, depreciation, amortization, depletion and net operating loss. From 2022 onwards, the 30 percent cap applies to a more binding definition of earnings before interest, tax and net operating loss. As under section 163(j), the disallowed interest deductions can be carried forward but there is no relief from these limits either for firms with a debt-equity ratio of under 1.5 or for interest payments that are subsequently subject to U.S. income tax. There are no specific provisions for financial services companies but the new limits do not apply to a range of real estate businesses¹³, farms, and regulated utilities.

An Appraisal

The full expensing of certain forms of capital investments goes a long way in reducing the marginal effective tax rate on new capital projects and moving the U.S. closer to a ‘cashflow’ form of corporate tax.¹⁴ As such, for investments financed by equity (whether retained earnings or new issues) it would tax only rents—any amount, that is, in excess of the minimum required by the investor—and not the normal return to capital arising from those forms of capital spending that qualify for the 100 percent write-off. In so doing, it would in principle eliminate the distortion that is currently inherent in the business tax’s treatment of capital spending. Said another way, the marginal effective tax rate (METR)¹⁵ on equity-financed investment would be zero, and the incentive to invest exactly as it would be in the absence of taxation. This does, however, presume full loss offset, while the TCJA moves in the opposite direction by eliminating loss carry backs and further limiting loss carry forwards. Averaging over all assets, including those for which immediate expensing is not available (but not taking account incomplete loss offset), Gravelle and Marples (2018)¹⁶ estimate that the average METR for equity-financed investment in the U.S. is reduced by the TCJA from around 16 percent to 3 percent.

That said, the announced phasing out of the expensing provision from 2023 will create timing distortions and complexity (for example, giving incentives for firms to front-load investments in order to take advantage of the more generous tax treatment). A less distorting approach would be to simply allow firms to expense all new capital investments (however long- or short-lived) and to make that expensing provision a permanent feature of the U.S. tax code, as a step toward making the business tax a full cashflow tax.

Matters are more complex in relation to debt-financed investments. Combining interest deductibility with generous depreciation allowances amounts effectively to allowing the cost of investment to be deducted more than once, and so leads to a negative METR: a marginal subsidy, that is, to capital formation. This appears to have been the case prior to the TCJA for a broad range of investments. As a result, the TCJA has two countervailing effects for firms unconstrained in their ability to deduct interest, both before and after the TCJA: immediate expensing points to an even more negative METR, but the reduced statutory rate implies a less negative one. The latter effect seems to be the dominant one: Gravelle and Marples (2018) report an increase in the average METR for debt-financed investments, across all assets, from -64 to -48 percent.¹⁷ The result likely hinges on the fact that (i) these types of investments were already getting a front-loaded depreciation and so the incremental benefit from full expensing is perhaps relatively small and (ii) the reduction in the statutory rate was large.

If, on the other hand, the restriction on interest deductibility bites, the situation is much the same as for equity finance,¹⁸ with an METR for plant and machinery of zero. For firms that were not constrained in their interest deductions, pre-TCJA however, and so initially had a negative METR, there will thus be an increase in the METR, albeit in the direction of reducing the distortion to capital investment decisions.

Broadly speaking, METRs thus fall for equity-financed investment and increase for debt-financed ones. Overall, and allowing for some non-deductibility of interest, Gravelle and Marples (2018) find only a modest reduction in the average METR across all types of investment and finance, from 1.7 percent to -3.6. Importantly, however, within this aggregate movement is a general convergence of METRs towards zero. The dispersion of METRs, that is, has fallen, implying an efficiency gain (possibly significant) from the reduced tax-induced dispersion in METRSs and resulting misallocation of capital across different types of investment.

Independent of the impact of METRs, the restriction on interest deductibility (along with the fall in the statutory rate itself) will tend to moderate ‘debt bias.’ This is a positive step, given the financial stability risks from excess leverage (especially in the financial sector).¹⁹ But thin capitalization rules are an imperfect and inevitably crude device for limiting debt—one stark instance being their failure to address these issues for financial institutions).

The new caps on interest deductibility also introduce a procyclical distortion in the system with the cap becoming more/less binding when earnings weaken/strengthen. In a downside scenario, this could exacerbate strains and bankruptcy pressures faced by leveraged corporate entities. This effect is magnified by the TCJA’s restrictions on the offsetting of losses.

A better approach would be simply to remove the underlying tax distortion itself. This could be done in a variety of ways. One is by moving towards an Allowance for Corporate Equity (ACE) form of corporate tax (as recommended for example by Clausing et al, 2016), which would maintain the deduction for interest but add a corresponding deduction for a notional cost of equity finance. Another possibility, toward which the TJCA may have tipped the balance, would be to maintain cash flow treatment of investment while eliminating the deduction for interest spending on newly-contracted debt. This would require transition rules, but does not seem unduly complex. Previously the argument against such a cash flow tax was the potential for losing significant revenues in the near-term that would be made up only over time. With the introduction of the TCJA much of this downside has already occurred with the introduction of temporary expensing for a subset of investments. More challenging, however—or at least much less familiar—would be the extension of cash flow treatment to financial operations.²⁰

An Appraisal

Eliminating the corporate AMT is a positive step. The revenue cost is moderate, and it will lessen costs for firms in complying with (and understanding the consequences of) the tax code as a result of having to run a parallel tax calculation each year (including keeping track over time of credits, the basis of depreciable property, carryovers, credits and operating losses).

An Appraisal

Static costing of the impact of various changes in the individual income tax described above (see JCT 2017) indicate that, in 2019, all income groups will be subject to a lower effective tax rate (Figure 2). However, not only does the absolute dollar reduction in tax payable increase as income rises, so does even the proportional reduction, at least until the top percentile (Figure 3). Various studies corroborate this finding. ²² Indeed approximately 25 percent of the revenue cost of the personal income tax changes accrues to those making over US$500,000 per year (Figure 3) By 2025 (i.e., even prior to the expiration of the various TCJA provisions), 43 percent of households at the lowest end of the income distribution would face the same or higher tax rate (in part a product of the lower rate of indexation of provisions in the code). These results will clearly change, however, once macroeconomic feedback effects are taken into account: (see Lizarazo et al (2017) for a discussion of potential dynamic effects of personal income tax reform in the U.S. in a heterogenous agent model).

View Full Size

Total Effective 2018 Tax Rates Change Due to the Tax Cuts and Jobs Act

(percent)

Citation: IMF Working Papers 2018, 185; 10.5089/9781484372548.001.A001

Source: Tax Policy Center.Note: refers to provisions affecting the individual income tax code.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Tax Cuts and Jobs Act: Effect by Income Category

(static costing; in 2019)

Citation: IMF Working Papers 2018, 185; 10.5089/9781484372548.001.A001

Source: Joint Committee on Taxation

• Download Figure
• Download figure as PowerPoint slide

The main beneficiaries of the tax reform are thus the same households that have seen the largest real income gains over the past 15 years. Since the last major tax reform in the U.S., gains in real incomes have largely accrued to those earning more than 150 percent of the median income.²³ Those earning the median income or less have stagnated. Indeed, the median household real income in 2016 was only 0.6 percent higher than it was in 1999 and, for the lowest 40 percent of the income distribution, real incomes were actually lower in 2016 than in 1999. On the other hand, the top 20 percent of the income distribution saw real incomes growing almost 10 times faster than the median over the same period (and faster still for the top 5 and 1 percent of the income distribution).

View Full Size

Share of Households in Different Income Brackets

(percent)

Citation: IMF Working Papers 2018, 185; 10.5089/9781484372548.001.A001

Source: Alichi, et al., 2016Note: Low income is defined as <50% of median income, high income as great than 150 percent of median income.

• Download Figure
• Download figure as PowerPoint slide

As such, given the changes in the effective rates of personal taxation it is hard to see the TCJA as achieving the administration’s stated objective of implementing a tax cut that is targeted at the middle-class. To achieve that objective, it would be possible to make some changes to concentrate tax relief more on those earning close to or below the median income while providing less for those earning above 150 percent of the median income. Some special regimes, including the carried interest provision, could be eliminated to recover some of the lost revenues and improve the progressivity of the system. Such changes would simplify the system, significantly lower the fiscal costs of the changes to the personal income tax, help mitigate income polarization, and provide a boost that is specifically targeted to those groups that have seen little increase in real incomes over the past 20 years.

To help support the working poor, the child tax credit could be made fully refundable. It could also could be phased out from a much lower household income level (i.e. close to the median income) to make the system more progressive and to lessen the revenue cost of the credit. In addition, to raise labor force participation and support low- and middle-income families, the tax code could help lower income families cover part of the costs for child and dependent care expenses (something that the administration itself had previously proposed).

Steps could be taken to expand eligibility for the Earned Income Tax Credit (EITC), notably to encompass workers without dependents, those under 25, and older workers that are not yet eligible for social security, and to increase its generosity. The revenue costs of such changes are likely to be relatively low²⁴ but they would have a significant impact on those working families that find themselves with incomes that are close to or below the poverty line.

These various changes could be legislated so as to make the personal income tax system both simpler and more progressive. There would then no longer be a need to maintain the AMT as a parallel regime to “patch” the main personal income tax regime. Full elimination of the AMT would somewhat lower compliance costs and further simplify the system.

The replacement of individual exemptions with a higher standard deduction and the elimination of various itemized deductions are positive steps that will simplify the system and result in fewer taxpayers itemizing their deduction. While the caps on the state and local tax deduction and mortgage interest deduction are in the direction of past staff advice they could have been taken further by a more assertive reduction in the mortgage interest cap and a full elimination of the deduction for state and local taxes.

The overall distributional effect of the TCJA will of course also depend on the incidence of changes to business taxation. These—and particularly the degree to which they will ultimately benefit workers—have been a focus of debate and continuing controversy.²⁵ To the extent that the CIT bears on rents, the rate reduction will directly benefit owners (including of course the non-residents who own something like one quarter of all U.S. equites), ²⁶ with any effect on workers coming through general equilibrium effects that may follow from that. More direct effects on workers can follow to the extent that the CIT bears on the use of capital itself.

The argument is clearest for the case of a small open economy, in which the after-tax tax return to capital available on world markets is taken as given: taxing that return cannot reduce capital owners’ income, but must instead must fall on less mobile factors, most obviously labor (and perhaps also land). Several considerations make this a difficult issue to resolve. Empirical evidence on the extent to which the corporate tax has been shifted to labor remains contentious (see Gravelle 2011). For the U.S., with no major rate change over decades, experience to draw on is necessarily scant. Some estimates for Europe put the incidence of corporate taxes on labor at around 30 percent. Since the U.S. is evidently not a small economy, one might expect a larger incidence on capital and hence a smaller benefit to labor from the business tax reform. Also, the extensive holding of U.S. corporations’ equity by nonresidents suggests that a considerable part of the benefit from a lower business taxes enjoyed by non-residents.²⁷ On the other hand, to the extent that the reform moves the system closer to a rent tax, that in itself will alleviate whatever burden is eventually borne by labor. The issue thus remains to a large degree open and likely to be a subject for future research (particularly since evidence on the eventual burden of the corporate tax would be expected to play out over the longer term).

An Appraisal

The TCJA sets the maximum effective rate on pass-throughs that qualify for the deduction (29.6 percent) below both the effective rate on distributed corporate profits (32.85 percent)³⁰ and the top marginal PIT rate (37 percent). The treatment of pass-through entities in the TCJA thus creates incentives to recharacterize C-corporation and personal income as pass-through income. As such, the policies create inequities between businesses that earn the same income but that have a different organizational form (or a different business area), with consequent incentive to change organizational form simply for tax reasons. The policies also create inequities between households that can and cannot arrange their activities and recharacterize their income to qualify for the deduction. In doing so, this creates significant incentives for some high-income employees to become independent contractors (forming partnerships or sole proprietorships, for example) to reduce their tax obligations, potentially eroding both the business and personal income tax bases.

The TCJA does contain guardrails that are intended to limit the use of the pass-through deduction. However, it is unclear how effective these will be in preventing an erosion of the personal income tax base from high income individuals or from stopping pass-throughs from redesigning their operations to maximize their ability to qualify for the 20 percent deduction. In addition, the guardrail provisions greatly increase the complexity of an already complicated set of provisions and may themselves add to the inequities and inefficiencies in the tax treatment of income depending on the nature of the business that derives it (e.g. real estate development versus providing accounting services).

It would be better not to have the 20 percent deduction for certain types of pass-through income. This would have ensured the highest tax rate for pass-throughs remains above that for distributed corporate profits, creating incentives for a broad range of entities to incorporate as C-corporations. In doing so, this would significantly simplify and make more neutral the treatment of business income. Further, such a change would eliminate the incentive for higher income individuals to become pass-throughs. A deduction could be provided solely for pass-through income that is below a certain threshold, to provide a preferential tax regime for truly small businesses and entrepreneurs.

An Appraisal

As with the reduction in the statutory rate, the movement towards territoriality restores the U.S. to what has become the advanced country norm. The U.K. and Japan, most notably, have moved in the same direction in recent years (though China, India, Brazil and other countries that may well become increasingly prominent as capital exporters remain closer to the worldwide model).

In efficiency terms, territoriality has the merit that U.S. firms will compete on equal tax terms in foreign countries with firms based elsewhere (at least, those that are not domiciled in jurisdictions that operate effective worldwide tax systems), including in the acquisition of other companies and productive assets: the criteria of capital import and ownership neutrality respectively (CIN and CON). At the same time, however, inefficiencies arise in that differences in third countries’ tax rates mean that territoriality can distort the choice of U.S. companies as to where to operate abroad: a failure of ‘capital export neutrality’ (CEN).

Theory provides little guidance, however, on which of these dimensions of efficiency is more important.⁴² It may be that neither of these efficiency effects, which operate in opposite directions, will be especially marked, given that the ability to defer U. S taxes is likely to have meant that for many firms the pre-reform system was already, in many respects, a form of de facto territorial system—though other considerations arise below.

With the move to territoriality, some tax on deemed repatriated profits in the transition to the territorial approach is necessary and advisable. However, given that this is a tax on past profits, and thus is non-distortionary, the rate was likely set at too low a level given the urgent need for federal budget revenues. This is particularly so since the previous system of tax deferral has already conveyed large tax benefits to those companies that chose not to repatriate profits. Finally, it seems hard to justify a differential tax treatment of past profits based on whether they are currently being held in cash or non-liquid assets, especially given that taxpayers have 8 years to meet their liability.

Worth noting too is that the evidence suggests that, while for tax purposes this offshore income is viewed as unrepatriated, the resources are actually not being held offshore in an economic sense and, rather, are likely largely invested in a range of (mostly fixed income) U.S. dollar assets. While there will be some portfolio rebalancing as firms readjust their capital structure to the new provisions of the tax code, this is likely to occur over an extended period of time. Also, because the repatriation is deemed, the low tax rate is available to all firms regardless of whether or not they “move” (in an accounting sense) their offshore profits. Based on these factors, it is not expected that the taxation of offshore profits will create a visible effect in international financial markets or lead to significant capital flows back into the U.S., or create measurable movements in currency markets.

By removing the backstop of taxation in the U.S., the move towards territoriality means there is potentially a heightened risk that profits are shifted out of the U.S. into lower tax jurisdictions. For this reason—and because of possible spillover effects discussed later—IMF staff and others⁴³ have argued that such a move would need to be accompanied by some form of minimum taxation of earnings abroad. The GILTI provisions introduce such a tax, though evidently targeted at those companies that derive a large part of the earnings from projects with returns in excess of 10 percent of tangible assets. How to tax such earnings has been a centerpiece of recent debates in international taxation, with a recognition that their inherent mobility has enabled firms with heavy reliance on intangibles—notably the leading ‘digital’ companies, but also for example some pharmaceutical companies and even some blue chips—to pay relatively little tax. The GILTI provisions take a sword to this Gordian knot, ensuring that such activities pay, somewhere in the world, a tax rate of at least 10.5 percent with no deferral.

The core incentive that the GILTI provision creates is towards establishing, in respect of operations outside the U.S., a low ratio of income (in the numerator) to tangible assets (in the denominator). Looking to the numerator, this points towards arranging to receive income in the U.S. rather than abroad, with income from intangibles (such as royalties and similar payments) evidently very much in the minds of those framing the TCJA (particularly given how much of such intangible income has come to be realized in low tax jurisdictions, so as to defer U.S. tax liability).

There is evidently, however, also an effect through the denominator, creating incentives to locate tangible assets outside rather than in the U.S., perhaps by acquiring foreign companies.⁴⁴ Indeed as will become clear below, there can be an incentive to locate tangible assets abroad to the point that the associated investments would be unprofitable in the absence of the tax. To put this another way, GILTI can mean that the METR for investment abroad is negative. This does not sit easily with the objective of improving the competitiveness of U.S. business, at least in respect of investment within the U.S.

There is no doubt that the GILTI provisions, whose effect is supported by the FDII discussed shortly, are an ingenious approach to the problems posed by intangibles in the prior regime Nevertheless, a case can be made that the approach would be even more effective if it were imposed on a jurisdiction-by-jurisdiction basis, rather than on the aggregate of activities abroad: foreign tax credits from higher ta jurisdictions could not then be mixed with more lightly taxed income to increase the available foreign tax credit. Increasing tangible assets in one jurisdiction would then not have the effect of reducing GILTI income arising in others. there would then be still less to gain by receiving GILTI income in higher tax jurisdictions. By the same token, as taken up later, a jurisdiction-by-jurisdiction minimum tax would support stability in the wider international tax system.

One other aspect of the GILTI provisions may prove important. To the extent that, for some firms, its effect may plausibly be to increase the U.S. tax payable on foreign earnings, it may actually increase the incentive to invert. The TCJA does incorporate additional provisions to discourage inversion,⁴⁵ but this may nonetheless in some cases still be a tax-efficient strategy.

The FDII, applicable to domestic corporations, is in important respect the mirror image and support of GILTI, establishing a lower tax rate for income derived from transactions with entities or persons abroad. The underlying logic is that without some more generous treatment of domestic investments that serve foreign markets, the elements of territoriality and the GILTI provision could leave it preferable to produce such items abroad by moving tangible assets offshore (where the intangible assets are located). As such, GILTI is sometimes referred to as the stick discouraging U.S corporations from locating intangibles abroad and FDII as the carrot to encourage their relocation to the U.S. The lower tax rate for exporters will have the effect of distorting location decisions based on tax considerations (which, in some sense, is what is intended).

There is concern, however, that the FDII provision could be subject to a WTO challenge on the grounds that it constitutes a subsidy that is contingent on export performance (normally prohibited under the WTO Agreement on Subsidies and Countervailing Measures). ⁴⁶ It may also run afoul of the minimum standard on harmful tax practices under the G20-OECD project on Base Erosion and Profit Shifting (BEPS), which aims to counter measures that offer advantageous tax treatment for activities that are not tied to real activity in the country concerned. Further, it is unclear, given the costs of transferring intangible assets and potentially countervailing effects from other provisions of the Act, that the preference will achieve its desired goals of incentivizing a repatriation of intangibles back to the U.S. The lower tax rate for exporters creates an economic distortion that arises from a more favorable tax treatment for exports than for domestic sales which will have the effect of distorting location decisions based on tax considerations (which, in some sense, is what is intended). As such, a less distortive system would eliminate this provision.

The third international tax novelty, the BEAT, is aimed directly and broadly at limiting opportunities to shift profits to low tax jurisdictions through inter-corporate payments. In this, it is notably more aggressive than measures coming out of the G20-OECD Base Erosion and Profit Shifting (BEPS) project. It may be that the BEAT will achieve its intended objective of curtailing various base erosion and profit shifting behaviors. There is concern, however, that it may prove punitive for a range of legitimate commercial activities—including in the financial sector—that do not constitute tax avoidance or aggressive tax planning. Crucial here is that the BEAT includes no test that the outgoing payments will be lightly taxed at the other end. The provision also creates a broad-ranging preference for domestic over foreign production (since it applies only to cross-border payments), although its scope is limited to large corporations. ⁴⁷ Further, new incentives are now created for companies to rearrange their operations—or at least their accounting—to avoid application of the BEAT (e.g. by recharacterizing cross-border transactions with related parties as cost of goods sold, or even by inverting). These shortcomings should be mitigated by only applying the provision to those transactions that are designed to transfer profits to low tax jurisdictions (as defined by regulations, or potentially for jurisdictions with statutory tax rates below a certain cut-off).⁴⁸

The GILTI, FDII and BEAT provisions have few if any precedents, and are certainly complex. Tax planners and academics are still in the process of fully understanding their implications. The complexity comes in part from a swathe of detailed provisions that are not set out above, and in some cases have not yet been defined.⁴⁹ Even within the broad parameters described above, however, deciding on the tax-efficient response by U.S. investors is not straightforward—and will depend sensitively on their circumstances. Companies may well, for some time, adopt a wait-and-see approach before they make decisions to change investment locations or restructure their operations (or even the accounting of their operations).

Take, for instance, a U.S. multinational company decision as to whether to locate in the U.S. or abroad a given set of tangible assets to serve foreign markets. Figure 5 shows how the tax-efficient decision depends both on the rate of return earned on those assets and the foreign tax rate⁵⁰. If the rate of return on tangibles is less than 10 percent, then the choice simply depends on whether the foreign tax rate exceeds 21 percent (in which case, given the element of territoriality, investing in the U.S. is preferred). Suppose, however, the return on tangible assets is higher than 10 percent, opening up the possibility that GILTI (if the investment is placed abroad) or FDII (if placed at home) will apply.

Specifically, the choice is then between (i) locating at home and paying 21 percent on that part of the return under 10 percent plus the FDII rate of 13.125 percent on the excess, or (ii) locating abroad and paying GILTI on the eligible income plus the foreign tax liability. If the rate of return is only slightly above 10 percent, then how that excess income is treated immaterial: If the foreign tax rate is not very much below 21 percent, investing in the U.S. is preferred. If, on the other hand, the excess return is extremely high, then in effect it is only the treatment of that excess income that matters. Since the firm can always achieve a 13.125 percent tax on that income by locating in the U.S. and being taxed under FDII, the U.S. would be the preferred location unless the foreign tax rate is only very slightly above that. Hence the bowed part of the boundary between the two choices shown in the figure.

Further complexities arise when considering not where to locate a given investment, but how much to invest there. ⁵¹ Take, for example, a firm that is benefiting from FDII today and expects to do so in the future. Suppose it is producing only for export. Then immediate expensing will generate a deduction at the FDII rate of 13.125 percent, and future cash flows will be taxed at the same rate. So it might seem that the tax will act effectively as a rent tax on that investment and that the marginal effective tax rate (METR) will be zero. However, investing gives rise to another effect: by increasing the stock of tangible capital to which the benchmark return of 10 percent is applied it automatically reduces the amount of income subject to the reduced FDII rate and increases the amount that is subject to the standard rate of 21 percent. This effect increases the tax liability associated with the investment and so implies an overall discouragement to investment (i.e. the METR is positive). But this is not the only possibility. As shown in Appendix I, the sign of the METR depends on both the regime in which the firm finds itself when investing and that in which it receives the associated income. Table 2 summarizes the results. Though there are exceptions, the overall impression is that FDII actually creates a disincentive to create tangible assets in the U.S., since a higher stock of such assets mechanically reduces the amount of FDII income to which the favorable treatment applies.

Table 2.

The METR for Domestic Investment under FDII

Note: Details in Appendix I.

Table 2.

The METR for Domestic Investment under FDII

		Period 2
		FDII	Standard regime
Period 1	FDII	+	+
	Standard regime	−	0

Note: Details in Appendix I.

View Table

Table 2.

The METR for Domestic Investment under FDII

		Period 2
		FDII	Standard regime
Period 1	FDII	+	+
	Standard regime	−	0

Note: Details in Appendix I.

For tangible investments abroad, the position is essentially reversed under GILTI (considering here only U.S. taxes⁵²), because it is a higher rate (not a lower rate) that is applied to the income in excess of a 10 percent return on tangible capital. A firm that is in the GILTI regime now and expects to remain there, for instance, gains an additional advantage by locating tangible assets abroad—beyond any income they generate—since by doing so it increases the amount of income that is excluded from taxation in the U.S. under the GILTI. Hence, as noted, the METR on investing in tangible capital abroad in such a case is negative.

And even all this does no more than touch on the complexities at stake.⁵³ In considering an investment decision within the U.S., a firm may, for example, conclude that the tax treatment (e.g. under FDII) would imply a much lower tax rate than it pays abroad. However, by locating in the U.S. that same investment would now become, in principle, subject to the BEAT on any transactions that subsidiary may have with offshore related parties. For example, take a firm that locates all its technology or human resource services in a subsidiary outside of the U.S. and then charges the multinational’s various corporate entities an internal payment to remunerate for providing those services. It may well become less advantageous to invest in the U.S. because doing so would give rise to BEAT-related tax liabilities on those internal corporate payments. But it may be even better strategy for that company to expand its activities in the U.S. and so generate enough deductible expenses not subject to the BEAT that it ceases to be liable to the BEAT. Again, the privately optimal response will be very company-specific.

Though the picture is complex, it may well be that the international provisions of the TCJA create incentives to locate more intangible assets in the U.S, and more tangible investments outside of the U.S., than before the TCJA. Though there are evidently other important aspects of the reform that encourage real investment in the U.S., this potential feature of the international provisions does not accord well with the stated objective of the reform to encourage investment and job creation within the U.S.

Macroeconomic Spillovers

The most direct and immediate impact of the U.S. tax reform on other countries is through the spur it provides to aggregate activity in the U.S (as analyzed above). This demand stimulus will not only add to U.S. growth but also create a faster normalization of U.S. monetary policy and a rise in both the US dollar and in interest rates. The net effect of this change is likely to provide a near-term boost to the U.S. and to translate into higher import growth, an increase in the U.S. current account deficit (to around 3½ percent of GDP by 2019–20), putting upward pressure on the dollar and worsening the international investment position. The higher U.S. current account deficit is expected to be matched by growing current account surpluses (or smaller current account deficits) in other systemic economies with the largest current account effect likely to be seen in Canada and Mexico. As a result, global imbalances are expected to rise, with the various attendant risks that such imbalances convey (including possibly catalyzing public support for increased protectionism).

However, outside of this baseline effect, the range and size of future risks, both for the U.S. and for the global economy, is likely to rise. The expectation is that the demand stimulus would lead the Federal Reserve to raise policy rates at a faster pace which could potentially create volatility and disruptions in U.S. asset markets. There could be a tightening of U.S. and global financial conditions, an abrupt decompression of term and other risk premia, and strains placed on leveraged corporates and households. Such a scenario could also create important adverse risks for non-U.S. corporates, households, and sovereigns (especially those that have borrowed heavily in U.S. dollars) and/or trigger a marked reversal of capital flows, particularly to emerging markets.

Tax-related International Spillovers⁵⁶

These baseline macroeconomic effects do not, however, capture the potential impacts of the reform on either investment abroad or the extent of profit shifting between the U.S. and the rest of the world. In practice, these effects are a major consideration for other countries as they consider the impact of the reform, and whether to respond to it through changes in their own tax system (and if so how). There is also the potential that policy changes abroad that are triggered by the TCJA could, in turn, have spillback effects for the U.S. itself. Indeed, the U.S is so significant to the international tax system, and so fundamental are the change in policies, that an ultimate reshaping of that system may prove one of the most lasting consequences of the reform. The complexity of the TCJA means, however, that quite how the reform will play out in the wider global context is far from easy to assess. Consider first the impact on real investment. As seen above, the TCJA likely lowers the METR for many investments in the U.S., so tending to increase investment there—including by foreign corporations (which can also benefit from the FDII provisions). That does not, however, necessarily mean a reduction in investment elsewhere. That may be the case, however, if companies are constrained in the overall amount of capital they can deploy worldwide or if they are serving integrated markets (within North America, for instance) from different locations. In such circumstances, companies would effectively have to allocate across locations, and the statutory rate at which profits are taxed becomes relevant. For this, the reduction in the average effective tax rate—which is more marked than the change in the METR—comes into play and will tend to reduce real investment abroad. Beer, Klemm and Matheson (2018) estimate a fairly wide range of possible real investment reductions: assuming no change in tax rates outside the U.S., their simulations imply a median reduction in multinational investment of between 1.1 and 2.7 percent of their capital stock,⁵⁷ and of 5 to 20 percent in the most affected percentile. Corresponding to this, median tax revenue from multinationals is estimated to fall by 0.6 to 1.3 percent (and 3 to 21 percent in the top percentile).

Three aspects of the reform may have countervailing effects on investment and revenue abroad:

• First, to the extent that the taxation of repatriated earnings had an impact under the prior regime, the move toward territoriality reduces the U.S. tax charge on foreign profits, and so (all else being equal) would tend to increase, U.S. corporations’ investments abroad. There is indeed evidence that the movement to territoriality in the U.K. led to an increase in outward investment to lower tax jurisdictions (see Liu, 2018).

• Second, as also noted above, the GILTI provisions which accompany the shift towards territoriality create a new incentive to locate real investments outside the U.S. The extent of these effects being very uncertain, however, it seems a reasonable presumption that the net effect will be reduction in real investment abroad, more marked in countries with closer relations to the U.S.

• Third, and acting to offset the first aspect, insofar as locating investments inside of the U.S. could be expected to give rise to taxation under the BEAT, this could make it less advantageous to invest in the U.S. rather than abroad.

The large reduction in the statutory tax rate in the TCJA can also be expected to reduce profit shifting out of the U.S., as discussed earlier, a counterpart to which would be a reduction in tax revenue accruing in other countries. Beer, Klemm and Matheson (2018) put the consequent median revenue loss outside the U.S., including also that from the reduction in investment discussed earlier (all still at unchanged tax rates outside the U.S.), at 1.3 to 2.8 percent of revenue from multinationals.⁵⁸ These effects may be exacerbated by both the BEAT measures (to the extent that it reduces taxable of foreign affiliates in the U.S. as a result of reduced cross-border payments to them.) and the effect of the GILTI insofar as it undercuts the tax advantage to U.S. corporations in relocating their income to low tax jurisdictions.

Second Round Effects and Strategic Responses

All this assumes unchanged policies outside the U.S. In practice, many countries are actively considering how the reform may affect them, and how, if at all, to respond.

Much attention has focused on the possibility that the TCJA will intensify the trend towards lower corporate tax rates that has been underway for many years—with likely consequent adverse effects on tax revenue, but perhaps also the effect of moderating any reduction in real investment. Assessing this raises several considerations.

One of these is the nature of the strategic interaction between the U.S. and the rest of the world. Suppose that one simply thinks of interactions in the setting of tax rates as a one-shot game, ignoring the complications that repeated interactions bring, and that there is only one other country, which behaves in Nash fashion, simply taking as given whatever tax rate is set by the U.S. One possibility then is that the U.S. acts as a Stackelberg leader, internalizing in its decision the likely response of others. Another is that the U.S. also plays Nash. Figure 16⁵⁹ contrasts the outcomes in these two cases. The initial equilibrium is at point A, with the U.S. setting a rate of 35 percent and the other country choosing the corresponding tax rate on its best response function.

View Full Size

Strategic Interactions in Tax-Setting

Citation: IMF Working Papers 2018, 185; 10.5089/9781484372548.001.A001

• Download Figure
• Download figure as PowerPoint slide

Take first the case in which the U.S. acts as a Stackelberg leader, choosing its own tax rate to pick off its preferred point on the foreign best response. The interpretation of the rate reduction under the TCJA would then be that. perhaps because of a change in policy preferences,⁶⁰ the preferred position of the U.S. is now at point B, where an iso-welfare locus is tangential to the foreign best response. In reaction to the U.S. cut, the foreign country will reduce its tax rate to the level at point B—and that will be the end of things.

Things are different, however, if the U.S. is playing Nash, so that the initial equilibrium is such that its best response passes though A. The rate cut of the TCJA can then be thought of as a downward shift of the U.S. best response as to take it through point C, implying, at the initial foreign tax rate, a U.S. rate of 21 percent. The new equilibrium is then found at point D, where the two best responses intersect.

Comparing the two cases, point D being south-west of B, the final reduction in tax rates—in both the U.S. and the rest of the world—is less if the U.S. is a Stackelberg leader than if it too plays Nash: the intensification of tax competition set in train by the U.S. tax cut, that is, is consequently less. Which view of the strategic game comes closer to capturing reality is by no means clear (and of course both are highly simplistic). It may be, for instance, that the apparent difficulty of achieving substantial tax change in the U.S. means that a leadership role is inevitable; and wise policy makers, recognizing that, will think through how others will respond to their decisions.⁶¹ Or it may be that substantive reform now being in place, simple changes in the rate will now be much easier.

The difference between the outcomes in these two strategic settings is likely to be greater, all else equal, the more responsive are tax rates set by each country to those set by others: the flatter, that is, is the foreign best response in Figure 16.

Taking a central estimate of this parameter from the (fairly sparse) literature, Beer, Klemm and Matheson (2018)⁶² estimate that in the Stackelberg case (or, more generally, when further changes in the U.S. rate are ruled out) rates in the rest of the world will fall by an average of 3.8 percentage points. They further estimate that (notwithstanding the dampening of investment effects this implies) the additional reduction in median revenue from multinationals will be in the order of 3 to 7 percent. This is larger than the effect at unchanged tax rates: policy responses elsewhere can thus be far from second-order in assessing the final effects of the TCJA.

Suppose, instead, that the U.S. is playing Nash. Within the same setting as Beer, Klemm and Matheson (2018), ⁶³ rates abroad fall instead by 4.6 points, while that in the U.S. falls not by the 14 points of the TCJA but by 16.8. The difference is thus modest. At levels of responsiveness that are higher but within the range of existing estimates, however, the difference is very marked. Such calculations can be no more than illustrative, but suggest, nonetheless, that doubts over the form of strategic interactions in tax-setting are a significant source of uncertainty in assessing the impact of the TCJA on competition in tax rates

All this, however, focuses solely on the statutory tax rate, and does not capture other central aspects of the TCJA that will likely shape the reactions of other countries.

One is the move towards territoriality. To some extent, the final total tax paid by U.S. multinational on their earnings abroad was independent of the tax rate charged abroad: foreign countries could tax at up to the U.S. statutory rate without any impact on the total tax paid by investors, but with the effect of increasing their own revenue at the expense of revenues to the U.S. Treasury. Historically, this has likely been an important way in which the U.S. policy has served to sustain higher statutory tax rates in the rest of the world. The widespread use of deferral, however, significantly blunted this effect, so this twist to downward rate competition may be limited.

The novel international provisions of the TCJA, however, in some important respects point in the opposite direction. To see this, consider again Figure 5, considered this time from the perspective of a foreign government, whose sole concern—for simplicity—is to attract U.S. multinational activity (the location of which is assumed, again for simplicity, to be driven solely by tax considerations). Setting a tax rate above 21 percent does not look like a wise strategy since U.S. corporations will then prefer to produce in the U.S. Those U.S. companies that are producing in other jurisdictions for export to third markets would take advantage of FDII if their return on tangibles is above 10 percent and relocate to U.S. as a base for exports.

More strikingly, supposing for the moment that there is only one foreign country, there is also little point in setting a tax rate below 13.125 percent, since that is the best that is available to an export-oriented U.S. corporation if it were to choose to locate in the U.S. A low tax jurisdiction that aimed to attract intangible income by setting low tax rates will now have an incentive to actually raise its tax rate.

The expectation might thus be, in broad terms, of a foreign tax rate being set between 13.125 and 21 percent.⁶⁴ Matters become more complex, however, when recognizing that there are many foreign jurisdictions. If one were to set its rate at 13.125, for example, another could by setting its own rate to (say) zero ensure that the U.S. corporation faced a total rate of 10.5 percent, rather than the 13.125 rate it would face in the other jurisdiction or if it opted for FDII. Another way to see the consequence of averaging is in the implication that a jurisdiction may be able to lower its statutory rate below 13.125 percent without triggering any GILTI liability for the U.S. multinational (particularly if there are significant investments taking place in jurisdictions with tax rates that are quite a bit higher). Clearly, however, that reduction in the tax payable by the U.S. corporation is fairly modest.

The pressures towards higher tax rates in low tax jurisdictions are further mitigated, of course, to the extent that they seek also to attract activity from higher-tax jurisdictions other than the U.S. Worth noting, nonetheless, is that any upward pressure on rates in low tax jurisdictions could have a significant spillover benefit to other countries that are also affected by profit shifting. That, in turn, once the dynamics fully play out, would limit the downward pressure.

The key point, however, is that the structural reforms introduced by the TCJA are likely have implications for the wider international tax regime that are profound, and subtler than a simple intensification of rate competition. These are, as yet, far from being perfectly understood, and doubtless much innovative thinking is going on. Countries might conceivably, for instance, choose to adopt schemes that vary the tax rate with the return on tangible investments, undoing the manner in which GILTI and FDII favor the locating of intangibles in the U.S.

Importantly too, countries have many other instruments, beyond the statutory rate, by which they might choose to respond. Broadly, they face two challenges: potentially increased outward profit shifting, and a potential disincentive to real investment by U.S. corporations. The extent of these will of course vary across countries. Many low tax jurisdictions, for instance, may not plausibly host significant real investments and their business model is based on attracting intangibles. For many of the more advanced economies, however, both considerations may plausibly apply since they have a large domestic market as well as potentially offer opportunities for reducing the tax burden on intangible assets.

More widely, the TCJA feeds into, and to a large degree may reshape, the wider continuing debate as to the future of the international tax system. The GILTI provision, for example, is a creative approach to the wider difficulties of taxing the income associated with intangibles. Other countries which see themselves as suffering from similar difficulties of a mobile tax base that is linked to intangible assets may look to creating their own form of GILTI. Similarly, countries may respond with a modified version of the BEAT if that gets traction as a widely acceptable response to base erosion and profit shifting.

The international provisions of the TCJA do not, however, resolve the fundamental problems of the current international system, which arise from the inherent and increasing difficulties associated with the implementation arms-length pricing.⁶⁵ The wider debate on the implications of digitalization, linked to broader arguments for asserting a power to tax business income in ‘market countries’, meaning those whose residents are the purchasers of goods and services sold—calls for some form of ‘destination based’ taxation. The TCJA recasts the debate, but does not close it.⁶⁶