The new section 951A rules that subject large chunks of active controlled foreign corporation income to immediate U.S. tax and the section 250 deduction that can cut that tax in half are like an automobile that came off the assembly line too fast. Regulation writers at Treasury and the IRS will do the best they can within their authority to tighten loose parts and add missing pieces. But to get this machine running smoothly will require substantive new legislation that goes well beyond anybody’s definition of technical corrections. That is the picture that emerges from the New York State Bar Association Tax Section’s 112-page report on the global intangible low-taxed income provisions. (For a simplified overview of the GILTI provisions, see the appendix to this article.)
First, let’s address some of the questions asked about GILTI. Yes, it seems certain that the regulations will clarify that the section 78 gross-up of taxes associated with GILTI will be included in the GILTI basket. Yes, it seems certain that expenses will be allocated and apportioned to the GILTI basket (although it is far from clear how that will be done). And yes, in line with the convincing arguments on policy and about Treasury authority made by the NYSBA report, regulations should allow affiliated groups that file consolidated returns to be treated as a single shareholder for purposes of GILTI.
Why so convincing? The main reason is that if you follow the statutory language literally and only let a single corporation within a consolidated group to be a shareholder, the hurdles that the group must clear to minimize its taxes would be unending. Moreover, as will often be the case with many international provisions of the new law, taxpayers unschooled in the intricacies of those provisions can be arbitrarily subjected to higher taxes without any good policy reason.
The haphazard nature of the GILTI provisions stems largely from the fact that CFCs with losses are treated differently (less favorably) than CFCs with positive income, and as a result you can get starkly different results by packing and cracking different groups of CFCs. Specifically, GILTI is not reduced by the routine return equal to 10 percent of qualified business asset investment (QBAI) — that is, depreciable, tangible property — of CFCs with losses. Moreover, any foreign income taxes paid by loss CFCs are not creditable. If a single corporation within an affiliated group can be a shareholder, corporations within an affiliated group can attain different groupings of CFCs by shifting ownership of those CFCs to different corporations within the group.
If consolidation of ownership is not allowed, the proverbial traps for the unwary would be rampant. For example, if M1 domestic corporation (part of an affiliated group) had no taxable profits and happened to be the owner of Z, a profitable CFC, it would not get a 50 percent GILTI deduction (because of the section 250 taxable income limitation). Meanwhile, an otherwise identical group could get the full section 250 deduction if ownership of Z was assigned to profitable domestic corporate shareholder and group member M2.
If consolidation of ownership is not allowed, an affiliated group could avoid expense allocations that limit foreign tax credits of a CFC chain by having a member with no allocable expenses own all its CFCs. Also, if consolidation of ownership is not allowed, the amount of QBAI (that reduces the GILTI inclusion) could be increased if a corporate shareholder in the group owns both loss CFCs and profitable CFCs. Alternatively, contrary to the interests of the taxpayer, QBAI could be reduced if ownership of loss CFCs is separated from ownership of profitable CFCs.
There is no tenable policy reason for treating similar groups differently because of arbitrary differences in legal structure. And there would be new administrative problems for the IRS, which would have to determine the taxable income of each member of the group. Compliance costs for taxpayers would unnecessarily rise because they would have to determine and defend the allocation of income across members of the affiliated group, and they would have to annually anticipate CFC earnings to avoid being caught with an unfavorable ownership structure. According to the NYSBA report, Treasury since 2004 has had regulatory authority under section 1502 to treat members of a group as a single taxpayer.
We Can Do Better
Here is the first reason the GILTI provisions need a legislative fix. For many of the same reasons that shareholders of an affiliated group receiving income should be aggregated, so should the calculation of income from the CFCs that are generating the income. It is true that for each shareholder, the aggregated income from profitable CFCs is subtracted from the aggregated losses of loss CFCs to determine net tested income. From that net amount is subtracted an amount equal to the aggregated difference of routine return over interest of all profitable CFCs. (The treatment of interest here is uncertain, as discussed below.) So there is a sort of quasi aggregation. However, the aggregation is incomplete because of the unfavorable treatment of loss CFCs: Their FTCs and routine returns are excluded from the GILTI calculations. Thus, one of the more obvious first steps in tax planning under the new regime is to consolidate loss CFCs with profitable CFCs. As the NYSBA report points out, this restructuring would occur solely for tax purposes with no policy justification. (Prior analysis: Tax Notes, Feb. 12, 2018, p. 845.)
Although regulations can allow affiliated groups that file consolidated returns to be aggregated as a single shareholder to avoid economically pointless tax planning, they cannot fix the problems of foreign corporations when the components that feed into the shareholder’s GILTI inclusion are not adequately aggregated. It is clear from the joint explanatory statement (at 518 n. 1538) and from the statute (section 951A(d)(2)(A)) that only tangible depreciable property used in the production of “tested income” is QBAI, and that tested income is positive income (excluding losses). Further, the statutory language clearly states that deemed paid tax credits are only available for (positive) tested income (section 960(d)(3)). Therefore, the NYSBA tax section recommends that legislation be enacted to treat all related corporations owned by a single shareholder as a single CFC for purposes of the GILTI calculations.
Here is the second reason for a legislative fix. For the purpose of calculating the FTC limitation, GILTI has its own new separate basket and — as has been a subject of widespread criticism — the new basket has no carryforwards or carrybacks. This means that otherwise identical taxpayers — the first with $1,000 of GILTI in year 1 and $500 of loss in year 2, the second with $250 of GILTI in both years — will be treated differently even though they have the same income ($500) over a two-year period. Not only is there no justification for this, but it violates the core principles of neutrality and fairness. Moreover, it produces arbitrary results and invites economically pointless tax planning.
The NYSBA report points out that a fix might be attainable by regulation because the statute states that the computation of nested income and tested loss includes deductions “properly allocable” to gross income and that losses from prior years could be considered properly allocable deductions. That’s a stretch in any case, especially given the legislative history: “The provision creates a separate foreign tax credit basket for GILTI, with no carryforward or carryback available for excess credits” (joint explanatory statement, at 518). After careful discussion of the pros and cons of various methods, the report recommends that GILTI losses be carried over, either by regulation or by legislation, at the shareholder level to offset future GILTI gains under rules similar to those for domestic net operating losses.
Here is the third reason for substantive legislative change. As is well known, noncorporate CFC shareholders receive particularly harsh treatment under the new law. They can pay tax on their GILTI at rates as high as 37 percent, and to add insult to injury they do not qualify for a section 250 deduction. Self-help may be available to these taxpayers. Section 962 allows individuals to elect to be taxed at the corporate rate on gross income of a CFC with no allowance for deductions, which includes the new section 250 deduction. The NYSBA report recommends that Treasury issue regulations making the section 250 deduction available to individuals who elect section 962 treatment. The tax section concedes that Treasury’s authority may be questionable, in which case the change should be made by statute.
Here is the fourth reason for cleanup legislation. Partnerships that own CFC stock pose particularly thorny issues under the GILTI provisions. The NYSBA report argues that the joint explanatory statement and the statute are ambiguous about whether the GILTI calculation should be made at the partner or partnership level. In this author’s opinion, that may be debatable for several reasons, including that the core statutory language (section 951A(a)) states that a U.S. shareholder must include GILTI in gross income. Given that the partnership is the shareholder, it would seem to follow that the calculation should take place at the partnership level. Despite statutory language to the contrary, the report makes strong policy arguments for the partner-level approach.
The partnership-level approach effectively makes aggregation elective and encourages tax planning that serves no policy purpose. Under the NYSBA tax section’s recommended alternative, the components of the GILTI calculation — tested income, tested loss, QBAI, and interest expense — flow from each CFC through the partnership and are used to calculate the GILTI inclusion in the partner’s taxable income and the partner’s section 250 deduction (which is available only for corporate partners and may be limited if the partner’s taxable income is insufficient). To implement this method, special rules would have to be developed to conform with new section 163(j) interest limits, which the statute explicitly states are calculated at the partnership level.
The first three of these possible legislative actions would entail revenue loss that would be scored by the Joint Committee on Taxation before enactment. If Congress is inclined to maintain revenue neutrality, it would be easy to obtain offsets within the GILTI framework. For example, the current arbitrary 10 percent rate applied to QBAI could be raised or the section 250 deduction currently set at 50 percent for GILTI can be reduced. It is already scheduled to decline to 37.5 percent beginning in 2026.
But Wait, There’s More
The principal authors of the NYSBA report are Kara Mungovan and Michael Schler, both at Cravath, Swaine & Moore LLP. They have left few stones unturned when it comes to the GILTI provision. The following paragraphs review some of their other recommendations, most of which can or probably could be addressed by regulation.
Beyond the regulatory relief already mentioned, one of the more significant regulatory actions the NYSBA tax section believes Treasury could make regarding the GILTI provisions involves the treatment of interest expense of a CFC with a tested loss. Recall that the basic calculation of GILTI is:
Sum of shareholder’s tested income – Sum of shareholder’s tested losses – Sum of shareholder’s amounts equal to (10% of QBAI - interest)
We know the QBAI of CFCs with tested losses is excluded from the GILTI calculation, but the statute seems to say the corresponding interest expense of loss CFCs would still be added to GILTI. This would create some aberrant results. For example, if a CFC had a small tested loss and a large amount of interest, all that interest expense (net of the small loss) would be a GILTI inclusion. The NYSBA report suggests that Treasury clarify (in the absence of a legislative proposal that would treat all CFCs of a shareholder as a single CFC) that interest expense of loss CFCs is excluded from the GILTI calculation.
The NYSBA report makes several recommendations, all of which seem to be clearly within Treasury’s regulatory authority regarding the all-important issue of expense allocation. Its findings are premised on the idea that all expenses should be allocated to some income and that expenses incurred in the generation of nontaxable categories of income should be allocated to other taxable income categories. And although the NYSBA tax section proposes that expenses be allocated to GILTI, it also takes the position that because the implications of excess credits are so severe for the GILTI basket (due to the absence of carryforwards and carrybacks), regulation writers should lean in the direction of minimizing expense allocations to the GILTI basket. So, for example, rules for allocation of research expenses should be rewritten to move the allocation away from the GILTI basket.
The NYSBA report points out that there is a loophole that in some cases can permanently exempt GILTI and subpart F income from tax in a year when CFC stock is sold. It takes no position on whether this may be addressed by regulation or legislation.
The NYSBA tax section points out that Treasury needs to specify what method of calculating income should be used for GILTI. For example, a nondeductible fine or some interest expense that exceeded the section 163(j) limitation would not be deductible under a taxable income method but would be deductible under the subpart F method (based largely on generally accepted accounting principles). The report recommends that either a variant of taxable income or the subpart F income method be adopted in regulations.
As if all this weren’t enough for the beleaguered regulation writers at Treasury, the NYSBA tax section assigns them more than a dozen additional tasks to clean up, clarify, and make choices on behalf of Congress. Surprisingly, among all the numerous recommendations, the report only mentions one pure technical correction, suggesting that the importance of technical corrections (at least regarding the GILTI provisions) is trivial compared with the relief and clarification provided by regulation or by substantive legislation.
It will be a minor miracle if Treasury and the IRS are able to issue proposed GILTI regulations anytime before Thanksgiving. For now, there is little doubt that they will closely consider much of the carefully reasoned analysis in the NYSBA report. The report is as good a guide as any for those who need to speculate on what Treasury can and might do in future regulations. As for Congress, it is not mere technical corrections that are needed, but a substantive revision of extremely complex rules that were drafted in relative haste. Unfortunately, the prospects for this type of congressional action are slim.
Appendix — Summary of GILTI Rules
Shareholders of CFCs must add a GILTI inclusion to their taxable income (new section 951A). They may be entitled to a deduction equal to a percentage of that inclusion (new section 250) and they may reduce their U.S. tax with credits, subject to a new GILTI separate basket limitation, generally equal to 80 percent of related foreign taxes (new section 960(d)).
The GILTI inclusion is the sum of net tested income minus net deemed tangible income return (NDTIR). Net tested income is the sum of the taxpayer’s allocable share of all tested income of CFCs in which the taxpayer is a shareholder with positive income less the taxpayer’s allocable share of all tested losses of the CFCs in which the taxpayer is a shareholder. Net tested income cannot be negative. Thus, the GILTI inclusion is computed on a shareholder-by-shareholder basis, and it also cannot be negative.
Tested income of any individual CFC is its gross income less properly allocable expenses and less five categories of income: subpart F income, dividends from another CFC owned by the taxpayer, income eligible for the high-tax exception (which does not include high-taxed income that is not subpart F income), foreign oil and gas extraction income, and income effectively connected with the conduct of a U.S. trade or business. NDTIR is calculated at the CFC level and then aggregated. The NDTIR of any CFC is 10 percent of QBAI minus interest expense paid outside the shareholder’s CFC chain. QBAI is the adjusted tax basis of depreciable tangible assets used in the production of net tested income. (So not all CFC tangible depreciable assets are necessarily QBAI.) If the CFC has tested loss, QBAI is set to zero. Also included in income is a section 78 gross-up of taxes allocable to the net tested return minus NDTIR.
To potentially reduce the taxpayer’s effective tax rate on GILTI to half of the 21 percent corporate rate, the taxpayer is allowed a deduction equal to 50 percent of the GILTI inclusion (which includes the section 78 gross-up) plus 37.5 percent of foreign-derived intangible income (FDII). (The percentages are reduced after 2025.) The deduction may not exceed the taxable income of the taxpayer/shareholder. Importantly, the deduction is only available to corporate shareholders.
FTCs are available for 80 percent of all foreign taxes allocable to (positive) tested income (not reduced by NDTIR). Foreign taxes attributable to other CFC income (for example, subpart F income) are not included in the calculation.
Finally, with the intent of focusing the burden of GILTI provisions on low-taxed foreign income, there is an FTC limit equal to the shareholder’s tax rate (21 percent for corporations) times 80 percent of the foreign income taxes attributable to the GILTI inclusion. For purposes of the FTC limit, the inclusion is adjusted downward (regardless of the foreign tax rate) dollar-for-dollar for any of the U.S. shareholder’s expenses (such as interest, research, or administrative cost) allocated or apportioned to GILTI. This is a new separate FTC limitation basket with no carryforwards or carrybacks.