Reuven S. Avi-Yonah is the Irwin I. Cohn Professor of Law at the University of Michigan Law School. He thanks Kimberly Clausing, Robert Goulder, Alexandra Klass, Michelle Layser, and Marty Sullivan for helpful comments.
In this installment of Reflections With Reuven Avi-Yonah, Avi-Yonah examines the use of tax incentives to encourage location-specific investment and offers a new approach that he says will be more successful.
Both the global intangible low-taxed income provisions of the Tax Cuts and Jobs Act and the substance-based income exclusion of pillar 2 provide incentives to shift real investment to low-tax locations. Under GILTI, a 10 percent return on tangible assets of the subsidiaries of U.S. multinationals (qualified business asset investment) is exempt from tax. Under pillar 2, the top-up tax calculation excludes 10 percent of payroll costs and 8 percent of the carrying value of tangible assets in a jurisdiction.1
How significant are these incentives? While economists have been critical of both QBAI and the substance-based income exclusion,2 there is little evidence that they have persuaded multinationals to move real investment to low-tax jurisdictions. Part of the reason is that many modern multinationals depend much more on intangibles than on real investment. Another is that because the affected tax rates are low (10.5 percent for GILTI and 15 percent for pillar 2), the size of the exclusion is also low.3
But many multinationals (for example, pharmaceutical and chip manufacturers) do invest in tangible facilities, and rates can increase over time. So the more basic question is, how effective are location-based tax incentives in shifting investment?
Traditionally, economists have been critical of location incentives.4 But more recent literature has been more positive.5 A Brookings Institution paper on the tax incentives of the CHIPS and Science Act and the Inflation Reduction Act concludes that:
Spurred in part by three significant pieces of federal legislation, since 2021, the United States has experienced an investment surge in “strategic sectors,” defined as clean energy, semiconductors and electronics, biomanufacturing, and other advanced industries. So far, economically distressed counties are receiving a larger-than-proportional share of that investment surge relative to their current share of the economy. With comparatively low prime-age employment rates and median household incomes, these counties account for about 8 percent of national GDP but have received 16 percent of announced strategic sector investments since 2021. Strategic sector investments are much more likely than private investment overall to target economically distressed counties, relative to recent years and the 2010-2020 recovery period — suggesting a significant departure from geographic patterns of prior investment. Distressed counties that have received a strategic sector investment currently have relatively high shares of employment in advanced industries — suggesting that such foundations continue to matter to private investors. Smaller regions (defined as “micropolitan areas”) account for about 25 percent of the nation’s employment-distressed population, but have secured 50 percent of all strategic sector investments going to distressed counties since 2021.6
The reason for these outcomes is that both the CHIPS and Science Act and the IRA include incentives to invest in distressed communities. For example, IRA tax credits include bonuses of at least 10 percent when the investment is in a low-income or “energy community.”7 The Department of Energy’s new Energy Infrastructure Reinvestment Financing program particularly benefits energy communities with loan guarantees to projects that repurpose or replace energy infrastructure that has ceased operations.
These data suggest that the location-based incentives under the CHIPS and Science Act and the IRA are effective. In that way they resemble prior section 936, which was remarkably effective in encouraging investment in Puerto Rico before it was terminated (leading to a significant economic decline and ultimately bankruptcy for the territory).8 They contrast with the Opportunity Zone program of the TCJA, which has not been effective and should be allowed to expire.9
Given the huge disparities in wealth between rich and distressed areas of the United States, a bolder experiment in encouraging investment in distressed communities can be imagined.10
Two issues need to be resolved for any U.S. regional tax relief proposal. The first is the geographic scope, and the second is what kind of tax incentive to provide.
Geographic Scope: Defining the Target
Many countries have very broad definitions of the areas that qualify for lower tax rates, such as the entire Chinese hinterland and the whole south of Italy.
The problem with using this kind of broad definition in the United States, like the whole area between the Mississippi and the eastern Appalachians, is that it is likely to be perceived as too partisan. Including all of it (and nothing else) would be rejected by Democrats as a blatant bribe to the reddest areas in the country. In addition, this area includes quite a few flourishing cities, such as Chicago; Nashville, Tennessee; Dallas; Austin, Texas; and Pittsburgh, which do not need federal aid.
A better option would be to build on an existing bipartisan consensus and use the definition of Opportunity Zones as enacted in the 2017 tax law, which in turn builds on the new markets tax credit proposed by President Clinton and enacted by Congress in 1996.11 An Opportunity Zone is defined as a census tract that (1) has at least 20 percent of its population with income below the poverty rate, and (2) is designated as such by the governor of the state.
Ideally, the target zone should include all census tracts eligible to be Opportunity Zones, not just the ones that were designated in 2018, even if some of them are close to rich areas. In that case, I would suggest that the employees that qualify would have to be residents of the zone, as is already the case for the work opportunity tax credit.12
Tax Incentive
Currently, the tax incentive offered by Opportunity Zones is rather limited. An individual investor who has unrealized appreciation in capital assets can invest the gain in a project in a zone approved by a nonprofit, and if the investor waits long enough, it will be possible to defer the original gain and exempt further gain from capital gains taxation.
To address the gap between the coasts and the heartland, something more radical is needed. The United States needs to create an incentive for large employers to move jobs to the targeted Opportunity Zones.
The obvious kind of tax incentive would be some kind of jobs tax credit, like the work opportunity tax credit, but to a larger extent. The work opportunity tax credit is available for wages paid by employers that hire individuals from targeted groups.13 The credit for a tax year equals 40 percent (or 25 percent for wages attributable to individuals meeting only minimum employment levels) of qualified first-year wages. The credit has limited application mainly because the qualified first-year wages are generally limited to only the first $6,000 paid to each member of the targeted group during the first year of employment.14
Generally, an individual is a member of a targeted group if a designated community resident.15 A designated community resident is any individual who, among others, has a principal place of abode within an empowerment zone, enterprise community, renewal community, or rural renewal county.16 The individual must continue to reside in those areas in order for wages to qualify for the credit. A rural renewal county is any county that is outside a metropolitan statistical area as defined by the Office of Management and Budget, and during the five-year periods 1990 through 1994 and 1995 through 1999 had a net population loss.17 An empirical study on the work opportunity tax credit showed that it has positive effects on targeted groups in terms of employment levels, but does not necessarily help in increasing wages in the long run (although the research found positive effects on wage increases in the short run).18
Under a tax credit, for every dollar of wages paid by a corporation to an employee within the target zone, it would get a full dollar reduction in its corporate tax liability, instead of the deduction for paying the wage (which is worth only 21 cents for every dollar paid). The problem is that this type of credit may not lead to new hires, but simply to a move of some highly paid executives to the target zone, with limited benefit for the local community. Moreover, as illustrated below, the effect of such a credit on tax liability is significantly lower than my preferred alternative.
Instead, the United States could adopt a formulary apportionment approach. An extensive economics literature has documented the effect of formulary apportionment on location decisions by multinationals. As Charles McLure pointed out, the traditional U.S. state formula of allocating corporate profits based on the location of tangible assets, payroll, and sales operates in practice like three taxes on the three elements of the formula, and it creates an incentive for the multinational to move jobs and assets to low-tax jurisdictions.19 This is generally viewed as a negative aspect of traditional formulary apportionment because it leads to job losses, which is why I have proposed repeatedly that the United States adopt a formulary apportionment method based on sales, which are harder to manipulate because consumers are relatively immobile.20
In the present context, however, the United States would like to encourage multinationals to move jobs to the target zone. Therefore, it should apportion worldwide corporate profit between the target zone and the rest of the world by using the following formula:
Total profit * (number of employees in target zone/total employees + wages in target zone/ total wages)/2 = profit in target zone
This formula differs from the traditional state formula in that it ignores sales and tangible assets, because we want to focus on jobs in the target zone. The averaging of wages and number of employees is based on the political compromise reached in the EU between the rich states (who want wages because their wage rate is higher) and the poorer states (who want number of employees because they have larger labor forces). The idea is to prevent Amazon from apportioning a lot of income to the target zone just by moving Jeff Bezos there.
From a development perspective, it doesn’t matter (and may even be better) if the multinational doesn’t hire locally but moves highly trained people into the target zone from other, richer locations. The goal is to create an incentive for people to move to poorer places, not away from them. Once people move to a place with jobs and salaries, they demand housing, and once they have housing, they demand good schools, good infrastructure, and the other amenities they had where they came from. Gradually, the poor area becomes more like the richer areas. This is what happened in Spartanburg, South Carolina, and in Tuscaloosa, Alabama, when BMW and Daimler located there in response to state and local tax incentives.
The tax rate on profits apportioned to the target zone should be significantly lower than the tax rates on other profits. Currently, the tax rate on domestic profits is 21 percent and on foreign profits 10.5 percent, or zero under QBAI if actual jobs are moved overseas. I would therefore recommend a corporate tax rate of zero on profits from the target zone (and an exemption from the corporate alternative minimum tax).
The following example illustrates how such a formulary apportionment method would work and why it is more effective than a jobs tax credit:
Assume a U.S.-based multinational that makes $1 billion in worldwide profits, pays $100 million in wages worldwide, and (for simplicity) has no other deductions and is subject to tax on worldwide income. The multinational has 10,000 employees worldwide, out of whom 5,000 work in the target zone. It pays the employees in the target zone $20 million in wages.
Under current law, the multinational would pay $189 million in tax:21
Gross income $1 billion - $100 million = $900 million * 21 percent = $189 million tax
Under a jobs tax credit, the multinational would pay $173.2 million in tax:
Gross income $1 billion - $80 million = $920 million * 21 percent = $193.2 million tax - $20 million credit = $173.2 million tax
Under the proposed formulary apportionment, the multinational would pay $122.85 million in tax:
Gross income $1 billion - $100 million = $900 million
In target zone: number of employees factor 5,000/10,000 = 50 percent
Wages factor 20/100 = 20 percent
Average of the two factors 50 + 20/2 = 35 percent
$900 million * 35 percent = $315 million taxed at zero
$900 million * 65 percent = $585 million taxed at 21 percent = $122.85 tax
Thus:
Net income (after deductions and tax) under current tax is $1 billion - $100 million - $189 = $711 million.
Net income under tax credit is $1 billion - $100 million - $173.2 = $726.8 million.
Net income under formulary apportionment is $1 billion - $100 million - $122.85 = $777.15 million.
The effective average tax rate on net income under current tax is 21 percent, under the tax credit 19.2 percent, and under formulary apportionment 13.65 percent.
The advantage of formulary apportionment is that the more employees you hire in the target zone, the lower your taxes, especially if you also pay them higher wages (in the example, the average wage in the target zone is lower, which is probably realistic at first). But this benefit lasts only as long as you maintain employment in the target zone.
Admittedly, such a location incentive would be inconsistent with pillar 2 and could result in the UTPR (formerly known as the undertaxed payment rule) being applied to U.S. multinationals that invest in the target zone. But this application of the UTPR to the domestic income of U.S. multinationals is controversial and has already resulted in the UTPR being suspended until 2025 for a country like the United States, whose statutory tax rate exceeds 20 percent. It is plausible that the United States will be able to obtain further delays in the imposition of the UTPR in the future.
FOOTNOTES
1 Both percentages decline to 5 percent over 10 years.
2 See Robert Goulder, “It’s 2024 and Trump’s Tariffs Are (Still) a Bad Idea,” Tax Notes Int’l, Feb. 12, 2024, p. 941: “The field of behavioral economics teaches us that a bad tax is one to which the predictable taxpayer responses are misaligned with national economic policy. For example, we might say that a tax provision is flawed if it encourages U.S. corporations to pour capital into foreign projects to the exclusion of domestic opportunities or encourages the hiring of foreign employees while doing nothing for domestic workers. The allowance for qualified business asset investments comes to mind because it rewards companies for making investments in tangible property held overseas. Professor Kimberly Clausing has previously referred to QBAI as the ‘America Last’ tax provision.”
3 Martin A. Sullivan, “The Not-So-Obvious Effects of Pillar 2 on Tangible Capital Investment,” Tax Notes Int’l, Jan. 1, 2024, p. 23.
4 See, e.g., Patrick Kline and Enrico Moretti, “Local Economic Development, Agglomeration Economies, and the Big Push: 100 Years of Evidence From the Tennessee Valley Authority,” 129 Q. J. Econ. 275 (2013); Edward L. Glaeser and Joshua D. Gottlieb, “The Economics of Place-Making Policies,” 2008 Brookings Papers on Economic Activity 155 (Spring 2008). See also Jahan Ganesh, “The Era of the Unfixable Problem,” Financial Times, Mar. 19, 2024 (arguing that there is no policy solution to location-based income disparities based on the U.K. and U.S. experience); Paul Krugman, “What’s the Matter With Ohio?” The New York Times, Mar. 21, 2024.
5 See Benjamin Austin, Edward Glaeser, and Lawrence H. Summers, “Saving the Heartland: Place-Based Policies in 21st Century America,” 2018 Brookings Papers on Economic Activity 151 (Spring 2018).
6 Joseph Parilla et al., “Strategic Sector Investments Are Poised to Benefit Distressed U.S. Counties,” Brookings, Feb. 13, 2024.
7 There is a 10 percent or 20 percent bonus credit for clean energy investments in certain low income communities, as well as a separate credit for projects in “energy communities” that are tied to the existence of a coal plant, coal mine, brownfield site, or specific percentage of fossil fuel-related jobs. There are also significant grant programs for clean energy projects in “rural and remote” areas of the country and on former “mine lands.”
8 Harry Grubert and Joel Slemrod, “The Effect of Taxes on Investment and Income Shifting to Puerto Rico,” 80 Rev. Econ. Stats 365 (1998); Zadia M. Feliciano and Andrew Green, “U.S. Multinationals in Puerto Rico and the Repeal of Section 936 Tax Exemption for U.S. Corporations,” National Bureau of Economic Research Working Paper 23681 (2017) (the elimination of section 936 had the effect of decreasing average manufacturing wages by 16.7 percent and decreasing the number of manufacturing establishments by 18.7 percent to 28 percent).
9 Michelle D. Layser, “Subsidizing Gentrification: A Spatial Analysis of Place-Based Tax Incentives,” 12 UC Irvine L. Rev. 1 (2021) (“Together, these results provide new evidence that gentrifying census tracts may draw tax-subsidized investment away from other eligible areas. They also suggest that a commonly proposed Opportunity Zones reform — to add statutory safeguards modeled after those in the NMTC [new markets tax credit] — would fail to prevent tax-subsidized investment in places that are already gentrifying.”); Layser and Andrew Greenley, “Structural Inequality and the New Markets Tax Credit,” 73 Duke L. J. 800 (2024).
10 The following is based on Reuven Avi-Yonah, Orli Avi-Yonah, Nir Fishbien, and Haiyian Xu, “Federalizing Tax Justice,” 53 Indiana L. Rev. 461 (2020).
11 Sections 1400Z-1 and 1400Z-2; Notice 2018-48, 2018-28 IRB 9.
14 See section 51(b)(2)(i). The wage limitations are set to $3,000 (instead of $6,000) for qualified summer youth employees and to 25 percent (instead of 40 percent) for individuals meeting only minimum employment levels.
16 Section 51(d)(5)(A) and (B).
18 Sarah Hamersma, “The Effects of an Employer Subsidy on Employment Outcomes: A Study of the Work Opportunity and Welfare-to-Work Tax Credits,” 27(3) J. Pol’y An. and Mgmt. 498, at 518 (Summer 2008).
19 Charles E. McLure Jr., “The State Corporate Income Tax: Lambs in Wolves’ Clothing,” The Economics of Taxation (1980). See also Kimberly A. Clausing, “The U.S. State Experience Under Formulary Apportionment: Are There Lessons for International Reform?” 69 National Tax Journal 353 (2016).
20 Avi-Yonah, Clausing, and Michael C. Durst, “Allocating Business Profits for Tax Purposes: A Proposal to Adopt a Formulary Profit Split,” 9 Fla. Tax Rev. 497 (2009); Avi-Yonah, “The Rise and Fall of Arm’s Length: A Study in the Evolution of U.S. International Taxation,” 15 Va. Tax Rev. 89 (1995); Avi-Yonah, “Slicing the Shadow: A Proposal for Updating U.S. International Taxation,” Tax Notes, Mar. 15, 1993, p. 1511.
21 For simplicity, I ignore the fact that the multinational may pay only 10.5 percent or less on its foreign profit. The incentive under the TCJA to move jobs overseas and pay 0 percent on the profit should be abolished. See Avi-Yonah, “Guilty as Charged: Reflections on TRA 17,” Tax Notes, Nov. 20, 2017, p. 1131.
END FOOTNOTES