Making Sense of Profit Shifting: Thomas Neubig

June 19, 2015

Thomas Neubig is Deputy Head of the Tax Policy and Statistics Division at the OECD’s Centre for Tax Policy and Administration.

Mr. Neubig is currently working on the OECD/G20 Base Erosion and Profit Shifting (BEPS) project’s “Action 11,” which aims to improve the quantitative analysis of profit shifting.

Mr. Neubig has over 30 years of experience as an economist in both government and the private sector. He has spent the last 20 years at Ernst & Young, where he founded and led a group of 34 tax policy economists, statisticians, and survey specialists. Mr. Neubig was EY’s Director of Quantitative Economics and Statistics, serving as an advisor to numerous public and private clients on federal, state, and global tax policy issues. Prior to joining EY, Mr. Neubig was National Director of Financial Sector Tax Policy Economics at PriceWaterhouseCoopers International Limited, from 1990 to 1994.

Previously, Mr. Neubig was Director and Chief Economist of the U.S. Department of the Treasury’s Office of Tax Analysis. He served as the top Executive Branch career economist in tax policy and directed a staff of 70 people, consisting of professional economists, revenue estimators, computer specialists, and support staff.

Mr. Neubig received his Ph.D. in Economics from the University of Michigan.

In this interview with the Tax Foundation, Mr. Neubig shares his insights into the forefront of what is known, in a quantitative sense, about profit shifting. Specifically, Mr. Neubig examines the current limitations to estimating the magnitude of profit shifting, the status quo of existing data sources used to measure profit shifting, key issues in determining where value added occurs on a geographic basis, and why using descriptive statistics to infer profit shifting is problematic. This interview is part of our 2015 Tax Foundation Forum series and has been edited for clarity and length.

Tax Foundation: How do we know that profit shifting exists?

Thomas Neubig: There have been literally hundreds of economic analyses using different methodologies and different databases, and they have found that multinational enterprises shift taxable income from higher-tax to lower-tax jurisdictions. A German researcher, Nadine Riedel, recently surveyed the empirical literature, and concluded: "Existing studies unanimously report evidence in line with tax-motivated profit shifting, despite using different data sources and estimation strategies. In terms of shifting channels, there is evidence consistent with strategic mispricing of intra-firm trade, the location of valuable intellectual property at low-tax affiliates, and debt shifting activities."

We have also seen evidence from several recent legislative inquiries in different countries that have put the spotlight on the base erosion and profit shifting (BEPS) behaviors of specific multinational enterprises. This information is often times not included in the databases that the academic researchers are using.

So there is no question that profit shifting exists. The issues are: the pervasiveness of the behavior, the magnitude of the profit shifting and base erosion, and how to correct the problem in an economically favorable manner.

TF: Does profit shifting matter?

Neubig: Absolutely, profit shifting matters. The leaders of the G20 countries would have not asked the OECD for an action plan to address profit shifting if it didn't matter. BEPS behaviors are artificial schemes to minimize taxes by exploiting legal differences in countries’ tax rules. Besides the wasteful misallocation of economic talent on socially unproductive tax engineering, ordinary citizens are harmed if tax rules permit multinational businesses to reduce their taxes by BEPS behaviors. As a result, citizens have to pay higher tax burdens to replace that lost revenue, other businesses are harmed if they have to pay higher taxes—plus, they may face competitors with “BEPS-reduced” cost structures—and whole countries are harmed if the integrity of their tax systems is compromised. Many developing countries have lost tax revenue, leading to underfunding of critical infrastructure important to their future growth. So, yes, stopping profit shifting matters, and that's why I came to the OECD rather than retiring.

What is known quantitatively about the magnitude, economic impact, and spillover effects of profit shifting?

There is an awful lot more that needs to be learned about the magnitude of the fiscal and economic effects of base erosion and profit shifting. Recently, the OECD released a discussion draft of the BEPS Action 11. Action 11 is designed to improve the analysis of BEPS. That discussion draft, which I was very involved in, addresses three aspects. First, it assesses the existing data sources and their many limitations. Second, it provides potential indicators of the scale and economic impact of BEPS to be able to track it historically and monitor it in the future. Third, it describes some of the existing empirical analyses of BEPS and proposes two complementary approaches to estimate the scale of the fiscal effects of BEPS.

The empirical literature clearly shows BEPS behaviors occurring, for example, in the location of valuable intangibles in low-tax countries, the placement of both external and internal debt in high-tax countries, and the tax planning to artificially separate taxable income from the activities that generate that income.

The international tax rules have not kept up with the rapid changes in the business sector—with its global supply chains, the increase in services and intangible investments, and also increasingly mobile capital and labor. Internationally coordinated and comprehensive tax rules need to address both double taxation and double non-taxation to ensure the continued important role of multinational enterprises in global growth.

Clearly, a lot more work needs to be done in terms of quantitatively knowing the magnitude and economic impacts. A critical part is improving the available data.

So at this point, there is no real estimate of the magnitude or the economic impact of profit shifting?

That's one of the goals of Action 11, and this discussion draft talks about two different approaches to estimating the fiscal magnitudes as well as some of the other economic effects. The final BEPS Action 11 2015 deliverable will provide policy makers and stakeholders with estimates of the potential magnitude, albeit limited by the available data.

Which metrics could be critical in monitoring and evaluating the efficacy of actions taken to address profit shifting?

The Action 11 discussion draft proposed seven specific metrics to monitor BEPS behaviors. Our view is that a single metric would not capture all of the important dimensions of BEPS behaviors, and we are certainly open to suggestions for alternative metrics. An important factor in measuring the efficacy of the proposed BEPS action items will be the enactment, implementation, and enforcement of those actions in the leading economies.

Also, corporate tax hopefully will not be headline news after the BEPS project is fully implemented, since there are many more important policy issues facing countries, including achieving peace and stimulating economic development. Achieving a business tax system that is conducive to economic growth, rather than paper shifting or legal schemes that divert policy makers from their attention on more important issues, will be one of my personal measures of the success of the BEPS project in the future.

You mentioned seven metrics in monitoring BEPS. Which areas are those seven metrics related to?

The seven that we proposed with the available data use national account statistics as well as individual company financial reports. They focus on foreign direct investment relative to GDP in different countries, effective tax rates relative to profit rates in companies, and the concentration of interest and R&D royalties. So they look at some of the key base erosion and profit shifting channels as well as some of the potential effects that could be observed when you are seeing shifting profits across jurisdictions.

I would like to talk more about measuring or perhaps inferring the existence of profit shifting. In the past decade, reports have been published using real economic activity as a proxy to measure profit shifting. What are some advantages and limitations in using real economic activity—such as assets, sales, employee compensation, and similar variables—in a descriptive way to gauge profit shifting?

Profit shifting is where there is a disconnect between where the taxable income is reported and where the underlying economic activity that created it actually occurred. Measuring income and measuring economic activity are both very hard, even when you are thinking about just a single domestic economy. The difficulties of measurement increase manyfold in the international tax context, where you are attempting to measure where income is created across different jurisdictions.

My former Ernst & Young colleague, Bob Cline, also joined the OECD to work on the BEPS Action Plan. Bob spent almost 40 years working on state and local tax issues in the United States. He taught me a lot about some of the measures state tax systems use in terms of assets, sales, and payroll that have conceptual and practical difficulties in terms of their correlation and use in the assignment of income by jurisdiction.

Recent studies have shown that intangible investments account for more than 50 percent of total investment in a number of major economies, yet most intangible assets don't show up on businesses balance sheets, since the cost of developing those intangible assets are expensed or deducted currently rather than capitalized. National Accounts only recently capitalized R&D investments, but not other intangibles.

And since intangible income is such a large fraction of business income, it's highly conducive to profit shifting and isn't properly measured in the financial reports. If you're just using depreciable assets, employees and land to measure economic activity, that's a real problem in terms of allocating a significant portion of business income that comes from intangible investments.

When you look at sales, there are questions, in terms of the reporting, about whether sales are measured where the production takes place or whether they're measured on a destination basis, where the services or goods are actually consumed. That has been an issue at the state and local level in the U.S.

Counting employees is also very different than measuring their labor contribution, because there are different levels of labor productivity across countries. And then if you attempt to measure total labor compensation, you should be including fringe benefits and profit sharing that would impose significant burdens on multinational businesses.

So you also have disconnects between good measures of profitability and some of these available measures of economic activity. Businesses have different production functions, and it’s not one size fits all, especially since an awful lot of the income is now earned from unique intangible assets and that's not even in the financial statements.

It's just very difficult to do a regression analysis that has high explanatory power relating reported income to some of these measurable economic factors, let alone with financial statement tax variables or just headline corporate statutory tax rates.

What are key issues in measuring profit shifting?

There are two different approaches that we outlined in the Action 11 discussion draft. One follows the approach that a lot of academic studies have taken that look at differences in tax rates within multinationals and where profits are reported. They clearly find that reported profits are very sensitive to tax rate differentials and that reported profits are lower in higher-tax rate countries and higher in lower-tax rate countries. Based upon the estimated responsiveness of reported profits to tax rates, one can get a rough estimate of what that would mean in terms of the total amount of global profit shifting.

Another approach would be a more standard revenue estimating approach in terms of identifying what specific changes in the law would raise in terms of revenue. There are going to be specific recommendations made by the OECD/G20 BEPS Action Plan in terms of changes in the tax rules for transfer pricing, interest, treaty abuse, permanent establishment, hybrid arrangements, and controlled foreign corporation areas, as well as several increased transparency recommendations. A number of countries are already estimating the revenue effects from those types of proposals that have already been enacted or proposed. So estimating the revenue effects from those specific changes in the law is another approach that can be done—but again, using limited data and information. Data and information limitations are a severe problem for analysts trying to do empirical estimates analyzing BEPS with currently available data.

What are the key issues or limitations in the two methodologies to estimate the extent or magnitude of profit shifting that you just highlighted?

The biggest issue is data limitations, and that's where I would recommend to your readers to look at the Action 11 discussion draft that talks about the available data and also about the significant limitations of the currently available data to analyze BEPS. BEPS behaviors affect all kinds of data, from affecting countries’ measurement of gross domestic product, to affecting companies' financial statement reports of profits as well as affecting what is reported on countries' tax returns. Important data are not reported and the information that is reported is often tainted by BEPS behaviors.

And then you also have the issue that what you're really wanting to do is separate the effects of BEPS behaviors from real economic factors. If you're looking at foreign direct investment across countries, probably the vast majority of that is driven by real economic factors—multinational businesses making important investments in the global economy. But there are also some elements of foreign direct investment that are driven by BEPS behaviors, often through special purpose entities and very low-tax countries.

We are going to see BEPS Action 13 make a significant step in ensuring that tax administrators have a more complete picture of the multinational enterprises operating in their countries. Many people have commented that when tax administrators are able to see the whole picture then that will affect what multinational enterprises do for tax purposes. So Action 13 and the country-by-country reporting will be a really important advance and will enable tax administrators to have much better information on how to administer their country's tax rules, and it will affect MNEs’ tax planning

I would say the data limitations are the most serious problem, and then there is the fact that you have to make estimates of BEPS that separate it from the real economic effects.

What is the current state of existing data sources used to measure profit shifting?

Recently there has been a preference among the academic researchers to use micro-data for individual firms to analyze BEPS. That clearly was an attempt to try to better separate out the BEPS behaviors from the real economic factors involved in where multinational enterprises are doing business across countries. So that was a very desirable change. However, I don’t believe academics’ studies have been as forthcoming in terms of the limitations of some of these micro-databases as would have been helpful to the readers, because the coverage of some of these micro-databases, especially the global ones, is very spotty once you get beyond the European countries, and even data for many European countries are missing important MNE entities.

So in the case of countries like the U.S. and especially among the developing countries, the percentage of companies that are actually in these databases is very, very low. So the question is: Can the tax responsiveness found in academic studies be extrapolated to the U.S., which has the highest corporate tax rate among the leading major economies? And can it be extrapolated to the developing countries that often have lower statutory tax rates, but rely upon the corporate income tax much more than the developed economies? So there are problems in terms of some companies not being included. Some of the companies that are mentioned in the news reports coming out of the legislative inquiries about BEPS behaviors are not included in the databases. You do not find a lot of the familiar names in these databases.

When you do find some of them listed in the database, you find that all the databases include is the name of the company but no financial information. So you don't know how many billions of dollars of income there is in particular affiliates and particular companies. It's an incomplete picture. And it's also using financial statement information about accounting tax expense rather than actual tax return information. This is where the U.S. has some best practices, both having corporate income tax return information of MNEs that the IRS releases in aggregate form plus information from the U.S. Bureau of Economic Analysis on multinationals’ affiliates’ investments. The U.S. also has rules that maintain confidentiality while allowing qualified academic researchers to analyze this data. I think that is a real best practice for understanding and analyzing MNE behaviors and BEPS.

Some of the focus on U.S. multinational enterprises is because it's like the drunk looking for his keys under the lamppost. The fact that the U.S. has and releases much better data about its multinational enterprises means that's where researchers gravitate. Yet we know that 75 percent of the global Fortune 500 multinational groups are non-U.S. groups. BEPS is a worldwide phenomenon. The differences in tax rates between 25 percent and zero percent are very large, and so BEPS is clearly not a U.S. phenomenon. It's a global phenomenon.

What is the difference between country-level and firm-level data? How significant is the difference, and what is the distinction in the strength of inferences one can draw?

As I mentioned before, a lot of the academics have been increasingly using individual company data to analyze BEPS behaviors, and they've been finding somewhat smaller measures of responsiveness to tax rate differentials. That is good in the sense that by using individual firm data you are able to hold a lot more constant when you are analyzing the location of profits than when you're working with aggregate country-level data and don’t have all the additional detail that is available from the financial statements.

I would not, however, discount the importance of some of the country-level statistics to be helpful in understanding the fiscal and economic effects of base erosion and profit shifting. Macro data complements the micro-data. Kim Clausing at Reed College has done some important work, both with micro-data and aggregate data. They both show important dimensions of BEPS behaviors.

Some people have concluded that—as a result of moving to micro-data, over time—some of the estimates of the magnitude of profit shifting have declined. I think that's more due to a change in methodology rather than an actual change in MNE behavior.

When you look at some of the metrics we proposed as indicators in the discussion draft, you can see that some of them are definitely increasing over the last 10 to 12 years. A number of empirical studies have also concluded that base erosion and profit shifting has been increasing over time.

Also, in some cases, there have been constraints on profit shifting through thin capitalization rules and tightening of transfer pricing rules and enforcement. A number of these empirical studies have found that those types of changes have had positive effects in terms of reducing profit shifting.

What is known about the relative magnitude of profit shifting between a parent firm and its foreign affiliates versus profit shifting between a firm’s foreign affiliates in high-tax and low-tax jurisdictions?

A number of empirical academic studies have looked at profit shifting from different dimensions and you are specifically asking about the profit shifting between parent and affiliates versus profit shifting between affiliates. All of them are finding that profit shifting is occurring. I would recommend several studies, such as a survey of the literature by Dhammika Dharmapala as well as the IMF’s Ruud De Mooij’s meta-analysis. These studies find profit shifting via intangibles and royalties as well as interest shifting is occurring across the different channels. Whether there is shifting from a higher-tax rate parent to a lower-tax country or whether from a higher-tax rate affiliate is having its profit shifted to a low-tax jurisdiction, both are profit shifting when you look at it from a global perspective, and that's the perspective the BEPS project is taking.

So the relative magnitude of profit shifting between a parent firm and its foreign affiliates versus profit shifting between a firm’s foreign affiliates in high-tax and low-tax jurisdictions isn’t a relevant distinction to delve deeper into?

From a global perspective, I don't think it is. A number of years ago, some U.S. policy makers said they didn't care about profit shifting from one foreign country to another. From a global perspective, we need to worry about the adverse effects of profit shifting no matter where it occurs.

Are multinationals a homogeneous group with respect to profit shifting?

After working in the private sector for over 20 years, I can attest that businesses, including multinational enterprises, cannot be treated as homogeneous. We'll be making a big mistake if we did that in terms of our analysis. Tax planning involves many issues of risk management and corporate governance. Those differ across the businesses that I've seen.

Several academic studies, such as a recent study by an Irish professor, Ron Davies, find that transfer pricing is concentrated among some of the largest multinationals in a select group of countries. Other studies have suggested that profit shifting could be much more widespread.

An important issue is that even if profit shifting was currently limited to a smaller number of multinational enterprises, we need to think about the dynamics such that if BEPS is not checked, if BEPS continues to be available in the future, then the number of companies doing profit shifting could increase very quickly. We saw that in the U.S. with corporate inversions, where it was relatively quiet and then all of a sudden there was this very large spike in the number of reported inversions. We certainly would not want to see that happen in the BEPS area. Whether BEPS is small or large, the key is addressing BEPS problems now on a coordinated international basis.

Some academic studies using European data suggest that profit shifting is decreasing in magnitude, and some studies using U.S. data suggest that profit shifting is increasing in magnitude. Has profit shifting increased over time, and why—in a directional sense—is there a difference between results using European and U.S. data?

More empirical analysis with better data is needed. When you look at current surveys of the literature, they do report smaller measures of multinational enterprises’ responsiveness to tax rate differences coming out of studies now than 10 or 20 years ago. That's more of a methodological change, where they are now looking more at micro-level, firm data.

My sense is that a lot of the European research, which is very good research, is working with these less than comprehensive proprietary financial statement databases. So they are not working with actual tax return information, and they have a focus on European countries and the particular economic and regulatory environment in the European Union. So when you look at U.S. studies, more of them are using actual tax return data. They are looking more specifically at U.S. parent multinational enterprises.

There is excellent research over a number of years by Harry Grubert at the U.S. Treasury and his co-authors. They are using actual income tax return information from the U.S. Others are using government data from the Commerce Department and Bureau of Economic Analysis that is very comprehensive and includes all of the affiliates of multinational enterprises, compared to some of the European studies using proprietary databases, where an awful lot of the affiliates are not included or there is only financial data. So differences in data sources are likely to explain the differences that good researchers, both in Europe and the U.S., are finding. I would note an important new study by Tim Dowd, Anne Moore, and Paul Landefeld of the U.S. Congressional Joint Committee on Taxation using a comprehensive database of tax return information from U.S. affiliates abroad, which finds a high responsiveness of profit shifting to tax rates, particularly to countries with very low average tax rates.

Tim Dowd, Anne Moore, and Paul Landefeld’s study, “Profit Shifting of U.S. Multinationals,” through a new, innovative methodology, examines the profit shifting behavior of U.S. multinationals and concludes that the standard approach for estimating profit shifting in the literature—the “log-linear specification”—may significantly understate the magnitude of profit shifting.

And in a directional sense, has profit shifting increased over time?

I'm working very hard on producing the 2015 Action 11 final report. We are still doing a lot of empirical work coming up with estimates of the scale and economic effects of BEPS, so I'm still holding judgment. Clearly, BEPS is a significant problem. Whether it's increasing or decreasing, it needs to be addressed because as I said before, if it's not addressed, that would be a really bad signal to companies that they can proceed on their merry way to the detriment of citizens, other businesses, and countries as a whole.

So is it fair to say that at this point we don't have conclusive evidence on that profit shifting has either increased or decreased over time?

I would say stay tuned for the final Action 11 report that will provide some additional insights on BEPS and the effects of potential countermeasures. But again, data limitations and the difficulties of estimating BEPS from real economic behaviors are going to be constraints on the level of precision of such estimates. Better data and increased access to available data for analysis, as well as to tax administrations, will be important outcomes of the BEPS Action Plan.


Related Articles