Making Sense of Profit Shifting: Dhammika Dharmapala
May 14, 2015
Dhammika Dharmapala is a Professor of Law at the University of Chicago Law School.
Professor Dharmapala is recognized as one of the leading experts on profit shifting, through his innovative research on the magnitude of profit shifting and his recent survey of the empirical profit shifting literature.
Professor Dharmapala’s research has been published in leading journals in the fields of law, economics, and finance and focuses on taxation, the economic analysis of law, and corporate finance and governance. His tax scholarship has featured areas such as the economic impacts of tax policy, the role of tax havens in corporate tax planning, and how dividend taxes affect portfolio choice.
Professor Dharmapala received his Ph.D. in economics from the University of California, Berkeley. He is an International Research Fellow at Oxford University Centre for Business Taxation and a Fellow of the CESifo Research Network. Professor Dharmapala serves on the Board of Directors of the American Law and Economics Association and on the Board of Management of the International Institute of Public Finance. Between years 2010-2013, he was Editor-in-Chief of the journal International Tax and Public Finance.
In this interview, Professor Dharmapala outlines the latest thinking on profit shifting, with a specific focus on the examination of the magnitude of profit shifting, the potential distinctiveness of firms engaging in profit shifting, and the political economy of international taxation and multilateral cooperation. This interview is part of our 2015 Tax Foundation Forum series and has been edited for clarity and length.
Tax Foundation: What is known about profit shifting?
Dhammika Dharmapala: I think that quite a lot is known, although there is much that remains unclear. The academic literature has been studying this topic since the 1990s. Overall, I would say that it finds strong evidence that multinational firms are responsive—in terms of the reporting of income—to tax differentials across jurisdictions.
This is not perhaps in itself all that surprising, and I think more relevant for policy makers, quite possibly, is the magnitude of profit shifting.
In terms of the magnitude, the early studies from the 1990s used aggregate country-level data. And they found quite large effects. More recently, scholars have used firm-level unconsolidated data, that is, data sets that provide separate data on each multinational’s affiliates. They have found a much smaller magnitude of profit shifting.
Just to give you a sense of what I would judge to be the consensus estimate that comes out of this academic literature: Imagine you have a multinational affiliate in a country that reduces its tax rate from 35 percent to 25 percent, i.e., by 10 percentage points. An affiliate that was previously reporting $100,000 of income would, according to these estimates, increase its reported income in that country, which now has a lower tax rate, to $108,000. This corresponds to what economists refer to as a ”semi-elasticity” of 0.8…
[The Heckemeyer and Overesch study] is one source for this. They find that to be the consensus estimate. There are actually some more recent studies that have found smaller estimates still, even though overall there's still evidence for profit shifting occurring. But I do think policy makers need to take into account the magnitude as well as the existence of profit shifting…
Lohse and Riedel have a working paper that’s primarily about the effects of transfer pricing regulations on profit shifting. But their estimates are that the overall magnitude of profit shifting is actually smaller than that. It's about half the magnitude of that 0.8. semi- elasticity.
So we know a fair amount about the phenomenon, but there are certainly many unanswered questions as well.
Additionally, I think there are two mechanisms that are thought to drive profit shifting. One has to do with the setting of transfer prices in ways that result in more income being reported in lower-tax jurisdictions. And the other uses inter-affiliate debt—interest payments from high-tax to low-tax affiliates.
TF: What are the most important unanswered questions at this point, in your opinion?
Dharmapala: I think there’s a broader question of how we can better understand profit shifting. More data is always helpful, but the primary constraint on our ability to understand the phenomenon more thoroughly is the limited amount of variation that exists in the world.
So, for instance, to find out how much income reported in Ireland is tax-motivated, we would need ideally another country that’s identical to Ireland except in its tax characteristics.
And we don’t observe that degree of fine-grained variation in the real world, which limits the strength of the inferences that we can draw.
You mentioned transfer pricing and debt as two vehicles or drivers of profit shifting. What are other unique factors or underlying drivers of this phenomenon? Is it primarily differences in countries’ tax law?
Differences, or perhaps unintended interactions, among the tax laws of different countries are an important part of the story. However, I would also emphasize the strategic incentives of governments. In particular, residence countries have an interest in encouraging their resident multinationals to avoid foreign taxes.
Indeed, the current discussion has been characterized by one very prominent observer as being a situation where people are complaining that everyone else's multinationals are not paying their fair share of tax.
So it’s in part driven by what you said, the unintended interactions of different countries' tax laws, but also by the residence countries’ incentives to encourage the avoidance of foreign taxes. And, of course, the source countries’ attempts to prevent that.
Let’s discuss firm behavior more specifically. How does profit shifting alter the behavior of firms?
Profit shifting is a phenomenon that in some ways doesn't involve a lot of change in firm behavior and in some respects has relatively small real economic costs. The costs of profit shifting are primarily the time and effort of tax planning professionals.
In contrast, if firms were to shift real economic activity, where they build factories and so forth across countries in response to tax differentials, that would create substantial economic distortions. So, in some respects, profit shifting does not alter the behavior of firms very much, if by behavior we are referring to real economic activity.
How about financial restructuring? What’s the extent to which firms alter their behavior in that regard?
There are transactions such as corporate inversions that involve substantial changes in ownership structure and perhaps in behavior. But profit shifting, or what you might think of as everyday profit shifting, is quite different.
It does not necessarily involve substantial changes in either behavior or ownership structures, except to the extent that multinational firms may set up affiliates in tax havens that they otherwise may not, in order to engage in tax planning strategies. But the costs of this are not necessarily that large.
But I do want to emphasize that profit shifting does have distributive consequences. Some taxpayers end up paying less and others end up paying more.
Does profit shifting matter?
I would view that as an issue for policy makers to decide. In the light of the evidence and of the various competing priorities, I see the role of academic researchers as being to provide and refine the evidence to the best of our abilities.
Certainly it's an issue of public concern. How it should rank in relation to other issues of public concern is not for me to say.
And the evidence that you have seen, what do they tell you in terms of for whom profit shifting matters the most?
The evidence suggests that the magnitude of profit shifting is perhaps somewhat smaller than we had previously thought. We have also seen relatively robust corporate tax revenues in most OECD countries.
However, the question of where this should rank in terms of policy preferences is somewhat beyond what that evidence can tell us. But I do hope for evidence-based policy making on this and other issues. And I hope that the evidence is taken into account.
I would like to focus for a minute on a passage in your paper, “What Do We Know About Base Erosion and Profit Shifting: A Review of the Empirical Literature.” One of your focal points there is on tax planning. You note: "While tax revenue obviously matters to governments and non-multinational corporations and taxpayers, from a global welfare perspective, the primary concern isn’t with the distribution of revenue across governments but with the real resources expended in tax planning and compliance." Can you explain that?
Well, if we want to think about how tax planning affects economic efficiency in terms of the size of the global pie, then we need to begin with the real resources that are expended. And these refer primarily to the time and effort of tax planning professionals who, in a world where tax planning was less prevalent, may be engaged in other activities. So that’s the primary economic efficiency cost of tax planning.
It also, as I said previously, redistributes revenue across governments… and redistributes tax burdens across different taxpayers. Those are important, but I think what I was trying to argue there was that we should start with these efficiency consequences.
There are other mechanisms that we can potentially use to redistribute revenue across governments, if that’s what we want to do. And we should focus on policies that reduce the deadweight costs of tax planning.
Has profit shifting increased over time?
There are some studies that claim that it has. The Lohse and Riedel paper that I mentioned earlier does contain a direct test of this. They actually find that it has not and that it may even have declined. I think it's an issue on which we need a broader range of evidence. But I wouldn’t find it at all surprising that it has declined over time because governments over the past decade or so have greatly expanded the scope of thin capitalization rules and transfer pricing regulations.
So that may well have reduced the extent of profit shifting, but it’s an issue on which I think we need more evidence.
If one would differentiate between the absolute dollar value of profits shifted and the responsiveness of firms to tax rate differentials, have those two trends moved in the same direction over time?
The problem is that the regulatory environment and the legal environment have not been constant. So it is possible that firms have become more inclined to shift profits, other things equal, but that stronger thin capitalization rules and transfer pricing regulations may have hindered them in this objective.
Do you see that as an important distinction? On the one hand the aggressiveness of firms and then on the other hand the absolute dollar value of profits shifted?
It would be nice to separate them out in some sense, because policy makers would like to know how effective particular rules are… [what] we observe is the net result of all of that. If there are changes in firms’ aggressiveness in profit shifting, that’s difficult to separate out from contemporaneous changes in the regulatory environment.
There's a divide in the academic literature on the direction of profit shifting. There’s your and Riedel’s 2012 study, and you mentioned previously the Lohse and Riedel study, that suggest profit shifting has decreased in over time.
And then there are studies using U.S. data—studies by Kimberly Clausing, Harry Grubert, and Ken Klassen and Stacie Laplante—that suggest that profit shifting has increased over time. How can one reconcile such opposing indicators and is there perhaps a difference in the underlying data that make them incomparable?
Some of those U.S. studies use tax return data that isn't widely available. Some use worldwide consolidated financial statement data from a database known as Compustat, as opposed to unconsolidated affiliate-level data, which has been used widely in European studies. So there are a variety of reasons why we might expect differences. In addition, there are differences in methodology that make it difficult to compare the various studies.
In my own survey, I focused on the studies coming out of Europe because it was possible to develop a narrative where there is a common methodology across a large number of studies and that makes the estimates very comparable.
Reconciling the results of these alternative methodologies is still something that we need to move towards.
Is there any estimate or range you would point to as the best estimate of profit shifting out of the U.S.?
I don't think we have a strong reason to think that it’s any different than profit shifting elsewhere. It may be, but I don't know that that has been firmly established. For instance, Kevin Markle has a paper that compares profit shifting in territorial and worldwide systems, and he doesn't really find that U.S.-based firms are substantially different from those based anywhere else.
What are implications for tax revenues for the U.S. Treasury when it comes to profit shifting?
Well, I have not done that calculation. I am aware of some estimates, but they are based on profit shifting magnitudes that come out of country-level data sets. They are based on a magnitude of profit shifting that’s a little bit larger than the current consensus estimate, which means that if we were to use those [more recent] firm-level estimates of the magnitude of profit shifting, the revenue impacts would be correspondingly smaller.
But I don't actually have a number off the top of my head. I do think the revenue losses are an important issue, but the ultimate importance is limited by the relatively small role that the corporate income tax plays in aggregate tax revenues.
Would it be helpful to think about the magnitude of profit shifting in terms of a percentage of corporate tax revenues?
One could think in those terms, and that's an important question to ask and to try to answer. I'm not sure that we have a great answer or a consensus answer to it at the moment.
But the profit shifting phenomenon is quite multifaceted. So, for instance, when U.S.-based multinationals shift income out of foreign high-tax countries to foreign low-tax countries, that has two different effects that arguably increases the well-being of the U.S. One is that it saves on foreign taxes. So the firms are and their shareholders are better off in after-tax terms.
Secondly, if and when the shifted income is eventually repatriated to the U.S., it results in a greater U.S. tax liability than if the higher rate of the high-tax foreign country been paid.
So it may not capture every facet of this phenomenon to simply look at it in terms of forgone revenue.
What’s missing for a better understanding of profit shifting?
I think the source of many of the differences of interpretation are not primarily a result of the data but of the extent of variation that we can use.
There are a limited number of countries in the world, and they don't vary in quite the fine- grained way that we would need to be able to resolve all of the uncertainties about profit shifting. And I'm not sure that there's a lot that can be done about that. But additional data would certainly be helpful. And I think data sources have become much better over time and hopefully will continue to do so.
What are possible solutions to address profit shifting?
I would start by reiterating the point earlier that different countries have varying interests with respect to this question. So multilateral cooperation may be helpful, but it’s difficult to achieve in part because residence countries benefit from their multinationals' avoidance of foreign taxes.
And so that limits the incentives that the residence countries have into various forms of multilateral cooperation or to enact strong CFC rules. That’s a caveat about the potential for multilateral cooperation. It's an issue that would need to be overcome.
As I mentioned before, many countries have unilaterally strengthened their thin capitalization rules and transfer pricing regulations. And this may be the direction in which additional countries go in the future. So that would be another possible solution. Tax rate reductions, or at least reductions in tax rate differentials across countries, would reduce firms' incentives to shift income across jurisdictions.
But obviously there are revenue and political constraints on that. We have seen countries increasingly move from worldwide to territorial systems and that has been discussed in the U.S. as well. The study by Kevin Markle that I mentioned earlier actually finds that profit shifting isn’t very different across worldwide and territorial residence-country multinationals. And that seems to be because of the important role that the deferral of home-country tax on foreign income plays in systems.
What could a worldwide versus a territorial system mean for the U.S.?
One benefit of the territorial system would be that U.S. multinationals would more easily be able to access cash held overseas in foreign affiliates.
So that would be one benefit. It may not very dramatically affect incentives for profit shifting, however, for the reasons I mentioned earlier in relation to Kevin Markle's work. He finds that firms based in both worldwide and territorial countries tend to shift income to about the same degree.
So if we mean specifically in relation to profit shifting, it's not completely clear how much difference it would make. But if, for instance, the U.S. would impose a minimum tax on foreign income—something that has been discussed recently—that may affect incentives particularly to shift from one foreign country to another.
In the public discourse, there's considerable discussion about multinationals as an almost entirely homogeneous group. To what degree are multinationals a homogenous group with respect to profit shifting?
One thing that has surprised me in working with this data is the heterogeneity among multinational firms.
For instance, in the paper with Nadine Riedel, we found that a very large fraction of these multinational groups in our data set had no tax haven affiliates.
And Desai, Foley and Hines using BEA data on U.S. multinationals have also found that there's a very substantial fraction, something like 40 percent or 50 percent, that do not have affiliates in tax havens, even though that's a low-cost way to facilitate tax planning. So that suggests that in relation to tax planning activity multinational firms are very heterogeneous.
And this isn’t something that’s fully understood, I think. Economic models tend to assume that firms will take advantage of whatever lawful tax avoidance opportunities exist. But that seems not to be the case in reality. And so we don't fully understand what’s driving that heterogeneity and those differences in behavior.
You have observed a fixed cost parameter that you, I believe, think is important in investigating the heterogeneity of multinationals. Can you elaborate on that?
This is more of a hypothesis, which I hope researchers will examine in the future. But, simply put, what I was saying there is that the failure of many multinationals to have haven affiliates and therefore presumably their failure to engage in tax planning is consistent with a substantial fixed cost, an upfront cost of engaging in tax planning.
And if that’s correct then that means that there will be many multinationals that are relatively insensitive to tax differentials, along with others that are. And so this is a way to understand that heterogeneity. But I think it's something that needs a lot more research.
In turn, that could suggest that there are quite significant barriers to profit shifting. Would you say that’s perhaps a correct way of looking at it?
Yes. Well, that would be another way to express a similar idea, that there are some barriers to profit shifting that are probably more important for smaller firms than larger ones. But understanding the nature of those barriers is important to formulating policy, and it's something that empirical researchers can perhaps benefit from learning from practitioners about.
From a researcher's perspective, what major developments on our understanding of profit shifting do you expect in the next 1-2 years?
From a researcher's perspective, it’s always good to have changes in the underlying set of rules because that gives us additional variation to test the impact of those rules on firms' behavior. I would welcome reforms from that perspective.
Given that the global economy is becoming more integrated, knowledge-intensive, and intangibles-based, how do you envision the future of the corporate tax base?
I think the first point to bear in mind is that the corporate income tax isn’t a very large source of revenue in most OECD countries.
However, that does not seem to be attributable to any recent increase in profit shifting activity because that has been the case for a long time. But, on the other hand… I recognize the force of the phenomena that you describe.
But these developments have been going on now for a while. And the corporate tax continues to play a robust, albeit small role in revenue raising for most countries. So I would be hesitant to predict that it would change substantially.
What are some of your own favorite papers or resources related to profit shifting?
There are a few papers that I would mention. One is the Hines and Rice paper from 1994, which in many ways has established a framework and methodology that researchers continue to use. So in that sense it has been very important.
The paper by Huizinga and Laeven, in 2008, was among the first to use affiliate-level data. And that's been an important advance.
I think among current working papers that I have seen, the paper by Kevin Markle is important because it very directly addresses the policy-relevant question of what happens when countries switch from worldwide to territorial systems.
I also think the Lohse and Riedel paper is of great interest, because it shows how the growth of transfer pricing regulations has limited profit shifting.
What’s next in terms of your contributions to a better understanding of profit shifting?
It's certainly something I hope to keep working on. I don't have any new results for the moment directly on these issues. But I do hope to keep working on those. It appears that policy makers will provide us with new sources of variation.