Profit shifting and tax‐rate uncertainty

• Author: Panagiotis I. Karavitis,Iftekhar Hasan,Manthos D. Delis
• Published date: 01 May 2020
• DOI: http://doi.org/10.1111/jbfa.12429
• Date: 01 May 2020

DOI: 10.1111/jbfa.12429

Profit shifting and tax-rate uncertainty

Manthos D. Delis1Iftekhar Hasan2,3 Panagiotis I. Karavitis4

1Montpellier Business School, Montpellier,

Occitanie, France

2Gabelli School of Business, FordhamUniversity,

New York,NY 10023, USA

3Bank of Finland, FI-00101 Helsinki, Finland

4Adam Smith Business School, University of

Glasgow, Glasgow, UK

Correspondence

IftekharHasan, Gabelli School of Business,

FordhamUniversity, New York,NY 10023, USA.

Email:ihasan@fordham.edu.

Fundinginformation

AustralianResearch Council, Grant/Award Num-

ber:DP170101413

Abstract

Using firm-level data for 1,084 parent firms in 24 countries and

for 9,497 subsidiaries in 54 countries, we show that tax-motivated

profit shifting is larger among subsidiaries in countries that have

stable corporate tax rates over time. Our findings further suggest

that firms move away from transfer pricing and toward intragroup

debt shifting that has lower adjustment costs. Our results are

robust to several identification methods and respecifications, and

they highlight the important role of tax-rate uncertainty in the

profit-shifting decision while pointing to an adjustment away from

more costly transfer pricing and toward debt shifting.

KEYWORDS

debt shifting, international taxation, multinational firms, profit

shifting, taxation uncertainty,transfer pricing

JEL CLASSIFICATION

F23, H25, H32, M41, M48

1INTRODUCTION

There is a growing perception, informed by empirical evidence, that governments lose a significant amount of tax rev-

enue to profit shifting within multinational enterprises (MNEs). Media reports provide multiple examples of alleged

profit shifting among large, well-established corporations that transfer income from high-tax-rate countries to low-

tax-rate countries in order to increase after-tax profits. The importance of this issue is highlighted in the 2012 Base

Erosion and Profit Shifting (BEPS) project by the Organisation for Economic Co-operation and Development (OECD)

and the G20 countries to fight profit shifting.

In this study, we examine how tax-rate uncertainty in MNE host countries affects firms’ profit-shifting behavior.

Tax-motivated profit (orincome) shifting involves two mechanisms that move income from high-tax jurisdictions to

low-tax jurisdictions: transfer pricing and international debt shifting. Transferpricing involves manipulating the prices

of transactions within the MNE group. For example, to lower taxes, an MNE might charge artificially low prices for

goods sold to a subsidiary in a low-tax country. Debt shifting (also called earnings stripping via interest in the United

States) occurs when a group’s companyin a high-tax jurisdiction borrows from a group’s lender in a low-tax jurisdiction.

The borrower pays interest and deducts that interest in the high-tax jurisdiction, and the lender receives interest and

recognizes taxable income in the low-tax jurisdiction.

J Bus Fin Acc. 2020;47:645–676. wileyonlinelibrary.com/journal/jbfa c

2019 John Wiley & Sons Ltd 645

646 DELIS ET AL.

We first hypothesizethat profit-shifting activity is higher for MNE subsidiaries in low-tax countries with stable cor-

porate tax rates(countries with low tax uncertainty) compared with those in low-tax countries that frequently change

their corporate tax rates (countries with high tax uncertainty). The reasoning is that low-tax jurisdictions with stable

tax rates increase the probability that an MNE benefits from profit-shifting activity.Empirical analysis of this hypothe-

sis is important to better identify countries that serve as hosts to profit-shifting flows and to determine the volume of

these flows within such countries.

Second, given that higher tax-rate uncertainty implies additional costs for MNEs, we hypothesizethat MNEs adopt

tax-planning strategies that have relatively low costs. Hence, we hypothesize that when tax-rate uncertainty is high,

MNEs shift away from costly forms of profit shifting (such as transfer pricing or relocation of intangible assets) and

toward intra-group debt arrangements.

We use a panel dataset with a maximum of 1,084 parent firms from 24 countries and 9,497 subsidiaries in

54 countries from 2010 through 2013. We limit our analysis to the post-global financial crisis period to avoid con-

taminating our findings with developmentsduring that period.1We measure tax-rate uncertainty using either changes

in corporate tax ratesor the volatility in those rates in subsidiary countries.

Our main empirical identification method builds on the difference-in-differences (DID) model introduced by

Dharmapala and Riedel (2013). This model identifies the aggressivenessof profit shifting by examining earnings shocks

at the parent-company leveland their propagation toward subsidiaries. Specifically, the main premise is that an exoge-

nous increase in a parent company’s profits implies increased profit shifting to subsidiaries in low-tax-rate countries.

We measure exogenous shocks to the parent companyvia a variable based on the pretax profits of other firms in the

same industry and country (Bertrand, Mehta, & Mullainathan, 2002). Toensure that we examine only true exogenous

shocks, we restrict our empirical analysis to subsidiaries in industries (and countries) other than those of their parent

companies.

We split our sample between subsidiary countries that changed their corporate tax ratesduring the previous three

or four years(countries with high tax-rate uncertainty) and those that did not (countries with low tax-rate uncertainty).

Our findings indicate significant profit shifting toward subsidiaries in countries with low tax uncertainty and insignifi-

cant profit shifting toward subsidiaries in countries with high tax uncertainty.Specifically, for parent firms experienc-

ing an earnings shock of 10%, their subsidiaries in countries with low tax-rate uncertainty report an approximately

0.92% increase in earnings before taxes compared to countries with high tax-rateuncertainty. Moreover, our estimate

ofprofit shifting toward countries with low tax-rate uncertainty is more than twice the equivalent effect in Dharmapala

and Riedel (2013), who do not examine the effects oftax-rate uncertainty.

This baseline finding is robust to the identification of profit shifting using the so-called tax-differential approach in

Hines and Rice (1994) and Huizinga and Laeven(2008) instead of the DID method. Furthermore, profit shifting is more

aggressive when we measure tax-rate uncertainty using corporate tax-ratevolatility or when using a forward-looking

measuresuch as the Economic Policy Uncertainty Index (Baker, Bloom, & Davis,2015). We also show that profit shifting

is weak in countries with high tax-rate uncertainty even when tax rates decrease. Our explanationfor this seemingly

counterintuitivefinding is that frequent tax changes in a subsidiary country make inaccurate predictions of tax benefits

more likely and/orraise the possibility of an imminent tax-rate increase.

Importantly,when tax rates are uncertain, firms conduct profit shifting using methods they can adjust quickly and at

low cost. In line with Dyreng and Markle (2016), we argue that transfer pricing is more costly because of severalnon-

negligible fixed and variable costs. For example,a company engaging in transfer pricing needs to invest in tax experts

(McGuire, Omer,& Wang, 2012) who in turn require additional administrative employees, whichsignificantly increases

the associated expenses.Further, transfer pricing entails compliance costs, as countries negotiate and/orform bilateral

agreements for tax cooperation. This pushes companies to hire executives with ‘know how’ of the countries where

major low-tax subsidiaries operate (e.g., Masulis, Wang,& Xie, 2012). Firms also incur initial expenses for shared-cost

agreements, which are related to the relocation of intangibles.

1Infact, we show that profit shifting is significantly weaker during 2007–2009.

DELIS ET AL.647

Given the costs of transfer pricing, our results might be due to adjustments, whereby MNEs use less costly intra-

group debt shifting instead. Debt shifting is much less expensive in our setting, because when tax rates increase or

fluctuate, firms just repay their debt. Our findings are fully in line with the cost-adjustment findings in Dyreng and

Markle (2016), given the strong and persistent effect that tax-rate uncertainty in low-tax subsidiary countries has on

MNE intragroup debt shifting. This finding is also in line with recent literaturethat examines firms’ trade-offsbetween

tax benefits and costs (e.g., Hopland, Lisowsky, Mardan, & Schindler, 2018; Nicolay et al., 2016; Saunders-Scott,

2015).

Our results contribute to the profit-shifting literature in three ways. Primarily,our research is the first to examine

how tax-rate uncertainty affects firms’ profit-shifting behavior.We find that, on average, profit shifting is aggressive

only as long as tax-rate uncertainty in subsidiaries’ countries is low, reinforcing the view that efforts to reduce MNEs’

profit shifting must focus on countries with stable corporate tax rates. Second, our findings provide a new explana-

tion for the relatively low level of profit shifting in global samples without differentiation bycountry type. By focusing

on tax-rate uncertainty,we obtain a much clearer perspective on the location and timing of profit-shifting volumes. In

other words, once we generate a relatively level playing field in terms of tax-rateuncertainty in subsidiary countries,

we find profit-shifting volumes that are substantially higher than in the current literature. Third, our study adds to the

literature examining the substitutability of various tax-planning strategies to lower potential costs. We provide evi-

dence for a cost-adjustment mechanism, whereby MNEs respond to high tax-rate uncertainty by using intracompany

debt shifting rather than transferpricing, given that the former is easier and cheaper to adjust.

The remainder of this study is organized as follows. Section 2 summarizes existing literature and provides testable

hypotheses. Section 3 discusses our empirical strategy and presents the dataset. Section 4 discusses the empirical

findings, and section 5 presents our conclusions.

2SETTING THE CONTEXT

2.1 Summary of the related literature and contribution of our study

The empirical literature documents that MNEs engage in tax-motivated profit shifting to subsidiaries in low-tax juris-

dictions. In general, profit shifting includes transferpricing (i.e., intragroup transactions), debt shifting (i.e., transferring

intragroup debt), and intangibles (e.g., patents) relocation. In an influential study,Hines and Rice (1994) suggest that

total reported subsidiary income has two parts: ‘true’ income from production (labor and capital) and ‘shifted’ income

from profit-shifting activities. Thus, shifted income is not attributable to a subsidiary’s own resources.

Since Hines and Rice’s (1994) study,the literature has significantly advanced in terms of procedures used to identify

profit shifting empirically; we outline here only the studies most closely related to our research.2Huizinga and Laeven

(2008), for example, identify profit shifting by constructing a weighted tax difference that uses information for all of

a multinational group’s affiliates instead of the simple tax difference between parent and subsidiary firms in Hines

and Rice (1994). This procedure accounts for the possibility of shifting income from a high-tax subsidiary to a low-tax

subsidiary, rather than generally shifting income from parent companies to subsidiaries. The results provide strong

evidence of profit shifting. In turn, Dharmapala and Riedel (2013) identify profit shifting through exogenous industry

shocks on the parent firm’s earnings and the propagation of these shocks toward their subsidiaries (a DID model).3

A number of studies examinethe aggressiveness of profit shifting using these two methods. For example, Dischinger

and Riedel (2011) and Karkinsky and Riedel (2012) examine the tax-motivated shifting of intangible assets toward

2Fora thorough review of the literature and empirical identification methods, see Dharmapala (2014).

3Manyalternative approaches exist in the tax accounting literature. For example,Collins, Kemsley, and Lang (1998) use consolidated data and assume that the

pretaxrate of return on foreign sales is a function of the return on worldwide sales in the absence of income shifting. If, instead, the return on foreign sales is a

functionof tax incentives after controlling for the worldwide return on sales, then this is attributable to profit-shifting activity.