Abstract
This study investigates tax-induced profit shifting and the effect of the availability of Advance Pricing Agreement (APA) rules on the shifting behaviour of firms. We provide a theoretical model which demonstrates that an APA eliminates the transfer pricing uncertainty related to the calculation of the arm’s length parameter, and it eliminates the incidence of a tax penalization rate. In this case, APA may be attractive for firms focusing on a profit shifting strategy, as it reduces the implicit costs related to the expected tax amendment to be imposed by tax authorities. On an empirical application of our model, we find that firms have an incremental volume of intrafirm transactions with related parties located in low-tax countries with available APA, which is consistent with our theoretical prediction.
Similar content being viewed by others
Notes
A recent study from Becker, Davies and Jakobs (2017) shows that the implementation of an APA solves a hold-up problem from a time inconsistent tax authority, as it increases firm’s activity. As a consequence, higher activity may indeed lead to an increase in profit shifting to low-tax countries.
The modelling of the firm’s revenues Ri(si) and costs Ci(si) as a function of sales si \(\in\) R+ is borrowed from the traditional study of Kant (1988). The amount of sales si \(\in\) R+ is theoretically a non-negative measure that may be modelled as a function of exogenous market factors and endogenous production factors.
We allow for any of the two countries i = {1,2} to be the high-tax country, for generalization of the model. The direction of the intrafirm transaction can be reversed with no prejudice to our analysis.
Any coordination between Countries 1 and 2 naturally implies the imposition of the same parameter price \(\overline{x }\), therefore our model is directly extended to this case, i.e. a simultaneous tax amendment imposed by the non-harmed Country j implies a reduction on the expected tax amendment cost. In our baseline analysis, we assume the non-harmed Country j does not impose any tax amendment because it will be strategically better-off by keeping the excess tax revenues to itself, instead of imposing the parameter \(\overline{x }\) so to share the tax revenues with the harmed Country i.
The transfer pricing deviation x – \(\overline{x }\) is a directional measure with respect to LTP and HTP cases. For a parameter price \(\overline{x }\) adopted by both countries, the tax amendment cost is defined as sgn(T) · (x –\(\overline{x }\))mτi ≥ 0, where sgn( ·): R \(\to\) {− 1,1} is the sign function. The sign of the tax differential T indicates whether the price deviation x – \(\overline{x }\) implies a LTP or HTP case, therefore it indicates which Country i is harmed. As a result, the tax amendment cost is always a non-negative amount, i.e. the sign equality sgn(x –\(\overline{x }\)) = sgn(T) is a necessary condition for profit maximisation, therefore we apply the simplification (x –\(\overline{x }\)) · sgn(T) =|x – \(\overline{x }\)|. The maximising condition sgn(x –\(\overline{x }\)) = sgn(T) may not necessarily apply if the parameter price \(\overline{x }\) is not regarded consistently by both countries. See Sect. 2.2.
For generalisation, we include the impact of the tax penalty rate zi ≥ 0 on the tax amendment cost within our model. To apply the analysis on cases where the mispricing x – \(\overline{x }\) does not incur in a tax penalty, set the penalty rate equal to zi = 0, with no further implications to our analysis.
Formally, we assume that the probability function pi defined on the support R+ has no modes at the boundary of the support, such that arg (pi’’ = 0) \(\in\) (0,∞) (Rathke, 2021).
The sign sgn(T) = sgn(x* – \(\overline{x }\)) provides relevant information about the maximizing direction of the solution, i.e. T < 0 implies LTP case, i = 1; T > 0 implies HTP case, i = 2.
It means that the maxima exists iff the mode of the probability function pi is not zero.
We assume a discrete set of comparable prices {x}, as we understand that this structure is most properly aligned with the real cases faced by firms and governments. Nonetheless, this set can also be defined as a continuum of prices on the non-negative reals R+, for the random parameter price X: R+ \(\to\) R+ is still a mapping to the non-negative real line, with no further implications to our analysis.
For the realisation of the random arm’s length parameter X so that we have sgn(x – X) = – sgn(T), the sign of the price deviation sgn(x – X) still indicates which one is the harmed Country i, i.e. the profit shifting formally becomes a random event depending on the realisation of X.
For generalization, we assume that both Countries 1 and 2 charge APA fees bi which are deductible for tax purposes. We may assume that APA fees are non-deductible, or that only Country i charges the APA fee, with no further implications to our analysis.
Remark that the APA indicator equal to Iy = 1 implies that the random arm’s length parameter X must be evaluated with respect to the APA rules y. On this account, Eq. (15) has an abuse of notation, for an extended specification of the generalized net expected profit of the MNE is equal to.
$$E\left(\Pi |\left\{x\right\},a\right)=\Pi -{p}_{i}\cdot \left|x-X\right|m{\tau }_{i}\left(1+{z}_{i}\right)$$$$+ {I}_{y}\cdot \left[{p}_{i}\cdot \left|x-X\right|m{\tau }_{i}\left(1+{z}_{i}\right)-{p}_{iy}\cdot \left|x-{X}_{y}\right|m{\tau }_{i}-bxm\right],$$where the variable Xy states explicitly the position where the random arm’s length X is evaluated under the APA rules y, and we have piy = pi(|x – Xy|m).
Equation (18) presents the difference operation (also called discrete differentiation) of the expected net profit in Eq. (16) with respect to the indicator variable Iy|a equal to.
$$\frac{\Delta E\left(\Pi |\left\{x\right\},a\right)}{\Delta {I}_{y|a}}=\frac{E\left(\Pi |\left\{x\right\},{I}_{y|a}=1\right)-E\left(\Pi |\left\{x\right\},{I}_{y|a}=0\right)}{\left({I}_{y|a}=1\right)-\left({I}_{y|a}=0\right)}=\frac{E\left(\Pi |\left\{x\right\},Y=y\right)-E\left(\Pi |\left\{x\right\}\right)}{1-0}.$$Equation (20) refers to the Law of Total Variance. Because of the structure of the sum of expected conditional variances plus variance of conditional expectations, it is also known as Eve’s Law.
Full derivation of Eq. (31) in the Appendix.
In a similar specification, Rathke (2020) investigates the impact of tax havens on profit shifting in Brazil. Our empirical model is not absolutely original. It derives from the baseline approach from Hines and Rice (1994), which is applied by virtually all papers on this subject (Beer, de Mooji & Liu, 2018; Dharmapala, 2014).
Due to missing information, our empirical model does not include the effect of APA fee.
Discounting the regularization and log transformation of variables, the semi-elasticity is equal to.
$$\frac{d{x}^{*}{m}^{*}}{{x}^{*}{m}^{*}+R}=\gamma \cdot dT$$where R is the total net revenues.
Discounting the regularization and log transformation of variables, the semi-elasticity is equal to.
$$\frac{\Delta {x}^{*}{m}^{*}}{{x}^{*}{m}^{*}+R}=\mu \cdot \Delta {I}_{a}$$where R is the total net revenues.
Economic and financial literature call it “earnings management” practices. Studies show that firms are able to misapply accounting rules and to create accounting accruals in order to adjust their reported earnings, according to specific management’s incentives. The predominant earnings-management practices in literature include income smoothing, target beating, loss avoidance, “big-bath” accounting, and “cookie-jar” accounting (Barth et al. 2008).
This expected gain may be obtained even if the harmed Country i is the low-tax country; the low-tax Country may be the harmed Country i imposing the costly tax amendment because the parameter price X is random. See Sect. 2.2.
The indicator function IA: R \(\to\) {0,1} can be generalized for any property, usually defined as a set inclusion, for the function is equal to one IA(x) = 1 if we have x \(\in\) A, and zero otherwise. The indicator function for an identity relation I(x = y) = 1 is also known as the Kronecker delta function δxy satisfying similar properties, such that we have δxy = 1 if x = y and zero otherwise. The indicator function IA satisfy the multiplicative property IA · IA = IA, therefore the rational property IA = IA \(\to\) IA/IA = 1.
Recall that sgn(T) · sgn(T) = sgn(T)/sgn(T) = 1.
References
Astorino PS (2015) Consequências das conexões políticas para as empresas de capital aberto no Brasil: desempenho e acesso a crédito do BNDES (Doctoral dissertation, Universidade de São Paulo).
Ball R, Robin A, Wu JS (2003) Incentives versus standards: properties of accounting income in four East Asian countries. J Account Econ 36(1–3):235–270
Barth ME, Landsman WR, Lang MH (2008) International accounting standards and accounting quality. J Account Res 46(3):467–498
Becker J, Davies RB, Jakobs G (2017) The economics of advance pricing agreements. J Econ Behav Organ 134:255–268
Beer S, de Mooij R, Liu L (2018) International corporate tax avoidance: a review of the channels, magnitudes, and blind spots. J Econ Sur. 34(3):660
Bernard AB, Jensen JB, Schott PK (2006) Transfer pricing by US-based multinational firms. NBER Working Paper No. 12493, Massaschussets.
Borkowski SC (1996) Advance pricing (dis) agreements: differences in tax authority and transnational corporation opinions. Intern Tax J 22:23
Bunn D (2018). Corporate Income Tax Rates around the World, 2018. Tax Foundation, Fiscal Fact Report no. 623, November.
Clausing KA (2003) Tax-motivated transfer pricing and US intrafirm trade prices. J Public Econ 87(9–10):2207–2223
Cristea AD, Nguyen DX (2016) Transfer pricing by multinational firms: new evidence from foreign firm ownerships. Am Econ J Econ Pol 8(3):170–202
Davies RB, Martin J, Parenti M, Toubal F (2018) Knocking on tax haven’s door: Multinational firms and transfer pricing. Rev Econ Stat 100(1):120–134
De Waegenaere A, Sansing R, Wielhouwer JL (2007) Using bilateral advance pricing agreements to resolve tax transfer pricing disputes. Nat Tax J. 60:173–191
Dharmapala D (2014) What do we know about base erosion and profit shifting? A review of the empirical literature. Fisc Stud 35(4):421–448
Diller M, Kortebusch P, Schneider G, Sureth-Sloane C (2017) Boon or bane? Advance tax rulings as a measure to mitigate tax uncertainty and foster investment. European Accounting Review 26(3):441–468
Eden L, Byrnes W (2018) Transfer pricing and state aid: the unintended consequences of advance pricing agreements. Transnat Corp J. 25(2):9
EY (2019) 2019 Transfer Pricing and International Tax Survey. Ernst & Young Global Limited.
Givati Y (2009) Resolving legal uncertainty: the fulfilled promise of advance tax rulings. Virginia Tax Rev 29:137
Grubert H, Mutti J (2000) Do taxes influence where US corporations invest? Natl Tax J 53(4):825–839
Hines JR Jr, Rice EM (1994) Fiscal paradise: foreign tax havens and American business. Q J Econ 109(1):149–182
Horst T (1971) The theory of the multinational firm: optimal behavior under different tariff and tax rates. J Polit Econ 79(5):1059–1072
Itagaki T (1979) Theory of the multinational firm: an analysis of effects of government policies. Intern Econ Rev. 20:437–448
Jacob M (2018) A note on tax research. Revista Contabilidade Finanças 29(78):339–342
Kant C (1988) Endogenous transfer pricing and the effects of uncertain regulation. J Int Econ 24(1–2):147–157
Lazzarini SG, Musacchio A, Bandeira-de-Mello R, Marcon R (2015) What do state-owned development banks do? Evidence from BNDES, 2002–09. World Dev 66:237–253
Lohse T, Riedel N (2013) Do Transfer Pricing Laws Limit International Income Shifting? Evidence from European Multinationals. CESifo Working Paper 4404.
PWC (2016). International Transfer Pricing 2015/16. PricewaterhouseCoopers.
OECD (2017) Transfer pricing guidelines for multinational enterprises and tax administrations. OECD, Paris
OECD (2019) Transfer pricing in brazil: towards convergence with the OECD standard. OECD, Paris
Ortmann R, Simons D, Voeller D (2021) Real effects of an international tax reform for MNEs. arqus Discussion Paper no. 265.
Rathke AAT (2020) Profit shifting in Brazil and the impact of tax havens. Revista Contabilidade Finanças 32:95–108
Rathke AAT (2021) Taxpayer deductions and the endogenous probability of tax penalisation. arXiv Working Paper no. 2202:13695.
Riedel N (2018) Quantifying international tax avoidance: a review of the academic literature. Rev Econ 69(2):169–181
Samuelson L (1982) The multinational firm with arm’s length transfer price limits. J Int Econ 13(3–4):365–374
Schindler D, Schjelderup G (2013) Profit Shifting and Corporate Profit Tax Evasion. In:8th Norwegian-German Seminar on Public Economies.
Tomohara A (2004) Inefficiencies of bilateral advanced pricing agreements (BAPA) in taxing multinational companies. Nat Tax J. 57(4):863–873
World Bank Group, PWC. (2020) Paying Taxes 2020. The World Bank Group and PricewaterhouseCoopers.
Yitzhaki S (1974) A note on income tax evasion: a theoretical analysis. J Public Econ 3:201–202
Funding
The author have not disclosed any funding.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
We confirm the absence of conflict of interests. We confirm that there is no funding or granting source associated to this study. Data may be available upon request, on the authors’ best assessment. We confirm that this study is original and is not submitted to another publication source. We confirm to cope with the Ethical Standards of the Journal of Business Economics.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix
Appendix
1.1 Definitions and equivalent relations
For this Appendix Section only, we assume the general real-valued functions f,g: R \(\to\) R with respect to any real variables x,y,z \(\in\) R
General product rule is ∂(fg)/∂x = f’g + fg’, and the general chain rule is ∂f[g(x)]/∂x = f’(g) · g’
The absolute value function | · |: R \(\to\) R+ maps any non-zero real variable to its absolute value, with the definition |0|= 0. The sign function sgn ( ·): R \(\to\) {− 1, 1} maps any non-zero real variable to its sign, with the definition sgn(0) = 0. The sign function has a multiplicative identity property such that sgn(x) = 1/sgn(x). Any real number x is a combination of an absolute value and a sign, for we have
The derivative of the sign function is defined for any non-zero real number and it is equal to zero, ∂sgn(x)/∂x = 0, since ∂1/∂x = 0, ∂(− 1)/∂x = 0. Using the product rule, the derivative of the absolute value function |x| is equal to the sign function sgn(x), for we have
For any difference x – y, we clearly have ∂(x – y)/∂x = ∂x/∂x = 1, ∂(x – y)/∂y = ∂(− y)/∂y = − 1, therefore the derivative of the difference x – y with respect to either x or y is equal to the sign of the argument \(\in\)within the difference, i.e. for any value z \(\in\) {x,y}, x ≠ y, we have
The derivative of the absolute value of the difference |x – y| with respect to the max value between x > 0, y > 0 is intuitively equal to 1. From Eqs. (A1)−(A3) and using the chain rule, we can show that this is the case, for we have
Results in Eqs. (A3) and (A4) have equivalent notations using the indicator function defined for real variables as I: R \(\to\) {0, 1}, where the indicator is equal to one IA(x) = 1 if the real variable x satisfies some general property A assigned to the function IA, and zero otherwise.Footnote 28 Eq. (A3) is equivalent to
The indicator function I is a jump function, for the derivative of I(x) with respect to x is not defined in the continuous case. The usual convention is to define ∂I/∂x = 0, x \(\in\) R. Nonetheless, the indicator function I is still useful for the specification of the derivative of a general real-valued function f with respect to some real variable x, regarding whether f is a function of x. For a general f(z), the derivative with respect to x is equal to
where we have I(z = x) = 1 and zero otherwise, i.e. Eq. (A6) is clearly equal to zero if f is independent of x, z ≠ x
On the other hand, it makes sense to refer to the difference operation of some function f(I) with respect to the indicator I, which is also called discrete differentiation. From the definition of the derivative in the usual sense, the discrete differentiation of f(I) with respect to I is equal to
The derivatives of the absolute value function and the sign function as in Eqs. (A2)−(A4) are applied directly for the solution in Sect. 2, Eqs. (5) and (6). The use of indicator functions I as in Eqs. (A5) and (A6) is applied directly for the solution in Sect. 2, Eqs. (9) and (10). The discrete differentiation defined in Eq. (A7) is applied directly for the solution in Sect. 2, Eqs. (18), (26) and (27)
1.2 Derivation of the key equations for the research hypotheses H1 and H2
We present the full derivation of the key equations in Sect. 2, which are the basis of our empirical hypotheses H1 and H2. Variables and equations used in this Appendix Section are the same as in Sect. 2, and all definitions are presented throughout that same Sect. 2
Let x* and m* be the maximum transfer price and the maximum internal output respectively, so they define the implicit functions equal to x* = x*(T) and m* = m*(T). The maxima levels in Eqs. (7) and (8) are equal to
where the maxima imply the sign equality sgn(T) = sgn(x* – \(\overline{x }\)) for any LTP or HTP cases. Differentiating Eq. (A9) with respect to either one of the tax rates τi, τj and solving for ∂m*/∂T, we have
where we assume that both I1 = 1 and I2 = 1 do not occur simultaneously, therefore τi = τ1 implies ∂T/∂τi = -1, τi = τ2 implies ∂T/∂τi = 1, ∂T/∂τi = I2 – I1 = sgn(T). Eq. (A10) refers to Eq. (9) in Sect. 2.
Differentiating Eq. (A8) with respect to either one of the tax rates τi, τj, we find an intricate expression equal to
Notice that for any non-negative value m* ≥ 0, the absolute value function implies |x* – \(\overline{x }\)|m* =|x*m* – \(\overline{x }\) m*|. Simplifying the impact of sgn(T)Footnote 29 and substituting Eq. (A8) into Eq. (A11), we obtain
Solving Eq. (A12) for ∂x*m*/∂T, we finally have
which refers to Eq. (10) in Sect. 2, with H* = τi(1 + zi)[2pi*’ + pi*’’|x* – \(\overline{x }\)|m*].
Now, let x* and m* be maxima, so they define the implicit functions equal to x* = x*(Ia) and m* = m*(Ia) respectively. The maxima levels in Eqs. (16) and (17) are equal to
where the maxima imply the sign equality sgn(T) = sgn(x* – X) for any LTP or HTP cases. Differentiating Eq. (A15) with respect to Ia and solving for ∆m*/∆Ia, we have
where the variable Xy states explicitly the position where the random arm’s length parameter X is evaluated under the APA rules y. Equation (A16) is always non-positive since we have Xy ≥ 0, b ≥ 0, C’’ ≥ 0. In Eqs. (A16), the random arm’s length price Xy is clearly evaluated under the APA rules y, since the complete equation is equal to zero if the MNE chooses for no APA, Iy|a = 0. Eqs. (A16) refers to Eqs. (26) in Sect. 2.
Now for the Eqs. (A14), the full expression of the maximum is equal to
where we have piy = pi(|x – Xy|m), and we use the indicator function Iy for the complete extended notation. Differentiating Eqs. (A17) with respect to Ia, we obtain the cumbersome expression
We go for the tedious solution of Eqs. (A18) with respect to ∆x*m*/∆Ia. First, we simplify the square of the sign sgn(T) · sgn(T) = 1, we offset the first and the last terms of the right-hand side of Eqs. (A18), and we isolate the terms sharing the discrete differential ∆Ia, so we obtain the equality
where the term X + Iy|a · (Xy – X) indicates explicitly the impact of the APA choice Iy|a over the random arm’s length parameter X for the output variation ∆m*/∆Ia. From the solution of Eqs. (A16), we obtain the simplification
which means that the random arm’s length parameter Xy will always be evaluated under the APA rules y if the MNE chooses for the APA.
Substitute Eqs. (A20) into Eqs. (A19) and apply the notation
so we obtain
Solving Eqs. (A22) for ∆x*m*/∆Ia, with the rational property of the indicator such that Iy|a/Iy|a = 1, we finally have
which refers to the Eqs. (27) in Sect. 2.
1.3 Derivation of Eqs. ( 31 )
Variables and equations used in this Appendix Section are the same as in Sect. 2, and all definitions are presented throughout that same Sect. 2.
The variation in probability p∆* defined in Eqs. (28) is equal to
Expand Eqs. (A24) in a second-order Taylor series with respect to X around the mean E(X), for the full expansion is equal to
Let the penalty rate be zero zi = 0, and assume the conditional mean E(X|y) is such that it produces the approximation pi ≈ piy within Eqs. (A25), while it preserves the sign equality sgn[x* – E(X)] = sgn[x* – E(X|y)]. We obtain
with the temporary notation for the second-order Taylor coefficients equal to
Finally, assume the approximation E(X) ≈ E(X|y) within Eqs. (A26), so the absolute difference |E(X) – E(X|y)| ≈ 0 is approximate to zero. We obtain
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Rathke, A.A.T., Rezende, A.J., Watrin, C. et al. Profit shifting and the attractiveness of Advance Pricing Agreements. J Bus Econ 93, 817–857 (2023). https://doi.org/10.1007/s11573-022-01125-5
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11573-022-01125-5
Keywords
- Profit shifting
- Transfer pricing
- Advance Pricing Agreement (APA)
- Base Erosion and Profit Shifting (BEPS)