The Power of Boilerplate: Bilateralism, Plurilateralism, and the International Tax System

Abstract

How do states exercise power in complex, decentralized governance regimes? This article sheds light on this question by analyzing one of the pillars of the post-war economic order: the international tax regime. The taxation of multinational enterprises is governed by a system composed of thousands of Bilateral Tax Treaties signed between pairs of national governments. We argue that despite the bilateral nature of these treaties, their legal content is largely controlled by a small group of economically powerful governments. By drafting and promoting “boilerplate” legal language, and by leveraging network effects, OECD countries were able to build a coherent, encompassing, quasi-multilateral regime to govern the taxation of multinationals in most of the world. We test and find support for our arguments about the role of network effects and the power of boilerplate using inferential network analysis and an automated text analysis of a corpus of 3200 treaty texts.

RESUMEN

¿Cómo ejercen el poder aquellos Estados que tienen regímenes de gobernanza complejos y descentralizados? Este artículo arroja luz sobre esta cuestión analizando uno de los pilares del orden económico de la posguerra: el régimen fiscal internacional. La tributación de las empresas multinacionales se rige por un sistema compuesto por miles de Tratados Fiscales Bilaterales (BTT, por sus siglas en inglés), los cuales son firmados entre pares de Gobiernos nacionales. Argumentamos que, a pesar de la naturaleza bilateral de los BTT, su contenido legal está controlado, en gran medida, por un pequeño grupo de Gobiernos económicamente poderosos. Los países de la OCDE lograron crear, mediante la redacción y promoción de un lenguaje jurídico “estándar” y el aprovechamiento de los efectos de redes, un régimen coherente, incluyente y cuasi multilateral con el que poder regir la tributación de las multinacionales en la mayor parte del mundo. Probamos y encontramos apoyo para nuestros argumentos en materia del papel de los efectos de las redes y del poder del lenguaje estándar mediante el uso del análisis inferencial de redes y un análisis de texto automatizado de un corpus de 3200 textos procedentes de tratados.

RESUME

Comment les États exercent-ils leur pouvoir dans des régimes de gouvernance complexes et décentralisés? Cet article tente d’éclaircir cette question en analysant l’un des piliers de l’ordre économique d’après-guerre : le régime fiscal international. La fiscalité des multinationales est régie par un système composé de milliers de conventions fiscales bilatérales (CFB) signées par deux gouvernements nationaux. Nous affirmons que malgré la nature bilatérale des CFB, leur contenu légal est largement régi par un petit groupe de gouvernements puissants sur le plan économique. En rédigeant et en promouvant un langage juridique type, et en exploitant les effets de réseau, les pays de l’OCDE ont pu construire un régime quasi multilatéral cohérent et inclusif pour régir la fiscalité des multinationales dans la majorité du monde. Nous testons et trouvons des éléments pour venir étayer nos arguments quant au rôle des effets de réseau et au pouvoir du langage type à l’aide d’une analyse de réseau par inférence et d’une analyse de texte automatique d’un corpus de 3 200 textes de traité.

Keywords:

• international political economy
• international relations
• international multinationals
• oecd
• taxationtaxation

Acknowledgments

We thank Ida Bastiaens, Cristina Bodea, Matthew Castle, Rasmus Christensen, Julia Gray, Lukas Hakelberg, Daniel Hanson, Martin Hearson, Haillie Lee, Abraham Newman, Lauge Poulsen, Gijs Schumacher, and seminar participants at Université Laval, the 2017 European Political Science Association meeting, the 2017 American Political Science Association meeting, Georgetown University, Michigan State University, the University of Pennsylvania.

Notes

1 Hearson (2017a) is a notable exception and Barthel and Neumayer (2012) is the most direct precursor to our work. In the latter study, the authors use spatial econometric models to show that competition over foreign direct investment drives BTT signing. Our findings on tax competition are consistent with theirs. In our paper, we push the ball forward by focusing on the role of institutional membership in the OECD as a vector of policy diffusion, by taking a broader set of network effects into account, and by considering 20 more years of data. Moreover, we use a text-as-data approach to study the power of boilerplate, the diffusion of legal text, and patterns of legal convergence over time.

2 While no study appears to have specifically investigated the use of multiple imputation and bootstrapping in TERGM models, simulations in Schomaker and Heumann (2016) suggest this procedure yields accurate results.

3 Along the lines of Mansfield et al. (2002), democracies could be more likely to cooperate than semi-democracies, or autocracies.

4 ALB, ARM, AZE, BGR, BIH, BLR, CZE, EST, GEO, HRV, HUN, KAZ, KGZ, LTU, LVA, MDA, MKD, MNE, POL, ROU, RUS, RUS, SRB, SVK, SVN, TJK, TKM, UKR, UZB.

5 TERGM coefficient estimates can be interpreted in terms of log-odds.

6 Downloaded from [http://ibfd.org](http://ibfd.org) on 2016-06-02. We exclude treaties that do not cover corporate income tax. Most treaties are written in English, and there are usually English translations available for the treaties concluded in other languages. We count non-English treaties in our binary analysis, but assume that the very small subset of BTTs not published in English are not OECD-compliant.

7 There are exceptions. For instance, the US uses a very different model as a baseline in its own negotiations.