‘Earnings shocks and tax-motivated income-shifting: evidence from European multinationals’ – revisited

ABSTRACT

This article revisits the study by Dhammika Dharmapala and Nadine Riedel on income shifting between European multinationals published in the Journal of Public Economics in 2013. It used a promising alternative causal identification strategy for profit-shifting based on earnings shocks, which has over a short time period been frequently cited in the literature. Using data from the same database, albeit for a period 10 years later, for 2006–2015, the significant causal evidence for profit-shifting from European parent firms to their lower taxed European subsidiaries cannot be reproduced. Neither can similar results be obtained by considering profit-shifting to subsidiaries located anywhere in the world. Results in line with those of the original study can however be found when using effective rather than statutory corporate tax rates. While these findings raise concerns about the external validity of earlier studies’ result, it should not put the existence and extent of profit-shifting into question. It rather raises concerns about the focus on statutory tax rates in measuring profit-shifting. Moreover, rather than a decrease in profit-shifting, it may illustrate the database’s and the methodology’s limitations and ability to capture all shifts and channels.

KEYWORDS:

• Tax avoidance
• profit-shifting
• government tax revenue
• effective tax rate
• multinational firms

JEL CLASSIFICATION:

• H26
• H71
• H87

Acknowledgements

This article forms part of the UNU-WIDER work programme ‘The economics and politics of taxation and social protection’. Additionally, I would like to thank the Nordic Tax Research Council as well as the University of Tampere for their additional funding to work on this study. Moreover, I would like to thank Jukka Pirttilä as well as Niels Johannesen, Nadine Riedel, Supawan Saelim, Ludvig Wier and the participants of the UNU-WIDER Internal Seminar and the FDPE PhD Labour and Public Economics workshop for their suggestions and/or comments on this and earlier versions. All remaining errors are mine.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

¹ Such as Huizinga and Laeven (2008).

² Such as in Fuest et al (2015), Johannesen, Tørsløv, and Wier (2016), Delis, Hasan, and Karavitis (2016), OECD (2015).

³ Calculated excluding the respective firm itself.

⁴ Access through Hanken School of Economics (April 2015–November 2015) restricted to very large, large and medium-sized firms’ data.

⁵ EU-25-3 refers as in the original study to the EU25 member countries excluding Cyprus, Malta and Slovenia.

⁶ To avoid endogeneity of profit levels.

⁷ Depending on the proportion of comparable firms making losses or reporting negative assets, the number of estimation sample firms varies, as logarithmic transformations for predicted profits are only defined for positive values.

⁸ Firms might manage to shift all their profits and report zero profits.

⁹ This is assumed to be a good proxy over the whole period.

¹⁰ The reduction in mean statutory tax difference at the onset of the economic crisis in 2008, from 3.6% in 2006 to 2% in 2014, is relatively small and cannot be claimed to be a driver, as even only analysing profit-shifting during the pre-crisis period leads to insignificant results.

¹¹ When using statutory rates for the average effective tax rates sample, the earlier insignificant results persist.

¹² Such duplicates stem from the continuous updating of the Orbis database during the downloading period, so that a firm might have shifted their core business to a different industry and the parent of a group may have changed.

Additional information

Funding

I would like to thank UNU-WIDER, the Nordisk Skattevidenskabeligt Forskningsråd (in English: Nordic Tax Research Council), as well as the University of Tampere for their funding to work on this study.