The Effect of Cross-Border Group Taxation on Ownership Chains

ABSTRACT

This study investigates the influence of taxation on ownership chains and specifically on the location decision for intermediate holding companies. By examining the effect of the introduction of a cross-border group taxation regime in Austria in 2005 on ownership chains of European multinational firms, I find evidence that foreign parent companies already invested in Austria restructured their ownership chains in order to meet the requirements of the group taxation regime. This effect is larger for foreign parent companies with loss-generating subsidiaries. Collectively, the empirical findings suggest that, when evaluating the effect of cross-border group taxation regimes, companies follow a detailed tax planning strategy that takes tax-base effects into account.

Keywords:

• Group taxation
• ownership chains
• intermediate subsidiaries
• Austria

Acknowledgments

I thank Martin Jacob (the editor), Jost Heckemeyer, Claire Estebanez, Peter Krenn, Martina Rechbauer and an anonymous reviewer for their helpful comments and suggestions. I also thank participants at the 2015 Meeting of the Business Taxation section of the German Academic Association for Business Research (VHB) in Passau, at the 2015 EAA Annual Congress in Glasgow and at the 2015 Workshop on Current Research in Taxation in Prague for helpful comments and suggestions on an earlier version of the paper.

Disclosure statement

No potential conflict of interest was reported by the author.

Supplemental Data and Research Materials

Supplemental data for this article can be accessed at https://doi.org/10.1080/09638180.2018.1564689.

Notes

1 A numerical example of the calculation of the timing effect can be found in an online Supplement.

2 Since the Austrian corporate tax rate applicable during that period was 25% this number is calculated as $3.1\cdot 0.25$.

3 This is the case if the parent company is located in a country that taxes corporate capital gains realized on the disposal of foreign subsidiaries. In my sample period, only 6 out of the 19 countries (Ireland, the Netherlands, Norway, Slovenia, Sweden and the UK) do not impose capital gains taxes on the disposal of foreign subsidiaries.

4 If however, the parent company is located in a country outside the EU or EEA area, the EU Merger Directive no longer applies and the question of whether the restructuring can be arranged net of taxes arises. Many non-EU countries allow for a tax free exchange of shares. For example, within the U.S. Internal Revenue Code 368(a)(1)(B), this is a tax free ‘B’ reorganization. Also, if the parent company is not an EU-company, dividends received from its subsidiaries might be subject to dividend taxation and withholding taxes on dividend payments may occur. In this case, any tax benefits of the regime may be outweighed by the tax costs associated with the restructuring.

5 In Austria, any contribution in shares requires a written contract according to the Austrian Corporate Restructuring Taxes Act, and this contract must be drawn up before a notary.

6 My main results do not change if I exclude parent companies from Germany, the Netherlands and Switzerland from my sample. Also, they do not change if I exclude parent companies from Denmark and Italy (both countries also offer a cross-border group taxation regime).

7 Alternatively, I could use the number of Austrian intermediate subsidiaries as my dependent variable. My results show that 97.43% of foreign companies with an Austrian intermediate subsidiary in my sample hold only one Austrian intermediate subsidiary and no foreign parent company holds more than two Austrian intermediate subsidiaries. I therefore do not run separate regressions on the number of Austrian intermediate subsidiaries.

8 Due to the way I construct my sample, parent companies already invested in Austria prior to 2005 could in principle add new subsidiaries in Austria and set them up as intermediate subsidiaries, even if the parent already has other subsidiaries in Austria. Those parent companies would be wrongly classified as shifting subsidiaries across existing chains. To address this problem, for all parent companies with an Austrian intermediate subsidiary after 2005, I cross-check whether the parent company has added new directly held subsidiaries in Austria after the introduction of the group taxation regime. Of the 1,295 parent companies already invested in Austria prior to 2005, I identify only 35 Austrian and 5 foreign parent companies that added new directly held subsidiaries in Austria after the introduction of the group taxation regime. Excluding these 40 parent companies from my sample does not change the results in Table 3.

9 In order to be included in the sample, a parent company needs to have an Austrian subsidiary in any of the years prior to 2005. The other alternative would be to require that a parent company has an Austrian subsidiary in year t before it is included in the sample for that year. This approach would allow me to estimate the probability of having an intermediate subsidiary in Austria, conditional on having a subsidiary in Austria, in each year. However, in such a setting, firms could enter the sample because of the reform. Narrowing down my sample to only those years in which a parent company has an Austrian subsidiary reduces my sample size to 3,631 observations. The DD estimator of the full logit model changes to 0.4027 and is significant at the 10% level.

10 I follow Equation (10) from Puhani (2012) and calculate the average marginal effect for the DD coefficient as the difference between the cumulative distribution function (CDF) of the logistic distribution evaluated at the estimated coefficients of the logit regression and the CDF of the logistic distribution with the same coefficients minus the DD interaction term. I set all co-variates at post-reform values. I cannot compute average marginal effects in Column (4), as $Foreig{n}_j$ and $Refor{m}_t$ are absorbed by the year and country fixed effects. I thus base the calculation of the average marginal effects in Column (4) on a specification without year and country fixed effects.

11 An example taken from my sample which shows how the parent company restructures its ownership chain and the description of the associated restructuring costs can be found in an online Supplement.

Additional information

Funding

Silke Rünger gratefully acknowledges financial support by the Austrian Science Fund (FWF) [P 22324-G11].