As of 2008, Germany has severely changed its thin capitalization rule by introducing the so-called interest barrier. This new rule aims at prohibiting tax avoidance of multinational firms by means of cross-border internal loans. For reasons of non- discrimination, the rule is, however, equally attributable on the national level and it is applicable to both internal and external debt. Since its beginning, the German interest barrier has had a very poor reputation as it was believed to distort financing decisions and hereby harm production efficiency.

Four years after its introduction, the time has come to empirically evaluate the interest barrier. In this paper, we trace to what extent the interest barrier impacted firms' nancing decisions. We distinguish between national and multinational firms as well as between the effects on internal debt to assets and external debt to assets.

Thin capitalization rules prevent firms from deducting excessive interest expenses from their tax base. Before 2008, the interest on internal debt going beyond 1.5 times the equity of the respective shareholder was not deductible. As of 2008, interest payments exceeding the interest earnings are generally only deductible at the amount of 30% of EBITDA once the exemption limit of an initial EUR 1 million is exceeded. In our empirical setup, we identify firms which would have been affected by the new interest barrier, had it already been in place in the years 2005 to 2007, i.e. before its actual coming into force. Then we analyze empirically how these firms adjusted their debt to assets ratios and their net interest payments as compared to the control group.

Our regressions show that the interest barrier drove firms to lower their debt to assets ratios and their net interest payments. Opposing its original intention, it seems to be, however, also the national firms which adjusted their capital structure and it was external rather than internal debt which was reduced. Therefore, we conclude that the interest barrier does indeed affect nancing decisions, but predominantly not in the intended way and not of the intended firms. In sensitivity analyses, we examine highly leveraged and low protable firms, which are likely to be subject to the interest barrier. The results suggest a debt-reducing interest barrier effect for these companies as well.

Our empirical evidence does not provide a positive evaluation of the new interest barrier. The legislator might have focused too much on the task of counteracting excessive and abusive internal lending by a few multinationals, whilst disregarding the effects on non-abusive firms.