Click on the above link to visit the KPMG Germany webpage on the Mondaq website
For disclaimer and copyright see end of this article.
At the time of the publication of the new 1995 OECD Guidelines, the German tax authorities were quick to issue a press statement declaring that "the Federal Republic of Germany considers the traditional methods to be adequate" and will not employ profit-oriented methods "except for purposes of verification [of another method] or estimation."
However, high officials in the German tax administration are known to have since had second thoughts on this matter. In a German language article (IStR 1999, 304) and again in the April 1999 issue of International Tax Review, Thomas Borstell and Michael Prick presented what they correctly identified as an important transfer pricing decision by the Düsseldorf Tax Court. It would appear to be the first reported court case in which the tax authorities have attempted to apply a comparable profits method.
Only excerpts of the decision have been published (IStR 1999, 311).
1. Facts, issues and grounds of the decision
The case involved the German marketing subsidiary of a foreign fashion clothes manufacturer. The subsidiary purchased mainly from related parties, but had also bought some of its goods from independent sources. It sold only to independent parties and reported an overall net loss for the 14 year period from 1980 to 1993.
Using the resale price method, the tax authorities claimed that the gross profit margin on goods purchased from related parties was too low and treated the income lost as a result of this deficiency as a constructive dividend. In order to determine what they saw as the appropriate gross profit margin, the tax officials relied on detailed comparables from the tax files of companies not party to the litigation (secret comparables) and to a lesser extent on the Betriebsprüfungskartei, a general body of economic data gathered by the tax authorities from audits throughout Germany. They also took some comparative data from public databases. In addition, the tax authorities also used the transactional net margin method (TNMM) to compare net margins as a check against their resale price method. Here they again relied on secret comparables.
The court did not follow the reasoning of the tax authorities and instead decided the case by comparing the gross margin which the subsidiary had earned on controlled transactions with the gross margin on its uncontrolled transactions. This approach may be viewed either as an "internal" comparable profits method based on a comparison of gross margins or as a resale price method using gross margin to calculate backwards from the uncontrolled selling price to the arm's length purchase price. It led to an upwards adjustment below that sought by the tax authorities.
Since the court compared the gross margins, not the prices, of the taxpayer's controlled and uncontrolled transactions, it is imprecise to characterise this approach as an "internal price comparison," as do Borstell and Prick in their article. However, these authors are certainly correct in finding that the "most remarkable aspect" of this decision is not the "internal" comparison of gross margins, although there is much to criticise in the court's handling of this part of the case.
2. TNMM and secret comparables
The most remarkable aspect of the case is the attempt by the tax authorities to corroborate their adjustments by means of the transactional net margin method (TNMM). Specifically, the tax authorities performed an external comparison of return on sales and return on assets using data taken from the tax files of allegedly comparable companies not party to the litigation. The tax authorities first used a public database to identify companies which might be comparable and then requested the confidential tax files of these companies from other tax offices. After a further elimination process, the tax authorities sought to present data taken from the confidential tax files of four firms to the court. The tax authorities were unable to obtain suitable comparative data from the external database itself.
The court rejected the "external" profit comparison by the tax authorities on two grounds:
Firstly, the court treated this method as permissible only under a provision in the international transactions tax act which permits estimation of the arm's length profit under certain circumstances (§ 1 (3) AStG). The court held that the requirements for such estimation were not met because an accurate result had been arrived at without estimation, namely using the "internal" profit comparison described above. Even if estimation were proper, the court implied that its "internal" profit comparison would be a superior type of estimation under the circumstances.
The court's reasoning here is dubious. Transactional methods, including the transactional net margin method, are not generally thought of as falling under § 1 (3) AStG.
Secondly – as Borstell and Prick stress – the court held that the data cited by the tax authorities in support of their comparison was inadmissible because it was taken from the tax files of taxpayers not party to the litigation (secret comparables). Direct introduction of the comparative profit information into evidence would entitle the taxpayer to examine its competitor's confidential data and cause the tax authorities and the court to violate the statutory prohibition on divulging information gained in the tax enforcement process. Revealing the information to the court alone would violate a variety of statutory and non-statutory due process considerations. Presentation of only the balance sheets in neutralised or anonymous form, so as to protect the identities of the firms involved, was likewise impermissible, since it was not possible to determine from the balance sheets alone whether the independent firms were truly comparable.
Furthermore, irrespective of the manner in which the non-public information was used, the court stated that it was impossible to determine whether the selection made by the tax authorities was truly representative, there being no way for either the court or the taxpayer to know whether data on other comparable firms had been withheld because it did not serve the tax authorities' case.
The court thus refused to consider confidential data taken from the tax files of taxpayers not party to the litigation. The court also stated that, even had it used the proffered evidence, it would have given the taxpayer the benefit of a certain "safety margin" to take account of the inherent uncertainty of such estimation procedures. The court did not address the use of data taken from public databases because the critical data presented to it was not of this nature.
3. Significance of the decision on secret comparables
In assessing the significance of the case, one must bear in mind that the litigation in question commenced in 1993. Great strides have since been made in compiling public quantitative data on German businesses. It is unclear whether the interim progress is sufficient to permit the tax authorities to assemble adequate comparable profits information from public sources. If not, it would appear to be only a matter of time – a year or two – before this is the case. Hence, even if the decision reported on is not reversed on appeal, it would only postpone the day when the tax authorities will have the – legally admissible – data they need to use comparable profits methods.
Borstell and Prick thus exaggerate when they write that, in Germany, there is "almost no available external data with respect to transactions or companies as a whole" (ITR April 1999, p. 11). Hence, their conclusion that the court's decision constitutes an effective bar to the use of comparable profit methods in Germany also goes too far.
It is true that the legal basis for the collection and publication of such data is not as good as that in, for instance, the United States. Practically speaking, however, Germany has far more independent middle-sized companies (companies which are not members of controlled groups) than does the United States.
Collection of data on these companies has progressed greatly in recent years. Even if data were to be assembled on only a fraction of all such independent companies, the absolute number of German independent companies on which data was available might well exceed the absolute number of independent companies on which data is available in the United States, assuming 100 % data availability there. Naturally, a database which included only a fraction of all German independent companies may not contain a representative sample. However, modern statistical sampling techniques often deliver reliable results using small percentage samples.
Borstell and Prick correctly note that the decision has implications for the anonymous taxpayer data compiled by the tax authorities on royalty fees and other transfer pricing issues (Lizenzgebührenkartei, Betriebsprüfungskartei). Adjustments of profits based solely on these non-public and hence non-verifiable sources are apparently incompatible with the court's holding, although the court does not address this issue specifically.
The decision is also interesting for its questionable treatment of the (external) transactional net margin method as a form of estimation subject to the material requirements of § 1 (3) AStG and for the view that, under this provision, the taxpayer is entitled to the benefit of a certain "safety margin" to take account of the inherent uncertainty of estimation procedures.
The tax authorities have petitioned the Federal Tax Court to hear an appeal from the lower court's decision. The Federal Tax Court has not yet ruled on this petition, but may be expected to grant it.
5. Research project: University of Freiburg and KPMG Germany
A three-year joint research project on the use of quantitative procedures, including the use of databases and quantitative comparable margin methods, for purposes of transfer pricing analysis was recently announced by the University of Freiburg and KPMG Germany. The results of this study are expected to have direct impact on the treatment of cases like that discussed here, which was approached by all concerned in a largely non-quantitative manner.
A slightly modified version of this article, authored by Alexander Vögele, first appeared in the July/August issue of International Tax Review.
For further information, please send a fax stating your inquiry to KPMG Frankfurt, attn. Christian Looks +49-(0)69-9587-2262 or an e-mail using the button appearing below. Please state your name and organisation in all inquiries.
Disclaimer and Copyright
This article treats the subjects covered in condensed form. It is intended to provide a general guide to the subject matter and should not be relied on as a basis for business decisions. Specialist advice must be sought with respect to your individual circumstances. We in particular insist that the tax law and other sources on which the article is based be consulted in the original, whether or not such sources are named in the article. Please note as well that later versions of this article or other articles on related topics may have since appeared on this database or elsewhere and should also be searched for and consulted. While our articles are carefully reviewed, we can accept no responsibility in the event of any inaccuracy or omission. Please note the date of each article and that subsequent related developments are not necessarily reported on in later articles. Any claims nevertheless raised on the basis of this article are subject to German substantive law and, to the extent permissible thereunder, to the exclusive jurisdiction of the courts in Frankfurt am Main, Germany. This article is the intellectual property of KPMG Deutsche Treuhand-Gesellschaft AG. Distribution to third persons is prohibited without our express written consent in advance.