On March 27, 2020 the Court of Appeals Den Bosch rendered an interesting judgment about the recovery of VAT in a case concerning a financial institution. The Court ruled that the bank’s ‘actual use method’ had been sufficiently substantiated. Therefore, the recovery of VAT on mixed costs does not have to be determined on the basis of the standard turnover-based pro rata method, but may be based on actual use.
This case is not only relevant for financial institutions, but also for other taxpayers performing VAT-taxed and VAT-exempt services, such as parties in the public, education and healthcare sectors.
1. Background and points of law
The taxpayer in this case is a bank that performs both exempt and VAT-taxed services. All the costs incurred by the bank can be regarded as mixed costs. Most of its turnover consisted of VAT-exempt interest income and commission income, the majority of which was subject to VAT. The bank had transferred part of its mortgage receivables to separate securitization companies. The interest received on these receivables was passed on to the aforementioned companies by the bank.
The taxpayer had made a financial analysis of the profit and loss (P&L) of each product. With regard to the P&L per product, the costs are apportioned to the various product groups by means of three interval-based allocation formulas (based on time registration, actual products purchased and proportional distribution). This resulted in an allocation of the bank’s mixed costs to the various product categories.
In its VAT returns, the taxpayer recovered VAT on its mixed costs in accordance with the turnover-pro rata method without taking account of i) the interest expenses paid, ii) the interest received on its notes in the securitization companies and iii) the interest passed on to the securitization companies. In essence, the taxpayer wanted to use these proceedings to apply the recovery of VAT in accordance with actual use.
In dispute is firstly whether the actual use alleged and calculated by the taxpayer results in a more accurate calculation of the actual use. Secondly, whether the interest expenses paid can be deducted from the interest income received. Thirdly, whether the interest that was passed on to the securitization companies can be kept outside the taxpayer’s turnover.
2. Judgment by Court of Appeals Den Bosch
Many market parties have been eagerly awaiting this judgment. The judgment rendered by the District Court dated from May 9, 2018 and the hearing before the Court of Appeals had taken place on December 13, 2018. The Court of Appeals (hereafter: Court) has now finally rendered judgment in this case involving fundamental matters of principle, which essentially revolves around the question how strictly the VAT recovery based on actual use should be applied in the Netherlands. The Court found the taxpayer’s actual use method permissible.
With regard to the application of the actual use, the Court firstly noted that in determining the actual use some leeway must be allowed in respect of accuracy and that such a determination is not an exact science. In accordance with European case law, it is not about the most accurate result.
The Court agrees with the taxpayer, and moreover also with the District Court, that the taxpayer’s turnover does not accurately reflect the actual situation. For example, interest rates have fallen sharply since 2011, which resulted in a significant decline in the VAT-exempt interest income. The Court deemed it plausible that the extent to which mixed costs were used for the activities of the taxpayer has not kept pace with this decline. According to the Court, the taxpayer had thus convincingly demonstrated that a calculation based on actual use can result in a more accurate determination of the VAT recovery, provided the actual use can be determined on the basis of objective and precisely determined data.
The Court saw no reason to doubt the P&L per product and therefore regarded the allocation formulas and the cost allocation arising from the P&L per product as a basis for determining actual use. Since the submitted P&L per product was used to allocate all the mixed costs to the various turnover categories, which contain both taxed and exempt turnover, the P&L per product produced sufficiently objective and precisely determined data. Compared to the standard turnover-based pro rata method, that data resulted in a more accurate determination of the actual use and thus of the VAT recovery. The fact that assumptions and presumptions were sometimes used does not change this. After all, it is not about the most accurate result, but a more accurate result. According to the Court, the assumptions and presumptions used were also sufficiently objective. For example, for commercial reasons the Register Controllers prepared the P&L per product in accordance with international accounting principles and this was reviewed by the Dutch Central Bank (De Nederlandsche Bank; DNB) and the Dutch Authority for the Financial Markets (Autoriteit Financiële Markten; AFM) as part of their regulatory supervision, while the external auditor moreover approved the financial statements.
The Court therefore did not follow the tax inspector in his assertion that the taxpayer had not convincingly demonstrated the actual use. The Court also rejected the tax inspector’s objection against the application of the ratio per category. According to the Court, this is in accordance with the CJEU’s Morgen Stanley judgment, from which it can be inferred that the determination of the VAT recovery must be based on the various turnover categories for which the costs are used.
With regard to the second disputed point, the Court saw no reason to allow the interest expenses paid to be deducted from the interest income received. According to the Court, the inclusion of an interest margin, unlike in the CJEU’s Banco Mais case where there was a direct relationship between interest income and interest expenses, does not lead to a more accurate result than that obtained under the standard pro rata method.
With regard to the third disputed point, the Court ruled that the interest that was passed on to the securitization companies cannot be kept outside the taxpayer’s turnover. According to the Court, the specific circumstances of this case mean that the taxpayer should still be regarded as a lender.
3. Impact on Dutch practice
This will be a welcome judgment for many market parties. EU case law has long offered more scope for applying an actual use method instead of a standard turnover-pro rata method. Due to the strict interpretation of actual use applied in Dutch case law, as well as the Dutch tax authorities’ interpretation based on this case law, this scope could not actually be used. The basic principle for recovering VAT should however be that it is as realistic as possible, is as much as possible in line with use and thus does justice to tax neutrality.
The Court Den Bosch deemed the applied method to be based on sufficiently objective and precisely determined data, which – in comparison to the standard turnover-based pro rata method – results in a more accurate determination of the actual use and thus of the VAT recovery. The Court explicitly noted that within an actual use approach there is room to apply assumptions, since mixed-use costs inherently mean that these cannot be directly related to taxed or exempt turnover. Methods based on cost accounting, which are comparable to the method applied by the taxpayer, should also be acceptable. The Court rightly ruled that the Morgan Stanley judgment also supports a more specific approach to costs with regard to VAT recovery.
The Netherlands has not taken advantage of the possibility offered in the EU VAT Directive to specifically allow a VAT recovery method based on sectors. A number of other Member States do allow this option and in practice many foreign taxpayers use it. It appears that, by adopting actual use, the judgment by the Court leaves room to allow a method based on sectors, provided it is based on objective and precisely determined data and leads to a more accurate result.
4. What can you do now?
This judgment is of crucial importance for the application of the actual use method to mixed costs by taxpayers that perform both VAT-taxed and VAT-exempt services. Although it is not inconceivable that this judgment will be appealed before the Supreme Court, it does give financial, public, educational and healthcare institutions cause to take a closer look at their VAT recovery on mixed costs.
We recommend that you take a further look at whether, and if so, to what extent it would be worthwhile to examine in more detail whether there is an actual use method that leads to a more accurate determination of the actual use of mixed costs than the standard turnover-pro rata method. For example, cost accounting methods already used within your organization. We conclude from the judgment that the bar for objective and precisely determined data is lower than Dutch courts and the Dutch tax authorities previously envisaged. In that respect, it is advisable to look at robust reference points, such as international accounting principles and the manner in which regulatory authorities and the external auditor deal with the allocation of mixed costs.
If you would like to discuss this judgment, feel free to contact the advisors of Meijburg & Co’s Indirect Tax Financial Services Group or your usual advisor.