Sweetheart tax deals trending in the EU

Added 25 Feb 2016
The amazing world of ‘sweetheart tax deals’i  keeps surprising us on quite a regular basis. More than one year ago, the LuxLeaks revelations exposed the secret world of tax rulings that, in some cases, benefited multinational corporations with tax rates lower than 1%. Last October the European Commission, through an initiative by Commissioner for Competition Margrethe Vestager, announced that selective tax advantages issued through tax rulings for Fiat in Luxembourg and Starbucks in the Netherlands are illegal under EU State aid rules. In early 2016 Vestager announced that the same was the case for a Belgian tax ruling system called the excess profit exemption. Now, after analysing the official data on EU Advance Pricing Agreements’ (APA)ii  published a few months ago, the conclusion is, in a nutshell, that EU ‘sweetheart tax deals’ are becoming a trend in EU member states’ tax systems. 

The numbers released show the developments from 2013-2014, and the increase in the number of ‘sweetheart deals’ is dramatic. One important question, which cannot yet be answered, is whether the LuxLeaks scandal has had an impact. Due to a delay in the publication of the data, we do not yet have the numbers for 2015 (the year after LuxLeaks). However, despite an intense debate in the EU about ‘sweetheart deals’, there has been no indication that Member States intend to slow down when it comes to issuing these deals. Calls from the European Parliament to publish the key elements of these deals have also so far been rejected.

The technical term for the sweetheart deals that multinational corporations make with governments is ‘APA’. The definition is “an arrangement that determines, in advance of controlled transactions, an appropriate set of criteria (e.g. method, comparables and appropriate adjustments thereto, critical assumptions as to future events) for the determination of the transfer pricing for those transactions over a fixed period of time”iii . As a law professor described, “it’s like taking your tax plan to the government and getting it blessed ahead of time”. Except – it doesn’t apply to you and me. Only multinational corporations get these deals.

The statistics on APAs – released by the EU Joint Transfer Pricing Forum every year – are interesting and revealing for several reasons if we compare the 2013 and 2014 figures .iv 

Let’s start with 2013:
  • The total number of APAs in force at the end of 2013 was 547. 
  • The number of APAs with EU countries was 369.
  • The number of APAs with non-EU countries was 178. 
  • The EU average, meaning the number of APAs in force per country, at the end of the year was 19,53.
Now let’s see what happened in 2014:
  • At the end of 2014, the total number of APAs had grown to 972 – an increase of 77%. 
  • The number of APAs with EU countries was 698. 
  • The number of APAs with non-EU countries was 274. 
  • The EU average number of APAs in force per country at the end of the year was 34,71.
These figures can be better appreciated in the following graph, where we can also see the number of APAs per country.

Source: Eurodad graph compiled with data from European Commission 2015 and 2014 .

As illustrated in the graph above, the documents also break down the data by Member State. Some points are worth highlighting:
  • There are seven EU countries which, according to the statistics, still have not introduced any APAs: Bulgaria, Croatia, Cyprus, Estonia, Greece, Malta and Slovenia.
  • The case of the Netherlands is a special one. The official number of APAs is not included in the released statistics, but in response to a question from the European Parliament, the Netherlands released information showing that it had granted 228 APAs in 2013 and 203 in 2014. 
  • While the Netherlands was ‘top of the league’ in 2013, Luxembourg was number one in 2014. 
  • During 2014 Latvia and Lithuania started issuing APAs.
  • 14 out of 21 EU countries offering APAs at the end of 2014 increased their number from 2013. The cases of Belgium and Luxembourg are the two most prominent. In the case of Belgium, the number of APAs in force skyrocketed. 
  • The EU average of APAs per country increased by more than 75% from 19.53 to 34.71.

APAs and similar agreements can be an effective way of making the tax system more efficient by bringing certainty to tax payers, but they can also be misused to legitimise significant tax avoidance, as the LuxLeaks scandal revealed. In October 2015, EU Member States decided that details of tax rulings would be automatically exchanged between Member States within the EU. However, important as this step could be, it does nothing to assist European citizens in accessing the tax rulings as they won’t be public. The continued secrecy of the EU can also have implications for developing countries because a significant part of APAs in force at the end of 2014 were with non-EU countries, who will not benefit from the decision to exchange information since it only addresses the exchange of tax rulings between EU Member States.

What is really needed is public access to the essential elements of tax rulings in order to see the content of these ‘sweetheart tax deals’ that are becoming a key element of the EU tax system, which is something the European Parliament has already suggested. So far, the LuxLeaks scandal has not lead to any progress on this point. Whether the scandal has caused the number of ‘sweetheart deals’ to decrease remains to be seen. It will thus be interesting to see the data for 2015 and compare it to previous years. 

i According to the Cambridge Online Dictionary a ‘sweetheart deal’ is “an agreement that you make in which you get something that it is to your advantage, especially by agreeing to give up something else”.
ii APAs are one type of tax ruling quite common among member states and where statistics are available.
iii EU APA Guidelines SEC (2007) 246 {COM (2007) 71 final}
iv Data from the Netherlands is not included in the EC documents, therefore the figures don’t include the data that the Dutch government provided afterwards.
v Data from the Netherlands is not included in the EC documents.