Amazon pays no UK corporation tax

The company attributes nearly all its income to its Luxembourg branch.

The Guardian has a major story today on the tax affairs of Amazon UK. The online retailer, Britain's biggest, pays no corporation tax in the UK, despite having between £2bn and £3bn sales here in 2010. The company avoids paying anything by registering the vast majority of its turnover in its Luxembourg office, which reported income of €7.5bn in 2010, despite having just 134 employees. The British office, which employed 2,265 people, reported a turnover of £147m.

The paper explains:

The UK operation avoids tax as the ownership of the main Amazon.co.uk business was transferred to a Luxembourg company in 2006. The UK business is now owned by Amazon EU Sarl and the UK operation is classed only as an "order fulfilment" business. All payments for books, DVDs and other goods go directly to Luxembourg. The UK business is simply a delivery organisation.

This arrangement saves them millions in tax:

At first glance, the corporation tax rates in Luxembourg and the UK are similar, but the Luxembourg authorities have a different view of costs that can be offset against income, which reduces taxable profit. So Amazon EU Sarl's €7.5bn of income in 2010 was almost entirely offset by €7.4bn of charges, enabling it to disclose a tax charge of just €5.5m. The charges are defined by the company as the "cost of product sales and other ongoing costs related to the operations of the company"…

This is in stark contrast to the performance of the UK fulfilment business which filed its 2011 accounts last month. For the first time since 2006, Amazon.co.uk Limited posted an after-tax profit of £1.2m, much better than the £3m after-tax loss reported a year earlier. The accounts show its turnover was £208m, a big improvement on the £147m recorded in 2010 but dwarfed by the £3.3bn of UK sales passed to Luxembourg.

The company still pays a fair amount of UK tax, because VAT is charged based on the location of the recipient, not the business. But a significant proportion of Amazon's sales are books, which are zero-rated for VAT; they also don't have to pay British sales tax on downloads, instead paying Luxembourgish rates. For ebooks, this is 3 per cent, rather than the 20 per cent they would be paying in the UK. Until last Sunday, the company also managed to not pay VAT on almost every sale under £18. It used a loophole, originally designed to protect flower sales, which allowed low-cost goods to be imported from the Channel Islands VAT-free.

Richard Murphy, the tax campaigning accountant, suggested how a revised tax code could more accurately assess the company's holdings:

First split the profit in three. One third is then allocated between the UK and Luxembourg based on where the sales really are. Well, all these sales are to UK customers so that ratio is 100% to the UK and 0% to Luxembourg. So that £125 divided by three = £41.66 million of profit allocated to the UK.

Then we split the next third on the basis of where the people are. That’s 2,265 here and 10 in Luxembourg. £41.66 million x 2,265/2,275 = £41.5 million to the UK and £166,000 to Luxembourg.

And then let’s do assets – admittedly the one I have had to guess. The guess is £100 million here and £5 million in Luxembourg so that is £41.66 million x 100/105 = £39.7 million of profit here and £1.96 million to Luxembourg.

Add it up and near enough £122.8 million of profit would be in the UK and £2.2 in Luxembourg. Instinctively that feels right of course - because that is exactly how the economics really are. Glaringly obviously, as Amazon’s accounts admit, the market is here in the UK, not in Luxembourg. But the game of abuse that is being played means that almost all the profit goes to Luxembourg on this one – and almost none to us.

Implementing such a change would be a massive undertaking, though, as well as being difficult to get through EU law. For now, the news is likely to remain just bad PR for Amazon.

The Amazon warehouse in Swansea, in the run-up to Christmas. Credit: Getty

Alex Hern is a technology reporter for the Guardian. He was formerly staff writer at the New Statesman. You should follow Alex on Twitter.

Photo: Getty
Show Hide image

The Prevent strategy needs a rethink, not a rebrand

A bad policy by any other name is still a bad policy.

Yesterday the Home Affairs Select Committee published its report on radicalization in the UK. While the focus of the coverage has been on its claim that social media companies like Facebook, Twitter and YouTube are “consciously failing” to combat the promotion of terrorism and extremism, it also reported on Prevent. The report rightly engages with criticism of Prevent, acknowledging how it has affected the Muslim community and calling for it to become more transparent:

“The concerns about Prevent amongst the communities most affected by it must be addressed. Otherwise it will continue to be viewed with suspicion by many, and by some as “toxic”… The government must be more transparent about what it is doing on the Prevent strategy, including by publicising its engagement activities, and providing updates on outcomes, through an easily accessible online portal.”

While this acknowledgement is good news, it is hard to see how real change will occur. As I have written previously, as Prevent has become more entrenched in British society, it has also become more secretive. For example, in August 2013, I lodged FOI requests to designated Prevent priority areas, asking for the most up-to-date Prevent funding information, including what projects received funding and details of any project engaging specifically with far-right extremism. I lodged almost identical requests between 2008 and 2009, all of which were successful. All but one of the 2013 requests were denied.

This denial is significant. Before the 2011 review, the Prevent strategy distributed money to help local authorities fight violent extremism and in doing so identified priority areas based solely on demographics. Any local authority with a Muslim population of at least five per cent was automatically given Prevent funding. The 2011 review pledged to end this. It further promised to expand Prevent to include far-right extremism and stop its use in community cohesion projects. Through these FOI requests I was trying to find out whether or not the 2011 pledges had been met. But with the blanket denial of information, I was left in the dark.

It is telling that the report’s concerns with Prevent are not new and have in fact been highlighted in several reports by the same Home Affairs Select Committee, as well as numerous reports by NGOs. But nothing has changed. In fact, the only change proposed by the report is to give Prevent a new name: Engage. But the problem was never the name. Prevent relies on the premise that terrorism and extremism are inherently connected with Islam, and until this is changed, it will continue to be at best counter-productive, and at worst, deeply discriminatory.

In his evidence to the committee, David Anderson, the independent ombudsman of terrorism legislation, has called for an independent review of the Prevent strategy. This would be a start. However, more is required. What is needed is a radical new approach to counter-terrorism and counter-extremism, one that targets all forms of extremism and that does not stigmatise or stereotype those affected.

Such an approach has been pioneered in the Danish town of Aarhus. Faced with increased numbers of youngsters leaving Aarhus for Syria, police officers made it clear that those who had travelled to Syria were welcome to come home, where they would receive help with going back to school, finding a place to live and whatever else was necessary for them to find their way back to Danish society.  Known as the ‘Aarhus model’, this approach focuses on inclusion, mentorship and non-criminalisation. It is the opposite of Prevent, which has from its very start framed British Muslims as a particularly deviant suspect community.

We need to change the narrative of counter-terrorism in the UK, but a narrative is not changed by a new title. Just as a rose by any other name would smell as sweet, a bad policy by any other name is still a bad policy. While the Home Affairs Select Committee concern about Prevent is welcomed, real action is needed. This will involve actually engaging with the Muslim community, listening to their concerns and not dismissing them as misunderstandings. It will require serious investigation of the damages caused by new Prevent statutory duty, something which the report does acknowledge as a concern.  Finally, real action on Prevent in particular, but extremism in general, will require developing a wide-ranging counter-extremism strategy that directly engages with far-right extremism. This has been notably absent from today’s report, even though far-right extremism is on the rise. After all, far-right extremists make up half of all counter-radicalization referrals in Yorkshire, and 30 per cent of the caseload in the east Midlands.

It will also require changing the way we think about those who are radicalized. The Aarhus model proves that such a change is possible. Radicalization is indeed a real problem, one imagines it will be even more so considering the country’s flagship counter-radicalization strategy remains problematic and ineffective. In the end, Prevent may be renamed a thousand times, but unless real effort is put in actually changing the strategy, it will remain toxic. 

Dr Maria Norris works at London School of Economics and Political Science. She tweets as @MariaWNorris.