The usual suspects have taken Amazon’s 25th birthday as the excuse to rehash accusations that the company is under-paying tax. Jeremy Corbyn’s message of ‘many happy tax returns’ was at least witty, but it still shows little grasp of basic economics, let alone how the tax system actually works. Unfortunately, the Conservative government also seems determined to exploit these misunderstandings with its own proposals for a ‘Digital Services Tax’.
Let’s deal briefly first with Amazon. I don’t claim to be an expert on the company’s tax affairs, unlike half of Twitter. I suggest that anyone with hard evidence that Amazon is under-paying tax should contact HMRC, rather than troll me. But it’s not difficult to spot some pretty elementary mistakes in much of the commentary.
For example, unfavourable comparisons are often made between the sales that Amazon makes in the UK, and the corporation tax it pays here. This is daft, because corporate taxes are paid on profits, not turnover. Any online marketplace might see a lot of business transacted between third parties on its website, but still be working on small profit margins, or even at a loss.
To pick another example, when a company only has a limited physical presence in a country, it is surely reasonable to expect it to pay less tax in that country. This is because it is not making the same demands on local taxpayer-funded public services or infrastructure.
Similarly, an online retailer might be able to gain a perfectly legitimate competitive advantage, including reducing their liability for business rates, by operating from a highly efficient out-of-town warehouse rather than a shop on a congested city-centre high street. This, after all, is one reason why Amazon is so successful.
You could go one step further and argue that the main job of any company is to provide goods and services that people want, and are prepared to pay for, not to generate revenue for the government. If they do that too, it’s a bonus. (Hat tip to Kristian Niemietz for this point.)
The UK Treasury does at least appear to understand most of these arguments. But it is still pressing ahead with plans for a Digital Services Tax (DST), due to be implemented in April 2020, which will charge an additional levy of 2% on the revenues of firms with a large digital presence, where these revenues can be linked to the participation of local consumers. The main targets will be social media platforms, search engines and online marketplaces (including Amazon).
To be fair, it’s not just the UK that’s planning to hit the tech sector. I recently contributed to a paper by EPICENTER, an independent network of think tanks from across Europe, which assessed proposals by the European Commission for an EU-wide tax on digital turnover, as well as unilateral initiatives in France, Italy and Spain.
These proposals are all based on the presumption that the tech sector does not pay its ‘fair share’ of tax. The justification for this claim is flimsy. Indeed, many advocates of a tech tax have resorted to what could be best be described as ‘policy-based evidence-making’.
In particular, the European Commission has claimed that multinational digital companies pay an average effective tax rate of only 9.5% in the EU, compared to 23.2% for more traditional businesses. These figures have been widely cited as evidence of systematic tax avoidance, or at least that the tech sector is ‘under-taxed’. But this is misleading, for two main reasons.
First, these estimates are based on stylised business models for hypothetical companies, rather than data from actual firms. They do not prove anything about the behaviour of the tech sector as a whole, let alone any individual firms. In reality, Bauer shows that the effective tax rates paid by traditional businesses are often lower than those paid by digital companies.
Second, where digital companies do pay less tax, this is usually for good economic reasons, or a result of tax breaks that governments themselves have promoted. Even in the stylised examples used by the European Commission, the main differences in effective tax rates are due to the more favourable treatment of R&D, software developed in-house and other intangible assets, as explained in the original work by PWC on which the analysis was based.
In the case of Amazon, the company has also made the most of tax breaks for employee share ownership schemes, again something that critics of the company would presumably support. (Of course I’m aware there have also been criticisms of how Amazon has treated its workers, especially warehouse staff, but the tax system is not the place to address allegations of poor labour practices.)
It might be that the nature of their business models allows tech companies to benefit more than traditional firms from these tax breaks. But that does not reflect any ‘bad behaviour’ on their part. If anything, the reverse could be true – it reflects the fact that the tax incentives are working to encourage the behaviours that the policymakers were seeking when they designed them.
In addition, UK officials have had to come up with a pretty dubious economic justification for targeting the tech sector with additional taxes. The DST seeks to extend the widely accepted principle that the profits of a business should be taxed in the countries in which it creates value to cover the participation of online customers in the UK.
However, goods and services with user-created value are nothing new, and certainly not restricted to the digital sector. Many traditional businesses, such as airlines and shipping companies, benefit from value contributed by foreign nationals, or have a substantial online element. There is a clear risk that decisions about what is and what is not within the scope of a digital tax become increasingly arbitrary and distortionary.
There are many practical problems too in deciding where and how much value is being created. The DST will require agreement on which activities, and which users, come within the scope of the new tax and which do not. This is hard enough for physical businesses, let alone the online world. For example, ‘smart car’ technology relies on collecting information from users, and more and more retailing and banking is now done online
There has also been next to no consideration of the broader economic impact of the DST. Companies, especially US-based tech giants, may be a popular target, but they are only legal entities and cannot bear the economic burden of taxes themselves. The reality is that all taxes are ultimately paid by people – including employees and customers, as well as shareholders. What’s more, turnover taxes have particularly large deadweight costs, are more likely to be passed on to consumers, and are a major deterrent to investment.
To cap it all, it’s not even certain that digital taxes are legal in international law. The EU proposals in particular seem to be designed to target large US-based firms. Indeed, even French officials describe their initiative as a ‘GAFA’ tax, aimed at Google, Apple, Facebook and Amazon, which is a bit of a giveaway. As Hufbauer and Lu explain, this appears to be a de facto tariff in breach of WTO rules. The DST certainly wouldn’t make a US-UK trade deal any easier.
In short, additional taxes on the turnover of companies with a large digital presence would be disproportionate and discriminatory, and further complicate the tax system for no obvious benefit. Fortunately, it’s not too late to press delete on the DST. And you don’t like Amazon’s business model, don’t buy from them.
This is an extended version of an article which first appeared on CapX