The Verge (“Amazon’s $300 million tax bill rejected by EU judges“):

Judges from the European Union’s second-highest court have rejected a €250 million ($300 million) tax bill lodged against Amazon in 2017 as part the bloc’s ongoing fight against US tech giants.

The case was one of a number spearheaded by Margrethe Vestager, the European Commissioner for Competition, in which sweetheart tax deals given to powerful corporations were framed as a form of illegal state subsidy. The most notable of these was a 2016 case in which Apple was ordered to pay Ireland €13 billion ($14.9 billion) in back taxes. This decision was annulled in 2020 by the same court involved in today’s ruling.

The Amazon case can be traced back to 2006, when the e-commerce giant established a labyrinthine tax structure in Europe that allowed it to funnel revenue from all EU sales through a subsidiary based in Luxembourg. Internally, Amazon referred to this as Project Goldcrest, named after Luxembourg’s national bird.

In 2017, the European Commission ruled that this structure was illegal and had allowed Amazon to avoid around €250 million in taxes. “Luxembourg gave illegal tax benefits to Amazon,” said Vestager at the time. “As a result, almost three quarters of Amazon’s profits were not taxed. In other words, Amazon was allowed to pay four times less tax than other local companies subject to the same national tax rules.”

In Amazon’s most recent financial filings it recorded revenue of €44 billion ($53 billion) passing through its Luxembourg subsidiary. But this subsidiary, which has 5,262 employees, also registered €1.2 billion in losses, and so paid zero corporation tax.

In the ruling this morning announced by the General Court of the European Union, judges found that the Commission “did not prove to the requisite legal standard that there was an undue reduction to the tax burden” of Amazon’s Luxembourg subsidiary. The ruling is a significant win for Amazon and a blow for EU politicians hoping to rein in US tech giants.

Although the 2017 ruling has now been annulled, today’s decision can still be appealed to the EU’s highest court, the European Court of Justice. The Apple case overturned in 2020 has already been appealed in this fashion and is awaiting a further ruling.

After details of Project Goldcrest were first revealed, the US Internal Revenue Service (IRS) also filed its own case against Amazon, seeking up to $1.5 billion in back taxes. A federal court judge rejected this case in 2017, saying the figure claimed by the IRS was “arbitrary, capricious, and unreasonable.”

I’m of two minds on this. On the one hand, corporations and people will naturally pay as little tax as they can legally get away with and, to the extent that Amazon cleverly structured their tax burden the maximize profits for shareholders, they’re fulfilling their fiduciary responsibilities. On the other, to the extent that we want to tax corporations as entities rather than their owners and employees, this race to the bottom is good for no one.

My longstanding instinct is that we simply shouldn’t tax businesses, large or small, as businesses. Doing so naturally requires a complicated assessment of profits and losses and practically invites gaming the system. It’s much easier to tax personal income or, even better, purchases.

But, given that practically every developed country attempts to tax business as business, the only way to do that effectively is to have a universal rate and set of rules. The OECD countries have been trying for more than a decade to do just that, with little success. Partly, because the project itself is hard:

In a globalised and digital economy, multinationals operate through centrally managed business models, and their global profits are largely the result of their global operations. Yet current international tax rules, developed nearly a century ago, treat subsidiaries of multinationals as legally independent firms which trade between each other using “arm’s length” or normal commercial prices to transfer goods and services.

But such prices are not always easy to find. Many markets are thin and dominated by the same multinationals, who then exploit this system to minimise their tax liability by shifting profits to jurisdictions with low or zero tax rates. This undermines the tax base of countries where real activities occur and, therefore, where the profits have been generated. These rules are also skewed in favour of rich countries because they help multinationals’ home countries get the biggest share of tax from global profits. This “transfer pricing” is exacerbated by tax competition to the point that the global average statutory corporate tax rate has fallen by more than half in three decades.

But mostly because there is a wide diversity of philosophical view on how to fix and even whether to fix it:

Comprehensive reforms have been hindered by dominant OECD member governments, which come to negotiations with the misplaced perception that national interest is served by protecting multinationals headquartered in their own countries. This has prevailed over genuine, global public interest.

The US has been part of that problem, although there are signs that the Biden Administration may pursue a different policy. Right now, though, as Tommaso Faccio and Jayati Ghosh argue, while a slow consensus is forming, the US is a major obstacle,

The negotiating process has nonetheless reached agreement that multinationals should be considered unitary businesses. This means that their worldwide profits should be taxed in line with their real activities in each country and allocated to different jurisdictions, based on a formula according to the key factors that generate profit: employment, sales and assets. Many states in the United States use a similar “formulary apportionment” system to determine their taxable shares of US corporate profits. In 2016, the EU Commission put forward a similar proposal for an EU Common Consolidated Corporate Tax Base, but it has not yet been approved by the European Council.

Formulary apportionment would remove the current artificial incentive for multinationals to shift reported income to low-tax locations. Tax liabilities, instead, would be allocated by measures of their real economic activity in each location. But the proposal currently being negotiated involves applying this to only a small share of a firm’s global profits (so-called “residual” rather than “routine” profits) and is mainly directed at mostly US-based highly digitalised multinationals.

That even the EU (now minus the UK, which is closely aligned with the US on the matter) can’t agree on this shows how hard the problem is. But Faccio and Ghosh want to go much further:

This is not sufficient to address the problem.

Instead, we need a more ambitious and comprehensive reform that replicates the US system at the international level, without distinction between digital and non-digital businesses. This would help to establish a more level playing field, reduce distortions, limit opportunities for tax avoidance, and provide certainty to multinationals and investors. To put an end to harmful tax competition between countries, this system should be supported by a global minimum tax on multinationals so as to reduce the incentive for multinationals to shift profits to tax havens.

Until recently, negotiations on a global minimum tax were benchmarked by the existing US minimum tax on US corporations’ foreign earnings (known as “GILTI”), which has a rate of 10.5%. As a result, public discourse centred around a possible minimum tax rate of around 12.5% (incidentally, the corporate tax rate in Ireland, one of the EU’s own tax havens). Such a low minimum tax rate could in fact over time become the global ceiling, in which case the laudable initiative to oblige multinationals to bear their fair share of taxes would end up doing the opposite.

Honestly, actually getting 10.5% to 12.5% effective taxation, and having it flow to where it’s earned, would be far preferable to the current situation, where corporations spend millions on accountants and lawyers to hide their earnings, inflate their losses, and shift what modest burden they owe to the lowest bidder.