Powerful Google tax opponent will urge UK to drop hostility to radical EU change

Multinationals would file single European tax return under plan proposed by EU tax commissioner to stamp out aggressive avoidance

If multinationals filed a single EU tax return, EU commissioner Pierre Moscovici believes, it would remove the temptation for them to divert income from one country to another. Photograph: Gareth Fuller/PA

Simon Bowers

@sbowers00

Wednesday 27 January 2016 17.39 GMT Last modified on Wednesday 27 January 2016 22.01 GMT

Share on Pinterest Share on LinkedIn Share on Google+

Shares

21

Comments

53

 Save for later

One of the most powerful opponents of Google’s controversial tax structures, European tax commissioner Pierre Moscovici, is expected on Thursday to call on Britain and Ireland to drop their objections to radical tax reform across the EU.

Moscovici, who has previously advocated a Europe-wide “digital tax” on companies such as Google, now wants to tackle aggressive tax avoidance among multinationals by requiring them to file a single European tax return.

Cameron defends Google tax deal in Commons clash with Corbyn

 Read more

He believes this reform – known as the common consolidated corporate tax base (CCCTB) – would remove the temptation for international firms to artificially divert income from one country to another. Member states, would still be free to set their own corporate tax rates.

Britain, however, is among a small band of countries fiercely opposed to the European commission’s plans, believing they would weaken the UK’s ability to tailor its tax system to attract jobs and investment from international businesses.

“The CCCTB [proposal] has been around a very long time,” Treasury minister David Gauke said last year. “It is a proposal still looking for a justification.”

Moscovici is due to give an update on other corporate tax reforms in Brussels on Thurday morning, but is expected to use the occasion to insist his CCCTB reforms are far from dead in the water – despite British opposition.

The former French finance minister has a long track record of challenging the tax affairs of internet companies – and Google in particular. Two years ago, he led calls for the G20 to create dedicated tax rules for digital companies, though his efforts were ultimately blocked by American pressure. He has also previously advocated a Europe-wide “digital tax” on internet companies that make money from consumers’ personal data.

France has consistently taken a tougher approach to aggressive tax planning by Google and other digital companies. In 2011, tax inspectors raided the search group’s Paris offices, and ever since have been challenging Google’s claims that its French sales can be legitimately booked in Ireland. French tax officials are said to be seeking £380m in back taxes.

Google tax deal: MPs launch inquiry after criticism of £130m settlement

 Read more

Google is next week expected to reveal that its controversial tax structures have boosted its offshore cash reserves to $43bn – up $4bn in 12 months.

Overnight on Tuesday, fellow US tech group Apple, which also uses controversial tax structures to pay less tax in Europe, revealed its offshore cash pile has now reached $200bn – largely held through companies in Ireland. Chief executive Tim Cook boasted Apple now had “the mother of all balance sheets”.

Earlier this month, Moscovici told MEPs that he wanted to make 2016 “the year of tax reform”, with CCCTB at the centre of his plans. “We have a serious problem with tax avoidance and lack of transparency. Too many people have looked the other way”, Moscovici said.

Since unveiling his reforms last summer, the European tax commissioner has been barraged with lobbying submissions. The majority have come from business trade bodies, law firms and multinationals, many based in the UK, Ireland and the Netherlands.

ut Moscovici insists widespread anger among voters at a string of tax scandals will play a vital role in winning round reluctant governments, often subject to lobbying from big business. He has described CCCTB as part of a [global] trend, drawing support from the pressure of public opinion.

Britain, meanwhile, has attempted to tackle Google’s tax avoidance in its own way. Chancellor George Osborne last year introduced a new tax on diverted profits, having promised to put a stop to technology companies such as Google going to what he called “extraordinary lengths to pay little or no tax” in the UK. Of those who used such structures, he said: “you abuse the trust of the British people”.

The stories you need to read, in one handy email

 Read more

But last week his crackdown pledge was left in tatters after Google confirmed it had struck a deal with HMRC that effectively allowed it to continue to route £4.6bn of UK sales via an Irish company that pays no tax in Britain. Google’s UK arm – which paid £21m in tax according to its latest accounts – will only be required to pay slightly more to HMRC under the settlement.

Critics have branded it a sweetheart deal, pointing out that the £130m in back taxes, which relates to a 10-year period, is tiny in comparison with the sums mounting up in Google’s coffers in Bermuda.

George Osborne has been vocal in supporting some initiatives on international tax reform, but has confused many tax experts by also slashing the UK tax rate – due to fall to 18% by 2020 – and introducing controversial tax breaks to attract multinationals to invest in Britain.

Google last year stalled its plans for a big new London headquarters in King’s Cross, London, insisting, according to reports, that the design proposals for a building – complete with a rooftop pool – to house 5,000 workers were “boring”. A new architect has since been hired, and fresh plans are now expected.

In 2011, Google increased its workforce in France by half and invested heavily in a new Paris head office near the Saint-Lazare train station amid public anger about its tax payments. One French senator accused the group of running its local business as a “charity”.

Google AI computer beats human champion of complex Go boardgame

Fan Hui, three-time champion of the east Asian board game, lost to DeepMind’s program AlphaGo in five straight games

 Fan Hui makes a move against AlphaGo in DeepMind’s HQ in King’s Cross. Photograph: Google DeepMind

Alex Hern

@alexhern

Wednesday 27 January 2016 18.15 GMT Last modified on Wednesday 27 January 2016 22.01 GMT

Share on Pinterest Share on LinkedIn Share on Google+

Shares

1,854

Comments

186

 Save for later

When Gary Kasparov lost to chess computer Deep Blue in 1997, IBM marked a milestone in the history of artificial intelligence. On Wednesday, in a research paper released in Nature, Google earned its own position in the history books, with the announcement that its subsidiary DeepMind has built a system capable of beating the best human players in the world at the east Asian board game Go.

Go, a game that involves placing black or white tiles on a 19x19 board and trying to remove your opponents’, is far more difficult for a computer to master than a game such as chess.

DeepMind’s software, AlphaGo, successfully beat the three-time European Go champion Fan Hui 5–0 in a series of games at the company’s headquarters in King’s Cross last October. Dr Tanguy Chouard, a senior editor at Nature who attended the matches as part of the review process, described the victory as “really chilling to watch”.

“It was one of the most exciting moments of my career,” he added. “But with the usual mixed feelings … in the quiet room downstairs, one couldn’t help but root for the poor human being beaten.”

It’s the first such victory for a computer program, and it came a decade before anyone expected it. As recently as 2014, Rémi Coulom, developer of the previous leading Go game AI, Crazy Stone, had predicted that it would take 10 more years for a machine to win against a top-rated human player without a handicap.

AlphaGo beat all expectations by approaching the challenge in a completely different way from previous software. Building on techniques DeepMind had employed in other feats of artificial intelligence, such as its system that could learn to play retro video games, AlphaGo used what the company calls “Deep Learning” to build up its own understanding of the game. It could then pick the moves it thought most likely to win.

When teaching a computer to play a game, the simplest method is to tell it to rank every possible move over the course of the game, from best to worst, and then instruct it to always pick the best move. That sort of strategy works for trivial games such as draughts and noughts and crosses, which have both been “solved” by computers that have fully examined every board state and worked out a way to play to at least a draw, no matter what the other player does.

However, for complex games such as Chess, the simple approach fails. Chess is just too big: in each turn there are approximately 35 legal moves, and a game lasts for around 80 turns. Enumerating every board position becomes computationally impossible very quickly, which is why it took so many years for IBM’s team to work out a way to beat Kasparov.

Go is bigger still. The definition of easy to learn, hard to master, it essentially has just two rules governing the core play, which involves two players alternately placing black and white tiles on a 19x19 board. The stones must be placed with at least one empty space next to it, or part of a group of stones of the same colour with at least one empty space, and if they lose their “liberty”, they are removed from the board.

While a game of chess might have 35 legal moves each turn, a game of Go has around 250 (including 361 legal starting positions alone); where Chess games last around 80 turns, Go games last 150. If Google had tried to solve the game in the same way noughts and crosses was solved, it would have had to examine and rank an obscene amount of possible positions: in the ballpark of 1,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000 of them.

That renders an exhaustive search impossible, and even a selective search, of the style used by Deep Blue to defeat Kasparov, tricky to run efficiently.

Adding to the woes of those trying to master Go is the fact that, unlike chess, it’s very difficult to look at the board and mathematically determine who is winning. In chess, a player with their queen will probably beat a player whose queen has been taken, and so on: it’s possible to assign values to those pieces, and come up with a running score that roughly ranks each player’s prospects. In Go, by contrast, counters are rarely removed from the board, and there’s no simple mathematical way to determine who is in the stronger position until the game is very far progressed.

So AlphaGo focused on a very different strategy. As David Silver, DeepMind’s co-lead researcher on the project, puts it: “AlphaGo looks ahead by playing out the rest of the game in its imagination, many times over.” The program involves two neural networks, software that mimics the structure of the human brain to aggregate very simple decisions into complex choices, running in parallel.

One, the policy network, was trained by observing millions of boards of Go uploaded to an online archive. Using those observations, it built up a predictive model of where it expected the next piece to be played, given knowledge of the board and all previous positions, that could accurately guess the next move of an expert player 57% of the time (compared to a previous record of 44.4% from other groups).

This “supervised learning” was then backed up by a bout of “reinforcement learning”: the network was set to play against itself, learning from its victories and losses as it carried out more than 1m individual games over the course of a day.

The policy network was capable of predicting the probability that any given move would be played as next, but the system also needed a second filter to help it select which of those moves was the best. That network, the “value network”, predicts the winner of the game given each particular board state.

Building AlphaGo isn’t just important as a feather in DeepMind’s cap. The company argues that perfecting deep learning techniques such as this are crucial for its future work. Demis Hassabis, DeepMind’s founder, says that “ultimately we want to apply these techniques in important real-world problems, from medical diagnostics to climate modelling”.

For now, the DeepMind team is focused on one final goal on the Go board: a match against Lee Se-dol, the world champion. Lee says that “regardless of the result, it will be a meaningful event in the baduk (the Korean name for Go) history. I heard Google DeepMind’s AI is surprisingly strong and getting stronger, but I am confident that I can win at least this time.”

Google Pays Britain $185 Million to Settle Back Taxes

A customer at a Google store in London. CreditAndrew Cowie/European Pressphoto Agency

SAN FRANCISCO — Google on Friday agreed to pay 130 million pounds, or about $185 million in back taxes to Britain, making it the latest United States technology company to settle claims that it does not pay its fair share of taxes in Europe.

The sum covers taxes from 2005 to 2015 and Google said it would change how it calculates its tax payments in Britain so they are based on a percentage of local sales derived from the country.

“We will now pay tax based on revenue from U.K.-based advertisers, which reflects the size and scope of our U.K. business,” a Google spokesman wrote in an email. “The way multinational companies are taxed has been debated for many years and the international tax system is changing as a result.”

Last April, Britain adopted a so-called Google tax that would impose a levy on any international company that did not fairly pay taxes on profits generated from its British operations.

Google, which is now owned by a holding company called Alphabet and has its European headquarters in Ireland, is hardly the only technology company with European tax problems.

Various countries, including Germany and France, have criticized the complicated tax structures tech companies use to reduce their local taxes. Many of them route sales through lower-tax countries like Ireland, even if the sales are made in other nations.

In May, Amazon, which had been funneling most of its sales taxes through Luxembourg, a low-tax haven, said it would start paying taxes in European countries where it has large operations. Apple reached a deal to pay local Italian tax authorities in December after authorities there looked into whether the company tried to lower its taxes by moving more than $1 billion in revenue from its Italian operations through an Irish subsidiary.

The executive arm of the European Union, the European Commission, is also investigating whether Apple and Amazon receive unfair state support through low-tax agreements in Ireland and Luxembourg.