Comments Guidelines

All comments are pre-moderated. No spam, slurs, personal attacks, or foul language will be allowed.

Wednesday, October 12, 2016

New study casts more doubt on data center subsidies

A new report by Good Jobs First confirms what has been long-suspected: Data center megadeals of over $50 million in subsidies create very few jobs at a cost per job that easily exceeds $1 million. Indeed, the average for 11 megadeals going to tech giants like Google, Apple, Facebook, and Microsoft came to $1.8 million ($2.1 billion/1174) nominal cost per job.

As I have discussed before, such a figure far exceeds what a typical automobile assembly plant will receive, even though the latter creates far more, and better-paying, jobs than server farms do. An auto facility will receive something around $150-200,000 per job, and it will bring along suppliers to boot (though, unfortunately, sometimes the suppliers will also receive incentives).

The new study finds that by far the most important site location consideration is the cost of electricity and, increasingly, whether the electricity is generated by renewable sources like wind or solar. Thus, many of the biggest data centers are located in states like North Carolina (cheap coal-fired plants), Oregon and Washington (cheap hydropower). States with cheap electricity do not need massive subsidies, but they provide them, anyway.

At least, they usually do. As I have related before, American Express in 2010 announced a $400 million data center in North Carolina, without incentives. But fear not, Amex had not forgotten about using the site selection process as a rent-seeking opportunity. The reason it did not seek incentives, as far as anyone can tell (don't forget about the inherent information asymmetry here), is that the company knew it was going to close a 1900-job call center in Greensboro, which would trigger clawbacks on the data center if it received subsidies for it. So in that case North Carolina gave no incentives for the server farm.

Not only that, Google knows how to build and expand data centers without incentives. Of course, that's in Europe. The Netherlands Foreign Investment Agency confirmed for me that it gave no subsidies to Google for a $773 million, 150-job center opening in Groningen province next year. I was unable to get affirmative confirmation on projects in Ireland, Finland, and Belgium, but none of them show up in the EU's Competition Directorate case database, so presumably they did not receive incentives either.

 The study concludes with sensible recommendations: Transparency where it doesn't exist, capping incentives at $50,000 per job, and knowing when to get out of subsidy auctions for these projects. Maybe simpler still, I would suggest that economic development officials just say no.

Friday, September 16, 2016

That's what I'm talkin' about! (Ireland)

I have argued many times (most directly here) that, contrary to the claims of nearly all Irish policymakers, low taxes are not what makes the Irish economy tick. The country experienced 30 years of low taxes with no gain on average European income; it was only after 1987 that other policy changes (education, EU-funded infrastructure, and Social Partnership) led to gains on the EU average. Thanks to a Tax Justice Network blog post, I now have a great illustration to show this in living color.

The graph below plots Irish income per capita as a percentage of the EU average from about 1955 to 2012, with important dates noted as vertical lines. Notice that Ireland doesn't get above 60-65% until after 1990. In addition, the Commission-enforced increase in the corporate income tax rate from 10% to 12.5%, which took effect in the early 2000s, had no impact on the Celtic Tiger's spectacular rise in income per capita relative to the EU average. This means Ireland had higher growth when the tax rate was 12.5% than when it was 0%!



Source: Tax Justice Network, link above
Cross-posted at Angry Bear

Tuesday, August 30, 2016

European Commission orders Apple to repay Ireland $14.5 billion in illegal tax benefits

Today the European Commission slapped down tax avoiding Apple and low-road economic developer Ireland with a $14.5 billion state aid repayment order. Yes, you read that right: Apple has to pay Ireland €13 billion in back taxes, plus interest, covering the years 2003-2014. According to the Commission's order, Apple's effective tax rate in Ireland ranged from 1% in 2003 to 0.005% in 2014, far below Ireland's statutory 12.5% corporate income tax rate.

This could not have happened to two more deserving parties. Apple, of course, constantly whines about taxes and is a pioneer in creating the most arcane tax avoidance strategies imaginable, such as the establishment of subsidiaries it claims are taxable *nowhere*.

Ireland has for over 50 years followed a low-road economic development strategy based on low taxes and high investment subsidies. When I interviewed numerous government officials there in 2009, almost all of them (literally, just one exception) were convinced that the country's economic success in the "Celtic Tiger" era (approximately 1990-2007) was due to its low-tax strategy. This argument overlooked the fact that the first 30 years of the strategy's use saw no gain whatsoever on EU average income, i.e., Ireland grew no faster than the average of the first 15 EU Member States (or EU-15, for short). It was only when Ireland used "higher road" strategies -- free high school, building new technological universities, social partnership, and a large infusion of EU-funded infrastructure -- that the country really began to take off.

While it is no surprise that Ireland and Apple reacted angrily to today's decision and will appeal the case to the Court of Justice of the European Union, the reaction of the U.S. Treasury is more puzzling. Basically, as John Judis explains at Talking Points Memo, the United States is defending Apple as a national champion that the Europeans shouldn't be charging additional taxes. Instead, Judis argues, the United States should be cheering on the EU pressure on tax avoidance and apply more of its own.

Critics of the decision argue that the Commission's Directorate-General of Competition has no expertise in tax cases. However, as one of the commenters there pointed out, this is properly a state aid (subsidy) case. Indeed, DG-Competition has been tackling fiscal aid cases such as this for 20 years. It was a previous DG-Competition investigation that in 1998 ruled Ireland's 10% tax rate to be not just a subsidy, but an operating subsidy, i.e., not based on investment but ongoing operations. The Commission really does know what it's doing here.

As a result, Ireland is now under pressure to abandon its well-known tolerance of murky tax arrangements, and Apple is finding itself at risk for aggressively creating tax avoidance gambits. It's a great day for honest taxpayers.

Mike Pence subsidizes offshorers

The Indianapolis Star reports (h/t Greg LeRoy) that the Indiana Economic Development Corporation (IEDC), which is headed by Governor/Vice-Presidential nominee Mike Pence, has given millions of dollars in incentives to companies that have subsequently offshored jobs to countries including China, Mexico, Taiwan, and Japan.

Not only that, but the 10 companies that outsourced the jobs sent more jobs abroad from Indiana, 3800, than their agreements with the IEDC required them to create, 1087. For this, the companies were initially awarded $24 million. Moreover, four of the 10 companies failed to reach their job commitments and were subject to clawbacks and termination of future payments under their agreements.

This is a startling mix of good practices and terrible practices. On the good side of the ledger, the state uses both performance-based incentives and clawbacks. Not only that, the cost per job is quite reasonable compared to most states, a mere 24,000,000/1087 = $22,079 per job.

The fly in the ointment is that Indiana does not measure job creation goals against a company's pre-existing state workforce, but against its workforce at the project site location only. Thus, Vera Bradley added about 30 jobs at its headquarters in the Fort Wayne area while laying off 250 at another Fort Wayne-area facility, and is considered to be in good standing on its incentives. How can I put this charitably? This is idiotic. As we can see, the ten companies ran rings around the IEDC and got $24 million for creating -2713 jobs.

It's not like other states aren't aware of this problem, so why isn't Indiana paying attention? According to one source quoted by the Star, among the states that prohibit what Indiana has allowed are immediate neighbors Ohio and Michigan, near neighbor Missouri, and also North and South Carolina.

To top it all off, IEDC does not make companies' job performance numbers public. (Banging my head on the table...) Thus, it is impossible to independently monitor the performance of subsidy awards. If not for the existence of federal trade adjustment assistance application data, the Star could not have done this study at all. Fortunately, the data were there, and we are treated to one more instance of the widespread underperformance of subsidy recipients.

And remember: Pence is the guy that Donald Trump chose as his running mate.

Friday, July 22, 2016

A bet with Tim Worstall on Apple's state aid repayment

Tim Worstall, a contributor at the Forbes Magazine website, wrote a column Thursday wherein he challenges all comers to a bet on the outcome of Apple's state aid case I discussed Sunday. He repeated the challenge to me specifically in the comments section of my post, and I accepted. It's for bragging rights only (I think I'd prefer for the prize to be a beer next time I go to London, oh well).

Worstall's angle is that he believes winning the bet shows he has a better understanding of the issues than I do.That's conceivable, but not necessarily true. In particular, he challenges my emphasis on
Ireland's creation of a corporate entity that is taxable nowhere, and the possibility that the company negotiated a tax rate far below the country's already-low 12.5% corporate income tax rate.
The thing is, we have strong evidence for what the case is about: The Commission decision that started this Article 108(2) detailed investigation. So let's turn to the decision and see what the Commission says it is investigating.

As paragraph (or "recital" in legalese) 6 states, this investigation revolves around transfer pricing (i.e., between different Apple subsidiaries) agreements called "advance pricing arrangements" or APAs. APAs are precisely the result of negotiations. Here is what the New York Times says about Apple Operations International:
Atop Apple’s offshore network is a subsidiary named Apple Operations International, which is incorporated in Ireland — where Apple had negotiated a special corporate tax rate of 2 percent or less in recent years — but keeps its bank accounts and records in the United States and holds board meetings in California.
Worstall is correct that these negotiations do not change the official tax rate of 12.5%; what they change is the effective tax rate. What the APAs do is reduce the amount of earnings of Apple Operations Europe (a wholly-owned subsidiary of Apple Operations International) and Apple Sales International (a wholly-owned subsidiary of Apple Operations Europe) -- I'll get to the details in a moment -- and these artificially derived profit numbers are then taxed at the standard rate of corporate income tax, 12.5%. To get the effective tax rate of under 2%, as cited by the NYT, is to claim that the tax paid (12.5% of the amount of profits as determined under the APAs) is actually less than 1/50th of the true profits these subsidiaries earned.

To take a numerical example, let's say that Apple Operations Europe made $625 million in 2015. However, under its advance pricing arrangement, let's say AOE is only deemed to have $100 million in profits for the year (we'll see how in a minute). AOE then pays 12.5% tax, or $12.5 million, in Irish corporation taxes. But 12.5/625=2%, so its effective tax rate is 2%.

So, when Worstall says this state aid case is not about a negotiated tax rate, he is not replying to what Apple's critics are actually saying.

How does Apple get its profit determined for the two subsidiaries with APAs? There are four separate rulings by the Irish government, two in 1991 (one for each company) and two in 2007 (one for each company). We'll take the 1991 APA for Apple Operations Europe as our example to show the main ideas. According to paragraph 31,
the net profit attributable to the AOE branch would calculated as 65% of operating expenses up to an annual amount of USD [60-70] million and 20% of operating expenses in excess of USD [60-70] million.
 When you think about this for a minute, you realize that this is a very odd formula for calculating "profit." Usually, we'd think about it as sales minus costs. This formula does not mention sales at all! So it's not surprising that it could give a figure far from what we would consider the true profit of Apple Operations Europe. For the record, AOE's 2007 formula does mention sales of intellectual property (paragraph 32), but neither the 1991 or 2007 formula for Apple Sales International uses sales in its calculation (paragraphs 33-34). Ho-kay!

Indeed, in its detailed analysis, the decision says that the formula for AOE was "reverse engineered" to allocate a specific profit number to the subsidiary (paragraph 61).

What about the non-tax resident company structure? Again, Worstall denies it's relevant here. But in fact, both Apple Operations Europe and Apple Sales International (and Apple Operations International, for that matter, though it's not a direct target in the case) are incorporated in Ireland but are not tax-resident there (paragraphs 25 and 27). These are the only Apple subsidiaries in Ireland that are not tax-resident in Ireland (paragraph 19). So this case wouldn't exist without this very unusual corporate structure.

To sum up, I believe that Worstall has misunderstood the reasons for the case given in the decision to open a detailed investigation. As I pointed out two years ago, if the Commission opens an investigation, it almost always finds that state aid was given, and that it is not compatible with the common market. So I am quite confident Apple will lose the case. Unless the Commission unexpectedly changes its mind on the importance of regulating fiscal aid, I am also confident that I will win my bet with Worstall. We'll know soon enough.

Sunday, July 17, 2016

We will soon know if Apple is ordered to pay billions in EU state aid case UPDATED

$19 billion? $8 billion? A few hundred million? Estimates are all over the map regarding the European Commission's imminent decision on whether Apple received illegal, unnotified state aid (subsidies) from Ireland. As I mentioned two years ago, two factors go into the Commission's claim: Ireland's creation of a corporate entity that is taxable nowhere, and the possibility that the company negotiated a tax rate far below the country's already-low 12.5% corporate income tax rate.

As you may remember from earlier posts, if the Commission rules against Apple, the sanction it will impose will be the repayment of the illegal subsidies, with interest. Some estimates of Apple's tax savings are astronomical (the New York Times reported an estimate Apple saved $7.7 billion in 2011 alone), which creates the possibility of an extremely high repayment order. On the other hand, recent cases have seen relatively low repayments, such as the mere 30 million the Commission ordered Starbucks to repay the Netherlands last year. Still, it is perfectly possible that the Commission used smaller cases as a way to establish the legal precedent regarding fiscal aid before dropping the bomb on Apple.

It's worth noting, as reported by Bloomberg, that the largest state aid repayment order ever made was approximately €1.4 billion charged to √Člectricit√© de France, followed by two orders in the €1 billion range. When this decision is announced, we may see a new record by a large margin. Or maybe we won't. Either way, the Commission will be giving us a clear sign regarding how seriously it intends to treat fiscal aid abuses.

Update: For a little perspective on fines, Chillin' Competition (via email) points out that the European Commission has just imposed an anti-trust fine on four truck manufacturers of over 2.9 billion , a new record.

Cross-posted at Angry Bear.

Wednesday, June 29, 2016

New book on investment incentives will help shape policies debates for years to come

This past week I received my chapter author's copy of a new book from Columbia University Press, Rethinking Investment Incentives: Trends and Policy Options. Based initially on the November 2013 conference on investment incentives at Columbia Law School, the contributors were put through their paces to upgrade their conference presentations into proper papers. The result is what Theodore Moran of Georgetown University calls in the Foreword "a who's-who of experts across this broad span of topics." He predicts, and I concur, that the work presented in this book will help drive policy discussions around the globe.

The book is divided into four parts. The first discusses theoretical debates on definitions and the effect of these incentives on (especially) foreign direct investment. The second section provides a global overview of the use of incentive incentives, both in major economies and in developing countries. Part III includes practical tools for ensuring program effectiveness as well as value for money. This includes a chapter on cost-benefit analysis, a methodology of which I am highly skeptical. As I have written before, if you end this analysis at the state (or city!) border, you miss many of the indirect job losses inflicted at competing companies by the addition of new subsidized competition. Indeed, according to economist Tim Bartik, very few subsidy programs have positive *national* effects, even if they have positive local effects that will be the only thing considered in the cost-benefit analysis.

Finally, the fourth part of the book considers ways to reduce the competitive use of investment incentives to attract investment. My chapter falls in this section, considering the control of subnational incentives in Australia, Canada, and the United States. (Spoiler: Most of the record is not pretty; Australia was an exception but the policy expired in 2011.) A variety of supranational regulatory efforts, including most notably that of the European Union, are considered in a chapter by Lise Johnson.

Have I teased you enough yet? This book is a must-have if you are interested in investment incentives and economic development; co-editors Ana Teresa Tavares-Lehmann (University of Porto, Portugal), Perrine Toledano, Lise Johnson, and Lisa Sachs (all of the Columbia Center for Sustainable Investment) are to be congratulated for the fine product.