America’s big tech firms are admired for their ingenuity — their products have improved our lives in myriad ways. But critics say they’ve also used that ingenuity to devise elaborate corporate structures that shift earnings to offshore tax havens. Apple and Microsoft have been pilloried before the U.S. Senate; the British Parliament has gone after Amazon and Google.
To be fair, aggressive tax strategies are the norm among U.S. multinationals, in part because our corporate tax rate of 35% is the highest in the world. But tech giants like Apple seem to have carried it further than most, contributing less to public coffers than other comparably sized U.S. corporations.
“There’s a lot of evidence that this is going on,” says Stanford Graduate School of Business professor Lisa De Simone. But there’s been surprisingly little analysis of how income shifting works. “Why are some companies doing it and others aren’t? How do they choose an optimal structure? What are the tradeoffs?”
For answers, she and Richard Sansing of the Tuck School of Business at Dartmouth modeled the decision problem facing a U.S. multinational. In their new paper, they find that a company’s ability to shift income depends on the type of assets it has. In particular, De Simone says, knowledge-based firms with global brands are at a distinct advantage under existing tax rules.
Export Profits, Not Products
To understand what’s meant by “income shifting,” consider a U.S. firm that exports a product to Mexico, earning a profit of $100. After tax, it keeps $65. Now suppose it sets up a subsidiary in Bermuda and books the sale there. The $100 is now foreign income, which the company can exclude on its U.S. tax return (at least until the money is repatriated, but in practice it rarely is).
The product can still go from the United States to Mexico, so there’s no extra shipping cost. And Bermuda has no corporate tax, so the firm keeps the full $100. No economic value has been added; there’s no business reason for the sale to come from Bermuda. The company has clearly shifted $100 of earnings offshore to avoid taxation — boosting its net income by 54% in the process.
Of course, if it were that simple, every company would do it and there would be no U.S. exports. “Some of the public debate gives the impression that multinationals can arrange things however they want,” De Simone says, but that’s far from being the case.
Assets on Wheels
First, the corporation needs to make a case that the subsidiary owns the products it sells. To do that, it has to move a share of its income-generating assets to Bermuda. Here’s where tech firms have an edge: Their core assets are things like software, patents, copyrights, and designs — all considerably more portable than, say, metal-stamping presses or farmland. By shuffling intellectual property around, these companies can locate income practically anywhere in the world, at least for non-U.S. sales, to minimize their overall tax bill.
Indeed, in another study last year, De Simone found that U.S. multinationals with mobile asset bases shifted a larger portion of their income offshore, achieving “high long-run levels of tax avoidance.” It’s not that tech firms are worse corporate citizens; they simply shift more income because they can. “The system creates certain incentives,” De Simone says, “and it’s hard to blame corporations for responding to those incentives.”
Drive a Soft Bargain
Second, any transfer of assets to a subsidiary is supposed to be paid for as if parent company and subsidiary were independent parties — in other words, at full market value. And that payment back to the parent company is taxable U.S. income, partially offsetting the benefit of the arrangement.
Of course, the transaction is a bit of elaborate theater: The subsidiary often pays with cash it gets from the parent, in exchange for its own initially worthless stock certificates. It’s not uncommon for the same corporate lawyer to play both roles, transferring money from one pocket to the other.
But when international tax rates differ, the terms of these deals really matter. To shift as much income as possible into a lower-tax jurisdiction, a multinational will want to undervalue the transferred assets. “It’s all about to what extent can the U.S. parent understate the payments it should receive from the foreign subsidiary,” De Simone says.
Choose a Structure
Here’s where it gets tricky. Intellectual property rights can be divvied up by any of three methods, and the corporation has to figure out, in essence, which one maximizes the difference between the true value of its assets and the transfer price it gets for them.
- Licensing Deal: The parent company can retain ownership of the intellectual property and license the use of it in exchange for annual royalty payments. Because this arrangement involves less planning and paperwork — and is less likely to invite an audit — it’s often preferred by smaller companies.
- Asset Sale: It can sell the intellectual property outright, in which case the subsidiary carries out any ongoing development work on it. This lines up with a common rationale for asset distribution: that products need to be adapted for foreign markets. In practice, IP is rarely moved to major-market countries, since they tend to have high tax rates.
- Cost-Sharing Agreement (CSA): Instead of splitting up research-and-development activities, this arrangement allows the subsidiary to simply pay a portion of ongoing development costs each year, after an initial buy-in contribution, in return for a commensurate share of the profits.
Data from an earlier study show that cost-sharing arrangements have become more common over time, which raises the question why. By modeling cash flows under alternative structures, De Simone and Sansing find that the choice turns, surprisingly, on name recognition. The reason is that under U.S. cost-sharing rules, companies can ignore marketing intangibles like trademarks and trade names in calculating the buy-in payment.
“The implicit assumption is that marketing assets have no value overseas,” De Simone says. But that’s often untrue in today’s global economy — think of the margins that Apple products command worldwide. “You’d expect the sub to pay for that,” she says. By leaving name recognition out, CSAs lower transfer prices and leave more income offshore. The bigger the brand, the larger the tax benefit.
According to Senate testimony in 2012, an analysis of cost-sharing arrangements by 15 major multinationals found that the foreign subsidiaries had an average return on assets of 268%, compared with 40% for their U.S. parents. Clearly those subsidiaries were not paying the market value of what they were getting.
An Exercise in Modesty
Chances are that’s not just because of the loophole on marketing intangibles. Under any of these methods, the IRS is at a big disadvantage in trying to judge whether transfer prices meet the arm’s-length standard. The company always has better information. It can assert, for instance, that the prospects for a new technology or product are more uncertain than it believes them to be.
The model shows that “if they can exploit that information asymmetry, it tends to favor a sale of assets over a cost-sharing arrangement,” De Simone says. Other studies have looked at the use of price manipulation to shift income. But they miss an important element, she says: The IRS can also come back later and change the terms if it finds they aren’t “commensurate with income.”
Incorporating that element of risk changes the calculation. “Maybe the taxpayer can get away with understating the value, but the punishment could be so severe if they get caught that it’s not worth it,” De Simone says. In such a case, it would have been better for the parent company to retain ownership, do all the R&D domestically, and license the IP to its subsidiaries.
“The company can do a sensitivity analysis on possible outcomes. In my experience doing some of this for multinationals, it’s something they look at. They try to take into account, how could this go wrong? What are the chances? What’s the rosiest picture? What’s the least rosy picture?”
Insights for Policy
Ultimately, the cause of all this artifice and scheming is the disparity in tax rates and rules from country to country, and particularly between the United States and the rest of the world. It creates an environment in which smart companies — just doing what they’re supposed to do, maximizing shareholder value — are able to game the system. In the process, it distorts business decisions and benefits some taxpayers at the expense of others, who have to make up the shortfall.
“That’s how we do things in the U.S.,” De Simone laughs. “We have our own set of rules for everything, not just taxes.” Perhaps the recent public debates will provide some impetus to finally move ahead on corporate tax reform and harmonization. “We’re starting to see multistate action in the European Union and the Organization for Economic Cooperation and Development to curb some of this behavior by multinationals,” she says. “It’ll be interesting to see whether it spurs some changes here too.”
If so, the insights provided by work like this, which gets beyond political grandstanding and moral exhortation to analyze companies as rational economic decision-makers, will be an essential guide.
Lisa De Simone is an assistant professor of accounting at Stanford GSB and the John S. Osterweis Faculty Scholar for 2014-2015.