CTJ’s Measurement Problem
May 28, 2014
This week, the Citizen’s for Tax Justice released a report titled “American Corporations Tell IRS the Majority of Their Offshore Profits are in 12 Tax Havens.” According to their analysis of IRS data, they found that U.S. controlled foreign corporations only paid an effective tax rate of 7 percent on their foreign earnings in 12 countries in 2010. They also find that the average effective rate across all countries was 17 percent.
Unfortunately, the underlying data the CTJ uses has significant flaws that make the analysis virtually meaningless. Specifically, the data substantially double-counts income earned by U.S. controlled foreign corporations. As a result, the amount of profits they measure in certain countries is overstated. This in turn greatly understates the effective tax rate corporations pay.
Double-Counted Foreign Corporate Profits
One of the most important things to keep in mind when measuring foreign corporate income, corporate taxes paid, and their effective tax rate is that there are several different data sets you can use. None of them are perfect and they all have their flaws. CTJ’s data source of choice certainly has them.
CTJ used data from IRS form 5471, which tracks profits and taxes paid by U.S. controlled foreign corporations (CFCs). While this data tracks current income and current income taxes paid, whether it is repatriated or not, it significantly overstates income by double-counting CFC-to-CFC dividend payments
An example adapted from an Altshuler and Grubert paper shows how this double-counting happens. Suppose a U.S. parent corporation has control of a foreign corporation (company A), which is located in a country with no corporate income tax. Then suppose company A owns a foreign corporation (company B) in another country with a 10 percent corporate income tax rate.
Assume then that company B earns $100 and pays a 10 percent tax rate to the foreign country in which it earned that income. Company B then transfers the $90 in after tax profits to company A through dividend payments. Company A collects those after-tax dividends. The U.S. parent company then needs to report the income of both majority-owned foreign affiliates. It reports $100 earned by company B and $90 earned by company A. On net, foreign affiliate income is measured to be $190 in this data, but in reality only $100 in pre-tax income was earned, the extra $90 is just an accounting illusion.
This drives down the effective tax rate of this corporation’s foreign earned income to 5 percent when in reality the effective rate on the income is 10 percent.
This would be akin to the CTJ saying that your income just doubled because you moved your paycheck from your right pocket to your left.
Furthermore, if U.S. affiliates in certain countries have majority-control of many foreign companies, they may receive a large amount of this after-tax dividend income from multiple sources. Their income would be drastically overstated in this case. Think of what the income of company A would look like if it also owned several other foreign affiliates that transferred after-tax dividends to it.
In fact, BEA data shows that more than $68.2 billion of this type of doubled-counted income is reported in Bermuda, $95 billion in Luxembourg, and $145 billion in the Netherlands. A Tax Notes International Article (here) also finds that most of the growth of income earned in these countries was due to this type of income.
There’s no doubt that U.S. corporations use a variety of legal methods to reduce their corporate tax bill on their overseas operations. It also may be the case that corporations use low-tax countries as a means to do so. But measuring the extent to which this happens is not an easy task. All data sources have their flaws and it is best to be open about the limitations when doing these types of analyses.