Comment on Desai and Hines, "old rules and new realities: corporate tax policy in a global setting".

INTRODUCTION

In the December 2004 issue of the National Tax Journal, Desai and Hines claim that the current U.S. tax burden on foreign income is in the neighborhood of $50 billion a year." This is, of course, in addition to the foreign host country tax burden, which can be credited against the initial tentative U.S. tax on the income. This alleged $50 billion loss in profits to U.S. corporations resulting from the U.S. tax system is made up of three related components. The first is the estimated $20 billion of U.S. tax collected in 1999 on all corporate income now defined as foreign source under the U.S. tax rules. Based on this initial $20 billion, they conclude that a further $10 billion should be attributed to the effect of the taxes owed on unrepatriated earnings. The final $20 billion is the additional after-tax profits U.S. companies would have been able to earn if all foreign income were totally exempt. U.S. firms would expand investment abroad and have a greater incentive to avoid foreign taxes. Desai and Hines suggest that these large 'efficiency' gains justify complete exemption of all foreign income. They also justify the exemption of all foreign income on the basis of their concept of 'ownership neutrality' among competing buyers of foreign assets that they introduce at the end of the paper.

Unfortunately, the measure Desai and Hines present seems to have no conceptual basis and cannot be used to address any relevant policy issue. This note shows that each step in their analysis is based on flawed theory and the misinterpretation of data. To begin with, the actual U.S. residual tax on all foreign source income was not $20 billion in 1999; it was substantially less, as we will see, because Desai and Hines use the published data for total foreign source income before 'adjustments' for domestic losses. More important, Desai and Hines seem to assume that all of the residual U.S. tax is obtained from dividend repatriations. In fact, a closer look at the data indicates that only a small part (perhaps 15 percent) of the direct tax revenue from 'foreign' income is attributable to dividends from active business income. Most of the remainder is derived from payments that are deductible abroad, such as royalties and interest, or activities performed exclusively in the United States, such as export sales and loans extended by U.S. banks to foreign borrowers. This suggests that the $10 billion attributed to the cost of deferring dividend repatriations is a vast overestimate and is closer to the $1-2 billion estimated in earlier papers, including one by the authors. The discussion of the costs of deferral also fails to consider the various strategies companies can engage in to mitigate them. Indeed, throughout the paper, Desai and Hines assume that companies are passive, naive victims of government policy without any ability to engage in tax avoidance strategies. In addition to their not accurately identifying the companies' income that actually results from operations abroad, the Desai-Hines assertions about the expenses that should be linked to this income have no justification. They implicitly assume that parent overhead expenses like interest or R&D never make any contributions to profits abroad.

The ownership neutrality concept does not clarify how foreign income should be taxed, because it arbitrarily assumes that U.S. investment abroad, to exploit a patent, for example, never has any effect on production anywhere else. Finally, the $20 billion third component is probably an underestimate of the gain in after-tax profits by U.S. companies as a result of exempting all foreign income, not because of the exemption of dividends but of all other foreign income as well. It goes far beyond even the Desai-Hines concept of ownership neutrality because U.S. companies would be put in a position much superior to their foreign competitors. No industrialized country exempts royalties and interest. Exempting all foreign income would not just eliminate the U.S. tax burden on foreign investment; it would create a huge subsidy for foreign investment and result in a major migration of the U.S. corporate tax base abroad.

Before going into more detail, it will be helpful to review the basics of the U.S. system for taxing 'foreign' business income. The United States imposes the corporate tax on all repatriated foreign income, which includes not only dividends, but also interest, royalties and other foreign payments such as compensation for services provided abroad. It also includes income defined as foreign source under the U.S. rules, even though, as explained below, it has no connection with any U.S. business activity abroad. The repatriation of equity income from an active business can be deferred, but, under subpart F in the Internal Revenue Code, the deferral privilege is not extended to certain 'tainted' income, including passive portfolio income received by foreign subsidiaries and the income from purely trading operations in tax havens.

Taxpayers receive a credit against this tentative U.S. tax for foreign taxes paid on the income, including a credit for the underlying foreign corporate tax linked to a direct dividend, but this credit is limited to what the U.S. tax would have been on the income. Furthermore two steps are important in this foreign tax credit limitation calculation. In the first, the foreign income is separated into 'baskets' to reduce cross-crediting, i.e., credits flowing over from a type of income that may be highly taxed to a type that has been lightly taxed. The three important baskets are general nonfinancial active income, financial services income, and passive portfolio income received by controlled foreign corporations. Within any basket, excess credits generated by one type of income (e.g., dividends) can flow over to other income (e.g., royalties) in the basket. In the second step, parent overhead expenses (e.g., interest) are allocated to each basket in order to calculate the net foreign source income on which the credit in the basket can be taken. This affects companies only if they cannot use all of the credits for the foreign taxes they have paid. With this introduction, we can proceed to review each step in the Desai-Hines analysis.

WHAT TYPE OF FOREIGN INCOME IS THE CURRENT REVENUE DERIVED FROM?

Using data in Raub (2003), Desai and Hines estimate that the U.S. tax actually collected on foreign income in 1999 was $20 billion. This is based on income of $166 billion, taxed at the corporate rate of 35 percent, and credits of $38 billion. But as is clearly stated in Raub (2003), the $166 billion is "before adjustments" for domestic losses. (1) Any measure of the tax burden on a given type of income should reflect the possibility that it can absorb losses from another type. We can evaluate the significance of the adjustments using tabulations of the U.S. Treasury data files for 2000. Income was $196.7 billion before and $174.6 billion after adjustments. The latter is the foreign income that is actually taxed. Accordingly, about one third of the Desai-Hines $20 billion results from this use of the unadjusted data rather than the adjusted data. The total residual tax on foreign income in 2000 was $12.7 billion, reflecting the $174.6 billion in income and $48.4 billion of allowable foreign tax credits. But the main issue here is the composition of the $12.7 billion. (2)

Desai and Hines effectively assume that all of the $20 billion of actual collections from foreign source income constitutes the residual tax on dividends. Their $10 billion of indirect efficiency loss arising from retained earnings is based on this erroneous assumption. For example, they state, "Conservatively, the $20 billion estimate can be increased by 50 percent, to incorporate the effect of taxes owed on unrepatriated earnings and the differing average rates of taxation on repatriated and unrepatriated income." In fact, dividends account for a relatively small amount (no more than 15 percent) of current revenue. For the remainder, deferring income was not an option. Let us, therefore, look at how foreign income and the tax derived thereon are distributed.

Of the $12.7 billion residual tax in 2000, the U.S. tax on the largest income basket, for general nonfinancial active income, was $5.6 billion. Of this $5.6 billion collected, only about $1.3 billion is obtained from dividends. The vast majority of dividends bear no residual tax, because either the parent has excess credits or the dividend carries a credit greater than the 35 percent U.S. tax rate. That leaves only low-tax dividends received by parents without excess credits. The remaining $4.3 billion collected on the income in this basket arises from royalties and interest received from active affiliates, both of which are deductible from host country taxable income, as well as foreign branch income and export sales source income. Under the U.S tax rules, 50 percent of export sales income can be classified as foreign source. Companies benefit from this provision because the export income can be shielded from U.S. tax by excess foreign tax credits, but a substantial share still remains taxable. Indeed, it accounts for $1.5 billion of the revenue in the nonfinancial active basket (more than dividends).