The foreign source income repatriation patterns of US parents in worldwide loss.

INTRODUCTION

On average 40 percent of US corporations with foreign source income are not taxable because they are in loss, and annually these loss firms are responsible for about 13 percent of foreign source dividend repatriations. (1) Yet the repatriation behavior of these US parents in loss has largely been overlooked, both in policy analysis and in the economic literature. This paper begins to address this oversight by carefully describing the repatriation incentives and behavior of loss firms. In doing so, we are expanding the preliminary analysis of the repatriation behavior of loss firms put forth by Altshuler and Newlon (1993) (hereafter A&N).

The main goal of this paper is to improve our understanding of how loss status impacts dividend repatriations. This study is unique in that it uses a balanced panel of US income tax returns from 1998 to 2002 in order to examine the repatriation behavior of US parents in loss. (2) The panel data allows us to relate the repatriation behavior of individual US parents to their income status. Firms in loss generally repatriate less foreign source income than firms with taxable income. This is true even though firms in loss may be able to repatriate foreign source income with no "residual" US tax and, thus, a zero "tax price." (3) The panel is used to determine whether the lower repatriation by loss firms is only prevalent for "chronic loss firms" or if it also pertains to a loss year for firms that are occasionally in loss. The former would suggest consistent repatriation behavior across income and loss states but systematic differences across US parents, while the latter would suggest systematic differences in repatriation for a given parent across its income and loss experiences.

The paper proceeds as follows. The second section reviews the importance of repatriations of US parents in loss in the context of empirical international tax research and policy analysis, and attempts to characterize the tax incentives facing loss firms. The third section describes the data sets used for this analysis. The fourth section provides an initial data analysis of the repatriation behavior of loss firms and describes limitations of the study, and the fifth section offers preliminary conclusions.

IMPORTANCE OF REPATRIATION OF LOSS FIRMS

Repatriated foreign source dividends are a key component of foreign source income, comprising about 55 percent of repatriated earnings in any given year, and have been of great interest to researchers and policy makers alike. (4) Because US parents can generally choose the level and timing of repatriations from their foreign subsidiaries, researchers have focused on the relationship between dividend repatriation and taxes, and have tried to ascertain if taxes affect dividend repatriation behavior. Because dividend repatriations are a key source of foreign tax credit usage, policy makers and revenue estimators are concerned with the precise level of repatriated dividends (as well as other foreign source income and foreign tax payments) so that accurate estimates of foreign tax credit usage in current law and under alternative policy scenarios can be determined. However, in both the research and policy arenas, the foreign source income repatriations of US parents in loss have largely been ignored. In this section we will highlight why the repatriations of loss firms are important and place this information in the context of the existing literature and analysis.

The "new view" analysis of the relationship between taxes and foreign source income repatriation has a 20 year history. (5) Hartman (1985) asserts that in theory repatriation taxes on foreign source income--if constant over time--should not affect the dividend payout decisions from a mature controlled foreign corporation (CFC) to its US parent and, in his model, should not affect the investment made by the subsidiary out of its retained earnings. (6) However, several microeconomic studies (Goodspeed and Frisch, 1989; Hines and Hubbard, 1990; Altshuler and Newlon, 1993) have tested the relationship and found negative correlations between taxes and repatriations in practice. (7)

Altshuler, Newlon, and Randolph (1995) reconcile the theory with the empirical findings by demonstrating that in practice the repatriation tax price facing individual firms varies over time due to changes in the firms' foreign tax credit positions, and it is these transitory components of the tax price that are negatively correlated with dividend repatriations. That is, when a US parent faces a "residual" US tax on its repatriated foreign earnings, the "tax price" is higher and the parent is less likely to repatriate. Foreign tax credit position refers to the relationship between US tax liability on foreign source income and the foreign tax credits available to offset that liability. A corporation whose US tax liability on foreign source income exceeds its supply of foreign tax credits is said to be "in excess limit," while a US corporation whose foreign tax credit sup ply exceeds its US tax liability on foreign source income is said to be "in excess credit."

The Hartman and Sinn model generalizes the original Hartman model to allow for multiple modes of income repatriation and demonstrates that the tax price of the lowest cost option does not affect investment out of retained earnings. (8) Grubert (1998) extends all previous work and provides loose support for the Hartman and Sinn model by analyzing repatriation incentives and behavior in a world in which there are multiple repatriation vehicles, and finds that taxes have a large and significant effect on the composition of repatriations. (9) Finally, Altshuler and Grubert (1998, 2001) perform theoretical and empirical analyses that consider more sophisticated repatriation alternatives, and again find evidence that taxes impact repatriation behavior. Thus, accurately understanding the relationship between taxes and foreign source income repatriations is important for many reasons. (10)

A&N perform the only separate analysis of loss firms to date using a 1984 cross section for a limited subset of US parents in loss. (11) They describe the apparent paradox that these loss firms appear somewhat less likely to repatriate dividends than their taxable counterparts. (12) This is a paradox because in principle loss parents face increased repatriation incentives because the fact that they are in loss means that they can temporarily avoid the "residual" US tax that normally accrues on dividends repatriated from a low tax jurisdiction. A&N postulate that the net operating loss deductions (NOLDs), which these parents must forego in order to repatriate dividends, are more valuable to them than the delayed residual tax payments and the extra foreign tax credits (FTC), which would accrue as a result of the repatriation. (13)

A&N go on to mention that NOLDs could be more valuable because their carryover period is longer than the FTC carryover period. We note further that NOLDs can be used to offset tax liability on US source income, whereas foreign tax credits can only offset liability on foreign source income, and, as long as the corporation has positive income in one of the two prior years, it can be carried back for an immediate tax refund. (14) By contrast, foreign tax credit usage can be quite restricted under current law due to the existence of nine different "baskets," each of which has separate FTC usage restrictions with special loss allocation rules across those baskets. In any case, if the dividend repatriation behavior of US parents in loss is systematically different than parents with income, then the formal analysis of the relationship between taxes and repatriation perhaps ought to be expanded to include the unique incentive facing US parents in loss. (15) Though this might not change the conclusions, it could improve the accuracy of the empirical estimates and remove any existing bias in the estimates in the studies that omit US parents in loss. (16)

In this study we examine the relationship between dividend repatriation and loss status using more complete panel data than was used by A&N. By doing so, we hope to determine whether the apparent bias against "loss repatriation" is pervasive, if it exists over time, and whether it exists for all firms or only a subset of firms.

DATA SOURCES AND ISSUES

The sources of information concerning foreign source income for corporations used in this paper are the US corporate tax return (Form 1120) and the information return filed by each controlled foreign corporation (CFC) (Form 5471) of a US parent. Form 1120, Schedule C, contains separate information on voluntary dividend repatriations and "deemed" dividend repatriations from foreign sources, but combines other forms of foreign source income such as interest, royalties or branch income with domestic source income. (17) Form 1120 also reports total foreign tax credits used in any given year.

Form 5471 is filed by a US parent for each CFC and contains the most comprehensive, detailed information on foreign source income of CFCs including the subsidiaries' earnings and profits, dividend, interest, rent and royalty repatriations to its US parent, and other information. However, complete information (including the items mentioned above) is only edited for the top 7,500 subsidiaries in even years. There is a surprising amount of turnover in top 7,500 subsidiaries so that less than half of the CFCs remain in the top 7,500 over a six year time window. Finally, Form 5471 only contains information on repatriations of related parties.