Examining Investor Expectations Concerning Tax Savings on the Repatriations of Foreign Earnings under the American Jobs Creation Act of 2004

The American Jobs Creation Act of 2004 was signed into law on October 22, 2004. One of the most significant aspects of this legislation is a temporary tax holiday for dividend repatriations from foreign subsidiaries. U.S. multinational corporations may elect during a one-year window to deduct 85 percent of extraordinary cash dividends received from foreign subsidiaries. In this study, we model the impact that this legislation has on a firm’s decision to either repatriate or reinvest foreign earnings from abroad. We then examine investors’ assessment of how U.S. multinational corporations will respond to the temporary tax holiday. Our results indicate that investors repriced the tax liability consistent with investors anticipating that U.S. multinational corporations will repatriate a significant portion of their permanently reinvested foreign earnings during the tax holiday. INTRODUCTION On October 22, 2004 the President signed into law the American Jobs Creation Act of 2004 (AJCA 2004, U.S. Joint Committee on Taxation 2004). The AJCA 2004 involves numerous changes to the existing tax law, among the most notable being a dividend repatriation tax holiday. U.S. multinational corporations (MNCs) have a oneyear window during which they can deduct 85 percent of qualifying dividends repatriated from their foreign subsidiaries. The primary beneficiaries of this dividend repatriation tax holiday are U.S. MNCs with significant accumulated foreign earnings that were previously taxed at a low foreign rate relative to their U.S. domestic earnings. Initial research and media reports indicate the impact of this tax holiday on some U.S. MNCs is substantial. Albring et al. (2005) estimate these firms will save $29 billion from the dividend repatriation tax holiday under the AJCA 2004. In reference to this legislation, Mitchell Oler is an Assistant Professor at Virginia Polytechnic Institute & State University, Terry Shevlin is a Professor at the University of Washington, and Ryan Wilson is an Assistant Professor at the University of Iowa. We acknowledge helpful comments from Michelle Hanlon, Susan Albring, and participants at the 2007 JATA Conference. We also appreciate the comments of two anonymous referees and Bryan Cloyd (the editor). All remaining errors and opinions are our own. Professor Shevlin gratefully acknowledges financial support from the Paul Pigott /PACCAR Professorship and Mr. Wilson from the Deloitte Foundation. Editor’s note: This paper was presented at the 2007 Journal of the American Taxation Association Conference. Submitted: July 2006 Accepted: March 2007 26 Oler, Shevlin, and Wilson The Journal of the American Taxation Association, Fall 2007 the Wall Street Journal reported on October 5, 2005 that ‘‘nine months into 2005, U.S. companies have announced plans to repatriate about $206 billion in foreign profits under a special one-year tax break’’ (Aeppel 2005). Under current U.S. tax law, earnings from foreign subsidiaries are not taxed in the U.S. until a dividend is repatriated back to the U.S. parent. However, to avoid double taxation, firms receive a tax credit for foreign taxes paid, subject to limitations. Thus, in effect, when a firm repatriates income from a low-tax country, it is required to pay only the difference between the U.S. tax rate and the foreign tax rate to the U.S. government. The temporary tax holiday in the AJCA 2004 results in firms being taxed in the U.S. at a maximum effective rate of 5.25 percent on repatriated foreign earnings as opposed to the top U.S. corporate statutory rate of 35 percent. This study examines investors’ assessments of how U.S. MNCs might respond to this tax holiday before the firms actually announce the extent of the foreign earnings they intend to repatriate during the holiday. Specifically, we estimate market valuation and stock return regressions to determine whether investors anticipate that firms with low-taxed offshore earnings will take advantage of this tax holiday and repatriate their foreign earnings at a significantly reduced tax rate. For financial reporting purposes, U.S. MNCs can designate foreign subsidiary earnings as ‘‘permanently reinvested’’ under APB Opinion No. 23 (Accounting Principles Board 1972). Permanently reinvested earnings (PRE) are earnings from foreign subsidiaries that have been invested abroad and that managers intend to reinvest indefinitely or that managers intend to remit in a tax-free liquidation. Firms with an average foreign tax rate below the statutory U.S. tax rate are able to defer recognizing the residual U.S. tax on their foreign earnings for financial reporting purposes by designating them as permanently reinvested. Despite this designation, previous research by Collins et al. (2001) is consistent with the market capitalizing the deferred repatriation tax liability into current stock prices for foreign earnings that managers have designated as permanently reinvested. This result suggests the market does not find the claim that these foreign earnings are permanently reinvested to be credible, and anticipates these firms will eventually pay the residual repatriation tax when those earnings are brought home to the U.S. Dhaliwal and Krull (2006) also investigate whether investors incorporate the deferred tax liability on PRE into stock prices. In contrast to Collins et al. (2001), Dhaliwal and Krull (2006) find that after controlling for size, the deferred repatriation tax on PRE does not directly affect market value. We employ a market-value (or price-level) regression model and supplement the analysis with a returns approach, which is less subject to econometric problems. We examine 1 Earnings of U.S.-owned foreign branches and subpart F earnings (basically earnings arising from passive activities such as investments in financial assets) are subject to U.S. taxation in the period earned. 2 We discuss foreign tax credits in more detail below. A firm may use excess foreign tax credits from subsidiaries based in high-tax countries to offset additional U.S. taxes on dividends from low-tax countries. 3 Given a top U.S. statutory rate of td 35%, the 85 percent deduction for dividends from foreign earnings results in a maximum effective rate of 5.25% (td tƒ) (1 85%)) (35% 0%) (1 85%)), assuming no foreign tax credits because tƒ 0. 4 Statement of Financial Accounting Standards No. 109 requires that firms disclose the amount of tax related to permanently reinvested earnings if it is material and if it is ‘‘practicable’’ to determine the liability. Many of the firms in our study do not disclose the extent of the tax liability associated with permanently reinvested earnings, indicating that the determination of such a liability is not practicable. 5 Note that in addition to using different specifications of the market-value regression model, the Collins et al. (2001) and Dhaliwal and Krull (2006) papers examine very different samples of U.S. MNCs. Collins et al. (2001) examine a set of U.S. MNCs reporting PRE in 1993. The Dhaliwal and Krull (2006) study examines a larger sample of firms reporting PRE, positive foreign assets, and foreign sales from 1993 to 1999. 6 We also check our results using both the Collins et al. (2001) and Dhaliwal and Krull (2006) model specifications and discuss these results below. Our inferences on the market pricing of the deferred repatriation tax on PRE are the same under both specifications. Investor Expectations Concerning Repatriation Tax Holiday 27 The Journal of the American Taxation Association, Fall 2007 U.S. MNCs with low-taxed offshore earnings to assess how the market pricing of this deferred tax liability changes when passage of a temporary tax holiday for dividend repatriations becomes probable. Specifically, we test whether, consistent with Collins et al. (2001), the market capitalizes into current stock prices the unrecognized deferred tax liability associated with PRE generated in low-tax jurisdictions prior to the time when passage of the tax holiday was probable. We then examine the same set of firms after passage of the tax holiday became probable to determine whether there is a significant reduction in the capitalization of the deferred tax liability associated with unrepatriated foreign earnings. Such a reduction would be consistent with the market anticipating that the firms will repatriate a significant portion of their permanently reinvested earnings during the window provided by the AJCA 2004. Thus, we provide additional evidence on whether investors capitalize expected but unrecognized repatriation taxes on earnings designated as permanently reinvested using an interrupted time-series setting. We begin with a simple model of a firm’s decision whether to reinvest foreign earnings abroad or to repatriate earnings to the U.S. (from Scholes et al. 2005) and extend the model to show how the decision changes with the introduction of the tax holiday under the AJCA 2004. We demonstrate the incentive that the AJCA 2004 provides for firms to repatriate foreign earnings during the tax holiday. Based on this model, we provide a numerical example that illustrates the incentive to repatriate within the window. However, the model also shows that as the firm’s investment horizon increases, the incentive to repatriate is substantially reduced. In addition, some firms might face attractive foreign investment opportunities that provide a rate of return sufficient to make repatriation under the Act unattractive. In short, the decision to repatriate during the holiday is going to be a function of a firm’s foreign investment opportunities and investment horizon. That said, the magnitude of the savings and the difficulties associated with bringing the earnings home tax-free make it very likely that repatriation during the holiday will be the most efficient choice for most firms. We test whether, consistent with this incentive, the market anticipates that the majority of U.S. MNCs subject to average foreign tax rates below the U.S. statutory rate will choose to repatriate a significant portion of their PRE during 2005. On October 27, 2003, Corporate Financing Week reported that ‘‘proposed legislation allowing companies to bring back earnings held in foreign subsidiaries to the U.S. appears to be gaining momentum in Congress.’’ The article quotes a tax and accounting analyst from Lehman Brothers as indicating the legislation was ‘‘gaining momentum very quickly and that passage in the first quarter of 2004 is highly likely.’’ A review of the Lexis-Nexis academic research database turned up no articles prior to 2003 referring to similar legislation as having a high probability of passing. For this reason, we focus our tests on examining the change in the extent to which the market capitalizes the residual repatriation tax at the 2003 and 2004 fiscal year-ends. Specifically, we analyze the change in the market’s pricing of the repatriation tax from the 2001 and 2002 fiscal year-end, before passage of the tax holiday was probable, to the 2003 and 2004 fiscal year-ends at which point passage of the holiday was probable. 7 IRC Code Section 965, added as a result of AJCA 2004, allows firms to take a one-time deduction in either 2004 or 2005 (for calendar year firms) of 85 percent of foreign earnings repatriated to the U.S. parent company through an extraordinary cash dividend. Due to the complexity of the new provision, combined with uncertainty regarding a number of issues related to the 85 percent dividend, the majority of firms chose to defer the use of the one-time deduction until 2005. We have eliminated from our sample any firms that chose to repatriate earnings under this provision in 2004. Omission of these firms from our analysis allows us to examine investors’ assessment of how firms will respond to the AJCA 2004 before the firms actually announce specifics regarding the amount of PRE they intend to repatriate during the tax holiday. 28 Oler, Shevlin, and Wilson The Journal of the American Taxation Association, Fall 2007 The results from our price-level regression tests are consistent with investors anticipating firms will repatriate a significant portion of their PRE during the window provided by the AJCA 2004. Consistent with Collins et al. (2001), we find that prior to the time when passage of the tax holiday became probable, investors capitalized into current stock prices the unrecognized deferred tax liability associated with unrepatriated foreign earnings designated as permanently reinvested earnings. However, for the periods after passage of the tax holiday became likely, we find a significant reduction in the extent to which investors capitalize the deferred repatriation tax into current stock prices. The results of our stock return regression tests are consistent with the price-level regressions: The estimated tax savings from the tax holiday are significantly positively associated with stock returns in 2003. We also run supplemental price-level tests to examine whether investors price the deferred repatriation tax as a function of a firm’s probability of repatriation during the holiday. Our results are consistent with investors cross-sectionally adjusting their pricing as a function of the probability of repatriation. Our findings are particularly notable because none of the firms in our sample had announced the extent of the PRE they intended to repatriate during the window. Our findings suggest that the market is relatively efficient in incorporating expectations about firms responding to the tax cut on repatriations and that the market expected firms to be responsive to the tax cut. Our results indicate a fairly sophisticated understanding on the part of investors of the factors a firm must consider in making the decision about whether to repatriate or reinvest, and of how the AJCA 2004 impacted that decision. The remainder of the paper proceeds as follows. The second section provides a discussion of the key provisions associated with the dividend repatriation tax holiday. The third section develops a simple model of a firm’s decision on whether to repatriate or reinvest foreign earnings from abroad and the impact that AJCA 2004 has on that decision. The fourth section details our sample selection and descriptive statistics. The fifth section presents our research design and findings. The last section concludes. KEY PROVISIONS OF THE REPATRIATION TAX HOLIDAY Pursuant to IRC Section 965, a provision of the AJCA 2004, firms are eligible for the 85 percent dividends received deduction (DRD) on cash dividends from controlled foreign corporations (CFCs). A CFC is a foreign subsidiary (i.e., a subsidiary that is located in some foreign country) of a U.S. corporation. However, this dividend is limited to the greater of $500 million or the amount shown as earnings permanently reinvested outside the U.S. on the firm’s financial statements. If the extent of earnings permanently reinvested outside the U.S. is not disclosed, but the tax liability attributable to such earnings is disclosed, then 8 This result is not consistent with Dhaliwal and Krull (2006) who find that PRE or deferred taxes on PRE are not significantly associated with stock prices. However, Dhaliwal and Krull (2006) examine a larger set of sample firms collected from 1993 to 1999. We restrict our analysis to firms identified by Albring et al. (2005) as having PRE $500 million in 2002. Our sample has an average PRE of $3,362 million, while the Dhaliwal and Krull (2006) sample has an average PRE of $401.08 million. We focus on a set of firms with very large PRE that are likely to benefit most from the passage of the AJCA. We believe our analysis of this set of firms is both interesting and economically significant. This difference in sample composition however could explain the differential results regarding the significance of the TAX variable in the two papers. 9 Shane and Stock (2006) investigate the extent to which investors correctly interpret the temporary reported book income effects of tax-motivated income shifting around the Tax Reform Act of 1986. They report evidence consistent with analysts and investors failing to correctly identify and price these temporary effects. This finding would seem to contrast with our finding that investors are relatively sophisticated in examining the impact of the repatriation tax holiday. However, Shane and Stock (2006) examine a setting where managers likely took deliberate care to conceal their earnings management from detection. In contrast, we examine a situation where managers have simply not announced how they will react to a major change in the tax law. Investor Expectations Concerning Repatriation Tax Holiday 29 The Journal of the American Taxation Association, Fall 2007 the financial statement amount is set equal to this liability divided by 0.35. For firms that file financial statements with the Securities and Exchange Commission (SEC), the applicable financial statement is the most recent audited financial statement filed on or before June 30, 2003. For firms that do not file with the SEC, it is the most recent financial statement certified on or before June 30, 2003. In addition to the requirements limiting the amount of the dividend eligible for the DRD, the dividend eligible for the benefit must be extraordinary. Firms that have been receiving dividends from their CFCs will only benefit from the provision to the extent the average of such annual dividends is increased by the dividend repatriated during the tax holiday. A base period amount is calculated as the average dividend received by the firm during three of the previous five taxable years ending before June 30, 2003. If there are less than five taxable years available, then the base period includes all taxable years. Any cash dividend received during the tax holiday that is in excess of the base amount is eligible for the DRD. Another critical provision of Section 965 stipulates that the DRD is available only to the extent that cash dividends received from a CFC are reinvested in the U.S. Section 965(b)(4) states that permitted investments include: (1) funding of working, hiring, and training (other than executive compensation), (2) infrastructure, (3) research and development, (4) capital investments, and (5) financial stabilization of the corporation for purposes of job retention or creation. The domestic reinvestment plan adopted by the firm must be approved by the CEO prior to the date the cash dividend is paid and must subsequently receive the approval of the board of directors. Many of the U.S. MNCs examined in this study indicate in their 2004 financial statements that they are awaiting technical guidance related to these reinvestment requirements before they could estimate the amount of earnings they intend to repatriate under the tax holiday. FASB issued Staff Position 109-2 in December 2004 to provide disclosure guidance related to the repatriation of foreign earnings under AJCA 2004. Pursuant to FSP 109-2, firms must disclose any planned repatriation, or if they are still in the process of evaluating repatriation under AJCA 2004, they must disclose the range of amounts being considered. The majority of the firms in our sample disclosed a broad range for the potential amount of PRE that they were considering repatriating during the tax holiday. The disclosed range often began with zero, giving investors little specific guidance on the extent to which the company intended to take advantage of the tax holiday. For example, Agilent Technologies, Inc. made the following disclosure related to AJCA (2004) in the tax footnote of its 2004 Form 10-K: On October 22, 2004, the AJCA was signed into law. The AJCA includes a deduction for 85 percent of certain foreign earnings that are repatriated, as defined in the AJCA, at an effective tax cost of 5.25 percent on any such repatriated foreign earnings. Agilent may elect to apply this provision to qualifying earnings repatriations in fiscal 2005. Agilent has begun an evaluation of the effects of the repatriation provision; however, we do not expect to be able to complete this evaluation until after Congress or the Treasury Department provide additional clarifying language on key elements of the 10 As noted by Albring et al. (2005) and Blessing (2004), this approach would significantly understate the amount of permanently reinvested earnings because the deferred tax would be recorded at the incremental tax rate net of any foreign tax credits as opposed to a rate of 35 percent. 11 Based on analysis of investment alternatives implied by the firms’ decision to accumulate foreign earnings overseas, Blouin and Krull (2006) predict that firms will actually use any repatriated funds under AJCA 2004 to repurchase shares. They provide evidence consistent with their predictions. 30 Oler, Shevlin, and Wilson The Journal of the American Taxation Association, Fall 2007 provision. We expect to complete our evaluation of the effects of the repatriation provision within a reasonable period of time following the publication of the additional clarifying language. The range of possible amounts that Agilent is considering for repatriation under this provision is between zero and $970 million. The related potential range of income tax is between zero and $51 million. We assume that any change in the extent that the market capitalizes the deferred tax on unrepatriated foreign earnings is driven exclusively by the repatriation tax holiday and not by other provisions in the AJCA 2004. However, the AJCA 2004 contained two other major provisions that might impact our tests. One of these provisions was a reduction in the number of income ‘‘baskets’’ used to calculate a firm’s foreign tax credit limitation. This change allows firms more flexibility in shielding income earned in low-tax jurisdictions from the repatriation tax. There are, however, some key differences between the repatriation tax holiday and the change in the calculation of the foreign tax credit limitation. The reduction in the number of baskets will make it easier for firms to shield earnings in lowtax countries from the repatriation tax using earnings from high-tax countries. For purposes of our study, we calculate our estimate of the repatriation tax using an average foreign tax rate. Only firms with an average foreign tax-rate below the U.S. statutory rate have a positive repatriation tax liability in our analysis. It is unlikely that these firms would have sufficient earnings from high-tax countries to significantly offset their foreign earnings from low-tax countries; otherwise their average foreign tax rate would not be below the U.S. statutory rate. This observation does not rule out a possible benefit for these firms associated with the reduction in baskets; however, this benefit is likely to be relatively small in comparison to the benefit provided by the repatriation tax holiday. The AJCA 2004 also includes a provision for a domestic manufacturing deduction (DMD) that allows manufacturing firms to deduct 3 percent (which increases to 6 percent in 2007 and 9 percent in 2010) of qualified production income (subject to limitations). The DMD will increase the domestic after-tax rate of return, rd, starting in 2005 (assuming input prices are not bid up too much such that after-tax rates of return to manufacturing are unchanged). With an expected increase in rd, firms will be encouraged to repatriate during the tax holiday and this reinforces our prediction that investors will view the tax holiday positively. In an effort to control for the possible impact of the DMD provision, we directly test whether, in a price-level regression, the multiplier on domestic net income changed after passage of the AJCA became probable. The results (not tabulated) are not consistent with the market increasing the multiplier on domestic income as a result of the DMD provision. In fact, we actually see a decline in the market’s pricing of domestic net income during the AJCA period. Some of the benefit associated with the introduction of the DMD provision might have been mitigated by the repeal of the Extraterritorial Income Exclusion (ETI), which was eliminated as a part of the AJCA 2004. Specifically, the AJCA 2004 reduced ETI benefits to 80 percent in 2005, 60 percent in 2006, and after 2006 the ETI no longer exists. Regardless of what caused the decline in the pricing of domestic net income, our inferences related to the pricing of the deferred repatriation tax remain unchanged. THE DECISION TO REPATRIATE OR REINVEST FOREIGN EARNINGS In choosing whether to repatriate foreign earnings U.S. MNCs must consider the taxes that would be paid if the earnings were repatriated currently, future taxes that would be paid on the earnings if repatriation is deferred, and any implicit taxes that would be paid by choosing a tax-favored option that has a lower pretax rate of return. The U.S. government Investor Expectations Concerning Repatriation Tax Holiday 31 The Journal of the American Taxation Association, Fall 2007 taxes U.S. firms on their worldwide income, however foreign subsidiaries of U.S. MNCs are not included on the consolidated U.S. tax return. The earnings from these foreign subsidiaries are not taxed until the cash is repatriated to the U.S. parent. When the earnings are repatriated, they are taxed at the U.S. statutory rate and the firm receives a credit for foreign taxes paid. In choosing to reinvest foreign earnings abroad, a firm avoids paying the residual U.S. tax until the foreign earnings are repatriated to the U.S. The following set of equations, adapted from Hartman (1985) and Scholes et al. (2005), models the decision a firm faces to repatriate or reinvest. In the equations below, td is the U.S. tax rate, tƒ represents the foreign tax rate, rd is the U.S. after-tax rate of return, and rƒ represents the foreign after-tax rate of return. 13 Letting DIV represent the amount of dividends that are repatriated in the current period, the amount remaining after paying the home country tax, in this case the U.S. tax, is: DIV DIV(1 t ) d DIV (t t ) . (1) d ƒ 1 t 1 t ƒ ƒ The second term on the left-hand side of Equation (1) denotes the calculation of additional U.S. taxes. DIV is grossed up by (1 tƒ) to derive an estimate of foreign taxable income which is then multiplied by the U.S. tax rate (to estimate U.S. taxes) less the amount of the foreign tax credit (foreign taxable income times tƒ representing the foreign taxes paid on the foreign income). If the firm then chooses to reinvest this amount in the U.S. for n periods at an after-tax rate of return of rd, then the accumulation in n periods is: DIV(1 t ) d n (1 r ) . (2) d (1 t ) ƒ However, if the firm instead chooses to reinvest the earnings and profits abroad for n periods and then repatriate, then the accumulation in n periods is: n DIV(1 r ) DIV(1 t ) ƒ d n n DIV(1 r ) (t t ) (1 r ) . (3) ƒ d ƒ ƒ 1 t 1 t ƒ ƒ In comparing Equations (2) and (3), the only difference is in the last term (1 rd) n versus (1 rƒ) . As a result, the decision to reinvest hinges on whether the after-tax rate of return on foreign investment, rƒ, exceeds the after-tax rate of return on home country 12 DeWaegenaere and Sansing (DS 2006) develop a more complex model than the model presented here. Our model assumes a finite-lived subsidiary with eventual repatriation of foreign earnings, whereas DS assume foreign subsidiaries investment in operating assets are infinite-lived allowing for the possibility of permanent deferral of repatriation tax on the earnings of its operating assets. Additionally DS separate out investments in operating assets and financial assets and among growth firms (the assumption in our model) and mature firms where (rd rƒ) and any foreign reinvestment is in financial assets. While such separation allows analytical insights into firms’ repatriation decisions and on possible coefficient values on deferred tax liabilities (DTL) and permanently reinvested earnings arising from operating assets and financial assets, data on PRE and DTL decomposed by asset type are neither disclosed nor available. Thus, we use our simpler model to make our point. However, consistent with the DS analysis, in sensitivity analysis we examine the pricing of the repatriation tax within the holiday period as a function of excess cash holdings. 13 Consistent with existing U.S. tax law, this model assumes the home country has a worldwide tax system with foreign tax credits rather than a territorial system. This model also assumes the home-country tax rate, td, exceeds the foreign-country tax rate, tƒ. Otherwise, no home-country tax would be due upon repatriation. 14 DIV / (1 tƒ) is equal to foreign source income (FSI), and the foreign tax credit (FTC) is limited to the min (FSI td, FSI tƒ). 32 Oler, Shevlin, and Wilson The Journal of the American Taxation Association, Fall 2007 investment, rd. Interestingly, the decision to repatriate is not impacted by the extent of the repatriation tax or the investment horizon because the firm bears the cost of the repatriation tax irrespective of whether it chooses to repatriate now or to reinvest and repatriate at some point in the future. If a firm did not repatriate foreign earnings before the AJCA 2004, then we can infer that rd rƒ. The results of the above model rely on two critical assumptions. The first assumption is that the foreign earnings will eventually be repatriated and subject to the tax on repatriation. This assumption appears to be supported by the work of Collins et al. (2001), who find the market capitalizes into stock prices the unrecognized deferred tax liability associated with unrepatriated foreign earnings despite management’s designation of the earnings as permanently reinvested. Scholes et al. (2005) discuss a number of ways or mechanisms that firms might use to bring foreign earnings home tax-free. However, our sample firms have already reported these earnings as taxable income in the foreign jurisdiction so it is not possible to employ the usual mechanisms (transfer pricing, interest payments to the U.S. parent on debt, royalties, and other payments to the U.S. parent for services provided) because these mechanisms affect future profits, not past reported profits. Thus the ability to avoid U.S. taxation on already reported foreign earnings is somewhat limited. Further, these mechanisms to transfer future profits usually relate to firms wishing to transfer profits from high tax jurisdictions to low tax jurisdictions whereas in our setting the U.S. tax rate is higher than the foreign tax rate (otherwise there would not be any additional U.S. tax due on repatriation). The second assumption of our model is that the tax price of repatriation remains constant over time; i.e., there is no time subscript on td and tƒ—they are assumed intertemporal constants. The introduction of the AJCA 2004 obviously violates this assumption. The AJCA 2004 allows MNCs to repatriate at a significantly reduced rate of a maximum 5.25 percent because 85 percent of the dividend is excluded from additional U.S. taxation. As a result, the tax price of dividend repatriation now impacts the final result, and firms face the following decision: If the firm decides to repatriate under the AJCA 2004, then the amount remaining after paying the U.S. tax is: DIV DIV 15% * (t t ) . (4) d ƒ 1 tƒ If the firm then chooses to reinvest the amount repatriated during the tax holiday in the U.S. for n periods at an after-tax rate of return of rd, then the accumulation in n periods is DIV n DIV 15% * (t t ) (1 r ) . (5) d ƒ d 1 tƒ 15 The model is obviously a simplification. For example, by designating foreign earnings as PRE, firms do not have to accrue or recognize the additional U.S. taxes due on repatriation. Thus managers can ‘‘manage’’ reported after-tax earnings by increasing or decreasing the amount designated as PRE. Krull (2004) provides evidence that firms appear to manage reported earnings via PRE. The value to the firm (or at least the managers) of this discretion is not captured in our model. Investor Expectations Concerning Repatriation Tax Holiday 33 The Journal of the American Taxation Association, Fall 2007 Under the AJCA 2004, the firm must now compare Equation (5) to Equation (3) to determine whether to repatriate under the tax holiday. The firm will choose to repatriate under AJCA if Equation (5) Equation (3): DIV DIV(1 t ) d n n DIV 15% * (t t ) (1 r ) (1 r ) . (6) d ƒ d ƒ 1 t 1 t ƒ ƒ The decision is now a function of the current repatriation tax under the AJCA 2004, the future expected U.S. tax rate, the expected after-tax rates of return on both foreign and domestic investments, and the length of the investment horizon. In our original analysis, without the tax holiday provided by the AJCA 2004, a firm would choose to continue to reinvest foreign earnings and profits abroad if rƒ exceeded rd. A simple numerical example demonstrates the impact the AJCA 2004 has on this decision and the incentive a firm has to repatriate during the tax holiday. Suppose a firm has foreign retained earnings (PRE) $100 and has a ten-year investment horizon. Further, assume the firm has investment opportunities abroad such that its foreign pre-tax rate of return (Rƒ) is 10 percent, and assuming a 15 percent foreign tax rate, the firm’s foreign after-tax rate of return (rƒ) is 8.5 percent. In addition, assume the firm faces a domestic U.S. tax rate of 35 percent. Under this scenario, before the tax holiday provided by the AJCA 2004, the firm would have had to earn a pre-tax rate of return on earnings repatriated and reinvested domestically of at least 13.08 percent to justify repatriation, calculated as follows: 10 10 (1 r ) (1 0.085) → r 0.085 → R 0.13077. d d d However, during the tax holiday, a firm facing the same scenario would have to earn a pre-tax domestic rate of return of only 9.24 percent to justify repatriating within the window provided by the tax holiday. This required rate of return on repatriated earnings is calculated from the inequality in Equation (6) as follows: 100 100(1 0.35) 10 10 100 15% * (0.35 0.15) (1 r ) (1 0.085) ; d 1 0.15 1 0.15 r 0.06008 → R 0.09243. d d The above example illustrates the incentive provided by the tax holiday for firms to repatriate low-tax foreign earnings. Notice that before the tax holiday the firm was indifferent to repatriating or reinvesting abroad when the after-tax domestic rate of return equaled the after-tax foreign rate of return. However, during the tax holiday the firm requires an after-tax rate of return abroad that is higher than at home by 2.5 percent (6.0% 8.5%) in order to justify not taking advantage of the tax holiday. This example helps to illustrate the importance that the spread between domestic and foreign rates of return plays in determining the extent to which a firm benefits from the tax holiday. Thus, firms with aftertax foreign investment opportunities that far exceed their domestic investment opportunities 16 Note that as the investment horizon increases, the tax incentive to repatriate under AJCA 2004 is reduced. Inequality (6) can be re-written as rd (1 rƒ) [(1 td) / (1 (0.85tƒ 0.15td))] / n 1. Note that as the investment period increases the importance of the tax holiday declines, and as n approaches the reinvestment decision again becomes a function of only rd and rƒ. 34 Oler, Shevlin, and Wilson The Journal of the American Taxation Association, Fall 2007 TABLE 1 Sample Selection and Industry Distribution Panel A: Summary of the Sample Selection Criteria: (2001–2004) Total sample of firm-year observations with PRE $500 million in 2002 468 Less firm-year observations that did not disclose PRE in their annual report 41 Less firm-year observations with missing regression variables 115 Complete firm-year observation 312 Panel B: Industry Distribution of Sample Firm-Year Observations SIC Code Industry Type Number of Companies Mean MVE Mean PRE 1–999 Agriculture, Forestry, Fishing 3 1,483 1,117 1000–1999 Mining, Building 8 10,749 2,50