Container Corp. of America v. Franchise Tax Board

Justice Brennan

delivered the opinion of the Court.

This is another appeal claiming that the application of a state taxing scheme violates the Due Process and Commerce Clauses of the Federal Constitution. California imposes a corporate franchise tax geared to income. In common with a large number of other States, it employs the “unitary business” principle and formula apportionment in applying that tax to corporations doing business both inside and outside the State. Appellant is a Delaware corporation headquartered in Illinois and doing business in California and elsewhere. It also has a number of overseas subsidiaries incorporated in the countries in which they operate. Appellee is the California authority charged with administering the State’s franchise tax. This appeal presents three questions for review: (1) Was it improper for appellee and the state courts to find that appellant and its overseas subsidiaries constituted a “unitary business” for purposes of the state tax? (2) Even if the unitary business finding was proper, do certain salient differences among national economies render the standard three-factor apportionment formula used by California so inaccurate as applied to the multinational enterprise consisting of appellant and its subsidiaries as to violate the constitutional requirement of “fair apportionment” ? (3) In any event, did California have an obligation under the Foreign Commerce Clause, U. S. Const., Art. I, §8, cl. 3, to employ the “arm’s-length” analysis used by the Federal Government and most foreign nations in evaluating the tax consequences of intercorporate relationships?

i — i

A

Various aspects of state tax systems based on the “unitary business” principle and formula apportionment have provoked repeated constitutional litigation in this Court. See, e. g., ASARCO Inc. v. Idaho State Tax Comm’n, 458 U. S. 307 (1982); F. W. Woolworth Co. v. Taxation & Revenue Dept., 458 U. S. 354 (1982); Exxon Corp. v. Wisconsin Dept. of Revenue, 447 U. S. 207 (1980); Mobil Oil Corp. v. Commissioner of Taxes, 445 U. S. 425 (1980); Moorman Mfg. Co. v. Bair, 437 U. S. 267 (1978); General Motors Corp. v. Washington, 377 U. S. 436 (1964); Butler Bros. v. McColgan, 315 U. S. 501 (1942); Bass, Ratcliff & Gretton, Ltd. v. State Tax Comm’n, 266 U. S. 271 (1924); Underwood Typewriter Co. v. Chamberlain, 254 U. S. 113 (1920).

Under both the Due Process and the Commerce Clauses of the Constitution, a State may not, when imposing an income-based tax, “tax value earned outside its borders.” ASARCO, supra, at 315. In the case of a more-or-Iess integrated business enterprise operating in more than one State, however, arriving at precise territorial allocations of “value” is often an elusive goal, both in theory and in practice. See Mobil Oil Corp. v. Commissioner of Taxes, supra, at 438; Butler Bros. v. McColgan, supra, at 507-509; Underwood Typewriter Co. v. Chamberlain, supra, at 121. For this reason and others, we have long held that the Constitution imposes no single formula on the States, Wisconsin v. J. C. Penney Co., 311 U. S. 435, 445 (1940), and that the taxpayer has the “‘distinct burden of showing by “clear and cogent evidence” that [the state tax] results in extraterritorial values being taxed . . . .’” Exxon Corp., supra, at 221, quoting Butler Bros. v. McColgan, supra, at 507, in turn quoting Norfolk & Western R. Co. v. North Carolina ex rel. Maxwell, 297 U. S. 682, 688 (1936).

One way of deriving locally taxable income is on the basis of formal geographical or transactional accounting. The problem with this method is that formal accounting is subject to manipulation and imprecision, and often ignores or captures inadequately the many subtle and largely unquantifiable transfers of value that take place among the components of a single enterprise. See generally Mobil Oil Corp., supra, at 438-439, and sources cited. The unitary business/ formula apportionment method is a very different approach to the problem of taxing businesses operating in more than one jurisdiction. It rejects geographical or transactional accounting, and instead calculates the local tax base by first defining the scope of the “unitary business” of which the taxed enterprise’s activities in the taxing jurisdiction form one part, and then apportioning the total income of that “unitary business” between the taxing jurisdiction and the rest of the world on the basis of a formula taking into account objective measures of the corporation’s activities within and without the jurisdiction. This Court long ago upheld the constitutionality of the unitary business/formula apportionment method, although subject to certain constraints. See, e. g., Hans Rees’ Sons, Inc. v. North Carolina ex rel. Maxwell, 283 U. S. 123 (1931); Bass, Ratcliff & Gretton, Ltd. v. State Tax Comm’n, supra; Underwood Typewriter Co. v. Chamberlain, supra. The method has now gained wide acceptance, and is in one of its forms the basis for the the Uniform Division of Income for Tax Purposes Act (Uniform Act), which has at last count been substantially adopted by 23 States, including California.

B

Two aspects of the unitary business/formula apportionment method have traditionally attracted judicial attention. These are, as one might easily guess, the notions of “unitary business” and “formula apportionment,” respectively.

(1)

The Due Process and Commerce Clauses of the Constitution do not allow a State to tax income arising out of interstate activities — even on a proportional basis — unless there is a “ ‘minimal connection’ or ‘nexus’ between the interstate activities and the taxing State, and ‘a rational relationship between the income attributed to the State and the intrastate values of the enterprise.’ ” Exxon Corp. v. Wisconsin Dept. of Revenue, supra, at 219-220, quoting Mobil Oil Corp. v. Commissioner of Taxes, supra, at 436, 437. At the very-least, this set of principles imposes the obvious and largely self-executing limitation that a State not tax a purported “unitary business” unless at least some part of it is conducted in the State. See Exxon Corp., supra, at 220; Wisconsin v. J. C. Penney Co., supra, at 444. It also requires that there be some bond of ownership or control uniting the purported “unitary business.” See ASARCO, supra, at 316-317.

In addition, the principles we have quoted require that the out-of-state activities of the purported “unitary business” be related in some concrete way to the in-state activities. The functional meaning of this requirement is that there be some sharing or exchange of value not capable of precise identification or measurement — beyond the mere flow of funds arising out of a passive investment or a distinct business operation— which renders formula apportionment a reasonable method of taxation. See generally ASARCO, supra, at 317; Mobil Oil Corp., supra, at 438-442. In Underwood Typewriter Co. v. Chamberlain, supra, we held that a State could tax on an apportioned basis the combined income of a vertically integrated business whose various components (manufacturing, sales, etc.) operated in different States. In Bass, Ratcliff & Gretton, supra, we applied the same principle to a vertically integrated business operating across national boundaries. In Butler Bros. v. McColgan, supra, we recognized that the unitary business principle could apply, not only to vertically integrated enterprises, but also to a series of similar enterprises operating separately in various jurisdictions but linked by common managerial or operational resources that produced economies of scale and transfers of value. More recently, we have further refined the “unitary business” concept in Exxon Corp. v. Wisconsin Dept. of Revenue, 447 U. S. 207 (1980), and Mobil Oil Corp. v. Commissioner of Taxes, 445 U. S. 425 (1980), where we upheld the States’ unitary business findings, and in ASARCO Inc. v. Idaho State Tax Comm’n, 458 U. S. 307 (1982), and F. W. Woolworth Co. v. Taxation & Revenue Dept., 458 U. S. 354 (1982), in which we found such findings to have been improper.

The California statute at issue in this case, and the Uniform Act from which most of its relevant provisions are derived, track in large part the principles we have just discussed. In particular, the statute distinguishes between the “business income” of a multijurisdictional enterprise, which is apportioned by formula, Cal. Rev. & Tax. Code Ann. §§25128-25136 (West 1979), and its “nonbusiness” income, which is not.¹ Although the statute does not explicitly require that income from distinct business enterprises be apportioned separately, this requirement antedated adoption of the Uniform Act,² and has not been abandoned.³

A final point that needs to be made about the unitary business concept is that it is not, so to speak, unitary: there are variations on the theme, and any number of them are logically consistent with the underlying principles motivating the approach. For example, a State might decide to respect formal corporate lines and treat the ownership of a corporate subsidiary as per se a passive investment.⁴ In Mobil Oil Corp., 445 U. S., at 440-441, however, we made clear that, as a general matter, such a per se rule is not constitutionally required:

“Superficially, intercorporate division might appear to be a[n]. . . attractive basis for limiting apportionability. But the form of business organization may have nothing to do with the underlying unity or diversity of business enterprise.” Id., at 440.

Thus, for example, California law provides: Even among States that take this approach, however, only-some apply it in taxing American corporations with subsidiaries located in foreign countries.⁶ The difficult question we address in Part V of this opinion is whether, for reasons not implicated in Mobil,⁷ that particular variation on the theme is constitutionally barred.

“In the case of a corporation . . . owning or controlling, either directly or indirectly, another corporation, or other corporations, and in the case of a corporation . . . owned or controlled, either directly or indirectly, by another corporation, the Franchise Tax Board may require a consolidated report showing the combined net income or such other facts as it deems necessary.” Cal. Rev. & Tax. Code Ann. §25104 (West 1979).⁵

(2)

Having determined that a certain set of activities constitute a “unitary business,” a State must then apply a formula apportioning the income of that business within and without the State. Such an apportionment formula must, under both the Due Process and Commerce Clauses, be fair. See Exxon Corp., supra, at 219, 227-228; Moorman Mfg. Co., 437 U. S., at 272-273; Hans Rees’ Sons, Inc., 283 U. S., at 134. The first, and again obvious, component of fairness in an apportionment formula is what might be called internal consistency — that is, the formula must be such that, if applied by every jurisdiction, it would result in no more than all of the unitary business’ income being taxed. The second and more difficult requirement is what might be called external consistency — the factor or factors used in the apportionment formula must actually reflect a reasonable sense of how income is generated. The Constitution does not “invalidat[e] an apportionment formula whenever it may result in taxation of some income that did not have its source in the taxing State . . . Moorman Mfg. Co., supra, at 272 (emphasis added). See Underwood Typewriter Co., 254 U. S., at 120-121. Nevertheless, we will strike down the application of an apportionment formula if the taxpayer can prove “by ‘clear and cogent evidence’ that the income attributed to the State is in fact ‘out of all appropriate proportions to the business transacted ... in that State,’ [Hans Rees’ Sons, Inc.,] 283 U. S., at 135, or has ‘led to a grossly distorted result,’ [Norfolk & Western R. Co. v. State Tax Comm’n, 390 U. S. 317, 326 (1968)].” Moorman Mfg. Co., supra, at 274.

California and the other States that have adopted the Uniform Act use a formula — commonly called the “three-factor” formula — which is based, in equal parts, on the proportion of a unitary business’ total payroll, property, and sales which are located in the taxing State. See Cal. Tax & Rev. Code Ann. §§25128-25136 (West 1979). We approved the three-factor formula in Butler Bros. v. McColgan, 315 U. S. 501 (1942). Indeed, not only has the three-factor formula met our approval, but it has become, for reasons we discuss in more detail infra, at 183, something of a benchmark against which other apportionment formulas are judged. See Moorman Mfg. Co., supra, at 282 (Blackmun, J., dissenting); cf. General Motors Corp. v. District of Columbia, 380 U. S. 553, 561 (1965).

Besides being fair, an apportionment formula must, under the Commerce Clause, also not result in discrimination against interstate or foreign commerce. See Mobil Oil Corp., supra, at 444; cf. Japan Line, Ltd. v. County of Los Angeles, 441 U. S. 434, 444-448 (1979) (property tax). Aside from forbidding the obvious types of discrimination against interstate or foreign commerce, this principle might have been construed to require that a state apportionment formula not differ so substantially from methods of allocation used by other jurisdictions in which the taxpayer is subject to taxation as to produce double taxation of the same income and a resultant tax burden higher than the taxpayer would incur if its business were limited to any one jurisdiction. At least in the interstate commerce context, however, the anti-discrimination principle has not in practice required much in addition to the requirement of fair apportionment. In Moorman Mfg. Co. v. Bair, supra, in particular, we explained that eliminating all overlapping taxation would require this Court to establish not only a single constitutionally mandated method of taxation, but also rules regarding the application of that method in particular cases. 437 U. S., at 278-280. Because that task was thought to be essentially legislative, we declined to undertake it, and held that a fairly apportioned tax would not be found invalid simply because it differed from the prevailing approach adopted by the States. As we discuss infra, at 185-187, however, a more searching inquiry is necessary when we are confronted with the possibility of international double taxation.

n

5»

Appellant is m the business of manufacturing custom-ordered paperboard packaging. Its operation is vertically integrated, and includes the production of paperboard from raw timber and wastepaper as well as its composition into the finished products ordered by customers. The operation is also largely domestic. During the years at issue in this case — 1963, 1964, and 1965 — appellant controlled 20 foreign subsidiaries located in four Latin American and four European countries. Its percentage ownership of the subsidiaries (either directly or through other subsidiaries) ranged between 66.7% and 100%. In those instances (about half) in which appellant did not own a 100% interest in the subsidiary, the remainder was owned by local nationals. One of the subsidiaries was a holding company that had no payroll, sales, or property, but did have book income. Another was inactive. The rest were all engaged — in their respective local markets — in essentially the same business as appellant.

Most of appellant’s subsidiaries were, like appellant itself, fully integrated, although a few bought paperboard and other intermediate products elsewhere. Sales of materials from appellant to its subsidiaries accounted for only about 1% of the subsidiaries’ total purchases. The subsidiaries were also relatively autonomous with respect to matters of personnel and day-to-day management. For example, transfers of personnel from appellant to its subsidiaries were rare, and occurred only when a subsidiary could not fill a position locally. There was no formal United States training program for the subsidiaries’ employees, although groups of foreign employees occasionally visited the United States for 2-6 week periods to familiarize themselves with appellant’s methods of operation. Appellant charged one senior vice president and four other officers with the task of overseeing the operations of the subsidiaries. These officers established general standards of professionalism, profitability, and ethical practices and dealt with major problems and long-term decisions; day-to-day management of the subsidiaries, however, was left in the hands of local executives who were always citizens of the host country. Although local decisions regarding capital expenditures were subject to review by appellant, problems were generally worked out by consensus rather than outright domination. Appellant also had a number of its directors and officers on the boards of directors of the subsidiaries, but they did not generally play an active role in management decisions.⁸

Nevertheless, in certain respects, the relationship between appellant and its subsidiaries was decidedly close. For example, approximately half of the subsidiaries’ long-term debt was either held directly, or guaranteed, by appellant. Appellant also provided advice and consultation regarding manufacturing techniques, engineering, design, architecture, insurance, and cost accounting to a number of its subsidiaries, either by entering into technical service agreements with them or by informal arrangement. Finally, appellant occasionally assisted its subsidiaries in their procurement of equipment, either by selling them used equipment of its own or by employing its own purchasing department to act as an agent for the subsidiaries.⁹

B

During the tax years at issue in this case, appellant filed California franchise tax returns. In 1969, after conducting an audit of appellant’s returns for the years in question, appellee issued notices of additional assessments for each of those years. The respective approaches and results reflected in appellant’s initial returns and in appellee’s notices of additional assessments capture the legal differences at issue in this case.¹⁰

In calculating the total unapportioned taxable income of its unitary business, appellant included its own corporate net earnings as derived from its federal tax form (subject to certain adjustments not relevant here), but did not include any income of its subsidiaries. It also deducted — as it was authorized to do under state law, see supra, at 167, and n. 1 — all dividend income, nonbusiness interest income, and gains on sales of assets not related to the unitary business. In calculating the share of its net income which was appor-tionable to California under the three-factor formula, appellant omitted all of its subsidiaries’ payroll, property, and sales. The results of these calculations are summarized in the margin.¹¹

The gravamen of the notices issued by appellee in 1969 was that appellant should have treated its overseas subsidiaries as part of its unitary business rather than as passive investments. Including the overseas subsidiaries in appellant’s unitary business had two primary effects: it increased the income subject to apportionment by an amount equal to the total income of those subsidiaries (less intersubsidiary dividends, see n. 5, supra), and it decreased the percentage of that income which was apportionable to California. The net effect, however, was to increase appellant’s tax liability in each of the three years.¹²

Appellant paid the additional amounts under protest, and then sued in California Superior Court for a refund, raising the issues now before this Court. The case was tried on stipulated facts, and the Superior Court upheld appellee’s assessments. On appeal, the California Court of Appeal affirmed, 117 Cal. App. 3d 988, 173 Cal. Rptr. 121 (1981), and the California Supreme Court refused to exercise discretionary review. We noted probable jurisdiction. 456 U. S. 960 (1982).

H-I I — I HH

, <T ^

We address the unitary business issue first. As previously noted, the taxpayer always has the “distinct burden of showing by ‘clear and cogent evidence’ that [the state tax] results in extraterritorial values being taxed.” Supra, at 164. One necessary corollary of that principle is that this Court will, if reasonably possible, defer to the judgment of state courts in deciding whether a particular set of activities constitutes a “unitary business.” As we said in a closely related context in Norton Co. v. Department of Revenue, 340 U. S. 534 (1951):

“The general rule, applicable here, is that a taxpayer claiming immunity from a tax has the burden of establishing his exemption.

“This burden is never met merely by showing a fair difference of opinion which as an original matter might be decided differently. ... Of course, in constitutional cases, we have power to examine the whole record to arrive at an independent judgment as to whether constitutional rights have been invaded, but that does not mean that we will re-examine, as a court of first instance, findings of fact supported by substantial evidence.” Id., at 537-538 (footnotes omitted; emphasis added).¹³

See id., at 538 (concluding that, “in light of all the evidence, the [state] judgment [on a question of whether income should be attributed to the State] was within the realm of permissible judgment”). The legal principles defining the constitutional limits on the unitary business principle are now well established. The factual records in such cases, even when the parties enter into a stipulation, tend to be long and complex, and the line between “historical fact” and “constitutional fact” is often fuzzy at best. Cf. AS ARCO, 458 U. S., at 326-328, nn. 22, 23. It will do the cause of legal certainty little good if this Court turns every colorable claim that a state court erred in a particular application of those principles into a de novo adjudication, whose unintended nuances would then spawn further litigation and an avalanche of critical comment.¹⁴ Rather, our task must be to determine whether the state court applied the correct standards to the case; and if it did, whether its judgment “was within the realm of permissible judgment.”¹⁵

B

In this case, we are singularly unconvinced by appellant’s argument that the State Court of Appeal “in important part analyzed this case under a different legal standard,” F. W. Woolworth, 458 U. S., at 363, from the one articulated by this Court. Appellant argues that the state court here, like the state court in F. W. Woolworth, improperly relied on appellant’s mere potential to control the operations of its subsidiaries as a dispositive factor in reaching its unitary business finding. In fact, although the state court mentioned that “major policy decisions of the subsidiaries were subject to review by appellant,” 117 Cal. App. 3d, at 998, 173 Cal. Rptr., at 127, it relied principally, in discussing the management relationship between appellant and its subsidiaries, on the more concrete observation that “[h]igh officials of appellant gave directions to subsidiaries for compliance with the parent’s standard of professionalism, profitability, and ethical practices.” Id., at 998, 173 Cal. Rptr., at 127-128.¹⁶

Appellant also argues that the state court erred in endorsing an administrative presumption that corporations engaged in the same line of business are unitary. This presumption affected the state court’s reasoning, but only as one element among many. Moreover, considering the limited use to which it was put, we find the “presumption” criticized by appellant to be reasonable. Investment in a business enterprise truly “distinct” from a corporation’s main line of business often serves the primary function of diversifying the corporate portfolio and reducing the risks inherent in being tied to one industry’s business cycle. When a corporation invests in a subsidiary that engages in the same line of work as itself, it becomes much more likely that one function of the investment is to make better use — either through economies of scale or through operational integration or sharing of expertise — of the parent’s existing business-related resources.

Finally, appellant urges us to adopt a bright-line rule requiring as a prerequisite to a finding that a mercantile or manufacturing enterprise is unitary that it be characterized by “a substantial flow of goods.” Brief for Appellant 47. We decline this invitation. The prerequisite to a constitutionally acceptable finding of unitary business is a flow of value, not a flow of goods.¹⁷ As we reiterated in F. W. Woolworth, a relevant question in the unitary business inquiry is whether ‘“contributions to income [of the subsidiaries] resulted] from functional integration, centralization of management, and economies of scale.’” 458 U. S., at 364, quoting Mobil, 445 U. S., at 438. “[Substantial mutual interdependence,” F. W. Woolworth, supra, at 371, can arise in any number of ways; a substantial flow of goods is clearly one but just as clearly not the only one.

C

The State Court of Appeal relied on a large number of factors in reaching its judgment that appellant and its foreign subsidiaries constituted a unitary business. These included appellant’s assistance to its subsidiaries in obtaining used and new equipment and in filling personnel needs that could not be met locally, the substantial role played by appellant in loaning funds to the subsidiaries and guaranteeing loans provided by others, the “considerable interplay between appellant and its foreign subsidiaries in the area of corporate expansion,” 117 Cal. App. 3d, at 997, 173 Cal. Rptr., at 127, the “substantial” technical assistance provided by appellant to the subsidiaries, id., at 998-999, 173 Cal. Rptr., at 128, and the supervisory role played by appellant’s officers in providing general guidance to the subsidiaries. In each of these respects, this case differs from ASARCO and F. W. Woolworth,¹⁸ and clearly comes closer than those cases did to presenting a “functionally integrated enterprise,” Mobil, supra, at 440, which the State is entitled to tax as a single entity. We need not decide whether any one of these factors would be sufficient as a constitutional matter to prove the existence of a unitary business. Taken in combination, at least, they clearly demonstrate that the state court reached a conclusion “within the realm of permissible judgment.”¹⁹

> > — (

We turn now to the question of fair apportionment. Once again, appellant has the burden of proof; it must demonstrate that there is “‘no rational relationship between the income attributed to the State and the intrastate values of the enterprise, Exxon Corp., 447 U. S., at 220, quoting Mobil, supra, at 437, by proving that the income apportioned to California under the statute is “out of all appropriate proportion to the business transacted by the appellant in that State,” Hans Rees’ Sons, Inc., 283 U. S., at 135.

Appellant challenges the application of California’s three-factor formula to its business on two related grounds, both arising as a practical (although not a theoretical) matter out of the international character of the enterprise. First, appellant argues that its foreign subsidiaries are significantly more profitable than it is, and that the three-factor formula, by ignoring that fact and relying instead on indirect measures of income such as payroll, property, and sales, systematically distorts the true allocation of income between appellant and the subsidiaries. The problem with this argument is obvious: the profit figures relied on by appellant are based on precisely the sort of formal geographical accounting whose basic theoretical weaknesses justify resort to formula apportionment in the first place. Indeed, we considered and rejected a very similar argument in Mobil, pointing out that whenever a unitary business exists,

“separate [geographical] accounting, while it purports to isolate portions of income received in various States, may fail to account for contributions to income resulting from functional integration, centralization of management, and economies of scale. Because these factors of profitability arise from the operation of the business as a whole, it becomes misleading to characterize the income of the business as having a single identifiable ‘source.’ Although separate geographical accounting may be useful for internal auditing, for purposes of state taxation it is not constitutionally required.” 445 U. S., at 438 (citation omitted).

Appellant’s second argument is related, and can be answered in the same way. Appellant contends:

“The costs of production in foreign countries are generally significantly lower than in the United States, primarily as a result of the lower wage rates of workers in countries other than the United States. Because wages are one of the three factors used in formulary apportionment, the use of the formula unfairly inflates the amount of income apportioned to United States operations, where wages are higher.” Brief for Appellant 12.

Appellant supports this argument with various statistics that appear to demonstrate, not only that wage rates are generally lower in the foreign countries in which its subsidiaries operate, but also that those lower wages are not offset by lower levels of productivity. Indeed, it is able to show that at least one foreign plant had labor costs per thousand square feet of corrugated container that were approximately 40% of the same costs in appellant’s California plants.

The problem with all this evidence, however, is that it does not by itself come close to impeaching the basic rationale behind the three-factor formula. Appellant and its foreign subsidiaries have been determined to be a unitary business. It therefore may well be that in addition to the foreign payroll going into the production of any given corrugated container by a foreign subsidiary, there is also California payroll, as well as other California factors, contributing — albeit more indirectly — to the same production. The mere fact that this possibility is not reflected in appellant’s accounting does not disturb the underlying premises of the formula apportionment method.

Both geographical accounting and formula apportionment are imperfect proxies for an ideal which is not only difficult to achieve in practice, but also difficult to describe in theory. Some methods of formula apportionment are particularly problematic because they focus on only a small part of the spectrum of activities by which value is generated. Although we have generally upheld the use of such formulas, see, e. g., Moorman Mfg. Co. v. Bair, 437 U. S. 267 (1978); Underwood Typewriter Co. v. Chamberlain, 254 U. S. 113 (1920), we have on occasion found the distortive effect of focusing on only one factor so outrageous in a particular case as to require reversal. In Hans Rees’ Sons, Inc. v. North Carolina ex rel. Maxwell, supra, for example, an apportionment method based entirely on ownership of tangible property resulted in an attribution to North Carolina of between 66% and 85% of the taxpayer’s income over the course of a number of years, while a separate accounting analysis purposely skewed to resolve all doubts in favor of the State resulted in an attribution of no more than 21.7%. We struck down the application of the one-factor formula to that particular business, holding that the method, “albeit fair on its face, operates so as to reach profits which are in no just sense attributable to transactions within its jurisdiction.” Id.,at 134.

The three-factor formula used by California has gained wide approval precisely because payroll, property, and sales appear in combination to reflect a very large share of the activities by which value is generated. It is therefore able to avoid the sorts of distortions that were present in Hans Rees’ Sons, Inc.

Of course, even the three-factor formula is necessarily imperfect.²⁰ But we have seen no evidence demonstrating that the margin of error (systematic or not) inherent in the three-factor formula is greater than the margin of error (systematic or not) inherent in the sort of separate accounting urged upon us by appellant. Indeed, it would be difficult to come to such a conclusion on the basis of the figures in this case: for all of appellant’s statistics showing allegedly enormous distortions caused by the three-factor formula, the tables we set out at nn. 11, 12, supra, reveal that the percentage increase in taxable income attributable to California between the methodology employed by appellant and the methodology employed by appellee comes to approximately 14%, a far cry from the more than 250% difference which led us to strike down the state tax in Hans Rees’ Sons, Inc., and a figure certainly within the substantial margin of error inherent in any method of attributing income among the components of a unitary business. See also Moorman Mfg. Co., supra, at 272-273; Ford Motor Co. v. Beauchamp, 308 U. S. 331 (1939); Underwood Typewriter Co., supra, at 120-121.

V

For the reasons we have just outlined, we conclude that California’s application of the unitary business principle to appellant and its foreign subsidiaries was proper, and that its use of the standard three-factor formula to apportion the income of that unitary business was fair. This proper and fair method of taxation happens, however, to be quite different from the method employed both by the Federal Government in taxing appellant’s business, and by each of the relevant foreign jurisdictions in taxing the business of appellant’s subsidiaries. Each of these other taxing jurisdictions has adopted a qualified separate accounting approach — often referred to as the “arm’s-length” approach — to the taxation of related corporations.²¹ Under the “arm’s-length” approach, every corporation, even if closely tied to other corporations, is treated for most — but decidedly not all — purposes as if it were an independent entity dealing at arm’s length with its affiliated corporations, and subject to taxation only by the jurisdictions in which it operates and only for the income it realizes on its own books.

If the unitary business consisting of appellant and its subsidiaries were entirely domestic, the fact that different jurisdictions applied different methods of taxation to it would probably make little constitutional difference, for the reasons we discuss supra, at 170-171. Given that it is international, however, we must subject this case to the additional scrutiny required by the Foreign Commerce Clause. See Mobil Oil Corp., 445 U. S., at 446; Japan Line, Ltd., 441 U. S., at 446; Bowman v. Chicago & N. W. R. Co., 125 U. S. 465, 482 (1888). The case most relevant to our inquiry is Japan Line.

A

Japan Line involved an attempt by California to impose an apparently fairly apportioned, nondiscriminatory, ad valorem property tax on cargo containers which were instrumental-ities of foreign commerce and which were temporarily located in various California ports. The same cargo containers, however, were subject to an unapportioned property tax in their home port of Japan. Moreover, a convention signed by the United States and Japan made clear, at least, that neither National Government could impose a tax on temporarily imported cargo containers whose home port was in the other nation. We held that “[w]hen a State seeks to tax the instru-mentalities of foreign commerce, two additional considerations, beyond those articulated in [the doctrine governing the Interstate Commerce Clause], come into play.” 441 U. S., at 446. The first is the enhanced risk of multiple taxation. Although consistent application of the fair apportionment standard can generally mitigate, if not eliminate, double taxation in the domestic context,

“neither this Court nor this Nation can ensure full apportionment when one of the taxing entities is a foreign sovereign. If an instrumentality of commerce is domiciled abroad, the country of domicile may have the right, consistently with the custom of nations, to impose a tax on its full value. If a State should seek to tax the same instrumentality on an apportioned basis, multiple taxation inevitably results. . . . Due to the absence of an authoritative tribunal capable of ensuring that the aggregation of taxes is computed on no more than one full value, a state tax, even though ‘fairly apportioned’ to reflect an instrumentality’s presence within the State, may subject foreign commerce ‘“to the risk of a double tax burden to which [domestic] commerce is not exposed, and which the commerce clause forbids.” ’ ” Id., at 447-448, quoting Evco v. Jones, 409 U. S. 91, 94 (1972), in turn quoting J. D. Adams Mfg. Co. v. Storen, 304 U. S. 307, 311 (1938) (footnote omitted).

The second additional consideration that arises in the foreign commerce context is the possibility that a state tax will “impair federal uniformity in an area where federal uniformity is essential.” 441 U. S., at 448.

“A state tax on instrumentalities of foreign commerce may frustrate the achievement of federal uniformity in several ways. If the State imposes an apportioned tax, international disputes over reconciling apportionment formulae may arise. If a novel state tax creates an asymmetry in the international tax structure, foreign nations disadvantaged by the levy may retaliate against American-owned instrumentalities present in their jurisdictions. ... If other States followed the taxing State’s example, various instrumentalities of commerce could be subjected to varying degrees of multiple taxation, a result that would plainly prevent this Nation from ‘speaking with one voice’ in regulating foreign commerce.” Id., at 450-451 (footnote omitted).

On the basis of the facts in Japan Line, we concluded that the California tax at issue was constitutionally improper because it failed to meet either of the additional tests mandated by the Foreign Commerce Clause. Id., at 451-454.

This case is similar to Japan Line in a number of important respects. First, the tax imposed here, like the tax imposed in Japan Line, has resulted in actual double taxation, in the sense that some of the income taxed without apportionment by foreign nations as attributable to appellant’s foreign subsidiaries was also taxed by California as attributable to the State’s share of the total income of the unitary business of which those subsidiaries are a part.²² Second, that double taxation stems from a serious divergence in the taxing schemes adopted by California and the foreign taxing authorities. Third, the taxing method adopted by those foreign taxing authorities is consistent with accepted international practice. Finally, our own Federal Government, to the degree it has spoken, seems to prefer the taxing method adopted by the international community to the taxing method adopted by California.²³

Nevertheless, there are also a number of ways in which this case is clearly distinguishable from Japan Line.²⁴ First, it involves a tax on income rather than a tax on property. We distinguished property from income taxation in Mobil Oil Corp., 445 U. S., at 444-446, and Exxon Corp., 447 U. S., at 228-229, suggesting that “[t]he reasons for allocation to a single situs that often apply in the case of property taxation carry little force” in the case of income taxation. 445 U. S., at 445. Second, the double taxation in this case, although real, is not the “inevitable]” result of the California taxing scheme. Cf. Japan Line, 441 U. S., at 447. In Japan Line, we relied strongly on the fact that one taxing jurisdiction claimed the right to tax a given value in full, and another taxing jurisdiction claimed the right to tax the same entity in part — a combination resulting necessarily in double taxation. Id., at 447, 452, 455. Here, by contrast, we are faced with two distinct methods of allocating the income of a multinational enterprise. The “arm’s-length” approach divides the pie on the basis of formal accounting principles. The formula apportionment method divides the same pie on the basis of a mathematical generalization. Whether the combination of the two methods results in the same income being taxed twice or in some portion of income not being taxed at all is dependent solely on the facts of the individual case.²⁵ The third difference between this case and Japan Line is that the tax here falls, not on the foreign owners of an instrumentality of foreign commerce, but on a corporation domiciled and headquartered in the United States. We specifically left open in Japan Line the application of that case to “domestically owned instrumentalities engaged in foreign commerce,” id., at 444, n. 7, and — to the extent that corporations can be analogized to cargo containers in the first place — this case falls clearly within that reservation.²⁶

In light of these considerations, our task in this case must be to determine whether the distinctions between the present tax and the tax at issue in Japan Line add up to a constitutionally significant difference. For the reasons we are about to explain, we conclude that they do.

B

In Japan Line, we said that “[e]ven a slight overlapping of tax — a problem that might be deemed de minimis in a domestic context — assumes importance when sensitive matters of foreign relations and national sovereignty are concerned.” Id., at 456 (footnote omitted). If we were to take that statement as an absolute prohibition on state-induced double taxation in the international context, then our analysis here would be at an end. But, in fact, such an absolute rule is no more appropriate here than it was in Japan Line itself, where we relied on much more than the mere fact of double taxation to strike down the state tax at issue. Although double taxation in the foreign commerce context deserves to receive close scrutiny, that scrutiny must take into account the context in which the double taxation takes place and the alternatives reasonably available to the taxing State.

In Japan Line, the taxing State could entirely eliminate one important source of double taxation simply by adhering to one bright-line rule: do not tax, to any extent whatsoever, cargo containers “that are owned, based, and registered abroad and that are used exclusively in international commerce . . . Id., at 444. To require that the State adhere to this rule was by no means unfair, because the rule did no more than reflect consistent international practice and express federal policy. In this case, California could try to avoid double taxation simply by not taxing appellant’s income at all, even though a good deal of it is plainly domestic. But no party has suggested such a rule, and its obvious unfairness requires no elaboration. Or California could try to avoid double taxation by adopting some version of the “arm’s-length” approach. That course, however, would not by any means guarantee an end to double taxation.

As we have already noted, the “arm’s-length” approach is generally based, in the first instance, on a multicorporate enterprise’s own formal accounting. But, despite that initial reliance, the “arm’s-length” approach recognizes, as much as the formula apportionment approach, that closely related corporations can engage in a transfer of values that is not fully reflected in their formal ledgers. Thus, for example, 26 U. S. C. § 482 provides: And, as one might expect, the United States Internal Revenue Service has developed elaborate regulations in order to give content to this general provision. Many other countries have similar provisions.²⁸ A serious problem, however, is that even though most nations have adopted the “arm’s-length” approach in its general outlines, the precise rules under which they reallocate income among affiliated corporations often differ substantially, and whenever that difference exists, the possibility of double taxation also exists.²⁹ Thus, even if California were to adopt some version of the “arm’s-length” approach, it could not eliminate the risk of double taxation of corporations subject to its franchise tax, and might in some cases end up subjecting those corporations to more serious double taxation than would occur under formula apportionment.³⁰

“In any case of two or more . . . businesses (whether or not incorporated, whether or not organized in the United States, and whether or not affiliated) owned or controlled directly or indirectly by the same interests, the Secretary [of the Treasury] may distribute, apportion, or allocate gross income, deductions, credits, or allowances between or among such . . . businesses, if he determines that such distribution, apportionment, or allocation is necessary in order to prevent evasion of taxes or clearly to reflect the income of any of such . . . businesses.”²⁷

That California would have trouble avoiding double taxation even if it adopted the “arm’s-length” approach is, we think, a product of the difference between a tax on income and a tax on tangible property. See supra, at 187-188. Allocating income among various taxing jurisdictions bears some resemblance, as we have emphasized throughout this opinion, to slicing a shadow. In the absence of a central coordinating authority, absolute consistency, even among taxing authorities whose basic approach to the task is quite similar, may just be too much to ask.³¹ If California’s method of formula apportionment “inevitably” led to double taxation, see supra, at 188, that might be reason enough to render it suspect. But since it does not, it would be perverse, simply for the sake of avoiding double taxation, to require California to give up one allocation method that sometimes results in double taxation in favor of another allocation method that also sometimes results in double taxation. Cf. Moorman Mfg. Co., 437 U. S., at 278-280.

It could be argued that even if the Foreign Commerce Clause does not require California to adopt the “arm’s-length” approach to foreign subsidiaries of domestic corporations, it does require that whatever system of taxation California adopts must not result in double taxation in any particular case. The implication of such a rule, however, would be that even if California adopted the “arm’s-length” method, it would be required to defer, not merely to a single internationally accepted bright-line standard, as was the case in Japan Line, but to a variety of § 482-type reallocation decisions made by individual foreign countries in individual cases. Although double taxation is a constitutionally disfavored state of affairs, particularly in the international context, Japan Line does not require forbearance so extreme or so one-sided.

C

We come finally to the second inquiry suggested by Japan Line — whether California’s decision to adopt formula apportionment in the international context was impermissible because it “may impair federal uniformity in an area where federal uniformity is essential,” 441 U. S., at 448, and “prevents the Federal Government from ‘speaking with one voice’ in international trade,” id., at 453, quoting Michelin Tire Corp. v. Wages, 423 U. S. 276, 286 (1976). In conducting this inquiry, however, we must keep in mind that if a state tax merely has foreign resonances, but does not implicate foreign affairs, we cannot infer, “[ajbsent some explicit directive from Congress, . . . that treatment of foreign income at the federal level mandates identical treatment by the States.” Mobil, 445 U. S., at 448. See also Japan Line, 441 U. S., at 456, n. 20; Michelin Tire Corp., supra, at 286. Thus, a state tax at variance with federal policy will violate the “one voice” standard if it either implicates foreign policy issues which must be left to the Federal Government or violates a clear federal directive. The second of these considerations is, of course, essentially a species of pre-emption analysis.

(1)

The most obvious foreign policy implication of a state tax is the threat it might pose of offending our foreign trading partners and leading them to retaliate against the Nation as a whole. 441 U. S., at 450. In considering this issue, however, we are faced with a distinct problem. This Court has little competence in determining precisely when foreign nations will be offended by particular acts, and even less competence in deciding how to balance a particular risk of retaliation against the sovereign right of the United States as a whole to let the States tax as they please. The best that we can do, in the absence of explicit action by Congress, is to attempt to develop objective standards that reflect very general observations about the imperatives of international trade and international relations.

This case is not like Mobil, in which the real issue came down to a question of interstate rather than foreign commerce. 445 U. S., at 446-449. Nevertheless, three distinct factors, which we have already discussed in one way or another, seem to us to weigh strongly against the conclusion that the tax imposed by California might justifiably lead to significant foreign retaliation. First, the tax here does not create an automatic “asymmetry,” Japan Line, supra, at 453, in international taxation. See supra, at 188, 192-193. Second, the tax here was imposed, not on a foreign entity as was the case in Japan Line, but on a domestic corporation. Although, California “counts” income arguably attributable to foreign corporations in calculating the taxable income of that domestic corporation, the legal incidence of the tax falls on the domestic corporation.³² Third, even if foreign nations have a legitimate interest in reducing the tax burden of domestic corporations, the fact remains that appellant is without a doubt amenable to be taxed in California in one way or another, and that the amount of tax it pays is much more the function of California’s tax rate than of its allocation method. Although a foreign nation might be more offended by what it considers unorthodox treatment of appellant than it would be if California simply raised its general tax rate to achieve the same economic result, we can only assume that the offense involved in either event would be attenuated at best.

A state tax may, of course, have foreign policy implications other than the threat of retaliation. We note, however, that in this case, unlike Japan Line, the Executive Branch has decided not to file an amicus curiae brief in opposition to the state tax.³³ The lack of such a submission is by no means dispositive. Nevertheless, when combined with all the other considerations we have discussed, it does suggest that the foreign policy of the United States — whose nuances, we must emphasize again, are much more the province of the Executive Branch and Congress than of this Court — is not seriously threatened by California’s decision to apply the unitary business concept and formula apportionment in calculating appellant’s taxable income.

(2)

When we turn to specific indications of congressional intent, appellant’s position fares no better. First, there is no claim here that the federal tax statutes themselves provide the necessary pre-emptive force. Second, although the United States is a party to a great number of tax treaties that require the Federal Government to adopt some form of “arm’s-length” analysis in taxing the domestic income of multinational enterprises,³⁴ that requirement is generally waived with respect to the taxes imposed by each of the contracting nations on its own domestic corporations.³⁵ This fact, if nothing else, confirms our view that such taxation is in reality of local rather than international concern. Third, the tax treaties into which the United States has entered do not generally cover the taxing activities of subnational governmental units such as States,³⁶ and in none of the treaties does the restriction on “non-arm’s-length” methods of taxation apply to the States. Moreover, the Senate has on at least one occasion, in considering a proposed treaty, attached a reservation declining to give its consent to a provision in the treaty that would have extended that restriction to the States.³⁷ Finally, it remains true, as we said in Mobil, that “Congress has long debated, but has not enacted, legislation designed to regulate state taxation of income.” 445 U. S., at 448.³⁸ Thus, whether we apply the “explicit directive” standard articulated in Mobil, or some more relaxed standard which takes into account our residual concern about the foreign policy implications of California’s tax, we cannot conclude that the California tax at issue here is pre-empted by federal law or fatally inconsistent with federal policy.

<

The judgment of the California Court of Appeal is

Affirmed.

Justice Stevens took no part in the consideration or decision of this case.

1 Certain forms of nonbusiness income, such as dividends, are allocated on the basis of the taxpayer’s commercial domicile. Other forms of nonbusiness income, such as capital gains on sales of real property, are allocated on the basis of situs. See Cal. Rev. & Tax. Code Ann. §§25123-25127 (West 1979).

2 See generally Honolulu Oil Corp. v. Franchise Tax Board, 60 Cal. 2d 417, 386 P. 2d 40 (1963); Superior Oil Corp. v. Franchise Tax Board, 60 Cal. 2d 406, 386 P. 2d 33 (1963).

3 See the opinion of the California Court of Appeal in this case, 117 Cal. App. 3d 988, 990-991, 993-995, 173 Cal. Rptr. 121, 123, 124-126 (1981). See also Cal. Rev. & Tax. Code Ann. § 25137 (West 1979) (allowing for separate accounting or other alternative methods of apportionment when total formula apportionment would “not fairly represent the extent of the taxpayer’s business activity in this state”).

4 We note that the Uniform Act does not speak to this question one way or the other.

5 See also Cal. Rev. & Tax. Code Ann. §26105 (West 1979) (defining “ownership or control”). A necessary corollary of the California approach, of course, is that intercorporate dividends in a unitary business not be included in gross income, since such inclusion would result in double-counting of a portion of the subsidiary’s income (first as income attributed to the unitary business, and second as dividend income to the parent). See § 25106.

Some States, it should be noted, have adopted a hybrid approach. In Mobil itself, for example, a nondomiciliary State invoked a unitary business justification to include an apportioned share of certain corporate dividends in the gross income of the taxpayer, but did not require a combined return and combined apportionment. The Court in Mobil held that the taxpayer’s objection to this approach had not been properly raised in the state proceedings. 445 U. S., at 441, n. 15. Justice Stevens, however, reached the merits, stating in part: “Either Mobil’s worldwide ‘petroleum enterprise’ is all part of one unitary business, or it is not; if it is, Vermont must evaluate the entire enterprise in a consistent manner.” Id., at 461 (citation omitted). See id., at 462 (Stevens, J., dissenting) (outlining alternative approaches available to State); cf. The Supreme Court, 1981 Term, 96 Harv. L. Rev. 62, 93-96 (1982).

6 See generally General Accounting Office Report to the Chairman, House Committee on Ways and Means: Key Issues Affecting State Taxation of Multijurisdictional Corporate Income Need Resolving 31 (1982).

7 Mobil did, in fact, involve income from foreign subsidiaries, but that fact was of little importance to the case for two reasons. First, as discussed in n. 5, supra, the State in that case included dividends from the subsidiaries to the parent in its calculation of the parent’s apportionable taxable income, but did not include the underlying income of the subsidiaries themselves. Second, the taxpayer in that case conceded that the dividends could be taxed somewhere in the United States, so the actual issue before the Court was merely whether a particular State could be barred from imposing some portion of that tax. See 445 U. S., at 447.

8 There were a number of reasons for appellant’s relatively hands-off attitude toward the management of its subsidiaries. First, it comported with the company’s general management philosophy emphasizing local responsibility and accountability; in this respect, the treatment of the foreign subsidiaries was similar to the organization of appellant’s domestic geographical divisions. Second, it reflected the fact that the packaging industry, like the advertising industry to which it is closely related, is highly sensitive to differences in consumer habits and economic development among different nations, and therefore requires a good dose of local expertise to be successful. Third, appellant’s policy was designed to appeal to the sensibilities of local customers and governments.

9 There was also a certain spillover of goodwill between appellant and its subsidiaries; that is, appellant’s customers who had overseas needs would on occasion ask appellant’s sales representatives to recommend foreign firms, and, where possible, the representatives would refer the customers to appellant’s subsidiaries. In at least one instance, appellant became involved in the actual negotiation of a contract between a customer and a foreign subsidiary.

10 After the notices of additional tax, there followed a series of further adjustments, payments, claims for refunds, and assessments, whose combined effect was to render the figures outlined in text more illustrative than real as descriptions of the present claims of the parties with regard to appellant’s total tax liability. These subsequent events, however, did not concern the legal issues raised in this case, nor did they remove either party’s financial stake in the resolution of those issues. We therefore disregard them for the sake of simplicity.

11 Total income of unitary business

Percentage attributed to California

Amount attributed to California

Tax (5.5%)

1963.... $26,870,427.00 11.041 $2,966,763.85 $163,172.01

1964.... 28,774,320.48 10.6422 3,062,220.73 168,422.14

1965.... 32,280,842.90 9.8336 3,174,368.97 174,590.29

See Exhibit A-7 to Stipulation; Record 36, 76, 77, 79, 104, 126.

12 According to the notices, appellant’s actual tax obligations were as follows:

Total income of unitary business

Percentage attributed to California

Amount attributed to California

Tax (5.5%)

1963.... $37,348,183.00 8.6886 $3,245,034.23 $178,476.88

1964.... 44,245,879.00 8.3135 3,673,381.15 202,310.95

1965.... 46,884,966.00 7.6528 3,588,012.68 197,340.70

See Exhibit A-7 to Stipulation; Record 76, 77, 79.

13 This approach is, of course, quite different from the one we follow in certain other constitutional contexts. See, e. g., Brooks v. Florida, 389 U. S. 413 (1967); New York Times Co. v. Sullivan, 376 U. S. 254, 285 (1964).

14 It should also go without saying that not every claim that a state court erred in making a unitary business finding will pose a substantial federal question in the first place.

15 AS ARCO and F. W. Woolworth are consistent with this standard of review. ASARCO involved a claim that a parent and certain of its partial subsidiaries, in which it held either minority interests or bare majority interests, were part of the same unitary business. The State Supreme Court upheld the claim. We concluded, relying on factual findings made by the state courts, that a unitary business finding was impermissible because the partial subsidiaries were not realistically subject to even minimal control by ASARCO, and were therefore passive investments in the most basic sense of the term. 458 U. S., at 320-324. We held specifically that to accept the State’s theory of the case would not only constitute a misapplication of the unitary business concept, but would “destroy” the concept entirely. Id., at 326.

F. W. Woolworth was a much closer case, involving one partially owned subsidiary and three wholly, owned subsidiaries. We examined the evidence in some detail, and reversed the state court’s unitary business finding, but only after concluding that the state court had made specific and crucial legal errors, not merely in the conclusions it drew, but in the legal standard it applied in analyzing the case. 458 U. S., at 363-364.

16 In any event, although potential control is, as we said in F. W. Woolworth, not “dispositive” of the unitary business issue, id., at 362 (emphasis added), it is relevant, both to whether or not the components of the purported unitary business share that degree of common ownership which is a prerequisite to a finding of unitariness, and also to whether there might exist a degree of implicit control sufficient to render the parent and the subsidiary an integrated enterprise.

17 As we state supra, at 167-169, there is a wide range of constitutionally acceptable variations on the unitary business theme. Thus, a leading scholar has suggested that a “flow of goods” requirement would provide a reasonable and workable bright-line test for unitary business, see Hellerstein, Recent Developments in State Tax Apportionment and the Circumscription of Unitary Business, 21 Nat. Tax J. 487, 501-502 (1968); Hellerstein, Allocation and Apportionment of Dividends and the Delineation of the Unitary Business, 14 Tax Notes 155 (Jan. 25, 1982), and some state courts have adopted such a test, see, e. g., Commonwealth v. ACF Industries, Inc., 441 Pa. 129, 271 A. 2d 273 (1970). But see, e. g., McLure, Operational Interdependence Is Not the Appropriate “Bright Line Test” of a Unitary Business — At Least Not Now, 18 Tax Notes 107 (Jan. 10, 1983). However sensible such a test may be as a policy matter, however, we see no reason to impose it on all the States as a requirement of constitutional law. Cf. Wisconsin v. J. C. Penney Co., 311 U. S. 435, 445 (1940).

18 See n. 15, supra. See also, e. g., F. W. Woolworth, 458 U. S., at 365 (“no phase of any subsidiary’s business was integrated with the parent’s”); ibid, (undisputed testimony stated that each subsidiary made business decisions independently of parent); id., at 366 (“each subsidiary was responsible for obtaining its own financing from sources other than the parent”); ibid. (“With one possible exception, none of the subsidiaries’ officers during the year in question was a current or former employee of the parent”) (footnote omitted).

19 Two of the factors relied on by the state court deserve particular mention. The first of these is the flow of capital resources from appellant to its subsidiaries through loans and loan guarantees. There is no indication that any of these capital transactions were conducted at arm’s length, and the resulting flow of value is obvious. As we made clear in another context in Corn Products Refining Co. v. Commissioner, 850 U. S. 46, 50-53 (1955), capital transactions can serve either an investment function or an operational function. In this case, appellant’s loans and loan guarantees were clearly part of an effort to ensure that “[t]he overseas operations of [appellant] continue to grow and to become a more substantial part of the company’s strength and profitability.” Container Corporation of America, 1964 Annual Report 6, reproduced in Exhibit I to Stipulation of Facts. See generally id., at 6-9, 11.

The second noteworthy factor is the managerial role played by appellant in its subsidiaries’ affairs. We made clear in F. W. Woolworth Co. that a unitary business finding could not be based merely on “the type of occasional oversight — with respect to capital structure, major debt, and dividends — that any parent gives to an investment in a subsidiary . . . .” 458 U. S., at 369. As Exxon illustrates, however, mere decentralization of day-to-day management responsibility and accountability cannot defeat a unitary business finding. 447 U. S., at 224. The difference lies in whether the management role that the parent does play is grounded in its own operational expertise and its overall operational strategy. In this case, the business “guidelines” established by appellant for its subsidiaries, the “consensus” process by which appellant’s management was involved in the subsidiaries’ business decisions, and the sometimes uncompensated technical assistance provided by appellant, all point to precisely the sort of operational role we found lacking in F. W. Woolworth.

20 First, the one-third-eaeh weight given to the three factors is essentially arbitrary. Second, payroll, property, and sales still do not exhaust the entire set of factors arguably relevant to the production of income. Finally, the relationship between each of the factors and income is by no means exact. The three-factor formula, as applied to horizontally linked enterprises, is based in part on the very rough economic assumption that rates of return on property and payroll — as such rates of return would be measured by an ideal accounting method that took all transfers of value into account — are roughly the same in different taxing jurisdictions. This assumption has a powerful basis in economic theory: if true rates of return were radically different in different jurisdictions, one might expect a significant shift in investment resources to take advantage of that difference. On the other hand, the assumption has admitted weaknesses: an enterprise’s willingness to invest simultaneously in two jurisdictions with very different true rates of return might be adequately explained by, for example, the difficulty of shifting resources, the decreasing marginal value of additional investment, and portfolio-balancing considerations.

21 The “arm’s-length” approach is also often applied to geographically distinct divisions of a single corporation.

22 The stipulation of facts indicates that the tax returns filed by appellant’s subsidiaries in their foreign domiciles took into account “only the applicable income and deductions incurred by the subsidiary or subsidiaries in that country and not. . . the income and deductions of [appellant] or the subsidiaries operating in other countries.” App. 72. This does hot conclusively demonstrate the existence of double taxation because appellant has not produced its foreign tax returns, and it is entirely possible that deductions, exemptions, or adjustments in those returns eliminated whatever overlap in taxable income resulted from the application of the California apportionment method. Nevertheless, appellee does not seriously dispute the existence of actual double taxation as we have defined it, Brief for Ap-pellee 114-121, but cf. Tr. of Oral Arg. 28-29, and we assume its existence for the purposes of our analysis. Cf. Japan Line, 441 U. S., at 452, n. 17.

23 But see infra, at 196-197 (discussing whether state scheme is preempted by federal law).

24 Note that we deliberately emphasized in Japan Line the narrowness of the question presented: “whether instrumentalities of commerce that are owned, based, and registered abroad and that are used exclusively in international commerce, may be subjected to apportioned ad valorem property taxation by a State.” 441 U. S., at 444.

25 Indeed, in Chicago Bridge & Iron Co. v. Caterpillar Tractor Co., No. 81-349, which was argued last Term and carried over to this Term, application of worldwide combined apportionment resulted in a refund to the taxpayer from the amount he had paid under a tax return that included neither foreign income nor foreign apportionment factors.

26 We have no need to address in this opinion the constitutionality of combined apportionment with respect to state taxation of domestic corporations with foreign parents or foreign corporations with either foreign parents or foreign subsidiaries. See also n. 32, infra.

27 Cf. Treasury Department’s Model Income Tax Treaty of June 16,1981, Art. 9, reprinted in CCH Tax Treaties ¶158 (1981) (hereinafter Model Treaty) (“Where ... an enterprise of a Contracting State participates directly or indirectly in the management, control or capital of an enterprise of the other Contracting State. . . and . . . conditions are made or imposed between the two enterprises in their commercial or financial relations which differ from those which would be made between independent enterprises, then any profits which, but for those conditions would have accrued to one of the enterprises, but by reason of those conditions have not so accrued, may be included in the profits of that enterprise and taxed accordingly”); J. Bischel, Income Tax Treaties 219 (1978) (hereinafter Bischel).

28 See generally G. Harley, International Division of the Income Tax Base of Multinational Enterprise 143-160 (1981) (hereinafter Harley); Madere, International Pricing: Allocation Guidelines and Relief from Double Taxation, 10 Tex. Int’l L. J. 108, 111-120 (1975).

29 See Surrey, Reflections on the Allocation of Income and Expenses Among National Tax Jurisdictions, 10 L. & Policy Int’l Bus. 409 (1978); Bischel 459-461, 464-466; B. Bittker & J. Eustice, Federal Income Taxation of Corporations and Shareholders ¶ 15.06 (4th ed. 1979); Harley 143-160.

30 Another problem arises out of the treatment of intercorporate dividends. Under formula apportionment as practiced by California, inter-corporate dividends attributable to the unitary business are, like many other intercorporate transactions, considered essentially irrelevant and are not included in taxable income. See n. 5, swpra. If the “arm’s-length” method were entirely consistent, it would tax intercorporate dividends when they occur, just as all other investment income is taxed. (In which State that dividend could be taxed is not particularly important, since the issue here is international rather than interstate double taxation. See Mobil, 445 U. S., at 447-448.) It could also be argued that this would not, strictly speaking, result in double taxation, since the income taxed would be income “of” the parent rather than income “of” the subsidiary. The effect, however, would often be to penalize an enterprise simply because it has adopted a particular corporate structure. In practice, therefore, most jurisdictions allow for tax credits or outright exemptions for intercorporate dividends among closely tied corporations, and provision for such credits or exemptions is often included in tax treaties. See generally Model Treaty, Art. 23; Bischel 2. No suggestion has been made here that appellant’s dividends from its subsidiaries would have to be exempt entirely from domestic state taxation. And the grant of a credit, which is the approach taken by federal law, see 26 U. S. C. § 901 et seq., does not in fact entirely eliminate effective double taxation: the same income is still taxed twice, although the credit insures that the total tax is no greater than that which would be paid under the higher of the two tax rates involved. Moreover, once the Federal Government has allowed a credit for foreign taxes on a particular intercorporate dividend, we are not persuaded why, as a logical matter, a State would have to grant another credit of its own, since the federal credit would have already vindicated the goal of not subjecting the taxpayer to a higher tax burden that it would have to bear if its subsidiary’s income were not taxed abroad.

31 At the federal level, double taxation is sometimes mitigated by provisions in tax treaties providing for intergovernmental negotiations to resolve differences in the approaches of the respective taxing authorities. See generally Model Treaty, Art. 25; 2 New York University, Proceedings of the Fortieth Annual Institute on Federal Taxation § 31.03[2] (1982) (hereinafter N. Y. U. Institute). But cf. Owens, United States Income Tax Treaties: Their Role in Relieving Double Taxation, 17 Rutgers L. Rev. 428,443-444 (1963) (role of such provisions procedural rather than substantive). California, however, is in no position to negotiate with foreign governments, and neither the tax treaties nor federal law provides a mechanism by which the Federal Government could negotiate double taxation arising out of state tax systems. In any event, such negotiations do not always occur, and when they do occur they do not always succeed.

32 We recognize that the fact that the legal incidence of a tax falls on a corporation whose formal corporate domicile is domestic might be less significant in the case of a domestic corporation that was owned by foreign interests. We need not decide here whether such a case would require us to alter our analysis.

33 The Solicitor General did submit a memorandum opposing worldwide formula apportionment by a State in Chicago Bridge & Iron Co. v. Caterpillar Tractor Co., No. 81-349, a case that was argued last Term, and carried over to this Term. Although there is no need for us to speculate as to the reasons for the Solicitor General’s decision not to submit a similar memorandum or brief in this case, cf. Brief for National Governors’ Association et al. as Amici Curiae 6-7, there has been no indication that the position taken by the Government in Chicago Bridge & Iron Co. still represents its views, or that we should regard the brief in that case as applying to this case.

34 See generally Model Treaty, Art. 7(2); Bischel 33-38, 459-461.

35 See Model Treaty, Art. 1(3); Bischel 718; N. Y. U. Institute § 31.04[3],

36 See Bischel 7.

37 See 124 Cong. Rec. 18400, 19076 (1978).

38 There is now pending one such bill of which we are aware. See H. R. 2918, 98th Cong., 1st Sess. (1983).