As American as Apple Inc.: International Tax and Ownership Nationality

I. Introduction

One of the most contentious issues in U.S. tax policy in recent years concerns the manner in which large U.S. multinational companies are taxed-or not.[1] Employing a multi-layered configuration of foreign subsidiaries-in a structure known as the "double Irish, Dutch sandwich"* [2]-banner U.S. companies like Apple, Cisco, and Google[3] have significantly reduced their global tax bills by shifting profitable aspects of their businesses to low-tax countries. Reformers point to the inefficiency and unfairness of conferring special tax benefits on the kind of large technology-intensive multinationals best able to avail themselves of these "loopholes," while other U.S. businesses pay full freight.[4] The multinationals themselves point to the intensity of global competition and the hobbling effect of full exposure to U.S. tax rates. [5]

The debate is, of course, more complex than this simple back and forth-indeed, far more complex. The tax-reducing structures themselves-whose colorful names make better press than how they actually work-are almost unsoundably elaborate and only rarely publicly visible.[6] The laws and regulations governing these structures are voluminous, scattered, and ever-shifting.[7]

But this Article is not so much about the debate as about its terms. And regarding these, it asks an apparently simple question: When we speak of "U.S. multinationals," what do we mean by "U.S."? More specifically, to what extent are these "U.S." companies owned by nonU.S. investors?

A staunch internationalist might come to this question in hope of finding that global capital markets are getting the better of narrowminded nationalism. A staunch nationalist might come to the question wary of finding that transnational economic interests are getting the better of elected national governments. I come to this question only indirectly, in a manner that is hopefully less ideological and admittedly more lawyerly. My interest in the question arises because, as I argue, ownership nationality plays an important role in several key sections of the brief for retaining, enhancing, or only partially limiting the tax benefits now available to large U.S. multinationals. My interest is in whether these arguments carry weight. As such, I am interested in the suppositions on which they rest.

Part II brings to the fore ownership nationality's[8] role in the current policy discussion regarding the taxation of large U.S. multinationals.[9] The competitiveness of U.S. business and the propriety of offering a tax holiday for "repatriated profits" are two highlighted issues.

Part III examines the most likely response to be heard by experts in the field when questions arise regarding the nationality of these companies' shareholders: the "robust stylized fact"[10] of "home country bias" in equity ownership. This is the generally accepted finding that investors, for reasons that may be partly rational and partly irrational, hold a disproportionate share of their equity portfolio in home country stocks as opposed to foreign country stocks. I tentatively suggest that this fact may be more stylized and less robust than has been supposed. But more to the point, and more pointedly, I argue that it is certainly less relevant in specific relation to large U.S. multinationals than its de rigueur invocation in the current international tax debate might lead one to believe. Indeed, as I emphasize, there is a sense in which the literature itself argues for its own inaptness.

Parts IV through VIII then conduct some reconnaissance for better data. The search is conducted in a way that is meant to be sensitive to the complex, organic, and ever-changing institutional structure of cross-border investment. The provisional report from this search is that such data are not available-certainly not to the public, perhaps not to the government, and probably not even to the companies themselves.

II. OWNERSHIP NATIONALITY IN THE CURRENT DEBATE

Several of the arguments concerning the U.S. taxation of large U.S. multinational companies take for granted that these entities are in some relevant sense "U.S. companies." This Part reviews a selection of these arguments highlighting the role played by the supposition of U.S. ownership. The claim is not that the ownership nationality of these companies is all that matters for any one of these issues, but rather that for all of these issues ownership nationality plays a key role, whether that be logical or merely rhetorical.

A. Win, America:[11] The Invocation of U.S. Competitiveness

Those in favor of reducing the U.S. tax burden on U.S. multinationals argue that the current U.S. tax code-and certainly a revised code that prevents tax schemes like the double Irish, Dutch sandwich-puts U.S. companies at a competitive disadvantage in the global arena.[12]

If the argument is that the competitiveness of these companies should guide U.S. international tax policy, it would seem to be a fair question who owns these companies. If the proposal is to sacrifice precious revenue to help someone "win," it seems worth knowing who that someone is.

It would be unfair to characterize competitiveness claims as concerned solely with the interests of large multinationals' shareholders, as opposed to their employees and other suppliers.[13] And it would be incorrect to ignore the fundamental economic insight that the benefits of a tax reduction are generally shared with counterparties and others who do not actually pay the tax-due to responsive price and quantity adjustments. The point here is simply that we may also wish to keep in mind the less interesting possibility that not all, perhaps not even most, of the benefits will shift in this manner; specifically, that some portion of the "winnings" generated by reducing the tax on the profits of large U.S. multinationals will go to the shareholders who own those profits-60%, if one extrapolates from one respected account.[14]

B. Win America: The Campaign for a Repatriation Holiday

Shareholder nationality is also relevant in the debate over whether multinationals should be offered a second "repatriation holiday." First, some background. When a U.S. multinational uses a tax device like the double Irish, Dutch sandwich to shift otherwise taxable profits to foreign subsidiaries in low- or no-tax jurisdictions, it is true, as noted in the Introduction, that the parent-subsidiary group pays little or no tax on the income at the time such income is earned. However, the group does in fact pay U.S. tax on the shifted profits when and if the parent extracts those profits from its foreign subsidiaries, by causing them to pay a dividend. The dividend is generally taxable to the parent at the full U.S. corporate rate of 35%.[15] This extraction of foreign subsidiary earnings is commonly (though, as discussed below, misleadingly) referred to as "repatriation."

But there is still a tax benefit. In the years before repatriation, the parent causes the subsidiary to invest those profits, and the subsidiary pays little or no tax on the resulting investment earnings, just as it paid little or no tax on the original earnings. In effect the U.S. parent uses its foreign subsidiary as an unlimited Individual Retirement Account (IRA). As with an IRA, the income is untaxed (or little-taxed) as earned, grows tax free, and is finally taxed upon "withdrawal," which in this case is by dividend. Depending on how long the parent waits before repatriating, the present discounted value[16] of its total tax bill may be significantly reduced.[17]

In fact the tax treatment of shifted profits has in the past been-and may again be-even more favorable than this. Around this time in the preceding decade, multinationals lobbied for and received a "repatriation tax holiday." During 2005 only, U.S. corporations were able to repatriate accumulated overseas profits at a significantly reduced tax rate-5.25% versus 35%.[18] As such, for these profits, the tax upon "withdrawal" was largely eliminated (to continue the IRA analogy), and the tax benefit went beyond the timing benefit described above.

Before the 2005 holiday, U.S. multinationals had accumulated $800 billion of "un-repatriated" earnings.[19]During 2005, they "repatriated" $312 billion at the holiday rate.[20] Since that time the retained earnings accounts of foreign subsidiaries have (perhaps predictably) re-swollen. Current estimates put the amount at $1.9 trillion (which is $1.7 trillion in 2005 dollars).[21] And (perhaps predictably) in recent years, multinationals have joined forces to lobby for a second holiday.[22]

Proponents of a second holiday point to the damaging phenomenon of "lock-out." The prospect of taxation on repatriation, they say, causes them to keep their earnings offshore-earnings, which if brought home, they would use to hire American workers and suppliers.[23] Opponents counter that the repatriating firms would not in fact bring on more workers and suppliers, but rather would simply distribute the money to their own shareholders. This, opponents say, is precisely what they did last time.[24]

Both positions are problematic. In the first place, the "foreign" earnings of large U.S. multinationals remain "abroad" only in the legalistic sense that the foreign subsidiary, which is incorporated abroad, holds title to the assets.[25] The U.S. taxable event is the passage of title to the parent-actual or constructive.[26] Before that happens, the assets themselves might be invested anywhere (at least financially).[27] In fact, thanks to a recent congressional hearing, we know that the vast majority of the "unrepatriated" foreign earnings of companies like Apple, Cisco, and Google already reside in U.S. bank accounts or other U.S. investments.[28]

On the other hand, the finding that foreign earnings were paid out to shareholders in 2005-rather than as wages, salaries, and payments to suppliers-does not seem to imply that the money simply exited the U.S. economy. One must ask: What did the shareholders then do with it? Quite possibly, they invested it elsewhere in the U.S. economy or spent it on U.S.-produced goods and services.

And here one arrives at the potential relevance of ownership nationality. Suppose, as opponents warn, that some portion of the foreign earnings of U.S. multinationals would in turn be distributed further up to the multinational's shareholders in the event of a holiday. (Given the fungibility of money this probably would be hard to prevent.) Understanding the impact on U.S. economic activity of this portion of the flow would seem to require an understanding of who these shareholders are. Imagine, for example, that foreign shareholders are more likely than U.S. shareholders to reinvest in foreign rather than U.S. businesses, which are in turn more likely to engage foreign rather than U.S. suppliers and workers. Then the positive impact of the holiday on U.S. economic activity would be attenuated to the extent that the repatriating companies are foreign-owned.

It might seem that there is an inherent contradiction in the scenario just described. The foreign investors under consideration are by definition owners of U.S. companies. Why would they be more likely to reinvest in non-U.S. businesses? They have demonstrated their globalist outlook.

But there is only a contradiction if one thinks that such foreign investors regard U.S. multinationals in the same way that they regard other purely domestic U.S. businesses and differently from how they regard businesses closer to home. Suppose, for example, that the relevant boundaries for investing are cognitive rather than political-that what matters is familiarity, not nationality. In that case, the set of "cognitively domestic" investments for foreign investors might include both nonmultinational own-country businesses and large multinationals. Symmetrically, the companies closest to home in the minds of U.S. investors might be nonmultinational U.S. businesses and large multinationals. Consequently, a foreign shareholder receiving a distribution from a large U.S. multinational and sprinkling it over her full portfolio might well return less to the U.S. economy than would a U.S. investor.

With this thought in mind, consider now the economics literature on "home country bias."

III. "Home Country Bias" and "Home Country Bias" Bias

Any investigation of shareholder nationality quickly runs up against the academic economics literature on "home country bias." This Article is no exception-although a point of emphasis here is that this literature is not the end of the line. Indeed, the need to look beyond the literature follows from a close reading of the literature itself.

A. Summary of the Literature

The home country bias literature identifies-and seeks to explain- the finding that equity investors tend to invest disproportionately in domestic as compared to foreign companies. The magnitude of this tendency is what has held the interest of researchers. Some degree of tilt toward home-country stocks is explainable by lingering tax disadvantages, transaction costs, and foreign exchange risk. The literature's raison d'être is the puzzle that these usual suspects appear to fall far short of explaining the data.

The emphasis on magnitude is evident in Kenneth French and James Poterba's 1991 article,[29] the root of much of the research on home country bias over the last quarter century. (The same emphasis continues in current work.[30]) French and Poterba begin with what they call "crude estimates"[31] of the foreign portion of the portfolio holdings of investors from various countries.[32]

They calculate that in 1989 U.K. investors, for example, held 18% of their equity portfolios (by value) in foreign companies and, specifically, 1.1% in U.S. companies. Japanese investors held 1.9% of their equity portfolios in foreign companies and 3.1% in U.S. companies.[33] The authors then use these estimates to calculate, for each country's investors, the expected returns for domestic versus foreign stocks that would rationalize such percentages.[34] They find that U.K. investors must have perceived that (inflation-adjusted) returns to U.K. stocks were (roughly speaking) 5.1 percentage points greater than for U.S. stocks, while Japanese investors must have believed that returns for Japanese companies were 2.5 percentage points greater than for U.S. companies.[35] To appreciate the impressive magnitude of these differentials, note that annual inflation-adjusted pretax stock returns (assuming reinvested dividends) were on the order of 12 percentage points over the 1980's and 1990's.[36] For purposes of further illustration, were the 5.1 percentage point U.K. differential caused entirely by differences in the tax rate for U.K. investors on investment in U.S. companies versus in U.K. companies, the tax rate for U.K. investors would have to be 5.1/0.12 = 42.5 percentage points higher on U.S. investments than on U.K. investments.[37] For the Japanese investor, the analogous tax rate difference is 2.5/0.12 = 20.8 percentage points. Keep in mind that these numbers, 42.5 and 20.8 represent additional percentage points of tax over and above the tax imposed on owncountry investments-this in a world in which foreign tax credits were generally available.

Subsequent contributors to the home country bias literature have updated these numbers and affirmed the continuing existence of the bias-though with some dampening over time.[38] A recent survey, for instance, reports[39] that in 2008, despite the fact that 94.5% of stocks by value were foreign from a U.K. investor's perspective, the portion of U.K. investors' equity portfolios consisting of foreign stocks was only 46.7%[40]-which is, albeit, greater than the 18% reported in French and Poterba two decades earlier.[41] From the perspective of Japanese investors, 90% of the world stock market by value consisted of foreign companies,[42]but only 12.6% of Japanese investors' portfolios by value consisted of foreign stocks[43]-up from the 1.9% reported by French and Poterba.[44]

What accounts for the bias? Researchers have posited and evaluated eight different explanations (some of which have already been mentioned):

1. Regulatory barriers. Regulatory barriers that prevent or constrict cross-border investment-such as restrictions on currency conversion.

2. Tax disincentives. Effective surtaxes on foreign investment returns-such as would result were cross-border investors taxed by both their own country of residence and the country of residence of the companies in which they invest, without being able to credit one tax against the other.[45]

3. Transaction costs. The greater transaction costs (literally, the fees charged by brokers and other intermediaries) incurred in buying and selling foreign as compared to domestic stocks.[46]

4. Foreign exchange risk. The additional risk of foreign exchange fluctuations that is attendant on investing in foreign currency-denominated stocks.[47]

5. Hedging local risk. The possibility that home-country stocks are a better hedge against risks specific to home-country investors-as when, hypothetically, layoffs or price increases at home tend to be associated with higher profits for home-country companies.[48]

6. Governance issues. Greater corporate or national governance risk faced by foreign investors.[49]

7. Information asymmetry. The possibility that investors have less information concerning foreign-country companies, making them riskier investments.[50]

8. Irrational biases. Various explanations rooted in behavioral biases, such as the possibility that "familiarity" leads to unjustified overconfidence while lack of familiarity leads to unjustified fear.[51] Most researchers seem to agree that no single explanation is sufficient, but that rather some selection may be at work.[52] Which selection and with what relative weights?

The first four explanations are generally downplayed in the literature. There seems to be longstanding general agreement in the home bias literature that, given financial opening over the last several decades, regulatory barriers and tax disincentives no longer provide an adequate explanation.[53] The transaction costs explanation is cast into doubt by data showing that investors trade in and out of foreign assets at least as frequently as they do domestic assets-despite their lower holdings of the former.[54] Foreign exchange rate risk is de-emphasized because of the availability of currency forward markets and other hedging instruments.[55]

The viability of the fifth set of explanations, hedging local risk, is also shaded in doubt, although this is somewhat more controversial.[56] It seems a fair assessment that explanations of this kind, however ingenious, are highly model-dependent.[57] Intuitively, it appears just as plausible that local fortunes would tend to move in tandem-as when new technologies increase both domestic stock returns and domestic wages. That would make domestic stocks a poor hedge for domestic investors and the result would be "foreign equity bias." Part and parcel with its theoretical fragility, the empirical evidence for hedging local risk is mixed at best.[58]

That leaves the last three explanations: governance issues, information asymmetries, and behavioral biases. Collectively, these receive a generally positive reception;[59] most of the disagreement concerning them relates to the difficulty of distinguishing among them.[60] These three explanations are discussed in more detail in the next Section.

B. Applicability

There are reasons to question the home country bias literature on its own terms. Recall, for example, that French and Poterba candidly refer to their tabulations of portfolio holdings as "crude estimates."[61] One might ask: What makes these point estimates crude? How wide a range of estimates is needed to take that crudity into account? Have estimates been refined over time? And, depending on how wide that range was and still is, could and can the magnitude of the bias-which is what makes the puzzle-really be reliably determined?[62]

These questions reappear briefly in the next Part, which takes a closer look at where most of the numbers behind the home country bias literature come from.[63] But a general critique of the home equity bias literature is not the primary objective of this Article. Rather, the focus is on whether the estimates presented in the home country bias literature, however rough, are specifically applicable to the large U.S. multinationals at the center of the current debate over U.S. international taxation.

It seems a fair hypothesis that they are not. Home country bias estimates are averages over the full set of potential equity investments. Yet the most well-accepted explanations for the bias-governance issues, information asymmetries, and behavioral biases-seem to justify regarding large U.S. multinationals as exceptional.

With regard to governance issues, large U.S. multinationals are generally publicly traded U.S.-chartered companies subject to well-established minority-protecting corporate laws, extensive disclosure requirements, and the (generally) reliable political governance of U.S. federal and state authorities. All of these regimes are imperfect. But they must compare reasonably well to the alternatives facing foreign investors.[64]

More importantly, with regard to information asymmetries and behavioral biases, it is important to remember that large multinational enterprises are, after all, multinational. This generally means that they have large commercial presences outside the United States. There are thirty-seven Apple retail stores in the United Kingdom and seven in Japan.[65] And 61% of Apple's net sales are outside the United States.[66] The Google brand is similarly transnational. At least politicians of other nations seem to think the name "Google" resonates well enough with their constituents to make that company's local taxation a hobbyhorse.[67] And while it is true that nonconsumer companies like Cisco operate behind the scenes, this is no more so in Europe and Asia than in the United States.

The multinationality of large U.S. multinationals presumably brings to potential foreign investors both rationally process-able information and irrationally process-able "familiarity." Recent research indicates that home equity bias, as between pairs of nations, is significantly lower the more the two nations engage in bilateral goods trade.[68] In particular, trade trumps distance. That is, when trade is included as a variable along with geographical distance, the bias-increasing impact of distance evaporates or is reversed.[69] This research is on a countryby-country basis; it measures how the holdings of one country's invest- ments by another country's investors, aggregated over all the assets originating in the former, is affected by the degree of bilateral trade between the two nations. It is a fair hypothesis that the effect is heightened if one zooms in on the particular companies within each country that are most engaged in trade with the other country.

Unfortunately, the extant literature on home equity bias does not break out large U.S. multinational enterprises from other companies.[70] It thus seems worth attempting to step beyond that literature to see if finer grained information is available.

IV. The Treasury International Capital System

Various agencies of the U.S. government collect data on cross-border securities holdings, including foreign holdings of U.S. stocks.[71] The bulk of the work is divided across agencies according to whether the foreign resident holds more than 10% of the voting power of the U.S. entity. The Treasury International Capital (TIC) system-a joint undertaking of Treasury, the Federal Reserve, and the Federal Re- serve Bank of New York-collects data on "portfolio holdings" (less than 10%). The Commerce Department's Bureau of Economic Analysis collects data on "direct investment" (10% or greater).

This Part focuses on TIC surveys of portfolio holdings. These numbers are probably more relevant than direct investment reports in relation to most large U.S. multinationals.[72] Furthermore, they and their foreign analogues[73] appear to be the most frequently employed in the literature on home bias.[74] Figure l[75] shows the TIC-reported portion of U.S. equity (by value) held by foreigners for each of the annual TIC surveys since Congress authorized the modern survey in the mid- 1970's. The percentage climbs from 3.8% in 1974 to 13.6% in 2012.

A. Where the Numbers Come From

In sum, the TIC system requires U.S. entities whose securities are held by foreign investors, or who are holding other U.S. entities' securities on behalf of foreign investors, to report those holdings to Treasury.

A review of the details of TIC reporting shows that, in certain re- spects, the TIC surveys are unusually comprehensive and reliable.

First, the TIC system's sample design is institutionally comprehensive. The surveys are not sent to some vast list of potential foreign investors. Such a list would be difficult to compile (and those on it, hard to compel). Rather, the surveys are sent to the various U.S.resident entities that lie along the chain from the U.S. issuer of the equity to the equity's ultimate foreign owner. The U.S.-resident entity that reports in each case is the last U.S. entity along the chain before it crosses the border.[76]

For example, if a foreign investor's U.S. stock is held in custody by a U.S.-resident broker-dealer, the U.S. broker-dealer reports the holding. The ultimate foreign owner at the end of the chain does not report her holdings. Neither does the U.S. issuer at the start of the chain.[77] On the other hand, the U.S. issuer does report foreign-held shares if no U.S.-resident custodians are interposed. This is so even if the reported holder is a foreign custodian, and so not the last link in the chain from issuer to owner.[78] Since the issuer is always a U.S. resident by definition, contingent reporting by issuers acts as a backstop: There always is a U.S. entity before the chain crosses the border.[79]

Most reports come via U.S.-resident broker-dealers and U.S.-resident custodian banks.[80] The danger that both a custodian and the issuer will report a single security holding is mitigated by reviewing the data on a security-by-security basis, in the eight months that separate (annual) collection and preliminary release.[81]

Second, response rates are high if not maximal. Survey recipients are required by law to respond. The requirement is enforced with meaningful penalties. And, by survey design, those subject to such requirements and penalties are within the legal and practical jurisdiction of the United States. To be sure, the fines are modest. On the other hand, for the officers, directors, employees, or agents of surveyed entities, knowing participation in the entity's willful failure to respond is punishable by up to a year in prison.[82]

Third, there is reason to think that response under the TIC survey is not only more likely than for the typical survey, but also more accu- rate. Survey recipients respond under oath,[83] which puts federal perjury statutes into play. Under Federal Sentencing Guidelines these felony statutes carry with them the possibility of up to twenty-one months in prison for a first-time offender.[84] To be sure, tax returns are also filed under oath and tax cheating does occur. Yet for TIC surveys the other side of the balance is perhaps not as full: What one chooses to report does not directly impact one's balance sheet.

Fourth, the sample size for the TIC surveys approaches the full universe. Every five years (next in June 2014),[85] a "benchmark survey" is sent to "all significant U.S.-resident custodians and issuers,"[86] the word "custodians" being very inclusively defined.[87] Annually,[88] surveys are sent to "those institutions that collectively report at least 95% of the market value of foreign holdings as measured by the preceding" benchmark survey.[89] Estimation techniques are then used to update the most recent benchmark results to the current year.[90]

FIGURE 1[91]

PERCENTAGE FORIEGN OWNERSHIP U.S. EQUITY U.S. TRESUREY TIC ANNUAL REPORTIS

Fifth, the data collected from each reporter is surprisingly detailed. A broker-dealer, for example, fills out a separate form (or rather submits a separate line of computer code) for every security held by every foreign resident (perhaps itself a custodian) in its accounts. The result is a data set with several million records.[92]Each record indicates, among other things, the CUSIP number of the equity security (from which the issuer and class of shares may be determined[93]), the country of the foreign holder, the number of shares the foreign resident holds, and whether the foreign holder is an official institution, natural person, or personal trust/investment vehicle.[94]

What more could one want? With a large enough spreadsheet it presumably would be possible to sort holdings by CUSIP number and find out the foreign ownership share of any U.S. stock, including that of any multinational. One would only need to compare the dollar value of foreign holdings so determined with each company's total market capitalization, which is easily discovered. However, there are problems-problems that are arguably disqualifying in this context.

B. The Acknowledged "Foreign" Denominator Problem

The problems of primary concern here are not those that are perennially recited in TIC staff reports[95](and sometimes mentioned in academy writing[96]). One of these, however, is relevant and worth mentioning-in part because it appears to cast some doubt on the ro-bustness of home equity bias as it pertains to foreign investment in the United States.

Note that the data collection just described, however comprehensive in certain respects, is insufficient to construct Figure 1. That figure plots a ratio (expressed as a percentage): foreign holdings of U.S. equity divided by total U.S. equity. The TIC survey collects information solely on the numerator. The denominator must be separately supplied. This is true of the government reports from which this figure is derived, and it is also true when TIC numbers (or their foreign analogues) are used to measure home country bias.[97]

The fact that the denominator is separately sourced from the numerator raises the risk that the two numbers are stated on a different basis. More precisely, the danger is that this discrepancy produces a systematic bias and is not fairly treatable as a kind of random error that washes out with averaging.

That this risk is real is partly evinced by the fact that every report of TIC holdings data since the first has included in a prominent position the warning-reminiscent of French and Poterba's "crude estimates" reference[98]-that the figures it presents on foreign ownership share are "rough indicators."[99] "It is not possible," the reports explain, "to obtain data on the total value outstanding by security type on exactly the same basis as the [TIC] survey data are collected."[100]

How bad a problem could this be? The TIC reports obtain their denominators (total U.S. equity) from the Federal Reserve Statistical Release Z.l Flow of Funds Accounts for the United States.[101] These denominators are estimates based on an amalgam of statistics provided by private data purveyors.[102] Some clue as to the full set of difficulties is provided by the surprisingly back-of-the-envelope methods used to value closely-held shares, which are included in total U.S. equity, and intercompany shares, which are not. The market value of shares in close corporations is calculated by multiplying book value by the ratio of market value to book value that is observed for public companies in similar industries.[103] The result is then multiplied by one quarter to account for illiquidity.[104] Intercompany holdings are valued at book multiplied by three.[105] The documentation for these series provides no indication of the appropriate range of error for these multiples. And it seems likely that many studies presenting estimates based on these data, including those in the home bias literature, do not include this error when (and if) they report an error range for their own estimates.

C. Problems Identifying Issuer Characteristics

The problems of primary importance in this Article, which receive little or no attention in the literature, concern the impossibility of breaking out large U.S. multinationals from the full set of U.S. companies. To some extent these problems are a matter of what the TIC system knows; to some extent, they are a matter of what it reports.

1. "Fund" Opacity

Suppose a foreign investor invests in a U.S.-organized "fund." The word "fund" is defined for TIC reporting purposes broadly enough to include mutual funds, hedge funds, and apparently even "separate accounts."[106]Suppose that this "fund" in turn invests in U.S. securities.

The foreign investor's investment in the "fund" shows up in the TIC survey as a foreign holding of U.S. equity.[107] But, looking through the fund, what kinds of U.S. securities does the foreign investor really own? How much in particular is she investing in the stocks of large U.S. multinationals?

Even with unlimited access to the data and unlimited time, one would not be able to determine the portion of the "fund" that is invested specifically in large U.S. multinationals. One could not even determine how much of the "fund" was invested in stocks, as opposed to bonds, money market instruments, or other securities,[108] or in U.S. securities versus foreign.[109] The TIC survey asks about the U.S. stock holdings only of foreign residents.[110] The "fund" is not a foreign resident.[111] Its holdings are not reported.

This problem is evident in a breakdown that appears in the appendix of the latest annual TIC report on foreign holdings.[112] Table All shows foreign ownership of U.S. equity by industry of U.S. issuer. This is, in fact, a significant coarsening of the raw data, since the TIC system itself knows CUSIP numbers. Yet the table is more informative regarding what the TIC system does not know. The table shows an outsized number for foreign investment in "Capital Markets (including Mutual Funds)." The number amounts to a full 25% ($1.1 trillion versus $4.2 trillion) of all foreign holdings in U.S. equities.[113]

2. Confidentiality

In any event, TIC reports provide no breakdown by U.S. issuer beyond the breakdown by broad industry group just described. In particular, there is no breakdown according to the U.S. issuer's multinational status-or any workable proxy thereof.[114] Thus, even if one were willing to extrapolate from the subset of foreign investment that does not go through "funds," this is not possible based on what is currently disclosed.

As noted, the TIC system does have detailed information on issuer characteristics. Might TIC someday report more of this information? Perhaps. The statutes and regulations authorizing the TIC system do specifically guard against the revelation of any information that might permit public discernment of individual holdings.[115]But it seems unership likely that finer-grained reporting about issuer characteristics would permit incisive deductions about any given investor's portfolio. If one knew, for example, that large U.S. multinationals exhibited a greater foreign ownership share than other U.S. companies (putting aside the fund opacity issue), this would indeed provide information about the average portfolio of foreign investors. But it would not say much about the portfolio of any particular foreign investor.

For the time being, however, restrictions on TIC reporting are layered on top of "fund" opacity, and the two together make it nearly impossible to get to the bottom of who owns large U.S. multinationals. But perhaps there are other places to look.

V. "INSTITUTIONAL INVESTMENT MANAGERS" UNDER THE SECURITIES EXCHANGE ACT

On the question "Who owns Google?" one apparently can get quite far with relatively little effort. Googling "Who owns Google?" brings up a 115-page list of Google Inc.'s 1716 "institutional owners" with information on each "owner's" share.[116] Unfortunately, it is not as easy to bring up an understanding of what these numbers mean.

The Securities Exchange Act of 1934 was amended in 1975 to add § 13(f), which requires large "institutional investment managers" (IIMs) to disclose on a quarterly basis the number of shares and the market value of the "§ 13(f) securities" held in accounts over which they exercise "investment discretion."[117] The class of § 13(f) securities includes most stocks traded on U.S. national securities exchanges, including the New York Stock Exchange (NYSE) and NASDAQ.[118]

The class is thus wide enough to include the shares of most large U.S. multinationals.

In making its quarterly report, each IIM submits to the SEC a table of stock holdings. The rows of the table correspond to individual § 13(f) securities. The columns include, among other things, the number of shares and their market value. A given cell of the table within these columns shows the number of shares or the market value, as the case may be, of a given § 13(f) security summed across all accounts over which the IIM exercises "investment discretion."[119]

Table 1 shows a portion of the table filed on November 14, 2013 by FMR LLC, the U.S. arm of the Fidelity investment management "complex."[120]

TABLE 1

Selected Rows and Columns from FMR LLCs Form 13F of November 14, 2013

TABLE 1

Selected Rows and Columns from FMR LLCs Form 13F of November 14, 2013

These tables are filed in electronic form with the SEC, and the SEC posts them on its public website as received.[121] Information intermediaries such as Yahoo! Finance and NASDAQ.com then convert the mass of tables as so disclosed-one table for each reporting IIM, each table listing many issuers-into a new collection of tables-one for each issuer, each listing many reporting IIMs.

These "inverted tables" are easily accessible online. And they are what come up when one Googles "Who owns Google?" Table 2 shows several columns from the first of the 115 pages of the inverted table for Google Inc., as it appears on NASDAQ.com.[122]

TABLE 2

First Rows and Selected Columns of Inversion of Form 13F filings for Google Inc from NASDAQ.com (as of Feb. 7, 2014)

Viewing this table with an eye toward determining Google's foreign ownership share, one's first reaction might be that the only barriers to enlightenment are time and research assistance. Presumably one could go through all 1700 entries-or a sample thereof-determine by other means whether the "owner" was U.S. or foreign, and then add up the shares within each group.

Unfortunately, there are two serious problems with this plan: gaps in reporting and a potentially multilayered opacity as between those who report and those who own.

A. Reporting Gaps

Two kinds of reporting gaps plague § 13(f) disclosures: those written into the statute and those that arise in practice.

1. Gaps in the Statute

Despite the heading of NASDAQ.com's first column, the reporting IIMs listed in Table 2 are not necessarily "owners." Rather, they are an amalgam of two kinds of legal persons. An institutional investment manager is either (1) A non-natural person investing for its own account-such as the California Public Employees Retirement System (CalPERS), or (2) A person, natural or non-natural, that exercises "investment discretion" over the account of some other person, natural or non-natural-such as mutual fund managers, hedge fund man- agers, and managers of separate accounts.[123] In other words, while non-natural persons always report-whether investing for their own account or someone else's-natural persons report only when investing for someone else's account.

Furthermore, there are two kinds of smallness exceptions. First, IIMs who exercise investment discretion over holdings that total less than $100 million, aggregated over all accounts and all § 13(f) securities, need not file Form 13F at all.[124] Second, even an IIM that must file Form 13F need not report holdings of a given § 13(f) security if the IIM's managed holdings of such security, aggregated over all accounts, consist of fewer than 10,000 shares and such holdings have a market value of less than $200,000.[125]

Lastly, though NASDAQ.com or Yahoo! Finance do not make this clear, it should be kept in mind that the § 13(f) reports from which these intermediaries derive their tables generally do not include shares that are not listed on a national securities exchange.[126] In exceptional cases (one in particular, per the next paragraph) a non-negligible portion of a multinational's equity in thereby neglected.

These statutory reporting gaps appear to have a substantial impact on § 13(f) reporting. In some cases the reporting gap can be filled with other disclosures. For Google, § 13(f) reports account for 84% of the company, leaving a 16% reporting gap.127 But the missing 16% is accounted for by an unlisted class of shares held by (U.S.) co-founders.[128] In other cases, the gaps are more mysterious. Cisco and Apple shares are all publicly traded.[129] Yet for Cisco, the gap is 27%;[130] for Apple, 40%.[131]

Of course, it is not clear that the statutory reporting gaps just described create a bias with respect to foreign versus U.S. ownership. Reporting gaps due to enforcement, however, would seem to do so a priori.

2. Enforcement Gaps

Given that a security falls within the subset of § 13(f) securities, the statute that requires reporting makes no distinction based on either the nationality of the IIM or the nationality of the person who has granted the IIM investment discretion. It is true that reporting is required only of IIMs who use a "means or instrumentality of interstate commerce" in the course of their business as an IIM. But, as in many other areas, this "interstate commerce" hook is actually a wide net. "Interstate" includes "international."[132] It also explicitly includes use of a U.S. national securities exchange.[133] Thus a large Turkish investment management company managing Saudi money invested in U.S. companies through the New York Stock Exchange must, at least technically, report these Saudi holdings.

There are reasons to believe, however, that the reality of the situation differs from this directive. Recall that the TIC system, discussed in Part IV, was designed so that the reporting obligation rests inside the United States. Given that reporting obligations under § 13(f) may well lie overseas, one might imagine that § 13(f) has a more robust enforcement apparatus than the TIC system. But the opposite appears to be true. A 2010 report by the SEC's Inspector General casts a fairly dark shadow on the accuracy of § 13(f) reporting generally.[134] It seems reasonable to suppose that the shadow grows larger, the farther one moves from SEC offices.

B. Intermediary Opacity

Section 13(f) reporting was not designed to reveal owner characteristics such as nationality. Rather it was designed to reveal the choices of large portfolio managers.[135] The precise justification for this revelation has never been entirely clear. Apparently, it was thought that individual investors might find it useful to know where the "smart money" was parked-whether their plan was to imitate or avoid.[136] Furthermore, it was anticipated that regulators would make use of such information in reining in the entities doing the parking.[137] Whose money was being parked was never the concern.

This focus of attention-whatever its justification-is clearly reflected in the design of § 13(f). The entity with investment discretion reports. That entity may well be a manager that organizes and/or manages several funds and separate accounts. Within each fund or account, the investor or investors may be U.S. or foreign. The IIM reports-in each cell of tables such as Table 1-merely the total holdings of a given stock across all the funds and accounts over which it has investment discretion.

In a few cases large investors of identifiable nationality can be readily and exclusively associated with a particular reporting investment manager. In most other cases, however, this is not possible. As discussed in the next three Subsections, the § 13(f) postures of two of the largest sovereign wealth funds and the single largest IIM reporter for Google illustrate the range of possibilities.

1. The Government Pension Fund Global of Norway

Norway's Government Pension Fund Global (Norway GPFG), sourced from taxes, fees, and state-earned profits from North Sea oil,[138] stood at $818 billion the end of 2013, making it the single largest sovereign wealth fund in the world.[139]

Norway GPFG's disclosures represent the high water mark for transparency. With specific regard to investment management, Norway GPFG reports that it is run by the Norwegian Ministry of Finance, which in turn delegates its management to the Norwegian Central Bank (Norges Bank), which in turn manages the fund out of a division, Norges Bank Investment Management (NBIM). Norges Bank files Form 13F.[140]

NBIM also hires "external equity managers."[141] But given NBIM's extensive disclosures under Norwegian law, it is possible to determine with fair confidence that these external managers are not managing Norway GPFG's holdings of major U.S. equities.[142] Consequently, such holdings would not show up in the § 13(f) reports, if any, of these managers.

One can thus be fairly confident that Norges Bank's Form 13F is exhaustive of, and consists solely of, the investments of Norway GPFG. The most recent public 13(f) disclosures (as of this writing) show that, as of September 30,2012, Norges Bank had investment discretion over $148 billion in holdings of § 13(f) securities.[143] Specifi- Fcally, the report lists 2,116,195 shares of Google Inc. worth $1.5 billion.[144]

Norges Bank's § 13(f) disclosures crosscheck (roughly) with its disclosures under Norwegian law. NBIM reports that, as of September 30,2013, Norway GPFG invests roughly 64% of its nearly $800 billion total in "equities"[145] and 30% of this (thus 19% of the $800 billion)[146] in U.S. equities in particular. This implies dollar holdings of about $149 billion (as compared to $148 billion) in U.S. equities. Further, NBIM reports that, at the end of 2012, Norway GPFG owned about $1.4 billion in Google stock (as opposed to $1.5 billion).[147]

2. China Investment Company

While the manner in which Norway invests its $818 billion of North Sea oil money is largely discernible from public disclosures, the same is not true of the manner in which China invests its $575 billion in foreign reserves. These reserves are managed by the China Investment Company (CIC) and constitute the fourth largest sovereign wealth fund.[148] Important for the present purpose, there appear to be no comparable laws in China mandating public disclosure of holdings, stock by stock, as in Norway.

We do know that CIC International hires external managers. But we do not know who these managers are and how much they are assigned to manage. The result is that CIC's holdings are scattered in indiscernible fashion among the § 13(f) reports of an unknown number of IIMs.[149]

In the limited reporting that CIC does, it discloses that 64% of its money is "externally managed."[150] It also discloses that, as of December 2012, it invests "32%" (of what we cannot be sure) in public equities,[151]and that nearly half of this is in the United States.[152] The size of the full fund was about $500 billion at the end of 2012,[153] but this includes domestic investment, and there appears to be no way to separate out the absolute amount invested outside of China.

3. The Fidelity "Complex"

The largest block of holdings of Google shares is reported by "FMR LLC," the U.S. arm of the Fidelity investment management "complex." FMR LLC reports a total of $653 billion in § 13(f) securities.[154] Of this, $16 billion is invested in Google, constituting 5.5%.[155]

If one scrolls down through pages of the "inverted table" that are not shown in Table 2, one also finds the reporter "FIL Ltd." FIL Ltd., the foreign arm of the Fidelity "complex," reports only $21 billion in § 13(f) securities, of which $207 million is invested in Google.[156]

Fidelity's structure well illustrates the haze surrounding the identity of ultimate owners in § 13(f) reports. A private company-or rather a set of private companies with uncertain connection-Fidelity's structure is byzantine and ever-shifting. Making deductions about the national composition of its reported holdings would itself constitute a substantial research undertaking, with no guarantee of success.

Fidelity's promotional materials and disclosures present two facts about how FMR LLC and FIL Ltd. are related. First, they are not. Or at least they do not regard each other as under common control. For this reason neither shows up in the other's § 13(f) filing,[157] as would be the case were one controlled by the other, or both controlled by a third entity.

Second, FIL Ltd., which was once part of Fidelity, but was separated in 1980, has as its specific directive the handling of investments by non-U.S. investors.[158]

Given the second fact, one is tempted to conclude that the $16 billion of Google holdings reported by FMR LLC is ultimately owned by U.S. investors, while foreign investors' Google holdings are contained in the meager $207 million reported by FIL Ltd. One might regard this as evidence pointing toward the inference that Google is mostly U.S.-owned.

There are problems, however, with this deduction. First, not all Fidelity's foreign investors need invest through funds organized by FIL Ltd. It is true that European investors in particular may have tax and regulatory reasons for investing through funds that are organized in Europe rather than in the United States.[159] However, not all foreign owners of Google need be from Europe (or regions with similar laws favoring foreign-organized funds). Furthermore, not all foreign owners of Google need invest through funds. If, for instance, FMR LLC manages some portion of CIC money in a "separate account," that account would be an (indiscernible) addend in FMR LLC's § 13(f) reporting.

Second, it is important to remember again that § 13f reporting is done by the entity with investment discretion. The entity with investment discretion may differ not only from the owner of the shares but also from the entity that organizes the fund. If FIL Ltd. organizes a fund and wishes to have some portion of it invested in U.S. stocks, it may well appoint FMR LLC as sub-adviser. In this case, FMR LLC might well have "investment discretion" for purposes of § 13(f), and the holdings would show up in the Section 13(f) reporting of FMR LLC, not FIL Ltd.[160]

It is difficult to know exactly how much this happens as between these two entities. But there is good reason to believe that it is prevalent.[161] After all, the two companies, despite their reporting posture, are surely not completely separate.[162] It seems a fair guess that FIL Ltd. would not endeavor to reinvent the wheel in choosing U.S. stocks.

VI. Disclosures Under the Investment Company Act and the Investment Advisers Act

Two other reporting regimes administered by the SEC are potential sources of information on the foreign ownership share of large U.S. multinationals. The first is the reporting regime for mutual funds under the Investment Company Act of 1940.[163] The second reporting regime, under the Investment Advisers Act of 1940, is imposed not on mutual funds, but on "investment advisers," including, but not limited to, those that manage mutual funds.[164]

A. The Investment Company Act

1. Mutual Funds

With a leap of faith-albeit one of uncertain and ever-shifting span-it is possible to make some headway on ownership nationality by studying the quarterly disclosures of mutual funds[165] on Forms NCSR and N-Q.[166] With regard to segregating large U.S. multinationals from other U.S. companies, these disclosures fully reveal the funds' portfolios. With regard to identifying the nationality of fund shareholders, tax and regulatory considerations make it not unreasonable- at least as a first approximation and subject to the important caveats discussed below-to suppose that any given U.S. mutual fund is owned mostly by U.S. residents.

To illustrate the last point, consider U.K. investors. There are reasons to believe that a U.K. investor interested in U.S. equity and specifically interested in making that investment through participating in a regulated collective investment scheme would eschew U.S. mutual funds and instead turn to funds organized in the United Kingdom or elsewhere in Europe.

First, although several tax disadvantages to investing through a U.S. mutual fund have gone by the wayside in recent years,[167] at least one tax-timing disadvantage remains-at least for investors expecting gains rather than losses. The details of this disadvantage are complex and ever-shifting. As of this writing the essential points are these: U.K. investors have available to them regulated collective investment schemes that are simultaneously income-transparent and capital-opaque.[168] Income transparency puts these funds on a par with U.S. mutual funds with respect to dividends received from U.S. portfolio companies-specifically, with respect to U.K. foreign tax credit relief for U.S. withholding tax.[169] The relative advantage of these funds over U.S. mutual funds lies in the implications of capital opacity for the taxation of gain from the sale of portfolio company stock.[170]

(Editor: Zhang Chi)

Personal details:

Chris William Sanchirico, University of Pennsylvania Law School; University of Pennsylvania Wharton School - Business Economics and Public Policy Department

Footnotes: 

[1] As of this writing (March 2014), the issue has sparked several high-profile congressional hearings, and figures large in tax reform proposals put forward by the President and key congressional leaders. It has also garnered front-page, Pulitzer-prize winning media coverage. Moreover, the issue is also salient in other developed nations.

[2] See note 6 for sources describing this structure and other similar structures.

[3] Senate Apple Hearing, note 1, exhibit 19, at 284-89, contains an appendix listing the twenty companies with the largest accumulation of shifted profits (or rather an imperfect measure thereof). The list, entitled "Top 20 Companies for 2012 with Foreign Indefinitely Reinvested Earnings" was compiled by Audit Analytics and contains the following companies (and amounts): General Electric ($108 billion ), Pfizer ($73 billion), Microsoft ($61 billion), Merck ($53 billion), Johnson & Johnson ($49 billion), IBM ($44 billion), Exxon ($43 billion), Citigroup ($43 billion), Cisco ($41 billion), Apple ($40 billion), Abbott Laboratories ($40 billion), Procter & Gamble ($39 billion), HP ($33 billion), Google ($33 billion), PepsiCo ($32 billion), Coca-Cola ($27 billion). Chevron ($27 billion), JPMorgan Chase ($25 billion), Amgen ($22 billion), and United Technologies ($22 billion). Id. at 288.

[4] Prominent criticisms include (but are not limited to): J. Clifton Fleming Jr., Robert J. Peroni & Stephen E. Shay, Worse Than Exemption, 59 Emory L.J. 79, 84-85 (2009) (argu- ing that the present U.S. tax system-which nominally taxes the worldwide income of U.S. residents, but in fact offers, among other things, opportunities for profit shifting-is "worse" than a system that altogether exempts foreign source (active business) income, with regard to "distort[ing] taxpayer decisions in bizarre ways, and [being] inequitable [in] allowing] residents who earn foreign-source income to avoid the tax burden borne by their fellow residents"); Edward Kleinbard, Stateless Income, 11 Fla. Tax Rev. 699, 702, 706 (2011) (defining stateless income as "the movement of taxable income within a multinational group from high-tax to low-tax source countries without shifting [real economic activity]" and arguing that, "[w]hen unchecked, stateless income strips source countries... of the tax revenues generated in those jurisdictions^]... distorts the investment decisions of multinational firms, and... distorts a U.S. multinational firm's decision whether to repatriate that stateless income back to the United States"); Stephen C. Loomis, The Double Irish Sandwich: Reforming Overseas Tax Havens, 43 St. Mary's LJ. 825, 845 (2012) ("[T]he effect of the Double Irish is clearly more negative than positive, and reform is needed to close the loophole.").

[5] On competitiveness, see Section II.A and notes 11-13.

[6] Most of what is publicly known about these structures (and others like it) appears ultimately to derive from three sources: (1) the three congressional hearings listed in note 1; (2) a 2007 article in the practice literature, Joseph B. Darby III & Kelsey Lemaster, Double Irish More Than Doubles the Tax Saving: Hybrid Structure Reduces Irish, U.S. and Worldwide Taxation, Prac. U.S./Int'l Tax Strategies, May 15, 2007, at 2; and (3) the investigative reporting of Jesse Drucker at Bloomberg News, as listed in note 1. Each of these accounts leaves important structural questions unanswered.

This account of primary sources is supported by Kleinbard who also draws on Drucker as well as Darby and Lemaster. See Kleinbard, note 4, at 707 n.ll ("The facts [describing Google's double Irish sandwich] are drawn principally from [Drucker, Google 2.4% Rate, note 1] as supplemented by inferences drawn from [Darby & Lemaster, supra]. Since Google's tax planning is not transparent to outside observers, it is possible that there are some slight mischaracterizations of details in the text, but these would not change the thrust of the points made therein.").

[7] The laws and rules governing these structures include the Code, Treasury regulations, various bilateral income tax treaties between the United States and other nations, the tax laws, regulations, and treaties of "tax havens" such as Ireland, the Netherlands, and Bermuda, and the tax laws, regulations, and treaties of countries where the goods and services are ultimately sold or used.

With regard to flux, recent changes to U.S. tax laws with important and in some cases uncharted implications in this area include: (1) the "contract manufacturing regulations," modified in 2008, extending an exception to the rules that require current parent taxation of foreign subsidiary income, Reg. § 1.954-3(a)(4)(iv) (providing a "substantial contribution test" for the "own-manufacturing" exception to the definition of "foreign base company sales income," which is passed through to the parent); T.D. 9438, 2009-5 C.B. 387, corrected in 74 Fed. Reg. 11,843 (Mar. 20,2009); and (2) the "platform contribution regulations," finalized in 2011 and meant to ensure arm's length accounting for certain in kind contributions of pre-existing intangibles in cost-sharing arrangements among affiliates for further intangibles development, Reg. § 1.482-7(b)(ii), (c), (g); T.D. 9568, 2012-12 I.R.B. 499, corrected in 77 Fed. Reg. 8143-44 (Feb. 14, 2012) and 77 Fed. Reg. 3605-06 (Jan. 25, 2012).

[8] What do I mean by "nationality" in relation to a given shareholder? I mean it in whatever sense it is being invoked in the current debate with regard to companies. See Section II.A for a discussion of arguments based on U.S. "competitiveness." For concreteness, however, one may consider nationality to be coincident with citizenship.

[9] What do I mean by "large U.S. multinationals"? I mean the companies whose taxes are in question in the current debate over U.S. international taxation. For a sense of the term's heartland, if not its boundaries, see the list in note 3 of the twenty companies with the largest permanently reinvested earnings and the list in note 11 of the members of the former Win America Campaign coalition.

[10] Nicolas Coeurdacier & Hélène Rey, Home Bias in Open Economy Financial Macroeconomics, 51 J. Econ. Literature 63, 64, 66 (2013) (providing a recent survey of the literature on home country bias and using the phrases "very robust portfolio fact" and "robust stylized fact" to describe the empirical phenomenon).

[11] The "Win America Campaign" is the name of a prominent lobbying coalition-whose operations are currently, perhaps temporarily, suspended-that argued for, among other things, a second repatriation tax holiday, as discussed in the next Section. The members of the Win America Campaign include(d): Adobe Systems, Apple, Broadcom, BrownForman, CA Technologies, Cadence Design Systems, Cisco Systems, Devon Energy, Duke Energy, EMC, Google, Eastman, Loews, Microsoft, Oracle, Pfizer, and Qualcomm. See Factbox: WIN America Members Include Microsoft, Others, Reuters, Apr. 24,2011, http://www.reuters.com/article/2011/08/24/us-usa-tax-holiday-factbox-idUSTRE77N43720110 824.

It is possible that the Win America Campaign has been superseded, at least partly and/or temporarily, by the "LIFT America Campaign," a similar but not identical coalition. ("LIFT" stands for "Let's invest for tomorrow.") See LIFT America, http://www.liftameri cacoalition.org/about-us/ (last visited Nov. 18, 2014).

[12] See, e.g., R. Glenn Hubbard, Tax Policy and International Competitiveness, Taxes, Mar. 2004, at 213, 219:

Increasingly, the markets for U.S. companies have become global, and foreign-based competitor companies operate under tax rules that are often more favorable than our own. The existing U.S. tax law governing the activities of multinational companies ... can result in circumstances that harm the competitiveness of U.S. companies. . . . [T]he current U.S. international tax rules should be reviewed with an eye to ... removing impediments to U.S. international competitiveness. ... If U.S. businesses are to succeed in the global economy, the U.S. tax system must not generate a bias against their ability to compete effectively against foreign-based companies.

For a critical analysis of competitiveness claims, see Jane G. Gravelle, Does the Concept of Competitiveness Have Meaning in Formulating Corporate Tax Policy?, 65 Tax L. Rev. 323, 347 (2012) ("International competitiveness is not only an inappropriate standard, but it also leads to poor policy outcomes.... It is important that public discourse move away from meaningless concepts such as international competitiveness and instead focus on what tax policies might be optimal, in that they fulfill the more appropriate objective of 'better off.'").

[13] See, e.g.. Tax Found., Corporate Taxes: Winning, YouTube (Feb. 14, 2012), https:// www.youtube.com/watch?v=ws36P3BkDF4 (arguing that American workers "win" when American companies do).

[14] The theoretical and empirical literature on the "incidence" of the corporate tax-that is, on the question of who bears the real economic burden of the corporate tax taking account of tax-responsive changes in wages, returns, and prices-is large, varied, and collectively inconclusive. See Jim Nunns, Urban-Brookings Tax Pol'y Ctr., How TPC Distributes the Corporate Income Tax 2 (2012), available at http://www.taxpolicycenter.org/ UploadedPDF/412651-Tax-Model-Corporate-Tax-Incidence.pdf ("The incidence of the corporate income tax is perhaps the least settled issue in distributional analysis."). The Tax Policy Center (TPC), a well-respected nonpartisan think tank that provides widely-quoted estimates of the distributional effects of tax changes, often needs to determine the incidence of the corporate tax in the process of constructing its distributional estimates. Id. at 1. For this purpose TPC uses the mid-range of incidence estimates in the literature. Id. at 8. According to its review of this literature, labor's share of the corporate tax burden ranges from 12% to 28%, the share of all capital (corporate and noncorporate) ranges from 9% to 24%, and the share borne by corporate shareholders alone ranges from 60% to 70%. Id. tbl.l.

[15] See IRC § 11 (corporate rates). The dividends received deduction in § 243, which allows parent corporations to deduct from their income dividends received from their subsidiaries, is generally not available when the subsidiary is a foreign corporation. Further, the dividends received deduction in § 245 applies only to the "U.S.-source portion" of dividends, which is essentially the portion attributable to that part of the foreign subsidiary's income that is either effectively connected to a trade or business in the United States or derived from dividends paid by a lower-tier U.S. subsidiary. IRC § 245(a)(5). At the time it receives a dividend, however, the parent will receive a tax credit for any foreign taxes paid by the subsidiary. IRC § 902 ("deemed paid credit").

[16] The present value of a future cash flow is equal to the dollar value of the cash flow reduced to take into account forgone investment earnings between the present and the time the flow is received.

[17] See Alvin C. Warren Jr., Income of Foreign Subsidiaries: A Review of the Basic Analytics, 145 Tax Notes 321, 321-22, 324 (Oct. 20, 2014) (describing the benefits of international deferral).

[18] See IRC § 965 (providing a "dividends received deduction" to U.S. shareholders equal to 85% of the cash dividends of "permanently reinvested earnings" received from a "controlled foreign corporation" up to $500 million, thus reducing the top statutory rate from 35% to 15% of 35%, or 5.25%).

[19] See Melissa Redmiles, The One-Time Dividends Received Deduction, 1RS Stat. Income Bull., Spring 2008, at 102,103 (1RS analysis based on data from Form 8895).

[20] Id.

Some 843 corporations, a relatively small number of corporations given that roughly 9,700 corporations had [controlled foreign corporations] in 2004, took advantage of the deduction. But these corporations repatriated almost $362 billion. Of that, $312 billion qualified for the deduction, creating a total deduction of $265 billion. In comparison, $804 billion of end-of-year, accumulated, nontaxable earnings and profits were reported for all controlled foreign corporations of all U.S. corporations for Tax Year 2004, the last tax year for which this statistic is available.

[21] Senate Apple Hearing, note 1, exhibit 19, at 286, shows, for 2012, $1,943 trillion of "total foreign indefinitely reinvested earnings" for the "Russell 3000." The Russell 3000 represents "the largest 3,000 U.S. companies representing approximately 98% of the investable U.S. equity market." Russell Investments, "Russell Indexes," https://www.russell .com/indexes/americas/indexes/fact-sheet.page?ic=US3000 (last visited Apr. 24, 2015). I convert $1,943 trillion to $1.65 trillion 2005 dollars using the Bureau of Labor Statistics CPI Inflation Calculator at www.data.bls.gov.

Update'. Since this Article was accepted for publication in March 2014, new estimates have become available. See Audit Analytics Staff, Overseas Earnings of Russell 1000 Tops $2 Trillion in 2013, (Apr. 1, 2014), available at http://www.auditanalytics.com/blog/overseas-earnings-of-russell-1000-tops-2-trillion-in-2013/ (showing, for 2013, $2,119 trillion for the Russell 1000 (contrast: Russell 3000)).

[22] See Richard Rubin, Cisco-Backed Repatriation Tax-Break Lobby Effort Ceases, Bloomberg News, Apr. 24, 2012, http://www.bloomberg.com/news/2012-04-23/repatriationtax-lobbying-campaign-said-to-disband.html (describing the lobbying coalition Win America Campaign, its recent suspension, the possibility of its reinstatement, and the ongoing efforts of individual coalition members): Martin A. Sullivan, The Economic Case for Unlocking Foreign Profits, 136 Tax Notes 7, 7-10 (July 2, 2012).

[23] For example, Claire Buchan Parker, a spokeswoman for the LIFT America Coalition, argues that "[a]s a result of our antiquated tax system, some $2 trillion is locked out of the United States-and that figure grows every day. Reforming the U.S. tax system would allow companies to avoid this lock-out effect, thus freeing up badly needed capital to invest in enhanced research and development, plant expansion, expanded services and products, better worker benefits and quality new jobs for American workers." Claire Buchan Parker, The Path to Jobs and Prosperity: Tax Reform, Roll Call, Jan. 27, 2014, http://www .rollcall.com/news/the_path_tojobs_and_prosperity_tax_reform_commentary-230384-l .html?zkPrintable=true.

[24] See, e.g., Majority Staff of S. Permanent Subcomm. on Investigations, 112th Cong., Repatriating Offshore Funds: 2004 Tax Windfall for Select Multinationals 3,22-23 (Comm. Print 2011) (citing academic literature indicating that repatriation primarily benefited shareholders through stock repurchases and arguing that overall, " [t]he evidence presented in this [report] shows that, rather than producing new jobs or increasing research and development expenditures, the 2004 repatriation tax provision was followed by an increase in dollars spent on stock repurchases and executive compensation").

[25] See Chris Sanchirico, A Repatriation Tax Holiday for US Multinationals? Four Contagious Illusions, TaxVox, Dec. 10,2014, available at http://taxvox.taxpolicycenter.org/2014/ 12/10/repatriation-tax-holiday-us-multinations-four-coutagious-illusions/ (challenging the widely held assumption that foreign earnings are "trapped outside of the country").

[26] See IRC § 956 (taxing parent as if it received a dividend when the subsidiary invests in or loans to the parent).

[27] See IRC § 956(c)(1), (2) (definition of "U.S. property" and exceptions thereto). Note, however, that "U.S. property" does include "tangible property located in the United States" and "any right to the use in the United States of [intellectual property] which is acquired or developed by the controlled foreign corporation for use in the United States." IRC § 956(c)(1)(A) (tangible), (D) (intellectual).

[28] See Majority Staff of S. Permanent Subcomm. on Investigations, 112th Cong., Offshore Funds Located Onshore: Addendum to Repatriating Offshore Funds 5 (Comm. Print 2011) (table showing percentage of undistributed accumulated foreign earnings held in U.S. bank accounts or investments by various large U.S. multinationals).

[29] Kenneth R. French & James M. Poterba, Investor Diversification and International Equity Markets, 81 Am. Econ. Rev. (Papers & Proceedings) 222, 222, 224-25 (1991).

[30] See Piet Sercu & Rosanne Vanpée, The Home Bias Puzzle in Equity Portfolios, in International Finance: A Survey 310, 310, 320 (H. Kent Baker & Leigh A. Riddick eds., 2012) (arguing "no single explanation seems to capture the full extent of international underdiversification" and rejecting, for example, regulatory barriers as an explanation of "massive" home country bias).

[31] French & Poterba, note 29, at 222, 222 n.l (perhaps referring in part to the fact that their holdings data are cumulated from transactions data).

[32] French and Poterba list two data sources for their estimates: the "U.S. Treasury Bulletin" and Michael Howell & Angela Cozzini, International Equity Flows-1990 Edition (1990). French & Poterba, note 29, at 221 tbl.l. "U.S. Treasury Bulletin" is not listed among the article's references but an email from one of the authors indicates that the reference was to tables such as appear in the Winter 1989 Treasury Bulletin. 1989-1 Treas. Bull. 141-45. E-mail from James Poterba to author (Feb. 5, 2014, 08:57 AM) (on file with the author). In turn, the Treasury Bulletin sources these tables to data (on transactions rather than holdings) collected by the Treasury International Capital (TIC) system, which is described in detail in Part IV. Treas. Bull., supra, at 124.

Howell and Cozzini's work is a publication from an investment adviser and data purveyor, which currently operates under the rubric "CrossBorder Capital." See CrossBorder Capital, http://www.crossbordercapital.com/Default.aspx (last visited Nov. 18,2014). Their proprietary data set on foreign holdings and global liquidity flows is apparently derived from an analysis of the International Monetary Fund's "Coordinated Portfolio Investment Survey" (CPIS), to which TIC is the U.S. contribution. See note 71; The Research, CrossBorder Capital, http://www.crossbordercapital.com/aboutus.aspx (last visited Jan. 15,2015) ("Based on IMF data, this research is led by Michael Howell, who has developed it from the early 1980s."); see also Richard Portes & Hélène Rey, The Determinants of CrossBorder Equity Flows, 65 J. Int'l Econ. 269, 276 n.17 (stating that CrossBorder Capital's data for the United States is "virtually the same as" TIC data).

[33] French & Poterba, note 29, at 222 tbl.l. It is not clear why the Japanese percentage for U.S. stocks (3.1) is larger than for stocks of all foreign countries (1.9). The reason may have something to do with how holdings data were deduced in part from transactions data, as mentioned in note 31.

As an aside: French and Poterba report that U.S. investors had 94% of their portfolio in U.S. stocks, 1.3% in Japanese stocks, and 5.9% in U.K. stocks. French & Poterba, note 29, at 222 tbl.l.

[34] Id. at 223. The authors do this based on assumptions about investor objectives and estimates of variances and cross-country correlations in stock returns. Further, they imagine that the global stock market consists only of the six largest national stock markets. Lastly, they assume that foreign currency risk is hedged. Id.

[35] Id. at 224. The numbers in the last two sentences of this paragraph are gleaned from Panel B of Table 2. The following explains both the derivation of the numbers in that panel and their precise relationship to the numbers (5.1 and 2.5) discussed in the text of this Article. In short, 5.1 and 2.5 are actually differences in differences, not mere differences.

In deriving their Panel B of Table 2, French and Poterba take as a benchmark the implied expected returns of a hypothetical investor holding the "global market portfolio"- that is, an investor whose proportional investment in the stock of any particular country is equal to the proportion of global stock value attributable to the stock value of that country's companies. Implied expected returns for this hypothetical investor are those that would make the global market portfolio optimal given estimates, separately supplied, of the actual variances and cross-country correlations of stock returns. Implied expected returns for this hypothetical globalist investor are then compared to implied expected returns, calculated in the same way, for actual portfolio holdings in each country. For U.K. portfolio holdings, for instance, French and Poterba report, in Panel B of Table 2, that implied expected returns for actual U.K. investors, given their tilt toward U.K. stocks, were 4.4 percentage points greater, and for U.S. stocks 0.7 percentage points less, than implied expected returns for the hypothetical globalist investor.

To relate these two numbers to the single U.K. number provided in the text of this Article, note that 5.1 = 4.4 + 0.7. Thus, despite how it is discussed in this Article, the number 5.1 is not actually the difference in implied returns across U.K. and U.S. stocks. Rather it is the difference in implied returns across U.K. and U.S. stocks in the difference in implied returns across the hypothetical globalist portfolio and actual U.K. portfolio holdings. The same caveat applies to the numbers for Japan.

[36] See S&P 500: Total and Inflation-Adjusted Historical Returns, Simple Stock Investing, http://www.simplestockinvesting.com/SP500-historical-real-total-returns.htm (last visited Nov. 18, 2014) (last column of table at bottom of page).

[37] I base this calculation on the following formula:

d = [rüai.pmx - r^ai.preu] + [tusi'üí.pmx ~ tuKrHíi,pMx where d is the amount by which, for a U.K. investor, the real (that is, inflation-adjusted) after-tax return to investment in the United Kingdom exceeds the real after-tax return to investment in the United States, rr,ai.pmx is real pretax return on stocks of country i, where i may represent either the United States or the United Kingdom, and r, is the tax rate on the U.K. investor from returns on stock in com- panies in Country i. I then set rreai.Pmx = r"aiipmx = 0.12, per the text accompanying note 36. A similar calculation was performed for Japan.

[38] See Coeurdacier & Rey, note 10, at 64, 66 fig.l, 67 tbl.l (based on data from the "Consumer Portfolio Investment Survey"); Sercu & Vanpée, note 30, at 315-17 tbl.15.3 (based on "CPIS" data). For more on the puzzling source of Coeurdacier & Rey's data, see note 58.

[39] These numbers are from Sercu & Vanpée, note 30, at 315-16 tbl.15.3. A similar, but not identical, table appears in another recent survey, Coeurdacier & Rey, note 10, at 67 tbl.l. See note 62 for the sources of these data.

Throughout the following discussion I cite to both surveys in parallel fashion, as they appear to come from different approaches within economics-Sercu & Vanpée from the finance literature, and Coeurdacier & Rey from the macroeconomics literature and specifically "Open Economy Financial Macroeconomics."

[40] See Sercu & Vanpée, note 30, at 315-16 (row for U.K.). Note that 100 - 5.5 = 94.5 and 100 - 53.35 = 46.65.

[41] French & Poterba, note 29, at 222 tbl.l.

[42] See Sercu & Vanpée, note 30, at 315-15. (row for Japan). Note that 100 - 9.93 = 90.07.

[43] See id. Note that 100 - 87.37 = 12.63.

[44] French & Poterba, note 29, at 222 tbl.l. As an aside, for U.S. investors, foreign stocks were 65.7% of the world stock market by value (100 - 34.29 = 65.71), but foreign stocks constituted only 21.4% of U.S. investors' equity portfolios by value (100 - 78.61 = 21.39).

[45] See Coeurdacier & Rey, note 10, at 91-95; Roger H. Gordon & James R. Hines Jr., International Taxation, in 4 Handbook of Public Economics 1935, 1961-62 (Alan J. Auerbach & Martin Feldstein eds., 2002); Sercu & Vanpée, note 30, at 320-21.

[46] See Sercu & Vanpée, note 30, at 320-21; see also Coeurdacier & Rey, note 10, at 9195.

[47] See Sercu & Vanpée, note 30, at 318-19; see also Coeurdacier & Rey, note 10, at 7679. This explanation should be distinguished from explanations based on hedging local risk, as discussed below, which often refer to "real exchange rate risk" in models that, in fact, have at most a single international currency.

[48] The large literature on this explanation is recently surveyed in Coeurdacier & Rey, note 10, at 71-91, and Sercu & Vanpée, note 30, at 317-20.

In addition to surveying the literature, Coeurdacier & Rey, note 10, at 81-87, present a "benchmark" model in which hedging local risk unambiguously generates home equity bias. In the model, there are two identical countries, home and foreign, and, respectively, two consumer/worker households, two firms, two capital goods, and two consumer goods. Further, for each country there are two assets: a "real bond" (that pays out in consumer goods rather than currency) and equity in the domestic firm that pays out the return to capital used in production. Households optimally use real bonds to hedge purchasing power risks and equities to hedge (bond return-conditional) labor income risks. The model generates a (bond return-conditional) negative correlation between labor earnings and domestic stock dividends, and this, in turn, produces home equity bias. Id. at 86 (discussion after "(ii)"). The authors find such a (conditional) negative correlation between wages and dividends in the data. Id. at 88.

[49] See R. Gaston Gelos & Shang Jin Wei, Transparency and International Portfolio Holdings, 60 J. Fin. 2987, 2989, 2992, 2996, 2998 tbl.IV, 3015 tbl.AII (studying portfolio allocations of "dedicated global emerging market funds" among "emerging markets" countries, as reported by eMergingPortfolio.com, and finding that, among other things, both "opacity"-as measured by PriceWaterhouseCoopers "O-factor"-and lack of minority shareholder rights adversely affect investment); Sercu & Vanpée, note 30, at 323-24.

[50] See Sercu & Vanpée, note 30, at 321-23; see also Coeurdacier & Rey, note 10, at 9598; Portes & Rey, note 32, at 276 & n.17 (2005) (using financial transactions data from "Cross-Border Capital," which for the United States is "virtually the same as the US Treasury TIC data," and finding evidence for the importance of information asymmetries and frictions, as proxied by geographical distance, telephone traffic, bank branches, and overlap in trading hours). But, with specific regard to geographical distance, see Antonin Aviat & Nicolas Coeurdacier, The Geography of Trade in Goods and Asset Holdings, 71 J. Int'l Econ. 22, 22-23 (2007) (substituting bilateral trade in goods as a proxy).

[51] See Sercu & Vanpée, note 30, at 324-26; see also Coeurdacier & Rey, note 10, at 98.

[52] See Sercu & Vanpée, note 30, at 310 ("[N]o single explanation seems to capture the full extent of international underdiversification on its own."); see also Coeurdacier & Rey, note 10, at 64 ("noting probably all contribute to part of the gap").

[53] See, e.g., Aviat & Coeurdacier, note 50, at 22-23 ("If twenty years ago the segmentation of financial markets could well explain the 'home bias puzzle,' it is not likely to be the case today."); Gordon & Hines, note 45, at 1963 ("[T]ax distortions may explain some part of the observed portfolio specialization, [but] are far too small to rationalize the substantial specialization in portfolios observed in the data."); Michael J. Graetz & Itai Grinberg, Taxing International Portfolio Income, 56 Tax L. Rev. 537, 551 (2003) ("[T]ax differentials [do not] have much explanatory power. In several countries, corporate-shareholder tax integration regimes favor domestic over foreign investment. But these tax effects are far too small to explain the home-country bias seen in the data."); Sercu & Vanpée, note 30, at 320 (similar). Indeed, French and Poterba, note 29, at 224, expressed the same view in 1991, noting that regulatory barriers and tax disincentives are difficult to identify.

[54] See Linda L. Tesar & Ingrid M. Werner, Home Bias and High Turnover, 14 J. Int'l Money & Fin. 467, 485-86 (1995) (relying on TIC data, discussed in Part IV, and FDI data in the U.S. Commerce Department's Survey of Current Business; concluding that "the high transactions rate on foreign investments suggests that investors frequently adjust the composition and the size of their international portfolios, even though much of this activity has little impact on net investment positions," which in turn suggests that "high transactions costs associated with trading foreign securities cannot be the reason for the observed reluctance of investors to diversify their portfolios internationally"); see also Coeurdacier & Rey, note 10, at 92 ("most studies . . . argue that costs need to be very large to explain portfolio holdings"); Graetz & Grinberg, note 53, at 551 ("transaction costs [do not] have much explanatory power"); Sercu & Vanpée, note 30, at 321 (discussing several papers on turnover). But see Coeurdacier & Rey, note 10, at 94-95 (casting some doubt on the consensus view).

[55] See Coeurdacier & Rey, note 10, at 79 ("[Fluctuations in the nominal exchange rate . . . could be hedged using positions in the forward currency market or the currency bond market."); French & Poterba, note 29, at 223 (assuming investor uses forward contracts to lock in exchange rates); Sercu & Vanpée, note 30, at 318-19 ("[E]xchange-rate instruments . . . do offer perfect hedges against the main component of purchasing power risk").

[56] See Graetz & Grinberg, note 53, at 551 ("Hedging explanations are to little avail."); Sercu & Vanpée, note 30, at 320 ("[T]he literature on hedging [including foreign exchange risk] as an explanation for the home bias in equities does not provide truly convincing results. That is, hedging does not seem to be a major factor behind international underdiversification."). Coeurdacier and Rey are similarly skeptical of this explanation as it applies specifically to purchasing power risk ("real exchange rate risk"), although they are more optimistic about this explanation as it specifically applies to labor income risk (more generally, "nontradeable income risk"). Coeurdacier & Rey, note 10, at 79, 81-91.

[57] For example, the "benchmark" two-asset model presented in Coeurdacier and Rey, note 10, at 81-87, and discussed in note 56 is highly stylized. The model has two countries only, which are identical; exogenously built-in home biases in purchases of both consumption goods and capital goods; specially located uncertainty (multiplicative shocks to the full production function and to the marginal disutility of leisure); specially structured uncertainty (independently and identically distributed across shocks and over time); two assets only-bonds and stocks; bonds that pay in goods not currency and are traded only among households with no borrowing by firms themselves; complete financial markets; no exchange rate risk (implicitly, a single world currency); a special form of production function (Cobb-Douglas) that, among other things generates no profits; consideration only of a symmetric steady state; and other limiting features.

Focus, specifically, on the model's incorporation of only two kinds of assets-real bonds (among households) and equity. It is interesting to reflect on the fact that the main point of the model, relative to the literature, is to show that introducing a second asset-real bonds-changes the hedging characteristics of home equity, making it a much better complement to labor income risk. Id. at 82-85. As the authors themselves hint, id. at 89-90, it seems possible that introducing yet a third asset, such as housing stock, might change the picture yet again, perhaps upsetting the authors' results.

[58] See Portes & Rey, note 32, at 271 (concluding from the data that "the diversification motive is dominated by the information effect: [w]e find weak support for a diversification motive ... once we control for informational frictions"); Sercu & Vanpée, note 30, at 320 (survey of the literature); Tesar & Werner, note 54, at 468-69 (concluding from the data that "the composition of the portfolio of foreign assets may reflect factors other than diversification of risk"); see also Coeurdacier & Rey, note 10, at 89-91 (presenting the authors' discussion of the shortcomings of their model, described in notes 48 and 57, including that the model and others like it "cannot replicate realistic moments [that is, means, variances, skews, and the like] of asset prices and exchange rates").

[59] With regard to governance issues, see Sercu & Vanpée, note 30, at 323-24 (providing a favorable review of governance issues explanation in survey of literature). But see generally Coeurdacier & Rey, note 10 (not discussing governance issues). With regard to information asymmetries, see id. at 95-98 (offering more favorable review than in Sercu and Vanpée of information asymmetry explanation in survey of literature); Graetz & Grinberg, note 53, at 551-52 ("Intuitively, the most convincing explanation is grounded in information asymmetries...."); Sercu & Vanpée, note 30, at 321-23 (providing mostly favorable review of information asymmetry explanation in survey of literature). With regard to irrational biases, see id. at 324-26 ("Behavioral biases probably play a role besides [perhaps meaning "alongside"] information asymmetries and governance issues"); see also Coeurdacier & Rey, note 10, at 98 (providing generally favorable discussion of behavioral explanations; Portes & Rey, note 32, at 271 ("[MJarket segmentation appears to be attributable mainly to informational asymmetries or familiarity effects.").

[60] See Coeurdacier & Rey, note 10, at 98 ("[I]t remains difficult to disentangle empirically informational frictions linked to distance and/or institutional differences from behavioral biases such as 'familiarity' and/or 'competence' effects."); Portes & Rey, note 32, at 271 n.7 ("Separating out familiarity effects from pure informational symmetries remains a challenge for the empirical literature."); Sercu & Vanpée, note 30, at 323, 326 (noting strong link between governance issues and information asymmetries; arguing behavioral biases are difficult to distinguish and measure; and observing contention regarding whether the real culprit is information asymmetries).

[61] See note 29 and accompanying text.

[62] Contributions to the home equity bias literature are not always as clear as they might be about where the underlying data come from. For example, a recent survey published in the reputable Journal of Economic Literature-perhaps the world's premier journal for papers surveying economic research-lists as a central data source the "Consumer Portfolio Investment Survey," by which apparently is meant the International Monetary Fund's (IMF) "Coordinated Portfolio Investment Survey," commonly known by its acronym "CPIS." See Coeurdacier & Rey, note 10, at 67 tbl.l. This guess is difficult to confirm: Among the almost 200 entries in its bibliography, Coeurdacier and Rey's paper contains no reference to either survey. Id. at 110-15. However, a Google search for the first, the "Consumer Portfolio Investment Survey," brings up nothing relevant. Moreover, in the appendix of an earlier draft of the same paper, Coeurdacier and Rey list as a source for the same "Table (1)" simply "IMF" within the table itself and, in an appendix, "Consumer Portfolio Investment Survey (CPIS, IMF)." Nicolas Coeurdacier & Hélène Rey, Home Bias in Open Economy Financial Macroeconomics 5, app. at 61 (Jan. 2012) (unpublished manuscript), available at http://econ.sciences-po.fr/sites/default/files/file/homebias-feb2013- finalJEL_0.pdf. See also Coeurdacier & Rey, note 10, at 99 (referring for another purpose to "CPIS ... data from the IMF"). These associations with the IMF make it likely that the authors mean "Coordinated Portfolio Investment Survey" rather than "Consumer Portfolio Investment Survey."

[63] See note 74 regarding the source of data for various contributions to the home country bias literature.

[64] See note 49 (regarding emerging markets and the measurement of governance quality).

[65] ifo Apple Store: News and Information about Apple Inc.'s Retail Stores, http://www .ifoapplestore.com (last visited Apr. 29,2015) (including Excel spreadsheet of Apple stores that can be sorted by country and date of opening.

Update: These numbers have increased to thirty-eight stores in the United Kingdom and eight in Japan. Apple Store Locations, https//www.apple.com/retail/storelist (last visited Jan. 16, 2015).

[66] Apple Inc., Annual Report (Form 10-K) 8 (Oct. 30,2013), available at https//www.sec .gov/Archives/edgar/data/320193/000119312513416534/d590790dl0k.htm (Foreign and Domestic Operations and Geographic Data).

Update: This number has increased to 62%. Apple Inc., Annual Report (Form 10-K), at 7 (Oct. 27, 2014), available at http://www.sec.gov/Archives/edgar/data320193/000119312514 383437/d783162dlOk.htm (Foreign and Domestic Operations and Geographic Data).

[67] The following is a sampling of news stories from January 2014 only: Hugo Duncan, Google Could Face Multi-Million Pound Bill After Paying Just ￡11.6 Million in Tax in 2012, Daily Mail (Jan. 5, 2014), available at http://www.dailymail.co.uk/news/article-2534193/ Google-face-multi-million-pound-bill-paying-just-ll-6-million-tax-2012.html; Google: Not Liable to Pay Taxes for Indian Operations, Times of India (Jan. 31,2014), available at http:/ /timesofindia.indiatimes.com/tech/tech-news/Google-Not-liable-to-pay-taxes-for-Indianoperations/articleshow/29669269.cms; John Reed, Call by Israeli Legislators to Tax Google More Heavily, Fin. Times (Jan. 26, 2014) (discussing similar attempts in Germany and France); Eric Sylvers, Italy Passes eurolbn Google Tax Law-Then Promptly Delays It for Six Months, ZDNet (Jan. 3, 2014), http://www.zdnet.com/italy-passes-lbn-google-tax-lawthen-promptly-delays-it-for-six-months-7000024759/.

[68] Aviat & Coeurdacier, note 50, at 24 (finding, based on CPIS data, that "a 10% increase in bilateral trade induces a 6% to 7% increase in bilateral financial assets holdings"). Regarding one explanation for this effect, see id. at 23 ("[D]ue to information asymmetries, entrepreneurs may learn about each other by trading goods and this information facilitates trade in financial assets (and vice versa)"). Regarding the possibility of reverse "causality" (that is, asset holdings causing the trade), see id. at 23-24 (finding that there is reverse "causality," but attempting to isolate the effect of trade on asset holdings by using a composite proxy for the former constructed from variables unlikely to directly affect the latter); see also Philip R. Lane & Gian Maria Milesi-Ferretti, International Investment Patterns, 90 Rev. Econ. & Stat. 538, 538-49 (2008) (using CPIS data to demonstrate "the strongly positive association between bilateral trade and portfolio allocations, even controlling for a host of gravity-type variables," such as time zone difference, com- mon language, colonial relationship, currency union, investment tax treaty, and common origin of legal system).

[69] Aviat & Coeurdacier, note 50, at 23 ("[T]he 'distance puzzle' documented by Portes and Rey [note 32] is drastically reduced once we control for trade in goods. We find that the distance effect on asset holdings is at least divided by two.").

[70] The closest the literature comes is the coarser breakdown into closely held versus publicly traded companies. Indeed one influential paper finds that this breakdown by itself explains a substantial portion of the bias. See Magnus Dahlquist, Lee Pinkowitz, Rene M. Stulz & Rohan Williamson, Corporate Governance and the Home Bias, 38 J. Fin. & Quantitative Analysis 87, 97 (2003) ("[Tjaking into account that some shares are not available for stock market trading reduces the extent of the home bias significantly for the U.S. and dramatically for most other countries."). For example, the authors report that one measure of home bias for U.S. investors decreases by 38% when closely held shares are taken out of the calculation although there is insufficient data "to compute the home bias in foreign countries". Id. at 100. The authors cite as a data source Treasury's "Report on U.S. Holdings of Foreign Long-term Investments" for 2000, which is in turn based on TIC data, as described in Part IV. See id. at 93-94.

[71] The International Investment and Trade in Services Act of 1976 is the main statutory authority for surveys of, among other things: (1) foreign holdings of U.S. securities, often referred to as liabilities surveys; (2) U.S. holdings of foreign securities, often referred to as asset surveys or claims surveys; and (3) cross-border securities purchases and sales, as opposed to holdings. See Pub. L. No. 94-472,90 Stat. 2059 (codified as amended at 22 U.S.C. §§ 3101-08 (2013)).

There is also older and apparently separate statutory authority for this kind of data collection deriving from the President's power to respond to requests from the IMF. See Bretton Woods Agreements Act of 1945, ch. 339, § 8, 59 Stat. 515 (codified as amended at 22 U.S.C. § 286f (2013)) (authorizing the President to collect information requested by the IMF); Articles of Agreement of the International Monetary Fund, Dec. 27,1945, art. VIII, § 5(vii), 60 Stat. 1412 [hereinafter IMF Agreement] (authorizing the IMF to require national data on international investment position).

Accordingly, TIC data on U.S. holdings for foreign securities (the U.S. asset survey) is the U.S. contribution to the Coordinated Portfolio Investment Surveys (CPIS) conducted under the auspices of the IMF. See Treasury Dep't, Frequently Asked Questions Regarding the TIC System and TIC Data, http://www.treasury.gov/resource-center/data-chartcenter/tic/Pages/ticfaq2.aspx (last visited Mar. 6, 2015) (Question 5); see also Int'l Monetary Fund, Coordinated Portfolio Investment Survey, http://cpis.imf.org (last visited Mar. 6, 2015).

The asset surveys of other countries participating in CPIS are another potential source of data on foreign investment in the United States. Unfortunately, as of this writing (March 2014), several key nations do not participate in CPIS, including the People's Republic of China, Taiwan, and most Middle Eastern oil exporters. See Carol C. Bertaut, William L. Griever & Ralph W. Tryon, Understanding U.S. Cross-Border Securities Data, 92 Fed. Res. Bull. A59, A72 & n.22 (2006), available at http://www.federalreserve.gov/pubs/bulletin/2006/crossbordersecurities/default.htm; see also Int't Monetary Fund, Economies Reporting CPIS Data, http://cpis.imf.org/docs/misc/CPIS_Data.pdf (last visited Mar. 6, 2015) (updated list of countries participating in CPIS).

With specific regard to U.S. inbound and outbound foreign direct investment, see Foreign Direct Investment and International Financial Data Improvements Act of 1990, Pub. L. 101-533, §§ 1-4,6-10,104 Stat. 2351 (codified as amended in 22 U.S.C. §§ 3101-18, 314146 (2013)) (authorizing additional information sharing across agencies and by-industry reports).

For the history of foreign investment in the United States, inclusive of the manner in which data has been collected in the past, see generally Mira Wilkins, The History of Foreign Investment in the United States to 1914 (1989); Mira Wilkins, The History of Foreign Investment in the United States, 1914-1945 (2004).

[72] Those who acquire 5% or more of a security traded on a national securities exchange must report their ownership share to the Securities and Exchange Commission (SEC). See Securities Exchange Act of 1934, § 13(d), 15 U.S.C. § 78m(d) (2013). Yahoo! Finance (but not NASDAQ.com) then compiles these shareholder reports by issuer and posts them at finance.yahoo.com, where one can find that the publicly traded shares of companies like Apple, Google, and Cisco tend not to have 5% owners, let alone 10% owners. See, e.g., Apple Inc. Major Holders, http://finance.yahoo.com/q/mh?s=AAPL+Major+Holders (last visited Feb. 6, 2014) (showing in first row of first table that "% of Shares Held by All Insider and 5% Owners" is zero). It can also be confirmed that the shareholders' equity of many large multinationals-such as Apple, Microsoft and Cisco-consists solely of pub- licly traded shares reportable by virtue of § 13(d). Apple Inc., Annual Report (Form 10K), cover page, 47 (Oct. 30, 2013), available at https//www.sec.gov/Archives/ edgar/data/320193/000119312513416534/d590790dl0k.htm; Microsoft Corporation, Annual Report (Form 10-K) cover page, 51 (July 30, 2013), available at https//www.sec.gov/ Archives/edgar/data/789019/000119312513310206/d527745dl0k.htm; Cisco Systems, Inc., Annual Report (Form 10K) cover page, 74 (Sept. 10, 2013), available at https//www.sec ,gov/Archives/edgar/data/858877/000085887713000049/csco-2013727xl0k.htm.

Google Inc. is an exception to the last statement because, as detailed in note 128,16.7% (updated to 7.8%) of Google's stockholders' equity consists of "Class B" shares, which are not listed on a national securities exchange. However, per the same note, more than 98% of such Class B shares are held by three U.S. residents. Thus Google is not in fact an exception to the statement in the main text that this note is meant to support: Namely that TIC's portfolio investment surveys are more relevant than direct investment surveys with regard to foreign ownership.

[73] See note 71 (discussing the IMF's CPIS).

[74] See note 32 (TIC data probably a source for French & Poterba, note 29); note 38 (CPIS inclusive of TIC a source for Sercu & Vanpée, note 30, and Coeurdacier & Rey, note 10, at 64 (but see note 62)); note 50 (TIC "virtually the same as" and perhaps ultimate source for the source used by Portes & Rey, note 32); note 54 (CPIS inclusive of TIC a source for Tesar & Werner, note 54); note 68 (CPIS inclusive of TIC a source for Aviat & Coeurdacier, note 50, and Lane & Milesi-Ferretti, note 68); note 70 (TIC data a source for Dahlquist et al., note 70).

[75] The figure is compiled from tables appearing in periodic reports of holding data (as opposed to transactions data) issued by the TIC system. The most current report as of this writing (March 2014) is U.S. Treasury Dep't, Fed. Reserve Bank of N.Y. Bd. of Gov. of Fed. Reserve Sys., Foreign Portfolio Holdings of U.S. Securities as of June 30,2012,5 tbl.2 row 3 (2014) [hereinafter 2012 SHLA Report] (showing percentage for 2006-2012). Reports going back to 1997 and data going back to 1974 are available at Treasury Dep't, Annual Cross-U.S. Border Portfolio Holdings, http://www.treasury.gov/resource-center/ data-chart-center/tic/Pages/fpis.aspx (last visited Mar. 6, 2015).

Update: Since this Article was accepted for publication (March 2014) a new report has been published. U.S. Treasury Dep't, Fed. Reserve Bank of N.Y., Foreign Portfolio Holdings of U.S. Securities as of June 30, 2013, 5 tbl.2 row 3 (2014) [hereinafter 2013 SHLA Report], This report revises the figure for 2012 up to 13.7%. It also adds a new percentage for 2013, which also happens to be 13.7%.

[76] See U.S. Treasury Dep't, Fed. Reserve Bank of N.Y., Foreign-Residents' Holdings of U.S. Securities, Including Selected Money Market Instruments (SHLA) as of the Last Day of June, at 4, 35-36 (2010), available at http://www.treasury.gov/ticdata/Publish/shlallin(628-ll).pdf [hereinafter 2013 SHLA Form and Instructions] (explaining who must report under what circumstances and providing forms and instructions applicable to reporting in 2013).

Two notes about this source: First, SEC regulations under the Survey Act are very general. See 31 C.F.R. §§ 128.1-.5,128.11-.13,128.21-.23 (Part 128, Reporting of International Capital and Foreign-Currency Transactions and Positions); 31 C.F.R. §§ 129.9-.6 (Part 129, Portfolio Investment Survey Reporting). Detailed information about survey design must be gleaned from the forms and instructions, such as SHLA 2013 Form and Instructions, supra, and the reports of the data collected, such as 2012 SHLA Report, note 75.

Second, as explained within, benchmark surveys with a larger sample are conducted every five years, last in 2009 (as of this writing). The benchmark survey instrument for 2009 is very similar to the most recent annual survey instrument cited above in this note, and I cite only to the latter.

Update: Since this Article was accepted for publication (March 2014) a new benchmark survey has been produced. U.S. Treasury Dep't, Fed. Reserve Bank of N.Y. Bd. of Gov. of Fed. Reserve Sys., Foreign Portfolio Holdings of U.S. Securities as of June 30, 2014 (Apr. 30, 2015), available at http://www.treasury.gov/ticdata/Publish/shl2014r.pdf [hereinafter 2014 Benchmark SHL Report]. The survey instrument for this new benchmark survey is Treasury Dep't, Fed. Reserve Bank, of N.Y., Foreign-Residents' Holdings of U.S. Securities, Including Selected Money Market Instruments (SHL) as of the Last Day of June 2014 (2014), available at http://www.treasury.gov/ticdata/Publish/shl2014in(4-28-2014).pdf [hereinafter 2014 SHL Form and Instructions].

[77] See 2013 SHLA Form and Instructions, note 76, at 4, 35.

[78] See id. at 4, 36.

[79] This discussion, which merely asserts that TIC is unlikely to be under-inclusive, begs the following question regarding the possibility of TIC's over-inclusiveness: With regard to the other end of the chain-the owner end as opposed to the issuer end-what if a foreign custodian holds for a United States owner? The potential mischaracterization of a domestic investor as a foreign investor would seem to render the TIC's sample design too "institutionally comprehensive." This is a manifestation of "custodial bias," a problem that is already well acknowledged (at least in some contexts) and which I discuss in detail in notes 95 and 96.

[80] See Erika Brandner, Fang Cai & Ruth Judson, Improving the Measurement of CrossBorder Securities Holdings: The Treasury International Capital SLT, Fed. Res. Bull., May 2012, at 6, available at http://www.federalreserve.gov/pubs/bulletin/2012/pdf/Cross-BorderSecurities_FINAL.pdf.

[81] See id. at 3, 6-7.

[82] Survey Act, 22 U.S.C. § 3105(a)-(c) (setting out criminal and civil penalties for failing to submit or comply); 31 C.F.R. §§ 128.4, 129.6 (2013) (same).

[83] 31 C.F.R. § 129.3 (2013) (portfolio survey responses under oath); see 15 C.F.R. § 801.3 (direct investment reports made under oath).

[84] 18 U.S.C. § 1621 (2013) (relevant perjury statute). Obstruction of justice statutes might also be invoked. See generally Chris William Sanchirico, Evidence Tampering. 53 Duke L.J. 1215, 1249-57 (2004) (discussing federal criminal statutes regarding obstruction and perjury). The base offense level for obstruction and perjury under the Federal Sentencing Guidelines is fourteen, which roughly means imprisonment from fifteen to twentyone months for a first-time offender and perhaps also fines ranging from $4000 to $40,000 (for individual defendants). U.S. Sentencing Guidelines Manual § 2J1.2-.3, ch. 5, pt. A, § 5E1.2 (2014).

[85] Update: June 2014 was the date of the next survey at the time this Article was drafted. The date of the next benchmark survey is now June 2019. See note 76.

[86] 2013 SHLA Form and Instructions, note 76, at 1. There is apparently no publicly available list of benchmark survey respondents.

[87] See id. at 4, 37.

[88] Recently a monthly survey of holdings was instituted (on top of a pre-existing monthly survey of transactions). See generally Brandner et al., note 80, at 1, 12-13 (describing the new monthly holding survey). However, the monthly surveys collect much less detailed information. In particular, the monthly surveys ask only for total U.S. equity owned by foreigners with no breakdown by issuer or even issuer type. See U.S. Treasury Dep't, Fed. Reserve Bank of N.Y. Bd. of Gov. of Fed. Reserve Sys., Aggregate Holdings of Long-Term Securities by U.S. and Foreign Residents (TIC SLT) (2011) [hereinafter 2013 Form SLT] (form in use in 2013).

[89] 2013 SHLA Form and Instructions, note 76, at 4. But see 2012 SHLA Report, note 75, at 1 ("90 percent"). In any event, this apparently results in a "substantial reduction in the number of reporters" (emphasis added). 2013 SHLA Form and Instructions, note 76, at 1.

[90] 2012 SHLA Report, note 75, at 1.

[92] Bertaut et al., note 71, at A62 (reporting 2.8 million records for, presumably, the mid 2000's).

[93] See CUSIP Global Services, About CGS Identifiers, http://www.cusip.com/cusip/ about-cgs-identifiers.htm (last visited Mar. 6, 2015).

[94] 2013 SHLA Form and Instructions, note 76, sched. 2, items 5, 6, 14-15, 18.

[95] This note reviews acknowledged problems with the TIC-aside from the "foreign denominator problem" discussed in the text-as such problems appear in staff reports and as they pertain to determining the foreign ownership of U.S. equity. The next note considers the academic treatment of these problems.

The relevant sources are (in chronological order): William L. Griever, Gary A. Lee & Francis E. Wamock, The U.S. System for Measuring Cross-Border Investment in Securities: A Primer with a Discussion of Recent Developments, 87 Fed. Res. Bull. 633, 637-38, 641 (2001 ) (describing "custodial center bias," problem with bearer securities, and problem with offshore funds), available at www.federalreserve.gov/pubs/bulletin/2001/10011ead.pdf; Carol C. Bertaut & William L. Griever, Recent Developments in Cross-Border Investment in Securities, 90 Fed. Res. Bull. 19, 20 (2004) (discussing custodial center bias, but not by that name, and problem with bearer securities), available at http://www.federalreserve.gov/ pubs/bulletin/2004/winter04_if.pdf; Bertaut et al., note 71, at A63-A66 (similar to Griever et al., supra, but using the term "custodial bias"); Brandner et al., note 80, at 4-6 (similar).

"Custodial bias" refers to the fact that it is often only the country of residence of the foreign custodian that is recorded, not the country of residence of the owner. This problem affects the utility of TIC data for determining foreign ownership share-even for all U.S. corporations as a group-because U.S. investors might hold their shares with foreign custodians. A clear example appears at Bertaut et al., note 71, at A63.

That said, if a U.S. investor holds shares through a foreign custodian this may show up on the converse U.S. assets survey as a U.S. investment abroad. Thus there is at least some chance of netting the problem out across surveys, at least if one is interested in the foreign ownership share of all U.S. equity taken together. On the other hand, such cross-netting is regarded as difficult. Brandner et al., note 80, at 5. Even more problematic-though certainly not more explicit-is the fact that the U.S. owner will not report his or her ownership in the U.S. asset survey if he or she is a natural person. U.S. owners report only if they are "end-investors"; their custodians report only if they are U.S. persons. U.S. Treas, Dep't, Report of U.S. Ownership of Foreign Securities, Including Selected Money Market Instruments (SHCA) 3 (2013) [hereinafter 2013 SHCA Report]. Though it is not entirely obvious from the survey form and instructions, one can confirm that a natural person is not an "end-investor." U.S. Treasury Dep't, Frequently Asked Questions for Aggregate Holdings of Long-Term Securities by U.S. and Foreign Residents (TIC SLT) 3 (2011) (from the answer to question 3: "Natural persons do not file TIC reports."), available at http://www .treasury.gov/ticdata/Publish/slt_faqs.pdf. It should also be noted that other reporting regimes-most notably, the Foreign Account Tax Compliance Act (FATCA) and the Report of Foreign Bank and Financial Accounts (FBAR) form-provide some scope for looking through foreign custodians to U.S. owners. See Hiring Incentives to Restore Employment Act of 2010, Pub. L. No. 111-147,124 Stat. 71, 97-115 (codified at 26 U.S.C. §§ 1471-1474 and in scattered sections of 26 U.S.C.); Bank Secrecy Act, Pub. L. No. 91-508, 84 Stat. 1114-24 (1970) (codified as amended at 12 U.S.C. §§ 1829b, 1951-1959; TD F 90-22.1, Report of Foreign Bank and Financial Accounts (Mar. 2011), available at http://www.fincen ,gov/forms/files/f9022-l_fbar.pdf. However, cross-netting across reporting regimes is likely to be even more difficult than cross-netting across TIC surveys.

Second, there is the similar problem that holdings of foreign-resident investment funds- in particular, those located in Caribbean financial centers-are counted as investment from the particular countries in which the funds are organized, regardless of the residence of the fund investors. Brandner et al., note 80, at 5 ("Caribbean banking centers are also a significant destination of U.S. investment overseas."). This specifically affects the determination of the foreign-ownership share of U.S. equity to the extent that U.S. rather than foreign investors are investing through these offshore funds.

Again, some cross-netting with the converse U.S. asset survey and other reporting regimes may be possible. Indeed, cross-netting with the TIC asset survey is likely to be more feasible in this second case. Plausibly, many such offshore funds are feeder funds invested in by U.S. university endowments and U.S. pension funds (who wish to avoid generating unrelated business taxable income from their participation in highly leveraged private equity and hedge fund investments). Such institutional investors would indeed be "end-investors" for purposes of the asset survey. 2013 SHCA Report, supra, at 3.

Lastly, bearer securities are reported by the issuer, which may have no information on the foreign investor's country of residence. See Part VIII. But this is not directly a problem if the object is to determine the foreign-ownership share of registered securities such as the equity of large U.S. multinationals.

To what extent do these problems-especially the first two-hinder an accurate determination of foreign ownership share of U.S. equity? Some staff reports express the view that they are mostly an issue for the U.S. assets survey as opposed to the U.S. liabilities survey. The U.S. liabilities survey is thought to be relatively problem free. See Bertaut et al., note 71, at A67 ("In general, the data on U.S. liabilities are considered to be reasonably comprehensive, as debt instruments [inclusive of equity] tend to be issued by and bought or sold through large institutions that can be fairly readily identified and included in the data reporting network."). Notwithstanding the regular participation of large institutions, it is not clear why "custodial bias" does not pose a substantial problem for the liabilities survey. In any event, it should be noted that reassurances with regard to the liabilities survey concern determining the absolute dollar amount of foreign investment in all U.S. assets, not the foreign ownership share of large U.S. multinationals, not even the foreign ownership share of all U.S. equity.

[96] This note reviews acknowledged problems with the TIC-aside from the "foreign denominator problem," which is discussed in the text, as such problems appear in academic writing and as they pertain to determining the foreign ownership of U.S. equity.

Relevant sources include Graetz & Grinberg, note 53, at 542 n.23 and accompanying text, and Gabriel Zucman, The Missing Wealth of Nations: Are Europe and the U.S. Net Debtors or Net Creditors?, 128 Q.J. Econ. 1321, 1322 (2013).

Zucman is focused on the possibility that data problems lead to the false impression that developed nations are more in "debt" than they actually are. In his core example, a French household holds its U.S. equities with Swiss custodians. Id. at 1327-28. Such holdings would not show on France's assets survey because, presumably: (1) France's asset survey, like its U.S. counterpart (see the fourth paragraph of note 95), does not require reporting by natural persons; and (2) there is no French custodian. Furthermore, the holdings also would not appear in Switzerland's assets (or liabilities) surveys-because no Swiss asset (or liability) is involved. And yet, the holdings would be counted in the U.S. liabilities survey-because a U.S. issuer is involved (see the discussion in the text accompanying notes 76-78). Thus, a developed-country liability is reported, but the countervailing developedcountry asset is not.

Consistent with this core example-and consistent with staff reports (see the last paragraph in note 95)-Zucman regards the liabilities surveys as relatively comprehensive. Id. at 1328 ("[C]entralization makes it relatively easy to estimate the amount of U.S. equities and bonds held by foreigners. (The country allocation of liabilities, however, is distorted: U.S. securities held by French savers through Switzerland are wrongly attributed to Switzerland, because seeing through the Swiss banks is not possible.)"). To reiterate from the last paragraph in note 95, it is unclear why custodial bias could not also cause U.S. ownership to be miscounted as foreign in the U.S. liabilities survey, if a foreign custodian holds for a U.S. owner.

[97] See note 74 regarding data sources for the home equity bias literature.

[98] French & Poterba, note 29, at 222 n.l; see Section III.B.

[99] See, e.g., 2012 SHLA Report, note 75, at 4. Across annual reports, this text typically appears in the discussion leading up to Table 2 directly under the heading "Shares of U.S. Long Term Securities that are Foreign Owned" at the beginning of the report. These Table 2's are the source for Figure 1 in this Article.

[100] Id.

[101] 2012 SHLA Report, note 75, at 5 (tbl.2 n.l).

[102] For example, total U.S. equity is defined to include "Corporate Equities, Issues at Market Value." See Bd. of Gov. of the Fed. Reserve Sys., Z.l Financial Accounts of the United States: Flow of Funds, Balance Sheets, and Integrated Macroeconomic Accounts 103 (2014), available at http://www.federalreserve.gov/releases/zl/current/zl.pdf (data for L.213 Corporate Equities). This series is defined by adding in, among other things, "[njonfinancial corporate business; corporate equities; liability," which is in turn determined using "data from the following commercial sources: cash mergers and acquisitions data from Thomson Financial Services SDC database; public issuance and share repurchase data from Standard and Poors Compustat database; and private equity issuance data from Dow Jones Private Equity Analyst and PriceWaterhouseCoopers Money tree report." Bd. of Gov. of the Fed. Reserve Sys., Financial Accounts Guide: Series Analyzer for FL103164103.Q (Sept. 12, 2014), http://www.federalreserve.gov/apps/fof/ SeriesAnalyzer.aspx?s=FL103164103&t=L.213&suf=Q.

[103] In some cases, revenue is used in place of book value.

[104] "Nonfinancial corporate business; corporate equities; liability," mentioned in note 102, is defined by adding in, among other things, "[njonfinancial corporate business; closely held corporate equities; liability," Bd. of Gov. of the Fed. Reserve Sys., Financial Accounts Guide: Series Analyzer for FL103164115.Q (Mar. 6, 2014), http://www.federalreserve.gov/ apps/fof/SeriesAnalyzer.aspx?s=FLl03164115&t= [hereinafter Series Analyzer], which is defined as noted in the text. See Bd. of Gov. of the Fed. Reserve Sys., Financial Accounts Guide: Series Analyzer for FL103164123.Q (Mar. 6, 2014), http://www.federalreserve.gov/ apps/fof/SeriesAnalyzer.aspx?s=FL103164123&t=.

[105] "Nonfinancial corporate business; corporate equities; liability," mentioned in note 102, is defined by adding in, among other things, "[njonfinancial corporate business; public corporate equities less intercompany holdings; liability." This item is, in turn, defined by subtracting out "[njonfinancial corporate business; intercompany holdings; liability," Series Analyzer for FL103164115.Q, note 104, which in turn is defined as described in the text. See Bd. of Gov. of the Fed. Reserve Sys., Financial Accounts Guide: Series Analyzer for FL103164193.Q (Jan. 29, 2009), http://www.federalreserve.gov/apps/fof/SeriesAnalyzer .aspx?s=FL103164193&t=&bc=yxqUlFEkiY6hooYlTC9CkjavRt0IOgkW+WyQCgQsFli JM2Wy9SUtCAbKlbXjkj/Ne22rDzFmlyQvQTeuXgjw/ngb0/UhjEr//dQ7/j3eaxI9GZAu5 0tyn7bUH51B03 JEd+OPRXNMIWQyky9awWtXLg==:FL103164115&suf=Q.

[106] 2013 SHLA Form and Instructions, note 76, at 69 (defining "funds" as "[pjooled, separate and general investment accounts, including mutual funds (open and closed end), country funds, exchange traded funds, unit investment trusts, collective-investment trusts, hedge funds, and all other similarly pooled, commingled investment funds. Foreign ownership of shares of U.S.-resident funds are reported as security type = 3 (fund shares), regardless of the types of securities held by the fund.") (emphasis added). For a discussion of the nature and importance of "separate accounts," see Subsection VI.B.l.

[107] 2013 SHLA Form and Instructions, note 76, at 11 ("Report foreign-residents' ownership of shares/units of funds and investment trusts legally established in the United States (U.S.-resident funds) as equity securities.").

[108] See id. ("All foreign-residents' ownership of U.S.-resident fund shares should be assigned security type = 3 (fund shares), and not categorized as a debt security, regardless of the types of securities held by the fund."); id. at 16 ("Foreign-resident limited partner own- interests in U.S.-resident limited partnerships should be reported as security type = 4 (other equity).").

[109] See id. at 12 ("The determination of whether a fund's shares are a U.S. security is based on the country in which the fund is legally established, not based on the residence of the issuers of the securities the fund purchases.").

[110] Id. at 8. ("What Securities Must Be Reported? This report collects information on securities issued by U.S.-residents that are owned by foreign residents.").

[111] Id. at 69 (defining "foreign-resident (foreign, foreigner)" as "any individual, corporation, or other entity legally established outside of the United States, regardless of the actual center of economic activity of the entity____Foreigners/foreign residents include:... any corporation or other entity legally established outside of the United States. . .").

[112] 2012 SHLA Report, note 75, app. tbl.All at A-54-55.

[113] Id. at tbl.All, col. "Equity."

[114] E-mail from Kurt Schuler, Office of Int'l Affairs, U.S. Treasury Dep't (Nov. 4, 2013, 11:35 EST) (on file with author).

[115] Survey Act, 22 U.S.C. § 3104(c) (2013) (permitting use of information only within U.S. government; disallowing publication or disclosure that would make it possible to identify survey respondent; prohibiting compelled disclosure without consent; see also id. §§ 3142(c), 3143(a), (c)(2), (d) (Comptroller General's review), § 3144 (immunity from process); IMF Agreement, note 71, art. VIII, § 5(b) ("Members shall be under no obligation to furnish information in such detail that the affairs of individuals or corporations are disclosed.") With specific regard to portfolio investment, see 31 C.F.R. § 128.3(a) ("Aggregate data derived from these forms may be published or otherwise publicly disclosed only in a manner which will not reveal the amounts reported by any individual respondent."), § 129.5 (similar to § 3104(c)). With specific regard to direct investment, see 15 C.F.R. § 801.5 (similar to § 3104(c)). See also 2013 SHLA Form and Instructions, note 76, at 2 ("The results of this survey will be made available to the general public at an aggregated level so that neither the U.S. persons or organizations providing information nor individual or organizational ownership of U.S. securities can be identified."); 22 U.S.C. § 3104(e) (penalty for willful unauthorized disclosure or use).

[116] See Google Inc. Institutional Ownership, NASDAQ, http://www.nasdaq.com/symbol/ goog/institutional-holdings (last visited Feb. 7, 2014) (referring to "institutional ownership" and "institutional holders"). Since this Article was accepted, the number of institutional owners has decreased to 1681, and the number of pages to 113. Id. (last visited Mar. 10, 2015).

[117] See Securities Exchange Act, § 13(d), (f), 15 U.S.C. § 78m(d), (f) (2013), as interpreted and implemented by Rule 13í-1; 17 C.F.R. § 240.13í-1 (2013). The program was implemented, after some delay, on June 15, 1978. See Exchange Act Release No. 15,292, 43 Fed. Reg. 52,697 (Nov. 2,1978); S. Rep. No. 94-75, at 78-81 (1975), for an account of the genesis of the provision, including pre-existing reporting requirements.

[118] Pursuant to § 78m(f)(4), the SEC compiles and regularly updates a list of all § 13(f) securities. Sec. & Exchange Comm'n, Official List of Section 13(f) Securities (Jan. 2, 2015), http://www.sec.gov/divisions/investment/13flists.htm. Rule 13f-l(c) permits the IIMs to rely on this list.

[119] Section 78(a)(35) defines "investment discretion" to include the authority to make decisions regarding what is bought and sold for the account as well as the making of such decisions "even though some other person may have responsibility for such investment decisions." Furthermore, "[a]n institutional investment manager shall also be deemed to exercise 'investment discretion' with respect to all accounts over which any person under its control exercises investment discretion." 17 C.F.R. § 240.13f-l(b). The term "control" is defined neither in § 78c nor in the rules thereunder.

[120] FMR LLC, Quarterly Report Filed by Institutional Investment Managers, Holdings (Form 13F) (Nov. 14, 2013), http://www.sec.gov/Archives/edgar/data/315066/000031506613 004720/00/xs/Forml3F_X01/FMRLLC.xml.

Each § 13(f) security-such as Apple Inc., CUSIP 37833100-appears in several rows because the holdings of such securities are broken out according to whether and with whom FMR LLC shares investment discretion. This breakout appears in columns that are not shown in Table 1.

[121] See EDGAR: Search Tools, Sec. Exch. Comm'n (May 30,2012), http://www.sec.gov/ edgar/searchedgar/webusers.htm.

[122] NASDAQ, Google Inc. Institutional Ownership, http://www.nasdaq.com/symbol/ goog/institutional-holdings (last visited Feb. 7, 2014).

[123] See 15 U.S.C. § 78m(f)(6); In re Quattro Global Capital, LLC, Exchange Act Re- lease No. 56252, Investment Advisers Act Release No. 2634, 2-3 (Aug. 15, 2007), http:// www.sec.gov/litigation/admin/2007/34-56252.pdf.

[124] See 15 U.S.C. § 78m(f)(l) (filing requirement exemption); 17 C.F.R. § 240.13f-l(a) (same).

[125] See 15 U.S.C. § 78m(f)(l) (exempting security for "insignificance"). There are two other reasons an IIM may not have to report, but these seem less significant. First, the reporting requirement applies only to IIMs who use "any means or instrumentality of interstate commerce" in the course of their business as an IIM. See id. The definition of "interstate commerce" in § 78c(17), however, includes communication between any foreign state and any U.S. state or territory, and includes the use of any national securities exchange. Thus, it seems likely that most IIMs with accounts that hold equities in large U.S. multinationals would qualify. Second, IIMs may request confidential treatment to avoid disclosing the holdings of a natural person. See § 240.24b-2; Form 13F, SEC, 1-2, https://www.sec.gov/about/forms/forml3f.pdf (last visited Mar. 8, 2015) [hereinafter 13F Form and Instructions],

[126] See notes 117-18.

[127] Since this article was accepted for publication these figures have changed to 66% and 34% respectively. See NASDAQ, note 116.

[128] This note provides background information on Google shares as of the time of this writing (with an update below).

Four classes of Google Inc. stock are authorized: Class A common stock, Class B common stock, Class C capital stock, and convertible preferred stock. Google Inc., Annual Report (Form 10-K) 76 (note 11 to Consolidated Balance Sheet) (Dec. 31, 2013), available at https://www.sec.gOv/Archives/edgar/data/1288776/000128877614000020/goog2013123110k.htm#s46244CABF3D2DB81A4AE2E719E4123AA. Only Class A common stock and Class B common stock are issued and outstanding. Id. The rights of holders of Class A and Class B common stock differ only in regard to voting: one vote per share for Class A and 10 votes per share for Class B. Id. Only Class A is publicly tradable; since August 2004 has been listed on Nasdaq Global Select Market under the symbol GOOG. Id. at 3, 21. But Class B shares are convertible one for one into Class A shares at the option of the holder, and automatically convert upon sale or transfer. Id. at 76. As of January 30, 2014, there were outstanding 279,883,488 shares of Class A and 56,167,343 shares a Class B. Id. at cover page. Thus, ignoring differences in voting rights, Class A shares constitute approximately 83.3% of equity and Class B shares, 16.7%. As of December 31, 2013, there were approximately 2536 "stockholders on record" of Class A (see note 230 regarding "street name" issues) and approximately 74 stockholders of record of Class B. Id. at 22. As of December 31, 2013, major holders of Class B shares were as follows: Google co- founder Lawrence E. Page owned 42.4% of Class B, constituting approximately 7.1% (42.4% of 16.7%) of Google. Google, Inc., Schedule 13G Amendment 7,3 (Dec. 31,2013), available at https://www.sec.gov/Archives/edgar/data/1288776/000119312514048184/ d670242dscl3ga.htm. Google co-founder Sergey Brin owned 41.6%, constituting approximately 6.9% of Google. Google, Inc., Schedule 13G Amendment 8, 3 (Dec. 31, 2013), available at http://www.sec.gOv/Archives/edgar/data/l 288776/000119312514048198/ d672690dscl3ga.htm. Google's executive chairman Eric E. Schmidt and the Schmidt Family Living Trust owned 14.1% of Class B shares, constituting approximately 2.4% of Google equity. Google, Inc., Schedule 13G Amendment 9, 5 (Dec. 31, 2013), available at https://www.sec.gov/Archives/edgar/data/1242463/000119312514048502/d677691dscl3ga .htm. Thus these three owners held 98.1% of Class B stock, constituting approximately 16.4% of Google shares.

Update: In April 2014 in a "two-for-one stock split effected in the form of a stock dividend," Google distributed Class C stock to holders of Class A and Class B stock, one share of Class C per one share of either Class A or Class B. Google Inc., Annual Report (Form 10-K) 48 (Dec. 31, 2014), available at http://www.sec.gov/Archives/edgar/data/1288776/ 000128877615000008/goog2014123110-k.htm [hereinafter Google 2014 Annual Report]. For a clearer explanation of the mechanics of the split, see Steven Russolillo, What Google's Stock Split Means for You, Wall St. J. (Apr. 3, 2014), http://blogs.wsj.com/ moneybeat/2014/04/03/what-googles-stock-split-means-for-you/. Class C shares are non- voting, and are listed on Nasdaq Global Select Market under the symbol GOOG, formerly the symbol for Class A shares, which are now identified as GOGGL. Id. at 17. As of January 29, 2015, there were outstanding 286,938,352 shares of Class A, 53,018,898 shares of Class B, and 340,665,532 shares of Class C. Id. at cover page. Thus, ignoring voting differences, Class A shares constitute 42.2% of equity, Class B 7.8%, and Class C 50.1%. As of December 31, 2014, the ownership shares of co-founders Lawrence E. Page and Sergey Brin had fallen to approximately 13.1% (from 14% = 7.1% + 6.9% a year earlier). Moreover, they had both adopted "stock trading plans" "in order to allow [them] to sell a portion of their Google stock over time as part of their long-term strategies for individual asset diversification and liquidity." Google Inc., Current Report (Form 8K) (Feb. 13, 2015), available at http://www.sec.gov/Archives/edgar/data/1288776/000128877615000014/ form8_kforfoundersl0b5xlpl.htm (indicating that these individuals together own 13.1% of Class A and B combined and 13.1% of Class C).

[129] See note 72.

[130] Since this Article was accepted for publication this figure has changed to 25%. See Cisco Systems, Inc. Institutional Ownership, NASDAQ, http://www.nasdaq.com/symbol/ csco/institutional-holdings (last visited Mar. 8, 2015).

[131] See, Apple Inc. Institutional Ownership, NASDAQ, http://www.nasdaq.com/symbol/ aapl/institutional-holdings (last visited Mar. 8, 2015).

[132] See 15 U.S.C. § 78c(a)(17).

[133] See id.; Filing and Reporting Requirements Relating to Institutional Investment Managers, Exchange Act Release No. 14,852, 43 Fed. Reg. 26,700, 26,700-05 (June 19, 1978).

[134] See SEC, Office of Audits, Review of the SEC's Section 13(f) Reporting Requirements vi (2010) [hereinafter Office of Audits Report].

[N]o SEC division or office conducts any regular or systematic review of the data filed on Form 13F. . . . [N]o SEC division or office has been delegated authority to review and analyze the 13F reports, and no division or office considers this task as falling under its official responsibility.... [N]o SEC division or office monitors the Form 13F filings for accuracy and completeness. As a result, many Forms 13F are filed with errors or problems, which may not be detected or corrected in a timely manner. . . . [N]o checks are built into the EDGAR system, through which the Forms 13F are filed, to scan for obvious errors in the forms. . . . The OIG's testing of a sample of CTRs [confidential treatment requests] processed by IM [the SEC's Division of Investment Management] revealed that files and supporting documentation could not be located for approximately one-half of the CTRs selected in our initial sample of 25 items. When we selected an additional 12 CTRs, files could not be located for two-thirds of the additional 12 items.

Id. Some of these problems may have been rectified since this report was filed.

Similar concerns are expressed in James D. Cox & Randall S. Thomas, Letting Billions Slip Through Your Fingers: Empirical Evidence and Legal Implications of the Failure of Financial Institutions to Participate in Securities Class Action Settlements, 58 Stan. L. Rev. 411, 446-48 (2005) ("Our review of the LexisNexis database reveals no SEC enforcement action for noncompliance with Section 13(f) in the legislation's thirty-six-year history. Either there is remarkable compliance with the provision or a similar level of inattention to this provision by the SEC."). But see, e.g., In re Quattro, note 123 (SEC enforcement action following publication of Cox & Thomas).

[135] See S. Rep. No. 94-75, at 82-85 (1975) (stating that purpose was "to create a central depository of historical and current data about the investment activities of institutional investment managers" to assist investors and government regulators).

[136] See id at 77-82.

[137] See id. at 83-84; In re Quattro, note 123, at 2.

[138] See Norway Government Pension Fund Global, Sovereign Wealth Fund Inst., http:// www.swfinstitute.org/swfs/norway-government-pension-fund-global (last visited Mar. 8, 2015).

[139] See Sovereign Wealth Fund Inst., Fund Rankings http://www.swfinstitute.org/fundrankings (last visited Feb. 7, 2014).

Update: By the beginning of 2015 this number had increased to $893 billion. Id. (last visited Mar. 8, 2015).

[140] Norges Bank, Quarterly Report Filed by Institutional Managers, Holdings (Form 13F) (Feb. 7, 2014), http://www.sec.gov/Archives/edgar/data/1374170/000137417014000002/ 0001374170-14-00d002-index.htm [hereinafter Norges Bank 4Q14 Form 13F].

Norges Bank files in effect less frequently than most other IIMs. The SEC appears to grant Norges Bank de facto "confidential treatment" of such limited duration for each report as to insure that Norway GPFG's annual reporting of its holdings under Norway law is the most recent accounting of Norway GPFG's holdings available. See id. (discussing confidential treatment); see also Office of Audits Report, note 134, at 24 (discussing de facto confidential treatment by "a particular large foreign institutional investment manager").

For example, several weeks after the end of the third quarter of 2013, Norges Bank provided in its Form 13F no specific holdings data for that end of quarter, citing confidential treatment. Norges Bank, Quarterly Report Filed by Institutional Managers, Holdings (Form 13F) (Nov. 12,2013) [hereinafter Norges Bank 3Q13 Form 13F], http://www.sec.gov/ Archives/edgar/data/1374170/000137417013000032/0001374170-13-000032-index.htm (including very limited disclosures and the following note: "Norges Bank has submitted its Form 13F to the Securities and Exchange Commission pursuant to a request for confidential treatment. This action is consistent with discussions between Norges Bank and the Staff of the Securities and Exchange Commission to coordinate the reporting required under Section 13(f) with the extensive public disclosure requirements under Norwegian law applying to Norges Bank."). Compare this to an amended report filed two days later containing detailed holdings information for the third quarter of the prior year, 2012. Norges Bank, Quarterly Report Filed by Institutional Managers, Holdings (Amended Form 13F) (Nov. 14, 2013), http://www.sec.gov/Archives/edgar/data/1374170/00013741701 3000034/0001374170-13-000034-index.htm [hereinafter Norges Bank 3Q13A Form 13F],

[141] NBIM, Government Pension Fund Global Annual Report 2012, at 19 (2013), available at http://www.nbim.no/globalassets/reports/2012/annual-report-2012.pdf ("NBIM awards investment mandates to external equity managers with expertise in specific markets. . . . The fund had 146 billion kroner, or 3.8 percent of its capital, under external management at the end of 2012.").

[142] See id. (stating that all fifty-one external managers were hired to manage investing in either emerging markets like Mexico, small capitalization companies in established markets, or "environmentally-related investments").

[143] See Norges Bank 3Q13A Form 13F, note 140. As explained in that note, because of the confidential treatment, Norges Bank 4Q13 Form 13F, note 140, reveals only total investment in § 13(f) securities, which was $164 billion as of Dec. 31, 2013.

Update: Since this Article was accepted for publication (March 2014), the detailed holdings data for the third quarter of 2013 have become available. Norges Bank, Quarterly Report Filed by Institutional managers, Holdings (Amended Form 13F) (Nov. 14, 2014), http://www.sec.gov/Archives/edgar/data/1374170/000137417014000044/0001374170-14000044-index.htm [hereinafter Norges Bank 3Q14A Form 13F[. The updated number for the number cited in the text is $171 billion.

[144] Norges Bank 3Q13A Form 13F, note 140.

Update: The latest disclosure shows that these numbers for 2013 are 2,030,939 and $1.8 billion, respectively. Norges Bank 3Q14A Form 13F, note 143.

[145] NBIM, Government Pension Fund Global Quarterly Report 3Q 2013, at 2 (2013) [hereinafter NBIM 3Q Report 2013].

Update: Since this Article was accepted for publication (March 2014) a new report has been issued, which shows that Norway GPFG invests roughly 61 % of its nearly $900 billion total in "equities." NBIM, Government Pension Fund Global Quarterly Report 3Q 2014, at 2 (2014) [hereinafter NBIM 3Q Report 2014],

[146] NBIM 3Q Report 2013, note 145, at 24.

Update: Since this Article was accepted for publication (March 2014) a new report has been issued, which shows that Norway GPFG makes 31% of equity investments (thus 19% of the $900 billion) in U.S. equities in particular. NBIM 3Q Report 2014, note 145 at 29.

[147] See NBIM, Annual Report 2012, http://www.nbim.no/en/press-and-publications/ Reports/2012/annual-report-2012/ (last visited Feb. 7, 2014). On this web page, under "Web only," follow "holdings," and then search "Google." The dollar figure in the text is the result of converting the figure that thus appears to dollars using Norges Bank's record of the exchange rate as of December 31, 2012. There appear to be no more recent numbers for holding of specific stocks, as of this writing.

Update: Since this Article was accepted for publication (March 2014) new figures have become available. At the end of 2013, Norway GPFG owned about $2.1 billion in Google stock (as opposed to $1.8 billion). See NBIM, Annual Report 2013, at 36 (2013), http:// www.nbim.no/globalassets/reports/2013/annualreport/eq_holdings_spu.pdf. The dollar figure is the result of converting the figure that thus appears to dollars using Norges Bank's record of the exchange rate as of December 31, 2013.

[148] See Sovereign Wealth Fund Inst., note 139; Sovereign Wealth Fund Inst., China Investment Corporation, http://www.swfinstitute.org/swfs/china-investment-corporation/ (last visited Feb. 7, 2014). Figures available after the acceptance date for this Article (March 2014) list the fund at $652 billion. See Sovereign Wealth Fund Inst., note 139; Sovereign Wealth Fund Inst., China Investment Corporation, http://www.swfinstitute.org/swfs/chinainvestment-corporation/ (last visited Mar. 10, 2015).

[149] See External Fund Managers, China Inv. Corp., http://www.china-inv.cn/cicen/investment/investment_external.html (last visited Nov. 17, 2014). (This link was active in February 2014 but has since vanished.)

CIC retains external fund managers to assist with the management of certain aspects of its international investment portfolio. External fund managers are selected either via open procurement processes or on a special project basis. External fund managers are expected to demonstrate: Strong financial background and credit record with risk control aligned to local judicial and regulatory requirements; An appropriate level of experience and performance related to assets under management (AUM); Investment staff that hold relevant professional qualifications; Established corporate governance and internal controls to ensure regulated operation; At least three-years without any regulatory issues or complaints. The appointment of a fund manager is through an open, fair, and well-ordered process which incorporates a phased approach, including: application, preliminary review, second review, fees & contract negotiation, and result announcement. CIC selects managers based on the application documents, the interview, the experience of the investment teams and track record in the specific asset class, the management fee proposal, and other relevant considerations.

[150] See China Inv. Corp., Annual Report 2012, at 36 (2013), available at http://www .china-inv.cn/cicen/include/resources/CIC_2012_annualreport_en.pdf [hereinafter CIC Annual Report 2012]. Figures available after the acceptance date for this Article (March 2014) list this number as 67%. See China Inv. Corp., Annual Report 2013, at 36 (2014), available at http://www.china-inv.cn/wps/wcm/connect/5c337cl3-8677-4862-aa364efbclb243ae/CIC_2013_annualreport_en.pdf?MOD=AJPERES&CACHEID=5c337cl38677-4862-aa36-4efbclb243ae [hereinafter CIC Annual Report 2013].

[151] CIC Annual Report 2012, note 150, at 36. Figures available after the acceptance date for this Article (March 2014) list this number as 40%. CIC Annual Report 2013, note 150, at 35.

[152] CIC Annual Report 2012, note 150, at 37. Figures available after the acceptance date for this Article (March 2014) show the same as of December 2013. CIC Annual Report 2013, note 150, at 36.

[153] CIC Annual Report 2012, note 150, at 50. Figures available after the acceptance date for this Article (March 2014) show that this figure had grown to $652 billion by the end of 2013. CIC Annual Report 2013, note 150, at 56.

[154] FMR LLC, Quarterly Report Filed by Institutional Investment Managers, Holdings (Form 13F) (filed Nov. 14, 2013 for period ending Sept. 30, 2013), http://www.sec.gov/ Archives/edgar/data/315066/000031506613004720/0000315066-13-004720-index.htm [hereinafter FMR LLC 3Q13 Form 13F],

Update: Figures available after the acceptance date of this Article (March 2014) list this number at $769 billion. FMR LLC, Quarterly Report Filed by Institutional Investment Managers, Holdings (Form 13F) (Feb. 17, 2015), http://www.sec.gov/Archives/edgar/data/ 315066/000031506613004720/0000315066-15-002484-index.htm [hereinafter FMR LLC 4Q14 Form 13F],

[155] FMR LLC 3Q13 Form 13F, note 154 (reporting 18,395,507 Google shares worth $16,113,684,000). The total number of Google shares for the same date was 276,681,000 + 57,476,000, representing Class A and Class B common respectively. Google Inc., Quarterly Report (Form 10Q) (filed Oct. 24, 2013 for period ending Sept. 30, 2013). See Consolidated Balance Sheets.

Update: Figures available after the acceptance date of this Article (March 2014) put this number at $16 billion constituting 5% FMR LLC 4Q14 Form 13F, note 154 (reporting 17,249,759 Google Class A common stock worth 9,153,757,000 and Class C capital stock worth $10,157,464,000). The total number of Google shares for the same date was 286,938,352 (Class A common), 53,018,898 (Class B common), and 340,665,532 (Class C capital). Google Inc., Quarterly Report (Form 10Q) (Jan. 29, 2015). See note 128 for a description of Google's capital structure.

[156] FIL Ltd., Quarterly Report Filed by Institutional Investment Managers, Holdings (Form 13F) (filed Nov. 14, 2013), http://www.sec.gov/Archives/edgar/data/318989/ 000031898913000005/0000318989-13-000005-index.htm [hereinafter FIL Ltd. 3Q13 Form 13F], See note 128 regarding the total number of Google shares.

Update: These figures are consistent with figures available after the acceptance date of this Article (March 2014). See FIL Ltd., Quarterly Report Filed by Institutional Investment Managers, Holdings (Form 13F) (Feb. 12, 2015), http://www.sec.gov/Archives/edgar/ data/318989/000031898918000033/0000318989-15-000033-index.htm; Google Quarterly Report, note 128 (explaining the total number of Google shares).

[157] See 13F Form and Instructions, note 125, at 4-6. The two entities do show up on each other's Form ADVs, discussed in Section VLB though each there disavows common control.

[158] See Fidelity Worldwide Investment, https://www.fidelityworldwideinvestment.com/ global/default.page (last visited Mar. 8, 2015); Moody's Investor Service, Company Profile: FIL Limited (Bermuda) 1 (Nov. 6, 2012), www.moodys.com.

[159] See Subsection VI.A. 1 for more on this.

[160] See 15 U.S.C. § 78c(a) (35). For the purposes of § 13(f):

A person exercises "investment discretion" with respect to an account if, directly or indirectly, such person (A) is authorized to determine what securities or other property shall be purchased or sold by or for the account, (B) makes decisions as to what securities or other property shall be purchased or sold by or for the account even though some other person may have responsibility for such investment decisions, or (C) otherwise exercises such influence with respect to the purchase and sale of securities or other property by or for the account as the Commission, by rule, determines, in the public interest or for the protection of investors, should be subject to the operation of the provisions of this chapter and the rules and regulations thereunder.

Id.; see also 17 C.F.R. § 240.13f-l(b).

For the purposes of this rule, "investment discretion" has the meaning set forth in section 3(a)(35) of the Act (15 U.S.C. 78c(a)(35)). An institutional investment manager shall also be deemed to exercise "investment discretion" with respect to all accounts over which any person under its control exercises investment discretion. Id.

Would not such a sub-advisory relationship be revealed in one or the other entity's § 13(f) report? Such co-appearance in reports is required when there is shared discretion between IIMs. The term "shared discretion" is defined (despite the common meaning of "shared") to include the case in which one IIM "controls" the other. It follows that FMR LLC and FIL Ltd. would also appear on a single § 13(f) report were they commonly controlled-the report of the common controller. But "control" for purposes of § 13(f) is nowhere defined. In any event, FMR LLC and FIL Ltd. do not regard themselves as commonly controlled. See note 157. The upshot is that the Fidelity complex might well be taking the position that FMR LLC is FIL Ltd.'s unaffiliated sub-adviser with sole discretion, and that this need not be indicated anywhere in the § 13(f) reporting of either entity.

[161] See, e.g., FMR Co., Inc., Form ADV Brochure 4 (2014) available at http://www.advi serinfo.sec.gov/iapd/content/viewform/adv/Sections/iapd_Adv2Brochures.aspx?ORG_PK= 108617&RGLTR_PK=50000&STATE_CD=&FLNG_vPK=017CD36800080179031704000 5389511056C8CC0 (disclosing the possibility of sub-advising in the other direction: "FMRC may, to the extent permitted by its advisory contracts, delegate investment discretion over all or a portion of the portfolio to one or more sub-advisers, including FMRC's affiliates and various subsidiaries and affiliates of FIL Limited ('FIL').").

[162] Confusingly, they show up on each other's Form ADVs and the accompanying brochures, although a connection is there disclaimed. Id. at 13 ("FMRC disclaims that it is a related person of FIL").

[163] See Investment Company Act of 1940,15 U.S.C. §§ 80a-l to 80a-64 (2013).

[164] See Investment Advisers Act of 1940, 15 U.S.C. §§ 80b-l to 80b-21.

[165] The term "mutual fund" is variously defined. Here it is used to mean "registered management investment company."

An "investment company" is essentially a company that issues securities to raise funds that it in turn invests in the securities of other companies. See 15 U.S.C. § 80a-3(a)(l) (providing basic definition of "investment company"), § 80a-2(a) (defining "company," "issuer," and "security). However, certain "private funds" are excluded from the definition of "investment company": notably those that make no public offering of their securities and either have fewer than 100 beneficial owners or are owned by "qualifying purchasers," who, in turn, are essentially wealthy individuals (at least $5 million invested). See id. at § 80a-3(c) (providing relevant exemptions from registration under the Investment Company Act), § 80a-2(a) (defining "qualifying purchaser" and referring to such funds as "private funds").

A "management investment company" is not a management company that sells its services to an investment company, but an investment company with managed investments, as opposed to an investment company with a fixed portfolio. Specifically, it is an investment company that is an "open-end mutual fund" or a "closed-end mutual fund". See 15 U.S.C. § 80a-4, 5 (defining "management [investment] company" and "open-end" and "closedend" companies). The other kinds of investment companies are "unit investment trusts" and "exchange traded funds". For a primer on open-end and closed-end mutual funds, including a discussion of how they differ from each other and from other types of investment companies, see Invest Wisely: An Introduction to Mutual Funds, SEC, http://www .sec.gov/investor/pubs/inwsmf.htm (last visited Nov. 17, 2014).

An investment company-whether or not a management investment company-must register with the SEC in order to engage in "interstate commerce" in the United States. See 15 U.S.C. § 80a-7, -8 (describing prohibited activities for unregistered investment companies and registration). "Interstate commerce" is broadly defined. See id. at § 80a2(a)(18). A management investment company that is so registered is a "registered management investment company".

Registered management investment companies, and specifically open-end mutual funds, account for the vast majority of dollars invested through registered investment companies. See Investment Company Inst., 2013 Investment Company Fact Book: A Review of Trends and Activities in the U.S. Investment Company Industry 8, 9 fig. 1.1 (2013), http:// www.ici.org/pdf/2013_factbook.pdf.

Note that "regulated investment company" (RIC) is a category under U.S. tax law that is largely, though not perfectly, coincident with the category of "registered investment company" under U.S. securities law. See IRC § 851 (defining "RIC"). As described in the second paragraph of note 169, regulated investment companies are effectively pass-through entities for tax purposes.

[166] A RIC must disclose to the SEC its complete portfolio at the end of each quarter of its fiscal year. The SEC then makes this information available online at EDGAR: Mutual Funds, Sec. Exch. Comm'n, http://www.sec.gov/edgar/searchedgar/mutualsearch.html (last visited Mar. 8, 2015).

For even quarters, this disclosure of the company's full portfolio is in answer to Item 6 on Form N-CSR. Sec. Exch. Comm'n, Form N-CSR: Certified Shareholder Report of Registered Management Investment Companies 7, available at http://www.sec.gov/about/ forms/formn-csr.pdf. For odd quarters, the disclosure is in answer to Item 1 on Form N-Q. Sec. Exch. Comm'n, Form N-Q: Quarterly Schedule of Portfolio Holdings of Registered Management Investment Company 3, available at http://www.sec.gov/about/forms/formn-q .pdf. Both forms are combined reporting forms with widely dispersed statutory roots. For Form N-CSR, the relevant authority is § 30(b)(2) of the Investment Company Act of 1940, 15 U.S.C. § 80a-29(b)(2) and Rule 30b2-l(a), 17 C.F.R. § 270.30b2-l(a). For Form N-Q, the relevant authority is § 30(b)(1) of the Investment Company Act of 1940, § 80a-29(b)(l) and Rule 30bl-5,17 C.F.R. § 270.30bl-5.

Registered management investment companies must also periodically submit reports to their shareholders. See § 30(e) of the Investment Company Act of 1940,15 U.S.C. § 80a29(e); Rule 30e-l, 17 C.F.R. § 270.30e-l. But these shareholder reports need only contain a summary report on portfolio holdings, provided reference is made to the availability of the full report on Form N-CSR. See Shareholder Reports and Quarterly Portfolio Disclosure of Registered Management Investment Companies, Securities Act Release No. 338393, Exchange Release No. 49,333, Investment Company Act Release No. 26,372,69 Fed. Reg. 11,244,11,248 (Mar. 9, 2004). In such summary, the company need only indicate its fifty largest holdings, and any holding that constituted more than 1% of the fund's net asset value.

Registered management investment companies must also file Form N-SAR with the SEC on a semi-annual basis. This form contains information on fees earned, brokerage commissions paid, related funds, custodial arrangements, sharing of advisers, and purchases and sales of the investment company's own shares. Form N-SAR is required under § 30(b)(1) of the Investment Company Act of 1940, 15 U.S.C. § 80a-29(b)(l) and Rule 30bl-l, 17 C.F.R. § 270.30bl-l.

[167] For example, the United Kingdom offers a credit to shareholders to partially mitigate entity-level taxation. See Income Tax (Trading and Other Income) Act of 2005, ch.5, § 397 (for most U.K. dividends); Income Tax (Trading and Other Income) Act of 2005, §§ 397AC, as amended by Finance Act of 2008, § 4(1), and Finance Act of 2009 § 40 (for most nonU.K. dividends held by minority shareholders). The U.K. shareholder receives a credit equal to one-ninth of the value of the (ungrossed-up) dividend. This benefit then is partly offset by requiring that the taxpayer increase her dividend income for income tax purposes by the amount of such credit. Income Tax (Trading and Other Income) Act of 2005, ch.5, § 398 (gross up by credit, for dividends from both U.K. and non-U.K. companies). This credit is sometimes explained by presenting it in this equivalent fashion: The dividend is first multiplied by the factor 100/90, income tax is imposed on this scaled-up amount, and a credit of 10% of this scaled-up amount is then provided. The equivalence is as follows:

td[100/90) d] - (l/10)[(100/90)d] = r,[(10/9)rf - (l/9)d = tAd+(l/9)d) - (l/9)d, where A is the dividend, t,i is the tax rate on the dividend, the first and last expressions are the after-credit tax on the dividend under the two equivalent statements of the tax calculation.

Until recently this credit was not available for dividends from non-U.K. companies, which would include dividends from U.S. mutual funds. See id. § 397; Michael J. Graetz & Alvin C. Warren, Jr., Dividend Taxation in Europe: When the ECJ makes Tax Policy, 44 Common Mkt. L. Rev. 1577,1611-12 (2007). The law was changed in response to claims of discrimination from within the European Union. See id. at 1595-99 (discussing decisions of the European Court of Justice concerning inbound dividends).

[168] Since July 2013 such funds have included the U.K.'s own Authorized Contractual Scheme (ACS). See The Collective Investment in Transferable Securities (Contractual Scheme) Regulations 2013, S.I. 1388 (U.K.); HM Treasury, Contractual Schemes for Collective Investment: Summary of Consultation Responses and Government Response 3 (2013), available at https://www.gov.uk/govemment/uploads/system/uploads/attachment_ data/file/202938/consult_contractual_schemes_collective_investment_summary_of_respon ses.pdf. But for some time these quasi-transparent funds have been available to U.K. investors elsewhere in the European Union-two prime examples being the Luxembourg Fonds Commun du Placement and the Irish Common Contractual Fund. See Deloitte LLP, UK Tax Transparent Funds, http://www.ukbudgetarchive.com/ukbudget2013/mea-sures/business/ukbudget2013-business-UK-tax-transparent-funds.cfm (last visited Feb. 11, 2015). In these notes, I compare investing in U.S. equity through a U.S. mutual fund to investing through the UK's new ACS. I assume that the U.K. investor is an individual and is not investing through a "tax wrapper" (a tax-preferred investment vehicle). Tax wrappers are discussed later in this Section.

[169] This note continues the hypothetical introduced in the previous note. It concerns the U.S. and U.K. taxation of U.S. portfolio company dividends. There are several points here.

First, neither the U.S. mutual fund nor the U.K. ACS will pay fund-level tax on the portfolio company dividend.

The U.S. mutual fund (as a RIC under U.S. tax law, IRC § 851), receives a deduction for dividends "paid," IRC §§ 561, 852(a), and such deduction may be taken for "consent dividends," which are "hypothetical" and not actually paid out to fund investors. See IRC §§ 561(a), 565; Reg. § 1.565-1; see also IRC § 852(a) (disallowing the deduction for dividends "paid" unless such dividends exceed 90% of RIC's income), § 4982 (imposing excise tax on certain undistributed income of a RIC).

The U.K. ACS is explicitly a pass-through entity. See HM Treasury, note 168, at 3 ("A co-ownership scheme has no legal personality, so that there is no entity at the level of the fund which can be subject to corporation tax, income tax or capital gains tax.").

Second, whether the U.K. investor invests through a U.S. mutual fund or a U.K. ACS, the portfolio dividend will enter the U.K. investor's U.K. income for tax purposes in the year it is paid by the portfolio company. For the U.K. ACS, this follows from explicit pass- through treatment. For the U.S. mutual fund, the U.K. investor will pay tax on the fund's induced dividend (consent or actual) of the portfolio company's dividend.

Third, in both cases, in calculating U.K. income tax, the U.K. investor will enjoy the dividend credit discussed in note 167.

Fourth, in both cases the U.K. investor will incur U.S. withholding tax on the dividend (which in one case comes from the fund, and in the other, from the portfolio company).

Fifth, in both cases the U.K. investor generally can credit U.S. withholding tax toward U.K. income tax on the dividend. The United Kingdom offers a foreign tax credit on an item-by-item basis, Taxation (International and Other Provisions) Act 2010, c. 8, § 18 (U.K.) (allowing credit for income tax, capital gain tax, and corporation tax), rather than a category-by-category basis as in the United States, see IRC § 904. This foreign tax credit will generally offset U.S. withholding tax because the United States withholding rate, which by treaty is 15%, is less than the effective rate on these particular dividends for most U.K. investors. Convention Between the Government of the United States of America and the Government of the United Kingdom of Great Britain and Northern Ireland for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income and on Capital Gains, U.S.-U.K., art. 10 1 2, July 24, 2001, 2224 U.N.T.S. 247, available at http://www.irs.gov/pub/irs-trty/uk.pdf [hereinafter 2001 U.S.-U.K. Convention]. This point about relative size is true even accounting, as one must, for the special dividend credit discussed in the preceding paragraph. But this assumes that the investor is not in the lowest "tax band" (tax bracket). See HM Revenue & Customs, Helpsheet 263, Calculating Foreign Tax Credit Relief on Income (2013), http://www.hmrc.gov.uk/help-sheets/hs263.pdf (explaining how U.K. foreign tax credit relief offsets U.S. withholding tax on dividends).

Sixth, as a historical note, it is worth mentioning that U.S. withholding tax is not foreign tax creditable under the U.K.'s older form of regulated fund, the Authorized Investment Fund (a genus which includes the well-known Open Ended Investment Company, or "OEIC," pronounced to rhyme with the first syllable of "troika"). The underlying tax opacity of these old-style funds means that it is the fund itself, separate from its unit holders, which is regarded as paying the U.S. withholding tax. The fund alone, therefore, is eligible for foreign tax credit relief. As always, such relief is available only up to the amount of U.K. tax paid on that particular item of income. See Taxation (International and Other Provisions) Act 2010, c. 8, § 36 (U.K.) (limiting credit to U.K. income tax for foreign tax paid on "any particular source" to the increment to U.K. income tax attributable to that "particular source," and providing a formula for the calculation of such increment). But, since 2009, such funds have owed no U.K. tax on the U.S. dividend. See Corporation Tax Act of 2009, pt.9A (added by Finance Act of 2009, § 34) c. 4, § 34 (U.K.); INTM651020-Distribution Exemption Overview, HM Treasury, http://www.hmrc.gov.uk/ manuals/intmanual/intm651020.htm (last visited Mar. 8, 2015) ("CTA09/Part 9A is designed to ensure that the great majority of dividends and other distributions will be exempt."). Therefore, the fund receives no credit for U.S. withholding tax on this dividend, and neither does the fund owner. A similar issue of fund opacity prevented U.K. investors who invested in U.S. equity through AIFs from enjoying the treaty-based exemption from withholding tax for workplace pensions and Self-Invested Personal Pensions (SIPPs). The opaque AIF, not the pension or SIPP, was regarded as the dividend recipient. These issues are presumably what drove U.K. investors interested in U.S. equity to funds in Luxembourg and Ireland. And this emigration is presumably what drove the United Kingdom to institute the new-style ACS.

[170] This note continues the hypothetical introduced in note 167 and analyzed in note 168. This note, along with notes 171 and 172, concerns the taxation of portfolio stock gain, while note 167 concerns the taxation of portfolio dividends. This note, in particular, describes how-aside from the difference discussed in the next two sentences in the main text of the Article and detailed in notes 171 and 172-the tax consequences of portfolio stock gain are the same for investment through a U.S. mutual fund as for investment through a U.K. ACS.

First, there is no fund-level taxation in either case. This is, in each case respectively, for the same reason that there is no fund-level taxation for portfolio company dividends, as discussed in note 167.

Second, there is no U.S. withholding tax in either case. For the pure-pass through U.K. ACS, this is because the United States does not generally impose such tax on gain. (The reasons are overlapping and convoluted. See, e.g., IRC § 871(a) (imposing withholding tax on capital gains only when the taxpayer is a nonresident alien and present in the United States 183 days during the current year); id. § 7701(b)(3) (considering an individual who is present in the United States 183 days during the current year "substantially present" in the United States and thus not a nonresident alien)). For the U.S. mutual fund the situation is still more complex. For reasons described in note 167, the U.S. mutual fund will pay out the portfolio stock gain as its own dividend. Dividends from U.S. corporations are generally subject to U.S. withholding tax. However, U.S. mutual funds are relieved of their obligation to withhold on fund dividends attributable to both long-term and short-term capital gains on portfolio stocks. With regard to long-term capital gains dividends, see IRC § 852(b)(3) (defining "capital gain dividends" in terms of "net capital gain"); IRC § 1222(11) (defining "net capital gain" as the excess if any of net long-term capital gain over net short-term capital gain, each of which are themselves defined in the form of "excess, if any"); Reg. § 1.1441-3(c)(2)(i)(D) (providing an election to not withhold under § 1441 and § 1442 on "capital gains dividends," which election is effected by simply not withholding). With regard to short-term capital gains dividends, see the following temporary, but (so far) perennially renewed statutory provision: §§ 871(k)(2)(A), (C), 881(e)(2) (exempting from tax a regulated investment company's "short-term capital gain dividend," defined so that it may not exceed the fund's excess of net short-term capital gain over net long-term capital loss); see also §§ 1441(c)(12), 1442(a) (corresponding exemption from withholding obligation of mutual fund).

To clarify the second point: U.S. withholding tax consequences do differ along the portfolio gain versus portfolio dividend dimension-see note 169 for a discussion of the fact that U.S. withholding is (effectively, in the case of U.S. mutual funds) imposed on portfolio dividends-but not along the U.S. mutual fund versus U.K. ACS dimension, which is the dimension of concern here.