Does Cross-Listing Lead to Functional Convergence?
                                      Empirical Evidence



                                           Meghana Ayyagari*




Abstract

I examine the effect of legal bonding on ownership and control structures of foreign firms cross-listing in
the U.S. Contrary to the predictions of corporate governance convergence theories, there is little evidence
of convergence-related migration to a dispersed ownership structure on cross-listing. The paper finds that
foreign firms use ADRs as a vehicle to sell control blocks, often to a new foreign owner, rather than as a
means to change their governance structure Firms that cross-list and sell stakes to domestic owners are
from large economies with high stock market liquidity. In contrast, firm-level characteristics are more
important predictors of a control change to a foreign owner. Cross-listing firms that sell control blocks to
foreigners tend to be smaller, have low levels of debt and have a high foreign income growth rate. The post
cross-listing performance of firms that undergo a control change is also different from firms that do not
experience a control change.

Keywords: Corporate Governance, Ownership, Legal Environment
JEL Classification: G32 (Financial Policy; Capital and Ownership Structure), K23 (Corporation and
Securities Law)



World Bank Policy Research Working Paper 3264, April 2004

The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the
exchange of ideas about development issues. An objective of the series is to get the findings out quickly,
even if the presentations are less than fully polished. The papers carry the names of the authors and should
be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely
those of the authors. They do not necessarily represent the view of the World Bank, its Executive Directors,
or the countries they represent. Policy Research Working Papers are available online at
http://econ.worldbank.org.



______________________________________________________________________

*PhD Candidate, Robert H Smith School of Business, University of Maryland, College Park, Maryland 20742.
Email: mayyagar@rhsmith.umd.edu. I am extremely grateful to Vojislav Maksimovic and Gordon Phillips for their guidance and
encouragement. I also thank Lemma Senbet, Mark Chen, Thorsten Beck, Inessa Love, Leora Klapper, Peter Murrell and seminar
participants at University of Pittsburgh, University of Virginia, University of Colorado, George Washington University, The
Brattle Group and the Bank of Canada for helpful comments and suggestions.

1. Introduction

          CEOs of foreign companies often cite the desire for increased corporate governance as

one of the motivations for cross-listing on US exchanges.1 This is supported by the Functional

Convergence Hypothesis developed most broadly by Coffee (1999, 2002), which states that

foreign firms incorporated in a jurisdiction with weak investor protection rights cross-list on US

securities markets to "legally bond" themselves to higher disclosure standards and stricter

enforcement. This helps them attract investors who would otherwise be reluctant to invest or

those who would discount such stocks because of the risk of fraud and embezzlement. However,

the empirical evidence on the bonding hypothesis remains mixed. While studies by Reese and

Weisbach (2002) and Mitton (2002) provide empirical support for the importance of legal

bonding, a more recent study by Seigel (2002) finds that US securities laws were quite ineffective

in deterring or punishing insider stealing in a sample of Mexican firms.2

          In this paper, I provide an alternative approach to test the bonding hypothesis by looking

at the effect of legal bonding on ownership and control structures. The seminal work in the area

of law and finance by La Porta, Lopez-de-Silanes, Shleifer, and Vishny (henceforth LLSV)

(1998) has established that the extent of legal protection provided to minority shareholders in a

jurisdiction determines if the firms incorporated in that country have a concentrated or a dispersed

ownership structure. A cross-listing by a foreign firm on a US securities market provides a unique

opportunity to test what happens to ownership structure when a firm migrates from a poor

investor protection environment to one with greater protection for minority investors and better

enforcement (the US market).

1 Following is the statement by Mr. H. N. Sinor, the Managing Director & CEO of ICICI Bank, one of India's largest
banks and the first bank from Asia (excluding Japan) to list on the NYSE : "...we wanted to be much more transparent
in our disclosures. We adopted US GAAP and the norms specified by the Securities and Exchange Commission (SEC)
of the US are much stricter than the norms specified by the Indian regulators - RBI and SEBI. After the ADR issue, we
have a class of investors who are more demanding as far as corporate governance and disclosures go. We wanted that
we should be constantly evaluated by foreign investors and benchmarked against global companies."
http://www.indiainfoline.com/comp/icba/2004.html
2Earlier studies such as Fanto (1996) even suggest that SEC requirements for foreign firms are meaningless. Most
other authors including MacNeil (2001), Cheung and Lee (1995) and Licht (2001) stress the need for additional
empirical work to determine the impact of a US listing on corporate governance.



                                                           1

         Recent literature suggests that there are several views on what could happen to ownership

structure on cross-listing. One view, supported by theories of functional convergence (Coffee

1999) and by the rent-protection theory of ownership (Bebchuk 1999), predicts that migrating to

a corporate law system that effectively limits private benefits of control can produce more

efficient choices of ownership structure.

         For instance, Coffee (2001) argues that dispersed ownership could arise and persist if

there is significant improvement in protection of minority rights in European and transitional

economies. Therefore, if cross-listing on the US market accords at least some of this

improvement in investor protection for firms listing from these countries, a change in their

ownership structures would be expected. John and Kedia (2000) in fact, make the prediction that

firms from weak legal regimes should experience rapid changes in corporate governance on

cross-listing.

         A second view is the path-dependence theory of ownership, proposed by Bebchuk and

Roe (1999), which states that the ownership structure of a firm at any point is influenced by the

initial ownership pattern due to complementarities, network externalities, and sunk costs. If a

firm starts out with a concentrated ownership structure, the controlling owner is likely to retain

control or sell the control block to a new incoming shareholder, rather than sell voting rights to a

dispersed group of shareholders. Therefore, the theory predicts that the initial ownership patterns

of foreign firms persist after cross-listing.

         This paper tests the two theories by using a unique database of 425 firms from 42

countries that cross-listed on a major exchange in the United States. I examine time-series

changes in ownership and control structures around the date of cross-listing. The aggregate

picture shows that a majority of the firms that cross-list have an initial controlling shareholder

(73%) and family ownership is the most dominant form of control­roughly one in two firms has a

family as the controlling owner. There also exists a high degree of separation of ownership and

control rights in these firms through the use of pyramidal structures and other control-enhancing


                                                   2

features. And as expected from La Porta, Lopez-de-Silanes, and Shleifer (1999), most of the

cross-listing firms that are widely held, to start with, are from English common law countries.

         When I examine the change in ownership structure over time, I find that there is a small

decline in the percentage of voting rights held by the controlling shareholder in cross-listing firms

as compared to a benchmark sample of firms that do not cross-list. This decline is largest for

firms from countries with poor investor protection (the French civil law countries) and for firms

that start out with initial owners as a state or another corporation. Interestingly, I find that the

decline decreases with time. The mean voting rights held by the controlling shareholders decrease

in the year of cross-listing and for three years thereafter. However, the controlling shareholders

start reconsolidating their position in the fourth and fifth years after cross-listing.

         The decline in voting rights around the time of cross-listing is not large enough to result

in a widely-held ownership structure, suggesting little evidence of any convergence-related

movement. This persistence in concentrated ownership patterns, even after listing on US markets,

lends support to the path-dependence theory of ownership structures.

         Although most of the companies continue to be controlled by an ultimate owner on

cross-listing, I document a high incidence of changes in the identity of the controlling

shareholder. Controlling shareholders in 35% of the sample of cross-listing firms approach US

markets as a way to divest their stake and sell their control block. Several reasons contribute to

why a US listing might facilitate such a sale, including greater visibility (Baker, Nofsinger, and

Weaver (2002)), superior earnings forecasts (Lang, Lins, and Miller (2003), and higher valuations

(Doidge et al. (2003)). I elaborate on the reasons in the next section of the paper. Many of the

exiting controlling shareholders happen to be either the state or a family and the new entrants are

predominantly widely held corporations. Further, I find that more than half of the control changes

are to foreign controlling owners.

         I then turn to firm and country-level data to get a better understanding of which

characteristics influence a change of control in companies on cross-listing. I use the Worldscope


                                                   3

database as a source of firm-level data for the cross-listing firms. I find that country-level

characteristics like size and liquidity of the home country's market and the extent of legal rules

and investor protection are significant predictors of a control change to domestic owners. In

contrast, individual firm characteristics, such as size, leverage and market valuation are important

predictors of control change to foreign owners.         Additional tests show that capital control

regulations and FDI restrictions in the home country decrease the probability of control changes

to foreign shareholders.

        Post-listing characteristics of firms that cross-list show that all firms have significantly

higher foreign income growth rates. So an ADR listing is able to expand the international

orientation of the foreign firms. I also find that firms that undergo control changes, especially to

foreign owners significantly increase their leverage ratios.

        In summary, the paper shows that the increased regulation associated with US markets is

not effective enough in changing the ownership concentration of foreign firms that cross-list on

US exchanges. The feature of the US corporate law system that limits private benefits of control

doesn't produce an increase in dispersed ownership structures in firms issuing an ADR. However,

cross-listing does provide an opportunity for existing shareholders to sell out to new owners as

evidenced by the high incidence of control transfers in the study sample. The evidence suggests

that rather than bonding there

        The paper is organized as follows. Section 2 examines related literature and discusses the

main hypothesis this paper tests. Section 3 discusses the data and summary statistics. Section 4

presents the main results. Section 5 concludes with implications for future research.




2. Motivation and Related Literature

        In this section I identify the different ways in which foreign companies can approach the

US markets and the reasons for them to do so. Section 2.1 discusses what an American




                                                  4

Depositary Receipt (ADR) is and the Securities Exchange Commission (SEC) rules and

regulations that a company issuing an ADR is subject to. Section 2.2 discusses the various

reasons why a company may cross-list on a US exchange including implications of the bonding

hypothesis. Section 2.3 lays out the main hypothesis I test in this paper regarding ownership

changes around cross-listing.




2.1.What are ADRs?

        Firms can migrate to US securities markets either by listing foreign shares through an

American Depository Receipt (ADR) on the NYSE, NASDAQ or AMEX or by listing shares

directly on these exchanges. An ADR is a negotiable instrument issued by a US commercial bank

acting as a depositary and represents a fraction or a multiple of one or more shares of the foreign

stock. The shares of foreign stock are on deposit with the depositary's correspondent bank (the

custodian bank) located in the home country of the issuer. Firms seeking the benefits of ADRs

can choose from four different types of sponsored ADR programs, each with its own set of

benefits as well as its own set of legal and regulatory requirements: Level I, Level II, Level III,

and Rule 144A/GDR.

        A Level I depositary receipt program is the simplest method for companies to access the

U.S. and non-U.S. capital markets. Level I depositary receipts are traded in the U.S. OTC market

with prices published in the "Pink Sheets" and on some exchanges outside the United States.

Establishment of a sponsored Level I program does not require full SEC registration and the

companies do not have to report their accounts under U.S. GAAP or provide full SEC disclosure.

A company cannot raise capital via a Level I ADR.

        Companies that wish to raise capital or make an acquisition using securities, use

Sponsored Level II or Sponsored Level III depositary receipts. Level II and Level III depositary

receipt programs require SEC registration and adherence to applicable requirements for U.S.

GAAP. Reporting requirements include Form F-6 registration statement to register the ADR,


                                                 5

Form 20-F registration statement that contains detailed financial information about the issuer, and

Form F-1 to register the equity securities underlying the ADRs and annual reports.

         Foreign companies can also access the U.S. and other capital markets through Rule 144A

and/or Regulation S Depositary Receipt facilities without SEC registration. Rule 144A programs

provide for raising capital through the private placement of Depositary Receipts with large

institutional investors (Qualified Institutional Buyers or QIBs) in the United States. Regulation S

programs provide for raising capital through the placement of Depositary Receipts offshore to

non-U.S. investors in reliance on Regulation S.

         Since I am interested in studying the effect of US laws on ownership changes for cross-

listing firms, I use only Level II and Level III ADRs for this study.




2.2. Why do Firms Cross-List? The Bonding Hypothesis

         There exist several reasons why a company might want to list abroad: to expand investor

base, to increase stock liquidity, to improve the terms on which they can raise capital, to increase

visibility of the company, and to achieve non-financial benefits such as increasing customer base

by broadening product recognition amongst investors of the host country. These and other

motivations are reviewed in detail in Karolyi (1998) and Pagano, Roell and Zechner (2002).

         This section examines the investor protection and corporate governance argument applied

to cross-listing. According to the Bonding Hypothesis, exchanges with strict regulation and

disclosure requirements are the ones attracting foreign firms rather than exchanges with

inadequate supervision. Below are some of the aspects of greater regulation that cross-listing on a

US stock exchange via a Level II or Level III ADR entails:

     · Issuer is subject to SEC enforcement

     · Issuer is committed to providing full information and to reconciling financial statements

         to GAAP




                                                  6

    · Issuer is subject to listing requirements imposed by the exchange it lists on

    · Issuer is exposed to the "scrutiny of reputational intermediaries including US

        underwriters, auditors, debt rating agencies and securities analysts" as suggested by

        Coffee (2002). Lang, Lins and Miller (2002) show that the earnings forecasts of cross-

        listing firms are superior to those that do not cross-list.

    Cantale (1996) and Fuerst (1998) show that firms signal quality by listing on strictly

regulated markets. They predict that companies that cross-list on a more demanding exchange

should exhibit higher profitability than those that list on exchanges with not so severe regulatory

standards. Stulz (1999) predicts that companies from countries with poor legal standards can

reduce the agency costs of external finance by subjecting themselves to tighter standards.

        However, empirical testing of these models has yielded mixed results. Reese and

Weisbach (2002) find that once they control for firm size, firms from weak investor protection

countries are less likely to list in the US. However, firms from these countries that do cross-list

issue more equity on cross-listing. Another paper that provides a direct test of this hypothesis is

by Doidge et. al. (2003), who find that corporate governance seems to be secondary compared to

other factors that drive large firms to tap US markets. Seigel (2002) finds that US securities laws

are ineffective in punishing expropriation and insider stealing especially if the foreign firm has no

assets in the US. So it all comes down to the extraterritorial reach of US securities laws.

        The rights accorded to holders of ADRs are determined by a complex interplay of the

issuer's home country corporate laws and the US federal securities law. Shareholder rights being

the purview of corporate law, the corporate law of the foreign issuer's home country determines

the rights and protections for shareholders of the foreign issuer. However, certain corporate

governance issues are regulated by US federal securities laws and the rules of the national

securities exchanges on which the companies are listed. This is the basis on which the bonding

hypothesis presumes that US securities law deters corporate malfeasance by foreign issuers.

Further, the recently instituted Sarbanes-Oxley Act of 2002, which is aimed at higher corporate


                                                  7

accountability to shareholders, extends to all public corporations, including ADR issuers, which

have their securities, listed on US exchanges.

       However, several factors undermine the importance of US securities law with respect to

ADR holders. I discuss some of these below:

   · The laws of many countries recognize the depositary bank as the shareholder of the

       securities underlying the ADR program and not the ADR holders. To be recognized as

       the shareholder of the underlying securities, many countries require the ADR holders to

       remove their ADR shares from the program and receive the underlying shares, thereby

       becoming a registered shareholder of the foreign issuer.

   · No NYSE rules regarding notice of shareholder meetings or disclosure of meeting agenda

       items to holders of ADRs

   · Foreign issuers are not subject to the SEC's proxy rules

   · Foreign issuers may obtain waivers of holding annual shareholder meetings, including

       quorum requirements for these meetings

   · Depository agreements: In addition to federal securities law, the depository agreements

       play an important role in determining voting rights of ADR holders. Under some

       agreements, the issuer agrees to recognize the holders of ADRS as shareholders of the

       underlying securities and not the depository bank. There is no difference made between

       holders of ADRs and domestic investors in the foreign country. However some

       depository agreements severely limit voting rights of the ADR holders. These include

       prohibiting ADR holders' voting rights other than in limited circumstances; some

       agreements provide that if ADR holders do not vote, shares are autoproxied to the issuer;

       many include a disclaimer that there is no guarantee that ADR holders will receive proxy

       materials in time too exercise their votes and in some agreements, a depository bank that




                                                8

         is recognized as the shareholder may not split its vote if ADR holders provide conflicting

         voting.

         There also exist several provisions in current SEC regulations to allow for

accommodations to foreign issuers. For instance, one of the requirements of Sarbanes Oxley is

that listed companies have audit committees consisting of external, independent directors. Some

foreign issuers are exempt from this requirement and allowed to have internal auditors, especially

when it is authorized by the issuer's home-country statutes. Foreign issuers are also exempt from

Regulation Fair Disclosure (Regulation FD) which prohibits the practice of selective disclosure of

material nonpublic information by U.S. public companies to securities industry professionals.

         In sum, the non-uniformity in regulations that apply to domestic and foreign issuers

emphasizes the need for empirical work at the firm level to establish whether legal bonding

influences the corporate governance structure of the foreign firms that cross-list on US

exchanges. The following section discusses the thought experiment behind this paper.




2.3. Legal Bonding and Ownership Structure

         This paper tests the effectiveness of increased corporate governance associated with US

legal rules on cross-listing firms. LLS (1999) have shown that the level of investor protection in a

country is an important determinant of ownership structure of firms. Proponents of the rent-

protection theory of ownership like Bebchuk (1999) argue that countries with poor investor

protection like French civil law countries have concentrated ownership structures because of the

high value of private benefits of control in these countries. In contrast, countries following

English common law have more dispersed shareholder structure because the legal rules and

enforcement mechanism in these countries limit the existence of control benefits for a large

shareholder.

         If the bonding hypothesis were to hold, controlling shareholders of cross-listing firms

would no longer be able to consume private benefits when subject to stricter US securities laws.


                                                 9

John and Kedia (2002) in discussing ownership as one of the instruments of corporate governance

predict that firms migrating to exchanges with stricter regulation should see a "rapid change in

their existing governance structures". Coffee (2001) also predicts a movement towards dispersed

ownership structure when there is a reform in the extent of minority shareholder rights protection

because of investors' preferences for ownership and liquidity.

          The above legal hypothesis of ownership structure is supported by corporate governance

convergence theories. According to Coffee (1999) and Gilson (2001), convergence in corporate

norms does not always require a formal legislative amendment of corporate laws and codes in

different countries. Firms listing on foreign stock exchanges opt into foreign governance

standards, thus developing a functional substitute to formal convergence. Therefore cross-listing

firms practice legal bonding as a form of functional convergence3.

          Related to this is the path dependence view of ownership structure, which says that there

are significant sources of path dependence in patterns of corporate ownership structure. Initial

ownership structures have an effect on subsequent structures because of sunk adaptive costs,

network externalities and complementarities. Further, if the initial pattern of ownership structure

provides one group of players with relatively more wealth and power, this group would have a

better chance to influence the corporate rules, which in turn would reinforce the initial patterns of

ownership. So the question comes down to whether the regulations provided by US securities

laws for cross-listing firms are strong enough to override the home country's lax regulatory

system.

        There seem to be other advantages associated with making the transition from a

concentrated ownership structure to a dispersed form. For example, Himmelberg, Hubbard and

Love (2002) find that ownership concentration increases the cost of capital and leads to


3Legal theorists distinguish between "formal convergence" and "functional convergence". Formal convergence in
corporate governance requires a legislative amendment of corporate codes and Coffee (1999) argues that legislative
inertia and role of special interest groups actually hinder formal convergence. Functional convergence, on the other
hand, doesn't require formal institutional change and is brought upon by globalization forces which compel firms to
revise their governance structure.



                                                             10

inefficiencies in investment. A dispersed ownership structure supported by US laws would help

overcome the associated under-investment problem for cross-listing companies.

         Cross-listing on a US exchange may also affect ownership structure by facilitating a

control transfer. Controlling shareholders who wish to sell their stakes might view issuing an

ADR as a way to increase the price at which they can sell their stakes. Several factors could

contribute to this: Recent work by Baker, Nofsinger and Weaver (2002) shows that a US cross-

listing increases the firm's visibility in the media and among analysts. Lang, Lins and Miller

(2002) show that the earnings forecasts of cross-listing firms are superior to those that do not

cross-list. This increase in information associated with cross-listing is helpful in making the

company attractive to more potential buyers. Further, Diodge et al (2003) have shown that there

is a valuation differential associated with cross-listing firms. Finally, just listing on a US

exchange could be a signal of quality and adherence to corporate governance for the cross-listing

firm. All these reasons contribute to possible changes in ownership and control structure after

cross-listing.

         This paper is also related to existing literature on ownership changes surrounding the IPO

process. Many cross-listing firms are also first-time issuers in that they IPO on US exchanges.

Rydqvist and Hogholm (1995) investigate the going public decision in a sample of family-owned

corporations in Sweden and find that the original owners liquidate their stakes in a majority of the

cases. In a more recent study employing a larger sample, Holmen and Hogfeldt (2002) find that

ownership remains controlled subsequent to an IPO offering. While several other studies examine

the identity of ownership structures in different countries (including LLS 1999, Claessens et. al.

(2000)), this paper is one of the first to study time series changes in ownership structure.




                                                   11

3. Sample Construction and Descriptive Statistics

         To study the relationship between cross-listing, corporate governance and ownership

structure, I create a sample of foreign firms that listed on a major US stock exchange since the

year 1990. The complete list of cross-listing firms is obtained from the Bank of New York's ADR

Index. The cross-listing year and type of issue details are also verified using Citibank's ADR

Universe. In the event of a discrepancy between the Bank of New York's ADR Index and

Citibank's ADR Universe, I use the information that corresponds closest to the list of cross-listing

firms obtained from the NYSE and the NASDAQ.

         To be included in the final sample, the following criteria are imposed: (1) Since I am

interested in testing the effect of US legal rules, I include only Level 2 ADRs and Level 3 ADRs

in the sample. Firms that list via Level 1 ADRs (OTC stocks) or Rule 144a are subject to little or

no SEC disclosure requirements and do not have to comply with US GAAP and so are excluded

from the sample. (2) I use only those cross-listing firms that are covered by Worldscope database.

Since most of the firms that cross-list are big companies with high market capitalization, there is

little danger of losing out a large sample of the cross-listing firms by using Worldscope. Further, I

have access to firm-level data in Worldscope only from the year 1990. So the year 1990 is the

initial cut-off date for the sample. (3) Finally, I include only firms for which I can get ownership

information from one year prior to cross-listing to one-year after cross-listing. So there is at least

three years of ownership information for each cross-listing firm. The final sample extends from

1990 to 2002 and contains 425 firms that cross-listed on a major US stock exchange.




3.1. Cross-listing by Year, Exchange and Country

         Panel A of Table I shows the distribution of the sample by year of listing and exchange.

More number of firms list on the NYSE, which has the strictest disclosure requirements than on

the NASDAQ. Very few firms list on AMEX stock exchange.




                                                  12

        Panel B of Table I presents data on cross-listing by country of origin. Following LLSV

(1998), the countries are further classified on the basis of their legal tradition into English

Common Law, French civil law, German civil law, Scandinavian civil law and Socialist Law

countries. I exclude cross-listing firms from China since it does not belong to any of the legal

traditions identified above. I also exclude cross-listing firms from Canada since none of them

issue ADRs and all are direct cross-listings. The sample confirms the findings of Reese and

Weisbach (2002) that more firms list from French civil law countries than from English common

law countries and they do so disproportionately on the NYSE than on the NASDAQ. While 81%

of firms from French civil law countries list on the NYSE, only 53% from the English common

law countries do so.

3.2. Ownership Characteristics--Prior to Cross-listing

        For each of the companies in the sample, I identify who the ultimate owner of the

company is, what his cash flow to voting rights are and whether there has been a change in

ultimate owner from two years prior to issuing an ADR to five years, post-listing.

        As a starting point for the data collection, I use Worldscope, which provides annual data

on the names and holdings of all owners that hold more than 5% of a company's stock. This

information is supplemented with data from a variety of sources to include all owners who have

more than 5% of outstanding stock and are missing in Worldscope. All the sources used are listed

country-wise in Appendix B.

        For the years prior to ADR listing, the main sources of data are Worldscope and company

annual reports filed either with the domestic stock exchange or published on the company web

pages. Other data sources include country specific company handbooks like the Brazil Company

Handbook and Venezuelan Company Handbook and international databases such as Mergent

International.

        The ownership information for the years after ADR listing is more easily available

because a large fraction of the ADRs file Form 20-F with the SEC. Section 7 of this form titled


                                                 13

Major Shareholders and Related Party Transactions (prior to 1999, it was Section 4 titled

Control of Registrant) requires the ADR issuer to list all persons known to the company to own

directly or indirectly more than 5% of the issued classes of stock. Further, the company is

required to mention if there have been any significant changes in ownership structure in the past

three years. Consequently, many companies provide detailed time-series information on major

shareholdings in their 20-F statements. The source for the 20-F filings was both Lexis-Nexis and

Hoovers Online. In all cases, the ownership information is collected for the end of the fiscal year

or the closest possible date. This is because ownership information typically lags by about six

months in the 20-F filings and by one or two years in the Company Handbooks and Worldscope.

         I include companies for which it is possible to determine that all the major shareholders

are listed or if it is possible to identify the shareholder who owns 50% or more of the cash flow

rights. For cases in which the shareholder owns less than 50% and the rest is held by nominee

accounts, I use the 20-F statements to decide whether or not to include the company in the

sample4.

         The ADR firms are divided into firms that are widely held and those with controlling

ultimate owners. The procedure for identifying the ultimate owners is similar to the one in LLS

(1999) and Claessens et. al. (2000). I use two cut-off levels, 20% and 10% to identify ultimate

owners. An ultimate owner is therefore defined as someone who has 20% (or 10%) or more of the

control votes of the company and is not controlled by anyone else. This sometimes involves

tracking down ownership along multiple chains to find the major shareholders in the major

shareholders and so on till we arrive at an ultimate owner who is not controlled by anyone else.

The identity of the ultimate owners is further classified into five types: widely held corporation,

widely held financial institution, families (or individuals with large stake), state (or government

or kingdom) and miscellaneous. The miscellaneous category typically includes cooperatives,


4This is different from Claessens et al (2000) who actually drop such companies. My justification for
including it is that I use 20-F statements to verify whether the nominee accounts are widely held or in the
hands of a controlling shareholder.


                                                      14

voting trusts (as is common in Netherlands) or a business group with no single controlling

investor. If a firm does not have an ultimate owner, it is classified as widely held. If a firm has

multiple ultimate owners, I pick the largest ultimate owner for the analysis.

    Panel A of Table II presents ownership statistics at the 20% cut-off level and Panel B

presents statistics at the 10% cut-off level. The firms are classified by the legal system they

belong to and by the identity of the controlling shareholder. The two panels show similar trends at

the two cut-off levels. The largest fraction of cross-listing firms that are widely held are from

English common law countries (61% and 49% at the 20% level and 10% levels respectively).

This is to be expected since LLS (1999) have shown that dispersed ownership is a characteristic

of only English common law countries.

    Further, family ownership seems to be the most predominant form of control across all other

legal traditions. For the whole sample, out of 425 cross-listing firms, at the 10% cut-off level, 310

firms have a controlling owner of which 151 firms are controlled by a family. The smallest group

of ultimate owners is widely held corporations (21 out of 310). The state controlled firms are

typically ones that used American Depositary Receipts as a privatization vehicle.




3.3. Degree of Separation of Ownership and Control--Prior to cross-listing

         For each firm in the sample, I make the distinction between ownership (cash flow rights)

and control (voting rights) using the procedures in Claessens et al (2000) and LLSV (1999). As

an example, consider the simplest case in which Firm A owns 25% of Firm B which in turn owns

60% of Firm C. Assuming that A is a widely held corporation, C is classified as being controlled

by a widely held corporation, which is A, who owns 15% of the cash flow rights (product of the

ownership stakes along the chain) and 25% of the voting rights (weakest control chain) of C. In

many cases, the ultimate owner (here A) could have several control right chains through which he

controls C. Then the ultimate control share is given by the sum of the weakest control shares

along each chain. For instance if A also held 50% of Firm D which held 35% of C, then ultimate


                                                 15

control share of A via B and D would be = 25+35 =60% while the total cash flow rights of A

would be = 0.25*60+0.50*35 = 32.5%.

        Panel A of Table III presents the cash flow and voting rights of the controlling

shareholders for the full sample and also on the basis of the identity of controlling shareholder

and legal tradition. The data are presented for the year of cross-listing. Full sample characteristics

show that the percentage of voting rights held by the controlling shareholder are significantly

different from the cash flow rights at the 1% level. This separation of ownership and control is

further reflected for each type of controlling owner. The separation is maximum in the case of

family controlled enterprises (-5.12 % at 1% significance) and least for firms controlled by a

widely held corporation. (-1.48 %). The degree of separation is also seen by looking at the ratio of

cash flow rights to voting rights (C/V). The mean C/V for the whole sample is 0.89, which

implies that the mean controlling shareholder can control 100% of the voting rights using only

89% of the cash flow rights. The same statistics are reported for different legal traditions.

Separation is least significant for firms listing from English common law countries (-1.8%) and

maximum for firms from French civil law countries (-6.8%). There does not appear to be much

separation of cash flow and voting rights in the firms listing from German and Scandinavian civil

law countries or from Socialist law countries.

        The ratios are also significantly different from each other across different legal traditions

as shown in Panel B of Table III. Panel B of Table III also reports the mean ratio of cash flow to

voting rights when countries are classified on the basis of Law and Enforcement, Financial

Development and Transparency. To investigate if the degree of law and enforcement in the home

country of the cross-listing firm affects the ratio of cash flow to voting rights, I use three alternate

measures from LLSV (1998): Shareholder Rights, Law and Order Score and Judicial Efficiency.

The countries are classified into high and low depending on the median value of each of these

measures.




                                                  16

        Shareholder Rights is a measure of shareholder rights protection in the country. It is an

index of `anti-director" rights ranging from 0 to 6 from LLSV (1998) which is formed by

aggregating six important shareholder rights. Law and Order Score is a measure of the law and

order tradition in the country ranging from 1 to 10 from the International Country Risk Guide.

Judicial Efficiency is another index scored 0 to 10 from the country-risk rating agency Business

International Corporation that is an assessment of the efficiency of the legal environment from the

investors' perspective.

        On all three measures, I find that firms need a larger portion of cash flow rights to control

100% of the firm in countries with good legal rules and enforcement. In comparison, the ratios

are significantly lower for cross-listing firms from countries that are not protective of investors.

        In addition to shareholder protection variables, I investigate if the ratio of cash flow to

voting rights in cross-listing firms also varies with the level of financial development of their

home country. Lins, Strickland and Zenner (2003) and Reese and Weisbach (2002) show that

firms from emerging markets are more likely to cross-list. I use three different but comparable

measures from the Financial Structure Database constructed by Beck et. al (2000) : Higher values

of The Index of Financial Structure indicate a more market based financial system. Developed is

a dummy that takes the value 1 if the firm's home market is developed and 0 if it is an emerging

market firm. Overall Size is a variable indicating the overall size of the financial system (both the

stock market and the banking sector). See Appendix A for a detailed description of the variables

definition and their sources.

        I find that ratio of cash flow to voting rights is significantly lower for emerging

economies and for firms coming from countries with small financial systems. Firms from

countries with more market based economies have higher cash flow to voting rights indicating

that financial structure of the home market is a significant determinant of the extent of control

benefits in firms.




                                                  17

        Finally since all firms that cross-list via a Level II or Level III ADR have to conform to

US GAAP, I investigate if the cash flow to voting rights ratio varies with home country's

accounting standards. I use an Accounting index produced by the Center for International

Financial Analysis and Research Inc that is a measure of disclosure requirements and

transparency in accounting statements produced by firms in different countries. Poorer the

accounting standards, one would expect controlling shareholders to be better able to divert funds

and expropriate requiring a lower amount of cash flow to voting rights. I find that the ratio is 0.86

for firms with poor accounting standards, which is significantly lower than the ratio of 0.95 for

firms with good accounting standards.




3.4. Use of Control Enhancing Features --Prior to cross-listing

        The ways in which cash flow rights can be different from control rights is through the use

of dual classes of stock, pyramidal structures and cross shareholdings. I create dummies for each

of these categories. Dual-Class equals one if the firm issues multiple classes of shares with

different voting rights, Pyramid equals one if the controlling owner exercises control through at

least one publicly traded company and Cross-Shareholding equals 1 if the company has a

controlling shareholder and owns a percentage of shares in either its controlling shareholder or in

any other company in that chain of control. I also define dummies to identify if there is just one

controlling owner, Single, and if any person from higher management is also from the controlling

family, Management. See Claessens et al (2000) for detailed description of these variables.

        Table IV presents details on the prevalence of various control enhancing features in

different cross-listing firms. As in Table III, firms are classified on the basis of their home

country's legal rules, law and enforcement, financial development and accounting standards. The

use of multiple classes of shares with different voting rights is significantly higher (lower) in

firms from French civil law (English common law) countries than in firms cross-listing from

other countries. The use of pyramidal structures seems to be quite high (around 50%) in most of


                                                  18

the companies irrespective of the legal system they belong to. Cross-shareholding structures are

more common in Scandinavian countries in this sample.

         When countries are classified on the basis of shareholder rights protection and legal

enforcement, I find that pyramiding is very common in firms whose home countries are not very

protective of minority shareholders (61%), have low law and order scores (57%) and rank low on

the score of judicial efficiency (53%). Similar statistics hold for the use of multiple voting classes

of shares. 37% of firms listing from countries with poor legal and judicial systems use multiple

classes of shares as compared to only 10% of firms coming from protective legal and efficient

judicial systems. Therefore, I find that the use of control enhancing structures in firms that cross-

list from different countries is very representative of the overall use of these structures in those

economies as discussed in LLS (1999).

         I also find that firms from developed countries and firms from countries with large

financial systems and market based economies are less likely to have multiple classes of voting

shares and less likely to have a single controlling shareholder. This is as expected because a

growing body of research has shown that a country's set of legal rules and legal institutions is

important for explaining the level of bank and stock market development (LLSV 1998). Hence

one would expect ownership concentration levels and general ownership characteristics to vary

with the level of financial development of the home market.

         In addition, the controlling shareholder belongs to upper management in most of the

firms listing from under-developed economies. On the account of level of transparency in the

accounting system of the country, I find that firms that cross-list from countries with poor

accounting standards have significantly higher pyramidal structures, multiple classes of shares

and single controlling shareholders.




                                                 19

4. Empirical results

4.1. Changes in Voting Rights

        In this section, I document the change in voting rights of cross-listing firms after they

cross-list on a major exchange in the US via a Level II or Level III ADR program. I first examine

these changes by classifying firms by their home country's legal tradition. Table III shows that

there is a significant difference in cash flow and voting rights, prior to cross-listing, between

firms belonging to different legal traditions and levels of shareholder protection. It is therefore

worthwhile to examine if the changes in the legal and regulatory environment that cross-listing

firms are subject to, upon issuing an ADR, translate into changes in ownership concentration as

predicted by the rent protection theory of ownership. Finally, I also assess if the changes in voting

rights are different for different groups of shareholders. The purpose is to examine if a US listing

affects certain groups of controlling shareholders more than other groups.

        Panel A and B of Table V report the percentage changes in voting rights at the two cut-

off levels of 20% and 10% ownership concentration during a six-year period (-1,+5) around the

cross-listing year. The statistics reported are group means for different groups of home countries

of the cross-listing firm classified on the basis of legal origin. I also report statistics for a

comparison sample of benchmark firms from each country that do not cross-list.

        A perfect control sample would be the Level I ADRs that trade on OTC markets because

while these firms cross-list, they are not subject to SEC enforcement or exchange listing rules.

However, the lack of SEC filings of these companies makes it extremely difficult to find their

ownership information. To construct the comparison sample, for each cross-listing firm, I find a

firm matched by industry (four-digit SIC) and nearest asset size from Worldscope. If there is no

match at the four-digit SIC level, I look at three or two-digit SIC codes before applying the size

criterion. The final restriction for the non cross-listing firm to be included in the sample is the

availability of good ownership information from Mergent International and country specific




                                                  20

sources. In about 20% of the cases, the availability of ownership information over-rides the size

criterion. Further, to minimize the data collection without losing out on the representative-ness of

the comparison sample, I adopt the following sampling method: If a country has five or less than

five firms issuing ADRs, I try to identify a match for each of the five firms. However, if the

country has more than 5 firms cross-listing in the US, then I randomly sample 5 of them and

identify comparison firms for each of those. I track ownership details for the comparison firms for

only two years, the year prior to cross-listing (of the firm that it is a match to) and four-five years

after cross-listing.

        One of the sample selection issues here is that the cross-listing firms are very different

from the home country firms that do not cross-list. To address this issue, I compare the ownership

characteristics of the domestic firms to the cross-listing firms prior to their listing in the US. I find

that both samples have high concentration of ownership structures and family ownership is

predominant in both the domestic non-crosslisting firms and the cross-listing sample. In addition,

the mean separation of cash flow and voting rights is similar in both samples. For instance, family

owned firms that cross-listed in the US had a mean ratio of 0.85 in the year prior to cross-listing

compared to a mean ratio of 0.83 for family owned firms in the control sample. Cross-listing

firms from French civil law countries had a ratio of 0.86 compared to 0.84 for control sample

firms from these countries. Test of means reveals all statistics to be significant at the 1% level.

These tests show that the ownership profile and the potential for private benefits of control is

similar in the domestic control sample firms and the firms that cross-listed, suggesting that the

control sample constructed here is a good match for the main sample.

        A comparison of the ownership structures of the control and the treatment sample exhibit

significant similarities. Both samples have concentrated ownership structures, predominantly

family ownership and similar separation of cash flow and voting rights. The ratio of cash flow to

voting rights is 0.86 for the cross-listing firms from French civil law countries and is 0.84 for the




                                                  21

firms that do not cross-list from these countries. The breakdown of statistics by type of owner is

not reported for brevity and is available from the author.

         Both panels A and B of Table V report negative but insignificant decreases in the year

prior to cross-listing (-2, -1). There are highly significant but small decreases in voting rights in

the year of the cross-listing and thereafter. The largest decrease in voting rights in both panels

occurs during the (-1, +3) period and this is slightly reduced during the (-1, +5) period. For

example at the 10% level, voting rights decrease by 6.03% during the period (-1, +3) and only by

5.46% during the period (-1, +5). Similar results are obtained when the sample size is kept fixed

over all periods (results not reported). In contrast, for the comparison sample, the results show a

slight but insignificant decrease of -2.96% (-3.45%) at the 20% (10%) ownership concentration

level.

         The results suggest that issuance of an ADR is accompanied by a decrease in control

rights held by the concentrated shareholder when compared to firms from the same country that

do not issue an ADR. However, the decrease is definitely not of a large enough magnitude to

result in a change to a widely held ownership structure. In fact, in the fifth year after cross-listing

there is actually a slight increase in percentage of control rights.

         Table V also reveals that the firms from French civil law countries experienced the

largest decrease in voting rights (-9.38% at 20% level and ­9.31% at 10% level) over the period

(-1, +3). These changes are significantly different from zero at the 1% level. In comparison,

corresponding numbers for firms from English common law countries were ­2.52% and -3.39%

respectively. The latter firms also registered a slight increase in voting rights over the period (-1,

+5). Ownership changes seem to be the least in Scandinavian and Transition economies.

         In Table VI, I examine the change in voting rights for different types of controlling

shareholders. The table reports changes in voting rights for the original controlling shareholder

over different periods of time surrounding the year of cross-listing. If the original controlling




                                                  22

shareholder loses control during the period of study, then he is no longer included in the

observations.

        Table VI reveals some important differences between family and state controlled firms. I

find that family controlled firms experience only a small decline of 5.8% over the period (-1, +3),

which reduces to 3.9% over the period (-1, +5). This suggests that family controlled firms reduce

their control rights up to year three after cross-listing but actually start increasing control

thereafter. This could reflect one of two things: The controlling shareholders reduce their stake at

the time of cross-listing to signal their interest in corporate governance and attract more investors

but tend to reconsolidate soon after. Alternatively, it could also mean that they take advantage of

the valuation premium associated with cross-listing (Doidge et. al (2003)) and sell a part of their

stake only to buy back after year three when valuations have returned to normal.

        State controlled firms, on the other hand, experience significant and large declines of

17.2% and 22% over the periods (-1, +3) and (-1, +5) respectively. So unlike the family

controlled firms' state controlled firms continue to divest their stake in year four and five. The

large decline in voting rights experienced by state controlled enterprises is not very surprising

because a lot of privatizations of state-owned assets have been carried out by issuing American

depositary receipts. This is especially true for countries in Latin America. The best example is

that of the Brazilian telecommunication giant Telebras, a $37 billion company. When the

Brazilian government decided to privatize Telebras in 1998 and break it up into 12 smaller units,

12 separate ADR issues were established on the New York Stock Exchange. One of the reasons

why governments decide to use global equity markets for privatization, apart from wanting to

increase investor base is that in many countries, the home market is unable to absorb such a large

tranche. Hence they are compelled to look at international markets to raise capital. On going

privatizations could also explain the finding that unlike family controlled firms, state controlled

firms continue to divest their stake in year four and five.




                                                   23

         Table VI also shows that companies controlled by corporations and widely held financial

institutions exhibit a similar pattern as family controlled firms. Firms controlled by corporations

and financial institutions experience large declines in control over the period (-1, +3) of 26.7%

and 14.5% respectively which reduces when over a longer period of (-1, +5). This seems to

suggest that cross-listing firms controlled by corporations and financial institutions also do a lot

of selling and buy-back around the cross-listing date to take advantage of valuation premiums.

The larger decline in the case of corporations and financial institutions could reflect the lower use

of multiple classes of shares in these types of firms. Family owned firms, on the other hand,

would prefer to use shares with low or no voting rights to take advantage of the valuation

premium since they value control more. Hence the study shows only a small decline in voting

right shares in the case of family owned firms.

         While the decline in voting rights is not significant to result in a widely-held ownership

structure, there exist other changes in ownership structures as described in the section below.




4.2. Changes in Control of Cross-listing Firms

         In this section, I examine the incidence of control changes taking places in the sample of

firms that cross list on a major exchange in the United States. A change in control is registered

when the original shareholder completely divests his stake to a new party or if the original

controlling shareholder sells a sufficient portion of his stake such that he no longer has the

maximum percentage of voting rights. A change in control is also registered if a widely held firm

sees the emergence of a controlling shareholder within five years after cross-listing.

         Appendix D shows an example of control change associated with an ADR listing in the

case of Vimpel-Communications (VimpelCom), which in 1996 became the first Russian company

to trade on the New York Stock Exchange. The controlling shareholder of the company in the

year prior to ADR issuance (1995) was the founder of the company, Dr Dmitri Zimin, who was

also president and CEO. Dr Zimin owned 45% of the cash flow rights of VimpelCom and


                                                 24

controlled 58.6% of the company's voting stock. By 2001, VimpelCom had two new majority

shareholders: Alfa group who controlled 25% plus two shares of VimpelCom's voting stock via

the company EcoTelecom Ltd. and Telenor, Norway's leading telecommunications company

which controlled 25% plus 13 shares of VimpelCom's voting stock.

    The full sample characteristics show that roughly 35% of the firms that had an original

controlling shareholder experience a change in control within five years of cross-listing in the US.

Panel A of Table VII provides a detailed breakdown of control changes by type of controlling

shareholder. The first column of Table VII provides the number of control changes taking place

for each type of original controlling shareholder and the last row of Table 8 shows the emergence

of new controlling shareholders. Out of the 78 control changes in the study, most occurred in

cases where the original controlling shareholder was a family (26 out of 78) or the state (22 out of

78). The 26 families that exited, transferred control to 7 new families, 8 public corporations, one

state, 4 widely held financial institutions and 3 changes were classified as miscellaneous. Three of

the 26 families completely divested their stake resulting in a widely held corporation. Overall, 11

of the 78 control changes resulted in the original firm becoming widely held.

    Table VII also shows that the largest number of new controlling shareholders are widely held

corporations and widely held financial institutions. Only 16 of the 78 control changes resulted in

a new family being in control compared to 38 changes to a widely held corporation or financial

institution.

    Panel B provides a better sense of net changes in controlling shareholder types. The table

shows the change in controlling shareholder type from prior cross-listing to post cross-listing for

the sample of firms that underwent control changes. There is a total decrease of 11.7% and 24.2%

in the presence of family controlled and state controlled firms respectively, over the period (-1,

+5) indicating that these types of controlling shareholders are most likely to sell their stakes on

cross-listing. On the other side, the controlling shareholder that emerges most after a control

change is a widely held corporation or a widely held financial institution. This implies there is a


                                                  25

degree of movement away from state and family ownership in some companies cross-listing in

the US.

    An interesting result from panel B is that 45.3 % of control changes result in a foreigner

being in control. This seems to indicate that a US cross-listing provides a way for the domestic

controlling shareholder to divest his stake to a foreign owner. Empirical studies have shown that

cross-listing on a highly liquid market like the US increases visibility (Merton's (1987) investor

recognition hypothesis) and also sends out a good signal that the firm believes in good corporate

governance because it is bonding to greater transparency (Coffee (1999) and Stulz (1999)). These

two features would make it easier for a controlling shareholder to sell his control block to a

foreign owner.

    Table VIII presents a first comparison of the companies that cross-list with those that do not

cross-list. The cross-listing sample is split up into firms that undergo a control change and those

that do not. The table computes the difference in median values for some important balance sheet

variables such as leverage, total assets, return on assets, total asset growth and foreign income

growth. The differences are computed by estimating a median quantile regression (LAV) on a

constant, a time dummy variable which captures time relative to the year of cross-listing as well

as control dummies for calendar year and country. I investigate three relative time periods:

PRELISTING (year before cross-listing), CLYEAR (year of cross-listing) and POSTLISTING

(five years after cross-listing).

    The table reveals some interesting patterns for the firms that undergo a control change and

those that don't. All cross-listing firms seem to be significantly larger than companies that do not

cross-list. This finding is consistent with Pagano, Roell and Zechner (2002) who argue that there

exist economies of scale in cross-listing which increases with company size. The company

growth variables show that there cross-listing is associated with high growth rates for the firms.

Total asset growth peaks in the year before (or in the year of) cross-listing for both samples of

companies that undergo a control change and those that do not. The foreign income growth


                                                  26

variable is also significantly larger for all cross-listing companies compared to the control group.

There is also some evidence linking international presence to a control change because the

foreign growth rates for the firms that undergo a control change are larger than the growth rates

for companies that do not undergo a control change.

     The table also shows the difference in leverage ratios for the two samples relative to the

control sample. A significant finding of the table is that the leverage ratios of the firms that

undergo a control change are significantly higher in the year of the cross-listing and five years

later. This seems to suggest that some of the firms cross-list to increase their debt capacity and

this effect of increase in debt capacity is experienced mainly by firms that undergo a change in

their control structure.

     Tobin's Q ratios are significantly higher for all cross-listing firms in the year prior to cross-

listing but this difference remains only for firms that do not undergo a control change. In fact,

firms that undergo a control change in this sample are found to have a decreased Tobin's Q ratio

in the year of the cross-listing. Finally, the return on asset measure is significantly higher for all

cross-listing firms only in the year of cross-listing. This is suggestive of window dressing by the

ADR issuing firms




4.3. Predicting Control Change from Company Characteristics

         The descriptive statistics discussed in the last two sections provide some evidence that

there is a change in the identity of the controlling shareholder in firms that cross-list in the US. In

this section I turn to more rigorous regression analysis to investigate the control changes in detail.

I use duration analysis to determine which company characteristics predict a control change and

multinomial logit analysis to predict when control change results in a foreign shareholder. I also

use multinomial logit analysis on a smaller sub sample of OECD countries to study the effect of

restrictions on foreign direct investment on control change.




                                                   27

         In Table IX, the determinants of a control change occurring within five years of cross-

listing are analyzed using a Cox proportional hazard rate model. This methodology is useful in

predicting discrete events (recorded as failures) in a panel setting and has been used by Pagano et.

al (2002) in examining the decision by firms to cross-list. It estimates the effects of a set of

covariates X influencing the probability of undergoing a control change at time t (the probability

is referred to as the hazard rate h(t)). The model is written as:

                        h(t)=ho(t) exp (X')                                                      (1)

where  is a vector of coefficients and ho(t) is the baseline hazard rate (that is the hazard rate

when all X variables are set to 0). The Cox proportional hazard rate estimator is a semi parametric

estimator and is referred to as proportional hazard because it assumes that the hazard ratio

h(t)/ho(t) is constant for any two firms at any point in time.

      Table IX reports the estimates as exponentiated coefficients (exp (1), exp(2),...) rather as

coefficients (1, 2, ....), because exponentiated coefficients can be readily interpreted as the

effect of unit change in independent variable X on the hazard ratio h(t)/ho(t).

         The set of determinants include both firm-level characteristics and characteristics of the

home country of the cross-listing firm. The set of firm characteristics include age of the firm, size

(measured as logarithm of total assets), leverage ratio, Tobin's Q, total asset growth and total

foreign income growth of the company. I also use lagged values of the following country-level

variables: GDPCAP is the log of GDP per capita of the home economy, TVT_GDP is stock

market total value-traded to GDP which is defined as total shares traded on domestic stock

market exchange to GDP, COMMON is a legal origin dummy which takes the value 1 for

English common law countries and 0 otherwise, PROPERTY is the property rights regulation

index from LLSV (1999) and the Index of Economic Freedom, and is a measure of the degree to

which the government protects private property, and CAPCONTROL is an index of capital

control regulations, as in Harrison et. al. (2003), constructed from the International Monetary




                                                 28

Fund's annual report, Trade and Exchange Restrictions. This is a composite index constructed by

taking into account five types of control restrictions on international transactions. Thus, this

measure is reflective of the openness of a country's economy. The standard errors and p-values

are adjusted for clustering on companies to allow for dependence of errors concerning the same

company.

         Table IX shows that the variables that have the greatest impact on a control change

occurring are size of the company and leverage. A unit increase in leverage decreases the relative

hazard of a control change by 83.2 % and a unit increase in size increases the probability of a

control change by 19.5% respectively. This suggests that larger companies and companies with

less debt are more likely to undergo a control change. This could be because investors are

reluctant to buy control blocks in highly levered companies and it is the larger companies who are

able to cross list outside the home market and take advantage of increased visibility to facilitate a

control change.

         An increase in Tobin's Q ratio decreases the probability of a control change by just 1.4%

(significant at 0.001 level). This seems surprising at first because one would expect investors

wanting to buy control blocks in companies with high market valuation. But a high market

valuation also implies that the concentrated shareholder is more reluctant to sell his control block

and hence there are a lesser number of control changes in these companies. I also find that foreign

income growth enters significantly into the regression. An increase in foreign income growth

results in an increase in the probability of a control change by 6.7%. This table however doesn't

reveal if the foreign income growth predicts a control change to domestic or to foreign

shareholders. This issue is addressed in Table X.

         Among the country-level characteristics, I find that the economic size and liquidity of the

home country's economy have a large positive and significant effect. This suggests that it is

easier to sell control blocks if the home country's economy is big and the stock exchange is very

liquid. Further, control change is less likely in English common law countries, which is to be


                                                 29

expected because most of the cross-listing firms from these nations do not have a concentrated

shareholder and Table VII reveals very few changes from a widely held ownership structure to

concentrated ownership on cross-listing. The extent to which a government protects private

property and enforces laws to protect private property also seems to be determining factor of

control changes. Cross-listing firms are less likely to sell their control blocks to a new controlling

owner on cross-listing if their home market is protective of property rights. So control changes

following ADR issuance are more common in countries that have poor property rights protection.

I also find the interesting result that a unit increase in capital control restrictions decreases the

probability of a control change by almost 8.8%. Therefore probability of a control change on

issuing an ADR is decreased if the home country's economy is not very open and is subject to

severe capital control restrictions. Firms from such economies are not very attractive to investors

even after they issue an ADR.




4.4. Emergence of a new foreign controlling shareholder

         This section investigates if control changes result in a foreign controlling shareholder.

Multinomial logit is used to predict whether a company is more likely to experience a control

change resulting in a new domestic owner, or in a foreign owner, or becomes widely held or does

not experience a control change at all. The regressors are lagged as before and standard errors are

adjusted to allow for clustering.

         Table X reports the relative risk ratios when control change results in a new domestic

owner, a new foreign owner, or in the firm becoming widely held. The base line case is when

there is no control change. The set of independent variables used are the same firm and country-

level characteristics used in Table IX with one change. I substitute capital control regulations

with the level of foreign direct investment, FDI, taken from the IMF's Balance of Payment

Statistics and YEARS_LIBER which is the number of years since the official date of




                                                 30

liberalization of the country (from Bekaert and Harvey 2000).     FDI is a measure of foreign

direct investment in the recipient country scaled by aggregate gross domestic investment.

        There exist interesting dis-similarities between the three groups. Country characteristics

like GDP per capita and TVT_GDP are significant predictors of control change to a new domestic

controlling owner. Odds that a controlling shareholder sells his stake to another domestic

shareholder rather than retaining his stake is 1.14 times higher if his company is located in a

developed economy and is 13.68 times higher if the stock exchange of the home country has high

liquidity. So greater the financial development of the home country's economy and greater the

liquidity of the home country's stock exchange, more companies sell their stakes to a domestic

controlling owner on cross-listing. On the other hand, odds that the controlling shareholder sells

his stake to a foreign owner decrease by 98% with each 1-unit increase in the liquidity of the

home market's stock exchange. This suggests that the size and liquidity of the home country's

stock exchange is a significant determinant of whether control is transferred to a domestic owner

or a foreign owner.

        Table X also reveals that increased property rights regulation increases the odds of a

dispersed ownership structure after cross-listing and decreases the odds of a control change to a

domestic controlling owner. It has a positive but insignificant effect on odds of a foreign

controlling owner emerging post cross-listing. This suggests that use of an ADR to sell a control

block to another shareholder is rare in countries that are protective of property rights. This is

broadly consistent with recent evidence on privatizations. Megginson, Nash, Netter and Poulsen

(2001) find that the sale of control blocks is easier where governments respect property rights.

Hence firms in these countries do not need to approach global markets to sell their control blocks.

        The level of foreign direct investment into the home country also has a significant effect

on the sale of a control block to a foreign owner. This seems to suggest that a foreign sale is

easier when the home country is already open to foreign investors. I don't find any of the country




                                                 31

level characteristics to be significant predictors of the emergence of dispersed ownership

structures on cross-listing.

        Firm-level characteristics, on the other hand, serve as important predictors for control

change to a widely held ownership structure. Odds of a controlling shareholder completely

divesting his stake, resulting in a widely held company, are higher for companies with higher

market valuations. Further, age comes in as a significant predictor only for companies moving

from a concentrated ownership to a dispersed ownership structure (ages decreases the odds by

71%). Further, larger the company, less likely it is to move to a dispersed ownership structure.

        Table X also shows the firm level characteristics that predict control change to a foreign

owner. Size and leverage of the cross-listing are highly significant indicators of a control change

to a foreign owner. The odds are decreased by 37% and 70% for large firms and highly levered

firms respectively. This seems to suggest that only small companies and companies with low debt

are attractive to foreign investors. This both supports and contrasts a recent study by Freund and

Djankov (2000) who find that foreign investment in Korea is focused on larger firms and firms

with low debt ratios. However, the authors in that study focus on direct foreign investment and

not on firms issuing ADRs. In addition, I find that the odds of a new foreign controlling owner

are 1.15 times higher (increases about 151%) among firms with high foreign income growth. The

number of years since official liberalization of the country does not enter significantly into any of

the regressions. In summary, Table X reveals that individual firm characteristics are important if a

firm wishes to use the ADR route to divest to a new foreign owner.

        One of the main policies that governments use to discriminate between foreign and

domestic investors is the use of foreign direct investment (FDI) restrictions. Hence, a good

measure to determine the emergence of a foreign controlling shareholder would be to use a

measure of restrictions on foreign direct investment and foreign ownership of equity in different

countries. However, data on this is available only for a small sample of OECD countries. Golub

(2003) presents a score of overall restrictiveness indicators, RESTRICTION, for OECD countries


                                                 32

based on regulations in three areas: restrictions on foreign ownership, screening and approval

procedures used by countries to limit FDI and other restrictions (that include constraints on the

ability of foreign nationals to manage or work in foreign companies, operational requirements,

stipulations on nationals forming a majority of the board of directors and so on). The score ranges

from 0 to 1 with 1 being the case when no foreign equity is allowed.

         Table XI uses the RESTRICTION measure in a multinomial logit regression to study its

effect on emergence of a foreign controlling shareholder. I find that greater the level of FDI

restrictions in the country, probability of a control change to a foreign owner on issuing an ADR

drops by 94%.

         In summary, sale of control blocks on cross-listing, to shareholders in the same country,

is largely dictated by the home-country's stock market development and legal environment.

However, sale of control blocks to a foreign owner seems to be more influenced by firm-level

characteristics than country-level characteristics of the home country. The tables show that an

ADR issuance helps small, young firms with low debt and a high foreign income growth rate

leapfrog the drawbacks of their home-country environment and complete a successful block sale

on the global market.




4.5. Ex-Post Characteristics

         In this section, I examine the effects of cross-listing and control change on the subsequent

performance of companies. Each variable to be estimated yit (eg: the leverage ratio of a company i

at time t) is regressed on a set of cross-listing dummies as shown below:

yit = 0 +1 f +1CLYEARit + 2YEAR3it + 3YEAR5it + it
                  i                                                                              (2)


where fi denotes a company fixed effect, CLYEARit is a dummy intended to capture the impact of

cross-listing of company I, YEAR3it is a dummy corresponding to the three years after listing,

and YEAR5it is a dummy capturing the permanent shift in the dependent variable after cross-




                                                   33

listing (takes the value 1 after year3 of cross-listing). To eliminate fixed effects, I difference both

sides of the equation and estimate Least Absolute-Value (LAV) regressions to distinguish

between the various impacts. The specification on first-differencing is now:

yit = 1CLYEARit + 2YEAR3it + 3YEAR5it + it                                                          (3)


         Table XII reports the LAV-regression results for cross-listing firms that do not undergo a

control change (Panel A) and for firms that do undergo a control change (Panel B). The two

panels reveal that all firms irrespective of control change experience a big increase in total asset

growth in the year of the cross-listing. The similarities however end here. Total asset growth is

seen to reduce in companies that do not undergo a control change but increases in companies that

do. The companies that undergo a control change experience a 23.6 % reduction in total assets

accompanied by a very high increase in leverage ratios. Companies that do not report a control

change, on the other hand, experience a 1.4% increase in total assets and a 2% decrease in

leverage, both of which are not significant.

     Both samples of cross-listing firms show a peak in total asset growth in the year of cross-

listing but this growth rate is not sustained in the long run. Foreign income growth, on the other

hand, is significantly higher for the long-term. This is consistent with evidence from Pagano et al

(2003) that cross-listing is pursued by more export oriented firms and firms wishing to expand in

international markets.

     There is also a significant difference in leverage ratios of firms that cross-list and undergo

control changes and those that don't. Table XII reveals that firms that undergo control changes

raise significant amounts of debt in the years after cross-listing. Two factors could be responsible

for this result. The high leverage ratios could be associated with foreign firms which issued an

ADR as part of privatization. Recent privatization literature (Frydman et al 2000) has shown that

massive restructuring is associated with firms that privatized and sold stakes to foreign owners

and raising debt could be part of this restructuring process. Another factor that could contribute to




                                                  34

this is the identity of the new owners. The transition matrix in Table VII shows that control

changes resulted in more financial institutions and corporations being in control than a state or a

family. So it could be that foreign equity investors such as banks also lend to the firm thereby

driving the result that firms with control changes raise more debt on cross-listing. This is

consistent with evidence presented in a recent paper by Aggarwal and Klapper (2003) where they

report a similar symbiotic relationship between foreign ownership and lending in a sample of

Indian IPOs.

    Further, companies that undergo a control change also show an increase in ROA and a

decrease in Tobin's Q ratio, which is not exhibited by firms that do not undergo a control change.

Overall, companies that undergo a control change after cross-listing seem to become more export-

oriented and increase their debt capacity substantially. While the accounting profitability

measures shows an increase in profitability in these companies, the market valuation of the

company is seen to reduce. One possible explanation for this could be that if the original

controlling shareholder is better informed of the future value of the company and he uses the

amount of equity retained as a signal of company's value, the divestment on the part of the owner

is interpreted as a negative signal by the market Leland and Pyle (1977).




5. Conclusions

    This paper examines if cross-listing on a US exchange results in a change of ownership and

control structure for foreign firms. The paper tackles this issue in two parts. First it examines if

there is a movement from concentrated ownership structures to dispersed ownership structures as

predicted by recent literature in law and finance. The bonding hypothesis by Coffee (1999)

predicts that foreign firms cross-list on US exchanges to bond themselves to higher disclosure

standards and stricter enforcement. This implies that cross-listing should be associated with some

change in governance structure for the foreign firms. The paper shows that there is no such mass




                                                 35

transformation of ownership structures. Most of the firms that cross-list have concentrated

shareholdings and continue to do so after cross-listing. This finding questions the hypothesis that

legal protections provided by cross-listing are effective enough to cause firms to change their

governance structure.

     Second the paper examines if there any changes in composition of the shareholders on cross-

listing. The paper finds that an ADR listing facilitates a control transfer in a large fraction of

firms and 45% of these transfers results in a foreign controlling owner on cross-listing. The

finding suggests that the ADR is used as a vehicle by the controlling shareholder to sell his

control block. A listing on a US exchange provides him with increased visibility and an increased

investor base required for selling the control block, both in the home country and to a foreign

owner.

     The post listing characteristics of the cross-listing firms reveals that the firms that undergo a

control change have a large foreign income growth and significantly increase their debt capacities

post cross-listing. While their return on assets shows a permanent positive increase, these firms

have a slightly negative Tobin's Q ratio in the long run. On the other hand, companies that do not

undergo a control change do not increase their leverage ratios and have high Tobin's Q ratios in

all years after cross-listing.

     A possible explanation for the absence of a change in ownership concentration is the presence

of complementarities between a firm's selection of governance structure and the prevailing

institutional environment. As long as foreign issuers are still held more accountable to home

country laws and are subject to different governance standards than domestic US firms, the

institutional environment that these firms operate in on cross-listing is not very different and

hence there is unlikely to be a transformation in their governance structures. I turn to a theoretical

exploration of these issues in Ayyagari (2003).




                                                  36

                                          References

Aggarwal, R. and Klapper,L. (2003) : Ownership Structure and Initial Public Offerings.
Unpublished working paper. World Bank

Ayyagari, M. (2003) : Effect of Investor Protection and Strategic Complementarities on
Organizational Design. Unpublished working paper.

Baker, K., Nofsinger, J., and Weaver, D. (2002) : International cross-listing and visibility. Journal
of Financial and Quantitative Analysis 37(3), 495-521.

Bebchuk, L. A. (1999) : A Rent Protection Theory of Corporate Ownership and Control. NBER
Working Paper No. 7203

Bebchuk, L. A. and Roe, M. (1999) : A Theory of Path Dependence in Corporate Ownership and
Governance, Stanford Law Review 52. 775-808.

Beck, T., Demirguc-Kunt, A., and Levine, R. (2000) : A New Database on Financial
Development and Structure. World bank Economic Review. 597-605.

Bekaert, G. and Harvey, C.R. (2000) : Foreign Speculators and emerging equity markets. Journal
of Finance 55, 565-614.

Cheung, C. S. and Jason, L. (1995) : Disclosure Environment and Listing on Foreign Stock
Exchanges. Journal of Banking and Finance 19(2), 347-62.

Claessens, S., Djankov, S. and Lang, L. (2000) : The separation of ownership and control in East
Asian corporations. Journal of Financial Economics 58, 81-112.

Coffee, J. C. (1999) : The Future as History: The prospects for global convergence in corporate
governance and its implications. Northwestern University Law Review 93, 641-708.

Coffee, J. C. (2002) : Racing Towards the Top? The Impact of Cross-listings and Stock Market
Competition on International Corporate Governance. Columbia Law Review 102(7).

Coffee, J. C. (2001) : The Rise of Dispersed Ownership: The Role of Law in the Separation of
Ownership and Control. Yale Law Journal 111, 1-82

Doidge, C., Karolyi, A. and Stulz, R. (2003) : Why are Foreign Firms Listed in the U.S. Worth
More? Journal of Financial Economics forthcoming.

Fanto, J.A. (1996) : The Absence of Cross-Cultural Communication: SEC Mandatory Disclosure
and Foreign Corporate Governance. Journal of International Law and Business 17, 119-207.

Foerster, S.R. and Karolyi, A. (1993) : International Listings of Stocks: The Case of Canada and
the US. Journal of International Business Studies 24: 763-784.

Foerster, S.R. and Karolyi, A. (1998) : Multimarket Trading and Liquidity: A Transactions Data
Analysis of Canada-US Interlistings. Journal of International Financial Markets 8, 393-412.




                                               37

Foerster, S.R. and Karolyi, A. (1999) : The Effects of Market Segmentation and Investor
Recognition on Asset Prices: Evidence from Foreign Stocks Listing in the United States, Journal
of Finance 54, 981-1013.

Freund, C. and Djankov, S. (2000) : Which Firms do Foreigners Buy? Evidence from Korea.
World Bank Working Paper #2450.

Frydman, R., Hessel, M.P., and Rapaczynski, A. (2000) : Why Ownership Matters?
Entrepreneurship and the Restructuring of Enterprises in Central Europe. New York University
Department of Economics Working Paper.

Fuerst, O. (1998) : A Theoretical Analysis of the Investor Protection Regulations Argument for
Global Listing of Stocks. Unpublished manuscript.Yale School of Management.

Gilson, R. (2001) : Globalizing Corporate Governance: Convergence of Form or Function. 39
American Journal of Comparative Law 329.

Golub, S. (2003) : Measures of restrictions on inward foreign direct investment for OECD
countries. OECD Economics Department Working Paper #357.

Harrison, A., McMillan, M.S., and Love, I. (2004) : Global Capital Flows and Financing
Constraints. Journal of Development Economics forthcoming.

Himmelberg, C., Hubbard, G., and Love, I. (2002) : Investor protection, Ownership and
Investment. World Bank Working Paper #2834.

Holmen, M. and Hogfeldt, P. (2002) : A Law and Finance Analysis of Initial Public Offerings.
SSRN Working Paper Series.

John, K. and Kedia, S. (2000) : Design of Corporate Governance: Role of Ownership Structure,
Takeovers, Bank Debt and Large Shareholder Monitoring. Harvard working paper.

Karolyi, A. (1998) : Why do companies list shares abroad? A survey of the evidence and its
managerial implications, Financial Markets, Institutions and Instruments Vol. 7 Blackwell
Publishers, Boston.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., and Vishny R. (1998) : Law and Finance. Journal
of Political Economy 106, 1113-1155.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., and Vishny R. (1999) : The quality of
Government. Journal of Law, Economics and Organization 15, 222-279.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., and Vishny R. (1999) : Corporate ownership
around the world. Journal of Finance 54, 471-517.

Lang, M., Lins, K., and Miller, D. (2003) : ADRs, Analysts and Accuracy: Does Cross-listing in
the US Improve a Firm's Information Environment and Increase Market Value? Journal of
Accounting Research 41(2)

Leland, H.E. and Pyle, D.H. (1977) : Informational Asymmetries, Financial Structure, and
Financial Intermediation. Journal of Finance 32, 371-387.


                                                38

Licht, A.N. (2000) : Genie in a Bottle: Assessing managerial opportunism in international
securities transactions. Columbia Business Law Review 51 (1).

Licht, A.N. (2001) : Managerial Opportunism and Foreign Listing: Some Direct Evidence.
University of Pennsylvania Journal of International Economic Law 22, 325-348.

Lins, K., Strickland, D., and Zenner, M. (2003) : Do non-US firms issue equity on US stock
exchanges to relax capital constraints. Journal of Financial and Quantitative Analysis
forthcoming.

MacNeil, I. (2001) : Competition and Convergence in Corporate Regulation: The Case of
Overseas Listed Companies. University of Aberdeen working paper.

Megginson, W.L., Nash, R.C., Netter, J.N., and Poulsen, A.B. (2001) : The Choice of Private
versus Public Capital Markets: Evidence from Privatizations. SSRN Working Paper Series.

Merton, R.(1987) : Presidential address: A simple model of capital market equilibrium with
incomplete information. Journal of Finance 47, 2035-2054.

Miller, D.(1999) : The Market Reaction to International Cross-listings: Evidence from Depository
Receipts. Journal of Financial Economics 103.

Mitton, T. (2002) : A cross-firm analysis of the impact of corporate governance on the East Asian
financial crisis. Journal of Financial Economics 64(2).

Pagano, M. and Roell, A.A. (1998) : The Choice of Stock Ownership Structure: Agency Costs,
Monitoring and the Decision to go Public. Quarterly Journal of Economics 113, 187-225.

Pagano, M., Roell, A.A., and Josef, Z.(2002) : The Geography of Equity Listing: Why do
Companies List Abroad?. Journal of Finance forthcoming.

Reese, W. and Weisbach, M. (2002) : Protection of Minority Shareholder Interests, Cross-
Listings in the United States, and Subsequent Equity Offerings. Journal of Financial Economics
66, 65-104.

Rydqvist, K. and Hogholm, K. (1995) : Going Public in the 1980s: Evidence from Sweden.
European Financial Management 1(3), 287-315.

Siegel, J. (2002) : Can Foreign Firms Bond Themselves Effectively by Renting US Securities
Laws ? MIT Working Paper

Stulz, R. M. (1999) : Globalization of Equity Markets and the Cost of Capital. Journal of Applied
Corporate Finance, 8-25.




                                                39

                                                                                       Table I
                                                                            Descriptive Statistics
The table presents summary statistics on the number of cross-listings in the United States by year and by country of incorporation. The sample consists of foreign firms that listed
their shares on one of the major US exchanges (Amex, NASDAQ or NYSE) via Level II or Level III ADR Programs. Only exchange listed ADRs are considered because they
require specific accounting and disclosure information for a firm to cross-list. Panel A shows cross-listing statistics by year. Panel B shows the number of firms that were cross-
listed from each country. The countries are also divided into five categories depending on the legal regime they belong to--English Common Law, French Civil Law, German
Civil Law, Scandinavian Civil Law and Socialist Law


Panel A: Cross-Listing Statistics by Year


                                                                       Cross-Listings
          Year                     AMEX                          NASDAQ                          NYSE                     Total
          1990                        0                                2                            3                       5
          1991                        0                                1                           10                       11
          1992                        0                                5                            5                       10
          1993                        1                                3                           19                       23
          1994                        1                                6                           25                       32
          1995                        0                                6                           15                       21
          1996                        0                               16                           25                       41
          1997                        0                               18                           36                       54
          1998                        0                               11                           37                       48
          1999                        1                               21                           23                       45
          2000                        0                               30                           48                       78
          2001                        0                                9                           29                       38
          2002                        1                                4                           14                       19
          Total                       4                               132                          289                     425

Panel B: Cross-Listing by Country


                                                                                 Cross-Listings
   Legal Origin       Nation                           AMEX                  NASDAQ                  NYSE              Total
 English Common Australia                                  0                       7                    6                13



                                                                                          40

                                              Cross-Listings
 Legal Origin   Nation               AMEX    NASDAQ         NYSE  Total
     Law        Ghana                 0        0              1     1
                Hong Kong, China      0        5              3     8
                India                 0        3              8    11
                Ireland               0        8              5    13
                Israel                0        5              3     8
                New Zealand           0        1              4     5
                Singapore             0        2              0     2
                South Africa          0        5              4     9
                United Kingdom        3        30            41    74
                Total                 3        66            75    144
French Civil LawArgentina             0        3              9    12
                Belgium               0        0              1     1
                Brazil                0        1             29    30
                Chile                 0        0             20    20
                Colombia              0        0              1     1
                Dominican Republic    0        0              1     1
                France                0        12            21    33
                Greece                0        2              3     5
                Indonesia             0        0              2     2
                Italy                 0        1             10    11
                Luxembourg            0        4              2     6
                Mexico                1        1             24    26
                Netherlands           0        7             18    25
                Peru                  0        0              2     2
                Philippines           0        1              1     2
                Portugal              0        0              3     3
                Spain                 0        1              5     6
                Turkey                0        0              1     1
                Venezuela             0        0              2     2



                                                    41

                                              Cross-Listings
 Legal Origin   Nation               AMEX    NASDAQ         NYSE   Total
                Total                 1        33            155    189
                Austria               0        0              1      1
                Germany               0        7             14     21
                Japan                 0        10            11     21
    German
   Civil Law    Korea, Rep.           0        3              5      8
                Switzerland           0        2             10     12
                Taiwan                0        2              3      5
                Total                 0        24            44     68
                Denmark               0        1              1      2
                Finland               0        0              4      4
Scandinavian LawNorway                0        1              3      4
                Sweden                0        6              1      7
                Total                 0        8              9     17
                Hungary               0        0              1      1

 Socialist Law  Poland                0        1              0      1
                Russian Federation    0        0              5      5
                Total                 0        1              6      7




                                                    42

                                                                                       Table II
                                                       Ownership Characteristics of the Cross-Listing Firms
The table classifies foreign firms that listed their shares on one of the major US exchanges (Amex, NASDAQ or NYSE) on the basis of their country of incorporation and their
ownership structures. The countries are divided into five categories depending on the legal regime they belong to--English Common Law, French Civil Law, German Civil Law,
Scandinavian Civil Law and Socialist Law. The ownership details are reported for the year the firm cross-lists in the US. A firm is classified as widely held if it does not have a
controlling shareholder in that year. A firm is classified as having a controlling shareholder if it has an ultimate owner who controls at least 20% (10%) of the firm's votes. Panel A
(B) presents statistics for the 20% (10%) cutoff level. The identity of the controlling shareholder is presented using the classification scheme in La Porta et. al (1999). Panel A also
reports the mean age of the firms that cross-list. Detailed variable definitions and the sources of ownership data are presented in the Appendix.

Panel A: At the 20% cut-off level


Legal Origin                    Number of Number of                            Widely                             Identity of Controlling Shareholder
                                 countries        firms           Age           Held                                    Widely held          Widely held
                                                                                          Family           State        corporation          financial       Miscellaneous
English Common Law                  10              144            22            88              32              9                3                 8                   4
French Civil Law                    19              189            31            35              77             29               16                 4                  28
German Civil Law                     6               68            28            36              10              8                4                 1                   9
Scandinavian Law                     4               17            11             7               5              4                1                 0                   0
Socialist Law                        3               7             24             1               2              4                0                 0                   0
Full Sample                         42              425                          167            126             54               24                 13                 41

Panel B: At the 10% cut-off level

                                                                                                                  Identity of Controlling Shareholder

Legal Origin                    Number of Number of                            Widely
                                 countries        firms           Age           Held
                                                                                                                        Widely held          Widely held
                                                                                          Family           State        corporation          financial       Miscellaneous
English Common Law                  10              144            22            71              45              8                7                 10                  5
French Civil Law                    19              189            31            18              79             36                6                 16                 33
German Civil Law                     6               68            28            22              17              8                7                 1                  11
Scandinavian Law                     4               17            11             3               7              5                1                 1                   0
Socialist Law                        3               7             24             0               3              4                0                 0                   0
Full Sample                         42              425                          115            151             61               21                 28                 49




                                                                                           43

                                                                                        Table III
                                                     Cash Flow and Voting Rights of Controlling Shareholders
The table presents the mean cash flow rights, voting rights and ratio of cash flow to voting rights of all firms that had a controlling shareholder in the year in which they cross-
listed on a major exchange in the United States. A firm is classified as having a controlling shareholder if it has an ultimate owner who controls at least 20% of the firm's votes.
Panel A presents statistics for firms classified on the basis of (1) Identity of controlling shareholder: The ultimate owner could be a family (or individual), the state, a widely held
corporation, a widely held financial institution or could be miscellaneous, which includes business groups, voting trusts, foundations et. Al) (2) Legal origin: The firms are
classified according to the legal system they belong to into English common law, French civil law, Scandinavian civil law, German civil law and Socialist law. The last column
presents the difference of means test between cash flow rights and voting rights.
Panel B presents the mean ratio of cash flow to voting rights for countries classified into several categories: (1) Legal Origin: Countries are classified on the basis of English
Common Law (dummy=1), French civil law (dummy=1) and German and Scandinavian Civil Law (dummy=1) (2) Law and Enforcement : Countries are sub-classified on the
basis of (a) High vs. Low Shareholder Rights-The Anti-director rights index (scored 1-6) is an index aggregating shareholder rights and is used to distinguish between countries
that are protective of shareholders and those that are not. (b) High vs. Low Law & Order Score-Law and Order is an index scored 1-10 where higher score implies strong law and
order tradition and lower score implies weak law and order tradition. The countries are classified into high and low based on median values of the Law and Order Scores for the
countries in the sample (d) High vs. Low Judicial Efficiency-A high score implies a country with a well functional judicial system and a low score implies a poor judicial system.
(3) Financial Development: Countries are sub-classified on the basis of (a) Index of Financial Structure-Higher Values of this Index imply a more market based financial system.
(b) Large vs. Small Financial System-Overall size of the financial system is measured as deposit money bank assets and stock market capitalization as a share of GDP. The
countries are classified into high and low based on median values of the overall size variable for the countries in the sample. (c) Developed vs. Under-developed-Develop is a
dummy variable which takes the value zero if both priv (claims on private sector by deposit money banks as a share of GDP) and tvt (Total Value traded on the stock market as a
share of GDP) are below the period mean, 1 otherwise. (3) Transparency: Countries are classified on the basis of the Accounting Standards Index, scored from 0 to 90, where a
higher score indicates greater disclosure requirement in the countries.

Panel A: Cash Flow and Voting Rights

                        Category                         Cash Flow Rights (%)       Voting Rights (%)
                                                                                                                    Ratio               Test of Means
                        Full Sample                                 44.76                   50.23                    0.89                   -6.82***


                        Family                                      45.67                   49.79                    0.85                   -5.12***

     Identity of        State                                       54.64                   57.37                    0.94                   -2.68***
     controlling        Corporation                                 47.13                   50.38                    0.94                   -1.80*
    shareholder         Financial                                   21.96                   28.97                    0.83                    -1.48
                        Miscellaneous                               40.21                   45.87                    0.88                   -3.23***


                        English Common Law                          34.86                   35.78                    0.97                   -1.87*

    Legal Origin        French Civil Law                            49.55                   57.47                    0.86                   -6.76***
                        German Civil Law                            42.21                   42.26                    1.00                    -0.05
                        Scandinavian Civil Law                      42.26                   55.36                    0.70                    -2.09*
*, ** and *** represent significance at 10%, 5% and 1% level respectively



                                                                                             44

Panel B: Ratio of Cash Flow to Voting Rights

Legal Origin
                                         Dummy=0                   Dummy=1      Test of Means
English Law                                  0.87                       0.97      -2.39**
French Civil Law                             0.95                       0.86      2.95***
Scandinavian Civil Law                       0.91                       0.70       2.55**


Law and Enforcement
                                             Low                       High     Test of Means
Shareholder Rights                           0.86                       0.95      -2.82***
Law & Order                                  0.85                       0.94      -2.80***
Judicial Efficiency                          0.85                       0.94      -3.07***


Financial Development
                                             Low                       High     Test of Means
Index of Financial Structure                 0.86                       0.96      -3.39***
Overall Size                                 0.83                       0.95      -3.61***
Develop                                      0.80                       0.95      -4.77***


Transparency
                                             Low                       High     Test of Means
Accounting                                   0.86                       0.94      -2.41**
*, ** and *** represent significance at 10%, 5% and 1% level respectively




                                                                                 45

                                                                                      Table IV
                                                       Control Enhancing Features of Cross-Listing Firms
The table classifies countries into two groups in each of four categories: (1) Legal Origin: On the basis of legal origin, countries are further classified into (a) English Common
Law countries or others (b) French civil countries or others and (c) Scandinavian civil law countries or others. (2) Law and Enforcement: Countries are sub-classified into (a) High
vs Low Shareholder Rights-The Anti-director rights index (scored 1-6) is an index aggregating shareholder rights and is used to distinguish between countries that are protective of
shareholders and those that are not. (b) High vs. Low Minority Rights-Minority Rights is a dummy variable which takes the value 1 if the Company Law or Commercial Code of
the country grants minority shareholders a judicial venue to challenge management decisions, 0 otherwise. (c) High vs. Low Law & Order Score-Law and Order is an index scored
1-10 where higher score implies strong law and order tradition and lower score implies weak law and order tradition. The countries are classified into high and low based on
median values of the Law and Order Scores for the countries in the sample (d) High vs. Low Judicial Efficiency-A high score implies a country with a well functional judicial
system and a low score implies a poor judicial system. (3) Financial Development: Countries are sub-classified on the basis of (a) Market vs. Non-market based-Market is a
dummy which equals one if country has a market based financial system, 0 otherwise. (b) Large vs. Small Financial System-Overall size of the financial system is measured as
deposit money bank assets and stock market capitalization as a share of GDP. The countries are classified into high and low based on median values of the overall size variable for
the countries in the sample. (c) Developed vs. Under-developed-Develop is a dummy variable which takes the value zero if both priv (claims on private sector by deposit money
banks as a share of GDP) and tvt (Total Value traded on the stock market as a share of GDP) are below the period mean, 1 otherwise. (4) Transparency: Countries are classified on
the basis of the Accounting Standards Index, scored from 0 to 90, where a higher score indicates greater disclosure requirement in the countries. In all cases, N represents the
number of firms with controlling shareholders in the category, Cross-Shareholding represents the percentage of firms which have cross-holdings (if the company owns any amount
of shares in its controlling shareholder or in another company in its chain of control), Pyramids represents the percentage of firms which have pyramidal ownership structures
(where controlling owner exercises control through at least one publicly traded company), Single represents the percentage of firms which have only one controlling shareholder,
Management represents the percentage of firms which have the CEO, board chairman, or vice-chairman from the controlling family and Multiple Shares represents the percentage
of firms which have multiple classes of shares with different voting rights.




                                                                                  Cross                                                                     Multiple
Category                                                                N         Shareholding          Pyramid            Single        Management classes
                             English Common Law                        57                 0.04             0.47             0.21              0.42               0.05
                             Others                                    205                0.07             0.50             0.42              0.30               0.29
                             Test of Means                                               -0.96            -0.43            -3.05***           1.57             -3.87***
                             French Civil Law                          157                0.06             0.52             0.44              0.32               0.34
      Legal Origin           Others                                    105                0.07             0.47             0.28              0.34               0.09
                             Test of Means                                               -0.31             0.71            2.72***            -0.41            4.91***
                             Scandinavian Law                          10                 0.30             0.40             0.20              0.40               0.40
                             Others                                    252                0.05             0.50             0.38              0.33               0.23
                             Test of Means                                               3.22***          -0.69             -1.20             0.44               1.19



                                                                                           46

                                                                              Cross                                     Multiple
Category                                                                  N   Shareholding  Pyramid  Single  Management classes
                              Low Shareholder Rights                     152       0.07      0.61     0.43     0.32         0.30
                              High Shareholder Rights                    110       0.04      0.41     0.29     0.34         0.15
                              Test of Means                                        0.38     3.26***  2.41**    -0.23       3.01***
                              Low Law & Order Score                      134       0.04      0.57     0.37     0.33         0.37
Law and Enforcement High Law & Order Score                               128       0.08      0.42     0.38     0.33         0.10
                              Test of Means                                       -1.14      2.25**  -0.08     -0.04       5.25***
                              Low Judicial Efficiency                    129       0.05      0.53     0.46     0.30         0.37
                              High Judicial Efficiency                   133       0.07      0.47     0.29     0.35         0.11
                              Test of Means                                       -0.47      0.88    2.74***   -0.92       5.31***
                              Small Financial System                     112       0.05      0.50     0.49     0.38         0.40
                              Large Financial System                     145       0.07      0.49     0.29     0.29         0.11
                              Test of Means                                       -0.50      0.22    3.23***   1.20        5.58***
                              Under-developed                            97        0.07      0.59     0.46     0.41         0.46
Financial Development Developed                                          158       0.06      0.43     0.33     0.28         0.09
                              Test of Means                                        0.42      2.17**  1.97**    2.04**      7.23***
                              Non-Market Based                           105       0.04      0.41     0.37     0.28         0.17
                              Market based                               157       0.08      0.55     0.38     0.36         0.28
                              Test of Means                                       -1.24     -2.26**  -0.03     -1.39       -1.98**
                              Low Accounting Standards                   131       0.05      0.55     0.46     0.32         0.36
      Transparency            High Accounting Standards                  131       0.07      0.44     0.29     0.34         0.11
                              Test of Means                                       -0.46      1.88*   3.01***   -0.13       4.98***
*, ** and *** represent significance at 10%, 5% and 1% level respectively




                                                                                   47

                                                                                       Table V
                                      Changes in Voting Rights Surrounding Cross-Listing-Comparison by Country
The table shows the changes in voting rights of the largest controlling shareholder for the sample of foreign firms that cross-listed on one of the major US exchanges during the
years 1990 to 2001. Each firm's ownership structure is tracked from the year before it lists in the U.S. to five years after the cross-listing. A minimum of three years of ownership
data is used as a constraint to include a firm in the sample. VR (-2,-1) is the change in mean voting rights of the controlling shareholder in the years prior to cross-listing
year(from two years prior to on year prior), VR (-1,0) is the change in mean voting rights of the controlling shareholder from year ­1 to the year of cross-listing, VR (-1,1) is the
change in mean voting rights of the controlling shareholder from year ­1 to one year after cross-listing, VR (-1,3) is the change in mean voting rights of the controlling
shareholder from year ­1 to three years after cross-listing and VR(-1,5) is the change in mean voting rights of the controlling shareholder from year ­1 to five years after cross-
listing . Changes in voting rights are also calculated for a comparison sample for the period VR (-1,3). For each country in the sample, if the number of firms cross-listing is less
than 5, then for each firm, ownership statistics are reported for a matching firm in the same industry (four-digit SIC) and nearest asset size which does not cross-list. If the number
of firms cross-listing is greater than 5, then the comparison sample is generated for a random selection of five cross-listing firms in the country. Results are presented for all firms
at the 20% cut-off level and at the 10% cut-off level. A mean test is used to test whether the mean change in voting rights is significantly different from zero. Detailed variable
definitions and sources are given in the appendix.
Panel A: At the 20% cut-off level

Year Relative to                                                            English Common                                                          Scandinavian and
Cross-Listing Year                Full Sample            Comparison                 Law            French Civil Law German Civil Law                  Socialist Law

VR (-2,-1)                              -0.41                                       0.05                   -0.82                    -0.43                    .
VR (-1,0)                            -2.63***                                     -2.25**                -3.63***                  -1.01*                  -2.36
VR (-1,1)                            -3.92***                                     -2.42**                -5.98***                   -2.77                  -2.76
VR (-1,3)                            -5.76***                  -2.96               -2.52*                -9.38***                   -4.46                  -1.16
VR(-1,5)                             -5.15***                                       1.96                 -9.11***                   -8.87                 -2.83*
*, ** and *** represent significance at 10%, 5% and 1% level respectively


Panel B: At the 10% cut-off level
Year Relative to                                                           English Common                                      German Civil         Scandinavian and
Cross-Listing Year                Full Sample            Comparison                 Law            French Civil Law                  Law              Socialist Law
VR (-2,-1)                              -0.48                                       0.13                   -1.01                    -0.45                     .
VR (-1,0)                            -2.33***                                      -1.71*                -3.41***                  -0.81**                 -2.28
VR (-1,1)                            -3.65***                                     -1.96**                -5.72***                   -2.72*                 -2.75
VR (-1,3)                            -6.03***                  -3.45             -3.39***                -9.31***                   -4.09                   -2.2
VR(-1,5)                             -5.46***                                       1.08                 -9.03***                   -9.28*                 -2.16
*, ** and *** represent significance at 10%, 5% and 1% level respectively




                                                                                           48

                                                                                      Table VI
                              Changes in Voting Rights Surrounding Cross-Listing-Comparison by Shareholder Type
The table shows the changes in voting rights of the original controlling shareholder for the sample of foreign firms that cross-listed on one of the major US exchanges during the
years 1990 to 2001. Each firm's ownership structure is tracked from the year before it lists in the U.S. to five years after the cross-listing. A minimum of three years of ownership
data is used as a constraint to include a firm in the sample. VR (-2,-1) is the change in mean voting rights of the controlling shareholder in the years prior to cross-listing
year(from two years prior to on year prior), VR (-1,0) is the change in mean voting rights of the controlling shareholder from year ­1 to the year of cross-listing, VR (-1,1) is the
change in mean voting rights of the controlling shareholder from year ­1 to one year after cross-listing, VR (-1,3) is the change in mean voting rights of the controlling
shareholder from year ­1 to three years after cross-listing and VR(-1,5) is the change in mean voting rights of the controlling shareholder from year ­1 to five years after cross-
listing . Changes in voting rights are also calculated for a comparison sample for the period VR (-1,3). For each country in the sample, if the number of firms cross-listing is less
than 5, then for each firm, ownership statistics are reported for a matching firm in the same industry (four-digit SIC) and nearest asset size which does not cross-list. If the number
of firms cross-listing is greater than 5, then the comparison sample is generated for a random selection of five cross-listing firms in the country. Results are presented for all firms
at the 20% cut-off level and at the 10% cut-off level. A median test is used to test whether the median change in voting rights is significant. Detailed variable definitions and
sources are given in the appendix.
Panel A: At the 20% cut-off level

    Year Relative to                                                                                                     Widely held             Widely held
  Cross-Listing Year             Full Sample           Comparison               Family                State              Corporation              Financial          Miscellaneous

VR (-2,-1)                           -0.41                                       0.03                 -2.55                    1.66                  0.85                  -1.13
VR (-1,0)                           -2.63***                                    -2.60**              -7.67***                 -6.57                 -5.82                 -2.31**
VR (-1,1)                           -3.92***                                    -2.80**             -11.38***                -17.89**               -12.54                 -3.65
VR (-1,3)                           -5.76***                 -2.96             -5.78***             -17.16***               -26.74***               -14.50                -3.161
VR(-1,5)                            -5.15***                                    -3.89*              -21.99***                 -23.76                -0.68                  -6.69
*, ** and *** represent significance at 10%, 5% and 1% level respectively


Panel B: At the 10% cut-off level

    Year Relative to                                                                                                     Widely held             Widely held
  Cross-Listing Year             Full Sample           Comparison               Family                State              Corporation              Financial          Miscellaneous

VR (-2,-1)                           -0.48                                      0.003                -2.88*                    1.37                  0.62                  -0.85
VR (-1,0)                           -2.33***                                    -1.75**              -7.06***                 -4.25                 -4.31                 -1.55*
VR (-1,1)                           -3.65***                                    -2.01**             -10.47***                -16.12**               -5.31                  -2.74
VR (-1,3)                           -6.03***                 -3.45             -4.82***             -16.21***                -21.57**                7.80                  -2.36
VR(-1,5)                            -5.46***                                    -3.09*              -20.59***                 -17.92                -6.27*                 -4.81
*, ** and *** represent significance at 10%, 5% and 1% level respectively



                                                                                           49

                                                                                         Table VII
                                                Controlling Shareholder Dynamics Surrounding Cross-Listing
The table shows the changes in identity of the controlling shareholder upon cross-listing on a major US exchange during the years 1990 to 2001. Each firm's ownership structure is
tracked from the year before it lists in the U.S. to five years after the cross-listing. The original controlling shareholder could be a family (or individual), the state, a widely held
corporation, a widely held financial institution or miscellaneous, which includes business groups, voting trusts and foundations. A change of control is registered in one of the
following cases: when the original controlling shareholder divests his entire stake; when the original controlling shareholder sells his stake such that a new ultimate owner now has
maximum percentage of the voting rights; if there is no original controlling shareholder but a controlling shareholder emerges post cross-listing. The new controlling shareholder
could again be one of five types: a family (or individual), the state, a widely held corporation, a widely held financial institution or miscellaneous, which includes business groups,
voting trusts and foundations. Panel A presents the dynamics in the identity of the controlling shareholder prior to and post cross-listing. Panel B presents the percentage change in
each type of controlling shareholder. Foreign is a dummy variable identifying if the controlling shareholder of a firm belongs to a country different from the country of
incorporation of the firm. Results are presented for all firms at the 10% cut-off level. Detailed variable definitions and sources are given in the appendix.
Panel A: Controlling Shareholder Dynamics
                                                                                                               Post Cross-Listing
                                                   Total
                                                Number of                                                                                                         No Controlling
                                                 Control                                                 Widely-held        Widely-held                             Shareholder
                                                 Changes            Family                State          corporation          financial       Miscellaneous
                         Family                     26                  7                   1                   8                  4                  3                     3
  -
   oss                   State                      22                  2                   1                   9                  4                  4                     2

      Cr                 Widely-held corporation     9                  2                   0                   2                  3                  0                     2
        to               Widely-held financial       9                  1                   0                   1                  4                  0                     3
                Listing  Miscellaneous               7                  2                   1                   0                  2                  1                     1
          Prior          No Controlling
                         Shareholder                 5                  2                   1                   0                  1                  1                     0
                         Total                      78                 16                   4                  20                 18                  9                    11

Panel B: Percentage Changes in Ownership Type from prior to post cross-listing
Controlling Shareholder Percentage
Type                                    Change
Family                                    -11.7
State                                     -24.2
Corporation                               17.5
Financial                                 14.5
Miscellaneous                              3.8
Foreign                                   45.3


                                                                                              50

                                                                                        Table VIII
                                               Cross Listing versus Domestic Companies: Difference of Medians
The table reports the differences in medians of companies that cross-list and undergo a control change, relative to the control sample, and of companies that cross-list but do not
undergo a control change, relative to the control sample. The differences are calculated using quantile median regressions (LAV) where the dependent variable is regressed on a
time dummy controlling for calendar year and country effects. The time dummies used are: PRELISTING is a dummy variable that takes value 1 for observations taken 1 year
before the company cross lists, CLYEAR is a dummy capturing the timing of the cross-listing (1 in the year of cross-listing and 0 elsewhere) and POSTLISTING is a dummy
taking the value 1 in the fifth year after cross listing. Each cell in the table below represents a separate LAV regression and presents the coefficients of the time dummies, which is
the difference in medians. TOTAL ASSETS is the Total Assets of the firm, LEVERAGE is the leverage ratio of the firm, Q is Tobin's Q ratio given by (market value of assets
divided by replacement cost), TASTGR is total asset growth, FINCGR is growth in foreign income, ROA is return on assets. Detailed variable definitions and sources are given
in the appendix.



                                                          Control Change                                         No Control Change

                                         PRELISTING           CLYEAR          POSTLISTING           PRELISTING        CLYEAR         POSTLISTING

 TOTAL ASSETS                               1.652***           0.968**             0.887*             1.134***        1.067***           4.311***

 TASTGR                                      4.487**           2.785**              1.007              3.67***        5.431***            3.561*

 LEVERAGE                                    1.86**`            5.58**           3.876***              6.67**            2.86              -1.98

 FINCGR                                   14.678***          15.142***           17.245***            10.987**       11.002***           13.714**

 ROA                                          1.91              0.91*              0.45**               -0.31           0.98*              -0.05

 Q                                            6.7**             -1.3**              1.562             11.87***       13.101***           9.876***
*, ** and *** represent significance at 10%, 5% and 1% level respectively




                                                                                             51

                                                        Table IX
                         Predicting Control Change in Cross-Listing Firms
The table presents the Cox estimates of the hazard ratio of a control change occurring within five years of foreign firms
cross-listing on a major US exchange. The dependent variable takes the value one in the year of the first control change
and zero otherwise. A change of control is registered in one of the following cases: when the original controlling
shareholder divests his entire stake; when the original controlling shareholder sells his stake such that a new ultimate
owner now has maximum percentage of the voting rights; if there is no original controlling shareholder but a
controlling shareholder emerges post cross-listing. The independent variables are as follows: AGE is the age of the firm
from its year of incorporation to the year of cross-listing, SIZE is log of firm's sales, LEVERAGE is the leverage ratio
of the firm, Q is Tobin's Q ratio given by (market value of assets divided by replacement cost), TASTGR is total asset
growth, FINCGR is growth in foreign income, GDPCAP is log of GDP per capita, TVT is total value traded as a
percentage of GDP, COMMON is the legal origin dummy which takes the value 1 for common law countries and 0
otherwise, PROPERTY is the Property Rights Regulation Index from The Index of Economic Freedom.
CAPCONTROL is index of capital controls from the IMF's Trade and Exchange Restrictions Report. Lag values of all
independent variables are used. Standard errors and resulting p-values are adjusted for clustering on companies.
Detailed variable definitions and sources are given in the appendix


No. of subjects : 1284
No. of failures : 78
Log Likelihood : -262.36
Wald 2 (11)         : 47.56
Prob>2 (11)         : 0.000
                                 Hazard Ratio             Z           P>Z
AGE                                   0.955             -1.43        0.125
SIZE                                  1.195             3.56         0.000
LEVERAGE                              0.168             -1.98        0.058
Q                                     0.986             -2.45        0.000
TASTGR                                1.035             1.35         0.146
FINCGR                                1.067             6.35         0.000
GDPCAP                                1.216             2.47         0.011
TVT_GDP                              16.567             4.65         0.000
COMMON                                0.183             -2.03        0.041
PROPERTY                              0.045             -1.44        0.000
CAPCONTROL                            0.912             -2.19        0.087




                                                             52

                                                          Table X
        Predicting the Emergence of Foreign Controlling Owners on Cross-Listing
The table shows results from a multinomial logit regression that predicts the probability of a control change taking
place to a new controlling shareholder who is foreign. The dependent variable is a dummy variable that takes one of
four values: 1 (for no control change), 2 (when there is a control change to a new domestic controlling owner), 3 (when
there is a control change to a foreign controlling owner) and 4 (when the firm becomes widely held after cross-listing).
Group 1 (no control change) is the comparison group. The explanatory variables used are described as follows: Q is
Tobin's Q ratio given by (market value of assets divided by replacement cost), SIZE is log of firm's sales, LEVERAGE
is the leverage ratio of the firm, AGE is the age of the firm from its year of incorporation to the year of cross-listing,
TASTGR is total asset growth, FINCGR is growth in foreign income, GDPCAP is log of GDP per capita, TVT_GDP is
total value traded as a percentage of GDP, FDI is the level of foreign direct investment taken from IMF's Balance of
Payments Statistics, PROPERTY is Property Rights Regulation Index, YEARS_LIBER is the number of years since
the official date of liberalization of the country (from Bekaert and Harvey 1999). Lag values of all independent
variables are used. The coefficients reported are the relative risk ratios. Significance levels of the Z-statistics are also
reported. Standard errors and resulting p-values are adjusted for clustering on companies. Detailed variable definitions
and sources are given in the appendix.

Number of observations: 1987
Log Likelihood: -135.17
Wald 2 (38): 194.60
Prob>2 : 0.00
Pseudo R2 : 0.42


                                                            Relative Risk Ratios

                                        Change to a new Change to a new
                                        controlling owner controlling owner Change to widely
                                           (domestic)              (foreign)               held
Q                                      0.999                 0.999                 1.002*
SIZE                                   1.046*                0.588**               0.025**
LEVERAGE                               0.873                 0.295***              1.022
AGE                                    0.992                 0.982                 0.295**
TASTGR                                 0.998                 0.999                 0.984
FINCGR                                 1.003                 1.151**               1.002
GDPCAP                                 3.321***              2.518                 1.381
TVT_GDP                                13.685**              0.003***              0.013
PROPERTY                               0.726***              1.101                 1.587***
FDI                                    1.121                 1.828***              1.231
YEARS_LIBER                            0.903*                0.884                 0.948
*, ** and *** represent significance at 10%, 5% and 1% level respectively




                                                               53

                                                         Table XI
              Effect of FDI Restrictions on Control Change in OECD Countries
The table shows results from a multinomial logit regression that predicts the effect of restrictions on foreign direct
investment on control change in OECD countries. The dependent variable is a dummy variable that takes one of four
values: 1 (for no control change), 2 (when there is a control change to a new domestic controlling owner), 3 (when
there is a control change to a foreign controlling owner) and 4 (when the firm becomes widely held after cross-listing).
Group 1 (no control change) is the comparison group. The explanatory variables used are described as follows: Q is
Tobin's Q ratio given by (market value of assets divided by replacement cost), SIZE is log of firm's sales, LEVERAGE
is the leverage ratio of the firm, AGE is the age of the firm from its year of incorporation to the year of cross-listing,
TASTGR is total asset growth, FINCGR is growth in foreign income, GDPCAP is log of GDP per capita, TVT_GDP is
total value traded as a percentage of GDP, FDI is the level of foreign direct investment taken from IMF's Balance of
Payments Statistics, YEARS_LIBER is the number of years since the official date of liberalization of the country (from
Bekaert and Harvey 1999), RESTRICTION is a composite OECD measure of restrictions on foreign direct investment
which includes restrictions on foreign equity ownership (from Golub 2003). Lag values of all independent variables are
used. The coefficients reported are the relative risk ratios. Significance levels of the Z-statistics are reported. Standard
errors and resulting p-values are adjusted for clustering on companies. Detailed variable definitions and sources are
given in the appendix.

Number of observations:
Log Likelihood: -67.349
Wald 2 (20): 75.39
Prob>2 : 0.00
Pseudo R2 : 0.36
                                               Relative Risk Ratios
                                   Change to a new              Change to a new
                                  controlling owner             controlling owner
                                        (domestic)                   (foreign)

Q                              0.999**                     0.999
SIZE                           1.389***                    0.531
LEVERAGE                       0.876                       0.583
AGE                            1.009                       0.966
TASTGR                         0.978                       0.998
FINCGR                         1.001                       1.063*
GDPCAP                         2.629***                    1.09
TVT_GDP                        18.409***                   1.335
PROPERTY                       0.586***                    0.726
FDI                            1.625*                      1.873**
YEARS_LIBER                    0.459***                    0.705
RESTRICTION                    5.303                       0.061**
*, ** and *** represent significance at 10%, 5% and 1% level respectively




                                                               54

                                                                                     Table XII
                                Ex-Post Regressions Distinguishing between Firms with and without Control Changes
The table shows the estimates of the ex-post effects of cross-listing distinguishing between firms that underwent a control change and those that did not experience a control
change. Panel A presents results for companies that did not have any control changes and Panel B presents results for companies that underwent a control change. Each column
gives the results of a LAV regression for a dependent variable. The explanatory variables are CLYEAR which is a dummy capturing the timing of the cross-listing (1 in the year of
cross-listing and 0 elsewhere), YEAR3 which is a dummy taking the value 1 in the three years after cross-listing and YEAR5 which is a dummy taking the value 1 after the third
year). A constant, country on incorporation and calendar year dummies are also included but the coefficients are not reported here. First differences of all variables are taken to
eliminate fixed effects. Detailed variable definitions and sources are given in the appendix.

Panel A: No Control Change

                                           1                      2                  3                    4                    5                      6

                                    Total Assets          Total Assets      Foreign Income
                                                              Growth             Growth              Leverage         Return on Assets           Tobin's Q

Year of Cross-Listing                    0.045               6.790***            0.225**             -0.090***                0.53                 0.987
Three-year effect                       -0.016                  0.06             0.312**                0.002               -0.33***               1.023*
Five-year effect                         0.014                 -2.35             0.556**                -0.025               -0.12                2.194**
Pseudo R-squared                         0.00                   0.02               0.00                  0.05                 0.00                  0.00
N                                        1567                   1470               1362                 1293                 1436                   1171
*, ** and *** represent significance at 10%, 5% and 1% level respectively


Panel B: Control Change
                                           1                      2                  3                    4                    5                      6

                                    Total Assets          Total Assets      Foreign Income
                                                              Growth             Growth              Leverage         Return on Assets           Tobin's Q

Year of Cross-Listing                    0.035               7.880***            2.178**                -0.086                0.42                 0.235
Three-year effect                        0.057                  0.16             1.222**               1.003*                 1.07                 -1.246
Five-year effect                      -0.236**                  1.14             2.289***            2.669***                0.580*              -1.632***
Pseudo R-squared                         0.01                   0.00               0.00                  0.04                 0.01                  0.00
N                                        412                    351                 368                  406                  365                   325
*, ** and *** represent significance at 10%, 5% and 1% level respectively




                                                                                         55

                                                                    Appendix A
                                                              Variable Definitions

Variable                 Definition


Ownership Variables
Widely held              Equals 1 if there is no controlling shareholder. Control is measured following LLSV (???)
Family                   Equals 1 if a person or family is controlling shareholder, and zero otherwise
State                    Equals 1 if domestic or foreign state is controlling shareholder, and zero otherwise
Widely-held corporation  Equals 1 if a widely held non-financial company is controlling shareholder, and zero otherwise
Widely-held financial    Equals 1 if a widely held financial company is controlling shareholder, and zero otherwise
Miscellaneous            Equals 1 if above five categories are zero. Includes pension funds, voting trusts, groups, non-profit organizations and
                         employees
Cross-Shs                Equals 1 if firm has a controlling shareholder and owns shares in its controlling shareholder or in firm that belongs to
                         the chain of control, zero otherwise
Pyramid                  Equals 1 if controlling shareholder exercises control through at least one publicly traded company, and zero otherwise


Single                   Equals 1 if firm has a 20% (or 10%) controlling shareholder and no other shareholder has control of at least 10 percent
                         of the votes
Management               Equals 1 if controlling shareholder is also CEO, Chairman, Vice-Chairman of the Board, and zero otherwise



Firm-level Variables
AGE                      Time since the date of incorporation of the company
SIZE                     Log of Sales. Source: Worldscope
LEVERAGE                 Total Assets/ Shareholder's Equity. Source: Worldscope
TASTGR                   Total Asset Growth. Source: Worldscope
FINCR                    Foreign Income Growth. Source: Worldscope
Q                        Tobins Q Ratio ((Market Value of Equity+Long Term Debt)/Replacement Value(Net property, plant &
                         equipment+Inventory). Source: Worldscope
ROA                      Return on Assets. Source: Worldscope




                                                                          56

Variable                 Definition
Country-level Variables
GDPGROWTH                Growth in GDP per capita. Source: World Development Indicators
TVT_GDP                  Total Value of Stocks Traded as a percentage of GDP. Source: World Bank Financial Structure Database
FDI                      Foreign Direct Investment. Source: IMF Balance of Payments Statistics
YEARS_LIBER              Time elapsed since the official liberalization date of the country. The official liberalization dates are reported in
PROPERTY                 The degree to which property rights are protected in the economy. Source: Heritage Foundation
RESTRICTION              OECD Measure of FDI restriction for each country. The total score ranges from 0 to 1 and is based on regulations in
                         each of three areas: (1) Restrictions on foreign ownership (2) Obligatory screening and approval procedures used to
                         limit FDI (3) Other formal restrictions including constraints on the ability of foreign nationals to manage or work in
                         affiliates of foreign companies and other operational controls on these businesses. Source: Golub (2003)

CAPITAL CONTROLS         Index of Capital Controls. Source: IMF Trade and Exchange Restrictions Report;Harrison, Love and McMillan(2002)
ACCOUNTING               Index created by examining and rating companies' 1990 annual reports on their inclusion or omission of 90 items in
                         balance sheets and income statements and published by the Center for International Financial Analysis & Research Inc.
                         the maximum is 90, the minimum is 0.
SHAREHOLDER RIGHTS       An index aggregating the shareholder rights. The index is formed by adding 1 if : (1) the country allows shareholders to
                         mail their proxy to the firm; (2) shareholders are not required to deposit their shares prior to the General Shareholders'
                         Meeting; (3) cumulative voting or proportional representation of minorities in the board of directors is allowed; (4) an
                         oppressed minorities mechanism is in place; (5) the minimum percentage of share capital that entitles a shareholder to
                         call for an Extraordinary Shareholders' Metting ie less than or equal to 10 percent; or (6) shareholders have preemptive
                         rights that can only be waived by a shareholder's vote. Source: LLSV 1999

RULE OF LAW              Measure of the law and order tradition of a country. Ranges from 1(weak) to 10(strong). Source: International Country
                         Risk Guide
JUDICIAL EFFICIENCY      Assessment of the efficiency and integrity of the legal environment as it affects business. Scale from 0-10. Source:
                         LLSV 1999
OVERALL SIZE             Overall size of the financial system (Deposit money bank assets and stock market capitalization as share of GDP).
                         Source: World Bank Financial Structure Database
MARKET                   Index of Financial Structure. Higher values indicate a more market based system. Source: Beck et. al(2000)
DEVELOPED                Equals zero if both priv(claims on private sector by deposit money banks as a share of GDP) and tvt_gdp are below the
                         period mean




                                                                          57

                                                      Appendix B
                                            Sources of Ownership Data

Country                     Years        Sources
All Countries               1990-2002    Worldscope, Hoovers Online, Lexis-Nexis, ISI Emerging Markets Database


Country Specific Sources
Argentina                   1995         Major Companies of Latin America and the Caribbean 1997, Graham and Whiteside Publication
Argentina                                Argentina Company Handbook 1995-1996
Argentina                                Emerging Markets Handbook 1997
Australia                                Asia Pacific Handbook 1999, Primark Publication
Australia                                Major Companies of Far East and Australasia 2002, Graham & Whiteside Gale Group
Australia                                Jobson's Year Book of Public Companies 99-00
Austria                     1998, 1999   European Handbook 2000, Primark Publication
Austria                                  Major Companies of Europe 2002
Austria                                  www.transnationale.org
Belgium                                  Major Companies of Europe 2002
Brazil                      1995         Major Companies of Latin America and the Caribbean 1997, Graham and Whiteside Publication
Brazil                                   Brazil Company Handbook 97/98, 01/02
Chile                       1995         Major Companies of Latin America and the Caribbean 1997, Graham and Whiteside Publication
Colombia                    1995         Major Companies of Latin America and the Caribbean 1997, Graham and Whiteside Publication
Denmark                     1998, 1999   European Handbook 2000, Primark Publication
Denmark                                  Major Companies of Europe 2002
Dominican Republic          1995         Major Companies of Latin America and the Caribbean 1997, Graham and Whiteside Publication
Finland                     1998, 1999   European Handbook 2000, Primark Publication
Finland                                  Major Companies of Europe 2002
Finland                                  Guide to Nordic Stock Markets
Finland                                  Europe's Top Quoted Companies
France                      1998, 1999   European Handbook 2000, Primark Publication
France                                   Major Companies of Europe 2002
France                                   Europe's Top Quoted Companies



                                                            58

Country        Years        Sources
Germany        1998, 1999   European Handbook 2000, Primark Publication
Germany                     Major Companies of Europe 2002
Germany                     Europe's Top Quoted Companies
Greece                      Major Companies of Europe 2002
Greece                      Europe's Top Quoted Companies
Hong Kong      1996, 1997   Asian Company Handbook 1999, Toyo Keizai Publication
Hong Kong                   Asia Pacific Handbook 1999, Primark Publication
Hong Kong                   Major Companies of Far East and Australasia 2002, Graham & Whiteside Gale Group
India          1996, 1997   Asian Company Handbook 1999, Toyo Keizai Publication
India
Indonesia      1996, 1997   Asian Company Handbook 1999, Toyo Keizai Publication
Indonesia                   Major Companies of Far East and Australasia 2002, Graham & Whiteside Gale Group
Indonesia                   Emerging Markets Handbook 1997
Ireland                     Europe's Top Quoted Companies
Israel                      Major Companies of Europe 2002
Italy                       Europe's Top Quoted Companies
Italy                       Major Companies of Europe 2002
Japan                       Asia Pacific Handbook 1999, Primark Publication
Japan                       Morningstar Japan, July 1992, 1994
Japan                       Major Companies of Far East and Australasia 2002, Graham & Whiteside Gale Group
Luxembourg                  Major Companies of Europe 2002
Malaysia                    Asia Pacific Handbook 1999, Primark Publication
Malaysia                    Kuala Lumpur Stock Exchange Annual Companies Handbook 1996
Mexico         1995         Major Companies of Latin America and the Caribbean 1997, Graham and Whiteside Publication
Mexico                      Mexico Company Handbook 1995-96
Netherlands                 Europe's Top Quoted Companies
Netherlands                 Dutch Company Handbook 1998-99
New Zealand                 Major Companies of Far East and Australasia 2002, Graham & Whiteside Gale Group
New Zealand                 Jobson's Year Book of Public Companies 99-00
Norway         1998, 1999   European Handbook 2000, Primark Publication



                                               59

Country         Years            Sources
Norway                           Europe's Top Quoted Companies
Norway                           Guide to Nordic Stock Markets
Peru            1995             Major Companies of Latin America and the Caribbean 1997, Graham and Whiteside Publication
Philippines     1996, 1997      Asian Company Handbook 1999, Toyo Keizai Publication
Philippines                      Major Companies of Far East and Australasia 2002, Graham & Whiteside Gale Group
Portugal                         Major Companies of Europe 2002
Portugal                        Portuguese Stock Exchange 1993
Portugal                         Europe's Top Quoted Companies
Portugal                        Emerging Markets Handbook 1997
Russia                           Russian Equity Guide 96-97, Brunswick Brokerage'
Russia                          Russian Equity Guide 2000-01
Singapore       1996, 1997      Asian Company Handbook 1999, Toyo Keizai Publication
Singapore                        Asia Pacific Handbook 1999, Primark Publication
Singapore                        Major Companies of Far East and Australasia 2002, Graham & Whiteside Gale Group
South Africa    1996, 1997, 1998 McGregors's Who Owns Whom in South Africa, 1998 Listed and Unlisted Companies, McGregor Publication
South Africa    1994, 1995, 1996 McGregors's Who Owns Whom in South Africa, 1996 Listed and Unlisted Companies, McGregor Publication
South Korea     1996, 1997      Asian Company Handbook 1999, Toyo Keizai Publication
South Korea                      Major Companies of Far East and Australasia 2002, Graham & Whiteside Gale Group
South Korea                      Korea Company Handbook Investment Guide, Daewoo Securities Co., Ltd.
South Korea                     Emerging Markets Handbook 1997
South Korea                     Korea Company Handbook 2000
South Korea                      Asia Pacific Handbook 1999, Primark Publication
Spain           1998, 1999      European Handbook 2000, Primark Publication
Spain                            Major Companies of Europe 2002
Spain                            Europe's Top Quoted Companies
Sweden          1998, 1999       European Handbook 2000, Primark Publication
Sweden                           Europe's Top Quoted Companies
Sweden                           Major Companies of Europe 2002
Switzerland     1998, 1999      European Handbook 2000, Primark Publication
Switzerland                     Major Companies of Europe 2002



                                                    60

Country        Years        Sources
Switzerland                 Europe's Top Quoted Companies
Taiwan         1996, 1997   Asian Company Handbook 1999, Toyo Keizai Publication
Taiwan                      Major Companies of Far East and Australasia 2002, Graham & Whiteside Gale Group
Taiwan                      Emerging Markets Handbook 1997
Thailand                    Asia Pacific Handbook 1999, Primark Publication
Thailand
UK             1998, 1999   European Handbook 2000, Primark Publication
UK                          Europe's Top Quoted Companies
UK                          Major Companies of Europe 2002
Venezuela                   Venezuela Company Handbook 1992-93




                                              61

                                                                               Appendix C
                                                                       Country Characteristics

GDP per capita is the real GDP per capita averaged over the period 1990-1999. TVT_GDP is Total Value of Stocks Traded as a percentage of GDP averaged over 1990-99.
COMMON is a dummy variable which takes the value1 for English common law countries and 0 for others. PROPERTY is the Property Rights Regulation Index from The Index
of Economic Freedom. CAPCONTROL is index of capital controls from the IMF's Trade and Exchange Restrictions Report, averaged over 1990-96. FDI is the level of foreign
direct investment taken from IMF's Balance of Payments Statistics, averaged over 1990-99. RESTRICTION is a composite OECD measure of restrictions on foreign direct
investment which includes restrictions on foreign equity ownership (from Golub 2003).


Nation                     GDP/CAP                                                                    CAPITAL                             (OECD) FDI
                           (US $)                 TVT_GDP              COMMON         PROPERTY        CONTROL          LEVEL OF FDI       RESTRICTION

Argentina                         7440.86                 0.04                0              4              3.00               2.59                  .
Australia                        20498.74                 0.29                1              5              0.08               1.82                0.33
Austria                          29344.73                 0.07                0              5              0.55               1.07                0.43
Belgium                          27304.92                 0.09                0              5              0.40               2.20                0.29
Brazil                            4298.29                 0.13                0              3              2.43               1.58                  .
Chile                             4389.20                 0.08                0              5              2.75               5.19                  .
Colombia                          2289.70                 0.01                0              3              3.89               2.14                  .
Denmark                          34167.82                 0.20                0              5              0.40               2.47                0.16
Dominican Republic                1588.07                   .                 0              2                 .               2.89                  .
Finland                          26289.28                 0.23                0              5              0.50               1.96                0.46
France                           26986.50                 0.23                0              4              1.40               1.66                0.23
Germany                          30004.38                 0.27                0              5              0.08               0.58                0.17
Ghana                              373.57                 0.00                1              3                 .               1.34                  .
Greece                           11455.06                 0.24                0              4                 .               0.96                0.33
Hong Kong, China                 21994.46                 1.15                1              5              0.00                                     .
Hungary                           4526.60                 0.10                0              4                 .               4.65                  .
India                              375.75                 0.08                1              3              3.50               0.39                  .
Indonesia                          966.93                 0.09                0              3              1.00               1.06                  .
Ireland                          18767.87                 0.28                1              5              1.30               5.22                0.25



                                                                                     62

Nation              GDP/CAP                                    CAPITAL               (OECD) FDI
                    (US $)          TVT_GDP   COMMON  PROPERTY CONTROL   LEVEL OF FDIRESTRICTION

Israel                    15343.58      0.18     1        4        2.00      1.24             .
Italy                     19050.73      0.14     0        4        1.40      0.35           0.26
Japan                     42285.18      0.29     0        5        0.09      0.06           0.24
Korea, Rep.               10261.01      0.58     0        5        1.89      0.67             .
Luxembourg                44015.98      0.03     0        5          .                        .
Mexico                    3346.84       0.12     0        3        2.08      2.21             .
Netherlands               27063.04      0.58     0        5        0.00      4.12           0.24
New Zealand               16165.58      0.13     1        5        0.00      4.05           0.24
Norway                    33227.32      0.18     0        5        1.50      1.74           0.47
Peru                      2121.17       0.04     0        3        2.43      3.07             .
Philippines               1090.42       0.16     0        4        3.00      1.72             .
Poland                    2880.74       0.04     0        4          .       2.39             .
Portugal                  10980.78      0.13     0        4        1.38      2.02           0.22
Russian Federation        2630.33       0.02     0        3          .       0.53             .
Singapore                 22510.88      0.75     1        5        0.00      10.44            .
South Africa              3921.59       0.21     1        3        4.00      0.96             .
Spain                     15025.36      0.42     0        4        1.54      1.89           0.23
Sweden                    27400.20      0.44     0        4        0.70      5.35           0.34
Switzerland               44520.43      1.22     0        5          .       2.04           0.28
Taiwan, China             12141.17      2.62     0         .         .                        .
Turkey                    2782.34       0.21     0        4        3.63      0.46           0.39
United Kingdom            19360.10      0.60     1        5        0.00      2.77           0.17
Venezuela                 3519.88       0.03     0        3        2.44      2.60             .




                                                     63

                                   Appendix D
                      Example of change in ownership structure

VIMPEL COMMUNICATIONS
Date of incorporation: 1992
ADR Listing: 11/01/1996
Listed on NYSE as Level III ADR
Depositary: Bank of New York

Ownership Structure prior to ADR Listing (Dec 1995) i.e. at t-1


                            VIMPEL
                   COMMUNICATIONS

      20.1 O,          4.7 O               20.2 O,
       27.5 V          3.5 V               27.6 V

     KB Impuls-TV                 SOTA-100


    100 O, V                             100 O,V


                    Dmitri B.
                    Zimin




Ownership Structure post ADR Listing (Dec 2001) i.e. at t+5


                           VIMPEL
                   COMMUNICATIONS

    28.98 O,                                 13.05 O,
     25 V                                     25 V

     Telenor East                 Eco Telecom
     Invest AS                    Ltd.

       77.66
       O,V


       Norwegian                  Alfa Group
       Government




                                        64