Kim 2016

G Model
MULFIN-519; No. of Pages 20 ARTICLE IN PRESS

J. of Multi. Fin. Manag. xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Journal of Multinational Financial

Management
journal homepage: www.elsevier.com/locate/econbase
The joint role of the bonding mechanisms and the reduction

in market segmentation in valuation of firms cross-listed as
Global Depositary Receipts (GDRs)
Oksana Kim
Minnesota State University, Department of Accounting and Business Law, MH 227, Mankato 56001, United States
a r t i c l e i n f o a b s t r a c t
Article history: Using a sample of GDRs cross-listed in London, we revisit the debate regarding the validity
Received 6 January 2016 of the market segmentation and the bonding hypotheses for a cross-listing phenomenon.
Received in revised form Unlike prior studies that relied on emerging/developed market partitioning of countries,
19 November 2016
we use equity trading costs to determine the degree of market segmentation, which is a
Accepted 16 December 2016
more direct and less noisy measure of this construct. Additionally, there is little correla-
Available online xxx
tion between this metric and the level of home-country investor protection for examined
GDRs, therefore providing stronger settings to distinguish between the segmentation and
JEL classification:
bonding explanations. We find that legal bonding mechanisms and reduction in segmen-
G15
G32 tation have a positive impact on changes in firm value upon cross-listing, when examined
as standalone frameworks. Next, we report that these two frameworks have a joint, com-
Keywords: plementary impact on changes in firm value upon cross-listing, based on the country-level
GDR Rule of Law metric. Conversely, we find that the positive association between the capital
Market segmentation
raising activity and changes in firm value upon cross-listing is less significant for countries
Bonding
from the most segmented markets. Finally, we find that analysts following [accuracy] is an
Firm value
Cross-listing
effective reputational bonding mechanism for firms from the most [least] segmented mar-
kets primarily after cross-listing. This study sheds light on the complexity of the interplay
between major valuation theories and different types of bonding mechanisms in the case
of cross-listing.
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
Cross-listing on an overseas market with a stringent regulatory environment confers significant benefits, such as increased
liquidity, reduced cost of capital, and enhanced market valuation (e.g., Doidge et al., 2004, 2009; Miller 1999). On the other
hand, cross-listing is associated with increased reporting and compliance costs that managers often cite as the major fac-
tor precluding eligible firms from pursuing an overseas listing (Karolyi 2012). Despite the abundance of empirical studies
examining international listings, the source of valuation benefits associated with cross-listing is not well understood. Var-
ious theories offered potential explanations for documented benefits (or lack thereof), the major of which are the market
segmentation and the bonding hypotheses. The former conventional wisdom framework suggests that investors bear direct
costs associated with the constraints imposed by the regulatory environment, such as foreign ownership restrictions, trad-
ing costs and taxes, and indirect monitoring costs due to lack of reporting transparency. Cross-listing allows overcoming
E-mail address: Oksana.kim@mnsu.edu
http://dx.doi.org/10.1016/j.mulfin.2016.12.003
1042-444X/© 2016 Elsevier B.V. All rights reserved.
Please cite this article in press as: Kim, O., The joint role of the bonding mechanisms and the reduction in market seg-
mentation in valuation of firms cross-listed as Global Depositary Receipts (GDRs). J. Multinatl. Financial Manage. (2016),
G Model
2 O. Kim / J. of Multi. Fin. Manag. xxx (2016) xxx–xxx
the negative impact of market segmentation by minimizing these costs for foreign investors (Alexander et al., 1987). The
competing bonding hypothesis posits that firms corporate governance attributes and the investor protection level are the
major causes of cross-listing effects (Coffee 1999). While the volumes of international listings have been steadily increasing,
there is no credible explanation to date regarding the major drivers of economic consequences of cross-listing.
The lack of agreement in the empirical literature regarding the legitimacy of the two theories provides a strong moti-
vation for this study. We revisit this debate in alternative cross-listing settings by examining valuation benefits for Global
Depositary Receipts (GDRs), which provides stronger experimental settings to assess the relevance of the two frameworks
to the cross-listing phenomenon. First, we document that the firm value, as proxied by Cumulative Abnormal Returns (CARs)
and measured over a 49-month window, increases as a result of GDR cross-listing. Second, we report that GDRs from the
most segmented markets experience greater changes in firm value upon cross-listing, which is in support of the market
segmentation hypothesis. Third, we show that GDRs from markets with low levels of investor protection (country-level
legal bonding metrics) and those that raise capital (firm-level legal bonding proxy) experience greater changes in firm value
after cross-listing, therefore supporting the legal bonding theory. Next, we report that the legal bonding mechanism and
the reduction in market segmentation play a complementary positive impact on changes in firm value upon cross-listing.
Conversely, the positive association between the capital raising activity and changes in firm value is less significant for firms
from the most segmented markets. Finally, reputational bonding proxies − analyst accuracy and following − have a markedly
different impact on firm value, compared to the legal bonding metrics. Particularly, we find the analysts following [accuracy]
and the reduction in market segmentation have a complementary positive impact on changes in firm value for firms from
the most [least] segmented markets primarily after cross-listing, suggesting that the extent of reputational bonding is more
limited than often anticipated.
The market segmentation theory found mixed support in the empirical literature that focused predominantly on American
Depositary Receipts (ADRs). For instance, abnormal returns around a cross-listing event should vary across the share stocks in
accordance with the differences in the degree of market segmentation between a local market and the US, while prior studies
do not confirm this (Foerster and Karolyi, 1999). In addition, the magnitude of the cost of capital decline documented in those
studies is unlikely to be solely attributable to the elimination of the barriers (Errunza and Miller, 2000). The segmentation
of markets represents an explicit barrier to foreign investment (Kang and Stulz, 1997). While those barriers were gradually
decreasing during the 1990s, there should have been a reduction in a number of overseas listings observed due to diminishing
net benefits (Stulz, 1999), while the volume of international listings increased over the past two decades. Researchers
put emphasis on other (implicit) barriers and turned to alternative frameworks that would help to explain cross-listing
consequences. As argued by Kang and Stulz (1997), the major implicit barriers are political risk and information asymmetries.
The latter offered a more promising explanation, and the bonding hypothesis replaced the market segmentation theory as a
major framework.
The bonding hypothesis originated in studies of Coffee (1999) and Stulz (1999). They suggested that by cross-listing in
the US, firms effectively “bond” themselves to the stringent governance and compliance requirements, such as increased
disclosure and high risk of litigation, of overseas markets that would put binding constraints on private actions of managers.
Due to this extended commitment, firms anticipate enhanced market valuation and reduced cost of capital. This alternative
framework, nevertheless, also found mixed support. Licht (1998, 2000) reported that the Securities and Exchange Commis-
sion (SEC) rarely enforced actions against foreign issuers and the effectiveness of the bonding mechanisms was questionable
for cross-listed firms. In line with this argument, Siegel (2005) documented that Mexican ADRs, were likely to take advantage
of the relatively lax US enforcement of the regulations for foreign issuers.
We propose that the main reason for the mixed findings of prior literature regarding the validity of the market segmenta-
tion hypothesis was due to reliance on the noisy measures of the segmentation construct that were grounded in partitioning
of countries based on levels of economic development. The main assumption in those studies was that emerging markets
were more segmented compared to developed markets, and hence capital markets effects were expected to be more pro-
nounced for firms from the former group of countries upon cross-listing (Foerster and Karolyi, 1999; Miller 1999). We argue
that the variation in the level of segmentation barriers between the emerging/developed markets groups can be as great
as it is within those groups, and this could potentially confound the findings and the interpretation of the results in prior
studies. According to the recent Association of Chartered Certified Accountants (ACCA) report (2012), within the emerg-
ing markets group frontier markets lag behind other emerging markets on many dimensions much more than the latter
lag behind developed markets. As a result of capital markets integration, some emerging markets occupy more advanced
positions than developed ones. For example, South Korea that recently transitioned from the emerging into the developed
market category offers the longest account opening timeframe and higher withholding taxes for foreign investors compared
to emerging markets, such as China or India.
In this study, we rely on trading costs that foreign investors would bear when investing in local stocks, which represents
an explicit investment barrier and is therefore a more direct measure of the degree of market segmentation, compared to a
simple emerging/developed market partitioning (Miller, 1999). Next, unlike prior studies that almost exclusively examined
ADRs listed on the US exchanges, our study focuses on the alternative cross-listing instruments of GDRs that originated in
the 1990s and for which the main cross-listing platform is the London Stock Exchange. The main advantage that the GDR
setting provides is that firms from emerging markets pursue these programs exclusively, unlike ADRs that represent a mix
of emerging and developed markets firms. In our study, all examined cross-listed firms represent emerging markets that
went through relatively similar stages of economic development and market reforms. Therefore, the issue that the economic
G Model
O. Kim / J. of Multi. Fin. Manag. xxx (2016) xxx–xxx 3
consequences of cross-listing may vary with the propensity of home-country institutions to adopt differential capital market
policies (see Leuz, 2006) is less of concern.
Next, the findings of prior ADRs-based studies that built on the bonding framework cannot be generalized to GDRs a
priori. GDRs are fundamentally different from ADRs in terms of the listing requirements and obligations. These are consid-
erably lower for GDRs when compared to ADRs, and GDRs attracted much criticism from market participants because of the
multiple reporting and corporate governance exemptions under their “light touch” regime. Accordingly, it is unclear if the
provisions of the bonding framework would indeed hold for GDR cross-listed firms due to the relaxed reporting require-
ments, which provides a valuable opportunity for empirical investigation of the validity of the bonding framework for these
alternative cross-listing instruments. Further, prior studies adopting the legal bonding framework to explain the cross-listing
consequences partitioned countries based on the investor/shareholder right variables (e.g., La Porta et al., 1998, 2000; Burns
et al., 2007) for which the civil/common law classification became pervasive. These investor protection constructs, never-
theless, are one reason for the market segmentation, which makes it difficult to attribute the documented results in the
prior studies to the validity of either the bonding or the market segmentation frameworks. Besides, the distinction between
common and civil law counties has become blurred, and the latter often adopt more progressive policies, despite legal
system restrictions. For instance, foreign investors would face greater account opening and settlement delays and limited
exchangeability between Depositary Receipts and underlying shares in the case of the Indian market compared to the Rus-
sian market, although the former is arguably more integrated with the developed market of the UK due to strong historical
ties and common law traditions. Our study addresses these limitations and we attempt to capture the market segmentation
using the trading cost metric that has little correlation with the investor protection and legal system constructs.1
Our main contribution to the literature, nevertheless, is not only in revisiting the cross-listing framework debate in novel
GDR settings using a more explicit measure of market segmentation (trading costs), but also in extending the previously
described analyses. An important feature of the debate regarding the validity of the two frameworks was that they were
often considered competing explanations for a cross-listing phenomenon, whereas we examine whether the two theories
have a joint, complementary or substitutive, role in firm valuation upon cross-listing. To our knowledge, this is the first study
to undertake such an investigation. Our study provides important insight into the complexity of an interplay between (i)
theories attempting to explain economic consequences of cross-listing and (ii) different types of bonding mechanisms.
The study proceeds as follows. Section 2 provides background information regarding GDRs and discusses related literature.
Section 3 is dedicated to research design, and Section 4 describes the data collection process. Section 5 reports the empirical
results, and Section 6 concludes the study.
2. Institutional details and related literature
2.1. GDRs compared to previously examined ADRs, and criticism of GDRs
The structure and mechanisms for cross-listing via ADRs (see Miller 1999) and GDRs (see Kim and Pinnuck 2014) are
similar. There are Level I over-the-counter traded ADRs and GDRs and there are Level II and III (capital-raising) exchange-
listed programs. Level I programs are subject to minimal reporting obligations, whereas Level II and III ADRs and GDRs must
follow strict reporting requirements. Nevertheless, the major difference between ADRs and GDRs is in the scope of pre-
listing disclosure requirements and post-listing reporting obligations. Particularly, the UK Listing Authority (UKLA) (2015)
established disclosure and reporting requirements for GDRs that are not as excessive as those for exchange-listed ADRs,
which allowed firms to cross-list on the LSE at lower compliance costs. As a result, GDRs became an attractive listing tool for
emerging market firms, unlike ADRs that are also prevalent among firms from developed markets. These minimal disclosure
requirements for GDRs collectively known as the “light touch” regulation initially met strong support among firms from
emerging markets and global investor groups. Appendix A lists comparative pre- and post-listing obligations for ADRs and
GDRs and shows that on average, GDRs report less information at lower frequency and are officially exempt from protective
governance regulations. Further, listing in the UK can be done directly or via GDRs, which is the focus of this study.2 GDRs,
as a rule, are listed on the Main Market of the LSE, whereas the Alternative Investment Market (AIM) and the Professional
Securities Market (PSM) do not provide legal and trading support for these instruments and are therefore not considered in
this study.3
Despite the increased popularity of GDRs, market participants were concerned about insufficient investor protection
for GDRs compared to ADRs as a result of minimal reporting obligations, as well as the “comply or explain” approach
1
In fact, the correlation between the investor protection and trading cost metrics in our study is negative 37 percent.
2
Direct listings are primarily done by the local UK firms, firms from developed markets, and firms incorporated in several tax haven territories such as
Ireland, Cyprus, BV Islands, etc. GDRs, on the other hand, are a convenient cross-listing vehicle for foreign issuers primarily from emerging markets. None
of GDRs are simultaneously listed as ADRs.
3
AIM and PSM-listed firms (unregulated segments) do not have to adhere to stringent reporting rules and can report under local standards, unlike
GDRs that have to comply with IFRS, which is why we do not examine them. We also omit details regarding history and development of ADRs and GDRs
over time, as prior studies cover those extensively (see Karolyi 2006; 2012). We do not examine change in firm value for ADRs because they received
significant coverage in prior cross-listing literature, unlike their GDR peers. Kim and Pinnuck (2014) provide a detailed discussion of history, development,
and requirements for London-listed GDRs that we do not repeat here, as this is not the main objective of the study.
G Model
that allowed GDR issuers to avoid litigations. In 2006, the year in which GDRs proved to be remarkably popular and took
considerable market share from ADRs, the SEC Chairman, Christopher Fox, proposed enhanced cooperation between the US
and UK regulatory regimes in order to ensure greater investor protection and create healthy international capital markets
(SEC, 2006). Chairman Fox made the following statement: “Some of the experiments in regulation that we have witnessed
around the world seem to have worked. Others have failed. So long as our experiments are aimed at providing high-quality
investor protection, we all stand to gain. But if our experiments are guided by the desire to beggar our neighbors, we will
all surely lose”. Yet the UKLA stood firm on its position that it fully achieves its consumer protection goal through principle-
based regulations that “provide firms with greater flexibility to decide for themselves how best to run their businesses,
while meeting [our] regulatory objective” (UKLA, 2015). Besides, GDRs have to comply with IFRS, which is considered a high
quality set of accounting standards.
In 2008, Deloitte reported that one-third of 2007 GDR issuers highlighted financial control weaknesses in the pre-listing
documents. This was not viewed as a deficiency and all issuers were admitted to trade (Deloitte, 2008). Deloitte issued a notice
against the general misconception about exchange-listed GDRs that are exempt from a number of reporting obligations,
including compliance with the Combined Code on corporate governance. In 2009, the UKLA initiated a two-year review of
the existing listing standards. The outcome of this review process was the recommendation to revise the UK listing regime
into “premium” and “standard” listings. Exchange-listed GDRs would fall into the standard listing category, whereas the
premium listing regime would create superficial compliance obligations for UK and foreign issuers. Foreign companies listed
as GDRs received invitations to upgrade their programs to premium listings. Companies admitted to Premium Listings are
eligible for inclusion in the FTSE indices, which guarantees participation of a new class of investors with greater liquidity.4,5
Appendix B provides details for GDRs and the direct Premium Listing standard and shows the extent of the reduced disclosure
and corporate governance obligations for GDRs. In summary, the preceding discussion reinforces the argument we made
in the previous section that it is not evident a priori whether or not the ADR-based findings regarding the bonding theory
would hold for GDRs due to their reduced disclosure regime; this motivates our empirical investigation.
2.2. Market segmentation hypothesis—origin and criticism
The theoretical work of Alexander et al. (1987) showed that for firms that list on two segmented markets the expected
returns are lower than for single-listed firms. The risk is more widely shared by investors and the required rate of return on
such a security will decrease upon dual listing. In line with these findings, Errunza and Losq (1985) showed that when the
investment barriers are asymmetric across two markets, firms from the markets with higher investment barriers would be
priced with a premium. The premium will vanish and the returns will decline as soon as the degree of market segmentation
drops (upon cross-listing). Next, the study by Alexander et al. (1988) was the first empirical work to demonstrate that
firms from completely or partially segmented markets have incentives to pursue a dual listing because it reduces their
required returns (cost of capital). The sample of the study contained both Canadian and non-Canadian foreign firms, and the
latter experienced a more dramatic decline in ex post returns than their Canadian peers, therefore supporting the market
segmentation hypothesis.
As noted by Karolyi (1998), the fair question in the studies examining market price reaction and changes in realized
returns after cross-listing is whether it is attributable to changes in cost of capital due to changes in firms’ risk exposure,
or to the reduction in the price bid-ask spread that occurs as a response to the increase in tradability (liquidity) of shares.
In other words, testing for the price/returns changes around cross-listing is a joint hypotheses test for the liquidity and the
risk decline effects. Miller (1999) suggested that an overlap of two events – a cross-listing intention announcement and the
actual cross-listing event – could drive results of the earlier studies. In his study that covered 181 dual-listed ADRs, Miller
documented positive abnormal returns around the announcement day followed by normal returns after the announcement.
Miller (1999) also documented differential changes in post-listing abnormal returns for exchange-listed liquid markets (Level
II and III ADRs) and for less liquid non-exchange listings (Level I and Rule 144A ADRs). Besides, firms from the emerging
markets experienced a more dramatic decline in the post-listing abnormal returns. Together, these findings provide support
for the liquidity effect and the market segmentation hypotheses.
Nevertheless, there were some intriguing findings and trends that this framework failed to explain. Foerster and Karolyi
(1999) documented a puzzling post-listing decline in firm value, which seems to contradict the market segmentation hypoth-
esis. Moreover, in their study the increase in firm value for Canadian stocks whose market is fairly integrated with the US
market was as dramatic as that for firms from other markets. Further, Foerster and Karolyi (2000) documented that for a
sample of Global Equity Offerings (GEOs), a magnitude of post-listing underperformance over a three-year window was not
a function of a firm origin, which is also inconsistent with the market segmentation framework. The authors put forward
alternative theories, such as the market incompleteness of Merton (1987) and liquidity of Amihud and Mendelson (1986), to
4
The results of the additional investigation that we omit from this study show that only four firms from Russia chose to upgrade their existing GDR
programs to the new Premium Listing regime with enhanced compliance requirements.
5
The official listing guide of the LSE cites potential benefits of Premium Listings: ‘As Premium Listed companies comply with the UK’s highest standards of
regulation and corporate governance, as a consequence they may enjoy a lower cost of capital through greater transparency and through building investor
confidence’ (LSE, 2013).
G Model
explain these findings. Stulz (1999) summarized the limitations of the market segmentation framework. First, he suggested
that the share price reaction around cross-listing documented in the above mentioned works was economically small, com-
pared to cost of capital decline implied by change in risk exposure as a result of elimination of barriers. Second, despite the
benefits of cross-listing, a fair number of firms choose not to cross-list their shares. Finally, the rapid growth in cross-listing
numbers is inconsistent with the argument that as markets become integrated over time, the net benefits of cross-listing
decline.
2.3. Bonding hypothesis—evidence in support and criticism
The theoretical works of Stulz (1999) and Coffee (1999) set the ground for alternative explanations for international
listings. Particularly, Stulz (1999) suggests that agency conflicts that arise between controlling and minority shareholders,
as well as information asymmetry, are the primary drivers for a cross-listing decision. A firm’s cost of capital depends on
its corporate governance attributes, which includes a firm’s internal controls, such as board of directors, as well as external
(market) controls, including monitoring from analysts, auditors, and institutional investors (see Chung et al., 2015). A firm
can improve its corporate governance attributes by credibly committing, or “bonding”, itself to stringent regulatory and legal
environments of a cross-listing market such as the NYSE and NASDAQ in the US. Coffee (1999) also highlighted the importance
of corporate governance in a firm’s cross-listing decision and suggested that valuation benefits accrue through legal bonding
mechanisms such as enhanced risk of litigation due to SEC enforcement, the investor ability to exercise class actions at
lower cost, and increased reporting requirements. Both Coffee (1999) and Stulz (1999) emphasized the significant role of
reputational intermediaries, including underwriters, analysts, auditors, and institutional investors who provide additional
scrutiny upon cross-listing.
The bonding hypothesis found support in empirical works. Reese and Weisbach (2002) reported that firms from French
Civil Law countries with poor protection for minority shareholders were more likely to cross-list in the US and these firms
selected major stock exchanges. Therefore, a US exchange listing provided a credible bonding mechanism. Doidge et al. (2004)
found that the cross-listing premium, as measured by Tobin’s q ratios, exists for firms with foreign listings and is higher for
exchange-listed firms and for those from countries with weak investor protection. In a follow up study, Doidge et al. (2009)
documented that the cross-listing premium existed in every year during the period of 1990–2005. The documented valuation
benefits, however, could stem from the cost of capital decline, consistent with bonding hypothesis, or from revisions in cash
flows growth due to improved operational performance of firms as a result of cross-listing.6 The most recent study of Hail
and Leuz (2009), based on a large panel of companies from 45 countries listed as ADRs on the NYSE and NASDAQ, as OTCs
and as private placements, found support to the bonding hypothesis. The documented decline in cost of capital was between
70 and 120 basis points, compared to substantially higher numbers documented by the earlier studies.
Several studies, nevertheless, challenged the validity of the bonding hypothesis. Licht (2001, 2003) argued that cor-
porate governance is of secondary importance in the case of cross-listing, and bonding, in fact, plays a negative role as
it allows foreign issuers to avoid some of the domestic regulations. Siegel (2005) examined the extent of legal actions
taken by the SEC against Mexican ADRs. He documented that in the case of asset tunneling, the SEC’s response was weak
and it failed to recover those assets. He argued this is because foreign authorities are either unwilling or unable to coop-
erate with US enforcement agencies. The author acknowledged, however, that the prior corporate governance violations
impeded the subsequent ability of firms to raise finance internationally. Thus, Siegel (2005) documented that firms received
a significant reputational penalty, which was more severe than the punishment they received from the US enforcement
agencies.
Bris et al. (2007) attempted to disentangle the market segmentation, liquidity, and the bonding hypotheses by examining
foreign stocks with dual-class shares cross-listed in the US during an early period of 1987–1996. The authors compared listing
premiums for listed/unlisted and low voting/high voting rights classes of shares. However, over-the-counter traded ADRs
dominated their sample, rather than exchange-listed firms that are the focus of our investigation and that entail the most
significant effects of elimination of trade barriers and bonding. In a study that is most relevant to our work, Bris et al. (2012)
performed tests of the impact of the market segmentation and liquidity hypotheses against information-based explanations.
The authors compared LSE listed and traded foreign firms and found evidence in favor of the latter effect. In Bris et al. (2012),
firms from Ireland dominated the sample of listed firms, whereas companies from Japan dominated the sample of traded
firms. We assume that these samples are predominantly direct listings, and not GDRs (no mentioning of it in the paper
but firms from emerging markets in general pursue GDRs), unlike our study that is exclusively GDRs-based. Different types
of listings (direct versus GDR) are subject to fundamentally differential reporting obligations and listing requirements, as
discussed above. Additionally, we also explore different types of bonding mechanisms − reputational and legal − which the
previously mentioned works did not do.
6
Hail and Leuz (2009) shed light on these complex issues: they estimated cost of capital and changes in cash flows simultaneously using implied cost of
capital models and found robust evidence of greater changes in cost of capital for exchange-listed firms and those from countries with a weak regulatory
regime.
G Model
3. Research design
3.1. Firm value estimation
We rely on the event-study methodology to examine changes in firm value upon cross-listing. We examine the monthly
values of Cumulative Abnormal Returns (CARs) measured over a 49-month window: 24 months prior to and 25 after cross-
listing, including a cross-listing month. Consistent with the methodology in the seminal study of Ritter (1991), the market-
adjusted (abnormal) return arit for firm i in month t is:
arit = rit − Rmt . (1)
The average market-adjusted return for the sample of n stocks in a month t is defined as:
1
n
ARt = arit . (2)

n
i=1
The cumulative abnormal returns for a period from month x to month y for a firm i and for a sample of n stock are,
respectively7 :

y
CARx−>y,i = art . (3.1)

t=x

y
CARx−>y,n = ARt . (3.2)

t=x
First, we use the FTSE UK All-Share equity index that covers constituents with the combined value of approximately 98
percent of the total UK market capitalization. This UK market index (Rmt ) was selected as a benchmark for CARs because all
examined GDRs cross-listed on the Main Market of the LSE. Second, we repeat all the tests using the local market indices to
compute CARs. Appendix C presents the list of these indices.
3.2. Market segmentation hypothesis test
Prior literature (Dimson et al., 2009) made an assumption about the degree of market segmentation based on the historical
correlation between regions’ stock market index returns. This methodology is fairly simplistic and noisy: the pairs-based
correlations may be reflective of the degree of market integration (the market segmentation hypothesis), the level of investor
protection (legal bonding hypothesis), or both. Consequently, we adopted an alternative measure of the degree of market
segmentation – trading costs – that represents an explicit barrier for foreign investors to transact with emerging markets
securities and does not include the investor protection component. We collected the trading costs information from the
State Street Global Advisors (2015)8 reports on emerging markets and the IFC Emerging markets report (2012). The trading
costs are reported as a percentage of the total gross consideration paid/received when purchasing/selling equities and may
include: exchange fee, stamp duty fee, transaction levy, trading fee, securities transaction tax, a local broker commission fee,
and other market charges. Emerging markets vary considerably in terms of the structure of trading costs and we calculated
this variable as the sum of any of the previously mentioned components, based on the Global Advisors report. The information
for Bahrain, Egypt, and Israel (three firms in total) was unavailable in the reports and we used Pakistan’s trading costs for
these markets, provided that experts place these countries in the same operational risks category. Appendix D reports the
total trading costs for the GDRs markets used in the analyses.
Next, one may argue that trading costs are one of the major but not the only component of the transaction costs imposed
on foreign investors. Foreign exchange risks, settlement terms, and account opening timeframes must also be considered
as factors diverting foreign investors from directly investing in emerging markets’ securities. Nevertheless, the State Street
experts suggest that these risks are residual in nature, compared to the actual trading costs, and forward-looking policies can
aid in management of those risks. For example, foreign exchange risks can be minimized via hedging, an account opening
framework is a one-time procedure/risk, and majority of the examined emerging markets have had the same T + 3 settlement
period as the LSE. In summary, trading costs are the primary component of total transaction costs that cannot be controlled
or avoided by foreign investors.
To test the market segmentation hypothesis, we estimate the following regression:
CARit = ˇ0 + ˇ1 DSegm + ˇ2 DCL + ˇ3 DCL ∗ DSegm + [firmcontrols] + eit . (4)
7
CARs for a month [-24] for a firm i are simply abnormal returns (ar) for that month.
8
Formerly, Elkins-McSherry consulting agency.
G Model
We control for firm-level factors associated with returns, such as size, book-to-market value of equity, leverage, and
industry (Fama and French 1992). Prior studies also suggest that stock tradability (liquidity) has a positive impact on a firm
value, and we therefore control for the trading volume throughout the tests (Cheung et al., 2015). In Model (4), CARs are
computed in accordance with (3.1). We partitioned the GDRs’ emerging markets into the low and high trading costs (low
and high investor transacting barriers) categories, based on the mean value of these costs for the examined markets. DSegm
– a dummy variable equal to one for firms with high trading costs and zero otherwise. DCL – a dummy variable equal to one
for observations after cross-listing, including the listing month. We expect changes in firm value to be more pronounced for
firms from the most segmented markets and hence, ␤3 should be significantly positive to support the market segmentation
hypothesis.
3.3. Legal bonding hypothesis test
Prior studies relied on the disclosure index created by the Center for International Financial Analysis and Research (CIFAR)
(Saudagaran and Diga 1997), a proxy for a country-level governance and investor protection level. Because the latest available
CIFAR edition is as of 1995, it might not fully capture the true variation in disclosure level for firms in our GDRs sample,
as the coverage of firms in this study ends in 2008 and it is highly unlikely that country-level disclosure metrics remained
unchanged over time. Next, the common/civil law partitioning of countries that was popular in prior studies (La Porta et al.,
2000) is also noisy and prior literature reports that more recently, this distinction blurs due to integration of the world’s
capital markets. For example, Russia, a civil law country, has adopted more progressive capital markets policies (such as
adoption of IFRS) compared to India, a common law country.
We therefore turn to an alternative country-level corporate governance indicator, Rule of Law, from the report of
Kaufmann et al. (2010) that, unlike CIFAR, is not static and is available from 1996 through 2014.9 This indicator “reflects
perceptions of the extent to which agents have confidence in and abide by the rules of society, and in particular the quality
of contract enforcement, property and investor rights, the police, and the courts, as well as the likelihood of crime and vio-
lence” (Kaufmann et al., 2010).10 Rule of Law is an annual metric (continuous), and a higher value suggests a better corporate
governance level. We follow prior literature that relied on Kaufmann et al. (2010) report (e.g., Daske et al., 2008) to test the
legal bonding hypothesis:
CARit = ˛0 + ˛1 Bond Rule Lawit + ˛2 DCL + ˛3 DCL ∗ Bond Rule Lawit + [firm controls] + it . (5)
In the above model, all the variables are as previously defined. BondRuleLaw is a reciprocal value of the Rule of Law
metrics. This modification is for convenience of interpretation: the lower the Rule of Law level on a local market and the
higher the reciprocal values of this metric, the greater the extent of bonding upon cross-listing and expected changes in firm
value. The coefficient ␣1 [␣1 + ␣3 ] determines the impact of the legal bonding on firm value before [after] cross-listing. Next,
we examine the impact of the firm-level bonding metric − a firm’s choice to raise capital through the Level III GDR program
that entails the most rigorous disclosure requirements and the highest extent of bonding − on firm value:
CARit = ˛0 + ˛1 IPOit + ˛2 DCL + ˛3 DCL ∗ IPOit + [firm controls] + it . (6)
In Model (6), all the variables are as previously defined and IPO is a dummy variable for GDRs raising capital through
Level III programs. We expect the coefficient ␣3 to be significantly positive to support the provisions of the legal bonding
framework.
3.4. Reputational bonding hypothesis test
According to Stulz (1999), upon cross-listing firms become subject to enhanced scrutiny from investors and financial
analysts, which is expected to put constraints on private actions of managers. Brennan and Hughes (1991) argue that investors
will be attracted only to stocks with significant analyst attention (following), and the greater the number of analysts issuing
forecasts, the larger the weight placed by investors on forecasts when valuing firms. Cowen et al. (2006) suggest that analysts
with less biased forecasts (higher accuracy) are more likely to develop positive reputations and attract investors. Greater
investor attention, in turn, would result in additional scrutiny and monitoring of a firm’s performance. Consistent with this
claim, Fang and Yasuda (2014) reported that the accuracy of analysts’ recommendations improves the ranking status (and
hence, reputation) of analysts in the institutional investors’ rosters. There is a positive association between analysts’ attention,
their reputation, and investor scrutiny of a firm’s performance. This discussion also suggests that enhanced monitoring of
cross-listed firms by analysts should manifest in improved accuracy and following.
The theory suggests several paths through which the properties of analysts’ forecasts, such as accuracy and following,
would improve after cross-listing. Lang and Lundholm (1996) argue that by adopting more forthcoming disclosure practices,
firms are able to improve properties of analysts’ forecasts such as following, accuracy, and dispersion. Hope (2003) posits
9
We use 1996 values for observations prior to 1996.
10
According to Cubbin and Stern (2005), this metric explains country-level institutional differences better than other corporate governance indicators
(see also Minkov 2011 for discussion).
G Model
that better enforcement of accounting disclosure standards makes firms’ reporting more transparent and less uncertain for
analysts, encouraging them to follow firms more actively. The increase in analysts’ following in this case is a result of the
improved disclosure and lower cost of following a firm (Lang et al., 2003). Alternatively, analysts might follow firms more
actively after cross-listing in response to increased investors’ demands for information. Next, improvement in accuracy
of analysts’ forecasts is expected to occur due to better communication of information by firms as a result of enhanced
disclosure requirements. Hope (2003) suggests that annual reports help analysts understand firms’ reporting practices at
a broad level, while disclosure notes assist them in assessing a firm’s future prospective in order to make forecasts more
accurately. Alternatively, the increased competition among analysts when the demand from investors rises should encourage
them to produce better quality forecasts (Lys and Soo 1995).
Further, Easley and O‘Hara (2004) developed a theoretical model demonstrating that when disclosure is of low quality,
the proportion of public versus private information could be increased through an improvement in properties of analyst
forecasts (accuracy, following, and dispersion). This should reduce the extent of information asymmetry and lower the cost
of capital. Therefore, analyst accuracy and following are valuable proxies for the extent of reputational bonding. In support,
findings in Botosan (1997) indicate the substitutive effect for the extent of disclosure and analyst forecasts on the cost
of equity capital. Kim and Pinnuck (2014) provide similar evidence for ADR and GDR cross-listed firms and document a
significant association between analyst following and cost of capital. In summary, the evidence in prior literature suggests
that analysts’ scrutiny of cross-listed firms is, first, an effective bonding mechanism, and second, it manifests in improved
accuracy of forecasts and following. We test the reputational bonding hypothesis:
CARit = 0 + 1 Accuracy/Follit + 2 DCL + 3 DCL ∗ Accuracy/Follit + [firm controls] + it . (7)
In Model (7), Accuracy − reciprocal of the forecast error, determined as the absolute value of [realized (I/B/E/S) EPS
−forecasted consensus EPS]/end of period price (Easton et al., 2002)11 ; Following − a number of reports issued by analysts
in a given month. The coefficient ␥1 [␥1 + ␥3 ] defines the impact of accuracy/following on firm value before [after] cross-
listing, which should be positive to support the predictions of the reputational bonding theory. Appendix E provides the
complete list of variables’ definitions.
3.5. The joint impact of the bonding mechanisms and the reduction in the degree of market segmentation upon cross-listing
on firm value
Once we examine the individual impact of the bonding mechanisms on firm value, we proceed to our main research objec-
tive, which is to examine whether the bonding mechanisms and the reduction in market segmentation have a complementary
or substitutive impact on firm value. We estimate the following regressions:
CARit = ı0 + ı1 DSegm + ı2 Bond Rule Lawit + ı3 DCL + ı4 DSegm ∗ Bond Rule Lawit + ı5 DSegm ∗ DCL+
(8.1)
+ı6 DCL ∗ Bond Rule Lawit + ı7 DSegm ∗ Bond Rule Lawit ∗ DCL + [firm controls] + it .
CARit = ı0 + ı1 DSegm + ı2 IPO + ı3 DCL + ı4 DSegm ∗ IPO + ı5 DSegm ∗ DCL + ı6 DCL ∗ IPO+
(8.2)
+ı7 DSegm ∗ IPO ∗ DCL + [firm controls] + it .
In the above models, we control for firm-level variables (size, etc.), as done in previous models. The coefficient ␦4 [sum
␦4 + ␦7 ] defines the joint impact of the bonding mechanisms and reduction in market segmentation on firm value before
[after] cross-listing. The positive [negative] values are interpreted as evidence in favor of complementary [substitutive]
effect. Finally, we estimate the following regressions to assess the joint impact of the reputational bonding mechanisms and
the reduction in market segmentation on firm value:
CARit = 0 + 1 DSegm + 2 Accuracyit + 3 DCL + 4 DSegm ∗ Accuracyit + 5 DSegm ∗ DCL+

(9.1)
+6 DCL ∗ Accuracyit + 7 DSegm ∗ Accuracyit ∗ DCL + [firm controls] + it .
CARit = 0 + 1 DSegm + 2 Follit + 3 DCL + 4 DSegm ∗ Follit + 5 DSegm ∗ DCL+

(9.2)
+6 DCL ∗ Follit + 7 DSegm ∗ Follit ∗ DCL + [firm controls] + it .
In the above models the coefficient 4 [4 + 7 ] defines the joint impact of accuracy and following and reduction in market
segmentation on firm value before [after] cross-listing.
11
The reciprocal value is used for convenience of interpretation of the results: the lower the forecast error and the higher the reciprocal value, the higher
the accuracy and the greater the expected changes in firm value.
G Model
CARs around the cross-listing event

0.9
0.8
0.7
0.6
0.5 CARsFTSE
0.4 CARslocal
0.3
0.2
0.1
0
-24 -22 -20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18 20 22 24
-0.1
Fig. 1. Cumulative Abnormal Returns (CARs) around the cross-listing event.

Cumulative Abnormal Returns (CARs) for the sample of 40 GDRs cross-listed on the London Stock Exchange over the period of 1994–2008. CARs are
benchmarked on (1) the FTSE ALL Share Index (UK) and (2) local market indices (Appendix C) and are computed in accordance with (3.2). CARs are
calculated over a 49-month window: 24 months prior to and 25 months after a cross-listing event, including a listing month.
4. Sample composition
The construction of the GDRs sample started with the complete lists of cross-listed firms from the two main sources
extensively used in prior literature—the “DR Universe” guide of the Citibank and the “DR Directory” of the Bank of New York
Mellon (thereafter BNY) (Bank of New York 2013). The initial sample included all active, as of 2008, GDRs with information
regarding a country of origin, effective listing year, listing exchange, a sponsor, etc. This sample is referred to as the Main
sample. For the GDRs that existed in the past and were cancelled prior to 2008 (Terminated GDRs), the complete list of those
programs was obtained from Citibank’s “DR Universe”. The records of the BNY and Citibank databases were verified against
the listing statistics of the LSE. The initial sample identification procedure resulted in 185 GDRs included in the Main (active)
sample of GDRs listed beginning in the 1990s, since the origin of GDRs, until December 2008, and remained active.12 A total
of 58 countries represented the initial GDR sample. The background check, nevertheless, revealed that out of 185 GDR-listed
firms in the Main sample only 75 firms met a definition of a cross-listed firm.13 Other GDR firms were either single-listed
on the LSE, or for them a GDR was the first listing. As for the sample of terminated GDRs, there were 5 cross-listed GDRs
identified that we added to the sample of active GDRs, thus the total sample of cross-listed GDRs included only 80 firms.
Next, we merged the sample of 80 GDRs with the data available in Datastream, which was a primary source for variables
data collection. Our analyses revealed that 14 GDRs missed data on the accounting variables, primarily in the pre-listing
period, 10 GDRs had errors in analyst forecasts prior to cross-listing, and 16 GDRs had insufficient data for the test window
that began 24 months prior to a cross-listing event. Thus, our final sample comprises 40 GDR cross-listed firms with data
available for all the empirical tests described in the previous section.
5. Empirical analyses
5.1. Descriptive statistics and CARs around cross-listing
Table 1 Panel A reports the distribution of the examined GDRs by country and industry. Cross-listed firms from India
dominate this sample, followed by firms from South Korea, Taiwan, and Turkey. Overall, 14 countries represent the GDRs
sample. It is also evident that financial institutions dominate the sample of GDRs, followed by firms conducting mining
and related business. Table 1 Panel B provides the descriptive statistics for the main variables of the study based on 1960
firm-month observations. All the variables have significant extreme values and have a skewed distribution. CARs experience
substantial variation based on both the UK market and the local market indices.
In Table 2, we report the CARs computed in accordance with (3.2) by month, also depicted on Fig. 1. The positive pre-listing
run up started in months 21 and 22 before cross-listing based on the UK and local market indices, respectively. After these
early months, CARs are significant in all months (at 1 percent or better). Therefore, unlike in the case of ADRs (Foerster and
12
We limit our investigation period by 2008 because of the global financial crisis that negatively impacted global markets, including the examined London
Stock Exchange post-2008. During the period of financial uncertainty, there can generally be a variety of institutional and regulatory changes implemented
within capital markets/exchanges that would affect firms’ decisions to cross-list or not, not limited by bonding requirements and the potential for elimination
of trade barriers. This would introduce noise to our analyses and put caveats on the results.
13
One may argue that the examined period of 1994–2008 is relatively long and that important institutional changes occurred in the European markets
during this period, such as introduction of the IFRS mandatory reporting requirements in 2005. Appendix A, based on the archival documentation sourced
from the LSE, shows that GDRs were reconciling their financials to IAS/IFRS from the inception of these programs and prior to 2005.
G Model
Table 1
Sample composition.
Panel A. Distribution of examined GDRs.
Country Number of companies Region
Bahrain 1 Middle East + Africa

Czech Republic 1 Eastern Europe
Egypt 1 Middle East + Africa
Greece 2 Eastern Europe
Hong Kong 1 Asia
India 15 Asia
Israel 1 Middle East + Africa
Pakistan 3 Middle East + Africa
Philippines 1 Asia
Poland 1 Eastern Europe
South Africa 1 Middle East + Africa
South Korea 4 Asia
Taiwan 4 Asia
Turkey 4 Eastern Europe
Total 40
Industry distribution (%):
[1] Banks and other financial institutions 42.5

[2] Mining and related business 22.55
[3] Retail and other industries that are individually not numerous, including 19.95
leisure and travel, airline, transportation, pharma, food processing, and
information technology
[4] Media 15
Total 100
Panel B. Descriptive statistics for the main variables of the study.
Variable No of firm-months obs. Mean Std. Dev. Minimum P25% Median P75% Maximum
CAR FTSE 1960 0.52 0.96 −2.54 −0.07 0.41 1.06 5.42
CAR Local ind 1960 0.47 0.79 −1.96 0.02 0.31 0.83 5.21
Size 1960 14.57 1.67 10.52 13.36 14.69 15.64 17.82
BM 1960 0.76 0.66 0.08 0.34 0.58 0.96 5.95
Lev 1960 4.60 7.34 −0.74 0.51 1.55 5.22 55.99
RuleLaw 1960 63.57 16.80 19.23 54.60 68.90 75.60 89.00
Accuracy 1960 179.09 556.74 0.001 19.04 54.20 134.09 10,170
Following 1960 8.18 5.85 1.00 3.00 7.00 12.00 27.00
The Table reports the distribution of 40 GDRs cross-listed on the London Stock Exchange over the period of 1994–2008 by country of origin and industry
(Panel A) and the descriptive statistics for the main variables of the study (Panel B). In Panel B, the analysis is based on the sample of 40 GDRs cross-listed
on the London Stock Exchange over the period of 1994–2008. CARs are benchmarked on the FTSE ALL Share Index (UK) and local market indices (Appendix
C). CARs are calculated over a 49-month window: 24 months prior to and 25 months after a cross-listing event, including a listing month, resulting in 1960
firm-month observations. All the variables are measured on a monthly basis to match monthly CARs and are defined as follows: Size − natural logarithm
of the Total Assets; BM − book-to-market value of equity; Leverage − natural logarithm of Total Assets divided by the Book Value of Equity; RuleLaw −
governance metric based on Kaufmann et al. (2010).
Karolyi, 1999), IPOs and Seasonal Equity Offerings, (Ritter 1991; Loughran and Ritter 1995), GDRs exhibit normal post-listing
performance, which is in line with the findings in Miller (1999). We elaborate on the nature of this pre-listing run up in the
next section.
5.2. Regression results
We report the results of the market segmentation hypothesis in Table 3. Prior to cross-listing, CARs of firms from the most
segmented markets are lower, compared to those of firms from the least segmented markets, based on both CAR metrics
(value = −0.10*** and −0.08*). The coefficient on the interaction term DSegm*DCL is significantly positive, suggesting that
changes in firm value as a result of cross-listing are more pronounced for firms from the most segmented markets. These
findings support the provisions of the market segmentation theory.
Table 4 reports the results of the legal bonding hypothesis tests. First, we focus on a country-level bonding proxy from the
report of Kaufmann et al. (2010), namely the Rule of Law variable. Prior to cross-listing, the extent of bonding has a positive
impact on firm value based on CARs benchmarked on the UK market index (value = 18.76***). After cross-listing, there is a
significant positive impact of the extent of bonding on firm value, as indicated by positive values of the Wald coefficient
tests (all significant at one percent or better). More importantly, the role of bonding is more pronounced after cross-listing,
as is evident from the significant positive coefficient on BondRulelaw*DCL interaction term. Next, we examine a firm-level
bonding proxy – a firm’s choice to cross-list through the most rigorous program – and the results are consistent with those
G Model
Table 2
Cumulative Abnormal Returns (CARs) around cross-listing.
Relative to a cross-listing month CARs based on FTSE ALL Share Index CARs based on local market indices
−24 −0.021 −0.002

−23 −0.002 0.032
−22 0.067 0.100**
−21 0.098** 0.116***
−20 0.136*** 0.160***
−19 0.155*** 0.184***
−18 0.181*** 0.207***
−17 0.185*** 0.204***
−16 0.261*** 0.255***
−15 0.325*** 0.319***
−14 0.369*** 0.347***
−13 0.382*** 0.359***
−12 0.437*** 0.384***
−11 0.459*** 0.404***
−10 0.460*** 0.405***
−9 0.461*** 0.416***
−8 0.517*** 0.452***
−7 0.524*** 0.464***
−6 0.537*** 0.485***
−5 0.586*** 0.520***
−4 0.609*** 0.543***
−3 0.622*** 0.576***
−2 0.692*** 0.627***
−1 0.766*** 0.677***
0 0.746*** 0.647***
1 0.688*** 0.595***
2 0.659*** 0.578***
3 0.690*** 0.602***
4 0.689*** 0.612***
5 0.670*** 0.604***
6 0.656*** 0.579***
7 0.675*** 0.595***
8 0.671*** 0.599***
9 0.691*** 0.598***
10 0.698*** 0.607***
11 0.716*** 0.595***
12 0.670*** 0.571***
13 0.655*** 0.566***
14 0.579*** 0.546***
15 0.612*** 0.557***
16 0.625*** 0.551***
17 0.615*** 0.561***
18 0.603*** 0.565***
19 0.641*** 0.569***
20 0.656*** 0.589***
21 0.656*** 0.589***
22 0.647*** 0.593***
23 0.635*** 0.577***
24 0.592*** 0.577***
The Table reports the Cumulative Abnormal Returns (CARs) for the sample of 40 GDRs cross-listed on the London Stock Exchange over the period of
1994–2008. CARs are benchmarked on the FTSE ALL Share Index (UK) and on local market indices (Appendix C) and are computed in accordance with (3.2).
CARs are calculated over a 49-month window: 24 months prior to and 25 months after cross-listing, including a listing month. The *, **, and *** denote a
statistical significance at 10, 5 and 1 percent, respectively, and indicates that CARs are different from zero.
reported for the country-level bonding proxy. This commitment has a positive impact on firm value before cross-listing
based on CARs benchmarked on the UK market index (value = 0.27***), and this effect is more pronounced after cross-listing,
as is evident from the significantly positive coefficient on the DCL*IPO interaction term (value = 0.31***; 0.45***).
In summary, the results reported in Table 4 are consistent with the proposition that commitment to rigorous reporting
requirements manifests in positive changes in firm value upon cross-listing. Importantly, the pre-listing run up in CARs
documented in the previous section does not contradict the market segmentation hypothesis, as our results in Table 3
suggest that most segmented markets firms have lower value before cross-listing, consistent with the market segmentation
theory. This intriguing run up that received limited attention in prior literature is well explained by the evidence in Table 4,
in combination with the information regarding the eligibility for admission for GDRs reported in Appendix B. Particularly,
firms must release through prospectus and other pre-listing documents up to three years of audited financial statements,
management reports, and statements of compliance with local market corporate governance norms. GDR issuers are required
to submit these pre-listing particulars for the UKLA approval as early as possible prior to the listing (LSE, 2015). Additionally,
G Model
Table 3
Market segmentation hypothesis test.
CARit = ˇ0 + ˇ1 DSegm + ˇ2 DCL + ˇ3 DCL ∗ DSegm + +[firmcontrols] + eit .
CARs benchmarked on local market indices CARs benchmarked on UK market index
Coefficient Coefficient
Constant 2.29 *** 1.51***
DSegm −0.10*** −0.08*
DCL 0.22*** 0.18***
DSegm*DCL 0.11* 0.23***
Firm controls, Industry effects Yes Yes
No. Obs. 1960 1960
Adj R-squared 0.38 0.38
The Table reports the results of the market segmentation hypothesis test. The analysis is based on the sample of 40 GDRs cross-listed on the London
Stock Exchange over the period of 1994–2008. CARs are benchmarked on the FTSE ALL Share Index (UK) and local market indices (Appendix C). CARs are
calculated over a 49-month window: 24 months prior to and 25 months after a cross-listing event, including a listing month, resulting in 1960 firm-month
observations. Control var: Size − natural logarithm of the Total Assets; BM − book-to-market value of equity; Lev − Debt-to-Equity ratio; Vol − logarithm
of the trading volume; DSegm − a dummy variable equal to one for firms from the most segmented markets (with highest trading costs); DCL − a dummy
variable equal to 1 for observations after cross-listing, including the listing month itself. The estimation is performed using robust standard errors. The *,
**, and *** denote a statistical significance at 10, 5 and 1 percent, respectively.
Table 4
Legal Bonding hypothesis test.
CARit = ˛0 + ˛1 BondRuleLawit + ˛2 DCL + ˛3 DCL ∗ BondRuleLawit + [firmcontrols] + it .
CARit = ˛0 + ˛1 IPOit + ˛2 DCL + ˛3 DCL ∗ IPOit + [firmcontrols] + it .
Coefficient Rule of Law IPO Rule of Law IPO
Constant 2.06*** 1.8*** 0.97*** 0.18

BondRuleLaw 2.72 18.76***
IPO 0.01 0.27***
DCL 0.17*** 0.16*** 0.22*** 0.14***
DCL*BondRuleLaw 7.75*** 8.08***
DCL*IPO 0.31*** 0.45***
Firm controls, Industry effects Yes Yes Yes Yes
Wald Test [␣1 + ␣3 ] = 0 10.47*** 26.84***
No. Obs. 1960 1960 1960 1960
Adj R-squared 0.41 0.42 0.44 0.46
The Table reports the results of the legal bonding hypothesis tests using, first, a country-level bonding metric, Rule of Law, from the report of Kaufmann
et al. (2010) and second, a firm-level bonding proxy, IPO. The analysis is based on the sample of 40 GDRs cross-listed on the London Stock Exchange over the
period of 1994–2008. CARs are benchmarked on the FTSE ALL Share Index (UK) and local market indices (Appendix C). CARs are calculated over a 49-month
window: 24 months prior to and 25 months after a cross-listing event, including a listing month, resulting in 1960 firm-month observations. Control var:
Size − natural logarithm of the Total Assets; BM − book-to-market value of equity; Lev − Debt-to-Equity ratio; Vol − logarithm of the trading volume;
BondRuleLaw − reciprocal of Rule of Law metric from Kaufmann et al. (2010); IPO − a dummy variable equal to 1 for Level III (capital raising) GDRs; DCL
− a dummy variable equal to 1 for observations after cross-listing, including the listing month itself. The estimation is performed using robust standard
errors. The *, **, and *** denote a statistical significance at 10, 5 and 1 percent, respectively.
firms applying for a GDR listing must fully comply with IFRS. Therefore, the commitment to enhanced reporting requirements
and hence, bonding, starts 15 − 36 months prior to this event, according to the UKLA GDR listing timeline guidance (LSE,
2013). The LSE’s guide to listings suggests that “pre-IPO preparation includes the critical review [by the Exchange and UKLA]
of a company’s business plan and growth prospects, assessing the management team, appointing an appropriate board,
tightening internal controls, improving operational efficiency and resolving issues that may adversely affect the listing early
on”. Additionally, the LSE cites pre-listing practices: “A company should expect to show investors a consistent pattern of top-
and bottom-line growth and a sound balance sheet” (LSE, 2015). These early bonding requirements explain the evidence on
the pre-listing run up in the CARs (Fig. 1) and the positive impact of legal bonding metrics on firm value prior to cross-listing
(Table 4).14
Lastly, Fig. 2 depicts the rank-order impact of the extent of legal bonding on changes in CARs based on the FTSE UK index
and the Rule of Law metric.15 The pre-listing run up in firm value is more pronounced for firms for which the extent of
bonding is the highest, lending support to the legal bonding hypothesis.
Next, we examine the impact of the reputational bonding on firm value in Table 5. We find some evidence that analyst
following has a negative impact on firm value before and after cross-listing based on CARs benchmarked on the UK market
14
In greater details, these pre-listing obligations for GDRs are described in the UKLA’s Listing Rules Chapter 18, available at:
https://www.handbook.fca.org.uk/handbook/LR/18/.
15
The results based on the local market indices are qualitatively similar.
G Model
CARs for GDRs from Low [high] Rule of Law markets for
which extent of bonding upon
1.6
cross-listing is High [low]
1.4
1.2
0.8
LowRuleLaw
0.6
HighRuleLaw
0.4
0.2
0
-24-23-22-21-20-19-18-17-16-15-14-13-12-11-10-9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 101112131415161718192021222324
-0.2
Fig. 2. Changes in CARs for firms from low and high quality regulatory environment.
Cumulative Abnormal Returns (CARs) for the sample of 40 GDRs cross-listed on the London Stock Exchange over the period of 1994–2008. Firms are
categorized by the extent of legal bonding upon cross-listing: GDRs from Low [high] Rule of Law markets experience high [low] changes in CARs before
and after cross-listing. CARs are benchmarked on the FTSE ALL Share Index (UK) and are computed in accordance with (3.2). CARs are calculated over a
49-month window: 24 months prior to and 25 months after a cross-listing event, including a listing month.
Table 5
Reputational bonding hypothesis test.
CARit = 0 + 1 Accuracy/Follit + 2 DCL + 3 DCL ∗ Accuracy/Follit + [firmcontrols] + it .
Coefficient Accuracy Following Accuracy Following
Constant 2.17*** 2.15*** 1.37*** 0.95**

Accuracy −0.00004 −0.00006*
Following 0.003 −0.013***
DCL 0.29*** 0.34*** 0.35*** 0.60***
Accuracy*DCL 0.00009* 0.00012**
Following*DCL −0.004 −0.02***
Wald Test: ␣1 + ␣3 = 0 0.00005 −0.0005 0.00006 −0.03**
No. Obs. 1960 1960 1960 1960
Adj R-squared 0.40 0.40 0.41 0.42
The Table reports the results of the reputational bonding hypothesis tests. The analysis is based on the sample of 40 GDRs cross-listed on the London
Stock Exchange over the period of 1994–2008. CARs are benchmarked on the FTSE ALL Share Index (UK) and local market indices (Appendix C). CARs are
calculated over a 49-month window: 24 months prior to and 25 months after a cross-listing event, including a listing month, resulting in 1960 firm-month
observations. Control var: Size − natural logarithm of the Total Assets; BM − book-to-market value of equity; Lev − Debt-to-Equity ratio; Vol − logarithm
of the trading volume; Accuracy − reciprocal of the forecast error, determined as absolute value of [consensus forecast − reported EPS]/beginning of period
price; Following − a number of reports issued by analysts in a given month; DCL − a dummy variable equal to 1 for observations after cross-listing, including
the listing month itself. The estimation is performed using robust standard errors. The *, **, and *** denote a statistical significance at 10, 5 and 1 percent,
respectively.
index. Accuracy of analyst forecasts has virtually non-detectable impact on firm value before and after cross-listing, although
there is some evidence that accuracy makes a more positive contribution towards firm valuation after cross-listing. This
evidence, nevertheless, is based on the pooled sample of firms with different origins, which prompts further investigation
into the role of financial analysts in providing an effective monitoring mechanism for cross-listing firms. Our joint hypotheses
tests will shed further light on this.
We proceed to our main research objective, which is to examine the joint role of the two theories – market segmentation
and bonding – in firm valuation. The results in Table 6 suggest that the legal bonding mechanism, as measured by the
country-level Rule of Law metric, and the reduction in market segmentation have a complementary positive effect on firm
value as a result of cross-listing. Particularly, the Wald test statistic is 94.38*** based on the CARs benchmarked on the UK
market index. Interestingly enough, this result does not hold when we replace the country-level metric with a firm-level
G Model
Table 6
The joint impact of the legal bonding mechanisms and the reduction in market segmentation on firm value upon cross-listing.
CARit = ı0 + ı1 DSegm + ı2 BondRuleLawit + ı3 DCL + ı4 DSegm ∗ BondRuleLawit + ı5 DSegm ∗ DCL+

+ı6 DCL ∗ BondRuleLawit + ı7 DSegm ∗ BondRuleLawit ∗ DCL + [firmcontrols] + it .
CARit = ı0 + ı1 DSegm + ı2 IPO + ı3 DCL + ı4 DSegm ∗ IPO + ı5 DSegm ∗ DCL + ı6 DCL ∗ IPO+
+ı7 DSegm ∗ IPO ∗ DCL + [firmcontrols] + it .
Coefficient Rule of Law IPO Rule of Law IPO

Constant 2.4*** 1.83*** 1.51*** −0.48
DSegm 0.25 0.19*** −0.07 0.62***
DCL 0.03 0.09** −0.02 0.02
BondRuleLaw/IPO 3.69 0.29*** 23.49*** 1.01***
DSegm*[BondRuleLaw/IPO] −22.1* −0.57*** 11.51 −1.18***
DSegm*DCL 0.03 0.04 −0.88** 0.17**
DCL*[BondRuleLaw/IPO] 8.94*** 0.27*** 9.46*** 0.42***
DSegm*[BondRuleLaw/IPO]*DCL 9.28 0.24** 82.78*** 0.13
Wald Test: ␦4 + ␦7 = 0 −12.81 −0.33*** 94.38*** −1.04***
Adj R-squared 0.39 0.42 0.44 0.54
The Table reports the results of the test of the joint impact of the legal bonding mechanisms and the reduction in market segmentation on firm value.
The analysis is based on the sample of 40 GDRs cross-listed on the London Stock Exchange over the period of 1994–2008. CARs are benchmarked on the
FTSE ALL Share Index (UK) and local market indices (Appendix C). CARs are calculated over a 49-month window: 24 months prior to and 25 months after
a cross-listing event, including a listing month, resulting in 1960 firm-month observations. Control var: Size − natural logarithm of the Total Assets; BM
− book-to-market value of equity; Lev − Debt-to-Equity ratio; Vol − logarithm of the trading volume; DSegm − a dummy variable equal to one for firms
from the most segmented markets (with the highest trading costs); BondRuleLaw − reciprocal of Rule of Law metric from Kaufmann et al. (2010); IPO − a
dummy variable equal to 1 for Level III (capital raising) GDRs; DCL − a dummy variable equal to 1 for observations after cross-listing, including the listing
month itself. The estimation is performed using robust standard errors. The *, **, and *** denote a statistical significance at 10, 5 and 1 percent, respectively.
capital-raising variable, IPO. Prior to cross-listing, firms’ commitments to raise capital has a positive impact on firm value but
this impact is less significant for firms from the most segmented markets, and the coefficient on DSegm*IPO is significantly
negative, based on both CARs estimations. Therefore, there is no evidence of the joint complementary impact of the bonding
commitment and the reduction in market segmentation on firm value. This result holds after cross-listing, and the Wald test
statistics are significantly negative at one percent or better.
In summary, the empirical tests provide support to the market segmentation and the legal bonding theories, when we
examine those as standalone frameworks. Nevertheless, country-level institutions and firm-level bonding commitment play
markedly different roles in the case of the firms from the most segmented markets – those characterized by the highest
trading barriers. Particularly, GDR cross-listed firms enjoy enhanced valuation benefits as a result of the elimination of trade
barriers and the improvement in investor protection level. On the other hand, investors do not value firms from the most
segmented markets as favorably as those from the least segmented markets when firms commit to the capital raising activity.
In Table 7, we report the results of the joint test for the reputational bonding metrics and the reduction in market
segmentation. Prior to cross-listing, analyst following has an incrementally positive impact on firm value for firms from the
most segmented markets (value = 0.031***; 0.071***), whereas accuracy of analyst forecasts does not affect it. Interestingly,
there is a negative impact of analyst following on firm value for firms from the least segmented markets prior to and after
cross-listing. After cross-listing, analyst following and the reduction in market segmentation play joint, complementary
roles in valuation of firms, whereas accuracy of analyst forecasts has an opposite effect on firm value. Conversely, accuracy
of forecasts has a positive impact on firm value for firms from the least segmented markets, but only after cross-listing. Our
extended investigation suggests that accuracy of analyst forecasts improved only marginally after cross-listing (based on the
median value and only at ten percent), which explains the mixed results regarding this metric. The documented negative
impact of accuracy of forecasts on firm value can be due to the learning effect, as after cross-listing firms report significantly
more information that analysts need to learn how to interpret. Our investigation is limited by 24 months after cross-listing,
and we do not examine whether or not the accuracy of forecasts would improve in later periods. Prior studies of Kim and
Pinnuck (2014) and Abdallah (2008) also documented that accuracy and following have distinctly different impacts on firm
valuation.
Overall it appears that enhanced analysts’ following [accuracy] benefits firms from the most [least] segmented markets to
a greater extent than firms from the least [most] segmented markets indicating that analysts are an important reputational
bonding mechanism, but only for selective categories of firms. We conclude that the extent of reputational bonding in the
case of cross-listed firms is limited.
5.3. Additional tests
As suggested by Foerster and Karolyi (1999) and Miller (1999), it would be theoretically more appropriate to examine
the changes in firm value around a firm’s cross-listing announcement rather than the actual listing event because the
first commitment to cross-listing occurs when firms signal their intentions to cross-list through public announcements. This
G Model
Table 7
The joint impact of the reputational bonding mechanisms and the reduction in market segmentation on firm value upon cross-listing.
CARit = 0 + 1 DSegm + 2 Accuracyit + 3 DCL + 4 DSegm ∗ Accuracyit + 5 DSegm ∗ DCL+

+6 DCL ∗ Accuracyit + 7 DSegm ∗ Accuracyit ∗ DCL + [firmcontrols] + it .
CARit = 0 + 1 DSegm + 2 Follit + 3 DCL + 4 DSegm ∗ Follit + 5 DSegm ∗ DCL+

+6 DCL ∗ Follit + 7 DSegm ∗ Follit ∗ DCL + [firmcontrols] + it .
Coefficient Accuracy Following Accuracy Following

Constant 2.26*** 2.46*** 1.38*** 1.44***
DSegm −0.07* −0.30*** −0.014 0.40***
Accuracy/Following 0.00007 −0.01 0.0001 −0.06***
DCL 0.21*** 0.31*** 0.22*** 0.71***
DSegm*[Accuracy/Following] −0.0001 0.03*** −0.0002 0.07***
DSegm*DCL 0.10 0.12 0.21** −0.10
DCL*[Accuracy/Following] 0.00008 −0.005 0.00009 −0.03***
DSegm*[Accuracy/Following]*DCL −0.00007 −0.007 −0.00012 0.01
Wald Test: ␦4 + ␦7 = 0 (most segm) −0.00021*** 0.02*** −0.0003*** 0.08***
Wald test: ␦2 + ␦6 = 0 (least segm) 0.00016** −0.015** 0.0002*** −0.09***
Adj R-squared 0.38 0.39 0.38 0.44
The Table reports the results of the test of the joint impact of the reputational bonding mechanisms and the reduction in market segmentation on firm
value. The analysis is based on the sample of 40 GDRs cross-listed on the London Stock Exchange over the period of 1994–2008. CARs are benchmarked on
the FTSE ALL Share Index (UK) and local market indices (Appendix C). CARs are calculated over a 49-month window: 24 months prior to and 25 months
after a cross-listing event, including a listing month, resulting in 1960 firm-month observations. Control var: Size − natural logarithm of the Total Assets;
BM − book-to-market value of equity; Lev − Debt-to-Equity ratio; Vol − logarithm of the trading volume; DSegm − a dummy variable equal to one for firms
from the most segmented markets (with the highest trading costs); Accuracy − reciprocal of the forecast error, determined as absolute value of [consensus
forecast − reported EPS]/beginning of period price; Following − a number of reports issued by analysts in a given month; DCL − a dummy variable equal to
1 for observations after cross-listing, including the listing month itself. The estimation is performed using robust standard errors. The *, **, and *** denote
a statistical significance at 10, 5 and 1 percent, respectively.
Table 8
Cross-listing announcements.
Company Country Listing announcement date Listing date Raised capital (Level III GDRs)
Gulf Finance House Bahrain June 4, 2007 December 7, 2007 Yes

Amtek Auto India Nov 23, 2004 November 25, 2004 Yes
Axis Bank India March 16, 2005 March 22, 2005 Yes
Federal Bank India Jan 27, 2006 January 31, 2006 Yes
PVP Ventures India March 13, 2000 March 30, 2000 Yes
Subex India Feb 24, 2006 March 9, 2007 Yes
Tata Tea India February 28, 2000 March 8, 2000 Yes
MCB Bank Pakistan July 4, 2006 October 10, 2006 Yes
Oil&Gas Development Pakistan December 23, 2005 December 6, 2006 Yes
United Bank Pakistan September 25, 2006 June 29, 2007 Yes
Banco de Oro Philippines November 30, 2005 January 19, 2006 Yes
Naspers South Africa May 17, 2007 August 23, 2007 No
Daishin Securities South Korea October 29, 2007 November 2, 2007 Yes
Samsung C&T South Korea December 6, 2006 December 7, 2006 No
Evergreen Marine Taiwan July 26, 1996 August 5, 1996 No
Hon Hai Precision Taiwan August 10, 1999 October 8, 1999 Yes
Sunplus Technology Taiwan March 9, 2001 March 16, 2001 Yes
Finansbank Turkey February 4, 1998 June 11, 1998 No
The Table reports the list of GDRs (n = 18) for which listing announcements were available in the Lexis/Nexis database and other public sources.
approach, nevertheless, has some drawbacks. First, firms could make several announcements regarding an intention to cross-
list, with a varying degree of details and assurance. Second, for a significant portion of cross-listed firms, announcements may
not be available, and in the study of Foerster and Karolyi (1999) more than half of the ADRs sample was lost. Additionally,
cross-listing commitments may be reversed with some or no impact on firm value, whereas delisting after an exchange
cross-listing event is potentially very costly. Finally, the coverage of emerging markets in databases is poor prior to cross-
listing, suggesting that the identified announcements may not be the first ones. Nevertheless, in Lexis/Nexis, we found the
announcement dates for 18 GDRs (out of 40), which is 45 percent of the examined sample.
Table 8 provides the details regarding these GDRs, including the announcements and actual listing dates. A total of
14 GDRs raised capital via the Level III program. We found that the most comprehensive coverage of the capital market
events, in terms of the extent of historic information, details, and the coverage in Lexis/Nexis, was for Taiwanese firms; the
earliest announcement dates back to 1996 for Evergreen Marine. Next, we identified the longest lag between the listing
announcement date and the listing event; it was for India’s Subex, Pakistani Oil & Gas Development, and United Bank.
For several firms, announcement dates were almost immediately before the cross-listing event (e.g., Amtek Auto of India,
G Model
CARs around the listing announcement day
0.35
0.3
CARsFTSE
0.25
0.2 CARsLocal
0.15
0.1
0.05
0
-100
-94
-88
-82
-76
-70
-64
-58
-52
-46
-40
-34
-28
-22
-16
-10
-4
2
8
14
20
26
32
38
44
50
56
62
68
74
80
86
92
98
-0.05
Fig. 3. Cumulative Abnormal Returns (CARs) around the cross-listing announcement day.
Cumulative Abnormal Returns (CARs) for the sample of 18 GDRs cross-listed on the London Stock Exchange, for which the listing announcement dates were
available (Table 8). CARs are benchmarked on (1) the FTSE ALL Share Index (UK) and (2) local market indices (Appendix C) and are computed in accordance
with (3.2) but on a daily basis: 100 days prior to and 100 days after a cross-listing announcement.
Samsung C&T of South Korea). These short lags raise a suspicion that the identified announcements were not the first ones,
as we expect firms to start preparing for listing long before the event, as was discussed previously. Fig. 3 depicts the behavior
of CARs around the announcement dates. Consistent with the main results reported in the preceding section, CARs increase
over the examined interval that starts 100 days before and ends 100 after the cross-listing announcement day. The major
increase in firm value, on the other hand, begins before the announcement.
We repeat the regression tests based on the reduced sample that includes the firms for which announcement informa-
tion was available. We do not find conclusive evidence regarding the individual impact of the market segmentation and
legal/reputational bonding proxies on changes in firm value upon cross-listing. This is not surprising, given the limited sam-
ple size and the possibility that the identified announcements may not be the first ones. Additionally, examined CARs are
based on daily returns, whereas firm controls and legal and bonding proxies are computed on a monthly or annual basis.
Nevertheless, the results of this additional test strongly support the conclusions regarding the joint complementary role of
the legal bonding metric, Rule of Law, and the reduction in market segmentation on firm value.
6. Conclusions
We revisit the debate regarding the legitimacy of the market segmentation and the bonding hypotheses for the cross-
listing phenomenon, using novel GDR settings. Unlike prior studies that used a simple emerging/developed countries
partitioning, we rely on a more explicit measure of the investment barriers – trading costs – to test the market segmentation
hypothesis. This measure has little correlation with the home market investor protection constructs, providing stronger
experimental settings to distinguish between the two frameworks. While we find that both frameworks have a positive
impact on changes in firm value upon cross-listing when examined individually, we also report that the legal bonding and
reduction in market segmentation play a complementary positive role in firm valuation. Interestingly, raising capital via
the most stringent GDR program (firm-level bonding mechanism) and the reduction in market segmentation have a joint
but substitutive role in firm valuation. Lastly, we document that the extent of reputational bonding in the case of cross-
listed firms is limited to a post-listing period and that analysts’ accuracy and following have a markedly distinct role in firm
valuation.
The study is subject to several limitations. We rely on the trading costs collected from a 2015 State Street report, while
Bekaert and Harvey (1995) suggest a complex theoretical model for estimation of the capital markets segmentation that
has a time-varying component. The proposed model is rather complex and is based on country-level parameters that are
unavailable to us. Further, prior literature examining the role of the reputation and composition of analysts relied on the
information from I/B/E/S to construct the reputational list (e.g., Fan and Yasuda 2014). We also referred to I/B/E/S but were
unable to find the equivalent information for the emerging markets stocks covered in our study. For the 1990s, the system
returned the error messages for most firms, whereas for the later period, we ended up with most fields missing data. As
was noted previously, the coverage of emerging markets firms is generally poor in the 1990s and particularly in relation to
G Model
analyst forecasts metrics. These limitations can be addressed in the future research when the coverage of emerging markets
in databases improves.
Appendix A. Differential reporting obligations for exchange-listed ADRs and GDRs.
Type Pre-listing US Securities Listing Submitted Reporting Accounting SOX or

requirements Acts/EU Alternatives forms with frequency reconciliation equivalent
Directive SEC/FSA requirements
Level II ADR Three-year 1933, 1934 NYSE, NASDAQ, F-6, 20-Fa , F-6K Quarterly Reconciliation Full
audited Amex (reviewed) and of the major compliance
financial annual Balance Sheet with SOX
statements; (audited) and Income provisions
working capital Statement
statement; items to the US
pro-forma GAAPa
statements due
to change in
business
Level III ADR 1933, 1934 NYSE, NASDAQ, F-1, F-6, F-6K, Quarterly Full Full
(capital raising) Amex 20-Fa (reviewed) and reconciliation compliance
annual of financial with SOX
(audited) statements to provisions
the US GAAPa
Level II & Level III 12–18 months EU Prospectus London Stock None/Annual Interim Reconciliation No equivalent
GDRs (capital pre-listing Directive, FSA Exchange Main report (unaudited), of the major of SOX; no
raising) history of Chapter 18 and Market annual Balance Sheet requirement to
accounts, no provisions of combined with (audited) and Income comply with
working capital Rule144A if offering on US Statement the Combined
statement or combined with PORTAL items to the IAS Code on
change in US private (prior to 2001) Corporate
business offering and IFRS (after Governance
pro-forma 2001)
requiredb .
Source: Bank of New York Mellon; Citibank; London Stock Exchange (website and archives); UK Listing Authority.
a In January 2008 the SEC adopted an amendment to the Form 20-F. In particular, it eliminated the requirement to reconcile
financial statements to the US GAAP for foreign private issuers reporting in accordance with IFRS as issued by the IASB. The
amendment applies to financial statements with financial year ending post November 15, 2007 (SEC – Federal Register –
Release Nos. 33-8879; 34-57026; International Series Release No. 1306; File No. S7-13-07).
b Please note that despite the fact that these statements are not required some GDR-listed companies have chosen to
voluntarily increase the scope of reported information and disclosure in order to satisfy investor information needs and
signal their commitments to transparent reporting.
Appendix B. Comparison of LSE listing options: Premium (direct) Listing versus Standard listing (GDR).
Criteria Premium Listing Standard Listing as GDR
ELIGIBILITY FOR ADMISSION:

Transferability of securities admitted to trade Full Full
Minimum market capitalization/free float GBP 700,000/25% GBP 700,000/25%
At least 75% of the business must be supported by a three-year Yes No
earnings records
Issuer has controlled the business for at least three years Yes No
Sponsor and publication of a prospectus Yes No sponsor required but companies
publish prospectus
Historical financial information for three years prior to the listing Yes Yes, or shorter period
(audited); disclosure of risk factors
Operating and financial review statement, capital resources disclosure, Yes Yes
and trend information disclosed in the prospectus
Description of the functions, activities, remuneration, and experience Yes Yes
of management
Compliance with corporate governance norms in the home jurisdiction Yes Yes
Related party transactions Yes Yes, special requirements for Russian
GDRs
Working capital statement Yes No
G Model
CONTINUING OBLIGATIONS:
Sponsor in connection to further issues Yes No
Compliance with Model Code that restricts dealings by managerial Yes No
staff in possession of insider information
Pre-emption rights on further issuance of shares Yes No
Compliance with UK Corporate Governance Code Yes No but companies can voluntarily
apply this or other codes and make a
disclosure
Significant and related party transactions disclosure Yes No
Disclosure of the insider list Yes Yes
Annual financial report prepared in accordance with IFRS or equivalent Yes Yes
(audited)
Interim financial reports Yes No
Interim management statement; major shareholding information Yes No
Cancellation of listing must be approved by 75% of shareholders Yes No
Source: UK Listing Authority, London Stock Exchange (LSE).
Appendix C. Local market indices (in USD) used in computations of CARs and in the regression tests.
Country Stock Market Index
Bahrain S&P Broad Market Index

Hong Kong Hang Seng
Czech Republic CZ PX-50
Egypt Share Price Index
Greece Athens Composite
India India BSE Sensex; MSCI India
Israel Israel TA100
Pakistan KSE 100; MSCI Pakistan
Philippines MSCI Philippines
Poland Warsaw MWIG 40
South Africa JSE All Share
South Korea Korea Composite; MSCI Korea
Taiwan Taiwan SE Weighted Index; MSCI Taiwan
Turkey ISE National 100; MSCI Turkey
Appendix D. Equity trading costs by country.
Country Trading costs as a percentage of the total consideration
Bahrain 0.02
Czech Republic 0.7079
Egypt 0.02
Greece 0.0325
Hong Kong 0.108
India 0.125
Israel 0.02
Pakistan 0.02
Philippines 0.01
Poland 1
South Africa 0.25
South Korea 0.5
Taiwan 0.3
Turkey 0.4668
The Table reports the trading costs as a percentage of consideration paid/received when purchasing/selling local equity
instruments. The information comes from the reports of the State Street Global Advisors and the IFC’s guide on Emerging
markets.
Appendix E. Variables definitions.
CAR Cumulative adjusted abnormal returns benchmarked on the local

market indices, monthly metric
CARFTSE Cumulative adjusted abnormal returns benchmarked on the UK
market index − FTSE All Share, monthly metric
r Firm’s realized returns, adjusted for stock splits and dividends,
monthly metric
Size Natural logarithm of total assets of a firm, annual metric
BM Book-to-market value of equity of a firm, monthly metric
Lev Debt-to-equity ratio of a frim, annual metric
G Model
DSegm Dummy variable equal to one for firms representing the most
segmented markets − those for which the trading costs exceed the
mean value of trading costs for all the examined markets (App. D)
DCL Dummy variable equal to one after cross-listing, including the listing
month
Rule of Law Country-level indicator of the strength of the local market corporate
governance, per the report of Kaufmann et al. (2010), annual metric
Defined as:
“perceptions of the extent to which agents have confidence in and abide by
the rules of society, and in particular the quality of contract enforcement,
property rights, the police, and the courts, as well as the likelihood of
crime and violence”
BondRuleLaw Reciprocal of the Rule of Law, annual metric
Industry Broad industry representation:
[1] Banks and other financial institutions;
[2] Mining and related business;
[3] Retail and other industries that are individually not numerous,
including leisure and travel, airline, transportation, pharma, food
processing, and information technology;
[4] Media
Accuracy Reciprocal of the forecast error, determined as absolute value of
[realized (I/B/E/S) EPS −forecasted consensus EPS]/end of period price
(Easton et al., 2002), annual metric
Following Number of reports issued by analysts in a given month, monthly metric
IPO Dummy variable equal to one for firms cross-listed as Level III GDRs
(capital-raising programs)
All the variables were downloaded in USD.
References
Abdallah, A., 2008. Do regulations matter? The effects of cross-listing on analysts’ coverage and forecast errors. Rev. Account. Finance 7 (3), 285–307.
Alexander, G., Eun, C., Janakiramanan, S., 1987. Asset pricing and dual listing on foreign capital markets: a note. J. Finance 42, 151–158.
Alexander, G., Eun, C., Janakiraman, S., 1988. International listings and stock returns: some empirical evidence. J. Financ. Quant. Anal. 23, 135–151.
Amihud, Y., Mendelson, H., 1986. Asset pricing and the bid-ask spread. J. Financ. Econ. 17, 223–249.
Association of Chartered Certified Accountants (ACCA), 2012. Available at: http://www.accaglobal.com.
Bank of New York Mellon, 2013. Available at: http://www.adrbnymellon.com/.
Bekaert, G., Harvey, C., 1995. Time-varying world market integration. J. Finance 50 (2), 403–444.
Botosan, C., 1997. Disclosure level and the cost of equity capital. Account. Rev. 72 (3), 323–349.
Brennan, M., Hughes, P., 1991. Stock prices and the supply of information. J. Finance 46, 1665–1691.
Bris, A., Cantale, S., Nishiotis, G., 2007. A breakdown of the valuation effects of international cross-listing. Eur. Financ. Manage. 13 (3), 498–530.
Bris, A., Cantale, S., Hrnjic, E., Nishiotis, G., 2012. The value of information in cross-listing. J. Emp. Finance 18, 207–220.
Burns, N., Francis, B., Hasan, I., 2007. Cross-listing and legal bonding: evidence from mergers and acquisitions. J. Bank. Finance 31, 1002–1031.
Cheung, W., Fung, S., Chung, R., 2015. The effects of stock liquidity on firm value and corporate governance: endogeneity and the REIT experiment. J. Corp.
Finance 35, 211–231.
Chung, C., Liu, C., Wang, K., Zykaj, B., 2015. Institutional monitoring: evidence from the F-score. J. Bus. Finance Account. 42 (7 & 8), 885–914.
Coffee, J., 1999. The future as history: the prospects for global convergence in corporate governance and its implications. Northwest. Univ. Law Rev. 93,
641–708.
Cowen, A., Groysberg, B., Healy, P., 2006. What types of analyst firms make more optimistic forecasts? J. Account. Econ. 41, 119–146.
Cubbin, J., Stern, J., 2005. Regulatory Effectiveness and the Empirical Impact of Variations in Regulatory Governance. World Bank Policy Research Working
Paper.
Daske, H., Hail, L., Leuz, C., Verdi, R., 2008. Mandatory IFRS reporting around the world: early evidence on the economic consequences. J. Account. Res. 46,
1085–1142.
Deloitte report on GDR issuers, 2008. Available from: http://www.deloitte.com/view/en GX/global/insights/deloitte-research/index.htm.
Dimson, E., Marsch, P., Staunton, M., 2009. The Triumph of the Optimists: 101 Years of Global Investment Returns. Princeton University Press.
Doidge, C., Karolyi, G.A., Stulz, R., 2004. Why are foreign firms that are listed in the US worth more? J. Financ. Econ. 71, 205–238.
Doidge, C., Karolyi, G.A., Stulz, R., 2009. Has New York become less competitive than London in global markets? Evaluating foreign listing choices over
time. J. Financ. Econ. 91, 253–277.
Easley, D., O‘Hara, M., 2004. Information and the cost of capital. J. Finance 59 (4), 1553–1583.
Easton, P., Taylor, G., Shroff, P., Sougiannis, T., 2002. Using forecasts of earnings to simultaneously estimate growth and the rate of return on equity
investment. J. Account. Res. 40 (3), 657–676.
Errunza, V., Losq, E., 1985. International asset pricing under mild segmentation: theory and test. J. Finance 40, 105–124.
Errunza, V., Miller, D., 2000. Market segmentation and the cost of capital in international equity markets. J. Financ. Quant. Anal. 35 (4), 577–600.
Fama, U., French, K., 1992. The cross-section of expected stock returns. J. Finance 47 (2), 427–465.
Fang, L., Yasuda, A., 2014. Are stars’ opinions worth more? The relation between analyst reputation and recommendation values. J. Financ. Serv. Res. 46
(3), 235–269.
Foerster, S., Karolyi, G.A., 1999. The effects of market segmentation and investor recognition on asset prices: evidence from foreign stocks listing in the
United States. J. Finance 54, 981–1013.
Foerster, S., Karolyi, G.A., 2000. The long-run performance of global equity offerings. J. Financ. Quant. Anal. 35, 499–528.
Hail, L., Leuz, C., 2009. Cost of capital effects and changes in growth expectations around US cross-listings. J. Financ. Econ. 93, 428–454.
Hope, O.-K., 2003. Disclosure practices, enforcement of accounting standards, and analysts’ forecast accuracy: an international study. J. Account. Res. 41/2,
235–272.
International Finance Corporation (IFC), 2012. Emerging Markets Report.
Kang, J.-K., Stulz, R., 1997. Why is there a home bias? An analysis of foreign portfolio equity ownership in Japan. J. Financ. Econ. 46 (1), 3–28.
Karolyi, G., 1998. Why do companies list abroad? A survey of the evidence and its managerial implications. Financial Markets. Inst. Instrum. 7, 1–60.
Karolyi, G.A., 2006. The world of cross-listings and cross-listings of the world: challenging conventional wisdom. Rev. Finance 10, 99–152.
G Model
Karolyi, A., 2012. Corporate governance, agency problems and international cross-listings: a defense of the bonding hypothesis. Emerg. Mark. Rev. 13,
516–547.
Kaufmann, D., Kraay, A., Mastruzzi, M., 2010. The Worldwide Governance Indicators: a Summary of Methodology, Data and Analytical Issues: World Bank
Policy Research. Working Paper No. 5430.
Kim, O., Pinnuck, M., 2014. Competition among exchanges through simplified disclosure requirements: evidence from the American and Global
Depositary Receipts. Account. Bus. Res. 44 (1), 1–40.
London Stock Exchange (LSE) listing regime overview, 2013. Retrieved from:
http://www.londonstockexchange.com/companies-and-advisors/main-market/companies/primary-and-secondary-listing/listing-categories.htm.
London Stock Exchange (LSE), 2015. Guide to Listings, Retrieved from: http://www.lseg.com/download-our-guide-listing.
La Porta, R., Lopez-De-Silanes, F., Schleifer, A., Vishny, R., 1998. Law and finance. J. Polit. Econ. 106, 1113–1155.
La Porta, R., Lopez-De-Silanes, F., Schleifer, A., Vishny, R., 2000. Investor protection and corporate governance. J. Financ. Econ. 58, 3–27.
Lang, M., Lundholm, R., 1996. Corporate disclosure policy and analyst behavior. Account. Rev. 71, 467–492.
Lang, M., Lins, K., Miller, D., 2003. ADRs, analysts and accuracy: does cross-listing in the US improve a firm’s information environment and increase
market value? J. Account. Res. 41, 317–345.
Leuz, C., 2006. Cross-listing, bonding, and firms reporting incentives: a discussion of Lang, Raedy, and Wilson. J. Account. Econ. 42, 285–299.
Licht, A., 1998. Regulatory arbitrage for real: international securities regulation in a world of interacting securities markets. Va. J. Int. Law 38, 563–638.
Licht, A., 2000. Genie in a bottle?: Assessing managerial opportunism in international securities transactions. Columbia Bus. Law Rev. 51, 351–377.
Licht, A., 2001. Managerial opportunism and foreign listing: some direct evidence. Univ. Pa. J. Econ. Law 22, 325–347.
Licht, A., 2003. Cross-listing and corporate governance: bonding or avoiding? Chic. J. Int. Law 4, 122–141.
Loughran, T., Ritter, J., 1995. The new issue puzzle. J. Finance 50 (1), 23–51.
Lys, T., Soo, L., 1995. Analysts‘ forecast precision as a response to competition Journal of Accounting. Audit. Finance 10 (4), 751–765.
Merton, R.C., 1987. A simple model of capital market equilibrium with incomplete information. J. Finance 42, 483–510.
Miller, D., 1999. The market’s reaction to international cross-listing: evidence from depositary receipts. J. Financ. Econ. 51, 103–123.
Minkov, M., 2011. Cultural Differences in a Globalizing World. Emerald Group Publishing.
Reese, W., Weisbach, M., 2002. Protection of minority shareholder interests, cross-listings in the United States, and subsequent equity offerings. J. Financ.
Econ. 66, 65–104.
Ritter, J., 1991. The long-run performance of initial public offerings. J. Finance 46, 3–27.
Securities and Exchange Commission (SEC), 2006. Available at: http://www.sec.gov/news/speech/2006/spch111606cc.htm.
Saudagaran, S., Diga, J., 1997. Financial reporting in emerging capital markets: characteristics and policy issues. Account. Horizons 97, 41–64.
Siegel, J., 2005. Can foreign firms bond themselves effectively by submitting to US law? J. Financ. Econ. 75, 319–360.
State Street Global Advisors report, 2015. Available at: https://www.ssga.com/home.html.
Stulz, R., 1999. Globalization, corporate finance, and the cost of capital. J. Appl. Corp. Finance 12, 8–25.
UK Listing Authority (UKLA) (Financial Services Authority) Handbook, Chapter 18 Admission and continuing obligations requirements for GDR issues,
2015. Available from: http://fshandbook.info/FS/html/FCA/LR/18.

Kim 2016

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Kim 2016

Uploaded by

Copyright:

Available Formats

G Model

MULFIN-519; No. of Pages 20 ARTICLE IN PRESS

Contents lists available at ScienceDirect

Journal of Multinational Financial

The joint role of the bonding mechanisms and the reduction

E-mail address: Oksana.kim@mnsu.edu

2. Institutional details and related literature

2.1. GDRs compared to previously examined ADRs, and criticism of GDRs

2.2. Market segmentation hypothesis—origin and criticism

2.3. Bonding hypothesis—evidence in support and criticism

3.1. Firm value estimation

arit = rit − Rmt . (1)

ARt = arit . (2)

CARx−>y,i = art . (3.1)

CARx−>y,n = ARt . (3.2)

3.2. Market segmentation hypothesis test

CARit = ˇ0 + ˇ1 DSegm + ˇ2 DCL + ˇ3 DCL ∗ DSegm + [firmcontrols] + eit . (4)

3.3. Legal bonding hypothesis test

CARit = ˛0 + ˛1 IPOit + ˛2 DCL + ˛3 DCL ∗ IPOit + [firm controls] + it . (6)

3.4. Reputational bonding hypothesis test

CARit = 0 + 1 Accuracy/Follit + 2 DCL + 3 DCL ∗ Accuracy/Follit + [firm controls] + it . (7)

CARit = 0 + 1 DSegm + 2 Accuracyit + 3 DCL + 4 DSegm ∗ Accuracyit + 5 DSegm ∗ DCL+

CARit = 0 + 1 DSegm + 2 Follit + 3 DCL + 4 DSegm ∗ Follit + 5 DSegm ∗ DCL+

CARs around the cross-listing event

Fig. 1. Cumulative Abnormal Returns (CARs) around the cross-listing event.

5.1. Descriptive statistics and CARs around cross-listing

Panel A. Distribution of examined GDRs.

Country Number of companies Region

Bahrain 1 Middle East + Africa

Industry distribution (%):

[1] Banks and other ﬁnancial institutions 42.5

Panel B. Descriptive statistics for the main variables of the study.

5.2. Regression results

−24 −0.021 −0.002

CARit = ˇ0 + ˇ1 DSegm + ˇ2 DCL + ˇ3 DCL ∗ DSegm + +[firmcontrols] + eit .

CARs benchmarked on local market indices CARs benchmarked on UK market index

CARit = ˛0 + ˛1 BondRuleLawit + ˛2 DCL + ˛3 DCL ∗ BondRuleLawit + [firmcontrols] + it .

CARit = ˛0 + ˛1 IPOit + ˛2 DCL + ˛3 DCL ∗ IPOit + [firmcontrols] + it .

CARs benchmarked on local market indices CARs benchmarked on UK market index

Coefﬁcient Rule of Law IPO Rule of Law IPO

Constant 2.06*** 1.8*** 0.97*** 0.18

CARit = 0 + 1 Accuracy/Follit + 2 DCL + 3 DCL ∗ Accuracy/Follit + [firmcontrols] + it .

CARs benchmarked on local market indices CARs benchmarked on UK market index

Coefﬁcient Accuracy Following Accuracy Following

Constant 2.17*** 2.15*** 1.37*** 0.95**

CARit = ı0 + ı1 DSegm + ı2 BondRuleLawit + ı3 DCL + ı4 DSegm ∗ BondRuleLawit + ı5 DSegm ∗ DCL+

CARs benchmarked on local market indices CARs benchmarked on UK market index

Coefﬁcient Rule of Law IPO Rule of Law IPO

5.3. Additional tests

CARit = 0 + 1 DSegm + 2 Accuracyit + 3 DCL + 4 DSegm ∗ Accuracyit + 5 DSegm ∗ DCL+

CARit = 0 + 1 DSegm + 2 Follit + 3 DCL + 4 DSegm ∗ Follit + 5 DSegm ∗ DCL+

CARs benchmarked on local market indices CARs benchmarked on UK market index

Coefﬁcient Accuracy Following Accuracy Following

Gulf Finance House Bahrain June 4, 2007 December 7, 2007 Yes

CARs around the listing announcement day

Appendix A. Differential reporting obligations for exchange-listed ADRs and GDRs.

Type Pre-listing US Securities Listing Submitted Reporting Accounting SOX or

Criteria Premium Listing Standard Listing as GDR

ELIGIBILITY FOR ADMISSION:

Country Stock Market Index

Bahrain S&P Broad Market Index

Appendix D. Equity trading costs by country.

Country Trading costs as a percentage of the total consideration

Constant 2.06* 1.8* 0.97*** 0.18

Constant 2.17* 2.15* 1.37* 0.95