Does Cross-listing Increase Firm Value?
Evidence from Dutch Firms Cross-listing in the U.S. and within Europe
DUCO S.W. MERKENS
January 2007
UNIVERSITY OF GRONINGEN
Faculty of Economics
International Economics and Business Department
Under the supervision of
H. Gonenc
Does Cross-listing Increase Firm Value?
Evidence from Dutch Firms Cross-listing in the U.S. and within Europe
Abstract
The cross-listing of shares has been suggested as a mechanism to voluntarily “bond” oneself to a stricter legal and disclosure regime on a foreign exchange, thereby making it harder for managers and controlling shareholders to take advantage and expropriate minority shareholders (Coffee, 1999 and Doidge et al., 2004). This thesis has investigated the relationship between firm value (Tobin’s Q) and cross-listing of Dutch companies, using a sample of 133 listed firms on the Euronext Amsterdam from 1996-2005. This study supports the “bonding” hypothesis and finds evidence that Dutch firms that are cross-listed in the U.S. are worth more than Dutch firms that are not cross-listed in the U.S.
Dutch firms cross-listed within Europe experience no significant price premium compared to non-cross-listed firms. However, this study finds evidence of informational “bonding”, since Dutch firms that are cross-listed within Europe and followed by a higher number of financial analysts are worth more.
When looking around the time of cross-listing, this study finds that Tobin’s Q is higher before the listing, decreases at the year of listing and drops further after the cross-listing. This unanticipated result shows that firms that decide to cross-list are already worth more before the actual cross-listing. This might indicate that management strategically decides to cross-list when their shares are overvalued.
Keywords: Cross-listing; bonding hypothesis; investor protection; Analyst coverage
TABLE OF CONTENTS
1. Introduction... 4
2. Theoretical Background... 7
2.1 Earlier Theory ...7
2.1.1 Capital Market Segmentation Hypothesis...7
2.2 Recent Theories...9
2.2.1 Shareholder Base Hypothesis...9
2.2.2 The Role of Capital Markets and the Liquidity Hypotheses...10
2.2.3 Cross-listing and the Legal “Bonding” Hypothesis ...11
2.2.3.1 The Role of the U.S. Market ...
12
2.2.3.2 The "Bonding" Hypothesis and Cross-listing within Europe ...
17
2.2.3.3 The Actual Act of Cross-listing...
17
2.2.4 Cross-listing and the Informational “Bonding” Hypothesis ...18
2.3 Hypotheses ...20
2.3.1 The “Bonding” Hypothesis ...20
2.3.2 The Actual Act of Cross-listing ...21
2.3.3 Informational “Bonding” Hypothesis ...21
3 Data and Methodology... 21
4 Empirical Results ... 27
4.1 Dutch Firms Cross-listing within Europe...27
4.2 Dutch Firms Cross-listing in the U.S. ...30
4.3 The Actual Act of Cross-listing in the U.S. ...32
5 Conclusion ... 34
6 Implications and Recommendations ... 36
7 References... 38
8 Tables ... 41
1. Introduction
In the past 30 years, global economic integration has at incredible speed advanced cross-border financial flows, not only among industrial economies, but also between industrial and developing economies. Spurred by liberalization of capital controls, global cross-border capital flows have risen to reach nearly $6 trillion in 2004. One of the most interesting parts of this incredible financial globalization boom is the increasing number of firms that have decided to list their shares on foreign exchanges. This process of listing a share abroad, on a foreign exchange, is called cross-listing. Cross-listing doesn’t have to be a new offering, but can also be a listing with existing stocks newly cross-listed on a foreign exchange. During the last two decades we have seen some interesting developments in the world of cross-listings. First, a peak of 4.700 cross-listings in 1997, and since the peak a steady decrease to about 2.300 cross-listings in 2002 (Karolyi, 2006). Second, continental European exchanges were the primary destination for (Dutch) firms cross-listing their shares in the 1980s. U.S. exchanges have been the main destination to cross-list in the 1990s and in the early 2000s. But lately we have also seen a rise in popularity for the London Stock Exchange.
Many well-known Dutch multinationals like Philips, Unilever and Ahold have decided to cross-list; first within Europe and later on U.S. exchanges. But why do Dutch firms decide to cross-list? Increased access to capital? Operational and strategic advantages? An enormous amount of literature has been written on the cross-listing subject, for instance, theories that deal with capital market segmentation. One of the more recent arguments, on why firms cross-list, is that firms, that decide to cross-list in the U.S., have lower controlling shareholder agency costs, because greater legal controls and disclosure requirements on management and controlling shareholders reduce agency costs.
stricter legal and disclosure regime on a foreign exchange, thereby making it harder for managers and controlling shareholders to take advantage and expropriate minority shareholders (Coffee, 1999 and Stulz, 1999). Due to this stricter U.S. legal regime, Dutch cross-listed firms face a more likely threat of litigation or formal investigation into deceptive financial statements. Institutional differences make it also much easier to penalize firms in the U.S., for instance, the accounting rules in the U.S. specify a more narrow range of permitted accounting practices for earnings manipulation.
The support for the “bonding” theory can be divided in one group that looks at the stock market implications of cross-listing and a second group that looks directly at the relation between corporate governance and cross-listing on a foreign exchange.
Doidge et al. (2004, 2005) find empirical evidence for the “bonding” theory. They find a “cross-listing premium” for firms that have decided to list in the U.S. In other words, companies that are cross-listed in the U.S. are worth more. Also Reese and Weisbach (2002), King and Segal (2003) and Burns et al. (2006) find evidence of a “bonding” hypothesis. Wójcik et al. (2005) find higher corporate governance ratings for cross-listed firms than for companies that do not cross-list in the U.S.
Numerous studies have suggested that information considerations are key factors in cross-listing decisions. However, in literature there is not much empirical evidence on the relationship between cross-listing and the information environment of a company. Lang et al. (2003) find that non U.S. firms that are cross-listed on U.S. exchanges have greater analyst coverage (number of analysts following the firm). They also find that firm value is positively and significantly related to analyst coverage. An important aspect of cross-listing from a firm value perspective is the pre-commitment (informational “bonding”) for more disclosure and information. More analyst coverage leads to an improved disclosure and information environment. Coffee (2002) mentions that firms under the scrutiny of these “financial watchdogs” will more likely respect the shareholder rights and will less likely expropriate minority shareholders. This will result in a further decrease in the cost of following the company (lower information asymmetries) and consequently the cost of capital.
Using a sample of 133 listed firms on the Euronext Amsterdam as of December 31, 2005, this thesis has investigated the relationship between firm value and the cross-listing of Dutch companies from 1996-2005. Firm value is measured by Tobin’s Q; Dutch firms that trade at a “cross-listing premium” or are worth more have a significantly higher Tobin’s Q than Dutch firm that do not cross-list.
This thesis finds that Dutch firms that are cross-listed in the U.S. are worth more. This result is in line with the findings of Doidge et al. (2005) who find that the “cross-listing premium” ranged from 12,9 to 53,9 percent with an average valuation premium of 23,6 percent in the Netherlands from 1997 to 2004. Pooled regressions results show a higher and significant premium for firms that are traded outside an exchange, when cross-listed firms are divided into exchange (NYSE, NASDAQ, AMEX) and non-exchange listed firms (SEC Rule 144a and OTC). When looking around the time of cross-listing in the U.S. this thesis finds that Tobin’s Q is higher before the listing, decreases at the year of listing and drops further after the cross-listing. This unanticipated result shows that firms that decide to cross-list are already worth more before the actual cross-listing, potentially indicating that management strategically decides to cross-list when their shares are overvalued. Dutch firms cross-listed within Europe experience no significant price premium compared to firms that are not cross-listed. However, this study finds evidence of informational “bonding”, because Dutch firms that are cross-listed and followed by a higher number of financial analysts are worth more.
This paper contributes to the literature in the following ways. Although several studies have been done in Europe, to my knowledge, a study on the “cross-listing premium” for Dutch firms cross-listing within Europe has never been done before. Secondly, I also look at valuation premiums for Dutch firms listed on the Euronext Amsterdam that have decided to cross-list in the U.S., not only from 1997-2004 (Doidge et al., 2005), but from 1996-2005. More importantly I include smaller firms (no threshold for companies smaller than $100 millions). Thirdly, I also examine whether Dutch cross-listed firms with better information environments are worth more (informational “bonding”). Finally, numerous studies have found the existence of a “cross-listing premium”, but only one study (Clarkson et al., 2006), although methodologically different, has studied whether the premium was created through cross-listing.
2. Theoretical Background
When companies have ideas and opportunities but need capital and resources, while global investors have capital and resources but need ideas and opportunities to grow their capital, then listing on a stock exchange is an option for both investors and companies to bridge capital and opportunities. As the world economy progresses and globalizes, more companies operate globally and an increasing number of companies choose to raise capital through equity issues beyond the borders of their home market. According to the OECD in 1995, private sector cross-border capital flows in equities have risen to 35 percent of the total flow in all securities, as compared with only 5 percent in the early 1980s (Karolyi, 1998). According to the Federal Reserve Board in 2002, the gross value of cross-border equity trades was 80 percent of worldwide equity market capitalization. Equity financing has obviously become an important vehicle in international cross-border capital flows (international cross-border capital flows are positively related to financial market liberalizations, which are in turn, associated with higher real per capita growth (Bekaert et al., 2001). An easy method for cross-border equity financing is through (direct) listings of shares, cross-listings, on a major world stock exchange.
There is a vast amount of literature on international cross-listings, especially on foreign firms cross-listing in the U.S. This literature review will start with a review of the old idea about cross-listings and continue with new ideas that incorporate the ideas on corporate governance. Also a distinction in this literature review will be made between cross-listing in the U.S. and cross-listing within Europe.
2.1 Earlier Theory
2.1.1 Capital Market Segmentation Hypothesis
firm’s risk and therefore the cost of capital of the firm falls.” In other words, those firms where the marginal benefit of a lower cost of capital is higher than the marginal cost of listing decide to undertake a U.S. listing. Consequently, the firm that is higher valued, keeping the expected cash flow equal, is the firm that has a cross-listing and makes it shares accessible for foreign investors.In conclusion, the lower cost of capital can come from diversification gains and segmentation gains, i.e. similar assets in different markets have different prices (excluding transaction costs).
Several event studies, especially event studies on foreign firms listing in the U.S., but also on Dutch firms listing in the U.S., have tested the capital market segmentation hypothesis. Perotti and Cordfunke (1997) find that Dutch companies experience abnormal returns when listed on a U.S. exchange. They experience, when listed on the NYSE, on average an increase in value of 1,2 percent. Other event studies, focused on listing in the U.S. find that the announcement of a listing is accompanied by a significant abnormal return that is higher for firms from emerging markets but also existent for firms from developed countries that list on a global stock exchange (Alexander et al. (1988), Foerster and Karolyi (1993, 1999), Jayaranam et al. (1993), and Miller (1999). For instance, Miller (1999) documents an abnormal return of 1.2 percent for the announcement effect of a U.S. listing. Firms from emerging markets experience an abnormal return of 1.5 percent. Firms from emerging markets that cross-list on an exchange experience an abnormal return of 2,6 percent. Also a price run-up preceding cross-listings is documented, and a price run-down after the cross-listing. Foerster and Karolyi (1999) find positive results (0,15 percent) per week for the year prior the cross-listing and negative results after the cross-listing.
René Stulz, was the first researcher who opposed to the conventional wisdom of the market segmentation hypothesis. He outlined in his 1999 article titled “Globalization, Corporate Finance, and the Cost of
Capital” five inconsistencies with the capital market segmentation hypothesis.
The first difficulty is that the event-study abnormal returns are extremely small (see above), compared to the large changes in the cost of capital implied by shifting market risk exposures. The run-up prior to listing (market anticipating an U.S. listing) has to be impressively big to compensate for the only small abnormal return, but the run-up does not seem to be big.
integrated in American capital markets. For example, Mittoo (1992) finds that Canadian firms experience the same “cross-listing premiums” for U.S. listings as European and Asian firms, although the North-American capital market should be more integrated.
Fourth, the market segmentation hypothesis cannot clarify the time-series pattern of listings in the second half of the 1990s; we have seen an increase in listings, instead of a decrease in cross-listings. Greater integration of markets should cause the company’s cost of capital to be more determined on a global level (firms cannot lower their cost of capital by cross-listing abroad), and should therefore decrease the demand for new cross-listings.
The final difficulty is that the market segmentation hypothesis cannot explain why the share-price reactions are largest for exchange-listed firms (Miller, 1999) opposed to those that use a private placement (Foerster and Karolyi, 1999).
Due to the inconsistencies of the market segmentation hypothesis, Stulz (1999) has developed a new “game plan” for researchers to explain the benefits of cross-listings better. This “game plan” focuses on corporate governance issues. Corporate governance issues might cause the limited capital market reactions observed during cross-listing. Of course this new “game plan” is in line with the increased attention by economists to explore the implications of agency cost, i.e. the financial loss to the “principal” (the shareholders) due to the abuse of discretion by the “agent” (the manager) hired to run the firm. Next Stulz (1999) states that there are six important mechanisms through which globalization can influence investors to monitor management better and where future research on cross-listing should aim at: an independent and active board of directors, certification in the capital markets (investment bankers and auditors (reputation risk)), stronger legal protections of minority investors (e.g. SEC regulations), an effective and vigorous market for corporate control, rigorous disclosure requirements to provide fuller financial information, and finally active large shareholders (e.g. pension funds). These mechanisms to monitor management better laid the groundwork for new ideas on the benefits of cross-listing.
Next I will discuss four new perspectives, after Stulz’s critique, on the benefits of cross-listing. I will focus the most on the two final and most recent perspectives.
2.2 Recent Theories
2.2.1 Shareholder Base Hypothesis
therefore lowers the firm’s cost of capital. Cross-listing shares can also add firm visibility and firm-awareness. Doidge et al. (2001) state that “the critical difference between the shareholder base hypothesis and the capital market segmentation hypothesis is that it recognizes that simply listing shares in the U.S. does not necessarily imply that U.S. investors will become significant shareholders. U.S. investors may choose to mostly stay away from shares issued by firms they know little about. For these firms, the risk premium benefit emphasized by the asset pricing literature is not relevant and a listing in the U.S. may have little value.” Thus, the shareholder base might increase when a firm decides to cross-list its shares, but not necessarily. For instance, certain investment funds are restricted to invest only in firms that are from their home market and don’t invest in firms that have their headquarters outside the U.S. If shareholders decide to stay away from the newly issued shares, then the cost of capital will not fall as much and that might be the reason why we only see small abnormal returns in event studies.
2.2.2 The Role of Capital Markets and the Liquidity Hypotheses
The role of capital markets hypothesis is based on the theory that firms that cross-list in the U.S. gain access to capital and increase in value while bypassing underdeveloped capital markets. Fanto and Karmel (1997) found that managers of a number of cross-listed firms in the U.S. cited increased access to new capital as one of the most important motivations for pursuing an overseas listing. Lins et al. (2000) have investigated whether firms that list in the U.S. become less credit-constrained. They find that firms become less credit-constrained and show that the firm’s investment opportunities depend less on the firm’s cash flows after the U.S. listing than before. Ammer et al. (2004) find that U.S. investors increase their holdings in foreign stocks that initiate a listing in the U.S., thereby helping the firm’s capital raising activity.
Global capital markets play an important role as a monitoring device for firms that decide to raise capital through the cross-listing of shares. The three monitoring mechanism advanced by Stulz (1999) of global capital markets are: the certification role of investment bankers for firms that decide to raise capital through a cross-listing, outside active shareholders with sufficient resources like institutional investors, and finally the market for corporate control (e.g. acquisitions). Obviously, all these monitoring devices play an important role to facilitate access to new capital.
investor. Hence, investors demand a higher expected return when investing in stock with higher trading costs. This higher expected return implies a higher cost of capital to the firm. An improvement of market liquidity after cross-listing can be measured by analyzing turnover ratios, turnover volume and bid-ask spreads on the domestic market. Foerster and Karolyi (1998) show that the beneficial effects of an increased competitive environment from another exchange through cross-listing can lead to increased trading activity by narrowing the spreads on the domestic market. Karolyi (2006) acknowledges that research in liquidity predates Stulz’s critique: “However, with better quality transactions data available for more markets around the world (beyond the U.S.) and with more rigorous research methodologies at our disposal, we are learning that liquidity changes (spreads, volume, volatility) for newly cross-listed firms may very well be related closely to changes they incur in the information environment, the firm’s ownership structure and perhaps even corporate governance systems.”
The liquidity hypothesis has received some support very recently. John Thain, the NYSE’s Chief Executive, supports the thought that cross-listing increases liquidity when firms decide to cross-list. He rationalizes the potential merger between the NYSE and Euronext on the believe that the merger will increase liquidity by giving firms the opportunity to automatically cross-list its shares on both exchanges (resulting in more trades for the exchange).
2.2.3 Cross-listing and the Legal “Bonding” Hypothesis
pecuniary benefits, be non-pecuniary benefits like the prettiness of the secretary and the ability to fly the corporate jet on private occasions). The trade-off between restricted access to external finance (loosing positive NPV projects) and private benefits pose a quandary for managers and controlling shareholders. This trade-off theory can be explained by a relatively simple example, think of a controlling shareholder (owning 50% of the shares) that has private benefits of control of 50 million per year. Although there are growth opportunities of 750 million in present value, the controlling shareholder prevents the firm from raising external capital because the shareholder cannot convince capital providers that it will stop or lower its consumption of private benefits of control. But if the controlling shareholder can make a commitment (“bond”) to consume only 25 million a year, then capital providers are willing to fund growth, and firm value will increase by 1 billion (750 million in present value and 250 million (discount rate ~ 10 %) of private benefits in present value). The controlling shareholder (discount rate ~ 10 %) loses 250 million but gains 500 million in present value. If managers or controlling shareholders “bond” themselves to avoid taking private benefits or in order to signal to minority shareholders that the firm or its management will not expropriate the firm’s resources it can also decide to issue debt to be subject to monitoring from debt-issuers (a bond also legally obliges the firm to repay interest and the principal on specified dates), but more importantly in this study it can decide to cross-list its shares abroad to make a promise to forego private benefits in the eyes of minority shareholders credible. Thus management receives a net utility gain as the increase in utility from growth opportunities outweighs the utility reduction of foregoing private benefits of control.
2.2.3.1 The Role of the U.S. Market
The NYSE was the first stock exchange were listing was used as a mechanism to bond to a certain governance mechanism in a domestic U.S. context. Jeffrey Gordon (1988) states: “insiders who seek to lower the cost of capital will find it valuable to bond a promise that the firm's single class capital structure will not be renegotiated.… the NYSE (one-share-one-vote listing) rule is the only secure “bond” available for such a promise.” In other words an American firm can make a credible signal by pursuing a listing on the NYSE.
In a 1999 article titled “The Future as History: The Prospects for Global Convergence in Corporate
Governance and its Implications,” Coffee mentions that not only American firms can “bond” themselves
opposite impulse: namely, to opt into higher regulatory or disclosure standards and thus to implement a form of "bonding" under which firms commit to governance standards more exacting than that of their home countries.” Coffee (1999) further claims that firms with dispersed ownership will list in markets that better protect minority shareholders, while markets with a low corporate governance regime will attract firms with a concentrated ownership shareholder structure.
When companies decide to list their shares in the U.S. they can either use American Depository Receipts (ADRs) or directly list on a stock exchange.1 Few European firms cross-list directly, while the requirements for listing directly are basically similar for ADRs (Reese and Weisbach, 2002). Depositary receipts were established in the 1927 to help U.S. investors wishing to purchase shares of Non-U.S. firms. A Depository Receipts is a negotiable instrument, issued by a U.S. depository bank, and represents one or more shares of a foreign stock or a fraction of a share, evidencing ownership of a Non-U.S. company. Especially, the ease of trading, clearing and settling are attractive for investors desiring to acquire securities issued by firms outside the investor’s home market. Other often mentioned benefits for investors of ADRs are that: it facilites diversification into Non-U.S. securities, it allows for comparison of securities of similar companies trading in the investor’s home market and it allows for lower dividend tax rates for exchange-listed ADRs.
There are four different types of ADRs that can be issued (See appendix II). ADRs are not all required to register with the Securities and Exchange Commission (SEC), but all ADRs must comply with the “Securities Act” and the “Exchange Act”.2 Issuers can broaden their shareholder base with existing shares by either Over-The-Counter (Level I) or Exchange Traded (Level II) shares. Level I ADRs are part of an unlisted program and are traded over-the-counter. Stocks are traded outside stock exchanges and are quoted in the Pink Sheets and/or on the OTC bulletin board. Level I ADRs are not subject to any SEC registration and don’t have to comply with the U.S. reporting requirements. Level II ADRs are part of a listed program on a recognized U.S. exchange and have to comply with all the U.S. reporting requirements (IFRS vs. U.S. GAAP, including timely submission of financial statements) and full SEC registration. Level II ADRs are traded on the NYSE, NASDAQ or the AMEX exchange. There is also a possibility for ADRs to raise capital with new shares by either a Public Offering (Level III) or a Private Placement (Rule 144a). Level III ADRs are offered and listed on a recognized U.S. exchange and Rule 144a are private placements to so-called Qualified Institutional Buyers. The former are traded on the NYSE, AMEX or the
1
In this paper no difference is made between ADRs and Global Depositary Receipts or between sponsored and unsponsored ADR (GDRs are DRs that are offered to investors in two or more markets outside the issuer’s home market).
2
NASDAQ exchange, the latter are quoted on PORTAL in the U.S. Level III face full SEC registration and also have to comply will all the U.S. reporting requirements.
According to Citigroup3 a registered IPO has several advantages compared to a 144a IPO. The advantages are: access to a wider investor base, the possibility of having an acquisition currency, and that shares are easily tradable post offering. A disadvantage might be the SEC registration process which is time consuming. The slow registration process might delay the access to a market that is receptive today. Level II and Level III ADRs also have to comply with stricter corporate governance requirements, which are also stricter than the Dutch corporate governance requirements, e.g. a majority of independent directors in the Supervisory board, the empowering of non-management directors, corporate governance committees and audit committees. The legal requirements, strengthened by the Sarbanes-Oxley Act, place tremendous pressure on companies to meet the (corporate governance) requirements, e.g. the requirements for attestations by officers and auditors on the soundness of the firm’s internal controls. However Licht (2003) argues and wonders to what extent U.S. securities laws are actually enforced against foreign firms listed in the U.S. Siegel (2005) shows that, in general, SEC actions against cross-listed firms have been enforced rarely and ineffective. He finds detailed evidence from cross-listed firms from Mexico that U.S. law enforcement neither deterred nor punished a group of controlling shareholders after stealing billions of dollars from their respective firms.
The number of empirical studies, interest and support for the “bonding” theory is still growing. Miller (1999) finds a higher announcement-day share price reaction for exchange listings versus SEC Rule 144a private placement and OTC listings. Coffee (2002) argues that this is an important finding because there are different legal and reporting requirements for the different types of listings. (Exchange listings impose tighter constraints on controlling shareholders while the constraints are lower for OTC and Rule 144). Also Foerster and Karolyi (1999) find that Asian firms experience a longer-run price reaction and less-dramatic post-listing declines, which, Coffee (1999) mentions is consistent with the corporate governance structure in numerous Asian countries where minority shareholders are more likely to be abused by controlling shareholders.
Korczak and Lasfer (2005) find that insiders of cross-listed companies will less likely trade on private information because of the “bonding” hypothesis and a stricter enforcement regime in the U.S. By studying the composition and post-listing behavior of foreign firms that cross-list in the U.S. Reese and Weisbach (2002) also find evidence for the “bonding” hypothesis. To evaluate the decision on why firms decide to cross-list they surveyed 1.158 cross-listings relative to a benchmark sample of 17.381 domestic firms. They find that the legal regime (English common law vs. French civil law), where the firm is
incorporated, influences the possibility of cross-listing: French civil law companies were more likely to list in the U.S. and on a U.S. exchange. Also one of their principal findings is that the increase in equity offerings outside the U.S. was especially strong for companies incorporated in a weaker legal system after cross-listing in the U.S. This is consistent with the legal “bonding” hypothesis and inconsistent with the capital market segmentation hypothesis (111 secondary offerings within two years after listing against 46 offerings before cross-listing).
Doidge et al. (2004) find empirical evidence for the “bonding” theory. They find at the end of 1997 that, foreign companies that are cross-listed in the U.S. have a valuation premium of 16,5 percent. In the Netherlands the premium ranged from a negative 24,7 percent to a positive 27,9 percent for exchange listed firms with an average valuation premium of 18,5 percent. They surveyed 712 cross-listings and benchmarked them with 4.078 globally listed companies, where Tobin’s Q (1997) was used as a measure to calculate the “cross-listing premium”.
Their main findings are that the “cross-listing premium” is related to future growth opportunities, the “cross-listing premium” is higher for companies from countries with poorer investor protections, and lastly the “cross-listing premium” is greater for exchange listed firms than OTC and SEC Rule 144a private placements. The “cross-listing premium” for exchange listed firms is even 37 percent and persists after controlling for country and firm specific characteristics.
Doidge (2004) identifies 137 dual class shares (different classes of shareholders with different rights) from 20 countries relative to a benchmark sample of 745 domestic firms. He finds evidence that supports the “bonding” hypothesis. He also finds that foreign firms that pursue a listing on a U.S. exchange have voting premiums that are on average 43 percent lower than other Non-U.S. firms that do not cross-list. This finding shows that cross-listing decreases the level of expropriation of minority shareholders by controlling shareholders and decreases the private benefits of control. Although Doidge (2004) finds a premium for exchange listed firms, he does not find a “cross-listing premium” for Level I and Rule 144a listings and concludes hence that the “bonding” hypothesis cannot be applied to Level I and Rule 144a listings.
Wójcik et al. (2005) find higher corporate governance ratings for companies listed in the US than for companies from the same countries that do not cross-list in the U.S.
2.2.3.2 The “Bonding” Hypothesis and Cross-listing within Europe
To understand whether there exist a “cross-listing premium” in relation with the “bonding” hypothesis within Europe we need to take a look at corporate governance mechanism and legal and financial disclosure requirements in different countries and on different exchanges.
Wójcik et al. (2005) find no significant relationship between corporate governance and cross-listing in Europe.
Relating financial disclosure, Cantale (1998) finds that the abnormal return, when cross-listing, is the highest for the NYSE, lower for the London Stock Exchange (LSE) and the lowest, but still positive for the Paris Stock Exchange. He mentions that the abnormal returns can be explained by disclosure requirements, which were stricter on the NYSE, less strict on the LSE, and even less on the Paris Stock Exchange. Another consideration is that the European legislation regulates cross-listings under the “mutual recognition principle”, in other words the listing requirements in one country should also be sufficient to list in another country (no divergence in corporate governance environments). In conclusion, there are sparse and diverse results for the existence of a “cross-listing premium” within Europe. Most of the legal part of “bonding” should be the same across Europe, because ultimately all legal problems can be brought to the Supreme Court of the European Union. Removal of capital market controls and a homogeneous regulatory regime in the E.U. will converge capital markets even more and will result in greater transparency and tighter disclosure requirements. This will make the access to capital markets easier and will provide more financing options for firms. Ultimately the lower cost of capital is not only advantageous for shareholders but for the economy as a whole. Thus the opportunity to pre-commit or “bond” managers not to consume excessive private benefits through cross-listing will diminish through the integration of capital markets within Europe.
2.2.3.3 The Actual Act of Cross-listing
Numerous studies show that cross-listed companies are worth more than non-cross-listed companies in the U.S., but fail to relate the “cross-listing premium” to the actual act of cross-listing. If the “bonding” hypothesis is correct, we should see an increase in the firm’s valuation around the actual act of cross-listing. If the “cross-listing premium” is not created at the time of cross-listing, then the premium is not really a “cross-listing premium”, but a premium that is already inherent in these firms.
during cross-listing, i.e. greater analyst coverage and increased forecast accuracy. Wójcik et al. (2005) find higher corporate governance ratings for European companies cross-listing in the U.S., but more importantly here also higher corporate governance ratings before cross-listing and even more so after cross-listing.
Clarkson et al. (2006) show, using a sample of Asian companies cross-listing on U.S. and European exchanges, that the actual act of cross-listing is not always value creating. They find evidence that the operating performance of firms drops significantly after cross-listing and argue that firms strategically choose to cross-list when they are traded at their peak value. Although, after controlling for the downward trend in value after listing and the value before listing, they also find that firms that list on global exchanges do retain more of their valuation premium.
2.2.4 Cross-listing and the Informational “Bonding” Hypothesis
Wójcik et al. (2005) divide “bonding” into legal “bonding” and reputational “bonding”. They basically explain reputational “bonding” as a “bonding” mechanism where a firm can sort of signal its reputation through an open information environment with the investment community (investment banks, investors, auditors, rating agencies and others).
that can better assess and appreciate a company’s business and growth opportunities (Pagano et al. 2002). Coffee (2002) also mentions that informational “bonding” can also be achieved by the following of financial intermediaries such as underwriters, analysts and institutional investors. He mentions that firms under the scrutiny of these “financial watchdogs” will more likely respect the shareholder rights and will less likely expropriate minority shareholders.
Lee and Swaminathan (2000) find that companies with more correct forecasts experience a lower implied cost of capital. Lang et al. (2003) state that “despite its theoretical importance, there is surprisingly little direct evidence on the relation between a firm’s information environment and cross-listing.”
The reason for this lack is that it is relatively hard to measure a firm’s information environment. Lee and Swaminathan (2000) take the characteristics of analyst forecasts as a measure for the information environment. Lang et al. (2003) use the accuracy of analyst forecasts but also the number of analysts following the firm. Why is there an increase in firm value when firms are followed by more analysts? When more analysts, but also more agents from the whole investment community (financial watchdogs), increase their attention and security around a firm then firm value can increase, because there is more security that the residual claim will actually flow to equity holders. Lang et al. (2003) find that Non-U.S. firms that are cross-listed on U.S. exchanges have greater analyst coverage and increased forecast accuracy in comparison to other Non-U.S. firms that are not cross-listed. They also find a change in analyst coverage and forecast accuracy around the time that firms decide to cross-list. Thirdly, they observe a change in firm value around the time of cross-listing that correlates with analyst following and forecast accuracy suggesting that firm value increases during cross-listing due to an improvement in the information asymmetry between the outsiders and insiders. Finally, firm value of a cross-listed firm is positively related to analyst coverage and forecast accuracy.
2.3 Hypotheses
2.3.1 The “Bonding” Hypothesis
Numerous articles in the cross-listing literature, dealing with the “bonding” hypothesis, already presented above, have focused on the U.S. They found that firms that cross-list in the U.S. are worth more, because of stricter corporate governance, legal and disclosure requirements.
Also Dutch firms cross-listed in the U.S. are regulated under these stricter corporate governance, legal and disclosure requirements. They also face a more likely threat of litigation or formal investigation into deceptive financial statements due to institutional differences between the U.S. and the Netherlands. For example, Huijgen and Lubberink (2002) state why penalizing firms for earnings manipulation is much easier in the U.S.: “first, the rules in the U.S. specify a more narrow range of permitted accounting practices while the Dutch accounting legislation leaves room for a considerable degree of flexibility, which makes it more difficult for the plaintiff to demonstrate earnings manipulation. Second, the SEC may start investigations on its own, while in the Netherlands the only way of alleging misleading reporting is to go to court which is perceived as cumbersome, time consuming and costly for the plaintiff.” Looking at the discrepancy between the consequences after fraudulent practices by upper management in the U.S. and in the Netherlands, we see that upper management is not harshly punished in the Netherlands. Sentences were mild for members of the Ahold board after being found guilty of fraud. Although hard to compare (Ahold was called the European equivalent of Enron), board members, found guilty of fraud have faced lifetime imprisonment in the U.S.
The collective evidence of previous research (Doidge et al. 2004, Coffee 2002, Burns et al. 2006) leads me to hypothesize that Dutch firms that are cross-listed in the U.S. are worth more than non-cross-listed Dutch firms. I come to the following hypothesis;
1a) Dutch firms that are cross-listed in the U.S. are worth more than non-cross-listed Dutch firms.
In this study also a distinction will be made whether the type of cross-listing, exchange or non-exchange, influences the “cross-listing premium” for Dutch firms. This brings us to the following hypothesis;
1b) The “cross-listing premium” for Dutch firms listed in the U.S. is higher for exchange than non-exchange cross-listings.
As can be seen from previous sub-sections, there has not been extensive research carried out on the “bonding” hypothesis and cross-listing decision within Europe. This leads to the following hypothesis; 2) Dutch firms that are cross-listed within Europe are worth more than non-cross-listed Dutch firms.
In contrast to my expectations for the “cross-listing premiums” in the US, and mainly due to the “mutual recognition principle”, I expect due to scarce and mixed evidence less or no significant effect of cross-listing within Europe on the value of Dutch firms.
2.3.2 The Actual Act of Cross-listing
We should see an increase in the firm’s valuation, if the “bonding” hypothesis is correct, around the actual act of cross-listing. I come to the following hypothesis;
3) Dutch firms experience an increase in market value around the time of cross-listing. In line with the “bonding” hypothesis, I expect to see an increase around the time of cross-listing.
2.3.3 Informational “Bonding” Hypothesis
3 Data and Methodology
The initial sample consists of all the Dutch firms that were listed on the Euronext Amsterdam as of December 31, 2005. To conduct this study, I obtain data on firms, from Amadeus and Thomson One Banker. To obtain the accounting data on firms, necessary in this study, this study extracts data from Amadeus, a database of 5.6 million public and private companies in thirty-one European countries published by Bureau van Dijk. Data on cross-listing locations within Europe were extracted from Amadeus. Price data was obtained from Thomson Financial’s Datastream. Missing accounting information is extracted from Thomson One Banker, an online financial database application of the Thomson Financial group that extracts its data from industry-leading sources such as First Call, Worldscope and Thomson SDC. Still missing accounting data was retrieved from Annual Reports. This study found a total of 140 firms listed on the Euronext in Amsterdam, as of December 31, 2005. After excluding those firms with no financial information in both databases, the final sample consists of 133 firms. The firms with a cross-listing in the U.S. were obtained from a joint-website from JPMorgan Chase Bank and Thomson Financial4 and cross-checked with data obtained from the Bank of New-York5. Other data on ADR programs were extracted from the annual report of the World Federation of Exchanges. In addition to this, some authors (e.g. Doidge et al., 2004) reduce their sample, by restricting their sample only to larger firms (firms with total assets of less than $100 million), and by excluding financial institutions.
Methodology
In this study a regression model relates the “cross-listing premium” to a dummy variable for cross-listing and several company control variables. Next a model will be presented that relates the “cross-listing premium” to firms that are cross-listed and have a higher number of analysts. The dependent variable is Tobin’s Q. The independent variables, besides a dummy variable for cross-listing and an interaction variable are the controlling variables: firm-size, firm debt structure and the sales growth rate.
Dependent Variable
“Cross-listing Premium” – Tobin’s Q
Tobin’s Q is a commonly used measure for firm value in academic literature. The Tobin’s Q ratio is the market value of a company divided by the replacement value of a firm's assets. In this study, Tobin’s Q is calculated as the book value of total assets, minus the book value of equity, plus the market value of
4http://www.adr.com
(Access date: 2006-08-31)
equity. For the denominator total assets are taken. I use a simplification of Q with book value of debt instead of market value of debt and total assets instead of replacement cost, assuming that market and book value of debt are equal and that the replacement costs of assets are similar as the assets in place. Independent Variables
Cross-listing Dummy
The cross-listing dummy equals one if a firm is cross-listed. Because multiple Dutch firms are cross-listed within Europe, I also subdivide the sample in three groups; one group with no cross-listings, the second group with 1 or 2 cross-listings and the last with 3 cross-listings or more. This study does the same for countries within Europe, because some firms are just multiple times listed on exchanges in the same country (especially in Germany). We should keep in mind that multiple cross-listing within one country can only legally “bond” oneself to one single institutional regime. Dutch firms that are cross-listed in the U.S. have a dummy that equals one if a firm is listed in the U.S. A cross-listing dummy variable is also used to indicate whether a firm is U.S. exchange (NYSE, NASDAQ, and AMEX) or non-exchange listed (SEC Rule 144a or OTC listing).
Number of Analysts
Lang et al. (2003) take the number of analysts as a proxy for the firm’s information environment. In this study the number of analysts is also used as a proxy for the firm’s information environment. The information environment variable is a dummy variable that takes a value of 1 if the number of analysts following is higher than the median number of 9 analysts in the sample, as of December 31, 2005. Analyst following is related to cross-listing, because cross-listing increases the information quantity obtainable to the market, which reduces the cost of following the company, and consequently could lead to additional coverage by investment analysts. Lang et al. (2003) also state why analyst following is related to cross-listing: “additional analyst following should also bring about more scrutiny, which, in the presence of agency costs, should improve firm value by increasing the cash flows that accrue to shareholders.” Finally, Lang et al. (2003) find evidence that firms that cross-list on U.S.-exchanges have greater analyst coverage, and, due to a better information environment, are worth more.
Interaction Variable: Number of Analysts * Cross-listing Dummy
following is higher than the median number of 9 analysts in the sample, as of December 31, 2005 (see above).
Company Characteristics
Pagano et al. (2002) perform a microeconomic analysis of the characteristics of European firms on their motivation of cross-listing their shares on a foreign exchange. They find that companies that list on other European exchanges and in the United States are bigger in size and more likely to be privatized recently. The size factor implies that the cross-listing decision involves fixed costs (underwriting costs and exchange costs) and thus economies of scales. They also find that European firms that decide to cross-list in the U.S. “pursue a strategy of rapid, equity-funded expansion”. In other words, companies increase their capital expenditures; there is an urge to fund growth and a need for foreign sales expansion. Accordingly, firms that decide to cross-list should have a high investment and growth rate. Ceteris paribus, the firm will raise capital at the time of cross-listing or rapidly after cross-listing. Higher sales growth should also lead to higher P/E ratios and a higher Tobin’s Q.
Growth Opportunities
Growth is widely known as an important value driver for the valuation of firms. As a proxy for growth opportunities, this study takes the yearly sales growth of firms. Doidge et al. (2004) find that the coefficient on sales growth is always positive and significant for the 1997-2004, except for the year 2002. The reason for including this controlling variable is that firms that decide to cross-list could be just firms that are growing faster and are therefore worth more and not worth more because of cross-listing.
Firm Size
As a proxy for size, the logarithm of total assets is taken. Doidge et al. (2004) find that cross-listed firms are larger than firms that do not cross-list. They find that cross-listed firms are four times bigger relatively than non-cross-listed firms in 1997 and almost six times bigger in 2004.
Leverage
cross-list are also worth more because of more efficient capital structures, e.g. reducing agency cost with the use of debt.
Multicollinearity
To avoid multicollinearity in the OLS regression, the independent variables are tested on their correlation. The correlation coefficients that are higher than 0.6 can be found in appendix III. Since, I also want to understand how the various independent variables impact the dependent variable, multicollinearity in the OLS regression is a problem. The variance inflation factors (VIF) are computed to measure the severity of any multicollinearity among the independent variables. The VIF measures the impact of multicollinearity among the independent variables in a regression model on the precision of estimation. If the VIF value is significantly greater than 5, it is very likely that there is severe multicollinearity. The variables with a VIF value higher than 5 will be excluded from the regression. The VIF value for the interaction variable number of analysts * listing dummy CCL 3 (which is equal to 1 if the firm is 3 times or more cross-listed in different countries)) is 6.47. Also the VIF values for the interaction variable for firms non-exchange cross-listed or non-exchange-listed in the US are too high with VIF values of respectively 13.87 and 8.24. The interaction variables in both regressions are excluded. The VIF value for the interaction variable for firms generally cross-listed in the U.S. takes a value of 11.16 and is also eliminated from the equation. In the regressions, where the interaction variable is excluded, the interaction variable is replaced by the dummy for the number of analysts following the firm.
Heteroscedasticity
The regressions are estimated with heteroscedasticity consistent standard errors. Regression
This study follows the models used in Doidge et al. (2004) for the primary specifications and uses Ordinary Least Squares (OLS) regression models of the following form.
Tobin’s Qt = á + â1 [Cross-Listing Dummy]t + â2 [log Total Assets]t + â3 [Leverage]t + â4 [Sales growth]t
+
tThe model for testing whether Dutch firms that are cross-listed and followed by more analysts within Europe are worth more is the following:
Tobin’s Qt = á + â1 [Cross-Listing Dummy]t + â2 [log Total Assets]t + â3 [Leverage]t + â4 [Sales growth]t
4 Empirical Results
Table 1 gives a breakdown on the number of Dutch firms cross-listed within Europe as of December 31, 2005. It can be seen that many Dutch firms have decided to cross-list in Germany in the past. It is interesting to see that Germany has been our number one trade partner and has been our main destination for cross-listings. This study finds, not surprisingly, that eminent Dutch multinationals like Royal Ahold, Royal Philips Electronics and Royal Dutch Shell are among the Dutch firms that are cross-listed the most within Europe.
Table 2 gives a breakdown on the number of listings in the U.S. as of December 31, from 1996-2005 and the U.S. listing status. Table 3a reports the descriptive statistics for the sample for Europe. They show that Dutch firms cross-listed within Europe have higher Tobin’s Q ratios, are larger, have higher leverage, higher sales growth and are followed by a higher number of analysts. The mean (median) Tobin’s Q for Dutch cross-listed firms within Europe is 1.22 (1.15), while a mean (median) 1.21 (1.17) Tobin’s Q is reported for non-cross-listed firms. Descriptive statistics for the U.S. for 1996-2005 are reported in table 3b. They indicate that Dutch cross-listed firms in the U.S. have higher leverage, are larger and are followed in 2005 by a higher number of analysts.
4.1 Dutch Firms Cross-listing within Europe
Table 4 shows the results for the OLS regressions in the year 2005 for Dutch firms cross-listing within Europe. All the specifications (t-statistics) are computed with heteroscedasticity consistent standard errors. Specification (1) investigates the Tobin’s Q of all firms in the sample on the Cross-Listing Dummy (CLD), taking a value of one if a firm is listed within Europe and a zero when the firm is not cross-listed. This study also controls for corporate characteristics affecting the “cross-listing premium” including sales growth, company size and leverage. The control variables used in this regression are growth opportunities (SG), Leverage (Lev) and Firm Size (LGTA). The CLD variable takes a value of 0.04, but is not significant. The adjusted R-squared is considerable with a value of 0.15.
less analysts (within Europe). The adjusted R-squared increases to 0.20. Specification (2) also shows that the CLD coefficient moves to a negative -0.02, although not significant, when controlling for firms that are cross-listed and followed by a higher number of analysts. This result potentially indicates that the “cross-listing premium” is more likely caused by informational “bonding” rather than legal “bonding” (Dutch governance laws are almost similar to European governance laws). This finding is in line with Lang et al. (2003) who find evidence that firms that cross-list on U.S. exchanges have greater analyst coverage, and, due to better information environments, are worth more.
Specification (3) investigates the regression when the sample is divided in three groups. It could be that Tobin’s Q is higher when firms are more than one-time cross-listed. The three classifications are zero cross-listings, one or two cross-listings and three or more cross-listings. When using these three classifications, the adjusted R-squared increases to 0.17. An insignificant result is found for one or two listings, but a significant result (5%) and a positive coefficient of 0.08 is found for three cross-listings or more. Thus, Dutch firms with three cross-listing or more within Europe trade at a premium. This “cross-listing premium” could be caused by additional corporate governance and compliance rules when listed on more exchanges. When a firm has to take up the strictest measure, but necessary in order to comply with an individual exchange, then the multiple listed firms will end up with an overall stricter corporate governance code. The stricter corporate governance code should decrease the chance of expropriation of minority shareholders by controlling shareholders. Hence, the stricter corporate governance code should improve firm value by increasing the cash flows that accrue to minority shareholders. Another potential reason for the higher premium is that firms from stricter governance regimes are under stronger reputational pressure. Wójcik et al. (2005) point out that reputational “bonding” works as a “bonding” mechanism where a firm can sort of signal its reputation through an open information environment with the investment community including investment banks, investors, auditors, rating agencies and others.
Specification (4) finds no significant results when two independent interaction variables are added.
explanatory power of the regression drops in specification (5) and no significant results are found. I cannot find evidence of an extra “bonding” effect, since firms that “bond” themselves three times to a legal and regulatory regime are not significantly worth more than firms that decide to cross-list only once or twice. In regression (6) the dummy variable for the number of analysts following is added as a proxy for the firm’s information environment. The coefficient variable is positive and significant, indicating that firms that have a better information environment are worth more. I also find no results for firms that are cross-listed in one or two other countries and three or more countries, when controlling for the number of analysts following.
4.2 Dutch Firms Cross-listing in the U.S.
Table 5 shows the results for Dutch firms that decide to cross-list in the U.S. and whether they experience “cross-listing premiums”. The coefficient on the cross-listing dummy in regression (1) is positive (0.11) and significant at the 10% level in the year 2005. When the predictor variable Analyst is added then the coefficient on CLD becomes more significant, but slightly lowers the “cross-listing premium”. The Analyst variable is really significant indicating that firms that are followed by more analysts are worth more. This result shows that after controlling for information environment effects that Dutch firms trade at a premium and that firms can legally “bond” themselves by cross-listing in the U.S. When observing specification (3-11) we see that the dummy variables for Dutch firms cross-listed in the U.S. are positive implying positive “cross-listing premiums”. But the premiums are rather small and the coefficients range from 0.05 to 0.20. The coefficient estimate is really significant at the 1% level in the year 1999 and 2003, significant at the 5% level in the year 2003 and also significant at the 10% level in the year 2001 and 2004. Especially the year 1999, in the middle of the US bull market, is a remarkable year with a coefficient that is positive (0.20) and significant (1%), and has a reasonably high adjusted R-squared of 0.23. In general, the explanatory power of these regressions is reasonable, ranging from 2.9 percent in 2002 to 23,3 percent in 1999, demonstrating that the specification is a reasonable benchmark for the analysis.
Overall, the dummy variable for cross-listed companies is positive and significant in many years, implying that Dutch firms that are cross-listed in the U.S. trade at a premium. The results are in line with the results of Doidge et al. (2005) who find that Dutch firms that are cross-listed in the U.S. are worth more. The higher trading premium could be the result of, as has been hypothesized in this study, tighter accounting standards, and better investor protection available to minority shareholders in the U.S., ensuing in lower controlling shareholder agency costs.
The next step in the analysis to determine whether Dutch firms experience “cross-listing premiums” is to look at Dutch companies that decide to cross-list on an U.S. exchange (NYSE, NASDAQ, and AMEX) or outside an exchange (OTC and SEC Rule 144a).
The coefficient on the cross-listing dummy in regression (1) is positive but not significant in the year 2005. When the predictor variable Analyst is added then the coefficient stays non significant but slightly lowers the “cross-listing premium”. The explanatory power of the regression grows after adding the independent interaction variable and increases to 22,7 percent. When observing specification (3-11), we see that the coefficients of the dummy variables for Dutch firms cross-listed in the U.S. are positive implying positive “cross-listing premiums”. But the premiums are rather small; the coefficient estimate ranges from 0.00 to 0.15 and is only two times significant at the 5% level and once at the 10% level. These findings are contrary to the results by Doidge et al. (2004, 2005). They find a Tobin’s Q premium for U.S. exchange listed firms and a significant “cross-listing premium” in most years. The higher protected U.S. exchange securities regime doesn’t result in a higher valuation for Dutch firms listed on a U.S. exchange.
The higher legal and reputational pressure on cross-listed firms after the introduction of the Sarbanes-Oxley Act (Sarbox) in 2002 should entail an extra “bonding” effect, i.e. the “cross-listing premium” should be higher and more significant after the year 2002. This study doesn’t find any indication of an extra “bonding” effect. An interesting observation through 1996-2005 is that “small seems to be beautiful”; firm size is almost every year negatively related to Tobin’s Q.
Table 7 reports the results for the pooled regressions. Specification (1) and (2) show that firms that are cross-listed in the U.S. whether on an exchange or outside an exchange are worth more. In specification (1) the coefficient on U.S. Listed is positive (0.11) and significant (1%), demonstrating that firms that are cross-listed in the U.S. experience higher Tobin’s Q ratios. When looking at the year dummies, when controlling for year effects, we see that the coefficient estimates from 2001 till 2004 are highly significant (1%) and negative with a dip in 2002, potentially showing effects of the Dutch bear market. In specification (2) the general U.S. listed variable is replaced for the exchange and non-exchange variables. The coefficient on the CLD US exchange variable and the CLD US Non-exchange are both positive, respectively 0.10 and 0.19, and both significant at the 1% level. These results are in line with the findings of Doidge et al. (2005) who also find that Dutch firms cross-listed in the U.S. trade at a premium from 1997 to 2004.
In general, this study finds and supports the hypothesis that cross-listed Dutch firms trade at a premium relative to firms that are not listed in the U.S, while no premium was found for Dutch firms cross-listing within Europe.
Overall the results imply that after controlling for other variables that the tougher U.S. securities laws do result in higher valuations for Dutch cross-listed firms in the U.S. The U.S. corporate governance regime increases the value of Dutch cross-listed firms through stricter and more regulation, supervision and enforcement. It seems that Dutch firms do experience a reduction in controlling shareholder agency costs as found in other literature after cross-listing in the U.S. Although the explanatory power of the regressions are reasonable, indicating that the specifications are reasonable benchmarks for the analyses, it should be noted that the explanatory power of the regressions are sometimes rather low implying that there are some essential country, capital markets and macroeconomic factors omitted. In general, the results show that “bonding” through cross-listing in the U.S. is a form of “piggybacking” on U.S. institutions (Black, 2001).
4.3 The Actual Act of Cross-listing in the U.S.
The pooled regression results, but also the year-by-year results of Dutch firms cross-listing in the U.S. indicate that Dutch firms that are cross-listed in the U.S. are worth more. But, in order to understand whether the premium is, in fact, related to the actual act of cross-listing and not already inherited in these companies, I scrutinize Tobin’s Q around the time of cross-listing.
The coefficient associated with CLD (t) is also positive (0.13), and significant at the 5% level, while the coefficients after the actual cross-listing are positive but insignificant. These results indicate that Tobin’s Q (t-1) is higher before the listing, decreases at the year of listing (t) and drops further after the cross-listing although firms remain worth more. These results are not anticipated because the results show that firms that decide to cross-list are already worth more before the actual cross-listing. Thus, firms decide to cross-list after a good period of time when they are traded at a premium. When including the year dummies, to control for year effects, the coefficient estimates on CLD (t-1 and t) fall and become less significant. This might indicate that firms decide to cross-list in years when stocks are higher valued by the market.
These surprising results are in line with the findings of Clarkson et al. (2006) who show while using a sample of Asian companies, listing on U.S. and European exchanges, that the actual act of cross-listing is also not always value creating. They also find that the operating performance of firms drops significantly after cross-listing. After digging deeper for the reason of the loss in value, they still find evidence that cross-listing firms appear to retain a long-term valuation premium. Miller (1999) also finds that the performance of global equity offerings in the U.S. is negative.
To fully evaluate whether there is value destruction at the time of cross-listing and whether management strategically decides to cross-list when their shares are overvalued, it would be interesting to look at multiple years before the firms actually decide to cross-list and not to only look at one year. Lang et al. (2003) mention that, for instance, by only looking at one year that “it can be argued that cross-listing firms will incrementally change their reporting strategies in advance of the formal adoption of U.S. GAAP” and will therefore already experience the incremental increase in value earlier.
5 Conclusion
This thesis has investigated the relationship between firm value and cross-listing of Dutch companies. Using a sample of 133 listed firms on the Euronext Amsterdam from 1996-2005, I examine whether Dutch firms that are cross-listed in the U.S. or within Europe are worth more, because of legal and informational “bonding”, than Dutch firms that decide not to cross-list.
A number of papers, as in Coffee (1999), Reese and Weisbach (2002), Doidge (2004), Doidge et al. (2004, 2005) and Wójcik et al. (2005) have argued that a practical way to “bond” managers not to consume excessive private benefits is to cross-list a company’s stock on a exchange that imposes higher legal and regulatory requirements than the firm’s primary exchange. In these papers cross-listing is regarded as a signaling mechanism to pre-commit to extract less private benefits from minority shareholders. In return controlling shareholders have not to forego positive NPV growth opportunities. These previous studies, for example Doidge et al. (2004), found empirical evidence for the “bonding” theory. They found, for instance, that foreign cross-listed firms in the U.S. experienced a valuation premium of 16,5 percent in 1997. Doidge et al. (2004) also found evidence of a “cross-listing premium” in the Netherlands.
In this study I argue that Dutch firms that are cross-listed in the U.S. have lower controlling shareholder agency costs, because greater legal controls and disclosure requirements on corporate insiders and controlling shareholders in the U.S. will reduce agency costs. As a result Dutch firms cross-listed in the U.S. will be worth more, i.e. the “cross-listing premium” (measured by Tobin’s Q) should be higher for Dutch firms that decide to cross-list in the U.S. Mainly due to the “mutual recognition principle”, I expect due to scarce and mixed evidence less or no significant effect of cross-listing within Europe on the value of Dutch firms.
