Corporate Governance and the value of cash
Evidence from Dutch Firms Cross-listed in the U.S.
RIENTS DE BOER
MAY 2011
Master’s Thesis under the supervision of:
Dr. H. Gonenc
UNIVERSITY OF GRONINGEN UNIVERSITY OF UPPSALA
Faculty of Economics and Business Faculty of Social Sciences
Corporate Governance and the value of cash
Evidence from Dutch Firms Cross-listed in the U.S.
ABSTRACT
By cross-listing shares in the U.S., foreign firms are assumed to come
under the scrutiny of U.S. laws resulting in limitations for those who
control the firm (insiders) to expropriate from minority shareholders
(Coffee, 1999, 2002; Stulz, 1999). Using a sample of 110 Dutch firms I
find that cash of firms cross-listed in the U.S. is valued higher in
comparison to that of non-cross-listed firms. This effect leads to a cash
valuation premium for firms cross-listed on an U.S. exchange (level 2 and
3 ADRs). For firms listed OTC and in accordance with SEC rule 144A a
large part of the higher valuation of cash stems from the increased amount
of analyst followings. In addition I find that the implementation of the
Sarbanes-Oxley Act in 2002 leads to a decrease in investors’ valuation of
cash for firms cross-listed in the U.S. This finding is especially strong for
firms that have to meet full SEC requirements due to their cross-listing on
an U.S. exchange.
Keywords:
Cross-listing, Agency theory, Cash valuation, Ownership
structure, Analyst coverage, Sarbanes-Oxley Act
Rients de Boer
SECTION 1 INTRODUCTION...4
SECTION 2 LITERATURE AND HYPOTHESIS ...7
2.1 The act of cross-listing ...7
2.1.1 Types of cross-listing...8
2.1.2 Reason for cross-listing...8
2.2 Cross-listing and corporate governance ...9
2.2.1 Internal Corporate Governance; Ownership structure ...9
2.2.1.1 Ownership Structure and cross-listing ...12
2.2.2 External Corporate Governance; Legislation...13
2.2.3 Corporate Governance in the Netherlands...15
2.3 Cash...16
2.3.1 Cash levels and cross-listing...17
2.3.2 The value of cash ...18
2.4 Informational bonding...19
2.5 Hypotheses development...20
2.5.1 Cross-listing and the value of cash ...20
2.5.2 Corporate Governance ...21
2.5.3 Information Environment...22
SECTION 3 DATA AND METHODOLOGY ...24
3.1 Sample...24
3.1.1 Sources for the necessary data ...24
3.2 Data type ...24
3.2.1 Dependent variable: Tobin’s Q ...24
3.2.2 Explanatory variables...25
3.2.3 Interaction variables consisting of two variables...26
3.2.4 Interaction variables consisting of three variables...27
3.2.5 Control variables...27
3.3 Description of the test ...29
3.4 Outliers ...30
3.5 Heteroskedasticity ...30
3.6 Multicollinearity...30
SECTION 4 EMPIRICAL RESULTS ...31
4.1 Results Panel data Regression...35
4.1.1 Cross-listing and the value of cash ...35
4.1.2 Corporate governance and the value of cash ...37
4.1.3 Cross-listing and the information environment ...38
4.1.4 SOX implementation ...40
SECTION 5 CONCLUSIONS...42
5.1 Cross-listing and the value of cash...42
5.2 Corporate governance cross-listing and the value of cash ...42
5.3 Information environment...43
5.4 SOX implementation...44
SECTION 1
INTRODUCTION
Throughout the former three decades, globalization in capital markets has evolved towards a market in which it has become easier to exchange ownership and to trade in securities from all over the globe. Until the recent global economic downturn there was an indistinct tendency; the world’s capital markets were undergoing great expansion, divergence, and integration. Between 1980 and 2007, the size of global cross-border capital flows enlarged
intensely, adding up to a total of $10.5 trillion in 2007(McKinsey, 2009).
Additionally, in the last decade the world has witnessed some huge scandals with regards to fraud, accounting manipulations and unethical behaviour (Rockness and Rockness, 2005). An important part of the enormous increase of cross-border capital flows are equities. Due to the huge capital market globalization firms have gotten the opportunity to raise capital across borders more easily. One way of doing this that has increased in popularity alongside the increase of the capital markets is cross-listing. This is the process of listing a firm’s stock on another exchange than its domestic one. Throughout the years the U.S. stock exchanges — the New York Stock Exchange (NYSE), NASDAQ and the American Stock Exchange (Amex) — have been the largest markets for Depositary Receipt (DR) trading, comprising 86% of all global DR trading value (BNY Mellon, 2008). A U.S. cross-listing has substantial impact on the corporate governance of a firm. By cross-listing in to the U.S., firms come under strict regulations which might decrease the opportunity for insiders to expropriate from minority shareholders. Building on this Coffee (1999, 2002) and Stulz (1999) created the ‘bonding hypothesis’ which states that foreign firms listed on U.S. stock exchanges come under the scrutiny of U.S. laws, resulting in restrains for those who control the firm (insiders) to expropriate from minority shareholders. The main cross-listing flow in the 1990s and in the early 2000s found its way to the U.S. markets (Pagano et al. 2002). Within Europe, after the U.K., firms from the Netherlands have been precursors when it comes to cross-listings (Benos and Weisbach, 2004).
This thesis will deal with the first and the last of these conflicts. Agency theory predicts that when control mechanisms are limited management or controlling shareholders might expropriate assets from minority shareholders and turn them in to private benefits. Myers and Rajan (1998) find that liquid assets are the easiest target for expropriation. Elaborating on it Dittmar and Mahrt-Smith (2007) note that cash holdings are the easiest accessed assets for management. Management can expropriate cash holdings relative easily since there is little scrutiny.
In this research the focus is on Dutch firms. Corporate governance and the protection of minority shareholders in the Netherlands are regarded weaker than that of the U.S. (World Bank and the International Finance Corporation, 2011). The advantage of focusing on firms from the Netherlands is the ability to see the implications cross-listing has on Dutch corporate governance. The focus of this thesis lays with the possibility of cross-listings of Dutch firms into the U.S resulting in a higher valuation of cash by investors.
Two aspects that are directly linked to cross-listing in the U.S. are the implementation of the Sarbanes-Oxley ACT (SOX) and exposure to a larger analyst community. In 2002 the U.S. implemented SOX in order to restraint fraud. As a consequence it has been found for companies from poorly-governed countries that are cross-listed in the U.S. the implementation of SOX has resulted in an increase in firm value (Litvak, 2008).
In addition, research discovered that cross-listing in the U.S. exposes companies to a large analyst community. Coffee (2002) states that under the scrutiny of a large amount of analysts, firms are more likely expected to value shareholder rights and less likely to expropriate from minority shareholders.
Since this research looks at the effects of Dutch firms cross-listed in the U.S., a specific overview is provided to the impact of exposure to a larger analyst community and the effect of SOX legislation on the value of cash.
Using a sample of 110 listed Dutch firms on Euronext Amsterdam at December 31, 2010, I find that cross-listing in the U.S. results in to higher valuation of cash reserves. In case of Dutch firms cross-listed on an U.S. exchange this valuation leads to a premium.
Regarding the effect of analyst' followings after cross-listing in the U.S. I find that a significant part of the increase in cash valuation for firms cross-listed OTC and by means of SEC rule 144A can be explained by the increase in analyst coverage after the moment of cross-listing.
In several articles it is mentioned that the Netherlands are a country with poor corporate governance (La Porta et al., 1998; Timmerman and Doorman, 2002; De Jong et al., 2005). For this reason it might be expected that the implementation of the Sarbanes-Oxley Act in the U.S. would have a positive effect on the valuation of cash of Dutch cross-listed firms in the U.S. However, I find proof for a reverse effect which is particularly strong for exchange cross-listed firms and can be explained by the high compliance cost of SOX on companies. In addition my results make me sceptical about the literature which states that the Netherlands is a country with poor investor protection. I belief that due to certain codes corporate governance in the Netherlands is becoming stronger from 2002 onwards.
The insights that may be derived from this thesis can be of interest since the research on linking cross-listing to cash reserve valuation is still in its preliminary phase. This paper complements the literature showing that cross-listing results into a higher valuation of cash (Frésard and Salva, 2010). As far as I know focussing on one country and its cross-listed and non-cross-listed firms has not been investigated yet when it comes to this topic. Focussing on one country allows me to specify my findings in relation to corporate governance in the Netherlands.
SECTION 2
LITERATURE AND HYPOTHESIS
Based on the main literature on cross-listing, the value of cash and the difference in corporate governance systems several hypotheses are formulated.
2.1 The act of cross-listing
Over the last 20 years a growing number of firms have sought to list their equity outside their country of origin. Since more places have become available for firms to list their shares, a consequence of the demand for cross-listings has been that major stock exchanges have been exposed to a growing level of competition to obtain listings (Chemmanur, 2003). Each of the major stock exchanges have tried to obtain as many listings and trading volume as possible. In 1997 the amount of cross-listings peaked at 4,700. After that there has been a steady decrease towards around 2,200 in 2002 (Karolyi, 2006). One of the reasons for this decrease might be the passage of the Sarbanes-Oxley Act (SOX) in the U.S. in 2002 (Doidge et al., 2009). SOX was intended to restraint fraud, however it also had a significant effect on global capital markets. Listing in the U.S. became less desirable due to the adoption of SOX. Firms and investors especially had problems with Section 302, which states that CEO and CFO must certify ‘‘the appropriateness of the financial statements and disclosures’’ and that the ‘‘financial statements and disclosures fairly present the operations and financial condition of the issuer.’’ In addition countless negative reactions came with the enclosure of Section 404, which states that management must issue an annual report with auditor attestation on the effectiveness of internal controls and procedures for financial reporting (Rockness and Rockness, 2005). The firms’ dislike comes from the point that complying with these rules brings high costs to firms due to the intensity of internal control systems (Zhang, 2005). Some firms complain that all the auditing and controlling leads their attention away from doing their business and therefore leads to loss of potential obtainable firm value (Solomon, 2004).
Although the overall motivation to cross-list seemed decreased by corporate governance regulations; these guidelines also provided several benefits. It has been found that for cross-listed companies from poorly-governed countries the implementation of SOX have resulted in an increase in firm value (Litvak, 2008). An explanation for this has been offered by the work of Coffee (1999, 2002) and Stulz (1999). They created the ‘bonding hypothesis’ stating that foreign firms listed on U.S. exchanges come under strict U.S. investor protection regulation which results in restrains for those who control a firm (insiders) to expropriate from minority shareholders.
According to the Bank of New York Mellon (BNY Mellon) over $2.4 trillion of American and Global depository receipts (ADRs and GDRs, respectively) were traded on exchanges across the globe in the first half of 2008 (BNY Mellon, 20081). Currently the amount of cross-listings is roughly 3.200 worldwide (JPMorgan, 20102 and BNY Mellon, 20103).
2.1.1 Types of cross-listing
Global cross-listings types may differ; though in the U.S. four categories of DRs exist. Levels 1 and 2 apply to cases where the DR is created using existing equity; in addition Levels 3 and 4 apply to cases where new equity is issued, such as an initial public offering (IPO). At level 1 shares are traded over the counter (OTC), this type of DR is the least costly and may result in to limited bonding, since firms do not have to comply with full SEC disclosure. Levels 2 and 3 cross-listings may lead to more legal bonding since in these situation firms must comply with U.S. GAAP, report quarterly, meet the listing requirements of the exchange, and meet full SEC disclosure requirements in case of level 3. Last of all, level 4 (also referred to as a 144A listing) allows firms to raise capital by means of a private placement. These DRs do not have to register at SEC and comply with U.S. GAAP4. (Lasfer, 2010) Although it seems that DRs at level 2 and 3 lead to most bonding, foreign firms accessing U.S. capital markets by OTC listings or through Rule 144a face governance constraints as well. These firms do not have to meet full SEC disclosure requirements, though they could be subject to additional informal monitoring provided by various financial intermediaries such as analysts or large investors (Frésard and Salva, 2010).
2.1.2 Reason for cross-listing
Even after acknowledging the various types of DRs, the reason for the existence of cross-listing has yet to be provided. To explain the phenomenon of cross-cross-listing numerous reasons can be found in the literature.
Karolyi (1998) describes several benefits of cross-listing, including improved access to capital, increased liquidity, a decline in capital costs, and intensified corporate prestige. Others complement this list by stating additional advantages such as:
• Greater risk sharing, lower transaction costs, and a greater exposure to specialised analysts community that can better appreciate the business and growth opportunities (Marosi and Massoud, 2008)
• Achieving greater corporate visibility (Pagano et al., 2002)
• Extended share’ trading hours by cross-listing within different time zones (Werner, Kleidon, 1996)
Besides these liquidity and market-segmentation-overcoming based arguments, corporate governance related benefits derived from cross-listing occur in the literature as well.
2.2 Cross-listing and corporate governance
An important aspect of corporate governance is dealing with the agency conflict, more specific the chance that managers and controlling shareholders (insiders) will expropriate assets from minority shareholders (Benos and Weisbach, 2004). One way of preventing this expropriation is strong corporate governance.
2.2.1 Internal Corporate Governance; Ownership structure
In numerous firms management’s commitment to a firm is stimulated by rewarding management with company’ shares. Claessens et al., (2002) utter that initially firm performance increases due to high managerial shareholdings, however soon managers become entrenched and start tunneling firm assets at the expense of outside investors, resulting in a decrease of firm value as seen by investors.
In the perspective of large outsider shareholders, two types can be distinguished: rich investors and institutional investors. Regarding the monitoring of management rich investors might be more reliable than institutional shareholders, mainly because they have their own money at stake. Institutional investors are run by managers which creates potential for an additional agency conflict (Gorton and Kahl, 1999). Besides this it has been found that large institutional investors do not care about long term perspectives of firms but focus on a rise in the share price and maximization of profits in the short term (European Commission, 2010). In addition a limitation of the monitoring authority of large institutional investors is mentioned by Denis et al. (1997), who state that large institutional investors can only improve corporate governance if they have access to the same non-public information as management has; however in most cases they do not and therefore are limited in their ability to effectively monitor management.
In addition to the two groups of large shareholder types, a differentiation can be made between independent and dependent firms.
Independent firms have shareholders with a minimum ownership of 5 per cent (Borokhovich et al., 2005). Even though these shareholders (occasionally called blockholders) are large they do not control the firm, and the firms is still considered to be widely held. In comparison to small shareholders, blockholders have stronger incentives and more opportunities to monitor management and improve firm performance. The monitoring activities of blockholders can lower agency costs in the firm and increase cash flows for all shareholders (Atanasov, 2004).
When firms have large shareholders who are in control of a firm, these firms are referred to as dependent firms. Faccio and Lang (2002) define controlling shareholders as shareholders who have ‘an absolute majority (>50%) of voting rights, or hold enough voting rights to have de facto control.’ De facto control might occur when a shareholder has more than 40 per cent of voting rights and no shareholder owes a higher percentage directly or indirectly (Faccio and Lang, 2002).
Vishny, 1997; Atanasov, 2004). Kalcheva and Lins (2007) discover that when firm’s ownership is mainly held by large insiders the value of these firms is lower. They find that this effect is stronger when cash reserves are large and when the firms come from countries with poor investor protection. In addition, Dittmar and Mahrt-Smith (2007) uncover that the value of cash holdings decreases when firms have several larger shareholders.
Even though controlling large investors are often considered as negative to firm value in the eyes of minority shareholders there is tendency to increase the monitoring activities of large shareholders. Regulators try to advance this by making efforts in several fields:
• “strengthening shareholder cooperation through discussion platforms; • disclosure by institutional investors of their voting practices at shareholders'
meetings;
• institutional investors adherence to best practice;
• identification and disclosure of possible conflicts of interest by institutional investors;
• Improve information on risk.” (European Commission, 2010)
Farinha and López-de-Foronda (2009) state that in civil law countries there is typically little separation between ownership and control, and therefore conflicts are mainly between large shareholders that control the decisions of the firm and minority shareholders. The Netherlands is classified as a civil law country so therefore conflicts between large and minority shareholders might be present.
In Europe large controlling shareholders are often families. Faccio and Lang (2002) prove that in Western Europe 36.93 per cent of firms are widely held (independent) and 44.29 per cent are family controlled. However when looking specifically at the Netherlands; Donker et al. (2009) discover that in continental Europe the Netherlands has the most market-oriented corporate governance structure that resembles the ones of the U.S. and the U.K.
In recent years laws have been adopted to protect minority shareholders. For example, certain laws prevent investors from holding very large stakes. (Franks et al., 2001). Taking this in account it might be expected that there is an increase in blockholders as opposed of controlling shareholders. However the size of the total share of a certain shareholder is not always easily noticeable and therefore it is hard to state if a large shareholder is a blockholder or a controlling shareholder. Faccio and Lang (2002) report that shareholders might use several devices that give them control rights in excess of cash-flow rights. These devices include pyramid holdings structures, holdings through multiple control chains and cross-holdings.
With a pyramid holding a controlling shareholder carries out control over one corporation by means of another corporation that he does not own totally. An example of this is Heineken NV which is owned by Heineken Holding NV which in its turn is owned by the Heineken family. In this way the Heineken family classifies as controlling shareholder. Another type of construction is a multiple control chain which gives controlling ownership to a shareholder by means of several control chains, each representing at least 5% of total voting rights. Finally referring to cross-holdings means that a firm controls its own shares either directly or indirectly.
So at first it might seem that a shareholder is no controlling shareholder, however by looking at ultimate ownership created through the different devices as described above a shareholder might still exercise control over a company.
2.2.1.1 Ownership Structure and cross-listing
The act of cross-listing is found to have an effect on a firm’s ownership structure and on agency conflicts within the firm.
controlling shareholders are often families (Faccio and Lang, 2002). However as mentioned before the consequence of such ownership might not be applicable for the Netherlands since Dutch firms are found to have a non-typical European ownership structure (Donker et al., 2009).
Doidge et al. (2004a) also state that most large foreign companies are typically controlled by large shareholders and therefore companies will only cross-list in the U.S. if these controlling shareholders can benefit from it. Doidge et al. utter that controlling shareholders focus their efforts on expropriating from minority shareholders or on exploiting growth possibilities. Expropriation mainly occurs when a firm lacks growth option. In the absence of growth opportunities, there is no need to access external capital markets and therefore controlling shareholders are indifferent about the consequences of expropriating from minority shareholders. In situations in which firms have growth opportunities that can be exploited there is a need to access capital markets in order to raise sufficient funding. Therefore it is important that cash can be raised against favourable rates. When outside investors expect controlling shareholders of expropriation of minority shareholders they will provide firms with capital against favourable rates. For this reason controlling shareholders limit the expropriation of minority shareholders in case of growth opportunities. It is argued that controlling insiders commit to limit their expropriation from minority shareholders after cross-listing and this increases the ability of firms to take advantage of growth opportunities (Doidge et al., 2004a).
2.2.2 External Corporate Governance; Legislation
Nowadays numerous laws and institutions try to limit the level of wealth insiders can extract from investors. When there is a high availability and applicability of laws and institutions, the protection for minority investors is considered better. For this reason several researchers have found a connection between the level of minority shareholders’ protection and the ability of firms to raise external capital from investors (e.g. La Porta et al., 1997).
for the weak protection of minority investors under their own jurisdictions' laws firms aim to achieve a higher market valuation (Coffee, 2002). The bonding hypothesis of Coffee (1999, 2002) and Stulz (1999) states that due to the strict U.S. investor protection regulation, shares foreign corporations cross-listed in the U.S. should trade at a premium, the so called cross-list premium. Over the years much empirical support has been provided for the bonding hypothesis. Doidge et al. (2004) prove that U.S. listed foreign firms are worth more than their non-U.S. listed counterparts. By analysing a sample of 398 firms from Indonesia, Korea, Malaysia, the Philippines, and Thailand, Mitton (2002) shows that firms with higher disclosure quality due to cross-listing in the U.S. outperform firms with low disclosure quality. Based on his findings he states that individual firms have problems ruling out expropriation of minority shareholders when legal protection is insufficient. Additionally, Dyck and Zingales (2004) find evidence of considerably lower control premiums for firms that are cross-listed in the U.S. in comparison to firms that are not cross-listed. In line with this Doidge (2004) discovers that prices of superior voting shares are traded against a discount as opposed to normal voting shares for firms that cross-list in comparison with those that are single listed.
All these findings suggest that by cross-listing on an exchange with higher disclosure demands than in the firm’s domestic market, the results are that there is a better protection against expropriation of assets from minority shareholders by insiders.
Even though much support has been found for the bonding theory there are also other researchers who have wondered if cross-listing actually leads to legal bonding.
state that regardless of the loss of bonding and other benefits, insiders of foreign firms may have become stimulated to delist due to the higher compliance costs and increased disclosure related to SOX (Marosi and Massoud, 2008).
Despite the critics, more empirical results exist in support of the bonding hypothesis. Therefore it can be assumed that by cross-listing in to the U.S., minority investors are better protected against expropriation of insiders. Nonetheless there is no consensus as to whether firms that are cross-listed in the U.S. actually bond with U.S. law (Burns et al., 2006).
2.2.3 Corporate Governance in the Netherlands
In 1998 a well-known research of La Porta et al. described minority shareholder protection in the Netherlands as low. A more recent paper by Timmerman and Doorman (2002) confirms this statement of La Porta et al. and goes further by stating that not only have minority shareholders a weak position but that on the contrary board of directors have a much stronger position. As example they mention a case in which Gucci Group NV issued a large number of shares to Pinault-Printemps-Redoute (PPR) to fight off a substantial stake that Moët Hennessy Louis Vuitton (LVMH) had acquired in to the company. Since Gucci neglected asking permission of the general meeting of shareholders, the Amsterdam Court of Appeal decided that this qualified as misconduct; however the share issue was not nullified because it was argued that it was too burdensome to reverse all the consequences of the transaction. As a result of this affair the Wall Street Journal stated that the Netherlands is not a shareholder friendly country, because it will not give fair and equal treatment to shareholders.
The code itself presents principles and best practices in five areas: (I) compliance with and enforcement of the Code, (II) the management board, (III) the supervisory board, (IV) the shareholders and general meeting of shareholders and (V) Financial Reporting. This already shows that the code is much smaller in size and spread compared to the 11 areas of SOX. The Code works with a ‘comply or explain’ principle. This means that all publicly traded companies are expected to comply with all provisions in the code or else specifically explain why they do not. In contrast to the previous Peters ‘code’, compliance to the code Tabaksblat is very high (Akkermans et al., 2007).
Five years after the introduction of the code Tabaksblat it was felt that the code became out of date. Therefore the Monitoring Committee Dutch Corporate Governance Code (also known as the 'Frijns Committee') made an updated version of the code which became effective as of 1 January 20095. The Dutch corporate governance code is now known as the code Frijns. In December 2009, the code Frijns was incorporated in the Dutch Civil law. The governance principles have basically remained the same. The code Frijns is much less demanding than SOX, though ‘’the code is intended to reassure the public that Dutch businesses are run using sound, transparent methods, while simultaneously providing a framework of principles that allow innovation and an entrepreneurial spirit to thrive’’.6 Compliance to the code Frijns has not been measured yet, however since it is basically the same as the code Tabaksblat, compliance is expected to remain high. Regardless of the high compliance to the Dutch governance code the Netherlands are still considered as a country with low minority shareholder protection. By looking at the extent of disclosure, the extent of director liability, and the ease of shareholder suits the World Bank and the International Finance Corporation (IFC) publish a yearly investor protection index. When it comes to ‘protecting investors’ in 2010 the Netherlands is ranked 109th out of 183 economies, while the U.S. is rank 5th out of those 183 economies (World Bank and IFC, 2011). This indicates below average investor protection and shows that the Netherlands is regarded as a country with poor investor protection in comparison to the U.S.
2.3 Cash
5http://www.ing.com/group/showdoc.jsp?docid=074107_EN (access date: 14-01-2011) 6
In recent years several researchers find that firms are holding significant amounts of cash (a.o. Foley et al., 2007 and Bates et al., 2009).
2.3.1 Cash levels and cross-listing
Over the years the literature has provided numerous explanations to why firms hold cash. These motives can be summarized in to four categories:
1. The transaction motive is based on the classic models in finance. These models state that
there are transaction costs when firms convert a noncash financial asset into cash and use cash for payments (Baumol, 1952, Miller and Orr, 1966). For this reason it might be less costly to just hold cash instead of converting assets in times of higher cash demand. The economies of scale associated with the transaction motive result in the assumption that lower amounts of cash are being held by sizeable companies. This way of reasoning is backed up by much empirical evidence (f.e. Mulligan, 1997.)
2. The precautionary motive mentions being able to deal with unexpected misfortune in
times when access to capital is hard as the main reason to hold cash. In this case cash holdings can be regarded as a buffer for potential difficult times in the future. Proof of the existence of the precautionary motive is delivered by Opler et al., (1999) who find that companies hold higher levels of cash when their cash flows are more volatile and when they have limited access to external funds. Several researches, also find support for another aspect of the precautionary motive, namely that firms with superior investment prospects maintain higher levels of cash because financial disturbance and detrimental calamities are more costly for them (Opler et al., 1999, Han and Qiu, 2007, Riddick and Whited, 2009.)
3. The tax motive is an explanation for firms holding cash in case of potential future
taxations. When firms send foreign earnings home this might lead to certain tax consequences, for this reason those firms hold more cash (Foley et al., 2007.) Especially when high tax consequences of repatriated earnings exist, firms are found to be holding more cash. As a result, high cash reserves may expect to be found under multinational enterprises (MNEs).
4. The agency motive is a result of the agency theory of Jensen (1986). Jensen (1986) argues
management will prefer to keep cash in the firm instead of paying it out to shareholders. These unrestricted cash holdings are referred to as the excess cash holdings. Jensen (1986) defines excess cash as “cash flow in excess of that required to fund all projects that have positive net present values when discounted at the relevant cost of capital”.
When assuming that firms are cross-listing in order to exploit growth opportunities (Doidge et al., 2004a) it might be expected that based on the transaction motive (Baumol, 1952, Miller and Orr, 1966) and the precautionary motive (Opler et al., 1999) cross-listed firms hold more cash.
2.3.2 The value of cash
Considering the four motives mentioned above, the first three can be considered as rational reasons for holding cash; however having a too high level of cash might be undesirable for two reasons. The first one is that holding too much cash makes firms unable to optimally utilize assets. This means that they will have a lower return on cash, relative to investing it in projects with a similar level of risk. This cost is often referred to as the cost of carry, which is the variance between the return on cash and the interest that would have to be paid to finance an additional dollar of cash (Dittmar et al., 2003).
The second reason explaining why holding a too high level of cash is detrimental is related to the agency problem. Agency theory mentions that there is a difference in interest between management and shareholders (Jensen and Meckling, 1976). When assuming that managers do not try to create shareholder wealth maximization, holding cash will become more expensive since managers will be provided with the chance to engage in wasteful capital spending (Dittmar et al., 2003), or expropriation of assets for personal gains (Myers and Rajan, 1998). When there is a low level of monitoring and minority shareholders’ legal protection is little, there is a higher risk that insiders will turn their firm’s assets into private benefits. Myers and Rajan (1998) find that liquid assets are the easiest target for expropriation; continuing on this Dittmar and Mahrt-Smith (2007) note that cash holdings are the least complicated accessed by management for this purpose, and that there is little scrutiny on it. Several circumstances might lead to a higher risk of an agency problem related to cash holdings.
management is using it for own personal gains. Other research suggests that deep-seated managers first build up cash in excess and after that spend it swiftly (Harford et al., 2008.) Considering these findings it appears that insiders (whether just management, management with significant amounts of shares or influential shareholders) will expropriate from minority shareholders in order to achieve personal gains. The main way to limit insiders’ expropriation of minority shareholder is monitoring. Dittmar et al. (2003) analyse a sample of 11,000 firms from 45 countries and find that in countries with poor shareholder protection firms hold more cash in excess. With regards to the valuation of cash linked to certain countries’ corporate governance Pinkowitz et al. (2006) found that in countries where legal protection is weak, investors value cash at a discount. Recent studies show that when minority shareholders’ legal protection is weak (Pinkowitz et al., 2006) and when monitoring demands on insiders is limited (Dittmar and Mahrt-Smith, 2007), investors value cash holdings at a sizable discount.
2.4 Informational bonding
When a firm decides to cross-list its shares, it also commits itself to disclosing more information. This process is often referred to as informational bonding. In addition to the demand of more information, several articles state that by cross-listing firms obtain a higher level of analyst coverage (a.o. Lang et al., 2003; Fernandes and Ferreira, 2008; Frésard and Salva, 2010). Next to increased disclosure and a larger information, environment also enlarged forecast accuracy can be regarded as a result of this higher level of analyst coverage (Lang et al., 2003). Under the scrutiny of a large analysts community firms are more likely expected to value shareholder rights and less likely to expropriate from minority shareholders (Coffee, 2002). Besides Coffee other researchers have acknowledged the role of analysts as external monitors of companies. Jensen and Meckling (1976) highlight the social function of analysts by stating that “security analysis activities reduce the agency costs associated with the separation of ownership and control”. Moreover, Healy and Palepu (2001) mention that information intermediaries, such as analysts and rating agencies, increase the production of public information that contributes to detecting managerial misconduct.
As a consequence of declining agency costs, the cost of capital decreases as well. Lang et al. (2003) find that firms with a higher level of analyst coverage and higher forecast accuracy have higher valuations.
Yu (2008) mentions several arguments why analysts are considered to be effective monitors of companies. To begin, analysts have a different target group to which they deliver information; they consider current shareholders, as well as potential future shareholders and other market participants.
Secondly, analysts have often had more training to deal with financial statements than most executives and board members of a firm had. Next to this, analysts have more resources which they can access to go through dreary financial statements and complex footnotes. For these reasons it can be said that in most situations analysts are more financially substantiated than the executives and board members of a firm.
Finally, the frequency of company’ controls per year done by analysts is much higher compared to that of other monitors, such as board members, auditors and the SEC. Analysts can guarantee a better oversight by continuously examining management behaviour and financial reporting. In contrast, board meetings are once in a quarter and a firm’s auditor only visits at the end of the fiscal year. When it comes to legal controls, in line with findings of Seigel (2005) it is reported that the SEC is only able to review financial reports partly. In 2002, the U.S. Government Accountability Office (GAO) stated that SEC staff only reviewed 16 per cent of annual reports filed in 2001. (Yu, 2008)
2.5 Hypotheses development
Based on the discussed literature several hypotheses are formulated.
2.5.1 Cross-listing and the value of cash
As described above in the literature, the cross-listing phenomenon discusses the bonding theory as a potential solution to the agency problem. According to the agency theory (Jensen and Meckling, 1976) low monitoring on insiders results in to higher risk of company assets expropriation by insiders. Looking at the type of assets that are most vulnerable to this, Myers and Rajan (1998) and Dittmar and Mahrt-Smith (2007) state that cash are the easiest assets that insiders can transfer into personal benefits.
capital is cross-listing (Eiteman et al., 2007). Several studies have found that besides attracting capital, cross-listing also leads to increased firm value (a.o. Lang et al., 2003 and Doidge, 2004). In addition, as a consequence of cross-listing firms receive additional monitoring. The bonding theory, introduced by Coffee (1999, 2002) and Stulz (1999), indicates that when firms cross-list, they become subject to strict U.S. corporate governance regulation which makes it harder for insiders to expropriate or tunnel firm assets for their private benefits.
The Netherlands is considered as a country with poor investor protection in comparison to the U.S. (La Porta et al., 1998; Timmerman and Doorman, 2002; De Jong et al., 2005; World Bank and IFC, 2011). So when Dutch firms cross-list in the U.S. and come under U.S. legislation this might result in stronger corporate governance systems for Dutch firms. Since cash is an asset that is easily expropriated by insiders, the valuation of cash might increase when it becomes harder for insiders to expropriate due to the increased monitoring that follows a cross-listing in the U.S. I hypothesize:
H1: Investors will value the cash of Dutch firms higher if they are cross-listed in the U.S.
Different types of cross-listing can be distinguished, which have different monitoring effects (see appendix I). Firms traded as levels 2 and 3 DRs obtain the highest increase in monitoring, therefore I hypothesize:
H1a: Investors will value the cash of Dutch firms higher if they are cross-listed on level 2
and 3 in the U.S.
2.5.2 Corporate Governance
According to theory the Netherlands is depicted as a country in which there is little separation between ownership and control, therefore conflicts are mainly between large shareholders that control the decisions of the firm and minority shareholders (Farinha and López-de-Foronda, 2009.)
investors recognize the potential agency problem and therefore discount cash holdings. However when firms with such shareholders cross-list in the U.S. this is expected to increase the value of cash holdings due to higher monitoring and the need for external funds. Therefore I hypothesize:
H2: Investors will value the cash of Dutch firms with large amounts of closely held shares
higher if they are cross-listed in the U.S.
In line with theory it is expected that level 2 and 3 ADRs come under more strict scrutiny of U.S. regulation. Therefore I expect that the effect of cross-listing is stronger for these firms, and I hypothesize:
H2a: Investors will value the cash of Dutch firms with large amounts of closely held shares
higher if they are cross-listed on level 2 and 3 in the U.S.
2.5.3 Information Environment
As mentioned before a greater exposure to a specialised analyst’s community that can better appreciate the business and growth opportunities is regarded as one of the benefits of cross-listing (Marosi and Massoud, 2008). Healy and Palepu (2001) mention that information intermediaries, such as analysts and rating agencies, increase the production of public information that contributes to detecting managerial misconduct. Coffee (2002) elaborates on this by stating that under the scrutiny of a large analyst’s community, firms are more likely expected to value shareholder rights and less likely to expropriate from minority shareholders. Therefore, it can be stated that firms that are followed by a large analyst community end up with having stronger governance. This leads to my last hypotheses:
H3: Investors will value the cash of firms higher when they are cross-listed in the U.S. and
have a large information environment.
In line with previous hypotheses I expect that the effect of an increased information environment will be stronger for cross-listings on level 2 and 3. I hypothesize:
H3a: Investors will value the cash of firms higher when they are cross-listed on level 2 and 3
2.5.4 SOX implementation
A common explanation for the decrease in foreign listings on U.S. exchanges is the implementation of the Sarbanes-Oxley Act (SOX) in 2002. The idea behind SOX is to lead firms towards honest and ethical conduct, transparency, compliance with laws, and accountability in order to protect investors (Rockness and Rockness, 2005). Nevertheless it is stated that the implementation of SOX has made cross-listing on U.S. exchanges so unattractive that firms would delist and deregister if that was an easy thing to do (Litvak, 2007, Doidge et al, 2009). Since delisting is not a real option the implementation of SOX has had its effect on firms listed in the U.S. The main discussed downside effect is that SOX makes a U.S. listing less advantageous due to the high costs it imposes on companies and their managers (Marosi and Massoud, 2008, Doidge et al, 2009). The effect of SOX on U.S. firms is hard to estimate since there is no control group of firms not affected by it (Litvak, 2008). However for Dutch cross-listed firms in the U.S. the impact of SOX can be measured by comparing them to their domestic counterparts. I hypothesize:
H4: The effect of cross-listing in the U.S. on the valuation of cash will be statistically more
significant after implementation of the Sarbanes-Oxley Act in 2002.
Since cross-listings on level 2 and 3 need to fulfil all SOX disclosure requirements the effect of SOX might be stronger on those listings. Therefore I differentiate between cross-listing types and hypothesize:
H4a: The effect of placing level 2 and 3 ADRs on the valuation of cash will be statistically
SECTION 3
DATA AND METHODOLOGY
3.1 Sample
The sample consists of all firms listed on Euronext Amsterdam as of the 31st of December 2007. I keep financial firms (SIC codes 6000 - 6999) out, because their business implies holding large amounts of liquid assets. After excluding the firms of which no sufficient data could be obtained, this study uses a sample of 110 Dutch firms. In the sample the cross-listed firms are diversified by the type of their cross-listing (see appendix 1). By type of listing, there are 12 firms directly listed on an U.S. exchange and 22 firms listing through private placements under Rule 144 or OTC. The benchmark sample contains 76 firms. In line with the work of Frésard and Salva (2010) I consider each firm only once, regardless of the number of cross-listed securities it has.
3.1.1 Sources for the necessary data
The sources that have been used to collect the data consist of Thomson Financials’ Datastream, Thomson Financials’ One Banker, Worldscope, the Institutional Broker’s Estimate System (IBES), Orbis (former Amadeus) and JPMorgan’s American Depositary Receipt (ADR) database7. One Banker, Datastream and Orbis are used to obtain varies price data. IBES is accessed to get the historical amount of analysts following a firm. Worldscope is used to obtain the historical percentage of closely held shares (CHS) of each firm. This study makes use of JPMorgan’s ADR database and Orbis to obtain the cross-listing data. To verify the data from JP Morgan, this study also checks the Bank of New York’s ADR database8. Missing accounting data is retrieved from annual reports.
3.2 Data type
This research deals with unbalanced annual panel data. For each firm, data has been collected for the period 1996-2007.
3.2.1 Dependent variable: Tobin’s Q
7http://www.adr.com/BrokerInvestor/drsearch.aspx (access date: 15-10-2010)
8
Consistent with prior studies in this field the dependent variable is the firm’s Tobin’s Q as a measure of firm value (Mura, 2007; Doidge et al., 2009). The Tobin’s Q ratio is the market value of a company (MV) divided by the replacement value of a firm's assets (RV).
Tobin’s Q = MV / RV
The market value of a firm in this study is calculated by taking the book value of total assets (BVTA), minus the book value of equity (BVE), plus the market value of equity (MVE) as also applied by Mura (2007) and Doige et al. (2009).
MV = (BVTA – BVE) + MVE
For the replacement value of a firm's assets research looks at total assets. Under the assumption that market and book value of debt are equal and that the replacement costs of assets are similar to the value of the firm’s current assets; this study uses book value of total assets as replacement value (Mura, 2007, Doige et al., 2009).
RV = BVTA
Sample domination of large firms and heteroskedasticity might lead to wrong inferences. By taking the book value of total assets as the replacement value of the firm and dividing the market value of the firm by this figure, sample domination of large firms and heteroskedasticity can be prevented (Fama and French, 1998).
3.2.2 Explanatory variables Cash (Cash)
The main independent variable is cash. Corresponding to Pinkowitz et al. (2006) and Frésard and Salva (2010), cash is defined as cash and short term investments divided by total assets. The coefficient of the cash variable displays the valuation of investors of one additional dollar in cash.
Cross-listing Dummy (CLD)
In line with Doidge et al. (2009) and Frésard and Salva (2010) I track the variable activity around the time of cross-listing. Since it has been found that abnormal returns end on the long term after a cross-listing (Miller, 1999 and Foerster and Karolyi, 1999) and are especially higher in the year before and during the cross-listing the CLDs only get a value of 1 in the 3 years surrounding the cross listing event.
Ownership Structure (CHS)
As seen in previous literature the ownership structure of the firm has an effect on its corporate governance. In this study the level of a firm’s corporate governance is measured by means of the ownership structure (a.o. Florackis and Ozkan, 2009; Holm and Schøler, 2010; Frésard and Salva, 2010).
In the Worldscope database I look for the data item named ‘‘closely held shares’’ in order to determine the presence of large shareholders. Closely held shares (CHS) is defined as “the percentage of shares held by senior corporate officers and directors, and their immediate families; shares held in trusts; shares held by another corporation (except shares held in a fiduciary capacity by financial institutions); shares held by pension and benefit plans; and shares held by individuals who hold 5% or more of shares outstanding’’ Frésard and Salva (2010).
Number of analysts (IE)
Theory predicts that an increase in analyst coverage should result in to a higher valuation of firms by investors. This measure is computed as the yearly number of analyst following a firm (as in Frésard and Salva, 2010; and Lang et al., 2003) and taken from the Institutional Broker’s Estimate System (IBES).
3.2.3 Interaction variables consisting of two variables Cash * Cross-listings in the U.S. (Cash * CLD)
Cash * Exchange listed ADRs (Cash * Exchange)
Cash * OTC and 144A ADRs (Cash * OTC & 144A)
3.2.4 Interaction variables consisting of three variables
Cash * Cross-listings in the U.S. * Closely Held Shares (Cash * CLD * CHS) Cash * Exchange listed ADRs * Closely Held Shares (Cash * Exchange * CHS) Cash * OTC and 144A ADRs * Closely Held Shares (Cash * OTC & 144A * CHS)
Since corporate governance and cross-listing both are expected to have an influence on the valuation of cash by shareholders, the various variables that are used as proxies for corporate governance and cross-listing are interrelated with each other in to the above mentioned interaction variables.
Cash * Cross-listings in the U.S. * Number of analysts (Cash * CLD * IE) Cash * Exchange listed ADRs * Number of analysts (Cash * Exchange * IE) Cash * OTC and 144A ADRs * Number of analysts (Cash * OTC & 144A * IE)
As mentioned in section 2.4 the information environment can contribute to monitoring the management of a firm. In line with among others Lang et al. (2003), Fernandes and Ferreira (2008), Frésard and Salva (2010) the number of analysts following a firm has been taken as a proxy to determine the information environment of a firm. Lang et al. (2003) state that more analyst following leads to more inspection, which might limit agency costs and improve the firm value in the eyes of shareholders.
To see if firms that are cross-listed and have a better information environment receive higher cash valuations, this interaction variable is applied. Since OTC and 144A cross-listed firms do not have to comply with full SEC requirements it is interesting to see if an increase in external monitoring by analysts leads to a higher company valuation for these types of listings.
3.2.5 Control variables
In addition to the explanatory variables and interaction variables, several firm related factors are included to control for.
Growth Opportunities (GRWTH)
Profitability (PROF)
Profitability is also often seen as a factor influencing the value of a firm. Following Frésard and Salva (2010) net income divided by total assets is taken as proxy for profitability.
Size (SIZ)
Doidge et al. (2004) state that larger firms are more likely to cross-list than smaller firms. In their study they show that cross-listed firms are relatively four times as large compared to non-cross-listed firms in 1997 and almost six times as large in 2004. Because of this difference in size this research controls for firm size. As in Lang et al. (2003) and Frésard and Salva (2010), the logarithm of total assets is included as a proxy for firm’ size.
Dependent firms (Dep. Firms)
As mentioned above, a firm’s ownership structure is found to effect its corporate governance. In the financial database Orbis it is assumed that firms are independent when none of the firm’s recorded shareholders has more than 25 per cent of direct or total ownership. In such cases the minority shareholder protection in a firm is regarded as decent. When a shareholders of a firm does have more than 25 per cent of control, chances of expropriation of minority shareholders are larger and such firms are marked as dependent firms. In this research one dummy variable is created which takes the value of 1 when the firm is regarded as dependent and zero otherwise.
As mentioned by Frésard and Salva (2010) the CHS measure does make a precise division between shares owned by controlling insiders and shares owned by large outside investors. Therefore I include the dependent firm variable to distinguish the presence of controlling outside investors in the sample.
Industry type
3.3 Description of the test
This study implements models based on the model of Fama and French (1998), similar to those used in Frésard and Salva (2010) and applies a panel regression model of the following form:
Tobin’s Q i,t = α + β1 [CLD]i,t + β2 [Cash]i,t + β3 [CHS]i,t + β4 [IE]i,t + β5 [GRWTH]i,t + β6 [PROF]i,t + β7 [SIZ]i,t + β8 [DEP. FIRM]i,t + β9 [Construction]i,t + β10 [Consumer products]i,t + β11 [Electrical and optical equipment]i,t + β12 [IT]i,t + β13 [Paper and paper products]i,t + β14 [Services]i,t + β15 [Transport/Storage]i,t + εi,t
A second model checks for the effects of including interaction variables between two explanatory variables:
Tobin’s Q i,t = α + β1 [CLD]i,t + β2 [Cash]i,t + β3 [CHS]i,t + β4 [IE]i,t + β5 [Cash * CLD]i,t + β6 [GRWTH]i,t + β7 [PROF]i,t + β8 [SIZ]i,t + β9 [DEP. FIRM]i,t + β10 [Construction]i,t + β11 [Consumer products]i,t + β12 [Electrical and optical equipment]i,t + β13 [IT]i,t + β14 [Paper and paper products]i,t + β15 [Services]i,t + β16 [Transport/Storage]i,t + εi,t
A third model is applied to check for the effects that results from the inclusion of interaction variables between three explanatory variables:
Tobin’s Q i,t = α + β1 [CLD]i,t + β2 [Cash]i,t + β3 [CHS]i,t + β4 [IE]i,t + β5 [Cash * CLD]i,t + β6 [Cash * CLD * CHS]i,t + β7 [Cash * CLD * IE]i,t + β8 [GRWTH]i,t + β9 [PROF]i,t + β10 [SIZ]i,t + β11 [DEP. FIRM]i,t + β18 [Construction]i,t + β19 [Consumer products]i,t + β20 [Electrical and optical equipment]i,t + β21 [IT]i,t + β22 [Paper and paper products]i,t + β23 [Services]i,t + β24 [Transport/Storage]i,t + εi,t
3.4 Outliers
As in Doidge et al. (2009) I also applied winsorization to reduce the impact of outliers. All dependent, independent and control variables are winsorized at the first and 99th percentiles.
3.5 Heteroskedasticity
Differing variance between variables is limited by using natural logarithms, dividing firm characteristics by total assets and applying percentages.
3.6 Multicollinearity
SECTION 4
EMPIRICAL RESULTS
As can be seen in figure 1 the distribution of the key variables in the sample is not entirely Gaussian. Consequently a non-parametric test is used to check for the medians of the variables in the sample. More specifically, the Kruskal-Wallis (K-W) test is applied to test for the equality of medians among the population groups. With significance levels at 1% and 5% the K-W test results show that all variables have an equally formed and scaled distribution for each group, except for the control variable growth. In addition the equality of means is tested by an F-test. The F-test confirms the dissimilar distribution among groups of the growth variable and finds comparable results for the control variable profitability.
Figure 1
Distribution of key variables.
average annual per cent increase in net sales. PROF measures profitability by dividing net income by total assets. SIZE is the natural logarithm of total assets.
0 40 80 120 160 0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 F re q u e n cy Tobin's Q 0 40 80 120 160 200 240 280 320 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 F re q u e n cy Cash 0 20 40 60 80 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 F re q u e n cy CHS 0 100 200 300 400 500 600 0 5 10 15 20 25 30 35 40 45 F re q u e n cy IE 0 40 80 120 160 200 240 280 320 -0.8 -0.4 0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 F re q u e n cy Grow th 0 100 200 300 400 500 -.8 -.6 -.4 -.2 .0 .2 .4 .6 .8 F re q u e n cy Prof itability 0 20 40 60 80 100 120 6 8 10 12 14 16 18 20 F re q u e n cy Size
listed, have higher Tobin’s Q values than non cross-listed firms. This indicates that investors value cross-listed firms higher compared to their domestic counterparts, which has been found before by a.o. Doidge et al. (2009).
Furthermore, table 1 shows that cross-listed firms hold more cash than solely domestic listed firms. An explanation for this can be the transaction motive (Baumol, 1952, Miller and Orr, 1966) and the precautionary motive (Opler et al., 1999), given that a firms’ reason to cross-list is assumed to be based on exploiting growth opportunities (Doidge et al., 2004a). The question that remains unanswered is if this higher amount of cash is also valued higher. When it comes to the percentage of closely held shares table 1 indicates that cross-listed firms have lower amounts of managerial shareholdings compared to non cross-listed firms. This finding can be considered as an improvement in a firm’s corporate governance as a result of cross-listing in the U.S. and is also found by Frésard and Salva (2010) and O’Connor (2007). In addition, table 1 signifies that cross-listed firms have a higher amount of analysts following compared to their domestic counterparts (benchmark). This confirms that when firms cross-list in the U.S. they not only have to comply with strict U.S. legislation, but additionally come under the scrutiny of a large analyst community as proven before in the work of a.o. Lang et al. (2003). Remarkably descriptive statistics of control variables indicate what theory predicts (Doidge et al., 2004a), namely that cross-listed firms are larger and have higher growth, however they appear to be less profitable than their domestic counterparts9, which might be explained by the high cost related to cross-listing in the U.S. (Zhang, 2005).
Table 1
Descriptive statistics.
This table presents mean, median, maximum, minimum and standard deviations for the key variables used in this study. Tobin’s Q, is computed as ((total assets - the book value of equity) + the market value of equity) / total assets). Cash is defined as cash and short term investments divided by total assets. CHS represents the percentage of closely held shares. IE is the yearly number of analyst following a firm. GRWTH measures a firms growth by taking the average annual per cent increase in net sales. PROF measures profitability by dividing net income by total assets. SIZE is the natural logarithm of total assets. I compute F-tests and K-W tests to determine if there are significant differences between grouping criteria for the mean and median values, respectively. ***, ** and * represent statistical significance at the 1%, 5% and 10% level, respectively.
Number of obs. Mean Median Maximum Minimum Std. Dev.
Tobin's Q
All 1203 1.38 1.24 2.45 0.13 0.55
Exchange 137 1.57 1.51 2.45 0.51 0.53
Non-Exchange 236 1.47 1.37 2.45 0.54 0.56
9 It should be noted that growth and profitability are found not to be equally formed and scaled among all
Benchmark 830 1.32 1.17 2.45 0.13 0.54 F test 16.70 *** K-W test 37.25 *** Cash All 1261 0.06 0.07 0.40 -0.84 0.15 Exchange 141 0.10 0.09 0.40 0.00 0.07 Non-Exchange 249 0.07 0.07 0.40 -0.84 0.15 Benchmark 871 0.06 0.06 0.40 -0.84 0.16 F test 5.81 *** K-W test 20.33 *** CHS All 1056 38% 36% 98% 0% 26% Exchange 128 21% 7% 90% 0% 25% Non-Exchange 231 31% 29% 98% 0% 23% Benchmark 697 43% 42% 98% 0% 25% F test 57.88 *** K-W test 119.26 ***
IE Number of obs. Mean Median Maximum Minimum Std. Dev.
4.1 Results Panel data Regression
This section discusses the main results of the Panel data regressions that have been conducted to test four hypotheses that have been formulated in section 2.5.
4.1.1 Cross-listing and the value of cash
The results of the regression that tests my first hypothesis can be found in table 2. A first noteworthy statistical significant result at a 10% level in columns 1 and 3 is that an increase in the level of cash is discounted. The regression result indicates that the overall value of € 1 of cash for Dutch firm is € 0.16. This is in line with findings of Pinkowitz et al. (2006) and Frésard and Salva (2010) who discover that liquid assets are generally valued at a discount. The results of the regression in columns 2 and 4 indicate that Dutch cross-listed firms in the U.S. receive a higher investor valuation of cash then their domestic counterparts.
To be specific, for Dutch cross-listed firms in the U.S., despite of the type of cross-listing the valuation of € 1 of additional cash is € 0.55, statistically significant at a 5 per cent level. In addition looking at the specific cross-listing types I find that € 1 of additional cash is valued at € 0.49 for firms cross-listed OTC and by means of SEC rule 144A and at € 1.08 for Dutch firms cross-listed on an U.S. exchange, statistically significant at a 10 and 5 per cent respectively. This finding indicates that shareholders value the cash of Dutch firms cross-listed on an U.S. exchange at a premium. Similar findings are reported by Frésard and Salva (2010) who find that on average, investors’ valuation of excess cash turned out to be near three times larger for cross-listed firms than for non cross-listed firms. Frésard and Salva found that the overall value of $ 1 is $ 0.58, which increases to $ 1.61 and to $ 1.42 for exchange cross-listed firms and OTC listings, respectively. In these results it is clear that excess cash of both types of listings are traded at a premium.
Even though my findings are of different size than those of Frésard and Salva (2010), it can be concluded that an increase of the valuation of cash due to cross-listing is noticeable and significant. Therefore I accept both hypotheses 1 and 1a. It is clear that investors recognize the constraining effect of legal and disclosure requirements related to U.S. cross-listings and therefore value cash higher for U.S. cross-listed firms than for their domestic counterparts. My findings indicate that this effect is the strongest for firms cross-listing on an U.S. exchange. An explanation for this is that they have to comply with full SEC disclosure requirements, while level 1 and level 4 listings do not have to meet all conditions.
Table 2
Investors’ valuation of cash: cross-listed firms versus domestic counterparts.
This table reports results from panel data regressions that estimate the valuation of cash by investors. The dependent variable is Tobin’s Q, which is computed as ((total assets - the book value of equity) + the market value of equity) / total assets). CLD is a general listing dummy that equals one if a firm is cross-listed, and zero otherwise. Exchange is a cross-listing dummy taking a value of one if firms are cross-listed at level 2 and 3 (Exchange ADRs), and zero otherwise. 144A & OTC is a cross-listing dummy that equals one if firms are traded over the counter and through private placement under SEC regulation 144A (Non-Exchange ADRs), and zero otherwise. Cash is defined as cash and short term investments divided by total assets. CHS represents the percentage of closely held shares. IE represents the residuals of the regression between size and the amount of analyst following the firm. GRWTH measures a firm’s growth by taking the average annual per cent increase in net sales. PROF measures profitability by dividing net income by total assets. SIZE is the natural logarithm of total assets. Dep. Firms is a dummy variable taking the value of 1 when firms have shareholders with ultimate ownership of at least 25 per cent, and zero otherwise. All non-dummy variables are winsorized at 1 and 99 per cent tails. The regression coefficients include time fixed effects.
Column 1 and 3 display the regression results of the models without the interaction variables. Column 2 and 4 display the coefficients of the models that include the interaction variables. ***, ** and * represent statistical significance at a 1, 5 and 10 per cent level, respectively.
Variable (1) (2) (3) (4)
CLD 0.180 *** 0.166 **
Exchange 0.004 -0.036
144A & OTC 0.237 *** 0.250 ***
Cash 0.162 * 0.035 0.161 * 0.013
Cash * CLD 0.555 **
Cash * Exchange 1.075 **
Cash * 144A & OTC 0.493 *
CHS -0.230 *** -0.248 *** -0.235 *** -0.254 *** RIE 0.008 *** 0.007 ** 0.008 *** 0.005 ** GRWTH 0.134 *** 0.125 *** 0.133 *** 0.124 *** PROF 0.323 ** 0.349 *** 0.322 ** 0.366 *** SIZE -0.026 *** -0.031 *** -0.026 *** -0.033 *** Dep. Firm -0.039 -0.046 -0.033 -0.043
Control for industry Yes Yes Yes Yes
Number of observations 951 951 951 951
