1 Cross Listing to the U.S. Stock Exchange During Crisis: The Solution for Stock
Liquidity?
University of Amsterdam Faculty of Economics and Business MSc Business Economics, Finance track
October 2013
Supervisor: N. Martynova
Arnaud van der Wijk 5768128
2 Did Cross Listing to the U.S. Lead to a Higher Level of Stock Liquidity for European Firms during the Current Financial Crisis, Compared to European Firms That Were Not Cross Listed to the U.S.?
Abstract
This research is based on the market liquidity hypothesis which implies that an important incentive for foreign companies to cross list, is that the cross listing premium (increase in firm value or Tobin’s Q) can be attributed to the increase in stock liquidity.
In this research, two hypotheses are constructed in order to test whether cross listing to the U.S. lead to a higher level of stock liquidity for European firms during the current financial crisis, compared to European firms that were not cross listed to the U.S. The relative bid-ask spread and the turnover ratio are used as measurers for stock liquidity. Both regression models that test the hypotheses do not find a significant difference in stock liquidity. Due to the fact that there has been a significant increase in the number of firms cross listing to the German Stock Exchange the last decade and developments in the international stock market, this may have decreased the value of the U.S. as a cross listing destination for European firms during a crisis.
Keywords: cross listing, stock liquidity, financial crisis, United States
Arnaud van der Wijk Nieuwe Leliestraat 131 1015 SN Amsterdam
arnaudvanderwijk@hotmail.com Student number: 5768128
3 Table of Contents 1. Introduction 5 2. Background 2.1 Recent Trends 9 2.2 Theory
2.2.1 Market Liquidity Hypothesis 11
2.2.2 Additional Hypotheses 12
2.3 Literature Review
2.3.1 Stock Liquidity 13
2.3.2 The U.S. Stock Market 15
3. Variables Description 17 3.1 Control Variables 18 3.2 Dependent Variables 19 4. Methodology 20 5. Data 22 6. Estimation Results 24 7. Robustness Tests 26 8. Conclusion 27
9. Limitations and Further Research 28
10. Bibliography 30
4 Table of Figures and Tables
Table 1: Industry Categorized 22
Table 2: Descriptive of Variables 23
5 1. Introduction
Cross listing has an upward effect on stock liquidity when firms cross list to deeper and more liquid stock markets (Fang, Noe and Tice, 2009). According to several authors (Foerster and Karolyi, 1999; Lins, Strickland and Zenner, 2005; Valero, Lee and Cai, 2009) an increase in stock liquidity attracts investors and can therefore be seen as a predictor of increased firm value for cross listed firms. High stock liquidity attracts more investors because they can buy and sell stock easily, quickly and at a low cost which generates a high level of trading activity. According to Vayanos (2004) the liquidity of a stock market could strongly decline during a crisis period due to the fact that investors tend to hold more liquid assets, which may cause that investors shift from the stock market to the market for Treasury Bonds. This is called the ‘flight to quality’ or the ‘flight to liquidity’. The decrease in demand for stocks could cause a decrease in stock liquidity for listed companies.
A high level of stock liquidity makes a firm less sensitive to the business cycle (Wuyts, 2007) which makes the U.S. stock market an interesting cross listing destination. According to Doidge, Karolyi and Stulz (2007) the U.S. stock market is a highly liquid stock market and has therefore been a popular cross listing destination for foreign firms. Also, a market survey by Mittoo (1992) indicated that managers of foreign firms cited the increase in liquidity as the primary reason to list their shares on the U.S. stock market.
This paper is related to and based on the market liquidity hypothesis, this hypothesis notes that an important incentive for foreign companies to cross list to the U.S. is that the cross listing premium (increase in firm value or Tobin’s Q) can be attributed to the increase in stock liquidity. This can be explained by the fact that high stock liquidity attracts more investors because they can buy and sell stock easily, quickly and at a low cost when a firm is listed on a foreign, liquid, stock exchange. This generates a high level of trading activity and can therefore be a predictor of the increase in firm value. Also, liquidity can be seen as risk reducing since during high volatile times investors tend to invest in more liquid assets. The aim of this research is testing whether European firms who are cross listed to the U.S. during the current financial crisis find a higher level of stock liquidity, compared to European firms that are not cross listed to the U.S.
Discussion about the value of the U.S. stock market as cross listing destination has been raised by the Sarbanes-Oxley act in 2002 and the developments in the international stock market over the years. For example, the Deutsche Börse (especially the Frankfurt Stock Exchange) has gained significant interest, along with the SEHK and the Singapore Stock
6 Exchange, of international investors and listed firms due to the fact that the German stock market has two different markets that both have its advantages: the Regulated Official Market with high corporate governance (firms have to satisfy high level of requirements in order to list shares) and the Open Market with low corporate governance (firm has to satisfy low level of requirements in order to list shares), where especially the Open Market is a highly liquid stock market. Also, when a firm is already listed on a foreign stock exchange it is
automatically listed on the FSE without needing permission of the firm. Many U.S. firms are already listed on the Frankfurt Stock Exchange (Brockman & Chung, 1998) as well as many other firms all over the world. Therefore, it is important to test whether European firms that are cross listed to the U.S. stock market find a higher level of stock liquidity compared to European firms that are not cross listed to the U.S. If the specific benefits of cross listing to the U.S. have eroded due to the developments of the stock market and changes in corporate governance, it can give firms the incentives to cross list to other foreign exchanges.
The main incentives for firms to decide to cross list is that a cross listing is accompanied with an expansion in shareholder base, it increases the visibility/transparency of a firm and there is a decrease in cost of capital. Also, when firms cross list to the U.S. there is also a highly liquid secondary market for their shares, such as the NYSE, which has an upward effect on the firm’s liquidity (Karolyi, 1998). There are three additional hypotheses from the empirical literature that discuss the incentives to cross listing. The market segmentation hypothesis, the signalling hypothesis and the bonding theory, which will be further elaborated in section two. An incentive for European firms for cross listing to the U.S. stock market is its highly liquid secondary market. High stock liquidity can have an upward effect on firm value and when firms cross list to the U.S. stock market it significantly increases the investors base (Fang, Noe and Tice, 2009). Cross listing to the NYSE or NASDAQ is done through issuing ADR’s (American Depositary Receipts) which is a tradable equity instrument which represents a stockholding in a security and allows investors to diversify without purchasing from international exchanges.
In order to give answer to the main research question, this study is for the most part based on the research of Chandar, Patro and Yezegel (2007) and Frino, Di Marco and Lepone (2009). Chandar, Patro and Yezegel have analysed the different stock price responses for firms from emerging countries that have cross listed to the U.S. through ADR’s during a crisis in their home countries and found that these cross listed firms react less negatively than non-cross listed firms. They attribute this significant difference to the market segmentation hypothesis. Frino, Di Marco and Lepone assess the impact of an American Depositary Receipt
7 programme on the issuing company. The paper carries out the research for 122 Australian listed firms covering the period 1983-2005, where they use the relative bid-ask spread and the turnover ratio as dependent variable and control for market capitalization and industry effects. They find that there is indeed a significant difference in domestic stock liquidity between Australian firms that cross listed to the U.S. and those that did not.
Note that in this study we investigate the domestic stock liquidity for European firms that are listed on the AEX, LSE, DAX and SBF. Other exchanges are not included in the sample because of the lack of accurate data. Therefore, when mentioning European firms in this study this does not include all European listed firms but rather European firms that are listed on the mentioned exchanges.
The contribution of this research is that there has not been empirical research on the
differences in the level of stock liquidity between European firms that are cross listed to the U.S. and those that are not, during the current global credit crisis. The paper of Frino, Di Marco and Lepone (2009) investigates the differences in domestic stock liquidity for a developed country (Australia) between companies that are cross listed to the U.S. and those that are not. However, the research was done during economic an economic stable time, which is different from this research where there has been a heavy global financial crisis. In the paper of Chandar et al. (2007) the authors investigate the difference in reaction of the stock price for emerging countries during a domestic currency crisis and conclude that cross listed firms react less negatively than non-cross listed firms. My research is a bit similar in spirit to Chandar et al. but my study will investigate the difference in the level of domestic stock liquidity between cross listed and non-cross listed firms for developed countries during the global credit crisis. The difference between the studied currency crises by Chandar et al. and the current financial crisis is that the financial crisis is global instead of domestic/regional, where a systemic shock affects all firms due to the demand factors and supply of credit, not just export and import firms as in a currency crisis. Also, the current financial crisis is considered as the heaviest economic crisis since the great depression in the 1930’s. Finally, the cross listed firms in the study of Chandar et al. were cross listed to a country where there was financial stability at that time; this is different in this research where there is a crisis in the cross listing destination as well.
It is important to note that the financial crisis that started in 2008 is characterized by the collapse of many large institutions and commercial banks and has caused a large downturn on stock markets around the world. Moreover, since in the recent decades there has also been a rapid growth in institutional ownership of common stocks and an increasing presence of
8 institutional traders in stock markets, the collapse of these financial institutions causes a decrease in stock market activity and uncertainty on the stock market (Blume & Keim, 2012). Finally, the global banks financing firms (with credit) faced funding problems and were unable to provide liquidity to firms during the financial downturn, which had severe effects on the economy.
In order to answer the research question, a regression model is constructed where a dummy variable that equals one if the firm is cross listed to the NYSE and zero otherwise, is used as the main independent variable (this creates a treatment- and a control group). The relative bid-ask spread (bid-ask spread divided by the price midpoint) and the turnover ratio (stock price times trading volume divided by the market capitalization) are used as dependent variables in order to capture several components of stock liquidity (section five of this study will give a short overview of the control variables and dependent variables that are used in the empirical literature). The data will come from DataStream. The control group (European firms that are not cross listed to the U.S.) is selected on the basis of the same level of market capitalization and the same type of industry as the treatment group (firms that are cross listed to the U.S.). This means that in order to control for industry specific effects, every firm of the control group was linked to a firm of the treatment group that was active in the same industry and had approximately the same market capitalization. Therefore, the number of firms that are in the control group is the same number as the treatment group, 31.
However, since the market capitalization of the firms in the control- and treatment group differed too much, it is important to still implement the market capitalization as a control variable in the regression. This gives a sample of 62 firms (where all 62 are listed on the FSE) over a 4-year time period (2008-2012) with daily data for the dependent and independent variables and a total of observations of 61,153.
Due to the fact that the financial crisis has hit financial institutions and firms harder than non-financial companies, cross listing to the U.S. might have a different effect for these non-financial firms in terms of stock liquidity compared to non-financial firms. In order to test whether non-financial firms that cross list to the U.S. find a different effect on their stock liquidity, a similar regression for only non-financial firms will be performed.
The results after the main regression analysis show that there is not a significant difference, at a 5% level, in domestic stock liquidity between European firms (financial and non-financial firms) that are cross listed to the U.S. and those that are not, for both regression models during the recent financial crisis (when looking at the results of the regression analysis for non-financial firms, the difference in stock liquidity measured through the relative bid-ask
9 spread, is in fact significant at a 5% level, which is interesting material for further research). The insignificance might be explained by the fact that the significance of the U.S. as a cross listing destination has declined due to the fact that almost every European firm is already cross listed to the German Stock Exchange which has a high level of corporate governance and is a highly liquid market. Also, it can be explained by the fact that the systemic crisis affects the access to credit for all firms (U.S. and European), which reduces stock liquidity. The paper is constructed as follows; in section two there is an overview of the recent trends on the international stock market, the empirical theories on cross listing and a literature review. Section three provides a variables description of the variables used in this research. The fourth section discusses the methodology. Section five gives an overview of the used data. Sections six presents the results of the regression analysis and section seven carries out the robustness tests. The conclusion is covered in section eight and section nine discusses the limitations, implications and the direction for future research.
2. Background
2.1 Recent Trends
In recent years many other exchanges have become a popular cross listing destination. At this moment almost every firm on the NYSE is cross listed to the Deutsche Boerse (Open Market) due to its economic stability and the low requirements. Since many listed firms all over the word do cross listing, it is important to understand the incentives of these companies and also the potential benefits and accompanying costs (Brockman & Chung, 1998).
The German equity market has gained a substantial number of listings on the stock exchange from companies all over the world the last years. Because the German stock exchange is the most developed and primary market of Europe, many European, Japanese and U.S. firms have decided to cross list their shares in Germany. Also, the listing itself is not expensive and quick and when a firm wants to list on several German equity exchanges it only has to
comply with one institutional environment.
However, it is important to make a distinction between the Open Market (Regulated Unofficial Market) and the Regulated market which are both German equity markets. The regulated market is a stock exchange where certain requirements have to be satisfied in order to list shares. On the Open Market, issuers are not obliged to meet requirements and is therefore characterized as an exchange with few corporate governance rules and regulations
10 which makes it an interesting destination for firms that are not in need to ‘signal’ their
credibility through cross listing to Germany (Boerse Frankfurt, 2010).
Due to the low requirements and low listing costs, the Open Market is a very popular destination for U.S. firms to issue their shares to European investors and increases their investors base. However, this also means that there is less information available on listed firms for the investors (Valero et al., 2009).
The most popular German equity market is the Frankfurt Stock Exchange, which is an
Unregulated Open market. The Frankfurt Stock Exchange accounts for over 90 percent of the turnover in the German market and a very large share of the European market. Almost half of the 300 market participants in Frankfurt are foreign investors and more than 80 countries list on the FSE, as of 2010. Many firms are listed on the FSE is due to the fact that when a firm is already listed on a foreign stock exchange, it is automatically listed on the FSE without needing permission of the firm (Boerse Frankfurt 2010).
Another important incentive for firms to be listed on the German Regulated Official Market is that the German equity markets have been under constant construction in order to create the optimal level of corporate governance for firms as well as investors, in contrast to the NYSE or NASDAQ where there has not been many changes recently. Also, the Regulated Market is constantly implementing rules and requirements, which increases the disclosure of information and tries to protect the small investors. Of all the European countries, Germany has implemented the highest amount of modifications in their legal framework so that it can match the level of corporate governance of the U.S. which keeps making Germany more and more interesting for foreign firms to issue their shares (Boerse Frankfurt, 2010)
However, note that there are many other exchanges up and coming. Take for example the NYSE Euronext, which includes over 8000 listed issues and comprises 90% of the Dow Jones Industrial Average and is the leading European derivatives exchange by value of trading. Also, the Hong Kong Stock Exchange (SEHK) and the Singapore Stock Exchange (SGX) have seen an increase in foreign firms that list their shares on their exchange (Peng and Su, 2013).
11 2.2 Theory
2.2.1 Market Liquidity Hypothesis and Liquidity Definition
The Market Liquidity Hypothesis notes that an important incentive for foreign companies to cross list to the U.S. is that the cross listing premium (increase in firm value or Tobin’s Q) can be attributed to the increase in stock liquidity. This can be explained by the fact that high stock liquidity attracts more investors because they can buy and sell stock easily, quickly and at a low cost when a firm is listed on a foreign exchange. This generates a high level of trading activity and can therefore be a predictor of the increase in firm value. Also, liquidity can be seen as risk reducing since during high volatile times investors tend to invest in more liquid assets.
Since there are multiple definitions of liquidity, it is important to first define the term ‘liquidity’ that is used in this research. There is a difference between the short-term firm liquidity and the liquidity of a stock. The short-term liquidity of a firm is the ability of a company to meet the short-term obligations i.e. whether a firm has enough liquidity to pay its short-term debts. An improvement in this type of liquidity is an important factor for firms since several studies have indicated that there is an inverse relationship with the cost of capital and that it has a positive effect on the stock return. Also, when a firm is unable to pay creditors on time and incapable of meeting obligations to the suppliers of credit, services, and goods it can be very problematic for firms. It may affect the firm’s operations and reputation (Lins, Servaes and Tufano, 2010).
The liquidity of a stock is an asset's ability to be sold without causing a significant movement in the price and with minimum loss of value. In other words it refers to how easy it is to buy and sell shares without seeing a difference price after the transaction. A market may be considered deeply liquid if there are buyers and sellers in large quantities. This is related to market depth which can be measured as the units that can be sold or bought for a given price impact. The opposite is that of market breadth which is measured as the price impact per unit of liquidity. This research focusses on the liquidity of the firm’s stock and will use estimators that can measure the several aspects of stock liquidity (Fang, Noe and Tice, 2009).
12 2.2.2 Additional Hypotheses
There are three wide-known additional hypotheses that explain the motivations for why firms cross list. First, the market segmentation hypothesis implies that there are certain barriers to international investment such as information asymmetry, differences in corporate governance or differences in taxes, which cause that international capital markets are segmented. Through cross listing, this segmentation of markets evaporates and, therefore, cross-listed securities may trade at different prices because internationalization can lower a firms’ cost of capital. Also, it can facilitate corporate expansion relative to firms that do not internationalize because through cross listing the risk is more widely shared (Gozzi et al. 2008). Chandar, Patro and Yezegel (2007) note, however, that even though there has been a
continuing increase in liberalization of the equity markets there is still a segmentation of the financial markets due to legal restrictions, a lack of investor protection and costs associated with obtaining information on foreign firms. Due to this segmentation of the markets many firms decide to cross list to stock exchanges abroad. Also, the authors agree with the findings of Stulz (1981) and Gozzi (2008) where they expect that cross listed firms are priced globally instead of only local (where investors meet a higher cost of capital because the risk is for the most part borne by the investors of the non-cross listed home country) and therefore will react different to changes in the business cycle.
Second, the signalling hypothesis states that when firms cross list they have to meet certain corporate governance or accounting requirements, which can be stricter on foreign exchanges. When a firm can satisfy these requirements and is willing to decrease the information
asymmetry between the firm and the investor, this gives a good signal to many investors. In other words, the firm ‘signals’ that it is willing to disclose information and meet higher requirements which makes it more interesting for investors to purchase their shares since they know the fair value of the asset. This increases the liquidity and can have a positive effect on firm value (Kim and Verrechia, 2004).
Finally, the bonding theory argues that firms cross list to bond themselves to a better corporate governance framework. Improved governance lowers a firm’s cost of capital, which reduces the chance of expropriation of corporate resources by firm insiders, who thrive for an increase in firm value, but also an increase in private benefits. Like the market
segmentation theory, the bonding theory predicts that cross listing lowers capital costs, causing the firm value to rise (Gozzi et al. 2008). The bonding theory has similarities with the signaling hypothesis since both hypotheses emphasize the importance of decreasing the
13 information asymmetry between investors and companies. Most empirical research has focused on this widely known hypothesis and finds mixed support. There is still an ongoing debate about which hypothesis can explain the motivations of the management to cross list and how cross listing creates a premium.
Cross listing during a crisis period can have significant benefits for firms since Karolyi (1998) and Chandar, Patro and Yezegel (2007) have analysed the different stock price responses for firms that have cross listed to the U.S. through ADR’s in their home countries during a crisis and found that these cross listed firms react less negatively than non-cross listed firms. Studies have shown that cross-border listing reduces a firm’s cost of equity capital when markets are liberalized since the securities are then re-priced according to the world market price of risk (Errunza & Miller, 2000). In other words, cross listed firms may be priced globally and therefore are less sensitive to local market risks. These findings are consistent with the international asset pricing model of Stulz (1981), where due to the absence of international investment barriers or segmentation of the market, the cross listed firms are priced globally and are therefore less sensitive to a local crisis.
2.3 Literature Review
2.3.1 Stock Liquidity
Several authors have discussed stock liquidity and its incidentals. Vayanos (2004) indicates that during crisis periods, investors tend to prefer holding liquid assets, which is called ‘flight to quality’ or ‘flight to liquidity’. Therefore, participation in small firms may decrease when the economy worsens or can even lead to a shift from the stock market to the market for Treasury Bonds. This means that during a downturn in the business cycle it can have severe effects for firms that are active on the stock market.
Pastor and Stambaugh (2003) show that liquidity is a risk factor that is priced in the market and that liquidity can also influence the capital structure of a firm since a company is more likely to issue shares to finance its investment when the stock market or the stock liquidity of the firm itself is high. Finally, the liquidity of a firm can have a stabilizing effect on the market since it is then less sensitive to changes in the business cycle (Wuyts, 2007).
Wuyts also describes the features of a liquid market since this also has a significant influence on the decision to cross list. A liquid market tends to have five characteristics: tightness (low transaction costs as well as implicit costs), immediacy (speed at which orders can be executed
14 and settled), depth (existence of abundant orders), breadth (the orders are large and high in volume with minimal impact on prices) and resiliency (the speed at which the market corrects imbalances). There are multiple indicators to measure liquidity for a stock which can be categorized into three categories: transaction cost measures, volume-based measures and price impact measures but there is no measure that unequivocally measures all the five characteristics. Therefore, multiple indicators for stock liquidity have to be used in order to measure the level of liquidity correctly.
Berkman and Nguyen (2010) concluded that when markets are transparent, cross listing can improve liquidity due to an additional order flow from foreign market participants and
foreign stock markets have different corporate governance standards that firms have to follow, so when they cross list this can increase transparency, which increases liquidity.
However, during crisis periods, which was the case during the period of their research, there was uncertainty and in certain extreme cases stocks may be subject to periods where liquidity suddenly evaporates, effectively eliminating the opportunity for a trader to enter or exit a position at all (Lang, Maffett, 2011). Brunnermeier and Pedersen (2009) provide a model where the ability of the provider of liquidity to provide actual liquidity depends on their margin requirements. In periods of a financial recession or crisis a reinforcing mechanism between market liquidity and funding liquidity leads to liquidity spirals. Reduced funding liquidity leads to a ‘flight to quality’ because liquidity providers shift their liquidity provision towards stock with low margins.
The paper of Vayanos (2004), finds the most support, which indicates that during crisis periods, investors tend to prefer holding liquid assets and that participation in small firms may decrease or will lead to a shift from the stock market to the market for Treasury Bonds. The theory on whether the change in liquidity of cross listed firms during a period of
financial distress compared to non-cross listed firms is different, is rather limited. However, the paper of Silva and Chavez (2007) studies whether local firms that are cross listed present, on average, a liquidity advantage due to lower trading costs or higher quality of information linkages over non-cross listed ones but this improvement is not uniform across all firms that were tested.
2.3.2 The U.S. Stock Market
During the 1980’s and 1990’s many firms chose to issue their shares on the U.S. stock market through listing (direct investment) or through a Depositary Receipt program. Doidge, Karolyi
15 and Stulz (2004) conclude that firms who cross list to the U.S. find a risk premium reduction because a cross listing to the U.S. means that the risk of the firm is shared more widely due to an increase in the shareholder’s base and it also can give access to more developed markets for emerging countries since the U.S. has deep and liquid secondary capital markets. This means that the firms can raise capital at a lower cost than at home. Also, due to the high requirements on the disclosure of information it can give a positive signal due to a decrease in information asymmetry, which can have an upward effect on firm value (signalling
hypothesis). Finally, they note that a U.S. listing enhances the protection of the firm’s
investors and, consequently, reduces the agency costs of controlling shareholders. Compared to the rest of the world, investors are extremely well protected in the U.S., which may also be due to the high level of monitoring, by the press and analysts. Lang, Lins and Miller (2003) find that when a firm decides to cross list to the NYSE the number of analysts that monitors the firm increased with 128%.
The reason why not more firms have cross listed to the U.S., according to Doidge (2004), is not because of the direct costs that are associated with the listing (the legal costs and
investment banking fees are in many cross listings paid by depositary banks) or the Securities and Exchange Commission (SEC) reporting and compliance requirements, but because of the fact that most large companies are controlled by a small share of large shareholders. These firms will only cross list to the U.S. when it comes with certain (personal) benefits for these large investors. Since this is not always the case, not all firms decide to cross list on the U.S. stock exchange. For example, the study of Southam and Sapp (2010) find that the executive compensation can be an important incentive for firms to cross list to the U.S. because the CEO’s of non-U.S. firms may earn compensations that are comparable to U.S. CEO’s. This of course could encourage the board of directors to cross list to the U.S.
Companies that want to list their shares on any of the U.S. stock exchanges, through a direct listing, have to satisfy two requirements. The firm’s pre-tax income must exceed $25 million in one of the three latest years and the market value of publicly held shares must be at least $100 million. Second, the firms must comply with the GAAP reporting methods and the SEC registration requirements, which require a higher level of (information) disclosure than many other exchanges around the world.
Indirect listing can be done through issuing Depositary Receipts which are negotiable equity instruments that indirectly represent ownership of the shares in the company for domestic investors and gives these investors the opportunity to diversify without purchasing from international exchanges. An American Depositary Receipt (ADR) represents a certain
16 number of underlying shares in the domestic market and these certificates were developed in order to earn dividends on foreign stock without direct access to the local market itself. The depositary bank holds the certificates for which the investors pays a fee and pays the
dividends in U.S. dollars to the investors, therefore the bank does not bear the currency risk. The difference between these two listing options is that with the Depositary Receipt program, the legal costs and investment banking fees is in many cases paid by depositary banks. Also, the D.R. program has to require the U.S. settling rules, which make the occurrence of a ‘trade fail’ very low (Karolyi, 1998).
Note that there are three categories or ‘levels’ in ADR’s that can be issued by a firm, where the first level means that the shares are traded over the counter (OTC) where firms do not have to be ADR listed on an exchange. Also, there are very low requirements by the SEC to trade shares OTC. The second and third level imply that the ADR is listed on an exchange (NYSE, NASDAQ, AMEX) and means that the firm is required to meet the SEC obligations and requirements. The difference between these two levels is that the level two DR program only uses existing capital as the underlying stock, where the level three DR program uses both existing and new capital (Frino et al. 2009).
A large amount of empirical researches have examined the effects of cross listing to the U.S. stock market and many found that there were indeed significant increases in firm value and stock liquidity. However, there has been much debate the recent years, about the significance of cross listing to the U.S. because of the Sarbanes-Oxley (Sox) act in 2002 and the fact that many U.S. and European firms have cross listed to the Deutsche Böerse.
The Sarbanes-Oxley act was passed by the U.S. congress because of dubious accounting procedures by WorldCom and Enron that have brought big financial and economic
instability). Therefore, this act was designed to raise the level of corporate governance and transparency of firms listed in the U.S. This has made the U.S. less popular for investors on one hand because they now had to satisfy certain requirements before they could increase their investors base. On the other hand it did give the listed firms more credibility due to the fact that they could satisfy the new requirements.
In the study of Doidge, Karolyi and Stulz (2007) they conclude that during 1990-2005 (before, during and after the Sox act) the (Tobin’s Q) premium has not fallen significantly in recent years and that the U.S. stock market still has unique governance benefits for foreign firms. In other words, the benefits of cross listing to the U.S. have not been eroded by the Sarbanes-Oxley act.
17 3. Variables Description
In order to give answer to the research question, it is important to construct several hypotheses that can be empirically tested. Due to the fact that liquidity has multiple
definitions and can be measured in several ways, the method of using several measurers for liquidity has been done by many researchers in order to measure the five characteristics of high stock liquidity; tightness (low transaction costs as well as implicit costs), immediacy (speed that orders can be executed and settled), depth (existence of abundant orders, breadth (the orders are large and high in volume with minimal impact on prices) and resiliency. As stated before, the research question wishes to empirically test whether European firms that cross list to the NYSE, through ADR’s, have had a significant higher level of stock liquidity than European firms that did not cross list to the U.S. during the current financial crisis. In the empirical literature, three types of measurers can measure the components of stock liquidity: transaction cost measures, volume-based measures and price impact measures. In order to construct a good framework to measure the level of liquidity, at least two estimators have to be used, where in this research the transaction cost measurer and volume-based measurer will be used. The transaction cost measure will be represented by the relative bid-ask spread of the stock (the bid-ask spread divided by the price midpoint) and the volume-based measure is represented by the turnover ratio (number of shares trade times price divided by the market capitalization).
In the empirical literature there has been much debate about what estimator to use in order to measure the level of stock liquidity. This comes from the fact that liquidity is somewhat abstract and is therefore defined different by many researchers. Because of this disagreement and the use of different estimators for stock liquidity, it is important to give a short overview of the dependent variables that have been used in the empirical literature before starting with the methodology. But first, the incentives for the used control variables in the regression model are discussed.
3.1 Control variables
The control variables used in this research will be the daily volume of traded shares (Volume), the market capitalization of the firm, the volatility of the stock return and a dummy variable (D_CrossMult.) which equals one if a firm is listed on more than one foreign exchange and zero otherwise.
18 perfect matched sample for a treatment- and control group and because variations in bid-ask spreads and turnover ratios are caused by variations in volume, volatility and price. Therefore, it is important to add these variables in the regression as control variables in order to have an unbiased model (Brockman, Chung, 1999).
The daily volume of traded shares has a significant influence on the bid-ask spread (as well as the turnover ratio). For example, a frequently traded share will have a positive relationship with liquidity because when a share is traded many times on a trading day, there will be a high level of supply and demand which increases the probability of fulfilment of an order and also means that a transaction will have little effect on the stock price. This little effect on the stock price after a transaction implies a high level of liquidity.
The market capitalization has a positive relationship with stock liquidity since it has a positive relation with transaction volume and it decreases the information asymmetry since firms with a high market capitalization usually have better analysts and media screening which decreases the adverse selection costs resulting. This makes a firm with a low(er) market capitalization less liquid and it is therefore required to implement the firm’s market capitalization in the regression (Rosch & Kaserer, 2013).
The volatility of the stock return can affect stock liquidity because a high volatility can mean that a transaction cannot be fulfilled due to adverse price changes of the stock i.e. it increases the risk of holding a stock which can increase the transaction costs. This can make investors less interested in the stock due to possible significant price changes, which will have its effect on stock liquidity.
Finally, due to the fact that many firms in the sample are listed on more than one foreign exchange, it is important to use a second dummy variable since multiple listings can already give a firm the advantages of a larger investors base and diversification of risk and can
therefore diminish the significant of cross listing to the NYSE. This larger investors base will, of course, increase trading which tightens the relative bid-ask spread and has an upward effect on the turnover ratio. Therefore, this control variable is also included in the regression.
3.2 Dependent variables
Even though liquidity seems a straightforward concept at first sight, it is still not entirely clear what comprises liquidity. Due to the lack of a uniform definition or estimator in the empirical literature, it is important give a short description of the measures for liquidity used in previous research. As noted before it is not possible to measure all the five characteristics of liquidity by using one dependent variable. Therefore, many researchers have used multiple
19 measurers for liquidity in their researches in order to have an unbiased estimate.
Berkman and Nguyen (2010) use four estimators for liquidity, the quoted bid ask spread, the price impact (average price impact of a transaction), PIN (measurer of symmetric order flows) and turnover ratio. However, due to the fact that the database DataStream or any other
database available at the UvA is used for this research and does have some limitations on the availability of certain data, it is not possible to acquire the price impact and the PIN as variables i.e. these variables are too complex to derive from DataStream or any other UvA database and are therefore not included in the regression.
Naes et al. (2011) use the relative spread, the LOT (Lesmond model, comprehensive measure of liquidity costs that includes the explicit, implicit, and opportunity costs of trading), the ILR (illiquidity ratio) and the Roll liquidity estimator (implicit spread estimator) as liquidity measure.
Sarr and Lybek (2002) use the relative bid-ask spread as an estimator as well. Also, the turnover and the Market-Efficiency Coefficient (Var of the logarithm of long period returns / (number of short periods in each longer periods * variance of the logarithm of short periods) are used as dependent variables, where the Market-Efficiency Coefficient is very complicated to derive.
As one can tell, even though there are many good estimators for liquidity in the available empirical literature, this does not mean that they are used in this research. However, with the dependent variable that will be used it is guaranteed to have a clear picture on the level of liquidity since many studies have used these estimator before, are still widely used and are considered good indicators of the liquidity of a firm’s stock.
In the paper of Lang & Maffett (2011) another estimator is used known as the Amihud price impact measure (DPI) which is defined as the daily percentage price change divided by the daily stock price times the daily trading volume. This is a good price impact measurer and it is possible to derive this measure from DataStream. However, when using two estimators (relative bid-ask spread and the turnover ratio) as measurers for liquidity, it will be enough to get a clear picture of the level of liquidity.
Note that the relative bid-ask spread will be calculated as (closing bid-ask spread / price midpoint) and the turnover ratio (number of daily traded shares * closing stock price / market cap) in the regression model.
20 4. Methodology
In order to test the hypothesis that European firms who are cross listed to the NYSE, during the current financial crisis, find a significant higher level of stock liquidity than European firms that are not cross listed to the NYSE, an OLS regression is used. The OLS estimator is consistent when the regressors are exogenous and there is no perfect multicollinearity. Also the residuals are required to be homoscedastic and serially uncorrelated. Finally, the errors are expected to be normally distributed. When these conditions are met, the OLS is an unbiased maximum likelihood estimator. As noted earlier, the methodology of the paper will have similarities with the paper of Brockman and Chung (1999) and the paper of Frino et al. (1998), as daily trading volume and volatility are used as control variables and the relative bid-ask spread and the turnover ratio are used as liquidity measures.
The control group is selected on the basis of market capitalization and industry. This means that for the 31 European firms of the dataset that are cross listed to the U.S. during the period of 2008-2012, each of these firms is matched with an European firm that is not cross listed to the U.S., is active in the same industry and has approximately the same market capitalization and is added to the control group. This is done in order to have, approximately, a similar treatment- and control group. In other words, this gives an equal amount of firms in the control- and treatment group, where each firm in the treatment group is matched with a firm that is not cross listed to the U.S., that has the same market capitalization and is active in the same industry.
Note, however, that firms that are cross listed to the U.S. have a higher market capitalization than firms that are not cross listed to the U.S. This creates a difference between the control- and treatment group. Therefore, it is important to still implement the market capitalization as a control variable in the regression. The other independent variables used are a dummy variable that equals one if the firm is cross listed to the NYSE and equals zero otherwise, the volatility of the stock return calculated over 10-trading days for each firm, another dummy variable that equals one if the firm is cross listed to multiple foreign exchanges and zero otherwise (in order capture the increase in investors base and other advantages of cross listing to multiple stock exchanges) and the number of shares traded on a trading day. Finally, there are five additional dummy variables implemented in the regression where each dummy variable represents a certain year of the dataset (2008, 2009, 2010, 2011, 2011, 2012). For instance, the dummy variable for the year 2008 will equal one if the data is from 2008 and zero otherwise. These year dummies are added since implementing these variables will get
21 more reliable information out of the dataset.
Note, however, that the volume and the market capitalization will not be used as control variables in the second regression formula. This is due to the fact that the turnover ratio is calculated with the use of the market capitalization and the daily trading volume. When including the market capitalization and the daily trading volume, it will cause problems for the regression. Therefore, the OLS regressions are constructed as follows:
(1) Relative Bid-Ask Spread = β0 + β1 D_CrossUS it + β2 Volume it
+ β3 D_CrossMult.it + β4 Volatility it + β5 Market Cap.it + β6 D_2008 + β7
D_2009 + β8 D_2010it + β9 D_2011 + β10 D_2012 + ε it
(2) Turnover Ratio = β0 + β1 D_CrossUS it + β2 Volatility it +β3 D_CrossMult it +
β6 D_2008 it + β7 D_2009 it + β8 D_2010 it + β9 D_2011 it + β10 D_2012 it + ε it These regressions have been used by Brockman & Chung (1998) and Frino, Di Marco and Lepone (2009) in order to compare the level of stock liquidity between firms that are cross listed and those that are not. Both studies find significant evidence for their hypotheses. It will be interesting to test whether cross listing to the U.S. still has a significant positive effect on stock liquidity for European firms or whether the developments in the German Stock Market and many other changes in the international stock market has made the NYSE abundant as a cross listed destination for European firms that wish to increase their stock liquidity.
In order to test these hypotheses, the data for the relative bid-ask spread, turnover ratio, the volatility of the stock return, the number of shares traded and the market capitalization is needed. Also, it is important to find out what European firms are cross listed to the NYSE and in what type of industry they are active.
With these two regressions, it is possible to test the constructed hypotheses and helps give answer to the research question: Did cross listing to the U.S. lead to a higher level of stock liquidity for European firms during the current financial crisis, compared to European firms that were not cross listed to the U.S.?
Note, however, that both regression models are not the typical fundamental economic models in the sense that it does not investigate how much the dependent variable changes after a change in the main independent variable while controlling for other factors. It rather tries to show whether there is a difference in stock liquidity for firms that are cross listed to the U.S. and those that are not, while using a dependent variable that tries to capture the level of
22 liquidity (on which there still is uncertainty and discussion in the empirical literature on how to measure this). This makes it hard to construct a solid, unbiased model and therefore, a low R-squared and potential biases is expected to be present. However, with the available
empirical literature it is possible to construct a good, solid regression model.
5. Data
The data that has been used in the research has come from the financial database DataStream and will cover the period 2008-2012 of 62 firms. With this database it is possible to find the European firms (companies listed on the AEX, DAX, LSE and SBF) that are cross listed to the NYSE through A.D.R.’s. In total, 33 of these European firms have cross listed to the NYSE. The reason why this research only focuses on the NYSE and does not include the NASDAQ or any other U.S. stock exchange is because almost all of the European firms listed on the NYSE are also listed on the NASDAQ (or any U.S. stock exchange) and vice versa. This means that when including European firms cross listing to the NASDAQ in the research, it would not give a much larger sample.
When looking at these firms in the years 2008-2012, a two of these listed companies have removed their shares from the NYSE during this period crisis which gives a sample of 31 European NYSE listed firms. When composing the control group for the sample, there was searched for firms with approximately the same market capitalization, even though the market capitalization will be used as a control variable, in order to keep the control and treatment group as similar as possible. Also, there is a possibility in DataStream to display the type of industry the company is in where the industries are categorized as followed:
Table 1: Industry Categorized 01 Industrial
02 Utility
03 Transportation
04 Bank/Savings & Loan 05 Insurance
06 Other Financial
In order to control for industry specific effects, every firm of the control group was linked to a firm of the treatment group that was active in the same industry. Therefore, the number of
23 firms that are in the control group is the same number as the treatment group, 31.
This gives a sample of 62 firms over a 4-year time period (2008-2012) with daily data for the dependent and independent variables and a total of observations of 61,153. Note, however, that since DataStream carries forward price data this means that there has not been a transaction that day; the price data at days where there was no transaction volume were removed (Rosch and Kaserer, 2013).
The reason why almost the entire timespan of the financial crisis has been chosen as the timeframe for the research and why European firms are studied instead of one country in particular, is because when using daily data for this type of empirical research it is important that liquidity is studied over relatively long timeframes and across multiple countries, as opposed to intraday data (Goyenko, Holden and Trzcinka, 2009).
Table 2 gives the descriptive of the variables and it is interesting to see that the standard deviation of the relative bid-ask spread is relatively high which suggests high volatility. Also, the market capitalization can differ strongly between firms even though the sample has been selected based on the same market capitalization.
Note that the market capitalization is presented in millions and the volume in thousands since the values of these variables differ too much from the dependent variable, which can give unclear coefficients.
Table 2: Descriptive of Variables
Variable Mean Std. Deviation Min Max
Relative Bid-Ask Spread 0,00098 0,0085 0 0,6635
Turnover Ratio 3,8434 3,3573 0,0319 33,2154 D_CrossUS 0,5326 0,4989 0 1 Volume 11,3756 28,841 0,053 894,6091 D_CrossMult. 0,6668 0,4713 0 1 Volatility 0,0228 0,0174 0,0007 0,3712 Market Cap. 1307,4963 2478,5033 2,0495 14006,3455
The total sample consists of 61 European firms, where there are 31 firms cross listed to the NYSE and 31 are not. Of those 61 firms, every firm is cross listed to the Frankfurt Stock Exchange (FSE) and 40 of these firms are cross listed to more than one stock exchange. Also, 20 of these firms are active in the financial sector, 42 are not. There are no large differences in the mean of the dependent variables of each group, however, the average relative bid ask spread for firms that are cross listed to the NYSE is in fact lower than firms that are not cross listed to the NYSE, which gives support to the empirical literature and the main hypotheses.
24 Also, the turnover ratio for firms that are cross listed to the NYSE is higher than firms that are not cross listed to the NYSE which also supports the main hypothesis. This might give evidence for the fact that there is in fact a significant difference in stock liquidity.
Table 3: Mean of the Relative Bid-Ask Spread and Turnover Ratio per group
Group N Mean Relative Bid Ask Spread Mean Turnover
Ratio Cross Listed to NYSE
31 0,0008 3,8096
Not Cross Listed to NYSE
31 0,0011 3,5674
Cross Listed only to FSE
22 0,0012 3,5159
Cross Listed to FSE and other
foreign stock exchanges 40 0,0008 3,3065
Financial Firms 20 0,0011 3,8543
Non-Financial Firms 42 0,00083 3,7934
6. Estimation Results
In the first regression of regression model 1 and 2, the main independent variable D_CrossUS is statistically significant at a 1% level, but there is an extremely low R-squared value since the turnover ratio and especially the relative bid-ask spread is determined by many other factors. When only adding this dummy variable as explanatory variable to determine the relative bid ask spread, it will give a low R-squared. This result means that cross listing to the U.S. has in fact had a significant effect on stock liquidity for European firms during the financial crisis (the same result holds for the regression where only non-financial firms are included in the sample as is shown in Appendix 3 and 4). Again, it is important to note that both regression models are not the typical fundamental economic models. It rather tries to show whether there is a difference in stock liquidity for firms that are cross listed to the U.S. and those that are not. This makes it hard to construct a solid, unbiased model and therefore, a low R-squared is expected to be present.
In the second regression the volatility of the stock return (Volatility) is added and in both regression models the dummy variable stays significant at a 1% level, where the volatility is
25 insignificant in regression model 1 (where the increase in R-squared can almost be neglected) but significant in regression model 2 (where there is a small increase in R-squared). This means that the volatility of the stock price affects the level of stock liquidity.
Regression 3 adds D_CrossMult. into the regression (dummy variable that equals one if a firm is cross listed to more than one foreign exchange and zero otherwise) which is
considered significant in both regression but reduces the significance of D_CrossUS and is considered significant at a 10% level in regression model 1 but insignificant in regression model 2. In both regression models there is a slight increase in R-squared. The significance of this variable means that cross listing to more than one exchange, other than the FSE, has in fact a significant effect on the level of stock liquidity for a firm during the current financial crisis.
In the fourth regression for regression model 1 (the dependent variable is the relative bid-ask spread) there is another variable added, namely the Volume (in thousands) which stands for the number of shares traded on a trading day of a particular stock. It is clear to see that adding this significant variable further reduces the significance of the main independent variable but does increase the R-squared. However, the R-squared is still considered low for an OLS regression model.
When adding the market capitalization (in millions) in the first regression model, the level of significance for D_CrossUS increases but is still not significant on a 10% level. Also, the R-squared increases but remains low. The significance of this variable means that the higher the market capitalization of the firm, the higher the level of domestic stock liquidity.
In the last regression for both regression models, the year dummies are added. This does not change the significance of the main independent variable. However, after adding the year dummies the market capitalization (regression model 1) becomes insignificant. Since market capitalization has been a significant factor in many empirical researches, this might indicate over specification of the model. Also, it is interesting to see that the dummy variables for the years 2008 and 2009, when the crisis was at its peak, are significant in regression model 1 and 2 but insignificant in Appendix 3 and 4 (dummy variables for 2010, 2011 and 2012 are insignificant in regression model 1 and 2 but significant in Appendix 3 and 4).
Also, when regressing for the sample of non-financial firms, D_CrossUS is in fact significant at a 1% level (when the turnover ratio is the dependent variable), which means that after adding several control variables there is in fact a significant difference in the level of stock liquidity for European firms that are cross listed to the U.S. and those that are not. This gives room for further research.
26 Since the D_CrossUS is not statistically significant at a 5% level in regression model 1 and 2 (when including financial and non-financial firms in the sample), it is possible to conclude that there is not a significant difference in domestic stock liquidity between European firms that are cross listed to the U.S. and those that are not cross listed to the U.S. during the current financial crisis. In other words, cross listing to the U.S. did not give significant benefits for European firms in terms of stock liquidity, during the current financial crisis. This may be due to the fact that there have been made many advancements in the
international stock exchange market, such as has been made on the Deutsche Börse which is a popular cross listing destination for European firms. Therefore, an additional cross listing to the U.S. may not have a significant effect for the stock liquidity during the current financial crisis.
As noted earlier, there are several hypotheses that describe the incentives for cross listing and the potential benefits that it can give to firms. The results show no significant difference in stock liquidity, which is not in line with the empirical literature. However, there is in fact a significant correlation between firms that are cross listed to more than one exchange and the stock liquidity. This could give evidence to the market segmentation theory since cross listing to multiple destinations has a significant positive effect on the domestic stock liquidity of a firm, which might be attributed to a higher investors base.
It is too farfetched to attribute the results of this study to the theory of Vayanos (2004) where he states that there is a ‘flight to quality’ or a ‘flight to liquidity’ during a crisis period. This event may cause that investors tend to shift their investments from the stock market to the Treasury Bond market which could create a decrease in demand for listed companies and therefore has its effect on stock liquidity.
In the study of Doidge, Karolyi and Stulz (2007) they conclude that after the Sarbanes-Oxley act in 2002 the benefits of cross listing to the U.S. have not been eroded. This research, however, shows that the U.S. stock market does not give significant benefits for European firms in terms of stock liquidity during a crisis.
7. Robustness Test
As discussed in section 6, the R-squared of the regression model in this study is rather low in comparison to other research that implies a similar specification. Brockman and Chung (1999) find an R-squared of 0,4807 for the first regression model. In this regression their dependent variable is the relative bid-ask spread and the control variables are the daily trading volume,
27 the volatility of the stock return and the stock price, which has large similarities with the first regression of this study. In their second regression they find an R-squared of 0,2565, where their dependent variable is the average dollar depth and the control variables are the daily trading volume, the volatility of the stock return and the stock price. The difference in R-squared might be explained by the fact that they use intraday data (5-minute time intervals), which can provide a better picture of certain patterns of the stock’s liquidity on a trading day than daily data.
Testing for autocorrelation can be done with a Durbin-Watson model specification test where (d = 1,7289 – 1,809). The interpretation of this is that when the Durbin-Watson statistic for one of the independent variables is below 1,7289 there is in fact autocorrelation. A value higher than 1,809 means that there is no sign of autocorrelation, a value between 1,7289 and 1,809 means that the test is inconclusive. The Durbin-Watson statistic for the independent variables shows significant residual autocorrelation in regression (2). The effect on the regression might be that the standard errors tend to be underestimated and the t-scores
overestimated. In the paper of Brockman and Chung the same regression model (also without lags) is used but there is no presence of autocorrelation. By performing the Breusch-Pagan test heteroskedasticity is detected. The resulting problems of heteroskedasticity, such as biased estimates of coefficient variances, the white heteroskedasticity-consistent standard errors are employed in the assessment of every regression.
There are no signs of strong multicollinearity between the independent variables as is shown in Appendix 5. However, there is in fact high correlation between D_CrossUS and
D_CrossMult but this is because of the choice of data (many firms in the sample that are cross listed to the U.S. are also listed on another exchange, for example the FSE) instead of a certain economic relationship that can give biased results for the regression.
8. Conclusion
This paper is based on the Market Liquidity Hypothesis which notes that an important incentive for foreign companies to cross list to the U.S. is that the cross listing premium (increase in firm value or Tobin’s Q) can be attributed to the increase in stock liquidity. This can be explained by the fact that high stock liquidity attracts more investors because they can buy and sell stock easily, quickly and at a low cost when a firm is listed on a foreign
exchange. This generates a high level of trading activity and can therefore be a predictor of the increase in firm value. Also, liquidity can be seen as risk reducing since during high volatile times investors tend to invest in more liquid assets.
28 In order to answer the main research question, a regression model is constructed where he relative bid-ask spread and the turnover ratio are used as dependent variables in order to get a good estimate of the domestic stock liquidity. In both regression, the dummy variable that equals one if the firm is cross listed to the NYSE and zero otherwise, is used as the main independent variable. The control variables implemented in the regression are the daily trading volume, a dummy variable that equals one if the firm is cross listed to more than one exchange (and zero otherwise), the volatility of the stock return and the market capitalization. The results of the regression analysis show that there is not a significant difference, at a 5% level, in domestic stock liquidity between European firms that are cross listed to the U.S. and those that are not, for both regression models.
The fact that there has not been measured a significant difference in domestic stock liquidity is not surprising due to the globalization of the stock market and the fact that there was a severe crisis in the country that the European firms were cross listed to. Also, the fact that especially the German Stock Market has made significant improvements in its corporate governance and has a highly liquid stock market, could make cross listing to the U.S less significant for European firms since most of them are already cross listed to the FSE. However, it is interesting to see that when firms are cross listed to more than one exchange this has a significant effect on stock liquidity, at a 5% level, in both regression models which gives evidence for the market segmentation theory.
9. Limitations and Further Research
As noted in paragraph seven, the model has it shortcomings when it comes to estimating the differences in domestic stock liquidity between European firms who are cross listed to the U.S. and those that are not. Apart from the fact that the four-year timespan is rather low and that the sample size is rather small (62 firms), there are also complications with the OLS regression model. Autocorrelation can give biased estimators as the standard errors can be underestimated. This may increase the significance of the variables in the model where it may cause OLS estimates of the variance and the standard error to be biased.
Apart from the above mentioned limitations of this study, there is still an ongoing debate on what the best estimator(s) is for stock liquidity. This study uses the relative bid-ask spread and the turnover ratio due to the fact that these are the most widely used estimators. However, these interpretations might change with future research.
29 This study is especially relevant for the management of European listed firms who have to decide whether to cross list to the U.S. considering the fact that there is the presence of the Frankfurt Stock Exchange. There is not enough evidence to rule the U.S. out as an interesting cross listing destination but the fact that the international stock market has made significant developments has to be acknowledged.
As a future research recommendation it can be interesting to also use a price impact measure (Amihud price impact measure) as the dependent variable for stock liquidity and check whether this gives different results. Another recommendation is to use time-intervals instead of daily trading data since the use of time-intervals can provide a better picture of certain patterns of the stock’s liquidity on a trading day, than daily data.
An interesting research will be to study what the additional significant effects are of cross listing to multiple stock exchanges. Since many firms are already listed on the FSE with a high liquid secondary market and a certain degree of corporate governance, or lack off (depends on which market is chosen to list on), it will be interesting to check the additional effects of another foreign listing. Also, it can be very relevant for firms to study what stock exchange, the FSE or the NYSE, gives the most advantages for a firm when cross listing. This has to be done with firms that are either cross listed to the FSE or the NYSE in order to measure the differences.
Finally, the fact non-financial firms that are cross listed to the U.S. find a significant higher level of stock liquidity than non-financial firms that were not cross listed to the U.S. is an interesting finding and is interesting material for further research.
30 Bibliography:
Berkman, H. and Nguyen, N.H. 2010. Domestic liquidity and cross-listing in the United States, Journal of Banking & Finance. Volume 34, Issue 6, Pages 1139-1151
Blume, M. and Keim, B. 2008, Institutional Investors and Stock Market Liquidity: Trends and Relationships, National Bureau of Economic Research, Vol 73, pp. 13-55.
Brockman, P. Chung, P. 1999, Cross-listing and firm liquidity on the stock exchange of Hong Kong, Managerial Finance, Vol. 25 Issue 1, pp.64 – 88
Brunnermeier, M. Pedersen, L. Market Liquidity and Funding Liquidity, Review of Financial Studies, Volume 22, Issue 6, pp. 2201-2238
Chandar, N. Patro D. Yezegel A. 2007, Crises, Contagion and Cross-Listings, Journal of Banking & Finance, Volume 33, Issue 9.
Doidge, C. 2004. U.S. cross-listings and the private benefits of control: evidence from dual-class firms. Journal of Financial Economics. Volume 72, Issue 3, Pages 519–553
Doidge, C., Karolyi, G.A., Lins, K.V., Miller, D.P. and Stulz, R. 2009. Private benefits of control, ownership, and the cross‐listing decision. Journal of Finance. Volume 64, Issue 1, Pages 425-466
Errunza V. Miller D. 2000, Market Segmentation and the Cost of the Capital in International Equity Markets, Journal of Financial and Quantitative Analysis, Volume 35, Issue 04
Fang, V., Noe, T. and Tice, S. 2009. Stock market liquidity and firm value, Journal of Financial Economics, Volume 94, Pages 150 – 169
Foerster, S.R. and Karolyi, G.A. 1999. The Effects of Market Segmentation and Investor
Recognition on Asset Prices: Evidence from Foreign Stocks Listing in the United States. Journal of Finance. Volume 54, Issue 3, Pages 981-1013
Fox, J. 1997, Applied Regression Analysis, Linear Models, and Related Methods. California, Sage Publications. Vol. 15 p. 306.
31
Frankfurt Stock Exchange Segments. 2010.
http://www.frankfurtstockexchange.de/the4frankfurtsegments.html
Frino, A. Di Marco, E. Lepone, A., 2009, The Impact of ADR Listing on Liquidity, Market Insights, Volume 28.
Goyenko, R. Holden, C. and Trzcinka, C. 2009, Do Liquidity Measures Measures Liquidity? Journal of Financial Economics Vol. 92, pp. 153–181.
Gozzi, J. Levine, R. Schmukler, S. 2008, Internationalization and the Evolution of Corporate Valuation, Journal of Financial Economics 88, pp. 607–632
Karolyi, G.A., 1998, Why Do Companies List Their Shares Abroad? A Survey of the Evidence and Its Mangerial Implications, Blackwell Publishers, Boston.
Kim, O. Verrecchia, R. 1994, Market Liquidity and Volume Around Earnings Announcements, Journal of Accounting and Economics, Volume 17, Pages 41-67.
Lang, M.H., Lins, K.V. and Miller, D.P. 2003. ADRs, Analysts, and Accuracy: Does Cross Listing in the U.S. Improve a Firm's Information Environment and Increase Market Value? Journal of Accounting Research. Volume 41, Issue 2, Pages 317–345
Lang, M. Maffet, M. 2011, Transparency and liquidity uncertainty in crisis periods, Volume 52, Issues 2–3, pp. 101–125
Lins K. Servaes H. and Tufano P. 2010, What Drives Corporate Liquidity? An international survey of cash holdings and lines of credit. Journal of Financial Economics, Volume 98, Issue 1, Pages 160-176.
Lins, K.V., Strickland, D. and Zenner, M. 2005. Do non-U.S. firms issue equity on U.S. stock exchanges to relax capital constraints? Journal of Financial Quantative Analysis, Volume 40, Pages 109-133
Mankiw, N. 1990, A Quick Refresher Course in Macroeconomics, Journal of Economic Literature, Vol. 28, 1645–1660.
