Impact of foreign listings for European
countries after implantation of the SOX.
University of Groningen
Niels Bijkerk University of Groningen
s1855921 Faculty of Economics and Business
Keppelseweg 432 MSc. Finance
ABSTRACT
This paper examines changes in risk characteristics, following a cross-listing and also if it is consistent with a higher degree of market integration. In addition, I analyze if cross-listings lead to a lower cost of capital explained by higher levels of integration. Investigating the risk changes with a single- and two-index model between 2002 and 2013, this paper found no evidence of significant changes in risk regardless of the risk characteristics analyzed. The results imply no diversification advantages due to higher levels of integration in the post-listing period. As a result, there is no reduction of the cost of capital.
Key words: International cross-listings, cost of equity, risk, market segmentation
1 Introduction
Extensive literature examines the effects and the possible advantages of international listings. An international listing can be seen as a listing on a nondomestic exchange, this is referred to as various terms in the literature: cross-listing, foreign listing, and overseas listing. In this paper I will use these terms interchangeably. Sarkissian and Schill (2016) highlighted three reasons why firms choose to use an international listing. The first is to generate more economic synergies with the country of listing. Second, a firm uses a cross-listing in order to increase the price efficiency of its securities. Third, to choose for countries with better investor protection, an international listing is used to direct management to investor-friendly behavior. The second objective will be highlighted in this paper. This paper examines in further detail the perceptions and advantages of cross-listings over a longer period after implementation of new governmental rules.
The paper by Sarkissian and Schill (2016) confirm that decisions for cross-listings are mainly driven by economic synergies between the domestic and host markets, under the most attractive conditions of timing and location instead of pricing efficiency and institutional differences between markets, as highlighted in the literature. This leads to the following questions: what are the main reasons for firms to list on an overseas market and are there long-term valuation gains for cross-listed firms? In addition, Sarkissian and Schill (2016) investigate the long-term valuation benefits (efficient timing) and short-term valuation benefits (market timing). The results suggest that cross-listings are likely to occur in periods of time in which the gains of cross-listings are particularly strong. Sarkissian and Schill (2016) only found evidence for short-term valuation gains as no long-term valuation gains were established. Other papers support short-term valuation gains; Sarkissian and Schill (2009) as well as Foerster and Karolyi (1999) also found evidence for short-term valuation gains. These valuations gains are measured within a year of listing. This paper aims to extend the already extensive literature by focusing on the long-term advantages of an overseas listing. In order to do so, a sample is compiled with 1,397 cross-listed firms with their primary listing in the Netherlands, Germany, or the United Kingdom. The foreign listings are distributed over 14 countries between 2002 and 2013.
Governmental changes (which will be explained in further detail later in this section) could lead to changes in the integration of markets. A large body of literature has examined the segmentation of international capital markets. For example, Howe and Madura (1990) examined the impact of international listings on common-stock risk. Although, a large amount of research focused on this topic, it remains controversial. For instance, Foerster and Karolyi (1999) indicated that segmentation could be reduced by cross-listings, which is in contrast to Howe and Madura (1990), who found no significant shift in risk. Errunza and Miller (2000) conducted research in order to test the theoretical models, which predict a reduction in the cost of capital from depositary receipt offerings.
Furthermore, segmentation and integration are important topics for corporate finance. As stated by Howe and Madura (1990, p. 1134), integration is an important determinant in the pricing of international assets, as the pricing is influenced by the capital flow across national boundaries and the degree of independence of different markets. Markets could be segmented between countries; corporate managers could engage investment or financing activities to overcome the effects of segmentation. In addition, Foerster and Karolyi (1999) illustrate with an event study that cross-listings reduce market segmentation. They document the effect on share value and the global risk exposure of these firms, and report a significant decline of the stock’s domestic market beta relative to the home market index. The results are consistent with the market segmentation hypothesis, which indicate lower expected returns for cross-listed companies. Despite the benefits of foreign listings to overcome mildly segmented markets, existing empirical studies indicate little or no evidence for a long-term net gain from a cross-listing. Gozzi et al. (2004) pointed out that the Tobin’s q does not rise after internationalization. The Tobin’s q will rise before the international listing. However, after the listing the Tobin’s q will fall significantly.
Turning to recent changes, the regulation signed by the U.S. government could change the perception of foreign listings to enable market segmentation. In 2002 the U.S. government signed the Sarbanes-Oxley (SOX) Act. Karoyli (2006, p. 143) mentions the following in relation the Act: “the Act contains a number of provisions to improve the corporate governance practices of publicly listed companies in the U.S. As a result, the SOX result in a significant increase in costs, redundancies with home-market requirement and difficulties of terminating reporting obligations”.
significant fraction of the market between 1990 and 1999. In 10 years the number of foreign equities doubled. However, the number of foreign equities retreated from 2002 to 2004 for the U.S. market to the level of 1990. The results globally indicate the same results. After a peak in 1995 for cross-listed firms, the number of such firms declined by 30 percent. This decline indicates that there is a shift in the perception of possible advantages for cross-listed firms. The SOX Act not only had an impact on the number of foreign listings in the U.S., it also led to a shift in the distribution of foreign listed firms globally. In 1995, the United States attracted almost 60 percent of global initial public offerings (IPOs) and Germany (as second best) attracted only 17 percent of global IPOs (Cetorelli and Peristiani, 2015, p. 159). Whereas in 2005 after the implementation of the SOX Act, Germany had 33 percent of IPOs, the United States secured 31 percent, followed by the United Kingdom closely thereafter. Overall, the SOX Act led to major changes in the perception of potential benefits associated with international listings in the U.S. As a result, the distribution of international listed companies has changed since cross-listings are mainly determinant by economic synergies between countries. The new regulations could lead to new barriers, which could affect pricing effects. The impact from the SOX Act does not only influence non-U.S. Firms that are cross-listed in the U.S. The impact of the new regulations has to be taken in account by European cross-listed firms as well due to spillover effects to Europe. The indirect effects from the SOX Act on the European firms could lead to new investment strategies for European firms. The indirect impact of the SOX Act for European firms raises an interesting question whether European firms can reduce the cost of capital due to higher levels of integration after a foreign listing. The purpose of this paper is to investigate the indirect effects on international listings for European firms after the SOX Act. In particular, this paper examines if the risk characteristics of an international listing changes as a result of the listing in the long run. The change should be consistent with a greater degree of market integration. As mentioned by Gagnon and Karolyi (2010), a foreign listing is a corporate financial strategy to reduce the negative effects of market segmentation. However, will the corporate financial strategy reduce the negative effects after the new regulations, which leads to a change in the distribution of international listings? Are there long-term advantages after the new regulations?
changes in a similar manner with the level of integration due to the international listing. And triggers changes in the beta to a decrease in the cost of capital? One other feature, which is new, is the focus on European countries instead of focusing mainly on the U.S. market.
The remainder of the paper proceeds as follows: Section 2 presents a literature review; Section 3 presents the methodology, hypothesis, and the data; Section 4 presents the results; and Section 5 sets forth a conclusion.
2 Literature review
Several papers have examined the advantages of cross-listed firms compared with pure domestic listed firms. Foerster and Karolyi (1999), Baker et al. (2002), Chambers et al. (2015), and Domowitz et al. (2001) all point out advantages of foreign listings. Karolyi (1998, 2006) uses these studies to highlight the general advantages of cross-listings. Therefore, these two studies will be described in detail.
Karolyi (1998) uses a survey of the existing literature to create a better understanding of the decision by firms to be cross-listed. The survey confirms the following results: share prices react favorably to cross-listings in the first month after listing; the post-listing price performance is highly variable the year after the listing; the post-listing trading volume increases as well as the home-market trading; the liquidity of the trading shares improves; the domestic market risk is significantly decreased whereas the global risk is only associated with a small increase in risk and strict disclosure requirements are the most important impediment for cross-listings. A lower level domestic market risk and a small increase in global market risk suggest a cost reduction of equity. Together with the increase of the share price, these are the most relevant results of this survey study from Karolyi (1998). The second study (Karolyi, 2006) uses a survey study to determine how important cross-listings are. The survey highlights multiple reasons why firms use cross-listings, the main reasons are to control and improve the corporate government system; also, foreign listed shares have a broader range of investors and a better information supply, and cross-listings can lead to lower the cost of capital.
a change in the equilibrium prices of the cross-listed securities (Howe and Madura, 1990). This leads to higher prices and lower returns for the listings.
Howe and Madura (1990) conducted a study focusing on the permanent shifts in risk caused by international listings. With a single- and two-index model, changes in the domestic beta are calculated. The data is constructed for U.S. foreign listed firms in Germany, France, Japan, and Switzerland between 1969 and 1984. In order to calculate changes in the domestic beta, 16 quarters prior to the listing (pre-listing) and 16 quarters after the listing (post-listing) are used. The international listings do not appear to cause significant changes in risk. Howe and Madura (1990) explain the insignificant results by noting the already reasonable integrated markets or observing that listings are ineffective mechanisms for reducing segmentation.
In contrast to the study by Howe and Madura (1990), Foerster and Karolyi (1999) found evidence for the segmentation hypothesis. Foerster and Karolyi (1999) investigated the stock price performance and changes in risk exposure associated with the cross-listings in the U.S. American Depositary Receipts (ADRs) between 1976 and 1992 are analyzed from the NYSE and AMEX. The final sample consists of 153 listings from 11 countries. The pre-listing consists of 52 weeks before the listing and the post-listing consist of 52 weeks after the listing. The main results imply a significant average excess return of 19 percent in the year following the listing. Second, the stocks’ market beta compared with its home market index declined from 1.03 to 0.74 on average. These results support the market segmentation hypothesis; this implies dramatic patterns in share values around international listings due to direct investment barriers.
A more recent study by Hail and Leuz (2009) examines whether dual-listed firms in the U.S. lead to a reduction in the cost of capital. This study is interesting since it also focuses on the period after implementation of the SOX Act in the U.S. The sample is constructed for a large panel of ADR firms from 45 countries. The main conclusion of the paper is that the SOX Act has not diminished the benefits of a foreign listing in the U.S. The cost of capital for cross-listed firms decreases between 70 and 120 basis points.
2005. The decline in the number of firms cross-listed on the NYSE can potentially be explained by changes in firm characteristics rather than changes in the benefits of foreign listings. There is still a significant premium (valuation difference between listed and non-listed firms) for U.S. exchange listings every year. In contrast, there is no premium for listings on the LSE. Another result, which is important for this paper, is that after controlling for firm characteristics, the decline in cross-listings on the NYSE is not caused by the SOX Act. In addition, with a sample period of 57 years, the paper of Sarkissian and Schill (2016) traces links between time-series characteristics of aggregate cross-listing decisions and a home or host country’s financial and economic development. The paper by Sarkissian and Schill (2016) examines market characteristics, whereas Doidge et al. (2009) investigate the influences of firm characteristics. Sarkissian and Schill (2016) use a sample with 3,589 exclusively foreign-exchange-traded listings distributed over 33 foreign markets. First, the results confirm that a substantial portion of host-market waves is due to underlying waves in proximate home markets. Second, they observe that foreign listing waves in host markets are not only driven by their home country’s economic and financial performance, but also by proximate home country markets. Third, the foreign listing waves are associated with systematic short-term value gains for cross-listed firms, however no long-term systematic valuation gains are observed. As a result, they conclude that cross-listing decisions most often attempt to profit from host-market outperformance.
3 Methodology and data
3.1 Methodology
In order to investigate how risk characteristics change as a result of a foreign listing, various risk characteristics and measures will be used and compared between the pre- and post-listing periods. In addition, the impact of different currencies will be analyzed and a robustness test will be applied to account for the financial crisis. However, the segmentation hypothesis must first be explained in more detail. This hypothesis implies that local investors demand a higher ‘super risk premium’ due to the inability to diversify risk as a result of segmented markets. Investment barriers (e.g. regulatory barriers, information constraints and taxes) lead to segmented markets, as Foerster and Karolyi (1999) mention. Segmented markets have two characteristics, defined by Howe and Madura (1990, p. 1136), which explain why local investors demand a higher risk premium: capital does not flow into or out of a perfectly segmented market, and an asset in a segmented market is priced in the context of that market. As a result, segmented markets and the higher risk premium are an incentive for firms to adopt financial strategies to overcome the negative effects of market segmentation. As different studies have confirmed (see e.g., Foerster and Karolyi, 1999; Errunza and Miller, 2000; Hail and Leuz, 2009), cross-listings are an effective financial strategy to overcome segmentation. The theory behind this strategy is that cross-listed firms can diversify their home market risk due to better-integrated markets. This leads to higher stock market prices and lower returns; consequently, the risk premium for segmented markets diminishes.
To begin, the first two risk measures are estimated with a single- and two-index model, respectively. The first risk measure is the domestic beta, which is estimated with the single-index model and the return of the domestic market single-index. The second risk measure is the foreign beta, which is estimated with the two-index model and the return of the foreign market index of the country of listing. The single- and two-index models are
!! = !! + !!!!,!"# + !! (1)
single-index model; !! is the domestic beta for the two-index model; and !! is the beta estimated using the market index return on the foreign market.
It is necessary to explain the reasoning for testing the segmentation hypothesis with the single- and two-index models. If markets are segmented and cross-listings are an effective mechanism to overcome segmentation, the firms’ stock sensitivity for the domestic market must decrease after a foreign listing. Due to higher levels of integration as a result of the cross-listing, the investment barriers diminish and firms are able to diversify the domestic market risk. Accordingly, the ability to diversify increases and therefore the firms’ stock sensitivity for the domestic market decreases; consequently a lower domestic beta in the model will indicate this. On the other hand, the firms’ stock sensitivity for the foreign market increases after a cross-listing, which leads to an increase in the foreign beta. As a consequence, the decrease in the domestic beta and the increase in the foreign beta indicate diversification effects due to higher levels of integration after a foreign listing. Because of the decrease in firms’ stock sensitivity for impacts of the domestic market, local investors demand a lower risk premium. As a result, the stock price of cross-listed firms will increase and the returns decrease. In order to test whether the market segmentation hypothesis holds, two hypotheses are proposed: first, the null hypothesis for the domestic beta is no change in the domestic beta after a foreign listing for the single- and two-index models, and the alternative hypothesis is a decrease in the domestic beta. Second, the null hypothesis for the foreign beta is no change, and the alternative hypothesis is an increase in the foreign beta after a cross-listing.
and eventually lead to a decrease in the standard deviation. On the other hand, when uncertainty increases after a foreign listing – as a result of language barriers, cultural differences or regulatory uncertainty (Howe and Kelm (1987) explain that regulatory overseas can change; see the SOX Act in the U.S. as an example) – this leads to an increase in the standard deviation. The null hypothesis is no change in the standard deviation and the alternative hypothesis is a decrease in the standard deviation after a cross-listing.
The next measure is the R-squared from the single- and two-index models. The R-squared indicates how much of the estimated variability is explained by the two models. For the single-index model the R-squared should decrease after a foreign listing. A foreign listing will lead to better integration and the domestic market risk can be diversified. As a result, less of the variability of the firm is attributable to the domestic market. On the contrary, the diversification effects can only be explained when the variability of the firm estimated with the two-index model is more attributable to the foreign market. In other words, the markets are more integrated. This will lead to diversification advantages. However, as mentioned in the discussion of the standard deviation as risk measure, when for both models the R-squared increases after a cross-listing, implying the variability of the firm is more attributable to the domestic market as well as to the foreign market, the uncertainty increases. The higher uncertainty will lead to higher risk premiums demanded by investors to compensate for the uncertainty. The null hypothesis for both models is no change in the R-squared. The alternative hypothesis for the single-index model is a decrease in the R-squared, whereas for the two-index model an increase in the R-squared after a foreign listing is expected.
Moreover, I investigate the impact of the difference in currencies on the domestic and foreign betas. Dutch and German foreign listings are obtained in Euros and British foreign listings are obtained in British Pounds. In order to estimate the betas, the different currencies are converted via exchange rates to compare them. The Euro is expected to be more stable and therefore lead to less pronounced effects from the foreign listed firms. This is since the Euro is a more stable currency due to the monetary union, while the British Pound is used by only one country. Formally, the null hypothesis is that there is no difference in the change of beta between the two currencies and the alternative hypothesis is a difference in the change of beta between the currencies.
constructed to create a robustness test and account for the financial crisis. Because the financial crisis started in 2008 it is likely that the higher levels of uncertainty would influence the domestic and foreign betas. This would lead to a higher cost of capital. In order to control for this potential bias, the sub-samples before and after the crisis are created and compared.
3.2 Data.
A large body of literature focuses on foreign listings in the U.S. However, most foreign listings are nowadays on the Frankfurt and London Stock Exchange. Therefore, for this paper data is collected for IPOs in Germany, the Netherlands, and the United Kingdom. The U.S. government signed the SOX Act in 2002; therefore, the sample period is selected from the beginning of 2002 until 2014. In order to explore whether the risk characteristics of the listing firms’ shares change due to the foreign listing, a two-year pre-listing period and a two-year post-listing period is selected. The timespan for the pre- and post-listing should be sufficiently long so as to obtain trustworthy and reliable beta estimates. However, as mentioned by Howe and Madura (1990), betas can be intertemporally unstable as a result of long pre- and post-listing periods, and as a result, could level out the changes in the betas. This leads to no observable change in the beta. Consequently, a two-year pre- and post-listing period is chosen as a trade-off. The sample is constructed until the end of 2013 since a two-year post-listing is needed.
The weekly stock index is obtained from DataStream by using the firm identifiers from the WorldScope database. Weekly stock returns are based on a Total Return Index. A Total Return Index is chosen instead of a Price Index, as this index indicates a theoretical growth in the value of a firm, assuming that dividends are reinvested to purchase additional units at the closing price applicable on the ex-dividends date (Nandha and Faff, 2008, p. 990). For this study, weekly returns are obtained since these are less affected by biases. Brown et al. (1985) investigates the advantages of weekly returns rather than daily or monthly returns. Brown et al. (1985) provide evidence that weekly returns are less sensitive to bid-ask effects, non-synchronous trading days, and day-of-the-week effects.
Firms with missing data over the full sample period between 2002 and 2013 are excluded from the sample. After the exclusion on the basis of missing returns the sample includes 96 Dutch listed firms, 610 German listed firms, and 691 British listed firms. Table 1 presents the sample of international and domestic listings by country.
Table 1
This table shows the sample of international and domestic listings by Dutch, German, and British companies.
Panel A. Number of selected firms by country
Domestic listed Cross-listed Total
Netherlands 28 68 96
Germany 463 147 610
United Kingdom 294 397 691
Total 785 612 1397
first sub-period between 2002 and 2007 contains 278 foreign listings. The second sub-period between 2008 and 2013 contains 311 foreign listed firms in total.
Panel A of Table 2 illustrates the distribution of listings by country; cross-listings are distributed over 14 countries. Germany and the United States are the most dominant in attracting secondary listings compared with the other countries. The dominance of the U.S. and Germany are in line with the mentioned change in distribution of cross-listings by Cetorelli and Peristiani (2015) and Karolyi (2006). In Panel B of Table 2, the cross-listings are distributed according to the year. Two hundred and thirteen international listings occurred in 2013, this is significantly more compared with the other years. The number of listings per year ranges between 10 and 60 on average. One explanation for the high number of international listings in 2013 could be the regained strength and trust in the economy after the 2008 financial crisis. As mentioned by Sarkissian and Schill (2016), cross-listing decisions are mainly driven to time outperformance of foreign markets. This could explain the high number of cross-listings for the U.S. and German markets in 2013 due to better economic prospects. The U.S. and German economies recovered much sooner from the financial crisis compared with other European countries. This could explain why firms choose to cross-list in the U.S. and Germany. This implies that despite the new regulations (the SOX Act), firms still see the U.S. as an opportunity to cross-list.
Table 2
This table shows the sample of foreign listings aggregated by year and country
Panel A. Distribution of international listings by country
Dutch Firms German Firms British Firms Total
Country City Austria Vienna 1 1 Canada Toronto 2 2 France Paris Germany Berlin 8 65 73 Frankfurt 18 105 123 Munich 1 1 Stuttgart 4 8 12 Hong Kong 1 1 Ireland Dublin 1 1 Italy Milan 4 12 16 Luxembourg Luxembourg 1 1 Netherlands Amsterdam 1 1 Spain Madrid 1 1
South Africa Johannesburg 2 2
Switzerland Bern
Zurich 6 6
United Kingdom London 15 1 16
United States New York 33 83 216 332
Total 84 103 402 589
Panel B. Distribution of international listings by year
Year Netherlands Germany United Kingdom Total
4 Empirical results
4.1 Domestic beta
The change in the domestic beta is estimated using two models. The estimated domestic betas from the single-index model (equation 1) are presented in Table 3. The results are presented for the entire sample as well as aggregated by country. The model uses the market return as an explanatory variable with a value weighted domestic index. The first column presents the number of observations. The second column presents the betas estimated for the pre-listing period, the third column presents the betas estimated for the post-listing period, and the fourth column presents the t-statistics for the difference in mean of the betas in order to investigate if there are significant changes in the estimated betas. The domestic beta for the entire, Dutch, and German samples inclines towards the foreign listing. However, none of the changes in the domestic beta are statistically significant at a 10 percent significance level. For the British sample, the domestic beta declines after the foreign listings. The change in domestic beta is statistically significant at a 5 percent significance level.
Turning to previous literature, the insignificant changes in domestic beta for the entire, Dutch, and German samples are in line with Howe and Madura (1990). The inclines in the domestic beta are supported by the results noted by Sarkissian and Schill (2016) of no long-term valuation gains. The insignificant changes in the domestic beta could be due to reasonably well-integrated markets. A foreign listing in a sufficiently integrated market would have no significant effect on the domestic market. One other explanation could be that a foreign listing is not an effective mechanism to decrease segmentation.
indicates higher levels of risk after a cross-listing. Higher levels of risk lead to higher returns, which eventually increase the cost of capital. However, this must be interpreted with caution; hence, the t-statistic is not statistically significant and measured differences are small.
Table 3
Comparison of pre- and post listing betas with !! = !!+ !!!!,!"#+ !!
Number of observations Pre-listing beta (!!) Post-listing beta (!!) t-statistic (difference) Entire sample 61,256 0.71 0.72 1.90 United Kingdom 41,808 0.77 0.75 4.67 Germany 10,712 0.71 0.72 0.98 Netherlands 8,736 0.63 0.68 1.32
Furthermore, the domestic betas are estimated with the two-index model (equation 2) in order to create more robust results. In this case, the explanatory variables are returns on the market index from the domestic country and the foreign return on the market index from the country of listing. The domestic beta (!!) is presented in Panel A of Table 4. The domestic beta estimated with the two-index model for the entire and Dutch samples increases after the foreign listing. Nevertheless, the results are not statistically significant. The results support the insignificant domestic beta changes estimated using the single-index model. In contrast, the domestic beta estimated does decline for the British and German sample; the changes are statistically significant. Although, there has to be noticed the decreases are relatively small and indicate for only limited diversification advantages. As mentioned above, the significant changes are in line with papers by Foerster and Karolyi (1999) and Hail and Leuz (2009). As a consequence, the estimated betas of the two-index model indicate that foreign listings could be an effective mechanism to overcome segmentation. However, a fully-integrated market could explain the insignificant changes in the domestic beta.
well-established large companies. This suggests that they have other mechanisms to overcome market segmentation and this could explain the insignificant results. Overall, there is evidence of market segmentation effects and these effects can be reduced by cross-listings. However, firms may also use other mechanisms to overcome market segmentation. Evaluating the insignificant and inclining results in the domestic beta reveals that in the long run no significant advantages due to a decrease in the domestic beta are gained.
4.2 Foreign beta
In addition to the regression model with just one control variable, the model is extended to include the overall return on the overseas market to estimate the foreign betas. The foreign betas are presented in Panel B of Table 4. The alternative hypothesis tested for the foreign beta is an increase in the foreign beta due to an international listing. The foreign beta is estimated with the return on the market index from the country of listing. As a result of the foreign listing, a small decline in the foreign beta is estimated for the entire sample. Although the declining foreign beta is in contrast with the expectation, the change is not significant and should therefore be interpreted with caution.
In comparison, the foreign beta for each country increases after a foreign listing. The increase in beta after the cross-listing is not statistically significant at a 10 percent significance level. The insignificant changes in the foreign betas imply reasonably well-integrated markets. In particular, the foreign betas for the entire and British samples are relatively small, which is supported by the results from Howe and Madura (1990). With a t-test the betas are tested and they are not significantly different from zero. Overall, the results indicate an insignificant increase in the foreign beta after an international listing. The results support the insignificant changes of the domestic beta.
not statistically significant and therefore this cannot be confirmed. The results are in line with the results of no long-term valuation gains established by Sarkissian and Schill (2016). Evaluating the changes in the domestic and foreign beta due to cross-listings, the cost of capital would not decrease on basis of the market segmentation hypothesis. Although there are some significant changes after a cross-listing the differences are relatively small and indicate for limited diversification advantages. The results imply that cross-listings would not lead to higher levels of market integration. This can be explained by sufficiently integrated markets or as cross-listed firms are large and well-established with effective mechanisms to overcome market segmentation.
Table 4
Comparison of pre- and post listing betas with !! = !!+ !!!!,!"#+ !!!!,!"# + !!
Panel A. Domestic beta
Number of observations Pre-listing beta (!!) Post-listing beta (!!) t-statistic (difference) Entire sample 61,256 0.66 0.69 1.73 United Kingdom 41,808 0.68 0.63 2.80 Germany ! 10,712 0.79 0.78 3.41 Netherlands ! 8,736 0.38 0.44 1.54
Panel B. Foreign beta
Number of observations Pre-listing beta (!!) Post-listing beta (!!) t-statistic (difference) Entire sample United Kingdom! 61,256 0.10 0.08 1.41 41,808 0.10 0.13 1.55 Germany ! 10,712 -0.34 -0.23 -1.52 Netherlands ! 8,736 0.39 0.43 0.94 4.3 Standard deviation
and future regulatory changes. The standard deviation could be interpreted as the total risk of the firms’ stock.
The results are presented in Table 5. There is a small decrease in the standard deviation (6.26 per cent for the pre-listing period compared with 6.19 percent for the post-listing period); however, this small decrease is not statistically significant at a 10 percent significance level. The results are in line with the insignificant results noted by Howe and Madura (1990). The results are contrary to studies by Foerster and Karolyi (1999) and Sarkissian and Schill (2009), both of which reveal significant shifts in total and systematic risk within a year of cross-listing. The contrasting results could be explained by two reasons. First, both papers use a timeframe within a year of listing, compared to the timeframe used in this present study of two years pre-listing and two years post-listing. Second, as mentioned by Sarkissian and Schill (2009), cross-listings in the U.S. have much stronger effects compared to other countries. Both papers use listings in the U.S., while this study also uses foreign listings from other countries. The insignificant decrease in the standard deviation supports the results from the single- and two-index models. Cross-listing effects cannot only be explained on the basis of market integration. The results not imply for better information flows and stock prices containing less noise as a result of better integration after a foreign listing.
4.4 R-squared
In sum, the change in the R-squared for the two-index model supports previous results that firms use other mechanisms to overcome market segmentation, as explained by Stapleton and Subrahmanyam (1977). Contrary, the change in the R-squared for the single-index model implies for more uncertainty due to a foreign listing. These results are in line with the previous results, which indicate for an increase in the firms’ stock sensitivity of the domestic market after a cross-listing. The increase of the R-squared for the single-index model can be explained due to language barriers, culture differences and future regulation changes in the country of listing.
Table 5
Additional pre- and post-listings comparisons Pre-listing Post-listing t-statistic (difference) Standard deviation 0.63 0.62 0.48 R2 single-index model 0.11 0.57 8.54 R2 two-index model 0.11 0.16 6.89 4.5 Currencies
Furthermore, the next test examines whether currencies have an impact on the estimated domestic and foreign beta. The domestic and foreign betas from the two models (equation 1 and 2) are presented in Table 6. For the single-index model the domestic beta (see Panel A from Table 6) for the Euro is in line with the domestic beta from the German and Dutch samples presented in Table 3. The domestic beta indicates an increase due to foreign listings. However, the change is statistically significant at a 10 percent significance level. The results support earlier findings of a reduction in the cost of capital due to diversification possibilities, which lead to a lower domestic beta.
differences are measured. Panel C of Table 6 presents the results of the changes in the foreign betas due to cross-listing. The results for the British sample do not change compared with the previous tests. The foreign beta increases after foreign listings, which indicates further integration within the markets. On the contrary, the Dutch and German samples reveal a decrease of the foreign beta and this change is not statistically significant.
The t-test, which compares the mean of the samples with different currencies, indicates no significant difference between the samples. Overall, there is no evidence of a different impact of foreign listings due to different currencies. The results support previous tests and are in line with papers by Sarkissian and Schill (2009) and Doidge et al. (2009). Both papers confirm that foreign listings in the U.S. have much stronger effects on the valuation gains compared with listings in other countries. Since this paper analyzes listings both in the U.S. and other countries, this could explain the insignificant changes in domestic and foreign betas.
Table 6
Comparison between currency for Euro and Pound.
Panel A. Domestic beta with single-index model
Comparison of pre- and post listing betas with !! = !!+ !!!!,!"#+ !! Number of observations Pre-listing beta (!!) Post-listing beta (!!) t-statistic (difference) Euro 50,544 0.68 0.72 3,94 Pound 41,808 0.77 0.75 4.67
Panel B. Domestic beta with two-index model
Comparison of pre- and post listing betas with !! = !!+ !!!!,!"#+ !!!!,!"# + !!
Number of observations Pre-listing beta (!!) Post-listing beta (!!) t-statistic (difference) Euro ! 50,544 0.66 0.70 ! 1.87 Pound ! 41,808 0.68 0.63 ! 2.80 ! !
Panel C. Foreign beta with two-index model
Comparison of pre- and post listing betas with !! = !!+ !!!!,!"#+ !!
4.6 Robustness test
In order to determine whether the financial crisis of 2008 had an impact on firms’ stock sensitivity for the domestic and foreign markets, the sample is split in two. The first sub-sample is constructed for the period before the crisis; it contains foreign listings between 2002 and 2007. The second sub-sample includes the period of the financial crisis; it contains foreign listings between 2008 and 2013. The results for the single- and two-index models are presented in Table 7.
Panel A of Table 7 shows the domestic beta estimated with the single-index model. In the period before the financial crisis, the domestic betas for the entire sample, British sample and Dutch sample increase after a foreign listing, whereas the domestic beta for the German sample decreases. For the second period, the domestic betas increase for the British, German and Dutch sample; the domestic beta for the entire sample remains the same. The difference in mean for the Dutch sample measured over the first sample period and the difference in mean for the British sample over the second sample period are statistically significant at a 5 per cent significance level. The other measured differences in mean are not statistically significant. The results are in line with the changes in domestic beta for the full sample period, presented in Table 3. When comparing the two sub-periods with a t-test for difference in mean, no significant differences are measured at a 10 per cent significance level. Overall, the sub-periods imply no different results compared with previous tests for the domestic beta estimated with the single-index model. Two samples are statistically significant; however, the changes in the domestic beta are relatively small and imply not very large changes due to higher levels of integration.
However, the other samples show no significant change and therefore overall the results imply there is no difference between the two sample periods.
Panel C of Table 7 shows the foreign betas, indicating firms’ stock sensitivity for the foreign market of listing. Over the first period, no significant evidence is found of higher sensitivity of the firms’ stock for the foreign market after a cross-listing. The second period indicates higher levels of sensitivity for the foreign market. The foreign beta increases for the British, German and Dutch samples, and the increase is statistically significant at a 10 percent significance level. Comparing the two sub-periods, the difference in mean is only significant for the entire sample at a 10 per cent significance level.
Table 7
Comparison between two sub-periods before and after the financial crisis
Panel A. Domestic beta with single-index model
Comparison of pre- and post listing betas with !! = !!+ !!!!,!"#+ !!
Number of observations Pre-listing beta (!!) Post-listing beta (!!) t-statistic (difference) Sub-sample (2002-2007) Entire sample 28,912 0.69 0.72 1.93 United Kingdom 20,176 0.73 0.78 2.12 Germany 2,808 0.72 0.66 1.86 Netherlands 5,928 0.56 0.62 2.41 Sub-sample (2008-2013) Entire sample 32,344 0.72 0.72 0.75 United Kingdom 21,632 0.47 0.70 4.57 Germany 7,904 0.70 0.75 1.23 Netherlands 2,808 0.57 0.58 0.87
Panel B. Domestic beta with two-index model
Comparison of pre- and post listing betas with !! = !!+ !!!!,!"#+ !!!!,!"# + !!
Number of observations Pre-listing beta (!!) Post-listing beta (!!) t-statistic (difference) Sub-sample (2002-2007) Entire sample 28,912 0.57 0.61 1.45 United Kingdom 20,176 0.65 0.70 1.79 Germany 2,808 0.78 0.71 1.36 Netherlands 5,928 0.35 0.41 1.96 Sub-sample (2008-2013) Entire sample 32,344 0.72 0.73 0.51 United Kingdom 21,632 0.71 0.56 2.68 Germany 7,904 0.80 0.80 0.79 Netherlands 2,808 0.39 0.37 1.12
Panel C. Foreign beta with two-index model
5 Summary and conclusion
In this study, I investigate the effects on international listings after the implementation of SOX Act in year 2002. Specifically, I analyze changes in risk characteristics following a cross-listing and whether it is consistent with a higher degree of market integration. In addition, I analyze if cross-listings lead to a lower cost of capital explained by higher levels of integration. Overall, the results indicate no significant risk changes regardless of the risk characteristics analyzed. The first test examines the domestic beta in the pre- and post-listing period. The results imply that there are no diversification advantages due to a decrease in firms’ stock sensitivity for the home market. Moreover, the results for the foreign beta indicate no statistically significant changes between the pre- and post-listing period. The firms’ stock sensitivity for the foreign market increases, however these changes are not statistically significant. This implies firms’ stock sensitivity for shocks in the foreign markets does not increase. Together with the insignificant inclines in the domestic beta this suggests no diversification advantages caused by higher levels of integration due to cross-listings. As a result, there is no reduction in the cost of capital.
In conclusion, there is no support for the market segmentation hypothesis. The results suggest that there are no gained advantages due to further integration. The increase in the domestic beta indicates that a foreign listing will not lead to lower sensitivity to the home market whereas sensitivity to the foreign market of listing should increase. As a result, there are no diversification opportunities; a cross-listing will not lead to a lower cost of capital. The results are supported by studies from Howe and Madurra (1990), Sarkissian and Schill (2009), and Sarkissian and Schill (2016). This study contributes to recent studies, which question the extent to which market integration can explain cross-listing effects (see e.g., Doidge et al., 2004; Karolyi, 2006; Hail and Leuz, 2009). As questioned by the papers, is the bonding hypothesis a more promising explanation for the cross-listing effects? The results of this study indicate that foreign listings can be an effective mechanism to overcome the market segmentation, as also stated by Foerster and Karolyi (1999). However, cross-listing can only be an effective mechanism to overcome market segmentation if a market is not already sufficiently integrated, as pointed out by Doidge et al. (2009). It could also be that firms use other mechanisms to overcome market segmentation. To conclude, no evidence is found of cross-listings effects, which can be explained by the market segmentation hypothesis. This study indicates that cross-listings will not lead to higher levels of integration, which lead to a lower cost of capital.
Despite the fact that for different risk characteristics measured similar results are estimated, which are supported by previous literature, some limitations must be taken in account. The single- and two-index models are based on the CAPM model. However, the CAPM model is premised on the estimated market portfolio – no taxes, no information costs, no transformation costs, and integrated markets. This premise does not hold and may be open to biases. Despite this limitation, the results can be seen as trustworthy and robust.
References
Alexander, G.J., Eun, C.S. and Janakiramanan, S. (1988). International listings and stock returns: Some empirical evidence. Journal of Finance. 42(2), 151-158.
Baker, H. K., Nofsinger, J. R. and Weaver, D. G. (2002). International cross-listing and visibility. Journal of Financial and Quantitative Analysis. 37 (3), 495-521).
Bartram, S.M., Brown, G, and Stulz, R.M. (2009). Why do foreign firms have less idiosyncratic risk than U.S. firms? NBR working paper no. 14931.
Bodurtha, J.N. (1994). International factors and U.S. equity excess returns. University of
Michigan Working Paper.
Brown, S.J. and Warner, J. B. (1985). Using daily stock returns the case of event studies.
Journal of Finance Economics. 14 (1), 3-31.
Chambers, D., Sarkissian, S. and Schill, M.J. (2015). Market and regional segmentation and risk premia in the first era of financial globalization. Cambridge University Working
Paper.
Cetorelli, N. and Peristiani, S. (2015). Firm value and cross-listings: the impact of stock market prestige. Journal of Risk and Financial Management. 8 (1), 150-180.
Doidge, C., Karolyi, A. and Stulz, R. (2004). Why are foreign firms listed in the US worth more? Journal of Financial Economics. 71 (2), 205-238.
Doidge, C., Karolyi, G.A. and Stulz, R.M. (2009). Has New York become less competitive than London in global markets? Evaluating foreign listing choices over time. Journal
of Financial Economics. 91 (3), 253-277.
Domowitz, I., Glen, J. and Madhavan, A. (2001). Liquidity, volatility and equity trading costs across countries and over time. International Finance. 4 (2), 221-256.
Errunza, V.R. and Miller, D.P. (2000). Market segmentation and the cost of capital in international equity markets. Journal of Financial and Quantitative Analysis. 35 (4), 577-600.
Foerster, S. R. and Karolyi, G. A. (1999). The effects of market segmentation and investor recognition of asset prices: evidence from foreign stocks listing in the United States.
Gagnon, L. and Karolyi, G.A. (2010). Do international cross-listings still matter? To be published as a chapter in evidence on Financial Globalization and Crises. Beck, T., Schmukler S. and Claessens, S.
Gozzi, J. C., Levine, R. and Schmukler, S. (2008). Internationalization and the evolution of corporate valuation. Journal of Financial Economics. 88 (3), 607-632.
Hail, L. and Leuz, C. (2009). Cost of capital effects and changes in growth expectations around U.S. cross-listings. Journal of Financial economics. 93 (3), 428-454.
Howe, J.S. and Kelm, K. (1987). The stock price impacts of overseas listings. Financial
Management. 16 (3), 51-56.
Howe, J.S. and Madura, J. (1990). The impact of international listings on risk. Journal of
banking and Finance. 14 (6), 1133-1142.
Karolyi, G.A. (1998). Why doe companies list shares abroad? A survey of the evidence and its managerial implications. Financial markets, Institutions and Instruments. 7 (1), 1-60.
Karolyi, G.A. (2006). The world of cross-listings and cross-listings of the world: Challenging conventional wisdom. Review of Finance. 10 (1), 99-152.
Sarkissian, S. and Schill, M.J. (2009). Are there permanent valuation gains to overseas listing? Review of Financial Studies. 22 (1), 371-412.
Sarkissian, S. and Schill, M.J. (2016). Cross-listing waves. Journal of financial and
quantitative analysis. 51 (1), 259-306.
Stapleton, R. and Subrahmanyam, M. (1977). Market imperfections, capital market equilibrium and corporation finance. Journal of Finance. 32 (2), 307/319.
