Competition among exchanges through simplified disclosure requirements: evidence from the American and Global Depositary Receipts

Abstract

In this study, we address the ongoing debate as to whether the competition among the world's major exchanges through simplified disclosure requirements is justified. Companies from across the globe have a choice of cross-listing shares as either American or Global Depositary Receipts (ADRs and GDRs, respectively). The former are primarily listed on the US exchanges – NYSE, NASDAQ and AMEX – whereas the latter are issued into non-US markets such as the London Stock Exchange (LSE). The GDRs listed on the LSE are subject to simplified disclosure requirements compared to their exchange-listed ADR peers that have to meet more stringent compliance standards. Proponents of the ‘light touch’ approach argue that firms cross-listing as GDRs are not subject to the higher reporting costs faced by ADRs yet still face similar valuation benefits. Those who challenge this approach argue that simplified disclosure requirements set by the LSE will ultimately be recognised by the market as ineffective, diverting traders from investing in GDRs. This study provides evidence that supports the LSE's ‘light touch’ approach and shows that the benefits of information risk reduction for ADRs and GDRs are comparable. The explanation for this finding is that the two avenues through which information asymmetry is expected to be resolved after cross-listing – disclosure and analysts – are substitutive and make equally important contribution to information risk reduction, eventually leading to similar cost of capital decline for ADRs and GDRs.

Keywords:

• analyst forecasts
• cost of capital
• cross-listing
• disclosure
• information risk

Acknowledgements

We thank the participants of the workshops at the University of Technology Sydney, the University of New South Wales, the University of Queensland, the University of Melbourne, the Journal of Contemporary Accounting and Economics (JCAE) 2011 symposium, and the American Accounting Association 2011 annual meeting for their helpful comments and suggestions. We are particularly grateful to the anonymous reviewers and the associate editor for their constructive feedback. Oksana Kim is grateful for research support provided by the Minnesota State University (Mankato) and the University of Melbourne. The manuscript has previously circulated under the title ‘The impact of cross-listing on the cost of equity capital: the case of ADRs and GDRs’. All remaining errors are ours.

Notes

1. The study focuses on the Main Market LSE-listed GDRs and does not cover GDRs listed on the Luxembourg Stock Exchange (LuxSE) and other exchanges due to the fact that, first, the LSE is the world's largest GDR market in terms of market capitalisation and liquidity and, second, the cross-sectional variety of companies trading as GDRs through the LSE's International Order Book (IOB) far exceeds that of the next largest GDR destination – the LuxSE. To compare, as of 2009 there were over 270 securities from 46 countries trading on the IOB service as GDRs (Russian IPO Guide 2010), while the majority, 88%, of the LuxSE GDRs were represented by companies from India and Taiwan and were single-listed (Bank of New York Mellon 2009/2013). The LuxSE is recognised as one of the leading markets for raising capital via debt rather than equity issuance.

2. This study focuses on exchange-listed Level II and III ADRs and GDRs for which cross-listing benefits are expected to accrue through enhanced disclosure requirements. The requirements concerning admission for listing and continuing reporting obligations for GDR issuers are detailed in Chapter 18 of the UK Listing Rules and are available at: http://fshandbook.info/FS/html/FCA/LR/18 (UK Financial Services Authority 2013).

3. In fact, in case of two exchanges with initially close reputations and differential investor base (assuming this is the case for NYSE and LSE) it is the exchange that lowers standards (not excessively) that ultimately sets the new reporting benchmark.

4. A comprehensive summary of the UK listing choices for overseas companies is also available at www.stikeman.com/en/pdf/UK_ListingChoices.pdf‎

5. Further, a closer look at the total DR market reveals that, first, by the end of 2012 the number of exchange-listed sponsored ADRs was about the same as GDRs and second, over the past five years the GDRs programmes have demonstrated a steady growth of 29% (increase in listings), despite the global economic uncertainty, while ADRs remained on the same level (Bank of New York Mellon 2009/2013). Finally, GDRs have consistently outperformed ADRs and provided a higher return on investment to shareholders, and several of the largest public offerings in history were via GDRs (e.g. Gazprom of Russia). The statistics indicate that raising capital as GDRs has become more attractive than listing as ADRs over time.

6. Larcker and Rusticus (2010) discuss problems with estimating the association between voluntary disclosure and cost of capital which can give rise to mixed results. They argue that while the earlier studies treated disclosure as exogenous, the relationship in fact suffers from endogeneity issue. It is possible that firms with lower disclosure levels have higher cost of capital and hence the association between cost of capital and disclosure level is negative. But if such firms decide to disclose more (voluntary disclosure) and do not succeed, the observed association will be positive. In the present study cross-listed firms are subject to mandatory disclosure and have little discretion over the extent of reported information.

7. It is often the case that one cross-listed firm enters an ADR/GDR sample several times when it has changed a name, is acquired, or is restructured. There are multiple cases where ADR-listed firms have changed cross-listing programmes and/or switched to a GDR listing (and vice versa), or prior to an ADR/GDR listing, have been directly listed on the exchanges that require extensive disclosure of accounting information. In addition, some ADR and GDR-listed firms never list on their domestic markets and hence, never become cross-listed. Such firms should not be treated as cross-listed firms and should be excluded from the ADR/GDR sample. The study pays specific attention to the institutional differences and the pre- and post-listing disclosure requirements of GDRs that have not been discussed in the prior literature. GDRs can be listed on more than one market of the LSE and the disclosure requirements for those markets are different. The Main Market of the LSE is a European Union (EU) regulated market that puts in place stringent disclosure requirements for all companies listed on it. Although exempt from a number of reporting obligations to which native (directly listed) UK firms are subject, the GDR programmes listed on the Main Market have been subject to more rigorous disclosure requirements compared to their domestic markets. The disclosure requirements, however, are different – minimal – for the GDRs listed on the exchange-regulated Professional Securities Market (PSM) and the Alternative Investment Market (AIM). Therefore generalising findings on all LSE-listed GDRs (and foreign listings) may not be appropriate.

8. One of the reasons for the shift away from the market segmentation theory as the main framework was due to an overall dissatisfaction with the market segmentation hypothesis and its failure to explain the variation in the cost of capital decline as a function of cross-listed firms and listing destinations. For instance, abnormal returns should vary as a function of the differences in degree of market segmentation, while the empirical evidence provides no support to it (see Foerster and Karolyi 1999). Therefore other frameworks were suggested to explain changes in cost of capital and firms value as result of cross-listing, including the bonding hypothesis and information risk theory.

9. The Final Rule on acceptance of the IFRS financial statements from foreign issuers is formulated in the SEC's Release Nos. 33–8879; 34–57026; International Series Release No. 1306; File No. S7–13–07. According to the Rule, foreign private issuers are allowed to include in their filings with the SEC financial statements prepared in accordance with IFRS as issued by the IASB without reconciliation to US GAAP. The exemption from the reconciliation requirement relates to both annual and required interim reporting periods and applies to financial years ending after 15 November 2007. In Section 4 we discuss how this amendment affected our sample composition.

10. For example, Sarkissian and Schill (2009) show that in order to estimate equilibrium returns around a cross-listing event, long time-series of data are required prior to and after a cross-listing event, putting substantial limitations on the data availability and making the cost of capital estimation noisy.

11. We omit a firm-specific subscript i throughout this section for simplicity.

12. The treatment effect model (3) requires a two-stage estimation. First, the ADR-versus-GDR cross-listing decision equation is estimated using the Probit model. The firm (analyst following, size, industry, growth, leverage and ROE)-, year (inflation)-, listing destination (velocity)- and local market (disclosure level and market performance index)- related effects are used as determinants for a firm's decision in favour of an ADR or a GDR programme. The estimated parameters from the first equation are used to calculate the inverse Mill's ratio, which is then included in the second, cost of capital, equation as an additional explanatory variable (see Leuz and Verrecchia 2000).

13. Note that we model a choice of cross-listing as an ADR or a GDR, given the commitment of firms to cross-list on a high-quality exchange, rather than a decision to cross-list or not to cross-list. Hence, our research design focuses on a choice of two alternative programmes: ADRs versus GDRs.

14. The complete list of country-specific indices and risk-free rate proxies that were used in the study is available from the authors upon request but is not reported in the paper.

15. The number of ADR programmes as per Citibank was substantially different from that as per BNY. For instance, the Citibank's list showed around 570 active exchange-listed ADR programmes, while the BNY database showed only 290, including NYSE, NASDAQ, AMEX (Level II and III programmes) and OTC listings. The effective listing dates for some ADR listings also differed between the two sources and it was unclear what caused these inconsistencies. Prior studies that used the two databases did not specify which data source supplied the most reliable information. As having the correct listing dates and programme types are crucial to the research, it was necessary to verify the effective listing date (including, importantly, listing month) for each company manually.

16. Investigating companies' background information, press releases, and other publicly available reports revealed that the above mentioned inconsistencies in the listing dates and programmes occurred for two reasons. First, in some cases ADR-listed firms changed the sponsor and switched from Citibank, for instance, to the BNY. In this case the BNY database showed the effective listing date as the day when it actually became the sponsor. This is not the actual ADR listing date of a company. Second, if a company switched from a NYSE listing to the OTC trading (the company delisted from an exchange), the BNY database would display OTC as the listing programme of that company along with the OTC-listing year, since this is the last trading programme for the company. For the purpose of our sample, however, such a company is a NYSE-listed Level II or III ADR-listed firm and should be included in the respective sample. To address these inconsistencies we tracked ADR-listed companies back to their first (original) ADR programme to find out whether or not their first listing programme was a Level II or III NYSE/NASDAQ/AMEX listing rather than an OTC trading programme.

17. For example, when analysing Chinese ADRs, we found that none of the 39 firms in the initial ADR sample met a definition of a cross-listed firm at the time of ADR listing. Further investigation revealed that those Chinese ADRs were either listed only on the NYSE without being listed on the local market (22 companies), or they became listed on the local market later and ADR listing is the first listing for them (17 companies). Thus, no Chinese ADRs meet the definition of a cross-listed firm and should be removed from the initial Main (Active) sample.

18. The countries that lost the greatest number of firms (as a percentage of their total ADRs) due to the fact that those firms did not meet the definition of a cross-listed firm were Brazil, Chile, China, Germany, Japan, Italy, Netherlands, Russia, Mexico, Spain, and Switzerland. The UK lost some of its sample due to the fact that different ADR tranches were listed as separate ADR programmes by the BNY and the NYSE, while for the purposes of the study we needed to select the earliest ADR-listed tranche for a particular company and ignore its subsequent ADR issues because changes in the information risk are expected to occur around the first listing.

19. A comparison with earlier cross-listing studies is not possible due to different time periods and sample compositions. For example, in Doidge et al. (2009) the number of firms cross-listed in the USA is substantially higher than in our sample. The authors include Canadian firms in the sample and in addition make a statement that foreign listing and cross-listing are used interchangeably. In our study we clearly distinguish between single-listed and cross-listed firms, and between foreign listings and cross-listings. Finally, our sample's time periods differ from those of Doidge et al. (2009). Our study also extends the time period in the study of Hail and Leuz (2009) and our sample is larger.

20. In 2007, the SEC eliminated the requirement to reconcile financial statements to the US GAAP for foreign private issuers reporting in accordance with IFRS as issued by the IASB and this amendment applies to financial statements with financial year ending post 15 November 2007. In our sample of ADRs, there were seven cross-listed firms that could potentially be affected by this regulation: 1 from Argentina, 2 from Brazil, 1 from Japan, 1 from India, 1 from Ireland and 1 from the UK. Only two firms – one from Ireland and one from the UK – were reporting in accordance with IFRS prior to cross-listing. Nevertheless, the IFRS version used in Ireland and the UK is as adopted by the EU rather than the IASB. Therefore, these two companies were not exempt from the US GAAP reconciliation requirement upon cross-listing.

21. This is a starting point for GDRs data collection process reported in Table 1.

22. In the final ADRs/GDRs sample the distribution by industry is as follows: (1) = 25%, (2) = 11%, (3) = 39.5%, (4) = 8%, (5) = 10%, (6) = 6.5%. Where: (1) financial institutions, (2) mining and related business, (3) retail business, (4) telecommunications and related business, (5) information technology, (6) biotech and pharmacy.

23. The Datastream coverage for non-US and non-Canadian firms dates to 1987.

24. Note we do not include Canadian firms cross-listed in the USA in our analysis because those firms are subject to the Multi-Jurisdiction Disclosure System agreement (1991), according to which Canadian firms listed in the USA are allowed to use disclosure and reporting requirements enforced by the home-market authorities. Therefore, the post-listing disclosure requirements for Canadian firms are different from those for companies domiciled in other countries that do not have similar agreements with the USA.

25. Note that throughout the study we assume that the US and the UK environments are reasonably similar with respect to investor protection level and law enforcement, consistent with discussion and evidence provided by the corporate governance literature (La Porta et al. 1997, 1998).

26. Although we posit that firms in the modified ADRs sample and GDRs are expected to experience similar changes in visibility, this remains just a theoretical claim and hence we still want to control for possible differences in home market disclosures and enforcement practices between the modified ADRs and the GDRs sample, and hence we include the CIFAR index validated in prior studies (Hope 2003).

27. For example, when model (8.2) is estimated for the sample that excludes common law developed countries the coefficient on Foll is −2.39** and the coefficient on the interaction term Disc*Foll is 2.17**.

28. Our final check addresses the issue of using small samples of ADRs and GDRs. We repeated the tests based on the individual OHS, ETSS and realised returns sub-samples that are larger than the overlap sample. The results of these tests leave our conclusions unchanged. Finally, we winsorised 1% of data and the results were in line with those documented above.