Δευτέρα 31 Αυγούστου 2009
Δευτέρα 17 Αυγούστου 2009
Κυριακή 16 Αυγούστου 2009
Τετάρτη 27 Μαΐου 2009
Σάββατο 23 Μαΐου 2009
Πέμπτη 9 Απριλίου 2009
Τετάρτη 8 Απριλίου 2009
Σάββατο 4 Απριλίου 2009
Which Firms Choose to List as American Depositary Receipts (ADRs)?:
Evidence from Mexican Cross Listings
Paquita Y. Davis-Friday*
University of Notre Dame
386 Mendoza College of Business
Department of Accountancy
Notre Dame, IN 46556
pfriday@nd.edu
and
Thomas J. Frecka
University of Notre Dame
305B Mendoza College of Business
Department of Accountancy
Notre Dame, IN 46556
tfrecka@nd.edu
December 2002
*Corresponding Author. The authors appreciate comments from Donal Byard, Tim Loughran, Darius Miller, workshop participants at the University of Cincinnati, the University of Notre Dame, and Rutgers University, and participants at the 2000 AAA International Accounting Section Mid-Year Meeting and the 2001 Annual Congress of the European Accounting Association. The authors are also grateful for financial support from PriceWaterhouseCoopers and Arthur Andersen.
Which Firms Choose to List as American Depositary Receipts (ADRs)?:
Evidence from Mexican Cross Listings
ABSTRACT: This study provides evidence that Mexican firms that choose to trade in the U.S. as exchange-listed American Depositary Receipts (ADRs) have significantly weaker financial performance subsequent to cross-listing than Mexican firms that are eligible to list in the U.S., but do not. We show that on average, Mexican cross-listed firms are smaller, more highly levered and less profitable than non-crosslisted firms (NCL). While supplemental tests of earnings quality suggest that NCL firms exhibit significantly smoother earnings, that evidence is not sufficient to explain the stronger reported financial performance of those firms relative to cross-listed firms. Finally, our tests of value relevance, using book value and earnings to explain price, show significantly higher explanatory power for the cross-listed firms. The value-relevance results suggest that benefits from U.S. regulation exist for investors in Mexican cross-listed firms and that reported market inefficiency in Mexico may result in low demand for the financial statements of NCL firms.
Key Words: American Depositary Receipt; Mexico; Cross listing
Data Availability: Data used in this study are available from public sources, as indicated in the text.
JEL classification: E31; G15; M41; N26
1. Introduction
This paper investigates whether there are systematic differences between Mexican firms that choose to trade in the U.S. as American Depositary Receipts (ADRs) and Mexican firms that are eligible to trade in the U.S., but do not. Specifically, we examine the roles of country-specific institutions and firm-specific factors in explaining the cross-listing decision for firms from an emerging market with close ties to the U.S. This question is especially important in light of recent research findings in Lang, Raedy and Yetman (2001). Their results indicate that cross-listed firms perform better, are less risky and less aggressive in terms of earnings management. They also find that cross-listed firms report accounting data that are more strongly associated with share prices. However, their evidence is based on samples from 26 countries representing both developed and emerging markets and differing corporate governance environments. LaPorta et al. (1998) suggest that country-specific legal rules regarding the protection of shareholders and creditors, and the quality of enforcement of those rules, help define the rights that various securities afford their owners. These rights in turn determine whether or not firms can raise external finance (LaPorta et al. 1998, 1114). Therefore, country-specific legal rules may create differing incentives for firms to seek financing outside of their home countries and as important, investors may respond differently to the capital raising efforts of firms from different countries. The law and the quality of its enforcement are likely especially important factors for firms from emerging markets that seek financing outside of their home countries (Lins, Strickland and Zenner 2000, 4).
We consider the implications of the LaPorta et al. (1998) study for firms cross listing in the U.S. by investigating firms from Mexico, the United States’ second largest trade partner and a country from which the majority of eligible firms have cross listed on U.S. exchanges. We 1
focus on an important Mexican institution, the use of segmented share classes to prevent foreign control of Mexican corporations, as well as other country and firm-specific characteristics that may cause the type of cross-listed firm from Mexico to differ from those from other countries. Evidence from Foerster and Karolyi (2000) suggests that companies cross listing on major U.S. exchanges from countries with significant investment barriers for foreigners, like those that exist in Mexico, outperform their benchmarks while those from the same countries that issue equity via Rule 144a private placements significantly under perform.
Our results indicate that, on average, Mexican cross-listed firms are smaller, more highly levered and less profitable than the non-crosslisted (NCL) firms. While supplemental tests of earnings quality suggest that NCL firms exhibit smoother earnings, the evidence is not sufficient to explain the stronger reported financial performance of Mexican NCL firms. Finally, our tests of value relevance, using book value and earnings to explain price, show significantly higher explanatory power for the ADR firms and generally non-significant explanatory power for the NCL firms. The value-relevance results may indicate that investors in Mexican ADR firms benefit from U.S. regulation and that evidence of market inefficiency in Mexico (Bhattacharya et al. 2000) may result in low demand for financial statements of NCL firms. This study has the advantage of focusing on a single, emerging-market economy in contrast to most previous ADR research that uses multi-country samples dominated by developed market countries. It is also one of the first studies to deal with selection bias issues by comparing cross-listed and non-crosslisted firms. These advantages are obtained at the cost of having to base tests on and draw conclusions from relatively small samples.
The rest of the paper is organized as follows. Section 2 reviews prior research and provides additional background information on Mexican institutions and the ways Mexican
2
companies have accessed U.S capital markets. Section 3 develops our hypotheses and describes the research design to test the hypotheses. Section 4 describes our data sources, the composition of our samples, and presents the results of our analyses. Section 5 concludes the paper.
2. Background and Prior Research
2.1 Accessing U.S. Capital Markets
Most non-U.S. exchange-listed companies trade in the form of American Depositary Receipts (ADRs). ADRs are dollar-denominated derivative instruments. A depositary bank (primarily the Bank of New York in the case of Mexico) obtains claims on foreign securities and issues dollar-denominated claims against them. The ADRs pay U.S. dollar dividends and are processed through the U.S. clearance and settlement system.1 A company that wishes to list and trade securities in the secondary market will choose either the Level 1 or Level 2 ADR. The advantages of Level 2 listing are the availability of price quotes and the greater visibility associated with exchange listing. The primary disadvantage is the need to meet U.S. disclosure requirements, including reconciliation to U.S. GAAP.
In order to raise new capital in the U.S., the non-U.S. company must choose a Level 3 ADR and meet the standard SEC requirements for issuing securities, such as issuing an F-1 Registration Statement, providing annual 20-F filings, and reconciling financial statements to U.S. GAAP, or choose a semi-private offering via Rule 144a. Rule 144a allows privately placed securities to be sold to qualified institutional buyers (e.g., large insurance companies, banks and investment companies). Such sales are exempt from the SEC registration requirements.
1 Pulatkonak and Sofianos (1999) note that NYSE-listed non-U.S. companies trade in several forms including ADRs, Global Depositary Receipts, home-market "ordinaries" in the case of Canada, and New York Registered Shares. The instruments differ in the degree of fungibility with the corresponding home market securities. For
3
However, such issuers must provide buyers with information about the nature of the business and recent audited financial statements, generally prepared under home country GAAP. Furthermore, Rule 144a offerings are restricted to smaller companies since registration with the SEC is generally required if the assets of the foreign issuer exceed $10 million or if it has 500 or more equity holders. We investigate firms that have chosen to raise capital across all of these dimensions as well as firms that are eligible to trade on U.S. exchanges, but have not cross listed.
2.2 Evidence Regarding Cross Listing
Saudaragan (1988, 1990) argues that the potential benefits of foreign listing relate to financial, marketing, political and employee motivation considerations. The primary financial considerations are access to larger capital markets, reduced dependence on domestic sources of capital, avoidance of a decline in stock prices due to saturation of domestic markets, access to foreign capital to finance foreign expansion, and the opportunity to meet operating and long-term debt obligations abroad. Market considerations relate to greater visibility and product identification abroad and enhanced corporate image. Political motivations include improved international relations and employee motivation considerations include improved labor relations and the use of stock in employee compensation plans abroad.
Baker, Nofsinger and Weaver (1990) investigate changes in visibility associated with cross listing, using analyst coverage and media attention as proxies. Their results from analyzing changes in analyst coverage are consistent with the “investor recognition” hypothesis of Merton (1987), which indicates that an increase in the size of the investment base lowers expected return, lowers the cost of capital and increases the market value of shares.
instance, Canadian ordinaries are fully fungible because a trade made in the U.S. will be denominated in U.S. dollars while the same trade made in Canada would be denominated in Canadian dollars.
4
We do not explore visibility issues in this paper, but instead focus on access to capital as the primary firm benefit from cross listing in the U.S. Limited access to external capital is a primary motive for foreign listing in Bruner, Chaplinsky and Ramchand (1999), Miller (1999), and Stulz (1999). Later, we provide evidence of substantial capital raising for our exchange-listed ADR firm sample.
Lins, Strickland and Zenner (2000) empirically investigate whether cross listing in the U.S. reduces the costs of market segmentation by improving access to capital.2 Using an ADR sample covering the period 1986-1996 and data from Worldscope, the authors document that the sensitivity of investment to cash flow decreases significantly for firms from emerging markets but does not change for developed market firms following cross listing and that firms access international capital markets more often following cross listing. They suggest that this may be the case because emerging markets are often segmented. Market segmentation may arise from direct (e.g. ownership restrictions) and indirect barriers (e.g. information production, accounting standards). These barriers impede the flow and formation of capital in non-U.S. markets (Lins et al. 2000, 1). Lins et al. (2000) argue that relative to their home markets, greater disclosure requirements, shareholders rights protection, liquidity, and analyst following reduce the information asymmetry for non-U.S. firms listed on U.S. equity markets, especially those from emerging markets (Lins et al. 2000, 2).
2 We borrow the following quote from Lins et al. (2000) because it is so relevant for our study: “Stock markets from Mexico City to Sao Paulo have sunk in importance recently with trading shifting to Latin stocks listed in New York as American Depositary Receipts, or ADRs. … The focus abroad means the only Latin American companies that can raise financing nowadays are those that have access to international markets. Thousands of small and medium-sized businesses—the bulk of Latin America’s productive capacity—are left behind.” (“Latin American Stock Trading Migrates North,” Wall Street Journal, October 27, 1999).
5
2.2 The Focus on Mexico
Over the past decade there has been a sizable amount of trading of Latin American securities within U.S. markets as well as capital issuance using ADRs and Rule 144a listings. For example, in 1995, ten (six) of the twenty-five most actively traded foreign stocks in the U.S. in terms of share volume (dollar volume) were issued by companies domiciled in Mexico, Brazil, Argentina, and Chile.3 Currently, of the 458 non-U.S. companies listed on the NYSE, 95 are from Latin America and 29 are from Mexico.4
Mexico offers a particularly rich environment for conducting research on corporate finance choices as it differs substantially from that for U.S. firms. First, the efficiency of the market is open to question based on recent evidence in Bhattacharya et al. (2000) that Mexican restrictions on insider trading are not being enforced. Second, Mexican accounting rules and investor protections are rooted in the civil law tradition of France in contrast with the common law roots of the U.S. (LaPorta et al. 1998). LaPorta et al. (1998) find that countries from the French civil law family generally provide the worst legal protections to shareholders. Third, Mexican companies use inflation accounting in their domestic financial statements and in reconciling to U.S. GAAP, they are not required to change the measurement model, thus reducing the cost of reconciliation to U.S. GAAP. Fourth, Mexico provides a contrasting environment, in comparison with the U.S., for considering the effects of ownership structure. As is the case for several other emerging market countries, Mexican firms typically issue multiple
3 See Cochrane, Shapiro and Tobin (1996) for this and additional evidence concerning the internationalization of U.S. investment.
4 NYSE website, www.NYSE.com, statistics as of November 2001.
6
classes of equity that differentiate between foreign and domestic traders.5 Domowitz et al. (1997) have documented significant price premia for Mexican shares not restricted to domestic investors.
These market efficiency, legal, disclosure and shareholder class features, coupled with Mexico’s traditionally heavier reliance on banking institutions as opposed to capital markets as a source of capital, and a tendency for companies to be closely held, may make Mexican firms significantly different than U.S. firms and firms cross-listed from other countries.
A reliable comparison of the financial performance of Mexican ADR and NCL firms may be impacted by the potential for differential information quality between the two groups. In this regard, Lang, Raedy and Yetman (2001) compare the characteristics of local-GAAP earnings for ADR firms relative to matched samples of NCL firms and U.S. firms.6 They analyze financial performance based on standard profitability and risk measures and analyze accounting quality differences based on earnings smoothing, accrual management and earnings frequency distribution tests. In general, they conclude that ADR firms are of higher quality than NCL firms.
The Lang, Raedy and Yetman (2001) results are interesting for two reasons. First, their evidence that ADR firms tend to be more financially healthy and of lower risk than NCL firms conflicts with the results of the global equity offering literature (in terms of returns) and is inconsistent with our hypothesis that Mexican ADR firms will exhibit weaker financial performance than NCL firms. Second, their study presents aggregate results for a sample
5 Mexican A shares represent a majority of corporate ownership and are restricted to Mexican nationals while B shares are available to all investors. See Davis-Friday (2001) and Gordon (1999) for more information regarding Mexican capital market and financial reporting institutions.
6 The Lang, Raedy and Yetman (2001) study is the only other study of which we are aware that deals with the selection bias issue by comparing ADR and non-ADR firms. However, their study differs from ours in several 7
spanning twenty-six countries, but dominated by United Kingdom (34.5% of sample) and Japanese (15.19% of sample) firms. The mixing of code law and common law countries and developed and emerging market economies makes it difficult to generalize the results and masks any potential individual country differences.
3. Research Design and Hypothesis Development
We develop and test four hypotheses. First, firms that cross list exhibit weaker financial performance than those that do not cross list. Second, the major motivation for cross listing is to reduce capital constraints. Third, because of the lack of enforcement of regulations in Mexico, Mexican cross-listed firms’ share prices will be more associated with their accounting information as a result of the additional U.S. regulation and finally, ADR firms’ earnings will be of higher quality than those of non-ADR firms.
3.1 Financial Characteristics of ADR Firms
Prior research does not compare the financial performance of Mexican cross-listed companies with eligible companies who have not cross listed, so there is no prior empirical justification for positing a directional hypothesis regarding the difference between the financial performance of ADR and NCL firms. However, some general guidance is provided by the global equity offerings literature.
Foerster and Karolyi (2000) analyze the long-run performance of a sample of 333 global equity offerings via U.S. ADRs over the period 1982-1996. They show that for their sample of Asian, Latin American and European issuers, three-year post-issuance returns are from 8-15% lower than for local market benchmarks. The results are consistent with the documented results
respects, including the fact that their sample of non-ADR firms is not restricted to eligible cross-listers. Readers are
8
for domestic equity offerings (e.g., Ritter 1991, Spiess and Affleck-Graves 1995, and Loughran and Ritter 1995). Foerster and Karolyi (2000) investigate regional differences in their overall results, and after separating the returns for private and public issues, they show that the key driver of poor performance in Latin America is the public issues program (e.g., Tubos de Acero de Mexico). Segmenting their sample into developed and emerging market groups, global equity offerings for developed markets outperform emerging markets for all horizons and benchmarks. No specific country results were reported.
Based on both the domestic and global equity offering literature, we expect weaker financial performance for our ADR population in comparison with NCL firms (benchmark). We also attempt to analyze before- and after-cross listing performance for a small sample of ADR issuers for which we have sufficient data. Again, based on prior literature, we expect to observe weaker financial performance after cross listing for these firms and this is the basis of our first hypothesis stated in the alternative form:
H1: ADR firms exhibit weaker financial performance relative to NCL firms and relative to their own performance prior to cross listing.
Consistent with the Lins et al. (2000) idea that foreign companies issue equity on U.S. exchanges to relax capital constraints, we expect our ADR firms to exhibit higher financial leverage than the benchmark NCL firms. Whether or not these firms are successful in loosening capital constraints subsequent to ADR listing depends on post-ADR listing performance. While we do not posit a reduction in financial leverage subsequent to ADR listing, we do provide evidence for a small sample of firms for which we have data. Our second hypothesis describes our expectations regarding the need for capital. It is stated in the alternative form:
referred to the discussion of selection bias issues discussed in that paper.
9
H2: ADR firms are more levered than NCL firms.
3.2 The Effect of Corporate Governance on Cross Listing
The cost of greater transparency is a potentially more binding constraint on cross listing for Mexican firms than are out-of-pocket costs. In a series of papers, LaPorta et al. (1997, 1998, 1999) show that the observed prevalence of widely held corporations in the U.S. does not hold as a pattern throughout the world and that widespread corporate ownership, where it does exist, is consistent with strong legal protection for minority shareholders. LaPorta et al. (1999) analyze the ownership of the top twenty market capitalization firms in each of 27 countries, including Mexico. Ownership structure information is not available for non-crosslisted Mexican companies, so they focus on the 20 largest ADR firms. All are classified as “family owned” by the authors.7 Thus, even though privacy concerns exist as a potential constraint on cross listing since Mexico has one of the highest incidences of family ownership, the majority of eligible Mexican firms has cross listed in the U.S.
In two papers focused on Mexico, Domowitz, Glen and Madhavan (1997, 1998) explore the effects of ownership restrictions and cross listing on stock prices and market liquidity. Their share price analysis of 24 equity issues on a daily basis and 46 equity issues on a weekly basis for 21 firms over the 1990-1993 period indicates significant price premia for shares held by foreign investors (B shares). They also examine the impact on the Mexican stock exchange (BMV) around the time of ADR issuance and present strong evidence of increased volatility and weak evidence of reduced liquidity (stronger evidence of reduced liquidity for B shares). In
7 We verified the high incidence of family ownership of Mexican firms by referencing a sample of 20-F registration statements available to us.
10
examining the impact of ADR listing on bid-ask spreads, they observe decreased spreads in 22 cases.
Additionally, Mexican firms are more likely to benefit from cross listing than are firms from many other countries because of time zone considerations. Pulatkonak and Sofianos (1990) indicate that the six countries with the greatest U.S. market share are all within one time zone of New York and that 68.7% of the trading of Mexican cross-listed firms occurs in the U.S. The importance of time zone relates primarily to trading session overlap facilitating arbitrage trading, simultaneous security pricing in the U.S. and home markets, and the ability to disseminate new information during business hours.
Finally, in their tests of the efficiency of the Mexican stock market, Bhattacharya et al. (2000) note that in spite of laws against insider trading, there has never been an indictment, trial, or conviction for insider trading by the Mexican equivalent of the SEC. Their examination of daily price reactions to corporate news announcements for a sample of Mexican A and B shares over the period July 1994-June 1997 shows no difference in returns, volatility, trading volume, or bid-ask spreads for event and non-event periods. They provide further evidence that the return volatility of A shares leads the return volatility of B shares (consistent with insider trading), but not strongly enough for trading rules to arbitrage it away.
These issues relate directly to the results in Lins et al. (2000). Their results demonstrate that U.S. listing reduces the indirect costs of market segmentation by improving access to capital, especially for emerging market firms. In the case of Mexico, ADR shares traded in the U.S. may represent underlying B shares or A shares (those previously restricted to ownership by Mexican nationals). Therefore, the cross listing reduces market segmentation and potentially improves the informational efficiency of the cross-listed share prices. Thus, a potential benefit
11
of cross listing for investors in Mexican securities may be improved market efficiency resulting from U.S. regulation.
3.3 Accounting Information and Regulation
Much of the prior research related to the SEC disclosure debate has focused on the value-relevance of GAAP reconciliations. Amir, Harris and Venuti (1993) examine the value-relevance of both the aggregate reconciliations of earnings and shareholders’ equity and the value-relevance of the individual components for a sample of 101 firms based on Form 20-F filings over the period 1981-91. They find value-relevance for both aggregate reconciliations and for the capitalized goodwill and asset revaluation components of the reconciliations. Their sample is dominated by firms from the U.K., the Netherlands, Australia, and Sweden and includes only three Mexican firms.
Rees and Elgers (1997) extend the work of Amir et al. (1993) by focusing on the reconciliations found in initial Form F-1 registration statements for a sample of 140 (mostly U.K., Australian and Canadian) firms for the period pre-1985 through 1991. The authors regress ROE (book value deflated earnings), and the reconciliations (change in stockholders’ equity and the change in earnings) on price-to-book value for up to three retrospective years and find statistically significant results for the stockholders’ equity reconciliation and some individual components of the reconciliations (goodwill, asset revaluations and “all else” categories). Regressions of reconciling items on returns were not significant.
Davis-Friday and Rivera (2000) investigate the value-relevance of Form 20-F reconciliations for a sample of Mexican ADRs. They find that while both Mexican GAAP and U.S. GAAP earnings and book values are significantly related to ADR prices, the reconciliation from Mexican to U.S. GAAP is not statistically significantly related to ADR prices. The authors
12
suggest that a likely reason for the apparent lack of value-relevance for the reconciliation is that Mexican financial statements are inflation-adjusted and the reconciliations exclude the inflation adjustments. The inflation adjustment is likely the largest difference between Mexican and U.S. GAAP.
To our knowledge, no prior study has compared the value-relevance of accounting information reported by Mexican ADR and NCL firms. We expect to observe significantly greater value-relevance for our ADR firms in comparison with NCL firms. This result is anticipated due to the relative nature of the information environments in which the two sets of firms operate. According to evidence from Bhattacharya et al. (2000), the U.S. capital market appears to be more highly regulated than the Mexican market. Therefore investors may be more willing to use financial statement information to set security prices for the ADR firms than for the NCL firms. Our third hypothesis, stated in the alternative, is as follows:
H3: Earnings and book values explain more of the cross-sectional variation in stock prices for ADR firms than for NCL firms.
Based on our limited data set, we perform quality of earnings tests for our ADR and NCL samples similar to those in Land and Lang (2002) and Lang, Raedy and Yetman (2001). Consistent with the Lang et al. (2001) results and the regulatory environment in Mexico, we expect to observe lower quality accounting for the NCL firms. Failure to reject this hypothesis will give some credence to the belief that any observed difference in financial performance between the two groups is reliably measured and ADR firms exhibit higher earnings quality than NCL firms. Our final hypothesis, stated in the alternative form is:
H4: ADR firms exhibit higher earnings quality than NCL firms.
13
4. Results
4.1 Data and Sample Selection
Our primary data source for financial statement and market data is Economatica, a commercial database for Latin American companies. Economatica obtains its firm financial statement data from the Mexican stock exchange (Bolsa). We obtained security market betas from Bolsa statistical summaries (Anuario Bursatil de la Bolsa Mexican de Valores) and data for return calculations from the same source. The number of analysts following our sample firms was obtained from IBES International.
The financial statement data used in subsequent analyses were prepared using Mexican accounting principles. The accounting data are general price-level-adjusted and translated to U.S. dollars using peso-to-dollar exchange rates at each year-end.8 We do not analyze reconciliations to U.S. GAAP for level 2 and level 3 ADR companies for two reasons. First, we want to be able to compare financial characteristics of ADR and NCL firms on the same basis. Second, previous research (e.g., Davis-Friday and Rivera (2000)) indicates that the reconciliations are not value-relevant for Mexican companies. We summarize the sample selection procedures in table 1.
[Insert table 1 here]
The ADR sample consists of all Mexican non-financial services firms with available data that were ever listed on U.S. exchanges during the 1995-1999 period.9 At the end of 1995, there
8 Two inflation adjustment accounts are normally shown in Mexican financial statements. First, any gain or loss on net monetary position is shown in the income statement as part of a comprehensive financing cost category that also includes interest and foreign exchange gains or losses. The offsetting account for the price level adjustments for nonmonetary items is “Cumulative effect of restatement” or “Gain or loss from holding non-monetary assets” and is a component of Shareholders’ Equity. Most of the sample companies had large cumulative losses from holding monetary assets during our test period.
9 Banks were eliminated due to their differing financial characteristics in comparison with the rest of the sample. This constraint eliminated two financial services firms (Banca Quadrum and Grupo Financiero Serfin) from the
14
were 22 Mexican ADRs listed on the New York Stock Exchange (NYSE); three traded on the NASDAQ, and two on the American Stock Exchange (AMEX). By the end of 1999 there were 34 Mexican firms trading on the NYSE, NASDAQ or AMEX. After excluding the two banks (footnote 10), the ADR sample ranges from 23 to 32 firms for a total of 140 firm years.10 However, in subsequent analyses, two firm-year observations with negative book values and three outliers (ROE greater than three standard deviations from the mean) were eliminated. Additional missing data (especially price data) further limited the sample in some years (see table 4). All of the sample firms have December 31 fiscal year ends. The sample includes slightly more level 3 ADRs than level 2 ADRs.
The non-ADR sample is also drawn from Economatica, and includes all additional Mexican firms that meet the NYSE international listing requirements, but are not exchange-listed.11 This sample includes firms that are listed over-the-counter (OTC) or as private placements (144a).12 It also includes firm years for three companies that later listed as Level 2 ADRs: Cemex (1995-1998), Fomento Economico Mexico (1996-1997), and Grupo Maseca (1995-1997). The annual non-ADR sample size ranges from 14 to 29 firms (40 different firms) and a total of 121 firm years.
ADR sample and three financial services firms (Grupo Financiero Banamex Accival, Grupo Financiero BBVA Bancomer and Grupo Financiero Inbursa) from the NCL sample.
10 The ADR level designations in panel A of Appendix B are the current level designations as disclosed by the Bank of New York. Note that three previous level 2 or 3 ADR firms have downgraded to level 1—Bufete Industrial, Grupo Mexicano de Desarrollo and Grupo Sidek. All three firms traded on the NYSE during our test period, but they presently trade OTC.
11 The alternate international requirements include: 5,000 worldwide round-lot holders, 2.5 million worldwide public shares, $100 million worldwide public market value, and aggregate three-year pre-tax earnings of $100 million together with a minimum of $25 million of earnings in each of the two most recent years. The financial conditions may also be fulfilled using operating cash flow.
12 We conduct our empirical analyses separately on firms that are not crosslisted at all (NCL), OTC & 144a firms, and exchange-listed ADR firms.
15
A summary of the listing status of the sample firms is provided in table 2. A key difference between the ADR and non-ADR firms (NCL and OTC/144a) is that the exchange-listed firms are much more stable in terms of meeting the NYSE listing requirements. Panel A shows that, of the 32 ADR firms in the sample, 23 were listed for all five years of the sample period. The average number of sample years per firm is 4.38. In contrast, panel B shows that only eleven of the 40 non-crosslisted (NCL) and OTC/144a firms in the eligible but not exchange-listed sample meet eligibility requirements for all five test period years. Further, sixteen of the 40 firms meet the eligibility requirements only intermittently and again, three of the firms later become ADRs. The average number of sample years per firm is only 3.03 for the non-ADR sample.
It appears that a substantial majority of eligible Mexican firms have cross listed their securities in the U.S. as level 2 or 3 ADRs. Many of the rest of the firms have only recently or intermittently met NYSE eligibility requirements. For others, we suspect there are idiosyncratic reasons for not cross listing. For example, Kimberly Clark de Mexico, Sears Roebuck and Wal Mart de Mexico have U.S. counterparts and there appear to be few benefits to cross listing for these firms. Grupo Bimbo, an eligible, but non-crosslisted firm, acquired a bakery in Texas in order to take advantage of distribution channels and capital resources in the U.S.
It appears that, at least for eligible Mexican companies, the SEC disclosure requirements are not a major constraint to cross listing. Apparently, the benefits of cross listing (primarily access to capital, we believe) exceed the costs. Over-the-counter (OTC) and 144a firms also have the opportunity to benefit from raising private capital, but without incurring the costs of SEC registration. Later, we provide evidence bearing on the viability of this option.
[Insert table 2 here]
16
4.2 Industry Classifications
Our analysis of the differences between ADR and non-ADR firms begins with a description of their respective industry (sector) distributions in table 3. Based on the sector names from Economatica, the ADR and non-ADR firms represent 18 different sectors. The 32 ADR firms are in 11 different sectors and the 40 non-ADR firms are in 16 different sectors. While we find broad industry representation for both samples, there are also substantial industry differences. The top four industries represent over 59% of the firm year observations in the ADR sample (communications 18.7%, construction 15.6%, steel 12.5%, and food 12.5%). In contrast, even though the non-ADR sample is more diverse in terms of industry representation, the top four industries (retail 22.5%, beverages 12.5%, cement 10.0%, and holding companies10.0%) comprise 55.0% of that sample.
The apparent sector self-selection for cross listing creates two potential problems. First, any observed differences in the overall financial characteristics of the samples have the potential to be industry-specific. Second, the lack of comparable industry data for the ADR and non-ADR samples makes it difficult to perform a matched sample analysis along the lines of previous studies (e.g., Ely and Pownall (1999) and Lang, Raedy and Yetman (2001)). In an attempt to deal with both problems, we present overall sample results and also discuss results for a reduced sample of five industries that have sufficient comparative data.
[Insert table 3 here]
17
4.3 Descriptive Statistics
Table 4, panel A, presents descriptive statistics pooled over the 1995-1999 test period for financial statement and market variable comparisons of three groups: ADR, OTC/144a and NCL firms. Panel B presents various ratio and growth performance measures for the three groups and significance tests are summarized in panel C. The data are somewhat skewed by larger firms, so the medians are better measures of central tendency than the means.
Panel A of table 4 shows that Mexican ADR firms tend to be nominally smaller than both the NCL and OTC/144a firms. For example, for the six financial statement variables, five of the ADR medians and four of the six means are smaller than those of the other two groups. In addition to the size differences, of particular interest is the apparent higher financial leverage and lower profitability of the ADR sample compared especially to the NCL firms.
Several market value measures are presented in panel A. Market value of equity (MVE) is nominally highest for the OTC/144a firms and lowest for the ADR firms. The Relative Size variable (ratio of MVE to total market capitalization of the Bolsa) is also highest for the OTC/144a firms and lowest for the ADR firms. We were only able to obtain small samples of beta and returns data, therefore it is difficult to generalize much from these results, but there is some evidence of lower returns for the ADR firms. Finally, analyst coverage tends to be highest for the OTC/144a firms with little difference in coverage between ADR and NCL firms.
An overall conclusion from panel A of table 4 is that, after meeting threshold size requirements for listing on the NYSE, the relative size of a firm is not a determinant of cross listing (i.e., relatively smaller eligible firms do not avoid cross listing) status for Mexican firms.
Panel B of table 4 summarizes financial and market performance measures for the three groups. Even after deleting outliers, the distributions remain somewhat skewed so we again focus more
18
heavily on the medians. Here, the overall conclusion is that the financial performance of the ADR firms is weaker than that of both the NCL and OTC/144a firms over the 1995-1999 test period. For example, the mean (median) ROE’s for the ADR sample of –10.5% (10.1%) are significantly lower than the mean (median) ROE’s of 13.7% and 14.7% (16.2% and 13.2%) for the respective OTC/144a and NCL samples. Further, the relatively lower ROE’s for the ADR firms are impacted by significantly higher leverage (Assets/Equity) in comparison with the NCL firms. The ROA measures are also significantly lower for the ADR firms. Further analysis shows that the ROA results are driven by both significantly lower profit margins and asset turnover for the ADR firms. Based on the growth in total assets variable, it also appears that historical growth has been lower for the ADR firms than for the other two groups.
Further analysis shows that an important difference in the distributions of ROE and ROA for the ADR firms in comparison to both the OTC/144a and NCL groups is the presence of 36 negative observations for the ADR group compared with only four negative observations for the other two groups combined. The negative observations remain even after deleting negative book value observations (2) and outliers greater than three standard deviations from the original means of the distributions (3). In fact, nineteen of the 32 ADR firms in the sample had at least one year of negative ROE/ROA. The wide distribution of net income/net loss particularly impacts the standard deviation of ROE (1.645) for the ADR group. The negative observations are indicative of weaker financial performance after cross-listing for the ADR firms, since the stock exchanges require at least three years of positive earnings prior to cross-listing.
The weaker financial performance for the ADR sample (and expectations of weaker future performance) is also reflected in the price-to-book (PB) and price-to-earnings (PE) ratios. The median PB ratio of 1.410 for the ADR sample is significantly lower than the median PB’s of
19
1.785 and 1.910 for the respective OTC/144a and NCL samples. While the differences in the PE ratio means and medians are not significant, the median ratio is nominally lower for the ADR sample.
Panel C of table 4 summarizes the results of pairwise comparisons of mean and median differences for the three groups. Most of the significant differences have been noted above. The primary additional insight is that there are significant differences in the variables for all three groups. Based on ROE and ROA measures, financial performance is strongest for the OTC/144a firms and weakest for the ADR firms. The NCL firms differ most prominently from the ADR and OTC/144a firms in terms of their lower financial leverage. The significant differences among all three groups cause us to analyze each group separately in tables 6 through 9.
[Insert table 4 here]
We now report on an untabulated analysis of ADR and NCL differences for a small sample of industry-matched firms. We include all firms with at least four years of data from the following industries (number of ADR and NCL firms, respectively, from each industry in parentheses): food (2,1), holding companies (3,3), paper products (1,1), retail (2,3) and steel (4,1). Thus, the reduced sample includes only 12 ADR and nine NCL firms, and at most, 56 (47) pooled observations for the ADR (NCL) groups.
The results for the reduced sample are very consistent with those for the full sample reported in table 4. For example, we still observe the significant size difference between the two groups. One indicator of the size difference is MVE (market value of equity). The mean (median) for the NCL group is 22,197(17, 817) in comparison with 5,351(3,317) for the ADR group. The results for the performance measures are also generally consistent with those in table 4. While the nominal difference in ROE (NCL lower) is no longer significant, the ROA and
20
leverage differences remain significant, indicating that ROE is impacted by higher leverage for the ADR group. For the full sample, the ROA difference was driven by significant differences in both profit margin and turnover; for the reduced sample, only the difference in turnover is significant. The results for PB and PE are consistent with those for the full sample. Finally, we note that the Revenue Ratio (ratio of domestic revenue to total revenue) difference is now significant for the reduced sample. This suggests that, after controlling for firms’ industry membership, the exchange-listed ADR firms earn more of their revenues outside of Mexico.
Using our non-ADR sample as a reasonable local market benchmark, the weaker performance for Mexican ADR firms, based on financial characteristics, is consistent with the results of Foerster and Karolyi (2000) for their combined 35-country sample based on returns.13 In an attempt to compare the before and after exchange listing performance of ADR firms, we work with a reduced sample of level 2 and 3 ADRs for which we can obtain data for at least one year prior to cross listing. Table 5 compares the financial performance of these firms for one or two years preceding and for two years after ADR listing. While the sample is very small and our tests lack power, an overall conclusion is that financial performance did not improve after cross listing. Although the results are not significant, ROE and ROA means and medians are nominally lower after cross listing. Also note that there does not appear to be any significant reduction in leverage after cross listing. Again, we caution that this analysis is based on extremely small samples.
13 Foerster and Karolyi (2000) also provide evidence that while companies issuing equity in the U.S. on major exchanges modestly outperform their benchmarks (in contrast to under-performance of private placements), firms that come from emerging markets with low accounting standards (Mexico included) significantly outperform their benchmarks. We do not have enough data to separately analyze private placements versus public placements in terms of financial performance. However, our ADR sample is dominated by public placements (see table 6), and we believe our results for Mexico are consistent with Foerster and Karolyi (2000) results for emerging market companies in general.
21
Based on the limited before and after cross listing evidence in table 5 along with the results of the benchmarking comparison of the performance of ADR firms with non-ADR firms in table 4, and the prevalence of negative earnings after cross listing as noted above, we conclude that the financial performance results for Mexican ADR firms appear consistent with the results of prior global equity offering research; that is, there is no evidence that greater access to capital is associated with improved financial performance for Mexican firms after cross listing.
[Insert table 5 here]
4.4 Capital Issuance Analysis
As previously noted, Lins et al. (2000) provide evidence that non-U.S. firms issue equity on U.S. stock exchanges to relax capital constraints. Evidence of this constraint is the significantly higher financial leverage of Mexican cross-listed firms in comparison with non- cross-listed (NCL) firms. Further, in a dynamic capital structure environment, Palepu, Healy and Bernard (2000) and Myers and Majluf (1984) argue that the issuance of new equity represents last resort financing.
Table 6 summarizes debt and equity issues by Mexican firms in the U.S. broken down by our sample categories of firms.14 Several insights are evident from the table. First, much more capital is raised by level 2 and 3 ADR firms than by NCL firms. Level 2 or 3 ADR firms raised over 83% of total debt and almost 87% of total equity. Improved access to capital markets, at the expense of meeting SEC disclosure requirements, is clearly a benefit to cross listing as a level 2 or 3 ADR. Second, substantially more equity ($11.8 billion) has been issued than debt ($4.3 billion). The substantially higher equity issuance is consistent with the idea of accessing U.S. markets to loosen financing constraints. At the same time, the issuance of relatively less
22
debt in the U.S. may be indicative of the close ties Mexican firms have to their domestic bankers (LaPorta et al. 2002). Finally, we note the limited amount of capital that has been raised by OTC and 144a firms.
[Insert table 6 here]
4.5 Logistic Regression Results
Tables 3 through 6 provide evidence of individual differences in various financial and market characteristics among the ADR, OTC/144a and NCL firms. We now use a Logistic regression model to examine the conditional relation between the variables we identified above and the likelihood of cross listing.
Under the Logit model, the odds of a firm’s cross listing in the U.S. are described by the conditional ratio of: P(Gj/Xj)/(1-P(Gj/Xj)), where P(Gj/Xj) is the probability of being a member of G given Xj. The log-linear function of this probability can be expressed as follows (Maddala 1991):
()()()jkjojjLogXLogXAAPPLogk11A...1/+++=− (1)
where: Pj = probability of firm j cross listing and Xjk = the explanatory variables for firm j.
We test the classificatory power of a model that includes the following independent variables from table 4: MVES (size), ROA (profitability), LEVERAGE (capital structure), GROWTH, REVRATIO (ratio of domestic to total revenue) and INDUSTRY. The variables are chosen based on theoretical considerations and previous empirical research; for example, LEVERAGE is based on the capital constraint hypothesis (Lins et al. 2000), ROA is a profitability proxy based on prior global equity offering research (Foerster and Karolyi 2000) and MVES is a proxy for the firm’s size and its need for foreign capital based on Saudagaran
14 Data were obtained from the Bank of New York. While Level 2 ADRs may not raise new capital in the U.S. by 23
(1988). GROWTH is also related to the firm’s need for capital. We include REVRATIO (the ratio of domestic to total revenue) as proxy for firms’ exposure outside of their home country (Saudagaran 1990) and INDUSTRY based on our previous analysis. We do not use the full set of variables in table 4 so as to avoid over-fitting models to our relatively small samples.15
The first column of table 7 summarizes the results for the ADR/NCL comparison. First, the model exhibits reasonable classificatory power, as evidenced by the concordant percentage of 80.80%, in comparison with the naïve hypothesis that all sample firms are ADRs (128 out of 209, or 61.2%). All variables except ROA and GROWTH are significant. As expected, the variable with the highest classificatory power is LEVERAGE. REVRATIO was generally not significant in the univariate tests in table 4. However, it is highly significant here, but not in the hypothesized direction. MVES is also significant and there is some evidence of an industry effect.
The second column of table 7 reports the results of the logistic regression for a comparison of ADR and OTC/144a firms. The variables do not improve the classification compared to the naïve hypothesis that all sample firms are ADR firms (concordant percentage, 72.6%, naïve percentage 73.1%). Importantly, the LEVERAGE variable is not significant. This suggests that these two groups are very similar in terms of the characteristics that we investigate.
The third column of table 7 reports the results of a comparison of OTC/144a firms with NCL firms. Compared to a naïve hypothesis that all sample firms are OTC/144a firms (47/108 = 43.52%), the concordant percentage of 89.5% shows the model is very effective in distinguishing
issuing new shares, they are not prevented from raising equity capital through private issues.
15 We also performed stepwise logistic regressions using various combinations of variables from table 4 without any improvement in results.
24
between the two groups. Again, the most highly significant variable is LEVERAGE, followed by REVRATIO. All other variables add classificatory power except for GROWTH.
Overall, these results support our first and second hypotheses that ADR firms are more capital-constrained (levered) than NCL firms and they exhibit weaker financial performance after cross listing.
[Insert table 7 here]
4.6 Value-Relevance Hypothesis
The final analyses investigate whether there are differential associations between accounting information and stock prices for ADR firms, which, due to cross listing, may operate in a different information environment than NCL firms. Our tests are based on standard regressions of accounting book value (BVS) and earnings (EPS) on market value of equity (MVS) as in Amir et al. (1993). In table 8, we present both individual year and pooled (1995-1999) results for the ADR, OTC/144a and NCL samples.16
While we report results for all three groups, we primarily focus on the NCL and ADR comparisons here. Table 8 provides evidence of a significant difference in the explanatory power of the accounting variables for the two groups.17 The pooled adjusted R2 for the ADR sample in panel A is 0.71 and individual year R2s range from 0.55 to 0.88. In contrast, the adjusted R2 for the pooled NCL sample in panel C is essentially zero, and only in one year, 1995, is there evidence of explanatory power on par with the ADR group.18 Both BVS and EPS are significant in the pooled regression for the ADR group. BVS for the ADR group is also highly
16 Correlation analyses for both pooled samples indicate significant associations among all the accounting and market variables and no significant association with the time period (year).
17 We estimated F-statistics for comparing nested regression models and found that except in 1995, the adjusted R2 is significantly higher in the ADR sample estimation. 25
significant in all of the individual regressions, while EPS is only significant in 1997. The anomalous result for the NCL sample is the highly significant result for 1995 and the corresponding low explanatory power for the other four years.19
We also estimate pooled regression results for separate Level 2 ADR, Level 3 ADR, OTC and 144a samples (untabulated). The results are interesting. There is significant explanatory power (adjusted R2 of 0.63) for BVS for the 144a group; however, the result is based on only fourteen observations. For the OTC group, the adjusted R2 is 0.44, and both BVS and EPS are significant. We find the strongest results for the level 2 ADR group (adjusted R2 of 0.83) where both BVS and EPS are significant. Finally, the adjusted R2 for the level 3 ADRs is 0.65, but only BVS is significant.
These results support the third hypothesis that ADR (and OTC/144a) firms’ accounting information is more highly associated with stock prices than that of NCL firms.
[Insert table 8 here]
4.7 Sensitivity Analysis: Quality of Earnings
So far, we have provided consistent evidence that Mexican NCL firms are financially stronger than ADR firms, based on the measures we examine. However, Lang et al. (2001) find evidence of greater income smoothing behavior by NCL firms. The authors’ results are for data pooled over twenty-six countries. In the case of Mexico, there is also reason to expect greater smoothing behavior on the part of NCL firms due to the belief that domestic regulatory control may not be very effective.
18 Again, we caution that individual year regressions, particularly as reported in panels B and C, are for extremely small samples.
19 However, this anomaly results from a regression with only eight observations.
26
In table 9, we summarize the results of smoothing tests based on relations between the cash flow and accrual components of earnings. First, in panel A, note that the variability of accruals and cash flows from operations is always nominally greatest for the ADR firms in comparison with the other two groups. Particularly in the case of cash flow from operations, the greater variability may indicate that Mexican ADR firms are riskier than Mexican NCL firms. Ceteris paribus, the NCL firms should exhibit smoother earnings than the ADR firms. Thus, evidence indicating a difference in smoothing measures between the groups is not necessarily indicative of a difference in the quality of earnings.
In panel B of table 9 we present the results of correlation tests between accruals and cash flows for the two groups. As in Lang, Raedy and Yetman (2001), a more negative cash flow/accrual correlation is indicative of earnings management in the sense of using accruals to smooth variability in cash flows. We find that the correlation between cash flows and accruals is significantly more negative for ADR firms compared to NCL firms. The differences for the ADR versus OTC/144a and NCL versus OTC/144a comparisons are not significant.
In untabulated results, we also examine the frequency distributions of earnings to assess the frequency of small, positive earnings for the two groups as in Burgstahler and Dichev (1997). There is no evidence of any difference in the distribution of small positive earnings between the groups. This result is contrary to the results for Lang et al. (2001) as expressed in hypothesis four. Taken together, these results suggest that the differences in financial characteristics we observe are probably not an artifact of accounting.
In summary, our tests of various financial characteristics indicate that Mexican ADR firms are more highly levered and that they have raised significantly more capital in the U.S. than non-exchange listed firms. Further, the results indicate that, after cross listing, Mexican
27
ADR firms are financially weaker than NCL firms. While there is some (weak) evidence that reported earnings for NCL firms are smoother than reported earnings for ADR firms, it is more likely that such differences reflect risk differences (e.g., higher variability of cash flows for ADR firms) rather than differences in the prevalence of earnings management for the two groups.
Lang et al. (2001) suggest that firms may self-select into cross listing; stronger firms may use cross listing to credibly signal their strength and (weaker) non-crosslisted firms may find the SEC disclosure requirements an impediment to cross listing. Again, our results for Mexico are inconsistent with those of Lang et al. (2001) for their 26-country sample; our results are consistent with those found for IPO samples and for global equity offerings as in Foerster and Karolyi (2000).
[Insert table 9 here]
5. Conclusion
This paper examines the characteristics of Mexican companies that cross list their securities in the U.S. It has been suggested that the SEC disclosure requirements serve the purpose of restricting the access of weaker foreign companies to U.S. capital markets. However, we provide evidence that Mexican ADR firms exhibit weaker financial performance than non-crosslisted (NCL) firms. For example, we document a high percentage of negative earnings observations for ADR firms and relatively few negative earnings numbers for NCL firms. Additionally, our comparisons of Mexican ADR and NCL companies indicate that ADR firms are more highly levered than non-crosslisted (NCL) companies and we document the significant amount of equity capital they have raised in the U.S. in an attempt to loosen capital constraints. Logistic regression models using financial variables as inputs are highly significant in correctly
28
classifying ADR and NCL firms. The results indicate that the substantially higher leverage for ADR firms is an important financial characteristic that differentiates the two groups. Finally, we test the relative quality of financial statement numbers for ADR and NCL firms, and conclude that accounting quality differences do not explain the differences in financial performance.
We also hypothesized that financial disclosures for Mexican ADR firms will be more value-relevant than for NCL firms. Our regressions of share price on book value per share and earnings per share reveal significant explanatory power for the ADR firms and generally insignificant results for NCL firms. The results indicate that financial statements are value relevant for ADR investors, but not for investors in Mexican companies that do not cross list in the U.S. The results may indicate that the SEC regulations add credibility to the financial statements of ADR firms. In contrast, the lack of value-relevance of financial statements for NCL firms may be indicative of market inefficiency (e.g., insider trading) in the Mexican market as suggested by Bhattacharya et al. (2000).
The results from our research underscore the importance of focusing on a single country where market institutions and corporate governance mechanisms can be carefully considered. Our focus on a single country is also important due to the tendency of previous studies to generalize based on multi-country samples, particularly samples dominated by Australian, Canadian and U.K. firms. Of course, the benefits of our single country, Mexican market focus are offset somewhat by data limitations and constraints from working with relatively small samples.
29
REFERENCES
Amir, E., T. Harris, and E. Venuti. 1993. A comparison of the value-relevance of U.S. versus non-U.S. GAAP accounting measures using Form 20-F reconciliations. Journal of Accounting Research (Supplement) 31: 230-264.
Baker, H., J. Nofsinger and D. Weaver. 1999. International cross listing and visibility. NYSE Working Paper 99-01.
Bank of New York (BNY). 1997. Depositary Receipt Directory. New York, NY: BNY.
Bhattacharya, U., H. Daouk, B. Jorgenson, and C. Kehr. 2000. When an event is not an event: the curious case of an emerging market. Journal of Financial Economics (55): 69-101.
Bruner, R., S. Chaplinsky and L. Ramchand. 1999. Coming to America: A clinical study of IPOs in the U.S. by foreign firms. Working paper, University of Virginia.
Burgstahler, D. and I. Dichev. 1997. Earnings management to avoid earnings decreases and losses. Journal of Accounting & Economics (December):99-126.
Cochrane, L., J. Shapiro, and J. Tobin. 1996. Foreign equities and U.S. investors: Breaking down the barriers separating supply and demand. Stanford Journal of Law, Business & Finance. (Vol. 2.2): 241-263.
Davis-Friday, P. 2001. Equity valuation and current cost disclosures: The case of Mexico. Journal of International Financial Management and Accounting (Autumn): 260-285.
Davis-Friday, P. and J. Rivera. 2000. Inflation accounting and 20-F disclosures: Evidence from Mexico. Accounting Horizons (June):114-135.
Domowitz, I., J. Glen, and A. Madhavan. 1997. Market segmentation and stock prices: Evidence from an emerging market. Journal of Finance (July): 1059-1085.
_____ 1998. International cross listing and order flow migration: Evidence from an emerging market. Journal of Finance (December): 2001-2027.
Ely, K. and G. Pownall. 1999. The value relevance of consolidated vs. unconsolidated earnings and book values for Japanese firms: Evidence from cross-listed and non-crosslisted firms. Working paper, Emory University.
Foerster, S. and G. Karolyi. 2000. The Long-run performance of global equity offerings. Journal of Financial and Quantitative Analysis (December): 499-527.
Gordon, E. 2001. Accounting for changing prices: The value-relevance of historical cost, price level, and replacement cost accounting in Mexico. Journal of Accounting Research (June): 177-200.
30
LaPorta, R., F. Lopez-De-Silanes, A. Shleifer, and R. Vishny. 1998. Law and finance. Journal of Political Economy (106): 1113-1155.
_____. 1999. Corporate ownership around the world. The Journal of Finance (April): 471-517.
LaPorta, R., F. Lopez-de-Silanes, and G. Zamarripa. 2002. Related lending. Working paper, Harvard University, Yale University and National Banking and Securities Commission (Mexico).
Land, J. and M. Lang. 2002. Empirical evidence on the evolution of international earnings. Working Paper, University of North Carolina.
Lang, M., J. Smith Raedy, and M. Yetman. 2001. How representative are cross-listed firms? An analysis of firm and accounting quality. Working Paper, University of North Carolina.
Lins, K., D. Strickland, and M. Zenner. 2000. Do non-U.S. firms issue equity on U.S. stock exchanges to relax capital constraints? Working paper, University of Utah.
Loughran, T. and J. Ritter. 1995. The new issue puzzle. Journal of Finance (50): 23-51.
Maddala, G. S. 1991. A Perspective on the use of limited-dependent and qualitative variables models in accounting research. The Accounting Review (October): 788-808.
Miller, D. 1999. The market reaction to international cross-listings: Evidence from depositary receipts. Journal of Financial Economics (51): 103-123.
Myers, S. and N. Majluf. 1984. Corporate financing and investment decisions when firms have information that investors do not have. Journal of Financial Economics (June):187-222.
Pulatkonak, M. and G. Sofianos. 1999. The distribution of global trading in NYSE-listed non-U.S. stocks. NYSE Working Paper 99-03.
Palepu, K., V. Bernard and P. Healy. 2000. Business Analysis and Valuation. (South-Western College Publishing).
Rees, L. and P. Elgers. 1997. The market's valuation of non-reported accounting measures: Retrospective reconciliations of non-U.S. and U.S. GAAP. Journal of Accounting Research (Spring): 115-127.
Ritter, J. J. 1991. The long-run performance of initial public offerings. Journal of Finance (46): 3-27.
Saudagaran, S. 1988. An empirical study of selected factors influencing the decision to list on foreign stock exchanges. Journal of International Business Studies (Spring): 101-127.
_____. 1990. An investigation of selected factors influencing multiple listing and the choice of foreign stock exchanges. Advances in Financial Planning and Forecasting (Volume 4, Part B): 75-122.
31
Smith, K. and G. Sofianos. 1997. The impact of an NYSE listing on the global trading of non-U.S. stocks. NYSE Working Paper 97-02.
Spiess, D. and J. Affleck-Graves. 1995. Under-performance in long-run stock returns following seasoned equity offerings. Journal of Financial Economics (38): 243-267.
Stulz, R. 1999. Globalization, corporate finance, and the cost of capital. Journal of Applied Corporate Finance. (47): 1425-1460.
32
TABLE 1
Sample Selection Procedures
1995
1996
1997
1998
1999
Number of Companies Traded on BMV
185
193
198
194
188
Number of Companies Eligible for NYSE
37
51
59
64
60
Number of Companies Traded on U.S. Exchanges1
25
28
30
33
34
Less firms missing data on Economatica or Bolsa
(2)
(2)
(2)
(2)
(2)
ADR Sample
23
26
28
31
32
Eligible Companies not Traded on NYSE
14
23
29
32
26
Companies Traded OTC2
3
5
8
13
8
Less Banks
(3)
Companies Traded Rule via 144a2
1
2
3
4
4
Eligible Companies not Traded in U.S.
10
16
11
15
14
Non-ADR Sample
14
23
29
29
26
1Teléfonos de Mexico is traded both on the NASDAQ (A shares) and NYSE (L shares). It is only included once in the NYSE totals.
2 Several of the companies traded on U.S. exchanges also have shares traded OTC or listed under Rule 144a. The number of OTC and 144a listings does not include firms that are also listed on ASE, NASDAQ, or NYSE.
33
TABLE 2
Sample Firms
Panel A: Mexican Firms Traded on U.S. Exchanges through December 1999
Company Name
Effective Date
Exchange
Level
Sector
95
96
97
98
99
Altos Hornos de Mexico2
NYSE
2
Steel
X
X
X
X
Biper
NASDAQ
2
Communications
X
X
Bufete Industrial2
NYSE
3
Construction
X
X
X
X
Cemex3
NYSE
2
Cement Production
X
Coca-Cola FEMSA
NYSE
3
Beverages
X
X
X
X
X
Consorcio G Grupo Dina2
NYSE
3
Transportation
X
X
X
X
X
Controladora Comercial Mexicana
NYSE
3
Retail
X
X
X
X
X
Desc
NYSE
3
Holding Company
X
X
X
X
X
Empresas ICA
NYSE
3
Construction
X
X
X
X
X
Fomento Economico Mexico (FEMSA)3
NYSE
2
Beverages
X
X
Gruma3
NYSE
2
Food
X
X
Grupo Casa Autrey (Saba)
NYSE
3
Food
X
X
X
X
X
Grupo Elektra
NYSE
2
Retail
X
X
X
X
X
Grupo Imsa
NYSE
3
Steel
X
X
X
X
Grupo Industrial Durango
NYSE
3
Paper Products
X
X
X
X
X
Grupo Industrial Maseca1
NYSE
2
Food
X
X
X
X
X
Grupo Iusacell
NYSE
2
Communications
X
X
X
X
X
Grupo Mexicano de Desarrollo11,2
NYSE
3
Construction
X
X
X
X
X
Grupo Radio Centro
NYSE
3
Communications
X
X
X
X
X
Grupo Sidek1,2
NYSE
3
Holding Company
X
X
X
X
X
Grupo Simec
AMEX
3
Steel
X
X
X
X
X
Grupo Televisa
NYSE
3
Communications
X
X
X
X
X
Grupo Tribasa
NYSE
3
Construction
X
X
X
X
X
Industrias Bachoco
NYSE
3
Food
X
X
X
Internacional de Ceramica1
NYSE
3
Construction
X
X
X
X
X
Pepsi-Gemex
NYSE
2
Beverages
X
X
X
X
X
Savia (Empresas la Moderna)
NYSE
3
Holding
X
X
X
X
X
Teléfonos de Mexico (A)
NASDAQ
2
Communications
Teléfonos de Mexico (L)
NYSE
3
X
X
X
X
X
Transportación Marítima Mexicana
NYSE
3
Transportation
X
X
X
X
X
Tubos de Acero de Mexico
AMEX
2
Steel
X
X
X
X
X
TV Azteca
NYSE
3
Communications
X
X
X
Vitro
NYSE
3
Mineral Products
X
X
X
X
X
TOTAL
23
26
28
31
32
1These firms trade shares on an organized exchange as well as over the counter (OTC). In the analysis that follows, they are only included in the “exchange” firm analysis.
2These firms delist after the test period. They are included in the analysis during the time period when they are listed on exchanges.
3These firms are categorized as exchange-listed firms after they cross list in the U.S. and they are also included in the non-crosslisted, but eligible sample during the years before they cross list. 34
TABLE 2 (Continued)
Sample Firms
Panel B: Mexican Firms Eligible to List on the New York Stock Exchange
Company Name
Exchange
Effective Date
144a
OTC
Amex
Nasdaq
NYSE
Sector
95
96
97
98
99
Alfa
X
Holding Company
X
X
X
X
X
Apasco
X
Cement
X
X
X
X
X
Ara Consorcio
X
Construction
X
Argos Embotelladora
Beverage
X
Bimbo Grupo
Food
X
X
X
X
X
Carso Global Telecom
X
Communications
X
Cementos Chihuahua
Cement
X
X
Cemex
X
Cement
X
X
X
X
Cintra
Airlines
X
X
X
X
Coppel
Retail
X
X
Corp Moctezuma
Cement
X
Cydsa
Holding Company
X
X
El Puerto de Liverpool
Retail
X
X
X
X
X
Empaques Ponderosa
X
Paper Products
X
X
X
Fomento Economico Mexico
X
Beverage
X
X
Grupo Azucarero Mexico
Food
X
Grupo Carso (96-99)
X
Holding Company
X
X
X
X
X
Grupo Continental
X
Beverage
X
X
X
X
X
Grupo Gigante
Retail
X
X
X
X
Grupo Industrial Saltillo
Holding Company
X
X
X
X
Grupo Maseca (Gruma)
X
Food
X
X
X
Grupo Modelo
Beverage
X
X
X
X
X
Grupo Palacio de Hierro
Retail
X
Grupo Sanborns
Retail
X
X
X
Hylsamex (96-99)
X
Steel Plant
X
X
X
X
Industrias CH
Metallurgy
X
Jugos del Valle1
X
Beverage
X
X
Kimberly Clark de Mexico
X
Paper Products
X
X
X
X
X
Mexichem
PetroChemical
X
X
X
X
Nuevo Grupo Mexico
Mining
X
X
Nutrisa
Retail
X
Panamco
Beverage
X
Penoles Industrias
Mining
X
X
X
X
X
Posadas Grupo
X
Tourism/Hotel
X
X
X
San Luis Corp
X
Automotive Parts
X
X
X
X
Sears Roebuck
Retail
X
X
Seguros Com America
X
Insurance
X
Situr Grupo
X
Tourism/Hotel
X
X
Soriana Organizacion
Retail
X
X
X
X
X
Wal Mart de Mexico (97-99)
X
Retail
X
X
X
X
X
TOTAL
5
10
3
14
23
29
29
26
35
TABLE 3
Industry Distribution of the Exchange-Listed ADR Firm and Eligible Firm Samples
Sector Name*
Exchange Listed Firms
Eligible Firms
#
%
FirmYears
%
#
%
FirmYears
%
1
Airline
0
0.00
0
0.0
1
2.5
4
3.31
2
Automotive Parts
0
0.00
0
0.0
1
2.5
4
3.31
3
Beverages
3
9.37
12
8.6
6
12.5
16
13.22
4
Cement
1
3.13
1
0.7
4
10.0
12
9.91
5
Construction
5
15.62
24
17.1
1
2.5
1
.83
6
Communications
6
18.75
25
17.9
1
2.5
1
.83
7
Food
4
12.50
15
10.7
3
7.5
9
7.43
8
Holding Company
3
9.37
15
10.7
4
10.0
16
13.22
9
Insurance
0
0.00
0
0.0
1
2.5
1
.83
10
Metallurgy
0
0.00
0
0.0
1
2.5
1
.83
11
Mineral Products
1
3.13
5
3.6
0
0.0
0
0.0
12
Mining
0
0.00
0
0.0
2
5.0
7
5.78
13
Paper Products
1
3.13
5
3.6
2
5.0
8
6.61
14
PetroChemicals
0
0.00
0
0.0
1
2.5
4
3.31
15
Retail
2
6.25
10
7.1
9
22.5
28
23.14
16
Steel
4
12.50
18
12.9
1
2.5
4
3.31
17
Tourism/Hotel
0
0.00
0
0.0
2
5.0
5
4.13
18
Transportation
(Machinery & Equipment)
2
6.25
10
7.1
0
0.0
0
0.0
Total
32
100.0
140
100.0
40
100.0
121
100.0
*The sector names are from Economatica.
36
TABLE 4
Descriptive Statistics (1995-1999)
Panel A: Financial Statement and Market Variables
Exchange Listed ADR Firmsa
OTC & 144a Firmsa
Non-Crosslisted Firmsa
Variable
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
Assets
135
2,105
3,040
1,172
48
2,477
2,810
1,634
73
1,936
2,094
1,389
Liabilities
135
997
1,071
682
48
1,151
1,293
600
73
786
1,142
467
Equity
135
1,027
2,098
488
48
1,021
855
983
73
975
830
742
Net Income
135
113
378
39
48
153
181
73
73
148
180
97
Operating
Cash Flow
135
227
663
81
47
265
500
147
73
233
296
161
Investing
Cash Flow
135
-143
339
-63
47
-220
461
-96
72
-192
260
-99
MVE
135
7,755
16,418
3,263
48
16,302
21,216
7,025
73
9,500
9,429
6,253
Relative Size
135
0.075
0.167
0.030
48
0.15
0.19
0.07
73
0.092
0.097
0.055
Beta
67
0.653
0.404
0.670
31
0.85
0.30
0.93
42
0.660
0.315
0.629
Returns
29
1.082
2.736
0.101
26
1.19
3.07
0.19
21
1.336
3.028
0.313
Analysts
91
12.396
6.031
13.000
37
16.51
6.48
19.00
42
13.214
7.380
12.500
a The data, except ratios and percentages, are totals in terms of millions of U.S. dollars.
37
TABLE 4 (Continued)
Descriptive Statistics
Panel B: Performance Measures
Exchange Listed ADR Firmsa
OTC & 144a Firmsa
Non-Crosslisted Firmsa
Variable
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
ROEa
135
-0.105
1.645
0.101
48
0.137
0.162
0.162
73
0.147
0.078
0.132
Leverage
135
3.724
11.042
2.372
48
2.599
1.745
2.090
73
1.952
0.945
1.616
ROA
135
0.037
0.086
0.042
48
0.075
0.058
0.078
73
0.081
0.040
0.074
Asset Turnover
135
0.697
0.482
0.604
48
0.643
0.292
0.602
73
0.865
0.365
0.777
Profit Margin
135
0.042
0.222
0.060
48
0.112
0.132
0.125
73
0.111
0.078
0.096
Growth (TA)
135
0.344
0.476
0.291
47
0.635
1.466
0.379
61
0.342
0.512
0.319
EPS
135
0.117
0.683
0.041
48
0.254
0.288
0.155
72
0.243
0.281
0.139
BVS
135
1.886
2.938
0.630
48
1.745
1.455
0.884
72
1.639
1.427
1.012
PE
88
19.895
24.293
13.760
38
15.339
10.212
13.355
55
20.250
25.316
15.580
PB
125
1.864
1.768
1.410
44
2.217
1.442
1.785
61
1.974
0.876
1.910
Revenue Ratio
135
0.767
0.233
0.822
48
0.838
0.216
0.907
72
0.772
0.225
0.815
aThere are 36 negative ROE observations in the ADR sample, three in the OTC & 144a sample and one in the NCL sample.
38
TABLE 4 (Continued)
Panel C: T and Z Statistics from Pairwise Comparisons of Mean and Median Differences
Non-Crosslisted Firms vs. Exchange Listed ADRsa
OTC & 144a Firms vs. Exchange Listed ADRsa
Non-Crosslisted Firms vs. OTC & 144 A Firmsa
Panel A Variable
Mean
Median
Mean
Median
Mean
Median
Assets
-1.48
0.58
-0.52
1.27
-0.84
0.90
Liabilities
**-2.76
##-2.35
-1.23
0.09
-1.38
1.58
Equity
-0.25
##2.63
-0.03
##2.03
-0.29
0.00
Net Income
0.90
###4.66
0.95
###3.39
-0.14
-0.20
Operating Cash Flow
1.53
###3.46
1.22
###2.81
-0.40
0.18
Investing Cash Flow
-0.52
###-3.10
-1.25
##-2.41
0.39
0.61
MVE
0.29
#1.87
**2.51
###4.35
**-2.21
##2.77
Relative Size
0.89
##2.13
**2.36
###3.25
*-1.91
1.23
Beta
0.10
-0.02
**2.74
##2.66
**-2.67
###2.97
Returns
0.30
0.80
0.14
0.63
0.16
0.57
Analysts
0.63
0.50
***3.32
###3.46
**-2.12
##2.12
Panel B Variable
Mean
Median
Mean
Median
Mean
Median
ROE
*1.78
###3.53
*1.69
###3.24
0.41
0.72
Leverage
*-1.85
###-4.54
-1.14
-1.23
**-2.35
##2.01
ROA
***5.00
###4.62
***3.37
###3.47
0.65
-0.09
Asset Turnover
***2.82
###3.90
-0.92
0.15
***3.71
###-3.33
Profit Margin
***3.23
###3.45
**2.57
###3.46
-0.04
1.26
Growth (TA)
-0.03
-0.09
1.33
#1.88
-1.31
#1.83
EPS
*1.86
###5.85
*1.90
###4.74
-0.21
0.25
BVS
-0.81
###3.29
-0.43
###2.83
-0.39
-0.14
PE
0.08
0.77
-1.48
-0.48
1.29
1.25
PB
0.57
##2.23
1.31
#1.93
-0.99
0.19
Revenue Ratio
0.16
0.60
*1.94
1.63
-1.63
0.77
a***,**,*(###,##,#) The test of the mean (median) difference is statistically significant at the 1%, 5%, or 10% level, respectively, two-tailed, based on a paired t-test (Wilcoxon Rank sum test).
39
TABLE 4 (Continued)
Variable Definitions
Assets
Book value of total assets;
Liabilities
Book value of total liabilities;
Equity
Book value of common equity;
Net Income
Total net income;
Operating Cash Flow
Net cash from operating activities;
Investing Cash Flow
Net cash from investing activities;
MVE
December 31 market value of common equity per share times the number of outstanding shares;
Relative Size
The ratio of the firm’s market value of equity in year t to the total Bolsa market capitalization in the same year.
Beta
The coefficient of variation between the firm’s stock price and the market as reported by the Bolsa;
Returns
Annual percentage change in the market value of common equity;
Analysts
The number of analysts providing estimates in the I/B/E/S international database;
ROE
Return on common equity (net income/equity);
Leverage
Total assets/Equity;
ROA
Return on assets (net income/assets);
Asset Turnover
Sales/Assets;
Profit Margin
Net Income/Sales;
Growth (TA)
The 36-month change in total assets;
EPS
Earnings per share (net income/outstanding common shares);
BVS
Book value per share (equity/outstanding common shares);
PE
Price-earnings ratio (market value of common equity per share/earnings per share);
PB
Price-to-book ratio (market value of common equity per share/book value of common equity per share);
Revenue Ratio
The ratio of domestic revenue to total revenue.
40
TABLE 5
Descriptive Statistics Before and After Cross Listing
Panel A: Financial Statement and Market Variables- Exchange Listed ADR Firmsa
Before
After
Difference Tests
Variablea
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
Tb
(p-value)
Zb
(p-value)
Assets
26
2.896
4.288
1.178
30
3.071
4.492
1.501
-0.15
(0.88)
-0.86
(0.39)
Liabilities
26
1.133
1.524
0.580
30
1.430
1.792
0.842
-0.66
(0.51)
-0.83
(0.41)
Equity
26
1.697
2.957
0.547
30
1.674
2.954
0.490
0.03
(0.98)
-0.42
(0.68)
Net Income
26
0.269
0.536
0.063
30
0.192
0.463
0.050
0.57
(0.57)
1.19
(0.23)
Operating Cash Flow
26
0.367
0.858
0.004
30
0.289
0.924
0.020
0.33
(0.75)
0.06
(0.95))
Investing Cash Flow
26
-0.323
0.543
-0.139
30
-0.318
0.765
-0.46
-0.03
(0.98)
-0.45
(0.65)
a The data are totals in terms of millions of U.S. dollars.
41
TABLE 5 (Continued)
Descriptive Statistics Before and After Cross Listing
Panel B: Performance Measures - Exchange Listed ADR Firms
Before
After
Difference Tests
Variable
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
T
(p-value)a
Z
(p-value)a
ROE
26
12.40
11.25
10.89
30
2.56
22.43
9.96
2.03
*(0.05)
1.08
(0.28)
Assets/Equity
26
1.97
1.19
1.97
30
2.36
1.56
1.83
-1.04
(0.30)
-0.43
(0.67)
ROA
26
5.63
3.88
5.35
30
3.65
7.01
3.98
1.28
(0.21)
1.17
(0.24)
Sales/Assets
26
0.44
0.75
0.17
30
0.88
1.17
0.45
-1.65
*(0.10)
-1.17
(0.24)
Net Income
/Sales
19
5.49
0.22
0.24
25
0.61
1.44
0.10
1.55
(0.13)
1.80
#(0.07)
Revenue Ratio
26
0.83
0.22
0.92
30
0.79
0.27
0.86
0.75
(0.46)
0.77
(0.44)
aRefer to table 4 for variable definitions.
b*(#) The test of the mean (median) difference between Non-ADR and exchanged-listed firms is statistically significant at the 10% level or better, two-tailed, based on a paired t-test (Wilcoxon Rank sum test).
42
TABLE 6
Debt and Equity Issuance
Debt
Equity
Classification
No. of Issues
Amount ($mil)
No. of Issues
Amount ($mil)
NCL
Total (number and amount)
3
315.2
14
4579.6
Average
105.1
327.1
Median
35.3
117.8
Standard Deviation
129.7
554.1
144a
Total (number and amount)
0
0.0
3
54.9
Average
18.3
Median
13.8
Standard Deviation
8.5
OTC
Total (number and amount)
1
165.9
4
215.9
Average
165.9
54.0
Median
165.9
41.6
Standard Deviation
49.4
Level 2 ADR
Total (number and amount)
0
0.0
13
1450.2
Average
111.6
Median
119.4
Standard Deviation
68.0
Level 3 ADR
Total (number and amount)
8
3603.9
29
8802.3
Average
450.5
303.5
Median
160.7
85.0
Standard Deviation
671.2
441.1
43
TABLE 7
Estimate of U.S. Cross Listing Probability during 1995-1999
Variable (Coefficient)
Exchange-Listed ADR Firms (n=128)
vs
Non-Crosslisted Firms (n=61)
Exchange-Listed ADR Firms (n=128)
vs
OTC and 144a Firms (n=47)
OTC and 144a Firms (n=47)
vs
Non-Crosslisted Firms (n=61)
Coefficient(Chi-square)
Coefficient (Chi-square)
Coefficient (Chi-square)
Intercept (α0)
-4.98 (11.97)***
3.31 (10.56)
-17.02 (19.75)***
MVES (β1)
0.02 (5.82)**
-0.02 (3.66)*
0.06 (8.05)***
ROA (β2)
-0.97 (0.09)
-6.20 (5.51)**
12.02 (2.73)*
LEVERAGE (β3)
1.75 (22.67)***
0.00 (0.00)
3.46 (17.30)***
GROWTH (β4)
-0.00 (0.00)
-0.00 (2.15)
0.00 (0.37)
REVRATIO (β5)
3.29 (9.52)***
-1.93 (3.74)*
11.25 (16.15)***
INDUSTRY (β6)
-0.07 (3.96)**
0.01 (0.05)
-0.14 (5.88)**
Model Statistics
Concordant Percent
80.80
72.60
89.00
Likelihood Ratio χ2 (p-value)
54.00 (<0.00)
19.90 (0.00)
60.93 (<0.00)
Pseudo R2
24.85%
10.75%
43.12%
Statistically significant coefficients (two-tailed test) are denoted by: *** less than .01 ** less than .05 * less than .10.
There are 32 firms listed on the NYSE, NASDAQ, or AMEX, ten firms traded over-the-counter, five firms traded via rule 144a and 25 firms that are not cross listed in the U.S. during the 1995-1999 period, but are eligible to list on the NYSE. P(ADR) ≡ Probability of cross listing in the U.S.; MVES is the market value of equity per share; ROA is the ratio of net income to total assets; LEVERAGE is the ratio of total assets to stockholders’ equity; GROWTH is the 36-month percentage change in total assets; REVRATIO is the ratio of domestic revenue to total revenue; and INDUSTRY is a dummy variable taking on the value of one when it is equal to category n (defined in table 3) and zero otherwise.
44
TABLE 8
Relation between Annual Stock Prices and Summary Financial Statement Measuresa
ititBVSitEPSitMVSεγγγ+++=210
Variable
Coefficient
Pooled
1995
1996
1997
1998
1999
Panel A: ADR Sample
Obs
132
21
22
24
28
29
Intercept
γ0
2.69
**(2.42)
3.91
*(1.76)
3.84
*(1.72)
1.52
(1.19)
2.41
(1.47)
1.11
(0.55)
EPS
γ1
2.71
*(1.66)
12.63
(1.39)
5.28
(0.93)
4.40
**(2.98)
3.01
(0.62)
-9.80
(-1.41)
BVS
γ2
5.98
***(15.49)
4.39
***(4.07)
3.94
**(2.86)
8.28
***(12.44)
4.54
***(5.95)
7.53
***(5.28)
Adjusted R2
0.71
0.69
0.60
0.88
0.55
0.64
Panel B: OTC & 144a Sample
Obs
47
2
6
10
13
12
Intercept
γ0
9.11
***(3.96)
NA
16.21
*(2.27)
9.78
**(2.65)
10.39
**(2.79)
3.34
(0.97)
EPS
γ1
30.45
***(4.13)
NA
131.43
(1.26)
67.09
***(4.07)
50.02
**(2.83)
66.81
***(3.96)
BVS
γ2
3.66
**(2.61)
NA
-29.10
(-1.09)
-3.42
(-0.97)
1.29
(0.62)
-1.63
(0.72)
Adjusted R2
0.61
NA
0.24
0.83
0.47
0.83
Panel C: Non-Crosslisted Sample
Obs
71
8
16
16
14
13
Intercept
γ0
13.49
***(5.88)
-1.94
(-1.02)
15.41
**(2.52)
12.36
**(2.40)
13.19
**(2.12)
19.35
**(3.43)
EPS
γ1
-2.76
(-0.36)
12.75
(1.43)
2.79
(0.16)
-25.90
(-1.38)
78.12
(1.12)
29.24
(0.53)
BVS
γ2
0.74
(0.48)
12.57
**(3.99)
-0.35
(-0.13)
4.52
(1.19)
-14.56
(-0.94)
-6.15
(-0.83)
Adjusted R2
NA
0.91
NA
NA
0.05
NA
aEach estimated coefficient is expected to be positive. MVS ≡ December 31 price, BVS ≡ Book value per share, and EPS ≡ Net income per share. T-statistics are in parentheses. Statistically significant coefficients (two-tailed test) are denoted by: *** less than .01 ** less than .05 * less than .10.
45
TABLE 9
Quality of Earnings Analysis
Panel A: Descriptive Statisticsa
Non-Crosslisted Firms
OTC & 144a Firms
Exchange Listed ADR Firms
Variable
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
Accruals
73
-0.037
0.054
-0.040
47
-0.020
0.076
-0.033
135
-0.043
0.102
-0.035
Cash Flows from Ops
73
0.118
0.052
0.130
47
0.095
0.080
0.110
135
0.080
0.111
0.078
Panel B: Spearman Correlations among Cash Flows and Accruals (1995-1999)a
Group
Coefficient
Z-statistic
(p-value)
Non-Crosslisted Firms
-0.75
Exchange-Listed Firms
-0.54
-1.92 (0.05)
OTC & 144a Firms
-0.67
Exchange-Listed Firms
-0.54
0.24 (0.81)
Non-Crosslisted Firms
-0.75
OTC & 144a Firms
-0.67
1.37 (0.17)
Variable Definitions
Accruals
The difference between net income and cash flows from operations scaled by total assets
Cash flows from operations
Net operating cash flows scaled by total assets 46
Evidence from Mexican Cross Listings
Paquita Y. Davis-Friday*
University of Notre Dame
386 Mendoza College of Business
Department of Accountancy
Notre Dame, IN 46556
pfriday@nd.edu
and
Thomas J. Frecka
University of Notre Dame
305B Mendoza College of Business
Department of Accountancy
Notre Dame, IN 46556
tfrecka@nd.edu
December 2002
*Corresponding Author. The authors appreciate comments from Donal Byard, Tim Loughran, Darius Miller, workshop participants at the University of Cincinnati, the University of Notre Dame, and Rutgers University, and participants at the 2000 AAA International Accounting Section Mid-Year Meeting and the 2001 Annual Congress of the European Accounting Association. The authors are also grateful for financial support from PriceWaterhouseCoopers and Arthur Andersen.
Which Firms Choose to List as American Depositary Receipts (ADRs)?:
Evidence from Mexican Cross Listings
ABSTRACT: This study provides evidence that Mexican firms that choose to trade in the U.S. as exchange-listed American Depositary Receipts (ADRs) have significantly weaker financial performance subsequent to cross-listing than Mexican firms that are eligible to list in the U.S., but do not. We show that on average, Mexican cross-listed firms are smaller, more highly levered and less profitable than non-crosslisted firms (NCL). While supplemental tests of earnings quality suggest that NCL firms exhibit significantly smoother earnings, that evidence is not sufficient to explain the stronger reported financial performance of those firms relative to cross-listed firms. Finally, our tests of value relevance, using book value and earnings to explain price, show significantly higher explanatory power for the cross-listed firms. The value-relevance results suggest that benefits from U.S. regulation exist for investors in Mexican cross-listed firms and that reported market inefficiency in Mexico may result in low demand for the financial statements of NCL firms.
Key Words: American Depositary Receipt; Mexico; Cross listing
Data Availability: Data used in this study are available from public sources, as indicated in the text.
JEL classification: E31; G15; M41; N26
1. Introduction
This paper investigates whether there are systematic differences between Mexican firms that choose to trade in the U.S. as American Depositary Receipts (ADRs) and Mexican firms that are eligible to trade in the U.S., but do not. Specifically, we examine the roles of country-specific institutions and firm-specific factors in explaining the cross-listing decision for firms from an emerging market with close ties to the U.S. This question is especially important in light of recent research findings in Lang, Raedy and Yetman (2001). Their results indicate that cross-listed firms perform better, are less risky and less aggressive in terms of earnings management. They also find that cross-listed firms report accounting data that are more strongly associated with share prices. However, their evidence is based on samples from 26 countries representing both developed and emerging markets and differing corporate governance environments. LaPorta et al. (1998) suggest that country-specific legal rules regarding the protection of shareholders and creditors, and the quality of enforcement of those rules, help define the rights that various securities afford their owners. These rights in turn determine whether or not firms can raise external finance (LaPorta et al. 1998, 1114). Therefore, country-specific legal rules may create differing incentives for firms to seek financing outside of their home countries and as important, investors may respond differently to the capital raising efforts of firms from different countries. The law and the quality of its enforcement are likely especially important factors for firms from emerging markets that seek financing outside of their home countries (Lins, Strickland and Zenner 2000, 4).
We consider the implications of the LaPorta et al. (1998) study for firms cross listing in the U.S. by investigating firms from Mexico, the United States’ second largest trade partner and a country from which the majority of eligible firms have cross listed on U.S. exchanges. We 1
focus on an important Mexican institution, the use of segmented share classes to prevent foreign control of Mexican corporations, as well as other country and firm-specific characteristics that may cause the type of cross-listed firm from Mexico to differ from those from other countries. Evidence from Foerster and Karolyi (2000) suggests that companies cross listing on major U.S. exchanges from countries with significant investment barriers for foreigners, like those that exist in Mexico, outperform their benchmarks while those from the same countries that issue equity via Rule 144a private placements significantly under perform.
Our results indicate that, on average, Mexican cross-listed firms are smaller, more highly levered and less profitable than the non-crosslisted (NCL) firms. While supplemental tests of earnings quality suggest that NCL firms exhibit smoother earnings, the evidence is not sufficient to explain the stronger reported financial performance of Mexican NCL firms. Finally, our tests of value relevance, using book value and earnings to explain price, show significantly higher explanatory power for the ADR firms and generally non-significant explanatory power for the NCL firms. The value-relevance results may indicate that investors in Mexican ADR firms benefit from U.S. regulation and that evidence of market inefficiency in Mexico (Bhattacharya et al. 2000) may result in low demand for financial statements of NCL firms. This study has the advantage of focusing on a single, emerging-market economy in contrast to most previous ADR research that uses multi-country samples dominated by developed market countries. It is also one of the first studies to deal with selection bias issues by comparing cross-listed and non-crosslisted firms. These advantages are obtained at the cost of having to base tests on and draw conclusions from relatively small samples.
The rest of the paper is organized as follows. Section 2 reviews prior research and provides additional background information on Mexican institutions and the ways Mexican
2
companies have accessed U.S capital markets. Section 3 develops our hypotheses and describes the research design to test the hypotheses. Section 4 describes our data sources, the composition of our samples, and presents the results of our analyses. Section 5 concludes the paper.
2. Background and Prior Research
2.1 Accessing U.S. Capital Markets
Most non-U.S. exchange-listed companies trade in the form of American Depositary Receipts (ADRs). ADRs are dollar-denominated derivative instruments. A depositary bank (primarily the Bank of New York in the case of Mexico) obtains claims on foreign securities and issues dollar-denominated claims against them. The ADRs pay U.S. dollar dividends and are processed through the U.S. clearance and settlement system.1 A company that wishes to list and trade securities in the secondary market will choose either the Level 1 or Level 2 ADR. The advantages of Level 2 listing are the availability of price quotes and the greater visibility associated with exchange listing. The primary disadvantage is the need to meet U.S. disclosure requirements, including reconciliation to U.S. GAAP.
In order to raise new capital in the U.S., the non-U.S. company must choose a Level 3 ADR and meet the standard SEC requirements for issuing securities, such as issuing an F-1 Registration Statement, providing annual 20-F filings, and reconciling financial statements to U.S. GAAP, or choose a semi-private offering via Rule 144a. Rule 144a allows privately placed securities to be sold to qualified institutional buyers (e.g., large insurance companies, banks and investment companies). Such sales are exempt from the SEC registration requirements.
1 Pulatkonak and Sofianos (1999) note that NYSE-listed non-U.S. companies trade in several forms including ADRs, Global Depositary Receipts, home-market "ordinaries" in the case of Canada, and New York Registered Shares. The instruments differ in the degree of fungibility with the corresponding home market securities. For
3
However, such issuers must provide buyers with information about the nature of the business and recent audited financial statements, generally prepared under home country GAAP. Furthermore, Rule 144a offerings are restricted to smaller companies since registration with the SEC is generally required if the assets of the foreign issuer exceed $10 million or if it has 500 or more equity holders. We investigate firms that have chosen to raise capital across all of these dimensions as well as firms that are eligible to trade on U.S. exchanges, but have not cross listed.
2.2 Evidence Regarding Cross Listing
Saudaragan (1988, 1990) argues that the potential benefits of foreign listing relate to financial, marketing, political and employee motivation considerations. The primary financial considerations are access to larger capital markets, reduced dependence on domestic sources of capital, avoidance of a decline in stock prices due to saturation of domestic markets, access to foreign capital to finance foreign expansion, and the opportunity to meet operating and long-term debt obligations abroad. Market considerations relate to greater visibility and product identification abroad and enhanced corporate image. Political motivations include improved international relations and employee motivation considerations include improved labor relations and the use of stock in employee compensation plans abroad.
Baker, Nofsinger and Weaver (1990) investigate changes in visibility associated with cross listing, using analyst coverage and media attention as proxies. Their results from analyzing changes in analyst coverage are consistent with the “investor recognition” hypothesis of Merton (1987), which indicates that an increase in the size of the investment base lowers expected return, lowers the cost of capital and increases the market value of shares.
instance, Canadian ordinaries are fully fungible because a trade made in the U.S. will be denominated in U.S. dollars while the same trade made in Canada would be denominated in Canadian dollars.
4
We do not explore visibility issues in this paper, but instead focus on access to capital as the primary firm benefit from cross listing in the U.S. Limited access to external capital is a primary motive for foreign listing in Bruner, Chaplinsky and Ramchand (1999), Miller (1999), and Stulz (1999). Later, we provide evidence of substantial capital raising for our exchange-listed ADR firm sample.
Lins, Strickland and Zenner (2000) empirically investigate whether cross listing in the U.S. reduces the costs of market segmentation by improving access to capital.2 Using an ADR sample covering the period 1986-1996 and data from Worldscope, the authors document that the sensitivity of investment to cash flow decreases significantly for firms from emerging markets but does not change for developed market firms following cross listing and that firms access international capital markets more often following cross listing. They suggest that this may be the case because emerging markets are often segmented. Market segmentation may arise from direct (e.g. ownership restrictions) and indirect barriers (e.g. information production, accounting standards). These barriers impede the flow and formation of capital in non-U.S. markets (Lins et al. 2000, 1). Lins et al. (2000) argue that relative to their home markets, greater disclosure requirements, shareholders rights protection, liquidity, and analyst following reduce the information asymmetry for non-U.S. firms listed on U.S. equity markets, especially those from emerging markets (Lins et al. 2000, 2).
2 We borrow the following quote from Lins et al. (2000) because it is so relevant for our study: “Stock markets from Mexico City to Sao Paulo have sunk in importance recently with trading shifting to Latin stocks listed in New York as American Depositary Receipts, or ADRs. … The focus abroad means the only Latin American companies that can raise financing nowadays are those that have access to international markets. Thousands of small and medium-sized businesses—the bulk of Latin America’s productive capacity—are left behind.” (“Latin American Stock Trading Migrates North,” Wall Street Journal, October 27, 1999).
5
2.2 The Focus on Mexico
Over the past decade there has been a sizable amount of trading of Latin American securities within U.S. markets as well as capital issuance using ADRs and Rule 144a listings. For example, in 1995, ten (six) of the twenty-five most actively traded foreign stocks in the U.S. in terms of share volume (dollar volume) were issued by companies domiciled in Mexico, Brazil, Argentina, and Chile.3 Currently, of the 458 non-U.S. companies listed on the NYSE, 95 are from Latin America and 29 are from Mexico.4
Mexico offers a particularly rich environment for conducting research on corporate finance choices as it differs substantially from that for U.S. firms. First, the efficiency of the market is open to question based on recent evidence in Bhattacharya et al. (2000) that Mexican restrictions on insider trading are not being enforced. Second, Mexican accounting rules and investor protections are rooted in the civil law tradition of France in contrast with the common law roots of the U.S. (LaPorta et al. 1998). LaPorta et al. (1998) find that countries from the French civil law family generally provide the worst legal protections to shareholders. Third, Mexican companies use inflation accounting in their domestic financial statements and in reconciling to U.S. GAAP, they are not required to change the measurement model, thus reducing the cost of reconciliation to U.S. GAAP. Fourth, Mexico provides a contrasting environment, in comparison with the U.S., for considering the effects of ownership structure. As is the case for several other emerging market countries, Mexican firms typically issue multiple
3 See Cochrane, Shapiro and Tobin (1996) for this and additional evidence concerning the internationalization of U.S. investment.
4 NYSE website, www.NYSE.com, statistics as of November 2001.
6
classes of equity that differentiate between foreign and domestic traders.5 Domowitz et al. (1997) have documented significant price premia for Mexican shares not restricted to domestic investors.
These market efficiency, legal, disclosure and shareholder class features, coupled with Mexico’s traditionally heavier reliance on banking institutions as opposed to capital markets as a source of capital, and a tendency for companies to be closely held, may make Mexican firms significantly different than U.S. firms and firms cross-listed from other countries.
A reliable comparison of the financial performance of Mexican ADR and NCL firms may be impacted by the potential for differential information quality between the two groups. In this regard, Lang, Raedy and Yetman (2001) compare the characteristics of local-GAAP earnings for ADR firms relative to matched samples of NCL firms and U.S. firms.6 They analyze financial performance based on standard profitability and risk measures and analyze accounting quality differences based on earnings smoothing, accrual management and earnings frequency distribution tests. In general, they conclude that ADR firms are of higher quality than NCL firms.
The Lang, Raedy and Yetman (2001) results are interesting for two reasons. First, their evidence that ADR firms tend to be more financially healthy and of lower risk than NCL firms conflicts with the results of the global equity offering literature (in terms of returns) and is inconsistent with our hypothesis that Mexican ADR firms will exhibit weaker financial performance than NCL firms. Second, their study presents aggregate results for a sample
5 Mexican A shares represent a majority of corporate ownership and are restricted to Mexican nationals while B shares are available to all investors. See Davis-Friday (2001) and Gordon (1999) for more information regarding Mexican capital market and financial reporting institutions.
6 The Lang, Raedy and Yetman (2001) study is the only other study of which we are aware that deals with the selection bias issue by comparing ADR and non-ADR firms. However, their study differs from ours in several 7
spanning twenty-six countries, but dominated by United Kingdom (34.5% of sample) and Japanese (15.19% of sample) firms. The mixing of code law and common law countries and developed and emerging market economies makes it difficult to generalize the results and masks any potential individual country differences.
3. Research Design and Hypothesis Development
We develop and test four hypotheses. First, firms that cross list exhibit weaker financial performance than those that do not cross list. Second, the major motivation for cross listing is to reduce capital constraints. Third, because of the lack of enforcement of regulations in Mexico, Mexican cross-listed firms’ share prices will be more associated with their accounting information as a result of the additional U.S. regulation and finally, ADR firms’ earnings will be of higher quality than those of non-ADR firms.
3.1 Financial Characteristics of ADR Firms
Prior research does not compare the financial performance of Mexican cross-listed companies with eligible companies who have not cross listed, so there is no prior empirical justification for positing a directional hypothesis regarding the difference between the financial performance of ADR and NCL firms. However, some general guidance is provided by the global equity offerings literature.
Foerster and Karolyi (2000) analyze the long-run performance of a sample of 333 global equity offerings via U.S. ADRs over the period 1982-1996. They show that for their sample of Asian, Latin American and European issuers, three-year post-issuance returns are from 8-15% lower than for local market benchmarks. The results are consistent with the documented results
respects, including the fact that their sample of non-ADR firms is not restricted to eligible cross-listers. Readers are
8
for domestic equity offerings (e.g., Ritter 1991, Spiess and Affleck-Graves 1995, and Loughran and Ritter 1995). Foerster and Karolyi (2000) investigate regional differences in their overall results, and after separating the returns for private and public issues, they show that the key driver of poor performance in Latin America is the public issues program (e.g., Tubos de Acero de Mexico). Segmenting their sample into developed and emerging market groups, global equity offerings for developed markets outperform emerging markets for all horizons and benchmarks. No specific country results were reported.
Based on both the domestic and global equity offering literature, we expect weaker financial performance for our ADR population in comparison with NCL firms (benchmark). We also attempt to analyze before- and after-cross listing performance for a small sample of ADR issuers for which we have sufficient data. Again, based on prior literature, we expect to observe weaker financial performance after cross listing for these firms and this is the basis of our first hypothesis stated in the alternative form:
H1: ADR firms exhibit weaker financial performance relative to NCL firms and relative to their own performance prior to cross listing.
Consistent with the Lins et al. (2000) idea that foreign companies issue equity on U.S. exchanges to relax capital constraints, we expect our ADR firms to exhibit higher financial leverage than the benchmark NCL firms. Whether or not these firms are successful in loosening capital constraints subsequent to ADR listing depends on post-ADR listing performance. While we do not posit a reduction in financial leverage subsequent to ADR listing, we do provide evidence for a small sample of firms for which we have data. Our second hypothesis describes our expectations regarding the need for capital. It is stated in the alternative form:
referred to the discussion of selection bias issues discussed in that paper.
9
H2: ADR firms are more levered than NCL firms.
3.2 The Effect of Corporate Governance on Cross Listing
The cost of greater transparency is a potentially more binding constraint on cross listing for Mexican firms than are out-of-pocket costs. In a series of papers, LaPorta et al. (1997, 1998, 1999) show that the observed prevalence of widely held corporations in the U.S. does not hold as a pattern throughout the world and that widespread corporate ownership, where it does exist, is consistent with strong legal protection for minority shareholders. LaPorta et al. (1999) analyze the ownership of the top twenty market capitalization firms in each of 27 countries, including Mexico. Ownership structure information is not available for non-crosslisted Mexican companies, so they focus on the 20 largest ADR firms. All are classified as “family owned” by the authors.7 Thus, even though privacy concerns exist as a potential constraint on cross listing since Mexico has one of the highest incidences of family ownership, the majority of eligible Mexican firms has cross listed in the U.S.
In two papers focused on Mexico, Domowitz, Glen and Madhavan (1997, 1998) explore the effects of ownership restrictions and cross listing on stock prices and market liquidity. Their share price analysis of 24 equity issues on a daily basis and 46 equity issues on a weekly basis for 21 firms over the 1990-1993 period indicates significant price premia for shares held by foreign investors (B shares). They also examine the impact on the Mexican stock exchange (BMV) around the time of ADR issuance and present strong evidence of increased volatility and weak evidence of reduced liquidity (stronger evidence of reduced liquidity for B shares). In
7 We verified the high incidence of family ownership of Mexican firms by referencing a sample of 20-F registration statements available to us.
10
examining the impact of ADR listing on bid-ask spreads, they observe decreased spreads in 22 cases.
Additionally, Mexican firms are more likely to benefit from cross listing than are firms from many other countries because of time zone considerations. Pulatkonak and Sofianos (1990) indicate that the six countries with the greatest U.S. market share are all within one time zone of New York and that 68.7% of the trading of Mexican cross-listed firms occurs in the U.S. The importance of time zone relates primarily to trading session overlap facilitating arbitrage trading, simultaneous security pricing in the U.S. and home markets, and the ability to disseminate new information during business hours.
Finally, in their tests of the efficiency of the Mexican stock market, Bhattacharya et al. (2000) note that in spite of laws against insider trading, there has never been an indictment, trial, or conviction for insider trading by the Mexican equivalent of the SEC. Their examination of daily price reactions to corporate news announcements for a sample of Mexican A and B shares over the period July 1994-June 1997 shows no difference in returns, volatility, trading volume, or bid-ask spreads for event and non-event periods. They provide further evidence that the return volatility of A shares leads the return volatility of B shares (consistent with insider trading), but not strongly enough for trading rules to arbitrage it away.
These issues relate directly to the results in Lins et al. (2000). Their results demonstrate that U.S. listing reduces the indirect costs of market segmentation by improving access to capital, especially for emerging market firms. In the case of Mexico, ADR shares traded in the U.S. may represent underlying B shares or A shares (those previously restricted to ownership by Mexican nationals). Therefore, the cross listing reduces market segmentation and potentially improves the informational efficiency of the cross-listed share prices. Thus, a potential benefit
11
of cross listing for investors in Mexican securities may be improved market efficiency resulting from U.S. regulation.
3.3 Accounting Information and Regulation
Much of the prior research related to the SEC disclosure debate has focused on the value-relevance of GAAP reconciliations. Amir, Harris and Venuti (1993) examine the value-relevance of both the aggregate reconciliations of earnings and shareholders’ equity and the value-relevance of the individual components for a sample of 101 firms based on Form 20-F filings over the period 1981-91. They find value-relevance for both aggregate reconciliations and for the capitalized goodwill and asset revaluation components of the reconciliations. Their sample is dominated by firms from the U.K., the Netherlands, Australia, and Sweden and includes only three Mexican firms.
Rees and Elgers (1997) extend the work of Amir et al. (1993) by focusing on the reconciliations found in initial Form F-1 registration statements for a sample of 140 (mostly U.K., Australian and Canadian) firms for the period pre-1985 through 1991. The authors regress ROE (book value deflated earnings), and the reconciliations (change in stockholders’ equity and the change in earnings) on price-to-book value for up to three retrospective years and find statistically significant results for the stockholders’ equity reconciliation and some individual components of the reconciliations (goodwill, asset revaluations and “all else” categories). Regressions of reconciling items on returns were not significant.
Davis-Friday and Rivera (2000) investigate the value-relevance of Form 20-F reconciliations for a sample of Mexican ADRs. They find that while both Mexican GAAP and U.S. GAAP earnings and book values are significantly related to ADR prices, the reconciliation from Mexican to U.S. GAAP is not statistically significantly related to ADR prices. The authors
12
suggest that a likely reason for the apparent lack of value-relevance for the reconciliation is that Mexican financial statements are inflation-adjusted and the reconciliations exclude the inflation adjustments. The inflation adjustment is likely the largest difference between Mexican and U.S. GAAP.
To our knowledge, no prior study has compared the value-relevance of accounting information reported by Mexican ADR and NCL firms. We expect to observe significantly greater value-relevance for our ADR firms in comparison with NCL firms. This result is anticipated due to the relative nature of the information environments in which the two sets of firms operate. According to evidence from Bhattacharya et al. (2000), the U.S. capital market appears to be more highly regulated than the Mexican market. Therefore investors may be more willing to use financial statement information to set security prices for the ADR firms than for the NCL firms. Our third hypothesis, stated in the alternative, is as follows:
H3: Earnings and book values explain more of the cross-sectional variation in stock prices for ADR firms than for NCL firms.
Based on our limited data set, we perform quality of earnings tests for our ADR and NCL samples similar to those in Land and Lang (2002) and Lang, Raedy and Yetman (2001). Consistent with the Lang et al. (2001) results and the regulatory environment in Mexico, we expect to observe lower quality accounting for the NCL firms. Failure to reject this hypothesis will give some credence to the belief that any observed difference in financial performance between the two groups is reliably measured and ADR firms exhibit higher earnings quality than NCL firms. Our final hypothesis, stated in the alternative form is:
H4: ADR firms exhibit higher earnings quality than NCL firms.
13
4. Results
4.1 Data and Sample Selection
Our primary data source for financial statement and market data is Economatica, a commercial database for Latin American companies. Economatica obtains its firm financial statement data from the Mexican stock exchange (Bolsa). We obtained security market betas from Bolsa statistical summaries (Anuario Bursatil de la Bolsa Mexican de Valores) and data for return calculations from the same source. The number of analysts following our sample firms was obtained from IBES International.
The financial statement data used in subsequent analyses were prepared using Mexican accounting principles. The accounting data are general price-level-adjusted and translated to U.S. dollars using peso-to-dollar exchange rates at each year-end.8 We do not analyze reconciliations to U.S. GAAP for level 2 and level 3 ADR companies for two reasons. First, we want to be able to compare financial characteristics of ADR and NCL firms on the same basis. Second, previous research (e.g., Davis-Friday and Rivera (2000)) indicates that the reconciliations are not value-relevant for Mexican companies. We summarize the sample selection procedures in table 1.
[Insert table 1 here]
The ADR sample consists of all Mexican non-financial services firms with available data that were ever listed on U.S. exchanges during the 1995-1999 period.9 At the end of 1995, there
8 Two inflation adjustment accounts are normally shown in Mexican financial statements. First, any gain or loss on net monetary position is shown in the income statement as part of a comprehensive financing cost category that also includes interest and foreign exchange gains or losses. The offsetting account for the price level adjustments for nonmonetary items is “Cumulative effect of restatement” or “Gain or loss from holding non-monetary assets” and is a component of Shareholders’ Equity. Most of the sample companies had large cumulative losses from holding monetary assets during our test period.
9 Banks were eliminated due to their differing financial characteristics in comparison with the rest of the sample. This constraint eliminated two financial services firms (Banca Quadrum and Grupo Financiero Serfin) from the
14
were 22 Mexican ADRs listed on the New York Stock Exchange (NYSE); three traded on the NASDAQ, and two on the American Stock Exchange (AMEX). By the end of 1999 there were 34 Mexican firms trading on the NYSE, NASDAQ or AMEX. After excluding the two banks (footnote 10), the ADR sample ranges from 23 to 32 firms for a total of 140 firm years.10 However, in subsequent analyses, two firm-year observations with negative book values and three outliers (ROE greater than three standard deviations from the mean) were eliminated. Additional missing data (especially price data) further limited the sample in some years (see table 4). All of the sample firms have December 31 fiscal year ends. The sample includes slightly more level 3 ADRs than level 2 ADRs.
The non-ADR sample is also drawn from Economatica, and includes all additional Mexican firms that meet the NYSE international listing requirements, but are not exchange-listed.11 This sample includes firms that are listed over-the-counter (OTC) or as private placements (144a).12 It also includes firm years for three companies that later listed as Level 2 ADRs: Cemex (1995-1998), Fomento Economico Mexico (1996-1997), and Grupo Maseca (1995-1997). The annual non-ADR sample size ranges from 14 to 29 firms (40 different firms) and a total of 121 firm years.
ADR sample and three financial services firms (Grupo Financiero Banamex Accival, Grupo Financiero BBVA Bancomer and Grupo Financiero Inbursa) from the NCL sample.
10 The ADR level designations in panel A of Appendix B are the current level designations as disclosed by the Bank of New York. Note that three previous level 2 or 3 ADR firms have downgraded to level 1—Bufete Industrial, Grupo Mexicano de Desarrollo and Grupo Sidek. All three firms traded on the NYSE during our test period, but they presently trade OTC.
11 The alternate international requirements include: 5,000 worldwide round-lot holders, 2.5 million worldwide public shares, $100 million worldwide public market value, and aggregate three-year pre-tax earnings of $100 million together with a minimum of $25 million of earnings in each of the two most recent years. The financial conditions may also be fulfilled using operating cash flow.
12 We conduct our empirical analyses separately on firms that are not crosslisted at all (NCL), OTC & 144a firms, and exchange-listed ADR firms.
15
A summary of the listing status of the sample firms is provided in table 2. A key difference between the ADR and non-ADR firms (NCL and OTC/144a) is that the exchange-listed firms are much more stable in terms of meeting the NYSE listing requirements. Panel A shows that, of the 32 ADR firms in the sample, 23 were listed for all five years of the sample period. The average number of sample years per firm is 4.38. In contrast, panel B shows that only eleven of the 40 non-crosslisted (NCL) and OTC/144a firms in the eligible but not exchange-listed sample meet eligibility requirements for all five test period years. Further, sixteen of the 40 firms meet the eligibility requirements only intermittently and again, three of the firms later become ADRs. The average number of sample years per firm is only 3.03 for the non-ADR sample.
It appears that a substantial majority of eligible Mexican firms have cross listed their securities in the U.S. as level 2 or 3 ADRs. Many of the rest of the firms have only recently or intermittently met NYSE eligibility requirements. For others, we suspect there are idiosyncratic reasons for not cross listing. For example, Kimberly Clark de Mexico, Sears Roebuck and Wal Mart de Mexico have U.S. counterparts and there appear to be few benefits to cross listing for these firms. Grupo Bimbo, an eligible, but non-crosslisted firm, acquired a bakery in Texas in order to take advantage of distribution channels and capital resources in the U.S.
It appears that, at least for eligible Mexican companies, the SEC disclosure requirements are not a major constraint to cross listing. Apparently, the benefits of cross listing (primarily access to capital, we believe) exceed the costs. Over-the-counter (OTC) and 144a firms also have the opportunity to benefit from raising private capital, but without incurring the costs of SEC registration. Later, we provide evidence bearing on the viability of this option.
[Insert table 2 here]
16
4.2 Industry Classifications
Our analysis of the differences between ADR and non-ADR firms begins with a description of their respective industry (sector) distributions in table 3. Based on the sector names from Economatica, the ADR and non-ADR firms represent 18 different sectors. The 32 ADR firms are in 11 different sectors and the 40 non-ADR firms are in 16 different sectors. While we find broad industry representation for both samples, there are also substantial industry differences. The top four industries represent over 59% of the firm year observations in the ADR sample (communications 18.7%, construction 15.6%, steel 12.5%, and food 12.5%). In contrast, even though the non-ADR sample is more diverse in terms of industry representation, the top four industries (retail 22.5%, beverages 12.5%, cement 10.0%, and holding companies10.0%) comprise 55.0% of that sample.
The apparent sector self-selection for cross listing creates two potential problems. First, any observed differences in the overall financial characteristics of the samples have the potential to be industry-specific. Second, the lack of comparable industry data for the ADR and non-ADR samples makes it difficult to perform a matched sample analysis along the lines of previous studies (e.g., Ely and Pownall (1999) and Lang, Raedy and Yetman (2001)). In an attempt to deal with both problems, we present overall sample results and also discuss results for a reduced sample of five industries that have sufficient comparative data.
[Insert table 3 here]
17
4.3 Descriptive Statistics
Table 4, panel A, presents descriptive statistics pooled over the 1995-1999 test period for financial statement and market variable comparisons of three groups: ADR, OTC/144a and NCL firms. Panel B presents various ratio and growth performance measures for the three groups and significance tests are summarized in panel C. The data are somewhat skewed by larger firms, so the medians are better measures of central tendency than the means.
Panel A of table 4 shows that Mexican ADR firms tend to be nominally smaller than both the NCL and OTC/144a firms. For example, for the six financial statement variables, five of the ADR medians and four of the six means are smaller than those of the other two groups. In addition to the size differences, of particular interest is the apparent higher financial leverage and lower profitability of the ADR sample compared especially to the NCL firms.
Several market value measures are presented in panel A. Market value of equity (MVE) is nominally highest for the OTC/144a firms and lowest for the ADR firms. The Relative Size variable (ratio of MVE to total market capitalization of the Bolsa) is also highest for the OTC/144a firms and lowest for the ADR firms. We were only able to obtain small samples of beta and returns data, therefore it is difficult to generalize much from these results, but there is some evidence of lower returns for the ADR firms. Finally, analyst coverage tends to be highest for the OTC/144a firms with little difference in coverage between ADR and NCL firms.
An overall conclusion from panel A of table 4 is that, after meeting threshold size requirements for listing on the NYSE, the relative size of a firm is not a determinant of cross listing (i.e., relatively smaller eligible firms do not avoid cross listing) status for Mexican firms.
Panel B of table 4 summarizes financial and market performance measures for the three groups. Even after deleting outliers, the distributions remain somewhat skewed so we again focus more
18
heavily on the medians. Here, the overall conclusion is that the financial performance of the ADR firms is weaker than that of both the NCL and OTC/144a firms over the 1995-1999 test period. For example, the mean (median) ROE’s for the ADR sample of –10.5% (10.1%) are significantly lower than the mean (median) ROE’s of 13.7% and 14.7% (16.2% and 13.2%) for the respective OTC/144a and NCL samples. Further, the relatively lower ROE’s for the ADR firms are impacted by significantly higher leverage (Assets/Equity) in comparison with the NCL firms. The ROA measures are also significantly lower for the ADR firms. Further analysis shows that the ROA results are driven by both significantly lower profit margins and asset turnover for the ADR firms. Based on the growth in total assets variable, it also appears that historical growth has been lower for the ADR firms than for the other two groups.
Further analysis shows that an important difference in the distributions of ROE and ROA for the ADR firms in comparison to both the OTC/144a and NCL groups is the presence of 36 negative observations for the ADR group compared with only four negative observations for the other two groups combined. The negative observations remain even after deleting negative book value observations (2) and outliers greater than three standard deviations from the original means of the distributions (3). In fact, nineteen of the 32 ADR firms in the sample had at least one year of negative ROE/ROA. The wide distribution of net income/net loss particularly impacts the standard deviation of ROE (1.645) for the ADR group. The negative observations are indicative of weaker financial performance after cross-listing for the ADR firms, since the stock exchanges require at least three years of positive earnings prior to cross-listing.
The weaker financial performance for the ADR sample (and expectations of weaker future performance) is also reflected in the price-to-book (PB) and price-to-earnings (PE) ratios. The median PB ratio of 1.410 for the ADR sample is significantly lower than the median PB’s of
19
1.785 and 1.910 for the respective OTC/144a and NCL samples. While the differences in the PE ratio means and medians are not significant, the median ratio is nominally lower for the ADR sample.
Panel C of table 4 summarizes the results of pairwise comparisons of mean and median differences for the three groups. Most of the significant differences have been noted above. The primary additional insight is that there are significant differences in the variables for all three groups. Based on ROE and ROA measures, financial performance is strongest for the OTC/144a firms and weakest for the ADR firms. The NCL firms differ most prominently from the ADR and OTC/144a firms in terms of their lower financial leverage. The significant differences among all three groups cause us to analyze each group separately in tables 6 through 9.
[Insert table 4 here]
We now report on an untabulated analysis of ADR and NCL differences for a small sample of industry-matched firms. We include all firms with at least four years of data from the following industries (number of ADR and NCL firms, respectively, from each industry in parentheses): food (2,1), holding companies (3,3), paper products (1,1), retail (2,3) and steel (4,1). Thus, the reduced sample includes only 12 ADR and nine NCL firms, and at most, 56 (47) pooled observations for the ADR (NCL) groups.
The results for the reduced sample are very consistent with those for the full sample reported in table 4. For example, we still observe the significant size difference between the two groups. One indicator of the size difference is MVE (market value of equity). The mean (median) for the NCL group is 22,197(17, 817) in comparison with 5,351(3,317) for the ADR group. The results for the performance measures are also generally consistent with those in table 4. While the nominal difference in ROE (NCL lower) is no longer significant, the ROA and
20
leverage differences remain significant, indicating that ROE is impacted by higher leverage for the ADR group. For the full sample, the ROA difference was driven by significant differences in both profit margin and turnover; for the reduced sample, only the difference in turnover is significant. The results for PB and PE are consistent with those for the full sample. Finally, we note that the Revenue Ratio (ratio of domestic revenue to total revenue) difference is now significant for the reduced sample. This suggests that, after controlling for firms’ industry membership, the exchange-listed ADR firms earn more of their revenues outside of Mexico.
Using our non-ADR sample as a reasonable local market benchmark, the weaker performance for Mexican ADR firms, based on financial characteristics, is consistent with the results of Foerster and Karolyi (2000) for their combined 35-country sample based on returns.13 In an attempt to compare the before and after exchange listing performance of ADR firms, we work with a reduced sample of level 2 and 3 ADRs for which we can obtain data for at least one year prior to cross listing. Table 5 compares the financial performance of these firms for one or two years preceding and for two years after ADR listing. While the sample is very small and our tests lack power, an overall conclusion is that financial performance did not improve after cross listing. Although the results are not significant, ROE and ROA means and medians are nominally lower after cross listing. Also note that there does not appear to be any significant reduction in leverage after cross listing. Again, we caution that this analysis is based on extremely small samples.
13 Foerster and Karolyi (2000) also provide evidence that while companies issuing equity in the U.S. on major exchanges modestly outperform their benchmarks (in contrast to under-performance of private placements), firms that come from emerging markets with low accounting standards (Mexico included) significantly outperform their benchmarks. We do not have enough data to separately analyze private placements versus public placements in terms of financial performance. However, our ADR sample is dominated by public placements (see table 6), and we believe our results for Mexico are consistent with Foerster and Karolyi (2000) results for emerging market companies in general.
21
Based on the limited before and after cross listing evidence in table 5 along with the results of the benchmarking comparison of the performance of ADR firms with non-ADR firms in table 4, and the prevalence of negative earnings after cross listing as noted above, we conclude that the financial performance results for Mexican ADR firms appear consistent with the results of prior global equity offering research; that is, there is no evidence that greater access to capital is associated with improved financial performance for Mexican firms after cross listing.
[Insert table 5 here]
4.4 Capital Issuance Analysis
As previously noted, Lins et al. (2000) provide evidence that non-U.S. firms issue equity on U.S. stock exchanges to relax capital constraints. Evidence of this constraint is the significantly higher financial leverage of Mexican cross-listed firms in comparison with non- cross-listed (NCL) firms. Further, in a dynamic capital structure environment, Palepu, Healy and Bernard (2000) and Myers and Majluf (1984) argue that the issuance of new equity represents last resort financing.
Table 6 summarizes debt and equity issues by Mexican firms in the U.S. broken down by our sample categories of firms.14 Several insights are evident from the table. First, much more capital is raised by level 2 and 3 ADR firms than by NCL firms. Level 2 or 3 ADR firms raised over 83% of total debt and almost 87% of total equity. Improved access to capital markets, at the expense of meeting SEC disclosure requirements, is clearly a benefit to cross listing as a level 2 or 3 ADR. Second, substantially more equity ($11.8 billion) has been issued than debt ($4.3 billion). The substantially higher equity issuance is consistent with the idea of accessing U.S. markets to loosen financing constraints. At the same time, the issuance of relatively less
22
debt in the U.S. may be indicative of the close ties Mexican firms have to their domestic bankers (LaPorta et al. 2002). Finally, we note the limited amount of capital that has been raised by OTC and 144a firms.
[Insert table 6 here]
4.5 Logistic Regression Results
Tables 3 through 6 provide evidence of individual differences in various financial and market characteristics among the ADR, OTC/144a and NCL firms. We now use a Logistic regression model to examine the conditional relation between the variables we identified above and the likelihood of cross listing.
Under the Logit model, the odds of a firm’s cross listing in the U.S. are described by the conditional ratio of: P(Gj/Xj)/(1-P(Gj/Xj)), where P(Gj/Xj) is the probability of being a member of G given Xj. The log-linear function of this probability can be expressed as follows (Maddala 1991):
()()()jkjojjLogXLogXAAPPLogk11A...1/+++=− (1)
where: Pj = probability of firm j cross listing and Xjk = the explanatory variables for firm j.
We test the classificatory power of a model that includes the following independent variables from table 4: MVES (size), ROA (profitability), LEVERAGE (capital structure), GROWTH, REVRATIO (ratio of domestic to total revenue) and INDUSTRY. The variables are chosen based on theoretical considerations and previous empirical research; for example, LEVERAGE is based on the capital constraint hypothesis (Lins et al. 2000), ROA is a profitability proxy based on prior global equity offering research (Foerster and Karolyi 2000) and MVES is a proxy for the firm’s size and its need for foreign capital based on Saudagaran
14 Data were obtained from the Bank of New York. While Level 2 ADRs may not raise new capital in the U.S. by 23
(1988). GROWTH is also related to the firm’s need for capital. We include REVRATIO (the ratio of domestic to total revenue) as proxy for firms’ exposure outside of their home country (Saudagaran 1990) and INDUSTRY based on our previous analysis. We do not use the full set of variables in table 4 so as to avoid over-fitting models to our relatively small samples.15
The first column of table 7 summarizes the results for the ADR/NCL comparison. First, the model exhibits reasonable classificatory power, as evidenced by the concordant percentage of 80.80%, in comparison with the naïve hypothesis that all sample firms are ADRs (128 out of 209, or 61.2%). All variables except ROA and GROWTH are significant. As expected, the variable with the highest classificatory power is LEVERAGE. REVRATIO was generally not significant in the univariate tests in table 4. However, it is highly significant here, but not in the hypothesized direction. MVES is also significant and there is some evidence of an industry effect.
The second column of table 7 reports the results of the logistic regression for a comparison of ADR and OTC/144a firms. The variables do not improve the classification compared to the naïve hypothesis that all sample firms are ADR firms (concordant percentage, 72.6%, naïve percentage 73.1%). Importantly, the LEVERAGE variable is not significant. This suggests that these two groups are very similar in terms of the characteristics that we investigate.
The third column of table 7 reports the results of a comparison of OTC/144a firms with NCL firms. Compared to a naïve hypothesis that all sample firms are OTC/144a firms (47/108 = 43.52%), the concordant percentage of 89.5% shows the model is very effective in distinguishing
issuing new shares, they are not prevented from raising equity capital through private issues.
15 We also performed stepwise logistic regressions using various combinations of variables from table 4 without any improvement in results.
24
between the two groups. Again, the most highly significant variable is LEVERAGE, followed by REVRATIO. All other variables add classificatory power except for GROWTH.
Overall, these results support our first and second hypotheses that ADR firms are more capital-constrained (levered) than NCL firms and they exhibit weaker financial performance after cross listing.
[Insert table 7 here]
4.6 Value-Relevance Hypothesis
The final analyses investigate whether there are differential associations between accounting information and stock prices for ADR firms, which, due to cross listing, may operate in a different information environment than NCL firms. Our tests are based on standard regressions of accounting book value (BVS) and earnings (EPS) on market value of equity (MVS) as in Amir et al. (1993). In table 8, we present both individual year and pooled (1995-1999) results for the ADR, OTC/144a and NCL samples.16
While we report results for all three groups, we primarily focus on the NCL and ADR comparisons here. Table 8 provides evidence of a significant difference in the explanatory power of the accounting variables for the two groups.17 The pooled adjusted R2 for the ADR sample in panel A is 0.71 and individual year R2s range from 0.55 to 0.88. In contrast, the adjusted R2 for the pooled NCL sample in panel C is essentially zero, and only in one year, 1995, is there evidence of explanatory power on par with the ADR group.18 Both BVS and EPS are significant in the pooled regression for the ADR group. BVS for the ADR group is also highly
16 Correlation analyses for both pooled samples indicate significant associations among all the accounting and market variables and no significant association with the time period (year).
17 We estimated F-statistics for comparing nested regression models and found that except in 1995, the adjusted R2 is significantly higher in the ADR sample estimation. 25
significant in all of the individual regressions, while EPS is only significant in 1997. The anomalous result for the NCL sample is the highly significant result for 1995 and the corresponding low explanatory power for the other four years.19
We also estimate pooled regression results for separate Level 2 ADR, Level 3 ADR, OTC and 144a samples (untabulated). The results are interesting. There is significant explanatory power (adjusted R2 of 0.63) for BVS for the 144a group; however, the result is based on only fourteen observations. For the OTC group, the adjusted R2 is 0.44, and both BVS and EPS are significant. We find the strongest results for the level 2 ADR group (adjusted R2 of 0.83) where both BVS and EPS are significant. Finally, the adjusted R2 for the level 3 ADRs is 0.65, but only BVS is significant.
These results support the third hypothesis that ADR (and OTC/144a) firms’ accounting information is more highly associated with stock prices than that of NCL firms.
[Insert table 8 here]
4.7 Sensitivity Analysis: Quality of Earnings
So far, we have provided consistent evidence that Mexican NCL firms are financially stronger than ADR firms, based on the measures we examine. However, Lang et al. (2001) find evidence of greater income smoothing behavior by NCL firms. The authors’ results are for data pooled over twenty-six countries. In the case of Mexico, there is also reason to expect greater smoothing behavior on the part of NCL firms due to the belief that domestic regulatory control may not be very effective.
18 Again, we caution that individual year regressions, particularly as reported in panels B and C, are for extremely small samples.
19 However, this anomaly results from a regression with only eight observations.
26
In table 9, we summarize the results of smoothing tests based on relations between the cash flow and accrual components of earnings. First, in panel A, note that the variability of accruals and cash flows from operations is always nominally greatest for the ADR firms in comparison with the other two groups. Particularly in the case of cash flow from operations, the greater variability may indicate that Mexican ADR firms are riskier than Mexican NCL firms. Ceteris paribus, the NCL firms should exhibit smoother earnings than the ADR firms. Thus, evidence indicating a difference in smoothing measures between the groups is not necessarily indicative of a difference in the quality of earnings.
In panel B of table 9 we present the results of correlation tests between accruals and cash flows for the two groups. As in Lang, Raedy and Yetman (2001), a more negative cash flow/accrual correlation is indicative of earnings management in the sense of using accruals to smooth variability in cash flows. We find that the correlation between cash flows and accruals is significantly more negative for ADR firms compared to NCL firms. The differences for the ADR versus OTC/144a and NCL versus OTC/144a comparisons are not significant.
In untabulated results, we also examine the frequency distributions of earnings to assess the frequency of small, positive earnings for the two groups as in Burgstahler and Dichev (1997). There is no evidence of any difference in the distribution of small positive earnings between the groups. This result is contrary to the results for Lang et al. (2001) as expressed in hypothesis four. Taken together, these results suggest that the differences in financial characteristics we observe are probably not an artifact of accounting.
In summary, our tests of various financial characteristics indicate that Mexican ADR firms are more highly levered and that they have raised significantly more capital in the U.S. than non-exchange listed firms. Further, the results indicate that, after cross listing, Mexican
27
ADR firms are financially weaker than NCL firms. While there is some (weak) evidence that reported earnings for NCL firms are smoother than reported earnings for ADR firms, it is more likely that such differences reflect risk differences (e.g., higher variability of cash flows for ADR firms) rather than differences in the prevalence of earnings management for the two groups.
Lang et al. (2001) suggest that firms may self-select into cross listing; stronger firms may use cross listing to credibly signal their strength and (weaker) non-crosslisted firms may find the SEC disclosure requirements an impediment to cross listing. Again, our results for Mexico are inconsistent with those of Lang et al. (2001) for their 26-country sample; our results are consistent with those found for IPO samples and for global equity offerings as in Foerster and Karolyi (2000).
[Insert table 9 here]
5. Conclusion
This paper examines the characteristics of Mexican companies that cross list their securities in the U.S. It has been suggested that the SEC disclosure requirements serve the purpose of restricting the access of weaker foreign companies to U.S. capital markets. However, we provide evidence that Mexican ADR firms exhibit weaker financial performance than non-crosslisted (NCL) firms. For example, we document a high percentage of negative earnings observations for ADR firms and relatively few negative earnings numbers for NCL firms. Additionally, our comparisons of Mexican ADR and NCL companies indicate that ADR firms are more highly levered than non-crosslisted (NCL) companies and we document the significant amount of equity capital they have raised in the U.S. in an attempt to loosen capital constraints. Logistic regression models using financial variables as inputs are highly significant in correctly
28
classifying ADR and NCL firms. The results indicate that the substantially higher leverage for ADR firms is an important financial characteristic that differentiates the two groups. Finally, we test the relative quality of financial statement numbers for ADR and NCL firms, and conclude that accounting quality differences do not explain the differences in financial performance.
We also hypothesized that financial disclosures for Mexican ADR firms will be more value-relevant than for NCL firms. Our regressions of share price on book value per share and earnings per share reveal significant explanatory power for the ADR firms and generally insignificant results for NCL firms. The results indicate that financial statements are value relevant for ADR investors, but not for investors in Mexican companies that do not cross list in the U.S. The results may indicate that the SEC regulations add credibility to the financial statements of ADR firms. In contrast, the lack of value-relevance of financial statements for NCL firms may be indicative of market inefficiency (e.g., insider trading) in the Mexican market as suggested by Bhattacharya et al. (2000).
The results from our research underscore the importance of focusing on a single country where market institutions and corporate governance mechanisms can be carefully considered. Our focus on a single country is also important due to the tendency of previous studies to generalize based on multi-country samples, particularly samples dominated by Australian, Canadian and U.K. firms. Of course, the benefits of our single country, Mexican market focus are offset somewhat by data limitations and constraints from working with relatively small samples.
29
REFERENCES
Amir, E., T. Harris, and E. Venuti. 1993. A comparison of the value-relevance of U.S. versus non-U.S. GAAP accounting measures using Form 20-F reconciliations. Journal of Accounting Research (Supplement) 31: 230-264.
Baker, H., J. Nofsinger and D. Weaver. 1999. International cross listing and visibility. NYSE Working Paper 99-01.
Bank of New York (BNY). 1997. Depositary Receipt Directory. New York, NY: BNY.
Bhattacharya, U., H. Daouk, B. Jorgenson, and C. Kehr. 2000. When an event is not an event: the curious case of an emerging market. Journal of Financial Economics (55): 69-101.
Bruner, R., S. Chaplinsky and L. Ramchand. 1999. Coming to America: A clinical study of IPOs in the U.S. by foreign firms. Working paper, University of Virginia.
Burgstahler, D. and I. Dichev. 1997. Earnings management to avoid earnings decreases and losses. Journal of Accounting & Economics (December):99-126.
Cochrane, L., J. Shapiro, and J. Tobin. 1996. Foreign equities and U.S. investors: Breaking down the barriers separating supply and demand. Stanford Journal of Law, Business & Finance. (Vol. 2.2): 241-263.
Davis-Friday, P. 2001. Equity valuation and current cost disclosures: The case of Mexico. Journal of International Financial Management and Accounting (Autumn): 260-285.
Davis-Friday, P. and J. Rivera. 2000. Inflation accounting and 20-F disclosures: Evidence from Mexico. Accounting Horizons (June):114-135.
Domowitz, I., J. Glen, and A. Madhavan. 1997. Market segmentation and stock prices: Evidence from an emerging market. Journal of Finance (July): 1059-1085.
_____ 1998. International cross listing and order flow migration: Evidence from an emerging market. Journal of Finance (December): 2001-2027.
Ely, K. and G. Pownall. 1999. The value relevance of consolidated vs. unconsolidated earnings and book values for Japanese firms: Evidence from cross-listed and non-crosslisted firms. Working paper, Emory University.
Foerster, S. and G. Karolyi. 2000. The Long-run performance of global equity offerings. Journal of Financial and Quantitative Analysis (December): 499-527.
Gordon, E. 2001. Accounting for changing prices: The value-relevance of historical cost, price level, and replacement cost accounting in Mexico. Journal of Accounting Research (June): 177-200.
30
LaPorta, R., F. Lopez-De-Silanes, A. Shleifer, and R. Vishny. 1998. Law and finance. Journal of Political Economy (106): 1113-1155.
_____. 1999. Corporate ownership around the world. The Journal of Finance (April): 471-517.
LaPorta, R., F. Lopez-de-Silanes, and G. Zamarripa. 2002. Related lending. Working paper, Harvard University, Yale University and National Banking and Securities Commission (Mexico).
Land, J. and M. Lang. 2002. Empirical evidence on the evolution of international earnings. Working Paper, University of North Carolina.
Lang, M., J. Smith Raedy, and M. Yetman. 2001. How representative are cross-listed firms? An analysis of firm and accounting quality. Working Paper, University of North Carolina.
Lins, K., D. Strickland, and M. Zenner. 2000. Do non-U.S. firms issue equity on U.S. stock exchanges to relax capital constraints? Working paper, University of Utah.
Loughran, T. and J. Ritter. 1995. The new issue puzzle. Journal of Finance (50): 23-51.
Maddala, G. S. 1991. A Perspective on the use of limited-dependent and qualitative variables models in accounting research. The Accounting Review (October): 788-808.
Miller, D. 1999. The market reaction to international cross-listings: Evidence from depositary receipts. Journal of Financial Economics (51): 103-123.
Myers, S. and N. Majluf. 1984. Corporate financing and investment decisions when firms have information that investors do not have. Journal of Financial Economics (June):187-222.
Pulatkonak, M. and G. Sofianos. 1999. The distribution of global trading in NYSE-listed non-U.S. stocks. NYSE Working Paper 99-03.
Palepu, K., V. Bernard and P. Healy. 2000. Business Analysis and Valuation. (South-Western College Publishing).
Rees, L. and P. Elgers. 1997. The market's valuation of non-reported accounting measures: Retrospective reconciliations of non-U.S. and U.S. GAAP. Journal of Accounting Research (Spring): 115-127.
Ritter, J. J. 1991. The long-run performance of initial public offerings. Journal of Finance (46): 3-27.
Saudagaran, S. 1988. An empirical study of selected factors influencing the decision to list on foreign stock exchanges. Journal of International Business Studies (Spring): 101-127.
_____. 1990. An investigation of selected factors influencing multiple listing and the choice of foreign stock exchanges. Advances in Financial Planning and Forecasting (Volume 4, Part B): 75-122.
31
Smith, K. and G. Sofianos. 1997. The impact of an NYSE listing on the global trading of non-U.S. stocks. NYSE Working Paper 97-02.
Spiess, D. and J. Affleck-Graves. 1995. Under-performance in long-run stock returns following seasoned equity offerings. Journal of Financial Economics (38): 243-267.
Stulz, R. 1999. Globalization, corporate finance, and the cost of capital. Journal of Applied Corporate Finance. (47): 1425-1460.
32
TABLE 1
Sample Selection Procedures
1995
1996
1997
1998
1999
Number of Companies Traded on BMV
185
193
198
194
188
Number of Companies Eligible for NYSE
37
51
59
64
60
Number of Companies Traded on U.S. Exchanges1
25
28
30
33
34
Less firms missing data on Economatica or Bolsa
(2)
(2)
(2)
(2)
(2)
ADR Sample
23
26
28
31
32
Eligible Companies not Traded on NYSE
14
23
29
32
26
Companies Traded OTC2
3
5
8
13
8
Less Banks
(3)
Companies Traded Rule via 144a2
1
2
3
4
4
Eligible Companies not Traded in U.S.
10
16
11
15
14
Non-ADR Sample
14
23
29
29
26
1Teléfonos de Mexico is traded both on the NASDAQ (A shares) and NYSE (L shares). It is only included once in the NYSE totals.
2 Several of the companies traded on U.S. exchanges also have shares traded OTC or listed under Rule 144a. The number of OTC and 144a listings does not include firms that are also listed on ASE, NASDAQ, or NYSE.
33
TABLE 2
Sample Firms
Panel A: Mexican Firms Traded on U.S. Exchanges through December 1999
Company Name
Effective Date
Exchange
Level
Sector
95
96
97
98
99
Altos Hornos de Mexico2
NYSE
2
Steel
X
X
X
X
Biper
NASDAQ
2
Communications
X
X
Bufete Industrial2
NYSE
3
Construction
X
X
X
X
Cemex3
NYSE
2
Cement Production
X
Coca-Cola FEMSA
NYSE
3
Beverages
X
X
X
X
X
Consorcio G Grupo Dina2
NYSE
3
Transportation
X
X
X
X
X
Controladora Comercial Mexicana
NYSE
3
Retail
X
X
X
X
X
Desc
NYSE
3
Holding Company
X
X
X
X
X
Empresas ICA
NYSE
3
Construction
X
X
X
X
X
Fomento Economico Mexico (FEMSA)3
NYSE
2
Beverages
X
X
Gruma3
NYSE
2
Food
X
X
Grupo Casa Autrey (Saba)
NYSE
3
Food
X
X
X
X
X
Grupo Elektra
NYSE
2
Retail
X
X
X
X
X
Grupo Imsa
NYSE
3
Steel
X
X
X
X
Grupo Industrial Durango
NYSE
3
Paper Products
X
X
X
X
X
Grupo Industrial Maseca1
NYSE
2
Food
X
X
X
X
X
Grupo Iusacell
NYSE
2
Communications
X
X
X
X
X
Grupo Mexicano de Desarrollo11,2
NYSE
3
Construction
X
X
X
X
X
Grupo Radio Centro
NYSE
3
Communications
X
X
X
X
X
Grupo Sidek1,2
NYSE
3
Holding Company
X
X
X
X
X
Grupo Simec
AMEX
3
Steel
X
X
X
X
X
Grupo Televisa
NYSE
3
Communications
X
X
X
X
X
Grupo Tribasa
NYSE
3
Construction
X
X
X
X
X
Industrias Bachoco
NYSE
3
Food
X
X
X
Internacional de Ceramica1
NYSE
3
Construction
X
X
X
X
X
Pepsi-Gemex
NYSE
2
Beverages
X
X
X
X
X
Savia (Empresas la Moderna)
NYSE
3
Holding
X
X
X
X
X
Teléfonos de Mexico (A)
NASDAQ
2
Communications
Teléfonos de Mexico (L)
NYSE
3
X
X
X
X
X
Transportación Marítima Mexicana
NYSE
3
Transportation
X
X
X
X
X
Tubos de Acero de Mexico
AMEX
2
Steel
X
X
X
X
X
TV Azteca
NYSE
3
Communications
X
X
X
Vitro
NYSE
3
Mineral Products
X
X
X
X
X
TOTAL
23
26
28
31
32
1These firms trade shares on an organized exchange as well as over the counter (OTC). In the analysis that follows, they are only included in the “exchange” firm analysis.
2These firms delist after the test period. They are included in the analysis during the time period when they are listed on exchanges.
3These firms are categorized as exchange-listed firms after they cross list in the U.S. and they are also included in the non-crosslisted, but eligible sample during the years before they cross list. 34
TABLE 2 (Continued)
Sample Firms
Panel B: Mexican Firms Eligible to List on the New York Stock Exchange
Company Name
Exchange
Effective Date
144a
OTC
Amex
Nasdaq
NYSE
Sector
95
96
97
98
99
Alfa
X
Holding Company
X
X
X
X
X
Apasco
X
Cement
X
X
X
X
X
Ara Consorcio
X
Construction
X
Argos Embotelladora
Beverage
X
Bimbo Grupo
Food
X
X
X
X
X
Carso Global Telecom
X
Communications
X
Cementos Chihuahua
Cement
X
X
Cemex
X
Cement
X
X
X
X
Cintra
Airlines
X
X
X
X
Coppel
Retail
X
X
Corp Moctezuma
Cement
X
Cydsa
Holding Company
X
X
El Puerto de Liverpool
Retail
X
X
X
X
X
Empaques Ponderosa
X
Paper Products
X
X
X
Fomento Economico Mexico
X
Beverage
X
X
Grupo Azucarero Mexico
Food
X
Grupo Carso (96-99)
X
Holding Company
X
X
X
X
X
Grupo Continental
X
Beverage
X
X
X
X
X
Grupo Gigante
Retail
X
X
X
X
Grupo Industrial Saltillo
Holding Company
X
X
X
X
Grupo Maseca (Gruma)
X
Food
X
X
X
Grupo Modelo
Beverage
X
X
X
X
X
Grupo Palacio de Hierro
Retail
X
Grupo Sanborns
Retail
X
X
X
Hylsamex (96-99)
X
Steel Plant
X
X
X
X
Industrias CH
Metallurgy
X
Jugos del Valle1
X
Beverage
X
X
Kimberly Clark de Mexico
X
Paper Products
X
X
X
X
X
Mexichem
PetroChemical
X
X
X
X
Nuevo Grupo Mexico
Mining
X
X
Nutrisa
Retail
X
Panamco
Beverage
X
Penoles Industrias
Mining
X
X
X
X
X
Posadas Grupo
X
Tourism/Hotel
X
X
X
San Luis Corp
X
Automotive Parts
X
X
X
X
Sears Roebuck
Retail
X
X
Seguros Com America
X
Insurance
X
Situr Grupo
X
Tourism/Hotel
X
X
Soriana Organizacion
Retail
X
X
X
X
X
Wal Mart de Mexico (97-99)
X
Retail
X
X
X
X
X
TOTAL
5
10
3
14
23
29
29
26
35
TABLE 3
Industry Distribution of the Exchange-Listed ADR Firm and Eligible Firm Samples
Sector Name*
Exchange Listed Firms
Eligible Firms
#
%
FirmYears
%
#
%
FirmYears
%
1
Airline
0
0.00
0
0.0
1
2.5
4
3.31
2
Automotive Parts
0
0.00
0
0.0
1
2.5
4
3.31
3
Beverages
3
9.37
12
8.6
6
12.5
16
13.22
4
Cement
1
3.13
1
0.7
4
10.0
12
9.91
5
Construction
5
15.62
24
17.1
1
2.5
1
.83
6
Communications
6
18.75
25
17.9
1
2.5
1
.83
7
Food
4
12.50
15
10.7
3
7.5
9
7.43
8
Holding Company
3
9.37
15
10.7
4
10.0
16
13.22
9
Insurance
0
0.00
0
0.0
1
2.5
1
.83
10
Metallurgy
0
0.00
0
0.0
1
2.5
1
.83
11
Mineral Products
1
3.13
5
3.6
0
0.0
0
0.0
12
Mining
0
0.00
0
0.0
2
5.0
7
5.78
13
Paper Products
1
3.13
5
3.6
2
5.0
8
6.61
14
PetroChemicals
0
0.00
0
0.0
1
2.5
4
3.31
15
Retail
2
6.25
10
7.1
9
22.5
28
23.14
16
Steel
4
12.50
18
12.9
1
2.5
4
3.31
17
Tourism/Hotel
0
0.00
0
0.0
2
5.0
5
4.13
18
Transportation
(Machinery & Equipment)
2
6.25
10
7.1
0
0.0
0
0.0
Total
32
100.0
140
100.0
40
100.0
121
100.0
*The sector names are from Economatica.
36
TABLE 4
Descriptive Statistics (1995-1999)
Panel A: Financial Statement and Market Variables
Exchange Listed ADR Firmsa
OTC & 144a Firmsa
Non-Crosslisted Firmsa
Variable
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
Assets
135
2,105
3,040
1,172
48
2,477
2,810
1,634
73
1,936
2,094
1,389
Liabilities
135
997
1,071
682
48
1,151
1,293
600
73
786
1,142
467
Equity
135
1,027
2,098
488
48
1,021
855
983
73
975
830
742
Net Income
135
113
378
39
48
153
181
73
73
148
180
97
Operating
Cash Flow
135
227
663
81
47
265
500
147
73
233
296
161
Investing
Cash Flow
135
-143
339
-63
47
-220
461
-96
72
-192
260
-99
MVE
135
7,755
16,418
3,263
48
16,302
21,216
7,025
73
9,500
9,429
6,253
Relative Size
135
0.075
0.167
0.030
48
0.15
0.19
0.07
73
0.092
0.097
0.055
Beta
67
0.653
0.404
0.670
31
0.85
0.30
0.93
42
0.660
0.315
0.629
Returns
29
1.082
2.736
0.101
26
1.19
3.07
0.19
21
1.336
3.028
0.313
Analysts
91
12.396
6.031
13.000
37
16.51
6.48
19.00
42
13.214
7.380
12.500
a The data, except ratios and percentages, are totals in terms of millions of U.S. dollars.
37
TABLE 4 (Continued)
Descriptive Statistics
Panel B: Performance Measures
Exchange Listed ADR Firmsa
OTC & 144a Firmsa
Non-Crosslisted Firmsa
Variable
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
ROEa
135
-0.105
1.645
0.101
48
0.137
0.162
0.162
73
0.147
0.078
0.132
Leverage
135
3.724
11.042
2.372
48
2.599
1.745
2.090
73
1.952
0.945
1.616
ROA
135
0.037
0.086
0.042
48
0.075
0.058
0.078
73
0.081
0.040
0.074
Asset Turnover
135
0.697
0.482
0.604
48
0.643
0.292
0.602
73
0.865
0.365
0.777
Profit Margin
135
0.042
0.222
0.060
48
0.112
0.132
0.125
73
0.111
0.078
0.096
Growth (TA)
135
0.344
0.476
0.291
47
0.635
1.466
0.379
61
0.342
0.512
0.319
EPS
135
0.117
0.683
0.041
48
0.254
0.288
0.155
72
0.243
0.281
0.139
BVS
135
1.886
2.938
0.630
48
1.745
1.455
0.884
72
1.639
1.427
1.012
PE
88
19.895
24.293
13.760
38
15.339
10.212
13.355
55
20.250
25.316
15.580
PB
125
1.864
1.768
1.410
44
2.217
1.442
1.785
61
1.974
0.876
1.910
Revenue Ratio
135
0.767
0.233
0.822
48
0.838
0.216
0.907
72
0.772
0.225
0.815
aThere are 36 negative ROE observations in the ADR sample, three in the OTC & 144a sample and one in the NCL sample.
38
TABLE 4 (Continued)
Panel C: T and Z Statistics from Pairwise Comparisons of Mean and Median Differences
Non-Crosslisted Firms vs. Exchange Listed ADRsa
OTC & 144a Firms vs. Exchange Listed ADRsa
Non-Crosslisted Firms vs. OTC & 144 A Firmsa
Panel A Variable
Mean
Median
Mean
Median
Mean
Median
Assets
-1.48
0.58
-0.52
1.27
-0.84
0.90
Liabilities
**-2.76
##-2.35
-1.23
0.09
-1.38
1.58
Equity
-0.25
##2.63
-0.03
##2.03
-0.29
0.00
Net Income
0.90
###4.66
0.95
###3.39
-0.14
-0.20
Operating Cash Flow
1.53
###3.46
1.22
###2.81
-0.40
0.18
Investing Cash Flow
-0.52
###-3.10
-1.25
##-2.41
0.39
0.61
MVE
0.29
#1.87
**2.51
###4.35
**-2.21
##2.77
Relative Size
0.89
##2.13
**2.36
###3.25
*-1.91
1.23
Beta
0.10
-0.02
**2.74
##2.66
**-2.67
###2.97
Returns
0.30
0.80
0.14
0.63
0.16
0.57
Analysts
0.63
0.50
***3.32
###3.46
**-2.12
##2.12
Panel B Variable
Mean
Median
Mean
Median
Mean
Median
ROE
*1.78
###3.53
*1.69
###3.24
0.41
0.72
Leverage
*-1.85
###-4.54
-1.14
-1.23
**-2.35
##2.01
ROA
***5.00
###4.62
***3.37
###3.47
0.65
-0.09
Asset Turnover
***2.82
###3.90
-0.92
0.15
***3.71
###-3.33
Profit Margin
***3.23
###3.45
**2.57
###3.46
-0.04
1.26
Growth (TA)
-0.03
-0.09
1.33
#1.88
-1.31
#1.83
EPS
*1.86
###5.85
*1.90
###4.74
-0.21
0.25
BVS
-0.81
###3.29
-0.43
###2.83
-0.39
-0.14
PE
0.08
0.77
-1.48
-0.48
1.29
1.25
PB
0.57
##2.23
1.31
#1.93
-0.99
0.19
Revenue Ratio
0.16
0.60
*1.94
1.63
-1.63
0.77
a***,**,*(###,##,#) The test of the mean (median) difference is statistically significant at the 1%, 5%, or 10% level, respectively, two-tailed, based on a paired t-test (Wilcoxon Rank sum test).
39
TABLE 4 (Continued)
Variable Definitions
Assets
Book value of total assets;
Liabilities
Book value of total liabilities;
Equity
Book value of common equity;
Net Income
Total net income;
Operating Cash Flow
Net cash from operating activities;
Investing Cash Flow
Net cash from investing activities;
MVE
December 31 market value of common equity per share times the number of outstanding shares;
Relative Size
The ratio of the firm’s market value of equity in year t to the total Bolsa market capitalization in the same year.
Beta
The coefficient of variation between the firm’s stock price and the market as reported by the Bolsa;
Returns
Annual percentage change in the market value of common equity;
Analysts
The number of analysts providing estimates in the I/B/E/S international database;
ROE
Return on common equity (net income/equity);
Leverage
Total assets/Equity;
ROA
Return on assets (net income/assets);
Asset Turnover
Sales/Assets;
Profit Margin
Net Income/Sales;
Growth (TA)
The 36-month change in total assets;
EPS
Earnings per share (net income/outstanding common shares);
BVS
Book value per share (equity/outstanding common shares);
PE
Price-earnings ratio (market value of common equity per share/earnings per share);
PB
Price-to-book ratio (market value of common equity per share/book value of common equity per share);
Revenue Ratio
The ratio of domestic revenue to total revenue.
40
TABLE 5
Descriptive Statistics Before and After Cross Listing
Panel A: Financial Statement and Market Variables- Exchange Listed ADR Firmsa
Before
After
Difference Tests
Variablea
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
Tb
(p-value)
Zb
(p-value)
Assets
26
2.896
4.288
1.178
30
3.071
4.492
1.501
-0.15
(0.88)
-0.86
(0.39)
Liabilities
26
1.133
1.524
0.580
30
1.430
1.792
0.842
-0.66
(0.51)
-0.83
(0.41)
Equity
26
1.697
2.957
0.547
30
1.674
2.954
0.490
0.03
(0.98)
-0.42
(0.68)
Net Income
26
0.269
0.536
0.063
30
0.192
0.463
0.050
0.57
(0.57)
1.19
(0.23)
Operating Cash Flow
26
0.367
0.858
0.004
30
0.289
0.924
0.020
0.33
(0.75)
0.06
(0.95))
Investing Cash Flow
26
-0.323
0.543
-0.139
30
-0.318
0.765
-0.46
-0.03
(0.98)
-0.45
(0.65)
a The data are totals in terms of millions of U.S. dollars.
41
TABLE 5 (Continued)
Descriptive Statistics Before and After Cross Listing
Panel B: Performance Measures - Exchange Listed ADR Firms
Before
After
Difference Tests
Variable
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
T
(p-value)a
Z
(p-value)a
ROE
26
12.40
11.25
10.89
30
2.56
22.43
9.96
2.03
*(0.05)
1.08
(0.28)
Assets/Equity
26
1.97
1.19
1.97
30
2.36
1.56
1.83
-1.04
(0.30)
-0.43
(0.67)
ROA
26
5.63
3.88
5.35
30
3.65
7.01
3.98
1.28
(0.21)
1.17
(0.24)
Sales/Assets
26
0.44
0.75
0.17
30
0.88
1.17
0.45
-1.65
*(0.10)
-1.17
(0.24)
Net Income
/Sales
19
5.49
0.22
0.24
25
0.61
1.44
0.10
1.55
(0.13)
1.80
#(0.07)
Revenue Ratio
26
0.83
0.22
0.92
30
0.79
0.27
0.86
0.75
(0.46)
0.77
(0.44)
aRefer to table 4 for variable definitions.
b*(#) The test of the mean (median) difference between Non-ADR and exchanged-listed firms is statistically significant at the 10% level or better, two-tailed, based on a paired t-test (Wilcoxon Rank sum test).
42
TABLE 6
Debt and Equity Issuance
Debt
Equity
Classification
No. of Issues
Amount ($mil)
No. of Issues
Amount ($mil)
NCL
Total (number and amount)
3
315.2
14
4579.6
Average
105.1
327.1
Median
35.3
117.8
Standard Deviation
129.7
554.1
144a
Total (number and amount)
0
0.0
3
54.9
Average
18.3
Median
13.8
Standard Deviation
8.5
OTC
Total (number and amount)
1
165.9
4
215.9
Average
165.9
54.0
Median
165.9
41.6
Standard Deviation
49.4
Level 2 ADR
Total (number and amount)
0
0.0
13
1450.2
Average
111.6
Median
119.4
Standard Deviation
68.0
Level 3 ADR
Total (number and amount)
8
3603.9
29
8802.3
Average
450.5
303.5
Median
160.7
85.0
Standard Deviation
671.2
441.1
43
TABLE 7
Estimate of U.S. Cross Listing Probability during 1995-1999
Variable (Coefficient)
Exchange-Listed ADR Firms (n=128)
vs
Non-Crosslisted Firms (n=61)
Exchange-Listed ADR Firms (n=128)
vs
OTC and 144a Firms (n=47)
OTC and 144a Firms (n=47)
vs
Non-Crosslisted Firms (n=61)
Coefficient(Chi-square)
Coefficient (Chi-square)
Coefficient (Chi-square)
Intercept (α0)
-4.98 (11.97)***
3.31 (10.56)
-17.02 (19.75)***
MVES (β1)
0.02 (5.82)**
-0.02 (3.66)*
0.06 (8.05)***
ROA (β2)
-0.97 (0.09)
-6.20 (5.51)**
12.02 (2.73)*
LEVERAGE (β3)
1.75 (22.67)***
0.00 (0.00)
3.46 (17.30)***
GROWTH (β4)
-0.00 (0.00)
-0.00 (2.15)
0.00 (0.37)
REVRATIO (β5)
3.29 (9.52)***
-1.93 (3.74)*
11.25 (16.15)***
INDUSTRY (β6)
-0.07 (3.96)**
0.01 (0.05)
-0.14 (5.88)**
Model Statistics
Concordant Percent
80.80
72.60
89.00
Likelihood Ratio χ2 (p-value)
54.00 (<0.00)
19.90 (0.00)
60.93 (<0.00)
Pseudo R2
24.85%
10.75%
43.12%
Statistically significant coefficients (two-tailed test) are denoted by: *** less than .01 ** less than .05 * less than .10.
There are 32 firms listed on the NYSE, NASDAQ, or AMEX, ten firms traded over-the-counter, five firms traded via rule 144a and 25 firms that are not cross listed in the U.S. during the 1995-1999 period, but are eligible to list on the NYSE. P(ADR) ≡ Probability of cross listing in the U.S.; MVES is the market value of equity per share; ROA is the ratio of net income to total assets; LEVERAGE is the ratio of total assets to stockholders’ equity; GROWTH is the 36-month percentage change in total assets; REVRATIO is the ratio of domestic revenue to total revenue; and INDUSTRY is a dummy variable taking on the value of one when it is equal to category n (defined in table 3) and zero otherwise.
44
TABLE 8
Relation between Annual Stock Prices and Summary Financial Statement Measuresa
ititBVSitEPSitMVSεγγγ+++=210
Variable
Coefficient
Pooled
1995
1996
1997
1998
1999
Panel A: ADR Sample
Obs
132
21
22
24
28
29
Intercept
γ0
2.69
**(2.42)
3.91
*(1.76)
3.84
*(1.72)
1.52
(1.19)
2.41
(1.47)
1.11
(0.55)
EPS
γ1
2.71
*(1.66)
12.63
(1.39)
5.28
(0.93)
4.40
**(2.98)
3.01
(0.62)
-9.80
(-1.41)
BVS
γ2
5.98
***(15.49)
4.39
***(4.07)
3.94
**(2.86)
8.28
***(12.44)
4.54
***(5.95)
7.53
***(5.28)
Adjusted R2
0.71
0.69
0.60
0.88
0.55
0.64
Panel B: OTC & 144a Sample
Obs
47
2
6
10
13
12
Intercept
γ0
9.11
***(3.96)
NA
16.21
*(2.27)
9.78
**(2.65)
10.39
**(2.79)
3.34
(0.97)
EPS
γ1
30.45
***(4.13)
NA
131.43
(1.26)
67.09
***(4.07)
50.02
**(2.83)
66.81
***(3.96)
BVS
γ2
3.66
**(2.61)
NA
-29.10
(-1.09)
-3.42
(-0.97)
1.29
(0.62)
-1.63
(0.72)
Adjusted R2
0.61
NA
0.24
0.83
0.47
0.83
Panel C: Non-Crosslisted Sample
Obs
71
8
16
16
14
13
Intercept
γ0
13.49
***(5.88)
-1.94
(-1.02)
15.41
**(2.52)
12.36
**(2.40)
13.19
**(2.12)
19.35
**(3.43)
EPS
γ1
-2.76
(-0.36)
12.75
(1.43)
2.79
(0.16)
-25.90
(-1.38)
78.12
(1.12)
29.24
(0.53)
BVS
γ2
0.74
(0.48)
12.57
**(3.99)
-0.35
(-0.13)
4.52
(1.19)
-14.56
(-0.94)
-6.15
(-0.83)
Adjusted R2
NA
0.91
NA
NA
0.05
NA
aEach estimated coefficient is expected to be positive. MVS ≡ December 31 price, BVS ≡ Book value per share, and EPS ≡ Net income per share. T-statistics are in parentheses. Statistically significant coefficients (two-tailed test) are denoted by: *** less than .01 ** less than .05 * less than .10.
45
TABLE 9
Quality of Earnings Analysis
Panel A: Descriptive Statisticsa
Non-Crosslisted Firms
OTC & 144a Firms
Exchange Listed ADR Firms
Variable
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
N
Mean
Std Dev
Median
Accruals
73
-0.037
0.054
-0.040
47
-0.020
0.076
-0.033
135
-0.043
0.102
-0.035
Cash Flows from Ops
73
0.118
0.052
0.130
47
0.095
0.080
0.110
135
0.080
0.111
0.078
Panel B: Spearman Correlations among Cash Flows and Accruals (1995-1999)a
Group
Coefficient
Z-statistic
(p-value)
Non-Crosslisted Firms
-0.75
Exchange-Listed Firms
-0.54
-1.92 (0.05)
OTC & 144a Firms
-0.67
Exchange-Listed Firms
-0.54
0.24 (0.81)
Non-Crosslisted Firms
-0.75
OTC & 144a Firms
-0.67
1.37 (0.17)
Variable Definitions
Accruals
The difference between net income and cash flows from operations scaled by total assets
Cash flows from operations
Net operating cash flows scaled by total assets 46
Εγγραφή σε:
Αναρτήσεις (Atom)