The impact of ADR activity on stock market liquidity: Evidence from Latin America

G Model

ARTICLE IN PRESS

QUAECO-769; No. of Pages 11

The Quarterly Review of Economics and Finance xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

The Quarterly Review of Economics and Finance journal homepage: www.elsevier.com/locate/qref

The impact of ADR activity on stock market liquidity: Evidence from Latin America Alma D. Hales a,∗ , André V. Mollick b,1 a b

Tennessee Tech University, Department of Economics, Finance and Marketing, 1 William L. Jones Drive, Cookeville, TN 38505, United States University of Texas-Pan American, Department of Economics and Finance, 1201W. University Drive, Edinburg, TX 78539, United States

a r t i c l e

i n f o

Article history: Received 27 March 2013 Received in revised form 11 February 2014 Accepted 9 March 2014 Available online xxx Keywords: ADR activity Latin America Liquidity United States

a b s t r a c t This paper examines the impact of ADR activity on liquidity of four major Latin American stock markets. We construct a measure of ADR activity in U.S. markets for a sample of ADRs trading during January 2003–December 2010, which is subsequent to the financial liberalization episodes and currency crises that shocked emerging markets in the 1990s. The sample lists 164 depositary receipt programs (Levels I, II, and III): 16 from Argentina, 81 from Brazil, 19 from Chile, and 48 from Mexico. Using System GMM methods to handle the potential effects from stock market development on economic growth and ADR issuance, we find that higher ADR turnover in U.S. markets has positive effects on domestic market turnover, particularly for issuance of exchange-listed (Levels II and III) ADRs. This positive relationship is not a statistical artifact created by the global financial crisis of 2008. © 2014 The Board of Trustees of the University of Illinois. Published by Elsevier B.V. All rights reserved.

1. Introduction This paper examines the impact of American Depositary Receipt (ADR) activity on the liquidity of four Latin American stock markets (Argentina, Brazil, Chile and Mexico) during the period 2003–2010. ADRs are negotiable instruments issued in the U.S., in dollars, but they represent ownership of foreign equities. ADRs offer U.S. investors familiar trade, clearance and settlement procedures in addition to competitive foreign-exchange rates on currency conversions for dividends and other cash distributions. More importantly, an ADR investment offers U.S. investors international diversification benefits as pointed out by Officer and Hoffmeister (1987), Wahab and Khandwala (1993), Choi and Kim (2000), Alaganar and Bhar (2001), and Arnold, Nail, and Nixon (2004). These benefits help explain the consistent rise in U.S. investment in ADRs from Latin America since 2003, except for a major contraction associated with the global financial crisis of 2008. The role of ADRs as convenient vehicles for U.S. investors seeking to hold foreign equity is particularly important in Latin America. For instance, according to the U.S. Treasury International Capital

∗ Corresponding author. Tel.: +1 931 372 3717; fax: +1 931 372 6249. E-mail addresses: [email protected], [email protected] (A.D. Hales), [email protected] (A.V. Mollick). 1 Tel.: +1 956 665 2494; fax: +1 956 665 5020.

system, 56% of U.S. holdings of Latin American equity, in 2007, were in the form of ADRs. This percentage is significantly higher than any other region in the world—Europe and Africa (20%) and Asia (14%). In addition, as reviewed by Karolyi (1998, 2006), ADRs also benefit their respective issuers because they offer an expanded shareholder base, higher liquidity, higher global visibility, and a lower cost of capital. Despite the documented benefits of flourishing U.S. investment in ADRs from Latin America and the relative macroeconomic stability in Latin American economies, De la Torre and Schmukler (2007) find that Latin American stock markets are smaller and less active than similar developing economies. The disparity between still flourishing ADR activity and dampened stock market conditions in Latin America generates the following question: Has growth in ADR activity inhibited the development of Latin American stock markets? The empirical evidence for earlier periods is mixed. Fernandes (2009) finds that despite the negative effects of ADR issuance on the liquidity of domestic stocks, the net impact of ADR issuance is positive, particularly for firms whose returns are highly correlated to those of the cross-listing firm. Moel (2001) and Levine and Schmukler (2007) find, however, that ADR growth is detrimental to the development of the domestic stock market, while Karolyi (2004) suggests that ADR activity does not hurt or benefit stock market development: it is an outcome of poor domestic conditions. These mixed results reflect the dual outcomes associated with ADR growth. On the one hand, ADR activity increases the liquidity, visibility and shareholder base of issuing firms, with the

http://dx.doi.org/10.1016/j.qref.2014.03.005 1062-9769/© 2014 The Board of Trustees of the University of Illinois. Published by Elsevier B.V. All rights reserved.

Please cite this article in press as: Hales, A. D., & Mollick, A.V. The impact of ADR activity on stock market liquidity: Evidence from Latin America. The Quarterly Review of Economics and Finance (2014), http://dx.doi.org/10.1016/j.qref.2014.03.005

G Model QUAECO-769; No. of Pages 11

ARTICLE IN PRESS A.D. Hales, A.V. Mollick / The Quarterly Review of Economics and Finance xxx (2014) xxx–xxx

2

potential to boost interest and confidence in the remaining domestic market stocks, thus spurring development. On the other hand, these potential benefits must be balanced against the costs of crosslisting, particularly the migration of trade from the domestic to the foreign market. Trade migration to the foreign market leads to a deterioration of local market operations causing more domestic firms to seek opportunities in foreign markets, further inhibiting development at home, other things the same. This paper contributes to the existing literature in the following ways. First, the analysis measures ADR activity with data from the ADR trading in the United States. Most existing literature measures ADR activity using data from the domestic market, but this approach is limited. Specifically, the traditional approach can overstate the importance of the ADR segment if the ADRs are dormant in U.S. markets (typically ADRs trading over- the- counter (OTC)) and understate the importance of the ADR segment for those ADRs for which the majority of trading has shifted to the U.S. (exchangelisted issues). By measuring U.S. trading activity, we update the quantification of ADR liquidity in four stock markets: Argentina, Brazil, Chile and Mexico. Second, this study distinguishes between types of ADRs (exchange listed issues versus OTC issues) with an additional distinction for exchange listed issues into Level II (straight cross-listings) and Level III (capital raising issues). Since Level III ADRs are typically issued by large, high-growth firms, as documented by Boubakri, Cosset, and Samet (2010), they can have a different impact on stock market development than ADRs aimed primarily at broadening the shareholder base. Third, this study incorporates dynamic panel data models to examine the relationship between ADR activity and stock market liquidity. Claessens, Klingbiel, and Schmukler (2006), for example, suggest that countries with better fundamentals allow for more internationalization. ADR activity may simultaneously affect and be an outcome of stock market development, which can be better handled by dynamic panel data models than static panels. Using system generalized method of moments (SGMM) for firmlevel data of 164 major Latin American ADRs between January 2003 and December 2010, the results indicate a positive impact on stock market turnover (measured by the value of shares traded scaled by market capitalization, as explained below) that arises as the turnover of the ADR segment trading in the United States increases. This finding suggests that failing to account for the ADR segment’s activity in the foreign market provides an incomplete picture of the role of ADR activity in domestic stock market development. Furthermore, we find that the positive effects of ADR activity on domestic market turnover arise primarily from the issuance of exchange-listed ADRs. Our results show a positive and statistically significant relationship between ADR turnover and market turnover for ADRs of Levels II and III but not for Level I ADRs. Finally, our results show that turbulence in financial markets during the peak of the global financial crisis impacted the relationship between ADR activity and market liquidity. However, the positive relationship between ADR turnover in U.S. markets and the turnover of Latin American markets is not a statistical artifact driven by the effects of the crisis. The remainder of this paper is organized as follows: Section 2 reviews the related literature. Section 3 describes the sample and measures employed. Section 4 details the methodology while Section 5 presents the empirical results. Section 6 concludes.

market development in one of two ways. First, if cross-listings act as market liberalization events, firms that cross-list attract global attention and bring increased visibility, credibility and enhanced liquidity to other local market stocks. Consequently, local financial intermediaries feel competitive pressure from global markets and begin improving the efficiency of trading systems, through greater transparency and more stringent disclosure requirements. Ultimately, this leads to integration with global markets, resulting in higher economic development and growth. Alternatively, if crosslistings divert trading away from the local market they negatively impact the market’s quality. In this case, the benefits of cross-listing accrue only to the large, cross-listing firms which enjoy greater visibility and enhanced liquidity but their strictly domestic peers suffer as the local stock market deteriorates. The existing evidence offers mixed results on the effects of international cross-listing on local stock market development. Fernandes (2009), for example, documents that despite the negative effects of trade migration, the net impact of ADR issuance, particularly (but not limited to) a country’s first ADR issue, is positive with spillover effects for other domestic firms. This suggests that ADR issuance acts as a firm-level liberalization effect: the positive spillover effects arise from improved risk sharing and accrue primarily to domestic firms whose returns are highly correlated with the cross-listing firm. However, Moel (2001) finds mixed results with respect to the impact of ADR issues on stock market development using annual data for a sample of 28 countries during the period 1988–1997. The results for three aspects of stock market development (liquidity, growth, and openness) suggest that ADRs negatively affect liquidity and growth but they increase disclosure and openness. The negative impact of ADR growth is higher for Africa and Latin America. Karolyi (2004) extends this analysis using monthly, firm-level data for twelve emerging markets and considers four measures of stock market activity: market capitalization, number of listed companies, turnover, and capital flows. His results suggest that ADR growth increases cross-border capital flows and stock market development, but only large, cross-listing firms reap the benefits. The remainder of the market (firms that do not issue ADRs) deteriorates. Levine and Schmukler (2007) adopt a different approach by examining the impact of internationalization (including ADR issues, cross-listing on a foreign exchange, and issuing capital abroad) on the domestic stock market activity of 55 countries. Their findings indicate that internationalization is negatively related to the turnover of domestic firms due to several factors, including trade migration to international financial markets. The empirical evidence on the impact of ADR activity on domestic stock market development is inconclusive. One important limitation of the approaches above is that internationalization is treated as exogenous to stock market development. Yet exogeneity is highly unlikely since internationalization and stock market development are not independent processes. While the level of internationalization may impact stock market development, it is also a consequence of different levels of development. For instance, Claessens et al. (2006) use annual data for 78 countries during the period 1984–2000 and find that countries with better economic fundamentals have greater stock market development and more internationalization.

2. Related literature

3. The sample

Because ADR issues increase the international exposure of domestic firms, they can potentially impact the stock market. As discussed by Karolyi (2006), ADR activity can influence stock

We begin the sample construction with the listing of ADR programs (only Levels I, II, and III, active and terminated) from Argentina, Brazil, Chile and Mexico available from DataStream

Please cite this article in press as: Hales, A. D., & Mollick, A.V. The impact of ADR activity on stock market liquidity: Evidence from Latin America. The Quarterly Review of Economics and Finance (2014), http://dx.doi.org/10.1016/j.qref.2014.03.005

G Model

ARTICLE IN PRESS

QUAECO-769; No. of Pages 11

A.D. Hales, A.V. Mollick / The Quarterly Review of Economics and Finance xxx (2014) xxx–xxx

3

Table 1 Descriptive statistics: stock market development in Latin America and the U.S. Capitalization

Value of shares traded

$U.S. Billions

$U.S. Billions

% Of GDP, value traded

% Of capitalization, turnover

Argentina

2004 2006 2008 2010

40.59 51.24 39.85 63.91

4.84 5.28 6.64 3.81

3.16 2.46 2.02 1.03

11.92 10.30 16.67 5.95

Brazil

2004 2006 2008 2010

330.35 710.25 591.97 1545.57

104.56 275.93 750.25 868.81

15.75 25.33 45.38 41.58

31.65 38.85 126.74 56.21

Chile

2004 2006 2008 2010

116.92 174.42 131.81 341.80

11.96 27.97 35.64 53.82

12.70 19.18 20.67 26.38

10.23 16.03 27.04 15.75

Mexico

2004 2006 2008 2010

171.40 348.35 234.05 454.35

45.07 95.65 112.80 119.12

5.93 10.05 10.21 11.50

26.30 27.46 48.19 26.22

United States

2004 2006 2008 2010

16,323.73 19,425.85 11,737.65 17,138.98

19,354.90 21,509.98 33,267.64 64,014.24

164 250 450 211

118.57 110.73 283.43 373.50

Notes: Annual stock market capitalization and value of shares traded (in current $U.S. billions) for Latin American markets are calculated from monthly data extending from January 2003 to December 2010. Values are computed at year end. For the Latin American markets, stock market data are from the World Federation of Exchanges and GDP data are from the International Monetary Fund’s Principal Global Indicators and International Financial Statistics databases. Annual data for the United States are from the World Bank.

International at the end of 2010.2 Level I ADRs are traded in overthe-counter (OTC) exchanges while Levels II and III ADRs are traded on the New York Stock Exchange (NYSE) or NASDAQ. Furthermore, Levels II and III ADRs differ significantly in purpose. The former are issued with the intent to meet U.S. investor demand for foreign equity while the latter raise capital in the U.S. market. We include only ADRs trading during the period January 2003 to December 2010 and whose activity is verifiable through the Bank of New York’s, Citibank’s, JP Morgan Chase’s and/or Deutsche Bank’s depositary receipt databases.3 ADRs with no verifiable underlying stock, and stocks with missing or distorted price and/or volume data are excluded. In cases where two (or more) ADRs correspond to the same underlying stock, the underlying stock is included only once. These filters result in 164 ADRs included in the sample: 16 from Argentina; 81 from Brazil; 19 from Chile; and 48 from Mexico.

2 While London and Frankfurt have become important outlets for cross-listings more recently, we utilize only U.S. trading data in the construction of the ADR activity variable due to the relevance of New York for Latin American countries, and for equity markets in particular. For simplicity the text references only ADRs but to obtain a complete picture of the impact of depositary receipt activity on the liquidity of the domestic stock market, Global Depositary Receipts (GDRs) with trading activity in the United States are also included. These GDRs are classified as Level III ADRs when they raise new capital and as Level II ADRs when they are straight crosslistings. Other types of ADRs include: private placements under Rule144A (in which shares are placed amongst qualified institutional buyers (QIBs)) and Regulation S issues in which capital is raised outside the United States. Our focus is on ADRs of Levels I, II, and III because trading data on Rule 144A and Regulation S issues are limited. 3 Where DataStream International data and depositary receipt databases differ, the New York Stock Exchange, Over-the-Counter Bulletin Board (OTCBB) and the Electronic Data Gathering, Analysis and Retrieval (EDGAR) system of the SEC verify the relevant data.

3.1. Measuring stock market development Our analysis focuses on the liquidity of Latin American stock markets. To capture the depth of the market, we use Value Traded which is the total value of shares traded in the stock market each month, scaled by GDP. This measure captures trading activity in the market given the size of the country’s economy. A potential problem with this variable is that it ignores the size of the stock market. A small but active market is better developed than a large, inactive market; yet both will exhibit low Value Traded. To account for this shortcoming, we include a second liquidity measure, Market Turnover, which is the total monthly value of shares traded in the stock market scaled by the market capitalization. This yields information about the liquidity of the market given its size. Table 1 shows an overview of stock market conditions in Latin America and the United States during 2003–2010. The values are calculated on an annual frequency to highlight general patterns of development. For the Latin American countries, we obtain stock market values from the World Federation of Exchanges and GDP data from the International Monetary Fund’s Principal Global Indicators and International Financial Statistics databases. We obtain all U.S. data from the World Bank. Table 1 suggests that Latin American stock markets are underdeveloped with respect to both size and liquidity. Both the capitalization and value traded of the four Latin American markets combined fall significantly below values for the U.S. market. Table 1 also displays average statistics for Value Traded. The depth of U.S. markets is still evident as the value of shares traded exceeds GDP across all years presented. This pattern does not hold for Latin American markets, but average Value Traded indicates that the most liquid stock market in Latin America is Brazil followed by Chile, Mexico and Argentina. To evaluate the value of shares traded relative to the size of the stock market, we employ Market Turnover, which is shown in the last column of Table 1. This variable suggests a slightly different pattern than Value Traded regarding the liquidity

Please cite this article in press as: Hales, A. D., & Mollick, A.V. The impact of ADR activity on stock market liquidity: Evidence from Latin America. The Quarterly Review of Economics and Finance (2014), http://dx.doi.org/10.1016/j.qref.2014.03.005

G Model

ARTICLE IN PRESS

QUAECO-769; No. of Pages 11

A.D. Hales, A.V. Mollick / The Quarterly Review of Economics and Finance xxx (2014) xxx–xxx

4

Table 2 Descriptive statistics: ADR activity measures for Latin America. Capitalization $U.S. Billions

% Of market

Value traded domestically

Value traded in the U.S.

$U.S. Billions

$U.S. Billions

% Of market

% Of capitalization

Argentina

2004 2006 2008 2010

24.391 31.632 29.853 27.273

67.0 68.2 58.9 56.6

0.143 0.117 0.125 0.125

35.8 26.2 22.5 37.4

0.196 0.339 0.463 0.408

0.79 1.07 1.51 1.44

Brazil

2004 2006 2008 2010

127.708 319.217 598.540 842.915

52.6 52.7 52.4 63.3

5.391 13.262 38.202 46.530

62.3 57.9 61.6 64.2

5.279 19.490 70.505 61.520

4.20 6.14 12.33 7.38

Chile

2004 2006 2008 2010

28.458 44.500 56.719 75.218

30.6 29.9 30.5 27.4

0.437 0.895 1.297 1.505

44.0 38.9 45.5 35.7

0.350 0.555 1.695 1.559

1.22 1.23 3.0 2.1

Mexico

2004 2006 2008 2010

94.949 193.336 202.727 227.120

65.3 70.2 57.0 59.5

2.495 4.727 6.461 6.338

67.4 59.6 69.1 64.0

3.789 8.996 15.916 9.342

4.02 4.73 7.92 4.21

Notes: Average monthly data on ADR activity in current U.S. dollars for the period 2003–2010. Includes all Level I, II, and III depositary receipt programs trading between 2003 and 2010, available from DataStream International and verifiable via the Bank of New York, Citibank, or Deutsche Bank. Monthly ADR capitalization is obtained by adding the market capitalization of all the ADR issues identified on a daily basis and taking the monthly average of these sums. The ADR value traded is obtained by first adding the daily value traded of each ADR identified in the sample (number of shares traded multiplied by the closing price (converted to U.S. dollars at the corresponding closing spot rate)) to obtain the daily value traded of the ADR segment, then the value traded of the ADR-segment is added across all trading days of each month. Closing prices, trading volume, and exchange rate data are from DataStream International. Data on total market capitalization and value of shares traded are from the World Federation of Exchanges.

across Latin America. Brazil continues to exhibit the highest levels of liquidity, but now it is followed by Mexico, then Chile and finally Argentina. The difference indicates a large number of listed but inactive stocks in the Chilean market. The statistics also show high values of Market Turnover across all countries in 2008, consistent with the global financial crisis. This supports the findings of Levy Yeyati, Schmukler, and Van Horen (2008), who document that emerging markets exhibit large price downturns and higher trading activity during crises. Overall, the summary statistics suggest that liquidity was on the rise in Brazil, Chile and Mexico before the shocks associated with the global financial crisis, but conditions after the crisis surpassed their pre-crisis levels only in Brazil. However, for Argentina liquidity conditions appear to deteriorate consistently through the sample period. 3.2. Measuring ADR activity Existing literature, such as Karolyi (2004), Moel (2001), and Claessens et al. (2006) has measured ADR activity by examining a stock market’s share of ADR listings, share of market capitalization and share of value traded domestically captured by ADR issuing firms. However, the benefits of ADR issuance do not arise by the mere issuance of the security. An ADR issue that is not traded by U.S. investors may as well be non-existent.4 In this study, we therefore focus on measuring ADR activity using information on the ADR segment’s liquidity in U.S. markets. For comparison, in Table 2 we provide descriptive statistics for liquidity of the ADR segment at

4 Consider the following possibility: First, an actively traded firm in Mexico taps the ADR market through a Level I ADR. The ADR trades for a few days upon issuance then becomes dormant in the U.S. market. The underlying stock remains actively traded. Measuring liquidity of the ADR segment using data only on the underlying stock would overstate the importance of the ADR issue. Alternatively, certain ADR issues from the four Latin American countries considered here, particularly exchange-listed programs, exhibit higher dollar trading volumes in the U.S. market than in the respective home market. For these firms, measuring liquidity of the ADR segment using only data on the underlying stock understates the importance of the ADR issue.

home and abroad, which we discuss below in conjunction with Fig. 1. To examine the liquidity of the ADR segment, we begin by obtaining the monthly value traded in the home and U.S. markets. The daily value traded of each ADR in the sample is the product of: the number of shares traded and the closing price each day (converted to U.S. dollars at the corresponding closing spot rate for shares trading in the home market). To obtain the monthly value, the daily value traded is aggregated across firms and across all trading days of each month. DataStream International provides all data necessary for calculating value traded of the ADR segment. Fig. 1 displays the trends in the monthly value traded of the stock market and the ADR segment. The figure shows that the ADR segment composes a significant amount of value traded in the domestic market. These patterns are confirmed in Table 2. In Brazil and Mexico domestic value traded is dominated by ADR-related stocks (by Column (4), more than 50% in all years). We find slightly lower values in the remaining countries: the ADR segment constitutes approximately 30% of the stock market in Argentina and 40% in Chile.5 More importantly, Fig. 1 highlights the difference in activity between the ADR trading in the U.S. and that of its corresponding stock trading in the home market. In Argentina, Brazil, and Mexico, the dollar trading volume of ADRs exceeded that of the underlying stocks through most of the sample period. The value of Chilean ADRs traded in the U.S. trailed that of the underlying stocks trading in the Chilean market during the early part of the sample period but surpassed it by 2008. Furthermore, Fig. 1 suggests that in several cases the value of ADRs traded in the U.S. exceeded the value of shares traded in the entire domestic market especially in

5 Overall, these summary statistics coincide with those reported for the year 2000 in Karolyi (2004). Regarding ADR Capitalization (% market), the means for this sample follow the same pattern but are smaller in magnitude. For ADR Value Traded Domestically (% market), the patterns are also similar except in Argentina. Because the sample in Karolyi (2004) stops in 2000 and the sample in this study begins in 2003, the difference likely reflects substantial changes occurring in the stock market as a result of the Argentinian crisis of 2002.

Please cite this article in press as: Hales, A. D., & Mollick, A.V. The impact of ADR activity on stock market liquidity: Evidence from Latin America. The Quarterly Review of Economics and Finance (2014), http://dx.doi.org/10.1016/j.qref.2014.03.005

G Model QUAECO-769; No. of Pages 11

ARTICLE IN PRESS A.D. Hales, A.V. Mollick / The Quarterly Review of Economics and Finance xxx (2014) xxx–xxx

5

Fig. 1. Value traded of cross-listed companies in Latin America and the United States ($U.S. billions). Notes: Average monthly data in current U.S. dollars for the period 2003–2010. Each graph displays the value of stocks traded in the stock market (Market), the value traded of the ADR segment in the United States (ADR (U.S.)) and the value traded market of the ADR segment’s underlying stocks in the domestic market (ADR (Home)). Data on the total value of stocks traded in the market are from the World Federation of Exchanges. Value traded for ADRs and their underlying stocks are based on author’s calculations using price and volume data from DataStream International. The ADR value traded in the domestic market is converted to U.S. dollars using the WM/Reuters closing spot rates. Only Levels I, II, and III ADRs are included.

Mexico where the relationship was observed through most of the sample period. Given the importance of trading in the U.S., we construct a variable that captures the liquidity of the ADR segment in the U.S. market. In Table 2, we also report data on the capitalization of the ADR segment. The values suggest that the largest ADR segment is from Brazil followed by Mexico, Chile and Argentina. Consistent with the value of shares traded, the ADR segment comprises a significant portion of the total market capitalization—over 50% in Argentina, Brazil, and Mexico and approximately 30% in Chile. We measure the liquidity of the ADR segment relative to its size so we construct the variable ADR Turnover which measures the value of ADR shares traded in the U.S. scaled by the market capitalization of the ADR segment.6 Table 2 suggests that the patterns for liquidity of the ADR segment in the U.S. resemble the liquidity patterns of the domestic markets. On average, based on ADR Turnover, the most liquid ADR segment was that of Brazil followed by Mexico, Chile and Argentina. In addition, across each Latin American country, ADR Turnover was

6 Constructing the monthly capitalization of the ADR segment requires adding the capitalization of all the identified ADR issues on a daily basis. Then the monthly average of these sums yields a monthly measure of ADR-related capitalization. Although ADRs represent claims to the same cash flows as their underlying stocks, because the two trade in different markets, their prices, even after adjusting for foreign currency conversions, are not always equal, as shown by Grossman et al. (2007), Chan, Hong, and Subrahmanyam (2008), Eichler, Karmann, and Maltritz (2009), and Gagnon and Karolyi (2010). Graphical inspection suggests that the capitalization of ADR-issuing firms does not differ substantially when measured using U.S. prices versus home market prices. To be consistent with existing literature, we use home share prices for calculating the capitalization of the ADR segment in the empirical analysis.

highest during 2008 which may reflect sell-offs during the 2008 financial crisis. However, unlike the patterns for domestic market liquidity, ADR liquidity appears to be on the rise throughout the sample period in all countries except Mexico where ADR Turnover is lower in 2010 than it was in 2006. 4. Methodology An important characteristic of the stock market and ADR segment activity variables is that their construction includes stock prices. Since stock prices generally follow a random walk, there is a possibility that the constructed series have unit roots. We conduct the Im, Pesaran and Shin (2003) panel unit root tests on value traded, turnover, and ADR turnover. The tests reject the null hypothesis that all panels contain unit roots for each of the series. We then conduct the Hadri (2000) LM tests of stationarity which test the opposite null hypothesis, namely that all panels included are stationary. These tests also reject the null that all panels are stationary. Therefore, these results suggest that some, but not all, of our panels contain a unit root. We also apply Augmented Dickey Fuller (ADF) tests on individual series. Consistent with the panel results, individual ADF tests are unable to reject the null hypothesis that the series have a unit root for value traded in all countries except Argentina. However, the ADF tests reject the nonstationarity of Market Turnover in all countries except Chile. The tests also reject the null of a unit root in ADR Turnover in all countries except Mexico.7

7

Panel and individual unit root tests are available from the authors upon request.

Please cite this article in press as: Hales, A. D., & Mollick, A.V. The impact of ADR activity on stock market liquidity: Evidence from Latin America. The Quarterly Review of Economics and Finance (2014), http://dx.doi.org/10.1016/j.qref.2014.03.005

G Model

ARTICLE IN PRESS

QUAECO-769; No. of Pages 11

6

A.D. Hales, A.V. Mollick / The Quarterly Review of Economics and Finance xxx (2014) xxx–xxx

To account for the persistence of the market liquidity variables the econometric model includes lagged stock-market liquidity as an explanatory variable. Since we examine the relationship between ADR activity and stock market liquidity and its variation across countries, our empirical strategy attempts to maximize the use of both the time and cross-country dimensions of the available data. The dynamic equation takes the form:  yi,t = ˇ0 + ıyi,t−1 + ˇ1 xi,t + zi,t + ui,t ,

(1)

where yi,t represents Value Traded or Market Turnover, xi,t represents ADR Turnover, zi,t is a vector of control variables, and ui,t is an error term that contains country and time specific fixed effects as follows: ui,t = i + εt + i,t , where the i,t are assumed to be independent and identically distributed with mean zero and variance 2 . The impact of international cross-listings on domestic market activity is ambiguous. Cross-listings can act as market liberalization events such that cross-listing firms attract global attention and bring increased visibility, credibility and enhanced liquidity to other local market stocks. Hargis (2000) predicts that after cross-listing the market capitalization and liquidity of the domestic market should increase.8 This coincides with the overview of the evolution of foreign participation (including ADR issues) and the corresponding stock market development of Latin American stock markets provided by Hargis (1998). In addition, more recent evidence in Halling, Pagano, Randl, and Zechner (2008) indicates that domestic turnover increases in the year of cross-listing and remains higher for cross-listing firms from developed countries. Therefore, the domestic market appears to benefit from the increased competition between foreign and domestic market makers and the additional information generated in the foreign market. Ultimately, this suggests that increases in ADR Turnover lead to higher liquidity in the home market or equivalently: from Eq. (1), ˇ1 > 0. However, Hargis and Ramanlal (1998) suggest that the benefits of internationalization on domestic market development depend on a complex interaction of variables including information transparency and the cross-listing firm’s ability to expand its shareholder base. Consistent with this view, Domowitz, Glen, and Madhavan (1998) use Mexican data to document that the impact of cross-listing on the liquidity and volume traded in the home market varies based on the foreign ownership restrictions of the share class. Internationalization can even be detrimental to the home market; if order flow migrates to the foreign market, liquidity and trading volume decrease at home. In the case of trade migration, the benefits of cross-listing accrue only to the large, cross-listing firms which enjoy greater visibility and enhanced liquidity but their strictly domestic peers suffer as the local stock market deteriorates. The empirical findings of Moel (2001) and Karolyi (2004) support this notion as they document that increases in ADR activity negatively affect the size and liquidity of non-ADR stocks. Furthermore, Levine and Schmukler (2007) document an inverse relationship between internationalization and the turnover of strictly domestic firms. The authors identify two transmission channels for the adverse effects of trade migration: spillover effects and trade diversion (a compositional shift in the domestic market as investors switch away from purely domestic firms into international firms increasing the relative importance of international firms as a share

8 Hargis (2000) also predicts that after cross-listing, the price and liquidity of a stock should increase. The empirical evidence supports this prediction; firms that issue ADRs experience the benefits of integration: higher trading volume, higher liquidity and a lower cost of capital: i.e., Foerster and Karolyi (1998), Karolyi (2004), and Miller (1999).

of the entire market).9 If the trade migration channel is operative, then the impact of ADR Turnover on domestic market liquidity should be negative: from (1), ˇ1 < 0. Eq. (1) also includes a set of control variables in the vector zi,t such as economic growth, inflation, trade openness, the size of each country’s stock market and the existence of any capital controls. While existing evidence in King and Levine (1993a, 1993b) and Levine and Zervos (1998) suggests that financial development leads to positive economic growth, there is the existence of bidirectional causality between financial development and growth. Financial development influences economic growth, but economic growth will affect a country’s financial development. As the economy grows, the financial sector will evolve to respond to the changing demands of the real sector. Therefore, economic growth should be positively related to stock market development. The monthly growth rate of each country’s Industrial Production Index (IP) controls for changes in economic growth. While GDP per capita is the typical control for macroeconomic development, since GDP data are only available on a quarterly basis, we employ the Industrial Production Index which is reported every month. In addition, the empirical approach controls for the impact of inflation because financial contracting is difficult in environments of high inflation since future real values are unpredictable. Even in low or moderate inflationary environments, inflation is problematic because it drives down the real rate of return on assets. This increases credit market frictions that result in credit rationing which can potentially distort financial market performance as shown by Boyd, Levine, and Smith (2001). Therefore, inflation, as measured by changes in each country’s Consumer Price Index (CPI) should be negatively associated with stock market development. Our analysis also controls for the potential impact of trade openness on stock market development. One strand in the literature argues that trade openness and financial openness are simultaneously necessary to deliver financial development: Rajan and Zingales (2003) and Baltagi, Demetriades, and Law (2009). Others argue that trade openness is a prerequisite for financial liberalization, such as Chinn and Ito (2006). Either way, trade openness should be positively associated with stock market development. The natural log of the monthly sum of exports and imports (scaled by GDP) captures trade openness. We also control for characteristics of the four Latin American stock markets, namely size and regulatory restrictions that may potentially affect their liquidity.10 We measure market size using the log of each market’s capitalization and expect larger markets to be more liquid. We also consider restrictions on capital market transactions because they can impact stock market activity as documented by Eichler (2012), who finds that capital controls help explain mispricing on a sample of 536 cross-listed stocks. We follow the approach in Eichler (2012) and use the dataset provided by Schindler (2009) which contains information on various capital controls including restrictions on (a) purchases or sales of equity domestically by nonresidents of each country and (b) purchases or

9 First, the authors document a negative relationship between the share of international firms and the domestic turnover of international firms and a positive relationship between the domestic turnover of international firms and the domestic turnover of domestic firms. Therefore, as internationalization increases, trade migrates away from the domestic market to major international financial centers. In turn, these negative effects of internationalization on the domestic turnover of international firms spill over and decrease the turnover of the strictly domestic firms. Second, they find that an individual firm’s share of the total stock market’s turnover increases when the firm becomes international: the trade diversion effect. 10 We thank an anonymous referee for this suggestion.

Please cite this article in press as: Hales, A. D., & Mollick, A.V. The impact of ADR activity on stock market liquidity: Evidence from Latin America. The Quarterly Review of Economics and Finance (2014), http://dx.doi.org/10.1016/j.qref.2014.03.005

G Model QUAECO-769; No. of Pages 11

ARTICLE IN PRESS A.D. Hales, A.V. Mollick / The Quarterly Review of Economics and Finance xxx (2014) xxx–xxx

7

Table 3 Fixed effects and SGMM estimates. Impact of ADR turnover on stock market liquidity in Latin America. Panel A: fixed effects

L.1 ADR Turnover IP growth Inflation Openness Market Capitalization Nonresident Purchases Resident Purchases Resident Sales Constant R2 AB (2) AB (2) p-value Sargan test Sargan test p-value

Panel B: SGMM

Value traded

Turnover

Value traded

Turnover

0.336*** (0.041) 0.185** (0.050) 0.049 (0.089) −0.194 (0.278) −0.001 (0.011) 0.025* (0.010) −0.006* (0.002) −0.002 (0.002) −0.001 (0.003) −0.628* (0.241) 78.04

0.132 (0.076) 0.150*** (0.012) 0.021 (0.023) −0.111 (0.050) 0.003 (0.003) 0.002 (0.001) −0.003*** (0.000) 0.001 (0.001) −0.000 (0.001) −0.036 (0.018) 64.23

0.714*** (0.105) 0.114* (0.064) 0.035 (0.088) −0.135 (0.111) −0.002 (0.004) 0.009* (0.005) −0.001 (0.002) −0.000 (0.001) −0.006*** (0.002) −0.234** (0.115)

0.260*** (0.048) 0.166*** (0.007) 0.017 (0.021) −0.076* (0.043) −0.007*** (0.001) 0.004*** (0.000) −0.003*** (0.000) 0.001* (0.001) 0.000 (0.001) −0.112*** (0.008)

0.50 (0.620) 687.69 (0.923)

0.61 (0.541) 853.70*** (0.003)

Notes: Latin American countries included are Argentina, Brazil, Chile and Mexico. Monthly data from January 2003 to December 2010. Panel A: Fixed effects regression of Value Traded-value of shares traded scaled by GDP and Turnover-value of shares traded scaled by market capitalization on their respective lagged values L.1–L.6 (due to space constraints, lags 2–5 are not shown, but are available upon request), ADR Turnover-value traded of ADR shares in the U.S. scaled by ADR capitalization), and control variables-IP growth which is the monthly change in each country’s industrial production index; Inflation based on the Consumer Price Index; Openness measured by the log of the sum of exports and imports scaled by GDP; and capital control variables that take on the value of 1 if the country imposes restrictions on domestic purchase by nonresidents, purchases abroad by residents, and sales abroad by residents, respectively and zero otherwise. Robust standard errors are shown in parentheses. Panel B: System GMM regression (with ADR Turnover and IP growth entered as endogenous variables and the number of instruments constrained by setting the maximum lags to 2) of the variables described above. * Statistical significance at the 10% level respectively. ** Statistical significance at the 5% level respectively. *** Statistical significance at the 1% level respectively.

sales of equity abroad by residents in each country.11 This approach requires creating binary variables that take on the value of 1 in the presence of restrictions and zero otherwise. In our estimations, we exclude controls on domestic sales by nonresidents because there was little variation for this measure in our sample. An additional component involves assessing possible impacts of the global financial crisis on the relationship between ADR activity and stock market development. Levy Yeyati et al. (2008) document that in emerging markets, crises are associated with higher trading activity and falling market capitalization. Therefore, one must ensure that changes in stock market conditions are not erroneously attributed to changes in ADR activity when they really reflect effects from the crisis. Consequently, empirical estimates are conducted across two subperiods: pre-crisis and crisis/post-crisis. We employ the following two dates to identify the beginning of the crisis: September 2008 because, as documented by Bartram and Bodnar (2009), this date reflects the collapse of financial markets after the bankruptcy of Lehman and bailout of AIG; and May 2008 because Dooley and Hutchison (2009) document evidence that conditions in Latin American markets began deteriorating significantly during this month. We also consider whether the impact of ADR activity on stock market liquidity varies based on the type of ADR-Level I, II, or III. Level I ADRs allow foreign issuers to reap the benefits of a

11 Schindler (2009) provides data from 1995 to 2005. Data for later dates is collected from various issues of the International Monetary Fund’s Annual Report on Exchange Arrangements and Exchange Restrictions.

publicly traded security without the costs of changing their reporting standards because they trade in OTC exchanges and do not require registration with the Securities and Exchange Commission (SEC) or adherence to U.S. Generally Accepted Accounting Principles (GAAP). Unlike Level I, issuers of Levels II and III ADRs must meet SEC disclosure requirements and conform to U.S. GAAP which enhances the legal protection of the firm’s investors and increases the value of the firm (Coffee, 2002; Doidge, Karolyi, & Stulz 2003). While Levels II and III ADR issuance is costly, these ADRs create higher visibility for the issuers and are more liquid than Level I ADRs. In addition, because Level III ADRs raise capital in U.S. markets, these issues receive more attention from the financial media and greater analyst coverage further increasing the visibility of the issuers. The greater visibility, particularly of Level III ADRs, suggests that they are more likely to affect the domestic stock market. Estimating (1) can be problematic because lagged stock market development yi,t−1 depends on ui,t−1 and ui,t−1 is a function of the country-specific effect, i . Since there is correlation between the regressors and the error term, Ordinary Least Squares (OLS) estimators are inconsistent. In addition, the fundamental factors that drive stock market development also drive internationalization as in Claessens et al. (2006). Internationalization can affect and be affected by different levels of development making the ADR activity variables endogenously determined. Furthermore, (1) utilizes the growth rate of the industrial production index to measure economic activity. As discussed previously, there is likely to be bidirectional causality between financial development and economic growth. Consequently, changes in stock market activity should elicit changes in industrial production (the so-called

Please cite this article in press as: Hales, A. D., & Mollick, A.V. The impact of ADR activity on stock market liquidity: Evidence from Latin America. The Quarterly Review of Economics and Finance (2014), http://dx.doi.org/10.1016/j.qref.2014.03.005

G Model

ARTICLE IN PRESS

QUAECO-769; No. of Pages 11

A.D. Hales, A.V. Mollick / The Quarterly Review of Economics and Finance xxx (2014) xxx–xxx

8

Table 4 SGMM estimates. Impact of ADR turnover on stock market liquidity in Latin America, pre- and post-global financial crisis. Crisis start: September 2008 Pre-Crisis Value traded L.1 ADR turnover IP growth Inflation Openness Market capitalization Nonresident purchases Resident Purchases Resident sales Constant AB (2) AB (2) p-value Sargan test Sargan test p-value

***

0.726 (0.107) 0.135 (0.126) 0.167* (0.095) −0.035 (0.178) −0.001 (0.006) 0.006 (0.005) −0.002 (0.003) −0.004 (0.003) −0.001 (0.001) −0.157 (0.104) 0.44 (0.658) 473.66 (0.919)

Crisis start: May 2008 Post-Crisis

Turnover ***

0.295 (0.055) 0.149*** (0.008) 0.054* (0.029) −0.037 (0.039) −0.007*** (0.001) 0.003*** (0.001) −0.002** (0.001) 0.000 (0.001) 0.000 (0.001) −0.088*** (0.029) 0.77 (0.442) 577.74** (0.035)

Value traded *

0.221 (0.121) 0.355*** (0.066) −0.031 (0.030) −0.315** (0.138) −0.015*** (0.002) 0.038*** (0.008) 0.018*** (0.002) 0.001 (0.003) −0.060*** (0.011) −0.994*** (0.195) −0.04 (0.972) 269.89*** (0.006)

Pre-Crisis Turnover

Value traded

0.060 (0.096) 0.224*** (0.021) −0.002 (0.006) −0.175*** (0.066) −0.010*** (0.001) 0.007*** (0.001) −0.001 (0.001) 0.002** (0.001) −0.000 (0.001) −0.196*** (0.039) 0.44 (0.658) 273.77 (0.004)

***

0.732 (0.106) 0.181 (0.139) 0.181* (0.096) −0.067 (0.152) 0.000 (0.006) 0.006 (0.004) −0.002 (0.002) −0.004 (0.003) −0.000 (0.002) −0.142 (0.092) 0.38 (0.706) 440.82 (0.930)

Post-Crisis Turnover ***

0.297 (0.049) 0.172*** (0.007) 0.054* (0.031) −0.027 (0.048) −0.006*** (0.001) 0.003*** (0.001) −0.002*** (0.001) 0.001 (0.001) −0.000 (0.001) −0.086*** (0.025) 0.84 (0.400) 547.12** (0.028)

Value traded *

0.270 (0.148) 0.250*** (0.060) −0.011 (0.054) −0.222** (0.111) −0.016*** (0.003) 0.037*** (0.009) 0.002 (0.002) −0.002 (0.003) −0.056*** (0.013) −0.945*** (0.237) −0.15 (0.881) 305.51*** (0.006)

Turnover 0.137* (0.081) 0.213*** (0.036) −0.004 (0.007) −0.158** (0.062) −0.009*** (0.001) 0.007*** (0.001) 0.002** (0.001) 0.002 (0.002) −0.001 (0.001) −0.179*** (0.031) 0.39 (0.694) 319.76*** (0.001)

Notes: Latin American countries included are Argentina, Brazil, Chile and Mexico. Monthly data from January 2003 to December 2010. System GMM regression (with ADR Turnover and IP growth entered as endogenous variables and the number of instruments constrained by setting the maximum lags to 2) of Value Traded-value of shares traded scaled by GDP and Turnover-value of shares traded scaled by market capitalization on their respective lagged values Lag 1, ADR Turnover-value traded of ADR shares in the U.S. scaled by ADR capitalization), and control variables-IP growth which is the monthly change in each country’s industrial production index; Inflation based on the Consumer Price Index; Openness measured by the log of the sum of exports and imports scaled by GDP; and capital control variables that take on the value of 1 if the country imposes restrictions on domestic purchase by nonresidents, purchases abroad by residents, and sales abroad by residents, respectively and zero otherwise. Robust standard errors are shown in parentheses. * Statistical significance at the 10% level respectively. ** Statistical significance at the 5% level respectively. *** Statistical significance at the 1% level respectively.

“wealth effect”). This casts doubt on exogenous economic activity and thus suggests an endogenous IP growth rate. To better handle these problems, we incorporate GMM estimators and allow for endogenous regressors. While two methods are available – the difference GMM (DGMM) estimator of Arellano and Bond (1991) and the system GMM (SGMM) estimator of Arellano and Bover (1995) and Blundell and Bond (1998) – the SGMM is selected because Blundell and Bond (1998) show that in situations where the lagged dependent and explanatory variables are persistent, the lagged-level instruments are weak, which compromises the asymptotic precision of DGMM estimator. The system approach jointly estimates the dynamic regression equation in first differences and levels, using different sets of instruments for each part. In generating the internal instruments necessary, this method employs the lagged levels and the lagged differences of all endogenous variables but only the lagged differences of exogenous variables. The extra moment conditions of the SGMM estimator improve precision and lower the finite sample bias.12 5. Results Panel A of Table 3 displays the empirical results of our benchmark model which resembles the approach employed by Karolyi

12

Roodman (2009) discusses the proliferation of instruments. One possibility is to collapse the instrument matrix. The estimations constrain the number of instruments by limiting the number of lags (to 2 and 1) and collapsing the instrument matrix. In our samples, this collapsing approach did not help the specification tests in most of the cases.

(2004). This approach takes a regression with fixed effects and augments it with six lags of the dependent variable to alleviate the bias that is associated with the dynamic nature of the equation.13 Due to space constraints, we report only the coefficients for the first lag and the remaining coefficients are available upon request. While Karolyi (2004) finds that ADR activity neither helps nor hinders stock market development, our results suggest that as ADR Turnover increases in U.S. markets, the domestic market benefits from increased liquidity. The positive relationship holds for both measures of stock market liquidity: value traded and turnover. Although these coefficients highlight the importance of accounting for activity in U.S. markets, a limitation of this approach is that it assumes exogeneity of all the regressors. Since stock market liquidity and the degree of ADR Turnover may be simultaneously determined, SGMM estimates of (1) allow for the endogenous determination of ADR Turnover and growth in Industrial Production. The results with the number of lags limited to 2, are reported in Panel B of Table 3.14 The SGMM estimates complement the benchmark fixed-effects model, in which ADR turnover and industrial production growth are now allowed to respond to fluctuations in the two dependent variables: value traded and turnover of the home market. The results continue to indicate that if ADR

13 We test the appropriateness of the fixed effects model using a Hausman test. The test statistics of 61.19 (0.0000) and 28.81 (0.0111) for Value Traded and Turnover, respectively, support the use of fixed effects. 14 The specification with the maximum lags set to 2 minimizes the number of instruments while maintaining acceptable diagnostics. Other specifications are available upon request.

Please cite this article in press as: Hales, A. D., & Mollick, A.V. The impact of ADR activity on stock market liquidity: Evidence from Latin America. The Quarterly Review of Economics and Finance (2014), http://dx.doi.org/10.1016/j.qref.2014.03.005

G Model QUAECO-769; No. of Pages 11

ARTICLE IN PRESS A.D. Hales, A.V. Mollick / The Quarterly Review of Economics and Finance xxx (2014) xxx–xxx

programs from Latin America are liquid in the foreign market, internationalization can increase liquidity in the domestic market as shown by the positive coefficients on both Value Traded and Turnover. Serial correlation tests in AB (2) are satisfactory in both cases. However, the SGMM results for Turnover should be interpreted cautiously as the Sargan test rejects the validity of the overidentifying restrictions. The SGMM estimates also confirm the expected negative impacts of inflation and restrictions on domestic purchases by nonresidents but the positive effect of market size on domestic market liquidity (as measured by turnover). Our results suggest that when one accounts for ADR activity in U.S. markets, there is a beneficial role to ADR issuance for domestic market development in Latin America. Interestingly, for Latin America, the value traded of international stocks is typically higher abroad than domestically (see the discussion of Table 2 earlier); the value traded of international stocks in the foreign market can even exceed the total value traded in the domestic market. While this pattern would predict the detrimental effects of trade migration, our results indicate that increasing trading activity in the foreign market stimulates domestic trading activity. Therefore, for these four Latin American countries, it appears that the foreign and domestic markets are complementary and not substitute trading venues. This contrasts with the findings by Levine and Schmukler (2007). The difference likely arises because our focus is on the impact of foreign turnover and not the concentration of ADR firms within a domestic market. Halling et al. (2008) find positive effects on domestic liquidity as a result of cross-listing but they find that these positive effects do not accrue to emerging markets. However, their emerging market sample extends beyond Latin America and includes only 27 Latin American cross-listed firms (as opposed to the 164 ADRs included here). The cross-listing dynamics may be different in Latin America in the 2000s, especially when coupled with the relative macroeconomic stability in the four countries during this time. To unveil any potential effects arising from the global financial crisis, the sample period is decomposed into two parts: the precrisis period begins in January 2003 and extends through August 2008 while the crisis/post-crisis period extends from September 2008 until December 2010. The results are presented in Table 4. The results indicate that the turmoil in financial markets during the peak of the global financial crisis significantly impacted the relationship between ADR Turnover and stock market liquidity. For example, the positive relationship between ADR Turnover and Value Traded is observed only during the crisis/post-crisis period suggesting that the relationship is possibly driven by the massive sell-offs occurring during this time period. However, the relationship between ADR Turnover and Market Turnover is robust to the effects of the crisis. The coefficients on ADR Turnover are positive and statistically significant in both the pre-crisis and crisis/postcrisis periods. Therefore, during tranquil periods, for Latin America, an increasingly liquid ADR segment in the foreign market boosts domestic market liquidity. Therefore, decomposing the sample periods into sub-periods to account for the global financial crisis does not eliminate the primary finding that increasingly liquid ADRs in the U.S. market can foster domestic market liquidity, as measured by turnover. The result is robust to different definitions of the beginning of the crisis but it is important to note that similar to the full-period results, across the different specification the Sargan test statistic rejects the validity of the overidentifying restrictions for market Turnover. Following Dooley and Hutchison (2009), May 2008 is also included to mark the beginning of the crisis period. To address the research question in more detail, estimates of Eq. (1) by ADR level are provided next. Based on the findings in Table 4, we include only the effect of ADR activity on market Turnover, by level and during the pre-crisis period. Table 5 displays the results

9

Table 5 SGMM Estimates. Impact of ADR turnover (by Level) on stock market turnover in Latin America.

L.1 ADR turnover IP growth Inflation Openness Market capitalization Nonresident purchases Resident purchases Resident sales Constant AB (2) AB (2) p-value Sargan test Sargan test p-value

Level I

Level II

Level III

0.336*** (0.068) −0.130 (0.092) 0.068* (0.036) −0.078* (0.044) −0.007*** (0.001) 0.006*** (0.001) −0.002*** (0.001) 0.000 (0.001) 0.001 (0.001) −0.153*** (0.019) 0.18 (0.855) 422.36 (0.803)

0.274*** (0.053) 0.095*** (0.009) 0.057* (0.034) −0.050 (0.052) −0.010*** (0.001) 0.004*** (0.001) −0.002*** (0.000) −0.001 (0.001) 0.001 (0.001) −0.119*** (0.024) 1.06 (0.290) 505.87 (0.268)

0.281*** (0.031) 0.098*** (0.009) 0.060* (0.033) −0.007 (0.046) −0.007*** (0.002) 0.004*** (0.001) −0.002*** (0.000) 0.001 (0.001) 0.000 (0.001) −0.109*** (0.026) 0.30 (0.767) 518.68 (0.155)

Notes: Latin American countries included are Argentina, Brazil, Chile and Mexico. Monthly data from January 2003 to May 2008. System GMM regression (with ADR Turnover and IP growth entered as endogenous variables and the number of instruments constrained by setting the maximum lags to 2 of Turnover-value of shares traded scaled by capitalization on its respective lagged value Lag 1, ADR Turnovervalue traded of ADR shares in the U.S. scaled by ADR capitalization), and control -IP growth which is the monthly change in each country’s industrial production index; Inflation based on the Consumer Price Index; Openness measured by the log of the sum of exports and imports scaled by GDP; and capital control variables that take on the value of 1 if the country imposes restrictions on domestic purchase by nonresidents, purchases abroad by residents, and sales abroad by residents, respectively and zero otherwise. Robust standard errors are shown in parentheses. * Statistical significance at the 10% level respectively. *** Statistical significance at the 1% level respectively.

using SGMM with endogenous ADR activity and IP growth, and limiting the maximum number of lags to 2. Across all specifications the diagnostic tests are satisfactory, including the test on instrument validity which indicated problems in some estimates of Tables 3 and 4. The coefficient for the impact of ADR Turnover on Market Turnover is statistically insignificant for ADRs of Level I. However, the impact is positive and statistically significant (at the 1% level) for Levels II (0.095 coefficient) and III ADRs (0.098 coefficient). Therefore, the positive impact of ADR Turnover on domestic market liquidity stems mainly from the turnover of exchange-listed ADRs in U.S. markets. The lack of discernible impact on Level I but positive impact of Levels II and III highlights two important points. First, an important part of measuring ADR activity is accounting for liquidity in U.S. markets. For positive effects from liquidity in the U.S. market to spill over to the home market, the ADRs must be liquid in the U.S. market. However, Level I ADRs tend to have low liquidity and high trading infrequency in U.S. markets. Across all markets the value of Level I shares traded in the home market exceeds that of the ADRs trading in the United States. Therefore, it is not surprising that we find no benefit arising from the Level I ADR segment. Second, the benefits from increased ADR turnover in U.S. markets are primarily due to the effects from large cross-listed firms that choose to adhere to the relatively strict listing requirements of U.S. markets. Both Levels II and III exhibit positive impacts because they have the strongest capacity for increasing visibility and credibility of the domestic market because they enjoy the highest coverage from media and analysts. If these firms are effectively

Please cite this article in press as: Hales, A. D., & Mollick, A.V. The impact of ADR activity on stock market liquidity: Evidence from Latin America. The Quarterly Review of Economics and Finance (2014), http://dx.doi.org/10.1016/j.qref.2014.03.005

G Model QUAECO-769; No. of Pages 11

10

ARTICLE IN PRESS A.D. Hales, A.V. Mollick / The Quarterly Review of Economics and Finance xxx (2014) xxx–xxx

“bonding” to the U.S. regulatory environment, then by subjecting themselves to the scrutiny of investors, analysts, the SEC, and the exchanges themselves, they improve the information environment. The results here indicate that these improvements also accrue to the domestic market improving its overall liquidity. 6. Concluding remarks This paper examines the impact of ADR activity on the liquidity of four Latin American stock markets. We take into account three elements not present in the literature for emerging markets: measurement of ADR activity using U.S. trading data, a re-examination of different levels of exchange-listed ADRs under these measures, and reverse causation since home market liquidity may affect economic growth at home and (very likely) the extent of liquidity of ADRs in the U.S. market. Our analysis shows that in these four Latin American countries the ADR segment represents approximately 50% of the home stock markets. Our analysis updates the quantification of ADR activity by examining the turnover of the ADR segments in U.S. markets. We find for these Latin American countries that the value of shares traded in the U.S. typically exceeds the value of shares traded at home. In some cases, the value of ADR shares traded in the U.S. even exceeds the total value of shares traded in the domestic market. By accounting for U.S. activity, we find that increases in ADR turnover boost turnover in the domestic market. These effects are not driven by effects of the global financial crisis as the positive relationship exists even in the tranquil period before the crisis (January 2003–April 2008). Furthermore, we document that the positive effects of ADR turnover on domestic market liquidity exist only for large, exchange-listed ADRs. This suggests that the potential benefits of ADR issuance in Latin American markets stems on their ability to enhance the visibility of other domestic stocks and their ability to ease capital constraints. Ultimately, our results do not indicate that internationalization has inhibited development, as shown by the low levels of liquidity in these four major Latin American markets. Instead, our results suggest that activity in the foreign market is a complement to domestic activity. If this is the case, even when domestic activity is stagnant, improvements in the international activity of these countries’ firms can bring the liquidity, risk diversification and information necessary for the stock market to fulfill its role in the efficient allocation of funds. Important questions left for future research include: What is the impact of internationalization on economic growth? and, provided such an impact exists, is it driven by Latin American firms issuing exchange-listed ADRs? Acknowledgments The authors wish to thank an anonymous Referee for comments that helped improve the paper and Hadi S. Esfahani for editorial comments. The usual disclaimer applies. References Alaganar, V. T., & Bhar, R. (2001). Diversification gains from American depositary receipts and foreign equities: Evidence from Australian stocks. Journal of International Financial Markets Institutions, and Money, 11, 97–113. Arellano, M., & Bond, S. (1991). Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. Review of Economic Studies, 58, 277–297. Arellano, M., & Bover, O. (1995). Another look at the instrumental variables estimation of error-components models. Journal of Econometrics, 68, 29–51. Arnold, T., Nail, L., & Nixon, T. D. (2004). Do ADRs enhance portfolio performance for a domestic portfolio? Evidence from the 1990. Research in International Business and Finance, 18, 341–359.

Baltagi, B. H., Demetriades, P. O., & Law, S. H. (2009). Financial development and openness: Evidence from panel data. Journal of Development Economics, 89, 285–296. Bartram, S. M., & Bodnar, G. M. (2009). No place to hide: The global crisis in equity markets in 2008/2009. Journal of International Money and Finance, 28, 1246–1292. Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics, 87, 115–143. Boubakri, N., Cosset, J., & Samet, A. (2010). The choice of ADRs. Journal of Banking and Finance, 34, 2077–2095. Boyd, J. H., Levine, R., & Smith, B. D. (2001). The impact of inflation on financial sector performance. Journal of Monetary Economics, 47, 221–248. Chan, J. S. P., Hong, D., & Subrahmanyam, M. G. (2008). A tale of two prices: Liquidity and asset prices in multiple markets. Journal of Banking & Finance, 32, 947–960. Chinn, M. D., & Ito, H. (2006). What matters for financial development? Capital controls, institutions, and interactions. Journal of Development Economics, 81, 163–192. Choi, Y. K., & Kim, D. (2000). Determinants of American depositary receipts and their underlying stock returns: Implications for international diversification. International Review of Financial Analysis, 9, 351–368. Claessens, S., Klingbiel, D., & Schmukler, S. (2006). Stock market development and internationalization: Do economic fundamentals spur both similarly? Journal of Empirical Finance, 13, 316–350. Coffee, J. C. (2002). Racing towards the top: The impact of cross-listings and stock market competition on international corporate governance. Columbia Law Review, 102, 1757–1831. De la Torre, A., & Schmukler, S. L. (2007). Emerging capital markets and globalization: The Latin American experience. Washington, DC: Stanford University Press and the World Bank. Doidge, C. A., Karolyi, G. A., & Stulz, R. M. (2003). Why are firms that list in the U.S. worth more? Journal of Finance, 53, 2107–2137. Domowitz, I., Glen, J., & Madhavan, A. (1998). Market segmentation and stock prices: Evidence from an emerging market. Journal of Finance, 52, 1059–1085. Dooley, M., & Hutchison, M. (2009). Transmission of the U.S. subprime crisis to emerging markets: Evidence on the decoupling-recoupling hypothesis. Journal of International Money and Finance, 28, 1331–1349. Eichler, S. (2012). Limited investor attention and the mispricing of American depositary receipts. Economics Letters, 115, 490–492. Eichler, S., Karmann, A., & Maltritz, D. (2009). The ADR shadow exchange rate as an early warning indicator for currency crises. Journal of Banking and Finance, 33, 1983–1995. Fernandes, N. (2009). Market liberalizations at the firm level: Spillovers from ADRs and implications for local markets. Journal of International Money and Finance, 28, 293–321. Foerster, S. R., & Karolyi, G. A. (1998). Multimarket trading and liquidity: A transaction data analysis of Canada-US interlistings. Journal of International Financial Markets Institutions and Money, 8, 393–412. Gagnon, L., & Karolyi, G. A. (2010). Multi-market trading and arbitrage. Journal of Financial Economics, 97, 53–80. Grossman, A., Ozuna, T., & Simpson, M. W. (2007). ADR mispricing: Do costly arbitrage and consumer sentiment explain the price deviation? Journal of International Financial Markets, Institutions, & Money, 17, 361–371. Hadri, K. (2000). Testing for stationarity in heterogeneous panel data. Econometrics Journal, 3, 148–161. Halling, M., Pagano, M., Randl, O., & Zechner, J. (2008). Where is the market? Evidence from cross-listings in the United States. Review of Financial Studies, 21, 725–761. Hargis, K. (1998). Do foreign investors stimulate or inhibit stock market development in Latin America? Quarterly Review of Economics and Finance, 38, 303–318. Hargis, K. (2000). International cross-listing and stock market development in emerging economies. International Review of Economics and Finance, 9, 101–122. Hargis, K., & Ramanlal, P. (1998). When does internationalization enhance the development of stock markets? Journal of Financial Intermediation, 7, 263–292. Im, K., Pesaran, H., & Shin, Y. (2003). Testing for unit roots in heterogenous panels. Journal of Econometrics, 115, 53–74. Karolyi, G. A. (1998). Why do companies list shares abroad? A survey of the evidence and its managerial implications. Financial Markets, Institutions and Instruments, 7, 1–60. Karolyi, G. A. (2004). The role of American depositary receipts in the development of emerging equity markets. Review of Economics and Statistics, 86, 670–690. Karolyi, G. A. (2006). The world of cross-listings and cross-listings of the world: Challenging conventional wisdom. Review of Finance, 10, 99–152. King, R., & Levine, R. (1993a). Finance and Growth: Schumpeter might be right. Quarterly Journal of Economics, 108, 717–737. King, R., & Levine, R. (1993b). Finance, entrepreneurship and growth: Theory and evidence. Journal of Monetary Economics, 32, 513–542. Levine, R., & Schmukler, S. (2007). Migration, spillovers, and trade diversion: The impact of internationalization on domestic stock market activity. Journal of Banking and Finance, 31, 1595–1612. Levine, R., & Zervos, S. (1998). Stock markets, banks, and economic growth. American Economic Review, 88, 537–558. Levy Yeyati, E., Schmukler, S. L., & Van Horen, N. (2008). Emerging market liquidity and crises. Journal of the European Economic Association, 6, 668–682. Miller, D. P. (1999). The market reaction to international cross-listing: Evidence from depositary receipts. Journal of Finance, 42, 483–510.

Please cite this article in press as: Hales, A. D., & Mollick, A.V. The impact of ADR activity on stock market liquidity: Evidence from Latin America. The Quarterly Review of Economics and Finance (2014), http://dx.doi.org/10.1016/j.qref.2014.03.005

G Model QUAECO-769; No. of Pages 11

ARTICLE IN PRESS A.D. Hales, A.V. Mollick / The Quarterly Review of Economics and Finance xxx (2014) xxx–xxx

Moel, A. (2001). The role of American depositary receipts in the development of emerging markets. Economía, 2, 209–273. Officer, D. T., & Hoffmeister, J. R. (1987). ADRs: A substitute for the real thing? Journal of Portfolio Management, 13, 61–65. Rajan, R., & Zingales, L. (2003). The great reversals: The politics of financial development in the 20th century. Journal of Financial Economics, 69, 5–49.

11

Roodman, D. (2009). A note on the theme of too many instruments. Oxford Bulletin of Economics and Statistics, 71, 305–9049. Schindler, M. (2009). Measuring financial integration: A new dataset. IMF Staff Papers, 56, 222–238. Wahab, M., & Khandwala, A. (1993). Why not diversify internationally with ADRs? Journal of Portfolio Management, 19, 75–82.

Please cite this article in press as: Hales, A. D., & Mollick, A.V. The impact of ADR activity on stock market liquidity: Evidence from Latin America. The Quarterly Review of Economics and Finance (2014), http://dx.doi.org/10.1016/j.qref.2014.03.005