Essay on foreign and domestic investors

... consists o f two essays The first essay examines the contribution o f foreign investor on domestic market microstructure The second essay investigates whether foreign or domestic investors are better... cross-sectional determinants o f the abnormal return during foreign presence, and hence sheds light on the contribution o f foreign investors on domestic markets We find that the presence o f foreign investors. .. permission of the copyright owner Further reproduction prohibited without permission Foreign Investors Contribution on Domestic Market Microstructure: The Indonesian Case Abstract This essay

INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. ProQuest Information and Learning 300 North Zeeb Road. Ann Arbor, Ml 48106-1346 USA 800-521-0600 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ESSAY ON FOREIGN AND DOMESTIC INVESTORS BY MAMDUH M HANAFI A DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN BUSINESS ADMINISTRATION UNIVERSITY OF RHODE ISLAND 2001 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number 3039077 __ ___ __ (g ) UMI UMI Microform 3039077 Copyright 2002 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DOCTOR OF PHILOSOPHY DISSERTATION OF MAMDUH M HANAFI APPROVED Dissertation Committee Major Professor OF THE GRADUATE SCHOOL UNIVERSITY O F RHODE ISLAND 2001 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ACKNOWLEDGEMENT I w ish to thank members o f my dissertation committee, Professors S. Ghon Rhee, Jacqueline Faught, Tim Tyrrell, for their guidance and support throughout this dissertation work. I am especially grateful to my major professor, Dr. S. Ghon Rhee, who sparked my interest in Asian markets and has generously and patiently counseled me on this subject. I also would like to thank Professors Shaw Chen and Cathy W essells, who have kindly consented to serve in the examination committee o f my dissertation defense. My sincere thanks also go out to Professor Shaw Chen, in his capacity as the Director o f Ph.D. program in College o f Business Administration, who has provided me with much guidance and encouragement. I am also indebted to the exceptionally gifted Finance faculties and the Ph.D. students in the University o f Rhode Island from where I gained much knowledge; and their suggestions and insights greatly enhance my understanding o f this work. The Pacific-Basin Capital Market Research Center (PACAP) and its director, Professor Shaw Chen, and its former director, Professor Rosita Chang, deserve my sincere thanks for assistance and availability o f databases which provide me with a wealth o f research opportunities. In the course o f my doctoral studies, I received financial support from Universitas Gadjah Mada, Yogyakarta, Indonesia and the University of Rhode Island. This work would not have been possible without their generosity, which is gratefully acknowledged. Finally, I wish to thank my family for their supports. iii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. PREFACE The author uses the manuscript format for this dissertation. In its current form , the dissertation consists o f two essays. The first essay examines the contribution o f foreign investor on domestic market microstructure. The second essay investigates whether foreign or domestic investors are better informed. Both essays use the sample o f stocks listing on the Jakarta Stock Exchange. iv Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLE OF CONTENTS PAGE TABLE OF CONTENTS ............................................................................................ v MANUSCRIPT Foreign Investors’ Contribution on Domestic Market Microstructure: The Indonesian C a se ........................................................................................................ 1 1. Introduction ................................................................................................................. 2 2. Related Literature ........................................................................................................ 4 3. The ‘Event ‘o f Foreign Presence .................................................................................. 7 4. The Jakarta Stock Exchange (JS X )................................................................................ 9 5. Data and Sample ........................................................................................................ 12 6. Empirical Findings .................................................................................................... 14 7. Heterogeneous Effect o f Financial Liberalization ................................................ 31 8. Conclusion ................................................................................................................. 32 References ..................................................................................................................... 33 Are Foreign Investors More Informative? An Examination o f The Price Movement in The Jakarta Stock E xchange......................................................................................... 56 1. Introduction ......................................................................................................... 57 2. Related Literature ................................................................................................. 62 3. The Jakarta Stock Exchange (JSX) .......................................................................... 66 4. Data and Methodology ................................................................................................68 5. Trading Characteristics o f Foreign and Domestic Investors .................................. 75 6. Price Movement During The Event Period ...............................................................78 7. Brokers, Trade Size, Liquidity, and The Crisis Period ........................................... 84 8. Trading Size o f Domestic and Foreign Investors .................................................... 89 9. Analysis in the Non-event period ............................................................................. 91 10. Sensitivity Tests ...................................................................................................... 95 11. Summary and Conclusion .................................................................................... 100 References ..................................................................................................................... 102 BIBLIOGRAPHY ........................................................................................................... 145 v Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Foreign Investors’ Contribution on Domestic Market Microstructure: The Indonesian Case Abstract This essay investigates foreign investors’ contribution on domestic market microstructure. Specifically we investigate whether foreign investors’ presence increases domestic m arket’s wealth as evidence by positive abnormal return during their presence; and whether the positive abnormal return is associated with improved efficiency and improved liquidity in the domestic market. We use two events as signs o f foreign presence: (1) the date when foreign board in the JSX (Jakarta Stock Exchange) is first created, and (2) the announcement o f foreign restriction removal on September 4, 1997. The second event takes place in the financial crisis period experienced by the Indonesian market in late 1990s. Using modified market model, we find that foreign presence is associated with positive abnormal return. We find that efficiency variables seem to relate more strongly with abnormal return than liquidity variables. The positive abnormal for the second event tends to be smaller than that for the first event, suggesting that foreign presence or liberalization during the financial crisis provides more limited benefit. We also find that the effect o f foreign presence tends to be heterogeneous across stocks. l Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Foreign Investors* Contribution on Domestic Market Microstructure: The Indonesian Case 1. Introduction From theoretical and policy perspectives, the presence o f foreign investors creates a continuing debate. Foreign investors can be expected to help develop a domestic economy. But their presence may lead to negative consequences. Financial liberalization attracts foreign capital to finance economic growth (Levine and Zervos, 1998); improves resource allocation from international risk sharing (Obstfeld, 1994); and reduces the costs o f external finance (Rajan and Zingales, 1998). On the other hand, foreign investors are associated with a more volatile market and a high dependency on foreign countries. Concerns about a volatile market prompt some governments to restrict their markets. For example, in light o f a financial crisis, M alaysia closed its financial market to foreign investors in September 1998. The Asian crisis o f 1997 revives the debate o f the role o f foreign investors. Some authors argue that foreign investors contributed to the crisis (Stiglitz, 1998; Krugman, 1998). Possible behaviors leading to the crisis were momentum trading or positive feedback trading and herding behavior among foreign investors. When a market is in good shape, positive feedback traders combined with herding behavior will drive the domestic market up, maybe to a level unwarranted by economic fundamentals. When the market turns bad, these two behaviors may cause a financial crash, as foreign investors are likely to abandon the declining market altogether. Thus foreign investors who behave this way may potentially destabilize the market. While this is the common perception, interestingly Choe, Kho, and Stulz (2000) do not find 2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. herding behavior among foreign investors in the Korean m arket during the crisis period. They suggest that foreign investors do not destabilize the m arket. In the normal period, foreign investors seem to engage in positive feedback trading or momentum trading (Choe, Kho, and Stulz, 2000; Grinblatt and Kelohaiju, 2000) and herding behavior (Choe, Kho, and Stulz, 2000). Empirical research focusing on the contribution o f foreign investors can contribute to the debate. This study investigates the effect o f foreign investors on a domestic m arket microstructure. Specifically, we investigate w hether foreign investors improve m arket efficiency and liquidity. While Henry (2000a) and Kim and Singal (2000) use aggregate data, we take a different approach. We use stock level data to investigate this issue. This approach enables us to investigate the cross-sectional determinants o f the abnormal return during foreign presence, and hence sheds light on the contribution o f foreign investors on domestic markets. We find that the presence o f foreign investors is associated with positive abnormal return. The changes in m arket efficiency seem to explain the abnormal returns better than the changes in m arket liquidity. Liberalization conducted during a financial crisis seems to provide m odest benefits. Finally, we find that the effect o f liberalization does not seem to be homogenous across stocks. We organize this paper as follows. The next section discusses related literature; sections 3 and 4 discuss the events used in this study and a brief history o f the Jakarta Stock Exchange; sections 5,6, and 7 discuss data, methodology, the analysis, and the results o f this paper. Finally, section 8 concludes this paper. 3 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2. Related Literature This paper extends the literature on the effect o f liberalization on domestic markets. Perhaps the spirit o f this paper is closest to that o f Henry (2000a), Kim and Singal (2000), and Domowitz and Madhavan (1998). Each o f these papers investigates the effect o f foreign investors on domestic markets using different approaches. Previous research on the effect o f liberalization focuses on the lower cost o f capital as the result o f market integration. When the market is opened to foreign investors, risk sharing between domestic and foreign investors occurs, resulting in a lower cost o f capital o f domestic firms (Errunza and Losq, 1985; Eun and Janakiramanan, 1986; Alexander, 1987; Stulz, 1991). Henry (2000a) pushes this argument further by investigating the valuation effect o f liberalization. If the cost o f capital is lower, holding cash flow constant, we can expect to have a higher valuation. He finds that countries initiating stock market liberalization experience abnormal returns o f 4.7 percent per-month in real dollar terms during the eight-month window leading up to the implementation o f the liberalization. Since stock market liberalization is usually conducted along with other liberalizations, the number may be upwardly biased. When Henry (2000a) controls for other liberalizations during his period o f study, he still finds a positive number o f about 3.3 percent per-month. Kim and Singal (2000) investigate the effect o f the stock market openings o f emerging economies. They find that stock returns increase immediately after the opening without causing additional volatility. Stock markets also become more efficient as shown by the evidence from the tests of random walk hypothesis. Inflation does not seem to increase -rather it tends to decrease- and the exchange rate does not seem to appreciate. The results seem to 4 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. suggest that the benefits outweigh the costs o f market openings. The revaluation effect of foreign investors seems to be consistent with the general prediction o f international capital asset pricing. Opening a domestic market to foreign investors allows risk sharing that results in a lower cost o f capital. From a market integration perspective, after the opening o f a domestic market, the cost o f capital will be proportional to the covariance o f a domestic company’s cash flow to the world market’s cash flow; on the other hand, before the market is open, it is proportional to the variance o f a domestic company’s cash flow. Since the covariance term is less than the variance, as generally assumed by the international capital asset pricing model, the opening o f the market lowers the cost o f capital. Holding cash flow constant, the lower cost o f capital leads to more projects with positive net present values, leading to real investment booms. Henry (2000b) investigates this issue and finds that there is a temporary investment boom following a stock market liberalization. Using event study framework, he finds that the growth rate mean o f private investment in the three years immediately following stock market liberalizations exceeds the sample mean by 22 percentage points. This result stands in a sharp contrast with previous study that finds that liberalization does not lead to a permanent increase in the growth rate o f capital stock (Levine and Zervos, 1998). The positive valuation effect may be caused by improved liquidity (Amihud and Mendelson, 1986). Higher liquidity lowers the cost o f capital, which leads to a higher asset price. Liberalization seems to increase market liquidity (Levine and Zervos, 1998; Kim and Singal, 2000).1 This may also explain the revaluation effect 1 Kim and Singal (2000) provide evidence that the trading frequency increases after the market 5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. associated with liberalization. Our focus on the microstructure variables parallels Domowitz and Madhavan (1998), when they investigate the effect o f inter-listing in the M exican market. They consider several scenarios. If inter-market linkage information is freely available, inter­ listing improves market quality. In an opposite case, where inter-m arket linkage information is poor, inter-listing reduces liquidity and increases volatility in domestic market. When they investigate empirically the effect o f inter-listing in the Mexican market, they find that the inter-listing improves efficiency, but tends to reduce trading volume, especially in B-series stocks, which were open to foreign investors prior to the listing. There seem to be trading migrations abroad in these stocks, leading to lower liquidity in the domestic market. Informed traders seem to prefer trading in a more efficient market. Other papers use efficiency and liquidity variables in different settings. Amihud et al. (1997) investigate the effect o f changing o f trading mechanism in the Tel Aviv Stock Exchange, from a call auction to a mechanism where the call auction was followed by continuous trading sessions. They find that the change is associated with a price appreciation o f about 5.5% during event days. When they investigate the crosssectional determinants o f average abnormal returns during the event days, they find that continuous trading improves market efficiency and liquidity, and generates positive liquidity externalities across related stocks. Berkman and Eleswarapu (1998) investigate foe effect o f foe changing o f forward trading regulation in foe Bombay opening to foreign investors. While they use this finding to explain improved efficiency associated with the market opening, we can also interpret the finding as an improved liquidity associated with the market opening. 6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Stock Exchange. They find that the abolition (reinstatement) o f forward trading is associated w ith negative (positive) price appreciation. The forward trading is associated with unproved liquidity and higher noise. Market seems to reward improved liquidity as evidenced by a positive association between liquidity changes from the post to the pre-event period and the abnormal returns during the event period. This result is consistent w ith the notion that forward trading improves liquidity but does not eliminate noise. 3. The ‘Event ‘o f Foreign Presence We use two events that took place in the JSX m arket to investigate the effect o f foreign presence (financial liberalization) on domestic market. First, we use the day when foreign board appears for the first time in the PACAP database. Before September 4, 1997, the Jakarta Stock Exchange (JSX) imposed foreign ownership restrictions o f 49% o f outstanding shares.2 Once the limits become binding, foreign investors m ust buy shares from other foreign investors. Foreign board was created to facilitate the trading o f foreign owned shares among foreign investors. We view the date when the foreign board appears for the first time as ’an official’ sign o f foreign presence. We focus on the first dates when foreign limits are reached for the following reasons. Errunza and Losq (1989) argue that a fully segmented market will a have 2 The phenomena o f foreign restrictions are relatively common in international market, either in developing or developed countries. For example, Thailand market has alien and main board (Bailey and Jagtiani, 1994). Alien board is similar to foreign board in the JSX. In Switzerland market, registered stocks could be held only by domestic investors, while foreign investors could trade in two classes o f stocks, voting and nonvoting bearer (VB and NVB stocks, see Loderer and Jacobs, 1995). 7 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. lower cost o f capital than a mildly segmented market when the m arket is opened. Henry (2000a) finds that stock market liberalizations following initial liberalization are associated with a m uch smaller and statistically insignificant revaluation. These papers suggest that the effect o f follow-up liberalizations tend to be weaker than that o f initial liberalization. We expect that when a company hits foreign limit for the first time, market impact will be higher than that o f subsequent hits. Second, we focus on the date o f financial liberalization announcement on September 4, 1997. On that day, the Indonesian government announced several economic measures that included the removal o f foreign ownership restriction. Under the new regulation, with the exception o f financial companies, foreign investors can buy up to 100% o f outstanding shares. On January 28, 1998, the Indonesian government announced the removal o f foreign ownership restriction for financial companies. A t this point, the Indonesian market was, at least officially, completely open to foreign investors. Both events have similarity and uniqueness with those in other studies. First, the events in this study are exogenous to companies. These are sim ilar to the changing o f foreign limits in the Singapore m arket (Lam, 1997), the announcement o f liberalization in Henry (2000a), but are different from the changing o f foreign limits in Stulz and W asserfallen (1995). In Stulz and Wasserfallen (1995), companies change foreign limit voluntarily, hence the change is endogenous to the companies. Second, the announcement o f liberalization measure o f September 1997 was conducted during a severe financial crisis. It is possible that policy makers behave like managers who issue shares when they are overvalued (Ritter, 1991), as pointed out by Summers (1994). If 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. any, then the announcement o f September 4, 1997 suggests that the policy makers behave like a manager o f a bankrupt company issuing shares. Henry (2000a) does not m ention explicitly the setting o f liberalization events he investigates. If policy makers liberalize when the market is doing well, we may expect to have results that are biased upwardly. In our case, the financial crisis setting may provide results that are biased downwardly. Although both events in our study involve the changing o f foreign restriction, we view them as the signs o f foreign presence. Hence we focus on the effect o f foreign presence, rather than on the effect o f foreign restriction changes. Our study belongs more closely to the literature on the effect o f liberalization on domestic markets (Henry 2000a,b; Kim and Singal, 2000; Domowitz and Madhavan, 1998), rather than to the literature on foreign restrictions (Stulz and Wasserfallen, 1995; Bailey and Jagtiani, 1994; Domowitz and Madhavan, 1997; Lam, 1997). The main theme in the literature on foreign restriction is the violation o f the law o f one price as a result o f the restrictions. 4. The Jakarta Stock Exchange (JSX) The history o f JSX dated back during the colonization era. The Dutch colonial government established the first stock exchange in Batavia3 in 1912. During the First W orld War, it was closed and then reopened again in 1925. The occupation o f Japan in Indonesia halted the exchange. Seven years after Indonesian independence, the exchange was opened again in 1952. The nationalization program in 1956 halted the 3 Batavia became Jakarta, which is the capital o f Indonesia. 9 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. trading again. The modem Jakarta Stock Exchange started in 1977 when President Soeharto reopened the exchange. The new institution called Capital Market Executive Agency (Badan Pelaksana Pasar M odal or Bapepam) managed the exchange. During the early years o f the exchange, Bapepam set a priority o f promoting and protecting domestic investors. The policy o f promoting domestic investors is intended for wealth distribution. Foreign com panies or joint venture companies were among the first companies to go public to satisfy the policy. The Indonesian government introduced a financial company called Danareksa. Danareksa served as a closed-end mutual fund. In this role, Danareksa helped implement the policy o f promoting domestic investors’ participation in the market, and hence wealth transfer objective. For investor protection, the government strongly discouraged speculations. Price movements were lim ited to 4% daily. Danareksa actively intervened the m arket when the limits reached 4%. At this stage, the market was closed for foreign investors. Such microstructure policy did not appeal much to potential market players. Macro economic policy did not help either. In the early 1980’s, the Indonesian government introduced a banking deregulation. This measure created a stiff competition in the banking industry leading to a higher interest rate, making investment in stock market less attractive. From 1977 to 1988, there were only 24 companies that w ent public. Most o f these companies went public to satisfy the policy o f promoting dom estic investors, rather than to obtain financing capital. The Indonesian stock market started to pick up in the late 1980’s when the Indonesian government introduced deregulation packages aim ed at promoting stock market. In 1988, the government 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. removed the 4% price limits, introduced easier requirements and procedures for going public, and, the m ost important o f all, allowed foreign investors to buy up to 49% o f outstanding shares o f listed stocks. The Indonesian government started to impose taxes on interest income in the same year making stock investment more attractive relative to deposit investment. In the two years following deregulation, the number o f companies that went public increased significantly from 24 to 67. Along with the growth o f Indonesian economy, the JSX index started to move up significantly. Figure 1 shows the development o f the JSX index from 1985 to 1998. Table 1 shows the number o f companies that went public and the amount o f fund raised since year 1977 to 2000. [ insert Table 1 and Figure 1 ] The Jakarta Stock Exchange was privatized in 1993. The JSX is a self-regulated institution owned by brokerages and is under the supervision o f Bapepam. Bapepam shifted the role from managing and organizing the exchange to a supervising one. The Jakarta Stock Exchange handles the operation o f the exchange. In 1995, The Jakarta Stock Exchange introduced automated trading system, called JATS (Jakarta Automated Trading System), to replace manual trading system. The Indonesian market suffered a setback when the financial crisis hit Indonesia in the middle o f 1997. The Jakarta Stock Exchange provides a ‘partial’ cycle o f foreign ownership regulation.4 Until December 1987, the Indonesian market was practically closed to 4 A new regulation to restrict foreign ownership would have completed a cycle (i.e. closed-open- 11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. foreign investm ent The Indonesian government started to open its market gradually. During the next 10 years, the Indonesian government introduced four key measures to open up the Indonesian stock market: (I) The Minister o f Finance decree o f September 16, 1989 that allowed foreign investors to buy up to 49% o f outstanding shares o f all listed non-financial companies, (2) The Minister o f Finance decree o f October 30,1992 that allowed foreign investors to buy up to 49% o f outstanding shares o f listed financial companies, (3) The M inister o f Finance decree o f September 11, 1997 that allowed foreign investors to buy up to 100% o f listed non-financial companies, and (4) The M inister o f Finance decree o f January 28,1998 that allowed foreign investors to buy up to 100% o f listed financial companies. These series o f events provide natural experiments where we can investigate the effect o f the events on the market behavior. Specifically we use the changing o f foreign ownership as a sign o f foreign presence, and use it to investigate the effect o f foreign investors on domestic market. S. Data and Sample 5.1. Data We use the data from PACAP database and the transaction data obtained directly from the JSX. We focus on regular board. The regular board is the most liquid board (about 83% [89%] o f the JSX’s trading value [volume] during our sample period) and is likely to be the place for typical investors to trade. The PACAP data we use contain daily closing price, daily return, daily trading volume, daily trading value, and financial statement from 1985 to 1997. The transaction data we use start from May closed market again). We do not expect to have such new regulation in the near future. 12 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1995 to August 1998. W e use the transaction data to investigate the announcement o f liberalization measure o f September 4, 1997, since the PACAP database ends by December 1997. 5.2. Comparison Between Stocks with Foreign Board and Stocks without Foreign Board As o f December 1997, out o f 291 companies listing in the Jakarta Stock Exchange, there were 80 companies that have trading on the foreign board. Since the stocks that have foreign board represent those that foreign investors are interested in, we compare their characteristics with those that do not have foreign board. Investigating the differences may help reveal answers to the questions o f home-bias puzzle. Table 2 shows the characteristic differences between these two groups. [Insert table 2] Stocks with foreign board tend to have better accounting performance as evidenced by higher current ratio and return on assets. They also have lower systematic risk and lower variance o f residual calculated from the market model.5 Foreign investors also seem to prefer more liquid stocks. M eans o f daily trading volume and value for stocks with foreign board are higher than those for stocks without foreign 5 The market model specification is R* = a + (In Rm, + e„, where R* = return o f stock i at day t, Rm/=market return calculated from the JSX index at day t, a=intercept, Pu =systematic risk, ert= residual. See also section 6 for market model specification used for this study. 13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. board.6 The most surprising result is that size does not seem to attract foreign investors. Stocks with foreign board tend to be smaller than stocks without foreign board. We use tw o variables to measure size: market capitalization and total asset. Both provide the sam e conclusion. Liquidity tends to have a positive association with foreign ownership suggesting that, in the Indonesian case, liquidity is a better proxy for foreign investors’ fam iliarity with domestic stocks (see also Bailey and Jagtiani, 1994).7 W hile other variables show the direction o f the sign as expected, only current ratio shows a statistically significant difference. 6. Empirical Findings 6.1. Price Impact o f Foreign Presence Price Movement Around the Event. Figure 2 shows cumulative abnormal return around the event o f the establishment o f foreign board for the first time (for the rest o f the paper we refer it as the first event).8 Figure 3 shows cumulative market return around the announcement o f liberalization measure o f September 4, 1997 (for the rest o f the 6 We do not investigate the causality between foreign investors and the variables presented here. For example, it is possible that foreign investors drive liquidity up, but it is also possible that liquidity attracts foreign investors. For some variables, it makes more sense to assume one direction than the other way around. For example, for accounting performances, it makes more sense to assume that investors are attracted to companies with better performances rather than foreign investors help improve performances. We tend to believe that foreign investors do not have sufficient control or [pressure on the way company business runs. Other measures could be stocks with ADR and degree o f export trading (Kang and Stulz, 1997). 8 For the first event, abnormal return is defined as follows. First we estimate the parameters from the following market model: R,, = a + P Rm, + e„, excluding days - 2 0 to +20. Then w e calculate “the residual” in the days - 2 0 to +20 by substituting market return in the model above using the parameters obtained from days -1 5 0 to +150, excluding days - 2 0 to +20 (Fama et al. 1969). 14 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. paper we refer it as the second event).9 Figure 2 shows an upward movement in stock prices starting from day -2 0 leading to the event date. Stocks move in a much more random fashion following the event date. Figure 3 shows a rather different picture. We observe an upward movement starting from day -2; however we also observe a much more volatile movement during the event windows. This seems to be the characteristic o f a financial crisis, in which the second event took place. In the financial crisis, the market index showed a declining trend and a high volatility. The variance o f daily market return during the financial crisis is about 9%, which is significantly larger than that in the normal period (about 1.9% ).10 The mean o f daily market return during the normal period is about 0.07%, while the number in the crisis period is about -23% . [Insert figure 2 and figure 3] The Price Impact from Market M odels. To measure the price impact o f foreign presence more formally, for the first event, we perform a modified market model as follow s:11 9 For the second event, since our focus is on all stocks in the JSX, we present market return calculated from the JSX Index. 10 The normal period is defined from January 1, 1995 to July 31, 1997, while the crisis period is defined from August 1, 1997 to August 31, 1998. Financial crisis starts from about the middle o f year 1997. The chart o f the JSX index (see figure I) shows a slowly increasing trend from early 1995 until the crisis started to hit Asian countries, including Indonesia, by about the middle o f 1997. 11 This methodology is basically a variant o f standard event-study analysis. In conventional eventstudy, parameter a and p are estimated by excluding days around the event (event-window). Then ‘residual’ is calculated using these parameters in the event window (Fama et al. 1969). However, this specification assumes that there is no exogenous shift in the parameters that may occur in evaluation period (i.e. event window). The modified specification above allows us to accommodate the possibility o f the shift, while still gives us the same pattern and timing o f excess return as those that would be obtained from a conventional event-study (Binder, 1985). Moreover, it is more convenient since we have only one step to perform. That is w e run the model (1) and (2) above to yield p4 (average abnormal return), while in a conventional event-study, we run market model using estimation period, 15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. R// = a + PiRm/+ p 2Rnv/+/>» + P3Rwf + p4eventl,+ psevent2,+ e n (1) For the second event, we use the following specification:12 R// = a + PiRm/+ P2 R11V/+/; + P3 RW, + p4eventl,+e,v Where R/r a Rm/ Rnv/+/; Rw/ event 1/ (2) = = = = = = daily return for stock i at day t intercept market return (calculated from the JSX index) at day t market return at day t+1 US market return as measured by S&P 500 return at day t dummy variable with certain values (depending on the horizon o f the days surrounding the events) during event days, and zero otherwise. event2/ = dummy variable with certain values (depending on the horizon o f the days surrounding the event) during event days, and zero otherwise. Qu = error term We use day -150 to + 150 to estimate the equation (1) and (2) above. To address the issue o f infrequent trading (Dimson, 1977; Williams and Scholes, 1977), we add lead and lag market return variable (Dimson, 1977). Prelim inary investigation shows that lead one period (t+1) for market return provides the most consistent (significant coefficient), so we use only one lead market return variable in both specifications. This result is similar to Berkman and Elaswarapu (1998) in their study. Note that in the specification (1) and (2) above, p 4 and ps measure cumulative abnormal returns over the event horizon. We define the horizon for eventl and event2 after spotting figure 2 and 3. For the first event, we use two definitions o f eventl and event2. First, we assign a value o f 1/3 for day -3 to -1 for eventl and zero otherwise, and we assign a value o f 1/3 for and then generate ‘the residual' in the event-window. 12 This specification provides better fit to the data in the second event. 16 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. day 0 to +2 for event2 and zero otherwise. The graph shows that in the short term, stock price tends to spike three days before the event and drifts low er after the event. Second, we take a som ewhat longer horizon to capture the movement o f stocks leading to the event Specifically, we assign a value o f 1/21 for day —20 to 0 for eventl and zero otherwise, and we assign a value o f 1/5 for day +1 to +5 for event2 and zero otherwise. For the second event w e use day -1 to +2 as the event window. Specifically we assign a value o f 1/4 during these days, and zero otherwise for variable eventl in the specification (2) above. Figure 3 shows that market moves up significantly from day -2 to +2, and then starts to drift downward. We pooled the observations across stocks from day -150 to +150 to obtain a better statistical power. [Insert table 3] Table 3 provides the result from the m arket model for the first event (specification (1) above). From column (1) o f table 3, we find that the announcement is associated with positive abnormal return from day -3 to -1 . During the three days leading to the institution o f foreign board, we have a positive abnormal return o f about 2.4% per-day. This num ber is similar to Henry (2000a) o f about 3.3% per month during the eight-month window leading to the implementation o f financial liberalization.13 The positive abnormal return during the pre-event period reverses to negative abnormal 13 For the sake o f comparison, we present the results from other papers investigating diiferent events using a similar methodology. Berkman and Eleswarapu (1998) find an average o f negative abnormal return o f 15% on Badla stocks when forward trading facility is abolished (event window starts from days 0 to days +15), and about 3% when forward trading facility is reinstated. Amihud et al. (1997) find an average o f positive abnormal return o f 5.5% when they investigate the value o f changing trading 17 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. return o f about 2.2% per-day during day 0 to +2, leaving the net impact during the six days from day —3 to +2 o f about 0.2% per-day. In column (2) o f table 3, when we extend the horizon for eventl to the period o f day —20 to 0, and day +1 to +5 for event2, we obtain stronger results. We have a positive abnormal return o f about 6% per-day during the 20 days leading to the event. This positive sign reverses to negative abnormal return o f about 1% per-day during the five days after the event. The net effect o f this event is about 5% per-day during day —20 to +5. Overall result strongly supports the finding o f w ealth increase associated w ith the presence o f foreign investors. Table 4 shows the result o f the m odified market model o f (2) above for the second event. Column (1) o f the table shows the result for 47 stocks used in the first event (stocks that have foreign board as o f December 1997), while column (2) shows the result for all stocks listed in the JSX (i.e. stocks with and without foreign board as o f December 1997). [Insert table 4] The table shows that liberalization announcement is associated with a positive abnormal return o f about 0.06% per-day during the three days leading to the announcement day. The numbers are statistically significant at conventional level. The magnitude o f the abnormal return in the second event is much smaller than that in the first event. This result seems to indicate that the liberalization during the crisis period provides a more lim ited price appreciation. mechanism (event window is days - 5 to days +30). 18 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. It is possible that there is an information leakage associated with the announcement o f the event This is especially the case for the announcement o f regulation (Binder, 1985). The process o f the establishment o f a regulation includes a lengthy discussion by regulators and other involved parties, which leaks the information to the m arket Since a government tends to introduce liberalizations simultaneously, Henry (2000a) points out the need to control for other liberalizations around the event date. It will be interesting to assess the effect o f these factors to investigate the direction o f possible bias related to our finding. In the case o f information leakage, if we can identify the start o f the leakage, then we may be able to isolate such effect. Because o f possible confounding effects, it does not seem possible to extend the event window into a much longer period. We try different definition for the length o f the event window. We use day -20 to 0 for event 1 to make it consistent with the specification for the first event. We find a negative coefficient for the event 1.14 For the case o f other liberalizations during the event, if we control for other liberalizations, we would have had smaller effects such as found in Henry (2000a). In either case, the analysis seems to support our conclusion o f a limited effect o f the liberalization during a financial crisis. Summary o f The Price Impact ofForeign Presence. The findings above suggest that the foreign presence is associated with positive abnormal returns (see also Henry, 2000a). The positive abnormal returns in the first event, which is exogenous to companies, 14 We also include event2 variable with the value o f 1/4 for day +3 to +6 and zero otherwise (similar to specification (1) above). We find an insignificant coefficient for eventl variable. Note that the cumulative abnormal return in figure 2 does not seem to suggest the information leakage long before 19 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. seem to be consistent with the prediction o f Stulz and W asserfallen (1995) model. But the positive abnormal returns in the second event, which is also exogenous to companies, seem to be inconsistent with Stulz and Wasserfallen (1995) prediction. Their model implies a negative abnormal return associated with foreign lim it reduction (see also a similar point by Lam, 1997). There is a significant difference, however. W hile Stulz and Wasserfallen (1995) view foreign limits as endogenous (established voluntarily), the changing o f foreign restriction in the Indonesian case is exogenously determined (by government regulation in this case). In both events, a more consistent interpretation is an association (a positive one) between the abnormal return and the presence o f foreign investors, rather than an association between the abnormal return and the change of foreign limits. The second event highlights the importance o f the setting o f liberalization. Liberalization conducted during a financial crisis seems to provide modest benefits. The price appreciation tends to be smaller than that o f the first event or that reported by other studies. Put differently, countries experiencing a financial crisis should explore other measures, rather than focusing on financial liberalization per-se.15 6.2. Liquidity, Efficiency, and The Abnormal Returns in the Event Period In this section, we investigate the determinants o f abnormal returns calculated in sub-section 6.1.16 Since our focus is on the potential benefits and costs associated the event. For example, a recent paper suggests that there seems to be an association between corporate governance in emerging markets and the severity o f the Asian financial crisis (see Johnson et al. (2001)). Thus improving corporate governance may help combat the severity o f the financial crisis. We estimate model (1) and (2) above for each company to obtain p 4 coefficient (abnormal return) for each company. 20 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. with foreign investors, we identify several variables that may proxy for the benefits and costs o f foreign investors. Specifically, we focus on efficiency and liquidity variables as a potential explanation to the abnormal returns. Since the events we use are about the changing o f foreign restrictions, as control variables, we also use several variables found to be the determinants o f the premium o f the prices on the foreign board over the prices on the regular board. 6.2.1. Liquidity We calculate measures o f liquidity by introducing several variables: daily trading volume, daily trading value, daily trading volume (value) scaled by market trading volume (value), and market depth. (1) Daily trading volume and value: To control for possible changes in the market, we also calculate adjusted trading volume (value), which is trading volume (value) divided by market trading volume (value). The adjustment is especially im portant for the second event, since this event was introduced when the Indonesian m arket experienced a financial crisis. We calculate liquidity variables for the pre-event period (day -150 to -31) and the post event period (day +31 to +150) for each company. (2) Market depth or liquidity ratio: M arket depth is calculated as follow s (Amihud et al. 1997): Mdit = 2 Volume it / 2 |Rit| Where 21 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Mdu Volume it I Rit | = market depth o f stock I at time t = trading volume at day t for stock I = absolute return o f stock I at day t M arket depth measures additional volume per-unit return. A higher market depth implies a higher liquidity (Amihud et al. 1997). This measure is similar to Kyle (1985) who defines market depth as the trading volume per-unit o f price change. As in trading volume, we calculate the market depth in the pre and post event periods for each company. 6.2.2. Efficiency variables We calculate measures o f efficiency variables using the variance o f daily returns, residual, and the variance o f residual o f the following market model (Amihud and M eindelson, 1997; Berkman and Eleswarapu, 1998): R/f = ai + (hi Rm/+ (hi Rmt+/ + eIf ...................... (3) We estimate separate regressions for the pre and post event periods for each company. The variance o f daily returns and the variance o f residual o f model (3) above are used to proxy for the level of noisiness in the market. For each company, we have pairs of noise measures in the post-event and pre-event periods. 6.2.3. Foreign restriction variables The literature on foreign ownership restriction includes several variables to explain the behavior o f the premium o f the prices on the foreign board over the prices 22 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. on the regular board. We include them in this study as control variables. (1) Size: This variable is used to proxy for information availability (Merton, 1987). Every year we calculate size as the closing price at the end o f the year times the number o f shares outstanding at the end o f the year. Then we average the numbers to obtain the size variable. It is well documented that foreign investors prefer large and well-known companies (Kang and Stulz, 1997). An asset with a larger base o f informed investors sells at higher price than that with a smaller base (M erton, 1987). We expect to have a positive association between size and abnormal returns for both events. (2) Trading volume on foreign board: Bailey and Jagtiani (1994) argue that trading volume on the alien board (foreign board) can be interpreted as a measure o f foreign investors' familiarity with domestic stocks and liquidity as well. They calculate trading volume on the alien board without adjustment and with the foreign and regular board adjustments (i.e. divided by the total trading volume on the regular and foreign board). We use unadjusted trading volume on foreign board since our preliminary investigation shows that this variable tends to provide stronger explanation on abnormal returns than the adjusted ones. We expect to have a positive association between this variable and the abnormal returns during the event period. (3) Relative Supply: Following Bailey and Jagtiani (1994), we calculate stock relative supply as follows: Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. n i .• n • Volifcreijn Relative Supply, = Vo, ^ - Volj Where subscript i refers to stock i, subscript foreign refers to trading volume on foreign board, Vol, refers to total trading volume o f stock i, TotVol refers to total volume across stocks. This number measures the degree o f tightness o f foreign demand relative to the supply o f stocks (Bailey and Jagtiani, 1994). The smaller the number for a stock, the tighter the demand for the stock relative to its supply. Investors are willing to pay premium for the stock with this characteristic. We expect to have a negative relationship between this variable and the abnormal returns, that is, the tighter the demand for a stock, the larger premium investors are willing to pay. 6.2.4. Descriptive Statistics o f Liquidity an d Efficiency Variables Table 5 presents descriptive statistics o f the liquidity and efficiency variables for the first event. The most surprising result is that the market becomes less liquid in the post event period. The variables used as proxy for liquidity show sm aller numbers in the post-event period than in the pre-event period. The differences for adjusted trading volume and trading value are statistically significant at conventional level, while unadjusted trading volume and trading value do not show strong differences. This result seems to suggest that foreign investors reduce liquidity, instead of increasing it. But since the setting is foreign presence that results in foreign ownership restriction, we are m ore inclined to conclude that foreign restriction leads to decreased liquidity. This finding is consistent with Lau et al. (1997) for Singapore market or 24 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Domowitz and Madhavan (1998).17 They find that stock liquidity tends to decrease after the introduction o f foreign limits or inter-listing. [Insert table 5] The information related variables show expected signs. W e expect that foreign investors create a more efficient market; hence we expect that the variables used to proxy for noises have sm aller values in the post-event period than in the pre-event period. We find that differences in the variables used to proxy for market noise show negative values. For example, the difference o f the variance o f residual shows a negative value o f 4.26 x 10'13 and statistically significant at 10% level. Mean o f the daily return in the post-event tends to be lower than that in the pre-event. Assuming that the actual daily return can be used as proxy for expected return, we can interpret that the cost o f capital tends to decrease in the post-event period. A lower risk premium, a lower risk-free rate, or both may lower the cost o f capital (Henry, 2000a). However we do not investigate this issue further. Tables 6a and 6b present descriptive statistics o f the efficiency and liquidity variables for the second event. Table 6a presents the result for the companies used in the first event (companies that have foreign board as o f December 1997). Table 6b presents the numbers for all companies in the JSX as o f August 1998. We exclude non­ stock securities. Table 6a shows that liquidity tends to increase in the post-event period 17 As in Domowitz and Madhavan (1998), it is possible to have a trading migration from one market to another one (in the JSX case, from regular board to other boards such as negotiated board). For example, foreign investors tend to be dominant participants in cross-trade board, which make up the 25 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. compared to that in the pre-event period. Trading volume and trading value, either adjusted or unadjusted, show higher values in the period after the announcement o f financial liberalization. But only unadjusted trading volume difference shows statistical significance at conventional level. In the post-event period, market tends to be noisier. Variables to proxy for market noise have larger values in the post-event compared to those in the pre-event period. [Insert tables 6a and 6b] Table 6b shows that liquidity tends to decrease in the post-event period compared to that in the pre-event period. Efficiency variables tend to increase significantly in the post period. The increase tends to be greater than that in table 6a. Table 6a and 6b seem to suggest that liberalization does not improve liquidity and efficiency (see for example, Kim and Singal, 2000). However the setting o f financial liberalization in our study is somewhat unique, that is, the liberalization conducted in a ■o financial crisis period. Our study highlights the importance o f the setting o f the liberalization when investigating the effect o f liberalization. largest negotiated board. Unfortunately PACAP database has only regular and foreign boards. 18 Henry (2000) and Kim and Singal (2000) do not specify specifically the setting o f the liberalization they investigate. Kim and Singal (2000) use Indonesia as one o f the sample in their study, but their date o f liberalization is September 6, 1989, when Ministry o f Finance allows foreigners to purchase up to 49% o f listed companies, excluding bank stocks. The late o f 1980’s and early 1990’s witness one o f the strongest 'bull' market in Indonesia. Thus the period is very different from ours. 26 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6.2.5. Cross-Sectional Regression o f Abnormal Returns on Efficiency and Liquidity Variables To investigate jointly the effect o f efficiency and liquidity variables on the abnormal returns, we perform a cross-sectional regression using abnormal return (coefficient 04 o f model (1) and (2) above for each company) as dependent variable and several variables explained in the previous section as independent variables. The Cross-Sectional Regression o f The First Event. Table 7 summarizes the result for the first event. N ote that we use the abnormal returns estimated from day —20 to day 0 since previous section shows that the abnormal returns estim ated from this period tend to provide a stronger result [Insert table 7] In regression (1) of table 7, we include only variables used to proxy for liquidity. Specifically we use the changes in adjusted daily trading volume CdTRDVOL/) and market depth (JM KTDEP,). Since liquidity tends to decrease in the post-event period, we define the changes in liquidity as the numbers in the pre minus those in the post event. We expect to have negative coefficients, suggesting that the market discounts the price for liquidity decrease. We do not observe significant results. In regression (2) o f table 5, we include only variables used to proxy for market noisiness. Specifically we use the changes in the variance o f residual estimated from the market model (3) above GdRESVAR,) and the variance o f daily return 27 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (^VARRET,). Since we expect to have a noisier market in the pre-event period, hence larger numbers in the pre than those in the post-event period, we define the changes in the efficiency variables as the numbers in the pre-event minus the numbers in the post­ event period. Under this definition, we expect to have positive coefficients, suggesting that the market rewards increased efficiency. We have m ixed signs for JRESVAR, (positive number) and £ rn ^ © fN —o Table 7 (Continued) VOo J3 a- O [...]... f foreign restriction, we view them as the signs o f foreign presence Hence we focus on the effect o f foreign presence, rather than on the effect o f foreign restriction changes Our study belongs more closely to the literature on the effect o f liberalization on domestic markets (Henry 2000a,b; Kim and Singal, 2000; Domowitz and Madhavan, 1998), rather than to the literature on foreign restrictions... f foreign restrictions are relatively common in international market, either in developing or developed countries For example, Thailand market has alien and main board (Bailey and Jagtiani, 1994) Alien board is similar to foreign board in the JSX In Switzerland market, registered stocks could be held only by domestic investors, while foreign investors could trade in two classes o f stocks, voting and. .. become binding, foreign investors m ust buy shares from other foreign investors Foreign board was created to facilitate the trading o f foreign owned shares among foreign investors We view the date when the foreign board appears for the first time as ’an official’ sign o f foreign presence We focus on the first dates when foreign limits are reached for the following reasons Errunza and Losq (1989)... voluntarily), the changing o f foreign restriction in the Indonesian case is exogenously determined (by government regulation in this case) In both events, a more consistent interpretation is an association (a positive one) between the abnormal return and the presence o f foreign investors, rather than an association between the abnormal return and the change of foreign limits The second event highlights the... adjustment and with the foreign and regular board adjustments (i.e divided by the total trading volume on the regular and foreign board) We use unadjusted trading volume on foreign board since our preliminary investigation shows that this variable tends to provide stronger explanation on abnormal returns than the adjusted ones We expect to have a positive association between this variable and the abnormal... different approaches Previous research on the effect o f liberalization focuses on the lower cost o f capital as the result o f market integration When the market is opened to foreign investors, risk sharing between domestic and foreign investors occurs, resulting in a lower cost o f capital o f domestic firms (Errunza and Losq, 1985; Eun and Janakiramanan, 1986; Alexander, 1987; Stulz, 1991) Henry (2000a)... behavior among foreign investors in the Korean m arket during the crisis period They suggest that foreign investors do not destabilize the m arket In the normal period, foreign investors seem to engage in positive feedback trading or momentum trading (Choe, Kho, and Stulz, 2000; Grinblatt and Kelohaiju, 2000) and herding behavior (Choe, Kho, and Stulz, 2000) Empirical research focusing on the contribution... the cross-sectional determinants o f the abnormal return during foreign presence, and hence sheds light on the contribution o f foreign investors on domestic markets We find that the presence o f foreign investors is associated with positive abnormal return The changes in m arket efficiency seem to explain the abnormal returns better than the changes in m arket liquidity Liberalization conducted during... copyright owner Further reproduction prohibited without permission 2 Related Literature This paper extends the literature on the effect o f liberalization on domestic markets Perhaps the spirit o f this paper is closest to that o f Henry (2000a), Kim and Singal (2000), and Domowitz and Madhavan (1998) Each o f these papers investigates the effect o f foreign investors on domestic markets using different... The revaluation effect of foreign investors seems to be consistent with the general prediction o f international capital asset pricing Opening a domestic market to foreign investors allows risk sharing that results in a lower cost o f capital From a market integration perspective, after the opening o f a domestic market, the cost o f capital will be proportional to the covariance o f a domestic company’s