Information content and policy implications of stock splits, new evidence from the saudi arabian capital market

... permission of the copyright owner Further reproduction prohibited without permission Information Content and Policy Implications of Stock Splits: New Evidence from the Saudi Arabian Capital Market. .. consisting of the minister of finance and national economy, the minister of commerce and industry, and the governor of the Saudi Monetary Agency The purpose of this research is to study the impact of. .. the market response to the announcement of the new policy and the magnitude of abnormal returns surrounding the announcement date The study then investigates whether the policy appears to have information

INFORMATION CONTENT AND POLICY IMPLICATIONS OF STOCK SPLITS NEW EVIDENCE FROM THE SAUDI ARABIAN CAPITAL MARKET Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Information Content and Policy Implications of Stock Splits: New Evidence from the Saudi Arabian Capital Market A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy By Ali Mofarreh Ali Serhan May 2005 University of Arkansas Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI N um ber: 3201536 C o pyright 2005 by Serhan, Ali M ofarreh Ali A ll rights reserved. IN F O R M A T IO N TO U SER S T h e quality o f th is re productio n is d e p e n d e n t upon th e q uality o f th e copy subm itted. B roken o r in distinct print, colored or p o o r q u ality illustrations and photog ra phs, print ble ed-thro ugh, su b stand ard m argins, and im proper a lig n m e n t can a d ve rsely a ffe ct reproduction. In the unlikely e ve n t th a t the a u th o r did not send a co m p le te m anuscript and th ere are m issing pages, th e se will be noted. A lso, if unauthorized cop yrig ht m aterial had to be rem oved, a note w ill indicate th e deletion. ® UMI UMI M icroform 3201536 C op yrig h t 2006 by P roQ uest Inform ation and Learning C om pany. A ll rights reserved. T his m icroform edition is protected a g ainst u nauth orized copying u n d e r Title 17, U nited S tates Code. P ro Q ue st Inform ation and Learning C om pany 300 North Z eeb Road P.O. Box 1346 A nn A rbor, Ml 4 8 106 -1 346 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. © 2005 by Ali Mofarreh Serhan All Rights Reserved Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. This dissertation is honorably dedicated to my parents. iv Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ACKNOWLEDGEMENTS Above all, all praises and thanks are due to Allah, the most merciful and the most gracious, the lord of mankind, who blessed me with his guidance, boundless bounties, ample mercy, and endless help and support. I am sincerely very thankful to him for his favors upon me in my life and throughout my educational journey. In him I trust. My parents deserve a special and sincere acknowledgement for keeping their hearts attached to me and my family overseas. Without their permanent support, patience, concerns, and constant prayers, I would not be able to attain my goals. During my academic journey that ultimately led to finishing this dissertation, there were many people who have made significant contributions to its success, and they deserve acknowledgement and appreciation. I am highly indebted to my dissertation advisor and committee chair, Professor Carolyn M. Callahan, for her profound insights, constructive ideas and valuable inputs, encouragement and support, concern about me and my family, and professional mentorship not only in my dissertation stage but also during my entire doctoral program. I am also grateful to the other members of my dissertation advisory committee, Professor Gary D. Ferrier, and Professor Rodney E. Smith for their valuable inputs, encouragement, and support. Though words are truly inadequate in capturing her real sacrifice, my wife, Nourah, deserves sincere thanks and great appreciation for her tireless support, assistance, patience, and sharing my pleasure and my pressure. My dissertation would have never been accomplished without her. Sincere thanks also go to my children Adel, Abdulkareem, Sarah, Hassan, Omar, Malik, and Yunis for their patience and sympathy. v Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. My gratitude goes to my brothers, sisters, and the rest of my broad family members and friends, who permanently were very concerned about me and my family. Their support and prayers for me are appreciated. Finally, I am also appreciative and thankful to all the staff of the Accounting Department at the Walton College of Business, especially the head of department, Professor Karen Pincus, for their help and support. vi Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLE OF CONTENTS ACKNOWLEDGEMENTS.................................................................................................... v 1. INTRODUCTION............................................................................................................... 1 2. MARKET OVERVIEW..................................................................................................... 4 2.1 Historical Glance of the Market....................................................................................4 2.2 Market Growth...............................................................................................................6 2.3 Market Structure & Performance................................................................................10 2.4 Trading Process............................................................................................................13 2.4.1 Types of Trading Orders...................................................................................... 18 2.4.2 Trading Commission............................................................................................ 19 2.5 Market Indexes........................................................................................................... 20 3. THE SPLIT POLICY AND ITS IMPORTANCE.......................................................... 21 4. LITERATURE REVIEW AND THEORY.....................................................................23 4.1 Information Content of Stock Splits.......................................................................... 24 4.1.1 Stock Split Decision Incentives.......................................................................... 25 4.1.1 .a Signaling Hypothesis...................................................................................26 4.1.1 .b Trading Range Hypothesis.......................................................................... 28 4.1.1. c Liquidity Hypothesis.................................................................................30 4.1.1 .d Survey Research on Stock Splits................................................................31 4.1.2 Factors Determining the Market Response to Split Announcements............... 32 4.2 Stock Changes around Stock Split..............................................................................36 4.2.1 Ownership Structure............................................................................................ 36 4.2.2 Volatility Changes............................................................................................... 37 4.2.3 Number of Trades, Turnover, and Volume........................................................ 37 5. THE FIRST STUDY........................................................................................................ 38 5.1 Hypotheses Development........................................................................................... 38 5.2 Research Methodology............................................................................................... 39 5.2.1 Study Sample Selection.......................................................................................39 5.2.2 Data Collection.................................................................................................... 40 5.2.3 Research Design.................................................................................................. 41 5.3 Analysis and Results................................................................................................... 43 5.3.1 Sample Characteristics.........................................................................................43 5.3.2 Discussion.............................................................................................................51 5.3.3 Results...................................................................................................................61 5.4 Sensitivity Tests...........................................................................................................63 6. THE SECOND STUDY................................................................................................... 64 6.1 Hypotheses Development......................................................................................... 64 vii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6.2 Research Methodology............................................................................................... 68 6.2.1 Study Sample Selection.......................................................................................68 6.2.2 Data Collection.................................................................................................... 68 6.2.3 Research Design and Model Specification........................................................ 69 6.3 Analysis and Results................................................................................................... 72 6.3.1 Sample Characteristics......................................................................................... 72 6.3.2 Descriptive Statistics of Model Variables..........................................................81 6.3.3 Correlations...........................................................................................................83 6.4 Regression Analysis and Discussion......................................................................... 86 6.4.1 Statistical Issues................................................................................................... 86 6.4.1.a Outliers..........................................................................................................86 6.4.1.b Multicollinearity.......................................................................................... 86 6.4.1 .c Heteroskedasticity..................................................................................... 87 6.4.2 Tests of Hypotheses............................................................................................. 87 6.5 Sensitivity Tests...........................................................................................................92 6.6 Conclusion and Implications......................................................................................92 7. THE THIRD STUDY........................................................................................................93 7.1 Hypotheses Development........................................................................................... 93 7.1.1 Ownership Structure............................................................................................ 93 7.1.2 Number of Trades, Turnover, and Volume.........................................................94 7.2 Research Methodology............................................................................................... 95 7.2.1 Study Sample Selection.......................................................................................95 7.2.2 Data Sources.........................................................................................................96 7.2.3 Research D esign.................................................................................................. 97 7.3 Analysis and Results................................................................................................... 98 7.3.1 Sample Attributes................................................................................................ 98 7.3.2 Tests of Hypotheses and Discussion................................................................. 107 7.3.2.a Ownership...................................................................................................107 7.3.2.b Trading Volume......................................................................................... I l l 7.3.2.C Number of Trades...................................................................................... 115 7.3.2.d Turnover......................................................................................................118 7.4 Conclusion and Implications.................................................................................... 121 8. LIMITATIONS................................................................................................................ 122 9. CONTRIBUTION........................................................................................................... 122 REFERENCES.................................................................................................................... 124 viii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1. INTRODUCTION “A stock split (or forward split) is an event in which the firm decides to divide each share of stock into multiple shares. Conversely, a reverse stock split is an event in which a number of shares is combined into one new share” (Wu and Chan 1997). In traditional theory, stock splits can be considered merely cosmetic changes in the equity that have no economic significance to shareholders. However, empirical research does not support this theory as stock split announcements are associated with positive abnormal returns. The impact of stock split signals and their implications on share returns and other investment aspects have been the focus of financial researchers for three decades as evidenced by the large body of accounting and finance literature associated with stock split issues. The accounting and finance literature repeatedly documents a market reaction to stock splits. The literature also provides plausible explanations of the motives driving firms to initiate stock splits. One of the explanations provided is that stock splits serve as a device to signal private information known by managers to market participants and other interested parties (Brennan and Copeland, 1988) and (McNichols and Dravid, 1990). Another explanation suggests that firm managers intend to maintain the stock price at a “preferred” or “optimal” trading range (Lakonishok and Lev 1987), whereas Brennan and Hughes (1991) and Ikenberry et al. (1996) suggest that stock splits cause investment analysts and dealers to reevaluate their firms and discover the private information conveyed by managers (dual purpose). The financial literature also reveals the factors that play a significant role in determining the magnitude of information content of stock splits such as firm characteristics, split characteristics, and surrounding 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. environment characteristics, including the degree of information richness in the trading environment. This study extends the literature on stock splits to an international stock market, the Saudi Arabian capital market. It investigates a unique stock split policy urging publicly traded firms to split their par values two-for-one. This policy was adopted on December 14,1997, by the Ministerial Committee for Share Trading Supervision (MCSTS), a governmental committee consisting of the minister of finance and national economy, the minister of commerce and industry, and the governor of the Saudi Monetary Agency. The purpose of this research is to study the impact of this new public policy on the market. I will approach this issue from three different perspectives. Each perspective in itself comprises an independent research study. The first study investigates the market response to the announcement of the new policy and the magnitude of abnormal returns surrounding the announcement date. The study then investigates whether the policy appears to have information content for market participants. The second study models and relates the anticipated information content found in the announcement of the split policy to the factors mentioned in prior literature to be responsible for driving split information content and determines the magnitude and the direction of relationships among them. The third study concentrates on investigating the changes made to the market structure and whether or not the stated goals of the policy have been achieved. In this research, decreeing stock split is empirically found to be associated with information content. However, evidence gathered shows that the market reaction is mixed, not in one direction. Some firms reacted positively (CAR=2.11%) considering the split 2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. decree as good news, while some other firms reacted negatively (CAR= -2.55%) considering it as bad news. At the sector level, the cement industry is the only industry that demonstrates positive cumulative abnormal returns. On the other hand, banking, manufacturing, and service sector, in general, demonstrated negative cumulative abnormal returns. Surprisingly, agriculture and electricity sectors demonstrated no reaction. No evidence that sectors differ in their reactions is found. Further, the magnitude of abnormal returns is not associated with share market price. Examining the determinants of the market reaction to decreased stock split reveals that only firm size and the book to market value ratio are significant. Firm size is positively related to the market reaction, while book to market ratio is negatively related to market reaction. These results imply that as the size of a firm increases, the magnitude of abnormal return associated with the market reaction to the decreed stock split increases, while as the book to market value ratio of a firm increases, the magnitude of abnormal return decreases. The impact of firm size is greater in magnitude than that of book to market ratio. Finally, the stock split policy has attained some of its intended goals. First, there is no empirical evidence that the total number of shareholders increased by the implementation of the policy. Dissimilarly, volume is found to increase significantly for some firms and decrease for other firms in the market. Similarly, number of trades is also found to increase significantly for some firms and decrease for other firms in the market. However, turnover, generally, has declined after the split. It can be concluded that the execution of stock split has dual effects. These results have important implications for Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. both policy makers and investors as they suggest that a stock split policy has differential market effects that may be industry specific. The remainder of this dissertation consists of eight sections. Section 2 provides an overview of the Saudi stock market, while section 3 presents a detailed description of the stock split policy enforced by the Saudi government. Section 4 furnishes a brief review of previous stock split literature and provides a theoretical background for all three papers’ research hypotheses. Section 5, section 6, and section 7 are devoted to discuss separately each study of the research papers of the dissertation. Each section presents research hypotheses, methodology, analysis, and findings for each of the three respective studies. Section 8 reports the caveats that could limit the scope of the results. Finally, section 9 indicates the incremental contribution made by conducting this research. 2. MARKET OVERVIEW 2.1 Historical Glance o f the Market The first Saudi joint stock company, the Arab Automobile Company, was established in the mid 1930’s. There were about 14 stock companies in 1975. In the late 1970’s, the number of large publicly traded companies and joint venture banks increased due to Saudization1 of foreign banks and the rapid economic expansion. During this period, major initial public offerings were made to generate the required capital for these new firms. In 1984, a Ministerial Committee, consisting of Minister of Finance and National Economy, Minister of Commerce, and Governor of Saudi Arabian Monetary Agency 1 Saudization in this context means changing the nationality o f firms to Saudi nationality through selling more than 50% o f their equity to Saudi citizens to become national companies as opposed to foreign firms. 4 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (SAMA), was formed by a royal decree to develop and set the standards, regulations, and rules that promote the market operations. SAMA was delegated and charged with the day-to-day regulation and supervision of the capital market. Share-trading intermediation was also restricted to commercial banks since they operate under the umbrella of SAMA in order to facilitate and improve the regulatory framework. Accordingly, commercial banks established the Saudi Share Registration Company (SSRC) for the purpose of providing central registration facilities for joint stock companies and settling as well as clearing all shares trading transactions occurring in the market. An automated clearing and settlement system was introduced in 1989 via sophisticated computer systems. The Electronic Securities Information System (ESIS), which was developed and operated by SAMA, was also introduced to the market in 1990 in order to provide market participants with all trading information such as bid-ask share prices, number of trades, volume,..., and daily market index. Moreover, ESIS was developed to enable investors to enter their buying or selling orders and to follow up their status. In October 6, 2001, “Tadawul,”2 the second electronic generation of securities trading, clearing, and settlement system was launched to replace the old system (ESIS). “Tadawul” is a highly sophisticated computer system. It allows electronic trading accounts for all investors and provides depository service of traded shares. This system also permits selling and buying several times during the day, along with other advanced capabilities. Tadawul thus integrated the ESIS daily trading system with that of SSRC for clearing and settlement into a comprehensive electronic system with more powerful capabilities. 2 Tadawul is an Arabic word, which means trading. It refers to the name o f the current security trading system. 5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.2 Market Growth The market growth in capital, number of transactions, volume, and Riyal’s value of traded volume increased significantly. Between 1990 and 1997, the year in which stock split was adopted by the government, market capitalization increased by 130%, and the all share index also increased by 99.8%. Volume also increased by 1747%, transactions by 441% and value of traded shares by 1309%. In contrast, between 1990 and 2003, those market indicators have increased dramatically. Market capitalization has increased by 508%, the all share index by 352.9%, volume by 32641%, transactions by 4327%, and value of traded shares by 13300%. Table 1 depicts market growth in these indicators between 1990 and 2003, showing the gradual year- by- year growth in the market. TABLE 1 Development in Market Indicators 3 Year 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Transactions (thousands) 85 90 272 319 357 292 284 460 377 438 498 605 1034 3763 Shares Traded (millions) 17 31 272 60 152 117 138 314 295 528 555 692 1736 5566 Value of traded shares (billions) 4.403 8.527 13.698 17.36 24.871 23.226 25.397 62.06 51.509 56.578 65.292 83.601 133.787 596.510 Market Cap.(billions) 97 181 206 198 145 153 172 223 160 229 254 275 281 590 All Share Index 980 1,788 1,889 1,793 1,282 1,368 1,531 1,958 1,413 2,058 2,258 2,430 2,518 4,438 3 Monetary values are in Riyal, which is the local currency o f Saudi Arabia. Exchange rate is 3.75 Riyals for a US Dollar. 6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The number of transactions has grown year after year from 85000 in 1990 to 460,000 in 1997 and eventually to 3,763,000,000 in 2003. The graph on Figure 1 reflects the steadfast growth in the number of transactions, and the huge jumps that happened in 2002 and 2003. FIGURE 1 Growth in Transactions w | | c H o | 4000 3500 3000 2500 2000 1500 1000 500 z 0 s | Transactions Volume (number of shares traded) increased between 1990 and 1992 from 17 million shares to 272 million shares, an increase of 15 times. From 1993 to 1996, it declined from the level reached in1992 with an increasing trend. From 1997 on, the volume increased dramatically until it reached five billion and five hundred sixty six million shares in 2003, an increase of almost 17 times from 1997. The chart on Figure 2 represents the growth took place in the number of shares traded. The huge jump happened in year 2002 and 2003. 7 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. FIGURE 2 Growth in Shares Traded Shares Traded Year Market capital also grew between 1990 and 1997. It grew from 97 billions Saudi Riyals (SR) in 1990 to SR 223 billions in 1997, an increase of 130%. Between 1997 and 2003, market capital grew vividly from SR 223 billions in 1997 to 590 billions in 2003, with an increase of 165 %. Market growth is represented in the chart on Figure 3. FIGURE 3 Growth in Market Capital 700 600 500 400 300 200 100 « -♦— Market Capital Years 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The value of traded shares grew as the volume in the market grew. This value grew from SR 4.403 billions in 1990 to SR 62.06 billions in 1997, with an increase of approximately SR 58 billions (1309 %). Furthermore, the value increased more between 1997 and 2003 to SR 596 billions with an increase of approximately SR 534 billions from the value in 1997 (860 %). The historical growth in the value of traded shares is depicted in Figure 4. FIGURE 4 Growth in the Value of Traded Shares Value of Traded Shares o 05 05 T— 05 CM o> T— a> T— 05 co 05 05 T— o> o> T— LO 05 05 v- CO 05 05 05 05 T — T— 05 o 05 o 05 T —o CM CO 05 05 vo o CM CM O O CM CO O O CM Year The market index rose by 99.8% from 1990 to 1997. Return on the market index, therefore, almost doubled in less than eight years. Surprisingly, the market index also increased 2480 points in the year 2003 from its level in 1997. This skyrocketing rise represents 127% increase over index reading in 1997. Totally, the market index rose 353% between 1990 and 2003. The chart depicted in Figure 5 reflects this trend. 9 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. FIGURE 5 Growth in the Market Index 5000 n * 4000 A 1 3000-] | 2000 -1 S 1000 -j 0 H index 1--------------1-------------- 1--------------1-------------- 1--------------1---------------1--------------1-------------- 1--------------1-------------- 1--------------1-------------- r Years 2.3 Market Structure & Performance The Saudi capital market is the largest capital market in the Middle East. It operates in the largest economy in the region. It consists of seven sectors: banking, manufacturing, cement, services, electricity, agriculture, and telecommunications sector, which has been launched recently with new two telecommunications firms. As of December 31, 2003, the total number of listed firms in the Saudi capital market was 70 firms that are distributed among the market sectors. The banking sector includes 9 national banks. The manufacturing sector includes 24 companies. The cement sector includes 8 companies. The service sector includes 18 companies. The telecommunications sector included one company. The agricultural sector includes 9 companies. Finally, the electricity sector includes one large company that came into existence from merging several working electricity companies in all provinces of the country. 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. At the end of 2003, based on Tadawul 2003 report, the entire market capitalization was approximately SR 590 billions. The largest sector by market capitalization is the banking sector, which at the end of this year was capitalized at SR 176.2 billion, 30% of the total market capitalization. The second largest sector by market capitalization was the manufacturing sector with SR 134 billions, 22.73%. The telecommunications sector occupied the third with SR 126.83 billions, 21.5%. The fourth sector in capitalization was the electricity sector with SR 84.4 billions, 14.3% of market capitalization. The Cement, service, and agricultural sectors occupied the fifth, the sixth, and the seventh respectively, with SR 43 billions (7.3%) for the cement sector, 23 billions (4%) for the service sector, and 2.5 billions (0.42%) for the agricultural sector. In terms of transactions, the number of transactions executed in the market during 2003, based on Tadawul report, was 3.8 million transactions. The most active sector was the service sector with about 1.22 millions transactions, 32% of total transactions executed in the market. The least active sector was the cement sector with about 109 thousands transactions, 2.9 % of transactions executed in the market. The volume traded in the market during 2003, based on the same previously mentioned report, was approximately 5.6 billion shares. The most active sector in volume traded was the service sector with about 2.3 billion shares, 41% percent of volume of the entire market, whereas the least active sector with respect to volume traded was the banking sector with approximately 87 million shares, 1.6%. On the other hand, the value of shares traded in the whole market was approximately SR 596.5 billions where the manufacturing sector had the biggest stake in this value by attaining SR 171.55 billion in 11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. value traded, 28.8 % of the market traded value. The least sector in value traded was the agricultural sector with 8.5 billions, 1.4% of the market. The total number of outstanding shares of the entire market as of December 31, 2003 is 2,347,147,371 common shares. The largest sector in number of shares was the electricity sector, which had about 833.318 million shares, 35.5% of the total number of shares in the market. The least sector in number of shares was the agricultural sector with 36 million shares, 1.5 % of the total number of shares in the market. During the fiscal year of 2003, the total earnings attained for all listed firms in the market added up to SR 18,991.85 millions, the portion of each market sector was as follows: The banking sector SR 8.259 billions, 43.4% of the market earnings The manufacturing sector SR 3.671 billions, 19.3% of the market earnings The cement sector SR 2.171 billions, 11.4% of the market earnings The service sector SR 276.83 millions, 1.4 % of the market earnings The electricity sector SR 1.077 billion, 5.6% of the market earnings The telecommunications sector SR 3.545 billions, 18.66% of the market earnings The agricultural sector SR -10.09 millions, a mixture of some gains and some losses The banking industry is, therefore, the most profitable sector in the market, followed by the manufacturing industry and the telecommunications industry. Shareholders’ equity at the end of 2003 for the entire market reached SR 182.8 billions. The distribution of this amount across market sectors makes the manufacturing sector at the top of all sectors with SR 48.05 billions, 26.3% of the market. The banking sector is second with SR 43.2 billion of shareholders’ equity, 23.6% of the market equity. 12 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The third sector is the electricity sector with shareholders’ equity of SR 42.8 billions, 23.4 % o f the market equity. The telecommunications sector, service sector, cement sector, and agricultural sector follow in terms of shareholders’ equity respectively. The dividends distributed to shareholders for the year 2003 by the entire market were SR 11.348 billions. The banking sector alone distributed SR 6.4 billions, the largest dividends among market sectors. This amount represents 56.4 % of the total market dividends. The least sector in dividend distribution was the agricultural sector with dividends of only 13.75 millions. In 2003, Earning per share (EPS) ranged from SR 41.05 to SR 5.51 for the banking sector. In the manufacturing sector, the range of EPS was from SR 16.78 to -SR 7.53. The cement sector had a range of EPS of SR from 35.56 to SR 5.63. EPS of the service sector ranged from SR 19.73 to SR -14.12. The electricity sector had EPS of SR 1.97 as it consists only of one large company. The telecommunications sector also consists of one company with EPS of SR 11.82. Finally, the agricultural sector’s EPS ranged from SR 2.86 to SR -2.32. The banking, cement, and telecommunications sectors were the best sectors in the market in terms of EPS. Furthermore, all listed firms in the three sectors generated profit. In contrast, the manufacturing, service, and agricultural sector had low EPS in general with some unprofitable firms. 2.4 Trading Process The trading process in the stock market has evolved gradually with the advancements introduced to the working mechanism of the market. So, if we are to understand the progress made in trading process, we need to trace back the advancements made to the market mechanism. 13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Before early 1980’s, the stock market was informal. The trading process during that period was primitive and manual. For investors to buy or sell at this time, they needed to meet personally or through informal intermediaries to brokers and trade with each other the number of shares they wanted to trade. The proof of ownership of shares was “certificate of shares”, which indicated the number of shares owned by an investor in the respective company. The certificate was issued by the joint stock companies. When a trade was agreed upon by the buyer and the seller, they or their representatives went to the company in which the shares were traded to transfer the ownership of shares from the seller to the buyer. A new certificate of shares would then be issued for the buyer including the number of shares traded. Also another certificate of shares would be issued for the seller if not all of the shares in the earlier certificate of shares were sold. The trading process thus took nearly a week, including the time needed by departments of shareholders relations in joint stock companies to transfer the ownership. Trading during this era was ineffective, inconvenient, impractical, and time consuming. Meanwhile, it did not enable investors to trade large number of shares easily. All these trading barriers drove high transaction costs. In 1984, a Ministerial committee for share trading supervision (MCSTS) was appointed by the Saudi government to include the minister of finance and national economy, the minister of commerce, and the Governor of SAMA. The goal of this committee was to supervise share trading and develop the rules and regulations that govern share trading market. MCSTS delegated the authority of daily trading supervision and regulations of this trading to SAMA. SAMA, in its turn, established a new division, Shares Trading Control, within its organizational structure to cope with the new 14 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. delegated task. In the meantime, SAMA delegated share intermediation to banking sector. “Banks are, however not allowed to buy and sell stocks for their own accounts, act as market makers, or maintain inventory for trading purposes”(Henry Azzam, 1997). SAMA delegated share intermediation to banks because the banking sector is under its direct supervision and owns a widespread network of banking branches in all over the country that can serve as the best intermediaries for share trading. The decision of delegation of intermediation to the banking sector, in my opinion, is effective and efficient for three reasons. First, it saved the efforts in establishing a new organizational body to be in charge of intermediation in the stock market. Second, It made it easier for individuals investors to approach the banking branch nearest to them and place an order. Finally, SAMA could supervise the banking sector easily in terms of intermediary activities since this sector had been under its supervision since 1952. In this stage, share trading became easier than it was in the earlier period. Investors could buy and sell through banking branches provided stock intermediation service without the burden of searching for stock owners to trade with. However, the procedures of finalizing trades were very slow and manually done. When a trade took place, the bank would document the transaction and send the documentation to the publicly traded firm whose shares were traded. The firms then would take the necessary procedures to transfer the ownership of traded shares from the seller to the buyer and reissue a “certificate of shares” as a proof of ownership. Although some trading convenience and a somewhat formal market had been created by banking intermediary role, the trading system still had more to improve. Keeping in permanent contacts with publicly traded firms to transfer the shares 15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ownership to the new owners after each transaction was impractical and ineffective for trading. It delayed the process of trading and at the same time caused diversions and distraction for both the banking sector and the publicly traded companies. This implied the need for a central independent depository unit. Consequently, the banks, as the official intermediaries decided to establish the Saudi Share Registration Company (SSRC) to facilitate fulfillment of their intermediary role by providing central registration facilities for joint stock companies beyond settling and clearing all shares trading transactions that occur in the market. A sophisticated computer system was introduced in 1989 to handle the automated clearing and settlement process. In 1990, a big improvement was made by launching Electronic Securities Information System as stock trading system (ESIS).The ESIS is an electronic screenbased system using a central host computer in SAMA head office in Riyadh4. The host computer is connected to remote terminals located in 500 branches of all banks spread throughout the country. In this new system, a bank branch receives orders from traders and submits them to its central trading unit, or dealing room. There, the bank’s dealers generate bids and asks by appropriately combining or splitting orders and entering them into the central stock trading system operated by SAMA (Henry Azzam, 1997). ESIS provides several electronic handling services, including maintenance of accurate client information, liberation of bids and asks into the stock market and execution of transactions, delivery of sold shares and money payments through SAMA clearings, and price dissemination to the central trading units of all participating banks and their branches. 4 Riyadh is the capital o f Saudi Arabia. 16 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. “The Saudi stock market’s payment and settlement systems are among the most advanced in the world. For example, share settlement in Saudi Arabia is at (T + 0) and will soon move to real-time settlement (In most advanced economies it ranges between three and seven execution days)” (Henry Azzam 1997). In spite of these huge advances in the trading process, certificates of shares ownership could not be cleared by SSRC on the same day of trading. Consequently, investors could not retrade their shares in the market until they received their shares certificates, generally, the next day. Furthermore, saving, organizing, and keeping track of shares certificates after each transaction, in fact, was another constraint. In October 2001, a new sophisticated trading system “TADAWUL” was launched by SAMA to replace the former system, ESIS. TADAWUL provides trading, clearing, settlement, depository, and registration services for Saudi Arabian Securities. The new system allows for straight-through-processing (STP) where 100% of all equity trades are settled instantaneously (T+0).5 In the new trading system, each investor has an electronic portfolio account against which all traded shares will be cleared and settled. The introduction of TADAWUL has increased transparency by enabling listed firms to submit their own announcements and financial statements via the internet. Upon their verification, they are disseminated to the investing public, via TADAWUL website, banks, and vendors. TADAWUL concentrates all trading of local shares in one single market. It also enables investors to buy and sell shares not only in banks trading rooms but also throughout internet from any place in the world. The execution of trade, settlement, clearing, and payment are real time. 5 This technical information was adopted from Tadawul website (www.tadawul.com.sa.) 17 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TADAWUL eliminated the deficiencies and shortcomings found in the older system, ESIS. It eliminated the need for issuing certificate of shares ownership after each transaction by immediately adding the traded shares to the electronic portfolio account of the buyer and subtracting them from the portfolio account of the seller portfolio account. Additionally, it enables investors to re-trade their shares several times in the same day of purchasing. Thus, the new system, TADAWUL, promotes market efficiency and transparency. 2.4.1 Types o f Trading Orders In the new version of trading system, TADAWUL, there are ten types of orders available that can be placed by investors to achieve their investment goals of trading and to create a reasonable level of flexibility in the market. These orders as they are stated in Tadawul web site are as follows. > Hit order: An order designed to enable investors to sell the total shares they have at the current best price available in the market. > Take order: An order designed to enable investors to buy the total shares offered at the current best price available in the market. > Match order: This type of orders initiates an opposing order to an existing order. > Market order: It is entered to the market without price to trade a share immediately at the current best price available in the market. It becomes a limit order once a price is generated automatically by the defined price protection formula. > Limit order: An order to buy and sell a specified number of shares at a specified price or better. 18 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. > Unpriced order: Similar to a market order in that it is unpriced order. It becomes a limit order once a price is generated by the trading system. The difference between the two is that an unpriced order is without price protection. > Undisclosed-volume order: An order with a portion of shares that is not displayed publicly. An order of this type receives a price based on prevailing market conditions. Undisclosed-volume order prevents the negative influence a large-size order could play on market price. > All-or- none order: In this order, all shares must completely trade; otherwise, the order will not be executed. > Minimum block order: The order trades in the minimum number of shares specified. The remaining shares are rolled in after each execution from the order. > Minimum fill order: The minimum shares must be filled before the order can trade. The TADAWUL Trading System is characterized by a high degree of flexibility as orders are retained in the market for one day or more, a week, or 30 days. If the orders are not executed within the specified time, they will automatically be cancelled. It also enables investors to cancel or modify their orders at any time. 2.4.2 Trading Commission Trading commissions have been subject to a couple of changes imposed by the government regulations during the last fifteen years. In October 6, 2001, a newly lower scheme of commission was implemented to stock trading. This scheme states that: 19 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. > The maximum brokerage commission taken by banks is (0.0015) of the traded value. A lower commission can be obtained through negotiation between investors and their banks. > The minimum commission on trades whose values are lower than or equal SR 10000 is (0.0015). > The commission is applicable to executed trades only, no commission on order submission, change, or cancellation. 2.5 Market Indexes There was no market index for gauging the aggregate market movement up and down when the market was informal. When the market became formal after government involvement in regulating and organizing in 1984, the Index of National Center for Economic and Financial Information (INCEFI) was established to measure the movement of market trading on a daily basis with instant reading. INCEFI is a weighted average index by the number of outstanding shares in each firm in the market. It also includes all listed firms in its computational process. Thus it is called all shares index. After the introduction of TADAWUL, the last version of electronic trading system, INCEFI was changed to Tadawul all share Index (TASI), which is merely an extension of INCEFI. Besides the all share market index, there are several sub-indexes for each market sector independently. Banking, manufacturing, cement, services, telecommunications, and agriculture sectors have their own indexes, which measure market trading activities within the designated sector. They are all share indexes within the listed firms in the sector. In addition, they are weighted average indexes by the number of outstanding shares in firms listed within each sector. 20 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3. THE SPLIT POLICY AND ITS IMPORTANCE On December 14, 1997, the MCSTS, which is a governmental committee consisting of the Minister of finance and national economy, the Minister of commerce, and the Governor of the Saudi Arabian monetary agency, issued a decree in which all publicly traded companies (companies listed in the stock exchange) with share nominal value (par value)6 of SR 100 are required to divide their shares’ nominal value into two shares, i.e. splitting nominal values to SR 50 and consequently doubling the number of outstanding shares, besides dividing market values for each share by 2 to reflect the split. Each listed company is also required by this decision to make the necessary amendments to its charter accordingly. In addition to these requirements, the new initial public offerings (IPO) must be offered in par value of SR 50 per share. The committee has the supervisory power over all listed firms in the stock market, including initiating and imposing regulations and rules necessary to provide the stock market with necessary information transparency. This policy gains its distinctiveness and uniqueness for research purposes from the fact that it was not expected by either the market or firms’ managements. Moreover, such a policy has never been previously issued before to enable the market to evaluate its market consequences and benefits. The declared goals of the MCSTS’ decree were to unify share par values of publicly traded firms, increase the number of traded shares (volume), expand shareholders base by making trading affordable for investors with low income in terms of 6 Nominal value is the accounting value o f share. It is used to determine the book value o f firm, and it is also used when firms distribute stock dividend. 21 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. number and value of shares, and promote the depth and efficiency of stock market by increasing the number of shareholders. In compliance with this decision, the listed companies started to take scheduled steps to respond and execute the imposed policy and adjust their accounting records to reflect the newly imposed nominal values. Market values were adjusted also. The process of adjustment by each listed firm started after the end of the fiscal year 1997 by inviting the general assembly of shareholders to amend the charter. The announcement of adoption of this policy or decree may be considered a significant event and convey information content to the market. This provides an opportunity to contribute to empirical literature on the adjustment of stock prices to new information, share splits. The importance of the policy (signal) comes from the fact that it is distinguishable in several features from previously similar researched signals used in studying the information content of stock splits. The first feature, which gives the signal of stock split policy its distinctiveness is that all previous stock split studies documented in empirical literature investigate stock split announcements initiated internally by firms’ management, which could announce such signals more than once a year. Signals of this nature help market participants predict such announcements and act upon them. Consequently, the response of the market to such signals is relatively reduced. The signal of stock split policy announcement, however, reached the market surprisingly by an external party, MCSTS. If this signal has information content, the response of the market to the announcement is expected to be higher than a response to a traditional stock split announcement, and the information content will be impacted literally in stock prices. 22 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The second feature is that the decree requires all firms listed in the stock market with a par value of SR 100 to adjust their par values from SR 100 to SR 50 and thus imposes a 2-for-l split of their shares. The firms and the entire market received the announcement (signal) in the same way and at the same time. Most listed firms held par values o f SR 100; hence the decree was applicable to almost the whole market. The investigation of this particular announcement adds more reliable empirical evidence to the share split literature than the previous research where split signals were collected from different firms across time in different contextual, economic, and political settings. The third feature is that the decree was not expected by market participants. This induced a high degree of uncertainty and volatility into the market, because this signal had never been implemented by any governmental agency before, and the laws do not permit firm managements to make a market policy degree. Thus, the decree is really novel and the expected leak of information to the market prior to the announcement is very low. Therefore, the response of the market to the signal is expected to be higher than similar frequently repeated signals. These critical features make the signal of stock split (decree) unique compared to the other stock split announcements previously investigated in literature. Furthermore, the stock split decree announcement with these unique features provides more reliable empirical evidence than management split announcements. Hence, competing hypotheses of a signaling effect can be eliminated. 4. LITERATURE REVIEW AND THEORY The stock split phenomenon has captured many researchers’ interests in the last three decades although it does not theoretically include real economic effects. The 23 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. phenomenon has also triggered streams of research that cover all aspects of the phenomenon. This section provides a review and a synthesis of the extant literature. This provides a basis for understanding the essence of this market effect phenomenon. 4.1 Information Content o f Stock Splits It was believed that stock splits were just cosmetic accounting changes that reduce par value and increase shares outstanding with no resultant change in stockholders’ claims in firm assets because they did not involve any additions to firms’ cash flows or change in owners’ equity. Basically, the same cake is just cut into smaller slices. However, empirical studies have shown contradictory evidence to this belief by demonstrating that stock splits are not merely cosmetic, but also convey information content and economic value. Statistically significant abnormal returns associated with stock splits announcements support this contention. In a seminal study, Fama et al. (1969) document that stock splits are not merely cosmetic accounting adjustments to accounting records but are associated with positive abnormal returns before, but not after the stock split, for their sample of 940 splits between 1927 and 1959. Forjan and McCorry (1998) also reach the conclusion that stock split announcements are associated with higher share prices by using a large sample of firms splitting their common shares. Similarly, Ikenberry et al. (1996) document significant abnormal returns of 3.38 percent following split announcements. Specifically, post-split excess (abnormal) returns of 7.93 percent in the first year and 12.15 percent in the first three years for a sample of 1275 two-for-one stock splits were documented. This result indicates market underreaction to split announcements, as the market does not absorb the whole event and translate it into an announcement return. 24 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. In terms of the studies that investigate the reaction of closed end funds to stock distributions announcements, Datar and Dubofsky (1999) find no statistical difference between the announcement day abnormal returns for closed end fund stock distributions and those of stock distributions made by “ordinary” firms. More importantly, in another unique split study of American Depositary Receipts (ADR) that are not associated with splits in their home-country stock, significant positive abnormal returns between 1% and 2% are found around ADR solo-split announcements. Similarly, in a study concerned with the reaction of the South African stock market to split announcements and capitalization, abnormal returns following splits or capitalization issues are also documented (Biger and Page 1992). Furthermore, stock splits also are associated with a positive and significant stock market response in the Hong Kong stock market (Wu and Chan 1997). Thus, it is clearly apparent from these well-documented studies that stock splits are no longer cosmetic accounting changes, but are actually events with economic value that deserve more in-depth understanding and explanation. In the following subsections, the incentives (motives) behind stock split decisions and the determinants of the magnitude of market response to split announcement are discussed. 4.1.1 Stock Split Decision Incentives The financial literature offers some explanations for the occurrence of stock split market effect and the incentives that drive this phenomenon. There are several available explanations of the real economic effects of stock splits in the literature, including the signaling hypothesis, trading range hypothesis, and liquidity hypothesis. These 25 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. explanations rely on financial market data to support each explanation. In addition, there are other explanations extracted from managers’ stock splits survey research. 4.1.1. a Signaling Hypothesis The signaling hypothesis suggests that managers possess better information about the prospects of the firms than do any outsiders. Therefore, they use their discretionary financial decisions, such as stock splits, as a tool to convey private positive and optimistic information to capital market participants. The signaling hypothesis thus requires implicitly that a cost or penalty be associated with a false signal to be valid; otherwise, it is hard to distinguish between undervalued and overvalued stocks. Szewczyk and Tsetsekos (1993) find support for this hypothesis when they find an inverse relationship between managerial ownership and stock split announcements abnormal returns by examining 175 stock splits of 5 for 4 and greater that occurred between 1972 and 1986 with no other firm-specific announcements either on the day before or on the day of the split announcement. Since stock splits have less information for firms with higher managerial ownership, these results provide evidence for the information effects, i.e. signaling hypothesis. Likewise, Szewczyk and Tsetsekos (1992) also provide support for the signaling hypothesis when they find an inverse relationship between the share price reaction to stock splits and the degree of institutional ownership. The abnormal returns for firms with low institutional ownership are almost double that of high degree institutional ownership. The information asymmetry is evidently greater for lower institutional ownership firms, and this discrepancy is due to the efficient information acquisition activities of higher institutional ownership. Hence, stock splits 26 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. convey more favorable private information to investors when the percentage of institutional ownership is lower. Another version of the signaling hypothesis, an attention-getting model, is developed where managers with favorable inside information attract the attention of security analysts by announcing stock splits (Brennan and Hughes 1991). The lower post­ split stock prices, according to this version, increase trading volume and trading commissions for brokers giving them greater incentive to analyze and promote the stocks of these firms. The support for this signaling hypothesis finds a negative relation to a firm’s share price when security analysts make earnings forecasts. Therefore, the positive abnormal returns associated with stock split announcements in this model are due to market expectations of favorable information by managers. Similarly, Brennan and Copeland (1988) develop a signaling model in which the splits serve as a costly signal of the manager’s private information, because of the administrative costs of issuing the split and the increased transaction costs for investors via share price and odd lots. Empirical evidence indicates that the number of shares outstanding after the split and the target share price provide useful signals of private information. The signaling hypothesis is also supported when the results of comparing earnings and dividend growth rates for a test group of splitting firms and a control group o f non-splitting firms show that investors may be concerned before the split that an earning reversal could follow the abnormally high earnings performance, which had occurred in the prior four years. Managers of these firms may issue the split to signal to investors the stability of earnings. 27 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4.1.1. b Trading Range Hypothesis This hypothesis is also called the “optimal” or “preferred” range hypothesis. Accordingly, managers utilize stock splits to reduce their firms’ share price to a preferred range. The underlying assumption under which optimal range hypothesis works involves the individual stockholders preference to buy round lots. The stockholder cannot afford to do so when the stock price is high. Moving the price into the preferred range, therefore, makes the market for trading in the stock wider or deeper by attracting more investors. An increase in investors should increase liquidity. Growing empirical evidence shows that managers utilize stock splits to restore share prices to optimal trading range. Lakonishok and Lev (1987) suggest that managers issue stock splits to return their stock prices to a normal or acceptable level when they compare and find that closing monthly stock prices for the splitting and non-splitting group are similar four to five years before the split announcement, and in the remaining period leading up to the split announcement, the gap between the two groups widens monotonically until their mean closing prices reach $54.12 and $32.37, respectively, at the time of the split announcement. Four months after the split announcement, the mean stock prices are almost similar and remain so for the remainder of the five-year post­ announcement period. The split factor is also used by managers to adjust their stock prices toward a market-wide price average and to an industry-wide price. In this direction, split factors are found to be an increasing function of pre-split share prices and decreasing function of market value of the firms (McNichols and Dravid 1990). Therefore, managers, when making the decision to split, have in mind some preferred trading range. These ranges can be predicted, using pre-split prices, market 28 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. value o f a firm’s equity to approximate split factor, which in turn determines to some degree the preferred trading range. Attention-getting can be included as part of optimal price range hypothesis (McNichols and Dravid 1990). Lower post-split price may increase the demand for a firm’s stock. In turn, lower post-split price will generate higher commissions for stockbrokers and give them an incentive to sell the firm’s stock more aggressively. Yet, institutional investors prefer high share prices to reduce transactions cost for a fixed dollar trade amount. Therefore, the optimal price range results from conflicting demand preferences from these three parties (splitting firms, brokers, and institutions). In the context of this hypothesis, the mean number of institutions owning shares and the mean percentage of institutional ownership for the test group of splitting firms increased 20.7% and 13.3%, respectively, versus only 8.1% and 5.1%, respectively, for the control group (Powell and Baker 1993). The test group and control group are not statistically different six months before the split. Yet, six months post split, the difference in the mean percentage institutional ownership between test group and control group is found to be statistically significant. The increase in these two institutional ownership variables is a function of firm size. That is, when the firm size is large, the increase in institutional ownership is low and insignificant, whereas the increase in institutional ownership is high and significant when the firm size is small. After controlling for contaminating announcements, Lamoureux and Poon (1987) find the mean number of shareholders for firms announcing stock splits increased 34.65% in the year of the split, while that of the non-splitting control group firms increased only 29 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.11%. This implies that the stock splits play a role in changing ownership composition (mix), and small firms may issue stock splits to attract attention from financial analysts. 4 .1.1 .C Liquidity Hypothesis The liquidity hypothesis states that stock distributions (stock split and stock dividend) enhance liquidity by increasing the proportion of shares traded and decreasing bid-ask spreads. Distribution of stock first brings the attention of investors to split shares and makes these shares more attractive to them in terms of prices. Thereby, the number of trades and the number of shareholders will increase and the increases in these two variables serve to increase a stock’s liquidity after the distribution. Empirical support for liquidity hypothesis is mixed. Copeland (1979), Conroy, Harris, and Benet (1990), and Lamoureux and Poon (1987) find that stock splits may reduce shareholder liquidity from different perspectives. The first two studies report increases in the bid-ask spread after a split, whereas the third study reports a decrease in the mean market-adjusted, splitadjusted volume after a split. Murray (1985) refutes the results of the first two studies by documenting that stock splits do not cause an increase in short-term trading activity. Furthermore, he cannot conclude that stock splits cause a statistically significant decrease in trading volume, and he finds no evidence of change in the percentage bid-ask spread for the splitting firms compared to a control sample of non-splitting firms. Lakonishok and Lev (1987), however, conclude that stock splits cause an increase in trading volume as measured by the monthly number of shares traded divided by the number of shares outstanding in the period around a split when they find statistical differences in the mean monthly trading volume between a sample of splitting firms and a control sample of non­ splitting firms. Month zero is signified with the greatest difference in monthly trading 30 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. volume with the mean monthly turnover of 5.4% for the splitting firms and 4.1% for non­ splitting firms. Ex-date price increases of splitting firms are documented for NASDAQ firms (Maloney and Mulherin 1992). These price increases are linked to changes in liquidity. An increase in the number of shareholders, higher dollar volume, and more trades following share splits are reported, and a direct relationship between increases in the institutional ownership following splits and the asymmetric behavior of the bid-ask quotes around the ex-date is reported also. 4.1.1.d Survey Research on Stock Splits The objective of survey research on stock splits aims at finding the reasons that drive managers to occasionally split their firms’ stock. Baker and Gallagher (1980) surveyed the chief financial officers of 100 NYSE-listed firms that issued a split of 1.25 for 1 and greater during 1978. 98.4% of the responding managers believe that stock split make it easier for small investors to buy round lots, 93.7% of the managers believe that stock splits keep a firm’s stock price in an optimal range, and 85.7% believe that stock splits increase the number of stockholders in a firm. In contrast, the responses from a control sample of non-splitting firms are only slightly less supportive of these three incentives. Only 93.8% believe that stock splits make it easier for small investors to buy round lots, 78.1% of the managers think that stock splits keep a firm’s stock in an optimal range, and 76.6% think that stock splits promote the number of stockholders in a firm. In a similar survey study (Baker and Powell 1993), 251 managers of firms that conducted splits between 1987 and 1990 were surveyed for the purpose of determining motives behind their decisions of splitting stocks. The main motive of the 136 responding 31 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. managers is to move the stock price into a better trading range. Improving trading liquidity comes next. Other important motives include signaling positive managerial expectation about the future and attracting investors. The preferred price range by the surveyed managers was from $20 to $35. 4.1.2 Factors Determining the Market Response to Split Announcements As previously stated, there is a great deal of empirical evidence that stock markets react positively to share splits. The magnitude of this reaction to the announcements varies and is influenced by several determinants as revealed in the literature. Brennan and Copeland (1988) develop and test a signaling model to explain the determinants of the abnormal returns associated with a stock split announcement. The log of the abnormal return over 2-day announcement period is estimated as a function of the log of target share price (pre-announcement share price/split factor), the log of firm value prior to split announcement, the log of contemporaneous market return, and the log of pre-announcement share price. The results provide strong empirical support for the main prediction of the signaling model, that the number of shares outstanding after the split provides a useful signal to investors about the private information of management. The signaling model, in conjunction with the contemporaneous market return, explains about 27% of the announcement date returns. This signaling model is extended further (Szewczyk and Tsetsekos 1993) who include managerial ownership and pre-announcement abnormal returns measured over day -60 to day -2 to capture any information leakage before the announcement. The results reveal that managerial ownership is inversely related to abnormal return, and pre­ announcement abnormal return is not significant. Consistent with the finding of Brennan 32 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and Copeland (1988), target price is found negative and significant in this model, but firm size is statistically significant, inconsistent with that of Brennan and Copeland (1988). These findings provide additional support of the information effects hypothesis. In other studies, however, the level of insider ownership is found to be positively related to cumulative abnormal return of split announcement (Han and Suk 1995). It is important to view the relationship within the context of information asymmetry. That is, the effect of insider ownership should become greater as the information asymmetry between insiders and outsiders increases. The market in general reacts more positively to split announcements from firms with higher insider ownership than from firms with lower insider ownership because a stock split signal by a firm with higher insider ownership is viewed as more credible and should receive a more positive market response. The discrepancy between these two studies in the direction of the relationship may be attributed to the first study using abnormal returns, while the second study uses cumulative abnormal returns as dependent variables. Split factor and firm size in the latter study are also positively related to cumulative abnormal returns. Ikenberry et al. (1996) improve the model predicting the determinants of abnormal returns upon split announcement based on Lakonishok et al. (1994) and Haugen’s (1995) suggestion of using book-to-market ratio as a measure of undervaluation, with high book-to market firms being more likely to be undervalued. It is believed that if a stock split is a signal of undervaluation, and if book-to-market ratio is a good proxy for the degree of undervaluation, the magnitude of the split announcement reaction should be positively correlated with book-to-market ratio. Nevertheless, contrary to what they speculate about the relationship between stock split effects and the degree of 33 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. undervaluation, the empirical evidence shows the opposite, negative relationship. The other variables in the model are consistent with the prior research. This result suggests that, at least in the short run, the market interpretation of a split announcement is more positive for glamour stocks. The fact that the post-announcement returns for the first and the third year exceed that of split event returns indicates the market underreacts to split announcements. In their attempt to remove the effect of an inaccurate classification rule between stock splits and stock dividends in estimation of abnormal return around distribution announcement, Rankine and Stice (1997) improve the previously estimated model by suggesting a new dummy variable (account) to uncover the incremental impact of accounting choice. Account takes on one if a distribution accounted for as a stock dividend and zero if a distribution accounted for as a stock split. The results indicate a positive relationship between accounting choice and abnormal returns. More specifically, the accounting treatment impacts the announcement period reaction; that is, distributions accounted for as stock dividends are associated with announcement returns of 2.7 %, whereas distributions accounted for as stock splits are associated with significantly lower announcement return of 0.93%. Other research demonstrates that split behavior is positively associated with institutional ownership (Mason and Shelor 1998). The association indicates that a higher level of institutional ownership increases stock split propensity. It implies that institutional owners encourage stock split behavior, or they prefer firms with stock split characteristics, or both. Consequently, stock split announcement abnormal returns must be influenced very little by the level of pre-split institutional ownership because the 34 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. surprise by the market is very low. In other words, the abnormal return is expected to be inversely related to the level of institutional ownership. In a more recent study, split factor and split factor surprise are also added to the model predicting the magnitude of abnormal return by Conroy and Harris (1999). Empirical evidence shows that these two determinants are positively related to abnormal returns of split events. Information richness as it is proxied for by the number of financial analysts making annual estimates of firm earnings also plays a role in determining the market anticipation of split announcements (Arbel and Swanson 1993). Further, the magnitude of the market response to stock split announcement is inversely related to the stocks’ information richness. The research shows that the post-announcement market price adjustment process is rapid and complete for information-rich stocks and slow and incomplete for information-poor stocks. Therefore, as information asymmetry between management and investors gets bigger, the market response to split announcements is expected to be stronger. In conclusion, based on a review of the literature, the variables determining the extent of information effect of stock splits can be classified in the following categories: 1- Stock split firm characteristics: This category encompasses firm size, stock price prior to split announcements, pre-announcement abnormal returns (pre-disclosure returns) to capture any information leakage before the announcement, managerial ownership, book-to-market ratio, and institutional ownership. 2- Stock split characteristics: 35 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. This category encompasses target share price (pre-announcement share price/split factor), split factor, split factor surprise, and accounting treatment choice. 3- Surrounding information environment characteristics: This category includes contemporaneous market return and information richness. 4.2 Stock Changes around Stock Split Regardless of the information content stock splits may have and the motives that drive managers to initiate stock splits, the literature documents changes subsequent to stock splits. The changes involve ownership structure, volatility, number of trades and turnover, and volume. 4.2.1 Ownership Structure Lamoureux and Poon (1987) and Maloney and Mulherin (1992) document an increase in the number of stockholders after stock splits. The number of both individuals and institutional shareholders increases, without change in the proportion of equity held by institutions, and this change is positively related to the split factor (Mukherji et al. 1997). This fact is consistent with prior research that suggests broadening the shareholder base may be a possible motive for stock splits by managers to decrease takeover threats (Lakonishok and Lev 1987). It is also consistent with the Baker and Gallagher (1980) suggestion that the primary motive of managers is to reduce the stock price so that small investors can afford round lots. On the other hand, Gary and Kent (1993) conclude that stock splits broaden a firm’s ownership base by increasing the number of shareholders. They show that the number of shareholders remains the same around the split, but the number and percentage o f shares owned by institutions increase. 36 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Integrating these results suggests that the ownership structure of a firm is affected by stock splits, and consequently the number of shareholders increases following a split. Therefore, a stock split’s impact on ownership structure has implications for organizational control, monitoring, and agency costs. 4.2.2 Volatility Changes One of the changes caused by stock splits is the increase in post-split return volatility of the stock split firms. Dubofsky (1991), for example, finds empirically that the variance of daily returns increases following stock splits for both NYSE and AMEX stocks. Yet, the increase in post-split daily variance is greater for NYSE stocks even after controlling for price and split size. When return is measured on a weekly interval, NYSE stocks still exhibit a lower post-split increase in volatility while there is no increase in the volatility of weekly returns subsequent to AMEX stock split. The remaining increase in volatility of NYSE stock returns can be attributed to the differences in structural features of the two exchanges. Likewise, volatility increases significantly after the split from both informed and noise traders (Desai et al. 1998). 4.2.3 Number o f Trades, Turnover, and Volume One of the stock trading activities that stock splits have an impact on is the number of trades and turnover. A significant increase in the number of trades and a significant decrease in the average volume turnover per trade after the splits are documented by Desai et al. (1998). 37 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5. THE FIRST STUDY 5.1 Hypotheses Development Conventional financial theory suggests that stock splits are only “cosmetic” accounting changes to accounting records leading to increasing the number of outstanding shares without adding any benefits to shareholders or to equity value. The theory suggests that this process is simply slicing the same cake into smaller pieces. However, it has been well documented that this theory is unsound as stock prices change positively around stock split announcements and execution dates (Brennan and Copeland 1988; Lakonishock and Lev 1987; Fama et al. 1969; Forjan and McCorry 1998; Biger and Page 1992; and Wu and Chan 1997). The consistency in market reaction to stock splits becomes clearly evident and well documented in the literature, leading to a viable theory on information content of stock splits. Managers issue stock splits to signal private good information to investors and market participants (Brennan and Copeland 1988) or to maintain optimal share price trading ranges. Either way, the split signals help investors reconsider their investments in light of the split signals they receive. The theory of “signaling”, restoring stock prices to “optimal” levels, or “liquidity” is considered sensible when a split is initiated by management because managers know more about the value and conditions of their firm than investors , thereby, the stock splits initiated by the management are the only split signals that should receive the response of the market. However, this notion is no longer valid because empirical research provides evidence against it. More specifically, a unique study of the effect of stock splits on stock 38 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. prices for American Depository Receipts (ADRs),7 which are not associated with splits in their home country, demonstrates significant positive abnormal returns between 1% and 2% around ADR “solo-split” announcements (Muscarella and Vetsuypens 1996). Similarly, an investigation by Datar and Dubofsky (1999) of stock splits of closed end funds, for which managers possess no inside information about their firms’ values,8 documents no statistical difference between the announcement day abnormal returns for closed end fund stock distributions and those of stock distributions made by “ordinary” firms. This paper analyzes the stock split return behavior of the Saudi Arabian governmental policy adopted on December 14, 1997. Previous discussion leads to the following hypotheses: HI: The abnormal return induced by the announcement of stock split policy is significantly positive. H2: Abnormal returns associated with stock splits are positively related to market price level of shares. H3: Market sectors differ in the magnitude of their responses to the split announcement due to the difference in market price levels of their shares. 5.2 Research Methodology 5.2.1 Study Sample Selection Since the MCSTS decree targets all publicly traded firms with nominal value (par value) of SR 100, and all listed firms in the market have a par value of SR 100 except 6 7 ADRs are bank-issued negotiable certificates that give U.S. investors ownership rights in a foreign company. They are issued by the depository bank that holds the underlying securities in their country o f origin. 8 Every week the values o f closed end funds assets are publicly disseminated in the financial press, and prices and bid-ask quotes are reported continuously during the trading day. 39 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. firms9 with par value of SR 50, the initial study sample consisted of 63 firms out of the 69 total firms listed in the Saudi Arabian capital market. The sample is comprehensive and represents all market sectors: banking, manufacturing, electricity, services, cement, and agricultural sector. After a careful investigation, it was found that 9 firms from the remaining sample were without enough weekly trading observations to permit statistically reasonable prediction of expected stock returns.10 Consequently, these firms were removed entirely from the sample.11 Additionally, two firms had no trading observations in the event window. These two firms12 were also excluded from the sample. Thus, the final study sample consists of 52 firms, which represent 75.4 % of the publicly traded firms in the stock market. 5.2.2 Data Collection The data were collected from several sources. Weekly stock prices for all firms traded in the market and weekly Index of National Center for Financial and Economic Information of stock prices from 1985-2001 were collected from the database of Saudi Monetary Agency (SAMA), the governmental body in charge of supervising daily activities of the Saudi stock market. Financial statements of sampled firms were gathered from board of directors’ annual reports of these firms and from the annual directory published by the Saudi Share Registration Company, the company that handles clearing and settling transactions of daily trading. Both sources were gathered in hard copy form. Other data items were extracted from the hard-copy reports issued by the Board of Saudi Chambers of Commerce about joint stock companies. These reports provide sorted 9 These firms include: 3010,4090,4070, 2160, 2150, and 2060. They already have par value o f SR 50. 10 Any firm has trading observations less than 30 weekly observations. 11 These firms have the following ID: 1130,2030, 2090,2170, 4060,4140,4160, 5050, and 5060. 12These firms hold the following ID: 1070, 1100 40 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. information about the performance measures of joint stock companies, general information about the companies and their activities, ownership structure, board of directors, and some brief financial information. The remaining data items were obtained from SAMA annual reports about local stock trading activities that contain brief summaries of market performance and historical news about traded firms in the market, such as important announcements, mergers, acquisitions, raising capital, etc. 5.2.3 Research Design To test the proposed hypotheses, standard event study methodology is used to calculate abnormal returns. A stock’s abnormal return is defined as the deviation of its actual return from a contemporaneous expected return generated by the market model. The market model is used to estimate abnormal returns (Szewczyk and Tsetsekos 1993). To permit for separation of the estimation window from the event window and remove contamination from estimation process, the model parameters are estimated over 47-week estimation period beginning 49 trading weeks prior to week 0, the week of announcement. The event window (-1,0, and +1) is defined as three weeks —the week of announcement, the week before, and the week after—as the process of announcement may have a delayed/pre-emptive effect. For the weeks of the event window, abnormal returns are estimated using the market model. The expected return for a stock is measured by the following market model: ERJt= a J+(3JMT1 (1) Where: ERjt is expected return for stock j on week t. 41 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Ml] is the return on the weekly Index of National Center for Financial and Economic Information of stock prices on week t. a j and J3 j are intercept and slope estimators from the market model estimated during estimation period for the stock j. Abnormal return for each stock in tim e, is then calculated using the following equation: ARjt - ACTjt - ( a j + MTt) (2) Where: ARjt is abnormal return for stock j on week t. ACTp is the actual return for stock j on week t. When abnormal returns are computed for the announcement week and the surrounding two weeks (event window) for each sample observation, the abnormal return (AR) is taken for the sample on that week, followed by summing the abnormal return across event window to calculate the cumulative abnormal return (CAR). Both AR and CAR are computed as follows: a r, = £ a r j , ( 3) 1 CAR , = X AR (4) „ t= - l Where: AR, is abnormal return on week t CAR j is cumulative abnormal return for stock j from event window week zero to week three. 42 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5.3 Analysis and Results 5.3.1 Sample Characteristics After the sample refinement process has been executed, 52 publicly traded firms comprise the study sample. Panel A in Table 1 contains the number and percentage of firms in each market sector. It also provides some insights into the nature of the sample and the distribution of firms across all market sectors. As Panel A in Table 1 depicts, a total of 52 firms, which represents 75.4% of all publicly traded firms in the stock market, comprises the firms selected for the study. The selected firms from each market sector represent from 65% (manufacture sector) to 100% (agriculture sector) of their respective sectors. Thereby, I can be reasonably assured that each sector is well represented in the study sample by the selected firms. Since the sampled firms form 75% of the entire stock market, the sample is considered representative for the population. TABLE 1 Sample Description Panel A: Sample determination process and firms included in the study per each sector Sector (industry) Description la 2b 3C 4d 5e 6f Total All listed firms 12 17 8 17 6 9 69 Firms with Par value of SR 50 - (3) (1) (2) - - (6) Firms with trading observation [...]... date The study then investigates whether the policy appears to have information content for market participants The second study models and relates the anticipated information content found in the announcement of the split policy to the factors mentioned in prior literature to be responsible for driving split information content and determines the magnitude and the direction of relationships among them... governor of the Saudi Monetary Agency The purpose of this research is to study the impact of this new public policy on the market I will approach this issue from three different perspectives Each perspective in itself comprises an independent research study The first study investigates the market response to the announcement of the new policy and the magnitude of abnormal returns surrounding the announcement... billions, 43.4% of the market earnings The manufacturing sector SR 3.671 billions, 19.3% of the market earnings The cement sector SR 2.171 billions, 11.4% of the market earnings The service sector SR 276.83 millions, 1.4 % of the market earnings The electricity sector SR 1.077 billion, 5.6% of the market earnings The telecommunications sector SR 3.545 billions, 18.66% of the market earnings The agricultural... company The certificate was issued by the joint stock companies When a trade was agreed upon by the buyer and the seller, they or their representatives went to the company in which the shares were traded to transfer the ownership of shares from the seller to the buyer A new certificate of shares would then be issued for the buyer including the number of shares traded Also another certificate of shares... literature on the adjustment of stock prices to new information, share splits The importance of the policy (signal) comes from the fact that it is distinguishable in several features from previously similar researched signals used in studying the information content of stock splits The first feature, which gives the signal of stock split policy its distinctiveness is that all previous stock split studies... before, and the laws do not permit firm managements to make a market policy degree Thus, the decree is really novel and the expected leak of information to the market prior to the announcement is very low Therefore, the response of the market to the signal is expected to be higher than similar frequently repeated signals These critical features make the signal of stock split (decree) unique compared to the. .. million shares, 35.5% of the total number of shares in the market The least sector in number of shares was the agricultural sector with 36 million shares, 1.5 % of the total number of shares in the market During the fiscal year of 2003, the total earnings attained for all listed firms in the market added up to SR 18,991.85 millions, the portion of each market sector was as follows: The banking sector... millions, a mixture of some gains and some losses The banking industry is, therefore, the most profitable sector in the market, followed by the manufacturing industry and the telecommunications industry Shareholders’ equity at the end of 2003 for the entire market reached SR 182.8 billions The distribution of this amount across market sectors makes the manufacturing sector at the top of all sectors with... decisions and the determinants of the magnitude of market response to split announcement are discussed 4.1.1 Stock Split Decision Incentives The financial literature offers some explanations for the occurrence of stock split market effect and the incentives that drive this phenomenon There are several available explanations of the real economic effects of stock splits in the literature, including the signaling... distribute stock dividend 21 Reproduced with permission of the copyright owner Further reproduction prohibited without permission number and value of shares, and promote the depth and efficiency of stock market by increasing the number of shareholders In compliance with this decision, the listed companies started to take scheduled steps to respond and execute the imposed policy and adjust their accounting