The Effect of the Revocation of the Short Sell Ban on the Performance of Stocks in Indonesia

The Effect of the Revocation of the Short Sell Ban on the Performance of Stocks in Indonesia Ardhitama Shaumarli ardhitama.shaumarli@ui.ac.id Faculty of Economics, Universitas Indonesia Abstract This research discusses mainly about the effect of the short sell ban on the performance of stocks. The performance is measured through the liquidity, overvalue, and price efficiency of the stocks. This research is conducted after the revocation of the short sell ban after the preceding crisis of 2008. The result of this research is the enactment of short sell will increase the liquidity of stocks. Besides that, it’s also found that stocks undergo overvaluing when the ban was still enacted. Lastly, the enactment of short sell will increase the efficiency of stocks. Keywords: short sell; liquidity; overvalue; efficiency; crisis 1. Introduction The European and American financial crisis of 2007-2009 caused the world economy to generally weaken, even worse when the Lehman Brothers collapsed on 15 th of September, 2008. This was led to the instability of the world market. Short Sell was assumed to be able to over-fluctuate prices, so the SEC announced the short sell ban on September 19 th, 2008, to avoid further failing of the exchange. The chairman of SEC, Christopher Cox, stated that the temporary ban of short sell on financial stocks will return the balance of the market. This triggered the short sell trading ban on many countries, starting from the US, which was then followed by other countries. Indonesia was included in the list of countries that enacted the ban on the crisis period (Suhartono, 2008). On the first two weeks of September, 2009, the IDX Composite dropped 400 points from level 2.145,62 on September 1st, to 1.719,25 on September 15th. Even though the people who exploited short sell on the crisis period was not the main factor of the drop, but they were the main suspect. In order to avoid further drop on the IDX Composite, Indonesia Stock Exchange blocked the short sell facility on October 6th 2008 until April 30th 2009. Like many other European countries, after the global crisis, Indonesian exchange also strengthen the regulations regarding short sell by revising the criteria of stocks that may be enlisted in the short sell list (Pekarek & Meseha, 2012). Along with the gradual stabilization of the global market, on May 1 st 2009 Indonesian Stock Exchange revoked the ban of short sell trading, expecting that the reopening of the short sell facilities, the marked will become even more liquid, and raise the IDX Composite. They expected by revoking Theban would improve IDX Composite. This research specifically tests the existing theory of the performance of stocks after the revocation of short sell ban. Specifically, the stock performance will be limited on the aspects of liquidity and price efficiency. Furthermore, there is also the assumption that the previous short sell ban would cause overvaluing, which is interesting to be examined if it happened in Indonesia. 2. Literature Review In Indonesian stock exchange, short sell is regulated by the Bapepam Regulation V.D. 6 of 2008. Generally, short sell involves six parties, which are (1) securities firms, (2) investors, (3) short seller, (4) broker, (5) buyer, and (6) clearinghouse. The investor would deposit his investment funds to the securities firm, expecting profit. The fund would be then used to buy stocks and the securities assets by other securities firms to gain return. The securities firms would collect stocks that belong to the investor, which would become stocks borrowing market to the short seller. On the other side, the investor can do short sell by opening a short sell account. 2.1 Short Sell Constrain and Liquidity To this day, the researches on the effects of short sell on liquidity yield different results. Short sell, which is identical with the utilization of negative information, was judged to be able to increase the price discovery of stocks and causes bid-ask spread to reduce (Diamond & Verecchia, 1987). With the assumption of the ban applied to investors that either have the information or not to trade with negative information will cause the decrease of speed of price discovery, which will increase bid-ask spread that funds the reduce of liquidity. But if the short sell ban only applies to investors who have access to information, the fraction of investors in selling position would decrease. This makes bid-ask spread to decrease, and liquidity to increase. The actions of the short sellers can help to increase the immediacy of stock transactions (Diether, Lee, & Werner, 2009). When good news about a certain firm is spread on the market, usually there are many investors that would respond excessively by doing order buy. On the other hand, investors that own those stocks would not quickly let go of the stocks, because they would gain bigger profits when the price rises. This triggers order imbalance. Seeing that, if short sellers assess that stocks are overvalued, they would do short actions that would cause match orders to happen. The number of buy order would then trigger the rise of stock prices, but if it is not accompanied by a good foundation, prices would gradually decrease back down. Generally by allowing short selling the number of players in the capital market would increase, which consequently will cause the increase in the number of transactions (Gruenewald, Wagner, & Weber, 2010). The increase of potential sellers in this market could trigger the increase of stock volume, which signifies the increase of efficiency. 2.2 Short Sell Constrain and Stock Overvaluation Theories on the effects of short sell ban towards stock prices yield different results, starting with the argument that the ban on short sell would push stock prices to the state of overpricing (Miller, 1977). This argument is based on the assumption that investors have heterogeneous interests. Pessimistic investors that actually do not possess stocks could not trade in short position, so they would automatically be out of the market. This cause the stock prices would rise in bias, reflecting the valuation of prices from solely bullish investors. Short sellers usually are investors who purchase securities on a higher price range than the fundamental value, which would be resold to optimistic investors (Harrison & Kreps, 1978). When short sells are banned, it would cause stock prices to move above the valuations of the most optimistic investor, according to the expectations of future profit. Short selling itself is needed in order to keep stock prices to reflect their intrinsic values (Norman G., 1993). When a stock is overvalued, the investors would use the opportunity by short selling. This would reduce the price of the overvalued stock to its intrinsic price. If short selling is banned, stock prices would experience overvaluing and could trigger bubbling in the future (Abreau & Brunnermeier, 2003) 2.3 Price Efficiency Price efficiency can be defined as a degree in which the price of a stock can reflect all of the available information, time-wise, or accuracy-wise (Saffi & Sigurdson, 2011). Generally, there are a lot of previous researches that found that the objections of short selling can affect price efficiency. When short selling is banned, pessimistic investors would leave the market, and prices would only be reflected by optimistic investors (Miller, 1977). Pessimistic investors are inclined to respond to negative information while optimistic investors tend to respond to the positive ones, using private information that they receive. When negative information leave the market, the speed of price discovery would slow down (Bai, Chang & Wang, 2006). This shows that the ban on short selling would render prices to be less efficient. According to the heterogeneous agent model, the decrease of stock prices because of short selling ban would trigger future market failures (Hing & Stein, 2003). In the model, it is explained that if investors are banned from short selling, negative information that are possessed by investors would not affect the prices. This information would accumulate until the time when the market declines, the accumulation of negative information would worsen the market’s condition after leaking out. 3. Research Method The criteria of stocks that are used in this research is stocks that had their short sell permit revoked. In other words, stocks that qualified the criteria are stocks that are listed in the Indonesian Stock Exchange’s short sell list for after the 2008 crisis for the first time. The period of the research is from May 2009 to April 2011. Table 1: The changing of short sell list Announcement Date Effective Date New Added Re-added Deleted Total List 30-Apr-09 01-Mei-09 18 0 0 18 29-Mei-09 01-Jun-09 4 0 3 19 30-Jun-09 01-Jul-09 7 0 0 26 31-Jul-09 03-Agu-09 3 2 2 29 31-Agu-09 01-Sep-09 5 0 0 34 30-Sep-09 01-Okt-09 2 1 1 36 30-Okt-09 02-Nov-09 3 0 4 35 30-Nov-09 01-Des-09 0 1 3 33 30-Des-09 04-Jan-10 0 0 5 28 29-Jan-09 01-Feb-10 0 2 1 29 25-Feb-09 01-Mar-10 0 0 0 29 31-Mar-10 01-Apr-10 0 0 0 29 30-Apr-10 03-Mei-10 0 0 0 29 31-Mei-10 01-Jun-10 0 1 1 29 30-Jun-10 01-Jul-10 4 1 0 34 30-Jun-10 02-Agu-10 0 0 0 34 31-Agu-10 01-Sep-10 1 0 2 33 30-Sep-10 01-Okt-10 0 0 1 32 29-Okt-10 01-Nov-10 1 0 0 33 30-Nov-10 01-Des-10 1 1 0 35 30-Des-10 03-Jan-11 1 0 0 36 31-Jan-11 01-Feb-11 5 1 0 42 28-Feb-11 01-Mar-11 2 1 3 42 31-Mar-11 01-Apr-11 2 0 1 43 Total 59 11 27 On the period of the research 59 moments of addition of new stocks was found on the short sell list, 11 were re-added, and 27 were omitted. This research will focus solely on the moments of addition of new stocks of the short sell list because the moments of omission provides only little sample. From the 59 moments of additions, it was found that 26 of the stocks were listed less than 136 exchange days. Because of the short span of time, those stocks were removed from the samples, leaving only 33 additions to be analyzed. 4. Results and Discussion 4.1 Short Sell and Liquidity The measuring of the relationship between short selling and liquidity was done by committing the regression of the data panel. The period that is used in this model is divided into two, which are the period before and after the stocks addition on the short sell list. The former period is (-136,0) exchange days and the latter is the (0,136) exchange days. Model that is used refers to Bohemer, Jones, and Zhang (2009), that is: �� = + �� +� �� + ��� .............................................. (1) Y is the dependent variable that explains liquidity. The proxies that are used are (1) frequency, (2) volume, (3), value, and (4) spread. Frequency is the unit that declares how many times stocks are traded each day. Volume refers to the number of stocks that are treaded in each day, per sheet. Relative quoted bid-ask spread is the difference from the highest price offered by buyers on an asset with the lowest price offered by seller. X is the control variable, which consists of (1) market capitalization, (2) value, (3) range price, and (4) daily-volume weighted average share price (VWAP). Market capitalization is the total price of the firm stock that circulates on a certain period of time. Range indicates the difference between the highest and the lowest price in one day, divided by its proportional price. VWAP is the average price according to the volume, each day. Table 2 The Result of Liquidity Model Frequency Value Volume *** *** Intercept 958 11.234 32.905*** Dummy 68*** 3.215*** 3.806*** Market Cap -0,000*** 1,214*** -0,309*** *** Value 0,003 0,403*** Range 15*** 3.894*** 531*** VWAP 0,033*** -0,549*** -1,472*** R-Squared 61,27% 68,37% 87,52% Table 2 explains the result of regression of data panel using the Spread 1,683*** -0,084*** 0,000*** 0,000*** -0,001*** 0,000*** 82,62% fixed effect model with GLS cross section and liquidity proxy. The *, **, and*** points to the signification on the levels 10%, 5%, and 1%. The independent variable dummy is worth 1 when short sell is allowed, and is worth 0 when it is not. The variables frequency, value, volume, and spread are dependent variables. The variable market cap, value, range, and VWAP are control variables. R-squared uses the conformation price, which shows how well the independent and control variables can explain the dependent variables. From the result of the regression the result was that stocks would become more liquid after short sell is allowed. Proven by the increase in frequency, the increase of value of the volume and value, and decrease in bid-ask spread of the stocks. The increase of frequency signifies the raise of immediacy of the stocks. The raise of volume and value of these stocks show that there are more stocks that are traded in the marked, which shows that investors are able to sell or buy stocks without being influenced by its price quotation. So does the result of the decrease of quoted spread shows that the tighter the difference of the ask and bid price of the stocks, which signs investors are able to sell or buy assets simultaneously with little difference between the price and the expectation price. This increases the chance for match ordering stocks. The strengthening of the liquidity is influenced by the increase of the speed of the price discovery of the stocks after short selling is allowed. Short sellers usually use negative information in trading, which causes an increase in price discovery. When investors can quickly adapt prices according to available information, the shaping of stock prices would be accomplished faster than indicated, according to the decrease of spread. 4.2 Short Sell and Overvalue Effect According to the research of Chang, Cheng, and Yu (2007) and Chang, Luo, and Ren (2013), measuring the hypothesis of overvalue effect is done by utilizing the event study method. Proxy that is used for the measurements is abnormal return (AR) and cumulative abnormal return (CAR). The t0 event date point that is used is the first effective day of the stocks’ short sell trading period. The estimation window period that is used is (-136, -13) days of exchange before event date. Event window that is used is (-20, +20) days. This range event windows is picked because it is considered to be unbiased by the announcement list of short sell on the month before and after. Here is the abnormal return equation using market model � �� = � + �� � ���� � = ��� − � ........................................................ (2) �� ................................................... (3) ���� � , � = ∑��� � ................................................. (4) The t-test was done to find out if the average value of abnormal return and the average cumulative abnormal return that was gotten was significantly different from zero. Table 3 Abnormal Return with moments of addition of new stocks on the short sell list Days Mean Median P-Value -10 0,16% -0,04% (0,73) -9 -0,19% -0,92% (0,79) -8 -0,10% -0,66% (0,85) -7 -0,19% -0,13% (0,71) -6 -0,04% -0,15% (0,95) -5 0,84% -0,06% (0,15) -4 -0,32% -0,41% (0,42) -3 -0,54% -0,63% (0,34) -2 -0,47% -0,39% (0,27) -1 -0,62% -0,89% (0,28) ** 0 -1,14% -1,40% (0,04) 1 -0,75% -0,92% (0,11) 2 0,26% 0,24% (0,62) 3 -0,23% -1,05% (0,69) 4 -0,42% -0,83% (0,53) * 5 -1,24% -1,33% (0,07) 6 0,31% 0,23% (0,51) 7 -0,50% -1,04% (0,29) 8 -0,33% -0,33% (0,39) 9 0,33% 0,81% (0,47) * 10 0,95% 0,30% (0,05) Table 3 shows that on the event date the 0 pint of average abnormal return on the moment of new stocks addition is -1,14% (p-value 0,04). This indicates that on the first day stocks are effectively wallowed to be sold in short sell trade, average return of actual stocks lower than the expectation of its return. From the testing of significance using the t-test method, the result of the pvalue of this 0 point was yielded that it was 0,04, which shows the abnormal return hypothesis has the same value with 0 can be rejected with the level of significance of 5%. On t0 the significant negative abnormal return proved that there is overreaction of short seller on negative information. When stocks are already overvalued, negative information would be react to in an exaggerated fashion by investors. This reaction could be caused by the opportunity for short seller to utilize the stocks condition, whose values are too high to be shortly sold. On the same table, it is also seen that the average of daily abnormal return that are significant negative, but no daily abnormal return before a significant event date. This indicates overreactions towards negative information are emphasized after the ban on short selling was revoked. Table 4 Cumulative Abnormal Return with Addition Moment of new stocks on short sell list Event Mean Median P-Value Windows (-10,-1) -1,45% -1,94% (0,16) ** (-1,+1) -2,50% -2,27% (0,01) ** (0,+2) -1,62% -1,22% (0,02) *** (0,+5) -3,51% -3,78% (0,00) ** (0,+10) -2,75% -4,56% (0,01) ** (0,+20) -4,67% -6,59% (0,04) Table 4 explains the cumulative abnormal return on the moment of addition of new stocks on the shor sell list. The *, **, and *** signs indicates signification on the level 10%, 5%, and 1%. The table shows that the cumulative abnormal return from day 0 to day 5 is -3,51% and significant on the level 1% (p-value=0,00). Cumulative abnormal return was also still significant negatively valued t 0 to t10 (-2,75%, p-value = 0,01) and also t 0 to t20 (-4,67%, p-value = 0,04). The CAR value that was the negative significant shows that accumulatively until 20 days after the revoke of the short sell ban, actual return of stocks is lower than expected. This indicates the return expectation before short sell is too high, triggering overvalue. However, the table also shows also when the average cumulative abnormal return on day t -10 and t-1 is worth -1.45 event date but not significant (p-value 0.16). In addition the average cumulative abnormal return on t -1 to t-1 of the event date is worth -2.50% and significant at the 5% level (p-value 0.01). This means that the return has been a significant abnormal occur up to 10 days before the event date. Graphic 1 Abnormal Return and Cumulative Abnormal Return with moments of addition of new stocks on the short sell list Graph 5.1 is the result of a plot of abnormal returns and cumulative abnormal return in period t-20 to T20 trading days of the event date. From this graph it is seen that a few inventions, namely the value of cumulative abnormal return that fell before the event date and persistent decline. Firstly by using the view stated when the price reversal cumulative abnormal return before the event date will be positive and decreased after the event date. But as seen from the pre-event windows from t-13 return on average continued to decline. This is presumably due to the stock information will be entered into the known short sell up to 13 days in advance. Investors can find stocks that are potentially included in the list of short-selling, the stock check compliance with the criteria set by the Indonesia Stock Exchange. With so before the stock is effectively included in the list of short-selling, investors have to make adjustments to the valuations that led to the return to be down. Second, the effect of negative cumulative abnormal return is supposed to be temporary, because the price will be corrected to its fundamental price alone. However, the effect of the decline in cumulative abnormal return is obtained after a short sell persistent allowed. In other words, cumulative abnormal return will continue to decline on a long period. Return expectations of investors worth more than the actual return, so that the value of the abnormal return becomes negative. This continues until the value of the accumulated abnormal return that is obtained is kept down. Rationalization of cumulative abnormal return down is caused by persistent overvalued stock that occurred in a period of time, so it takes a long time to get back to the fundamentals. Besides analyzing the price efficiency in the subsequent discussion indicates when short selling is allowed, so the influx of pessimistic investors will make the negative information becomes more sensitive to price formation. Moreover, the perpetrators short sell will use negative information specific to a particular company to short sell. As a result of market information becomes less capable of predicting stock prices. As a result, the information of this market became the main base in the formation of the expected return. The bias is the expected return value indicates that expectations before the short sell is done will be different after the short sell is allowed. Therefore, investors need to adjust their expectations to a particular stock if the shares are allowed for in short sell. 4.3 Short Sell and Efficiency Measurement of price efficiency and stability of return is done by comparing the stock return before and after getting into the short sell list. The before period that will be used is (-136,-31) trading days and the after period that will be used is (+31, +136) trading days. Table 5 Changes in Price Efficiency Indicator Before After P-value R-squared 39,21% 33,66% (0,09)* R-squared + 25,49% 15,62% (0,10)* R-squared – 28,20% 22,00% (0,01)*** Beta 1,24 1,13 (0,22) Beta + 1,17 1,12 (0,70) Beta – 1,26 1,13 (0,22) Correlation 9,06% 1,03% (0,02)** Correlation + 8,87% 3,64% (0,30) Correlation – 13,61% -5,14% (0,00)*** The table explains the change indicator of price efficiency before and after the stocks may trade in short sell. The "Before" column explains the result of cross-sectional calculation using preevent windows (-136, -31) days trading with and "After" column use post - event window (+31, +136) days trading after stocks got into short shell list. P-value explains the results from different tests using t -test to determine the level of significance of the difference value between before and after short-selling is allowed. Mark *, **, and *** indicate significance at the level of 10 %, 5 % and 1 %. Winsorized is done on stock returns that have more than three times of value from the deviation standard of the average in order to make the regression results are not biased. By notice of a change in the table, it can be concluded that the stock price will tend to be more efficient after allowed for short sell. This is evidenced by a significant decrease of the Rsquared and correlation after short-selling is allowed. Increased efficiency of price is more affected when the market condition is down, whereas when the market conditions are rising, it will not significantly increasing efficiency. When short selling is allowed, specific information related to the company also affects to establish stock prices. Moreover, the negative information is more often used by the trader short selling when trading. As more information becomes involved in shaping the price, then the price will be more difficult to predict and become more efficient. Table 6 Descriptive Change Statistics Distribution of Return Before Mean 0,15% Stdev 4,67% Stdev + 12,05% Stdev – 9,38% Skewness 0,75 Kurtosis 3,85 Extreme + 2,54% Extreme – 0,74% Max 17,04% Min -13,27% From the return distribution can be seen after After P-value 0,19% (0,70) 2,54% (0,00) *** 5,67% (0,00) *** 4,68% (0,00) *** 0,38 (0,03) ** 1,27 (0,00) *** 0,09% (0,00) *** 0,06% (0,00) *** 8,01% (0,00) *** -6,62% (0,00) *** short sell effectively allowed, the return of the stock tends to approach normal distribution. Seen from the skewness and kurtosis that are getting closer to zero. Fluctuation of return that shown from volatility, volatility +, and volatility - will be lower. Similarly, the value of the extreme + and extreme - are declining. When short sell allowed, pessimistic investors will go into the market and start doing transactions in accordance with their beliefs. This pessimistic investor action affected the overvalued stock price got back to the fair prices, thus increasing the possibility of a negative return. As a result, the obtained skewness results reduced their levels of positive, but not to be negative. This indicates that the level of positive return after a short sell is allowed is not higher than before. The coming of pessimism investor also makes the valuation of stock returns become more balanced. Optimism investors assume that the stock price will rise, but pessimism investors have valuation that the stock price will fall in the future. The combination of both investors’ valuation makes return fluctuations become more stable, which is demonstrated from the reduction volatility of the stock after short sell is allowed. This declining of volatility also associated with the lower frequency of extreme returns, so that the deviation of stock returns are not widen. The fall of volatility and extreme levels shows that the stock return is becoming more stable. By seeing the return distribution, it can be concluded that the individual stock level, the allowing of short sell makes stock return move to the normal distribution. It shows that the stability of returns from the stock after a short sell is allowed. However, the increasing stability of the stock return can’t be interpreted generally that the market becomes more stable. 5. Conclusion This study sought to test the performance of stock in Indonesian Stock Exchange after the prohibition of short sell stock has erased. By assuming stable market conditions, stock performance that especially measured is liquidity, overvalued effect, and price efficiency. From the obtained results, can be shown that if the prohibition of short sell is erased, it would increase the liquidity of the stock. It is proved by the value of the frequency, volume and the increasing of stocks value, and the declining value of the bid-ask spread after short sell stocks has allowed. This increasing of liquidity is due to the coming of pessimism investor in the market, thus increasing investor who trade in the market will increase the chances of stocks trading transactions. Besides that, short sell can stimulate price discovery becomes faster that causes the formation of the price to be faster. This is proved by the value of the bid-ask spreads which are become closer when the short sell is allowed. Subsequent analysis found that the stock will have overvalued when the short sell is prohibited. This is proved by the negative value of the cumulative abnormal returns after prohibition on short selling is erased. This negative abnormal return occurred due to overreaction from the short sell trader about negative information. Cumulative abnormal return decreased after short selling is allowed. But since 13 days before the event date, the stock has experienced negative significant cumulative abnormal return. Investors already know stocks that have fulfilled the criteria are allowed for short sell according to the terms of Indonesian Stock Exchange. When they respond to this information overload, the cumulative abnormal return becomes negative before the point of the event date. In the price efficiency side, after stocks entered short sell, the stocks price would be more efficient. This is measured by significant decrease of R-squared value and correlation of stocks return, especially when the market is down. The performance of the stock after the stock is allowed entered short sell, the distribution of its return closer to normal distribution. Seen from the skewness and kurtosis are moving approach to a zero value. Standard deviation is reduced and the values of the frequency of extreme returns are declining. This indicates that the level of individual short sell can stabilize return from the stocks. 6. References Abreu, D., & Brunnermeier, M. K. (2003). Bubbles and Crashes. Econometrica, 71, 173-204. Bai, Y., Chang, E. C., & Wang, J. (2006). Asset Prices Under Short-Sale Constraints. Working Paper, MIT Sloan School of Management. Retrieved from http://web.mit.edu/wangj/www/pap/BCW_061112.pdf Beber, A., & Pagano, M. (2013). Short Selling Bans around the World: Evidence from the 2007-09 Crisis. Journal of Finance, 68, http://ideas.repec.org/p/sef/csefwp/241.html 343-381. Diambil kembali dari Chang, E. C., Cheng, J. W., & Yu, Y. (2007). Short-sales constraints and price discovery: Evidence from the Hong Kong market. Journal of Finance, 62, 2097–2121. Retrieved from http://dx.doi.org/10.1111/j.1540-6261.2007.01270.x Chang, E. C., Luo, Y., & Ren, J. (2013). Short-selling, margin-trading, and price efficiency: Evidence from the Chinese market. Journal of Banking & Finance. Retrieved from http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2135067 Clifton, M., & Snape, M. (2008). The effect of short-selling restrictions on liquidity: Evidence from the London Stock Exchange. Working Paper. Diamond, D. W., & Verrecchia, R. E. (1987). Constraints on Short-Selling and Asset Price. Journal of Financial Economics, 18, 277-311. Diether, K. B., Lee, K.-H., & Werner, I. M. (2009). It’s SHO time! Short-sale price tests and market quality. Journal of Finance, 64, 37-73. Retrieved from http://onlinelibrary.wiley.com/doi/10.1111/j.1540-6261.2008.01428.x/abstract Fragnière, E., & Markov, I. (2009). Short Selling in France during the Crisis, the Bans, and What Has Changed since the Euro Correction. Handbook of Short Selling. Gruenewald, S., Wagner, A. F., & Weber, R. H. (2010). Short Selling Regulation After the Financial Crisis - First Principles Revisited. Harrison, J. M., & Kreps, D. M. (1978). Speculative Investor Behavior in a Stock Market with Heterogeneus Expectations. The Quarterly Journal of Economics, 92. Hong, H., & Stein, J. (2003). Differences of opinion, short-sales constraints and market crashes. Review of Financial Studies, 16(2), 487-525. Miller, E. M. (1977). Risk, uncertainty and divergence of opinion. Journal of Finance, 32, 1151– 1168. Retrieved from http://www.jstor.org/stable/2326520 Norman G., F. (1993). Stock Market Logic: A Sophisticated Approach to Profits on Wall Street. Dearborn Financial. Pekarek, E., & Meseha, M. (2012). Short Selling—The Ambrosia or Kryptonite of Emerging Markets? Handbook of Short Selling, 365-380. Saffi, P. A., & Sigurdson, K. (2011). Price Efficiency and Short Selling. Review of Financial Study, 24 (3), 821-852. Retrieved from http://papers.ssrn.com/sol3/papers.cfm?abstract_id=949027 Suhartono, H. (2008). Indonesia’s bourse bans short selling for October. Reuters. Retrieved from http://www.reuters.com/article/idINJAK15697520080930