THE ACADEMY OF ECONOMIC STUDIES MASTER DAFI

Transcription

1 THE ACADEMY OF ECONOMIC STUDIES MASTER DAFI Test events on the capital market in Romania. Comparative study with the capital markets in Bulgaria and Austria Author: Ionesi Claudiu Coordinator: Prof. univ. dr. Anamaria CIOBANU Bucharest 2009

2 ABSTRACT This paper will analyze the Romanian investors reaction to events of global significance. The questions asked for an answer are: How can influence an event the capital markets? and Is the Romanian capital market mature enough not to overreact to significance events? In order, to do this, we will choose the BET index, considered to be the most important benchmark of the Romanian capital market, and we will if and how it reacts, when economical or political global events take place. The reaction will be explained using of three hypotheses: the Efficient Market Hypothesis, the Overreaction Hypothesis and the Uncertain Information Hypothesis. Simultaneously, we will analyze the SOFIX index, considered to be the most important benchmark of the Bulgarian capital market, and the ATX index, considered to be the most important benchmark of the Austrian capital market. We have chosen these two countries because we intended comparing analyses with a more developed country than Romania, and a country less developed than Romania, as Bulgaria is considered. 1. Introduction The Romanian capital market In the first years of trading, after 1995, the liquidity had a low level, mostly because in 1995 there were listed only 9 stocks, and in the next year the number of stocks grew to 17. A spectacular growth regarding the number of stocks, the capitalization and liquidity could be observed in 1997, when the number of traded stocks grew to 76. Nowa, on the Bucharest Stock Exchange the number of traded stocks is 70. The fulminating evolution of the Bucharest Stock Exchange regarding the capitalization can be observed from this table. Year No. of traded stocks (volume) Capitalization (RON) No. of companies No. of new companies No. of unlisted companies , , , , , , , , , , , , , ,

3 In order to determine, whether the investors can obtain abnormal returns and how they react when an unexpected event occurs we will take in consideration the three hypotheses: The Efficient Market Hypothesis; The Overreaction Hypothesis The Uncertain Information Hypothesis The Efficient Market Hypothesis The Efficient Market Hypothesis is associated with the idea of random walk. Random walk is a term used to describe the random evolution of stock prices from one day comparing to the precedent day. The explanation is the following: if a piece of news induces an instant stock price in a certain day, the next day stock price will be only influenced by the news in that day. So, the stock prices will reflect totally all the information, and, even the unadvised investors that buy a stock will obtain the same return as the advised investors. The Overreaction Hypothesis The Overreaction Hypothesis is usually observed after an event took place. Concretely, after a positive event the stock prices tend to grow more than the fundamental value and after a negative event the stock prices tend to decrease more than the fundamental value. In the future period, we assume that the underappreciated stocks will have a higher return than the market return and the over appreciated stocks will have a lower return than the market. The Uncertain Information Hypothesis The Uncertain Information Hypothesis is based on the fact that, often, the investors speculate before an important event will take place, without knowing precisely the consequences, but only the time of the event. Practically, the rumors before an event make the risk-averse investors to protect their investments of volatility and they take positions under the fundamental value of the stocks. After, the uncertainty is over, the prices tend to grow no matter the type of event. Usually, the abnormal return for negative events is higher than the ones for the positive events. In 2008, Ciobanu, Mehdian and Perry made a similar analyze. They researched for the Romanian investors reaction in the time periods after economical and political events took place. They took in consideration the BET and BET-C index. They concluded that after an event, the BET index react according to the overreaction hypothesis, meaning that the investors take positions below the fundamental values after negative events and above after positive events. If we take in consideration the BET-C index, after a positive event, they remarked a positive correction of prices, according to the uncertain information hypothesis, meaning that the investors took positions below the market before an important took place. Also, after negative events, the stock prices instant tend to the correct. This evolution is made according to the efficient market hypothesis. 3

4 2. The methodology The database contains the daily returns of the three indexes during the 2008 year for the Romanian, Austrian and Bulgarian stock markets. The analyzed indexes are: BET for Romania, SOFIX for Bulgaria and ATX for Austria. The daily return of an index is calculated as: K it = ln(i it / I it-1 ) x 100, where K it daily return of index i on day t; I it closing value of index i on day t; I it-1 - closing value of index i on day t-1. For the three indices we calculated the mean return and standard deviation recorded in Results are presented in the table below: Index No. of Mean return Standard deviation BET (Romania) 249-0,49% SOFIX (Bulgaria) 248-0,64% ATX (Austria) % All the three indices had negative evolutions during On average, the most depreciated was the Bulgarian SOFIX index, followed by the Romanian BET index and the Austrian ATX index. The indices of the two countries in Eastern Europe were more impaired, which was somewhere expected considering that until 2007 the East-European stock exchanges had consistently better returns. Secondly, the database includes a series of events that had a determining role on the global economic markets in the year These events are mostly financial news, but we have some news that are non-financial, yet had an impact on financial markets. Date Event January 28 Societe Generale Fraud March 2 July 11 August 8 September 7 September 15 September 17 September 28 September 30 October 8 November 4 November 7 November 26 December 12 Dmitri Medvedev - elected president The intention of nationalization of Fannie Mae and Fred Mac if the situation gets worse Georgia-Russia War Nationalization of Fannie & Mae Fred Mac Lehman Bankruptcy AIG Plan - 85 billion Insurance Company Fortis is helped by the Belgian, Dutch and Luxembourg government Dexia Bank nationalized by the French and Belgian authorities 6 central banks lower interest rate Obama president Iceland close to bankruptcy Mumbai hostages Madoff Scandal 4

5 As you can see, the vast majority of news is represented by negative events. This is normal if you keep in mind that 2008 was marked by strong decreases of capital markets. We will use all these events just to calculate the volatility and post-event returns. When we move to calculate the abnormal returns we eliminate those events whose period of 10 after the events triggers overlap. To calculate the normal return of the market we have several methods, including the CAPM method (capital asset price model) or the method of calculating average daily return on the returns obtained daily during the year. In this paper we will use the CAPM method. Summarizing, the relationship between expected return and risk can be written using the following formula: E(R i ) = R f +β i (E(R m ) - R f ), where E(R i ) the normal return of the asset; R f the risk-free rate; E(R m ) the market return; β i beta indicator. In this case, for the three countries will comprise the following variable first to determine risk for each country: - Difference of inflation; - Country premium risk; - U.S. capital market premium risk. In this study we use the following values: Inflation Country premium risk 2 Romania 7.90% 0.00% Bulgaria % 3.90% Austria 2.858% 3.90% We will also consider the risk for the U.S. capital market is 5%. After summing these three variables, they are added to the risk-free rate and we get what might be considered normal return for each of the three markets calculated for To find the daily normal return will divide the annual return to 365 (the number of in a year). If you return from a day exceeds the normal return calculated using the CAPM method then we consider the difference between the two returns as an abnormal return. AR it = K it E(R it ), where AR it the abnormal return of an index on day t; K it the return of an index i on day t; E(R it ) the normal return using the CAPM model for index on day t. 1 Valori preluate de pe site-ul: 2 Valorile pentru ianuarie 2009 preluate de pe site-ul: 5

6 To test whether an event has led to increasing volatility on one of the three capital markets, we calculate the variance for the daily returns for the after the analyzed events, and variance for the ordinary. The variance formula is the following: N j 2 Var = K K ij / N 1 t1 K it the return of an index i on day t; K the mean return for every type of day (post-event day or ordinary day); N j - the number of day for each type of day; j type of day(post-event day or ordinary day). After carrying out these tests of volatility, we will calculate the abnormal returns. In order to calculate the cumulative abnormal returns we will apply a procedure of three steps, implemented by Mehdian, Nas and Perry (2008): 1. The calculation of abnormal return for each day of analysis. 2. The calculation of average abnormal return for each day. 3. The calculation of cumulative abnormal return for all post-event analysis. The calculation of abnormal return for each day of analysis is done using the formula: AR itd = K itd E(R it ), where AR itd the abnormal return of an index i, on a day t, for an event d; R itd the return of an index i, on a day t, for an event d; E(R it ) the normal return using the CAPM model of an index i, for day t. The calculation of average abnormal return for each day is done using the formula: n AR it 1/ n ARitd, t d 1 The calculation of cumulative abnormal return for all post-event analysis is done using the formula: CAR it = CAR i(t-1) + AR it To check if the cumulative abnormal returns are statistically different from zero, we make a t-test using the following formula. CARit t VAR CAR * n it it j 6

7 3. Empirical results Table 3.1 presents the normal returns expected by investors and how they were calculated using the risk-free rate, the difference of inflation for each currency in the three countries against the euro, the country risk premium and U.S. capital markets risk. Romania Bulgaria Austria Risk-free rate Difference of 3,9 8,2 0 inflation Country premium 3,9 3,9 0 risk U.S. capital market risk Normal return Daily normal return 0,057% 0,061% 0,025% Table 3.1. We can see that the normal annual returns expected by investors are similar in the case of Romania and Bulgaria, but much smaller in the case of Austria. This difference may be explained by the fact that these two countries represents emerging markets while Austria is a country developed. Austria is a market with a rich history and a low volatility, compared with Romania and Bulgaria who just started to build their capital markets and manifest a high volatility. Index Ordinary Post-event Post-positiveevent Post-negativeevent BET (Romania) -0,30% -0,64% -0,52% 0,51% SOFIX (Bulgaria) ATX (Austria) -0,42% -0,92% -2,12% -0,20% -0,05% -0,84% -0,90% -0,44% Table 3.2. First, we must observe that all the three indices had negative return in The best performance was obtained by ATX index with an average daily return of %, which can be explained by the higher degree of maturity of the Austrian capital market. The other two indices SOFIX and BET had evolutions ATX below ATX, but similar (-0.30% and -0.42%), which confirms that the perspective of investors on the two capital markets is the same. The post-event return is consistently lower than the normal return. This can be explained by the fact that 2008 was a year marked in the great majority of negative news, which is observed when seeing the performances of all capital markets around the world. 7

8 The most interesting is that the return of post-positive events is lower than the post-negative events. A possible explanation may be offered by over-reaction theory. After the trigger of a positive event the evolution of the capital markets is positive. In the first 3-4 after the news is published, the market investors speculate the moment and the indices increase considerably. After those first, the indices tend to resume their previous trend and correct to the initial values and, in most cases fall below these values. Index Day Standard deviation BET Ordinary 0, (Romania) Post-event 0, SOFIX (Bulgaria) ATX (Austria) Post-positive-event 0, Post-negative-event 0, Ordinary 0, Post-event 0, Post-positive-event 0, Post-negative-event 0, Ordinary 0, Post-event 0, Post-positive-event Post-negative-event Table , ,

9 We can see that the variance for post-event is greater than the variance for the ordinary, which indicates that volatility is higher in these post-event periods. This observation may be explained by the uncertain information hypothesis which predicts a high volatility around unexpected events. D BET(Romania) SOFIX(Bulgaria) ATX(Austria) a y Positive event Negative event Positive event Negative event Positive event Negative event 1 4,68% -2,99% 0,15% -2,57% 5,51% -3,69% 2-2,22% -0,03% 0,91% 0,75% -3,58% 0,32% 3-0,86% -2,44% -1,91% -1,33% -3,86% -1,90% 4-0,67% -0,47% -1,54% 0,07% 1,65% 0,68% 5 2,72% 1,36% -0,62% 0,28% 0,35% 1,20% 6-4,45% -0,53% -2,55% 0,19% -3,81% -1,15% 7-4,76% -0,62% -4,47% -0,30% -5,17% -0,71% 8-1,67% -0,34% -3,57% 0,02% -4,82% -0,47% 9-1,22% 0,17% -5,04% 0,10% -0,21% 1,12% 10 2,67% 0,25% -3,20% 0,18% 4,70% 0,01% Table 3.4. In the case of Romania, after the first 5 when an event occurs we obtain positive returns. After passing this short period of 5, news began to be understood fully by the market and the BET index corrects to the usual returns before the event. Often these decreases tend to correct the stock price more than the average. As I said before, this can be explained by the theory of over-reaction. Thus, taking into account the rare positive moments, the market speculators take advantage of this moment and artificially inflate their importance. When they close their positions, they make the prices to fall below average due to over-all reaction. A similar thing happens and the market in Austria. The first 4-5 after a positive event triggers generate good performances of the market but this overreaction will be corrected later. The Bulgarian market is atypical in this regard. Although the returns are positive, after positive events trigger they are not very high. The resemblance consists in the fact that after the passing of 4-5 ( noticed on the other markets), the SOFIX resumes its descending trend. If we review the negative returns we note that the investors' reaction is less unpredictable. Thus, although returns are below the average daily rates, they don t exceed in amplitude the response after positive events. After the first three of steep decreases, markets tend to be correct, even recording positive returns. This phenomenon can be best explained by efficient market theory, which says that after any increase or decrease the magnitude of correction following a smaller scale to restore equilibrium in the market. 9

10 D a Pozitive event Negative events BET SOFIX ATX BET SOFIX ATX CAR t CAR t CAR t CAR t CAR value value value value y CAR t value t value 1 4,68% 0,305 0,15% 0,006 5,51% 0,260 2,99% 0,818 2,57% 2,175 3,69% 1, ,46% 0,160 1,06% 0,041 1,94% 0,091 3,02% 0,827 1,82% 1,540 3,37% 1, ,60% 0,104 0,85% 0,033 1,92% 0,091 5,46% 1,494 3,15% 2,669 5,27% 2, ,93% 0,060 2,39% 0,093 0,28% 0,013 5,93% 1,622 3,08% 2,607 4,59% 1, ,65% 0,238 3,01% 0,117 0,08% 0,004 4,57% 1,250 2,80% 2,372 3,39% 1, ,80% 0,052 5,56% 0,215 3,74% 0,177 5,09% 1,394 2,62% 2,214 4,54% 1, ,56% 0,362 10,03% 0,389 8,90% 0,421 5,71% 1,563 2,92% 2,471 5,25% 2, ,23% 0,471 13,60% 0,527 13,72% 0,648 6,05% 1,656 2,90% 2,458 5,73% 2, ,45% 0,550 18,63% 0,722 13,94% 0,658 5,88% 1,610 2,80% 2,372 4,61% 1,808 5,78% 10 0,376 21,84% 0,846 9,24% 0,436 5,64% 1,543 2,63% 2,223 4,60% 1,806 Table 3.5. Analyzing the table we can see a very high volatility of the cumulative abnormal return, especially in case of positive events. In the case of BET, and in the ATX case these cumulative abnormal returns are maintained positive until the fifth day, which means that the market considered as positive the events, especially as the general trend in 2008 was a negative for all three indices. We can say that after an event triggers we witness an over-reaction of investors, whom after hearing some positive news have over-bid in the hope of a possible conversions of the generally negative trend. However, this positive market reaction is not one of duration, because indications correct and return to negative values. The smallest value in the last day of analysis (the tenth) of a cumulative abnormal return can be seen at the SOFIX index. This, though it respects the rule of the few of increases after positive news (the amplitude is lower than ATX and BET), it corrects the most drastic of all, emphasizing pessimism investors on capital markets in general, and the Bulgarian capital markets in particular for the year In the case of negative events, the cumulative abnormal return has developed in a more peaceful way than positive events. The negative slope of the cumulative abnormal returns is smoother than the slope for positive events; however it is more abrupt than usual market returns, which signals us that these events are incorporated by the market. At the end of the of 10 analysis, the cumulative abnormal returns have a slight positive correction which tends to bring the market to a certain balance. It is interesting to see that these negative corrections make cumulative abnormal returns for negative events to take in the end higher values than the cumulative abnormal returns for positive events. The graphic representations of this information give us a more practical perspective of the abnormal returns. 10

11 Graphic 1 This graph shows that the Romanian market has the most consistent reaction after an event is triggered in the first day. This reaction is maintained in the first 4-5. The difference between reactions occurs after the news is understood by the market. Thus, after positive events, the corrective trend is strong, while for negative events, the reaction is to return to the return close to the normal one. Graphic 2 In comparison with Romania, Bulgaria has an atypical behavior, especially for positive events. Thus, although the trigger of positive events, the market offers positive returns, they are lower than the returns on the Romanian market. The lack of optimism about the future benefits of the announced event is seen fairly quickly, exactly two after the event, when the market suffers a consistent decline. In terms of negative events, they influence the market in a consistent manner in the first 2-3, after which the market tends to return close to the normal. From this point of view we can see some similarities with the Romanian market that stabilizes after

12 Graphic 3 The Austrian market is better alike with Romania in terms of post-event evolutions. Thus, after the positive events triggers, in the first 4-5, the market reacts very well, obtaining high return. After this period of 4-5, ATX records negative returns of the same magnitude of increases after the event. This high volatility can be explained as in the case of Romania, an over-reaction, both positive and negative from the market waiting for some positive news. For the negative events we can see a succession of negative returns on a range of three after the event triggers, followed by a positive correction, followed by market stabilization. This evolution is similar to the evolution of the Romanian capital market. Graphic 4 The evolution line of cumulative abnormal return for positive events presents a general downward trend, starting with the first day after announcing the event and ending with the tenth day. The highest value is found in the very first day. Abnormal return is maintained until the fifth day, when the CAR falls below the zero line, the line forming a rather abrupt slope. This could be a sign of over-reaction from the market speculated that when positive news announcement. After the fall of the CAR line, although is still below 0, there are signs that it finds its way to this value, that is to balance, according to the efficient markets theory. In case of negative events, we first observe, that during the 10 analyzed the cumulative abnormal returns line position is always below the value 0. The highest value is found in the first day, after that it follows a pretty serious decline until the fourth day, which shows us an over-reaction of the market. Interestingly, with the end of ten the effects of the event are still felt in the market, the line CAR not providing signs that would point towards the value 0. 12

13 Graphic 5 As in the other analysis, the analysis of cumulative abnormal returns, Bulgaria reacts atypical in positive events case. We can see that although we are dealing with a positive event, the CAR line lies very little (only two ) over zero value. This evolution can be explained by the fact that the market had anticipated these events and embedded the value-addition of these them earlier and then, finding the news, the market was over-early response, in comparison with other markets analyzed. Another explanation is the perception of investors on the Bulgarian market on the happened events, or the lack of optimism towards possible advantages brought by the positive news. If for the positive events the SOFIX behavior was different from the BET, for the negative events the CAR line has about the same evolution. Thus, the cumulative abnormal returns line decreases, in the SOFIX case, steepen in the first three analyzed, and then develops lateral without giving signs of returning to 0. Graphic 6 The evolution of cumulative abnormal returns line for the ATX index is quite similar to the evolution of BET index. Thus, the highest value is obtained in the very first day, and then, during the 10 analyze, the trend is downward. The ATX CAR line reaches sooner 0 after the event is triggered, in only three, but at 5 after the event is triggered (as in the case of Romania), the CAR corrects strongly decreasing below 0. As in the case of Romania, we have to do with an over-reaction positive and negative. In the end, the returns begin to increase, indicating that our market is trying to return to balance. In case of negative events, the evolution is similar to Romania, in the sense that the cumulative abnormal returns line remains always below 0. However, we 13

14 should note an evolution especially lateral for all 10 analyzed, without giving any sign that might close to Conclusions and recommendations Although, we analyzed two emerging markets and a developed, the result was the same, no matter how mature they are, the capital markets react to events. In understanding these results, we should keep in mind the year chosen for analysis. We are in a period in which globalization has reached an incredible level. Thus, one of the news all over the world can do around the world in seconds thanks to advance technology and communication. Therefore, financial markets, especially those who operate based on market economy principles, tend to experience similar effects after an important event. Also, 2008 was noted by big decreases of capital markets, which staggered seriously the plans of long-term investors and encourage the type of speculative investment, which in turn increased market volatility. Firstly, the results of our tests indicate an increased volatility around the events for all three markets analyzed. Interesting is that if we analyze by type of events, the higher volatility is met around positive events. This should not be a surprise if we think that 2008 has been sprinkled with very few positive events, most events being negative. The returns obtained in particular after the first of trading, are beyond the normal returns for the positive events and are well below the normal returns for the negative events. These over-reactions were observed on all three markets, the difference is at the number of required for the market to return to the normal state. The longest periods of speculations have been met in the case of Romania and Bulgaria, the market in Austria giving signs that it returns to normal state sooner after an event. These speculations can be explained using the theory of over-reaction. Thus, immediately after an event is triggered, the speculators set prices very high for the positive events and are very low for the negative events to take advantage of moments of uncertainty, immediately after the event is triggered. Interesting is that after the over-reaction, the market tends to be corrected quite drastically, especially after positive events. This is to emphasize the theory exposed above, explaining the rare positive events in 2008, and their role in the strategies of short term investors. Another explanation for higher returns after a positive event is triggered is offered by the uncertain information theory, which says that investors prefer to take positions below the fundamental value of shares due to uncertainty over what the event will happen. In the case of strong corrections occurred after large increases induced by positive events, but, also in the case of decreases after negative events, the indices return tend to return as close to the normal return, meaning it tends towards equilibrium. This evolution is best explained by the theory of efficient markets. In conclusion, the Romanian market generates abnormal returns and increased volatility following important global events. Things are the same for the other two capital markets considered. Therefore, a possible recommendation to investors on the capital markets is to avoid long-term investments and invest on short or medium term, especially in these times marked by high volatility. This can be inferred from differences between normal return computed using CAPM model, a model considered to be a tool used mainly for long-term analysis and the effective return calculated on daily returns in Also, 14

15 in an attempt to interpret as correctly as possible the market reaction after an event is triggered we recommend using the technical analysis over fundamental analysis. Bibliography: 1. Anamaria Ciobanu, Seyed Mehdian, Mark J. Perry (2008) An Analysis of Investors Behavior in the Romanian Capital Market, Susmita Dasgupta, Benôit Laplante, Nlandu Mamingi (1997) Capital Market Responses to Environmental Performance in Developing Countries, 12-17, Burton G. Malkiel (2003) The Efficient Market Hypothesis and Its Critics, CEPS Working Paper No. 91, Laura Obreja Braşoveanu, Anamaria Ciobanu (2008) Estimating Equity Risk Premium for Valuating a Romanian Company. The Case of Antibiotice SA, Central and Eastern European Library, James Ryan, Mark Donnelly (2000) The Overreaction Hypothesis: An Examination in the Irish Stock Market, DCU Business School Paper No. 38, Keith C. Brown, W.V. Harlow, Seha M. Tinic (1988) Risk Aversion, Uncertain Information and Market Efficiency, Journal of Financial Economics 22,