##
تقنيات الاقتصاد القياسي لقياس التقلبات في اسعار الاسهم المدرجة في سوق فلسطين للاوراق المالية خلال مراحل الصعود والهبوط وقياس العلاقة السببية ما بين سوق فلسطين وسوق عمان وسوق تل ابيب
Econometric Techniques to Examine Volatility in PEX Bulls and Bears and the Causal Relationship between PEX, ASE and TASE

تقنيات الاقتصاد القياسي لقياس التقلبات في اسعار الاسهم المدرجة في سوق فلسطين للاوراق المالية خلال مراحل الصعود والهبوط وقياس العلاقة السببية ما بين سوق فلسطين وسوق عمان وسوق تل ابيب
Econometric Techniques to Examine Volatility in PEX Bulls and Bears and the Causal Relationship between PEX, ASE and TASE

##### Date

2012-06-27

##### Authors

عبد الرحمن موسى حسين العويسات

Abdel Rahman Mousa Hussein Aliwisat

##### Journal Title

##### Journal ISSN

##### Volume Title

##### Publisher

AL-Quds University

جامعة القدس

جامعة القدس

##### Abstract

This study is empirically aimed at conducting three tests; testing volatility persistent in PEX
bulls and bears, testing market efficiency for PEX, ASE, and TASE, and testing the causality
relationship between the three markets. That is, it attempts to explore whether stock market
volatility present a different behavior during PEX bulls and bears phases and explore whether
PEX, ASE, and TASE are efficient at weak level. For this purpose, long memory measure is used
to indicate volatility persistence and market efficiency. In order to define bull and bear phases,
we employed the 200-day moving average, already used by practioners and we found three
cycles including 3 bulls and 3 bears. Thus, the study employed Rescaled Range (R/S) to
calculate the values of difference parameter d to find evidence of long memory behavior for the
daily data observations from August, 1997 to March, 2012. In addition to a long memory
measure, the study used nonparametric ADF and PP tests to test market efficiency of PEX, ASE,
and TASE at weak level.
According to Jarque–Bera test, the closing values of Al-Quds Index of PEX in each bull
and bear don’t follow the normal probability distribution. So, the study used nonparametric tests
of ADF and PP to determine whether the time series are stationary. The time series are found to
be non stationary at level in each phase implying that PEX is efficient at weak level in each
phase. Further, according to R/S results, the study found that the estimates of parameter d are
above 0 and below 0.5 for bear phases, while the values are above 0.5 for the bull phases
implying long memory stationarity for the volatility process. This means that volatility is more
persistent in the PEX bears markets than in the PEX bull markets. Further, the PEX bears
markets are longer than PEX bulls markets. As a result, volatility persistent in PEX bears and
risk associated with it should be considered by investors. Added to this, the overall market-adjusted performance measurement indicates that PEX has average levels of returns and risk
more than ASE and TASE. To avoid that, investors and other decision makers should consider
both fundamental and technical analysis.
For market efficiency test, ADF and PP test are also used to find whether time series data
of Al-Quds index, ASE index and TA-100 index are stationarity. In the three cases, means and
variances seem to be not constant. This indicates that the three indices are found to be nonstationary
at level implying that the three markets are efficient at weak level. For further
investigation, R/S statistic is used to calculate the difference parameter to indicate market
efficiency. The estimates of d are above 0.5 for the PEX and TASE cases implying that time
series data are non-stationary, and there is no evidence of long memory behavior (long range
dependence) in the time series data. For ASE, the value of d is above 0 and below 0.5 implying
that the time series has long memory behavior. This indicates that ASE isn’t efficient at weak
level. So, we accept that PEX, and TASE are efficient at weak level but ASE isn’t. Therefore,
regulators and policy makers should support market efficiency.The study further investigates correlation and causality relationship among PEX, ASE
and TASE. It analyzes whether there is a long run linkage or interdependency between the three
markets. The data sample includes daily observations for the January, 2000-March, 2012 time
period. As mentioned before, the data are non-stationary at level, while the data are stationary at
first difference and therefore conducting Granger causality tests isn’t restricted. The correlation
matrix indicates that the three markets aren’t highly correlated. The correlation is verified for the
direction of influence by the Granger causality test between the three markets. However, the
study found that there is no significant causal relationship between the three markets except theunilateral causality relationship of ASE over PEX, and the relationship of TASE over ASE,
whereas reverse causality doesn’t hold true.
In general, the study finds that there is no multilateral causal relationship among the three
markets and they are being highly correlated. Therefore, Palestinian investors don’t have to
consider changes in TASE index, while changes in ASE index must be considered.

This study is empirically aimed at conducting three tests; testing volatility persistent in PEX bulls and bears, testing market efficiency for PEX, ASE, and TASE, and testing the causality relationship between the three markets. That is, it attempts to explore whether stock market volatility present a different behavior during PEX bulls and bears phases and explore whether PEX, ASE, and TASE are efficient at weak level. For this purpose, long memory measure is used to indicate volatility persistence and market efficiency. In order to define bull and bear phases, we employed the 200-day moving average, already used by practioners and we found three cycles including 3 bulls and 3 bears. Thus, the study employed Rescaled Range (R/S) to calculate the values of difference parameter d to find evidence of long memory behavior for the daily data observations from August, 1997 to March, 2012. In addition to a long memory measure, the study used nonparametric ADF and PP tests to test market efficiency of PEX, ASE, and TASE at weak level. According to Jarque–Bera test, the closing values of Al-Quds Index of PEX in each bull and bear don’t follow the normal probability distribution. So, the study used nonparametric tests of ADF and PP to determine whether the time series are stationary. The time series are found to be non stationary at level in each phase implying that PEX is efficient at weak level in each phase. Further, according to R/S results, the study found that the estimates of parameter d are above 0 and below 0.5 for bear phases, while the values are above 0.5 for the bull phases implying long memory stationarity for the volatility process. This means that volatility is more persistent in the PEX bears markets than in the PEX bull markets. Further, the PEX bears markets are longer than PEX bulls markets. As a result, volatility persistent in PEX bears and risk associated with it should be considered by investors. Added to this, the overall market-adjusted performance measurement indicates that PEX has average levels of returns and risk more than ASE and TASE. To avoid that, investors and other decision makers should consider both fundamental and technical analysis. For market efficiency test, ADF and PP test are also used to find whether time series data of Al-Quds index, ASE index and TA-100 index are stationarity. In the three cases, means and variances seem to be not constant. This indicates that the three indices are found to be nonstationary at level implying that the three markets are efficient at weak level. For further investigation, R/S statistic is used to calculate the difference parameter to indicate market efficiency. The estimates of d are above 0.5 for the PEX and TASE cases implying that time series data are non-stationary, and there is no evidence of long memory behavior (long range dependence) in the time series data. For ASE, the value of d is above 0 and below 0.5 implying that the time series has long memory behavior. This indicates that ASE isn’t efficient at weak level. So, we accept that PEX, and TASE are efficient at weak level but ASE isn’t. Therefore, regulators and policy makers should support market efficiency.The study further investigates correlation and causality relationship among PEX, ASE and TASE. It analyzes whether there is a long run linkage or interdependency between the three markets. The data sample includes daily observations for the January, 2000-March, 2012 time period. As mentioned before, the data are non-stationary at level, while the data are stationary at first difference and therefore conducting Granger causality tests isn’t restricted. The correlation matrix indicates that the three markets aren’t highly correlated. The correlation is verified for the direction of influence by the Granger causality test between the three markets. However, the study found that there is no significant causal relationship between the three markets except theunilateral causality relationship of ASE over PEX, and the relationship of TASE over ASE, whereas reverse causality doesn’t hold true. In general, the study finds that there is no multilateral causal relationship among the three markets and they are being highly correlated. Therefore, Palestinian investors don’t have to consider changes in TASE index, while changes in ASE index must be considered.

This study is empirically aimed at conducting three tests; testing volatility persistent in PEX bulls and bears, testing market efficiency for PEX, ASE, and TASE, and testing the causality relationship between the three markets. That is, it attempts to explore whether stock market volatility present a different behavior during PEX bulls and bears phases and explore whether PEX, ASE, and TASE are efficient at weak level. For this purpose, long memory measure is used to indicate volatility persistence and market efficiency. In order to define bull and bear phases, we employed the 200-day moving average, already used by practioners and we found three cycles including 3 bulls and 3 bears. Thus, the study employed Rescaled Range (R/S) to calculate the values of difference parameter d to find evidence of long memory behavior for the daily data observations from August, 1997 to March, 2012. In addition to a long memory measure, the study used nonparametric ADF and PP tests to test market efficiency of PEX, ASE, and TASE at weak level. According to Jarque–Bera test, the closing values of Al-Quds Index of PEX in each bull and bear don’t follow the normal probability distribution. So, the study used nonparametric tests of ADF and PP to determine whether the time series are stationary. The time series are found to be non stationary at level in each phase implying that PEX is efficient at weak level in each phase. Further, according to R/S results, the study found that the estimates of parameter d are above 0 and below 0.5 for bear phases, while the values are above 0.5 for the bull phases implying long memory stationarity for the volatility process. This means that volatility is more persistent in the PEX bears markets than in the PEX bull markets. Further, the PEX bears markets are longer than PEX bulls markets. As a result, volatility persistent in PEX bears and risk associated with it should be considered by investors. Added to this, the overall market-adjusted performance measurement indicates that PEX has average levels of returns and risk more than ASE and TASE. To avoid that, investors and other decision makers should consider both fundamental and technical analysis. For market efficiency test, ADF and PP test are also used to find whether time series data of Al-Quds index, ASE index and TA-100 index are stationarity. In the three cases, means and variances seem to be not constant. This indicates that the three indices are found to be nonstationary at level implying that the three markets are efficient at weak level. For further investigation, R/S statistic is used to calculate the difference parameter to indicate market efficiency. The estimates of d are above 0.5 for the PEX and TASE cases implying that time series data are non-stationary, and there is no evidence of long memory behavior (long range dependence) in the time series data. For ASE, the value of d is above 0 and below 0.5 implying that the time series has long memory behavior. This indicates that ASE isn’t efficient at weak level. So, we accept that PEX, and TASE are efficient at weak level but ASE isn’t. Therefore, regulators and policy makers should support market efficiency.The study further investigates correlation and causality relationship among PEX, ASE and TASE. It analyzes whether there is a long run linkage or interdependency between the three markets. The data sample includes daily observations for the January, 2000-March, 2012 time period. As mentioned before, the data are non-stationary at level, while the data are stationary at first difference and therefore conducting Granger causality tests isn’t restricted. The correlation matrix indicates that the three markets aren’t highly correlated. The correlation is verified for the direction of influence by the Granger causality test between the three markets. However, the study found that there is no significant causal relationship between the three markets except theunilateral causality relationship of ASE over PEX, and the relationship of TASE over ASE, whereas reverse causality doesn’t hold true. In general, the study finds that there is no multilateral causal relationship among the three markets and they are being highly correlated. Therefore, Palestinian investors don’t have to consider changes in TASE index, while changes in ASE index must be considered.

##### Description

##### Keywords

المحاسبة والضرائب,
Accounting & Taxation