Search Articles

Home / Articles

AN ANALYSIS OF LONG-RANGE CORRELATION AND HEAVY TAIL: KSE – 100 INDEX CLOSING BASED ON NEWS SENTIMENTS AS CASE STUDY

. Asma Zaffar, ovais Siraj and Muhammad Ehsanullah


Abstract

KSE-100 index (1st -12th) cycles are stationary behavior. Heavy tail analysis behavior is examined. Long-range correlation persistency is calculated in the perspective of heavy tail analysis. Each value of KSE-100 index data is strongly correlated to previous data. The differencing parameter of all KSE-100 index cycles have ranging from 0< d < 0.5 in both self-affine (dA) and self-similar (dS) fractal dimension. Heavy tail parameter ( ) asymptotically adhere to the Pareto law that indicate the dynamics is periodic and regular. Heavy tail parameter ( ) and differencing parameter (d = H-0.5) is developed from the Hurst Exponent 0.5 < H < 1(persistent). The self-similar long-range correlation strength is (1 < ) and self-affine long-range correlation strength is (-1< ). KSE-100 index cycles are strongly long-range correlated besides cycles 9th. The Phillips-Perron (PP) unit root tests is applied in KSE-100 index cycles which are shown, every cycle has stationary nature. This study confirms that the six months break up shown that KSE-100 index market was persistent in evaluate years. The economy is steadily progress in the direction of established state of the structure of market. The concept of efficiency would accordingly integrate time varying, evolutionary and behavioral aspect of the development of market.

 

Keywords: Long-range correlation, Heavy Tail Analysis, Hurst Exponent (H), Persistent, KSE-100 index cycles.

Download :