ﺑﺮرﺳﯽ رﻓﺘﺎر رﻣﻪ اي ﻣﺘﻐﯿﺮ زﻣﺎن در ﺑﻮرس اوراق ﺑﻬﺎدار ﺗﻬﺮان ﺑﺎ اﺳﺘﻔﺎده از ﻣﺪل اﻣﺘﯿﺎز ﺧﻮدرﮔﺮﺳﯿﻮﻧﯽ ﺗﻌﻤﯿﻢ ﯾﺎﻓﺘﻪ

Investigating Time Varying Herd Behavior in Tehran Stock Exchange: Generalized Autoregressive Score Approach

رﻓﺘﺎر ﺗﻮده وار ﯾﺎ رﻣﻪ اي ﯾﮑﯽ از ﻣﻬﻤﺘﺮﯾﻦ ﺳﻮﮔﯿﺮي ﻫﺎي رﻓﺘﺎري اﺳﺖ ﮐﻪ در ﺑﺎزارﻫﺎي ﻣﺎﻟﯽ وﺟﻮد دارد و از ﻋﻮاﻣﻞ ﺷﮑﻞ دﻫﻨﺪة ﺑﺤﺮان ﻫﺎي ﻣﺎﻟﯽ اﺳﺖ. ﺑﺎﺗﻮﺟﻪ ﺑﻪ اﯾﻨﮑﻪ رﻓﺘﺎر رﻣﻪ اي ﺑﻪ ﺻﻮرت ﻣﺴﺘﻘﯿﻢ ﺑﺮ ﻗﯿﻤﺖ اﺛﺮ ﻣﯽ ﮔﺬارد، از اﯾﻦ رو ﭘﯿﺶ ﺑﯿﻨﯽ ﻗﯿﻤﺖ ﺑﺮ اﺳﺎس داده ﻫﺎي ﻗﯿﻤﺘﯽ ﮔﺬﺷﺘﻪ، ﻧﺸﺎن از وﺟﻮد رﻓﺘﺎر رﻣﻪ اي در ﺑﺎزار دارد. اﯾﻦ ﻣﻘﺎﻟﻪ ﺑﺎ ﻫﺪف ﺑﺮرﺳﯽ وﺟﻮد رﻓﺘﺎر رﻣﻪ اي در ﺑﻮرس اوراق ﺑﻬﺎدار ﺗﻬﺮان، ﻣﺪل زﻣﺎن ﻣﺘﻐﯿﺮ ﻏﯿﺮﺧﻄﯽ ﻧﻮﯾﻨﯽ ﺑﻪ ﻧﺎم اﻣﺘﯿﺎز ﺧﻮدرﮔﺮﺳﯿﻮﻧﯽ ﺗﻌﻤﯿﻢ ﯾﺎﻓﺘﻪ )GAS( اراﺋﻪ ﮐﺮده و ﺑﺎ ﻣﺪل ﻫﺎي ﻏﯿﺮﺧﻄﯽ ﺳﻨﺘﯽ GARCH و AR ﻧﯿﺰ ﻗﯿﺎس ﺷﺪه اﺳﺖ. در راﺳﺘﺎي ﭘﯿﺶ ﺑﯿﻨﯽ ﺑﺎزدﻫﯽ ﺷﺎﺧﺺ ﺑﻮرس اوراق ﺑﻬﺎدار ﺗﻬﺮان ﺟﻬﺖ ﺗﺸﺨﯿﺺ وﺟﻮد رﻓﺘﺎر رﻣﻪ اي، از داده ﻫﺎي ﻗﯿﻤﺘﯽ ﻃﯽ ﺑﺎزة 1390 اﻟﯽ 1399 اﺳﺘﻔﺎده ﺷﺪه اﺳﺖ. ﺗﻮان و دﻗﺖ ﭘﯿﺶ ﺑﯿﻨﯽ ﺗﻮزﯾﻊ ﺑﺎزدﻫﯽ ﻣﺪل ﻧﻮﯾﻦ GAS ﺑﺎ ﻧﺘﺎﯾﺞ ﻣﺪل ﻫﺎي ﻏﯿﺮﺧﻄﯽ GARCH و AR در داده ﻫﺎي درون و ﺑﺮون ﻧﻤﻮﻧﻪ اي ﺟﻬﺖ ﺗﺸﺨﯿﺺ وﺟﻮد رﻓﺘﺎر رﻣﻪ اي ﻗﯿﺎس ﺷﺪه اﺳﺖ. ﻧﺘﺎﯾﺞ ﭘﮋوﻫﺶ در آزﻣﻮن ﻫﺎي درون و ﺑﺮون ﻧﻤﻮﻧﻪ اي ﻧﺸﺎن دﻫﻨﺪة دﻗﺖ ﺑﺎﻻﺗﺮ ﻣﺪل ﻧﻮﯾﻦ GAS ﻧﺴﺒﺖ ﺑﻪ ﻣﺪل ﻫﺎي ﺳﻨﺘﯽ GARCH و AR در ﭘﯿﺶ ﺑﯿﻨﯽ ﺗﻮزﯾﻊ ﺑﺎزدﻫﯽ روزاﻧﮥ ﺷﺎﺧﺺ ﮐﻞ ﺑﻮرس اوراق ﺑﻬﺎدار ﺗﻬﺮان و وﺟﻮد رﻓﺘﺎر رﻣﻪ اي در ﺑﺎزار ﺳﺮﻣﺎﯾﻪ اﯾﺮان ﺑﻮده اﺳﺖ.

Herd behavior is one of the most important behavioral biases in financial markets and is one of the determinants of financial crises. Given that herd behavior directly affects price, presenting a model based solely on past prices, with good predictability, indicates the existence of market herd behavior. This article aims to investigate the existence of herd behavior in Tehran Stock Exchange and presents a new nonlinear variable time model called Generalized Autoregressive Score (GAS) and to compare with traditional GARCH and AR nonlinear models, in order to predict the returns distribution of the total index of the stock exchange during the period 2010 to 2020. The results of modeling for the asset by the new GAS model are compared with the results of the GARCH and AR models and their performance are tested for inside and outside the sample. Sample in the internal and external tests show that the new GAS model is more accurate than the traditional GARCH and AR models in predicting the daily returns distribution of the total index of the Tehran Stock Exchange and also the presence of herd behavior in Iran's capital market has been approved.