ﭘﯿﺶ ﺑﯿﻨﯽ روﻧﺪ ﺣﺮﮐﺖ ﻗﯿﻤﺖ ﺟﻬﺎﻧﯽ ﻃﻼ ﺑﺎ روﯾﮑﺮد ﻣﺪلﺳﺎزي ﺗﻮزﯾﻊﻫﺎي ﺣﺎﺷﯿﻪاي: ﮐﺎرﺑﺮدي از ﻣﺪلﻫﺎي ﮔﺎرچ ﮐﺎﭘﻮﻻي ﮔﻮﺳﯽ و ﺗﯽ

Forecasting the Global Gold Price Movement with Marginal Distribution Modeling Approach: An Application of the Copula GARCH Gaussian and t

ﺑﺎ ﺗﻮﺟﻪ ﺑﻪ اﻫﻤﯿﺖ ﻗﯿﻤﺖ ﻃﻼ در ﺑﺎزارﻫﺎي ﻣﺎﻟﯽ و اﺛﺮات اﻗﺘﺼﺎدي ﺣﺎﺻﻞ از ﻧﻮﺳﺎﻧﺎت ﻗﯿﻤﺘﯽ آن، روﻧﺪ ﺗﻐﯿﯿﺮات ﻗﯿﻤﺖ ﻃﻼ در اﻗﺘ ﺼﺎد ﻣﻠﯽ و ﺟﻬﺎﻧﯽ، ﺗﻮﺟﻪ ﺑ ﺴﯿﺎري از ﻣﺤﻘﻘﺎن و ﺗﺤﻠﯿﻠﮕﺮان اﻗﺘ ﺼﺎدي را ﺑﻪ ﺧﻮد ﺟﻠﺐ ﮐﺮده اﺳﺖ. از اﯾﻦ رو ﻫﺪف اﺻﻠﯽ اﯾﻦ ﻣﻄﺎﻟﻌﻪ، ﭘﯿﺶﺑﯿﻨﯽ روﻧﺪ ﺣﺮﮐﺖ ﻗﯿﻤﺖ ﻃﻼي ﺟﻬﺎﻧﯽ ﻣﯽﺑﺎﺷﺪ اﯾﻦ ﭘﮋوﻫﺶ ﺑﺎ ﻫﺪف ﻣﻌﺮﻓﯽ ﯾﮏ اﻟﮕﻮي ﺗﺮﮐﯿﺒﯽ از ﻣﺪلﻫﺎي ﮐﺎﭘﻮﻻ و ﻣﺪل ﮔﺎرچ ﮐﻼﺳـــﯿﮏ -GARCH( )Copula و ﻣﻘﺎﯾﺴـــﻪ آن ﺑﺎ ﻣﺪل ﻫﺎي ﺧﺎﻧﻮاده ﮔﺎرچ، ﺟﻬﺖ ﭘﯿﺶ ﺑﯿﻨﯽ روﻧﺪ ﺣﺮﮐﺖ ﻗﯿﻤﺖ ﺟﻬﺎﻧﯽ ﻃﻼ در ﺑﺎزه زﻣﺎﻧﯽ 2000 /01 /04 ﺗﺎ2018 /06 /26 ، در اﻓﻖ ﻫﺎي ﭘﯿﺶ ﺑﯿﻨﯽ 10 ،5 ،1، و 22 روزه، ﺻـــﻮرت ﭘﺬﯾﺮﻓﺘﻪ اﺳـــﺖ. دﻗﺖ ﭘﯿﺶ ﺑﯿﻨﯽ ﻣﺪل ﻫﺎي ﻣﺬﮐﻮر ﺑﺎ اﺳـــﺘﻔﺎده از ﻣﻌﯿﺎر ﺧﻄﺎي RMSE، ﻣﻮرد ارزﯾﺎﺑﯽ و ﻣﻘﺎﯾﺴـــﻪ ﻗﺮار ﮔﺮﻓﺘﻪ اﻧﺪ . ﻧﺘﺎﯾﺞ ﺑﺪﺳـــﺖ آﻣﺪه ﺣﺎﮐﯽ از آن ﺑﻮد ﮐﻪ در اﻓﻖ ﻫﺎي ﭘﯿﺶﺑﯿﻨﯽ ﮐﻮﺗﺎهﻣﺪت ﻣﺪل ﮐﺎﭘﻮﻻي ﻧﺮﻣﺎل ﺑﺎ ﺗﻮزﯾﻊ ﺣﺎﺷــﯿﻪاي GARCH-t و در اﻓﻖ ﭘﯿﺶﺑﯿﻨﯽ ﺑﻠﻨﺪﻣﺪت ﻣﺪل ﮐﺎﭘﻮﻻي ﺗﯽ ﺑﺎ ﺗﻮزﯾﻊ ﺣﺎﺷﯿﻪ اي GARCH-t از ﻋﻤﻠﮑﺮد ﺑﻬﺘﺮي ﻧﺴﺒﺖ ﺑﻪ ﻣﺪل ﻫﺎي رﻗﯿﺐ ﺑﺮﺧﻮردار ﺑﻮدﻧﺪ. ﻣﺪل ﺗﺮﮐﯿﺒﯽ ﻣﻌﺮﻓﯽ ﺷـــﺪه در اﯾﻦ ﭘﮋوﻫﺶ داراي ﭘﺘﺎﻧﺴـــﯿﻞ ﺑﺎﻻﯾﯽ در ﺟﻬﺖ ﭘﯿﺶ ﺑﯿﻨﯽ روﻧﺪ ﺣﺮﮐﺖ ﻗﯿﻤﺖ ﻃﻼي ﺟﻬﺎﻧﯽ ﻣﯽﺑﺎﺷ ﺪ، ﺑﻨﺎﺑﺮاﯾﻦ اﺳﺘﻔﺎده از اﯾﻦ ﻣﺪل ﺑﺮاي ﺳﺮﻣﺎﯾﻪ ﮔﺬاران ﺑﺨﺶﻫﺎي ﻣﺨﺘﻠﻒ، ﺗﺤﻠﯿﻠﮕﺮان اﻗﺘﺼﺎدي و ﻧﯿﺰ ﺑﺮﻧﺎﻣﻪرﯾﺰان ﮐﻼن ﮐﺸﻮر ﻧﺘﺎﯾﺞ ارزﻧﺪه اي را ﻣﯽﺗﻮاﻧﺪ داﺷﺘﻪ ﺑﺎﺷﺪ.

Given the importance of gold prices in financial markets and the economic effects of price fluctuations, the trend of gold price changes in the national and global economy has attracted the attention of many researchers and economic analysts. Therefore, the main purpose of this study is to predict the trend of the global gold price movement. The purpose of this study was to introduce a combined model of the GARCH-Classic and the GARCH-Copula models and to compare them with the Garch family models in order to predict the global gold price trend in the period 01/04/2002 to 26/06/2018. The forecast horizons are 1, 5, 10, and 22 days. The prediction accuracy of these models has been evaluated and compared using RMSE error criterion. Results showed that in short-run prediction horizons, the normal Capula model with GARCH-t distribution and in long run prediction horizon, Capula- t model with the distribution of GARCH-t performs better than competing models. The hybrid model presented in this study has a high potential for predicting the trend of global gold price movement, so using this model for different sector investors, economic analysts, as well as country macro planners, can have valuable results.