黃金價格與不同區域黃金ETF關聯性之研究

Abstract

本文探討不同區域的黃金 ETF是否會影響黃金價格的變動。研究樣本資料期間自2007年3月到2022年7月止，運用時間數列分析方法估計，實證研究結果顯示：（1）變數間存有一組共整合關係，黃金價格與亞洲區以及北美區的黃金ETF呈正相關，而與歐洲區以及全區的黃金ETF呈負相關；（2）考慮模型中所有的變數間會彼此相互影響的情況下，亞洲區的黃金ETF與黃金價格之間有互為領先的反饋關係。歐洲區的黃金ETF會「Granger領先」黃金價格，而北美區以及全區的黃金ETF與黃金價格則不具有領先、落後或是反饋關係的現象存在；（3）當向量誤差修正模型(VECM)受到黃金價格本身以及亞洲區的黃金ETF的衝擊時，黃金價格會呈現負向反應。向量誤差修正模型受到歐洲區、北美區以及全區的黃金ETF的衝擊時，黃金價格會呈現正向反應；（4）預測誤差變異分解的結果顯示解釋能力最大的前三個解釋變數分別是黃金價格本身、亞洲區黃金ETF、歐洲區黃金ETF。

This article investigates whether the holdings of gold ETF in various regions could explain the changes of gold price or not. The sample period is from March 2007 to July 2022. Using time series analysis methods, the major empirical results show that: (1) There is one cointegration relationship among variables. The gold price is positively related to Gold ETFs in Asia and North America, and is negatively related to Gold ETFs in Europe and Total. (2) Considering the interaction between all variables in the model, Gold ETFs in Asia has a feedback relationship with the gold price. Gold ETFs in Europe do “Granger cause” the gold price. As for Gold ETFs in North America and total, it didn’t relate to the gold price. (3) When the VECM model faces the shock from the gold price and Gold ETFs in Asia, the gold price will react negatively. However, when the model encounters the shock from Gold ETFs in Europe, North America and total, the gold price will react positively. (4) The result of the forecast error variance decomposition shows that the gold price, Gold ETFs in Europe and Asia have better explanatory power among all the other variables.