波動與外溢效應之關聯性分析： 以鋼鐵、煤炭及天然氣指數型基金為例

The Volatility and Spillover Effects for The Correlation Analysis: The Study of Steel, Coal and Natural Gas Exchange-Traded Funds

馬韶鴻(Shao-Hong Ma)

指導教授 ： 陳若暉

中原大學/商學院/企業管理研究所/碩士(2014年)

https://doi.org/10.6840/cycu201400587

摘要

過往研究報告多數是探討食品跟石油或股票指數跟期貨間的關係，少有探討鋼鐵、煤炭跟天然氣ETF波動性變化及傳染效應。本研究以波動傳遞觀點開始，利用多變量GARCH模型，以美國紐約證券交易所掛牌的鋼鐵、煤和天然氣共6支ETF為樣本檢測，透過多變量波動模型建構，評估ETF波動性及報酬變動異質性及序列相關影響。本研究用單根檢定(ADF)進行恆定分析，尋求最適模式。分析煤炭、天然氣與鋼鐵ETF各市場之間的短期和長期報酬關係，使用對角線Baba, Engle, Kraft and Kroner (BEKK)、固定條件相關性Constant Conditional Correlation (CCC)、和動態條件相關，Dynamic Conditional Correlation (DCC)，並比較各模型差異的性質，以確定鋼鐵、煤炭跟天然氣ETF最適合模型及完整分析。
歸納本研究實證分析得知:
利用Akaike information criterion (AIC)最小值，檢測自回歸條件異方差的ARCH與GARCH存在效應，鋼鐵、煤炭及天然氣ETF的確存在價格相關及波動傳染效應，反映彼此間互動關係。短期或許會受到不同市場供需影響，長期而言具有穩定均衡關係。資產報酬波動會隨時空環境改變而產生不同變化，指數型基金Exchange-Traded Funds( ETF)在不同市場或資產下波動性會隨時間環境具有群聚效應。另利用多變量MGARCH其中三種模型CCC、DCC、BEKK模型檢測。發現固定條件相關(CCC)模型比動態條件相關(DCC)的配適度佳，在CCC模型中發現鋼鐵及煤炭ETF，大部分的ARCH(α)和所有的GARCH(β)，存有顯著之波動性效果，跨波動外溢效應影響程度亦強，且顯著。當煤炭價格上揚、供需產生變化、匯率及運費都會影響鋼鐵價格。本文針對波動關係加以分析，在不同階段時期，變數間存在波動傳遞效果，各國政府都可參考數據模型實施策略穩定物價，在投資策略上提供投資人或避險者參考。

關鍵字

鋼鐵、煤炭、天然氣、ETF、多變量GARCH

並列摘要

In the past, most of the studies were to evaluate food with,oil and stock indexes, few explored the ETF volatility and contagion effect on steel, coal and natural gas. The study began with volatility transmission point of view, it used the multivariate GARCH model to analyze changes in volatility and ETF returns in the New York Stock Exchange in the steel, coal and natural gas ETF as of 6 samples tested. Through constructing the multivariate volatility models to estimate the volatility and related impact of sequence heterogeneity, this study used a Unit Root Test (ADF) for examining a stationary analysis, seeking the optimum model.The paper user the minimum of Akaike information criterion (AIC) to detect auto-regressive conditional heteroskedasticity- ARCH and GARCH existence effect. To analyze the relationship between short-term and long-term returns of coal, natural gas and for steel ETF among various markets.
This paper used diagonal Baba, Engle, Kraft and Kroner (BEKK), Constant Conditional Correlation (CCC), and Dynamic Conditional Correlation (DCC), and compared the differences between the various models to completely determine the most appropriate model steel, coal, natural gas ETFs.
This study found the volatility contagion effect on steel, coal and natural gas prices related ETF. It reflected the interaction among them. Perhaps the short-term market will be differently affected by the influence of supply and demand. It has a stable equilibrium relationship in the long run. Asset return volatility follows the change of time and environment,while Exchange-Traded Funds (ETF) will have the clustering of volatility in different markets or assets following the change of time or environment.
Using three of multivariate MGARCH model, that is CCC, DCC,and BEKK model. This study found that the goodness of fit for the constant conditional correlation (CCC) model was better than the dynamic conditional correlation (DCC). We also found in the CCC model of steel and coal ETF,while most of the ARCH (α) and all the GARCH (β) are strong and significant fluctuations in the effect of cross-volatility spillovers.
During the 2007-2008 financial crisis, due to a substantial rise in oil prices, it enlarged volatility of the raw material prices, investors and consumers face the risk of price fluctuations. Therefore, the results the relation between various volatility to determine whether there exist the effect of volatility transmission, cam be as a reference for investors or hedgers to make their investment strategy.

並列關鍵字

Natural Gas ； multivariate GARCH ； Iron ； ETF ； Coal

參考文獻

1.林筱鳳、林問一、劉亞秋(2011)，「連結台灣加權股價指數、美國道瓊工業指數及上海综合指數進行連動分析研究」，智慧科技與應用統計學報，第9卷第1期 ，頁61-79。

4.呂安惇(2009)，「全球指數股票型基金之季節性外溢效果與槓桿效果分析」，中原大學碩士論文。

5.林詩彥(2006)，「鋼鐵價格決定機制及影響因素分析」，中原大學碩士論文。

1.Bali, Turan G., and Robert F. Engle (2010) "The intertemporal capital asset pricing model with dynamic conditional correlations. " Journal of Monetary Economics, 57(4), 377-390.

2.Bollerslev, T. Engle R. F. Wooldridge J. M. ( 1988) "A capital asset pricing model with time varying covariances." Journal of Political Economy, 96, 116–131.