Structural breaks, financial bubbles and dual long memory: An empirical analysis on the energy market

Energy has been a resource needed in every aspect of life in every period. While the world's richness in energy resources makes its use in heating, industry, transportation and trade widespread, it causes a wide range of people to be affected by the crises that will affect the prices. Therefore, the factors that can affect prices in energy markets have been a subject of interest to many researchers, investors and academics. The unequal distribution of energy resources in the world causes the supply of resources, their processing, distribution and the cost of going through many stages until they reach the end consumer. Energy prices, supply and demand, OPEC's decisions, macroeconomic indicators, etc. It has a very volatile structure for various reasons and prices are difficult to determine. For this purpose, daily data of Brent (Europe), No.2 Heating Oil (New York Harbor), Natural Gas (Henry Hub), WTI (Cushing, Oklahoma) were used to represent energy commodities. Data on energy commodities were evaluated using SADF and GSADF tests, and the existence of financial bubbles was examined. Financial bubbles were found in Brent, Natural gas and WTI data at different levels of significance. In the continuation of the analysis, it was examined whether energy commodities have long memory, different models and distributions were tried, and the FIGARCH model and Skewed Student-t distribution were found to be the most appropriate. It was concluded that Brent, Heating Oil, Natural Gas and WTI series exhibited a long memory period and Heating oil and WTI were the closest to efficiency in terms of market efficiency. Brent and natural gas were found to be very resistant to returning to the mean. Within the scope of the analysis, the VAR-EGARCH model was used to examine whether energy commodities affect each other or not, and the spread of return and volatility. It is concluded that there is a multi-return and volatility spread among commodities. In addition, in order to examine whether it is possible to eliminate the risks that energy commodities may encounter in the spot markets through future contracts, the hedge ratio was calculated and the DCC-GARCH model was applied to see the change over time. It has been determined that the natural gas series, which is one of the energy commodities, has a low correlation compared to other commodities and is suitable for portfolio diversification. It was concluded that other commodities have a high correlation and tend to move together.