6. Conclusion
This study suggested a new forecasting model approach to find causes of recent extreme oil price declines. The proposed models include combinations of real demand, speculative demand, and supply factors. The models were used to forecast monthly crude oil prices in three global oil markets and their out-of-sample predictive performances were evaluated. Using the forecasting outcomes, we suggested the factors that are the most important in explaining the oil price declines.
Our empirical results showed that, around the 2008 oil price decline, the real demand reduction measured by stock indices was an important factor in explaining the decline. In terms of model performance, M1 tended to outperform other models, implying that the reduced demand factor owing to the occurrence of the global financial crisis mainly led the collapse of oil prices. In particular, the S&P 500 index measured real demand factors seemed to have significantly influenced the oil price drop among other things. The 2008 financial crisis starting in the U.S. may have directly impacted the U.S. economy, which in turn lowered global oil prices. In contrast, the causes of declining prices during 2014–2016 were different from those of the first period. In terms of model performance, M4-2 tended to provide a better prediction result than other models. Compared with M4-1, this result indicates that U.S. shale oil development was more significant than the increased oil supply by OPEC in dropping oil prices. In summary, real demand reduction was the most important factor in the first oil price drop. Whereas, increasing shale oil supply as well as reductions in real demand and speculative demand played important roles in the second decline.
