Analysis Of Oil Price Fluctuations And Influences

Crude oil, the most important non-renewable energy, is the main input of the world industry. Therefore, crude oil is always called as the “black blood” or “black gold” of the economy and even the whole society. Oil price fluctuations have essential effects on global real economy and the effects always spread to financial markets such as stock and exchange rate markets. As the linkage between Chinese economy and global economic activity becomes closer, the dependence of Chinese economic growth on crude oil is greater over time. In this sense, the effects of oil price shocks on Chinese macroeconomy and financial markets will become more significant.More importantly, due to many factors such as financial crisis, geopolitical events and speculations in derivative markets, international oil price changes are more volatile in recent years. Under this background, people may ask whether oil prices are predictable. How to obtain more accurate forecasts of oil price volatility? How to better hedge oil price risk? What are the effects of oil shocks on financial markets? The goal of this paper is to do with these questions. In this thesis, we will address these questions in two ways. First, we will investigate the characteristic of oil price changes, including price predictability, volatility predictability and hedging. Second, we investigate the effects of oil price shocks on financial markets.In Chapter 3, we innovatively use a dynamic model averaging approach with time-varying parameter models to forecast oil prices. We find that the short-term oil price is unpredictable. However, the significant predictability of long-term oil price is also revealed. Specifically, relative to the random walk model, our MSPE reduction is as high as 50% and the success ratio is as high as 80%. The statistical test shows that the improvement of predictive ability over random walk is significant. This finding breaks the traditional wisdom that the asset prices are unpredictable.In Chapter 4, we use an innovative Markove regime switching multifractal(MSM) volatility model to capture and forecast crude oil price volatility. Based on this model, we can consider both the stylized facts of multifractality and structural break in oil volatility. The empirical findings suggest that both in-sample and out-of-sample performance of MSM is better than traditional GARCH-class models. According to the analysis of forecasting performance, we also find that the worse predictive ability of GARCH is due to overfitting caused by too many parameters.In Chapter 5, we investigate futures hedging under the framework of minimum-variance. Using 19 different popular models including the simple na?ve strategy, we find that none of the more complex models under consideration can significantly beat the na?ve one. We use an economic model allowing for the difference of demand elasticity of crude oil spot and futures to explain this finding. Furthermore, we investigate futures hedging in additional 23 markets and also find that it is very difficult to significantly outperform the na?ve strategy. Two factors, model misspecification and estimation, are used to explain this result. Our finding of the superiority of na?ve strategy puts forward a challenge to the traditional minimum-variance hedging.In Chapter 6, we analyze the impacts of oil shocks on world stock indices. A common wisdom in the literature is that oil price increases have negative influences on stock prices. Differently, we show that the responses of stock prices to oil price changes differ greatly depending on the driving forces of oil shocks and the relative position of a country in the world oil trading. We also find that oil price uncertainty can affect stock prices and stock market comovement. The findings in this chapter improve the understanding on the effects of oil shocks.In Chapter 7, we investigate the effects of oil price shocks on foreign exchange rate markets from the out-of-sample perspective. There is a consensus in the literature that it is very difficult to forecast exchange rate. In other words, the structural economic models cannot beat the random walk model. This is the seminal “Meese-Rogoff” puzzle. We innovatively use oil market information to predict exchange rate and find the significant improvement of short-term predictive ability over random walk. Moreover, the revealed predictability is robust to the different sample periods, the different benchmark models, the alternative evaluation methods and the choices of currency. This finding partly resolves the “Meese-Rogoff” puzzle.