Research On The Analysis And Prediction Of International Oil Price Volatility Under The Impact Of COVID-19

The oil industry has developed rapidly in various countries in the world,and plays a key role in various production processes.Therefore,changes in its prices affect the world's economy and the stability of the country at all times.From 1984 to today,the oil market has experienced four huge shocks that led to its price collapse.The final result has had a profound impact on the oil industry and global economic development.These four huge shocks before 2020 have caused many negative impacts on the economies of all countries in the world.At the same time,they have also had negative impacts on market stability and the maintenance of international relations.In the first half of 2020,the international oil price crashed again,and the US stock market broke down many times.The main reason for its occurrence is the emergence of the new crown pneumonia epidemic,which has caused severe damage to the global economy.Under the impact of this outbreak,the national oil market was sluggish,and consumer markets in many countries were shut down,resulting in a sudden drop in oil consumption and a sharp drop in international oil prices.The rapid collapse of international oil prices shows that the situation in the oil market is already very severe and supply and demand are no longer balanced.However,whether the final situation favors opportunities or risks still needs to be explored.In this paper,the spot price of Brent crude oil represents the international crude oil price as the research object.First,the research data selected during the new crown epidemic is diagnosed and processed for structural mutations,so as to eliminate the possible impact of mutation points on the results.Then analyze the volatility characteristics of the logarithmic return sequence before and after the retreat.Under the assumption of GED distribution,the ARIMA((2,4),0,0)-EGARCH model is constructed for these two sets of sequences.On this basis,the quantile regression model is further introduced.According to the constructed ARIMA((2,4),0,0)-EGARCH model and QAR-EGARCH model,short-term forecasts of the trend of international oil price returns during the epidemic period are made within the sample.Finally,MAE and RMSE are selected as the predictive performance evaluation indicators to compare and analyze the predictive results of the sequence before and after the retreat.The empirical research results show that: from the perspective of volatility characteristics,the international crude oil prices before and after the structural mutation and retreat during the new crown epidemic and the logarithmic return series(take Brent crude oil as an example)all show volatility clusters and "spikes and thick tails".At the same time,both sets of sequences have asymmetry and leverage effects.From the perspective of model application,due to the weak robustness of the ARIMA((2,4),0,0)-EGARCH model,this article will perform quantile regression and ARIMA((2,4),0,0)-EGARCH Combine,and carry out a weighted combination for certain specific quantile points.Different quantile points have different variable selection or prediction effects.Finally,the most significant quantile point is selected to predict the two sets of sequences before and after the retreat.The estimation of each quantile in quantile regression is done by multiplying different weights on the basis of all samples.From the prediction results,this article mainly uses the ARIMA-GED-EGARCH model and the QARGED-EGARCH model to analyze and predict data.For simple models,the combined model has better prediction performance.At the same time,according to the results,it can be seen that the fitting effect of the two models before the retreat is better than the fitting effect after the retreat,indicating that the prediction based on structural mutation is more accurate,so we cannot predict the time series.Ignore the impact of structural mutations on the data.