The Coal Price Forecast Based On VAR Model

The coal industry is the most important basic energy industry in China, kept about 70% of the proportion in primary energy consumption. It is the pillar industry of national economy and has a pivotal strategic position. As with the raw materials property and energy property, changes in the price of coal not only directly affects the production of their own, the operational status of related industries, but also involves all aspects of the national economy, related to the macroeconomic comprehensive, coordinated, sustainable and healthy development. As the coal industry's most important indicators, coal price trend related to the survival of thousands of companies and its ups and downs. After the financial crisis, more volatility in coal prices makes people aware that predicting coal price risk effectively and properly has important practical significance.Based on the latest statistics data from 2005 to 2011, and combined with the background of macroeconomic trends in recent years, this article analyze empirically and forecast the market price of coal using VAR model based on econometrics and statistics. Innovation of this paper is confined in the shackles of the theoretical framework, starting from the spot market trading conditions. We have stressed the fundamental role of supply and demand theory, studying the long-term trend of the impact of supply and demand on coal prices; also we pay attention to short-term market reaction to the accidental conditions, measure impacts of the market liquidity and expected changes in psychological on prices, combined with significant fluctuations in the economy in recent years.The first chapter is an introduction, mainly on the purpose of this study, background and significance. Literature Review reviews the price situation of the coal market, focusing on comparing various mathematical models in the role of coal price forecasting. The second chapter establishes prediction model based on screening and analyzing various economic variables which affect the price of coal, according to economic theory causal relations and statistical significance test to find a suitable explanatory variable. The third chapter is the main body of this article. After selecting explanatory variables, this paper discusses the possibility of establishment of multiple linear regression models, after pointing out the existence of the multicollinearity problem and the sequence is not stable we gives up the use of multiple linear regression model, instead we use a longer time series of macroeconomic VAR model. After analyzing the advantages and reliability of the vector autoregressive model in forecasting economic time series, we test the stability of the system and residuals using VAR the model, then lead to cointegration model on the basis of stability of various indicators. The fourth chapter is the analysis and testing of the model. Through analysis of variance and pulse, coal prices would get great influence by its own lagged terms; and then variable from January 2010 to March 2011 VAR model would be substituted into the VAR formula, then we could obtain valid conclusions by comparing the predictions difference and trends changes between the predicted and actual values. Chapter V summarizes the full text and points out the existence of paper defects and the direction of further improvement such as missing important explanatory variables or the uncertainty of lag period and so on.