能源价格波动对中国物价水平的潜在与实际影响

目的/意义原油作为重要的工业原料、大宗商品和战略储备物资，其价格波动对全球经济发展具有重大影响，对原油价格进行预测是能源经济领域的热点研究课题。对国际原油价格预测的主要方法进行系统梳理并展望未来研究方向，有助于该研究领域的纵深发展。设计/方法基于295篇国际期刊文献，对国际油价预测的发展阶段、发表期刊、研究机构等进行归纳总结，然后对近二十年油价预测领域的主要研究方法进行系统梳理和分析，最后对油价预测工作进行评述和趋势展望。结论/发现自2008年金融危机后，油价预测研究快速发展，相关文章主要发表于能源经济领域的权威期刊《Energy Economic》。现有油价预测方法主要包括计量模型、机器学习和混合预测模型。展望未来，利用高频日内交易数据，融合计量和机器学习前沿方法，对原油价格进行区间预测或概率密度预测是值得探索的方向，此外还应深化油价预测结果的实际应用价值。

Nonlinearities in the oil price-output relationship

[Purpose/Significance] As a kind of important industrial raw material, commodity and strategic reserve, crude oil price fluctuations have significant impact on the global economy. Forecasting crude oil price is a hot research topic in the field of energy economy. To promote the in-depth development of the crude oil price prediction, this study systematically reviews the crude oil price prediction methods and provides future research direction. [Design/Methodology] Firstly, based on 295 papers from international journals, this paper analyzes the development stages, published journals and research institutions, etc, of international crude oil price forecasting research. Then, it systematically combs and analyzes the main forecasting methods for crude oil price in the past two decades. Finally, this paper gives a few comments and points out trend on crude oil price forecasting. [Conclusions/Findings] Since the financial crisis in 2008, oil price forecasting research has experienced rapid development, and related papers are mainly published in the authoritative journal Energy Economics in the energy economy field. The existing oil price forecasting methods mainly include econometric approaches, machine learning and hybrid prediction models. In future, it is worth making interval or probability density forecast of the crude oil price by using high-frequency intraday data and combining it with advanced methods of econometric models and machine learning. In addition, the practical application value of oil price forecasting results should also be deepened.