原油具有战略和金融双重属性。原油价格波动分析一直是全球的研究热点，特别是油价大幅波动的拐点对能源金融行业的相关人员至关重要。基于此，本文对国际原油价格拐点分析及统计推断进行了探索性研究，以原油月度价格作为研究对象，集成构建PPM-KM国际原油价格拐点分析模型以适应国际原油价格拐点后验概率的测算、聚类及识别。首先，基于PPM模型测算出国际原油价格序列突变的后验概率，并结合K-Means聚类方法给出原油价格突变后验概率识别阈值，对原油价格的历史突变进行识别和分析。其次，以比较符合描述突变规律的泊松分布，对数-正态分布，幂律分布三种分布，构建国际原油价格拐点统计推断模型，对原油月度价格的突变规律进行概率模拟并比较分析。结果表明，1986年-2015年期间共发生37次显著的油价突变。在不同的时点，市场供需结构的失衡、突发地缘政治事件、美元指数、全球经济发展情况分别成为油价突变的主因。通过对油价突变点时间间隔的分布拟合，本文初步认为国际原油月度价格拐点的时间间隔服从幂律分布的假设是合理的。
The crude oil is of dual attributes, strategic and financial. Oil price fluctuations attract attention around the world, especially the oil price fluctuation change point is considered more important for energy finance industry. Based on this idea, an exploratory research direction is introduced in this paper. Monthly international crude oil prices were taken as the study objective and the PPM-KM integration model was established by extending product partition model (PPM) to adapt to measure, cluster, and identify the posterior probability of change points of international crude oil price. First, this paper measured the mutations posteriori probability of the oil price based on the PPM model in order to distinguish and analyze the mutations of the oil price, and gave the result of the tolerance threshold and mutations of commodity price combined with K-Means clustering method. Second, the Poisson distribution, power-law distribution, and logarithmic-normal distribution were used to build statistical inference model to the catastrophes description, and then corresponding probability distribution functions for simulation and analyses of the monthly crude oil price change point trends were constructed. The results showed that there were 37 significant breaking points in the period of 1986 to 2015. At different time points, the imbalanced structure of market supply and demand, unexpected events, the dollar index, the global geopolitical economic development situation the main oil mutations as the main cause of crude oil price fluctuation respectively. By fitting the distribution of the time interval of change points, this paper preliminary think the time interval of monthly international crude oil price change points obeys power-law distribution assumption is reasonable.
[1] Hou Aijun, Suardi S. A nonparametric GARCH model of crude oil price return volatility[J]. Energy Economics, 2012, 34(2):618-626.
[2] Mohammadi H, Su Lixian. International evidence on crude oil price dynamics:Applications of ARIMA-GARCH models[J]. Energy Economics, 2010, 32(5):1001-1008.
[3] 柴建,郭菊娥,龚利,等. 基于Bayesian-SV-SGT模型的原油价格‘Value at Rsik’估计[J]. 系统工程理论与实践. 2011, 31(1):8-17.
[4] Askari H, Krichene N.Oil price dynamics (2002-2006)[J]. Energy Economics, 2008, 30(5):2134-2153.
[5] Kilian L. Not all oil price shocks are alike:Disentangling demand and supply shocks in the crude oil market[J]. American Economics Review, 2009, 99(3):1053-1069.
[6] Salisu A A, Fasanya I O. Modelling oil price volatility with structural breaks[J]. Energy Policy, 2013, 52:554-562.
[7] Ji Qiang, Fan Ying. How does oil price volatility affect non-energy commodity markets?[J]. Applied Energy, 2012, 89(1):273-280.
[8] 吴刚,魏一鸣.突发事件情景下的中国战略石油储备应对策略研究[J].中国管理科学,2011, 19(2):140-146.
[9] 吴振信,薛冰,王书平. 基于VAR模型的油价波动对我国经济影响分析[J].中国管理科学,2011,19(1):21-28.
[10] 王书平,吴振信.布伦特原油价格季节性波动分析[J].中国管理科学, 2008,16(1):48-52.
[11] 张殉,余乐安,黎建强,等.重大突发事件对原油价格的影响[J].系统工程理论与实践,2009, 29(3):10-15.
[12] Chai Jian, Guo J E, Meng Lei, et al. Exploring the core factors and its dynamic effects on oil price:An application on path analysis and BVAR-TVP model[J]. Energy Policy, 2011, 39(12):8022-8036.
[13] Canova F.Detrending and turning points[J]. European Economic Review, 1994, 38(3-4):614-623.
[14] Choy J H C, Broemeling L D. Some Bayesian inferences for a changing linear model[J]. Technimetrics, 1980, 22(1):71-78.
[15] Dey D K, Ghosh S K, Chang Hong.Measuring the effect of observations using the posterior and the intrinsic bayes factors with vague prior information[J]. Sankhyā:The Indian Journal of Statistics,1997,59(3):376-391.
[16] Inclan C. Detection of multiple changes of variance using posterior odds[J]. Journal of Business and Economic Statistics, 1993, 11(3):289-300.
[17] Inclan C, Tiao G C. Use of cumulative sums of squares for retrospective detection of changes of variance[J]. Journal of the American Statistical Association, 1994, 89(427):913-923.
[18] Ferreira P E. A Bayes analysis of a switching regression model:Known number of regimes[J]. Journal of the American Statistical Association, 1975, 70(350):370-374.
[19] Hamilton J D. A new approach to the economic analysis of nonstationary time series and the business cycle[J]. Econometrica, 1989, 57(2):357-384.
[20] 柴建,朱青,张钟毓,等.国际油价识别及分析[J].中国人口资源与环境, 2014, 24(1):109-117.
[21] D'Agostino R B, Stephens M A.Goodness-of-fit techniques[M]. New York:Marcel Dekker, 1986.
[22] Barabási A L. The origin of bursts and heavy tails in human dynamics[J]. Nature, 2005, 435(7039):207-211.
[23] Vázquez A, Oliveira J G, Dezsö Z, et al. Modeling bursts and heavy tails in human dynamics[J]. Physical Review E Statistical Nonlinear & Soft Matter Physics, 2006, 73(3):80-98.
[24] Cunado J, Gracia F P D. Oil price shocks and stock market returns:Evidence for some European countries[J]. Energy Economics, 2014, 42(1):365-377.
[25] Kilian L, Park C. The impact of oil price shocks on the U.S. stock market[J]. International Economic Review, 2009, 50(4):1267-1287.
[26] Xavier G, Parameswaran G, Vasiliki P, et al. A theory of power-law distributions in financial market fluctuations[J]. Nature, 2003, 423(6937):267-270.
[27] Zhou Weixing. Universal price impact functions of individual trades in an order-driven market[J]. Quantitative Finance, 2007, 12(8):1253-1263.
[28] Loschi R H, Cruz F R B.An analysis of the influence of some prior specifications in the identification of change points via product partition model[J]. Computational Statistics & Data Analysis, 2002, 39(4):477-501.
[29] Loschi R H, Cruz F R B, Iglesias P L, et al. A Gibbs sampling scheme to the product partition model:An application to change point problems[J]. Computers & Operations Research, 2003, 30(3):463-482.
[30] 周涛,韩筱璞,闫小勇,等.人类行为时空特性的统计力学[J].电子科技大学学报,2013, 42(4):481:540.
[31] Adamic LA.Zipf,power-laws,and Pareto-AranRingtutorial[R/OL].http://www.hpl.hp.com/research/idl/papers/ranking/ranking, 2000.
[32] 张济中.分形[M].北京:清华大学出版社,1997,348.
[33] Albert R, Jeong H, Barabási A L. Internet:Dianmeter of the worldwide web[J]. Nature, 1999, 401(6749):130-131.
[34] Barabási A L, Albert R.Emergence of scaling in random networks[J]. Science, 1999, 286(5439):509-512.
[35] Clauset A, Shalizi C R. Newman M E J. Power-law distributions in empirical data[J]. Annals of Applied Statistics, 2009, 51(4):661-703.
[36] Vuong Q H. Likelihood ratio tests for model selection and non-nested hypotheses[J]. Econometrica, 1989, 57(2):307-333.
