我国猪肉价格波动的实证分析

发布时间：2018-06-10 01:57

本文选题：ARCH类模型 + 价格波动　；参考：《江西财经大学》2017年硕士论文

【摘要】：20世纪以来,随着经济的发展和科学水平的进步,我国猪肉供给量和消费量总体均持续稳定增长,中度规模及以上养殖户占比缓慢增加、猪肉的生产周期呈下降趋势。在总量持续上涨的同时,也呈现了一系列问题:生产上散养户占比历年高居不下但市场议价能力极低、抗风险性极弱;中度规模化及以上养殖户比重低,生产成本中劳动力要素比重增加迅速,新型疫病种类逐年增加、传播迅速、传染性强多呈爆发性趋势,流通环节机制不完善、疫情防治机制不完善,散养户信息交流合作机制不健全等。在这种大背景下,研究猪肉价格波动规律对政府制定宏观政策,稳定猪肉市场、提升生猪养殖者收益具有很重要的现实意义。文章主要运用了统计学、经济学和数理统计学等相关知识理论,将猪肉价格序列季节因子、不规则因子、猪肉价格趋势和循环因子以及影响猪肉价格波动等因素,运用ARCH类模型及其他数理统计模型进行了多方面的实证分析。论文对猪肉价格波动长期实证观测发现,猪肉价格序列波动具有集聚性,即大波动后往往会有小波动,小波动之后慢慢趋于更小波动直到趋于平稳,呈现出长记忆特性。传统计量模型的条件假设“同方差”用于研究猪肉价格波动序列并不合理,因此本文多次尝试用ARCH类模型来建模和分析猪肉价格波动收益序列的特征并试图找出最佳模型,继而利用相关研究结论对猪肉价格波动因素进行实证分析。而为了更好的验证和观测ARCH类模型展现出的我国猪肉价格波动的相关特征,引入季节调整和H-P滤波模型,利用了X12调整方法进行季节调整,将猪肉价格时间序列进行分解,将季节因子和不规则因子进行分离,将趋势和循环要素视为一体,进行了短期波动的分析,利用H-P滤波模型分析我国猪肉价格波动周期,波动周期平均为34.5个月,呈下降趋势,与孕育仔猪到生猪出栏所需要的时间基本保持一致。本文使用ARCH(P)多次尝试,发现猪肉价格波动收益序列具有显著的群聚特征,但并没有找出拟合效果较好的模型;GARCH(p,q)模型具有ACRH(P)高阶的性质,因此本文尝试使用GARCH(p,q)进行统计分析,依据拟合效果、AIC和SC值,发现GARCH(2,2)模型的拟合效果较好;并使用GARCH(p,q)—M,发现猪肉价格收益率并不存在高风险伴随高收益的特点;尝试TGARCH(1,1)效果不佳,而使用TGARCH(2,2)拟合效果尚可,发现猪肉价格收益序列具有杠杆效应。综合得出:猪肉价格波动特征:具有集聚性、杠杆效应,并不存在高风险高收益性;而模型拟合方面GARCH(2,2)、TARCH(2,2)尚可。本文基于供给、需求和外部冲击这三个角度从理论上分析猪肉价格波动的影响因素。猪肉价格波动是多因素共同作用的结果:生产成本是推动猪肉价格波动内在动力;散养户的机会成本是决定散养户是否退出猪肉供给市场的重要决策因素;猪肉供给量短期内剧烈变化直接影响猪肉价格;居民可支配收入和替代品价格也是影响猪肉价格波动的重要因素,生猪疫病等突发事件、养殖户对未来市场的预期也是影响猪肉价格的重要因素。结合ARCH类相关结论利用逐步回归建模、分析发现:猪肉的价格和自身滞后两阶显著相关,与滞后一阶显著正相关,与滞后两阶显著负相关;仔猪价格对猪肉价格具有直接推动作用并具有放大效果;长期回归方程中猪肉的产量与猪肉价格显著正相关(而用引用误差修正模型进行短期修正):猪肉的产量与猪肉价格呈负相关;猪肉价格与农村居民可支配收入显著相关;与替代品鸡肉显著正相关,鸡肉的替代性非常强;与替代品牛肉、羊肉不相关;与疫情显著相关;综上所述,笔者在文章最后结合本文的研究和国家政策文件以及散养户角度提出下列政策建议:规范猪肉价格波动研究,实现生猪市场管理现代化、着力提高生猪养殖户生产技术、强化生猪市场储备调节、加快建立生猪养殖生产合作社,建立公共信息公布交流平台、在市场波动规律基础上建立生猪市场波动预警机制。
[Abstract]:Since twentieth Century, with the development of economy and the progress of scientific level, the supply and consumption of pork in China have been steadily increasing, the proportion of the medium scale and above the farmers' occupation ratio is increasing slowly, and the production cycle of pork is declining. While the total amount continues to rise, there are also a series of problems. But the market price is very low, the market bargaining power is very low, the risk resistance is very weak; the proportion of the medium scale and above farmers is low, the proportion of labor factors in the production cost is increasing rapidly, the new epidemic diseases are increasing year by year, the spread is rapid, the infectious mechanism is not perfect, the mechanism of the epidemic prevention and control is imperfect, and the scattered household letter is not perfect. In this context, it is of great practical significance for the government to make macro policies, stabilize the pork market and improve the income of the pig farmers under this background. The article mainly uses the theory of statistics, economics and mathematical statistics to make the season of pork price sequence. Factors, irregular factors, pork price trend and circulation factor, and the fluctuation of pork price, the ARCH model and other mathematical statistics models are used to carry out an empirical analysis. The long-term empirical observation of pork price fluctuation shows that the volatility of pork price sequence is concentrated, that is, after large fluctuation, it often has small fluctuation. After small fluctuations, small fluctuations tend to tend to smaller fluctuations until they tend to be stable, showing a long memory characteristic. The conditional hypothesis of the traditional econometric model is not reasonable for the study of the pork price fluctuation sequence. Therefore, this paper attempts to model and analyze the characteristics of the pork price volatility income sequence with the ARCH model and try to find out many times. In order to better verify and observe the related characteristics of the fluctuation of pork price in China, the seasonal adjustment and the H-P filtering model are introduced, and the seasonal adjustment method is used to adjust the price of pork in order to better verify and observe the related characteristics of the fluctuation of pork price in China's ARCH model. In line decomposition, the seasonal factor and the irregular factor are separated, the trend and the cycle elements are considered as one. The analysis of short-term fluctuation is carried out. The fluctuation cycle of the pork price in China is analyzed by H-P filtering model. The fluctuation cycle is 34.5 months on average, which is basically consistent with the time needed to breed piglets to the pig. This paper uses ARCH (P) to find that the pork price volatility income sequence has significant clustering characteristics, but does not find a better fitting model; GARCH (P, q) model has the high order properties of ACRH (P). Therefore, this paper attempts to use GARCH (P, q) to carry out statistical analysis, according to the fitting effect, AIC and SC values, find out the model. The effect is good; and using GARCH (P, q) - M, it is found that the rate of pork price return does not have the characteristics of high risk accompanied by high income; the effect of trying TGARCH (1,1) is not good, while the use of TGARCH (2,2) fitting effect is still available, and it is found that the pork price return sequence has leverage effect. There is no high risk and high profit; model fitting is GARCH (2,2) and TARCH (2,2). This paper analyzes the influence factors of pork price fluctuation based on the three angles of supply, demand and external shock. The price fluctuation of pork is the result of the common effect of multiple factors: production cost is the internal motive force to promote the fluctuation of pork price; The opportunity cost of the farmers is an important decision factor to determine whether the farmers withdraw from the pork supply market or not; the short term violent change of the pork supply directly affects the pork price; the residents' disposable income and the replacement price are also the important factors affecting the pork price fluctuation, such as the pig epidemic disease and other sudden events, and the farmers' expectation of the future market. It is also an important factor affecting the price of pork. Using the stepwise regression modeling with the ARCH related conclusions, it is found that the price of pork is significantly related to the lagging two order, which is significantly positively correlated with the lagging first, and is negatively correlated with the lagging two order; the price of piglets has a direct driving effect on the pork price and has a magnifying effect. The output of pork in the equation was positively correlated with the pork price (and the short-term correction by the reference error correction model): the pork production was negatively correlated with the pork price; the pork price was significantly related to the disposable income of the rural residents; it was significantly positively correlated with the substitute chicken, and the replacement of chicken was very strong; it was not related to the substitute beef and mutton. In conclusion, the author put forward the following policy suggestions in the end of the article: to standardize the study of the price fluctuation of pork, to modernize the pig market management, to improve the production technology of the pig farmers, to strengthen the regulation of the pig market reserve and to speed up the establishment of the pig. Aquaculture production cooperatives, the establishment of public information disclosure and exchange platform, based on the market fluctuations, the establishment of pig market fluctuations early warning mechanism.
【学位授予单位】：江西财经大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F323.7

【参考文献】