STATISTICAL REVISIT TO THE MIKE-FARMER MODEL: CAN THIS MODEL CAPTURE THE STYLIZED FACTS IN REAL WORLD MARKETS?
STATISTICAL REVISIT TO THE MIKE-FARMER MODEL: CAN THIS MODEL CAPTURE THE STYLIZED FACTS IN REAL WORLD MARKETS?
复制标题
对迈克-法默模型的统计回顾：该模型能否捕捉现实世界市场中的典型事实？
DOI：
10.1142/s0218348x13500084
复制
发表时间：
2013-06-01
影响因子：
4.7
通讯作者：
Wen, Xing-Chun
中科院分区：
文献类型：
--
作者：
He, Ling-Yun;Wen, Xing-Chun
根据现有文献，Mike-Farmer（MF）模型1是基于订单驱动市场中的连续双向拍卖机制的经验构建的，该模型可以成功地捕获交易水平上股票价格的扩散行为。在我们的论文中，我们重新审视了基于MF模型的价格序列的统计特性，以澄清它是否可以重现真实的世界市场中的程式化事实。然而，去趋势波动分析（DFA）的波动率标度指数Hv为<$0. 6，可能略低于真实的市场的波动率标度指数;而Gu和Zhou提出的修正MF模型2可以提高DFA的波动率标度指数Hv为<$0. 75，更接近实证结果。最后，我们检验了另一个在真实的世界中常见的两个程式化事实：波动聚集和杠杆效应的存在性。
According to current literature, the Mike-Farmer (MF) model1 is constructed empirically based on the continuous double auction mechanism in an order-driven market, which can successfully capture the diffusive behavior of stock prices at the transaction level. In our paper, we revisit the statistical properties of the generated series of prices based on the MF model to clarify whether it can reproduce the stylized facts in real world markets. However, the Detrended Fluctuation Analysis (DFA) scaling exponent of volatility Hv ≈ 0.6, which may be slightly lower than that in real markets; while a modified version of the MF model proposed by Gu and Zhou2 can improve the DFA scaling exponent of volatility Hv ≈ 0.75, which is closer to the empirical findings. Finally, we test the existence of another commonly found two stylized facts in the real world: the volatility clustering, and leverage effect.