Sample path properties of an explosive double autoregressive model

被引:6
|
作者
Liu, Feng [1 ]
Li, Dong [2 ]
Kang, Xinmei [1 ]
机构
[1] Chongqing Univ Technol, Sch Math & Stat, Chongqing, Peoples R China
[2] Tsinghua Univ, Ctr Stat Sci, Beijing 100084, Peoples R China
关键词
Double AR model; explosion; geometric Brownian motion; Lyapunov exponent; nonstationarity; random coefficient AR model; EXPONENTIAL LIKELIHOOD ESTIMATORS; DOUBLE AR(P) MODEL; AR-ARCH MODELS; CONDITIONAL HETEROSCEDASTICITY; ASYMPTOTIC INFERENCE; AR(1) PROCESS; ERRORS; STATIONARY; BEHAVIOR; GARCH;
D O I
10.1080/07474938.2015.1092841
中图分类号
F [经济];
学科分类号
02 ;
摘要
This article studies sample path properties of an explosive double autoregressive (DAR) model. After suitable renormalization, it is shown that the sample path converges weakly to a geometric Brownian motion. This further strengthens our understanding of sample paths of nonstationary DAR processes. The obtained results can be extended to nonstationary random coefficient autoregressive (RCA) models. Simulation studies are carried out to support our results.
引用
收藏
页码:484 / 490
页数:7
相关论文
共 50 条