摘要
考虑一类"中度偏离"单位根过程,y_t=q_ny_t-1+u_t,其中qn=1+c/(k_n),k_n=o(n),c为一非零常数,{u_t}为随机扰动项序列.在允许扰动项方差无穷的条件下,构造q_n的复合分位数估计,并得到了该估计的渐近分布.最后通过数值模拟,在扰动项服从t(2)分布下,说明了该估计的稳健和有效性.
Under the mildly integrated and the mildly explosive cases, the asymptotic distributions of composite quantile estimation for moderate deviations from a unit root model with possibly infinite variance errors are obtained, respectively. Some simulation studies are also given to show that the composite quantile estimation has a good performance.
引文
[1]Giraitis L,Phillips P C B.Uniform limit theory for stationary autoregression[J].Journal of Time Series Analysis,2006,27:51-60.
[2]Phillips P C B,Magdalinos T.Limit theory for moderate deviations from a unit root[J].Journal of Econometrics,2007,136:115-130.
[3]Zhou Zhiyong,Lin Zhengyan.Quantile inference for moderate deviations from a unit root model with infinite variance[J].Journal of Korean Statistical Society,2015,44:280-294.
[4]Koenker R,Bassett G.Regression quantiles[J].Econometrica,1978,46:33-49.
[5]Zou Hui,Yuan Ming.Composite quantile regression and the oracle model selection theory[J].The Annals of Statistics,2008,36:1108-1126.
[6]Huang Saihua,Pang Tianxiao,Weng Chengguo.Limit theory for moderate deviations from a unit root under innovations with a possibly infinite variance[J].Methodology and Computing in Applied Probability,2014,16:187-206.
[7]Cs¨org?o M,Szyszkowicz B,Wang Qiying.Donsker’s theorem for self-normalized partial sums processes[J].The Annals of Probability,2003,31:1228-1240.
[8]Davis R A,Knight K,Liu Jian.M-estimaton for autoregressions with infinite variance[J].Stochastic Processes and their Applications,1992,40:145-180.