ACD模型自加权LAD估计的渐近性质

Asymptotics for the self-weighted LAD estimator of ACD models

针对高频数据建模中常用的自回归条件持续期(ACD)模型,在允许误差方差无穷的条件下,构造模型参数的自加权最小一乘(SLAD)估计,并证明了该估计的相合性和渐近正态性.数值模拟显示SLAD估计比拟极大似然估计和最小一乘估计更稳健,最后将其应用于青岛海尔和宝信软件这两只股票的价格持续期建模.

The self-weighted least absolute deviation(SLAD) estimation for autoregressive conditional duration models with possibly infinite variance is proposed in this paper, and is shown to be consistent and asymptotically normal distributed. A large number of simulation studies confirm our theoretical results and suggest that the SLAD estimation is more robust compared to quasi-maximum likelihood estimation and least absolute deviation estimation. An application to two stocks’ price durations shows that the performance of SLAD estimation is more better.