中文摘要：

一、引言　从现存文献可以看出,综列（panel data,或译为面版数据）单位根检验已成为近期计量经济理论与应用研究的前沿领域之一.我们知道,标准的非平稳检验是针对总量时间序列的单位根检验,而综列单位根检验则是将多个观测对象即横截面单元组合而形成面版数据的综列单位根检验,由此导致了实现这种检验方法论的特有困难：如横截面单元是否相关、同质或异质性、N和T趋于无穷大的次序等,均会影响到检验统计量的构造和渐近分布.

英文摘要：

Choi（2001 ） proposed the combining p-value tests for panel data. This paper modify the test to allow for the cross-sectional dependence with autoregressive errors and replace Choi＇s DF-GLS with ADF when DGP include an intercept or/and linear time trend. The Monte-Carlo simulation shows that the empirical size of the extended tests is very close to the nominal size of the asymptotic distributions, the power of our tests is very high, such simulation results show that our modification is feasible. Appling our tests to Chinese securities market gives the results that dependent panel price indexes of the markets is a panel unit mot process, this conclusion implies that the Chinese securities market is general weak efficiency.