连老师,您好。
我对广义最小二乘估计(GLS)和可行广义最小二乘估计(FGLS)不明白其适用条件,如适用于T》N 或者N》T,或者对N和T大小并无规定呢?
在Stata中的xtgls命令,一般要求是T》N结果才准确,特别是附加选项
panels(correlated) 时,帮助文件提到:
specifies heteroskedastic error structure with cross-sectional correlation. If p(c) is specified, you must also specify a time variable (use xtset). The results will be based on a generalized inverse of a singular matrix unless T>=m (the number of periods is greater than or equal to the number of panels).
也就是说这个命令更常用是在T》N时,这样才能保证运算中矩阵非奇异。
所以我有以上2个具体问题:
1、如果要求是T》N,在Stata中估计一般面板数据模型的xtreg命令,在估计随机效应(RE)时,其也是利用GLS进行估计的,为什么这时不要求T》N呢?
2、我看Woodridge的Econometric Analysis of Cross Section and Panel Data中
Page 281/770 10.4.3 A general FGLS Analysis
这部分内容提到了:
In fact,with large N asymptotics,the general FGLS estimator is just as efficient as the random effects estimator under assumptions RE1-RE3.
If N is not several times larger than T,an unrestricted FGLS analysis can be have poor finite sample properties because has T(T+1)/2 estimated elements.
从上面两段话可以看出,应该是说N》T时,使用FGLS方法更好。
所以我很疑惑,不知道哪里理解错误。