题目是这样的:
原题是:
A naive econometrican is running a regression in time series annual data. She computes a test for autocorrelation, which rejects the null hypothesis decisively. (Assuming no lagged variables.)
Consider the model set out now. However, assume in this case that cov(u,e) is not equal to zero.
(a) show that the least squares estimator b^ is biased and inconsistent.
(b) the instrumental variables in which Xt-1 is used as an instrument, obtain the formula for b~. Is is consistent and asympototically normal?
(c) demonstrate why, if the least square estimator b^ were CAN, you would prefer it to b~.
后两题我感觉能做,第一题由于题目条件看不懂,实在无法。我们上课时候一般表示是 Y=Xt' b + Ut, Ut= Ut-1 + et。。我的想法是推导biased必须要证明E(U|X)≠0,或者说Xt是内生变量,可能要用 cov(u,e) 这个条件,但我始终做不出。当然依照题意,这里不一定是AR(1)。。
悬赏价格再加一点,虽然我知道这里的高手都不缺论坛币,但我为表诚意~~~
感谢 hugebear 哥哥先 |