在stata怎么实现GMM？GMM在什么情况更适用？

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

在stata怎么实现GMM？

望高手指教

[此贴子已经被作者于2005-11-11 22:41:02编辑过]

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏2 回帖

关键词：Stata tata GMM Stata GMM

还是panel的GMM

xtabond2

but maybe you need an instrument variable at first. but it is hard to find a good iv yet.

i am not sure my idea is right and expecting other suggestions as well.

[此贴子已经被作者于2005-11-11 0:17:26编辑过]

GMM只是我们借以到处相关统计量的手段，灵活多变，在你没有导出具体的统计量之前，你的那个问题不好说呀。

---------------------------------

http://jinhe.xjtu.edu.cn/bbs/list.asp?boardid=5

ivreg2

我是在考虑动态panel模型时用GMM，这是参考萧政的新书

一般选IV就是选二阶被解释变量滞后？

可以参考baltige 的书

实际上，只要数据量大就可以

XTABOND or XTABOND2 could be used in dynamic panel data GMM estimation (xtabond2 is better - it can do what xtabond can do, but not the reverse). For the instruments, by default, XTABOND2 uses all lagged values (you can certainly restrict the instrument set) of dependent and independent variables in the first differenced equation, and the lagged first-differerces of the series (dependent and independent variables) dated t-1 to be used in the level equations.

By the way, the IVREG2 mentioned by maoxinshu above is to be used for cross-section estimation for which a suitable instrument is needed - the instrument is always hard to get, not the lagged vales as I mentioned above.

Hope these help.

 ivgmm0 我看到过用这个命令做GMM