请选择 进入手机版 | 继续访问电脑版
楼主: zilong728
1265 1

[Panel Data专题] Panel data issue.. [推广有奖]

  • 0关注
  • 1粉丝

博士生

40%

还不是VIP/贵宾

-

威望
0
论坛币
6 个
通用积分
3.7500
学术水平
0 点
热心指数
0 点
信用等级
0 点
经验
368 点
帖子
148
精华
0
在线时间
168 小时
注册时间
2009-11-1
最后登录
2023-11-18

zilong728 在职认证  发表于 2013-12-7 02:20:34 |显示全部楼层 |坛友微信交流群

+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~ !

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
Hi, Dr. Lian,
I have a quick question regarding my research really struggling me. I watch your vedio of panel data. It's really a good tutoring.

Thanks for any suggestion in advance.

I have a fixed effect model to examine the firm's performance in a panel data set.

The dependent variables is firm's Return on asset(ROA).

ROA= f(Firm level control, state level control)

Can I put the state level control on the right hand side? I cluster the standard error by firm ID, however, if I put the state level variables on the right hand side, Do I  need to do cluster the standard error by the state.

Or this is totally wrong by put two different level control in a single model?

Thanks
Best regard.

二维码

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

关键词:panel data Panel issue Data pane regarding research question standard however

arlionn 在职认证  发表于 2013-12-11 16:12:19 |显示全部楼层 |坛友微信交流群
If the state level control variables are time invariant (such as location), then they will be dropped when you excute the -xtreg y x, fe- command. The reason is that the time invariant variables are perfect collineary with fixed effects (FE) dummies.

Otherwise, they can be added to the FE model.

Whether you specify  -vce(cluster id)- or -vce(cluster state)- is determined by your theorical analysis.
-vce(cluster id)- implies that the error terms are indenpendent for different firms, but can be correlated within firm.

使用道具

您需要登录后才可以回帖 登录 | 我要注册

本版微信群
加好友,备注jltj
拉您入交流群

京ICP备16021002-2号 京B2-20170662号 京公网安备 11010802022788号 论坛法律顾问:王进律师 知识产权保护声明   免责及隐私声明

GMT+8, 2024-3-29 02:10