help文件如下：

*Version October 2015
----------------------
basic syntax:

egi var, in(industry) st(state)index(#) mult(weight_var) by(year)

Description:

EGI calculates Ellision and Glaeser's index of geographical concentration of various
industries. The module computes industry and region specific share of employment and
also suitable for survey data where information up to plant level is available. EGI
also evaluate other components of Ellision and Glaeser's index like square deviation
of employment share  and Herfindhal index of plant level employment concentration.

Ellision and Glaeser's index

EGIi = (GIi - HIi)/(1-HIi)

where, GIi = Sum of squared deivation of ith industry's state employment share from
a measure Ss.

Ss = employment share ith state in the average industry.

Hi = Herfindhal style measure of the plant level concentration of employment in ith
industry.

EGI syntax return and store the Ellision and Glaeser's index for each industry variety
as a new variable. By choosing suitable index option it may also return other components.
Option


Option
-------
by(varlist): allows groups defined by `varlist' e.g. (year)

mult : allows for computing weighted index

Examples:
----------
egi worker , in(nic2) st(stcode) index(EGI) by(year) mult(mult)
egi worker , in(nic2) st(stcode) index(HHI) by(year)
egi worker , in(nic2) st(stcode) index(GI)
egi worker , in(nic2) st(stcode) index(SS)

Author:
Jay Dev Dubey
Jawaharlal Nehru University
New Delhi                                       
jaydevdubeyster@gmail.com

Acknowladgemnt:
---------------
Pritam Datta (PHFI)

References:
---------------

Ellison, Glenn, and Edward L. Glaeser. Geographic concentration in US manufacturing industries:
a dartboard approach. No. w4840. National Bureau of economic research, 1994.

Dumais, Guy, Glenn Ellison, and Edward L. Glaeser. "Geographic concentration as a dynamic process."
Review of Economics and Statistics 84.2 (2002): 193-204.

楼主你好，请问EG指数计算的stata命令中权重是什么意思啊，该怎么计算？

具体计算过程我没有研究过，权重可以不设定，不是必选项，我没有用这个option

谢谢您的回答，那直接将命令后面中括号里面的去掉就可以了吧，这样算出来就是表示产业集聚的EG指数了吧？感觉计算的过程过于简单了，结果中同一行业所有年份的EG指数数值相同，这是正常的吗？

对，权重那个option直接去掉就行，不确定的话可以打开链接看看他的命令具体的代码，计算过程与所需的公式是否符合，然后同一地区同一行业eg指数相同应该是对的，所有年份都一样感觉上好像不太对，如果你分年计算，加了by(year)的option

以上是个人理解，不一定正确。