STATA CORE CODE

most useful STATA code here.
partial contribution credit to PhD candidate: Sun Junqin, XJTU

Notes

use *…

import/export file

• 表明在哪个文件夹下做
  cd path
• 导入excel
  insheet using .csv,clear
  import excel .xlsx, firstrow clear *get firstrow as var name
• 导出excel
  export excel .xlsx //注意无长字符串
  export delimited using “xxx.csv” * export csv

变量处理

*-主要变量 winsor 集中处理 //处理无规律循环

local xx "asset revenue saleexpense"   
    foreach v of varlist `xx'{  
     winsor `v' , p(0.01) gen(`v'_x)   //按年进行winsor
     rename `v'  `v'_y
     rename  `v'_x  `v'
}  

* winsor 单个变量

winsor var, gen(var_1) p(0.01)

*append // 处理有规律循环

clear
forvalues i=1/23 {
import excel `i'.xlsx,  firstrow clear
drop in 1/2
save `i'.dta ,replace
}
use 11.dta 
forval x=12/17 {
append using `x'.dta
}

*检验独立董事是否连锁

forval i=1/1012 {
use qq.dta, clear
keep if mm==`i'
rename stkcd id
rename indepname name
drop mm
save `i'.dta, replace
merge 1:m nn using qq.dta
drop _merge
keep if name==indepname
save `i'.dta, replace
}

* 变更数据形式

reshape long inc, i(id) j(year)   // inc 为变量名

*输出结果

ereturn list r2_a

输出回归结果

outreg2 [m52 m53 m51 ] using tab011,  addstat(Adj. R2,     e(r2_a)) bdec(3) tdec(2) title(table1 full sample ) keep(     sum1 CSTR  pro  inter  cash DOMIN STATE size lev )          sortvar( CSTR pro inter) word replace tstat    

输出描述性统计

logout, save(mytable) word replace: tabstat  , statistics( n MIN  mean median  MAX sd  ) c(s) f(%10.4f)  

* spearman和pwcorr相关系数

eststo clear
sysuse auto, clear
local vlist "price mpg weight length"
    local upper
local lower `vlist'
expand 2, gen(version)
foreach v of local vlist {
    egen rank = rank(`v') if version == 1
    replace `v' = rank if version ==1
    drop rank
}
foreach v of local vlist {
    estpost correlate `v' `lower' if version == 0
    foreach m in b rho p count {
        matrix `m' = e(`m')
 }
if "`upper'"!="" {
estpost correlate `v' `upper' if version == 1
    foreach m in b rho p count {
        matrix `m' = e(`m'), `m'
     }
}
    ereturn post b
    foreach m in rho p count {
        quietly estadd matrix `m' = `m'
 }

    eststo `v', title(`v')
    local lower: list lower - v
    local upper `upper' `v'

    }
esttab using 20170220.rtf, replace rtf nonumbers mtitles     noobs not b(a2) 
logout, save(mytable1) excel replace:pwcorr_a     
logout, save(mytable2) excel replace:spearman  ,star(0.01)  
logout, save(mytable3) excel replace:spearman ,star(0.05)
logout, save(mytable4) excel replace:spearman  ,star(0.1)

** 配对

use 10.dta,clear
drop if area1==.   //和西部进行配对
bys year indus plate area1: g exp=_n if area1
bys year indus plate: egen m=max(exp)
drop if m==.
expandcl m if area1!=1,cl(year Stkcd) gen(g)
bys year Stkcd: replace exp=_n if area1!=1
bys year indus plate exp: egen r=sum(size*area1)
g q=abs((size-r)/r) if area1!=1
bys year indus plate exp: egen u=min(q) if !area1
drop if u!=q&!area1
egen mat=group(year indus plate exp)  //划分组
bysort mat:gen fre1=_N
drop if fre1==1
drop exp-fre1

* 两样本均值检验

ttest wTobinQ, by (area1)

* 两样本中位数检验

ranksum lev , by(SOE) 

* 解决自变量集体定义标签

（1）将第二行的内容转置，并作为一个变量放入同一个文件数据中，起名，比如叫“varName”； （2）运行下面的代码：
local i=0
foreach v of var _all {
local i=i'+1 local l= varname ini’
la var v' "l’”
}
drop varname

*数字文字对应表

label define repair 1 "好" 2 "较好" 3 "中" 4 "较差" 5 "差" 
rencode place, replace //文字转数字便于使用 i.place 

将数字转化为文字
tostring id ,replace 将文字转化为数字
destring id, replace

gen year=substr(time,1,4) //从前往后提取
gen id= substr(name,-7,6)  // 从最后开始提取
gen mm=1 if strmatch( Profession ,"*教授*")  
gen m333= real(regexr(m6,"分",""))  //提取22.3分此类格式数据
gen m=length(var1)  //变量的长度

*记录ttest

ttest mkv1,by(disclose)
ret list  //查看暂元
logout, save(mytable1) excel replace

• suest 检验

  suest m3 m7
  test [m3_mean]revenuechange_dummy_PCM = [m7_mean]revenuechange_dummy_PCM

*安装最新命令

ssc install ranktest, replace  //ranktest为ado名

*适合小样本

twoway(scatter  intangible  liability) (lfit intangible  liability)  //拟合线，lfit为直线，qfit为曲线
scatter intangible  liability, mlabel(n) mlabpos(6)   // 散点图

*循环加返回值

use  "C:\Users\lenovo\Desktop\1.dta" ,clear
set obs 733441
gen new=.
gen n=_n
gen t=0
forval i=1/3 {
forval j=1/3{
corr x`i' x`j'   
replace t=t+1 
replace new=r(rho) if n==t
replace new=. if `i'>=`j'&n==t
replace t=t-1 if `i'>=`j'
}
}

*便于循环中直接引用上述值

sort id1
sum id1
return list
global ff=r(max)      //便于下一步直接代入，这一步重点
forval i=1/$ff {      
qui reg earning returnwith1 if id1==`i'
replace afa=_b[_cons] if id1==`i'
replace beta=_b[returnwith1] if id1==`i'
}
esttab e1 f1 g1 h1 ,star(* 0.1 ** 0.05  *** 0.01)  //     在屏幕上整体呈现

*将变量等分

xtile income5q=income, nquantile(5)

*以逗号进行分列

split var1,parse(,)destring

• 下载文件

  copy http://www.cninfo.com.cn/finalpage/2016-08-30/1202650633.PDF F:\stata/650633.pdf ,replace

*单变量DID

diff var, t(year) p( list)

use test.dta, clear
levelsof A , clean local(stock)
foreach m in `stock' {
capture copy     "http://ichart.finance.yahoo.com/table.csv?s=`m'&a=0&b=1&c=1990&d=11&e=31&f=2014&g=w&ignore=.cs"    `m'.csv, replace
}

*排序

gsort id year -comp  
areg BTD_new accurals i.year, absorb(id)  //固定效应

*算引入后边际效应

xi:areg rd_sale_w  c.Ltrading#c.Llev Llev Ltrading Lmtb     LLNTA   Lprofit Ltangi Lrated Lindus_growth_w Lindus_RDS     i.fyear ,absorb(gvkey) cluster(gvkey) robust
est store m2
margins, dydx(*)   //所有的变量
margins, dydx(Llev) at(Ltrading==1)

clear
forvalues i=1/1208 {
use finalfinal.dta,clear
drop in `i'
reg ROA1  treatpost politicalnsoe post logasset             wleverage BTM1 SALLGR1 tenure2  Acc1-Acc12                  indus1-indus23 ,robust
gen cof1=_b[treatpost]
gen cof2=_se[treatpost]
gen t=cof1/cof2
keep in 1
gen mmmm=`i'
save `i'.dta ,replace
}
use 1.dta,clear
forval x=2/1208 {
append using `x'.dta
} 
sort t

*计算CAR

**download index *

 clear
 capture mkdir "D:/事件研究/"

 cap which cnindex
 if _rc!=0{
    ssc install cnindex
    }

 cap which cntrade
 if _rc!=0{
    ssc install cntrade
    }


 cnindex 000300, path("D:/事件研究/")  /*由于为00开头，所以选取深圳成指*/
 keep date rit 
 rename rit rmt 
 sort date
 save "D:/事件研究/index.dta",replace

*******Events 
 clear
 input stkcd str10 Eventdate
 2 "2014-04-14"
 600900 "2015-04-14"
 28 "2016-04-14"
      600000 "2016-04-14"
          601898 "2016-04-14"
              601988 "2016-04-14"
                  601666 "2016-04-14"

 end


 set matsize 500
 cap gen Eventdate1=date(Eventdate,"YMD")
 drop Eventdate
 rename Eventdate1 Eventdate
 format Eventdate %dCY-N-D

 sort stkcd Eventdate

 *******Events Matrix
 local N=_N

 mkmat stkcd Eventdate, mat(event)

 **save the event into a matrix called event 

 capture postclose event 
 postfile event id stkcd CAR1 CAR2 CAR3 CAR4 CAR5 CAR6 CAR7 CAR8 CAR9 CAR10 CAR11  CAR12  CAR13 CAR14 using  "D:/事件研究/事件研究.dta",replace

 forval i=1/`N' {
 local date = scalar(event[`i',2])
 local stkcd = scalar(event[`i',1])
 dis `stkcd'

 *生成rit（计算个股日收益率）
 clear
 cntrade  `stkcd', path("D:/事件研究/")
 keep stkcd date rit 
 destring rit, force replace 
 replace rit = rit/100
 drop in 1
 sort date 
 merge date using "D:/事件研究/index.dta", nokeep
 drop _m 
 order stkcd date rit rmt 

 preserve 
 keep if date<`date'
 gsort -date
 gen time=-_n
 keep if time>=-200

 save "D:/事件研究/pre_event.dta", replace 
 restore 

 keep if date>=`date'
 sort date 
 gen time = _n-1
 keep if time<=10
 append using  "D:/事件研究/pre_event.dta"
 qui reg rit rmt if time<=-10
 predict abr if time>=-3, r

 keep if time>=-3 

 sort time 
 gen CAR = sum(abr)

 keep stkcd time CAR

 local CAR1 = CAR[1]
 local CAR2 = CAR[2]
 local CAR3 = CAR[3]
 local CAR4 = CAR[4]
 local CAR5 = CAR[5]
 local CAR6 = CAR[6]
 local CAR7 = CAR[7]
 local CAR8 = CAR[8]
 local CAR9 = CAR[9]
 local CAR10 = CAR[10]
 local CAR11 = CAR[11]
 local CAR12 = CAR[12]
 local CAR13 = CAR[13]
 local CAR14 = CAR[14]

post event (`i') (`stkcd') (`CAR1') (`CAR2') (`CAR3') (`CAR4') (`CAR5') (`CAR6') (`CAR7') (`CAR8') ///
   (`CAR9') (`CAR10') (`CAR11') (`CAR12') (`CAR13') (`CAR14')

}
postclose event

use "D:/事件研究/事件研究.dta", clear

*分割

split v,parse(/)destring