Question

当我尝试在Stata中设置每月动态面板数据时：

clear all
drop _all
cls
scalar mnpe     = 3                     
scalar cssize   = 6                        
scalar ldataset = mnpe * cssize         

set obs `=ldataset'

//Define variables for the Timewindow to be analized
scalar tstart   = 480                   //From: 2000/January
scalar tend     = `=tstart' + `=cssize' //To  : 2000/June

//Generate id and time for the Dynamic Panel
egen id         = seq(), block(`=cssize')
egen time       = seq(), from(`=tstart') to(`=tend') 
format %tm time                         //Period definition tm: monthly

//Declare THE Dynamic Panel
tsset id time, monthly                  //Period definition   : monthly
//Generate endogenous & exogenous variables
gen y           = rnormal(5,1)
gen x1          = rnormal(10,1)

//List them up, by id
list, sepby(id)

xtline x1, tlabel(#3) overlay

我遇到了以下丢失的时间数据：

     +-----------------------------------+
     | id     time          y         x1 |
     |-----------------------------------|
  1. |  1   2000m1   3.821689   10.56352 |
  2. |  1   2000m2   4.572973   12.12826 |
  3. |  1   2000m3   5.798653   8.558586 |
  4. |  1   2000m4   8.021777   12.33319 |
  5. |  1   2000m5   3.917288   9.168543 |
  6. |  1   2000m6   4.780022      10.94 |
     |-----------------------------------|
  7. |  2   2000m1    4.70402   9.462132 |
  8. |  2   2000m2   4.164953   11.23561 |
  9. |  2   2000m3    7.09228   10.03154 |
 10. |  2   2000m4   4.429367   10.36254 |
 11. |  2   2000m5   3.171889   9.664372 |
 12. |  2   2000m7   4.455237   10.33883 |
     |-----------------------------------|
 13. |  3   2000m1   3.643499   12.20954 |
 14. |  3   2000m2   4.201683   9.507583 |
 15. |  3   2000m3   4.809181   11.44474 |
 16. |  3   2000m4    5.51936   10.25571 |
 17. |  3   2000m6   5.279757   9.676065 |
 18. |  3   2000m7   5.379889   11.40392 |
     +-----------------------------------+

如您所见，还有几个月缺失。不确定到底发生了什么，尽管我不认为这是由于夏令时造成的差异。

Answer 1

没有毛刺-标量tend中的值是错误的。

您需要从中减去1：

clear

scalar mnpe     = 3                     
scalar cssize   = 6                        
scalar ldataset = mnpe * cssize         

set obs `=ldataset'

scalar tstart   = 480                   
scalar tend     = `=tstart' + `=cssize' - 1 

egen id         = seq(), block(6)
egen time       = seq(), from(`=tstart') to(`=tend') 

format %tm time

这将产生所需的结果：

list, sepby(id)

     +-------------+
     | id     time |
     |-------------|
  1. |  1   2000m1 |
  2. |  1   2000m2 |
  3. |  1   2000m3 |
  4. |  1   2000m4 |
  5. |  1   2000m5 |
  6. |  1   2000m6 |
     |-------------|
  7. |  2   2000m1 |
  8. |  2   2000m2 |
  9. |  2   2000m3 |
 10. |  2   2000m4 |
 11. |  2   2000m5 |
 12. |  2   2000m6 |
     |-------------|
 13. |  3   2000m1 |
 14. |  3   2000m2 |
 15. |  3   2000m3 |
 16. |  3   2000m4 |
 17. |  3   2000m5 |
 18. |  3   2000m6 |
     +-------------+

在egen的seq（）函数中出现毛刺？

1 个答案: