Creating dummy variable in Panel Data

[Obaid Ur Rehman]

Hi, i want to create a dummy variable (group-wise). my data ranges between 2001 to 2010. Features of the dummy variables are below:
Condition 1. the dummy variable equals one if variable AQA value for firm i in year t is non-missing or non-zero
Condition 2. in condition 1, for example AQA value is non-zero in year 2004 for firm i, then i want dummy variable for firm i equals one from year 2004 to end year 2005, 2006, 2007, 2008, 2009, and 2010.

[Carlo Lazzaro]

Obaid:
without any data excerpt (please see -dataex-) I find difficult to sketch something more than a temptative reply (by the way: my guess is that your 2 conditions may contradict each other in some instances).
Anyway, I do hope that the following toy-example can be helpful:

. set obs 4
number of observations (_N) was 0, now 4

. g id=1

. g year=2010+_n

. g AQA=0 in 1

. replace AQA=0 in 3

. replace AQA=runiform() in 2

. expand 2

. replace id=2 in 5/8

. replace AQA=runiform() in 5/8

. bysort year: gen dummy1=1 if AQA!=0 & AQA!=.

. sort id year

. list

    +-------------------------------+
     | id   year        AQA   dummy1 |
     |-------------------------------|
  1. |  1   2011          0        . |
  2. |  1   2012   .3488717        1 |
  3. |  1   2013          0        . |
  4. |  1   2014          .        . |
  5. |  2   2011   .2668857        1 |
     |-------------------------------|
  6. |  2   2012   .1366463        1 |
  7. |  2   2013   .0285569        1 |
  8. |  2   2014   .8689333        1 |
     +-------------------------------+

.

[Obaid Ur Rehman]

Thanks Carlo

The following picture show what i exactly need

the dummy variable 'AOA' capture 1 when x1 has non-zero value. Also the AOA keep one for rest of its observations for each "sn".

note: following is example data set. The original dataset is unstructured panel.

sn	year	x1	AOA
1	2002		0
1	2003	0.614976	1
1	2004	.	1
2	2002	0	0
2	2003	0	0
2	2004	.	0
3	2002	0.038193	1
3	2003	0	1
3	2004	0	1
4	2002	.	0
4	2003	.	0
4	2004	0.141908	1
5	2002	0.845105	1
5	2003	.	1
5	2004		1
6	2002	0.37749	1
6	2003	..	1
6	2004	0.83081	1

Currently I'm using following command:

gen AOA = 0
replace AOA = 1 if x1 != 0 & x1 != .

but i don't know how expand one to rest of observation for individual "sn" in a panel dataset.

[Obaid Ur Rehman]

Currently, I have the following output:

set obs 5
number of observations (_N) was 0, now 5

. gen id = 1

. set obs 10
number of observations (_N) was 5, now 10

. replace id = 2 if id == .
(5 real changes made)

. set obs 15
number of observations (_N) was 10, now 15

. replace id = 3 if id == .
(5 real changes made)

. set obs 20
number of observations (_N) was 15, now 20

. replace id = 4 if id == .
(5 real changes made)

. bysort id: gen year = 2010+_n

. gen AQA = 0

. replace AQA=runiform() in 2 if id == 1
(1 real change made)

. replace AQA=runiform() in 4 if id == 2
(0 real changes made)

. replace AQA=runiform() in 9 if id == 2
(1 real change made)

. replace AQA=runiform() in 13 if id == 3
(1 real change made)

. replace AQA=runiform() in 16 if id == 4
(1 real change made)

. gen dummy = 0

. replace dummy = 1 in 2 if id == 1
(1 real change made)

. replace dummy = 1 in 9 if id == 2
(1 real change made)

. replace dummy = 1 in 13 if id == 3
(1 real change made)

. replace dummy = 1 in 16 if id == 4
(1 real change made)

. sort id year

. list

     +------------------------------+
     | id   year        AQA   dummy |
     |------------------------------|
  1. |  1   2011          0       0 |
  2. |  1   2012   .8689333       1 |
  3. |  1   2013          0       0 |
  4. |  1   2014          0       0 |
  5. |  1   2015          0       0 |
     |------------------------------|
  6. |  2   2011          0       0 |
  7. |  2   2012          0       0 |
  8. |  2   2013          0       0 |
  9. |  2   2014   .3508549       1 |
 10. |  2   2015          0       0 |
     |------------------------------|
 11. |  3   2011          0       0 |
 12. |  3   2012          0       0 |
 13. |  3   2013   .0711051       1 |
 14. |  3   2014          0       0 |
 15. |  3   2015          0       0 |
     |------------------------------|
 16. |  4   2011    .323368       1 |
 17. |  4   2012          0       0 |
 18. |  4   2013          0       0 |
 19. |  4   2014          0       0 |
 20. |  4   2015          0       0 |
     +------------------------------+

Following is desired result for my large panel dataset. thanks

replace dummy = 1 in 3 if id == 1
(1 real change made)

. replace dummy = 1 in 4 if id == 1
(1 real change made)

. replace dummy = 1 in 5 if id == 1
(1 real change made)

. replace dummy = 1 in 10 if id == 2
(1 real change made)

. replace dummy = 1 in 14 if id == 3
(1 real change made)

. replace dummy = 1 in 15 if id == 3
(1 real change made)

. replace dummy = 1 in 17 if id == 4
(1 real change made)

. replace dummy = 1 in 18 if id == 4
(1 real change made)

. replace dummy = 1 in 19 if id == 4
(1 real change made)

. replace dummy = 1 in 20 if id == 4
(1 real change made)

. sort id year

. list

     +------------------------------+
     | id   year        AQA   dummy |
     |------------------------------|
  1. |  1   2011          0       0 |
  2. |  1   2012   .8689333       1 |
  3. |  1   2013          0       1 |
  4. |  1   2014          0       1 |
  5. |  1   2015          0       1 |
     |------------------------------|
  6. |  2   2011          0       0 |
  7. |  2   2012          0       0 |
  8. |  2   2013          0       0 |
  9. |  2   2014   .3508549       1 |
 10. |  2   2015          0       1 |
     |------------------------------|
 11. |  3   2011          0       0 |
 12. |  3   2012          0       0 |
 13. |  3   2013   .0711051       1 |
 14. |  3   2014          0       1 |
 15. |  3   2015          0       1 |
     |------------------------------|
 16. |  4   2011    .323368       1 |
 17. |  4   2012          0       1 |
 18. |  4   2013          0       1 |
 19. |  4   2014          0       1 |
 20. |  4   2015          0       1 |
     +------------------------------+