Dynamic DiD model

[ALKEBSEE RADWAN]

Hello everyone

I am facing a problem with Dynamic DiD model. The issue is

1- I am using csdid command

csdid  gb roll , ivar(code) time(year) gvar( group ) method(dripw)

gb = green inovation, roll= liquidity, code= firm ID, year= time, and group2= is the time after treated , zero before.

when I include ivar option which is firm ID here I get the following error message

Panel is not balanced
Will use observations with Pair balanced (observed at t0 and t1)
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
xxxxxxxxxxxxxx
Difference-in-difference with Multiple Time Periods

Number of obs = 0
Outcome model : least squares
Treatment model: inverse probability
------------------------------------------------------------------------------
| Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
g2016 |
t_2015_2016 | 0 (omitted)
t_2015_2017 | 0 (omitted)
t_2015_2018 | 0 (omitted)
t_2015_2019 | 0 (omitted)
t_2015_2020 | 0 (omitted)
t_2015_2021 | 0 (omitted)
t_2015_2022 | 0 (omitted)
t_2015_2023 | 0 (omitted)
-------------+----------------------------------------------------------------
g2017 |

I think it is becuase this code only works with one observation for every firm ( I am not sure 100%). So, I excluded it

the error message has gone but the outcomes are the same see below

. csdid  gb roll ,  time(year) gvar( group2 ) method(dripw)
Units always treated found. These will be ignored
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
xxxxxxxxxxxxxx
Difference-in-difference with Multiple Time Periods

                                                             Number of obs = 0
Outcome model  : least squares
Treatment model: inverse probability
------------------------------------------------------------------------------
             | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
g2016        |
 t_2015_2016 |          0  (omitted)
 t_2015_2017 |          0  (omitted)
 t_2015_2018 |          0  (omitted)
 t_2015_2019 |          0  (omitted)
 t_2015_2020 |          0  (omitted)
 t_2015_2021 |          0  (omitted)
 t_2015_2022 |          0  (omitted)
 t_2015_2023 |          0  (omitted)
-------------+----------------------------------------------------------------
g2017        |
 t_2015_2016 |          0  (omitted)
 t_2016_2017 |          0  (omitted)
 t_2016_2018 |          0  (omitted)
 t_2016_2019 |          0  (omitted)
 t_2016_2020 |          0  (omitted)
 t_2016_2021 |          0  (omitted)
 t_2016_2022 |          0  (omitted)
 t_2016_2023 |          0  (omitted)
-------------+----------------------------------------------------------------
g2018        |
 t_2015_2016 |          0  (omitted)
 t_2016_2017 |          0  (omitted)
 t_2017_2018 |          0  (omitted)
 t_2017_2019 |          0  (omitted)
 t_2017_2020 |          0  (omitted)
 t_2017_2021 |          0  (omitted)
 t_2017_2022 |          0  (omitted)
 t_2017_2023 |          0  (omitted)
-------------+----------------------------------------------------------------
g2019        |
 t_2015_2016 |          0  (omitted)
 t_2016_2017 |          0  (omitted)
 t_2017_2018 |          0  (omitted)
 t_2018_2019 |          0  (omitted)
 t_2018_2020 |          0  (omitted)
 t_2018_2021 |          0  (omitted)
 t_2018_2022 |          0  (omitted)
 t_2018_2023 |          0  (omitted)
-------------+----------------------------------------------------------------
g2020        |
 t_2015_2016 |          0  (omitted)
 t_2016_2017 |          0  (omitted)
 t_2017_2018 |          0  (omitted)
 t_2018_2019 |          0  (omitted)
 t_2019_2020 |          0  (omitted)
 t_2019_2021 |          0  (omitted)
 t_2019_2022 |          0  (omitted)
 t_2019_2023 |          0  (omitted)
-------------+----------------------------------------------------------------
g2021        |
 t_2015_2016 |          0  (omitted)
 t_2016_2017 |          0  (omitted)
 t_2017_2018 |          0  (omitted)
 t_2018_2019 |          0  (omitted)
 t_2019_2020 |          0  (omitted)
 t_2020_2021 |          0  (omitted)
 t_2020_2022 |          0  (omitted)
 t_2020_2023 |          0  (omitted)
-------------+----------------------------------------------------------------
g2022        |
 t_2015_2016 |          0  (omitted)
 t_2016_2017 |          0  (omitted)
 t_2017_2018 |          0  (omitted)
 t_2018_2019 |          0  (omitted)
 t_2019_2020 |          0  (omitted)
 t_2020_2021 |          0  (omitted)
 t_2021_2022 |          0  (omitted)
 t_2021_2023 |          0  (omitted)
-------------+----------------------------------------------------------------
g2023        |
 t_2015_2016 |          0  (omitted)
 t_2016_2017 |          0  (omitted)
 t_2017_2018 |          0  (omitted)
 t_2018_2019 |          0  (omitted)
 t_2019_2020 |          0  (omitted)
 t_2020_2021 |          0  (omitted)
 t_2021_2022 |          0  (omitted)
 t_2022_2023 |          0  (omitted)
------------------------------------------------------------------------------
Control: Never Treated

Here is a sample of the data

input double(code year) float(gb roll group2)
 2 2015 0 .0015    0
 2 2016 0 .0017    0
 2 2017 0 .0045    0
 2 2018 0  .004    0
 2 2019 0 .0024    0
 2 2020 0 .0016    0
 2 2021 0 .0021    0
 2 2022 3 .0026 2022
 2 2023 0 .0033    0
 4 2015 0 .0546    0
 4 2016 0 .0868    0
 4 2017 0 .0925    0
 4 2018 0 .2977    0
 4 2019 0 .2918    0
 4 2020 0 .0543    0
 4 2021 0 .0697    0
 4 2022 0 .0937    0
 4 2023 0 .0454    0
 5 2015 0 .0126    0
 5 2016 0 .0344    0
 5 2017 0  .068    0
 5 2018 0 .2144    0
 5 2019 0 .0793    0
 5 2020 0 .1106    0
 5 2021 0 .1476    0
 5 2022 0 .2271    0
 6 2015 0 .0096    0
 6 2016 0 .0205    0
 6 2017 0 .0166    0
 6 2018 0  .049    0
 6 2019 0  .051    0
 6 2020 0 .0244    0
 6 2021 0 .0751    0
 6 2022 0 .0276    0
 6 2023 0 .0211    0
 7 2015 0 .0365    0
 7 2016 0 .0286    0
 7 2017 0 .0754    0
 7 2018 0 .0655    0
 7 2019 0  .075    0
 7 2020 0 .0685    0
 7 2021 0 .0952    0
 7 2022 0 .0558    0
 7 2023 0 .1564    0
 8 2015 0 .0274    0
 8 2016 0 .0362    0
 8 2017 0 .0342    0
 8 2018 0 .0388    0
 8 2019 0 .0335    0
 8 2020 0 .0417    0
 8 2021 0 .0406    0
 8 2022 0 .0539    0
 8 2023 0 .0397    0
 9 2015 0 .0055    0
 9 2016 0 .0087    0
 9 2017 0 .0126    0
 9 2018 0 .0503    0
 9 2019 0 .0269    0
 9 2020 0 .0096    0
 9 2021 0 .0055    0
 9 2022 0 .0066    0
 9 2023 0 .0127    0
10 2015 0  .027    0
10 2016 0  .054    0
10 2017 0 .0507    0
10 2018 0 .1357    0
10 2019 0 .2967    0
10 2020 0 .1918    0
10 2021 0 .1192    0
10 2022 0 .0898    0
10 2023 0 .0459    0
11 2015 0 .0265    0
11 2016 0 .0267    0
11 2017 0 .0256    0
11 2018 0 .0913    0
11 2019 0 .0961    0
11 2020 0 .0263    0
11 2021 0 .0541    0
11 2022 0 .0512    0
11 2023 0 .0453    0
12 2015 0 .0119    0
12 2016 0 .0195    0
12 2017 0  .019    0
12 2018 0  .056    0
12 2019 0 .0588    0
12 2020 0 .0204    0
12 2021 0 .0122    0
12 2022 0 .0175    0
12 2023 0 .0202    0
14 2015 0 .0363    0
14 2016 0 .0651    0
14 2017 0 .0539    0
14 2018 0 .0546    0
14 2019 0 .0834    0
14 2020 0 .0896    0
14 2021 0 .2143    0
14 2022 0 .0446    0
14 2023 0 .0319    0
16 2015 0 .0106    0
16 2016 0 .0389    0
end

Can anyone help why I get zero observations

how I can fixe this issue ?
Thank you very much

[ALKEBSEE RADWAN]

One more question
I want to construct seven indicator variables Before(- 3), Before(- 2), Before(- 1), Current, After( + 1), After( + 2), and After(3 + ) for the six years around a (gb) event. Specifcally, Before(- 3)
indicates a frm-year observation from three years prior to the gb, Before(- 2) for two years before, and Before(- 1) for the year right before the gb; Current represents the gb year, After( + 1) is the year right after the gb, After( + 2) is the second year after the gb, and After(3 + ) is an indicator equal to one if a frm-year is at least three years after the gb, and zero otherwise.

[ALKEBSEE RADWAN]

Clyde Schechter
Andrew Musau

Could you please help me at least for the second question (how to create the dummies above)

Thank you

[Clyde Schechter]

As I do not use -csdid- and know little about it, I won't attempt to respond to #1.

With regard to #2, the variable names you are requesting are not legal in Stata. Instead this code will give you variable names that are legal and somewhat resemble what you asked for.

xtset code year
by code (year): egen gb_event_year = max(cond(gb > 0, year, .))
gen years_from_gb_event = year - gb_event_year
forvalues i = 1/3 {
    gen byte before`i' = (years_from_gb_event == -`i')
    gen byte after`i' = (years_from_gb_event == `i')
}
gen byte current = (years_from_gb_event == 0)

[ALKEBSEE RADWAN]

As I do not use -csdid- and know little about it, I won't attempt to respond to #1.

With regard to #2, the variable names you are requesting are not legal in Stata. Instead this code will give you variable names that are legal and somewhat resemble what you asked for.

xtset code year
by code (year): egen gb_event_year = max(cond(gb > 0, year, .))
gen years_from_gb_event = year - gb_event_year
forvalues i = 1/3 {
gen byte before`i' = (years_from_gb_event == -`i')
gen byte after`i' = (years_from_gb_event == `i')
}
gen byte current = (years_from_gb_event == 0)

Dear Clyde Schechter,
Thank you very much for your kind reply

I got your point
I will use this code, if I face any issue with this code I will get back to you
once again thank you very much