staggered Difference-in-Differences analysis using the `jwdid`

hussein bataineh

Join Date: Apr 2023
Posts: 155

staggered Difference-in-Differences analysis using the `jwdid`

17 Apr 2025, 17:59

I’m aiming to examine the impact of Chief Sustainability Officer presence (`csopresence1`) on the relationship between pension freezes (`hard_final_Exact_new`) and market-adjusted returns (`market_adjusted_return_w`). I adopt a staggered difference-in-differences (DiD) framework, following a published study design. The treatment sample includes a symmetric event window covering two years before and two years after each firm's defined benefit (DB) pension freeze. Matching is implemented to improve comparability between treated and control firms.

- `hard_final_Exact_new` is a binary variable equal to 1 in the year a firm freezes its DB pension plan for all employees, and 0 otherwise.
- `csopresence1` is a binary indicator equal to 1 if the firm has a Chief Sustainability Officer in a given year, and 0 otherwise. This variable may vary over time depending on the presence or departure of a CSO.

I am running a staggered Difference-in-Differences analysis using the `jwdid` command in Stata to evaluate the impact of a hard freeze on market-adjusted returns, with CSO presence as a moderator. However, when I run:

```stata
estat event, window(-2 2)
```
I only get estimates for event years 0, 1, and 2 — but event times -2 and -1 are missing. This occurs both in the overall sample and within subgroup comparisons using `ores(csopresence1==1)` and `ores(csopresence1==0)`.

I suspect the problem lies in how I generated or restricted the sample around the freeze year. Here's a simplified version of my code pipeline for reproducibility:
---

* STEP 1: Keep only firms that report a freeze status

bysort id: gen has_pension_data = sum(inlist(hard_final_Exact_new, 0, 1))
keep if has_pension_data > 0

* STEP 2: Generate gvar (treatment year)

bysort id (year): gen freeze_year = year if hard_final_Exact_new == 1
bysort id (year): replace freeze_year = freeze_year[_n-1] if missing(freeze_year)
bysort id (year): replace freeze_year = freeze_year[_n+1] if missing(freeze_year)
gen gvar = freeze_year
drop freeze_year

bysort id: egen ever_treated = max(hard_final_Exact_new == 1)
replace gvar = . if ever_treated == 0

* STEP 3: Matching

ssc install kmatch, replace
kmatch ps ever_treated Fund_Status_w ROA_w Firm_Size_w Platn_Size_w Leverage_w, ematch(year ff_12) comsup nn(1) wor idgenerate replace
keep if !missing(_ID_1)

* STEP 4: Create event time and restrict to ±2 years
gen event_time = year - gvar
tab event_time
keep if inrange(event_time, -2, 2)

* STEP 5: Drop observations with missing covariates
```stata
foreach var in market_adjusted_return_w csopresence1 Firm_Size_w ROA_w Leverage_w Market_book_four_w Non_pension_CFO_w STD_CFO_w board_size_w GenderRatiogenderratio_w independent_percentage_w sustainability2_w SUSTAIBILITY_COMITEE_FU Fund_Status_w FUNDING_RATIO_w Platn_Size_w ff_12 year {
drop if missing(`var')
}

* STEP 6: Estimate Staggered DiD (control for industry only, no firm or year FE)

jwdid market_adjusted_return_w, gvar(gvar) tvar(year) hettype(twfe) xattvar(csopresence1) exovar(Firm_Size_w ROA_w Leverage_w Market_book_four_w Non_pension_CFO_w STD_CFO_w board_size_w GenderRatiogenderratio_w independent_percentage_w sustainability2_w SUSTAIBILITY_COMITEE_FU Fund_Status_w FUNDING_RATIO_w Platn_Size_w i.ff_12)

* STEP 7: Estimate Event-time ATTs

estat event, window(-2 2)
estat event, window(-2 2) ores(csopresence1==1)
estat event, window(-2 2) ores(csopresence1==0)

Despite this, I do not get coefficients for `event_time = -2` and `-1`. What am I missing? Could it be an issue with the `gvar` construction, trimming too early, or `jwdid` not recognizing untreated periods?

Any guidance on ensuring the pre-treatment years appear in the event study table would be appreciated.

Thanks in advance!

---

HTML Code:

  * STEP 1: Keep only firms that report a freeze status

. 
. 
. 
. bysort id: gen has_pension_data = sum(inlist(hard_final_Exact_new, 0, 1))

. 
. keep if has_pension_data > 0
(3,020 observations deleted)

. 
. 
. 
. * STEP 2: Generate gvar (treatment year)

. 
. 
. 
. bysort id (year): gen freeze_year = year if hard_final_Exact_new == 1
(5,716 missing values generated)

. 
. bysort id (year): replace freeze_year = freeze_year[_n-1] if missing(freeze_year)
(816 real changes made)

. 
. bysort id (year): replace freeze_year = freeze_year[_n+1] if missing(freeze_year)
(91 real changes made)

. 
. gen gvar = freeze_year
(4,809 missing values generated)

. 
. drop freeze_year

. 
. 
. 
. bysort id: egen ever_treated = max(hard_final_Exact_new == 1)

. 
. replace gvar = . if ever_treated == 0
(0 real changes made)

. 
. 
. 
. 
. 
. * STEP 3: Matching

. 
. 
. 
. ssc install kmatch, replace
checking kmatch consistency and verifying not already installed...
all files already exist and are up to date.

. 
. kmatch ps ever_treated Fund_Status_w ROA_w Firm_Size_w Platn_Size_w Leverage_w, ematch
> (year ff_12) comsup nn(1) wor idgenerate replace
(51 observations with PS outside common support)

Propensity-score nearest-neighbor matching (without replacement)

                                                         Number of obs = 5,109
                                                Neighbors:    min =          1
Treatment   : ever_treated = 1                                max =          1
Covariates  : Fund_Status_w ROA_w Firm_Size_w Platn_Size_w Leverage_w
Exact       : year ff_12
PS model    : logit (pr)

Matching statistics
------------------------------------------------------------------------------
           |             Matched             |            Controls           
           |       Yes         No      Total |      Used     Unused      Total
-----------+---------------------------------+--------------------------------
   Treated |      1525         95       1620 |      1525       1964       3489
 Untreated |      1525       1964       3489 |      1525         95       1620
  Combined |      3050       2059       5109 |      3050       2059       5109
------------------------------------------------------------------------------

Stored variables
Variable      Storage   Display    Value
    name         type    format    label      Variable label
----------------------------------------------------------------------------------------
_ID_1           long    %12.0g                ID of control 1

. 
. keep if !missing(_ID_1)
(2,759 observations deleted)

. 
. 
. 
. 
. 
. * STEP 4: Create event time and restrict to ±2 years

. 
. gen event_time = year - gvar
(2,175 missing values generated)

. 
. tab event_time

 event_time |      Freq.     Percent        Cum.
------------+-----------------------------------
         -1 |         79        9.03        9.03
          0 |         83        9.49       18.51
          1 |         87        9.94       28.46
          2 |         80        9.14       37.60
          3 |         74        8.46       46.06
          4 |         68        7.77       53.83
          5 |         64        7.31       61.14
          6 |         59        6.74       67.89
          7 |         55        6.29       74.17
          8 |         53        6.06       80.23
          9 |         42        4.80       85.03
         10 |         36        4.11       89.14
         11 |         29        3.31       92.46
         12 |         26        2.97       95.43
         13 |         19        2.17       97.60
         14 |         10        1.14       98.74
         15 |          4        0.46       99.20
         16 |          4        0.46       99.66
         17 |          2        0.23       99.89
         18 |          1        0.11      100.00
------------+-----------------------------------
      Total |        875      100.00

. 
. keep if inrange(event_time, -2, 2)
(2,721 observations deleted)

. 
. 
. 
. 
. 
. 
. 
. 
. 
. * STEP 5: Drop observations with missing covariates

. 
. ```stata
` is not a valid command name
r(199);

. 
. foreach var in market_adjusted_return_w csopresence1 Firm_Size_w ROA_w Leverage_w Mark
> et_book_four_w Non_pension_CFO_w STD_CFO_w board_size_w GenderRatiogenderratio_w indep
> endent_percentage_w sustainability2_w SUSTAIBILITY_COMITEE_FU Fund_Status_w FUNDING_RA
> TIO_w Platn_Size_w ff_12 year {
  2. 
.     drop if missing(`var')
  3. 
. }
(58 observations deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)
(2 observations deleted)
(1 observation deleted)
(0 observations deleted)
(4 observations deleted)
(0 observations deleted)
(19 observations deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)

. 
. 
. 
. 
. 
. * STEP 6: Estimate Staggered DiD (control for industry only, no firm or year FE)

. 
. 
. 
. jwdid market_adjusted_return_w, gvar(gvar) tvar(year) hettype(twfe) xattvar(csopresenc
> e1) exovar(Firm_Size_w ROA_w Leverage_w Market_book_four_w Non_pension_CFO_w STD_CFO_w
>  board_size_w GenderRatiogenderratio_w independent_percentage_w sustainability2_w SUST
> AIBILITY_COMITEE_FU Fund_Status_w FUNDING_RATIO_w Platn_Size_w i.ff_12)
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 11 iterations)

HDFE Linear regression                            Number of obs   =        231
Absorbing 2 HDFE groups                           F(  27,    175) =       1.62
                                                  Prob > F        =     0.0357
                                                  R-squared       =     0.2951
                                                  Adj R-squared   =     0.0735
                                                  Within R-sq.    =     0.1995
                                                  Root MSE        =     0.2159

---------------------------------------------------------------------------------------
market_adjusted_ret~w | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
----------------------+----------------------------------------------------------------
    __post__#c.__tr__ |
            Post-Trt  |   .0218056   .0564926     0.39   0.700     -.089689    .1333001
                      |
    __post__#c.__tr__#|
    c._x_csopresence1 |
            Post-Trt  |   .1800449   .0437251     4.12   0.000     .0937485    .2663413
                      |
          Firm_Size_w |  -.0333592   .0343346    -0.97   0.333    -.1011224    .0344039
                ROA_w |   .5523711   .4756693     1.16   0.247    -.3864159    1.491158
           Leverage_w |  -.2668604   .1902003    -1.40   0.162    -.6422422    .1085213
   Market_book_four_w |   -.001762   .0019322    -0.91   0.363    -.0055754    .0020513
    Non_pension_CFO_w |   .5187462   .5404702     0.96   0.338    -.5479326    1.585425
            STD_CFO_w |   1.842451   1.379336     1.34   0.183    -.8798238    4.564727
         board_size_w |   .0038435   .0109931     0.35   0.727    -.0178525    .0255396
GenderRatiogenderra~w |   .3756645   .2342125     1.60   0.111    -.0865802    .8379092
independent_percent~w |   .2041363   .3158852     0.65   0.519    -.4192986    .8275712
    sustainability2_w |   -.000783   .0012713    -0.62   0.539     -.003292     .001726
SUSTAIBILITY_COMITE~U |    .000147    .043223     0.00   0.997    -.0851586    .0854525
        Fund_Status_w |   .5561143   .8662319     0.64   0.522    -1.153492     2.26572
      FUNDING_RATIO_w |   .1249579   .1798537     0.69   0.488    -.2300036    .4799194
         Platn_Size_w |    .019746     .02518     0.78   0.434    -.0299495    .0694416
                      |
                ff_12 |
                   2  |  -.0047594   .1222563    -0.04   0.969     -.246046    .2365271
                   3  |   .0431572   .0990414     0.44   0.664    -.1523121    .2386265
                   4  |  -.2774121   .2604377    -1.07   0.288    -.7914151     .236591
                   5  |    .020742   .1198638     0.17   0.863    -.2158226    .2573067
                   6  |  -.0001728   .0915409    -0.00   0.998     -.180839    .1804934
                   7  |   .0761291   .2242298     0.34   0.735    -.3664137    .5186719
                   8  |   .1842103   .1038355     1.77   0.078    -.0207208    .3891414
                   9  |   .0405306   .0936152     0.43   0.666    -.1442295    .2252907
                  10  |    .027867   .0838528     0.33   0.740    -.1376259    .1933598
                  11  |   .0446025   .0995131     0.45   0.655    -.1517978    .2410029
                  12  |  -.0096438   .1109468    -0.09   0.931    -.2286098    .2093222
                      |
                _cons |  -.5591376   .4552884    -1.23   0.221      -1.4577    .3394252
---------------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
        gvar |        15           0          15     |
        year |        15           1          14     |
-----------------------------------------------------+

. 
. 
. 
. * STEP 7: Estimate Event-time ATTs

. 
. 
. 
. estat event, window(-2 2)
------------------------------------------------------------------------------
             |            Delta-method
             | Coefficient  std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
   __event__ |
          0  |   .0278923   .0564498     0.49   0.621    -.0827473     .138532
          1  |   .0238476   .0564739     0.42   0.673    -.0868393    .1345345
          2  |   .0138538   .0566067     0.24   0.807    -.0970932    .1248008
------------------------------------------------------------------------------

. 
. estat event, window(-2 2) ores(csopresence1==1)
------------------------------------------------------------------------------
             |            Delta-method
             | Coefficient  std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
   __event__ |
          0  |    .121249    .060498     2.00   0.045     .0026751    .2398228
          1  |    .121249    .060498     2.00   0.045     .0026751    .2398228
          2  |    .121249    .060498     2.00   0.045     .0026751    .2398228
------------------------------------------------------------------------------

. 
. estat event, window(-2 2) ores(csopresence1==0)
------------------------------------------------------------------------------
             |            Delta-method
             | Coefficient  std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
   __event__ |
          0  |  -.0587959     .06056    -0.97   0.332    -.1774914    .0598995
          1  |  -.0587959     .06056    -0.97   0.332    -.1774914    .0598995
          2  |  -.0587959     .06056    -0.97   0.332    -.1774914    .0598995
------------------------------------------------------------------------------

Despite this, I do not get coefficients for `event_time = -2` and `-1`. What am I missing? Could it be an issue with the `gvar` construction, trimming too early, or `jwdid` not recognizing untreated periods?

Any guidance on ensuring the pre-treatment years appear in the event study table would be appreciated.

Thanks in advance!

Tags: None

Jeff Wooldridge

Join Date: Apr 2014

Posts: 2124
#2

17 Apr 2025, 18:42

I believe you just need to add "never" at the end of the jwdid command.
Comment

hussein bataineh

Join Date: Apr 2023
Posts: 155

18 Apr 2025, 04:00

Thanks professor Jeff for your response and advice

I generate the treatment year (gvar) as the first year a firm freezes its DB plan, and define event time as year - gvar. My event window spans from -2 to +2.

However, when I run estat event, window(-2 2), Stata omits event_time == -1 and treats it as the reference category. As a result, I get coefficients for -2, 0, 1, and 2, but I want event_time = 0 to be the omitted base period instead. I need coefficients for pre-treatment years (-2, -1) and post-treatment years (1, 2), with event_time = 0 as the reference.

Despite using a balanced event window and specifying the correct gvar, the omitted period remains -1, not 0.

these are the alll codes i used

xtset id year

* STEP 1: Keep only firms with freeze data
bysort id: gen has_pension_data = sum(inlist(hard_final_Exact_new, 0, 1))
keep if has_pension_data > 0
drop has_pension_data

* STEP 2: Generate gvar = year of first pension freeze per firm
gen freeze_year = year if hard_final_Exact_new == 1
bysort id (year): egen gvar = min(cond(hard_final_Exact_new == 1, year, .))

* STEP 3: Create treatment flag and remove never-treated gvar
bysort id: egen ever_treated = max(hard_final_Exact_new == 1)
replace gvar = . if ever_treated == 0

* STEP 4: Matching using kmatch (on pre-treatment variables)
ssc install kmatch, replace
kmatch ps ever_treated Fund_Status_w ROA_w Firm_Size_w Platn_Size_w Leverage_w , ematch(year ff_12) comsup nn(1) wor idgenerate replace

* Keep only matched firms
keep if !missing(_ID_1)

* STEP 5: Generate event time based on original gvar
gen event_time = year - gvar

* STEP 6: Keep ±2 window and recenter (so 0 is baseline)
keep if inrange(event_time, -2, 2)
gen event_time_recentered = event_time - 1

* STEP 7: Drop observations with missing variables
foreach var in market_adjusted_return_w csopresence1 Firm_Size_w ROA_w Leverage_w Market_book_four_w Non_pension_CFO_w STD_CFO_w board_size_w GenderRatiogenderratio_w independent_percentage_w sustainability2_w SUSTAIBILITY_COMITEE_FU Fund_Status_w FUNDING_RATIO_w Platn_Size_w ff_12 {
drop if missing(`var')
}

* STEP 8: Run staggered DiD using jwdid (no firm FE, includes industry FE, with never-treated)
jwdid market_adjusted_return_w, gvar(gvar) tvar(year) never xattvar(csopresence1) hettype(twfe) exovar(Firm_Size_w ROA_w Leverage_w Market_book_four_w Non_pension_CFO_w STD_CFO_w board_size_w GenderRatiogenderratio_w independent_percentage_w sustainability2_w SUSTAIBILITY_COMITEE_FU Fund_Status_w FUNDING_RATIO_w Platn_Size_w i.ff_12)

* STEP 9: Estimate event-time treatment effects centered on event_time = 0
estat event, window(-2 2)
estat event, window(-2 2) ores(csopresence1==1)
estat event, window(-2 2) ores(csopresence1==0)

HTML Code:

 . xtset id year

Panel variable: id (unbalanced)
 Time variable: year, 2004 to 2022, but with gaps
         Delta: 1 unit

. 
. 
. 
. * STEP 1: Keep only firms with freeze data

. 
. bysort id: gen has_pension_data = sum(inlist(hard_final_Exact_new, 0, 1))

. 
. keep if has_pension_data > 0
(3,020 observations deleted)

. 
. drop has_pension_data

. 
. 
. 
. * STEP 2: Generate gvar = year of first pension freeze per firm

. 
. gen freeze_year = year if hard_final_Exact_new == 1
(5,716 missing values generated)

. 
. bysort id (year): egen gvar = min(cond(hard_final_Exact_new == 1, year, .))
(4,043 missing values generated)

. 
. 
. 
. * STEP 3: Create treatment flag and remove never-treated gvar

. 
. bysort id: egen ever_treated = max(hard_final_Exact_new == 1)

. 
. replace gvar = . if ever_treated == 0
(0 real changes made)

. 
. 
. 
. * STEP 4: Matching using kmatch (on pre-treatment variables)

. 
. ssc install kmatch, replace
checking kmatch consistency and verifying not already installed...
all files already exist and are up to date.

. 
. kmatch ps ever_treated Fund_Status_w ROA_w Firm_Size_w Platn_Size_w Leverage_w , ematch
> (year ff_12) comsup nn(1) wor idgenerate replace
(51 observations with PS outside common support)

Propensity-score nearest-neighbor matching (without replacement)

                                                         Number of obs = 5,109
                                                Neighbors:    min =          1
Treatment   : ever_treated = 1                                max =          1
Covariates  : Fund_Status_w ROA_w Firm_Size_w Platn_Size_w Leverage_w
Exact       : year ff_12
PS model    : logit (pr)

Matching statistics
------------------------------------------------------------------------------
           |             Matched             |            Controls           
           |       Yes         No      Total |      Used     Unused      Total
-----------+---------------------------------+--------------------------------
   Treated |      1525         95       1620 |      1525       1964       3489
 Untreated |      1525       1964       3489 |      1525         95       1620
  Combined |      3050       2059       5109 |      3050       2059       5109
------------------------------------------------------------------------------

Stored variables
Variable      Storage   Display    Value
    name         type    format    label      Variable label
-----------------------------------------------------------------------------------------
_ID_1           long    %12.0g                ID of control 1

. 
. 
. 
. * Keep only matched firms

. 
. keep if !missing(_ID_1)
(2,759 observations deleted)

. 
. 
. 
. * STEP 5: Generate event time based on original gvar

. 
. gen event_time = year - gvar
(1,525 missing values generated)

. 
. 
. 
. * STEP 6: Keep ±2 window and recenter (so 0 is baseline)

. 
. keep if inrange(event_time, -2, 2)
(2,643 observations deleted)

. 
. gen event_time_recentered = event_time - 1

. * STEP 7: Drop observations with missing variables

. 
. foreach var in market_adjusted_return_w csopresence1 Firm_Size_w ROA_w Leverage_w Marke
> t_book_four_w Non_pension_CFO_w STD_CFO_w board_size_w GenderRatiogenderratio_w indepen
> dent_percentage_w sustainability2_w SUSTAIBILITY_COMITEE_FU Fund_Status_w FUNDING_RATIO
> _w Platn_Size_w ff_12 {
  2. 
.     drop if missing(`var')
  3. 
. }
(70 observations deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)
(2 observations deleted)
(1 observation deleted)
(2 observations deleted)
(4 observations deleted)
(0 observations deleted)
(23 observations deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)

. 
. 
. 
. * STEP 8: Run staggered DiD using jwdid (no firm FE, includes industry FE, with never-t
> reated)

. 
. jwdid market_adjusted_return_w, gvar(gvar) tvar(year) never xattvar(csopresence1) hetty
> pe(twfe) exovar(Firm_Size_w ROA_w Leverage_w Market_book_four_w Non_pension_CFO_w STD_C
> FO_w board_size_w GenderRatiogenderratio_w independent_percentage_w sustainability2_w S
> USTAIBILITY_COMITEE_FU Fund_Status_w FUNDING_RATIO_w Platn_Size_w i.ff_12)
WARNING: Singleton observations not dropped; statistical significance is biased (link)
(MWFE estimator converged in 10 iterations)

HDFE Linear regression                            Number of obs   =        294
Absorbing 2 HDFE groups                           F(  29,    234) =       1.57
                                                  Prob > F        =     0.0374
                                                  R-squared       =     0.2625
                                                  Adj R-squared   =     0.0765
                                                  Within R-sq.    =     0.1628
                                                  Root MSE        =     0.2188

----------------------------------------------------------------------------------------
market_adjusted_retu~w | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-----------------------+----------------------------------------------------------------
     __post__#c.__tr__ |
              Pre-Trt  |   .0060173   .0487266     0.12   0.902    -.0899815    .1020161
             Post-Trt  |   .0044865   .0556677     0.08   0.936    -.1051875    .1141605
                       |
     __post__#c.__tr__#|
     c._x_csopresence1 |
              Pre-Trt  |  -.0288511   .0734471    -0.39   0.695    -.1735531    .1158509
             Post-Trt  |   .1734737    .042552     4.08   0.000     .0896398    .2573076
                       |
           Firm_Size_w |  -.0341696   .0301904    -1.13   0.259    -.0936493      .02531
                 ROA_w |    .576091   .4393868     1.31   0.191    -.2895685    1.441751
            Leverage_w |  -.1448874   .1738411    -0.83   0.405    -.4873811    .1976063
    Market_book_four_w |   -.002368   .0016926    -1.40   0.163    -.0057028    .0009668
     Non_pension_CFO_w |    .167836   .4973129     0.34   0.736    -.8119467    1.147619
             STD_CFO_w |   1.406346   1.225945     1.15   0.252    -1.008955    3.821647
          board_size_w |    .008878   .0097473     0.91   0.363    -.0103256    .0280817
GenderRatiogenderrat~w |   .4153175   .2044771     2.03   0.043     .0124661    .8181688
independent_percenta~w |  -.1462571   .2698755    -0.54   0.588    -.6779533    .3854392
     sustainability2_w |  -.0004903   .0011047    -0.44   0.658    -.0026667     .001686
SUSTAIBILITY_COMITEE~U |  -.0162136   .0390411    -0.42   0.678    -.0931306    .0607033
         Fund_Status_w |   .2611689   .7194284     0.36   0.717    -1.156216    1.678553
       FUNDING_RATIO_w |   .0444909    .158408     0.28   0.779    -.2675971     .356579
          Platn_Size_w |   .0210667   .0224707     0.94   0.349     -.023204    .0653374
                       |
                 ff_12 |
                    2  |  -.0791235   .1038881    -0.76   0.447     -.283799     .125552
                    3  |   .1041036   .0872338     1.19   0.234    -.0677605    .2759676
                    4  |  -.1770176   .2189702    -0.81   0.420    -.6084225    .2543873
                    5  |   .0384563   .1062548     0.36   0.718     -.170882    .2477946
                    6  |  -.0065967   .0775324    -0.09   0.932    -.1593474     .146154
                    7  |   .0234437   .2129313     0.11   0.912    -.3960636     .442951
                    8  |   .1782273   .0938512     1.90   0.059    -.0066739    .3631286
                    9  |   .0196391   .0855431     0.23   0.819     -.148894    .1881721
                   10  |   .0429274   .0749074     0.57   0.567    -.1046516    .1905064
                   11  |   .0609769   .0875059     0.70   0.487     -.111423    .2333769
                   12  |   .0659639   .0963597     0.68   0.494    -.1238796    .2558074
                       |
                 _cons |  -.2710939   .3950664    -0.69   0.493    -1.049435    .5072475
----------------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
        gvar |        16           0          16     |
        year |        16           1          15     |
-----------------------------------------------------+

. 
. 
. 
. * STEP 9: Estimate event-time treatment effects centered on event_time = 0

. 
. estat event, window(-2 2)
------------------------------------------------------------------------------
             |            Delta-method
             | Coefficient  std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
   __event__ |
         -2  |   .0058625   .0487279     0.12   0.904    -.0896424    .1013674
         -1  |          0  (omitted)
          0  |   .0101638   .0556221     0.18   0.855    -.0988536    .1191811
          1  |    .006501   .0556476     0.12   0.907    -.1025662    .1155682
          2  |  -.0030507    .055782    -0.05   0.956    -.1123813      .10628
------------------------------------------------------------------------------

. 
. estat event, window(-2 2) ores(csopresence1==1)
------------------------------------------------------------------------------
             |            Delta-method
             | Coefficient  std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
   __event__ |
         -2  |  -.0136975   .0699261    -0.20   0.845    -.1507501     .123355
         -1  |          0  (omitted)
          0  |   .1016317    .059554     1.71   0.088     -.015092    .2183555
          1  |   .1016317    .059554     1.71   0.088     -.015092    .2183555
          2  |   .1016317    .059554     1.71   0.088     -.015092    .2183555
------------------------------------------------------------------------------

. 
. estat event, window(-2 2) ores(csopresence1==0)
------------------------------------------------------------------------------
             |            Delta-method
             | Coefficient  std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
   __event__ |
         -2  |   .0151535    .054008     0.28   0.779    -.0907002    .1210072
         -1  |          0  (omitted)
          0  |   -.071842   .0595295    -1.21   0.227    -.1885177    .0448337
          1  |   -.071842   .0595295    -1.21   0.227    -.1885177    .0448337
          2  |   -.071842   .0595295    -1.21   0.227    -.1885177    .0448337
-----------------------------------------------------------------------------

Announcement

staggered Difference-in-Differences analysis using the `jwdid`

Comment

Comment