PR for 109 xtevent 220 (#146)

* #109 Update from main (#145)

* PR for #111: drop of collinear variables (#112)

* #111 use _rmcoll to check collinearity

* #111 collinearity among the event-time dummies

* #111 consider combinations of event-time dummies

* PR for #108: warning if robust se (#117)

* #108 add warning when robust is specified

* #108 correct a typo

* #108 edit the warning message

* Add files via upload

* #108 remove comment

* Add files via upload

* #108 remove comment

* PR for #113: examples without collinearity (#115)

* #113 examples without collinearity

* #113 help files' examples without collinearity

* #113 add collinearity example to test file

* #116 specify normalization to zero (#118)

* PR for #103: emphasize impute option (#121)

* #103 emphasize impute option on homepage

* #103 github doesn't support font colors

* #103 alternative for not allowing colors

* #103 add note about impute option

* #103 update link to note about impute

* #103 correct typo

* #103 edit bullet about impute option

* #103 indicate markstat to create the html file

* #103 delete issue folder

* PR for #114: proxyiv doesnt allow varnames (#122)

* #114 loop over only numbers

* #114 case when no lead order is specified

* #114 remove comment

* #114 normalize one coeff per external instrument

* #114 edit error message

* #114 examples to check implementation

* #114 examples' log file

* #114 edit normalization messages

* #114 examples to check changes

* #114 correct examples

* #122 Erase issue folder

Co-authored-by: Jorge Eduardo Perez Perez <[email protected]>

* PR for #35: implement estimation robust treatment heterogeneity (#124)

* #35 Create experimental file

* #35 Create indicator for never-treated

* #35 Estimate basic event study

* #35 Implement Sun's suggested event study plot

* #35 compare sun_abraham vs EventStudyInteract

* #35 include imputation of the endpoints

* #35 embed EventStudyInteract into xtevent

* #35 enable postestimation options

* #35 examples to check implementation

* #35 cohort effects and save interactions

* #35 add examples

* #35 examples' log file

* #35 delete to update examples

* #35 suboption to save the interactions

* #35 update examples to check implementation

* #35 example with treatment heterogeneity data

* #35 add examples to help file

* #35 add example to the home page

* #35 add examples to test file

* #35 edit the help file

* #35 message about SA's estimator

* #35 examples and check with larger dataset

* #35 specify needed option for SA in help file

* #35 add reference's year

* #35 examples controlling for eta

* #35 description for suboption to save interactions

* #35 add avar as a dependency

* #35 compare imputation rules

* #35 change title of SA's examples in readme

* #35 add depedencies' versions

* #35 delete issue folder

* #35 cite as SA estimator instead of IW estimator

* #35 use _regress to compute cohort weights

* #35 check if avar is installed

* #35 check of existing interaction variables

* #35 wording of Sun and Abraham explanation

* #35 specify that avar is needed

* #35 typo control cohort

* #35 comment about scaling factor

* #35 typo scaling factor

* #35 remove extra space

* #35 specify opt instead of cmd

Co-authored-by: rayhuang11 <[email protected]>

* Issue 59 allow for data structures that cannot be xtseted (#127)

* #59 Create issue folder

* #59 Add dataset creation dofile to issue folder

* #59 Two step estimation

* #59 increase number of obs per state-time cell

* #59 Upload data generation example

* #59 create large version of Example31

* #59 larger version of Example31

* #59 large version of repeated cross-sectional data

* #59 estimation for different sample sizes

* #59 compare estimation time

* #59 one-step estimation with repeated-cs data

* #59 small version of repeated-cs

* #59 test the implementation

* #additions and corrections to the 1st approach

* #59 additions and corrections to the 1st approach

* #59 implement 2nd approach

* #59 delete small version dataset

* #59 upload small repeated-cs dataset

* #59 check need of touse

* #59 estimate FE with areg

* #59 edit get_unit_time_effects' help file

* #59 option to replace dataset in memory

* #59 describe load option

* #59 implement repeated cs in IV setting

* #59 delete examples old file

* #59 examples to check implementation

* #59 add examples to test file

* #59 add repeatedcs data to test folder

* #59 new data url in help file

* #59 describe repeatedcs option in help file

* #59 remove get_unit_time_effects from issue folder

* #59 remove get_unit_time_effects help file

* #59 place get_unit_time_effects in xtevent folder

* #59 make get_unit_time_effects downloadable

* #59 saving option and documentation inside code

* #59 option saving and documentation inside code

* #59 enable in static versions

* #59 update examples to check implementation

* #59 option saving in test file's examples

* #59 check saving in execution time

* #59 correct typos and wording

* #59 flexible number of leads to test

* #59 static examples

* #59 quiet generation of leads

* #127 fix eventiv brace bug

* #127 Minor changes to test file

* #127 Default to -reghdfe- for fixed effects estimation in repeatedcs

* #127 Informative varname for imputed policyvar in static estimation

* #127 Informative varname in getunittimeeffects

* #127 Minor help file changes for getunittimeeffects

* #127 erase absorbed var in getunittimeeffects

* #127 Delete issue folder

---------

Co-authored-by: Constantino-Carreto-Romero <[email protected]>

* #128 update the imputation note's link (#129)

* #130 Fix typo (#133)

* #132 Fix bug in overid testing when overidpost was active (#134)

* 131 check that the smoothest paths wald critical value is correct (#135)

* #131 Fix Wald critical value for smoothest line

* #131 fix -polyline()- call

* #137 Fix bug in static estimation without imputation (#138)

* Issue126 excluding units but eventtime is defined (#139)

* #126 verify the bug

* #126 remove sample marker

* #126 example after correction

* #126 dataset to verify the bug

* #126 marksample with missings in policyvar and eta

* #126 no marksample with missings in z and eta

* #126 with marksample and missings in z and eta

* #126 without marksample and missings in z and eta

* #126 with marksample and missings in z

* #126 with marksample and missings in z

* #126 without marksample and missings in z

* #126 test file with missing values

* #126 separate marksample for _eventgenvars

* #126 Tempvar generation for mark

* Update 126 from main (#140)

* #128 update the imputation note's link (#129)

* #130 Fix typo (#133)

* #132 Fix bug in overid testing when overidpost was active (#134)

* 131 check that the smoothest paths wald critical value is correct (#135)

* #131 Fix Wald critical value for smoothest line

* #131 fix -polyline()- call

* #137 Fix bug in static estimation without imputation (#138)

---------

Co-authored-by: Constantino Carreto Romero <[email protected]>

* #139 Add mark to xtevent.ado

---------

Co-authored-by: Constantino Carreto Romero <[email protected]>

* #139 Delete issue folder

* 136 check smoothest line solution (#142)

* #136 Create issue folder

* #136 Upload markdown

* #136 Update xteventplot.ado

* #136 Upload test log

* #136 Upload data and plots

* #136 Update from main  (#141)

* #137 Fix bug in static estimation without imputation (#138)

* Issue126 excluding units but eventtime is defined (#139)

* #126 verify the bug

* #126 remove sample marker

* #126 example after correction

* #126 dataset to verify the bug

* #126 marksample with missings in policyvar and eta

* #126 no marksample with missings in z and eta

* #126 with marksample and missings in z and eta

* #126 without marksample and missings in z and eta

* #126 with marksample and missings in z

* #126 with marksample and missings in z

* #126 without marksample and missings in z

* #126 test file with missing values

* #126 separate marksample for _eventgenvars

* #126 Tempvar generation for mark

* Update 126 from main (#140)

* #128 update the imputation note's link (#129)

* #130 Fix typo (#133)

* #132 Fix bug in overid testing when overidpost was active (#134)

* 131 check that the smoothest paths wald critical value is correct (#135)

* #131 Fix Wald critical value for smoothest line

* #131 fix -polyline()- call

* #137 Fix bug in static estimation without imputation (#138)

---------

Co-authored-by: Constantino Carreto Romero <[email protected]>

* #139 Add mark to xtevent.ado

---------

Co-authored-by: Constantino Carreto Romero <[email protected]>

* #139 Delete issue folder

---------

Co-authored-by: Constantino Carreto Romero <[email protected]>

* #136 Upload plot testing file

* #136 Do not display path when optimization fails

* #142 Delete issue folder

---------

Co-authored-by: Constantino Carreto Romero <[email protected]>

* PR for Issue120 integrate impute documentation (#144)

* #120 documentation of the impute option

* #120 Update from main (#143)

* PR for #103: emphasize impute option (#121)

* #103 emphasize impute option on homepage

* #103 github doesn't support font colors

* #103 alternative for not allowing colors

* #103 add note about impute option

* #103 update link to note about impute

* #103 correct typo

* #103 edit bullet about impute option

* #103 indicate markstat to create the html file

* #103 delete issue folder

* PR for #114: proxyiv doesnt allow varnames (#122)

* #114 loop over only numbers

* #114 case when no lead order is specified

* #114 remove comment

* #114 normalize one coeff per external instrument

* #114 edit error message

* #114 examples to check implementation

* #114 examples' log file

* #114 edit normalization messages

* #114 examples to check changes

* #114 correct examples

* #122 Erase issue folder

Co-authored-by: Jorge Eduardo Perez Perez <[email protected]>

* PR for #35: implement estimation robust treatment heterogeneity (#124)

* #35 Create experimental file

* #35 Create indicator for never-treated

* #35 Estimate basic event study

* #35 Implement Sun's suggested event study plot

* #35 compare sun_abraham vs EventStudyInteract

* #35 include imputation of the endpoints

* #35 embed EventStudyInteract into xtevent

* #35 enable postestimation options

* #35 examples to check implementation

* #35 cohort effects and save interactions

* #35 add examples

* #35 examples' log file

* #35 delete to update examples

* #35 suboption to save the interactions

* #35 update examples to check implementation

* #35 example with treatment heterogeneity data

* #35 add examples to help file

* #35 add example to the home page

* #35 add examples to test file

* #35 edit the help file

* #35 message about SA's estimator

* #35 examples and check with larger dataset

* #35 specify needed option for SA in help file

* #35 add reference's year

* #35 examples controlling for eta

* #35 description for suboption to save interactions

* #35 add avar as a dependency

* #35 compare imputation rules

* #35 change title of SA's examples in readme

* #35 add depedencies' versions

* #35 delete issue folder

* #35 cite as SA estimator instead of IW estimator

* #35 use _regress to compute cohort weights

* #35 check if avar is installed

* #35 check of existing interaction variables

* #35 wording of Sun and Abraham explanation

* #35 specify that avar is needed

* #35 typo control cohort

* #35 comment about scaling factor

* #35 typo scaling factor

* #35 remove extra space

* #35 specify opt instead of cmd

Co-authored-by: rayhuang11 <[email protected]>

* Issue 59 allow for data structures that cannot be xtseted (#127)

* #59 Create issue folder

* #59 Add dataset creation dofile to issue folder

* #59 Two step estimation

* #59 increase number of obs per state-time cell

* #59 Upload data generation example

* #59 create large version of Example31

* #59 larger version of Example31

* #59 large version of repeated cross-sectional data

* #59 estimation for different sample sizes

* #59 compare estimation time

* #59 one-step estimation with repeated-cs data

* #59 small version of repeated-cs

* #59 test the implementation

* #additions and corrections to the 1st approach

* #59 additions and corrections to the 1st approach

* #59 implement 2nd approach

* #59 delete small version dataset

* #59 upload small repeated-cs dataset

* #59 check need of touse

* #59 estimate FE with areg

* #59 edit get_unit_time_effects' help file

* #59 option to replace dataset in memory

* #59 describe load option

* #59 implement repeated cs in IV setting

* #59 delete examples old file

* #59 examples to check implementation

* #59 add examples to test file

* #59 add repeatedcs data to test folder

* #59 new data url in help file

* #59 describe repeatedcs option in help file

* #59 remove get_unit_time_effects from issue folder

* #59 remove get_unit_time_effects help file

* #59 place get_unit_time_effects in xtevent folder

* #59 make get_unit_time_effects downloadable

* #59 saving option and documentation inside code

* #59 option saving and documentation inside code

* #59 enable in static versions

* #59 update examples to check implementation

* #59 option saving in test file's examples

* #59 check saving in execution time

* #59 correct typos and wording

* #59 flexible number of leads to test

* #59 static examples

* #59 quiet generation of leads

* #127 fix eventiv brace bug

* #127 Minor changes to test file

* #127 Default to -reghdfe- for fixed effects estimation in repeatedcs

* #127 Informative varname for imputed policyvar in static estimation

* #127 Informative varname in getunittimeeffects

* #127 Minor help file changes for getunittimeeffects

* #127 erase absorbed var in getunittimeeffects

* #127 Delete issue folder

---------

Co-authored-by: Constantino-Carreto-Romero <[email protected]>

* #128 update the imputation note's link (#129)

* #130 Fix typo (#133)

* #132 Fix bug in overid testing when overidpost was active (#134)

* 131 check that the smoothest paths wald critical value is correct (#135)

* #131 Fix Wald critical value for smoothest line

* #131 fix -polyline()- call

* #137 Fix bug in static estimation without imputation (#138)

* Issue126 excluding units but eventtime is defined (#139)

* #126 verify the bug

* #126 remove sample marker

* #126 example after correction

* #126 dataset to verify the bug

* #126 marksample with missings in policyvar and eta

* #126 no marksample with missings in z and eta

* #126 with marksample and missings in z and eta

* #126 without marksample and missings in z and eta

* #126 with marksample and missings in z

* #126 with marksample and missings in z

* #126 without marksample and missings in z

* #126 test file with missing values

* #126 separate marksample for _eventgenvars

* #126 Tempvar generation for mark

* Update 126 from main (#140)

* #128 update the imputation note's link (#129)

* #130 Fix typo (#133)

* #132 Fix bug in overid testing when overidpost was active (#134)

* 131 check that the smoothest paths wald critical value is correct (#135)

* #131 Fix Wald critical value for smoothest line

* #131 fix -polyline()- call

* #137 Fix bug in static estimation without imputation (#138)

---------

Co-authored-by: Constantino Carreto Romero <[email protected]>

* #139 Add mark to xtevent.ado

---------

Co-authored-by: Constantino Carreto Romero <[email protected]>

* #139 Delete issue folder

* 136 check smoothest line solution (#142)

* #136 Create issue folder

* #136 Upload markdown

* #136 Update xteventplot.ado

* #136 Upload test log

* #136 Upload data and plots

* #136 Update from main  (#141)

* #137 Fix bug in static estimation without imputation (#138)

* Issue126 excluding units but eventtime is defined (#139)

* #126 verify the bug

* #126 remove sample marker

* #126 example after correction

* #126 dataset to verify the bug

* #126 marksample with missings in policyvar and eta

* #126 no marksample with missings in z and eta

* #126 with marksample and missings in z and eta

* #126 without marksample and missings in z and eta

* #126 with marksample and missings in z

* #126 with marksample and missings in z

* #126 without marksample and missings in z

* #126 test file with missing values

* #126 separate marksample for _eventgenvars

* #126 Tempvar generation for mark

* Update 126 from main (#140)

* #128 update the imputation note's link (#129)

* #130 Fix typo (#133)

* #132 Fix bug in overid testing when overidpost was active (#134)

* 131 check that the smoothest paths wald critical value is correct (#135)

* #131 Fix Wald critical value for smoothest line

* #131 fix -polyline()- call

* #137 Fix bug in static estimation without imputation (#138)

---------

Co-authored-by: Constantino Carreto Romero <[email protected]>

* #139 Add mark to xtevent.ado

---------

Co-authored-by: Constantino Carreto Romero <[email protected]>

* #139 Delete issue folder

---------

Co-authored-by: Constantino Carreto Romero <[email protected]>

* #136 Upload plot testing file

* #136 Do not display path when optimization fails

* #142 Delete issue folder

---------

Co-authored-by: Constantino Carreto Romero <[email protected]>

---------

Co-authored-by: Constantino-Carreto-Romero <[email protected]>
Co-authored-by: rayhuang11 <[email protected]>

* #120 Update text

* #120 edit html file about imputation

* #120 Upload html to root folder

* #120 Update link to impute documentation

* #120 Render html

* #144 Delete issue folder

---------

Co-authored-by: Constantino-Carreto-Romero <[email protected]>
Co-authored-by: rayhuang11 <[email protected]>

---------

Co-authored-by: Constantino-Carreto-Romero <[email protected]>
Co-authored-by: rayhuang11 <[email protected]>

* #109 Update home page readme

* #109 Further updates to readme

* #109 Update version statements in ado files

* #109 Initial changes to home page examples

* #109 Remaining fix for #126 due to repeated calls to _eventgenvars in _eventiv

* #109 Further changes to example in home page readme

* #109 Fix smoothest line bug when there are multiple normalizations

* #109 Update xteventplot version date

* #109 Change xteventplot error messages and optimization output

* #109 Minor typo in homepage

* #109 Updates to help files

* #109 Update acknowledgements

* #109 Update install.txt and release.txt

---------

Co-authored-by: Constantino-Carreto-Romero <[email protected]>
Co-authored-by: rayhuang11 <[email protected]>

README.md

@@ -7,14 +7,22 @@
 `xtevent` is a Stata package to estimate linear panel event-study models. It includes three commands: `xtevent` for estimation; `xteventplot` to create event-study plots and; `xteventtest` for post-estimation hypotheses testing. 
 
 
-- Last version: 2.1.1 (12aug2022)
+- Last version: 2.2.0 (15mar2023)
 - Current SSC version: 2.1.0 (1aug2022)
 -----------
 
 ### Updates
 
+* **Version 2.2.0 (15mar2023)**:
+    - Add `cohort` and `control_cohort` to obtain estimates using [Sun and Abraham's (2021)](https://www.sciencedirect.com/science/article/pii/S030440762030378X) method.
+    - Add `repeatedcs`option and `get_unit_time_effects`command to estimate event-studies in repeated cross-section settings.
+    - Add `noestimate` option to `savek()` to generate event-time dummies without estimating the regression model.
+    - Enable clustered and robust standard erros for IV estimation without `reghdfe`.
+    - Fixed bugs present in version 2.1.1.    
+    - See [here](https://github.com/JMSLab/xtevent/releases/tag/v2.2.2) for the complete update list.
+
 * **Version 2.1.1 (12aug2022)**:
-    - Fixed bugs present in xtevent 2.1.0.
+    - Fixed bugs present in version 2.1.0.
     - Updates in the help files and other documentation.
     - See [here](https://github.com/JMSLab/xtevent/releases/tag/v2.1.1)  for the complete update list.
 
@@ -28,7 +36,7 @@
     - **To produce equivalent results as with xtevent 1.0.0, where the default was to impute the endpoints, the user should use *impute(stag)*.** The **impute** option imputes missing values in the *policyvar* following different rules. For instance, specifying **impute(stag)** indicates the program to check before imputing if the *policyvar* follows staggered adoption. For a detailed explanation of the **impute** option, see this [detailed example](https://rawcdn.githack.com/JMSLab/xtevent/cf16d12f90ddf363df62c397cf0e9dc05bbd9875/impute_option_description.html).
     - The option `nonstaggered` has been depreciated. The default option is now not to impute missing values or endpoints.   You should now choose any of the imputation rules in the `impute` option. To get results using imputation consistent with staggered adoption, as in version 1.0.0 you should use `impute(stag)`.
     - Now the option `trend` allows for trend adjustment by either OLS or GMM.
-    - Fixed several bugs present in version 1.0.0
+    - Fixed several bugs present in version 1.0.0.
     - See [here](https://github.com/JMSLab/xtevent/releases/tag/v2.0.0)  for the complete update list.
 
 -----------
@@ -83,7 +91,7 @@ help xtevent
 
 ### Examples
 
-Using xtevent 2.1.1
+Using xtevent 2.2.0
 
 #### xtevent
 ```stata
@@ -94,70 +102,75 @@ webuse nlswork, clear
 by idcode (year): gen time=_n
 xtset idcode time
 
-*Generate a policy variable that follows staggered-adoption
+
+*Generate a policy variable that follows staggered adoption
 by idcode (time): gen union2=sum(union)
 replace union2=1 if union2>1 
 order time union union2, after(year)
 
-*** examples
+*** Examples
 *Estimate a basic event study with clustered standard errors 
+*Impute the policy variable assuming no unobserved changes
 xtevent ln_w age c.age#c.age ttl_exp c.ttl_exp#c.ttl_exp tenure , ///
-            pol(union) w(3) cluster(idcode) 
-            
-*Omit fixed effects
+            pol(union) w(3) cluster(idcode) impute(nuchange)
+	   
+*Omit unit and time fixed effects
 *Impute the policy variable verifying staggered adoption
 xtevent ln_w age c.age#c.age ttl_exp c.ttl_exp#c.ttl_exp tenure , ///
             pol(union2) w(3) cluster(idcode) nofe note impute(stag)
 
-*Adjust the pre-trend by estimating a linear trend by GMM
+* Bring back unit and time fixed effects
+*Adjust for a pre-trend by estimating a linear trend by GMM
 xtevent ln_w age c.age#c.age ttl_exp c.ttl_exp#c.ttl_exp tenure , ///
-            pol(union) w(3) cluster(idcode) trend(-2, method(gmm))
+      pol(union2) w(3) cluster(idcode) trend(-2, method(gmm)) ///
+	    impute(stag)
 			
-*FHS estimator with proxy variables
+*Freyaldenhoven, Hansen and Shapiro (2019) estimator with proxy variables
 xtevent ln_w age c.age#c.age ttl_exp c.ttl_exp#c.ttl_exp tenure , ///
-            pol(union) w(3) vce(cluster idcode) proxy(wks_work) 
+      pol(union) w(3) vce(cluster idcode) proxy(wks_work) ///
+	    impute(stag)
 			          
 *reghdfe and two-way clustering
 xtevent ln_w age c.age#c.age ttl_exp c.ttl_exp#c.ttl_exp tenure , ///
-            pol(union) w(3) impute(nuchange) cluster(idcode year) reghdfe ///
-            proxy(wks_work)
+            pol(union2) w(3) impute(stag) cluster(idcode year) reghdfe ///
+            proxy(wks_work) 
             
-*Sun and Abraham Estimator
-*generate the variable that indicates cohort
-gen timet=year if union==1
+*Sun and Abraham (2021) Estimator
+*Generate the variable that indicates cohort
+gen timet=year if union2==1
 by idcode: egen time_of_treat=min(timet)
-*generate the variable that indicates the control cohort. 
-*we use the never treated units as the control cohort. 
+*Generate the variable that indicates the control cohort 
+*We use the never-treated units as the control cohort 
 gen never_treat=time_of_treat==.
-*estimate the event-time coefficients with the Sun-and-Abraham Estimator.
+*estimate the event-time coefficients with the Sun-and-Abraham (2021) Estimator
 xtevent ln_w age c.age#c.age ttl_exp c.ttl_exp#c.ttl_exp tenure, ///
-            policyvar(union) window(3) impute(nuchange) vce(cluster idcode) ///
+            policyvar(union2) window(3) impute(stag) vce(cluster idcode) ///			
             reghdfe cohort(time_of_treat) control_cohort(never_treat) 
-
 ```
 
 #### xteventplot
 ```stata
-*** setup
+
+*** Setup
 webuse nlswork, clear
 * year variable has many missing observations
 * Create a time variable that ignores the gaps
 by idcode (year): gen time=_n
 xtset idcode time
 
-*** examples 
-*Basic event study with clustered standard errors. 
-*Impute policy variable without verifying staggered adoption.
+*** Examples 
+*Basic event study with clustered standard errors 
+*Impute the policy variable assuming no unobserved changes
 xtevent ln_w age c.age#c.age ttl_exp c.ttl_exp#c.ttl_exp tenure , ///
             pol(union) w(3) cluster(idcode) impute(nuchange) 
 
-* simple plot
+* Simple plot
 xteventplot
 
 *Plot smoothest path in confidence region
 xteventplot, smpath(line)
 
-*FHS estimator with proxy variables
+*Freyaldenhoven, Hansen and Shapiro (2019) estimator with proxy variables
 xtevent ln_w age c.age#c.age ttl_exp c.ttl_exp#c.ttl_exp tenure , ///
             pol(union) w(3) vce(cluster idcode) impute(nuchange) ///
             proxy(wks_work)
@@ -177,7 +190,7 @@ xtset idcode year
 
 *** examples 
 *Basic event study with clustered standard errors. 
-*Impute policy variable without verifying staggered adoption.
+*Impute the policy variable assuming no unobserved changes
 xtevent ln_w age c.age#c.age ttl_exp c.ttl_exp#c.ttl_exp tenure , ///
             pol(union) w(3) cluster(idcode) impute(nuchange) 
 

xtevent/_eventiv.ado

@@ -281,7 +281,7 @@ program define _eventiv, rclass
 
 	if "`gen'" != "nogen" {	
 		*If impute was specified, this is the second call to _eventgenvars: this time, both the policyvar and the event-time dummies will be imputed. Additional computations will happen as well  (e.g., macros, etc.).
-		_eventgenvars if `touse', panelvar(`panelvar') timevar(`timevar') policyvar(`policyvar') lwindow(`lwindow') rwindow(`rwindow') `trend' norm(`norm') impute(`impute') `repeatedcs'
+		_eventgenvars if `tousegen', panelvar(`panelvar') timevar(`timevar') policyvar(`policyvar') lwindow(`lwindow') rwindow(`rwindow') `trend' norm(`norm') impute(`impute') `repeatedcs'
 		loc included=r(included)
 		loc names=r(names)	
 		loc komittrend=r(komittrend)

xtevent/get_unit_time_effects.ado

@@ -1,4 +1,4 @@
-* get_unit_time_effects.ado 1.0.0 Nov 14 2022
+*! get_unit_time_effects.ado 2.2.0 March 15 2023
 
 version 11.2
 

xtevent/get_unit_time_effects.sthlp

@@ -1,14 +1,13 @@
 
 {smcl}
-{* *! version 1.0.0 Nov 14 2022}{...}
+{* *! version 2.2.0 Mar 15 2023}{...}
 {cmd:help get_unit_time_effects}
 {hline}
 
 {title:Title}
 
 {phang}
-{bf:get_unit_time_effects} {hline 2} Generate Group & Time effects in a Repeated Cross-Sectional Dataset
-
+{bf:get_unit_time_effects} {hline 2} Generate Group and Time Effects in a Repeated Cross-Sectional Dataset
 
 {marker syntax}{...}
 {title:Syntax}
@@ -28,7 +27,7 @@
 {synopthdr}
 {synoptline}
 {syntab:Main}
-{synopt: {opth p:anelvar(varname)}} variable that identifies the groups at which effects will be computed {p_end}
+{synopt: {opth p:anelvar(varname)}} variable that identifies the groups{p_end}
 {synopt: {opth t:imevar(varname)}} variable that identifies the time periods{p_end}
 {synopt: {opt saving(filename, [replace])}} save results to {it:filename}{p_end}
 {synopt:{opt noo:utput}} omit regression table{p_end}
@@ -45,24 +44,27 @@
 {title:Description}
 
 {pstd}
-{cmd: get_unit_time_effects} estimates group-time fixed effects in a repeated cross-sectional dataset. It produces a  Stata data file with the variables {it:panelvar}, {it:timevar}, and {it:_unittimeeffects}. The variable 
-{it:_unittimeeffects} contains the group-time effects. Hansen (2007) describes a two-step procedure to obtain the coefficient estimates of covariates that vary at the group level within a repeated cross-sectional framework. The
- two-step procedure can be used to obtain the coefficient estimates of an event-study when the data is repeated cross-sectional. {cmd:get_unit_time_effects} implements the first part of the two-step procedure. Then, 
- {cmd: xtevent} can be used for the second part of the procedure to obtain the event-study coefficient estimates. See {help xtevent}.{p_end}
+{cmd: get_unit_time_effects} estimates group-time fixed effects in a repeated cross-sectional dataset. It produces a Stata data file with the 
+variables {it:panelvar}, {it:timevar}, and {it:_unittimeeffects}. The variable {it:_unittimeeffects} contains the group-time effects. Hansen (2007)
+ describes a two-step procedure to obtain the coefficient estimates of covariates that vary at the group level within a repeated
+ cross-sectional framework. The  two-step procedure can be used to obtain the coefficient estimates of an event-study when the data
+ is repeated cross-sectional. {cmd:get_unit_time_effects} implements the first part of the two-step procedure. Then, {cmd: xtevent} can
+ be used for the second part of the procedure to obtain the event-study coefficient estimates. See {help xtevent}.{p_end}
 
 {marker options}{...}
 {title:Options}
 {synoptline}
 
 {phang}
-{opth panelvar(varname)} specifies the group variable. For the Hansen (2007), the policy variable should vary at this group level.
+{opth panelvar(varname)} specifies the group variable. For the Hansen (2007) estimator, the policy variable should vary at this group level.
 
 {phang}
 {opth timevar(varname)} specifies the time variable. 
 
 {phang}
-{opt saving(filename, [replace])} specifies the name of the Stata data file which contains the unit-time effects estimates. If {opt saving} is not specified, the file will be saved in the 
-current directory with the name {it: unit_time_effects.dta}. The suboption {it:replace} overwrites the unit-time effects file.
+{opt saving(filename, [replace])} specifies the name of the Stata data file which contains the unit-time effects estimates. If {opt saving} is
+ not specified, the file will be saved in the current directory with the name {it: unit_time_effects.dta}. The suboption {it:replace} overwrites
+ the unit-time effects file.
 
 {phang}
 {opt nooutput} omits the regression table. 
@@ -73,28 +75,47 @@ current directory with the name {it: unit_time_effects.dta}. The suboption {it:r
 {title:Examples}
 
 {hline}
-{pstd}Load the small version of the repeated cross-sectional dataset example31{p_end}
+{pstd}Load example repeated cross-sectional dataset{p_end}
 {phang2}{cmd:. use "https://github.com/JMSLab/xtevent/blob/main/test/small_repeated_cross_sectional_example31.dta?raw=true", clear}{p_end}
 {phang2}{cmd:. {stata xtset, clear}}{p_end}
 
-{pstd}Get unit-time effects and save them as a dta file with the name and directory indicated through the {bf:saving} option. Add the {bf:replace} suboption to overwrite the file.{p_end}
-{phang2}{cmd:. get_unit_time_effects y u eta, panelvar(state) timevar(t) saving("My_directory/effect_file.dta", replace)}
+{pstd}Get unit-time effects and save them as a dta file with the name and directory indicated through the {bf:saving} option. Add the
+ {bf:replace} suboption to overwrite the file.{p_end}
+{phang2}{cmd:. {stata get_unit_time_effects y u eta, panelvar(state) timevar(t) saving("effect_file.dta", replace)}}
 {p_end}
 
 {pstd}Proceed with {bf:xtevent}{p_end}
 {phang2}{cmd:. {stata "bysort state t (z): keep if _n==1"}}{p_end}
 {phang2}{cmd:. {stata "keep state t z"}}{p_end}
 
 {pstd}Merge with the file that was created with {bf:get_unit_time_effects}{p_end}
-{phang2}{cmd:. merge m:1 state t using "My_directory/effect_file.dta"}{p_end}
+{phang2}{cmd:. {stata "merge m:1 state t using effect_file.dta"}}{p_end}
 {phang2}{cmd:. {stata drop _merge}}{p_end}
 {pstd}Use {bf:xtevent} to estimate an event-study{p_end}
-{phang2}{cmd:. {stata xtevent effects, panelvar(state) t(t) policyvar(z) window(5)}}{p_end}
+{phang2}{cmd:. {stata xtevent _unittimeeffects, panelvar(state) t(t) policyvar(z) window(5)}}{p_end}
 {phang2}{cmd:. {stata xteventplot}}{p_end}
 
+{title:Authors}
+
+{pstd}Simon Freyaldenhoven, Federal Reserve Bank of Philadelphia.{p_end}
+       [email protected]
+{pstd}Christian Hansen, University of Chicago, Booth School of Business.{p_end}
+       [email protected]
+{pstd}Jorge Pérez Pérez, Banco de México.{p_end}
+       [email protected]
+{pstd}Jesse Shapiro, Harvard University and NBER.{p_end}
+       [email protected]	   
+
+{title:Support}    
+
+{pstd}For support and to report bugs please email Jorge Pérez Pérez, Banco de México.{break} 
+       [email protected]   
 
+{pstd}{cmd:xtevent} can also be found on {browse "https://github.com/JMSLab/xtevent":GitHub}.
 
 {title:References}
 
-{pstd}Hansen, C. (2007) . "Generalized Least Squares Inference in Panel and Multilevel Models with Serial Correlation and Fixed Effects" Journal of Econometrics, 140(2), 670-694. 
+{pstd}Hansen, C. (2007) . "Generalized Least Squares Inference in Panel and Multilevel Models 
+with Serial Correlation and Fixed Effects" Journal of Econometrics, 140(2), 670-694.{p_end}
 
+

xtevent/release.txt

@@ -1,4 +1,4 @@
-xtevent 2.1.1 - August 12, 2022
+xtevent 2.2.2 - Mar 15, 2023
 
 
 Authors:
@@ -7,9 +7,9 @@ Authors:
        [email protected]
     Christian Hansen, University of Chicago, Booth School of Business.
        [email protected]
-    Jorge Pérez Pérez, Banco de México
+    Jorge Pérez Pérez, Banco de México.
        [email protected]
-    Jesse Shapiro, Harvard University and NBER
+    Jesse Shapiro, Harvard University and NBER.
        [email protected]       
 
 
@@ -18,8 +18,10 @@ Simon Freyaldenhoven, Christian Hansen, Jorge Pérez Pérez, and Jesse M. Shapir
 
 
 Changelog:
-- Fixed bugs present in xtevent 2.1.0.
-- Updates in the help files and other documentation.
+- Add cohort and control_cohort to obtain estimates using Sun and Abraham's (2021) method.
+- Add repeatedcsoption and get_unit_time_effectscommand to estimate event-studies in repeated cross-section settings.
+- Add noestimate option to savek() to generate event-time dummies without estimating the regression model.
+- Enable clustered and robust standard erros for IV estimation without reghdfe.
+- Fixed bugs present in version 2.1.1.
 
-
-For more information, see https://github.com/JMSLab/xtevent/releases/tag/v2.1.1.
+For more information, see https://github.com/JMSLab/xtevent/releases/tag/v2.2.0.

xtevent/xtevent.ado

@@ -1,4 +1,4 @@
-* xtevent.ado 2.2.0 Sep 29 2022
+*! xtevent.ado 2.2.0 Mar 15 2023
 
 version 11.2