Error while replicating FlexPanelDID command

[Scott Rick]

Hi. I am trying to run the flexpaneldid command as described in the paper that introduced it (https://papers.ssrn.com/sol3/papers....act_id=3692458) but it gives me the error message below. I then ran the flexpaneldid_preprocessing just in case, but the resulting dataset no longer has the dependent variable 'patents' or the treatment variable 'treatment' as described in the paper anymore. As a result the flexpaneldid command as in the paper no longer can be run.

Has anyone else had success in replicating this code? And if so, what should I be doing?

Error message: #1

 use "http://fmwww.bc.edu/RePEc/bocode/f/flexpaneldid_example_data.dta", clear

. flexpaneldid patents, id(cusip ) treatment(treatment ) time(year ) statmatchin
> g(con(employ stckpr rnd sales ) cat(pats_cat rndstck_cat )) outcometimerelstar
> t(3 ) outcomedev(-2 ) test


********************************************************************************
************************* flexpaneldid *****************************************
********************************************************************************

--------------------------------------------------------------------------------
outcome:                patents
id:                     cusip
treatment:              treatment
time:                   year
outcome_time_start:     3
outcome_time_end:       .
outcome_dev:            -2
cemmatching:            
statmatching:           con(employ stckpr rnd sales ) cat(pats_cat rndstck_cat )
test:                   test
outcomemissing:         
didmodel:               
--------------------------------------------------------------------------------
invalid file specification
r(198);




Error message: #2

use "http://fmwww.bc.edu/RePEc/bocode/f/flexpaneldid_example_data.dta", clear

flexpaneldid_preprocessing, id(cusip) treatment(treatment) time(year) matchvar
> s(employ stckpr rnd sales pats_cat rndstck_cat) matchtimerel(-1) matchvarsexac
> t(sic_cat) prepdataset("preprocessed_data.dta") replace


********************************************************************************
************************* flexpaneldid - preprocessing *************************
********************************************************************************

--------------------------------------------------------------------------------
id:                     cusip
treatment:              treatment
time:                   year
matchvars:              employ stckpr rnd sales pats_cat rndstck_cat
matchvarsexact:         sic_cat
match_time:             -1
prepdataset:            preprocessed_data.dta
--------------------------------------------------------------------------------


********************************************************************************
************************* Preprocessing ****************************************
********************************************************************************

Preprocessing of 61 treated: 
.................................................. 50 
........... 61

********************************************************************************
************************* Preprocessing - Summary ******************************
********************************************************************************

Number of treated:                                    61
Number of treated dropped during preprocessing:        0
Number of treated after preprocessing:                61
Mean size of selection groups:                   45.7377

. flexpaneldid patents, id(cusip ) treatment(treatment ) time(year ) statmatchin
> g(con(employ stckpr rnd sales ) cat(pats_cat rndstck_cat )) outcometimerelstar
> t(3 ) outcomedev(-2 ) test
variable patents not found

[Scott Rick]

Does anyone know how to do this? I've tried several different approaches, but they all seem to fail

[Sayantani Sayantani]

I think I got this to work for me by doing the following:

1. Run the pre-processing first using the following code and save it in the same file where you have saved the example dataset. The example dataset can be obtained from "http://fmwww.bc.edu/RePEc/bocode/f", under the link flexpaneldid_example_data.dta.

To the run the pre-processing, first load the flexpaneldid_example_data.dta and the use the code
flexpaneldid_preprocessing, id(cusip) treatment(treatment) time(year) matchvars(employ stckpr rnd sales return pats_cat rndstck_cat rndeflt_cat ) matchtimerel(-1 ) matchvarsexact(sic_cat ) prepdataset('preprocessed_data.dta') replace

Save this newly created file in the same folder as your flexpaneldid_example_data.dta.

2. Load the flexpaneldid_example_data.dta again. Now run the flexpanel did code. But remember that under
"prepdataset("preprocessed_data.dta") replace" at the end, you need the whole path of where the preprocessed_data.dta is stored within the parenthesis. Make sure the filename of the preprocessed data is exactly correct and then it should work.