Re: st: How to generate dummy for the following years

[Ronnie Babigumira <rb.glists@gmail.com>]

Umut
Reading through your email again, I don't think I answered your question so the only useful part of my earlier response may be how to generate always_exporter.

Ps: For what you were trying to do with

**
gen starter=0
replace starter=1 if export==0 in 2003 & export==1 in 2004

***

I am wondering if what you are looking for is a -reshape-, so

reshape wide export, i(firm_id) j(year)

This would give you

firm_id export2003 export2004 export2005 export2006 export2007 export2008 export2009 export2010 export2011
00028073 1 1 0 1 0 1 1 1 1 
00036721 0 0 0 0 0 0 0 0 . 
00047152 1 1 1 0 0 0 0 0 1 
00063351 . . . . . . . 1 1 
00144277 . . . . 1 1 0 1 . 
00144614 . . . . . . 0 0 . 
00145362 1 1 1 . . . . . . 
00147034 1 1 1 1 1 1 1 1 .

You can then 
generate starter == .
replace starter == 1 if export2003 == 0 & export2004 == 1

Or, to create a dummy, you can
gen starter = export2003 == 0 & export2004 == 1

Ronnie
-- 


010100100110111101101110011011100110100101100101




On Friday, July 27, 2012 at 12:28 AM, Ronnie Babigumira wrote:

> Umut,
> Regarding your original question on generating dummies. There will be more direct ways but here is one way to do it. 
> 
> clear
> input str8 firm_id year export
> 00028073 2003 1
> 00028073 2004 1
> 00028073 2005 0
> 00028073 2006 1
> 00028073 2007 0
> 00028073 2008 1
> 00028073 2009 1
> 00028073 2010 1
> 00028073 2011 1
> 00036721 2003 0
> 00036721 2004 0
> 00036721 2005 0
> 00036721 2006 0
> 00036721 2007 0
> 00036721 2008 0
> 00036721 2009 0
> 00036721 2010 0
> 00047152 2003 1
> 00047152 2004 1
> 00047152 2005 1
> 00047152 2006 0
> 00047152 2007 0
> 00047152 2008 0
> 00047152 2009 0
> 00047152 2010 0
> 00047152 2011 1
> 00063351 2010 1
> 00063351 2011 1
> 00144277 2007 1
> 00144277 2008 1
> 00144277 2009 0
> 00144277 2010 1
> 00144614 2009 0
> 00144614 2010 0
> 00145362 2003 1
> 00145362 2004 1
> 00145362 2005 1
> 00147034 2003 1
> 00147034 2004 1
> 00147034 2005 1
> 00147034 2006 1
> 00147034 2007 1
> 00147034 2008 1
> 00147034 2009 1
> 00147034 2010 1
> end
> 
> * Always exporter
> tempvar foo
> bys firm_id : egen `foo' = mean(export)
> gen always_exporter = `foo' == 1
> 
> * Start and end data for exporters
> * Split data
> tempfile fdata nedata
> save `fdata'
> drop if export == 1
> save `nedata'
> use `fdata'
> keep if export == 1
> 
> bysort firm_id (year): gen start= year[1]
> bysort firm_id (year): gen end= year[_N]
> 
> * Append non-export data and fillin missing data
> append using `nedata'
> bys firm_id (start): replace start = start[_n-1] if start >= . 
> bys firm_id (end): replace end = end[_n-1] if end >= . 
> 
> sort firm_id year
> list, sepby(firm_id)
> 
> 
> 
> Hope this is helpful
> 
> Ronnie
> 
> -- 
> 010100100110111101101110011011100110100101100101
> 
> 
> 
> 
> On Thursday, July 26, 2012 at 6:07 PM, umut senalp wrote:
> 
> > Dear Stata Listers,
> > 
> > I am using StataSE 12. I have an unbalanced data for over 2000 firms (covering the period 2003-2011) and I have a dummy variable for the firms' export status (Export=1 if firm have positive export value that year). While, a few firms observed for all 9 years (2003-2011), majority of them observed for minimum 2 years consecutively. 
> > I want to create a new set of dummy variables indicating the year that a firm started to export (Starter), 
> > one year before starting to export (Year_before),and for the following two years that the firm started to export which are Year_After and Two_Years_After. 
> > So, I want to generate 4 new dummies based on the firm's Export status. In addition to that, I have some firms which are exporter for the all period that they are observable (means I do not know when they started to export), and I want to name them as Always_Exporter.
> > 
> > Since I am not so familiar with programming in Stata I tried following command but it failed, it was just for the firms that I have observation for 2003 onwards, anyway.
> > gen starter=0
> > replace starter=1 if export==0 in 2003 & export==1 in 2004
> > 
> > And I realised that this approach will not work since I need to do it for many times for following years to catch all starters.
> > 
> > Well, I think I need to use a program that includes loop but I don't know where to start to write the code. I would be glad if you could help me on this issue. I hope I've the issue clear enough.
> > 
> > Kind regards
> > 
> > Umut SENALP 
> > *
> > * For searches and help try:
> > * http://www.stata.com/help.cgi?search
> > * http://www.stata.com/support/statalist/faq
> > * http://www.ats.ucla.edu/stat/stata/
> 



*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/