List some few observations of a variable in panel data

[Priel Paul]

Hello @all,
I am trying to estimate a panel growth model of 50 countries, using quarterly data from 2001 to 2020.

My aim is to create block period of time (i.e, 2001-2005, 2006-2010, 2011-2015, 2016-2020) based on my dependent variable(as shown in the model below) and consider each block period as a single observation by taking/listing the first year together with its quarters for each block (i.e, 2001q1, 2001q2, 2001q3, 2001q4, 2006q1, 2006q2, 2006q3, 2006q4, 2011q1, 2011q2, 2011q3, 2011q4, 2016q1, 2016q2, 2016q3, 2016q4)

(ln(y_it-20 )- ln(y_it))/20= A + Bln(y_it-1) + CX+ u_it

I have already created a column (Growth~e) based on my dependent variable(as can be seen in the table below), since I have an interval of 5 years for each block period, I have divided by 20 the difference ln(f20.y_it )- ln(f.y_it).

Please note that the first few values of Growth~e column are missing entries.

I am kindly asking if there is a command available that allows me to make some of these observations in that way.

Kind regards,

Priel.

[Andrew Musau]

This article discusses how to bin: https://journals.sagepub.com/doi/pdf...867X1801800311. It is difficult to provide exact code because we do not know whether your dates are quarterly dates or whether they are strings. But an approach would involve passing a quarterly date through the -dofq()- and -year()- functions and subsequently binning.

clear
set obs 50
gen qtr= 163+_n
format qtr %tq
gen wanted= ceil(year(dofq(qtr))/5)*5

Res.:

. l, sepby(wanted)

     +-----------------+
     |    qtr   wanted |
     |-----------------|
  1. | 2001q1     2005 |
  2. | 2001q2     2005 |
  3. | 2001q3     2005 |
  4. | 2001q4     2005 |
  5. | 2002q1     2005 |
  6. | 2002q2     2005 |
  7. | 2002q3     2005 |
  8. | 2002q4     2005 |
  9. | 2003q1     2005 |
 10. | 2003q2     2005 |
 11. | 2003q3     2005 |
 12. | 2003q4     2005 |
 13. | 2004q1     2005 |
 14. | 2004q2     2005 |
 15. | 2004q3     2005 |
 16. | 2004q4     2005 |
 17. | 2005q1     2005 |
 18. | 2005q2     2005 |
 19. | 2005q3     2005 |
 20. | 2005q4     2005 |
     |-----------------|
 21. | 2006q1     2010 |
 22. | 2006q2     2010 |
 23. | 2006q3     2010 |
 24. | 2006q4     2010 |
 25. | 2007q1     2010 |
 26. | 2007q2     2010 |
 27. | 2007q3     2010 |
 28. | 2007q4     2010 |
 29. | 2008q1     2010 |
 30. | 2008q2     2010 |
 31. | 2008q3     2010 |
 32. | 2008q4     2010 |
 33. | 2009q1     2010 |
 34. | 2009q2     2010 |
 35. | 2009q3     2010 |
 36. | 2009q4     2010 |
 37. | 2010q1     2010 |
 38. | 2010q2     2010 |
 39. | 2010q3     2010 |
 40. | 2010q4     2010 |
     |-----------------|
 41. | 2011q1     2015 |
 42. | 2011q2     2015 |
 43. | 2011q3     2015 |
 44. | 2011q4     2015 |
 45. | 2012q1     2015 |
 46. | 2012q2     2015 |
 47. | 2012q3     2015 |
 48. | 2012q4     2015 |
 49. | 2013q1     2015 |
 50. | 2013q2     2015 |
     +-----------------+

.

To get a single observation per period, it appears that you want to collapse the data.

help collapse

[Priel Paul]

Hi Andrew,

Thank you for reply, if this help, below is a structure of my data. I am a new Stata user. Thanks for the article, let me check it.