I am working on the Bacon decomposition (using bacondecomp command), which requires a strongly balanced panel.
The panel dataset I am working on was unbalanced, so I performed the following code to convert it into a balanced panel :
Code:
bysort district: gen nyear=[_N] tab nyear keep if nyear==18
I came across this definition of a "strongly balanced" panel: "When the dataset contains a time variable, panels are said to be strongly balanced if each panel contains the same time points, weakly balanced if each panel contains the same number of observations but not the same time points, and unbalanced otherwise." (https://www.stata.com/statalist/arch.../msg00101.html).
However, after the conversion I made, my panel dataset contains 585 districts and each district counts 18 years of data. Why is it still "weakly" balanced? How can I fix this issue?
Please find below a sample the data I am using:
Code:
* Example generated by -dataex-. To install: ssc install dataex clear input double district float(year lny x) 10101 2000 15.752553 0 10101 2001 16.193884 0 10101 2002 16.471481 0 10101 2003 16.605242 0 10101 2004 17.15857 0 10101 2005 17.22689 0 10101 2006 17.527142 0 10101 2007 18.355291 0 10101 2008 18.943258 0 10101 2009 18.69728 0 10101 2010 18.725645 0 10101 2011 18.640745 0 10101 2012 19.338213 0 10101 2013 19.44976 0 10101 2014 18.819197 0 10101 2015 19.055973 0 10101 2016 19.063976 0 10101 2017 19.14196 0 10103 2000 13.91526 0 10103 2001 13.926852 0 10103 2002 14.537033 0 10103 2003 14.863282 0 10103 2004 15.23191 0 10103 2005 15.82604 0 10103 2006 15.940092 0 10103 2007 16.163109 0 10103 2008 16.602633 0 10103 2009 16.625645 0 10103 2010 16.94231 0 10103 2011 17.234108 0 10103 2012 17.08096 0 10103 2013 17.589808 0 10103 2014 17.507406 0 10103 2015 17.655422 0 10103 2016 17.713999 0 10103 2017 17.98462 0 10105 2000 17.176313 0 10105 2001 17.218447 0 10105 2002 17.22451 0 10105 2003 17.351425 0 10105 2004 18.095589 0 10105 2005 18.538553 0 10105 2006 18.711836 0 10105 2007 19.08575 0 10105 2008 19.27568 0 10105 2009 19.96377 0 10105 2010 19.67612 0 10105 2011 19.225895 0 10105 2012 19.719824 0 10105 2013 19.645073 0 10105 2014 19.229593 0 10105 2015 19.285046 0 10105 2016 19.239407 0 10105 2017 19.5521 0 10107 2000 16.896315 0 10107 2001 17.19038 0 10107 2002 17.37088 0 10107 2003 17.51199 0 10107 2004 18.074612 0 10107 2005 18.092892 0 10107 2006 18.189089 0 10107 2007 18.554443 0 10107 2008 19.16435 0 10107 2009 18.798866 0 10107 2010 19.15159 0 10107 2011 19.48421 0 10107 2012 19.636 0 10107 2013 19.558506 0 10107 2014 19.60922 0 10107 2015 19.698847 0 10107 2016 19.85667 0 10107 2017 20.01661 0 10109 2000 17.288967 0 10109 2001 17.53965 0 10109 2002 17.691305 0 10109 2003 17.912817 0 10109 2004 18.181335 0 10109 2005 18.984133 0 10109 2006 19.162714 0 10109 2007 18.920141 0 10109 2008 18.706972 0 10109 2009 19.65674 0 10109 2010 19.695604 0 10109 2011 20.01785 0 10109 2012 20.11707 0 10109 2013 19.93405 0 10109 2014 19.91495 0 10109 2015 19.680166 0 10109 2016 19.603456 0 10109 2017 19.917625 0 end
Marina
0 Response to Converting panel data from WEAKLY balanced to STRONGLY balanced (Bacon decomposition)
Post a Comment