Hi,

I work on unbalanced panel data (26 states ("cant"), time period around 1970-2017) on wealth shares ("P90"). I want to generate a new variable that measures for each state the overall change in P90 from the last to the first observation (over time). Look at the excerpt from my dataset for an example:

For cant=1, I want to create a variable that is equal to: 0.755419 – 0.7490362

Note that the dataset doesn't include the same first year for all cantons. The first observation for cant=1 is 1975. For cant=2, it is 1977.

How do I create such a variable for all 26 states?

Thanks!

input float cant double year float P90
1 1975 .7490362
1 1981 .7319754
1 1983 .7209229
1 1987 .7390646
1 1991 .724773
1 1995 .7064956
1 1999 .7529215
1 2003 .734058
1 2004 .7299312
1 2005 .7453706
1 2006 .7487444
1 2007 .7550854
1 2008 .7425138
1 2009 .7330889
1 2010 .7377599
1 2011 .7376866
1 2012 .741205
1 2013 .742767
1 2014 .7461374
1 2015 .7498978
1 2016 .7495821
1 2017 .755419
2 1977 .6740181
2 1981 .6523791
2 1991 .6468652
2 1997 .6702179
2 2003 .6489691
2 2004 .6559614
2 2005 .6677032
2 2006 .6725178
2 2007 .6955403
2 2008 .6790535
2 2009 .6876738
2 2010 .6895582
2 2011 .6878965
2 2012 .6927519
2 2013 .6989366
2 2014 .6949768
2 2015 .7029172
2 2016 .7052284
2 2017 .7107742
3 1973 .743027
3 1981 .6784297
3 1983 .672592
3 1991 .7058665
3 1993 .714372
3 1995 .6840838
3 1997 .7190968
3 1999 .7196878
3 2003 .7256334
3 2004 .7265204
3 2005 .7271922
3 2006 .7188094
3 2007 .7376335
3 2008 .7220148
3 2009 .7229825
3 2010 .7288389
3 2011 .7396641
3 2012 .7448026
3 2013 .7533194
3 2014 .758797
3 2015 .7646412
3 2016 .7693793
3 2017 .769841
4 1981 .6057187
4 1991 .6259277
4 1997 .57855046
4 2003 .51805264
4 2004 .53431404
4 2005 .53986645
4 2006 .545574
4 2007 .55528235
4 2008 .5523054
4 2009 .5599864
4 2010 .56871885
4 2011 .52957606
4 2012 .54068273
4 2013 .54666424
4 2014 .5533657
4 2015 .55440116
4 2016 .56613487
4 2017 .5771765
5 1979 .685352
5 1981 .33659706
5 1983 .3615786
5 1985 .3325868
5 1987 .3805338
5 1989 .4322503
5 1991 .4656941
5 1993 .4792115
5 1995 .5525481
5 1997 .6400445
5 1999 .7036961
5 2001 .7184193
5 2002 .6966427
5 2003 .8024087
5 2004 .7394421
5 2005 .7702557
5 2006 .7694538
5 2007 .8118458
end
format %ty year
[/CODE]