Hi,
I would greatly appreciate any help on the following problem. The problem
is similar to one previously discussed, but I am struggling with an added
layer or two of complexity.
I have a time series of data that looks generally like the following, with
many more unitid’s in actuality than those listed below:
unitid year approp
100 1985 127893
100 1986 134596
100 1987 147345
. . .
. . .
. . .
100 2004 189234
107 1985 192345
107 1986 200334
107 1987 234345
. . .
. . .
. . .
107 2004 257832
112 1985 192345
112 1986 213999
112 1987 223444
. . .
. . .
. . .
112 2004 287862
etc
.
.
.
I wish to compute a semi-log growth rate for the variable <approp> for
each <unitid>. More specifically, for each <unitid> and <year> combination
within the panel, I wish to compute a semi-log growth rate for the
variable <approp> for the trailing three year period beginning two years
in the past. As such, the first non-missing growth rate in the panel would
appear in 1990 for each <unitid>, and be computed by regressing the log of
<approp> for that <unitid> over the three year period between 1988 and
1985. The growth rate for <approp> in 1991 for each <unitid> would then be
computed by regressing the log of <approp> for that <unitid> over the
three year period between 1989 and 1986... and so forth and so on.
I have attempted a number of different approaches, but continue to hit a
wall. Any guidance would be greatly appreciated; many thanks in advance
for your consideration.
Cheers, jas
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/