What you seek are often called "spells", in this case periods in which
prices remain constant. There are many other names too, including
"runs".
My reason for using the first term is that Stata resources are available
under that heading:
SJ-7-2 dm0029 . . . . . . . . . . . . . . Speaking Stata: Identifying
spells
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . N.
J. Cox
Q2/07 SJ 7(2):249--265 (no
commands)
shows how to handle spells with complete control over
spell specification
tsspell from http://fmwww.bc.edu/RePEc/bocode/t
'TSSPELL': module for identification of spells or runs in time
series /
tsspell examines the data, which must be tsset time series, to /
identify
spells or runs, which are contiguous sequences defined / by some
condition. tsspell generates new variables indicating / distinct
spells,
With -tsset- data and -tsspell- from SSC, a solution is
tsspell P_t
gen change = _seq - 1
Nick
[email protected]
Rijo John
I have a time series data set as follows.
ID year P_t-1 P_t
1 1990 16 16
1 1991 16 16
1 1996 16 16
1 2005 42 42
1 2006 42 42
1 2007 42 42
2 1990 29 29
2 1991 29 29
2 1992 29 29
Where P_t and P_t-1 represents price in year t and t-1 respectively.
What I wish to achieve is to create a new variable "Change". This
variable should show the number of years passed since the last change
in P_t. In the above example the data should finally look like below.
ID year P_t-1 P_t Change
1 1990 . 16 0
1 1991 16 16 1
1 1996 16 16 2
1 1997 16 16 3
1 2005 16 42 0
1 2006 42 42 1
1 2007 42 42 2
2 1990 . 29 0
2 1991 29 29 1
2 1992 29 29 2
2 1993 29 35 0
If I do
bysort ID P_t : gen Change=_n-1 I get the varible numbered but it
doesnt sort the year. If I sort ID and year then I cant use the _n to
generate these numbers.
*
* 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/
