For a panel of 60 countries I need to calculate the average for every 'segment' of time of the budget surplus for that country. The data that I already have are the country, year, surplus and break year. (The break year is determined by a Bai and Perron structural break filter in another program and I entered them manually already for all countries). Now I need to know if the average surplus in the segment after the break is larger than the average surplus before the break (if it is, the country implements a consolidation).
Therefore, I am trying to make the new 'segment' variable based on the 'break' variable. Segment is 1, and if there is a break, the following year start a new segment. Having the segment variable will allow me to quickly compare the averages of segments.
Is there a way to make Stata create this segment variable?
| country | year | surplus | break | segment |
| Australia | 1990 | 1 | 0 | 1 |
| Australia | 1991 | -1 | 0 | 1 |
| Australia | 1992 | -2 | 0 | 1 |
| Australia | 1995 | -2 | 0 | 1 |
| Australia | 1996 | -2 | 1 | 1 |
| Australia | 1997 | -2 | 0 | 2 |
| Australia | 1998 | 4 | 0 | 2 |
| Australia | 1999 | 4 | 0 | 2 |
| Australia | 2000 | 4 | 0 | 2 |
| Australia | 2001 | 3 | 1 | 2 |
| Australia | 2002 | 3 | 0 | 3 |
| Australia | 2003 | 4 | 0 | 3 |
| Australia | 2004 | 4 | 0 | 3 |
| Australia | 2005 | 4 | 0 | 3 |
0 Response to Identifying segments based on binary variable
Post a Comment