st: RE: insufficient observation in panel data/time values repeated within panel‏

[Nick Cox <n.j.cox@durham.ac.uk>]

Much depends on what you are trying to do, which isn't at all clear. 

But, if as your example implies, you are assigning an individual identifier to each observation, then your data are not panel data in any strong sense. Indeed, there is no point in defining them as such because necessarily you can do almost nothing with them in that form. 

Nor it seems do counties -- not indicated in this example -- define panels in any useful sense as presumably the number of houses sold may easily vary from year to year even for a given county. County, community: same or different? 

There is still much you can do with these data, but you need to say much more about what you think is the generating process to allow really good advice to be given by others, meaning here economists or econometricians. But I think you have not panels, but clusters. 

Nick 
n.j.cox@durham.ac.uk 

ajita

I have panel data on housing sales price  that looks like the following:
id        saleyear    saleprice community
1          2002        80000       1517
2          2001        90000       1517
3          2001        100000      1560
4          2001         80000       1560
Id is created by myself using generate id command for each house sold in a
particular county in a particular year. When i do tsset id saleyear it works
fine but after that when I do xtreg dependent var independent var I get a
message as few observations.  When I used community as my unique id and did
Xtset community saleyear I got r message as time values repeated within
panel since I have observation on different house sold in same year in the
same community. I appreciate your effort.

*
*   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/