st: panel data with three levels

[Mintewab Bezabih Ayele <M.Bezabih@sussex.ac.uk>]

Dear Statalisters, 

I have a four year panel data of farm households and I want to do a fixed effects analysis. Then within each farm  household, I have multiple plots that are not replicated across years, which is where my problem lies.

The reason the plots are not replicated across years is as follows.  Because I am studying a multicropping farming system in which plots are defined by the crops they are planted on, a potential crop rotation may mean what that a plot the grew wheat in any given year may not be the same plot that grows wheat the next year. Moreover, the size of the plots may also change as in one year the entire farm may be divided into four plots and the next year it could only be two plots. My problem is complicated by the fact that similar crops of two varieties could be planted in different plots within the household. As you could imagine I cannot have a panel  with year, farm household, and plot.

 What I want to do is to have a panel analysis with year and household id but somehow cluster the standard errors for the plots. Is there any way od doing that? 

Any help appreciated
cheers
minti


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/