st: estimate elasticity using LA-AIDS model

[jongin Kim <filetana@gmail.com>]

Dear members of this mailing list~
 I’m trying to estimate elasticity of wheat import in Japan using LA-AIDS model.
I used SUR(seemingly unrelated regression). I succeeded to get a
result, but the problem is when I try to compute elasticity by using
the regression result, most of the own price elasticity for USA, CAN,
AUS becomes plus. Could you explain me what’s wrong with my analysis
as follows? Or the thing that own price elasticity take plus value can
be rationalized?
Stata code
--------------------------------------------------------------
*Homogeneity condition*
 constraint define 1 [qusa]pusa+[qusa]pcan+[qusa]paus+[qusa]prow=0
 constraint define 2 [qcan]pusa+[qcan]pcan+[qcan]paus+[qcan]prow=0
 constraint define 3 [qaus]pusa+[qaus]pcan+[qaus]paus+[qaus]prow=0

  *Symmetry condition*
 constraint define 4 [qusa]pcan=[qcan]pusa
 constraint define 5 [qusa]paus=[qaus]pusa
 constraint define 6 [qcan]paus=[qaus]pcan
   *SUR with constraints*
  sureg (qusa: wusa=pusa pcan paus prow xwp ) (qcan: wcan=pusa pcan
paus prow xwp ) (qaus: waus=pusa pcan paus prow xwp ), constraints(1 2
3 4 5 6)
--------------------------------------------------------------------------------------

Results

Compensated elasticity
　 USA               CAN              AUS
USA 0.8056718 -0.335045 -0.4599036
CAN -0.676183 0.1640552 0.47014533
AUS -1.312479 0.6648092 0.66463825


Uncompensated elasticity
　 USA              CAN    AUS
USA 0.311614 -0.579848 -0.6330256
CAN -1.424014 -0.206492 0.20809894
AUS -1.74234 0.4518152 0.51401138

Data set& result are as follows

http://www.adrive.com/public/32740513c587c507d08390dcb846c6bb753142996103fb30c173f21c7b357a91.html

*
*   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/