Dear all,
I am estimating an Almost Ideal Demand System to obtain own-price elasticities. I have done that successfully, but am not sure which restrictions to use. The aidsills command allows to include two restrictions, namely symmetry and homogeneity. One can estimate the model including one or the other, none or both. If NO restriction is included, a Chi^2 homogeneity test is performed automatically, whereas a Chi^2 test for symmetry is performed when the homogeneity restriction is included. I estimated all possible combinations of restrictions and always obtained own-price elasticities that were significant with a p-value < 0.01.
The homogenity and symmetry test both showed a Prob > chi2 =0.0000, which means that symmetry and homogeneity can be rejected at any conventional level. Does that mean I need to use the unrestricted model or the one with both restrictions?
I am thankful for any response.
Best,
Julian
I am estimating an Almost Ideal Demand System to obtain own-price elasticities. I have done that successfully, but am not sure which restrictions to use. The aidsills command allows to include two restrictions, namely symmetry and homogeneity. One can estimate the model including one or the other, none or both. If NO restriction is included, a Chi^2 homogeneity test is performed automatically, whereas a Chi^2 test for symmetry is performed when the homogeneity restriction is included. I estimated all possible combinations of restrictions and always obtained own-price elasticities that were significant with a p-value < 0.01.
The homogenity and symmetry test both showed a Prob > chi2 =0.0000, which means that symmetry and homogeneity can be rejected at any conventional level. Does that mean I need to use the unrestricted model or the one with both restrictions?
I am thankful for any response.
Best,
Julian
