In this study, θ is the seemingly unrelated regressions (SUR) estimator, and θ * is the
three-stage least squares (3SLS) estimator. Under the exogeneity assumption of the expenditure
variable of the demand system, the SUR estimator is consistent and asymptotically efficient. If
the expenditure variable is endogenous, the SUR estimator is no longer consistent or efficient,
whereas the 3SLS estimator is inefficient but consistent.
The calculated values of the chi-square for all meat and fish products in the system are
larger than the critical value of chi-square with 9 degrees of freedom at the 5 percent significance
level. Thus, the null hypothesis that the expenditure variable is exogenous is rejected (Table 3).
The result indicates that the expenditure variable in the LA/AIDS should be endogenous.
Therefore, the 3SLS estimator is used to estimate the LA/AIDS for meat and fish
products in Korea. The instruments employed in the estimation are the first lags of all prices and
expenditure variables, disposable income, and the consumer price index (Eales, Durham, and
Wessells).
Estimation Results
Table 4 presents the 3SLS estimates of the LA/AIDS for meat and fish products in Korea
using quarterly data from 1980 to 1998. The system weighted R2 is 0.788, indicating that
independent variables in the LA/AIDS explain 78.8 percent of the data variation. Prices of
Hanwoo beef, pork, chicken, and fish exhibit significant effects in their corresponding equations
but those of other products (imported beef, crustacean, and mollusk) are not significant. The
expenditure variable is significant in the equations for imported beef, pork, and fish; but not in
the share equations of Hanwoo beef, chicken, and crustacean.
11