135
40 and 21-25 size class shrimp.. The results are discussed by size class
in terms of initial structural, reduced form and final form estimates
and final form margin estimates. Tests for asymmetry were not performed
due to the simultaneous nature of the quarterly models and the fact that
asymmetry tests require recursiveness.
As pointed out by Christ (1966) and Kmenta (1971), conventional
tests of statistical significance; i.e. t-statistics, are not valid for
most simultaneous estimation procedures. This is particularly true for
small sample studies such as the present study. Thus, the discussion of
the findings from the quarterly analysis emphasizes all estimates
possessing the anticipated sign, regardless of the relative sizes of a
given parameter estimate and the corresponding standard error as found
in the initial structural estimation. In the analysis of quarterly
data, the structural estimates represent only partial effects; i.e., the
impact of a predetermined variable only on the level, or structure, in
the market for which it has a direct effect. The reduced and final form
estimates provide, in addition, the indirect effects on the other market
levels.
The 31-40 Size Class
Price models for retail, wholesale, and ex-vessel market levels are
presented, in terms of the structural, reduced form, and final form
2
parameter estimates. The structural model had a system R of .9660.
Final form expressions for retail/wholesale and wholesale/ex-vessel
margins are also discussed.
Retail structural estimates
Empirical estimates of the parameters and corresponding standard
errors for the retail price model are given as