134
for these two products, thus increasing the demand for shrimp and
eventually bidding up the price for shrimp products. Due to the
relatively small percentage share that shrimp accounts for in the per
capita consumption of meat, poultry, and fish, however, the measured
impact may be simply the correlation over time between the prices of
shrimp and other meat products. The parameter estimate for real
disposable income was not found to be statistically significant. Thus,
real income has little effect on the determination of retail price of
either size class on a monthly basis. This is verified by simply
observing the data over the time period of the analysis (1972-1981)
prices have fluctuated as real disposable income remained relatively
stable* Finally, the constant term was also found to be insignificant.
Quarterly Data
Structural expressions for quarterly retail, wholesale, and ex
vessel demand are estimated using the three stage least squares
method. This approach is warranted due to (1) the simultaneous nature
of the price relationships found in the causality results for the quar
terly data and (2) the presence of significant contemporaneous cross
correlation of the residuals from two stage least squares analysis of
the simultaneous system. Once the structural estimates of the coeffi
cients are obtained, derived reduced form parameters are calculated. In
addition, the presence of lagged dependent variables in the reduced form
expressions necessitates the derivation of the final form estimates of
the parameters. These final form estimates are then used to construct
margin expressions in terms of the full set of exogenous variables of
the structural model. Final form expressions are computed for both 31-