Empirical catch equations for the GMRFF states were characterized

as a system of seemingly unrelated regression equations with first order

autoregressive disturbances and cross-equation parameter restrictions.

The presence of contemporaneous correlation across states was expected

due to the productive interdependence of vessels operating out of dif-

ferent states. The autoregressive portion of the disturbances was

anticipated to result from the omission of explicit information on

changes in the resource stock. The estimator chosen to estimate the

system of catch equations was a four state feasible Aitken's estimator.

All parameter estimates were of the appropriate sign, reasonable magni-

tude and demonstrated acceptable statistical significance.


 Estimated Parameters


 The catch equations estimated were reduced from expressions which

resulted from the special treatment of fishing effort. Fishing effort

was defined to be the product of nominal effort (vessels) and fishing

power. Fishing power was expressed as a function of average crew size

and average vessel size. The amount of fishing effort attributed to

vessels in any given state and time period was then determined by the

corresponding average crew and vessel size.

 Parameters of the double log fishing power function were con-

strained to be equal across states due to the extreme similarity of

production practices in all states participating in the GMRFF. The

fishing power function parameters could not be estimated independently

of catch due to the definitional nature of the fishing power function.

Thus, an expression relating catch in each state to fishing effort, as

defined above, was developed. The result was a set of reduced form