Properties of the Stochastic Distance Function and its Role in Fisheries
Analyses
In this work we examine the ability of the normalized stochastic output
distance function to properly approximate the curvature properties of different
underlying production technologies. The distance function is a natural tool for
fishery production models in that it does not require cost data (only data on
input and output quantities are necessary), which is often unavailable, and it
can accommodate multi-input, multi-output relationships (common in fisheries).
However, Monte Carlo experiments have indicated that data characterized by
certain characteristics may not be properly approximated by the normalized
stochastic distance function. Thus, one could erroneously conclude that there
are economies of scope in a multi-species fishery, when in fact vessels may have
lower marginal costs targeting only one species (but restrictive TACs on one
species or trip limits lead vessels to target several species to recover fixed
costs). This work is being conducted jointly with Dan Holland at the University
of Massachusetts, Dartmouth. For more information or to comment on this project,
contact ron.felthoven@noaa.gov.
|
