Sorting Models in Discrete Choice Fisheries Analysis

One of the greatest challenges facing empirical fisheries researchers is to endogenize fishing effort in bioeconomic models in a way that accounts for fleet heterogeneity. Such heterogeneity can manifest in a wide range of both observable and unobservable characteristics of fishing vessels and individual fishermen. Without accounting for heterogeneity, we simply have an incomplete understanding of how pressure on fish resources responds to policy instruments that are available, the states of fish stocks, and exogenous shocks to the system. Largely due to data limitations, the discrete choice fisheries literature has focused on modeling unobserved heterogeneity through random parameters. In this paper, we draw on the industrial organization literature on product differentiation and the public economics literature on spatial sorting to estimate sorting models of observable heterogeneity. Models of this type estimate individual-specific structural coefficients based on observable individual characteristics and choice-specific constants using contraction mapping. We apply the methods to location choices and target species choices in the Gulf of Mexico reef-fish fishery. For this application, we have an unusual data set that couples daily observations from logbooks with demographic information from a mail survey of captains. We use contraction mapping to control for spatially-, and species-explicit stock information. The models are used to explore spatial and inter-temporal species effort substitution in response to two marine reserves, which are implemented in sample.