Halibut Sport Fishing Survey Data Collection Completed
An economic survey developed by Dr. Dan Lew (EESR Program) and Professor Doug Larson (University of California, Davis) to collect the data necessary to understand recreational angler preferences for saltwater fishing in Alaska has been completed. The survey collected information about each angler’s 2006 Alaska sport fishing activities (primarily saltwater) and preferences that will be used to estimate several economic models for assessing the impacts of potential regulatory changes on sport angler behavior.
The mail survey was initially distributed to a stratified sample of 4,000 Alaska resident and nonresident anglers who were licensed to sport fish during 2006. Several follow-up mailings were conducted, in addition to telephone contact, to maximize response rates for this voluntary survey. In the end, an overall response rate of approximately 57% was achieved.
Over the summer the data will be edited and summarized, and supplemental data will be collected to facilitate construction of key economic variables that are necessary inputs in the economic models. A summer research assistant, Jeff Ferris, will assist in these activities. Although some preliminary modeling will occur over the summer, the full data analysis cannot proceed until after these activities have been completed.
By Dan Lew
Using Bayesian Vector Autoregression (BVAR) to Identify Inter-industry Relationships for Alaska Fisheries
Using monthly borough-level employment data from 1990 to 2005, Dr. Chang Seung (ESSR Program) and Professor Sung Ahn (Washington State University) are developing a time-series model called a vector autoregression (VAR) for the fisheries off Alaska. The goal of this project is to identify the relationship between industries (with particular interest in seafood industries) for select fishery-dependent boroughs or regions so that researchers can better understand the socioeconomic impacts of fishery-related changes as well as shocks in other industries.
The modeling approach developed here will avoid many of the problems that often arise in applying VARs such as insufficient degrees of freedom, multicollinearity, and poor out-of-sample forecast performance. This will be accomplished by utilizing nonsample Bayesian prior information contained in the input-output tables within the commercially available IMPLAN data set (which defines the linkages and multiplier effects among various sectors throughout the United States for a subset of our industries of interest).
These priors will be incorporated as stochastic restrictions to the system. The Bayesian methods are expected to increase the precision of the estimates, which
will help produce more useful models for future applied analyses.
