Minimum information management and price-abundance relationships in a fishery

Abstract

We explore the information content of dockside prices and fishing costs in the estimation of stock abundance. Our approach is two-pronged: we first examine whether the implied biomass, that is, the biomass that is consistent with a simple microeconomic model calibrated with observed prices and costs, offers an approximation of actual stock assessments—both agree over the first 20 years of observation, but diverge over the last five. In a second approach, we use annual data in Vector Autoregressive (VAR), Bayesian VAR (B-VAR), and Vector Error Correction (VEC) frameworks and add monthly data in a mixed-frequency data analysis including Mixed-Frequency Bayesian VAR (MF-BVAR) and Mixed-Data Sampling (MIDAS) frameworks for log-differenced time series. Parameter uncertainties are addressed through Bayesian regression and forecasting methods. We find a statistically significant correlation between biometric estimates and changes in a price-based indicator that is robust to the inclusion of confounding factors. We conclude that the combination of price data and per-trip landings, when interpreted with care, can serve as a complementary, but comparatively affordable and timely, source of information for stock assessments.