Traditional surplus production models are based on restrictive assumptions of fish population dynamics and distribution, fish catch, and fishing effort, and can provide misleading guidelines for fisheries management. Following a review of two conventional models and later contributions, this article first reassesses the rationale for unique sustainable fishery equilibrium and its estimation, in light of complex responses associated with non-constant marginal returns, target switching, and stochastic variability. A new modelling approach is then formulated, with specifications estimated by classical or cointegration regression techniques accounting for possible switches in long-run equilibrium parameters and varying strengths of short-run readjustments. This approach is applied to marine fisheries in Indonesia (Java Sea) and Papua New Guinea, and results are compared with traditional model specifications. Parametric and nonparametric estimates highlight specific problems in the conventional approach, in terms of diagnostic tests and long-term equilibrium analysis.

JEL Classification Codes: C14, C22, Q22