| Abstract: | Optimal harvesting policies are commonly derived by assuming a time-invariant relationship between some productivity measure and the resource variable under control. Yet, long-term trends in the environment appear to induce persistent changes in spawning success in several fish stocks. I show that when predictable trends in environmental effects are incorporated into stock-recruitment models, optimal policies respond to changing environmental conditions in a way that depends very much on the management objective. When the goal is to maximize expected discounted yield, resulting risk-neutral policies computed for a model of a cyclic iteroparous population respond by continuously adjusting optimal spawning targets in phase with the environmental cycle: escapements are raised when favorable conditions are anticipated and they are lowered when poor environments are expected. These feedback responses reinforce recruitment fluctuations and lead to a sequence of boom and bust periods in the fishery. Policies shift diametrically when a risk-averse objective is pursued such as maximization of the expected sum of discounted logarithms of catches. Optimal escapements closely parallel fluctuations in population abundance, with harvest rates and catches much less variable than in the risk-neutral policy. Harvest rates respond in a compensatory way to changes in population abundance, anticipated environmental conditions, and expected strength of incoming year-classes. Depending on the specific model used, a constant harvest rate strategy may perform nearly as well as the optimal. Analytical results are provided that characterize risk-neutral optimal policies for stochastic delay-difference population models. Results show that knowledge of current environmental conditions can be used to construct harvest policies which are nearly as good as those “optimal” ones based on long-term environmental forecasts. |
