The evolution of forecasting for decision-making in dynamic environments

Abstract

Climate change and technological advances are reshaping ecosystems and societies. Strategic choices that were best yesterday may be sub-optimal tomorrow; and environmental conditions that were once taken for granted may soon cease to exist. In dynamic settings, how people choose behavioral strategies has important consequences for environmental dynamics. Economic and evolutionary theories make similar predictions for strategic behavior in a static environment, even though one approach assumes perfect rationality and the other assumes no cognition whatsoever; but predictions differ in a dynamic environment. Here we explore a middle ground between economic rationality and evolutionary myopia. Starting from a population of myopic agents, we study the evolutionary viability of a new type that forms environmental forecasts when making strategic decisions. We show that forecasting types can have an advantage in changing environments, even when the act of forecasting is costly. Forecasting types can invade but rarely overtake the population, producing a stable coexistence with myopic types. Moreover, forecasting fosters collective intelligence by providing a public good which reduces the amplitude of environmental oscillations and often increases mean payoffs to forecasting and myopic types alike. We interpret our results for understanding the evolution of different modes of decision-making such as forecasting. And we discuss implications for the management of environmental systems of societal importance.