Are long-horizon expectations (de-)stabilizing? Theory and experiments

We consider boundedly rational agents who do not plan over the infinite future but make trading plans at a finite, arbitrary horizon. We investigate the role of that horizon in the price dynamics of an asset in a Lucas tree model. We then design a laboratory experiment to test our theoretical predictions against the behaviors of human subjects. Short-horizon markets are prone to substantial and prolonged deviations from rational expectations (RE). By contrast, markets populated by even a modest share of long-horizon forecasters exhibit convergence towards the fundamental price. Longer-horizon forecasts do display more heterogeneity and thus some departure from RE; however this same heterogeneity also prevents the coordination of subjects on wrong anchors - a pattern of behavior that leads to mispricing in short-horizon markets. Long-horizon forecasts are well-described by adaptive learning, which delivers convergence to RE equilibrium, while short-horizon forecasts exhibit destabilizing trend-chasing.