On the (In)consistency of RE modeling[Positive feedback investment strategies and destabilizing rational speculation]

The notion of rational expectations (RE) as usually understood seeks to encompass two different propositions: (i) perceived law of motion equals actual law of motion: an equivalence between the probability distributions of future outcomes which inform the decisions of agents and the objective distributions which generate those outcomes; (ii) perceived law of motion equals model law of motion: a correspondence of the subjective distributions underlying the choices of agents and the distributions generated by professionally validated models (particularly that which the analyst proposes contemporaneously). Both definitions are quite different unless absolute validity is counterfactually attributed to the provisional and fallible models constructed by economists. A further ambiguity arises with the model-consistent notion since the constructs built by economists have certainly evolved and will continue to change. If an economist imputes current-model-consistent expectations to agents in the past when trying to validate that model, she attributes to those individuals different beliefs from those that the analyst held at the time. These issues condition the logic and significance of large segments of macroeconomic theory. They seem particularly relevant for the study of phenomena like the processes leading to financial crises, where unsustainable patterns of behavior may have once found support in influential bodies of macroanalysis.