Some Comments on the Current State of Econometrics

Regarding the current econometric scene, in this review I argue that (a) traditional econometric modeling approaches do not provide a reliable basis for making inferences about the causal effect of a supposed treatment of data in observational and quasi-experimental settings; and (b) the focus on conventional reductionist models and information recovery methods has led to irrelevant economic theories and questionable inferences and has failed in terms of prediction and the extraction of information relative to the nature of underlying economic behavior systems. Looking ahead, a nontraditional econometric approach is outlined. This method recognizes that our knowledge regarding the underlying behavioral system and observed data process is complex, partial, and incomplete. It then suggests a self-organized, agent-based, algorithmic-representation system that involves networks, machine learning, and an information theoretic basis for estimation, inference, model evaluation, and prediction.