Bimodal t-ratios: The Impact of Thick Tails on Inference

This paper studies the distribution of the classical t-ratio with data generatedfrom distributions with no nite moments and shows how classical testing is affectedby bimodality. A key condition in generating bimodality is independenceof the observations in the underlying data generating process (DGP). The paperhighlights the strikingly different implications of lack of correlation versus statisticalindependence in DGPs with infinite moments and shows how standard inferencecan be invalidated in such cases, thereby pointing to the need for adapting estimationand inference procedures to the special problems induced by thick-tailed (TT)distributions.[...]