Model Selection for Causal Models: The Global Procedure With AICc and AICu This paper introduces a model selection approach for causal models, and compares its performance with that of Hsiao's (1981) approach. Model selection criteria, AICc and AICu, for the selection of causal regression models are developed. In contrast to Akaike's (1973) AIC, and Akaike (1969) and Hsiao's (1981) FPE criteria, ...