Akaike-type criteria and the reliability of inference: Model selection versus statistical model specification

作者：Spanos Aris^*

来源：Journal of Econometrics, 2010, 158(2): 204-220.

DOI：10.1016/j.jeconom.2010.01.011

• Akaike Information Criterion
• AIC
• BIC
• GIC
• MDL
• Model selection
• Model specification
• Statistical adequacy
• Curve-fitting
• Mathematical approximation theory
• Simplicity
• Least-squares
• Gauss linear model
• Linear regression model
• AR(p)
• Mis-specification testing
• Respecification
• Double-use of data
• Infinite regress and circularity
• Pre-test bias
• Model averaging
• Reliability of inference

Since the 1990s, the Akaike Information Criterion (AIC) and its various modifications/extensions, including BIC, have found wide applicability in econometrics as objective procedures that can be used to select parsimonious statistical models. The aim of this paper is to argue that these model selection procedures invariably give rise to unreliable inferences, primarily because their choice within a prespecifled family of models (a) assumes away the problem of model validation, and (b) ignores the relevant error probabilities. This paper argues for a return to the original statistical model specification problem, as envisaged by Fisher (1922), where the task is understood as one of selecting a statistical model in such a way as to render the particular data a truly typical realization of the stochastic process specified by the model in question. The key to addressing this problem is to replace trading goodness-offit against parsimony with statistical adequacy as the sole criterion for when a fitted model accounts for the regularities in the data.

• 出版日期2010-10
• 单位美国弗吉尼亚理工大学(Virginia Tech)