Ανάρτηση Ερευνητικού Δοκιμίου no 08/24

Ερευνητικό Δοκίμιο no 08/24 με τίτλο "Asymptotics of a QLR-type Τest for Οptimal Predictive Ability"

του Στέλιου Αρβανίτη 

Περίληψη

The limit theory of a Gaussian Quasi-likelihood Ratio (QLR)-type test for the hypothesis  of Optimal Predictive Ability is developed in the present note. This hypothesis which  generalizes the Superior Predictive Ability hypothesis from a single given loss function to an entire class of loss functions was considered in APPK21 (1). The results are developed  indicatively for the class of Convex Loss functions. As in APPK21 (1), the research hypothesis is formulated in terms of moment inequality conditions, the empirical versions of which are reducible to a set of piece-wise linear functions. A consistent and exact test is constructed based on a QLR-type test statistic, moment selection and subsampling.

Ο Στέλιος Αρβανίτης είναι Καθηγητής του τμήματος Οικονομικής Επιστήμης του Οικονομικού Πανεπιστημίου Αθηνών.