A0527
Title: New robust inference for predictive regressions
Authors: Anton Skrobotov - Russian Presidential Academy of National Economy and Public Administration and SPBU (Russia) [presenting]
Rustam Ibragimov - Imperial College London and St. Petersburg State University (United Kingdom)
Abstract: New simple approaches are proposed to robust inference in predictive regressions. First, we utilize instrumental variable estimators such as Cauchy estimator to guarantee the asymptotic normality of the estimator regardless of order of integration of the variables in regression models and endogeneity. Second, we apply previous results that show that robust inference on unknown parameters of interest under heterogeneity and dependence may be conducted by partitioning the data into some number of groups and performing the standard t-test with the parameter's group estimates and critical values of Student-t distributions. The proposed approach compares favorably with widely used alternative inference procedures in terms of its finite sample properties and is robust to different forms of nonstationary volatility and heavy tails in terms of size control whereas its power is comparable to other tests.
Journal