A1915
Title: Initial-condition-robust inference in autoregressive models
Authors: Ming Li - National University of Singapore (Singapore) [presenting]
Donald Andrews - Yale University (United States)
Yapeng Zheng - Chinese University of Hong Kong (Hong Kong)
Abstract: A new confidence interval (CI) for the autoregressive (AR) parameter in an AR model with an AR parameter that may be close to or equal to one is introduced. Existing CIs rely on the assumption of a stationary or fixed initial condition to obtain correct asymptotic coverage and good finite sample coverage. When this assumption fails, their coverage can be quite poor. The proposed CI for the AR parameter has coverage probability that is completely robust to the initial condition, both asymptotically and in finite samples. This CI incurs only a small price in terms of its length when the initial condition is stationary or fixed. The CI is also robust to conditional heteroskedasticity of the errors.
