Second order bias in a forecast evaluation statistic (in Russian)
We derive the second order asymptotic bias for a statistic evaluating the quality of out-of-sample forecasts by parametric models. The existing literature admits that the quality of the first order asymptotic approximation may be unsatisfactory. We find that the second order asymptotic bias allows one to explain insufficient precision of the asymptotic approximation. Simulations confirm the obtained analytical results.
When requesting a correction, please mention this item's handle: RePEc:qnt:quantl:y:2009:i:6:p:77-91. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Stanislav Anatolyev)
If references are entirely missing, you can add them using this form.