Out-of-Sample Forecast Tests Robust to the Window Size Choice

• Activity Number: 20
• Type: Topic Contributed
• Date/Time: Sunday, July 31, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Barbara Rossi*+ and Atsushi Inoue
• Companies: Duke University and North Carolina State University

Abstract

This paper proposes a new methodology for comparing the relative out-of-sample forecasting performance of two competing models that is robust to the choice of the estimation window size. The methodology involves evaluating the relative predictive ability of the competing models over a wide range of window sizes. We show that the tests proposed in the literature may lack power to detect predictive ability, and might be subject to data snooping across different window sizes if used repeatedly. An empirical application to models of exchange rate determination shows that the recent findings of economic models forecasting ability are not robust to the choice of the estimation window size.