Examining Diagnostics for Trading-Day Effects from X13-ARIMA-SEATS

• Activity Number: 373
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Osbert Pang* and Brian Monsell
• Companies: U.S. Census Bureau and U.S. Census Bureau
• Keywords: diagnostics ; trading-day effects

Abstract

This study examines some diagnostics available in X13-ARIMA-SEATS for detecting a trading-day effect. The diagnostics of interest are the model-based chi-squared test of fixed trading-day effects, the model-based F-test that is derived from the chi-squared test, and the spectrum peaks at the usual trading-day frequencies (under both the default last 8 years of data as used by the program and the full data). We simulate sets of seasonal series without trading-day effects in order to assess the significance level of the aforementioned diagnostics, and to determine the appropriate size-adjusted critical values. We subsequently simulate sets of seasonal series with trading-day effects and use those size-adjusted critical values to estimate the power of the diagnostics of interest.