Seasonal Adjustment and the Great Recession: Implications for Statistical Agencies

Richard B. Tiller and Thomas D. Evans

Abstract

The Great Recession (Dec07-Jun09) and its aftermath have raised major challenges to the practice of seasonal adjustment by statistical agencies around the world. Recent studies have questioned the adequacy of the standard approach to handling recession related disturbances and have offered different interpretations and solutions. A problem with evaluating these alternatives is that we can never know the true recession effects. A useful way to obtain objective information is to directly test the robustness of commonly used seasonal adjustment filters to pre-specified test series designed to reflect plausible recession effects. This study focuses on sensitivity testing of X-11 filters using artificial test series based on the behavior during the Great Recession of U.S. total non-farm payroll employment from the Current Employment Statistics (CES) program. The results show that the standard methodology is reasonably robust to recession effects.