Residual Seasonality in U.S. GDP Data

Rudebush et al (2015a, b) and the Bureau of Economic Analysis find the presence of residual seasonality in the official estimates of U.S. real gross domestic product (GDP).Directly seasonally adjusting official seasonally adjusted GDP, which we refer to as double seasonal adjustment, could revise the first quarter growth in the past several years upward by an average of about 1.5 percentage points.The presence of residual seasonality can significantly distort current analysis of national and regional economies.In this paper we look more closely at the U.S. GDP data and study the quality of the seasonal adjustment when it is applied to data that has already been indirectly seasonally adjusted.We find that double seasonal adjustment can lead to estimates that are of moderate quality.While the optimal method would be to directly seasonally adjust the aggregate not seasonally adjusted data, if this is not possible, double seasonally adjusted data would likely lead to better estimates.1 We thank Dominique Ladiray, National Institute of Statistics and Economic Studies -France, and session participants in the Seasonal Adjustment 2 session of the International Symposium on Forecasting in Santander, Spain June, 2016, for their comments and suggestions.The views expressed in this paper are those of the authors alone and do not necessarily reflect the views of the Federal Reserve Bank of Dallas or the Federal Reserve System.