Seasonal Adjustment and Measuring Persistence in Output.

We examine the robustness of persistence measurement to different seasonal adjustment procedures. The empirical analysis is based on U.S. data for real gross national product (GNP). Officially published data on U.S. GNP are seasonally adjusted by Census X-11. We find that this series gives significantly higher measures of persistence than GNP data adjusted by alternative seasonal adjustment methods. Some Monte Carlo evidence in support of this conclusion is also presented. Copyright 1990 by John Wiley & Sons, Ltd.