ABSTRACT
Obtaining measures of uncertainty for seasonal adjustment is a long-standing problem (President’s Committee to Appraise Employment and Unemployment statistics, 1962). Wolter and Monsour (1981) propose two variance measures for X-11 seasonal adjustment that account for sampling error (SE), one better suited for the typical case of nonstationary time series. Pfeffermann (1994) and Bell and Kramer (1999) develop measures capturing additional uncertainty. Pfeffermann, Morry, and Wong (1995) apply the Pfeffermann method with ARIMA (autoregressive-integrated-moving average) extrapolation and the multiplicative mode of adjustment. Pfeffermann and Scott (1997) further
K12089 Chapter: 8 page: 185 date: February 14, 2012
K12089 Chapter: 8 page: 186 date: February 14, 2012
Modeling and
extend the method by proposing modifications that use all the X-11 irregular terms, not just the central ones, and simplify the equations for estimating the error variances when SE autocovariances are available. Scott, Sverchkov, and Pfeffermann (2005) treat month-to-month change where the series are similar to many index series.