A Comparison of Methods for Seasonal Adjustment of the Monetary Aggregates

Abstract:     
This paper compares the performance of a four methods of seasonal adjustment for monthly monetary aggregates. The methods compared are 1. GLAS - the method currently employed by the Bank. 2. STL - a recently developed flexible non-parametric adjustment method. 3. X-11 ARIMA (a slightly modified version of X-11) an established adjustment method widely used by government agencies including the CSO. 4. STAMP a more structural method of adjustment than the other three developed by Andrew Harvey of the LSE.

The paper compares these approaches against a number of criteria including the ability to give the best estimate of current data, the ability to deal with adding up constraints and trading day adjustments, and ease of use. Predictability perhaps, the paper finds that no one method does best on all criteria and so a simple conclusion on the best method cannot be drawn on the evidence of this paper. However, future work will seek to reach a conclusion about the most appropriate seasonal adjustment method for the Bank of England to apply to these series, drawing perhaps from the discussion generated by this paper.

JEL Classification: C22, C32, E51

working papers series