1504.0 - Methodological News, Dec 2001  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 15/11/2001   
   Page tools: Print Print Page Print all pages in this productPrint All

EASTER PROXIMITY EFFECT CORRECTION FOR RETAIL TIME SERIES

As part of the annual retail reanalysis (performed in August 2001) Time Series Analysis introduced an improved Easter correction method.

The date of Easter varies from one year to another. This variation may impact on the figures of a monthly survey. This impact is referred to as an Easter proximity effect.

The improved Easter correction method takes into account the proportion of activity in March and April which may be due to the date of Easter. A regression analysis is applied to test if the Easter related activity is significant. If the test is significant then an Easter proximity correction would be applied in the seasonal adjustment process to remove the Easter proximity effect. The improved Easter correction method gives a superior Easter correction when the dates for Easter fall on or around the March and April boundary (ie. approximately 25 March to 7 April). The next occurrence of an Easter proximity effect will occur in 2002 when Good Friday falls on March 29.

The application of the improved Easter correction method will cause revisions to the seasonally adjusted and trend estimates when compared with previous estimates. In practice, the seasonally adjusted and trend estimates for March and April would be expected to be revised along the length of the time series. Of course there are the other "normal" or "expected" revisions associated with the seasonal adjustment process which may affect all months of the series to varying degrees.

More details on the Easter correction method can be obtained from 'Adjusting for an Easter Proximity Effect': ABS Working Papers in Econometrics and Applied Statistics - Working Paper no.99/3 - December 1999 (Catalogue number 1351.0, www.abs.gov.au) or the journal article 'An Easter Proximity Effect: Modelling and Adjustment' in Volume 43, No. 3, September 2001 issue of the Australian and New Zealand Journal of Statistics.

For more information, please contact Craig McLaren on (02) 6252 6540 or Mark Zhang (02) 6252 5132.

email: craig.mclaren@abs.gov.au, mark.zhang@abs.gov.au