Calendar adjustment is a statistical method for removing the calendar effect from an economic time series. The calendar effect is the variation caused by the changing number of particular week days or holidays in different months or other time periods (quarters, years).
Calendar adjustment is mainly used in the calculation of short-term statistics (STS), for converting gross (unadjusted/raw) figures or indices into their calendar adjusted equivalent. In order to adjust a figure or an index, the calendar nature of a given month is taken into account and calendar effects are removed, whatever their nature. The calendar effect may for example depend on:
- the timing of certain public holidays (Easter can fall in March or in April, depending on the year);
- the possible overlap of certain public holidays and non-working days (1 May can fall on a Sunday);
- the occurrence of a leap year.
Working-day adjustment is a part of the calendar adjustment which focusses on the changing number of working days (Monday - Friday) in the various months and their effect on statistical indicators (e.g. industrial production, production in construction) for these months.
It should be noted that the calendar adjustment can have effects on the yearly average of the index value in the reference year. Generally, the yearly average of the quarterly or monthly index values for the reference year is set to 100 for the unadjusted data (see article Short-term business statistics - compiling indices at European level). This average might be changed by the calendar adjustment. Consider a year in which several holidays fall on a weekend. In this year the number of working days is higher than in other years. The average calendar adjusted index values will therefore be a bit lower than 100 in the reference year since the production (or turnover) was generated with a higher input of working days.
The effect of the calendar adjustment on the yearly average of the index values in the reference year is generally quite small. It depends on the calendar constellation in the reference year and the effect one additional working day has. As a rule the variation between the yearly average of the adjusted and the unadjusted index should not be more than +/- 2 percentage points.
Since seasonal adjustment also includes the calendar adjustment (as in the case of Eurostat STS data, where the calculation of the seasonally adjusted data is based on calendar adjusted data) the calendar adjustment effect on the yearly index average is also visible in the seasonally adjusted data.
It is possible to eliminate the effect after the calendar adjustment by re-referencing the results from the calendar adjustment again to an average of 100 in the reference year and some Member States of the EU do this while others prefer to show the effect in their data as an additional information to users.
The decision of re-referencing or showing the additional information has only an impact on the level of the time series, but has no effect on the growth rates.