Seasonal adjustment

Method

The series are made by seasonally adjusting monthly figures from the LFS. The levels are then corrected to make the yearly averages of the seasonally adjusted figures equal to the corresponding unadjusted figures from the LFS.

In order to reduce the uncertainty, the presented series are three months moving averages of the seasonally adjusted figures. For instance the figure from September is the average of the estimates from August - October.

Previous to the seasonal adjustment, adjustments are made for outliers and for holidays not falling on weekdays in the same month in the LFS every year.

Major change in the LFS from January 2006

With effect from January 2006, some improvements have been made in the Norwegian Labour Force Survey (LFS) due to new ILO recommendations. The changes improve the comparability with the EU countries. The changes consist of some minor revisions in definitions, adjustments in questioning about working hours, and the classification of the lay-offs and lowering the lower age limit from 16 to 15 years.

Comparable time series are necessary to get seasonally adjusted figures of good quality.

We make the 2006 LFS data for seasonal adjustment as comparable as possible with the rest of the time series by using 16 years as the age limit as well as the classification of lay-offs used prior to the changes. The seasonally adjusted employment figures are therefore without any breaks in the series, while the seasonally adjusted unemployment figures are partly adjusted for the changes in the LFS in 2006.

A slightly new way of questioning about working hours are gradually introduced in the sample for 2006. Still, the data on man-weeks worked for seasonal adjustments are based on the whole sample. The new way of questioning about working hours does not seem to systematically affect the figures for man-weeks worked that we are seasonally adjusting.

Seasonal adjustment of the Labour Force Surveys (LFS)

Seasonally adjusted figures describe the development in time series smoothed for normal seasonal variations caused by institutional, climatic or calendar matters. The purpose of adjusting for seasonal variations is to better describe the development during the last year, to give figures for the change between the last two 3-month periods, and to make the interpretation of the figures easier.

Method in use

X12ARIMA

About the input data

Due to changes in the LFS from 2006, the samples are modified to the lower age limit of 16 years and the classification of lay-offs is according to the pre-2006 definition, in order to make the time series as comparable as possible.

Indirect adjustments

We seasonally adjust 2 age-divided series (more/less than 24 year)

separately for employed persons and unemployed persons, and add the two series afterwards.

For man-weeks worked we seasonally adjust 3 gender- and age-divided series: persons 16-24 years, males 25-74 years and females 25-74 years.

Registered unemployed and registered unemployed + government measures are seasonally adjusted for 4 gender- and age-divided series separately.

Structure

Additive or multiplicative form and the ARIMA model for each series is selected by X12ARIMA automatically based on the data from 1997. This structure is not fixed.

Preadjustments

Regression models in X12ARIMA preadjust the series, where we define the explanatory variables. The series are preadjusted for different breaks in the series, outliers, and for holydays not falling on weekdays in the same month in the LFS every year. Specifically the series for man-weeks worked is preadjusted for many effects.

The employment and unemployment series, seasonally adjusted are preadjusted when Easter is in March.

The man weeks-worked series are also sensitive to single holydays, so they are also preadjusted for 2nd Easter day in March, May 1, May 17, Whit Monday, Ascending day, Dec31 and Jan1. We also take account of the effect when two of these days occur on the same date (Ascension Day on May 17 in 2007 and on May 1 in 2008).

Registered unemployed people at the Employment Offices and registered unemployed +government measures are only preadjusted for a break in the series due to a change in the definition of registered unemployed in January 1999. We let X12ARIMA determine the size of the level shift automatically for each of the subseries.

Outlier treatment

Outliers are detected automatically and removed by X12ARIMA

Construction of gender divided figures

Gender divided figures are constructed from seasonally adjusted figures by utilizing annual averages from the LFS.

This is done for employment and unemployment series, and for the man-weeks worked series for persons 16-24 years.

Revisions

Monthly. The series are updated every month. We let X12ARIMA estimate new seasonal weights and parameters in the preadjustment regression models every month. Therefore, also distant figures may be revised, but due to re-estimated seasonal weights, the changes diminish backwards in time. Any automatic change in the choice between additive or multiplicative form and ARIMA model may cause the figures to be revised.

Quarterly. The figures from the first two months in the quarter are preliminary. They are re-estimated at the end of the quarter with the extra interviewees that we get after the preliminary deadlines.

Yearly. After seasonal adjustment, the levels are corrected to make the annual averages of the seasonally adjusted figures equal to the corresponding unadjusted figures from the LFS. The annual averages of a year are not ready until we publish the 4^th quarter figures. At that point we use new factors of level correction. Until the 4^th quarter publication is available, preliminary level correction factors from last year have to be used.

Seasonally adjusted figures divided by age groups are broken down into gender-divided figures by utilizing annual averages from the LFS. When we publish the 4^th quarter figures, we use the new annual proportions by gender. Until the 4^th quarter publication is available, preliminary proportions by gender from last year have to be used.

More rare evaluation; Re-evaluation of the model structure and explanatory variables that are used in preadjustment regressions for each of the series.

Routines updating

Routines are updated continually.

Annual restrictions

After the seasonal adjustment, the levels are corrected to make the yearly averages of the seasonally adjusted figures equal to the corresponding unadjusted figures from the LFS.

Definition relations

We seasonally adjust employment and unemployment. The labour force, seasonally adjusted, is calculated as the sum of the seasonally adjusted employment and the seasonally adjusted unemployment.

Qualitative indicators

Q-values, M1-M11, and F-tests for stable and movable seasonality from X12ARIMA for each of the indirectly seasonally adjusted totals
Q -values, M1-M3, M5-M7 and F-tests for stable and movable seasonality from X12ARIMA for each of the sub series

Graphic evaluation

Publication

At http://www.ssb.no/akumnd_en/ 3-months moving averages of the seasonally adjusted totals are published. For instance the figure from November is the average of the estimates from October, November and December.

Persons in charge

Dag Rønningen or Jørn Ivar Hamre

About X12ARIMA

http://www.census.gov/srd/www/x12a/

Seasonal adjustment of the Labour Force Surveys (LFS)	Seasonally adjusted figures describe the development in time series smoothed for normal seasonal variations caused by institutional, climatic or calendar matters. The purpose of adjusting for seasonal variations is to better describe the development during the last year, to give figures for the change between the last two 3-month periods, and to make the interpretation of the figures easier.
Method in use	X12ARIMA
About the input data	Due to changes in the LFS from 2006, the samples are modified to the lower age limit of 16 years and the classification of lay-offs is according to the pre-2006 definition, in order to make the time series as comparable as possible.
Indirect adjustments	We seasonally adjust 2 age-divided series (more/less than 24 year) separately for employed persons and unemployed persons, and add the two series afterwards. For man-weeks worked we seasonally adjust 3 gender- and age-divided series: persons 16-24 years, males 25-74 years and females 25-74 years. Registered unemployed and registered unemployed + government measures are seasonally adjusted for 4 gender- and age-divided series separately.
Structure	Additive or multiplicative form and the ARIMA model for each series is selected by X12ARIMA automatically based on the data from 1997. This structure is not fixed.
Preadjustments	Regression models in X12ARIMA preadjust the series, where we define the explanatory variables. The series are preadjusted for different breaks in the series, outliers, and for holydays not falling on weekdays in the same month in the LFS every year. Specifically the series for man-weeks worked is preadjusted for many effects. The employment and unemployment series, seasonally adjusted are preadjusted when Easter is in March. The man weeks-worked series are also sensitive to single holydays, so they are also preadjusted for 2nd Easter day in March, May 1, May 17, Whit Monday, Ascending day, Dec31 and Jan1. We also take account of the effect when two of these days occur on the same date (Ascension Day on May 17 in 2007 and on May 1 in 2008). Registered unemployed people at the Employment Offices and registered unemployed +government measures are only preadjusted for a break in the series due to a change in the definition of registered unemployed in January 1999. We let X12ARIMA determine the size of the level shift automatically for each of the subseries.
Outlier treatment	Outliers are detected automatically and removed by X12ARIMA
Construction of gender divided figures	Gender divided figures are constructed from seasonally adjusted figures by utilizing annual averages from the LFS. This is done for employment and unemployment series, and for the man-weeks worked series for persons 16-24 years.
Revisions	Monthly. The series are updated every month. We let X12ARIMA estimate new seasonal weights and parameters in the preadjustment regression models every month. Therefore, also distant figures may be revised, but due to re-estimated seasonal weights, the changes diminish backwards in time. Any automatic change in the choice between additive or multiplicative form and ARIMA model may cause the figures to be revised. Quarterly. The figures from the first two months in the quarter are preliminary. They are re-estimated at the end of the quarter with the extra interviewees that we get after the preliminary deadlines. Yearly. After seasonal adjustment, the levels are corrected to make the annual averages of the seasonally adjusted figures equal to the corresponding unadjusted figures from the LFS. The annual averages of a year are not ready until we publish the 4^th quarter figures. At that point we use new factors of level correction. Until the 4^th quarter publication is available, preliminary level correction factors from last year have to be used. Seasonally adjusted figures divided by age groups are broken down into gender-divided figures by utilizing annual averages from the LFS. When we publish the 4^th quarter figures, we use the new annual proportions by gender. Until the 4^th quarter publication is available, preliminary proportions by gender from last year have to be used. More rare evaluation; Re-evaluation of the model structure and explanatory variables that are used in preadjustment regressions for each of the series.
Routines updating	Routines are updated continually.
Annual restrictions	After the seasonal adjustment, the levels are corrected to make the yearly averages of the seasonally adjusted figures equal to the corresponding unadjusted figures from the LFS.
Definition relations	We seasonally adjust employment and unemployment. The labour force, seasonally adjusted, is calculated as the sum of the seasonally adjusted employment and the seasonally adjusted unemployment.
Qualitative indicators	Q-values, M1-M11, and F-tests for stable and movable seasonality from X12ARIMA for each of the indirectly seasonally adjusted totals Q -values, M1-M3, M5-M7 and F-tests for stable and movable seasonality from X12ARIMA for each of the sub series Graphic evaluation
Publication	At http://www.ssb.no/akumnd_en/ 3-months moving averages of the seasonally adjusted totals are published. For instance the figure from November is the average of the estimates from October, November and December.
Persons in charge	Dag Rønningen or Jørn Ivar Hamre
About X12ARIMA	http://www.census.gov/srd/www/x12a/