Labour force survey, seasonally-adjusted figures, May 2003

Norwegian unemployment has gradually increased since the autumn 1999, according to seasonally adjusted figures from the LFS. In per cent of the labour force the number of unemployed persons increased by 0.4 percentage points, which is outside the error margins. Seasonally adjusted figures for registered unemployed persons at the Employment Offices also increased during the same period.

After a period of strong growth the employment has just shown a slight increase from 1999, and in 2002 the number of employed persons started to fall. From the first to the second quarter 2003 the number of employed persons decreased by 12 000.

Increased unemployment also in the EU and the OECD

The seasonally adjusted unemployment rate in Norway increased from 4.1 to 4.6 per cent between February 2003 and May 2003. In the same period unemployment increased from 7.9 to 8.1 per cent as an average in the EU member countries. The unemployment rate in the OECD-countries as a whole went up from 7.0 to 7.2 per cent. In May 2003 the unemployment rate amounted to 6.1 in USA, 0.3 percentage point higher than in February 2003. The unemployment rate in Denmark increased from 5.0 per cent in February 2003 to 5.2 per cent in May 2003, and from 5.2 to 5.4 per cent in Sweden. In Finland the rate stayed at 9.0 per cent. In the same period unemployment increased both in France and Germany. This is according to figures from the OECD and Eurostat .

Man-weeks worked have shown a decreasing trend since the autumn 1999. The last 3- month period man-weeks worked decreased by 6 000, which is inside the margin of errors.

About the man-weeks estimates

Previous to the seasonal adjustment, we make adjustments for holidays that fall in random months every year.. Due to the fact that the holiday of With Monday in 2003 fell on a weekday, we have adjusted the figure of man-weeks worked up by 191 000 before seasonal adjustment. The adjustments are made so that the relative change in average weekly working hours from November 2002 to June 2003 is like last year's. In addition the holiday absence rates are set equal to the corresponding absence rates in the reference year 1991 in order to get comparable figures with the time series before the break in 1996. Before 1996 the Norwegian LFS was not a continuous survey, but based on one week each month.

In June in the reference year 1991 the survey week was the first week of the month, where the proportion on holiday is low. As a consequence, the proportion of the population on holiday in June the years after 1995, when all weeks were surveyed, was high compared with June 1991. Due to this, June has had too many leave substitutes since 1996. The employment figures for June this year were therefore adjusted downward by 24 000 before seasonal adjustment. For June last year the employment figures were pre-adjustment down by 25 900 leave substitutes.

Uncertainty

Quality tests show that the seasonally adjusted LFS unemployment figures are uncertain. The seasonal-adjustment method has problems identifying a stable seasonal pattern for this series. The random component is relatively large compared with the seasonal component. The figures should therefore be used with caution.

The purpose of adjusting for seasonal variations is to describe the development during the last year and to give figures for change between the last two 3-month periods, cleared for normal seasonal variations. In order to reduce uncertainty, the presented series are three months moving averages of the seasonally adjusted figures. For instance the figure for May is the average of the estimates from April, May and June.

Tables:

• Table 1. Employment and unemployment for persons aged 16-74 years, seasonally adjusted, three-months moving average. Absolute figures in 1000 and unemployed persons in per cent of the labour force.