Vinnumarkaður - 01.06.1995, Side 29

Synopsis of methods and concepts
27
3.1.5. Weighting, estimation of totals and annual averages
Annual averages are calculated from both surveys within the
year. Totals and percentages are estimated by applying
weights to each case. Weights are calculated with Equation
2.1, where
N = mean population aged 16-74 years,
n = number of respondents,
n* = number of sample individuals with domicile in
Iceland but residence abroad,
k = sex of respondent,
a = agegroupofrespondent {16, 17,18-19, 20-24,25-29,
.., 65-69, 70-73, 74 years) and
m = (April survey, November survey}.
3.2 Unemployment registration and estimated
number of man-years
3.2.1 Origin of data
The number of unemployment days, as registered at the
public unemployment agencies, is collected each month by
the Social Ministry - Office of Labour. These data are
published each month in a special bulletin Yfirlit um
atvinnuástandið [Employment Overview]. Since 1986 Sta-
tistics Iceland has also collected quarterly data from the
same source on the number of registered unemployed per-
sons, by length of registered unemployment, sex and age
groups. The reference points are the end of February, May,
August and November each year. From 1986 to February
1988 the reference day was the last Friday in the respective
month. Since May 1988 the reference day has been the last
working day in the reference month. The National Economic
Institute (NEI) estimates the number of man-years each
year.
3.2.2 Coverage and reliability
Registered unemployment. Registration at a public unem-
ployment agency is a prerequisite for unemployment com-
pensation. Not all unemployed persons make use of this
right. Self-employed persons and persons not previously in
employment, such as homemakers and students, have lim-
ited compensation rights. Persons may also be registered
without being available for work. These problems diminish
the value of using registration data to estimate true unem-
ployment.
Estimated number of man-years. The NEI's estimation
of number of man-years is used to calculate relative unem-
ployment as registered by public employment agencies.
These estimates derive from data on the number of working
weeks supplied by the tax authorities. As the production of
these figures tends to be late, the number of man-years in
1993 and 1994 are the NEI’s own estimates. In 1992 the
number of working weeks of employees was counted in
selected firms. The total wage outlay for these firms was
more than 60% of all wage outlays that year. The number of
working weeks for the remaining employees was estimated
by taking into account average wages per economic sector.
The number of working weeks of own-account workers was
estimated by using data from monthly tax returns.
The reference period for the number of man-year is the
whole year. The monthly variations are calculated by using
coefficients which simulate seasonal variations. The distri-
bution by sex and region for the years 1993 to 1994 is
calculated by using the same proportional distribution as in
1992.
Number of registered unemployed persons by length of
unemployment, sex and age group. Between 1986 and May
1993 these data were collected from municipalities with
more than 500 inhabitants. From August 1993 all municipali-
ties registering unemployment separately have reported to
SI.
Age groups are defined by the calendar year. In February
and May information is collected for the number of unem-
ployed persons who will be 16 years or older at the end of the
year. In August and November the age limit is persons who
will be 15 years or older at the end of the year.
Until August 1993 agency non-response has been higher
for the quarterly data than for the monthly data of registered
unemployment days.
3.2.3 Concepts
Number of man-years. The number of full-time job equiva-
lents, estimated from the number of working weeks during
the year. Two part-time jobs, totalling 26 weeks each,
constitute one man-year. Registered unemployment is cal-
culated in a similar way and added to the number of full-time
equivalents.
Average monthly unemployment. Average monthly
unemployment is calculated as the number of unemploy-
ment days divided by the average number of working days
per month (21,67).
Length of registered unemployment. The number of
successive weeks an individual has been registered as
unemployed on the reference day. Loss of compensation
rights does not interrupt the count of weeks in unemploy-
ment. Casual work or part-time work does not affect the
count of unemployment weeks unless this leads to the
individual’sremoval from the registerfor one weekor more.
3.3 Labour force projection
3.3.1 Origin of data
Every three years Statistics Iceland creates population pro-
jections based on various combinations of mortality rates,
fertility rates and net migration. The latest projection dates
from 1991. One variant of the population projections is
published in Landshagir [Statistical Abstract of Iceland].
The present labour force projection is based upon the
population projection published in Landshagir 1993, and
the labour force surveys 1991-1993.
3.3.2 Projection assumptions
The labour force projection is based on the following
assumptions:
Mortality rates decrease by 0.5% per year up to the year
2000, based on the average mortality rates of 1985-89 taken
as mortality rates for 1987. Fertility rates are assumed to
decrease to a total fertility rate of 1.9 by 1997. Migration
rates for 1986-90 are applied to external migration, modi-
fied as to result in a net migration of zero.
Throughout the period, the total number of economically