Skógræktarritið - 15.05.2001, Side 200

Figure 1. Forest industries’ share of mainland Norway’s man-year empioyment
1962-99.
Regional Importance
In our assessment of the forest
industries' importance as an
employment factor in the coun-
ties relative to the national level,
localisation coefficients were
used. These indicate the impor-
tance of an industry in a region,
or how specialised the region is
in any industry, compared to the
national average. The localisa-
tion coefficient at point of time í
for industry i in the region r is
calculated as:
Here Erj represents employ-
ment in industry i at region rand
Pr the total employment in
region r. Ej represents employ-
ment in industry i at the national
level and P total overall employ-
ment. An LQ-value of 1 therefore
means that the industry has the
same representation (or impor-
tance) in the region as national-
ly. An LQ>1 (<1) can be inter-
preted as meaning that the
industry is relatively more impor-
tant (less important) in the
region than nationally.
Our analysis involves the com-
parison of data for forest-based
employment at the municipal
level aggregated to Statistics
Norway’s 101 prognosis regions.
Table 1 shows the regions in
which the total forest industries
are most important. For forestry
this is calculated on the basis of
national accounts data for nor-
mal man-years on the national
level (5400) and forest cutting on
the municipal level.
Winner and loser regions -
shift-share analysis
In this section we look at
changes in employment patterns
at the regional level, focusing at
the upswing in the early 1990s
(1990-95).
There are two possible ways of
defining employment winner and
employment loser regions. By (a)
looking at changes in absolute
values we will, for industries in
general growth, get "large"
regions at the top of the winner
list, even though these may have
had a significantly weaker rate of
growth than other regions. The
contrasting picture will show
large regions topping the list of
losers for industries undergoing
general recession even though
the region has coped relatively
speaking better than others. By
(b) only looking at relative
changes, both winner and loser
lists will be easily dominated by
regions often categorised as
insignificant and of little interest.
One way to combine these
methods is to perform a shift-
share analysis. This involves
splitting changes in absolute
value into a structure component
which tells us how large the
change would have been if an
industry in the region had had
the same relative change as the
national average, and a shift
component which is the differ-
ence between observed change
in absolute value and the struc-
ture component. The shift com-
ponent therefore expresses the
lost or gained market share cal-
culated as the number of jobs. If
an industry is in general reces-
sion, for example, regions that
show either progress or a minor
decline relative to the national
average will have a positive shift
component. if the industry is in
general growth, the shift compo-
nent will be positive only if the
region has a better percentage
development than the national
average.
More formally the shift-share
model can be written as
Table 1. Regions where forest industries are most significant in employment.
Prognosis region Forest industry total Locali Forestry sation coefficient Wood and wood products Pulp and paper
Flisa 8.3 11.5 13.4 0.0
H0nefoss 6.9 3.8 2.3 14.3
Sarpsborg 6.3 0.3 0.3 17.3
Halden 5.1 2.3 0.3 12.9
S0r-0sterdal 4.6 11.5 3.2 3.0
Egersund 4.5 0.2 9.6 0.0
Selhii/Tvrial 4.2 12 is QJ}
198
SKÓGRÆKTARRITIÐ 2001 l.tbl.