change in the industry structure related to the
increasing share of ICT in these countries.
Regarding utilities and business sector
services, only hotels and restaurants and
financial intermediation remained spatially
concentrated during the second half of the
1990s. These results relate to the relative
importance of Balassa indices in tourism in
Spain, Austria, Italy and Greece27 on the one
hand and the strong weight of the financial
sector in Luxembourg on the other hand. While
changes in concentration are usually slow, they
have been discernible for financial
intermediation and construction in some
countries. Where construction is concerned,
this may reflect a strong need for
infrastructures and the expansion of housing
investment in some of these countries stemming
partly from the decline in interest rates in these
countries associated with the EMU process.
Comparing the two periods 1985-1989 and
1996-2001, manufacturing industry appears to
be somewhat more evenly spread (standard
deviation declines from 0.04 to 0.03). This
reflects opposing forces at work at the sector
levels. Sectors with a low to medium
technology content and/or low growth
prospects and highly tradable products have
achieved productivity gains mainly through
consolidation or concentration - this
consolidation process making it possible to
inject additional technological content. This has
resulted in changes of concentration indices at
the country level and in an increase in the
standard deviation (see textiles, for example).
On the other hand, sectors with fewer tradable
goods - requiring proximity between the
manufacturing base and the consumer markets -
have remained more evenly spread (see food
products, for example). The relative weight of
the sectors with a location driven by the “home-
market bias” has naturally increased faster in
countries where this market segment was
converging towards the average euro area
pattern (see transport equipment in Spain and
Portugal, for example). Finally, sectors with a
high technology intensity are widespread across
the euro area. This highlights the fact that the
location of innovative sectors depends on
multiple parameters, and that concentration
dynamics usually do not shape emerging
sectors, but play a greater part later, when the
new sectors enter the maturity phase (for
example, the standard deviation of “high tech
intensity sectors” changed slightly from 0.05 to
0.04, reflecting the fact that ICT innovations
shaped the 1990s more than the 1980s).
Similar developments have been experienced by
some of the EU acceding countries, as described
in Box 2.
27 Data for hotels and restaurants are missing for Greece for the
1980s.
Box 2
EVOLUTION OF SPECIALISATION AND CONCENTRATION IN THE NEW MEMBER STATES
Since the start of the transition in the late 1980s, the countries of central and eastern Europe
have experienced fundamental restructuring processes leading to sectoral convergence towards
western European patterns. Nevertheless, despite the relative and absolute decline of
agricultural output and the increasing importance of services during the 1990s, the share of the
primary (tertiary) sector is still higher (smaller) in the CEEC-8 than in the EU-15. The relative
reduction of industrial production led to output shares matching those of the EU, with the
exception of the higher shares in the Czech Republic.
Compared to the EU-15, specialisation is more pronounced in the CEEC-8. Mirroring western
European development, the majority of CEECs became more specialised in the second half of the
20
ECB
Occasional Paper No. 19
July 2004