A tale of two bazaar economies:

1
A tale of two bazaar economies An
Input-Output analysis for Germany and Italy by
Emanuele Breda and Rita Cappariello (Banca
dItalia Economic and financial statistics
Department)
2
International relocation of production
ACQUISITION OF INTERMEDIATES FROM FOREIGN
SUPPLIERS (OUTSOURCING)
FDI (OFF-SHORING)

• both phenomena have similar macroeconomic
  implications
• a change in import composition
• a rising share of imports in the value of
  domestic production.

3
Main references (Europe and rest of the world)
Feenstra and Hanson (1996) in US the share of
imported inputs on the total purchase of
intermediate products grew from 5.5 per cent in
1972 to 11.6 in 1990. Egger and Egger (2003)
increase, on average, of the imported inputs as a
share of gross output for 11 European counties
between 1990 and 1997. Hummels et al. (1998,
2001) increasing share of imported goods and
services content in exports for some OECD
countries during the 1980s. ECB (2005) increase
of the import content of exports from 1995 to
2000 for an aggregate of economies of the euro
area.
4
Main references (input-output)
The indicators for international outsourcing are
usually obtained by using trade data and/or
information from the INPUT-OUTPUT TABLES

• The definition of international outsourcing
  underlying all these measures is the one in
  Feenstra and Hanson (1996)
• By outsourcing we mean the import of
  intermediate inputs by domestic firms.
• This definition is much broader than the one
  considering just trade resulting from specific
  relationships between a domestic firm and a
  supplier abroad (Grossman and Helpman, 2005).

5
Main references (input-output) (2)

• The advantages of this methodology
• to provide an exhaustive measure of the
  phenomenon, regardless of relationships between
  domestic producers and foreign suppliers.
• to avoid an arbitrary dichotomy between
  intermediate inputs and other categories of goods
  which is very common in trade statistics.

6
Main references (input-output) (3)

• However, these indicators
• if constructed on technical coefficients of
  production drawn from annual I-O tables and the
  trade data, are based on the strong assumption
  that the share of imported inputs is the same in
  all the industries (Feenstra and Hanson, 1996
  Amiti and Wei, 2005).
• if based upon direct data on imported inputs
  drawn from the symmetric I-O tables, use
  information having a significant delay since
  these tables are compiled only every 5 years
  (ECB, 2005).

7
References for Germany and Italy

• Germany Sinn (2003, 2006), after having observed
  a strong increase in industrial production and
  exports but a much slower growth in domestic
  value-added between 1995 and 2003, used the
  expression bazaar effect for German economy
• Belke et al. (2007) and Danninger and Joutz
  (2007) interpreted in a more optimistic fashion
  these same facts, concluding that Germany gained
  from internationalisation and intensified its
  trade links with fast-growing markets.
• Italy (manufacturing firms) Bugamelli et al.
  (2008) found a positive correlation between
  value-added growth and (lato sensu)
  internationalisation
• Barba Navaretti and Castellani (2004) and
  Federico and Minerva (2007) found no evidence of
  negative effects of internationalisation on the
  domestic employment level.

8
Motivations of the paper

• Answer to the questions
• Was the strong increase in outsourcing from 1995
  to 2000 biased by the cyclical peak of
  international trade in 2000?
• Is the Italian economy a bazaar economy too?

Measuring international outsourcing by a set of
different indicators for more recent years for
Italy and Germany
9
Motivations of the paper (continued)

• To reach our objectives
• We use instead the annual use I-O tables of
  imported inputs updated up to 2006 and available
  only for a limited groups of countries.
• This data have been used recently to evaluate the
  development of outsourcing in Italy (Falzoni and
  Tajoli, 2007 Daveri and Jona-Lasinio, 2007) but
  without any consideration on methodological
  issues regarding neither the measurement of the
  phenomenon nor any cross-country comparison.

10
International outsourcing concepts and measures
Four different groups of indicators for
outsourcing
Share of Imported Inputs on Total Inputs (IITI
broad and narrow)
Share of Value-Added on Production (VAP)
Import Content of Exports (ICE and DIICE)
Import Content of Production (ICP and DIICP)
11
International outsourcing concepts and measures
(2)

• Share of Imported Inputs on Total Inputs (broad
  version Feenstra and Hanson, 1996)

where mji and dji correspond respectively
to the imported and the domestically produced
inputs from industry j1,.., J used to produce
output in industry i1,..,n, and M and D are the
corresponding totals.
12
International outsourcing concepts and measures
(2)

• Share of Imported intra-industry Inputs on Total
  intra-industry Inputs (narrow version Bracci,
  2006, Falzoni and Tajoli 2007),

with mii and dii corresponding respectively
to imported and domestically-produced
intra-industry inputs (i.e. the diagonal elements
of matrices) miidii to total intra-industry
inputs of industry i M and D to total
intra-industry inputs of the whole economy.
13
International outsourcing concepts and measures
(3)

• Import Content of Domestic Production (Egger and
  Egger, 2003)

• Direct and Indirect Import Content of Domestic
  Production

with yi representing the gross output of
industry i 1,.., n, 0 makj 1 representing a
multiplicative coefficient of the imported input
from sector k that is embodied in the domestic
production of sector j, and subsequently utilised
as an input in industry i (dji), and Y being the
total output.
14
International outsourcing concepts and measures
(4)

• Import Content of Exports (Hummels and al., 2001
  Chen and al., 2005)

• Direct and Indirect Import Content of Exports
  (BCE, 2005)

with xi representing the exports of industry
i 1,.., n and X the total exports of the
economy.
15
International outsourcing concepts and measures
(5)

• Share of Domestic Value-Added on Production

where vi is the value-added of industry i1,.., n.
16
International outsourcing concepts and measures
(6)

• IITI (broad and narrow) indices focus on the
  firms choice between domestic and external input
  markets.
• ICP and DIICP intend to capture the firms
  substitution of domestic production with
  production phases relocated abroad. Since they do
  not measure IO (international outsourcing)
  regardless of the degree of vertical integration,
  they can be useful to evaluate the effects of IO
  on macroeconomic variables, but less reliable for
  comparing indices across industries.
• ICE and DIICE are measures of vertical
  specialisation (or proxies for measuring
  globalisation), that is the production chain
  that link different countries, acting either as
  producers in intermediate stages or as exporters.
• VAP index is decreasing in IO, and its ability
  to capture industries IO is less
  straightforward.

17
Data

• Two sets of annual input-output tables (Destatis
  and Istat) including the use matrices with a
  distinction between domestically produced inputs
  and imported inputs.
• Indices calculated including total purchases of
  non-energy products and market services (46
  NACE-sectors).

18
The extent of internationalisation
19
The extent of internationalisation (2)

• In 2006 the overall level of international
  outsourcing appears to be comparable between the
  two countries.
• The level of international outsourcing in market
  services in Germany is higher than in Italy
  because of a more intensive use of imported
  inputs in the financial and banking sector.
• Manufacturing sector seems to be more
  internationalised in Italy than in Germany.

20
The extent of internationalisation in
manufacturing

• For every 100 euros of goods produced in
  high-tech sectors, the direct ICP was 26 and 22
  euros respectively for Italy and Germany, whereas
  the DIICP was 33 and 29 euros.
• The low-tech sectors are the least
  internationalised ones in both countries.
• In 2006 the highest values of international
  outsourcing indicators are found in high and
  medium-tech industries.

21
The extent and development of international
outsourcing Total economy
22
The extent and development of international
outsourcing Manufacturing
23
The development of international outsourcing
24
The dynamics of internationalisation

• In the 1995-2000 period very intense growth in
  both countries, especially in Germany (and, for
  Germany in services, which, however, started from
  extremely low levels of internationalisation and
  have a limited weight on the whole economy).
• Between 2000 and 2004 weaker increase in Germany.
• During the same period we can observe a reduction
  for the whole Italian economy, which reflects the
  increasing relative importance of market services
  industries.
• In the last years under analysis (2005-06),
  growth resumed at a fast pace, especially in
  Italy.

25
A shift and share analysis

• The absolute variation of the indices have
  been broken down in two parts
• the change of intensity in industries
  international outsourcing (the within component).
  
• the change relative to the economy structure (the
  between component).
• The shift and share analysis is based on 46
  sectors (NACE codes) for the all period
  (1995-2006).

26
International outsourcing and sectoral change
between 1995 and 2006
27
International outsourcing and sectoral change
between 1995 and 2006 (2)

• For Italy the ss analysis confirms an increase
  in the international fragmentation for both the
  whole economy and the manufacturing sector. This
  rise is at least partially counterbalanced by a
  production shift towards less internationally
  integrated industries.
• For Germany, both component move in the same
  direction, showing a clear increase in the level
  of outsourcing.
• The firms propensity to use imported inputs grew
  within each sector at a similar pace in both
  countries, but the shift of the economic
  structure towards more internationalised sectors
  was stronger in Germany than in Italy. This holds
  for the all economy and for manufacturing.

28
Conclusions

• In 2006 Germany and Italy show a similar level of
  international outsourcing. Considering the
  manufacturing only, Italian firms seems to be
  ever more internationalised than the German ones.
  Therefore, we could consider, by a static point
  of view, the Italian economy as a bazaar
  economy too.
• Also the dynamic evolution was broadly similar in
  the two countries stronger in the years
  1995-2000, weaker (if not negative) during
  2001-04
• in the last years under study (2005-06) the
  growth of indicators resumed at a fast pace,
  especially in Italy. This seems to suggest a
  change in strategies and a reorganisation of
  production in Italian firms.
• However, German firms started from a lower level
  and experienced a stronger growth in the 1st
  sub-period and in the whole period.
• For both countries, the highest levels of
  international outsourcing are found in high and
  medium-tech sectors.

29
Conclusions (2)

• Italian lira crises in the first half of the 90s
  temporarily boosted price competitiveness of
  Italian goods, whereas corresponding Deutsche
  mark appreciation triggered German firms effort
  to compensate for the loss of competitiveness.
• In the beginning of the last decade, the decline
  of IO indicators for Italian economy was caused
  by a structural shift in the production towards
  services, while in manufacturing sectors IO
  continued to grow (albeit slowly).
• The growth of indicators in the last years
  analysed suggests that, before the international
  crisis of 2008-09, at least a significant part of
  Italian firms was trying to cope with the new
  challenges posed by globalisation, by the
  diffusion of ICT, and by the adoption of the euro.