Home Uncategorized Week-1

Uncategorized

Week-1

After reading chapter-1 from the attached text book explain about the research topic ‘Globalization of Production’.

Answer should be based on the attached three journal articles and text book.

Please answer in own words and as thorough as possible. APA format is must

APA format mustNo Plagiarism

Agricultural Economics Research Review
Vol. 24 July-December 2011 pp 301-308

* Author for correspondence,
Email: sangles@rediffmail.com

Impact of Globalization on Production and Export of Turmeric
in India – An Economic Analysis

S. Angles*, A. Sundar and M. Chinnadurai
Department of Agricultural Economics, Tamil Nadu Agricultural University,

Coimbatore – 641 003, Tamil Nadu

Abstract

India is a major supplier of turmeric to the world with more than 60 per cent share in turmeric trade. The
production and export performance of turmeric in India have been examined using secondary data for the
period from 1974-75 to 2007-08 and exponential form of growth function has been used for the analysis.
The growth in production and export of turmeric has been reported significant, because of the high
demand coupled with inflation. Instability index has been worked for the production and export for pre-
liberalization and post-liberalization periods. Instability has been observed high for production, export and
prices of domestic and international markets and domestic and international prices have shown high
integration. For the assessment of direction of trade, the Markov chain model has been used. The data
regarding country-wise export of turmeric has shown that the previous export share retention for Indian
turmeric has been high in minor importing countries (pooled under others category) (87 %), followed by
UAE (49 %), Iran (41 %) and UK (35 %). The countries such as USA and Japan have not been the stable
importers of Indian turmeric. The plans for export may be oriented towards these two countries and also
plans should be formulated for stabilizing the export of turmeric to other countries. The farmers should be
provided training on production of a quality product.

Key words: Turmeric, Export of turmeric, Indian turmeric, Markov chain model

JEL Classification: Q13, Q17

Introduction
India is popularly known as the “Spice Bowl of the

World” as a wide variety of spices with premium quality
is grown in the country since ancient times. In Vedas,
as early as 6000 BC, scruples evidences are available
regarding various spices, their properties and utility.
Among the commodities that were traded during that
period, spices occupied a major portion due to thei

superior quality and diversity which attracted foreigners
to India. Turmeric — the Golden Spice — is widely
cultivated in different countries such as India, China,
Myanmar, Nigeria, Bangladesh, Pakistan, Sri Lanka,
Taiwan, Burma, Indonesia, etc. Among these countries,

India occupies the first position in area, viz. 1,75,300
ha and also in production, viz.7,94,400 tonnes during
2007-08. In India, turmeric is grown in its 18 states.
The states like Andhra Pradesh, Tamil Nadu,
Karnataka, Orissa and West Bengal are the major
turmeric-producing states in India. The major countries
that export turmeric are: India, China, Myanmar and
Bangladesh. Indian turmeric fetches a premium price
due to its superior quality in the international market.
India has occupied around 60 per cent of the world
trade in turmeric.

Raveendran and Aiyaswamy (1982) had analysed
the growth in quantity exported and export prices of
turmeric in India. They had observed a cyclic pattern
of variation in prices with a length of three to seven

302 Agricultural Economics Research Review Vol. 24 July-December 2011

years. They had also found a high correlation between
export price and domestic price of turmeric. Mamatha
(1995) has estimated the growth rates of production
and export of selected spices including turmeric. She
has observed a positive growth rate in respect of
production and export of these spices. Kumar and
Sankaran (1998) have analysed the instability in turmeric
production in India and have concluded that decrease
in area instability has been compensated by the marginal
increase in the yield instability during 1980s. The
resulting reduction in production instability indicated that
the yield instability was the dominant factor
compensating production instability. Nair (2002) has
studied the impact of monsoons on the prices of the
spices in two states — Andhra Pradesh and Karnataka,
which are the leading suppliers of chilli, turmeric and
ginger as raw materials for the processing industries.

India’s share of 90 per cent of world trade in
turmeric during the pre-liberalization period was
drastically reduced to 60 per cent during 2007-08.
During this era of globalization, it is imperative to re-
assess the nations supplying potential, domestic and
international demand scenarios and export potential.
Keeping in view the above points, the present study
has analyzed production, price behaviour and export
potential of turmeric in India. The specific objectives
of the study were: to estimate the growth and instability
in area, production, productivity and export of turmeric
in India, to study the extent of price integration of
turmeric in domestic and international markets, and to
analyze the direction of trade of turmeric in India.

Data and Methodology
The study was mainly based on the secondary data

from various sources, which included Annual Reports,
Yearbooks, Statistical Data publications of Spices
Board, Indiastat.com, Ministry of Commerce and
Industries and Arecanut and Spices Development
Board. The study period was divided into two sub-
periods, viz. pre-liberalization (1974-75 to 1990-91) and
post-liberalization (1991-92 to 2008-09).

Compound Growth Rate

The annual compound growth rates for area,
production, productivity and export of turmeric were
computed separately for the two sub-periods and
compared in the form of Equation (1):

( )T ut oY =Y 1+r e …(1)

where,

Yt = Value at time’t’,
Yo = Initial value,
r = Growth rate,
T = Time in years; 0, 1, 2 …….., n, and
u = Random error-term.

Instability Index

The instability associated with turmeric area,
production, yield, export quantity, value and unit value
of export, domestic and international market prices was
estimated using the Instability Index of the form:

Instability Index = STDEV of ln (Yt+1/Yt)
…(2)

where,

STDEV = Standard deviation,

Yt = Crop area / production / yield / export quantity
/ export value / export unit value in the current
year, and

Yt+1 = Crop area / production / yield / export quantity
/ export value / export unit value in the next
year.

This index is unit free and very robust and it
measures deviations from the underlying trend (log linear
in this case). When there are no deviations from the
trend, the ratio Yt+1/Yt is constant, and thus standard
deviation in it is zero. As the series fluctuates more,
the ratio of Yt+1/Yt also fluctuates more, and standard
deviation increases.

Direction of Trade

The structural change in exports was examined
using the Markov chain approach. Central to Markov
chain analysis was the estimation of the transitional
probability matrix P. The element Pij of this matrix
indicates the probability that exports will switch from
country i to country j with the passage of time. The
diagonal Pij measures the probability that the export
share of a country will be retained. Hence, an
examination of the diagonal element indicates the
loyality of an importing country to a particular country’s
export.

Angles et al. : Impact of Globalization on Production and Export of Turmeric in India 303

In the context of the current application, there were
five main turmeric importing countries. The average
export to a particular country was considered to be a
random variable which depended only on its past
exports to that country and which could be denoted
algebraically as Equation (3):

Ejt = ∑ Eit-1 Pij + ejt …(3)
i=1

where,

Ejt = Exports from India during the year t to jth
country,

Eit-1 = Exports to ith country during the year t-1,

Pij = The probability that exports will shift from ith
country to jth country,

ejt = The error-term which is statistically independent
of Eit-1, and

r = The number of importing countries.

The transitional probabilities Pij, which can be
arranged in a (c × r) matrix, have the following
properties:

0 ≤ Pij ≤ 1

n
∑ Pij=1, for all i
i-1

Thus, the expected export shares of each country during
period t were obtained by multiplying the exports to
these countries in the previous period (t–1) with the
transition probability matrix.

The transitional probability matrix is estimated in
the linear programming (LP) frame work by a method
referred to as minimization of Mean Absolute Deviation
(MAD). The LP formulation is stated as

Min O’ P* + Ie

Subject to

X P* + v = y

GP* = I

P* ≥ 0

where, P* is a vector of the probabilities Pij, 0 is a
vector of zero, I is an appropriately dimensioned vector
of country, e is the vector of absolute errors (|U|), y is
the vector of exports to each country, x is a block
diagonal matrix of lagged values of y, and v is the vector
of errors and G is a grouping matrix to add the row
elements of P arranged in P* to unity.

Results and Discussion

Growth Analysis of Area, Production and
Productivity of Indian Turmeric

A perusal of Table 1 reveals that in all the periods,
the growth rates of production were higher than of
productivity and area. In all the periods, turmeric had
the productivity-led growth. The growths in area,
production and productivity were found higher during
pre-liberalization period than post-liberalization or overall
period. The lower growth in area and productivity in
post-liberalization period might be due to stability in area
under turmeric, i.e. no scope to allocate more area under
new planting. Growth recorded in all periods was
significant at one per cent level, except in the
productivity during the post-liberalization period, which
was significant at 5 per cent level.

India virtually has a monopoly in supplying of
turmeric to the world with a share of about 78 per cent
in the total global output and 60 per cent in the global
trade. Favourable weather conditions prevailing in the
major turmeric growing areas in the country (Andhra
Pradesh, Tamil Nadu, Orissa, Karnataka and West
Bengal) and the important steps taken by the Spices
Board, such as providing drying sheets to small and
marginal growers of turmeric and other spices for drying

Table 1. Compound growth rates of area, production and productivity of Indian turmeric

Year Area Production Productivity

1974-75 to 1990-91(Pre-liberalization period) 3.41* 7.97* 4.40*
1991-92 to 2007-08(Post-liberalization period) 1.88* 3.42* 1.50**
Overall1974-75 to 2007-08 2.74* 5.86* 3.04*

Note: *, ** denote significance at 1 per cent and 5 per cent levels, respectively.

304 Agricultural Economics Research Review Vol. 24 July-December 2011

under hygienic conditions, providing subsidies for the
small and marginal farmers for the construction of
concrete drying yards and warehouses, organization
of educational programmes for growers on improved
technologies, have led to increased productivity of
turmeric. Besides, release of high-yielding varieties over
the years also has made a significant contribution.

Growth Analysis of Turmeric Export from India

The growth in turmeric export in both quantity and
total value was found to be higher in the post-
liberalization period than pre-liberalization or overall
period (Table 2). It implies that a higher quantity of
turmeric is being exported after export liberalization,
which reduced the unit price of turmeric. The growth
in unit value has been found to be higher in pre-
liberalization than to post-liberalization period. The main
hurdle to turmeric export was the quality; if quality was
maintained; the growth rate would have increased by
many fold. Turmeric being a multi-use product of natural
origin, it is used in many fields such as culinary, medicine,
cosmetics and textiles.

Though a lower growth in unit value was seen in
post-liberalization period, the total export value of
turmeric export had the growth rate of 9.89 per cent
per annum which showed the rise in demand for
turmeric. The growth rate of the total export value was
higher due to higher growth of the export quantity of
turmeric. The overall export quantity, export value and
unit value of turmeric exported were significant at one
per cent level over the study period.

About less than 10 per cent of the turmeric
produced in the country is exported. The main hurdle
to the export is the quality, because the processing is
not done properly. For getting a good quality product,
there is the need of adoption of improved technologies,
such as stream boiling and mechanical drying instead
of conventional cooking and sun drying. If proper
processing and pre-limitation of pesticide residue is
maintained, then there would be ample scope for
increasing export in the years to come.

Growth Analysis and Correlation Coefficients of
Domestic and International Prices of Turmeric

The growth rates and correlation coefficients of
domestic and international prices of turmeric have been
presented in Table 3. The results revealed that during
pre-liberalization period the domestic market prices had
a high growth rate of 10.47 per cent per annum,
whereas the international prices had a growth rate of
11.70 per cent per annum. The coefficients of domestic
and international prices were significant at one per cent
level. However, during post-liberalization period, the
growth rate in turmeric price was estimated to be 2.08
per cent per annum for domestic prices and 5.71 per
cent per annum for international prices. These were
lower compared to the post-liberalization period. This
is due to higher quantity of export due to liberalization
which brought down the unit price of turmeric exported.
Correlation studies indicated that the domestic prices
were positively associated with the international market
prices (r = 0.96) during pre-liberalization period. During

Table 3. Compound growth rates and correlation coefficients of domestic and international prices of turmeric

Year Domestic International Correlation
market price market price coefficient

1974-75 to 1990-91 (Pre-liberalization period) 10.47* 11.70* 0.96
1991-92 to 2007-08 (Post-liberalization period) 2.08 5.71* 0.64
Overall (1974-75 to 2007-08) 8.77* 9.18* 0.90

Note: *Significant at one per cent level

Table 2. Compound growth rates of export quantity, total value and unit value of Indian turmeric

Year Export quantity Export value Unit value

1974-75 to 1990-91 (Pre-liberalization period) 1.34 8.13* 6.70*
1991-92 to 2007-08 (Post-liberalization period) 5.74* 9.89* 3.92*
Overall (1974-75 to 2007-08) 5.26* 12.70* 7.08*

Note: * Significant at one per cent level

Angles et al. : Impact of Globalization on Production and Export of Turmeric in India 305

the post-liberalization period also, the domestic prices
had a positive correlation but with a lower degree with
international prices (r = 0.67). There were many factors
other than price, which affected the international and
domestic prices during post-liberalization period.

Instability Analysis of Turmeric Production and
Trade in India

The instability index was worked out for turmeric
production and trade in India for the three periods to
analyse the extent of instability. It was observed from
Table 4. that the production was almost stable in all the
periods compared to area and productivity. The
fluctuations in yield of turmeric were mainly influenced
by the rainfall and other climatic factors. The release
of new varieties and innovative cultural practices
developed in recent years were also responsible for
the variations in productivity, which affected the levels
of production in different years. The fluctuations in the
export quantity of turmeric were very high during the
pre-liberalization period (0.49), whereas during the post-
liberalization period, there is less instability (0.15). This
indicates that the export growth during post-liberalization
did not fluctuate much due to less restrictions and
growing demand of Indian turmeric.

The instability in total value of turmeric export is
very high during pre-liberalization period (0.57)

compared to post-liberalization period (0.19). Compared
to the quantity and value, the unit value showed a lower
instability during the pre-liberalization period, but the
unit value became stable during the post-liberalization
period. Moreover, there are no stiff competitors in the
international market for turmeric due to comparative
advantage or agro-climatic advantage. Even though
there was not much fluctuation in domestic prices of
turmeric in all the periods, the extent of variation was
relatively high (0.39) in post-liberalization period. This
might be due to the changing demand for turmeric
products in foreign countries. The instability in
international price of overall period is very high (0.30)
compared to pre-liberalization and post-liberalization
periods. These result implied that there was a high
instability in pre-liberalization than post-liberalization
period.

Direction of Trade of Turmeric Export from India

The transitional probability, presented in Table 5,
depicts a broad idea of change in the direction of trade
of Indian turmeric. The five major countries which
imported Indian turmeric were: UAE, USA, UK, Iran
and Japan. The export to remaining countries was
pooled under the category of other countries. It can be
seen from Table 5 that USA was not a stable importer
of Indian turmeric even though the quantity imported

Table 5. Transitional probability matrix of Indian turmeric export: 1989-90 to 2007-08

Country USA UK Iran Japan UAE Others

USA 0.00 6.97 0.00 25.88 42.28 24.87
UK 0.00 35.20 32.77 6.43 25.60 0.00
Iran 0.00 0.00 41.88 8.20 0.00 49.92
Japan 83.89 0.00 0.00 0.00 16.11 0.00
UAE 2.22 8.15 13.43 12.03 49.14 15.03
Others 2.00 3.05 0.00 3.12 4.28 87.55

Table 4. Instability in Indian turmeric production and trade

Particulars Pre-liberalization period Post-liberalization period Overall period
(1974-75 to 1990-91) (1991-92 to 2007-08) (1974-75 to 2007-08)

Area 0.09 0.12 0.10
Production 0.21 0.20 0.20
Yield 0.14 0.18 0.16
Export quantity 0.49 0.15 0.36
Export value 0.57 0.19 0.42
Unit value 0.33 0.18 0.27
Domestic price 0.39 0.34 0.34
International market 0.20 0.17 0.30

306 Agricultural Economics Research Review Vol. 24 July-December 2011

by USA was higher. The USA would lose its share of
42.28 per cent to the UAE, 25.88 per cent share to
Japan and 24.87 per cent share to other countries, even
though USA gained considerable share from Japan
(83.89 %). In future, its share may be reduced from
the total turmeric traded from India. The countries such
as China give a stiff competition to India in turmeric
trade. The UK was found to be one of the stable
importers of Indian turmeric because it retained its
original share of around 35.20 per cent over the period.
It lost its major share of 32.77 per cent to Iran and
25.60 per cent to UAE.

Iran is another stable importer of Indian turmeric
because it retained its original share of 41.88 per cent.
It lost its major share to other countries to the extent of
49.92 per cent. It gained from the share of UK to the
extent of 32.77 per cent and 13.43 per cent from UAE.
Hence, in future Iran will be one of the most stable
importer and its growth may be higher in turmeric import
from India. Japan has not retained its original share
and it lost a major share of 83.89 per cent of its original
share to USA, followed by UAE (16.11 %). It gained
25.88 per cent from USA, followed by UAE (12.03%).
Hence, Japan may not be regarded a stable importer
of Indian turmeric in future. The reason may be that it
imports turmeric from Burma and Thailand.

The UAE has retained 49.14 per cent of its original
share and it is a stable importer of Indian turmeric. It
lost its major share to other countries category (15.03%)
and to some extent to Iran, Japan, UK and USA. But it
gained high share of 42.28 per cent from USA, followed
by UK (25.6 %) and Japan (16.1%). Being a major
importer of Indian turmeric, if it loses its share, it will
create a high instability in the export of turmeric from
India in future. The countries pooled under the other
category retained 87.55 per cent of its original share,
which implied that even though they import in lower
quantities, there is high stability, they have retained most
of its original share. It gained 49.92 per cent of the
Iran share, 24.87 per cent of USA share and 15.03 per
cent of UAE share. Hence, compared to major
importing countries at present, the countries pooled
under ‘others category’ would import more turmeric
from India in near future.

Thus, it is clear that the countries pooled under
‘others category’, UAE, Iran and UK would be the
stable importers of the Indian turmeric in future and
countries like USA and Japan are not the stable

importers. Hence, it would be necessary to give more
stress on the USA and Japan. The plans for export
should be oriented towards these two countries and
also plans should be formulated for stabilizing the export
to other countries. Mamatha (1995) has assessed the
direction of trade of turmeric. The countries such as
UAE (25 %), UK (65 %) and Singapore (15.71 %)
were the stable importers of the Indian turmeric. The
countries such as Japan, USA and Iran were found to
be not stable importers of Indian turmeric. The results
endorsed the present study, except in the case of Iran,
which indicated retention of 41.88 per cent of its original
share in the present study. Singapore was stable
importer, in past studies which had retained 15.71 per
cent. But, in the present study its share reduced
drastically and hence it was clubbed under other
countries category. The reason may be that Singapore
imports turmeric from its neighboring countries such
as Thailand and Burma where the cost was
comparatively lower than of Indian turmeric.

The countries pooled under ‘others category’ had
87.55 per cent of the retention of its original share in
the present study, which was 74 per cent in the earlier
study, which implied that the retention of the countries
pooled under ‘other category’ gained its original share
over the period. The reasons may be that in many areas
such as, food, textiles and cosmetics, turmeric is being
replaced by synthetic chemicals, as a coloring agent.
In medicine, turmeric is a naturally available medicine
at a lower cost. As a result, the retention of its original
share was increasing over the period. The other reasons
are that the properties of turmeric are being explored
continuously and its usage is increasing along with the
demand for fast food shops in the major importing
countries.

Projections of Indian Turmeric Export to Major
Importing Countries

The projection of the Indian turmeric export to
different countries was computed using the transitional
probability matrix and the results of actual and projected
exports of Indian turmeric have been presented in Table
6. The market share projections of turmeric exports to
different countries have been computed up to 2020.

Even though the total quantity increased, the
percentage share of actual and estimated export of
turmeric to USA declined between 1999-00 and 2007-
08. However, the projected value suggests that the

Angles et al. : Impact of Globalization on Production and Export of Turmeric in India 307

Table 6. Actual and projected exports of Indian turmeric to major importing countries
(in tonnes)

Year USA UK Iran Japan UAE Others

Actual Estimated Actual Estimated Actual Estimated Actual Estimated Actual Estimated Actual Estimated

1999-00 2427 2783 1676 2074 2077 2086 1878 2355 8162 6455 21555 21543
(6.43) (7.46) (4.44) (5.56) (5.50) (5.59) (4.97) (6.31) (21.61) (17.31) (57.06) (57.76)

2000-01 2584 2188 1837 2081 2971 2515 3027 2560 6044 6691 28165 21741
(5.79) (5.79) (4.12) (5.51) (6.66) (6.66) (6.78) (6.78) (13.54) (17.71) (63.11) (57.55)

2001-02 2739 3236 1842 2177 2724 2658 2559 2636 5272 6226 22641 27694
(7.25) (7.25) (4.88) (4.88) (7.21) (5.96) (6.78) (5.91) (13.95) (13.95) (59.93) (62.06)

2002-03 3914 2717 2006 1959 949 2453 2614 2391 4724 5602 18196 22657
(12.08) (7.19 (6.19) (5.19) (2.93) (6.49) (8.07) (6.33) (14.58) (14.83) (56.16) (59.97)

2003-04 3880 2662 2060 1918 488 1689 2694 2355 7239 5689 20683 18089
(10.47) (8.21) (5.56) (5.92) (1.32) (5.21) (7.27) (7.27) (19.54) (17.56) (55.83) (55.83)

2004-05 2508 2835 2576 2216 800 1851 2686 2693 5215 7044 29312 20406
(5.82) (7.65) (5.98) (5.98) (1.86) (5.00) (6.23) (7.27) (12.10) (19.02) (68.02) (55.09)

2005-06 2635 2955 2772 2399 1447 1879 2608 2422 7361 5970 29582 27472
(5.68) (6.86) (5.97) (5.57) (3.12) (4.36) (5.62) (5.62) (15.86) (13.85) (63.75) (63.74)

2006-07 2461 2943 2896 2660 6095 2503 2632 2787 7824 7127 29593 28385
(4.78) (6.34) (5.62) (5.73) (11.83) (5.39) (5.11) (6.01) (15.19) (15.36) (57.46) (61.17)

2007-08 2649 2973 2461 2730 3709 4552 2797 3187 5151 7317 32485 30741
(5.38) (5.77) (5.00) (5.30) (7.53) (8.84) (5.68) (6.19) (10.46) (14.21) (65.96) (59.69)

2008-09 3110 2460 3110 2780 6122 31726
(6.31) (4.99) (6.31) (5.64) (12.42) (64.34)

2009-10 3103 2548 2931 2944 6759 31024
(6.29) (5.17) (5.94) (5.97) (13.71) (62.92)

2010-11 3240 2609 2970 2988 7088 30414
(6.57) (5.29) (6.02) (6.06) (14.37) (61.68)

2015-16 3373 2690 3242 3131 7579 29294
(6.84) (5.46) (6.57) (6.35) (15.37) (59.41)
2020-21 3391 2695 3278 3146 7624 29175

(6.88) (5.47) (6.65) (6.38) (15.46) (59.17)

Note: Figures within the parentheses indicate percentage to total.

percentage of quantity would slightly increase from 5.77
per cent in 2007-08 to 6.88 per cent by 2020-21 AD. In
the case of UK, the actual export had increased from
1999-00 to 2007-08 and the estimated value showed
that the share of UK was increased for the same period
and the projected market share is expected to increase
marginally during 2007-08 to 2020-21 from 5.30 per
cent to 5.47 per cent. In the case of Iran, the actual
and estimated export had increased from 1999-00 to
2007-08. The projected market share was expected to
decrease marginally from 8.84 per cent to 6.65 per
cent during 2007-08 to 2020-21. In the case of Japan,
the actual and the estimated value export had increased
marginally between 1999-00 and 2020-21. The actual
export share of turmeric to UAE had decreased
drastically from 21.61 per cent in 1999-00 to 10.46 per
cent in 2007-08. However, the estimated value
decreased to the extent of 17.31 per cent in 1999-00 to

14.21 per cent in 2007-08. The projected market share
is expected to increase from 14.21 per cent to 15.46
per cent during 2007-08 to 2020-21. The actual export
share to the countries pooled under ‘others’ showed
an increase during 1999-00 to 2007-2008 from 57.06
to 65.96 per cent respectively. The projected market
share was expected to increase from 57.76 per cent to
59.69 per cent from the year 1999-00 to 2007-08, but
the share is expected to reduce to less extent from
2007-08 to 2020-21.

Keeping in view of the foregoing discussions, more
stress has to be given on the countries such as UK,
Iran, UAE and other countries category for maintaining
present status of export and the government has to
give more importance to the countries such as USA
and Japan to maintain the market share in the future.
The countries such as Bangladesh and Srilanka are
gaining the status of major importers of Indian turmeric.

308 Agricultural Economics Research Review Vol. 24 July-December 2011

The policies have to be drawn based on the problems
faced by the importing countries, so that the export of
turmeric would increase in future and India may earn
more foreign exchange through turmeric export.

Conclusions and Policy Implications
The analysis of the rate of growth in export and

direction of trade in turmeric in India has revealed that
the growth of turmeric export is satisfactory but the
direction of trade gives a warning. The liberalization
and globalization had a well-defined impact on the
turmeric export and this gives a positive signal. The
study has suggested that more importance should be
given to the R&D on quality of turmeric. Looking into
the importance of international demand, export earnings
and domestic needs, government should increase and
stabilize its outlay of funds for research on turmeric
under the spice development programs. The
government should be more conscious regarding the
policies pertaining to the above aspects and also WTO
implications to protect our farmers and to maintain our
monopoly in international markets. Appropriate export
promotion strategies and policies have to be evolved to
maintain the market share of Indian turmeric.

The policy implications emerging out of the study
are outlined below:

• High priority has to be assigned to increase the
production and productivity of turmeric.

• To maintain quality of turmeric, trainings should
be organized for the farmers on the way to produce
good quality turmeric. The facilities such as stream
boilers and mechanical driers need to be provided
by the government and spice industries to marginal
and small farmers.

• The result of Markov chain analysis has indicated
that India is likely to loose its export markets in
some of the countries like USA and Japan. Our
exports are likely to be concentrated in minor
importing countries, Iran, UAE, and UK in the
future. A high dependence on one or two export
markets will increase the trade risk in the long-
run. Therefore, more importance has to be given
to the minor importing countries such Bangladesh,
Sri Lanka, etc. and appropriate export promotion
strategies have to be evolved to diversify the
geographical concentration. Appropriate steps and

policies have to be evolved to maintain the market
share of Indian turmeric.

• There is a need to disseminate information on
international markets, price behaviour and other
trade matters to Indian farmers and institutions to
reap the benefits. This calls for strengthening of
information technology and providing forecasts in
products and prices.

References
Buckewell, A.E., Shucksmith, D.M. and Young, D.A (1983)

Structural projections of the Scottish dairy industry
using micro and macro Markov transitional matrices.
Journal of Agricultural Economics, 34(1): 57-68.

Deepa. K.M. (2010) Turmeric: The golden spice, Facts for
You, Sept: 19-20

GoI (Government of India) (1998) Spices Statistics, Spice
Board, Ministry of Commerce, Cochin, Kerala.

GoI (Government of India) (1999) Cocoa, Arecanut and Spice
Statistics, Department of Agriculture and Co-operation,
Directorate of Cocoa, Arecanut and Spice Development,
Calicut, Kerala.

Karvy Comtrade Limited (2008) Turmeric Seasonal report,
Karvy Ltd, March.

Karpagam, C. (2000) Cost of cultivation: A study on the
knowledge and adoption behavior of turmeric growing
farmers, Master of Science Thesis, University of
Agricultural Sciences, Dharwad.

Kumar, N.A. and Sankaran, P.G. (1998) Instability of turmeric
production in India. Journal of Spices and Aromatic
Crops, 7(1): 19-22.

Mamatha, B.G. (1995) Export trade of selected spices in India
— An economic analysis, Master of Science Thesis,
University of Agricultural Sciences, Bangalore.

Nagarajan, S. S. (2000) Turmeric cultivation : A hurdles race
on the farm fields. Kisan World, pp.42-43.

Nair, G.K., (2002) Monsoon delay may impact oleoresin trade
Financial Daily from The Hindu group of publications
Wednesday, 7 August.

Raveendran, N. and Aiyaswamy, P.K. (1982) An analysis of
export growth and export prices of turmeric in India.
Indian Journal of Agricultural Economics, 37(3): 323-
325.

Spices Statistics (2004) Spice Board, Ministry of Commerce
and Industries, Govt of India, Cochin.

Received: March 2011; Accepted: June 2011

Review and its content may not be copied or emailed to multiple sites or posted to a listserv without the

individual use.

Globalization of Production and Innovation:
How Outsourcing is Reshaping an Advanced

Manufacturing Area

LUCIA CUSMANO�†, MARIA LUISA MANCUSI�‡ and ANDREA MORRISON‡§
�Centro di Ricerca sui Processi di Innovazione e Internationalizzazione (CESPRI) – Bocconi University, Via Sarfatti, 25,

I-20136 Milan, Italy
†Department of Economics, Insubria University, Via Monte Generoso, 71, I-21100 Varese, Italy.

Email: lcusmano@eco.uninsubria.it
‡Department of Economics – Bocconi University, Via Sarfatti, 25, I-20136 Milan, Italy.

Email: mariahuisa.mancusi@unibocconi.it
§Urban and Regional Research Centre Utrecht (URU), Faculty of Geosciences, Utrecht University, Heidelberglaan 2,

NL-3508 TC Utrecht, the Netherlands. Email: a.morrison@geo.uu.nl

(Received July 2007: in revised form April 2008)

CUSMANO L., MANCUSI M. L. and MORRISON A. Globalization of production and innovation: how outsourcing is reshaping an

advanced manufacturing area, Regional Studies. This paper investigates the determinants and the spatial and functional dimensions

of firms’ outsourcing. Based on a large survey of manufacturing firms in Lombardy, Italy, the analysis shows that outsourcing is

remarkably wide across sectors and has a clear regional dimension, concerning highly skilled firms at most. Offshoring is still a

minor fraction of the deverticalization process, largely related to wider strategies of internationalization by foreign group subsi-

diaries at intermediate stages of the value chain. The evidence suggests the regional system is inserting onto global knowledge

networks, but also points at the risk of ‘branch plant effects’ in high-technology segments.

Outsourcing Offshoring Regional production system Manufacturing industry Italy

CUSMANO L., MANCUSI M. L. et MORRISON A. La mondialisation de la production et de l’innovation: comment l’approvision-

nement à l’extérieur réorganise une zone industrielle avancée, Regional Studies. Cet article cherche à examiner les déterminants et

la portée géographique et fonctionnelle de l’approvisionnement à l’extérieur des entreprises. A partir d’une enquête détaillée des

entreprises industrielles situées en Lombardie en Italie, l’analyse laisse voir que l’approvisionnement à l’extérieur s’avère très gén-

éralisée à travers les secteurs et a une portée nettement régionale en ce qui concerne notamment les entreprises dont la main-

d’oeuvre est hautement qualifiée. Les activités offshore représentent toujours une proportion négligeable du processus de désinte-

gration verticale et se rapporte étroitement aux stratégies d’internationalisation des filiales des groupes étrangers aux étapes inter-

médiaires de la chaı̂ne des valeurs. Les preuves laissent supposer que le système regional s’insère dans des réseaux de connaissance

mondiaux, mais indique également la menace que pose des ‘effets établissement’ dans les secteurs à la pointe de la technologie.

Approvisionnement à l’extérieur Activités offshore Système de production régional Industrie Italie

CUSMANO L., MANCUSI M. L. und MORRISON A. Die Globalisierung von Produktion und Innovation: Wie sich eine fort ges-

chrittene Produktionsregion durch Outsourcing verändert, Regional Studies. In diesem Beitrag untersuchen wir die Determinan-

ten sowie die räumlichen und funktionellen Dimensionen des Outsourcing von Firmen. Ausgehend von einer umfangreichen

Erhebung unter produzierenden Firmen in der Lombardei, Italien, geht aus der Analyse hervor, dass das Outsourcing in den

verschiedenen Sektoren bemerkenswert weit verbreitet ist und eine eindeutig regionale Dimension aufweist, die vor allem

Firmen mit hohem Qualifikationsniveau betrifft. Die Verlagerung ins Ausland stellt weiterhin einen kleinen Bruchteil des Dever-

tikalisierungsprozesses dar und ist größtenteils mit den breiter angelegen Internationalisierungsstrategien von Filialen ausländischer

Regional Studies, Vol. 44.3, pp. 235–252, April 2010

0034-3404 print/1360-0591 online/10/030235-18 # 2010 Regional Studies Association DOI: 10.1080/00343400802360451
http://www.regional-studies-assoc.ac.uk

Konzerne auf den mittleren Stufen der Wertschöpfungskette verknüpft. Die Belege lassen darauf schließen, dass sich das regionale

System in die globalen Wissensnetzwerke einfügt, weisen aber auch auf das Risiko von ‘Zweigwerkseffekten’ in Hightech-

Segmenten hin.

Outsourcing Verlagerung ins Ausland Regionales Produktionssystem Produzierende Industrie Italien

CUSMANO L., MANCUSI M. L. y MORRISON A. Globalización de producción e innovación: cómo la contratación externa remo-

dela un área manufacturera avanzada, Regional Studies. En este artı́culo investigamos los determinantes y las dimensiones espacial y

funcional de la contratación externa de empresas. Basándonos en un importante estudio de empresas manufactureras de Lombar-

dı́a, Italia, en este análisis mostramos que la contratación externa está muy extendida en todos los sectores y tiene una clara dimen-

sión regional, sobre todo con respecto a las empresas altamente cualificadas. La externalización de servicios representa todavı́a una

fracción menor del proceso de desverticalización, y en gran medida relacionada con estrategias más extensas de la internacionaliza-

ción por parte de filiales de grupos extranjeros en fases intermedias de la cadena de valores. La evidencia indica que el sistema

regional se inserta en las redes de conocimiento globales pero también señala el riesgo de ‘efectos de las sucursales’ en segmentos

de alta tecnologı́a.

Contratación externa Externalización de servicios Sistema de producción regional Industria manufacturera Italia

JEL classifications: D21, F23, L23, O32

INTRODUCTION

Over the last few decades, industrial restructuring in the
form of outsourcing has been emerging as a defining
character of the capitalist dynamics, transforming
business models and affecting the spatial structure of
industrial systems. In particular, the international
dimension of outsourcing (offshoring) has been lately
drawing much attention at both the analytical and the
policy levels, as a key driver of changes in the competi-
tive position of advanced and emerging regions
(UNITED NATIONS CONFERENCE ON TRADE AND
DEVELOPMENT (UNCTAD), 2004; ORGANISATION
FOR ECONOMIC CO-OPERATION AND DEVELOP-
MENT (OECD), 1998; AMITI and WEI, 2005).

The outsourcing phenomenon in advanced regions
dates back to the mid-1970s and accelerated during
the 1990s (GEREFFI and STURGEON, 2004), signalling
the ‘deverticalization’ of the modern corporation
(CHANDLER, 1977). Moreover, the structure of out-
sourcing has been widening in functional terms, as out-
sourcing strategies no longer concern only, or mostly,
fairly specialized repetitive tasks in production and
assembly. Rather, outsourcing increasingly involves ser-
vices of various type and content, including sensitive
functions and knowledge-intensive tasks, such as
design and research and development (R&D)
(HOWELLS, 2000; LEIBLEIN et al., 2002). As a conse-
quence, the increasing ‘distributedness’ of production
processes is followed (and affected) by a growing
‘distributedness’ of knowledge-intensive functions and
innovation processes, so that value-creating resources
and capabilities ever more frequently reside across the
boundaries of the firm (COOMBS and METCALFE,
1998).

The functional breadth of the outsourcing phenom-
enon is but one dimension of the complex emerging

trend, to which the spatial dimension should be
added. On the one hand, the internationalization of
value chains, or global ‘fragmentation’, has been attract-
ing much media hype, but also increasing theoretical
interest, because of its consequences on the positioning
of countries in the international division of labour
(for example, FEENSTRA and HANSON, 1996;
ARNDT and KIERZKOWSKI, 2001; GROSSMAN and
HELPMAN, 2002). On the other hand, agglomeration
advantages and cluster-centred flexible specialization
in core-regions are being reinterpreted (for example,
SCOTT, 1988; GAROFOLI, 2002; BOSCHMA, 2004), as
their relevance and geographical scale are affected
themselves by post-Fordist dynamics (PHELPS, 2004;
TORRE and RALLET, 2005).

Although lively, the theoretical and policy debate has

found still limited empirical application for two main

reasons. First, empirical investigations have been

mainly directed at specific sectors or local production

systems (for example, CORÒ and GRANDINETTI,

1999; AMIGHINI and RABELLOTTI, 2006), specific

functions, as in the case of the growing literature on

business service externalization (for example,

O’FARRELL et al., 1993; BEYERS and LINDAHL, 1996;

COE, 2000) or specific actors, such as multinational

branch plants or service-related headquarters (for

example, PHELPS, 1993; PERKMANN, 2006). Second,

quantitative studies based on large panel data sets have

been mostly based on very broad definitions of outsour-

cing, rarely differentiating the externalization of activi-

ties from more general purchasing strategies in a ‘make

or buy’ framework, and have often employed data at a

high (mostly industry) level of aggregation.1

The present paper contributes to fill this gap by

investigating the diversified patterns of externalization

across manufacturing industries and business actors

236 Lucia Cusmano et al.

in an advanced area, Lombardy (the Italian leading

economic region), which represents a mature and

highly heterogeneous industrial system, where large cor-

porations specialized in high-technology sectors coexist

with traditional industrial districts populated by small

firms.
Drawing on original and representative firm-level

survey data, the paper explores the extent of the externa-

lization practices, detailing direction, breadth and depth

of outsourcing strategies, thus providing an original

empirical contribution to the outsourcing debate. In

particular, the international outsourcing of production,

services and R&D activities is confronted with region-

ally contained dynamics, and the characteristics of

business actors driving the process at different spatial

levels are explored. In doing so, the paper adds to the

sparse empirical literature on the determinants of out-

sourcing and offshoring at the firm level (GIRMA and

GÖRG, 2004; GÖRG et al., 2004; GROSSMAN and
HELPMAN, 2002; SWENSON, 2004; TOMIURA, 2005),
and provides original insights for discussing both the
implications at the system level and the related arguments
proposed by the relevant literature.

The paper is organized as follows. The second
section summarizes the main issues emerging from the
literature and policy debate about outsourcing and off-
shoring, focusing on the motives for

outsourcing

and their relationship with its direction (local versus
international outsourcing), depth (total versus partial
outsourcing), and breadth (scope of functional outsour-
cing). The third section presents the survey method-
ology and the data set. The fourth section provides an
extensive description of outsourcing patterns in Lom-
bardy across industries and activities. The fifth section
focuses on the characteristics of firms driving the
process of deverticalization, presenting an econometric
assessment which differentiates between regional and
international outsourcing. The sixth section concludes,
discussing implications of the observed trend for the
regional system evolution and competitiveness.

OUTSOURCING: ECONOMIC DRIVERS,

SPATIAL DIMENSION AND FIRM

CHARACTERISTICS

Different strands of the literature, ranging from manage-
ment approaches to transaction cost economics and
more regional oriented studies, have investigated the
factors underpinning firms’ decisions to outsource
their internal activities, the spatial dimension of the
externalization process, and the associated firms’ charac-
teristics. However, while the motives for outsourcing
and its geographical scope are often (and naturally)
studied together in the literature, the interpretation of
fragmentation trends and spatial restructuring in terms
of firm-level characteristics is more recent and mostly
discussed in empirical contributions.

Cost factors have featured prominently in the debate
about vertical disintegration and its spatial dimension.
The transaction cost analytical framework represents in
this sense the main theoretical reference, suggesting
that firms externalize activities when and where external
provision is less expensive than internal procurement
(WILLIAMSON, 1985). SCOTT (1988) argues that in capi-
talist societies the organization of production, including
its spatial distribution, is constantly scrutinized by firms
with the purpose of reducing costs. This often implies
seeking for factor price differences across locations,
countries or regions, particularly, though not exclusively,
when labour-intensive production and assembling are
concerned. Accordingly, the spatial distribution of
outsourcing reflects factor cost differentials, involving
peripheral areas of advanced countries or developing
regions, which attract routinized unskilled production,
while core-regions dominate in unstandardized skilled
labour or contact-intensive activities, characterized by
high unit linkage costs (LEUNG, 1993). The recent
integration of international markets and the increasing
competitive pressure they have brought about help to
explain the late upsurge in international subcontracting
towards low-cost areas (FEENSTRA, 1998). Evidence of
total outsourcing at the international level comes
especially from traditional manufacturing sectors
heavily hit by competition from emerging economies.
Cost-cutting strategies have been favouring the emer-
gence of ‘lean and mean’ global players, transforming
producers into international buyers which coordinate
global production networks of subcontractors in many
different countries (GEREFFI, 1999).

The transaction cost perspective also emphasizes the
additional cost burden associated with international
outsourcing, as spatial dispersion can result in longer
lead times, larger inventories, communication and
coordination problems, difficulties in contractual speci-
fication and monitoring, which tend to rule out distant
subcontracting of non-standardized functions (GILLEY
and RASHEED, 2000). The transaction cost approach
therefore suggests that outsourcing to local suppliers is
to be preferred when market relationships are less
expensive at closer distance. In the case of advanced
high-cost areas, proximity is an advantage in terms of
contractual specification and monitoring, which are
all the more relevant when non-standardized tasks or
specific assets are concerned. As a consequence,
advanced services tend to locate much closer to their
primary source of demand since they entail significant
customization, frequent contacts between users and
providers, or even simultaneous production and
consumption (HOWELLS, 2000).

However, as standardization and asset specificity
evolve, international outsourcing concerns a wider
range of functions and products, including apparently
strategic activities, such as design and R&D. In this
respect, FREEMAN and SOETE (1997) underline that
not all R&D has high degrees of uncertainty and

Globalization of Production and Innovation 237

complexity attached to it. Indeed, several knowledge-
intensive activities have been undergoing a process of
‘commoditization’, generally reflected in declining
terms of trade and harsher price competition, even in
segments of high-technology industries (MINIAN,
2006). As such, firms find it preferable to outsource
these activities to suppliers who can offer standardized
products or services at a lower cost. In addition,
improved communication technologies make codifica-
tion easier and increase the ability of firms to monitor
and compare the quality of external suppliers, thus
creating alternatives to direct or close control and mini-
mizing the need for close user–producer interactions
(TETHER et al., 2001; NARULA, 2001). ‘Organized
proximity’ (TORRE and RALLET, 2005), that is,
common behavioural rules and routines and the
means for sharing information and knowledge, offers
powerful mechanisms for long-distance coordination,
thus widening the scope for outsourcing relational
intensive activities at the international level.

Cost advantages related to standardized input pro-
vision can also be found in relatively high-cost areas if
providers serve a large market and enjoy economies of
scale and specialization, as in the case of territorial
agglomeration of clients. This might explain why off-
shoring is still, in absolute terms, a limited phenomenon
(AMITI and WEI, 2005) and also why, in core regions,
outsourcing has been contributing to the expansion of
service complexes or thickening of local business
service markets (WOOD et al., 1993; ONO, 2007).

SCOTT (1988) relates the cost advantages of
subcontracting at the local level to self-reinforcing
marshallian externalities, as those which characterize
urban agglomerations (ILLERIS, 2005) or manufacturing
clusters. Marshallian externalities are a multidimen-
sional concept, comprising both pecuniary externalities
and knowledge externalities. These are characterized by
different tendencies.

On the one hand, pecuniary externalities are con-
sidered to be increasingly less important in driving the
agglomeration of suppliers and, as a consequence, in
explaining localized vertical disintegration (PHELPS,
2004; PHELPS and OZAWA, 2003). In fact, improve-
ments in transport and communication technology
and infrastructure have reduced the need for geographi-
cal proximity. Location in one area does not preclude
access to externalities generated in another area if the
two are strongly connected. In this sense, pecuniary
externalities are increasingly related to ‘accessibility’
rather than simply to ‘proximity’. Indeed, the wider
availability of pecuniary externalities tends to act as a
centrifugal force, deconstructing traditional industrial
agglomerations and changing the scale at which
agglomeration advantages are perceived (MARTIN,
1999; PHELPS, 2004).

On the other hand, knowledge externalities and
benefits from labour market pooling continue to act as
a significant centripetal force, favouring agglomeration

of specialized suppliers and flexible specialization models
in core regions (GAROFOLI, 2002). In this context, the
externalization of production and service activities is
mainly driven by motivations other than costs, such as
production smoothing, core-competence focus, or
expertise- and knowledge-searching strategies.

Production smoothing and the search for flexibility
are, according to BEYERS and LINDAHL (1996), ‘quasi-
cost’ factors in the sense that they are indirectly related
to cost-reduction strategies. In environments character-
ized by unstable market conditions, subcontracting
emerges as a mechanism for rapidly adjusting to
changes in the market, without harmful effects on the
level of efficiency (AJAYI, 2005). It stands as a defining
character of flexible regimes of capital accumulation,
in which internal economies of scale are largely replaced
by external economies (SCOTT, 1988; STORPER and
SCOTT, 1989). Production smoothing often takes place
at the local level, as rapidity and monitoring of quality
control are greatly important, unless bulky and highly
standardized activities are involved. However, it is a
strategy which typically involves ‘capacity’ or concur-
rent subcontracting (IMRIE, 1986, p. 956), rather than
downsizing through externalization.

Externalization is more likely to occur in rapidly
evolving markets, which require innovative responsive-
ness, fed by specialized providers and the integration of
different mixes of information and expertise (COFFEY
and BAILLY, 1992). As products become more sophisti-
cated and production relies on an increasing range of
specialized technological understanding, firms can
hardly develop internally all the capabilities and compe-
tences required to bring a product to the market.
Especially in environments characterized by strong
competition and short product life cycle, firms devote
internal resources to strengthen their core business,
while outsourcing non-core activities. This occurs, for
instance, in the case of ancillary services, which are
usually labour intensive (ABRAHAM and TAYLOR,
1996), but also for those complex activities in which
firms would be unable to keep the pace with changes
and challenges posed by specialized suppliers. In this
case, subcontracting to external specialized providers
responds to the related needs of strengthening core
competences, diverting resources and attention from
non-core activities, and accessing highly specialized
expertise, which complement in-house capabilities.

The expanding need for specialized knowledge also
explains the widening functional scope of outsourcing
decisions, which increasingly involve non-manufactur-
ing functions (WOOD, 1991). Indeed, outsourcing of
service activities to specialized suppliers has been a hall-
mark of recent industrial restructuring in advanced
regions, concerning an ever-larger range of service func-
tions.2 Business service functions are becoming increas-
ingly sophisticated and manufacturing firms generally
lack resources and strategic incentives to invest in their
development (COE, 2000). Total outsourcing of services

238 Lucia Cusmano et al.

is commonplace for small and medium-sized enter-
prises, which, by definition, have a limited amount of
resources to invest and little scope for economies of
scale in the intra-organizational provision. However, in
advanced areas, where manufacturing competitiveness
increasingly depends on knowledge contents, even
large corporations might be unable to produce innova-
tive services and normally refer to external knowl-
edge-intensive providers for expertise and consultancy
(WOOD et al., 1993), although the resulting relationship
rarely takes the form of ‘pure’ service externalization
(BEYERS and LINDAHL, 1996). More often, and
especially when knowledge-intensive or strategic activi-
ties are involved, complementary relationships between
in-house departments and specialized suppliers are
observed (MAHNKE, 2001). In such cases, outsourcing
responds to the need for reaping specialization gains
while exposing to a variety of learning experiences.
The risk associated with this strategy is that, if it
implies dismissal of strategic capabilities, it might also
undermine firms’ absorptive capacity (MAHNKE,
2001). This is one of the firm-level characteristics that
have attracted the attention of recent contributions
interpreting the trends of fragmentation and spatial
restructuring in terms of features of business players
which are driving the outsourcing dynamics.

Firm-specific factors such as size (ABRAHAM and
TAYLOR, 1996; MARTINEZ ARGÜELLES and
RUBIERA MOROLLÓN, 2004; KIMURA, 2002; GIRMA
and GÖRG, 2004; TAYMAZ and KILIÇASLAN, 2005;
MAZZANTI et al., 2006), productivity (KIMURA,
2002; TOMIURA, 2005; OLSEN, 2006), R&D intensity
(BARNEY, 1999; MAHNKE, 2001; MOL, 2005), human
capital (TOMIURA, 2005; MAZZANTI et al., 2006),
export or foreign direct investment (FDI) strategies
(GEREFFI, 1999; GROSSMAN and HELPMAN, 2002;
TOMIURA, 2005) are discussed and related to the cost
arguments, specialization or knowledge-searching
strategies commented above.3 Evidence on the matter
is, however, mostly anecdotal or based on case studies.
Investigation based on large firm-level data sets is in
its early stages, often referring to specific industries or
local production systems.

SAMPLE IDENTIFICATION AND SURVEY

METHOD

The empirical analysis draws on a representative and
large data set concerning the main manufacturing
sectors of Lombardy. The region represents a fully
fledged and mature industrial system, recently affected
by substantial tertiarization, although still exhibiting
important remnants of a manufacturing core. The
region accounts for about one-fifth of the Italian gross
domestic product and is leading the country in most
of the rankings related to innovation and internationa-
lization, although such leadership has been gradually

eroding at the national level, and the region has been
lately losing ground with respect to other advanced
European areas (CUSMANO and MALERBA, 2005). Its
openness makes it particularly exposed to international
changes and pressures, which affect in different
manners its highly heterogeneous sectors of specializ-
ation and productive milieux, characterized by a signifi-
cant presence of both high-technology multinationals
and small firm-based traditional industrial districts.4

The target sample of 1200 firms is drawn from the
national firm Census (ISTAT, 2001) and is stratified
according to geographical location, manufacturing
activity, and firm size.

Geographical stratification groups into four macro-
areas neighbouring provinces that exhibit significant
within-group similarities in terms of productive
specialization:

Milan.

. North-East: Varese, Como, Lecco, and Sondrio.

. North-West: Brescia and Bergamo.

. South: Pavia, Lodi, Cremona, and Mantua.

Stratification based on manufacturing activity is
obtained with reference to eight macro-sectors:

. Energy and Chemistry: mining, extraction of crude
petroleum and gas, coal and lignite, chemistry,
rubber and plastic, electricity, gas and water supply.

. Food and Tobacco: food products, beverages and
tobacco.

. Textile and Clothing: textile, wearing apparel,
tanning and leather, footwear.

. Wood and Furniture: wood and product of wood,
furniture.

. Paper and Publishing: publishing, printing and repro-
duction of recorded media.

. Mechanics and Transport: basic metals, other non-
metallic mineral products fabricated metal products,
machinery and equipment, motor vehicles, jewellery.

. Electronics and Optics: electrical machinery, radio
communication equipment and apparatus, precision
and optical instruments, watches and clocks, account-
ing and computing machinery.

. Construction: Construction and housing.

Size dimension stratification is based on the number of
employees and is built around five cells:

. 6–9.

. 10–49.

. 50–249.

. 250–499.

. Equal to or more than 500.

These size classes are based on the European Union
classification, but explicitly exclude micro-firms (i.e.
firms with fewer than six employees).

The number of firms in each stratum of the target
sample was obtained assuring proportionality to the

Globalization of Production and Innovation 239

total number of employees in the same stratum of the
population. However, appropriate balancing criteria
have been adopted in order to avoid strata with
small or medium-sized firms to have an insufficient
number of firms and ensure a satisfactory estimates’
precision.

Data were collected through an original firm-level
survey conducted in 2005. Each firm in the target
sample was contacted by a survey agency, which inter-
viewed via telephone either the chief executive
officer, or the managing director, or the chief adminis-
trative officer. A second target sample was available to
the survey agency to replace non-respondents. This
allowed one to obtain a final sample of 1148 regionally
based firms, which corresponds to a response rate equal
to 96%. The sample industry and size composition is
reported in Table 1, which shows that the Mechanics
and Transport macro-sector accounts for the relative
highest share of firms in the sample (34.8%), followed
by Textile and Clothing (14.5%), Energy and Chemistry
(14.5%), and Construction (12.5%). Table 1 also reports
the response rate by sector, which shows that firms from
Wood and Furniture and Construction were used by the
survey agency to replace non-respondents in other
sectors. As a consequence, appropriate survey esti-
mation methods are employed in the empirical analysis
to control for the potential bias originating from this
non-response/over-response bias.

The sample is mostly composed of small and
medium-sized firms (about 50% of the firms belong
to the 10–49 employees class). The share of small and
medium-sized enterprises is particularly dominant in
the Wood and Furniture industry and in Construction,
where about two-thirds of the firms have fewer than 50
employees. On the other hand, a non-negligible share
of large firms characterizes a few sectors, such as
Energy and Chemistry, Paper and Publishing, and
Mechanics and Transport (Fig. 1).

OUTSOURCING PATTERNS IN LOMBARDY:

BREADTH, DEPTH, AND

INTERNATIONALIZATION

The survey conveys information on firms’ outsourcing
decisions, where outsourcing is intended here as the

procuring of activities originally performed internally.
More specifically, the respondent was first asked to indi-
cate which functions the firm performs in-house, differ-
entiating among the following functional categories:

. Production and assembling.

. R&D and design.

. Services (information technology, personnel adminis-
tration, logistics and distribution, packaging,
maintenance).

For each function, the respondent was then asked to
specify if activities originally performed within firm
boundaries had been contracted out. If so, the respon-
dent was also asked to indicate whether the contractor
is located in Lombardy, in another Italian region, or
abroad. These sets of questions allow one to draw a
picture of both the geographical dimension of outsour-
cing and the depth of the phenomenon across functions.

The outsourcing phenomenon appears to have
pervasively affected the manufacturing system in
Lombardy. In fact, outsourcing involves nearly half of
the firms in the sample, and is uniformly distributed
across industries. The two significant exceptions are
Paper and Publishing and Electronics and Optics,
which represent the upper (60.7% of firms outsourcing)
and the lower (42.5%) tails of the distribution, respect-
ively (Table 2).

Direction

Outsourcing has a clear regional dimension: on average
more than 40% of firms (83% of outsourcers, that is,
firms outsourcing at least one function) refer to a
regional supplier for some of the functions they have
decided to contract out. This pattern prevails in
sectors which are at the heart of regional industrial
districts, such as Wood and Furniture, Textile and
Clothing, and Mechanics and Transport, or which are
mostly a locally based business, such as Construction.
This evidence is consistent with contributions stating
that local knowledge and supply chains, inter-firm
and interpersonal networks substantially increase the
scope of outsourcing (for example, MORGAN, 1997).

Furthermore, relying exclusively on regional con-
tractors is effectively the most common option
(Fig. 2). This strategy is indeed followed by a 30.8%

Table 1. Sample composition

Industry 6–9 10–49 50–249 250–499 �500 Total Response rate

Energy and Chemistry 19 63 62 9 10 163 0.91

Food and Tobacco 10 19 19 2 0 50 0.82

Textile and Clothing 26 78 54 4 5 167 0.92

Wood and Furniture 21 40 12 1 0 74 1.35

Paper and Publishing 6 30 15 4 1 56 0.79

Mechanics and Transport 59 193 113 21 14 400 0.89

Electronics and Optics 18 45 27 3 1 94 0.79

Construction 43 79 19 3 0 144 1.76

Total 202 547 321 47 31 1148 0.96

240 Lucia Cusmano et al.

share of firms in the sample (61.8% of all outsourcers).
Again, it is the locally based Construction industry
which is mostly involved in self-contained local net-
works of contracting. Regional outsourcing is also an
exclusive strategy for a significant share of firms in
Mechanics and Transport. Thus, restructuring
through externalization mainly generates localized lin-
kages. As mentioned, this can be related to the

presence of large local clusters, which are a distinctive
feature of the competitive system of Lombardy in this
and other traditional sectors and which are likely to
capture the outsourced functions. In this sense, the
trend seems to be driven by highly localized advantages
from division of labour and complementary specializ-
ation, rather than by the search for cost differentials
across space.

Table 2. Outsourcing, by industry (percentage of firms)

Industry

Percentage

outsourcers

Of which (share)

Regional

outsourcersa
National

outsourcersb Off-shorersc

Energy and Chemistry 51.53 0.82 0.33 0.18

Food and Tobacco 50.00 0.76 0.52 0.16

Textile and Clothing 47.90 0.83 0.29 0.25

Wood and Furniture 50.00 0.84 0.41 0.03

Paper and Publishing 60.71 0.71 0.53 0.09

Mechanics and Transport 48.25 0.84 0.30 0.16

Electronics and Optics 42.55 0.75 0.30 0.33

Construction 54.17 0.95 0.15 0.04

Total 49.74 0.83 0.31 0.16

Notes: aFirms outsourcing at least one activity within the region.
bFirms outsourcing at least one activity in other Italian regions.
cFirms outsourcing at least one activity abroad.

Fig. 1. Sample composition, by size class (%) across industries

Globalization of Production and Innovation 241

The share of firms that rely on outsourcing to
other Italian regions stands at a significant distance,
both in traditional sectors and in high-technology ones
(Table 2). Indeed, in Electronics and Optics,
international value chains attract outsourced activities
to a larger degree than national producers. This
high-technology industrial area is, however, quite an
exception in the regional fragmentation dynamics.
Off-shoring is, in fact, still a limited phenomenon,
accounting for a minor part of the overall outsourcing
trend. The participation to international fragmentation
processes concerns fewer than 8% of the firms in the
sample (and 16% of the actual outsourcers) and is
unevenly distributed across industries. Two highly differ-
ent sectors stand well above the average: Electronics and
Optics, and Textile and Clothing. The share of outsour-
cers that have been contracting activities abroad is equal
to nearly one-third in the first and one-quarter in the
latter. It is, therefore, in these two sectors that the exter-
nalization process mostly reflects integration into the
increasingly international division of labour.

Breadth and depth

It is to be expected that the regional or international
dimension of the outsourcing process is related to the

content of the activities concerned, and that those activi-
ties are (performed and) outsourced differently across
industries, reflecting industry differences in terms of
competitive factors, competitive strategies of the firms,
and comparative advantages of the territories. Accord-
ingly, the functions being outsourced are differentiated,
and the relationship between outward orientation (or
regional embeddedness) and type of activity is explored,
always taking into account industry differences.

The breadth and depth of outsourcing is analysed in
relation to the three functional categories of production
and assembling, R&D and design, and services. First of
all, it is important to underline that firms have been
performing these functions to a different degree.
Production/assembling activities are, as expected, the
defining character of the manufacturing system in
Lombardy, and even when firms decide to turn to
external suppliers for activities they used to perform
in-house, outsourcing rarely implies complete disinvest-
ment. As shown in Table 3, only a very minor share of
firms (3%) never carried out any production activities,
whereas 94% exhibit dedicated functions. For nearly
74% of firms, production or assembling have not been
affected by any type of outsourcing, while the remain-
ing 20% of firms with dedicated functions have partially
outsourced them. Only 3% of firms in the sample

Fig. 2. Regional outsourcing, by industry (percentage of firms)

242 Lucia Cusmano et al.

have been going through complete disinvestment
in manufacturing activities. ‘Hollow companies’
(FREEDMAN, 2004), that is, firms turned into pure
coordination structures, are therefore still rather
infrequent.

Indeed, the strategy of full disinvestment (total out-
sourcing) concerns a small share of firms across all the
functions examined. The area in which turning to
external suppliers most frequently implies that firms
dismiss the related function is R&D and design.
Among firms that have been performing some type of
research activity (75% of the sample),5 very few out-
sourced any of them, preferring to keep this strategic
(and sensitive) function close to the in-house core.
However, when outsourcing took place, it was more
likely, compared with other functions, that it turned
into total out-contracting. In other words, partial
outsourcing is relatively less common for R&D than
for other functions.

This evidence is at odds with the prediction of the
resource-based view of the firm, which would suggest
that complementary relationships between in-house
departments and specialized suppliers are more likely
to be observed in the case of knowledge-intensive
activities (MAHNKE, 2001), as dismissing these functions
undermines firms’ absorptive capacity. This is less the
case if R&D-related activities present low degrees of
uncertainty and complexity (FREEMAN and SOETE,
1997) or in environments characterized by ‘open inno-
vation models’ (LAURSEN and SALTER, 2004), which
would, however, require close ties and proximity
(SORENSON et al., 2006).

This evidence can be further qualified by considering
the direction of outsourcing, in both spatial and organ-
izational terms. Table 4 presents the relevance of
regional versus international suppliers, considering

‘potential outsourcers’ only (that is, firms that have/
had functions related to the area under investigation).
For instance, nearly 40% of the firms with in-house
services experienced deverticalization. Among them,
85% referred, at least for one of the externalized activi-
ties, to regional suppliers. This suggests that the local
markets for services are thick, although exclusively
regional outsourcing is less common, that is, most
firms resort to service providers located at various
sites, across the regional and, to a lower degree, the
national border.

This evidence is consistent with the observation in
the literature that localized externalization is one of
the driving forces of service markets growth in core
regions (O’FARRELL et al.,1993, COE, 2000).

As far as R&D is concerned, over 82% of the firms
performing some type of R&D activity have not
experienced any (even partial) outsourcing. Regional
markets are nevertheless relevant as not only the
location of suppliers, but also the share of firms refer-
ring to foreign contractors, is similar to that observed
in production and assembling. This result, however,
hides significant inter-industry differences (Fig. 3).
Industries exhibiting a relatively higher percentage
of firms outsourcing abroad include such different
areas as Textile and Clothing, and Electronics and
Optics. When uncovering the details about the
nature of the outsourced activities, it is, however,
evident that the similar trend is related to different
strategies (and, supposedly, determinants). For
Textile and Clothing, offshoring consists mainly in
international outsourcing of production and assem-
bling activities and, to a lesser extent, services. The
Electronics and Optics industry is significantly more
oriented towards international outsourcing of R&D
and design, although the offshoring of production

Table 4. Direction of outsourcing, by localization (percentage of potential outsourcersa)

Function

Percentage of

outsourcers
Of which (share)

Percentage of

regional

outsourcersb
Percentage of

off-shorersc

Production/assembling 24.06 0.73 0.17
Research and development/design 17.71 0.72 0.18
Services 39.51 0.85 0.11

Notes: aFirms performing or having performed the function.
bFirms outsourcing at least one activity within the region.
cFirms outsourcing at least one activity abroad.

Table 3. Depth of outsourcing, by function (percentage of firms)

Function

Never

performed

Totally

outsourced

In-house and

outsourced In-house only

Production/assembling 2.96 3.05 20.30 73.69
Research and development/design 25.70 6.36 7.40 60.54
Services 15.33 4.36 29.09 51.22

Globalization of Production and Innovation 243

activities is also non-negligible. If nearly half of
the R&D outsourcers in the industry outsourced
abroad, about one-third of those outsourcing
production turn to international suppliers. Energy
and Chemistry is the other industry whose inter-
national outcontracting is above the average in all
functions and, particularly, in R&D. At the other
extreme, the Construction and Wood and Furniture
industries refer almost entirely to the domestic
market.

The high share of international outsourcing in
knowledge-intensive activities is partially to be related
to strategies of multinational groups. In fact, the
present data also show that, when offshored, the R&D
function is indeed transferred to another group affiliate
abroad (or the foreign headquarters) more frequently
than in the case of production and services.6 This
seems to suggest that regional high-technology indus-
tries are affected by a sort of ‘branch plant effect’,
which is generally associated with peripheral areas
(PHELPS, 1993). Multinational branch plants are ‘out-
sourced’ of their R&D functions by headquarters,
which implies a lower degree of regional embeddedness
in the forms of localized knowledge-intensive linkages.

FIRM-LEVEL CHARACTERISTICS AND THE

GEOGRAPHICAL DIMENSION OF

OUTSOURCING

Econometric model and description of the variables

Following recent empirical contributions to the deter-
minants of outsourcing (GIRMA and GÖRG, 2004;
MAZZANTI et al., 2006; MOL, 2005; TOMIURA,

2005), outsourcing decisions by firms are modelled as
a function of a number of variables reflecting firm-
specific characteristics, while accounting for sectoral
specificities. Different probit models are estimated in
order to assess the possible distinct relevance of these
characteristics for the inward (regional) and the
outward (foreign) orientation of the outsourcing strat-
egy. In the first model, regional outsourcing is con-
sidered exclusively: the dependent variable is a
dummy equal to one when the firm has been undertak-
ing regional outsourcing only. The second model
describes the probability of a firm performing inter-
national outsourcing: here the dependent variable is a
dummy equal to one when the firm has outsourced
some activity to another country. Finally, for the sake
of comparison, a third model is also estimated where
the dependent variable is a dummy equal to one when
the firm performs any kind of outsourcing, that is, inde-
pendently of the localization of the contractor.

All

models are estimated while accounting for the effects
of sampling design and response on population estimates
by using pseudo-maximum likelihood methods and
allowing for probability sampling weights and
stratification.

The explanatory power of a basic set of quantitative
variables, also obtained from the survey, is first tested
(Table 5). The first explanatory variable, PRO-
DUCTIVITY (sales over employees), is intended to
test the hypothesis that firms engaged in outsourcing
have higher productivity than vertically integrated
firms (OLSEN, 2006). The rationale is that firms out-
source activities in which they are less efficient or for
which they do not enjoy much competitive advantage,
based, for instance, on unique knowledge or skills,

Fig. 3. Breadth of domestic outsourcing and offshoring, by industry (percentage of firms)

244 Lucia Cusmano et al.

while focusing on their core competencies or reallocat-
ing resources towards activities with greater value added,
and thus gaining in productivity. Only very few studies
have analysed the reverse direction of causality, provid-
ing, however, no clear-cut evidence (KIMURA, 2002;
TOMIURA, 2005). For example, TOMIURA (2005)
points to a greater marginal relevance of productivity
for international outsourcing than for generic outsour-
cing that would be explained by the high fixed costs
for foreign contracting, which makes this alternative
viable for rather productive firms. Notice that the
present measure of labour productivity could be
positively related to outsourcing also because firms
contracting out their activities usually reduce the
number of employees, while sales remain constant.

The second explanatory variable, RDI (R&D over
sales), measures R&D intensity, whose impact on
outsourcing cannot be straightforwardly signed
(MAHNKE, 2001; MOL, 2005). The conventional view
would argue that R&D-intensive industries tend to be
vertically integrated in order to recover the high sunk
costs generated by R&D investment. A further and
complementary argument, based on the transaction
cost approach, maintains that industries dealing with
complex products face severe incentive and appropria-
bility problems, which they tend to solve through
vertical integration (MOL, 2005; TEECE, 1986). On
the other hand, R&D intensity would be associated to
extensive outcontracting in the literature conceiving
the firm as an open platform, developing external
networks, in particular international ones, to access
relevant capabilities, rather than building them internally
(BARNEY, 1999). The growing complexity of technol-
ogies is one of the key reasons for firms to search for
external sources of knowledge (BRUSONI et al., 2001)
and relationships with suppliers represent important
channels for accessing capabilities. Local outsourcing
provides advantages of generally lower transaction costs
and continuous interaction, favouring interactive learn-
ing and incremental change. On the other hand, inter-
national outsourcing can be aimed at entering global
knowledge networks, and recent advances in communi-
cation technologies have made this easier (MOL, 2005).

A further explanatory variable is a measure of firm
size (SIZE ), here evaluated in terms of number of
employees. There seems to be disagreement on the
direction of the impact of size. The core competence
literature, for instance, would indeed suggest that
small firms are more likely to outsource, since they

have a strong incentive to devote their limited internal
(physical, financial, and intangible) resources to core
activities and bring out non-core ones (ABRAHAM
and TAYLOR, 1996; CORÒ and GRANDINETTI,
1999). In the case of local systems, regionally confined
outsourcing by small and medium-sized enterprises is
to be interpreted in the framework of a strong division
of labour, which allows local producers to enjoy increas-
ing returns from specialization, and the local system to
achieve a high degree of ‘flexible specialization’
(GAROFOLI, 2002). On the other hand, small firms
are expected to outsource fewer activities, as they
have a smaller scope to start with, and, especially as far
as service activities are concerned, fewer and simpler
needs than large firms (MARTINEZ ARGÜELLES and
RUBIERA MOROLLÓN, 2004). The positive relation
between firm size and outsourcing is supported by the
idea that subcontracting is a strategy of ‘production
smoothing’, which allows large firms to reduce costs
and enhance flexibility (IMRIE, 1986; KIMURA, 2002;
TAYMAZ and KILIÇASLAN, 2005). The effect of size
has been tested by some recent empirical works
(GIRMA and GÖRG, 2004; MAZZANTI et al., 2006),
which, however, do not provide clear-cut evidence.

The explanatory variable HK (share of employees
having secondary education) is meant to characterize
the firm human capital endowment. Cost-saving strat-
egies would suggest a positive relationship between
human capital and outsourcing in general. The rationale
is that firms employing highly skilled workers pay effi-
ciency wages and are not able to pursue different wage
strategies. However, in order to save costs, they would
be keen to outsource peripheral activities for which
they employ workers that are paid above the market
rate efficiency wage. On the other hand, the competence
perspective points to differentiated effects and to ambig-
uous empirical outcome. MAZZANTI et al. (2006) under-
line that a high level of skills can represent an important
incentive to specialize in knowledge-intensive activities,
while outsourcing more standard production. However,
firms with high skills are less willing to outsource if this
creates the risk of losing some distinctive capabilities,
thus impoverishing the organizational competences
which are built upon them. Finally, other authors have
emphasized that human capital is particularly relevant
for contracting internationally. Qualified human skills
are in fact deemed essential for engaging in contracting
abroad since this requires human capital intensive activi-
ties such as negotiating with partners in foreign languages

Table 5. Explanatory variables (logs): descriptive statistics

Variable Measure Mean Standard deviation Minimum Maximum

SIZE Employees 3.42 1.25 1.79 8.52

PRODUCTIVITY Sales/employees 11.88 1.31 4.29 17.73
RDI Research and development/sales 0.69 1.02 0 4.62
HK Percentage of secondary-educated employees 3.11 1.21 0 4.62

EXPI Export/sales 2.015 1.74 0 4.62

Globalization of Production and Innovation 245

and concluding contracts under different legal systems
(TOMIURA, 2005).

In all specifications industry specific effects are con-
trolled and a dummy variable FINAL PRODUCT,
which identifies firms engaged in the production of
final goods, is introduced.7 This variable is included
to account of the differentiated behaviour of firms oper-
ating at different stages of the value chain and a positive
relationship is expected between outsourcing and
downstream production activities (for example, final
transformation or assembling). This is because down-
stream producers tend to exhibit a greater scope of
activities or functions for which outsourcing can take
place. In addition, it is especially in downstream pro-
duction that outsourcing represents an effective strategy
for smoothing production over different subcontractors,
thus coping with seasonal or demand peaks (IMRIE,
1986). More broadly, this finding would be consistent
with the evidence on cost-saving strategies, as pointed
out in recent works on subcontracting relationships
(TAYMAZ and KILIÇASLAN, 2005).

The present authors further control for the outward
orientation of the firm and for organizational specific
effects. The outward orientation of the firm seems to
be an important control for offshoring behaviour and
is captured by two indicators that should convey infor-
mation on the firm business experience in foreign
countries. EXPI (export/sales) represents export inten-
sity, while FDI is an indicator variable, which takes the
value of one if the firm has undertaken foreign direct
investment.8 The paper first controls for the relevance
of FDI per se, and then it further distinguishes between
production and/or R&D FDI (FDI_PLANT), on the
one hand, and the mere opening of a sales office
abroad (FDI_SALES), on the other hand. The empiri-
cal literature suggests a positive relationship between
outward orientation and international outsourcing
(TOMIURA, 2005), which is consistent with the likely
reduction of fixed costs of foreign contracting when
firms already have business experience in foreign
countries. In addition, GÖRG et al. (2004) underline
the potential advantage for exporters in accessing
extensive knowledge on where to procure competi-
tively priced inputs. This is in line with the model of
GROSSMAN and HELPMAN (2002), which stresses the
relevance of search costs for international outsourcing.
In accordance, both kinds of internationalization
strategies are expected to affect positively the prob-
ability of foreign outsourcing, and the productive
type of FDI to have a larger effect than the investment
in sales units or offices.

As far as organizational-specific effects are con-
cerned, a GROUP dummy is introduced that takes a
value equal to one if the firm is part of an economic
group, and zero otherwise. This variable is further
split into the variable SUBSIDIARY, which identifies
firms that are subsidiaries in a group, and GROUP_-
HEAD, which identifies the headquarters.9 It is

expected that being part of a group positively affects
outsourcing, as firms are, in principle, embedded in a
larger network of providers and potential clients. In
addition, subsidiary firms are expected to be more
likely than headquarters to perform outsourcing, as
the first are generally more involved in production
activities, whereas the latter are likely to host adminis-
trative and often strategic (for example, R&D) func-
tions, enjoying economies of scale and scope in the
provision of group-wide services. At the same time,
because Lombardy is a core region characterized by a
dynamic business environment and a major tertiary
area (Milan), branch plants are expected to rely signifi-
cantly on local service markets.

Results

Table 6 reports correlations among the main explana-
tory variables; and Table 7 shows the results from differ-
ent specifications of probit estimation for survey data for
each of the three dependent variables specified in the
previous section (any outsourcing, exclusively regional
outsourcing, and international outsourcing).10

Interestingly, in all regressions human capital appears
to be a good and highly significant predictor of outsour-
cing behaviour. The result is consistent with both the
cost-saving explanation (GIRMA and GÖRG, 2004)
and the idea of specialization driven by skills. Human
capital significance is observed when focusing both on
regionally oriented outsourcers and on offshorers,
even when outward orientation variables (EXPI, FDI)
are taken into account. One can infer that human
capital represents an important asset for operating in
foreign markets, but also a relevant driver in the region-
ally based flexible specialization.

By contrast, productivity does not appear to be sig-
nificant in any of the specifications tested, which is
not surprising since the reverse causality has been in
fact rarely observed in the empirical literature.11 Fur-
thermore, the result might be affected by the prevalence
of small and medium-sized enterprises in the sample
(and in the Lombardy, and more generally Italian, pro-
duction system), as the effect on productivity induced
by the possible employment reduction is likely to be
marginal for small and medium-sized enterprises.

The R&D intensity variable (RDI) exhibits a similar
pattern of significance. Differently from MOL (2005),
no robust evidence is found that R&D-intensive firms

Table 6. Correlations among explanatory variables

SIZE PRODUCTIVITY RDI EXPI HK

SIZE 1.000

PRODUCTIVITY 0.145 1.000

RDI 0.251 0.043 1.000

EXPI 0.415 0.066 0.296 1.000

HK 0.263 0.181 0.190 0.187 1.000

246 Lucia Cusmano et al.

Table 7. Estimation results

Variable

(1) (2) (3) (4) (5) (6) (7)

All
outsourcing

Exclusively

regional
outsourcing

International

outsourcing
All
outsourcing
Exclusively
regional
outsourcing
International
outsourcing
International
outsourcing

SIZE 20.00 20.01 0.19�� 20.01 20.00 0.10 0.07
(0.05) (0.05) (0.05) (0.05) (0.06) (0.07) (0.07)

PRODUCTIVITY 0.03 20.00 0.03 0.03 0.00 0.02 0.02
(0.04) (0.04) (0.06) (0.04) (0.04) (0.06) (0.06)

RDI 0.08� 20.01 0.05 0.08 20.02 0.04 0.04
(0.05) (0.05) (0.06) (0.05) (0.05) (0.06) (0.06)

HK 0.14�� 0.10�� 0.21�� 0.14�� 0.10�� 0.19�� 0.19��

(0.04) (0.04) (0.08) (0.04) (0.04) (0.08) (0.08)

FINAL PRODUCT 0.16 0.25� 20.49�� 0.16 0.25� 20.52�� 20.52��

(0.12) (0.13) (0.18) (0.12) (0.13) (0.19) (0.19)

EXPI 0.03 0.03 0.11�� 0.12��

(0.03) (0.03) (0.05) (0.05)

FDI 20.09 20.32�� 0.31��

(0.13) (0.14) (0.16)

FDI_PLANT 0.65��

(0.30)

FDI_SALES 0.41

(0.34)

CONSTANT 20.82� 20.53 22.95�� 20.81� 20.59 22.54�� 22.48��

(0.46) (0.47) (0.73) (0.46) (0.48) (0.79) (0.78)

Industry dummies Yes Yes Yes Yes Yes Yes Yes

F-statistic 2.33 2.14 6.27 2.05 2.18 7.31 6.88

Prob . F 0.01 0.01 0.00 0.01 0.01 0.00 0.00
Observations 1099 1099 1099 1099 1099 1099 1099

(8) (9) (10) (11) (12) (13) (14)

SIZE 20.01 0.01 0.04 20.01 0.01 0.06 0.05

(0.06) (0.06) (0.07) (0.06) (0.06) (0.07) (0.07)

PRODUCTIVITY 0.03 0.00 0.01 0.04 0.01 0.00 0.00

(0.03) (0.04) (0.06) (0.04) (0.04) (0.06) (0.06)

RDI 0.07 20.02 0.04 0.07 20.02 0.04 0.04
(0.05) (0.05) (0.06) (0.05) (0.05) (0.06) (0.06)

HK 0.14�� 0.10�� 0.17�� 0.14�� 0.10�� 0.17�� 0.17��

(0.04) (0.04) (0.07) (0.04) (0.04) (0.07) (0.07)

FINAL PRODUCT 0.16 0.25� 20.51�� 0.16 0.25� 20.51�� 20.53��

(0.12) (0.13) (0.19) (0.12) (0.13) (0.19) (0.19)

EXPI 0.03 0.03 0.12�� 0.03 0.03 0.12�� 0.12��

(0.03) (0.03) (0.05) (0.03) (0.03) (0.05) (0.05)

FDI_PLANT 20.19 20.75�� 0.55� 20.20 20.78�� 0.58� 0.62��

(0.22) (0.22) (0.30) (0.22) (0.22) (0.30) (0.30)

FDI_SALES 20.17 20.37 0.32 20.16 20.37 0.34 0.39
(0.32) (0.30) (0.36) (0.32) (0.30) (0.36) (0.35)

GROUP 0.10 20.03 0.32�

(0.15) (0.15) (0.17)

GROUP HEAD 0.25 0.39 20.30 20.31
(0.25) (0.29) (0.27) (0.27)

SUBSIDIARY 0.07 20.13 0.40�� 0.26
(0.16) (0.16) (0.18) (0.20)

SUBSIDIARY �

FOREIGN

0.59�

(0.34)

CONSTANT 20.79� 20.63 –2.34�� 20.80� 20.68 –2.27�� –2.21��

(0.46) (0.48) (0.76) (0.46) (0.48) (0.78) (0.78)

Industry dummies Yes Yes Yes Yes Yes Yes Yes

F-statistic 1.83 2.41 6.83 1.76 2.34 6.59 6.56

Prob . F 0.02 0.00 0.00 0.03 0.00 0.00 0.00
Observations 1099 1099 1099 1099 1099 1099 1099

Notes: Standard errors are given in parentheses; �significant at 10%; ��significant at 5%; and ��significant at 1%.

Globalization of Production and Innovation 247

have a higher overall probability to outsource, thus the
findings do not support the perspective of increasing
specialization and reliance on external sources by
R&D-oriented firms. If R&D-intensive firms outsource
in order to search for competent external suppliers that
provide complementary resources, as suggested by the
‘relational view’ (MAHNKE, 2001), the present findings
might suggest that the regional scale is too small for creat-
ing a market for competent suppliers (PHELPS and
OZAWA, 2003), while in international markets, where
competition is harsher, knowledge-intensive firms tend
to be more sensitive to the appropriability problems
entailed by outsourcing.

Size does not appear to be a good predictor for the
general strategy of outsourcing. This is in line with the
evidence provided by MOL (2005). When specifying
the direction of outsourcing, the variable appears to
exhibit a positive effect on international outsourcing,
that is large firms appear to be significantly more likely
to engage in offshoring. This result, which confirms
findings by TOMIURA (2005), seems to point to an
apparently trivial implication, that is, availability of a
large pool of resources is relevant for sourcing at the
international level. However, the effect of size disappears
once one properly accounts for international orientation
of firms through export intensity and FDI. It is found
that a strong relationship exists between international
outsourcing behaviour and foreign business experience,
as represented by both export activity (EXPI ) and FDI.
It is foreign investment in a production or R&D unit that
really matters, while the mere opening of a sales office
abroad does not show any effect on the probability to
do any kind of outsourcing.

A striking difference emerges between firms out-
sourcing at the regional level only and those going
abroad when focusing on the final product dummy.
This is significant in both cases (under all the specifica-
tions tested), but takes a positive value, as expected, only
when considering exclusively regional outsourcing. On
the contrary, it exhibits a significant negative relation-
ship with offshoring. It is maintained that this result is
only apparently counterintuitive. In fact, it is consistent
with findings in the trade literature showing that intra-
industry trade currently dominates international trade
flows (GROSSMAN and HELPMAN, 2002; MINIAN,
2006). In particular, multinationals develop large net-
works among their affiliates that source factories all
over the world (GEREFFI, 1999). In this respect, it is
worth noticing that the final product variable is still
significant and negative for the estimation on the off-
shoring dependent variable when controlling for
group dummies.

Related to the above, being part of a group matters
for international outsourcing: support is found for the
evidence that group subsidiaries rather than headquar-
ters are the drivers of international outsourcing activi-
ties. Furthermore, once one controls for foreign
ownership by interacting the dummy SUBSIDIARY

with a dummy FOREIGN taking a value of one if
the firm is controlled by foreign actors,12 it is found
that it is foreign ownership of subsidiaries that posi-
tively affects the probability of international outsour-
cing. This suggests that foreign-controlled units
located in Lombardy ‘have been outsourced’, that is,
headquarters of foreign multinational enterprises have
either appropriated the function or transferred it to
another subsidiary/external firm. This adds to the
evidence about R&D total offshoring by firms in
high-technology industries in pointing to a sort of
‘branch plant effect’ in knowledge-intensive segments
(PHELPS, 1993).

CONCLUSIONS

This paper has explored the extent of outsourcing from
the perspective of individual firms located in an
advanced manufacturing area. In doing so, it contrib-
utes to the vast literature on firms outsourcing strategies
and in particular to the understanding of the determi-
nants of firms’ outsourcing decisions. Results shed
light on a number of relevant dimensions of outsourcing
(i.e., direction, breadth, depth) and thereby comp-
lement and corroborate the prevalently qualitative and
anecdotal evidence on this topic.

The first outcome of this investigation clearly indi-
cates that, in the case of Lombardy, outsourcing is
remarkably widespread and concerns to a similar extent
all industrial sectors. The present findings support pre-
vious evidence showing that subcontracting in core
regions (for example, WOOD et al., 1991; COE, 2000;
ILLERIS, 2005) is mainly local. Indeed, overall outsour-
cing has a clear and predominant local dimension in
those sectors which are highly rooted in regional clusters.
In this sense, it appears that marshallian externalities are
still relevant in driving deverticalization and feeding
the local dense web of productive relationships, which
constitutes the defining character of the flexible special-
ization model (SCOTT, 1988; STORPER and SCOTT,
1989). The present evidence suggests that in this core
region externalization is not leading to a loss of density
of those traded relationships that represent the source
of local competitive advantage (GAROFOLI, 2002).

Subcontracting in Lombardy (and in core regions)
involves services to a large degree. The externalization
of service functions contributes to thickening of the
local service markets and feeds positive agglomeration
effects in advanced areas. In this sense, core-competence
focus, or ‘quasi-cost’ factors (BEYERS and LINDAHL,
1996), and the search for complementary expertise at
the territorial level seem to be more relevant than cost
rationales. The importance of human capital in explain-
ing regionally based specialization further suggests that
local fragmentation is driven by knowledge-oriented
players. These actors, mostly downstream producers,
maintain some related in-house capabilities which

248 Lucia Cusmano et al.

support the flexibility and governance of close ties with
selected suppliers (O’FARRELL et al.,1993).

Interestingly, outsourcing in value-added services
(i.e. R&D and design) is less common – suggesting
that firms are still very much concerned with appropria-
tion problems – but is relatively more likely to span
across regional or even national boundaries.

The implications of the results for regional develop-
ment – and the broad debate about the impact of glo-
balization on regional clusters – are of particular
interest. The risk of local networks impoverishment
has been the focus of much recent debate about
distant outsourcing. Regions are depositories of tangi-
ble and intangible resources, which mostly reside in
local firms and on which localized capabilities are
built (BOSCHMA, 2004; MASKELL and MALMBERG,
1999). Distant outsourcing can seriously undermine
these capabilities, reduce the local relational density,
and the internal cohesion of the regional system.

In the case of Lombardy, it seems that externalization
has been adding new ties or reinforcing existing ones in
the local production system. Especially in the area of
business services, opportunities have been created for
focused niche players to enter regional markets. There-
fore, it appears the adjustment to global trends has been
taking place mostly within the model of flexible special-
ization, driven by highly localized advantages from div-
ision of labour and complementary specialization.
Insertion into global production networks seems not
to have been disruptive of industrial clusters, deepening,
rather than weakening, local linkages.

However, the strong inward orientation of the
process raises other questions, in relation to the long-
term development trajectory. In fact, dismantling and
relocating activities outside a region does not necessarily
imply negative effects for the local economy, in terms of
efficiency or development dynamics. Outward externa-
lization can lead to a better allocation of resources,
which shift from low-yielding activities, or declining
sectors, to more productive ones. Furthermore, regions
benefit from international outsourcing when firms sub-
contract to high competent producers and/or markets,
which become channels for accessing new knowledge
and preventing lock-in effects (CAMAGNI, 1991).

Overall, the present analysis indicates that inter-
national outsourcing is still a minor part of a wider frag-
mentation trend, which concerns mainly traditional
sectors, such as Textile and Clothing. In this case, inter-
national outsourcing consists mainly of the externaliza-
tion of production and assembling activities and seems
to respond to efficiency-seeking strategies, which may
positively affect regional dynamics.

However, international outsourcing also touches upon
knowledge-intensive and large-scale sectors. On the one
hand, this might favour or strengthen the insertion of
the regional system into global knowledge networks.
On the other hand, if it implies dismissal of high value-
added functions or loss of strategic assets at the local

level, it might seriously hinder the growth potential of
the regional system. The evidence suggests that the
region might indeed suffer from a sort of ‘branch plant
effect’ in knowledge-intensive segments, as R&D func-
tions, when outsourced, are more likely to be entirely
appropriated by foreign headquarters or research units.

In this late phase of world capitalism, which SCOTT
(2005) describes as marked by intensified regionalization
of production overlaid by a global division of labour, tra-
ditional industrial core regions, such as Lombardy, face
the challenge of preserving their internal cohesion,
while governing their insertion into evolving global net-
works, counteracting the centripetal forces which are
redistributing knowledge-intensive functions and rede-
fining the global geography of core and periphery.

Acknowledgements – The authors would like to thank
the participants at the following: ProAct Conference

(Tampere, Finland), International Schumpeter Society

Conference (Nice, France), EARIE Conference (Amsterdam,

the Netherlands), EMAEE Conference (Manchester, UK),

and in seminars and workshops at the Università di

Roma Tre (Rome, Italy), CESPRI-Università Commerciale

‘L. Bocconi’ (Milan, Italy), Università del Salento (Lecce,

Italy), Università del Piemonte Orientale (Novara, Italy), and

Beta-Université Louis Pasteur (Strasbourg, France). Financial

support from IReR-Lombardia is gratefully acknowledged.

Andrea Morrison also acknowledges funding from the Italian

Ministry of Education, University and Research (PRIN 2005

on ‘Fragmentation and Local Development’).

NOTES

1. TOMIURA (2005) provides a concise review.

2. O’FARRELL et al. (1993) define this process of vertical

disintegration as ‘service unbundling’.

3. For the sake of clarity and to ease interpretation, these

relationships will be presented in detail in the fifth

section when discussing the structure of the empirical

analysis on firm-level characteristics and the geographical

scope of outsourcing.

4. Lombardy has been the focus of early investigations about

flexible production systems and industrial district models,

emerging in the late 1970s as a peculiar case of diffused

industrial development based on small and medium-

sized enterprises (GAROFOLI, 1983).

5. The survey adopted a broad definition of R&D, which is

intended as any activity oriented towards research and

experimentation.

6. A total of 69% of firms offshoring R&D activities do so,

at least partially, to a group affiliate. The corresponding

figures for production and services are 64% and 52%,

respectively. The difference is, however, more striking

when considering firms offshoring exclusively to group

affiliates: these are 35% of firms offshoring R&D, but

only 2% of firms offshoring production and 7% of

firms offshoring services.

7. The survey specifically asked whether the firm produces

a final good or an intermediate good.

Globalization of Production and Innovation 249

8. Here the survey first asked if the firm has pursued any

foreign direct investment, distinguishing between: (1)

production facilities, (2) an R&D laboratory, and (3) a

sales office.

9. In the survey, whenever the respondent declared the firm

belonged to a group, he/she was asked to specify the pos-
ition of the firm within the group.

10. In the regressions, all continuous variables are measured

in logs: to treat observations with zero value in some of

the relevant variables, a ‘1’ was added before taking the

logarithm.

11. MOL (2005) finds a positive and significant effect of pro-

ductivity on outsourcing. However, in his regressions the

coefficient of productivity is effectively zero.

12. This is again obtained from a survey question asking for

the country of residence of the subjects retaining either

the ownership or the control of the firm.

REFERENCES

ABRAHAM K. and TAYLOR S. (1996) Firms’ use of outside contractors: theory and evidence, Journal of Labor Economics 14, 394–424.

AJAYI D. D. (2005) Integration, industrial linkages and production subcontracting: an overview, European Journal of Social Sciences 1,

42–61.

AMIGHINI A. and RABELLOTTI R. (2006) How do Italian footwear industrial districts face globalization?, European Planning Studies

14, 485–502.

AMITI M. and WEI S. (2005) Fear of service outsourcing: is it justified?, Economic Policy 20(42), 308–347.

ARNDT S.W. and KIERZKOWSKI H. (Eds) (2001) Fragmentation: New Production and Trade Patterns in the World Economy. Oxford

University Press, Oxford.

BARNEY J. B. (1999) How a firm’s capabilities affect boundary decisions, Sloan Management Review 40, 137–145.

BEYERS W. B. and LINDAHL D. P. (1996) Explaining the demand for producer services: is cost-driven externalization the major

force, Papers in Regional Science 75, 351–374.

BOSCHMA R. A. (2004) Competitiveness of regions from an evolutionary perspective, Regional Studies 38, 1001–1014.

BRUSONI S., PRENCIPE A. and PAVITT K. (2001) Knowledge specialization, organizational coupling, and the boundaries of the firm:

why do firms know more than they make?, Administrative Science Quarterly 46, 597–621.

CAMAGNI R. (1991) Innovation Networks: Spatial Perspectives. Belhaven, London.

CHANDLER A. (1977) The Visible Hand: The Managerial Revolution in American business. Belknap/Harvard University Press,
Cambridge,

MA.

COE N. M. (2000) The externalisation of producer services debate: the UK computer services sector, Services Journal 20, 64–81.

COFFEY W. and BAILLY A. (1992) Producer services and systems of flexible production, Urban Studies 29, 857–868.

COOMBS R. and METCALFE J. S. (1998) Distributed Capabilities and the Governance of the Firm. Discussion Paper No. 16, Centre for

Research on Innovation and Competition (CRIC), University of Manchester, Manchester.

CORÒ G. and GRANDINETTI R. (1999) Strategie di delocalizzazione e processi evolutivi nei distretti industriali italiani, L’Industria 4,

897–924.

CUSMANO L. and MALERBA L. (2005) Le sfide strategiche per l’innovazione in Lombardia, in Società, Governo e Sviluppo del Sistema

Lombardo: dieci anni di esperienze, pp. 151–154. Guerini e Associati for the Istituto Regionale di Ricerca (IRER) della

Lombardia, Milan.

FEENSTRA R. C. (1998) Integration of trade and disintegration of production in the global economy, Journal of Economic Perspectives

12, 31–50.

FEENSTRA R. C. and HANSON G. (1996) Foreign investment, outsourcing and relative wages, in FEENSTRA R. C., GROSSMAN G. M.

and IRWIN D. A. (Eds) Political Economy of Trade Policy: Essays in Honor of Jagdish Bhagwati, pp. 89–127. MIT Press, Cambridge,

MA.

FREEDMAN P. (2004) The age of the hollow company, The Sunday Times, 25 April.

FREEMAN C. and SOETE L. (1997) The Economics of Industrial Innovation. Pinter, London.

GAROFOLI G. (1983) Industrializzazione diffusa in Lombarida. Istituto Regionale di Ricerca (IRER) della Lombardia/Franco Angeli,

Milan.

GAROFOLI G. (2002) Local development in Europe: theoretical models and international comparisons, European Urban and Regional

Studies 9, 225–239.

GEREFFI G. (1999) International trade and industrial upgrading in the apparel commodity chain, Journal of International Economics

48, 37–70.

GEREFFI G. and STURGEON T. J. (2004) Globalization, employment and economic development: a briefing paper. Paper presented

at the Sloan Workshop Series in Industry Studies, Rockport, MA, USA, 14–16 June 2004.

GILLEY K. M. and RASHEED A. (2000) Making more by doing less: an analysis of outsourcing and its effects on firm performance,

Journal of Management 26, 763–790.

GIRMA S. and GÖRG H. (2004) Outsourcing, foreign ownership, and productivity: evidence from UK establishment-level data,

Review of International Economics 12, 817–832.

GÖRG H., HANLEY A. and STROBL E. (2004) Outsourcing, Foreign Ownership, Exporting and Productivity: An Empirical Investigation with

Plant Level Data. Research Paper No. 2004/98, Leverhulme Centre for Research on Globalisation and Economic Policy.
University of Nottingham, Nottingham.

GROSSMAN G. M. and HELPMAN E. (2002) Outsourcing in a Global Economy. Working Paper No. 8728, January, National Bureau of

Economic Research (NBER), Cambridge, MA.

250 Lucia Cusmano et al.

HOWELLS J. (2000) The Nature of Innovation in Services. Report presented to the Organisation for Economic Co-operation and

Development (OECD) ‘Innovation and Productivity in Services’ Workshop, Sydney, Australia (available at: http://www.
oecd.org/dsti/sti/industry/indcomp).

ILLERIS S. (2005) The role of services in regional and urban development: a reappraisal of our understanding, Services Industries

Journal 25, 447–460.

IMRIE R. F. (1986) Work decentralization from large to small firms: a preliminary analysis of subcontracting, Environment and Plan-

ning A 18, 949–965.

ISTAT (2001) Ottavo Censimento Generale dell’Industria e dei Servizi. ISTAT (Italian National Institute of Statistics), Rome.

KIMURA F. (2002) Subcontracting and the performance of small and medium firms in Japan, Small Business Economics 18, 163–175.

LAURSEN K. and SALTER A. J. (2004) Searching high and low: what types of firms use universities as a source of innovation?,

Research Policy 33, 1201–1215.

LEIBLEIN M., REUER J. and DALSACE F. (2002) Do make or buy decisions matter? The influence of organizational governance on

technological performance, Strategic Management Journal 23, 817–833.

LEUNG C. K. (1993) Personal contacts, subcontracting linkages, and development in the Hong Kong–Zhujiang Delta region,

Annals of the Association of American Geographers 83, 272–302.

MAHNKE V. (2001) The process of vertical dis-integration: an evolutionary perspective on outsourcing, Journal of Management and

Governance 5, 353–379.

MARTIN R. (1999) The new ‘geographical turn’ in economics: some critical reflections, Cambridge Journal of Economics 23, 65–91.

MARTINEZ ARGÜELLES S. R. and RUBIERA MOROLLÓN F. (2006) Outsourcing of advanced business services in the Spanish

economy: explanation and estimation of the regional effects, Service Industries Journal 26, 267–285.

MASKELL P. and MALMBERG A. (1999) Localised learning and industrial competitiveness, Cambridge Journal of Economics 23,

167–185.

MAZZANTI M., MONTRESOR S. and PINI P. (2006) The General Profile of the Outsourcing Firm: Evidence for a Local Production System of

Emilia Romagna. Danish Research Unit for Industrial Dynamics (DRUID), Working Paper No. 06-20, Department of Indus-

trial Economics and Strategy, Copenhagen Business School/Department of Business Studies, Aalborg University, Aalborg.
MINIAN I. (2006) Nueva división internacional del trabajo: redes, segmentación y localización, Economia Marche, Review of Regional

Studies 25, 27–52.

MOL M. J. (2005) Does being R&D intensive still discourage outsourcing? Evidence from Dutch manufacturing, Research Policy

34, 571–582.

MORGAN K. (1997) The learning region: institutions, innovation and regional renewal, Regional Studies 31, 491–503.

NARULA R. (2001) In-house R&D, outsourcing or alliances? Some strategic and economic considerations, in CONTRACTOR F.

(Ed.) Valuation of Intangible Assets in Global Operations, pp. 101–122. Quorum, Westport, CT.

O’FARRELL P. N., MOFFAT L. A. R. and HITCHENS D. M. W. N. (1993) Manufacturing demand for business services in a core and

peripheral region: does flexible production imply vertical disintegration of business services?, Regional Studies 27, 385–400.

OLSEN K. B. (2006) Productivity Impacts of Offshoring and Outsourcing: A Review. Technology and Industry Working Papers No.

2006/1, Organisation for Economic Co-operation and Development (OECD) Directorate for Science, Technology and
Industry, Paris.

ONO Y. (2007) Market thickness and outsourcing services, Regional Science and Urban Economics 37, 220–238.

ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT (OECD) (1998) Internationalisation of Industrial R&D: Patterns

and Trends. OECD, Paris.

PERKMANN M. (2006) Extra regional linkages and the territorial embeddedness of multi-national branch plants: evidence from the

South Tyrol region in North-East Italy, Economic Geography 82, 421–441.

PHELPS N. A. (1993) Branch plants and the evolving spatial division of labour: a study of material linkage change in the northern

region of England, Regional Studies 27, 87–101.

PHELPS N. A. (2004) Clusters, dispersion and the spaces in between: for an economic geography of the banal, Urban Studies 41,

971–989.

PHELPS N. A. and OZAWA T. (2003) Contrasts in agglomeration: proto-industrial, industrial and post-industrial forms compared,

Progress in Human Geography 27, 583–604.

SCOTT A. J. (1988) Flexible production systems and regional development: the rise of new industrial spaces in North America and

Western Europe, International Journal of Urban and Regional Research 12, 171–186.

SCOTT A. J. (2005) Les moteurs régionaux de l’économie mondiale, Géographie, économie, société 7, 231–253.

SORENSON O., RIVKIN J. W. and FLEMING L. (2006) Complexity, networks and knowledge flow, Research Policy 35, 994–1017.

STORPER M. and SCOTT A. J. (1989) The geographical foundation and social regulation of flexible production complexes, in

WOLCH J. and DEER M. (Eds) The Power of Geography, pp. 21–40. Unwin Hyman, Boston, MA.

SWENSON D. (2004) Entry costs and outsourcing decisions: evidence from the U.S. overseas assembly provision, North American

Journal of Economics and Finance 15, 267–286.

TAYMAZ E. and KILIÇASLAN Y. (2005) Determinants of subcontracting and regional development: an empirical study on Turkish

textile and engineering industries, Regional Studies 39, 633–645.

TEECE D. J. (1986) Profiting from technological innovations, Research Policy 15, 285–305.

TETHER B. S., HIPP C. and MILES I. (2001) Standardisation and particularisation in services: evidence from Germany, Research Policy

30, 1115–1138.

TOMIURA E. (2005) Foreign outsourcing and firm-level characteristics: evidence from Japanese manufacturers, Journal of the

Japanese and International Economies 19, 255–271.

Globalization of Production and Innovation 251

TORRE A. and RALLET A. (2005) Proximity and localization, Regional Studies 39, 47–59.

UNITED NATIONS CONFERENCE ON TRADE AND DEVELOPMENT (UNCTAD) (2004) The Shift Towards Services. World Investment

Report, United Nations, Geneva.

WILLIAMSON O. E. (1985) The Economic Institutions of Capitalism. Free Press, New York, NY.

WOOD P. A. (1991) Flexible accumulation and the rise of business services, Transactions of the Institute of British Geographers n.s 16,

160–172.

WOOD P. A., BRYSON J. and KEEBLE D. (1993) Regional patterns of small firm development in the business services: evidence from

the UK, Environment and Planning A 25, 677–700.

252 Lucia Cusmano et al.

multiple sites or posted to a listserv without the copyright holder’s express written permission. However, users

may print, download, or email articles for individual use.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Production Globalization and Income Inequality

The Globalization of Production and Income
Inequality in Rich Democracies

Matthew C. Mahutga, University of California, Riverside
Anthony Roberts, California State University, Los Angeles
Ronald Kwon, University of California, Riverside

Despite prominent and compelling theoretical arguments linking manufacturingimports from the global South to rising income inequality in the global North,the literature has produced decidedly mixed support for such arguments. We
explain this mixed support by introducing intervening processes at the global and
national levels. At the global level, evolving characteristics of global production net-
works (GPNs) amplify the effect of Southern imports. At the national level, wage
coordination and welfare state generosity counteract the mechanisms by which
Southern imports increase inequality, and thereby mitigate their effects. We conduct
a time-series cross-section regression analyses of income inequality among eighteen
advanced capitalist countries to test these propositions. Our analysis addresses
alternative explanations, as well as validity threats related to model specification,
sample composition, and measurement. We find substantial variation in the effect of
Southern imports across global and national contexts. Southern imports have no sys-
tematic effect on income inequality until the magnitude of GPN activity surpasses its
world-historical average, or in states with above-average levels of wage coordina-
tion and welfare state generosity. With counterfactual analyses, we show th

Southern imports would have led to much different inequality trajectories in the
North if there were fewer GPNs, and if the prevailing degrees of wage coordination
and welfare state generosity were higher. The countervailing effects of GPNs and
institutional context call for theories of inequality at the intersection of the global and
the national, and raise important questions about distributional politics in the years to
come.

The authors thank John P. Boyd, David Brady, Lane Kenworthy, Christopher Kollmeyer, Jonus
Pontusson, Evan Schoefer, Andrew Schrank, Frederick Solt, David Swanson, attendees of the panel
on globalization and inequality at the 2014 annual meeting of the American Sociological Association
and writing workshop at the Social Science Research Center, Berlin (WZB) and anonymous Social
Forces reviewers. This research was funded by the National Science Foundation, grant number
1528703. Send questions and comments to Matthew C. Mahutga:matthew.mahutga@ucr.edu
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Social Forces

96(1)

181–214, September 2017
doi: 10.1093/sf/sox041

Advance Access publication on 25 May 2017

Production Globalization and Income Inequality 181

mailto: matthew.mahutga@ucr.edu

Introduction
Rising inequality is one of the most salient social changes among rich democra-
cies. From 1980 to 2007, the Gini coefficient for post-tax and transfer household
income inequality increased by roughly 24 percent in the United States, 19 per-
cent in Australia, 13 percent in Belgium, 11 percent in Canada, 20 percent in
Finland, 15 percent in Germany, 34 percent in the UK, and just 2 percent in
Austria. While rising inequality was the norm in rich democracies, income
inequality declined by 8 percent in Denmark, 4 percent in France, 2 percent in
Norway, and 5 percent in Switzerland (author’s calculations from Solt [2009]).
After decades of detailed research, however, leading economic policymakers
admit that “understanding the sources of the long-term tendency toward greater
inequality remains a major challenge,” a point echoed more recently by the US
Congressional Budget Office (Bernanke 2007; CBO 2011).

One of the earliest explanations for rising inequality was the globalization of
production, and in particular increases in Southern manufacturing imports gener-
ated by the offshoring behavior of Northern firms. As we detail below, well-
developed theories of trade suggest that Southern imports should increase inequal-
ity in rich democracies by widening the wage gap between high- and low-skill
labor, and by reducing the labor share of income as a whole. Paradoxically, how-
ever, the empirical evidence on the distributional effects of Southern imports is
decidedly mixed. Among the earliest proponents, Adrian Wood (

199

4) estimated
that imports from the global South were the most important driver of rising
inequality in the North, driven primarily by their effect on the relative demand for
skilled and unskilled labor. Others observed more muted effects—Southern im-
ports increase inequality, but primarily in non-European countries (Gustafsson
and Johansson 1999; also see Lin and Tomaskovic-Devey [2013]) or have smaller
effects than labor market institutions, sectoral composition, and other factors (e.g.,
Alderson and Nielsen 2002; Krugman 1995). Still others observe no effect of
Southern imports on inequality (e.g., Lee, Kim, and Shim 2011; Mahler 2004). For
these reasons, former Fed Chairman Bernanke (2007) concluded that the distribu-
tional consequences of production globalization remain an “open question.”

In this article, we take the mixed empirical support for the distributional ef-
fects of production globalization as a puzzle in need of explanation. We argue
that two types of intervening processes moderate its distributional consequences.
The first is the expansion and consolidation of global production networks
(GPN) worldwide, which have become a modal form of industrial organization.1

Manufacturing is increasingly embedded within networks of interfirm relations
that incorporate a greater proportion of the global South over time, and under-
mine the bargaining position of Southern firms. Both processes heighten the
downward pressure of Southern imports on low-skill wages in the North. Thus,
the expansion of GPNs exacerbates the effect of Southern imports on income
inequality in rich democracies. The second is egalitarian institutions—wage
coordination and welfare states—at the national level. The direct, equalizing ef-
fects of these institutions are well understood (Wallerstein 1999; Alderson and
Nielsen 2002; Allan and Scruggs 2004). However, we introduce a new set of

182 Social Forces 96(1)

moderating mechanisms by which both processes interrupt the market forces by
which production globalization should increase inequality. In short, we argue
that production globalization has inconsistent effects in previous research
because it interacts with these intervening processes.

To subject these arguments to empirical scrutiny, we conduct a time-series
cross-section regression analysis of post-tax and transfer income inequality
among eighteen rich democracies. The results support our interventions. The
inequality effect of Southern imports increases with the worldwide consolidation
of GPNs, and decreases with the degree of wage coordination and welfare state
generosity across countries. These results are robust to a host of socioeconomic
and sociopolitical explanations for income inequality, econometric, measure-
ment, and sampling considerations. Moreover, these moderating effects are sub-
stantial. When we decompose the effect of Southern imports across levels of
each moderating condition, we find they have no significant effect on inequality
until GPN consolidation surpasses the world-historical average, or in states with
above-average levels of wage coordination and welfare state generosity.
Similarly, inequality would have followed a much different trajectory if the rate
of GPN consolidation and the prevailing degrees of wage coordination and wel-
fare state generosity were different from what we observe.

As we elaborate in the concluding section, the countervailing moderating ef-
fects of global and national context force us to move beyond debates about the
relative importance of domestic and global drivers of inequality. That is, our re-
sults suggest the need for theories of inequality at the intersection of the global
and the national. They also raise important questions about distributional poli-
tics in the years to come.

Theories of the Distributional Effects of

Production
Globalization

There are two key mechanisms by which production globalization should
increase inequality in the North. The first draws largely from Heckscher–Ohlin
(H-O) trade theory. International trade reduces the price of production factors
toward that which prevails in the countries where they are most abundant.
Because unskilled labor is relatively abundant in the global South, Southern im-
ports reduce the demand for (relatively more expensive) unskilled labor in the
North (Alderson and Nielsen 2002; Wood 1994). At the same time, Southern
imports increase the demand for skilled labor in the North. In tandem, these
changes in the relative demand for skilled and unskilled labor increase inequality
by reducing the relative wages of low-skilled labor. That is, they increase
inequality within the working class.

The second mechanism is rooted in sociological theories involving the social re-
lations among labor, management, and capital. Here, Southern imports effectively
expand the size of the labor market beyond national borders. Because this expan-
sion includes workers in the global South, where workers have lower wages and
social protections on average, it increases labor market competition among

Production Globalization and Income Inequality 183

industrial workers in the Northern countries. That is, Southern imports incorpo-
rate large reserves of “surplus” industrial labor in Southern countries. This re-
duces the aggregate bargaining power of labor in developed countries. Because
reductions in the bargaining power of labor reduce the labor share of income vis-
à-vis capital and/or management, they also increase inequality (Elsby, Hobijn, and
Şahin 2013; Lin and Tomaskovic-Devey 2013; Tomaskovic-Devey and Lin 2011).
Southern imports should also increase inequality between labor and capital.

In theory, then, Southern imports should have large distributional effects.
Despite these strong theoretical expectations, however, empirical investigations
are less than conclusive. Some analyses find substantial effects, while others
observe relatively small or no significant effects (e.g., Alderson and Nielsen
2002; Elsby, Hobijn, and Şahin 2013; Gustafsson and Johansson 1999;
Krugman 1995; Lee, Kim, and Shim 2011; Mahler 2004; Massey 2009; Spence
and Hlatshwayo 2011; Wood 1994). We attribute this mixed empirical support
to intervening processes at the global and national levels. In the next two sec-
tions, we introduce global production networks and institutional context as key
intervening factors that produce variation over time and space in the distribu-
tional consequences of the globalization of production.

Global Production Networks and the Inequality Effect
of Southern Imports
Global production networks are increasingly central to the organizational strate-
gies of leading firms in nearly all manufacturing industries (e.g., Bair 2009; Gereffi,
Humphrey, and Sturgeon 2005; Mahutga 2014b). Social scientists attempt to mea-
sure this dynamic in various ways, including intra-firm trade as a percentage of
total trade and industry-specific trade-based metrics that capture particular models
of network “governance” (Feenstra 1998; Mahutga 2012; Milberg 2004).
Figure 1 graphs the trend in a very general metric of GPN consolidation—the ratio

Figure 1. Consolidation of globally networked models of economic organization

20
40

60
80

10
0

12
0

W
or

ld
M

an
uf

ac
tu

rin
g

Im
po

rt
s/

W
or

ld
M

an
uf
ac
tu
rin
g

V
al

ue
A

dd
ed

197

0 1980 1990 2000 2010

Note: Trade data are from UNCOMTRADE. Value-added data are from UNIDO (2015).

184 Social Forces 96(1)

of world manufacturing trade to world value added in manufacturing (Feenstra
1998; Mahutga 2012). The ratio of global trade to global value added increases
with the degree of production globalization because “intermediate inputs cross
borders several times during the manufacturing process… [and] while the denomi-
nator is value-added, the numerator is not, and will “double count” trade in com-
ponents and the finished product” (Feenstra 1998, p. 34; Mahutga 2012). That is,
the divergence of global trade from value added is proportional to the degree that
finished and intermediate inputs cross national borders multiple times in the pro-
duction process. The greater the divergence, the more manufacturing is organized
via GPNs.2 According to figure 1, GPNs are increasingly consolidated, and much
of this has occurred in the past thirty years. In 1970, 26.74 percent of world value
added in manufacturing was traded. This ratio climbed to 43.5 percent by 1980,
56.33 percent by 1990, 84.79 percent by 2000, and 126.55 percent by 2008.

The worldwide consolidation of GPNs should exacerbate the distributional
effect of Southern imports. First, recall that, in theory, Southern imports increase
inequality in part by driving down the wages of low-skill labor in the North, a
dynamic that should increase with the low-skill wage gap in the North and
South. The diffusion of GPNs has led to “industrial upgrading” in the global
South, where the number of capable suppliers and their geographic distribution
has increased dramatically over time. As such, factories migrate from higher- to
lower-wage Southern countries (Schrank 2004), which increases the low-skill
wage gap between the North and South directly. The greater supply of capable
suppliers also increases this gap through indirect channels: holding the number
of leading firms fixed, an increase in the number of capable suppliers generates
asymmetrical bargaining relations between leading firms and their Southern sup-
pliers (Mahutga 2014a). This allows lead firms to secure price concessions from
Southern suppliers, which decreases Southern low-skill wages even further
(Anner, Bair, and Blasi 2013; Schrank 2004). In short, GPN consolidation inte-
grates increasingly lower-wage countries into GPNs, and reduces the bargaining
power of Southern firms. Both of these processes increase the downward pres-
sure of Southern imports on low-skill wages in the North.

Second, the amount of economic activity coordinated by GPNs has
increased over time (Gereffi, Humphrey, and Sturgeon 2005; Mahutga
2014b; Milberg 2004; Yeung and Coe 2015). This should interact with the
second primary mechanism by which production globalization increases
inequality—its negative effect on the bargaining power of labor—even among
Northern workers who are not in direct competition with Southern workers.
Standard theories of wage variation start with negotiations between workers
and management over the terms of employment (Fernandez and Glazer
1991). Workers who possess skills that are relatively scarce, or who reside in
occupations with high demand, possess more bargaining power, and there-
fore command higher remuneration, than workers who possess abundant
skills or reside in occupations with little demand (Wright 2000). However,
the labor market return to these resources depends on individual variation in
bargaining behavior. As an increasing amount of economic activity becomes
coordinated via GPNs, workers come to believe that jobs are increasingly

Production Globalization and Income Inequality 185

vulnerable to offshoring, and therefore experience heightened perceptions of
economic insecurity (Milberg and Winkler 2009; Scheve and Slaughter
2004). Heightened perceptions of economic insecurity cause workers to
accept lower rates of remuneration on average, which reduces the labor share
of income and increases income inequality (Riedl 2013).

We formalize our arguments about the moderating effect of GPN consolida-
tion with the following hypothesis:

H1: The effect of Southern imports should increase with the consolida-
tion of networked forms of economic organization at the global level.

Institutional Context and the Inequality Effects
of Southern Imports

Wage Coordination

Rich democracies vary along institutional dimensions known to matter for a
range of political economic outcomes (e.g., Epsing-Anderson 1990; Hall and
Soskice 2001; Western 1997). One institution stands out as important for
income inequality—wage coordination among labor, capital, and sometimes the
state (Alderson and Nielsen 2002; Kenworthy 2001; Mahler 2004; Wallerstein
1999). Examples of wage coordination include industry-level wage bargaining
through formal relations between capital, peak labor confederations (Austria), or
large unions from influential industries (Germany); between employer confedera-
tions and large firms (Japan and Switzerland); or by government imposition of
wage schedules or freezes (e.g., Belgium, Denmark, and the Netherlands) (Traxler
1999). Wage coordination limits wage variation within the private sector as well
as the income gap between labor and capital. Indeed, a negative association
between wage coordination institutions and income inequality has been a persis-
tent finding in the comparative political economy literature (Alderson and Nielsen
2002; Bradley et al. 2003; Kenworthy and Pontusson 2005; Wallerstein 1999).

As a point of departure, we introduce additional channels though which these
institutional arrangements should lower inequality.3 Recall that production
globalization should increase inequality by reducing both (a) the relative wages
of unskilled workers and (b) the bargaining power (and thus income share) of
labor as a whole. In terms of the distribution of income between low- and high-
wage workers, the distributional effects of production globalization should
depend critically on the extent that wages respond freely to changes in labor
demand (Mahutga and Jorgenson 2016). In countries where wage coordination
is the norm, changes in output and productivity brought on by competition from
Southern imports are, to varying degrees, “decoupled” from wages: “a wage
agreement covering a work force of any size must specify a general rule” by
which wages will be determined over the agreement period (Wallerstein 1999,
p. 673). Even in the hypothetical (and unobserved) scenario where wage coordi-
nation is regressive (i.e., results in a higher degree of dispersion than would be
the case in the absence of wage coordination), the fact that wages are set through

186 Social Forces 96(1)

institutional negotiations means they cannot respond instantaneously to changes
in demand for particular segments of labor.

In terms of the labor share of income, strong wage-coordinating institutions
shift the locus of control over remuneration from firms to labor, and foster col-
lective identity among differentiated workers (Wallerstein 1999). This represents
an institutional source of bargaining power that should reduce the downward
pressure of Southern imports on the labor share of income. The moderating
effect of wage coordination should be particularly strong with respect to the
labor share of income because it has been shown to benefit the wages of
those most negatively impacted by Southern imports—low-skill workers—
disproportionately (Wallerstein 1999). Thus:

H2: The effect of Southern imports should decline with increases in
wage coordination.

However, recent scholarship argues that wage coordination systems are declining
in their significance for inequality. Capitalist firms in coordinated states might opt
out of wage bargaining altogether, refuse to extend bargained wage increases to
unrepresented workers, or the very nature of coordinated bargaining systems may
change in significant ways. Here, the working-class solidarity underlying the mod-
erating effect of wage-coordinating institutions breaks down between “core work-
ers who have jobs and who are intent on preserving their relatively privileged
position within the labor market, and labor market ‘outsiders’ who either do not
have jobs or are in more precarious forms of employment and thus do not enjoy
the same package of wages and benefits as insiders” (Thelen 2012, p. 149; Rueda
2007). As a result, historically strong wage-coordinating systems might produce
labor market dualism, where wage coordination may only equalize the core seg-
ment labor market “insiders,” which may also enjoy higher average wages than
labor market “outsiders.” Both scenarios would tend to push the moderating effect
of wage coordination toward zero and thus suggest a theoretically informed null
hypothesis (also see Huber and Stephens [2014]; Scheve and Stasavage [2009]).

The Welfare State
It is widely known that welfare transfers income from affluent to poor households
(Bradley et al. 2003; Kenworthy and Pontusson 2005). While these direct, egalitar-
ian effects are rather clear, we argue that strong welfare states should also weaken
the link from Southern imports to wage dispersion between skilled and unskilled
workers, and to the bargaining power (and thus income share) of labor. First,
strong welfare states should boost the disposable income of those most harmed by
production globalization—low-skill workers. Here, eligibility requirements under-
lying transfer payments in advanced industrial democracies are intrinsically pro-
gressive (to varying degrees), and thus disproportionately affect low-income
households. Because skills are highly correlated with incomes, transfer payments
increase the post-transfer incomes of low-skill vis-à-vis high-skill workers. Put dif-
ferently, generous welfare states reduce the impact of the wage effects of Sothern
imports on the post-transfer income gap between low- and high-skill workers.

Production Globalization and Income Inequality 187

Second, recall that, in theory, production globalization reduces the bargaining
power of labor, and exacerbates perceptions of economic insecurity among
Northern workers. In a simplified bargaining game, unemployed workers can either
come to terms on a given employment package or remain unemployed. In countries
with strong welfare states, the income penalty to unemployment is less pronounced
than in countries with weaker welfare states. Because unemployment comes with a
weaker income penalty, workers should be more willing to bargain better—they
have less to lose by asking for more. Indeed, micro-level evidence suggests that
strong welfare states mitigate perceptions of economic insecurity (Anderson and
Pontusson 2007; Mughan 2007). If strong welfare states facilitate more strategic
bargaining behavior among workers in the labor market, production globalization
should have a smaller negative effect on the labor share of income (and therefore
income inequality) in countries with strong welfare states. Thus, we expect that

H3: The effect of Southern imports should decline with increases in the
size and strength of the welfare state.

Data and Methods
Dependent Variable
Income inequality
Gini coefficients of post-tax and transfer income inequality are available in vari-
ous forms, but the most complete and cross-nationally/temporally comparable is
the Standardized World Income Inequality Database (SWIID) (Clark 2013; Solt
2009). The cross-national and temporal comparability of Gini coefficients is
made problematic by definitional variation across national surveys in terms of
the units of observation (household vs. individual), the definition of income, and
because of differences in survey quality. Solt’s approach utilizes all of the infor-
mation available from the World Income Inequality Database (WIID), regional
inequality databases, national statistical offices, and the scholarly literature,
along with high-quality estimates from the Luxembourg Income Study (LIS), to
inform a Monte Carlo multiple imputation procedure that harmonizes multiple
estimates of Gini, and gives a sense of the reliability of those harmonized
estimates.

Unlike other data sources, Solt’s Ginis (1) do not require the assumption that
Gini incomparability is constant across countries/time; (2) are benchmarked to
the most reliable Luxembourg Income Study estimates available; (3) treat
“quality” with continuous (rather than dichotomous) reliability estimates; and
(4) include many more cross-national and temporally comparable Ginis than
other sources. Because the LIS provides more Gini coefficients for developed
countries, the SWIID estimates for our sample are even more reliable than the
full sample. Nevertheless, we restrict our analysis to post-tax and transfer Gin

coefficients with standard errors less than 1 and assess the robustness of our re-
sults to this threshold and to alternative sources of Gini coefficients (see Solt
2009, p. 238).

188 Social Forces 96(1)

Independent Variable
Southern imports
A common measure of production globalization among advanced industrial
countries is the value of manufacturing imports from Southern countries (see
Alderson and Nielsen [2002]).4 However, trade scales linearly with country size,
which complicates comparisons across countries of vastly different economic
and geographical weight. A common approach to facilitate international com-
parisons of Southern imports is to normalize imports from Southern countries
(typically defined as non-OECD and non-COMECON countries) by gross
domestic product (GDP). We utilize an alternative procedure to facilitate inter-
national comparisons—we divide manufacturing imports from Southern (non-
OECD) countries by total imports. Our data on manufacturing imports from
Southern countries and total imports come from the OECD (2011a). We prefer
this normalization to GDP for two interrelated reasons.

First, normalizing Southern imports by total imports captures the pattern
rather than the level of trade, since total imports represent the maximum amount
of Southern imports possible for a given country (United Nations 2014a, p. 332;
see Beckfield [2006] on measuring EU economic integration with trade). Second,
recent empirical work finds that Southern imports increase GDP by increasing
profit rates among offshoring firms (Kollmeyer 2009a). This relationship allows
for the possibility that GDP increases disproportionately with increases in
Southern imports, such that the ratio of Southern imports to GDP could either
under- or overstate the degree that firms in a given country integrate Southern
workers into their supply chains.5 Contrarily, temporal variation in the ratio
Southern imports/total imports will depend only on the relative rate of growth
in Southern manufacturing to other types of imports.6 We nevertheless show the
robustness of our results to alternative measures below.

Moderating Variables
Global Production Network Consolidation
To measure the worldwide consolidation of GPNs, we follow Feenstra (1998)
and Mahutga (2012) by employing the ratio of world trade in manufacturing to
world value added in manufacturing, as displayed in figure 1. Data on world
trade come from the United Nations (2014b). Data on value added come from
the UNIDO’s Industrial Statistics database (UNIDO 2015). This covariate varies
over time, but not across countries.

We measure wage-coordination with Kenworthy (2001), and updated by
Huber et al. (1997, 2004, 2014). Scores ranged from 1 to 5, with 1 indicating
fragmented bargaining at the plant level and 5 indicating centralized bargaining
among large union and business confederations, or government-imposed wage
schedules. This is the most preferred measure for capturing the institutionaliza-
tion of wage-coordination processes because it is a measure of the institutional
capacity to coordinate rather than the degree of achieved coordination, and

Production Globalization and Income Inequality 189

because of its ability to capture the diversity of institutional arrangements con-
ducive to coordination (Kenworthy 2001).

Welfare State Generosity

We measure the welfare state with the updated generosity index (Scruggs, Jahn,
and Kuitto 2014), which expands on and updates Epsing-Anderson’s (1990) de-
commodification index. As opposed to measuring transfer payments directly,
the “generosity index” combines information on benefit replacement rates, qual-
ifying conditions, and elements of the insurance coverage or take‐up rates for
unemployment, sickness, and retirement programs. More generous welfare
states are those that provide relatively large outlays for longer periods of time,
and have minimal eligibility requirements.

Control Variables
Baseline Controls
There are a host of factors that influence cross-national and temporal levels of
income inequality. First, we control for the harmonized unemployment rate
(OECD 2011b). Unemployment should correlate positively with income inequal-
ity insofar as a loss of income among a large portion of the economically active
population should inflate the lower end of the income distribution.

Existing explanations for the inequality upswing in developed countries evoke
changes in the age and gender composition of the labor force. Given the positive
correlation between age and income, the aging of the labor force should expand the
gap between older and younger citizens (Rubin, White-Means, and Daniel 2000).
Alternatively, competing theoretical narratives argue that an increase in female
labor force participation might either increase or decrease inequality (e.g., Alderson
and Nielsen 2002). Thus, we also control for the elderly population (% 65+) and
female labor force participation from data available from the OECD (2011a).

Three important changes to the economic and labor market governance struc-
tures of advanced industrial countries have occurred simultaneously with eco-
nomic globalization. Financialization has been shown to contribute to income
inequality in the United States (Lin and Tomaskovic-Devey 2013; Tomaskovic-
Devey and Lin 2011) and other advanced industrial countries (Kus 2013). Thus,
we follow Lee, Kim, and Shim (2011) by controlling for the percentage of the
labor force in the Finance, Insurance, and Real Estate (FIRE) sector (OECD
2011b). Similarly, rich democracies experienced varying rates of deindustrializa-
tion and union decline, both of which have been shown to matter for inequality
elsewhere (Alderson and Nielsen 2002). Thus, we employ data on the percent of
the labor force in industry (OECD 2011b), and union density (Visser 2011).

Additional Controls
A venerable tradition in sociology finds that inequality is a function of internal
developmental processes, operationalized as the percent of the labor force
in agriculture, sector dualism, the natural rate of population increase, and

190 Social Forces 96(1)

secondary education (Alderson and Nielsen 2002). The first two control for the
distributional effects of the migration of labor from the agricultural to the
manufacturing sector. The latter two control for the distribution of skills and
the size of the surplus labor pool. Data on these measures come from the World
Development Indicators (World Bank 2016).

Finally, we also control for institutional and political processes associated
with the distribution of income (see Bradley et al. 2003; Huber and Stephens
2014; Lee, Kim, and Shim 2011). Power resource theory suggests that partisan
politics play a key role in distributional outcomes. Leftist governments, in partic-
ular, reduce post-tax and transfer income inequality by enacting policies to redis-
tribute wealth (see Bradley et al. [2003, pp. 195–96]). Thus, we control for the
relative strength of leftist parties with the cumulative cabinet share of leftist par-
ties (Huber et al. 1997, 2004, 2014).

Correlations and descriptive statistics appear in table A1.

Time-Series Cross-Section Regression
We conduct a time-series cross-section regression analysis of income inequality
among eighteen rich democracies. The sample of advanced industrial countries,

Table 1. Country-Years Included

Country Year

Austria 1993–2006

Belgium 1993–1999, 2001–2002

Canada 1975–1993

Denmark 1975–1979, 1981–2007

Finland 1975–2007

France 1975–2007

Germany 1991–2007

Ireland 1981–1999, 2001–2006

Italy 1975–2006

Japan 1975–1977, 1986, 1987, 1989, 1990, 2000–2006

Netherlands 1975–1979, 1981–2007

New Zealand 1990–1998

Norway 1975–2007

Portugal 2003, 2006

Sweden 1975–2007

Switzerland 1991–2006

United Kingdom 1975–2004

United States 1975–2002

Total: N = 18; n = 411

Production Globalization and Income Inequality 191

listed in table 1, includes most of Western Europe, Japan, the United States,
Canada, Australia, and New Zealand (e.g., Alderson and Nielsen 2002; Lee,
Kim, and Shim 2011; Western 1997). The unit of observation in the time-series
cross-section regression is the country-year. As is evident in table 1, the panels
are unbalanced. Due to missing data on the left- and right-hand side, and our
exclusion cases with Gini standard errors greater than 1, our models contain a
maximum of 411 observations, as described in table 1.

A clear strength of this design is that it allows us to control for omitted
unobservable covariates that vary across countries but not over time. The
inclusion of fixed country effects in the models that follow correct for this
source of omitted-variable bias. Such data typically yield heteroskedastic and
serially correlated disturbance terms (Wooldridge 2002). Employing the avail-
able identification tests in Stata 14.0, we examined the error structure of our
models and rejected the null hypothesis of homoscedasticity and zero serial
correlation. To correct heteroskedasticity, we employ robust standard errors.
To correct for serially correlated errors, we employ a first-order (AR1) auto-
correlation correction with a Prais–Winston transformation that accounts for
the unbalanced panels. To address biases owing to unobserved covariates that
vary across time but not over countries, we also include decadal dummies (Lee,
Kim, and Shim 2011).

Testing the hypotheses that the consolidation of GPNs, wage coordination,
and welfare states moderates the impact of Southern imports on income
inequality is straightforward in this design. We regress income inequality on
interaction terms between Southern imports and each covariate, along with
relevant controls (Friedrich 1982; Lee, Kim, and Shim 2011). To mitigate
collinearity between constituent and interaction terms, we mean-deviated
Southern imports, GPN Consolidation, Wage Coordination, and Welfare
Generosity. Because this amounts to subtracting a constant from each term,
this does not affect the coefficients on the interaction terms (but see note 7).
Hypotheses 1–3 are explicitly directional, and we therefore employ direc-
tional hypothesis tests.

Table 2. Zero-Order Correlation between Southern Imports and GINI across High and Low
Global Production Network Consolidation, Wage Coordination, and Welfare State Generosity

Low High

GPN consolidation 0.186*** 0.217***

Wage coordination 0.422*** 0.315***

Welfare state generosity 0.551*** −0.027

Note: Observations country-mean deviated. Low GPN consolidation and welfare state
generosity is below median; high is median and above. GPN consolidation is drawn from
UNCOMTRADE (see figure 1). Welfare state generosity is from Scruggs, Jahn, and Kuitto
(2014). Low wage coordination is less than or equal to 3; high is 4 or 5. Wage coordination is
drawn from Huber et al. (1997, 2004, 2014). *** p < 0.001

192 Social Forces 96(1)

Results
Table 2 reports the correlation between Southern imports and inequality as it var-
ies across “low” and “high” levels of GPN consolidation, Wage Coordination,
and Welfare State Generosity. Consistent with hypothesis 1, the effect of Southern
imports increases as global production networks become consolidated among
Northern manufacturing firms. Similarly, the bivariate association between
inequality and Southern imports is smaller among countries with high levels of
wage coordination and generous welfare states. Indeed, Southern imports are
almost uncorrelated with income inequality in countries with high welfare state
generosity. Does apparent variation in the effect of Southern imports hold in
conservative econometric models that control for additional correlates of
inequality?

Model 1 in table 3 introduces Southern imports and our controls. Consistent
with the panoply of previous research, the effect is relatively small but significant
at conventional thresholds. Model 2 introduces an interaction term between
Southern imports and GPN consolidation to test whether the effect of Southern
imports increases with the worldwide consolidation of GPNs. The interaction
term is positive and highly significant. Model 3 introduces the interaction term
between Southern imports and wage coordination. The moderating effect of
wage coordination is negative and highly significant. Finally, model 4 includes
the interaction between Southern imports and welfare state generosity. The
interactive effect is negative and highly significant.7 Given the positive coefficient
on Southern imports in model 1, these results suggest that GPN consolidation
exacerbates the effect of Southern imports on income inequality, while wage
coordination and welfare state generosity ameliorate this effect.

Robustness
Alternative Estimator and Measure of Southern Imports
To assess the robustness of our results, we begin with models employing alter-
native corrections for unobserved country-invariant processes that vary over
time and the traditional measure of Southern imports in table 4. Relative to
this first concern, we estimated conservative two-way fixed effects models by
including a full set of T-1 dummy variables. These results (models 1–3) were
substantively and almost numerically identical.8 Relative to the second, we
examine interaction terms involving each of our three moderators and the
ratio of Southern imports to GDP. Data on the ratio of Southern imports to
GDP are drawn from the International Trade database (OECD 2011c).
Consistent with our discussion above, an unreported model including this
covariate without moderation yields a coefficient close to zero (b = 0.013;
p < 0.893). Moreover, the t-statistics for the interaction terms involving
Southern imports/GDP in table 4 are much smaller than those reported above.
Still, the interaction effects are significant and correctly signed: the effect of
Southern imports/GDP increases with GPN consolidation and decreases with
wage coordination and welfare state generosity.9

Production Globalization and Income Inequality 193

Table 3. Coefficients from Fixed Effects Regression of Gini on Southern Imports, Moderators,
and Select Independent Variables

(1) (2) (3) (4)

Southern imports (SPEN)a 0.075** 0.054* 0.070* 0.073**
(0.031) (0.031) (0.030) (0.030)

SPEN*GPN consolidation 0.004***
(0.001)

SPEN*Wage coordination −0.049***
(0.014)

SPEN*Welfare state generosity −0.014***
(0.003)

GPN consolidationa −0.026*
(0.014)

Wage coordinationa −0.240**
(0.079)

Welfare state generositya −0.172***
(0.033)

Unemployment 0.076* 0.090* 0.076* 0.077*
(0.041) (0.043) (0.041) (0.040)

Union density −0.035 −0.026 −0.025 −0.007
(0.023) (0.023) (0.022) (0.022)

Industrial employment 0.101* 0.113* 0.123* 0.088
(0.057) (0.058) (0.057) (0.056)

Female labor force participation 0.054*** 0.057*** 0.051*** 0.065***
(0.014) (0.014) (0.013) (0.013)

Elderly population 0.270** 0.123 0.249** 0.124
(0.105) (0.113) (0.102) (0.103)

FIRE sector employment 0.199* 0.326*** 0.208* 0.175*
(0.096) (0.099) (0.094) (0.093)

Agricultural employment 5.563** 4.765** 4.573** 4.004*
(1.918) (1.969) (1.918) (1.900)

Sector dualism −0.134 −0.147 −0.121 −0.156
(0.110) (0.112) (0.109) (0.106)

Secondary education 0.009 0.012 0.008 0.008
(0.008) (0.009) (0.008) (0.008)

Natural rate of population increase 0.236 0.052 0.015 0.356
(0.692) (0.688) (0.687) (0.670)

Cumulative left cabinet share 0.005 0.063 0.014 0.035
(0.052) (0.055) (0.050) (0.050)

1980s −0.197 −0.084 −0.270 −0.192
(0.237) (0.245) (0.241) (0.234)

1990s −0.013 0.240 −0.121 −0.104
(0.312) (0.330) (0.317) (0.310)

2000s −0.003 0.184 −0.181 −0.039
(0.386) (0.418) (0.395) (0.386)

(Continued)

194 Social Forces 96(1)

Gini Coefficient Quality
We also consider the extent that our results are robust to the quality of the Gini
coefficients we employ. We estimate six additional models, reported in table 5.
Models 1, 3, and 5 restrict the analysis to Gini coefficients with standard errors
less than 0.75. Models 2, 4, and 6 restrict the analysis to Gini coefficients with
standard errors less than 0.5. Reassuringly, our results are substantively, and
nearly numerically, identical to those reported in table 3.

Sample Composition and Gini Coefficient Measurement
Table 6 considers the sensitivity of our results to the composition of our sample
(models 1–12), and the source of our Gini coefficients (models 4–12). In models
1–3, we employ bootstrap confidence intervals in lieu of parametric hypothesis
tests. Bootstrap confidence intervals are constructed by estimating coefficients
on 500 unique samples, in which the entire set of each country’s observations
are randomly sampled with replacement. These coefficients then form a distribu-
tion of coefficients with which to calculate confidence intervals around our
observed coefficients. Hypothesis tests based on bootstrap confidence intervals
are reported next to the parametric standard errors in models 1–3 of table 6.
These confidence intervals are wider than those in table 3, but each interaction
term remains significant at conventional levels.

There are vastly fewer Gini coefficients available in the LIS data than in
SWIID, which makes it difficult to observe significant coefficients in the fairly
saturated models in table 3. Thus, models 4–6 report coefficients from a fixed ef-
fects regression of the LIS Ginis on Southern imports, each moderator, the inter-
actions, and the decadal dummies. In these models, there are roughly 75 percent
fewer observations than in table 3 and the sample composition varies signifi-
cantly, but the interaction terms are significant and correctly signed. Models 7–9
add the controls that were most consistent in table 3, which reduces the number
of observations to 103–6. The interaction terms remain significant. Models
10–12 include the full set of controls, which reduces the number of observations
to 86, and stretches the degrees of freedom beyond reason (the ratio n/k = 4.1,

Table 3. continued

(1) (2) (3) (4)

Constant 10.580* 13.182** 10.173* 14.318**
(5.043) (5.224) (5.049) (5.347)

ρ 0.755 0.718 0.731 0.719
N 411 411 411 411

R2 0.956 0.957 0.957 0.959

Note: Heteroskedasticity and serial correlation consistent standard errors in parentheses; * p < 0.05; ** p < 0.01; *** p < 0.001 (one-tailed tests). ρ is the first-order (AR1) auto-regressive term. aThis coefficient is the effect of the focal covariate when the other term in the interaction equals the sample mean.

Production Globalization and Income Inequality 195

Table 4. Sensitivity to Model Specification and Alternative Measure of Southern Imports

(1) (2) (3) (4) (5) (6)

Southern imports/total imports (SPEN1)a 0.062* 0.092** 0.097**
(0.035) (0.033) (0.033)

Southern imports/GDP (SPEN2)a −0.107 0.165 0.096
(0.115) (0.113) (0.103)

SPEN1*GPN consolidation 0.004***
(0.001)

SPEN1*Wage coordination −0.057***
(0.014)

SPEN1*Welfare state generosity −0.016***
(0.003)

SPEN2*GPN consolidation 0.005*
(0.003)

SPEN2*Wage coordination −0.179**
(0.065)

SPEN2*Welfare state generosity −0.022*
(0.012)

GPN consolidationa −0.072*** −0.003
(0.021) (0.014)

Wage coordinationa −0.243*** −0.213**
(0.077) (0.083)

Welfare state generositya −0.140*** −0.139***
(0.032) (0.033)

Female labor force participation 0.074*** 0.075*** 0.085*** 0.056*** 0.049*** 0.059***
(0.014) (0.014) (0.014) (0.014) (0.013) (0.014)

Unemployment 0.072* 0.054 0.062 0.075* 0.064 0.059
(0.044) (0.043) (0.042) (0.042) (0.040) (0.040)

Elderly population 0.242* 0.343*** 0.186* 0.349*** 0.315*** 0.302**
(0.116) (0.106) (0.110) (0.099) (0.096) (0.099)

FIRE sector employment 0.393*** 0.337*** 0.285** 0.218* 0.202* 0.150
(0.111) (0.109) (0.109) (0.100) (0.096) (0.097)

Union density −0.010 0.002 0.015 −0.042* −0.037 −0.034

196
SocialForces

96(1)

(0.024) (0.024) (0.023) (0.024) (0.022) (0.023)

Industrial employment 0.083 0.086 0.048 0.087 0.095* 0.059
(0.059) (0.059) (0.059) (0.058) (0.057) (0.057)

Agricultural employment 2.632 1.295 1.113 4.573* 4.021* 3.223*
(2.137) (2.141) (2.072) (1.967) (1.883) (1.883)

Sector dualism −0.053 −0.026 −0.063 −0.128 −0.088 −0.110
(0.111) (0.108) (0.105) (0.112) (0.107) (0.106)

Secondary education enrollment 0.008 0.010 0.011 0.011 0.009 0.008
(0.009) (0.009) (0.009) (0.009) (0.008) (0.008)

Natural rate of population increase −0.807 −1.185* −0.657 0.131 −0.068 0.131
(0.726) (0.719) (0.695) (0.702) (0.686) (0.686)

Cumulative left cabinet share 0.105* 0.070 0.094* −0.006 −0.020 −0.026
(0.055) (0.053) (0.052) (0.055) (0.050) (0.051)

1980s −0.184 −0.271 −0.121
(0.242) (0.241) (0.235)

1990s 0.041 −0.088 −0.007
(0.328) (0.315) (0.308)

2000s 0.098 −0.142 0.080
(0.419) (0.392) (0.379)

Fixed yearly effects Included Included Included — — —

Constant 9.087* 12.467** 14.719** 10.991* 13.032** 15.093**
(5.050) (5.124) (5.127) (5.046) (5.052) (5.063)

ρ 0.737 0.747 0.743 0.746 0.746 0.758

N 411 411 411 411 411 411

R2 0.961 0.961 0.963 0.956 0.956 0.927

Note: Heteroskedasticity and serial correlation consistent standard errors in parentheses; * p < 0.05; ** p < 0.01; *** p < 0.001 (one-tailed tests). ρ is the first-order (AR1) auto- regressive term. aThis coefficient is the effect of the focal covariate when the other term in the interaction equals the sample mean.

Production
Globalization

and
Incom

e
Inequality

197

Table 5. Sensitivity to Gini Coefficient Quality

(1) (2) (3) (4) (5) (6)

Southern imports (SPEN)a 0.050 0.038 0.066* 0.054* 0.075** 0.065*
(0.031) (0.032) (0.030) (0.030) (0.030) (0.030)

SPEN*GPN consolidation 0.004*** 0.004***
(0.001) (0.001)

SPEN*Wage coordination −0.051*** −0.052***
(0.013) (0.013)

SPEN*Welfare state generosity −0.014*** −0.014***
(0.003) (0.003)

GPN consolidationa −0.028* −0.023
(0.014) (0.014)

Wage coordinationa −0.228** −0.199**
(0.079) (0.078)

Welfare state generositya −0.193*** −0.199***
(0.035) (0.034)

Female labor force participation 0.060*** 0.060*** 0.056*** 0.056*** 0.071*** 0.071***
(0.013) (0.013) (0.013) (0.013) (0.013) (0.013)

Unemployment 0.129** 0.127** 0.117** 0.114** 0.121** 0.116**
(0.045) (0.045) (0.043) (0.043) (0.042) (0.042)

Elderly population 0.058 0.089 0.181* 0.202* 0.049 0.071
(0.113) (0.108) (0.103) (0.097) (0.103) (0.098)

FIRE sector employment 0.379*** 0.373*** 0.264** 0.266** 0.231** 0.227**
(0.100) (0.101) (0.095) (0.092) (0.094) (0.091)

Union density −0.043* −0.048* −0.037* −0.043* −0.016 −0.022
(0.023) (0.023) (0.022) (0.022) (0.022) (0.022)

198
SocialForces

96(1)

Industrial employment 0.151** 0.140** 0.165** 0.151** 0.132* 0.119*
(0.060) (0.060) (0.058) (0.058) (0.057) (0.057)

Sector dualism −0.061 −0.052 −0.048 −0.034 −0.096 −0.095
(0.120) (0.121) (0.116) (0.115) (0.111) (0.110)

Agricultural employment 3.617* 3.729* 3.564* 3.515* 2.987 2.844
(2.011) (2.029) (1.945) (1.922) (1.922) (1.896)

Secondary education enrollment 0.011 0.012 0.008 0.008 0.007 0.007
(0.009) (0.009) (0.008) (0.009) (0.008) (0.008)

Natural rate of population increase −0.781 −0.797 −0.807 −0.873 −0.470 −0.549
(0.721) (0.710) (0.721) (0.704) (0.704) (0.684)

Cumulative left cabinet share 0.050 0.029 −0.000 −0.012 0.018 0.006
(0.054) (0.054) (0.049) (0.048) (0.049) (0.049)

1980s −0.187 −0.211 −0.383 −0.386 −0.279 −0.301
(0.247) (0.242) (0.245) (0.241) (0.237) (0.230)

1990s 0.194 0.140 −0.203 −0.224 −0.156 −0.191
(0.335) (0.334) (0.325) (0.325) (0.316) (0.312)

2000s 0.096 0.028 −0.309 −0.312 −0.137 −0.152
(0.427) (0.434) (0.411) (0.416) (0.397) (0.398)

Constant 10.409* 10.691* 11.470* 11.748** 7.01 7.587
(5.114) (4.976) (5.133) (4.925) (5.126) (4.980)

ρ 0.688 0.677 0.687 0.676 0.683 0.673
N 401 381 401 381 401 381

R2 0.964 0.973 0.963 0.973 0.966 0.975

Note: Models 1, 3, and 5 include cases with Gini standard errors less than 0.75. Models 2, 4, and 6 include cases with Gini standard errors less than 0.5.
Heteroskedasticity and serial correlation consistent standard errors in parentheses; * p < 0.05; ** p < 0.01; *** p < 0.001 (one-tailed tests). ρ is the first- order (AR1) auto-regressive term. aThis coefficient is the effect of the focal covariate when the other term in the interaction equals the sample mean.

Production
Globalization
and
Incom
e
Inequality
199

Table 6. Sensitivity of Results to Sample Composition and Gini Source

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)

Bootstrap resamplingb Luxembourg Income Study Gini coefficients

Southern imports (SPEN)a 0.054* 0.070* 0.073** 0.095 0.090 0.128 0.003 0.080 0.053 0.084 0.251** 0.228**
(0.031)N.S. (0.030)N.S. (0.067)N.S. (0.098) (0.079) (0.079) (0.111) (0.079) (0.085) (0.106) (0.095) (0.085)

SPEN*GPN consolidation 0.004*** 0.003* 0.005* 0.008***
(0.001)++ (0.002) (0.002) (0.002)

SPEN*Wage coordination −0.050*** −0.084*** −0.056* −0.049†
(0.014)++ (0.025) (0.028) (0.032)

SPEN*Welfare state
generosity

−0.014*** −0.013** −0.016** −0.018**

(0.003)+ (0.005) (0.006) (0.007)

GPN consolidationa −0.026* −0.012 −0.039 −0.058
(0.014)N.S. (0.023) (0.035) (0.043)

Wage coordinationa −0.240** −0.340 0.010 −0.316
(0.079)+ (0.255) (0.222) (0.268)

Welfare state generositya −0.173*** −0.192** −0.350*** −0.321***
(0.033)+++ (0.071) (0.069) (0.084)

Controls Full Full Full Decades Decades Decades Decades, female labor, elderly Full Full Full
Population, agriculture employ

ρ 0.718 0.731 0.719 0.016 0.031 0.004 0.029 0.028 −0.007 0.012 0.010 −0.002

N 411 411 411 116 113 113 106 103 103 86 86 86

R2 0.958 0.957 0.959 0.880 0.872 0.888 0.887 0.878 0.905 0.923 0.916 0.930

Note: Heteroskedasticity and serial correlation consistent standard errors in parentheses; † p < 0.10; * p < 0.05; ** p < 0.01; *** p < 0.001 (one-tailed tests). ρ is the first-order (AR1) auto-regressive term. aThis coefficient is the effect of the focal covariate when the other term in the interaction equals the sample mean. bHypothesis tests based on bias corrected and accelerated (BCa) bootstrap confidence intervals reported next to parametric standard errors: +95 percent BCa confidence interval does not include zero; ++99 percent BCa does not include zero.

200
SocialForces

96(1)

where k is the number of covariates in the model). Nevertheless, the interaction
terms are significant and properly signed, though the significance level of the
wage coordination interaction drops (p < 0.10). In toto, the analyses reported in table 6 suggest that our results are robust to sample composition (models 1–12) and to the measurement income inequality (models 4–12).

Additional Concerns
The null hypothesis on our interaction terms is that the effect of Southern im-
ports does not vary across observed levels of the moderators. We can also test
the null hypothesis that the coefficients on Southern imports and interaction
terms are jointly equal to zero. Such a test amounts to testing the null hypothesis
that the effect of Southern imports is equal to zero at any level of the moderator.
These (inherently non-directional chi square) tests were significant at the 0.05
level or greater in all models reported above except models 4 and 6 in table 4
(also see figure 2 below).10

While we have focused upon the way GPN consolidation, wage coordination,
and welfare states moderate the distributional effects of Southern imports, the
interaction terms have a symmetric interpretation. Auxiliary analyses suggest
that Southern imports have a positive effect on the slope of GPN consolidation
(see note 8 and table A2 in the online appendix). Global organizational pro-
cesses matter more for inequality when the national economy is more deeply
articulated with them. Conversely, Southern imports have a negative effect on
the slope of wage coordination and welfare state generosity. These results
appear in figures A3–A5 in the online appendix. We explicate the theoretical im-
plications of the latter two in the concluding section.

Finally, we examine whether our results are robust to the rise of single-headed
households. We calculated the (weighted) percent of single-headed households
using the Luxembourg Income Study for as many of our cases as possible (73). The

Figure 2. Marginal effect of Southern imports across GPN consolidation, wage coordination,
and welfare state generosity

−
.2

−
.1

5
−

.1
−

.0
5

0
.0

5
.1

.1
5

.2
.2

5
.3

.3
5

M
ar

gi
na

l E
ffe

ct
o

f S
P

E
N

o
n

G
in

36 56 76 96 116

GPN Consolidation

−
.2
−
.1
5
−
.1
−
.0
5
0
.0
5
.1
.1
5
.2
.2
5
.3
.3
5
.4
M
ar
gi
na
l E
ffe
ct
o
f S
P
E
N
o
n
G
in
i
−
.2
−
.1
5
−
.1
−
.0
5
0
.0
5
.1
.1
5
.2
.2
5
.3
.3
5
.4
M
ar
gi
na
l E
ffe
ct
o
f S
P
E
N
o
n
G
in
i

1 2 3 4 5

Wage Coordination

17.9 25.075 32.25 39.425 46.6

Welfare State Generosity

Note: The y-axes display the marginal effects obtained from models 6–8 of table 2. x-axes display
the observed range of each moderator. Upper and lower lines are 95 percent confidence intervals.

Production Globalization and Income Inequality 201

http://sf.oxfordjournals.org/lookup/suppl/doi:10.1093/sf/sox041/-/DC1

results were largely consistent with those in table 6. The only difference was in the
interaction involving wage coordination, which was correctly signed but non-
significant. However, models on this reduced sample excluding single-headed
households produced the same result. In combination with the non-significance of
single-headed households, this result likely owes to sampling rather than household
composition, but does suggest a degree of caution (also see model 11, table 6).11

Substantive Significance
Our analysis supports our argument that the distributional consequences of
Southern imports vary by global (GPN consolidation) and national (wage coordi-
nation and welfare state generosity) context. However, it sheds little light on the
importance of this variation for the overall effect of Southern imports, or for trends
in inequality more generally. To address the first question, we examine how much
the effect of Southern imports varies across each condition. Figure 2 displays the
marginal effect of Southern imports as it varies across the observed range of our
moderating covariates.12 These marginal effects are obtained from the coefficients
reported in models 2–4 of table 3. Zero is denoted with the solid horizontal line.

The left pane displays the marginal effect of Southern imports as it varies
across GPN consolidation. The effect increases by just over 650 percent from
the minimum to maximum observed value of GPN consolidation, is null at the
lowest levels of GPN consolidation, and becomes significantly positive as GPNs
become more consolidated. The middle pane displays the marginal effects of
Southern imports across wage coordination. The effect decreases by 104.3
across the full range, is significantly positive among countries with a range of
wage coordination from 1 to ~3, and is null thereafter. The right pane displays
the marginal effect of Southern imports as it varies across welfare state generos-
ity. Here, the effect declines by 137.9 percent from the minimum and maximum,
is positive and significant at low levels of generosity, but becomes null at moder-
ate levels of generosity. The increase in inequality per unit increase of Southern
imports varies considerably across periods of greater/lesser GPN consolidation,
and across countries with different labor market and welfare state institutions.

To understand the importance of these moderating effects for overall trends
in inequality, we ask how inequality would have changed, on average, if
Southern imports took place in a world characterized by varying global and
national contexts. That is, how would inequality have changed if the overall
effect of Southern imports equaled that observed at minimum, mean, and maxi-
mum values of GPN consolidation, wage coordination, and welfare state gener-
osity? To proceed, we begin with the following equation:13

θ= + + + −
+ + +
+ + + −
+ + +
+ − + − + −

GINI SPEN UNEMP UD
IEMP FLFP ELDP
FIRE AGEMP DUAL
ED NRPI LCUM

T T T

12.170 .076 .035
.101 .054 .270
.199 5.563 .134
.009 .236 .005

.197 .013 .003

t t t

t t t
t t t
t t t

s s s80 90 00

202 Social Forces 96(1)

This equation is identical to that in model 1 of table 3, except that we aver-
aged across the country-specific intercepts and the coefficient on Southern im-
ports (θ) is allowed to vary. We then estimate nine counterfactual models by
manipulating θ to equal its marginal effect at the minimum, mean, and maxi-
mum value of each of our three moderators (see Alderson [1999]).14

Figure 3 reports the results of these counterfactual equations. The dashed line
in the middle of the three graphs is the observed trend. The dotted lines refer to
the counterfactual equation when the coefficient on the focal moderator equals
its minimum throughout the period. The solid lines are the estimated counterfac-
tual trends when the coefficient on the focal moderator equaled its mean
throughout, and the dash-dot-dot-dash line is the counterfactual trend estimated
when the focal coefficient on the moderator equaled its maximum throughout.
To compare the magnitude of these counterfactual trends across moderators, the
y-axis (predicted Gini) is fixed across the three graphs.

On average, income inequality increased by 5.22 percent among the countries in
our sample, and Gini reached a level of 28.94 in the most recent year examined. If
there were fewer GPNs worldwide such that the mean ratio of trade to value added
were equal to the minimum observed, Southern imports would have produced a
2.12 percent increase in inequality, and a Gini score nearly two points lower than
observed. Conversely, if GPNs had consolidated earlier such that the ratio of world
trade to world value added equaled its maximum throughout the period, Southern
imports would have increased the Gini by 12.24 percent. The level of inequality
would have been nearly two points higher than observed in the most recent period.

Both wage coordination and welfare state generosity paint the opposite picture.
Southern imports would have produced a 9.45 percent increase in inequality if the
prevailing degree of wage coordination equaled the minimum observed, and a
level of inequality about a point higher in 2007. If wage coordination were equal

Figure 3. Counterfactual trends in income inequality
Welfare State Generosity

26
28

30
32

26
28
30
32
26
28
30
32

1970 1980 1990 2000 2010 1970 1980 1990 2000 2010 1970 1980 1990 2000 2010

GPN Consolidation

Observed Min Mean Max

Wage Coordination

Note: Observed is the observed inequality trend. Min, Mean, and Max are the trends that
would have been observed if Southern imports occurred in a world characterized by the
minimum, mean, and maximum observed level of GPN consolidation, wage coordination, and
welfare state generosity.

Production Globalization and Income Inequality 203

to the maximum throughout, Southern imports would have increased inequality
by 3.45 percent, and inequality would be a point and a half lower than observed
in 2007. Welfare state generosity has the biggest counterfactual variance. If the
prevailing level of welfare state generosity were equal to the minimum observed,
Southern imports would have increased inequality by 13 percent, and produced a
level of inequality over two points higher than observed in 2007. Contrarily,
Southern imports would have slightly reduced inequality (−0.001 percent) if the
maximum observed welfare state generosity were the norm, and observed levels
of inequality would be almost three points lower than observed in 2007.

Discussion and Conclusion
We argue that the effects of production globalization on income inequality
vary by global and national context. At the global level, the consolidation of
GPNs amplifies the distributional consequences of Southern imports. As
inter-firm linkages intensify across Northern and Southern countries, both an
increasingly low-wage labor pool and an increasing array of economic activ-
ity become integrated into GPNs. This intensifies the downward pressure of
Southern imports on low-skill wages and labor bargaining power. At the
national level, wage coordination and welfare state generosity mitigate the
distributional effect of Southern imports. Wage coordination decouples
changes in skill-specific labor demand from changes in wages, provides an
institutional source of labor bargaining power, and encourages worker soli-
darity, the latter two of which benefit low-skill workers disproportionately.
Welfare states reduce the post-transfer income gap between low- and high-
skill workers, and improve the bargaining position of labor as a whole.
Thus, GPN consolidation intensifies the link from Southern imports to the
skill-wage premium and labor share of income, while wage coordination and
welfare state generosity weaken these links.

Our analysis provides a compelling explanation for the inconsistent effects of
Southern imports. First, Southern imports did not have a significantly positive
effect on inequality until the ratio of global trade to global value added sur-
passed 64.52 percent, and this did not occur until 1995. It is not surprising,
then, that early research (or research using older data) finds small or inconsistent
effects for Southern imports, while more recent research suggests larger effects
(e.g., Bernanke 2007; Elsby, Hobijn and Şahin 2013; Spence and Hlatshwayo
2011). Second, Southern imports increase inequality only when wage coordina-
tion occurs at or below the industry level and is not patterned across different
industries (i.e., is less than 4 on the five-point scale), and when welfare state gen-
erosity is less than 33.89. But, less than half the country-years analyzed here
have wage coordination scores less than 4. Only 46 percent have welfare state
generosity scores less than 33.89.15 It is not surprising, then, that analysts typi-
cally find a greater role for production globalization when studying liberal coun-
tries like the United States than when they conduct comparative work including
countries with more active labor market policies and larger welfare states (Elsby,
Hobijn, and Şahin 2013; Lin and Tomaskovic-Devey 2013; Massey 2009;

204 Social Forces 96(1)

Spence and Hlatshwayo 2011; c.f. Gustafsson and Johansson 1999; Lee, Kim,
and Shim 2011; Mahler 2004). In sum, the distributional effects of production
globalization appear inconsistent because they depend on organizational and
institutional processes that vary across time and space.

Our findings further illuminate recent sociological explanations for the
inequality upswing in rich democracies. Lee, Kim, and Shim (2011) find that a
growing productivity gap between the public and private sector, driven in part
by the differential exposure of the public and private sector to international com-
petition, undermines the equalizing effect of public sector employment. Theories
on the causes of global production network formation contend that leading
firms build these networks to solidify their own competitive positions within an
industry (Bair 2009; Ponte and Gibbon 2005). Because these strategic considera-
tions inform decisions about which phases of a production processes to retain
“in-house,” globalized production networks concentrate highly productive
value-adding activities within the developed countries where leading firms are
located (Mahutga 2012, 2014b). Thus, at least some of the productivity gap that
dampens the egalitarian effect of public sector employment is related to the
boost to private sector productivity provided by GPNs in manufacturing.

Our findings also move the sociological literature on inequality beyond debates
about the relative importance of domestic and global factors to an understanding of
how they work together to produce distinct inequality trajectories across time and
space. For example, recent scholarship implies that the impact of wage coordination
on inequality should be on the decline either because these institutional arrange-
ments are retrenching, or, where core segments of the labor force preserve wage
coordination, labor market dualism occurs (Rueda 2007; Thelen 2012). As a point
of departure, we find that wage coordination matters for the distributional impact of
a global driver of inequality (Southern imports) in spite of the well-documented
dynamics in this scholarship (also see Oskarsson [2009]). Thus, we introduce a new
mechanism by which wage coordination can reduce inequality. Nevertheless, it is
possible that the moderating effect of wage coordination might be smaller in coun-
tries where dualization interrupts traditional class-based political projects underlying
wage coordination (Palier and Thelen 2010). Such an outcome appears inconsistent
with our results at first glance, however, because wage coordination has an increas-
ingly large negative effect on inequality as Southern imports increase (see above).

Similarly, our findings add to our understanding of the mechanisms by which
welfare states are “the single most important determinant for reducing inequal-
ity across advanced industrial democracies” (Lee, Kim, and Shim 2011, p. 118).
Welfare state generosity has the second largest moderating effect on Southern
imports, and produces the most egalitarian counterfactual scenario, where
inequality would have declined in response to production globalization if the
prevailing degree of welfare state generosity were closer to the maximum
observed. While this finding might seem counterintuitive, it isn’t: transfers asso-
ciated with the maximum level of generosity more than offset the effects of
Southern imports. Indeed, as we noted above, welfare state generosity has an
increasingly large negative effect on inequality as Southern imports increase.
Thus, welfare states both limit the magnitude with which global social change

Production Globalization and Income Inequality 205

can lead to distributional change, and become more important domestic deter-
minants of the distribution of income as globalization proceeds.

Finally, our findings have implications for the future of distributional politics in
the global North. On one hand, welfare states appear to be the most plausible
way to actively mitigate the distributional consequences of production globaliza-
tion in the future. Firms will do what firms will do. Wage coordination developed
over long and protracted periods that are somewhat unique to particular national
contexts, and many doubt their long-term viability. Increasing the size and scope
of welfare states across countries may be the most viable and efficacious way to
redistribute the gains from production globalization in the years to come.

On the other, the countervailing effects of GPN consolidation, wage coordi-
nation, and welfare state generosity also raise important questions regarding the
longer-term viability of egalitarian institutions in the global North. The argu-
ment that globalization pressures states to undermine corporatist patterns of
labor relations and adopt austerity measures is frequently made, but this intui-
tion is controversial. Some find retrenchment in welfare states and corporatist
labor relations since the 1980s (Allan and Scruggs 2004; Huber and Stephens
2001; Thelen 2012), others find expansion (Kenworthy 2007; Kenworthy and
Pontusson 2005), and still others find little systematic effect in any direction
(Brady, Beckfield, and Seeleib-Kaiser 2005). Theories linking globalization to re-
trenching egalitarian institutions have perhaps underspecified the mechanisms
by treating globalization as a static causal category. Instead, the ability of trans-
national actors to impact the regulatory and institutional behavior of nation-
states must depend upon the extent that these actors can themselves transcend
the confines of the authority structures they wish to change. The consolidation
of GPNs is one example of just such a dynamic process: as GPNs become modal
organizational forms over time, the reliance of Northern capital on Northern
labor declines, which in turn undermines postwar class compromises in the
North. If the inconsistent effects of economic globalization on egalitarian institu-
tions are explicable by such a dynamic relationship, then the conditional effects
we identify above may understate the total effect of production globalization on
inequality, which is a key question for future research (e.g., Kollmeyer 2009b).

Notes
1. To be clear, we use the term “global production network” generically to encompass

literatures on global commodity chains (GCC), value chains (GVC), and production
networks (e.g., Gereffi, Humphrey, and Sturgeon 2005; Yeung and Coe 2015).

2. The GPN/GVC/GCC literature explicates multiple modes of network “governance,”
understood as a characteristic of the inter-firm ties within a particular production
network. The ratio of trade to value added is a strategic measure of GPN consolida-
tion because it captures offshoring as carried out in all of these modes, some of which
include a high degree of trade in intermediate components, and others of which
involve multiple exports of relatively finished products (Mahutga 2012).

3. Theoretical expectations consistent with our argument that labor market institutions
should condition the effects of economic globalization have been formulated

206 Social Forces 96(1)

elsewhere (Kenworthy 2007). To our knowledge, none have directly tested this prop-
osition (c.f. Oskarsson 2009).

4. Previous research operationalizes production globalization with both outflows of foreign
direction investment (FDI) and imports from Southern countries. We restrict our analy-
sis to Southern imports because (a) the replacement of Northern with Southern labor
does not motivate the vast majority of FDI (Alderson and Nielsen 1999), (b) Southern
imports capture labor-saving FDI, and (c) production networks are increasingly orga-
nized via non-equity inter-firm relations rather than FDI (Milberg 2004). Unreported
analyses show that our results are robust to the inclusion of FDI (and other common in-
dicators of globalization), which does not interact with our three conditional processes.

5. Disproportionality will depend on temporal variation in the effect of Southern im-
ports on GDP. First, denote SI with X and GDP with Y. Each will have an observed
growth rate equal to

= =dx
dt

a
dy
dt

bx and y,

where t is the number of years. These have a well-known solution, which is the
compound growth rate of X eat0 andY ebt0 , respectively. Ywill also have a growth
rate attributable to X, which we can write as βX eat0 . The ratio is thus equal to

β
=

+
r

X e
Y e X e

.t
at

bt at
0

0 0

To see that the ratio depends both on the relative growth rate of Y and X and on β,
we can divide by the numerator

β
=

+( − )
r

Y X e
1

/
.t b a t

0 0

If β is constant, then r changes only with the relative growth in Y andX (i.e., b − a)
as t increases. If β increases with t, then, holding a and b constant, r decreases with
t. If β decreases with t, then, holding a and b constant, r increases with t.

Kollmeyer’s (2009a) argument assumes a wage gap between the North and
South, and GPN/GVC theorists contend that GPN consolidation widens this gap
over time (see above). In a basic growth model controlling for human capital and
the initial level of GDP, we observe a significantly positive β across all t, and a sig-
nificant increase in β of about eight-tenths of 1 percent per year. This result is avail-
able upon request.
6. Southern imports will add to total imports automatically, but not disproportionately.

Denoting Southern imports with X and other imports with Y, the ratio is

=
+

r
X e

Y e X e
.t

at
bt at
0
0 0

Dividing by the numerator,

= ( − )r Y X e
1

/
.t b a t
0 0

r changes only with the relative growth in Y and X (i.e., b − a) as t increases.

Production Globalization and Income Inequality 207

7. See figure 2 below for observed variation in the effect of Southern imports across the
full range of observed values for each moderator.

8. A related temporal concern is the association between our measure of GPN consoli-
dation and time (see figure 1). We re-estimated models 3–6 from table 3 and included
a linear time trend. These results are substantively identical except in one case: the
interaction involving Southern imports/GDP and GPN consolidation dropped in sig-
nificance (p < 0.10) (see table A2 in the online appendix). We thank an anonymous Social Forces reviewer for raising this issue.

9. The distributional effects of Southern imports could be larger in countries more
economically dependent on Southern imports. This raises the possibility that the
level of trade “may confound the relationship between [Southern imports/total im-
ports] and income inequality” (Beckfield 2006, note 7). Thus, we estimated addi-
tional versions of the models in table 3 that control for Southern imports/GDP and
the sum of imports and exports/GDP. In each case, the results were substantively,
and almost numerically, identical (see table A3 in the online appendix). Moreover,
the effect of Southern imports/total imports may vary with the level of Southern im-
ports/GDP. We do observe a positive and significant interaction between these two
covariates net of controls (b = 0.047; p < 0.001). Southern imports/total imports have a positive effect over the full range of Southern imports/GDP, but the effect of Southern imports/GDP is negative over nearly half the range of the former (see figure A1 in the online appendix). This may provide another explanation for the varied findings in the literature, and we thank an anonymous reviewer for raising these points.

10. In both cases, auxiliary analyses suggest that the effect of Southern imports/GDP var-
ies more steeply across GPN consolidation and welfare state generosity than the
effect of Southern imports/total imports. The lack of evidence against the joint null
hypothesis thus owes to uncertainty about the point estimate for Southern imports/
GDP at any level of these moderators. See figure A2 in the online appendix.

11. We thank anonymous Social Forces reviewers for bringing these additional concerns
to our attention.

12. GPN consolidation varies from 36.52 to 116.36, wage coordination varies from 1 to
5, and welfare state generosity varies from 17.89 to 46.6.

13. SPEN is Southern Imports, UNEMP is Unemployment, UD is Union Density, IEMP
is Industrial Employment, FLFP is Female Labor Force Participation, ELDP is
Elderly Population, FIRE is FIRE Sector Employment, AGEMP is Agricultural Sector
Employment, DUAL is Sector Dualism, ED is Secondary Education Enrollment,
NRPI is the Natural Rate of Population Increase, and LCUM is the Cumulative
Share of Left Cabinet Seats.

14. The marginal effect of Southern imports at the minimum, mean, and maximum
observed value of each moderator is as follows. GPN consolidation: −0.044, 0.059*,
0.246***; wage coordination: 0.164***, 0.068*, −0.007; welfare state generosity:
0.269***, 0.076, −0.102*. * p < 0.05; ** p < 0.01; *** p < 0.001.

15. To put these thresholds into perspective, countries that typically receive wage
coordination scores below the above limit are Canada, France, Luxembourg,
New Zealand, the UK, and the United States. Those who typically receive wel-
fare state generosity scores falling below the above limit are Australia, Canada,
Italy, Japan, New Zealand, the UK, and the United States. The liberal countries
of Canada, New Zealand, the UK, and the United States are uniquely low on
both dimensions.

208 Social Forces 96(1)

http://sf.oxfordjournals.org/lookup/suppl/doi:10.1093/sf/sox041/-/DC1

Appendix
Table A1. Correlations and Descriptive Statistics

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

1 Gini

2 Southern importsa 0.400

3 GPN consolidationa 0.097 0.575

4 Wage coordinationa −0.521 −0.225 0.081

5 Welfare state
generositya

−0.694 −0.304 0.178 0.494

6 Unemployment 0.301 −0.133 −0.083 −0.190 −0.217

7 Union density −0.611 −0.434 −0.155 0.370 0.415 −0.019

8 Industrial employment −0.055 −0.341 −0.503 0.

209

−0.192 −0.144 0.050

9 Female labor force
participation

0.177 0.109 0.043 −0.331 −0.155 −0.101 0.001 0.040

10 Elderly population −0.173 0.219 0.510 0.188 0.473 −0.090 0.174 −0.043 0.207

11 FIRE sector
employment

0.312 0.592 0.608 −0.265 −0.015 −0.016 −0.550 −0.527 0.335 0.250

12 Agricultural sector
employmentb

−0.059 −0.441 −0.447 0.229 −0.137 0.080 0.197 0.363 −0.510 −0.436 −0.801

13 Sector dualism 0.093 −0.166 −0.103 0.227 −0.220 −0.079 0.001 0.434 −0.313 −0.178 −0.550 0.732

14 Secondary education −0.308 0.171 0.498 0.230 0.422 0.091 0.227 −0.603 −0.093 0.216 0.367 −0.335 −0.337

15 Natural rate of pop
increase

0.292 0.009 −0.317 −0.330 −0.442 0.034 −0.259 −0.195 −0.213 −0.825 −0.117 0.335 0.084 −0.137

16 Cumulative left cabinet
share

−0.479 −0.150 0.331 0.188 0.528 −0.161 0.571 −0.190 0.174 0.508 0.009 −0.239 −0.283 0.419 −0.455

mean 27.5 0 0 0 0 7.29 43.7 26.4 5.5 14.3 12.4 0.667 1.84 102 0.312

SD 4.22 5.46 22.4 1.39 6.93 3.72 22.4 4.25 37.3 2.35 3.89 0.241 1.64 15.9 0.274

aSample mean deviated.
bBase-10 logarithm.

Production
Globalization
and
Incom
e
Inequality
209

About the Authors
Matthew C. Mahutga is Associate Professor of Sociology at the University of
California, Riverside. His research examines the global determinants of eco-
nomic organization and their consequences for a range of political and socio-
economic outcomes. His work appears in interdisciplinary outlets, including
Europe-Asia Studies, Global Networks, Review of International Political
Economy, Social Forces, Social Networks, Social Problems, Social Science
Research, Urban Studies, and elsewhere, and has been supported by the
National Science Foundation.

Anthony Roberts is an Assistant Professor of Sociology at California State
University, Los Angeles. His research interests include financialization, global
production, income inequality, industrial relations, and comparative capitalism.
His most recent work appears in Socio-Economic Review, Sociology of
Development, and the International Journal of Comparative Sociology. A cur-
rent project examines financialization and wage inequality in OECD and trans-
ition countries.

Ronald Kwon is a PhD candidate at the University of California, Riverside.
His research interests include immigration and political economy. Recent articles
appear in Population Studies, the Korea Journal, and the International Journal
of Comparative Sociology.

Supplementary Material
Supplementary material is available at Social Forces online, http://sf.
oxfordjournals.org/.

References
Alderson, Arthur S. 1999. “Explaining Deindustrialization: Globalization, Failure or Success?” American

Sociological Review 64(5):701–21.
Alderson, Arthur S., and Francois Nielsen. 1999. “Income Inequality, Development, and Dependence: A

Reconsideration.” American Sociological Review 64(4):606–31.
———. 2002. “Globalization and the Great U Turn: Income Inequality Trends in 16 OECD Countries.”

American Journal of Sociology 107(5):1244–99.
Allan, James, and Lyle Scruggs. 2004. “Political Partisanship and Welfare State Reform in Advanced

Industrial Societies.” American Journal of Political Science 48(3):496–512.
Anderson, Christopher, and Jonas Pontusson. 2007. “Workers, Worries and Welfare States: Social

Protection and Job Insecurity in 15 OECD Countries.” European Journal of Political Research 46(2):
211–35.

Anner, Mark, Jennifer Bair, and Jeremy Blasi. 2013. “Towards Joint Liability in Global Supply Chains:
Addressing the Root Causes of Labor Violations in International Subcontracting Networks.”
Comparative Labor Law and Policy Journal 35(1):1–43.

Bair, Jennifer Lynn, ed. 2009. Frontiers of Commodity Chain Research. Palo Alto, CA: Stanford University
Press.

Beckfield, Jason. 2006. “European Integration and Income Inequality.” American Sociological Review 71
(6):964–85.

210 Social Forces 96(1)

http://sf.oxfordjournals.org/

Bernanke, Ben, Chairman of the Federal Reserve. 2007. Speech before the Greater Omaha Chamber of
Commerce, Omaha, NE, February 6.

Bradley, David, Evelyne Huber, Stephanie Moller, Francois Nielsen, and John Stephens. 2003.
“Distribution and Redistribution in Postindustrial Democracies.” World Politics 55(1):193–228.

Brady, David, Jason Beckfield, and Martin Seeleib-Kaiser. 2005. “Economic Globalization and the Welfare
State in Affluent Democracies,1975–2001.” American Sociological Review 70(6):921–48.

Clark, Rob. 2013. “Convergence in National Income Distributions.” Social Forces 92(2):413–36.
Congressional Budget Office (CBO). 2011. “Trends in the Distribution of Household Income Between 1979

and 2007.” Congress of the United States Congressional Budget Office, Washington, DC.
Elsby, Michael W. L., Bart Hobijn, and Ayşegül Şahin. 2013. “The Decline of the US Labor Share.”

Brookings Papers on Economic Activity, no. 2 (Fall): 1–63.
Epsing-Anderson, Gosta. 1990. Three Worlds of Welfare Capitalism. Princeton, NJ: Princeton University

Press.
Feenstra, Robert C. 1998. “Integration of Trade and Disintegration of Production in the Global Economy.”

Journal of Economic Perspectives 12(4):31–50.
Fernandez, Raul, and John Glazer. 1991. “Striking for a Bargain Between Two Completely Informed

Agents.” American Economic Review 81(1):240–52.
Friedrich, Robert J. 1982. “In Defense of Multiplicative Terms in Multiple Regression Equations.”

American Journal of Political Science 26(4):797–833.
Gereffi, Gary, John Humphrey, and Timothy Sturgeon. 2005. “The Governance of Global Value Chains.”

Review of International Political Economy 12(1):78–104.
Gustafsson, Bjorn, and Mats Johansson. 1999. “In Search of Smoking Guns: What Makes Income

Inequality Vary over Time in Different Countries?” American Sociological Review 64(4):
585–605.

Hall, Peter A., and David Soskice, eds. 2001. Varieties of Capitalism: The Institutional Foundations of
Comparative Advantage. Oxford: Oxford University Press.

Huber, Evelyne, and John D. Stephens. 2001. Development and Crisis of the Welfare States: Parties and
Politics in Global Markets. Chicago: University of Chicago Press.

———. 2014. “Income Inequality and Redistribution in Post-Industrial Democracies: Demographic,
Economic and Political Determinants.” Socio-Economic Review 12(2):245–67.

Huber, Evelyne, Charles Ragin, John D. Stephens, David Brady, and Jason Beckfield. 1997, 2004, 2014.
Comparative Welfare States Data Set.

Kenworthy, Lane. 2001. “Wage-Setting Measures: A Survey and Assessment.” World Politics 54(1):
57–98.

———. 2007. “Inequality and Sociology.” American Behavioral Scientist 50(5):584–602.
Kenworthy, Lane, and Jonas Pontusson. 2005. “Rising Inequality and the Politics of Redistribution in

Affluent Countries.” Perspectives on Politics 3(3):449–71.
Kollmeyer, Christopher. 2009a. “Consequences of North-South Trade for Affluent Countries: A New

Application of Unequal Exchange Theory.” Review of International Political Economy 16(5):803–26.
———. 2009b. “Explaining Deindustrialization: How Affluence, Productivity Growth, and Globalization

Diminish Manufacturing Employment.” American Journal of Sociology 114(6):1644–74.
Krugman, Paul. 1995. “Growing World Trade: Causes and Consequences.” Brookings Papers on Economic

Activity, vol. 1.
Kus, Basak. 2013. “Financialisation and Income Inequality in OECD Nations: 1995–2007.” Economic and

Social Review 43(4):477–95.
Lee, Cheol-Sung, Young-Bum Kim, and Jae-Mahn Shim. 2011. “The Limit of Equality Projects: Public-

Sector Expansion, Sectoral Conflicts, and Income Inequality in Postindustrial Economies.” American
Sociological Review 76(1):100–124.

Production Globalization and Income Inequality 211

Lin, Ken-Hou, and Donald Tomaskovic-Devey. 2013. “Financialization and U.S. Income Inequality,
1970–2008.” American Journal of Sociology 118(5):1284–1329.

Mahler, Vincent A. 2004. “Economic Globalization, Domestic Politics, and Income Inequality in the
Developed Countries: A Cross-National Study.” Comparative Political Studies 37(9):1025–1153.

Mahutga, Matthew C. 2012. “When Do Value Chains Go Global? A Theory of the Spatialization of Value-
Chain Linkages.” Global Networks 12(1):1–21.

———. 2014a. “Global Models of Networked Organization, the Positional Power of Nations and
Economic Development.” Review of International Political Economy 21(1):157–94.

______. 2014b. “Production Networks and the Organization of the Global Manufacturing Economy.”
Sociological Perspectives 57(2):229–55.

Mahutga, Matthew C., and Andrew K. Jorgenson. 2016. “Production Networks and Varieties of
Institutional Change: The Inequality Upswing in Post-Socialism Revisited.” Social Forces 94(4):
1711–41.

Massey, Douglas S. 2009. “Globalization and Inequality: Explaining American Exceptionalism.” European
Sociological Review 25(1):9–23.

Milberg, William. 2004. “The Changing Structure of International Trade Linked to Global Production
Systems: What Are the Policy Implications?” Working Paper No. 33, Policy Integration Department,
World Commission on the Social Dimension of Globalization. Geneva: International Labour Office.

Milberg, William, and Deborah Winkler. 2009. “Globalization, Offshoring and Economic Insecurity in
Industrialized Countries.” DESA Working Paper No. 87. New York: United Nations.

Mughan, Anthony. 2007. “Economic Insecurity and Welfare Preferences: A Micro-Level Analysis.”
Comparative Politics 39(3):293–310.

OECD. 2011a. Historical Statistics. Paris.
———. 2011b. Employment Database. Paris.
______. 2011c. International Trade by Commodity Statistics Database. Paris.
Oskarsson, Sven. 2009. “Divergent Trends and Different Causal Logics: The Importance of Bargaining

Centralization When Explaining Earnings Inequality Across Advanced Democratic Societies.” Politics
and Society 33:359.

Palier, Bruno, and Kathleen Thelen. 2010. “Institutionalizing Dualism: Complementarities and Change in
France and Germany.” Politics and Society 38(1):119–48.

Ponte, Stefano, and Peter Gibbon. 2005. Trading Down: Africa, Value Chains and the Global Economy.
Philadelphia: Temple University Press.

Riedl, Maximilian. 2013. “Wage Bargaining, Job Loss Fears and Offshoring.” Center for European
Governance and Economic Development Research, Discussion Paper No. 174.

Rubin, Rose M., Shelley I. White-Means, and Luojia Mao Daniel. 2000. “Income Distribution of Older
Americans.” Monthly Labor Review (November):19–30.

Rueda, David. 2007. Social Democracy Inside Out: Government Partisanship, Insiders and Outsiders in
Industrialized Democracies. Oxford: Oxford University Press.

Scheve, Kenneth, and Matthew J. Slaughter. 2004. “Economic Insecurity and the Globalization of
Production.” American Journal of Political Science 48(4):662–74.

Scheve, Kenneth, and David Stasavage. 2009. “Institutions, Partisanship and Inequality in the Long Run.”
World Politics 61(2):215–53.

Schrank, Andrew. 2004. “Ready-to-Wear Development? Foreign Investment, Technology Transfer, and
Learning-by-Watching in the Apparel Trade.” Social Forces 83(1):123–56.

Scruggs, Lyle, Detlef Jahn, and Kati Kuitto. 2014. “Comparative Welfare Entitlements Dataset 2. Version
2014-03.” University of Connecticut and University of Greifswald.

Solt, Fredrick. 2009. “Standardizing the World Income Inequality Database.” Social Science Quarterly 90
(2):231–42.

212 Social Forces 96(1)

Spence, Michael, and Sandile Hlatshwayo. 2011. “The Evolving Structure of the American Economy and
the Employment Challenge.” Working Paper, Council on Foreign Relations, Center for Geoeconomic
Studies.

Thelen, Kathleen. 2012. “Varieties of Capitalism: Trajectories of Liberalization and the New Politics of
Social Solidarity.” Annual Review of Political Science 15(1):137–59.

Tomaskovic-Devey, Donald, and Ken-Hou Lin. 2011. “Income Dynamics, Economic Rents, and the
Financialization of the U.S. Economy.” American Sociological Review 76(4):538–59.

Traxler, Franz. 1999. “The State in Industrial Relations: A Cross-National Analysis of Developments in
Socioeconomic Effects.” European Journal of Political Research 36(1):55–85.

UNIDO. 2015. INDSTAT2 Vienna: United Nations.
United Nations. 2014a. Indicators of Sustainable Development: Guidelines and Methodologies, 3rd ed.

New York: United Nations.
———. 2014b. COMTRADE. New York: United Nations.
Visser, Gelle. 2011. “Data Base on Institutional Characteristics of Trade Unions, Wage Setting, State

Intervention and Social Pacts, 1060–2010.” Amsterdam Institute for Advanced Labour Studies:
University of Amsterdam.

Wallerstein, Michael. 1999. “Wage-Setting Institutions and Pay Inequality in Advanced Industrial
Societies.” American Journal of Political Science 43(3):649–80.

Western, Bruce. 1997. Between Class and Market: Postwar Unionization in the Capitalist Democracies.
Princeton, NJ: Princeton University Press.

Wood, Adrian. 1994. North-South Trade, Employment and Inequality: Changing Fortunes in a Skill-Driven
World. Oxford: Oxford University Press.

Wooldridge, Jeffrey M. 2002. Econometric Analysis of Cross Section and Panel Data. Cambridge, MA:
MIT Press.

World Bank. 2016. World Development Indicators. Washington, DC.
Wright, Erik Olin. 2000. “Working-Class Power, Capitalist-Class Interests, and Class Compromise.”

American Journal of Sociology 105(4):957–1002.
Yeung, Henry Wai-chung, and Neil M. Coe. 2015. “Toward a Dynamic Theory of Global Production

Networks.” Economic Geography 91(1):29–58.

Production Globalization and Income Inequality 213

Copyright of Social Forces is the property of Oxford University Press / USA and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright
holder’s express written permission. However, users may print, download, or email articles for
individual use.

Turn in your highest-quality paper
Get a qualified writer to help you with

“ Week-1 ”

Get high-quality paper

NEW! AI matching with writer

Hire a Writer

Client Reviews

4.9

Sitejabber

4.6

Trustpilot

4.8

Our Guarantees

100% Confidentiality

Information about customers is confidential and never disclosed to third parties.

Original Writing

We complete all papers from scratch. You can get a plagiarism report.

Timely Delivery

No missed deadlines – 97% of assignments are completed in time.

Money Back

If you're confident that a writer didn't follow your order details, ask for a refund.

New to Your Trusted Assignment Help Service? Sign up & Save

Calculate the price of your order

Type of paper needed:

Pages:

You will get a personal manager and a discount.

Academic level:

We'll send you the first draft for approval by at

Total price:

$0.00

Power up Your Academic Success with the
Team of Professionals. We’ve Got Your Back.

Power up Your Study Success with Experts We’ve Got Your Back.

Order Now Order Now