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

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

* Author for correspondence,
Email: [email protected]

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 their
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 …

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: [email protected]
‡Department of Economics – Bocconi University, Via Sarfatti, 25, I-20136 Milan, Italy.

Email: [email protected]
§Urban and Regional Research Centre Utrecht (URU), Faculty of Geosciences, Utrecht University, Heidelberglaan 2,

NL-3508 TC Utrecht, the Netherlands. Email: [email protected]

(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 …

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

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 that
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:[email protected]
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

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: [email protected]

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 (1994) 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

1970 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 Gini
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 …

Why us

Why wait? place your order now