Global Value Chains and Employment - Recherche en IA spécialisée au MIT

Global Value Chains and Employment
Growth in Asia
Neil Foster-Mcgregor∗

This paper considers the sources of employment demand in Asian economies.
Using data from the World–Input Output Database, I examine the relative
importance of domestic and foreign demand in generating employment. Malgré
some degree of heterogeneity across the sample, domestic demand is found to
be the major driver of employment in all cases. Plus loin, the relative importance
of ﬁnal and intermediate exports in generating employment varies by economy,
with some economies relying on intermediate exports to generate employment
to a greater extent than others, reﬂecting their importance as suppliers of
intermediate inputs in global value chains, while others rely to a greater extent
on ﬁnal exports, reﬂecting their role as assemblers within global value chains.
Considering developments over time, I ﬁnd that employment is driven by two
offsetting factors: (je) ﬁnal demand (either domestic or foreign) et (ii) labor
productivity, with changes in interindustry structure also being important in the
case of intermediate exports.

Mots clés: crise, decomposition, employment, global value chains
Codes JEL: F14, F62, O19

je. Introduction

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

International

trade and the globalization of supply chains have been
important drivers for the growth and development of a number of economies in
Asia and beyond, most notably the People’s Republic of China (RPC). Global value
chains (GVCs) are often considered a relatively easy route to industrialization
(Baldwin 2016). By enabling economies to contribute to certain stages of
production, they avoid the problems associated with developing whole industries.
En outre, GVCs are considered important channels for technology spillovers
and technology upgrading. Despite concerns about potential negative consequences
for workers (ILO 2011), there is also an expectation that GVC participation can
promote employment in developing economies. With some notable exceptions
(par exemple., Los, Timmer, and de Vries 2015; Portella-Carbo 2016), cependant, there is a

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

∗Neil Foster-Mcgregor: Deputy Director, UNU-MERIT, Maastricht, Netherlands. E-mail: foster@merit.unu.edu. Le
author would like to thank the managing editor and the journal referees for helpful comments and suggestions.
ADB recognizes “China” as the People’s Republic of China and “Korea” as the Republic of Korea. The usual ADB
disclaimer applies.

Revue du développement en Asie, vol. 36, Non. 2, pp. 100–130
https://doi.org/10.1162/adev_a_00133

© 2019 Asian Development Bank and
Asian Development Bank Institute.
Publié sous Creative Commons
Attribution 3.0 International (CC PAR 3.0) Licence.

Global Value Chains and Employment Growth in Asia 101

paucity of evidence on the impact of trade generally, and GVC participation in
particular, on employment, both in absolute terms and relative to domestic sources
of employment.

This issue of the relationship between trade and employment generation is
even more relevant in the context of the recent global ﬁnancial crisis that has
been associated with weak and volatile demand in world markets. En effet, there is
evidence that levels of global trade and GVC participation have plateaued since the
ﬁnancial crisis and in recent years have actually declined (Timmer et al. 2016). UN
number of potential explanations for these recent developments in trade and GVC
performance have been proposed: weak demand in response to the crisis; the lack
of progress in multilateral and bilateral trade policy negotiations—in particular
with regard to standards, règlements, and rules of origin, among others—that
have limited further fragmentation; the possibility of overfragmentation that is now
being corrected; an increase in the extent of reshoring; and the possibility that
the beneﬁts of major technological developments, particularly in information and
communication technology, have been exhausted.

En particulier,

In this paper, I am interested in identifying recent (2000–2014) developments
in employment in a sample of six Asian economies and decomposing these
developments along a number of dimensions.
I decompose
employment developments into effects due to domestic and foreign demand,
where this latter effect is further split into effects due to foreign ﬁnal demand
(ﬁnal exports) and foreign intermediate demand (intermediate exports). I use a
structural decomposition analysis to decompose employment growth in the sample
of economies into effects due to labor productivity growth, ﬁnal demand growth,
changes in interindustry structure, and changes in the ratio of value-added growth.1
Enfin, I identify differences in employment growth and the drivers of employment
growth between the precrisis and postcrisis periods.

The analysis is linked to a number of different strands of literature. C'est
linked, Par exemple, to the small existing literature on measuring the employment
effect of exporting.2 A number of these papers consider this relationship in the
context of the PRC, with Feenstra and Hong (2010) using input–output analysis
to estimate the employment effects of exporting. They ﬁnd that export growth
contributed around a third of overall employment growth during 1997–2005, avec
most growth coming from nontraded goods like construction. While exports grew

1This study includes ex post analysis and does not identify the cause of changes in labor productivity or ﬁnal

demand, which are exogenously given.

2The analysis is also linked to the literature estimating econometrically the relationship between offshoring
and employment (par exemple., Hijzen and Swaim 2010; Foster-McGregor, Poeschl, and Stehrer 2016). Dans ces études,
offshoring is considered to have two offsetting effects: (je) a substitution effect that leads to the destruction of jobs,
et (ii) a productivity effect that increases overall output and impacts positively upon employment. The approach
adopted in these studies is to identify the overall effect of offshoring on employment, with a positive (negative) effect
being interpreted as implying that the productivity effect is stronger (weaker) than the substitution effect. In the
decomposition in section IV.B, I split up the effects of labor productivity and ﬁnal demand, isolating the effects of
each of these variables on employment demand.

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

102 Revue du développement en Asie

much faster during 2000–2005, developments in domestic demand continued to
dominate. Similar results were found by Chen et al. (2012), who show that the
impact on employment of processing exports was lower than that of nonprocessing
exportations, which in turn was signiﬁcantly lower than the effect due to domestic
demand. Los, Timmer, and de Vries (2015) adopt a similar approach to identify
the direct and indirect employment effects of exports in the PRC. Their results
indicate that between 1995 et 2001, rapid growth in foreign demand was offset by
large increases in labor productivity, with the result being that foreign demand had
essentially no effect on net employment. During 2001–2006, cependant, a strong rise
in foreign demand resulted in about 70 million jobs being added. In the most recent
period for which data were available (2006–2009), domestic demand became more
important than foreign demand, which the authors argue may signal a rebalancing
of the global economy.

Other studies conduct similar analyses for different economies. Feenstra and
Sasahara (2017), Par exemple, estimate the impact on employment in the United
États (NOUS) from exports and imports during 1995–2011. They ﬁnd that growth
in US exports led to increased demand for workers that generated around 6.6
million new jobs. Imports into the US from the PRC reduced labor demand by
à propos 2 million jobs, which still resulted in an overall net gain for the US during
the review period in terms of employment demand from importing and exporting.
Kiyota (2014) considers the impact of exports on employment in the cases of the
RPC, Indonésie, Japan, and the Republic of Korea, with the results indicating a
positive and increasing effect of exports on employment in Japan, la RPC, et le
Republic of Korea. The study further shows the important role of machinery-related
industries in generating employment through exports, with indirect employment
effects through vertical interindustry linkages also playing an important role.

In a recent paper, Portella-Carbo (2016) extends this kind of analysis to
consider the super-multiplier effects of international trade. In his model, trade
impacts employment in a manner similar to the studies mentioned above, mais
has additional impacts by stimulating household consumption, private business
investissement, and the production of intermediate goods. In short, autonomous
demand impacts employment through the Keynesian multiplier and the accelerator
mechanism. His results indicate that the effects of trade on employment vary
greatly across economies, with the effects in upswings being particularly large in
economies such as the PRC and Germany, but smaller in other economies such as
France and the United Kingdom, which rely to a greater extent on domestic demand.
The current paper is also linked to recent work on the effects of the crisis
on trade patterns. Timmer et al. (2016) consider the decline in world trade in the
postcrisis period, arguing that there are two competing explanations for the changed
circonstances. The ﬁrst is that the composition of demand has changed, in particular
from durable investment and consumer goods toward services that are less trade
intensive (Bems, Johnson, and Yi 2011, 2013; and Bussière et al. 2013). The second

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

Global Value Chains and Employment Growth in Asia 103

relates to the possible decline of GVC activity that may have occurred for a number
of reasons already mentioned above (Evenett and Fritz 2015; Kee and Tang 2016;
Harms, Lorz, and Urban 2012). Timmer et al. (2016) ﬁnd that both channels have
played a role in the stalled development of international trade. My analysis of the
precrisis and postcrisis periods in this study’s dataset extends the analysis of Timmer
et autres. (2016) to consider the employment implications of changes in international
trade activity.

The starting point for the approach that I adopt in this paper is the analysis
of Los, Timmer, and de Vries (2015), who consider the impact of foreign demand
on employment creation in the PRC. The approach builds upon recent applications
of input–output tables to issues of international trade (Timmer et al. 2013, 2014;
Foster-McGregor and Stehrer 2013; Koopman, Wang, and Wei 2014; Wang et al.
2017) by estimating the consequences of developments in international trade for
domestic employment. The analysis is based upon the recently released update to
the World Input–Output Database (WIOD) (Timmer et al. 2015), which has data
depuis 2000 à 2014. I concentrate on the six Asian economies included in the WIOD:
India; Indonésie; Japan; la RPC; la République de Corée; and Taipei,Chine.

The main interest of the analysis is to identify to what extent the crisis has
impacted labor demand and whether there has been a shift away from foreign
demand toward domestic demand. En même temps, there are a number of
additional hypotheses that are implicit in the analysis. En particulier, I expect that
the smaller economies in the sample are reliant to a larger extent on exports for
generating labor demand than larger economies, which in turn are likely to have a
more diversiﬁed production structure. En tant que tel, the smaller economies in the sample
may be affected to a greater extent in the postcrisis period by the observed declines
in trade and GVC activity. En outre, theory indicates that changes in the terms of
trade between primary products and manufactures will affect natural resource-poor
economies in the opposite way as resource-rich economies such as Indonesia. Le
period 2000–2014 was one in which the terms of trade improved dramatically
for resource-rich economies such as Indonesia and deteriorated substantially for
the other ﬁve Asian economies in the sample. En tant que tel, I may expect that the
resultant real exchange rate changes would favor growth in the demand for
nontradables—and as a result the demand for domestic goods—in Indonesia, avec
the opposite being the case in the other ﬁve sample economies.

The results of the analysis indicate that domestic demand is the major driver
of employment in all six economies, albeit to a lesser extent in some economies.
The relative importance of ﬁnal and intermediate exports in generating employment
also varies by economy. Some economies such as Taipei,China rely on intermediate
exports to generate employment to a greater extent than others, likely reﬂecting their
importance as suppliers of intermediate inputs in GVCs. Other economies such as
the PRC rely to a greater extent on ﬁnal exports, reﬂecting their role as assemblers
within GVCs. Considering developments over time, I ﬁnd that employment is driven

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

104 Revue du développement en Asie

by two offsetting factors: (je) ﬁnal demand (either domestic or foreign) et (ii) labor
productivity. In the case of intermediate exports, changes in interindustry structure
have also played a role in many economies.

The remainder of the paper is set out as follows. Section II describes the
methodology used in the analysis. Section III brieﬂy discusses the data. Section IV
presents the results. Section V concludes.

II. Méthodologie

The initial approach to identify the employment effects of international trade
and GVCs follows closely that adopted by Los, Timmer, and de Vries (2015), lequel
in turn was based upon the contribution of Johnson and Noguera (2012). Ici,
the approach of Los, Timmer, and de Vries (2015) is described in detail before
discussing the extension of this approach.

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

I begin by assuming that there are N economies, S industries in each
economy, and F production factors in each economy–industry. Industry output
in a particular economy is determined using domestic production factors, in this
case capital and labor, and intermediate inputs, which may be sourced either
domestically or from foreign sources. The output produced in each industry can
be used as either ﬁnal demand or as intermediate inputs in the production of
other goods. Demand for ﬁnal goods is assumed to come from three sources: (je)
households, (ii) government, et (iii) ﬁrms.3 When considering shipments of ﬁnal
goods and intermediates, both within and across economies, there needs to be a
distinction between the source and destination economy–industry. Following Los,
Timmer, and de Vries (2015), I use i to denote the source economy, j the destination
economy, s the source industry, and t the destination industry. It is assumed that
markets clear and the additional assumption of a single price irrespective of a
product’s use is imposed. By deﬁnition, when markets clear, the product-market-
clearing condition can be written as

(cid:2)

yi(s) =

fi j(s) +

mi j(s, t )

(1)

Ici, yi(s) is the value of output in industry s in economy i, fi j(s) is the value
of goods sold by this industry for ﬁnal use in economy j, and mi j(s, t ) is the value
of products sold by this industry for intermediate use by industry t in economy j.

Los, Timmer, and de Vries (2015) go on to express the market-clearing
conditions for each of the SN industries using matrix algebra. Pour faire ça, let y

3In the WIOD 2016 release, ﬁnal demand is split into ﬁve sources: (je) ﬁnal consumption expenditure by
households, (ii) ﬁnal consumption expenditure by nonproﬁt organizations serving households, (iii) ﬁnal consumption
expenditure by government, (iv) gross ﬁxed capital formation, et (v) changes in inventories and valuables.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

Global Value Chains and Employment Growth in Asia 105

be the output vector of dimension (SN × 1), the elements of which represent output
levels in each economy–industry. A global input–output matrix A of dimension
(SN × SN ) is also deﬁned. The matrix has elements ai j(s, t ) = mi j(s, t )/y j(t ),
which capture the ratio of intermediate inputs per unit of output and are termed
the technical coefﬁcients. Los, Timmer, and de Vries (2015) describe these terms
as giving the cost shares of output from industry s in economy i used by industry t
in economy j. The matrix A can be written as

⎡

A ≡

⎢
⎢
⎢
⎢
⎣

A11 A12
A21 A22
…
…
AN1 AN2

⎤

⎥
⎥
⎥
⎥
⎦

· · · A1N
· · · A2N
…
. . .
· · · ANN

with Ai j being an S × S matrix with typical element ai j(s, t ). Given this setup, it
should be clear that the submatrices on the main diagonal contain the cost shares of
domestically produced intermediate inputs, while those on the off-diagonal contain
the cost shares of foreign intermediate inputs. The matrix A thus summarizes the
input requirements of all intermediate goods across industries and economies.

Using the matrix A, equation (1) can be expressed in matrix form as

⎡

⎢
⎢
⎢
⎣

≡

⎤

⎥
⎥
⎥
⎦

⎡

⎢
⎢
⎢
⎣

y1
y2
…
yN

A11 A12
A21 A22
…
…
AN1 AN2

· · · A1N
· · · A2N
…
. . .
· · · ANN

⎤

⎡

⎥
⎥
⎥
⎦

⎢
⎢
⎢
⎣

⎤

⎥
⎥
⎥
⎦

⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎢
⎣

y1
y2
…
yN

⎡

(cid:9)

⎤

⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎥
⎦

f1 j

j
(cid:9)

(cid:9)

f2 j
…
fN j

with yi being the S-vector with production levels in economy i, and fi j being the
S-vector of ﬁnal demand in economy j for the products of economy i. This can be
written as

y = Ay + F

which can further be expressed as

y = (I − A)−1f = Bf

(2)

(3)

with I being an SN × SN identity matrix and B = (I − A)−1 being the well-known
Leontief inverse (Leontief 1936) that captures the gross output values in all stages
of production that are generated in the production process of one unit of ﬁnal
output.

To complete their model, Los, Timmer, and de Vries (2015) assume that the
quantity of output in an industry is a function of the quantities of the labor, capital,

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

106 Revue du développement en Asie

and intermediate inputs used in production. The value of output in an industry is
then equal to the value of all inputs used. Denoting the value of output in industry s
of economy i as yi(s), and letting li(s) denote the number of workers in this industry,
I can then deﬁne pi(s) as the number of workers required per US dollar of output in
industry s in economy i:

pi(s) = li(s)/yi(s)

which can be written as an SN × 1 column vector, p. As Los, Timmer, and de Vries
(2015) point out, the elements in this vector are economy and industry speciﬁc.
The requirement vector p indicates the labor per US dollar of output needed in
one particular production stage. To consider the labor required in all the stages of
production of a ﬁnal product, a new vector, k, is deﬁned. This can be constructed
for any ﬁnal demand vector by postmultiplying the labor requirement vector by the
gross outputs needed for production of the ﬁnal good, c'est:

k = ˆp(I − A)−1f = ˆpBf

(4)

with a hat indicating a diagonal matrix with the elements of a vector on the
diagonal. For a particular economy–industry, k represents the labor required in each
economy–industry in the world to meet this demand, both in the economy–industry
of interest and also in upstream industries—both domestically and abroad—
delivering intermediates to produce the ﬁnal output of the economy–industry of
interest.

Équation (4) is the main equation of interest in capturing the importance of
ﬁnal demand for employment generation. I extend this approach in two ways. Firstly,
I use the recent approach of Wang et al. (2017) to further decompose the sources
of employment generation. Secondly, I use a structural decomposition approach to
identify the sources of recent developments in employment generation.

The approach of Wang et al. (2017) decomposes production activities into
three terms: (je) production for domestic demand, (ii) ﬁnal goods exports, et (iii)
intermediate exports (what they term GVC activities and which may also involve
the reimportation of intermediates at later stages of the production process). Le
approach of Wang et al. (2017) begins by rewriting equation (2) comme

y = Ay + f = ADy + f D + AF y + f F = ADy + f D + E

(5)

where AD =

⎢
⎢
⎢
⎣

⎡

⎤

· · ·
A11
0
0
0 A22 · · ·
0
…
…
…
. . .
· · · Agg
0
0

⎥
⎥
⎦ is a GN × GN diagonal block matrix of
⎥

domestic input coefﬁcients; AF = A − AD is a GN × GN off-diagonal block matrix
of imported input coefﬁcients; f D = [f11f 22 · · · f gg](cid:4) is a GN × 1 vector of ﬁnal

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

Global Value Chains and Employment Growth in Asia 107

production for domestic consumption; f F = f − f D is a GN × 1 vector of ﬁnal
h(cid:5)=m Egr](cid:4) is a GN × 1 vector
G
product exports; and E = [
of gross exports.

h(cid:5)=2 E2h · · ·

G
h(cid:5)=1 E2h

(cid:9)

Rearranging equation (5) gives

y = (I − AD)−1f D + (I − f D)−1E = Lf D + LE = Lf D + Lf F + LAF y

where L = (I − AD)−1 is the local Leontief inverse. Premultiplying by the
(diagonal) labor requirements vector and replacing y = (I − A)−1f = Bf gives

ˆpBf = ˆpLf D + ˆpLf F + ˆpLAF Bf

(6)

This equation decomposes the labor used in each economy–industry into the

following:

(je) Employment generated domestically for domestic ﬁnal demand ( ˆpLf D)

(ii) Employment generated domestically for ﬁnal production exports ( ˆpLf F )
(c'est à dire., employment generated for production that crosses borders for ﬁnal
consumption only)

(iii) Employment generated through the export of intermediate goods ( ˆpLAF Bf)

Équation (6) indicates that labor induced through these different channels of
demand depends upon three sets of factors: (je) changes in labor requirements ( ˆp);
(ii) changes in the intermediate input structure (L, B, and AF ); et (iii) changes
in the level of ﬁnal demand (F, f D, and f F ). For ease of exposition, I can rewrite
equation (6) comme

k = kD + kE + kI

(7)

where the superscripts D, E, and I refer to domestic ﬁnal demand, export ﬁnal
demand, and the export of intermediates, respectivement.

The description above is based closely upon that provided by Los, Timmer,
and de Vries (2015) and has been extended to introduce the approach of Wang
et autres. (2017). Within this framework, I now use the time series that I have available
to decompose the growth in employment due to ﬁnal demand. Pour faire ça, I begin
by noting that 1 plus the growth rate of employment between two time periods
(0 et 1) (or the ratio of employment due to ﬁnal demand) can be expressed as
follows:
1 + ˙k = k1
k0

= ˆp1B1 f 1
ˆp0B0 f 0

(8)

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

108 Revue du développement en Asie

where similar expressions for the three subterms of equation (7) can also be written
down.

Using standard decomposition methods, it is possible to decompose this
expression into an effect due to changes in (je) employment to gross output
(c'est à dire., employment requirements); (ii) interindustry structure, both domestically and
internationally); et (iii) foreign ﬁnal demand. The resulting decomposition is as
follows:
ˆp1B1 f 1
ˆp0B0 f 0

= ˆp1B1 f 1
ˆp0B1 f 1

× ˆp0B0 f 1
ˆp0B0 f 0

× ˆp0B1 f 1
ˆp0B0 f 1

(9)

This expression can be further decomposed to capture the role of labor
productivity changes in driving employment changes. To see this, the standard
deﬁnition of labor productivity (c'est à dire., value added per worker) can be written as

l pi(s) = vi(s)/li(s)

with l pi(s) being labor productivity in economy i in industry s, and vi(s) being value
added in economy i in industry s. This can further be expressed as

vi(s)/li(s) = vi(s)/yi(s) × yi(s)/li(s)

which can be rearranged to give

li(s)/yi(s) = li(s)/vi(s) × vi(s)/yi(s)

and expressed in matrix form as

ˆp = ˆq ˆw

(10)

This states that the ratio of employment to gross output is equal to the
inverse of labor productivity multiplied by the ratio of value added to gross output.
Substituting equation (10) into equation (8) gives

1 + ˙k =

ˆq1 ˆw1B1 f 1
ˆq0 ˆw0B0 f 0

which can be decomposed as

ˆq1 ˆw1B1 f 1
ˆq0 ˆw0B0 f 0

ˆq1 ˆw1B1 f 1
ˆq0 ˆw1B1 f 1

ˆq0 ˆw1B1 f 1
ˆq0 ˆw0B1 f 1

ˆq0 ˆw0B1 f 1
ˆq0 ˆw0B0 f 1

ˆq0 ˆw0B0 f 1
ˆq0 ˆw0B0 f 0

The ﬁrst

term captures the impact of changing labor productivity on
employment, the second captures the impact of changing value added to gross
output ratios, the third captures the impact of changes in technical coefﬁcients (c'est à dire.,
intermediate use), and the ﬁnal term captures changes in foreign ﬁnal demand. Ce
is the ﬁnal decomposition of overall employment demand. Note, cependant, that the

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

Global Value Chains and Employment Growth in Asia 109

approximate growth rate of employment (due to foreign demand) between two time
periods can be written as

(cid:10)

˙k ≈ ln (1 + ˙k) = ln

(cid:11)

(cid:10)

ˆq1 ˆw1B1 f 1
ˆq0 ˆw1B1 f 1
(cid:10)

+ ln
(cid:11)

+ ln

ˆq0 ˆw0B0 f 1
ˆq0 ˆw0B0 f 0

ˆq0 ˆw1B1 f 1
ˆq0 ˆw0B1 f 1

(cid:11)

(cid:10)

+ ln

(cid:11)

ˆq0 ˆw0B1 f 1
ˆq0 ˆw0B0 f 1

(11)

which allows the growth rate of employment (due to ﬁnal demand) to be expressed
as the sum of the four logged decomposition terms:

(je)

an effect due to changes in labor productivity (q);

(ii)

an effect due to changes in the ratio of value added to gross output (w), lequel
is often considered to be driven by upgrading within GVCs;

(iii)

an effect due to changes in the interindustry structure (B), including the
role of a changing trade structure in intermediate goods, that may be due to
such things as factor substitution, technological change, et (GVC) sourcing
motifs; et

(iv)

an effect due to changes in ﬁnal demand.

The studies of Feenstra and Hong (2010) and Chen et al. (2012) mentioned
above argued that the employment impact of (foreign) demand growth crucially
depend on the levels and growth rates of the labor productivity of the various
activités. En particulier, they argue that the employment effect of foreign demand
is mainly a race between increases in productivity levels and demand levels, avec
higher productivity reducing the induced demand for labor (holding foreign demand
constant). The decomposition deﬁned above allows this hypothesis to be addressed
in a more formal way.

A similar decomposition can be written for the individual terms of equation

(7) as follows:

(cid:10)

˙k

D ≈ ln

˙k

E ≈ ln

ˆq1 ˆw1L1 f D
1
ˆq0 ˆw1L1 f D
1

(cid:11)

(cid:10)

+ ln

ˆq0 ˆw1L1 f D
1
ˆq0 ˆw0L1 f D
1

(cid:11)

(cid:10)

+ ln

ˆq0 ˆw0L1 f D
1
ˆq0 ˆw0L0 f D
1

(cid:11)

(cid:10)

+ ln

ˆq1 ˆw1L1 f F
1
ˆq0 ˆw1L1 f F
1

(cid:11)

(cid:10)

+ ln

ˆq0 ˆw1L1 f F
1
ˆq0 ˆw0L1 f F
1

(cid:11)

(cid:10)

+ ln

ˆq0 ˆw0L1 f F
1
ˆq0 ˆw0L0 f F
1

(cid:11)

(cid:10)

+ ln

(cid:11)

ˆq0 ˆw0L0 f D
1
ˆq0 ˆw0L0 f D
0
(12)
(cid:11)

ˆq0 ˆw0L0 f F
1
ˆq0 ˆw0L0 f F
1
(13)

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

110 Revue du développement en Asie

(cid:10)

˙k

I ≈ ln

(cid:11)

(cid:10)

(cid:11)

(cid:10)

(cid:11)

ˆq1 ˆw1L1AF
ˆq0 ˆw1L1AF
(cid:10)

1 B1 f 1
1 B1 f 1

+ ln
(cid:11)

ˆq0 ˆw1L1AF
ˆq0 ˆw0L1AF
(cid:10)

1 B1 f 1
1 B1 f 1

+ ln
(cid:11)

ˆq0 ˆw0L1AF
ˆq0 ˆw0L0AF
(cid:10)

1 B1 f 1
1 B1 f 1

ˆq0 ˆw0L0AF
ˆq0 ˆw0L0AF

1 B1 f 1
0 B1 f 1

+ ln

ˆq0 ˆw0L0AF
ˆq0 ˆw0L0AF

0 B1 f 1
0 B0 f 1

+ ln

ˆq0 ˆw0L0AF
ˆq0 ˆw0L0AF

0 B0 f 1
0 B0 f 0

(cid:11)

+ ln

(14)

Note that equation (14) includes additional terms capturing the effects of
AF (the imported input coefﬁcients) et B (the full Leontief matrix). While I
do calculate the full decomposition in this case, the results combine the three
terms capturing the interindustry structure (L, AF , et B), which allows for a ready
comparison with the results from the other decompositions. En outre, foreign
ﬁnal demand can be further decomposed by the source of that demand. In the
analysis that follows, I decompose foreign ﬁnal demand across four regions—Asia,
the Americas, Europe, and the rest of the world—to examine the regional sources
of employment growth in Asia.

III. Data and Descriptive Statistics

The analysis is conducted using the 2016 version of the WIOD, which builds
upon and extends the 2013 version of the dataset.4 WIOD constructs a world
input–output table for 43 économies (plus the rest of the world) for the period
2000–2014 and includes data on 56 sectors, mainly at the 2-digit ISIC Revision
4 level (see Timmer et al. [2015, 2016] for details on construction and coverage).
In addition to world input–output tables, WIOD also reports a set of socioeconomic
accounts that include information on employment levels (in terms of both hours
worked and persons engaged).

The WIOD therefore has all the relevant information needed to conduct
the analysis that follows. I make use of (je) the international input–output tables
(reporting the values of intermediate ﬂows between all industries and economies),
(ii) value added by economy and industry, (iii) gross output by economy and
industry, (iv) domestic and foreign ﬁnal demand by economy and industry, et (v)
employment by economy and industry.5 In this analysis, I am forced to measure
employment in thousands of workers rather than hours worked. This is largely
because data on hours worked are not available for the PRC.

4The major drawback of the 2016 version compared with the 2013 version is the lack of information on
employment by skill type, which thus does not allow for an analysis of the changing role of demand on employment
composition.

5The sources for data on employment, which can be found in Gouma et al. (2018), are varied and do not

always coincide with ofﬁcial national data.

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

Global Value Chains and Employment Growth in Asia 111

Tableau 1. Employment Levels (Million), 2000

Employment Due To:

Employment Consumption

Domestique

Final
Exports

Intermediate
Exports

Économie

India

Indonésie

Japan

RPC

Republic of Korea

Taipei,Chine

410.1

96.9

65.3

719.6

18.2

16.9

372.9
91%
76.3
79%
59.5
91%
620.5
86%
14.6
80%
12.6
74%

19.8
5%
8.6
9%
2.6
4%
57.8
8%
1.7
9%
1.8
11%

17.5
4%
12.0
12%
3.1
5%
41.3
6%
2.0
11%
2.6
16%

PRC = People’s Republic of China.
Source: Author’s calculations based on Timmer et al. (2015).

The time series of world input–output tables on which the empirical analysis
is based are expressed in current US dollars. Given the interest in developments
au fil du temps, this creates an issue when variables are expressed as ratios of quantities
(par exemple., number of employees) to values (par exemple., the value of gross output). To deal
with this, I follow the approach of Los, Timmer, and de Vries (2015) and deﬂate
all nominal values using the gross domestic product deﬂator from the World
Development Indicators database.

IV. Results

UN.

Employment Developments in Asia

Dans ce

subsection,

initial descriptive results, describing
developments in employment generation in Asian economies over the period
2000–2014 and the sources of demand for this employment.

I present

Tableau 1 presents the initial level of employment in 2000 for the six sample
économies, along with the decomposition of employment given by equation (5)
into employment due to domestic demand and the two sources of foreign demand
(ﬁnal exports and intermediate exports). The table shows that employment in
2000 was around 720 million and 410 million in the two economies with the
largest populations, the PRC and India, respectivement. The two smallest economies in
population terms, Taipei,China and the Republic of Korea, had employment levels
dans 2000 of around 17 million and 18 million, respectivement. Domestic consumption
accounted for the vast majority of employment in all of the sample economies,
being as high as 91% in India and Japan, with lower levels reported in the Republic

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

112 Revue du développement en Asie

Chiffre 1.1. Employment Developments in the People’s Republic of China

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

Source: Author’s calculations based on Timmer et al. (2015).

of Korea (80%); Indonésie (79%); and Taipei,Chine (74%).6 The fact that these
latter economies rely to a greater extent on trade for their labor demand may
reﬂect the fact that they are relatively small in terms of gross domestic product
and are therefore generally more open to trade. Among trade channels, ﬁnal
exports accounted for the greatest share of employment in the PRC and India,
possibly indicating the importance of assembly activities for the PRC in particular.
Intermediate exports accounted for the greatest share in the other four economies,
with the difference between employment due to intermediate and ﬁnal exports being
relatively large in Taipei,Chine.

Figures 1.1–1.6 report employment levels for each of the six economies over
the full review period (2000–2014), as well as the decomposition of employment
into three sources (domestic consumption, ﬁnal exports, intermediate exports).
Considering each economy in turn, Chiffre 1.1 shows that employment in the PRC
rose steadily during the review period, despite a small decline around the time of the
global ﬁnancial crisis, with a notable increase in employment during 2010–2011.
The role of domestic demand dropped somewhat during the early 2000s, mais
increased after the ﬁnancial crisis. This may be related to a general reorientation
of the PRC’s economy toward domestic demand. The declining share of domestic
demand in generating employment in the early 2000s was offset by a rising share of

6This study only accounts for the direct effect of the different trade channels on employment. The adopted
approach is not able to account for indirect effects on employment through increases in the incomes of workers and
ﬁrms that can raise domestic demand and demand for employment through increased household consumption and
ﬁrm investment.

Global Value Chains and Employment Growth in Asia 113

Chiffre 1.2. Employment Developments in Indonesia

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

Source: Author’s calculations based on Timmer et al. (2015).

both forms of exports, with demand due to ﬁnal exports rising to a larger extent
than demand due to intermediate exports. In the case of Indonesia, Chiffre 1.2
shows increasing employment, with a relatively rapid rise after 2011. Unlike
la RPC, there was no observed decline in employment around the time of the
ﬁnancial crisis. Aussi, in contrast to the PRC, there is a rising share of domestic
ﬁnal consumption in employment generation across the whole review period, avec
declining contributions of both ﬁnal exports and intermediate exports over time.
This provides some support to the view expressed above that developments in the
terms of trade may have led to increased demand for nontradables in relatively
resource-rich Indonesia. The results for India in Figure 1.3 also conﬁrm a positive
employment trend, with a noticeable increase after 2011. En général, there is little in
the way of changes in the composition of employment sources.

In the case of Japan, Chiffre 1.4 shows a different pattern of generally
declining employment, which was more pronounced after the global ﬁnancial crisis.
In terms of the sources of employment demand, there was a slight decline in the
role of domestic consumption over time, with small increases in the role of all trade
activités. Developments in the Republic of Korea, as shown in Figure 1.5, ﬁt with
the pattern found for all sample economies other than Japan, with a rising trend
in employment demand. The importance of domestic consumption for employment
generation tended to decline over time, with a rising share of employment demand
for the two forms of exports, particularly intermediate exports, observed. Le
results for Taipei,China in Figure 1.6 largely mimic those for the Republic of

114 Revue du développement en Asie

Chiffre 1.3. Employment Developments in India

Source: Author’s calculations based on Timmer et al. (2015).

Chiffre 1.4. Employment Developments in Japan

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

Source: Author’s calculations based on Timmer et al. (2015).

Korea, with a rising trend in employment demand and a declining role for domestic
consumption in generating this demand during the review period.

To emphasize changes in the contributions of the different sources of
employment demand over time, Chiffre 2 reports changes (in percentage points)

Global Value Chains and Employment Growth in Asia 115

Chiffre 1.5. Employment Developments in the Republic of Korea

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

Source: Author’s calculations based on Timmer et al. (2015).

Chiffre 1.6. Employment Developments in Taipei,Chine

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

Source: Author’s calculations based on Timmer et al. (2015).

in the different sources of employment demand for the six sample economies
entre 2000 et 2014. In most economies, the share of domestic consumption
in generating employment declined during the review period. The only exception to
this was Indonesia where the share of employment due to domestic consumption

116 Revue du développement en Asie

Chiffre 2. Changes in Contributions to Employment, 2000–2014

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

PRC = People’s Republic of China.
Source: Author’s calculations based on Timmer et al. (2015).

increased by nearly 7 percentage points, with both forms of international trade
seeing a declining share, most notably ﬁnal exports. In the other economies, un
declining share of domestic consumption was observed, with declines ranging from
a low of around 1.5 percentage points in India to a high of nearly 6 percentage
points in the Republic of Korea. In all of these cases, the contributions of both trade
channels tended to increase, with the exception of Taipei,China for ﬁnal exports,
where employment due to intermediate exports accounted for much of the increase.

Decomposition Results

The previous subsection described developments in employment demand
and the sources of this demand. This subsection reports results from decomposing
the different sources of employment demand as shown in equations (12), (13),
et (14).7 Chiffre 3 shows the average annual growth rate of employment during
2000–2014 due to domestic ﬁnal demand, along with the decomposition of this
growth. In most economies, the average growth rate of employment due to domestic
ﬁnal demand was low, and even negative in the case of Japan. The exceptions to
this were Indonesia and India, which reported average growth rates of employment
due to domestic ﬁnal demand of 4.2% et 3%, respectivement. In terms of the

7The decomposition as described above decomposes employment growth at the economy-sector level. Dans
order to aggregate up to the economy level, sectoral employment weights are used in equations (12), (13), et (14).

Global Value Chains and Employment Growth in Asia 117

Chiffre 3. Decomposition of Employment Growth Due to Domestic Final Demand,
2001–2014

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

PRC = People’s Republic of China.
Source: Author’s calculations based on Timmer et al. (2015).

decomposition of employment growth, I tended to ﬁnd a relatively large positive
effect of domestic ﬁnal demand and a relatively large negative effect of labor
productivity growth. In the PRC, la République de Corée, and Taipei,Chine, ces
two effects tended to offset each other, resulting in a muted effect of domestic ﬁnal
demand on employment growth. In Indonesia and India, the growth in domestic
ﬁnal demand was higher than the growth in labor productivity during the review
period, resulting in relatively strong employment growth. The ﬁnal case of Japan
is interesting, with essentially no growth in either domestic ﬁnal demand or labor
productivity during 2000–2014. The other thing to note from this ﬁgure is the almost
complete lack of a role for either the growth in the ratio of value added to gross
output or the interindustry trade structure.

Figures 4 et 5 report information on employment growth due to the two
forms of international trade, ﬁnal exports and intermediate exports, along with
the corresponding decomposition of these effects. Considering ﬁnal export demand
(Chiffre 4), relatively low growth rates of employment due to ﬁnal export demand
are again observed. The growth effects range from a low of –1.7% in Indonesia
to a high of 2.7% in India. Encore une fois, the overall effects are driven by the
(offsetting) growth effects of ﬁnal demand and labor productivity, with the effects
of changes in the interindustry structure and the growth rate of value added to gross

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

118 Revue du développement en Asie

Chiffre 4. Decomposition of Employment Growth Due to Final Export Demand, 2001–2014

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

PRC = People’s Republic of China.
Source: Author’s calculations based on Timmer et al. (2015).

Chiffre 5. Decomposition of Employment Growth Due to Intermediate Exports, 2001–2014

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

PRC = People’s Republic of China.
Source: Author’s calculations based on Timmer et al. (2015).

Global Value Chains and Employment Growth in Asia 119

output being generally small. Despite similar and muted growth effects, there are
important differences across the sample economies. In the case of the PRC, le
growth of foreign ﬁnal demand was relatively high (autour 14%), with relatively
high growth rates also being observed in India (8.9%) and the Republic of Korea
(7.6%). Labor productivity growth in these economies was also relatively high,
cependant, leading to a small overall growth effect. In other economies, most notably
Japan and Taipei,Chine, the growth of foreign ﬁnal demand was relatively low at
autour 2%.

Turning to the case of intermediate exports and their impact on employment
growth (Chiffre 5), somewhat similar results are observed.8 Firstly, employment
growth due to intermediate exports tends to be higher than the rates observed
for the case of foreign ﬁnal demand. Growth rates range from 1.6% (Indonésie)
à 4.2% (India). Secondly, ﬁnal demand is usually the main driver of employment
growth due to intermediate exports, with the growth effects of ﬁnal demand ranging
entre 6.5% et 8.1%. Thirdly, labor productivity growth has a strong (negative)
impact on employment growth due to intermediate exports across economies,
with the main exception being Japan, where labor productivity growth has been
muted, and to a lesser extent Taipei,Chine. Fourthly, and unlike the previous cases,
changes in the interindustry structure play an important role in driving employment
growth due to intermediate exports in a number of economies. The most important
example in this context is the PRC, where the effect of interindustry structure is
large and dominates the effect of ﬁnal demand. This suggests that in the PRC,
developments in sourcing patterns have strongly impacted employment growth due
to GVC participation. This may indicate the movement from a reliance on upstream
intermediate imports to an increasing reliance on domestic upstream intermediate
inputs in the PRC. Changes in interindustry structure also impacted positively upon
employment growth in India, Indonésie, and the Republic of Korea. Negative effects
were observed in the cases of Japan and Taipei,Chine, with changes in sourcing
patterns reducing employment growth in both of these economies.

Precrisis and Postcrisis Developments in Employment

The previous subsection showed that of the different forms of international
trade considered, the employment effects of intermediate exports tended to grow
more rapidly during 2000–2014. This subsection considers whether there were
differences in employment developments (and the corresponding decompositions)
between the precrisis and postcrisis periods. As already discussed, developments in
trade ﬂows and GVC participation tended to differ in the precrisis and postcrisis

8As discussed above, the three terms that capture the effects of changes in interindustry structure—broadly

deﬁned—are combined in order to provide a ready comparison with the other two decompositions.

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

120 Revue du développement en Asie

Chiffre 6. Changes in Contributions to Employment, 2000–2008

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

PRC = People’s Republic of China.
Source: Author’s calculations based on Timmer et al. (2015).

periods, with this subsection examining whether there have been similar effects on
employment.

Figures 6 et 7 report results that are similar to Figure 2 but for the two
subperiods, 2000–2008 and 2009–2014, respectively.9 Changes in the composition
of employment demand tended to be larger during 2000–2008 than 2009–2014.
During the former period, results indicate that employment demand due to
trade tended to increase relative to demand due to domestic
international
consommation. This was true for all cases except Indonesia. In most economies,
the rise in employment demand resulting from international trade was due largely
to intermediate exports, with the effect of ﬁnal exports being negative for India;
Indonésie; and Taipei,Chine. The major exception was the PRC, where demand due
to ﬁnal exports was dominant.
While smaller

the changing composition of
employment demand during 2009–2014 followed a similar pattern in the Republic
of Korea and Taipei,Chine, with an increasing contribution of intermediate exports
at the expense of domestic consumption. In the PRC, there was a change in the
relative roles of international trade and domestic consumption, with the role of

than during 2000–2008,

9Average growth rates are reported for the periods 2000–2008 and 2009–2014, implying that the growth rate
de 2009 includes information on employment levels in 2008. Using 2007 as the cutoff year rather than 2008 does not
alter the qualitative results signiﬁcantly.

Global Value Chains and Employment Growth in Asia 121

Chiffre 7. Changes in Contributions to Employment, 2009–2014

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

PRC = People’s Republic of China.
Source: Author’s calculations based on Timmer et al. (2015).

Tableau 2. Growth Rate of Employment Due to Trade Channels in the Precrisis and
Postcrisis Periods (%)

Domestic Consumption

Final Exports

Intermediate Exports

2001–2008

2009–2014

India
Indonésie
Japan
RPC
Republic of Korea
Taipei,Chine

1.34
2.89
−0.70
−0.20
1.57
1.05

6.20
7.18
−0.80
2.68
1.64
0.66

2001–2008
−0.17
−4.62
3.45
7.10
2.59
−1.16

2009–2014

2001–2008

2009–2014

10.28
5.23
−0.47
−3.54
3.81
1.09

5.78
1.77
4.95
7.00
5.23
5.08

5.89
3.88
−0.01
−0.10
4.84
2.33

PRC = People’s Republic of China.
Source: Author’s calculations based on Timmer et al. (2015).

domestic consumption increasing and the role of international trade, most notably
ﬁnal exports, falling. The pattern in Indonesia during 2009–2014 was similar to
that during 2000–2008, albeit with intermediate exports accounting for most of the
decline in the contribution of international trade. In India and Japan, the overall
changes and the changes in composition were relatively small.

Tableau 2 reports the average growth rates of employment for three different
demand sources during 2001–2008 and 2009–2014. The table reveals a number of
interesting outcomes. In a number of economies, the growth rate of employment
generation in the postcrisis period actually exceeded that in the precrisis period

122 Revue du développement en Asie

Chiffre 8.1. Employment Due to International Trade in the People’s Republic of China,
2000–2014

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

Source: Author’s calculations based on Timmer et al. (2015).

for all three sources of demand. This is true for Indonesia and India, et, except
in the case of intermediate exports, also for the Republic of Korea. For two of
these economies—Indonesia and India—employment growth due to ﬁnal exports
was actually negative in the precrisis period. In the cases of the PRC and Japan,
there is negative employment growth due to ﬁnal exports in the postcrisis period,
following relatively rapid growth in the precrisis period, with a small negative
growth rate of employment for Japan (−0.5%) and a larger negative effect for the
RPC (−3.5%). For these two economies, a similar pattern is also observed in the
case of intermediate exports, with a small negative growth rate of employment in the
postcrisis period after experiencing relatively rapid growth in the precrisis period.
For the PRC, this relatively poor performance in the postcrisis period is offset by
relatively rapid employment growth due to domestic consumption; while in the case
of Japan, employment demand due to domestic consumption was also sluggish.

Given the interest in the effect of the crisis on employment and the channels
of employment generation, Figures 8.1–8.6 report the employment generated by
international trade for each of the six economies during 2000–2014. Considering
each economy in turn, Chiffre 8.1 shows that in the PRC, employment due to
international trade dropped dramatically around the time of the crisis. Entre 2008
et 2009, employment due to foreign demand dropped 14.9%, following a drop of
à propos 6% entre 2007 et 2008. The decline was driven by changing demand
due to both ﬁnal and intermediate exports, with a somewhat larger percentage

Global Value Chains and Employment Growth in Asia 123

Chiffre 8.2. Employment Due to International Trade in Indonesia, 2000–2014

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

Source: Author’s calculations based on Timmer et al. (2015).

Chiffre 8.3. Employment Due to International Trade in India, 2000–2014

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

Source: Author’s calculations based on Timmer et al. (2015).

124 Revue du développement en Asie

Chiffre 8.4. Employment Due to International Trade in Japan, 2000–2014

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

Source: Author’s calculations based on Timmer et al. (2015).

Chiffre 8.5. Employment Due to International Trade in the Republic of Korea, 2000–2014

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

Source: Author’s calculations based on Timmer et al. (2015).

Global Value Chains and Employment Growth in Asia 125

Chiffre 8.6. Employment Due to International Trade in Taipei,Chine, 2000–2014

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

Source: Author’s calculations based on Timmer et al. (2015).

decline in employment demand due to intermediate exports. Cependant, the recovery
in employment following the crisis also tended to be more rapid in the case
of intermediate exports. Chiffre 8.2 shows that in Indonesia, there was a more
prolonged drop in employment due to ﬁnal exports, though at the time of the crisis,
the percentage decline in demand due to intermediate exports was much stronger
(though, as with the PRC, it also recovered more quickly). This was also the case
with India, as shown in Figure 8.3, which saw relatively large percentage declines in
employment demand due to ﬁnal exports in the lead-up to the crisis and a large drop
in demand due to intermediate exports at the time of the crisis during 2008–2009.
Chiffre 8.4 shows that the case of Japan is somewhat different, with the decline in
demand at the time of the crisis due to ﬁnal exports being signiﬁcantly larger than
that due to intermediate exports. The recovery in Japan was also slower, with levels
of employment due to ﬁnal exports and intermediate exports not having reached
their precrisis levels by 2014. The Republic of Korea also exhibited a pattern
dissimilar to some of the other economies. Chiffre 8.5 shows rising employment
demand due to ﬁnal exports continuing to grow throughout the crisis alongside a
relatively minor contraction in employment due to intermediate exports. Results
thus suggest that the Republic of Korea was largely shielded from the effects of the
crise, at least in terms of the employment impact of foreign demand. The negative
impacts of the crisis on employment demand were also relatively muted in the case
of Taipei,Chine. Chiffre 8.6 shows the declines in demand being fairly evenly split
between ﬁnal and intermediate exports.

126 Revue du développement en Asie

Chiffre 9. Decomposition of Employment Growth Due to Domestic Final Demand in the
Precrisis and Postcrisis Periods

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

PRC = People’s Republic of China.
Note: The precrisis period refers to 2001–2008, while the postcrisis period refers to 2009–2014.
Source: Author’s calculations based on Timmer et al. (2015).

The ﬁnal part of this subsection reports the decomposition results for the
precrisis and postcrisis periods separately. The decomposition results are reported
for domestic consumption in Figure 9, for ﬁnal exports in Figure 10, et pour
intermediate exports in Figure 11. In these ﬁgures, the left-hand side of each pair
of bars for each economy reports the decomposition for the precrisis period (2001–
2008) and the right-hand side bars are for the postcrisis period (2009–2014).

I begin with the decomposition of employment growth due to domestic ﬁnal
demand. Chiffre 9 reveals that employment growth due to domestic demand tended
to be higher in the postcrisis period than in the precrisis period. The difference in
growth rates between the two periods was particularly large in the case of Indonesia
(2.3% versus 6.7%) and India (1.1% versus 5.6%). Decomposition results indicate
largely offsetting effects of domestic demand growth and labor productivity growth,
with only small effects of either changes in interindustry structure or the ratio
of value added to gross output. The majority of cases show a declining role for
labor productivity growth in the postcrisis period, with the difference between the
precrisis and postcrisis periods being particularly large in the case of Indonesia and
India. The role of domestic demand growth in the postcrisis period also declined in
many of the sampled economies—and became negative in the case of Japan—with
a slight increase in the contribution of domestic demand growth observed only in
Taipei,Chine.

Global Value Chains and Employment Growth in Asia 127

Chiffre 10. Decomposition of Employment Growth Due to Final Exports in the Precrisis
and Postcrisis Periods

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

PRC = People’s Republic of China.
Note: The precrisis period refers to 2001–2008, while the postcrisis period refers to 2009–2014.
Source: Author’s calculations based on Timmer et al. (2015).

Turning to the trade channels, Figures 10 et 11 reveal that the (negative)
effect of labor productivity on employment growth diminished in most economies
(most notably Indonesia and India) in the postcrisis period relative to the precrisis
period, with declines in labor productivity increasing employment in the case of
Japan (implying a reduction in labor productivity). Le (positive) effect of ﬁnal
demand growth also tended to decline in the postcrisis period relative to the precrisis
period. This is particularly the case for employment growth due to intermediate
exportations (Chiffre 11), with large changes in the case of ﬁnal exports limited to the
PRC and Japan (Chiffre 10). In the case of intermediate exports, there was an
important role for changes in the interindustry structure in many economies, most
notably the PRC and Japan. Comparing the precrisis and postcrisis periods, changes
in interindustry structure tended to lead to a decline in employment growth in the
postcrisis period. This is true for all of the sample economies except the Republic
of Korea and Taipei,Chine.

V. Conclusion

There is evidence of declining engagement in GVCs since 2008 à la
global level, as well as for a number of Asian economies, most notably the PRC;

128 Revue du développement en Asie

Chiffre 11. Decomposition of Employment Growth Due to Intermediate Exports in the
Precrisis and Postcrisis Periods

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

PRC = People’s Republic of China.
Note: The precrisis period refers to 2001–2008, while the postcrisis period refers to 2009–2014.
Source: Author’s calculations based on Timmer et al. (2015).

la République de Corée; and Taipei,Chine. This paper considers the employment
implications of these developments. While it is difﬁcult to generalize across the
diverse set of Asian economies considered, the results suggest that despite declining
trade and GVC participation, growth in employment demand has been stable,
and in some cases has increased, depuis 2008. Domestic demand accounts for the
vast majority of employment demand, though effects differ by economy. Domestique
demand is found to be relatively less important in the Republic of Korea and
Taipei,Chine, possibly reﬂecting their relatively small size, and it has declined
in importance for a number of economies in the postcrisis period. The case of
the PRC is particularly interesting and suggests a reorientation toward domestic
sources of demand. Indonesia is also relying increasingly on domestic sources
of demand, possibly driven by global and regional terms of trade developments.
Such results may raise questions about the overall importance of GVC involvement
for employment, though the employment (and incomes) generated through foreign
ﬁnal demand may have relatively large multiplier effects, indirectly impacting
employment through domestic demand.

In terms of the role of trade in generating employment, results differ by
economy. Some economies rely on intermediate exports to generate employment to
a greater extent than others, reﬂecting their importance as suppliers of intermediate
inputs (par exemple., Taipei,Chine) or raw materials and primary products (par exemple., Indonésie).

Global Value Chains and Employment Growth in Asia 129

Others, most notably the PRC, generate more employment through ﬁnal exports,
reﬂecting their role as assemblers within GVCs. Developments in employment
tend to be driven by two offsetting factors: (je) ﬁnal demand (either domestic or
foreign) et (ii) labor productivity. The positive role of ﬁnal demand and changes in
interindustry structure have tended to more than offset the negative effects of labor
productivity growth. In the case of intermediate exports, changes in interindustry
structure have also played a role in many economies, suggesting that there has been
a reorientation in supply chains that have impacted employment positively in some
économies (par exemple., la RPC) and negatively in others (par exemple., Japan). Since the global
ﬁnancial crisis, developments in interindustry structure have in general tended to
impact negatively upon employment growth relative to the precrisis period.

The importance of foreign demand for employment growth has diminished
for most economies since 2008. This has been offset by weaker labor productivity
growth in the postcrisis period for most economies, which has minimized the effect
of weak foreign demand on employment growth. En tant que tel, and despite the weak
foreign demand growth, the declines in labor productivity growth have resulted in
an increased growth rate of employment due to foreign demand in economies such
as India and Taipei,Chine (in the case of ﬁnal demand).

Les références

Baldwin, Richard. 2016. The Great Convergence: Information Technology and the New

Globalization. Cambridge: Presse universitaire de Harvard.

Bems, Rudolfs, Robert C. Johnson, and Kei-Mu Yi. 2011. “Vertical Linkages and the Collapse
of Global Trade.” American Economic Review Papers and Proceedings 101 (3): 308–
12.

_____. 2013. “The Great Trade Collapse.” Annual Review of Economics 5 (1): 375–400.
Bussière, Matthieu, Giovanni Callegari, Fabio Ghironi, Giulia Sestieri, and Norihiko Yamano.
2013. “Estimating Trade Elasticities: Demand Composition and the Trade Collapse of
2008–09.” American Economic Journal: Macroeconomics 5 (3): 118–51.

Chen, Xikang, Leonard K. Cheng, K.C. Fung, Lawrence J. Lau, Yun-Wing Sung, K. Zhu, C. Lequel,
J.. Pei, and Y. Duan. 2012. “Domestic Value Added and Employment Generated by Chinese
Exports: A Quantitative Estimation.” China Economic Review 23 (4): 850–64.

Evenett, Simon J., and Johannes Fritz. 2015. The Tide Turns? Trade, Protectionism, and Slowing

Global Growth. Londres: CEPR Press.

Feenstra, Robert C., and Chang Hong. 2010. “China’s Exports and Employment.” In China’s
Growing Role in World Trade, edited by Robert C. Feenstra and Shang-Jin Wei, 167–202.
Chicago: The University of Chicago Press.

Feenstra, Robert C., and Akira Sasahara. 2017. “The ‘China Shock’, Exports and US
Employment: A Global Input–Output Analysis.” NBER Working Paper No. 24022.
Foster-McGregor, Neil, Johannes Poeschl, and Robert Stehrer. 2016. “Offshoring and the

Elasticity of Labour Demand.” Open Economies Review 27 (3): 515–40.

Foster-McGregor, Neil, and Robert Stehrer. 2013. “Value Added Content of Trade: UN

Comprehensive Approach.” Economics Letters 120 (2): 354–57.

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3

130 Revue du développement en Asie

Gouma, Reitze, Wen Chen, Pieter Woltjer, and Marcel P. Timmer. 2018. World Input–Output
Database Socio-Economic Accounts 2016: Sources and Methods. Mimeo. http://www
.wiod.org/protected3/data16/SEA/SEA16_Sources.pdf.

Harms, Philipp, Oliver Lorz, and Dieter Urban. 2012. “Offshoring along the Production Chain.”

Canadian Journal of Economics 45 (1): 93–106.

Hijzen, Alexander, and Paul Swaim. 2010. “Offshoring, Labour Market Institutions and the

Elasticity of Labour Demand.” European Economic Review 54 (8): 1016–34.

International Labour Organization (ILO). 2011. Trade and Employment: From Myths to Facts.

Genève.

Johnson, Robert C., and Guillermo Noguera. 2012. “Accounting for Intermediates: Production

Sharing and Trade in Value Added.” Journal of International Economics 86 (2): 224–36.

Kee, Hiau Looi, and Heiwai Tang. 2016. “Domestic Value Added in Exports: Theory and Firm

Evidence from China.” American Economic Review 106 (6): 1402–36.

Kiyota, Kozo. 2014. “Exports and Employment in China, Indonésie, Japan and Korea.” OECD

Trade Policy Papers No. 166.

Koopman, Robert, Zhi Wang, and Shang-Jin Wei. 2014. “Tracing Value-Added and Double

Counting in Gross Exports.” American Economic Review 104 (2): 459–94.

Leontief, Wassily. 1936. “Quantitative Input and Output Relations in the Economic Systems of

the United States.” Review of Economics and Statistics 18 (3): 105–25.

Los, Bart, Marcel P. Timmer, and Gaaitzen J. de Vries. 2015. “How Important Are Exports for
Job Growth in China? A Demand Side Analysis.” Journal of Comparative Economics 43
(1): 19–32.

Portella-Carbo, Ferran. 2016. “Effects of International Trade on Domestic Employment: Un
Application of a Global Multiregional Input–Output Supermultiplier Model (1995–2011).»
Economic Systems Research 28 (1): 95–117.

Timmer, Marcel P., Erik Dietzenbacher, Bart Los, Robert Stehrer, and Gaaitzen J. de Vries.
2015. “An Illustrated User Guide to the World Input–Output Database: The Case of Global
Automotive Production.” Review of International Economics 23 (3): 575–605.

Timmer, Marcel P., Abdul Azeez Erumban, Bart Los, Robert Stehrer, and Gaaitzen J. de Vries.
2014. “Slicing Up Global Value Chains.” Journal of Economic Perspectives 28 (2): 99–118.
Timmer, Marcel P., Bart Los, Robert Stehrer, and Gaaitzen J. de Vries. 2013. “Fragmentation,
Incomes and Jobs: An Analysis of European Competitiveness.” Economic Policy 28 (76):
613–61.

_____. 2016. “An Anatomy of the Global Trade Slowdown Based on the WIOD 2016 Release.”

GGDC Research Memorandum No. 162. University of Groningen.

Wang, Zhi, Shang-Jin Wei, Xinding Yu, and Kunfu Zhu. 2017. “Measures of Participation
in Global Value Chains and Global Business Cycles.” Mimeo. https://www.nber.org
/papers/w23222.pdf.

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
un
d
e
v
/
un
r
t
je
c
e
–
p
d

F
/

3
6
2
1
0
0
1
6
4
4
2
9
4
un
d
e
v
_
un
_
0
0
1
3
3
p
d

b
oui
g
toi
e
s
t

o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
Télécharger le PDF