Technology and Investment, 2013, 4, 236-243
Published Online November 2013 (http://www.scirp.org/journal/ti)
http://dx.doi.org/10.4236/ti.2013.44028
Open Access TI
Inter-Industry Productivity Spillovers from Japanese and
US FDI in Mexico’s Manufacturing Sector
Leo Guzmán Anaya
University of Guadalajara, Guadalajara, Mexico
Email: leo.guzman8@gmail.com
Received April 24, 2013; revised May 24, 2013; accepted May 30, 2013
Copyright © 2013 Leo Guzmán Anaya. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
ABSTRACT
Foreign Direct Investment can have positive effects on host countries by generating spillovers to domestic firms and
contributing to increases in their productivity. These productivity spillovers1 can take place within an industry (intra-
industry spillovers) and across industries (inter-industry spillovers) as in the case of technology or knowledge transfer
to domestic suppliers (backward productivity spillovers) or customers (forward productivity spillovers). Using unpub-
lished economic census data from Mexico’s manufacturing sector this study differs from others by comparing inter-
industry productivity spillovers from Japanese and US FDI. Results show that Japanese FDI increases the productivity
of upstream sectors; however these gains seem to be shared only among foreign suppliers, while US FDI does not seem
to generate backward productivity spillovers. Results show no presence of forward productivity spillovers.
Keywords: Foreign Direct Investment; Productivity Spillovers; Backward Linkages; Forward Linkages; Mexican
Manufacturing Sector
1. Introduction
Foreign Direct Investment (FDI) has been high on the
agenda of many policy makers in developing countries,
especially after the decrease of bank lending of the
1980’s, which prompted most of these countries to invite
FDI by reducing restrictions on foreign capital and by
offering subsidies and tax incentives to foreign investors
[2]. Reference [3] mentions that in 1998, 60 countries
made regulatory changes towards FDI, where 94% were
aimed to generate more favorable conditions to foreign
investors. Mexico has also eased restrictions on FDI,
especially after the debt crisis of the 1980’s where for-
eign investment became the main source of financing
after 1988 [4]. The Mexican government has since then
pursued active policies to lower the entry barriers to in-
vestment from Multinational Enterprises (MNEs) in the
hope that FDI would promote economic development
through technological, knowledge and productivity spill-
overs as well as through export growth. The country has
also transitioned from an import substitution to an export
promotion strategy where FDI has played a central role.
Since the middle of the 1980’s, and more particularly
since the implementation of the North American Free
Trade Agreement (NAFTA) in 1994, FDI flows to Mex-
ico increased2. Despite the increases of FDI inflows to
Mexico and other developing countries, the spillover
effects of FDI still remain unclear [2].
The economic foundations to offer special incentives
to attract FDI derive from the idea that foreign invest-
ment is a composite bundle of capital, know-how, and
technology which produces externalities (known as “spill-
overs”), mainly through technology transfer as well as
technology and skill diffusion in the countries that import
FDI, which will increase the productivity of local firms
([3,5,6]). Also, the literature distinguishes between spill-
overs to firms in the same industry (intra-industry spill-
overs) from spillovers to firms across industries (inter-
industry spillovers). This implies that local firms may
benefit from presence of MNEs in their same sector or
from FDI presence in upstream or downstream sectors
that have linkages with local firms.
Early empirical literature focused on intra-industry spill-
2After receiving FDI inflows worth around 5.5 billion constant 2000
US dollars in 1993, the amount in 1994 was over 11 billion US dollars
in constant 2000 prices.
1Productivity spillovers are defined according to [1] as “the average
p
roductivity gains of domestic firms due to foreign firm presence in the
host country”.
L. GUZMAN 237
overs in several countries, generally bypassing the possi-
bility of linkages between local suppliers and MNEs. Re-
cently however, interest has been more oriented towards
analyzing inter-industry spillovers in countries recipients
of FDI. Results seem more robust in terms of inter-in-
dustry spillovers compared with intra-industry spillovers.
This study differs from previous ones by analyzing inter-
industry productivity spillovers from Japanese and US
FDI in Mexico’s manufacturing sector. In particular, the
interest lies in whether or not Japanese FDI generates lar-
ger inter-industry productivity spillovers than US FDI
due to the nature and organizational characteristics, which
according to theoretical studies are embodied in each
type of investment. By doing so, the present study found
that inter-industry productivity spillovers do differ by
origin of the investor.
The present article is organized as follows: Section 2
analyses the theoretical and empirical framework related
to productivity spillovers, presenting the main arguments
why host countries are more likely to benefit at the inter-
industry level than at the intra-industry level. Also, why
spillover effects from Japanese and US FDI may differ
due to individual characteristics. Section 3 describes the
methodological strategy applied as well as the data used
in the analysis. Section 4 presents the results obtained
and discusses the main findings. The fifth section ends
with some concluding remarks.
2. Theoretical and Empirical Framework
Theoretically, productivity spillovers from FDI represent
a type of positive effects from foreign firm presence on
the host country’s local firms3. The productivity gains
may come in several ways. Reference [7] recognizes three
channels for productivity spillovers: worker mobility, com-
petition effect and demonstration effect. The first occurs
when highly trained and skilled staff from foreign firms
moves to domestic firms, the second when the entrance
of MNEs forces domestic firms to upgrade techniques to
remain competitive and productive4, and the third refers
to technological and know-how transfer to local firms
due to imitation of the more advanced practices of the
MNEs.
These benefits on local firms’ productivity represent
the intra-industry spillovers gains from FDI. However,
the spillover phenomenon is not only confined to the
same industry. FDI presence may benefit local econo-
mies at an inter-industry level through linkages between
foreign and domestic firms that form part of a production
chain. Inter-industry spillovers occur when productivity
benefits reach upstream or downstream sectors via cus-
tomer-supplier relationships. This means that inter-in-
dustry spillovers can occur when domestic firms serve as
suppliers of MNEs (backward productivity spillovers), or
when local firms buy inputs from foreign firms (forward
productivity spillovers).
Reference [8] identifies several channels through which
backward and forward productivity spillovers occur. The
types and characteristics of these channels are summa-
rized in Table 1.
Theoretical work has pointed out that spillovers are
more likely to be vertical (inter-industry) than horizontal
(intra-industry) in nature. For example [10] mentions that
MNEs may have a strong incentive to prevent knowledge
transfer to their competitors but may benefit from trans-
ferring expertise and know-how to their local suppliers
with the goal of reducing input costs and increasing input
quality. Also, the arrival of foreign firms increases the
demand for intermediate inputs and services, which in
turn increases productivity of downward related indus-
tries. Finally, [8] point out that when demand in a host
country is inelastic due to substitute goods absence, MNEs
will prefer those countries with limited domestic compe-
tition and numerous local suppliers, resulting in limited
intra-industry spillovers. Therefore, productivity gains from
MNEs’ presence are more likely to occur in non-com-
peting and complementary sectors.
With respect to differences in inter-industry productiv-
ity spillover effects from Japanese and US FDI, two
theoretical reasons give support to the claim that spill-
overs from Japan and the US must be different for the
case of Mexico. First, the theoretical model developed by
[11] on vertical linkages states that the share of interme-
diate inputs provided by MNEs is positively correlated
with the distance between host and home country5. Lar-
ger shares of indigenous sourcing imply more interac-
tions between local and MNEs and increase the potential
for spillovers to occur. Second, [15] theoretically com-
pares Japanese and US FDI spillover effects on develop-
ing countries and claims that Japanese FDI generates
larger spillovers than US FDI because Japanese enter-
prises invest in sectors that are more suitable for these
economies, in terms of a more accessible, standardized
and mature technology that can be easily absorbed by
3It is important to note the distinction between productivity spillovers
and technological spillovers, where the former refers to average pro-
ductivity gains of domestic firms due to foreign firm presence in the
host country, while the latter requires the productivity increase to be
associated with an improvement in the techniques used by local firms
[1]. For this analysis we focus on productivity spillovers according with
the previous definition.
4In this sense, there might also be a negative effect on domestic firms i
f
the MNE entry attracts demand away from the local counterparts. The
average productivity of local firms may increase because only the most
p
roductive firms survive the competition (“selection effect”).
5This has been empirically tested for Japanese and US MNEs. For US
MNEs, [12] show that there is an inverse relationship between trade
costs and sales of intermediate inputs from headquarters to overseas
affiliates. Japanese MNEs also seem to increase local sourcing in their
US affiliates when transportation distance and shipping delays increase
([13,14]).
Open Access TI
L. GUZMAN
Open Access TI
238
Table 1. Vertical spillover characteristics.
Type of spillover Channel Characteristics
Expansion of producer services Expansion of demand for local inputs due to foreign firm entry.
Backward Linkage externalities Provision of some type of assistance to endogenous suppliers from foreign firms.
The assistance can be technical, training or transfer of knowledge.
Better quality of inputs Availability of better and modern products to the local economy.
Forward Training for input use Provision of guidelines and assistance for the most effective usage of products.
Source: Author’s compilation based on [8,9].
local firms.
Empirically, work searching for productivity spillovers
has extended and refined the original model pioneered by
[16]. The methodology used has usually analyzed intra-
industry spillovers by searching for changes in produc-
tivity of local firms due to FDI presence in a sector/in-
dustry. Positive spillovers from foreign presence have
generally been found when cross-sectional data has been
used, which may be due to estimation bias6. When panel
data estimation techniques were introduced, mixed re-
sults appeared and casted doubt on the existence of in-
tra-industry productivity spillovers from FDI. An expla-
nation for these results is that productivity spillovers are
more likely to be vertical than horizontal in nature.
Recent empirical work seems keener to search for in-
ter-industry spillovers as opposed to intra-industry spill-
overs. Results from inter-industry studies seem to sup-
port the claim that domestic suppliers gain from foreign
presence. Specifically, most studies show that for devel-
oping countries, domestic firms usually benefit from spill-
overs through backward linkages. A summary of empiri-
cal work at the inter-industry level is presented in Table
2.
For Mexico, the previous studies by [28,29] support
the presence of positive backward productivity spill-
overs7. Reference [28] also reports positive and signifi-
cant forward productivity spillovers for Mexico.
Empirical studies comparing productivity spillovers from
Japanese FDI and US FDI have primarily searched for
intra-industry spillovers. Reference [35] found that pro-
ductivity spillovers from Japanese FDI were strongest for
the electronic industry in the UK, while no positive ef-
fects on productivity were found for US FDI. Similarly,
[36] found positive effects from Japanese FDI on Total
Factor Productivity Growth (TFPG) of domestic firms in
India, while US FDI did not gender significant results.
On the other hand, [37] found no positive spillover ef-
fects from Japanese or US FDI for a panel of 26 devel-
oping countries. In fact, under some specifications, the
study found negative effects from US and Japanese FDI
on labor productivity and value added of these countries.
One of the reasons for contradicting intra-industry re-
sults may be that foreign firms have the incentive to pre-
vent any kind of positive spillovers to competing firms in
the same industry. On the other hand, foreign companies
that benefit directly or indirectly may promote spillovers
to domestic firms with an input-output linkage so in-
ter-industry spillovers are more likely to occur [33]. Ref-
erence [21] searched for inter-industry spillovers by dif-
ferentiating origin of the investor for Romania. The study
found that productivity of local supplying firms was
positively correlated with FDI presence from Asian or
US investors, but was negatively correlated with FDI
presence for the European Union (EU). The main argu-
ments are that investors outside of the EU do not enjoy
preferential terms and would source more of their inputs
locally to comply with the “rules of origin”, especially
for firms that use Romania as an export platform to other
EU members. Also, the distance between the home and
host country will play a part in the sourcing decisions of
the multinationals8. Interestingly, in the study US and
Asian FDI represented only 8% and 10% of total FDI in
Romania, while EU represented 61%. Similar results were
found in [23] for China, where country of origin also
seemed to matter in the existence and magnitude of inter-
industry spillovers.
The results from previous studies seem to indicate that
country of origin matters in the existence of vertical pro-
ductivity spillovers. However, no previous inter-industry
study has been carried out separating the source of FDI
flows from the US and Japan. Also, previous studies that
have compared these two types of FDI have only focused
on intra-industry spillovers, which could be the reason
for the mixed results. The present study attempts to fill
this gap, and analyze the inter-industry linkages from
Japanese and US FDI in Mexico’s manufacturing sector.
7These studies have several limitations. Reference [28] had to assume
that foreign ownership remained unchanged for the period of analysis
(1993-1999), since it was only available for 1993. [29] used cross-
sectional data so productivity spillovers were not observed through
time, which is an important part of the externality transmitting process
[34].
8This is in accordance with the theoretical argument from [11], which
states that the degree of procurement is expected to be positively cor-
related with distance.
6Using cross section data does not allow investigating the development
of firms’ productivity over time, which is an important part of the ex-
ternality transmitting process. Also, the direction of causality is hard to
establish, since it is possible that the positive effects on productivity are
due to the fact that MNEs tend to locate in the most productive sectors
and not be the cause of the productivity gains.
L. GUZMAN 239
Table 2. Inter-industry spillover empirical evidence from FDI.
Reference No. Study Sample Year HorizontalVertical Backward Forward
[8] Reganati and Sica (2007) 40,000 firms in Italy 1997-2002Not found (+) Not tested
[9] Javorcik (2008)
Enterprise surveys Czech
Republic and Latvia 2003, 2004Mixed Mixed Mixed
[10] Lall (1980) Case Study India 1979 Not tested (+) Not tested
[17] Blalock (2001) 20,000 firms in Indonesia 1988-1996Not found (+) Not tested
[18] Blalock and Gertler (2008) 20,000 firms in Indonesia 1988-1996Not found (+) Not tested
[19] Schoors and Van der tol (2002) 1000 firms in Hungary 1997, 1998(+) (+) ()
[20] Javorcik (2004) 4000 firms in Lithuania 1996-2000Not found (+) Not found
[21] Javorcik, Saggi and
Spatareanu (2004) 50,000 firms in Romania 1998-2000(+) (+) & () Not tested
[22] Liu and Lin (2004) 1700 firms in China 1999-2002() (+) Not found
[23] Du, Harrison and
Jefferson (2011) 250,000 firms in China 1998-2007Not found (+) (+)
[24] Bwalya (2006) 125 firms in Zambia 1993-1995() (+) Not tested
[25] Jabbour and Mucchielli (2007) 2000 firms in Spain 1990-2000() (+) (+)
[26] López and Südekum (2009) 5000 firms in Chile 1990-1999(+) (+) Not found
[27] Iyer (2009) 9000 firms in India 1989-2004Mixed Mixed Mixed
[28] López-Córdoba (2003) 5300 firms in Mexico 1993-1999() (+) (+)
[29] Jordaan (2008) 6-digit industry level data Mexico1993 () (+) Not tested
[30] Yudaeva et al. (2003) 23,000 firms in Russia 1993-1997(+) () ()
[31] Stancik (2007) 4000 firms in Czech Republic1995-2003() () Not found
[32] Vacek (2010) 671 firms in Czech Republic 1995-2004Not found (+) Not found
[33] Kugler (2001) Firm-level data, 10
manufacturing sectors Colombia1974-1998Not found(+) Not tested Not tested
Source: Author’s compilation.
It is important to note that similar to the study from [21]
for Romania, FDI in Mexico is predominantly from one
source, where the US accounts for 52% of FDI inflows
for the period 1999-2010. Japan on the other hand, only
accounts for 1% during the same time period. However,
as presented in [21], Japanese FDI may have the incen-
tives to generate backward and forward linkages to a
larger extent than US FDI due to the characteristics em-
bodied in each type of investment.
3. Methodology
The present study uses a balanced industry-level panel
data from INEGI9. The data consists of unpublished eco-
nomic census data for the years 1999, 2004 and 2009 for
the manufacturing sector at a 6-digit NAICS10 level. Ide-
ally, it would be advisable to work with plant-level data;
however, INEGI is prevented by law to disseminate plant
level information from economic censuses to avoid plant
identification. In total, 282 classes of manufacturing ac-
tivities were included in the analysis11. INEGI also pro-
vided the inter-industry input-output (IO) matrix12 used
to derive the measure for backward and forward linkages
from FDI, as well as the information on producer prices
adopted to deflate those values presented in nominal
terms. The dataset includes information on foreign par-
ticipation (from Japan or US origin), production, capital,
labor and material inputs by industry.
The econometric model used to search for inter-in-
dustry spillovers from Japanese and US FDI in Mexico is
based on the model pioneered by [17] and extended by
[20,21]. The previous models are extended in this analy-
sis by allowing backward and forward linkages to be
present depending on the nationality of the investor. The
model takes the following form:
123 4
56 7
8
ln
lnlnlnHorizontal
Backward BackwardForward
Forward
jt
jt jtjtjt
jtjt jt
jtj t jt
Y
KLM
USJAP US
JAP

 

 
 

Where,
j
t, Y
j
t
K
,
j
t and L
j
t
M
represent output,
capital, labor and material inputs for industry j at time t.
All variables were converted to constant 2010 prices us-
ing industry information on producer prices. Output repre-
12The IO matrix was made available for 2003. Ideally, it would be
etter to use multiple IO matrices since relationships between sectors
may change over time or with FDI activity. However, radical changes
among these relationships are unlikely and we can assume that indus-
trial structures do not change rapidly over time ([20,21]). IO matrices
for other years were unavailable. While coefficients from the IO table
remain fixed for the entire period, horizontal values and output do
change over time, so backward and forward values are time-varying,
industry-specific variables [20].
9Mexico’s National Institute of Statistics and Geography.
10North American Industry Classification System.
11The manufacturing sector under NAICS classifications has a total o
f
293 classes; 11 classes were dropped due to missing observations.
Open Access TI
L. GUZMAN
240
sents gross total output in an industry. Capital is meas-
ured as total fixed assets in an industry. Labor is meas-
ured by the total number of workers in an industry and
material inputs represent the value of intermediate mate-
rial inputs used by industry.
The Horizont al
j
t variable represents foreign pres-
ence in the same industry and is measured as the share of
industry j’s output produced by foreign firms in time t.
This variable measures intra-industry spillovers from FDI.
The Backward
j
t variable was added to test the in-
ter-industry productivity effects from US and Japanese
FDI. It represents the share of an industry’s output pur-
chased by downstream foreign enterprise. It is a proxy
for foreign presence in downstream sectors that intends
to capture the effect of foreign enterprises as customers
of domestic suppliers [17]. It is defined as:
Backward Horizontal
j
tjk
kif kj

kt
Where
j
m
is the share of upstream industry j’s out-
put used by downstream sector k taken from the 2003 IO
matrix. Two different measures of this variable are cal-
culated, one for Japanese FDI and one for US FDI re-
spectively.
The Forward
j
t variable measures the share of an in-
dustry’s output sold by upstream foreign enterprises to
domestic customers. It is a proxy for foreign presence in
upstream sectors; it intends to capture the effect of for-
eign enterprises as suppliers of domestic producers. It
takes the following form:
Forward Horizontal
j
tjm
mifm j

mt
Where
j
m
is the share of downstream industry j’s
output used by upstream sector m taken from the 2003 IO
matrix. This variable is also calculated for the case of
Japan and US foreign presence.
Finally,
j
and t
are industry and time specific
effects, and
2
~0,II
jt D
is the error term.
4. Results
The test for the spillover hypothesis was carried out on
the full sample and on a sub-sample excluding foreign
firm production, in other words only focusing on domes-
tic production13. Also, the Levinsohn-Petrin14 correction
was used to deal with the endogeneity problem that arises
when estimating production functions15. Using the Le-
vinsohn-Petrin correction allowed obtaining Total Factor
Productivity (TFP) values that were later used as de-
pendent variable in the basic model. We also extended
the model with first differences16. Results from both es-
timations are presented in Tables 3 and 4 respectively.
Table 3 shows that the variables jointly considered are
significant at a 1 per cent level. From the variables of
Table 3. Results from fixed effects estimation model.
Dependent variable: InTFPjt
Regressors All Domestic
InKjt 0.53***
(30.51)
0.34***
(7.30)
InLjt 0.27***
(10.45)
0.26***
(3.90)
InMjt 0.78***
(39.67)
0.68***
(13.17)
Horizontaljt 0.07*
(1.79)
0.04***
(23.66)
BackwardUSjt 0.09
(0.84)
0.01
(0.51)
BackwardJAPjt 1.94***
(2.86)
0.67
(1.09)
ForwardUSjt 0.11
(0.74)
0.05
(1.25)
ForwardJAPjt 1.76
(1.00)
0.42
(0.51)
Constant 2.88***
(9.10)
3.94***
(4.90)
R2 0.78 0.55
F 246.74*** 84.90***
nOBS 846 846
Notes: *Significance at 90% interval; **Significance at 95% interval;
***Significance at 99% interval. Numbers in parenthesis indicate t statistics.
interest, the variable “Horizontal”, which measures intra-
industry spillovers results positive and significant when all
production is considered and negative and significant
when only domestic production is tested. This result is
interesting and seems to indicate that FDI arrival in
Mexico has been accompanied by competitive pressure
that has pushed less efficient domestic firms from the
same industry out of the market. In this sense productivity
spillover gains are shared only among foreign firms.
Inter-industry spillovers were tested in the “Backward”
and “Forward” variable. The backward variable for the
case of the US showed negative but not statistically sig-
nificant values, the same was observed for the US for
13The presence of an unobserved effect in the data was confirmed
through the Breusch-Pagan Lagrange multiplier test. Afterwards, the
chi-squared Hausman test was conducted to test for inconsistency in the
random effects model and determined the appropriate estimation
method.
14Reference [38].
15The main problem is that the estimates might be biased due to corre-
lation between input levels and productivity shocks. Reference [39]
developed a semiparametric method to account for endogeneity o
input demand. Specifically, they proposed an estimator that uses in-
vestment as a proxy to account for the unobservable shocks. However,
[38] mention that using the investment proxy might not always work.
They mention that investment is a control on a state variable, which has
adjustment costs that may affect the responsiveness to productivity
shocks and thus violate the monotonicity condition. Also, the invest-
ment proxy is only valid for industries that report nonzero investment.
16The main reason for differencing the equation is to address the poten-
tial problem of omission of unobserved variables, which could in turn
affect the relationship between foreign presence and industry produc-
tivity [8]. First differencing permits the removal of these potential
unobservable effects.
Open Access TI
L. GUZMAN 241
Table 4. Results from fixed effects and random effects dy-
namic estimation model.
Dependent variable: ΔInTFPjt
Regressors All Domestic
ΔInKjt 0.52***
(28.59)
0.74***
(6.62)
ΔInLjt 0.27***
(9.40)
0.06
(0.44)
ΔInMjt 0.76***
(36.51)
0.87
(10.34)
ΔHorizontaljt 0.02
(0.68)
0.07***
(14.37)
ΔBackwardUSjt 0.04
(1.01)
0.05*
(1.66)
ΔBackwardJAPjt 2.41***
(2.78)
0.23
(0.11)
ΔForwardUSjt 0.08
(0.94)
0.02
(0.13)
ΔForwardJAPjt 0.69
(0.24)
2.76
(0.52)
Constant 0.02
(0.88)
8.90
(6.28)
R2 0.74 0.56
F 42.85***
X2 1666.62***
nOBS 564 564
Notes: *Significance at 90% interval; **Significance at 95% interval; ***Sig-
nificance at 99% interval. Numbers in parenthesis indicate t (or z for random
effects) statistics. Fixed Effects was carried out for “domestic production”
and Random Effects for “all production”.
Forward productivity spillovers. This seems to indicate
that US FDI in an industry has not created any type of
inter-industry linkages during the period of analysis. A
reason for these results may be that US companies in
Mexico import most of their inputs from abroad and use
Mexico as an assembly location where production is
again exported to foreign markets. In this sense, no in-
teraction between US affiliates and Mexican domestic
suppliers/buyers is present.
For the case of Japanese FDI, results showed positive
and significant backward linkages with all production and
not statistically significant results with domestic produc-
tion. More precisely, the point estimate suggests that an
increase of 1 per cent in Japanese presence in an industry
is followed by an increase in productivity of upstream-
related industries of 1.94 per cent. The lack of significant
results for domestic production seem to indicate that Japa-
nese firms use other foreign firms as suppliers, which
could be because these MNEs have prearranged contracts
with a supplying network or because Mexico’s support-
ing industry does not meet the requirements needed by
these firms. The forward variable was not statistically
significant for Japan under any specification; this means
that Japanese suppliers are not increasing productivity of
either Mexican or foreign firms in Mexico.
Finally, Table 4 shows the estimation under first dif-
ferences of the TFP variable. Results remained consistent
with the ones presented in Table 3. Negative and sig-
nificant intra-industry spillovers from FDI were found
with domestic production, showing that an increase from
0 to 10 per cent in foreign presence in an industry reduces
productivity growth of domestic firms in that industry by
0.6 percent. Similarly, domestic suppliers reduce their
productivity growth by 0.7 percent when US firms in-
crease their presence from 0 to 10 per cent. Japanese FDI
reported positive and statistically significant effects on
all production. The point estimate indicates that an in-
crease of 1 per cent in Japanese presence in an industry
genders an increase in the productivity growth rate of
2.41 per cent in supplying industries. However, no ef-
fects were found over domestic production. Finally, no
forward productivity growth spillovers were reported for
either Japanese or US FDI.
5. Conclusions
Using unpublished economic census data from Mexico’s
manufacturing sector, this study analyzed whether the
nationality of the foreign investor mattered for produc-
tivity spillovers at the inter-industry level. Intra-industry
productivity spillovers were also tested due to foreign
firm presence.
Results have shown that vertical productivity spill-
overs from Japanese and US FDI differ. Japanese FDI
had positive and significant effects on productivity and
productivity growth of backward related industries with
elasticities of 1.94% and 2.41% respectively. US FDI on
the other hand, was only statistically significant in terms
of productivity of upstream sectors with negative effects.
Results seem at odd at first, since US FDI is the main
source of foreign investment in Mexico, representing
over 50% of total FDI inflows between 1990 and 2010.
However, results are consistent with the empirical find-
ings from [21] for Romania and with the theoretical
model from [11] that indicates that inter-industry link-
ages are not determined by the amount of the investment,
but are rather positively affected by the distance between
headquarters and production plant in the recipient coun-
try.
Also, since spillover gains from FDI might be ob-
served only among foreign firms, the specifications were
tested on domestic production. Results changed signifi-
cantly. No significant results were found for Japanese
FDI in backward productivity spillovers of domestic
firms. US FDI showed a negative and significant result in
backward productivity spillovers with an elasticity of
0.05%. No forward productivity spillovers were re-
ported for either Japanese or US FDI presence. These
results seem to indicate that Japanese firms in Mexico are
Open Access TI
L. GUZMAN
242
using other foreign suppliers, where the productivity gains
remain among the more efficient foreign firms supplying
Japanese producers. Since most Japanese MNEs form
part of a “Keiretsu”17, it can be argued that the usage of
suppliers usually comes from members of the same busi-
ness group where supplier networks and prearranged con-
tracts are established. US FDI showed detrimental effects
on the development of Mexico’s supplying industry.
These results could be a reflection of US enterprises im-
porting most of their inputs under NAFTA’s preferential
terms and taking advantage of programs that encourage
temporal imports of intermediate products. Also, Mexi-
can firms do not appear to use US firms as suppliers,
which might be due to the fact that most of the produc-
tion from US MNEs is for export purposes.
Horizontal spillovers were also tested and interesting
results were obtained. The initial specification showed
positive and significant intra-industry productivity spill-
overs, indicating that FDI in an industry increases pro-
ductivity of firms within the industry. However, when
only domestic production was analyzed, the results showed
negative productivity spillover effects from FDI. These
findings seem to indicate that foreign firm presence in
Mexico’s manufacturing sector has been accompanied by
competitive pressure that has pushed less efficient do-
mestic firms from the same industry out of the market. In
this sense, as in the case of inter-industry productivity
spillovers, the productivity gains from FDI seem to be
shared only among foreign firms.
Finally, the results reported here have several policy
implications. First, governments cannot treat all FDI from
different origins in the same manner, results have shown
that differences exist and governments have to encourage
FDI that is the most adequate to their development goals
and brings positive spillovers to the local economy. Sec-
ond, governments also have to provide the right envi-
ronment for spillovers to occur, encourage the develop-
ment of the local supplier base that is adequate to serve
established and prospective foreign firms and thus facili-
tate the technological exchange between foreign and do-
mestic firms [1]. Third, it is important to point out that
developing local suppliers may not be enough, since for
many MNEs the input sourcing decision is taken at head-
quarters and generally follows the use of international
supplier networks. Governments have to treat each in-
vestment project as a different phenomenon and don’t
expect to have the same benefits from all FDI received.
In this sense, the present analysis has shown the national-
ity of the investor matters in the presence of inter-indus-
try productivity spillovers, in other words, FDI from dif-
ferent sources has different effects on the host country.
REFERENCES
[1] T. Perez, “Multinational Enterprises and Technological
Spillovers,” Hardwood, Amsterdam, 1998.
http://dx.doi.org/10.4324/9780203362914
[2] M. Carkovic and R. Levine, “Does Foreign Direct In-
vestment Accelerate Economic Growth?” Working Paper,
University of Minnesota Department of Finance, 2002.
[3] H. Görg and D. Greenaway, “Much Ado about Nothing?
Do Domestic Firms Really Benefit from Foreign Direct
Investment,” Discussion Paper No. 944, Institute for the
Study of Labor, 2003.
[4] J. A. Gurría, “Flujos de Capital: El Caso de México,”
Serie Financiamiento del Desarrollo CEPAL, No. 27,
1994.
[5] V. N. Balasubramanyam, M. Salisu and D. Sapsford,
“Foreign Direct Investment and Growth in EP and IS
Countries,” The Economic Journal, Vol. 106, No. 434,
1996, pp. 92-105. http://dx.doi.org/10.2307/2234933
[6] M. Blomström, S. Globerman and A. Kokko, “The De-
terminants of Host Country Spillovers from Foreign Di-
rect Investment: Review and Synthesis of the Literature,”
In: N. Pain, Ed., Inward Investment, Technological Chan-
ge and Growth, Palgrave, London, 2001, pp. 34-66.
[7] M. Blomström and A. Kokko, “How Foreign Direct In-
vestment Affects Host Countries,” World Bank Working
Paper No. 1745, 1997.
[8] F. Reganati and E. Sica, “Horizontal and Vertical Spill-
overs from FDI: Evidence from Panel Data for the Italian
Manufacturing Sector,” Journal of Business Economics
and Management, Vol. 8, No. 4, 2007, pp. 259-266.
[9] S. Lall, “Vertical Inter-firm Linkages in LDCs: An Em-
pirical Study,” Oxford Bulletin of Economics & Statistics,
Vol. 80, No. 3, 1980, pp. 203-226.
[10] B. Javorcik, “Can Survey Evidence Shed Light on Spill-
overs from Foreign Direct Investment?” The World Bank
Research Observer, Vol. 23, No. 2, 2008, pp. 139-159.
http://dx.doi.org/10.1093/wbro/lkn006
[11] A. Rodríguez-Clare, “Multinationals, Linkages and Eco-
nomic Development,” The American Economic Review,
Vol. 86, No. 4, 1996, pp. 852-873.
[12] G. Hanson, R. Mataloni and M. Slaughter, “Vertical Pro-
duction Networks in Multinational Firms,” The Review of
Economics and Statistics, Vol. 87, No. 4, 2005, pp. 664-
678. http://dx.doi.org/10.1162/003465305775098080
[13] W. Chung, W. Mitchell and B. Yeung, “Foreign Direct
Investment and Host Country Productivity: The American
Automotive Component Industry in the 1980s,” Journal
of International Business Studies, Vol. 34, No. 2, 2003,
pp. 199-218.
http://dx.doi.org/10.1057/palgrave.jibs.8400017
[14] X. Martin, W. Mitchell and A. Swaminathan, “Recreating
and Extending Japanese Automobile Buyer-Supplier Links
in North America,” Strategic Management Journal, Vol.
16, No. 8, 1995, pp. 589-619.
http://dx.doi.org/10.1002/smj.4250160803
[15] K. Kojima, “Japanese Direct Foreign Investment, A
Model of Multinational Business Operations,” Charles E.
17The Keiretsu represents a group of companies with interlocking busi-
ness relations and shareholdings.
Open Access TI
L. GUZMAN
Open Access TI
243
Tuttle Company, Tokyo, 1978.
[16] R. E. Caves, “Multinational Firms, Competition and Pro-
ductivity in Host-Country Markets,” Macmillan, London,
1974.
[17] G. Blalock, “Technology from Foreign Direct Investment:
Strategic Transfer through Supply Chains,” Ph.D. Disser-
tation, Haas School of Business, University of California,
Berkley, 2001.
[18] G. Blalock and P. J. Gertler, “Welfare Gains from For-
eign Direct Investment through Technology Transfer to
Local Suppliers,” Journal of International Economics,
Vol. 74, No. 2, 2008, pp. 402-421.
http://dx.doi.org/10.1016/j.jinteco.2007.05.011
[19] K. Schoors and B. van der Tol, “Foreign Direct Invest-
ment Spillovers within and between Sectors,” Working
Paper No. 157, Department of Economics, Ghent Univer-
sity, Belgium, 2002.
[20] B. Javorcik, “Does Foreign Direct Investment Increase
the Productivity of Local Firms? In Search of Spillovers
through Backward Linkages,” American Economic Re-
view, Vol. 94, No. 3, 2004, pp. 605-627.
http://dx.doi.org/10.1257/0002828041464605
[21] B. Javorcik, K. Saggi and M. Spatareanu, “Does It Matter
Where You Come From? Vertical Spillovers from For-
eign Direct Investment and the Nationality of Investors,”
World Bank Policy Research Working Paper No. 3449,
2004.
[22] Z. Liu and P. Lin, “Backward Linkages of Foreign Direct
Investment,” 2004.
http://www.docin.com/p-98810674.html
[23] L. Du, A. Harrison and G. H. Jefferson, “Testing for
Horizontal and Vertical Foreign Investment Spillovers in
China, 1998-2007,” Journal of Asian Economics, Vol. 23,
No. 3, 2012, pp. 234-243.
http://dx.doi.org/10.1016/j.asieco.2011.01.001
[24] S. Bwalya, “Foreign Direct Investment and Technology
Spillovers: Evidence from Panel Data Analysis of Manu-
facturing Firms in Zambia,” Journal of Development Eco-
nomics, Vol. 81, No. 2, 2006, pp. 514-526.
http://dx.doi.org/10.1016/j.jdeveco.2005.06.011
[25] L. Jabbour and J. L. Mucchielli, “Technology Transfer
through Vertical Linkages: The Case of the Spanish Man-
ufacturing Industry,” Journal of Applied Economics, Vol.
10, No. 1, 2007, pp. 115-136.
[26] R. López and J. Südekum, “Vertical Industry Relations,
Spillovers, and Productivity: Evidence from Chilean Plants,”
Journal of Regional Science, Vol. 49, No. 2, 2009, pp.
721-747.
http://dx.doi.org/10.1111/j.1467-9787.2009.00631.x
[27] C. Iyer, “Foreign Firms and Inter-Industry Spillovers in
Indian Manufacturing: Evidence from 1989 to 2004,” The
Journal of Applied Economic Research, Vol. 3, No. 3,
2009, pp. 297-317.
[28] E. López-Córdoba, “NAFTA and Manufacturing Produc-
tivity in Mexico,” Economía, Vol. 4, No. 1, 2003, pp.
55-98. http://dx.doi.org/10.1353/eco.2004.0007
[29] J. A. Jordaan, “Intra- and Inter-Industry Externalities
from Foreign Direct Investment in the Mexican Manu-
facturing Sector: New Evidence from Mexican Regions,”
World Development, Vol. 36, No. 12, 2008, pp. 2838-
2854. http://dx.doi.org/10.1016/j.worlddev.2008.02.006
[30] K. Yudaeva, K. Kozlov, N. Melentieva and N. Ponoma-
reva, “Does Foreign Ownership Matter? The Russian
Experience,” Economics of Transition, Vol. 11, No. 3,
2003, pp. 383-409.
http://dx.doi.org/10.1111/1468-0351.00157
[31] J. Stancik, “Horizontal and Vertical FDI Spillovers: Re-
cent Evidence from the Czech Republic,” CERGE-EI
Working Paper Series No. 340, 2007.
[32] P. Vacek, “Panel Data Evidence on Productivity Spill-
overs from Foreign Direct Investment: Firm-Level Meas-
ures of Backward and Forward Linkages,” IES Working
Paper No. 19, 2010.
[33] M. Kugler, “Externalities from Foreign Direct Investment:
The Sectoral Pattern of Spillovers and Linkages,” Uni-
versity of Southampton, Mimeo, 2001.
[34] S. Rosenthal and W. Strange, “Evidence of the Nature
and Sources of Agglomeration Economies,” In: V. J.
Henderson and J. F. Thisse, Eds., Handbook of Urban
and Regional Economics, Vol. 4, North Holland, Am-
sterdam, 2004.
[35] S. Girma and K. Wakelin, “Regional Underdevelopment:
Is FDI the Solution? A Semi-Parametric Analysis,” Dis-
cussion Paper No. 2995, Centre for Economic Policy Re-
search, 2001.
[36] R. Banga, “Do Productivity Spillovers from Japanese and
US FDI Differ,” In: R. Jha, Ed., Indian Economic Reform,
Palgrave McMillan, New York, 2003, pp. 256-273.
[37] A. Lemi, “Foreign Direct Investment, Host Country Pro-
ductivity and Export: The Case of US and Japanese Mul-
tinational Affiliates,” Journal of Economic Development,
Vol. 29, No. 1, 2004, pp. 163-186.
[38] J. Levinsohn and A. Petrin, “Estimating Production Func-
tions Using Inputs to Control for Unobservables,” Review
of Economic Studies, Vol. 70, No. 2, 2003, pp. 317-342.
http://dx.doi.org/10.1111/1467-937X.00246
[39] S. Olley and A. Pakes, “The Dynamics of Productivity in
the Telecommunications Equipment Industry,” Econome-
trica, Vol. 64, No. 6, 1996, pp. 1263-1297.
http://dx.doi.org/10.2307/2171831