FDI Spillovers Effect, Environmental Pollution and Total Factor

Report
FDI Spillovers Effect,
Environmental Pollution and
Total Factor Productivity
Guoqing ZHAO1 and Zhongyuan ZHANG2
1. School of Economics, Renmin University of China,
Beijing, 100872 ,China
2. Institute of Asia-Pacific Studies, Chinese Academy of Social Sciences,
Beijing, 100007,China
2012-6
Outlines:
1. Introduction
2 Empirical Analysis Framework and Data
2.1 Basic Specification
2.2 Data and definitions of key variables
2.3 Measuring Forward and Backward FDI spillover
3 Estimation results
3.1 The Effects of FDI Spillover and Environmental
Pollution on Productivity
4 The FDI Spillover Effect and Environment Pollution
5 Conclusions
1. Introduction
•Some empirical studies confirm positive
productivity spillovers from FDI.
Blomstrom and Sjoholm(1999)
Sadik and Bolbol(2001)
•But others find negative or no spillovers.
Aitken and Harrison(1999)
Veugelers and Cassiman (2004)
•The mixed evidence intuitively implies that there is no
universal relationship between FDI and domestic firms’
productivity.
•Javorcik(2004) argues that researchers have been
looking for FDI spillovers in the wrong place since
multinationals have an incentive to prevent information
leakage that would enhance the performance of their
local competitors.
• the presence of FDI creates negative externalities
within industries and positive externalities between
industries through vertical linkages
Bwalya(2006) finds no intra-industry productivity
spillovers from FDI but significant inter-industry
knowledge spillovers occurring through linkages in
Zambia’s firm.
Jordaan(2008) finds negative externalities within
industries but positive externalities between industries
that in several Mexican regions in the early 1990s.
Using a large panel of Chinese manufacturing firms,
Liu(2008) finds that spillovers through backward and
forward linkages between industries
Lin et. al. (2009) find strong and robust vertical
spillover effects on both state-owned firms and non-state
firms in China’s manufacturing firms (above a minimum
scale) from 1998 to 2005.
•In this paper we also investigate the relationship
between FDI and the environmental performance of
industries
the pollution haven hypothesis (PHH), which states that
FDI will be attracted to those countries with less
stringent environmental regulations thus inducing a
regulatory “race to the bottom” in order to attract higher
FDI inflows from dirty sectors to the detriment of the
host country’s environment
Esty and Geradin, 1997;
Mani and Wheeler, 1998.
•In contrast, the pollution halo hypothesis argues that
the presence of foreign-owned firms may yield
substantial environmental benefits to developing
countries since FDI has been known to directly
encourage the dissemination of environmental related
knowledge and technologies
Albornoz et al., 2009.
2 Empirical Analysis Framework and
Data
2.1 Basic Specification
To examine the correlation between industry
productivity and FDI in the same sector
(intra-industry) and inter-industry, the
empirical framework can be described as
following:
Yit    1 K it   2 Lit   3 FHit   4 FWit
  5 BWit  Z it   i   it
(1)
Insert t heenvironmen
t al pollut ionindices
in equat ion (1) in order t oinvest igate t heir
effect son indust riesproduct ivit y :
Yit    1 K it   2 Lit   3 FHit 1   4 FWit 1
  5 BWit 1   6 EPit  Z it   i   it
(2)
2.2 Data and definitions of key variables
The datum were obtained from the China Statistical
Yearbook (NBSC, 2000–2009) which cover 28
industries from 1999 to 2008(2004 was the year for
which the data was unavailable).
Table 1 presents definitions of the key variables used in
the empirical estimations.
Table 2 Descriptive statistics(omitted)
Panel A Summary statistics
Panel B Correlation matrix
2.3 Measuring Forward and Backward FDI
spillover
calculate forward FDI spillover
effect in the following way:
FW jt   u  jut  FH ut
Where
 jut 
m1 jut

u
m1 jut
(3)
calculate backward FDI spillover
effect in the following way:
BW jt   b  jbt  FH bt
(4)
Where
 jbt 
m2 jbt

b
m2 jbt
This paper uses
the 2002 Input-Output Table for 1999-2002,
the 2005 Input-Output Table for 2003 and 2005 and
the 2007 Input-Output Table for 2006-2008.
3 Estimation results
3.1 The Effects of FDI Spillover and
Environmental Pollution on Productivity
•Table 3 presents the estimations of equation (2).
•use the Hausman test for our regression model to select
the proper specification between fixed-effect and
random-effect approach.
• first entry FH and FW, FH and BW (excluding
environment pollution variables) into the equation
alternatively.
•Focusing our attention to the spillover effects of FDI
firstly, we do not find any statistically significant effects
of horizontal FDI spillover on industry productivity
since the coefficients on FH, though positive, are
insignificant.
• In contrast, we find strong positive effects of backward
FDI spillover on industry’s productivity for the
coefficients on BW are significant positive, suggesting
effectively backward linkages across industries.
• we also find uneven positive spillovers from forward
FDI spillover, which is positive significant in column (1)
and loss its power in column (3) when we include BW
variable,
• Overall, when we explicitly separate horizontal
spillovers, backward linkages and forward linkages, our
results support that FDI spillovers are more likely
existing the backward linkages across industries.
•There are four environment pollution indices :
industry wasted water discharge per revenue from
principal business (WA),
total volume of industrial sulphur dioxide emission per
revenue from principal business (SO2),
total volume of industrial soot emission per revenue
from principal business (SMO) and
total volume of industrial dust emission per revenue
from principal business (DI),
Table 3 presents the estimations of the full specification
of equation (2).
•The most striking result from Table 3 is that, in all
specifications (from column (4) to (7)),
•the coefficients on horizontal FDI effect variable (FH)
are insignificant positive and the coefficients on
backward FDI spillover (BW) are significant positive
which echo the results of no statistically significant
effects of horizontal FDI spillover on industry
productivity and FDI spillovers are more likely existing
the backward linkages across industries.
the coefficients on WA, SO2, SMO and DI are all
significantly negative at conventional significant level,
suggesting environment pollutions do have
disadvantages on industry’ productivity progress.
3.2 The Dynamic empirical model of FDI
Spillover Effects and Environmental Pollution
on Productivity
specify a dynamic equation which includes a lagged dependent variable:
Yit     Yit 1  10 Kit  11 Kit 1   20 Lit   21 Lit 1
 3 FH it 1   4 FWit 1  5 BWit 1   6 EPit
 Zit   i   it
(5)
It is well known that OLS estimates are biased
and inconsistent
the generalized method of moments (GMM)
techniques developed by
Arellano and Bond (1991)
Arellano and Bover (1995) and
Blundell and Bond (1998),
the coefficients on backward FDI spillover variables (BW) are
significant positive except column (2), suggesting FDI spillovers
are more likely existing the backward linkages across industries.
There are some differences of the coefficients on forward FDI
spillover variables (FW), which are positive and significant in 3
columns now suggesting the possibility of spillover effect exists
among the forward linkage.
However, there are essential changes on the coefficients on
horizontal FDI spillover variables (FH), which are all significant
negative now, suggesting the competition effect is important
which preponderates over horizontal FDI spillover effect.
As for environment pollution variables, the
coefficients on WA and SO2 are significantly
negative, suggests environment pollutions do
have disadvantages on industry’ productivity
progress.
However, the coefficients on SMO is
insignificantly negative,
the coefficients on DI is significantly positive,
are different from the results of Table 3.
4 The FDI Spillover Effect and
Environment Pollution
4.1 The FDI Spillover Effect on the Marginal
Effect of Environment Pollution
Assume FDI spillover effect affects the marginal
effect of environment pollution on industry's
productivity progress.We then test whether the coefficients
on EP depend on either the horizontal FDI effect or the vertical FDI effect,
so that we have
 6   60   61  FH it 1   62  FWit 1   63  BWit 1
(6)
By substituting (6) into equation (2) ,
we derive the model:
Yit    1 K it   2 Lit   3 FH it 1   4 FWit 1
  5 BWit 1   60 EPit   61 EPFH it 1
  62 EPFWit 1   63 EPBWit 1  Z it 
 i   it
(7)
Table 5 reports the regression results which include the
interaction terms between environment pollution indices
and FDI spillover effect variables.
the coefficients on the interaction variable between FH
and WA, SO2 are significantly positive, the coefficients
on the interaction variable between FH and SMO, DI are
insignificantly positive, which suggest horizontal FDI
spillover effect mitigates the disadvantage effect of
environment pollution on industry’ productivity
progress.
The coefficients on the interaction variable
between FW and WA, SO2, SMO, DI are all
insignificantly negative, while the coefficients on
the interaction variable between BW and WA,
SO2, DI are all insignificantly positive,
suggesting no significant evidence of vertical
FDI spillover effect mitigates the disadvantage
effect of environment pollution on industry’
productivity progress.
4.2 The Level Effect of FDI Spillover on
Environment Pollution
•To test the level effect of FDI spillover on
environment pollution, in this paper we introduce
the multivariate linear regression model which is
a natural generalization of a linear regression
model.
•That is, two or more possibly correlated
dependent variables are simultaneously modeled
as the linear functions of the same set of
predictor variables.
There are four environment pollution indices
as proxies for environment pollution:
WA, SO2, AMO and DI in our model:
EPM it   0   1  FH it 1   2  FWit 1
  3  BWit 1  i  it
(8)
•Table 6 reports the regression results of equation which
include the contemporaneous environment pollution
indices and lag all the right hand side regressors by one
period.
•The coefficients on horizontal FDI spillover effect
variable (FH) in Panel A and B are all significantly
negative, suggesting horizontal FDI spillover decreases
the emission of environment pollution.
•However, the coefficients on backward FDI spillover
effect variable (BW) in Panel A and B are significantly
positive except in column (1), suggesting backward FDI
spillover increases the emission of environment
pollution.
•The results of forward FDI spillover effect are mixed,
•The consistently significant coefficient on horizontal
FDI spillover effect variable indicates that foreign firms
may adopt low-power consuming and low-environmentpollution-intensity technologies which low the emission
of environment pollution.
•However, it is not mean that they be willing to transfer
environmental knowledge within the same industry
because their generosity does not appear to extend to
direct competitors.
•Though we have strong evidences of the positive
vertical FDI spillover effect (especially through
backward linkages) that promotes industry’ productivity
progress,
•the backward FDI spillover increases the emission of
environment pollution. This maybe due to the spillover
of backward FDI linkages are high-power consuming
and high-environment-pollution-intensity technologies.
5 Conclusions
•examines the effects of the foreign direct investment
(FDI), which is distinguished as horizontal, forward
linkage and backward linkage spillovers, and
environmental pollution on total factor productivity
(TFP).
•the existence of positive spillovers from FDI taking
place through backward linkages, but there are no
significant evidences of spillovers occurring through
either the horizontal or the forward linkage channel.
•Environmental pollution has significant negative effect
on TFP.
•horizontal FDI spillover effect mitigates the
disadvantage effect of environment pollution on industry’
productivity progress and decreases the emission of
environment pollution
•the vertical FDI spillover effect (especially through
backward linkages) doesn’t mitigate the disadvantage
effect of environment pollution on industry’ productivity
progress and increases the emission of environment
pollution.
Q&A
Thank You!

similar documents