Transcript Does FDI Harm Host Country Environment? Evidence from Coastal

Does FDI Harm the Host Country’s
Environment?
Evidence from Coastal and Interior
China
Helen Feng Liang
University of California, Berkeley
April 12, 2006
Why study trade and pollution?

Environmental issues are global


global warming, ozone depletion, acid rain, etc.
Trade could have impact on environment


WTO, NAFTA
Are dirty industries moving to the South?

Trade could lead to growth, growth could be good or
bad for the environment

Trade policy and environmental policy
2
What makes China an interesting unit of
analysis?

Received a large amount of FDI in recent years

A popular destination of manufacturing outsourcing

Vast number of cities, large variety in location, economic
development, access to foreign investment, and policy

Increasing pollution problems
3
4
5
Research Question?

What’s the impact of foreign direct investment
on the environmental quality in China?

What are the channels of the impact?
6
Outline






Introduction
Theory
Foreign Direct Investment and Environment
in China
Data, Measurement, and Empirical Strategy
Result and Conclusion
Discussion
7
How Does Trade Influence Pollution?

Openness  Growth  Less or More Pollution


Trade  More output  More pollution: Scale Effect
Trade  Higher income  Higher demand for environ: Income Effect



Openness  Tech Efficiency  Less pollution: Technique Effect



Antweiler et al 2001, Frankel & Rose 2002
Wang & Jin 2002
Limitation of Cross Country studies:


Environmental Kutznet Curve: an inverse-U shaped relation between per capita GDP
and Pollution level
Grossman & Krueger 1993, 1995, Frankel & Romer 1999
Endogeneity of Trade Policy and Environmental Policy
Contribution of my study: within-country study to overcome
endogenous trade policy
8
Foreign Direct Investment in China:

What’s driving the distribution of FDI?

Government policy


Geographic location


4 special economics zones and 14 coastal open cities
Closeness to sea ports/trade hubs: Shanghai, Hong Kong, etc
FDI’s Effect on the Environment:

FDI  more pollution:


FDI  less pollution:




more output  more pollution
FDI  Tech Spillover  Improve Energy Efficiency
FDI  Competition  Crowd out less efficient firms  Improve Overall Energy Efficiency
FDI  Higher Income  Higher Demand for Environment
Ambiguous:

FDI could go to dirty or clean industries
9
Hypothesis

Hypothesis 1


EKC – Pollution level in China’s cities increases
with per capita GDP, but at a decreasing speed.
Hypothesis 2

Pollution level in China’s cities decreases with
foreign direct investment, for given industrial
composition and the level of output.
10
Environmental issues in China:
Primary Consumption (Mtce)
SO2 is a major source of pollution due to the structure of fuel
consumption
1,600
Primary Electricity
Natural Gas
1,400
Petroleum
1,200
Coal
1,000
800
600
400
200
0
1980
1985
1990
1995
2000
Source: China Energy Databook V6.0 Figure 4.A.1
11
Environmental issues in China:
China's Sulfur Dioxide Emission by Industry Sectors
China's Sulfur Dioxide Em ission by Industry Sectors in 2002
Source: China Statistic Yearbook 2003
Chemical M aterials &
Products, 4.71%
Nonmetal M ineral
Products, 9.92%
Others, 27.83%
Smelting and Rolling of
M etals, 9.52%
Electricity&Water
Production & Supply,
48.02%
Source: China Statistic Yearbook 2003
12
Data Source

1996-2003 panel of 231 cities in China: city-year observations

City level pollution



Industry composition and ownership


China National Bureau of Statistics
Other city social and economic variables


Estimated industrial SO2 emission: China Urban Statistics Yearbook 19972003
SO2 ambient concentration: China Energy Databook v6.0
China Urban Statistics Yearbook
Pollution intensity by industry sectors

World Bank Industrial Pollution Projection System (IPPS)
13
Variables

Dependent Variables – Sulfur Dioxide Emission:


Explanatory Variables:




Factory SO2 emission (tons)
FDI sector: asset and employment of foreign invested plants
Domestic sector: asset and employment of domestic plants
Income: per capita GDP and per capita GDP square
Control Variables:


Industrial output
Estimated SO2 emission based on industry composition



Employment * SO2 emission per employee
Land area, capital labor ratio
Year effects, provincial dummies, city fixed effects
14
FDI v.s. Domestic Factories:
Capital labor ratio
 FDI are more capital intensive
15
FDI v.s. Domestic Factories:
Expected SO2 emission per employee based on
industry composition
 FDI probably goes to “cleaner” industries
16
Empirical Strategy –
A reduced form regression with IV

SO2 it = β0 + β1* FDIit + β2*Domestic assetit +
β3*incomeit + γ*Xit + αi + Yeart + εit
Where,
FDIit = ρ0 + ρ1* coastali + ρ2* distance to
sea portsi + µit

All the terms are in logs
17
Empirical Strategy – Geography and Trade
policy as instruments for FDI

Challenge:


Solution: Geography and central government policy influence FDI at city
level, but not the other way around


Unobserved variables influencing openness and income at the same time,
and influencing pollution via income
Similar technique used in Wei (2001) cross-city and Frankel and Rose
(2002) cross-country studies
Instruments:


Coastal Open Policy dummies: coded 1 if the city is one of the 18 open
cities and special economic zones designated in the 1980s
Distance to Sea Ports and hubs of incoming FDI, Shanghai, HongKong,
Dalian, QinHuangDao, and Taiwan
18
Cross-sectional Analysis
Dependent Variable: LogSO2 in tons
Log FDI asset
Log Domestic asset
Log FDI employment
Log domestic employment
Log expected SO2 Emission
from FDI factories
Log expected SO2 Emission
from domestic factories
Log per capita GDP
Log per capita GDP squared
OLS
(1)
-.0314
.3847†
OLS
(2)
OLS
(3)
IV
(1)
-.6264 *
-.0221
-.0901
.7734**
IV
(2)
-.7624 *
.5750 †
-.1420**
.2571 *
6.0812 **
-.3326**
5.5370 **
-.2991**
6.0814 *
-.3283 **
6.9149 *
-.3801 **
5.3650 **
-.2940 **
*** p<0.001, ** p<0.01, * p<0.05
All regressions include a constant, city level control variables, and year and provincial dummies.
19
Fixed Effects
Dependent Variable: LogSO2 in tons
Log FDI asset
Log Domestic asset
Log FDI employment
Log domestic employment
Fixed Effect
(1)
.0152
.0367
Fixed Effect
(2)
-.0271
.1801
Fixed Effect & IV
(1)
1.6044
.6912
Fixed Effect & IV
(2)
.2003
.1152
*** p<0.001, ** p<0.01, * p<0.05
All regressions include a constant, city level control variables, and year and city fixed effects.
20
Conclusion and Caveats

Conclusion

FDI 


GDP 


shows a negative effect on SO2 factory emission
Supports Kutznet Curve, with all the cities on the left
side of the hump
Caveats

Other Pollutants: water, soil, etc

21
Future Research

Look for technology spillover effects



Within industry sectors
Between upstream-downstream sectors
Look for crowd-out effects



Within industry sectors
Export vs domestic market
Labor supply and wage increase
22