Transcript Folie 1

WIOD Conference, Wien, 26th – 28th May 2010
Linking
Environmental Indicators
with Economic Data
(work in progress)
Andreas Löschel
Sascha Rexhäuser
Michael Schymura
Centre for European Economic Research, Mannheim Germany
WIOD Conference, Wien, 26th – 28th May 2010
Table of Contents
1. Introduction
2. Environmental Indicators
3. Econometric Analysis
4. Additional Indicators
WIOD Conference, Wien, 26th – 28th May 2010
1. Introduction
Work Package 7 of the WIOD Project:
 Impact of globalization and structural change on environmental issues
 Linking environmental indicators with economic data,
using:
 Panel data econometrics
 Structural Decomposition Analysis (SDA)
WIOD Conference, Wien, 26th – 28th May 2010
1. Introduction
WP 7: SDA of indicators: “The change in each
indicator is broken down into changes in its
consistent parts (
 Effects of changes in production technology,
 Changes in trade patterns
 Changes in consumption patterns
 Changes in environmental efficiency at industry level ).“
Production based indicators:
 Current Presentation
Consumption based indicators:
 Future Work
(Pollution Haven/Carbon Leakage)
(Ecological Footprint Analysis)
WIOD Conference, Wien, 26th – 28th May 2010
1. Introduction
On the search of adequate indicators:
 The requirements of adequate environmental indicators crucially
depend on the central question of research, and on
 The method used.
 General requirements of environmental indicators:
 Unambiguous
 Relevant
 Measurable
 Comparability
 Free of value judgments
 see Ebert & Welsch (2004) for a more detailed discussion
WIOD Conference, Wien, 26th – 28th May 2010
2. Environmental Indicators
 Definitions
 Requirements
 Choice
WIOD Conference, Wien, 26th – 28th May 2010
2. Environmental Indicators
Definition:
 Environmental indicators quantify and simplify information on
environmental issues in order to make this information useable and
publicly known (Hammond et al. 1995)
Types of indicators:
 Pressure indicators
(i.e. pollutant emissions to the atmosphere)
 State indicators
(i.e. pollutant concentration within the atmosphere)
 Response indicators
(i.e. energy intensities)
WIOD Conference, Wien, 26th – 28th May 2010
2. Environmental Indicators
Pressure Indicators:
 We focus on pressure indicators because:
 They serve best the criterion of policy relevance of environmental
indicators and allow for a comparison of cross-country differences
(Hammond et al. 1995)
 Data are available from the WIOD input-output tables
 The data directly attribute environmental damages to the production
process
WIOD Conference, Wien, 26th – 28th May 2010
2. Environmental Indicators
Addressing the Question of Responsibility of
Environmental Pressure:
 Producer responsibility accounting
(total emissions from resident institutional units)
 Consumer responsibility accounting
(total emissions from production - exports + imports)
 Which accounting scheme is fair?
We cannot answer this question precisely without relying on strong
value judgments.
Lenzen et al. (2007) offers an overview on this debate.
WIOD Conference, Wien, 26th – 28th May 2010
2. Environmental Indicators
Examples of consumer based indicators:
 Ecological footprint, carbon footprint
 Material Flow Analysis (MFA) (ecological „rucksack“ indicators)
 Total Embodied Material Requirements (TEMR) and Total Material
Consumption (TMC)
 Life Cycle Assessment (LCA)
 Very popular in input-output frameworks
WIOD Conference, Wien, 26th – 28th May 2010
2. Environmental Indicators
Examples of producer based indicators:
 Total emissions of pollutants (per GDP, or per industry output)
 Carbonization index
 (Energy intensity)
 Popular in econometric analysis
WIOD Conference, Wien, 26th – 28th May 2010
2. Environmental Indicators
Discussion:
 The choice of adequate indicators depends on the question of research
(globalization and structural change impacts on the environment) and
on:
 The method used (SDA and panel data econometrics)
 SDA: a change in an indicator will be broken down in its consisting
parts.
In order to reach the goals of WP 7, constructed indicators like the
ecological footprint are not our preferred choice because we are
interested in the impact of changes in world trade patterns!
WIOD Conference, Wien, 26th – 28th May 2010
3. Econometric Analysis
 SDA
 Problems and Discussion
WIOD Conference, Wien, 26th – 28th May 2010
3. Econometric Analysis
Introductory Comments:
 WP 7 is interested in globalization, structural change and
environmental issues.
 The econometric implementation of the SDA allows to shed light on this
central research question.
 WIOD input-output tables provide sectoral disaggregated panel data for
all countries.
WIOD Conference, Wien, 26th – 28th May 2010
3. Econometric Analysis
Structural Decomposition Analysis
(econometric approach):
A change in an environmental pressure indicator will be decomposed in
its:
 Scale effect
 Technique effect
 Composition effect
 Trade induced composition effect
EMit  0  1INCit  2  INCit   3KLit  4 it TIit  it
2
it
 (REL.INC, REL.KL)
Source: own construction following Antweiler et al. (2001) and Cole & Elliot (2003)
WIOD Conference, Wien, 26th – 28th May 2010
3. Econometric Analysis
Scale Effect
1INCit
 How do total pollutant emissions change due to an increase in national
economic activity?
 Technique and composition effect are held constant.
 The scale effect can be measured by using a country’s GDP, income or
total output.
WIOD Conference, Wien, 26th – 28th May 2010
3. Econometric Analysis
Technique Effect
1INCit 2  INCit 
2
 How do total pollutant emissions change due to a change in production
technologies (i.e. triggered by environmental regulation)?
 Composition and scale effect are held constant.
 The technique effect is measured by most authors (Antweiler et al.
2001, Cole & Elliott 2003, and many more) by relying on the
environmental Kuznets Curve relationship as a proxy for technological
progress.
 Thus, GDP per capita is a measure for the technique effect.
 Remember that GDP is also used to measure the scale effect!
 Thus, it is difficult to distinguish between scale and technique effect
WIOD Conference, Wien, 26th – 28th May 2010
3. Econometric Analysis
Composition Effect
3KLit
 How do total pollutant emissions change due to a shift in the mix of
industries within an economy?
 Technique and scale effect are held constant.
 The composition effect can be measured by using capital-to-labor ratios
as an indicator of the mix of dirty and clean industries within an
economy.
 This requires the assumption that capital-intensive industries have
higher pollution or energy intensities.
WIOD Conference, Wien, 26th – 28th May 2010
3. Econometric Analysis
Trade Induced
Composition Effect
4  it TIit
it   (REL.INC, REL.KL)
 How do total pollutant emissions change due to a shift in the mix of
industries within an economy relative the world’s average?
 Or in other words: is there an “outsourcing” of dirty industries to other
countries (Pollution Haven Effect)?
 Technique and scale effect are held constant.
 The trade induced composition effect can be measured by using
relative capital-to-labor ratios and relative incomes.
 Why is relative income needed in addition to relative capital-to-labor
ratios?
(relative) income is a proxy-variable for environmental stringency,
further details later!
WIOD Conference, Wien, 26th – 28th May 2010
3. Econometric Analysis
Problems of this approach 1/2:
 The Kuznets Curve relationship is used here as a proxy for
environmental stringency to measure the technique effect
 The EKC method has some crucial requirements
(Antweiler et al. 2001):
 indicator (or pollutant) should be a by-product of goods production
 by-product should vary across industries
 should be subject to regulation
 well known abatement technologies for this special pollutant should
be available
 The pollutant should have strong local effects
 Thus, possible indicators are SO2 and NOx, but not CO2!
WIOD Conference, Wien, 26th – 28th May 2010
3. Econometric Analysis
Problems of this approach 2/2:
Endogeneity between:
 The volume of trade flows and economic growth (GDP)
(Frankel & Rose 2005, Frankel & Romer 1999)
 Use of instrument variables like distance, land area, etc.
 The stringency of environmental regulation and trade flows (and
political economy problems of trade policy)
(Ederington & Minier 2003)
 time dummies and/or instrument variables for unemployment, union
membership, etc.
WIOD Conference, Wien, 26th – 28th May 2010
3. Econometric Analysis
Short Summary:
 The standard empirical pollution haven approach limits the choice of
environmental indicators (local pollutants: SO2, NOx).
 Problems for employing global pollutants (EKC relationship is an
inappropriate measure of the technique effect).
 Use of instrument variables necessary.
 SDA is only useful by applying pressure indicators and the producer
accounting principle.
WIOD Conference, Wien, 26th – 28th May 2010
4. Additional Indicators
 Dynamics of Indicators
WIOD Conference, Wien, 26th – 28th May 2010
4. Additional Indicators
Mobility Indices
 Here we want to shed some light on the time dynamics of
environmental indicators.
 Appropriate indicators for this analysis are for instance response
indicators like relative energy intensities by industry or country.
 This allows us to compare different countries or industries and sheds
some light on technological progress.
WIOD Conference, Wien, 26th – 28th May 2010
4. Additional Indicators
The Shorrocks-Index
 Introduced by Shorrocks (1979)
 Needs rearranging of the data into transition matrices P (transition
probabilities) for a transition from one time period to another.
 Environmental indicators have to be classified into n groups (i.e.
relative energy intensities from 90-80%, 80-70%, …)
 The index for the amount of dynamics (not the direction) has the form:
M(P) 
n  trace(P)
n 1
 A graphical illustration can be given by using Salter-Curves (Salter
1960) or Kernel-Density-Estimators.
WIOD Conference, Wien, 26th – 28th May 2010
5. Conclusion
WIOD Conference, Wien, 26th – 28th May 2010
5. Conclusion
 Our work is focused on production based indicators.
 We pronounced the preliminary state of this paper, final results will be
presented in Amsterdam (December 2010).
 Currently no data from the WIOD project.
 SDA combined with mobility indices will give substantial insights into
the effects of economic development on environmental indicators.
WIOD Conference, Wien, 26th – 28th May 2010
Thank you for your
attention !