Download dr Jacek Batóg[1], dr Barbara Batóg Department of Econometrics

dr Jacek Batóg1, dr Barbara Batóg
Department of Econometrics and Statistics
Faculty of Economics and Management
University of Szczecin
CONVERGENCE OF RELATIVE ENVIRONMENTAL POLLUTIONS
IN THE BALTIC SEA REGION
Abstract
In this research the concept of absolute and marginal convergence was applied to
evaluate the tendencies in relative emissions of basic pollutions in the countries of the Baltic
Sea Region. The proposed measure of relative pollution is an indicator of the ecological
effectiveness of analyzed economies. An attempt was made to verify the hypothesis whether
the relative levels of environmental pollution are characterized by the same long-term
equilibrium and whether their dispersion diminishes over time. Using the new concept of
marginal vertical convergence the contribution of individual countries into the process of
overall convergence was estimated.
Key words: environmental pollution, convergence
1. Introduction
The issues related to the impact of environmental pollution and climate change are
frequently discussed in terms of contemporary social and economic processes (e.g. Ebi &
McGregor 2008, Harris 2008, Nishioka 2006). The impact of environmental factors on the
dynamics of economic growth is so big that there is a need to revise the most basic measures
1
Email: [email protected]
1
of that growth which have been used so far. According to the most recent propositions of the
European Commission, the new “green” sustainable development indicator formula, next to
the social factors, will also incorporate climate change, energy consumption, level of
biodiversity, air and water pollution, and their impact on human health, water management
and the use of natural resources (see e.g. Goossens et al. 2007, Proposal 2010, Stiglitz, Sen,
Fitoussi 2010).
The research studies into the level of pollution and other forms of degradation of the
natural environment have been based so far on absolute values of the analysed variables (EEA
2004, Kerr 2007, Alvarez, Marrero, Puch 2004). In the analyses which attempt to define
whether there is a long-term balance point describing all the countries in question in terms of
a particular type of pollution, relative measures of pollution per capita have been used
(Stegman 2005, Heil & Woden 1999, Strazicich & List 2003).
The approach recommended in this study involves the employment of relative
measures of pollution in which the denominator describes the size of the economy expressed
in its GDP. As a result, we obtain a measure defined as the amount of pollution per GDP unit,
which allows assessment of the ecological effectiveness of the analysed economy. Low values
of this measure indicate a high potential for reduction of the amount of pollution. This relative
pollution approach takes into account both the size and the ecological effectiveness of
economies in the analysis of economic processes’ impact on the degradation of the natural
environment.
The main goal of the current study was to evaluate and compare the reduction processes of
relative environmental pollution in the developed and developing countries within the Baltic
See Region. The relative approach is linked to the assumption that less developed countries
are characterized by bigger speed of environmental improvement. If such assumption is not
true we can observe a weak or even lack of income convergence among the examined
2
countries in case when the environmental factor is incorporated into the green accounting of
GDP2.
The former analyses of the pollution level convergence were based, among others, on
the distributional approach in the sense of stochastic kernel estimation which suggested that
the cross-country distribution of emissions per capita was characterized by persistence. Since
there is a lack of convergence in an absolute sense for emission per capita across countries,
projection models that generate convergence in emissions per capita are inconsistent with the
empirical behavior (Stegman 2005)3.
Some authors used also the stochastic approach (see e.g. Bernard & Durlauf 1995) to
examine convergence in emissions per capita by applying panel unit root tests (Strazicich &
List 2003) and decompositions of the Gini Index (Heil & Wodon 1999) to analyze
convergence in projected emissions per capita.
In order to verify the main hypothesis, the authors tried to assess whether the relative
levels of environmental pollution are characterized by the same long-term equilibrium and
whether their dispersion diminishes over time. Using the new concept of marginal vertical
convergence, they identified the contribution of individual countries to the process of total
convergence.
2. Methodology and sample
In the present research three kinds of convergence analysis are applied. The first one,
called σ-convergence, is an alternative to the Gini index and occurs if the dispersion of
2
The recent study of income convergence among EU countries was presented in (Batóg, Batóg 2006).
Some researchers came to quite different conclusions. For instance Criado, Valente and Stengos analyzed the
spatial distributions of per capita emissions and showed that cross-country pollution gaps have decreased over
the period for NOX and SOX within the Eastern as well as the Western European areas (Criado, Valente, Stengos
2009).
3
3
environmental pollution levels for economies tends to decrease over time. In the paper the
examination of σ-convergence is based on the estimation of a linear trend of standard
deviations of relative environmental pollution:
σ  δ0  δ1t  εt ,
(1)
where:
σ – standard deviations of logarithms of relative environmental pollution,
0, 1 – parameters of linear trend,
t – random error.
The second one, referred to as absolute -convergence, is an alternative to the
distributional and stochastic approaches, while the third one is a new concept of marginal
vertical convergence.
β–convergence exists if countries with low and high level of environmental pollution
tend to reach the same long-run equilibrium (see e.g. Aghion & Howitt 1999, Barro & Sala-iMartin 2004). This kind of convergence is described by the equation of absolute β–
convergence for discrete periods of unit length:
1  y i ,T
ln
T  y i ,0
 T


  a  (1  e

T


)
 ln( y i ,0 )  u iT ,

(2)
where:
left-hand side – the average annual growth rate of relative environmental pollution in
country i at time T,
yi0 – initial level of relative environmental pollution in country i.
The rate of β–convergence can be calculated using the formula:
 
1
 ln1  1  T  .
T
(3)
The term α1 can be obtained by rearranging formula (3) in the following manner:
4
 1  e  T 
.

 T

 1  
Therefore the value of α1 is obtained by the estimation of the equation (2).
The half-life of convergence T1/2 is the time for which the convergence process is half
way between the initial and the steady-state level and is given by formula:
T1 / 2 
ln 2

.
(4)
It should be pointed out that the absolute -convergence analysis is not robust to the
choice of the examined period, especially its initial year.
The classical convergence approaches ( and σ) do not differentiate between the speed
of convergence for individual countries. As a solution to this problem, the marginal vertical
convergence could be applied. Its idea is described by the formula:
 i     iN 1 ,
(5)
where:
βi – the value of marginal vertical convergence for country i,
β – the rate of convergence calculated for N countries,
 iN 1 – the rate of convergence calculated for N-1 countries (excluding country i).
The marginal vertical convergence allows to asses the contribution of a given country
to the process of total convergence for the whole set of examined objects.
The research concerned four kinds of relative pollutions: A – emissions of acidifying
substances (tonnes of acid equivalent per 1000 units of GDP4), B – emissions of greenhouse
gas (tonnes of CO2 equivalent per unit of GDP), C – emissions of ozone precursors (tonnes of
ozone-forming potential per 1000 units of GDP) and D – emissions of particulate matter
(tonnes of particulate-forming potential per 1000 units of GDP). The emissions of acidifying
pollutants, greenhouse gas and ozone precursors are defined according to the Convention on
4
GDP is expressed in 2007 US$ (converted to 2007 price level with updated 2005 EKS PPPs).
5
Long-range Transboundary Air Pollution (LRTAP Convention), the UNFCCC and the EU
Greenhouse Gas Monitoring Mechanism. The emissions of acidifying pollutants include
nitrogen oxides (NOx), sulphur dioxide (SO2) and ammonia (NH3). The emissions of
greenhouse gas include all the greenhouse gases covered by the Kyoto Protocol (CO2, CH4,
N2O, SF6, HFCs and PFCs). They do not include the greenhouse gases that are also ozonedepleting substances and which are controlled by the Montreal Protocol. The emissions of
ozone precursors include ground-level ozone precursor (NOx, non-methane volatile organic
compounds – NMVOCs, CO and CH4). Emissions of primary particulate matter and
secondary particulate precursors (NOx, SO2 and NH3) – PM10 and PM2.5 – are defined by
European Environment Agency (EEA) as fine solid or liquid particles added to the
atmosphere by processes at the Earth's surface. Particulate matter includes dust, smoke, soot,
pollen and soil particles.
The primary annual data of absolute levels of emissions for 8 countries from the Baltic
See Region (BSR): Denmark, Estonia, Finland, Germany, Latvia, Lithuania, Poland and
Sweden for the period 1990-2006 were collected from the Eurostat Database. GDP data come
from The Conference Board and Groningen Growth and Development Centre, Total Economy
Database, January 2008. The spatial range of data was constrained to BSR countries to assure
the homogeneity of the examined countries according to their anthroposphere characteristics.
The analyzed period was limited by the availability of the latest data on pollutions.
3. Empirical results
Time series of examined relative pollutions are presented in Figures 1-4. The data
show that all considered relative emissions of pollutions were characterized by diminishing
trends. In the early 90’s Estonia and Poland reported the highest levels of all examined
6
relative pollutions. The levels of relative pollutions in Lithuania and Latvia were smaller than
in Estonia and Poland but higher than in the remaining analyzed countries (4 old EU
members). The lowest relative pollutions were reported in Sweden and Germany. At the
beginning of examined period the dispersions of values of all kinds of pollution were much
bigger than in 2006.
Relative emission of acidifying
substances
0.0008
0.0007
0.0006
0.0005
0.0004
0.0003
0.0002
0.0001
0.0000
1990
1992
1994
1996
1998
2000
2002
2004
2006
Years
Denmark
Germany
Poland
Estonia
Latvia
Sweden
Finland
Lithuania
Average EU
Fig. 1. Relative annual emission of acidifying substances in the Baltic See Region in 19902006 [tonnes per 1000 US$ of GDP]
Source: calculations based on Eurostat and The Conference Board and Groningen Growth and Development
Centre, Total Economy Database data.
7
Relative emission of greenhouse gas
0.0030
0.0025
0.0020
0.0015
0.0010
0.0005
0.0000
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
Years
Denmark
Germany
Poland
Estonia
Latvia
Sweden
Finland
Lithuania
Average EU
Fig. 2. Relative annual emission of greenhouse gas in the Baltic See Region in 1990-2006
[tonnes per US$ of GDP]
Source: calculations based on Eurostat and The Conference Board and Groningen Growth and Development
Centre, Total Economy Database data.
Relative emission of ozone
precursors
0.014
0.012
0.010
0.008
0.006
0.004
0.002
0.000
1990
1992
1994
Denmark
Germany
Poland
1996
1998
Years
Estonia
Latvia
Sweden
2000
2002
2004
2006
Finland
Lithuania
Average EU
Fig. 3. Relative annual emission of ozone precursors in the Baltic See Region in 1990-2006
[tonnes per 1000 US$ of GDP]
Source: calculations based on Eurostat and The Conference Board and Groningen Growth and Development
Centre, Total Economy Database data.
8
Relative emission of particulate matter
0.018
0.016
0.014
0.012
0.010
0.008
0.006
0.004
0.002
0.000
1990
1992
1994
Denmark
Germany
Poland
1996
1998
Years
Estonia
Latvia
Sweden
2000
2002
2004
2006
Finland
Lithuania
Average UE
Fig. 4. Relative annual emission of particulate matter in the Baltic See Region in 1990-2006
[tonnes per 1000 US$ of GDP]
Source: calculations based on Eurostat and The Conference Board and Groningen Growth and Development
Centre, Total Economy Database data.
Meaningful conclusions could be drawn based on the observation of changes in the
components of the relative pollution i.e. in absolute levels of pollution and in the levels of
GDP. The highest absolute levels of all 4 pollutions were registered for Germany and Poland.
In the case of absolute level of pollution A the significant decrease was observed in
Germany and Poland, whereas the reduction was much slower in Estonia, Lithuania and
Latvia. In 2001-2006 the decrease of absolute pollution A in these three countries was almost
invisible. In the case of pollution B its absolute level in the last years was similar in all the
examined countries apart from Germany. Therefore, one can deduce that the decrease of
relative level of pollution B was caused by the growth of GDP. The absolute level of pollution
C hardly changed in Estonia, Latvia, Lithuania and Poland contrary to the rest of the
examined countries in which the decrease of pollution C was registered. In the case of
pollution D the absolute level of pollution did not decrease only in Latvia, Lithuania and
Poland.
9
Even in those cases where the reduction in absolute emissions of pollutions was not
observed, we can draw a conclusion that the process of GDP generating within BSR countries
becomes more environmentally friendly.
Equations 6-9 present the estimated models of absolute –convergence of the relative
emissions of pollutions A, B, C and D (t-statistics in brackets).
1  AT 
   0.230  0.018 ln A0 ,
 ln
( 4.91 ) ( 3.23 )
T
 A0 
R 2  0.635
1  BT
 ln
T  B0
R 2  0.714 β  3.86%
1  CT
 ln
T
 C0

   0.247  0.029 ln B0 ,
( 4.64 ) ( 3.87 )


   0.087  0.006 ln C0 ,
( 1.20 ) ( 0.43 )

1  DT 
   0.152  0.016 ln D0 ,
 ln
( 4.61 ) ( 2.54 )
T
 D0 
R 2  0.030
R 2  0.519
β  2.09% T 12  33.17
T12  17.94
(6)
(7)
β  0.65% T12  106.36
(8)
β  1.81% T12  38.33
(9)
It could be observed that there exists the convergence process for A, B and D
pollutions. For these three kinds of pollution all the parameters are statistically significant.
The speed of convergence varies from 1.81% for relative emission of particulate matters to
3.86% for relative emission of greenhouse gas. It means that in 20-40 years the relative
pollutions A, B and D will reach half-way to their steady-states (long-term equilibriums). The
resulting estimates of  for relative emission of pollution C do not allow to reject the lack of
convergence hypothesis.
All the above conclusions are confirmed by the results of σ-convergence analysis.
Figure 5 shows the behavior of standard deviations of examined relative pollutions (in
logarithms) in 1990-2006. The negative tendencies are evident for pollutions A, B and D over
the whole analyzed period while in case of pollution C the dispersion has started decreasing
since 1997.
10
Standard deviations of
logarithms of relative
pollutions
0.9
0.8
0.7
0.6
0.5
0.4
0.3
19901991199219931994199519961997199819992000200120022003200420052006
Years
A
B
C
D
Fig. 5. Trends of standard deviations (σ) of logarithms of relative annual pollutions A, B, C
and D
Source: calculations based on Eurostat and The Conference Board and Groningen Growth and Development
Centre, Total Economy Database data.
The convergence results reported by estimated linear trends 10, 11 and 13 demonstrate
the occurrence of diminishing dispersion for relative pollutions A, B and D (t-statistics in
brackets). The estimated models of linear trends are characterized by very high values of the
coefficient of determination and by statistically significant slopes. Only in the case of
emissions of ozone precursors (C) the convergence is not visible – see equation 12.
 A  0.82228  0.013304  t , R 2  0.898
(10)
 B  0.618586  0.015308  t , R 2  0.946
(11)
 C  0.438183  0.001685  t , R 2  0.073
(12)
 D  0.728535  0.009201 t , R 2  0.764
(13)
( 69.32 )
( 63.73 )
( 27.56 )
( 53.91 )
( 11.49 )
( 16.16 )
( 1.09 )
( 6.98 )
The results of calculation of the marginal vertical convergence for the two selected
relative pollutions A and B are presented in Tables 1-2 and Figures 6-7.
Table 1. Marginal vertical convergence for relative emission of acidifying substances (A)
Country
α1i
 iN 1
T1i/ 2 [years]
i
Δ T1i/ 2 [years]
11
-0.0203
-0.0202
-0.0184
-0.0174
-0.0168
-0.0168
-0.0160
-0.0149
Poland
Germany
Lithuania
Latvia
Denmark
Finland
Sweden
Estonia
0.0246
0.0244
0.0218
0.0203
0.0196
0.0196
0.0185
0.0171
28.19
28.37
31.83
34.10
35.33
35.34
37.57
40.63
0.0037
0.0035
0.0009
-0.0006
-0.0013
-0.0013
-0.0024
-0.0038
-4.98
-4.80
-1.34
0.93
2.16
2.17
4.40
7.46
α1i –slope of linear model (2) estimated for N-1 countries (excluding country i),
 iN 1 – rate of convergence calculated for N-1 countries (excluding country i),
T1i/ 2 – half-life of convergence calculated for N-1 countries (excluding country i),
i – value of marginal vertical convergence for country i (5),
Δ T1i/ 2 – difference between the half-life of convergence calculated for N countries and for N-1
10.0
8.0
6.0
4.0
2.0
Estonia
Sweden
Finland
Denmark
-6.0
Latvia
-4.0
Lithuania
-2.0
Germany
0.0
Poland
Changes in half-life of
convergence for relative
emission of acidifying substances
countries (excluding country i).
Countries
Fig. 6. Changes in half-life of convergence (T1/2) for relative annual emission of acidifying
substances (A) [years]
Source: calculations based on Eurostat and The Conference Board and Groningen Growth and Development
Centre, Total Economy Database data.
Table 2. Marginal vertical convergence for relative emission of greenhouse gas (B)
Country
Sweden
Poland
Germany
Latvia
Lithuania
Denmark
Finland
Estonia
α1i
-0.034
-0.032
-0.028
-0.028
-0.028
-0.027
-0.027
-0.026
 iN 1
0.0500
0.0448
0.0379
0.0377
0.0372
0.0361
0.0348
0.0328
T1i/ 2 [years]
13.88
15.46
18.30
18.36
18.63
19.22
19.92
21.15
i
0.0113
0.0062
-0.0008
-0.0009
-0.0014
-0.0026
-0.0038
-0.0059
Δ T1i/ 2 [years]
-4.06
-2.48
0.36
0.42
0.69
1.28
1.98
3.21
α1i –slope of linear model (2) estimated for N-1 countries (excluding country i),
 iN 1 – rate of convergence calculated for N-1 countries (excluding country i),
T1i/ 2 – half-life of convergence calculated for N-1 countries (excluding country i),
12
i – value of marginal vertical convergence for country i (5),
Δ T1i/ 2 – difference between the half-life of convergence calculated for N countries and for N-1
countries (excluding country i).
2.0
Estonia
Finland
Denmark
Latvia
Lithuania
-4.0
Germany
-2.0
Poland
0.0
Sweden
Changes in half-life of
convergence for relative
emission of greenhouse gas
4.0
-6.0
Countries
Fig. 7. Change in half-life of convergence (T1/2) for relative annual emission of greenhouse
gas (B) [years]
Source: calculations based on Eurostat and The Conference Board and Groningen Growth and Development
Centre, Total Economy Database data.
The above results indicate that some countries have a stronger contribution to the
whole process of convergence of relative environmental pollutions. For the relative emission
of acidifying substances, Poland and Germany tend to diminish the speed of absolute
convergence, while Estonia and Sweden are inversely affected. In the case of relative
emission of greenhouse gas a diminishing rate of convergence is observed for Sweden and
Poland, while an increasing rate is observed for Estonia and Finland.
4. Conclusions
One of the main findings of the conducted research is that the trends in the relative
emissions of pollution in the Baltic See Region are diminishing. It is important to mention
that the decrease in relative pollution levels in the new EU members was caused by the
13
growth of their GDP, while the changes in the absolute pollution levels were very small5.
These results confirm the increase in ecological effectiveness of the analysed economies.
The biggest decrease in relative pollution was observed for countries with higher
initial levels of pollution. The highest absolute values of pollutions were noticed in Germany
and Poland.
There also exists significant convergence for the three types of pollution as
demonstrated by the methodology of  and σ convergence. The convergence is equivalent to
the economies of BSR countries becoming environmentally friendly with the same degree.
This is consistent with the opinion of Rao and Riahi (Rao & Riahi 2006)6 who emphasised
that “the scenarios of long-term stabilization of the pollution levels correlated to the limit set
in various regulations assume that the levels of environmental pollution in individual
countries are convergent”.
However, the influence of individual countries on the overall speed of convergence is
different. These different speeds were demonstrated using the idea of marginal vertical
convergence. The above conclusion agrees with the empirical findings obtained by Alvarez,
Marrero and Puch who stated that heterogeneity of pollution levels among countries does not
appear to be associated with substantial differences in the production technologies or the
sources of emissions of pollutants but rather is implied by region-specific differences that can
be related to the level of economic development and to the output growth (Alvarez, Marrero,
Puch 2004).
Literature cited
5
See also results obtained by Alvarez, Marrero, Puch (2004).
6
See also discussion in (Jiang et al. 2000).
14
Aghion P, Howitt P (1999) Endogenous Growth Theory. MIT Press, Cambridge
Massachusetts
Alvarez F, Marrero GA, Puch LA (2004) Air Pollution Convergence and Economic Growth
across European Countries. Universidad Complutense de Madrid, Facultad de Ciencias
Económicas y Empresariales in its series Documentos del Instituto Complutense de Análisis
Económico with number 0406
Barro R, Sala-I-Martin X (2004) Economic Growth. MIT Press, Cambridge Massachusetts
Batóg J, Batóg B (2007) Income Convergence in the European Countries: Empirical Analysis.
In:
Folia
Oeconomica
Stetinensia,
5(13).
Wydawnictwo
Naukowe
Uniwersytetu
Szczecińskiego, Szczecin:129-142
Bernard AD, Durlauf SN (1995) Convergence in International Output. Journal of Applied
Econometrics 10(2):97-108
Criado CO, Valente S, Stengos T (2009) Growth and the pollution convergence hypothesis. A
nonparametric approach. CEPE Working Paper 66. Centre for Energy Policy and Economics.
Swiss Federal Institutes of Technology, Zurich
Ebi KL, McGregor G (2008) Climate Change, Tropospheric Ozone and Particulate Matter,
and Health Impact. Environmental Health Perspective 116(11):1449-1455
EEA (2004) Impacts of Europe's Changing Climate. An indicator-based assessment. EEA
Report 2, Copenhagen
Goossens Y, Mäkipää A, Schepelmann P, van de Sand I, Kuhndtand M, Herrndorf M (2007)
Alternative progress indicators to Gross Domestic Product (GDP) as a means towards
sustainable development. Study No. IP/A/ENVI/ST/2007-10. European Parliament, Policy
Department Economic and Scientific Policy, PE 385.672, Brussels
Harris PG (2008) Climate Change and Global Citizenship. Law and Policy October:481-501
15
Heil MT, Wodon QT (1999) Future Inequality in Carbon Dioxide Emissions and the
Projected Impact of Abatement Proposals. World Bank Policy Research Working Paper 2084
Jiang K, Morita T, Masui T, Matsuoka Y (2000) Global Long-term Greenhouse Gas
Mitigation Emission Scenarios Based on AIM. Environmental Economics and Policy Studies
3:239-254
Kerr A (2007) Serendipity is not a strategy: the impact of national climate programmes on
greenhouse-gas emissions. Area 39(4):418–430
Nishioka Y, Levy JI, Norris GA (2006) Integrating Air Pollution, Climate Change, and
Economics in a Risk-Based Life-Cycle Analysis: A Case Study of Residential Insulation.
Human and Ecological Risk Assessment 12:552–571
Proposal for a Regulation of the European Parliament and of the Council on European
environmental economic accounts (Text with EEA relevance), European Commission,
Brussels, 9.4.2010, COM(2010)132 final, 2010/0073 (COD). Accessed 25 Aug. http://eurlex.europa.eu /LexUriServ/LexUriServ.do?uri=COM:2010:0132:FIN:EN:PDF
Rao S, Riahi K (2006) The Role of Non-CO2 Greenhouse Gases in Climate Change
Mitigation: Long-term Scenarios for the 21st Century. The Energy Journal, Multi-Greenhouse
Gas Mitigation and Climate Policy Special Issue:177-200
Stegman A (2005) Convergence in Carbon Emissions Per Capita. Macquarie University,
Centre For Applied Macroeconomic Analysis, The Australian National University CAMA
Working Paper 8
Stiglitz JE, Sen A, Fitoussi JP (2010) Report by the Commission on the Measurement of
Economic Performance and Social Progress. Accessed 25 Aug. www.stiglitz-senfitoussi.fr/en/index.htm
Strazicich MC, List JA (2003) Are CO2 Emission Levels Converging Among Industrial
Countries? Environmental and Resource Economics 24:263-271
16