WORKING PAPERS
Col·lecció “DOCUMENTS DE TREBALL DEL DEPARTAMENT D’ECONOMIA”

“Relationship between technological progress, capital elasticity and emissions of industrial pollutants for the production sectors in Catalonia” Laia Pié Dols Marc Sáez
Document de treball nº -18- 2009

DEPARTAMENT D’ECONOMIA Facultat de Ciències Econòmiques i Empresarials

Edita: Departament d’Economia http://www.fcee.urv.es/departaments/economia/public_html/index.html Universitat Rovira i Virgili Facultat de Ciències Econòmiques i Empresarials Avgda. de la Universitat, 1 432004 Reus Tel. +34 977 759 811 Fax +34 977 300 661 Dirigir comentaris al Departament d’Economia.

Dipòsit Legal: T - 2117 - 2009 ISSN 1988 - 0812

DEPARTAMENT D’ECONOMIA Facultat de Ciències Econòmiques i Empresarials

Relationship between technological progress, capital elasticity and emissions of industrial pollutants for the production sectors in Catalonia
Laia Pié1, Marc Saez2,3

1 2

Department of Economics, Universitat Rovira i Virgili, Reus, Spain Research Group in Statistics, Applied Economics and Health (GRECS), University of CIBER of Epidemiology and Public Health

Girona, Spain
3

Author for correspondence: Laia Pié Department of Economics Faculty of Bussiness and Economics University Rovira i Virgili. Av. Universitat 1, 43204 Reus 34-977 759884, Fax 34-977 300661 e-mail: laia.pie@urv.cat

Abstract

As is known, the Kyoto Protocol proposes to reinforce national policies for emission reduction and, furthermore, to cooperate with other contracting parties. In this context, it would be necessary to assess these emissions, both in general and specifically, by pollutants and/or among productive sectors. The object of this paper is precisely to estimate the polluting emissions of industrial origin in Catalonia in the year 2001, in a multivariate context which explicitly allows a distinction to be made between the polluter and/or the productive sector causing this emission. Six pollutants considered, four directly related to greenhouse effect. A multi-level model, with two levels, pollutants and productive sectors, was specified. Both technological progress and elasticity of capital were introduced as random effects. Hence, it has been permitted that these coefficients vary according to one or other level. The most important finding in this paper is that elasticity of capital has been estimated as very non-elastic, with a range which varies between 0.162 (the paper industry) and 0.556 (commerce). In fact, and generally speaking, the greater capital the sector has, the less elasticity of capital has been estimated.

Key words: Kyoto protocol, multilevel model, technological progress

1

1. Introduction

In recent years societies have been consuming natural resources inappropriately, provoking serious environmental problems. At this moment in time, as we know, the main problem is climate change. Human activity, mainly due to the burning of fossil fuels, is producing artificial emissions which add carbon dioxide to the atmosphere. Thus, the “greenhouse effect”, which is actually a natural phenomenon caused by several gases present in the atmosphere and is responsible for the temperatures which makes Earth inhabitable, has escalated, producing to a greater or lesser degree, global warming. On the one hand, climate change is one of the main existing threats for sustainable development, and on the other, it represents one of the main environmental challenges affecting global economy, health and social welfare. In 1997, during the United Nations Framework Convention on Climate Change, in the Japanese city of Kyoto, the ‘Kyoto Protocol’ was signed. It entails an agreement in which industrialized countries and transition economies agree to reduce their collective emissions by 5.2% in six greenhouse gases of human origin, including carbon dioxide (CO2), methane (CH4) and nitrous oxide (N2O), as well as three industrial gases, hydrofluorocarbons (HFC), perfluorocarbons (PFCS) and sulphur hexafluoride (SF6). This reduction should take place between 2008 and 2012, taking the 1990 levels as a reference base. The European Union has had to reduce the aforementioned emissions by 8%1. In the case of Spain, it was agreed not to increase greenhouse gas emissions in the 2008-2012 period beyond 15%. In Catalonia, in accordance with the Catalan Convention for Climatic Change, a reduction of 5.33 million tons of greenhouse gas emissions has been planned for the 2008-2012 period (Department of the Environment and Housing, 2008).

1

The commitments undertaken by each Member State vary according to a series of reference parameters.

2

In order to reach these goals, the Protocol proposes, on the one hand, to reinforce or establish national policies for emission reduction (increase in energy efficiency, promotion of sustainable agricultural methods, development of renewable energy sources, etc) and, on the other hand, to cooperate with other contracting parties (exchange of experiences or information, coordination of national policies with a view to greater efficiency through cooperation mechanisms, such as the emission license, joint application mechanisms and the clean development mechanism). However, it must be taken into account that to be able to reduce polluting emissions and maintain, at the same time, a high standard of living in the society, a balance must be kept between different policies and this often implies important economic effects. In this context, it would be necessary to assess these emissions, both in general and specifically, by pollutants and/or among productive sectors. The object of this paper is precisely to estimate the polluting emissions of industrial origin in Catalonia in the year 2001, four directly related to greenhouse effect (CH4, CO, CO2, N2O) and two photochemical air pollutants (NMVOC, NOx). This is done in a multivariate context which explicitly allows a distinction to be made between the polluter and/or the productive sector causing this emission. Following a brief introduction in this section, the methods, variables and data used are described in Section 2. In Section 3, the results are shown and, finally, these are discussed and some conclusions are made in Section 4.

3

2. Methods 2.1. Data All the data was obtained from the input-output Tables for Catalonia in the year 2001, TIOC-01 (IDESCAT, 2007).2

Pollutants considered were six, five related to greenhouse effect and one (nonmethane volatile organic compounds) that corresponds to photochemical pollution. 2.2. Statistical analysis In accordance with Mizobuchi and Kakamo (2007), it can be assumed, in the first place, that the pollutant emission is a function of production,
POLi = f (Yi )

[1] of the pollutants,

where

POL

indicates

the
'

vector

POL = (CH 4 , CO, CO2 , NMVOC , NOx , N 2O ) , CH4 being methane; CO carbon monoxide;

CO2 carbon dioxide; NMVOC non-methane volatile organic compounds; NOx nitrogen oxides; N2O nitrous oxide; and the sub-index i indicates the sector (i=1,…,26, see Table 1). Following the hypothesis of Environmental Kuznets Curve, this function should not be specified in a linear way, but mainly as an inverted U (Panayotou, 2000; Strazicich and List, 2003). In a preliminary analysis, however, both this and other non-linear specifications3 had a worse adjustment than the following linear specification4:

2

Atmospheric polluting emissions of Catalonia by sectors, from the Corine-air inventory (Ministerio de Medio Ambiente), produced by Vicent Alcántara for the Project on Environmental Accounts in Catalonia. Information provided by the author. 3 In particular, restricted cubic splines with different degrees of freedom (see Harrell et al., 1988). 4 Results not shown can be supplied by the authors.

4

POLi = γYi

[2]

where γ is a vector of unknown parameters.

Following Romer’s production function (1986),
β

Yi = Ai K i

[3]

it is also assumed that there is no growth in population.5 Here A represents technological progress and K the capital.

Following several analyses, the econometric model was specified as the following multi-level model (see Saez, 2001) (see footnote 4):
log(POL )ij = log( A) j + β i log( K )ij + uij

[4]

Indicating a perturbation end with u and in which both technological progress (A), and elasticity of capital (β ) were introduced as random effects, the technological progress specific-pollutant (j being the sub-index which indicates the pollutant, and the elasticity of capital specific-sector (i being the sub-index which indicates the sector).
log( A) j = log( A) + ε j

β1i = β + ωi

where ε and ω are random perturbations with a mean of zero and constant variance.

5

This is an acceptable assumption in this paper, since transversal cuts were used.

5

The model was estimated based on restricted maximum likelihood (Patterson and Thompson, 1971; Lindstrom and Bates, 1988; Bates and Pinheiro, 1998). Estimates were carried out in the R environment of free software (version 2.6.0) (R Development Core Team, 2007). 3. Results

In Table 1 some descriptions of gross data, both of the analysed pollutant emissions and the capital of the 26 productive sectors, are shown. The pollutant with greater volume of emissions was methane, CH4 (11,044 million tons), followed by non-methane, volatile, organic compounds, NMVOC (10,040 million tons). Carbon dioxide, CO2 (1,176 million tons) and nitrous oxide, N2O (3,122 million tons) were the pollutants with fewer emissions from among those analysed in this paper. Dispersion was significant in all cases, with coefficients of variation in a range between 2.008 (carbon dioxide, CO2) and 3.557 (carbon monoxide, CO). Regarding the capital, the average capital was 2,190 million euros (the median equals 1270 million euros, first quartile equals 632, and third quartile equals 2092). Only six sectors had an above-average capital, the public services sector standing out with 12,616 million euros. Of the 20 sectors with a below-average capital, the chemistry sector (57 million euros) and, to a lesser degree, electrical equipment, electronics and optics (163 million) stand out. In Table 2 the results of the model estimates are shown. The estimator of technological progress (average) was estimated as equivalent to the emission of 5,974 thousand tons (median equals 2620 thousand tons). In NMVOC, nitrogen oxides (NOx) and nitrous oxide (N2O), above-average technical progress was estimated (see Fig. 1a), although the difference was only statistically significant in

6

the latter case (p<0.05). In the rest of pollutants, a below-average technical progress was estimated, although only in the cases of methane (p<0.05) and, marginally in the case of carbon dioxide (p<0.1), were these differences statistically significant (with respect to the median). Capital elasticity was estimated as clearly non-elastic, 0.3637. Commerce, agriculture, personal services, food, financial intermediation and construction had an elasticity of capital greater than 0.5 (Table 3). Transport and communications, homes that employ domestic staff, energy products, minerals, coke, petroleum and fuels, education and electrical energy, gas and water were the other sectors with an above-average elasticity of capital (Table 3). Among the sectors with a belowaverage elasticity of capital, the paper sector (elasticity equals 0.162) and rubber and plastic products (in this case elasticity was estimated at 0.189) stand out. However, a great variability can be observed, both in technical progress per pollutant (Fig. 1a) and in capital per sector (Fig. 1b) (see also standard error confidence interval of random errors in Table 2). Finally, in Table 4, estimations of the emissions of the pollutants per productive sector are shown. Thus, in a decreasing order of emissions, the estimated emission of NMVOC was 6772 million tons, with construction (71,378 million), Housing (20,437 million), homes that employ domestic staff (19,917 million), personal services (11,418 million) and the materials transportation industry (7,819 million) with above-average estimated emissions. The productive sectors, on average, were estimated to emit 6387 million tons of methane (CH4). It was estimated that agriculture (93,715 million tons), transport and communications (50,217 million) and construction (17,486 million) emitted much more methane than the average. An emission of 4285 tons of carbon monoxide (CO) was estimated, with paper (69,661 million), homes that employ domestic staff (11,447 million), the chemistry industry (8917 million) and hotel management (6035 million) being the sectors in which above-average emissions were estimated. An emission of 4137

7

million tons of nitrogen oxides (NOx) was estimated. In this case, the plastics industry (35,950 million), the materials transportation industry (15,885 million), electrical equipment, electronics and optics (11,539 million), education (11,136 million), hotel management (9261 million) and commerce (5541 million) were the sectors with above-average emission estimations. The estimated emissions of nitrous oxide were 2620 million tons. The above-average estimated nitrous oxide emissions were energy products, minerals, coke, petroleum and fuels (19,553 million), manufacture of wood and cork (12,423 million), social services, sanitary and veterinary activities (11,624 million), machinery (8101 million) and financial intermediation (4546 million). Carbon dioxide (CO2) emission was estimated at 1196 million tons. Sectors with above-average CO2 emissions were public services (7753 million), hotel management (7673 million), education (4646 million), food (3048 million) and electrical energy, gas and water (2194 million). The fishing industry, the metal industry, the sector of other industries and that of non-metallic mineral products were the only sectors with below-average estimates in each and every one of the pollutants. 4. Discussion

In this paper, emissions from pollutants of industrial origin in Catalonia in 2001 in a multivariate context have been estimated. In particular, a multi-level model has been used with two levels, the pollutant and the productive sector. The parameters of the model, the technical progress and the elasticity of capital have denoted the random effects. Thus, they have permitted as much one as the other, to vary according to one or other level, the technical progress according to the pollutant, and the elasticity of capital according to the productive sector. The most important finding in this paper is that elasticity of capital has been estimated as very non-elastic, with a range which varies between 0.162 (the paper industry) and 0.556 (commerce). In fact, and generally speaking, the greater

8

capital the sector has, the less elasticity of capital has been estimated. As a logical exception, the sectors related to services do not follow this general rule.6 On average, the estimated volume of emissions was, in decreasing order, NMVOC with 6772 million tons; methane (CH4), 6387 million tons; carbon monoxide (CO), 4285 million tons; nitrogen oxides (NOx), 4137 million tons; nitrous oxide (N2O), 2620 million tons; and carbon dioxide (CO2), 1196 million tons. The hotel management sector is the one which has above-average estimated emissions for a greater number of pollutants, nitrogen oxides (NOx), carbon dioxide (CO2) and carbon monoxide (CO). The above-average emissions in two of the analysed pollutants were estimated in construction (NMVOC and CH4), education (HFC and CO2); homes that employ domestic staff (NMVOC and CO), the materials transportation industry (NMVOC and NOx). The rest of the sectors either exceed the estimated emissions in one of the pollutants or the estimated volume of emissions is below-average in all cases, as in the fishing industry, the metal industry, the sector of other industries, and that of other non-metallic mineral products. Take note that in the services sectors a quite high volume of emissions of nitrogen oxides (NOx), carbon dioxide (CO2) and carbon monoxide (CO) (in that order) was estimated. However, in the industrial sectors, high emissions were estimated in all pollutants, the predominance of one or another depending on capital intensity. For example, for the sector of energy products, minerals, coke, petroleum and fuels, with a capital of 1047 million euros, an emission of 19,533 million tons N2O was estimated. However, in industrial sectors with greater capital, the metal industry, the sector of other industries and the non-metallic mineral products industry, the estimated emission never exceeded the average.

6

Thus, see other services and social activities with a capital of 5,079 million euros and an elasticity of capital of 0.542.

9

Acknowledgements This study has been supported by the Spanish Ministry of Education and Culture (Grant SEJ2007-66318).

Referentes Alcántara-Escolano V, Padilla E. Subsistemas input-output y contaminación: una aplicación al sector servicios y las emisiones de CO en España. II Jornadas
2

de Análisis Input-Output, Zaragoza 5-7 de Septiembre de 2007. Bates DM, Pinheiro JC. Computational methods for multilevel models. University of Wisconsin, 1998 [citado el 24 de Enero de 2008]. Disponible en: http://franz.stat.wisc.edu/pub/NLME/ Departament de Medi Ambient i Habitatge. Convenció Catalana del Canvi Climàtic. [citado html Harrell FE, Lee KL, Pollock BG. Regression models in clinical studies: determining relationships between predictors and response. Journal of National Cancer Institutes 1988; 80(15):1198-1202. IDESCAT. Taules input-output per a Catalunya 2001. Barcelona: Institut d’Estadística de Catalunya, 2007. [citado el 15 de Febrero de 2008]. Disponible en: http://www.idescat.net/Cataleg/?tc=c&idp=70&lang=es Lindstrom MJ, Bates DM. Newton-Raphson and EM Algorithms for Linear MixedEffects Models for Repeated-Measures Data. Journal of the American Statistical Association 1988; 83:1014-1022. Mizobuchi K, Kakamu K. Simulation studies on the CO2 emission reduction efficiency in spatial econometrics: a case of Japan. Economics Bulletin 2007; 18(4):1-9. el 29 de Febrero de 2008]. Disponible en: http://mediambient.gencat.net/cat/el_medi/C_climatic/occc/html/index_occc.

10

Panayotou T. Economic growth and the environment. CID Working Paper 56, 2000. Patterson HD, Thompson R. Recovery of inter-block information when block sizes are unequal. Biometrika 1971; 58:545-554. R Development Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing, 2007. ISBN 3-900051-07-0 [citado el 24 de Enero de 2008]. Disponible en: http://www.R-project.org. Romer PM. Increasing returns and long-run growth. Journal of Political Economy 1986; 94:1002-1037. Saez M. El problema de las medidas repetidas. Análisis longitudinal en epidemiología. Gaceta Sanitaria 2001; 15(4):347-352. Strazicich MC, List JA. Are CO2 emission levels converging among industrial countries? Environmental and Resource Economics 2003; 24:263-271.

11

Table 1.- Description of emissions and capital of the productive sectors. Emissionsa
Pollutants Methane, CH4 Carbon monoxide, CO Carbon dioxide, CO2 Non methane volatile organic compounds, NMVOC Nitrogen oxides, NOx Nitrous oxide, N2O
a

Mean 11,044.2 5561.7 1175.6 10,040.3 4747.8 3122.2

Std. deviation 36,143.4 19,784.0 2361.1 22,717.8 10,022.7 6597.5

Median 21.5 214.0 176.5 2688.5 981.0 444.5

Q1 (25 ) 8.0 83.0 86.0 37.0 247.0 243.0

th

Q3 (75 ) 211.0 708.0 492.0 8,788.0 2314.0 1049.0

th

Thousands of tons

Capital
Productive sectors Agriculture Chemistry Commerce Construction Education Electrical energy, gas and water Electrical equipment, electronics and optics Energy products, minerals; coke, petroleum and fuels Financial intermediation Fishing Food Homes that employ domestic staff Hotel management Machinery Manufacture of transport material Manufacture of wood and cork Metal Other industries Other non-metallic mineral products Other services and social activities; personal Paper Public services Real estate activities and entrepreneurial services Rubber and plastic products Sanitary and veterinary activities; social services Transport and communications Thousands of Euros 1,969,524 56,608 307,468 1,189,201 1,730,576 991,339 162,784 1,046,972 2,092,418 631,507 769,894 1,448,833 992,051 1,057,297 1,349,838 277,352 3,991,980 7,929,001 4,095,448 5,079,222 2,738,336 12,615,685 524,738 569,510 1,415,289 189.499,1

12

Table 2.- Results of the estimation of the model
Coefficients Fixed effects Technical progress a Capital Random effects Technical progress (per pollutant) a Capital (per sector) Standard error of the model AIC BIC
a

5.97400 0.36370 Standard error IC 95%
1.61954 0.12990 0.71349 681.07 696.25 0.02656-98.7487 0.09361-0.18027

Elasticity

13

Table 3.- Elasticity of capital per productive sector
Productive sectors Commerce Agriculture Other services and social activities; personal Food Financial intermediation Construction Transport and communications Homes that employ domestic staff Energy products, minerals; coke, petroleum and fuels Education Electrical energy, gas and water Estimated elasticity (mean) Manufacture of wood and cork Manufacture of transport material Other industries Fishing Chemistry Electrical equipment, electronics and optics Metal Machinery Hotel management Public services Other non-metallic mineral products Real estate activities and entrepreneurial services Sanitary and veterinary activities; social services Rubber and plastic products Paper
Descending order

Elasticity 0.5562694 0.5427474 0.5423769 0.5362701 0.5131578 0.5053491 0.4435312 0.4261068 0.4019234 0.3986370 0.3917263 0.3637000 0.3599079 0.3491512 0.3394910 0.3376152 0.3202694 0.3142722 0.3091455 0.3046418 0.3023085 0.2516396 0.2269864 0.2243526 0.2070013 0.1894189 0.1620125

14

Table 4.- Estimated emissions of pollutants per productive sector
Productive sectors Agriculture Chemistry Commerce Construction Education Electrical energy, gas and water Electrical equipment, electronics and optics Energy products, minerals; coke, petroleum and fuels Financial intermediation Fishing Food Homes that employ domestic staff Hotel management Machinery Manufacture of transport material Manufacture of wood and cork Metal Other industries Other non-metallic mineral products Other services and social activities; personal Paper Public services Real estate activities and entrepreneurial services Rubber and plastic products Sanitary and veterinary activities; social services Transport and communications Averages NMVOC
49.072 397.150 --71377.581 41.859 2238.999 5321.881 2525.730 4916.215 229.481 5717.271 19917.228 901.596 1968.795 7819.121 1045.744 803.675 5491.003 2507.950 11417.670 4130.298 12.710 20437.085 4.267 26.448 7.161

CH4
93715.159 6.496 518.592 17486.398 90.030 51.825 12.122 63.822 1701.077 24.421 331.622 64.680 14.284 17.040 28.853 17.848 9.509 49.059 16.173 1561.734 5.430 33.282 9.492 5.527 9.457 50217.401

CO
--8917.250 82.344 3262.421 696.496 711.295 435.358 78.106 492.164 3303.621 153.781 11446.774 6035.025 91.805 102.748 175.095 165.368 393.357 441.208 101.843 69660.501 30.167 206.459 59.775 35.610 56.368

NO
4.343 7.135 5540.822 --11135.905 890.065 11538.887 3529.517 1917.707 1169.132 214.323 1342.715 9261.090 421.498 15885.222 1222.834 251.880 281.370 480.954 439.921 1010.327 677.820 73.363 35949.702 21.771 148.407

N2O
42.565 25.644 40.707 1279.272 --608.663 75.643 19553.207 4545.701 1341.307 768.603 192.297 964.083 8101.442 376.903 12423.178 990.281 224.939 251.696 429.870 254.737 29.085 928.349 332.106 11623.528 98.921

CO2
283.953 --952.117 87.300 4645.969 2194.137 524.413 305.978 70.970 390.006 3047.672 138.425 7673.390 542.891 87.150 95.372 185.443 117.127 143.702 284.631 88.234 7753.476 26.036 179.505 51.720 31.533

6772

6387

4285

4137

2620

1196

Thousands of tons higher than the average are highlighted in bold.

15

Figure 1.- Estimations of random effects
Figure 1a.- Technical progress per pollutant
1

log of Technical progress

-2

-1

0

1

2

3

4

CH4

CO

CO2 Pollutants

COVMN

HFC

N2O

1

log (technical progress)

Figure 1b.- Elasticity of capital per sector

Elasticity of capital 0.2 Agriculture 0.3

0.4

0.5

Education

Fishing

Machinery Productive sector

Other industries

Public services

Textile

16