رشد اقتصادی, صنعتی سازی و محیط زیست

In this paper, I argue the compositional shift from agricultural to industrial production – industrialization – is a central determinant of changes in environmental quality as economies develop. I develop a simple two-sector model of neoclassical growth and the environment in a small open economy to examine how industrialization affects the environment. The model is estimated using sulfur emissions data for 157 countries over the period 1970–2000. The results show the process of industrialization is a significant determinant of observed changes in emissions: a 1% increase in industry's share of total output is associated with an 11.8% increase in the level of emissions per capita.

Over the past thirty years, emissions levels of key industrial pollutants have decreased in the developed world, but have increased in developing countries. This observation, known as the Environmental Kuznets Curve (or EKC), has dominated how researchers and policy makers think about the relationship between economic growth and the environment.1 While there have been many attempts to explain the EKC, existing theories have not come to grips with three other puzzling features of the data: (i) there has been a great deal of cross-country convergence in pollution emissions over time, (ii) there is substantial variation in the emission intensities (the level of emissions produced per unit of output) of industrial pollutants both over time and across countries and (iii) as a fraction of GDP, pollution abatement costs have been small and constant over time in the industrialized world. This paper provides a theory of economic growth and the environment that explains these features of the data, and offers new testable implications. Specifically, the theory predicts cross-country convergence in pollution emissions as economies industrialize. The empirical results in turn demonstrate the process of industrialization is a significant determinant of observed changes in sulfur emissions: a 1% increase in industry's share of total output is associated with an 11.8% increase in the level of emissions per capita.2 I develop a simple two-sector neoclassical model of economic growth and the environment in a small open economy in which growth is driven by a combination of capital accumulation and technological progress. The model features two goods, each of which is produced using a combination of capital and labor: a clean agricultural good, and a dirty industrial good that produces pollution as a joint output. I assume the agricultural good is consumed while the capital intensive industrial good is used in investment. I adopt a simple Solow-type framework with a fixed savings rate and abatement intensity. Technological progress in the production of goods and abatement is exogenous. In this context, the compositional shift from agricultural to industrial production as an economy grows – industrialization – drives changes in pollution levels during the transition to the balanced growth path. Development begins with rapid economic growth as capital is accumulated and this growth increases emissions in two ways. With growth, more output is produced and this increase in the scale of production causes emissions to rise. As capital becomes relatively more abundant, the composition of output shifts towards pollution intensive industrial production, leading to a further increase in pollution emissions. At the same time, improvements in the techniques of production arising from ongoing technological progress in abatement work to lower emissions. If growth is initially rapid, then compositional shifts towards industrial production overwhelm technological progress in abatement, so emissions levels rise. As development proceeds, diminishing returns to capital cause growth and compositional changes to slow. Technological progress in abatement then occurs faster than emissions growth, so emissions levels fall. Together, changes in the scale, composition and techniques of production during industrialization give rise to the EKC.3 While this interaction explains why an EKC could arise, it is important to note that the EKC is not a necessary result. Whether an EKC is observed depends on the initial capital stock and rate of technological progress in abatement; moreover, even when EKC patterns are produced, they differ across countries. This finding is consistent with the evidence; the EKC is not a robust feature of the data.4 The process of industrialization does, however, generate convergence in cross-country emissions levels during the transition to the balanced growth path. Economy-wide diminishing returns to capital cause the scale and composition effects to decrease as capital accumulates. As a result, countries that differ only in their initial capital stock will exhibit convergence in pollution emission levels; the growth rate of pollution changes faster in poor countries than in rich countries. This takes place regardless of whether pollution levels are increasing or decreasing along the balanced growth path; and arises regardless of the trade pattern. Moreover, the model tells us that convergence occurs through industrialization. There is, in fact, considerable evidence of convergence in pollution emissions over time, both within and across countries.5 As development proceeds, and more of the industrial good is produced domestically, expenditures on pollution abatement increase. However, because of diminishing returns to capital the growth rate of pollution abatement costs falls as industrialization occurs, meaning that growth rate of pollution abatement costs and income are roughly the same once an economy is industrialized. This fits with the data: the available evidence indicates that for members of the OECD, pollution abatement costs have been a small and constant fraction of GDP over time.6 To evaluate the theory, I log-linearize the model around the balanced growth path to derive an estimating equation linking emissions per capita in any period to emissions per capita in the previous period and additional controls.7 These controls include typical determinants of the balanced growth path, such as the savings rate and population growth, but also include a measure of industrialization. I formulate the estimating equation as a dynamic panel data model and estimate it using the Least Squares with Dummy Variables (LSDV) estimator suggested by Islam (1995). This approach allows me to directly estimate the effect of industrialization on pollution levels and evaluate the testable restrictions implied by the theory. I estimate the model using a unique panel data set obtained by combing data on sulfur emissions (Stern, 2006), with data on population, savings and income from the Penn World Tables (Heston et al., 2009), and data on sectoral composition from the World Bank's World Development Indicators. There are two reasons for using data on sulfur emissions. First, while sulfur emissions have been studied extensively in the context of the EKC, there is little support for an EKC type relationship in the data (Stern and Common, 2001 and Harbaugh et al., 2002). Hence, little is known about what forces are driving changes in sulfur dioxide pollution across countries. The second reason for doing so is data availability. Sulfur is one of two pollutants (carbon dioxide being the other) for which there is data on emissions for a large number of countries over a substantial period of time. The data set is an unbalanced panel and includes observations for 157 countries over the period 1970–2000.8 My main empirical results indicate that the process of industrialization is a significant determinant of observed changes in sulfur emissions. A 1% increase in industry's share of total output is associated with an 11.8% increase in the level of emissions per capita. This finding is robust to including unmodeled determinants of the balanced growth path, endogeneity caused by using a lagged dependent variable as a regressor, endogeneity caused by simultaneity, and restricting the sample to exclude outliers.9 This paper contributes to the literature on economic growth and the environment in three ways. First, it makes a theoretical contribution by developing a model of the EKC that explains other features of cross-country pollution data not considered before. Specifically, I explain why, over time: (i) there has been cross-country convergence in pollution emissions, (ii) there has been variation in emission intensities across countries, and (iii) pollution abatement costs have been a small and constant fraction of the GDP in the industrialized world. Most existing theories focus solely on explaining the inverted-U shaped relationship between income and pollution (see for example, Selden and Song, 1994, Lopez, 1994, John and Pecchenino, 1994, Stokey, 1998 and Andreoni and Levinson, 2001), but do not match other features in the data. This paper also contributes to the theoretical literature by developing a simple model of economic growth and the environment that allows for composition effects. Although it has long been recognized economic growth can affect the environment through changes in the scale, composition and techniques of production, most theories have adopted a one good framework, which eliminates the possibility of compositional effects (see for example, Brock and Taylor, 2010 or Criado et al., 2011). While one good models may be useful for studying the behavior of pollutants that are produced by all or most economic activity (such as carbon dioxide), they may not be as useful for studying the behavior of other pollutants which are mainly produced by industrial processes tied to specific sectors. In these cases, sectoral shifts brought about by development may be critical to consider. Importantly, previous two-sector models of economic growth and the environment (such as Bovenberg and Smulders, 1995 and Bovenberg and Smulders, 1996) have not explicitly examined how compositional changes affect environmental quality as an economy develops. The third contribution of this paper is empirical. This paper is the first to examine cross country convergence in sulfur emissions using a dynamic panel model derived from theory. While many authors have previously examined the cross-country sulfur emissions data (see for example, Grossman and Krueger, 1995, Stern and Common, 2001 and Harbaugh et al., 2002), this study is the first to employ an empirical approach that is tied tightly to theory. By introducing pollution into a simple neoclassical growth model, this paper bears close resemblance to the work of Brock and Taylor (2010). They develop an augmented version of the Solow model in which the EKC is generated through the interaction of scale and technique effects. In their one good model, emissions intensities decline at a constant rate over time. This means that their model is unable to explain the substantial variation in the emission intensities of industrial pollutants across time and countries, as depicted in Fig. 1 for the case of sulfur emissions. Fig. 1 plots sulfur emission intensities by income group over the period 1960–2000. Given that emission intensities are declining at a roughly constant rate for high income countries, and there is significant variation in emission intensities over time for low income countries that are currently in the process of industrializing, this figure suggests that the changes in emission intensity induced by industrialization are an important mechanism driving changes in industrial This paper is also closely related to the work of Smulders et al. (2011). They develop a simple endogenous growth model in which an EKC is generated through changes in the scale, techniques and composition of production. Unlike the model presented here, in which the composition effect arises as capital accumulates, in their work the composition effect arises through changes in technology. This means that their model is unable to explain why there has been a great deal of cross-country convergence in pollution emissions over time or why pollution abatement costs have represented a small constant fraction of GDP over time in the industrialized world. The rest of this paper proceeds as follows. Section 2 describes the model. Section 3 establishes its equilibrium and discusses pollution in the long run. Section 4 describes how the process of industrialization affects pollution. Section 5 outlines the empirical methodology and results. Section 6 concludes. Proofs to all propositions are included in Appendix A.

This paper presented a simple two-sector model of neoclassical growth in a small open economy to investigate the relationship between growth and environmental outcomes. Most existing research in this area has come through the lens of the EKC, but existing theories have not come to grips with three relevant features of the data: (i) there has been a great deal of cross-country convergence in pollution emissions over time, (ii) there is substantial variation in the emission intensities of industrial pollutants across both over time and across countries, and (iii) pollution abatement costs have been a small and constant fraction of GDP in the industrialized world. This paper provides a theory of growth and the environment that also explains these other features. The theory showed how the EKC can arise during the transition from agricultural production to industrial production as economies develop. As countries develop and accumulate capital, pollution levels increase as a result of increases in the scale of production as more output is produced, and through shifts in the composition of output towards pollution intensive industrial production. Initially these increases overwhelm the pollution reducing effect of technological progress. As development proceeds and diminishing returns to capital set in, growth and compositional changes slow; as a result technological progress in abatement occurs faster than emissions growth and emissions levels fall. Such changes in scale, composition and technique as countries industrialize generate an EKC. Although the theory showed why an EKC could arise through industrialization, it is not a necessary result. Instead the theory predicted cross country convergence in emissions as economies industrialize. To evaluate this prediction I derived an estimating equation directly from the theory by log-linearizing the model around the balanced growth path. The empirical results showed that the process of industrialization is a significant determinant of observed changes in sulfur emissions, supporting the theory's prediction: a 1% increase in industry's share of total output is associated with an 12% increase in the level of emissions per capita.