An analysis of the relationship between economic growth and the environment: consequences on climate change and sustainable development
- Autores: Yurena Mendoza Lemes
- Directores de la Tesis: María Amparo Camarero Olivas (codir. tes.), Javier Ordóñez Monfort (codir. tes.)
- Lectura: En la Universitat Jaume I ( España ) en 2015
- Idioma: inglés
- Tribunal Calificador de la Tesis: Andrés José Picazo Tadeo (presid.), Aurelia Bengochea Morancho (secret.), Juan Carlos Cuestas Olivares (voc.)
- Materias:
- Enlaces
- Tesis en acceso abierto en: TDX
- Resumen
General Introduction The huge increase in the world¿s greenhouse gases has led to increase the ¿threat¿ of climate change and global warming. While there may still be conflicting views on addressing the treatment of the climate change as a threat instead of as a reality, there is full consensus about the need to apply environmental protection measures for the sake of a sustainable development. However, the design of such measures is, at least, complicated due to the fact that policies will only be effective if they are global since it should be taken into account that greenhouse gases (hereinafter GHG) contribute to the environmental degradation regardless of the place where they are emitted. For this reason local, national or regional environmental policies won¿t be enough and there is a need for plans which should be as global as possible.
This feature turns the elaboration of measures to fight against climate change into a hard task, since it needs, first, to assess the necessary amount of reduction of emissions released into the atmosphere. In this first step, the complexity lies in the question of which is the target percentage that guarantees GHG¿s concentration at safe levels for sustainability, in fact there is an extensive debate on that subject among the environmental experts themselves. After this first step, in the design of international treaties individual targets should be assigned to each of the countries and this is where the most troubled point arises, since the potential candidate states to join disagree on the assignment criterion for allocating responsibilities.
One of these international treaties, and the most important one designed up to now is the Kyoto protocol, and the scheme to assign emission rights is the conflicting point among the parties, as representatives of developing countries claim that developed countries should be the ones to take the initiative in the fight against the reduction of emissions due to responsibility, equity and justice reasons. Any global program about Climate Change should be fair for all parties involved if it wants to be a successful program. In this regard, non-industrialized countries consider the current system to allocate emissions rights to be unfair, since the criterion used to design the individual targets in the Kyoto protocol was the reduction of emissions by an average of 5% for the 2008-2012 period on the basis of the levels the industrialized countries had in 1990.
This principle of allocation generates nonconformity in developing countries claiming two main reasons: the first one is that it is not fair, having in mind the developed countries are mainly responsible for the levels of concentration of gases in the atmosphere that have caused the feared climate change caused so far; the second one is that such criterion ignores the stage of development of the different countries as well as to other important criteria like population quantity of the countries that will sign that treaty.
These disagreements have caused emerging countries like China and India, which are key for the success of the mitigation of emissions plans, due to the amount of emissions they release, to be opposite to join the international treaties where developed countries are involved. These countries reject joining arguing that allocation criterion are unfair, alleging the responsibility of the industrialized countries, that are the ones that up to now have generated the high rates of pollution and therefore should be the ones to assume contractive measures of energy consumption. While the commitment of industrialized countries would help to the environmental conservation, it wouldn¿t be enough in any case, since the emerging countries in lower development stages, specialized in highly polluting sectors and with obsolete production structures, would be excluded, and the plan¿s effectiveness would be very limited.
Becoming clear that the success of the preservation policies will depend upon the comprehensiveness achieved with these, the international debate is focused on how to engage emerging economies that claim to grow without energy impositions just as the advanced economies did in the past.
Most authors of the environmental economics and sustainable development literature agree in pointing out certain key issues matters for achieving a global climate agreement such as:
- Which is the environmental performance of the developed countries?, Are their levels of emissions stable? Are they converging in their levels of emissions among the different industrialized countries? - Do the environmental conservation policies cause negative effects on the economic growth of a country? - Have the international environmental policies developed so far been effective? - Is energy consumption the critical variable for economic growth? This thesis attempts to answer these questions drawing conclusions about the behavior pattern of countries emissions, as well as from the key variables that must be taken into account to help in achieving the design of effective environmental policies.
Thus seeking to answer the first of the questions set out, Chapter II analyzes the CO2 emissions behavior of industrialized countries, 22 OECD countries, to examine if there is convergence in the levels of emissions of the different countries, as well as the stochastic convergence among them.
There is extensive literature that analyzes if the countries are converging in their CO2 levels, finding mixed results. The analysis we performed in this chapter contributes to the field of environmental economic attempting to clarify the possible sources of the mixed evidence in the convergence of CO2 emissions among the countries. Specifically, we analyzed two possible sources of inconclusive results: firstly, the fact of neglecting the step prior to convergence analysis, may be misleading for researchers since the stationary nature of the original series would negate the possibility of stochastic convergence; secondly, have not taken into account the presence of potential non-linearities in CO2 emissions series can lead to an unsuitable specification of the functional form in the analysis of these series and therefore draw the wrong conclusions.
To analyze these two hypothesis that may lead to distort the results, we relied on a detailed methodological strategy. First, we started performing linear tests proposed by Ng and Perron (2001). Next, using the [Lee2003] test, we will keep in mind the possible existence of structural changes in the series. Finally, we apply a nonlinear test within a Smooth Transition Autoregressive framework designed by [Kapetanios:2003aa]. The empirical evidence obtained by our methodological strategy suggests that the original series of CO2 emissions for the longest time span considered, from 1870 to 2006, are stationary, therefore to continue evaluating the convergence in this context would be inappropriate. However, if we consider the period 1950-2006, CO2 emissions CO2 per capita are in a non-stationary local regime. Therefore, in this case, we proceed to study convergence. Bearing in mind the possible nonlinearities, convergence in CO2 emissions are evaluated using the [Kapetanios:2003aa] test, obtaining strong evidence of divergence among the 22 OECD countries. It is surprising that after analyzing the behavior of emissions from the other OECD countries against the US as benchmark country, a strong empirical evidence is shown in favor of convergence. This result is alarming in the sense that the US is one of the leading countries in emissions which leads us to conclude that environmental performance of OECD countries is not as expected and therefore this makes it difficult to involve emerging countries in signing international treaties.
After the above results it is clear the need to impose measures that aim to stabilize the current levels of emissions release from advanced economies. However, at this point, the fear of whether said limitation of energy consumption damage or restricts the economic growth arises. This causes decision making in favor of environmental conservation involving much controversy. Therefore it is necessary to analyze to what extent there is a link between energy consumption and economic growth of countries (GDP is mostly used in the literature as proxy for economic growth). Thereby, it is capital to understand the behavior of series of CO2 emissions, energy consumption and economic growth, as these variables are critical for the effective and appropriate design of environmental policies in pursuit of sustainable development. More specifically, key information for policy decision makers in the design of effective measures against climate change may be the causal relationship between these variables. This is what we will do in Chapter III, where the relationship between the ratio of emissions and GDP in developed countries will be assessed. The methodological strategy to to analyze the above involves checking if the variables GDP and CO2 emissions are cointegrated and if it is so, to study the direction of causality of the relation CO2-GDP, that is, if energy consumption, and therefore the emissions, are causing growth, known as the growth hypothesis, or the opposite case that GDP causes emissions, called the conservation hypothesis. Alternatively the neutrality hypothesis would indicate that there is no causal link between both variables, and the feedback hypothesis postulates bidirectionality causality between both variables.
According to the prevailing hypothesis has different and important implications of economic policy which are key in the design of suitable policies to combat climate change. If evidence confirms the growth hypothesis, this means energy consumption is a critical input for growth, so policymakers are limited when implementing measure that encourage GDP, limiting, in turn, energy consumption and hence emissions. However if it is possible to verify that the conservation hypothesis exists, measures of environmental policy could be taken without adversely affecting growth, since the GDP causes the release of emissions. In the event that there is interdependence between CO2 emissions and GDP energy efficiency policies might not affect economic growth, and the ¿feedback hypothesis¿ would be verified. Finally, evidence in favor of the neutrality hypothesis states that neither a restrictive policy for energy consumption nor an expansive one would have any effect on GDP.
In Chapter III, unlike many studies studying the link between CO2 emissions and economic growth (GDP), we analyze causality between them using a Granger nonlinear methodology. For this purpose we perform the test proposed by [Skalin:1999aa] to 10 highly developed OECD countries during the 1850 to 2008 period. For these countries, the empirical evidence clearly supports the relation bidirectional Granger causality between CO2 and GDP, thus validating the "feedback hypothesis". This result warns about the design of restrictive environmental policies in the release of emission emitted since these may have negative effects on the growth path.
From the design of these environmental policies, arises the most comprehensive and important agreement on reduction of greenhouse gases existing so far, the Kyoto Protocol, which sets as objective the reduction of emissions between 2008-2012. More specifically, the countries of the EU-15 were required to emit 8% less than what they did in 1990. After this target period, before a new design of a new protocol to expand the existing one, it is interesting to know to what extent progress has been made in fighting against emissions. At this point we consider if the long term pattern of the relationship energy consumption-GDP can reveal performance regarding reduction targets set in the Kyoto Protocol, thus replying to the third question set out at the be beginning this thesis. Methodologically we rely on the concept of cointegration. Cointegration implies that the difference between the series of energy consumption and growth is stable over time and shocks do not affect these differences because they are transitory. However, if GDP shows an increasing trend but the emissions level is not proportional, the long term relationship hypothesis, this is cointegration must be rejected. Thus, countries where cointegration not exits may comply with what was signed in Kyoto at relatively lower costs than those whose GDP trends and energy use share comovements over time, that is, when both variables are cointegrated. Again, in Chapter IV of this thesis, we rely on nonlinear methodology to assess the existence of cointegration. More specifically we use the test introduced by [Kapetanios:2003aa] and extended by [Chong2008a] that not only identifies the deterministic cointegration, but also a stronger concept such as the stochastic cointegration. The result reached reveals a clear pattern: Austria, Denmark, Italy, Netherlands, Portugal and Spain should achieve greater reductions of emissions between 2009 and 2012 if they want to achieve their respective targets set in the Kyoto Protocol.
Alternatively in the Appendix to Chapter IV we consider how to engage the emerging countries to achieve agreements as global as possible setting out two important questions when designing treaties that can be seen as fair by these countries: Does economic growth of the countries depend upon energy consumption?, Could a dependence pattern be established based on the stage of development of the country analyzed?. This is, it must examine whether the series of CO2 and GDP of the countries are in equilibrium on the long term which means that any measure of energy consumption restraint designed by any agreement will have negative consequences on the energy growth. However, in this point is very important to know not only the relationships between variables in the long term but also the existence of catching up between both series since otherwise policies can be designed for a country that is not in equilibrium on the long-term that will have very damaging consequences for the growth path as despite that country not being in equilibrium the common trend of the country shows the existence of catching-up between the two variables. Methodologically speaking it is about assessing whether permanent movements in CO2 emissions of a country are associated with permanent movements in GDP, that is, it is about to discern if the stochastic common elements are important and if there is a catching up process in the levels of both series. Therefore from the point of view of convergence between CO2 and GDP we would be following the more robust version, stochastic convergence, developed by [Bernard:1995aa]. The results allow us certain classification of countries from the point of view of the degree of development achieved so far by them. For OECD economies, we found strong evidence of equilibrium on the long term while evidence of catching-up is clearer for emerging countries.
Following the analysis in Chapter IV, there is abundant literature which focuses on finding out if the energy consumption is the critical variable for economic growth. The evolution of this literature has consisted in trying to solve problems and criticism found in previous studies. One of the most common criticism is that previous studies focused on bivariate relationship between energy-consumption economic growth. Many authors attempt to solve this criticism using control variables. However the control-variables have frequently been chosen ¿ad hoc¿, following the subjective economic rationality of the authors. Our contribution to this literature is to apply a probabilistic model to select the explanatory variables between a long set of potential variables for the United States from 1949-2010, not only to perform an aggregate study but also a sector analysis. In other words we wonder, Is energy consumption which drives the path of US GDP or are there other more relevant variables?. This subject is addressed in Chapter V and we do it for USA. The choice of this country is due to mainly to three reasons: First, from the conclusions of Chapter I we can derive what already has been set out in other economic areas and that is that USA economy influences the behavior of the rest of countries; additionally, the importance of studying and analyzing one of the countries that is leading the world ranking emitting countries; thirdly and finally the access and availability of the data this country allows. The results found demonstrate the critical role of public expenditure and energy intensity when explaining the GDP. Besides, it becomes clear the importance of the sectoral study since the variables are different for each of the sectors.
Summary, Conclusions and Further Research Greenhouse gases are emitted locally around the different countries in the world, but their negative effects, resulting in the very worrying climate change, are an international problem since the fight to reduce them requires international treaties and solutions. Thus to effectively address the environmetal damage is critical to involve as many countries as possible in reducing emissions.
These treaties and agreements are designed by, for example, one of the most important institution in the fight against climate change, the United Nations Convention on Climate Change (UNFCCC), which has the primary objective of stabilizing and reducing the levels of gas emissions in the largest possible number of countries. Nonetheless, for these agreements to be as multilateral and global as they can, developing countries have to find the environment protection policies to be fair and equal. An aspect considered to be a key factor in this is the convergence in their levels of emissions among developed countries, since they are all in a similar stage of development, relatively superior to that of developing countries, which can then afford to apply emissions reduction policies through the investment in cleaner technologies and in turn specialize in less polluting economic sectors.
We analyze, in chapter II, the existence of convergence among 22 OECD countries and observe that said countries actually diverge. This fact is aggravated when studying the environmental performance of countries using the United States as a benchmark. The results show that the countries analyzed are found to converge with the US, that is, they are converging with the world leading country in emissions levels.
Two kinds of conclusions can be inferred from such study. On one hand, conclusions of economic policy, and on the other, those referring to methodology. In regards to the implications of political economy it is clear that a much stronger effort is necessary from developed countries and that the so far implemented policies have not had the expected results, as evidence shows that their emissions levels are not stable. Thus it becomes very complicated to ask emerging countries - which development is just beginning and which have the incumbent need to generate wealth in more contaminating sectors and with more polluting machines ¿ a depletion in their emissions when richer countries do not comply with these requirements.
On what concerns the methodology used in this branch of literature related to convergence, our analysis in chapter II shows the existence of several works that they are not robust on the convergence assessment among countries. We can observe in different cited articles in the previous literature how a critical previous step is neglecting in the analysis of convergence, that is, the stochastic nature of the original series of emissions. If this analysis is omitted and country's emissions and the emissions average of the studied countries are originally stationary, an assessment of convergence using this series lacks statistical significance, which in turn can lead to misleading conclusions concerning crucial policy decisions aimed at combating Climate Change.
The second key feature worthy of mention from a methodological point of view in this chapter is the criticism we offered in relation to the use of the Carlino & Mills measure to study convergence, which consists in using the average of the sampled countries as a reference. Although the study incorporates a sample of countries at a certain level of heterogeneity in their economic level and development, an average which includes countries such as Luxembourg and the United States will inevitably be characterized as a high dispersion measurement. In fact, as demonstrated in chapter II, when we apply the Carlino & Mills measure with the United States as benchmark we end up concluding that a convergence exists in regard to the United States, as US substantially bias the results. Therefore, searching for an alternative measure to analyze convergence of emissions in a time series framework, rather than the average of the studied countries, is something of great interest in a future analysis within this branch of literature.
Finally within chapter II, we introduced in detail the main contribution of this thesis, the importance of considering nonlinear methodology in the study of variables such as emissions and the GDP. Although recently different authors have used nonlinear methods to draw conclusions within environmental economics literature, examples were very limited when this dissertation was started.
Back to the political economy implications that have been highlighted throughout the chapter II, the need to develop more strict policies for environmental issues becomes significant for better results to fight global warming. However, the main problem for implementing climate change mitigation measures is the uncertainty about how these measures could affect economic growth. Therefore, it becomes critical to understand whether these variables, emissions and the GDP, are related, and if they are, which of the two causes effects over the other or if we face a relation of interdependence. The articles in this field of literature are based on the methodological concepts of causality and cointegration to provide an answer to this issue.
In chapter III, we verified that within a sample of 10 OECD countries, for 7 out of them the variables emissions and economic growth were cointegrated, so we went on to study causality between them. The empirical evidence is clear in favor of the interdependence between emissions and growth. As we have done, we will differentiate consequences in economic policy from those which are strictly methodological and of literature research.
Concerning policy implications, the interdependence shown by both variables implies that a decrease (or an increase) in the allowed emissions levels would entail a decrease (or an increase) in economic growth. Nevertheless, to assess the level of dependence between emissions and the GDP, it is necessary to carry out an individualized and detailed research of each country to understand the sign of the relation. This detailed analysis is beyond the scope of this dissertation, but that is a crucial modelization that remains outstanding in this thesis for a future research studies. If the limitation of emissions have negative effects on the growth path, policymakers would have a very critical challenge, as it implies that energy acts as a limiting factor for growth. Therefore in order to design strategies for prevention Climate Change, it must be taken into account that the policies with effects on the reduction and stabilization of emissions will also generate consequences for the GDP.
Regarding methodology, again, as we have done in chapter II, we compare the results after employing linear tests with those obtained using nonlinear methodology, remaining clear that the results depend on the chosen strategy, yet another time. Our contribution, besides the adoption of nonlinear methods, consists in overcoming the criticism in not analyzing whether the variables are cointegrated before studying the causality between them. The existence of cointegration guarantees causality in one of two ways, but not vice versa. Thus, directly analyzing causality might lead to an error interpretation, such as if the gap between the series is not stationary causality analysis lacks statistical significance.
While previously we concentrated on studying the importance of understanding the behavior of the emissions and the GDP variables in order to design effective environmental policies, in chapter IV we focused on what has been the performance on one of the most important among these policies, the Kyoto protocol. Our research was concretely centered on EU-15 countries. Base on the long term relation between energy consumption and the GDP, we will assess if they are cointegrated. Methodologically, the contribution we bring into this analysis is to identify a stronger concept of cointegration, stochastic cointegration instead of just deterministic cointegration. If the GDP shows a growing trend but the level of emissions is not proportional, the hypothesis of a long-term relation must be rejected. For these countries where cointegration does not exist, as the GDP and emissions values are not stable, accomplish with the Kyoto targets entail higher costs. This is what is reached for countries such as Austria, Denmark, Italy, the Netherlands and Portugal.
In regards of economic implications, policymakers must take into account that there are in fact certain countries that are meeting, or are on the right path to, the objectives established by Kyoto, while others, those listed before, that will have to make a greater effort to reduce emissions in order to satisfy the requirements. Subsequently, an improvement in environment conservation can be distinguished, but it is still not as close as expected to the ultimate goal, sustainable development.
The need for the modelization of both variables in future studies that was mentioned earlier can also be of great interest for the evaluation of its trends. This way, we could find, in the cases where stochastic cointegration does not exist, which variable grows disproportionately in relation to the other.
After analyzing by univariate approach the stochastic nature of emissions series in chapter II, and later in a bivariant framework in chapter III and IV, studying the relation between emissions and economic growth, finally, we considered which variables could provide information and were therefore also important in the modelization of this bivariant relationship. The main criticism in the literature referring to the GDP-emissions relation is focused on the macroeconomic nature of both variables which are, thus, conditioned by the effects of a myriad of variables that wouldn¿t be studied by employing a bivariant analysis. So far empirical proposals to overcome this criticism is based on the incorporation in the analysis GDP-emissions nexus some control variables selected by the authors subjectively.
In chapter V, we provide a probabilistic model to identify whether it is energy consumption that explains the GDP or if there are other macroeconomic variables that help us better understand this relation that we presume to be exposed to countless interactions with other economic variables. Therefor, in this dissertation, instead of analyzing what occurs with specific pre-selected variables, we statistically select from a large group of possible candidates those with the higher probability of contributing to the explanation of the GDP. Thus this analysis allows us to recognize if it is necessary and important to take other variables into account in the modelization of the GDP-emissions relationship. We do this for the United States and the study confirms what we were expecting, the critical importance of energy consumption in the explanation of the GDP. Additionally, it is very interesting to see how the public consumption and the energy intensity are two crucial variables in the American economic growth. Lastly, we must also highlight the different results in explanatory variables for each economic sectors.
These results show the arduousness that the policy makers face in order to design measures in favor of sustainable development, since to understand the effect that enviromental preventive measures have on economic growth they must not only concentrate on the consequences that a decrease in energy consumption has, but they also have to pay attention to how the GDP interacts with macroeconomic variables such as the Public Expenditure or the Energy Intensity. Furthermore, the complexity of the matter increases as we found that general measures cannot be applied within a country, but they must be specific to each economic sector.
In conclusion, the challenge that economic and environmental policymakers embark in pursuit of sustainable development is a difficult task, given the need for cooperation to be as global as possible in order for the adopted measure to truly be effective and not only simultaneously address the idiosyncratic characteristics of each of the committing country, but go a step further to design these policies in each generator of wealth sector.
Regarding methodology, environmental economics literature is relatively recent but growing very fast. Nonetheless, the limitation of existing data related to emissions as well as environmental performance, has forced some investigations to be less robust than it would be desirable statistically. This fact causes that the literature accept certain methodologies or to be somewhat permissive in the selection of methodologies used, such as, panel methods due to the limited availability of data, or even some studies skip steps in methodological procedures based on these limitations. Despite all this, we must be able to gain rigour in this kind of studies in the next decades, and produce results that could help the design of policies to fight against the very serious issue of climate change and achieve the desired sustainable development at world level.
