12.2.2 Lower emissions pathways are not necessarily associated with lower economic growth

Section 12.2.1 has demonstrated that business-as-usual futures in countries with similar characteristics can result in very different emission profiles, depending on the development path adopted. Since economic growth figures prominently among the objectives of policy-makers worldwide, the relationship between economic growth and emissions at the national level is reviewed in Section 12.2.2. Consideration is given to whether lower emissions pathways are necessarily associated with lower economic growth The conclusion that there are degrees of freedom between economic growth and GHG emissions is further explored in Section12.2.3 and Section 12.2.4.

Economic activity is a key driver of CO₂ emissions. How economic growth translates into new emissions, however, is ambiguous. On one hand, as the economy expands, demand for and supply of energy and of energy-intensive goods also increases, pushing up CO₂ emissions.. On the other hand, economic growth may drive technological change, increase efficiency and foster the development of institutions and preferences more conducive to environmental protection and emissions mitigation (see Chapter 3). Also, economic growth may be associated with specialization in sectors high) emissions per unit of output, such as services (manufacturing and heavy industries, respectively), thus resulting in a faster strong or weak relationship between domestic emissions and GDP. Unlike technological change or efficiency, however, specialization does not affect the level of global emissions: it only modifies the distribution of emissions across countries.

The balance between the scale effect of growth and the mitigating factors outlined above has generated intense scrutiny since the early 1990s. Much of the literature focuses on the ‘environmental Kuznets curve’ (EKC) hypothesis, which posits that at early stages of development, pollution per capita and GDP per capita move in the same direction. Beyond a certain income level, emissions per capita will decrease as GDP per capita increases, thus generating an inverted-U shaped relationship between GDP per capita and pollution. The EKC hypothesis is compatible with several, and possibly joint, explanations: structural shift towards low carbon-intensity sectors; increased environmental awareness with income, policy or technology thresholds; and increasing returns to abatement (Copeland and Taylor, 2004). The EKC hypothesis was initially formulated for local pollutants in the seminal analysis of Grossman and Krueger (1991) but was quickly expanded to CO₂ emissions. Even so, it recognized that some of the theoretical explanations for local pollutants, namely that higher income individuals would be more sensitive to environmental concerns, are less relevant for GHGs that do not have local environmental or health impacts. The EKC hypothesis has generated considerable research, and the field is still very active. Recent summaries can be found in Stern (2004), Copeland and Taylor (2004) or Dasgupta et al. (2004). With regard to carbon dioxide, three conclusions can be drawn, as discussed below.

First, using GDP and emissions data over multiple countries and time periods, studies consistently find that GDP per capita and emissions per capita move in the same direction among most or all of the sample (Schmalensee et al., 1998; Ravallion et al., 2000; Heil and Selden, 2001; Wagner and Müller-Fürstenberg, 2004). A 1% increase in GDP per capita is found to lead to an increase in CO₂ emissions per capita of 0.5% to 1.5%, depending on the study. All studies also find evidence that this coefficient, elasticity of per capita CO₂ emissions relative to per capita GDP, is not constant but decreases as per capita income rises. Until recently, empirical studies consistently found a relationship between per capita GDP and per capita CO₂ emissions such that, beyond a certain level of GDP per capita, per capita CO₂ emissions would decrease as income increases - thus confirming the EKC hypothesis for carbon dioxide. However, the reliability of these estimates has been challenged recently on technical grounds. For a general discussion, see Harbaugh et al. (2002) and Millimet et al. (2003); and for a critical review focusing on carbon dioxide, see Wagner and Müller-Fürstenberg (2004). Two main points emerge from the most recent reviews: (1) they cast doubt on the idea that the EKC hypothesis could be validated based on existing data; (2) they conclude that the relationship between GDP and emissions data is less robust than previously thought.

Second, studies using time series at the country level find less robust relationships between GDP per capita and CO₂ emissions per capita. For example, Moomaw and Unruh (1997) show that international oil price shocks, and not per capita GDP growth, explain most of the variations in per capita emissions in OECD countries. Similarly, Coondoo and Dinda (2002) find a strong correlation between emissions and income in developed countries and in Latin America, but a weaker correlation in Africa and Asia. Recent work on the EKC (Dasgupta et al., 2004) also shows that the relationship between GDP per capita and pollution is not as rigid as it seems, and in fact, mostly disappears when other explanatory variables, notably governance, are introduced.

Third, including trade among the explanatory variables of CO₂ emissions usually yield EKC curves peaking farther in the future (Frankel and Rose, 2002), although there are methodological issues associated with this approach (Heil and Selden, 2001). Using trade-corrected emissions data for USA, Aldy (2005) also shows that taking trade into accounts leads to curves that peak much later. Neither taking trade into account as a new explanatory variable nor correcting emissions for trade effects, however, significantly increases the robustness of the correlation between observed levels of GDP per capita and observed emission levels.

To sum up, the econometric literature on the relationship between GDP per capita and CO₂ emissions per capita does not support an optimistic interpretation of the EKC hypothesis that “the problem will take care of itself” with economic growth. The monotonically increasing relationship between economic activity and CO₂ emissions emerging from the data does not appear to be econometrically very robust, especially at country level and at higher GDP per capita level. The pessimistic interpretation of the literature findings that growth and CO₂ emissions are irrevocably linked is not supported by the data. There is apparently some degree of flexibility between economic growth and CO₂ emissions. For example, CO₂ emissions from fossil-fuel combustion in China remained essentially constant between 1997 and 2001. This was despite a +30% growth in GDP (IEA, 2004a) due to the combination of closing small-scale, inefficient power plants, shift in industry ownership away from the public sector, and introduction of energy efficiency and environmental regulation (Streets et al., 2001; Wu et al., 2005). However, these econometric studies do not distinguish between structural emissions and emissions that result from policy decisions. Thus, limited information is provided about how future policy choices may or may not influence CO₂ emissions paths. To explore these choices, a more disaggregated approach is necessary, as discussed in the following section.