$Ranran Wang^{123}$ $Valentina A.Assenova^4$ $Edgar G.Hertwich^{356}$

1. Institute of Environmental Sciences (CML), Faculty of Science, Leiden University, 2333 Leiden, the Netherlands
2. University of Twente, Faculty of Engineering Technology, Drienerlolaan 5, 7522 Enschede, the Netherlands
3. Yale School of the Environment, 195 Prospect Street, New Haven, CT 06511, USA
4. Management Department, The Wharton School, University of Pennsylvania, Philadelphia, PA 19104, USA
5. Department of Energy and Process Engineering, Norwegian University of Science and Technology, Kolbjørn Hejes v 1B, Trondheim 7491, Norway
   https://www.sciencedirect.com/science/article/pii/S2590332221005492?via%3Dihub
   

   Table of Contents

Highlights

• We estimate impacts of five mechanisms on nations’ CO2 emissions over 1970–2016
• Without these mechanisms, emissions grow as fast as the economy
• Energy system decarbonization was the primary mechanism for high-income countriesd
• Productivity gains were the primary mechanism for low-income countries

Science for society

Fearing negative repercussions (/ˌriːpəˈkʌʃən/)[^reper] for economic growth, climate policy is not on track to meet the Paris Agreement temperature target. However, the academic literature disagrees on both the impact of economic growth on climate mitigation and the effectiveness of various mitigation mechanisms on CO2 emissions reduction. Here, we investigate the relationship between CO2 emissions and economic growth in 73 countries during the period 1970–2016. We find that in the absence of mitigation mechanisms, emissions would have indeed grown at the same rate as the economy. However, these five mechanisms—energy system decarbonization, electrification, increased economic productivity, deindustrialization, and winter warming—are identified as successfully reducing emissions by 19 gigatonnes, mostly during periods of economic growth. Yet, observations indicate that emissions reduction rates consistent with the Paris Agreement could be achieved while maintaining economic growth only if energy systems are more rapidly decarbonized.

Summary

Nations must curtail carbon dioxide (CO2) emissions by 7% per annum to meet the Paris Agreement temperature targets. A perceived economic growth-climate mitigation trade-off has diminished political will to act. However, there is no scholarly consensus regarding the magnitude of the trade-off between economic growth and CO2 mitigation and a lack of ex post evidence regarding the extent to which mitigation measures can effectively lower CO2 emissions. Here, we present a structural equation model integrating emissions and economic and energy system characteristics over the period 1970–2016 to empirically assess mechanisms that influence the GDP-CO2 relationship for 73 countries. Robust to various model specifications and statistical tests, we found a simple unitary scale effect between per capita GDP and per capita CO2 emissions, while five emission-reduction mechanisms, principally energy system decarbonization and productivity gains, collectively contributed to global emission reductions by 19 petagrams. Within the observed year-to-year emissions development, reductions at a rate consistent with the Paris Agreement can be achieved in about 10% of instances while maintaining economic growth.

Keywords

climate change mitigation; ex post analysis; renewable energy; electrification; service economy; structural change;;panel analysis; first differences model; winter warming

Discussion

The emission-reduction mechanisms we tested and estimated are constrained by the plausible and theoretical ranges of the economic and non-economic factors. Such factors include but are not limited to the mix of low-carbon energy sources in the electrical grid, the electrification rate, and the extent of deindustrialization. Economic productivity reduces but does not entirely offset the positive unitary effect GDP growth rate has on the CO2 emission rate. The explanatory variables we tested explained less than 50% of the variance observed among countries and across time, indicating that a considerable fraction of the emission variations either remains unexplained or may be related to noise in the data. It is subject to further research. All of the variables used in this study are subject to some level of noise resulting from changes in data collection practices, reporting standards, and definitions. Such noise invariably influences the explanatory power of a model. Some of the unexplained variances may be attributable to factors not included in the model, such as various influences of the weather, regional differences in the strength of various mechanisms, structural changes not captured by the simple indicator of industry’s share of GDP, or changes in emissions not related to energy consumption, such as from cement production. Moreover, to reduce problems introduced by country heterogeneity and poor data quality, we omitted countries with short time series (<40 years) and small populations (<1 million in the year 2010) in our analysis.

Using a structural equation model in a first-difference form and a large number of observations, we showed a clear and unitary relationship between GDP and CO2 emissions. At the same time, we identified five mechanisms that explain deviations from this unitary relationship and can hence serve as potential reasons why previous empirical studies focusing on GDP as the only driver of CO2 emissions have yielded such divergent results. More importantly, the estimated emission-reduction effects of the mechanisms offer crucial empirical evidence for pursuing mitigation strategies during economic growth. The observed strong coupling between economic growth and growing CO2 emissions can be weakened by increased economic productivity and mediated by energy system decarbonization, electrification, increasing winter temperatures, and a shift from industry to services. Based on an extensive global dataset and verified by various statistical tests, our ex post analysis confirms the emission reductions from energy system decarbonization and electrification suggested by ex ante studies. Previous ex post research suggested that renewables, in particular, did not contribute to reduced use of fossil fuels and would instead increase energy consumption40 was based on smaller datasets. Our results further highlighted that shifting to renewables is about 2.5 times as effective as the coal-to-gas switching for reducing CO2 emissions. This is good news for climate change mitigation, as a shift toward low-carbon energy sources is the only one of the investigated mechanisms not limited by an upper (electrification) or lower (shift to gas, deindustrialization) bound.

Our empirical findings are broadly supportive of recent critiques of the economic growth model,51 but they also allow for a compromise position. Our results indicate that countries, such as Germany, Denmark, Finland, New Zealand, and Uruguay, have managed to achieve decoupling, i.e., reducing CO2 emissions during periods of economic growth, primarily through decarbonization of the energy system. Our model, however, suggests that a continued and timely decoupling will require further decarbonization and structural change and that, as soon as the shift toward lower-carbon energy sources stops, emissions will increase as the economy grows. While some, mostly European, countries serve as a model for how to achieve a temporary decoupling of emissions from economic growth, their rates of emission reduction have not been rapid enough to halt global warming at 1.5°C. Thus, one can understand the full significance of the empirical analysis provided with reference to humanity’s fixed emission budget to stay within the 1.5°C of the Paris Agreement on climate change. Reducing economic growth, increasing the share of services in the GDP, and electrifying the energy system are all mechanisms that can reduce the emissions while the energy system is decarbonized and hence limit the rate at which this transition needs to happen. The reverse is also true: economic growth per se and investments in construction and manufacturing, in particular, make the Paris target harder to achieve, requiring even faster decarbonization of the energy system.

The policy implications of the above results are as follows. First, continued economic growth leads to a growth in emissions from the present level; the higher the carbon intensity and the lower the productivity of an economy, the higher the emission increase resulting from an input-driven economic growth. It is hence crucial that decarbonization and productivity improvements happen first. Second, some European countries offer a successful model for decarbonizing the economy in which emissions decline while the economy still grows (such as Denmark, Finland, France, Germany, and Sweden, see Note S6)60; however, rates of decarbonization even in these leading economies need to be accelerated substantially to reach the Paris climate target. Third, while developing countries have received the economic and emission mitigation benefits from increased productivity, the historical development of their energy mixes has mostly contributed to increasing emissions. Recent trends of electrification and energy system decarbonization resulted in considerable emission reductions in the developing countries, but a dramatic upgrade in their energy system is still needed to mitigate climate change and meet the global temperature goals.

The findings have clear policy implications. As examples, in the US, the infrastructure bill likely to be approved by Congress will result in an increase in industry’s share of the GDP and help to grow the GDP while contributing little to decarbonization. To offset the emission increases resulting from the expansion of infrastructure and economic growth, additional investments in clean energy need to be implemented. In China, the continued construction boom with flats being used as a vehicle for savings leads to high emissions and makes it harder to achieve the desired peak in emissions. For developing countries, our findings highlighted the economic and climate benefits of pursuing productivity gains and suggest ending fossil fuel subsidies and increasing climate-related funding targeting clean energy investment and electrification.

While politicians tout the “green recovery” and “building back better,” a recent tally of energy-oriented expenditure in recovery packages indicates higher subsidies for fossil fuels than for clean energy.61 Support for fossil fuel production and emission-intensive industries, such as aviation and construction, will make it more challenging to reconcile future economic growth with the need to stabilize the climate. It will inadvertently strengthen the argument for a different economic model.62 Furthermore, the effect of public policy on the economic structure is often not recognized as climate relevant. However, our research clearly underlines the climate benefit of both a shift from industry toward services and an increase in productivity. The recovery is being led by growth in manufacturing and construction, while service industries are still suffering. A stimulus directed at construction and manufacturing cannot help in the required transition unless the expenditure is explicitly directed at mitigating steps, such as building refurbishment and transmission grid upgrades, needed to absorb higher shares of renewables