An assessment of different transition pathways to a green global economy

The urgent need for a green global economy to meet climate targets necessitates increased investment in renewable energy and emission mitigation from economic activities. Green industrial policies should minimize negative consequences of capital and human resource diversion from fossil fuel activities. Green growth policies, like the European Green Deal, aim to decouple economic growth from fossil fuel consumption. This paper examines the requirements of green transition pathways regarding physical and human capital resources, R&D investment differences across pathways, and the impact of transition speed on mitigation efforts. It explores how capital and labor transfers can facilitate the transition and lower GHG abatement costs. The green transition requires investment in activities and technologies reducing fossil fuel dependence, offering cleaner products/services. While full decoupling of GDP growth from resource use is debated, green growth policies can succeed if they foster mitigation. Reliance solely on renewable energy faces challenges due to spatial and temporal mismatches between resource availability and demand, potentially hindering net-zero emissions. The green transition doesn't automatically equate to GHG reductions; socioeconomic and technology-specific factors influence the long-term relationship between renewable energy expansion and emissions. Therefore, this research broadens the concept to include renewable and low-carbon energy system development, coupled with emission reduction policies from direct and indirect fossil fuel consumption. The framework investigates investment in clean energy technologies, their spillover effects on decarbonization methods, and resultant employment opportunities. Empirical evidence supports positive (but relatively small) impacts of clean technology investment and environmental regulations on employment and GHG emission reduction. A systematic review revealed a positive but small impact on net employment, a negative impact on income distribution, and that new green jobs often demand higher skills and training. Technological change driven by R&D investment in physical and human capital productivity is another key factor. Human capital development strengthens R&D capacity, and technological progress increases labor productivity. R&D investment is seen as an engine for economic growth and a facilitator of the green transition. The research focuses on the role of R&D investment allocation in determining the speed and intensity of the green transition and its implications for decarbonization.

Integrated assessment models (IAMs) of climate and the economy are widely used for analyzing climate change policies. They employ general equilibrium analysis, capturing economic dynamics from production to consumption. Models vary in aggregation levels, from global (DICE) to disaggregated regional (RICE), sectoral (IMAGE), and technological (WITCH, MIND) dimensions. A key premise is that decision-makers are rational profit/utility maximizers. IAMs are powerful tools for analyzing and optimizing climate policies, assessing long-term impacts, and tracing variable interactions. However, critiques exist; compared to input-output models, IAMs often place less emphasis on material and natural resource use, providing more detail on technological change. Input-output analyses, while often lacking economy-climate feedback loops, can offer insights into equitable outcomes of green growth policies. IAMs are favored for their ability to couple climate change dynamics with macroeconomic processes, assessing trade-offs between abatement efforts and economic impacts. Agent-based macroeconomic frameworks offer an alternative, but typically don't allow for optimal policy derivation. A simple IAM framework is deemed suitable for analyzing green transition mechanisms and trade-offs between transition pathways and optimal mitigation efforts. A key feature of IAMs is their modeling of technological change through targeted R&D investment, which reduces abatement costs and facilitates the green transition. Some prominent IAMs lack endogenous mechanisms for investment in capital and labor productivity, and the interplay between R&D, technological change, education, and productivity. For example, DICE assumes an exogenous technological change path. Several attempts have been made to address this limitation by introducing endogenous technical change, induced innovation in the energy sector (in DICE and RICE), or explicit modules of technological change and innovation diffusion through R&D investment and learning-by-doing (WITCH, MIND). These improvements substantially reduce abatement costs and highlight the importance of learning and inertia, leading to smoother abatement with higher early investment efforts. This study builds upon these, applying an endogenous model of abatement cost depending on productivity growth induced by R&D investment. While literature exists on technological change through R&D investment, less effort has been dedicated to endogenizing human capital development and labor productivity in IAM frameworks. This research aims to fill this gap.

The study modifies the 2016 version of the DICE model, a finite horizon dynamic model with 60 periods (300 years), including a representative agent model with endogenous technological growth. A two-sector economy (green and brown industries) is assumed, modeling the transition as a growing share of green output in the composite final consumption good. In each period (5 years), a central planner divides the capital stock and labor force between sectors through transfers. Capital and labor are inputs to the aggregate production function in each sector. Capital and labor productivity are determined by allocating R&D investment to each factor in each sector. The abatement rate controls carbon emissions in both sectors (assuming lower emission intensity in the green sector). Emissions accumulate, and a portion sinks into the ocean. Radiative forcing is a function of atmospheric carbon stock, global temperature is a function of radiative forcing and past temperatures, and economic and climate models are linked through a damage function. Increased global mean temperature reduces net economic output in each sector. Damages can be avoided by fostering the green transition and increasing abatement measures. The first mechanism (green sector expansion) is exogenous (transition pathways), the second (abatement) is a decision variable. Three transition pathways (Linear, Delayed, and Fast) and a Baseline pathway (constant green output share) are analyzed. The central planner aims to maximize utility from composite consumption over the time horizon. Optimal abatement rates, capital formation shares, capital and labor transfers, and R&D investment shares in capital and labor productivity are calculated for each sector. The model doesn't include material resource dynamics; results should be interpreted cautiously. The three transition pathways (Linear, Delayed, and Fast) and the Baseline pathway with a constant share of green output are defined and their characteristics presented in Table 1. The central planner maximizes utility from composite consumption over the modeling time horizon with a fixed discount rate. The resulting dynamic optimization provides the optimal abatement rate in each sector, capital formation allocation, capital and labor transfers between sectors, and R&D investment shares in each sector's capital and labor productivity. A schematic diagram of the modified DICE model is shown in Figure 1. The model does not include material resources dynamics.

The analysis presents results for three green growth scenarios (Linear, Delayed, Fast) and a Baseline scenario (no transition). Figure 2a shows the share of green consumption in total consumption. In the Linear case, this share increases steadily to 90% by 2100. The Delayed case shows zero growth initially, then accelerates to reach 90% by 2100. The Fast case reaches 90% by 2050 and remains there. Figure 2b displays the share of green output in total output. All pathways show an increase, starting from 20% in 2015 and reaching around 70% in 2100, with less difference compared to Figure 2a due to the model’s structure and optimization results. Figures 2c and 2d show the shares of capital and labor inputs in the green sector. The labor share increases slowly across pathways, while the green capital share follows the consumption patterns in Figure 2a, emphasizing the role of capital formation. Figure 3 shows abatement rates and saving rates in both sectors. The Fast pathway entails higher abatement rates in both sectors, especially in the brown sector (reaching 70% by 2100). Saving rates show fluctuations; the Delayed pathway has increased green sector saving rates initially, while the Fast pathway shows declining brown sector saving rates due to capital transfer to the green sector. Figure 4 shows capital and labor transfers between sectors. Capital and labor transfers are from brown to green sectors, with the Fast pathway showing the highest initial capital transfer. Later, the capital transfer decreases under the Fast pathway as the green sector stabilizes. The Delayed pathway shows a decrease in capital transfers due to increased green sector saving rates. Labor transfer patterns follow the overall pathway patterns. Supplementary Figure 1 presents transfer costs, abatement and climate damage costs; Supplementary Figure 2 shows sectoral economic outputs. Figure 5 displays the optimal allocation of R&D output to capital and labor productivity. All pathways require full allocation of R&D to green sector productivity, except for the initial stage of the Delayed pathway where some R&D is allocated to the brown sector. Supplementary Figure 3 shows that heavy green sector R&D investment results in higher green capital and labor productivity compared to the brown sector, and higher green capital than green labor productivity. Supplementary Note 1 and Supplementary Figures 4, 5, and 6 analyze sensitivity to transfer costs. Lowering transfer costs facilitates greater transfers, but the overall transfer patterns and saving rates remain unchanged. Figure 6 shows climate indicators. The Fast pathway results in lower and stabilizing carbon emissions. The Delayed pathway shows initially rising and then declining emissions. The Linear pathway shows similar (but without initial increase) trajectory. The end-of-century temperature change is 3.1 °C in the Delayed case and 2.7 °C in the Fast case.

The findings highlight key policy insights for the green transition. First, the main burden lies on capital formation, accumulation, and transfer, with green capital shares dramatically increasing. Second, pathways rely on different mechanisms. The Fast pathway uses capital transfers from the brown sector, while the Delayed pathway relies on increased green sector saving and endogenous capital formation. Third, the Fast pathway requires high brown sector decarbonization to limit reliance on the brown sector. The Linear pathway lies between the Fast and Delayed pathways. Regardless of speed, greening requires resource transfer and R&D investment in green sector productivity. The model's limitations include the absence of material resource dynamics, different skill levels, structural differences between sectors, and detailed analysis of abatement options and disaggregated climate impacts. Future research should address these limitations and incorporate aspects of unemployment, inequality, and upskilling. Also, the model excludes natural resources, which necessitates future improvements.

This study develops an IAM framework for analyzing green transition pathways. Results show that capital formation is crucial, with pathways relying on different mechanisms (transfers vs. endogenous formation). The Fast pathway requires high abatement, while the Delayed pathway shows higher reliance on endogenous green capital formation. Greening requires resource transfer and R&D investment. Future research should address model limitations and incorporate more detailed analysis of economic and social factors.

The model has several limitations. It excludes material resource dynamics, doesn't distinguish between skill levels in the workforce, ignores potential structural differences between the green and brown sectors, and lacks detailed analysis of different abatement options or disaggregated climate change impacts on various parts of the economic production system. Additionally, the model doesn't address issues of unemployment, inequality, social justice, and upskilling needs. Finally, the exclusion of natural capital from the analysis is a significant shortcoming that needs to be rectified in future model versions.