An assessment of different transition pathways to a green global economy

The study addresses how the global economy can transition from fossil-fueled (brown) to low-emission (green) sectors while meeting climate targets and sustaining economic growth. It asks: what physical and human capital resources are required for green transition pathways; how do R&D investments in capital and labor productivity differ across pathways; and how are mitigation efforts affected by the speed and intensity of transition. The context is the need to decouple growth from fossil fuel use to meet the SDGs and Paris Agreement, recognizing limitations of relying solely on variable renewables and the importance of mitigation and carbon removal. The paper situates the analysis within integrated assessment modeling, extending DICE to explore green growth–decarbonization interactions, capital and labor reallocation, and endogenous technological change.

The paper reviews evidence on green growth and decarbonization, noting challenges of full decoupling of GDP from resource use and the reliability constraints of high shares of solar and wind. It highlights mixed long-term links between renewable expansion and GHG reductions and the need to pair green transitions with mitigation and carbon removal. Employment effects are generally positive but modest, with potential adverse distributional impacts and higher skill requirements for new green jobs. The IAM literature is surveyed, contrasting aggregate IAMs (e.g., DICE, RICE, IMAGE, WITCH, MIND) and input–output approaches, and discussing critiques of exogenous technological change in DICE and advances with endogenous/induced innovation and learning-by-doing. The paper identifies gaps around endogenized human capital and labor productivity in IAMs, motivating a two-sector DICE extension with endogenous R&D-driven productivity and learning, and sectoral capital and labor transfers.

The authors modify DICE2016 into a two-sector (green and brown) finite-horizon dynamic IAM with 5-year time steps over 300 years (reporting results to 2100). The central planner maximizes discounted utility of a composite consumption good. Key elements: (1) Two-sector Cobb–Douglas production with capital and labor inputs, each with sector-specific, endogenously evolving productivities. Total output equals the sum of sectoral outputs. (2) Decision variables each period include: sectoral abatement rates, sectoral saving rates, bilateral capital and labor transfer rates (incurring convex transfer costs), and R&D output allocation shares between sectors and between capital vs labor productivity within each sector. (3) R&D output equals a fixed 2% of total output each period and is allocated to increase sectoral capital and labor productivity via growth functions; learning-by-doing reduces the marginal abatement cost as productivity rises. (4) Emissions are produced by both sectors with the green sector having lower emission intensity; emissions are reduced by abatement, and an exogenous component for land-use change is included. Carbon cycles and temperature follow standard DICE carbon cycle and two-box temperature dynamics; a sector-agnostic damage function (aligned to DICE) links temperature to output losses. (5) Composite consumption is a CES aggregator of green and brown consumption, with an exogenously specified share path for the brown good that defines transition pathways. (6) Calibration maintains comparability to DICE2016, including a baseline with two identical sectors, and parameterization for transfer costs, abatement costs, productivity scaling/exponents, depreciation (10%), and elasticity parameters as summarized in Table 2 of the paper. Transition pathways are defined by the exogenous trajectory of the brown share in composite consumption: Baseline (constant 80% brown/20% green to 2100), Linear (brown share declines linearly from 80% in 2015 to 10% by 2100), Delayed (80% brown until 2050, then declines to 10% by 2100), and Fast (brown share declines to 10% by 2050, then remains). The optimization yields time paths for abatement, savings, factor transfers, and R&D allocations under each pathway. Sensitivity analysis explores transfer cost parameters.

• Across all transition pathways, the main burden of the green transition falls on capital: formation, accumulation, and transfers from the brown to the green sector. The green sector’s capital input share grows dramatically (approximately tripling by 2100), while green labor share increases modestly and more homogeneously across pathways.
• All pathways require full allocation of R&D output to the green sector’s productivity (both capital and labor). Any early allocation to the brown sector occurs only briefly under the Delayed pathway and declines to zero by about 2040.
• Within the green sector, productivity gains are larger for capital than for labor, implying greater effectiveness of R&D directed at green capital productivity.
• Factor transfers: Net transfers occur from brown to green for both capital and labor; there are no transfers from green to brown. Under the Fast pathway, capital transfers are initially highest and then decline as the green sector stabilizes. Under the Delayed pathway, capital transfers decrease over time as higher green saving rates enable endogenous capital accumulation.
• Saving rates: Under the Delayed pathway, the green sector’s saving rate increases in the first half of the century (anticipating later rapid expansion), then supports growth with less reliance on transfers. Under the Fast pathway, the brown sector’s saving rate falls, reflecting capital outflows to the green sector.
• Abatement: The Fast pathway entails higher abatement rates in both sectors, particularly in the brown sector, where abatement reaches about 70% by 2100. Decarbonization of the brown sector mirrors transition speed and is strongest under Fast.
• Emissions and temperature: Total CO2 emissions are lowest and stabilize earlier under the Fast pathway. The end-of-century global mean temperature increase is about 3.1 °C in Delayed versus just below 2.7 °C in Fast. The Fast pathway delivers roughly a 0.5 °C reduction relative to the Baseline with a fixed green share.
• Sensitivity: Lower transfer costs facilitate larger transfers of the targeted factor. Reducing labor transfer costs increases labor reallocation and can reduce or eliminate the need for capital transfers by end-century, but overall qualitative patterns of transfers and saving rates remain unchanged across transfer cost scenarios.

The findings address the research questions by identifying that capital dynamics—rather than labor reallocation—dominate the requirements of green transition pathways. Early and full R&D allocation to the green sector, especially toward capital productivity, is essential to accelerate green output growth and reduce abatement costs via endogenous technological change and learning. Transition mechanisms differ by pathway: the Fast pathway achieves early green dominance through large, near-term capital transfers from the brown sector coupled with strong abatement, especially in the brown sector, to limit climate damages during the transition. The Delayed pathway relies more on higher green saving rates early on and a ramp-up of R&D-driven productivity to reduce dependence on transfers later. The Linear pathway lies between these extremes, combining gradual decline in brown capital transfers with steadily rising green saving, and still requires full R&D commitment to the green sector. These dynamics underscore a strong coupling between green growth and decarbonization: faster transitions necessitate higher near-term abatement and resource shifts, yielding earlier emissions reductions and lower end-of-century temperatures. Policy-relevant implications include prioritizing investment frameworks and institutions that enable efficient capital reallocation and sustained green-directed R&D, while managing reliance on brown-sector capacity during the early phases of rapid transitions.

The paper introduces a two-sector DICE-based IAM with endogenous R&D-driven capital and labor productivity and learning-by-doing to study green transition pathways (Linear, Delayed, Fast) against a Baseline. It shows that: (1) capital formation, accumulation, and transfers from brown to green are pivotal to the transition; (2) full and early R&D allocation to green productivity—especially capital productivity—is required across all pathways; (3) pathways rely on different mechanisms—Fast depends on large early capital transfers and high abatement, while Delayed emphasizes increased green saving and subsequent productivity-driven growth with declining reliance on transfers; (4) faster transitions reduce emissions sooner and lower temperature outcomes (≈0.5 °C below Baseline; ≈0.4 °C lower than Delayed by 2100). Future research should incorporate natural resource dynamics, heterogeneity in labor skills and sector structures, detailed abatement and energy transition options, and disaggregated climate impact channels, and address deep uncertainties through sensitivity and robustness analyses.

• The green share of consumption is exogenously imposed by pathway, not an endogenous outcome.
• The model omits material resource dynamics and natural capital, limiting analysis of material decoupling and resource constraints.
• Damage functions are sector-agnostic adaptations of DICE’s aggregate damages; more granular impact channels (labor markets, capital assets, financial systems) are not modeled.
• No differentiation of workforce skill levels or structural sectoral differences; issues of unemployment, inequality, just transition, and up-skilling are beyond scope.
• Results may reflect DICE production function features (e.g., higher capital share), potentially biasing emphasis toward capital.
• Deep uncertainties in climate–economic parameters are acknowledged; only limited sensitivity (e.g., transfer costs) is shown.
• Reliance on theoretical modeling; no empirical datasets are analyzed; model outputs are illustrative and calibrated to DICE2016 comparability.