Carbon emissions cap or energy technology subsidies? Exploring the carbon reduction policy based on a multi-technology sectoral DSGE model

The study addresses how different carbon reduction policies—carbon emission cap (CEC), fossil fuel technology subsidies (FTS), and renewable energy technology subsidies (RTS)—affect both short‑term macroeconomic fluctuations and long‑term equilibria when energy technology progress is explicitly modeled. Motivated by China’s dual‑carbon goals (peak by 2030, neutrality by 2060) and the pivotal role of technological innovation, the paper seeks to clarify heterogeneous effects across policies, their interactions within a carbon trading market, and the dynamics under uncertainty. It aims to determine which policy mix most effectively reduces emissions while sustaining economic growth and how shocks (TFP, government spending, energy efficiency, and energy R&D productivity) propagate through an economy with dedicated fossil and renewable technology sectors.

Prior research has examined carbon taxes, trading, and caps, noting risks like carbon leakage and trade‑offs in welfare. Carbon taxes incentivize firm abatement but can raise social welfare losses at higher rates, while trading systems reduce emissions and promote innovation, especially in energy sectors. Synergies between carbon trading and caps can channel investment to low‑carbon technologies via price signals. Evidence on energy technology subsidies is mixed, with some studies finding growth and energy structure benefits and others noting potential negative macroeconomic effects, potentially due to differing short‑ vs long‑term impacts and subsidy intensities. DSGE literature spans RBC and New Keynesian frameworks; earlier models often treated production as homogeneous and exogenous TFP shocks as stand‑ins for technology, omitting sectoral heterogeneity and explicit energy technology dynamics. Subsequent work separated energy and non‑energy sectors, or fossil and renewable energy sectors, to analyze policy effects, but typically did not endogenize energy technology sectors’ roles in innovation, factor demands, and emissions. This paper fills these gaps by embedding fossil and renewable technology producers within an NK‑DSGE, enabling comparative assessment of cap vs subsidies under uncertainty.

The NK‑DSGE model comprises six sectors: households, intermediate goods, final goods, fossil energy technology, renewable energy technology, and government, with pollutant emissions, energy productivity, and carbon policies embedded. Households own capital, labor, and energy, derive utility from consumption and disutility from labor and energy use, face taxes on factor incomes, receive lump‑sum transfers, choose consumption, investment (with generalized quadratic adjustment costs), and bond holdings. Energy technology sectors (fossil and renewable) use Cobb‑Douglas production with capital and labor to produce technologies that augment energy efficiency in intermediate goods production; technology productivity follows AR(1) processes, and learning‑by‑doing links technology use to energy efficiency. The intermediate goods sector operates under monopolistic competition with Calvo price rigidity, uses capital, labor, fossil energy, and renewable energy, and incorporates energy technology inputs to enhance efficiency. Firms face emission costs and can abate via fossil technology; pollutant emissions depend on fossil energy use and abatement effort. Pollutant stock accumulates over time and affects labor efficiency via a damage function. The final goods sector aggregates differentiated intermediates using a CES aggregator under perfect competition. The government budget is balanced by taxes on labor, capital, fossil energy, and fees for emission permits; it purchases final goods and provides transfers, with a carbon trading mechanism and, depending on scenario, caps or subsidies. Market clearing holds and an aggregate resource constraint equates output to consumption, investment, government spending, and abatement costs. Policy scenarios: BAU (no policy), CEC (mandatory cap with permit pricing), FTS (exogenous fossil technology subsidy), and RTS (exogenous renewable technology subsidy). Calibration uses values consistent with China (e.g., β=0.97; capital depreciation 0.025; emissions per energy unit 0.6; Calvo price parameter 0.75; production elasticities for intermediate and technology sectors). Bayesian estimation identifies persistence of AR(1) shocks using China quarterly data (2000Q1–2020Q4) for output, government expenditure, and energy technology proxies based on energy patents (CPC Y02E). Shocks include: TFP, government expenditure, energy efficiency, energy technology research productivity; and policy shocks: cap stringency, fossil subsidy, renewable subsidy.

Long‑run steady state comparisons (holding total output equal across policy scenarios): • Consumption falls under carbon policies: −15.04% in CEC; −19.80% in FTS/RTS. • Factor inputs rise: labor +8.55% (CEC) and +11.73% (FTS/RTS); capital +9.55% (CEC) and +10.78% (FTS/RTS). • Renewable energy use rises: +17.69% (CEC) and +23.02% (FTS/RTS); renewable energy efficiency +28.65% (CEC) and +40.26% (FTS/RTS); renewable technology +24.83% (CEC) and +34.69% (FTS/RTS). • Fossil fuel use falls: −16.31% (CEC) and −22.82% (FTS/RTS). • Emission outcomes: pollutant emissions −27.84% (CEC) vs −33.45% (FTS/RTS); pollutant stock −27.84% (CEC) vs −33.45% (FTS/RTS). Carbon intensity Z/Y decreases from 0.648 (BAU) to 0.445 (CEC, −31.26%) and 0.411 (FTS/RTS, −36.61%). • Emission permit price is lower with subsidies than under the cap (example steady state: 0.1408 vs 0.1527). • To achieve comparable economic and environmental outcomes, required subsidy intensity is lower for fossil technology (≈0.0475) than for renewable technology (≈0.05), implying higher cost‑efficiency of fossil technology subsidies in the current structure. Short‑run dynamics under shocks: • TFP shock: output, consumption, investment, and capital rise in all scenarios; responses are most pronounced in BAU, then CEC, and weaker in subsidy scenarios. Pollutant emissions increase in BAU, FTS, RTS; remain fixed under CEC’s cap. Demand initially shifts toward fossil technology (given dominance of fossil energy), with substitution away from renewable technology. • Government expenditure shock: boosts output but crowds out private investment and capital; increases demand for renewable technologies and reduces demand for fossil technologies; emissions increase in BAU/FTS/RTS but stay fixed under CEC; CEC exhibits stronger variable responses than subsidy scenarios. • Energy technology research productivity shock: raises R&D sector productivity, reallocates factors from intermediate goods to technology sectors, leading to a decline in total output; renewable technology demand increases, fossil technology demand first rises briefly (complementarity) then falls; emissions decline (BAU/FTS/RTS) and remain fixed under CEC. • Energy efficiency shock: increases output, consumption, investment, capital, and labor; due to LBD and relative competitiveness, demand for fossil technology rises initially while renewable technology demand falls; emissions tend to rise with economic expansion in BAU/FTS/RTS and stay constant under CEC. Policy shocks: • Relaxing the emissions cap raises output, consumption, capital—and emissions (pro‑cyclical under cap policy). • Increasing fossil or renewable technology subsidies raises output, consumption, capital—while emissions fall (counter‑cyclical under subsidy policies). Overall: subsidies deliver stronger long‑run emission reductions than a cap at comparable output levels; fossil technology subsidies are more cost‑efficient; subsidy policies smooth macro‑fluctuations and emissions more than cap policy.

Embedding fossil and renewable energy technology sectors reveals heterogeneous and interacting channels through which policies affect macroeconomy and emissions. Compared to a cap alone, technology subsidies reduce emissions more at the same output level by lowering technology adoption costs, improving energy efficiency, and accelerating the shift in the energy mix. Fossil technology subsidies deliver near‑term abatement at lower fiscal cost given fossil energy’s current predominance, while renewable technology subsidies more strongly stimulate short‑term output and employment via induced sectoral activity and infrastructure investments. Under uncertainty, caps make emissions pro‑cyclical when relaxed and induce sharper macro responses; subsidy policies dampen volatility and render emissions counter‑cyclical to positive policy shocks, supporting more stable transitions. These findings support a policy package that phases from initial emphasis on fossil technology efficiency (to rapidly curb emissions) toward growing support for renewables (to optimize the energy mix and sustain growth), coordinated with a carbon trading market to provide price signals and with public investment to ease adoption frictions.

A multi‑technology sector NK‑DSGE model for China shows that carbon emission caps underperform energy technology subsidies in long‑run emission reduction at comparable output. To achieve identical targets, fossil technology subsidies require lower fiscal intensity than renewable subsidies, indicating higher cost‑efficiency in the current energy structure. In the short run, loosening caps increases both output and emissions, while subsidy shocks raise output and consumption and concurrently reduce emissions. TFP, government spending, and energy efficiency shocks stimulate macroeconomic activity, whereas energy technology research productivity shocks can temporarily lower output via factor reallocation while boosting renewable technology demand. Policymakers should pair caps with phased subsidies: prioritize fossil technology subsidies initially for efficient abatement, then shift toward renewable technology subsidies to restructure energy use and foster growth. Coordination with other macro and environmental policies and external factors (e.g., trade frictions) can maximize economic and environmental gains.

The model uses representative households; future work could incorporate heterogeneous households and labor types, especially distinguishing energy technology R&D labor from production labor. Policy interactions beyond those modeled (e.g., with monetary policy, pollution control) merit exploration. The current setup is a closed economy calibrated to China; cross‑regional and open‑economy dynamics and spillovers should be analyzed. Additional determinants of energy technology progress and emissions (e.g., international technology diffusion, sectoral trade, financial constraints) could enrich mechanisms and enhance external validity.