Pricing indirect emissions accelerates low-carbon transition of US light vehicle sector

The study investigates how including indirect (supply chain) emissions in emissions pricing affects optimal decarbonization pathways for the US light-duty vehicle (LDV) sector. While LDVs are major contributors to energy use and CO2 emissions, policies typically target tailpipe emissions and often neglect upstream emissions from electricity generation, fuel supply, and vehicle manufacturing, which account for roughly a quarter of total LDV life-cycle emissions today. The authors hypothesize that fully pricing both tailpipe and indirect emissions could materially change relative technology costs and consumer/manufacturer choices, thereby altering the optimal transition pathway. The work addresses gaps in integrated energy models (IEMs), which often lack detailed representation of cross-sectoral indirect effects and vehicle technology-specific life-cycle impacts, by linking a detailed LCA to a national-scale energy-economy model to evaluate policy scenarios relevant to meeting US climate goals.

Prior studies and assessments indicate that reaching Paris-aligned mitigation targets is challenging and that IEMs may under-represent cross-sector interactions and material/resource dimensions (e.g., Pauliuk et al., Creutzig et al.). Earlier work on electricity systems found a limited role of indirect emissions in optimal decarbonization scenarios, whereas for vehicles, indirect emissions represent a larger share of total life-cycle emissions. Research has highlighted variability in gasoline upstream emissions and the importance of considering materials and resource efficiency in integrated scenarios, yet these aspects are often marginal in major assessments. Policies like California’s Low Carbon Fuel Standard partially regulate supply chain emissions, but no comprehensive transport policy prices all life-cycle emissions. This paper builds on and addresses these gaps by fully incorporating supply chain emissions into vehicle choice and system-wide outcomes within an integrated modeling framework.

• Modeling framework: The study soft-links the US Energy Information Administration’s National Energy Modeling System (NEMS; customized ‘Yale-NEMS’) with a detailed vehicle life cycle assessment (LCA) model. NEMS provides an integrated representation of US energy supply and demand sectors (residential, commercial, transport, industry) to 2050, with endogenous energy prices, technology deployment, and CO2 accounting across sectors. The LDV submodule includes multiple vehicle sizes, 86 efficiency technologies, and 16 propulsion options (e.g., ICEV, HEV, PHEV-10/40, BEV-100/200, HFCEV) using a discrete choice structure for consumers and production decisions for manufacturers.
• Scenarios and carbon pricing: A transport-sector carbon price starts in 2021 and increases linearly to 150 USD/tCO2 (2016$) by 2050, a level consistent with an 80% reduction in LDV direct CO2 emissions by 2050 relative to 2005 in the shared-effort framing. Main comparison: (1) pricing direct (tailpipe) emissions only; (2) pricing full life-cycle emissions (tailpipe + indirect). For simplicity, other sectors have no carbon price in main cases; side and uncertainty cases explore variations (e.g., hydrogen decarbonization, relative competitiveness of HFCEVs, pricing only well-to-wheel, high renewable/battery costs).
• Technology cost and power sector assumptions: Vehicle cost components updated bottom-up; Li-ion battery costs decline from ~465 USD/kWh (2016) to ~83 USD/kWh (floor) by 2050. Power sector includes fossil and renewable technologies with dispatch at sub-annual resolution; overnight capital costs for solar PV and onshore wind decline markedly (to ~370 and ~540 USD/kW by 2050). Resulting scenarios feature rapid grid decarbonization: renewables supply over half of electricity well before 2030 and >75% by 2050; electricity emissions fall ~88% (carbon intensity ~546 to ~45 gCO2/kWh). Methane leakage and non-CO2 GHGs are excluded.
• Soft-linking and calibration: First, Yale-NEMS outputs calibrate the LCA for US conditions, including baseline vehicle weights, lightweighting shares, on-road energy consumption, battery sizes, electricity carbon intensity, and carbon prices. The LCA computes per-vehicle life-cycle CO2 and converts to life-cycle carbon prices for each technology. These are returned to Yale-NEMS via a feebate mechanism applied to (a) vehicle production emissions and (b) expected lifetime energy chain emissions relative to a gasoline ICEV. Fees (for higher-than-ICEV emissions) or rebates (for lower) are added to purchase prices, influencing manufacturer and consumer choices. Example for BEVs: production incurs a fee (battery-intensive), but a growing rebate accrues for electricity use as the grid decarbonizes (e.g., ~400 USD/BEV in 2021 to ~2600 USD/BEV in 2050).
• Iteration and accounting: A second LCA iteration uses Yale-NEMS outputs on sales by technology/segment and energy use by carrier to compute total indirect emissions over time. Tailpipe and electricity use emissions come directly from Yale-NEMS. System boundaries include fuel production/combustion, electricity generation, material production and recycling, assembly, component reuse, and material substitution for lightweighting; vehicle production is assumed to occur in the US; minor materials are excluded.
• Uncertainty and side cases: Explore constant post-2021 battery costs; higher renewable costs; constant renewable costs with limited grid decarbonization; hydrogen decarbonization by 2050; HFCEV cost competitiveness; pricing only well-to-wheel emissions; and combinations thereof. Findings’ sensitivity to these assumptions is reported in side-case ranges.

• Technology adoption and sales: Full life-cycle emissions pricing accelerates the phase-out of ICEVs beyond tailpipe-only pricing and avoids HFCEVs entirely (due to penalties from NG-based hydrogen). Substitutions peak ~2040 with ~2.4 million additional BEV sales/year under full pricing; cumulatively, ~29 million ICEVs and ~9 million HFCEVs are avoided relative to tailpipe-only pricing.
• Emissions impacts to 2050 (cumulative differences, full vs. tailpipe-only): Total life-cycle CO2 reduced by ~1.6 Gt; tailpipe combustion reduced by ~1.4 Gt; fuel upstream (gasoline and hydrogen production) reduced by ~0.5 Gt; electricity emissions increase by ~0.25 Gt; vehicle production emissions increase by ~0.03 Gt (≈30 Mt). Net reductions are dominated by less fuel combustion and lower gasoline/hydrogen upstream emissions, with the electricity increase more than offset by reductions in the fuel supply chain.
• Robustness and ranges: Across side cases, additional cumulative life-cycle emission reductions from full pricing range from ~1.4–1.7 Gt CO2 (on top of reductions from pricing tailpipe only). Exception: in scenarios with constant renewable electricity costs (limited grid decarbonization), full pricing can yield higher emissions than pricing tailpipe-only.
• Fleet efficiency: Average real-world fleet fuel economy improves in all scenarios due to strong BEV penetration; full pricing further increases average MPG-equivalent relative to tailpipe-only pricing. Scenarios with higher HFCEV shares show lower average fuel economy.
• Resource and energy use (2050 levels, full vs. tailpipe-only): Gasoline −29% (range 26–32%); diesel −32% (30–39%); hydrogen −99.9% (−98.0 to +327.0% depending on side cases); electricity +18% (2–18%); total energy use −7% (1–6%). Materials demand slightly higher overall (+2.1%, range 0.8–2.0%), notably copper +4.7% (−0.7 to +5.0%), driven by +9.1% BEV sales (−6.1 to +9.3%).
• Cumulative quantities (to 2050, full vs. tailpipe-only): Gasoline consumption lower by ~0.15 trillion gallons (~0.6 trillion liters; ≈ one year of US gasoline use in 2020); hydrogen use lower by ~0.8 PWh; electricity use higher by ~3.2 PWh (≈ current annual US end-use electricity). Cumulative vehicle material use is ~24 Mt higher (BEV-intensive), partially mitigated by slight weight reductions; largest absolute increases: stainless steel (+13.2 Mt), copper (+3.1 Mt), aluminum (+1.6 Mt), plastics (+1.2 Mt); automotive steel and cast iron −1.0 Mt. Ambitious recycling/reuse could more than offset virgin material demand (~740 Mt potential).
• Emissions shares: Indirect emissions comprise ~24–29% of cumulative LDV life-cycle emissions (2010–2050) and ~44–49% of 2050 emissions, reflecting increased electricity and battery roles; however, total emissions are lower and indirect increases in some stages are offset by larger upstream fuel reductions.
• Carbon budget framing: Under analyzed cases, the US LDV sector uses ~3–5% of the remaining global carbon budget consistent with 1.5–2 °C pathways, roughly proportional to its current emissions share.

Including indirect (supply chain) emissions in emissions pricing materially changes the cost calculus and accelerates the shift to BEVs, yielding lower total life-cycle emissions than pricing tailpipe emissions alone. Despite higher electricity use and added battery manufacturing impacts, the reduction in gasoline production and upstream fuel emissions more than compensates, leading to simultaneous reductions in direct and indirect emissions—a notable win-win. These outcomes rely on continued grid decarbonization, which strengthens the BEV advantage. Policy implications are clear: expanding the scope of transport policies (e.g., performance standards, pricing schemes) to cover full life-cycle emissions can enhance decarbonization efficacy and align market incentives with system-wide climate benefits. HFCEVs are penalized under present hydrogen pathways; they become viable only with low-carbon hydrogen and significant cost declines. The integrated approach demonstrates that indirect emissions can substantially influence optimal decarbonization pathways in sectors where they are a large share of life-cycle impacts, such as LDVs, contrasting with prior electricity-sector-focused findings.

The paper integrates a detailed LCA with a national energy-economy model to fully price life-cycle emissions in the US LDV sector. Contrary to concerns about ‘dirty’ electricity and batteries, pricing indirect emissions accelerates BEV adoption, further reduces ICEV/HFCEV sales, and lowers both direct and indirect emissions cumulatively through 2050 when the grid decarbonizes. Resource and energy use patterns shift markedly away from gasoline/diesel and hydrogen toward electricity, with modest increases in certain materials that could be offset by recycling and reuse. The results support broadening policy scope to include supply chain emissions and confirm that large-scale BEV deployment is a no-regrets strategy under ongoing power sector decarbonization. Future research should incorporate additional processes and pathways—such as CCS at refineries, heterogeneous hydrocarbon emission intensities, synthetic fuels, net-negative energy options, low-carbon steel via renewable hydrogen, regional differences, and inclusion of pollutants beyond CO2—to further refine optimal mitigation strategies.

• Main scenarios apply carbon pricing only to the transport sector; other sectors are not priced, potentially underrepresenting economy-wide interactions.
• The LCA assumes vehicle production in the US and includes major materials while excluding minor ones; this introduces scope-related uncertainty.
• Non-CO2 GHGs (e.g., methane leakage) are not included; only CO2 is accounted for.
• Consumer-related factors such as charging infrastructure costs and discounting of future costs are not modeled, which can influence vehicle choice.
• Results hinge on continued electricity grid decarbonization and technology cost trajectories; in scenarios with constant renewable costs and limited decarbonization, full supply chain pricing can yield higher emissions than tailpipe-only pricing.
• Parameter uncertainty addressed via side and sensitivity cases (battery costs, renewable costs, hydrogen pathways), but deeper uncertainties remain regarding technology breakthroughs, policy designs, and market responses.