OEO Blog: Harmonizing energy models for better climate policy

Energy systems models are increasingly guiding U.S. energy policy. Analysts use these models to explore which energy sources, technologies, and policies can meet our evolving energy demands. However, each model is built independently by different teams—often behind closed doors and using different assumptions and modeling approaches. This raises a critical challenge: how should policymakers interpret model results when those results differ? Distinguishing between genuine differences in expected policy outcomes and those related to varying model assumptions is essential for robust policy decisions.

Over the last two years, four leading research teams set out to harmonize their electric power system model assumptions to produce more robust policy support. Teams harmonized model results for two scenarios: one that reflects only current policies and one that constrains power sector emissions to an economy-wide net-zero emissions scenario. For these scenarios, models were configured with the same numerical inputs. For example, each team assumed U.S. electricity demands would increase by 72% between 2027 and 2050, consistent with strong electrification. They also made similar assumptions related to the numerical values of parameters—such as technology costs and the economic discount rate—as well as the alternative technologies available to meet growing electricity demands. 

Teams estimated that building and operating a U.S. energy system under current policy would cost about $3.73 trillion (in net present value) and emit about 26.3 billion tonnes of CO₂ between 2027 and 2050. In contrast, a net-zero energy system would cost $5.15 trillion and emit about 4.46 billion tonnes of CO₂ over this period, declining to about 110 million tonnes per year in 2050. Why does a net-zero energy system still have positive emissions? In a net-zero scenario, emissions from some hard-to-abate activities, like those in heavy industry, are offset by  negative emission technologies like carbon dioxide removal.  

Once the models were harmonized, the teams were ready to experiment with different decarbonization policies and model configurations, where they evaluated alternatives against the costs (or net present value) and cumulative emissions associated with current policy over the planning horizon of 2027 to 2050.

Carbon Buyout Prices

Teams configured a "carbon buyout price" on emissions exceeding the net-zero limit. The buyout price acts similarly to a carbon offset but, instead of being voluntary, is required if emissions exceed a given limit. 

• With a carbon buyout price of $50/ton, models showed a clear—but short-lived—decrease in emissions by 2030, followed by a steady rebound as natural gas generators ramped up to meet growing electricity demand. Relative to current policies, a $50/ton buyout price reduced emissions by about 44% and increased costs by about 24%.  
• At a higher buyout price of $200/ton, cumulative emissions dropped about 83% below current policy projections, responding to the higher penalty for fossil fuel use. Costs increased by about 38%. 

• At a very high carbon buyout price of $1,000/ton, the system decarbonized even more, reducing cumulative emissions by 85% and 2050 emissions by 98% driven by increased adoption of wind, solar, and batteries. System costs increased by 44%. 

These results, consistent across the four models, highlight a crucial insight for policymakers: mandating emissions limits alone may be insufficient unless the penalty for exceeding those limits (the buyout price) is set sufficiently high. Achieving sustained, deep decarbonization appears contingent upon substantially higher emissions penalties than currently typical in policy discussions. To achieve the deep decarbonization the world needs, penalties should be close to the social cost of carbon, currently estimated to be around $200/ton. 

Transmission Constraints

The research teams also explored how constraining transmission expansion between regions affected emissions and costs. Relative to the current policy scenario, allowing unlimited transmission expansion (roughly 3x in practice) reduced emissions by about 83% between 2027 and 2050. Preventing any transmission expansion increased cumulative emissions slightly, resulting in a net reduction of 79%. Why? Without the ability to move renewable energy across regions, models relied more heavily on local fossil-fueled generation, increasing emissions and also slightly increasing costs. Relative to current policies, decarbonizing with unlimited transmission expansion versus no transmission system expansion increased costs by 38% or 43%, respectively. The results suggest that allowing transmission expansion across regions can slightly reduce the costs of decarbonization where local substitutes constrain the scale of impact. Again, the fact that all four models produced similar results makes this policy insight more robust.

Carbon Capture Constraints

What would happen if CCS were completely removed as an available technology? The models showed that in a net-zero case without CCS, cumulative emissions decreased by approximately 80% relative to the current policy scenario, which is slightly fewer reductions than the base net-zero case (83%). Additionally, costs increased by about 47% relative to current policies, which was a higher increase than when CCS was allowed (38%). Without CCS as an option, the models leaned much more on renewables and battery storage to meet emissions goals. 

Configurations Matter

In addition to harmonizing numerical model assumptions, the teams also examined how common “structural” assumptions—called “configurations” in the study—influence results. Model configurations are necessary simplifications of real-world systems. Testing different assumptions can help decision-makers understand their influence on model results. For example, energy systems models often simplify how variability in space and time is represented due to computational constraints. Comparing results between different configurations can clarify the influence of such assumptions, leading to better models and stronger policy support.

Interestingly, these configuration choices had minimal impact on system costs but notable impacts on emissions outcomes and technology portfolios selected by the models.

For example, dispatching generation using a unit commitment configuration altered the impact of carbon buyout prices. Unit commitment better captures the operations of thermal generators. Under a net-zero scenario, modeling unit commitment more accurately reduced expected emissions. Why? The combination of unit commitment and carbon buyout prices disadvantaged thermal generators relative to clean energy technologies like batteries and wind. However, under current policies (without emissions targets), the unit commitment configuration actually increased emissions. Without the emissions buyout price, when unit commitment was included, the models found that it was cheaper to build fewer natural gas generators but run them more often.

Retirement assumptions proved even more impactful. Under net-zero scenarios, economically driven retirements swiftly eliminated coal generation, leading to large and sustained emissions reductions. But under current policies, economic retirement phases out only about half of the coal fleet, resulting in immediate but short-lived emission reductions. Within a decade or so, many existing nuclear plants also become economically unviable, creating a supply gap partially filled by fossil fuels, causing emissions to rebound significantly.

These findings highlight the nuanced ways in which model configurations impact results. While the configurations had a near-negligible impact on costs, their impact on emissions was significant—but conditioned upon the modeled scenario. The same configuration can either increase or decrease emissions, depending on the scenario. Clearly documenting and understanding these choices is essential for robust policy planning.

The Benefits of Intercomparisons 

The benefits of harmonized modeling are clear: teams can draw conclusions with greater confidence, knowing that observed differences in results genuinely reflect policy choices, technological uncertainties, or deliberate simplifications required for computational tractability—not undisclosed modeling artifacts.

Few teams undertake this kind of open, collaborative benchmarking, but the payoff is real: increased trust in results, sharper policy insights, and models that improve together. Funded by the Sloan Foundation, this effort demonstrates how transparency and teamwork can help policymakers navigate the complex path to decarbonizing our power systems.

See the full preprint of the article here. 

Acknowledgments: The authors want to thank the Sloan Foundation for supporting this work and acknowledge the entire project team, including Patricia Hidalgo-Gonzalez4, Jesse Jenkins, Oleg Lugovoy, Qian Luo, Michael J. Roberts, and Rangrang Zheng.