Emissions and energy impacts of the Inflation Reduction Act

If the United States passes the largest climate investment in its history, how much does the atmosphere actually notice? The law is on the books, the tax credits are real, and the money is moving. So what does the science say? Bistline and colleagues, leading a team spanning nine independent modeling groups, ran that exact experiment — and the number they landed on is between 43 and 48 percent below 2005 emissions levels by two thousand thirty-five. Nine different computer models, built on different assumptions and using different methods. That range is what they agreed on. Getting to that number required grappling with something genuinely unusual about the Inflation Reduction Act, or IRA. This is not a carbon tax. It is not an emissions cap. It is a portfolio of investment incentives — expanded tax credits, grants, loans, and rebates — designed to make clean energy and low-emissions equipment cheaper relative to the alternatives. Key provisions include making the investment tax credit and production tax credit more flexible and technology-neutral, letting standalone energy storage claim the investment tax credit for the first time, and raising the forty-five Q credit for captured and stored carbon dioxide from fifty dollars per metric ton to eighty-five dollars. Bonus credits layer on top for domestic content and energy communities. Many of these credits are uncapped. That structure is what makes the IRA hard to model. A carbon price produces a fixed cost signal. A tax credit produces an incentive that only bites if firms and households respond to it — and how fast they respond depends on infrastructure availability, supply chains, permitting timelines, and administrative guidance that agencies had not yet issued when this analysis was conducted. Modeling teams had to make judgment calls about which bonus credits to include, how quickly deployment could scale, and whether credits would be captured by actors who would have invested anyway. Those choices compound with uncertainty about fuel prices and technological change. The methodological response was a nine-model intercomparison. The nine participating models — EPS-EI, GCAM-CGS, Haiku-RFF, IPM-NRDC, MARKAL-NETL, NEMS-RHG, ReEDS-NREL, REGEN-EPRI, and RIO-REPEAT — each ran a common reference scenario representing the world without the IRA, then an IRA scenario built from each team's central interpretation of the law. Three of the models focus specifically on the power sector, capturing fine temporal and operational detail. The other six are broader energy-system or economy-wide models that link electricity with transport, buildings, industry, and macroeconomic feedbacks. The teams were not forced to harmonize all their inputs — technology cost trajectories and natural gas price paths differ across models — so the comparison surfaces both robust insights and honest uncertainty. That honest uncertainty shows up in the numbers. Across economy-wide models, the IRA reduces emissions relative to the reference by anywhere from one hundred ninety to nine hundred ninety million tonnes of carbon dioxide equivalent per year in two thousand thirty. That is a wide range, and the paper does not hide it. But the direction is unambiguous across all nine models, and the central tendency is clear. The electricity sector is where the IRA hits hardest. Models attribute between thirty-eight and eighty percent of two thousand thirty economy-wide reductions to the power sector alone — sixty-four percent on average. With the IRA, the power sector reaches between forty-seven and eighty-three percent below 2005 levels by two thousand thirty, compared to forty-one to sixty percent in the reference case. Wind and solar additions accelerate dramatically — an average of fifty-eight gigawatts per year under the IRA, versus twenty-seven gigawatts per year without it. Low-emitting generation's share of electricity rises to between forty-nine and eighty-two percent by two thousand thirty. Coal collapses: models show coal generation falling between thirty-eight and ninety-two percent from two thousand twenty-one levels by two thousand thirty under the IRA, compared to only three to sixty percent in the reference. Electrification amplifies the power sector story. New light-duty electric vehicle sales rise to between thirty-two and fifty-two percent of new vehicle sales by two thousand thirty under the IRA — up from roughly seven percent in two thousand twenty-two. Electricity's share of final energy grows from about twenty-one percent today to between twenty-three and twenty-six percent by two thousand thirty. Models with endogenous demand show electricity consumption growing about sixteen percent between two thousand twenty-one and two thousand thirty under the IRA. That is more demand on the grid, but it is demand being served by a much cleaner grid. Carbon capture plays a supporting but uncertain role. Models estimate captured carbon dioxide at between ten to three hundred fifty million tonnes per year by two thousand thirty, with an average of one hundred fifty million tonnes — driven partly by the higher forty-five Q credit making some coal-to-CCS retrofits viable. By two thousand thirty-five that range widens to between ten and eight hundred ten million tonnes per year. The spread tells you something important: this is one of the places where modeling assumptions diverge most sharply, and where real-world outcomes will depend heavily on infrastructure that does not yet exist. All of this adds up to the headline: economy-wide emissions reach between thirty-three and forty percent below 2005 by two thousand thirty, and between forty-three and forty-eight percent below 2005 by two thousand thirty-five. But the U.S. officially committed to fifty to fifty-two percent below 2005 by two thousand thirty. The remaining gap — between zero point five and one point one gigatonnes of carbon dioxide equivalent per year — persists across all nine models even with the IRA fully implemented. The law closes the gap substantially. It does not close it completely. The accounting goes beyond greenhouse gases. On air quality, studies cited in the analysis put monetized health benefits from power-sector reductions in sulfur dioxide and nitrogen oxides at between nine and twenty-two billion dollars per year by two thousand thirty, and economy-wide benefits from particulate matter reductions at roughly fifty-three billion dollars per year. The paper flags estimating these co-benefits from their own emissions trajectories as important future work — they relied on external estimates rather than running air-quality modeling themselves. On climate benefits, using central social-cost-of-carbon values at a two percent near-term discount rate, the IRA's emissions reductions generate between forty-four and two hundred twenty billion dollars in annual climate benefits by two thousand thirty. Average abatement costs across models range from twenty-seven to one hundred two dollars per tonne of carbon dioxide, with an all-model average of sixty-one dollars. Bistline and colleagues note these average costs are generally lower than many updated social-cost-of-carbon estimates — meaning the IRA clears a basic cost-effectiveness bar even before you count health benefits. Households, on average, come out ahead on energy spending. Economy-wide models project net energy spending declines of between two and twenty-six billion dollars per year by two thousand thirty relative to the reference — between thirteen and one hundred ninety dollars per household. Residential electricity prices drop roughly eighteen percent by two thousand thirty compared to the reference, even as total electricity expenditures can rise in some models because electrification expands demand. The fiscal scale is large. Modeled tax credit utilization reaches between three hundred thirty and eight hundred seventy billion dollars through two thousand thirty, with an average of five hundred ten billion. Add in direct IRA expenditures of one hundred twenty-one billion and the ten-year fiscal cost lands at roughly four hundred fifty to one thousand billion dollars across models — six hundred thirty billion on average. That is a significant public investment. The paper's conclusion is that when you set those costs against the climate and health benefits, the math generally works — but the range is wide enough that implementation details matter enormously. And implementation is exactly where the key unknowns cluster. Transmission buildout, hydrogen networks, carbon dioxide pipelines, supply chains for critical materials, siting and permitting — these are the physical and regulatory infrastructure that the IRA's incentives assume will materialize. Bistline and colleagues are explicit: the law does not guarantee those outcomes; it makes them cheaper to pursue. Whether they actually happen depends on decisions outside any model's reach. The paper points toward complementary policies as the mechanism for closing the remaining gap. Environmental Protection Agency standards targeting power plants, cars, and trucks are cited as regulations designed to work alongside IRA incentives rather than replace them. State policies and private-sector responses to IRA-driven cost reductions could push further. But none of that is in the models. What is in the models is clear: the IRA is the most significant single piece of U.S. climate legislation ever enacted. It materially reshapes what the energy system looks like by two thousand thirty-five, and it is still not enough on its own to meet the two thousand thirty target. That is not a criticism of the law. It is the most honest thing the analysis finds. A landmark achievement and an unfinished one, in the same breath. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.