Energy system optimization models facilitate analyses on a national or regional scale. However, understanding the impacts of climate policy on specific populations requires a much higher spatial resolution. Here, we link an energy system optimization model to an integrated assessment model via an emission downscaling algorithm, translating air pollution emissions from nine U.S. regions to U.S. counties. We simulate the impacts of six distinct policy scenarios, including a current policy and a 2050 net-zero target, on NO x , SO 2 , and PM 2.5 emissions from on-road transportation and electricity generation. We compare different policies based on their ability to reduce emission exposure and exposure disparity across racial groups, allowing decision-makers to assess the air pollution impacts of various policy instruments more holistically. Modeled policies include a clean electricity standard, an on-road ICE vehicle ban, a carbon tax, and a scenario that reaches net-zero GHG emissions by 2050. While exposure and disparities decrease in all scenarios, our results reveal persistent disparities until at least 2040, particularly for Black non-Hispanic Americans. Our estimates of avoided deaths due to air pollution emphasize the importance of policy timing, showing that thousands of lives can be saved by taking action in the near-term.