How much blame do leaders deserve for counterfactual catastrophes: disasters that could have occurred under their watch but never did? We propose that when people oppose (vs. support) a leader, they are more likely to answer this question by considering how close the catastrophe came to occurring. In two preregistered experiments (total N = 1,927), U.S. participants assigned greater blame for counterfactual catastrophes (e.g., an imagined nuclear attack) to Presidents Trump and Biden when these catastrophes were framed as having been close to occurring rather than distant – but especially when judging the president they opposed (vs. supported). The results reveal a novel way in which counterfactual thinking facilitates motivated moral reasoning. Wanting to condemn a leader may decrease people’s focus on how a catastrophe never occurred and increase their focus on how it nearly occurred. We discuss how imagination can fuel political polarization.