By consequentialist standards, the only thing that should matter is how successful policies are at reaching their objective, relative to their cost. Yet, across 5 online studies (N = 1515), we found that French participants regarded a policy driven by an altruistic intention but that turned out to reach its objective very poorly at a huge cost, as being more commendable than (Experiment 1-4), and as deserving equal support as (Experiment 5), a policy motivated by selfishness but that dramatically helped the issue while saving lots of money. This preference was observed whether the decision was made by a CEO or minister, and across four issues. Independent manipulation of intent and efficiency (Experiment 5) suggested that folk judgments of policies are characterized both by low sensitivity to huge differences in efficiency expressed in numerical format, and high sensitivity to actors’ intentions. Participants’ moral commitment to the issue predicted greater support for any policy that somehow contributed to help the issue, whatever its level of efficiency and the intention driving it (Experiment 5), suggesting that moralization makes people more deontological in their judgments.