A puzzle of the modern welfare state is that a large fraction of social benefits is not takenup. Using a laboratory experiment, we present evidence that stigmatization through publicexposure causally reduces the take-up of a redistributive transfer by 30 percentage points.We build a theoretical model that interprets welfare stigma as unfavorable inferencesabout the claimant's type. Our design exogenously varies the informativeness of the takeupdecision by varying whether transfer eligibility is based on ability or luck. We find thatsubjects avoid the inference both of being low-skilled and of being willing to live offothers. Contrary to conventional wisdom, stigma may thus also contribute to low take-upif eligibility is not linked to economic performance.