<p>This paper reassesses the alleged moral difference between killing and letting die by shifting the analysis from permissibility to the allocation of consequence responsibility. Within Fischer and Ravizza’s guidance-control framework, responsibility is two-staged: agents are conduct-responsible when their behaviour issues from a mechanism they own that is at least moderately reasons-responsive; and they are consequence-responsible when, in addition, the death arises along a non-deviant, normatively intelligible route from that behaviour. I then operationalise the second stage via a disciplined counterfactual-subtraction procedure that is neutral between actions and omissions and that distinguishes consequence-particulars from consequence-universals. Applied to canonical contrasts, the framework yields patterned results: where both killing and letting die satisfy (or both fail) the same responsibility conditions, they are moral equals with respect to the death; where they come apart, any asymmetry tracks ownership, reasons-responsiveness, or decisive causal contribution rather than the surface doing/allowing label. The upshot is a unified, practice-usable standard for foreseeable-death contexts in end-of-life medicine and clinical governance, clarifying documentation, role delineation, and the evidential focus for retrospective appraisal without presupposing a default moral privilege for omissions.</p>