Selective Centralization: Governing Capability Redeployment Across Differentiated Business Units

When a firm redeploys a capability into an adjacent business, who governs it? Giving the receiving unit full authority fragments the capability; keeping it with the source unit freezes it. We show the choice is false. Drawing on a longitudinal case study of a European clean energy and electric mobility firm—one platform traced end-to-end, two further transfers, one failed attempt—we find that authority is not allocated over a capability but partitioned within it: the source unit keeps the codifiable core, the receiving unit adapts the context-dependent periphery. The partition is constituted by three decision rights: who defines the core, who releases it, who may change what surrounds it. Where no core can be separated, there is nothing to partition and the transfer is abandoned. We term the pattern selective centralization and carry the logic of the mirroring hypothesis inside a single capability in motion—the first process-level account of how redeployment is governed