This paper argues that the modeling of positive inter-jurisdictional public spending spillovers may benefit from an analysis that involves three stages: the production of local public inputs, the production of local public outputs (accounting for input spillovers and non-additive aggregations), and the consumption of public outputs (accounting for output spillovers and preferences for variety). We map the existing literature onto this three-stage model and identify combinations of the features that generate strategically equivalent models of inter-jurisdictional public spending spillovers.