This paper surveys recent contributions to the study of fiscal decentralization which adopt a political economy approach. It is argued that this approach can capture, in a variety of formal models, the plausible and influential ideas (increasingly, supported by empirical evidence) that fiscal decentralization can lead to improved preference-matching and accountability of government. In particular, recent work on centralized provision of public good provision via bargaining in a legislature shows how centralization reduces preference-matching, and recent work using "electoral agency" models formalizes the accountability argument. These models also provide insights into when decentralization may fail to deliver these benefits.