This paper examines the impact of fiscal decentralization on economic growth through the channels of (1) efficiency in allocating resources, and (2) counterbalancing effect between investment and tax competitions brought by capital mobility.

To do so, I construct theoretical models with four scenarios assuming a country’s regional disparities in endowment and preferences for public goods and services, and derive the maximization of output based on different utility functions and budget constraints of central and subnational governments. I then examine the relationship through these two channels with empirical analyses using a panel dataset with 67 countries during 1990-2016, and use Difference- and System-GMM as well as Structural Equation Modeling to conduct robustness checks.

Empirically, I find that under fiscal decentralization, when subnational governments allocate functional spending more differently over years, and that when subnational governments allocate functional spending more differently than the central government, they are correlated with higher economic growth. Theoretically, I also find that under fiscal decentralization, endowment effect plays an important role—well-endowed regions will be made better off in the process of investment and tax competitions, while poorly-endowed regions will be made worse off. The magnitude of this difference also depends on average tax rate and the government's elasticity of substitution between public spending and rent-seeking. Though this argument has yet been backed by empirical results.