Determinants of fiscal decentralization: a neural network approach for OECD countries

Le résumé fourni par la source

Fiscal decentralization is a complex phenomenon driven by several forces, including socioeconomic, institutional, and historical issues. We analyze the determinants of fiscal decentralization through the percentage of expenditure or revenue of the subcentral governments – over the total – in 30 OECD countries for the period 1995–2021. We employ an estimation and hypothesis testing methodology that relies on nonparametric artificial neural network regressions and allows detecting potential nonlinear patterns in the data. Our results show the presence of nonlinearities in the case of expenditure, with weaker evidence found for revenue when less decentralized than expenditure. We find population, urban population, GDP per capita, and corruption as determinants of fiscal decentralization, whereas density and economic openness are nonsignificant.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.