Thesis: This PhD thesis covers several topics in the fields of public finance and economic geography. The first article was written with Aurélie Sotura. It is the first known attempt at structurally estimating local public good spillovers. We develop a simple and flexible theoretical framework that allows to empirically test for the presence of public good spillovers between fiscally autonomous jurisdictions – i.e, how much residents of a given jurisdiction bene fit from public goods provided by close-by jurisdictions – and to investigate potential welfare gains from increased centralization of public good provision. Applying our framework to French data, we estimate strong public good spillovers between French municipalities. The second article was written with Benjamin Carantino. We study the normative implications of public good agglomeration externalities and heterogeneous location preferences. We present new graphical evidence suggesting substantive scale economies in the consumption of local public goods. We then build an economic geography model and derive a computable efficiency test for observed spatial transfers. Applying it to French data, we show that ignoring location preferences leads to recommending reforms that mistakenly redistribute towards the densest and richest places. Our findings also suggest that the government uses spatial transfers to compensate unequal public good agglomeration gains. The last article provides novel quasi experimental evidence on behavioral responses to property transaction taxes, using a unique administrative dataset on the universe of housing transactions in France and exploiting a 2014 tax reform.