This paper uses spatial econometric methods to investigate property-tax competition among local governments. The theoretical model is drawn from the literature on tax competition, in which local jurisdictions choose property-tax rates taking into account the migration of mobile capital in response to tax differentials. Using a "spatial lag" econometric model, the paper estimates the reaction function of the representative community, which relates the community’s property-tax rate to its own characteristics and to the tax rates in competing communities. A nonzero reaction-function slope indicates the presence of strategic interaction in the choice of tax rates. The estimation uses cross-section data on property taxes and other socio-economic variables for cities in the Boston metropolitan area. The results, which are presented for two periods before and after imposition of Proposition 2 1/ 2 (a tax limitation measure), indicate the presence of strategic interaction.