Association for Public Policy Analysis & Management

The attractiveness of a location depends on the combination of natural amenities and local government public goods. I propose that locations differ in their stock of local government public goods because of differences in existing natural amenities. Some local government public goods are either necessary to overcome natural characteristics or less costly to provide in certain locations. More specifically, I propose that some local government public goods are substitutes to natural amenities while others are complements.

In this research, I define substitutes as local government public goods that are provided in the absence or low presence of certain natural amenities. Meanwhile, complements are local government public goods that are provided when certain natural amenities are abundant. Existing studies do not consider the interaction between these two types of common resource. If natural amenities affect the provision of local government public goods, the effect of the latter on the attractiveness of a location may be overestimated (for substitutes) or underestimated (for complements). A gap in the literature is the lack of studies that explore how local government public goods vary with the existing natural amenities in locations.

Moreover, there could also be spillover effects of local government public goods among adjacent locations. I propose to test whether some local government public goods in the United States are substitutes to natural amenities, while others are complements considering the spillover effects among adjacent locations.

The unit of analysis in this study is the county government. The theoretical approach of my research is a combination of the Tiebout (1956), Wildasin (1988), Koethenbuerger (2011), and Janeba-Osterloh (2013) models. To test my hypothesis, I will explore how a county’s tax policies and expenditures vary with respect to natural amenities. I will analyze merged datasets at the county-level using climatological data and coastal data from the National Climatic Data Center (NCDC), topographic type and water area percentage from the United States Department of Agriculture Economic Research Services (USDA-ERS), recreational areas from ArcGIS, and tax policies and government expenditures from the Government Finance Database (GFD; Pierson et al. 2015) for the period 1970-2012. I propose to use spatial regression, particularly simultaneous autoregressive (SAR) models, in the determination of which natural characteristics of locations are important for the determination of tax policies and levels of county expenditures. I will also test for spillover effects.

I expect that my estimation results may show which local government public goods are substitutes to natural amenities and which ones are complements. Moreover, my expected results may also show that there are spillover effects among adjacent counties in the provision of certain local government public goods.