Spatial statistical methods in political science provide a tool to deal with spatial and other forms of interdependence in observational data. In this article, I derive a statistical model from a game of impure public goods provision. The resulting strategic autoregressive model (StratAM) allows the researcher to systematically explore the sources of free-riding behavior in the provision of public goods. The StratAM model is tightly related to the well-known spatial autoregressive (SAR) model and can be estimated in a maximum likelihood framework. I demonstrate the use of the StratAM model by analyzing free riding in the provision of foreign aid. Indicators of developmental needs and good governance strongly increase free-riding during the 1990s. Free-riding patterns in the 2000s are more similar to Cold War patterns.