Trade-offs arise between spatial dependence and choice of functional form in agricultural land price hedonic models. We discuss these trade-offs and how they can create spurious spatial dependence. Using a land sales dataset with apparent spatial dependence, we implement a semiparametric approach avoiding potential problems with the functional form. The results show that in addition to being nonlinear, the impacts are also characterized by significance thresholds that are difficult to capture in a parametric model. More importantly, we fail to detect any spatial dependence demonstrating that inappropriate functional form can indeed be responsible for finding spatial dependence in hedonic models.
Uncontrolled Keywords (separate with ;):