Logarithmic spatial heteroscedasticity model with exogenous variables: Ohio’s housing prices

The investigation of the spatial dependence in variance is obtaining increased attention. In this paper, we propose a new generalised logarithmic spatial heteroscedasticity model with exogenous variables. For this model, we provide its estimators and tests for parameter constraints or model checking. Moreover, we establish the asymptotics of all estimators and tests using the tool of spatial near-epoch dependence, and examine their finite-sample performance by simulation. Finally, we give an empirical application for investigating the impact of spatial dependence and house characteristics on both the mean and variance of house prices.