Abstract
A hedonic equation considers OLS models with independently and identically distributed errors. However, quality of property and location tend to exhibit highly auto-regressive correlation due to spatial dependence and heterogeneity. A modification is made to multiple regression analysis based on the land-rent concept so that the modified hedonic house price function can be determined. This paper develops a stochastic approach which is able to correct autocorrelation bias in the hedonic function. Others tend to model the spatial autocorrelation through the error terms, while this study models it through the constant term. The sample was orientated by location to reflect the neighbourhood effects. Thus the neighbourhood effects can be separated from the random disturbance. Our model, using data from Hong Kong, incorporates adjustments reflecting net floor area ratio, age, floor level, views, transport accessibility and amenities such as availability of recreational facilities. The stochastic model is a more flexible application of hedonic price regressions, making it an appealing alternative to spatial problems.