The analysis of Istanbul housing prices by hedonic price theory: Insights from parametric and semiparametric models

Abstract

Most researchers use parametric models involving specific functional forms in estimating hedonic housing price functions although related literature emphasizes intrinsic nonlinearity in the relationship between housing prices and characteristics. The rigid structure of parametric models might cause to be misspecification errors arising from a wrong functional form and so, misleading inferences. In contrast to the shortcomings of the parametric model, semiparametric models offer significant advantages for hedonic price function estimation by providing functional form flexibility. This paper investigates that how semiparametric models can be used to assist in the specification of a hedonic house price function using a dataset of apartment flat sales in Istanbul for 2013 October–December. The findings based on semiparametric model demonstrate that all housing attributes in nonparametric form have a significant nonlinear impact on housing prices. Among all housing attributes considered, Bosphorus view has the strongest positive impact on housing prices in Istanbul.