We study the extent to which house price dynamics differ across market segments and possible drivers of this heterogeneity. We construct a data set for individual houses and mortgages, based on a survey of about 500 Dutch households conducted over the period 2003-2016. We estimate a dynamic panel data model of house price dynamics by means of the Arellano-Bond estimator. Three main empirical results emerge. First, we generally find that house price dynamics imply a convergence towards their long-run equilibrium value, as indicated by a low serial correlation coefficient and a positive estimated mean reversion coefficient. Second, there is evidence that the housing market in the Netherlands is inefficient. Third, there is important heterogeneity across different market segments, with some markets being more œcyclical than others. In particular, the speed of convergence of house price dynamics and the efficiency of housing markets depends on the geographical location, the degree of urbanization, and the type and year of construction of a house. We do not find evidence of significant heterogeneity across different types of mortgage financing and households income.
Keywords: Housing market dynamics, house prices, heterogeneity, survey data, panel analysis.
JEL classifications: D14, G12, R32. Working paper no. 564