Modeling Regional House Prices in the Netherlands

This study applies panel data techniques to explore models to explain regional differences in house price developments, and to explore if the models vary per type of house or type of living environment. The Sale Price Appraisal Ratio (SPAR) method is applied to sales data from the Netherlands to create prices indices. The SPAR method uses house appraisal values to correct the price index for the mix of houses sold per time period. Several variables for which per region time series are available are used as explanatory variables. Used are income, the total number of households, the total number of houses and the number of jobs. The models proposed assume cointegration between prices and income. The development of prices is explained using lagged prices, the short-term shocks, and an error correction term, representing the deviation from the long-term equilibrium. The creation of the indices from the sales data was successful, albeit challenging for the indices that were split per type of house and per type of living environment. The empirical data confirms the existence of the cointegration relationship between prices and income. Due to the presence of cross-sectional dependency, the CCEP estimation method was used. Here the crosssectional average to the explanatory variables are added to the model to capture the unobserved variables representing the cross-sectional dependency. The standard errors for the estimated coefficients are high. The values and signs of the estimated coefficients were generally in line with expectations from economic reasoning. The estimations on the detailed indices showed that the estimated model coefficients are different depending on the type of house and the type of living environment.