Reza Taleblou; Teymour Mohammadi; Hadi Pirdayeh
Abstract
Researches in housing sector show that the effect of various economic factors on housing prices might be different in separate areas of a country and housing prices in different regions of the country have internal connections with each other. Modelling of these effects is done in the form of spatial ...
Read More
Researches in housing sector show that the effect of various economic factors on housing prices might be different in separate areas of a country and housing prices in different regions of the country have internal connections with each other. Modelling of these effects is done in the form of spatial Econometrics. In this study, data related to 28 provinces of Iran during the period 2000 -2013 is used to estimate and compare the dynamic spatial SDM panel models with spatial SDM panel models and also to estimate the direct and indirect effects (space overflows) related to the explanatory variables in both the short and long term by using population spatial weight matrix in Matlab software. In order to choose the best spatial pattern consistent with the theoretical model of housing price determination, we have used Elhorst methodology and at every step, Likelihood ratio test (LR) and Lagrange multipliers tests (LM) are used to compare the spatial patterns. We found out that dynamic spatial model shows the best specification. By comparing the results of the dynamic spatial panel models, lagged housing price variable and spatial effects of this variable have a significant role in determining house prices. The results also show that only the spatial effects of household spending variable have a significant effect on housing prices and other variables such as land price, construction costs, rental housing prices, have significant effect on housing prices in the provinces of Iran both directly and in the form of space overflow effects.
