REGIONAL HOUSING DEMAND: PANEL DATA ANALYSIS

Abstract

This paper analyses the variables affecting housing demand in Turkey with regional data. For this purpose, housing demand is estimated with data from 26 regions of Turkey according to the Classification of Territorial Units for Statistics using 3 models and robust panel data estimators with the variables of per capita income, house price difference, general price level, housing production and real estate loans. According to the findings, an increase in house prices decreases housing demand, while increases in other explanatory variables increase housing demand. In the first model, where housing prices, income and prices of other goods and services are used as explanatory variables, the price elasticity of housing is estimated as 0.06, income elasticity as 1.505 - 1.791 and cross-elasticity of demand as 0.83-1.71. In the second model, after the inclusion of housing production, price elasticity (0.143) and cross-elasticity of demand (1.211 in Beck-Katz - 1.396 in Driscoll-Kraay) increase, while income elasticity (1.066-0.866) decreases. In the third model, when real estate loans per household are added to the model, the parameters persist at approximate values. In the first model, where only house prices, household income and prices of other goods and services are included in the model, per capita income is the variable that affects housing demand the most. With the addition of housing production and real estate loans to the model, movements in prices of goods and services other than housing have larger parameter values compared to other variables.