ABSTRACT
This article attempts to explain and predict housing prices by constructing a model based on the variables that most influence demand: the theoretical purchase effort index without tax deductions as well as a new and innovative indicator that includes the excess of mortgages granted. The Johansen methodology for cointegration analysis reveals the existence of long-run equilibrium and the model’s subsequent ECM, to verify the statistical significance of the variables, confirms the validity of the model concerning this Spanish case study.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1 Data published by Bank of Spain and Ministry of Development. EM is own elaboration from data of Bank of Spain and Ministry of Development.The variables have been deseasonalized using Tramo Seats methodology (TSW). The housing price variable has been deflated by using the CPI.The rate of correlation between the variables, which might cause multicollinearity problems, has been analysed.Ramsey’s reset test was used to verify that the functional form is adequate.The Breusch–Godfrey test was used to analyse whether or not there were any problems with autocorrelation, the autoregressive conditional heteroskedasticity (ARCH) test was used to analyse heteroscedasticity, and the Jarque–Bera test was used to analyse normality.
2 The Bank of Spain develops this indicator based on the quotient between the cost during the first year of a fixed-rate mortgage, those currently being used, with a term of 15 years and which allow 80% of the price of an average home to be financed, and the average annual salary. An average home is considered to be a property of 93.75 m2 built space (approximately 75 m2 of usable floor space).
3 The Breusch–Godfrey test was used to analyse whether or not there were any problems with autocorrelation, the ARCH test was used to analyse heteroscedasticity.
4 Due to the peculiarity of the real estate market which works at slower pace, the same test has been performed with eight lags, with the same result of long-term equilibrium. On the subsequent estimate for ECM the explanatory variable mortgage excess with longer lags has resulted in high levels of significance.