TFL Vsebine / Revija SIR*IUS

Econometric model for house price forecast

Gradivo predstavlja bayesianski model vektorske avtoregresije za napovedovanje gibanja cen nepremičnin v Sloveniji. Ključne ugotovitve so, da so cene nepremičnin v Sloveniji odraz obsega investicij v nepremičnine, povprečne neto plače, stopnje brezposelnosti, gradbenih stroškov, rasti slovenske ekonomske aktivnosti, merjene z bruto domačim proizvodom, ter obrestnih mer za stanovanjske kredite. Model ob napovedih zmerne rasti slovenskega gospodarstva nakazuje zaustavitev padanja cen nepremičnin in kaže na sicer le postopno okrevanje nepremičninskega trga. Napovedovanje gibanja cen nepremičnin lahko poleg tega predstavlja tudi pomembno orodje v izvajanju makrobonitetnega nadzora, zato je nadaljnji razvoj ekonometričnih orodij za napovedovanje gibanja cen nepremičnin potreben ne glede na omejitve, ki jih predstavljajo razmeroma kratke podatkovne časovne serije za cene nepremičnin v Sloveniji.

This paper presents a bayesian VAR model for house prices forecasting in Slovenia. The key findings of the paper are that the house prices in Slovenia are driven by the following 6 variables: real housing investment, real average net wage, unemployment rate, real construction costs, real gross domestic product, and real housing interest rate. The applied econometric framework predicts that the declining trends in Slovene housing market will probably come to a standstill. As the solid economic growth in Slovenia continues, presented econometric model predicts a positive but only gradual growth of house prices over the projection horizon. In addition, forecasting housing prices can become a necessary tool for a prudent implementation of macro-prudential regulation. Further development of econometric tools for predicting developments in Slovene housing market is necessary irrespective to the limitations posed by relatively short Slovene house-price time-series data.