Assessing housing market dynamics across a sample of European cities: A Random Forest Machine Learning Approach

Luca Begatti & Corrado Mauceri

Studenteropgave: Kandidatafhandlinger

Abstrakt

This paper studies the housing market across European cities by combining regression analysis with a machine learning process. We identify that despite the limited data availability, there are time, national and local effects associated with real housing returns. According to our choice of explanatory variables, we document that mortgage rate, GDP per capita and unemployment rate are important determinants of housing returns. Generally speaking, we can infer that the European housing market does not show evidence of bubbles but there are some markets which deserve particular attention. We witness the presence of Shiller's Irrational Exuberance only with respect to the city of Nurnberg, Germany. We further witness in some cities a substantial adjustment rate suggesting market efficiency as well as strong mean reversion effects. Collectively, these results support the view that most of the European cities are not currently experiencing a bubble but in some instances the foundation for such inexplicable behaviour have been building up following the recent financial downturn.

UddannelserCand.merc.oecon Advanced Economics and Finance, (Kandidatuddannelse) Afsluttende afhandling
SprogEngelsk
Udgivelsesdato2018
Antal sider105