Forecasting the German Real Estate Market with Dictionary-based Textual Analysis

Abstract

A growing body of research uses textual analysis to examine the tone or sentiment of newspaper articles, press releases and investor message boards in order to characterize the relationship between media reports and market movements. However, the applications of textual analysis to real estate markets are rather rarely especially to the German market.Most of the time, the starting point to conduct textual analysis is a sentiment annotated wordlist. Therefore, in a first step, a survey among more than 1,600 real estate experts is conducted in order to objectively reveal positive-neutral-negative sentiment word classifications. The results form a discipline-specific “German Real Estate Sentiment Dictionary”, which consists of about 2,000 positive and negative words. Hence, in a second step, the German Real Estate Sentiment Dictionary is applied to media data of the most important German real estate newspaper in order to investigate whether there is a potential relationship between the generated sentiment index and the German real estate market.This research opens up not only a whole new dataset and its application to the German real estate profession, it is a unique example how dictionary-based textual analysis can be transferred to non-English-speaking countries.