P.I.: Silvia Muzzioli 

Budget: € 44.580,00 Euro

Funded by: University of Modena and Reggio Emilia

National Departments involved: Department of Economics “Marco Biagi”, Department of Physics, Computer Science and Mathematics, Department of Communication and Economics (University of Modena and Reggio Emilia, Italy).

International Departments involved: Department of Data Analysis and Mathematical Modelling, Ghent University, Ghent (Belgium); Department of Finance, Fox School of Business and Management – Temple University, Philadelphia (USA); Department of Computer Science and Artificial Intelligence, University of Granada, Granada (Spain).

ABSTRACT:

The project takes as its starting point the observation that, despite the importance attributed to market sentiment as a predictor of stock market returns, there is a lack of sentiment measures for the European stock market, particularly at the individual market level. Moreover, there is no consensus in the literature on the construction of sentiment indicators, and the techniques adopted to analyze, process, and aggregate the data in sentiment indicators call for a radical change. To fill this gap, the project aims to develop and compute composite indicators of sentiment for most EU countries by aggregating several sources of information. The first type of information included in our sentiment indices consists of investors' expectations embedded in option prices. The second pillar of our composite sentiment indicators is backward-looking and is based on several economic and financial variables. The third group of variables in our composite sentiment index will be extracted from text in social media, microblogging sites, and web searches exploiting supervised and unsupervised approaches enabling us to account for the rapid evolution of investor sentiment in response to unexpected developments. The final aim of the project will be to evaluate the relationship between the proposed sentiment measures and future market returns. The analysis will be carried out through in-sample and out-of-sample analyses by using machine learning methods and trading strategies. The results of the project are expected to be of great importance for a wide range of stakeholders: investors, firms, and policy-makers.