CMStatistics 2017: Start Registration
View Submission - CFE
A1367
Title: On possible causal links between Twitter sentiment and banks financial ratios Authors:  Paola Cerchiello - University of Pavia (Italy) [presenting]
Juri Marcucci - Bank of Italy (Italy)
Giancarlo Nicola - UNIPV (Italy)
Giuseppe Bruno - Bank of Italy (Italy)
Abstract: The aim is to study the relationships between Twitter sentiment and various financial indicators (such as for example stock return or volume) of some of the major Italian banks. Moreover, we test the current technology for analyzing and evaluating the sentiment of short web-text messages written in Italian, such as those published on the Twitter micro-blogging platform. In fact, gauging the sentiment among financial investors is of paramount importance for both market participants and regulation authorities. Behavioural finance posits that stock market investors define their purchasing strategies considering arbitrage bounds and collective sentiments. Regulation and market authorities can address critical situations by collecting and analyzing the sentiment mood inferred from investors action on social media. The goal is twofold: on one hand, we provide an empirical method to evaluate the polarity of short text messages in Italian. Secondly, we establish a sound statistical framework to measure the causal links between sentiment extracted from Twitter and financial market variables even in presence of no stationarity and cointegration in the data. A quantitative evaluation of the impact of sentiment on financial indicators is relevant to increase the timely awareness of regulators with respect to potentially critical microeconomic conditions.