Abstract

Transparency Measure with Twitter Activity: Gender Related Messages as Predictive Content for Company Social Scores

Belma Ozturkkal and Alessandra Tanda

Social media use and link to financial performance is investigated from different perspectives. Twitter tweets and comments have been investigated in the literature as a tool to proxy investors sentiment by analysts and their relation with company valuation is becoming trending in the financial studies (Albarrak et al., 2020; Othan and Kilimci, 2021; Alkan et al., 2022; Garcia, 2022). Nevertheless, a gap remain in the research on the topic from the perspective of transparency. To fill this gap, we study twitter activities of listed companies from four Mediterranean countries (Turkey, Italy, Spain and Portugal) and show that tweets increase transparency, and they can be used as proxy for their social scores within the ESG scoring. One of the novelty of this study is to analyse the transparency in four different markets, which have certain similarities. Miller and Skinner (2015) show that change in technology changes the disclosure of companies in different forms to a new media platforms. Amin et al. (2021) note Twitter is an important social media for CSR (Corporate Social Responsibility) disclosure. Additionally, they note women in boards are important for social media disclosure. We use data between 2018 and 2022 and retrieve the tweets with python Natural Language Toolkit. We use itertools for counting the list of words. We then estimate Refinitiv company social scores by employing the gender related Twitter activity in the native languages of chosen countries. The list of keywords in initially compiled in English, then translated and validated by native
speakers. Preliminary evidence is based on 3,299 gender tweets out of 111,224 tweets related to 195 companies. This contributes to 3% of the tweets in Turkey. In Italy, there are 1,316 gender tweets out of 171,628 tweets for 94 companies and this contributes to 0.75% of the tweets. In Portugal, there are 17 gender tweets out of 15,075 tweets related to 5 companies and this contributes to 0.1% of the tweets in the country. Preliminary results show a strong relation between gender tweets and Refinitiv social score.