Article: Sentiment and banks' performance: evidence from Musk's tweets Journal: International Journal of Banking, Accounting and Finance (IJBAAF) 2025 Vol.15 No.4 pp.381 - 420 Abstract: This study aims to examine the relationship between sentiment and bank performance by considering a sample of 225 US-listed banks during the period 2017-2021. The analysis uses RoBERTa (a pre-trained transformer model) to extract emotions from Elon Musk's tweets and then apply the GMM model to assess the impact of sentiment on banks' profitability. The findings suggest that the overall sentiment has a negative impact on ROA, with negative emotions - such as, anger and sadness - having a statistically significant negative impact on ROA, while positive emotions appear to be insignificant. Moreover, sentiment has no effect on ROE, but negative emotions negatively affect ROE. The findings survive geographical and political regimes. These results carry certain implications for banks, enabling more informed decision-making regarding risk management, loan approval processes, and customer engagement strategies. Additionally, policymakers should consider social media sentiment as a potential source of economic volatility and its implications for financial stability. Inderscience Publishers - linking academia, business and industry through research

Title: Sentiment and banks' performance: evidence from Musk's tweets

Authors: Nicholas Apergis; Lorenzo Calò; Pierluigi Toma; Valeria Stefanelli

Addresses: Centre for Economics, Energy and Public Policy, Faculty of Economics, Prague University of Economics and Business, 120 00 Praha 3-Žižkov, Czechia ' Department of Economics and Management, School of Finance and Technology, University of Salento, Piazza Tancredi, 7, 73100 Lecce LE, Italy ' Department of Economics and Management, School of Finance and Technology, University of Salento, Piazza Tancredi, 7, 73100 Lecce LE, Italy ' Department of Economics and Management, School of Finance and Technology, University of Salento, Piazza Tancredi, 7, 73100 Lecce LE, Italy

Abstract: This study aims to examine the relationship between sentiment and bank performance by considering a sample of 225 US-listed banks during the period 2017-2021. The analysis uses RoBERTa (a pre-trained transformer model) to extract emotions from Elon Musk's tweets and then apply the GMM model to assess the impact of sentiment on banks' profitability. The findings suggest that the overall sentiment has a negative impact on ROA, with negative emotions - such as, anger and sadness - having a statistically significant negative impact on ROA, while positive emotions appear to be insignificant. Moreover, sentiment has no effect on ROE, but negative emotions negatively affect ROE. The findings survive geographical and political regimes. These results carry certain implications for banks, enabling more informed decision-making regarding risk management, loan approval processes, and customer engagement strategies. Additionally, policymakers should consider social media sentiment as a potential source of economic volatility and its implications for financial stability.

Keywords: sentiment; bank profitability; Musk's tweets; US banks.

DOI: 10.1504/IJBAAF.2025.151899

International Journal of Banking, Accounting and Finance, 2025 Vol.15 No.4, pp.381 - 420

Received: 10 Apr 2025
Accepted: 10 Oct 2025
Published online: 25 Feb 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Sentiment and banks' performance: evidence from Musk's tweets

Keep up-to-date