In this study, we conduct a textual sentiment analysis of hand-collected U.S. banks’ stress test disclosure reports. Our findings reveal that sentiment factors, including positive wording and strong modals, convey significant financial information about banks. Both factors are positively associated with banks’ projected minimum and end-of-horizon Tier 1 capital ratios. In addition, strong modal sentiment reflects banks’ financial standing, as indicated by z-scores. These results remain consistent across a series of robustness checks and a difference-in-differences analysis. Furthermore, we find that textual sentiment provides informative insights into banks’ performance measures, including net revenue rate and net interest income rate. In contrast, negative wording contains little information regarding banks’ capitalization and financial condition, which supports our argument that banks tend to avoid negative language in their stress test disclosures. Overall, the linguistic tone of banks’ self reported stress test disclosures offers meaningful information about their capital adequacy, financial strength, and operational performance.
Zheng, Yi. Do Banks’ Self-Reported Stress Test Disclosures Convey Critical Financial Information? A Textual Sentiment Analysis. // Journal of Financial Services Research. – 2026. – DOI: 10.1007/s10693-026-00475-3. – URL: http://link.springer.com/openurl/fulltext?id=doi:10.1007/s10693-026-00475-3
Zheng & Yi (2026). Do Banks’ Self-Reported Stress Test Disclosures Convey Critical Financial Information? A Textual Sentiment Analysis. Journal of Financial Services Research. https://doi.org/10.1007/s10693-026-00475-3
Zheng and Yi. “Do Banks’ Self-Reported Stress Test Disclosures Convey Critical Financial Information? A Textual Sentiment Analysis.” Journal of Financial Services Research, 2026. doi:10.1007/s10693-026-00475-3.
@article{Zheng2026,
title = {Do Banks’ Self-Reported Stress Test Disclosures Convey Critical Financial Information? A Textual Sentiment Analysis},
author = {Zheng and Yi},
journal = {Journal of Financial Services Research},
year = {2026},
doi = {10.1007/s10693-026-00475-3},
url = {http://link.springer.com/openurl/fulltext?id=doi:10.1007/s10693-026-00475-3}
}
Using data on publicly traded U.S. bank holding companies from 1992 to 2019, we examine whether disparities between CEO and non-CEO executive pay affect banks’ liquidity creation. We find that banks with larger CEO pay gaps create more...
In this study, we use data from 22,822 women-led firms across 50 countries spanning 2010 to 2020 to challenge the belief that corporate political activities (CPAs) facilitate credit access. We demonstrate that CPAs do not improve access...
We study the effect of public-guaranteed loans (PGLs) on bank risk-taking during the COVID-19 pandemic in France. The presence of guarantee schemes may foster riskier lending, pushing banks to lend to riskier borrowers or worsening...
Our aim in this paper is to gain a better understanding of the potential moderating role of the peer effect in loan officers’ decision-making. We use data from a French cooperative bank with several business branches from 2012 to 2016...
We examine whether the discretionary governance practices of pension board of trustees have a significant impact on the returns of the plan’s invested assets. We construct a unique database of board practices by analyzing meeting...