We study the informational content of qualitative disclosures in private equity performance reporting. Using more than 20,000 interim reports of roughly 2,000 portfolio companies, we analyze narrative updates and interim performance measures. We quantify report tone using large language models trained on financial text and test whether tone predicts realized performance. More positive tone predicts higher subsequent performance: a one standard deviation increase in tone is associated with about 8% higher annualized returns from the interim report to exit, controlling for interim valuations, portfolio company fundamentals, and fund characteristics. Tone remains informative throughout the investment lifecycle and across geographies, institutional settings, and accounting regimes. Machine learning algorithms show that the informational value of these reports is also substantial when tested out of sample. The results indicate that private equity narratives convey forward-looking information that is not fully incorporated into reported interim valuations, and have implications for delegated monitoring and performance measurement in private markets.
Borja Fernández Tamayo, SKEMA Business School & Unigestion SA
Reiner Braun, Technical University of Munich School of Management
Florencio Lopez-de-Silanes, SKEMA Business School & NBER
Ludovic Phalippou, University of Oxford Said Business School
Natalia Sigrist, Unigestion SA