Papers

This study investigates whether the digestibility of corporate disclosures mitigates investors' information-processing costs and influences market outcomes. While emerging literature highlights the value of data visualizations, the cognitive friction caused by a filing's narrative structure and formatting remains largely unexplored. We extract key financial statement notes from S&P 500 10-K filings (2014–2024) with a custom pipeline that fully preserves the disclosures' overall structure (including tables, lists, and formatting) using markdown syntax. This allows us to utilize a Large Language Model (LLM) to score their structural and semantic clarity and provide supporting reasoning. We then consolidate these scores to construct a firm-year digestibility metric, categorizing disclosures as High Digestibility Filings (HDF) or Low Digestibility Filings (LDF) according to whether they fall in the top or bottom quartile of the sample distribution, respectively. Our determinant analysis reveals that firms use structural formatting to prevent severe opacity and to strategically offset the cognitive friction of complex disclosures. We find that HDF firms restore efficient price discovery, enabling immediate filing-date returns to predict future profitability. Furthermore, by allowing investors to rapidly synthesize disclosure implications, HDF firms mitigate the post-filing volatility typically triggered by information shocks. We also compare the explanatory power of our digestibility metric against its counterpart based on extracted words without retaining the disclosures’ overall structure. The results confirm that the word-based digestibility metric does not have the same explanatory power. Taken together, our findings support that disclosure digestibility is a distinct, primary driver of information processing and market efficiency.

We investigate whether investors incorporate peer narratives when assessing managerial sentiment in the Management Discussion & Analysis (MD&A) sections, and whether deviations from this peer benchmark influence firm valuation. Leveraging FinBERT to measure managerial tone across 26,056 10-K filings (2006 to 2023), we construct a measure of peer-relative tone deviation. Consistent with the accounting conservatism literature, we document a significant, asymmetric valuation effect: firms exhibiting positive tone deviation (relative optimism) suffer a 2.5 to 3.5 percentage point reduction in their one-year-ahead Market-to-Book multiple and relative valuation gap. In contrast, relative pessimism signals credibility and avoids this penalty. We demonstrate that this valuation discount is not an artifact of underlying fundamental performance but is highly conditional on disclosure readability and litigation risk, consistent with the disclosure processing cost framework. Finally, we show that relative optimism triggers a persistent skepticism discount around subsequent earnings announcements, adversely affecting the information environment for the focal firm. Our findings indicate that investors actively penalize relative optimism, illustrating how markets filter complex soft information by evaluating narratives through a peer-relative lens rather than in isolation.

Leveraging FinBERT, we separately measure tone in both the Management Discussion and Analysis (MD&A) and Risk Factor (RF) sections, across 40,116 10-K filings from 2006 to 2023. We demonstrate that the interaction between tones in the two sections is significant in predictive regressions of annual earnings. Our findings suggest that although the MD&A and RF sections are intended to convey different types of information, their joint information content enhances our understanding of the interaction between managerial sentiment and firm outlook. This joint informativeness is driven by tone spillover between the MD&A and RF sections, reflecting their interdependence due to shared firm-level determinants and managerial effort to present a cohesive narrative. We demonstrate that other key sections of the 10-K do not share the same relationship with MD&A, highlighting the unique contribution of the RF section. We demonstrate that (dis)agreement in changes in tone across the two sections contains meaningful signals about future earnings trends, shedding light on an avenue that firms may use to formulate their narrative disclosures strategically.

This study examines how investors recalibrated their reliance on accounting vs macroeconomic signals during the COVID-19 crisis. We study how stock price responsiveness to earnings and macroeconomic news varied before, during, and after the crisis. Using a sample of FTSE-350 firms, we find that investor reaction to earnings, as measured by the earnings response coefficient (ERC), was elevated during the pandemic, while the reaction to macroeconomic news remained unchanged. To understand the increase in ERC, we review the informativeness of stock prices during the pandemic, using firm size and stock price synchronicity as proxies. A decrease in price informativeness during the pandemic combined with a decline in the reliability of macroeconomic forecasts might explain the greater sensitivity to earnings news. To this end, we also show that the weakened investor response to macroeconomic news is associated with heightened uncertainty rather than increased economic risk. Our findings indicate that heightened uncertainty alters the hierarchy of information investors rely on, increasing the importance of timely accounting releases, especially during crises.