Knowledge that Transforms

We compare the earnings information produced by the five largest forecast data providers (FDPs)—Bloomberg, Capital IQ, FactSet, I/B/E/S, and Zacks—and observe substantial differences across FDPs in both forecasted and actual street earnings values, and thus the earnings surprise, for the same firm‐quarter. We provide evidence that differences in the earnings surprise across FDPs for the same firm‐quarter (i.e., “FDP differences”) have economically meaningful implications for price responsiveness and liquidity around earnings announcements. We also find that, for announcements where FDPs disagree about whether the announcing firm missed or beat earnings expectations, investors are more likely to side with higher‐quality FDPs but may not fully impound the implications of FDP quality differences during the announcement window. On average, relative to the other FDPs, I/B/E/S ranks highly in our measure of FDP quality, such that investor reactions are likely to align with I/B/E/S earnings information, validating its use as a representative FDP in academic research. Taken together, our results are consistent with FDPs pursuing differentiated information production strategies that generate capital market frictions when these strategies lead to material FDP differences.

We are among the first to investigate how Generative AI (GenAI) shapes investors' trading activities. Using an AI‐sentiment measure extracted from earnings‐call transcripts to proxy for textual signals, we find notable shifts in trading behaviors around earnings calls. Before the wide deployment of ChatGPT, short selling was aligned with AI‐sentiment, whereas retail trading was not. However, following ChatGPT's deployment, the alignment of retail traders with AI‐sentiment significantly increases, while the alignment of short sellers weakens, albeit insignificantly. Stocks with higher information processing costs exhibit a more pronounced increase in retail trading alignment, scenarios where retail investors are likely to benefit more from AI. Using retail‐AI alignment as a proxy for the extent to which retail investors trade based on AI signals, we show that information asymmetry declines and retail investors' trading profitability improves, whereas short sale profitability declines in high retail‐AI alignment stocks. Exogenous outages reduce the alignment between retail trading and AI‐sentiment, allowing us to draw causal inferences. Collectively, this study suggests that AI is a promising technology for narrowing the information gap in the trading of complex textual financial disclosures between investor classes with clear disparities in the ability to process public disclosures.

The engagement quality review is a key component of an audit firm's quality control system. This study leverages unique data on individual engagement quality reviewers (EQRs) to examine how previous shared working experience between EQRs and engagement partners affects audit quality. While prior research suggests that within‐firm network ties between predecessor and successor partners facilitate knowledge transfer, we find that prior shared working experience between EQRs and engagement partners is associated with lower audit quality. Mechanism tests indicate that such experience is associated with a higher likelihood of regulatory enforcement actions related to deficiencies in audit procedures, insufficient evidence, and a lack of professional skepticism. We also find that such experience corresponds with higher materiality thresholds, suggesting reduced scrutiny during audit planning and execution. Additional analyses reveal that these adverse effects primarily arise when previous shared working relationships did not produce adverse outcomes, when EQRs are not audit industry leaders, and when they face lower reputational risk. These findings are especially salient given that EQRs’ previous shared working experience with engagement partners appears to weigh heavily in EQR assignments. Overall, our study provides important insights into the implications of EQR independence and the determinants of engagement quality review effectiveness.

We examine whether and, if so, how retail consumers change their shopping in response to firm‐specific negative environmental, social, and governance (ESG) news. Using an event study methodology, we do not find significant changes in consumer foot traffic in response to negative ESG news, on average. However, the average consumer reacts negatively when such news is covered by national or global media outlets, which elevates consumer awareness. In addition, we provide evidence of the heterogeneity in responses to negative ESG news across consumer groups. Consumers in more ESG‐conscious counties, as measured by county ESG preferences, income, education, and political ideology, reduce store visits in response to negative ESG news. In contrast, consumers in the least ESG‐conscious counties increase their visits in response to negative ESG news. These opposing reactions explain the insignificant average consumer response to negative ESG news. Furthermore, the ESG‐conscious consumers' negative reaction is, at most, modest, dissipating within six weeks. Overall, our findings suggest that firms face divergent responses to ESG activities from different consumer groups, underscoring the divisive nature of ESG issues.

We study how the assignment of intellectual property rights between inventors and their employers affects innovation. Incomplete contracting theories predict that stronger employer property rights reduce the threat that employee inventors hold up their employers, thereby affecting inventor and invention outcomes. We test these predictions using a U.S. appellate court ruling that shifted the assignment of property rights from inventors to their employers. Within‐employer‐year analyses demonstrate that affected inventors are less likely to retain patent rights, assign patents to new employers, or leave their current employer, all consistent with reduced inventor ability to hold up their employers. Due to the reduced possibility of hold‐up, affected inventors’ innovations are revealed more promptly when disclosed, draw from a broader set of prior patents, and spread more to subsequent patents. If affected inventors do leave their employer, they are more likely to relocate to unaffected states. Furthermore, employers affected by the ruling are more likely to locate their inventors in agglomeration economies and alter their innovation strategy by reallocating activity across states and expanding their innovation portfolios. Our collective evidence suggests that shifting intellectual property rights to employers affects inventor and invention outcomes by reducing the threat of employee hold‐up from the employer's perspective.

This study investigates the relation between audit competition, audit quality, and auditor labor hours. Using proprietary data on auditor realization rates, we construct new measures of competition based on theory predicting that abnormal profits will quickly disappear when competition is high but persist over multiple periods when competition is low. We find consistent evidence of persistent abnormal profits among U.S. Big 4 engagements and that individual offices earn persistent abnormal returns, suggesting that the market is not perfectly competitive. Examining the consequences of lower competition, we find that profit persistence is negatively related to audit hours and positively related to audit quality. Although we are cautious about inferring causality, our findings suggest that lower competition is associated with more efficient and effective audits.

Using a novel data set of misconduct records for Finnish CEOs and directors and their parents, we explore whether corporate executives’ financial misconduct is associated with similar behavior by their parents. Controlling for various other factors of executive financial misconduct, we find that executives are significantly more likely to engage in financial misconduct, including accounting, tax, and other financial offenses, if their parents have a history of financial misconduct. This intergenerational association is stronger when the parental misconduct is more severe. Additional findings reveal that growing up in high‐misconduct municipalities and cohabiting with spouses who engage in financial misconduct are also associated with a higher likelihood of executive misconduct, indicating that such behaviors may be shaped by broader socialization processes that extend beyond the immediate family. Although our analyses do not establish causal relationships, the collective evidence presented in this study offers insights into why some corporate executives engage in misconduct while others do not.

We examine the usefulness of machine learning approaches for measuring vocal tone in corporate disclosures. We document a substantial mismatch between the widely adopted actor‐based training data underlying these approaches and speech in corporate disclosures. We find that existing models achieve near‐perfect vocal tone classification within their training domain. However, when tested on actual executive speech during conference calls, their performance declines to chance levels. We thus introduce FinVoc2Vec, a deep learning model that adapts to audio recordings of conference calls and classifies the vocal tone of executive speech significantly more accurately than chance. FinVoc2Vec estimates are associated with future firm performance and can be used to construct profitable stock portfolios. Throughout our analyses, estimates from previous vocal tone models are largely unrelated to firm performance. Our findings emphasize the importance of a domain‐specific approach to voice analysis in accounting and finance.

We investigate the dynamics of internal forecasting in multi‐location firms and the relations between forecast characteristics and investment. Using U.S. Census microdata on plant‐level growth expectations, we find that plants within multi‐location firms make forecasts that are both more certain and less accurate than those of standalone plants. We provide evidence suggesting that headquarters infers uncertainty from inter‐plant forecast disagreement, and that headquarters is able to facilitate the use of relevant information held by one plant but applicable to another. Differences between peer and focal plants' forecasts predict forecast errors and relate to investment decisions at the focal plant, suggesting that information from multiple sources is integrated into capital allocation decisions. Headquarters' heavier use of peer plant forecast information when focal plants are more uncertain likely weakens focal plant managers' incentives to consider extreme scenarios when forecasting.

We study the use of generative AI for firm‐specific financial analysis on the Seeking Alpha platform. After the initial launch of ChatGPT in November 2022, the share of AI‐generated articles rose sharply to 13.5% of all articles, then declined in late 2023 after Seeking Alpha equated the use of AI to plagiarism and announced a prohibition on its use. We organize our study around two questions: (1) Does AI use increase author productivity? and (2) does AI use have capital market consequences and ultimately affect the informational landscape? We find that authors who adopt AI become more productive, publishing more articles and covering more new firms than non‐adopters. Findings on AI article informativeness are more nuanced. On average, AI articles are less informative than human‐written articles, eliciting smaller trading volume and abnormal return responses. However, AI use leads to increased firm coverage and in turn to improved liquidity and faster price discovery. Our findings suggest that, while AI‐generated articles are currently perceived as less informative than human‐written articles, their comparatively low cost enables increased firm coverage and thereby improves the overall informational landscape.