Selected Working Papers
Are Managers 'Under-the-Weather' During Earnings Conference Calls?
with Bill Francis and Wenyao Hu [SSRN]
Abstract: Earnings conference calls represent an important communication channel for investors to observe managerial behavior. We examine the impact of executive mood using seasonally-adjusted weather conditions at corporate headquarters during these calls. Using 66,724 earnings calls from 2006 to 2017, we find that managers speak more negatively and with less (more) quantitative information (uncertainty) when local weather conditions are poor. Managers also exhibit more (less) extreme negative (moderate) language and executive mood persistently influences returns after earnings calls. Our results remain significant after adding controls for investor mood, separating firms from big and small states, mediation tests, firm and executive fixed effects, and propensity score matching. Our findings suggest exogenous effects of weather significantly impact managerial mood and how management communicates in unprepared corporate disclosures, which influences firm value in a manner not fully captured by fundamental financial accounting information.
Gender and Analyst Reports
with Bill Francis, Gilna Samuel, Kate Suslava, and Daqi Xin [SSRN]
Abstract: We examine gender differences in characteristics of sell-side analyst reports. We find that female analyst reports are shorter and more readable. Consistent with an “ethical standard” explanation, the textual sentiment of female analyst reports is less optimistic. Moreover, female analyst reports contain less financially oriented content, are more long-term oriented, and are less likely to be issued in response to coverage of firm earnings announcements. Female analyst reports elicit more negative reactions than those of their male counterparts and female analysts who write shorter reports are more likely to become Institutional Investor All-Stars. Female analysts improve report readability and increase objectivity over their career more than male analysts do. Inconsistent with in-group favoritism, firms run by female CEOs have more negative reports written about them by female analysts. Our results provide evidence of gender stereotyping in the analyst profession.
Gender and Earnings Conference Calls
with Nerissa Brown, Bill Francis, Wenyao Hu, Tengfei Zhang, and Daqi Xin [SSRN]
Abstract: Using quarterly earnings conference call transcripts, we investigate gender issues in interactions between sell-side analysts and executives. We find that women are generally less “visible” on conference calls. Specifically, female analysts appear on calls less frequently, speak less, and ask fewer follow-up questions. Female analysts and executives exhibit less uncertainty, less numerical information, and fewer hesitations in their dialogue than their male counterparts. Female executives (analysts) are interrupted more (less) frequently. The above relations are mitigated with greater state-level attention to #MeToo. Further, we introduce a new measure of firm gender equality/similarity using the first principal component of nine gender-related features on conference calls. We find that gender equality is associated with positive market reaction and lower bid-ask spreads. Earnings call gender equality is positively associated with analysts’ following stock recommendation, forecast accuracy, and forecast speed, and negatively associated with analysts’ following dropped coverage and the number of revisions. Overall, our results suggest that gender equality on earnings conference calls improves market efficiency and benefits analysts’ forecasts by enhancing calls’ information content and flow.
How Does Payment for Order Flow Influence Markets? Evidence from Robinhood Crypto Token Introductions
with Tom Boulton and Michael Walz [SSRN]
Abstract: Compared to payment for order flow (PFOF) in equity and options markets, PFOF in crypto asset markets lacks transparency and generates significantly higher fees (10.5-105 times higher). Using Robinhood Crypto token introductions as a PFOF shock, we find that trading volume decreases for most crypto assets except Bitcoin and Ethereum after the token introductions. In addition, after the shock, order imbalances shift to net sales, while average trade sizes, implied spreads, and return volatility generally rise. These changes increase daily trading costs for participants by approximately $4.8 million. With the exception of order imbalance, these changes are permanent and do not reverse when tokens are delisted from Robinhood. These findings highlight the broader impact of PFOF on crypto market quality.
Special Conference Calls
with Wenyao Hu [SSRN]
Abstract: This paper investigates the determinants, content, and consequences of special conference calls. Special conference calls, meetings held by publicly-traded companies in which executives provide an overview of specific events related to the firm, have increased in frequency by 375% since the COVID-19 pandemic. Using 11,692 special calls held by U.S. public firms from 2005 to 2022, we find that hosting decisions are related to firm-level uncertainty and investor information demand. Firms that are larger, covered by more analysts, and operating in more complex business segments are more likely to hold these events. Equity markets treat special calls as incrementally informative relative to 8-Ks, with significantly positive abnormal returns and trading volume. Textual analysis of transcripts using latent Dirichlet allocation identifies four main topics: customer and growth discussions, strategic outcomes, biotech-related business updates, and clinical trial announcements. Market reactions differ across topics, with positive responses to growth and negative responses to trial announcements. A robustness analysis using BERTopic confirms these findings. Overall, special calls represent an important discretionary disclosure channel.
The Value of Data: Analyst Vs. Machine
with Stefano Bonini, Majeed Simaan, and Guofu Zhou [SSRN]
Abstract: We assess the investment value of analysts in the age of prediction machines. Using 25 years of data on over 2,500 stocks and 1.3 million analyst reports, we find that portfolios based on analyst signals deliver negative returns, while machine-driven strategies generate robust alphas exceeding 20% annually. Analysts’ information adds value only when embedded in machine forecasts, with portfolios often tilting opposite to analysts’ forecasts. Even among stocks with less publicly available information, their data has limited economic value relative to that of machines. Our results suggest that analysts’ standalone usefulness has eroded, and future portfolio construction is best served by hybrid models that combine human valuation with machine-learning capabilities.
with Bill Francis and Wenyao Hu [SSRN]
Abstract: Earnings conference calls represent an important communication channel for investors to observe managerial behavior. We examine the impact of executive mood using seasonally-adjusted weather conditions at corporate headquarters during these calls. Using 66,724 earnings calls from 2006 to 2017, we find that managers speak more negatively and with less (more) quantitative information (uncertainty) when local weather conditions are poor. Managers also exhibit more (less) extreme negative (moderate) language and executive mood persistently influences returns after earnings calls. Our results remain significant after adding controls for investor mood, separating firms from big and small states, mediation tests, firm and executive fixed effects, and propensity score matching. Our findings suggest exogenous effects of weather significantly impact managerial mood and how management communicates in unprepared corporate disclosures, which influences firm value in a manner not fully captured by fundamental financial accounting information.
Gender and Analyst Reports
with Bill Francis, Gilna Samuel, Kate Suslava, and Daqi Xin [SSRN]
Abstract: We examine gender differences in characteristics of sell-side analyst reports. We find that female analyst reports are shorter and more readable. Consistent with an “ethical standard” explanation, the textual sentiment of female analyst reports is less optimistic. Moreover, female analyst reports contain less financially oriented content, are more long-term oriented, and are less likely to be issued in response to coverage of firm earnings announcements. Female analyst reports elicit more negative reactions than those of their male counterparts and female analysts who write shorter reports are more likely to become Institutional Investor All-Stars. Female analysts improve report readability and increase objectivity over their career more than male analysts do. Inconsistent with in-group favoritism, firms run by female CEOs have more negative reports written about them by female analysts. Our results provide evidence of gender stereotyping in the analyst profession.
Gender and Earnings Conference Calls
with Nerissa Brown, Bill Francis, Wenyao Hu, Tengfei Zhang, and Daqi Xin [SSRN]
Abstract: Using quarterly earnings conference call transcripts, we investigate gender issues in interactions between sell-side analysts and executives. We find that women are generally less “visible” on conference calls. Specifically, female analysts appear on calls less frequently, speak less, and ask fewer follow-up questions. Female analysts and executives exhibit less uncertainty, less numerical information, and fewer hesitations in their dialogue than their male counterparts. Female executives (analysts) are interrupted more (less) frequently. The above relations are mitigated with greater state-level attention to #MeToo. Further, we introduce a new measure of firm gender equality/similarity using the first principal component of nine gender-related features on conference calls. We find that gender equality is associated with positive market reaction and lower bid-ask spreads. Earnings call gender equality is positively associated with analysts’ following stock recommendation, forecast accuracy, and forecast speed, and negatively associated with analysts’ following dropped coverage and the number of revisions. Overall, our results suggest that gender equality on earnings conference calls improves market efficiency and benefits analysts’ forecasts by enhancing calls’ information content and flow.
How Does Payment for Order Flow Influence Markets? Evidence from Robinhood Crypto Token Introductions
with Tom Boulton and Michael Walz [SSRN]
Abstract: Compared to payment for order flow (PFOF) in equity and options markets, PFOF in crypto asset markets lacks transparency and generates significantly higher fees (10.5-105 times higher). Using Robinhood Crypto token introductions as a PFOF shock, we find that trading volume decreases for most crypto assets except Bitcoin and Ethereum after the token introductions. In addition, after the shock, order imbalances shift to net sales, while average trade sizes, implied spreads, and return volatility generally rise. These changes increase daily trading costs for participants by approximately $4.8 million. With the exception of order imbalance, these changes are permanent and do not reverse when tokens are delisted from Robinhood. These findings highlight the broader impact of PFOF on crypto market quality.
Special Conference Calls
with Wenyao Hu [SSRN]
Abstract: This paper investigates the determinants, content, and consequences of special conference calls. Special conference calls, meetings held by publicly-traded companies in which executives provide an overview of specific events related to the firm, have increased in frequency by 375% since the COVID-19 pandemic. Using 11,692 special calls held by U.S. public firms from 2005 to 2022, we find that hosting decisions are related to firm-level uncertainty and investor information demand. Firms that are larger, covered by more analysts, and operating in more complex business segments are more likely to hold these events. Equity markets treat special calls as incrementally informative relative to 8-Ks, with significantly positive abnormal returns and trading volume. Textual analysis of transcripts using latent Dirichlet allocation identifies four main topics: customer and growth discussions, strategic outcomes, biotech-related business updates, and clinical trial announcements. Market reactions differ across topics, with positive responses to growth and negative responses to trial announcements. A robustness analysis using BERTopic confirms these findings. Overall, special calls represent an important discretionary disclosure channel.
The Value of Data: Analyst Vs. Machine
with Stefano Bonini, Majeed Simaan, and Guofu Zhou [SSRN]
Abstract: We assess the investment value of analysts in the age of prediction machines. Using 25 years of data on over 2,500 stocks and 1.3 million analyst reports, we find that portfolios based on analyst signals deliver negative returns, while machine-driven strategies generate robust alphas exceeding 20% annually. Analysts’ information adds value only when embedded in machine forecasts, with portfolios often tilting opposite to analysts’ forecasts. Even among stocks with less publicly available information, their data has limited economic value relative to that of machines. Our results suggest that analysts’ standalone usefulness has eroded, and future portfolio construction is best served by hybrid models that combine human valuation with machine-learning capabilities.
Selected Publications
- Riskless Principal Trades in Corporate Bond Markets (with Louis Craig and Larry Harris) Journal of Fixed Income (2026) Vol 35., No 3., pp. 18-29
- Buy the Dip? (with Stefano Bonini and Majeed Simaan) European Financial Management (2024) Vol. 30, No. 4, pp. 2033-2070 [SSRN]
- Do Green Business Practices License Self-Dealing or Prime Prosociality? Cross-Domain Evidence from Environmental Concern Triggers (with Melanie I. Millar, Mason C. Snow, and Roger M. White) Accounting, Organizations, and Society (2024) Vol. 113, 101497
- Fixed Income Conference Calls (with Gus Ddoi.org/10.3905/jfi.2025.1.219e Franco, Da Xu, and Zhiwei (Vivi) Zhu) Journal of Accounting and Economics (2023) Vol. 75, No. 1, 101518 [SSRN]
- Does Native Country Turmoil Predict Immigrant Workers’ Honesty in Markets? (with Roger M. White) Journal of Economic Behavior & Organization (2022) Vol. 197, pp. 150-164 [SSRN]
- Do Sell-Side Analysts Play a Role in Hedge Fund Activism? Evidence from Textual Analysis (with Huimin (Amy) Chen) Contemporary Accounting Research (2022) Vol. 39, No.3, pp. 1583-1614 [SSRN]
- Do Sin Taxes Spur Cheating in Interpersonal Exchange? (with David Kenchington, Jared D. Smith, and Roger M. White) Accounting, Organizations, and Society (2022) Vol. 96, 101281 [SSRN]
- Unbridled Spirit: Illicit Markets for Bourbon Whiskey (with Conor J. Lennon) Journal of Economic Behavior & Organization (2021) Vol. 191, pp. 1025-1045 [SSRN]
- Which Buy-Side Institutions Participate in Public Earnings Conference Calls? Implications for Capital Markets and Sell-Side Coverage (with Andrew C. Call and Nathan Y. Sharp) Journal of Corporate Finance (2021) Vol. 68, 101964 [SSRN] [Internet Appendix: Participant Classifications] [Internet Appendix: Participant Data]
- Investor Awareness or Information Asymmetry? Wikipedia and IPO Underpricing (with Thomas J. Boulton, Bill Francis, and Daqi Xin) Financial Review (2021) Vol. 56, No. 3, pp. 535-561 [SSRN] [Internet Appendix]
- Bank Loan Renegotiation and Credit Default Swaps (with Brian Clark, James Donato, and Bill Francis) Journal of Banking & Finance (2020) Vol. 151, 105936 [SSRN]
- The Dark Side of Individual Blockholder Philanthropy (with Roger M. White) Financial Management (2020) Vol. 49, No. 3, pp. 741-767 [SSRN]
- Bulk Volume Classification and Information Detection (with Marios A. Panayides and Jared D. Smith) Journal of Banking & Finance (2019) Vol. 103, pp. 113-129 [SSRN] [Internet Appendix]
- Angels or Sharks? The Role of Personal Characteristics in Angel Investment Decisions (with Thomas J. Boulton and Pengcheng Zhu) Journal of Small Business Management (2019) Vol. 57, No. 4, pp. 1280-1303 [SSRN]