Does Corporate Purpose Conflict With Shareholder Returns?
Vivian W. Fang, Alex Edmans, Lijun (Gillian) Lei
This paper studies the link between purpose statements and future shareholder returns. We use deep neural networks to identify a “purpose statement” as one that views stakeholder value as a means to ultimately improve shareholder value, in contrast to “purpose-like” statements that put shareholders and stakeholders on an equal footing. A value-weighted portfolio of companies with purpose statements earns a 0.3% monthly alpha above characteristics benchmarks; a long-short portfolio that buys firms with purpose statements and sells those with purpose-like statements earns a 0.28% monthly alpha. Purpose statements are positively linked to future earnings surprises, suggesting a channel through which they lead to higher stock returns, but purpose-like statements are not. They are positively associated with unvested but not vested CEO equity ownership, suggesting that long-term equity more effectively aligns managers with shareholders’ long-term interests.
Frank Zhou, Bong-Geun Choi, Jung Ho Choi, Maureen McNichols
Given the essential role accountants play in the economy, the dynamics of accountant supply and demand have significant consequences for businesses, their stakeholders, and the functioning of capital markets. This paper analyzes the employment patterns, career trajectories, and earnings of college graduates in business fields to understand these dynamics while overcoming empirical challenges: unobservable factors, sectoral movements, and equilibrium outcomes. Utilizing a large dataset of resume information, we specify and estimate a dynamic labor market equilibrium model that jointly characterizes (i) individuals’ choices among six sectors: Big 4, Non-Big 4, Internal Accountants, Finance & Consulting, Technology, and Others, considering current and future wage and non-wage attributes; and (ii) employers’ within- and across-sector oligopsonistic competition for labor. Our estimates uncover the magnitude of entry barriers, switching costs, wage markdowns, and marginal product of labor across sectors, which are essential hidden factors influencing accountants’ career choices. Our counterfactual analysis suggests a competitive labor market increases the labor market share of external accounting despite reducing that of internal accounting.
Enhancing Corporate Communication with AI-summarized Disclosures
Gwen Yu, T.J. Wong, Yang Yi, Shubo Zhang, Tianyu Zhang
We conduct a field experiment where we provide investors with AI-generated summaries of annual reports during virtual conference calls. We find that providing annual report summaries increases both the quantity and quality of investor engagement during the calls. Treatment firms experience a 46% increase in the number of investor questions, relative to control firms. Investors’ questions for treatment firms are (i) longer, (ii) contain more numerical content, and (iii) address issues requiring higher-level cognitive processing. Questions are more likely to address topics covered in the summaries, particularly when investors are less experienced or when the annual reports are longer and contain negative sentiment. Additionally, management’s responses during calls of treatment firms become longer and more detailed. The capital markets also react more strongly to these calls. The findings suggest that AI-generated summaries can help investors process annual reports’ content by highlighting key information and enhancing the reports’ saliency.
Relative Performance Feedback With Bounded Peer Comparison: Evidence From A Field Experiment
Henry Eyring
This study uses a field experiment set in online professional education to examine whether relative performance feedback (RPF) has a more positive effect when the scope of ranks provided is restricted. Providing RPF on a randomly selected subset of measures yields higher performance than providing no RPF or RPF on the full set of measures. Thus, there are negative returns to displaying additional measures for peer comparison. The benefit of constraining RPF is consistent with theory that information overload impedes learning; when RPF shows only a subset of ranks, this yields roughly twice the increase in accuracy. The results also document cross-sectional variation. When the subset of ranks shown are more predominantly the recipient’s own lower
ranks, this boosts (damps) grades for higher (lower) initial performers, and lower performers reduce their actions by roughly 18%. These results provide evidence on whether and how to restrict access to measures for peer comparison.
The Information Content Of 10-K And 10-Q Amendments
MaryJane Rabier, Sarah McVay, Simmi Mookerjee
Amendments to Forms 10-K and 10-Q—filed as 10-K/As and 10-Q/As—occur in 12.8 percent of firm-years between 2004 and 2023. These SEC filings are largely overlooked as indicators of financial reporting quality because they often occur as a result of routine compliance updates or minor technical corrections, although they also include more substantive corrections. We find that firms issuing any 10-K/Q amendment are significantly more likely to misstate future financial statements. We use large language models (LLMs) to classify the 10-K/Q amendments according to whether the underlying correction appears more or less material. Amendments tied to more material issues are most predictive of future non-reliance restatements. These amendments complement, rather than substitute for, Audit Analytics data (considering both non-reliance restatements and immaterial error revisions), with an overlap of only 18 percent, growing the overall proportion of firm-year revisions from 6.9 percent to 17.4 percent. These findings extend existing frameworks of reporting quality and highlight amendments as a valuable information source.