May 2024

Do Risk Disclosures Matter When It Counts? Evidence from the Swiss Franc Shock
Journal of Accounting Research, Volume 59(1)
(Joint with Luzi Hail and David Oesch)
[Abstract] We examine the relation between disclosure quality and information asymmetry among market participants following an exogenous shock to macroeconomic risk. In 2015 the Swiss National Bank abruptly announced that it would abandon the longstanding minimum euro-Swiss franc exchange rate. We find evidence suggesting that firms with more transparent disclosures regarding their foreign exchange risk exposure ex ante exhibit significantly lower information asymmetry ex post. The information gap in bid-ask spreads appears within 30 minutes of the announcement and persists for two weeks, during which new information gradually substitutes for past disclosures. We validate the information dynamics of past risk disclosures with three field surveys: (1) Sell-side analysts emphasize the importance of existing (risk) disclosures in evaluating the translational and transactional effects of the currency shock. (2) Lending banks’ credit officers rely on past disclosures as the primary information source available for smaller (unlisted) firms in the immediate aftermath of the shock. (3) Investor-relations managers use existing financial filings as a key resource when communicating with external stakeholders. The results suggest that historical disclosures help investors attenuate information asymmetry in light of unexpected news.

Financial Transparency of Private Firms: Evidence from a Randomized Field Experiment
Forthcoming at the Journal of Accounting Research
(Joint with Joachim Gassen)
[Abstract] This paper examines why private firms choose to be financially transparent or opaque by conducting a field experiment with more than 25,000 firms in Germany. We inform a randomly chosen set of firms about a disclosure option that allows eligible firms to restrict access to their otherwise publicly available financial statements. We also vary the messaging in subtle ways to induce experimental variation in the probability that firms take transacting (capital providers or customers and suppliers) versus non transacting stakeholders (competitors or general interest parties) into consideration when making their filing decision. Based on each firm’s actual filing decision, we find that treated firms are 15% more likely to restrict access to their financial statements. This intention-to-treat effect is persistent and concentrated among firms that should derive lower net benefits from disclosure (smaller, more mature firms in less capital intensive industries). These findings indicate that informational constraints affect firms’ disclosure practice. Additionally, we show that the treatment effect is almost 40% larger for firms that have a higher, exogenously induced, probability of considering non-transacting stakeholders when making their disclosure decision. We also provide suggestive evidence that disclosure requirements put an undue burden on very small private firms.

---

---

How Do Consumers Use Firm Disclosure? Evidence from a Randomized Field Experiment
Available on SSRN
(Joint with Sinja Leonelli, Thomas Rauter and Gurpal Sran)
[Abstract] We combine a large-scale field experiment with a customized survey to study whether and how consumers use firm disclosure. In a sample of more than 24,000 U.S. households, we first establish several stylized facts: (i) the average consumer has a moderate preference to purchase from ESG-responsible firms; (ii) consumers typically have no preference for more or less profitable firms; (iii) consumers rarely consult ESG reports and virtually never use financial reports to inform their purchase decisions. In our field experiment, we then inform households about real firm-disclosed profitability and ESG activities through seven randomized information treatments. Consumers increase their purchase intent when exogenously presented with firm-disclosed positive signals about environmental, social, and—to a lesser extent—governance activities. Full ESG reports only have an impact on consumers who choose to view them, whereas financial reports and earnings information do not have an effect. After the experiment, consumers increase their actual product purchases, but these effects are small, short-lived, and only materialize for viewed ESG reports and positive social signals. Through a follow-up survey, we provide explanations for why consumers (do not) change their shopping behavior after our information experiment.

Decoding Social Disclosure Decisions: A Field Experiment with Workforce Diversity Data
Available Soon
(Joint with Jung Ho Choi and Maureen McNichols)
[Abstract] In recent years, U.S. public companies have increasingly begun to voluntarily disclose official workforce diversity data (i.e., EEO-1 reports), which they previously only confidentially filed with the U.S. Equal Employment Opportunity Commission (EEOC). To understand the factors leading these corporations to release this information publicly, we conduct a field experiment by reaching out to Investor Relations (IR) and Human Resources (HR) personnel at about 4,000 large US firms that currently do not publicly disclose their EEO-1 reports. We experimentally vary the information content of our requests and find that companies are more likely to respond when directly considering investors’ rather than employees’ informational needs. On the other hand, we do not find any evidence that companies are more likely to respond when directly considering S&P 100 firms’ disclosure decisions. We also show that IR departments are significantly more involved in this disclosure process relative to HR departments and that a temporary regulatory action by the Office of Federal Contractor Compliance Programs does not lead to continuous workforce diversity disclosures. Our follow-up survey indicates that companies consider both shareholder welfare and equity value, as well as the potential litigation costs, when making social disclosure decisions. Taken together, our results are relevant in the current regulatory debate about the single (shareholder) and double (stakeholder) materiality of non-financial disclosure.

---

---

Bloated Disclosures: Can ChatGPT Help Investors Process Information?
Available on SSRN
(Joint with Alex Kim and Valeri Nikolaev)
[Abstract] Generative AI tools such as ChatGPT can fundamentally change the way investors process information. We probe the economic usefulness of these tools in summarizing complex corporate disclosures using the stock market as a laboratory. The unconstrained summaries are dramatically shorter, often by more than 70% compared to the originals, whereas their information content is amplified. When a document has a positive (negative) sentiment, its summary becomes more positive (negative). More importantly, the summaries are more effective at explaining stock market reactions to the disclosed information. Motivated by these findings, we propose a measure of information "bloat." We show that bloated disclosure is associated with adverse capital markets consequences, such as lower price efficiency and higher information asymmetry. Finally, we show that the model is effective at constructing targeted summaries that identify firms' (non-)financial performance and risks. Collectively, our results indicate that generative language modeling adds considerable value for investors with information processing constraints.

From Transcripts to Insights: Uncovering Corporate Risks Using Generative AI
Available on SSRN
(Joint with Alex Kim and Valeri Nikolaev)
[Abstract] We explore the value of generative AI tools, such as ChatGPT, in helping investors uncover dimensions of corporate risk. We develop and validate firm-level measures of risk exposure to political, climate, and AI-related risks. Using the GPT 3.5 model to generate risk summaries and assessments from the context provided by earnings call transcripts, we show that GPT-based measures possess significant information content and outperform the existing risk measures in predicting (abnormal) firm-level volatility and firms’ choices such as investment and innovation. Importantly, information in risk assessments dominates that in risk summaries, establishing the value of general AI knowledge. We also find that generative AI is effective at detecting emerging risks, such as AI risk, which has soared in recent quarters. Our measures perform well both within and outside the GPT’s training window and are priced in equity markets. Taken together, an AI-based approach to risk measurement provides useful insights to users of corporate disclosures at a low cost.

Financial Statement Analysis with Large Language Models
Available on SSRN
(Joint with Alex Kim and Valeri Nikolaev)
[Abstract] We investigate whether an LLM can successfully perform financial statement analysis in a way similar to a professional human analyst. We provide standardized and anonymous financial statements to GPT4 and instruct the model to analyze them to determine the direction of future earnings. Even without any narrative or industry-specific information, the LLM outperforms financial analysts in its ability to predict earnings changes. The LLM exhibits a relative advantage over human analysts in situations when the analysts tend to struggle. Furthermore, we find that the prediction accuracy of the LLM is on par with the performance of a narrowly trained state-of-the-art ML model. LLM prediction does not stem from its training memory. Instead, we find that the LLM generates useful narrative insights about a company’s future performance. Lastly, our trading strategies based on GPT’s predictions yield a higher Sharpe ratio and alphas than strategies based on other models. Taken together, our results suggest that LLMs may take a central role in decision-making.

Who Falls Prey to the Wolf of Wall Street? Investor Participation in Market Manipulation
Available on SSRN, NBER Working Paper
(Joint with Christian Leuz, Steffen Meyer, Eugene Soltes and Andreas Hackethal)
[Abstract] Price distortions created by so-called “pump-and-dump” schemes are well known, but relatively little is known about the investors in these frauds. By examining 470 “pump-and-dump” schemes and a large data set of trading records for over 110,000 individual investors from a major German bank, we provide comprehensive evidence on the participation rate, magnitude of the investments, the losses, and the characteristics of the individuals who invest in such schemes. Participation is quite common with nearly 8% of active retail investors participating in at least one “pump-and-dump” losing on average nearly 30%. Next, we identify several distinct types among participating investors, some of which (i.e., speculating day trader) should not be viewed as falling prey to the schemes. Recognizing this heterogeneity is key when designing investor protections because we find investor types respond differently to market manipulation. We also show that portfolio composition and past trading behavior better explain scheme participation than demographics. Lastly, we document longer lasting effects on participating investors that go beyond the immediate financial losses.

Do Conflict of Interests Disclosures Work? Evidence from Citations in Medical Journals
Available on SSRN, NBER Working Paper
(Joint with Christian Leuz, Anup Malani and Laszlo Jakab)
[Abstract] Financial ties between drug companies and medical researchers are thought to bias studies published in medical journals. To enable readers to account for such bias, most medical journals require authors to disclose potential conflicts of interest. We examine whether disclosure reduces article citations, indicating a discount. A challenge to estimating this effect is selection as drug companies may seek out higher quality authors. Our analysis confirms this positive association. Including observable controls for article and author quality attenuates but does not eliminate this relation. We perform three tests. First, we show that the positive association is weaker for review articles, which are more susceptible to bias. Second, we examine article recommendations to family physicians among articles that are a priori more homogenous in quality. We find a significantly negative association between disclosure and expert recommendations, consistent with discounting. Third, we conduct an analysis within author and article, exploiting journal policy changes that result in conflict disclosure by an author. We examine the effect of this disclosure on citations to a previously published article by the same author. This analysis reveals a negative citation effect. Overall, our evidence is consistent with the notion that other researchers discount articles with disclosed conflicts.

---

---

How Private Companies Attract the Market's Attention
(Joint with Thomas Bourveau and Matthias Breuer)

The Relevance of Peer Information for Private Firms: Evidence from a Field Experiment
(Joint with Jeppe Christoffersen, Mike Minnis, Thomas Plenborg and Morten Seitz)

Cash Versus Accrual Accounting: Evidence from a Randomized Field Experiment
(Joint with Pietro Bonetti and Matthias Breuer)

Corporate Footprints
(Joint with Thomas Barry, Matthias Breuer, Rongchen Li, Thomas Rauter and Harm Schütt)

Large Language Models and Financial Reporting Oversight
(Joint with Alex Kim, Valeri Nikolaev and Irene Tan)

---