This paper considers the puzzle of increased trading volume around information releases through the lens of canonical models of disagreement. I use a unique data set of clicks on news by key finance professionals to simultaneously measure gradual information diffusion and differences of opinion. I find that neither channel subsumes the other and that the two are complementary in generating trading volume around news events. Their relative strengths depend on the characteristics of the underlying information: gradual information diffusion matters more for straightforward news, while differences of opinion play a larger role around textually ambiguous news. (JEL G12, G14, G41, D84) Received January 12, 2020; editorial decision: November 24, 2020 by Editor: Jeffrey Pontiff.

This paper estimates the effect of news positioning on the speed of price discovery, using exogenous variation in prominent (“front‐page”) positioning of news articles on the Bloomberg terminal. Front‐page articles see 240% higher trading volume and 176% larger absolute excess returns during the first 10 minutes after publication than equally important non‐front‐page articles. Overall, the information in front‐page articles is fully incorporated into prices within an hour of publication. The response to non‐front‐page information of similar importance eventually converges but takes more than two days to be fully reflected in prices.

How is technically skilled human capital reflected in firm performance? We leverage a uniquely detailed employer–employee matched dataset to measure US firms’ technical human capital in information technology (IT), Software Engineering, Mobile Networks, Data Analysis, and Web Development. All five technical skillsets are associated with higher firm valuations. However, they negatively forecast both financial and operational performance in the future. For example, a one-standard-deviation increase in employees with IT skills corresponds to 2.2% higher Tobin’s q but predictable future returns of –10 basis points per month. Our results are stronger in tighter labor markets, in firms with more cash, and during time periods when each technical skillset is especially popular. These patterns suggest that the market expects too much from popular technologies, leading to over-valuation. Overall, our results highlight how corporate over-investment can extend to intangible capital such as skilled employees.

What drives the puzzle of market reactions to old news? Motivated by theories of correlation neglect, we conduct an experiment on finance professionals and show that even sophisticated investors have difficulty identifying old information that recombines content from multiple sources. We evaluate the market implications of this mechanism using a unique dataset of 17 million news articles from the Bloomberg terminal. Recombination of old information prompts larger price moves and subsequent reversals than direct reprints. This effect persists across news sentiment, ambiguity, and investor attention. Furthermore, while overall reactions to old information decline over time, differential reactions to recombinations increase.

How does artificial intelligence (AI) impact audit quality and efficiency? We explore this question by leveraging a unique dataset of more than 310,000 detailed individual resumes for the 36 largest audit firms to identify audit firms’ employment of AI workers. We provide a first look into the AI workforce within the auditing sector. AI workers tend to be male and relatively young and hold mostly but not exclusively technical degrees. Importantly, AI is a centralized function within the firm, with workers concentrating in a handful of teams and geographic locations. Our results show that investing in AI helps improve audit quality, reduces fees, and ultimately displaces human auditors, although the effect on labor takes several years to materialize. Specifically, a one-standard-deviation change in recent AI investments is associated with a 5.0% reduction in the likelihood of an audit restatement, a 0.9% drop in audit fees, and a reduction in the number of accounting employees that reaches 3.6% after three years and 7.1% after four years. Our empirical analyses are supported by in-depth interviews with 17 audit partners representing the eight largest U.S. public accounting firms, which show that (1) AI is developed centrally; (2) AI is widely used in audit; and (3) the primary goal for using AI in audit is improved quality, followed by efficiency.

We study the use and economic impact of AI technologies. We propose a new measure of firm-level AI investments using employee resumes. Our measure reveals a stark increase in AI investments across sectors. AI-investing firms experience higher growth in sales, employment, and market valuations. This growth comes primarily through increased product innovation. Our results are robust to instrumenting AI investments using firms' exposure to universities' supply of AI graduates. AI-powered growth concentrates among larger firms and is associated with higher industry concentration. Our results highlight that new technologies like AI can contribute to growth and superstar firms through product innovation.

This paper investigates how consulting services affect audit quality, from the perspective of knowledge- and expertise-sharing between employees. Semistructured interviews with 16 audit partners reveal that consulting expertise is used in 60–80 percent of audit engagements, with the main rationale for such collaboration being knowledge-sharing and improved audit quality. We leverage a comprehensive office-level dataset of employment profiles covering 86 percent of all employees at large U.S. public accounting firms to systematically investigate the effect of consulting employees on audit quality. We document that a one standard deviation increase in the share of consulting employees in an office results in a 2.6 percentage point reduction in restatements (a decrease of 19 percent relative to the baseline). This effect is strongest when consulting employees have skills complimentary to auditors, e.g., technical and human resources skills, and when consultants have specific industry expertise in the same industry as the audit client. Data Availability: Data from common sources, such as Audit Analytics, can be purchased from the providers. Data on firm-level measures of workforce characteristics and skills are available from the authors upon request, conditional on approval from the data provider (Cognism).