We document strikingly opposite time-series patterns of analyst forecast errors (FEs) and associated market reactions, illustrating that analyst forecasts have become a less useful benchmark of the market’s earnings expectations in recent years. The mean FE has increased from negative one to two cents in the 1990s to positive one to two cents in the 2010s, whereas average earnings announcement returns have declined from 0.30% in the 1990s to −0.30% in the 2010s, turning negative in the past 17 years. Underlying the time-series pattern of increasing FEs is a secular trend where firms move away from just meeting or beating, to which the market reaction has become increasingly negative, toward a large beat, while the frequency of meeting or beating the consensus analyst forecast remains stable during the same period. We develop a parsimonious predictive model of earnings surprises based on peer and past analysts’ FEs and find that our predicted FE closely mirrors reported FE, with the average value hovering around one to two cents in most years of the past two decades. The market reaction to “around zero” unexpected FE (FE minus predicted FE) is indistinguishable from zero over time, suggesting that our model serves as a good benchmark of the market’s expectation. Our evidence has broad implications for appropriate earnings benchmarking, for the disappearing discontinuity of the earnings surprise distribution around zero, for earnings management to beat analysts’ forecasts, for empirical designs when examining the earnings-return relation, and for the disappearing earnings announcement premium.

Extant literature documents that aggregate accruals positively predict future market returns and attributes this relation to either changes in discount rates or systematic earnings management. We offer an alternative explanation: aggregate merger and acquisition (M&A) activity drives this relation. M&A activity affects the magnitude of accruals, which in turn drives the market return predictability of aggregate accruals. We find that the ability of both aggregate accruals and discretionary aggregate accruals (a measure of systematic earnings management) to predict market returns disappears after controlling for aggregate M&A activity. Furthermore, aggregate M&A activity predicts future market returns, consistent with a price response to improvements in macroeconomic outcomes due to aggregate M&A activity.

We study whether relative power in the CEO-CFO relationship influences CEO compensation. To operationalize relative power of a CEO over a CFO, we define CFO co-option as the appointment of a CFO after a CEO assumes office. We find that CFO co-option is associated with a CEO pay premium of about 10%, which is concentrated more in the early years of the co-opted CFO’s tenure and in components of compensation that vary with the achievement of analyst-based earnings targets. Our evidence also indicates that a primary channel through which CEO power over a co-opted CFO yields the achievement of earnings targets is the use of earnings management to inflate earnings. Co-opted CFOs rely primarily on using discretionary accruals to manage earnings prior to the Sarbanes-Oxley regulatory intervention and switch to real activities manipulation afterwards. The evidence thus suggests that the form of earnings management depends on costs imposed on the CFO to inflate earnings.

Drawing from established models in psychology, we examine whether managers’ response latency—the pause before a manager begins answering—during conference call question-and-answer sessions serves as a cue to management deception. We provide a three-pronged validation of response latency as a deception cue. First, managerial statements directly quoted in settled securities class action lawsuits exhibit longer response latency, whereas statements quoted in dismissed lawsuits do not. Second, response latency predicts linguistic patterns consistent with deceptive communication theory. Third, analysts respond to high-latency answers with greater skepticism, as evidenced by challenging follow-ups. We then examine whether managers’ high-latency responses predict capital market distortions. We find that a more positive disclosure tone following high response latency leads to an immediate positive stock price reaction that reverses over time, consistent with managers communicating distorted information. Additional analyses suggest that while response latency also increases with observable non-deceptive sources of cognitive strain, such as uncertainty or question dynamics, these factors do not explain our results. Overall, our findings suggest that response latency captures meaningful information about management deception in interactive disclosures.

In this paper, we document a previously unknown cost of stock splits: failure to sufficiently beat earnings targets and its associated capital markets punishment. We show that both firms’ earnings announcement returns and likelihood of beating analysts’ expectations by at least two cents decline post-stock split. This patterned decline in both split activity and post-split returns only occurs for publicly-listed firms, whereas abnormal returns for closed-end funds do not consistently vary over time. Overall, the results suggest that declining signaling benefits and increasing costs led to fewer stock splits in recent years.

We investigate two competing hypotheses for why clustering of trades occur around short sale disclosure in the UK: herding- and information-based trading. First, we use a matched sample of firms with similar short interest to compare firms with and without short selling disclosure and find that firm-level short interest exhibits a much smaller reversal for disclosure stocks than for matched non-disclosure stocks. A smaller reversal for disclosure stocks is consistent with short position disclosure inducing investor herding behavior and thus making short interest more persistent. Second, we explore the role of corporate news on the premise that information-based trading is related to news about firm fundamentals. We find that short sale disclosure occurs with similar frequency across the pre-earnings announcement, post-earnings announcement, and no-information windows, suggesting that information shocks related to firm fundamentals are not a major factor of short sale disclosure clustering. More importantly, the clustering of short sale disclosure does not vary significantly between good and bad earnings news. Overall, the evidence is most consistent with investor herding behavior for the clustering of short sale disclosure.

We decompose daily, signed firm-level returns into five components: private information revealed through trading, public information, market-wide information, noise and discount rate. The private information component of returns positively predicts returns and fundamentals and spans the momentum factor. However, momentum does not span the private information factor. Around earnings announcements, private information absorbs significantly more alpha from post-earnings announcement drift (PEAD) than the public information component. Noise predicts return continuations (reversals) when aligned with (opposed to) private information. Overall, our results suggest that gradual diffusion of private information drives more of momentum and PEAD than underreaction to public information.

We investigate the effects that mutual fund liquidity requirements have on fragility. In 2018, SEC Rule 22e-4 restricted ownership of illiquid securities in funds. As expected, post-rule, funds hold more liquid securities. Firms issuing illiquid securities face higher costs due to a smaller investor pool. However, higher liquidity does not ameliorate adverse shocks. Facing outflows, funds maintain cash levels and sell illiquid securities. This is because liquidity requirements are not sufficiently countercyclical: funds must maintain cash even when they should use it to mitigate flow pressures. Hence, outflows force funds to sell more illiquid securities post-rule change, unintentionally increasing fragility.

We examine the consequences of divergence between a firm’s principal disclosure outlets across mandatory and voluntary venues. We develop a decomposable measure of semantic distance using sentence-level transformer embeddings through a Blinder-Oaxaca framework. Using the semantic wedge between disclosures in the 10-K’s required MD&A and the voluntary disclosures made during the earnings call’s presentation, we find larger gaps between these disclosures result in a negative drift in returns following the 10-K filing. Decomposing the semantic distance into differences in topic coverage (i.e., topic selection and emphasis) and framing (i.e., how the two channels characterize the same topics) reveals that the drift is attributable to selective differences in topic coverage across disclosure channels. Our evidence indicates that cross-channel omission is a priced but slowly incorporated signal: what a firm declines to say “twice” is more informative than how it reframes what it repeats.

I document that mandated performance disclosure increased managerial risk-taking, resulting in significant unintended consequences and agency conflicts. After the SEC required that all mutual funds disclose a self-selected primary benchmark, I find that most fund managers chose a benchmark that was not the best fit index. Furthermore, I provide evidence that actively managed funds increased risk-taking relative to their disclosed benchmarks in response to the disclosure change. I also find that the mandated disclosure requirement exacerbated the well-documented tendency of managers who underperform during the first half of the year to increase the risks they take in the second half of the year.