Do political connections add value to audit firms? Evidence from IPO audits in China

We examine the role of cash flow from operations (CFO) in chief executive officer (CEO) cash compensation. We predict that CFO is contract‐relevant in the presence of earnings, and more so when (1) the quality of earnings relative to the quality of CFO as a measure of performance is low and (2) the need for CFO as a financing source is high. Our analysis is motivated principally by normative arguments and anecdotes from financial disclosures linking CFO to managerial effort and contracts, notwithstanding the traditional role of earnings in performance measurement. We find that the weight of CFO in the compensation model is positive and significant in the presence of earnings and stock returns. We also find that the relative quality of CFO compared with that of earnings has a positive (negative) impact on the weight of CFO (earnings). We further find that the relative weight of CFO is enhanced substantially when enterprise activities crucially depend on internally generated cash flow. These findings are unaltered when we include CEO age, firm size, and risk in the model and allow the coefficients to vary across industries.

We develop a mixed‐duopoly model in which a private firm competes against a state‐owned enterprise (SOE) who cares about social welfare and is privately informed about market demand. When the SOE's social concerns are sufficiently important and when the market competitiveness is sufficiently low, the SOE commits to fully disclose its private information. Otherwise, the SOE commits to withhold its private information. When the disclosure equilibrium prevails, the private firm can be more profitable competing against an SOE than against another private firm. In this mixed‐duopoly setting, the equilibrium social welfare is maximized when the SOE puts a positive weight on both social welfare and its own profit. Our analysis has further implications for both mandatory disclosure and market entry.

This paper examines how overconfidence affects the properties of management forecasts. Using both the “over‐optimism” and “miscalibration” dimensions of overconfidence to generate our predictions, we examine three research questions. First, we examine whether overconfidence increases the likelihood of issuing a forecast. Second, we examine whether overconfidence increases the amount of optimism in management forecasts. Third, we examine whether overconfidence increases the precision of the forecast. Using both options‐ and press‐based measures to proxy for individual overconfidence, we find support for all three research questions.

This paper provides empirical evidence that underreaction in financial analysts' earnings forecasts increases with the forecast horizon, and offers a rational economic explanation for this result. The empirical portion of the paper evaluates analysts' responses to earnings‐surprise and other earnings‐related information. Our empirical evidence suggests that analysts' earnings forecasts underreact to both types of information, and the underreaction increases with the forecast horizon. The paper also develops a theoretical model that explains this horizon‐dependent analyst underreaction as a rational response to an asymmetric loss function. The model assumes that, for a given level of inaccuracy, analysts' reputations suffer more (less) when subsequent information causes a revision in investor expectations in the opposite (same) direction as the analyst's prior earnings‐forecast revision. Given this asymmetric loss function, underreaction increases with the risk of subsequent disconfirming information and with the disproportionate cost associated with revision reversal. Assuming that market frictions prevent prices from immediately unraveling these analyst underreac‐tion tactics, investors buying (selling) stock on the basis of analysts' positive (negative) earnings‐forecast revisions also benefit from analyst underreaction. Therefore, the asymmetric cost of forecast inaccuracy could arise from rational investor incentives consistent with a preference for analyst underreaction. Our incentives‐based explanation for underreaction provides an alternative to psychology‐based explanations and suggests avenues for further research.

This study extends extant research on the determinants of financial analyst bias by examining the role that political incentives play. Using a series of scheduled provincial political events in China, we document that analysts are significantly more likely to issue favorable recommendations or revise their recommendations upward during political event periods, and the effect of political events on optimism is larger for analysts employed by brokerage firms affiliated with politicians. Cross‐sectional evidence suggests that the impact of political events on analyst optimism is concentrated in those provinces where capital market development is a more important performance indicator for politicians or where the incumbent politicians face a pending promotion. Stock return analyses reveal that favorable recommendations issued during political event periods are significantly less profitable in the long run and are less credible according to investor perceptions. Reinforcing our main evidence, we also find that financial analysts are more likely to issue optimistic earnings forecasts during political event periods. Collectively, our results imply that political incentives distort analyst opinions and political‐economic factors affect the corporate information environment in China.

Since the Split Share Structure Reform took effect in China in 2005, holders of nontradable shares (controlling shareholders) have had to negotiate with holders of tradable shares (minority shareholders) to gain the liquidity right. In a typical deal reached, the controlling shareholder agrees to pay share compensation to minority shareholders and, in many cases, also pledges to meet a specific firm performance target (performance commitments). Using this reform setting, we examine the impact of performance commitments on earnings management behavior, and find the following results. First, less profitable firms have greater incentives to make performance commitments that help to reduce the share compensation that controlling shareholders have to pay. Second, firms entering into such commitments engage in earnings management to meet the promised performance target when actual performance falls short, and firms facing greater default costs tend to manage earnings more aggressively. Third, depending on the performance metric stipulated in the commitment contract, firms employ varying methods to manage earnings. We also find that firms that rely on earnings management to meet their performance targets display inferior performance in the postcommitment years relative to firms that do not. Overall, our evidence is consistent with performance commitment contracts (with costly defaults) between a firm's controlling and minority shareholders causing incentives for earnings management.

We develop FinBERT, a state‐of‐the‐art large language model that adapts to the finance domain. We show that FinBERT incorporates finance knowledge and can better summarize contextual information in financial texts. Using a sample of researcher‐labeled sentences from analyst reports, we document that FinBERT substantially outperforms the Loughran and McDonald dictionary and other machine learning algorithms, including naïve Bayes, support vector machine, random forest, convolutional neural network, and long short‐term memory, in sentiment classification. Our results show that FinBERT excels in identifying the positive or negative sentiment of sentences that other algorithms mislabel as neutral, likely because it uses contextual information in financial text. We find that FinBERT's advantage over other algorithms, and Google's original bidirectional encoder representations from transformers model, is especially salient when the training sample size is small and in texts containing financial words not frequently used in general texts. FinBERT also outperforms other models in identifying discussions related to environment, social, and governance issues. Last, we show that other approaches underestimate the textual informativeness of earnings conference calls by at least 18% compared to FinBERT. Our results have implications for academic researchers, investment professionals, and financial market regulators.