Knowledge that Transforms

In many areas, practitioners seek to use observational data to learn a treatment assignment policy that satisfies application‐specific constraints, such as budget, fairness, simplicity, or other functional form constraints. For example, policies may be restricted to take the form of decision trees based on a limited set of easily observable individual characteristics. We propose a new approach to this problem motivated by the theory of semiparametrically efficient estimation. Our method can be used to optimize either binary treatments or infinitesimal nudges to continuous treatments, and can leverage observational data where causal effects are identified using a variety of strategies, including selection on observables and instrumental variables. Given a doubly robust estimator of the causal effect of assigning everyone to treatment, we develop an algorithm for choosing whom to treat, and establish strong guarantees for the asymptotic utilitarian regret of the resulting policy.

We consider estimation and inference on average treatment effects under unconfoundedness conditional on the realizations of the treatment variable and covariates. Given nonparametric smoothness and/or shape restrictions on the conditional mean of the outcome variable, we derive estimators and confidence intervals (CIs) that are optimal in finite samples when the regression errors are normal with known variance. In contrast to conventional CIs, our CIs use a larger critical value that explicitly takes into account the potential bias of the estimator. When the error distribution is unknown, feasible versions of our CIs are valid asymptotically, even when <a:math xmlns:a="http://www.w3.org/1998/Math/MathML" display="inline"> <a:msqrt> <a:mi>n</a:mi> </a:msqrt> </a:math>‐inference is not possible due to lack of overlap, or low smoothness of the conditional mean. We also derive the minimum smoothness conditions on the conditional mean that are necessary for <c:math xmlns:c="http://www.w3.org/1998/Math/MathML" display="inline"> <c:msqrt> <c:mi>n</c:mi> </c:msqrt> </c:math>‐inference. When the conditional mean is restricted to be Lipschitz with a large enough bound on the Lipschitz constant, the optimal estimator reduces to a matching estimator with the number of matches set to one. We illustrate our methods in an application to the National Supported Work Demonstration.

We study repeated independent Blackwell experiments; standard examples include drawing multiple samples from a population, or performing a measurement in different locations. In the baseline setting of a binary state of nature, we compare experiments in terms of their informativeness in large samples. Addressing a question due to Blackwell (1951), we show that generically an experiment is more informative than another in large samples if and only if it has higher Rényi divergences. We apply our analysis to the problem of measuring the degree of dissimilarity between distributions by means of divergences. A useful property of Rényi divergences is their additivity with respect to product distributions. Our characterization of Blackwell dominance in large samples implies that every additive divergence that satisfies the data processing inequality is an integral of Rényi divergences.

This study uses experimental economic markets to investigate the impact of risk and the potential for loss on managers' demand for audit quality. We posit that these two important contextual factors influence managers' audit quality preferences. We study these factors because they are ubiquitous to companies, and we focus on their influence on managers because managers continue to play a significant role in the auditor hiring process and we know relatively little about their auditor preferences. We predict that risk, the potential for loss, and their interaction will each decrease manager demand for high audit quality due to a desire to achieve greater reporting flexibility. Experimental results are consistent with our predictions; specifically, increased risk, the potential for loss, and to a lesser extent their interaction, significantly reduce managers' likelihood of hiring the best available auditor in the market. Path analysis indicates that this reduction in audit quality demand leads to increases in misreporting. Finally, we observe investors overpaying for assets to a greater extent when managers hire lower‐quality auditors. Our results show that the contextual factors of risk and the potential for loss, which are ubiquitous to companies, can reduce demand for audit quality, which can increase misreporting behavior and ultimately harm investors.

We examine the relation between low‐quality internal controls and audit fee premiums. Using a novel data set of audit hours and audit fees we find, consistent with the audit risk model, that auditors increase their effort (hours) owing to low internal control quality. We find that auditors also charge a significant fee premium to clients with internal control weaknesses. This premium is observed for severe internal control weaknesses and companies with low‐quality alternative governance mechanisms. The results are robust to multiple methods to address endogeneity, including company fixed effects, difference‐in‐differences design, and a propensity score‐matched sample. Taken as a whole, low internal control quality leads to fee premiums, which are a deadweight loss to client companies.

This study investigates why auditors leave public accounting and the consequences of auditor departures, both of which have been of great concern for audit firms and regulators worldwide. Using data from China and controlling for demographics, we find that audit partners and managers, auditors who generate more revenue, and auditors who provide higher audit quality have a lower likelihood of departure, while non–Big 4 auditors have a higher likelihood of departing public accounting. We also find that an audit firm is likely to lose clients when an auditor departs. Clients who stay with the same firm pay lower audit fees but with no drop‐off in audit quality after a signing auditor departs. In supplementary analyses, we also demonstrate that the determinants and consequences of auditor departures are different for auditors who leave the firm but not the profession (i.e., auditor turnover). Specifically, high audit quality is associated with a lower likelihood of auditor departure, but it increases auditor turnover. We also observe that audit quality is reduced for the clients who stay with an audit firm after highly skilled auditors depart for high‐visibility corporate positions. Our study provides insights that should be of interest to the audit profession, audit firms, and regulators.

Staggered boards (SBs) are one of the most potent common entrenchment devices, and their value effects are considerably debated. We study SBs' effects on firm value, managerial behavior, and investor composition using a quasi‐experimental setting: a 1990 law that imposed SBs on all Massachusetts‐incorporated firms. We find that relative to a matched control group of companies, for treated companies the law led to an increase in Tobin's Q, investment in capital expenditures and R&D, patents, and higher‐quality patented innovations, resulting in higher profitability. These effects are concentrated in innovating firms, especially those facing greater Wall Street scrutiny. An increase in institutional and dedicated investors also accompanied the imposition of SBs, facilitating a longer‐term orientation. The evidence suggests that SBs can benefit early‐life‐cycle firms facing high information asymmetries by allowing their managers to focus on long‐term investments and innovations.

Understanding the role of bank lending in corporate innovation is important to policymakers, practitioners, and academics. We provide new evidence on such a role by exploiting the implementation of SFAS 166/167, which removed the off‐balance sheet status of certain securitized assets of banks. The regulation affects bank lending and thus represents a credit supply shock to borrowing firms. We find that affected banks raise spreads and cut loan amounts after the regulation. Firms that borrow from affected banks reduce R&D investment and the number and quality of the patents they generate. The reduction is concentrated among firms whose banks experience more downward pressure on capital ratios and greater market discipline, and firms that are more dependent on external financing. Additional analyses reveal that information asymmetry between incumbent banks and outsiders with respect to borrowing firms prevent them from switching. The overall findings suggest that bank lending promotes borrowing firms' innovation activities. Policies that restrict bank lending are likely to hurt innovation of borrowers.

We examine the extent to which deferred vesting of stock and option grants (deferred pay) helps firms retain executives. To the extent an executive forfeits all deferred pay if they leave the firm, deferred vesting will increase the cost (to the executive) of an early exit. The impact of deferred pay on executive retention, a key ingredient for firms to create shareholder value is hence an important empirical issue. Using pay duration proposed in Gopalan et al. (2014) as a measure of the extent of deferred equity, we find that CEOs and non‐CEO executives with longer pay duration are less likely to leave the firm voluntarily. The talent retention role of deferred pay is mitigated by performance‐vesting provisions and signing bonuses offered by industry peers. Moreover, we also find that voluntary turnover is less sensitive to pay duration for executives who are perceived to be more talented and have more firm‐specific skills. Overall, our study highlights a strong link between compensation design and turnover of top executives. It suggests that firms take into account the need for retaining managerial talent in designing executive compensation.

High investor sentiment has been linked with opportunistic managerial behavior in the face of more optimistic investors and analysts. We extend this line of work by documenting that the likelihood of misstatements is higher when sentiment is high. Although this would suggest elevated audit risk, we posit that a contemporaneous reduction in auditors' litigation cost could drive down audit fees and going concern opinion (GCO) reporting conservatism in order to please clientele. Consistent with this notion, we document that auditors charge lower fees and report GCOs less conservatively when sentiment is high. However, this reduction in reporting conservatism is unwarranted; results reveal that auditors are less likely to issue GCOs to clients which subsequently file for bankruptcy during high sentiment periods. We conduct additional tests to examine whether auditors' litigation costs indeed vary with sentiment and document that auditors are less likely to be sued and the market reacts less negatively to misstatement announcements when sentiment is high. Collectively, our findings suggest that, although misstatement risk is increasing with sentiment, auditors' litigation risk actually declines.