Malcolm
Wardlaw

Confounding events can cause false positives when testing the relationship between short-term returns around a quasi-experimental event and firm characteristics. We show that this risk is severe in practice: return-characteristic relationships are often statistically significant at the 1% level on over 30% of all trading days. Benchmarking a relationship against the distribution of the same relationship on pre-event days is effective at addressing the problem. We introduce a novel GLS variation of this approach that achieves large gains in statistical power relative to OLS and provide Stata and Python modules that implement both procedures.

We study depositor behavior and capital flight using data from the Freedman’s Savings Bank, a bank established after the U.S. Civil War to support formerly enslaved individuals. White depositors, who generally had better access to financial information, begin providing significant funding as interest terms are enhanced. They respond faster to postpanic stabilization efforts and negative bank-specific information, and are twice as likely to close accounts before failure, passing expected losses to Black depositors, who were the bank’s primary philanthropic target. Our results show how demand deposit funding of bank-like institutions can create systemic disadvantages for less experienced borrowers.

Using detailed depositor arrival data from the Freedman's Savings and Trust, we examine how the failure to prevent racial violence impacts bank participation. From 1866 to 1873, we find that events of racial violence perpetrated against Black Americans decrease new account openings at branches in the same region by 25% relative to other branches in the 30 days after the event.  Alternatively, events that increased political representation and protections increase the relative arrival of new depositors at the bank by around 50% for the affected branches. We also show that those who opened accounts in the wake of a violent event were less likely to close an account before the bank's eventual failure, suggesting that those who were resistant to the impact of political violence may have tragically been more exposed to other institutional failures.

This paper assesses different econometric approaches to working with count-based outcome variables and other outcomes with similar distributions, which are increasingly common in corporate finance applications. We demonstrate that the common practice of estimating linear regressions of the log of 1 plus the outcome produces estimates with no natural interpretation that can have the wrong sign in expectation. In contrast, a simple fixed-effects Poisson model produces consistent and reasonably efficient estimates under more general conditions than commonly assumed. We also show through replication of existing papers that economic conclusions can be highly sensitive to the regression model employed.

This paper presents evidence of a large, persistent decline in establishment-level workplace injury rates after private equity (PE) buyouts of publicly traded U.S. firms. We find that firms experience fewer OSHA safety violations after buyouts and that a larger decline in injury rates is associated with an increased probability of exit via IPO. Employment reductions after buyouts are concentrated in relatively low-injury-risk establishments. Overall, our results suggest that buyouts improve workplace safety and that PE acquirers benefit from this improvement. We explore possible causes of these changes through interviews with executives of companies acquired in buyouts and through cross-sectional analysis.

A large and rapidly growing literature examines the impact of misvaluation on firm policies by using mutual fund outflow-induced price pressure to isolate nonfundamental price variation. I demonstrate that the standard approach to computing outflow-induced price pressure produces a measure that is inadvertently a direct function of a stock's actual realized return during the outflow quarter, raising doubts about its orthogonality to fundamentals. After removing these direct measurements of return, outflows generate a fairly negligible quarterly decline in returns, with no subsequent reversal, and many established results in this literature no longer hold. I provide suggestions for future analysis.

We present evidence that financing frictions adversely impact investment in workplace safety, with implications for worker welfare and firm value. Using several identification strategies, we find that injury rates increase with leverage and negative cash flow shocks, and decrease with positive cash flow shocks. We show that firm value decreases substantially with injury rates. Our findings suggest that investment in worker safety is an economically important margin on which firms respond to financing constraints.

We estimate a dynamic model of the decision to close a troubled bank. Regulators trade off an aversion to closing banks against the risk that allowing a bank to continue will raise the eventual costs to the deposit insurance fund. Using a conditional choice probability approach, we estimate the costs associated with closing banks, both in direct costs to the insurance fund and in other costs perceived by regulators, either social or personal. We find that delayed closures were driven by a desire to defer costs, an aversion to closing the largest and smallest troubled banks, and political influence.

This dataset contains bank transaction and dividend records for the Freedman's Savings and Trust Bank, which operated between 1865 and 1874 in the United States and served primarily newly freed, formerly enslaved people. The branch locations ranged from New York to Shreveport and from St. Louis to Jacksonville. The data consists of two parts: a random sampling of about 500 passbooks, which contain transaction data, and a record of all dividend repayments to account holders following the dissolution of the bank. The latter dataset contains over 40,000 observations. The result is a robust picture of ending account balances and deposit and withdrawal activity for the Freedman's Savings and Trust Bank.

Syndicated loan covenants exhibit substantial heterogeneity even among observably similar borrowers. We ask whether this variation reflects economically meaningful differences in contracting outcomes shaped by bargaining power or is largely incidental once observable risk is accounted for. Using a new measure of covenant tightness, defined as a contract's position within a deal-specific distribution of feasible covenant packages generated by large language models and validated against independent programmatic methods, we document a sharp asymmetry across borrower credit quality. Among investment-grade borrowers, tighter covenants are associated with significantly higher loan spreads but do not predict subsequent violations. Among speculative-grade borrowers, tightness has no incremental pricing power but strongly predicts violations. This cross-outcome asymmetry is difficult to reconcile with a residual risk interpretation and instead reflects two distinct contracting regimes: negotiated outcomes driven by bargaining power for investment-grade borrowers, and lender risk management and proximity to binding constraints for speculative-grade borrowers. More broadly, we show how counterfactual contract distributions can be used to measure contract design in settings with substantial contractual heterogeneity.

We use natural language processing to measure the complexity and standardization of the events of default and covenants sections of 7,934 private loan agreements filed with the SEC from 1996 to 2017. Using word counts, a legal and financial dictionary, and LDA topic models, we construct measures of contractual detail and "distance to boilerplate." We document a complexity-standardization tradeoff: larger loans with longer maturities are simultaneously more detailed and more standardized, while riskier borrowers and those with complex financial structures receive more customized contracts. Pairwise distance analysis reveals that the form of the contract is determined primarily at the firm level, not the bank level, rejecting the hypothesis that default provisions are off-the-shelf templates. More standardized contracts are renegotiated more frequently, while more customized contracts are renegotiated less, suggesting that boilerplate language serves as a flexible reference point adapted through ex post amendment.

This paper introduces a novel method for estimating occupation-specific workplace injury rates using establishment-level administrative data. Contrary to the belief that the U.S. job mix has shifted towards safer jobs, our findings reveal that changes in occupational composition overall did not contribute to the 57% decline in the U.S. workplace injury rates since 1999. Instead, injury rates declined substantially across many occupations, with the largest declines in blue-collar and healthcare support occupations. While declining employment in production occupations and growing employment in high-skilled occupations lowered the aggregate injury rate, these effects are small and offset by other compositional changes.

We show that a platform-pays mechanism can address ratings inflation and ratings shopping with minimum regulatory oversight. While we focus on ratings industry, the mechanism also applies to a setting where firms seek unbiased reports from external auditors. The mechanism has two necessary and sufficient features: strategic delegation and outcome-contingent contracts. First, an issuer strategically delegates the task to acquire ratings from credit ratings agencies (CRAs) to a pass-through non-monitoring platform (the "trust"). The trust operates as a commitment mechanism, assuring investors that issuers did not shop for ratings. Second, the ex-ante determined publicly available fees schedule is partially outcome-contingent, allowing most of the fees to be paid upfront. We provide the contractual space of fees that support the unbiased ratings equilibrium. A contract calibrated to data shows that the cost of implementation of such a mechanism is small. Results encourage policymakers to undertake reforms.

This study examines the joint choice of bond features that corporations make in defining their bonds. Using data on new corporate bond issues from 1990 through 2012, we find that corporate bonds are packages of different provisions and restrictions reflecting complementarities between bond features that change with issuer characteristics. Broadly, we find evidence that asymmetries of information between issuer and investors, agency conflicts between stockholders and bondholders, credit risk, the nature of the firm’s growth prospects, and bond market conditions all play a role in the use or non-use of different corporate bond features.

We examine the response of investment to peers' stock prices. While the response to average peer-Q is typically positive, the response to prices of peer firms that are more threatening and those of industry leaders is reliably negative. The responses are more strongly negative when the prices contain more firm-specific information. We also show that different measures of peer price informativeness can capture either positive or negative investment signals. Thus, the response to peer prices varies with industry competition and whether the prices reflect firm-specific or industry information, clouding the traditional interpretation of variation in the responses to peers prices.