This study implements topical analysis to identify the extent to which managers shift their responses from analysts’ questions in earnings conference calls. We refer to this behavior as managerial topic-shifting. Using a sample of conference calls from 2002 to 2017, we find that managers in firms with better performance, more powerful CEOs, and a weaker information environment shift more from the topics of analysts’ inquiries. Managers are also more likely to shift topics when analysts’ questions display a more positive tone or lower specificity. Moreover, we find that managerial topic-shifting provides incremental information to capital markets by facilitating the incorporation of earnings information into stock prices. Our study documents a previously unexplored dimension of managerial disclosure strategy.

We study the labour market effects of permanent 30%–64% reductions on unemployment insurance benefits available in seven states. Leveraging linked firm-establishment data, we find that establishments based on reform states experience employment increases that are 0.8%–1.3% larger than those of the same firm’s establishments in other states. Using a similar multi-state firm design, starting salaries are 1.2%–5.5% lower in reform states and posted salaries for the same job fall by 3.2%–3.5%. The negative co-movement of employment and wages after the reform suggests a labour supply shock and mitigates against confounding changes in labour demand driving the results. Our findings are consistent with workers lowering their reservation wages as outside options fall, and employers take advantage of this by offering lower wages and increasing employment.

Using Chicago police data, we train a machine learning model to predict the risk of being shot in the next 18 months. Out-of-sample accuracy is strikingly high. A central concern with using police data is “baking in” bias, or overestimating risk for groups likelier to interact with police conditional on behavior. Our predictions, however, accurately recover risk across demographic groups. Legal, ethical, and practical barriers should prevent using victimization predictions to target law enforcement. But using them to target social services could increase both the potential for interventions to reduce shootings and the available statistical power to detect those reductions.

We consider estimation and inference for a regression coefficient in panels with interactive fixed effects (i.e., with a factor structure). We demonstrate that existing estimators and confidence intervals (CIs) can be heavily biased and size-distorted when some of the factors are weak. We propose estimators with improved rates of convergence and bias-aware CIs that remain valid uniformly, regardless of factor strength. Our approach applies the theory of minimax linear estimation to form a debiased estimate, using a nuclear norm bound on the error of an initial estimate of the interactive fixed effects. Our resulting bias-aware CIs take into account the remaining bias caused by weak factors. Monte Carlo experiments show substantial improvements over conventional methods when factors are weak, with minimal costs to estimation accuracy when factors are strong.

We study social preferences in settings where someone who chooses on behalf of others knows how those individuals rank the available options but may lack cardinal information concerning those comparisons. Contrary to majoritarian principles, most people place more weight on preventing least preferred outcomes for others than on enabling most preferred outcomes. Ranks matter both intrinsically and because they provide a basis for inferring cardinal utility. Ordinal aggregation principles are stable across domains and countries with divergent political traditions. Designing attractive social choice mechanisms is challenging in practice partly because aggregation principles that make manipulation difficult yield outcomes people consider normatively unappealing.

Why is workplace sexual harassment chronically underreported? We hypothesize that employers coerce victims into silence through the threat of a retaliatory firing, and test this theory by estimating whether external shocks that reduce the value of a worker's outside options exacerbate underreporting. Under mild assumptions, a rise in the severity of formal complaints is indicative of increased underreporting. Combining this insight with an objective measure of the quality of charges filed with the Equal Employment Opportunity Commission (EEOC), we perform two analyses. First, we assess whether workers report sexual harassment more selectively during recessions, when outside labor market options are limited. We estimate the fraction of sexual harassment charges deemed to have merit by the EEOC increases by 0.5-0.7% for each one percentage point increase in a state-industry's monthly unemployment rate. The effect is amplified in industries employing a larger fraction of men and in establishments with a higher share of male managers. Second, we test whether less generous UI benefits create economic incentives for victims of workplace sexual harassment to remain silent. We find the selectivity of sexual harassment charges increases by more than 30% in response to a 50% cut to North Carolina's Unemployment Insurance (UI) program following the Great Recession.Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.