Dividends are clustered in increments of 5, such as 25, 50, and 75. Firms that gravitate towards these ‘prominent’ amounts have lower operating performance and lower annual five-factor alphas of 77 b.p. Consistent with agency frictions that lead to lazy decisions, clustering effects are stronger for entrenched firms, with more market power, and less shareholder activism. Dividend increases also cluster more than cuts, consistent with saliency bias. In a counterfactual exercise, we find no similar patterns in a sample of ADRs. Our results complement a number of recent studies showing the economic importance of simple decision heuristics.

We test whether bank loans change public bond yields. A 25% increase in bank debt raises bond yields by 8 bps, reflecting a trade-off between the benefits of bank cross-monitoring and higher bond risk. This effect is smaller for firms with no credit default swaps (CDSs) and with junk debt—scenarios where bank monitoring is most valuable. It is unlikely that firms with bank debt are riskier, because they are less likely to be downgraded and have lower loan spreads. We find similar results using a natural experiment around the 2014 oil shock. Our results highlight how bond yields depend on incentive conflicts among creditors.