Detecting Earnings Information in Repurchase Announcements Using the Luck of the (Double) Irish

Abstract

Previous studies disagree about whether open market repurchase announcements contain positive, negative, or no information about earnings expectations. In this comprehensive study, I use a novel measure of analyst earnings forecast revisions that corrects for undetected biases found in prior studies. While I find no evidence of average changes in earnings expectations across a 28-year sample of repurchase announcements, I find positive post-announcement forecast revisions in a subsample of firms likely to be using “Double Irish” offshore subsidiary structures to delay or avoid U.S. taxes on firm profits. Tests of this subsample utilize a quasi-natural experiment related to a 2007 IRS repurchase regulation which imposed additional repurchase-related costs on these firms. The results are robust to a variety of test specifications and proxies for earnings. This link between repurchase costs and positive earnings revisions is broadly consistent with signaling models of payouts and suggests that managers may announce especially costly repurchases when earnings expectations are underestimated by outsiders.