Information Asymmetry and Self-Selection Bias in Loan Announcement Studies

Abstract

We question the validity of the broad consensus in the literature that loans are unique relative to other financial contracts. Research in this area is event-study driven and implicitly assumes that relatively small samples of loan announcements adequately represent all bank loans. Our analysis begins with a sizeable loan universe and asks whether some credits are more likely to be announced than others. We find that loan announcements are relatively rare, and that factors such as information asymmetry and materiality affect borrowers’ decisions to announce loans. Using Billet, Flannery, and Garfinkel’s (1995) data, we show that their sample fails to represent the loan universe and that significant abnormal announcement returns are confined to the smallest firms. The abnormal return for our own sample, which better represents the loan population, is insignificantly different from zero. We conclude that if all loans were announced, the average abnormal return would likely be insignificant. Our findings, which suggest self-selection bias affects extant loan announcement research, do not support the views that loans generally are a special form of finance or that private and public debt differ in significant ways.

Keywords: loans, announcements, selection bias, event studies, information asymmetry

JEL Classification: G14