We discuss the evolution of research on textual attributes in accounting and suggest potential directions for future research using Cazier and Pfeiffer (2017) to illustrate the challenges and opportunities in the literature. We categorize the literature based on whether a given textual attribute is assumed to be “discretionary” versus “nondiscretionary” and whether it is assumed to be “helpful” versus “harmful.” The literature to date has been largely descriptive, with limited focus on causality, and we argue that there are substantial opportunities for research that better isolate the discretionary components of textual disclosure, as well as identifying specific contexts in which a given attribute is more likely to be helpful than harmful.

JEL Classifications: D8; M4.

You do not currently have access to this content.