A Plain English Measure of Financial Reporting Readability

This study highlights the advantages of using multifaceted measures of readability based on the plain English attributes specifically recommended by the SEC (e.g., passive voice, hidden verbs). Over the past decade a burgeoning literature has emerged using textual analysis to examine the impact of financial reporting readability on the capital markets. Loughran and McDonald (2014) provide evidence that components of the most commonly used measure of readability, the Fog index, are poorly specified in financial applications and suggest an alternative proxy for reporting readability based on the file size of a 10-K document. We show that a significant portion of the variation in file size is driven by the non-text components of the 10-K (e.g., HTML, XML) that have little to do with the textual readability of the reports. As such, we create of measure of readability based on plain English attributes and then validate our measure using a quasi-exogenous mandatory shock provided by the 1998 Plain English Mandate requiring firms issuing prospectuses to communicate in plain English. We show that compared to the annual filings of a matched set of firms, the firms filing prospectuses had significantly greater improvements in plain English measures of readability. In contrast, we find no evidence of decreases in file size, which suggests that total file size does not capture the attributes of readability promoted by the SEC. In sum, we provide evidence consistent with file size having substantial measurement error as a proxy for readability and suggest that measures based on plain English attributes better capture the financial readability attributes highlighted by the SEC.