Research has begun to analyze the tone and narrative structure of earnings announcements after decades of focusing on market reactions to the earnings news itself. One conclusion from this literature is that language matters – the tone (i.e., the excess of optimistic over pessimistic language) and opacity of the earnings announcement text is associated with future firm performance and with the market reaction to the earnings press release. In other words, managers’ language choices convey information beyond that captured by earnings, and investors respond accordingly.
In a recent paper, we study another aspect of managers’ disclosure choice: the decision to emphasize or downplay information by placing it in certain spots within an earnings announcement. In particular, we ask two questions. First, does the placement of news within an earnings announcement reflect managerial efforts to inform or mislead investors? Second, do investors react appropriately to the news emphasized by managers?
Our study is motivated by the considerable discretion managers have in preparing earnings announcements, and the fact that managers may use this discretion for good or for ill. In the past, many managers emphasized performance measures that presented the firm in a favorable light. The Securities and Exchange Commission’s concern about managers abusing this discretion led to Regulation G, which prohibits the presentation of non-GAAP measures in a way that gives them more prominence than GAAP earnings. However, managers continue to have significant discretion in what they emphasize in the more general language in their earnings announcements. They can discuss positive news before negative news, discuss more important news before less important news, or discuss historical results before forward-looking results, etc. Our interest is in how the placement of positive information in earnings announcements influences investors, and whether it should.
We measure variation in information placement based on the extent to which positive information is concentrated in the earliest part of the earnings announcement, rather than being spread evenly throughout the document. Our definition is consistent with the SEC’s view that items are more prominent when they are presented earlier in a document. Specifically, we use textual analysis to calculate the net tone of earnings announcements on a sentence-by-sentence basis, and then compare the tone of the earliest sentences to the tone of the sentences in the entire document. We describe earnings announcements as emphasizing positive information when the tone (i.e., the excess of positive language over negative language) of the earliest portion is greater than the tone of the overall document (i.e., when positive information is concentrated in the earliest portion of the document). This approach allows us to test for the effect of disclosure information placement while holding the language of the document (the net tone) constant.
We begin by noting that managers, on average, emphasize positive news: the first part of an earnings press release is, on average, more positive than the overall document. We then investigate the determinants of information placement, with the goal of assessing whether managers place information within a document opportunistically or whether managers use relative placement to inform investors. Based on several proxies for opportunism, we find no evidence that managers emphasize good news for opportunistic purposes. In contrast, we find consistent evidence that relative information placement conveys useful information. Specifically, when managers report good news or when investors are too pessimistic about future earnings, positive information tends to be more concentrated in early portions of the document than when managers report bad news or when investors are too optimistic about future earnings. We interpret this as evidence that information placement is primarily driven by the underlying economic news being reported by the firm, rather than representing an opportunistic tendency to obfuscate bad news.
We next investigate whether investors act as if information placement reflects information content. We find that investors do indeed respond as if the placement of news conveys information – market reactions to earnings announcements are more positive when positive information is presented more prominently (i.e., earlier in the document). This association is consistent with investors interpreting the placement of information as an indication of the importance of that information. In other words, investors place a greater weight on language that occurs earlier in the document.
Finally, we investigate whether investors overreact or underreact to information placement. To do so, we examine post-earnings returns to assess whether they are systematically correlated with the degree to which positive information is emphasized. We find that post-earnings returns are positively correlated with emphasized tone. We interpret this result as evidence that investors do not give too much weight to information at the beginning of earnings announcements. Instead, the evidence is consistent with a more general underreaction to the information provided at the earnings announcement.
Taken together, our evidence suggests that the placement of information within earnings announcements conveys useful information to the market, and that investors respond to that information, though incompletely. Our study adds to the large body of literature related to the structure of financial disclosures and the placement of information within those disclosures. In particular, our results speak to long-standing regulatory concerns that managers will systematically mislead investors by giving greater prominence to favorable news, while hiding bad news at the end of a disclosure. While we do find that investors are influenced by information placement, our results do not support those concerns. The investor response to emphasized information seems appropriate, given how managers choose to emphasize information in our sample; if anything, investors underreact to the news implicitly conveyed by information placement.
This post comes to us from Professor Lin Cheng at the University of Arizona’s Eller College of Management and professors Darren Roulstone and Andrew Van Buskirk at Ohio State University’s Fisher College of Business. It is based on their recent article, “Are Investors Influenced by the Order of Information in Earnings Press Releases?” available here.