Print
E-MailAndrew Sutherland is Ford International Career Development Professor of Accounting, John Core is Nanyang Technological University Professor of Accounting, and Inna Abramova is a PhD student at at MIT Sloan School of Management. This post is based on their recent paper.
Much research on voluntary disclosure focuses on decisions stemming from persistent factors. For example, the economic forces that give rise to a firm’s level of proprietary costs are persistent, and how the voluntary disclosure decision is affected by proprietary costs is persistent. Institutional investor ownership (IO) is another relatively stable determinant of disclosure. Prior research indicates that increases in disclosure associated with increases in IO can decrease information asymmetry and improve liquidity. In this paper, we hold IO constant, and examine how short-term changes in IO attention affect the firm’s short-term disclosure choices, and the resulting information asymmetry and liquidity consequences.
Our paper follows recent research on how disclosure and management behavior responds to changes in IO. Boone and White (2015) and Bird and Karolyi (2016) find that large increases in IO caused by index reconstitutions lead to more voluntary disclosure, which increases analyst following, reduces information asymmetry, and improves liquidity. Kempf et al. (2017) show that IO distractions lead to less participation in conference calls, fewer shareholder proposals, and more opportunistic behavior by management. These papers interpret their results as increases in IO resulting in improved monitoring.
However, recent shifts in the composition of IOs have raised questions about the effectiveness of IO monitoring. In 2016, passive funds attracted $563 billion of new investments while active funds experienced nearly $326 billion of withdrawals, accelerating a trend that has doubled the share of passive investment since 1998 (Lauricella and Lamy 2017). Passive investing has grown in popularity in part because it allows investors to gain exposure to a broader set of firms at a lower cost than actively managed investments. These same features impose constraints on the extent of passive IO monitoring activities. Passive IOs have to oversee more firms at a lower cost than their actively managed counterparts, and are less likely to specialize by industry. And yet, monitoring is perhaps more important to passive than active investors, because passive investors have limited ability to sell shares in underperforming firms (Romano 1993). When voting on “easy to monitor” matters, such as poison pills and board independence, passive IOs appear to rely on rules of thumb advocated by proxy advisors (Appel et al. 2016; Malenko and Shen 2016). In overseeing complex decisions requiring more situation-specific due diligence, such as M&A transactions and individual director appointments, passive investor monitoring is often ineffective (Schmidt and Fahlenbrach 2017).
Following this work, our paper investigates whether IOs appear to express a preference for more disclosure under the paradigm that “more disclosure is better”, without understanding which firms or circumstances are not suited for the paradigm. Firms incur costs when disclosing and wish to limit these costs. These costs include information processing costs of collecting information, opportunity costs of adding disclosures within space limitations, and potential reputation and litigation costs arising from incorrect predictions of uncertain future outcomes (Graham et al. 2005; Chapman and Green, 2017). Moreover, additional disclosures can increase proprietary, agency, and reputation costs. Thus, one hypothesis is that if IOs pester firms for costly but unnecessary disclosures, then firms will respond to IO attention by providing disclosure that is likely least informative. If this is the case, these marginal disclosure increases will not affect information quality or liquidity.
To investigate whether fleeting investor attention helps explain transient disclosure behavior, we employ the Kempf et al. (2017) IO distraction measure. The intuition behind this measure is that a firm’s IOs have other investments as well, and when major economic events affect those other investments, the IO will pay less attention to the firm. In other words, the strength of institutional attention to a given firm will vary depending on the returns on other positions in the investors’ portfolios. This approach is akin to recent work exploiting shocks in one part of a bank’s portfolio to study how monitoring changes for the bank’s other clients (e.g., Gopalan et al. 2011, Murfin 2012).
We find that disclosure has a negative relation with IO distraction. The number of forecasts, 8-K filings, and total disclosures all significantly increase with attention. Because variation in distraction comes from developments in other industries, and we control for industry-quarter and firm-calendar quarter effects, our results cannot reflect industry-specific shocks (e.g., economic conditions or an M&A wave), or firm-specific disclosure habits (always forecasting in the first quarter). The effects of IO attention on disclosure are economically significant. Moving from the 25th to 75th percentile of distraction within a given firm, our estimates indicate that the number of forecasts declines by 7.8%.
Next, we study the ways in which disclosure is sensitive to IO attention. First, we find disclosures per day are nearly twice as sensitive to IO attention than are days of disclosure. Thus, firms are much more likely to simply add or take away disclosures, conditional on disclosing on a given day, than they are to change their decision about whether to disclose on a given day. Second, we show that the most informative types of disclosures (e.g., earnings forecasts) do not change with IO attention. Last, we decompose our attention measure into portions related to passive and non-passive investors. We show that the variation in disclosure we find is primarily driven by passive investors, the group that the literature argues has the least monitoring resources. Together, these findings are consistent with firms slightly modifying their disclosure to placate the least informed segment of IOs.
Next, we study the consequences of attention-driven changes in disclosure, beginning with information quality. We examine whether distraction is related to the horizon or precision of forecasts, and whether the market reacts any differently to disclosures in periods of distraction. Although our earlier tests find that the quantity of disclosure is quite sensitive to IO attention, we find no response in forecast properties, or abnormal volume or abnormal volatility despite there being less disclosure. Finally, we examine how IO attention affects liquidity. We find no relation between attention and the bid-ask spread or price impact of trade on disclosure days.
Our paper contributes to research concerned with understanding firm disclosure choices (Beyer et al., 2010; Core 2001; and Healy and Palepu 2001). A common theme in this literature is that firms develop a sustained disclosure policy that incorporates their competitive environment, proprietary and agency costs, and investor base. Identifying whether IO attention could alter the disclosure policy has been challenging, because attention-grabbing events for the firm (e.g., capital raising, losses) can cause disclosure changes for their own reasons. Our analysis builds on the insight that IOs have limited attention, and that attention-grabbing events occurring in other industries can exogenously shift oversight away from the firm. Although we find that firms rarely adjust their decision to forecast, we show that the number of forecasts provided regularly changes, and this variation is predictably related to IO attention. One particularly novel aspect of this finding is that firms can reduce disclosure without worsening information quality or liquidity.
Our paper is also relevant to the literature studying IO monitoring. The shift of IOs from active to passive has generated interest in how IOs influence firm behavior, including disclosure. Recent research finds significant changes in disclosure and governance following index reconstitutions that increase the level of IO in the firm. Debate has followed about whether these ownership increases improve or worsen monitoring of the firm. One line of work finds that IOs cause more voluntary disclosure and higher disclosure quality, which leads to reductions in information asymmetry and improvements in liquidity. Others caution that passive investors often employ uniform rules-based monitoring techniques that are not effective for more complicated situations, and in some cases impose unnecessary costs on management. For example, Schmidt and Fahlenbrach (2017) find that firms experiencing an increase in passive ownership make poorer M&A decisions and appoint independent directors that are negatively received by investors. Consistent with disclosure being costly and passive owners pestering management for disclosures, we find that firms reduce disclosure when less attention is paid to them, and that such adjustments have no effect on information quality.
The complete paper is available for download here.
