Why do firms disclose analyst following on their corporate websites?

Abstract

About two-thirds of S&P500 firms disclose their analyst following on their corporate websites. Half of these firms disclose their analyst following in an unbiased fashion, while the rest manage this disclosure by selectively omitting analysts with pessimistic views (selective disclosers). Consistent with facing stronger incentives to manage expectations, coupled with weaker information environments and monitoring by external stakeholders, selective disclosers exhibit lower profitability, higher growth opportunities, stronger financing needs and a greater percentage of retail investor ownership. Moreover, selective disclosers experience lower future returns and exhibit higher income smoothing and lower accrual quality. Taken together, our evidence is consistent with some firms attempting to serve their strategic needs by selectively excluding from their websites analysts with less favourable views of their firm.

Keywords:

• voluntary disclosure
• analyst following
• internet financial reporting

JEL Classifications:

• G30
• G32
• G38
• M40

Acknowledgements

We thank Pawel Bilinski, Ester Einhorn, Marios Panayides, Grace Pownall, Aneesh Raghunandan, seminar participants at the 2019 EAA Annual Congress in Paphos and the 2nd Cass CeFARR conference and seminar participants at the University of Cyprus for their constructive comments and suggestions. We would also like to thank Ioanna Philippou and Alexandros Antonopoulos for their excellent research assistance.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 This expectation is based on the assumption that managers know both who the analysts that follow their firm are and each analyst’s expectation about the company. This assumption is in line with general corporate practices and supported by empirical findings which suggest that managers discriminate among analysts attending earnings conference calls by only allowing more favourable analysts to ask questions during the call (Mayew 2008).

2 This is clearly stated in such website disclosures. Most firms explain the provision of this information by stating that they are providing this listing as a service to their shareholders.

3 Although the disclosure of analyst following can be useful to both retail and institutional investors, it is less clear if institutional investors will similarly be misled by biased disclosure. To the extent that sophisticated investors refer to the provided list as a resource that will help them identify which analyst reports to seek, it is possible that they will also be misled by the selective disclosure of analysts covering the firm, although this is not an essential nor a necessary premise of the paper.

4 There is a large literature that documents the significant benefits to firms stemming from increased disclosures. For example, Diamond and Verrecchia (1991) and Lambert et al. (2007) show that increased disclosure reduces the cost of capital, while the enhanced information environment following the adoption of IFRS is associated with increased stock liquidity (Daske et al. 2008) and reduced cost of equity capital (Li 2010).

5 In a similar vein, Doyle et al. (2003) find that the exclusion of expenses from pro-forma earnings predictably leads to lower future cash flows. Cohen et al. (2020) predict and find that firms that hide bad news during conference calls exhibit future negative abnormal returns as the bad news is eventually released.

6 The uniqueness of the institutional environment of regulated firms makes them incomparable to the rest of our sample. For almost all our variables, we find a significant difference between the means and medians of the regulated sample and our firm sample.

7 Relative to earnings forecasts, stock recommendations do not change as often. Related literature shows that the average time period between two recommendations is more than one year. To ensure that we include all analysts that follow a given firm, we assume that an analyst follows the firm if that analyst has issued a recommendation within a two-year period prior to July 2014. Results are not changed significantly if we only retain recommendations issued within a one-year period (see footnote 16). 

8 Recommendations span a scale of 1 to 5, where 1 represents a Strong Buy and 5 a Strong Sell. Therefore, more favourable recommendations translate to lower numerical values.

9 We exclude initiations for this analysis as analysts face different incentives when issuing their first report (Irvine 2003, O'Brien et al. 2005)

10 For expositional simplicity, our reference to disclosure date reflects the time the data was collected (July 2014) and not when firms initiated making such disclosures since the latter is unknown.

11 Related literature documents that institutional ownership is an important firm-monitoring mechanism that can be attributed to an institutions’ more sophisticated investors (Walther 1997, Jiambalvo et al. 2002) and their use of securities class actions to effectively enforce their monitoring role (Cheng et al. 2010).

12 This relationship is not consistently positive as asset pricing theory would predict. Dichev (1998) for example finds that high bankruptcy risk is associated with lower than average returns. Vassalou and Xing (2004) find that default risk is priced but that high default risk earns higher future returns only in small firms with high book-to-market ratios. Penman et al. (2007) decompose the book-to-market ratio, and after controlling for systematic risk, document a negative relation between leverage and future returns. Based on this finding, Caskey et al. (2012) hypothesize and find that leverage can explain future returns only if measured in relation to the firm’s optimal capital structure. Based on these studies it is evident that leverage can be an important factor that explains future returns, but the direction of such association is less evident.

13 Only the IBES recommendation file includes the identification information of both the investment bank and the analyst. Due to regulatory compliance concerns, effective October 2018, IBES stopped the practice of disclosing the names of contributors and analysts in their detailed files for a number of their key contributors.

14 18 firms both added and dropped analysts from their websites and are thus included in both columns.

15 Results do not change significantly if we classify firms into selective and non-selective disclosers based on recommendations issued within one year (rather than two years) preceding July 2014. Specifically, we continue to find that the mean recommendation of analysts included on (excluded from) websites of selective disclosers is lower (higher), i.e., more (less) favourable, than the corresponding value for non-selective disclosers. Logistic regression results also provide consistent evidence that firms with greater financing needs, higher growth opportunities, and weaker debt-holder monitoring are more likely to selectively disclose analyst following, although firm profitability and retail ownership are no longer significant. Retail ownership once again is significantly higher for selective disclosers when we exclude the 49 firms that got reclassified when we switched to only including recommendations issued within one year of July 2014.

16 The fact that the coefficient on Selective Disclosers is negative and significant in year 2 (column 2 of Table 4) suggests that selective disclosers continue to exhibit a drop in their value in the second year after the disclosure and that the two-year overall return is not driven by a reduction in value from the first year only.

17 Proxying for firm size using the log of the market value of equity instead of total assets does not qualitatively change results.

18 The issue of endogeneity is mitigated since these are additional measures of opportunistic behaviour measured contemporaneously with the strategic disclosure of analyst following.

19 In the first model, we lose one observation from the non-discloser sample (118) for a total of 245 observations. In the second model, we lose an additional 24 observations from requiring four years of data to compute the standard deviation of both net income and operating cash flows.

20 We are grateful to an anonymous reviewer for pointing this out.

21 A disadvantage of using median values is that the discrete nature of analyst recommendations renders the median recommendation less informative and hence, a cruder consensus measure.