افشای منصفانه مقررات SEC، اطلاعات و هزینه سرمایه

Regulation Fair Disclosure (“Reg FD”), adopted by the U.S. Securities and Exchange Commission in October 2000 was intended to stop the practice of “selective disclosure”, in which companies give material information only to a few analysts and institutional investors prior to disclosing it publicly. Our analysis shows that the adoption of Reg FD caused a significant shift in analyst attention, resulting in a welfare loss for small firms, which now face a higher cost of capital. The loss of the “selective disclosure” channel for information flows could not be compensated for via other information transmission channels. This effect was more pronounced for firms communicating complex information and, consistent with the investor recognition hypothesis, for those losing analyst coverage. Moreover, we find no significant relationship of the different responses with litigation risks and agency costs. Our cross-sectional results suggest that Reg FD had unintended consequences and that “information” in financial markets may be more complicated than current finance theory admits.

We empirically investigate the effects of the adoption of Regulation Fair Disclosure (“Reg FD”) by the U.S. Securities and Exchange Commission (“SEC”). Reg FD, which took effect on October 23, 2000, was intended to stop the practice of “selective disclosure,” in which companies give material information only to certain selected analysts and institutional investors prior to disclosing it publicly. To provide an evaluation of Reg FD it is important to understand both how information is transmitted from firms to capital markets and how the allocation of information-producing resources affects securities prices. We find that the adoption of Reg FD caused a significant reallocation of information-producing resources, resulting in a welfare loss for small firms, which now face a higher cost of capital. The loss of the “selective disclosure” channel for information flows could not be compensated for via other information transmission channels. This unintended consequence of Reg FD shows that “information” in financial markets is more complicated than current theory admits and has some important public policy implications. There were three reasons given by the SEC for the adoption of Reg FD. First, it was argued that selective disclosure leads to a loss of investor confidence. If small investors fear that insiders will regularly profit at their expense, they will not be nearly as willing to invest. A second rationale concerned the link between corporate governance and the incentives to engage in selective disclosure. This problem was the use of information by management to essentially bribe analysts, perhaps in exchange for a quid pro quo. Finally, the SEC stated the view that selective disclosure is not required for market efficiency because of technological change. The basic idea seems to be that companies can now use websites and/or “webcast” conference calls, making the channel of information flow from firm management to analysts less important. Is Reg FD a desirable public policy? We focus on assessing the effects of Reg FD on information production and transmission in capital markets; we ask whether the same information is transmitted to capital markets since the passage of Reg FD, just via a different channel now (albeit in a possibly “fairer” way), or whether, for some reason, there is less information transmitted to capital markets. We view this issue as of independent interest. Little is known about whether the sources of information or the costs of producing the information have any impact on firm values and security prices. Ultimately, we care about the effects on security prices and the resulting allocation of resources. We use the natural experiment of the adoption of Reg FD to investigate these issues. Information can be transmitted from firms to markets via four channels: (1) firms, in addition to mandatory disclosures, can disclose information to the public voluntarily (e.g., earnings pre-announcements); (2) firms can selectively disclose information, e.g., phone calls, or one-on-one meetings; (3) “sell-side” analysts can produce research which is released to the public, e.g., analysts reports; (4) private information can be produced by outsiders, “informed traders,” who then trade on the basis of their information. Reg FD sought to eliminate the second channel of information flow, under the implicit assumption that the same information would still flow into markets but by the other channels, particularly channel (1). But, Reg FD also affects channel (3) because the biggest impact of Reg FD is on analysts. Some have predicted that Reg FD will either lead to a diminished role for analysts since the information will now be available for everyone or a large-scale reduction in analysts' jobs since many simply cannot perform an adequate analysis without the aid of selective disclosure (Coffee, 2000). If analysts' previous role is subsumed by the other channels of information flow, then market efficiency is not changed. This is the logic of Reg FD that is our focus. Our main empirical strategy is to explore various cross-sectional differences among firms pre- and post-Reg FD. Our main findings are that there was a reallocation of information-producing resources and that this reallocation had asset-pricing effects. We document that small firms on average lost 17 percent of their analyst following, while big firms gained 7 percent, on average. Moreover, while the odds of the big firms voluntarily disclosing information (through pre-announcements) after Reg FD are twice as large as in the period before Reg FD, small firms did not increase their pre-announcements significantly. We find consistent effects of this reallocation on earnings forecast errors and market responses to earnings announcements: small firms experience higher forecast errors (Agrawal et al., 2006 find similar results) and volatility at earnings announcements, consistent with a higher information gap; no significant increases occur for big firms. These results suggest that big firms were able to replace the loss of channel (2) with channels (1) and (3), but that small firms were not able to do so. We demonstrate that this reallocation resulted in a higher cost of capital for small firms (and no significant change for big firms). We further investigate other characteristics of firms that may explain the cross-sectional impact of Reg FD. We explore Merton's (1987) investor recognition hypothesis saying that firms which are not covered by analysts or receive little publicity have higher costs of capital. We look at the complexity of information disclosed by firms, differences in litigation risks among firms, and we examine the link between agency problems in firms and incentives to make voluntary disclosures. Consistent with the investor recognition hypothesis, we find that the stocks of small firms that completely lost analyst coverage after Reg FD experienced significant increases in the cost of capital, while small stocks with no previous analyst coverage – which presumably did not have any analysts benefiting from selective disclosure pre-FD – experienced no significant change in the cost of capital. Moreover, we find that more complex firms (using intangible assets as a proxy for complexity) are more adversely affected by Reg FD than less complex firms. The post-Reg FD period coincides with the onset of a recession in early 2001 and the bursting of the technology bubble. We conduct various robustness tests to show that our results are driven by Reg FD and not by other contemporaneous effects. In particular, we conduct our analysis excluding high tech firms, which were the ones most affected by the bursting of the bubble, and include a recession dummy interacted with the post-FD period. The cross-sectional results and the robustness tests, while not conclusive proof that the effects are exclusively due to Reg FD, provide strong evidence of Reg FD being an important driving force behind the results. Several papers examine the impact of Reg FD in the information environment. There is a consensus in the literature that the quantity of voluntary public disclosures increased after Reg FD as reported in, for example, Heflin et al., 2003 and Bailey et al., 2003, and Straser (2002). The evidence is mixed on other relevant aspects such as the dispersion and accuracy of analyst forecasts, volatility around earnings announcement, and the degree of information asymmetry and informed trading. For example, with respect to the accuracy or dispersion of analyst forecasts, Shane et al. (2001) and Heflin et al. (2003) find no significant change in either, and Bailey et al. (2003) find no increase in the accuracy but significant increases in dispersion. Agrawal et al. (2006) and Mohanram and Sunder (2006) find increases in both, and Irani and Karamanou (2003) find increases in dispersion. On measures of the information gap, like return volatility around earnings announcement, Heflin et al., 2003 and Shane et al., 2001, and Eleswarapu et al. (2004) find decreases, but Bailey et al. (2003) and Ahmed and Schneible (2007) find no increase after controlling for important factors. Similarly, with respect to the level of information asymmetry reflected in trading costs, Eleswarapu et al. (2004) find that it decreases, consistent with SEC goals to level the playing field, but Straser (2002) finds opposite results.1 Our results contribute to the extensive literature in essentially two dimensions. First, we identify several firm characteristics that contribute to cross-sectional variations in the information environment pre- and post-Reg FD. In particular, we explore the effect of cross-sectional variations on firm size, complexity – as measured by the amount of intangible assets – exposure to litigation risk, and proxies for the intensity of use of the selective disclosure channel pre-FD. By breaking down the sample into subgroups with different exposures to the closure of the selective disclosure channel, we are able to better pinpoint Reg FD as the driving force behind the results. Our second main contribution is to examine the asset pricing implications of the regulatory changes. Our findings indicate that firms more exposed to the selective disclosure channel experienced a higher cost of capital after the passage of the regulation. The paper proceeds as follows. In Section 2 we investigate the reallocation of information-producing resources following the enactment of Reg FD. In Section 3 we look at the effects of this reallocation, in terms of the effects on earnings forecast errors, market responses to earnings announcements, and the cost of capital. In Section 4 we investigate whether characteristics aside from size help explain the results. Some robustness results are discussed in Section 5. The implications for inferences about information and public policy are in Section 6.

Overall, our results suggest that Reg FD had unintended consequences and that “information” in financial markets may be more complicated than current finance theory admits. We found that small firms were adversely affected by Reg FD; their cost of capital rose. Some small firms just completely stopped being followed by analysts, and consistent with the investor recognition hypothesis, the cost of capital increased for those firms. Moreover, public communication does not seem to be a good substitute for private one-on-one communication, where there is a free give and take, especially for firm communicating complex information. This finding is reminiscent of the “small-firm effect” which was first documented by Banz (1981) and Reinganum (1981). Our finding of the return of a small firm effect suggests a reason for this finding, namely, there was a lack of information, perhaps a certain kind of information about small firms. Schwert (2003) suggests that this information was subsequently forthcoming because investors responded to the academic studies. But that was prior to Reg FD. Are our results driven by Reg FD? One final piece of evidence concerns recent developments. The decline in analyst research documented in here has not gone unnoticed by public firms losing coverage or by the marketplace generally. Notably, there has been a significant increase in paid-for-research, in which companies pay research providers for coverage.20 Interestingly, two new firms, the National Exchange Network (NRE) and the Independent Research Network (IRN), have recently been created to act as intermediaries between research providers and public companies. A press release launching the IRN – a joint venture between NASDAQ and Reuters formed June 7, 2005 – provides some further confirmation for the main results in our paper: “Regulation FD (Reg FD) and the Global Analyst Settlement have brought sweeping changes to the research industry. As major broker dealers have focused resources on large capitalization stocks, a growing number of small- or mid-cap companies and ADRs have lost research sponsorship. 38% of all NASDAQ-listed companies and approximately 17% of NYSE companies have no analyst coverage. Over 50% of all publicly listed U.S. companies have two or fewer analysts covering them.”