Effect of regulation FD on disclosures of information by firms

Abstract

Critics of Regulation Fair Disclosure (FD) have argued that its enactment would result in not only a decrease in asymmetric information but a decrease in total amount of information disclosed by firms. We investigate this conjecture and find (1) no change in market risk premium, (2) an increase in risk premiums for size and (3) an increase in the distress risk premium in the post-FD period. These findings lead us to conclude that in the post-FD period there is a significant decrease in the dissemination of (1) overall information by small firms and (2) unfavourable information by firms in general.

Keywords:

• regulation FD
• disclosure of information
• Fama–French three-factor model
• structural changes

JEL Classification:

• G12
• G30

Acknowledgements

The data on the monthly returns of the 25 FF portfolios and the explanatory variables R_m , SMB, and HML data are taken from the website of Dr. Kenneth French, http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html Final Rule: Selective Disclosure and Insider Trading, Securities and Exchange Commission, 17 CFR Parts 240, 243 and 249, Release Nos. 33-7881, 34-43154, IC-24599, File No. S7-31-99, RIN 3235-AH82, http://www.sec.gov/rules/final/33-7881.htm

Notes

¹ For example, see, Brown et al. (2002), Sunder (2002), Heflin et al. (2003), Bushee et al. (2004), Chiyachantana et al. (2004), Eleswarapu et al. (2004), Gintschel and Markov (2004), Findlay and Mathew (2006), Francis et al. (2006), Cornett et al. (2007), Bagnoli et al. (2008), Choi et al. (2009), DeJong and Apilado (2009) and Jong (2011).

² Because of regulation FD, one of the following four situation may arise: (1) the informational asymmetry between investors increases even though firms increase overall disclosure of information, (2) the informational asymmetry between investors increases and firms have decreased disclosure of information, (3) the informational asymmetry between investors decreases because the firms have increased overall disclosure of information and (4) the informational asymmetry between investors decreases even though there is decrease in the overall disclosure of information by firms. In congruence with the findings of the current literature we assume that there is a decrease in informational asymmetry due to regulation FD and then examine regulation FD's effect on the overall disclosure of information by firm.

³ Researchers have found out that SMB and HML factors may also act as a proxy for some other risk factors besides representing the size and distress condition of the firm. Lakonishok et al. (1994) argue that book to-market measure proxies for investor bias in earnings-growth extrapolation. Liew and Vassalou (2000) show that SMB and HML predict future economic growth in the presence of traditional business cycle variables. Lettau and Ludvigson (2001) document the sensitivity of HML to bad news at times of economic distress. Ferguson and Shockley (2003) and Rolph (2003) argue that the FF factors proxy for leverage effects. They suggest that the FF factors are mimicking portfolios for risk factors associated with time variation in risk premia. No matter what the interpretation, SMB and HML act as proxies for the riskiness of the firm. In this article, following Fama and French (1996) we assume that SMB and HML represent, respectively, size and distress risk premiums.

⁴ The Snedecor's F-statistic stipulates that the numerator sample variance should be higher than the denominator sample variance.

⁵ We are not testing for the normality of the population because the sample is large enough to be asymptotically normal.

⁶ Using the pre- and post-FD variances the F-statistic is computed by dividing the higher variance with the lower variance.

⁷ To compare the pre- and post-FD periods we are using the time when the FD law was proposed (December 1999) not the time when the law was implemented (October 2000). We believe that the market received the information on FD law in December and started adjusting then (efficient market hypothesis). We have done similar tests for the pre and post October 2000 and get similar results leading to same findings.

⁸ We used the Toyoda's test (1974) rather than the more common Chow test (1960). The Chow test assumes that the homoscadesticity of error term in the two nonoverlapping regimes whereas Toyoda's test accounts for the heteroscadisticity as well. Note, however, that if the two nonoverlapping regimes are homoscadestic then Toyoda's test reduces to Chow test.