The reduction in CoC post FD is mostly due to medium and large firms.

Regulation Fair Disclosure
and the Cost of Equity Capital
Zhihong Chen
City University of Hong Kong
Dan S. Dhaliwal
University of Arizona
Hong Xie
University of Illinois at Urbana-Champaign
1
Research Question
 Has Reg FD reduced the cost of capital as the SEC
believed?
2
Motivation
 Reg FD prohibits selective disclosure of material
information to a subset of market participants.
 The SEC is concerned that with selective disclosure
individual investors will lose confidence in the integrity of
capital markets
 By curtailing selective disclosure, the SEC believed that
Reg FD would

“encourage continued widespread investor participation in our
markets, enhancing market efficiency and liquidity, and more
effective capital raising” (SEC 2000)
 i.e., SEC believed that Reg FD could ultimately reduce
CoC by leveling the information playing filed.
3
Motivation – continued
 Recent theory papers (Easley and O’Hara 2004; Hughes et
al. 2006; Lambert et al. 2006) suggest that CoC


(1) increases in information asymmetry
(2) decreases in quality and quantity of information
 If Reg FD curtails selective disclosure and reduces
information asymmetry, theory predicts a decrease in CoC
post Reg FD, ceteris paribus.
 If Reg FD produces a “chilling effect” – deterioration in
quality and quantity of firms’ communication, theory
predicts an increase in CoC post Reg FD, ceteris paribus.
4
Motivation – continued
 Prior studies investigating the effect of Reg FD on firms’
information environment produce mixed implications for
the effect of Reg FD on CoC.
 Studies imply a decrease in CoC post Reg FD
 Eleswarapu et al. 2004; Sunder 2002; Gintschel and
Markov 2004; Heflin et al. 2003.
 Studies imply an increase in CoC post Reg FD
 Bailey et al. 2003; Irani and Karamanou 2003; Gomes
et al. 2006.
 SIA 2001; AIMR 2001.
 The effect of Reg FD on CoC is ultimately an empirical
question.
5
Hypothesis Development
 To the extent that Reg FD is effective in curtailing selective disclosure
and reducing information asymmetry without a significant adverse
effect on overall disclosure quality and quantity, we expect CoC to
decrease post Reg FD.
 Formally,

H1: CoC is lower in the post-FD period than pre-FD period on
average for the full sample, ceteris paribus.
 Gomes et al. (2006) find


Small firms lose analyst following after Reg FD and are unable to
replace lost information production from selective disclosure by
increased information production in public channels.
Medium and large firms gain analyst following after Reg FD and are
able to compensate lost information production from selective
disclosures by increased information production in public channels.
 Formally,

H2: The reduction in CoC post FD is mostly due to medium and large
firms.
6
Hypothesis Development - continued
 To the extent that Reg FD curtails selective disclosure, reduction in
CoC post Reg FD should be greater for firms with more severe
selective disclosure before Reg FD than firms with less severe
selective disclosure.
 Based on prior literature, we identify following firm characteristics
indicative of selective disclosure:



Lower book-to-market ratios (Smith and Watts 1992; Matsumoto 1999;
Hutton 2005)
With R&D activities (Gintschel and Markov 2004)
High levels of institutional, especially transient institutional, ownership
(Ke and Petroni 2004; Ke et al. 2005)
 Formally,

H3: The reduction in CoC post FD is greater for firms with low bookto-market ratios, positive R&D expenses or high levels of institutional,
especially transient institutional, ownership, relative to their
respective counterparts, ceteris paribus.
7
Research Design
 Estimates of the implied Cost of Capital.



Gebhardt, Lee and Swaminathan (2001)
Claus and Thomas (2001)
Gode and Mohanram (2003)
 Advantage of using implied cost of capital
 Theoretically correct (Dhaliwah et al. (2005))
 Attempt to separate cash flows effect and discount rate effect
from stock prices (Hail and Leuz (2006))
 We use the average of three measures as our proxy for CoC
to mitigate measurement errors (Dhaliwah et al. (2005)).
8
Research Design – continued
 Regression Specification
R AVGit     1TranFDit   2 PostFDit   1  MKTit   2  SMBit   3  HMLit   4 LogMVit
  5 LogBM it   6 MMTit   7 Lev it   8 Ferrit   9 Growthit   10 LogDisp it
  11 Less3it   12 IndCoCit 1   it
 Sample period from June 1998 to March 2002
 TranFD = 1 if an observation is from 12/1999 to 9/2000 and zero




otherwise.
PostFD = 1 if an observation is from 12/2000 to 3/2002 and zero
otherwise.
H1: β2 < 0 for the full sample
H2: β2 < 0 mostly due to medium and large firms
H3: β2 more negative for firms with low BTM ratio, positive R&D
expenses or high levels of institutional, especially transient institutional,
ownership
9
Results –
Descriptive Statistics
Figure 1 Median Cost of Capital by Quarter during June 1998 to March 2002
Median Implied Cost of Capital
13.00
Median Implied Cost of Capital
12.00
11.00
10.00
9.00
8.00
7.00
199806 199809 199812 199903 199906 199909 199912 200003 200006 200009 200012 200103 200106 200109 200112 200203
Date
RAVG
RGLS
RCT
RGM
10
Results –
The Effect of Reg FD on the Cost of Capital
Table 4 The Effect of Regulation Fair Disclosure on the Cost of Equity Capital
Predicted Sign
Full Sample
Small Firms
Medium Firms
Large Firms
Intercept
?
TranFD
?
9.723***
(35.76)
0.414***
(8.68)
11.263***
(15.01)
0.629***
(4.54)
8.097***
(12.56)
0.438***
(5.57)
8.676***
(17.51)
0.488***
(8.71)
PostFD
–
-0.780***
(-12.55)
-0.012
(-0.07)
-0.871***
(-8.89)
-0.667***
(-8.04)
MKT
+
SMB
+
HML
+
LogMV
–
LogBM
+
MMT
–
Lev
+
Ferr
+
Growth
+
LogDisp
–
Less2
+
IndCoC
+
0.216***
(3.04)
-0.046
(-0.88)
0.231***
(5.93)
-0.467***
(-17.77)
0.645***
(10.23)
-2.695***
(-44.42)
1.129***
(10.64)
0.191***
(17.65)
10.038***
(23.43)
-2.862***
(-15.62)
0.606***
(5.55)
0.309***
(17.28)
0.502***
(3.31)
0.024
(0.21)
0.011
(0.12)
-0.988***
(-7.52)
0.047
(0.32)
-3.382***
(-20.16)
0.777***
(3.97)
0.189***
(12.50)
11.612***
(19.34)
-2.437***
(-7.12)
0.297**
(2.06)
0.323***
(6.62)
0.352***
(3.65)
0.059
(0.86)
0.141**
(2.51)
-0.284***
(-2.69)
0.479***
(4.21)
-2.541***
(-31.34)
1.355***
(8.56)
0.156***
(10.00)
8.894***
(14.49)
-2.805***
(-11.31)
0.213
(1.30)
0.335***
(12.90)
0.501***
(3.93)
0.237***
(2.80)
0.088
(1.38)
-0.238***
(-4.83)
0.736***
(9.48)
-2.263***
(-26.28)
1.393***
(7.74)
0.147***
(5.89)
1.427
(1.38)
-3.187***
(-10.47)
-0.233
(-0.66)
0.318***
(12.58)
0.502
38,461
0.533
7,827
0.385
14,549
0.330
16,085
Adj.R2
N
11
Results –
Cross-Sectional Variation in the Reduction in the Cost of Capital
Table 5 Cross-Sectional Variation in the Effect of Regulation Fair Disclosure on the Cost of Equity Capital
R AVGit     1TranFDit   2 PostFDit   1  MKTit   2  SMBit   3  HMLit   4 LogMVit
  5 LogBM it   6 MMT it   7 Lev it   8 Ferrit   9 Growthit   10 LogDisp it
(1)
  11 Less3it   12 IndCoC it 1   it
Partition Variable
Book-to-Market
Low
(t-stat)
-1.179***
(-6.87)
R&D Expenditure
-0.311***
(-3.99)
Q2
(t-stat)
-0.562***
(-5.52)
Q3
(t-stat)
-0.065
(-0.52)
Q4
(t-stat)
-0.008
(-0.05)
High
(t-stat)
0.201
(1.01)
High-Low
(2)
1.380***
(27.82)
-1.285***
(-12.75)
-0.974***
(58.71)
Institutional Holdings
0.322
(0.95)
-0.199
(-1.19)
-0.723***
(-5.70)
-0.988***
(-7.24)
-0.956***
(-10.10)
-1.278***
(13.30)
Transient Institutional Holdings
0.438
(1.31)
-0.086
(-0.54)
-0.599***
(-4.67)
-0.898***
(-8.64)
-1.168***
(-8.00)
-1.606***
(19.53)
Non-transient Institutional Holdings
0.363
(1.03)
-0.372**
(-2.30)
-0.836***
(-4.64)
-1.122***
(-9.51)
-0.787***
(-8.69)
-1.150***
(10.02)
12
Results – ADRs and Foreign Firms
Table 7 Changes in the Cost of Capital for American Depository Receipts and Foreign Firms (ADRs)
R AVGit     1TranFDit   2 PostFDit   1  MKTit   2  SMBit   3  HMLit   4 LogMVit
  5 LogBM it   6 MMT it   7 Lev it   8 Ferrit   9 Growthit   10 LogDisp it
(1)
  11 Less3it   12 IndCoC it 1   it
Panel A: Changes in the Cost of Capital for ADRs in the Post-FD Period Relative to the Pre-FD Period
Variable
All ADRs
Small ADRs
Large ADRs
PostFD
0.310
0.328
0.38
(1.40)
(0.80)
(1.55)
Adj. R2
0.470
0.412
0.376
N
4,065
1,542
2,380
Panel B: Cross-Sectional Variation in Changes in the Cost of Capital for ADRs
High-Low
Low
High
(t-stat)
(t-stat)
Partition Variable
(2)
Book-to-Market
0.163
0.821**
0.658
(0.65)
(1.94)
(1.81)
R&D Expenditure
0.822**
(2.43)
-0.041
(-0.16)
-0.863**
(4.03)
Institutional Holdings
0.911**
(2.45)
-0.144
(-0.58)
-1.055**
(5.61)
Transient Institutional Holdings
1.127***
(3.01)
-0.174
(-0.68)
-1.301***
(8.30)
Non-transient Institutional Holdings
0.996**
(2.38)
-0.194
(-0.97)
-1.190***
(6.65)
13
Summary of Sensitivity Checks
 Repeat tests in Table 4 and Table 5
Additional control for analysts’ forecasts properties
 Excluding observations followed by less than three analysts.
 Add annual GNP, unemployment rate, 9/11 dummy to control for
confounding events.
 Excluding NASDAQ firms.
 Control for percentage of buy recommendations.
 Use risk premiums as the dependent variables.
 Different sample periods
 Use average value of dependent and independent variables in
regression
 Use constant sample
 Results are qualitatively similar.

14
Conclusions and Contributions
 The results support that Reg FD has lowered CoC as the SEC believed
The cost of capital is decreased in the post-FD period relative to
the pre-FD period on average for the full sample (a broad crosssection of firms).
 The reduction in CoC post Reg FD is significant for medium and
large firms but is insignificant for small firms.
 The reduction in CoC is more pronounced for firms with more
severe selective disclosure before Reg FD.
 We contribute to the literature on the likely effects of Reg FD
 We also contribute to the literature on the role of legal institutions and
securities regulations in affecting CoC.
 We corroborate Bhattacharya and Daouk (2002) – CoC is lower for
firms in countries where insider trading laws are enforced.
 We complement Hail and Leuz (2006) – CoC is lower for firms in
countries with more strict and extensive disclosure requirements.

15